&EPA
United States
Environmental Protection
Agency
Field Demonstration of Innovative
Condition Assessment
Technologies for Water Mains:
Leak Detection and Location
Office of Research and Development
National Risk Management Research Laboratory -Water Supply and Water Resources Division
-------
FIELD DEMONSTRATION OF INNOVATIVE CONDITION ASSESSMENT TECHNOLOGIES
FOR WATER MAINS: LEAK DETECTION AND LOCATION
by
Bruce Nestleroth, Stephanie Flamberg, Wendy Condit, and John Matthews
Battelle
Lili Wang and Abraham Chen
Alsa Tech, LLC
Contract No. EP-C-05-057
Task Order No. 0062
for
Michael D. Royer
Task Order Manager
Water Supply and Water Resources Division
National Risk Management Research Laboratory
2890 Woodbridge Avenue (MS-104)
Edison, NJ 08837
National Risk Management Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
March 2012
-------
DISCLAIMER
The U.S. Environmental Protection Agency, through its Office of Research and Development, funded and
managed, or partially funded and collaborated in, the research described herein under Task Order (TO)
0062 of Contract No. EP-C-05-057 to Battelle. It has been subjected to the Agency's peer and
administrative review and has been approved for publication. Any opinions expressed in this report are
those of the author (s) and do not necessarily reflect the views of the Agency, therefore, no official
endorsement should be inferred. Any mention of trade names or commercial products does not constitute
endorsement or recommendation for use. The quality of secondary data referenced in this document was
not independently evaluated by EPA and Battelle.
-------
ABSTRACT
Three leak detection/location technologies were demonstrated on a 76-year-old, 2,057-ft-long portion of a
cement-lined, 24-in. cast iron water main in Louisville, KY. This activity was part of a series of field
demonstrations of innovative leak detection/location and condition assessment technologies sponsored by
the U.S. Environmental Protection Agency (EPA). The main goal of the demonstrations was to acquire a
snapshot of the current performance capability and cost of these innovative technologies under real-world
pipeline conditions so that technology developers, technology vendors, research-support organizations,
and the user community can make more informed decisions about the strengths, weaknesses, and need for
further advancement of these technologies.
Leak detection was one part of a comprehensive water pipeline condition assessment demonstration
where six inspection companies operated 12 technologies that were at various stages of development and
provided different types and levels of leak and/or structural condition data. Technologies were included
for wall-thickness screening (i.e., average wall loss over many tens of feet), for detailed mapping of wall
thickness, and for leak detection. Both in-line and external inspection technologies were demonstrated.
The inspection technologies used visual, mechanical, acoustic, ultrasonic, and electromagnetic methods
for acquiring leak and pipe condition data. The inspection results for each technology were compared to
the leak rates or dimensions of introduced and naturally occurring anomalies, as well as their location
along the pipeline.
This report presents the results from three leak detection technologies: Pressure Pipe Inspection
Company's (PPIC's) Sahara®, Pure's SmartBall™, and Echologics' LeakfinderRT. Simulated leaks using
calibrated orifices in combination with natural leaks that already existed in the test pipe were used to
evaluate the performance of each leak detection system. The natural leaks were used to assess detection
and location capabilities, while the calibrated orifices were used to evaluate the leak rate assessment
capabilities for each technology. The combination of natural leaks and simulated leaks provided an
assessment of the capability of each leak detection system to detect, locate, and prioritize leak rates. Each
company provided a written report on the location and general size of natural leaks detected in the test
pipe, as well as leak rate estimates for the simulated leaks. Additional results from the acoustic pipe wall
assessment, internal inspection and external inspection technologies will also be made available in a
companion report.
in
-------
ACKNOWLEDGMENTS
The authors wish to express their sincere appreciation for the excellent cooperation and support provided
by Andy Williams, Keith Coombs, Dennis Pike, and other Louisville Water Company (LWC) colleagues,
as well as the LWC contractor, MAC Construction and Excavating, Inc. We would also like to
acknowledge the following technology vendors for their participation and in-kind support to this
demonstration:
The Pressure Pipe Inspection Company
Pure Technologies, Ltd.
Echologics Engineering Inc.
Lastly, we would like to acknowledge the technical reviewers for providing valuable review comments on
the final report:
James Thomson, Jason Consultants Group
Frank Blaha, Water Research Foundation
Dave Hughes, American Water Company
George Kunkel, Philadelphia Water Department
Mike Royer, EPA/ORD/NRMRL/WSWRD/UWMB
Ari Selvakumar, EPA/ORD/NRMRL/WSWRD/UWMB
Y. Jeffrey Yang, EPA/ORD/NRMRL/WSWRD/WQMB
Marc Bracken, Echologics Engineering Inc.
Stewart Day, Pure Technologies, Ltd.
Xiangjie Kong, The Pressure Pipe Inspection Company.
IV
-------
EXECUTIVE SUMMARY
The state of the art in condition assessment technologies for water mains is still developing and water
utilities are interested in third-party, independent sources of information on the capabilities of innovative
inspection technologies. Technology demonstrations with a range of real-life defects and conditions are
particularly valuable to water utilities and can play a vital role in accelerating the adoption of appropriate,
innovative condition assessment technologies. A field demonstration program was conducted to evaluate
condition assessment technologies applicable to the inspection of cast iron water mains. It is critical that
utilities have the capability to undertake reliable condition assessment of cast iron pipelines in order to
prevent failures and/or premature rehabilitation or replacement.
The main goal of the demonstration program was to acquire a snapshot of the current performance
capability and cost of applicable inspection technologies under real-world pipeline conditions so that
technology developers, technology vendors, research organizations, and the user community can make
more informed decisions about the strengths, weaknesses, and need for further advancement of these
technologies. As part of this research effort, several emerging and innovative inspection technologies
were demonstrated on a 76-year-old, 2,057-ft-long portion of a cement-lined, 24-in. cast iron water main
in Louisville, KY. This report presents the results from three leak detection technologies including the
Pressure Pipe Inspection Company's (PPIC's) Sahara®, Pure's SmartBall™, and Echologics'
LeakfinderRT. A companion report discusses the results of the acoustic pipe wall assessment and
internal/external inspection demonstrations.
The three leak detection technologies were demonstrated to assess their capabilities to detect, locate, and
size leaks on a straight, cement-lined, 24-in. cast iron water main. The test pipe had a burial depth
between 3.5 and 6.0 ft and wall thicknesses ranging from 0.68 to 0.78-in., as measured periodically during
routine maintenance activities. The test pipe historically operated at pressures between 45 and 50 pounds
per square inch (psi), while transmitting 4 to 6 million gallons per day (MOD) of flow. Under the
Louisville Water Company's (LWC) Main Replacement and Rehabilitation Program, a portion of 24-in.
diameter cast iron transmission water main along Westport Road was scheduled for replacement. LWC
agreed to make this portion of the pipe available for field demonstration, as well as provide necessary on-
site assistance.
Many aspects of the leak detection technologies were observed and documented over the course of the
demonstration. This included documenting the logistical and operational requirements encountered
during the demonstration, which are summarized in the report including the number of technicians
needed, any need for operator intervention, the number and spacing of pipe contact points, access
requirements, and more. This information will help utilities to gauge the feasibility of using these
technologies at their site. Sahara® and SmartBall™ require internal pipe access, but are non-disruptive in
nature and can be performed while the pipeline is in service. LeakfinderRT does not require internal pipe
access, is non-disruptive, and can be performed on a live main with or without flow. While each
technology used some form of acoustic listening device, the implementations were quite different:
• PPIC's Sahara® mounted a hydrophone sensor at the end of a cable tether. The hydrophone,
which was inserted and pulled through the pipeline using the water flow, provided real-time
assessment of leaks. The hydrophone sensor was tracked by an operator from ground level.
• The Pure SmartBall™ sensor and data-recording device were placed within a foam ball. The
sensor and ball were inserted in the pipeline and propelled by the water through the pipeline
to a downstream extraction point where a net inserted into the pipe caught and removed the
unit.
-------
• Echologics LeakfinderRT demonstrated two types of sensors. Pairs of accelerometers were
mounted on the outside of the pipe at discrete locations to detect and locate unknown leaks.
Then pairs of hydrophones in contact with the water at discrete locations were used to
estimate simulated leak rates.
A combination of simulated leaks and natural leaks were used to provide an assessment of the capability
of each leak detection system to detect, locate, and quantify leaks.
Simulated leaks were chosen to assess the leak sizing capability of each technology. By using artificial
leaks, changes in leak rate could be made by changing one variable (e.g., the leak diameter) enabling the
direct correlation of reported leak rates to actual leak rates. The restriction ports used for the simulated
leaks ranged in size from 0.25-in. down to a 0.02-in. diameter hole. The simulated leak rates ranged from
0.06 to 8.2 gpm. The test-pipe pressure was monitored and recorded during the demonstration and used
in combination with leak calibration curves to estimate the actual leak rate. The results for each
technology were compared with the pre-determined leak rates under each test condition to evaluate the
vendor reported leak rates.
Sahara® and SmartBall™ were able to detect the smallest single, simulated leaks, which were 0.06 gpm.
LeakfinderRT was only capable of locating leaks with a flow rate greater than 0.6 gpm. None of
technologies could discern two separate leaks in close proximity (less than 2.7 ft apart) for any of the
simulated leak clusters and appear only to report the larger leak rate for the cluster of leaks. The ability to
identify whether a signal is from an isolated leak or multiple leaks in close proximity is helpful in judging
the general condition of a pipeline. However, it is also important to accurately identify the location and
size of the largest leaks for repair purposes. Leaks within a foot or two of each other will likely both be
excavated if they are large enough to merit remediation. For this demonstration, the technology was
considered to be successful if it detected at least one leak in a cluster of leaks because of their close
proximity (with spacing from 0.6 to 2.7 ft).
For the simulated leaks, Sahara® , SmartBall™, and LeakfinderRT were able to find all of the leak
clusters and estimate the approximate magnitude of the largest leaks for over half of the leak clusters.
The results are summarized as follows:
• Sahara® reported 11 of 19 total simulated leaks, but reported 11 of 11 leak clusters. Within
each leak cluster, Sahara® accurately characterized the leak range for 6 of the 11 clusters. For
leak sizes that were not accurately characterized, four were off by one size category, while
one leak was off by two size categories as defined by the vendor.
• SmartBall™ reported 11 of 19 total simulated leaks, but reported 11 of 11 leak clusters.
Within each leak cluster, SmartBall™1 accurately characterized the leak range for 7 of the 11
clusters. All four of the leaks not accurately characterized were off by one size category as
defined by the vendor.
• LeakfinderRT reported 11 of 19 total simulated leaks and reported 11 of 11 leak clusters.
LeakfinderRT accurately characterized the leak range for 8 of the 11 clusters. Three of these
eight leak rates were reported as "negligible," meaning either close to, or less than, the 0.6
gpm detection threshold defined during calibration. The three leak rate ranges not accurately
characterized differed from the simulated leak rates by approximately one to four gpm.
The natural leaks were used to assess detection and location capabilities. Eight potential leak locations
identified by the technology vendors were excavated and examined. During the excavation, the soil was
examined for excessive moisture and erosion. When each leak site was fully uncovered, visual
VI
-------
assessment was used to determine whether the leak was from a bell-and-spigot joint or the body of the
pipe at an anomaly such as corrosion or a crack. After the potential leak sites were uncovered, the pipe
was pressurized to qualitatively assess the leak sizes by examining the amount of water leaching/spraying
from the pipe. During this process, EPA's contractor was able to definitively confirm naturally occurring
leaks in four of the eight locations that were excavated; all occurred at the bell-and-spigot joint. The
other four excavation locations could not be examined under pressure because of field conditions, but soil
moisture in the vicinity of the reported leak was visually assessed during pipe removal as a possible
indication that the test pipe had been leaking.
For the natural leaks, Sahara®, SmartBall™, and LeakfinderRT reported 6, 12, and 3 natural leaks,
respectively. Each was able to detect the two largest natural leaks. SmartBall™ reported the most natural
leaks, many of which were categorized by them as small (approximately 0.1 to 5.5 gpm), but not all were
verified due to time and budget constraints. The results are summarized as follows:
• Sahara® reported six natural leaks in real time. Except for one very small leak at 1,696 ft that
was not excavated and therefore could not be verified, the remaining five leaks were directly
(leak pinpointed) or indirectly (wet soil in the general vicinity) verified based on visual
evidence. However, Sahara® initially missed a small leak at one location. After the leak was
verified by EPA's contractor and reported to the vendors, PPIC performed additional post-
processing and subsequently reported that they were able to detect this leak.
• For the 12 natural leaks reported by SmartBall™, six were excavated and directly or
indirectly verified based on visual evidence; two other reported leaks were excavated, but the
existence of small leaks were not conclusive. The remaining four locations were not
excavated.
• LeakfinderRT reported the largest verified natural leak at 341.5 ft and another two leaks near
1,912 ft and 1,930 ft, which were also found by Sahara® and SmartBall™1. However, it failed
to identify natural leaks at bell-and-spigot joints near 53 ft, 195 ft, 556 ft and 640 ft, which
were confirmed to exist. It is not clear whether LeakfinderRT would have found these leaks
had the larger leaks been repaired and their noise signatures removed.
The cost of inspection is dependent on a number of variables including the length and diameter of pipe to
be inspected, pipe accessibility, and number of services requested (some vendors offer multi-service
discounts). Based on vendor quotes for inspecting 10,000 ft of 24-in. diameter cast iron pipe along the
same route as the demonstration site in Louisville, KY, the cost for a leak detection survey ranges from $2
to $5/ft. The three leak detection platforms that were demonstrated can also be used for pipe wall
thickness screening surveys; the cost for both leak detection and pipe wall thickness survey ranges from
$2.7 to $9/ft. Cost savings can be achieved when combining the leak detection with pipe wall thickness
survey to reduce time, labor, and equipment costs for inspection.
The inspection costs presented above do not include the cost for the water utilities to prepare the line and
provide traffic control and other logistical support. This site preparation cost for line modification and
field support is highly site-specific. It will depend upon regional costs for construction labor, along with
factors such as the access requirements, availability and condition of existing hydrants/valves, length of
deployment, days on site, and more. Based on typical construction costs (RSMeans, 2011), it is estimated
that the site preparation costs for a leak detection inspection of 10,000 ft, 24-in. diameter cast iron pipe
may range in magnitude from $0.12/ft (for traffic control only with use of existing taps) to $0.43/ft
(including traffic control, pit excavation, tapping, backfill, and surface restoration).
vn
-------
CONTENTS
DISCLAIMER ii
ABSTRACT iii
ACKNOWLEDGMENTS iv
EXECUTIVE SUMMARY v
APPENDICES ix
FIGURES ix
TABLES x
ABBREVIATIONS AND ACRONYMS xi
1.0: INTRODUCTION 1
1.1 Background 1
1.2 Project Objectives 2
1.3 Organization of Reports 3
2.0: SUMMARY AND CONCLUSIONS 4
3.0: MATERIALS AND METHODS 13
3.1 Site Description 13
3.1.1 Site Location 13
3.1.2 Test Pipe Condition 13
3.1.3 Leak History 13
3.2 Technology/Vendor Selection 17
3.3 Technology Description 18
3.3.1 PPIC Sahara® Leak Detection 18
3.3.2 Pure Technologies SmartBall™ 19
3.3.3 Echologics LeakfinderRT 20
3.4 Site/Test Preparation 22
3.4.1 Access Requirements 24
3.4.2 Safety, Logistics, Excavation, and Tapping 24
3.4.3 Simulated Leaks 34
3.5 Test Configuration 37
3.5.1 PPIC Sahara® Leak Detection 37
3.5.2 Pure Technologies SmartBall™ 39
3.5.3 Echologics LeakFinderRT 42
3.6 Post-Demonstration Leak Confirmation 44
4.0: RESULTS AND DISCUSSION 49
4.1 PPIC Sahara® Systems 49
4.1.1 Summary of Results 50
4.1.2 Leak Evaluation 51
4.2 Pure Technologies SmartBall™ 53
4.2.1 Summary of Results 53
4.2.2 Leak Evaluation 54
4.3 Echologics LeakfinderRT 56
4.3.1 Summary of Results 56
4.3.2 Leak Evaluation 57
4.4 Cost of Leak Detection/Location 59
4.4.1 Leak Detection/Location Services 59
Vlll
-------
4.4.2 Site Preparation 63
5.0: REFERENCES 65
APPENDICES
Appendix A: THE PRESSURE PIPE INSPECTION COMPANY SAHARA® LEAK DETECTION
REPORT
Appendix B: PURE SMARTBALL™ LEAK DETECTION REPORT
Appendix C: ECHOLOGICS LEAKFINDER LEAK DETECTION REPORT
Appendix D: INSPECTION VENDORS COMMENTS TO FINAL REPORT
FIGURES
Figure 3 -1. Location Map of Westport Road Transmission Main Replacement Proj ect 15
Figure 3-2. Locations and Details of Pipe and Joint Breaks and Leaks 16
Figure 3-3. Pipe Break along Westport Road Adjacent to Test Area in August 2008 17
Figure 3-4. Sahara® Inspection System 19
Figure 3-5. Aluminum Case and Foam Housing for SmartBall™ Acoustic Acquisition Device,
Data Storage, and Power Supply 20
Figure 3-6. SmartBall™ Insertion and Extraction Tubes 21
Figure 3-7. Echologics Proprietary Hydrophone Technology 21
Figure 3-8. Wireless Transmitter 23
Figure 3-9. Example Cross-Correlation Plot with Spike Indicating a Leak 23
Figure 3-10. Construction Trailer for Equipment Storage and Work Space 26
Figure 3-11. Location of Pits for Demonstration 27
Figure 3-12. Location of Pit 1 -Near Chenoweth Lane 29
Figure 3-13. Location of Pit 2 -Near St. Matthews Ave 29
Figure 3-14. Approximate Location of Pit 3 - Near Ridgeway Ave 30
Figure 3-15. 1-in. Corporation Valve with 3/t-in Threaded Plug with Leak Orifice 30
Figure 3-16. Tap Locations in Pit 4 32
Figure 3-17. Tap Locations in Pit 2 32
Figure 3-18. Tap Locations in Pit 5 33
Figure 3-19. Test Pipe Discharge to Storm Sewer Configuration 33
Figure 3-20. Monitoring Pipeline Pressure 36
Figure 3-21. Sump Pump System for Simulated Leak Locations 36
Figure 3-22. Example Position Profile of the SmartBall™1 vs. Time of Day 41
Figure 3-23. Example Velocity Profile of the SmartBall™1 vs. Time of Day 41
Figure 3-24. Example SmartBall™1 Receiver Tracking Points vs. Time of Day 41
Figure 3-25. Example Leak Calibration Curve Used to Size Leaks 42
Figure 3-26. Calibration Curves for Restriction Ports 44
Figure 3-27. Leak ID L7 - Small Leak at Bell-and-Spigot Joint 48
-------
TABLES
Table 2-1. Comparison Data for the Logistical and Operational Variables Associated with the
Field Demonstration 5
Table 2-2. Summary of Simulated Leak Detection Results by PPIC Sahara®, Pure SmartBall™,
and Echologics LeakfmderRT 6
Table 2-3. Summary of Natural Leak Detection Results by PPIC Sahara®, Pure SmartBall™, and
Echologics LeakfmderRT 10
Table 3-1. Summary of Historical, Operational, and Environmental Characteristics of Test Pipe 14
Table 3-2. Summary of Test Pipe Access Requirements for LWC Demonstration 25
Table 3-3. Summary of Access Pits - Description and Purpose 28
Table 3-4. Summary of Corp Valve Locations and Orientations 31
Table 3-5. Orifices Used to Simulate Various Leak Sizes During the Demonstration 35
Table 3-6. Daily Activities for Each Leak Detection Technology Vendor 38
Table 3-7. SmartBall™ Receiver (SBR) Locations 41
Table 3-8. Leak Test Matrix for PPIC Sahara® and Pure SmartBall™ 45
Table 3-9. Leak Test Matrix for Echologics LeakfmderRT 46
Table 3-10. Natural Leak Verification Results 47
Table 4-1. Sahara® Leak Classification Table for 24-in. to 60-in. Diameter Pipe 50
Table 4-2. Natural Leaks Detected by Sahara® Leak Detection 50
Table 4-3. Simulated Leaks Detected by Sahara®Leak Detection 51
Table 4-4. Evaluation of Natural Leaks Detected by Sahara® Leak Detection 52
Table 4-5. Evaluation of Simulated Leaks Detected by Sahara® Leak Detection 52
Table 4-6. SmartBall™ Leak Classification Table 53
Table 4-7. Natural Leaks Detected by SmartBall™ 53
Table 4-8. Simulated Leaks Detected by SmartBall™ 54
Table 4-9. Evaluation of Natural Leaks Detected by SmartBall™ 55
Table 4-10. Evaluation of Simulated Leaks Detected by SmartBall™ 55
Table 4-11. Hydrophone-to-Hydrophone Distances for Detection of Simulated Leaks 56
Table 4-12. Accelerometer-to-Accelerometer Distances for Detection of Natural Leaks 56
Table 4-13. Natural Leaks Detected by LeakfmderRT with Accelerometers 56
Table 4-14. Simulated Leaks Detected by LeakfmderRT with Hydrophones 57
Table 4-15. Evaluation of Natural Leaks Detected by LeakfmderRT with Accelerometers 58
Table 4-16. Evaluation of Simulated Leaks Detected by LeakfmderRT with Hydrophones 58
Table 4-17. PPIC Sahara® Cost Estimates for Inspection of a 24-in. Diameter, 10,000 ft Long Cast
Iron Pipeline 60
Table 4-18. Pure SmartBall™1 Cost Estimates for Inspection of a 24-in. Diameter, 10,000 ft Long
Cast Iron Pipeline 60
Table 4-19. Echologics LeakfmderRT Cost Estimates for Inspection of a 24-in. Diameter, 10,000 ft
Long Cast Iron Pipeline 62
Table 4-20. Estimated Site Preparation Costs for SmartBall™ Inspection of 10,000 ft pipe 63
Table 4-21. Estimated Site Preparation Costs for Sahara® Inspection of 10,000 ft pipe 64
-------
ABBREVIATIONS AND ACRONYMS
ANSI American National Standards Institute
EPA United States Environmental Protection Agency
gpm gallons per minute
GPS global positioning system
LWC Louisville Water Company
MOD million gallons per day
MJ mechanical joint
MRRP Main Replacement and Rehabilitation Program
NDT non-destructive testing
NPT National Pipe Thread
NRC National Research Council
NRMRL National Risk Management Research Laboratory
PPIC Pressure Pipe Inspection Company
psi pounds per square inch
PVC polyvinyl chloride
QA/QC quality assurance/quality control
QAPP Quality Assurance Project Plan
RF radio frequency
SBR SmartBall™ receiver
SOTR State of the Technology Review
TO Task Order
WERF Water Environment Research Foundation
XI
-------
1.0: INTRODUCTION
Three leak detection/location technologies were demonstrated on a 76-year-old, 2,057-ft-long portion of a
cement-lined, 24-in. cast iron water main in Louisville, KY. This activity was part of a series of field
demonstrations of innovative leak detection/location and condition assessment technologies sponsored by
the U.S. Environmental Protection Agency (EPA) from July through September 2009. The main goal of
the demonstrations was to acquire a snapshot of the current performance capability and cost of these
innovative technologies under real-world pipeline conditions so that technology developers, technology
vendors, research-support organizations, and the user community can make more informed decisions
about the strengths, weaknesses, and need for further advancement of these technologies.
Leak detection was one part of a comprehensive water pipeline condition assessment demonstration
where six inspection companies operated 12 technologies that were at various stages of development and
provided different types and levels of leak and/or structural condition data. Technologies were included
for wall-thickness screening (i.e., average wall loss over many tens of feet), for detailed mapping of wall
thickness, and for leak detection. Both in-line and external inspection technologies were demonstrated.
The inspection technologies used visual, mechanical, acoustic, ultrasonic, and electromagnetic methods
for acquiring leak and pipe condition data. The inspection results for each technology were compared to
the leak rates or dimensions of introduced and naturally occurring anomalies, as well as their location
along the pipeline.
This report presents the results from three leak detection technologies, i.e., Pressure Pipe Inspection
Company's (PPIC's) Sahara®, Pure's SmartBall™, and Echologics' LeakfinderRT1. For all of the
technologies evaluated in this demonstration, the leak detection methods are the most mature and
therefore are somewhat less novel in their approach than the newer condition assessment techniques (e.g.,
pipe wall metal loss). However, each of the leak detection/location technology platforms presented in this
report is also being explored as a screening tool for detection of pipe wall metal loss. These technology
implementations were also demonstrated, but are discussed in a separate report.
1.1 Background
To gain a better understanding of the available technologies for condition assessment of water mains, a
Technology Forum was held on September 9 and 10, 2008, in Edison, NJ under Task Order (TO) 62. The
Forum indicated that the state of the art in condition assessment technologies is still developing and that
water utilities could benefit from third-party, independent sources of information on the capabilities of
innovative inspection technologies. Technology demonstrations on real systems are particularly valued
by water utilities and can play a vital role in accelerating the adoption of appropriate, innovative condition
assessment technologies. A range of real-life defects and conditions should be present when undertaking
these types of demonstrations to maximize the benefit to utilities.
After participating in the Forum, the Louisville Water Company (LWC) offered an approximately 2,500-
ft-long, 24-in. diameter, cement-lined, cast iron pipe for field demonstrations of water main inspection
technologies. LWC treats 135 million gallons per day (MOD) of water and transmits water to 270,000
service taps through 3,500 miles of water main ranging from 1 to 60-in. in diameter. Under its Main
Replacement and Rehabilitation Program (MRRP), the company annually replaces over 35 miles of water
mains to maintain the water transmission system. As part of this program, the 2,500-ft portion of 24-in.
diameter cast iron transmission water main along Westport Road was scheduled for replacement in
1 Since the time of the demonstration in July 2009, the Pressure Pipe Inspection Company (PPIC) was acquired by
Pure Technologies Ltd., and Echologics Engineering, Inc. was acquired by Mueller Water Products.
-------
September 2009. LWC agreed to make this portion of the pipe available for field demonstration, as well
as provide necessary on-site assistance.
The field demonstration occurred between July 6 and September 4, 2009. This field demonstration
program presented an opportunity to (1) apply inspection technologies under nearly normal operating
conditions; (2) thoroughly compare measurements of parameters via non-destructive testing (NDT) with
direct measurements of those same parameters; and (3) remove sections of the pipe for comparative
testing at a later date with other technologies.
Cast iron pipe is the oldest and largest part of the water network. It is critical that utilities have the
capability to undertake reliable condition assessment of cast iron pipes to prevent failures and premature
rehabilitation or replacement. Innovative technologies are available for condition assessment of cast iron
mains, but only limited third-party performance and cost data are available, which inhibits their effective
consideration by the user community.
The suite of technologies considered for demonstration was based on a state of the technology review
report prepared under TO 62 on inspection technologies of water mains for ferrous pipes (Thomson and
Wang, 2009) and Forum input. Consistent with the focus of the state of the technology review and the
Forum, only leak detection/location and structural condition assessment technologies for ferrous pipes
were considered for the field demonstrations. Six vendors providing 12 different technologies including
leak detection/location and condition assessment technologies (both internal and external) agreed to
participate in the field demonstration program with substantial in-kind support.
The EPA contractor, in coordination with the leak detection technology vendors and the LWC, was
responsible for the planning, coordination, oversight, and execution of this field demonstration project.
The major tasks associated with the field demonstration project are described below:
• Task 6.1: Pre-Demonstration Activities. Pre-demonstration activities included planning and
coordination of project activities among EPA, LWC, and participating technology vendors;
preparation of a Quality Assurance Project Plan (QAPP); development of test protocols (with
vendor input); and communication of project schedules and testing requirements to all project
participants.
• Task 6.2: Field Demonstration. EPA's contractor coordinated with the participating vendors
and LWC for all on-site demonstration activities, communicated safety requirements,
planned/adjusted test schedules, monitored test progress, and documented field observations.
In performing the field demonstration, the technical and quality assurance/quality control
(QA/QC) procedures were followed as specified in the EPA-endorsed QAPP.
• Task 6.3: Post-Demonstration Evaluation and Reporting. This task included the
preparation of technical reports and photo documentation in order to summarize the results of
the field demonstration.
1.2 Project Objectives
The main goal of the demonstrations was to acquire a snapshot of the current performance capability and
cost of these innovative technologies under real-world pipeline conditions so that technology developers,
technology vendors, research-support organizations, and the user community can make more informed
decisions about the strengths, weaknesses, and need for further advancement of these technologies.
The ultimate desired outcome from these demonstrations is to detect problems in large diameter, cast iron
water mains prior to their failure, as well as to reduce premature replacement of sound buried water
-------
infrastructure. These outcomes are expected to arise from improved decision making regarding location,
time, and types of water main inspection, maintenance, and renewal activities. Improved asset
management decision making is expected to occur due to expanded and accelerated acceptance and use of
effective condition assessment devices, systems, and procedures and better decisions regarding
development and use of innovative condition assessment devices, systems, and procedures. This field
demonstration program directly supports the goals of EPA's Sustainable Water Infrastructure Initiative
and National Risk Management Research Laboratory's (NRMRL's) Aging Water Infrastructure Research
Program.
1.3 Organization of Reports
This report is divided into four main sections that include introductory material (Section 1.0), summary
and conclusions from the results of the field demonstration (Section 2.0), description of the materials and
methods used to manage the field demonstration (Section 3.0), and discussion of results provided by each
technology vendor (Section 4.0). This report covers leak detection and location. A companion report
discusses the results of the acoustic pipe wall assessment and internal/external inspection demonstrations.
-------
2.0: SUMMARY AND CONCLUSIONS
The PPIC Sahara®, Pure SmartBall™, and Echologics LeakfinderRT leak detection, location, and sizing
technologies were demonstrated on a 76-year-old, 2,057-ft-long portion of a straight, cement-lined, 24-in.
diameter cast iron water main in Louisville, KY. The test pipe had a burial depth between 3.5 and 6.0 ft
and wall thicknesses ranging from 0.68 to 0.78-in., as measured periodically during routine maintenance
activities. The test pipe typically operated at pressures between 45 and 50 pounds per square inch (psi)
while transmitting 4 to 6 MGD of flow. A closed-circuit television video inspection of the entire test pipe
indicated that the cement liner was uniform and no through-wall anomalies were detected in the pipe wall.
While each technology used some form of acoustic listening device, the implementations were quite
different:
• PPIC's Sahara® mounted a hydrophone sensor at the end of a cable tether. The hydrophone,
which was inserted and pulled through the pipeline using the water flow, provided real-time
assessment of leaks. The hydrophone sensor was also tracked by an operator from ground
level and leaks were marked on the pavement.
• The Pure SmartBall™1 sensor and data-recording device were placed within a foam ball. The
sensor and ball were inserted in the pipeline and propelled by the water through the pipeline
to a downstream extraction point where a net inserted into the pipe caught and removed the
unit.
• Echologics LeakfinderRT demonstrated two types of sensors. Pairs of accelerometers were
mounted on the outside of the pipe at discrete locations to detect and locate unknown leaks.
Then pairs of hydrophones in contact with the water at discrete locations were used to
estimate simulated leak rates.
All leak detection systems on the market have the ability to listen for leaks; however, quantifying the leak
rate is not as straightforward or as broadly applied. The leak detection technologies selected for the
demonstration are all developing innovative, proprietary methods for not only detecting the leak, but also
interpreting the acoustic signals to quantify the leak rate.
Quantifying the leak rate in the pipe's natural condition is challenging because the acoustic signal
generated by the leak can be greatly affected by the leak size and geometry, internal pipeline pressure,
and backpressure created by soil or water outside the pipe. Nature rarely provides a sufficient range of
conditions that span the performance parameters to be tested. Furthermore, because excavation would
disturb the natural conditions, accurate verification of the leak rate cannot be performed for natural leaks.
Therefore, for practical reasons, artificial leaks were chosen to assess the leak sizing capability of each
technology. By using artificial leaks, changes in leak rate could be made by changing one variable, the
leak diameter, enabling the direct correlation of reported leak rates to actual leak rates. Quantification of
artificial leak rates is just one measure of the performance of each leak detection technology. Detection
of natural leaks still provides valuable information on a technology's capability to detect and locate leaks
as well as its ability to qualitatively assess the leak rate.
Many aspects of the technologies were observed over the course of the demonstration. Table 2-1
provides comparison data for the logistical and operational variables encountered during the
demonstration. On-site preliminary reports were provided by all, some instantaneously, with the longest
delivery time being the next morning. Preliminary reports were requested within 1 week and final reports
within 5 weeks of the demonstration. These vendor reports are an important source of data presented in
this summary report and are provided in Appendix A (Sahara®), Appendix B (SmartBall™), and
-------
Appendix C (LeakfmderRT). Users of this report can refer to these Appendices to review the original
format and organization of the inspection data as issued by the individual vendors.
Since this demonstration was a snapshot in time, new developments may have taken place since
completion of the demonstration. Therefore, the findings in this report may not be wholly representative
of the current operational capabilities of the demonstrated technologies. For this reason, the vendors were
asked to provide formal comments on the final leak detection report to highlight advancements since
completion of the demonstration and/or clarification on what was reported. These comment letters are
contained in Appendix D.
Table 2-1. Comparison Data for the Logistical and Operational Variables
Associated with the Field Demonstration
Logistical/Operational
Variables
Equipment logistics
Requires internal pipe
access?
Utility preparation
Number of technicians
needed for operation
Pipe contact points
Sterilization
Real-time data display
Leak positioning
Onsite report
Quick-look report
Final report(a)
Operator intervention
PPIC Sahara®
Dedicated truck
Yes
Requires one access point
and a controlled flow rate
2-3
One; Supplied equipment
for LWC could inspect up
to 2,500 ft; Sahara® has a
6,000 ft maximum cable
length.
Yes
Yes
Leak position directly
known; technician marks
leaks at ground level
Verbal as leaks were found
10 days after left site
12 weeks after left site
When leak detected,
operator moved device
back and forth to locate
leak
Pure SmartBall™
Overnight shipping
company
Yes
Requires two access
points and a controlled
flow rate. Large offtakes
on the pipe must be
closed
2
Two; Distance depends
on flow rate
Yes
No
Post analysis used to
locate leaks
Verbal, overnight
7 days after left site
1 week after left site
No operator tasks after
ball is launched until it is
received
Echologics LeakfinderRT
Operator transported two
cases
No (accelerometers)
Yes (hydrophones)
Requires two access points,
but can be accomplished
with hydrants or common
pipeline appurtenances
1
Two per test; Every 1,000
ft for leak rate; Every 300-
400 ft for location/detection
and condition assessment (b)
Yes
Yes
Leak position determined
analytically
Written within 1 hour of
test completion
19 days after left site
7 weeks after left site
Manual setting of filters;
automatic detection and
location
(a) PPIC provided one combined report for leak detection and condition assessment technologies; Pure provided an
individual report for their leak detection technology and one for their pipe wall assessment technology;
Echologics provided one combined report for leak detection and wall thickness assessment technologies.
(b) The sensor spacing for locating and detecting leaks was shorter than what is typically used by Echologics so
that they could collect data for leak location/detection and pipe condition at the same time. The 300-400 ft
spacing was needed to demonstrate their condition assessment technology, which is not part of this report; see
Appendix D for additional information.
Simulated leaks using calibrated orifices in combination with natural leaks that already existed in the test
pipe were used to evaluate the performance of each leak detection system. The natural leaks were used to
-------
assess detection and location capabilities, while the calibrated orifices were used to evaluate the leak rate
assessment capabilities for each technology. The combination of natural leaks and simulated leaks
provided an assessment of the capability of each leak detection system to detect, locate, and quantify
leaks.
Simulated Leaks
A summary of the simulated leak results are provided in Table 2-2. The results for each technology were
compared with the pre-determined leak rates under each test condition to evaluate the accuracy of the
vendor reported leak rates and locations. The simulated leak rates ranged from 0.06 to 8.2 gpm.
Table 2-2. Summary of Simulated Leak Detection Results by PPIC Sahara®, Pure
SmartBall™, and Echologics LeakfinderRT2
Demo
No.
1
2
o
6
4
Pit
ro(a)
Pit 4
Pit 2
Pit 4
Pit 2
Pit5
Pit 4
Pit 2
Pit5
Pit 4
Pit 2
Pit5
Corp
Valve
ID
CV1
CVS
CV2
CV4
CV6
CV7
CVS
CV1
CV2
CVS
CV4
CV6
CV7
CVS
CV1
CV2
CVS
CV7
CVS
Dist. (ft)
577.4
1,082.2
578.4
1,082.8
1,084.9
1,583.0
1,585.7
577.4
578.4
,082.2
,082.8
,084.9
,583.0
,585.7
577.4
578.4
1,084.1
1,583.0
1,585.7
Simulated
Leak
Rate
(gpm)(e)
0.59
8.2
0.06
0.15
0.59
1.1
8.2
1.0
2.0
0.14
0.57
4.6
7.9
0.57
0.06
4.6
1.0
7.9(d)
2.0
Sahara®
Leak Rate
(gpm)(b)
Very Small
(0 to 1.8)
Small
(1.8 to 18)
Very Small
(0 to 1.8)
Very Small
(0 to 1.8)
Large
(75 to 128)
Small
(1.8 to 18)
Small
(1.8 to 18)
Medium
(18 to 75)
Very Small
(0 to 1.8)
Small
(1.8 to 18)
Medium
(18 to 75)
Dist.
(ft)(c)
Pit 4
Pit 2
Pit 4
Pit 2
Pit5
Pit 4
Pit 2
Pit5
Pit 4
Pit 2
Pit5
SmartBall™
Leak
Rate
(gpm)
(f)
0.57
8.0
0.3
2.8
15
1.8
7.2
30
4.5
0.1
40
Dist.
(ft)
579
1,080
579
1,080
1,580
579
1,080
1,580
579
1,080
1,580
LeakfinderRT
Leak
Rate
(gpm)
Negligible
8.0
Negligible
Negligible
5.0 to 8.0
2.0 to 5.0
5.0 to 8.0
5.0 to 8.0
Oto 1.0
2.5 to 5.0
Negligible
(d)
Dist.
(ft)
-
1,077.1
-
-
1,580.2
577.6
1,082.2
1,578.2
560.7
1,092.6
-
(a) See Figure 3-11 for the location of the simulated leak pits along the test pipe.
(b) Sahara® only reported qualitative leak rates. In subsequent correspondence Sahara® provided additional details on the
approximate quantitative leak rate which is presented in Table 4-1 of this report. The values in parentheses represent the
leak rate range for the specific leak classification defined by Sahara®.
(c) The general procedure for a Sahara® inspection is to track the device with an aboveground sensor and mark the leak location
on the pavement. For the demonstration, Sahara® reported the simulated leak location as a pit number marked on the
pavement rather than the actual distance.
(d) CV7 was closed for LeakfinderRT demo.
In Table 2-2 the text with a red background signifies leak rate estimates that are off by one or more categories as
defined by each individual vendor.
-------
(e) The leak rate values represent the average rate over the entire demonstration. The pipe pressure varied from 50 psi to 58 psi
which will slightly impact these leak rates.
(f) The leak size categories defined by SmartBall™ are small (0 to 2 gpm), medium (2 to 10 gpm) and large (>10 gpm). If the
predicted leak rate by SmartBall™ was in the same leak category as the actual leak rate they were given credit as accurately
sizing the leak cluster.
During each demonstration, the leaks were turned on or effusing the corp valve and the leak rate was
controlled by the size of the orifice installed in the corp valve. Simulated leaks were placed in three pits
(4, 2, and 5). Pit 4 was first, followed by Pits 2 and 5 approximately 500 ft and 1,000 ft downstream of
Pit 4, respectively. Within the pits there were two to four corp valve-orifice assemblies for simulating
leaks, and these assemblies were axially spaced along the pipeline anywhere from 0.6 to 2.7 ft apart.
In Demo 1, the simulated leak rate was provided to the vendors to use as a calibration point, while the
remaining demos were blind, in that the technology had to report both the location and the approximate
size of the leaks. LeakfinderRT was able to establish a detection limit of 0.6 gallons per minute (gpm) at
a sensor spacing of 1,081 ft in Demo 1. Detection limits were not determined for Sahara® and
SmartBall™ , but they were each able to detect the smallest simulated leaks, which had a rate of 0.06
gpm. Each technology conducted four demos following the same test metrics. With extra time available,
additional demos were conducted upon vendor's request to gather additional data for technology
advancement. These additional tests are not part of the demonstration results evaluation - they were only
provided as a courtesy to the vendors.
In general, Sahara® and SmartBall™ were able to find the general location of all simulated leak clusters3.
Sahara® and SmartBall™ were able to detect the smallest single, simulated leaks, which were 0.06 gpm.
LeakfinderRT located and estimated flow rates for simulated leaks with a flow rate greater than 0.6 gpm;
leaks with flow rates <0.6 gpm were reported as "negligible."
• Sahara® reported 11 of 19 total simulated leaks, but reported 11 of 11 leak clusters. Within
each leak cluster, Sahara® accurately characterized the leak range for 6 of the 11 clusters. For
leak sizes that were not accurately characterized, 4 of 5 were off by one size category, while
the leak in Demo 2, Pit 5 was off by two size categories. The location accuracy could not be
evaluated as Sahara® only reported the pit number in which they found the leak.
• SmartBall™1 reported 11 of 19 total simulated leaks, but reported 11 of 11 leak clusters.
Within each leak cluster, SmartBall™1 accurately characterized the leak rate range for 7 of the
11 clusters. All four of the leak rates not accurately characterized were off by one size
category. The location accuracy could not be evaluated as SmartBall™1 only reported the
location of the pit and not where the actual leaks were located.
• LeakfinderRT reported 11 of 19 total simulated leaks and reported 11 of 11 leak clusters.
Within each leak cluster, LeakfinderRT accurately characterized the leak range for 8 of the 11
clusters. Three of these eight leak rates were reported as "negligible," meaning either close
to, or less than, the 0.6 gpm detectdion threshold defined by LeakfinderRT during calibration.
The other three leak rates not accurately characterized differed from the simulated leak rates
by approximately 1 to 4 gpm. The location accuracy was within 0 to 5 ft of the actual leak
location except for Demo 4 where the distances were off by a maximum of 17 ft.
3 A cluster of leaks is defined as 1 to 3 leaks within the same demonstration pit (less than 2.7 ft apart).
-------
None of the technologies was able to discern two separate leaks in close proximity (less than 2.7 ft apart)
for any of the simulated leak clusters and appear only to report the larger leak rate for the cluster of leaks.
The ability to identify whether a signal is from an isolated leak or multiple leaks in close proximity is
helpful in judging the general condition of a pipeline. However, it is also important to accurately identify
the location and size of the largest leaks for repair purposes. Leaks within a foot or two of each other will
likely both be excavated if they are large enough to merit remediation. For this demonstration, we
considered the technology to be successful if it detected at least one leak in a cluster of leaks because of
their close proximity and the likelihood that the larger leaks masked detection of the smaller leaks.
Sometimes the process calls for making the repair of the leak found (usually the larger) and then returning
to sound again in hopes that no leaks are missed. This was not done as part of the test protocol, but could
be an area for future research to improve leak discrimination capabilities.
Natural Leaks
Over a dozen naturally occurring leaks were reported by the leak detection technologies and were marked
as LI to L13 on the surface for excavation to verify the leaks (see Table 2-3). Eight of the 13 locations
were excavated and examined between August 19 and 24, 2009 for indications of a leak. Upon
observation of wet soils at L3, L4, L6, and L7, the pipe was pressurized to visually verify the leak
locations and qualitatively estimate the leak rates by examining the amount of water leaching/spraying
from the pipe. Leaks L4, L6, and L7 all occurred at the bell-and-spigot joints; leak L3 appeared to
originate from a bell-and-spigot joint, but could not be directly pinpointed. It should be noted that none
of the reported natural leaks occurred within the regions where the simulated leaks were generated.
The other four excavation locations, i.e., LI, L10, LI 1, and L12, could not be examined under pressure
because the water supply valve broke in the closed position and could not be repaired within the time and
budget constraints. As an alternative, soil moisture in the vicinity of the reported leak was visually
assessed during pipe removal as a possible indication that the test pipe had been leaking. For LI and L12,
the soil was definitely wetter in the area excavated; however, the actual leak location could not be
identified. For L10 and LI 1, the soil was relatively dry, but could not be used to exclude the possible
existence of a very small leak.
• Sahara® reported six natural leaks in real time. Except for one very small leak at 1,696 ft
which was not excavated and therefore could not be verified, the remaining five leaks were
directly (leak pinpointed) or indirectly (wet soil in the general vicinity) verified based on
visual evidence. However, Sahara® initially missed a small leak at L6. After the leak was
verified by EPA's contractor and reported to the vendors, PPIC performed additional post-
processing and subsequently reported that they were able to detect this leak.
• For the 12 natural leaks reported by SmartBall™, six were excavated and directly or
indirectly verified based on visual evidence; two other reported leaks (L10 and LI 1) were
excavated but the existence of small leaks were not conclusive. The remaining four locations
were not excavated due to time and budget constraints and therefore could not be verified.
• LeakfinderRT reported the largest verified natural leak at L4 (341.5 ft) and another two leaks
near 1,912 ft (L12) and 1,930 ft (L12), which were also found by Sahara® and SmartBall™.
However, it failed to identify natural leaks at bell-and-spigot joints near 53 ft, 195 ft, 556 ft
and 640 ft (i.e., LI, L3, L6 and L7), which were confirmed to exist. It is not clear whether
LeakfinderRT would have found these leaks had the larger leaks been repaired and their noise
signatures removed.
Additional measurements were taken for the one natural leak found in Pit L to understand what may have
caused the leak. The joint rotation was measured — that is, the angle between the two pipe joints where
-------
the leak was found. Researchers at the National Research Council (NRC) of Canada suggested measuring
the joint rotation to test a hypothesis that an angle of more than 2° would cause a joint leak and that an
angle of more than 5° would crack the bell. While the test pipe was generally level, this leak occurred
near a storm sewer. The angle measured was approximately 1.5°, indicating that the bell would not be
cracked (as was verified visually) or leaking due to joint rotation. Therefore, the large leak was most
likely due to degradation in the leadite seal for the bell-and-spigot joint rather than joint rotation.
-------
Table 2-3. Summary of Natural Leak Detection Results by PPIC Sahara , Pure SmartBall , and Echologics LeakfinderRT
Location
ID
LI
L2
L3
L4
L5
L6
L7
L8
L9
L13
L10
Lll
L12
PPIC Sahara®
ro#
1
Distance
(ft)
50
Description (a)
Very small leak
(0-1.8gpm)
Did not report leak
2
3
194
338
Very small leak
(0-1.8gpm)
Large leak
(75-128 gpm)
Did not report leak
Initially did not report leak, but found
after verification results were provided
to PPIC. The leak was masked by an
artificial leak at 578 ft.
4
638
Small leak
(1.8-18 gpm)
Did not report leak
Did not report leak
5
1,696
Very small leak
(0-1. 8 gpm)
Did not report leak
Did not report leak
6
1,906
Small leak
(1.8-1 8 gpm)
Pure SmartBall 1M
ro#
i
2
3
4
5
6
7
8
9
Distance
(ft)
53
125
199
341
414
556
641
966
1,210
Description
Small leak
(-0.1 5 gpm)
Small leak
(-0.1 gpm)
Small leak
(-0.8 gpm)
Medium leak
(-15 gpm)
Small leak
(-0.2 gpm)
Small leak
(-1.0 gpm)
Small leak
(-2.0 gpm)
Small leak
(-0.1 gpm)
Small leak
(-1.0 gpm)
Did not report leak
10
11
12
1,724
1,809
1,930
Small leak
(-1.5 gpm)
Small leak
(-2.0 gpm)
Small leak
(-5.5 gpm)
Echologics LeakfinderRT
Distance
ID# (ft)
Description
Did not report leak
Did not report leak
Did not report leak
2a 341.5
-2.5-5.0 gpm
Did not report leak
Did not report leak
Did not report leak
Did not report leak
Did not report leak
Did not report leak
Did not report leak
Did not report leak
-7 1,912
7C 1,930
-1.0-2.5 gpm
-1.0-2.5 gpm
Visual Verification
Verification attempted; soil was wet, but
there was a nearby storm sewer at 52 ft; leak
not pinpointed, but elevated moisture
indicative of leak.
No verification attempted
Verification attempted; water in pit near
bell-and-spigot joint at —195 ft; did not
pinpoint leak
Verification attempted; leak at bell-and-
spigot joint -339 ft
No verification attempted
Verification attempted; leak at bell-and-
spigot joint at -556 ft
Verification attempted; leak at bell-and-
spigot joint -640 ft
No verification attempted
(at intersection with St. Matthews)
No verification attempted
No verification attempted
Verification attempted, but no wet soil was
found at —1 ,724 ft; inconclusive
Verification attempted, but no wet soil was
found at —1,809 ft; inconclusive6
Verification attempted; soil was moist at
-1,906 ft or 1,930 ft; leak not pinpointed,
but elevated moisture indicative of leak.
(a) Sahara only reported qualitative leak rates. In subsequent correspondence, Sahara provided additional details on the approximate quantitative leak rate,
which is presented in Table 4-1 of this report. The values in parentheses represent the leak range for the specific leak classification.
4 Sahara® also found a large air pocket at 900-ft; this finding is not included in the results as detection of air pockets was not requested or verified.
5 In Table 2-3, the gray background signifies natural leaks that were verified in the field (by pinpointing leak or elevated moisture), the yellow background
signifies leaks that we attempted to verify but could not find any evidence of a leak; the red background signifies false negatives, and the white background
signifies leaks found by one or more leak detection technologies, but were not verified in the field.
6 Inconclusive because the pipe could not be pressurized at the time of excavation (see Section 3.6).
-------
Acoustic Interference
In their most basic form, leak detection technologies work by detecting and analyzing acoustic signals.
Unfortunately, there are numerous other causes of acoustic signals that can create problems when
interpreting the data for leaks. Pipeline flow conditions, road conditions and traffic, and construction are
a few of the sources of unwanted acoustic signals or noise that can affect the results. The propagation of
these numerous noise sources in the pipeline is a function of the pipeline geometry. For this
demonstration, factors that could have affected the acoustic signal from a leak include:
• Water was diverted to a sanitary sewer though a 12-in. line to create flow during the
demonstrations. While the noise was clearly audible, it is not known how much of this noise
could be detected in the main pipe.
• One lane of traffic was blocked for the demonstration, but through traffic was still permitted
in the other lane.
• Air pockets were observed in the pipeline even though efforts were made to eliminate air
from the line.
• Train tracks paralleled the pipe with several trains passing by the demonstration site each day.
• Excavation equipment was being used to lay new water pipe within a mile of the
demonstration site.
In addition, the acoustic signal produced by larger leaks could mask the sound generated by the smaller
leaks located nearby. In this demonstration, two or more leaks were often in close proximity, which can
influence a technology's ability to distinguish separate leaks.
Demonstration Summary
PPIC, Pure, and Echologics demonstrated their technologies to detect, locate, and size leaks on a straight,
cement-lined, 24-in. cast iron water main operating nominally at 50 to 58 psig. The capabilities of each
technology were demonstrated, with many aspects of the technologies observed over the course of the
demonstration. For the simulated leaks, Sahara® and SmartBall™ were able to find all of the leak clusters
and estimate the approximate magnitude of the largest leaks for over half of the leak clusters.
LeakfmderRT was also able to find all of the simulated leak clusters, but four leak rates were designated
as negligible, i.e., less than the detection limit of 0.6 gpm determined during calibration. However, none
of the technologies were able to discriminate individual smaller leaks within a leak cluster when the
spacing between leaks was anywhere from 0.6 to 2.7 ft. For the natural leaks, Sahara®, SmartBall™, and
LeakfmderRT reported 6, 12, and 3 natural leaks, respectively. Each was able to detect the two largest
natural leaks. SmartBall™ reported the most natural leaks, many of which were categorized by them as
small (approximately 0.1 to 5.5 gpm), but not all were verified due to time and budget constraints.
Each vendor uses different terminology for leak sizes, which makes it difficult to directly compare the
results. The demonstration program was set-up to simulate what utilities consider typical leak rates that
might warrant inspection yet not so large that they would be found without the need for inspection.
Currently, there are no industry guidelines that define what is considered a small, medium, or large leak.
Each technology has its own advantages and limitations, providing utilities with options to choose one
that best fits their needs and expectations. Sahara® and SmartBall™1 require internal pipe access, but are
non-disruptive in nature and can be performed while the pipeline is in service. LeakfmderRT does not
require internal pipe access, is non-disruptive, and can be performed on a live main with or without flow
(LeakfmderRT was operated under no-flow conditions for this demonstration).
11
-------
The cost of inspection is dependent on a number of variables including the length and diameter of pipe to
be inspected, pipe accessibility, and number of services requested (some vendors offer multi-service
discounts). Based on vendor quotes for inspecting 10,000 ft of 24-in. diameter cast iron pipe along the
same route as the demonstration site in Louisville, KY, the cost for a leak detection survey ranges from $2
to $5/ft. The three leak detection platforms that were demonstrated can also be used for pipe wall
thickness screening surveys; the cost for both leak detection and pipe wall thickness survey ranges from
$2.7 to $9/ft. Cost savings can be achieved when combining the leak detection with pipe wall thickness
survey to reduce time, labor, and equipment costs for inspection.
The inspection costs presented above do not include the cost for the water utilities to prepare the line and
provide traffic control and other logistical support. This site preparation cost for line modification and
field support is highly site-specific. It will depend upon regional costs for construction labor, along with
factors such as the access requirements, availability and condition of existing hydrants/valves, length of
deployment, days on site, and more. Based on typical construction costs (RSMeans, 2011), it is estimated
that the site preparation costs for a leak detection inspection of 10,000 ft, 24-in. diameter cast iron pipe
may range in magnitude from $0.12/ft (for traffic control only with use of existing taps) to $0.43/ft
(including traffic control, pit excavation, tapping, backfill, and surface restoration).
12
-------
3.0: MATERIALS AND METHODS
3.1 Site Description
3.1.1 Site Location. Louisville is located in the north-central portion of Kentucky, immediately
south of Indiana along the Ohio River. Its climate can be described as humid sub-tropical with yearly
temperatures ranging from 0°C in January to 25°C in July. The city's estimated population, as of 2006,
was just fewer than 600,000; the Louisville Metropolitan Area's population was approximately
1,250,000. Supplied by the Ohio River, the source water is treated and transmitted to service taps by
LWC, which was granted a charter from the Kentucky Legislature in 1854. Under this charter, water was
first provided to the citizens of Louisville by LWC in 1860. Currently, LWC treats and transmits 135
MGD of water to 270,000 service taps through 3,500 miles of water mains, ranging in diameter from 1 to
60-in. Under its MRRP, the company replaces over 35 miles of pipe every year as either a preventive or
reactionary effort to maintain the water transmission and distribution system.
As part of LWC's pipe replacement and rehabilitation program, a 2,500-ft length (2,057-ft were used for
the leak detection technologies) of 24-in. diameter pipe that was scheduled for replacement was made
available for the demonstrations of inspection and condition assessment technologies. The pipeline right-
of-way is in the north lane of Westport Road, from the intersection of Westport Road and Chenoweth
Lane, to the intersection of Ridgeway Avenue and Westport Road (see Figure 3-1). At Ridgeway
Avenue, the 24-in. diameter line goes under a set of CSX railroad tracks.
3.1.2 Test Pipe Condition. The portion of the 24-in. diameter transmission main along Westport
Road between Chenoweth Lane and Ridgeway Avenue was made available for the field demonstration
project (referred to herein as "the test pipe"). The test pipe is Class 150 deLavaud spun cast iron that is
lined with a factory-installed cement mortar and represents approximately 2,500 ft of transmission line.
The test pipe was installed in September 1933 and had a burial depth between 3.5 and 6.0 ft. Wall
thicknesses of the pipe range from 0.68 to 0.73-in., as measured periodically during routine maintenance
and inspections or during repairs. During a site visit in May 2009, wall thicknesses of pipe samples
removed during the installation of a 24-in. by 12-in. tee were measured and ranged from 0.76 to 0.78-in.
The test pipe typically operates at pressures between 45 and 50 pounds per square inch (psi) while
transmitting 4 to 6 MGD of flow. Table 3-1 summarizes the historical, operational, and environmental
characteristics of the test pipe.
In preparation for the new installation and prior to the demonstration, all taps and offtakes on the 24-in.
diameter test pipe were removed. The test pipe was bypassed and taken offline, but could be filled or
drained as needed for each demonstration. During the demonstration, traffic was restricted to a single
lane and traffic flow was sporadic. The amount of traffic during a demonstration was not separately
measured or recorded.
3.1.3 Leak History. Seven joint leaks and one pipe break have been reported along the test pipe
from May 1973 to August 2008; however, no information exists regarding the test pipe leak history prior
to 1973. Figure 3-2 shows the location and date of the recorded leaks and breaks: two near the
intersection of Ridgeway Avenue and Westport Road on May 22, 1973 and March 2, 1977; three near the
intersection of St. Matthews Avenue and Westport Road on December 14, 1995, August 23, 2001, and
February 17, 2002; and two near the intersection of Sherrin Avenue and Westport Road on November 18,
1985 and December 27, 2003. All of the seven joint leaks occurred at leadite joints.
Since no evidence of wall loss was noted at the time of the repairs, most of these joint leaks are assumed
to have been induced by settling/consolidation of underlying fill material or natural soils or as a result of
13
-------
the freeze/thaw cycle causing differential movement of pipe segments attached to the common joint. The
exception to this was the December 14, 1995 joint leak at the intersection of St. Matthews Avenue and
Westport Road in which evidence of corrosion was observed.
Table 3-1. Summary of Historical, Operational, and Environmental Characteristics of Test Pipe
Historical
Pipe Material
Installation Date
Pipe Segment Length (ft)
Pipe Inner Diameter (in.)
Pipe Class
Pipe Thickness (in.)
Approximate Total Pipe Length (ft)
Burial Depth (ft)
Pipe Lining
Pipe Lining Thickness (in.)
External Coating
Type of Joints
Land Use over Main
Leak History (recorded)
Date of First Joint Break (recorded)
Date of First Pipe Break (recorded)
Cast iron
09/1933
12
24
deLavaud Spun Cast; Cement lined; Class 150
0.68-0.78
2,000
3.5-6.0
Factory Applied Cement Mortar
Variable, on the order of 0.25
Bitumen paint
Leadite
Residential traffic; bituminous paving
Eight leaks since 1973 (see Figure 3-2)
05/22/1973
08/29/2008 (not within 2,057-ft test pipe)
Operational
Typical Operating Flow (MOD)
Typical Operating Pressure (psi)
Water pH (S.U.)
4-6
• Flow throttled due to concerns of main breaks
• Available flow for inspection ranging from 1,400 to 2,800
gpm (or 1 to 2 ft/sec) due to sewer restrictions
45-50
8.2
Environmental
Soil Parameters (moisture, pH,
resistivity, redox potential, etc.)
Average Monthly Temperature (°C)
No historical data(a)
January through December: 0, 2, 8, 14, 19, 23, 25, 24, 21, 14,
8,3
Minimum - 0 (January)
Maximum - 25 (July)
(a) Soil characterization was performed during the demonstration project.
The only recorded pipe break occurred on August 29, 2008, approximately 12 ft north of the centerline of
Ridgeway Avenue and 40 ft east of the centerline of Westport Road. The break appears to have occurred
near a joint and propagated longitudinally along the pipe (see Figure 3-3), resulting in complete failure.
The pipe break was caused by an attempt to operate the line at its full capacity, which indicated that the
pipe might have lost part of its original structural integrity due to aging. It should be noted that the
location of the pipe break is outside the test area (that is, not strictly part of the test pipe). However, it is
noteworthy because of the nature of the break and because it occurred just a few days before the EPA
Forum, which prompted LWC to offer the pipe for this demonstration.
14
-------
Pit #2 (1080' from Pit #1; St. Matthews Ave.)
Corp Valve 3 4 5 & 6, machine defects
12 Discharge to
Sanitary Sewer
1750'from Pit #1
PitE
(1439'from Pit#1)
Pit #3 (2057' from Pit #1 near Ridgeway Ave.
One 24" x 12" tee (with adapters to 2" and 6")
Pit #4 581 'from Pit #1
Corp valve 1 & 2, machine defects
PitD
(1173* from Pit #1)
Pit#1 (nearChenoweth Ln.)
Pit #5 (1580'from Pit#1)
One 24 x 12 tee (with adapters to 2 and 6
Corp Valve 7 & 8. machine defects
PitC
(809'from Pit#1)
PitB
(510'from Pit#1)
PitL
(338'from Pit #1)
Machine defects
Pit A
(250'from Pit #1)
WESTPORT RD
— 24 Pipe
^— 30™ Pipe
I I 10' W x 16' L Trench Box
D 6' x 6' Sensor Pit
Fire Hydrant ° 125 25°
Discharge Point SCALE IN FEET
(P) Pressure Gauge
Chenoweth Ln
to
Ridgeway Ave
DATE: 10/22/08
Figure 3-1. Location Map of Westport Road Transmission Main Replacement Project
-------
Dale: August 29.2
Exacl Localion: +0 R. F. C/L Weapon Rd ami 12 ft. N C/L Ridgcwav Avc
L:ak Type: Pipe
Depth 10 Pipe: 6 ft,
bvidence of Corrosion: Yes
Dale: November 18, 1985
L:\ael Location: Wcslport Rd- and Shcrrin Avc.
Leak Type: Joint
lo Pino: 4 II.
El iili.-in.-i.- nf Cumaiun; No
port Rd and Rtdgcway Ave,
[tale: March 2. W77
Exact l.tviilion: Westptin RU- and Ridgewav Ave.
Look Type: Joint
to Piiv: -1 ft.
.\ idcncc of Corrosion: No
Dale: December 27, JOUj
Exact Localion: 6 fl. N C/L Wcslnon Rd and 4fi ft W OL Sherrin Ave
Leak Type: Joini
Depth to Fine: 3 ft.
Dale: February ! 7. 2002
bud Location: 9 ft. N. C/L Westpon Rd. and 139 ft. E. C/L St. Matthews Avc
Leak Tvpc Joint
[Icplli in Pipe: .1.5 ft.
F.videnti_-4>fComjskm: N
WESTPORT RD
Chenowoth Ln
to
Ridgeway Ave
Figure 3-2. Locations and Details of Pipe and Joint Breaks and Leaks
-------
3.2
Note: arrow pointing to longitudinal propagation of crack
Figure 3-3. Pipe Break along Westport Road Adjacent to Test Area in August 2008
Technology/Vendor Selection
The TO 62 State of the Technology Review (SOTR) report (Thomson and Wang, 2009) provides an
overview of the state of inspection technologies for ferrous pipes. The technologies selected for
demonstration at Louisville, KY were based on the TO 62 SOTR report, feedback from the Technology
Forum, and an additional literature search on relevant reports prepared by organizations such as Water
Research Foundation (formerly American Water Works Association Research Foundation), Water
Environment Research Foundation (WERF), and EPA, as well as vendors' web sites. A list of potential
candidate technologies was compiled, which included acoustic-, magnetic-, electromagnetic-, and
ultrasonic-based technologies. Technologies that require the removal of coatings and preparation of the
pipe surface (such as ultrasonic tools for wall thickness measurement) are well established and were not
considered in this field demonstration. Innovative and emerging ultrasonic tools can be demonstrated
offsite after the pipe is exhumed.
The candidate technologies were further screened based on (1) suitability of the technologies for the test
pipe diameter and material, (2) readiness of the technologies within the field demonstration timeline, and
(3) potential to yield useful data for interested utilities. It is also important that the technologies
considered not only represent those that are commercially available, but also those that are in the stage of
development that could be demonstrated in the field. An added benefit of this demonstration project is to
bring new technologies to the forefront of condition assessment research and allow utilities to become
familiar with these technologies.
After the technology screening, an e-mail transmittal was sent to prospective vendors in February 2009 to
solicit expression of interest. Most vendors responded promptly and expressed their keen interest in
participating in the demonstration. Several vendors were eliminated from further consideration due to
17
-------
either lack of interest or financial constraints. Six vendors agreed to participate in and provide partial in-
kind contributions to the field demonstration project.
3.3 Technology Description
Various methods are available for detecting and locating leaks within water mains. These methods
typically involve acoustic leak detection equipment that 'listens' for the noise created by water escaping a
pipe. These devices can include individual fixed listening devices (such as accelerometers) in contact
with the pipe, valves and/or hydrants, or mobile/fixed listening devices in the water column (such as
hydrophones).
Several factors affect the loudness and frequency range of sounds made by water main leaks, including
leak size and shape, pipeline pressure, backpressure at the leak, distance from the noise source or pipe
discharge, pipe material and diameter, soil type and compaction, and ground cover. The pipeline
diameter was the main factor that limited the number of technologies available for this demonstration.
While many fixed leak detection systems are available for small-diameter pipes, only one fixed system
and two mobile systems specified functionality for a 24-in. diameter pipeline.
Technology developers are also advancing leak detection technologies to quantify leak rates. The rate at
which water leaks from a pipeline depends on the leak size, internal pipeline pressure, and backpressure
created by soil or water outside the pipe. All of these factors affect the acoustic signal that is produced by
the leak and what is ultimately detected by the leak detection technologies. The leak detection
technologies selected for the demonstration are all developing innovative methods for not only detecting
the leak, but also interpreting the acoustic signals to quantify the leak rate. The following sections detail
the leak detection technologies that were demonstrated at LWC.
3.3.1 PPIC Sahara® Leak Detection. Sahara® is a platform for several inspection techniques that
were demonstrated, such as Sahara® Video, Sahara® Leak Detection, and Sahara® Condition Assessment.
The Sahara® Leak Detection system is a technology that identifies the location and estimates the
magnitude of leaks in large diameter, 12-in. and above, water transmission mains of all material types.
This technology works on the principle that leaks cause pressure differential, which makes noise that can
be detected. This passive form of leak detection/location uses a 1-in. diameter piezoelectric hydrophone
(frequency range of 1 Hz to 170 kHz) tethered to a calibrated umbilical cable (see Figure 3-4). The
system is then inserted into a live main through a standard 2-in. tap. The insertion mechanism combines a
glanding arrangement to form a seal around the cable, and a retractable guide to protect the cable from
damage as it passes into the pipe. A winch and cable drum control the deployment and retrieval of the
umbilical, which ensures that the leak sensor can be reliably and consistently removed from the pipeline.
Once inserted, a drogue (parachute) attached in front of the hydrophone captures water flow to control the
inspection speed and guide the tool through the pipeline. As the sensor head travels along the pipeline,
leak sounds within the pipe are identified and confirmed in real time by the hydrophone in the sensor head
and then transmitted through the cable to the processing equipment for interpretation.
During operation, an operator stands by at the controller station to control hydrophone deployment, listen
to the hydrophone signal for leaks, and visually monitor the signal using specialized spectrogram
software. The initial indication of a leak is audibly detected by the operator and then verified by the
spectrogram. Once a leak is detected, the hydrophone can pass over the leak multiple times to classify
and pinpoint the leak. At the same time, a second operator travels the pipeline above ground using a tool
to detect the exact location of the sensor. When a leak is detected this operator will make a mark on the
ground identifying the location and record a global positioning system (GPS) point for reference.
18
-------
Csblt Drum
(In Disinfectant Batti)
Locating Tool Signals
Processing Unit
Figure 3-4. Sahara® Inspection System (courtesy of PPIC)
The Sahara® system can provide information on the location and qualitative size estimates of leaks in real
time. An electronic processing unit with audio and visual output is used for data analysis. The distinctive
acoustic signal produced by a leak is recorded by the sensor and processed into a visual signal. The visual
signal is then analyzed, along with the audio signal, to quantify the leak. In no-flow situations, a second
tethering line (mule tape) can be used to pull the hydrophone through a pipeline.
The Sahara® survey distance is not only affected by the amount of available cable (usually 1.2 miles [2
km]), but also by factors like the flow velocity, the number and degree of pipeline bends, pipeline
diameter, and internal pipe conditions (e.g., butterfly valves). In 2009, the specifications provided for the
Sahara® system indicated allowable operating pressures from 7 to 230 psi and flow velocities from 1 to 5
ft/s. In 2010, after the field demonstration, PPIC reported that the Sahara® specification was for an
operating pressure range from 5 to 200 psi and a flow rate range from 1 to 12 ft/s (Appendix D).
Since the demonstration, PPIC has configured Sahara® to operate from 5 to 200 psi and 1 to 12 ft/s
(Appendix D). A typical pressure insertion requires a minimum water flow of 1 ft/s to propel the sensor
head and drag the drogue through the pipe. Pipeline pressures above 200 psi may cause deployment
difficulties. The Sahara® system is claimed to be capable of locating a leak to less than 1 meter in
pipelines that are less than 30-ft deep.
Calibration is performed by testing each hydrophone and comparing it to a standard frequency response.
The Sahara® hydrophone has sensitivity to leaks as small as 0.005 gpm (located in 72-in. PCCP at 87 psi).
Data are interpreted and analyzed in real time by on-screen spectrogram and audio listening. The Sahara®
hydrophone uses dual analysis methods to distinguish leaks from ambient noise. Factors such as low
water pressure, electrical noise, air pockets, and external ambient noise can all affect the real-time
analysis of the sensor signal. During the demonstration, some leaks were masked by external factors and
required post-analysis to detect them. Post analysis methods include filtering noise to improve leak
detection.
3.3.2
Pure Technologies SmartBall . SmartBall is an autonomous in-line system that uses
TM
acoustic technology to detect and locate leaks and gas pockets in a pipeline. SmartBall consists of two
primary components: an aluminum alloy core and lightweight foam outer shell. The core is 2.5-in. in
diameter and houses the acoustic acquisition device, tracking equipment, data storage equipment, and
power supply. The aluminum core is placed within the foam shell (see Figure 3-5) that can vary in
diameter depending on the size, operation, and configuration of the pipeline to be surveyed (usually less
than one third of the diameter of the pipe). The SmartBall™ system is claimed to be operable for up to 12
19
-------
hr and is applicable for pipes greater than 4-in. in diameter. While not tested in this demonstration,
significant flow in a path different from that heading towards the net can be a problem; offtakes with
sizes and flows should be reported to field personnel. The potential for the SmartBall™ to be lost exists
if the direction of flow suddenly changes or another activity (e.g., high customer use or hydrant flow)
diverts the sensor from the planned inspection path.
Figure 3-5. Aluminum Case and Foam Housing for SmartBall™ Acoustic
Acquisition Device, Data Storage, and Power Supply
During an inspection, SmartBall™ is inserted into the pipeline through hydrants and any valve
configuration with clearance 4-in. diameter or greater using a specialized insertion tube that can be bolted
to the appurtenance (see Figure 3-6). Once in the pipeline, the foam shell absorbs water to allow acoustic
activity to penetrate the foam for recording by the core. SmartBall™ has negative buoyancy in water
which allows the ball to settle on the bottom of the pipe and traverse the pipe by rolling with the flow. As
it travels through the pipe, sound from leaks is measured by hydrophone and position data magnetometers
feedback position data, which is stored in the ball. Sound pulses transmitted between the ball and
external devices (SmartBall™1 receiver [SBR]) on the pipe are used to determine the position of the ball.
Absolute position reference points obtained from the SBR are applied to time-stamped data to generate a
position-versus-time relationship for the entire length of inspection. After the survey, data are
downloaded from the ball and analyzed to generate the acoustic information relative to the distance the
ball had traveled. SmartBall™1 is then retrieved through another 4-in. diameter or greater valve using an
extraction tube bolted to the valve, which contains a specialized net that captures the SmartBall™1,
compresses the foam shell, and removes it from the pipeline (see Figure 3-6).
3.3.3 Echologics LeakfinderRT. Echologics' proprietary leak detection system, LeakfinderRT,
uses a patented, cross-correlation technology to passively listen for noise created by a leak in large
diameter metal and plastic pipes. The system hardware consists of leak sensors, a wireless signal
transmission system, and a personal computer. The system places sensors on two water system fittings ~
such as valves, hydrants, or exposed pipe— that are separated by some distance and that bracket the leak.
If a leak is present, the software then uses the difference in leak signal arrival times at the two sensors; the
distance between the two sensors; and, the sound velocity in the pipe to identify the leak location. Two
types of sensors were used:
• Echologics' proprietary hydrophones for direct measurement of the water column (see Figure
3-7)
• Echologics' piezoelectric accelerometers, with a sensitivity of 1 V/g.
20
-------
Each sensor has its own specific attributes that make it preferable in certain situations. Echologics reports
that the hydrophone is particularly well-suited to measuring asbestos cement and medium- to large-
diameter mains (12-in. and larger), since leaks on these pipes generally are dominated by lower-frequency
noise (200 Hz and below).
Figure 3-6. SmartBall Insertion and Extraction Tubes
Figure 3-7. Echologics Proprietary Hydrophone Technology
21
-------
The accelerometers are used to sense leak-induced vibration while hydrophones are used for sensing leak-
induced sound. The personal computer calculates the cross-correlation function of the two leak signals to
determine the time lag, TmaX) between the two sensors. Then the location of the leak can be derived from
Equation 3-1 below:
J — max QnH J — /~) J
^1 ~ 2 dna^2-^ ^i (Equation 3-1)
LI and L2 are the positions of the leak relative to sensors 1 and 2, respectively, c is the propagation
velocity of sound in the pipe. D is the distance between location 1 and 2. Propagation velocity needs to
be determined experimentally or is estimated based on the type and size of the pipe.
The LeakfmderRT enhanced cross-correlation function is calculated indirectly in the frequency domain
using the inverse Fourier transform of the cross-spectral density function rather than using the shift-and-
multiply method in the time domain (Hunaidi et al., 2004). The enhanced correlation function provides
improved resolution for narrow-band leak signals. This is very helpful for plastic pipes (low frequency
sound emission), small leaks, multiple leaks and situations with high background noise. Moreover, a
major advantage of the enhanced function is that it does not require the usual filtering of leak signals to
remove interfering noises (Hunaidi et al., 2004). The enhanced correlation function allows for improved
detection of leaks over traditional leak detection technologies.
For this demonstration, the hydrophone sensors for detecting and quantifying the simulated leaks were
placed about 1,000 ft apart, while the accelerometer sensors for detecting any naturally occurring leaks
were placed about 300 ft apart. The signals emitted by the sensors were detected by wireless transmitters
(460 MHz or 433 MHz analogue units manufactured by Echologics, as shown in Figure 3-8), which send
the signal to a computer to record the data. According to Echologics, the wireless transmitters should be
at least 1 ft above ground to eliminate radio frequency (RF) signal interference. The LeakfmderRT
software requires pipe material, pipe diameter, and sensor spacing as key input variables. Leak sounds
are recorded and correlated by LeakfmderRT for a period of time determined in the field. The cross-
correlation results are displayed on screen and continuously updated in real time while leak signals are
being recorded.
Any potential leaks will appear as a spike in the cross-correlation plot. The position of the spike on the x-
axis corresponds to the time difference it takes for the signal to arrive at the two sensor locations. The
position relative to either of the sensors can be computed using the wave velocity for the material under
inspection and time difference. Figure 3-9 provides an example of a cross-correlation plot and spike
indicating a leak. LeakfmderRT also incorporates an enhanced correlation function, which allows
narrow-band leak noise to have more well-defined peaks. This is important when multiple leaks occur,
when leak sensors are placed in close proximity to each other, and/or the pipe contains very small leaks.
3.4 Site/Test Preparation
Several activities were necessary prior to, during, and after the field demonstration to accommodate the
various technology vendors/visitors and to verify the inspection conditions (leaks). The following
sections detail specific measures taken in order to conduct the field demonstration.
22
-------
Figure 3-8. Wireless Transmitter
Correlation Function
-0.08 -0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08
Time (second)
Figure 3-9. Example Cross-Correlation Plot with Spike Indicating a Leak
(courtesy of Echologics)
To assess the performance of the leak detection systems, leaks of a known location and size were
required. As such, EPA's contractor developed a plan to create simulated leaks, using calibrated orifices
in combination with any natural leaks that might already exist in the test pipe. The natural leaks were
used for detection and location assessment; however, they could not be used for leak rate assessment,
because the leak rate could not be accurately determined once excavated. The combination of natural
leaks and simulated leaks enabled an assessment of the capability of each leak detection system to detect,
locate, and estimate the flow rate of leaks.
Although it was intended to keep the locations of the simulated leaks hidden, this could not be reasonably
accomplished during the demonstration. As discussed in more detail below, large excavations were
necessary to install leak taps for 1-in. corporation (commonly referred to as corps or corp valves) valves
that would contain the calibrated leak orifices. In addition, each excavation pit had to be open for access
during the demonstration to turn the simulated leaks 'on" or 'off." It was not possible to completely hide
the location of the simulated leaks because one of the leak detection technology vendors used
23
-------
aboveground tracking devices to mark the location of the leak during the inspection, while the other two
vendors routinely walked the length of the test pipe during the demonstration.
Prior to the actual demonstration, the condition of the test pipe was relatively unknown, aside from basic
pipeline location data and information obtained during previous leak investigations. In June, the valves at
both ends of the 2,057-ft test pipe were closed to evaluate if there was any significant pressure drop in the
system. This assessment showed that the line maintained a nominally constant pressure for a full day, so
it was quite possible that there were no large natural leaks in the test pipe.
3.4.1 Access Requirements. The internal leak detection/location inspection technologies required
only the installation of relatively small taps (2 to 4-in. in diameter) for insertion and extraction. For the
in-line inspection technology demonstration, a 12-in. diameter tap and gate valve with a mechanical joint
(MJ) fitting were installed at each end of the test pipe for insertion and retrieval of equipment (see
companion report on internal inspection tools). Reducers were used to match the leak detection
equipment requirements to the 12-in. MJ fitting for launching and receiving the leak detection
technologies. For the Sahara® and LeakfinderRT hydrophone technologies, a 12-in. MJ to 6-in. MJ
reducer and a 6-in. MJ cap with a 2-in. National Pipe Thread (NPT) tap were used. Echologics supplied
an additional 2-in. to 1.5-in. reducer for its equipment. SmartBall™ required either a 4-in. or 6-in.
American National Standards Institute (ANSI) flange for a gate valve to launch its equipment. To
achieve this setup, a 12-in. MJ to 6-in. MJ reducer and a 6-in. MJ to 6-in. ANSI flange were used because
this equipment could be easily provided by LWC. LWC supplied all pipe fittings for the demonstration.
Video inspection methods confirmed the pipe did not have any internal obstructions such as tuberculation
which may have impeded the application of internal inspection technologies.
Echologics' LeakfinderRT technology uses two sensor types: hydrophones that require direct contact with
the water column and accelerometers that are glued to the outside of the pipe. Therefore, two
demonstration configurations were needed: (1) direct access to the water in the pipe for placement of
hydrophones at approximately 1,000-ft intervals by the access method described in the previous
paragraph, and (2) direct access to the pipe exterior for placement of accelerometers at approximately
300-ft intervals. The 300-ft intervals were achieved through five large excavation sites and six smaller
excavated holes. A summary of all access requirements is provided in Table 3-2.
3.4.2 Safety, Logistics, Excavation, and Tapping
Safety and Logistics
During the demonstration, MAC Construction (LWC's contractor) was responsible for traffic rerouting
and control. All technology demonstrations occurred on weekdays during normal business hours. While
the demonstration was ongoing, portions of Westport Road were closed to through traffic, with some
access allowed for local businesses. At the end of each day, MAC Construction plated all open
excavations to help avoid accidents during the evenings and weekends and reopened both lanes of traffic
on Westport Road.
A construction trailer (see Figure 3-10) equipped with electrical power provided a work space for the
inspection technology vendors, as well as equipment storage during the demonstration. At least one EPA
contractor was onsite each day of the demonstration and coordinated the dissemination of safety and
contact information to the technology vendors and visitors. All logistical and operational questions were
handled by the EPA contractor in charge. The EPA contractor also coordinated daily activities with the
technology vendors, MAC Construction foreman, and LWC inspectors to ensure that the demonstration
ran efficiently and effectively.
24
-------
Several visitors, including representatives of the EPA and utility companies, came to the site during the
demonstration. Visitors were instructed to pre-register via e-mail and sign in with the EPA contractor at
the construction trailer before going onsite. Safety gear including hard hats, steel-toed shoes and safety
vests was required before visitors could gain access to the demonstration site.
Table 3-2. Summary of Test Pipe Access Requirements for LWC Demonstration
Vendor
The Pressure
Pipe
Inspection
Company
(PPIC)
Pure
Technologies
Echologics
Engineering
Type of
Inspection
Internal;
tethered
Internal
External
Technology/
Product
Sahara® Leak
Detection/Location
System
SmartBall1M Leak
and Gas Pocket
Location
LeakfinderRT
Flow Requirements/
Pipeline Constraints
Flow must be >1 ft/s for
single 2-in. diameter tap;
Mule tape is required in
no-flow situations or when
flow is insufficient.
At lower flows, the
parachute is unable to
overcome the drag of the
cable for a given distance.
Requires appurtenances
along pipeline to place
receivers
Flow range reported at
time of demonstration was
> -0.8 ft/s, but < -1.5 ft/s;
Note: Pure reports in
Appendix D inspections as
low as 0.5 ft/s and as high
as 7 ft/s.
Requires appurtenances
and/or pipe access to place
sensors
Requires air to be removed
from the line; no flow
requirements (tests
conducted without flow)
Pipe Access
Requirements
One per inspection
interval (every 2,500 ft for
LWC demonstration; up to
6,000 ft based on Sahara®
maximum cable length).
A 2-in-diameter (or larger)
tap with female NPT
thread reducer located at
upstream to the section to
be inspected;
-10 ft clearance to mount
insertion equipment.
Two per inspection
interval (at beginning and
end of inspection).
4-in. or 6-in. diameter
clear bore gate valve.
> 8 ft vertical clearance at
launch tap and > 12 ft
vertical clearance at
retrieval tap.
Both taps at 12 o'clock
position
Two per inspection
interval. Hydrophones
require direct contact with
the water. Accelerometers
require solid contact with
the pipe exterior.
Pipe access every 300 to
2,500 ft
25
-------
Figure 3-10. Construction Trailer for Equipment Storage and Work Space
Excavation
Five large excavations were provided for the leak detection technologies during the demonstration; these
included Pits 1 through 5 as shown in Figure 3-11 and described in Table 3-3. These sites were selected
based solely on location along the test pipe. Since the condition of the pipe was initially unknown, EPA's
contractor installed eight 1-in. taps in Pit 2 (4 taps), Pit 4 (2 taps), and Pit 5 (2 taps) to ensure that leaks
were available for calibration and inspection during the demonstration. An additional six small
excavations, identified as Pits A, B, C, D, E, and F in Figure 3-11, were used to demonstrate one leak
detection system and several other condition assessment technologies. Pictures of locations for Pits 1, 2,
and 3 are shown in Figures 3-12 through 3-14.
Tapping
Several taps were provided for the demonstration, either to facilitate operation of internal assessment
tools or to simulate leaks. Pits 1 and 3 each contained a 12-in. diameter tap to install a gate valve for
insertion and extraction of internal inspection tools. Several reducers for the 12-in. gate valve were
provided to launch PPIC's Sahara® technologies (12-in. x 2-in. reducer), Pure's SmartBall™ system (12-
in. x 6-in. reducer), and Echologics' LeakfmderRT (12-in. x 2-in. reducer).
Pits 2, 4, and 5 contained taps into which corp valves were installed to simulate pipeline leaks (see Figure
3-15). Corp valves are 1-in. diameter valves with a 3/t-in. internal threaded outlet port. Pit 2 contained
four corp valves (labeled CV3, CV4, CVS, and CV6), while Pit 4 and Pit 5 contained two corp valves
each (labeled CV1 and CV2; CV7 and CVS, respectively).
26
-------
Pit #2 (1080' from Pit #1; St. Matthews Ave.)
Corp Valve 3 4 5 & 6, machine defects
.. 12" Discharge to
Sanitary Sewer
~~
PitF
(1750'from Pit #1)
Pit #3 (2057' from Pit #1 near Ridgeway Ave.)
One 24" x 12" tee (with adapters to 2" and 6")
Pit #4 (581'from Pit #1)
(1439'from Pit #1)
Corp Valve 1 & 2, machine defects
(1173'from Pit #1)
Pit #1 (near Chenoweth Lnj
One 24" x 12" tee (with adapters to 2" and 6"
Pit #5 (1580'from Pit #1)
Corp Valve 7 & 8, machine defects
PitC
(809'from Pit #1)
PitB
(510'from Pit #1)
PitL
(338'from Pit #1)
Machine defects
WESTPORT RD
(250'from Pit #1)
24 Pipe
— 30" Pipe
I I 10'Wx 16'L Trench Box
D 6' x 6' Sensor Pit
• Fire Hydrant
Discharge Point
Pressure Gauge
Chenoweth Ln
to
Ridgeway Ave
0 125 250
^^^^^
SCALE IN FEET
DATE: 10/22/08
Figure 3-11. Location of Pits for Demonstration
-------
Table 3-3. Summary of Access Pits - Description and Purpose
Pit ID
Pitl
Pit 2
Pit3
Pit 4
Pit5
Pit A
PitB
PitC
PitD
PitE
PitF
Description
• Near Chenoweth Lane at location of first 24-
in. x 12-in. tee
• 8ft of pipe exposed
• Reference point - 0 ft
• Intersection of Westport Road and St.
Matthews Avenue
• -1,080 ft from first 24-in. x 12-in. tee in Pit
#1
• ~8 ft of pipe exposed; ~2-ft circumferential
clearance
• Near Ridgeway Ave. at location of second
24-in. x 12-in. tee
• -2,057 ft from first 24-in. x 12-in. tee
• 8ft of pipe exposed
• -581 ft from first 24-in. x 12-in. tee
• 3 ft of pipe exposed; top half only
• ~l,580-ft from first 24-in. x 12-in. tee
• 3 ft of pipe exposed; top half only
• -250 ft from first 24-in. x 12-in. tee in Pit #1
• -3 ft of pipe exposed; top portion only
• -5 10 ft from first 24-in. x 12-in. tee in Pit #1
• -3 ft of pipe exposed; top portion only
• -809 ft from first 24-in. x 12-in. tee in Pit #1
• -3 ft of pipe exposed; top portion only
• -1,173 ft from first 24-in. x 12-in. tee in Pit
#1
• -3 ft of pipe exposed; top portion only
• -1,439 ft from first 24-in. x 12-in. tee in Pit
#1
• -3 ft of pipe exposed; top portion only
• -1,750 ft from first 24-in. x 12-in. tee in Pit
#1
• -20 ft of pipe exposed; -2 -ft circumferential
clearance
Purpose
• Launch internal inspection technologies
• Install 12-in. service tap (May 2009); attach
12-in. x 2-in. and 12-in. x 6-in. reducers to
allow access for internal tools
• Install four 1-in. service taps for leak
simulations
• Install two calibration metal loss defects*
• Install nine additional metal loss defects for
condition assessment*
• Retrieve internal inspection technologies
• Install 12-in. service tap (May 2009); attach
12-in. x 2-in. and 12-in. x 6-in. reducers to
receive internal tools
• Install 12-in. tee to divert flow to
storm/sanitary sewer
• Install two, 1-in. service taps for leak
simulations
• Install pit-like metal-loss defects for condition
assessment*
• Install two, 1-in. service taps for leak
simulations
• Install pit-like metal-loss defects for condition
assessment*
• Small excavation for LeakfmderRT, keyhole
condition assessment technologies and soil
sampling*
• Small excavation for LeakfmderRT, keyhole
condition assessment technologies and soil
sampling*
• Small excavation for LeakfmderRT, keyhole
condition assessment technologies and soil
sampling*
• Small excavation for LeakfmderRT, keyhole
condition assessment technologies and soil
sampling*
• Small excavation for LeakfmderRT, keyhole
condition assessment technologies and soil
sampling*
• Small excavation for LeakfmderRT, keyhole
condition assessment technologies and soil
sampling; significant graphitization was found
when excavated*
• Install one large calibration defect (metal-loss
defect - 6 1/8 in long; 0.28 to 0.45 in depth)*
* These pits were created for demonstration of condition assessment technologies, but were also used to
demonstrate the external leak detection technology (LeakfmderRT).
28
-------
Figure 3-12. Location of Pit 1 - Near Chenoweth Lane
Figure 3-13. Location of Pit 2 - Near St. Matthews Ave.
29
-------
Figure 3-14. Approximate Location of Pit 3 - Near Ridgeway Ave.
Corporation
Valve
J
Threaded Plug
with Leak Orifice
Figure 3-15. 1-in. Corporation Valve with %-in Threaded Plug with Leak Orifice
30
-------
The positioning of the corp valves was based on access restrictions within each pit (location of bell-and-
spigot joints for longitudinal placement and trench boxes for clock position). The manner in which the
trench boxes were placed in Pit 2 and Pit 4 to avoid a parallel fiber optic communication conduit and
other buried utilities only allowed taps for the corp valves to be drilled on one side of the pipe. In
addition, the length of the arm on the mechanical tap did not allow for a majority of the taps to be
installed at 270°, which was originally planned. Therefore, the orientation of the corp valves was placed
as close to 270° as practically possible. In Pit 5, it was possible to install one tap on each side of the pipe
at orientations of 45° and 315° from the top of the pipe. The location and orientation of the corp valves in
each pit is provided in Table 3-4.
Table 3-4. Summary of Corp Valve Locations and Orientations
Corp Valve
ID
CV1
CV2
CVS
CV4
CVS
CV6
CV7
CVS
Pit ID
Pit 4
Pit 2
Pit5
Distance (ft)
577.4
578.4
1,082.2
1,082.8
1,084.1
1,084.9
1,583.0
1,585.7
Approx.
Orientation
(degrees)*
315
315
315
270
315
320
45
315
* Counter-clockwise from the direction of flow.
The corp valves in all three pits were used in conjunction with various %-in. restriction ports, with drilled
holes of various sizes inserted into the valve outlet to simulate a range of leak sizes during the
demonstration (see Section 3.4.3 for details). The corp valve locations for the leak simulations are shown
in Figures 3-16 through 3-18.
The two 12-in. taps for launching and receiving the internal inspection technologies were installed in May
2009. The 1-in. corp valves were installed in June 2009 prior to commencement of the field
demonstration. The restriction ports were also fabricated and tested several weeks prior to the field
demonstration.
Generating Flow
While some of the inspection technologies require flow to detect and locate leaks, the test pipe was no
longer supplying water to customers in anticipation of the pending replacement project. Therefore, to
create flow during the demonstration, water was supplied to the test pipe through a valve near Chenoweth
Lane connected to a 30-in. diameter line with a pumping station within a mile. At the end of the test pipe,
the flow was diverted to the sanitary sewer through a 12-in. gate valve and polyvinyl chloride (PVC) line
located downstream of Pit 3 (see Figure 3-19).
There were two drawbacks to this arrangement. First, the discharge was essentially a very large leak that
created noise during the demonstration and interfered with the acoustic sensors; the effects of which
became more pronounced as technologies neared the discharge point. Second, because the discharge was
diverted to the storm sewer, it could not be used immediately after heavy rainfall to prevent sewers from
overflowing. Rain delayed several of the demonstrations with a record rainfall of 6.5-in. on August 4,
2009, causing a 2 !/2-day delay.
31
-------
Figure 3-16. Tap Locations in Pit 4
Figure 3-17. Tap Locations in Pit 2
32
-------
Figure 3-18. Tap Locations in Pit 5
Figure 3-19. Test Pipe Discharge to Storm Sewer Configuration
33
-------
3.4.3 Simulated Leaks. The goal of a leak simulation demonstration is to determine the
capabilities of different technologies at detecting and locating leaks of sizes that are a concern to water
utilities. Ideally, the demonstration should test both tool sensitivity for discerning leak rates and location
accuracy. As such, the demonstration was designed to identify capabilities for detecting/locating various
sizes of leaks, as well as multiple leaks in close proximity. In addition, and per vendor recommendations,
the simulated leaks were discharged close to the pipeline surface (rather than piped away) to generate a
similar acoustic signature to actual pipeline leaks.
The criticality of a leak can vary, depending on the pipeline operational needs and location of the leak.
Leaks that initially appear benign from a lost revenue standpoint may still be critical for water utilities to
detect and repair to avoid further consequences like icing on roadways or contamination of natural
waterways. However, if this same leak is in a non-critical location, it may only be an indication of
potential future pipeline condition issues and not critical for immediate repair.
The various leak rates were selected based on correspondence with EPA and the LWC. Leak rates less
than 1 gpm were selected to demonstrate technology potential as an early detection warning method.
Leak rates between 1 and 5 gpm were selected based on the potential for the leak to be economical to
repair. Leaks greater than 10 gpm are typically detected without the need for inspection and therefore
were not included as part of the demonstration.
The 1-in. taps in Pits 2, 4, and 5 were used to create leaks within a range of sizes from as low as 3 gallons
per hour up to 8 gpm. Removable restriction ports (Table 3-5) in each tap were used to create the various
sizes of leaks. The restriction ports ranged in size from 0.25-in. down to a 0.02-in. diameter hole. A
narrow saw-cut restriction port was also used to simulate a leaking crack in demos 2 through 4; in demos
2 and 3 it was combined with a larger leak orifice nearby while in demo 4 it was isolated. Since different
pipeline pressures give different flow results from each orifice, the pipeline pressure was monitored and
recorded during the leak inspections (see Figure 3-20). The typical line pressure during the demonstration
was between 52 and 55 psi, with the highest pressure attained in the morning and a gradual drop in
pressure during the day.
To keep water from filling the pits and changing the leak noise level, a slotted vertical standpipe
containing a sump pump was placed in the corner of each excavation (see Figure 3-21). Each pit was
backfilled with gravel to a level slightly above the leak taps to prevent flow from spraying out of the pit.
The goal was to provide a consistent signal level and mimic acoustic signals similar to leaks covered by
backfill. During each demonstration, the water level in the pits was monitored to ensure it did not raise
above the outlet of the corp valves. Between demonstrations, the slotted standpipe was pumped down as
necessary. Input from the leak detection vendors was requested, and the simulated leak design was
adapted from their supplied procedures and phone conversations. The leak simulation approach was
accepted by the leak detection companies at a meeting prior to installation of the corp valves.
Since Echologics' LeakfinderRT technology did not require flow to operate, the discharge valve
downstream of Pit 3 was kept in the closed position during the demonstration while the supply valve at
Chenoweth Lane was partially opened to pressurize the pipe and to maintain constant leak rates during the
LeakfinderRT demonstration.
34
-------
Table 3-5. Orifices Used to Simulate Various Leak Sizes During the Demonstration
Orifice Size
(in.)
Leak Rate (gpm)
at 30 psi
Leak Rate (gpm)
at 50 psi
Photo
0.250
5.9
7.5
0.187
0.125
3.3
1.4
4.3
1.8
0.062
0.43
0.54
0.032
0.11
0.14
0.020
0.05
0.06
Crack
0.09 wide
0.375 long
0.72
0.97
35
-------
B9fc^<*WŁ
Figure 3-20. Monitoring Pipeline Pressure
Sump pump
located within
stand pipe
•
Figure 3-21. Sump Pump System for Simulated Leak Locations
36
-------
3.5 Test Configuration
Three vendors participated in the water main inspection demonstration for leak detection technologies on
the following dates:
• Echologics LeakfinderRT leak detection/location with surface-mounted sensors and
hydrophones. Onsite from July 6, 2009, through July 8, 2009, and again August 10, 2009,
through August 12, 20097
• PPIC Sahara® leak detection/location. Onsite from July 13, 2009, through July 17, 2009
• Pure SmartBall™ leak and gas pocket detection/location. Onsite from August 3, 2009,
through August 7, 20098
The activities conducted each day are provided in Table 3-6.
3.5.1 PPIC Sahara* Leak Detection. Five Sahara® insertions were performed from July 13 to
July 17 for three different inspection technologies (leak detection, video, and condition assessment) that
used the same tether, insertion equipment, and tracking method as the leak detection technology. The
equipment arrived by a custom vehicle on the morning of the inspection. The vehicle contained the
sensors, cable deployment system, support electronics, and electrical power for conducting video, leak,
and condition assessment surveys. The Sahara® video inspection was performed first, on July 13, to
inspect the inside of the pipeline. This inspection identified potential obstacles for other internal
inspections, as well as internal corrosion and air pockets. The Sahara® video head was inserted into Pit 1
and traversed the line using the pipeline flow. In its initial launch, the Sahara® video parachute caught
during insertion and failed to deploy; it was replaced rather than repaired (as noted by the EPA
contractor's field observations). Once re-inserted, the Sahara® video head traveled the length of the test
pipe. After reaching Pit 3, the video head was then retracted and taken out of Pit 1.
Sahara® Leak Detection's activities were performed on July 14, 15, and 17, 2009. Three full surveys of
the pipeline were performed to test different arrangements of simulated leaks. Like the Sahara® Video
head, the Sahara® Leak Detection sensor head was inserted and retracted out of Pit 1 to conduct the leak
survey. With the proper fittings being installed prior to the inspection, setup required about 2 hours and
teardown required about 1 hour. Setup and tear down were faster on subsequent days of the
demonstration. All fittings that touched the water were sprayed with a chlorine solution for sterilization.
On July 15, a thunderstorm required that flow in the pipeline be stopped due to reduced storm sewer
capacity, so the survey ended before completion. Also, during several inspections, at the request of the
operator, the pipeline flow rate was increased to maintain the inspection rate.
After the initial inspections, the Sahara® hydrophone was tested onsite and found to have technical
problems and as such did not detect some of the small, simulated leaks. Subsequently, that particular
hydrophone was replaced on the last day of the demonstration with an alternate hydrophone confirmed to
pass QA/QC tests. Two of the very small leaks were re-simulated and were detected onsite using the new
7 Because a significant amount of air was in the line during their first visit to the demonstration site, Echologics was
unable to get accurate data from their LeakfinderRT technology. The test pipe was dewatered and cut a few weeks
prior to the demonstration to install tees at both ends of the test pipe. While the test pipe was filled and flushed for a
few hours upon completion of the tee installation, a video assessment showed that air pockets remained throughout
the pipeline. Attempts were made by LWC to remove air from the line and Echologics was permitted to return at a
later date to complete their demonstration.
8 Heavy rain fall occurred on August 4, 2009, preventing LWC from discharging to the storm sewer for 2-1/2 days.
As such, Pure was unable to access the pipeline for leak assessment until August 6, 2009.
37
-------
Table 3-6. Daily Activities for Each Leak Detection Technology Vendor
Date
Daily Activities
Echologics LeakfinderRT - One operator
July 6
July?
JulyS
Aug. 10
Aug. 11
Aug. 12
• Checked-in at demonstration site and set-up equipment
• Unable to complete noise test; background levels appeared low
• Installed sensors (accelerometers) in Pits 1 and 3 with receiver in Pit C
• Assessed background noise; added filters
• Reconfigured to 1,000ft
• Pipe pressure at 53 psi
• Suspected that the pipe had air pockets because could not get a clear signal;
tried to swap RF transmitters
• Still unable to get a good signal; prior experience by the vendor suggested that
the cause of the poor signal may have been air in the line.
• Opened fire hydrant to purge air from line. Milky water observed.
• Did not get any data; arranged to come back at a later date
• Checked-in at demonstration site and setup equipment
• Condition assessment for pipe from Pits 1, 2, and 3 using accelerometers
• Found one large leak and one or two smaller leaks
• Hydrophones placed in various pits to conduct leak detection
• Pipe pressure between 52 and 54 psi
• Road traffic over pits caused noise interference increasing inspection time
• Packaged equipment for shipping
PPIC Sahara* - 2-3 operators9
July 13
July 14
July 15
July 16
July 17
• Checked-in at demonstration site and setup Sahara® Video equipment
• Pipe pressure at 56 psi; flow rate ~ 2.6 ft/s with three valve turns
• Launched Sahara® Video; parachute failed to deploy and was replaced
• Started video inspection; increased flow to keep camera from bouncing (-2-2-
!/2 hours)
• Retrieved Sahara® Video equipment (~45 minutes)
• Launched Sahara® leak detection equipment for calibration survey; natural
leaks and simulated leaks detected during all surveys
• Conducted second leak detection survey
• Pipe pressure at ~58 psi
• Launched Sahara® leak detection equipment for third and fourth leak surveys
• Installed accelerometers for condition assessment
• Launched Sahara® condition assessment equipment - hydrophone
• Pipe pressure at ~55 psi
• Finished condition assessment
• Pipe pressure at ~55 psi
• Conducted leak detection survey with new hydrophones
• Prepared for PPIC PipeDiver inspection
• Packaged equipment for shipping
Pure SmartBall™ - Two operators
Aug. 3
Aug. 4
Aug. 5
• Check-in at demonstration site and set up equipment
• Significant rainfall; demonstration canceled
• Significant rainfall; demonstration canceled
More were onsite for the demonstration. PPIC used the demonstration to train new operators.
38
-------
Table 3-6. Daily Activities for Each Leak Detection Technology Vendor (Continued)
Date
Aug. 6
Aug. 7
Daily Activities
Installed sensors in Pits 1, C, and 3
Installed insertion and extraction tubes
Launched SmartBall™ (-45 minutes)
Conducted second SmartBall™ run (~50 minutes)
Dismantled insertion and extraction tubes
Installed insertion and extraction tubes
Conducted first SmartBall™ run (-75 minutes)
Conducted second SmartBall™ run (~53 minutes)
Conducted third SmartBall™ run (~44 minutes)
Dismantled insertion and extraction tubes
Packaged equipment for shipping
hydrophone. As a precaution, now all Sahara® hydrophones are reported to be tested following standard
QA/QC procedures and an on-site test protocol is implemented prior to inspection. The conclusion from
the last day was that the hydrophone used on the first three days had lower sensitivity to small leaks.
Sahara® Leak Detection verbally reported leaks as they were detected during the survey. As the
hydrophone transited the test pipe, the operator listened for leaks and stopped the hydrophone to isolate
the location of the leak. The hydrophone would periodically be moved back and forth to better assess
whether a leak was found and its potential size. When a leak was confirmed, an aboveground tracker
would locate the exact position of the tool and mark where the leak was found.10 Throughout the
demonstration, observers could listen to the hydrophone output, watch data on computer screens, and
speak with analysts about the real-time results. A preliminary report of leak detection was provided to
EPA's contractor on July 27, 2009. A final report with the leak detection and structural integrity
demonstration results was submitted to EPA's contractor. This document in Appendix A was resubmitted
on October 14, 2009 after leak verification information was released by EPA's contractor. Information
was added by PPIC on a 7th leak in close proximity to a calibration leak, which was claimed to mask the
natural leak signal. The rest of the report was not changed including the cover page and original
submission date of July 2009.
3.5.2 Pure Technologies SmartBall™. Five SmartBall™ insertions were performed from August
6 to August 7, 2009, for leak detection and pipe-wall thickness assessment. Seven cases of equipment,
five suitcase-sized and two long, thin boxes arrived by common overnight delivery service the week prior
to the demonstration.
SmartBall™ leak detection was performed by launching the equipment in Pit 1, allowing the SmartBall™1
to travel with the water flow to conduct the inspection, and then extracting the equipment using an
extraction tube in Pit 3. LWC provided a 6-in. ANSI flange on the top of the gate valve in Pits 1 and 3 to
which Pure mounted its 4-in. diameter insertion and extraction tubes. Prior to the insertion, Pure verified
that adequate flow was available to carry the SmartBall™1 the full length of the test pipe in a reasonable
amount of time. Flow rates between 1 and 2 ft/s were maintained, resulting in inspection times between
45 minutes and 1 hour.11 The inspection procedure involved first placing the extraction net in the
pipeline, then inserting the SmartBall™1. With the proper fittings being installed prior to the inspection,
the setup and tear down process for SmartBall™1 required about an hour each. All fittings that touched
the water were sprayed with a chlorine solution for sterilization.
10 The typical PPIC procedure is to mark the road above the pipeline with marking paint at the exact position of the
leak; however, this was not done during the demonstration to avoid giving away results to subsequent vendors
conducting leak assessment surveys.
11 The SmartBall™ typically travels at about 90 percent of the flow rate.
39
-------
Knowing the position of the SmartBall™ within the pipeline is critical for locating important pipeline
features, such as leaks, and multiple locating methods are used by SmartBall™. Distance profiles are
generated to give a rough estimate of the SmartBall™ position over time. Data obtained from the
accelerometers and magnetometers on board the SmartBall™ are used to obtain a velocity profile for
tracking the tool. Also, absolute position reference points obtained from the SBR are applied to the time-
stamped data to track the position of the SmartBall™1. Individual SBRs tracked the ball's progress
through the pipeline for over 850 ft; the distance and location of these SBRs were based on information
provided to Pure by EPA's contractor. The result of the rotation profile and SBR tracking is a position-
versus-time relationship for the entire run of the tool. The exact location of where each SBR was placed
along the test pipe during each run is detailed in Table 3-7.
Figure 3-22 shows an example of the position data recorded for each run. The position of the
SmartBall™1 indicated by the red line was fixed by fitting the position profile to known locations along
the pipeline. The slope of the red line indicates the instantaneous velocity of the tool. An example of the
velocity of the SmartBall™ as it travels through the pipeline is shown in Figure 3-23. Figure 3-24 shows
an example of the ball's position as it was tracked in real time by the SBRs. The combined use of travel
time (a coarse measure of position), velocity, and SBR position tracking (data can be noisy) provides an
acceptable solution for determining the SmartBall™1 position.
Once the ball was launched, observers and technicians waited for the ball to be received at Pit 3. The
vendor verbally reported on leaks to EPA's contractor the day after each inspection. There were no
ongoing activities for the operators to perform as the SmartBall™1 traveled through the pipeline. A final
report of leak detection results was provided on August 14, 2009.
To quantify the approximate leak rate documented during the inspection, Pure compared the leak
indication power of a detected leak with that of a known leak rate. The previously established calibration
curve12 used by SmartBall™1 is shown in Figure 3-25. Additional calibration leaks (Demo 1) were
provided by EPA's contractor during the demonstration (shown in green in Figure 3-25) to help Pure size
and locate blind leaks during subsequent runs.
Pure noted that because the simulated leaks are controlled and released through a threaded outlet, the
comparison to actual field-condition leaks may vary. This is because the acoustic frequency and power
indication of any leak will vary with many factors, including pressure, pipe diameter, anomaly size, and
anomaly configuration (pin-hole, rolled gasket, split pipe, etc.). However, the leak calibration curve
provides a useful tool to approximate leak rates for identified leaks. The reported leaks detected during
the inspection are shown as red circles in Figure 3-25.
Pure reports actual leak rates, but also provided classification of their leak sizes as follows:
• 0 to 2 gpm (0-7.5 liters per minute) = small,
• 2 to 10 gpm (7.5 to 37.5 liters per minute) = medium,
• > 10 gpm (37.5 liters per minute) = large.
Subsequent results only show the actual leak rates as provided by Pure.
12 The calibration curve was developed by Pure outside of this demonstration using a Vi-in valve attached to the
extraction stack and a calibrated bucket to measure the leak rates. Further details are provided in their report
included in Appendix B.
40
-------
Table 3-7. SmartBall1 V1 Receiver (SBR) Locations
Location ID
Insertion
Midpoint
Extraction
Distance from
Launch (ft)
0.0
809.0
2,057.0
Time of Oay(hh:m
Figure 3-22. Example Position Profile of the SmartBall™ vs. Time of Day from Run
#1 on August 6 (courtesy of Pure)
'^f^^!^^^ w^Vv\AVW^^/Ąyv'
13:59:55 14:08:18
Time of Day (hh:mm:s5)
Figure 3-23. Example Velocity Profile of the SmartBall™ vs. Time of Day from Run #5 on
August 7 (courtesy of Pure)
13:50:58
Time of Day(hh:mm:ss)
Figure 3-24. Example SmartBall™ Receiver Tracking Points vs. Time of Day from Run #2 on
August 6 (courtesy of Pure)
41
-------
•A calibration leak/point member
of this calibration curve, not part
of default calibration curve.
• Marked Leak.
Leak Rate (gal/min)
Figure 3-25. Example Leak Calibration Curve Used to Size Leaks (courtesy of Pure)
3.5.3 Echologics LeakfinderRT. From July 6 through 8, 2009, Echologics was onsite to demonstrate its
ThicknessFinder and LeakfinderRT technologies. These initial inspections were unsuccessful.13 Echologics was
allowed to return August 10 through 12 to have a second chance at demonstrating these technologies. The leak
assessment was conducted on August 11 and 12. One Echologics technician arrived the day of the inspection
with two cases of equipment the size of a common suitcase in the back of a small rented vehicle. This report
describes the LeakfinderRT demonstration. The ThicknessFinder demonstration will be reported in a subsequent
report.
LeakfinderRT used two types of sensors: (1) hydrophones (1.5-in. NPT threads) that required contact with the
water column, and (2) accelerometers that were glued to the outside of the pipe. The distance between sensors is
a function of many variables, including local noise considerations. The simplest configuration would have been
to examine the entire distance from Pit 1 to Pit 3, using the 12-in. taps as the sensor locations. However, initial
tests showed that this configuration was not feasible due to excessive noise levels. Instead, taps installed for the
simulated leaks in Pit 2 were made available to shorten the hydrophone distance intervals for assessing the
simulated leaks to 1,000-ft. For detection of natural leaks, accelerometers were used to record and correlate data
between neighboring sensors spaced 300 ft apart using all the access pits (Pits 1 through 5 and Pits A through F).
Echologics initially performed background measurements with LeakfinderRT accelerometers over the shorter test
pipe lengths of 300 ft in order to find any natural leaks and to collect pipe wall thickness data at the same time
(using ThicknessFinder, which is discussed in a companion report). This was followed by the simulated leak
assessment using LeakfinderRT hydrophones over the larger distance intervals (e.g., 1,000 ft). The assessment
lengths were a field decision made by Echologics based on the test pipe configuration.
Echologics presents the need for using the two sensor types in its report (see Appendix C). They state that, in
general, it is more challenging for a leak noise correlator to survey for water main leaks than it is
13 Because a significant amount of air was in the line during their first visit to the demonstration site, Echologics was unable
to get accurate data from their LeakfinderRT technology. The line was dewatered and cut a few weeks prior to the
demonstration to install tees at both ends of the test pipe. While the line was filled and flushed for a few hours upon
completion of the tee installation, a subsequent video assessment showed that air pockets remained throughout the pipeline.
42
-------
to locate a known leak, since there will be a high incidence of negative (no leak) results. When many
negative results are encountered, the surveyor may begin to question the operation of the equipment or his
procedures. Therefore, one of the main issues with testing pipes where there is no known leak is the need
to take steps to ensure that the results are properly analyzed so that the presence (or lack) of a leak may be
definitively decided. Based on Echologics' previous experience with leak detection surveys and their
familiarity with acoustic technology, procedures were implemented onsite, and follow-up analyses were
performed to make a definitive decision on whether a leak was present. They performed the following
activities:
• Hydrophones were attached on valves or hydrants available at each site. Where
measurements were performed on valves, the sensors were placed on the tops of valve keys
that had been lowered onto the valves or placed directly on the valve nut when possible (if the
valve chamber was clear of debris).
• After placement of the sensors on the appropriate valve or hydrant, the fitting was tapped and
listened to at the radio receiver to ensure that the sensor was functioning and that the radio
signal was reaching the receiver properly. This is called a scratch test.
• Sensor spacing was measured using a calibrated measuring wheel.
• A correlation measurement was performed, and the signal was saved to the computer, so that
further analysis could be performed later in the office, and so that the client could have a
permanent record of the raw noise file, if needed.
• Where a positive signal was detected (a correlation peak with good signal coherence), the
location was immediately checked to determine if it corresponded to a service line or other
notable draws from the pipe. If this was the case, several more correlations were conducted
to see if the "usage" stopped.
• Where negative results were obtained (no clear correlation peak was obtained), a series of
checks was completed, including a review of coherence and of the two communication
frequency spectra, to detect the presence of a PVC repair or some other anomaly in the test
section. Such checks are part of Echologics' protocol for leak detection surveys.
Echologics also presents several possible sources of error in its demonstration results documentation (see
Appendix C). These include inaccurate measurement of distances along the test pipe and errors in
manufacturing wall-thickness tolerances. A much smaller source of errors from electronic hardware and
digital processing is also identified by Echologics in Appendix C.
Once the acoustic sensors (either hydrophones or accelerometers) were set up, a few minutes of data were
recorded. Observers of the demonstration could watch the real-time data analysis and discuss the findings
with the LeakfinderRT technician. A preliminary field-written leak report was provided within an hour of
the inspection. With the proper fittings being installed prior to the inspection, setup and tear down
required about 1A hour each. All equipment that touched the water column was wiped with a chlorine
solution for sterilization. A detailed preliminary report was provided on August 31, 2009, and a final
report was provided on September 30, 2009, with minor revisions submitted on November 4, 2009 and
again on November 13, 2009.
43
-------
3.6
Post-Demonstration Leak Confirmation
Simulated Leaks
The leak rates for each restriction port discussed in Section 3.4.3 were established prior to the
demonstration. EPA's contractor quantified the leak rates at pipeline pressures ranging from 20 to 50 psi
for the various restriction port sizes, using laboratory facilities in West Jefferson, OH. A large pipe was
filled with water, and then pressurized to the various levels. Each restriction port was placed on a valve
attached to the pipe, and the leak rate was calculated based on the weight of water released over a specific
period of time. Figure 3-26 provides these leak calibration curves for each restriction port. Given the
orifice sizes in Figure 3-26, the calibration curves were nominally linear over the range of pressures used.
The conditions at the exit of the orifice can increase or decrease the flow rate. Water at the exit can
decrease flow rate by creating back pressure; the stone backfill and sump system installed was used to
reduce this effect. However, the stone can potentially increase the flow rate. The flow through an orifice
naturally contracts, referred to as the vena contracta; backpressure and cavitation caused by the backfill
can increase the effective orifice size. Testing was performed under controlled conditions; the variation
with water and backfill could increase or decrease the flow 2 to 10%. Because measuring the leak rate
during the demonstration may change the acoustic properties of the leak, the exact leak rate during each
live inspection run was not measured. Instead, the test-pipe pressure was monitored and recorded during
the demonstration and used in combination with the leak calibration curves (see Figure 3-26) to estimate
the actual leak rate. Since testing pressures were up to 8psi above the expected maximum pressure of 50
psi, the flowrates were determined by extrapolating the nominally linear curves in Figure 3-26.
I
E
• 0.25" orifice
0.187" orifice
0.125" orifice
Crack
0.062" orifice
.0.02" orifice.
10
20
30
Pressure (psig)
40
50
Figure 3-26. Calibration Curves for Restriction Ports
44
-------
The leak matrix used for the demonstration is provided in Table 3-8 for the PPIC Sahara® and Pure
SmartBall™ technologies. Upon completion of the leak test matrix, both vendors had extra time available
to conduct additional simulated leak runs to help in advancing their technologies. SmartBall™ conducted
one additional run (called Demo 5) with a 0.25-in. orifice in CV1 and a 0.063-in. orifice in CVS. Sahara®
conducted three additional runs (called Demo 5, Demo 6, and Demo 7) with a 0.187-in., 0.062-in., and
0.032-in. orifice in CV2.
Table 3-8. Leak Test Matrix for PPIC Sahara® and Pure SmartBall
TM
Pit#
Pit 4
Pit 2
Pit5
Corp
Valve
ID
CV1
CV2
CVS
CV4
CVS
CV6
CV7
CVS
Leak Configuration
Demo 1
(Calibration)
Orifice
Size (in)
0.063
—
0.250
—
—
—
—
-
Leak
Rate00
(gpm)
0.57
~
7.8
~
~
~
~
~
Demo 2
Orifice
Size (in)
—
0.02
—
0.032
—
0.063
Crack
0.25
Leak
Rate(a)
(gpm)
—
0.06
—
0.14
—
0.57
1.0
7.8
Demo 3
Orifice
Size (in)
Crack
0.125
0.032
0.063
~
0.188
0.25
0.063
Leak
Rate(a)
(gpm)
1.0
1.9
0.14
0.57
—
4.6
7.8
0.57
Demo 4
Orifice
Size (in)
0.02
0.188
~
~
Crack
~
0.25
0.125
Leak
Rate00
(gpm)
0.06
4.6
~
~
1.0
~
7.8
1.9
(a) Leak rate is for a pressure of 54 psi.
testing from approximately 50 psi to
The pipeline pressure,
58 psi.
and consequently the leak rates, varied during
The test matrix is slightly different for LeakfmderRT due to the methods used to collect the data. Rather
than inspecting the entire length of pipe in one run, LeakfmderRT used two hydrophones placed
approximately 1,000 ft apart to assess the line for leaks. As such, the simulated leak demonstration had to
be conducted in three stages to cover the entire test pipe length. The first set of four leak scenarios was
established using CV1 and CV2 in Pit 4; the two hydrophones that bracketed Pit 4 were in Pits 1 and 2.
The second set of four leak scenarios was established using CV7 and CVS in Pit 5; the two hydrophones
that bracketed Pit 5 were in Pits 2 and 3. The third set of four leak scenarios was established using CV3,
CV4, CVS, and CV6; the two hydrophones that bracketed Pit 2 were in Pits 4 and 5. The LeakfmderRT
test matrix is presented in Table 3-9.
For each simulated leak demonstration, a cluster of leaks was used to determine the leak detection,
location, and sizing capabilities for each technology. A leak cluster is defined as anywhere from one to
three leaks within the same pit that are axially spaced anywhere from 0.6 to 2.7 ft. In some of the
demonstrations, only one leak orifice was opened within a pit while other pits may have had two or three
orifices open. The demonstration was designed to identify capabilities for detecting/locating various sizes
of leaks, as well as multiple leaks in close proximity.
45
-------
Table 3-9. Leak Test Matrix for Echologics LeakfinderRT
Demo#
Demo 1
(Calibration)
Demo 2
Demo 3
Demo 4
Demo 1
(Calibration)
Demo 2
Demo 3
Demo 4
Demo 1
(Calibration)
Demo 2
Demo 3
Demo 4
Pit#
Pit 4
Pit 4
Pit 4
Pit 4
Pit5
Pit5
Pit5
PitS
Pit 2
Pit 2
Pit 2
Pit 2
Corp Valve
ID
CV1
CV2
CV1
CV2
CV1
CV2
CV1
CV2
CV7
CVS
CV7
CVS
CV7
CVS
CV7
CVS
CVS
CV4
CVS
CV6
CVS
CV4
CVS
CV6
CVS
CV4
CVS
CV6
CVS
CV4
CVS
CV6
Leak Configuration
Orifice
Size (in.)
0.063
0.02
Crack
0.125
0.02
0.188
~
Crack
0.25
0.25
0.063
0.25
0.125
0.25
0.032
0.063
0.032
0.063
0.188
Crack
LeakRate(a)
(gpm)
0.57
0.06
1.0
1.9
0.06
4.6
—
1.0
7.8
7.8
0.57
7.8
1.9
7.8
0.14
0.57
0.14
0.57
4.6
1.0
(a) Leak rate is for a pressure of 54 psi. The pipeline pressure varied
during testing from approximately 50 psi to 58 psi. The mean pressure
was used to determine the flow rate from the pressure-flow rate graph.
46
-------
Naturally Occurring Leaks
The leak rates and locations for the simulated leaks were recorded by EPA's contractor for later
comparison to the reports provided by the individual vendors. In addition, the leak detection/location
capabilities for the various technologies were qualitatively verified through focused excavations of the
pipeline to find naturally occurring leaks.14 Eight identified leak locations were excavated and examined.
During the excavation, with the assistance of MAC Construction, the soil was examined for excessive
moisture and erosion. When each leak site was fully uncovered, visual assessment was used to determine
whether the leak was from a bell-and spigot joint or the body of the pipe at an anomaly such as corrosion
or a crack. After the potential leak sites were uncovered, the pipe was pressurized to qualitatively assess
the leak sizes by examining the amount of water leaching/spraying from the pipe.
During verification, EPA's contractor was able to definitively confirm naturally occurring leaks in four of
the eight locations that were excavated; all occurred at the be 11-and-spigot joint (see Table 3-10 and
Figure 3-27). This does not mean that the other four leaks that could not be verified did not exist. For
two of the unverified leaks, the soil was definitely wetter in the area excavated; however, EPA's
contractor was unable to pinpoint the leak location. For the other two unverified leaks, where the soil was
relatively dry, it is quite possible that the reported locations used to determine the excavation location
were off by several feet so that EPA's contractor could not find evidence of a leak.
Additional measurements were taken for the one natural leak found in Pit L to understand what may have
caused the leak. The joint rotation was measured — that is, the angle between the two pipe joints where
the leak was found. Researchers at the NRC of Canada suggested measuring the joint rotation to test a
hypothesis that an angle of more than 2° would cause a joint leak and that an angle of more than 5° would
crack the bell. While the test pipe was generally level, this leak occurred near a storm sewer. The angle
measured was approximately 1.5°, indicating that the bell would not be cracked (as was verified visually)
or leaking due to joint rotation. Therefore, the large leak was most likely due to degradation in the leadite
seal for the bell-and-spigot joint rather than joint rotation.
Table 3-10. Natural Leak Verification Results
14
Leak
ro#
LI
L3
L4
L6
L7
L10
Lll
L12
Leak Excavation
Location (ft)
50-53
194-199
338-341
556
638-641
1,724
1,809
1,906-1,933
Location Where
Leak Found (ft)
52
195
339
556
640
—
—
1,909
Description
Soil was wet, but there was a nearby storm sewer at 52 ft;
leak not pinpointed, but elevated moisture considered as an
indirect indication of potential leak.
A lot of water in pit near bell-and-spigot joint; did not
pinpoint leak
Large leak at bell-and-spigot joint
Small leak at bell-and-spigot joint
Small leak at bell-and-spigot joint
No wet soil found at this location; inconclusive
No wet soil found at this location; inconclusive
Soil was moist at -1,906 ft or 1,930 ft; leak not pinpointed,
but elevated moisture considered as an indirect indication
of a potential leak.
14 EPA's contractor was only able to pressurize the line to witness four of the leaks (L3, L4, L6, and L7). Shortly
after verifying these leaks, the stem broke on the control valve near Chenoweth Lane, prohibiting the line from
being pressurized for remaining leak verification. As such, other external cues, such as wet soil and possibly odor
from leaching leadite in joints, were used to indirectly decide if a leak could have been present in the line.
47
-------
Figure 3-27. Leak ID L7 - Small Leak at Bell-and-Spigot Joint
48
-------
4.0: RESULTS AND DISCUSSION
PPIC Sahara®, Pure SmartBall™, and Echologics LeakfmderRT leak detection, location and sizing
technologies were demonstrated on a 76-year-old, 2,057-ft-long portion of a cement-lined, 24-in. diameter
cast iron water main in Louisville, KY. While each technology used some form of acoustic listening
device, the implementations were quite different:
• PPIC's Sahara® mounted a hydrophone sensor at the end of a cable tether. The hydrophone,
which was inserted and pulled through the pipeline using the water flow, provided real-time
assessment of leaks. The hydrophone sensor was also tracked by an operator from ground
level and leaks were marked on the pavement.
• The Pure SmartBall™1 sensor and data-recording device were placed within a foam ball. The
sensor and ball were inserted in the pipeline and propelled by the water through the pipeline
to a downstream extraction point where a net inserted into the pipe caught and removed the
unit.
• Echologics LeakfmderRT demonstrated two types of sensors. Pairs of accelerometers were
mounted on the outside of the pipe at discrete locations to detect and locate unknown leaks.
Then pairs of hydrophones in contact with the water at discrete locations were used to
estimate simulated leak rates.
After the demonstration was complete, a closed-circuit television video inspection of the entire test pipe
length was performed. The inspection report indicated that the cement liner was uniform and no through-
wall anomalies were detected in the pipe wall. The joints at the reported natural leak locations were
closely examined as well as the joints before and after these leak locations and no significant differences
(such as larger gaps) were observed.
The implementation of the leak detection/location technology demonstration could not have been
accomplished without the significant efforts of LWC, MAC Construction, and the technology vendors
PPIC, Pure, and Echologics. Each vendor participated by mobilizing their technology and crews onsite,
setting up the equipment, operating the technology, collecting data, and providing the requested
inspection reports. Detailed results for all three leak detection technologies are discussed in the
subsequent sections with a summary of the results provided in Section 2. The individual leak inspection
reports provided by each vendor are included in Appendix A (Sahara®), Appendix B (SmartBall™1), and
Appendix C (LeakfmderRT).
Since this demonstration was a snapshot in time, new developments may have taken place since
completion of the demonstration. Therefore, the findings in this report may not be wholly representative
of the current operational capabilities of the demonstrated technologies. For this reason, the vendors were
asked to provide formal comments on the final leak detection report to highlight advancements since
completion of the demonstration and/or clarification on what was reported. These comment letters are
contained in Appendix D.
4.1 PPIC Sahara® Systems
PPIC presented two sets of leak detection/location results: (1) for naturally occurring pipeline leaks; and
(2) for simulated leaks. All results provide a qualitative evaluation of the pipe condition and leak sizes
(very small, small, medium, and large). As with the other technologies, the location accuracy of the
anomalies is dependent on the accuracy of the pipe distance as measured on the surface and lay
information as the pipe may not precisely follow the road surface.
49
-------
4.1.1 Summary of Results. For Sahara® Leak Detection, PPIC has defined their leak rate
classification scheme based mainly on the distance away from a leak that the leak can first be detected.
Table 4-1 shows the leak rate classification scheme developed by PPIC based upon their own data for
pipes ranging from 24-in. to 60-in. in diameter.
Table 4-1. Sahara® Leak Classification Table for 24-in. to 60-in. Diameter Pipe
Classification
Very Small
Small
Medium
Large
Very Large
Distance
Detected
[m]
0-2
2-5
5-15
15-50
50+
Approximate Measured Leak Size
Min
[m3/hr]
0
0.4
4
17
29
Max
[m3/hr]
0.4
4
17
29
42
Median
[m3/hr]
0.2
2
10
23
35
Min
[gpm]
0
1.8
18
75
128
Max
[gpm]
1.8
18
75
128
185
Median
[gpm]
0.88
8.8
44
101
154
The Sahara® Leak Detection inspection identified six natural leaks in real time and 11 simulated leak
clusters15. A seventh natural leak was reported after verification results were sent to PPIC by EPA's
contractor; the additional leak signal was reported to be masked by a larger artificial leak 22 ft away.
Details of the natural leaks are presented in Table 4-2 with specific information on the direction, distance
from the insertion point, and estimated leak rate. Details of the detected simulated leaks are presented in
Table 4-3, specifically the corp valve ID and estimated leak rate for each simulated leak.
Table 4-2. Natural Leaks Detected by Sahara® Leak Detection
Sahara®
ro#
i
2
3
4
5
6
Post(a)
Distance from
Start (ft)
50
194
338
638
1,696
1,906
558
Description as
Provided by Vendor
Very small leak
Very small leak
Large leak
Small leak
Very small leak
Small leak
Very Small
Direction from
Insertion Point
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
(a) Initially, PPIC did not report a leak at this location, but later identified a signal
after receipt of the verification data. PPIC reported that the signal was initially
masked by an artificial leak at 578 ft.
15 With extra time available, additional demos were conducted upon the vendor's request. These additional demos
are not included in the simulated leak verification numbers as they were done as a common courtesy to allow the
vendors to gather more data for technology development.
50
-------
Table 4-3. Simulated Leaks Detected by Sahara® Leak Detection
Pit#
Pit 4
Pit 2
Pit5
Corp
Valve ID
CV1
CV2
CVS
CV4
CVS
CV6
CV7
CVS
Estimated Leak Rates (gpm)
Demo 1
(calibration)
Veiy small(a)
Small
-
Demo 2
Veiy small(a)
Very small
Large
Demo 3
Small
Small
Medium
Demo 4
Very small(a)
Medium
Small
-
Demos 5, 6, and 7°°
Small
Very small
Very small
-
(a) While larger leaks were orally reported during the demonstration, the detection of these very small
leaks required post-operation analysis.
(b) Demos 5, 6, and 7 were only conducted for Sahara® Leak Detection. Demo 5 (0.187-in. orifice),
Demo 6 (0.062-in. orifice), and Demo 7 (0.032-in. orifice) leaks originated from CV2. As time
permitted, vendors were extended the courtesy of conducting additional tests for improving their
technology.
4.1.2 Leak Evaluation. Sahara® reported six natural leaks (LI, L3, L4, L7, LI3, and LI2) in real time,
and a seventh (L6) after verification results were provided to PPIC by EPA's contractor. The additional
leak signal was reported to be masked by an artificial leak 22 ft away. Except for one very small leak at
1,696 ft, which was not excavated and therefore could not be verified, the remaining five leaks were
directly or indirectly verified based on visual evidence and other vendor leak reports (leak pinpointed, wet
soil in the general vicinity, or another vendor reported a leak in the same vicinity).
Sahara® Leak Detection also detected and qualitatively estimated the leak rate for 11 of 19 simulated
leaks and 11 of 11 leak clusters. Each simulated leak cluster was a combination of one to three
consecutive leaks, from orifices of different sizes, arranged 0.6 to 2.7 ft apart. Within each leak cluster,
Sahara® accurately characterized the leak range for 6 of the 11 clusters. For leaks that were not
accurately characterized, 4 out of 5 were off by one size category, while the leak in Demo 2, Pit 5 was off
by two size categories. The location accuracy could not be evaluated as Sahara® only reported the pit
number in which they found the leak.
Sahara® was not able to discern two separate leaks in close proximity (less than 2.7 ft apart) for all of the
simulated leaks. Identifying whether a signal is from an isolated leak or multiple leaks in close proximity
is helpful in judging the general condition of a pipeline. However, it is also important to accurately
identify the location and size of the largest leak(s). As reported by PPIC, when individual leaks are at
close proximity, the leak signatures combine and are difficult to differentiate.
Details of the detected natural leaks and subsequent visual verification are presented in Table 4-4. The
detected simulated leaks vs. the actual test conditions are presented in Table 4-5.
51
-------
Table 4-4. Evaluation of Natural Leaks Detected by Sahara® Leak Detection
16
Leak
ID#
LI
L3
L4
L6
L7
L13
L12
Sahara
ID#
1
2
3
Post
Distance from
Start (ft)
50
194
338
558
Description as
Provided by Vendor
Very small leak
Very small leak
Large leak
Very small leak
Initially, did not report a leak, but found after
verification results were provided to PPIC. The
leak signal was reported to be masked by an
artificial leak at 578 ft.
4
5
6
638
1,696
1,906
Small leak
Very small leak
Small leak
Visually Verified by EPA Contractor?
Verification attempted; soil was wet, but
there was a nearby storm sewer at 52 ft; leak
not pinpointed, but elevated moisture
indicative of leak.
Verification attempted; a lot of water in pit
near bell-and-spigot joint at -195 ft; did not
pinpoint leak
Verification attempted; leak at bell-and-spigot
joint at -33 9 ft
Verification attempted; leak at bell-and-spigot
joint at -556 ft
Verification attempted; leak at bell-and-spigot
joint at -640 ft
No verification attempted
Verification attempted; soil was moist at
-1,906 ft; leak not pinpointed, but elevated
moisture indicative of leak.
Table 4-5. Evaluation of Simulated Leaks Detected by Sahara® Leak Detection17
Corp
Valve
ID
CV1
CV2
CV3
CV4
CV5
CV6
CV7
CVS
Distance to
Leak (ft)
A
577.4
578.4
1,082.2
1,082.8
1,084.1
1,084.9
1,583
1,585.7
S
Pit 4
Pit 2
Pit5
Estimated Leak Rates (gpm)(a)
Demo 1
(calibration)
A
0.59
-
8.2
—
—
—
—
-
S
Very
small
(0-1.8)
Small
(1.8-18)
-
Demo 2
A
~
0.06
—
0.15
—
0.59
1.1
8.2
S
Very
small
(0-1.8)
Very
small
(0-1.8)
Large
(75-128)
Demo 3
A
1.0
2.0
0.14
0.57
—
4.6
7.9
0.57
S
Small
(1.8-18)
Small
(1.8-18)
Medium
(18-75)
Demo 4
A
0.06
4.6
—
—
1.0
—
7.9
2.0
S
Very
small
(0-1.8)18
Small
(1.8-18)
Medium
(18-75)
Demo 5, 6, 7
A
~
4.6
0.57
0.14
—
—
—
—
—
~
S
Small
(1.8-18)
Very small
Very small
(0-1.8)
-
-
(a) Pipeline pressure was 58 psi for Demos 1 and 2; unknown for Demos 3 and 4 (assumed 55 psi); and 55 psi
for Demos 5, 6, and 7.
A = actual, S = Sahara®
16 The gray background signifies natural leaks that were verified in the field (by pinpointing leak or elevated
moisture); the red background signifies false negatives; and the white background signifies leaks found by one or
more leak detection technologies, but were not verified in the field.
17 The text with a red background signifies leak rate estimates that are off by one or more categories as defined by
each individual vendor.
18 After rates were disclosed, PPIC reported a natural air pocket near the leak and reported that this may have
masked the leak, thus minimizing its signature.
52
-------
4.2
Pure Technologies SmartBall
TM
SmartBall™ presented two sets of leak detection/location results: (1) for naturally occurring pipeline
leaks; and (2) for simulated leaks. Both sets of results provide a qualitative evaluation of the leak size
(small, medium, large), as well as an estimated leak rate. As with the other technologies, the location
accuracy of the anomalies is dependent on the accuracy of the pipe distance as measured on the surface
and lay information (as the pipe may not precisely follow the road surface).
4.2.1 Summary of Results. For SmartBall™, the leak rate classifications as defined by Pure are
shown in Table 4-6.
Table 4-6. SmartBall1 V1 Leak Classification Table
Classification
Small
Medium
Large
Ap
Min [gpm]
0
2
proximate Leak Size
Max [gpm]
2
10
Median [gpm]
1
6
>10
The SmartBall inspection reported 12 natural leaks and 11 simulated leak clusters . Details of the
natural leaks are presented in Table 4-7 with specific information on the distance from the insertion point,
leak description, and estimated leak rate. Details of the detected simulated leaks are presented in Table 4-
8, specifically the corp valve ID, distance, and estimated leak rate for each simulated leak.
Table 4-7. Natural Leaks Detected by SmartBall
TM
Leak
ID#
1
2
3
4
5
6
7
8
9
10
11
12
Distance from
Start (ft)
53
125
199
341
414
556
641
966
1,210
1,724
1,809
1,930
Description as
Provided by Vendor
Small leak
Small leak
Small leak
Medium leak(a)
Small leak
Small leak
Small leak
Small leak
Small leak
Small leak
Small leak
Small leakw
Approx. Size
(gpm)
0.15
0.1
0.8
15
0.2
1.0
2.0
0.1
1.0
1.5
2.0
5.5
(a) According to Table 4-6 these leaks would be classified as large and
medium, respectively; however, the description in Table 4-7 is what
was reported in Pure's inspection report as included in Appendix B.
19 With extra time available, one additional demo (i.e., demo 5) was conducted upon the vendor's request. The
additional demo is not included in the simulated leak verification numbers as they were done as a common courtesy
to allow the vendors to gather more data for technology development.
53
-------
Table 4-8. Simulated Leaks Detected by SmartBall
iTM
Pit#
Pit 4
Pit 2
Pit5
Corp
Valve
ID
CV1
CV2
CVS
CV4
CVS
CV6
CV7
CVS
Distance
to Leak
(ft)1
579
1,080
1,580
Estimated Leak Rates (gpm)
Demo 1
(calibration)
0.57
(small)
8
(medium)
0
(small)
Demo 2
0.3
(small)
2.8
(medium)
15
(large)
Demo 3
1.8
(small)
7.2
(medium)
30
(large)
Demo 4
4.5
(medium)
0.1
(small)
40
(large)
Demo 5
8
(medium)
0.57
(small)
0
(small)
Note: 1) One location was reported for all of the simulated leaks associated with a specific pit.
4.2.2 Leak Evaluation. SmartBall™ reported 12 natural leaks (LI through LI2) within the
inspected area. For the 12 natural leaks reported by SmartBall™, six were excavated and directly or
indirectly verified (LI, L3, L4, L6, L7, and LI2) based on visual evidence (leak pinpointed, wet soil in
the general vicinity, or another vendor reported a leak in the same vicinity); two other reported leaks (L10
and LI 1) were excavated, but the existence of small leaks were not conclusive. The remaining four
locations were not excavated due to time and budget constraints and therefore could not be verified.
SmartBall™ reported 11 of 19 total simulated leaks but reported 11 of 11 leak clusters. Each simulated
leak was a combination of one to three consecutive leaks, from orifices of different sizes, arranged 0.6 to
2.7 ft apart. Within each leak cluster, SmartBall™1 accurately characterized the leak range for 7 of the 11
clusters. All four of the leaks not accurately characterized were off by one size category. The location
accuracy could not be evaluated as SmartBall™1 only reported the location of the pit and not where the
actual leaks were located.
SmartBall™1 was not able to discern two separate leaks in close proximity (less than 2.7 ft apart) for any
of the simulated leaks. The locations of the leaks in Pits 4, 2 and 5 were reported at only one point in the
excavation regardless of which leak orifice was open, not where the actual leak occurred.
The findings of the pipeline inspection are summarized in Table 4-9, along with the field verification
results for the naturally occurring leaks. Table 4-10 summarizes the simulated leak results as compared to
the actual leak rate based on the orifice size and pipeline pressure at the time of the inspection.
54
-------
Table 4-9. Evaluation of Natural Leaks Detected by SmartBall
TM20
Leak
ro#
LI
L2
L3
L4
L5
L6
L7
L8
L9
L10
Lll
L12
SmartBall™ Results
SmartBall™
LeakID#
1
2
3
4
5
6
7
8
9
10
11
12
Distance
from Start
(ft)
53
125
199
341
414
556
641
966
1,210
1,724
1,809
1,930
Description
Small leak
Small leak
Small leak
Medium leak
Small leak
Small leak
Small leak
Small leak
Small leak
Small leak
Small leak
Small leak
Approx.
Size (gpm)
0.15
0.1
0.8
15
0.2
1.0
2.0
0.1
1.0
1.5
2.0
5.5
Visually Verified by EPA Contractor?
Verification attempted; soil was wet, but
nearby storm sewer at 52 ft; leak not
pinpointed, but elevated moisture indicative
of leak.
No verification attempted
Verification attempted; water in pit near bell-
and-spigot joint at -195 ft; did not pinpoint
leak
Verification attempted; leak at bell-and-
spigot joint -339 ft
No verification attempted
Verification attempted; leak at bell-and-
spigot joint at -556 ft
Verification attempted; leak at bell-and-
spigot joint -640 ft
No verification attempted
No verification attempted
Verification attempted, but no wet soil was
found at -1,724 ft; inconclusive
Verification attempted, but no wet soil was
found at -1,809 ft; inconclusive
Verification attempted; soil was moist at
-1,930 ft; leak not pinpointed, but elevated
moisture indicative of leak.
Table 4-10. Evaluation of Simulated Leaks Detected by SmartBall
TM21
Corp
Valve
ID
CV1
CV2
CV3
CV4
CV5
CV6
CV7
CVS
Distance to
Leak (ft)
Actual
577.4
578.4
,082.2
,082.8
,084.1
,084.9
,583
,585.7
SB(b)
579
1,080
1,580
Estimated Leak Rates (gpm)
Demo 1
(calibration)
Actual(a)
0.57
~
7.8
~
~
~
~
~
SB
0.57
8
0
Demo 2
Actual(a)
—
0.06
—
0.14
—
0.57
1.0
7.8
SB
0.3
2.8
15
Demo 3
Actual(a)
1.0
1.9
0.14
0.57
—
4.6
7.8
0.57
SB
1.8
7.2
30
Demo 4
Actual(a)
0.06
4.6
~
~
1.0
~
7.8
1.9
SB
4.5
0.1
40
Demo 5
Actual(a)
7.8
—
0.57
—
—
—
—
-
SB
8
0.57
0
(a) Actual leak rates assume a pipeline pressure of 54 psi as representative of the operating pressures.
(b) SB stands for SmartBall™; One location was reported for all simulated leaks associated with a specific pit.
20 The gray background signifies natural leaks that were verified in the field (by pinpointing leak or elevated
moisture); the yellow background signifies leaks that we attempted to verify, but could not find any evidence of a
leak; and the white background signifies leaks found by one or more leak detection technologies, but were not
verified in the field.
21 The text with a red background signifies leak rate estimates that are off by one or more categories as defined by
each individual vendor.
55
-------
4.3
Echologics LeakfinderRT
LeakfinderRT presented two sets of leak detection/location results: (1) for the simulated leaks, including
an estimated leak rate; and (2) for naturally occurring pipeline leaks.. Leakfmder RT first attempted to use
hydrophones in Pit 1 and Pit 3, which are located near each end of the test pipe, to assess the simulated
leaks in Pits 2, 4, and 5. However, detection of the relatively small calibration leak (0.6 gpm) over the
2,057 ft pipe length was not possible due to ambient noise masking the signal. Instead, LeakfinderRT
placed the hydrophones at approximately 1,000 ft intervals and collected leak rate data on three sections
of the test pipe (see Table 4-11). The three sections were chosen so that a pit containing simulated leaks
would be bracketed by the two hydrophone sensors. As shown in Table 4-12, LeakfinderRT used a
second configuration which involved placing accelerometers at much shorter distances to detect natural
leaks. The accelerometers were installed, with distances between sensors of approximately 250 ft to 361
ft, in Pits 1 through 3 and Pits A through F.
Table 4-11. Hydrophone-to-Hydrophone Distances for Detection of Simulated Leaks
Location
of Leak
(Pit#)
Pit 4
Pit5
Pit 2
Location of
Hydrophones
Pit 1 & Pit 2
Pit 2 & Pit 3
Pit 4 & Pit 5
Sensor-to-Sensor
Spacing (ft)
1,080.7
979.3
1,001.6
Table 4-12. Accelerometer-to-Accelerometer Distances for Detection of Natural Leaks
Location of
Leak (ft)
0-250
250-510
510-809
809-1,080
1,080-1,439
1,439-1,750
1,750-2,057
Location of
Accelerometers
Pit 1 & Pit A
Pit A & Pit B
PitB&PitC
Pit C & Pit 2
Pit 2 & Pit E
PitE&PitF
Pit F& Pit 3
Sensor-to-Sensor
Spacing (ft)
250
260
299
271
361
295
313
4.3.1 Summary of Results. LeakfinderRT reported three natural leaks and 7 simulated leak
clusters. Details of the natural leaks reported by LeakfinderRT (with accelerometers) are presented in
Table 4-13 with specific information on the distance from the contact point and estimated leak rate.
Details of the detected simulated leaks are presented in Table 4-14, specifically the corp valve ID and leak
rate reported by LeakfinderRT (with hydrophones) for each simulated leak.
Table 4-13. Natural Leaks Detected by LeakfinderRT with Accelerometers
Location of
Accelerometers
Pit A & Pit B
Pit F& Pit 3
Pit F& Pit 3
Upstream
Accelerometer
Location (ft)
250
1,750
1,750
Downstream
Accelerometer
Location (ft)
510
2,057
2,057
Distance
(ft)
260
307
307
Leak
Location
(ft)
341.5
1,912
1,930
Estimated
Leak
Rate
(gpm)
2.5-5.0
1.0-2.5
1.0-2.5
56
-------
Table 4-14. Simulated Leaks Detected by LeakfinderRT with Hydrophones
Pit#
Pit 4
Pit 2
Pit5
Corp
Valve
ID
CV1
CV2
CVS
CV4
CVS
CV6
CV7
CVS
Est. Leak
Rates
(gpm)
Dist. to
Leak
(ft)
Demo 1
(calibration)
Negligible
8.0
-
-
1077.1
Est. Leak
Rates
(gpm)
Dist. to
Leak
(ft)
Demo 2
Negligible
Negligible
5.0 to 8.0
-
-
1580.2
Est. Leak
Rates
(gpm)
Dist. to
Leak
(ft)
Demo 3
2.0 to 5.0
5.0 to 8.0
5.0 to 8.0
577.6
1082.2
1578.2
Est. Leak
Rates
(gpm)
Dist. to
Leak
(ft)
Demo 4
Oto 1.0
2.5 to 5.0
Negligible
560.7
1,092.6
-
4.3.2 Leak Evaluation. LeakfinderRT (with accelerometers) reported three natural leaks within
the inspected area. LeakfinderRT (with accelerometers) reported the largest natural leak at L4 (341.5 ft)
and another two leaks near 1,912 ft (L12) and 1,930 ft (L12), which were also found by Sahara® and
SmartBall™. However, it failed to identify natural leaks at bell-and-spigot joints near 53 ft, 195 ft, 556 ft
and 640 ft (i.e., LI, L3, L6 and L7), which were confirmed to exist. It is not clear whether LeakfinderRT
would have found these leaks had the larger leaks been repaired and their noise signatures removed.
LeakfinderRT (with hydrophones) reported 7 of 19 total simulated leaks and reported 7 of 11 leak
clusters. The simulated leaks were placed in clusters of one to three consecutive leaks, from orifices of
different sizes, arranged 0.6 to 2.7 ft apart. Within each leak cluster, LeakfinderRT accurately
characterized the leak range for 5 of the 11 clusters. For leak clusters not accurately characterized, 3 of 6
were below the 0.6 gpm threshold defined by LeakfinderRT during the demonstration. The other three
leaks not accurately characterized were off by approximately one size range. The location accuracy was
within 0 to 5 ft of the actual leak location, except for Demo 4 where the distances were off by a maximum
of 17 ft. As with the other technologies, LeakfinderRT was not able to discern two or three individual
leaks in close proximity (less than 2.7 ft apart) for any of the simulated leaks.
With a leak rate of 0.6 gpm and a sensor spacing of 1,080.7 ft or greater in the calibration run,
LeakfinderRT (with hydrophones) was unable to detect the leak during the correlation test. Although the
LeakfmderRT(with hydrophones) was unable to detect the small calibration leak, Echologics still
considered the calibration run to be successful, since it defined a boundary of leak rate-hydrophone
spacing beyond which leaks in the test pipe cannot be correlated with the existing equipment.
Table 4-15 summarizes the findings of the LeakfinderRT (with accelerometers) assessment of naturally
occurring leaks compared with the field verification results. Table 4-16 presents a summary of the
simulated leak results reported by LeakfinderRT (with hydrophones) compared with the actual leak rate
based on the orifice size and pipeline pressure at the time the inspection was conducted.
57
-------
Table 4-15. Evaluation of Natural Leaks Detected by LeakfinderRT with Accelerometers
22
EPA
Contractor
Leak ID
LI
L3
L4
L6
L7
L12
L12
Location of
Sensors
Pit 1 (0 ft) and Pit
A (250 ft)
Pit 1 (0 ft) and Pit
A (250 ft)
Pit A (250 ft) and
Pit B (5 10 ft)
Pit B (5 10 ft) and
Pit C (809 ft)
Pit B (5 10 ft) and
Pit C (809 ft)
Pit F (1,750 ft) &
Pit 3 (2,057 ft)
Pit F (1,750 ft) &
Pit 3 (2,057 ft)
Leak
Location
(ft)
Estimated
Leak Rate
(gpm)
No leak detected
No leak detected
341.5
2.5-5.0
No leak detected
No leak detected
1,912
1,930
1.0-2.5
1.0-2.5
Visually Verified by EPA Contractor?
Verification attempted; soil was wet, but there was a
nearby storm sewer at 52 ft; leak not pinpointed, but
elevated moisture indicative of leak.
Verification attempted; water in pit near bell-and-
spigot joint at -195 ft; did not pinpoint leak.
Verification attempted; leak at bell-and-spigot joint
-339 ft
Verification attempted; leak at bell-and-spigot joint
at -556 ft
Verification attempted; leak at bell-and-spigot joint
-640 ft
Verification attempted; soil was moist in this area
but could not locate the leak at -1,912 ft;
inconclusive
Verification attempted; soil moist at -1,906 ft; leak
not pinpointed; elevated moisture indicative of leak.
Table 4-16. Evaluation of Simulated Leaks Detected by LeakfinderRT with Hydrophones
23
Demo
No.
1
2
3
4
Pit
ID
Pit 4
Pit 2
Pit 4
Pit 2
Pit5
Pit 4
Pit 2
Pit5
Pit 4
Pit 2
Pit5
Corp
Valve
ID
CV1
CV3
CV2
CV4
CV6
CV7
CVS
CV1
CV2
CV3
CV4
CV6
CV7
CVS
CV1
CV2
CVS
CV7
CVS
Actual
Dist.
(ft)
577.4
1,082.2
578.4
1,082.8
1,084.9
1,583.0
1,585.7
577.4
578.4
,082.2
,082.8
,084.9
,583.0
,585.7
577.4
578.4
1,084.1
1,583.0
1,585.7
Actual
Leak
Rate
(gpm)(a)
0.57
7.8
0.06
0.14
0.57
1.0
7.8
1.0
1.9
0.14
0.57
4.6
7.8
0.57
0.06
4.6
1.0
.Jb)
1.9
LeakfinderRT
Leak Rate
(gpm)
Negligible
8.0
Negligible
Negligible
5.0 to 8.0
2.0 to 5.0
5.0 to 8.0
5.0 to 8.0
Oto 1.0
2.5 to 5.0
Negligible
Dist. (ft)
-
1,077.1
-
-
1,580.2
577.6
1,082.2
1,578.2
560.7
1,092.6
-
(a) Pressure assumed to be 54 psi for actual leak rate calculations.
(b) CV7 was closed for LeakfinderRT demo.
22 The gray background signifies natural leaks that were verified in the field (by pinpointing leak or elevated
moisture) and the red background signifies false negatives.
23 The text with a red background signifies leak rate estimates that are off by one or more categories as defined by
each individual vendor.
58
-------
4.4 Cost of Leak Detection/Location
The cost of leak detection/location has two main components: (1) the cost of the leak detection/location
service provided by the inspection vendor; and (2) the cost for the water company to prepare the line and
support the leak detection/location vendor, which is often more difficult to quantify.
4.4.1 Leak Detection/Location Services. The leak detection/location vendor's cost to conduct a
leak detection/location survey is dependent on a number of variables including the length and diameter of
pipe to be inspected, pipe accessibility, and types of services requested (some vendors offer volume
discounts for leak detection and condition assessment services). Costs usually include
mobilization/demobilization, inspection (per ft or mile), tap installation (if required), travel, and data
analysis and reporting. Inspection service providers will readily provide cost proposals for specific lines
to be inspected, however, it is rare that a water company will only inspect a short segment of pipe such as
the one used for this demonstration.
To supplement the cost information gathered for the demonstration, EPA's contractor also requested that
the vendors provide a cost estimate for inspecting 10,000 feet of 24-in. cast iron pipe along the same route
as the demonstration in Louisville, KY. They were asked to include in their cost estimates:
• The cost of conducting a leak survey alone
• The cost of conducting a pipe wall thickness assessment alone
• The cost of conducting both (leak and pipe wall thickness survey) at the same time.
Each vendor was given drawings of the 30-in. diameter pipeline that replaced the test pipe used for the
demonstration. The vendors were instructed that the pipeline for the cost estimate would follow the route
of the 30-in. line, but to assume that the line is 24-in. diameter and 10,000 ft in length.
To the extent possible, the vendors were asked to supply with their cost estimates:
• Costs for line modifications to perform the inspection (and who is responsible for the
modifications)
• Mobilization/demobilization costs
• Inspection costs (including data analysis and reporting)
• Factors that can affect pricing, such as diameter, length, risers, valves, bends, tees, insertions,
etc. and how these factors might impact the cost
Since some details regarding the pipeline and its location were not well defined, the vendor was informed
that a range of costs was acceptable.
PPIC Sahara®
For a 24-in. diameter, 10,000 ft long cast iron pipe, the cost estimates for a Sahara® leak and/or pipe wall
thickness inspections are provided in Table 4-17. Costs were not broken out by individual activity (e.g.,
data acquisition, data analysis, reporting, etc.). Charges for mobilization/demobilization are $4,000 while
data analysis and reporting are included in the price of the survey.
As reported by PPIC, each site inspection has different factors that may result in modification costs for
either the client or inspection vendor. Pipeline and operational parameters, such as pipeline length, access
preparation, features, flow condition, etc. can affect pricing. Proper pre-inspection preparation (drawings,
access preparation, flow rate control, etc.) by the client can significantly increase productivity while
59
-------
reducing the overall cost of the inspection. Inspecting longer lengths of pipe at the same time can benefit
from long-term program pricing discounts.
Table 4-17. PPIC Sahara* Cost Estimates for Inspection of a 24-in. Diameter,
10,000 ft Long Cast Iron Pipeline
Type of Survey
Leak and gas pocket survey (includes data acquisition, data
analysis, and final report)
Pipe wall thickness survey (includes data acquisition, data
analysis, and final report)
Leak and gas pocket AND pipe wall thickness survey
(includes data acquisition, data analysis, and final report)
Cost Estimate24
$22,000
$33,000
$44,000
TM
Pure SmartBall
Pure provided a range of costs to conduct three types of surveys: (1) a leak and gas pocket survey, (2) a
pipe wall thickness survey, and (3) both leak and pipe wall thickness surveys on one mobilization. Line
modifications would be required of the client to install two 4-in. taps, one at the beginning and one at the
end of the survey length. Pipeline flow would also need to be maintained between 1.5 and 2 ft/s and
pipeline pressure above 10 psi. Pure stated that it was possible to conduct a leak survey at lower pipeline
pressures, but the accuracy of the results could sometimes be compromised. Pure also stated that these
prices were to be used as a guideline and not as fact for inspection projects of this size.
For a 24-in. diameter, 10,000 ft long cast iron pipe, the cost estimates for a SmartBall™ inspection are
provided in Table 4-18. Costs were not broken out by individual activity (e.g., mobilization, data
acquisition, reporting, etc.). Charges for mobilization, demobilization, data acquisition, data analysis, and
final report run between $25,000 and $40,000 per inspection depending on which technology is used.
Technology charges run between $12,000 and $20,000 per mile of survey, again depending on which
technology is used.
Table 4-18. Pure SmartBall™ Cost Estimates for Inspection of a 24-in. Diameter,
10,000 ft Long Cast Iron Pipeline
Type of Survey
Leak and gas pocket survey (includes mob/demob, data
acquisition, data analysis, technology charges, and final
report)
Pipe wall thickness survey (includes mob/demob, data
acquisition, data analysis, technology charges, and final
report)25
Leak and gas pocket AND pipe wall thickness survey
(includes mob/demob, data acquisition, data analysis,
technology charges, and final report)
Cost Estimate24
$40,000 to $50,000
$55,000 to $65,000
$80,000 to $90,000
24 The Sahara®, SmartBall™, and LeakfmderRT cost estimates do not include utility preparation and support costs.
25 Pure requires feedback on the pipe wall assessment results from the Louisville demonstration before they are
comfortable quoting more specific cost estimates.
60
-------
This type of survey would require two days onsite, one to do a site review with the client and an actual
day of work with the tool in the pipeline. Pure can produce an on-site interim report and the final report
within two weeks of completing the survey. The interim report generated just after the survey, while the
field crew is still onsite would cost an additional $3,000 to $5,000.
Echologics LeakfinderRT
Echologics provided a fairly detailed cost proposal describing the work to be done for executing leak and
condition assessment surveys for a 24-in. diameter, 10,000 ft long cast iron pipeline. Preparation work
would be required by the client before the arrival of Echologics field technicians and includes:
• Assess traffic management requirements and prepare a traffic management plan.
• Identify confined space entry locations and provide a confined space entry plan and necessary
equipment.
• Identify all fittings to be used for the inspection and mark with blue spray paint or the
equivalent.
• All fittings should be in working order with no leaking seals or joints when under pressure.
Any leaking fittings must be repaired before the inspection. Failure to do so prevents
accurate data from being acquired in this location.
• Any valves installed on the pipe to be surveyed should be operated, if possible, to make sure
they are fully open. Any boundary/closed valves should be acoustic sounded to make sure
the valve is not passing water.
• Valve boxes, chambers, and vaults are to be cleared of debris prior to the inspection. Failure
to meet this requirement will prompt the need for an on-call VAC truck for the duration of the
project.
• Provide detailed maps, plans, and as-built drawings, if possible, showing all pipe fittings and
any other essential distribution information to establish a data acquisition plan.
• Provide all repairs and rehabilitation history, if possible, on the section of pipe to be
surveyed.
• Air must not be present in the main and all air relief valves must be in good working order
and inspected prior to the start of the survey. If air is present, flushing must be undertaken to
eliminate any trapped air.
• Pipe pressure must be maintained at a minimum working pressure of 25 psi with a maximum
pressure of 150 psi. Anything outside of these limits will require special consideration.
Echologics also requires the provision of an experienced water operator with a fully equipped truck for
the duration of the project. These requirements are necessary to accomplish the project within the
proposed timeline and budget.
For the leak detection survey, Echologics will mount hydrophones on air valves, pitot taps, or wash outs,
as available. The hydrophones require a 1.5-in. NPT male nipple for installation. Echologics can supply
all fittings if they are provided with the thread sizes. The sensor-to-sensor spacing shall not exceed 1,000
ft (305 m).
For the condition assessment survey, Echologics requires access to the pipe every 300 to 500 ft through
the use of vacuum excavated potholes. The potholes should measure 6 to 8-in. in diameter and provide
61
-------
access to the top of the pipe. Data acquisition will be performed using magnetic surface mounted sensors
attached to available fittings or the pipe surface. Fire hydrants will need to be flushed to take the water
temperature at each measurement site. Pipeline installation date and site-specific pipe manufacturer data
must be provided prior to field work.
Echologics provided cost estimates for mobilization, data acquisition, data analysis, and final reporting.
Mobilization includes all of the preparation work required by Echologics field technicians along with
travel and shipping expenses. Data acquisition will take approximately 3 to 5 days with two field
technicians. Generally, it is possible to cover between 2,500 ft and 5,000 ft of pipe per day. If any leaks
are discovered during the data acquisition process it will be the decision of the client as to whether or not
a detailed investigation will be performed to pinpoint the location of the leak. Data analysis includes the
time required to analyze the acoustic recordings upon completion of data acquisition using proprietary
processes. The analysis time will depend on the pipe size and total length of pipe surveyed. The final
report will summarize all of the results and include background, methodology, sources of error, data
interpretation methods, analysis, results, and final recommendations. A draft report will be submitted to
the client prior to its finalization.
For a 24-in. diameter, 10,000 ft long cast iron pipe, the cost estimates for a LeakfinderRT inspection are
provided in Table 4-19.
Table 4-19. Echologics LeakfinderRT Cost Estimates for Inspection of a 24-in.
Diameter, 10,000 ft Long Cast Iron Pipeline
Type of Survey
Leak detection survey
Mobilization
Data Acquisition
Data Analysis
Reporting
Total
Condition assessment and leak detection26
Mobilization
Data Acquisition
Data Analysis
Reporting
Total
Rate
$3,000
$1.25
$0.25
$165
$3,500
$1.50
$0.50
$165
Units
flat
ft
ft
hrs
flat
ft
ft
hrs
Cost Estimate26
$3,000
$12,500
$2,500
$2,310
$20,310
$3,500
$15,000
$5,000
$3,630
$27,130
For a leak detection survey only, Echologics estimated a total of three to four days onsite and an
additional 14 hours of data analysis and final report preparation. The rate proposed is based on a 10 hour
workday. If overtime is needed the client will be invoiced accordingly. A standby rate of $1,500 per
person per day is incurred if Echologics field technicians are delayed for reasons out of their control (not
including weather).
For a condition assessment and leak detection survey, Echologics estimated a total of four to five days
onsite and an additional 22 hours of data analysis and final report preparation.
' Leak detection is automatically performed during the condition assessment process.
62
-------
4.4.2 Site Preparation. The inspection costs presented above do not include the cost for the water
utilities to prepare the line and provide traffic control and other logistical support. This site preparation
cost for line modification and field support is highly site-specific. It will depend upon regional costs for
construction labor, along with factors such as the access requirements, availability and condition of
existing hydrants/valves, length of deployment, days on site, and more. Based on typical construction
costs (RSMeans, 2011), it is estimated that the site preparation costs for a leak detection inspection of
10,000 ft of 24-in. diameter cast iron pipe may range in magnitude from $0.12/ft (for traffic control only
with use of existing taps) to $0.43/ft (including traffic control, pit excavation, tapping, backfill, and
surface restoration).
During an inspection, SmartBall™ can be inserted into the pipeline through existing hydrants or any
valve configuration with greater than 4-in. diameter clearance. SmartBall™ is then retrieved through
another 4-in. or greater valve. For purposes of this cost estimate, it is assumed that the two required
access points must be installed for a 10,000 ft pipe inspection (e.g., no existing hydrants or valves are
used). Table 4-20 estimates the site preparation costs as approximately $4,290 based upon 2 access pits
and installation of two 4" taps for a SmartBall™ inspection (with pits located at 0 ft and 10,000 ft).
Table 4-20. Estimated Site Preparation Costs for SmartBall Inspection of 10,000 ft pipe
Cost Item
1
2
3
4
5
6
7
Setup Costs
2 - Rented 6 ft x 8 ft trench boxes
4-in. taps w/ valve and 150 Ib
standard flange with extension tube
1 CY of stone backfill
Traffic control
3 Persons - Labor (excavate, install
taps, backfill, restoration)
1 Person - Equipment Operator
(excavate, remove plates, backfill)
1 - 5/8 CY Wheel Mounted Backhoe
Quantity
2 boxes x 2 days =
4 days
2 taps
ICY
1 person x 2 days x
8 hrs/day = 16 hrs
3 persons x 1 day x
8 hrs/day = 24 hrs
1 person x 1 day x
8 hrs/day = 8 hrs
1 day
Unit
Cost
$93.00
$525.00
$46.50
$50.00
$52.70
$67.75
$215.00
Unit
4
2
1
16
24
8
1
Total
Total Cost
$372.00
$1,050.00
$46.50
$800.00
$1,264.80
$542.00
$215.00
$4,290.30
During a Sahara® inspection, a 1-in. diameter hydrophone is inserted into a live main through a 2-in. tap.
The maximum length of inspection is 6,000 ft based on the umbilical cable length. For purposes of this
cost estimate, it is assumed that two required access points must be installed for a 10,000 ft pipe
inspection (e.g., no existing taps are used). Table 4-21 estimates the site preparation costs as
approximately $3,933 based upon 2 access pits and the installation of two 2-in. taps for a Sahara®
inspection (with pits located at 0 ft and 6,000 ft).
63
-------
Table 4-21. Estimated Site Preparation Costs for Sahara® Inspection of 10,000 ft pipe
Cost Item
1
2
o
J
4
5
6
7
Setup Costs
2 - Rented 6 ft x 8 ft trench boxes
2-in. taps w/ valve and 150 Ib
standard flange with extension tube
ICY of stone backfill
Traffic control
3 Persons - Labor (excavate, install
taps, backfill, restoration)
1 Person - Equipment Operator
(excavate, remove plates, backfill)
1 - 5/8 CY Wheel Mounted Backhoe
Quantity
2 boxes x 2 days =
4 days
2 taps
ICY
1 person x 2 days x
8 hrs/day =16 hrs
3 persons x 1 day x
8 hrs/day = 24 hrs
1 person x 1 day x
8 hrs/day = 8 hrs
1 day
Unit
Cost
$93.00
$346.23
$46.50
$50.00
$52.70
$67.75
$215.00
Unit
4
2
1
16
24
8
1
Total
Total
Cost
$372.00
$692.46
$46.50
$800.00
$1,264.80
$542.00
$215.00
$3,932.76
Echologics mounts hydrophones on hydrants, air valves, pitot taps, wash outs, or other fittings as
available. The hydrophones require a 1.5-in. NPT male nipple for installation. The sensor-to-sensor
spacing for the hydrophones was 1,000 ft (305 m) for the Louisville, KY demonstration, so 11 contact
points with direct access to the water in the pipe would be needed in a 10,000 ft span. Fire hydrants are
typically located at spacings of 500 to 1,000 ft apart, so it is assumed that in atypical application where
Echologics was selected as an appropriate option that no excavation would be required and existing
fittings could be used. Traffic control for a 3-day inspection would be approximately $1,200.
64
-------
5.0: REFERENCES
Hunaidi, O., A. Wang, M. Bracken, T. Gambino, and C. Fricke. 2004. Acoustic Methods for Locating
Leaks in Municipal Water Pipe Networks. International Conference on Water Demand
Management, Dead Sea, Jordan, 1-14.
RSMeans. 2011. CostWorks® Cost Books Online, http://meanscostworks.com/. Norcross, GA: Reed
Construction Data.
Thomson, J. and Wang, L. 2009. Condition Assessment of Ferrous Water Transmission and Distribution
Systems: State of the Technology Review Report. EPA/600/R-09/055, prepared for U.S. EPA,
Office of Research and Development, National Risk Management Research Laboratory.
http://www.epa.gov/nrmrl/pubs/600r09055/600r09055.pdf
65
-------
APPENDIX A
THE PRESSURE PIPE INSPECTION COMPANY SAHARA® LEAK DETECTION* REPORT
(See Appendix A, pp. 4 to 7; 10 to!4; and 29 to 36)
A-l
-------
TABLE OF CONTENTS
1. EXECUTIVE SUMMARY 1
2. PROJECT BACKGROUND 2
2.1 Project Background 2
2.2 Purpose of Inspection 4
2.3 Test Pipe Line Description 4
3. SAHARA TECHNOLOGY 5
3.1 Background and Theory 5
3.2 Sahara Tests 10
3.3 Sahara Results 11
4. PIPEDIVER TECHNOLOGY 17
4.1 PipeDiver Background and Theory 1 7
4.2 PipeDiver Testing 20
4.3 PipeDiver Results 22
5. SUMMARY 28
5.1 Combined Test Results 29
5.2 Inspection Conclusions 30
5.3 Advantages and Limitations 31
5.4 Future Developments 32
6. PHOTOGRAPHS 33
-------
1. EXECUTIVE SUMMARY
Over the course of July 1 3th to 29th, 2009, the Pressure Pipe Inspection Company (PPIC)
performed non-destructive condition assessment of a cast iron main using two non-
disruptive inspection platforms, Sahara and PipeDiver. The assessment was conducted on a
2057 foot long, 24 inch diameter, cast iron section of the Westport Rd. Transmission Main
between Pit 1 (Launch/Insertion Pit) and Pit 3 (Receive/Extraction Pit).
PPIC used its patented Sahara Technology, including Sahara Leak Detection, Sahara Video,
and Sahara Wall Thickness Testing. In addition, PPIC conducted a Remote Field Eddy Current
(RFEC) pilot test for metallic pipe wall condition assessment using the PipeDiver inspection
platform. Both technologies are non-disruptive and allow the pipeline to remain in service
during the inspection. PPIC's inspections are part of a study conducted by the U.S.
Environment Protection Agency (EPA).
Sahara Leak Detection identified six natural leaks and an air pocket within the inspected
area and detected all simulated leaks. Sahara Video identified several corrosion spots,
outlets, and air pockets within the pipeline. Analysis of the Sahara Wall Thickness Testing
data revealed several areas of suspected wall thickness loss. PipeDiver RFEC testing was
performed over the full scope (2057 ft) under live conditions and identified 41 pipe sections
with anomalous data signals. Verification and further calibration are recommended to
confirm the exact nature of these anomalies and help in further refinement of the PipeDiver
analysis procedures. Each individual technology provides a particular service but their
combined results provide a complete overall condition assessment of the pipeline.
-------
2. PROJECT BACKGROUND
2.1 Project Background
The U.S. Environmental Protection Agency (EPA) contracted the Battelle Memorial Institute
(BMI) to demonstrate selected innovative leak detection/location and structural condition
assessment technologies. This study emphasizes the need for non-invasive, non-
destructive, "inexpensive" techniques to help utilities assess the condition of their lines to
allow them to make good decisions regarding capital replacements, rehabilitation or
monitoring of their pipe infrastructure.
The Pressure Pipe Inspection Company (PPIC) is one of the several companies contracted by
BMI to demonstrate their non-destructive condition assessment techniques of metallic
pipes. These include PPIC's patented Sahara Leak Detection, Sahara Video, Sahara Wall
Thickness Testing and PipeDiver RFEC Testing. All these technologies are invasive,
requiring internal pipe access, but are non-disruptive in nature and are performed while
the pipeline is in service. Each technology has its own set of advantages and limitations
which allows utilities an option on which inspection technique best fits their needs and
expectations. Additionally, multiple techniques can be applied to a single pipeline to
provide successive levels of detail about the pipe condition.
The condition assessment technologies deployed by PPIC are at various stages of
commercial deployment. The Sahara leak detection system, for example, has been
successfully used commercially worldwide for over 1 0 years. While PipeDiver has been
successfully used in PCCP for live condition assessment, PipeDiver RFEC for metallic pipes
is still undergoing development and in the process of becoming a commercially available
service.
The Westport Rd. Transmission Main is a 24 inch diameter cast iron pipe that has been
taken out of service. EPA has acquired this pipeline for a non-destructive condition
assessment study, which PPIC is a part of. A map showing the approximate location of the
inspected pipeline is shown in Figure 2.1.
-------
Figure 2.1 Pipeline Plan
Westport Road
Transmission Main
Additional features were created along the inspection scope for various test procedures.
These features are listed in Table 2.1 (distances provided by the Battelle Memorial
Institute).
Table 2.1 Feature List
Feature
Pit 1 (Launch/Insertion Pit)
Pit A
Pit B
Pit 4
Pit C
Pit 2
Pit D
Pit E
Pit 5
Pit F
Pit 3 (Receive/Extraction Pit)
Distance from Pit 1 (ft)
0
250
510
581
809
1080
1173
1439
1580
1750
2057
STA
160+55
163+05
165+65
166+36
168+64
171+35
172+28
1 74+94
176+35
178+05*
181+12
*Approximate STAs are in relation to fire hydrant STA of 178+05 (hydrant listed in same location as
Pit F from Battelle chart).
-------
2.2 Purpose of Inspection
The purpose of this inspection is to demonstrate PPIC's various non-destructive condition
assessment services on metallic pipe which, together, provide an overall condition
assessment of the pipeline. These services include:
• A visual inspection of the inside of the pipeline
• Identifying and quantifying the presence of leaks
• A pipe wall assessment including wall thickness loss and irregularities
All services are performed using PPIC's patented Sahara technology platform and the
PipeDiver platform, both of which are live inspection platforms that operate while the
pipeline is in service.
2.3 Test Pipe Line Description
The non-destructive condition assessment inspections of the Westport Rd. Transmission
Main were conducted from July 1 3th to 29th, 2009. The test details are summarized in Table
2.2.
Pipeline
Inspection Dates
Total Distance
Table 2.2 Test Summary
Westport Rd. Transmission Main
July 13th to 29th, 2009
2057 feet
In order to produce sufficient flow in the pipeline for inspection purposes a 1 2 inch tee
past the extraction point was used to temporarily create flow by diverting water into a
nearby storm drain.
Figure 2.2 Pipeline Flow Setup
Extraction
Point
Storm Sewer
Diversion
Insertion
Point
_F i Y
Flow
The flow amount and duration was limited by the capacity of the storm sewer. In the event
of rain, the storm sewer's capacity would be reduced or eliminated entirely which, in turn,
would likewise affect the flow available in the 24 inch cast iron line.
-------
3. SAHARA TECHNOLOGY
3.1 Background and Theory
3.1.1 Sahara Platform
The first tool designed for live inspection of large diameter water mains, the Sahara
Pipeline Inspection System, is capable of detecting leaks, pockets of trapped gas, and
structural defects in large mains. Sahara is a critical component of condition assessment
and water loss management programs for utilities around the world. The unique Sahara
platform allows adaption of multiple technologies such as leak detection, video inspection,
and wall thickness assessment.
Advantages to the Sahara inspection system include:
• No disruption to pipeline service
• Use existing 2 inch (50 mm) taps
• A tethered system allows complete control of the sensor's position along the pipe
and ensures no lost sensors
• Accurate surface tracking to map pipelines and leak locations
• Usable in mains of all material types, as small as 4 inches in diameter, and with
pressures up to 200 PSI
3.1.2 Sahara Leak Detection
The Sahara system is a non-destructive condition assessment technology that pinpoints the
location and estimates the magnitude of leaks in large diameter, 1 2 inch and above, water
transmission mains of all construction types. With over 1,000 miles (1,600 km) of
inspections Sahara Leak Detection has proven sensitive to leaks as small as 0.005 gal/min
(located in 72" PCCP at 87 psi). Leaks are located above ground in real-time and marked to
within 1 foot of accuracy.
In operation, the system is inserted into a live pipeline through any tap that is at least 2
inches in diameter. Carried by the flow of water, the tethered sensor head can then travel
through the pipe for distances up to 6,000 feet per survey detecting each leak as it is
found. The leak's position is then located and marked on the above ground surface
facilitating subsequent repairs.
An electronics processing unit with audio and visual output is used for data analysis. A leak
produces a distinctive acoustic signal which is recorded by the sensor and processed into a
visual signal. The visual signal is then analyzed along with the audio signal to quantify the
leak.
-------
In no flow situations a second tethering line (mule tape) can be used to pull the
hydrophone through a pipeline.
Figure 3.1 Sahara Inspection System
Cab* Drum
(tnDI*lor*et«nlBath>
LocBtmgTooi SionaJi
Proc**! ing Unrt
{V»dto.f Audio)
An operator stands by at the controller station to control hydrophone deployment and
listen to the hydrophone signal for leaks in real time. Once a leak is detected the
hydrophone can pass over the leak multiple times to classify and pinpoint the leak. A
second operator travels the pipeline above ground using a tool to detect the exact
location of the sensor. When a leak is detected this operator will make a mark on the
ground identifying the location and record a GPS point for reference.
The capable survey length of the Sahara system is limited not only by the amount of
available cable, usually 1.2 miles (2 km), but also by the pipeline geometry
(horizontal/vertical elbows and bends), the pipeline flow rate, and the internal pipe
conditions.
Sahara Leak Detection is a proven technique in identifying the smallest leaks in pipelines.
Figure 3.2 below depicts some verified leaks and the corresponding pressures the leaks
were detected at.
-------
Figure 3.2 Sahara Verified Leaks
Sahara Verified Sensitivity to Small Leaks
- Estimated Seepage Threshold
Sahara Verified Small Leaks
10 20 30 40 50 60 70
90 100 110 120
Pressure (PSI)
Calibration is performed by testing each hydrophone and comparing it to a standard
frequency response. The Sahara hydrophone has sensitivity to leaks as small as 0.005
gal/min (detected on 48" PCCP pipeline at 87 psi).
Data is interpreted and analyzed in real time by on screen spectrogram and audio
listening. Using dual analysis methods provides high accuracy and can clearly
distinguish leaks from ambient noise.
Factors such as low water pressure, electrical noise, air pockets, and external ambient
noise can all affect the real time analysis of the sensor signal. During the inspection,
some leaks were masked by external factors and required post analysis to detect the
leaks.
3.1.3 Sahara Video Description
Sahara Video provides real time, in-service CCTV inspection through a 2 inch or larger tap.
Real-time video inspection enables visual inspection of features including:
-------
• Cement and other liners
• Internal corrosion and tuberculation assessments
• Valve location and inspection
• Debris and blockages
Figure 3.3 Sahara Video Head
The Sahara video system utilizes the same
control system and tethered cable as the
Sahara Leak Detection system but the
hydrophone sensor head is switched to a
video camera head that traverses a pipeline
after begin inserted through a standard 2
inch tap. A drogue (parachute) is attached
just behind the camera which captures
water flow and carries the camera and
cable down the pipeline.
An operator stands by at the controller
station to control camera deployment and
views the video output in real time. A
second operator traverses the pipeline above ground using a tool to detect the exact
location of the camera. When an item of interest is seen the second operator will make
a mark on the ground identifying the location and record a GPS point for reference.
Like the Sahara leak detection, the Sahara video system has a limited survey length
from the pipeline configuration and available flow rate. One circumstance or factor
affecting accuracy is video clarity. Video image becomes less clear in larger diameter
pipes, due to diffuse lighting and reduced field of view, and unclear water. To calibrate
the video system, each video camera is tested and compared to a standard frequency
response. Video is interpreted and analyzed in real time, but also recorded for future
examination.
3.1.4 Sahara Wall Thickness Testing
Sahara Wall Thickness Testing can be performed in conjunction with a Sahara Leak
Detection inspection. Testing requires a secondary acoustic sensor, either an external
accelerometer attached to the pipe surface or an additional internal hydrophone.
Reference signals (e.g., test strikes at access points or sounds produced by a speaker)
are generated within the pipe for testing.
The sound waves propagate through the pipeline in a specific manner bouncing
repeatedly off of the pipe walls. As the sound wave travels in this manner they gather
information about the pipe wall. By measuring the speed of sound multiple times in a
section of pipe the average wall thickness can be deduced. By using multiple acoustic
sensors separated by a known distance time of arrival data from the reference signal
can be used to calculate the speed of sound within the pipe and thus the average wall
thickness.
-------
Figure 3.4 Sahara Wall Thickness
Accelerometer
and Acoustic Unit
for Reference
Reference
Signal
Sahara
Hydrophone
Time Delay
The tethered control of the Sahara system allows the hydrophone to stop at precise
locations for each interval. Time of arrival data is then used to calculate the average
wall thickness over each interval. Since the wall thickness average intervals are defined
by hydrophone location there are infinite interval possibilities limited only by the
amount of time and resources available for the inspection.
Sahara wall thickness has the same limitations on survey as the leak detection system.
Also like the leak detection, air pockets can significantly interfere with the wall
thickness measurements as they affect the acoustic signal propagation. It is important
to note that the wall thickness measurements resulting from this technique are only an
average thickness over a range of pipes
Average wall thickness results need detailed pipe information and fluid parameters for
calculations. Current testing procedure requires an access (i.e. hydrant, flange, or exposed
pipe surface) a minimum of every 400 feet to generate reference acoustic signals.
Some factors affecting wall thickness accuracy include:
• Distance of a given section (the shorter, the more uncertain)
• Distance readings of the sections
• Accuracy of the pipeline and fluid parameters
• Unknown pipe features
• Rehabilitation, or large stationary air pockets
However, many pipeline related factors can be eliminated through a repeat inspection.
-------
Before each Sahara Wall Thickness test adequate calibration and preparation is performed
to ensure high quality. This includes:
• Calibration of Sahara sensor's sensitivity and distance reading
• Calibration of reference acoustic sensor for synchronization with Sahara
• Repeatability tests
A relative result is obtained based on all calculated results in every 30 foot interval. A
nominal pipe wall thickness would be calculated from a group of intervals that shows
similar wall thickness results (< 2% difference from the mean), and the result of other
portions would show the wall thickness change ratio to this nominal value. This relative
result is provided instead of calculated wall thickness to eliminate and minimize possible
uncertainties introduced by composite pipe material and alterable fluid parameters.
3.2 Sahara Tests
3.2.1 Sahara Test Schedule
A total of five Sahara insertions were performed from July 1 3th to July 1 7th for all the
different inspection technologies. The Sahara video inspection was performed first, on July
1 3th, to inspect the inside of the pipeline. This inspection identifies potential obstacles for
other internal inspections as well as internal corrosion and air pockets. The Sahara video
head was inserted into Pit 1 and traversed the line using the pipeline flow. After reaching
Pit 3 the video head was then retracted and taken out of Pit 1.
Sahara Leak Detection was performed on July 14th, 1 5th, and 1 7th. Three full surveys of the
pipeline were performed to test different arrangements of simulated leaks and perform a
repeatability survey under varying conditions. Like the Sahara video head, the Sahara
sensor head was inserted and retracted out of Pit 1. The leak detection survey was
conducted during the deployment and retrieval of the sensor through the pipeline. On July
1 5th a thunderstorm required that flow in the pipeline be stopped due to reduced storm
sewer capacity and the survey ended before completion.
Sahara Wall Thickness Testing was performed on July 1 5th and 16th in conjunction with
Sahara Leak Detection. The Sahara sensor head was inserted into Pit 1 and secondary
external sensors were installed at Pits A, C, E, and 3.
10
-------
Multiple test reference signals were generated at each of the pits to conduct the wall
thickness measurements.
Table 3.1 Insertion Details
Date
July 13th
July 14th
July 15th
July 16th
July 17th
Insertion
Point
Pit 1
Pit 1
Pit 1
Pit 1
Pit 1
End Point
Pit 3
Pit 3
After Pit F
Before Pit 3
Pit 3
Survey
Length (ft)
2057
2050
1797
1984
2050
Flow
Direction
East
East
East
East
East
Description
Video
Leak Detection & Leak
Simulations
Leak Simulations
Wall Thickness
Repeat Leak Detection,
Simulations & Wall
Thickness
3.3 Sahara Results
3.3.1 Sahara Video Survey Results
The Sahara Video inspection of Westport Rd. Transmission Main successfully identified
several significant observations. Details of the observations are presented in Table 3.2,
specifically the direction and distance the observation was found from the insertion point
(Pit 1).
Table 3.2 Observation Details
#
1
2
3
4
5
6
7
8
9
10
11
12
Description
Outlet
Outlet
Air pocket
Large air pocket
Outlet
Large air pocket
Outlet
Corrosion
Outlet
Large area of corrosion
Outlet
Outlet
Estimated Distance
from
Pit 1 (ft)
154
677
886
1024
1061
1237
1552
1565
1628
1637
1755
1946
Direction from
Insertion
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Potential
Correlated Pipe
Feature
Pit 2 (1080ft)
Pit 5 (1 580 ft)
Pit F (1750ft)
11
-------
Many additional air pockets, ranging from small to large in size, were discovered
during the video inspection. Both air pockets and wall corrosion could be clearly
distinguished in the video inspection.
Figure 3.5 Sahara Video Examples
Close-Up of a Joint Cap at the Insertion
Example of an outlet
Extraction Point, 24x24x1 2" Tee
Example of Large Air Pocket
3.3.2 Sahara Leak Detection Results
The Sahara Leak Detection of Westport Rd. Transmission Main successfully identified 6
natural leaks and 14 simulated leaks. Details of the natural leaks are presented in
Table 3.3, specifically the direction and distance the leak was found from the insertion
point. The most accurate method to locate a leak is from the mark created above
ground by the inspection team during the survey.
12
-------
Table 3.3 Natural Leak and Air Pocket Details
Leak*
1
2
3
4
5
-
6
7
Feature
Very Small Leak
Very Small Leak
Large Leak
Very Small
Small Leak
Large Air Pocket
Very Small Leak
Small Leak
Distance from Pit 1
(ft)
50
194
338
558
638
900
1696
1906
Direction from
Insertion Point
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Downstream
Simulated leaks were rearranged several times. Details of the detected simulated
Leaks are presented in Table 3.4, specifically the arrangement number, direction,
and location.
Table 3.4 Simulated Leak Details
Arrangement #
1*
1
2
2
2*
3
3
3
4
4
4*
5
6
7
Date
July 14th
July 14th
July 14th
July 14th
July 14th
July 1 5th
July 1 5th
July 1 5th
July 1 5th
July 1 5th
July 1 5th
July 1 7th
July 1 7th
July 1 7th
Leak Classification
Very small
Small
Large
Very small
Very small
Small
Small
Medium
Medium
Small
Very small
Small
Very small
Very small
Location
Pit 4
Pit 2
Pit 5
Pit 2
Pit 4
Pit 4
Pit 2
Pit 5
Pit 5
Pit 2
Pit 4
Pit 4
Pit 4
Pit 4
These leaks required post analysis. Leak signal could be masked by air pockets, water discharge,
and/or electrical issues.
13
-------
Figure 3.6 Examples of Sahara Leak Signals
jignal Generated by a Small Leak
Signal Generated by 3 Large Leak
Gane a:ed ny a Medium Leak
Generated by the Discharge Line
3.3.3 Sahara Wall Thickness Results
The Sahara Wall Thickness Assessment of Westport Rd. Transmission Main successfully
identified specific areas of wall thickness loss. Details of the wall thickness loss are
presented in Table 3.5, specifically the pipeline interval and average result over that
interval.
Table 3.5 Wall Thickness Details
Distance from Pit 1 (ft)
0-17
17-33
33-66
66-98
98-131
131-164
164-197
Average Wall Thickness
Loss Ratio (%)
N/A
< 15%
Nominal
< 15%
Nominal
Nominal
Nominal
14
-------
197-230
230-295
295-328
328-361
361-394
394-426
426-459
459-492
492-525
525-558
558-590
590-623
623-656
656-689
689-722
722-754
754-787
787-1640
1640-1673
1673-1706
1706-1738
1738-1771
1771-1804
1804-1837
1837-1870
1870-1902
1902-1935
1935-2057
15-30%
N/A
>30%
> 30%
>30%
Nominal
< 15%
15-30%
< 15%
< 15%
< 15%
Nominal
< 15%
Nominal
15-30%
15-30%
Nominal
N/A
Nominal
Nominal
< 15%
< 15%
< 15%
< 15%
Nominal
Nominal
15-30%
N/A
Pipeline intervals with an average wall thickness loss of less than 2% are listed as
nominal. The average wall thickness loss ratio is in relation to the nominal mean value.
The section from 295 to 328 feet shows the highest wall thickness loss.
Increased error margin in the section from 230 to 295 feet is due to the close
proximity of internal and external sensors. Subsequently, a wall thickness loss ratio
cannot be calculated for this interval. From 787 to 1 640 feet a wall thickness ratio
cannot be calculated due to presence of large air pockets and/or the proximity of
sensors. The pipeline discharge masked acoustic activity after 1 935 feet.
15
-------
4. PIPEDIVER TECHNOLOGY
4.1 PipeDiver Background and Theory
4.1.1 PipeDiver Platform
The PipeDiver system has been specifically designed for use in pipelines that are live or
can not be taken out of service due to lack of redundancy or operational constraints.
PipeDiver provides accurate condition assessment of critical infrastructure, specifically
detecting prestressing wire breaks in Prestressed Concrete Cylinder Pipe (PCCP). This
solution offers significant cost savings as the pipeline remains in service eliminating
the need for service shutdown and dewatering. The system has been proven effective
for the inspection of live PCCP lines from the verification of its pilot inspection of 30
inch diameter pipe in Halifax in 2007.
PipeDiver is a non-tethered, free swimming inspection platform for in-service water
mains. The inspection vehicle allows inspection of pipelines from 24 inch in diameter
and larger through two 1 2 inch diameter taps installed on the pipeline, one at each
end of the inspection region. Alternatively, reservoirs or open channels can be used as
insertion and extraction points.
Figure 4.1 The PipeDiver Inspection System
For a standard launch the insertion tube containing the PipeDiver vehicle is attached to
the 1 2 inch tap before being filled with water, pressure equalized, and opened to the
pipeline. The internal insertion piston pushes the PipeDiver vehicle into the pipe and,
once fully in the pipe, the vehicle is released and begins to travel with the flow. For a
standard retrieval, once the PipeDiver vehicle reaches the extraction side, a robotic
16
-------
claw and net which blocks the entire pipe diameter grabs the front of the vehicle and
secures it before pulling up out of the pipe and into the retrieval tube.
The PipeDiver vehicle travels at
approximately 90% of the pipeline's
flow rate, the neutrally buoyant
inspection vehicle can run for up to
30 hours in a single insertion.
Flexible fins are used to center the
tool within the pipe and provide
propulsion. Its flexible design
ensures that PipeDiver can navigate
through most butterfly valves and
bends in the pipeline while travelling
long distances.
Figure 4.2 PipeDiver Retrieval Arm
Figure 4.3 The PipeDiver Inspection System
Inspection
Extraction
Catch Point
Electronics
Module
Battery
Module
Transmitter
Module
The PipeDiver inspection tool is inserted into a live main through a 1 2" tap directly on
top of the main, then retrieved using a robotic arm inside a similar chamber at the end
of each inspection run. The modular system includes an electronics module, battery
module, and transmitter module for above ground tracking.
17
-------
4.1.2 PipeDiver RFEC Testing Description
The Remote Field Eddy Current (RFEC) is a proven technique for non-destructive
inspection of metallic pipelines. The PipeDiver is similarly a proven platform for
insertion into live pipelines and inspection using the RFTC technique. While the RFTC
and RFEC techniques are similar in nature there are several challenges involved in
modifying the PipeDiver platform to support RFEC technology:
• Detectors have to be closer to the wall
• More detectors are required
• Signal levels are significantly lower than RFTC
• Exciter to detector axial separation is much larger
To modify the PipeDiver for a RFEC inspection the exciter coil was moved from the rear
body near the center detector into the first body to achieve the minimum 1.5-2 pipe
diameters required for the RFEC technique. Six additional detector coils were added to
petals at the rear of the vehicle to provide increased sensitivity to wall thickness loss
while still permitting the the vehicle to be inserted and extracted through a 1 2 inch
diameter opening.
Figure 4.4 PipeDiver Coil Locations
Coil Setup for the RFTC Inspection Technique
Coil Setup for the RFEC Inspection Technique
18
-------
The future challenges for PipeDiver RFEC development will be to increase the number
of detectors close to the pipe wall, especially for larger diameter pipes, to increase the
resolution and accuracy of the wall thickness measurements.
Common factors affecting accuracy for any RFEC system include the pipeline
design and composition (i.e. metallic variations), inspection tool calibration,
inspection tool riding quality, the type and position of the defect. Calibration
details include running standard RFEC tests (with various coil separation/frequency
setups) on pipes with a set of defects (size and shape) to achieve the best detection
and sensitivity.
4.2 PipeDiver Testing
4.2.1 PipeDiver Inspections
PipeDiver RFEC Testing and trials were performed from July 21st to July 29th and four
successful runs were completed. This was a pilot inspection using the RFEC technique
in metallic pipe to obtain additional field data for analysis.
Table 4.1 shows the details of actual inspections, specifically the survey length and
description of the inspection.
Table 4.1 Insertion Details
Date
July 23rd
July 24th
July 27th
July 28th
Insertion
Point
Pit 1
Pit 1
Pit 1
Pit 1
End Point
Pit3
Pit3
Pit3
Pit 3
Survey
Length
(ft)
2057
2057
2057
2057
Flow Direction
and Speed
East, 1 ft/sec
East, 0.5 ft/sec
East, 1 ft/sec
East, 1 ft/sec
Description
PipeDiver RFEC
PipeDiver RFEC
PipeDiver RFEC
PipeDiver RFEC
4.2.2 PipeDiver Insertion Issue
On July 21st, the first insertion attempt, the PipeDiver vehicle became stuck during the
insertion process and that day's inspection had to be stopped and the vehicle retrieved
from the pipe. An investigation of the issue with the help of Sahara video (Figure 4.5
and 4.6) led to the conclusion that the front of the PipeDiver has become stuck in a
large, unfilled gap estimated to be 3 to 4 inches in width between joints just
downstream of the insertion point.
19
-------
Figure 4.5 Sahara Video of the Joint Gap
Tee
Estimated
Axial Distance
= 3-4"
Cast Iron
Figure 4.6 PipeDiver Insertion Schematic
PipeDiver
Insertion Tube
PipeDiver
Insertion Tube
12" HOI Tap
Original PipeDiver Design
Using Standard Hot Tap
24x24x12° Tee
Westport Rd. 24" Cast Iron Pipeline
Using a 24x24x12" Tee
An alternate insertion process was designed and implemented and the following four
insertions were successful.
PipeDiver is designed for live inspections using standard accesses including 1 2 inch
diameter hot taps, tees with minimum joint gaps, or similar features. For certain
accesses such as tees with large unfilled joint gaps or accesses with unknown internal
conditions Sahara Video is recommended to identify the exact layout of the insertion
point. The insertion design and process can then be modified for a successful insertion
if required.
20
-------
4.3 PipeDiver Results
4.3.1 PipeDiver RFEC Result Description
PipeDiver RFEC Testing was conducted as a pilot project to obtain field data for
analysis. Data was analyzed and characterized based on basic pattern recognition
from simple models of wall thickness variations.
Remote Field Eddy Current works on the basic theory that when a time harmonic
magnetic field is generated inside a metallic pipe it has two paths from the exciter to
detector coils (see Figure 4.7).
Figure 4.7 RFEC Signal Paths
The direct path remains inside the pipe and couples the coils directly while the remote
path remains outside of the pipe as long as possible. When the exciter-detector coil
separation exceeds 1.5 pipe diameters the signal from the remote field significantly
dominates the total signal received at the detector. Since the remote field path has
passed twice through the pipe wall any variation in magnetic wall properties including
wall thickness, conductivity, and magnetic permeability will result in a change in the
detector signal.
4.3.2 PipeDiver RFEC Results Overview
Table 4.2 lists the location of pipe sections PipeDiver data characterized as anomalous
and their distance from Pit 1.
21
-------
Table 4.2 PipeDiver Anomalous Pipes
Distance from Pit 1 (ft)
Start
216
264
276
324
360
384
444
504
516
576
612
864
936
948
1044
1056
1176
1212
1284
1308
1332
1356
1368
1416
1452
1512
1584
1608
1620
1644
1656
1704
1740
1752
1788
1812
1860
1872
1908
1956
1992
End
228
276
288
336
372
396
456
516
528
588
624
876
948
960
1056
1068
1188
1224
1296
1320
1344
1368
1380
1428
1464
1524
1596
1620
1632
1656
1668
1716
1752
1764
1800
1824
1872
1884
1920
1968
2004
22
-------
4.3.3 PipeDiver RFEC Pipe Signals
Figure 4.8 below shows the center detector signal amplitude (red) and phase (green)
from the July 23rd inspection of a section of pipeline which is classified as containing
normal pipes.
Figure 4.8 PipeDiver RFEC Nominal Pipes
v/
240 26D 280 300 320 340 360 380 400
Each joint is composed of a double signal due to the remote field effect. One signal is
from the exciter passing the joint and one from the detector passing. The first signal in
a joint is generally higher and longer due to the relative lengths of the pipe and axial
exciter-detector coil separation, 1 2 and 5.5 ft respectively (Figure 4.9).
23
-------
Figure 4.9 PipeDiver RFEC Joint Detection
I
1130 il
imn
-------
a comparison for a section of four pipes from the center detector. One of the
objectives of this inspection was to verify the validity of the PipeDiver RFEC technology
by performing such repeatability tests. The results from the multiple PipeDiver scans
show good repeatability.
Figure 4.11 RFEC Repeatability
0)
I
"o.
E
D)
co
Pipe 135
Pipe 136
1060 1070 1080 1090
Time (seconds), July 23rd
1100
1110
A known feature from the pipeline that is readily seen in the PipeDiver RFEC data is the
hydrant outlet that is located near Pit F (Figure 4.1 2). While the signal is relatively small
as compared to the joint signal it can be distinguished by having a double signal
occurring the exact distance as the PipeDiver's detector-exciter coil separation
distance.
25
-------
Figure 4.12 PipeDiver RFEC Hydrant Signal
0
T5
3
c
g;
CO
2960
2980 3000
Time (seconds), July 23rd
3020
Four new defects were machined into Pit F on July 28th (Figure 4.1 3). By comparing the
RFEC signals from the data before and after the defects were created we have the best
possible chance of seeing this relatively small amount of wall thickness loss in the data
(Figure 4.1 4).
26
-------
Figure 4.1 3 New Pit F Defects
Figure 4.14 Comparing RFEC Data Before and After
Defects
Befo
Aft.
fore Pit F Defects \
:er Pit F Defects
2940
2950
2960
Time (seconds), July 23rd
2970
2980
27
-------
The PipeDiver RFEC results show good repeatability between multiple scans using the
same configuration which validate it as a non-destructive inspection technique. The
RFEC data clearly shows joint signals, known features and anomalous signals which
may be potentially due to wall thickness loss. Further verification and calibration is
needed to confirm the nature of these anomalous signals.
5. SUMMARY
5.1 Combined Test Results
The following figure 5.1 combines all results including Sahara Leak Detection, Sahara
Video, Sahara Wall Thickness, and PipeDiver RFEC, showing their relative locations
along the pipeline.
28
-------
Figure 5.1 Combined Results
Nominal Wall Thickness
< 15% Wall Loss
16-30% Wall Loss
=•30% Wall Loss
A Leak Positions
+ Video Features
X PipeDiver Anomalies
I
IA
I.
I-
XX
X X
600
sou
XX
I x>i
X X
X X X XX X X
x x;
1500
1800
rn
lx
>
A
X
o
!x x
1800 1900
Distance (ft)
29
-------
The combined results make it easier to identify potential areas of interest within the
pipe. For example, the section between 300 to 400 ft contains a large leak, several
PipeDiver RFEC anomalies and has a high average wall thickness loss and is one of the
areas recommended for further verification and calibration. Similarly, the area between
1 560 to 1 640 ft contains several identified corrosion spots and PipeDiver RFEC
anomalies.
5.2 Inspection Conclusions
PPIC's evaluation of the Westport Rd. Transmission Main between Pit 1 and Pit 3 (2057
foot section) provided an overall condition assessment of the metallic pipeline.
The Sahara platform was used to provide three critical non-destructive condition
assessment services, including:
• Internal video inspection
• Leak detection
• Sahara and PipeDiver wall thickness assessment
All Sahara services were successfully inserted using a 2 inch tap in live conditions not
requiring the line to be shut down. The tethered system allowed the sensor to be
stopped at precise locations which enabled operators to make accurate and repeatable
identifications regarding pipeline condition discoveries.
Sahara Leak Detection detected six unidentified leaks and one air pocket, recorded and
marked their above ground position, and estimated the leak size all in real time.
Several simulated leaks were also detected in real time, and post analysis was able to
identify all leaks that had been masked by external noise factors such as the pipeline
discharge.
Sahara Video's tethered CCTV inspection was also successfully deployed using a 2 inch
tap. Real time analysis of the video provided insight into the internal condition of the
pipeline and clearly distinguished two areas of corrosion. Air pockets and outlets were
also clearly identifiable from the real time inspection. The second purpose of a video
inspection, to discover possible obstacles for a PipeDiver inspection, showed that
PipeDiver could be used with no risk from unidentified obstacles. Video recordings
were used for post analysis and helped identify a previously unknown risk: a joint gap
just downstream of the insertion point. These video results can now be used to
improve and change aspects of the PipeDiver system.
Sahara Wall Thickness was performed in conjunction with leak detection thus
minimizing extra resources and time. Analysis of the results uncovered specific
intervals of the pipeline showing higher wall thickness loss than others. By utilizing the
tethered Sahara system and being able to stop the hydrophone at precise locations,
consistent and multiple pipe intervals could be set to calculate average wall thickness
readings.
The PipeDiver platform is poised to becoming the industry standard for in-service
pipeline inspections. The technology can be modified for different services and
30
-------
eliminates the need to take pipelines out of service during inspections. PipeDiver was
successfully inserted and retrieved via two 1 2 inch Tees installed into the live main.
Results obtained form the Westport Rd. Transmission main inspection have identified
anomalous signals and processes that will allow PPIC to further improve the PipeDiver
system, specifically RFEC Testing.
5.3 Advantages and Limitations
The significant advantage to the overall Sahara inspection technologies is that its
tethered cable design brings the sensor as close as possible to the leak and allows
unlimited control of the sensor position. For Sahara Leak Detection this means that the
farthest the hydrophone sensor will be from a leak is the pipe diameter, or more
realistically the pipe radius, which permits very small leaks to be detected. Leaks are
detected in real time and immediately accurately located and marked above ground.
The primary limitation of the Sahara system is the same as its main advantage: its
tethered cable design. The inspection length possible from an insertion point is limited
by the amount of available cable as well as the amount of flow in the pipe line and how
far this flow can carry the hydrophone and cable through the pipe before friction stops
it.
Sahara Video permits a real time video inspection of a live pipeline and only requires a
2 inch access although it has the same cable and inspection limitation and the video
quality is reduced in larger diameter pipes.
The Sahara Wall Thickness technique allows flexible distance and better interval
resolution from the cable control but can only indicate the average wall thickness in a
section and not specific defects.
PipeDiver is a proven platform designed for live inspection of PCCP using the RFTC
technology but has been adapted to use the RFEC technique to provide wall thickness
loss in metallic pipelines. The detection sensitivity is limited by the number of sensor
channels but since the significant challenge of non-disruptive inspection has been
overcome future development can focus on increasing the number of available
detectors.
The Sahara and PipeDiver techniques are complementary technologies that offer a
spectrum of solutions to utilities. By detecting very small leaks and accurately
pinpointing the leak position, Sahara leak can provide pinhole corrosion in pipe wall
and joint problems, which are a good indication of pipe condition. For wall thickness
issues, including graphite, wall thinning, but not yet leaking, Sahara Wall Thickness can
provide average sectional wall thickness info during the same time with Sahara leak
and PipeDiver RFEC will be able to provide more detailed information. Also, Sahara
Video provides internal line condition and visual corrosion information. All are live
inspections that take place while the pipeline remains in service.
31
-------
5.4 Future Developments
Sahara Leak Detection is a mature technology used successfully for many years and
future development of the technique will focus on making it even easier to use. The
main challenge with Sahara Video is to improve its video and lighting quality in larger
diameter pipes and to possibly combine the video and leak techniques into a single
sensor head which would reduce the amount of insertions required and make the
overall inspection more efficient. The Sahara Wall Thickness technique will continue to
fine tune its field and analysis procedure in addition to more verification and
calibration.
PipeDiver is a proven platform for entering a pipe through a standard access in live
conditions and for inspection of PCCP. Using the data and experience obtained from
this first PipeDiver RFEC inspection pilot PPIC will be able to further improve the
PipeDiver system for metallic pipeline inspections. Technical components will be
reviewed for possible advancements including improved detectors and detector
placement. As well, the analysis process will be reviewed for new analysis techniques
and improved software. Specifications and implementations of standard accesses will
be reviewed to prevent future insertion and retrieval issues. Results need to be
compared to actual pipe calibration and verification from the Westport Rd.
Transmission Main in order to review and improve the current analysis techniques.
32
-------
6. PHOTOGRAPHS
Sahara insertion site with valve and tap
in Pit 1.
Sahara control center (truck) and Sahara
insertion setup at Pit 1.
Valve creating a simulated leak in Pit 4.
33
-------
Pits were constantly flooded due to
ground water and rain storms.
PipeSpy locating a simulated leak at Pit 4
Orifice used to create simulated leaks
34
-------
The Sahara Video sensor head and
drogue.
Technicians inserting the Sahara
hydrophone into the pipe in live
conditions.
The Sahara insertion tube setup in Pit
1.
35
-------
Acoustic unit recording reference
sound signals at the insertion point.
Accelerometer acoustic sensor attached
to the Sahara insertion tube.
Carrying the PipeDiver tool ready to be
installed into the insertion tube.
36
-------
Preparing the PipeDiver insertion and
retrieval tubes.
PipeDiver insertion tube setup at the
launch site.
•I
Attaching the PipeDiver extraction tube
on the gate valve.
37
-------
Setting up the PipeDiver extraction
tube.
Technicians locating the PipeDiver vehicle
from above ground.
38
-------
APPENDIX B
PURE SMARTBALL™ LEAK DETECTION REPORT
B-l
-------
SmartBall® Leak Detection Survey
August 6 & 7, 2009
Prepared For the Environmental Protection Agency
Prepared By:
Pure Technologies Ltd
August 14, 2009
JeffKler
-------
1 Executive Summary
The SmartBall was deployed to inspect the24 inch cast iron mortar lined pipeline on Thursday August 6th and Friday August 7th,
2009. The SmartBall was run through the pipe and was able to detect acoustic anomalies likely caused by leaks at 15 locations.
The identified leaks have been summarized below.
Summary of Pipeline Details
Total Length of Pipe Surveyed:
Type of Pipe:
Diameter of Pipe:
Number of Leak Locations
Number of Simulated Leak Locations
Total Number of Leak Locations
2057.0 ft
Cast Iron Mortar Lined
24 inch
12
3
15
2 Pipeline Summary
The ap proximate 1 ayout o f t he 2 4 i nch cast iron W ater p ipeline in spected s tarting a t the i ntersection o f Chenoweth L ane a nd
Westport Road, to the intersection of Ridgeway Avenue and Westport Road. The approximate line location is displayed on the
aerial photograph below in Figure 2.1.
Figure 2.1: General layout of the pipeline inspected.
Sensor Locations ( •>'" ) of the pipe inspected
Page 2 of25
-------
3 Tracking the Position of the SmartBall
The position of the SmartBall within the pipeline is critical for locating important features, such as leaks. The methodology used
to track the tool involves obtaining a velocity profile using data obtained from the accelerometers and magnetometers on board the
SmartBall. Then, absolute position reference points obtained from the SmartBall Receiver (SBR) are applied to time stamped data.
Individual SBR's were able to track the ball's progress through the pipeline for over 850 feet and the distance and location of
these SBR's were based on the information provided to Pure by Battelle. The result of the rotation profile and SBR tracking is a
position versus time relationship for the entire run of the tool. The exact location of where each SBR was placed along the
pipeline during the run is detailed in Appendix A.
Figure 3.1 shows the position data for the runs. The position of the SmartBall indicated by the red line was fixed by fitting the
position profile to known locations along the pipeline. The slope of the blue line indicates the instantaneous velocity of the tool.
The velocity of the ball as it travelled through the pipeline is shown in Figure 3.2. Figure 3.3 displays the position of the ball as it
was tracked in real time on site by the SBR's.
Run#l(Aug6):
1969 - -
13:59:58 14:08:13
Time of Day (hh:mm:ss)
Run#2(Aug6):
1523 IS
1331:3B
-5 43 1Ł
Tintc* Caylhfi Fim:n|
100153
10:13:19
Page3 of 25
-------
Run #3 (Aug7):
Timed Daf[hh:rnm,ss)
Run#4(Aug7):
1969 --
1640
1312 - -
Run#5(Aug7):
~T, in- 11 Piny (hi v
th ,1 -7th
Figure 3.1: Position Profile of the SmartBall vs. Time of Day for the August 6 , and 7 , 2009 inspections
Page 4 of25
-------
Run#l (Aug6):
1,0--
^yS*~?<*^s<;^J\^ fj^\/f^-\^\^t^^^
13:59:58
Tl me of Day (hh: r
Run#2(Aug6):
I
VwrfW-\(^w!^^
Run #3 (Aug7):
1.0
Ł• 0.7 - -
o
3
I
"^^^AsS^tVvd****^^^
0.0
08:10:25
09:27:05
09:43:45
10:00:25
10:17:05
Time of Day [hh:mm:ss]
Page 5 of25
-------
Run#4(Aug7):
o.:
^V
-------
820 -F
Run#l (Aug6):
13:43:18
13:51:38
13:59:58
Time of Day (hh:rnm:ss)
14:08:18
14:18:38
14:24:58
820
898
4B2
I
I 328
| 164 • •
Run#2(Aug6):
15:23:19
1531:38
1556:23
160M.53
16:13:18
Run#3(Aug7):
09:J7 03
09 43:4fl
Page 7 of25
-------
Ł820 +
Run#4(Aug7):
Run#5(Aug7):
•820 - -
> 656
g 4®2 -.'-
o :'
1 328 - -
o
c
ro
U 164 --
-r-
th
Figure 3.3: SBR Tracking Points vs. Time of Day for the August 6 , and 7 , 2009 inspections
Page 8 of25
-------
4 Results
Upon retrieval of the tool, the acoustic data recorded by the SmartBall was analyzed and cross-referenced with the position data
from the SBR to determine location. A summary of the leaks found in the runs is detailed below. The location accuracy of the
anomalies is dependant on the accuracy of the pipe distance and lay information provided to Pure.
4.1 Summary of Results
Figure 4.1 shows the value of the leak indication power as detected by the SmartBall with respect to the position of the SmartBall
along the pipeline. The severity of any leaks found can be estimated by correlating the value of the leak signal (a calculated
parameter) with calibrations performed by the SmartBall and are detailed in section 4.2 titled Leak Calibration Curve. The general
upward slope toward the right side of each graph has resulted from a large amount of flow noise generated by the water pressure
relief valve and disposal at the downstream end of the run.
Run#l (Aug6):
fi
Ł -30
Distance (ft)
Run#2(Aug6):
1-30
Distance (ft)
Page 9 of25
-------
Run #3 (Aug7):
984
Distance (ft)
Run#4(Aug7):
984
Distance (ft)
Run#5(Aug7):
a -20
6
Distance (ft)
Figure 4.1: Acoustic Profile of the SmartBallvs. Time of Day for the August 6th, 7th, 2009 inspections
The critical findings of the pipeline inspection are summarized in table 4.1.
Page 10 of 25
-------
Table 4.1 - Summary of Acoustic Anomalies Resembling Leaks That Are Not Simulated
Leak ID # [Distance from Start [Description
1
2
3
4
5
6
7
8
9
10
11
12
53ft
125ft
199ft
341ft
414ft
556ft
579ft
641ft
966ft
1080ft
1210ft
1580ft
1724ft
1809ft
1930
Leak (Small)
Leak (Small)
Leak (Small)
Leak (Medium)
Leak (Small)
Leak (Small)
Simulated Leak Site
Leak (Small)
Leak (Small)
Simulated Leak Site
Leak (Small)
Simulated Leak Site
Leak (Small)
Leak (Small)
Leak (Small)
Approximate Size (US gpm)
0.15
0.1
0.8
15
0.2
1.0
Varying
2.0
0.1
Varying
1.0
Varying
1.5
2.0
5.5
Table 4.2 - Simulated Leaks Detected by the SmartBall
Distance
from Start
579ft
1080ft
1580ft
Inspection 1
Estimated Leak
Rate (US gpm)
0.57
8
0
Inspection 2
Estimated Leak
Rate (US gpm)
0.3
2.8
15
Inspection 3
Estimated Leak
Rate (US gpm)
1.8
7.2
30
Inspection 4
Estimated Leak
Rate (US gpm)
4.5
0.1
40
Inspection 5
Estimated Leak
Rate (US gpm)
8
0.57
0
Page 11 of 25
-------
4.2 Leak Rate Calibration
To assist in identifying the approximate leak rate of any identified leak for the inspection performed on Thursday Aug. 6th and 7th,
2009, Pure Technologies Ltd has compared the leak indication power of a detected leak with that of a known leak rate. The
calibration curve applied to gauge the size of the leaks detected on this inspection is shown in Figure 4.2. Known leak rates and
their leak indication power (in dB) are usually developed by holding the SmartBall in the extraction net at the end of surveyed
runs. For these inspections leaks were created for the SmartBall while it passed through the pipe (shown in green in Figure 4.2
below). The leak indication power is the single most important indicator of a leak's size and presence. In order to confirm that an
acoustic anomaly is actually a leak, a frequency analysis tool is used and is shown with each identified leak in Section 5.
Leaks of varying rates are produced using a 1/2 inch ball valve attached to the extraction stack and a graduated bucket was used to
collect and measure the water created by each of the leaks over a measured period. Because the simulated leaks are controlled and
released through a threaded outlet, the comparison to actual field condition leaks may vary. This is because the acoustic frequency
and power indication of any leak will vary with many factors, including pressure, pipe diameter, size and configuration (pin-hole,
rolled gasket, split pipe, etc.). However, the leak calibration curve provides a useful tool in approximating leak rates for identified
leaks. These calibration leaks are shown in the below graph as green squares. The actual leaks detected during the inspection are
shown as red circles.
As an approximation, the leaks in the range of 0-2 US Gallons per Minute (0-7.5 Liters per Minute) have been classified as small,
2-10 US Gallons per Minute (7.5 to 37.5 Liters per Minute) have been classified as medium, and above 10 US Gallons per Minute
(37.5 Liters per Minute) have been classified as large.
•A calibration leak/point member
of this calibration curve, not part
of default calibration curve.
•Marked Leak.
Leak Rate (gal/min)
Figure 4.2: Leak Calibration Curve used to size the leaks on the Thursday Aug. 6, 2009 inspection
Page 12 of 25
-------
5 Sites of Interest - Details
Leak #1
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
53ft
53 ft after Insertion
Small
0.15
Leak Indicator
Frequency Spectrum
Location
,-
Figure 5. la: Leak Indication Power of Leak
Figure 5.1b: Frequency Spectrum of Leak
Figure 5.1c: Approximate Location of Leak
Page 13 of 25
-------
Leak #2
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
125ft
125 ft after Insertion
Small
0.1
Leak Indicator
Frequency Spectrum
Location
Figure 5.2a: Leak Indication Power of Leak
11* •111., 1.1... 1.1
Figure 5.2b: Frequency Spectrum of Leak
Figure 5.2c: Approximate Location of Leak
Page 14 of 25
-------
Leak #3
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
199ft
199 ft after Insertion
Small
0.8
Leak Indicator
Frequency Spectrum
L,
Figure 5.3a: Leak Indication Power of Leak
Location
Figure 5.3b: Frequency Spectrum of Leak
Figure 5.3c: Approximate Location of Leak
Page 15 of 25
-------
Leak #4
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
341 ft
341 ft after Insertion
Medium
15
Leak Indicator
Frequency Spectrum
Figure 5.4a: Leak Indication Power of Leak
-.''''• •
"
'
"^**iit"
•
r
<
. - .
•
-.
;, ,
Location
Figure 5.4b: Frequency Spectrum of Leak
Figure 5.4c: Approximate Location of Leak
Page 16 of 25
-------
Leak #5
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
414ft
395 ft before Mid-Point Sensor
Small
0.5
Leak Indicator
Frequency Spectrum
Figure 5.5a: Leak Indication Power of Leak
Location
Figure 5.5b: Frequency Spectrum of Leak
Figure 5.5c: Approximate Location of Leak
Page 17 of 25
-------
Leak #6
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
556ft
253 ft before Mid-Point Sensor
Small
1.0
Leak Indicator
Frequency Spectrum
Figure 5.6a: Leak Indication Power of Leak
Location
Figure 5.6b: Frequency Spectrum of Leak
Figure 5.6c: Approximate Location of Leak
Page 18 of 25
-------
Leak #7
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
641 ft
168 ft before Mid-Point Sensor
Small
2.0
Leak Indicator
Frequency Spectrum
Figure 5.la: Leak Indication Power of Leak
Location
Figure 5.7b: Frequency Spectrum of Leak
Figure 5.7c: Approximate Location of Leak
Page 19 of 25
-------
Leak #8
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
966ft
157 ft after Mid-Point Sensor
Small
0.1
Leak Indicator
Frequency Spectrum
Figure 5.8a: Leak Indication Power of Leak
Location
Figure 5.8b: Frequency Spectrum of Leak
Figure 5.8c: Approximate Location of Leak
Page 20 of 25
-------
Leak #9
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
1,210ft
401 ft after Mid-Point Sensor
Small
1
Leak Indicator
-I I-
Frequency Spectrum
Figure 5.9a: Leak Indication Power of Leak
Location
Figure 5.9b: Frequency Spectrum of Leak
Figure 5.9c: Approximate Location of Leak
Page 21 of 25
-------
Leak #10
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
1,724ft
333 ft before Extraction
Small
1.0
Leak Indicator
^
Frequency Spectrum
Figure 5.10a: Leak Indication Power of Leak
Location
Figure 5.1 Ob: Frequency Spectrum of Leak
Figure 5.10c: Approximate Location of Leak
Page 22 of 25
-------
Leak #11
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
1,809ft
248 ft before Extraction
Small
2.0
Leak Indicator
Frequency Spectrum
Figure 5.11a: Leak Indication Power of Leak
Location
Figure 5. lib: Frequency Spectrum of Leak
Figure 5. lie: Approximate Location of Leak
Page 23 of 25
-------
Leak #12
Distance from Insertion Point:
Distance to Nearest Sensor:
Estimated Size of Leak (small/med/large):
Estimated Size of Leak (US gpm)
1,930ft
127 ft before Extraction
Small
5.5
Leak Indicator
Frequency Spectrum
Location
Figure 5.12a: Leak Indication Power of Leak
' • -'
'
-I ' —I 1 1-
Figure 5.12b: Frequency Spectrum of Leak
Figure 5.12c: Approximate Location of Leak
Page 24 of 25
-------
Appendix A: Ball Tracking Sensor Locations
Sensor Locations for August 6th and 7th, 2009 Inspections
AGM Location ID
Latitude
Longitude
Distance from Launch
Insertion
38.2536
-85.6549
0.0 ft
AGM Location ID
Latitude
Longitude
Distance from Launch
AGM Location ID
Latitude
Longitude
Distance from Launch
Vlidpoint Sensor
38.2547
-85.6525
809.0 ft
Extraction
58.2566
-85.6489
2057.0 ft
Page 25 of 25
-------
APPENDIX C
ECHOLOGICS LEAKFINDER LEAK DETECTION* REPORT
(*See pp. 1,2,3, 10, 15, 16, 17, 18, 19, 20, 21, 22, and 23)
C-l
-------
Condition Assessment Field
Demonstration
Echologics Engineering Inc.
This report outlines the results of non-destructive condition assessment testing performed on 24-
inch concrete-lined cast iron cylinder pipe in Louisville, Kentucky
50 Ronson Dr, Unit 155
Toronto. Ontario. M9W IBS
-------
echo^logics
Summary
Echologics Engineering Inc.
50 Ron son Drive, Unit 155
Toronto, Canada M9W 1B3
T:+1[416]249.6124
F:+ 1[416] 249.3613
www.echologics.com
The purpose of this study is to assess the performance of Echologics proprietary non-
destructive acoustic condition assessment technology for leak detection and condition
assessment on cast iron pipes. Data acquisition was performed on a 24-inch cast iron
pipe that runs beneath Westport Rd in Louisville Kentucky on August 11th and 12th
2009. This report summarizes the results of the data acquisition and the corresponding
analysis.
Acknowledgements
Battelle
Abraham Chen - Program Manager
Bruce Nestleroth - Research Leader
Lili Wang - Program Assistant
Louisville Water Company
Keith Coombs - Program Manager
Revisions
0.1 Sept 7, 2009
0.2 Sept 18, 2009
0.3 Sept 25, 2009
0.4 Sept 30, 2009
1.0 Nov4, 2009
1.1 Nov13, 2009
Dave Johnston - Draft Report
Ellen Turner - Draft Report Revision
Marc Bracken - Draft Review
Dave Johnston - Draft Submittal to Client
Dave Johnston - Final Revision
Dave Johnston - Update
-------
echo^logics
Echologics Engineering Inc.
50 Ron son Drive, Unit 155
Toronto, Canada M9W 1B3
T:+1[416]249.6124
F:+ 1[416] 249.3613
www.echologics.com
Contents
1. Introduction 1
2. Background 2
2.1. Signal Processing 2
2.2. Leak Detection 3
2.3. Non-Destructive Condition Assessment 3
2.4. Metallic Pipe 5
2.5. Concrete Lining 6
2.6. Nominal Data 7
2.7. Sensitivity Analysis 7
Distance Measurement 7
Pipe Manufacturing Tolerances 8
Repair Clamps on Previous Leaks 8
Variation on Young's Modulus 8
Replacement of short Pipe Sections for Leak Repairs 9
Inaccurate Records 9
2.8. Sources of Error 10
2.9. Negative Correlation Signals 10
2.10. Condition Assessment Data Interpretation 11
2.11. Results of Pipe with 5% degradation 11
2.12. Results of Pipe with 9% degradation 12
2.13. Results of Pipe with 47% degradation 12
Guidelines for Interpretation of Results 14
3. Methodology 15
3.1. Leak Detection 15
3.2. Condition Assessment 16
3.3. Instrumentation 17
4. Results and Discussion 19
4.1. Demonstration Results 20
Section 1: PMl to PM2, Demonstration in Pit#4 21
Section 2: PM2 to PM3, Demonstration in Pit #5 21
Section 3: Pit#4 to Pit#5, Demonstration in Pit#2 21
General Comments 22
4.2. Leak Detection Results 22
File #2a- Pit A to Pit B 22
File#7c-PitFtoPit3 23
-------
echo^logics
Echologics Engineering Inc.
50 Ron son Drive, Unit 155
Toronto, Canada M9W 1B3
4.3. Condition Assessment Results I:±!Ł!$.2.4.9:6.L2? 24
F:+ 1 416] 249.3613
5. Concluding Remarks ™.™:*:222?™ 26
6. Appendix 27
Figures
Figure 1: Photos of pipe with 4.2% measured loss 13
Figure 2: Photos of pipe with 8.9% measured loss 13
Figure 3: Photos of pipe with 47.3% measured loss 14
Figure 4: Correlation result for File #2 23
Figure 5: Correlation Result for File #7 24
Figure 6: Pipe Wall Cross-Section i
Figure 7: Site Layout ii
Figure 8: Correlation Report for File #2a -PitAto PitB iii
Figure 9: Correlation Report for File #7c - PitF to Pit3 iv
Tables
Table 1: Nominal Dimensions 7
Table 2: Excavation Locations 19
Table 3: Sensor-to-Sensor Distances 20
Table 4: Demonstration Results 20
Table 5: Condition Assessment Results 25
IV
-------
Cast Iron Pipe Pilot Study
1. Introduction
Echologics Engineering was invited to conduct a pilot study on selected cast iron pipes
in Louisville, Kentucky. The intent of the study is to test the feasibility of Echologics
proprietary non-destructive condition assessment technology both for condition
assessment and leak detection on a 24-inch cast iron pipe along Westport Rd.
Data acquisition was performed on several sections of the 24-inch main. There are
three sets of results presented in this report. First, the results of the background leak
detection results will be discussed. Locations of any already existing leaks will be
presented in this section. Second, The results of the leak detection demonstration will
be presented. This will include whether or not the demonstration leak was discovered
and what the estimated flow rate is. Finally, the results of the condition assessment will
be presented.
Background measurements were performed in section lengths between 250-feet and
360-feet in length. The background measurements were performed with the purpose of
finding any already existing leaks and performing the condition assessment
measurements. Typically, the same methods are used when Echologics is performing
commercial assessment services.
The demonstration measurements were performed using different sensors
(hydrophones) and longer section lengths, approximately 1000-feet. Again, this
arrangement was chosen because it would be typical for commercial leak detection
projects.
As a warning to the reader, it should be noted at the outset that for completeness, we
have included fairly extensive technical information, some of which will be beyond the
technical knowledge base of some of the readers of this report. It is not our intent to
educate readers in signal processing theory, although we have provided some layman's
explanation of the background theory.
-------
Cast Iron Pipe Pilot Study
2. Background
2.1. Signal Processing
Time differences are measured using fast Fourier transforms (FFTs) and advanced
cross-correlation algorithms. There are also a number of other acoustic tools that aid in
data analysis processes. For the purposes of understanding this report, there are
several signal processing functions that should be understood:
Coherence Function: The coherence function is a measure of how similar the vibration
signals are on a frequency basis. When two signals are perfectly similar at a given
frequency (for example, two sine waves), the coherence function value is 1 at that
frequency. Good coherence would be considered anything at 0.5 and above.
Transfer function: The transfer function is a frequency based plot of the relative strength
of the two measurement channels. This function shows the relative vibration level of the
blue and white stations, and can be given in log or linear format. Many vibration
engineers prefer to see both formats, as a log plot is easier on initial read, however a
linear plot will show more detail.
Frequency plot (FFT): The frequency plots given in this report are fast Fourier
transforms of the raw level vs. time signals. Very simply, these plots show the frequency
content of the vibration signals measured. It is often possible to pick out leak noise on
the frequency plots, and these can be used to analyze the leak detection signals. For
example an FFT from the blue station may show a spectrum consistent with leak noise
with significant higher frequency vibration, while the white station signal may show no
high frequency content indicating a possible PVC repair (the PVC repair may filter out
high frequency content).
Correlation Function: The correlation function is the level vs. time function that will
indicate a leak, and in the case of condition assessment measurements will show the
out-of-bracket peak or time difference. Ideally a good correlation peak should be very
sharp, and very prominent. The LeakfinderRT software will present a warning for an out-
-------
Cast Iron Pipe Pilot Study
of-bracket signal when the time delay of the signal approaches the total time delay of
the entire measurement distance (i.e when t=>d/v).
2.2. Leak Detection
The leak detection methodology used is the cross correlation method. A correlator
listens passively for noise created by a leak. Two sensors are mounted on fire hydrants,
exposed pipe, or valves in such a way that the leak lies between them, or is 'bracketed'
by the sensors. A leak that lies outside the area spanned by the sensors is known as an
'out-of-bracket' leak. Any active leaks or draws or other sources of noise on the pipe will
vibrate the pipe and detected by the sensors.
The signals will be recorded and the cross-correlation plot will be analyzed. Any
potential leaks will appear as a spike in the cross-correlation plot. The position of the
spoke on the x-axis corresponds to the time difference it takes for the signal to arrive at
the Blue and White stations. The wave velocity is known and therefore the position
relative to either of the stations can be computed.
2.3. Non-Destructive Condition Assessment
An acoustic signal induced in the pipe may be used to determine the acoustic wave
velocity in a section of pipe, which can in turn be used to back calculate the average
wall thickness of the pipe. Knowing the distance between two sensors mounted some
distance apart on valves or fire hydrants, the acoustic wave velocity will be given by v =
d/t, where d is the distance between the sensors, and t is the time taken for the
acoustical signal to propagate between the two sensors. If an accurate measurement of
the acoustic wave velocity is made, it is possible to back-calculate the remaining
average thickness of the pipe between the two sensors.
The wall thickness measured represents an average between the two sensors. Typically
the length of the pipe section over which the acoustic velocity is measured 100 to 300
metres (300'-1000!), however this distance can be decreased to anywhere between 30-
100 m to increase the resolution.
-------
Cast Iron Pipe Pilot Study
Echologics proprietary leak noise correlator, LeakfindeRT was used to determine the
acoustic velocity. An acoustic source outside the area spanned by the sensors (an 'out-
of-bracket' source) was used to induce an acoustic wave in the pipe, and the time delay
difference was measured. At each site the noise source to induce the acoustic wave;
was either operation of a fire hydrant, or a valve or hydrant was impacted.
The average wall thickness of the pipe section between the acoustic sensors is then back calculated from
a theoretical model. As the pipe wall thickness decreases over time, the acoustical wave velocity
decreases. From an intuitive perspective, this is akin to trying to run on a trampoline versus solid ground;
as the bounding layer becomes more flexible the propagation velocity decreases. The acoustical wave
velocity is given in Equation 1: Wave Velocity - Thickness Model below. It should be noted that
there are other factors that affect the propagation velocity such as water temperature and pipe
wall inertia. These factors are not shown here but have been accounted for in the final results.
V = l'
where
v: Propagation velocity of leak noise in pipe
v,,: Propagation velocity of sound in an infinite body of water
f: Internal diameter of pipe
~j\ Thickness of pipe wall
KWJur\ Bulk modulus of elasticity of water
Er-,F;. Young's modulus of elasticity of pipe material
Equation 1: Wave Velocity - Thickness Model
The acoustic propagation wave (the water hammer mode) propagates as a
compression wave in the fluid, and a dilatational wave in the pipe. Therefore the pipe
will breathe on a microscopic level, and therefore the pipe will go into stress. There are
two key implications to this:
-------
Cast Iron Pipe Pilot Study
1. Only the structural part of the pipe that can carry load will contribute to the
structural stiffness of the pipe, therefore deposits on the pipe wall such as
tuberculation or graphite will not be included in the average wall thickness
measurement.
2. We will measure the minimum structural thickness of the pipe, as the level of
strain of the pipe will be dependent on the minimum wall thickness at any point
around the circumference the pipe.
As noted, the pipe wall thickness calculated from these measurements represents an
average value for the pipe section over which the acoustic velocity is measured. At first
glance, this may appear to be a limitation of the technology, as the question could be
reasonably asked as to whether the method can find pockets of corrosion. In practice
this has not been the case. The technology has been applied to generally much greater
sample lengths of pipe than could be done with random sampling or electro-magnetic
technologies. Therefore when surveying long lengths of type, the operators begin to
look for anomalies in the measurements that could indicate degraded sections of pipe.
When these are seen, the distance between the sensors may be decreased and more
resolution obtained. Generally, pipes will have a more-or-less uniform thickness profile
with isolated pockets of corrosion over significant lengths, say 50 to 100 meters, as soil
and bedding conditions are unlikely to change significantly over such distances. Also,
average wall thickness values are suitable to evaluate the residual life of pipes for the
purpose of long-term planning of rehab and replacement needs. The use of techniques
such as evaluation of stray currents, and soil corrosivity studies and main break history
may be used in conjunction with our data to evaluate overall pipe condition.
2.4. Metallic Pipe
The primary degradation mechanism in buried metallic pipes is corrosion. Corrosion
occurs in many different forms and can be accelerated or inhibited based on soil
properties, water properties and characteristics of the pipes surroundings.
Two main forms of corrosion occur in buried pipelines: uniform corrosion and pitting
corrosion. Uniform corrosion occurs when general, constant corrosion occurs on all
-------
Cast Iron Pipe Pilot Study
surfaces of the pipeline. This can occur from the inside out and is caused by the
properties of the water that the pipe is carrying. Or it can occur from the outside in if the
pipe is in submerged or semi-submerged conditions.
Pitting corrosion occurs on the inside and outside surfaces of the pipe. This is when
small areas corrode preferentially leading to cavities or pits, and the bulk of the surface
remains unaffected. Pitting corrosion can be accelerated under stagnant conditions,
which is why it is generally more severe on the outside surface of the pipe.
Other forms of corrosion can occur including: galvanic (dissimilar metals), De-Alloying
(graphite), inter-granular and erosion corrosion. All of these can contribute to the overall
degradation of the pipe but they are considered to be relatively insignificant compared
to the impact of uniform and pitting corrosion.
2.5. Concrete Lining
The wave propagation velocity is a function of the thickness of the pipe wall and the
corresponding material elastic modulus. Therefore, if a pipe is concrete lined the
structural stiffness of the pipe is increased via the addition strength of the concrete. The
wave velocity then becomes a function of the structural stiffness of the metal and the
concrete lining.
In order to account for this, it is necessary to calculate the nominal thickness of the pipe
as if it was not lined with concrete i.e. the equivalent structural thickness of a metallic
pipe without the concrete lining. This will be referred to as the equivalent thickness and
generally it is 2-3mm thicker than the thickness of the base metal. This value can also
be considered as the 'effective' or the 'structural' thickness of the pipe.
The measurement will then be compared to this value, the equivalent thickness rather
than the thickness of the metal alone.
-------
Cast Iron Pipe Pilot Study
2.6. Nominal Data
Battelle provided original specifications for both diameters of pipe. The details are
presented below in Table 1: Nominal Dimensions. There is also an image of the cross-
section of the pipe shown in Figure 6: Pipe Wall Cross-Section. It closely matches the
values presented here.
YOI
1932
Type
Pit Cast Iron
Dia
(inch)
24
Equivalent Thickness
=
=
Dia
(mm)
610
of Cast I
22.2
0.874
Cast Iron
Thickness
(inch)
0.75
ron without
mm
inch
Cast Iron
Thickness
(mm)
19.05
Concrete Lin
Lining Lining
Thickness Thickness
(inch) (mm)
0.25 6.35
ing
Table 1: Nominal Dimensions
2.7. Sensitivity Analysis
Echologics has committed a substantial amount of effort to reduce sources of error in
our assessments. However there are still variables that strongly affect the final result.
They are as follows:
Distance Measurement
A calibrated wheel is used for obtaining our distances, and distance measurements
were repeated 3-4 times for each location to ensure the best possible accuracy. For
example, on a total distance of 150m, an error of +2.5m resulting in a measured
distance of 152.5m will cause a positive error in the final result of approximately 17.5%.
An accurate distance measurement is therefore crucial to an accurate assessment. For
this reason, our preference is always to use line valves, as these provide the most
accurate distance measure, as it is a point-to-point measurement. If the pipe has
multiple bends and elevation changes between the sensor connection points, error in
the distance measurement increases, as it is not always easy to identify where the
bends occur. As a result, if a situation existed where the pipe location was in question, it
-------
Cast Iron Pipe Pilot Study
was requested that Bristol Water re-measure the distance after a pipe location was
performed. In some cases this improved the measurement result.
Pipe Manufacturing Tolerances
The pipe laid will have small differences in thickness and due to manufacturer and
tolerances. This factor is usually 5-10% dependent on the manufacturer and the
material. This may lead to a pipe growing by a small percentage (5-10%) compared to
the nominal thickness used. This is particularly true of the older vintages of pipe
measured in this study. Generally, the materials data used for the calculation is chosen
using conservative estimates. The purpose of this is to provide a worst-case scenario to
the client i.e. assume that the pipe is manufactured to the better side of the tolerances
and calculate the remaining thickness based on this. This is not considered to be error
because the presented result actually represents the current condition of the pipe.
Variation in internal diameter of the pipe can also affect the final result. If the
manufacturing tolerances for the diameter are approximately 5-10% the corresponding
results on the calculated value will also vary by approximately 5-10%. This is
considered to be relatively insignificant if, in fact, the information provided by the client
is correct. This is not always the case and it will be discussed later in this section.
Repair Clamps on Previous Leaks
A small number of repair clamps should have an insignificant effect on the test results,
since the acoustic wave is primarily water borne and will bypass the clamps. It should
be noted that although the acoustic wave is primarily water-borne, it is a coupled wave
that moves simultaneously in the pipe (in an axi-symmetrical breathing mode), and in
the water as a compression wave. Thus the wave will generally skip across
discontinuities such as clamps, and reestablish itself in the pipe material beyond.
Variation on Young's Modulus
In general, a change in elastic modulus of 10% will cause a change in the calculated
thickness by approximately 10%. Therefore it is necessary to account for this variation.
The elastic modulus is known for common materials used in the manufacturing of
pressure pipe but this value can vary from manufacturer to manufacturer. This depends
on the manufacturing process and the quality of the material.
-------
Cast Iron Pipe Pilot Study
Replacement of short Pipe Sections for Leak Repairs
The effect of short pipe replacements will depend on the material used. For example, a
new 6-metre long ductile iron repair in a 100-metre long / 152 mm-diameter cast iron
pipe section of average condition, will produce a small error of +3.5% in predicted wall
thickness. However, the same repair made with PVC pipe would produce an
unacceptable error of -41%. Preferably, pipe sections selected for testing should be free
of repaired segments. However, if this condition does not exist, the effect of new pipe
segments can be accounted for provided that accurate information is available for the
location, length, material type and class of new pipe segments.
Inaccurate Records
In some cases the possibility exists that inaccurate information was provided by the
client, specifically referring to the pipe diameter and the pipe material. As described
above, small manufacturing variations in elastic modulus and internal diameter only
affect the final result by 5-10% but if the information supplied by the client is incorrect, it
is flawed by much greater magnitudes. For example, a common error would be to
mistake a 200mm pipe for a 250mm pipe. When the calculation is performed using an
internal diameter of 250mm, the remaining thickness may be 12.5mm. If the same
calculation is performed using an internal diameter of 200mm, the remaining thickness
is reduced to 9.3mm, a change of 3.2mm! In this case, the error caused a 35% over
estimation of the pipe wall thickness.
Another common problem arises when improper pipe material information is provided.
For example, if a pipe was thought to be spun cast iron when, in fact, it is ductile iron.
When the calculation is performed using the elastic modulus for spun cast iron
(131Gpa), the remaining thickness may be 11.6mm. If the same calculation is
performed for a ductile iron pipe (169Gpa), the remaining thickness drops to 8.9mm, a
change of 2.7mm! The error caused a 30% over estimation of pipe wall thickness.
It becomes obvious that accurate records from the client are an essential requirement
for providing accurate condition assessment results.
-------
Cast Iron Pipe Pilot Study
2.8. Sources of Error
The results of the sensitivity analysis provide insight into how the various material
properties and pipe dimensions can affect the final result. If one ignores error introduced
by manufacturing tolerances and inaccurate nominal information, the main source of
error is cause by improper sensor-to-senor distance measurements.
The average section of pipe tested during this project was 150m. If one assumes that
the sensor-to-sensor spacing can be measured accurately to within 1m, the resulting
error in the thickness calculation is approximately 5%. If however, there are multiple
bends in the pipe or significant elevation changes, the error in the distance
measurement may increase. For example, one bend in the pipe may introduce an
additional error of 1m. With a total distance error of 2m, the resulting error in the final
calculation is approximately 10%.
2.9. Negative Correlation Signals
There were several locations where correlation signals could not be acquired, or they
were of poor quality. This can happen for a number of reasons, and we typically find
that this occurs on a percentage of all of our projects. Although we have never had the
opportunity to fully explore the reasons for this, the following are some of the conditions
that we have encountered that have affected our measurements:
1. The presence of plastic repairs in metallic pipes can cause poor correlation
signals, and will also cause inaccurate thickness
2. Loose or worn components in fittings used for the measurements, such as valve
or hydrant stems.
3. Heavily tuberculated pipe, particularly old cast iron or unlined ductile iron may
attenuate the acoustic signals to such an extent that a correlation is of very low
quality.
10
-------
Cast Iron Pipe Pilot Study
2.10. Condition Assessment Data Interpretation
The condition of a pipe may be assessed based by judging it based on other pipes that
we have measured and then exhumed to determine the condition. For a full condition
assessment, it is recommended that our data be used in conjunction with soils
information, any ground potential measurements done, along with any pipe samples
exhumed during leak repairs. Acoustic non-destructive condition assessment cannot
pinpoint the source of degradation. For example, a reading of -20% pipe wall could
mean that the pipe is generally degraded along it's entire length, or the pipe could have
significant degradation at only one or two locations.
In the absence of other parameters, we have provided a gradation scale based on our
previous project experience and pilot studies. Based on our previous experience, we
have provided background on typical results found during the course of our condition
assessment surveys. Please note that the sample photos shown in the following section
are from a previously performed pilot study. They are to be used only to demonstrate
the typical levels of degradation found from previous testing. This is meant to act only
as a guideline in assessing the results of this study.
The images presented below show four pictures in each. The top left picture shows the
as-found condition of the pipe. The top right image shows an overview shot. The bottom
left shows a close up of the surface after it was sandblasted. The bottom right shows
the internal surface after it was sandblasted.
The descriptions below described results measured by Echologics, given by an
averaged measured loss in percent. The physical results given are the average
measured value at either end of the pipe, the average pit depth on the outside surface /
inside surface and the qualitative condition on the outside surface / inside surface.
2.11. Results of Pipe with 5% degradation
A section of pipe where 4.7% measured loss is shown in Figure 1. The nominal
thickness of this pipe was 12mm (0.47in), whereas the lab measured physical thickness
at either end of the sample was 11.4+/-2.7mm (0.45in +/-0.1in). The average pit depth
was 1.5mm / 1.9mm. The pipe was qualitatively described as very good / very good.
11
-------
Cast Iron Pipe Pilot Study
This again is an indication that the acoustic wave velocity from the acoustic mode of the
pipe that we are measuring is based on the average minimum structural thickness, not
the average physical thickness.
The sample was taken from an area with corrosive clay based soil. The figures indicate
that although there are local areas of corrosion, the pipe wall is generally in good
condition. Based on this type of result, a pipe at this level of degradation may have
occasional failures from corrosion holes but it is structurally sound.
2.12. Results of Pipe with 9% degradation
Figure 2 shows photographs of a section of pipe measured at 8.9% average loss. The
physical thickness of this pipe was measured at 8.8+/-0.8mm (0.35in +/-0.03in)(nominal
was 9mm), with average pit depth at 2.5mm / 3.0mm. The condition of the pipe was
rated as very good / moderate. The corrosion of this pipe was primarily localized
internally on the bottom of the pipe as can be seen in the right photo. The corrosion
appeared in this case more continuous perhaps due to sediment build up at the bottom
of the pipe. Overall the structural integrity of the pipe is good.
2.13. Results of Pipe with 47% degradation
Figure 3 provides photographs of a pipe with a measured 47.3% average loss of pipe
wall thickness (11.0mm, 0.43in nominal). In the lab the average physical thickness was
measured as 11.6+/-3.3mm (0.456in, +/-0.13in) and an average pit depth of 3.8mm /
2.5mm. The physical condition of the pipe was described as very poor / poor. Note that
there were also numerous through holes in the pipe evident after sand blasting. It is
interesting to note that the pipe was not leaking when measured, probably due to the
build up of tuberculation.
12
-------
Cast Iron Pipe Pilot Study
Figure 1: Photos of pipe with 4.2% measured loss
Figure 2: Photos of pipe with 8.9% measured loss
13
-------
Cast Iron Pipe Pilot Study
Figure 3: Photos of pipe with 47.3% measured loss
Guidelines for Interpretation of Results
Based on the results, we recommend the following guidelines for the interpretation of
our data:
• 10% or less: The pipe is in very good condition, but may still have minor levels of
uniform corrosion. Some localized areas of pitting corrosion may exist but it is
expected that the areas are isolated.
• 10-20%: Pipe is in good condition, there may be some moderate uniform surface
or internal corrosion, or more localized areas of pitting corrosion.
• 20-35%: Pipe may have significant localized areas of pitting corrosion, or
moderate uniform corrosion throughout.
• >35%: Pipe is in poor condition and may have numerous areas of pitting
corrosion, including significant uniform thinning of the pipe.
14
-------
Cast Iron Pipe Pilot Study
3. Methodology
3.1. Leak Detection
In general, it is more challenging to survey for water main leaks with a leak noise
correlator than using it to pinpoint a leak, which is known to exist, as there will be a high
incidence of negative (no leak) results. When many negative results are encountered,
the surveyor may begin to question the operation of the equipment, or his procedures.
Therefore, one of the main issues with testing pipes where there is no known leak is to
ensure that the proper steps are taken to ensure that the results are properly analyzed
so that the presence (or lack of) a leak may be definitively decided. Based on our
previous experience with leak detection surveys, and our familiarity with acoustic
technology, procedures were implemented for both on site, and follow-up analyses were
performed in order to make a definitive decision on whether or not a leak was present.
1. Sensors were attached on valves or hydrants as available at each site. Where
measurements were performed on valves, the sensors were placed on the tops
of valve keys that had been lowered onto the valves or placed directly on the
valve nut when possible (if the valve chamber was clear of debris).
2. The LeakfinderRT radio channels are color-coded blue and white, where blue is
always the right audio channel and white the left. For all measurements, the
locations of the blue and white channel were noted.
3. In general, all leak detection measurements were taken on the same segments
of pipe where the condition assessments were performed.
4. After placement of the sensors on the appropriate valve or hydrant, the fitting
was tapped, and listened to at the radio receiver to ensure that the sensor was
functioning, and that the radio signal was arriving properly at the receiver. This is
called a scratch test.
15
-------
Cast Iron Pipe Pilot Study
5. Where possible, sensor spacing was accurately measured using a calibrated
measuring wheel.
6. A correlation measurement was performed, and the signal was saved to the
computer, so that further analysis could be performed later in the office, and so
that the client could have a permanent record of the raw noise file if needed.
7. Where a positive signal was detected (a correlation peak with good signal
coherence), the location was immediately checked to determine if it
corresponded to a service line or other notable draws from the pipe. If this was
the case, several more correlations were conducted to see if the 'usage' stopped.
8. Where negative results were obtained (no clear correlation peak was obtained), a
series of checks was completed, including a review of coherence and of the blue
and white frequency spectra, to detect the presence of a PVC repair or some
other anomaly in the test section. Such checks have become part of our protocol
for leak detection surveys.
3.2. Condition Assessment
The following survey methodology was used:
1. For each location surveyed, the distance between the sensors was measured. A
very accurate measurement of the distance between sensors is required.
Although less important for leak detection measurements, an error in
measurement of even 3 feet over a 300 foot distance can lead to errors of 15% in
wall thickness estimation. The margin of error acceptable will be dependent on
the pipe type and the distance between sensors. Typically, for a cast iron pipe,
we have not found it difficult to obtain this measurement accuracy. There were
some cases where accurate pipe geometry was not available. For example,
elevation changes and curves in the road may create discrepancies between our
distance measurement along the surface and the physical distance of the pipe
underground. Any locations that presented this difficulty were noted and will be
discussed in the final results.
16
-------
Cast Iron Pipe Pilot Study
2. Sensors were placed on the fittings, either hot taps that were previously installed
or in potholes on the surface of the pipe, and a noise source was created,
typically at a location out-of-bracket (beyond one of the sensors). The noise
sources were either a running well, light impacting on valves or use of the
shaker. Some sites permitted the use of all 3, others were limited to 1 based on
space restrictions
3. The temperature of the water was recorded, generally at the time of testing, for
each of the test sites.
4. The data was stored as a raw wave file for further analysis and confirmation in
our offices. Data was reanalyzed and filtered to obtain an optimum correlation
peak.
3.3. Instrumentation
The leak detection was completed using Echologics' proprietary leak detection system,
LeakfinderRT. The system works by placing sensors on two water system fittings such
as valves or hydrants bracketing the leak. If a leak is present, the software then uses
the time difference it takes the leak noise to reach the two sensors to pinpoint the leak
location. The sensors used for the purposes of this project were surface mounted,
either on hydrant flanges, hydrant secondary valves or line valves. There were two
types of sensors used in this study:
• Echologics' proprietary Hydrophones for direct measurement of the water column
• Echologics' piezoelectric accelerometers, with a sensitivity of 1 V/g
Each sensor has its own specific attributes that make it preferable in certain situations.
The Hydrophone is particularly well suited to measuring asbestos cement and medium
to large diameter mains (12in and larger), as leaks on these pipes generally are
dominated by lower frequency content (200Hz and below). The standard piezoelectric
accelerometer has a slightly higher noise floor, and has better high frequency response,
making them more suitable for some measurements on smaller diameter (10in and
17
-------
Cast Iron Pipe Pilot Study
lower) metallic pipes that typically have higher frequency content (200 Hz and higher).
Radios used were 460 MHz or 433 MHz analogue units manufactured by Echologics.
18
-------
Cast Iron Pipe Pilot Study
4. Results and Discussion
First, general information regarding the site location and the pipe will be discussed.
Following this, the results of the demonstration will be presented first, followed by the
results of the background measurements and the corresponding condition assessment.
A map showing the site location and the general layout can be found in Figure 7: Site
Layout.
Table 2: Excavation Locations presents a list with the locations of the excavation pits. It
shows the approximate distance between pits and a corresponding description of the
type of excavation. The distances presented were not the same distances used when
performing data analysis.
For the Leak Detection Demonstration, the pipe was broken up into three longer
sections. For the Background and Condition Assessment measurements the pipe was
broken up into seven sections. More sections were chosen for the assessment
measurements in order to provide a better representation of the pipe condition.
Pipe excavation locations
EPA technology Demonstration
ID
1
A
B
4
C
2
D
E
5
F
3
Distance
Feet
0
250
510
581
809
1080
1173
1439
1580
1750
2057
2100
Name
Lauch Pit
Sensor Pit A
Sensor Pit B
CorpValvel&2
Sensor Pit C
Corp Valve 3 45 & 6
Sensor Pit D
Sensor Pit E
CorpValve7&8
Sensor Pit F
Fire Hydrant
Receive Pit
12" Discharge
Type
6x8 with trenchbox, 12" T, Reducer
3x3 to top of pipe
3x3 to top of pipe
6x8 with trenchbox, stone backfill
3x3 to top of pipe
6x8 with trenchbox, stone backfill
3x3 to top of pipe
3x3 to top of pipe
6x8 with trenchbox, stone backfill
3x3 to top of pipe
Pressure gage
6x8 with trenchbox, 12 inch T, Reducer
Table 2: Excavation Locations
19
-------
Cast Iron Pipe Pilot Study
Location
Pitl to Pit2
Pit2 to Pit3
Pit4 to Pit5
Pitl to PitA
PitA to PitB
Sensor-to-
Sensor
Spacing (ft)
1080.7
979.3
1001.6
250.7
260.5
PitB to PitC | 298.6
PitC to Pit2
Pit2 to PitE
271
360.9
PitE to PitF 294.6
PitFtoPitS 312.7
Table 3: Sensor-to-Sensor Distances
4.1. Demonstration Results
The results of the demonstration tests are presented below in Table 4: Demonstration
Results. The column titled File # corresponds to the WAV file number in the name of the
file when it was recorded. It can be cross-referenced with the screenshots presented in
the Appendix. The column titled Type corresponds to the type of test that was provided
by Battelle. At each location there was four demonstrations the first of which, Demol
Cal, was a calibration test where the induced flow rate was known. The column titled
Location presents where the sensors were attached to the pipe. The column titled
Flowrate (GPM) presents either the known flow rate for calibration tests or the estimated
flow rate for the others. The column titled Result presents the outcome of the correlation
measurement, either negative or positive.
File
#
Id
If
ig
lh
2b
2c
2d
2e
3b
3c
3d
3e
Type
Demol Cal
Demo2
Demo3
Demo4
Demol Cal
Demo2
Demo3
Demo4
Demol Cal
Demo2
Demo3
Demo4
Location
Pitl
Pitl
Pitl
Pitl
Pit2
Pit2
Pit2
Pit2
Pit4
Pit4
Pit4
Pit4
to
to
to
to
to
to
to
to
to
to
to
to
Pit2
Pit2
Pit2
Pit2
Pit3
Pit3
Pit3
Pit3
Pits
Pit5
Pit5
Pits
Flowrate
(GPM)
0.6
Negligible
2.0 to 5.0
O to 1.0
None
5.0 to 8.O
5.0 to 8.O
Negligible
8.0
Negligible
5.0 to 8.O
2.5 to 5.0
Result
Negative
Negative
Positive -
Positive -
Negative
Positive -
Positive -
Negative
Positive -
Negative
Positive -
Positive -
577
560
6ft
7ft
476.8ft
478
502
497
487
8ft
9ft
8ft
4ft
from
from
from
from
from
from
from
Pitl
Pitl
Pit3
Pit3
Pits
Pits
Pits
Table 4: Demonstration Results
20
-------
Cast Iron Pipe Pilot Study
Section 1: Pit#1 to Pit#2, Demonstration in Pit#4
The calibration test, Demo 1, was performed with a known flow rate of O.SGpm. The
resulting correlation test presented a negative result. This suggests that a flow rate of
O.SGpm or less cannot be detected with hydrophones at a sensor spacing of 1080.7ft or
greater. Although the final result was negative this is still considered to be a successful
calibration test as it has defined a range that cannot be successfully correlated.
The flow rates in Demo 2, 3, and 4 were unknown. Demo 2 presented a negative
correlation test. This suggests that the flow rate is negligible and most likely to be close
to or below the calibration value, O.SGpm. Demo 3 presented a positive result at a
distance of 577.6ft from Pit #1. The character of the noise sources suggested a
moderate sized flow rate in the range of 2.0 to S.OGpm. Demo 3 presented a positive
result at a distance of 560.7ft from Pit #1. The coherence was very low and the
correlation peak was weak suggesting that the flow rate was low. It is estimated that this
flow rate is between 0 and I.OGpm but probably closer to I.OGpm as it is known that
0.6Gpm yielded a negative correlation.
Section 2: Pit#2 to Pit#3, Demonstration in Pit #5
The calibration test, Demo 1, presented a negative result with no flow out of the test
valves. This is as expected. Demo 2 and Demo 3 presented very similar results. The
correlated distances were within two feet of each other, 476.8ft and 478.8ft from Pit#3
respectively. Also, the character of the recordings was very similar suggesting that the
flow rates are almost the same. It is estimated that the flow rates are both between 5.0
and S.OGpm but the similarity in the signals suggests that it may be flowing from the
same orifice. Demo 4 presented a negative correlation result meaning that the flow rate
is close to or below 0.6Gpm.
Section 3: Pit#4 to Pit#5, Demonstration in Pit#2
The calibration test, Demo 1, was performed with a known flow rate of S.OGpm. The
corresponding correlated distance was 502.9ft from Pit#5. The coherence was very
strong and the correlation peak was prominent. Overall this test presented the loudest
of all file recorded suggesting that it is the highest flow rate of all the demonstrations.
Demo 2 presented a negative correlation result meaning that the flow rate is close to or
below 0.6Gpm. Demo 3 presented a positive correlation result at a distance of 497.8ft
21
-------
Cast Iron Pipe Pilot Study
from Pit#5. The recording had good coherence and a good correlation peak suggesting
that there was a high flow rate. It is estimated that the flow rate for Demo 3 was
between 5.0 and S.OGpm. Demo 4 presented a positive correlation result at a distance
of 487.4ft from Pit#5. The coherence was lower than the previous test but the
correlation peak was strong. It is estimated that the flow rate was between 2.5 and
S.OGpm for Demo 4.
General Comments
In some cases distance discrepancies between 2ft and 17ft is seen when the simulated
leak is being generated in the same excavation pit. It is known that there is more than
one valve in each of the demonstration pits but the distance between valves in the pit is
unknown. It is assumed that the discrepancies are mainly due to the fact the valves are
approximately 5ft apart, thus accounting for the difference. However, some of the
difference may actually be due to signal processing error, which can get worse as the
signal-to-noise ratio decreases. This may be the case for Demo 4, in Section 1: Pit#1 to
Pit#2, Demonstration in Pit#4.
4.2. Leak Detection Results
There were two positive leak locations discovered over the duration of the testing.
File #2a-Pit A to Pit B
File 2a was recorded with the Blue station on the pipe in Pit B and the White station in
Pit A with sensor spacing of 260.5ft. The correlation function shown for this file indicates
a leak at a position was 91.5ft from the White sensor. A sharp correlation peak and
moderate levels of coherence indicates a flow rate of 2.5 - 5.0 Gpm for this leak.
The evidence presented here strongly indicates the presence of a leak and if this pipe
were to remain in service, it would be suggested to perform remedial action.
22
-------
Cast Iron Pipe Pilot Study
Correlation Function
0.5-
•LI
T3
I 0.0 i
-0.5
-0.08 -0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08
Time (second)
Figure 4: Correlation result for File #2
File#7c-PitFtoPit3
The correlation function shown for File 7c was recorded with the Blue station mounted
to the pipe in Pit F and the White station mounted to the pipe in Pit #3 with a sensor-to-
sensor spacing of 312.7ft. The character of the signal suggests that there may be two
leaks at this location at a distance of 126.6ft and 144.6ft from the White station. The
weaker signal and wider correlation peak indicates a small leak, which sets the
estimated flow rate at 1.0 - 2.5 Gpm for each leak.
The evidence presented here is not entirely conclusive because the correlation peak is
not defined. If this pipe were to remain in service, it would be suggested to perform
further investigation by either using a ground-microphone to confirm a noise source or
potholing to confirm the presence of water.
23
-------
Cast Iron Pipe Pilot Study
Correlation Function
JLJ
-0.10 -0.05 0.00 0.05
Time (second)
0.10
4.3.
Figure 5: Correlation Result for File #7
Condition Assessment Results
The results of the condition assessment measurements are presented in Table 5:
Condition Assessment Results. Starting from Pit #1, three sections in a row presented
remaining equivalent thickness greater than 0.875-inches. This suggests that there is
minimal deterioration in these sections and the pipe is in good structural condition. Of
the remaining four sections of pipe between Pit C and Pit #3, three of them presented
remaining thickness below 0.875-inches. These are marked with an asterisks in the
table. This suggests that these sections of pipe have experienced slightly higher
corrosion rates although the pipe is still in good structural condition. The section
showing the highest losses is between Pit F and Pit #3. It presented a remaining
equivalent thickness of 0.85-inches.
It should be noted that none of the sections tested presented results significantly below
the nominal values. This suggests that, overall; the pipe is still in good condition.
24
-------
Cast Iron Pipe Pilot Study
File #
la
2c
3c
4b
5d
6c
7b
Location
Pitl to PitA
PitA to PitB
PitB to PitC
PitC to Pit2
Pit2 to PitE
PitE to PitF
PitF to Pit3
Sensor-to-
Sensor
Spacing
(ft)
250.7
260.5
298.6
271
360.9
294.6
312.7
Measured
Average
Thickness
(inch)
0.89
0.91
0.91
0.87
0.87
0.88
0.85
Condition
Good
Good
Good
Good*
Good*
Good
Good*
Table 5: Condition Assessment Results
25
-------
Cast Iron Pipe Pilot Study
5. Concluding Remarks
We thank you again for the opportunity to test the technology and we trust that this is
acceptable. Please do not hesitate to contact us if there are any questions regarding the
study.
Sincerely,
Echologics Engineering Inc.
Marc Bracken, M.A.Sc., P.Eng.
Dave Johnston, B.Eng. Materials Engineering
26
-------
Cast Iron Pipe Pilot Study
6. Appendix
27
-------
Cast Iron Pipe Pilot Study
Figure 6: Pipe Wall Cross-Section
-------
Cast Iron Pipe Pilot Study
Pit tt2 (at 171*00: intorsoction of Wostport Rd and St. Matthows Ave.)
Pit Dimensions: 8 ft long x 5 it wide with standard [ranch MX
Pipe Length Exposure: 8 ft at bottom
inj M.Hnl If slrtlUI Oil Min
Pit #3 (near Ridgeway Ave. at location of 24" x 12" tee)
Pi; Dimensions: 8 ft long x 5 ft wide with standard trench box
Pipe Length Exposure: 8 ft at bottom
Tap Irstallatipn: One 24" x 12" tee (with adapters to 2" and 6")
Equipment Installation: 24" x 12' tap tor flushing purposes
P'pe Lengn -xpcsure: 3 ft at centertire
Eiiuii iiimii li siHlrttun None
Pit #1 (near Chenowetti Ln. at location of 24" x 12" tee)
^Pit Dimensions: B it long x 5 ft wide with standard trench box
Pipe lengtti Exposure: 8 ft at bottom
Tap Installation: One 24" x 12" tee (with adapters to 2" and 6"|
Pipe Length Exposure: 3 ft at cenlerllne
Equipment Installation None
EflLJjpjnen! Installation: A 100 psi Bourdon lube pressure gauge
WESTPORT RD
Chenoweth Ln
to
Ridgeway Ave
9 vacuum Excavstlon Keyhwea (6* lo 8" ciamotei
RreHyonrt
Figure 7: Site Layout
-------
Cast Iron Pipe Pilot Study
LF*r LeakFinderRT - Correlation Results (8.5 x 8.6 inches)
Prin
t Exit
a
fe
S
Blue Station Frequency Spectrum White Station Frequency Spectrum
2.5+«T
2JQ+B2-
1 15«2-
jc' 1 .Q44E- 1 1
S^'$/LL
0 100 200 300 400 500 600 700 800 900 1000
Frequency (Hi)
1.5+c
.g 1.0+c
±±
c
S 5.0+0
0.0+i
Coherence Function Correlation
0.2 T
o.i • 1
1 1 ll 1
S mlfi^^tJ^L*^
0 100 200 300 400 500 600 700 800 900 1000
Frequency (Hi)
1.0
0.5'
0)
T3
& 0.0-
03
-•
J
J]y
^L,____
j.
0 100 200 300 400 500 600 700 800 900 1000
Frequency (Hi)
Function
Jj
*
AW»
)r4!nNJta
ifr
0.08 -0.06 -0.04 -0.02 0.00 0.02 0.04 0.06 0.08
Time (second)
Leak is 169.0 ft from Blue station and 91 .5 ft from White station, Time Delay = -0.02041 s
FileName:2aPitAto PitB_24inch_CICL_79.4m_LD_edit2.wav
Pipe Type: Cast Iron
Pipe Diameter: N/A
Pipe Length: 260.5 feet
Wave Velocity: 3794 ft/s
Frequency: 225 to 550 Hi
Figure 8: Correlation Report for File #2a - PitA to PitB
-------
Cast Iron Pipe Pilot Study
LeakFinderRT - Correlation Results (8.2 x 8.6 inches)
Print Exit
Blue Station Frequency Spectrum
White Station Frequency Spectrum
2.5+«-
2.Q-WB-
Oj
B 1 S-HE-
«
ra1 1 .O-KE-
C
2
k
) 200 400 600 800 1000
Frequency (Hz)
Correlation Function
400 600
Frequency (Hz)
800
1000
-0.5
-1.0
-0.10 -0.05 0.00
Time (second)
0.05
0.10
Correlation Result
Leak is 186.1 ft from Blue station and 126.6 ft from White station, Time Delay = -0.01587 s
Input Data
FileName: 7c PitF to Pit3_24inch_CICL_95.3rn_LD_edit2.wav
Pipe Type: Cast Iron
Pipe Diameter: N/A
Pipe Length: 312.7 feet
Wave Velocity: 3751 ft/s
Frequency: 225 to 550 Hz
Figure 9: Correlation Report for File #7c - PitF to Pit3
IV
-------
APPENDIX D
INSPECTION VENDORS COMMENTS TO FINAL REPORT
D-l
-------
Lessons learned from the demonstration:
After passing over simulated leaks, the Sahara hydrophone was tested on-site
and found to have technical problems. Subsequently, that particular
hydrophone was replaced with an alternate hydrophone confirmed to pass
quality control/assurance tests. Two of the very small leaks were re-simulated
and were detected on-site using the new hydrophone. As a precaution, all
Sahara hydrophones are tested onsite following standard QC/QA procedures
prior to inspection.
Improvement to the equipment used for leak detection since
demonstration:
Sahara Leak Detection is a mature technology used successfully for many years
and future development of the technique will focus on making it even easier to
use. The main challenge with Sahara Video is to improve its video and lighting
quality in larger diameter pipes and to possibly combine the video and leak
techniques into a single sensor head which would reduce the amount of
insertions required and make the overall inspection more efficient.
www.ppic.com
-------
TECHNOLOGIES
~"
Pure Technologies' Comments on Report
to 62-fmal reportj>artl leak detection revl 09152010-fmal.pdf
This document outlines the comments Pure T echnologies Ltd would 1 ike t o s ee acknowledge
regarding the report issued by Battelle/Alsa tech for the US Environmental Protection Agency
regarding the SmartBall technology titled "to 62-fmal reportj>artl leak detection revl
09J520JO-fmal.pdf".
Section 3.3.2
The report states that "The potential for the SmartBall® to be lost exists if the direction of flow
suddenly changes or another activity (high customer use or hydrant flow) diverts the sensor from
the planned inspection path." Please note that SmatBall® have never been lost due to activities
such as hydrant flow or high customer demands.
Table 3-2
The table states that the required flow for SmartBall is ~0.8ft/s-l.5ft/s. Please note that
SmartBall have done inspections at flow as low as 0.5ft/s and as high as 7ft/s.
Table 3-2
The final column of Table 3-2 describes "Pipe Access Requirements" for each technology. It
describes access frequency for LeakfinderRT but does not but not discuss the requirements for
SmartBall® or Sahara. Such additional columns might resemble:
Technology
Sahara
SmartBnll
Leakfinder RT
Pipe Appurtenance Required
2" diameter or larger full
port valve with vertical
clearance of 10 feet
4" diameter or larger full
port valve with vertical
orientation and clearance of
12 feet
(?)
Frequency of Required Appurtenance.
Every 2500 ft upstream of the section to be
inspected
Twice, one at beginning and one at the end of
inspection
Every 300 to 2500 ft
We continue to reference the data from this test to refine our algorithms and methods, it's a great
source since we rarely (if ever) get to run the same line so many times.
Pure Technologies
Unit 300, 705 11th Ave SW
Calgary Alberta, Canada
T2SOJ1
TECHNOLOGIES
-------
au nonson unve. unn 133
Toronto. Canada M9W 163
F: 41 6 249 8833
Toll Free; 1844 3246564
echolo^ics
O
Echologics Comments on the Report:
FIELD DEMONSTRATION OF INNOVATIVE CONDITION ASSESSMENT
TECHNOLOGIES FOR WATER MAINS AT LOUISVILLE, KENTUCKY
PART 1: LEAK DETECTION AND LOCATION
Below are Echologics's direct comments on the report:
Section 2.0, Paragraph 1, item 3:
The statement: "pairs of hydrophones that contacted the water at discrete locations to
estimate simulated leak rates" is not correct. The leak rates can be estimated using
either hydrophones or surface mounted sensors. In this case, surface mounted sensors
were used to demonstrate leak detection and condition assessment capabilities.
Typically, surface mounted sensors are placed at shorter intervals to increase the
resolution of condition assessment results. This also has the added benefit of increasing
the sensitivity of the leak detection.
Section 2.0 Table 2-1:
In the pipe contact points row: "Every 1,000 ft for leak rate Every 300-400ft for
detection" is incorrect and should read: "typically every 1,000ft for leak detection and
rate, every 300-400ft for condition assessment". In this case, surface mounted sensors
were used at shorter intervals because condition assessment measurements were
already being performed.
Lessons learned from the demonstration:
This demonstration allowed us to confirm several of our hypotheses. Specifically when
dealing with large quantities of air pockets. As our technology cannot directly test for air
pockets within the pipe, we must test for it indirectly. The demonstration allowed for the
confirmation of a long term theory that air pockets are the cause of attenuation of
vibrations within the water column. Since the date of the demonstration, this theory has
been applied and confirmed at several other project locations, much to the appreciation
of our clients.
-------
au nonson unve. unn 133
Toronto. Canada M9W 163
F: 41 6 249 8833
Toll Free; 1844 3246564
echolo^ics
O
Also, in the past, Echologics technology was only used to qualitatively estimate the size
of a leak. The demonstration allowed us gain insight into the energy / leak size
relationship and create a more accurate model for predicting leakage rate.
Improvement to the equipment used for leak detection since the demonstration:
As Echologics is continuously striving to improve on our leak detection and condition
assessment technology and methodology, we have made several improvements since
this date, specifically within the realm of passive signal filtering. We have developed
new hardware that better allows us to focus only on leak noise and attenuate
background noise.
Echologics enjoyed the opportunity to demonstrate our technology and greatly
appreciated the efforts the EPA and LWC in organizing this endeavor. We hope to be
able to participate alongside our competitors in any future projects of this nature.
Sincerely,
n
Dave Johnston, B. Eng
September 23rd 2010
------- |