Electrical Resistivity Technique to
Assess the Integrity of Geomembrane Liners
Southwest Research Inst.. San Antonio, TX
Prepared for
Municipal Environmental Research Lab.
Cincinnati, OH
Nov 84
PB85-122414
of
rsal fc&rma&m Service
-------
PB85-122JMU
EPA-600/2-84-180
November 198/.
KLRCrRICAL RESISTIVITY TECHNIQUE
TO ASSESS THK INTEGRITY OF
CEOMKHBRANE LINERS
by
David U. Shultt
Bob M. Duff
Wendell R. Peters
Southwest Research Institute
San Antonio, Tew? 78284
Contract No. 68-03-3033
Project Officer
Carlton C. Wiles
Solid and Hazardous Waste Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGKNCY
CINCINNATI, OHIO 45268
-------
TECHNICAL REPORT DATA
( iY rcd IRJ(IIC *, o m, pt,nu Co”!Nennt
I.NIPQRyp 2 . .AIC1PIINV5*. I
EPA—600/2-84—180 PH 5 1
4. TITLI AMO L IPO T aISTI
Electrical Resistivity Technique to Assesa the ____________________________
Integrity of Geomembrane Liners
. AUTHQ 3) — - LPU PO sNQ OmOANIZATIOp 1PO Y NO.
David W. Shultz Wendell R. Peters
Robert H. Duff, Ph.D.
PI PQMu,NQ O OAI.uiATIOp, NAMS ANO AOO 53g 6 IMLNyNc.
Southwest Research Institute BRD1A
!‘.O. Box Drawer 28510 IT.CONY ACT1AS TNO
6220 Culebra Road 68 03 . .3 033
Sun AntonF,, Texas 78284
I 2. SPON3O INQ AGENCY ,dAU5 * 140 AOORI$J — IS P1f’ I o T
Munjct al Environmental Research Laboratory —-Cjn.,OH ___________________________
Office of Research and Development 50 Q* O ENi’vcoos
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268 EPAI60 0/14
IS. SIJPI .IMSMTARY NOTES
Carl on C. Wiles, Project Officer — 513/684—7795
Tø. AUS?NACY
Two-dimensional electrical modeling of a liner system was performed using computer
techniques. The modeling effort examined the voltagu distributions in cross sections
of lined facilities with different leak locations. Results confirmed that leaks in
the liner influenced voltage distributions on the surface of the lined fac lLty.
Based on this, a small physical scale mod. 1 of a lined facility was constructed.
Tests were conducted to measure the influence of leaks on the surface voltage distri-
butions of water—f illed and earth—filled liner systems. Several leak and current
electrode configurations were examined. Plots of measured surface voltages indicated
the presence and location of a leak. These tests verified that the electrical
resistivity technique could be used for detecting and locating leaks in liners.
Larger—scale testing was performed at a 1—acre lined water impoundment. An
experimental measurement system was designed and constructed to facilitate data
acquisition for various leak locations, water depths, and leak sizes. These data
showed the technique to be capable of detecting a 1—inch diameter hole in the liner
with an error in location of potentially less than 1 foot.
7. SV WORDS AND OCCIJUENT
QISCRIFT0RS
ANALYSIS
b.IOsNTIFutp5/op5
EP4OsoTeu s
C. COSATI F eldJCto p
99
IS. OISTRIOUflQN STATEMENT
Release to Public
I 35CU ITV Ct,A53 9ThuRepo r
Unclassified
2?. NO.01 lACES
78
20. ESCUPITY CLASS Th&s pigs’ 22. PRIcE
Unclassified
EPA P 2220..I (5... 4 —I7 •R*v,OuI tOulION 1 OOSOI.LTE
I
-------
DISCLAIMER
The fnformj tion in this document has been funded wholly or in part by the
United States Environmental Protection Agency under Contract No. 68—03—3033 to
South’s’est Research Institute. It has been suhiect to the Agency’s peer and
admin stratjve review, and It has been approved for publication as en EPA
document. Mention of trade names or commercial products does not constitute
endorsement or recom’ endat(on for use.
it
-------
FOREWORD
The U.S. Environmental Protection Agency was created because of increas-
ing public and government concern about the dangers of pollution to the health
and welfa•:e of the American people. Noxious air, foul water, and spoiled land
are trng tc testimonies to the deterioration of our natural environment. The
cornplexl.ty of that environment and the interplay of its components require a
concentr3ted and tntegrnted attack on the problem.
Y.esearch and development is that necessary first step in probl2m solu-
tion, and it involves defining the problem, measuring its impact, and search-
ing for solutions. The Municipal Environmental Research Laboratory develops
new and improved technology and systems to prevent, treat, and manage waste—
water and solid and hazardous waste pollutant discharges from municipal and
coumunity sources, to preserve and treat public drinking water suppliEs, and
to minimize the adverse economic social, health, and aesthetic effects of pol—
lation. This publication is one of the products of that research and is a
most vi-tal communications link between the researcher and the user community.
This report d’,cumertts the development and demonstration of a electrical
resistivity technique co assess the physical integrity of geomembrane 1iner8.
The technique works by detecting and locating potential fluid leak pathB in
the liner. The developmental, procedures implemented to create and test this
technique are described and test results are presented.
Francis T. Mayo
Director
Municipal Envirotimental Research
Laboratory
iii
-------
PREFACE
Ceomembrane liners (often called flexible membrane liners) are sheets of
plastic or rubber materials used to contain fluids, most often in earthen
structures. Facilities where these liners are commonly used include water
reservoirs, landfills, and surface impoundments. At certain types of facili-
ties, it is important to know whether the liner is intact and performing its
intended containment function. Examples are hazardous waste landfills and
fluid impoundments. Loss of fluid from these facilities may create ground-
water and surface water contamination. Ceomenbrane liners are ln8pected to
various degrees or physical integrity during installation. Visual inspection
combined with F’eam testing is typical. However, no ‘nethod has been available
to test the entire geomembrane liner system. Such testing is desirable
because repairs could easily be made before placing the liner into service.
Efforts to monitor the performance of liners have typically relied on ground-
water sampling .:round the facility as an indirect indication of leakage. This
method is not precise enough for monitoring liner performance, sinca ground-
water contamination may take years to occur. In addition, monitoring wells
may never detect contamination if they are not located in the contqminated
groundw ,ter aquifer.
Recogniziig the need for a method to eteet and locate leaks in geomem—
brane liners, the U.S. Environmental Protection Agency intttiated a program to
develop and demonstrate several techniques or echniques to detect and locate
leaks in geomembrane liners. The techniques included time domain ref lectom—
etry, acoustic emissicn monitoring, and electrical resistivity. This report
presents the results of the project cc.nducted to develop and demonstrate the
electrical resistivfty technique.
iv
-------
ABSTRACT
This report documents a surface electrical resistivity technique to
detect arid locate leaks in geoinembrane liners. These materials exhibit resis-
tance to the flow of electrical current. Thus a fluid path through the geo—
membrane will form a detectable electric current path by which a leak may be
detected and located.
Two—dimensional electrical modeling of a liner system was performed using
computer techniques. The purpose of this modc.ling effort was to examine the
voltage distributions in cross sections of lined facilities with different
leak locations. Results confirmed that leaks in the liner influenced voltage
distributions on the surface of the lined facility.
Based on these results, a small, three—dimensional, outdoor, physical
scale model of a lined facility was designed and constructed, Te8ts were con-
ducted to measure the -influence- of leak -paths on—the surface voltage distribu-
tions of both water—filled and earth—filled liner systems. Several leak and
current electrode configurations were examined. Contour plots of measured
surface voltages indicated the presence and location of a leak. These model
tests verified that the electrical resistivity technique could be used for
detecting and locating leak paths in liners. Large testing of the technique
was necessary for further defining performance and developing measurement and
data analysis procedures.
This large—scale testing was performed at a 1—acre lined water impound-
ment located at Southwest Research Institute, An experimental measurement
system was designed and constructed to facilitate data acquisition for various
leak locations, water depths, and 1 .eak sizes. These data showed the technique
to be exceptionally well suited to the applicatton and ccpable of detecting a
1—inch diameter hole in the liner with an error in loca ton of less than 1
foot.
This report was submitted in fulfillment of Contract No. 68—03—3033 by
Southwect Research Institute under the sponsorship of the U.S. Environmental
Protection Agency. This report coverq the period September 1980 to July 1984,
and work was completed as of November 1983.
V
-------
• . . .
• . I • I •
• . . I • • • • • •
Conversion Table.
•
1. Introduction . . . • . . . .
2. Conclusions . • • . • • • . .
3. Recommendations • . .
4. Research Approach
Background
Concept Development . . .
Large—Scale Testing
5. Results and Discussion. • • . •
Overview. •
iJo—Leak Survey Results
Simul ,ted Leak Survey Results
Seam Leak Simulation Results.
Summary of Large—Scale Testing.
• I • • • • • • . . . . . . iii
• . . . . . . V
• . . . . . . . . . . . • . viii
. • . xi
• . . . . . . . . . . . . • xii
• . . I • • • •
•
• . . S S • • • • • • . . I •
• . . . I I • • • • • • S I S
S • • • • • I S • • • • • •
. I I • • I
• . . I • I I • I • • • • • 34
• . . . • . . . . . • . . . 53
• ••. . . 53
I • • • 53
• . . . . . . . . . . . . . 60
• 69
• 75
CONTENTS
Foreword . . • • • .
Abstract
Figures •
Abbreviations and Metric
Acknowledgments. • .
1
2
6
8
8
9
References
. . . . . . . . . • • . . . . 77
vii
-------
FIGURES
Number Page
I Conceptual electrical resistivity testing technique applied
to detect and locate leaks in a geomembrane linet system . • • • 11
2 Two—dimensional computer model of a liner flawing no penetra-
tion leaks but with an electrical current path over the
upper edges. . . . . . . . . . . . . . 19
3 A rectilinear array netwr rk (partially showa) of 21 x 11
resistors used to model a geomembrane liner. . • . • . • . . . . 19
4 Two—dImensional computer model of a liner with a leak. . • • . . . 20
5 Two—dimensional computer model of a liner with a leak at the
bottom of the—berm • • • • • • • • 20
6 Instrumentation used to conduct physical scale model teats . . . . 23
7 Block diagram of In8trumentatjon used to conduct model tests . . . 24
8 Phys ca1 scale model shoving the plastic liner and shunt
reaiatcrs. • . . . . 25
9 Potential measurerrents being taken u8ing the fiberglass beam
which supports electrodes . . 25
10 Equipotentja plot of voltages on the surface of the water
with a singl.e leak 28
11 Equipotentjaj plot of voltages on the surface of the water
with a highly conductive aluminum block and one leak 28
12 Equipotentjaj plot of voltages on the surEae of the water
with a single leak near the edge of the liiier 30
13 Equipotentja plot of voltages on the surface of the water
with two leaks 200 apart 30
14 Equipote . tjal plot of voltages on the surface of the water
with two leaks 400 apart 30
15 Equipoten j 5 plot of voltages on the surface of the water
with a single leak 31
viii
-------
FIGURES (ContInued)
Number
16 EquIpotentja plot of voltages on the surface of the soil
with a single leak • • 31
17 Equipntentj plot of voltages on the surface of the soil
with a single leak
18 Equipoten j 8 plot of voltages on the surface of the soil
with two leaks 400 apart 33
19 Plan view of the test impoundment • . . , . 35
20 Cross section A—A’ of the test impoundment . • 35
21 Test impoundment construction activity • 36
22 LIner installation activities • 37
23 Installation of a pipe designed as one of five leaks- through-
the liner 38
24 Rings used to hold plastic inserts flosigned to serve as leak8 . . 40
25 Simplified block diagram of the equipment used to test the
electrical resistivity technique 41
26 A mobile laboratory trailer, shown at the upper r. ght, is used
to hcuse the instrumentation during testing 43
27 Testing instrumentation inside the mobile laboratory 44
28 The electrically operated logging winch used during testing to
control the locattoii ot the measurement electrode 46
29 The circular brass disc below the end of the platform is used
as the current source eiectrode during teuting 47
30 The reference electrode and pr?ampllfier housing are shown to
the left of the potential measurement electrode box 49
31 Conceptual drawing of the test imnoundsent and operation of
the Lest equipment 50
32 A plot of tIe putential measurements taken along the 900
radial for a no—leak cage 54
33 An equIpotentta contour plot of no—leak measurements 54
ix
-------
FIGURES (Continued)
Number Page
34 A plot of the potential measurements taken along the 900 radial. . 56
35 An equipotentlal o1tour plot shoving the distortions from a
1—foot diameter leak on the 9Q0 radial . . • . . . . . . . . . . 56
36 A plot of the potential measurements taken along the 90° radial. • 58
37 An equipotential contour plot showing the diet jrtiona from a
1—inch d 4 anLcer leak . • • • • . 58
38 A plot of the potential measurements taken along the 115’ radial • 60
39 - An equ 4 raotential contour plot showing the distortions from a
1—foot diameter leak 60
40 A plot of the potential mea8urements taken along the 900 radial. . 61
41 An e ’iipotentja1 cortour plot showing the di8tortions from a
1—foot- dtameter leak . • • • 61
42 A plot of the potential measurements taken along the 90° radial. • 63
43 .n equipotential contour plot showing the distortions from a
1—foot diameter simulated leak 63
44 A plot of the potential measurements taken along the 90° radial. . 64
45 An equipotential contour plot showing the distortions from a
1—foot square simulated leak in 5 feet of water 54
46 A plot of th potential measurements taken along the 90° radial. • 65
47 An equipotential. contour plot showing the distortions from a
[ —foot square simulated leak in 6.5 feet of water 65
48 A plot of the potential measuraments taken along the 90° radial. • 67
49 An equipotential contour plot showing the distortions from a
0.1—foot square simulated leak without rnet’l. pans 67
50 A plot of the potential measure nenrs taken along the 9(J° radial. . 68
51 An equipotenttal contour plot showing the distortions from a
0.1—foot square simulated leak with metal cans surrounding
the leak 68
52 A plot of the potential measurements taken along the 900 radial. . 70
x
-------
FIGURES IContlnued)
Number Page
53 An equipOtefltj contour plot shoving the distortions from a
O.4—foot square, 3 0—foot long wire used to 8imulate a seam
leak . • • • • . . . . . . . . . . . . . . . . . . . . • • . . . 70
54 Plots of measurements along the 900 and 1100 radials 72
55 An equipoten j contour plot showing the distortions from two
1—inch diameter leaks on two different radlaig • ., ., 72
56 Plots of measurements along thc 90°, 110° and 115° radlaig 73
57 An equipoten j 8 contour plot showing the distortion 8 ftom
three I—foot diameter leakR n three different radials . 73
58 A plot of measurements along the 90° radial 74
59 An equtpote j 8 contour plot showing the distortions from
three different sized leaks along the 900 radial 74
TABLES
Nuriber
Electrical Propertj of Selected Geornembrane Liner Materials... 10
xi
-------
LIST OF AB REV1ATlONS
— Environmental Protection Agency
— flexible membrane liner
— one thousandth of an inch
— hertt
— alternating current
— high—density polyethy&ene
— decibel
— dir ’ct current
inches to centimeters (cm)
feet to meters (m)
ntis to centimeters (cm)
ntis to millimeters (mm)
acres to heetares (ha)
* 2.54
* 0.3048
* 2.54x103
x 2.54x102
x 0.4
EPA
FML
nit
Hz
AC
HDPE
dB
DC
FACTORS FOR Q)UVERTtNC DATA 1’J U.S. CUSTOMARY UNITS
TO St METRIC UNITS
xi i
-------
This report s prepared by the Department of Geoectencee, Southwest
Research Institute. The author8 wish to express appreciation to ( ‘anton C.
Wiles, whose vaj able assistance as Protect Off1ce helped guide this project
to a Buccee fuI Completion.
The following persona were responsible ror writing and editIng jor
po:ti ns of the text:
David U. Shultz
endell R. Peters
Bob M. Duff
Thomas E. Owen
The following persons were responsible for field data collection and
analyses:
Bob P4. Duff
David U. Shult
Frank X. Herzig
Kay Kuestpr
John Coo’ v1n
Cliff Q iley
xiii
-------
SECT(ON I
I NTRODUCT I ON
The United States generates large quantttieg of solid and hazardous
industrial and municipal wastes each year. The U.S. Environmental Protection
Agency (EPA) estimates that the annual production of hazardous wastes exceeds
57 miLlion metric tons. This amount represents approximately is percent of
all industrial wastes (Hanrahan, 1979: EPA, 1980).
The moat common method of disposal of these wastes is on land. Land
disposal includes landfills, surface impoundments, and land treatment. A
nationwide surve) reports there are 12,627 active landfills and 180,973 active
surface impoundments in the United States (Surface Impoundment Assessment
National Report, 1983). Fluids produced by landfills (leachateg) or contained
in surface impoundments are frequently toxic and contain substances that could
cause contamination if released into nearby surface or groundwater resources.
To prcvent contaninatton, geomembrane liner systems are often installed.
Placing these liners befleath the landfill or surface impoundments creates a
barrier that prevents or greatly inhibits migration of contaminating liquids.
Ceomembrane liners crad of plastic or rubber are useful because of their low
permeability. Geomemb-ane liners, often termed flexible membrane liners
(F?fl.’8), are sheets of polymeric materials fabricated in a factory and seamed
together at the field site to create a continuous liner system. Installation
practices as well as operational factors can lead to punctures or separated
seams, causing loss of physical integrity. Physical damage can also result
from exposure to incompatible waste fluids.
Sin e the goal of geomernbrane liner systems is only achieved when the
liner retains its physical integrity, initial inspection and in—service test-
ing of the physical integrity is important. In 1980, the U.S. Environmental
Protection Agency initiated a research prolect to develop a liner physical
integrity assessment technique. This effort has resulted in the development
and demonstration of an electrical resistivity tec’tnique t. determine liner
integrity. The technique detects and locates areas ot the liner system where
physical integrity has been breached by t.e passage of a conducting liquid.
The technique takes advantage of the high electrical tnsulating properties f
the liner compared w-ith the waste contained above the liner and the soil under
the liner. If the liner Is physicatly punctured or separated, the electrical
Conductivity of the fluid and underlying sot] form a current path by which the
leak may be detected and located. The technique can be applied on the surface
of an exiqtJnc l ndftll OC fluid impoundment and will not damage the liner.
It can be used to test new geomembrane liner systems before they arcs put into
service. The technique incorporates digital data arqirfc tjo. e’ ipncat,
-------
which eliminates human errors in manual reading and recording of measured
data, and it is portable to facilitate onsita application by tro or three
field personnel.
Southve t Research Inetitute ñeveloped and demonstrated the electrical
resistivity liner assessment techn’ e in three phases. The first phase
involved formulation of mathematical concepts and the computer modeling of a
liner system. Tvo—dimensional models vete used to examine the-electrical cur-
rent distributic,n in the cross sections of simulated landfills or surface
impoundments having electrically resistive geomembrane liner systems in place.
Results of the conceptual development and computer model studies demor.strated
the soundness of the electrical resistivity technique. The second phase
involved the construction of a small to— by 10—foot three—dimensional physical
scale model using a geometnbrane in a wooden frame. Results of the physical
scale model tests demonstrated that the electrtcal resistivity technique could
be used to detect and locate leaks in a geonembrane liner for both soil—filled
and water—filled test conditions. Based on these investigative studies, the
hardware needed to test the techniq’ie under large—scale conditions was
designed and aesen bled for testing in the third phase of the work. These
large—scale tests were performed in a 0.4—ha (1—acre) geomembrane—lined
facihty at Southwest Research Institute.
2
-------
SECTION 2
CONCLUSIONS
A surface electrical resistivity technique to examine the physical integ-
rity of a geomembrane liner system has been developud and demonstrated. The
technique performs by detecting and locating electrical current paths through
a geomembrane. The technique has excellent poteut 4 al for meeting industry
needs for testing liners installed in hazardoug solid waste and surface
impoundments. The technique can also be used to inspect liner integrity
imeediately after installation at new facilities, and to inspect in—service
geomembrancq without interruption of facility use.
Conclusions regarding the techniziue are:
(1) A geomembrane liner survey using the technique will not damage the
liner.
(2) The instrumentation and hardware required co apply the technique are
portable.
(3) The technique has been perforni2nce tented at a 1—acre, water—filled,
polyethylene—lined test impjundment. The tE.chnique has detected a
i—inch—diameter leak pat)’. in the liner and located it to within 1
foot of the actual loci tion In 5 feet of water. Similar results
have been obtained for n iltiple leak path configurations and tests-
in water depths ‘f 6.5 feet.
(4) The results of scale-model testing with soil fill suggest that a the
system may be used to survey liners installed at existing landfills.
(5) The results of experiments designed to determine the influence of
the depth of water on the measurements suggest that increasing depth
may reduce the magnitude of the measurements.
(6) The results of experiuents designed to measure the effect of highly
conductive metal objects located close to a leak path indicate that
the performance f the technique is not significantly influenced by
the presence of the objects.
(7) The results of an experiment simulating a seam leak in the liner
indicate that the geometry of a leak may influence the measurements
obtatned o. the surface of the water.
-------
SECTION 3
RECOMMENDATIONS
Based on the results of this prolect, the following recommendations are
made regarding the electrical resistivity technique:
(1) Evaluate th€. electrical resistivity technique at lined surface
impoundments and landfills where design and liner integrity status
information is available or can be experimentally determined. It
would be desirable to coi-.duct the first test at a fluid impoundment
to allow comparison of the results with the tests performed at the
1—acre test impoundment.
(2) Use the technique to test the integrity of an installed liner before
that liner is placed in service. This would allow ir.mediate veri-
fication of the test results, since access to any part of the liner
would be possible.
(3) Determine the applicability of the technique for testing other types
of l 4 ner materials such as asphalt, sprayed—on membranes, soil
cemert, clays, concrete, and soil sealants.
(4) DetermIne the applicability of the technique for testing double
liner systems (i.e., those in which two membrane liners are in-
stalled with a drain layer between the liners).
(5) Specify the modifications to existing equipment, instrumentation and
survey procedures that would be necessary to test liners at hazard-
ous waste impoundments or landfill sites.
(6) Continue the development of a computer model to predict the response
of liner systems to electrical resistivity testing, given facility
design information such as geometry, lirer type, waste type, surface
area, and depth of waste.
(7) Continue to develop computer analyses of liner survey results to
enhance the determination of leak signatures.
(8) Define and evaluate liner design and installation techniques that
would enhance th . applicability and usefulness of the electrical
resistivity testing technique.
4
-------
(9) Evaluate the applicability of the electrical reeiattvity eurvey
technique for conttnuous automattc r nitoring of geo emb ane liner
8y8te s.
5
-------
SECTION 4
RESEARQI APPROAQ
BACKCRO
Uners are installed in hazardous waste treatment and disposal facilities
such as landfills and surface impoundments to impede the flow of contaminating
fluidB into surface and groundwater resources. Since liners made from poly—
merle materials such as plastics and rubbers have very low water permeabilj—
ties, they are effective fluid containment membranes.
If the liner fails to provide phyotcal integrity, then the liner cannot
function as intended. Tears, ruptures, imperfect seam joints, or areal. degra-
dation would allow fluid to penetrate into the underlying soil, violating the
intent of the liner. Thus if a technique were available to test the liner
during installation before facility startup, during the operation of the
facility, and after the facility is closed, the goals of water resource pro-
tection would be better served,
Ideally, the testing method should be appropriate for assessing a brc j
range of liner terjalg. It should be inexpensive, portable, and suitable
for inspecting liners installed in any type of facility. The technique should
not require burial of sensing elements below the liner, since this would not
allow inspection of existing liner installations. The method must not only be
capable of detecting the presence of a leak, but also of locating the leak so
that apprc)priate remedial measures can be implemented. The method must oper-
ate Without damage to the liner.
In 1980, Southwest Research Institute initiated a project funded by the
U.S. Environmental Protection Agency to evaluate a surface electrical resis-
tivity technique applicable to these inspection and monitoring require 8 •
This technique wag based in part upon previ,ug work performed by the Institute
to develop a earth resistivity method for detecting small sub-
surface cavities. When this project was initiated, no techniques were avail-
able to test the physical integrity of installed liners, In addition, no
practjcaj methods were available to test the entire liner system after ins al—
lation but prior to facility startup. The rcmaining body of this report
presents the research studies conducted to develop and demonstrate an elec-
trical resistivity technique to meet the liner testing objectives.
6
-------
CONCEPT DEVELOPMENT
Theoretical and Mathenatjea Forn ulatjon
Ceomembrane liners made from impervious plastics a 1 d rubbers exhibit very
high electrical resistance. When a liner is installed in a lar.lfill or sur-
face impoundment, it effectively acts as an electrical insulator between the
materials contained in the facility and the surrounding earth. Electrical
properties of selected geomembra liner materials commonly used are presented
in Table 1. If the liner is physically punctured or separated, conductive
fluid flow through the leak establishes an electrical shunt through the liner.
This low resistance shunt forms an electrically detectable region wh!ch is the
basis by which a leak may be detected and located.
TABLE 1. ELECTRICAL PROPERTIES OF SELECT j _CEOMEMBRANE LINER MATERIALS
Thickness Area of Electrical Volume
of Specimen Specimen Resistance Pesistivity
Material Type (mus) (cm 2 ) — (ohms) (ohm — cm )
High Density
Polyethylene (A) 55.0 100.0 4.00 x 1013 2.86 x 1016
High Density 45.5 J.25 x 1013 4.81 x 1015
Polyethylene (B) 121.0
Pclyethylene 6.0 100.0 1.80 x 1011 1.18 x iØ 15
Chiorosulfonated
Polyethylene (A) 40.0 100.0 3.20 x 3.14 x 1013
Chiorosulfonated
Polyethylene (8) 36.0 100.0 3.75 x 1O 0 4.10 x 1013
Chiorosulforated
Polyethylene
— nylon reinforced 34.0 100.0 3.60 x 4.16 x 1012
Polyvinyl Chloride 29.0 100.0 1.52 x i 8 2.06 x 1011
Polyvinyl Chloride
— oil resistant 30.0 100.0 1.70 x 108 2.23 x 1011
Chlorinated
Polyethylene 32.5 9.0 7.20 x 1010 7.84 x 1012
chlorinated
Polyethylene
— reinforced 36.0 100.0 6.20 x 108 7.32 x 1011
Urethane Asphalt 69.0 100.0 2.20 x 1011 1.25 x 1014
- - —— --== -- - -— —--—
Source: Lahordtory tests performed by SwRI.
7
-------
Figure 1 shows thc basic surface electrical resistivity testing tech-
nique f,r detecting and locating a leak in a geomembrane liner. A current
8ource is used to inject current across the boundary of the liner. The liner
has an electrical leak path, as shown. When a voltage is applied between the
source and remote current return electrc es, current flows through the leak
as shown in the figure. 1’ a soil cover were present over the edges of the
liner, urrent would also flow through Lhis soil cover to the remote current
return electrode, bypassing the liner. Potentials measured on the surface
are affected by the current distributions near the leak and can be used to
locate the leak.
The success of the electrical resistivity technique in detecting leaks
depends upon the liner material having a high resistance per unit area coin—
pared with that. of a leak path. The current flow lines shown in Figure 1 for
a liner having a leak assume an esPentially perfect insulator for the liner
material. If the liner were a perfect insulator and there were no current
conducting paths through leaks over the top of the liner, no current would
flow. However, geomembrane liner materials have a high but finite resistiv-
ity. Therefore, some current will penetrate the liner under any conditions.
If the current density through the liner is assumed to be uniform, then the
total resistance which the liner presents a a load across the source cPn be
computed from the formula:
PLt
(1)
where;
the total liner ‘esl3tance (ohms);
P1. the volume resistivity of the liner material (ohmmeters);
t the thickness of the liner (meters); cnd
A a the total surface area of the liner (square meters).
For a liner 7.62 x 10 4 m (30 mils) thick and 2.01x10 4 m2 (five acres) in area
consisting of a material having a volume resistivity of 2 x 1010 ohmmeters,
Equation (1) gives a total liner resisteace of 750 ohms. Thus, even though
the resistivity of the liner material is high, the total circuit resistance
caused by the insulating liner can be relatively small for a large f. ility.
For practical liner materials, the location of the current electrode within
the lined landfill or impoundment snd the shape of the containment area will
result in a nonunifrom curtent density through the liner even for the no—Jeak
case. Equation (I) must therefore be considered as a first approximation to
the total resistance of the liner. Ni verthe1ecs, it is useful for estimating
performance f the technique.
The total electrical resistance load across a direct current sou’-ce con—
nected between the source cJrrent electrode inside of the impoundment region
and a remote current return electrode located at effective infinity outside of
the impoundment will consist of the series combination of a resistance, RF, of
the fill inside of the liner, the resistance, RL, of the liner and shunting
leak paths, and a resistance, RE, of tne earth surrounding the impoundment.
8
-------
REMOTE CURRENT
RETURN ELECTRODE
Figure 1. Conceptual electrical resistivity tes:ing technique applied to detect and locate Ieak In a
geomembrane liner system.
CURRENT SOURCE
ELECTRODE
-------
Each of these resistances is dependent upon the volume resistivity of the
medium and the geometry of the installation. The resistance, L , is the par-
allel combination of the resistance across the liner material, and the resis-
tance of any conducting leak paths. If an alternating current source is used,
the liner will exhibit a capacitive reactance In parallel with the total
resistance described above. At a large liner site, this capacitance can be
aignificant and its reactance even at freauencies as low as 10 Hz may be com-
parable to the resistance of the liner. This capacitive impedance of the
liner is primarily of importance in the design of the current source system
rather than in the interpretation of the measurements.
A mathematical model capable of predicting the potential at the surface
of an impoundment as a function of leak size and location will be beneficial
in evaluating electrical resttvity technique performance, as well as aiding
the interpretation of measurements. The field equations which must be solved
for the direct current case are:
V • E q /c 0 (2)
VXE O (3)
v.j=o (4)
where:
E electric field intensity;
J volume density of electric current;
qf volume density of tree charge; and
permittivity of free space (8.854 x I0 ”11 farads/meter).
Equation (3) allows the ‘ectric field to be expressed in terms of a scalar
potential function, 4, s
E —V4 (5)
Combining Equations (5) and (2) results in the familiar Poisson’s equation of
potential theory
V 2 d —qffc 0 (6)
Since free charge can exist only at the source and on surfaces of discontin-
uity, Equation (6) reduces to Laplace’s equation everywhere within the regions
of interest
(7)
Equations (2) through (7) are strictly valid only for the direct current
case. When a low frequency alternating current is used, a quastsLatic form of
Max ’ell’s equations would be more accurate. Since the frequency used for the
meaaurements Is 17 Hz, even for the largest impoundments this change only
requires that Equation (4) be replaced by
10
-------
V J —jwqf (8)
where w angular frequency of the curreflt. This madtfjeatjon includes the
Capacitive impedance effect of the liner.
A useful understanding of the problem can be obtained fros an examination
of the boundsry integral equation:
T Pi o i J (R’ )
•(r) — f $(r’) da’ — P I n’ (9)
SL
where:
r the distance from the current injection electrode (inside the
f.icility) to the point on the surface at which the potcntial is
measured;
— the potential at point r on the surface;
I = the total injected current;
r’ the distance from the current inlection electrode to the point
of integration on the liner;
R t ie distance between the potential measuring point r nd the
integratiofl point r’;
J (r) the component of current density perpendicular to the surface
of integration, i.e., liner;
SL = the surface SL with leak area8 removed;
the volume resistivity of material within the facility, i.e.,
liquid or solid waste; and
n a unit vector normal to the surface SL.
The integratia -ig in Equation (9) are surface integrals ovet a surface
defined by the liner and leak areas. The potential •(r) on the liner surface
can be shown to be directly related to the current density crossing the sur-
face. Thus, both integrals in Equation (9) are dependent on the distribution
ot current and the resulting potential at the surface is determined by this
current distribution. while the total resistance of the liner at a large site
may be small, the resistance per unit area of liner material, is still very
high compared with a leak. The current density at a leak will therefore be
many Limes as great as that at any point on the liner. A leak may then be
expected to have a strong effect on the l ,ocaliz ’d potential mea -ured at the
surface of the impoundment.
A second feature which u13” be obser”ed in Equation (9) is that the inte—
grands on the right—hand side of the equation are inve’ sely related to the
11
-------
distance between the measurement point and the integrat on po ; .t on the liner
surface. The greatest perturbation—of-the surface potential by a leak will
therefore oécur directly over the leak. The conclusions which may be drawn
from Equation (9) are that the high current density flowing through a leak
viii be detected from potential measurements on the surface and that the dis-
tribution of potential on the surface will indicate the location of the leak.
Yor prediction of system performance at large impoundment sites, the
question of leak detectabuity as a function of depth of fill arises. An
accurate anawer to this question will require detailed numerical solution of
the mathematical model. However, an approximate analysis o 9 this problem can
be obtained from examination of scaling laws. In other vorde, rather than
attempting to vary only the depth of a given site, co- sider sites of similar
geometry but different overall dimensions. For leaks which are not too close
to the edges of the liner this aiialysis should provide a rcasonable prediction
of the dependence of potential distribution on the depth of fill.
If all length dimensions, t, are changed by a factor n so that,
LnnL (10)
where:
t — the original length unit; and
— the scaled length unit;
then, froii examination of Equation (9) or other basic equations it can be
shown that
— •(nr) (11)
Since this relationship is based on e scaling of all dimensions, the thickness
of the liner and size of the leak are scaled by the same factor as the depth.
The miltiplicative ft ctor 1/n may be expected to be signifi,antly changed jf
one liner thickness is retained. The distrIbution of potential as given by
Equation (11) is a good estimate. Thus, the oeaguremeflts at two sites, ne
being ri times as deep as the other and having a leak of n times the area, will
be similar if the data from the deeper site is plotted with a 1/n change of
length scale. For example, measurements colIecte(j for a one—inch diameter
leak at a depth of 5 feet are expected tn be similar to measurements for a 5—
inch diameter leak at a depth of 25 feet.
Model studies were performed to validate the electrical resist!-iity liner
survey technique. These studies provided an opportunity to experimentally
analyze the current distribution and sjrfac2 potentials resulting from various
leak configurations. This Information was needed to develop field survey
equipment. The modeling studies and results are presented in the next
sect ion.
12
-------
Co puter Modelfn Studies
Two—dimensional resistive network computer modeling studies were per-
formed to predict the tt,f!uence of an electrical current petietration through a
geomembr liner and aseuciated surface potential voltages inside a landfill
or fluid impound nt. The Influences sought were the gnhtude of the surface
potentials at given locations for cases where fluid leaks o olcctrical cur-
rent paths e ieted through the liner. The tvo-dimeflejo l resistor network
designed to afmulate a liner we. modeled using a general purpoet circuit simu-
lation computer program called s ic , This eoftvace allows alcstlat(on of cir—
cuit containing resfatore, enpflcitors, Inductors, and voltage end current
sources. The resistivity of the fill Ingide the liner (either soil or fluid)
and the earth surrounding the liner waa modeled using noriMlired resistance
values of ne ohm. The liner was characterIzed in the model by using parallel
I 0 00—ohm resistors along the path of the liner. To simulate a leak or conduc-
tive path through the liner, one of the bOO—ohm resistors in the liner was
replaced by a I—ohm resistor. Therefore, the leak path ha the same resis-
tance as the earth. In effect, the liner le thuø removed at tha. acation.
A soil cover over a geomembra iiner will serve as an etectrioit conduc-
tive shunt for eIectr( L current into he surrounding earth. This hunting
effect was modeled by using 2 ohm resistors at the surface. It was included
in the model since landfills ani surface impoundments may have soil covers. A
current Source wag used in the model to inject current Into the surfa,e of the
earth as modeled In the program.
The output of the model analysig is the voltage at each of the nodes
connecting the sensor elements. These values are stored in an array which is
then used in a second program which plots the results in the form of equi—
potei.tjal contours. Figure 2 illugtrat,.s a contour plot shoving the two—
dimensional potential distribtjtton sro; ,nd a liner without leaks. The current
flow over t1 e top of the liner into the surrounding earth is shown. The net-
work used to model th1 cross section was a rectilinear array of 2! by fl
resistors, shown in part in Figure 3, for a total of 494 elements. T”,’ cur—
rent injection point Is at the top center of the figure wIth the other refer-
ence electrode connected to a conducting path along the bottom and sides of
the cross Hection. This reference is far enough from the liner to be located
at effective infinity.
Computer Results——
Figure 4 illustrates the results of one of the computer generated two—
dimensional analyses of a liner having a leak in the bottom. The eg ipoten—
tial lines which appear in the figure are generated and plotted automatically
by the computer program described earlier. The outline of the liner las been
sketched in the figure and is represented by the three—sided trapezoidal fig-
ure in the center of the plots. The cquipotenr4 lines showing the voltage
distribution patterns were computer generate4 , The current flow paths are at
right angles to the equipotentja lines and were sketched In by hand.
This figure shows that, as predicted, when a leak penetration Is present
in the liner, current flow betwee 1 electrodes located Inside and outside the
13
-------
FIgure 2. Two-djmer sio l computer model of a liner having no penetration leaks but with an electrical
current path over the upper edges.
\
\
\
‘N
Figure 3. A rectilinear array network (partially shown) 0(21 x 11 resistors used to model a geomembrane
liner.
14
-------
Figure 4. Two-dimensionai computer model of a liner with a leak. Equipotentjal lines at the top are non-
symmetrical, indicating the presence of a leak.
Figures. Two-dimensional computer model of a liner with a leak at the bottom of the berm.
15
-------
liner will follow two paths, namely, through the leak and over-the buried
edges of the liner. Note the non—symmetry of the equipotentja line8 termi-
nating on the surface of the facility, above the liner. Tue voltage gradient
is clearly steeper along the surface on the left Bide of the current injection
electrode above the leak.
Figure 5 shows the results of another computer generated two—dimensional
analysis of a liner having a leak near the edge of the liner. The equipoten—
tial lines at the surface are influenced by the -leak in a different manner
than in the previous example. Thus, the location of the leak changes the
relative distribution of voltages on the surface. tt is this effect which was
examined and investigated further to develop hardware and data reduction tech-
niques to locate liner leaks.
The relatively inexpensive nature of the two—dimensional computer model
and its precise control make it a very productive method of investigating
electrical testing concepts related to liner leaks. Even though it is a two—
dimensional model that does not accurately simulate the three—dimensional
volume of a landfill or surface impoundment, the ability to accurately predict
the potential distr{bution in various simulations of liner conditions makes it
a valuable tool. The sensitivity of this model to a simulated leak is greater
than for a true landfill Condition because of the two—dimensional character of
the model. Nevertheless, the results obtained are, in general, indicative of
the full—scale equipment sensitivity requirements.
Based upon the results of the computer analyses, a three—dimensional
physical scale model was designed and constructed. The follo: ing section
describes the model and presents the experimental results.
. !! Ysical Scale Modeling Studies
After completion of the computer modeling studies, a three—dimensional
physical scale model was designed end constructed. The objectives of using
this model were to: (1) obgerve the surface voltage distribution patterns
created by a fixed liner geometry, injeccior, current location, fill material
inside the liner, and leak configuration; (2) test the system instrumentation
and assess its Suitability for full scale testing: and (3) determine the accu-
racy of the technique in locating a leak.
Model Design——
The model was designed to allow studies of the resolution capabilities of
the technique, i.e., how accurately it could locatp a leak in the liner. Out-
side dimensions were established at 10 feet on each side, for a total lined
floor area of 100 ft 2 . Macmum depth was established at one foot, allowing
variation of water depth during the experiments. A black polyethylene sheet
6 mils thick was selected as the geomembrane.
Instrumentatjon__
The instrumentation used to perform the test measurements is showi in
Figure 6. A block diagram of the instrumentation is shown in Figure 7. The
16
-------
: • a
T .
. :
Figure 6. Instrumentation used to conduct physical scale model tests.
17
-------
Figure 7. Block diagram of instrumentation tsed to conduct model tests.
1 g
TO CURRENT
SOURCE ELECTRODE
TO REMOTE
RETURN ELECTRODE
0
-------
equipment includes an oscillator to generate AC injection current ac a fre-
quency of 17 Hz or 25 Hz. A power ampitfier increases the power-of the oscIl—
lator outpt’t to 10 watts. The step—up transformer takes the output of the
amplifier and drives the current into the liner model. A current meter wired
in series with the source electrode Is used to measure the injected current.
The model was constructed outdoors and contained a plastic liner. The
model La shown in Figure 8. It was constructed in a sqeare with 2 x 12 foot
boards, each 10 feet long. A 6—nil polyethylene sheet was installed as the
liner. To simulate the electrical shunting effect of a soil cover over the
liner, l 0 00—ohm resistors along the edge of the model connected the water and
soil inside the liner to the earth outside the liner. These resistors are
shown In Figure 8 along the sides of the model.
Data Measuremente.....
Two electrodes were connected to the outputs of the AC current source.
The source electrode was placed at the center of the fiberglass beam shown in
Figure 8. This electrode was inserted into the water and soil fill during the
studies to inject :,rren . An infinity electrode wag driven into the earth
appro dmetely 300 feat away from the model. This distance was sufficient
to be Considered at effective !nfinity with respect to the scale model.
Surface potential measurements were made using the fiberglass beam which
supported 52 electrodes spaced 2.5 Inches apart. The fiberglass beam and
potential electrodes are shown in Figure 9. The stationary reference elec-
trode was positioned j the corner of the scale model. Potential measurements
were made between each of the electrodes on the fiberigass beam and the
reference electrode.
To obtain measurements, the fiberglass beam supporting the potential mea-
suring electrodes was positioned across the water surface with the current in-
jection electrode located on the beam. The beam was rotated arour’d the pond
at 100 Increments, beginning with 0° and ending at 340°. For each placement
of the beam, potential data were obtained by making measurements at increasing
radial distances away from the current injection electrode. By rotating the
beam 100 about the cucrent injection point after measurements were obtained,
measurements covering the entire surface of the scale model were obtained.
Figure 9 shows a technician taking measurements along the fiberglass beam.
A total of 34 sets of surface potential measurements were obtained using
the scale model, 23 with water fill and 11 with soil fill. The water—filled
measurements were designed to simulate a fluid surface impoundment. The soil—
filled measurements were designed to simulate a landfill.
Measurements were made with no leaks, one leak, and two leaks. Leaks
were created by driving a 1/2—foot diameter copper clad s eel stake through
the liner into the soil underlying the liner. Thig stake p . ovided a good con-
ducting path betwee the water or soil and the earth, whIle minimizing the
loss of water.
19
-------
Figure 8. Physical scale model showing the plastic liner and shunt resistors,
Figure 9. Potential measurements being tak i using the fiberglass beani
which supports electrodes.
20
-------
The water level was varied from 3.25 to 8.75 inches. This was done to
observe the influence of water depth on leak signatures. A metal block was
placed in the water for five measurement sets to observe the influence of a
highly conductive anomaly on leak detection and location.
After water—filled tests were completed, the model was drained and 2.75
inches of soil were placed into the model. Three tests were performed at this
depth. MdLti n 1 soil was added until the depth was 5.5 inches. Five tests
were conducted at this depth. The remaining three tests were performed with a
soil depth of 10 inches. A new current electrode con8tructed from one—half of
a metal sphere approximately 3 inches in diameter was used for the soil tests.
The increased surface area o this electrode over the rod shaped water elec-
trode provided good current lnj ’ction stability.
All 34 tests were run with the current shunts in place. This simulated
the possible effects that a soil cover over a liner would have on the current
flow. It was anticipated that leak detection and location would be more dif-
ficult with these shunts i’l place. The 1 r effect would be to diminish the flow
of current through a leak. thu8 reducing the strength of the leak 8ignature.
Results——
Surface potential measurements weri made with 4 inches or water in the
acale model. Figure tO pre;entg a polar coordinate equipotentj plot of the
measured data fro’i one experimental configuration. The current injection
electrode is shown in the center of the plot. A leak is shown on the 3600
radial. The distortion of the equipotential lines indicates increased flow
of current around the leak.
Figure 11 is a similar plot for a dIfferent current electrode position.
The current injection point is toward the corner of the coordinate plot. To
examine the influence of a highly conducting metal object on the leak signa-
ture, an aluminum block was osittoned along the same radial between the cur-
rent injection point and the leak. The equipoten jal lines exhibit the same
distortion characteristics as shown in Figure 10. No observable distortion
was created by the aluminum block. This test was one of five experiments per-
formed using the aluminum block. All five tests confirmed that the conducting
block did not noticeably affect the experimental signatures of leaks in the
liner.
Figure 12 shows the influence o. a leak near the edge of the liner on the
surface equipotential contour lines. This plot demonstrates the fact that the
edges of the liner do not interfere with leak detection. This was an impor-
tant determination as part of the process of defining the limitations of the
technique.
Figure 13 shows a plot with two leaks an radlals 200 apart. The equt—
potential lines were distored by both leaks. This test was conducted with 4
inches of water in the model. Figure 14 at -ova a similar two—leak test, with
the leaks on radials 400 apart. Here both leaks exhibit similar distortions
of the contour lines.
21
-------
I .
Figure 10. Equipotential plot of voltages on the surface of the water
with a single leak.
Figure II. Equipotentjal plot of voltages on the surface of the water
with a highly conductive aluminum block and one leak.
240
14
28
t . Ic
100 ’
eo’
00 ’
240’
320’
Go,
40 ’
20 ’
340’
0’
320 ’
-------
IRRENT INJECTION
EDGE
Figure 12. Equipotential plot of voltages on the
sur(ace’of the wate, with a s ngle leak
near the edge of the liner.
CURRENT
Figure 13. Equipotential plot of voltages on the
surface of the water with two leaks 20°
apart.
Figure 14. Equipotential plot of voli. ges on the surface of the water with two leaks 40° apari.
#1
INJECTION
23
-------
To assess the Impact of increased water depth on the leak signatures, the
water level in the model was increased to 9 inches. The results of this test
at the increased water level are shown in Figure 15. As shown, the contour
lines are less distorted with the dricper water. However, the di8tortions are
evident for several contour lines closer to the current electrode than with
the shallower water fill. This suggests that the signatures of the leaks may
be smaller In magnitude but larger in areal extent as water depth is increased.
Testing on the water—filled model (simulating a fluid impoundment) con-
fIrmed that leaks can be detectd and located with the technique. To determine
if the system would achieve the same performance on a simulated landfill, the
water in the model was replaced with a sandy loam. Figure 16 presents the
results of a t 8t wIth 2.75 inches of soil In the model and one leak. The cur-
rent Injection electrode Is shown in the center of the plot. The equipotential
contours are distorted in the area of the leak as with the water—filied condi—
tiori. The same test was performed after the soil depth was increased 100 per-
cent to 5.5 inches. Figure 17 shows the results of this test. The contour
lines close around the leak location, clearly identifying its location.
To determine the influence of two leaks with coil fill, a test was per-
formed with two leaks on radials 400 apart. The depth of soil was 5.5 inches.
Figure 18 shows the location of the leaks and their influence on the equi—
potential contour lines. Both leaks can be clearly Identified on the plot by
the distortions in the contour lines.
These studies further demonstrated the fundamental concepts of this
approach which were first defined by the computer modeling efforts. Based
upon these results, the measurement equipment, data processing software, and
electrodes were specified for testing at a 1—acre lined facility located on
Institute grounds. Testing at this scale was performed to measure the per-
formance of the technique at conditions which more accurately simulate actual
field conditions. The design of the 1—acre facility, instrumentation specifi-
cations, and experimental procedures are presented in the following section.
1..ARGE-SCALE TESTING
The electrical resistivity leak detection technique was tested at a 1—
acre lined water- filled impoundment. The overall goal of the testing program
was to determine how well the technique worked under conditions simulating a
field environment. To accomplish this goal, the instrumentation required to
apply the electrical resistivity technique was assembled. A 100—mu geo—
membrane liner made of high—density polyethylene was Installed in a 1—acre
impoundment. Controllable leaks were tn talled In the liner for detection end
location studies. Experiments were then performed to evaluate the technique
and the instrumentation.
Impoundment Design
A 1—acre Impoundment was utilized to test the liner assessment technique.
The facility was designed to accommodate up to 6.5 feet of water. Figure 19
shows a plan view of the faculty. Overall dimensloqs are approximately 216
feet x 216 feet from the top of the bering. Side slopes arc ’ approximately
24
-------
Figure 15. Equipotential plot of voltages on the surface of the water with a single leak.
ligure 16 Equipotential plot of voltages on the i. race of the soil with a single leak
f
S —.
25
-------
• --.
___ J • _4_ •
,• ,— .. . —.
• ,•‘
• .. .•. .. .- . .
... •i,2 _.- , . ‘• ‘
..t. ‘ . .• .
• .*
Figure 17. Equipotentjal plot of voltages on the surface of (he soil with a single leak.
rCURRENT INJECTION
I
• • ç’
• —LEAJ( #1
I C
Figure 18. Equipoten ial plot of voltages on the surface of the soil with two leaks 400 apart.
26
-------
Figure 19. Plan view of the test impoundment.
Figure 20. Cross section A-A’ of the test impoundm t.
ROADWAY-
SURFACE
27
-------
three to one. A cross Section is shown in Figure 20. An access road exists
around the facility to allow for vehicular traffic during_testing. Construc—
tion of the- facility is shown in Figure 21.
A IOO—aiii high density polyethylene liner was installed in the facility
to serve as the test liner. Installation of the liner is shown ir, FIgure 22.
The liner was anchored at the top of the berm In a two—foot deep trench. The
trench was backfij led with soil.
It was necessary to construct leaks in the liner to facilitate testing
of the leak detection technique. Five pipes, each one foot in diamete , were
installed in the floor of the liner. Figure 23 shows the installation of one
of these pipes. The pipe was made of HDPE to allow the liner to be welded to
the pipe, creating a water—tight seal. A cap was ennstructed to fit over this
pipe. The cap is made from two rings which are bolted together. To create
various leak sizes, an HDPE insert was placed between the rings and then the
rings were bolted together nd placed over the top of the pipe. Figure 24
shows the rings and a plast:. insert. Five of the pipes werc installed in the
liner at the locations shown in Figure 19. These leak points allowed control
of leak location during the performance tests.
Instrumentation and Hardware
The . trumentatjon and hardware assembled to test the electrical resis-
tivity technique at field cO nd it i o ng included various electronj instrumen-
tation components, an electrically operated logging winch with 500 feet of
cable, a support platform, and four electrodes of different deslcne.
Electronic Ir.strumcntation.....
InstrumentatIon requiremen included a stable, low frequency, alternat-
ing current source: precisfon voltage sensing amplifiers; an analog_to_digjtgj
converter; and a data logging System capabis of Previewing the field measure-
ments to an accuracy of 72 da with a dynamic range of 120 dB. These rvqutre—
ments were based upon the results of the physicel. scale model testing and
experience with other field rest tIvtty measure ent instrumentation. The mea-
surement or receiver side of the Instrumentation was designed to obtain data
from the surface of the water, oince initial tests were planned for a water—
filled lined facility.
A Simplified block diagram of the electronic components used during the
teats is shown in Figure 25. The transmitter, consisting of a current source
connected to two electrodes, provides the necessary current needed to generate
potentials on the curface of the impoundment. The receiver measures the But—
face voltage potentials which are then analyzed to detect and locate leaks in
the liner. The current source electrode Ig located in the wat er. The remote
current return electrode is placed in the earth away from the impoundment.
Major components of the receiver include the System Con rol Unit and the
Multiprogranraer. These two components provide for automatic data acquisition.
- 28
-------
“1
- _
•
p ______
• ____
Figure 2L Test impoundment construction activity.
29
-------
I. SI ‘4 4
3 50
Figure 22. Uner inst*II&tj activities.
30
-------
Figure 23. Inst&llation of an HDPE pipe designed as one of five leaks through the liner.
31
-------
Figure 24. Rings used to hold plastic inserts designed as leaks.
32
-------
IRA NSMITrER
RECEIVER
TO
CURRENT RETURN
ELECTRODE IN EARTH
RECORDED DATA
TO PROCESSING
‘ -I
TO CURRENT
SOURCE ELECTRODE
IN WATER
DATA FROM FLOATING DATA FROM FIXED
MEASUREMENT REFERENCE ELECTRODE
ELECTRODE IN WATER
IN WATER
Figure 25. Simplified block diagram of the equipment used to test the electrical resistivity technique.
-------
With the HP 9825 computer, the data is automatically stored on tape for of f—
site processing on an HP 9826 computer and associated printer—plotter.
As indicated in Figure 25, measured surface potential voltages are sent
in analog form via cable through the winch to the system control unit. The
control unit 8ets the signal gain, filters the signal to remove ambient noise,
and generates a DC voltage proportional to the magnitude of the AC signal
measured by the potential electrode. The DC signal is coupled to a 12—bit
analog—to—digital high speed converter in the Mul.tiprogrammer. The digital
output is then sent to the HP 9825 computer for display and storage on mag-
netic tape. The recorded data is processed off—site using an HP 9826 computer.
The instrumentation is operated from a mobile field laboratory. The
laboratory is equipped with an onboard generator powered by a gasoline engine.
This generator supplies 60 Hz AC power required to operate the instrumenta-
tion. The mobile laboratory is shown in Figure 26. Th’ i atrumentatjon in-
side the trailer is shown in Figure 27. The HP oscillator and ‘ rown amplifier
are on the left. The system control unit and HP Multiprograinmer are shown t
the right., The oscilloscope is used to monitor output voltages. The HP 9825
computer, to the far right, records field data oi’ magnetic tape for off—site
proceasing.
Electric Logging Winch——
An electric logging winch is used to control the Jocatton of the poten-
tial measuring electrode on the-surface of the water. The winch is equipped
with variable speed control, a photoelectric odometer, and an electric slip—
ring. The winch drum, powered by the mobile laboratory AC generator, i
equipped with 500 feet of data logging cable. The cable is reeled in and Out
to make surface potential measurenenta. The potential electrode is attached
to the end of this cable. The odometer reading is monitored by the computer
and it8 value, which indicates the position of the measurement electrode, is
used to determine when to trigger a new potential reading. The odometer
allows calcula . jon of the location of leaks and therefore is a critical mea-
surement system component. The winch is shown in Figure 28.
System Elect rodes——
The liner testing system requires the use of a current source electrode,
a potential reference electrode, a remote current return electrode, and a
floating potential measurement elect odc. The current inlection electrode Is
a 3—foot diameter brass disc 1/16 inch thick. It is mounted on a platform
which extends out into the pond. The electrode is positioned 3 to 4 inches
below the water surface, directly above the bottom of the pond side slope.
Brass was selected for the electrode due to its excellent conductive proper-
ties and resistance to corrosion. The large dIameter of the disk increases
the surface area and reduces the voltage drop between the electrode and the
water. The circular shape is compatible with the polar coordinate data acqui-
sition system and helps reduce voltagc anonaltes in the rc- ,ion of the elec-
trode. Electrical current is applied to the electrode via a wire from the
output of the current source In the trailer. A phot-% of the current source
electrode is shown in Figure 29.
34
-------
Figure 26. A mobile laboratory trailer, shown at the upper right, is used to house the instrumentation
during testing.
35
-------
J
I.
Figure 27. Testing ln trumentatlon Inside the mobile laborstory.
It,
-------
Figure 23. The electrically operated logging winch used during testing to control the location of the
me iuremen$ electrode.
37
-------
Figure 29. The circular brass disc below the end of the platform is used at the current source electrode
during testing.
38
-------
The remote current return electr-de consisr of three copper clad steel
stakes driven into the earth to a depth of inches and wired together using
a common conductor. The three stakes. .vnich are approximately one foot apart,
were used to increase surface are’ and reduce the voltage drop between the
electrodes and the earth. The electrodes are located 600 feet to the north
of the test impoundment.
Surface potential meas . .rementg are made u8ing a stationary reference
electrode and a mobile potential electrode which is floated on the surface
of the water. The reference electrode is a copper bar approximately 1.5 feet
long, 2 in. wide and 0.25 in. thick. This electrode is positioned below the
water surface on the side slope of te pond adjacent to the platform. During
operation of the testing system, this electrode remains stationary, serving as
a reference point. All surface potential voltage measurements are made with
respect to this electrode. A preamplifier with unity gain is wired in—line
between this electrode and the system control unit. The preamplifier serves
as an Impedance matching circuit and electrical buffer to the control unit.
The reference electrode and preamplifier are shown to the left in Figure 30.
The potential asurement electrode consists of a plastic box approxi-
mately one—foot square with a preamplifier mounted Inside. A 5—inch welding
rod 1/16 inch in diameter Is inserted into the bottom of the box, directly
into a connector in the preamplifier. Figure 30 sI’ows this potential measure-
ment electrode box. The electrode is visible, pro rudtng up from the center
of the box. The preamplifier contained in the box is identical in design and
function to the one used for the reference electrode.
Potential Measurement Procedures——
The operation of instrumentation and hardware is shown in the conceptual
drawing in Figure 31. The step—by—step procedure for measuring surface poten-
tials on the water is as follows:
(1) Position all electrodes and check cables and connections for proper
Installation.
(2) Clean electrodes in the water to eliminate possible deposit
buildup.
(3) BrIng instrumentatIon up to stable operating conditions.
(4) Set the current source to the pproprtate current output (normally
0.05 ampere).
(5) PosItion the potentini measurement electrode at a selected refer-
ence odometer reading (usually 6.0 feet from the current
electrode).
(6) Load operating program and data tape into the HP 9825 computer.
39
-------
‘It;’
- : :t±%j• ; - - . • • • • —
• : :jf: *.
L ’
Reference
Electrode
and Amplifier
Float ing
Potentja1
Electrode
Figure 30. Tb. reference electrode and preamplifier housing are shown to the left of the potential
mesyjrcment cl ctrød box.
I
40
-------
MOBILE LABORATORY
ONE OF S LEAKS
TEcHNIcIAN MAINTAINS TENSION
ON SURVEY CABLE
N
N
CURRENT RETURN
ELECTROOE 600 AWAY
M
WATER
SURFACE
;- ‘If
LINER
Figure 31. Concepiual drawing of the test impoundm ent and operation of the test equlpmcm.
-------
(7) Manually pull the floating potential electrode away from the
current electrode, along a preselected measurement line, to the
desired location where measurements arc tc begin.
(8) Set the cosputer to receive and record data.
(9) Turn on the winch motor and reel the potential electrode back to
the initial starting point.
(10) Manually ma 1 ntain tension on the cable during winch operation to
prevent ca e fouling.
(11) When the potential electrode returns to its original position,
reset the computer.
(12) Manually move the rope attached to the potential electrode along
the berm of th pond to the next neasureim nt line.
(13) epeat steps (5) through ( I I) for each eurvey line.
The HP 9825 computer controls the recording of measured data. During
step (6) listed above, the computer will prompt the operator for the collowing
input data: current Injection position, water depth, leak location, logging
Interval, current water temperature, wind velocity, and relevant remarks.
After appropriate responses are entered, the computer will then instruct the
potential et ctrode to be positioned. After this is performed, the computer
will record data. The winch is suitc hed on and the electrode is pulled out
and into position. The computer is reset to log dat , and the winch is
started to reel the electrode back toward the current electrode. The entire
procedure is repeated for each survey line.
42
-------
SECTION 5
RESULTS .aD DISCUSSION
OVERVIEW
All measurements at the one—acre test impoundment were taken along radial
lines 5° or 100 apart, beginning at the current source electrode out to mark-
ings on the berm. The 00 through 180° radIal linea’i were located by marking
the liner as wag shown In Figure 31. To mark the berm, a transit was posi-
tioned over the current source electrode. The angle marks were painted on the
liner at S° increments, startIng at 0° to the left around 180° to the right.
The logging cable was positioned on these marks during data acquisition rune.
The location of any anomaly, such aq a leak, could be calculated from the
radial and the odometer data. To estah1i h a baseline condition, the entire
surface of the pond was surveyed with all lsnkg cIoo d. Then, vartous leaks
were created through the liner and measurements taken. A contour plotting
computer program va used to produce equ1potenti i contour plots of the no—
leak and leak data. The results of the testing are presented in the following
sections.
NO—LEAX SURVEY RESULTS
The computer generated plot shown in Figure 32 presents the no—ler k base-
line surface voltage potent a1g measured along the radial. The solid curve
shows the measured no—leak condicton surface pctential voltngps at the corre-
sponding odometer readings. The dottcd curve is a least square curve fit to
the measured data. The minimum voltage shown on the plot near an odometer
reading of 17 feet occurred when the measurement electrode passed over the
refere.ice electrode. Subsequent to this measurement run, the reference el c—
trode was moved closer te the current electrode to prevent this null from
occurring. No other perturbations along thlq radial were observed, indicating
th..t no anomolies such as a leak were detected. These data are typical of the
no—leak data obtained for other radial lines across the surface of the water.
A contour plotting computer proçrag was used to produce an cquipotent [
contour plot of these data. The data presentp In Figure 32, along with simi-
lar data for the other radial lines, were proceqced by the Contour plotting
program to produce the equipotent [ lines shown in FEgure 33. The X and Y
axes define the approxj te north and east water level boundaries of the
impoundment. The angles of the rad [ als are Indicated. The current source
electrode is identified by a dot above the X axis at the 92—foot point.
The equtpoten contour lines show insignificant distortions of the
surface potentials across the water surface. Theqe results are similar to the
43
-------
.
a. ,.
i.)e4.
/
al
I
a . 4e
(FO
Figure 32. A plot of the potential measurements taken along the 900 radial for a no-leak case.
Figure 33. An equIpote J CO.ItOUr plot of no-leak mcasurements
I U
44
-------
no—leak contour plots of data taken with the physical scale model. The con-
tour lines tend to be concentric seolcircies close to the current injection
point. Moving-away from this region, the contour lines begin to atr&ighten
out, due to the effect of snail amounts of current flow across the liner.
Part of this current flow is due to the capacitive effects of the liner.
Overall, this plot indicates no sudden, unexpected changes or perturbations
in the current flow on the surface of the water.
ONE—LEAJ( SURVEY RESULTS
To determine if the survey technique could detect and locate a single
leak, a one—foot diameter (0.79 ft 2 ) leak in 5 feet of water was opened.
The surface potentials were measured and the results plotted as before. The
solid line in Figure 34 Is a plot of tne mea8ured data along the 90° radial.
Examination of this plot indicates a peak in the sur ace potential8 occurring
at an odometer reading of approximately 72 feet. Based upon the conceptual
and physical scale model studies previously cor.ducted, this peak should occur
near a leak location. The actual leak location In the pond was along the 90°
radial, 72.3 feet from the current source electrode. Thus, the results of
this experiment demonstrate the success of this technique to detect and locate
a one—foot diameter leak in 5 feet of water.
An examination of the equipotentja contour plot in FLg re 35 shows that
the contour lines close around the location of the leak. This indicates that
the flow of current converges in the area of the leak, aho ,n by the “bull’s—
eye” pattern centered on the leak. The hull’s-eye potential pattern located
on the surface of the water serves to-graphically reveal the location of th
leak. The plots of the data for each radial along with the equipotentja con—
tout plots serve a8 excellent analytical tools to detect and locate a leak.
To determine If the techniq•ie could detect and locate a smaller leak
a one—iL ch dia.,eter leak (0.005 ft 2 ) was installed on the 90° radial
72.3 feet from the current electrode. Surface potential measurements were
taken at one—foot increments along the 60° through 120° radlals. Figure 36
shows the results of the n’. Inurement taken along the 90° radial. A small peak
was observed on the 900 radial directly over the leak. The equ1potentIg con-
tour plot of these data shown in Figure 37 indicates a significant distortion
on the 90° radial in the proximity of the true leak location. From these re-
sults, it appears that, as the defects are diminished due to increased water
depth, .1 radial plot suen as that shown In Figure 36 can be used to locate the
leak with more accut cy than the contour p]ot. These results indicate that
the electrical resistivity technique can be used 10 successfully detect and
locate a 0.005 ft 2 leak in approximately 40,)00 ft 2 of liner surface area to
an accuracy of one foot. The corito’r line shifts along the 110” radial in
Figure 37 do not inaicate an anamaly, since the shiFt is consistently the same
magni tude.
An expertiient was performed to detormine the response of the technique
when a leak was lceated near the edge o the berm. Figure 38 is a plot of the
deta measured along the 115° radial. A one—foot diameter (0.79 ft 2 ) leak was
located approximately 154 feet from the current source elecirode. A peak
pot ’ ntja1 reading occurred at an odometer reading of approximately 156 feet.
45
-------
. ?P
5.fl4
S. 3?t
).Ifl
2.’?’
1.
/
1.42 ’
‘ I
I e .s a’. AS i t b
ODOMETER IFtI
i ae
i a
I FOOT DIAMETER
LEAK LOCATION
Figure 34. A plot of the potenial measurements taken along the 90° radial. The peak occurs over a I-foot
diameter leak in approximately 5 feet of water.
Figure 35. An equipotential contour plot showing thc distortions from a I-foot diameter leak on the 90°
radial in approximately 5 feet of water.
46
-------
‘.4
e.
I. 14
a 4
t.I1
l.saf
1.14
/
/
/
S .• I. IS ,.• ••
000Mmq u
Figure 36. A plot of the potential mea3ureme taken along the 90 radial. The peak occurs over a I Inch
diameter leak in approatmately 5 feet of water.
Figure 37. An equipotenftal conlour plot showing the distortions from a 1-inch diameter leak
in approxiniat ly 5 feet of water.
I C
71
Ill
0
CURRENT SOURCE ELECTRODE
47
-------
000&IETEA Wi)
Figure 38. A plot of the potential measurements taken along the lii ’ radial. The peak occurs over a I -foot
diameter leak in approximately 5 feet of water.
1 FOOT DIAMETER LEAK LOCATION
$0 20 30 40 50 E0 0 80
CURRENT SOURCE ELECTRODE
I 30
I 35
90 iou no 120 $30 $40 i so iou ii i iou
FEET
Figure 39. An equipotential contour plot showing the distortions rom a I-foot diameter leak
in approximately 5 feet of water.
‘ O-
IL
I..
0• IS •s •
110 115
I S
“I
I SI
I SI
141
131
121
II I
50
40
30
30
I0
a
S
48
-------
The equipotential contour plot of the data is shown in Figure 39. D stortio
of the contour lines occur in the area of the actual leak location.- There is
an unexplained shift in the contour lines between the 115° and 1200 radials.
This may have been caused by a change in the current source output. However,
the data suggest that leaks can be detected at this distance from the current
source electrode and in close proximity to the edge of the liner. It s
predicted that the berm of the impoundment would have an effect on the mea-
surement results due to the resistive nature of the liner to current flow.
This experiment indicates that the magnitude of this effect is small enough
not to prevent detection of a leak this size. Subsequent tests with multiple
leaks located near the berm substantiate this conclusion. These tests are
described later.
SIMULATED LEAK SURVEY RESULTS
Tests were performed to determine if simulated leaks would produce re—
suits similar to actual leaks. It was desirable to use simulited leaks to add
flexibility to the experimental testing of the technique.
The leak signature of a one—foot diameter (0.79 ft 2 ) leak located on the
90° radial 72.3 feet from the current source electrode in 5 feet of water is
shown in Figure 40. The surface potential measure , . ent9 peak at an odometer
reading of 73 feet. The equipotentta contour plot of these measurements,
presented in Figure 41, clearly shows the presence and location of the leak.
A simulated leak of the identical surface area (0.79 ft 2 ) vasfabrjcated
from a brass disc. The disc was placed at the same location end grounded to
the earth outside the liner. A plot of the measured potentials along the 90°
radial is shown in Figure 42. This plot is identical in shape to the actual
leak plot previously shown in Figure 40. The equipotential contour plot in
Figure 43 clearly shows the location of the simulated leak. Based upon these
results, it was concluded that simulated leaks could be used in place of
actual leaks for testing of the electrIc i resistivity technique. The use of
simulated leaks provided greater t lexibility during subsequent testing of the
technique.
Tests were run using simulated leaks to determine the influence of depth
on the leak signatures. An examination of the potential data for the 90°
radial shown in Figure 44 Indicates a peak voltage of approx 4 ntately 11.0
V01t8. Figure 45 shows the equipotential contour lines for a one ft 2 leak in
5 feet of water. The contour line intervals indicate a relatively high flow
of current at the surface near the leak. Figure 46 presents the raw data for
the 90° radial line for the same leak configuration. However, the water depth
is 6.5 feet. Note that the voltage at the peak is 3.9 vo rs. The equipoten—
tial contour plot is shown in Figure 47. The contour line intervals indicate
a lower current flow near the leak. T1’. se two tests were conducted at the
same currant source amperage and position of the current source electrode.
These results suggest increased water depth reduces the surface potential
voltages near a leak. However, in both cases, the leaks are clearly distin-
guishable from both the raw data plots and the coatcur plots. Leak signatures
are not available for depths greater than 6.5 feet. Therefore, the exact
influence of depth on signatures is unknown. It Is indicated, however, that
49
-------
I
9.7
S..
0 .I
7.34
S
4.J
‘.4
‘•‘I
• I .31 - - . .
• 10 40 40 • IS O 04
- 0 00METIRW,J
Figure 40. A plot of the potential measuremenU taken along the 900 radial. The peak occurs over a I-foot
diameter leak in approximately 5 feet of water.
in
57 5
sea
$5 9
•40
IS
I,
sue
‘03
e s
I ’
S.
‘I
69
So.
40
30
79
so
0
o
CURRE’dT SOURCE ELECTRODE
— I— - ; I I I I I I I I__ I
Ia 29 39 49 50 60 79 90 90 193 lIe 129 130 140 I SO 199 170 uoo
FEET
Figure 41. An equ!potential contour plot showing the distortions from a I-loot diameter leak
in approximately 5 feet of water.
73
50
-------
ODOMETER (Ft
Figure 42. A plot of the potential measuremen taken along the 90° radial. The peak occurs over a simulated
I-foot diameter leak in approximately S feet of water.
‘OF
“I
Is’
I SI
141
I U
I 21
III
00
00
so
so
40
30
20
I I
0 Ia 20 30 40 50 00 0 00 30 100 110 120 130 140 15$ 160 I O JOB
FEET
Figure 43. An equipolential Contour plot showing the distortions from a -foot dIameter simulai 1 leak
in approximately 5 feet of water.
9.01
S
0.2
if
oef
5.1+
5 1
2.0 .
105
S CURRENT SOURCE ElECTRODE
51
-------
Figure 44. A plot of the potential measurements taken along the 90’ radial. The peak occurs over a I -foot
square simulai d leak in approximately 5 feet of water.
1 FOOV SQUARE
SIMULATED LEAK LOCATION
lb
98 le O 1.2 128 130 140 15R I SO 170 188
FEET
Figure 45. An equipotential contour plot showing the distortions from a I-foot square simulated leak
in approximately 5 feet of water.
I’
9.2
B 4
6.1
‘.7
4.1
ODOMEVER IFt)
I ”
“a
! I
I ”
13i
121
I II
Be
“a
U.
B ’
7.
I .
5 .
48
30
I I
0—
a
a ø 48 50 60 70 00
a CURRENT 531 RCE ELECThuOE
52
-------
ill
I SI
‘‘ I
I SI
I 5$
141
I a ’
I 3$
II I
—S.
so
7.
S.
50
40
30
20
0
Figure 46. A plot of the potential measurements taken along the 90° radial. The peak occurs over a I-foot
square simulated ieak in 6.5 feet of water.
120
I I .1 I I I I
0 90 00 110 $20 $30 140 I S O IS O 270 100 $90
FEFT
Figure 47. An cquupotcntial contour plot showing the distortions from a I-foot square simulated leak in 6.5
feet of water.
:-‘
I.
1.1
ODOMETER IF4I
I FOOT SQUARE
SIMULATED LEAK LOCATION
CURRENT SOURCE ELECTRODE
$0 20 30 40 50 00 70 em
53
-------
the effects are a linear function of depth. That is, as depth increases, a
one—Coot leak signature -will result In-equipotential plot s—similar to a one—
inch leak signature at shallower depths. A discussion of this subject appears
in the previous section.
An experiment was conducted to observe the influence of highly conductive
metal objects close to a leak. Tests made with the physical scale model indi-
cated no measurable influence of an aluminum block on leak signatures. To
assess this effect, a 0.1—ft 2 simulated leak was positioned on the 90° radial,
60 feet from the current electrode. Measurements were made on the 60° to 120°
radials. The raw data plot t.or the 90° radial is shown in Figure 48. These
data established a baseline measurement without any metal objects. A peak
occurs directly over the leak at the 60—foot odometer reading. Figure 49
eh3ws the closing contours around the leak.
To assess the effects of metal objects, six 2—gallon metal cans were
placed around the 0. 1—ft 2 simulated leak in a circular pattern. Each cs.i was
positioned approximately 9 inches from the leak. Measurements were retaken
under these conditions. A plot of the raw data along the 900 radial is 8hown
in Figut e 50. The shape of this curve 1.8 nearly identical to the shape of
the 90° radial curve in Figure !.8 for the test without cans. The only notice-
able difference is that the magnitudes of the surface potential voltages are
slightly lower with the cane. The contour plot of the can data presented in
Figure 51 is nearly identical to the plot without the cans. It can be con-
cluded from these results that the cans had a slight influence on the ma ni—
tude of the measurements, but did not change the signature of the 0.1—ft 1 leak
to any great extent.
SEAM LEAK SIMULATION RI SULTS
Ceomembrane liners are typIcally joined at seams using various methods.
An experiment was set up to determiqe the leak signature of a simulated seam
leak 30 feet in length. An uninsulated copper wire was connected between
two permanent leak locations in t.e pond. The ends of the wire were pushed
through each leak cover and grounded using 12—Inch steel stakes. The only
metal surface exposed to the water was the wira. The surface area of the wire
was calculated to equal 0.4 ft 2 . The wire was located along the 900 radial,
starting at 110 feet and ending at 140 feet away from the current electrode.
The raw data plot for the 90° radIal is shown in Figure 52. The shape of
the curve is significantly diffetent from the point leak curves. It is much
broader and the voltage gradient in the simulated leak zone is much smalle .
The equipotential contour plot in Figure 53 shows these differences in the
shape of the contour lines. The wire influer’ces tr surface potent 4 alq on all
radial lines. The contours converge in the immediate location of ie wire.
As expected. the leak sigr’ature of an elongated seam failure is spread out
along the radial and influencec the current dlstrtbutionp in a much more
significant way than do point leaks of the same surface area. This test dem-
onstrates that the leak detection technioue is senaitive to different leak
geometries. These results indicate that the technique h59 the potential to
distinguish between point leaks and seam leaks. The thredhold limits where a
54
-------
Figu, 48. A plot of the potental measurements taken along the 900 rudial. The peak occurs over a O.I-foot
square simulated leak without metal ns.
I,.’ .
I..’-
imf
us
I,
‘4
I i
i i
ii I i.
sb..
S.
7.
so
SI
4S
3.
20 • CURRENT SOURCE E ECTROoE
i0
• I I I I 4—....— j I I I I I i
0 is is so Se so s us uo isa uje is no 140 use isa u;o ioe use
FEET
Figure 49. An equipotential contour plot sh’,wung the du Ioriions prom a 0.1 -foot square simufajed leak
ithoui me’al canc.
ODOMETER v
S.
‘a
is.
lie
-------
‘ 5
I.
I 71
I SI
‘SI
‘4’
1)1
“I
‘I’
Ii.,,
S.
70
I.
5.
3.
21
is
U
S
‘•1
4.
.1
I L
SI
S.’
I..
‘.4
.. I. 41 I
OOOV(Ti I I
Figure 50. A riot of the potential measurements taken along the 9(r radal. The peak occurs over a 0.1 •fooi
square simulated leak with metal cans surrounding the leak.
‘So
0 t FOOT SQUARE SIMULATED
LEAE LOCATW?Pd WiTH METAL CANS
SURROUHDINZ THE LEAX
CURRENT SOURCE LCTRODE
I I I I I I I I I I I I
II’ ZO 30 40 35 15 75 15 30 I I ’ S III’ 170 13C 140 i S O 10 l?S ISO IS O
FEET
Figure 31. An equlporenilal conlour plot showing the distortions from a 0.1-loot square stm lated leak with
metal cans surrounding the leak.
lie
56
-------
I
1
I .
/
/
f
I.
ii a.
I..
I.-
‘.4 ’
- . a — - ____
• • 44 1 •
000METN i fl
Figure 52. A plot of the potenti’ l measurements taken along the 90° radial. The peak occurs over a 0.4-foot
square, JO1oo i long wire u cd to simulate a seam leak.
I . t I I I t 4
O ii 20 30 40 SR 60 70 00 90 lea 110 120 i o 140 iso i&o o
FEET
Figure 53. An equipotepipal contour plot showing the distortions from a 0 4-foot square, 30-foot long wire
used to simulate a ream leak.
“I•
171
I II
1 5 1
141
121
I I I
I
Is
SO
Is
S.
40
10
20
Is
0
1 59
CURRENT SOu cr ELECTRODE
o 4-FOOT OQUARE
SIMULATED LEAk LOCATION
57
-------
seam leak can be detected will, ae demonstrated with the circular leaks,
depend on surface area of thd leak and depth of water over the liner.
PIULTIP’,E LEAK SURVEY RESULTS
Multiple leak configuration expertLent6 were performed to determine the
leak signatures and sensitivity of the electrical resistivity technique. For
the first experiment, one—inch diameter leaks were installed at leak locations
01 and 04. Leak 01 is 72.3 feet away from the current electrode along the 90°
radial. Lkak 04 is 118 feet away from the current electrode along the 110.
radial. Measurements of the surface potentials were mede along the 60 to
140° radjals from the water line back to 6 feet from the current electrode.
Figure 54 nhowa the raw data plots for the 90’ and 1100 radialo. As with
the single leaks, peaks of the measured data occur over the leaks. These
peaks are, however, less pronounced than the corresponding peaks for the
single leak configuration. Nevertheless, they are sufficient in ‘ gnttude to
distort the equipotentisi contour lines in the area of each leak, as shown in
Figur 55. The equIpotent q contour lines clearly show the presence of the
two leaks, The plots in Figure 4 can be used to establish their locations to
within o ie foot.
Another experiment wa performed in which three one—foot diameter leaks
were detected and located. These leaks wpre established at locations 01, 04,
and 05. Surface potential measurements were at one••foot increments along the
700 through 1 5° radial lines. t igure 56 shows plots of the measured data
along the 90°, 110° anc 115° radials. Voltage peaks occur directly over each
of the three leaks. The equipotential contour plot shown in Figure 57 clearly
shows the presence of each leak.
A second experiment was conducted to detormine the leak signatures of
th:ee different sized leaks located along the 90° radial. One—foot, 5—inch.
and one—inch diameter leaks were established at leak locations 0!, 02, and 03,
respectively. Measurements were taken at one—foot increeents along radial
‘inca 75° through 105°. A plot of the measured data along the 90° radial in
shown n Figure 58. The plot shows two large potential peaks at 70 ft and 110
ft and a small peak at 140 ft. The first two peaks correspond to the one—foot
and 5—inch diameter leaks. The influence of the third one—inch dtacwter leak
is very slight in terms of showing a perturbation in the curve. The roughness
of these plots Indicates noise in the data. Under these conditions, it would
not be reasonable to predict that a leak exisled at the 03 location based on
the 900 plot of measured data in Figure 58. It appears that the effect of the
first two leaks was to “mask or dininisl ” the effect of the third leak on
the surface measurements.
The equipotentlal contour plot shown in Figure 59 clearly indicates the
presence of th’ first tw3 leaks. The shape of the contour line around the
small one—tnc’ diameter leak suggests that the leak did have an effect on the
measurements which was not easily identified from examination of the raw data
plots. 1 the presence of the leak at 03 had not been known, this plot would
have been sufficient ustificatjon to resurvey the pond using a different
electrode position. By repo8itioning the current inlection electrode at a
58
-------
4.
L 4
>3t$ /
a.
- a. ,
/Z
- -4 -*- I - -
S • S • • S IS
ODOMETER (FtJ
Figure 54. Plots of measuremenic along the 90° and 110° radials. Peaks occur over I-inch diameter leaks.
I4NCH DIAMETER
LEAK LOCATION
I INCH DIAMETER
140 LEAK LOCATION
Figure 55. An equipotential contour plot showing the distortions (rom two I-inch diameter leaks on two
different radials.
. 1 if
4. 4
4.I6
5 ‘4
ODOMETER IFI)
FEET
59
-------
Figure 56. P1015 of measurements along the 900, I 100 and 1150 ridials. Peaks occur over I .Ioot diameter
leaks.
01
171
Is’
i5
141
131
I O U
I I’
I0
00
M i
I L
•0
70
00
50
40
30
20
10
0
0 lB 20 30 40 50 00 ?0 00 90 lOB 110 120 130 40 150 l E O 179 lOB
F EEl
Figure 57. An cquipotentual contour piot sho ; ng the ‘ tstortions from three I -loot diameter leaks on three
different radials.
f
000MmR wo
PS OS
1 FOOT DIAMETER
LEAK LOCATION
1 FOOT DIAMETER
LEAK LOCATiON
• CURRENT SOURCE ELECTRODE
60
-------
u.si
U
Ie,
l e o’
11.14
I
Q 4 4 . ,j .1
Figure 58. A plot of measurements along the 900 radial. Peaks occur over three different sized leaks.
I INCH DIAMETli1 LLAK LOCATiOw
5-INC.4 DIAWETEq LLAX LOCATION
I SOOT DIAMET(R LE.’K LOCATION
I SI
I 21
III
PSi
hi
III
iii
I C
‘ S
so
so
40
so
• CURRENT SOURCE ELECTRODE
a I I I I I h _ - _ I I f—t— I I I
0 10 20 30 40 50 90 0 00 90 I S O 110 120 30 140 I S O 10 170 IS O
EET
Figure 59. An equipotential contour plot showing the distortions from three different sized leaks along the
90° radial.
61
-------
right angle to the 900 radial and moving closer to the leak location, the
presence of the leak could possibly have been confirmed,
SU)*(ARY OF LARGE—SCALE TESTING
The results of the large—scale testing indicate single and esiltiple leaks
in a geomembrane liner as small as one inch in diameter can be detected and
located vith an accuracy of one foot or less. This wag achieved for leaks
located in excess of 150 feet away from the current source electrode. The
location of the leak does not appear to influence its detection or location to
any significant extent. Increasing depth reduces the magnitude of the effects
that a leak has on surface measurements. However, leak detection and location
were not adversely affected. Depth of fluid over a liner System will
influence leak detection and location. Measurements obtajncd for a one—inch
diameter leak in 5 feet of water are predicted to be similar to measurements
obtained for a 5—inch diameter leak in 25 feet of water. Future testing of
the technique as well as improved computer modeling will further verify the
validity of this prediction.
Metal cans positioned around lea. influenced the magnitudes of the mea—
surelnents taken on the water surfac over the leak. However, they did not
prevent detectIon or location of the leak. Therefore, metal objects in lined
surface impoundments are not expected to prevent detection or location of
leaks.
Results of simtelated sean leak tests suggest that the geometry of a leak
can be determined from analyses of equipotential plots. The seam leak, simu— -
lated with a wire, elongates equipotential contour lines. The rypicaj bull’s—
eye pattern produced by single point leaks is stretched along the l ngrh f
the wire. Knowledge of this characteristic will be helpful when interpreting
the results of surveys at other lined facilities.
62
-------
FE RE NCE S
Hanrahan, 0. 1979. Hazardous Wastes: Current Problems and Near—Term
Solutions. Technol. Rev. 82(2):20—31.
EPA. 1980. Research Outlook 1980. EPA 600/9—80—006. U.S. Environmental
Protection Agency, Washington, DC. 224 pp.
Surface tmpundment A.ssement National Report. U.S. Environmental Protection
Agency, Office of Water, Office of Drinking Water. December, 1983.
154 pp.
63
------- |