600284007
SPILL ALERT DEVICE
FOR EARTH DAM FAILURE WARNING
Robert M. Koerner and
Arthur E. Lord, Or.
Drexel University
Philadelphia, Pennsylvania 19104
at an I Nu. R-8G251
Project Officer
John E. Brugger
Oil & Hazardous Materials Spills Branch
Municipal Environmental Research Laboratory
Edison, New Jersey 08837
MUNICIPAL ENVIRONVENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental
Research Laboratory, U. S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U. S. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
FOREWORD
The U.S. Environmental Protection Agency was created because of
increasing public and government concern about the dangers of pollution
to the health and welfare of the American people. Noxious air, foul
water, and spoiled land are tragic testimonies to the deterioration of
our natural environment. The complexity of that environment and the
interplay of its components require a concentrated and integrated attack
on the problem.
Research and development is that necessary first step in problem
solution; it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research
Laboratory develops new and improved technology and systems to prevent,
treat, and manage wastewater and solid and hazardous waste pollutant
discharges from municipal and community sources, to preserve and treat
public drinking water supplies, and to minimize the adverse economic,
social, health, and aesthetic effects of pollution. This publication is
one of the products of that research and provides a most vital communica-
tions link between the researcher and the user community.
The subject of this report is the development of a spill-alert
device for earth dam safety warning systems to identify dams that are in
danger of failing and emptying their contents into downstream waters.
For those dams that are in need of repair, the device can also act as a
construction design aid to identify the adequacy of these repairs. This
report will be valuable to governmental (Federal, state, and local),
industrial, and private owners and operators of earth dams, dikes,
embankments, lagoons, and impoundments that contain liquid materials and
semi-solid sludge. The impounded material can be of any type, but the
work described in this report focuses on liquid hazardous materials. The
report is also of interest to researchers investigating the fundamental
aspects of soil strength in relation to its "noise" generation during
stressing. These noises, or more appropriately, acoustic emissions, are
at the heart of this detection system. Further information on the
subject may be obtained by contacting the Oil and Hazardous Materials
Spills Branch of the Municipal Environmental Research Laboratory
(Cincinnati) at Edison, N. J. 08837.
Francis T. Mayo, Director
Municipal Environmental
Research Laboratory
iii
-------
ABSTRACT
A spill alert device for determining earth dam safety based on the
monitoring of the acoustic emissions generated in a deforming soil mass
was developed and field tested. The acoustic emissions are related to
the basic mechanisms from which soils derive their strength. Laboratory
feasibility tests, conducted under widely varying conditions, have re-
sulted in an instrument package consisting of a wave guide (a steel rod
projecting into the earth mass), a transducer (to convert the mechanical
waves transmitted from the deforming soil into an electrical signal), an
amplifier (to increase the signal level), and a counter (to quantify the
signal). The resulting monitoring system has been field tested at 19
field sites and found to portray accurately the stability of the
particular site in question. Additional detail has been added that
enables the following categorization of the relative stability of the
soil mass being monitored:
No emissions: soil mass is at equilibrium and safe.
Low emissions: continue to monitor soil mass.
High emissions: soil mass requires remedial work.
Very high emissions: this situation requires evacuation of
downstream residents.
This report was submitted in fulfillment of EPA Grant No. R-802511 by
Drexel University under the sponsorship of the U. S. Environmental
Protection Agency. This report covers the period from July 1, 1973, to
June 30, 1979, and work was completed as of June 30, 1979.
IV
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables xi
Acknowledgements xii
1. Introduction 1
2. Conclusions 2
3. Recommendations 3
4. Background and Project Design 4
5. Acoustic Emissions Fundamentals 7
6. Fundamentals of Acoustic Emissions in Soils 14
7. Applications of Acoustic Emissions in Soils .......... 51
8. Spill Alert Device Details. .................. 74
References 81
Appendices 86
A. Published and/or submitted technical papers
on acoustic emission monitoring 86
B. Spill alert device users manual 89
C. Application of acoustic emission monitoring
in seepage 95
D. Application of acoustic emission monitoring
in pipelines 100
E. Application of acoustic emission monitoring
in concrete .............. 109
F. Glossary 116
-------
FIGURES
Number Page
1 Schematic diagram of acoustic emission monitoring system
showing typical oscilloscope trace of series of emissions 8
2 Frequency distribution for silty sand soil tested in
(a) unconfined compression and (b) triaxial shear at
69 kN/m^ (10 psi) confining pressure 16
3 Frequency versus attenuation response of dry granular
soils using various techniques indicated 18
4 Granular soils tested 21
5 Schematic diagram and photograph of acoustic emission
monitoring setup on stressed soil specimen tested in
triaxial shear 22
6 Isostatic test results (time versus acoustic emission
in units of 10,000 counts) for four granular soils
listed in Table 2 23
7 Triaxial shear test results (deviator stress versus
acoustic emission in units of 100,000 counts) for four
granular soils listed in Table 2 25
8 Triaxial shear test results (deviator stress versus
acoustic emission in units of 100,000 counts) for four
granular soils listed in Table 2 27
9 Average amplitude of acoustic emissions (measured as peak
signal voltage output) for various soils as function of
percentage failure stress in triaxial creep at 34 kN/m2
(5 psi) confining pressure 32
10 Frequency distribution of acoustic emissions from
kaolinite clay (soil No. 6) tested in unconfined
compression at 33% water content 33
11 Attenuation of acoustic emissions in clayey silt (soil
No. 5) at varying water contents at frequency of about
1 kHz 35
VI
-------
Number Page
12 Triaxial creep response of clayey silt (soil No. 5)
at varying confining pressures ................................... 36
13 Triaxial creep response of kaolinite clay (soil No. 6)
at varying confining pressures ................................... 37
14 Stress/acoustic emission response of clayey silt (soil No. 5)
at varying water contents in unconfined compression .............. 39
15 Stress/acoustic emission response of four cohesive soils
in triaxial creep tests showing significance of
plasticity index....... ....... . ...... .......... ....... ........... 4°
16 Unconfined compression test results for undisturbed
sample of silty clay (soil No. 7) at 56% water content ........... 41
17 One-dimensional consolidation response of sandy, silty
clay at constant pressure on log-time scale ...................... 43
18 One-dimensional consolidation response of sandy silty
clay over range of pressures showing strain and acoustic
emission responses ............................................... 44
19 Effect of varying parameters in Equation (10) to observe
behavior in maximum acceleration of emissions.... ................ 49
20 Experimental setup and location of wave guide/accelerometer's
first resonance as a function of length considering different
diameter and geometry of steel rods .............................. 53
21 Elevation and plan views of site 3 near McCook, Nebraska,
showing horizontal wave guide location scheme .................... 57
22 Elevation view of site 5 in Philadelphia, Pa., showing
surcharge load and compressible soil along with
different types of wave guides ................................... 59
23 Time/settlement and time /acoustic emission response
curves from site No. 5.... ...... . ..... . ........ .. ................ 60
24 Acoustic emission count rate versus time of cut for
site No. 13 showing failure after fourth cut ..................... 63
vii
-------
Number Page
25 Schematic diagram of site No. 14 showing approximate
boundaries of five cuts made and photographs after
Cut Nos. 1 and 4 65
26 Acoustic emission response after Cut No. 1 67
27 Acoustic emission response after Cut No. 2 67
28 Acoustic emission response after Cut No. 3 68
29 Acoustic emission response after Cut No. 4 68
30 Acoustic emission response after Cut No. 5 69
31 Summary of acoustic emission rates after each cut 71
32 Settlement and acoustic emission response curves from
site No. 18, showing response at various locations
along slide area • 73
33 Photograph of acoustic emission field system 75
34 Photograph of acoustic emission field system 76
35 Photograph of acoustic emission laboratory system 77
36 Photograph of acoustic emission laboratory system 78
37 Photograph of acoustic emission laboratory system 79
B-l Photographs of spill alert device components 90
B-2 Details of wave guides used in acoustic emission
monitoring 92
6-3 Sample monitoring sheet 94
C-l Flow rates and acoustic emission rates compared for
seepage study at site No. 15 97
C-2 Experimental setup for study of acoustic emission
results from soil void seepage 98
viii
-------
Number Page
C-3 Acoustic emission rates for flow of water through
a col umn of Ottawa sand 99
D-l Acoustic emission count rate versus internal pipe
pressure for air leaking from 15.2-cm (6-in.)
diameter pipe 102
D-2 Acoustic emission counts versus internal pipe
pressure for water leaking from 15.2-cm (6-in.)
d i ameter pi pe 103
D-3 Acoustic emission counts versus internal pipe
pressure for oil leaking from 15.2-cm (6-in.)
di ameter pi pe. „ 104
D-4 Field results of signal amplitude and acoustic emission
count rate for a constant-source leak in a 7.6-cm (3-in.)
diameter pipeline as a function of distance from the leak 106
D-5 Field results of acoustic emission count rate for a
pulsating leak in a 7.6-cm (3-in.) diameter pipeline as
a function of distance from the leak and on both sides
of the leak 107
D-6 Data of Figure D-5 replotted to illustrate the method
of leak source location using the acoustic emission
monitoring technique 108
E-l Load versus acoustic emission response of 3-day-old
concrete specimens showing effect of load cycling on
acoustic emissions. Ill
E-2 Acoustic emission response of concrete cylinders as a
function of age (curing time) at various percentages
of ultimate fracture load 112
E-3 Acoustic emission versus time response for creep tests
(sustained-load tests) at various percentages of
ultimate failure load 113
IX
-------
Number Page
E-4 Acoustic emission rate versus time response for creep
tests (sustained-load tests) at various percentages
of ultimate failure load over long-term monitoring 114
E-5 Load versus acoustic emission response of concrete
beams tested in three-point loading tests (flexure
tests) with transducer mounted either on the compression
face or on the tension face 115
-------
TABLES
Number Page
1 Categorization of Acoustic Emission Level as Obtained
from Spill Alert Device on Numerous Earth Dams 2
2 Effect of Particle Characteristics on Acoustic
Emission in Granular Soil 20
3 Properties of Cohesive Soils Used in this Study 30
4 Influence of Medium Surrounding Wave Guide on
Frequency and Amplitude of First Resonance 54
5 Overview of Sites Being Monitored Using the
Acoustic Emission Method....... 56
6 Acoustic Emission from Neb-200 Dam Site 58
7 Commercially Available Acoustic Emission Equipment 80
XI
-------
ACKNOWLEDGEMENTS
The authors express their sincere appreciation to all those who
contributed so generously to the completion of this project. Special
thanks are due to John E. Brugger, EPA Project Officer, Ira Wilder, Oil &
Hazardous Materials & Spills Branch Chief, and W. Martin McCabe, Research
Assistant, Drexel University, for their excellent cooperation, interest,
and enthusiasm.
The laboratory work was performed mainly by W. Martin McCabe, John
W. Curran, Shirley L. McMaster, and John V. Lima, with technical
assistance by Albert Bangs and Kenneth Whitlock.
The difficult and costly field work aspects of the study would not
have been possible without the cooperation of the following public and
private organizations.
USDA Soil Conservation Service
Bethlehem Steel Corporation
City of Philadelphia, Division of Aviation
Site Engineers, Inc., Cherry Hill, N. J.
Les Mines Madeleine, Ltd., Quebec, Canada
Borough of Boyertown, Boyertown, Pa.
City of Hopewell, Va.
U. S. Coast Guard
Thomas M. Durkin and Sons, Inc., Philadelphia, Pa.
Raymond International, Inc., Soiltech Division
Our thanks are also expressed to Sidney Mathues and Richard Spotts
of the General Electric Co. (Philadelphia), who lent and/or donated
instrumentation to Drexel University.
The preparation of this report was accomplished through the efforts
of John J. McElroy, who drew all figures, and Elizabeth T. Fox, who typed
the manuscript; to them, we express our sincere thanks.
xii
-------
-------
SECTION 1
INTRODUCTION
A research and development program was undertaken to understand the
fundamentals, investigate the feasibility, and refine the development of
a field device—based on the detection and measurement of acoustic
emissions—to monitor the stability of earth dams. This program included
the laboratory investigation of a wide range of soil types (sands, silts,
and clays) under varying conditions (density, moisture, stress state,
etc.) in relation to acoustic emission behavior. After this information
was acquired and analyzed, a field-use system was assembled that met the
combined objectives of portability, ease-of-use, rapid data acquisition
and anaylsis, and low cost. This system is known as a spill alert device
and consists of a steel rod wave guide, transducer, amplifier, and
counter. Included in this report are the essential elements of the
laboratory program (Sections 5 and 6), the field program (Section 7), and
the final unit as currently used (Section 8).
Parallel studies that have spun off from this project (e. g.,
acoustic emission monitoring of seepage, pipelines, and concrete) are
included as appendices.
-------
SECTION 2
CONCLUSIONS
The principle of the spill alert device described in this report is
based on the detection and analysis of the acoustic emissions generated
by a deforming soil mass. These internal sub-surface acoustic emissions
are brought to the ground surface by a steel rod wave guide, converted to
an electrical signal by a transducer, amplified, shaped, and finally
counted on a frequency counter. The acoustic emission counts are
directly related to the stability of the earthen mass being monitored
(Table 1). With the acoustic emission count level known, the stability
assessment is immediately available—a major goal of the study. Other
goals of the study were also realized in that the device is portable (all
components are battery operated), is lightweight (less than 18 kg (40
pounds)), consists of commercially available components that are
reasonably priced (the entire system cost about $2,000 in 1979), is rapid
(each monitoring station requires only 3 to 10 min), and yields
easy-to-interpret data.
TABLE 1. CATEGORIZATION OF ACOUSTIC EMISSION LEVEL AS OBTAINED
FROM SPILL ALERT DEVICE ON NUMEROUS EARTH DAMS*
Acoustic emission
level Soil Relative
(counts/min.) deformation safety Recommendation
Negligible
(0 to 10)
Low
(10 to 100)
High
None
Slight
Large
Good
Marginal
Poor
Visit periodically
Continue to monitor
Remedial measures
(100 to 1,000) required
Very high Very large None Evacuate downstream
(greater than 1.000)
* These results are based strictly on the monitoring done in the project.
-------
SECTION 3
RECOMMENDATIONS
As a result of this 6-year project, the spill alert device is based
on a firm technical background. Its feasibility has been verified by
numerous small-scale laboratory tests and by extensive work in the field.
Nineteen sites have been, or are in the process of being, monitored by
the techniques. Persons and organizations other than EPA or Drexel
University are being encouraged to use the system to assess its utility
and to discover its limitations and/or flaws. To this end, we have been
actively publishing, lecturing, and making the equipment available to
those with valid uses for such a monitoring system. Thus, technology
transfer to the intended user remains as the final, currently indenti-
fiable goal of this project.
Among other candidates for acoustic emission monitoring are
above-grade stockpiles of non-soil, industrial materials (e.g., tailings,
fly ash, phosphate residues (slimes), gypsum). Such mounds can yield or
fail during or following rainstorms. Additionally, certain above-grade
sanitary landfills and poorly engineered dumps (both of which have large
non-soil components) are subject to disintegration for which advance
warning can be obtained by acoustic emission methods.
The detailed relationship between flow patterns and acoustic
emissions of water-rich clays, silts, and thixotropic materials should be
further investigated.
Additional lab and field work should be done on the use of sound-
attenuating jackets or shields bonded to the metal waveguides so that
emissions will be only transmitted from the stratum where the soil or
other material is in contact with the uncovered rod.
To keep potential and actual users of acoutic emission techniques
in dike integrity assessments current with the state-of-the-art, the con-
vening of topical conferences and symposia (preferably with published
proceedings) should be fostered.
Consideration should be given to the development by ASTM or similar
associations of guidelines or, better, standards for the use of acoustic
emission in soil applications.
Acoustic emission techniques are recommended for (and in some
instances, have already been applied to) slope stability of cuts and
fills, subsurface seepage (piping), and related civil engineering
concerns.
-------
SECTION 4
BACKGROUND AND PROJECT DESIGN
The problem of earth dam failures, which includes earthen dikes,
holding ponds, lagoons, embankments, etc., is ageless, but it probably
entered the technical literature in 1889 with the failure of the South Fork
Dam in Pennsylvania, which caused so much death and destruction in Johnstown.
Since that time, so many earth dam failures have occurred that a categoriza-
tion is possible: 30% of failures were structural, 40% were seepage, and
30% were hydraulic (1). The problem is far from being solved. Since 1972,
when the authors began working in this area, four major failures have
occurred, all of which have been widely publicized by the news media. These
failures included Buffalo Creek, West Virginia (February 26, 1972); Grand
Teton, Idaho (June 5, 1976); Johnstown (Laurel Run), Pennsylvania (July 20,
1977), and Taccoa, Georgia (November 4, 1977).
Along with the failures of these major dams have been the failures of
innumerable small dams of both private and public ownership. The latter
category (small dams) has received little attention, since such failures
have usually produced no loss of life and only minor property damage. Their
environmental damage, however, has often been devastating. This is particu-
larly the case with failure of dams containing hazardous materials or
industrial wastes, which cause extensive fish kills and water pollution,
depending on the original water quality of the receiving stream or river.
The category of small earth dams is the one focused on by this research and
development project, but the results are applicable to all types of unstable
earth masses.
Initial feasibility tests on acoustic emission generation in soils
were carried out in 1970 and 1971 and published in a short paper in January
1972. By July 1, 1973, the first part of a 6-year research and development
effort was funded by the U. S. Environmental Protection Agency (EPA). This
report summarizes this effort.
Work carried out under the project has been brought to the attention
of the technical community through publications in various journals and
conference proceedings. In all, 30 papers have been written by the
principal investigators on the subject. The complete reference list is
given in Appendix A of this report.
The most significant achievement is that the effort resulted in the
fabrication of a usable and workable earth-mass monitoring system known as
an earth dam spill alert device. The device has been field tested and
calibrated, and it is available for both government and private use. The
-------
system consists of components that are all commercially available and easy
to assemble, install, and use. The version of a user's manual is given in
Appendix B of this report.
Several industrially engineered and packaged acoustic emission-
based systems for field use in dam and soil stability evaluations are now
available commercially. We are convinced that the monitoring system
resulting from this research and development program will find widespread
use in evaluating the stability of earth masses in the near future.
In 1977, the acoustic emission system was entered in the prestigious
IR-100 competition, in which awards are made annually on the basis of the
100 most significant advances in industrial research and development. The
ceremony is held at Chicago's Museum of Science and Industry and the winning
entries are on display for a month. The contest is sponsored by INDUSTRIAL
RESEARCH (now INDUSTRIAL RESEARCH AND DEVELOPMENT), a widely distributed
publication that reports on those advances in R&D that have special and
practical application to industrial problems.
Koerner and Lord began their joint work (which ultimately led to the
work described here) by studying the relation between acoustics and soils in
1970. The first research covered measuring the strength of the returned
ultrasonic echo from an Al^Os/soil interface as the soil dried. They
expected to detect the shrinkage limit, etc. in this manner. Results were
poor. The next project was an attempt to measure the dynamic Young's
modulus of a soil in the composite resonator (bending mode vibration of
steel strip plus soil layer) apparatus of Bruel and Kjaer. Again, results
were inconclusive.
Fortunately, the next joint venture was quite productive from the
start. Transducers and amplifiers most sensitive in the kilohertz region
were kindly loaned to the authors by Sidney Mathues and Richard Spotts of
the General Electric Company of Philadelphia.
Soil samples were axially loaded to failure, and the receiving trans-
ducer picked up the generated noises throughout the deformation process, up
to and including failure. When the transducer output was fed into an ampli-
fier, shaped, and then counted on a frequency counter, the response revealed
a basic similarity to the typical stress-versus-strain response. A series
of tests at varying water contents on clayey silt soil samples (locally
called Delaware River silt) resulted in logical trends, since the lower the
water content, the greater the emissions and also the greater the strength.
These data and the resulting family of curves were reported in a paper which
was published as a Technical Note (2) in the Geotechnical Engineering
Division Journal of the American Society of Civil Engineers in January, 1972.
Applications of this newly found phenomenon seemed numerous, since
soil masses are known to deform before reaching a failure state. Such
problems are encountered with retaining walls, footing foundations, pile
deformations, underground tunnels, pipelines, etc. and were all reasonable
target areas for application. None seemed so promising as the slope
stability area in general and earth dams in particular. During the proposal
-------
writing stage, we became aware that our "noise" monitoring of soils had much
in common with microseismic monitoring of rock by geologists and geophysi-
cists and the acoustic emission monitoring for flaws of pressure vessels and
metals by metallurgists and aerospace engineers. This parallel body of
literature was reported in a state-of-the-art review by Lord (3), which was
in preparation at about the same time that a sponsoring agency was being
sought. The information that follows in this report was all done under EPA
sponsorship and, for the most part, is taken from various sections of the
papers listed in Appendix A, particularly references 4, 5, 6, and 7.
-------
SECTION 5
ACOUSTIC EMISSION FUNDAMENTALS
Introduction
Acoustic emissions are the internally generated sounds that a material
produces when it is placed under certain stress conditions. Sometimes these
sounds are audible (wood cracking, tin crying, ice expanding, soil and rock
particles abrading against one another, etc.), but more often they are not
heard by humans, either because of their low magnitude or high frequency or
both.
Normally a piezoelectric sensor (an accelerometer or transducer) is
used to detect the acoustic emissions. These sensors, when mechanically
stimulated, produce an electrical signal. (The casual reader may note a
similarity to what is commonly called a microphone). The signal is then
amplified, filtered, shaped, counted, and displayed or recorded. Figure 1
shows a schematic drawing of a typical acoustic emission monitoring system
being used as a stress is applied to a soil sample in unconfined
compression. Also shown is an oscilloscope trace of a typical set of
acoustic emission bursts from a stressd soil sample. The counts or
recordings of the emissions are then related to the basic material
characteristics to determine the relative stability of the specimen being
tested. (Counts refer to electric pulses above a threshold level.) When
no acoustic emissions are present, the material is in equilibrium and thus
stable under that condition. When emissions are observed, however, a
nonequilibrium situation is present that, if continued, can ultimately lead
to specimen failure.
Literature Survey
With respect to Figure 1, it should be noted that two types of soil
strength tests are commonly performed; one called unconfined compression
tests, the other, triaxial shear tests. Some explanation of these tests
may be of value. The soil sample is usually in the form of right circular
cylinder. It is taken from the sampling tube (or made in the laboratory)
and placed on a metal base especially prepared for such tests. (The test
is an ASTM Standard.) The upper base plate is placed on top of the sample
and the assembly is then fitted with a thin rubber membrane and made
leakproof using 0-rings. A plastic cylinder of approximately twice the
diameter of the sample is then installed and the chamber is bolted together
to the upper assembly which has a free moving piston in it. Water is next
introduced into the plastic chamber surrounding the rubber membrane encased
soil sample. When the sample comes to equilibrium, the pressures are ready
to be imposed.
-------
TYPICAL OSCILLOSCOPE TRACE-
OF A SERIES OF ACOUSTIC EMISSIONS
Figure 1. Schematic diagram of acoustic emission monitoring system
showing typical oscilloscope trace of series of emissions.
-------
If no confining pressure is applied (the weight of the water is
almost negligible), the piston imposes vertical load to the sample until it
eventually fails. The vertical load divided by the sample area is the
major principal stress (&]}. Since there is not confining pressure, the
horizontal stress, called the minor principal stress ("
can be obtained, where
c= cohesion (soil strength at zero confinement)
-------
Historically, acoustic emission work began in the mining industry to
detect instability in mine roof, face, and pillar rock to predict when
failure might occur. This work was initiated by Obert (8), and Obert and
Duvall (9) in the United States, and by Hodgson (10,11) in Canada. Their
monitoring of rock emissions, which they called microseisms, began in the
early 1930's and has continued to the present by the Bureau of Mines'
scientists (12) and others (13,14). Although these pioneering workers were
hampered by a lack of sophisticated and reliable equipment, their ideas and
goals were certainly in the right direction and set the tone (no pun
intended!) for many modern projects.
Beginning in the 1950's, acoustic emission research was initiated in
the metals area. Kaiser (15,16) worked with steel, copper, aluminum, lead,
and zinc. He discovered many fundamental properties that relate stress
behavior to acoustic emissions. Tatro and Liptai, (17,18) in the early
1960's, used the technique as a yield detector in metals and also did
pioneering work in analyzing the fundamental characteristics of acoustic
emissions in metals. Recently, the most active acoustic emission work has
been in the area of nuclear pressure vessel proof-testing (19,20). A large
number of transducers are placed on the vessel, which is pressurized. Any
flaws that may be present are detected and evaluated by their acoustic
emission response. These flaws can be source-located to within inches of
their actual locations.
While the previously mentioned materials (rocks and metals) have been
the major subjects of acoustic emission research, other materials have also
been evaluated. These include composites, concrete, ceramics, ice, and
wood, and the results have been summarized in a number of review articles
written by Liptai elt a]_ (21), Dunegan and Tatro (22), Knill et al_ (23), and
Lord (3). In addition, a recent bibliography on the subjectTas been
compiled by Drouillard (24).
Information regarding the acoustic emission response of soils is
noticeably lacking in the literature. The original soils reference, stem-
ming from a rock monitoring program (25), appears to have been by Cadman
and Goodman (26), who addressed soils, per se, in a relatively preliminary
manner. Subsequent work has been done by the authors at Drexel University
over the past 6 years and is summarized in this report.
ACOUSTIC EMISSION SOURCES
To familiarize the reader with the concept of acoustic emission, it
is important to examine the initiation or source of the emissions in stress
materials of different types (including metals, single crystals, and rocks)
and to hypothesize behavior in granular and cohesive soils.
Metals
A wide variety of mechanisms can generate acoustic emissions in
metals. According to Pollock (27), the formation and propagation of
dislocations (defined by Van Vlack (28) as highly stressed crystalline
10
-------
imperfections usually appearing as line defects) and the fracture of
brittle, dissimilar particles locked within the crystalline structure
(i.e., inclusions (29)) can produce weak signals. Stronger signals are
produced by initiation and propagation of macroscopic cracks (i.e., those
greater than 0.01 mm (3.9 x 10-4 on.) in length). Though metallurgists
generally agree on the foregoing as being probable acoustic emission
sources, substantiation of this theory is difficult since few controlled
experiments of a fundamental nature have been conducted.
Ionic Crystals
Engle (30) used single crystals of lithium fluoride oriented for
minimum resistance to displacement. Acoustic emission and displacement
as small as 10-7 cm (4 x 108 in.) were measured during stress
application. In general, he found that acoustic emission activity was
directly related to the cause and nature of piled up dislocations.
Sedgwick (31) tested both lithium flouride and potassium chloride
in compression within the elastic range. He found that the rapid
dislocation movement that occurs in hard lithium flouride crystals
produced greater acoustic emission activity than the typically slow
dislocation movement occurring in the softer potassium chloride
crystals. In addition, his analysis of the acoustic emission
distribution formed the basis for a macroscopic deformation model for
lithium flouride. The model predicts values of dislocation strain,
dislocation density, and ultrasonic attenuation that agree well with the
experimental data.
Additional aspects of the study of acoustic emission initiation in
metals and crystals can be found in the review article by Lord (3).
Rock
Audible noises form cracking pillars and roofs in mines provided
the initial impetus for acoustic emission monitoring by Obert (8).
Laboratory testing of rock specimens convinced Scholz (32) that the first
signals received after the application of stress were caused by crack and
pore closure. Both the amplitude and number of emissions recorded then
increased continuously as macroscopic cracks were initiated and propagated
in first a stable, then an unstable manner. He finally concluded that,
when rupture of the specimen was near, friction along crack surfaces—as
well as crack propagation and coalescence—were contributing to the
acoustic emission activity.
Mineral and lithological differences among rock specimens have been
shown by Knill e_t aj_ (23) to affect the amount of acoustic emission
activity recorded. Chugh et jal_ (33) have noted changes in the nature of
emissions caused by varying moisture and stress conditions. The effects
of factors such as these will be addressed later in this study.
11
-------
Soils
The mechanisms responsible for the shear strength of soils appear to
be the basic generators of acoustic emissions in soils. These mechanisms
in granular soils are the fundamental components of the angle of shearing
resistance, including sliding and rolling friction, degradation, and
dilation (34,35). In simple cases, sandy (non-cohesive) soils, the angle
of shearing resistance, , is defined by the equation: shear strength (T)
= normal effective stress on the failure surface (Oj,) multiplied by tan ^ ;
in the case of a cohesive (clay) soil, a term "c", cohesion, in stress
units is added to the right side of the equation.) Evidence for such a
conclusion will be provided here to show that conditions producing the
greatest number of interparticle and therefore frictional contacts (i.e.,
well-graded soils) also produce the greatest level of acoustic emission
activity. The tendency of a granular soil to generate more emissions with
higher confining pressures and consequently higher frictional forces is
further evidence of a friction-based emission source.
Horn and Deere (36) have shown that the frictional characteristics of
soil particles vary with mineral type. One would then logically conclude
that mineral type will also affect acoustic emission activity, although
this hypothesis has not yet been tested.
The strength mechanisms for most cohesive soils in a drained test
include both friction and cohesion. Some perspective on the relative
contribution of these mechanisms to the acoustic emission behavior of soils
will be discussed later.
\
ACOUSTIC EMISSION APPLICATION WITH EMPHASIS ON CIVIL ENGINEERING
The civil engineering community has recently taken an interest in the
acoustic emission technique and is nondestructively monitoring a wide
variety of structures. This activity is important since this particular
group is the most likely user of the earth dam warning system developed in
this project; the greater their familiarity with the technique, the more
favorable will be their response.
Continuing with the classic work orginated in rock monitoring, Hardy
and Khair (37,38) have adapted the technique to determine the safety of
over-pressurizing underground gas storage facilities, and Mearns and Hoover
(14) have continued a long-term project of monitoring the stability of rock
highway slopes begun by Goodman and Blake (25). Closely related is the
work of Wisecarver ejt ail_.(39), who have used the technique to determine the
stability of large, open-pit mine walls and concluded that the technique is
a satisfactory means for monitoring slope stability in rock and for
determining the adequacy of corrective measures.
Liptai (40) and Hutton (41) report use of the technique to inspect
the safety of large crane rails and wooden roof trusses, and the compression
effects of tendons in prestressed concrete beams and even in bridges.
Regarding highway bridge inspection, Galambos and McGogney (42) include
12
-------
acoustic emission monitoring as a possible nondestructive testing method in
their recent state-of-the-art review.
With rapid growth predicted in materials transportation by pipeline
(43), leak detection and overstressing (44) become significant economic and
environmental problems. Acoustic emission techniques have been used with
considerable success on buried (45) and underwater pipelines (46) to
determine whether they are leaking and to determine the actual location of
the leak.
Another area recently studied using the acoustic emission technique
is that of monitoring stressed wire rope (47,48). Tensile tests have shown
that there is a direct correspondence between wire breakage and acoustic
emission events, and that damaged cables are more emittive at a given load
than undamaged ones (49).
13
-------
SECTION 6
FUNDAMENTALS OF ACOUSTIC EMISSION IN SOILS
In this project, soils were broadly classified as sandy soils and
cohesive soils.
Sand is a soil composed largely of silicaceous particles ranging in
nominal diameter from 0.074 mm (74 microns = 0.003 in. = retained by No. 200
sieve) to 4.76 mm (ca. 0.2 in. = passed by No. 4 sieve) and thus covering a
ratio of diameters (largest:smallest) of 65 : 1. The sands studied included
the following types: round (Ottawa sand), subround (beach sand), angular
(concrete-making sand), and subangular (sand drain soil).
Cohesive soils include clays, silty clays, and clayey silts. Clays
are soils capable of remaining in a plastic state over a relatively wide
range of water contents. A silt is a fine-grained soil of low plasticity.
Commonly, a silt is a fine sand that can float in a watercourse, but the
term is sometimes indicative of organic content.
Iti In is bluuy, all uuheSi've soils pabscli a Nu. 20C Sieve (-200
In classifying the cohesive soils, an important parameter is the liquid
limit, which is defined as the water content of a cohesive soil in which a
cut closes under specified test conditions. "L" means less that 50% water;
"H", greater than 50% water. Soils are also classified as "M" (silts and
silty clays), "C" (clays), and other types of no interest here. The co-
hesive soils tested included a clayey silt (ML); a kaolinite clay (MH)
(kaolin is a white clay of low plasticity); a silty clay (CL); and a
Bentonite clay (CH) — (Bentonite is highly plastic, results from decompo-
sition of volcanic ash, and swells considerably on wetting).
Additional description and characterization is included in the text
of this report (vide infra).
GRANULAR SOILS
This section describes the behavior of velocity, frequency, and
attenuation in granular soils such as gravels and sands and the effects of
several important physical properties of these soils.
Velocity of Acoustic Emissions (Elastic Waves) in Granular Soils
Although the subject of velocity of elastic waves in soils is not
used directly in our acoustic emission studies, it is of significance to
know how fast such waves travel from their source to the pickup
accelerometer or wave guide.
14
-------
Velocity measurements in soils were made in the following manner: a
small hammer regulated by a timing device was used to generate an impulsive
mechanical wave in a large tank containing silty sand. This reasonably
reproducible signal was then monitored by using two strategically placed
accelerometers from which the wave velocity was computed. This test was
actually developed to determine soil attenuation as obtained from the
difference in magnitudes of the two accelerometer responses, a topic that
will be examined more fully later in this section. Measurements in the
granular soil produced velocity values from 120 to 240 m/s (400 to 800
ft/s), depending on density and water content. These values are consistent
with the literature, which is quite abundant on this particular topic.
Thus, only one general reference by Hardin (50) is cited.
For the acoustic emission study presented here, this velocity
information leads to the conclusion that detection of wave pulses in small
laboratory samples is essentially simultaneous with their initiation.
Furthermore, the pulses are probably accumulations of all types of waves
(P, S, and R) generated at the individual sites with the soil mass, where
P = longitudinal elastic wave (primary wave); S = shear elastic wave
(secondary wave), and R = Rayleigh surface elastic wave (see Glossary).
Frequency of Acoustic Emissions in Granular Soils
Of considerable interest with regard to accelerometer selection,
sensitivity, monitoring procedure, etc., are the predominant frequencies of
waves emanating from stressed soil samples. To determine the rrequency
composition, a series of unconfined compression creep tests was performed
on dry, silty sands 70 mm (2.8 in.) in diameter and 150-mm (6.0-in.) high.
The emissions were converted to electric analogs, taped, and then played
back through a Bruel and Kjaer octave band filter. Tests resulted in the
response shown in Figure 2(a), where emissions are predominantly in the
500-Hz to 2-kHz region (Hz = cycles per second).
In addition, triaxial shear creep tests were performed on the same
soil at 17% water content and at a 69 kN/m2 (kilo-Newtons per square
meter) (10 Ib/in^ = 10 psi) confining pressure. Figure 2(b) shows this
response, where the dominant frequencies are now in the 4- and 8-kHz
bands. The mode of generation of the acoustic emissions has thus changed.
Our tentative explanation is that densification as a result of confinement
has allowed more of the higher frequency signals to pass through the soil
structure in a less attenuated manner than with loose density soils. Thus
the dominant frequencies have shifted upward.
Additional tests led to the conclusion that accelerometers having a
band width from 500 Hz to 15 kHz are adequate for acoustic emission studies
in soils. At frequencies lower than 500 Hz, background noise becomes very
troublesome whereas, at higher frequencies, essentially no undamped
emissions are present.
Attenuation of Acoustic Emission in Granular Soils
Although the tendency of soils to attenuate stress waves (especially
15
-------
10
LJL§30
§520
10
0
y ; I
£030
I/)
si
LULU 20
o
10
125 250 500 1k
FREQUENCY BAND
(b)
2k Ak
(Hz)
8k 16k
Figure 2. Frequency distribution for silty sand soil tested in
(a) unconfined compression and (b) triaxial shear at
69 kN/nr (10 psi) confining pressure
16
-------
in comparison to other construction materials) is generally known, the
actual values of attenuation and their dependence on frequency are largely
unknown and of great importance in this study, since high attenuation
decreases the volume of soil from which emissions can be transferred to the
wave guide rod and thus requires the use of more rod and/or greater signal
amplification. At the low frequency range, Hardin's (51) logarithmic
decrement data on sands (obtained in resonant column tests) can be
converted to attenuation of approximately 0.007 dB/cm (0.2 dB/ft) at 200
Hz. Further review of the literature shows that Cadman and Goodman (26)
measured attenuation of approximately 0.09 dB/cm (2.7 dB/ft) at 500 Hz
using a sand embankment model in which failure (sand movement) was produced
by tilting the supporting surface. These differences suggest a frequency
dependence that must be explored further, considering the frequency range
of soil emissions previously analyzed.
Mentioned in the section on velocity (vide supra) was a soil tank
assembly wherein a pulse was generated and multiple signal pickups were
used to compute attenuation values. Tests conducted at a frequency of
approximately 1,000 Hz on a dry, silty sand resulted in an attenuation of
approximately 1.3 dB/cm (40 dB/ft).
A test setup described by Nyborg, Rudnick, and Shilling (52) was
duplicated to determine attenuation values at still higher frequencies. In
this method, a loudspeaker generates a continuous signal. A layer of soil
is placed between this loudspeaker and a microphone pickup. The microphone
response, measured in decibels as a function of soil layer thickness, thus
determines the attenuation in the soil. The frequency capability of this
system ranges from a few kilohertz up to the frequency limits of the
transducer system (limited by the speaker), which is about 18 to 20 kHz.
Frequencies of 1 to 2 kHz and below cannot be reliably tested because of
the large physical dimensions of the frame required at these long wave
lengths.
Initial tests using the loudspeaker/microphone technique compared
favorably with the published results of Nyborg et _al_ (52), and use of the
method was extended into the frequency regime o7~our interest. In general,
the attenuation values are very high and quite sensitive to changes in
water content. For example, a change in water content from 0% to 12%
decreased attenuation in silty sand by approximately 200%. The approximate
attenuation recorded by using this method varied from 5 dB/cm (150 dB/ft)
at 4 kHz, to 10 dB/cm (300 dB/ft) at 16 kHz, which are the highest values
observed in this investigation.
After considering the various methods used to determine attenuation
(each of which has been determined in a different regime), it is possible
to look at frequency versus attenuation on a unified basis. Figure 3 shows
the approximate response curve for granular soils. A pronounced difference
in behavior is to be noted at approximately 1 kHz. Below this level, atten-
uation is relatively low, and above it it is high. Since most of the acous-
tic emissions are in the 500 Hz to 8 kHz region, the pickup accelerometer
must be placed directly at the source of the emissions in laboratory
specimens; but special treatment will be required when dealing with the
17
-------
1000
i r
i i i r
100
CD
10
LU
0.1
iLDSPEAKER TECHNIQUE-
PULSE METHOD
-CADMAN AND GOODMAN
•HARDIN
J I
J I
68 10 12 1A
FREQUENCY (kHz)
J I
16 18 20
Figure 3. Frequency versus attenuation response of dry
granular soils using various techniques.
18
-------
monitoring of large earth masses in the field. When acoustic emissions are
monitored in the field, metal rod wave guides must be used to conduct the
emissions to the ground surface, where they can be monitored and recorded.
Effects of Physical Characteristics on Acoustic Emissions in Granular Soils
During this phase of the study, four granular soils with different
physical characteristics were tested under drained (no excess pore water
pressure) conditions in both isostatic and triaxial creep modes. The
samples were tested in the creep (sustained stress) mode so that, compared
to conventional strength testing, machine noise would be eliminated. (Refer
to Appendix F. for definitions.) The soils varied in shape, uniformity,
and size. Table 2 lists the physical characteristics of the granular soils
tested (Figure 4).
All tests were performed on 70 mm-(2.8-in.) diameter by 150-mm
(6.0-in.) high samples in a consolidated, drained condition. The pickup
accelerometer was 12.7 mm (1/2-in.) in diameter by 19 mm (3/4-in.) in
length and was embedded in the center of the sample as it was prepared
(Figure 5) (see Glossary). The connecting coaxial cable was taken out
through a port in the cell to an amplifier and counter. The gains were set
equal for all tests in this series so that acoustic emission levels could
be compared. Stress and strain data were taken in a conventional manner.
In the first series of tests, hydrostatic pressure was applied to the
specimen, producing isostatic conditions. Cumulative acoustic emission
counts were recorded with time after the pressure increment was applied.
Figure 6 shows the response curves for these tests. Other than the final
level of acoustic emission counts (see Table 2), the time for the acoustic
emissions to cease (i.e., equilibrium of particle reorientation) varied
primarily with particle shape. Samples containing the rounder particles
(soil labeled No. 2 and No. 4) ceased emitting much before those with
angular particles. Further comparisons will be deferred until later in
this section.
Using the same soil samples and experimental test setup as with the
isostatic test results just covered, a series of triaxial shear creep tests
was performed. The deviator stress (or principal stress difference) versus
strain behavior is given in Figure 7, and the deviator stress versus
acoustic emission behavior is given in Figure 8 for the four soils under
consideration. Note the almost identical behavioral patterns of stress/
strain and stress/acoustic emission curves at all levels of confining
pressure. This behavior indicates a basic relationship between strain and
acoustic emission, the determination of which was a fundamental goal noted
earlier. In addition to listing the limiting.acoustic emission counts at
failure (last column of Table 2), a modulus of emittivity was also calcu-
lated. This is the slope of the initial portion of the deviator stress vs.
acoustic emission curves shown in Figure 8. The value listed in Table 2,
however, is the inverse of the slope and is expressed in units of counts
per kilo-Newtons per square meter (kN/m2) since the value is intuitively
more helpful on a unit stress basis. It is designated as the coefficient
of emittivity and has potential use in field monitoring studies.
19
-------
TABLE 2. EFFECT OF PARTICLE CHARACTERISTICS ON ACOUSTIC EMISSION IN GRANULAR SOIL
N)
Soil Number Particle Coefficient of
and type shape8 uniformity^
Sand drain soil Subangular 8.4
No. 1
Ottawa sand Round * 2.0
No. 2
Concrete sand Angular 2.4
No. 3
Beach sand Sub round 1.5
No. 4
a
Based on a relative scale of angular, sub angular,
jj
Defined as CU • dgQ/djj).
d^g, the particle size at which 10% of the entire
Effective Friction Cell pressure
sizec angle,0 kN
0.45 35 34.5
69.0
138.0
0.20 35 34.5
69.0
138.0
0.21 39 34.5
69.0
138.0 •
0.24 42 34.5
69.0
138.0
sub round, round, or very round.
sample is finer, given in millimeters.
SR
1.7
7.0
15.0
0.2
0.5
1.2
0.04
0.2
1.8
0.01
0.10
0.38
(x!02)
16.1
6.5
6.5
5.2
5.2
2.5
8.1
5.5
4.8
1.0
1.0
0.87
f
TRIAX
2
3
12
2
3
4
8
9
14
1
2
4
Cumulative acoustic emission counts under isostatic conditions at cell pressure equilibrium.
Coefficient of emittivity, i. e., slope of initial
counts/M.
m
Cumulative acoustic emission counts under triaxial
portion of AE versus deviator «tress
creep conditions at failure.
curve in
units of
- 6.895 pal
25.4 mm - 1 in.
-------
t-l AJ
T3 >H
0)
T3 r-l •
C "X3
cd M C
03 0) CO
<4-l U
0) ••> (0
tH T3 V
0) 05 «
&Cd^
3 §4?
•• 4J h
T3 4J
CD O M
W 0)
03 • S
4) *J 5
0) -H ••>
r-l tl -O
•H C
O V- «
(0 0) CO
0.
H (X 0)
cd 9 u
r-l h
SrH O
•H C
K O O
U « U
•3-
V
60
•H
fe
21
-------
t
-AXIAL STRESS
\r_
1BZ2
•CONFINING PRESSURE
ACCELEROMETER
"FLOATING"
ACCELEROMETER
"FIXED"
r/77777/77//JmV###r~ TO ACOUSTIC EMISSION
INSTRUMENTATION
Figure 5. Schematic diagram and photograph of acoustic emission monitoring
setup on stressed soil specimen tested in triaxial shear.
22
-------
-o15
CO
o 10
00
to
5
UJ
CO
<
SOIL N0.1
cT= 138 kN/ntf
cT= 69 kN/m2
cT= 345
5 10
TIME (min.)
15
CO
g tO
W
CO
1—4
LU
SOIL NO. 2
cT= 138 kN/m2
CT= 69 kN/m2
cT= 345 kN/m2
5 10
TIME (min.)
15
Figure 6. Isostatic test results (time versus acoustic emission in units
of 10,000 counts) for four granular soils listed in Table 2.
23
-------
I 2
to
to
H-«
Z
LU
to
SOIL NO. 3
=138 kN/m'
kN/m2
,345 kN/m2 i
5
TIME
10
(mia)
15
ACOUSTIC EMISSION 1x10*)
P P P
O Is) $x o*
SOIL .N(U
r «T= 138 kN/m
(^
-------
GJ 100
Q
SlOO
o
SOIL N0.1
* 138 kN/m2
CT =34.5 kN/m2
STRAIN (%)
SOIL NO. 2
= 138 kN/m2
= 69 kN/m2
= 3A5 kN/m2
STRAIN (%)
Figure 7. Triaxial shear test results (deviator stress versus
strain) for four granular soils listed in Table 2.
25
-------
SOIL NO. 3
= 138 kN/m2
7"= 69 kN/m2
2 k
STRAIN (%)
= 138 kN/m2
SOIL NQ4
3""= 69 kN/m2
s 34.5 kN/m2
STRAIN (%)
Figure 7. Continued
i
26
-------
600-
J2
ACOUSTIC EMISSION (x10*)
= 138 kN/m2
SOIL NO. 2
CT = 69 kN/m2
<7 = 3A.5 kN/m2
0 2 4 . _ 6
ACOUSTIC EMISSION (x105)
Figure 8. Triaxial shear test results (deviator stress versus
acoustic emission in units of 100,000 counts) for
four granular soils listed in Table 2.
27
-------
>100
o
£= 138 kN/m2
SOIL NO. 3
69 kN/m2
= 34.5 kN/m2
0 4' 8 12
ACOUSTIC EMISSION (x105)
kN/m2
= 34.5 kN/m2
0 2 A 6
ACOUSTIC EMISSION (x105)
Figure 8. Continued
28
-------
Because this portion of the project was aimed at the effect of
particle characteristics on acoustic emission generation in stressed soil
samples, a comparison can' now be made. The comments that follow are made
after a review of the data in Table 2 and Figures 6 and 8, and after
observations made during the actual testing and subsequent data evaluation.
Particle Shape—
The more angular the soil particles contained within the total sample,
the more emittive is the sample under stress. Samples Nos. 3 and 1 (angular
and subangular, respectively) are significantly more emittive in both the
initial and final stages of triaxial testing than the other two samples. In
general, the isostatic behavior of these angular soils particularly of soil
No. 1, is also more emittive. As previously noted, the time for them to
reach equilibrium is significantly longer than for those samples having
rounded soil particles.
Coefficient of Uniformity—
As the coefficient of uniformity increases, so does the level of
cumulative acoustic emissions. This is a strong conclusion for the
triaxial test behavior and is in nearly perfect agreement with the
isostatic test results. However, the more angular soils also happen to
have the highest coefficient of uniformity. The actual cause of greater
emissions may therefore be a combined effect.
Effective Size—
Because the range of effective size is quite limited (0.20 mm to 0.45
mm (0.0079 to 0.0177 in), little in the way of a firm conclusion can be
stated.
COHESIVE SOILS: CLAYS, SILTY CLAYS, CLAYEY SILTS
This section describes the behavior of amplitude, frequency, and
attenuation of acoustic emissions in cohesive soils as listed in Table 3
along with the effects of many macroscopic variables that influence
fine-grained, i.e., cohesive soil behavior.
Signal Amplitude in Cohesive Soils
Of primary interest is the comparison of levels of acoustic emission
signals emanating from three different soil types (soil No. 2 and 3
(granular) and No. 6 (cohesive). The soils to be evaluated were tested in
triaxial shear at a confining pressure of 34 kN/m2 (5 psi). The output
signal from an embedded Columbia 476-R accelerometer range in kHz was
amplified at a gain of approximately 30 dB (0.01 "g" sensitivity setting,
where "g" is the acceleration due to gravity) for all tests and fed
directly into an oscilloscope. The peak signal level was measured in volts
at different levels of axial stress, and an average was calculated for
signals observed at each stress level. For each load increment in both
granular and cohesive soils, both the rate and amplitude of observed
29
-------
TABLE 3. PROPERTIES OF COHESIVE SOILS USED IN THIS STUDY9
U)
o
Soil
numberb
5
6
7
8
Liquid Plastic Plasticity Specific gravity Optimum Cohesion
Soil description Typec limit limit Index (X) of solids water (kN/m?)
(X water) (X water) (Hi - WD) (g/cm3) Content(X)
(HI) (wp)
Clayey silt ML 47 37 10 2.62 23 41
Kaolinite clay MH 52 33 19 2.60 29 28
Sllty clay CL 43 24 19 2.64 34 48
Ben ton He clay CH 570 58 512 2.20 43 62
Friction**
angle
(o)
29
29
10
5
a All soils passed No. 200 sieve, unit weights varied slightly according to test series (see text).
b Soils 1
c Unified
1 kN .
i a -
through 4 are found In Table 2.
Soil Classification System.
6.895 psl
0.016 pcf
Friction angle » ^ In degrees. ^ is defined by the equation ~( • c + "fntan 0 where "t • shear stress,
c • cohesion, •(„ • normal stress on shear plane.
-------
emissions decreased with time from the start of the increment. Figure 9
shows the results of these tests on three different soils, for which the
following observations can be made:
— Both sands tested (soil Nos. 2 and 3) showed the same general
response, in which the amplitude of the emissions increased with
increasing stress levels up to failure. The average signal
amplitudes are 100 times stronger at failure than in the initial
stress range value. The response appears to substantiate the
commonly observed phenomenon of "hearing" a granular soil sample
when it approaches the failure state.
— The more angular concrete sand generally produced emissions of
slightly higher amplitude than the rounded Ottawa sand.
— The clay response is markedly different in a number of respects.
First, the signal levels from kaolinite clay (soil No. 6) are 1/2
to 1/400 those of granular soils at corresponding stress levels.
Second, the general nature of the response is different.
Initially, the signal level increases, as with sands; then it
levels off and, finally, decreases as the maximum stress is
approached. The explanation appears to involve particle
reorientation. The initial random orientation of the particles
is increasingly emittive up to 50% to 80% of the maximum stress.
At this point, a plastic state is fully mobilized, and the
particles become aligned with planes of maximum shear stress. As
additional load is applied, the amplitude of the acoustic
emissions begins to decrease until failure is reached. The
slickensided surfaces of the failed specimens tend to
substantiate this behavior. But the behavior described here is
likely to apply to cohesive1 soils failing by shear plane
development only, and not to those failing by bulging. (See
Glossary.)
Frequency Distribution of Acoustic Emissions from Stressed Cohesive Soils
To select the proper pickup transducer for acoustic emission
monitoring, a series of unconfined compression (see Glossary) tests were
performed on kaolinite clay (soil No. 6) samples at different water
contents. A typical response is shown in Figure 10, where the predominant
frequencies are seen to be in the 2 to 3-kHz regions. Similar tests were
conducted on confined kaolinite clay samples ranging in water content from
22% to 38%. The response was essentially the same. Note that these
acoustic emission frequencies are the same as those resulting from
unconfined tests on granular soils. However, when granular soils were
tested in a confined state, an upward frequency shift to 8 kHz was
observed. Such a shift did not occur in cohesive soils.
Though such information is of basic interest, its main practical
importance is in the selection of the proper pickup transducer.
Accelerometers that are responsive over the range of frequencies from 100
Hz to 8 kHz are judged to be well-suited to monitor all soil types. When
31
-------
—LEGEND—
D = CONCRETE SAND - SOIL NO. 3
O = OTTAWA SAND - SOIL NO. 2
O = KAOLIN1TE CLAY - SOIL NO. 6
0.001
20 40 60 80
PERCENT FAILURE STRESS
Figure 9. Average amplitude of acoustic emissions (measured
as peak signal voltage output) for various soils as
function of percentage failure stress in triaxial
creep at 34 kN/m (5 psi) confining pressure.
32
-------
i30
1— 1
z
LJJ
^20
o
1—
LL
O
UJ
O
on
LU
Q.
-
|
1
6 8 10 12 U
FREQUENY (kHz)
16
18 20
Figure 10. Frequency distribution of acoustic emissions from kaolinite clay
(soil No. 6) tested in unconfined compression at 33% water content.
33
-------
background noise at these monitoring frequencies is a problem, a high-pass
filter of about 1 kHz can be used to eliminate some ambient noise at low
frequencies.
Attenuation of Acoustic Emissions
The tendency of soils to attenuate various forms of elastic waves is
well-known. In the preceding section on granular soils, attenuation was
investigated, and its strong frequency dependence was quantified. Though
such data are pertinent in this study, the relationship between water
content and attenuation in cohesive soils should also be examined. Figure
11 shows how attenuation in clayey silt (soil No. 5) is affected as water
content varies from 0% to 15%. The tests were conducted using a stress
wave with a frequency of about 1 kHz. Details of the method were presented
earlier. The attenuation decreases from its highest value of 1.9 dB/cm (57
dB/ft) in the dry state to a low, and perhaps asymptotic, value of about
1.0 dB/cm (30 dB/ft) at a water content of 15%.
Though these large attenuation values are not considered to be a
dominant factor in laboratory monitoring of acoustic emissions (since the
pickup accelerometer can be placed right at the source of the emissions
within the sample), they do offer a severe challenge in field monitoring.
For this reason, the use of long metal wave guides to bring the emissions
to the ground surface is necessary in the field. The characteristics and
features of these wave guides will be examined in Section 7 of this report,
which describes field testing of the acoustic emission monitoring technique.
Macroscopic Behavior of Acoustic Emission in Cqhesive_Soils
Effect of Confining Pressure — The effect of confining pressure on
the acoustic emission behavior of cohesive soils was evaluated for two of
the four soils listed in Table 3. The clayey silt (soil No. 5), with a
total unit weight of 1.69 g/cc (105 Ib/ft3) and a void ratio of 0.95, and
Kaolinite clay (soil No. 6) with a total unit weight of 1.81 g/cc
(113/lb ft3) and a void ratio of 0.84, were each tested at confining
pressures of 34, 69, and 138 kN/m2 (5, 10, and 20 psi). As previously
noted, the tests were consolidated-drained sustained load (creep) tests.
The response curves are given in Figures 12 and 13. The close parallel in
the behavior of stress/strain and stress/acoustic emission curves can be
readily seen. Also, the fact that the overall acoustic emission count
levels are slightly higher for the clayey silt with its silt-sized particle
component than for the kaolinite clay is in agreement with the signal
amplitude study previously cited (vide supra). The reader may wish to
compare the level of acoustic emissions (counts) as a function of deviator
stress and confining pressure for clayey silt (soil No. 5), kaolinite clay
(soil No. 6), and the granular soils (Nos. 1 - 4, Fig. 8) to obtain a
better appreciation of the relative "noisiness" of these various soils
under different stress conditions.
This analogous behavior of strain and acoustic emission indicates
that the two parameters are related and that either can be used in
conjunction with stress to characterize and/or monitor a given soil.
34
-*».-
-------
5 10
WATER CONTENT (%)
15
Figure 11. Attenuation of acoustic emissions in clayey silt (soil No. 5)
at varying water contents at a frequency of about 1 kHz.
35
-------
-*
400
$300
LU
200
in
I
£100
Q
CLAYEY SILT - SOIL NO. 5
-- 34.5 kN/m2
10 15
STRAIN (%]
20
25
-------
KAOLINITE CLAY- SOIL NO. 6
(7 = 345 kN/m2
10 15
STRAIN (%)
138 kN/m2
20
25
30
= 69 kN/m2
CT = 34.5 kNAn2
12345
ACOUSTIC EMISSION COUNTS (x 1,000)
Figure 13. Triaxial creep response of kaolin!te clay
Csoil No. 6) at varying confining pressures.
37
-------
Effect of Water Content — The initial series of acoustic emission
tests reported in 1972 (2) were conducted on the clayey silt labeled soil
No. 5 in this study. The samples were compacted at different water contents
and tested in unconfined compression. Figure 14 shows the results, which
indicate a decrease in strength and acoustic emissions with increasing
water content. The extremely low number of emissions recorded at higher
water contents in cohesive soils emphasizes the susceptibility of the
technique to experimental error and noise interference as water content
approaches the liquid limit when cohesive soils are being monitored. At
the liquid limit, the soil changes from a plastic material that will
deform, but not crack, to a viscous liquid or slurry that will fill and
conform to the shape of a container. An earthen dam constructed of soil
that attains the liquid limit will already be failing (visual observation).
Effect of Plasticity Index — Table 3 indicates that the four
cohesive soils tested in this study had plasticity indices (PI) of 10%,
19%, 19%, and 512%. Each soil was compacted to achieve a void ratio of
0.89 and tested in consolidated-drained triaxial creep at 34 kN/m^ (5
psi) confining pressure. The results are presented in Figure 15.
Cumulative acoustic emission counts are plotted versus percentage failure
stress so that soils of different strengths can be directly compared. The
most emittive soil is the clayey silt (soil No. 5), which has the lowest
plasticity index and, correspondingly, the greatest amount of larger
silt-sized particles. The least emittive soil is bentonite clay (soil No.
8), with an extremely high plasticity index and no silt-sized material. As
shown in Figure 15, the kaolinite clay (soil No. 6) and silty clay (soil
No. 7) have approximately the same emission response and plasticity index.
Thus, a strong correspondence exists between acoustic emission response and
plasticity of fine-grained soils.
Effect of Sample Structure — All testing considered up to this
point has been on remolded samples prepared in the laboratory under closely
controlled and thus ideal conditions. Since one of the case histories to
be examined later provided the opportunity of obtaining undisturbed samples,
this soil (the silty clay labeled No. 7) was tested in the as-received
condition. The significant properties were 1.97 g/cc(123 Ib/ft^) total
unit weight, 1.14 void ratio, 56% water content, 100% saturation, and an
average penetration resistance of 10 to 20 blows/m (3 to 6 blows/ft).
The pickup accelerometer was embedded in the lower central portion of the
sample by augering a 12-mm (0.5-in.) diameter, 25-mm (1.0-in.) deep hole in
the soil sample and inserting the accelerometer. The sample was tested in
unconfined compression in the creep mode. Results are shown in Figure 16.
Note that the acoustic emission level is low, partly because of the cohesive
character of the predominantly clay soil and its relatively high water
content. However, the acoustic emission response closely resembles the
stress/strain behavior shown in the upper part of the figure.
Effect of Stress History — Well-established in acoustic emission
literature (53) is the so-called Kaiser effect, in which acoustic emission
levels are low until a material is stressed beyond that level which it has
experienced in the past. Thus, many materials retain a record of their
38
-------
600
^500
z
-400
to
co
LU
to
<200
x
<
100
CLAYEY SILT - SOIL NO. 5
400 600 800 1000
ACOUSTIC EMISSION COUNTS
1200 1400
Figure 14. Stress/acoustic emission response of clayey silt (soil No.5)
at varying water contents in unconfined compression.
39
-------
a
A
o
O
CLAYEY SILT - SOIL NO. 5
KAOLINITE CLAY - SOIL NO. 6
SILTY CLAY - SOIL NO. 7
BENTONITE CLAY - SOIL NO. 8
0
1JOOO 2,000
ACOUSTIC EMISSION COUNTS
3,000
Figure 15. Stress/acoustic emission reponse of four cohesive soils in
triaxial creep tests showing significance of plasticity index.
-------
AOO
«M
.£
1300
to
to
£200
<100
x
SILTY CLAY - SOIL NO. 7
2
STRAIN
4)0
C-l
§300
to
CO
£200
< 100
I— I
x
100 200 300
ACOUSTIC EMISSION COUNTS
400
500
Figure 16. Unconfined compression test results for undisturbed
sample of silty clay (soil No. 7) at 56% water content.
41
-------
stress history. This concept has been recognized in geotechnical
engineering through the identification of the preconsolidation pressure as
determined in a standard consolidation test.
In this phase of the study, we have adapted such a stress history
test for acoustic emission monitoring by fixing an accelerometer to the
upper load platen of a consolidation odeometer. (A consolidation odeometer
is a device for conducting confined compression tests as a function of time
to determine the compression characteristics of soil.) Observation showed
that the upper porous stone of the odeometer did not significantly
attenuate the signals. Tests were conducted in a standard manner, which
deflection/time and acoustic emission/time data sets generated for each
pressure increment. The soil was a sandy silty clay known locally'as a
preconsolidated marl of low plasticity. It was used only for this test
series and therefore is not listed in Table 3. The pertinent properties of
this soil were 26% water content, 100% saturation, 1.59 g/cc (99 Ib/ft3)
dry unit weight, 2.65 specific gravity, 0.69 void ratio, a liquid limit of
30%, and a plastic limit of 27%. Figure 17 shows the typical response at a
given pressure increment using a log-of-time fitting method. The standard
deflection plot is roughly reflected in the curve of acoustic emission
counts; that is, during periods of low deflection, acoustic emission count
rates were low, and during periods of high deflection, rates were high.
The fact that the time of transition from low to high rates of deformation
does not completely coincide with the time of transition from low to high
acoustic emission rates cannot presently be explained without further
investiaation. From a series of plots such as these, a nrpssurp/strain
curve was developed (see the upper portion of Figure 18), and a
preconsolidation pressure of 408 kN/m2 (53 Ib/in? (psi) = 3.8
tons/ft2 (tsf)) was determined by the Casagrande technique.
Additionally, the time for 50% consolidation, t$Q, of each pressure
increment was used to obtain an acoustic emission count at 50%
consolidation. The acoustic emission data were normalized by dividing the
accumulated emission count at t$Q for each pressure increment by the
total emission count registered during all pressure increments. The
results are given on the lower portion of Figure 18. The response consists
of two nearly straight lines intersecting at about 858 kN/m2
(8.0 tons/ft2 = 111 psi. Though this value does not coincide with the
preconsolidation pressure, it does coincide with the beginning of the
straight line portion of the virgin compression curve. Most important,
however, is that the acoustic emission levels are generally lower at stress
levels below the preconsolidation pressure, Pc, than they are at stress
levels that exceed pc. Thus stress history seems to be identifiable using
the acoustic emission monitoring technique.
COMPARISON OF ACOUSTIC EMISSIONS FROM GRANULAR AND COHESIVE SOILS
The following important observations were made:
Both for granular (sand) and for cohesive (clay and clayey silt)
soils, stress vs. cumulative emissions curves corresponds closely to
stress vs. strain curves. Thus, acoustic emissions are an indicator
of deformation.
42
-------
0.6
.£ Q5
o
LJL
LU
O
0.3
120
£ 100
z
8 80
z
o
to 60
to
o
o
o
<
40
20
ill MIT
150
11 ml i i i 11 ml
0.1
10
TIME
100
Imin)
ml
1000 10,000
Figure 17. One-dimensional consolidation response of sandy silty
clay at constant pressure on log-time scale.
43
-------
z^ 10
t—I O
cc
^5
p20
<
30
Pc - 408 kN/m2
0.1
1 10
PRESSURE (kN/m2)
100
CO
co
2
LJ
co
S 860
M LLJ
cr
o
z
80
100
01
.
"
= 858 kN/m2
1 10
PRESSURE (kN/m2)
100
Figure 18. One-dimensional consolidation response of sandy silty clay over
range of pressures showing strain and acoustic emission responses,
44
-------
Most cohesive soils, and certainly all granular soils, can be
monitored successfully using the acoustic emission technique.
Possible exceptions are highly plastic clays with high water content
(poor candidates for dam construction materials).
All granular soils (sands) tested showed the same general response.
The behavior of cohesive soils is more dependent on water content and
soil characteristics.
The amplitude of the emissions for the sands increased with stress up
to failure, and the rate of increase was greatest as failure stress
was approached.
The average signal amplitude for sands is 100 times greater near
failure than at 20 percent of failure stress.
The cohesive soil (clay and silt) response is markedly different.
Significant signals are not emitted from clay until a higher percent
failure stress is reached than with granular soils.
Signal levels in clay are from 1/2 to 1/400 the level of signals from
the sands at corresponding stress levels.
Initially, the signal level for the clay increases, as with sands;
but then it levels off and decreases as the maximum stress is
approached.
The decrease in signal level in clay as failure is approached is
believed to result from reorientation of the pi ate!ike clay
particles. At first they are randomly oriented, and emissions
increase with increasing stress. But then the particles become
aligned with the direction of maximum shear stress, and emissions
decrease.
Emissions in cohesive soil decrease with increasing water content,
and are reduced to very low levels as the liquid limit (water content
beyond which a mass of soil cannot sustain a shear force) is
approached.
A strong correspondence exists between acoustic emission response and
plasticity (ability to be reshaped without developing surface
cracks). Cohesive soils with the highest plasticity index give least
response.
The frequency distributions of the acoustical emissions from granular
and cohesive soils peak in the 2- to 3-kHz region at low stress. As
confining pressure increases, the frequency shifts upward to 4 to 8
kHz for sands but remains nearly unchanged in cohesive soils.
There are marked differences in sound velocity, and attenuation
coefficient, for granular and cohesive soils.
45
-------
In sands, the acoustic emission increases monotonically with water
content both because of lower attenuation and because of continuing
contact between particles (at least until the sand liquifies and
loses all shear strength).
In cohesive soils, the particles tend to move away from one another
as water content increases, thus reducing the frictional interaction,
which is a major cause of the emissions. Cohesion of particles is
the major factor in the (shear) strength of clays, but release of
cohesive energy does not provide as many emissions (acoustic energy)
as does the overcoming of friction in sands (conversion of potential
to various forms of kinetic energy). (The primary shear strength in
sand arises from sliding and rolling friction and from structural
resistance.)
ESTIMATED MAGNITUDE OF ACOUSTIC EMISSIONS IN SOIL
This section (54) presents a method to estimate the magnitude of the
acoustic emissions at their source and at various distances from the source,
e.g., at a transducer or wave guide pickup. It is based on very basic phys-
ical concepts and, as such, is intended to give a first order of approxima-
tion only. Furthermore, some of the values required for the numeric solu-
tion are rough estimates, which require a much more extended effort for a
more exact value.
Theory
Consider a volume, V, of soil that is under stress
and subsequently deforms elastically. From simple elastic
tneory, tne elastic energy, u, stored in this volume is:
U = 1 M e2V (1)
7
where M = elastic modulus and
e s elastic strain.
If this energy is released in a time interval, t, the average
elastic power, P, released as waves during this interval is:
P= UR (2)
At
where R is the radiation efficiency, i.e., the fraction of
total energy released that is converted into elastic waves
(acoustic emissions). If this energy spreads uniformly in
three dimensions, the intensity, I, (i.e., the power per
unit area) at a distance r from the source will be:
46
-------
where r is the distance from the source to the monitoring
point. The simplified relationship between pressure, p,
and intensity is (55):
I = P2 (4)
^
where (^ = density of the material
c = wave velocity.
It is easily shown (55) for sinusoidal waves that the maximum
displacement is:
Pmax
where pmax is the maximum pressure, k is the wave number
(27T/X), and Xis the wave length. Thus the monochromatic,
spherical displacement wave spreads out as:
sin (tot - kr) (6a)
y = PniSlsin (wt - kr) (6b)
where o> * 2iff is the angular frequency, and f is the frequency
of the wave. Using Eq. (4) the above can be written as:
sin (wt - kr) (7)
The particle acceleration is then:
— - "2 s1n (wt • kr)
dt2 = " ^C2k
and using o>> a kc (55) one obtains for the maximum particle
acceleration:
max
Using Equations (1) (2) and (3) substituted into Equation (9)
along with the relationship M = pc2, (55) the resulting
maximum acceleration is:
2Atr2
47
-------
Equation (10) thus represents the maximum acceleration of the
acoustic wave produced by the source at some finite distance away.
Parametric Evaluation
To illustrate the significance of the developed theory, Equation
(10) is solved using the following estimated values:
f = 500 Hz (from references 3 to 6)
e = 0.002 (typical elastic strain in soils at the end of elastic
range)
V = 3.5 x 106 cc, i.e., a 5 ft. cube (an estimated soil volume
undergoing elastic deformation)
R = 0.001 (seismology estimate from Cook (56))
c = 18,300 cm/sec (600 ft/sec)
At = 0.1 sec (a rough estimate)
r = 760 cm (25 ft ) (for waves to be isotropic, r must be
significantly larger than the soil volume under consideration)
Use of the above data results in a maximum acceleration of 42 cm/sec2, and
using the acceleration of gravity, g, as 980 cm/sec2, the maximum acceler-
ation becomes 0.042g. This value of acceleration appears to be compatible
with the accelerations that have been measured in the field. However, due
to the many variables involved (and the very real possibility of compen-
sating errors), a parametric study of each variable is presented in Figures
19(a) to 19(g). In each figure, all the variables are kept constant (as per
above) except the one being studied. (The figures should be read with a
certain degree of caution, for in any real source modeling, there will
certainly be interrelations between the above parameters. The most obvious
is that as At decreases, f will increase. Also R will depend on f and c in
some complicated manner.)
Acceleration at the Source
The previous theoretical development and numeric example were for
the elastic spreading of the acoustic emissions as they propagate beyond
the strained zone into the adjacent soil mass. The soil attenuation can
now be superimposed onto the problem as follows: Consider a soil
attenuation of 1.0 dB per foot, which corresponds to a dry granular sand at
about 500 Hz. Over a distance of 762 cm (25 ft), this means that 25 dB of
signal strength has been lost in traveling from its source to the
monitoring station. Therefore, in the example problem stated previously,
which resulted in a maximum acceleration of 0.042g at 762 cm (25 ft) from
the source, the source acceleration would have been:
(dB) = 20 log (amax at source/amax at station)
25 = 20 log (amax at source/0.042g),
thus, amax at source = 0.75g.
48
-------
.15g
o
pj
cc
.10g
UJ
o
.05g-
x
<
0 5 10 15
DISTANCE FROM SOURCE (m)
Figure 19(a). Variation of amax with
distance from source
in equation 10.
z
o
5
tr
0.3g
UJ
aig
I
0.1 0.2
TIME INTERVAL (sec.)
03
Figure 19(b). Variation of a with
TTlflX
time in equation 10.
Q3gr
o
I
Ul
x
<
1,000 2JOOO 3JOOO
FREQUENCY (Hz)
03g
UJ
Q2g-
x
<
_L
J_
.002
STRAIN (ratio)
.006
Figure 19(c). Variation of 3^^ with Figure 19(d). Variation of a^g^ with
frequency in equation 10. strain in equation 10.
49
-------
z
o
Q2g-
UJ
o
o
x
-------
SECTION 7
APPLICATIONS OF ACOUSTIC EMISSIONS IN SOILS
In a soil mass, acoustic emissions are the self-generated noises that
develop as a result of a deformation. The emissions are intimately related
to the mobilization of shear strength components within the soil itself.
Such components as sliding friction, rolling friction, degradation, dilata-
tion, and probably cohesion all play a role in generating acoustic emis-
sions. The resulting acoustic emissions are received by a metal rod wave
guide, which is embedded in the soil and which transmits the emissions to
an accelerometer (in this case a piezoelectric transducer with a relatively
flat frequency response from 500 to 5,000 Hz) attached to the wave guide at
the ground surface. The accelerometer then converts these mechanical waves
into electrical pulses that are amplified and counted to obtain a numeric
result. (See Figure 1 for a schematic of a typical acoustic emission moni-
toring system.) The use of electrical bandpass filters is optional and
depends to a large degree on the level of background noise. A recorder is
used if a hard copy of the results is required.
EQUIPMENT
For the field work described in this section, a monitoring system
consisting of a Columbia 476-R accelerometer, Columbia VM-103 amplifier, and
Hewlett-Packard 5300A counter was used. These components are all DC-opera-
ted and thus require no external power source—an important consideration
for field work in remote areas. Other details concerning the instrumenta-
tion and relationships of acoustic emissions to basic soil properties were
presented earlier. Photographs of the equipment are included in Section 8
and Appendix B.
Unlike some field structures (rock formations), soils require an
extrinsic mechanism to bring the acoustic emissions from within the soil
mass, where they are generated, to the ground surface, where they can be
monitored. Such transmission element (called a wave guide) is necessary
because of the high attenuation of elastic waves in soils. In most other
non-soil structures, the pickup sensor can be mounted directly on the
material being monitored and then retrieved upon completion of the work.
The wave guides may simply be lengths of low-carbon-steel rod (e.g., bar or
rod stock), reinforcing bar, bailing wire, instrument pipe, drain or outlet
pipe, etc. that are driven into place when existing soil masses are to be
monitored or, whenever possible, are placed in an earthen dam during
construction. The wave guide must be placed in or near to a highly
stressed zone in the soil being monitored. Choosing the best location is a
difficult decision, not unlike the selection of instrumentation sites
irrespective of the particular technique. The acoustic emissions generated
51
-------
in the soil as it deforms will travel relatively unimpeded along the wave
guide to the pickup accelerometer, which is threaded onto the wave guide at
the ground surface. The following series of tests was performed to verify
the conducting quality of the wave guide.
A steel wave guide was set up in the laboratory with a small
minishaker attached to one end and a pickup accelerometer, amplifier, and
readout oscilloscope to the other end. The shaker and pickup accelerometer
were both mounted axially. Therefore this study is predominantly one of
longitudinal wave propagation. Two different-size wave guides were
used--3.2 mm (1/8 in.) and 12.7 mm (1/2 in.) in diameter—each with
different lengths, surface conditions, and coupling mechanisms. Figure 20
shows the response obtained for wave guides up to 4.9 m (16 ft) long, the
larger diameter being associated with the longer of the two sizes tested.
Conclusions drawn from this portion of the study are as follows:
— Longer wave guides lower the frequency of the first resonance,
making the system more sensitive in the 500 to 1,000 Hz range,
where a large number of soil emissions actually occur (see
Figures 2 and 10).
— Different diameter rods do not appear to influence the first
resonant frequency of the system.
— Different surface conditions (threaded versus smooth) do not
appear to affect the location of the first resonance.
— The method of connecting one rod to another does not appear to
influence the resonances as long as such connections are solid
and firm in their metal-to-metal contact.
These four conclusions are consistent with the fact that the first
and higher resonances are caused by standing congressional elastic waves in
the rod. From wave propagation theory, the values of the lowest rod
resonances can be written simply as f0 = v/2L, where f0 is the resonant
frequency, v is the velocity of sound in the rod, and L is the length of
the rod. The velocity of sound in the rod was determined by measuring the
time necessary for an elastic pulse to traverse the rod. The measured
velocity of 4.5 x 105 cm/sec (1.5 x 104 ft/s) for a rod 4.9 m (16 ft)
long gives a theoretical resonant frequency of 465 Hz, a value reasonably
close to the measured value.
Still of concern regarding the resonant response of the wave guide/
accelerometer system was the influence of the soil medium around the wave
guide. To study this effect, a sequence of tests was conducted on a 1.2-m
(4-ft) long, 12.7-mm (1/2-in.) diameter rod surrounded by silty sand of
various densities. The results (Table 4) indicate that the location of the
first resonant frequency is only slightly varied by the influence of the
surrounding medium. A slight peak is evident with the soil at 10% water
content and at the highest density. This higher density condition has the
effect of maximizing the particle contacts on the rod, which in turn lowers
the amplitude of the first resonance. In general, it appears that the soil
52
-------
\\\\\\\v
B&K
MINI- SHAKER
WAVE GUIDE x ,
ACCELEROMETER-^
AMPLIFIER
/\ /•/**! 1
OSCILLOSCOPE-
1.27cm DIA SMOOTH ROD
0.318cm DIA THREADED ROD
1 23 4 5
FREQUENCY OF FIRST RESONANCE (kHz)
Figure 20. Experimental setup and location of wave guide/accelerometer's
first resonance as a function of length considering different
diameter and geometry of steel rods.
53
-------
TABLE 4. INFLUENCE OF MEDIUM SURROUNDING WAVE GUIDE ON
FREQUENCY AND AMPLITUDE OF FIRST RESONANCE
Item
Wet density (gm/cm3)
Dry density (gm/cm3)
Resonant Frequency (Hz)
Amplitude (V.)
Loose
dry
soil
1.36
1.36
1,630
28
Dense
dry
soil
1.46
1.46
1,670
15
Soil at water content
10% 20% 30% Water
1.57
1.46
1,840
5
1.70
1.31
1,730
4
1.89
1.44
1,690 1,632
11 83
Air
-
-
1,630
96
ui 1 g/cm3 = 0.016 pcf (pounds per cubic foot)
V. = volts
-------
mass around the wave guide has a negligible influence on the detector's
frequency response, but does reduce the signal magnitude to a certain extent.
FIELD STUDIES
Nineteen field sites have been or are in the process of being moni-
tored using the acoustic emission technique. A brief description of each and
some comparative details are given in Table 5. Twelve are earth dams, two
are surcharge fills, two are embankments, one is a gypsum dam, and two are
seepage studies. Each site will be described briefly, along with the princi-
pal finding. The more important and informative sites will be examined in
greater detail.
Site No. 1 is a 9.1-m (30-ft) high homogeneous earth dam near Doyles-
town, Pennsylvania. The 3.0-m (10-ft) deep foundation soil was instrumented
for potential settlement by driving twenty 12.7-mm (1/2-in.) diameter steel
rods to the underlying rock. During placement of the fill, no emissions
were recorded, undoubtedly because the foundation soils were very dense (130
Ib/ft^ = 2.08 g/cc) and had very high strength (standard penetration
resistance of approximately 164 blows/m (50 blows/ft)). When coupled with
the fact that the dam is relatively small with reasonably flat side slopes,
this lack of emission data seems justified.
Site No. 2 is a 20.1-m (66-ft)-high zoned earth dam near Doylestown,
Pennsylvania. The dam is founded just above rock and was completed before
its downstream slope was instrumented with twelve 12.7-mm (1/2-in.) diameter
steel rod guides 3.0 to 4.6 m (10 to 15 ft) long. The purpose of this
instrumentation was to monitor lateral embankment movement as the reservoir
filled with water. For a number of reasons, the filling was very slow, so
the embankment was subjected to relatively small increments of lateral
pressure. This reservoir has taken 3 years to fill. At no time were
emissions recorded, suggesting little or no deformation and a stable dam.
Site No. 3 is a 20-m (67-ft)-high homogeneous earth dam near McCook,
Nebraska, and is shown in schematic form in Figure 21. This dam is founded
on approximately 61 m (200 ft) of wind-blown silt generally known as loess.
The dam was instrumented with 9.5-mm (3/8-in.) diameter reinforcing rods
placed horizontally on the foundation soil at four stations, each consisting
of a set of three rods of varying length. At one point during construction,
it was found that, for two of these stations, the emissions shown in Table 6
occurred. The data appear to indicate that the longer rods on the far side
of the crest of the dam give higher emission counts than the shorter rods,
which terminate in the slope area. This response is logical because less
loading, and hence less deformation, occurred in these regions. The fact
that set No. 1 responded more than set No. 2 is not completely understood,
since fills were slightly greater in the area of set No. 2.
However, emissions from the dam were very low overall (the data in
Table 6 were the maxima recorded) during the entire course of construction,
which lasted for 1-1/2 years. The soil's fine particle size, the prewetting
of the construction site, the horizontal rod placement, and the short-term,
random collection of data are all believed to have contributed to the low
levels of emission.
55
-------
TABLE 5. OVERVIEW OF SITES BEING MONITORED USING THE ACOUSTIC EMISSION METHOD
ON
Height
No.
1
2
3
4
5
6
7
8
9
10
1)
12
13
14
15
16
17
IB
19
Designation
Pa-616
Pa-617
Neb-200
Md-BSC
Pa-PIA
Neb-390B
Can-LMM
Del-GOC
Pa -BOB
NJ-RR
Va-KPN
NY-OSW .
Pa-OSPl
Pa-DSP2
Pa-LN
Tex-OC
Ky-WC
Del-CW
NY-ASP
Purpose
Flood control
Recreation
Flood control
Ore stockpile
Surcharge load
Flood control
Tailings dam
Contain dredging
spoil
Water supply
Contain chemical
wastes
Contain chemical
wastes
Contain petroleum
wastes
Stockpile for highway fill
Stockpile of highway fill
Seepage beneath earth dam
Contain chem. waste
Waste water storage
Water supply
Recreation
ft
30
66
67
40
6
68
95
15
40
120
8
20
4
15
8
20
15
15
12
too
28
18
60
ID
9
20
20
12
1
20
29
4
12
36
2
6
1
4
2
6
4
4
3
30
9
6
20
Length
ft
2600
2500
900
300
120
600
900
6 mi
600
4 mi
500
450
20
60
1200
3 mi
3 mi
200
1500
ffl
800
760
270
90
37
180
270
10 km
180
7 km
150
140
6
18
370
5 km
5 km
60
500
Relative
Embankment
Design and Const.
Excellent
Excellent
Excellent
Good
Good
Excellent
Good
Poor
Excellent
Poor
Poor
Poor
• Poor
Poor
Good
Good
Good
Good
Good
Relative
Foundation
Stability
Excellent
Excellent
Compressible
Very poor
Very poor
Compressible
Good
Very good
Excellent
Very poor
Unknown
Unknown
Good
Good
Poor
Satisfactory
Unknown
Good
Good
-------
Ln
•-J
WAVE GUIDE LOCATION TO
MONITOR EMBANKMENT AND
FOUNDATION MOVEMENT
(WAVE GUIDES HORIZONTAL)
SET-1
T-3
Figure 21. Elevation and plan views of site No. 3 near McCook, Nebraska,
showing horizontal wave guide location scheme.
-------
TABLE 6. ACOUSTIC EMISSIONS FROM NEB-200 DAM SITE
Set No.
1
1
1
2
2
2
Rod No.
3
2
1
3
2
1
Approx.
length
ft m
250 76
210 64
140 63
270 82
230 70
160 49
Acoustic
emission
counts/mi n
1958
639
80
16
7
0
Site No. 4 is a 12.2-m (40-ft) high stockpile of iron ore at
Sparrow's Point, Maryland. It is adjacent to the location of a previous
failure that resulted from overloading of the soft foundation soils by the
ore. Four wave guides were placed: Two of them were horizontal beneath
the fill, and two were vertical, beside the fill, penetrating deeply into
the foundation soil. In addition to the standard type of instrumentation
shown in Figure 1, an oscilloscope was included in the system monitoring
the emissions. As fill was being placed, the emissions were noticeable on
the oscilloscope but were not strong enough to trigger the counter. This
lack of measurable emissions may be an indication that the Kaiser effect
occurs in soils as it does in other materials. For cyclically loaded
materials, the Kaiser effect predicts an absence of emissions during stress
re-application until the highest previous stress level experience by the
material has been attained. This particular site had been previously
loaded with 12.2 to 21.3 m (40 to 70 ft) of iron ore many times in the
past. Thus, the relatively low emission readings (and settlements for that
matter) may be attributable to this preloading or stress history condition.
Low emission levels at stresses less than the preconsolidation pressure were
also observed in laboratory consolidation tests, as reported in Section 6.
Site No. 5 is a field study at the Philadelphia International Airport
in Philadelphia, Pennsylvania, and is shown schematically in Figure 22. A
surcharge fill has been placed around a previously installed, end-bearing
pile to determine how much load will be added to the pile as a result of
soil consolidation. This test constitutes a full-scale, negative skin
friction or downdrag test. The test piles and settlement anchors were
employed as acoustic emission wave guides to monitor the deformation of the
settling soil. Figure 23 presents these results, where the similarity
between the settlement/time and acoustic emission/time response curves
should be noted. The fact that the acoustic emission response dissipated
after 5 to 15 days actually agrees better with theoretical computations,
using standard consolidation theory, than the 2 to 3 days for the settlement
response. This significant case history illustrates the effectiveness of
the acoustic emission monitoring techniques.
Site No. 6 is a 20.7-m (68-ft) high earth dam in the same watershed
as site No. 3. Based on the experiences of that previous case history,
vertical settlement plates were chosen for wave guides. These were 25.4-mm
58
-------
ACOUSTIC
EMISSION
EQUIPMENT
1.8m
SOIL SURCHARGE FORCING SETTLEMENT
Ki
3-4.5 m
8
z
o
SAND AND GRAVEL
m
6-9m
6-9m
COMPRESSIBLE CLAYEY SILT
I
o
m
\ /
\/
DENSE SAND
Figure 22. Elevation view of site No. 5 in Philadelphia, Pa., showing surcharge
load and compressible soil along with different types of wave guides.
-------
TIME (days)
10 15 2Q_
LEGEND-
TOP OF LAYER O
MID LAYER •-
BOTTOM OF LAYER
10 15 20
TIME (days)
30
Figure 23. Time/settlement and time/acoustic emission
response curves from site No. 5.
60
-------
(1-in.) diameter rods within a 7.6-cm (3-in.) casing and would
therefore serve both purposes. Though settlements were indeed recorded,
acoustic emissions were not because of our inability to separate the actual
signal from the banging of the rods within the casing. Thus no information
was available from this site—a fact that illustrates the frequent problem
of distinguishing signals from ambient noise.
Site No. 7 is an existing mine tailings dam in Quebec, Canada, that
is being raised an additional 7 m (23 ft), making the total height 29 m
(95 ft). Because a failure occurred at the site previously, greater than
normal concern is being given to its stability. Both newly installed
vertical wave guides and horizontal steel pipe drains are being used as
wave guides. The site, however, is currently dormant because of mine
inactivity.
Site No. 8 is a 4.6- to 12.2-m (15- to 40-ft) high homogeneous earth
dam in Delaware City, Delaware, containing dredging spoils. Though the
embankment itself appears stable, it is founded on a high-water content
clayey silt of standard penetration resistance as low as 3 to 7 blows/m
(1 to 2 blows/ft). The foundation soils are instrumented with 12.7-mm
(1/2-in.) vertical rods, and acoustic emission counts vary from 0 to 10
counts/min. No noticeable long-term trends have been observed over the 18
months that this site has been monitored.
Site No. 9 is a 36.6-m (120-ft) high, zoned earth dam in Boyertown,
Pennsylvania, constructed on rock containing an old inactive fault. The
site is monitored with four sets of vertically placed, 12.7-mm (1/2-in.)
diameter reinfnrcing rods. Each set has three bars of different lengths.
The reservoir has recently been filled with no resulting acoustic emissions.
Site No. 10 is a system of holding ponds for various chemical waste
liquids in New Jersey. The embankments vary in height from 2.4 to 6.1 m (8
to 20 ft), have steep side slopes (about 1 on 1), and are founded on
extremely poor foundation soils. These foundation soils are silty clays
and clayey silts of standard penetration resistance from 0 to 16 blows/m
(0 .to 5 blows/ft). A deep-seated base stability failure had occurred at
the site before acoustic emission monitoring began. The site has since
been monitored with twelve 12.7-mm (1/2-in.) diameter wave guides that were
easily installed by pushing them, as 1/2-m (4-ft) sections, into the founda-
tion soils to depths up to 6.1 m (20 ft). Acoustic emission activity is
usually present, but count rates vary considerably. As an example, wave
guide No. 7 (at the toe of the slope in the vicinity of the failure) has
given the following response:
October 8, 1975 - 10 to 30 counts/min
November 5, 1975 - 0 to 5 counts/min
December 4, 1975 - 0 to 5 counts/min
February 25, 1976 - 20 to 40 counts/min
March 17, 1976 - 5 to 10 counts/min
June 3, 1976 - 10 to 15 counts/min
November 9, 1977 - 5 to 20 counts/min
June 7, 1978 - 10 to 30 counts/min
61
-------
August 2, 1978 - 5 to 10 counts/min
October 6, 1978 - 0 to 20 counts/min
December 1, 1978 - 0 counts/min
March 27, 1979 - 10 to 40 counts/min
The site is an active one in which a permanent, on-line monitoring system
has been recommended to the owners. Until such time as a continuous
monitoring system is installed, periodic visits will be made.
Site No. 11 is the site of a small earth dam in Hopewell, Virginia,
that bounds a lagoon containing an aqueous chemical Kepone in solution and
as a sediment. Until such time that this extremely active substance can be
"neutralized", the integrity of the impoundment, which is adjacent to the
James River, must be assured. Four vertical wave guides (12.7-mm (1/2-in.)
diameter steel rods) were placed through the embankment and into the
foundation soil. Only one of these, at the location of obviously poor
construction, gave any emission response (0-3 counts/min); maintenance work
on the embankment was recommended to the owners to correct the situation.
Site No. 12 is almost identical to Site No. 11 except that the stored
liquid consists of contaminated water, waste petroleum products, and sludges
from various industries in and near Oswego, New York. There had been a
failure of the dam, partly from erosion, and partly from overtopping as a
result of inadequate freeboard,, The break was repaired, and acoustic
emission monitoring via four vertical wave guides was conducted. Signifi-
cant acoustic emission activity was measured and corrective action was
recommended. The contents of the lagoon were subsequently removed.
Equilibrium of the earth dam was thus restored, and the acoustic emissions
ceased.
Site No. 13 is a 4.6-m (15-ft) high stockpile of fill that was
eventually used in the construction of an interstate highway in
Philadelphia, Pennsylvania. Though it was not an engineered embankment, it
did provide us with the opportunity of bringing a site to failure. A large
front-end loader was used to excavate the toe of the slope for a length of
about 6.1 m (20 ft) in a series of cuts, thereby incrementally decreasing
the stability of the slope. One vertical wave guide about 3.0-m (10-ft)
deep was installed at the top of the slope.
After each cut, the engine that powered the loader was shut off so
that acoustic emission readings could be made without high background
noise. Figure 24 gives the count rates for the four cuts made in bringing
the slope to failure. For the first two cuts, the acoustic emission count
rates—recorded as soon as the loader engine was stopped—attained their
maximum values and rapidly decreased thereafter. No data were obtained
during the third cut because of wind noise effects on the accelerometer at
the relatively unsheltered site. The accelerometer was subsequently
wrapped in a foam blanket. The fourth cut resulted in the same trend as
the earlier ones until 20 minutes after the cut was made, when the count
rate increased rapidly. The increased count rate was accompanied by the
detachment of a large mass of soil from the top of the slope, an event that
could easily be classified as a failure. Though this set of data is far
62
-------
U)
LU
200
150
o
o
o
HH
18
100
o
t-«
in
I
50
0
ICUT 1
CUT 2
NO DATA FOR CUT 3
BECAUSE OF WIND
INTERFERENCE
iCUT 4
50 100 150 200
TIME FROM FIRST CUT (min)
CUT1 CUT 2 CUT 3
CUT
FAILURE
250
Figure 24. Acoustic emission count rate versus time of cut
for site No. 13 showing failure after fourth cut.
-------
from complete or clean, it gave us the encouragement needed to proceed to
the more detailed and carefully controlled case history that follows.
Site No. 14 was a 4.6-m (15-ft) high stockpile of earth that was
being stored for future highway construction near the Philadelphia
International Airport. The soil was sampled, tested, and subsequently
found to be a well-graded, silty sand with some clay (approximately 17%).
The in-situ water content was approximately 12%, and the average unit
weight was 2.00 g/cc (125 lb/ft3). Consolidated-drained triaxial shear
tests resulted in an angle of shearing resistance of 16° and a cohesion
of 11.0 kN/m2 (1.6 lb/in2 (psi)).
The site had a slope of approximately 1:1. A relatively uniform
18-m (60-ft) long section was selected for excavation. The excavation was
made using a large front-end loader, which made successive cuts from the
toe of the slope (Figure 25). Before excavation, the site was instrumented
with the following systems:
— A grid of surface stakes was installed to be monitored using
standard surveying methods.
— Soil strain gages were embedded in the slope and at the top of
the slope to measure deformation and thus obtain soil strain.
-- Slope inclinometers were installed to measure horizontal
movements along a line at the top of the slope.
-- Steel-rod wave guides were installed vertically at the top of the
slope for acoustic emission monitoring.
The excavations were made as shown schematically in Figure 24, where the
first cut of 17 m3 (22 yd3) produced little in the way of visual
movement of the slope. The second cut of 55 m3 (67 yd3) was made 4
days later. A small tension crack was noted slightly above the cut and
extended for a length of approximately 10.7 m (35 ft). The third cut of 72
m3 (94 yd3) was made 3 days later, and tension cracks were again noted
one meter or so above the top of the cut. While this cut was open, a
relatively heavy rain occurred. The fourth cut of 98 m3 (128 yd3) was
made 8 days later, and tension cracks were very evident extending up to and
beyond the top of the slope for the entire length of the slope. Rain again
fell during the time period when this cut was open. The fifth and last cut
of 110 m3 (144 yd3) was made 6 days later. Thirty-seven minutes after
the cut was made, a large wedge of soil separated from the main embankment
and collapsed into the area where the previous cuts had been made. This
event was considered to be an actual earthen bank failure; monitoring was
discontinued shortly thereafter.
Throughout this excavation process, monitoring was continued for as
long as possible using the techniques described. Relevant comments
regarding the results follow:
64
-------
o>
FAILURE WEDGE
TO A.E. READOUT
EQUIPMENT
TENSION CRACK
FROM CUT
Figure 25. Schematic diagram of site
No. 14 showing approximate
boundaries of five cuts
made and photographs after
cuts Nos. 1 and 4.
CUT
-------
1. The horizontal movement of the surface stakes indicated a
gradually increasing movement away from the slope during the
first three cuts. The movement immediately after the third cut
averaged about 0.51 cm (0.20 in.). As noted earlier, both the
second and the third cuts resulted in tension cracks that caused
a loosening of the wooden stakes in the slope area above the cut
to the point where readings were no longer reliable. The slope
stake readings were discontinued at this point.
2. Soil strain gage readings were also interrupted by the tension
cracks, since the coils used in this technique were hand-placed
near the surface of the slope. They were judged to be reasonably
accurate, however, until shortly after the fourth cut was made.
At that time, the soil strain gages indicated an average strain
of approximately 0.4%. The readings were not uniform, however,
and they initially showed a slight compression before indicating
tension. The sensors were easy to install and to calibrate
initially, and they probably gave a reasonable assessment of the
strain conditions up to the point of large-scale cracking of the
embankment.
3. The slope inclinometers responded after the first cut was made,
showing a movement of 0.76 to 1.02 cm (0.3 to 0.4 in.) ranging
from zero at the top of the slope and zero at the bottom. After
the first cut, and for most of the cuts thereafter, little
additional movement was detected. Apparently, the deeper soil
beneath the near surface did not deform enough to be accurately
measured by this method. All three inclinometers gavp
essentially the same information.,
4. The acoustic emission response curves for each of the five cuts
are shown in Figures 26 through 30. From these curves, the
following observations can be made:
a. Each response from the first four cuts indicates a high
acoustic emission response initially, then an approximately
exponential decay with time until stability of the particular
cut is reached.
b. The fifth and last cut follows this general trend, but 30
minutes after the cut was made, the acoustic emission rate
began to increase rapidly. When the count rate reached its
maximum, about 7700 counts/min, a large section of soil
pulled away from the intact mass and slid down the remaining
slope. Thereafter, the count rate began to subside and
eventually came to equilibrium. The post-failure count
curve rate appears to rejoin the original curve (shown as
the dashed line in Figure 30).
c. Not indicated on these figures is the effect of rain on the
acoustic emission count rate. Approximately 8,200 min (5.7
days) after the third cut was made, a heavy rainfall caused
66
-------
14
£ „
o
LLJ
\
0 10 20 X
TIME (min.)
Figure 26. Acoustic emission response
2QOr after cut No. 1.
10
20 30 2000
TIME (min.)
2020
Figure 27. Acoustic emission response
after cut No. 2.
67
-------
80
60
8 *
LU
<: 20
10 20 30 40
TIME (min.)
Figure 28. Acoustic emission response
after cut No. 3.
eooo
2-400
co 300
§200
uJ
< 100
) 120 160
TIME (min.)
Figure 29. Acoustic emission response
after cut No. 4.
68
-------
20 40
TIME (min.)
Figure 30. Acoustic emission response
after cut No. 5.
69
-------
the count rate to rapidly increase to 200 counts/min. After
1,300 minutes (0.9 day), the count rate was back to its
former level of 2 to 5 counts/min. Rain again interrupted
the testing program after the fourth cut was made. Approxi-
mately 3,000 min (2.1 days) after the cut was made, rainfall
occurred and the count rate increased to 350 counts/min.
After an additional 2,500 min (1.7 days), the count rate
decreased to zero. Thus it took a longer period for the
slope to readjust to equilibrium when the rainfall ceased
after cut No. 4, a fact that may be due to the gradual
decrease in the slope's factor of safety. Regardless of the
relative magnitudes involved, it can be concluded that the
two rainfalls did have at least a temporary effect on the
slope's stability.
Additional data can be obtained from this particular site by plotting the
acoustic emission count rate of each cut (Figure 31). Here are presented
curves for both the maximum count rate and the average count rate during
the 1-hour period after monitoring began. The response curves are plotted
for the first four cuts, and thereafter the count rates increase rapidly, •
as indicated. This type of behavior that loss of stability in slopes is
not a linear process, but rather one in which instability occurs at a
rapidly increasing rate as failure is approached.
Site No. 15 is an acoustic emission study of seepage beneath a 3.6-m
(12-ft) high earth dam in northeastern Pennsylvania. Since this application
of the technique is slightly different, it will be described separately in
Appendix C of th^s reports
Site No. 16 involves the stability monitoring of piles of waste
gypsum material, some of which was used to form a dike containing waste
liquid. The site is near Houston, Texas, adjacent to the Houston Ship
Channel. Acoustic emission wave guides have been installed at numerous
locations around the area (which is actually in the form of three separate
piles) and data are being collected by the owner. Acoustic emission
monitoring equipment has also been purchased by the owner who, after a
number of field visits, is monitoring the site with his own personnel.
This particular site is of additional importance since the data are being
compared to other geotechnical monitoring systems, i.e,, piezometers. We
are in regular correspondence with the owner's representatives on this
particular site.
Site No. 17 is a waste water storage facility including sludge
lagoons, aeration ponds, and stabilization lagoons in Kentucky near the
Mississippi River. Embankment heights vary from 4 to 9 m (13 to 28 ft)
with some relatively steep slopes of up to 40° from horizontal. Erosion
of the slopes is easily observable.
Eight wave guides were installed to depths ranging from 1 to 4 m (4
to 12 ft). The highest acoustic emission count rate recorded was 2
counts/min with most locations registering 1.4 counts/min or less. We feel
that no deformation at the site is presently occurring and, with proper
70
-------
800-
to 7700
to 3100
AVE. A.h. KATtb
10 15
TIME (days)
20
CUT NUMBER
Figure 31. Summary of acoustic emission rates
after each cut.
71
-------
embankment maintenance, a stability failure is not likely to occur.
Site No. 18 is a water reservoir in Delaware, which was drained for
inspection of the bottom. Subsequent to drawdown, a localized section of
the sloped innerside slid into the empty reservoir. After this initial
movement had occurred, we were asked to monitor the slope to see whether
instability was a continuing problem and to assess remedial measures. As
the data of Figure 32 indicate, settlements were still continuing and did
so up to 127 cm (50 in.) of settlement. Acoustic emission monitoring wave
guides were installed immediately and indicated initially high count rates
(from February 3 to 10, 1978), followed by a periodic decreasing count rate
until no emissions were detected after March 2, 1978. This latter stage
corresponded to the lack of recorded settlement. The relatively high
acoustic emission behavior between February 21 to 28 was a result of
localized sloughing of the soil in the failure wedge's falling against the
wave guides and resulting in high emission count rates. This behavior was
not due to instability of the main failure wedge itself. This case history
illustrates the need for continuous monitoring in many natural situations.
During and after remedial work to the failed slope, no acoustic emission
counts were recorded.
Site No. 19 is an earth dam in western New York that bounds a
reservoir used for recreation purposes. It was inspected and found to be
leaking at and beyond its downstream toe in a number of locations. Ten
acoustic emission wave guides were installed to measure whether soil
deformation was accompanying the seepage and to detect exactly where the
seepage paths were located. At this time, it appears that the dam is
stable but requires adaitional monitoring for actual seepage detection.
Appendix C will further elaborate on the application of acoustic emission
monitoring to seepage problems.
72
-------
£ 0
gjO.5
TIME (1978)
2/10 2/20
-LEGEND
—• SP-1 W6-1
—o SP-2 WG-2
—A SP-3 WG-3
—a SP4 WG-4
c
'£
16
I/)
I/)
UJ
O
'
2/10
2/20
TIME (1978)
Figure 32. Settlement and acoustic emission response
curves from site No. 18, showing response
at various locations along slide area.
73
-------
SECTION 8
SPILL ALERT DEVICE DETAILS
A series of photographs of the acoustic emission system used for
field and laboratory monitoring is given in Figures 33 through 37. The
field (basic) system (Figures 33 and 34) consists of an accelerometer,
amplifier and counter. (A list of equipment suppliers with approximate
prices as of December, 1978 is given in Table 7.) For the operation of
this basic system, a user's manual has been prepared and is included as
Appendix B of this report.
The laboratory (modified) system (Figures 35, 36, and 37) is
intended for laboratory work where power is available and portability is
not a critical factor. This modified system has, in addition to the basic
system, two alternative high pass filters—one at 500 Hz and the other at
1,500 Hz. These filters eliminate some of the background noise that often
accompanies work in crowded areas. Also provided for in this modified
system is a chart recorder for a permanent record of the emission levels.
Any one of a number of commercial recorders can be used.
The possibility of signature analysis (from either basic or modified
systems) from a taped emission or entire test sequence is also possible
using advanced computer techniques common to the acoustic emission industry
in aerospace and nuclear applications. The literature abounds with such
methods, but their application was judged to be beyond the scope of this
research and development program because of high costs. The emphasis in
this program has been to obtain an approximate, qualitative assessment of
earth dam stability. Any count rate readout and recording (beyond visually
observing a meter reading and recording the data in a notebook) raises the
system's cost considerably; the cost of such improvements can, certainly,
be justified in industrial and commercial applications.
-------
Figure 33. Photograph of acoustic emission field system, showing
(right to left) steel wave guide rod with coupling
and attached accelerometer (note that, for purposes
of illustration, the components have not been fully
screwed together), thin coaxial cable of moderate
length, amplifier (center, reading full scale) with
coaxial cable connection to battery-powered electronic
counting system (left, reading 000969).
75
-------
Figure 34. Photograph of acoustic emission system in actual field
use, showing (right to left) waveguide, coupling, and
accelerometer (an unused wooden stake is also shown),
coaxial cable connecting the accelerometer to the
amplifier (center, lying on the ground), and battery-
powered electronic counting unit with coaxial cable
connection to amplifier. The carrying case for all the
equipment is shown to the right of the amplifier.
76
-------
EPA Sponsor^ -
Figure 35. Photograph of laboratory version of acoustic emission
system, showing (right to left) disassembled triaxial
cell test unit where accelerometer and short waveguide
can be seen supported by rubber band assembly, cable,
and instrument chassis that contains amplifier, counting
unit, and strip-chart recorder.
77
-------
EPA Sponsored Resell
Figure 36. Photograph of front view of laboratory-use acoustic
emission system, showing (left to right) strip-chart
recorder, counting unit (displaying a count of 000120)
and amplifier. Not visible (rear, behind counter) are'
band-pass filters.
-------
Figure 37. Photograph of side and front of laboratory-use acoustic
emission system, showing (left to right) side panel
controls (for power, band-pass filters, etc.), strip-
chart recorder, counting unit (without battery power
component, but with count selector), and amplifier.
79
-------
TABLE 7. COMMERCIALLY AVAILABLE ACOUSTIC EMISSION EQUIPMENT
Price as of
Equipment December 1978
Accelerometer:
Columbia; Model No. 476 $175 (less than 6)
nominal resonance =7.5 KHz $155 (6 to 10)
Amplifier:
Columbia; Vibration Meter
model VM-103 $395
Electronic counting system:
Hewlett-Packard;
5300 A Measuring System $500
5304 A Timer/Counter $385-
5310 Battery Pack $275
Cable Connectors:
B & K Instruments;
Coaxial microdot cable,
item AO-0037 $2/ft
Microdot to Microdot
JP-0012 connectors $3
Addresses of vendors cited: -
B & K Instruments, Inc.
5111 W. 164th Street
Cleveland, Ohio 44142 (216)267-4800
Columbia Research Laboratories, Inc.
McDade Blvd. & Bullens Lane
Woodlyn, Pa. 19094 (215)872-0381
Hewlett-Packard Co.
King of Prussia Industrial Park
King of Prussia, Pa. 19406 (215)265-7000
80
-------
REFERENCES
1. Sowers, 6. F., "The Use and Misuse of Earth Dams," Consulting Engr.,
July 1962. See also: Sowers, G. B., and Sowers, G. F., Introductory
Soil Mechanics and Foundations, 3rd Edition, Macmillan, New York, 1970.
2. Koerner, R. M. and Lord, A. E., Jr., "Acoustic Emissions in a Medium
Plasticity Clayey Silt," Jour, of Soil Mechanics and Foundations Div.,
ASCE, Tech. Note, Vol. 98, No. SMI, January 1972. pp. 161-165.
3. Lord, A. E., Jr., "Acoustic Emission - A Review," in Physical
Acoustics, Vol. 11, W. P. Mason and N. Thurston, Eds., Academic Press,
1975. pp. 289-353.
4. Koerner, R. M., Lord, A.E., Jr., McCabe, W. M. and Curran, J. W.,
"Acoustic Emission Behavior of Granular Soils," Journal of the
Geotechnical Engineering Division, ASCE, Vol. 102, No. GT7,
July 1976. pp. 761-773.
5. Koerner, R. M., Lord, A. E., Jr., and McCable, W. M., "Acoustic
Emission Behavior of Cohesive Soils," Journal of the Geotechnical
Engineering Division, ASCE, Vol. 103, No. GTS, Auyust 1377.
pp. 837-850.
6. Koerner, R. M., Lord, A. E., Jr., and McCabe, W. M., "Acoustic
Emission (Microseismic) Monitoring of Earth Dams," Proceedings of
Engr. Fdtn. Conf. on the Evaluation of Dam Safety, Calif.,
November 1976.
7. Koerner, R. M., Lord, A. E. Jr., and McCabe, W. M., "Acoustic Emission
Monitoring of Soil Stability," Journal of the Geotechnical Engineering
Division, ASCE, Vol. 104, No. GT5, May 1978. pp. 571-582.
8. Obert, L., "Use of Subaudible Noises for Prediction of Rockburst,"
RI-3555, United States Bureau of Mines, 1941.
9. Obert, L., and Duval, W. I., "Microseismic Method of Predicting Rock
Failure in Underground Mining: Part I and Part II," RI-3797 and 3803,
United States Bureau of Mines, 1946.
10. Hodgson, E. A., Bulletin of the Seismological Society of America,
Vol. 32, 1942. p. 249.
11. Hodgson, E. A., Transaction of the Canadian Institute of Mining and
Metallurgy, Toronto, Ontario, Canada, Vol. 46, 1943. p. 313.
81
-------
12. Blake, W., Leighton, F., and Duvall, W. I., "Microseismic Techniques
for Monitoring the Behavior of Rock Structures," Bulletin 665, United
States Department of the Interior, Bureau of Mines, 1974.
13. Hardy, H. R., Jr., "Evaluating the Stability of Geologic Structures
Using Acoustic Emission," ASTM-STP-571, American Society for Testing
and Materials, Philadelphia, Pa., 1975.
14. Mearns, R., and Hoover, T., "Subaudible Rock Noise (SARN) as a Measure
of Slope Stability," Final Report CA-DOT-TI-2537-1-73-24, United
States Department of Transportation, Federal Highway Administration,
August 1973.
15. Kaiser, 0., "Untersuchungen uber das auftreten Geraushen Beim
Zugversuch," Thesis presented to Technische Hochschule at Munich,
Germany, 1950.
16. Kaiser, Jr., "Erkenntnisse und Folgerungen aus der Messung von
Gerauschen bezugbeanspruchung von Metal!ischen Werkstoffen," Arkiv.
fur das Eisenhuttenwesen, Vol. 25, 43, 1953.
17. Tatro, C.A., and Liptai, R. G., "Acoustic Emission from Crystalline
Substances," Proceedings of the Symposium on the Physics of
Nondestructive Testing, Southwest Research Institute, San Antonio,
Texas, 1962. pp. 145-174.
18. Tatro, C. A., and Liptai, R. G., Proceedings of the 4th Symposium on
Nondestructive Testing of Aim-aft and Missile Components, Southwest
Research Institute, San Antonio, Texas, 1963. pp. 287-346.
19. Green, A. T., "Detection of Incipient Failure in Pressure Vessels by
Stress-Wave Emissions," Nuclear Safety, Vol. 10, 1969. pp. 4-18.
20. Nakamura, Y., "Acoustic Emission Monitoring System for the Detection
of Cracks in a Complex Structure," Materials Evaluation,
January 1971. pp. 8-12.
21. Liptai, R. G., "Acoustic Emission Techniques in Materials Research,"
International Journal of Nondestructive Testing, Vol. 3, 1971.
22. Dunegan, H. L., and Tatro, C. A., "Acoustic Emission Effects During
Mechanical Deformation," in Techniques of Metal Research, Vol. 5,
R. F. Bunshah, Ed., Interscience, New York, 1971.
23. Knill, J. L., Franklin, J. A., and Malone, A. W., "A Study of Acoustic
Emission from Stressed Rock," International Journal of Rock Mechanics
and Mining Sciences, Vol. 5, 87, 1968.
24. Drouillard, T. F., "Acoustic Emission: A Bibliography of
1970-1971-1972," ASTM-STP-571, American Society for Testing and
Materials, 1975.
82
-------
25. Goodman, R. E., and Blake, W., "Rock Noise in Landslides and Slope
Failures," Highway Research Board, Vol. 119, 1966. pp. 50-60.
26. Cadman, J. D., and Goodman, R. E., "Landslide Noise," Science,
Vol. 15, December 1, 1967. pp. 1182-1184.
27. Pollock, A. A., "From Metals to Rocks: Physics and Technology in
Common and in Contrast," Proceedings of the Conference on Microseismic
Activity in Geologic Materials, Pennsylvania State University,
University Park, Pa., June 9-11, 1975, Academic Press, Inc., New York,
1976.
28. Van Vlack, L. G., "Elements of Materials Science and Engineering,"
3rd ed., Addison-Wesley, New York, 1975.
29. Clark, D. S., and Varney, W. R., Physical Metallurgy for Engineers,
2nd ed., Van Nostrand, New York, 1962.
30. Engle, R. B., "Acoustic Emission and Related Displacements in Lithium
Fluoride Single Crystals," Thesis presented to Michigan State
University, Ann Arbor, Mich., 1966, (available through University
Microfilms 48160-6707535, Ann Arbor, Mich.).
31. Sedgwick, R. T., "Acoustic Emission from Single Crystals of LiF and
KC1," Journal of Applied Physics, Vol. 39, No. 3, 1968. pp. 1728-1740.
32. Scholz, C. H., "Mechanism of Creep in Brittle Rock," Journal of
Geophysical Research, Vol. 73, 1968. DP. 3295-3302.
33. Chugh, Y. P., Hardy, H. R., Jr., and Stefanko, R., "Investigation of
the Frequency Spectra of Microseismic Activity in Rock Under Tension,"
Proceedings of the Tenth Rock Mechanics Symposium, Austin, Texas,
May 1968.
34. Koerner, R. M., "Behavior of Single Mineral Soils in Triaxial Shear,"
Journal of the Soil Mechanics and Foundations Division, ASCE, Vol. 96,
No. SM4, Proc. Paper 7432, July 1970. pp. 1373-1390.
35. Lee, K. L., and Seed, H. B., "Drained Strength Characteristics of
Sand," Journal of the Soil Mechanics and Foundation Division, ASCE,
Vol. 93, No. SM6, Proc. Paper 5561, November 1967. pp. 117-141.
36. Horn, H. M., and Deere, D. U., "Frictional Characteristics of
Minerals," Geotechnique, London, England, Vol. 12, 1962. pp. 319-335.
37. Hardy, H. R., Jr., "Application of Acoustic Emission Techniques to
Rock Mechanics Research," STP-505, ASTM, Philadelphia, 1972.
pp. 41-83.
38. Hardy, H. R., Jr., and Khair, A. E., "Applications of Acoustic Emission
in the Evaluation of Underground Gas Storage Reservoir Stability,"
Proc. 9th Can. Rock Mech. Symp., Montreal, December 1973. pp. 77-111.
83
-------
39. Wisecarver, D. W., Merrill, R. and Stateham, R. M., "The Microseismic
Technique Applied to Slope Stability," Soc. of Mining Engineers,
Trans. A.I.M.E., Vol. 224, 1969. pp. 378-385.
40. Liptai, R. G., "Acoustic Emission from Composite Materials," Report
URCl-72657, Lawrence Radiation Lab., Livermore, Calif., 1971.
41. Mutton, P. H., "Acoustic Emission Applied Outside of the Laboratory,"
STP-505, ASTM, Philadelphia, 1972. pp. 114-128.
42. Galambos, C. F., and McGogney, C. A., "Opportunities for NOT of
Highway Structures," Materials Evaluation, ASTM, Vol. 33, No. 7,
July 1975. pp. 169-175.
43. Williams, D. R., Jr., "Five Decades of Progress in Pipelining," Jour.
of Const. Div., ASCE, Vol. C04, December 1975. pp. 751-767.
44. Koerner, R. M., Lord, A. E., Jr., and Deisher, J. N., "Acoustic
Emission Leak and Stress Monitoring to Prevent Spills from Buried
Pipelines," Proc. of 1976 Natl. Conf. on Control of Hazardous Matls.
Spills, New Orleans, pp. 8-15.
45. Parry, D. L., "Industrial Applications of Acoustic Emission Analysis
Technology," STP-571, ASTM, Philadelphia, 1975. pp. 150-183.
46. van Reimsdijk, A. J., and Bosselaare, H., "On Stream Detection of
Small Leaks in Crude Oil Pipelines," Proc. 7th World Petroleum Conf.,
Wnl C Mnvir-~ 1QC7 -r OOO OCO
• W I . \/ , I iWS* IWtS, I^W«. f^p. C. «•>•/ t» W •
47. Laura, P. A., Vanderveldt, H., and Gaffney, P., "Acoustic Detection of
Structural Failure of Mechanical Cables," Jour. Acoust. Soc. of Amer.,
Vol. 45, No. 3, 1969. pp. 791-793.
48. Laura, P. A., Vanderveldt, H. H., and Gaffney, P. G., "Mechanical
Behavior of Stranded Wire Rope and Feasibility of Detection of Cable
Failure," Marine Technology Society Jour., Vol. 4, No. 3, 1970.
pp. 19-32.
49. Harris, D. 0., and Dunegan, H. L., "Acoustic Emission Testing of Wire
Rope," Tech. Report DE-72-3A, Dunegan/Endevco, Livermore, Calif.,
October 1972.
50. Hardin, B. 0., "The Nature of Damping in Sands," Journal of the Soil
Mechanics and Foundations Division, ASCE, Vol. 91, No. SMI, Proc.
Paper 4206, January 1965. pp. 63-97.
51. Hardin, B. 0., "Elastic Wave Velocities in Granular Soils," Journal of
the Soil Mechanics and Foundations Division, ASCE, Vol. 89, No. SMI,
Proc. Paper 3407, January 1963. pp. 33-65.
84
-------
52. Nyborg, W. L., Rudnick, I., and Shilling, H. K., "Experiments on
Acoustic Absorption in Sand and Soil," J. Acous. Soc. Amer., Vol. 22,
No. 4, July 1950. pp. 422-425.
53. Kaiser, J., "Erkenntnisse und Folger ungen aus der Messung von
Gerauschen bei Zugbeanspruchung von metal!ischen Werkstoffen," Arkiv
fur das Eisenhuttenwesen, Vol. 1/2, January/February 1953. pp. 43-45.
54. Lord, A. E., and Koerner, R. M., "Estimated Magnitude of Acoustic
Emissions in Soil," Tech. Note, Journal of the Geotechnical
Engineering Div., ASCE, 1979.
55. Halliday, D., and Resnick, R., Physics, Wiley, New York, 1966.
56. Cook, N. G. W., "Seismicity Associated With Mining," Engineering
Geology, 10, 1976. pp. 99-122.
57. U.S. Environmental Protection Agency Proposal Solicitation, "Petroleum
Pipeline Leak Detection Study," RFP No. Cl-76-0145.
58. Proc. First Intl. Conf. on the Internal and External Protection of
Pipes, Univ. of Durham, England, (BHRA Fluid Engr., Pub!.),
September 9-11, 1976.
59. Koerner, R. M., Lord, A. E., Jr., and Deisher, J. N., "Acoustic
Emission Stress and Leak Monitoring to Prevent Spills from Buried
Pipelines," Proc. 1976 Nat!. Conf. on Control of Hazardous Materials
Spills, New Orleans, La., April 25-28. 1976. DD. 761-773.
60. Lord, A. E., Jr., Deisher, J. N., and Koerner, R. M., "Attenuation of
Elastic Waves in Pipelines as Applied to Acoustic Emission Leak
Detection," Materials Evaluation, ASTM, Vol. 35, No. 11,
November 1977. pp. 49-60.
61. McCabe, W. M., Koerner, R. M., and Lord, A. E., Jr., "Acoustic
Emission Behavior of Concrete Laboratory Specimens," ACI Journal,
July 1976. pp. 367-371.
85
-------
APPENDIX A
PUBLISHED AND/OR SUBMITTED TECHNICAL PAPERS
ON ACOUSTIC EMISSION MONITORING
1. Koerner, R. M., and Lord, A. E., Jr., "Acoustic Emissions in a Medium
Plasticity Clayey Silt," Tech. Note, ASCE, Journal of Soil Mechanics
and Foundations Div., Vol. 98, January 1972. pp. 161-165.
2. Lord, A. E., Jr., and Koerner, R. M., "Acoustic Emission Response of
Dry Soils," Jour, of Testing and Evaluation, ASTM, Vol. 100, No. 3,
May 1974. pp. 159-162.
3. Koerner, R. M., and Lord, A. E., Jr., "Earth Dam Warning System to
Prevent Hazardous Material Spills," AIChE/EPA Conference on Control of
Hazardous Material Spills, San Francisco, Calif., August 1974.
4. Koerner, R. M., and Lord, A. E., Jr., "Acoustic Emission in Stressed
Soil Samples," J. Acoust. Soc. Am., Vol. 56, No. 6, December 1974.
pp. 1924-1927.
5. Lord, A. E., Jr., and Koerner. R. M.. "Acoustic Fmissions in Soils and
Their Use in Assessing Earth t)am Stability," Jour. Acoust. Soc. Am.,
Vol. 57, No. 2, February 1975. pp. 416-419.
6. Lord, A. E., Jr., and Koerner, R. M., "Application of Acoustic
Emission Techniques to Materials Studies - Soils," Amer. Soc. for
Nondest. Testing Handbook, R. C. McMaster, Ed., to be published.
7. Koerner, R. M., and Lord, A. E., Jr., "Application of Acoustic
Emission Monitoring to Earth Dam and Foundation Stability," Amer. Soc.
for Nondest. Testing Handbook, R. C. McMaster, Ed., to be published.
8. Lord, A. E., Jr., and Koerner, R. M., Fundamental Studies of Acoustic
Emissions in Soil and Laboratory Specimens, First Conf. on Acoustic
Emission in Geologic Structures and Materials, Pennsylvania State
Univ., Vol. 2, No. 3, Trans. Tech. Pub!., Switz., June 9-11, 1975.
9. Koerner, R. M., and Lord, A. E., Jr., Applied Studies of Acoustic
Emissions in Soil Masses at Field Sites, First Conf. on Acoustic
Emission in Geologic Structures and Materials, Pennsylvania State
University, Vol. 2, No. 3, Trans. Tech. Pub!., Switz., June 9-11, 1975.
86
-------
10. Koerner, R. M., Lord, A. E., Jr., and McCabe, W. M., "Acoustic
Emission Monitoring in Concrete and Foundation Soils," Conf. Proc. of
Analysis and Design of Foundations for Tall Buildings, Lehigh Univ.,
Aug. 4-8, 1975. pp. 637-653.
11. Koerner, R. M., and Lord, A. E., Jr., "Acoustic Emission Monitoring of
Earth Dam Stability," Water Power and Dam Construction, Vol. 28,
No. 4, London., April, 1976. pp. 45-49.
12. Lord, A. E., Jr., Koerner, R. M., and McCabe, W. M., "Acoustic
Emission Behavior of Sand as Used in Foundation Bearing Capacity,"
ASTM, Materials Evaluation, May 1976. pp. 103-108.
13. McCabe, W. M., Koerner, R. M., and Lord, A. E., Jr., "Acoustic
Emission Behavior of Concrete Laboratory Specimens," American Concrete
Institute, ACI Jour., July 1976. pp. 367-371.
14. Koerner, R. M., Lord, A. E., and Deisher, J. N., "Acoustic Emission
Stress and Leak Monitoring to Prevent Spills from Buried Pipelines,"
Proc. 1976, Natl. Conf. in Control of Hazardous Material Spills, New
Orleans, La., April 25-28, 1976. pp. 8-15.
15. Koerner, R. M., Lord, A. E., Jr., McCabe, W. M., and Curran, J. W.,
Acoustic Emission Behavior of Granular Soils," Jour, of Geotechnical
Div., ASCE, Vol. 102, No. GT7, July 1976. pp. 761-773.
16. Koerner, R. M., Lord, A. E., Jr., and McCabe, W. M., "Acoustic
Emission Behavior of Cohesive Soils," Jour, of Geotechnical Engr.
Div., ASCE, Voli 103, No. GTS, August 1977. pp. 837-850.
17. Koerner, R. M., Lord, A. E., Jr., and McCabe, W. M., "Acoustic
Emission Monitoring of Soil Stability," Jour, of Geotechnical Engr.
Div., ASCE, Vol. 104, No. GT5, May 1978. pp. 571-582.
18. Lord, A. E., Jr., Curran, J. W., and Koerner, R. M., "A New Transducer
System for Determining Dynamic Mechanical Properties and Attenuation
in Soils," J. Acous. Soc. of Amer., Vol. 60, No. 2, August 1976.
pp. 517-520.
19. Lord, A. E., Jr., Diesher, J. N., and Koerner, R. M., "Attenuation of
Elastic Waves in Pipelines as Applied to Acoustic Emission Leak
Detection," ASNT, Materials Evaluation, Vol. 35, No. 11, November
1977. pp. 49-54 and Proc. of 1977 ASNT Conf. in Phoenix, Arizona.
20. Koerner, R. M., Lord, A. E., Jr., and McCabe, W. M., "Acoustic
Emission (Microseismic) Monitoring of Earth Dams," Conf. Proc. of the
Evaluation of Dam Safety, Engr. Fdtn. Conf., Pacific Grove, Calif.,
December 1976. pp. 274-291.
21. Koerner, R. M., Lord, A. E., and Deisher, J. N., "Acoustic Emission
Detection of Underground Gasoline Storage Tank Leaks," Proc. ASNT
Conf. in Phoenix, Arizona, March 28-30, 1977.
87
-------
22. Koerner, R. M., and Lord, A. E., Jr., "Acoustic Emission Response of
Coal and Charcoal Briquettes," 15th Biennial Conf. of the Inst. for
Briquetting and Agglomeration, Vol. 15, August 1977.
23. Koerner, R. M., McCabe, W. M., and Lord, A. E., Jr., "Advances in
Acoustic Emission Monitoring," Vol. 30, No. 10, Water Power and Dam
Construction, London, October 1978. pp. 38-41.
24. Lord, A. E., Jr., and Koerner, R. M., "Acoustic Emission Generation in
Soil Masses," invited paper for Acoustic Societies of America and
Japan Conference in Hawaii, November 1978 (in Conference Proceedings).
25. Koerner, R. M., and Lord, A. E., Jr., "Predicting Dam Failure,"
Research Direction, Vol. 1, No. 1, Winter 1978, Drexel University.
pp. 1-4.
26. Koerner, R. M., Lord, A. E., Jr., and McCabe, W. M., "The Challenge of
Field Monitoring of Soil Structures Using A. E. Methods," Second Conf.
on Acoustic Emission/Microseismic Activity in Geologic Structures and
Materials, The Pennsylvania State University, November 13-15, 1978 (in
Conference Proceedings), pp. 275-290=
27. McCabe, W. M., "Acoustic Emission in Coal: A Laboratory Study,"
Second Conf. on Acoustic Emission/Microseismic Activity in Geologic
Structures and Materials, The Pennsylvania State University, November
13-15, 1978 (in Conference Proceedings), pp. 35-54.
26. K.oerner, k. M., ''Uverview of A. t. Monitoring OT Kock structures," bth
Proceedings of Philadelphia Section, ASCE Geotechnical Group, 1979.
29. Lord, A. E., Jr., and Koerner, R. M., "On the Magnitude of Acoustic
Emissions in Soil and/or Rock," Geotechnical Engineering Division,
ASCE, August 1979. pp. 1249-1253.
30. McCabe, W. M., and Koerner, R. M., "Acoustic Emission (Microseistnic)
Monitoring for Ground Control in Tunnels," Presented at Rapid
Excavation and Tunneling Conference, June 18-21, 1979, Atlanta, Georgia
(in Conference Proceedings).
88
-------
APPENDIX B
SPILL ALERT DEVICE USERS MANUAL*
Theory of Operation
The system described in the following pages and shown in Figure B-l
is designed to sense very small vibrations that occur within the soil mass
and that are propagated along the steel rod (wave guide) to the ground
surface. The accelerometer is attached to the wave guide, where it converts
the vibrations into electrical impulses that are amplified by the Columbia
model VM-103 vibration meter and then counted by the Hewlett-Packard 5300
series counter/timer. The preparation and actual use of this equipment
during the monitoring process are outlined below.
Equipment Preparation
The accelerometer must be connected to the wave guide through a
small brass fitting called a coupler (Figure B-2). The coupler has male
threads at each end, the larger end (5/16-NC thread) of which must be
screwed into the end of the wave guide protruding from the ground. This
rnnnprtinn should be made tightly with pips wrenches for best reSullb. Trie — •«-
accelerometer is then threaded just barely hand-tight onto the smaller end
(10/32-NF thread). Caution: over-tightening the accelerometer may cause
damage.
Use thin coaxial cables with "Microdot" connectors on each end to
join the accelerometer to the top of the VM-103. To connect the VM-103
amplifier to the Hewlett-Packard counter, insert the banana-plug end of the
special coaxial cable into the VM-103 and attach the BNC connector end to
INPUT A (lower connector) on the counter panel. The initial setting on
these instruments should be as follows:
* Prepared by W. Martin McCabe and Robert M. Koerner, Department of Civil
Engineering, Drexel University.
89
-------
Hewlett-Packard counter
VM-103 amplifier
Figure B-l. Photographs of spill alert device components.
9Q
-------
VM-103
HEWLETT-PACKARD COUNTER
Turn range dial fully counter
clockwise (O.Olg).
Push in ACC button.
After turning on, push BAT TEST
swtich up. If red light below
switch goes on, batteries are good.
If not, remove back of VM-103 and
replace the two 9-volt batteries.
These batteries can be obtained at
any radio or electronics supply
shop.
Monitoring Operation
Turn battery pack switch to BATTERY
when operating in field, and to
CHARGE when recharging batteries.
Turn COM-SEP-CHK switch to SEP.
Turn ATTEN switches to xl.
Turn AC/DC switches to AC_.
Turn SLOPE switches to +/-.
Turn LEVEL dials to PSET.
Turn inner mode control dial to
OPEN/CLOSE A and outer DELAY dial
fully counterclockwise.
Turn on by turning OFF-SAMPLE
RATE-HOLD dial just far enough
clockwise to produce a click.
Push OPEN/CLOSE button to activate
counting operation (small red £
will appear to right of the
once more to deactivate the
counting operation (£ will
disappear).
Once all the connections have been made at the desired monitoring
location, place the instruments firmly on the ground and do not touch
unless to re-zero the display or change the sensitivity settings. If the
wind is blowing moderately, place a bucket or box over the top of the wave
guide and accelerometer (or wrap the components in flexible polyurethane
foam sheet padding) to prevent wind-induced noise from being counted. Turn
all instruments on, activate counting, and wait 2 to 3 min for warm-up.
The instruments at this point are on the most sensitive settings. If
counts are registering continuously after warm-up, the sensitivity must be
reduced. This step is accomplished by turning the RANGE dial on the VM-103
one click clockwise (from .Olg to .03g). If continuous counting is still
observed, turn the RANGE dial to its original setting (.Olg) and push the
ATTEN switch on the counter to X10. If continuous counting still occurs,
recheck all the connections or wait for a quieter period of the day.
When a sensitivity setting is found that produces only intermittent
counts or no counting at all for 1/2 min, zero the display, activate the
91
-------
COUPLER
FOR CONNECTING ACCELEROMETER
TO WAVE GUIDE
r-10/32 NF
0.5cm
3.0cm
1.5 cm
5/16 NC
STEEL WAVE GUIDE
WAVE GUIDE
SEGMENT
CONNECTION
sin, 1.5cm
1.27cm did.
COLD ROLLED
STEEL
5/16 NC
Figure B-2. Details of wave guides used in acoustic emission monitoring.
Wave guide is driven in 4 ft segments and extended with threaded
connections to the desired depth or refusal. (1 ft = 0.3048m)
92
-------
counting, and record the number of counts at 1-min intervals for 5 to 15
min. With little or no wind and otherwise quiet environmental conditions,
the most sensitive initial settings should be satisfactory. Such conditions
are usually achieved early in the morning. Remain as motionless as possible
during the monitoring period. This sequence should be repeated at each
monitoring location.
All instrument settings and counting data should be entered on the
monitoring sheets provided, one for each location. Any reading believed to
have been caused or affected by environmental conditions (gust of wind, low
flying airplane, movement of yourself or nearby bird or animal, etc.) should
be so marked on the monitoring sheet in the COMMENTS column. This observa-
tion is very important and should be strictly attended to. A sample
monitoring sheet is shown in Figure B-3.
Please note that when fully charged, the battery pack in the
Hewlett-Packard counter will provide approximately 5 hr of continuous field
use. When the batteries have fully discharged, the LOW BATTERY light will
come on and the pack must be recharged for a full 18 to 24 hr. CAUTION:
do not exceed 24 hr of charge. To charge, attach AC power cord to back of
counter chassis, plug into 110-volt outlet, switch battery pack to CHARGE,
and leave counter OFF. A battery-use tag is attached to the battery pack
and should be marked after each use to determine when recharge is necessary.
93
-------
ACOUSTIC EMISSIONS MONITORING POINT
Location:
Depth:
Date
Time
A.E.
A.E.
Rate
Amp
Counter
Notes
Figure B-3. Sample monitoring sheet.
94
-------
APPENDIX C
APPLICATION OF ACOUSTIC EMISSION MONITORING IN SEEPAGE
Sowers (1) suggests the 40% of all earth dam failures are caused by
seepage in whole or in part. Seepage can be of the controlled variety, as
analyzed in flow net studies by use of Laplace's Equation, or can lead
progressively to high-velocity flow, then piping, collapse of soil arches,
and subsequent seepage failure. Seepage flows also occur around outlet
pipes placed beneath the dam for control of the upstream reservoir level or
through holes initiated by burrowing animals. Whatever the source, the
fact remains that water does indeed flow through soil voids and that this
flow may be an emittive phenomenon. The first potential application of
acoustic emission monitoring of seepage was site No. 15, which is cited in
Table 5 of the text. The problem was brought to our attention by a site
developer who was losing water from a lake created by a small earth dam.
The dam had a maximum height of 3.6 m (12 ft) and was approximately 370 m
(1,200 ft) long. While grouting was the obvious solution to the problem,
the cost involved in grouting the entire length off the dam was prohibitive.
Thus a series of borings was made along the axis of the dam, and seepage
tests were conducted with the results shown in Figure C-l. The results
indicated that the 62-m (200-ft) section between borings B-3 and B-4 seemed
most iiKeiy to oe the major contributor to LJie loss of water.
Since open borings were available, acoustic emission monitoring was
also attempted. However, the plastic casing of the boring could not conduct
emissions and was not therefore suitable as a wave guide. Instead, a heavy
steel wire was inserted down to the bottom of the borehole where the seepage
was presumably occurring. Acoustic emission count rates were recorded, and
the AE counts per minute were also plotted (Figure C-l). The general agree-
ment between seepage and acoustic emission activity in the zone from B-3 to
B-4 should be noted. The actual mechanism causing the emissions is not
known (perhaps it was the turbulent flow of the seepage against and around
the casing), but the use of the acoustic emission technique in monitoring
for seepage seems to hold great promise.
Current efforts in evaluating acoustic emissions emanating from
seepage flow are being directed at laboratory studies for the following
reasons: to understand the basic phenomena involved; to help determine
whether an acoustic emission detected in an embankment or earth dam is due
to soil movement alone or to some combination of soil movement and seepage;
and to determine the location and extent of subsurface leaks from reser-
voirs, lagoons, deep-well pumping, pipeline leakage, etc.
95
-------
A seepage apparatus was constructed to systematically vary soil
types and seepage velocity and to examine the tendency to cause emission.
The apparatus (Figure C-2) is a plastic cylinder 20 cm (8 in.) in diameter
into which 46 cm (18 in.) of soil is placed. Water is introduced at the
bottom of the column of soil through a perforated base plate and flows
upward where it is collected and measured at the top. The velocity of
fluid flow is controlled by regulating the pressure under which the fluid
is introduced. Forcing the flow upward ensures complete saturation and a
more uniform velocity profile. The wave guide is a shortened version of
the actual 12-mm (0.5-in.) steel wave guides used in the field and has been
cut into two segments. The first segment extends from the exterior through
the plastic wall and terminates just inside the column. This short segment
is threaded to receive an accelerometer on the exterior end, and an 18-cm
(7-in.) extension on the interior end. The longer segment is embedded in
the soil column and serves as the primary receiver for acoustic emission
signals.
Each test must be run at least twice. During the first run, only
the short wave guide segment is in place. The resulting acoustic emission
rate (counts/sec) represents boundary effects and extraneous noise. The
soil is then removed, the wave guide extension is inserted,' the soil is
replaced at the same density as before, and the test is repeated at the
same flow velocity. The difference in acoustic emission rates registered
for the two tests is that rate transmitted by the longer extension alone,
exclusive of any boundary noises.
The soil being tested is Ottawa sand with an in-place density of
1 7f) n/rm3 (1Q6 Ib/ft^J and 2 corresponding void ratio of C.55. Tht
acoustic emission count rate as a function of the velocity of water flowing
through the voids is given in Figure C-3. There is a general tendency of
increasing emission rate (of both noise and seepage) with increasing water
flow velocity, but considerable scatter exists. Action of these seepage
tests is under way using acoustic emission filters and different pickup
transducers sensitive to higher frequencies than those used in tests
described. Additional information will be published.
96
-------
VO
\
UJ
C£
u-1
SEEPAGE STUDY AT SITE 15
ACOUSTIC EMISSION RATE
0
50 100 150 200
DISTANCE ALONG DAM AXIS (m)
250
11
10
9
8 1
BORING
3 6
NUMBER
2
5 A
7
Figure C-l. Flow rates and acous:ic emission rates compared
for seepage study at site No. 15.
-------
GRADUATED
CYLINDER "
i
SOflL
COLUMN
S! EPAGE
SAND AND GRAVEL
FILTER
OVERFLOW
CATCH BASIN
WAVE GUIDE
TO MONITORING
EQUIPMENT
\-ACCELEROMETER
PERFORATED
BOTTOM PLATE
WATER UNDER
PRESSURE
INLET
Figure C-2. Experimental setup for study of acoustic
emission results from soil void seepage.
-------
vo
3.0
8
LU
o:
z
o
*-I
t/)
I/)
LU
ro
o
0
O.U
LEGEND
WAVE GUIDE ON TANK
o WAVE GUDE IN SOIL
REPRESENTS ACOUSTIC
EMISSION LEVEL FROM
SEEPAGE
«
NOISE
0.16 0.18 O..ZO 0.22 0.24
FLOW VELOCITY (cm/s)
0.26
0.28
Figure C-3. Acoustic emission rates for flow of water
through a colomn of Ottawa sand.
-------
APPENDIX D
APPLICATION OF ACOUSTIC EMISSION MONITORING IN PIPELINES
A critical feature of the acoustic emission technique in monitoring
earth dams is a method of transmitting the emissions from their source within
the soil mass to the ground surface where they can be monitored. This trans-
mission is being accomplished by means of steel rods that in some instances
are 76 m (250 ft) long. The similarity of long steel rods to pipelines is
obvious, and thus an extension of the acoustic emission monitoring into pipe-
line leak detection seems natural. Pipeline leakage is indeed a serious
concern, for in 1974 there were 557 oil pipeline spill incidents reported by
the U.S. Coast Guard (57). These breaks resulted in an economic loss of 24
million liters (6.3 million gal) of oil, along with the attendant environ-
mental consequences. The trend toward installing more pipelines of larger
diameter only increases the need for leak detection systems. An equally
important area for the application of pipeline leak detection methods is the
rapidly growing number of chemical pipelines, which are mainly found in
internal plant systems. Actual data on flow rate and volume of material
transported are not so well quantified as with interstate petroleum systems
but are significant to this study, for stress corrosion problems are
abundant in this type of pipeline (58).
Various commercially available leak detectors and monitoring systems
are designed to detect leaks immediately after they occur or to detect pipe-
line cracks that would eventually lead to leaks. These systems can be cate-
gorized as follows: flow monitoring (quantity, rate); pressure monitoring
(drop, wave); acoustic methods (passive, impact); mobile systems (active,
magnetic flux); eddy current methods (inertia, probolog); and radioactive
methods.
Clearly, no one system will serve all situations, and thus it is
necessary to characterize each technique for its range of applicability.
This appendix concentrates entirely on acoustic methods and on the acoustic
emission technique (a passive method) in particular. In this technique, the
propagation of elastic waves along the pipeline is the basis of the method.
Just what type (mode) of elastic wave is generated in the pipeline is rarely
discussed by those using the technique. Furthermore, the attenuation
(damping) and velocity of the elastic wave as it traverses the pipeline is
usually not mentioned explicitly and is left for on-site experimentation
(which is both expensive and time consuming). For more details, the reader
is directed to references 59 and 60. The object of the first phase of the
study was to monitor leaks from small pipe sections in the laboratory, and
that of the second phase was to field-monitor leaks and to investigate the
possibility of leak location by the acoustic emission signals.
100
-------
Though it is recognized that any type of leak in a pressurized pipe-
line is -undesirable and serious, many of these leaks do occur and often per-
sist for considerable periods of time. Such leaks range from small (hence,
nuisance type) to sufficiently large to warrant concern that crack propaga-
tion and pipeline rupture will eventually occur.
To evaluate the possibility of leak detection using acoustic emission
techniques, a 1.27-cm (1/2-in.) diameter hole was drilled near the center of
a 1.2-m (4-ft) long, closed-end, 15.2-cm (6-in.) diameter steel pipe. Into
this hole was placed a plug containing a small hole carefully drilled to a
known diameter. Several such plugs were fabricated, with hole diameters
ranging from 0.33 to 1.98 mm (0.013 to 0.078 in.). The holes were tempor-
arily capped, and the pipeline was pressurized using air, water, and light
machine oil separately as the internal pipe media. The acoustic emission
monitoring scheme was essentially the same as that shown in Figure 1 (text),
except that the pickup accelerometer used for this set of tests had a flat
frequency response from a few Hz to 10 kHz. Using this system, it was
possible to detect accelerations as low as O.Olg.
When sufficient pressure was reached, usually 1380 kN/m (200 psi),
the temporary cap was removed from the leak, and pressure and acoustic
emission data were recorded as the material escaped. Figures D-l, D-2, and
D-3 show these results for air, water, and oil as the escaping medium. A
number of tests were run using a variety of hole sizes. For the air response
(Figure D-l), an acoustic emission rate was monitored (i.e., counts/sec), but
as a result of the rapid loss of pressure in the water and oil tests (Figures
D-2 and D-3), a cumulative acoustic emission count was recorded. From these
response curves, a series of nhservations can be made: .. -
1. Greater internal pipe pressures cause greater acoustic
emissions to occur.
2. In all cases, the acoustic emission response is approximately
linear (each curve is the average of about five tests).
3. At a given internal pressure, the larger the hole size the
greater is the acoustic emission response.
4. From these data, it appears that air is more emittive than
water, which in turn is more emittive than oil. This conclusion
is reasonable because the acoustic emissions are in reality
noise created by the escaping medium within the pipeline.
Because this is a friction phenomenon, it seems reasonable for
the liquids to act as lubricants (the oil more than the water),
which has the effect of diminishing the emission levels.
The first field study was made on an insulated 7.6-cm (3-in.) steam
line that had a constant source leak in the packing of a valve stem. The de-
tection system consisted of an accelerometer attached to a 6.3-mm (1/4-in.)
diameter, 30-cm (12-in.) long steel rod wave guide that was pushed through
the insulation making firm contact with the pipe being monitored. The
response of the pickup transducer was resonant at 5 kHz and was attached to
an amplifier and then to an electronic counter.
101
-------
10
o
ro
O
O
O
JQ
3
O
O
LU
cc
*
H-t
to
to
I-H
2:
LU
o 2
0
200 400 600 800
INTERNAL PIPE PRESSURE (kN/m2)
1000
Figure D-l. Acoustic emission count rate versus internal pipe pressure for
air leaking from 15.2 cm <6-in) diameter pipe.
-------
20
16
CD 12
o
(/) o
i— i O
Z
LU
O
O
o
0
^00
600 800 1000
INTERNAL PIPE PRESSURE
1200
(kN/m2)
1AOO
Figure D-2 Acoustic emission counts versus internal pipe pressure
for water leaking from 15.2 cm (6-in) diameter pipe.
-------
3.5
-3.0
25
o
in
co
2.0
I_U
a 15
< 1.0
Q5
1.194mm'
0 400 800 1200
INTERNAL PIPE PRESSURE (kN/m2)
Figure D-3. Acoustic emission counts versus internal pipe pressure
for oil leaking from 15.2 cm (6-in) diameter pipe.
104
-------
A series of readings was taken at successive distances from the leak.
These readings were obtained from the amplifier (for signal level in
acceleration units, i.e., "g's"), and from the counter (for acoustic
emission data in counts per second). Figure D-4 shows these results, where
the amplitude response is seen to be approximately linear and the acoustic
emission response is exponential. It is significant that the leak signal
becomes indistinguishable from the background noise at this particular site
beyond approximately 30 m (100 ft).
The second field study was conducted at the same site on a similar
steam pipe that functioned as a pulsating bleeder line. The pulse was set
at 15-sec intervals, and the leak lasted for a 5-sec duration. Instrumenta-
tion was the same as with the first field study. Data were taken on both
sides of the leak. Although signal amplitude levels were obtained, they
were not so sensitive as the acoustic emission rate readings taken from the
counter. These latter data are plotted in Figure D-5, where the response
on each side of the leak is seen to agree and is approximately exponential.
The signal exists in a relatively strong state for 30 m (100 ft), becomes
weaker in the next 15 m (50 ft). The background noise at this site was
approximately 50 to 55 dB as registered on a sound level meter. It should
also be mentioned that the pulsing leak was located between straight pipe
on the north side and pipe with a series of five bends on the south side.
Thus, the significance of pipe bends seems to be nominal as far as signal
attenuation is concerned.
Use of Equation 1 allows a graphical determination of the location
of the actual leak source. The attenuation coefficient can be computed
fruiVi tiic follow my furiuula.
(1)
a =
where a = attenuation coefficient in dB/distance, x = distance between
measured wave peaks, AQ = initial wave amplitude, and AT = subsequent
wave amplitude.
The computation is done by plotting the log (AQ/AI) response
against distance from some arbitrary field location data (see Figure D-6).
The points fall in two straight lines that intersect at the approximate
location of the leak. Furthermore, the slope of the lines gives an average
attenuation coefficient of 0.98 dB/m (0.30 dB/ft), which agrees with
published laboratory test data (57).
Throughout the field monitoring phase of the project, the signifi-
cance of background noise (i.e., background vibration levels) cannot be
overstated. When such ambient noise levels are great with respect to the
signal levels being monitored, the technique is quite limited in its
application. For the site described here, sound level readings of 50 to 55
dB (A-scale) were measured to give a general idea of background noise. At
a more remote site, where there would be less background noise, greater
equipment sensitivity could be used, thereby increasing the distances over
which this type of monitoring could be utilized.
105
-------
S.06g
o
LU
ce
|.02g
t—I
Sj
<
5 10 15 20 25 30
DISTANCE FROM LEAK SOURCE (m)
35
a
5 10 15 20 25 30
DISTANCE FROM LEAK SOURCE (m)
Figure D-4. Field results of signal amplitude and acoustic emission
count rate for a constant-source leak in a 7.6 cm (3-in)
diameter pipeline as a function of distance from the leak.
106
-------
LEGEND
• NORTH OF LEAK - STRAIGHT PIPE
O SOUTH OF LEAK-FIVE BENDS
0
10 20 30
DISTANCE FROM LEAK SOURCE (m)
50
Figure D-5. Field results of acoustic emission count rate for a
pulsating leak in a 7.6 cm (3-in) diameter pipeline
as a function of distance from the leak and on both
sides of the leak.
-------
2.5,
20
o
00
o"
3
1.0
0.5
0
\x
ACTUAL LEAK AT
FROM REFERENCE
10 20 30 40 50 60 70
DISTANCE FROM ARBITRARY REFERENCE LOCATION (m)
80
Figure D-6. Data from Figure D-5 replotted to illustrate the method of leak
source location using the acoustic emission monitoring technique.
-------
APPENDIX E
APPLICATION OF ACOUSTIC EMISSION MONITORING IN CONCRETE
When one considers a material that owes part of its strength to a
noise-emitting, frictional-based composite and a natural, in-place, wave-
carrying steel rod system, and one can readily visualize that reinforced
concrete is an appropriate candidate for the acoustic emission monitoring
technique. By attaching pickup sensors to the reinforcing bars (which are
embedded only a few inches inside the concrete), a large zone of concrete
can be monitored at a single pickup station. Insofar as the emittive
nature of concrete itself is concerned, many mechanisms could interact.
Possible acoustic emission mechanisms in plain concrete as a function of
stress state are as follows:
Pure compression:
crushing of matrix
crushing of aggregate
Pure tension:
bond breaking in matrix
hnnH h»"paif i on betwee" fn?.trix 2nd aggregate
Shear or torsion:
sliding friction
rolling friction
crushing of matrix
crushing of aggregate
bond breaking in matrix
bond breaking between matrix and aggregate
Prior acoustic emission work in this area of concrete monitoring (61)
gives an overview of the past work, and we have attempted to supplement
these efforts. Of particular interest is that the Kaiser effect ("memory"
effect) seems applicable to concrete (Figure E-l), that aging of concrete
can be monitored (Figure E-2), and that sustained load sequences can be an-
alyzed (Figure E-3). In Figure E-3, the cumulative acoustic emissions in-
creased rapidly as the load came closer to the ultimate failure load.
Figure E-4 presents additional information from the same test sequence, but
now the rate of emissions is plotted against time on a log-log scale. The
figure shows that acoustic emissions are still being generated at a rate of
about 100 counts/min after 80 hr of loading. This fact is significant,
since field testing of concrete structures (as in geologic and metal
structures) will require use of acoustic emission rates as opposed to cumu-
lative acoustic emission counts to assess structural stability. The linear
response of these creep tests on a log-log scale is of great fundamental
109
-------
interest. The three curves have nearly identical slopes of approximately
-1.0. Whether or not this behavior is a property of the concrete and its
respective load state remains a subject for future research. Note that
this sequence of tests (and the following series) was made using a trans-
ducer resonant at 175 kHz, a filter band width from 125 to 250 kHz, and a
total gain of approximately 80 dB.
The next series of tests was performed on 10.2 x 15.2 x 76.2-cm (4 x
6 x 30-in.) beams in a flexure mode (third point loading). Figure E-5 shows
the results with the transducer mounted on the compression face (single
test) and on the tension face (average of three tests). In comparison, the
curves are related to the total failure load. Tension, however, was ob-
served to be the governing failure mode in the unreinforced beams. Note
that the acoustic emissions detected at the tension face were considerably
more numerous than those detected at the compression face at all stress
levels up to tensile failure. Furthermore, above 85 percent of fracture
load (tension), the acoustic emissions begin to increase rapidly. This
increase is significant because it indicates that the technique may be
valuable as a good precursive indicator of failure in a field monitoring
scheme.
The previous experimental work dealt with plain concrete, which was
seen to be emittive while in a deforming state. Future work will be
directed at reinforced concrete, which is more typical of actual field
structures. Initial tests will be with laboratory-sized beams in flexure,
and subsequent work will be conducted on large-scale members and actual
field structures.
110
-------
100r
0
10 15 20 25 30
ACOUSTIC EMISSION COUNTS (x103)
35
Figure E-l. Load versus acoustic emission response of 3-day-old concrete
specimens showing effect ol load cycling on acoustic emissions.
-------
30r
i
8
2
O
CO
20
15
O
i—i
CO
o
10
A
100% FRACTURE LO/*D
85% FRACTURE LOAD
50% FRACTURE LOAD
0
8 12 16 20 24
CONCRETE AGE (days)
28
32
Figure E-2. Acoustic emission response of concrete cylinders
as a function of aga (curing time) at various
percentages of ultimate fracture load.
-------
u>
0
84V, FRACTURE LOAD
6AV. FRACTURE LOAD
48 V. FRACTURE LOAD
FRACTURE: LOAD
20
60 80 100 120 UO
TIME (min,)
160
Figure E-3. Acoustic emission versus time response for creep
tests (sustained-load tests) at various percentages
of ultimate failure load.
-------
100,000
LU
!< 1000
on
LU
o
t—t
8
<
100
10
TTTII I I I I UN I I I I 11
I I IIHJ
FRACTURE LOAD
27V.
L_ FRACTURE V
r LOAD
0.1
I I I I I 14.
1 ml
8V. FRACTURE LOAD
1 1 Hi
10
100
1000
(WIN,)
Figure E-4. Acoustic emission rate versus time response for creep tests
(sustained-load tests) at various percentages of ultimate
failure load over long-term monitoring.
114
-------
COMPRESSION
MONITORING
H!
TENSION
MONITORING
50 100 1EO 200 250
ACOUSTIC EMISSION COUNTS (x103)
300
350
Figure E-5. Load versus acoustic emission response of concrete beams tested
in three-point loading tests (flexure tests) with transducer
mounted either on the compression face or on the tension face.
-------
APPENDIX F
GLOSSARY
P-Waves - A P-wave is a longitudional elastic wave (primary wave).
S-Waves - An S-wave is a shear elastic wave (secondary wave).
R-Waves - An R-wave is a Rayleigh surface elastic wave.
Angle of shear resistance - the same as the friction angle, fy (q.v.).
C/'s - stresses:
O"l s major principal stress
#"2 = intermediate principal stress
<3"3 - minor principal stress
-------
t (Mtau") - c + )
where "tau" = shear stress
c = cohesion (kN/m2) (may equal zero in sand)
0^,= normal stress on shear plane (kN/m2)
4> = friction angle
Isostatic compression tests - the external stresses (cf],^* and CTg
are all equal.
-------
TECHNICAL REPORT DATA
/Please read Jtutncnons on the reverse before completing)
1. REPORT MO.
2.
4 T1TLS AND SUBTITLE
Spill Alert Device for Earth Dam Failure
WSrning
3. RECIPIENT'S ACCESSlOfWNO.
S. REPORT DATE
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION R6r>C*~ N
Robert M. Koerner
Arthur E. Lord. Jr.
9. 'SRPORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
Drexel University
Philadelphia, PA 19104
11. CONTRACTVGHANT NO.
R-802511
12. SPONSORING AGENCY NAMi AND ADDRESS
Municipal Environmental Research Laboratory—Cin, OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
13. TYPE Of REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
Project Officer: John E. Brugger (201)321-6634
A spill alert device for determining earth dam safety based on the monitoring of the
acoustic emissions generated in a deforming soil mass was developed and field tested.-
The duuublic enribbiurtb are related to the basic mechanisms rrom wnicn soils derive tneir .—v,
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lOSNTI PIERS/OPEN ENDED TERMS
c. COS ATI Field/Group
iIBSJTION STATeMcr
19. SECURITY CLASS I This Report/
Unclassified
21. NO. OP PAGES
Release to Public
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
------- |