;
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
Technology Transfer
Capsule Report
Acoustic monitoring
To Determine the
Integrity of Hazardous
Waste Dams
*»-
'•'ll'--
.-.; _- _ ,.:.,'.-.-.---
-------
-------
Technology Transfer
EPA 625/2-79-024
Capsule Report
Acoustic Monitoring
To Determin
the
Integrity of Hazardous
Waste Dams;
August 1 979
This report was developed b/ the
Industrial Environmental Research Laboratory
Cincinnati OH 45268
-------
Acoustic monitoring system: waveguide, accelerometer, amplifier, and counter
-------
1. Significance
There are as many as 500,000 diked
areas in the Unr ed States contain-
ing potentially hazardous wastes.
This total includ3s small waste
ponds at minor chemical manufac-
turing plants as
tailings lagoons
and phosphoric
The U.S. Enviror
Agency's (EPA's
/veil as mile-square
at mines, smelters,
acid plants.
mental Protection
Office of Research
and Development has long been
aware of the large number of
marginally safe impoundments. A
seemingly secur
3 dam may suddenly
fail with provocction as slight
as a heavy rain.
leaks and spills
dam waste pones is not known.
because many s
rhe full extent of
rom small, earthen-
jch spills go
unreported or may be known only
locally. The ovenall damage to the
environment, however, is probably
great.
Those spills resulting from earth
dam failure that|have been reported
often have had serious environ-
mental impact, for example:
• In 1 967, an alkaline waste lagoon
failed, sending 400 acre-feet
(493,000 m3}|of fly ash into
the Clinch River in Virginia. These
wastes traveled downstream to
Norris Lake, Tennessee, where it
has been estimated that 21 6,000
fish were killed.
• A few years later, another
lagoon was responsible for
seriously contaminating the
James River with kepone when
the restraining dam gave way.
• In 1976, in Oswego, New York,
the failure of a lagoon 100
feet X 130 feet X 1 5 feet (30
meters X 40 meters X 5 meters)
holding chemical manufacturing
wastes resulted in the material
ending up in Lake Ontario.
Many similar spills have been
reported in recent years. Releases
of contaminants such as the fore-
going pollute waterways, kill
aquatic species] adversely impact
drinking water systems, and
despoil scenic areas. Even small
spills from dikes that hold hazardous
wastes can have long-lasting
environmental impact.
It is also necessary to consider
the danger to human life posed by
marginally stable tailings dams
that hold back large quantities of
liquid refuse. Failure of such
dams is much less frequent, but
the consequences are far more
immediate. When danger of failure
is increasing, some means of pro-
viding warning is badly needed.
Because the problem is extensive,
involving thousands of small
facilities, an inexpensive and
simple technique for monitoring
the stability of earthen dams was
desired. It seemed unlikely that
conventional methods of inspection,
which require expensive instrumen-
tation and trained geotechnical
engineers, could meet the need.
Under this impetus, a system—
acoustic emission monitoring—was
developed based on the phenome-
non that soils emit sounds under
stress.
The EPA's Oil and Hazardous
Materials Spills Branch, in Edison,
New Jersey, awarded a grant to
Drexel University to develop
an acoustic emission monitoring
system to determine the stability of
dams. The Drexel investigators
found that the resulting signals, when
properly amplified and quantified,
can be a valuable guide in evaluating
the stability of earthen dams.
-------
2. Theory
Small waste dams are constructed of
a variety of soils, as a rule using
the materials at hand. The resulting
structure may be susceptible
to failure by one or more mechan-
isms. The high cost of conventional
methods of monitoring dam
stability usually rules them out
as candidates for evaluating failure
potential of small waste darns.
Shear Strength of Soil
The acoustic emission mechanisms
within soil are related closely
to the mechanisms of shearstrength.
The shear strength of soil is the
resistance to deformation when a
tangential (shear) stress is
experienced. This shear strength
is made up of:
• The structural resistance pro-
vided by interlocking of the soil
particles (Figure 1)
» The frictional resistance (rolling
and sliding) at the contact
points of the soil particles (Figure 1)
« True cohesion between the
surfaces of the particles resulting
from intermolecular attraction
(important in clays)
• Apparent cohesion caused by
capillary forces at suitable water
content
The frictional resistance to failure
increases with an increase in
the normal force on the shearing
plane (Figure 1). Only the normal
force supported by the solid soil
skeleton, however, results in
frictional resistance. The part of the
normal force that is supported by
the pore water pressure reduces the
friction component of the soil
strength.
Failure Mechanisms in Dams
Earthen dams consist of an embank-
ment (with an upstream and down-
stream slope) and a base foundation.
The point at which the downstream
slope meets the base is termed
the toe of the dam. If the dam
is built across an existing valley,
the intersection of the dam and
the valley slope (abutment) is termed
the groin of the dam. The stability
of a slope is somewhat indeter-
minable, inasmuch as the integrity
of a slope made of soil can never
be guaranteed because of changing
climatic conditions, varying
loads, and the activity of man in
the area. In particular, the possibility
of the soil becoming saturated
over time must be considered,
because pore water (water filling
the voids of the soil skeleton)
can have an important deleterious
effect on the strength of soils.
Dams that fail usually do so by one
or more of the following mechanisms:
• Groin failure (Figure 2a)
• Downstream slope failure above
or through the toe (Figure 2b, 2c)
• Base failure, if the rupture
surface is deep seated and passes
through the supporting soil
below the toe (Figure 2d)
• Seepage of such a nature
that soil particles are carried
away with the flow, and piping
(erosion by percolating water
resulting in conduits) develops
• Overtopping (impounded wastes
overflowing the embankment)
and subsequent erosion caused
by the fluid flow
• Upstream failure during draw-
down
Even if piping does not occur,
seepage may reduce slope strength
and contribute to failure, because
it brings fluid to the downstream
slope and thereby adds weight,
increases the load, and decreases
apparent cohesiveness and
effective normal force. Failure
through the base must be con-
sidered, particularly when the soil
beneath the dam is softer than the
slope-forming soil. In addition,
unless adequate freeboard
(embankment height above the
maximum expected water level) is
provided, failure can occur through
overtopping.
-------
P = force normal to the
shearing plane
S = shear force along the
shearing plane
<7 = normal stress = P/A
Figure 1.
Soil Shear Strength Components of Friction and Interlocking: (a) Shear in Granular Mass Showing Potential Particle
Movements and (b) Mechanisms of Resistance, Deformation, and Movement in Grains
Groin
Downstream slope
Abutment
Sliding
soil wedge
Base foundation
•Toe
(b)
Slope stability may be analyzed
based on measured strength
properties of the soil. The usual
procedure is to postulate a failure
surface, often taken to be the arc of
a circle. The total shear force
tending to slide the soil along the
failure surface is calculated. Then the
Downstream
slope
Sliding
soil wedge
Opposing
force
Firm base material
(d)
Figure 2.
Some Modes of Dam Failure: (a) GroinlFailure, (b) Slope Failure Above Toe, (c) Slope Failure Through Toe, and (d) Base
Failure (Deep-Seated Rupture Surface Below Toe)
resisting force i i the soil along
the surface is computed, usually
under the worst expected conditions,
for example, me
water pressure
ximum pore
a seismic event.
The ratio of resisting force to
the driving force is termed the
factor of safety. The surface having
the lowest factor of safety is
considered the most dangerous
rupture surface.
-------
30 i—
£ 20
te
10
STRAIN I
30
5 20
CO
CO
cc
1-
co
10
J_
I
100 200 300 400
ACOUSTIC EMISSION COUNTS
500
Figure 3.
Unconfined Compression Test Results for Undisturbed Sample of Silty Clay at 56-Percent Water Content
Monitoring Soil Stability
Conventional Methods. Soil engi-
neers and earth scientists use
a number of conventional devices
to determine the deformation and
strain of a mass of soil, including:
• Piezometers (devices to measure
the pore water pressure)
• Soil strain gages (linear
potentiometers measuring dis-
stance between reference anchor
plates)
• Slope inclinometers (pendulum-
actuated transducers lowered
down flexible near-vertical
tubes)
• Grids of stakes (monitored by
standard surveying methods)
• Settlement plates within the
soil mass (to determine
foundation settlement)
Each of these devices is expensive
to install, or requires consid-
erable expertise to monitor and
interpret, or both. Piezometers and
settlement plates must be installed
during construction or drill holes
must be provided. Piezometers
must be installed in large numbers
to provide a clear picture of the
pore water distribution.
Surface instruments require
extensive surveys to obtain data
or require maintenance when
monitoring over a long period of
time. Moreover, the data must be
interpreted by experts. Never-
theless, major dams are often so
instrumented because proper
instrumentation can provide warning
of potential failure. It has long
been recognized that warning
signals—for example, strain
discontinuities, pore-pressure
buildup, leakage, or an increase in
deformation rate—almost always
precede embankment failure,
except when the failure is caused by
an unanticipated catastrophic
event such as overtopping or an
earthquake. But, to observe the
signals, someone must be watching,
and until recently there has been
no economically viable method
of instrumenting the multitude of
small waste dams that exist
across the country.
Acoustic Emissions. Because of
the difficulty of implementing
each of the foregoing methods for
long-term monitoring of dam
stability, EPA's Office of Research
and Development sponsored studies
to develop an alternative moni-
toring technique. Acoustic emission
was a promising area. Acoustic
emission is the sound generated
internally by the strain resulting
within a material when a stress
is applied. The emission is the result
of the transient elastic energy
that is released when materials
undergo deformation, fracture, or both.
To apply the acoustic emission
technique to monitoring dam
soil stability, Drexel University
researchers have studied the acoustic
emissions produced by soils
under stress. They found that
curves of applied stress vs. cumu-
lative noise counts look very much
like curves of applied stress vs.
resulting soil strain, as illustrated in
Figure 3. These results contributed
to the conclusion that acoustic
emissions do, in fact, display
a one-to-one correspondence
with soil movement and instability.
This conclusion holds for granular,
sandy soils as well as for cohesive,
clayey soils, even though the
emission levels and mechanisms
involved vary with soil type.
Considerable data have been
obtained on emissions within soils.
-------
3. Acoustic Emission
Technology
by stress waves
Acoustic emissicns are generated
/ithin materials
during dynamic processes. The
particular dynam c process may be
the result of an externally applied
stress, or it may be the result of
some other unstable situation. At
present there is no proven theory for
the actual mechanism of acoustic
emission in any material, even
though such emissions are quite
common. Some examples are the
cracking of wood when it is over-
stressed, the "crying" of tin as it
is bent, the cracking of ice, and the
crunching of snow. A stress (force) is
applied, something gives (a strain
is produced), and energy is released
that appears partly as sound energy.
A large volume of information on
emissions from soils of various
types has been obtained by laboratory
studies using the triaxial shear
test apparatus (a
tory tool for eval
strength of soil)
standard labora-
jating the shear
with a transducer
(accelerometer) embedded in the
center of the soil sample being tested.
Some results ob
tests and by other means are
discussed in the
follow.
Sources of Emissions in Soils
It has long been
:ained from such
subsections that
realized that when
individual soil particles move with
respect to one another they pro-
duce a small, but nevertheless
detectable, noise. This noise,
which is designated acoustic emis-
sion, is generated within soils by the
mobilization of the same mech-
anisms, such as
cohesion, that a
for the shear stn
application of st
potential energy
cohesion are ovi
of this potential
friction and
•e responsible
mgth of soils. The
ess produces
As friction and
ircome, some
energy is converted
to other forms, including acoustic
energy.
The primary shear strength
mechanisms in granular soils
(Figure 1) include sliding friction,
rolling friction, and structural
resistance (overcome by particle
degradation). Friction between
the individual particles is the
predominant mechanism in granular
soils. Studies have shown that
conditions producing the greatest
number of interparticle and,
therefore, frictional contacts,
that is, well-graded uniform soil
conditions, also produce the greatest
amount of acoustic emission
activity. Cohesion of particles
also is a mechanism of the strength
of clay soils, but cohesion does
not provide as many emissions as
does friction.
Tension crack in highway stockpile,
induced during testing
-------
Cap
XL
Mass
Crystal used
in compression
Hemispherical spring
Imposed
vibration
Lead wire attached
to upper surface
of crystal
\
Metallic case (serves as
one of the electrodes
and lead wires)
Figure 6.
Typical Piezoelectric Accelerometer Construction
Laboratory Studies
The resistance to shear deformation
depends on different mechanisms
in different types of soil. Because
the mobilization of these mech-
anisms is the source of acoustic
emissions, it is important to know
the different signal strengths of the
different soils.
Figure 7 shows the results of tests
with three soils—two sands
and one clay—tested in triaxial
shear. The load was increased in
several increments until the soil
sample failed. After each load
increment was applied, both the
rate and the amplitude of the
emissions decreased with time as
relative stability was approached.
At each load increment, the
average signal strength was repre-
sented by summing the peak
voltages for all signals obtained
and dividing by the number of signals.
The following important obser-
vations were made:
• Both sands tested showed the
same general response.
• 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 clay response is markedly
different. Significant signals
are not emitted from clay until a
higher percent failure stress
is reached.
• Signal levels in clay are from
Vz to Ywo the level of signals
from the sands at corresponding
stress levels.
• Initially the signal level for
the clay increases, as with sands,
butthen 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 platelike 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.
There are other important features
of cohesive soils (clays and
clayey silts):
• Emissions decrease with increas-
ing 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.
• As is the case for granular
soils, stress vs. cumulative
emissions curves corresponds
closely to stress vs. strain
curves (Figure 3). Thus, acoustic
emissions are an indicator of
deformation.
• 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).
-------
40 60
PERCENT FAILURE STRESS
Figure 7.
Average Amplitude of Acoustic Emissions (Measured as Peak Signal
Voltage Output) for Various Soils as a Function of Percentage Failure Stress
in Triaxial Creep at 5-lb/in2 Confining Pressure
Field Studies
In addition to laboratory tests on
soils, 1 8 field installations (listed
in Table 1) have been monitored.
The table shows the considerable
range in size of dam, stability of
foundation, and quality of embank-
ment. Although the dams being
monitored are of several types,
most do hold back wastes of one
sort or another.
Site 14, an earthen stockpile
15 feet (4.6 meters) high, is of par-
ticular interest because it was
brought to failure as a planned
experiment while being monitored
by acoustic emission techniques.
Five successive cuts were made
in the bank, as shown in Figure 8,
before a large wedge of soil
separated; this result was con-
sidered to be bank failure.
Each response from the first four
cuts indicated a high acoustic
emission response initially, then an
approximately exponential decay
with time until relative stability
was reached. In general, the more
precarious and unstable the situation
became with each successive
cut, the greater the acoustic
emissions count rate experienced.
Approximately 30 minutes after the
fifth and last cut was made, the
acoustic emission rate rapidly
began to increase. When this count
rate reached its maximum, about
7,700. counts per minute, a large
section of soil pulled away from
the intact mass and slid down
the remaining slope. Thereafter, the
count began to subside and
eventually returned to a low level.
-------
Table 1.
Overview of Sites Being Monitored Using the Acoustic Emission Method
Site No. and
location
1 . PA
2. PA
3. NE
4. MD
5. PA
6. NE
7. PQ
8. DE
g. PA
10. NJ
1 1 . VA
12. NY
1 3. PA
14. PA
15. pA
1 6. TX . .
1 7. KY
18. DE
Purpose
Flood control
Recreation
Flood control
Ore stockpile
Stockpile for highway fill
Sludge and wastewater lagoons
Height
(ft)
30
66
67
40
6
68
95
15-40
120
8
4-15
8-20
15
15
12
150
13-28
25
Length
2,600 ft
2,500 ft
900 ft
300 ft
120 ft
600 ft
900 ft
6 mi
600 ft
500 ft
450 ft
20 ft
60 ft
1 200ft
2 mi
2 mi
1 000ft
Embankment
design and
construction
Excellent
Excellent
Excellent
Good
Good
Good
Poor
Excellent
Poor
Poor
Poor
Poor
Good
Good
Foundation
stability
Excellent
Excellent
Compressible
Poor
Poor
Good
Good
Good
Acoustic
emission
waveguides3
20 rods*
1 2 rods*
12 re-bars+
2 pipes*
1 pipe+
1 re-bar+
1 pile*
3 rods*
6 rods*
3 rods*
3 pipes-f-
1 1 rods*
12 rods*
4 rods*
6 rods*
1 rod*
4 rods*
8 rods*
(d)
8 rods*
3 rods*
Range of
acoustic
emission
count rate
(counts/min)
0
0
0-200
0-20
2-750
(b)
(b)
2-10
0-5
0-40
0-3
2-100
10-190
2-7 700C
,bi
( )
0-4
"Asterisk (*) = vertical; plus (+) = horizontal.
""Monitoring in process.
cHigh count occasioned by intentional destabilization.
dlnstallation in progress.
Failure wedge
To acoustic emissions
readout equipment
Tension crack
from cut No.
Cut No. 1
Figure 8.
Earth Stockpile Intentionally Brought to Failure
-------
Highway stockpile (Site 14) deliberately brought to destruction during testing
The effect of rain on the acoustic
emission count rate was seen
clearly during this experiment. Heavy
rainfall occurred twice, following
the third and the fourth cuts.
Both times the count rate increased
substantially and required about
a day to return to the former level.
It took longer for readjustment
of the slope back to equilibrium
after the rain of cut No. 4, possibly
because of the gradual decrease in
the slope's factor of safety.
Irrespective of the relative mag-
nitudes involved, it can be concluded
that the two rainfalls did have
an adverse effect on the slope's
stability, at least temporarily, and
that the destabilizing effect of
the additional water in the soil
was detected by acoustic emissions.
The experience
testing verifies •
gained from field
hat unstable
soil produces large quantities
of acoustic emissions. Conversely,
when the soil is stable, emission
rates are low or
nonexistent.
A full correlation between some
index of dam instability and the
emission count rate is not yet avail-
able; further experience with
field installations should provide
needed data. Nevertheless, a
qualitative guide has been devel-
oped for assessing the dangerofdam
failure based on acoustic emission
readings.
-------
5. Prospects and
Costs
Acoustic emission waveguides
have been installed at a number of
sites (see Table 1) with locations
as far west as Nebraska, north
to upper Quebec Province,
and south to Texas. The size of the
embankments or dams varies
greatly, with heights from 6 feet
(1.8 meters) to 150 feet (45.7
meters), and lengths from 20 feet
(6.1 meters) to 6 miles (10 km).
Acoustic monitoring systems have
been installed privately by
several corporations in the United
States that have waste chemical
lagoons, and a large chemical
manufacturer in England is in
the process of procuring the
components. At least three
American companies will now
supply ready-for-use systems.
How well has the system per-
formed? Some results are indicated
in the following:
• Failure of a suspect reservoir
wall was easily predicted
before the wall collapsed inward
during drawdown of the im-
pounded water.
• The sliding of an embankment
was predicted hours before
the actual collapse.
• A dangerously unstable
industrial dike, containing a
large quantity of dissolved toxic
heavy metals (lead, cadmium)
as a sludge, was reinforced in
time to prevent breaching.
• An abandoned lagoon at a
chemical waste disposal facility
was detected as in danger
of failure during a storm. Prompt
action in shoring up the wall
and adding more earth prevented
collapse.
Table 3.
Acoustic Emission Equipment
Supplier and address
Columbia Research Laboratories,
McDade Boulevard and
Bullens Lane
Woodlyn PA 1 9094
Equipment
Acceler-
ometer
Amplifier
Model
No. 476, nominal
resonance = 7.5 kHz
Vibration meter.
, Approxi-
mate
cost3
$1 75 (less
then 6),
. $155 (6
to 1 0)
$395
Hewlett-Packard Co Electronic
King of Prussia Industrial Park counting
King of Prussia PA 19406 system
Model VM-103
Measuring system, $500
5300A
Timer/counter, 5304A $385
Battery pack, 5310 $275
B&K Instruments, Inc. ...
5111 W. 164th Street
Cleveland OH 44142
Cable Coaxial Microdot Cable $2/ft
connectors AO-0037 (with JP-
0012 connectors)
Microdot to Microdot $3
connector, JJ-0032
a1978 prices.
-------
I
Not only has the system proven
effective, it is also inexpensive
and nearly maintenance free when
used periodically. Some suppliers
whose components have been
used in work to date are listed
in Table 3, with the components
and addresses of the suppliers.
Manufacturers having packaged
systems available for acoustic
emission monitoring of earth struc-
tures are:
Acoustic Emission Technology
Corporation
1828A Tribute Road
Sacramento CA 95815
Dunegan/Endevco
Rancho Viejo Road
San Juan Capistrano CA 92675
Geotechnical Instruments, Ltd.
Geotechnical House
Hatton, Warwick CV357JL
England
Weston Geophysical Engineers, Inc.
Post Office Box 550
Westboro MA 01 581
Typical waste pond retained by an earthen dam
Packaged systems intended to be
used in the portable periodic
sampling mode, handling a
single channel of information,
are available in the $2,000 to
$4,000 price range. A microcom-
puter-based system for continuous
operation with eight-channel
output, programmable alarm level,
and camera for hard copy is
available at about $10,000. More
sophisticated warning systems
use a minicomputer and can scan
an unlimited number of channels
continuously. The cost of such
Manufacturers should be contacted
for details concerning their
systems. Each of the companies
will provide advice and assistance
in the installation of an acoustic
emission monitoring system.
systems is on th
5 order of $25,000.
-------
This report was prepared for the U.S. Environmental Protection Agency
by the Centec Corporation, Reston VA. Dr. John E. Brugger of the
EPA Industrial Environmental Research Laboratory's Oil and Hazardous
Materials Spills Branch provided assistance and reviewed the report. Photo-
graphs were provided by the Centec Corporation, Drexel University, and
Mason & Hanger-Silas Mason Co., Inc., of Edison NJ.
Additional information or reference material may be requested from EPA
or from Drexel University. The Drexel University contact is:
Robert M. Koerner, Ph.D.
or
Arthur E. Lord, Ph.D.
Department of Civil Engineering
Drexel University
Philadelphia PA 19104
EPA's Oil and Hazardous Materials Spills Branch can provide information
on acoustic emission monitoring and other programs dealing with spill
prevention. Comments or questions should be addressed to:
Ira Wilder, Chief
Oil and Hazardous Spills Branch
Resource Extraction and Handling Division
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Edison NJ 0881 7
This report has been reviewed by the Industrial Environmental Research
Laboratory, U.S. Environmental Protection Agency, Edison NJ, and
approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the U.S. Environmental
Protection Agency. The mention of trade names or commercial products
is intended only to assist readers requiring acoustic monitoring information.
The Agency does not endorse or recommend the products or companies
indicated in this capsule report over those not mentioned or not in existence
at the time of publication.
------- |