United States
Environmental Protection
Agency
Municipal Environmental Research
Laboratory
Cincinnati OH 45268
EPA 600/2-79-084
August 1979
Research and Development
Field Testing of
Prototype Acoustic
Emission Sewer
Flowmeter
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution-sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-084
August 1979
FIELD TESTING OF PROTOTYPE ACOUSTIC
EMISSION SEWER FLOWMETER
by
K. M. Foreman
Research Department
Grumman Aerospace Corporation
Bethpage, New York 11714
Contract No. 68-03-2525
Project Officer
Hugh Masters
Storm and Combined Sewer Section
Wastewater Research Division
Municipal Environmental Research Laboratory (Cincinnati)
Edison, New Jersey 08817
MUNICIPAL ENVIRONMENTAL 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 publica-
tion. 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 of commercial products constitute endorsement or
recommendation for use.
LL
-------
FOREWORD
The 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 testimony to the deterioration of our natural environment.
The complexity of the environment and the interplay between its components
require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem
solution and it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research Laboratory
develops new and improved technology and systems for the prevention, treat-
ment, and management of wastewater and solid and hazardous waste pollutant
discharges from municipal and community sources, for the preservation and
treatment of public drinking water supplies, and to minimize the adverse
economic, social, health, and aesthetic effects of pollution. This publi-
cation is one of the products of that research; a most vital communications
link between the researcher and the user community.
The need exists to develop innovative, passive, nonintrusive, and low
cost solutions to the problems of continuous measurement and recording
of flows in storm and combined sewers. This experimental investigation
is of one such technique that monitors the pseudosound produced by flow past
a channel discontinuity. The results of laboratory and field tests demon-
strate the feasibility of this method of flow measurement using an accel-
erometer transducer attached to the outside surface of a flow channel.
Francis T. Mayo, Director
Municipal Environmental Research
Laboratory
iii
-------
ABSTRACT
This research project was designed to verify the acoustic emission flow-
meter concept in the natural operating environment of different storm sewer
field sites. The flowmeter is a novel, passive, nonintrusive technique that
uses the local sound resulting from partial transformation of the pressure
loss of flow at a channel or conduit discontinuity. In this context, a
discontinuity is any significant change in channel cross section or flow
direction. Other objectives were to examine the feasibility of calibrating
flowmeter field installations by means of geometrically similar small scale
laboratory models, and to explore the suitability of sewer manholes as
flowmeter sensor installation locations.
The three field sites used and the test results correlating acoustic
characteristics and physical flow rate of stormwater are described in the
body of this report.
The investigation demonstrated that the flowmeter principles hold true
in large storm sewers of 60 inch (1.5m) diameter and for flow rates up to
about 7500 gpm. The measured sound power, in decibels, is related to mass
flow rate to the 1.4 to 1.7 power, depending on channel discontinuity
characteristics. A manhole appears suitable for sensor installation. Small
scale laboratory models appear to simulate fairly well the sound intensity
to flow rate relationship of full scale sites according to theoretical
scaling laws. However, the spectral features of the acoustic signature
appear in conflict with presently formulated theory; additional research
into this anomaly is indicated. Recommendations are offered for future test-
ing and system design activities.
This report was submitted in fulfillment of Contract 68-03-2525 by
Grumman Aerospace Corporation under sponsorship of the U.S. Environmental
Protection Agency. This report covers the period July 6, 1977 to October
6, 1978, and work was completed as of October 6, 1978.
iv
-------
TABLE OF CONTENTS
Disclaimer ±±
Foreword ±±±
Abstract iv
List of Illustrations vi
Acknowledgment ix
1. Introduction 1
2. Conclusions 3
3. Recommendations 5
4. Scope of Work 7
Field Test Phase 7
Laboratory Test Phase 15
Theory 19
5. Experimental Equipment 25
Acoustic Measurements 25
Physical Measurements 26
6. Test Results 29
Cutter Mill Drain Sites (Lake Success Area) .... 29
Baldwin Creek Site 49
7. Discussion 62
References • 66
Glossary. • • • 67
-------
FIGURES
Number Page
1 Cutter Mill Drain field test sites, Lake Success Area, N.Y. . . 10
2 Plan view map of Cutter Mill Drain field test sites, Lake
Success Area, N.Y 11
3 Map of Baldwin Creek field test site, Baldwin, N.Y 13
4 Baldwin Creek field test site, Baldwin, N.Y 14
5 Grumman Research water supply facility 16
6 l/20th scale model of the Cutter Mill Drain field test
sites , 17
7 l/20th scale model of the Baldwin Creek field test site
installed in the Research facility 18
8 V-notch weir at Cutter Mill Drain outfall for measuring
low flow rates 27
9 Approximate water height in the 60-inch diameter Cutter Mill
Drain sewer pipe corresponding to measured flow rate 28
10 Sensor installation positions at the Cutter Mill Drain outfall. 30
11 Acoustic signal vs flow rate at position 1 for the l/20th
scale model of the Cutter Mill drain outfall 32
12 Acoustic signal vs flow rate at field site position 1 - Cutter
Mill Drain location 33
13 Spectral distribution of acoustic signal for two flow rate
differentials at field site position 1 - Cutter Mill
Drain outfall 34
14 Acoustic signal vs flow rate at position 3 for the l/20th scale
model of the Cutter Mill Drain outfall with different step
sizes 35
VL
-------
FIGURES (Continued)
Number Page
15 Additional data of acoustic signal vs flow rate at
laboratory model position 3 - for Cutter Mill Drain
outfall geometry 36
16 Comparison at Fourier Analyzer - processed acoustic signals
with General Radio 1933 Sound Meter measurements at laboratory
model position 3 - for Cutter Mill Drain outfall geometry. . . 37
17 Acoustic signal vs flow rate at field site position 3 -
Cutter Mill Drain outfall : . . . 39
18 Spectral distribution of acoustic signal for three flow rate
differentials at field site 3 - Cutter Mill Drain outfall. . . 39
19 Acoustic signal vs flow rate at laboratory model position 5
for three step sizes at a frequency of 2760±1QO Hz - for
Cutter Mill Drain outfall geometry 41
20 Comparison of acoustic signals at positions 3 and 5 of the
l/20th scale model of the Cutter Mill Drain outfall
geometry at a frequency of 2900±100 Hz 42
21 Locations of acoustic sensor mounting stud installations
at the Cutter Mill Drain check dam ,. . . . 44
22 Acoustic signal vs flow rate for laboratory model position
20 for two upstream step sizes - Cutter Mill Drain check dam
geometry 45
23 Additional data of acoustic signal vs flow rate at laboratory
model position 20 - for Cutter Mill Drain check dam geometry . 45
24 Acoustic signal vs flow rate at field site position 20 -
Cutter Mill Drain check dam 47
25 Spectral distribution of acoustic signal for three flow rate
differentials at field position 20 - Cutter Mill Drain
check dam 48
26 Acoustic signal vs flow rate for laboratory model positions 6
and 7 for two upstream step sizes - Cutter Mill Drain check
dam geometry 49
27 Sensor installation positions at the Baldwin Creek 100 degree
turn section field test site 50
28 Acoustic signal vs flow rate for l/20th scale laboratory
model position 4 - Baldwin Creek field site geometry. Data
processed by Fourier analyzer 53
vii
-------
FIGURES (Continued)
Number Pa&
29 Comparison of laboratory and field data of acoustic signal
vs flow rate at sensor position 4 - Baldwin Creek field site
geometry. Data processed by Fourier analyzer 53
30 Comparison of laboratory and field data of acoustic signal vs
flow rate at sensor position 4 - Baldwin Creek field site
geometry. Data obtained by General Radio 1933 Sound Meter
with filter unit 55
31 Comparison of spectral distribution of intensity of acoustic
emission signal obtained by General Radio 1933 Sound Meter
vs recorded data processed by Fourier Analyzer. Sensor
position 4 on Baldwin Creek field site geometry model 56
(a) Zero flow rate (laboratory background noise) 56
(b) Flow rate of 0,175 pps 56
(c) Flow rate of 1.15 pps 57
(d) Flow rate of 2.65 pps 57
32 Spectral distribution of acoustic signal at several sensor
positions of the Baldwin Creek field test site for a flow rate
differential of 526 pps. Recorded data processed by Fourier
Analyzer 58
33 Spectral distribution of acoustic signal at four sensor
positions of the Baldwin Creek field test site for a flow
rate differential of 526 pps. Data obtained by a General
Radio 1933 Sound Meter with filter unit 59
34 Comparison of field and laboratory data of acoustic signal vs
flow rate at a frequency of 480 Hz. Three sensor installation
positions of the Baldwin Creek field test site geometry.
Data obtained by a General Radio 1933 Sound Meter with filter
unit 60
35 Comparison of measurement method on acoustic signal vs flow
rate for two sensor installation positions on the laboratory
model of the Baldwin Creek field test site geometry. Data
at a frequency of 480 Hz 61
viii
-------
ACKNOWLEDGMENT
We acknowledge the assistance of Messrs. Arthur Speidel, Noe Areas, and
Richard Yoos of Grumman Aerospace Corporation (GAC) in conducting the test
program and processing the test data.
We are also indebted to Mr. Hugh Masters and Mr. Richard Field of
EPA, who provided helpful guidance in keeping the project in correct
perspective.
ix
-------
SECTION 1
INTRODUCTION
The investigations and results described in this final technical report
for EPA Contract 68-03-2525 concern a novel, passive, nonintrusive, sewer
flowmeter approach called the acoustic emission flowmeter (U.S. Patent
3,958,458, May 25, 1976). This kind of new technology is needed to solve
the urgent problems of measurement and recording of storm and combined
sewer flows for management, preservation, and treatment of national water
resources. With the documentation presented in this report, the technical
and operational base of data has been vastly expanded, thereby enhancing
the credibility of a highly unconventional and innovative flowmeter concept.
The acoustic emission flowmeter utilizes the local, nonpropagating
sound resulting from the partial transformation of flow pressure loss at a
discontinuity in a channel or conduit. Earlier work, principally in the
research laboratory (Ref. 1) demonstrated the basic mechanism of the flow-
meter concept for a large variety of open and closed channel flows and
channel materials representative of actual sewer networks. However, the
much reduced geometric size and flow quantity imposed by laboratory
conditions provided uncertainty about the extrapolation to more realistic
full scale field conditions of physical size, quantity, and flow quality.
Somewhat mitigating these doubts were the confirmatory indications of a
very limited preview field test involving monitoring sanitary sewage flow
entering an industrial waste treatment plant.
The present investigation has been geared principally to seeking veri-
fication of the earlier examinations, but in the natural environment of
three different storm sewer field sites. The geometric scale of the new
flow channel is twenty times the laboratory scale, and the maximum storm
sewer flows are almost two orders of magnitude larger than the greatest
laboratory flow. Other objectives of the current work have been a) to
examine the feasibility of calibrating field installations of the flowmeter
-------
system by means of geometrically similar, small scale models in the con-
venience of the laboratory environment, and b) to explore the suitability
of using sewer manholes as sensor installation locations. The passive
sensor used in the acoustic emission flowmeter is an accelerometer, and
its attachment to the nonwetted surfaces of the conduit is through the
intermediary of a dedicated mounting stud cemented to the channel wall.
The principal investigator has endeavored also to address problems of
future implementation of the technique, during the present research phase.
Several alternative adhesives and mounting stud installation procedures have
been evaluated under conditions of long term natural environmental exposure.
During the last quarter of the test program, a portable, simplified, real
time, sound monitor system also was used concurrent with the more usual
procedure of recording sound data and transferring the taped sound to a
Fourier Analyzer at a later time for spectral distribution processing;
comparative results are given here.
The body of this report describes the three field sites and the test
results of acoustic emission and physical flow rate. Measurements were
made in the field and for one-twentieth scale, geometrically similar,
laboratory models. Conclusions of the investigation and recommendations
for future action directly follow this introductory section.
-------
SECTION 2
CONCLUSIONS
Field investigations, involving unprecedented storrawater flow rates
and storm sewer sizes, have verified the operational principles of the
acoustic emission flowmeter in real working environments. Using a
prototype flowmeter system, acoustic measurements were made at three
different sites of dry weather and stormwater flow rates up to about 7500
gpm over a 10-month test period. Small acoustic sensor mounts installed on
weathered concrete surfaces remained operationally intact and secure at
normally unprotected public works sites during this period.
Unambiguous acoustic signals have been correlated with physically
measured flow rates at these three different field sites. Sound power has
been found to be related to flow rate to the 1.4 to 1.7 power, depending
on the channel discontinuity geometry. This is equivalent to an increase
of average sound level intensity of 4 to 5 decibels for each doubling of
flow rate. The suitability of low cost, rugged and small accelerometers
has been confirmed for monitoring in adverse working environments, including
a manhole installation, the dipole-like acoustic radiation caused by flow
near a channel discontinuity.
The successful demonstration of principles and extension of operations
to realistic full scale field conditions has enhanced the credibility of
the innovative and unconventional acoustic emission flowmeter concept.
Precalibration of field installations by using inexpensive, geometri-
cally similar, small scale laboratory models and theoretical scaling laws
has been partially demonstrated. The laboratory flow model simulates
fairly well the relationship of acoustic emission intensity to flow rate
of the full scale field site. However, the sound frequency band in which
the modeling agreement was achieved appears at variance with current
predictive theory. Without further research this problem could impede a
-------
full understanding of the process and could pose a potential barrier to
future implementation without in situ calibrations.
Measurements made with a breadboard, real time, portable acoustic
emission monitoring system compared generally favorably with data recorded
by the prototype system and processed through a high resolution Fourier
Analyzer. This encouraging experience with a quasi-commercial system also
provided guidance for signal filtering specifications of an eventual
operational flowmeter system.
-------
SECTION 3
RECOMMENDATIONS
The storm sewer demonstration testing reported here has furthered the
attractiveness of the passive acoustic emission flowmeter concept for
municipal and regional sewer systems. Some basic studies still appear
desirable to trace the conversion of flow energy into a spectral distribution
of sound intensity for different channel disturbances. This information
is particularly needed to characterize better the frequency of expected
maximum acoustic intensity and usable harmonics. The prevailing consensus
toward combined sewer systems because of economics, suggests that further
field examination of the acoustic emission flowmeter should be directed to
that application. Finally, it is becoming evident that efficient narrow
band signal filtering is needed, particularly above 2000 Hz, to make a
real time flowmeter system compatible with the likely spectral quality of
accelerometer sensor output for real installations.
Therefore, we recommend the following three step activity as the next
phase in the flowmeter development cycle.
• Conduct field tests of the prototype acoustic emission flow-
meter system at a well -instrumented combined (sanitary plus
stormwater flow) sewer facility. Independent flow rate and
flow quality instrumentation should be available at the
facility and testing should be conducted at a suitable channel
discontinuity for a sufficient time (e.g., months) to ex-
perience a wide variety of flow conditions
• Conduct further laboratory and field testing with
sufficient fluid dynamics and acoustics measurements
to resolve anomalies of acoustic emission characteristic
frequency described in the current test phase. Examine
possible structural resonance excitation of sensor installations
-------
and response characteristics for at least three different
accelerometer designs. Develop theory modifications to
reflect empirical results
Design a prototype flow measurement system combining
elements of a portable direct readout, real time sound
meter and a narrow band filter circuit compatible
with signal outputs of commercially available low cost
accelerometer sensors. Purchase components and construct a
portable system and conduct shakedown tests at an
instrumented field site. After these initial
preliminary checkouts, install system at the field
site for long term evaluation of performance and
reliability relative to EPA objectives.
-------
SECTION 4
SCOPE OF WORK
The primary objective of this project has been to verify the novel
acoustic emission flowmeter technique in real field installations. However,
it is perceived that the operational utility of the flowmeter would be
enhanced if in situ calibration of each installation were unnecessary, and if
the ubiquitous sewer manhole could be used for flow sensor locations. The
basic theory of the acoustic emission flowmeter requires the empirical
determination of a flow calibration parameter associated with the character-
izing dimensions of the candidate channel discontinuity.
Our initial approach to preclude full field calibrations has been to
employ geometric similarity principles to justify using small scale models
-for each candidate sewer configuration to be instrumented. Thus, with the
convenience and in the controlled conditions of the laboratory, low cost
modeling holds the promise of a technically acceptable and economically
viable alternative to direct field calibrations. For this reason, the test
project has consisted of two phases; one, involving measurements of acoustic
emission and associated flow rates entirely in a field environment, and a
second phase conducted with one-twentieth scale models of the field sites,
in the Grumman Research Laboratory. Data of the latter phase then can be
compared to field results and our similarity law hypothesis tested.
Also within the scope of this project has been an exploration of the
suitability of sewer manholes for sensor installation.
FIELD TEST PHASE
Three field test sites were selected, at the contract's initiation, with
the concurrence of the EPA Technical Monitor. All were within Nassau
County, New York, which allowed us to minimize travel distance from the
Grumman main facility in Bethpage, New York. The two general areas, Cutter
Mill Drain, on the north shore and Baldwin Creek on the south side of Long
-------
Island, previously had been instrumented each with a recording Parshall
flume by Department of Agriculture (DoA) Project 208 personnel. Because
of several coordination talks with cognizant DoA engineers it had been
assumed during precontract negotiations with the EPA that the Project 208
flow measurement equipment would remain in place and be available for this
effort; however, the flumes were removed by DoA personnel at an indeterminate
time prior to our actual contract award. Therefore, we proposed, as an
alternative, physically to measure the flow rate at our test sites by a
velocity traverse-integration method. With the EPA Technical Monitor's
concurrence, a Marsh McBirney Model 201 flow velocity measurement system
was purchased, and a probe support-manual traversing mechanism was constructed
for attachment at each test site. A V-notch wier also was used particularly
for very low, dry weather flows.
Cutter Mill Drain Site
The Cutter Mill field test location is an historic natural stream in the
University Gardens section of the Lake Success area of northwestern Nassau
County, New York. The storm drainage area encompasses 260 acres consisting
of high value suburban developments (59% area), golf courses (35% area),
and major arterial roads (6% area). Dry weather flow is small but continuous
and originates from the spilloff of Lake Success which is at about 70 feet
(21 m) higher elevation.
The focus of our measurements in this area involved two sites at an
open channel conducting the flow away from a 60-inch (1.52 m) diameter
underground, reinforced concrete, sewer pipe exiting at a concrete headwall
just north of Sussex Road. The open drain continues north for about 530
feet (162 m) before changing into a man-made culvert beneath North Hempstead
Turnpike (Northern Boulevard). A reinforced concrete check dam intercepts
the drainage ditch about 195 feet (59 m) downstream of the headwall. Both
locations, the headwall and check dam, provide well defined but different
type discontinuities in the channel that lend themselves to sensor installa-
tion for the acoustic emission flowmeter. The flow quantity measured at
the headwall is, for practical purposes, the same as that flowing over the
checkdam because additional runoff is very small from the surrounding
grounds between the two sites.
-------
The 7 foot wide by 7 foot high (2.1 x 2.1 m) headwall location is a
sudden enlargement type channel discontinuity for the 60-inch (1.5 m)
diameter storm sewer line. Sloping wing walls, made of reinforced concrete,
expand the channel to a full 14 foot (4.2 m) width downstream of the headwall,
and a concrete apron runout section of about 50 feet (15.2 m) length
maintains a uniform and flat drainage bed. At the sewer pipe centerline,
there is a step of 0.9 feet (0.27 m) between the bottom of the sewer pipe
and the concrete apron. As is evident from Fig 1, the 5 foot (1.5 m)
wide paved sloping sides downstream of the wing walls are partially covered
by a carpet of grass, weeds, and soil from the adjoining grounds. Thirty
feet upstream of the headwall is a manhole where an underground 24-inch
(0.61 m) diameter sewer line connects orthogonal to the underground 60-inch
(1.5 m) pipeline. The smaller pipe conducts storm water from two curbside
catch basins at a naturally low (dip) point of Sussex Road.
Sensor mounting studs were cemented to the sewer pipe roof (at center-
line), the headwall, and to the wing walls. The velocity traversing
equipment was mounted to the 1 foot (0.3 m) wide top of the concrete headwall.
The reinforced concrete check dam creates an upstream waterpool about
2.7 feet (0.82 m) deep and a downstream free water fall height of about
2 feet (0.61 m). A 6.5 foot (2 m) wide reinforced concerete apron provides
a spillway over the dam and the overflow falls into a downstream pool of
varying depth caused by a severely eroded drainage bed and the remains of a
concrete pavement. As shown by Figs. 1 and 2, the sloping banks of the
channel downstream of the check dam are paved with concrete to form two
20 foot (6.1 m) long strips, each 5 feet (1.5 m) wide, paralleling the
stream. This check dam provides a second type of channel discontinuity.
Sensor mounting studs were cemented to the downstream face of the dam, to
the unwetted sloping flanks of the spillway, and to the paved bank of the
drain channel downstream of the check dam.
Baldwin Creek Site
The Baldwin Creek study site is located in the community of Baldwin in
south central Nassau County, New York. The open channel conveys the effluent
of a 60 inch (1.5 m) diameter submerged reinforced concrete storm sewer for
-------
CHECK DAM SITE
HEADWALL-OUTFALL SITE
Figure 1. Photo views of the Cutter Mill Drain
field test sites. Lake Succes Area, N. Y.
-------
60" REINFORCED
CONCRETE PIPE
REINFORCED
CONCRETE
PAVED SLOPE CHECK DAM
MANHOLE
NASSAU COUNTY
TYPE 4
BRICK CHIMNEY
REINFORCED
CONCRETE
HEADWALL
REINFORCED
CONCRETE
WINGWALL
CATCH BASIN
NASSAU COUNTY
TYPE 1
2-STORY
BRICK
HORIZONTAL SCALE
I I I . I . I . I . I . I
4 FT CONCRETE WALK
BITUMINOUS
MACADAM
ROAD
Figure 2. Plan view map of Cutter Mill Drain field test sites. Lake Success area, N.Y.
-------
about 500 feet (150 m) before re-entering a submerged 60 inch (1.5 m)
diameter sewer pipe which conducts the flow southward into Parsonage Cove
of Hempstead Bay. The open channel is a public works improvement of an
originally meandering natural stream through a lowlands swamp area.
The underground storm sewer runs south beneath a public street,
Howard Place, (see Fig. 3) before taking a 90 degree eastward turn, runs
60 feet (18.3 m) beneath a grassy land parcel owned by the county, and
then takes a 100 degree turn southward, and emerges at a reinforced concrete
headwall. Immediately downstream of the headwall is a 20 foot (6.1 m) long
concrete channel bed which forms a stepless transition with the bottom
of the D-shaped flow channel of the 100 degree turn. This turn is of
reinforced concrete construction with a rectangular roof shape and includes
a manhole midway through the turn and above the outer radius wall. The
90 degree turn further upstream, is also of reinforced concrete construction
and has a manhole rising above the mid-turn position. The sloping banks
of the open channel are paved with concrete, to a width of about 10 feet
(3. m) as shown by Fig. 4.
The double turn configuration of the submerged sewer offers a third
channel discontinuity type to be studied. Sensor mounting studs were ce-
mented to the concrete headwall, at various places along the lower part of
the internal surface of the 100 degree turn, at various places along the
inner surface of the manhole, and at the inner surface of the junction between
the 60 inch diameter pipe and the 100 degree turn.
The Baldwin Creek channel drains storm waters from about 430 acres
of an area consisting mainly of single family residential buildings. About
10 acres of two-story apartment dwellings and 40 acres of commercial struc-
tures are included in this drainage basin, primarily along two major
thoroughfares, Grand Avenue and Woodside Avenue.
Dry weather flow is so slight as to be practically unmeasurable; how-
ever, there are retained water pools in the silted creek bed downstream
of the headwall, and some trickle movement is visually evident. There is a
maximum elevation differential of 25 feet (7.6 m) between the headwater of the
Grand Avenue storm sewer feed branch and the open channel culvert, and a
12
-------
u>
CEMENT
PAVEMENT
NASSAU COUNTY
STD. MANHOLE
TYPES WITH TYPE 1
CHIMNEY
CONCRETE
HEADWALL
NASSAU COUNTY
STD. MANHOLE
HOWARD ./
PLACE
60 R. C. PIPE
ASPHALT
PAVEMENT
Figure 3. Map of Baldwin Creek field test site, Baldwin, N. Y.
-------
GENERAL AREA VIEW
TOP SURFACE OF 100 DEGREE TURN
SECTION SHOWING MANHOLE COVER
HEADWALL OF 100 DEGREE TURN SECTION
Figure 4. Photo views of Baldwin Creek field test site,
Baldwin, N. Y.
14
-------
maximum of 15 feet (4.6 m) head differential between the beginning of the
Woodside Avenue storm sewer branch and the entrance to the open channel.
Historic records credit the Baldwin Creek channel with a maximum storm
flow rate of about 60,000 gpm (225 m /min), or over 8300 pounds per second.
LABORATORY TEST PHASE
In support of the field investigations, models were constructed with the
essential geometric features of the three field sites. The water was
supplied from the same water flow research facility (see Fig. 5) used in
the prior project (Ref. 1). It consists of a 240 gallon (908 m) cylindrical
tank, which can be pressurized by a regulated air supply, and a manually
controlled flow valve upstream of a 20 feet (6.1 m) length of 3 inch (7.6 cm)
diameter steep pipe. This long pipe was taken to simulate the 60 inch
(1.5 m) diameter storm sewer pipe at the two field sites, and established
the l/20th scaling size employed for the models. Construction of the models
was of a vinyl-mix cement to permit thin sections and with five-mesh (per inch)
stainless steel screen reinforcement to simulate the steel reinforcing rods
of the actual installations. Molds of wood or aluminum were used to
retain the wet cement until hardening. The three-inch diameter water supply
pipe was attached by a clamped flange with gasketing to preclude leakage.
The entire assembly was supported in a wooden box which was mounted on
rails at the facility as shown, typically, by Fig. 6 for the Cutter Mill
Drain test site model. In similar manner, as shown by Fig. 7, the 100
degree turn portion of the Baldwin Creek test site, was reproduced
in a l/20th goemetric scale version. Because of the space limitations in
the laboratory, the Cutter Mill check dam model was only about 3 feet (0.9 m)
downstream from the headwall model instead of a more accurately scaled 10
foot (3.0 m) distance. It was believed that the possible local sound
distortion at either discontinuity caused by flow interaction at the other
channel change could be discounted in the laboratory environment. Also,
because of the small test equipment size the maximum flow rate that could
be retained in the model open channel was just under 4.5 Ib/sec (2.0 I/sec).
To maintain this flow the upstream control valve had to reduce the pressure
head in the 3 inch (7.6 cm) line to a value much below the usual tank
15
-------
PLANT
WATER
SUPPLY
V SIGHTTUBE
I
AIR SUPPLY
VALVE /
FLOW
CHECK
VALVE
PLANT
AIR
SUPPLY
AIR PRESSURE
REGULATOR WATER
SUPPLY
VALVE
GRIT
INJECTOR
FLOW DIVERSION
TROUGH
SENSOR
MOUNTING
STUD
(TYP.
TOLEDO 55 GAL. DRUMS PLANT
SCALE FOR COLLECTING AND DRAIN
(2500 LB CAP) MEASURING TEST FLOW
Figure 5. Grumman Reserch water supply facility.
-------
COMBINED MODELS AS FABRICATED
FLOW DIRECTION
MODEL INSTALLED IN RESEARCH FACILITY
DETAIL OF CHECK DAM MODEL
Figure 6. 1/20th scale model of the Cutter Mill Drain field test sites.
17
-------
Figjre 7. 1/20th scale model of the Baldwin Creek field test site
installed in the research facility.
18
-------
pressure level of about 10-15 psi (69 - 103 kPa). Therefore, the specific
energy of the field site open channel flow was not rigorously simulated
in the laboratory.
Another imprecise scaling aspect was the large size of the sensor rela-
tive to the model's dimensions compared to the field test geometry instru-
mentation. Thus, each sensor on the model intercepted and was excited by a
greater relative portion of the flow's acoustic emission field than in the
full scale field installation.
As in the prior project, average flow rate was determined by timing
a measured collected weight of water (see Ref. 1); the flow rate usually
was averaged over a 20 second interval. Sound recordings were made during
the timed interval.
THEORY
Almost every real sewer flow line has turns, steps, junctions of two or
more unequal sized lines, and flow control equipment that introduce dis-
continuities in the channel. In the near field of each discontinuity, the
adjustment of the flow to new downstream conditions produces local pressure
pulsations and flow unsteadiness. In many cases the periodicity of the
pressure pulses is large enough to create audible sound. But the propagation
of these pulsations depends on the local mean directed flow velocity rather
than the fluid sonic speed. Therefore, this sound predominates only near
its source and becomes attenuated with distance from the discontinuity.
The near field, nonpropagating sound emitted by the flow is called
pseudosound (Refs. 2 and 3). The acoustic emission can be detected by
appropriate sensors, such as accelerometers, at the channel boundaries of
the flow because the pseudosound radiates as a dipole source, orthogonal
to the fluid flow direction. Eventually, the unsteadiness created at the
discontinuity becomes dissipated in the complex downstream flow field and
becomes coupled to the fluid's far field acoustic radiation pattern.
o
Sound pressure of pseudosound, p , is of the order of pu , where p is
the average fluid density and u is the average fluid velocity. The dipole
source has a total sound power, SP , in a free field of
19
-------
£z (1)
T PC
where £ is the characteristic dimension of the flow channel discontinuity,
and c is the fluid sonic speed. Substituting for sound pressure
•j 4
z. u
cr> ~ P^ ( 0\
SPT - — (2)
and Eq. (2) becomes the theoretical basis for the ideal fourth power
dependence of sound emission on flow rate. For a particular channel geometry,
the volumetric flow rate, Q, is proportional to velocity, so
SPT = Q4 (3)
In the pseudosound process, the direct link to the channel discontin-
uity is discerned by the sound at a characteristic frequency, f , where
fc - c/6 (4)
and 6 is the physical dimension characterizing the discontinuity. By
analogy to the fundamental physical definition f = c/A, where A is the wave-
length of sound, <5 can be equated to the wavelength of the characteristic
frequency in the fluid. Thus, the physical geometry of the discontinuity
provides a guide to determining the characteristic frequency and harmonics
where flow-related sound emission can be monitored.
Where the shape of the discontinuity shadows the sound from being
radiated the net radiated power is diminished by a factor, e. For bends
and branch junctions, e tends to unity. For valves and metering sections, values
of e are less than one and must be determined empirically. Furthermore,
for sensors mounted to the outer surface of the conduit, there is a sound
power loss through the conduit walls, thus transmission coefficient, T, is
approximately proportional to the ratio of channel hydraulic radius to wall
thickness, R/t. In cylindrical channels, such as sewer pipes, T ~ D/t,
for source excitation wavelengths larger than IT D = A . Loss factors for
noncylindrical channels and at frequencies less than A are discussed in
r
Ref. 4. The acoustic coupling of sound from one media to another involves
20
-------
the acoustic impedance, r = pc/s, in the two media, where s = the interface
surface area.
The acoustic coupling between water and concrete or iron surfaces
is about 53 and 15%, respectively (Ref. 1). The poor acoustic coupling of
_3
water to air (about 10 %) predicts that the airborne sound from the flow
process at the discontinuity theoretically is an insignificant mechanism
compared to the direct coupling of the sound created in the flow with the
solid boundaries of the flow channel.
Since the total sound power at the emitting source is of a finite
amount per unit time, the further the sound is monitored from the source,
the less the sound intensity in each unit area being monitored. However,
the integrated sound intensity over the entire radiating surface (at a
given distance from the source) never exceeds the total sound power of the
source itself plus the zero flow background noise. Total sound power equals
intensity times radiating surface area.
SCALING LAWS
The mean square pressure of sound at the inner boundary of a flow
channel is
CO
P2 = / F1(n) dn (5)
o
where F-, (n) represents the spectral intensity distribution of the pressure
fluctuations in the fluid, for n frequencies.
The pressure variations associated with pseudosound result from
velocity fluctuations, du, of the directed pipe flow, U . The pressure
changes, dp, equal (pdu) (du) and are proportional to the dynamic pressure
2
pU . Then the mean square pressure:
o
-2 24
p <* p v (6)
The spectral distribution can now be written as
00
-2 2 4 . ,n ^ .,n ^
P = p Uo / ^(—)d(—) (7)
O 00
21
-------
where ^ (n/n ) is an empirical function that must be determined for a range
of frequencies relative to the frequency, n <* c/d where d is the depth of
flow in an open channel, or the pipe diameter of a continuous, full flowing,
cylindrical line.
Using Eqs. (5) and (7) gives
^ (n/n ) = <|> (n/U /d)
1 o 1 o (8)
F/n)
Mn/n ) = (— - ) (c/d)
2TT 4
p Uo
The elastic properties and constraints of the channel wall material
attenuates the fluid pressure variations in a frequency dependent way- T(n) ,
such that the spectral density at the outer surface of the channel structure
is
F2(n) = T(n) F± (n) =
then
p2 = / F2(n)dn = P2UQ4 /(no/n1»)2(n/no)(t.1(n/no)d(n/no) (9)
o o
where n^ = c /d.,; c is the sonic speed of the channel wall material and d..
is the ringing diameter of the pipe (nominally the same as the average pipe
diameter for relatively thin-walled cylindrical pipes).
Also,
F2(n) = p2Uo4(no/n1)(J)3(n/no) (10)
where cj)_(n/n ) is an empirically determined factor that combines the sound
source radiating, and wall transmitting, dependencies on sound frequency.
For constant density of the fluid
noU 4 4 d1
F?(n) cc ° ,(n/n ) (11)
^ n.. o o o c d 3 o
1 m
22
-------
and
_ U 4n
P - ~9-^- f <)>3(n/no)d(n/no) (12)
For a flow channel discontinuity characterized by a dimension, h, we
can now define a pseudosound characteristic frequency as n = Uc/h and
the spectral distribution function of intensity takes on a value F(n ) .
From Eq. (11)
F2(nc)
o
U/h
(c/c) (d
m
Therefore, for geometrically scaled pipelines, if the same pipeline
construction material and fluid velocity is used, at the pseudosound
characteristic frequency the intensity function is
F2(n ) d
-fr- - ^
o
while at other frequencies
F?(n)
d $ (n/n) (13a)
U
o
Equation (13) is the theory that guides the promotion of small scale geo- •
metrically similar models for calibrating full size field installations.
For model flow velocities that duplicate the mean full scale field
velocity, the pseudosound intensity for the model should be the same as
the geometrically similar field installation. On the other hand, model
flow speeds that are lower (e.g., 1/2 to 1/5) than the mean real flow case
should result in much lower intensity signals (e.g., by 10 to 30 dB lower)
even with otherwise geometric similarity. However, channel dissipative
23
-------
mechanisms downstream of the discontinuity modify the theoretical fourth
power variation of sound intensity with flow velocity; our experience
reported here indicates a velocity exponent more in the nature of between
1 and 2, over a broad flow range. The sound intensity signal of the model
than should be only about 3 to 15 dB lower than the full scale installation
for model flow velocities of 1/2 to 1/5 of the original.
With full geometric similarity and fully duplicated mean flow velocities,
the model's quantitative flow rate should be approximately proportional
to the square of the linear scaling factor compared to full scale
conditions when the sound intensities from the model and full scale channel
setups are equal.
24
-------
SECTION 5
EXPERIMENTAL EQUIPMENT
ACOUSTIC MEASUREMENTS
Two approaches were used in processing the acoustic emission of the flow
in this project. The initial method used essentially the same laboratory
prototype measurement system previously employed in Ref. 1. This system
records signals from an accelerometer sensor mounted to the conduit walls at
normally unwetted areas. The tape recorded data are processed through a
Fourier Analyzer computer (HP-5465A) at some later date and the spectral
distribution of sound intensity is plotted for correlation with physical flow
measurements.
The second system also accepts the accelerometer output signal but
provides a real time display of sound level amplitude for a particular
frequency band. A General Radio (GR) Model 1933 Sound Level Analyzer Meter
displays signal amplitude at the various frequency bands of interest that
are obtained with a Krohn-Hite bandpass filter interposed between the sensor
and the meter. Both the Model 1933 Meter and the filter device are battery
powered and portable to field sites as well as usable in the laboratory.
All GR 1933 readouts have been manually recorded. A full description of the
Fourier analysis processor has been given in Ref. 1.
Accelerometers are mounted onto the solid walls of the channel
(usually concrete) by being screwed onto cheap, dedicated, stainless steel
mounting studs that have been cemented with commercially available epoxy
based adhesives to the concrete surfaces. The obvious advantage of this
approach is that many inconspicuous studs may be left at any field test site
without concern for potential theft or damage. The prototype system and
accelerometer is brought to the test site and activated by test personnel.
(A more compact operational system would be mounted on a utility pole).
25
-------
Our experience with these mounting stud installations in the field
includes the October 1977 to August 1978 exposure period. The first stud
attachments deteriorated rapidly because the weathered top layer of the
concrete surface first had not been removed before cementing; the epoxy
adhesive layer was stronger than the underlying cement layer. Our later
experience from about mid-November 1977 indicated excellent retention of the
mounting stud except where ground level attachment made it subject to
accidental mechanical impact by, for example, grounds maintenance machinery.
Although the stud face is only about 1/2 inch (1.25 cm) in diameter, it
adheres as strongly to a well prepared concrete surface as when it is
cemented to a 2 inch (5 cm) square metal plate prior to mounting on the
environmental surface.
Accelerometers, of three different output signal sensitivities, pro-
vided superior vibration detection (amplitude and frequency range) on the
concrete surfaces, nearby to the channel discontinuity, than, for example,
neighboring microphones supported to receive primarily airborne sound.
In the latter measurement equipment situation, the higher frequency range
(e.g., above ~8 kHz) was undetectable.
The most useful accelerometer was a B&K 4332 model because its high
sensitivity of 62 mv/g provides strong sensor signals, and its flat
frequency response to 50 kHz encompasses the needed working range.
PHYSICAL MEASUREMENTS
The physical measurement of the storm sewer flow in the field used
measurements of water velocity at a grid of many points, and the integration
of these data, each as representative of a small stream tube, over the
entire channel cross-sectional area. The velocity was measured with a Marsh-
McBirney Model 201 portable battery-powered, water current meter. This
system uses an electromagnetic sensing head. When immersed in water, a
voltage field is established around the probe. Electrodes imbedded in
the probe body sense the voltage field whose amplitude is proportional
to the water velocity around the probe. The electrical signals are transmitted
through a cable to a portable processing and meter display unit. The
26
-------
velocity readouts in the 10 ft/sec (3.05 m/sec) full scale range are real
time measurements and certified accurate to ±2%. The meter time constant
is only 6 seconds so practical flow rate changes are measurable. A marked
mounting rod and support structure allowed the sensing head to traverse,
in the vertical and horizontal directions, within 1.5 feet (0.5 m)
(upstream) of the exit face of the 60 inch diameter (1.5 m) sewer pipe
at the Cutter Mill and Baldwin Creek channels during wet weather flows.
The requirement of maintaining at least one inch (2.5 cm) of water around
the 1.5 inch (3.8 cm) maximum diameter of the velocity sensor limited
its usefulness in dry weather flows. For these cases where the maximum
water height in the sewer pipe is only 3 to 5 inches (7.5 to 12.5 cm), a
Y-nctch weir was used (See Fig. 8). Measurement of the maximum height of
water crest, H, for the Vee, and use of the equation
0 = 2.50 H
2.50
(14)
allows the volumetric flow rate to be computed. The dry weather flow at
Baldwin Creek is so slight, however, that even the V-notch weir was not
effective; the flow was estimated as essentially zero, although a slow
visible trickle existed.
Figure 8. V-notch weir at Cutter Mill Drain outfall for measuring low flow rates.
27
-------
From measurements of the maximum water height in the free-flowing
60 inch (1.5 m) diameter Cutter Mill sewer pipe, concurrent with the
velocity and measurements, a correlation has been made of flow rate to the
water height. The approximate relation, shown by Fig. 9, has been found
useful in making quick check assessments of mass flow during an acoustic
measurement session.
MAXIMUM
HEIGHT
OF WATER,
INCHES
14
12
10
^
DRY WEATHER FLOW
10
12
14
16
FLOW RATE, 10^ PPS
Figure 9. Approximate water height in the 60-inch diameter Cutter Mill
Drain sewer pipe corresponding to measured flow rate.
28
-------
SECTION 6
TEST RESULTS
The format of presentation in this section is to report on the
acoustic and flow data by field test site. For each particular site, first
the scale model laboratory results are given, and then the field measure-
ments documented. The implications of these combined results then are
discussed in a preliminary manner. These results are put into an overall
perspective in the following Section 7.
CUTTER MILL DRAIN SITES (LAKE SUCCESS AREA)
Outfall Site
As shown by the diagram and photograph (Fig. 10), three sensor mounting
studs (positions 1, 2, 2A) were attached to the head wall and two to the
wing wall (positions 3, 3A). Sensor positions 2A and 3A were intended
to be closer to the water surface then positions 2 or 3 and therefore to
improve the signal strength. However, sample data indicated in actuality
no appreciable difference in relative signal intensity. Also, data at
position 2 were of a similar nature to that at position 3, therefore,
no separate presentation are included here for positions 2, 2A, or 3A.
Electrical cables connected the screwed-on accelerometer to the preamplifier,
amplifier, and Nagra FM tape recorder. All the electronic components of
the prototype system rested on the concrete ledge atop the headwall.
Comparable sensor mounting locations to field positions, 1, 2, and 3
were provided on the l/20th scale laboratory model. Model position 1,
however, was on the model headwall, just above the supply pipe exit face,
instead of within and on the roof of the pipe as in the field installation.
In addition, the model was provided with sensor position 5, which is a near-
mirror image location to position 3, to check flow symmetry; no field sensor
location comparable to model position 5 was provided.
29
-------
SENSOR POSITION
to
O
VIEW OF
CORNER OF
OUTFALL
WALL
OUTFALL SENSOR
INSTALLATION POSITIONS
Figure 10. Sensor installation positions at the Cutter Mill Drain outfall.
-------
In the laboratory test series, the height of the step discontinutiy
between the supply pipe and the downstream channel model was changed from zero
to 1 inch (2.54 cm) by 1/2 inch steps. This investigation was designed
to demonstrate the premise of acoustic emission signal amplitude being
dependent on a frequency that characterizes the discontinuity. However,
data for all three discontinuity heights were not taken at all sensor
positions. The changing of the headwall discontinuity was independent
of the geometry of the check dam modeled site located about 3 feet (1 m)
downstream. It should also be observed that the altering of the vertical
step had a very minor influence on the contours of the horizontal (or
lateral) discontinuity; the latter channel change also stimulates acoustic
emission from the flow but at different characteristic frequencies than
the vertical step.
Sensor Position 1
The data obtained at model position 1 (above the supply pipe exit plane
centerline) is shown by Fig. 11 in terms of increase in sound decibels (dB)
above ambient background noise, for increased weight rate of water flow.
(The decibels are computed from the equation dB = 10 log (SP/SP ), where
Riir
SP is the sound power of the flow acoustic signal and SP is a referenced
KbT
sound power signal.) The five frequency bands compared by Fig. 11 were
taken from the full spectral range processed by the Fourier Analyzer. In
general, especially at the lower flow range (0-3 pps) the signal change in
dB appears to increase with flow at a higher rate for the 1/2 inch (1.27 cm)
step than for the zero step arrangement. At the higher flow rates the
difference in sound signal for the two step sizes tends to diminish. The
theoretical characteristic frequency in water for the 1/2 inch step is about
59,000 Hz, and it is indeterminately greater for the zero step condition.
Therefore, the observed peaks in the overall spectral distribution shown
by Fig. 9 must represent higher harmonics, or characteristic frequencies
arising from other dimensions of the model such as the lateral change
in flow channel area. In the latter case the increased signal from the
larger vertical drop condition represents the greater pressure fluctuations
(or local pressure drop) associated with increased impact effect from higher
31
-------
AdB
— f = 400, h = 0.5"
f = 680, h = 0.5"
f = 400, h = 0"
' _. ^ f = 20 - 20,000 Hz, h = 0"
-7^ — ^r
X ^Nf = |
f = 200, h = 0.5"
f = 20-20,000 Hz, h= 1/2"
f = 680, h = 0"
f = 4000, h = 0.5"
2. 3.
W, PPS
Figure 11. Acoustic signal vs flow rate at position 1 for the 1/20th scale model of the Cutter Mill Drain outfall.
water falls. A portion of these pressure losses are converted into
acoustic energy that radiates as a dipole. The greater signal at the lower
frequencies probably arises from the lower transmission loss of sound
through the model walls at the lower frequencies. For example, the indicated
actual exponent relating sound power to mass flow rate, from these model
tests with a 1/2 inch step is 1.24 at a frequency of 400 Hz and 0.62 at
a frequency of 4000 Hz. Both determinations were made for the flow
rate range of about 0.8 to 4.2 pps, corresponding ideally to full scale
flows of 320 to 1680 pps (2300 to 12,000 gpm), prespectively. The 50%
lower exponent at the higher frequency is attributable to the wall trans-
mission loss as well as different acoustic conversion mechanism efficiency.
The corresponding field site data at position 1 is given by Fig. 12
for five frequency bands. The sound power, in dB, increases with increased
flow rate above dry weather flow (22 pps).
32
-------
30
20
SOUND
POWER,
AdB, REL.
TO DRY
WEATHER
FLOW
10
11,000 Hz
20-20,000 Hz
700
800
900 • 1000 1100
100 200 300 400 500 600
DRY WEATHER FLOW W, PPS
Figure 12. Acoustic signal vs flow rate at field site position 1-Cutter Mill Drain Location.
In the frequency band between 5600 and 7200 Hz, the sound power varies
with flow rate approximately as the 1.15±0.03 power over the experienced
flow rate range. As becomes evident by studying Fig. 13, which displays
the approximate spectral distribution of acoustic emission differential
between the maximum wet weather flow (.1040 pps ~ 7500 gpm) and dry weather
flow (22 pps ~'160 gpm), the peak or characteristic frequency of the field
site is at about 11,000 Hz. The sound power varies overall at the 1.6 power
of flow rate at this characterizing frequency. The emitted sound is greatly
attenuated at higher frequencies (up to 18,000 Hz) because the higher flow
rates impact the channel bed further away from position 1 than at lower
flows; the longer transmission path in the concrete reduces the signal
strength. Over a lower flow rate range, between 230 and 22 pps (1650 and
160 gpm), the flow impacts closer to the exit face and the channel wall
33
-------
30 i-
25 -
20
AdB 15
10
SP~W1.6
BETWEEN W= 1040
AND 22 PPS
SP~W1'2
BETWEEN W= 230
AND 22 PPS
100
1000
FREQUENCY, Hz
10000
2x 104
Figure 13. Spectral distribution of acoustic signal for two flow rate differentials at field site position 1
Cutter Mill Drain outfall.
attenuation is relatively less severe at high frequencies. The exponent of
flow rate in this lower flow rate range is 1.2 at a characteristic frequency
of 18,000 Hz.
Sensor Position 3
Model test results monitored at wing wall position 3 are presented by
Fig. 14 for three frequencies bands and three step discontinuity heights.
As we evidenced at position 1, the higher step (i.e., 1 inch (2.54 cm))
produces greater sound intensity above background noise for all flow rates
compared to the lower step height conditions. The frequency band-center
yielding the greatest increase in flow acoustic emission intensity, above
the laboratory background noise, is at about 2900 Hz for position 3.
The increase in intensity with step height for the high range laboratory
34
-------
h=step height
40 r
AdB
1" = h,f = 2880 Hz
1", f = 360
1", f = 920
1/2", f = 360
1/2", f = 920
A ZERO, f = 2880
ZERO, f = 360
ZERO = h, f = 920
0.5 1.0 1.5 2.0 2.5
W, PPS
Figure 14. Acoustic signal vs flow rate at position 3 for the 1/20 scale model of the
Cutter Mill Drain outfall with different step sizes.
flows (i.e., ~ 3-4 pps), is approximately linear at 10 ± 2 dB per 0.5 inch
(1.77 cm) of step. Model flow sound power varies with flow rate to the
1.2 power at the l/20th scaling, over the range of flows tested, but at
a 1.95 power at l/10th scaling of the discontinuity step, for a frequency
of 2900 Hz. For no step discontinuity, the flow sound power varies as
flow rate. These results and the observation that 2900 Hz corresponds to
a characteristic wavelength in water of about 20 inches (51 cm), or even
numbered fractions thereof for harmonics, makes it evident that this
frequency probably is not linked directly to step height. More likely
2900 Hz results from the increase in flow channel (horizontal plane) width
at the discontinuity (i.e., approximately 5 inches (12.7 cm). The larger
signal intensity at larger step heights then represents a modulation (or
augmentation) of the basic acoustic emission signal because of the greater
level of pressure fluctuations (greater pressure loss) produced by the
higher water falls. Further support of this conclusion is the fact that
(see Fig. 15) that the signal intensities at about 6000 Hz (i.e., second
35
-------
AdB
f = 2900 Hz
(FUNDAMENTAL)
360 Hz >(1"STEP)
6000 Hz
(2ND HARMONIC)
360 Hz (1/2" STEP)
f = 2900 Hz (1/2" STEP)
~12000 Hz
(4TH HARMONIC) -(T'STEP)
f = 6000 Hz (1/2" STEP)
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
W, PPS
Figure 15. Additional data of acoustic signal vs flow rate at laboratory model position 3-for Cutter
Mill Drain outfall geometry
harmonic) and 12,000 Hz (4th harmonic) are virtually at about fixed
decrements below 2900 Hz; these decrements can be attributed to the channel
wall transmission loss coefficients which are only frequency dependent and
which exhibit greater loss at higher frequency. On the other hand, at a
frequency of 360 Hz (l/8th the assumed fundamental) the signal falls off
by a variable amount depending on flow rate above about 1 pps. The sensor
signal at Model position 3 also was processed by the GR 1933 signal
analyzer. A comparison of the GR 1933 meter (used with a Krohn-Hite
Model 3700 filter) readings to the Fourier Analyzer results (see Fig. 16)
indicates fairly good agreement for all frequencies below about 4000 Hz.
The data at about 2600 Hz has the best agreement. The spread in results at,
for example, 6000 Hz and above probably can be attributed to the filter
36
-------
AdB 20 -
GR 1933 ANALYZER
FOURIER ANALYZER
f = 1050 Hz
OVERALL (20 - 20K Hz)
40 i-
30
AdB
20
10
AdB
GR 1933 ANALYZER
GR 1933 ANALYZER
FOURIER ANALYZER
f = 2600 Hz
— f = 10,000 Hz
FOURIER ANALYZER
•f = 6000 Hz
GR 1933 ANALYZER;
f = 10,000 Hz
Figure 16. Comparison of Fourier Analyzer - processed acoustic signals with GR-1933
meter measurements at laboratory model position 3-for Cutter Mill Drain outfall geometry
37
-------
unit's cut-off pattern. One of the operating characteristics of the
Krohn-Hite filter is that if the adjustable high frequency and low
frequency filtering limits are set too close (e.g., on a central frequency)
the net output signal is 6 dB lower than the actual input signal. The
comparative signal values for the model's overall frequency band between
20 and 20,000 Hz) are approximately 10 dB apart over most of the flow rate
range, with the GR 1933 measurements lower than the Fourier Analyzer results.
The exponent for I/20th scale model flow rate in the sound power
relationship is 1.89 at a frequency band centered about 1050 Hz, 1.35 at 2600
Hz, and 2.43 at 6000 Hz. In the field, sensor position 3 on the wing wall
is about 6.7 feet (2.0 m) from the bottom lip of the sewer pipe centerline
instead of a few inches as in the model. Then, the distance correction
needed to be added for the field data is about 9 dB for very low flows and
about 8 dB for the high flow rates in order to correlate properly with
laboratory model test data.
The field test data given by Fig. 17 at two frequencies show an increase
in sensed signal which increases to the 0.95 power of flow rate (for 18,000
Hz). If an 8 dB correction for distance is made, the source of the acoustic
emission signal would appear to vary with flow rate to the 1.4 power at a
frequency of 18,000 Hz. Because the characteristic frequency in water of
the 10 inch (25.4 cm) step at the outfall is about 6000 Hz, the data given
at 18,000 Hz would be third harmonic results. However, an examination of
the approximate spectral distribution of intensity of field signal, shown
by Fig. 18, indicates that the peak intensity at about 6000 Hz is about
3 dB lower than the peak at 18,000 Hz. Normally, the intensity of harmonics
are less than at the fundamental mode. While it is true that sensor
installation resonance could be the cause of the higher signal, this point
has not been demonstrated a posteriori. On the other hand, the second
harmonic which would occur at about 12,000 Hz is not evident from the field
data. However, this omission, in itself, is not an invalidating reason
because, often, only odd harmonics are excited in large structural systems.
38
-------
30 r-
20
dB REL
TO DRY
WEATHER
FLOW
10
SP~ W
A/.95
.18,000 Hz
.4000 Hz
/DRY WEATHER FLOW
200
400 600 800
W, FLOW RATE, PPS
1000
1200
Figure 17. Acoustic signal vs flow rate at field site position 3-Cutter Mill Drain outfall.
20
15
AdB 10
BETWEEN W= 1040
AND 22 PPS
SP~W-!
BETWEEN W= 155
AND 22 PPS
"
BETWEEN W = 230
AND 22 PPS
_L
J
100
1000
FREQUENCY, Hz
10000
2x 104
Figure 18. Spectral distribution of acoustic signal for three flow rate differentials at field site 3-
Cutter Mill Drain outfall.
39
-------
It should be noted that the laboratory model's dimensions would create
a 1-inch (2.5 cm) step-induced characteristic frequency of almost 60,000 Hz
which could not be recorded because it exceeds by a considerable amount,
the working range of the tape recorder and sensor. For this reason, the
frequency-simulation of the field installation could not be accomplished by
geometrically similar small scale laboratory models. Exaggerating the
vertical scale at the discontinuity, to create a step of 3.5 inches (8.9 cm)
or more, would bring the frequencies into the working range, but might have
created other problems of geometric distortion. For this reason, in
analyzing the small scale model data, the other and larger dimensional
discontinuity features that also could induce concurrent acoustic emission
mechanisms, but at measurable frequencies, were used.
Any other explanation of the field data trend at 18,000 Hz except for
sensor installation resonance, fails to explain the variation of signal
intensity from the dry weather (22 pps) to 1040 pps flow range. Clues to
another source mechanism is suggested by the data for the 230 to 22 pps
flow range in Fig. 18. Intensity peaks at 1800, 6000, and 18,000 Hz
suggest the effective flow discontinuity to be the horizontal spread of
the pipe flow into the open channel (i.e., from a 2.5 foot (1.5 m) radius
to a 5 foot (3.0 m) channel half-width). Then the decrease of relative
intensity with third harmonic (at 6000 Hz) and with ninth harmonic
(at 18,000 Hz) are consistent with a generally expected trend. However,
this alternative explanation breaks down for the higher flow range data.
Because of- the inability to model signal frequency, the model and
field data cannot be completely and neatly correlated. However, the fact that
source sound power for both size scales of channels are proportional to
flow rate to a power of between approximately 1.0 and 2.0 seems to confirm
the similarity of source mechanism. The location of the sound emission gen-
erally in the audible frequency range, and below ultrasonic, also confirms
the pseudosound, dipole type, origins of the measurements.
Sensor Position 5
This mounting stud location is close to the mirror image of position 3
on the opposite side of the laboratory scale wing wall. Howevers it is
40
-------
slightly more downstream than position 3. No comparable data were taken
at the field site.
The variation of sound intensity with flow rate is shown for a
frequency of 2760 Hz by Fig. 19. In similar fashion to data at position
3, the intensity at position 5 increases about 10 ± 2 dB per 1/2 inch
(1.27 cm) of model discontinuity step. The exponent of variability with
flow rate is 1.54 compared to a 1.95 exponent for position 3 and similar
frequency (2900 Hz) with a 1 inch (2.54 cm) step. Comparison of other
step sizes for the two positions is given by Fig. 20, and show excellent
correlation. These data indicate the uniform distribution of the acoustic
emission with geometric symmetry of the flow channel. Other data at
frequencies of 360 and 7200 Hz, not shown, also yield similar results.
AdB 20 -
10 -
W, FLOW RATE, PPS
Figure 19. Acoustic signal vs flow rate at laboratory model position 5 for three step
sizes at a frequency of 2760 ± 100 Hz - for Cutter Mill Drain outfall geometry.
41
-------
AdB
POS 3 - 1" STEP
*r—
POS 5-1" STEP
POS 3 - 1/2" STEP
POS 5 - 1/2" STEP
POS 3 - NO STEP
•MtfMM
POS 5 - NO STEP
W, PPS
Figure 20. Comparison of acoustic signals at positions 3 and 5 of the 1/20th scale model
of the Cutter Mill Drain outfall geometry at a frequency of 2900 ±100 Hz.
4.5
42
-------
Check Dam Site
This location was the second of three field channel discontinuity sites
investigated. As shown by the diagram and photograph (Fig. 21), sensor
mounting studs were attached to the downstream vertical wall of the check
dam (position 12) and on the right cemented bank (position 20) about 5 feet
(1.5 m) downstream of the spillway lip. In addition, sensor mounting studs
were cemented to the concrete flanks (positions 6 and 7) of the spillway.
However, the only field position for which acoustic data were obtained over
the entire range of flow rates was at position 20. For some of the other
locations the mounting stud did not remain attached to the check dam during
a test because of what proved to be inadequate preparation of the concrete
substrate.
Sensor position 20
Data taken at model position 20 are given in Figs. 22 and 23 for five
different frequencies. Although the step in the upstream channel discon-
tinuity should have had no influence on the acoustic emission at the
check dam, evidently there was a carryover effect as the plotted data
clearly indicates. Because no data were taken in the laboratory with the
check dam section replaced by a straight section, we cannot accurately
assess how much of the pseudosound radiation is converted into sound carried
along by the flow to the vicinity of the check dam, about 3 feet (0.91 m)
downstream. However the 9-10 dB sound difference (at the highest of flow
rates) for the 1 inch (2.54 cm) step differential indicates a possible
50% conversion efficiency of the original measured outfall source sound
production rate of 20 dB per inch. At low flows, (i.e., at up to 1 pps)
there appears to be very little effect of the upstream step because,
presumably, the absolute level of acoustic emission was weak enough to be
completely dissipated.
The strongest acoustic signature at the check dam model is at 1920 Hz
frequency and a second harmonic is observed at 3800 Hz. The signal at
960 Hz is about 1/3 less than at 1920 Hz, indicating that it is not likely
to be the fundamental frequency of the check dam's acoustic signature.
43
-------
VIEW OF SENSOR
POSITION 20 .
(LOOKING
DOWNSTREAM)
CHECK DAM SENSOR
INSTALLATION POSITIONS
Figure 21. Locations of acoustic sensor mounting stud installations at the Cutter Mill Drain check dam.
-------
AdB 20 -
ALL CURVES FOR SENSOR
POSITION 20 EXCEPT IF
NOTED OTHERWISE
960 - 1.0
Figure 22. Acoustic signal vs flow rate for laboratory model position 20 for two
upstream step sizes - Cutter Mill Drain check dam geometry.
50 r
AdB
Figure 23. Additional data of acoustic signal vs flow rate at laboratory model position
20 - for Cutter Mill Drain check dam geometry.
45
-------
The sound power appears to vary to the 2.9 power of flow rate at 1920 Hz;
the exponent of flow rate is about 2.5 at 1400 Hz, 2.2 at 960 Hz, and 1.9
at 3800 Hz. These observed exponents are the closest to the theoretical
fourth power law for dipole sound radiation, and indicate the closest
approximation to an ideal pseudosound production situation created in the
laboratory. Once again, the characteristic dimension in water of the
model check dam step is associated with a frequency of about 59,000 Hz.
Because this value is well above the instrumentation capabilities, other
acoustic emission frequencies observed in the test data have been used for
correlation with physical flow rates.
The field site sensor position 20 test data are presented by Figs. 24
and 25. The latter is an approximate spectral distribution curve that
reveals peak trends of fundamental frequencies and their harmonics. Over
the full maximum range of flow rates (1040 to 22 pps), the fundamental
frequency appears to be 3800 Hz which corresponds to a characteristic
wavelength in water of 15.5 inches (39.4 cm). This dimension is about equal
to the actual free waterfall distance from the check dam spillway to the
downstream pool, and is confirmatory data on the basic pseudosound
mechanism creating the monitored acoustic signal. A third harmonic peak at
11,200 Hz also is evident from Fig. 25 and it exhibits the expected lower
intensity, in this case about 7.5 dB less. A fifth harmonic which would
be at 19,000 is just beyond the actual range of data presented but the
trend of Fig. 25 appears to indicate a reasonable expectation of encountering
46
-------
40 i-
30 -
f = 3800 Hz (CORRECTED FOR SOURCE
DISTANCE FROM SENSOR)
AdB REL
TO DRY
WEATHER 20
FLOW
10 -
200
400
600
800
1000
1200
FLOW RATE, W, PPS
Figure 24. Acoustic signal vs flow rate at field site position 20 - Cutter Mill Drain check dam.
this frequency if the instrumentation and data processing capabilities
so permitted. The exponent relating flow rate to sound power at 3800 Hz
is 1.07 if no correction is made for physical distance of the sensor from
the effective source of the acoustic emission (i.e., the water
fall). With a correction of about 9 dB for the 9.5 foot (2.6 m) distance,
the flow rate exponent becomes 1.6 for the sound power at the source.
Coincidently, this value is in close agreement with the laboratory scale
model exponent of 1.9 at 3800 Hz, where distances were too short to require
correction.
Sensor Position 6
Data obtained with the laboratory scale model at sensor position 6 is
shown by Fig. 26 for several frequencies and two upstream step sizes.
Representative comparative data for sensor position 7 also are given. These
results are in general agreement with the trends observed at position 20
but are of lower intensity. This smaller signal is understandable for
a number of reasons, most notably a) the sensor positions are upstream of
47
-------
1-06
00
AdB 5 -
AdB
BETWEEN W= 155
TO 22 PPS
BETWEEN W = 230 TO
22 PPS
BETWEEN W = 1040
TO 22 PPS
Figure 25. Spectral distribution of acoustic signal for three flow rate differentials at
field position 20 - Cutter Mill Drain check dam.
-------
SOLID LINE- 1/2" STEP
DASHED LINE - T'STEP
AdB
4080 AND 2500 Hz - POS 7
ALL FOR POS 6 EXCEPT
WHERE NOTED
W, PPS
Figure 26. Acoustic signal vs flow rate for laboratory model positions 6 and 7 for two
upstream step sizes - Cutter Mill Drain check dam geometry.
the actual discontinuity; b) the positions are further away from the water-
fall impact location than at position 20. In addition, at high flow
rates position 6 can become wetted by water level surges.
In view of the less favorable circumstances attendant to installation
of sensor mounting pads on the exposed flanks of the check dam, only a
limited examination was made at positions 6 and 7; greater attention was
directed to position 20.
BALDWIN CREEK SITE
100 Degree Turn
The locations of sensor mounting studs at the Baldwin Creek field site
are shown by the outline drawing (Fig. 27). Sensor position 4 is located
49
-------
COVER
GROUND LEVEL __
24"
Figure 27. Sensor installation positions at the Baldwin Creek
100 degree turn section field test site.
-------
on the reinforced concrete headwall, at the outfall to the open channel,
and about 3 inches (7.6 cm) from the roof and lip of the outfall. Positions
1, 2, and 3 are upstream, within several inches, of the outfall and on the
roof centerline, east and west walls (outer and inner radii), respectively,
of the reinforced concrete turn section.
Sensor locations 5 through 8 are located roughly midway through the turn
with the odd numbered location near to the roof and the even numbered
positions at about mid-channel height of the outer and inner radii walls,
respectively. Our visual evidence of severe erosion of the lower part
of the east (outer radius) wall indicates that for severe stormwater
flow conditions the high velocity profile is skewed toward the turn
location near position 6.
Sensor positions 9, 10, and 12 are located in the vicinity or on
the inner walls of the manhole chimney. Any of these three locations could
serve as an indicator of the general suitability of manholes for field
sensor installation. Finally, sensor location 11 is on the downstream
surface of the headwall joining the 60 inch (1.5 m) diameter concrete
sewer pipe to the 100 degree turn section.
In the laboratory the general geometric features of the Baldwin Creek
turn section have been duplicated (see Fig. 7). However, because of small
size and spatial limitations, faithful duplication of the inner wall
installations of sensor mounting studs was compromised; all sensor studs
on the model were cemented to the outer surfaces of the turn model at the
approximate relative location of the field installation. The relatively
thin wall of the laboratory model was expected to make transmission loss
small enough to make the external surface pickups closely measure the
sound excitation they would monitor at the internal surface, if physically
feasible. The manhole chimney also was not reproduced because attachment
of a mounting stud at positions 9, 10, and 12 would have been impractical.
Model positions 5 and 9 effectively substitute for the manhole vicinity
sensor positions in the field. As shown by Fig. 7, the 60 inch (1.5 m)
diameter concrete pipe was modeled by a l/20th inch linear scale polyvinyl-
chloride (PVC) plastic pipe which was cemented to the upstream face of the
51
-------
thin wall concrete model turn section. Upstream of the attachement the
supply pipe was configured to have the same geometric relationship (i.e.,
a 90 degree turn preceding the straight run) for the model as in the
field. Not simulated, however, was the underground (buried) nature of the
field installation. This was not done because the focus of the investigation
was the 100 degree turn at the exit of the pipeline and not the pipeline
itself.
Field site test data were limited to the single wet weather flow of
526 pps (3780 gpm) and dry weather flow (zero pps) conditions because of
a prevailing (Spring) relatively low rainfall situation in the field test
region during the normal work week; project budgeting made premium pay
periods (e.g., weekends) unavailable for additional measurements during
rainy episodes outside of scheduled work hours.
Sensor Position 4
Fourier Analyzer-processed sound data for this sensor mounting
location, obtained with the laboratory scale model, are presented by Fig.
28 for eight different frequency band centers. These frequencies all
represent cases of major sound intensities with clearly unambiguous signal
variations with flow rate. The strongest trend is at a frequency of 480 Hz
which corresponds to a wavelength of about 10 feet in water. There are
no physical dimensions in the lab model, involving water flow, that are so
large. Therefore, the source of the measured, apparently flow-related,
sound is not presently determined. On the other hand, sound intensities
at higher frequencies (e.g., 4000 and 6000 Hz) do also have usable
signal-flow rate characteristics and do correspond to discontinuity dimensions
that can be identified with the outfall channel enlargement of the laboratory
model.
Sensor positions 3 and 5 on the headwall of the Cutter Mill Drain
laboratory model bear superficial geometric similarity to position 4 on the
Baldwin Creek model. However, a comparison of similarly processed sound
data from these similar geometric positions on both models does not yield
similar frequency or intensity trends. Therefore, the upstream channel
52
-------
AdB
W, PPS
Figure 28. Acoustic signal \i$ flow rate for 1/20th scale laboratory model position 4
Baldwin Creek field site geometry. Data processed by Fourier analyzer.
AdB
/ xx/^ X^CT*^ FIELD f • 6000 H
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
_l MODEL W, PPS
50
FIELD W, PPS
Figure 29. Comparison of laboratory and field data of acoustic signal vs flow rate at sensor
position 4 - Baldwin Creek field site geometry. Data processed by Fourier analyzer.
53
-------
differences of the two scale models obviously affect the exit flow profile
of the outfall, and the sound emission characteristics.
The stormwater field data obtained at position 4 display a spectral
distribution with a peak intensity at about 10,000 Hz. This frequency
corresponds to a 6 inch (15 cm) dimension in water which can be approximately
associated with the effects of a sudden increase in channel area, at the
outfall, on the storwater flow level. Unfortunately, this hypothesis cannot
be thoroughly tested because only one storm flow rate was documented. The
laboratory scale tests provided no further insight into the sound source
mechanism that was detected by the sensor at field position 4.
The Fourier-analyzed sound data obtained in the field and laboratory
are compared in Fig. 29. It is evident that at any particular flow rate
the lower range of frequencies have higher signal intensities for the model
and approximately the opposite tendency for the field site. These trends,
frankly, are counter to the expected behavior.
On the other hand, the comparable sound data from the GR 1933
monitor, shown by Fig. 30, is more in keeping with theoretical expectations.
The model data measured by the GR 1933 meter at 480 Hz is about 3 dB below
the field measurement if the associated respective flow rates are related
by the flow area scale factor (i.e., 20 squared). The GR meter signal
strength of the model measurements tend to decrease at higher frequencies
(see Figs. 31a, b, c, and d), and the same trend is observed in the field.
However, the field installation signals at the higher frequencies are of
lower intensity than the model data for geometrically scaled similar
flow rates. The GR 1933 monitor tends to miss fine detail of the sound
spectra but detects the overall trend. Thus, the concurrence of the GR 1933
monitor's readings with general theoretical expectations seems to imply
that the anomalous results obtained with the fine resolution Fourier
processing probably are caused by inclusion of local resonances at the
sensor position rather than by radically different physics. A more rigorous
explanation would require a detailed sound survey to obtain direct and full
casual relationships. However, because the primary objective of this
investigation is to obtain field signals and sensor locations that can
54
-------
30
20
AdB
© FLD -480 Hz
MODEL 480 Hz = f
MODEL 1100 Hz
\
MODEL - 10,400 Hz
\
i
0.5 1.0 1.5 2.0 2.5 3.0
1/20th SCALE MODEL W, PPS
3.5
4.0
1200
FIELD SITE W, PPS
Figure 30. Comparison of laboratory and field data of acoustic signal vs flow rate at sensor
position 4 - Baldwin Creek field site geometry. Data obtained by GR1933 Sound
Meter with filter unit.
prove usable for flow measurements, we have succeeded in our quest without,
frankly, being completely sure of the reason.
Spectral Distribution of Intensity
The comparative field intensity data for sensor positions 1, 4, and 7
shown by Fig. 32, indicate (through Fourier processing) that peak amplitudes
are excited between 6000 and 11,000 Hz for these locations in the flow
range encountered. Thus, the observations at position 4 seem more likely to
have their origins in the flow distortion near the midpoint of the turn
rather than with the end of the turn. This is another restatement of the
principle that the most significant momentum loss affecting acoustic
emission is where the flow vector changes the most. The single points spotted on
55
-------
50 r 62 ,-
45
40
35
30
52 -
AdB
NOTE:
DISPLACED
AXIS FOR
GR DATA
10dB
42
32
GR SOUND
ANALYZER
(FILTERED)
RESULTS
FOURIER
ANALYZER
RESULTS
10
1 03
f/4
5x 10J
49
44
39
34
'29
64 r-
54
AdB
44
NOTE:
DISPLACED
AXIS FOR
GR DATA
34
10
a. Zero flow rate (lab background noise).
GR SOUND
ANALYZER
(FILTERED)
RESULTS
FOURIER
ANALYZER
RESULTS
*•-*
f/4
5x
b. Flow rate of 0.175 pps.
Figure 31. Comparison of spectral distribution of intensity of acoustic emission
signal obtained by General Radio 1933 Sound Meter vs. recorded
data processed by Fourier Analyzer. Sensor position 4 on Baldwin
Creek field site geometry model .
56
-------
68
58
AdB
48
38
GR SOUND
ANALYZER
(FILTERED)
FOURIER
ANALYZER.
RESULTS
28
10
f/4
5x 103
c. Flow rate of 1.15PPS.
61
r 74
56
51
- 64
46
41
AdB
- 54
36
NOTE:
DISPLACED
AXIS FOR 44
GR DATA
FOURIER
ANALYZER
RESULTS
GR SOUND
ANALYZER
(FILTERED)
RESULTS
10
10J
f/4 5x 103
d. Flow rate of 2.65 PPS.
Figure 31. Comparison of spectral distribution of intensity of acoustic emission
signal obtained by General Radio 1933 Sound Meter vs. recorded
data processed by Fourier Analyzer. Sensor position 4 on Baldwin
Creek field site geometry model
57
-------
Fig. 32 for positions 5, 6, 8, 9 and 11 are the peak amplitudes in the
20-20,000 Hz range spectra at each position. Position 8 data reinforce
the conclusion about the major source being at or near the midpoint of the turn,
and the higher amplitude of position 8 compared to position 7 can be
attributed to the former's closer proximity to the water flow. Position 11's
relatively high signal at 8000 Hz may be associated with the abrupt channel
cross sectional change at the beginning of the 100 degree turn section.
Some representative spectral distributions of the signals measured with
the GR 1933 meter are shown by Fig. 33. In general, the filtered GR
measurements obscure the fine features at the higher frequencies that are
revealed with Fourier processing, as can be appreciated by reviewing Figs.
31a through d for the model tests. At the higher range of the frequency spec-
tra, the intensity of the sound measured with the GR 1933 instrument appears
14
AdB 12
10
POS 7
'"" 1000 10000 ,05
p. ,0 - FREQUENCY Hz
ngure 32. Spectral distribution of acoustic signal at several sensor positions of the Baldwin Creek
tield test site for a flow rate differential of 526 PPS. Recorded data processed by Fourier Analyzer
58
-------
AdB
POS9
FREQUENCY. Hz
Figure 33. Spectral distribution of acoustic signal at four sensor positions of the Baldwin Creek field
test site fo.r a flow rate differential of 526 PPS. Data obtained by a GR1933 Sound Meter
with filter unit.
higher than observed from Fourier processing. The apparent explanation for this
is that the GR meter integrates the signal over a broader frequency band
than the Fourier processing procedure normally includes. On a comparative
basis, Fig. 33 indicates that the signal taken in the manhole chimney
(position 12) is nearly as good in quality for flow measurement use as
the outfall location (position 4). The signals immediately upstream or
downstream of the manhole chimney (positions 9 and 5), but along the walls
of the channel turn proper, are essentially equal and of somewhat lower
intensity than in the manhole chimney itself. GR meter data at position 11
(not shown by Fig. 33) are similar to the position 9 signature, but are
not quite as useful for our purposes, as by the full Fourier processed data
method.
59
-------
With the filtered GR 1933 meter, the lowest frequency (480 Hz) of
peaking sound intensity, appears to give the best data agreement between
field and model results at sensor position 9. With the theoretical mass
flow scaling ratio used (i.e., 400), the field data of sound intensity at
positions 4 and 7 are greater than the I/20th scale model intensity
results (see Fig. 34). However, a comparison at 480 Hz of model data re-
sults obtained by the Fourier analyzer and by direct GR 1933 readout process-
ing techniques (Fig. 35) shows the two to be similarly unambiguous. However,
the former gives greater intensity trends with flow rate. Therefore, the
Fourier processed data has higher sensitivity for flow measurement purposes
than the GR 1933 method with the Baldwin Creek geometry features. Further-
more, the frequency band center in which signal amplitude should be
monitored is the best one (i.e., the one with peak signal intensity) which is
20
AdB
RE:ZERO 10
FLOW
MODEL POS4
©FIELD -4
MODEL POS7
I
0.5 1.0 1.5 2.0 2.5 3.0
1/20th SCALE MODEL W, PPS
3.5
4.0
0 400 800 1200
FIELD SITE W, PPS - (SCALE ALIGNMENT IS FOR 400 SCALING FACTOR)
Figure 34. Comparison of field and laboratory data of acoustic signal vs flow rate at a frequency of 480 Hz.
Three sensor installation positions of the Baldwin Creek field test site geometry. Data obtained
by GR1933 Sound Meter with filter unit.
60
-------
associated with the best measurement position and data processing
technique (i.e., at position 4 the 10,400 Hz signal of the Fourier process
rather than the 480 Hz signal of the GR 1933 process); compare Figs. 32
and 33.
30 i-
AdB
POS 9 (FOURIER)
POS4 (FOURIER)
POS4(GR1933 METER)
POS9(GR1933 METER)
1.0
1.5
2.0 2.5 3.0
W, FLOW RATE, PPS
3.5
4.0
4.5
Figure 35. Comparison of measurement method on acoustic signal vs flow
rate for two sensor installation positions on the laboratory model
of the Baldwin Creek field test site geometry. Data at a frequency
of 480 Hz.
61
-------
SECTION 7
DISCUSSION
Because of the voluminous data obtained in the field and laboratory
tests it is appropriate to place this investigation into perspective:
Notable from the viewpoint of the previous laboratory-bound investigation
(Ref. 1) is the fact that the acoustic emission concept has been found to
exist at three different field installations involving different channel
discontinuities. The confirmation of the concept principle has been
extended to flows that are two orders of magnitude greater than prior
laboratory results (Ref. 1) and with real stormwater including Autumn-through-
Summer seasonal changes of water quality. Sensor installations have been
successfully deployed, without special safeguard, on public works structures
for over ten months without being compromised by weather or human activity.
Data were obtained in a manhole installation and found to provide usable
signals for flow rate determination. The bulk of useful acoustic data were
obtained with a prototype system that does not allow real time measurements.
However, an approach was made toward an eventual compact, portable, real
time, commercialized-type field apparatus utilizing a signal filtering unit
and a rapidly responding sound meter. While the performance of.the
frequency filter interfered with a fully successful demonstration,
especially for the upper portion of the audible frequency range (i.e.,
between 2000 and 20,000 Hz), the evidence is clear that such a system can
work; elaborate Fourier processing facilities such as employed by the
laboratory test and prototype instrumentation system are useful but not
essentially for a flow measurement system based on flow acoustic emission
at a channel discontinuity.
The field tests did reveal an aspect of large-dimensioned channels that
was only partially appreciated beforehand. This involves the need to
correct acoustic measurements for spatial radiation of the sound if the
location of the sensor emplacement is distant from the apparent sound source.
62
-------
However, such distance corrections are only required to adjust the signal
of a single sensor monitoring low as well as high flow rates at an outfall.
Biasing the sensor' s installation location to the projection of one extreme
rate of flow into the downstream channel requires a distance correction
for the other flowrate extreme. This problem can be overcome by redundant
sensors whose locations can bracket the full range of likely flows; sensor
redundancy also improves system reliability. For manhole or channel
directional change type discontinuity locations, there does not appear to
be a need for a distance correction factor.
The use of scaled down, geometrically similar laboratory models to
calibrate field installation was successful in being demonstrated. At
certain select positions for sensor emplacement, there was good agreement
of the signal intensity scaling law, using the geometric scale factor for
mass flow correlation. The sound power was found related to flow rate to
an exponent of between about 1.4 and 1.7, throughout the full flow rate
range as indicated by Table 1. This power law relation translates into an
increase of average sound intensity of 4 to 5 decibels for each doubling
of flow rate. (The theoretical dipole relation for idealized, non-
dissipative flow processes predicts a 12 decibel change for each doubling
of flow rate.) Displaying the flow rate by a meter that is responsive
to sound power, in absolute units, would give meter scale displacement
ratios of 2.6 to 3.2 for each doubling of flow rate. However, the current
theoretical expectations concerning the characteristic frequency band, at
which the measurements were well correlated, were not encountered. There is
a suspicion that this observation is associated with local resonances of
the channel structure or sensor installation. Also, the relaxation of full
dynamic similarity of the model flow may have contributed to this result.
The prime focus of the current investigation precluded our following this
question beyond the conjectures already raised. While our reasonably
successful scaling demonstration without complete understanding of certain
anomalies is uncomfortable, it is not without precedent for as new and
unexplored a technology as we have employed in this innovative approach
to sewer flow measurement. Further research, particularly to examine
some unresolved questions such as revealed through this reported
63
-------
investigation, has a high probability of indicating new and better ways
to employ the full potential of the still largely unexploited flow acoustic
emission principle.
64
-------
TABLE 1 SELECTED MAJOR TEST RESULTS - ACOUSTIC EMISSION FLOWMETER
1 1
Channel Identification
Cutter Mil]
Drain
Lake Success,
N.Y. area
Baldwin Creek
Baldwin, N.Y.
area
Outfall
Wingwall
Wingwall
Check Dam
Outfall -
(downstream
of 100 degree
turn section)
Base of manhole
chimney in
100 degree
turn section
Midway in turn
of 100 degree
turn section
Sensor
Position
1
3
5
20
4
9
5
Band
Center-
Frequency
Hz
400
4000
1100
5600-7200
18000
1050
2600
2900
18000
2760
960
1400
1920
3800
3800
3800
480
1100
480
6000
480
480
480
Flow Rate Range, pps(w)
Laboratory
Model
0-4.2
0-4.2
0-4.2
0-4.2
0-4.2
0-3.6
0-5.1
0-5.1
0-5.1
0-5.1
0.1-2.65
0.1-2.65
0.1-2.65
0.1-2.65
O.J-2.3
0.15-2.3
0.15-2.9
Full
Scale
0-1680*
0-1680*
22-1040
22-1040
22-230
0-1680*
0-1680*
0-1680*
22-1040
0-1440*
0-2040*
0-2040*
0-2040*
0-2040*
22-1040
22-1040
40-1060*
40-1060*
40-1060*
40-1060*
0-526
40-920*
0-526
60-920*
60-1160*
Exponent "n"
in SP ~ wn
1.24
0.62
1.60
1.15
1.20
1.89
1.35
1.95
1.20
1.0
0.95
1.4
1.54
2.2
2.5
2.9
1.9
1.09
1.6
1.0 -|
1.0 J
1.69
1.1
1.1
1.4
1.4
1.85
1.69
Remarks
Notes 1, 2
Lab data
Lab data
Field site
Field site
Field site
Lab data
Lab data
Lab data 1 in. step
Lab data ^ in. step
Lab data no step
Field-w/o sensor
distance correction
Field -w sensor
distance correction
Lab data 1 in. step
Lab data
Lab data
Lab data
Lab data
Field-w/o sensor
distance correction
Field-w sensor
distance correction
Lab data - General Radio
1933 Sound Meter + Filter
Lab data
Lab data
Field data - coincides
with Lab data
Lab data - General Radio
1933 Sound Meter + Filter
Field data - coincides
with Lab data
Lab data
Lab data
Notes 1 All data processed by Fourier Analyzer unless otherwise noted. * Equivalent full scale flows applying
2 All laboratory data with l/20th scale models. 400 x scaling factor.
-------
REFERENCES
1. Foreman, K.M. A Passive Flow Measurement System for Storm and Combined
Sewers. EPA-600/2-76-115, U.S. Environmental Protection Agency,
Cincinnati, Ohio, May, 1976. 123 pp.
2. Ffowcs-Williams, J. Hydrodynamic Noise. In: Annual Reviews of
Fluid Mechanics. Vol. 1 Annual Reviews, Inc. Palo Alto, CA, 1969.
3. Blokhintsev, D.I. Acoustics of a Non-Homogeneous Moving Medium.
NACA TM 1399, National Advisory Committee for Aeronautics,
Washington, B.C. February 1956.
4. Ver, I.L. and C.I. Holmer. Interaction of Sound Waves with Solid
Surfaces. In: Noise and Vibration Control, L.L. Beranek, ed.,
McGraw-Hill Book Co., New York, N.Y., 1971. pp. 270-361.
66
-------
GLOSSARY
accelerometer: An electromechanical transducer that generates an electrical
output when subjected to acceleration. Piezoelectric discs" clamped
between a mass and base develop a potential field when the acceleration
of the mass exerts a force on the discs. The ratio of electrical
output to mechanical input is called sensitivity.
acoustic emission: The radiation of sound generated by the interaction of
fluid flow with a solid surface.
acoustic reflection: The change of direction of sound pressure waves
impinging on a less than perfect sound absorbing surface.
conduit discontinuity: Any change in a flow channel because of channel
cross section or shape, or where flow direction is significantly changed,
decibels (dB): A measure of the ratio of two amounts of sound power. The
range of sound pressure or intensity is so large that it is more
convenient to use the logarithm to the base ten to express this ratio
(bel). Decibel equals one tenth of a bel. When other quantities
(e.g., voltage) are related to the square root of power, the number,
n, of dB are: n = 20 log-.,.(v/v ), where v is the referenced quantity.
dipole: The type of sound source created when a fluid interacts with
a solid surface to produce unsteady forces. Because of its oscillating
nature, this source is analogous to two point-sources equal in
strength but opposite in phase and separated by a very small distance.
The radiated power is proportional to the fourth power of flow speed.
Because of the pressure cancellation in the plane normal to the dipole
axis, the directionality of radiation is strongest along the dipole
axis which is normal, to the flow direction.
67
-------
hertz (Hz): An international unit of frequency equal to the number of
cycles per second.
nonintrusive: Not penetrating the fluid flow boundary.
normalized acoustic signal: When transducers of different sensitivities
measure the same sound source, their dB sound signals are different
by the ratio of sensitivities. Similarly, when a constant sound
source signal is measured against different background noise levels,
the total signals are different by the relative difference in back-
grounds. When using decibel units for sound level, the irrelevant
variables of measurement such as background noise or sensor sensitivity
can be eliminated by subtracting their dB contribution from the total
signal. The resulting dB level then is the normalized signal, and
is a more valid measure of the sound source alone.
overburden: The soil or backfill covering a buried sewer pipe or flow
conduit.
passive flow measurement: A method of determining the mass or volumetric
rate, of flow by using energy normally radiated by the fluid flow as
opposed to imposing external energy sources or flow energy dissipating
devices.
pseudosound: The pressure pulses produced in locally disturbed fluid flow
that have characteristics of sound in the near field but do not
propagate into the far field of the fluid. The radiation pattern
of pseudosound is like a dipole sound source.
sound power(SP): The total amount of energy radiated by a sound source
throughout a spherical envelope in a period of time (watts). In
practice, the sound power level, L , is used to relate the ratio of
w
sound power to a reference power. By international agreement, this
—12 —1?
reference power is 10 watts, and L = 10 lognn(Pm/10 ), dB.
w 1U T
unambiguous signal: A sensor output signal that can be interpreted as re-
lating to only one flow quantity. Over a continuous range of signal out-
put there are no intermediate minima or maxima with regard to the
dependent parameter.
68
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-7Q-n84
3. RECIPIENT'S ACCESSI Of* NO.
4. TITLE AND SUBTITLE
Field Testing of Prototype Acoustic Emission Sewer
Flowmeter
5. REPORT DATE
August 1979 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
K.M. Foreman
8. PERFORMING ORGANIZATION REPORT NO.
RE- 566
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Research Department
Grumman Aerospace Corporation
Bethpage, New York 11714
10. PROGRAM ELEMENT NO.
1BC822. SOS 1, Task 47
11. CONTRACT/GRANT NO.
EPA Contract 68-03-2525
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory—Cin.,OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final Report 7/6/77-10/6/78
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES See also EPA-600/2-76-115, "A Passive Flow Measurement System
for Storm and Combined Sewers"
PO: Hugh Masters, (201) 321-6678, (8-340-6678)
15. ABSTRACT
This investigation concerns verifying the operating principles of the acoustic
emission flowmeter (U.S. Patent 3,958,458) in the natural environment of three
different storm sewer field sites in Nassau County, New York. The flowmeter is
a novel, passive, nonintrusive method that uses the local sound resulting from
the partial transformation of the flow energy at a channel or conduit discontinuity.
Any change of sewer cross section or flow direction qualifies as a discontinuity.
The results show that the flowmeter principles hold true in large storm
sewers of 60 inch (1.5 m) diameter and for flow rates up to about 7500 gpm. A
manhole appears to be suitable for sensor installation. The relation of sound
signal intensity to flow rate at full scale sites appears amenable to small scale
laborabory model simulation according to scaling laws.
Recommendations are offered for future testing and design activities.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Flow, Acoustics, Sewers, Flowmeters,
Experimental design, Acoustic signatures,
Sound level meters
b.IDENTIFIERS/OPEN ENDEDTERMS
Acoustic emission,
Flowmeter
c. COSATI Field/Group
20C, 20A
13B
14B
17A
9A, 9B
13. DISTRIBUTION STATEMEN1
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
79
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
* US GOVERNMENT PRINTING OFFICE'1979 -657-146/5465
------- |