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

-------