EPA-600/2-76-006
March 1976
Environmental Protection Technology Series
DESIGN AND TESTING OF A
PROTOTYPE AUTOMATIC SEWER SAMPLING SYSTEM
Municipal Environmental Research Laboratory
Office of Research and Development
U.S.Jnyirpnmental Protection Agency
Cincinnati, Ohio 45268
-------
:K
\ I- mi.
MM; i .si \ ;ia ; i : a | £$$ \ ig ii
IK iii^
si{;t» i r ;Vr'; I-?.,, 'i:^;;;i;;'f»;f"i.)
i^i^!^^^^^^^^^^^^^^
,:L,J i !„ !„„!! ill !„!„ ! llh.i!,,,,!,,! ! ,ID!,,I! !„ LH ,"£,«,,, ULn ! !!
SERIE;
Iggegrch reports of the Office of Research and Development, U.S. Environmental
"action A,qency_, have been grouped into five series. These five broad
:||orjes were jestablishecTto facilitate further dgyejopment and application of
' L. " [ " ' ." ," " , of traditional grouping was consciously
foster technology transfer and a maximum interface in related fields.
serigs are:
gfecjs
Protection Te
Research
iflMBJiaHffii'ft'^iW11"
If, Iggort has been assigned to the ENVIRONMENTAL PROTECTION
i lli,i§ §Sfl§§ S3,§ScI,ifee§ [SSSatSb. performed to develop and
iguipment, and methodology to repair or prevent
Iueririorisfratfi in^tn impntatinn
Iggradation from point and non-point sources.,of pollution. This
J5rov;d§§ the new or improved technology required for the control and
to meet environmental quality standards.
I
I
ii i
E33EE:
^1
g ft'iS
-------
EPA-600/2-76-006
March 1976
DESIGN AND TESTING OF A PROTOTYPE
AUTOMATIC SEWER SAMPLING SYSTEM
by
Philip E. Shelley
?•
EG&G WASHINGTON ANALYTICAL SERVICES CENTER, INC,
Rockville, Maryland 20850
Contract No. 68-03-0409
Project Officer
Hugh E. 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
EERU-TSX
RECEIVED
APR 1 8 1989
EERU-TIX
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
ii
-------
FOREWORD
Man and his environment must be protected from the adverse effects?;;.
of pesticides, radiation, noise, and other forms of pollution, and T
the unwise management of solid waste. Efforts to protect the envi-
ronment require a focus that recognizes the interplay between the
components of our physical environment—air, water, and land. The
Municipal Environmental Research Laboratory contributes to this
multidisciplinary focus through programs engaged in
• studies on the effects of environmental contaminants
on the biosphere, and
• a search for ways to prevent contamination and to
recycle valuable resources.
The deleterious effects of storm and combined sewer overflows upon
the nation's waterways have become of increasing concern in recent
times. Efforts to alleviate the problem depend upon accurate char-
acterization of these flows in both a quantity and quality sense.
This report describes the consideration that led to the design and
fabrication of a prototype automatic sewer sampling system speci-
fically intended for application in a storm or combined sewer. It
also presents the results of preliminary field testing and con-
trolled laboratory testing, and will be of interest to those who
have a requirement for the quality characterization of wastewater
flows.
Louis W. Lefke
Acting Director
Municipal Environmental
Research Laboratory
iii
-------
ABSTRACT
A brief review of the characteristics of storm and combined
sewer flows is given, followed by a discussion of the re-
quirements for equipment to sample them, noting features
that are desirable in such equipment and problem areas.
When considered from a systems viewpoint, there are five
functional subsystems. Design considerations for each of
these are discussed,,followed by a description of the de-
sign implementation used for each subsystem in the fabrica-
tion and assembly of a prototype automatic sewer sampling
system intended for storm and combined sewer application and
other adverse sewer flow conditions.
The prototype sampler is described from an installation and
operation viewpoint, and the results of preliminary field
testing are discussed. The device was also tested under con-
trolled laboratory conditions and found to be capable of gath-
ering reasonably representative samples (i.e., within 10%)
over a fairly wide range of flow characteristics, even for
particles somewhat outside the regime of Stokes' law. Four
different commercially available samplers were tested under
the same flow conditions in a side by side fashion. Their
behavior was rather erratic, and they were not able to gather
representative samples consistently. None of them was capable
of good performance when appreciable bed load was present.
Results from these commercial units ranged from an overall
understatement of pollutant loading by 25% or more, to over-
statements of 200% and more.
This report was submitted in partial fulfillment of Contract
Numbers 68-03-0155 and 68-03-0409 under the sponsorship of the
Municipal Environmental Research Laboratory, Office of Research
and Development, United States Environmental Protection Agency.
Work was completed in February, 1975.
iv
-------
CONTENTS
Section
I
II
III
IV
V
VI
VII
VIII
IX
X
Figure
1
2
3
4
Page
CONCLUSIONS 1
RECOMMENDATIONS 4
INTRODUCTION 5
SAMPLER DESIGN CONSIDERATIONS 10
DESIGN IMPLEMENTATION 30
INSTALLATION AND OPERATION 46
PRELIMINARY FIELD TESTING 54
CONTROLLED LABORATORY TESTING 59
REFERENCES . * 85
APPENDIX '..'." 87
LIST OF ILLUSTRATIONS
Page
Sediment Distribution at Sampling
Station 12
Region of Validity of Stokes' Law 14
Orientation Angles . . . ; 15
Effect of Sampling Velocity on
Representativeness of Suspended
Solids 17
Effect of Lateral Orientation of
Sample Intake 18
Prototype Automatic Sewer Sampling
System Schematic 31
-------
LIST OF ILLUSTRATIONS (Cont'd)
Figure
7
8
Prototype Sampler Intake Schematic
Page
'33
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Alternate Sampler Intake Schematic
for Large Channels 35
General Equipment Arrangement Inside
the Pump Housing . . . ' 39
Backflush Water Solenoid Valve and
Check Valve Manifold . • . ' 39
Fluidic Diverters and Air Control
Valves 40
Views of Sample Cup Mounting and
Weighing Control 42
Distribution Arm Drive Wheel 43
Electronic Control Panel and Case 44
Timing Diagram 49
Sampling Site for Prototype Sampler .... 56
General Test Facility Arrangement 61
Test Facility Details 62
Intake Location of Prototype Sampler ... 65
Relative Concentration Profile 67
Variation of Sand Concentration with
Time 68
Bed Load Effects 69
Variation of Gilsonite Concentration
with Time 71
i
Effect of Turbulence Enhancer- on Rela-
tive Concentration Gradient 72
Vt
-------
LIST OF ILLUSTRATIONS (Cont'd)
Figure
25
26
27
Effect of Turbulence Enhancer on
Variations with Time
Pumice Concentration Profiles
Sampling Representativeness .
Page
73
75
79
LIST OF TABLES
Table
Effect of Shape Factor on Hydraulic
Size (in CM/SEC)
Average Sampling
Representativeness (%)
Page
25
78
vii
-------
ACKNOWLEDGEMENTS
The cooperation and support of the Fairfax County Department of
Public Works and especially Mr. Richard Meyer is acknowledged
with deep appreciation. They allowed the use of the Scott Run
site for preliminary field testing of the prototype sampler
and provided assistance and encouragement throughout that phase
of the project.
Appreciation is expressed to the staff of LaSalle Hydraulic Lab-
oratory for their expeditious support of all aspects of the con-
trolled laboratory testing, and especially to Mr. F. E. Parkinson,
Vice President and Development Manager.
The support of this effort by the Storm and Combined Sewer Section
(Edison, New Jersey) of the USEPA, Municipal Environmental Research
Laboratory, Cincinnati, Ohio and especially Mr. Richard Field and
Mr. Hugh E. Masters, Project Officer, for their guidance, sugges-
tions and inputs, and thorough manuscript review is acknowledged
with gratitude.
viii
-------
PREFACE
Readers of .this report who are interested in more detail about
commercially available automatic liquid samplers, custom de-
signed units, and project and field experience with such devices
in storm and combined sewers are referred to a collateral effort
reported in an Environmental Protection Technology Series Report
entitled "An Assessment of Automatic Sewer Flow Samplers - 1975."
It presents a brief review of the characteristics of storm and
combined sewer flows followed by a general discussion of the pur-
poses for and requirements of a sampling program. A compendium
of 82 model classes covering over 200 models of commercially
available and custom designed automatic samplers is given with
descriptions and characterizations of each unit presented along
with an evaluation of its suitability for a storm and/or combined
sewer application. A review of field experience with automatic
sampling equipment is also given covering problems encountered
and lessons learned.
ix
-------
-------
SECTION I
CONCLUSIONS
4.
Storm and combined sewer flows are the most difficult
of all wastewater streams to accurately characterize
in a water quality sense due to the extremely wide
•range, of pollutants that can be found in them, the
possible spatial variations of these pollutants within
a given cross-section of the flow, the widely varying
flow rates commonly encountered that can even include
reverse, flow in certain situations, the frequent pres-
ence of significant bed load which may be highly pol-
luted, the presence of every imaginable kind of debris
which poses a physical threat to any obstacle in the
flow, and so on.
In view of the large number of highly varying param-
eters associated with the storm and combined sewer
application, a systems approach to the design of an
automatic sampler intended for such use is virtually
mandatory; the five essential elements of such a sys-
tem are the sampler intake subsystem, the sample
gathering subsystem, the sample transport subsystem,
the sample storage subsystem, and the controls and
power subsystem.
Based upon a functional analysis of each subsystem
of an automatic sewer sampler, in the light of the
present state-of-the-art, a number of improvements
are both warranted and achievable, including areas
such as intake design, sample intake and transport
velocity, sample train line sizes, and sample
capacity.
A prototype automatic sewer sampling system intended
expressly for the storm and combined sewer applica-
tion was designed and fabricated using a modular
approach, which allows the basic design implementation
to be tailored to suit a wide variety of sampling pro-
gram and site requirements. The design features
include all solid-state electronics, a clock to allow
time-of-day correlation, high sample intake and trans-
port velocities, minimum line size of 0.95 cm (3/8 in.)
inside diameter, large peristaltic pumps and fluidic
-------
diverters to avoid having any moving parts in the
sample train, the return of the first flow to waste,
a fresh water purge and backflush before and after
each sample is taken, multilevel sample intakes with
non-intrusive mounting, and large sample capacity.
5. The prototype sampler was given some preliminary field
testing in an actual sewer following extensive bench
testing and check-out. It performed very well over-
all with very little de-bugging required as a result
of actual field operation. It was discovered that
the pumps are powerful enough to pick up relatively
large stones that could damage the tubing in the pumps
and so intake screens with a maximum aperture size of
0.48 cm (3/16 in.) are used. It was also learned that
if an intake is not completely submerged, the result-
ing air bubbles in the flow may cause erratic fluidic
diverter action. Finally, attention must be given to
the orientation and routing of the sample return
(drain) line so that it does not kink as the sample
distribution arm rotates.
6. The prototype sampler was tested in synthetic sewage
flows under laboratory conditions that were controlled
as to flow velocity, depth, and suspended solids den-
sity, size, and concentration in order to evaluate its
ability to gather samples that are representative in
terms of suspended solids. Although good results were
achieved when most of the particles were in suspension
in the flow (even for those falling somewhat outside
the range of validity of Stokes1 Law), results tended
to be erratic when any appreciable quantity of mate-
rial was transported as bed load.
7. Four different types of commercially available automa-
tic samplers from four different manufacturers were
tested side by side with the prototype sampler under
selected, controlled laboratory conditions. Perform-
ances, as measured by laboratory analysis of their
samples, were quite erratic. Although the testing
was far from exhaustive, enough data were gathered to
demonstrate that there can be marked differences in
results obtained with different types of sampling
equipment, even under identical, controlled flow
conditions.
-------
10,
It was not the purpose of this brief series of compara-
tive tests to evaluate the different design approaches
used by commercial manufacturers of automatic sampling
equipment. The results of the tests do point out the
critical urgency for such work to be performed, however.
As suggested by other work and demonstrated by the pres-
ent effort, some water quality data now in existence may
be considered suspect, at least insofar as suspended sol-
ids are concerned. Thus, there is a need for determining
the capabilities of various types of sample collection
systems to gather representative samples of wastewater
flows over a wide range of characteristics. This is es-
pecially critical for equipment to be used in storm and
combined sewer flows which often contain large amounts
of heavier suspended solids.
The prototype sampler was not designed to sample bed
load or floatables in a truly representative sense,
and no claims should be inferred as to its ability to
do so. Such sampling is necessary, especially in a
storm or combined sewer, if one is to be able to account
for all pollutant loadings in a time-mass-discharge
sense, but no equipment exists today that is well suited
for these purposes.
The new technologies implemented in the design of the
prototype sampler have demonstrated efficacy, and should
be considered in the design of commercially available
equipment. One possible exception is the fluidic di-
verter. A simple pinch-tube valve arrangement could of-
fer similar design advantages at less cost. The
trade off lies between the likelihood of debris in the
flow interfering with the operation of the pinch valve
(i.e., preventing it from closing completely) and the
likelihood of air bubbles in the flow causing sluggish
fluidic diverter action.
-------
SECTION II
RECOMMENDATIONS
The ability of the prototype automatic sewer sampling
system to gather representative samples over a wide
variety of flow conditions and to operate in an actual
sewer setting has been demonstrated. It is recommended
that it be installed at some on-going USEPA demonstra-
tion project where comparative data are desired.
When combined with earlier field experience, the er-
ratic results obtained with different types of com-
mercially available automatic samplers under controlled
laboratory conditions points out the urgent need for
determining the capabilities of various types of sam-
ple collection systems to gather representative samples
of wastewater flows over a wide range of characteris-
tics. This must be done under controlled conditions
if results are to be quantified in any way other than
as relativistic comparisons. It is recommended that
a number of sample collection systems of the types
that represent the majority of present day equipment
be assembled and tested under controlled laboratory
conditions representing a wide range of wastewater
flow characteristics.
Representative sampling of bed load and floatables
(including oil and grease) continues to be an extremely
difficult problem. It is recommended that a program to
develop equipment that is suitable for these purposes
be initiated in the near future.
-------
SECTION III
INTRODUCTION
GENERAL
The phenomenal growth of urban areas in recent times and the
rapid expansion of industrial operations to meet society's
ever increasing demands for more goods, energy, etc., have
heightened the pollutional potential of man's existence and
have contributed to his increasing awareness of and concern
for his environment. One of the impacts of the population
explosion is that sanitary engineering practices that ap-
peared tolerable even as recently as a few decades ago are
no longer acceptable today in many locales. The pollutional
effects of stormwater and combined sewer overflows on re-
ceiving water quality are becoming less and less tolerable,
and a research program to mitigate or ameliorate the situa-
tion has been underway at the USEPA (and its predecessor
agencies) for the last sever'al years.
Knowledge of the character of the urban environment leads
one to the expectation that stormwater draining from it will
be of poor quality, and more and more data are being gath-
ered that verify this expectation. Although a discussion of
urban runoff quality is beyond the scope of this presenta-
tion, it should be noted here that the nature of urban run-
off, its temporal and spatial fluctuations, and the
characteristics of urban catchment systems all contribute
to make gathering accurate flow characterization data, both
quantity and quality, a very difficult task. Uncontrollable
parameters include meteorological and climatological factors,
topography, hydraulic characteristics of the surface and
subsurface conduits, the nature of the antecedent period,
and the land use activities and housekeeping practices
employed.
An automatic liquid sampler is a highly desirable tool to
employ in determining the water quality characteristics of
storm and combined sewer flows. A review (1) of the major-
ity of commercially available automatic liquid samplers re-
vealed that none had been specifically designed for this
severe application and, in fact, many manufacturers do not
recommend their equipment for such use. As a result, a new
automatic sewer sampling system was designed incorporating
several new technological advances.
If one considers all of the vagaries of the storm and com-
bined sewer sampling problem presented in (1), it is intui-
tively obvious that a single piece of equipment cannot
-------
exist that is ideal for all sampling programs in all storm
and combined sewer flows of interest. One can, however,
set down some general requirements for sampling equipment
that is to be used in a storm or combined sewer application.
EQUIPMENT REQUIREMENTS
The success of an automatic sampler in gathering a repre-
sentative sample starts with the design of the sampler in-
take. This obviously will be dependent upon conditions at
the particular site where the sample is to be extracted.
If one is fortunate enough to have a situation where the
sewer flow is homogeneous with respect to the parameters
being sampled, then a simple single point of extraction for
the sample will be adequate. In the more typical case, how-
ever, there may be a spatial variation in the concentration
of a particular constituent that is to be examined as part
of a sampling program, and then the sampling intake must be
designed so that the sample which is gathered will be nearly
representative of the actual flow. Several different de-
signs have been utilized by researchers in an attempt to
meet this objective, but none can be considered as ideal or
universally applicable.
The automatic sampler must be capable of lifting the sample
to a sufficient height to allow its utilization over a rather
wide range of operating heads. It would appear that a mini-
mum sample lift of 5 meters (15 ft) or so is almost manda-
tory in order to give a fairly wide range of applicability.
It is also important that the sample size not be a function
of the sample lift; that is, the sample size should not be-
come significantly less as the sample lift increases.
The sample line size must be large enough to give assurances
that there will be no plugging or clogging anywhere within
the sampling train. However, the line size must also be
small enough so that complete transport of suspended solids
is assured. Obviously, the velocities in any vertical sec-
tion of the sampling train must well exceed the settling
velocity of the maximum size particle that is to be sampled.
Thus, the sample flow rate and line size are connected and
must be approached together from design considerations.
The sample capacity that is designed into the piece of
equipment will depend upon the subsequent analyses that the
sample is to be subjected to and the volumetric requirements
for conducting these analyses. However, in general, it is
desirable to have a fairly large quantity of material on
hand, it being safer to err on the side of collecting too
much rather than too little. As a minimum, it would seem
that at least 500 mSL and preferably 1000 mH of fluid would
-------
be desirable for any discrete sample. For composite sam-
ples at least 4 to 8 liters should be collected.
The controls on the automatic sampler should allow some
degree of freedom in the operation and utilization of the
particular piece of equipment. A built-in timer is desir-
able to allow preprogrammed operation of the equipment.
Such operation would be particularly useful, for example,
in characterizing the buildup of pollutants in the early
stages of storm runoff. However, the equipment should also
be capable of taking signals from some flow measuring de-
vice so that flow proportional operation can be realized.
It is also desirable that the equipment be able to start
up automatically upon signal from some external device that
might indicate the onset of storm flow phenomena such as an
external rain gauge, flow height gauge, etc. Flexibility
in operation is very desirable.
A power source will be required for any automatic sampler.
It may take the form of a battery pack or clock type spring
motor that is integral to the sampler itself. It may be
pressurized gas, air pressurized from an external source,
or electrical power, depending upon the availability at the
site.
In addition to being able to gather a representative sample
from the flow, the sampling equipment must also be capable
of transporting the sample without pre-contamination or
cross-contamination from earlier samples or aliquots and of
storing the gathered sample in some suitable way. As was
noted in (1), chemical preservation is required for cer-
tain parameters that may be subject to later analyses, but
refrigeration of the sample is also required and is stated
as the best single means of preservation.
DESIRABLE FEATURES
In addition to the foregoing requirements of automatic sam-
pling equipment, there are also certain desirable features
which will enhance the utility and value of the equipment.
For example, the design should be such that maintenance and
troubleshooting are relatively simple tasks. Spare parts
should be readily available and reasonably priced. The
equipment design should be such that the unit has maximum
inherent reliability. As a general rule, complexity in
design should be avoided even at the sacrifice of a certain
degree of flexibility of operation. A reliable unit that
gathers a reasonably representative sample most of the time
is much more desirable than an extremely sophisticated com-
plex unit that gathers a very representative sample 10 per-
cent of the time, the other 90 percent of the time being
-------
spent undergoing some form of repair due to a malfunction
associated with its complexity.
It is also desirable that the cost of the equipment be as
low as practical both in terms of acquisition as well as
operational and maintenance costs. For example, a piece of
equipment that requires 100 man-hours to clean after each
24 hours of operation is very undesirable. It is also
desirable that the unit be capable of unattended operation
and remaining in a standby condition for extended periods
of time.
The sampler should be of sturdy construction with a minimum
of parts exposed to the sewage or to the highly humid, cor-
rosive atmosphere associated directly with the sewer. It
should not be subject to corrosion or the possibility of
sample contamination due to its materials of construction.
The sample containers should be capable of being easily
removed and cleaned; preferably they should be disposable.
PROBLEM AREAS
The sampler by its design must have a maximum probability
of successful operation in the very hostile storm and com-
bined sewer environment. It should offer every reasonable
protection against obstruction or clogging of the sampling
ports and, within the sampler itself, of the sampling train.
It is in a very vulnerable position if it offers any signif-
icant obstruction to the flow because of the large debris
which are sometimes found in such waters. The unit must be
capable of operation under the full range of flow conditions
which are peculiar to storm and combined sewers and this
operation should be unimpeded by the movement of solids
within the fluid flow. If the unit is to be designed for
operation in a manhole, it almost certainly should be cap-
able of total immersion or flooding during adverse storm
conditions which very frequently cause surcharging in man-
hole areas. It is also necessary that the unit be able
to withstand and operate under freezing ambient conditions,
and that it be able to withstand the high flow velocities
and the associated high momentums found in storm and com-
bined sewer flows.
Probably one of the most significant problem areas lies in
the attempt to gather a sample that is representative of
low as well as high specific gravity suspended solids. The
different momentum characteristics call for differing ap-
proaches in sampler intake design and in intake velocities.
Another problem area arises in a sampling program where it
-------
is desirable to sample floatable solids and materials such
as oils and greases as well as very coarse bottom solids
and bed load proper.
For samples which are to be analyzed for constituents which
require chemical fixing soon after the sample is collected,
there are other problems. Although it is true that the re-
quired amount of fixing agent could be placed in the sample
container prior to placing it in the field, for composite
samples in particular, where the eventual total sample is
built up of smaller aliquots gathered over an extended
period of time, the initial high concentrations of the fix-
ing agent as it becomes mixed with the early aliquots may
well be such as to render the entire sample unsuitable for
its intended purpose.
Another problem area is the selection of materials in the
sampling train, i.e., those that come in contact with the
sample in any way. Adsorption of certain pollutants by
these materials could affect sample representativeness, but
there is no universally acceptable material due to the wide
diversity of pollutant activities.
Probably the,greatest problem areas lies in the inability to
produce a single design that is all things to all people.
This means that a number of extensive trade-off studies are
required. Care must be taken, however, that the resultant
design is not so "optimized" (i.e., compromised) that it is
not really well suited for any particular application or use,
-------
SECTION IV
SAMPLER DESIGN CONSIDERATIONS
SAMPLER
In a system breakdown of an automatic sewer sampler by func-
tional attributes, it may be divided into five basic ele-
ments or subsystems. These will be identified and discussed
in turn from the viewpoint of their design considerations.
SAMPLER INTAKE SUBSYSTEM
The sample intake of many commercially available automatic
liquid samplers is often only the end of a plastic suction
tube, and the user is left to his own ingenuity and devices
if he desires to do anything other than simply dangle the
tube in the stream to be sampled. In the following para-
graphs we wish to examine the functions of a sampler intake
that is intended to be used in a storm or combined sewer ap-
plication and the design considerations that arise therefrom.
Pollutant Variability
A general discussion of the character of storm and combined
sewage is given in (1) where the variability of pollutant
concentration is also treated. We wish to consider the
latter factor here in somewhat more detail. Let us con-
sider first some empirical data from (2). In the study, a
special pressurized circulating loop was assembled contain-
ing a 25 cm (10 in.) square test section some 4.6m (15 ft.)
long. Careful measurements of the velocity contours were
made and near uniformity was observed. Since the vari-
ability of a pollutant will be a function of velocity vari-
ations (among other factors), it is of interest to note the
horizontal and vertical variations of sediment distribution
observed experimentally in this test section with its very
small velocity variation.
Four readily available commerical sands, differing princi-
pally in size, were used in the study. They are referred
to by mean particle size (50 percent finer by weight) as
0.45 mm, 0.15 mm, 0.06 mm and 0.01 mm. Observed sediment
10
-------
distribution for the three coarsest sands are indicated in
figure 1. For all practical purposes the 0.01 mm sand was
uniformly distributed. It should be noted here that the
vertical variation is probably enhanced due to the design
of the test loop, which would tend to enhance concentrations
of heavier particles to the outside (the bottom of the test
section in this case) due to the action of centrifugal
forces. Observations made in (3) markedly indicate this
effect.
The observation was made in (3) that, in addition to varia-
tions in sediment concentration within the cross-section at
a given time, the sediment concentration at any point in the
cross-section was highly variable with respect to time, es-
pecially for the coarser sediments (0.45 mm). This observa-
tion was also made in (2) where data are presented on
concentration variation with respect to time as a function
of sampling interval. The concentration of successive 20-
second samples was found to vary over a range of 37 percent
of the mean, and the concentration of successive 60-second
samples varied over a range of 10.5 percent. Such varia-
tions arise from the natural turbulence of the flow as would
be encountered in an actual sewer and from the non-uniform
nature of recirculated flows in test loops which is peculiar
to laboratory simulations.
So far we have focused our attention on relatively heavy
(specific gravity approximately 2.65) solids and their dis-
tribution in a flow. For the lighter organic solids with
specific gravities near unity, the particle distribution
will be more nearly uniform in a turbulent flow. It would
appear that one can expect a reasonable degree of uniformity
in the distribution of particles which fall in the Stokes'
Law range of settling velocities, i.e., for values of the
external Reynolds' number less than unity. If one describes
a .particTe "in terms of Tts hydraulic size W, defined as the
velocity of uniform fall in a fluid at rest, Stokes' Law can
be written as
W.
[1]
where d is mean particle diameter, s.g. is the specific
gravity of the particle material, v is the kinematic viscos-
ity of the fluid, and g is the acceleration of gravity. The
external Reynolds' number (so called because the linear di-
mension upon which it is based is a particle dimension
rather than a flow dimension) can be expressed as
Re = Wd/v
[2]
11
-------
UJ
o
OL.
Ul
O
O
O
1.2
1.1
1.0
0.9
o 0.8
E 0.7
A.
40RIZONTAL DISTRIBUTION AT MID-DEPTH
4321012345
INCHES FROM CENTER
0
1
2
D-
0 o
f_ 0
1 «
CO
LU
1 5
t— »
6
ac
t—
Q. 7
Ul /
a
8
9
\
V\
\
B.
\
\
\
VERT]
\
\
.
\\
N!
Vi
\s
I ^
jl
CAL D
[STRIB
s
o- c
A- C
X- C
D- C
\
V
\*
JTION
EDIME^
SIZE
IT SI
Mil
[REAM
.OCITY
.06 mm - 5 ft/sec
.15 mm - 5 ft/sec
1.15 mm - 3 ft/sec
.45 mm - 5 ft/sec
\
\
AT CEN
TER
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4. 1.6 1.8 2.0
pflTTn CONCENTRATION AT POINT
KAIJ.U -- CONCENTRATION AT CENTER
Figure 1. Sediment Distribution at Sampling Station*
* Taken from reference 2.
12
-------
Combining equations [1] and [2] we can express the range of
validity of Stokes* Law as
Re = gd3 (s.g.-l)/18v2 < 1
[3]
If one considers water at 16°C (60°F) as the fluid (v=1.217
~5 2
xlO ft /sec), a plot of equation [3] over the range of
interest is given in figure 2. Here it can be noted that,
within the range of Stokes' Law, the maximum particle diam-
eter for sand with a specific gravity of 2.65 is less than
0.1 mm while for organic particles with a specific gravity
of 1.05 it is about 0.3 mm.
Since the kinematic viscosity of water is temperature de-
pendent, the Stokes1 Law particle diameter limit will also
be a function of temperature. For sand, a decrease in water
temperature from the upper eighties to the mid-forties re-
sults in a 50 percent increase in the maximum particle
diameter.
Sampler Intake Functions
The operational function of a sampler intake is to reliably
allow gathering a representative sample from the flow stream
in question. Its reliability is measured in terms of free-
dom from plugging or clogging to the degree that sampler
operation is affected, and invulnerability to physical damage
due to large objects in the flow. It is also desirable,
from the viewpoint of sewer operation, that the sampler in-
take offer a minimum obstruction to the flow in order to
help prevent blockage of the entire sewer pipe by lodged
debris, etc.
Let us first consider the ability of the intake to gather a
representative sample of dense suspended solids in the sedi-
ment range, say up to 0.5 mm with specific gravity of 2.65.
The results of a rather thorough examination of rel-atively
small diameter intake probes 0.63 and 0.32 cm (1/4 and
1/8 in.) are given in (2). The argument is developed that,
for a nozzle pointing directly into the flow (figure 3a),
the most representative sample of a fluid/suspended-solids
mixture will be obtained when the sampling velocity is equal
to the flow velocity at the sampling point. Using this as
the reference criteria, investigations were conducted to
determine the effects of a) deviations from the normal sam-
pling rate, b) deviations from the straight-into-flow posi-
tion of the probe, c) deviations in size and shape of the
13
-------
Jill I I
co
U4
CO
I I I
CO
Ul
CO
I I I I
LU
Q.
CO
~ UJ
CJ
1—<
I—
-. G£
i 2
IS
n)
O
4-1
CO
M-l
O
4-1
•H
t3
O
•H
60
(U
CN
(1)
H
3
00
_• o o"
t-. 10 K>
— lf>
o q q q q q
(ww) a yniHwvia Hioiiavd
14
-------
3a. Normal Orientation
Directly Into Flow
3b . Orientation at an
Angle to Head-on
3c. Vertical Orientation (0°)- 3d. Horizontal Orienta-
Orifice in Flat Plate tion (90°) - Orifice in
Flat Plate
Figure 3. Orientation Angles
15
-------
probe, and d) disturbance of sample by nozzle appurtenances.
The effect of the sampling velocity on the representative-
ness of the sample is indicated in figure 4, which presents
the results for 0.45 mm and 0.06 mm sand. For the latter
size, which falls within the Stokes' Law range, less than
±4 percent error in concentration was observed over sampling
velocities ranging from 0.4 to 4 times the stream velocity.
For the 0.45 mm particles, the error at a relative sampling
•rate of 0.4 was +45 percent, and at a relative sampling rate
of 4 the error was -25 percent.
)
For probe orientations up to 20° to either side of head-on,
(figure 3b), no appreciable errors in concentration were
observed. Similarly, introduction of 0.381 and 0.952 cm
(0.150 and 0.375 in.) probes showed comparatively little
effect on the representativeness of the sample. The probe
inlet geometry, i.e., beveled inside, beveled outside, or
rounded edge, also showed little effect on the representa-
tiveness of the sample, when compared to the standard probe.
Finally, in instances where a sampler body or other appur-
tenance exists, the probe should be extended a short distance
upstream if a representative sample is to be collected. In
summary, it was found that for any sampler intake facing
into the tream, the sampling rate is the primary factor to
be controlled.
Tests were also run with the sampling intake probes in the
vertical position (figure 3c) to determine the effect such
as orientation had upon the representativeness of the sample.
With such intakes, the sample entering them must undergo a
90° change of direction, and consequently there is a tendency
for segregation and loss of sediment to take place. Tests
were run with the standard probe, a 0.63 cm (1/4 in.) diam-
eter orifice in the center of a 2.5 x 5.1 cm (1 x 2 in.)
flat plate oriented so that its longest dimension was in the
direction of flow, and with an orifice in a crowned (mush-
room shaped) flat plate 3.2 x 5.1 cm (1.25 x 2 in.). The
results all showed negative errors in concentration, increas-
ing with particle size and increasing with intake velocities
less than the stream rate but nearly constant for intake
velocities higher than the stream rate.
Since the smallest errors were found for the orifices in the
flat and mushroom shaped plates (whose performances were
nearly identical for intake velocities greater than one-half
the stream velocity), it was decided to investigate the ef-
fect of lateral orientation, i.e., to rotate the plate 90°
so that it might represent an orifice in the side of a con-
duit rather than in the bottom (figure 3d). The results for
0.15 mm sand are presented in figure 5. It can be noted
that while side orientation caused greater errors (as was
-------
CJ
o
CO
, CO
co
fi
CO
>
•H
4-J
cd
CO
0)
a.
CO
C
o
4J -H
•H i-H
O O
O CO
rH
CO T3
> CO
T3
«> a
C CO
•H a<
M CO
eu 3
g CO
cd
CO <4-l
o
4J
o
CO
(50
•H
CM
co
o
c
CO
co
M
0
O
n)
p
17
-------
I-tt-
Kt-
q
fi
o o>
- o
t-
o
in
d
•U 1
o
o
LU
o
MOI1V^1M33M03 M01J
NOIlVHiNaONOO HldWVS
0)
A!
n)
4J
0
S
rt
to
o
•H
4-J
rt
4-J
d
QJ
•rl
OJ
4-1
Cfl
4-1
O
0)
w
•
in
0)
o
0
0)
M
QJ
M-l
0)
M
s
o
M
M-l
0)
A!
n)
td
•p
tfl
O
18
-------
to be expected), these errors approached the nearly constant
error of the 0° orientation as the relative sampling rate was
increased above unity.
The work reported in (3) was a laboratory investigation of
pumping sampler intakes. Nine basic intake configurations,
all representing an orifice of some type in the side wall
of the flume, were examined. Sand sizes of 0.10 mm and
0.45 mm were used in the study. Reference samples were taken
with a probe located near the wall and pointing into the di-
rection of the flow. The reference sample intake velocity
was equal to the stream velocity. The primary measurement
was sampling efficiency, the ratio of the sediment concentra-
tion in the test sample to that of the reference sample com-
puted for a point 1.3 cm (1/2 in.) from the wall. The
reference sample was taken just before and just after the
test sample was gathered. Although the data exhibited con-
siderable scatter, several conclusions were drawn. With re-
gard to the intake velocity, greater than 0.9 m/s (3 fps) is
generally desirable and, for sands coarser than 0.2 mm, an
intake velocity equal to or greater than the stream velocity
is desirable. With regard to intake configuration, for intake
velocities greater than about 0.9 m/s (3 fps) the sampling
efficiencies showed little effect of size of intake (range
was 1.3 to 3.8 cm diameter), of rounding the intake edges,
or of shape and orientation of the axis of an oval intake. "
Sampling efficiency was found to decrease with increasing
particle size above 0.10 mm for all intakes tested. Similar
observations were made in field tests with river water sam-
ples at St. Paul and Dunning, Nebraska, reported in (4).
To summarize the foregoing as it relates to the sampler in-
take function of gathering a representative sample we note
the following:
1) It becomes difficult to obtain a one-to-one
representation, especially for inlets at 90°
to the flow, for large, heavy suspended
solids.
2) For particles that fall within the Stokes1
Law range, consistent, representative samples
can be obtained.
3) The geometry of the sampler intake has little
effect on the representativeness of the sample.
4) The sample intake velocity should equal or
exceed the velocity of the stream being
sampled.
19
-------
Sampler Intake Design
The foregoing suggest certain directions that the design of
a. sampler intake for storm and combined sewer flows should
take. At the outset, it appears unwise to attempt to sample
suspended solids that fall much outside the Stokes1 Law
range. A realistic maximum size for sand with specific
gravity of 2.65 would appear to be around 0.1 mm to 0.2 mm.
High sample intake velocities will ba required, perhaps in
excess of 3 m/s (10 fps), if the sample is to be
representative.
Although the flow may be nearly homogeneous, except for very
coarse solids and large floatables, more than one sample in-
take is desirable for reliability of operation as well as
insurance against some unforeseen gradient in the pollutant.
In view of the changing water levels in the conduit with
changing flows, the changing velocity gradients within the
flows, and the possibility of changing pollutant gradients
not only with respect to these but also with type of pollut-
ant; not even a dynamically adaptive sampler intake can be
designed to gather a sample that is completely representa-
tive in every respect at the same time. In the absence of
some consideration arising from the particular installation
site, a regular distribution of sampling intakes across the
flow, each operating at the same velocity, would appear to
suffice. Since the intakes should be as non—invasive as
possible in order to minimize the obstruction to the flow
and hence the possibility of sewer line blockage, it seems
desirable to locate them around the periphery of the conduit,
In order to prevent unwanted material (rocks, sticks, rags,
and similar debris) from entering the sampling train and
possibly clogging or damaging the equipment, the sampler in-
take must provide some sort of a screening function. Due
to the nature of stormwater and combined sewer flows, a fine
mesh screen does not seem desirable. Instead, a screen made
up of a number of rather large holes, say 0.3 to 0.6 cm
(1/8 to 1/4 in.) in diameter, appears more desirable.
SAMPLE GATHERING SUBSYSTEM
Three basic sample gathering methods or categories can be
identified; mechanical, forced flow, and suction lift. Sev-
eral different commercial samplers using each method are
available today. The sample lift requirements of the partic-
ular site often play a determining role in the gathering
method to be employed.
20
-------
Mechanical Methods
There are many examples of mechanical gathering methods used
in both commercially available and one-of-a-kind samplers.
One of the more common designs is the cup on a chain driven
by a sprocket drive arrangement. In another design, a cup
is lowered within a guide pipe, via a small automatic winch
and cable. Other examples include a self closing pipe-like
device that extracts a vertical "core" from the flow stream,
a specially contoured box assembly with end closures that
extracts a short length (plug) of the entire flow cross-
section, a revolving or oscillating scoop that traverses
the entire flow depth, etc.
Some of the latter units employ scoops that are characterized-
for use with a particular primary flow measurement device
such as a weir or Parshall flume and extract an aliquot vol-
ume that is proportional to the flow rate. Another design
for mechanically gathering flow proportional samples involves
the use of a sort of Dethridge wheel with a sample cup
mounted on its periphery. Since the wheel rotation is pro-
portional to flow, the effect is that a fixed volume aliquot
is taken each time a certain discharge quantity has passed,
and total discharge can be estimated from the size of the
resultant composite sample.
The foregoing designs have primarily arisen from one of two
basic considerations. First, site conditions that require
very high lifts have dictated the use of mechanical gather-
ing units due to the limitations of suction lift pumps and
space considerations. Some mechanical units are capable of
lifts of 61m (200 ft.) or more. Second, the desire to
gather samples that are integrated across the flow depth
has led to some of the different mechanical approaches men-
tioned above. Unless vertical velocity and pollutant gradi-
ents are quantified and accounted for, their presence makes
the results of such depth integrated samples questionable,
at least in a mass discharge sense.
One of the penalties that one must trade-off in selecting a
mechanical gathering unit is the necessity for some obstruc-
tion to the flow, at least, while the sample is being taken.
The tendency for exposed mechanisms to foul, together with
the added vulnerability of many moving parts, means that suc-
cessful operation will require periodic inspection, cleaning,
and maintenance.
21
-------
Forced Flow Methods
All forced flow gathering methods require some obstruction
to the flow, but usually it is less than with mechanical
gathering methods,. It may only be a small inlet chamber
with a check valve assembly of some sort or it may be an
entire submersible pump. The main advantage of submersible
pumps is that their high discharge pressures allow sampling
at greater depths, thereby increasing the flexibility of the
unit somewhat insofar as site depth is concerned. Pump mal-
function and clogging, especia'lly in the sizes often used
for samplers, is always a distinct possibility and, because
of their location in the flow stream itself, maintenance is
much more difficult and costly to perform than on above
ground or more easily accessible units. They also necessar-
ily present an obstruction to the flow and are thus in a
vulnerable position as regards damage by debris in the flow.
Pneumatic ejection is a forced flow gathering method used by
a number of commerical samplers. The gas source required by
these units varies from bottled refrigerant to motor driven
air compressors. The units that use bottled refrigerant
must be of a fairly small scale to avoid an enormous appetite
for the gas and, hence, a. relatively short operating life be-
fore the gas supply is exhausted. Furthermore, concern has
recently been expressed about the quantities of freon that
are being discharged into the atmosphere. The ability of
such units to backflush or purge themselves is also necessar-
ily limited. The advantages of few moving parts, inherent
explosion proof construction, and high lift capabilities
must be weighed against low or variable line velocities, low
or variable sample intake velocities, and relatively small
sample capacities in some designs. Another disadvantage of
most pneumatic ejection units is that the sample chamber
fills immediately upon discharge of the previous sample.
Thus, it may not be representative of flow conditions at the
time of the next triggering and, if paced by a flowmeter,
correlation of results may be quite difficult.
Suction Lift Methods
Suction lift units must be designed to operate in the envi-
ronment near the flow to be sampled or else their use is
limited to a little over 9m (30 ft.) due to atmospheric
pressure. Several samplers that take their suction InLft
22
-------
directly from an evacuated sample bottle are available to-
day. Vacuum leaks, the variability of sample size with
lift, the requirement for heavy glass sample bottles to
withstand the vacuum, difficulty of cleaning due to the
requirement for a separate line for each sample bottle, the
necessity of placing the sample bottles near the flow stream
(and hence in a vulnerable position), the varying velocities
as the sample is being withdrawn, etc., are among the many
disadvantages of this technique.
Other units are available that use a.vacuum pump and some
sort of metering chamber to measure the quantity of sample
being extracted. These units in some designs offer the
advantages of fairly high sample intake and transport veloc-
ities, the fluid itself never comes in contact with the
pump, and the pump output can easily be reversed to purge
the sampling line and intake to help prevent cross-
contamination and clogging. Their chief disadvantage,
shared with certain other suction lift designs arises from
the following consideration.
With all suction lift devices a physical phenomenon must be
borne in mind and accounted for if sample representative-
ness is to be maintained. When the pressure on a liquid
(such as sewage) which contains dissolved gases is reduced,
the gases will tend to pass out of solution. In so doing
they will rise to the surface and entrain suspended solids
in route. (In fact, this mechanism is used to treat water;
even small units for aquariums are commercially available.)
The result of this is that the surface layer of the liquid
may be enhanced in suspended solids, and if this layer is
a part of a small sample aliquot, the sample may not be at
all representative. In the absence of other mitigating
factors, the first flow of any suction lift sampler should
therefore be returned to waste.
A variety of positive displacement pumps have been used in
the design of suction lift samplers, including flexible im-
peller, progressive cavity rotary screw, roller or vane, and
peristaltic types. Generally these pumps are self-priming
(as opposed to many centrifugal pumps), but some designs
should not be bperated dry because of internal wearing of
rubbing parts. The desirability of a low-cost pump that is
relatively free from clogging has led many designers to use
peristaltic pumps. A number of types have been employed in-
cluding finger, nutating, and two- and three-roller designs
using either molded inserts or regular tubing. Most of
-------
these operate at such low flow rates, however, that the
representativeness of suspended solids is questionable.
Newer high-capacity peristaltic pumps are now available and
should find application in larger automatic samplers. The
ability of some of these pumps to operate equally well in
either direction affords the capability to blow down lines .
and help remove blockages. Also, they offer no obstruction
to the flow since the transport tubing need not be inter-
rupted by the pump, and strings, rags, cigarette filters
and the like are passed with ease.
All in all, the suction lift gathering method appears to
offer more advantages and flexibility than either of the
others for a storm or combined sewer application. The limi-
tation on sample lift can be overcome by designing the pump-
ing portion of the unit so that it can be separated from the
rest of the sampler and thus positioned within 6m (20 ft.)
or so of the flow to be sampled. For many sites, however,
even this will not be necessary.
SAMPLE TRANSPORT SUBSYSTEM
The majority of the commercially available automatic samplers
have fairly small line sizes in the sampling train. Such
tubes, especially at 0.3 cm (1/8 in.) inside diameter and
smaller, are very vulnerable to plugging, clogging due to
the build-up of fats, etc. It is also imperative that ade-
quate sample flow rate be maintained throughout the sampling
train in order to effectively transport the suspended solids.
In horizontal runs the velocity must exceed the scour ve-
locity, while in vertical runs the settling or fall velocity
must be exceeded several times to assure adequate transport
of solids in the flow.
The complexities inherent in the study of a two-phase mix-
ture such as soil particles and water are such that rigorous
analytical solutions have not yet been obtained except in
certain limiting cases such as the work of Stokes mentioned
earlier. The use of hydraulic size, which is the average
rate of fall that a particle would finally attain if falling
alone in quiescent distilled water of infinite extent, as a
descriptor for a particle involves its volume, shape, and
density. It is presently considered to be the most signifi-
cant measurement of particle size. However, there are no
analytical relationships to allow its computation; recourse
must be made to experiment.
-------
An excellent discussion of the fundamentals of particle
size analysis is given in (6). Table 1, which is taken from
data presented therein, illustrates the effect of shape fac-
tor on hydraulic size for sand particles with specific grav-
ity of 2.65 in water at 20°C. It can be noted that while
a sphere with a nominal diameter of 0.2 mm will fall only
about one-third faster than a similar sized particle with a
shape factor of 0.3; a sphere with a nominal diameter of
4.0 mm falls over 2-1/2 times faster than a particle with
SF=0.3. For curves showing temperature effects, correction
tables, etc., the reader is referred to (6).
TABLE 1. EFFECT OF SHAPE FACTOR ON HYDRAULIC
SIZE (IN CM/SEC)*
Nominal Diameter
(mm)
0. 20
0.50
1.00
2.00
4.00
Shape Factors
0.3
1.78
4.90
8.49
12.50
17.80
0.5
1.94
5.63
10.10
15.50
22.40
0.7
2.11
6.31
12.10
19.30
28.00
0.9
2.26
7.02
14.00
23.90
35.60
Spheres
2.43
7.68
15.60
28.60
46.90
Taken from reference 6,
The transport of solid particles by a fluid stream is also an
exceedingly complex phenomena, and no complete theory which
takes into account all of the parameters has yet been
formulated. Empirical formulae exist, however, some of
which have a fairly wide range of applicability. An ex-
pression for the lowest velocity at which solid particles
heavier than water still.do not settle out onto the bottom
of the pipe of channel has been developed by Knoroz (8) on
the basis of numerous experiments carried out under his di-
rection at the All-Union Scientific Research Institute for
Hydraulic Engineering.
-------
A somewhat simpler expression for the adequate self-
cleaning velocity of sewers derived by Camp from experimen-
tal findings of Shields as given in (9) is:
/6.4gd
R1/6 /0.8d(s.g.-l) [4]
where f is the friction factor, n is Manning's roughness
coefficient, and all other terms are as previously identi-
fied. Using equation [4], for example, it is seen that a
velocity of 0.6 m/s (2 fps) is required to adequately trans-
port a 0.09 mm particle with a specific gravity of 2.65 and
a friction factor of 0.025. By comparison, the fall veloc-
ity of such a particle is around 0.06 m/s (0.2 fps).
In summary, the sampling train must be sized so that the
smallest opening is large enough to give assurance that
plugging or clogging is unlikely in view of the material
being sampled. However it is not sufficient to simply make
all lines large, which also reduces friction losses, without
paying careful attention to the velocity of flow. For a
storm or combined sewer application, minimum line sizes of
0.95 to 1.3 cm (3/8 to 1/2 in.) inside diameter and minimum
velocities of 0.6 to 0.9 m/s (2 to 3 fps) would appear war-
ranted. Finally, sharp bends and twists or kinks in the
sampling line should be avoided if there is any possibility
of trash or debris in the sample that could become lodged
and restrict or choke the flow. The same is true of valve
designs. It also appears desirable to deliver the sample
under pressure all the way from the pump to the sample con-
tainer to further help reduce clogging.
SAMPLE STORAGE SUBSYSTEM
For storm and combined sewer applications, discrete sampling
is generally desired. This allows characterization of the
sewage throughout the time history of the storm event. If
the samples are sufficiently large, manual compositing can
be performed based on flow records or some other suitable
weighting scheme. Although the quantity of samples re-
quired will be a function of the subsequent analyses that
are to be performed, in general at least 1 liter and pre-
ferably 2 liters will be desired. An additional benefit
arises because such relatively large samples are less vul-
nerable to errors arising from cross-contamination.
26
-------
The sample container itself should either be easy to clean
or disposable. The cost of cleaning and sterilizing makes
disposable containers attractive, especially if bacteriolog-
ical analyses are to be performed. Although some of today's
better plastics are much lighter than glass and can be auto-
claved, they are not so easy to clean or inspect for clean-
liness. Also, the plastics will tend to scratch more easily
than glass and, consequently, cleaning a well-used container
can become quite a chore. The food packaging industry, es-
pecially dairy products, offers a wide assortment of poten-
tial disposable sample containers in the larger sizes. Both
the 1.9& (1/2 gal.) paper and plastic milk cartons can be
considered viable candidates, and their cost in quantity is
in the pennies-each range.
The requirements fbr sample preservation are enumerated in
ClO) and will not be repeated here except to note that re-
frigeration is stated as the best single preservation and
will, in all likeiihood, be required unless the sampling
cycle is brief and samples are retrieved shortly after being
taken. It should be mentioned, however, that if the samples
are allowed to become too cold, they may no longer be repre-
sentative. For example, destruction of the organisms neces-
sary for the development of BOD may occur, or freezing may
cause serious changes in the concentration of suspended
solids. Light can also affect samples and either a dark
storage area or opaque containers would seem desirable. Un-
less disposable containers are used, however, it will be
difficult to inspect an opaque container for cleanliness.
Again the paper milk carton is attractive since not only is
it relatively opaque, but its top opens completely allowing
visual inspection of its contents.
CONTROLS AND .POWER SUBSYSTEM
The control aspects of some commercial automatic samplers
have come under particular criticism as typified by comments
reported in (1). It is no simple matter, however, to provide
great flexibility in operation of a unit while at the same
time avoiding all complexities in its control system. The
problem is not only one of component selection but packag-
ing as well. For instance, even though the possibility of
immersion may be extremely remote in a particular installa-
tion, the corrosive highly-humid atmosphere which will, in
all likelihoodj be present makes sealing of control elements
and electronics desirable in most instances.
-------
The automatic sampler for storm and combined sewer applica-
tion will, in all likelihood, be used in an intermittent
mode; i.e., it will be idle for some period of time and
activated to capture a particular meteorological event. If
field experience to date is any indication, the greatest
need for an improved control element is for an automatic
starter. While the sensor is not a part of the sampler
proper, its proper function is essential to successful
sampler utilization. Although remote rain gages, etc., can
be used for sensing elements, one of the most attractive
techniques would be to use the liquid height (or its rate of
increase) to start a sampling cycle. This will avoid the
difficulties associated with different run-off times due to
local conditions such as dryness of ground, etc.
The controls determine the flexibility of operation of the
sampler, its ability to be paced by various types of flow
measuring devices, etc. Built-in timers should be repeatable
and time periods should not be affected by voltage varia-
tions. The ability to repeatedly gather the required ali-
quot volume independent of flow depth or lift is very
important if composite samples are to be collected. Pro-
visions for manual operation and testing are desirable as
is a clearly laid out control panel. Some means of deter-
mining the time when discrete samples were taken is
necessary if synchronization with flow records is contem-
plated. An event marker could be desirable for a sampler
that is to be paced by an external flow recorder. Reli-
ability of the control system can dominate the total system
reliability. At the same time, this element will, in all
likelihood, be the most difficult to repair and calibrate.
Furthermore, environmental effects will be the most pro-
nounced in the control system.
The above tasks can probably be best executed, in the light
of the current electronics state-of-the-art, by a solid
state controller element. The unit should-be of modular con-
struction for ease of modification, performance monitoring,
fault location, and replacement/repair. Such an approach
also lends itself to encapsulation which will minimize en-
vironmental effects. Furthermore, solid state controllers
can be easily designed with sufficient flexibility to accept
start commands from a variety of types of remote sensors,
telephone circuits, etc. Finally, one of the attributes
essential to the control system of an automatic sampler to
be used in a storm or combined sewer application is that it
be able to withs-tand power outages and continue its program.
Such power interruptions appear to be increasingly common as
demand for electricity continues to grow.
28
-------
The foregoing discussion as it relates to problems asso-
ciated with interruptions in electrical service is, of
course, directed to samplers that rely upon outside power
for some aspect of their operation. The need for high sam-
ple intake and transport velocities, larger sample lines
and capacities, together with the possible requirement for
mechanical refrigeration make it unlikely that such a sam-
pler can be totally battery operated today. Other approaches
to self-contained power such as custom designed wet-cell
packs, diesel generators, etc., while within the current
state-of-the art, introduce other problems and complexities
that must be carefully weighed before serious consideration
can be given to their incorporation in an automatic sampler
design.
39
-------
SECTION V
DESIGN IMPLEMENTATION
GENERAL
From the foregoing discussion, it can be seen that it is
indeed a formidable task to design a single piece of equip-
ment that is eminently suitable for all storm and combined
sewer applications. For this reason, a functional module
approach was taken in the design of a prototype automatic
sewer sampling system aimed at overcoming deficiencies or
inadequacies in much of the extant equipment. Using this
building block type approach, wastewater samplers can be
assembled from basic system components, but tailored for
a particular installation or program.
The intention was not to develop a "finished" design that
was production engineered and ready for quantity manufac-
turing but, rather, to embody as many desirable features
and incorporate as many new technologies as possible, in
order to be able to assess whether or not the anticipated
benefits had, in fact, been realized. Consideration was
given to ultimate producibility, however, in that cost con-
siderations were heavily weighed. For example, a pumping
system that appeared attractive in many respects was re-
jected because of its nearly $1,500 cost. Even given quan-
tity and original equipment manufacturer (OEM) discounts,
it was felt that the price could never be reduced to a
practicable level. Similar considerations were obtained
in the design of the control subsystem. The near revolu-
tion in cost and availability of digital integrated circuit
devices was anticipated at the time of the design (mid 1973),
but implementation techniques available today were not ob-
tainable at that time. Functionally, however, the control
subsystem can be essentially duplicated today in a much
more compact package. It might not be necessary to incor-
porate all the provisions for adjustment that were felt
desirable for the prototype in a production model, however.
The various elements of the prototype automatic sewer sam-
pling system are schematically indicated in figure 6. Each
of the functional subsystems identified in Section IV will
be discussed in turn as implemented in the prototype unit.
30
-------
31
-------
SAMPLER INTAKE SUBSYSTEM
Due to the virtual mandate that there be no rigid obstruc-
tion to the flow in view of the very large solids that are
often encountered, the prototype sampler intake for use in
smaller sewer pipes, say up to 1.5m (5 ft) in diameter, is
comprised of a number of nozzles located around the circum-
ferential periphery of the pipe. Because of the modular
design of the prototype system, the actual number of in-
takes employed can be varied depending upon the size of the
pipe in which the sampler is to be used. The use of a num-
ber of discrete intake nozzles is to allow for redundancy
in the event of clogging of an individual nozzle by rags,
paper, etc., and also contributes to the gathering of a
more representative sample when vertical stratification of
the flow is encountered.
The prototype intake was designed for installation in a nom-
inal 0.6m (2 ft) diameter sewer pipe of arbitrary material.
For this size range, four intake nozzles were provided. Each
plastic intake nozzle (figure 7a) has a slot which is aligned
with three 0.5 cm (3/16 in.) diameter holes or ports in the
mounting strap that are spaced linearly in the -direction of
the flow. These serve a preliminary screening function to
keep out large, unwanted solids such as stones that might
possibly damage other portions of the unit. The four intake
nozzles are spaced around the periphery, two on each side of
the pipe as indicated in figure 7b.
As shown in the schematics of figure 7, the 30.5 cm (12 in.)
wide stainless steel mounting strap encloses the nozzles and
intake tubing to help protect them from damage due to the
movement of solids in the flow. The leading and trailing
edges are split back for 7.6 cm (3 in.) and bent down
slightly, forming "fingers" that contact the inside wall of
the sewer pipe and grip it securely when the turnbuckle is
tightened. The intake lines are routed to and along the
crown of the pipe into the manhole area. The "ring" formed
by the sampler intake projects less than 2.5 cm (1 in.) from
the wall of the pipe and, thus, does not offer undue resist-
ance to the flow. The intake would normally be mounted ap-
proximately 0.6-0.9m (2-3 ft) up the pipe from the manhole
in the direction of the flow. This can be accomplished by
a single technician in 15 minutes or less, thus, minimizing
the need for flow interruption.
For installation in larger rectangular conduits, especially
those with open top sections, a different mounting arrange-
ment would be desired. Cylindrical plastic intake screens
32
-------
FLOW DIRECTION
INTAKE PORTS
PLASTIC
INTAKE NOZZLE
SEWER WALL
7a. Cross Section through an
Intake Nozzle
STAINLESS STEEL
MOUNTING STRAP
SPACE FOR
TUBING
STAINLESS
STEEL
MOUNTING
STRAP
MOUNTING
TURNBUCKLE
SAMPLER INLET
NOZZLES
SEWER
PIPE
7b. Installation Schematic
Figure 7. Prototype Sampler Intake Schematic
33
-------
perforated with a large number of 0.5 cm (3/16 in.) holes
could be used on the end of each intake line. When mounted
on a rod hinged to a tethered float (the pump box could be
used for this purpose) and extending at an angle to the
channel floor (figure 8), samples could be taken at constant
percentages of depth regardless of flow level. Alternately,
the intake screens could be weighted and suspended by the
tubing at appropriate depths. In either scheme, the intake
is free to swing away if struck by a solid object in the
flow, which will minimize damage possibilities.
SAMPLE GATHERING SUBSYSTEM
The suction lift gathering method was chosen for the proto-
type automatic sewer sampling system. Among the features
wanted in the pump were a self-priming capability; the abil-
ity to operate dry without damage; the ability to operate
either forward or reverse to allow for pressurized back-
flushing and blowdown; the ability to pass strings, papers,
cigarette filters, etc., with ease; and a minimum inside
diameter of around 1 cm (3/8 in.).
The foregoing requirements led to the further investigation
of peristaltic pumps in the larger sizes. The principal
requirement was a capacity of at least 7.6 £pm (2 gpm) to
assure adequate sample intake and transport velocities.
Although not strictly a, peristaltic tubing pump, the Vanton
flex-i-liner sealless plastic pump was also given prelimi-
nary consideration. In this valveless unit, which avoids
the use of gaskets, pumping is accomplished by a rotor,
mounted on an eccentric shaft, which rotates within a liner
creating a progressive squeeze action on the fluid trapped
between the liner and the body block. Thus, the only two
parts in contact with the fluid being pumped are the outer
surface of the low cost moulded flexible liner and the inner
surface of the plastic body block. Both of these components
are available in a variety of materials. This pump's action
resembles that of a single-lobed peristaltic pump with a
very oblong moulded tube inside. Cost and high power re-
quirements, e.g. a 375¥ (1/2 HP) motor is required for a
single pump capable of delivering 7.6 Upm (2 gpm) against
a 2.5 kg/sq cm (36 psi) head, led to dropping this unit
from further consideration, but the concept is available in
models that deliver up to 151 £pm (40 gpm).
34
-------
\\\\\\
PUMP ENCLOSURE
OR FLOAT
A. PLAN VIEW
X \ \\ \. V\ \ \ \ V\u\> \
S AMPLE LINES TO REMAINDER
OF UNIT
\\\\\\\\
B. ELEVATION
Figure 8. Alternate Sampler Intake Schematic for
Large Channels
35
-------
There are a number of types of peristaltic pumps available
on the market including finger, nutating, and two- and
three-roller designs using either moulded inserts or regular
tubing. The nutating type was eliminated because of its
typical low flow rate (1.5 £pm maximum). Finger type pumps
were given more serious consideration in view of the capa-
bility of some designs to handle up to four tubes simulta-
neously. In order to obtain the desired flow rate, however,
it is necessary to use tubing with a minimum inside diameter
of 1.9 cm (3/4 in.). This is undesirable because of the
desire to minimize changes in line size, and flow velocities
in such large tubes are inadequate to assure good solids
transport. Furthermore, these large units are expensive
(from $450 to over $1,200, depending upon features) and re-
quire large motors, e.g. 525W (3/4 HP).
A number of two-rotor designs were investigated. Some of
these are now available in large sizes, e.g. up to 78 £pm
(21 gpm), but are priced from $375 to $565 per head without
motor. The alternative is to use a smaller pump at a higher
rotational speed. Such pump heads cost less than $100. A
typical unit was extensively tested and found to be gener-
ally satisfactory with two exceptions. First, the flow was
highly pulsating, and when the intake line was deliberately
plugged, rather violent motions resulted. Second, and even
more important, tubing life was quite short (e.g., 6-8 hours)
for all but small lifts.
During the course of the pump investigation, a new high-
capacity three-rotor peristaltic pump was introduced. Until
that time, all three-rotor designs had been of insufficient
capacity to warrant serious consideration. One of these
units was extensively tested and found to be very satis-
factory in all regards. The pump easily passed such delib-
erately introduced debris as cigarette filters, string,
paper, rope, rags, etc. It was possible to clog the intake
line by packing it with paper toweling and compacting it
with a steel rod while the pump was off. When run in the
forward direction, the pump was unable to d'islodge such an
artificial plug (ultimately the tubing would collapse), but
when operated in reverse in the backflush mode, the pump
was always able to clear the line.
\
Based upon favorable testing experience, the quality of
workmanship, favorable design features including the ability
to "stack" pump heads, and low cost (under $70 per head),
the decision was made to utilize this pump in the prototype
design. The four pump heads used in the prototype were
36
-------
especially built for this application by the manufacturer,
the nonstandard feature being the staggered alignment of
shaft keys and rotors of the respective heads so that the
load from each head would be more evenly distributed as
seen by the motor. After considerable experience with the
prototype, it is not felt that this latter sophistication
is necessary.
These pump heads are rated at approximately 9.5
(2.5 gpm) at their normal operating speed of approximately
600 rpm. They use 1 cm (3/8 in.) inside diameter fubing
with a 1/3 cm (1/8 in.) thick wall. Experience to date has
indicated that, even with this high capacity operation,
tubing life on the order of 30 to 40 hours can be expected.
It requires less than 30 minutes to change the tubing in
the present design, and so the standard maintenance proce-
dure is to install fresh tubing after 20 hours of pump
operation. This corresponds to some 20 storm events.
The pump heads are driven through common keyed shafts which
eliminate all requirements for interconnecting gears, pul-
leys, etc. Although the prototype unit uses four sample
lines, this number can be increased or decreased simply by
adding or deleting pump heads. The pumps are driven by an
inexpensive 250 watt (1/3 horsepower) electric motor through
a simple V-belt drive. They are completely enclosed in a
sealed plastic box that can withstand immersion without
leaking. Marine connectors and fittings have been used
throughout. This means that the pumps and motor can be lo-
cated near the actual flow to be sampled and operate even
though the sewer becomes surcharged or flooded. As a pro-
tective measure, a liquid level sensor has been included
inside the pump/motor housing. It will automatically shut
the system off should liquid begin to accumulate inside the
housing due to pump tubing failure or a leak due to housing
damage. The general equipment arrangement inside the
housing is shown in figure 9 (the plastic cover is removed) .
The anodized aluminum mounting plate serves as a heat ex-
changer as well as the main structural member of the housing,
A small "muffin" fan provides air circulation over the plate
and motor, resulting in quite acceptable ambient tempera-
tures within the sealed box.
SAMPLE TRANSPORT SUBSYSTEM
The sample transport subsystem of the prototype sampler con-
sists of four identical but discrete sample trains. Each
begins with a flexible Tygon tube connected to the intake
37
-------
at one end and the pump housing at the other. Since the
pump housing is located very close to the flow to be sam-
pled, these suction lines are relatively short. Inside
the pump housing, 0.95 cm (3/8 in.) I.D. silicone tubing
with a 0.32 cm (1/8 in.) thick wall is used. These lines
are approximately 0.6m (2 ft) long. The connecting lines
between the pump housing and the sample distribution arm
are 1.3 cm (1/2 in.) I.D. Both semi-flexible PVC pipe
(for longer runs) and tygon tubing (for relatively short
runs) have been successfully used for this purpose, de-
pending upon the test site conditions. This is the maxi-
mum recommended size in order to keep sample transport
velocities high enough, over 1.2 mps (4 fps) at the 10 &pm
(2.5 gpm) pumping rate, to assure adequate transport of all
suspended solids.
In order to allow for backflushing to minimize cross-
contamination of samples and to help dislodge any matter
that might be blocking an intake nozzle, a fresh water
tank is connected via a solenoid valve and manifold (fig-
ure 10) to each sample line. A 208& (55 gal) drum is used
for this purpose and is mounted on a stand above the sample
distribution arm. A check valve in each line assures that
the fluidic diverters will function when some lines are not
flowing full due to lower water levels in the sewer or
partial blockages, and provides a redundant assurance that
sewage cannot be pumped into the fresh water tank should
the solenoid valve malfunction.
In order to minimize the possibility of clogging or mal-
function due to strings, etc., in the flow, fluidic divert-
ers were custom designed to avoid the necessity of using
valves or other moving parts in the sample lines. Func-
tionally, the fluidic diverters allow the pump flow normally
to be returned to waste. When a sample is to be taken, the
diverters are automatically switched to fill the sample cup.
This is accomplished by using a small solenoid-operated air
valve to control aspiration of the fluidic diverter control
ports. These are illustrated in figure 11.
Short sections of Tygon tubing are used at the discharge
side of the fluidic diverters. The biased side (normally
return to waste) lines lead to a manifold block on the
distribution arm, from which a single, large tube carries
the flow back to the sewer. The lines from the sample
side of the fluidic diverters are held by the distribution
arm so that on signal they will discharge directly into the
next sample container.
38
-------
Figure 9. General Equipment Arrangement Inside the
Pump Housing
Figure 10. Backflush Water Solenoid Valve and
Check Valve Manifold
39
-------
Figure 11. Fluidic Diverters and Air Control Valves
40
-------
Sample Storage Subsystem
For characterization of storm and combined sewer flows,
discrete sampling is almost mandatory. In order to provide
a sufficient quantity of sample for a rather complete labo-
ratory analysis, the equipment has been designed to provide
approximately 2& (1/2 gal) samples. Since there are a
number of individual lines each operating independently,
varying flow stages and possible plugging may result in one
or more lines not containing full flow. In view of this, a
provision was incorporated for weighing each sample con-
tainer to assure that the desired amount of sample had, in
fact, been taken. This is accomplished by mounting each
sample cup in an individual counterweighted arm (figure 12)
that is pivoted near its middle. A tab on the arm contacts
a, cantilever spring connected to a limit switch which is
mounted on the distribution arm. When a precalibrated
quantity of sample has been gathered in the cup, the weight
of the fluid overcomes the spring tension, thereby closing
the limit switch which causes the fluidic diverter to switch
the flow back to waste.
The fluidic diverters and discharge lines are located on a
distribution arm that rotates 350° to fill the 12 sample
cups. This rotation is accomplished via a small electric
motor and friction drive wheel (figure 13). After filling
the twelfth cup, the arm contacts a limit switch that shuts
the entire system down until manually reset by an operator.
The sample containers themselves are 2.52, (84 oz) disposable
plastic cups with leakproof snap-on lids. These have a
panel that allows field annotations to be written with al-
most any writing instrument including pencil, ballpoint or
felt tip pens, etc.
Controls and Power Subsystem
Completely solid-state electronic control elements have been
used throughout in the prototype design. They are housed
within a weather—proof case (figure 14), which also contains
an isolation transformer to preclude any possible shock
hazard while working in the sewer with the pump motor. Con-
trols are adjustable in all aspects of the operation, and
logic sequences are protected against upset due to power
outages. The design will accept an automatic start command
from a suitable external' sensor. The unit is designed to be
paced by its built-in timer, but may also be paced by an
41
-------
Figure 12.
Views of Sample Cup Mounting And
Weighing Control
42
-------
Figure 13. Distribution Arm Drive Wheel
43
-------
Figure 14. Electronic Control Panel and Case
44
-------
external flowmeter where 'one is available. A clock is in-
cluded in the design to allow time synchronization with flow
records where desired. One-hundred-ten volt AC electrical
power is required to operate the unit. The actual set up,
programming, and operation of the unit are described in
Section VI.
45
-------
SECTION VI
INSTALLATION AND OPERATION
INSTALLATION
The sample intake ring is made of stainless sheet steel and
sized to fit a 61 cm (24 in.) diameter sewer line. Its di-
ameter can be adjusted by two stainless steel turnbuckles
to make it fit securely inside the sewer pipe. Four intakes,
positioned between the upstream and downstream ramps of the
ring, are connected to the pump enclosure through hoses that
are routed to the uppermost point of the ring and along the
top of the sewer line to the manhole area. The pump enclo-
sure, although watertight, should be located and secured
above the normal high water level if possible (to prevent
buffeting damage) and oriented within 10° of level. However,
this enclosure should not be placed at an elevation greater
than 4.6m (15 ft) above the lowest intake. Use hose clamps
on all tubing connections.
Route the four output tubes and the pump motor electrical
cable from the pump enclosure to the table set-up location,
and attach the tubes to the diverter connections and the
electrical cable to the control box. Connect the drain re-
turn tube to the table and to prevent kinks or any severe
flow restrictions, carefully route it back to the sewer.
Locate the freshwater flush tank between 0.9 and 1.5m (3 and
5 ft) above the elevation of the top of the sample distribu-
tion arm and fill with water. The electrically operated
stop valve and four check valves are attached thereto.
Route the four flush lines to the "T" connections in the di-
verter lines avoiding sagging places where water could set-
tle. Connect the stop valve electrical cable to the control
box. Place the sample table such that it is within 2° of
level. If necessary, to avoid kinking the drain line, ele-
vate the table several inches by placing blocks under the
three support legs. Connect the electrical leads to the
control box. Place sample containers in the twelve basket
containers and manually lift slightly and rotate the arm
such that the four sample tube outlets are over the No. 1
container. Connect the control box power cable to 110 VAC.
The system is now ready for automatic or manual sampling as
described in the following paragraphs.
46
-------
AUTOMATIC LIQUID SAMPLING SYSTEM OPERATION
Timing Overview
In operation, the system generally performs as follows. The
initial conditions are:
1. Sample bottles empty and in place.
2. Distribution arm set in position number 1.
3. Backflush tank full.
4. All components interconnected and checked out
(ready light on).
Upon signal, the sampling sequence begins with the pump
turned on in the reverse mode to provide a blow-down that
will force any fluid in the lines out through the.intake
nozzles to clear the sampling train of any possible block-
ages, previous samples, and/or backflush water. After a
predetermined (0-60 seconds) time period, the pump is momen-
tarily stopped and then started in the forward direction.
The pump is now extracting sewage from the flow, pumping it
through the diverters on the distribution arm, and returning
the fluid to waste through the drain provided. After a pre-
set time (0-5 minutes), which is selected to assure that a
representative flow is passing through the diverter, the di-
verter is switched to the sample position and its output
flows directly into the sample container. When a predeter-
mined weight of sample (up to 2 liters) has been- collected,
a switch closes and the diverter redirects the flow to
waste. As a backup measure, in the event of a blockage or
some other flow restriction, the diverter is also returned
to waste at the end of a predetermined time period. This
is simply to ensure that the system does not hang up trying
to fill one particular sample container.
Two events now occur in parallel. In the first, the back-
flush valve is opened and, after a momentary dwell, the pump
is started in the reverse direction. This allows the com-
plete sampling train to be backflushed for a preset time
period to purge the lines of sewage and dislodge any possi-
ble blockage that may have occurred. At the end of this
predetermined time period (0-60 seconds), the pump will turn
off and the freshwater valve will close.
In the second event, the index motor will move the distribu-
tion arm assembly to the next sample container position to
be ready for the next cycle, which is a repeat of the above
sequence. After the twelfth sample container has been
47
-------
filled, a switch will shut the complete system off, and it
will no longer function until the system has once again
been initialized.
Timing Sequence
The timing diagram shown in figure 15 indicates the logic
operation of the system. Each element is explained in the
following discussion. There are manual overrides and manual
switch controls which allow for system initialization and
for exercising the system in whatever manner may be desired
for testing. Certain simple precautions must be taken dur-
ing the initialization in that some timed events include
other events and, if the events are timed too low, they will
not occur, e.g., if the PT-2 timer is set too high and the
diverter timer is set too low, a sample will not be obtained
because the PT-2 time will shadow the diverter time and the
diverter will never switch.
Sample Interval - The total sample time includes the time,
Tl, set on the Sample Timer, CT-1 (0-15 minutes) and the
time, T2, set on the Cycle Timer, CT-2 (.016 sec-9 hours).
The total time between samples can be adjusted to approxi-
mately 9 hours.
Pump Motor - The pump motor is turned on momentarily, T3, in
reverse by the Blow-Down Timer, PT-2 (0-60 sec); when the
Blow-Down Timer de-energizes, the motor will stop for a pe-
riod of time, T7, set by the Motor Delay Timer (0-10 sec).
The off time set by the Delay Timer is to allow the fields
in the motor to collapse with power off. At the end of the
delay time the motor will start in the forward direction.
The motor will pump forward until CT-1 times out; then after
the delay time, T7, determined by the Delay Timer it will
reverse and backflush the system with fresh water until the
Back-Flush Timer, PT-1 (0-60 sec) times out.
Blow-Down Timer - The Blow-Down Timer is used to reverse the
pump to purge the intake lines of remaining fresh water and
any sewer liquid or blockage that might be in the sample
lines as well as to dislodge material which may be blocking
intake orifices.
Back-Flush Timer - The Back-Flush Timer reverses the motor
and opens the freshwater flush valve after the Sample Timer
times out. It controls the back flush time after the sample
has been taken.
Diverter Timer - The Diverter Timer (0-5 minutes) is used to
switch the output of the diverter from the drain position to
48
-------
CM
CM
UJ
O
>-
I i
-ES2T
H-S
i
I
"«^
^N
6
a
M
W)
o
M
•H
'&
•H
m
H
60
UJ
o
«=c
oo
CM
I
O
O
o
_J
CQ
o;
UJ
c±i
n:
oo
O
I—
O
X
.LU
a
CQ
a:
o
H-
O
Q_
s:
^
D-
49
-------
the sample position when a representative sample flow has
been obtained. When the sample container is filled, as de-
termined by sample weight, the diverter is returned to the
drain position. If the container is not filled when CT-1
times out, as might be the case when all intakes are above
the fluid level in the sewer or all intakes are blocked,
the diverter is also switched to drain.
Flush Valve - The Flush Valve is on the freshwater tank and
is opened by PT-1 to allow fresh water to purge the system.
Index Motor - The index motor moves the index arm assembly
from one sample position to the next in approximately
13 seconds. This time is not adjustable. The index motor
will turn on when CT-1 times out and moves the arm to the
next position where it stops. After the twelfth position,
the automatic cycling will stop.
Timer Adjustment - The initial set up of the Sample Cycle
Timer must allow sufficient time to ensure that a represent-
ative sample will be collected. This time must include the
time required to blow-down the lines, to stop the pump motor,
and to reverse its rotation direction so that it will pump a
sample from the sewer. This time should also be long enough
so that a sample container will be filled if only one line
is collecting a sample.
The Blow-Down Timer and the Motor Delay Timer must be set
for a time period that is less than the Diverter Timer time
period so that the pumps are pumping samples through the di-
verters before they are switched to the sample containers.
The Diverter Timer should be adjusted to a time long enough
to allow the Blow-Down and motor change of direction to oc-
cur, and short enough to enable it to collect a sample be-
fore the Sample Cycle Timer times out.
The Back-Flush Timer plus the Motor Delay Timer should be
adjusted for a total time of less than 13 seconds. If the
total time is greater than 13 seconds, the index arm will
pass the next position and there will not be a sample col-
lected for that position.
The Delay Cycle Timer may be set for any time desired from
1 minute to 9 hours; however, the minimum time must be
greater than the Back-Flush time plus the Motor Delay time
to allow the previous cycle to be completed. The Delay
Cycle Time determines the time between samples.
50
-------
Typical Timer Settings
Tl - 45 seconds
T2 - 1 hour
T3 - 10 seconds
T4 - 20 seconds
T5 - 10 seconds
T6 - 13 seconds (nonadjustable)
T7 - 2 seconds
At the start of a series of samples, and prior to automatic
enabling of the system, the system should be manually back-
flushed with fresh water to clear the lines of any possible
blockage.
Manual Mode Operation
The following steps will manually operate the system:
1. Power switch on.
2. Push power trip button twice (light on).
3. Using index motor switch, visually position
four sample hoses directly over a container.
4. Pump control right switch to reverse position.
5. Pump control left switch to on position.
6. Flush valve switch to full on position (up).
7. Wait approximately 10 seconds then place flush
valve switch to full down position.
8. Place pump control left switch to auto
•position.
9. Place pump control right switch to auto
position.
10. Place pump control left switch to on position.
11. Wait approximately 10 seconds then hold di-
verter switch in on position until desired
51
-------
sample quantity is collected in container.
Release diverter switch.
12. Place pump control left switch in.auto position.
13. Repeat steps 4 through 9.
14. Turn power switch to off position.
System has completed one,cycle and is ready to be set up in
manual or automatic modes.
Automatic Mode Start-Up Sequence
The following steps will manually start an automatic twelve-
sample sequence:
1. Set up timing typically as indicated earlier.
2. Power on.
3. Push power trip button twice (light on).
4. Reset (ready light on).
5. Perform operations 4 through 9 as above.
6. Operate automatic enable switch and release.
7. Operate manual start switch.
The system will proceed to take 12 samples arid automatically
shut down. Repeat step 5 above to flush system.
8. Turn power off.
Sampler Initiation From External Signal
The sampler may be operated in one of two preset modes upon
initiation from an external signal.
Up to twelve on-command samples may be taken by setting the
single/multiple sample switch to the "single" position and
supplying a 1- or 2-second short circuit between the two
leads connected to the external jack mounted on the control
box. Set up panel as follows:
1. Perform steps 1 through 4 as in automatic mode
above.
2. Reset start time counter to zero.
52
-------
3. Actuate set timer switch, noting time and date
for later correlation.
4. Perform steps 4 through 9 as in manual mode
above.
5. Operate automatic enable switch and release.
System is now ready to take samples on command from remote
source.
If it is desired to collect twelve consecutive samples at
preset time intervals, such as during a storm event, the
single/multiple switch is set to "multiple" and a 1- to
2-second short circuit signal, as above, is supplied. The
first sample will be taken at the end of the first interval
as set on T2 with subsequent samples taken at this same in-
terval T2. When using this mode, the clock timer mounted in
the control panel will start running when the system is set
up and will stop when the first sample is taken. This will
provide the time count required to determine when the sam-
ples were taken. Set up panel as directly above, steps 1
through 4.
53 ,
-------
SECTION VII
PRELIMINARY FIELD TESTING
PURPOSE
Subsequent to the fabrication, assembly, and laboratory
checkout of the prototype automatic sewer sampling system,
it was taken to a nearby site for preliminary field test-
ing. The purpose of this brief series of tests was to
demonstrate the ability of the prototype to reliably gather
samples in an actual field setting prior to conducting con-
trolled laboratory testing to determine its ability to
gather representative samples. Up to this point in time,
the prototype had not been used with actual sewage, and
it was felt that such use would be beneficial to verify
its operational characteristics, especially with regard
to freedom from plugging or clogging, etc.
SITE AND INSTALLATION
The site selected for the preliminary field testing was on
the Scott Run sewer line which carries sewage from an upper
portion of Fairfax County, Virginia under the Potomac River
to a main trunk in Maryland leading to the Blue Plains
treatment plant in the District of Columbia. The actual
sampling site was at the final metering vault located near
Virginia State Route 193, just before the line goes under
the Potomac. Although under the jurisdiction of the Fairfax
County Department of Public Works, the site is actually on
Federal park land. The metering vault is entirely below
ground, and access to it is controlled by a lock on the
vault hatch as well as by a chain link fence with a locked
gate that surrounds the vault area.
Thus, the site offered accessability (a truck could be
driven to within a few meters of the gate), safety (the
site was very well maintained by Fairfax County), protec-
tion from possible vandalism, a ready source of fresh
water (Scott Run was nearby), and electric power. The
sewer itself was a 61 cm (24 in.) diameter concrete line
which typically carried flows of around 5.7 to 22.7 mil-
lion liters per day (1.5 to 6 MGD). There was a slight
drop where the sewer entered the metering vault, and a bar
screen just upstream of the Parshall flume. The sampling
54
-------
point was between the entry to the vault and the bar screen
(see figure 16) in a fairly turbulent region. The pump
housing was placed on the vault floor as shown, which re-
sulted in a lift requirement of slightly less than one
meter. The rest of the system was located above ground
resulting in a pump discharge head of around 5 meters.
OPERATION
For the first few days of testing, the system was used with
no intake other than the open end of each intake tube which
was fastened, facing vertically down, to a strut which
could be mounted in the flow. This was done intentionally
in order to see what plugging or clogging effects might be
observed. In earlier attempts to deliberately choke the
complete system by artificially introducing a variety of
foreign objects, the only difficulty encountered was with
a nominal 0.48 cm (3/16 in.) diameter piece of hemp rope
approximately 30 cm (12 in.) long. Although this repre-
sents a rather unlikely piece of debris to be naturally
acquired by the sampler (more than one attempt was required
to artificially introduce it in the lab), it was passed by
the pump without difficulty. It did prove too large and
stiff to enter the diverter line, however, and as a result
simply oscillated between the check valve in the fresh
water line and the diverter take-off. The system continued
to operate, but it is felt that the rope could have acted
as a partial filter to some suspended solids resulting in
a slight depletion of them in the sample delivered to the
container by that particular line.
In the field testing with the open end of the tube as an
intake, ho difficulties were encountered the first day.
On the second day, however, one line picked up a relatively
large stone, major dimension approximately 0.63 cm (1/4 in.)
The flow at this time was high in heavy suspended solids as
witnessed by visual examination of collected samples. The
entire bottom of the sample container was covered with set-
tleable solids within less than a minute after collecting
the sample. When the stone was ingested it traveled as
far as the pump, where it ripped the tubing, causing a leak
inside the pump housing. This occurr.ed near the end of the
day, and so the entire pump housing was returned to the
laboratory for cleaning, inspection, and tube replacement.
On the next day of testing the sampler was installed as
before, but monitored closely from within the meter vault.
55
-------
Figure 16. Sampling Site for Prototype Sampler
56
-------
No difficulty was encountered. Fairly early in the fol-
lowing day the operator observed another large stone moving
up one of the intake lines and into the pump housing. Be-
fore the system could be shut down, it entered the pump
and stalled the motor, again tearing the tubing. Careful
examination revealed no damage to the motor or pump rollers
or body other than the ripped tube.
Although it was interesting to note that such large stones
(with their high momentum characteristics) could in fact
be picked up by the sampler, they are of little use in the
typical analysis of wastewater. To avoid a repetition of
such incidents, a plastic intake screen was fabricated and
used during the remainder of the field testing. This in-
take screen was a simple cylinder approximately 10 cm
(4 in.) in diameter and 30 cm (12 in.) long and containing
a large number of 0.48 cm (3/16 in.) diameter holes in its
lateral surface to assure free circulation of sewage.
Using this screen, no further problem with unwanted debris,
stones, etc., was experienced. The intake tubes terminated
at different depths within the screen so that effects of
varying water levels could be simulated by repositioning
the screen in the vertical.
The tests to simulate various water levels were run with
one, two, and three intakes above the water level, and the
prototype system worked well under all of these conditions.
A slight problem exists when an intake is deliberately
positioned just at the water-air interface. Since the
water surface is not tranquil, the intake is alternately
submerged and in the air, with the result that the sample
stream in that line is a mixture of air and sewage. The
fluidic diverter action is more sluggish or even erratic
when the flow stream is high in entrained air. One of two
extreme results may occur. If the amount of sewage versus
air is sufficiently low, the diverter for that line simply
will not switch'from drain. This means that the sample
collected will have no contribution from that particular
intake position, a desirable effect since such a surface
skimmed sample may not be as representative as those taken
from deeper in the flow. A similar effect may occur when
a submerged intake line is partially blocked resulting in
a sample stream that is free of entrained air but moving
at a low flow rate. In laboratory tests this phenomenon
was observed at flow rates around 3.8 £pm (1 gpm) or less.
Again this is seen as beneficial, since such lower velocity
flows may not be as representative in the heavier suspended
solids.
57
-------
The second result occurs when there is relatively more
sewage than air but still a considerable quantity of
bubbles in the line. Under such conditions the diverter
may switch, but its action is very sluggish. This means
that when the sample container is filled to the desired
level and the signal is given to switch the diverters
back to drain, the diverter in the line with entrained air
may not switch immediately, and that line will continue to
empty into the sample container. . The result is that the
sample container will be overfilled or, in some cases, even
overflow slightly. In cases where such overflows occurred,
they did not appear to be large enough to affect sample
representativeness due to the resulting decanting action.
They did require an additional c^ean-up step at the end of
a complete event cycle, namely the wiping of the table under
the sample container that had overflowed.
One other minor difficulty was encountered during the pre-
liminary field testing of the prototype automatic sampling
system. Since the fluidic diverters are mounted on the
rotating distribution arm, the single tube that returns the
flow to waste must undergo a rotation of approximately 350°
over the course of a complete event cycle. During one in-
stallation test series at Scott Run the system overflowed,
an event that was subsequently traced to a kink that had
developed in this drain line and restricted the flow
returning to waste. Consideration was given to the instal-
lation of a swivel fitting, but it was found that the prob-
lem did not develop if careful attention was paid to the
position and routing of the drain tube at installation.
In summary, the prototype automatic sewer sampling system
appeared to work well under all conditions encountered
during the preliminary field testing when the maximum ad-
missable particle size was reduced to 0.48 cm (3/16 in.)
with an intake screen. There were no instances of plugging
or clogging, and all automatic aspects of the control sys-
tem functioned satisfactorily. Since the quality charac-
teristics of the flow were unknown, none of the collected
samples was subjected to laboratory analysis. Visual
inspection revealed a surprising quantity of light, float-
able solids on several occasions as well as the heavy,
settleable solids mentioned earlier. As a result of the
overall satisfactory operation of the prototype system,
it was decided to proceed with the controlled laboratory
testing in order to examine the ability of the system to
extract a representative sample from flows of known charac-
teristics .
58
-------
SECTION VIII
CONTROLLED LABORATORY TESTING
OBJECTIVES
Although considerable data have been collected that purport
to reflect the ability of a particular piece of automatic
sampling equipment to gather a representative sample from a
wastewater stream, in no instance have the actual quality
characteristics of the stream been known. Similarly, there
have been a few side-by-side comparative tests run using
automatic samplers of different designs and manufacturers,
always with differences in results, but since the wastewater
streams used were ones of opportunity having unknown charac-
teristics, only relative comparisons were possible. For
example, Harris and Keffer (11) report differences of over
200 percent among samplers of different designs when applied
simultaneously to the same wastewater stream. In view of
this state of affairs, it had been decided to perform a
series of tests using a synthetic wastewater stream which
could be controlled as to such parameters as flow depth and
velocity and suspended solids characteristics including
specific gravity, size, and concentration.
There were three primary objectives to the controlled labo-
ratory testing activities. First, the extreme difficulty
in creating a tightly controlled (e.g., ±5%) synthetic waste
stream was well recognized. Both continuous-flow closed-
loop tests and open-loop, batch-fed tests had been performed
in the past, primarily in relation to sediment transport and
sampling studies, and considerable data scatter was observed,
Therefore, one objective was to determine the capability to
produce stable, controlled flows containing synthetic sus-
pended solids that simulated those found in storm and
combined sewers.
The second objective arose from the design and fabrication
of the prototype automatic sewer sampling system. It was
desired to quantify and evaluate the capability of this
device to gather samples representative of the wastewater
stream in question, especially as regards suspended solids,
under various flow conditions. Again the use of a con-
trolled, synthetic flow seemed to be the only means of
obtaining absolute rather than relative data.
59
-------
Finally, it was desired to comparatively evaluate a number
of the more popular commercially available automatic sampler
designs in a side-by-side fashion with the prototype sampler,
This would indicate whether or not, in fact, any actual
improvements in the ability to gather a representative sam-
ple had been achieved with the prototype design. A second
benefit, of course, would be the opportunity to compare
these commercially available designs with each other to see
which ones were best suited for gathering samples that are
representative as to suspended solids, and to do this
against a known base.
FACILITY
A number of requirements for a facility to be used for the
sampler controlled laboratory testing were identified.
These included a test flume with a semi-circular invert, the
ability to create stable flows over the velocity range of
0.3 to 2.4 m/s (1 to 8 fps) while maintaining a constant
depth in the test section, an accurate means of determining
flume discharge (i.e., flow velocity), a method for con-
stantly adding synthetic suspended solids to the flow so as
to create and maintain a known solids concentration in the
flume, the ability to provide suitable synthetic solids
representative of those encountered in storm and combined
sewer flows, and the ability .to provide laboratory analysis
of samples taken during the testing program.
The LaSalle Hydraulic Laboratory Ltd. (LHL) facility was
used for the controlled laboratory testing. Briefly, it
consisted of a water supply taken from a fixed pumping sta-
tion in the laboratory, the flow channel or flume itself, a
settling basin with a calibrated overflow weir, and an exit
to a return channel to the pump. The flow channel was
12.2m (40 ft) long with a cross-section 0.3m (1 ft) wide by
0.6m (2 ft) deep, including a semicircular invert (see fig-
ures 17 and 18). A test section was provided 3.7m (12 ft)
from the downstream end where a 2.54 cm (1 in.) recess in
the wall was provided to allow routing the 1.6 cm (5/8 in.)
O.D. tubes from the intakes of the prototype sampler to its
pump box. This was done, in view of the channel width, to
minimize any effects of these lines on the flow stream it-
self. Point gages upstream from the test section could
measure the water level to ensure that depth control was
maintained.
60
-------
Channel Outlet
Showing Diversion
Chute and Catch Basin
General View of
Test Channel
Figure 17. General Test Facility Arrangement
61
-------
1
View Inside Chute
Showing Cross Section
and Turbulence
Enhancer in Place
Solids Materials
Injection Vibrator
Figure 18. Test Facility Details
62
-------
The normal water supply was taken directly from the pump.
The flume and overflow weir were calibrated by a temporary
supply which took its flow directly from one of the cali-
brated V-notch weir towers in the laboratory. The solids
injection system consisted of a dry solids vibratory feeder
with a plexiglass hopper fixed over it. The rate of vibra-
tion (and hence solids injection) could be controlled by a
rheostat. A turbulence enhancer made of sheet metal with
alternating twisted blades was used for many of the runs in
an attempt to reduce vertical concentration gradients.
A wide range of suspended solids were available for creating
the synthetic flows. They were:
a. Silica sand, specific gravity 2.65
fine - 120 mesh _> d _> 140 mesh (.105-. 125 mm)
medium - 30 mesh _> d >_ 35 mesh (.500-. 595 mm)
coarse - 10 mesh _> d _> 12 mesh (1.68-2.00 mm)
b. Pumice, specific gravity 1.35
A single broad grain size distribution used in
earlier storm and combined sewer flow synthesiza-
tion was tested;
6 mesh ^_ d >_ about 100 mesh (.149-3.36 mm)
c. Gilsonite, specific gravity 1.06
fine - 12 mesh >^ d _> 30 mesh (.595-1.68 mm)
medium - 10 mesh 2L d ±. 12 mesh (1.68-2.00 mm)
coarse - 6 mesh >^ d j>_ 8 mesh (2.38-3.36 mm)
-d. Alathon, specific gravity 0.99
Uniform size of 3.0 mm
e. Polythene, specific gravity 0.92
Uniform size of 4.0 mm
These synthetic suspended solids may be considered as repre-
senting, either directly or through hydrodynamic simulation,
a wide range of suspended solids typically encountered in
storm and combined sewer flows. For example, the sand is
typical of many mineral and soil particles, sediments washed
from construction projects, etc.; the pumice may be consid-
ered representative of a host of heavier organics, rubber
and asphalt particles,' etc.; the gilsonite is representative
of the lighter organic suspended solids, many plastics, etc.;
and the alathon and polythene represent the range of many
floatables, organic as well as inorganic. For the majority
of the testing program, fine sand and medium gilsonite were
used, for reasons that will be discussed subsequently.
63
-------
INITIAL PROTOTYPE SAMPLER TEST RESULTS
The initial phase of the controlled laboratory testing pro-
gram involved using the prototype automatic sewer sampler
over a wide range of test parameters (flow velocity, solids
type and size, and concentration). The positions of the
four sampling intakes used for the large majority of the
testing are indicated in figure 19. As can be noted, two
were located near mid-depth while the other two were posi-
tioned near the water surface and near the invert respec-
tively. This arrangement is essentially similar to the
preferred three-point method of sampling discussed in (12).
The notion is that the mean of the upper half of the con-
centration will be approximated by the average of the surface
and mid-depth values, and the mean of the lower half by the
average of the mid-depth and bottom. Thus, as noted in (12),
"a composite sample made up of two samples from mid-depth
and one each from the surface and bottom, all of equal vol-
ume, approximately represents both concentration and size
distribution."
In addition to the prototype sampler, a so-called "reference"
sampler was used in part of this initial testing phase. The
reference sampler consisted of a "standard" sedimentation
probe (provided through the courtesy of the Federal Inter-
Agency Sedimentation Project Office at St. Anthony Falls)
connected to a peristaltic pump with a variable speed drive
arrangement. This reference sampler had been calibrated so
that any desired sample intake velocity could be set in
order to allow isokinetic sampling to be achieved. The in-
take probe could be positioned at,any desired point in the
cross-section of the flow with its inlet pointed directly
upstream into the flow. The reference sampler was used
primarily to investigate vertical and horizontal concen-
tration profiles.
As mentioned earlier, a number of suspended solid types were
tested only briefly. The two floatables, alathon and poly-
thene, were essentially confined to the surface layer of the
flow, and the prototype sampler, with its uppermost intake
around 3.4 cm (1.35 in.) below the surface, was unable to
gather a representative sample. The quantities of these
solids in the samples were so small that they could virtually
be counted by hand. The prototype sampler was not designed
for use with floatables, and these brief tests serve to
point out the need for specially designed intakes if repre-
sentative samples of floatables are to be gathered. Deter-
mining mass discharges (say Kg per day) will be complicated
64
-------
Note: Intake height above invert as a percent of total
water depth as indicated in parenthesis.
Figure 19. Intake Location of Prototype Sampler
65
-------
by sharp vertical concentration gradients and the fact that
transport rates of surface-carried solids may well differ
from those of the main stream due to vertical velocity
gradients.
The coarse sand was tested only briefly also. Even at flow
velocities of 2.4 m/s (8 fps), virtually all of the material
was transported as bed load. Very little could be gathered
with the reference probe when it was positioned over 1.3 cm
(1/2 in.) above the bed load. Consequently, samples taken
with the prototype sampler were not at all representative.
As long as the bed load depth was below the lowest intake,
no appreciable quantity of these solids would be picked up.
Conversely, when the bed load level reached the lowest in-
take, a sand slurry was collected, and the combined sample
would reflect concentrations an order of magnitude or more
greater than those of the flume.
A similar effect, although less pronounced, was observed
when using the medium sand. A relative concentration pro-
file is presented in figure 20. The mean concentration is
the average of four samples taken with the prototype sampler.
Their variation with time is also indicated in figure 20.
The relative concentration profile is the plot of points
representing the concentration of a sample taken at a par-
ticular depth with the reference sampler divided by the
mean concentration. As can be noted, the concentration
gradient is very sharp, the average occurring only some
6.4 cm (2.5 in.) above the invert.
The variation of sand concentration with time is better
illustrated by figure 21. Here the results of samples taken
by the prototype sampler at 60 second intervals are presented
for both fine and medium sand with a flow velocity of 1.2 m/s
(4 fps) in the flume. The standard deviation for the medium
sand is 202, almost twice the 110 for the fine sand. The
wider range of the results using medium sand is attributed
to the fact that at this velocity, a good portion of it
moves as bed load, and the more random dune movement near
the vicinity of the lowest intake was at least partially
responsible for the larger variation in results.
There was also considerable bed load build up with fine sand
at 1.2 m/s (4 fps). This is best illustrated by figure 22
which presents the results of a brief look at bed load
effects. Here a fairly heavy feed rate (a little over
1200 mg/Jt equivalent concentration) was established, and the
66
-------
M-I
O
fd
O
•H
4-J
n)
0
•H
1-1
-------
i
U- SI'S-
•H
H
4-1
•H
13
P5
O
•H
4J
4-1
(3
cu
CJ
(3
O
O
M-l
O
c
O
rt
•H
!-l
cd
CM
0)
M
S
00
•r)
s
§
CO
8
8
10
8
8
CVJ r—
68
-------
(20,180) (34,262) (24,985)
o
o
8,000
7,000
6,000
5,000
. 4,000
3,000
2,000
1,000
FINE SAND (.105
-------
prototype sampler was used in normal operation to take sam-
ples at 30 second intervals as the bed load began to ap-
proach the lowest intake. The procedure was then repeated
with the sample from the lowest intake being directed into
one container and that from the upper three intakes directed
into another. They were analyzed separately, and the results
are shown in figure 22. It can be noted that the upper three
intakes contribute virtually nothing to the overall sample
concentration. The combined results are obtained by adding
three times the concentration gathered by the upper intakes
to the concentration gathered by the lowest intake and
dividing by four to reflect the weighting that would be pro-
vided by normal operation. The curve so obtained is seen to
be qualitatively similar to the earlier results from normal
operation and reflects the effect of bed load inundating the
lowest intake.
Limited testing was also performed with the coarse and fine
gilsonite. Much of the fine gilsonite tended to remain in
the surface layer (surface tension effects outweighed
gravitational ones) and, consequently, could not be sampled
effectively. It also tended not to settle out in the set-
tling chamber, and material would blow over the weir. There-
fore, further use of this material was abandoned. There did
not seem to be any great differences in the results obtained
using coarse and medium gilsonite so, due to the greater
availability of the latter, it was used extensively. The
variation of gilsonite concentration with time is depicted
in figure 23 for flume velocities of 0.6 and 1.2 m/s (2 and
4 fps). The flume concentration was normally determined 'by
taking five samples directly from the shaker feed at the be-
ginning and end of a run. The average of these ten readings
was used as the flume concentration. The shaded areas in
figure 23 indicate the range within which these calibration
samples fell for each run.
As mentioned earlier, a turbulence enhancer was installed
to attempt to reduce the vertical concentration gradients.
It did not produce as great a change as had been hoped for
as can be seen in figure 24, which indicates relative con-
centration versus depth for medium gilsonite both with and
without the turbulence enhancer installed.. There was also
no great effect on the relative concentration variation with
time as is indicated by figure 25. None the less, it was
decided that the turbulence enhancer would be used for the
remainder of the tests.
70
-------
C3
CO
o
o
CO
to
co
a.
03
s
co
Q.
CM
I
a)
e
•H
H
a
o
•H
4-J
4-1
a
0)
o
s
o
o
0)
4J
•H
a
o
CO
r-l
•H
CD
a
o
•H
4->
cd
•H
(U
M
3
00
•H
O
o
CO
o
o
ID
O
CD
O
O
CM
NOIlVHiN30NOO
71
-------
MEDIUM GILSONITE (1.68
-------
CM
VO
^{°. -1
CM UJ
-------
Finally, a relatively small amount of testing was conducted
using ungraded pumice. Typical vertical concentration pro-
files taken with the reference sampler for two similar flume
concentrations (near 600 mg/£) are shown in figure 26. The
variation of samples taken with the prototype sampler at
30 second intervals from the actual flume concentration is
also depicted in figure 26. The average of each 12-sample
set is indicated as a straight line for each concentration.
The main thing to note here is that the prototype sampler
tended to understate the pumice concentration at this low
velocity. For higher velocities, the average concentration
from the sampler tended to approach that of the flume.
COMPARATIVE TESTING RESULTS
In view of the growing quantity of data suggesting that the
use of automatic samplers of different designs may result
in widely varying results, see Harris and Keffer (11), it
had been decided to devote a few days of controlled labo-
ratory testing to side-by-side comparisons of some of the
more popular models of commercially available equipment.
This would allow direct comparisons to be made of perform-
ance in a known stream. The prototype automatic sampling
system was also used to ascertain if any actual improvements
over commercially available equipment had been achieved as
regards the ability to gather a representative sample with
respect to suspended solids.
There is a wide variety of gathering methods and different
design implementations available in the automatic sampler
equipment marketplace today. For the limited .comparative
testing phase of this project, it was decided to use
examples of some of the more popular gathering methods. It
must be emphasized that the intention was not to study a
particular design in detail but, rather, to attempt to as-
certain how successfully the selected models would perform
when sampling suspended solids. The gathering methods em-
ployed by the four selected commercial samplers were (1) slow
speed peristaltic pump, 0.15 £pm (0.04 gpm); (2) moderate
speed peristaltic pump, 1.5 S-pm (0.4 gpm); (3) pneumatic
ejection; and (4) vacuum pump.
Each of the commercial units was equipped with a different
type of intake. The pneumatic ejection device comes equipped
with a special intake that must be used in order for the unit
to operate. The two peristaltic pump devices were equipped
with weighted cylindrical screens for intakes, and the vacuum
pump unit simply had a brass weight on the end of the intake
74
-------
o o
-r, *
V|M
o
o
to
o
o
co
0)
t-H
•H
U-l
O
^
PH
O
•H
; ij
cd
S-l
w
C
0)
o
C3
o
o
0)
o
•H
vO
CM
0)
V-l
bO
•H
•o
o
75
-------
hose. These latter three intakes could be modified by a
prospective user to meet his particular needs, but the pneu-
matic ejection unit's intake must be used as supplied.
In order to avoid any suggestion of equipment misuse, each
of the manufacturers invited to participate in the testing
were asked to provide their own personnel to operate and
maintain their respective equipment. Similarly, each manu-
facturer was allowed to install his intake in the test chan-
nel in whatever way he felt would be most advantageous to
his particular piece of equipment.
Space limitations precluded the simultaneous installation
of all four commercial samplers at the same time. There-
fore, two were tested during the week of July 22 and the
other two were tested during the week of July 29. The same
test parameters were used for both pairs of commercial
samplers, and the prototype sampler was used throughout
both series of tests.
Based upon the results of the earlier tests that had been
run using the prototype sampler, it was decided to concen-
trate on fine sand and medium gilsonite for the comparative
testing. Average flow velocities of 0.6 and 1.2 m/s (2 and
4 fps) were used with the gilsonite and 1.2 and 2.4 m/s
(4 and 8 fps) with the sand. Two concentrations, nominally
300 and 600 mg/£, were used for each velocity/solids com-
bination. Five samples were taken with each device for each
combination of parameters. The raw data are plotted in the
Appendix.
One measure of performance of an automatic sampler is its so
called "sampling efficiency" or, as it will be termed here,
sampling representativeness. This number, which is usually
expressed on a percentage basis, is the concentration of the
sample gathered by the unit divided by the actual concentra-
tion of the flow. Thus, exact one-to-one representation
results in a sampling representativeness of 100 percent.
Sampling representativeness numbers of less than 100 percent
indicate that concentrations (and hence mass discharges,
total pollutant loadings, etc.) will be understated, while
numbers greater than 100 percent mean that concentrations
are being overstated.
Obviously a device that consistently performs with a sampling
representativeness of 100 percent is ideal. However, a de-
vice that consistently performs with a sampling representa-
tiveness of, say, 80 percent can also be very useful. The
-------
important thing is its repeatability and its consistency of
representativeness, which can be viewed as a sort of cali-
bration factor.
Summary results of the comparative testing phase of the pro-
gram are presented in terms of sampling representativeness
averages in Table 2. The unaveraged sampling representa-
tiveness data are plotted in figure 27. Since it was not
the intention of this project to endorse or critize any par-
ticular piece of sampling equipment, the commercial units
are simply designated as Models A, B, C, and D. A few in-
terpretative comments are in order.
The overall performance of the prototype sampler appears to
be relatively good, especially for the gilsonite flows. For
the fine sand its performance was more erratic. At 1.2 m/s
(4 fps), much of the solids were flowing along the invert
as bed load and were beneath the level of the lowest intake
port, accounting for the general understatement of concen-
tration at this velocity. When the flow velocity was doubled
to 2.4 m/s (8 fps), the quantity of sand transported as bed
load was reduced, and higher concentrations were seen by'the
lowest intake port. This explains the average 48 pe-fce"nt
overstatement of concentration experienced by the prototype
sampler under this set of conditions.
The Model A commercial sampler was a forced flow, pneumatic
ejection-type using pressurized gas to force the sample from
the intake chamber in the flow up to the sample container
housed within the instrument case. The sample intake used
had a nominal 250 mH capacity and was gravity filled, after
discharge of the previous sample, through eight ports evenly
spaced around its periphery. As is typical of this type
sample, it must be used with its own specially designed in-
take. A 0.64 cm (1/4 in.) ID line was used to transpor-t the
sample from the intake to the container. The performance of
this unit for gilsonite flows was relatively good, being
more erratic than the prototype, but better in this respect
than the other three commercial units. The unit was unable
to gather any appreciable quantity of the fine sand, however,
apparently because of its intake design. "
The Model B commercial sampler was a suction lift unit that
utilized a .moderate speed peristaltic pump and 0.64 cm
(1/4 in. ID) sample transport tubing. The intake provided
was a weighted plastic cylindrical strainer with four rows
of five 0.32 cm (1/8 in.) holes evenly spaced around its pe-
riphery. This unit suffered from a kinked hose in the in-
take line which probably occurred early in the morning of
77
-------
co
CO
w
>
H
H
co
PM
W
o
J3
W
co
cs
-. CU
4-> CJ
CU -H fi
P. O O
>. 0 O
H rH
CU rH
co > cd
*Q ti
•rl |5 "rl
rH O a
OHO
co PM !a
r»- CM co vo co
in H co H
H
CO o> vO O CM
CO CM VO H
H
•K «K
r~ o co co o
CO -31 co
CO
o CM m o- vo
*3* r*^ co
CTi H H CO vo
en vo vo in
H
-3" O -3- O CO
O O H Ol CT>
H H CM
VO <)• CO O% H
rH !**• Q^ Q^ t-^
rH H -sT
•K
O VO -3- O"> O
H r^ I-N. in -vt
H H -*
cu
P.
>". «i M O Q
4-1
0 . H H H H
4-1 CU CU CU CU
O TJ T3 T3 TJ
rl O O O O
PM s a g g
!>»
i-Q
*O
cu
TJ
•rl
J>
•H
13
fi
0
•H
4-1
cd
p
4J
fi
cu
CJ
fi
o
CJ
cu
H •
PiO
a- o
cd H
CO
CO
•o o
cu a
N -rl
>• 4->
H
cd fi
fi 0
cd 'H
4J
co cd
•H rl
4-1
CO fi
co cu
CU CJ
fi fi
cu o
^ CJ
iH
4J CU
cd a
4J 3
fi H
cu in
CO
CU H
rl Cd
P. 3
CU 4->
rl CJ
cd
60
fi CU
•H 60
H Id
Oj Lj
a cu
cd >
co cd
CO
cu
60
cd •
M 0)
CU 60
> cd
Cd rl
cu
4-> >
fi cd
•H
O 4-1
p. ti
•rl
m o
p.
cu
!H ^3-
cd
cd
CO
rl 60
CU fi
43 -H
•M >
0 cd
CU
H H
H
Cd TI)
cu
• •s *X3
CU rl
60 Cd
cd 60
rl CU
CU rl
> CO
Cd -rl
^rj
4J
fi CO
•H Cd
O |3
Pi
4-1
0 fi
H -H
O
O
PI cd
3 4-1
cd
rc3 *cf
cu
CO CU
cd fi
o O
cd cu
4J rl
cd cu
•V 43
[5
cu
Pi*
>1
4-1 4J
O Pi
4J CU
0 CJ
rl X
p^ cu
13
fi
cd
CO
cu
ti
•H
MH
CO
•H
CO
ptl
• A
^-^
a
a
o
o
•
CM
v|
trj
v|
co
VO
•
H
* •
VO ,-s
o a
• a
H
in
II CM
H
• 1
o o
•
co v\
S^ /
T3
cu
4J V 1
•H
fi in
0 O
CO H
H •
•H 0
60
•V
a •n
3 vo
•H •
T3 CM
CU
a n
co •
•H O
*
O CO
jsj *^
.
*^)
fi
o
CJ
CU
CO
rl
cu
p.
co
rl
CU
4J
CU
a
ti
•rl
cd
4J
43
O
O
4-1
CO
.
O
!>i
43
rrf
a
0
CJ
cu
CO
rl
CU
Pi
4-1
CU
CU
<4H
^>
H
Pi
•H
4J
H
3
CO
EH
O
&
78
-------
4-1
O
T-H
4-J
0)
co
en
•H
4-1
d
(U
0)
Q)
f-t
ft
•cu
60
d
•H
J-l
3
M
•rl
9NIldW\/S
79
-------
-*
o
0)
0)
CO
CO
CO
-------
o
en
v|
C
•a,
v| ii
o
en
03
co
co
•H
4-1
CB
4J
C
(1)
0}
0)
S-l
ex
cu
60
(3
•H
t-l
PL,
0
rt
r--
CN
(U
M
S
t>0
o
S3
o
00
(%) SS3N3AIlVlN3S3cld3a SNHdWVS
81
-------
•u
-------
July 26 (beginning with run 26-1) and was not detected until
after run 26-12. This accounts for its poor performance
during this period. Otherwise, its performance with gilson-
ite was fairly good. There was a problem with intake orien-
tation (a result of its design) during runs 25-1 through
25-10 which accounts for the units' poor performance with
fine sand at 1.2 m/s (4 fps). Correcting this situation by
repositioning the intake resulted in some improvement in
operation with fine sand at 2.4 m/s (8 fps), but the unit
was still unable to gather appreciable quantities of this
mater ial.
The Model C commercial sampler was a suction lift type de-
vice that used a slow speed peristaltic pump with 0.64 cm
(1/4 in.) ID tubing. It was used both with and without its
plastic intake screen, which was a weighted cylinder whose
lateral surface was formed by rolling up a sheet of plastic
perforated with a large number of 0.32 (1/8 in.) holes. The
vertical position of the end of the suction tube within the
intake screen was adjustable. The performance of this unit
tended to be erratic, and no general explanation for its be-
havior can be made at the present time.
The Model D commercial sampler was also a suction lift type
device that used a vacuum pump to evacuate a metering cham-
ber into which the sample was drawn prior to expulsion into
the sample container. Its intake was simply the unscreened
end of the plastic suction tube with a weight attached. The
unit's performance would appear related to the intake design.
As flow velocities increase, the elevation of the intake
above the invert increases and, therefore, the intake can be
located in a region of relatively lower concentration.
This, combined with the difficulty of gathering the higher
density sand particles, would appear to explain this piece
of equipments' gross overstating of flow concentrations with
gilsonite and understating flow concentrations with sand.
DISCUSSION
The results of the controlled laboratory testing phase of
the project are encouraging, except for the ability to
representatively sample some flows containing heavy sus-
pended solids. Fairly good results were obtained with the
prototype sampler when most of the particles were in sus-
pension in the flow (even for those falling somewhat outside
the range of validity of Stokes' Law). Results tended to be
erratic, however, when any appreciable quantity of material
was transported as bed load.
83
-------
Of greater concern are the differences in performance of the
commercial automatic samplers. The data conclusively demon-
strate that there can be marked differences in the results
obtained with different types of sampling equipment, even
under identical, controlled flow conditions.
It was not the purpose of this brief series of tests to
evaluate the different design approaches used by commercial
manufacturers of automatic sampling equipment. The results
of the tests do point out the critical urgency for such work
to be performed, however. As suggested by other work and
demonstrated by the present effort, some water quality data
now in existence may be considered suspect, at least insofar
as suspended solids are concerned. Thus, there is an urgent
need for determining the capabilities of various types of
sample collection systems to gather representative samples
of wastewater flows over a wide range of characteristics.
This is especially critical for equipment to be used in storm
and combined sewer flows which often contain large amounts
of heavier suspended solids.
Another great area of concern is in the ability to sample
floatables and bed load solids. The prototype sampler was
not designed to sample either in a truly representative
sense, and no claim should be inferred as to its ability to
do so. Such materials (including undissolved or nonemulsi-
fied oil and grease) may, and in all likelihood will, be
present in many storm and combined sewer flows, and sampling
of them is necessary if one is to be able to account for all
pollutant loadings in a time-mass-discharge sense. No equip-
ment exists today that is well suited for these purposes,
and a program to develop devices that can be used for such
wastewater characteri-zation appears to be sorely needed.
84
-------
SECTION IX
REFERENCES
1. Shelley, P. E. and Kirkpatrick, G. A., "An Assessment
of Automatic Sewer Flow Samplers," USEPA Research and
Monitoring Environmental Protection Technology Series
Report No. EPA-R2-73-261 (1963). \
2. Inter-Agency Committee on Water Resources, "Laboratory
Investigation of Suspended Sediment Samplers,"
Report No. 5 (1941). ? tt"
3. Inter-Agency Water Resources Council, "Laboratory .
Investigation of Pumping Sampler Intakes," Report No. T
(1966).
4. Inter-Agency Water Resources Council, "Investigation of
a Pumping Sampler With Alternate Suspended Sediment
Handling Systems," Report No. Q (1962).
5. Schulz, E. F., Wilde, R. H. and Albertson, M. L.,
"Influence of Shape on the Fall Velocity of Sedimen-
tation Particles," Missouri River Division Sedimen-
tation Series Report No. 5, U.S. Army Corps of
Engineers, Omaha, Nebraska (1954).
6. Inter-Agency Committee on Water Resources, "Some
Fundamentals of Particle Size Analysis" Report No. 12
(1957).
7. Mkhitaryan, A. M., Gidravlika i Osnovy Gazodinamiki,
Gosudarstvennoe Izdatel'stvo Tekhnicheskoi Literatury
UkrSSR, Kiev (1959).
8. Knorz, V. S., "Beznapornyi gidrotransport i ego
raschet," Izvestiya Vsesoyuznogo Naucho-
Issledovatel'skogo Instituta Gidravliki, Vol. 44
(1951),
9. Water Pollution Control Federation, Design and Con-
struction of Sanitary and Storm Sewers, Manual of
Practice No. 9 (1970).
10. U.S. Environmental Protection Agency, "Methods for
Chemical Analysis of Water and Wastes - 1971,"
Report No. 16020 07/71.
85
-------
11. Harris, D. J. and Keffer, W. J., "Wastewater Sampling
Methodologies and Flow Measurement Techniques", USEPA
Region VII Surveillance and Analysis Division, Report
No. EPA 907/9-74-005 (1974).
12. Inter-Agency Committee on Water Resources, "Determina-
tion of Fluvial Sediment Discharge", Report No. 14
(1963).
86
-------
SECTION X
APPENDIX
CONTROLLED LABORATORY TESTING DATA
This Appendix contains plots of the raw data that were taken
during the comparative testing phase of the project. They
are plotted by run number, the first two digits of which are
the day of the month (July, 1974) when that run was made,
while the digit(s) across the bottom of the plot were as-
signed in chronological order for the day in question.
Five data replicates were taken for each set of test param-
eters involved. The concentrations for each piece of equip-
ment, as determined by laboratory analysis, are plotted, and
the simple arithmetic average of the five points is indicated
as a straight line. The straight line representing the aver-
age flume concentration is, in most cases, the simple average
of ten calibration samples, five taken just before and five
taken right after each run. The shaded area indicates the
band within which all of the flume calibration data fell.
All concentrations are given in parts per million on a weight
basis and may alternatively be considered as being given in
milligrams per liter.
Also given on each page are the velocity of the flow and the
suspended solid involved. Flow velocities are given in feet
per second; multiply by 0.3 to convert to meters per second.
The characteristics of the suspended solids are:
Medium gilsonite: 1. 6 8 <_d£2 . 0 0 mm, S.G. - 1.06
Fine sand: 0.105
-------
en
LO U1
g 3
Q. U-
8 S
co r~.
I
\O
(Wdd) NOIi«llN33N03
88
-------
o
o
oo
10
8
to
:>
•a:
•a:
o
o
CM
VO
oo
cvj csj
CO
CS
CQ
1 => O
> —i a.
: u. Q.
(Wdd)
89
-------
C\J .—
I
»0
04
O
S3
(Wdd) NOIlVMiN30NOD
90
-------
CM
CM CTi
CD CD
£ £
o o
i H-
11-
I f'
o
8
CM'
g
to
o
o
CM
O O
O O
O 00
g
vo
Ul CX.
z u-
U- CO
1
1
o o o o
(Wdd) NOIlVcliN30NOD
i %l
J "
it
IU
O
^
O
&
a
91
-------
o
en
is
CD
>
•a;
§0
o
0
o
00
o o
i— o
s
a iaaow
S S
o o
CO CM
8
o
I
O
oo
•a-
CO
3
CD
ti
3
£
0 13QOW (WV 3dAiOiOJid g
13QOW
o
o
•a-
(Wdd) NOIl\ftiiN33N03
-------
«£
a
nv)
o
O
CO
CM i—
I
O
CO
CO
>
=1
O
!S
S
3 13QOW QNV 3dA10iOcld o
<"
(wdd)
93
-------
o
S3
a
CD
5
»
__—(r
LO ' O LO
CO CO CO
(Wdd) NOUtfNiNH'JNOD
CJ i—
-------
O
13
I— CM
i— CD
s g
CNJ I
(Wdd) NOIiVaiN30NOO
95
-------
TECHNICAL REFURT DATA
(Please read lustructions on the > •. •••crse before completing)
1. REPORT NO.
EPA-600/2-76-006
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Design and Testing of a Prototype Automatic Sewer
Sampling System
5. REPORT DATE
March 1976 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTKOR(S)
Philip E. Shelley
8. PERFORMING ORGANIZATION REPORT NO.
3. PERFORMING ORGANIZATION NAME AND ADDRESS
EG&G Washington Analytical Services Center, Inc.
2150 Fields Road
Rockville, Maryland 20850
10. PROGRAM ELEMENT-NO. 1BB034 ',
ROAP 21 ASY: Task 039
11.CONTRACTK5H4S6KTNO.
68-03-0409
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati. Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final Report
14. SPONSORING AGENCY CODE
EPA-ORD
IS, SUPPLEMENTARY NOTES
See also EPA-600/2-75-065, "An Assessment of Automatic Sewer Flow Samplers - 1975'
16. ABSTRACTA brief review of the characteristics of storm and combined sewer flows is
given, followed by a discussion of the requirements for equipment to sample them, not-
ing features that are desirable in such equipment and problem areas. When considered
from a systems viewpoint, there are five functional subsystems. Design considerations
for each of these are discussed, followed by a description of the design implementation
used for each subsystem in the fabrication and assembly of a prototype automatic sewer
sampling system intended for storm and combined sewer application and other adverse
sewer flow conditions.
The prototype sampler is described from an installation and operation viewpoint, and
the results of preliminary field testing are discussed. The device was also tested
under controlled laboratory conditions and found to be capable of gathering reasonably
representative samples (i.e., within 10%) over a fairly wide range of flow character-
istics, even for particles somewhat outside the regime of Stokes' Law. Four different
commercially available samplers were tested under the same flow conditions in a side
by side fashion. Their behavior was rather erratic, and they were not able to gather
representative samples consistently. None of them was capable of good performance when
appreciable bed load was present. Results from these commercial units ranged from an
overall understatement of pollutant loading by 25% or more, to overstatements of 200%
and more.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Sampling, Samplers, Sewage, Water analysis,
Water quality, Water pollution, Effluents,
Storm sewers, Combined sewers, Overflows,
Manholes, Outfall sewers, Sanitary engi-
neering, Urban areas, Field tests
Prototype automatic
combined sewer sampling
system, Laboratory tests
Comparison tests
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURIVY CLASS (TillsReport!
UNCLASSIFIED
21. NO. OF PAGES
106
2OrSECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
96
•5^-U.S. GOVERNMENI PKIHIING OFFICE: 1976-657-695/539't Region No. 5-11
-------
-------
-------
-------
m
^^
-^V
!'rn
O m
=s 2
o' <
CD DO
sa
o 5.
CD m
u — CD
Q) _ Q)
~
— 3
O
=• 3 5} DO
o s a g
& § o m
% cff
CD o
00 S> ^
CD
CD 3
O
m
z
o
------- |