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
       	



-------