United States
Environmental Protection
Agency
Water Engineering Research
Laboratory
Cincinnati OH 45268
EPA-600/9-85-005
January 1985
Research and Development
Proceedings:

Seminar Workshop on
Aeration System
Design, Testing,
Operation, and Control

-------
                                               EPA-600/9-85-005
                                               January 1985
           PROCEEDINGS:  SEMINAR WORKSHOP ON
AERATION SYSTEM DESIGN, TESTING, OPERATION, AND CONTROL
      University of Wisconsin, Madison, Wisconsin
                    August 2-4, 1982
Cosponsored by the U.S. Environmental Protection Agency
                 and Environment Canada
                       Edited by

                    William C. Boyle
                University of Wisconsin
                Madison, Wisconsin 53706
      Cooperative Research Agreement No. CR809975
                    Project Officer

                   Richard C. Brenner
              Wastewater Research Division
         Water Engineering Research Laboratory
                 Cincinnati, Ohio 45268
         WATER ENGINEERING RESEARCH LABORATORY
           OFFICE OF RESEARCH AND DEVELOPMENT
          U.S. ENVIRONMENTAL PROTECTION AGENCY
                 CINCINNATI, OHIO 45268

-------
                                 DISCLAIMER


     Although the seminar/workshop reported herein was funded in  part by the
U.S. Environmental Protection Agency under Cooperative Research Agreement
No. CR809975 to the University of Wisconsin-Madison,  most of the  papers  have
not been subjected to the Agency's required peer and  administrative reviews
and, therefore, do not necessarily reflect the views  of the Agency; no
official endorsement should be inferred.  Mention of  trade names  or commercia"
products does not constitute endoresement or recommendation for use.
                                     n

-------
                                   FOREWORD

     The U.S. Environmental Protection Agency is charged by Congress with
protecting the Nation's land, air, and water systems.   Under a mandate of
national environmental laws, the agency strives to formulate and imple-
ment actions leading to a compatible balance between human activities and
the ability of natural systems to support and nurture  life.  The Clean
Water Act, the Safe Drinking Water Act, and the Toxics Substances Control
Act are three of the major congressional laws that provide the framework
for restoring and maintaining the integrity of our Nation's water, for
preserving and enhancing the water we drink, and for protecting the
environment from toxic substances.  These laws direct  the EPA to perform
research to define our environmental problems, measure the impacts, and
search for solutions.

     The Water Engineering Research Laboratory is that component of EPA's
Research and Development program concerned with preventing, treating, and
managing municipal and industrial wastewater discharges; establishing
practices to control and remove contaminants from drinking water and to
prevent its deterioration during storage and distribution; and assessing
the nature and controllability of releases of toxic substances to the
air, water, and land from manufacturing processes and  subsequent product
uses.  This publication is one of the products of that research and
provides a vital communication link between the researcher and the user
community.

     The seminar/workshop proceedings documented herein represent an effort
to summarize current research and development activities in the United States,
Canada, and Europe to improve aeration system design,  operation, performance,
and testing.  This information will find direct application in the search
for more cost effective wastewater treatment systems and more energy
efficient operation.  The brief workshop summaries presented at the end of
these proceedings provide insight into the further research needs of the
profession for this important unit process.
                                    Francis T. Mayo, Director
                                    Water Engineering Research Laboratory

-------
                                   PREFACE


     In the last few years, substantial effort has been devoted  to  the
search for more energy efficient wastewater treatment systems.   Since the
aeration of wastewater represents one of the more energy intensive  operations
in wastewater treatment, research and development efforts have  been aimed
toward improvements in design, manufacture, testing, operation,  and control
of these systems.  Among the research and development programs  that are
currently completed or underway are

     o clean water oxygen transfer test analyses and verification,
     o in-process oxygen transfer test development and analyses,
     o alpha factor test procedure development,
     o oxygen sensor evaluations,
     o fine bubble diffuser 0 & M evaluations,
     o dissolved oxygen control strategy analyses,
     o cost-benefit studies on aeration systems and aeration control, and
     o field-scale studies to optimize fine bubble design and operation.

     The major goals behind these broad-based research and development acti-
vities are aeration system energy reduction and improved criteria for sizing
oxygen transfer equipment.  In extracting maximum performance from  both
traditional and emerging aeration approaches, four principal technical  areas
must be addressed;  system design including specifications for  compliance
testing, shop and field testing, operation and related maintenance, and air
supply control.

     This Canadian-American seminar/workshop has brought together experts
and practitioners in the field of oxygen transfer to review and  discuss the
current status of aeration system design, testing, operation, and control.
Sufficient workshop time was provided to discuss in some depth  new  technology
and concepts in aeration system design and operation.  Areas of  agreement
and disagreement were identified and research needs described.

     The proceedings of this seminar/workshop will provide the  profession
with a current state-of-the-art document on aeration system design, testing,
operation,  and control.  It is anticipated that future workshops will update
this work and, perhaps, provide even more details in specific areas addressed
in this meeting.

                                      William C. Boyle, Project Coordinator
                                      University of Wisconsin
                                      Madison, Wisconsin
                                     iv

-------
                                  ABSTRACT


     The objective of this seminar/workshop for consulting engineers,
planners, and other professionals was to exchange information on recent
research and development activities in the United States, Canada, and Europe
on the design, operation, control, and testing of aeration systems.

     The 3-day seminar/workshop consisted of short, formal discussions
followed by an afternoon workshop session.  Papers and current technical
data on selected topics were presented in several technical sessions
including aeration system design, operation, and maintenance; aeration system
control; specification and testing of aeration equipment; and translation of
oxygen transfer data from clean water to process conditions.

     A summary of the afternoon workshop sessions is also presented  as a
portion of these proceedings.

     These proceedings were submitted in partial fulfillment of Cooperative
Research Agreement No. CR809975 under the cooperative sponsorship of the  U.S.
Environmental Protection Agency and Environment Canada.

-------
                                   CONTENTS
                                                                        Page

FOREWORD                                                                 iii

PREFACE                                                                   iv

ABSTRACT                                                                   v

ACKNOWLEDGEMENTS                                                          ix

INTRODUCTION TO SEMINAR/WORKSHOP

     An Overview of Oxygen Transfer Systems
          Edwin L. Barnhart                                                 1

CHARACTERISTICS OF AERATION SYSTEMS

     Diffused Aeration - Types and Applications
          Jerome D. Wren                                                    6

     Mechanical Aeration Systems - Types and Characteristics
          Terence D. Rooke                                                 14

     Oxidation Ditch Aeration Systems - Types and Characteristics
          George W. Smith                           ,                      29

AERATION SYSTEM DESIGN

     Translation of Clean to Dirty Water Oxygen Transfer Rates
          Michael L. Doyle and William C. Boyle                            45

     Aeration System Scale-Up
          Norbert W. Schmidtke                                             62

     Aeration System Design Protocols:  A North American Perspective
          Daniel W. Smith                                                  82

     Design Protocol for Aeration Systems - UK Perspective
          Arthur G. Boon and Brian Chambers                                99

     Experiences with Different Aeration Systems in Germany
          Rolf Kayser                                                     139

     Improvements of Air Diffusion Systems Applied in the Netherlands
          Johannes H. Popel                                               156

                                     vii

-------
     Blower Design Considerations
          G. 6. Powell

AERATION SYSTEM CONTROL

     Automatic Dissolved Oxygen Control in the Activated Sludge Process
          Joe P. Stephenson                                               201

     Evaluation of a Dissolved Oxygen Field Test Protocol
          Walter W. Schuk                                                 214

     The Calibration and Evaluation of Dissolved Oxygen Sensors
       in a Pilot Scale Activated Sludge Plant
          G. Speirs, D. Chapman, E. Luxon, and J. Matthews                238

OPERATION AND MAINTENANCE  OF AERATION  SYSTEMS

     Operation and Maintenance of Ceramic Fine Bubble Diffusers in
       Activated Sludge Systems
          Daniel H. Houck                                                 257

     Operation and Maintenance of Aeration Systems - An Operator's
       Perspective
          J. W. Gerald Rupke                                              269

     Operation and Maintenance/Troubleshooting
          David Redmon                                                    274

     Selected Experience with Aerators Used in the Treatment of
       Paper Industry Wastewaters
          James J. McKeown                            ,                    286

TESTING OF AERATION SYSTEMS

     Proposed Standard for Measurement of Oxygen Transfer
       in Clean Water
          C. Robert Ballod and Wayne Paulson                              295

     Oxygen Transfer Testing Under Process Conditions
          H. J. Campbell, Jr.                                             345

     Aerator Performance Evaluation - The Ontario Perspective
          Stephen A. Black                                                364

NEW DIRECTIONS

     Comparison of Dual  Nonsteady State and Steady State Testing
       of Fine Bubble Aerators at Whittier Narrows Plant, Los Angeles
          James A.  Mueller                                                375

     New Directions in Aerator Evaluation
          J.  S. Hovis and James J. McKeown                                400


                                   vi 11

-------
     New Directions - Off  Gas Methods
          Lloyd Ewing                                                      410

     Fundamental Research  in Biological Wastewater Systems for
       Advanced Design of  Mechanical Aerators
          Franz Nestmann                                                   431

WORKSHOP SUMMARIES

     Aeration System Design                                               438

     Operation and Maintenance                                            440

     Aerator Testing                                                      441

     Oxygen Transfer in Streams                                            445

APPENDICES

     A.  Seminar/Workshop Program                                         449

     B.  Seminar/Workshop Participants                                    451
                                   1x

-------
                               ACKNOWLEDGEMENTS

     The program for this seminar/workshop was conceived and designed through
the efforts of Nobert Schmidtke and Bruce E. Jank of Environment Canada,
Richard C. Brenner of the U.S. Environmental Protection Agency, and William
C. Boyle of the University of Wisconsin-Madison.   The local  arrangements and
research agreement administration were provided by John T.  Quigley of the
Engineering Extension of the University of Wisconsin.  The  contributions of
the seminar participants and authors during the 3 days of meetings in Madison,
Wisconsin, are gratefully acknowledged.

-------
AN OVERVIEW OF OXYGEN TRANSFER SYSTEMS

Edwin L. Barnhart, Professor
Southern Methodist University
Dallas, Texas  75275


     When I was first asked to present the opening remarks to this conference,
I was both flattered and delighted.  Flattered to be asked to address such an
eminent group and delighted to have people sit still and listen to a disserta-
tion on my favorite topic, oxygen transfer.  The reality of the situation
became obvious, however, as I sat down to actually prepare my remarks.  In
most cases, an audience knows a little about oxygen transfer, and even the
most fundamental observations are accepted as brilliant insights into the
process.  Here, however, the absolute reverse is true.  This audience is
probably one of the best informed in the world in dealing with the topic of
oxygen transfer.  What then does one say to inform or inspire such a group.

     The answer I found lies in the old expression concerning our ability to
recognize and separate the forest from the trees.  We, in this group, in our
professional activities, deal daily with the so-called "trees" of oxygen
transfer.  In fact, many of us spend a good part of our time dissecting the
leaves.  Under such circumstances, we may from time to time lose sight of the
bigger picture and of the role that oxygen transfer plays in the overall
science 
-------
     The size of the worldwide market cannot be estimated with any accuracy
at this point in time, but the data presented here is enough to set the
prospective.  We are dealing with a large and expanding market of significant
economic size.  Available data suggests that the overall transfer efficiency
of oxygen transfer equipment in place is quite low.  In fact, we probably
have the potential to save about one-half of all the power costs presently
employed for the transfer of oxygen if systems could be optionally designed,
built and operated.

     The fundamental requirement to begin on such a program of optimization
is agreement on a method of measuring oxygen transfer so that one can obtain
a valid comparison between systems.  Development and acceptance of such a
test protocol is one of the primary objectives of this workshop and hopefully
will be one of its principal accomplishments.


IMPORTANCE OF THE PROCESS

     No process is more fundamental to the successful operation of an aerobic
biological treatment system than is the transfer of dissolved oxygen.  Unless
dissolved oxygen is available where and when the bacterial system requires
it, the process will not function.  And, if adequate oxygen is available,
the process will function almost in spite of all other upsetting conditions.
Recent studies have shown that the absolute concentration of oxygen is also
of significant concern and influences the type of organisms that will dominate
and the general behavior of the process.

     Oxygen transfer capability is fundamental to proper employment of the
one hundred to two hundred billion dollars water pollution control program
that is underway on this continent.  Without adequate and correct oxygen
transfer, our capital investment will fail to achieve the goals that have
been set forth and, in fact, our society will be forced to live in an
unacceptable environment.  Oxygen transfer has a major impact on the way we
live and on the quality of life we can expect.


ASPECTS OF THE PROCESS

     The total function of aeration equipment must be viewed when we talk
about oxygen transfer.  A perfectly efficient, properly designed transfer
system must do more than simply dissolve gas in water.  It must provide an
environment which is optimum for bacteria to carry on our intended processes.
As such, the device is called on for a much broader class of service.

     The oxygen transfer device is the essential mixer which carries on all
the basic mixing functions within the aeration tank.  Mixing is required in
several senses.   A velocity must be introduced in the tank which is "suffi-
cient to hold the solids in suspension and provide sufficient shearing
forces that the floe does not become overly dense.  The mixing must  also be
gentle enough that the floe is not totally dispersed and the colonies
created by the bacteria are maintained and allowed to function.  Further,
the mixing must transport the bacteria and the food into close proximation

-------
so that reaction can occur.

     Another function of mixing involves the macro-system of the tank.  The
method of mixing will determine the degree to which the process approaches a
plug flow or completely mixed system.  While the nature of the system has
distinct implication with respect to treatment, it also influences, in many
cases, the energy required for oxygen transfer.  Consider, as an example, the
case where the raw waste exhibits a low a, say 0.6.  Whereas the treatment
effluent a approaches 0.9.

     If the system operates in a plug flow mode, the energy requirements for
oxygen transfer will be approximately 25% more than that in a completely
mixed system.  A dispersion of the organics through the tank is principally
affected by the mixing patterns induced by the aeration equipment.

     A second point that must be considered by this conference is that oxygen
transfer systems must achieve optimum transfer without a sacrifice of mixing
characteristics.  In fact, the design of all equipment must take into account
mixing otherwise the function of the system will not be accomplished.
FLEXIBILITY

     Another major area that must be considered in designing optimum aeration
systems is flexibility.  The oxygen requirements of any biological system
will vary widely during a typical day, week or month.  Typical designs today
often provide aeration equipment capable of meeting -the maximum four-hour
demand that occurs 10% of the time.  If such a design is chosen, a system
operating at that capacity will be capable of overaerating more than 95% of
the time.  Obviously, significant turn-down must exist both in the oxygen
transfer and in the mixing capability of the system.

     Many of the systems in place today that achieve high efficiency do so at
a significant sacrifice of flexibility.  Surface aeration systems are very
difficult to design in such a manner that a significant number of units can
be taken out of service.  Equally, diffused aeration systems have an absolute
limit on the minimum flow rate at which the necessary roll and hydraulic
dispersion takes place within the tank.  Clearly, this lack of flexibility
has a significant impact on energy efficiency.

     To install a device that is capable of transfer oxygen at the three or
even four pounds of oxygen per horsepower hour and then to be forced to
operate the device at double or triple its required capacity, a substantial
percent of the time can reduce the overall oxygen transfer efficiency
observed by the client to a very low level.  The lack of flexibility, com-
bined with the lack of sensing and control systems that allow operators to
actually adjust oxygen levels within the tank,  is  probably the single
largest contributor to energy waste in the field.  A small amount of attention
has historically been paid to this very important problem.

     Emphasis then, must be placed not only on providing a device that under
some ideal condition of flow, geometry and like can transfer oxygen  in an

                                      3

-------
 efficient mode,  but  also  to  the  development of a system that can maintain  the
 efficiency  over  the  wide  range of  conditions that would normally be  encounter-
 ed  in  an operating system.

     This conference should  address  this  requirement as one of the signifi-
 cant concerns  in oxygen transfer.  Consideration to examining a range of
 operating conditions should  also be  given in any test method so that the
 efficiency  of  an aeration system can be defined over the normal operating
 range  of the system  and not  at the single point at which it is most  efficient.


 FACTORS  RELATED  TO THE LIQUID

     Investigators have delineated the influences of oxygen saturation,
 temperature and  organic components on the ability to transfer oxygen.  The
 coefficient $  (beta) has  been defined to  estimate the approximate deteriora-
 tion of  saturation in the presence of various inorganic components while 6
 (theta)  is  used  to estimate  temperature impacts and the coefficient  a (alpha)
 has been used  to estimate the effect of organic compounds on the basic
 transfer mechanism itself.

     To  date,  the work concerning  oxygen  saturation has been quite good and
 although there is no absolute agreement on the coefficient 3 and all the
 factors  that effect  it there are generally acceptable techniques and
 estimating  procedures available  for  this  coefficient.  Similarly, although
 not all  totally  accept the mechanism for  estimating the impact of temperature,
 the range of variance between the  investigators is narrow and a reasonable
 estimate of this variance can be made.  The major remaining questions con-
 cern themselves  with alpha.

     Probably  in no  area  of  science  in which I am involved do we more behave
 like the blind men examining the elephant  than in the measurement of a.  Each
 set of experiments and each  investigator  seem to create conditions where a
 particular  influence on oxygen transfer is observed.  I have in my own ex-
 perimentations seen  alphas ranging from 0.3 to 1.9.  I" have seen changes in
 the physical conditions of my experimental system, change the measured
 results  by  40% to 50%.  I have seen  people I consider competent investigators
 perform  the same experiments as  I  did and  get significantly different results.
 These  variances  and  frustrations have been experienced by almost all who
 have worked extensively in this  area.  To  this date, we do not have  a
 reasonable  theory for the impact of  organic materials on oxygen transfer.
 What is  necessary, is the development of  some basic thesis which explains
 the behavior and allows us to compare our  experience on a meaningful
 scientific  basis.

     It  is  likely that organic agents influence oxygen transfer in at least
 two ways.   First, by  collecting  at the liquid-gas interfact and influencing
diffusitivity  and second,  on influencing  the very nature and size of the
transfer surface.  There  may, in fact, be  other more subtle effects  including
minor changes  in viscosity,   pumping  ability and the like which further
influence the  overall  transfer of  a  device.

-------
     It should be a goal of the group meeting here to attempt to develop a
basic thesis relating to the impact of organic materials on oxygen transfer,
so that the scientific community can compare their experiments to that thesis
and provide criticism and input that will lead to the evolution of an
acceptable understanding of the phenomena.
SUMMARY

     Oxygen transfer should be seen as a major industry in North America and
in the world.  Significant monies and energies are being expended to carry
out this most important process.  The protection of our environment is highly
dependent upon the successful implementation of this process.  We, gathered
here, to a large degree, have responsibility for seeing that continual
progress is made in more efficiently employing these resources and reliably
carrying out the intended process.  Several major problems face us - we must
design flexible, efficient systems under a variety of ever-changing conditions.
We must better define technology and the methods of testing that are to be
applied.  We have before us a significant challenge, but we have around us
an elite group of skilled scientists who I am sure are up to the challenge.


DISCLAIMER

      The work described in this paper was not funded by the U.S. Environmental
Protection Agency.  The contents do not necessarily reflect the views of the
Agency, and no official endorsement should be inferred.

-------
DIFFUSED AERATION - TYPES AND APPLICATION

Jerome D. Wren, Registered Professional Engineer
Vice President - Chief Engineer
Water Pollution Control Corporation
Milwaukee, Wisconsin  53201

     In the 1982 product guide section of the Journal, WPCF,
(1) under the heading "Diffusers, Air, over 30 suppliers
are listed.  The list is probably not all inclusive and many
of the suppliers listed offer several types, my company for
example, offers four types.  So, the variety of available
air diffusion devices is likely many times 30.

     It has been customary to suggest the relative clean
water performance characteristics of the devices by cat-
egorizing them as fine bubble (best ?), coarse bubble (good ?),
or even medium bubble (better ?).  This categorization is
somewhat subjective and relative and in some cases misleading
in terms of field performance.  MOP 5 (2), (3) and MOP 8
(4) distinguish only two categories; porous and non-porous.
Porous diffusers are defined as being constructed of either
natural ceramic or synthetic media and non-porous are de-
scribed by specific example.

     Porous diffusers are available in square or rectangular
flat plates, round discs either flat or profiled, domes
and tubes.  Non-porous diffusers are available in a wide
variety of shapes and materials.  They are constructed of
metal such as galvanized steel or stainless steel, a variety
of plastics, rigid or flexible.  They can be oriented
horizontally or vertically, have single or multiple air
release ports, fixed or variable area air flow passages,
baffled or plain.

     Several types of systems are a combination of air
diffusion and mechanical mixing.  Systems such as jets,
submerged turbines, and mechanically rotated air curtains
or baffle walls are examples.

     The general function of all the devices in activated
sludge aeration applications  is  to:

     - Transfer oxygen
     - Promote mixing of the oxygen with the aerated  solids
       and suspend the solids
     - Promote flocculation
     - Resist deterioration of  performance with time.

-------
     A brief review of the performance history of U.S.
aeration practice giver some insight into the reasons for
the multitude of devices.

     Oxygen transfer theory and early clearwater testing
indicated that porous type diffusers are relatively more
efficient than other types.  Thus, early systems consisted
of flat, square or rectangular porous plates.  Plates were
typically constructed of fused or bonded aluminum oxide or
silica.  They were installed in fixed plenums on the tank
floors, one plenum accommodating multiple individual plate
units.  Drop pipes supplied air to the plenums.  Adjustment
of air supply among the plenums was possible, but within a
plenum air was distributed among the plates in an inverse
relationship to their resistance to flow.

     The operating history of such systems indicated a wide
range of service life, but generally there was a loss of
performance with age or time in service  (2) (3).

     Loss of performance was characterized by system pressure
increase, a decrease in oxygen transfer efficiency, or both.

     Loss of performance was attributed to fouling or
clogging, air side or liquid side.

     A variety of methods to attempt restoration have been
tried, with varying degrees of success.

     The general lack of sustained success of those cleaning
methods and the expense of the techniques relative to the
then prevailing cost of power, encouraged the development of
more clog resistant systems.  Similar development proceeded
toward systems that could be removed from a tank without
draining, to make cleaning easier and non-process interrupting.

     Porous tubes or non-porous tube type diffusers were
developed to go along with the removable header systems. The
tubes consisted of a diffusion element and detachable mounting
hardware for connecting to the air header.

     Porous tubes were constructed of material similar to the
porous plates or bonded glass, synthetic beads, etc. Non-
porous tubes were available in many varieties, including
woven socks, wrapped cords, perforated polymers, etc.  Of
necessity the aeration equipment was located near the side
or sides of the tank, a position which clear water tests
showed to be less favorable than other locations   (5) (6).
To varying degrees the tube type diffusers were generally
vulnerable to the same operating problems as the porous  plates.

-------
     Developments continued to be in the direction of an
economical,, non-clog, efficient, low maintenance system.

     Much recent work (6)  (7) (8) f9) (10) (11) has been
done in an effort to define and refine the operating
characteristics of such devices.

     More or less, the variety of devices available today
were developed to overcome the historical problems associated
with high performance systems.

     The recent change in  the relationship between power
costs and maintenance costs has caused a renewed interest
in systems that obtain relatively high oxygen transfer but
require relatively more maintenance than others.

     The starting point for system performance comparisons
has traditionally been clear water tests.  Recent advances
toward a standard procedure (12) for clear water testing and
data evaluation will provide increased information on the
relative performance characteristics of the various systems.

     Clear water transfer  data of the several systems vary
quite widely but can be characterized in a few general ways.
For a fixed condition of submergence, and diffuser placement:

     - Some diffusion systems decrease in OTE with
       increased applied air or power/volume.
     - The mass of oxygen  transferred increases, but
       less than proportional to applied air rate.

     - Some diffusers exhibit more or less constant OTE
       with applied air rate, and the mass of oxygen
       transferred is proportional to applied power.

     - Some systems increase in efficiency with an increase
       in applied power.   For these the mass of oxygen
       transferred is more than proportional to applied
       power.

     - Some systems exhibit an optimum point of operation
       or a concave downward performance curve of OTE vs.
       air rate.  From the optimum point, the mass of oxygen
       transferred also increases with increase air, but
       less than proportionally.

     Field performance is  presumed to be related  to clear
water by the application of appropriate  conversion factors
such as alpha, beta, theta.


                              8

-------
     In the past, attempts to determine the relationships
between field and clear water have yielded unclear and
inconsistent results.  Recent activities  (13) in improving
field evaluation techniques show promise  of being  able to
determine the relative performance of the various devices
in service.

     With the renewed interest in porous  diffusers, new
techniques have been developed to characterize and the
performance and refine the design of the  diffusion element (14).
A discussion of the techniques is included in the agenda of
this conference.

     The characteristics can also be used to monitor the
field performance of the system at the diffuser.  Monitoring
the diffusers in service aids in evaluating performance with
time and gives an early indication of the necessity for
taking corrective measures or cleaning, if required.
Comparing the characteristics after cleaning to new, gives
a direct indication of the effectiveness  of the cleaning
technique.

     One of the most significant system differences between
currently available porous grid systems and past systems is
the incorporation of an individual flow control orifice for
each diffusion element.  Characteristically a plot of air
flow vs.  A P for the wetted media only,  is very flat. A
variation from average of +_ 10% in specific permeability
can result in a 200% range in air flow.   The flow control
orifice compensated for this and assures  almost equal air
flow among the diffusers in a grid.

     Other developments relate to improving air flow
uniformity or air flux rate across an individual element.
Uniformity of air flow results in more efficient use of
the diffuser's active area and has the potential for more
uniform restoration of performance upon cleaning.

     Another recent development is a method of in-place
non-process interruption cleaning   (16).  The process was
designed primarily to restore performance after plugging
due to inorganic scaling. On pilot scale  tests it has also
demonstrated some effectiveness in restoring performance
after other types fouling also  (15).

-------
     Recent  field  operating data suggests the desirability
of an inlet  mix  system,  to reduce the effects of surface
active agents  and  the suspected effect of high oxygen
uptakes at the inlet.  Other similar concepts suggest
applying the inlet load to the tank over a long inlet
weir (16) .

     Recent  studies (7)  show that the porous grid system
can be designed  to accommodate highly loaded processes.
However, flow, or  more appropriately load equalization
is probably  a  more desirable feature for the porous grid
system than  for  other types.  An example of diurnal BOD
load variation is  shown in Figure 1  (17).  The oxygen
demand may be  expected to vary in a similar sinusoidal
manner.  Dividing  the curve into areas above and below
daily average  shows a greater than 4:1 variation in demand.

     Since the transfer rate of porous grid is less than
proportional to  air rate, a 4:1 load variation requires a
greater than 4:1 air application rate.

     Also significant is that 74% of the total daily load
is being transferred at a rate above the average daily rate,
£ Aft
Q.6OO
^500
CD
i
0400
Z
-•»00
0)
100





\




AVE





RA






E










0-^







X





A
f


60 «


/





/





/


JU



/


/


-74%

%





/


OF






K
K
T
PI
U
0




TAI

N


. D/

\


iUY


\

LO



\
AD-



^
— ^




s,





s^








84%
l\
)URC£:
UPCRADMC EXISTMB «ASTE»ATtR
REATHENT PLANT -
ROCESS DESEN MANUAL
S.E.P.A. TECHNOLOGY THAJtSrtS
CT. 'M RP. >-• .
1 1
\~ i

          HID NIGHT
9:00 A.M.  12 NOON
   TIME OF DAY
                                            e:OO RH. MDNIGHT
                              FIG. I
          EXAMPLE OF DIURNAL LOAD VARIATION TO A U.S.  W.W.T.P
                               10

-------
     Some other types of systems do not require such a wide
range in air flow capability and would perhaps be more
efficient than a porous grid under peak loading conditions.

     Non-process considerations are also important to
the successful long term operation of any system.  These
include characteristics of the liquid environment, such as
temperature and ranges of change in temperature, corrosion
or scaling tendencies, presence of chemical constituents
which may degrade system components; reliability and cost
of power; 0 § M skills and cost relative to power; and
climate.

     In addition to activated sludge aeration, other unit
processes employ diffused air.  Some examples are pre and
post aeration, liquid transfer channels, sludge digestion,
sludge conditioning, aerated grit chambers, flow and load
equalization tanks, and transfer of other gases such as
ozone.

     Presently a great interest exists in high performance
systems, but there appears to be sufficient variety of
applications that no single system is likely to evolve as
the universal system of choice.
REFERENCES
        Journal,  Water Pollution  Control  Federation, Volume
        54,  No.  3,  1982 Yearbook,  Page  382,  Part  2.

        "Air Diffusion In Sewage  Works".  Manual  of Practice
        No.  5,  Federation of Sewage  §  Industrial  Wastes
        Associations,  Champaign,  Illinois (1952)

        "Aeration In Wastewater Treatment".   Manual of Practice
        No.  5,  Water Pollution Control  Federation, Washington
        B.C. (1971)

        Wastewater  Treatment Plant Design, Manual of Practice
        No.  8,  Water Pollution Control  Federation, Washington
        D.C. (1977)

        Bewtra,  J.K.,  and Nicholas,  W.R., "Oxygenation From
        Diffused Air in Aeration  Tanks."   Journal, Water
        Pollution Control Federation,  36, No. 10, 1195  (1964)
                              11

-------
 6.   Schmit,  F.L.,  Wren,  J.D.,  Redmon,  D.T.,  "The Effect
     Of Tank Dimensions § Diffuser Placement  On Oxygen
     Transfer'~Journal, Water Pollution Control Federation,
     Volume 50,  No. 7,  1750 (1978)

 7.   Yunt,  et.al.,  "An Evaluation Of Submerged Aeration
     Equipment - Clean Water Test Results", A Summary
     of Photographic Slides For the Presentation, WWEMA
     Industrial Pollution Conference,  Houston, Texas
     (June, 1980).

 8.   Ewing, et.al., Testing and Data Analysis of Diffused
     Aeration Equipment,  Journal, Water Pollution Control
     Federation, Volume 51, No. 10, 2384, (1979)

 9.   Rooney, T.C.,  Huibregtse,  G.L., "Increase Oxygen
     Transfer Efficiency With The Coarse Bubble Diffusers".
     Journal, Water Pollution Control Federation, Volume 52,
     No. 9, 2315, (1980)

10.  Gilbert, G., and Chen, S.J., "Testing for 02 Transfer
    Efficiency in a Full Scale Deep Tank."  Proc. 51st
    Ind. Waste Conf., Purdue Univ., Ann Arbor Science
    Publishers, Ann Arbor, Michigan, 291 (1977).

11.  Schmit, F.L., and Redmon, D.T., "Oxygen Transfer
    Efficiency In Deep Tanks",Journal,  Water  Pollution
    Control Federation, Volume 47, No.  11, 2586, (1975).

12.  Proceedings: Workshop Toward An Oxygen Transfer
    Standard, Boyle, W.C., ed., EPA-600/9-78-021,
   (April 1979).

13.  Redmon, D.T., Boyle,  W.C.,  "Preliminary Findings:
    Offgas Analysis", A Report To The ASCE Oxygen Transfer
    Standard Committee, Detroit, Michigan, (1981).

14.  Boyle,  W.C., Redmon,  D.T.,  "Biological Fouling of
    Fine Bubble Diffusers", EPA National Conference:
    Operation § Maintenance of P.O.T. W   Chicago,
    Illinois,(Jan.  1982).

15.  Schmit, F.L., "In-Place Ceramic Diffuser Cleaning
    System", EPA: National Conferences on Innovative
    § Alternative Technology, Various Locations, (1980)

16.  Houck,  D.H., Boon, A.G., "Survey § Evaluation of
    Fine Bubble Dome Diffuser Aeration Equipment",
    EPA 600/2-81-222 (1981)
                          12

-------
17.  "Upgrading Existing Wastewater Treatment Plants".
     "Process Design Manual, U.S. Environmental Protection
     Agency Technology Transfer, (October 1974).


DISCLAIMER

     The work described in this paper was not funded by the U.S.
Environmental Protection Agency.  The contents do not necessar-
ily reflect the views of the Agency, and no official endorsement
should be inferred.
                               13

-------
MECHANICAL AERATION SYSTEMS  -  TYPES AND CHARACTERISTICS

Terence D.  Rooke,  P.  Eng.
T.  D.  Rooke Associates  Limited
Toronto, Ontario  M6B 3W8

ABSTRACT

     Many types of mechanical  devices have been invented for the purpose of
aerating liquids.   These devices are categorized and their performance
characteristics summarized.

     While it is possible to rank various aeration devices in order of
energy efficiency  under standard test conditions, the optimum selection  of
aeration equipment for  a specific application must take into account other
factors.  Waste water characteristics,  oxygen up take rate and turndown
ratio, basin depth and  climatic conditions can each have a significant
impact on the final selection  of the aeration system.


INTRODUCTION

     In the waste  water treatment process, the consumption of soluble
organics by aerobic microbes requires the dissolution of large quantities of
oxygen into the mixed liquor.  Because  the solubility of oxygen in water at
atmospheric pressure  is limited, the driving force is small.  These
circumstances dictate the need for a large interfacial area in order to
achieve a mass transfer rate which will satisfy the process demand.

     The only other means of increasing transfer rate is to increase the
driving force.   This  can be  accomplished by conducting the mass transfer
operation at higher pressure or by using an oxygen enriched gas supply.  In
either case,  any increases in  the gas-liquid interfacial area which can  be
accomplished by the aeration device will provide proportionally higher rates
of  mass transfer.   Therefore one essential goal of any aeration system must
be  to  maximize the gas-liquid  interfacial area.

     Another consideration in  the design of an aeration system is the mixing
capability  of the  aeration device.  The direct pumping capacity plus the
induced flow must  be  capable of maintaining a uniform suspension of
biological  solids  and a positive dissolved oxygen concentration throughout
the aeration basin.   Mixing  can lengthen the duration of contact time
between the bubble and  the liquid and thus increase the oxygen stripping
efficiency.   Rapid blending  of toxic or other feed upsets will dilute their
impact  on the biological process.

     Moderate fluid shear rates can minimize the negative affect of
surfactants on mass transfer rates.  The velocity of the fluid relative  to
the bubble produces a washing  effect at the interfacial film.  Evidence  of
this phenomenon has emerged  in "dirty water" test results which indicate
                                     14

-------
different alpha factors for different aeration devices operating in the sane
waste water.

     Not only do different aeration devices exhibit differing transfer
efficiencies, circulation rates and fluid shear rates, but different
biological process designs place different requirements on the aeration
system.   A high rate system treating high strength industrial waste in a
relatively small aeration basin obviously places a different requirement on
the aeration system than does a low rate aerobic lagoon system treating a
dilute municipal waste.  Similarly, different aeration devices have a
different impact on the thermal balance of the process and the performance
of the process may be either enhanced or hindered by the choice of aeration
device.
TYPES OF AERATION DEVICES

All aeration devices can be categorized as follows:

 (1)  Diffused  air.

      (a)   Coarse Bubble Diffusers.

      (b)   Fine Bubble Diffusers.

      (c)   Tubes or  "Guns".

 (2)  Mechanical Aerators.

      (a)   Low  speed surface aerators  (LSSA).

      (b)   Motor speed surface aerators (MSSA).

      (c)   Brush type surface aerators  (BA).

      (d)   Down draft aerators (DDA).

      (e)   Submerged turbine aerators  (STA).

      (f)   Submersible aerators  (SMA).

      (g)   Jet  aerators  (JA).

 (3)   Special Aeration Devices.

      (a)   Trickling filters.

      (b)   Rotating  biological contactors.

      (c)   Miscellaneous proprietary devices.
                                     15

-------
This paper will be concerned only with Mechanical Aerators.


CHARACTERISTICS OF MECHANICAL AERATORS

Low Speed Surface Aerators (LSSA)

     A surface aerator is one which has an impeller operating at, or near
the free liquid surface.  It produces an umbrella shaped discharge of liquid
into the atmosphere, creating a large interfacial area through which the
mass transfer of oxygen takes place.  In the gas-liquid mixture, the gas is
the continuous phase and the liquid the discontinuous phase.  The supply of
oxygen is infinite.  The transfer rate is limited only by the rate at which
the impeller can expose new liquid interfaces to the atmosphere.

     "Low Speed" means that a speed reduction takes place between the motor
and the impeller, usually by means of a gear type reducer.  Operating speeds
are typically in the range of 100 RPM in the smallest sizes and down to 35
RPM in the  largest sizes.  Low speed aerators employ large impellers,
typically A ft diameter at 5 HP and up to 12 ft diameter at 150 HP.

     LSSA's produce extremely high pumping capacities.  For a given
expenditure of energy, a lower operating speed requires a larger diameter
impeller and the net result is an increase in pumping capacity.  A speed
decrease of 25% necessitates an increase of 18% in the impeller size which
in turn produces an increase of 25% in the direct discharge capacity of the
aerator.  A 5 HP LSSA will typically have a direct discharge capacity of
13,000 gpm  while a 150 HP unit will pump over 150,000 gpm.  Total flow
developed in the basin is typically three to six times the direct discharge
rate due to entrainment and induced flow.

     Due to their size and weight, LSSA1 s are usually fixed mounted but can
be float mounted where necessary.  (Refer to Figure 1.)

Advantages

     - LSSA's are simple in design, easy to install and to maintain as they
have no submerged components.

     - Their transfer efficiency is relatively high.

     - Their operation and maintenance costs are relatively low.

     - Their high pumping capacity allows them to provide adequate mixing
throughout very large volumes.

     - A turndown of 30% to 50% is attainable with liquid level sensitive
impellers and long term increases in plant loading can be accommodated by
the use of adjustable blade tips.
                                    16

-------

-------
Disadvantages

      - LSSA' s using only a surface  impeller can provide mixing only to
limited  depths.   The depth limitations range from about 10 ft for the
smaller  sizes to 17 or 18 ft for the largest sizes.  LSSA's fitted with a
lower mixing impeller or a draft tube are suitable for depths up to 30 ft.

      - The turndown in transfer rate and power consumption can not be done^
independently of pumping capacity and mixing.  Oxygen uptake rates can limit
the system design when dealing with exceptionally high strength wastes in a
high rate system.

      - LSSA's can be designed  to operate in the most severely cold climates
when mounted on a properly designed platform or fitted with spray
deflectors. The disadvantage  is that heat loss can reduce the system
kinetics.

Motor Speed Surface Aerators (MSSA)

      These devices  are similar in operation to LSSA1 s except that no speed
reduction is provided.   The impeller rotates at the same speed as the motor.
Operating speeds are usually 1800 RPM up to 10 HP, 1200 RPM from 10 HP to 40
HP and 900 RPM  in sizes 50 HP  and larger.

      Due to the high operating speed, the impeller size is small.  Modified
marine type propellers are used which are typically 8 in. diameter at 5 HP
and 24  in. diameter at 75 HP.   The  smaller impellers have a. lower pumping
capacity for any given motor size.  Typically, the direct discharge capacity
of a MSSA is 25% of the pumping capacity of the LSSA. '

      In  order to improve the effectiveness of the smaller impellers, a draft
tube is  provided which draws mixed  liquor from well below the surface of the
basin.   The vertical upwards discharge of the impellers is diverted by a
flow deflector which controls  the discharge trajectory.

      The elimination of the speed reducer makes MSSA1 s both lighter and more
economical than LSSA's for a given motor size.  Because they are lighter and
have a very limited shaft length, they are better suited for float mounting
and are  seldom  fixed mounted.   (Refer to Figure 2.)

Advantages

      - MSSA's are portable and simple to install.  They can be easily added
to  an existing system.  They have low initial cost and low installation
cost.

      - There are no  submerged  fittings and no auxiliary equipment such as
blowers.
                                    18

-------

19

-------
Disadvantages

      - MSSA' s have a relatively low transfer efficiency and a lesser mixing
capability compared to LSSA's.  There is no turndown capability other than
turning off individual units.

      - Cooling of the waste may decrease system kinetics.  Flotation
stability requires special consideration when installed in cold climates.

Brush Aerators (BA)

      These devices were developed specifically for Ditch Treatment Systems.
The  "Brush" type impeller rotates on a horizontal shaft which spans the
width of  the channel.  The brush simultaneously brings liquid into contact
with the  atmosphere and induces circulation in the ditch.

Advantages

      - The uni-directional discharge of the brush aerator is ideal for
inducing  circulation in a ditch or channel type system.  The transfer
efficiency is comparable to that of LSSA's.

Disadvantages

      - Best suited for one specific type of system.

      - High initial cost, no turndown capability.

Sub-Surface Aeration Systems

      All  systems which rely on a supply of compressed air introduced below
the  free  liquid surface are generically different from surface aerators.  In
these "sub-surface" aeration systems the fluid becomes the continuous phase
and  the gas the discontinuous phase.  The oxygen supply is limited by the
rate at which air is compressed and fed to the system.  The transfer rate is
dependent upon both the supply of air and the oxygen stripping efficiency of
the  aeration device.

      Because the saturation value of oxygen in water increases with
pressure, the mass transfer driving force increases with liquid depth.  The
stripping efficiency is observed to be significantly greater at greater
basin depths thus reducing the air flow requirements.  However, the
increased discharge pressure required from the compressor tends to  reduce
the energy savings associated with reduced air flow requirements.

Down Draft Aerators (PDA)

     These units consist of a vertical shaft driving an impeller which is
located inside a draft tube in the upper regions of the basin.  Compressed
air at very low pressure is injected into the draft tube imnediately below
the impeller.   The down pumping impeller provides circulation throughout the


                                    20

-------
basin and the downward velocities inside the draft tube carry the gas
bubbles to the bottom of the basin.  The fine bubbles provide a large
interfacial area and the depth of the basin results in an elevated partial
pressure of the oxygen.  The net result is a stripping efficiency in the
order of 40-50% of the available oxygen in the compressed air.

     The combination of the lower air flow requirement and the lower static
pressure at the point of injection results in a considerable reduction in
the size of the blower.  Consequently they are more energy efficient than
submerged turbine aerators.  Because they do not require an extremely long
shaft they can operate at higher speeds than the STA which in turn reduces
the torque and results in a smaller, less costly drive assent)ly.  (Refer to
Figure 3.)

Advantages

     - High stripping efficiency, small blower low discharge pressure, high
efficiency.

     - High pumping capacity, ideal flow pattern for suspension of solids.

     - Oxygen transfer and energy consumption can be turned down independent
of mixing.

     - No depth limitation.

     - No icing, spray or misting problems in cold climates.

Disadvantages

     - Draft tube increases installed cost.

     - Less efficient in shallow basins.

Submerged Turbine Aerators (STA)

     A gas dispersing impeller is located in the lower regions of the basin
and is driven by a vertical shaft.  Compressed air is fed below the impeller
and is dispersed by the hydraulic shear created by the impeller discharge.
The resulting fine bubbles are transported throughout the basin and provide
a large interfacial area and a long bubble detention time to facilitate the
mass transfer rate required.

     The fluid forces acting on the long overhung shaft and the critical
speed considerations both dictate a low operating speed, thus requiring a
speed reducer.  The rotational speed of the impeller is generally in the
range of 50 to 100 RPM.

     The impeller diameters and the resulting pumping capacities are
slightly less than those of LSSA's.  However, they do afford the ability to
independently control oxygen input without affecting the mixing capability


                                    21

-------
mm* <»**<* M: MM* ; :S Wil|

-------
of the device.  This can be accomplished by operating the aerator at
constant speed while increasing or decreasing the supply of compressed air.
(Refer to Figure 4.)

Advantages

     - No cooling of the waste or icing.

     - Energy and oxygen transfer can be turned down independently of the
mixing.

     - Suitable for deep basins.

     - Nearly unlimited oxygen uptake rate capability.

Disadvantages

     - Requires blowers and submerged piping.

     - Deep basins involve higher civil costs.

     - Lower transfer efficiency.

Submersible Aerators (SMA)

     A submersible aerator employs a submersible motor to drive an-impeller
located near the bottom of the basin.

     The impeller may operate in a draft tube or in a pump casing.  It may
operate at motor speed or the speed may be reduced through a submersible
gear box.  In either case, the impeller generates flow and compressed air is
piped to the unit to be dispersed and circulated through out the basin.
(Refer to Figure 5.)

Advantages

     - High pumping capacity and good mixing due to use of speed reduction
and larger diameter impellers.

     - Independent turndown of energy and oxygen transfer without reducing
mixing.

     - Good transfer efficiency.

     - No problems with cooling, spray or ice.

Disadvantages

     - Requires a blower.  Most maintain submersibe motor.
                                    23

-------
units at Sr«w gay   ^jsewj,
                *^**jr *  «¥ J i. So-iwTS! fl
        -

-------
                    FIGURE  5   SUBMERSIBLE  AERATOR
Jet Aerators  (JA)

     Instead  of a rotating impeller to aid in the dispersion of the
compressed gas, these units utilize a venturi tube through which mixed
liquor is pumped.  The gas is admitted into the throat of the venturi at the
point of highest velocity and lowest hydraulic pressure.  This promotes a
rapid dispersion of the incoming gas and the hydraulic flow pumped through
the venturi provides the circulation necessary to achieve satisfactory
mixing.

     The air  supply is normally from a centrally located blower room.  The
hydraulic flow through the venturi can be generated by either a remote dry
well pump or by locally located submersible pumps.    (Refer to Figure 6.)
                                     25

-------
                                                    Safety
                                                  disconnect
                                                    switch
               Guide bars
               Submersible
                pump with
               self-cleaning         ^f^iP^  ^Vortex jets
               intake screen
                                                             Air line
 Submersible
electrical cable
  FIGURE  6   JET AERATOR

              One of several configurations available with or without
              submersible motors.


Advantages

     - No cooling or ice problems in cold climate operation.

     - Can turndown energy and oxygen transfer independent from mixing.

     - Can be adapted to all types of basin geometry.

Disadvantage

     - Requires blower and submerged piping.

     - Relatively high installation cost.

     - Must maintain submersible pump.

     - Transfer efficiency comparable to STA.
                                  26

-------
MIXING

     Mixing continues to be a poorly understood yet very important aspect of
aeration system design.  Some of the misunderstanding stems from textbooks
on the subject  of biological treatment written by authors who had little
knowledge of the technology of fluid mixing.  In the absence of any better
criteria, these textbooks suggest guidelines based on one parameter only,
energy input per unit of volume.

     Mixing requires that flow streams be developed within the basin.  The
greater the flow, the faster the mixing.  Flow development can be either
forced and/or induced.  Some aeration devices develop high direct discharge
rates while some rely entirely on induced flow.  Even within the generic
category of surface aerators there exists a A or 5 to 1 range in the
discharge rates per unit of energy.  Such wide variations in performance
make any guideline calling for a minimum value of HP/1000 ft3 too general to
be of much value.

     Flow generation capacity per unit of basin volume or "Turn-over rate"
is much more pertinent, but still does not define the quality of mixing.  A
small diameter  high velocity stream and a large diameter low velocity stream
may have the same mass flow rate but they will produce different mixing
results.  Similarly, having all the mixing energy injected at one point in
the basin will  produce a different result than having the energy injected at
multiple points.  The direction, location and number of streams developed
will have a significant affect on the resulting fluid regime in the basin.

     Due to the difficulties in specifying equipment size or energy levels
to assure adequate mixing, it makes more sense to specify the mixing
performance required.  Velocity specifications are popular but velocity is
less definitive of mixing performance than one may be led to believe.  For
example, if the entire basin were rotating in the same direction at a
specified velocity, there would be little mixing.  The development of random
mixing requires varying velocities and eddy currents so it is quite possible
to find a zero  velocity reading at some point in a well mixed basin.  Nor is
the measurement of dissolved oxygen a technically correct measure of mixing.
Oxygen is a consumable commodity and the concentration must diminish the
further the flow travels from point of injection.

     Consideration of these facts suggest that the most practical and
effective measure of mixing in an aeration basin is the degree of uniformity
of solids distribution throughout the basin.  Solids suspension is a basic
process requirement and the fluid motion necessary to achieve a reasonably
uniform suspension will be more than adequate to distribute dissolved
oxygen.  A reasonable mixing specification might call for a uniform
suspension of biological solids throughout the basin within ± 10% of the
average.  Samples should be taken at several different depths in each of
several locations in the basin.
                                    27

-------
CONCLUSION

     There are many generic types of aeration devices available,^each having
its own peculiar characteristics.  Important variables include mixing, alpha
factor, basin geometry constraints, ease of installation and maintenance,
suitability for cold weather operation, initial cost, energy efficiency and
turndown capabilities.

     Various process designs place different demands on the aeration system
in terms of mixing and oxygen uptake rate.  Site conditions may dictate the
use of deep or shallow basins.  Climatic or aesthetic concerns may dictate
the use of covered basins.  High peaking factors may require a large
turndown capability.

     It is incumbent upon the treatment plant designer to select an aeration
system best suited to the needs of that particular project.  An awareness of
the characteristics of all generic types of aeration devices is necessary in
order to arrive at an optimum selection.
 DISCLAIMER

      The work described  in  this  paper was not funded by the U.S. Environ-
 mental  Protection Agency.   The contents do not necessarily reflect the
 views of the Agency,  and  no official endorsement should be inferred.
                                    28

-------
OXIDATION DITCH AERATION SYSTEMS - TYPES AND CHARACTERISTICS

George W. Smith, Process Section Manager
Envirex Inc., A Rexnord Company
Waukesha, Wisconsin   53186

ABSTRACT

     During the last ten years, the oxidation ditch aeration
system has gained in popularity.  Various types of basin con-
figurations have been proposed and used, along with numerous
types of aerator/mixing devices; the advantages and benefits
claimed with these widely varying configurations and aerator/
mixing devices have been very widespread and sometimes
conflicting in nature, especially concerning advantages of
oxygen transfer efficiency and nitrification/denitrification
capabilities.  The low-loaded activated sludge treatment modes
associated with these systems do achieve consistent degrees of
high treatment performance.  Oxidation ditch systems do offer,
because of their continuous flow circuits and uni-direction
flow pattern, low-energy mixing requirements, resulting in more
turn-down flexibility for both the small and large size
treatment plants.
INTRODUCTION

     The oxidation ditch aeration system, a treatment mode
chosen more and more frequently in recent years for a wide
range of treatment applications, is typically classified as a
low-loaded, looped-reactor, activated sludge system where
aerators of a specific type provide oxygen and establish an
uni-directional mixing pattern to the basin contents.  While
designs have been proposed for plants over 100 MGD in the
United States  (and are in operation for plants over 100 MGD in
Europe) the system, because of its ideal mixing regime, is
especially suitable and attractive for the small to medium size
extended aeration plant where nitrification is required.  Part
of the systems current popularity is undoubtedly due to the
favorable EPA publication released several years ago containing
a survey of the operations and performance of a large number of
oxidation ditches  (1).  According to this report, the oxidation
ditch, when compared to alternate treatment system modes, showed
a higher and more consistent performance level.  These results,
however, may be more due to the forgiving nature of low-loaded
systems rather than the unique basin or aerator characteristics
associated with looped reactor systems.  Low-loaded conventional
aeration systems, though lacking the mixing  (and in many cases,
flexibility) advantages of a looped reactor, should have the
same degree and dependability of treatment performance, at least
under normal influent conditions.


                               29

-------
     Developed in Holland in the fifties, the oxidation ditch
was brought over to the United States in the sixties and
successfully applied in various regions of the country.  These
early ditches were all single-loop basins, usually in a race
track configuration with sloped walls and an elongated straight-
away section, and employed horizontal rotary aerators, either of
the cage or brush rotor type.  Most of the brushes were manu-
factured by Lakeside or Passavant and used the trade names of
Magna Rotor or Mammoth Rotor.  The orbal system, a multi-channel
oxidation system using aerator discs was introduced in the late
sixties as were jet aeration channels (a single loop system)
using subsurface jet aerators.  Later the Carrousel System, a
folded-loop reactor using slow-speed, large diameter, vertical
surface aerators at the channel ends, were introduced.  In all,
there are at least seven variations of the oxidation ditch
system in operation in the United States, with the before-
mentioned four being by far the most dominate.  The list of
ditch-type systems is as follows:

     a)  Single-loop systems using cage or brush aerators.
     b)  Orbal  (multi-channel system using disc aerators).
     c)  Jet aeration channels.
     d)  Carrousel  (folded-loop system using vertical surface
         aerators).
     e)  U-tube ditches.
     f)  Counter-current ditches using diffused aeration.
     g)  Burns & McDonnell Treatment System.

     Each system will now be described in more detail.

Single Loop System

     There are presently more than three hundred single-loop
oxidation ditch plants using horizontal rotary aerators in
operation in the United States and Canada.  While various types
of horizontal rotary aerators have been tried, the dominate
aerator used is the brush rotor, usually operating in the speed
range of 50 to 70 rpm with an overall aerator diameter of
approximately 42 in.  The horizontal rotor of the brush consists
of a. 14 in.  diameter support torque tube on which 14 in.  long
steel blades are secured.   Their horizontal lengths can be as
long as 30 ft.

     While most of the early oxidation ditches were designed
with shallow depths and sloped side walls, deep tank  ditches
with vertical walls have now been used.  Many of the  early
ditches were for plants smaller than 1 MGD; low-loaded
oxidation ditches are now being proposed for plants  larger  than
20 MGD.

     High-loaded oxidation ditches have been used  in  the past
for larger size installations.  Their process results  are very

                               30

-------
similar to conventional aeration systems.  Since mixing is
usually never critical for high-loaded systems, it would seem
that oxidation ditches would not necessarily have strong
advantages for this type of application.  Care also must be
taken in these types of designs that velocities established by
the ditch aerators do not become excessive, resulting in a loss
of aerator oxygen transfer capability.

     Brush aerators have good oxygen transfer efficiency rates,
being approximately 3 to 3% pounds of oxygen/hp-hr at standard
conditions for a wide range of rotor speeds and blade immersions.
Mixing can also give a requirement of 21,000 gallons of basin
volume for every foot of rotor in order to maintain proper
channel velocities  (2).  Since actual channel velocity and mixing
efficiency is dependent on the frictional resistance of the walls
and floor, the shape and size of the ditch, and other mixing
factors such as channel bends and constrictions, the 21,000
gallon figure is probably conservative for many of the larger
ditch applications.  Baffle deflectors, used immediately down-
stream of the brush aerator in the deeper ditches, offer a
considerable flow constriction and significantly reduces the
velocity; usually this is beneficial since the higher velocity
rates can reduce the oxygen transfer capability of horizontal
rotary aerators.

     A loop reactor is a continuous flow circuit and this in
itself lends to the mixing efficiency.  A uni-directional mixing
pattern obviously requires less energy than a haphazard mixing
pattern in maintaining minimum velocity levels for solids
suspension.  In ditch applications where several brush or
horizontal rotary aerators are located in the same basin, it is
possible to remove one or more from service without creating
localized mixing inadequacies such as would be found in
conventional aeration tanks with mechanical surface aerators.

     The blade immersion of the brush can vary from 5-14 in. to
alter the oxygen delivery rate; rotor speeds of brushes can also
be varied to alter the oxygen output.  In many cases, this
variance is only possible because of the mixing superiority
offered by the brush aerator in an uni-directional flow basin;
in other words, oxygen delivery can be lowered to match oxygen
need without being penalized for mixing.  Again, this is a
strong advantage for low-loaded systems.

Orbal

     Over one hundred orbal oxidation ditch type plants using
aeration discs are in operation in the United States.  Developed
in South Africa, the orbal is a multi-channel system where the
degree of oxygen input can vary from channel to channel by means
of the number of aeration discs provided in each channel.
Considerable process.flexibility is allowed and numerous medium-

                               31

-------
load activated sludge process modes (such as contact stabili-
zation or step aeration) have been used.  The extended aeration
mode however, is the process mode most frequently used.  This
mode has the process advantage of nitrification/denitrification
occurring within the same sludge system.

     The aeration discs used are only % in. thick and 4 1/2 ft.in
diameter.  They are split in two semi-circular half sections for
ease of removal or relocation on the shafting.  Typical rotor
speeds are between 45 to 60 rpm.  Like the brush, the oxygen
transfer efficiencies of the disc are good, being in the range
of 3.0 to 3.4 pounds of oxygen/hp-hr at standard conditions.
The aeration discs consists of % in. diameter aeration holes and
numerous triangular protrusions which provide the oxygen and
liquid propulsion.  Despite the narrow thickness of the disc, it
has a very efficient mixing capability; design figures used are
1 hp worth of disc aerators for 50,000 gallons of basin volume
(3).

     While each channel of the orbal system duplicates the
characteristics of a loop reactor system, the kinetics of
several complete mix channels in series has some pronounced
differences over a single channel arrangement.  The oxygen uptake
rate, for instance, is most likely to vary considerably from
channel to channel, with the influent channel having the highest
rate and the last channel having the lowest rate.  In this way,
the orbal system approaches the characteristics of a plug flow
reactor; the desired effect is to eliminate any premature
exiting of incoming organic matter, especially the ammonia.

     It should be mentioned that the single channel reactor is
not a true complete mix system despite the rather quick dis-
persion of the influent flow over the entire contents of the
tank.  Actually the channel system has a  "moving-belt" type
circulation.  The degree of complete mixing can be very dependent
upon the design of the basin, especially  the size.  For a small
channel where the detention time is long, an individual flow
particle may, on the average, make over 500 circuits within the
basin.  A larger, more elongated channel, having the same 1 fps
channel velocity, and a reduced detention time, will have
considerably less circuits; as the number of circuits  approach
one, the characteristics of the system will approach that of  a
plug flow reactor.

     With the orbal system, the basin configuration is usually
kept oval on circular to assure a more complete mix charac-
teristic of each channel.  Usually the first or influent  channel
is largest in volume in order to gain the buffering advantage of
complete nix.

     The isolated zones of the orbal are  ideal  for achieving  a
consistent high-rate nitrification/denitrification performance

                                32

-------
(4).  In the orbal system, a high degree of process work will
occur in the first channel resulting in a depletion of dissolved
oxygen despite high oxygen input.  Nitrification has been shown
to occur in these oxygen depleted zones; because of the readily
available carbon source, all of the nitrates formed in the first
channel will be denitrified allowing the oxygen input of the
discs to be reduced.  The nitrification rate can be more than
90% complete in this oxygen-depleted or anoxic zone; complete
nitrification for diurnal ammonia peaks is provided by the
remaining channels where the dissolved oxygen levels, because of
the low uptake rates, can be kept excessively high.

     A high dissolved oxygen level can be kept in the last
channel of an orbal basin without a significant energy penalty.
Because of the efficient mixing provided by discs in a channel
reactor, a low oxygen input can be provided; because the uptake
rate is exceedingly low, the resultant dissolved oxygen level
will still be quite high.

     The zero dissolved oxygen level associated with the first
channel is an important energy consideration because of the high
driving force provided with oxygen delivery.  In the first
channel of the orbal, the dissolved oxygen level of the liquid
passing into the disc aerators is zero while the liquid passing
from the aerators also is near zero.  This is accomplished by
having wide-spread oxygen delivery application points of low
intensity; the low intensity delivery rate coupled with the
relatively high oxygen uptake rate of the liquid allows most of
the delivered oxygen to be immediately used instead of being
carried long distances in the channel flow in the form of
increased dissolved oxygen levels.  Keeping the oxygen delivery
rate below the uptake rate allows this to take place in this
channel.

     Like the brushes, the discs operate with a variable
immersion level, usually in the range of 9 to 21 in. Typically
the immersion level is allowed to vary over this range over a 24
hour period; a submerged orifice and dual effluent weirs are
used to achieve this purpose.

     The maintenance requirements for disc aerators and brush
aerators are similar.  Frequent greasing of bearings supporting
the horizontal shaft sections is a necessity.  Disc aerator
assemblies used with the orbal system usually have more bearings
than the brush aerator assemblies used with the single channel
ditch.  The steel aeration blades of the brush are more subject
to corrosion problems than are the plastic aeration discs.

     Aeration disc equipment is manufactured by Envirex Inc.
The orbal system is not a proprietary system and no license
fees are charged for the use of the process.  Figure 1 is of an
orbal basin showing the varying design dissolved oxygen levels.

                               33

-------
                        influent
                                        Return Sludge
High Dissolved Oxygen (2 ppm)
Medium Dissolved Oxygen (1 ppm)
Low Dissolved Oxygen (0 ppm)
                Figure 1.  The Orbal System

-------
Jet Aeration Channels

     Horizontal jet aerators have in the last decade been used
successfully in a looped reactor configuration.  These aerators
use both pumps and compressors to propel and aerate the liquid
contents of the ditch.  Because the aerator is a sub-surface
aerator rather than a surface aerator, the channel depths can be
much deeper.  This allows an advantage over the horizontal
rotary aerator described in the previous two sections which for
practical purposes, can only work, with depths between 5 ft. to
15 ft.  Optimum depths for a jet aeration channel would be
between 15 ft. to 20 ft.  Deeper channels are an advantage in
cold weather climates, allowing less heat loss.

     The typical jet aeration device consists of a directional
mix header with aerator nozzles on one side of the header.  The
header has a manifold containing separate air and liquid conduits
which distributes each flow stream uniformly to the nozzles.
Figure  2  shows the typical jet aerator device.

     The jet channels have found numerous applications for
industrial wastes where high uptake rates are a frequent design
occurrence.  The oxygen transfer efficiency is similar to that
of diffused aeration systems with rates between 4 to 5 pounds
oxygen/hp-hr being achieved in clean water.  Since the device is
a sub-surface aerator, it is probably subject to lower transfer
rates in dirty water, making the efficiency rate more equal to
that of a brush or a disc.  However, the looped reactor regime
in which it operates in, does give it a strong mixed efficiency
advantage over diffused aeration devices in conventional tanks.

Carrousel

     The Carrousel System, developed in the Netherlands in 1971,
has made a strong entry in the oxidation ditch market during the
last few years in this country.  Presently there are several
plants over 5 MGD operating successfully.

     The Carrousel basin uses a vertical shaft aerator with a
large diameter impeller and a low operating speed, usually less
than 30 rpm.  The aerator is located at every 180  turn of the
channel.  Using units as large as 100 hp, this aerator unit
would seem to have a cost advantage over the brush or disc
aerator in that it involves a lot less machinery and does not
span the full channel.  Another advantage is that, while the
Carrousel reactor might be quite new to the American consultant,
the low-speed vertical aerator is of a generic type that has
been familiar to the consultants in this country for a very long
time.

     Despite the numerous possible folds with the Carrousel basin
(which evidently depends upon the size of the basin and the

                                35

-------
CO
CTl
                  PRESSURIZED
                     AIR
                PRESSURIZED
                   LIQUID
                                                CONSTANT VELOCITY
                                                      CORE

                                                   HORIZONTAL JET FLUX
                                 Figure  2.   Jet  Aeration  Mixing Pattern

-------
number of individual  aerators  to be  used)  the  reactor  is  still  a
single loop system.   An  identical  arrangement  could be achieved
with a brush oxidation ditch by elongating this  ditch's straight-
away section and bending it over.  An elongated  travel section
does have the advantage  of allowing  some plug  flow charac-
teristics; for  large  basins, the potential for short-circuiting
is greatly reduced.   And if the aeration input zones are  kept
far apart  (as would be the case if high- oxygen input rates are
kept per volume of passing liquid) the  potential for anoxic
zones and denitrification between  the aerators are greatly
enhanced.  However, this same  advantage can result in  a dis-
advantage if the resultant effect  is high  in dissolved oxygen
levels immediately downstream  of the aerator.  Intense oxygen
input per volume of liquid passing through the aeration zone can
result in aerator driving force inefficiencies.

     The Carrousel is promoted by  Envirotech;  a  sketch of the
basin is shown  on Figure 3.   The oxygen  transfer rate of the
aerator is very good, being above  3.5 pounds of  oxygen/hp-hr.
Because of the  vertical  aerators location  at the end ,of the
channel and the partial  loss of mixing  expenditure, the mixing
efficiency is most likely slightly less than the brush or disc.

     The maintenance  requirements  of the low-speed aerator
device used with the  Carrousel is  similar  to that of other
conventional low-speed vertical-shaft aerators.

U-Tube Ditches

     The U-Tube or draft-tube  channel is promoted by Lightnin
and is unique in that it does  not  provide  an unrestricted flow
circuit like that found  with the other  types of  ditch  systems.
Oxygen delivery and mixing are independent with  this system.  A
draft-tube turbine forces the  channel flow through a conduit
below the channel floor  and behind a cross-sectional wall that
acts as a dam across  the channel width.  In order to achieve
sufficient channel velocities, the pumping rate  of the turbine
can easily be calculated; usually  the rate would be 100 to 200
times the incoming flow  rate.  Air spargers located just  below
the impeller of the turbine results  in  fine bubbles being forced
through the U-Tube.   Usually there is only one air application
point within the channel; this should enhance  the possibility of
anoxic zones at some  point further downstream.  Figure  4.

     Oxygen transfer  efficiencies  should be good because  of  the
increased pressures and  forced routing  through the U-Tube.
Mixing efficiencies,  however,  might  be  poor because of the dam
constriction across the  width  of the channel.  One strong
advantage with  this system is  the  fact  that all  of the channel
contents is forced through the U-Tube and  must mix with the
applied oxygen  after  each flow circuit.  As of yet, however,
there is no published data to  indicate  that passing the total

                                37

-------
co
00
                     R.A.S.
                                                               EFFLUENT
                                                          -^EFFLUENT

4
X SkCllUN A-A
V






                                    Figure 3.   Carrousel Layout

-------
                                          • Aerator
CO
VQ
               Baffles
               Air
                       Figure 4.  Aerator/U Tube  -  Longitudinal Section

-------
contents of the basin through the U-Tube has extra process
treatment benefits.

     Maintenance requirements with the U-Tube ditch can be
divided between the diffused aerators and turbine.  Maintenance
requirements on the diffusers should be low because of their
coarse bubble nature and because of the high air volume per
diffuser.

Counter Current Ditch

     This system, recently developed in Germany, is a newcomer
on the ditch market.  Using a circular ditch wrapped around a
clarifier, the Counter Current Ditch consists of a fine bubble
diffuser header rotating in a circular path around the channel.
The rotation provides an uni-directional mixing pattern and
localized upflow mixing each time the header passes through a
different zone.  Figure  5.

     The clean water oxygen transfer rate of fine bubble systems
have been shown to exceed 7 pounds of oxygen/hp-hr (5).  The
rotating arm of this system would probably slightly reduce the
energy savings over a conventional fine bubble system, although
higher oxygen transfer efficiencies have been claimed because of
the directional flow patterns and the resulting longer distance
that the bubble travels before surfacing.  Longer travel dis-
tance does not necessarily mean longer bubble contact time if
the upward velocity of the bubble component has not been altered.
More data will hopefully be available on transfer efficiencies
of this system in the future.  Like all fine bubble aeration
systems, the transfer efficiency rate is considerably less in
dirty water because of lower alpha factors  (6).

     The maintenance requirement for the Counter Current Ditch
can be potentially high when compared to other types of ditch
systems.  This would be because of the cleaning requirements
associated with the fine bubble diffusers.  Also, the rotating
air header has areas of potentially high maintenance.

Burns & McDonnell Treatment System

     This treatment system design is based upon a 1 MGD
demonstration study conducted in Missouri.  The ditch used an
in-channel clarifier to reduce capital costs.  The demonstration
program was quite successful under a wide variance of  load con-
ditions despite the lack of normal operating parameters
associated with external clarifiers - such  as return  sludge
pumping rates and sludge blanket level control.  Figure  6.

     One important development with this system is the unique
combination of fine bubble diffusers with low speed submersible
mixers.  The mixer used in the program had  an eight foot

                               40

-------
  Return Sludge
                                                    Effluent
Influent
  Return
  Sludge
                                                      X Effluent
             Figure  5.   Counter  Current Ditch
                                 41

-------
Mechanical
Skimming
         Hydraulic
         Skimming
                                                                Plant
                                                            >- Effluent
Waste
Sludge
            Figure  6.   Burns and McDonnell  Treatment System
                        With Intrachannel  Clarifier

-------
diameter impeller  and  a  30  rpm rotation  speed.   The motor  used
was only 3% hp  for a 300,000  gallon  basin;  considering  the con-
striction of the in-channel clarifier, the  mixing  efficiency  of
this device is  very impressive.  With the in-channel  clarifier
eliminated, the same 3^  hp  mixer could be used  for a  1,000,000
gallon basin and proper  mixing would be  provided.

     Although the  treatment system is very  new,  plants  are now
being designed  by  Burns  & McDonnell  for  full  scale operation.


DISCUSSION

     In discussing comparisons of the various systems,  it  is
important to note  that the  types of  equipment and  basin con-
figurations can overlap.  Several single loop oxidation ditches
operate using disc aerators,  while there are  some  orbal multi-
channel plants  operating using brush aerators.   There is even
one operating ditch plant in  Georgia with both  disc aerators  and
brush aerators  in  the  same  basin.  And while  all of the systems
previously noted,  except for  the orbal were single compartment
reactors, it is not difficult to imagine designs using  two or
more oxidation  ditches or Carrousels or  jet aeration  channels in
series, duplicating the  process flexibility advantages  of  the
orbal.  In fact, such  type  designs have  been  used  in  Europe (7).

     Many of the systems have particular areas  of  strength.   The
brush and disc  ditches are  more ideal for the smaller plants
 (though this by no means eliminates  them from consideration for
larger plants)  where basin  depths can be shallow and  the oxygen
input per aerator  is small.  The Carrousel  is more practical  for
plants larger than 1 MGD; it  would seem  that  plants larger than
50 MGD would have  concrete  intensity disadvantage  because  of  the
numerous folds  to  the  basin.   Jet aeration  channels can be
ideally used where deep  basins are required.

     The aerator types can  be divided between surface types and
sub-surface types. Sub-surface types will  usually give the
higher transfer rates  in clean water; in dirty  water, the
transfer rates  are not that far apart.   If  power costs
associated with oxygen delivery is going to be  an  important con-
sideration in the  design, aerator flexibility becomes important
for all of the  systems.  All  of the  looped  reactors have the
benefit of low  power required for mixing -  an important con-
sideration when taking into account  the  low-load nature of the
basin.

     Maybe the  most important area of energy  savings  is that  of
operating dissolved oxygen  levels.   Excessive levels  will  mean
an energy loss. Near-zero  levels are required  to  establish
anoxic zones and oxygen  recovery through denitrification.
Running looped  reactors  in  series can be an important energy

                               43

-------
saver; this can even be done with conventional low-loaded
aeration tanks in series  (7).  The savings can be as high as 30
to 40% if the first reactor is kept near zero.  If the last
reactor is kept small in volume and high in dissolved oxygen,
the process results will be maximized and the power costs
optimized.
REFERENCES

1.  "Comparison of Oxidal Plant to Competing Processes For
    Secondary and Advanced Treatment of Municipal Wastes", EPA-
    600/2-78-051  (March 19, 1978).
2.  "Rotor Aeration in the Oxidation Ditch - Bulletin 143",
    Lakeside Equipment Corporation, Bartlett, IL.
3.  "PFT Orbal System For Biological Treatment - Bulletin No.
    340-221", Envirex Inc., a Rexnord Company, Waukesha, WI.
4.  Applegate, C.S., Wilder, B. and DeShaw, J.R., "Total Nitrogen
    Removal in a Multi-Channel Oxidation System", Journal Water
    Pollution Control Federation, 52_, 568  (1980).
5.  Yunt, F., et al., "An Evaluation of Submerged Aeration Equip-
    ment - Clean Water Test Results", Paper presented at WWEMA
    Industrial Pollution Conference, Houston, TX (1980).
6.  Boon, A.G., "Oxygen Transfer in the Activated Sludge Process"
    In Proceedings:  Workshop Toward an Oxygen Transfer Standard,
    (Ed. W.C. Boyle), EPA-600/9-78-021, U.S. EPA, Cincinnati,
    OH  (1979).
7.  Matsche, N.F., "Nitrification-Denitrification in Full-Scale
    Treatment Plants in Austria", In Proceedings:   International
    Seminar on Control of Nutrients in Municipal Wastewater
    Effluents, EPA, Volume II, Coronado, CA  (1980).


DISCLAIMER

     The work described in this paper was  not funded by the  U.S.
Environmental Protection Agency.  The contents do not necessar-
ily reflect the views of the Agency, and no  official endorsement
should be inferred.
                                44

-------
TRANSLATION OF CLEAN TO DIRTY WATER OXYGEN TRANSFER RATES

Michael L. Doyle, Engineer, Rexnord Inc.,  Milwaukee,  Wisconsin  53706
William C. Boyle, Professor, Civil  Engineering,  University  of Wisconsin,
Madison, Wisconsin   53706

ABSTRACT

     The translation of clean water oxygen transfer rates to field conditions
represents a critical  step  in the design of aeration  systems.  This paper
reviews the parameters required  to  make this translation, discusses common
difficulties that arise and proposes methods that may facilitate more sound
design decisions.
INTRODUCTION

     Whereas  the  aerator  manufacturer  reports the performance of aeration
equipment as  the  oxygen transfer  efficiency  in  clean water at 0 D.O., one
atmosphere  pressure,  and  20°C,  it is the  transfer efficiency under actual
process  conditions  that is  of  practical concern in the design of aeration
systems.  It  is generally the  responsibility of the design engineer to trans-
late the manufacturer's clean  water transfer rate to that expected under
actual design conditions.   This paper  (a)  addresses the fundamental aspects
of  oxygen transfer  that should be recognized by the design engineer,
(b) points  out common misconceptions and  difficulties that arise when
translating clean water performance to field conditions and, (c) proposes
methods  which may facilitate more sound design  decisions.  Although geometry
effects  are important when  relating shop  test data to the actual aeration
tank design,  the  problem  of scale-up from one clean water system to another
is  not discussed  in detail  here.   However, the  influence of geometry on the
determination of  the field  oxygen transfer rate relative to the clean water
rate is  examined.   This paper  is  not written to provide correction factors
to  permit the quick and easy assessment of process performance of an aeration
device or system.   The evaluation of dirty water performance is a complicated
task with many areas of uncertainty that  still  need to be resolved.  Rather,
the authors hope  to shed  some  light on the variables that affect the perfor-
mance of aeration equipment in the field  and the manner in which those
variables may be  assessed.


FACTORS  AFFECTING DIRTY WATER  PERFORMANCE

     The standard clean water  transfer rate  (SOTR) as measured at 20°C with
a dissolved oxygen  concentration  of zero  may be related to actual field
conditions  (OTRf) according to the following equations:
                                  T-20
     OTR. =  (SOTR) (a) (   _*     )(9i  /U)                              (1)
         f                C20
                                      45

-------
           K    (field)
where  a = „    , . - r-                                        (2)
           K-   (clean water)


       B=   C*  (field)                                              (3)
          C*  (clean water)

             S20

 and     6T-20
                  la(20)

 (Note:   See  appendix  for  definition of the terms used above.)

 Beta

      The beta  factor,  sometimes referred to as the salinity correction,
 arises  due to  the "salting-out" effect of wastewater.  That is, dissolved
 salts and organics tend to lower the saturation concentration of oxygen
 wastewater compared to that of distilled water.  Although the activity of a
 saturated oxygen-water solution is by definition independent of dissolved
 contaminants,  the concentration changes as the activity coefficient is
 altered by the salting-out effect.

      The fact  that the activity is independent of the purity of the water has
 important implications in the measurement of the saturation concentration
 under field  conditions.   Direct determination of the oxygen concentration in
 wastewater by  the Winkler Method (1) is often complicated by oxidizing or
 reducing compounds which  prohibit use of this method.  Unfortunately a
 membrane covered  probe theoretically responds to activity which is dependent
 on the  degree  of  saturation - not the absolute concentration.  Therefore a
 probe standardized in  clean water will not yield a true reading of the
 dissolved oxygen  concentration in a contaminated water solution.

      To  overcome  these analytical difficulties, Bass and Shell  (2) recommend
applying  a correction  factor for temperature, barometric pressure, and dis-
solved solids  concentration.  Similarly, Standard Methods proposes an
equation  to adjust  the saturation D.O. to field conditions  (.1).

Theta

     Since the overall oxygen transfer coefficient, K- , is a function of
liquid viscosity  (among many other factors), a change in the liquid

                                     46

-------
temperature which brings a concomitant  change  in  viscosity will alter  the
transfer coefficient.  Since very  few aeration systems operate strictly at
20°C, a temperature correction  in  the form of  equation (6) is commonly used.

     Conflicting reports in the literature may be found concerning the value
of theta.  Temperature affects  not only the physical properties of the gas-
liquid interface, but also the  diffusivity of  oxygen in water.  Since  the
gas-liquid interface is influenced in different ways by various aeration
devices according their degree  of  turbulence,  mixing patterns, bubble  reten-
tion times and velocities, degree  of droplet formation and spraying, it may
be expected that temperature would affect each generic device to a varying
degree thereby producing different theta values.  Indeed, Hunter (3) reports
that the water temperature correction factor depends on the level of turbu-
lence and is, therefore, a function of  the aerator.  Hunter further cites
several articles by other researchers to support  his findings.  Kalinske (4)
claims that high energy systems will not exhibit  temperature effects to the
degree that low turbulence systems will and hence high turbulence aerators
will have a lower value of 9 than more  quiescent  aeration systems.

     In contrast to those findings are  the observations of Boon (5) and
Gilbert  (6).  Boon states that  there is approximately a 2% increase in K-
per  degree centigrade in the range from 0°C to  30°C for all types of
aerators.  Gilbert (6) presents data to show that the value of 6 ranges
between 1.024 and 1.026 which is roughly equivalent to a 2%/°C relationship.
Gilbert did not specify the type of aerator used  to generate this data.

     Despite the reported variation of  6 with  aerator type, there is cur-
rently wide acceptance of a value  of 1.024 for all aeration systems.
Caution should be exercised by  the engineer, however, when designing for
conditions outside of the range from 10°C to 30°C.  As the temperature
approaches these extremes the theta value becomes critical.
      The  overall  mass  transfer  coefficient, K   , is often dramatically
 altered under  field  conditions  due to dissolvea and suspended contaminants.
 This  ratio  of  the dirty water KL.  to the  clean  water K-  is probably the most
 controversial  and researched parameter that is  used in the translation of
 clean water data  to  actual  field  conditions.  Alpha factors as low as 0.2
 (7) and as  high as 1.5 (8)  have been reported in the literature.  Because of
 the extreme range of alpha, it  will normally have  the greatest effect on the
 translation of clean water  to dirty water transfer rates.  Variables
 affecting the  value  of alpha are  numerous and complex and include aerator
 type, nature of the  contaminants  in wastewater, level of turbulence, mixing
 patterns, liquid  depth, and air flow rate.

      Since  oxygen transfer  is a mass transfer operation involving both the
 dissolution of a  slightly soluble gas into a liquid as well as the transport
 of the oxygen  throughout the bulk of the  liquid, it is necessary to  examine
 the effect  of  contaminants  on both of these components of the mass transfer
 coefficient.
                                     47

-------
     Basically, it may be stated that contaminants do not affect the bulk
transport of oxygen (i.e. eddy diffusivity) to a great extent.  Although some
researchers (4) have pointed out that suspended solids and other compounds
alter the pumping characteristics of an aeration tank, this effect is small
compared to the impact of surfactants on the gas-liquid interface.  Casey
and Karmo (9) and Poon and Campbell (10) have indicated that suspended solids,
especially bio-solids, alter the clean water transfer coefficient, but this
effect is relatively minor over the range of mixed liquor solids maintained
in typical activated sludge systems.  In contrast, surfactants can exert a
significant influence on the gas-liquid interface by virtue of their hydro-
phobic end which is attracted to the gas phase while the hydrophilic head
remains drawn to the liquid.  This surface activity decreases the surface
tension of the interface which can result in the formation of smaller
bubbles in a diffused air system, and the production of smaller droplets and
greater gas entrainment in surface aeration systems.  More importantly,
these surfactants alter the hydrodynamic activity of the gas-liquid inter-
face.  Mancy and Barlage (11) have claimed that surface active agents form a
highly viscous layer (on the order of 10^ poise; approximately that of
butter) which causes a greater path length for oxygen in the liquid film
surrounding an air bubble (or liquid droplet).  This viscous layer also
reduces the rate of shear around the bubble which lowers the rate of surface
renewal (12).

     This surface effect will influence the transfer rate of different types
of aerators in vastly different ways.  Holroyd and Parker (13) found that
turbulent surfaces are less susceptible to surface active agents than are
more quiescent surfaces.  Kessener and Ribbius (7) reported alpha values of
0.2 for diffused air systems compared to 0.82 for brush aerators.  Systems
that continuously form fresh air-liquid interfaces through violent mixing
are usually not adversely affected by surfactants.  In fact, the formation of
smaller droplets along with greater bubble entrainment in a mechanical
aeration system can often result in alpha values greater than unity (8).  On
the other hand, oxygen transfer devices that promote transfer from the gas
phase to the liquid phase by virtue of a relatively gentle gas-liquid mixing
action may allow surfactants to accumulate at the interface with deleterious
results.

     Thus, surface aerators would not be expected to experience the sub-
stantially depressed alpha values that are frequently measured for fine
bubble diffusers.  At the same time, submerged diffusers may be affected to
varying degrees by surface active agents.  As previously mentioned, oxygen
transfer consists of gas to liquid transfer followed by transport throughout
the bulk liquid.  An aeration tank with a non-uniform diffuser arrangement
such as a spiral-roll coarse bubble system may be divided into an aeration
zone (in the immediate vicinity of the diffusers) and a pumping, or mixing
zone.  Since surfactants have only a minor effect on the transport of  oxygen
in the pumping zone, the reduction of alpha will depend on the relative con-
tribution of the aeration zone to K.. .  For aerators where the mixing
component of IL  is a significant factor, the alpha value may be  close to
unity.   However, it should not be assumed that a coarse bubble aerator will
necessarily be characterized by an alpha value greater than a fine bubble
aerator.   Downing and Bayley (14) have measured similar values of alpha for

                                     48

-------
both large and small bubbles rising  in  a narrow column.   The degree of bulk
mixing and the eddy diffusivity of oxygen  are  important  determining factors
of alpha, not just the bubble size.

     The design engineer should also  recognize that  alpha is subject  to
variation.  This variation may occur  on a  seasonal,  daily,  or even  hourly
basis.  Gilbert (6) has noted a variation  in alpha from  0.8 to 0.5  from hour
to hour in a brewing wastewater.  Eckenfelder  (15) has also shown a similar
variation for Kraft mill wastes.  While industrial waste variations may be
particularly extreme due to fluctuations in wastewater strength and composi-
tion, it is also possible for municipal treatment plants  to experience
variation in alpha values.  However,  the nature  of the wastewater (i.e. indus-
trial contributions, etc.) will make  the degree  of variability  site specific.
Doyle (16) conducted tests in a 10 foot deep,  1  foot diameter  column using  a
fine bubble ceramic diffuser and found  little  daily or hourly variation in
Milwaukee primary effluent.  Alpha values  were measured  in  the  range of
between 0.45 to 0.55 (see Figure 1).
        CL
       < .2
                         I   I
                   2   1   I   I   I   2

                  11/21    12/27 12/29  1/14
                                       1   I   2
                                      1/16   1/17
       I   I
 2111

I/IB    3/6 3/17
 Figure  1.
                       Sample  No.
                            Date
Fine Bubble System Alpha Measurements  - Mixed Liquor Influent;
Average of Replications (Depth - 10 ft.)
 These  tests were  performed  at  the Milwaukee  Jones  Island wastewater treatment
 plant  which receives  a  heavy industrial  load and,  hence, conclusions relative
 to  the variability  of alpha for  strictly domestic  wastewater cannot be made.

     An explanation for the relative  constancy of  alpha at this plant was
 provided by conducting  dilution  tests in which the mixed liquor feed was
 progressively  diluted with  tap water.  As Figure 2 reveals, even  small
 amounts of this wastewater  added to tap  water will lower alpha substantially.
 Apparently even low concentrations of the compound(s)  responsible for
                                     49

-------
    1.0

     .8
Alpha
     .4

     2.
    1.0

     .8

     .6

     .4
Alpha
    1.0

     £

     .6

     .4

     2

     .0.
             20      40      60
           %Clarifier Effluent
                                  80     100
             20     40     60
          % Mixed Liquor Effluent
                                  80
                                                       100
Alpha
                                 80
                                        100
                            20      40      60
                         % Mixed Liquor Influent
Figure 2.  Dilution Test Results  - Fine  Bubble  System (Tap Water as Diluent,
           Depth - 10  ft.)

depressing alpha in this waste  decreased K  .   Although these dilution tests
are primarily of academic interest, this method of  combining a "clean"
stream and a "dirty" waste  may  have a practical application - especially for
industrial wastes which are discharged on a  known schedule.  Alpha deter-
mination in a suitably designed column (see  further discussion in this paper
on bench scale testing) could be  performed on one component of the waste
discharge.  Then, various quantities  of  an intermittent, "dirty" stream could
be blended into the vessel  and  further additional alpha values measured.  If
the discharge volume and duration is  known for  the  "dirty" stream, a design
estimation of alpha could be made.

     In addition to temporal variation,  alpha has been found to change
during the course of treatment.   Lister  and  Boon (17) show a recovery in
alpha as the biodegradable  agents responsible for lowering the field K   are
removed.  The linear relationship shown  in Figure 3 is intended only to
illustrate a concept.  The  increase in alpha with treatment may not be linear
and will depend not only on the treatment kinetics  and tank hydraulics, but
also on the biodegradability of the contaminants.  Doyle (16) has noted a
general increase in alpha with  treatment as  shown by Figure 4.  These suc-
cessive batch tests were made with a  single  sample  drawn from the mixed
                       50

-------
        0-8
     <
     u_
     o

     UJ
     ID
       0-4
            Inlet
               Outlet
                    100       200      300       400

            OXYGEN  ABSORBED  BY SEWAGE DURING

            TREATMENT  (mg/l)

Figure 3.  Variation of Alpha with Degree of Treatment  (After Lister and
         Boon  (17)).
      1.0l-
       J8
       .6
 ALPHA
       .4
       .2
       .0
                                           Q   Q
                    12/29

                      ° 12/27
              I    I    I
I    I   I
           0123456

                   AERATION  TIME - Mrs.

 Figure 4.  Effect of Treatment on Alpha - Batch Tests (Depth - 10 ft.)
                             51

-------
liquor feed to an aeration tank and are representative of treatment in a per-
fect plug flow tank.  Note that for this waste, alpha does not seem to
increase in a linear fashion as treatment proceeds.  One test actually showed
a decrease and then a recovery in alpha as time progressed.  Other data
collected in this study failed to show a recovery to unity in alpha with
treatment, and even final clarifier effluent was found to yield alpha values
no greater than 0.7.  This finding is not unusual in light of McKinney's (18)
observation that many surfactants, for example the glycols, are resistant to
bio-degradation.  While treatment plants which experience depressed alpha
values due to biodegradable surfactants may experience a recovery in alpha
similar to the relationship shown in Figure 3, those plants treating waste-
water with biologically "hard" surfactants may show little, if any, improve-
ment in alpha.  Again, this phenomenon is highly site specific.

     In summary, the alpha factor is a complex parameter influenced by a
number of process variables:  type, tank geometry, mixing of contaminant,
and level of turbulence.  Although some generalizations may be made regarding
the effect of contaminants on alpha, the lack of a unifying theory for these
variables and a paucity of knowledge concerning the mechanism that deter-
mines alpha makes the prediction of the field K   very difficult.  The fact
that alpha is not only different for different generic types of aerators,
but also can vary within a given system is an important design concept.
TRANSLATING CLEAN WATER PERFORMANCE TO THE FIELD

     The ideal method by which to evaluate the field performance of an
aeration system is to test the equipment at design conditions in the aeration
basin.  For plants undergoing expansion this method is often quite feasible.
In that case, the choice of a suitable method for measuring oxygen transfer
in a respiring system must be addressed.  Detailed discussion of possible
methods are beyond the scope of this paper and may be found elsewhere (19).
In most cases the engineer must be able to adapt the manufacturers' informa-
tion to the field through the use of equation (1) and the judicious choice
of alpha, beta, and theta.  Assuming that beta can be determined by a tem-
perature, pressure, and salinity correction or some other proper method, and
theta  is taken as approximately 1.024, the critical design factor will be
alpha.  Basically two methods are being utilized by many engineers.  One
"technique" is the use of past design experience and/or reported literature
values for alpha.  The other technique consists of performing bench or pilot
scale aeration tests in clean water and dirty water to determine alpha.
Both methods clearly have their benefits when adequately applied but also
have severe drawbacks when used inappropriately.

Bench Scale Studies

     In an attempt to gain information specific to the waste to be treated,
especially unique industrial wastes, bench scale tests are often performed.
Various authors (2, 20, 21) have described bench scale alpha units.  A bench
test is useful only if it can be translated to full scale operation.
Unfortunately this is a difficult undertaking due to the extremely complex
components of mass transfer.  Accordingly, the literature is replete with

                                     52

-------
reported failures to duplicate the full scale alpha on a small scale.
Barnhart (22), Otoski et al. (23), Schmit et al.  (24), and Gilbert  (6) have
all experienced an inability to match alpha on the large and small scale.
While Stukenberg (20) states that equilibrating the clean water oxygen
transfer coefficient in the small vessel and large tank will yield similar
alpha values during dirty water testing, it must be realized that most
researchers have found that greater air/volume ratios must be used in the
small scale units due to decreased transfer efficiency at lower liquid depths
(for diffused air systems, see Figure 5).  Dissimilar air/volume ratios can
     30
     25
SOTE
     20


     15

     10
       0           2          4           6           8          10
                                 DEPTH (Ft.)
 Figure 5.   Standard Oxygen Transfer Efficiency vs.  Depth - Fine Bubble
            (Air Flow - 0.5 SCFM/Diffuser).

 cause disparate mixing intensities which will result in different determina-
 tions of alpha.  Mixing, in general, is a difficult property to scale,
 especially when attempting to maintain similitude with other aspects  of mass
 transfer.  The scaling of a surface aerator is complicated by variables such
 as bubble entrainment and drop size.  Shell (25) has stated that a mechanical
 aerator is impossible to scale correctly.  It is fortuitous that the  deter-
 mination of alpha for a mechanical aerator is not highly critical since the
 value is fairly close to unity.

     Although simply matching K.  on the large and small scales does  not
 appear to be an adequate criteril for the design of a bench scale model, and
 although mechanical mixing may not lend itself to modeling, the design of a
 small scale vessel for certain aerators appears to be feasible.  For  a fine
 bubble diffuser system employing uniform floor coverage, the diffusers have
 often been observed to act as separate mixing cells (26).   This type  of
 mixing may lend itself to modeling in a tall, narrow column.  Field condi-
 tions such as type of diffuser, air flow rate, percent floor coverage of the
                                     53

-------
stones, and liquid depth should be duplicated as closely as possible.   The
use of a shallow test vessel can produce misleading estimates of alpha.
Doyle (16) found that alpha was a function of depth in studies conducted
in a one foot diameter column using a ceramic fine bubble diffuser (See
Figure 6).  Values of alpha (0.45-0.55) in the mixed liquor f§ed (measured at
       1.01-
        .8
        .6
 ALPHA
        .4
        .2
        JO
                                                   8
10
                    246
                                 DEPTH (Ft.)
 Figure  6.  Alpha vs.  Liquid Depth - Fine Bubble (Final Effluent).

 a  10  foot  liquid depth) were similar to those found by Lister and Boon (17)
 in full scale  tests.  Although it may be desirable to operate at a liquid
 depth even closer  to  the actual aeration tank depth, clearly a tall column is
 an improvement on  the apparatus typically recommended in the literature.

      A  tall, narrow column will not be a suitable vessel for modeling all
 types of diffused  air systems, however.  Morgan and Bewtra (27) stated that
 a  spiral-roll,  coarse bubble aerator cannot be successfully modeled in a
 narrow  column.  Results from Doyle's work substantiate this finding.  A
 coarse  bubble  aerator was simulated by installing a plate with two 1/16"
 holes in the bottom of the column.  A value of alpha (0.4-0.5) equivalent to
 that  of  the fine bubble system resulted.  Since bulk mixing was largely
 eliminated by  the  restrictions of a narrow column, this component of K
 which is largely unaffected by surfactants was not incorporated in the
 column tests.   Hence, alpha was probably lowered below levels that would have
 been  observed  had  this mixing component been correctly modeled.

      To summarize, the bench scale apparatus typically recommended in the
 literature for  the determination of alpha has often been found to be inade-
 quate.  However, there are certain applications for small scale test vessels
which can be used  to  model some types of aeration systems.  In particular, a
                                     54

-------
tall, narrow column utilizing an actual commercial diffuser and operated at
close to the design conditions of liquid depth and air flow rate has been
found to be useful In assessing alpha for a fine bubble diffuser.  Attempts
to model aeration systems where mixing and liquid pumping comprises a
significant portion of K^  are more complex.  Proper scaling of these
properties is vital when attempting to evaluate alpha on the small scale.
Further research into this type of modeling is necessary if small scale
estimates of alpha for these aeration systems are to be made with reliability.

Engineering Estimation

     The use of past experience and/or reported literature values of alpha, a
technique referred to herein as "engineering estimation", is probably the
most common method currently employed for selecting a design value for alpha.
A thorough review of the literature indicates that there is very little
information available on actual field measurements of alpha in any type of
wastewater.  It is likely that a significant number of engineers use alpha
values quoted in standard textbooks or journal articles.  One of the chief
references often cited is Eckenfelder (15).  Eckenfelder, in 1959, listed
alpha values for many industrial wastes as well as municipal wastes, these
results all gathered from tests in a 1 liter (or 4 liter) vessel.  There is
no evidence in the literature that these values have ever been corroborated.

     Unfortunately, the verification of alpha values collected from the
laboratory is most difficult to achieve since methods used to estimate field
transfer efficiency in dirty water are often imprecise and inaccurate.
Furthermore, in order to accurately assess alpha one must also have field
clean water transfer data as well.  Probably one of the earliest field tests
for  alpha was performed by Kessener and Ribbius (7) in 1933.  They found
alpha values of 0.2 for a diffused air system and 0.8 for a brush aerator,
both in sterilized wastewater.  Since that time, little work was reported
until the Water Pollution Research Laboratory in Stevenage, England began
extensive testing in large tanks in the late 1950's.  Much of this work was
done with anionic detergents since it was felt that the surfactant in waste-
water played the most important part in depressing oxygen transfer rates.
Parallel tests were also performed with mixed liquor.  Since that time,
numerous other investigators have tested a variety of aeration devices in
large tanks employing anionic detergents (17, 28, 29, 30, 31).  A summary of
much of this data is presented in Table 1.  It should be emphasized that the
values of alpha obtained from detergent tests may give an indication of the
depression of the oxygen transfer rate that will be observed in wastewater,
but  they should not be considered as absolute.

     Field testing in dirty water, as mentioned earlier, is a far more
complex problem than detergent testing.  The biological uptake rate in the
wastewater creates an analytical problem which may result in significant
error (19).  Testing with or without mixed liquor solids adds to the diffi-
culty.  Finally, spatial and temporal wastewater variability creates further
uncertainty as to the true value of alpha to use for design purposes.

     Results of large scale tests for alpha in dirty water appear in Table  2.
Steady state testing by measuring oxygen uptake rates in mixed liquor  during

                                     55

-------
Table 1.  Alpha Values Determined in Large Systems Employing Detergents'
Aerator
Type
Fine
Fine
Fine
Bubble
Bubble
Bubble
Tank Size
(gal)
A8000
48000
39000
Characteristics
Domes,
Domes,
Saran
Subm -
Subm -
20 ft.
27 ft.
& Brandol
Alpha
0.
0.
0.
48
37
6
Reference
Lister
Lister
von der
& Boon
(17)
& Boon
(17)
Emde
Medium/Coarse       39000
 Bubble
Coarse Bubble       35000
Coarse Bubble       35000
Coarse Bubble       87000
tubes; Plastic box;
Subm s 12 ft.
Perforated Tubes
Subm ~ 13 ft.
Air Flow-25 scfm/
1000 ft3; Subm - 20
ft.
Air Flow-90 scfm/
1000 ft3; Subm - 20
ft.

Air Flow - 80-190
scfm/1000 ft3;
Subm - 10 ft.
                     (28)

0.7-0.8   von der Emde
                     (28)
  0.65    Schmit et al.
                     (24)

  0.75    Schmit et al.
                     (24)

0.7-0.8   Otoski (23)
Cage Rotor 50000
Simplex Aerator 500000 Compared against
Secondary Effluent
Simplex Cone Full Scale
Kessner Brush Full Scale
Surface Aerator Full Scale
Static Tubes 87000 Gs 10-190 scfm/
1000 ft3
1.2 Downing et al.
(29)
> 1.1 Downing et al.
(29)
1.1 von der Emde
(28)
1.0-2.0 Baars (30)
0.93 Boon (31)
1.1 Otoski (23)
  All tests with tap water and detergent  (5 mg/1); oxygen  transfer  deter-
  minations by unsteady state procedure.
                                     56

-------
    Table 2.  Alpha Values  Determined  in  Large  Systems Employing  Wastewater
en
—i
Aerator Test Procedures Tank Size Characteristics Alpha
Type
** Clean Water Dirty Water (gal)
Fine Bubble NSS
(Domes)
Fine Bubble NSS
(Domes)
Fine Bubble NSS
(Discs)
Fine Bubble ?
(Tubes)
Fine Bubble Calc.
(Domes)
Fine Bubble Calc.
(Discs)
Fine Bubble Calc.
(Tubes)
Coarse Bubble Calc.
Coarse Bubble Tracer
Oxidation Ditch NSS
(Mech. Aer.)
Oxidation Ditch NSS
(Mech. Aer.)
SS/OG 4800 Mixed Liquor,
inlet , Subm -
20 ft.
SS/OG 4800 Mixed Liquor,
inlet, Subm -
27 ft.
Tracer Full Scale Mixed Liquor
? 39000 Final Effluent
Subm - 13 ft.
O.G. Full Scale Mixed Liquor
Subm - 15 ft.
O.G. Full Scale Mixed Liquor
O.G. Full Scale Mixed Liquor
Cross roll
O.G. Full Scale Mixed Liquor
Cross roll
0.49-0.56
0.43-0.49
0,4-0.5
0.6
0.42-0.98
(inf -> eff)
0.34-0.62
(inf •> eff)
0.46-0.85
(inf •> eff)
0.83-1.19
(Random)
Tracer 660,000 Chemical Waste; 0.6-0.65
Mixed Liquor; Floor
grid; Subm - 19 ft.
SS 194,000 Mixed Liquor
(Extd Aer)
Subm - 1.5 ft.
SS 179,000 Mixed Liquor;
Refinery; Disc
Subm - 1.75 ft.
0.95-1.05
0.98-1.03
Reference
Lister & Boon (17)
Lister & Boon (17)
Kiiskinen (35)
von der Emde (.28)
Redmon et al. (34)
Redmon et al. (34)
Redmon et al. (34)
Redmon et al. (34)
Campbell (19)
Huibregtse & Doyle
(32)
App legate &
Huibregtse (33)
     NSS  - Non  steady  state;  SS  -  steady  state; O.G. - off gas

-------
steady state  conditions was performed by Lister and Boon  (17), Huibregste
and Doyle  (32) and Applegate and Huibregste  (33).  Off gas analysis whereby
inlet and  outlet oxygen and CO- are analyzed was employed by Redmon et al.
(34), Campbell (19), and Lister and Boon (17).  The radioactive tracers
Krypton-85 and tritiated water were used in field tests reported by
Kiiskinen  (35) and Campbell (19).

     From  the standpoint of accuracy and precision of testing under dirty
water conditions, it is felt that tracer techniques and off gas methods
offer great potential.  Recent work with a dual non-steady state procedure
employing  hydrogen peroxide may also be effective (36, 37).  Since laboratory
scale tests have very limited usefulness in estimating dirty water alpha,
field dirty water testing  (following clean water tests) with a variety of
aeration devices and wastewaters will have to supply the data base needed
to provide the field with  useful alpha values.


SUMMARY

     It is clear that the  state of the practice in translation of clean to
dirty water oxygen transfer rates is still in its infancy.  With the escala-
tion of energy costs, the  need for efficient aeration systems is apparent.
However, aeration system design appears to be lagging behind aeration device
technology.   Many new questions are being raised regarding the design and
operation  of  aeration systems in light of changing constraints on that
design.  Conservative over-design is not necessarily the answer since it may
lead to serious energy consuming inefficiencies.  It will be necessary to
better refine design practice and specifications so that designs more closely
fit field  situations.  This means that the profession must follow up on
designs with  careful field evaluation.  This is not a new concept - Conway
and Kumke  (21) made a similar plea in 1966 to the profession.  Improvements
in testing of oxygen transfer devices under clean and dirty water conditions
will go a  long way in providing the tools necessary to accomplish a part of
this job.
APPENDIX

 *
Con  - dissolved oxygen saturation concentration at 20 C at given submer-
 JL\J             I T ^
       gence, m/LJ.                                  3
C*   - dissolved oxygen saturation concentration, m/L .
C    - dissolved oxygen saturation concentration at 1.0 atmosphere and  100%
       relative humidity, m/L  .
C    - effective average dissolved oxygen concentration in process liquid,
       m/L3.
K.    - overall liquid phase mass transfer coefficient for oxygen in water,
 la
d    - effective saturation depth at infinite time, L.
P,    - atmospheric pressure at field condition, f/L .
P    - standard atmospheric pressure, usually 1.0 atmosphere,  at  100%  rela-
 S     tive humidity, f/L2.
                                     58

-------
 P    - saturated vapor pressure of water,  f/L2.
 T    - temperature,  °C.
 OTR  - oxygen transfer rate under field  conditions,  M/T.
 SOTR - standard oxygen transfer rate, M/T.
 Y    - weight density, M/L3.


 REFERENCES

 1.  "Standard Methods for the Examination  of Water and Wastewater."   15th
     Ed. Amer. Publ.  Health Assn., Washington, D.C. (1981).
 2.  Bass, S. J. and  Shell, G., "Evaluation of Oxygen Transfer  Coefficients  of
     Complex Wastewaters", Proceedings of the 32nd Annual Purdue  Industrial
     Waste Conference, 953 (1977).
 3.  Hunter, J. S., "Accounting for the Effects of Water Temperature  in  Aera-
     tor Test Procedures". In  "Workshop Toward an Oxygen Transfer Standard",
     Pacific Grove, California, EPA 600/9-78-021, 85  (1979).
 4.  Kalinske, A. A., "Problems Encountered in Steady State Field Testing of
     Aerators and Aeration Systems".  In  "Workshop Toward an  Oxygen Transfer
     Standard", Pacific Grove, California,  EPA 600/9-78-021,  208  (1979).
 5.  Boon, A. G. , "Oxygen Transfer in the Activated Sludge Process".   In
     "Workshop Toward an Oxygen Transfer  Standard", Pacific Grove,  California,
     EPA 600/9-78-021, 232 (1979).
 6.  Gilbert, R. G.,  "Measurement of Alpha  and Beta Factors".   In "Workshop,
     Toward an Oxygen Transfer Standard", Pacific Grove, California,  EPA 600/
     9-78-021, 147  (1979).
 7.  Kessener, H. S.  N. and Ribbius, F- J., "Practical Activated  Sludge
     Research", J.  Institute of Sewage Purification,  Part I,  50 (1935).
 8.  Downing, A. L. and Boon, A. G., "Oxygen Transfer in the  Activated Sludge
     Process".  In  "Proceedings of the 3rd  Manhattan  College  Conference  on
     Biological Waste Treatment", Pergamon  Press, 123 (1963).
 9.  Casey, T. S. and Karmo, 0. T., "The  Influence of Suspended Solids on
     Oxygen Transfer  in Aeration Systems",  Water Research, j8, 805 (1974).
10.  Poon, C. P. C. and Campbell, H., "Diffused Aeration in Polluted  Water",
     Water and Sewage Works, 114, 461 (1967).
11.  Mancy, K. H. and Barlage, W. E., "Mechanism of Interference  of Surface
     Active Agents  in Aeration Systems".  In "Advances in Water Quality
     Improvement" (Ed. by Gloyna, E. F. and Eckenfelder, W. W.),  Univ. of
     Texas Press, 262 (1968).
12.  Mancy, K. H. and Okun, D. A., "Effect  of Surface Active  Agents on
     Aeration", J.  Water Pollution Control  Federation, 37, 212  (.1965).
13.  Holroyd, A. and  Parker, H. G., "Investigations on the Dynamics of
     Aeration", J.  Institute of Sewage Purification,  280  (1952).
14.  Downing, A. L. and Bayley, R. W., "Aeration Processes for  the Biological
     Oxidation of Waste Waters", The Chemical Engineer,  (Oct.,  1961).
15.  Eckenfelder, Jr. W. W., "Factors Affecting the Aeration  Efficiency  of
     Sewage and Industrial Wastes", J. Water Pollution Control  Federation,
     31, 60 (1959).
16.  Doyle, M.L.,"Small Scale Determination of Alpha  in a Fine  Bubble Diffuser
     System". Master's Thesis, University of Wisconsin, Madison,  Wisconsin
     (1981).


                                      59

-------
17.  Lister, A. R. and Boon, A. G., "Aeration in Deep Tanks:  An Evaluation
     of a Fine Bubble Diffused-Air System", J. Institute of Water Pollution
     Control. 72_, 590 (1975).
18.  McKinney, R. E., "Syndets and Waste Disposal", Sewage and  Industrial
     Wastes Journal, 29_, 654 (1957).
19.  Campbell, H.  "Specification and Compliance Testing - Dirty Water
     Testing".  In "Seminar/Workshop - Aeration System Design,  Testing,
     Operation and Control", Madison, WI, Aug. 2-4, 1982.
20.  Stukenberg, J. R., Wahbeh, V. N. and McKinney, R. E., "Experiences  in
     Evaluating and Specifying Aeration Equipment", J. Water Pollution
     Control Federation. 49_, 66 (1977).
21.  Conway, R. A. and Kumke, G. W., "Field Techniques for Evaluating
     Aerators", J. ASCE - Sanitary Engineering Division, 92, SA2, 21 (1966).
22.  Barnhart, E. L., "Transfer of Oxygen in Aqueous Solutions", J. ASCE -
     Sanitary Engineering Division, 95, SA3, 645 (1969).
23.  Otoski, R. M., Brown, L. C. and Gilbert, R. G., "Bench and Full Scale
     Studies for Alpha and Beta Coefficient Variability Determination",
     Proceedings of the 33rd Annual Purdue Industrial Waste Conference,
     May 9-11, 835  (1978).
24.  Schmit, F. L., Wren, J. D. and Redmon, D. T.,  "The Effect  of Tank
     Dimensions and Diffuser Placement on Oxygen Transfer", J.  Water Pollu-
     tion Control Federation. 50, 1750 (1978).
25.  Shell, G. L., "Discussion of Working Group Summary Reports".   In
     "Workshop, Toward an Oxygen Transfer Standard", Pacific Grove, Califor-
     nia, EPA 600/9-78-021, 258 (1979).
26.  Pasveer, A., "Considerations of the Efficiency of the Aeration Process",
     J. Air and Water Pollution Institute, 10  (1966).
27.  Morgan, P. F. and Bewtra, J. K. , "Air Diffuser Efficiencies",  J. Water
     Pollution Control Federation, 32, 1047  (1960).
28.  von der Emde, W., "Aeration Developments  in Europe".  In "Advances  in
     Water Quality Improvement" (Ed. Gloyna, E.F.  and Eckenfelder,  Jr. W.W.)
     Univ. of Texas Press, 237  (1968).
29.  Downing, A. L., Bayley, R. W. and Boon, A. G., "The Performance of
     Mechanical Aerators", J. Inst. of Sewage  Purification, 3  (1960).
30.  Baars, J. K.  "The Effect of Detergents on Aeration:  A Photographic
     Approach", J. Proc. Inst. Sewage Purification, 358  (1955).
31.  Boon, A. G., "Measurement of Aerator Performance",  Symp. on  The
     Profitable Aeration of Wastewater, London, April 25,  1980, BHRA Fluid
     Engineering, Cranefield, Bedford UK, 13  (1980).
32.  Hubregtse, G. L. and Doyle, M. L., "Full  Scale Alpha  Determination  for
     Orbal Aeration Systems", Rexnord, Inc. Test Center  Report  (Jan.,  1982).
33.  Applegate, C. S. and Huibregtse, G. L.,  "Orbal Performance Tests  -
     Comparison of Two Types of Discs", Rexnord, Inc. Test Center Report
     (July, 1978).
34.  Redmon, D. T., Boyle, W. C. and Ewing, L., "Off  Gas Analysis for  Oxygen
     Transfer Testing in the Field", In Press.
35.  Kiiskinen, S., "Comparison of Different Aerators for  Diffused Aeration",
     Prog, in Water Technology, 11,  139  (1979).
36.  Mueller, J. A. and Rysinger, J. J., "Diffused Air  Testing Under Process
     Conditions", Proc. 36th Purdue  Industrial Waste  Conference,  Purdue
     Univ., 747 (1981).
                                     60

-------
37.  Mueller, J. A., "Comparison of Dual Non Steady State and Steady State
     Testing of Fine Bubble Aerators at Whittier Narrows Plant, Los Angeles",
     In "Seminar/Workshop - Aeration System Design, Testing, Operation and
     Control", Madison, WI, Aug. 2-4, 1982.


 DISCLAIMER

      The work described in this paper  was  not  funded  by the U.S.  Environ-
 mental Protection Agency.   The contents  do not necessarily  reflect  the views
 of the Agency,  and no  official endorsement should  be  inferred.
                                     61

-------
AERATION SYSTEM SCALE-UP

Norbert W. Schmidtke, President
Norbert W. Schmidtke & Associates Ltd.
178 Louisa Street, Kitchener, Ontario
Canada N2H 5M5

ABSTRACT

     The approach to aeration system scale-up and the degree of scale-up
success for different types of aeration systems is discussed.   Scale-up
examples  for surface turbines, cage rotors, diffusers, spargers, static
systems as well as RBC system aeration are presented and discussed.

     Progress in the area of aeration system scaling has been made,
Predictions of oxygen transfer in full scale systems based on models
developed from laboratory scale data have met with varying degrees of
success.
 INTRODUCTION

      Scale-up formulations for full-scale process design from data generated
 by way of prototype plant simulation at bench-scale have been the topic of
 many  investigations.  This is particularly true in the design of a variety
 of chemical engineering processes.  Advances in this direction in the
 wastewater treatment process design field are few.  Aside from the obvious
 benefits of having a reliable scale-up transform for a particular process,
 a further and perhaps even more important application would be the ability
 of scaling down process difficulties encountered in prototype plant operation
 and replicate the field condition in the laboratory for further study under
 controlled conditions.

      Since this workshop concerns aeration, it would be appropriate to direct
 the focus of this presentation to the problems of and directions various
 attempts of devising scale-up functions for oxygen transfer into water have
 taken.  By addressing the problem of oxygen transfer into water, and even
 if we are successful in scaling this process up or down, only one half of
 the problem is solved.  For aerobic biological treatment to be successful
 the oxygen must also be transported to the cell surface for diffusion into
 the cell.  There appears to be agreement that for biochemical reactions this
 represents the rate limiting step, provided of course that adequate substrate
 and nutrients are available.  The transfer of mass into and out of cells can
 be enhanced, up to a limiting value, through the creation of turbulence.
THE ROLE OF TURBULENCE

     It is this turbulence phenomenon which results not only in the transfer
of oxygen into cells but keeps the mixed liquor solids in suspension.  We
know that the energy input through an aeration device into a liquid system
is translated into liquid movement, and that beyond a certain minimum energy

                                     62

-------
input turbulent flow results.  One manifestation of turbulence is eddies
which come in all sizes.

     In considering turbulence, two aspects are dominant...the scale of
turbulence and the intensity of turbulence.  The scale of turbulence refers
to the mean size of eddies, while intensity of turbulence to the speed of
angular rotation of the eddies.

     In models, the scale of turbulence is small and the intensity of
turbulence is high.  For prototype systems, the opposite is true.  Most of
the energy imparted to a system through an aeration device will result in
large eddies.  They produce the main flow in the aeration tank.  But large
eddies have a geometric limitation.  For instance they can not be larger
than the liquid depth in the aeration tank.  Ultimately eddies are
dissipated.  Thus large eddies must become smaller arid smaller until finally
they become extinct.  It would stand to reason then that a whole continuum
of eddy sizes between these two extremes exists.  The dissipation of this
energy is through friction (heat) caused by the general break-up of eddies
due to collisions.  Of greatest importance in biochemical processes are the
smallest eddies which are capable of reaching the cells and thus are
instrumental in the more rapid diffusion of oxygen, nutrients, substrate
into and metabolic waste products out of the cells.  The total turbulence
energy is distributed among the large and these small eddies.

     Requirements for successful process scaling concern meeting a variety
of similarity criteria between model and prototype namely geometric,
dynamic, kinematic and environmental similarity.  One requirement for equal
process results in small and large sized tanks would be geometric similarity
in both turbulence scale and intensity.  This similarity should of course
extend to oxygen bubble size, the microorganisms and the floe particles.
However, we are unable to scale or control the geometry of any of these.

     Having recognized the importance of turbulence in mass transfer
processes and the difficulties associated with its geometric similarity we
are then faced with our greatest challenge in process scaling.  How can we
ensure that when examining a process at different scales the same turbulence
level is retained?  Is it possible to scale turbulence?  If the answer is
no, can these differences in turbulence levels between different sized
systems and their concomittant effects on process results be accounted for
in some other way?

     It would appear reasonable that by involving a process response
parameter, such as the overall oxygen transfer coefficient, K^a, successful
process scaling might be possible.

     If environmental and biological conditions can be kept identical, one
should be able to scale bio-hydraulic systems provided mass transfer rates
as expressed by the overall oxygen transfer coefficient (Ki,a) in both model
and prototype are equal.  Equal process results should be  attainable.
                                     63

-------
KLa AS A SCALING PARAMETER

     The overall oxygen transfer coefficient, I^a, is dependent on a variety
of variables, and correlations of variables affecting K,a have been
presented (l-ll).  The number of variables influencing the magnitude of
K^a are also dependent on the type of aeration device.  In recent years a
number of researchers have made an attempt at using K^a in a process scale
up function.

     A brief review of approaches to scale-up for surface aeration devices,
diffused aerators, static aerators and cage rotors is presented.  As well,
considerations for scale-up of rotating biological contactors are high-
lighted.  Fundamental to any investigation of this nature is an identi-
fication of operating and system variables for the various aeration system
alternatives.
SURFACE AERATION

Scale-Up Function

      It would be well then, to start by looking at a surface aeration
system, identify and examine all possible variables which influence
and  then go on to the other aeration system alternatives.  It will become
apparent that a number of justified, simplifying assumptions can be made.

      Variables influencing fluid motion in an unbaffled, surface turbine-
agitated square tank can be identified as:
            fn (W, D, HI, DI, WI,  LI, B, S,  p,  v,  g, N,  P)   =0         Q)

For  oxygen transfer, the corresponding influencing variables are:
            fn (KLA, DI, HI, N,  v,  DL>  g,  P, a,  P)   =  0               (2)

      In order to arrive at a general relation between variables influencing
Ki^a  in a surface aerated turbulent flow regime, variables associated with
fluid motion and mass transfer were combined into one functional
relationship:
  fn  (W, D, HI, DI, WI, LI, B, S,  p, v, g,  N, P,  KLA, DL, a)  =  0      (3)

where:

           W = a significant tank dimension              (L)
           D = liquid depth                              (L)
          HI = impeller immersion in liquid              (L)
          DI = impeller diameter                         (L)
          WI = impeller blade width                      (L)
          LI = impeller blade length                     (L)
           B = number of impeller blades                 dimensionless
           S = impeller pitch                            (L)
          p = fluid density                             (FL~ TZ)
           1> = fluid kinematic viscosity                 (LT1' )
           g = gravitational constant                    (LT   )

                                     64

-------
           N = impeller speed                            (T  )
       .  KLA = ICa = overall oxygen transfer coefficient(T~ )
           P = power input                               (LFT~1)
          DL = molecular diffusivity of oxygen           (L2!'1)
          
-------
                      -  -HI  M
                      £n %r NFr»

 If  one  goes one  step further and  assigns  a  constant  value  to  the
 immersion simplex (HI/Dl)  for both model  and  prototype,  the equation  is
 further simplified.   The equation also  indicates  that  I^a  as  a  dependent
 variable, is a function of the Froude number  (Npr) containing the  independent
 variables of impeller speed (N),  impeller diameter (Dl)  and the gravitational
 constant, g.
                         KLA  =   C1   (Npr)                             (6)

 Equation (6) can be  used to describe the  oxygen transfer characteristics of
 each system.  For the development of a  scale-up prediction model one  can
 explore the suitability of a power relationship such as:


                         KLA  =   0 N* DIb                              (?)
 where:
            N = impeller speed              (T~l)
           DI = impeller diameter            (L)
     0,  a, b  = constants

      If one postulates  that scale-up of surface agitated (turbine), complete-
 mixing,  unbaffled, square,  geometrically  similar  reactors  can be achieved
 when the overall oxygen transfer  coefficient  in model  and  prototype are
 equal,  then a scale-up  equation can be  developed  as  follows:


                     P  model   '   (KLA2(P prototype
from (7)

                 a                    "prototype
                                  ^"'"'rotote                     (9)
which  upon  simplification  reduces to —
                                       DI  C
                           Nm  =  N   (-£)                            (10)
 ,
where:
          Nm = model  impeller  speed              (T  )
          N  = prototype  impeller  speed          (T~ )
         DID = prototype  impeller  diameter       (L)
             = model  impeller  diameter           (L)
             = b/a =  a constant

     In order to ascertain whether a scale-up relationship such as equation
(10) can be developed and whether  conditions for equal process results
between different sized but geometrically similar vessels can be generated,
a total of 189 experiments using three geometrically similar laboratory- scale
                                     66

-------
                 f  Sh°Wn1in FiSure 1 were conducted.  These  studies  are
               n  ^      el^here  (7).   Table  1  indicates  the ranges  for the
               conditions imposed  for the  study.
Reactor sizes used for
scale -up mode) development
i
10°
* *
M'1 M~2
i
10'
I
M3
i
102
Liquid
Reactor size used to test
validity of scale-up model
P1
t
103
volume filters)
. REACTOR SIZE USED TO TEST
i VALIDITY OF SCALE-UP MODEL
| FOR A fiEOMETRICALLY
j DISSIMILAR SYSTEM
! /
j-^. Full scale
reactors
I i
10s 10s

Figure  1.   Spectrum of reactor sizes  for  scale-up model testing.
     TABLE 1.  RANGES FOR THE EXPERIMENTAL CONDITIONS
MODEL
M.I
M.2
M.3
P.I
NO. or
EXPERIMENTS
106
50
33
26
ENVIRONMENTAL
CONDITIONS
TS*
17.70-19.88
17.60-19.33
17.97-19.80
16.88-21.10
LIQUID MEDIUM
Tapwater
Tapwater
Tapwater
Tapwater
HYDRAULIC
CONDITION
VOLUME
LITERS
5.13
19.7
113
607
HYDRODYNAMIC
CONDITION
PPM
234-473
195-331
144-202
50-132
NFr
0.084-0.342
0.091-0.261
0.090-0.176
0.019-0.132
NO. OF
D1FF.
SPEEDS
57
33
5
19
 General Scale-Up Model

      It had been shown in the analysis of variables influencing mass  transfer
 for this study,  that within specific boundary conditions,  KLa  has  a functional
 dependence on the impeller speed N and the impeller diameter DI.   The
 independent variables N and DI were incorporated  in the  Froude number Vi,x
 in equation (6;.  Further considerations,  with the assumption  of a power
 relationship between variables,  led to equation (7).

     Upon  correcting KLa for temperature  effects  to 20°C,  the  data from
 models  M.I,  M.2  and  M.3 were correlated with  Kj^Q at  constant immersion
 depth simplex HI/DI.   From 189 observations for 3  models covering  a
 hydrodynamic (NFr) range of 0.084 to 0.342 the following correlation  was
 determined:
                                                    rl,52
                    KLA
                       20
=  1.527 x 10"8 N2'34
DIJ
     The calculated correlation coefficient of 0.9908 accounts for 987. of the
correlation by using the two independent variables N and DI.

Scale-Up Model Development

     Based on the postulate of equal mass transfer rates in model and
prototype as a prerequisite to the scaling of bio-hydraulic systems, the
scale-up equation (10) was developed.  Substitution of the exponents for N

                                     67

-------
and DI of  equation  (11)  into  the  scale-up equation  (10) establishes  the
relationship:                           , c-,  ,0  ,..
                                   m
                      N    =  N
                      m       p
                          '  Np
                              v      m
                                    m
                                        0.65
 If  the  geometric scale-up  ratio  DI/D^  is defined by Lr, then:
                     N    =   N   L °'65                               (12)
                      m        p   r
 Equation (12)  is  the  scale-up  function  for  the  systems examined.  Lr may
 also  describe  the ratio  of  any other  geometric  system variable provided that
 similar systems are studied.   Equation  (12) also allows one to predict from
 prototype data the hydrodynamic condition in a  scaled-down model which will
 give  equal Kj^Q  values. Equation  (12) was used to predict hydrodynamic
 conditions which  would give equal mass  transfer rates between laboratory
 and pilot- scale model.

 Scale-Up Model Testing

      In order  to  test the  scale-up model, a larger pilot-scale reactor (P*l)
 was designed.   The reactor  liquid volume of approximately 600 liters was
 selected as a  size which is relatively  easy and inexpensive to construct.
 This  reactor size was also  nearly one order of  magnitude larger than the
 largest laboratory- scale reactor used to develop the prediction model.
      The  mathematical model  for Kj^O prediction, equation  (11), was tested
 as  shown  by example  calculation using the  laboratory- scale  reactor  system
 M.I.   For M.I,  the  impeller  diameter (D!^)  is  5.40  cm.  An  impeller speed of
 250 rpm was arbitrarily  selected.   Substituting these values  into equation
 (11),  KLa20m is calculated to  equal 0.080 min"1.  The impeller  speed (Np)
 for the model P.I model  which  will  give  a  KLa2Qp value of 0.080 min"-"- must
 now be established.

      It is  known that the pilot-scale reactor  impeller diameter DL, =  26.67  cm.
 Substituting these data  into the equation  (12).


              250    =  N  (4.94)0'65
               N    =90 rpm
                P

     This means that      at impeller  speeds  of  90  rpm for the pilot- scale
reactor  (P.l) and  250 rpm in model M.I the  KX^Q values should equal 0.080
min~l.   In order to see  if  indeed this is true,  the calculated prototype
speed (Np = 90 rpm) is substituted into equation (13)  which describes the
experimentally derived Kj^Q values  for the pilot- scale reactor and then

                                     68

-------
calculating the prototype
                                 value.
     The linearized data  from the experimental KLa2Q determinations for the
pilot-scale reactor (P^l) may also be described by a power function as:
                     KLA,
                                2.57
                        20  '    "•-"   "Fr

The Froude number  (NFr)  is calculated as:
                                     (13)
             N.
              Fr
                       N2DI
                         g
,90 ,    26.67
 602   ( 98° ^
                                                       0.061
and substituting  into  equation (13) yields:
                                    .1.244
              KLA
                 20
                        2.57  (0.061)
             0.080 min"
     The actual  and  predicted ^320 values for P*l (equations 13 and 11,
respectively)  are  in excellent agreement,  thus confirming the validity of the
oxygen transfer  coefficient prediction equation (11) and the scale-up
equation (12).   Similar calculations are summarized in Table 2 for geometric
ratios of  4.94,  3.18 and 1.75 at three different impeller speeds.  Figure 2
illustrates more dramatically the excellent agreement between the in situ
and predicted  Kj^o  values for P.I based on equation (11).
     The  individual  in situ measured KJ^Q values for P.I and the predicted
values based on  equation (ll)  as developed from M.I, M. 2 and M.3 data, are
summarized  elsewhere (7).
                      0.400
                      o.ooo
                         0.000  0.100
                                     0.200   0.300  0.400
                                               .-1
                            Measured  KLA2o 'or P-1 (min' )
         Figure 2.  Comparison between in-situ  KLa£0  for  P.I  and

                     predicted KLa2Q fr°rn M.I, M.2 and M.3 data
                                     69

-------
 TABLE  2.   COMPARISON BETWEEN  PREDICTED AND MEASURED KLa
          VALUES FOR P,l CALCULATED FROM M.I, M.2 AND M.3  DATA

MODEL

M.I


M.2


M.3


PARAMETER
"„
rpm
250
350
450
200
250
300
140
170
200
DIm
cm
5.40


8.40


15.25


DIP
cm
26.67


26.67


26.67


Lr

4.94


3.18


1.75


L0.65

2.795


2.120


1.439


P.I
"PC
rpm
90
125
161
94
118
141
97
118
139
KLAZQ m
fflin-lP
0.080
0.177
0.322
0.090
0.157
0.244
0.099
0.157
0.235

KU20(12
. -1™
nnn
0.080
0.182
0.340
0.097
0.163
0.252
0.104
0.159
0.224
Scale-Up Model Refinement

     By  including the data from P.I in the mathematical model previously
obtained  from the M.I, M.2 and M.3  data, a further refinement of the scale-
up prediction model from regression analysis resulted  in:
                 KLA2() = 0.98 x 10
                                      "8 N2'41
DI
                                               1.55
(14)
     The multiple regression coefficient increased slightly from 0.9908 for
equation (11)  to 0.9949 for (14) -  explaining 99% of the  correlation by
using the  two  independent variables N and DI.

     Figure  3  shows the grouped data of 215 observations  for all four
reactor sizes  while comparing predicted to measured values.   The high
correlation  coefficient of 0.9949 is illustrated by the lack of data scatter,
even at the  extremes.
Figure 3.  Graphic
   illustration of
   correlation equation

   showing  predicted and
measured
   for all models
                   values
                                0.350 -
                             0.300
                          ^  0.250
                             <
                             _l
                             y.
                                0.150
                                0.100
                                0.050
                                                      Correlation Equation-
                                                *./*   KLA20 =0.98x10-8 N2408DI1551 -

                                                           o M-1
                                                           • M 2
                                                           * M-3
                                                           o P-1
                                                             R= 0.9949
                                                    11 = Number  of Observations
                             0.000
                               0.000  0.050
                                              0.100   0.150  0.200   0.250  0.300  0.350
                                              KLA20 Predicted (mirf1)
                                     70

-------
Error Calculation
      During the course of any experimental observation, a systematic error
 is made.   In order to assess the magnitude of the possible systematic error
 that  could be included in a calculation of KLa20, the partial differential
 of equation (14) was determined as follows:

                    6KIA20      „  ..   6N    ,  „    6DI                  ,,s)
                                                  ~Df
KLA
       =  2.41
                        20
N
     1.55
      The KLa2Q value for each model at low and high impeller speeds was
 calculated from equation (14).  For all experiments the impeller speed was
 determined within *1 rpm while the tolerance for the impeller geometry was
 -0.10 cm.  Calculations carried out for the maximum and minimum hydrodynamic
 condition of each of the four models summarized in Figure 4 illustrates
 that  the systematic error for KLa2o prediction is directly proportional to
 the Froude number as:
                   -   6 KLA
       =  0.034 N
                                      Fr
                                  (16)
     With model M.I the
min    at  low speed to
  ie error in KLa2Q due to N and DI ranges from *0.003
  -0.013 min"1 at high speed.  For model P.I, the err
at  low and high speeds decreased to 10.001 and +_ 0.005 min
                                  error
                         respectively.
                   0.020
                 .£
                 e
                 o 0,015
                   0.010
                   0.005
                   0.000
                                   8 KLA,o = 0.034 NFr
                      0.000     0.100     0.200    0.300
                               FROUDE NUMBER (NFr)

  Figure 4.  Possible systematic  error  in  Kj.a^n  due  to  N  and
      It would be of  considerable  importance  if  equation (12)  would also
hold  for geometrically  dissimilar surface  aeration systems.   As well,  the
boundary conditions  for applying  equation  (12)  would have to  be defined.

Scale-Up for a Geometrically  Dissimilar  System
     The objective of this  analysis  then, was  to  examine under what
conditions, if any, the  scale-up  transform
                              N
                                           0.65
                               m
                                                   (12)
is applicable to a geometrically  dissimilar  surface  aeration system.
     Because the scale-up equation  (12) was  developed  under conditions of a
constant aeration device immersion  simplex,  any  comparison with another
                                     71

-------
 system must  also  be made  at  the  same  simplex value.   It was decided to  see
 if  the equation would  hold for data  (12)  concerning oxygenation experiments
 in  tap water to determine the oxygen  transfer efficiency of a bladed  rotor
 (cage rotor).  These data were developed  from studies conducted in a
 rectangular  aeration tank 603 cm in  length, 152 cm in width at a liquid
 depth of 239 cm.   The  surface aeration device was a bladed rotor, 99  cm  in
 length and  700 mm diameter.

     Table 3 summarizes the  various  geometric simplexes identified in
 equation (4),  sizes them  to  those used in the development of the scale-up
 equation (12)  and compares them  with  the  geometrically dissimilar system
 under analysis.

TABLE 3-  SUMMARY COMPARISON OF AERATION SYSTEM GEOMETRIC SIMPLEXES

1- •!- C
Ol 4J (O
<£ 1—
c Type
O 01
•r- O
re >
01 O
Aerator
Immersion
Simplex
W/DI
W/D
WI/DI
LI/DI
S/DI
B
HI/DI
(7)
Square
4.0
2.0
Rushton Turbine
0.20
0.25
6
0.202
(12)
Rectangular
2.17
0.63
Cage Rotor
0.07
0.15
0.202
      Clearly, not only were the data from studies (12) conducted with a
 geometrically dissimilar aeration device to the one used for the development
 of  equation  (12), but aeration tank geometry was dissimilar as well.  The
 common  basis for data comparison was the aeration device immersion simplex
 (HI/Dl) which was set at a value of 0.202.

      The  reoxygenation data (12) developed under various hydrodynamic
 conditions were converted and expressed in terms of NFr, HI/DI and Kj^Q-
 Since the data were obtained at different rotor immersion simplexes, it was
 plotted as shown in Figure 5 in order to mark the constant aeration device
 immersion simplex of 0.202 which was used as a boundary condition for the
 development of the scale-up equation (12).  This then was the basis for
 comparison between the geometrically dissimilar system and the experiments
 carried out in the earlier study (7).

     As indicated on Figure 5 at a constant aeration device immersion simplex
 of  0.202, KLa£Q values were selected for various hydrodynamic conditions as
 defined by Npr.  These data are summarized in Table 4 and plotted in Figure
 6.  The regression equation was calculated as:

                        KLa20 = 2.31 NFr - 0.056
where :
          ^La20 = standard overall oxygen transfer coefficient (min  )
          NFr   = Froude number
     The Froude number can also be expressed in terms of an aerator peripheral
velocity and  the  liquid  volume  of  the  aeration  tank.   From a  practitioners'

                                      72

-------
         TABLE 4.   STANDARD OVERALL  OXYGEN TRANSFER COEFFICIENTS UNDER
                     VARYING  HYDRODYNAMIC CONDITIONS  AT HI/DI  = 0.202
                           rpm
                                           Fr
                                                       min-'
59
73
88
100
120
0.069
0.105
0.153
0.193
0.285
0.118
0.189
0.265
0.410
0.610
                              1.000
                              0.050 -
                                         0.200     0.400   0.700
                                       Immersion Simplex (HI/Dl)
Figure 5.
Standard overall  oxygen  transfer  coefficient at various  immersion
                                                              simplexes.
                               0600 -
                               0.500
                               0,300
                               0.200
                               0 100
                               0.000
                                             Kla20=2.3lNpr-0.056

                                               r= 0.9954
                                               HI/DI- 0.202
                                 0.000  0.100  0.200   0.300 0.400  0500
                                      FROUDE NUMBER (NFr)
Figure 6,
Standard overall oxygen transfer coefficient at  various
hydrodynamic conditions.
                                          73

-------
 point of  view the  concept  of peripheral  velocity rather than Froude number
 may be more meaningful.

       It was determined that  the  data  from  all models  including the study
 data  (12)  can be fitted with the transform of equation  (14)  as:
                                                                        (18)
 where:
           Vp = the aeration device peripheral velocity  (cm.s   )
           V  = aeration tank liquid volume  (1)
      These data can be depicted in Figure 7 so that if it is desirable  to
 operate any sized liquid system at a specific overall oxygen transfer
 value, the surface aeration device peripheral velocity required to attain
 this KS   value is indicated.
 600
— 400

i»
           £
           E
           S 20
                                                                  KLA20.--.Oi02..
                                                                 HI/DI.0.202
              10°
                          10'
                                     10'
                                                 10s
                                                             104
                                      TANK LIQUID VOLUME (LITRES)
 Figure 7.  Surface aeration device peripheral velocity and  standard  overall
            oxygen transfer  coefficient  for  various liquid volumes  at constant
            immersion  simplex.
      It  is interesting to  note  that  a straight line  correlation between the
 geometrically dissimilar system data (12)  and the data used  in the
 development of the  scale-up  relation, equation (12)  exists.   This is highly
 significant in that  it gives a  preliminary indication that  geometric
 similarity for the aeration  tanks  and surface aeration devices may not be
 required,  provided that  a  high  rate  of air-liquid interface  renewal occurs
 and the aeration device  immersion  simplex  is constant between all systems.

     Essentially, this specific analysis has shown that,  for a geometrically
dissimilar  aeration device and  aeration  tank under conditions of intensive
air-liquid  interface renewal and equal aeration  device immersion simplex
 (HI/DI), the overall oxygen  transfer  coefficient (KLa20)  could be predicted
from equation  (14) as  developed from  the experimental work  (7).  The
inference that can be  drawn  from this is:  while system geometry may not be
critical, the degree of turbulence generation and aeration device immersion
simplex is critical.   Cage rotors have these characteristics.
                                     74

-------
Recent Surface-Aeration  Scale-Up  Study

     Extensive  investigations and review  concerning  surface-aerator  scale-up
were recently reported on  (10).   With a scale  factor of 10 between two
aerators it was observed that the deviation  in oxygen transfer results
between small and  large  scale systems was within ^20%.  This was considered
to be acceptable.  Encouraged by  these results, it would appear then that a
similar approach can be  taken in  scaling  other aeration systems such as
diffused, turbine  (sparger), static and rotors  when equal environmental
conditions between model and prototype exist.


KLa AND SYSTEM  VARIABLES FOR OTHER AERATION  SYSTEMS

JC^a, and Diffused Aeration  System  Variables

     For a diffused aeration system the oxygen transfer capability may be
expressed as a  function  of gas  flow (G),  liquid depth above the diffusers
(H) and the width  of the aeration tank (W) as:

                \a =  fn (G, H, W)                                   (19)

K-,a and Turbine (Sparger)  Aeration System Variables

     One would  suspect that for a turbine aeration system the overall
oxygen transfer coefficient is  a  function of the combined variables for
diffused aeration  and  surface aeration.   This  indeed is true and can be
expressed as:
                KLa =  fn (G, H, N, Dl)                                (20)

Kja and Static  Aeration  System  Variables

     A static aeration system consists of submerged vertical cylindrical
tubes which may contain  a  helix to lengthen  the bubble path and hence
contact time.   Compressed  air is  fed through a sparger at the cylinder
bottom.  Variables affecting K^a  are gas  flow  (G), sparger design (K),
cylinder length (L) and  diameter  (D) and  liquid depth (H).  This may be
expressed in a  functional  relationship as:

                KLa =  fn (G, L, D, H)                                 (21)

KT a and Cage Rotor Aeration System Variables

     A rotor aerator could be likened to  a turbine aerator on its side
and thus have variables  influencing the overall oxygen transfer coefficient
KLa which are identical:
                KLa =  fn (N, DI,  Hi)                                 (22)

     This appears  to be  substantiated by  the analysis of data (13) between
two geometrically  dissimilar aeration systems  summarized earlier.
                                    75

-------
     Studies have shown (14) that 06 is also a machine variable, being higher
for surface than for diffused aeration systems.  It would be appropriate to
highlight some of the studies conducted by others concerning scale-up.

Diffused Aeration System Scale-Up

     Using dimensional analysis as a starting point a general relationship
for the prediction of the overall oxygen transfer coefficient, KLa was
developed (5):
                                                             -0.446
       .a  -  9.35
•pQa(Ndo)2-
PL
0.046
"Qa (Nd0)'
g
-U.lZb
V
Td
u.o?
~T "
w
_v
                                                                      (23)
where:
Qa
 P
 P
 N
do
 S
           w
              =  airflow/unit volume
              =  density of liquid
              =  viscosity of liquid
              =  number of air bubbles
              =  diffuser orifice diameter
              =  gravitational constant
              =  diffuser depth
              =  aeration tank depth
              =  aeration tank width
    -1
  T

F1T6T3

  L
LT
-2
I
L
L
     The  coefficients for the correlation were developed from reaeration
 experiments  conducted in 3 geometrically similar but different sized
 aeration  tanks.  The equation describes his data with a mean error of less
 than 10 percent.

     The  same  researcher (5) also developed a general relationship relating
 key parameters directly to K^a.
                              r\ f\ r\   i  rt /*    ^T'"*/
                                                                       (24)
He  found  that this equation predicted K^a for his data with a. mean error of
less than 87o.

Turbine (Sparger) System Scale-Up

     Another researcher (9) investigated parameters influencing K^a  for a
turbine (sparger) system.  While his correlations described his data,
successful scale-up remained elusive.

Static Aeration System Scale-Up

     The  problem of bench and full scale tests for alpha and beta coefficient
variability determination was recently addressed (14).  As  shown in  Figure 8
which compares bench and full scale K^a values with corresponding airflow
per unit  volume, approximately 2.5 to 3 times more air is required for the
                                      76

-------
bench scale unit than  for the prototype unit in order to produce equal KLa
values.  The primary reason  for this  is the difference in liquid depth
between the two systems,  i.e. 90  cm  vs 305 cm.  This would suggest that
distortion of the model geometry to allow for equal bubble hold-up time may
be in order.  A scale-up function  was not developed in this work (14).
 figure 8.   Comparison
   between  full-scale and
   bench  scale  KLS  at unit
   air flows.
Xtt-
3
O
0
Q »o-
5
|"
100-
>ll SCIM SCfM, / SCfM./SCfU.
-"• — -• — — ' t _~-i. !'..». ^«J
01 ISO 00 .' II
0 II l«» ID ' )l
01 110 HI rf 14
/
/
/ ^- ••"" *"*• "" -A'"
/ ,~'
/ *••
/ ^.x-"'
/ X '
^'''
	 1 	 1 	 1 	 1 	 1 	 1 — -
                                                                       01
                                                        mm
 Cage Rotor Aeration System Scale-Up
      An interesting study (6) reports on the successful  development of a
 predictive equation for oxygen transfer using "cage  rotors.   These  researchers
 started with bench top size models and validated their predictive  equation
 using data from existing full-scale installations at Aberdeen.Scotland and
 Ames, Iowa.  The predictive equation derived with the  assistance of
 dimensional analysis was:
7.42x10
    N

 where:
                        0.86
r  -nO.18 p -I -0.28 p  -,  0.70 r   -•
H     M       H       [%rl
                                                                  -0.19
                                                                       (25)
          K
           •La  =
            N  =
           HI  =
           DI  =
            H  =
          N-
           Fr  =
        overall oxygen transfer coefficient
        rotor speed
        depth of rotor immersion
        rotor diameter
        rotor blade width
        liquid depth
        dimensionless Reynolds  Number
        dimensionless Froude  Number
                            L
                            L
                            L
                            L
                                      77

-------
     Equation (25) was developed and coefficients evaluated using a single
I/12th scale model operated under a great variety of operating conditions.
Correlation coefficients for this data varied from 0.983 to 0.996 for each
of the components.  The equation shown above can be put into a more general
form and used for scale-up purposes to estimate oxygenation capacities for
rotors in general even though system geometry differs.  For a geometrically
similar class of rotors the relative values of K,a are:
                                                 -0.19
                           N,
                   K
                               N
                                Rei
                               N-
                                Re,
                                     0.70
                                                                       (26)
      Analysis of equation  (26)  shows  that  K^a  varies  directly  as  the  4/3
 power of the rotor speed and  as the % power of the  rotor  diameter.  The
 authors (6)  used data from two  full scale  installations to  validate their
 predictive equation.   While the measured O.C.  was 2.15 Ibs  oxygen per day
 their equation (25) predicted a value between  2.26  and 2.37 Ibs oxygen per
 day.  This variation  was attributed to the fact that  some degree  of
 uncertainity existed  as to the  actual depth of rotor  immersion.   Nevertheless
 their prediction is uncannily close (within less than 117o),  especially when
 considering that their equation was developed  using pure  water and compared
 to data generated from an  oxidation ditch  treating  swine  wastes.  No
 corrections  for  differences in  substrate were made.   This leads one to
 suspect that the choice of system used  for model validation might have  been
 most fortuitous.   Nevertheless,  the same authors (6)  went on to validate
 their model  using rotor performance information generated (15) for a  full-
 scale oxidation  ditch system.   Knight  determined that after one hour
 aeration of  the  initially  deoxygenated  ditch water, the dissolved oxygen
 concentration had increased to  6.33 mgL  .  Application of equation (25)
 estimated a  dissolved oxygen  concentration of  7.08  mgL" .  This again
 represents agreement  to within  less than 127o.

 Rotating Biological Contactor Aeration  Scale-Up

      A very  preliminary stab  at  examining oxygen transfer for  RBC can be
 made using limited data (16).  Again one can consider that K^a is some
 function of  the  contactors' rotational  speed, N, and  its  diameter, DI.
 As well,  disc immersion depth,  HI, is  important.  A function describing
 oxygen transfer  in an RBC  system could  then be written as:
                 KLa  =  fn (N, DI, HI)
                                                                      (27)
It can be postulated that a scale-up function at equal immersion  simplex
assumes the form:
                              DI
                  N
                   m
                                m
and that:
                 KLa  =  0NFrn
                                                                      (28)


                                                                      (29)
                                     78

-------
     Table  5  summarizes the oxygen  transfer data calculated to produce
figure y.   The data are most interesting in that they illustrate that for
this particular immersion depth  simplex (constant between models) a limiting
overall oxygen transfer coefficient is  attained.  It would appear from this
albeit sparse data that the maximum value of the overall oxygen transfer
coefficient is around 0.050 min"1-
 TABLE 5.   RBC  OXYGEN TRANSFER DATA AT  DIFFERENT HYDRODYNAMIC CONDITIONS
                           DI

                           cm

                          15.25
                          30.5
                          61
N
rpn
40
20
10
20
10
5
10
5
2.5
VP,
on s"1
33
16.5
8.25
33
16.5
8.25
33
16.5
8.25
V
-
0.0074
0.0018
0.0005
0.0037
0.0009
0.0002
0.0018
0.0005
0.0001
V
ram-1
0.049
0.036
0.021
0.042
0.031
0.011
0.028
0.015
0.008
   0.050




   0.040




e.  0.030
'i

(0
-I
X  0.020



   0.010
   o.ooa__»__L
                           1 ,  i	L
                             J	L.
                                                      DI
                                                      on

                                                     15.25 •
                                                     30.5  A
                                                     61   •
          0.0010   °-°020   0.0030
                                            0.0050   °-0060  0.0070
 Figure  9.   K,a as a function of Npr  for  different sized RBC systems


      Information and correlations of this type would  greatly assist in the
generation  of process design and scaling  information.  Admittedly only the
surface  has been scratched.
                                      79

-------
SUMMARY

     The information presented clearly indicates that progress in the area
of process scaling has been made.  Specifically, our ability to predict full
scale oxygen transfer capabilities of various aeration devices and systems
from small scale systems has increased.  But it is equally apparent that the
amount of data scatter in investigations concerning system scale-up, lack of
suitable field data for model validation is still formidable.  Differences
between various formulations are equally extreme.

     It is evident that we still have a long way to go before oxygen transfer
data generated at bench-scale can be used with great confidence for prototype
aeration system scale-up.  Equally, the transfer functions for scale-down
remain to be developed and validated.

     There is little doubt that in order to minimize scale-effects between
laboratory and prototype systems every attempt at attaining the goal of
environmental, geometric, kinematic and dynamic similitude must be made.
Successful scaling correlates directly with the degree to which we attain
this goal.


REFERENCES

 1.  Eckenfelder, W.W. "Factors Affecting the Aeration Efficiency of Sewage
     and Industrial Wastes", J. W.P.C.F., 31, (l), 60-70  (1959)
 2.  Tsao, G.T. and Kempe, L.L.  "Oxygen Transfer in Fermentation Systems.
     1.  Use of Gluconic Acid Fermentation for Determination of Instantaneous
     Oxygen Transfer Rates", J. Bio-chemical and Microbiological Technology
     and Engineering, II, (2), 129-142  (1960)
 3.  Kaplovsky, A.J.,  "Artificial Aeration of Canals in Chicago", J.
     W.P.C.F., 36, (4), 463-474  (1964)
 4.  Horvath, I.,  "Modelling of Oxygen Transfer Processes in Aeration
     Tanks", 3rd Int. Conf. Water Pollution Research, 2, 195-227, Munich,
     West Germany  (1966)
 5.  Smith, D.W.,  "Modelling Oxygen Transfer in Diffused Aeration Tanks",
     Ph.D. Thesis, University of Kansas, Lawrence, Kansas  (1970)
 6.  Nelson, G.L., et al,  "Basic Performance Parameters for Oxygenation in
     Rotor-Aerated Liquid Waste Systems", Transactions of the ASAE, 1972.
 7.  Schmidtke, N.W.,  "Scale-Up Methodology of Surface Aerated Model
     Reactors"s Ph.D. Thesis, University of Waterloo, Waterloo, Ontario
     (1974).
 8.  Schmidtke, N.W. and I. Horvath,  "Scale-Up Methodology for Surface
     Aerated Reactors", Prog. Wat. Tech. 9, 477-493  (1976)
 9.  Hunter III,  J.S.,  "A Basis for Aeration Design", Ph.D. Thesis,
     Colorado State University, Fort Collins, Colorado  (1977)
10.  Zlokarnik, M.,  "Scale-Up of Surface Aerators for Waste Water Treatment",
     Advances in  Biochemical Engineering II, T.K. Ghose, A. Fiechter and N.
     Blakebrough  editors.  Springer-Verlag, Berlin, Heidelberg, New York
     (1979)
                                    80

-------
11.  Harremoes, P.,  "Dimensionless Analysis of Circulation, Mixing and
     Oxygenation in Aeration Tanks,"  Prog. Wat. Tech. Vol. II, No. 3 (1979)
12.  Cleasby, J.L. and E. Baumann,  "Oxygenation Efficiency of a Bladed
     Rotor",  J.W.P.C.F. 40, (3).  Part 1, 412-424  (1968)
13.  Schmidtke, N.W.,  "Scale-Up Methodology for Geometrically Dissimilar
     Surface Aeration Systems",  unpublished document  (1982)
14.  Otoski, R.M. et al,  "Bench and Full Scale Tests for Alpha and Beta
     Coefficient Variability Determination", Proc. 33rd Purdue Industrial
     Waste Conference (1978).
15.  Knight, E.S.,  "Performance of a Cage Rotor in an Oxidation Ditch",
     M.S. Thesis, Iowa State University, Ames, Iowa  (1965)
16.  Personal communication with Professor Moloff (1976).
DISCLAIMER

      The  work described  in this  paper was not funded by the U.S. Environ-
mental  Protection Agency.   The contents do not necessarily reflect the
views of  the Agency,  and no official endorsement should be inferred.
                                     81

-------
AERATION  SYSTEM DESIGN PROTOCOLS:  A NORTH AMERICAN PERSPECTIVE

Daniel W.  Smith, Professor
Department of  Civil Engineering
University of  Alberta
Edmonton,  Alberta, Canada

ABSTRACT

      Aeration  system sizing  for suspended growth systems requires careful
evaluation of  organic and hydraulic loading parameters.  Experience based
and  rationally developed model alternatives for estimating oxygen demand
are  presented.  Physical requirements and conditions, including mixing,
tank geometry  and aeration system type and layout are covered.  The sensitivity
of the design  to various parameters is also investigated.  The results of a
small survey on aeration system design approaches is included.
 INTRODUCTION

      All  biological  treatment processes make use of oxygen during the
 metabolic steps  involved  in the conversion of the organic matter to cellular
 mass  components  and  energy.  The design of biological treatment systems can
 be  accomplished  making use of varying amounts of knowledge of the system
 biochemistry.  The objective of waste treatment research is to achieve a
 better  understanding of the biochemical processes involved and, using this
 information, develop rational design relationships for sizing of process
 reactors  and components.  A variety of such approaches have been followed
 in  the  development of aerobic suspended growth design models.  The application
 of  such models requires the user to determine a number of parameters for
 the design of wastewater  treatment systems.

      This discussion will focus on the approaches used in North America for
 ultimately identifying the oxygen requirement in aerobic suspended growth
 systems.   Topics to  be covered include estimating the organic loads and its
 variations;  estimating oxygen requirements and its variations; the effects
 of  tank geometry, aeration system layout, and mixing; the sensitivity of
 the design to various parameters; and the results of a survey on current
 practice.
SYSTEM PERFORMANCE MODELLING

     The development of models for suspended growth system performance
prediction have taken two basic forms:
          - experience based rules for design and operation, and
          - rationally developed models for design and operation.
For each of these approaches, the following basic informational components
are required:
          - organic loading and its variation with time,
          - hydraulic loading and its variation with time,
          - nutrient complement and its variation with time, and
          - definition of the physical environment.
                                      82

-------
Each of these components will vary with  the hour of the day, the day of the
week, the week or month of  the year, and from year to year. For municipal
systems the pattern of variation  is dependent on the social and industrial
infrastructure of the community.  For design purposes the values and
variations at the time of design  and, more importantly, the future pro-
jections of the values to the design life of the facility are required.

Organic Loading

     The organic loading to a treatment  plant can be evaluated using one of
the following parameters:
          - biochemical oxygen demand (BOD)
          - chemical oxygen demand  (COD)
          - total organic carbon  (TOC),  and
          - total oxygen demand (TOD)
Each of these measurements  can be related to the substrate requirements of
the microorganisms in the aeration  system.  However, the BODs value of the
wastewater most closely relates to  the actual oxygen requirements for the
metabolic processes involved during the  conversion of the useable organics
to cell components.  Therefore, the design parameter used for expressing
the organic load to a wastewater  suspended growth treatment process is the
BOD.  In some cases, only the soluble fraction is considered, however, the
realistic approach is to include  the total primary effluent BODs as the
design value.  One method of evaluating  existing operating conditions is to
make use of frequency distributions to describe organics and suspended
solids concentration to the treatment unit.  Data points used for such
distributions should assist in identifying loading rates for design.  The
optimum approach is to use  BOD data average the maximum significant design
retention time of the aeration tank.  Figure 1 presents a typical frequency
distribution for BODs, COD  and total suspended solids for raw wastewater.
The daily variation is represented  in Figure 2.

     The reduction in the raw wastewater by the primary sedimentation
process normally varies from 20 to  35 percent for BODs and 50 to 65 percent
for suspended solids.

     If nitrification is allowed  to occur in the aeration tank then this
oxygen requirement must be  estimated and allowed for in the design of the
aeration equipment.

Hydraulic Loading

     The hydraulic flow rate through an  activated sludge unit is equal in
importance to the organic loading.  Three flow values are important to
design:  average, design, and peak.  Here the design flow rate will be
assumed.  This value is that of some maximum significant period such as 4,
6 or 8 hours, depending on  the actual conditions.  Figure 3 illustrates the
type of flow pattern common to municipal systems.  Again this must be
estimated for the design life of  the system and/or design stages.
                                     83

-------
    999 99 S 98
              8° 60  <0 20 10 3 2  05
          5   20 40  60  80 90 95 98 995 999
             FREQUENCY OF OCCURRENCE
                                                              TIMf.h
Figure 1:  Typical frequency distribution     Figure 2:  Typical daily variation
            for domestic wastewater                       in wastewater flow and
            (after ref.  2)                                 BODs  (after ref. 2)
              0.08 -
              0.06 -
             rO.04
              0.02 r
              0.00
Typical daily
flow during
wet periods
                        Typical daily
                        flow during
                        dry periods
    Figure 3:  Typical municipal wastewater flow pattern  (after ref.8)

Nutrient Complement

     A large number  of nutrients are  required for the  optimum growth of  a
biological population.   Municipal wastewater is usually adequately supplied
with nutrients.  However,  industrial  and mixed industrial and domestic
wastewaters may be deficient in some  nutrient.  This should be investigated
prior  to design.
                                       84

-------
Physical Environment

     Although many  factors impact on the physical environment of a biological
system, temperature is  among the most important.   Temperature estimation
must be incorporated  into  the information estimating process since it
effects the  rate  of biological activity and the rate of oxygen transfer.
Other physical  components  such as tank geometry and mixing will be discussed
later.
EXPERIENCED  BASED ESTIMATES OF OXYGEN DEMAND

     A variety of experienced based design relationships  have been used in
North America.   These have varied from 1.0 kg 02  per  kg carbonaceous  BODs
removed  to relationships relating air to  be supplied  to BODs  applied  to
more elaborate functions of BOD5  and solids retention time.   Air  application
rates for the  more frequently used activated sludge processes are shown in
Table 1.

     A more  comprehensive approach was proposed in 1954 by Eckenfelder  and
O'Connor which related the oxygen requirement to  the  BODs removed and the
mass of  mixed  liquor suspended solids (MLSS) in the aeration  tank (3).   In
SI units the equation takes the form
           kg Oa  supplied
                day
           A'
kg BODs removed
       d
+  B' (kg MLSS in aeration
                 tank)
     where
          A'  = Oa  required for BODs  synthesis,  mg 02/mg  BODs  removed
          and
          B'  = 02  required for endogenous  respiration, mg  Oa/(mg  MLSS'd)

Several  other researchers also developed values for  A' and B'  following
this relationship.   Converted values are shown  in Figure 4.
                                 0.2  0.3 0.4 0.5 06 0.6 1.0  1.5 iO
                                  f M RATIO." «BOD^< «MLSE
        Figure 4:
Relationship between F:M ratio, oxygen requirements
and different predictive relationships (after ref. 2)
                                      85

-------
     Stensel and Shell related oxygen requirements to BODs removed  and
solids retention time (SRT)  (4).  Their relationship was of the  form:



          *.  -  [jrf\Mj   [A + ** <«« * * H
     where
          r0 = oxygen uptake rate, mg/(L*d)
          a  = cell yield coefficient, mg cell/mg BODs
               (range of value = 0.5 to 0.7)
          b  = endogenous decay coefficient, 1/d
               (range of value = 0.04 to 0.1)

The effect of temperature on the oxygen requirement is shown  in  Figure  5
for a domestic wastewater.
                                     Table 1
                    Air Application Rates for Various Activated Sludge Options
Activated Sludge
Alternative
Convention
Extended Aeration
Step Aeration
Contact Stabilization
Hinh Ratp
Gas Applied
1.0
3.7 to 15
30 to 55
94
up to 125
90
90
71 tn on
Units
Kg o2
Kg BOD5
m air
m wastewater
m3 air
Kg BOD5 applied
m3 air
Kg BODg applied
m3 air
Kg BOD5 removed
m3 air
Kg BOD5 applied
m3 air
Kg BOD5 applied
m3 air
Reference
2
2
2
1
2
2
•>
                                      86

-------
               D  , a
               O  '•»
               CO
               f
               "-„  1-6
                  °-2
               X
               O
1.4

1.2

1.0

0.8

0.6

0.4
            T~T   i—i—i—ri i 1111
                                          i—r
       -1 - 1 — LJ
                  1
                                             DOMESTIC WASTEWATER
                                              COD/BOD  =1.6- 2.0
                                              TSS/BODC =0.8- 1.2
                       i   i i i i i ii   ii   ii
                        0.4  0.60.81.0 1.5 2   346 678910 15 20  30 40 50

                                SOLIDS RETENTION, SRT, days
           Figure 5:   Relationship between solids retention time,
                      temperature and oxygen required  per unit
                      BODs removed (after ref. 2)
RATIONAL ESTIMATES OP OXYGEN DEMAND

     A number  of performance and design models have been presented during
the last 20  years (5,6,7).  The fundamental approach followed  by most of
the models is  the same.   This discussion will be limited to  the model
components related to oxygen uptake rate as developed by McKinney (5),
Lawrence and McCarty (6) and Eckenfelder (7).  The fundamental difference
between these  models lies in the relationship between substrate concentration
and the growth rate of microbial mass.  Discussions of  these differences
are presented  elsewhere (8,9).

     The objective of these rational models is to provide  a  reference base
for sizing aeration tanks and equipment, and predicting sludge production
and effluent substrate concentration.  Table 2 presents the  equation for
oxygen uptake  rate proposed by the three models included here.

     The questionnaire responses indicated a variable amount of interest in
these models as shown in Table 3.
                                      87

-------
                                                               Table 2.   Oxygen Demand Models
oo
oo
Model
A02 = Q [1 - 1.42 YT(S0-Se)] + 1.42kdXVa
1.5 (SQ-Se) - 1.42 (Ma+Ne)
rr - t
1.5 (S0-Se) 1.42 (Ma+Me)
t 0C
rr = 1.42 boxd + a'Sr

&09
Q *
S°
£
va
\
\
»0
Terms
= oxygen demand, g/d
= flow rate, m3/d
= solids yield coef. mg/mg
=• infl. substrate, mg/L
= effl. substrate, mg/L
= end, resp. factor, 1/h
» MLVSS, mg/L
= aeration tank volume, m
= oxygen uptake rate, mg/(L*h)
= active mass in aeration
= endogenous mass in aeration
= hydraulic detention time
= solids retention time
= fraction of degradable SS
Reference
Lawrence (6)
McKinney (5)
single pass system
recycle system
Eckenfelder and Ford (7)
                                                                     Xd  = degradable MLVSS
                                                                     a   = fraction of substrate used
                                                                           for oxidation
                                                                     S_  = substrate removed

-------
                             Tables
         Summary of  Aeration System Design  Survey
                   (15 completed  responses)
Aeration Systems  Used
     Diffused Air
          fine bubble                             15
          coarse  bubble                           66
     Mechanical
          surface turbine                         56
          brush                                  1 5
          submerged  turbine                       10  (+3 combined)
     Other
          lagoon  tubing                            2
          static                                  11
          jet                                     3
          0;                                       3
Design Data Generation  (number using indicated  approach)
Parameter
Hydraulic
BOD
Suspended solids
Surface active agents
Load variation
11 terature
5
7
7
4
9
Field
Measurement
12
11
IE
7
13
Calculator
9
9
9
0
5
                    Oxygen Uptake Rate Estimation
              Method                             Number
Air Supply/unit BOD
Rational Models
Lawrence and McCarty
McKinney
Eckenfelder
Own
Codes
6

4
4
4
1
4
Aeration Basin Geometry
Ratio     Length:width:Depth

     Plug flow
          6 to 20 .  2  to 3   1,  sidewater depth  3-5.5 m

     Aerator pattern
          Roll     5
          Grid     3

     Complete mix
          L:N = 1  to 3 : 1
               side  water depth 3 - 4.6 m
               aerator  mechanical
     Circular tanks  only noted for package plants
Mixing
     Minimum liquid  velocity
          0.15 m/s                               5
          basin turnovers    12-15 times/h       1
          dye testing                            1
          power/unit volume                      1
          air rate/unit  volume                   2
          scour calculation                      2
          velocity meter                        2
Process Control
     Use of D.O.  probes
          Number
               2/tank                            3
               I/tank                            5
          Location
               point of minimum D.Q.             1
               1/3  point from inlet              2
               midway  from aerator to wall        1
               adjustable                        1
          Other Methods
               portable D.O. probes              3
               field 0; uptake                   1
               Winkler test                      1
          Feedback  Loops
               record  D.O. only                  2
               record  0.0. manual control         3
               record  D.O. feedback loop         8
                                   89

-------
AERATION EQUIPMENT PERFORMANCE

     Establishment of the oxygen uptake rate is one of the first steps in
the design process.  Another early step is establishment of the oxygen
delivery capability of alternative types of aeration equipment.  The amount
of oxygen transferred by an aeration device must be expressed in a form
which allows comparison with other units.  Oxygen transfer tests are normally
conducted with  clean water and the results adjusted to standards conditions
of 20°C and one atmosphere pressure (2, 11).  This type of testing, when
conducted in tanks of the same shape and volume, yields information which
can be used to  compare aerator oxygen transfer efficiency.  Performance
values are expressed in units of KgOa transferred per kilowatt-hour (wire
to water basis).

     Using the  clean water non-steady state testing procedure (11) and
solving the general equation


                ff =  kLa (Cs - CL)

     where
          dc      change in Oa concentration
          dt  ~           time
          Kj^a =  overall oxygen transfer coefficient, per hour

          CL  =  saturation cone, of 0? in water under test conditions, mg/L
           o
          CL  =  concentration of Oa in water at corresponding time, mg/L

for K^a, yields        1  Cs - C±

                i           Cq - Co
                kLaT  '      S    2
which  can be used to determine k^a.  Temperature adjustment is made using


               ^LaT  =  ^La20 C  ®
     where
          1
          0  =  temperature correction coefficient, 1.024
T  =  test temperature, °C
     When the information is for comparison of two units tested in the  same
tank then k^a or a calculated oxygen transfer rate value in kg per hour can
be used.  When the information from tests in different size and shape tanks
must be used then comparison becomes much more difficult.  It has been
suggested (12) that the term (K^a • tank volume) can be used.  This leaves
only the interface area and the mixing factors to be evaluated.
                                     90

-------
     When the results  are  to  be  translated  into  the wastewater system it is
necessary to use a and 3 to adjust  for  the  effects of organics on the
wastewater on kLa and  of salts,  organics  and  solids on  Cs, respectively.
The general equation then  becomes


          -£  =  a kLa (3  Cs  - CL)

This equation can be changed  in  a variety of  ways to account for temperature
and units adjustments  (2).

     For development of a  design it is  necessary to predict the value of a,
the correction factor  for  obtaining a wastewater kLa value, and g, the
correction factor for  obtaining  a wastewater  Cs  value.  Determination and
evaluation of these factors has  been discussed in detail elsewhere (13) .
It is known that the values vary from one type of wastewater to the next,
from one degree of treatment  to  another and from one type of aeration
equipment to another.   For municipal wastewater  values  for a and 3 can be
selected from the literature  with careful consideration of the conditions
during which they were measured.  Only  a  limited amount of information has
been published regarding industrial waste values of these factors.  Normally,
a conservative  (or low) value should be selected.  Due  to the problems with
performance prediction caused by different  aeration basin geometry and with
a and 3> it is recommended that  the aeration  equipment  specifications
provide for either pre-start-up  testing or  testing within the first year of
start-up.  Furthermore, penalties for failure to meet oxygen transfer
requirements should be specified.   Specifications and testing procedures
are discussed elsewhere (14,  15, 16, 17).
PHYSICAL  COMPONENTS

     The  physical  components  of  an  aeration system have a major impact on
the overall  performance  of  the aeration device used as well as the performance
of the biological  process.  The  tank  geometry and aeration system layout
within the tank  control  the flow regime and mixing within the reactor.

Tank Geometry

     The  extremes  of  tank geometry  are from long, narrow, nearly tubular tanks
for a plug flow  regime to circular, complete mixing flow regime tanks.
These extremes allow  easy modelling of mixing and prediction of aeration
system design requirements.   With modifications to tank configuration, these
two extremes are used almost  exclusively  in aeration tank design.

     The  mixing  patterns in complete  mixing reactors is controlled by the
aeration  pattern.  Perfect  mixing is  difficult and operationally expensive
to achieve.  Surface  or  submerged turbine units are frequently used.  With
mechanical aerators point feed and  discharge can be used.  With diffused
air systems  a diffused feed and  discharge design is used.  Tank configurations
vary from circular to square  to  rectangular about 3:1 length to width ratios.
                                     91

-------
     Plug flow reactors range from 5 to 10 m in width, 5 m in depth and up
to 125 m in length.  Diffused air systems are frequently used along with
point input and outlet.  The survey results generally conform with these
values.

Aeration System Layout

     The arrangement of the aeration equipment has a major effect on achieving
the desired mixing pattern.  Figure 6 shows a few alternative arrangements.
The wide variety of equipment, tank configurations and arrangements makes
comparison of systems difficult.

Mixing

     Mixing has two basic purposes, maintaining the solids in suspension
and assuring that the oxygen is well distributed throughout the tank.  For
some types of aeration devices it is possible to provide adequate oxygen to
all parts of the tank without providing adequate suspension of solids and
without complete mixing.  Complete mixing assumes that the influent is
spread throughout the aeration tank in a very short time relative to the
retention time.  Some researchers have proposed using turnover time (18) or
pumping capacity (19) as a measure of the mixing achieved in a reactor.
Both values fall short of good description of the operation without specifica-
tion of baffling and other physical features of the system.  Power per unit
volume is also a poor indicator of mixing (2).

     Mixing is best indicated by tracer studies.  Another good indicator in
complete mix reactors is the oxygen uptake rate.  In an ideally mixed tank
the value would be the same at all points.  Suspended solids measurements
does not provide a good indication of mixing except in extremely poorly
mixed conditions.
STRATEGIES FOR OPERATION AND CONTROL

     The design of the aeration system for an activated sludge plant will be
significantly influenced by the method of operation selected.  The most
consistant performance of a biological system can be achieved with uniform
organic, hydraulic and recycle sludge loading.  This can be approached with
flow equalization and the use of parallel reactors with hydraulic loading
control.

     The aeration system lends itself to continuous or online control by use
of a dissolved oxygen monitor linked by a data processor to the drive motors
or control valves.  To achieve efficient operation variable speed motors or
adjustable blade turbines should be used.

     The information survey indicate the major of those responding felt
feedback loops was desireable.  Some respondants felt operator checking and
adjusting was more practical.
                                      92

-------
W:D = 1 :1
one side aeration
•:
r
Brush Aerator

 W:D  -  2:1
 one  side aeration

                       Surface  Aerator
                                                                       //,
 W:D - 2:1
 two side aeration
                                               One  Turbine
J 	 1

• *
vV-
^•c-^-3

'• t
fc /
i " ,
£
>
      Fine Bubble
                                              Two Turbines
                                                              /•
                                                                '
      Jet Aerator

     -. •
                  Figure 6:  Aeration tank and equipment layouts
                                      93

-------
 SENSITIVITY  OF DESIGN PARAMETERS

     The major components of an activated sludge design which have been
 identified above are the organic and hydraulic loading on the aeration
 chamber and  the operating temperature.  The design of aeration systems
 also relies  on correct assessment of the oxygen saturation value, oxygen
 concentration to be maintained in the reactor, and the effects of organics,
 salts  and suspended solids on the oxygen transfer rate (a and 3)•

     Analysis of the aeration process design can follow several approaches.
 A very basic comparative approach using the reference conditions shown
 in  Table 4a, the adjusted conditions shown in Table 4b, and the equations
 shown  in Table 4c, allows the generation of values which can be used to
 evaluate the relative importance of the variables to the design process.
 Figure 7 briefly summarizes the relative energy requirements estimated in
 this way.

     A 20 percent error in selection of the design temperature is not
 nearly as serious as a similar magnitude error in estimating organic loading
 or  the hydraulic loading to the aeration tank.  A 20 percent error in
 estimating the SRT also has considerably less impact.

     The most significant impact on the estimated energy requirement was
 caused by the 3 value, assuming a 20 percent lower value.  The importance
 of  a was similar to that of the organic and hydraulic loadings.

     The serious question which arises from this type of analysis is,
 what is the  potential for error in predicting the future characteristics of
 wastewater.  The characteristics of domestic wastewater historically have
 been the easiest to predict.  Water consumption and dietary habits have
 slowly changed in a predictable fashion.  However, the increasing cost of
 energy and the related increased cost of all utilities will slowly bring
 about  significant changes in attitudes about conservation.  As this occurs,
wastewater characteristics will change.  Wastewater strength can be expected
 to  increase.  This may carry with it a lowering of a and 3 values.  Such an
occurrence may mean a number of closely designed plants will find aeration
system deficiencies as they approach design life and/or design hydraulic
loading.

     Industrial waste characteristics are predictable as long as process
operations remain relatively static.  Industry is much more sensitive to
economic benefits of process adjustment through conservation if process
performance can be assured.  The trend toward increased conservation in
industry is well documented. This trend will require careful evaluation on
a site by site basis for selection of proper design parameters.
CONCLUSIONS

     The design of aeration systems in North America can follow a variety
of approaches.  System sizing relationships based on experience and on the
rationally developed system models are used.  The trend is toward improved

                                     94

-------
Table 4 .  Condition and Relationships used to Evaluate  Design Sensitivity
a.   Reference Values Used

     Parameter
                                                       Value
                                                                     Units
Influent BODs, So
Temperature
Temperature corrections
Biological rates
Aeration rates
Metabolism rate constant, km*
Synthesis rate constant, ks*
Endogenous rate constant, ke*
Aeration time, t
Solid retention time, 8C
Saturation Oa concentration, C$
Mixed liquor Oa concentration, C|_
b. Adjusted Values Used (20% error)
Parameter

S«
Temperature
km

k|
t
0c













250
20

1.02
1.024
15.0
10.4
0.01
5
120
9.1
2.0

Values
Lower
200
16
13.85
9.61
0.0092
4
9.6
Upper
300
24
16.24
11.26
0.0108
6
144
mg/L
°C


-
1/h
1/h
1/h
h
h
mg/L
mg/L

Units

mg/L
c
1/h
1/h
1/h
h
h
c.    Equations and Relationships  Used

     S  =  1  +S^j          Ma  = y-^S—         He     0.2 keMa0c

                                  0C  + ke


     _   .  1.5 (So - S)      1.42 (Ha  + Me)
     rr         t         "        0C


     ||  =  a kLa (B Cs  -  CL) - rr            kLaT  =  kLa2o


       _                           1,^. n
     Oa'Transfer Efficiency   -  1.5
                                                             JT-20)
                                      95

-------
                       Figure 7: Results  of parameter  analysis for rational design  of aeration  systems.
                   120
IO
CT>
                                                                                                         Ref 0.8
            Parameter

            Cond i t i ons
                           « c
                           u o

-------
understanding of the requirements of the biological systems and  the aeration
system.  There is still a major problem with translating the results of
testing an aeration system in one tank configuration and size to another.

     Review of design parameters and data requirements for sizing aeration
systems shows that organic and hydraulic loading parameters are  of major
importance.  However, the 3 factor and the CL value to be maintained in the
aeration tank are the most significant parameters to the design.

REFERENCES

1.   "Recommended Standards for Sewage Works."  Great Lakes-Upper Mississippi
     River Board of State Sanitary Engineers, Publ. by Health Education
     Service, Albany, H.Y. (1971).
2.   Wastewater Treatment Plant Design.  Joint Committee of the  WPCF and ASCE,
     Publ. by Water Poll. Control. Fed., Washington, B.C. (1977).
3.   Eckenfelder, W.W., Jr. and O'Connor, D.J., "The Aerobic Treatment of
     Organic Wastes." Proc. 9th Ind. Waste Conf., Purdue Univ.,  Ext. Ser. 89,
     512 (1955).
4.   Stensel, H.D. and Shell, G.L., "Two Methods of Biological Treatment
     Design."  Jour. Water Poll. Cont. Fed., 46_, 1, 271 (1974).
5.   McKinney, R.E. and Ooten, R.J.,  "Concepts of Complete Mixing Activated
     Sludge."  Transactions, 20th Sanitary Engineering Conf., Univ. of
     Kansas, Lawrence (1970).
6.   Lawrence, A.W. and McCarty, P.L.,  "Unified Basis for Biological Treatment
     Design and Operation."  Jour. Sanit. Eng. Div., ASCE, 96, SA 3, 757 (1970).
7.   Eckenfelder, W.W., Jr. and Ford, D.L., Water Pollution Control.  The
     Pemberton Press, Austin, Texas (1970).
8.   Metcalf and Eddy, Inc., Wastewater Engineering.  McGraw Hill Book Comp.,
     New York (1979).
9.   Benefield, L.D. and Randall, C.W.,  Biological Process Design for
     Wastewater Treatment.  Prentice-Hall, Inc., Englewood Cliffs, N.J. (1980).
10.  Lawrence, A.W.,  "Modeling and Simulation of Slurry Biological Reactors."
     In Mathematical Modeling for Water Pollution Control, Ann Arbor Science
     Publ., Inc., Ann Arbor, Mich. (1975).
11.  APHA, AWWA, WPCF, Standard Methods for the Examination of Water and
     Wastewater, 15th ed., Amer. Public Health Assoc., Washington, D.C. (1980).
12.  Eckenfelder, W.W., Jr. and O'Connor, D.J., Biological Waste Treatment.
     Pergamon Press, New York (1961).
13.  Gilbert, R.G., "Measurement of Alpha and Beta Factors", Proceedings,
     Workshop Toward an Oxygen Transfer Standard, U.S. Environ.  Prot. Agency
     Rep. No. EPA-600/9-78-021 (1979).
14.  Benjes, H.H. and McKinney, R.E.,  "Specifying and Evaluating Aeration
     Equipment".  Jour. San.  Engr. Div., ASCE, 93_, SA 6, 55 (1967).
15.  Stukenberg, J.R., Wahbeh, V.N. and McKinney, R.E., "Experiences in
     Evaluating and Specifying Aeration Equipment".  Jour. Water Poll. Control
     Fed..  49, 66 (1977).
16.  Stukenberg, J.R. and Wahbeh, V.N.,  "Surface Aeration Equipment:  Field
     Testing Versus Ship Testing".  Jour. Water Poll. Control Fed., 50, 2677
     (1978).
                                      97

-------
17.  Conway, R.A. and Kumke, G.W.,  "Field Techniques for Evaluating Aerators'
     Jour. San. Engr. Div.. ASCE. 92, SA2, 21 (1966).
18.  Shell, G. and Cassady, T., "Selecting Mechanical Aerators", Industrial
     Water Engineering. July/Aug., 21 (1973).
19.  Busch, A.W.,  Aerobic Biological Treatment of Wastewaters, Olygodynamics
     Press, Houston, Texas (1968).
20.  Salzman, R.N. and Lakin, M.B., "Influence of Mixing in Aeration",
     Proceedings, Workship Toward an Oxygen Transfer Standard, U.S. Environ.
     Prot. Agency Rep. No. EPA-600/9-78-021 (1979).
DISCLAIMER
      The work described  in  this  paper was not funded by the U.S.. Environ-
mental  Protection  Agency.   The contents do not necessarily reflect  the
views of the  Agency,  and no official endorsement  should be inferred.
                                      98

-------
DESIGN PROTOCOL FOR AERATION  SYSTEMS  -  UK PERSPECTIVE

Arthur G Boon and Brian Chambers, Water Research Centre,  England.

ABSTRACT

     The successful design  of an  activated-sludge aeration system depends
upon calculating the mass of  oxygen that must  be supplied to the process in
order to produce an effluent  of specified quality.   Maximum aeration
efficiency is achieved by taking  into account  the variations in oxygen
demand that will occur and  designing  the aeration system  to meet these
requirements.
                                                                t
     Factors that must be considered  include the effects  of aeration tank
geometry and aerator configuration on the performance of  the overall
aeration system.  In order  to achieve optimum  aeration efficiency it is also
necessary to match the supply of  oxygen with demand such  that excessive
dissolved oxygen concentrations do not  occur within the aeration tank.

     Recent surveys of aeration systems have shown that potentially the
fine-bubble diffused-air  system is capable of  achieving high aeration
efficiencies.  Details of process modifications  carried out at a large
sewage treatment works in the UK  to enhance the  aeration  efficiency of a
fine-bubble system are given  in an appendix.


1.  INTRODUCTION

     The rate of  treatment  of waste water by the activated-sludge process is
limited by the rate at which  micro-organisms in  the sludge can biochemically
oxidize the impurities,  it  cannot be  increased simply because dissolved
oxygen (DO) is readily available. The  maximum rate will  depend on a number
of  factors including treatability of  the waste water, viability of micro-
organisms in the sludge,  temperature  of aerated  sludge, and concentration of
dissolved oxygen which should be  maintained above critical values.  For
carbonaceous oxidation the  rate-limiting concentration is normally about 0.5
mg/1, and for nitrification the optimal concentration is  about 2 mg/1.

     The rate of treatment  will also  be affected by the need to produce
consistently an effluent of required  quality.

     Having established  the BOD loading rate of  the sludge to produce an
effluent of required BOD, with or without nitrification,  it is then possible
to  calculate the concentration of suspended solids that should be maintained
in  the aeration tank and  the  period of  aeration.  From these values the
wastage rate of sludge can  be calculated and from assumptions about the
settleability of the sludge the size  of final  settlement  tank may be
estimated.
                                      99

-------
       The  total oxygen  requirement  can  be  calculated  from the BOD to be
satisfied and the ammoniacal nitrogen  to be oxidised.   The aeration
intensity (g 02/m3 of  aeration  tank  per  hour)  can  then be  estimated and the
aeration system designed  to provide  the  required intensity at  expected
loading rates which will  vary diurnally  and maybe  seasonally.

     In this paper the method adopted  in the UK for  design of  an aeration
system is described, taking into account the factors outlined  above in order
to achieve  minimum use of energy for aeration  and  hence maximum aeration
efficiency.


2.  OXYGEN  REQUIREMENTS

2.1.  Total Demand

     The total rate of oxygen consumption by the micro-organisms in an
aeration tank is equal to the rate oxygen is required  for  carbonaceous
oxidation by heterotrophic bacteria, plus the  rate of  oxidation of
ammoniacal  nitrogen by autotrophic bacteria, less  the  rate at  which nitrate
is denitrified for carbonaceous oxidation.

     To satisfy BOD, the  heterotrophic bacteria consume oxygen for their
growth and  for respiring  endogeneously.   To achieve  full nitrification, the
additional  amount of dissolved  oxygen  consumed is  4.3  times the concen-
tration of  ammoniacal  nitrogen  oxidised  to  nitrate (assuming 6% of the
nitrogen is assimilated by the  bacteria).   Where anoxic conditions exist in
an aeration tank, by design or because of inadequate aeration, the amount of
oxygen available to satisfy BOD is equal to 2.83 times the concentration of
nitrate-nitrogen denitrified(l).

     Experience in the UK and the USA(2,3)  has indicated that  the total
amount of oxygen required for treatment  of  settled sewage  can  be calculated
from the following empirical equation.


   R = 0.75 F (cBS-cBE) + 2 x io~3 CM  v  + 4.3  (CANS-CANE)?   (i)

                              - 2.83
where R     is the mass of  oxygen  required  (g/h)

      F     is the flow rate  of  settled  sewage (nrVh)

      CBS   *s the BOD °f settled  sewage (mg/1)

            is the BOD of effluent  (mg/1)

            is the concentration of suspended solids in the aeration  tank
              (mg/D
                                     100

-------
            is the concentration  of  suspended solids in the aeration tank
            (mg/1)

            is the volume of  the  aeration tank (m^)

            *-s t^ie ammoniacal-N content  of settled sewage (mg/1)

      CANE  *s fche ammoniacal-N content  of the effluent (mg/1)

and   CONE  *s the nitrate-N  content of  the effluent (mg/1)

     If nitrite were formed as a  significant proportion of the oxidised
nitrogen in the effluent, the amount of  oxygen required for nitrification
and available from denitrification would be reduced  accordingly.  However,
it is unlikely that nitrite will  constitute a high proportion of the
oxidised nitrogen in an effluent  unless  the waste water contains an
inhibitor to the autotrophic  bacteria which oxidise  nitrite to nitrate.  Any
nitrate or nitrite present in the settled sewage  which is reduced to
nitrogen by anoxic conditions in  the aeration tank will also reduce the
amount of oxygen required for treatment.

     The above equation can be simplified, so that


     R = 0.75 F (CBS-CBE) + 2 x io~3 CM  v + 1.47  F (CANS-CANE) +

                          2.83 F  CONE                                 (2)


     When treating screened and degritted raw sewage,  it has been found(4)
that the amount of oxygen required is greater than that calculated from the
above equation.  However, by  changing the first term from 0.75F to l.OF the
oxygen used for such treatment can be calculated  with  reasonable accuracy
using the modified equation.

     Data have been obtained  which indicate that  the above equations can be
used to predict the oxygen demand of plants operated within the range of
sludge loadings from about 0.05 to 1 g BOD/g sludge  suspended solids day.

2.2.  Variations in Demand

     The total quantity of oxygen will vary depending  on the mass of sludge
in the aeration tank (Cj^V), which will vary with  the growth rate of the
micro-organisms (and hence the wastage rate of the sludge), which will
depend on the sludge loading  rate (g BOD/g sludge day), and this will vary
with the effluent quality required.

     A relation between the BOD of effluent and the  loading rate of sludge
has been obtained (Fig. 1) with pilot plant treating settled sewage of
average BOD 250 mg/1 at temperatures of  about 15°C(5).  This relation has
been substantiated by data obtained  from full-scale  plants treating settled
sewage with BOD in the range  from 150 to 250 mg/1.

                                      101

-------
 LU
 LL
 o
     100
      80
      60
      40
 8   20
 CD
     Figure 1.
        123^

        SLUDGE  LOADING  RATE  ( g BOD/g d )

Relation between loading rate of settled domestic sewage
(average BOD 250 mg/1 and  BOD of final effluent
     From the curve (Fig. 1) it can be seen that to produce an effluent  of
BOD of 20 mg/1, the sludge loading should not exceed about 1 g BOD/g
sludge.  Assuming the settleability of the sludge and the horizontal  area  of
the final settlement tank were sufficient to operate the plant with a mixed
liquor suspended solids (MLSS) of 3000 mg/1, the rate of oxygen demand to
satisfy the BOD of 230 mg/1 (excluding nitrification) would be about  92
g/nH h. This calculation is based on substitution of the appropriate  data
into the first two terms in Equation 1 which relate to carbonaceous
oxidation.  Thus the rate of oxygen demand (g/m3 h) is equal to


             0.75(250-20) + 2 x 10~3 x 3000 x 2  or 92, g/m3 h
                               2

             where V/F has the value of 2 h.

     If nitrification were required and the maximum loading rate  to ensure
the growth of nitrifying bacteria was assumed to be 0.2 g/g d,  the rate  of
oxygen demand to oxidise ammoniacal nitrogen and satisfy BOD could be
calculated from the first three terms of Equation 1.  Assuming  the
                                     102

-------
ammoniacal N to be  oxidised  to nitrate was 30 tng/1 and the MLSS were 3000
mg/1, the rate of demand would be equal to

             0.75X250-10) +  2  x 10~3 x 3000 x 10 + 4.3 x 30 e/m3 h
                                10                                 '

            or 37 g/m3  h, where V/F has the value of 10 h.

     If an anoxic zone  were  provided to denitrify the nitrate in the
recycled  sludge and the rate of sludge recycle was equal to the flow-rate of
sewage so that 50%  of  the nitrate was reduced,  the total rate of demand for
oxygen calculated from  Equation 2 (with the same values for the other
conditions assumed  for  the nitrifying plant) would be equal to

   0.75 (250-10) +  2 x  10"3  x 3000 x 10 + 1.47 x 30 + 2.83 x 15  g/m3 h
                                  10

             or 32.6 g/m3 h, where V/F has the value of 10 h.

     Diurnal variations in rate of flow of waste water and in concentrations
of BOD to be satisfied  and  ammoniacal nitrogen to be oxidised will result in
significant variations  in oxygen demand.  These variations can be calculated
using Equation 2  provided  the relevant data were available.  In practice, it
has  been  found that, to avoid lack of dissolved oxygen limiting the rate of
treatment,  the period  for which the total demand should be calculated is
equal to  the retention  time  of mixed liquor (recycled sludge plus sewage) in
the  aeration tank.  The  demand thus calculated could be used to determine the
rate of oxygen transfer required from the aeration system in a 'uniformly-
mixed* aeration  tank.   However, most activated-sludge plants have aeration
tanks which have  'plug-flow* mixing characteristics in order to produce
sludges of  good  settleability.

     In  'plug-flow* aeration tanks the demand for oxygen will be greater at
the  inlet end of  the aeration tank where the availability of substrate does
not  limit the  rate  of  oxidation compared with that at the outlet end where
the  rate  of  oxidation  is limited because most of the BOD will have been
satisfied.   In  order to estimate the variation in demand along the length of
a 'plug-flow' aeration tank  it is necessary to have some information about
the  treatability  or biodegradability of the sewage as the presence of
substances  which  are potentially inhibitory to biochemical oxidation will
have an  effect.   In the absence of such information it would have to be
assumed that data obtained  from the treatment of sewage which contain low
proportions  of  industrial  effluent could be applied.

     In  the  UK,  data have  been obtained from three large treatment works(6)
and  from  pilot-scale studies(7) which show that the rate of oxygen demand
for  nitrification is constant from the inlet to the outlet of a  'plug-flow*
aeration  tank provided the  concentration of dissolved oxygen is always above
about 2 mg/1 and  the concentration of ammoniacal nitrogen has not reached
the  rate-limiting value of  about 0.5 mg/1.
                                      103

-------
     To satisfy BOD, the demand at  the  inlet will  depend  on dilution rate
with recycled sludge and with the contents of  the  aeration  unit  which will
depend on the dispersion rate (and  hence  the degree  of  'plug-flow')  of the
mixed liquor.  Little information has been published to enable accurate
calculation to be made of the effect of such variables on the demand for
oxygen.  Observations at a number of treatment works in the UK,  having
aeration tanks in which there exists 'plug-flow' because  there is  a  high
length to width ratio (>20) and the dispersion number (8)  is very low (about
0.01 to 0.1), indicate that on average  about 60% of  the BOD was  satisfied in
the first 20% of the aeration tank volume.  In the remaining volume  of the
aeration tank about 15% of the BOD was  satisfied in  the next 20% of  the
aeration tank volume followed by 10, 10,  5% for each following 20% of the
volume.  It was also found that these values varied  diurnally as can be seen
from the figures in Table 1 which indicate the range of oxygen demand for
each zone .

             Table 1.  Variation in proportion of  oxygen  demand
                along the length of a plug-flow aeration  tank

             Proportion of           Proportion of oxygen demand
          aeration tank volume       for  carbonaceous oxidation
                                     Average     Range  diurnally

                 20                    60            40-85
                 20                    15             5-20

                 20                    10             5-15
                 20                    10             5-15

                 20                     5            <1-10
     If the results shown in Table 1 were related  to  the  examples calculated
previously for a non-nitrified effluent and  a  fully nitrified effluent,
without and with denitrification, the following distribution of oxygen
demands within the aeration tank can be calculated.

1.   Non-nitrified effluent (BOD <20 mg/1)

     BOD of settled sewage       250 mg/1

     MLSS                       3000 mg/1

     Sludge loading                1 g/g d

     Period of aeration of sewage  2 h

     Total oxygen demand         184 g/m^
                                     104

-------
            Proportion of tank      Rate of oxygen demand in each
               volume (%)                portion of tank (g/m^ h)
             Inlet     20

                       20

                       20

                       20

             Outlet    20
Average
276
69
46
46
23
Range
184-391
23-92
23-69
23-69
<5-46
2.   Nitrified effluent (BOD <20 mg/1 Ammoniacal N <1 mg/1)

     BOD of settled sewage                      250 mg/1

     Ammoniacal nitrogen of settled sewage       30 mg/1

     MLSS                                      3000 mg/1

     Sludge loading                             0.2 g/g d

     Period of aeration of sewage                10 h

     Total oxygen demand                        370 g/m^

     Oxygen demand for nitrification            129 g/m^

     Oxygen demand for carbonaceous             241 g/m-'
       oxidation
     If conditions in the aeration  tank  (DO  and  concentration  of  ammoniacal
N) were such that the rate of oxygen demand  for  nitrification  were  constant
throughout the period of aeration,  the rate  of demand  in  each  portion of the
tank would be 12.9 g/m^ h.  It can  be assumed that  the rates of oxygen
demand for carbonaceous oxidation would  again be divided  in  the proportions
shown in Table 1 so that the rates  can then  be calculated and  they  are given
below.
                                     105

-------
         Proportion of tank           Rate of oxygen demand  in each
            volume (%)                  portion of  tank  (g/m3 h)
                                          Average        Range

        Inlet      20                       72         48-102
                   20                       18           6-24

                   20                       12           6-18

                   20                       12           6-18

        Outlet     20                        6         <1-12
     The total rates of demand for oxygen  (to satisfy  BOD  and  oxidise
ammonia) in the aeration tank would be as  follows.

           Proportion of tank       Rate of oxygen demand  in each
             volume (%)               portion of tank  (g/m3 h)

                                        Average        Range

           Inlet    20                     85        61-115

                    20                     31        19-37

                    20                     25        19-31

                    20                     25        19-31

           Outlet   20                     19        13-25
2.3.  Nitrified effluent with denitrification (BOD <20 mg/1 Ammoniacal  N
      <1 mg/1 N trate N <15 mg/l)~

     In this case, the 15 mg/1 of nitrate N that would be used  to  satisfy
BOD would reduce the oxygen demand for carbonaceous oxidation from 241  g/m3
to 197 g/m3.  This would reduce the rates of demand in each portion of  the
tank although, like the example above, the rate of oxygen demand for
nitrification would remain constant at 1-2.9 g/m3 h.


3.  SUPPLY OF OXYGEN

     In most treatment plants in the UK, dissolved oxygen is  supplied to the
activated-sludge process by either fine-bubble diffused-air or  mechanical
surface aeration.  There are 6 oxygenated plants using the Unox process,
about 10 with oxygen injection using the Vitox process to uprate existing
aeration equipment and 2 Deep-Shaft plants.

     Air is normally diffused through ceramic diffusers  into  the mixed
liquor near to the bottom of the tank to produce fine-bubbles  (2 to 5 mm
diameter) so as to achieve adequate mixing and maximum contact  period.

                                     106

-------
     There are two basic  types  of  mechanical  surface device,  one which has a
vertically mounted (cone-type)  aerator  and  the  other a horizontally mounted
(brush-type) aerator.  Tests  in the  UK  of  these aeration devices have
shown(9) that the fine-bubble diffused-air  system can have the greatest
aeration intensity (g  Q£  supplied/m3 aeration tank capacity per hour) and
efficiency (g 02 supplied/Wh),  provided the system is correctly installed,
operated and maintained  (Table  2).

     The aeration intensity  and aeration efficiency are directly affected by
changes in the deficit of dissolved  oxygen.  Thus an aeration device
operating in mixed liquor to  maintain the  concentration of dissolved oxygen
at 20% of the air-saturation  concentration  will have only 80% of the
intensity and efficiency of  oxygen transfer shown in Table 2.
    Table  2.    Aeration intensity and efficiency of conventional systems
        of  aeration as  measured in tap water (2 to 8 m deep)  containing
        surfactant  (about 5 mg/1).  Similar results have been obtained
                            in aerated mixed liquor
        Aeration system       Maximum aeration     Aeration efficiency
                             intensity (g/m3 h)           (kg/kWh)

    Diffused-air

       Fine-bubble
       (<5mm diameter)               200                   1.5-3.6

       Coarse-bubble                100                   0.9-1.2
       (>5 mm diameter)

    Mechanical surface
    aerators

       Rotating vertically          125                   1.5-2.2

       Rotating horizontally        100                   1.2-2.4
 All values obtained at maximum deficit of dissolved oxygen
 Aeration efficiency calculated from total energy supplied.


      Comparison of the aeration intensities given in Table 2 with the
 examples calculated in Section 2.2 shows that the rates of oxygen transfer
 by the aerators could limit the rate of treatment at the inlet end of
 'plug-flow1  aeration tanks particularly with mechanical surface aerators.
                                      107

-------
     The methods used in the UK of testing an aeration  system  to  determine
its ability to transfer oxygen into solution have been  described(9). ^  The
results obtained have shown(9) that conducting tests in 'clean water'  using^
the unsteady-state method is not entirely satisfactory  because 'clean  water
can easily be contaminated by traces of surface-active  substances (including
some that might be added with chemicals used to deoxygenate  the water).   It
is now widely accepted in the UK that one or two tests  should  be  duplicated
with the second test conducted in water to which anionic surface-active
agent has been added to achieve an average concentration during the  test of
about 5 mg/1.  The surface-active agent is added to the water  used for
testing aerator for three important reasons.

1.   It minimises the effects of traces of other surface active contaminants
     (such as oil and greases).

2.   It affects the rate of oxygen transfer so that the rate measured  is
     similar to that in mixed liquor at the inlet end of the aeration  tank.
     At the outlet end the rate is usually similar to that in  'clean tap
     water'.

3.   It enables reproducible results to be obtained irrespective  of  the
     purity of the water - thus tap water, river water, or final  effluent
     may be used.

     To ensure that the rate of supply of oxygen by an  aerator is equal  to
lished over a wide range of conditions.  This is because the rate of oxygen
transfer is affected by changes in the conditions such  as degree  of  puri-
fication of sewage, configuration of aerators in the aeration  tank and
geometry (volume, depth and shape) of the aeration tank, and variation in
concentration of dissolved oxygen.  Considerable data are now  available  in
the literature which describe the effects of these changes in  conditions on
the performance of many different types of aeration device(10-13).  In this
paper are summarised the effects such changes in conditions  have  on  the
performance of an aerator with particular reference to  the fine-bubble
diffused-air system because most recent research in the UK has concentrated
on this system.

3.1.  Effect of Changes in Degree of Purification

     The performance of an aerator can be measured in aerated  activated-
sludge during treatment of sewage using the steady-state method(9).   Such
measurements have been used to establish the effects of a surface-active
agent as outlined above.  As most tests of aerators are carried  out  in
'clean water' a factor of proportionality, termed  'alpha* is used to relate
the performance of an aerator in 'clean water' to  that  in aerated
activated-sludge.

     The value of 'alpha' for a given aeration system is equal to the
performance of the system (aeration efficiency or  intensity) in  aerated
liquor under specified conditions divided by the corresponding performance
under similar conditions in clean water.  Published  results  show  that the


                                      108

-------
performance of an aeration  system varies significantly according to  the
nature of theTliquor,  the contaminants present, and the intensity of
aeration(ll).

     The value of 'alpha' for  a given aerator does not remain constant.  The
efficiency of  transfer of oxygen into aerated liquor varies with the degree
of purification of the waste water.  Experiments with a fine-bubble  aeration
system, operating in an essentially  'plug-flow* unit, have shown (Fig. 2)
            0-8
         I
         CL
         LL
         O
         HI
         z>
         _J
            (X
              0
                 Inlet
Depth of
water (m)
  3-7   o
  6-1   v
  8-1   a
.   Outlet
                          100        200        300       £00
                 OXYGEN  ABSORBED BY  SEWAGE  DURING
                 TREATMENT  (mg/l)
 Figure 2.  Variation of  alpha  factor with degree of purification of sewage
variations in 'alpha1  from 0.3 at  the start of the treatment  (when the waste
water was first brought  into  contact with recycled sludge)  to 0.8 at
completion of treatment  (when a high-quality, fully nitrified effluent had
been produced).  The results  showed that the performance  of the aerator at
the inlet end of the plug-flow unit was similar to that measured in tap
water with 5 mg of added anionic surfactant per litre,  under  similar
conditions of air-flow rate,  temperature, tank geometry and diffuser
configuration.  The performance at the outlet of a plug-flow  unit
corresponds approximately to  that measured in clean water under similar
conditions.

     The performance of  mechanical surface aerators installed in a plug-flow
aeration unit also changes as the  sewage becomes purified.  With such
aerators the value of 'alpha' may be about 1.2 initially  and  may decrease to
about 1, corresponding to a decrease in the concentration of  contaminants.
However, in some circumstances, particularly with hori-zonatally rotating
aerators, the initial value of 'alpha' may be about 0.8 and it may increase
to 1 as the concentration of  contaminants decreases.
                                    109

-------
     'Alpha* values do not indicate the relative  efficiency  of  aeration
systems.  A fine-bubble diffused-air system may have an  'alpha1  value as low
as 0.4 at the inlet end of a plug-flow aeration unit,  and  in a  mechanically
operated system it may be as high as 1.2.  The relative  efficiencies of
different aerators can be judged only by comparison of their mass-transfer
coefficients, aeration intensities, or aeration efficiencies when tested
under similar conditions in aerated liquor (or in  tap  water  with or without
added surfactant).  For example, a fine-bubble aeration  system  may have an
aeration efficiency of 2 g/Wh at the inlet of a plug-flow  aeration tank and
4 g/Wh at the outlet, whereas a mechanical surface aerator could have
corresponding values of 1.8 and 1.5 g/Wh respectively.   In clean water the
fine-bubble system may have an efficiency of about 6 g/Wh(9).

3«2-  Effect of Aerator Configuration

     Tests carried out by WRC showed that spacing  and  positioning
(configuration) of diffusers in an aeration tank  had significant effect on
aeration intensity of a fine-bubble aeration system while  changes in depth
of  immersion had  little effect(ll).  The results  (Fig. 3)  show  that the
intensity was independent of depth except with the greatest  density of
diffusers (Fig. 4), when it increased slightly with increased depth.

     The depth of mixed liquor above the diffusers of  most fine-bubble
plants in the UK  is within the range 2.5-5 m.  Within  this range, the
variation in aeration intensity with depth is slight for all configurations
of  diffusers (Figs 3 and 4) so that average values of  intensity can be used
confidently to examine the effect of changes in configuration and air-flow
rates (Table 3).

     From these results (Table 3) the following assumptions  can be made
within the limits of the data presented.

1.  For a given flow rate of air per diffuser, the change  in aeration
    intensity is  proportional to the change in number  of diffusers/m^
    surface area.

2.  For a given number of diffusers/m2 surface area, the aeration intensity
    changes in proportion to changes in flow-rate  of air/dome.

     The aeration intensity of surface aeration systems  has  been shown to
vary with rotation rate and depth of immersion of  the  aerator and with
geometry (depth and surface area) of the aeration  tank(10,13).
                                      110

-------
            240
             220
             200
             180
             160
           - 100
           <  80
           LU
              60
              20
No. of diffusers
  •   5
  •  10
  o  15
  7  10
  A  25
  a  25
Nominal air flow-rate  per dome (mJ'h)
    1-7
    0-85
    0-85
    1-7
    0-85
    1-7        D
                       1234567

                        DEPTH  OF WATER  ABOVE  DIFFUSERS  (m)
    Figure  3.   Variation of aeration intensity with depth  of water for
                different configurations of diffusers  (as shown in Fig.4)
                and air flows when  aerating tap water  containing surfact-
                ant (about 5 tng/1)  with standard dome  diffusers (178 mm
                diameter) sited on  the bottom of the tank.
                    T
                   1-5m
                   1
Figure 4.   Configuration of diffusers in aeration tank  (plan not to scale)

                                     111

-------
  Table 3.  Variation  of  aeration intensity of fine-bubble aeration with
           configuration  of  diffusers (depth of immersion 2.5-5 m)

 Total flow-   Number  of  diffusers   Flow rate of air  Aeration intensity
 rate of  air   per m2  surface area    (m3/h per dome)       (mg/1 h)
    (m3/h)

     42                 11.1                  1.7                193

     21                 11.1                  0.85               108
      8.4              4.4                  1.7                 70

     25.2              6.7                  0.85                65
      8.4              4.4                  0.85                50

      8.4              2.2                  1.7                 29

3.3. Effect of Concentration  of Dissolved Oxygen

     It  is normal for the  maximum performance of an aeration system  to  be
quoted   assuming that the  system is used to aerate water containing  zero
concentration of dissolved oxygen.  However, the rate of transfer of oxygen
is directly proportional to the dissolved-oxygen saturation deficit.  Hence
the rate at which oxygen is transferred into mixed liquor in an aeration
tank and the aeration efficiency of the aerator will vary, depending on the
concentration of dissolved oxygen.  For example, when nitrification  is
required,  and hence an aerator is operated in mixed liquor to maintain
consistently 20% of the saturation concentration of dissolved oxygen, the
rate and efficiency of oxygen transfer will be 0.8 of the quoted maximum.
At times of low loading, if 60% of the saturation concentration of dissolved
oxygen were present in the mixed liquor, the rate and efficiency of  oxygen
transfer would be only 0.4 of the maximum.

     An important reason for using commercial oxygen instead of air  in  the
activated-sludge process is that it makes possible a high rate of oxygen
transfer when operating with high concentrations of dissolved oxygen (40-60%
of the air saturation value)  in the mixed liquor.  However, operation of an
activated-sludge plant at such high concentrations of dissolved oxygen  has
little effect on the rate of treatment, and hence the size of the plant, or
the quality of effluent obtained(15).
4.  OPTIMISATION OF AERATION EFFICIENCY

     Tests of aeration systems have shown (Table 2) that potentially the
fine-bubbled diffused-air system should use less electrical  energy to
dissolve a given weight of oxygen when compared with most other  systems,
particularly coarse-bubble aeration.

     As a result of studies carried out by WRC to evaluate a fine-bubble
aeration system(9,ll), and further work undertaken by the US EPA in


                                     112

-------
co-operation with the WRC(3),  Thames  Water Authority, Severn-Trent
WaterAuthority and others,  the following factors have been determined as
important in order to obtain  optimum  performance (lowest energy input).

(i)    The geometry of  the  aeration basin and configuration of the diffusers
       are probably the most  significant factors to affect aeration
       efficiency.  It  is recognised  that there are certain advantages to
       plug-flow of liquor  along  the  aeration tank, particularly in relation
       to improved settleability  of activated sludge(8), but a length-to-
       width ratio not  exceeding  12:1 could give optimal aeration effici-
       ency. Tapering of aeration intensity by varying the configuration of
       diffusers along  the  tank should be provided but not to such an extent
       that, at times of low  BOD  and  ammoniacal nitrogen load, the required
       flow-rate of air is  below  the  minimum specified by the manufacturers
       or the intensity of  mixing is  inadequate to keep the activated sludge
       in suspension.  The  optimum depth of liquor in the aeration tank is
       probably between 4.5 and 6 m(3).

(ii)   Excessively high concentrations of dissolved oxygen (DO) have little
       benefit to the treatment process but significantly reduce the
       aeration efficiency.  Monitoring and control of DO by varying the
       air-flow rate should result in an increase in aeration efficiency and
       a saving of energy.  To achieve effective control of DO, a minimum of
       3 independently  controlled air-grids, each with air-metering and
       linked to a DO probe via a programmable controller, would be
       required.  The control system  could be provided by a mini-computer
       capable of varying the output  of air supplied by a variable number of
       blowers one of which could be  driven by a speed-controlled DC
       electric motor.

(iii)  The intensity of aeration  should correspond to the requirement for
       oxygen (Para. 2.2).  Changes in temperature and treatability of waste
       water will affect the  rate of  biochemical oxidation so that the
       maximum intensity of aeration  should be based on the maximum rate of
       carbonaceous and nitrogenous oxidation anticipated, preferably as
       determined from  experience rather than based on rigid application of
       such data as are presented in  this paper.
5.  DISCUSSION

      It  is  probable  that  the saving of energy in the activated-sludge
process  requires  additional  capital expenditure.  The benefits can only then
be  judged by  economic  evaluations,  taking into account any differences in
capital  and operating  costs.  Such  costs are beyond the scope of this paper,
and data, such  as capital costs,  are best obtained from quotations of
manufacturers in  specific circumstances.
                                     113

-------
     Experience in the UK, over many  years,  has  shown that most aeration
systems operated reliably, have few significant  problems,  and are not
expensive to maintain.  Contrary  to popular  belief  regarding clogging of
fine-bubble diffusers, the data obtained  from a  survey of  19 full-scale
sewage-treatment plants, the  smallest  treating 0.15 m-Vs and the largest
treating 9 m3/s, showed that  the  frequency of cleaning diffusers ranged from
once every 4 years to once in 15  years; many of  the plants had been in
operation for 6 to 12 years without the need to  clean the  diffusers(S).
Facilities were provided at all the works to clean  the air supplied to the
diffusers, and their capital  and  operating costs were relatively small
compared with total costs of  installation and operation of the aeration
system.

     It was unfortunate that  many of  the  plants  surveyed(3) were being
operated at aeration efficiencies significantly  lower than the maximum
anticipated from pilot and full-scale  testing of the fine-bubble
aerators(9,ll,12).  In most cases, lack of facilities to control air-flow
rates resulted in inefficient operation of the aerators, but at a few works
operation at sludge-loading lower than design, and/or the  adverse effects of
industrial waste waters, were thought  to  influence  aerator performance.  All
the plants produced effluents of  a high quality, most with oxidation of
ammonia and some with controlled  denitrification in anoxic zones.  The
average oxygen-transfer efficiency (calculated to correspond to the rate of
oxygen transfer at a DO of 20% of the  air-saturation concentration) ranged
from 2.9 to 1.0 g/Wh, with a  median value of 2.2 g/Wh.   At maximum deficit
of dissolved oxygen (corresponding to  operation  at  zero DO) the median value
of efficiency would be 2.75 g/Wh, which is in good  agreement with the
results obtained in water plus 5  mg anionic  surfactant/1 and in aerated
liquor(9).  At the works surveyed, the observed  aeration efficiencies could
be increased by changes in the operation  and control of the aeration
equipment which could save about  a third  of  the  energy previously used.

     The energy needed to dissolve oxygen can be minimised by the use of
fine-bubble aeration, but there are other factors that  should be taken into
account in the selection of an aeration system.  These  include capital cost
and size of plant, settleability  and  rate of production of activated sludge,
and also reliability, flexibility, and durability of the aeration
equipment.  In respect of some of these factors  there could be advantages in
the use of new processes such as  the  Deep-Shaft  and systems for injection of
oxygen which can be used to supplement the aeration intensity provided by
existing equipment, particularly  during periods  of  high loading(15).
However, to produce an effluent of low BOD,  the  rate of treatment will be be
low and hence there is no need to supply  oxygen  at  a high  intensity, except
perhaps at the inlet of a 'plug-flow'  aeration tank.

     Overloading at the inlet of  a 'plug-flow' aeration plant could be
avoided by making changes to  the  conditions  under which sewage and recycled
sludge are brought into contact.  For  example, such changes could include
the provision of a separate tank  at the inlet which would  be about 15% of
the total aeration volume.  In the case of a nitrifying plant, this tank
should be mixed by sub-surface mechanical stirrers  (using  energy of about
10 w/m-*) to achieve controlled anoxic  conditions so that nitrate, available

                                    114

-------
from the recycled sludge fed  to  the  tank;  this would reduce the total amount
of air required to treat the  sewage  and  may also improve the settleability
of the activated sludge(8).   In  the  case of a non-nitrifying plant, the
first tank could be used to aerate recycled activated sludge, before the
addition of sewage to the  second aeration tank.  This would minimise the BOD
loading at the inlet of the second tank, and reduce the overall sludge
loading of the plant for a given concentration of suspended solids in the
aerated liquor, so reducing the  risk of  slime formation on the diffusers and
improving the effectiveness of DO control.

     To demonstrate the low use  of energy, consistent quality of effluent,
and reliability, flexibility  and cost-effectiveness of using a fine-bubble
diffused-air system the Thames Water Authority (TWA) and WRC have set up a
demonstration at Rye Meads Sewage Treatment Works.  This demonstration is
financed by WRC at a total cost  of about £500 000 with contributions from
the UK Department of Energy (£145 000),  United States-EPA ($72 000) and
Environment Canada ($60 000).

     At Rye Meads two aeration channels  have been modified, one to be
operated to produce a fully-nitrified effluent (with partial denitrification
in an anoxic zone) and  the other to  produce a non-nitrified effluent of
30:20 (SS:BOD) standard for 95%  of the time, and their performance compared
(particularly load treated and energy used) with an existing channel which
produces a fully-nitrified effluent  (also with partial denitrification).

     The design protocol described in this paper and the philosophy of
design and operation outlined below  were used to provide the basis for
modifications to the aeration units  at Rye Meads.  Details of the process
modifications carried out  at  Rye Meads to enhance the aeration efficiency
are described in the Appendix.
6.   SUMMARY

     The  philosophy of  design and operation to achieve maximum aeration
efficiency can  be  summarised as follows.

1.   The  rate of oxygen supply has to equal the rate of biochemical
     oxygen  demand.

2.   To obtain  the correct rate of supply requires the correct configuration
     of diffusers  and design of aeration tank geometry.

3.   To vary the rate of supply to correspond with changes in oxygen demand
     requires  facilities for monitoring and control of dissolved oxygen by
     varying the flow-rate of air.

4.   To reduce  the energy used when the air-flow rate is reduced requires a
     number  of  blowers, one or two of which would be of controlled and
     variable  speed.
                                      115

-------
5.   There are other factors which must be considered to avoid  loss  of
     efficiency and problems with operation and these are listed below.

(i)    Minimum flow-rate of air, at minimum demand rate, should never be
       lower than the minimum rate specified by the manufacturers  if
       problems of clogging of diffusers, inadequate mixing of  sludge and
       sewage, and low aeration efficiency are to be avoided.

(ii)   At all flow-rates of air, the valves used for control of flow rate
       should be near fully open as possible - one valve should always be
       almost fully open.  Control of flow-rate should be largely  achieved
       by varying the pressure via a variable-speed blower.

(iii)  Guidelines specified by the manufacturers for maintenance and
       cleaning of equipment should be fully observed.

(iv)   Growths of slimes on diffusers, resulting in inefficient coarse-
       bubble aeration can be avoided by improved methods of mixing  of
       sewage and recycled sludge at the inlet of the aeration  tank  to
       reduce high loading rate of oxygen demand.  The provision of  anoxic
       conditioas at the inlet of a nitrifying plant, or aeration  of
       recycled sludge for a non-nitrifying plant, would assist to avoid
       such problems.

(v)    Maximum aeration efficiency would be obtained in  'uniformly-mixed'
       tanks although aeration tanks with 'plug-flow' mixing characteristics
       tend to produce activated sludge of good settleability.  A  length-
       to-width ratio not exceeding 12:1 could give optimal aeration
       efficiency without poor settleability.  The optimum depth of  liquid
       in the aeration tank is probably between 4.5 and 6 m.
7.  ACKNOWLEDGEMENTS

     The authors acknowledge the support and encouragement  they  have
received from staff of the Eastern Division of Thames Water Authority for
the project at Rye Meads Sewage Treatment Plant.

     This paper is published by permission of Dr J Cuthbert,  Director,
Process Engineering, Water Research Centre, UK,

8.  REFERENCES

1.  Cooper, P.P., and Wheeldon, D.H.V., "Fluidized and  Expanded-Bed Reactors
    for Waste-water Treatment."  Wat. Pollut.  Control,  ]9_ (2), 286 (1980).
2.   Eckenfelder, W.W. and O'Connor, D.J. ."Biological Waste Treatment1,1 Perga-
    mon Press, New York  (1961).
3.   Houck, D.H. and Boon, A.G., "Survey and Evaluation  of Fine-Bubble Dome-
    Diffuser Aeration-Equipment", US EPA,  Cincinnati, EPA-600/52-81-222,
    8 pp. (1981).
                                     116

-------
 4.   Johnstone,  D.W.M. and Carmichael, W.F., "Cirencester Carrousel Plant:
     Some Process Considerations", Wat. Pollut. Control. In press  (1982).
 5.   Boon, A.G.  and Burgess, D.R., "Effects of Diurnal Variations  in Flow of
     Settled Sewage on the Performance of High-Rate Activated-Sludge Plants",
     Wat. Pollut. Control. _71, 493 (1972).
 6.   Downing, A.L., Boon, A.G. and Bayley, R.W., "Aeration and Biological
     Oxidation in the Activated-Sludge Process", J. Proc. Inst. Sew. Purif.,
     66 (1962).
 7.   Downing, A.L., Jones, K. and Hopwood, A.P., "Some Factors of  Importance
     in the Design of Activated-Sludge Plants", In Joint Symposium on New
     Chemical Engineering Problems in the Utilization of Water, American
     Institute of Chemical Engineers and Institution of Chemical Engineers
     (1965).
 8.   Chambers, B. and Tomlinson, E.J., "Bulking of Activated Sludge:  Preven-
     tative and Remedial Methods", Ellis Horwood Ltd., Chichester  (1982).
 9.   Boon, A.G., "Measurement of Aerator Performance", In The Profitable
     Aeration of Waste Water, BHRA Fluid Engn, Cranfield, Bedford MK43 OAJ
     England, pp 13-18 (April, 1980).
10.   Downing, A.L., Bayley, R.W. and Boon, A.G., "The Performance  of Mechani-
     cal Aerators", J. Proc. Inst. Sew. Purif., 231 (1960).
11.   Lister, A.R. and Boon, A.G., "Aeration in Deep Tanks:  An Evaluation of
     a Fine-Bubble Diffused-Air System", Wat. Pollut.  Control, 72, 590 (1973).
12.   Schmit, F.L., Wren, J.D. and Redman, D.T., "The Effect of Tank Dimensions
     and Diffuser Placement on Oxygen Transfer," J. Wat. Pollut. Control Fed.,
     50, 1750 (1978).
13.   Crabtree, H.E. and Wood, A.J., "Uprating Aeration Processes," Wat. Pollut.
     Control. ^8_, 27  (1979).
14.   Blachford, A.J., Tramontini, E.M. and Griffiths,  J.A.,  "Oxygenated
     Activated-Sludge Process-Evaluation at Palmersford", Wat. Pollut.  Control
     In press (1982).
15.   Downing, A.L. and Boon, A.G., "Activated-Sludge Type Processes - High
     Intensity Systems," In Comprehensive Biotechnology Part 3, Ed. C.W.
     Robinson and J.A. Howell.  Pergamon Press, Oxford, New York,  Toronto,
     (To be published in 1983).


DISCLAIMER

     The work described in this paper was not funded by the U.S. Environ-
mental Protection Agency.  The contents do not necessarily reflect the views
of the Agency, and no official endorsement should be inferred.
                                      117

-------
                                   APPENDIX

Details  of  Process Modifications carried out at Rye Meads  STW to enhance
aeration efficiency
Al.   INTRODUCTION

      There are 12 activated-sludge aeration tanks at Rye Meads  arranged as 3
groups,  or stages, of 4 tanks.  Each stage is associated with a set of six
final sedimentation tanks as shown in Figure 1A.  Tanks 5 and 8 of Stage 2
were  selected for modification for the project.  Tank 5 was modified to
produce  a fully nitrified effluent of 15:10:5 (SS:BOD:Amm N) for 95% of the
time  and Tank 8 was modified to produce a non-nitrified effluent of 30:20
(SS:BOD) standard for 95% of the time.
 unit.
     Aeration Tank 9 of Stage 3 was selected to be monitored  as  the control
                  FIGURE 1A   RYEMEADS STAGE 2  LAYOUT  OF AERATION AND

                  FINAL TANKS (DIAGRAMMATIC I
   Tank 8
   Tank 7
   Tank 6
  Tank  5
                                                 Mixed liquor
           Settled sewage
                                           Recycled ^xj
                                           sludge     L-
A2.  SEPARATION  OF  PROCESS  STREAMS IN TANKS 5 AND 8

     In order  to make  valid comparisons between the  two modified aeration
tanks it was necessary to isolate them so that each  received  a separate
stream of settled sewage and recycled sludge.  Under the  previous
configuration  a  single pump was used to return sludge  to  a  common channel at
the inlet of Tanks  5-8 inclusive, whence it was distributed as shown in
                                       118

-------
Figure 1A.  A second  pump  was  installed to return sludge from Tank 8 to the
inlet of that tank only  and  the existing pump was used to return sludge to
Tank 5.  Tanks 6 and  7 were  drained and sealed for the period of the project
and unused mixed liquor  and  recycled-sludge channels were blanked off.  This
arrangement allowed one, two or three final sedimentation tanks to be used
for the settlement of sludge from each of Tanks 5 and 8.  It was also
necessary to install  an  additional recycled-sludge flowmeter, a new
recycled-sludge channel  and  a  surplus-sludge metering pump for Tank 8 and a
surplus-sludge metering  pump for Tank 5.


A3.  REDUCTION OF AERATION TANK LENGTH TO WIDTH RATIO AND PROVISION OF
     INCREASED SEWAGE FLOW TO  AERATION TANKS

     Each aeration tank  in Stage 2 consisted initially of 4 passes connected
in series as shown in Figure 2A.  Each pass is 70 m in length, 4.3 m wide
and 3.2 m deep.  The  total volume of each was therefore about 3800 m3.

     The length to width ratio of Tanks 5 and 8 was reduced to half its
original value by using  only the first two passes of each tank.   To achieve
this the outlet from  Pass  2  to Pass 3 of each tank was sealed and new
mixed-liquor outlets  were  cut  in the tank walls to allow flow to the
settlement tanks.  It was  not  considered that the reduction in length to
width ratio would lead to  any  problems with sludge settleability.
     FIGURE 2A  RYE  MEADS STW   AERATION  TANK  LAYOUT


I
PASS 4 —
f r
1
PASS 3
PASS 2
t,
i
«»
1 U
PASS1 	 	
in


                                                                .MIXED LIQUOR
SETTLED
SEWAGE
                                                  RECYCLED
                                                  SLUDGE
                                    119

-------
     Each unmodified aeration  tank at Rye Meads receives  on  average about
100 1/s of settled sewage.  Although the modified 2-pass  aeration tanks are
only half of the original volume  it was not  proposed  to decrease  the sewage
flow proportionally.  Since it is known that increasing the  rate  of
treatment leads to potential increases in aeration efficiency  and since tne
amount of oxygen required per unit BOD satisfied decreases with increasing
load, it was decided to design initially for an average sewage flow-rate of
85 1/s in the nitrifying plant and 140 1/s in the non-nitrifying  plant.
This represents an uprating, in terms of flow per unit tank  volume, of about
70% and 180% respectively.

     The reduction in length to width ratio and the proposed increase in
sewage flow-rates were chosen  after consideration of  possible  effects on
sludge settleability and the consequence of such effects  on  operation of
final settlement tank.  It was felt that the increase in  aeration efficiency
resulting from the change in length to width ratio would  be  small in
comparison to that resulting from correct configuration of diffusers, the
increased sewage flow-rate, and effective control of  aeration. However, to
obtain a substantial increase  in aeration efficiency  in an unmodified 4-pass
tank would have required an increase in sewage flow to a  rate  where
overloading of the settlement  tanks might have occurred.  The  conversion to
2-pass operation brought about the desired increase in plant loading by
reducing the aeration tank volume.  The effect of increasing the  sewage
flow-rate on the operating characteristics of the final sedimentation tanks
is summarised in Table 1A.  Table 1A shows in parts (a) and  (b) the proposed
operating conditions for the tanks modified to produce nitrified  and
non-nitrified effluent.  In part (c) of Table 1A it can be seen that in
order to achieve the same sludge loading it would be  necessary to increase
the sewage flow-rate to the values shown.  Under such conditions  the sludge
settleability would have to remain at values less than those shown (as SSVI
at 3.5 g/1) otherwise solids overflow with the effluent would  occur.

     The settleability of sludge at Rye Meads has traditionally been very
good (SSVI 3.5  50 ml/g) and it was not envisaged that the process
modifications would lead to any substantial deterioration in settling
properties.  However, it was considered that the greater  permissible
increase in SSVI resulting from 2-pass operation as compared with 4-pass
operation justified the choice of a 2-pass system, and provided an adequate
safety margin.
                                    120

-------
        Table 1A.  Effects of increased  sewage  flowrate  on  operating
                characteristics of final  sedimentation tanks
(a)  Modified Nitrifying Plant
Number of Final Settlement Tanks
Aeration tank volume (m^)               1900
Final tank surface area (m^)            380
Sewage flowrate (1/s)                    85
Recycled sludge flowrate  (1/s)           70
MLSS (mg/1)                             5000
Hydraulic residence time  (b)            3.4
Sludge loading (d-1)                    0.1
  (for BOD influent of 120 mg/1)
Max. permissible SSVI3.5  before          90
   solids loss (ml/g)

(b)  Modified Non-Nitrifying Plant

Aeration tank volume (m^)               1900
Final tank surface area (m^)            380
Sewage flowrate (1/s)                   140
Recycled sludge flowrate  (1/s)          105
MLSS (mg/1)                             3700
Hydraulic residence time  (h)            2.2
Sludge loading (d"1)                    0.2
 (BOD influent of 120 mg/1)
Max. permissible SSVIs-s  before         105
 solids loss (ml/g)

(c)  Unmodified 4-pass Aeration Tank
              1900
               760
                85
                70
              5000
               3.4
               0.1

               110
              1900
               760
               140
               105
              3700
               2.2
               0.2

               145
1900
1140
  85
  70
5000
 3.4
 0.1

 130
1900
1140
 140
 105
3700
 2.2
 0.2

 165
                                 Nitri-  Non-     Nitri-  Non-     Nitri- Non-
                                 fying   nitri-   fying   nitri-   fying  nitri-
                                        fying           fying          fying

Aeration  tank  volume  (m-*)
Final  tank surface  area  (nr)
Sewage  flowrate  (1/s)
Recycled  sludge  flowrate (1/s)
MLSS (mg/1)
Hydraulic residence time (h)
Sludge  loading (d~*)
Maximum permissible 88713.5
before  solids  loss  (ml/g)

* Recycled sludge flowrate per  unit  tank area  approaching critical value
3800
380
170
140
5000
3.4
0.1
70*
3800
380
280
210
3700
2.2
0.2
90*
3800
760
170
140
5000
3.4
0.1
90
3800
760
280
210
3700
2.2
0.2
105
3800
1140
170
140
5000
3.4
0.1
105
3800
1140
280
210
3700
2.2
0.2
125
                                     121

-------
 A4.  DESIGN AND INSTALLATION OF MODIFIED DOME-DIFFUSER AERATION SYSTEM

 A.4.1.  Estimation of Parameters Required for Calculation of Oxygen Demand

      The average flow-rates of 85 1/s and 140 1/s were taken as the basis
 for the calculation of the oxygen requirements of the  nitrifying and
 non-nitrifying plants respectively.  Other design parameters were obtained
 by examination of several years analytical and operational data provided by
 TWA staff at Rye Meads.

      It was found that the ranges of the parameters  required were
 distributed in such a way as to give approximate straight line relations
 when plotted on probability graph paper.  Figure 3A  shows, as an example,
 the relation found for the ammonia nitrogen concentration in settled
 sewage.  From such relations the values shown in Table 2A were obtained.
        FIGURE 3A  DISTRIBUTION  OF NH,-N CONCENTRATION
        IN  SETTLED SEWAGE
       005
02 05  1   2
  PERCENTAGE
5  10  20 30 tf) 50 60 70  80   90  95  98
OF VALUES NOT  EXCEEDING STATED VALUES
                                                                995   990
9999
    The range of  settled  sewage flow-rates shown in Table 2A was  calculated
by assuming the proposed  flow-rates of 85 1/s and 140 1/s were median
values.  The range  of  flow-rates expected for these median values was
                                      122

-------
estimated by assuming the flow-rates  would  vary in the same manner as
historical data indicated.

     Similarly it "was assumed  that  the modified nitrifying plant would
produce a distribution of effluent  nitrate  nitrogen which was approximately
the same as existing aeration  tanks.
     Table 2A.  Parameters  used  for  design of modified aeration systems

                                  Nitrifying  plant     Non-nitrifying plant
                                  5%< Median   95%<     5%<     Median   95%<


Settled sewage flowrate  (1/s)     64      85    160     100       140    260

Settled sewage BOD  (mg/1)         75     130    220      75       130    220

Settled sewage NH3~N  (mg/1)       20      33     44      20        33     44

Effluent N03-N (mg/1)             14      22     32     N.A       N.A    N.A


A.4.2.  Calculation of Oxygen  Demand Range in Modified Plants

     The range of oxygen demands to  be  satisfied in the modified aeration
tanks was  calculated  using  Equation  2 and the values shown in Table 2A.  In
addition it was assumed  that:-

(i)    The MLSS concentrations in the nitrifying plant and the non-nitri-
       fying  plant  would be maintained, on average, at 5000 mg/1 and
       3,700  mg/1 respectively-

(ii)   The BOD in the effluent from  the modified plants would be 5 mg/1 and
       20  mg/1 respectively for 95%  of  the time.

(iii)  The effluent ammoniacal N concentration in the nitrifying plant would
       be  less than 1 mg/1  or  effectively zero.

     The calculated range of oxygen  demands is shown in Table 3A.

    Table  3A. Range  of  oxygen demands  to be  satisfied in  nitrifying and
                         non-nitrifying  aeration tanks

                                 Nitrifying             Non-nitrifying

                             5%<  Median   95%<     5%<   Median    95%<

   Oxygen  demand  (kg/d)      1120    1975    4910     675    1320     3690
                                      123

-------
A.4.3.   Determination of Number and Arrangement of Dome-Diffusers

      The determination of the number of domes required  and  their arrangement
on  the  floor of the modified aeration tanks was carried out  according to the
following general procedure, taking into account  the  criteria mentioned.

(i)    The number of domes in each tank should be such  that  the minimum
        oxygen demand can be satisfied by a dome operating slightly above the
        recommended minimum air flow-rate per dome of  0.9 m  /h.

(ii)    The maximum oxygen demand should be satisfied  by the  domes operating
        with acceptable pressure-drop.

(iii)  The arrangements of the domes in the aeration  tanks  should be tapered
        from inlet to outlet such that the median  oxygen demand is satisfied
        by the domes operating at average air flow-rate,  resulting in the
        desired uniform dissolved-oxygen concentration throughout the tank.

(iv)    The arrangement of domes into separate 'zones' along  the length of
        the aeration tank should facilitate the proposed strategy of
        dissolved-oxygen control.

      Consideration of the above criteria generally results  in the calcu-
lations required being of an iterative nature and since the  criteria are
interactive some element of compromise is usually necessary  before a
reasonable, practical design results.
 a,  Nitrifying






c
o
r
ft
anoxic
wage
2


2


2


i.



•'

              I
               recycled  sludge
b. Non-nitrifying
                   1  sewage
                                        End of pass 1

recycled
sludge —


^
1

I
1


1


1


i.

	 	


    FIGURE AA PROPOSED  P.O. CONCENTRATIONS  AND DIVISION OF AERATION TANKS INTO
             CONTROL ZONES

                                      124

-------
     The basic concept of  the  layout  of the modified tanks at Rye Meads was
originally envisaged as shown  in  Figure 4A.  This figure shows the desired
dissolved oxygen concentrations considered necessary for efficient operation
in the nitrifying and non-nitrifying  systems, and the proposed division of
the aeration tanks  into separate  control zones each with its own fully
modulating air supply.

     In the nitrifying plant an anoxic  zone was incorporated as shown at the
inlet to the aeration tank.  Anoxic  zones result in the following process
advantages:-

(i)    The settleability of sludge in plants incorporating anoxic zones has
       been shown to be superior  to  that from similar plants without anoxic
       zones.

(ii)   The use of anoxic zones results  in nitrate being reduced to gaseous
       nitrogen and hence  the  concentration of nitrate discharged in plant
       effluent is  reduced.

(iii)  The use of nitrate  oxygen  to  satisfy BOD in anoxic zones results in a
       reduction  in the overall oxygen  requirements of the process and hence
       a saving in  energy.

(iv)   Problems associated with denitrification in final settlement tanks
       are minimised.

(v)    Slime growths on the diffusers in the aerated zone downstream of the
       anoxic zone  should  be minimal.
     Following  the  anoxic zone the aeration tank was divided into 3 separate
zones by  the  use  of baffles.   Within these zones the layout of diffusers and
the  action  of the control system was such that the dissolved-oxygen
concentration was maintained  at about 2 tng/1.   This is considered adequate
for  the maximum rate of conversation of ammonia to nitrate by nitrifying
bacteria.

     The  fourth aerated zone,  occupying only a small fraction of the total
tank volume would be used to  maintain a higher concentration of dissolved
oxygen (4 mg/1).  It was recognised that a lower aeration efficiency would
result in this  zone but this  would involve only low flow-rates of air since
the  design  was  such that treatment would normally be complete at the
beginning of  this zone.  It has been shown that high concentrations of
dissolved oxygen  are desirable to stimulate the action of activated-sludge
protozoa.   Protozoa are instrumental in removing finely-divided suspended
material  of bacterial origin  and their presence is necessary for the
production  of high-quality effluents.  It would also provide extra aeration
capacity  in the event of an unexpectedly high demand for oxygen when the
rate of supply in that zone could increase by about 60% corresponding to a
decrease  in dissolved oxygen  to 0.5 mg/1.
                                     125

-------
     The initial proposals  for  configuration  of  diffuser-domes in the
non-nitrifying tank were similar  to  those  for the  nitrifying tank.  However,
in the former case air would  be supplied to all  five  sections of the
aeration tank with the first  compartment used only for  re-aeration of
recycled sludge.  Settled sewage  would  enter  the aeration tank at the
beginning of the second compartment  as  shown  in  Figure  4A(b).

     In aeration tanks treating sewage  at  a high rate,  as proposed in the
non-nitrifying plant,  it has  often been found that microbial slimes grow on
the diffusers near the tank inlet.   This process causes clogging of the
pores of the diffusers and  the  production  of  coarse bubbles.  As a result
there is a  reduction in rate  and  efficiency of oxygen transfer and also an
increase in back-pressure across  the diffusers and hence energy is wasted
and effluent quality deteriorates.   It  has been  found in the laboratory that
re-aeration of recycled sludge  can prevent the sliming  of the diffusers in
the downstream when sewage  is mixed  with the  re-aerated sludge.

     It was considered that dissolved-oxygen  concentrations  of about 1 mg/1
in the next 3 aerated  compartments of the  non-nitrifying plant would
normally be adequate for carbonaceous oxidation.  The final  zone would be
the same as in the nitrifying plant  for the same reasons.

     At present  there  is no reliable theoretical method for  predicting the
degree of aeration taper required in a  plug-flow activated-sludge aeration
tank with a given oxygen demand.  The variation  in oxygen demand along a
tank is reflected by the change in respiration rate of  sludge that occurs.
The respiration  rate is a function of the  nature and  biodegradability of the
sewage treated and also depends on the  length to width  ratio of the aeration
tank.  A typical curve of respiration rate for a non-nitrifying sludge
versus length along an aeration tank from  inlet  to outlet might look similar
to that shown in Figure 5A(a).  The  respiration  rate  and hence the oxygen
demand, decreases as treatment  progresses, from  an initially high value to a
low level which, if treatment is  complete, represents the endogenous
respiration rate of the sludge.

     The amount of air required in any  particular  zone  of an aeration tank
is therefore proportional to  the  relevant  area beneath  a respiration rate
curve such  as that shown in Figure 5A(a).  Unfortunately it  is impossible to
derive accurately such curves before the aeration  tank  has been constructed.
This presents no disadvantage where  exact  replicas of existing tanks are
being added to a works to accommodate increases  in flow but  presents
problems for a new works design or for  the design  of  aeration tanks to
operate under substantially different conditions from existing aeration
tanks.  In  such cases the variation  in  demand shown in  Table 1 can be used
with reasonable confidence.

     In a nitrifying plant  the  total oxygen requirement includes an amount
of oxygen required for oxidation  of  ammonia.   As the  demand  for oxygen for
nitrification is normally exerted at a  constant  rate  throughout the tank
then the respiration rate curves  for a  nitrifying  and a non-nitrifying
aeration tank treating the  same sewage  might  appear as  shown in Figure
5A(b).   The intensity of aeration required in a  nitrifying plant is less


                                      126

-------
  RESPIRATION
  RATE (mgO»/lh

    (R)
          Air  required is proportional
         " to area under curve
Inlet
TANK  LENGTH
Outlet
                 Nitrification + Carbonaceous
                                             •Carbonaceous
Inlet
TANK  LENGTH
 Outiet
 FIGURE 5A  RESPIRATION RATE  CURVES   IN AERATION  TANKS
                            127

-------
extreme than that required in a non-nitrifying plant.  This  assumption has
been confirmed by measurements made on full-scale aeration tanks.

     The prediction of the air flows required to satisfy  the oxygen
demands shown in Table 3A and result in  the residual dissolved  oxygen
concentrations shown in Figure 4A was made using a computer  program made
available by Ames Crosta Babcock Ltd based on the protocol described in this
paper.  The design and construction of the necessary aeration tank
modifications was undertaken by Hawker Siddeley Water Engineering  Ltd.

     The initial layouts of dome diffusers are shown in Figure  6A  and Figure
7A.  In the nitrifying plant the number  of domes required to satisfy the
given oxygen demands were situated in control zones 5.1,  5.2 and 5.3.  Zone
5.4 was not considered as normally supplying air for treatment  but merely
for increasing the dissolved oxygen concentration as previously described.
In the non-nitrifying plant similarly, treatment was only considered
normally to occur in zones 8.2, 8.3 and  8.4.

     The designs adopted allow easy division of the aeration tank  volume
into control zones corresponding to regions with different configuration of
domes and this was achieved using standard items of aeration pipework.

     The diffuser domes employed were the normal Hawker Siddeley Water
Engineering type selected and installed  after refiring according to the
procedure recommended by the manufacturers.  The diffusers were fitted with
a  'double-orifice' type bolt which results in lower pressure drop  across the
assembly during operation.  Thus it is possible to employ higher air
flow-rates per diffuser without incurring severe energy losses.

     The criterion, for deciding that the initial layout  of  diffuser domes
in the modified tanks was correct, was that the desired dissolved  oxygen
concentrations should be achieved in the control zones when  the plant was
operating under the design average conditions with an equal  air flow-rate
per diffuser in each of the zones.

     The design dissolved oxygen set-points are 2 mg/1 in zones 5.1, 5.2 and
5.3, 1 mg/1 in zones 8.1, 8.2, 8.3 and 8.4 and 4 mg/1 in  zones  5.4 and 8.5.

     It was found that in order to maintain 2 mg/1 in zone  5.3, a  higher air
flow-rate per diffuser was required in that region than in  zones  5.1 and
5.2.  Moreover, increases in load to the plant did not result in increases
in respiration rate in zones 5.1 and 5.2 beyond a certain upper limit.
Increases in load appeared as increases  in respiration rate  towards the
second half of zone 5.3.  This was attributed to the nature  of  the sewage
being treated and it was considered that the short aeration times  existing
in zones 5.1 and 5.2 did not allow certain components in  the sewage to be
fully biodegraded.

     It was also found that the number of diffusers  in  zone 5.4 was not
sufficient to achieve the desired dissolved oxygen concentration of 4 mg/1
at reasonable air flow-rates per diffuser.  After  discussions with Hawker
Siddeley Water Engineering Ltd., it was  decided to modify the dome layout  in


                                     128

-------
FIGURE 6A  INITIAL LAYOUT  OF DIFRJSERS IN AERATION TANK NO 5  I NITRIFYING)


Zone 5'2





Zone 5-3 a










Zone 5-1









i

Zone 5-3 b


ANOXIC ZONE

Zone 54

ZONE      DONE LAYOUT
 51        5 Rows88 per row
 5-2        5  -   64  -   •
 53a       4  -   64  •   '
 5-3 b       4  •   45  •   '
 5-4        3  •   25  •   '
TOTAL NO




75

DOMES
440
320
256
180

1196
% TOTAL
368
26-8
21-3
15-1

100-0
1271
                                                                    Max  Median  Min.
                                                 Oxygen demand (ka/d) 4910  1925   1120
                                                 Total air flow (l/s)   1525   525    290
                                               Air flow/diffusedl/min)   77   26     15
  FIGURE 7A   INITIAL LAYOUT OF  DIFFUSERS  IN  AERATION  TANK NO 8  (NON-NITRIFYING)

Zone 84a














Zone 8-3




Zone 8-4 b



7r\r\a Q O
i.one rvi








Zone 8-5



[| Zone 8-1




ZONE
 81
 8-2
 8-3
 8-4a
 8-4 b
 8-5
              DOME  LAYOUT    TOTAL NO. DOMES    '/.TOTAL
4 Rows,  25 per row
                                100
5
4
H
3
3
88
48
40
45
25
                                   440
                                   192
                                   295 (
                                      (160
                                      (135
                                75
                                   927
                                       47-5
                                       207
                                       17-3
                                       14-5

                                      1000
                    Max.  Median  Mm.

Oxygen demand (kg/d) 3690   1320  675
Total air flow (l/s)   1365    400  185

Air f low/diff user (I/mini   88    26    12
                              1102
                                              129

-------
 zones 5.3.b and zone  5.4.   Accordingly, Tank  5  was  drained and additional
 domes were added to give  the layout shown in  Figure 8A.  Subsequent
 operation of the plant  has revealed that the  layout of the diffusers  is now
 very close to that required and hence no further  modifications were
 considered necessary.
                                                                     168




•*HV«W^— ^^•^WM^^^^^^^^^BV^H^^H^^M^^^
t
ZONE 5 2 '
i
»


'
ZONE 53a '
i
i
_____________
i >
i *
'
>



ZONE 5-t i
' - - 1
•,,,-- - 1
C ' ' ' J
OMVBIM^
1-4
1 >
1 »


»
p ZONE 5-3b
f


RECYCLED SLUOCF
ANOXIC ZONE
SETTLED
SEWAGE
	 ; ;
•
ZONE 5-4
h
1 *








     ZONE
           DOME  L/WXJT   TOTAL NO. DOMES   1, OF TOTAL
5-1
5-2
5-3a
5-3b
5.4
5 ROWS 88 PER ROW
5 -
4 -
4 ••
5 -
64 - -
64 ' '
59 - "
25 - ••
440
320
256>)492
236)
1252
125
                                      35-1
                                      »6

                                      205
                                      166
                                     1000
             MAX.

OXYGEN DEMAND   A910

TOTAL AIR FLOW    1525
AIR FLOW/DOME   73
MEDIAN  MIN.

1975    1120 kg/d

525     290 l/s
 25     U I/mm
                            1377
       FIGURE 8A    FINAL LAYOUT OF DOME DIFFUSERS IN TANKS  INtTRIFYINGI


      At the time of writing Tank 8 has only just been  converted to a  fully
non-nitrifying plant  and  tests  on the suitability of the diffuser layout  are
not  yet complete.
A.5.   PROCESS CONTROL, MONITORING,  DATA LOGGING AND  DATA PROCESSING  SYSTEMS

A.5.1.   Introduction

     The dissolved-oxygen control  system and process monitoring arrangements
associated  with the project are more complex than those  that would be
required were the design philosophy to be replicated at  another site.   Thus
the computer  installed on site is  only necessary to  handle large quantities
of on-line  data and provide a degree of data processing  which would  be
impractical manually.
                                     130

-------
     A schematic diagram of the overall system is  shown  in Figure 9A.
Process instrumentation provides information about  the plant  operating
conditions to the  computer via a telemetry system.  The  programmable logic
controllers (PLC's) which are responsible for maintaining  the dissolved
oxygen at the required  concentrations receive information  from other
instrumentation and take the necessary control action independently of the
computer.  The link from the PLC's to the computer  is only necessary to
provide the computer  with data for processing purposes.
       CONTROL ROOM
         ALARMS
         PLANT STATUS
         BLOWER
         CHANGES
         T.I. 990/10
         COMPUTER

         256 kBYTE
         MEMORY
         10 M BYTE
         DISC
         2 VDU
         2 SILENT 700
         TERMINALS
         1 LIME PRINTER
COMPUTER
ROOM
             TELEMETRY
             SYSTEM

           40 ANALOG «  80
            DIGITAL INPUTS

            176  DIGITAL
            OUTPUTS
            PLANT STATUS
            PRINTOUTS
            ALARMS/ERRORS
            OFFLINE DATA
            PROCESSING
INSTRUMENTS
NOT IN
CONTROL LOOP
                                     INSTRUMENTS
                                     IN CONTROL
                                     LOOP
    PLCs
  FIGURE 9A   PROCESS  CONTROL  AND DATA PROCESSING  BLOCK DIAGRAM

     The  computer processes the information received and  provides various
summaries,  some of which are directed to the Rye Meads  Control Room for the
information of Thames Water Staff and some of which  are for direct
comparisons of plant performance.

     The  instrumentation and control configuration for  the two modified
plants  is  shown in Figure 10A.  Each zone of the aeration tanks (as
described  previously) has a dissolved-oxygen probe,  a separate, metered,
controllable air supply and a PLC.

     The  air supply to the aeration tanks is provided by a set of six  ,
fixed-speed Rootes blowers, each of which is rated at about 1650 1/s (at
NTP).   The design of Rootes blowers is such that they can be considered as
constant-volume machines (i.e. positive displacement).   Thus increases in
resistance to flow downstream of the blowers are compensated by increases in
output  pressure such that similar volumetric output  is  maintained.
                                      131

-------
                                                         Monlarrg end
                                                         dalo procnsmq
                                                         comoul*'
        FIGunelO A INSTRUMENTATION ANO CONTROL COHPIGUBAriON - A.NON NITRIFYING 9. MITRIFYIMG
                                          KEY TO  FIG IDA

                                          pic   -  programmable logic  controller
                                          A     -  ammonia monitor
                                          F     -  flow recorder
                                          P     -  pressure recorder
                                          T     -  temp,  recorder
                                          0.0.  -  dissolved oxygen recorder
      It  was not considered  justified for the  purposes of this  project to
install  a variable-speed compressor in the system.   Hence the  required
variations in output are provided by the vent-valve shown in Figure 10A.
With  this arrangment the fixed  volumetric output  of the compressors is
modulated by the vent-valve to  provide a variable volume air supply to the
aeration tanks.
                                      132

-------
A.5.2.  Dissolved  Oxygen Control System

     The functions of the dissolved  oxygen control system are:-

(i)  to control  the dissolved oxygen concentration in each compartment
     of each  aeration tank to within ± 0.5 mg/1 of the design DO
     concentration;   and

(ii) to achieve  the required control such that the energy used for aeration
     is minimised  irrespective of variations  in demand for dissolved oxygen.

     In addition the control system  is also able to provide information  for
monitoring  and data processing purposes.

     The layout  of the DO control PLC's  is shown in Figure 10A and the
detailed arrangement of the control  loop  which allows the vent valve
arrangement to interact with the control  system is shown in Figure HA.
             FIGURE 11A  VENT  VALVE AND DO.' CONTROL LOOPS
                                       VENT
            VENT
            VALVE
            PLC
                       CALCULATE
                       PRESSURE
                       REQUIRED
                         PI.D.
            D.O.
            PLC
0.0
CONTROL
RID.
                                             VALVE POSITION
                                                              TANK
                                                              VALVE
                          10.0.
      It is recognised that  the  number of DO control PLC's  installed may be
 greater than the number required  by a potential replicator but  they were
 included to obtain evidence of  the number actually needed  to  achieve
 satisfactory DO control.
                                       133

-------
     Each  control PLC  compares  the  DO  concentration measured in the
appropriate aeration zone with  the  required  set  point.   The errors in this
comparison are  then used  to  calculated desired valve positions using a PID
control algorithm derived for this  application by  Hawker Siddeley Dynamics
Engineering Ltd.  This control  loop cycles every 20-40  seconds.

     The fixed  volume  characteristics  of  the blowers means  that changes in
tank valve positions do not  result  in  the desired  changes in air flowrates
unless  the system interacts  with  the vent-valve.   The PLC which controls the
vent-valve cycles every 5-10 seconds around  another control loop and the
vent valve is opened to the  correct position such  that  a constant pressure
in  the  air supply main results.

A.5.3.  Monitoring, Data Logging and Data  Processing

     In order that the objectives of the  project are met it is necessary to
collect data regarding the operation of the  modified aeration tanks.  An
accurate economic comparison between the  nitrifying plant and the
non-nitrifying  plant and  also the unmodified control plant  requires the
installation of instrumentation and monitoring equipment which would not be
necessary  for the normal  operation  of  a sewage treatment works.

     The on-site computer gathers the  following  information during a cycle
     time  of 10-15 min.

     Settled sewage flow  to  each aeration tank.

     Recycled sludge flow to each aeration tank.

     Volumetric air flow-rate to each  aeration tank.

     Volumetric air flow-rate to each  compartment  of the aeration tank.

     Air pressure in supply  to each aeration tank.

     Air temperature of supply to each aeration  tank.

     Dissolved  oxygen  attained in each compartment of the aeration tanks.

     Ammoniacal N concentration in  the effluent  from the non-nitrifying
     plant.

     Volumetric air flow-rate used  for aeration  of channels carrying mixed-
     liquor to  sedimentation tanks.

     Position of air vent-valve.

     The data gathered by the computer are continuously displayed in the
computer room.  An example of this  display is shown in  Table 4A.  A
print-out of plant conditions similar  to  that shown in  Table 4A is also
produced every  cycle on a terminal  situated  in  the plant control room-  This
terminal is also used  to signal major  faults to  the control room staff and
to request changes in  the number  of compressors  in use.

                                     134

-------
                     Table 4A.  Display of plant status
                Status of tanks 5 & 8 at 15.39, 23 April 1982

                         Tank 5 Nitrifying         Tank 8 Non-nitrifying
                        1234   Zone   12345
DO Actual
   Previous
   Previous
   Previous
   Previous

Air flow 1/s
Previous
Previous
Previous
Previous
Air supply P
Air supply flow
Air supply T
2.2
2.2
2.1
2.0
1.9
183
180
190
185
186

2
2
2
2
2
133
145
165
140
130

.0
.1
.2
.1
.0





4
1.9
2.0
2.0
1.9
1.9
205
212
215
223
217
.2
575


42
.0
4.
4.
4.
4.
4.
52
57
53
55
48
M.
3
2
1
0
1





WG
0.
0.
0.
0.
1.
41
45
40
43
45

1/s
DEGC
9 0.9
9 1.0
8 1.1
9 1.0
0 0.9
182
187
195
180
183
4.
460
42.
1
1
1
0
1
80
82
78
75
77
2
.1
.0
.0
.9
.0





M.
1.1
1.1
1.0
0.9
0.9
123
131
128
125
121
WG
4.2
4.0
3.9
3.8
3.9
31
35
28
25
30

1/s
0
DEGC
ML Air flow rate 5 & 8    Total 65. 1/s
Sewage Flow 1/s Tk 5  90.2  88.5  84.2  87.1  Tk 8 121.0  118.6  122.3 125.9
RAS Flows               87. 1/s            120.  1/s
Air vent valve posn    45% Open  Tk 8  Effluent ammonia concn 15. mg/1
No. compressors in use 1.
                                    135

-------
     Analytical data pertaining to the operation of  the  modified aeration
tanks are archived on a disc storage system by manual  input.   All the
relevant data relating to the unmodified control plant are  also  archived
manually.  These data include:

(i)   the results of analysis of composite samples of  settled  sewage and
      effluent for the nitrifying plant, non-nitrifying  plant, and
      unmodified control plant. (The parameters measured are BOD, COD,
      TOC, total nitrogen, ammoniacal nitrogen, oxidised nitrogen, soluble
      phosphate, anionic surfactant and suspended solids);

(ii)  the results of spot sample analysis to determine mixed liquor
      suspended solids, recycled sludge suspended solids, sludge
      settleability - as SSDI, sludge filtrability - as  GST, and volatile
      matter in the mixed liquor suspended solids; and

(iii) the results of analysis for BOD and ammonia on two-hourly  grab samples
      of settled sewage and for BOD, ammonia and oxidised nitrogen on
      two-hourly grab samples of each plant effluent.

     A computer program has been written to allow derived parameters of
interest for the comparison of the plants, to be calculated.   The program
involves interaction between manual input data and on-line  data  and permits
calculation of the variability of parameters during  the  course of a day.
For example, the variation in oxygen demand experienced  by  each  modified
plant during a day is calculated by combining the results of analysis of
grab samples of settled sewage and effluent with corresponding values of
settled sewage flow-rate.

     A list of the derived parameters and their relevance to the project
objectives is given below.

(i)    % BOD Removal - This parameter allows comparison  of  the performance
       of each plant and is required to demonstrate  that the modifications
       to the nitrifying plant do not result in any  deterioration in
       effluent quality in comparison with the control plant.

(ii)   Surplus sludge production - The increases in  aeration efficiency
       resulting from the plant modifications may be offset if increases in
       surplus sludge production also result.  The costs of sludge disposal
       are a major proportion of operating costs at  sewage  works and depend
       not only on the total amounts produced but also on their
       dewaterability.  Hence the amounts of surplus activated sludge
       produced by the modified plants and the dewatering characteristics of
       the sludge are being closely monitored.

(iii)   Oxygen demand - This parameter is a measure of  the mass of oxygen
       that must be supplied via the flow of air to  an aeration  tank in
       order to satisfy the oxygen requirements as described by Equation 2.
       Oxygen demand varies with time according to the usual diurnal and
                                    136

-------
      seasonal variations in sewage  flow-rate  and  composition.   The oxygen
      demand satisfied by unit input of electrical energy  used  by  the air
      blowers is the basic parameter for  assessing the  efficiency  of
      the aeration systems.
                                                                     r^
(iv)  The ratio of oxygen demand to  oxygen supplied - This parameter
      provides a measure of the efficiency of  the  mass  transfer of oxygen
      from the gas phase to the liquid phase in  the aeration  tank.  The  mass
      flow-rate of oxygen to the aeration tanks  is calculated from on-line
      measurements of air flow-rate, temperature and pressure.

(v)   Energy used for air compression - This parameter  is  a function of  the
      mass flow-rate of air supplied to the aeration tanks and  the pressure
      of the supply.

A6.  SUMMARY OF INITIAL RESULTS

     Delays encountered during the commissioning of the control system have
meant that to date few results have  been  obtained. At  the present time  Tank
8 has still not been completely converted to a non-nitrifying unit and the
total amount of air produced by 1 compressor is  being passed  into  the
aeration tanks, i.e. the vent-valve  and control  system  are not  in
operation.  The results of approximately  1 month's operation  of the plants
are summarised in Table 5A.  It can  be calculated  that  the modified plants
treat a far greater quantity of sewage (expressed  in proportion to the
relative aeration volumes) than the  control plant; the  percentage  increase
relative to the control plant was 84 for  the nitrifying plant and  124 for
the 'non-nitrifying* plant (which was partially  nitrifying -  Table 5A).
Sludge settleability has also improved in the  modified  plants and  there  has
been a slight reduction in the amount of  surplus sludge produced.

     The low aeration efficiencies reflect the lack of  control  of  dissolved
oxygen, however, under manual control of  dissolved oxygen  it  has been
possible to achieve an aeration efficiency of  about 2 kg 02/kWh in Tank  5.
  Table 5A.  Summary  of  average  results  obtained  in modified plants -
                                 March 1982
Settled sewage

BOD (mg/1)
COD (mg/1)
Ammoniacal-N  (mg/1)
Soluble phosphate  (mg P/l)
Anionic surfactant (mg/1)
                                  Nitrifying
                                   Tank 5
113
300
 32
  7.4
 10.5
        'Non-nitrifying1^- Control plant
             Tank 8            Tank 9
113
300
 32
  7.4
 10.5
113
300
 32
  7.4
 10.5
                                    137

-------
Table 5A continued
Operating conditions

Volume of aeration tank (m3)        1900            1900
Settled sewage flowrate (1/s)         88             106
Recycled sludge flowrate (1/s)        95              93
MLSS (mg/1)                         5050            3790
Air flowrate (1/s)                   870             780
Surplus sludge flowrate (1/s)          1.1             2.3
Temperature of aerated liquid  (°C)    12.5            12.5

Effluent quality

BOD (mg/1)                             4.5             6.1
COD (mg/1)                            38              41
Ammoniacal-N (mg/1)                    0.6             5.7
Oxidised nitrogen (mg/1)              18.8            21.31
Soluble phosphate (mg/1)               8.7             8.6
Detergent (mg/1)                       0.3             0.3

Derived data

SSDI (at 3.5 g/1)                       1.75           1.64
Volumetric loading (1/s per  1000 m3)   46             56
Sludge loading (d"1)                    0.09           0.14
Sludge age (d)                          9.1            4.6
Sludge production (g/g BOD applied)     1.222          1.552
                                    3800
                                      96
                                     125
                                    4580
                                    1130
                                       1.7
                                      12.5
                                       3.9
                                      30
                                      <0.2
                                      21.6
                                       9.2
                                      <0.2
                                       1.23
                                      25
                                       0.05
                                      13.1
                                       1.532
Hydraulic residence  time  (h)
Oxygen demand satisfied (kg/d)
Theoretical power consumption
                        (kWh/d)
Aeration efficiency  (kg 02/kWh)
Nitrifying
  Tank 5

    2.9
 1840

 1240
    1.53
                                             'Non-nitrifying1^  Control  plant
                                                   Tank 8          Tank 9
   2.6
2000

1125
   1.83
   4.7
2400

1885
   1.3
1.  Partial nitrification occurring.
2.  These values are probably too high because  the  period  for  data
    collection was  too  short.
3.  The results are low because the system  to control  flow-rate  of air
    and maintain low dissolved-oxygen concentrations was not  in  operation.
                                     138

-------
EXPERIENCES WITH DIFFERENT AERATION SYSTEMS IN GERMANY

Rolf Kayser, Professor
Technische Universitat Braunschweig
Institut fur Stadtbauwesen, Abt. Siedlungswasserwirtschaft
3300 Braunschweig, W. Germany

ABSTRACT

     Results from measurements under process conditions of the oxygen trans-
fer and the aeration efficiency, in conjunction with experiences with
different aeration systems in use in Germany have been examined. It is found
that the aeration efficiency of well-designed diffused air and surface
aeration systems is of the order of 1.5 to 2.0 kg 02/kWh, with the exception
of diffused air in spiral flow tanks, which has a considerably lower aeration
efficiency.

     For the selection of aeration systems numerous factors have to be
considered. Since in future denitrification has to be considered it is shown
that circulation tanks with surface aerators are preferable.
MEASUREMENT OF THE OXYGEN TRANSFER UNDER PROCESS CONDITIONS

      In order to measure the oxygen transfer rate in tap water or mixed
liquorjthe D.O. has  to be unbalanced. In tap water the D.O. is usually
brought to zero by the addition of sodium sulfite. Similary D.O. of the
mixed liquor is reduced to  zero by switching off the aeration. After re-
starting aeration in both clean water and mixed liquor similar reaeration
curves are obtained. They can be evaluated by use of the same procedure.

      An interruption of the aeration in circulation tanks with surface
aerators is illogical since it takes several minutes to build up the
circulation velocity after  aeration has been restarted. Due to the higher
differential velocity between water and aerator during this period the
oxygen transfer is higher.  Another problem associated with the reaeration of
mixed liquor is that at high oxygen uptake rates the D.O. may only increase
by 2-3 mg/1 thus producing  unreliable results. In order to overcome these
problems the reaeration test was inverted: without stopping aeration
hydrogen peroxide was added to the aeration tank, which immediately decays
to water and oxygen  thus raising the D.O. as desired by approximately
10 to 15 mg/1. By continued aeration the excess oxygen is stripped off and
a desorption curve is obtained which is homologous to the reaeration curve
(Figure 1). If the liquid film coefficient k^ has the same value for oxygen
absorption and desorption,  respectively, the slopes of the absorption and
desorption curves at the same value of the apparent D.O. saturation deficit
(cs*-c) and D.O. supersaturation (c-cs*) must then have same value. The
desorption tests therefore, can be evaluated like reaeration tests. The
theory behind this is described in (1). For the evaluation we would like to
thank Brown and Fisette (2) for the use of their computer program. The oxygen
transfer rate under  process conditions is calculated without considering
the  IS-factor. We use the clean water standard (book) saturation value.

                                    139

-------
                    c[mg02/U
                                 c     c_k-a (Cc*-c)
                                             L  l  s
                                                t(min)
               Figure 1.   Oxygen absorption (lower curve)  and
                          oxygen desorption (upper curve)
     The assumptions of testing with peroxide have been verified by  per-
forming a series of several  tests with reaeration and desorption of  D.O.  in
laboratory and full scale tanks.

     Procedures for testing  full scale tanks were developed once the theory
was worked out. After first  considering dilution of peroxide with tap water
it was finally decided to pump technical grade 35% hydrogen peroxide directly
into the aeration tank.

     An even distribution of the peroxide in the whole aeration tank is
essential for the reliability of the results. In tanks with cone aerators
peroxide was dosed at each cone in mid-depth between the cone and the wall.
In spiral flow tanks with diffused air the points of influx should be
separated by a distance of about the tank width. The procedures for  the
other tank types shall not be stressed here.

     In order to check the reproducability of the method, at the same plants
with the same airflow the tests were repeated on different days. Values of
SOTE obtained at one plant,  for example, are:  8.8; 7.3; 7.3; 9,7; 9-1 and
9-1 g/m3-m.  The average is 8.23 and the standard deviation 8.1%. The smallest
value was 11% lower and the  highest was 18% higher than the average. It has
                                    140

-------
to be stressed that these measurements were performed in an extended aeration
plant under full operating conditions without interrupting sewage and excess
sludge flow.

     When the sewage and return sludge flow was interrupted at another plant
on two different days the measurements for one airflow yielded 7.7; 7.9 and
7.7 g/m3-m and for another airflow 9.3; 9.0 and 9.7 g/m3-m. The results were
less scattering than that from measurements under full operating conditions.


REQUIRED OXYGEN TRANSFER

     The oxygen uptake rate depends mainly on the loading conditions and the
degree of nitrification. The loading rate to be selected is a function of
the required effluent quality. In Germany the discharge requirements are
laid down in a federal regulation: with respect to the biological process
only BOD5 is relevant: the BODc of 24 hour composite samples should be below
20 mg/1 80% of the time. This means that in general the effluent BOD has to
be in the range of 10 to 15 mg/1. In order to advise consultants and planners
the German Water Pollution Control Federation (ATV) has issued guidelines
for the design of activated sludge plants (3). Recommended loading rates and
required oxygen transfer rates are collected in Table 1.  The hourly c*.SOTR
is the value of the peak hour. Because most plants are not fully loaded when
they start operation and because of the daily and weekly oscillation of the
oxygen uptake rate it is recommended to be able to vary 
-------
with the tap water data but for process conditions in two different plants
a value of about 5.5 g 02/(m3 air)•(m diffuser depth)   was found. Assuming  a
gross power requirement of about 5 Watthours/m3 air •(m diffuser depth) the
aeration efficiency becomes 1.1 kg 02/kWh (1.8 lb 02/hp-hr).


        Table 2.  Recommended oxygen transfer of fine bubble
                  aeration in spiral flow tanks (Ref. 4)

Clean water
Process conditions
Good conditions
SOTR3 AEb
12 (1.3) 2.2 (3.6)
10 (1.1) 1.8 (3.2)
Fair conditions
SOTE3 AEb
10 (1.1) 1.7 (2.8)
8 (0.9) 1.3 (2.1)
        a g 02/m3 air •(m diff.depth)   x 0-1089 = (% 02/ft diff.depth)

        b kg 02/kWh x 1.643 =  lb 02/hp-hr


     The disadvantage of diffused air in spiral flow tanks is well known:
due to the similar direction of the water movement and the rising air bubbles
the contact time of the bubbles is much lower than if they were rising in
quiescent water.
OXYGEN TRANSFER RATE
Kg
20-
-
10-
.
0
02)/(m3AIRHm DIFF.-DEPTH)] 1%/ft]
— 	 . 	
"""*'"•••—...
BOTTOM COVERED PIPE DIFFUSERS """"" 	
3.5m DIFF/m2 .
TANK DEPTH 3.AO m
SPIRAL FLOW TANKS
0.5 •/• 1.0 m DIFF/m2

2 4 6 8 10

- 2.0
• 1.5
- 1.0

- 0.5
- 0
AIR FLOW RATE [(m3 AIR)(m3TANK) (HOUR)I
           Figure  2.   Specific oxygen transfer rates of a bottom
                      covered diffused air system in clean water
                                     142

-------
     Much better results  therefore, are obtained when the diffusers  are
evenly distributed  on  the tank bottom (5). In Fig. 2 results from  clean
water measurements  in  a tank covered with tubular diffusers are  compared
with the generally  accepted values for spiral flow tanks (6). The  efficiency
of the system tested was  about 70 to 100% higher than that of the  con-
ventional spiral flow  arrangement.
                £  2.0
                t/i
                cc
                LU
                to
t  1.5 -
Q
(T
                m
                LU
                o
                    1.0 -
                    0.5 -
             20
            —L_
                                              [cuft AIR/(ft DIFFl-h]
                                           60     80      100
                        REQ..C SOTR
                        26g/m3 h
                        (1.63 (b/
                        11000 cuft)-h)
                                                    63(3.94.)
                                             i- 6
                         F/M.-0.05
                              0.3
                                              [ft]
                                             - 2
                              2       L       6       8      10
                            AIR FLOW PER DIFFUSER l(m3AIR)/(m DIFFlh]
           Figure 3.   Diffuser spacing and air flow per diffuser
                       for an assumed specific oxygen transfer
                       rate of 10 g/m3 *m
     In order  to^demonstrate the problems associated with an even diffuser
distribution in  Fig.  3 the dependence of diffuser spacing, air flow per
diffuser and required oxygen transfer rate is shown. It was assumed that
the oxygen  transfer rate under process conditions is about 10 g 02/m3  air-m
(p 1%/ft,<*~0.5) ;  values for the required oxygen transfer rate were taken
from Table  1.  As  a design air flow per meter of tubular diffusers 8 to
10 m3/h (80-100  cu ft/ft-hr) are used generally, which would lead to  a
distance between diffusers of about 1.5 m (5 ft) for plants with F/M  =
0.3-0.15. Such an arrangement will create a flow pattern which might  be
closer to spiral  flow than to a bottom cover type; furthermore, sludge
deposits between the diffusers may not be avoided. If, on the other hand,
a spacing of 0.5  m (""1.5 ft) is utilised the air flow per diffuser reduces
to 3.0 m3/m-h  (30 cu ft/ft/hr) at F/M=0.3-0.15, or half the value for
extended aeration.  Subsequently the investment costs for diffusers and
piping would be  fourfold or more. Dome type diffusers may be more useful.
                                     143

-------
 The investment costs however, will be of the same order of magnitude.

      The problem was solved by Pasveer and Sweeris (7), who split aeration
 and mixing.  At flow velocities between 0.3 and 0.5 m/s (1.0 and 1.5 ft/sec;
they found in a one meter  (3 ft) deep tank oxygen transfer efficiencies  ot
about 30 g/m3-m (3%/ft).   The two systems based on this principle which  are
manufactured in Germany are shown in Fig. 4.  In the tank on the left  side tne
diffusers are fixed on a header on the tank bottom.  The water is rotated m
the tank by means of a propeller with submerged gearmotor.
                PROPELLER SYSTEM
                                     ROTATING BRIDGE SYSTEM
                                                        ADDITIONAL
                                                        DIFFUSERS
                 GENERAL SCHEME
                                              AUXILIARY BAFFLE
                                              TO REDUCE VELOCITY
             Figure 4.  Diffused air systems with mechanically
                        forced water flow
      The other system was used only for extended aeration in past. The
 aeration tank is constructed around the final clarifier. The diffusers  are
 fixed at the bridge which rotates with a peripheral velocityof 0.6  to
 0.8 m/s (2-2.5 ft/sec).  Due to the rising bubbles the water in the circular
 tank rotates at about 0.3 m/s depending on the air flow, thus leaving a
 differential velocity of 0.3 to 0.5 m/s. In larger tanks a baffle has to be
 installed in order to reduce the water velocity. In plants with higher
 oxygen uptake rates additional bottom-fixed diffusers have to be installed
 which reduce the water velocity; a baffle in this case is not needed.

      In both plant types we made oxygen transfer measurements under  process
 conditions  (Table 3). The oxygen transfer efficiency of the propeller type
 plant M was  excellent but the aeration efficiency was not as good as one
 may  have expected because the propeller draw 2.5 kW and the blowers  were
 overdesigned with respect to pressure. In plant H in addition to clogged
 diffusers the propellers draw 7 and 5 kW, therefore the efficiency is as
 low  as in spiral flow systems. It has to be mentioned that the manufacturer
                                      144

-------
of this system has improved the propeller. They claim that with the new
propeller the power requirement for circulation is only 1 W/m3 (0.4 hp/
1000 cu ft).
  Table 3.  Results from measurements under process conditions for four
            plants with diffused aeration

Tank diam. (m)
Tank vol. (m3 )
Diff. length (m)
Diff. depth (m)
Air fl. rate (m3 /m3 ••h)
per diff. (m3/m«h)
otSOTR (g/m3-h)
oCSOTE (g/m3 -m)
Pd (kW)
AE (kg/kWh)
Propelle
M
15.5
632
52
3.05
0.8 0.4
9.6 4.8
34 21
14.2 17.8
13.9 8.8
1.56 1.54
r system
17.9
860
66
3.2
1.4 0.6
18 7.5
28 14
6.3 7.7
27.2 12.5
0.95 1.0
Rotating bridg
N
26/38
1735
72 + 120b
2.8
1.1 0.7 0.4
9.5 6.4 3.2
23 18 14
7.8 9.1 13.9
27.5 19.3 10.7
1.44 1.62 2.26
e
B
22/38
2300
90
2.8
0.4
10
9.3
8.2
17.1
1.25
    clogged diffusers
    Fixed on bridge + fixed on the bottom

    Air discharge from blower nameplate

    Wire energy uptake of blower plus propeller or bridge
     The rotating bridge of both plants N and B drew approximately 2.0 to
2.5 kW, which is close the value of 1 W/m3 mentioned previously. The
oxygen transfer efficiency however, is much lower than that of the propeller
system (14.2 - 17.8 g/m3-m as compared with 7.8 - 13.9 g/m3«m). Only when
the air flow rate is far below the design capacity, the aeration efficiency
is good. The reason for the low efficiency at high air flow rates is the
high air load at one point (the bridge). In smaller plants with tank
volumes below 500 m3 the aeration efficiency will be better because the air
load at the bridge is much lower.

     Another invention comes close to a perforated tank bottom; steel
boxes having a perforated rubber sheet on the upper side about
0.25 cm thick. Recommended air flow rates are 0 to 3 m3/m2-h (0 to 10 cu ft/
sq ft/hr). In tap water tests the results were very promising; the transfer
efficiency was about 30 g/m3-m   (3%/ft). Because of the small bubbles
however, the ot-factor was close to 0.3. The efficiency under process con-
                                     145

-------
ditions  therefore, was  of  the  same order to magnitude as other diffused  fine
bubble systems,  excepting  the  spiral flow system which may run at an
efficiency which is  50% or more  lower.

      The so  called "tower" aeration tanks which have been developed by the
chemical industry in which the air flow is low due to the depth and there
fore, the treatment  of  exhausted air becomes cheaper, will not be stressed
here.

Vertical shaft  surface  aerators

      After the  first cone  aerator was imported from England in 1962, another
cone-aerator has been thrown in  the market every year. By the end of the
sixties  Germany had  become the "cone-country".
                                               (E, VIEW FROM TOP)
          Figure 5.  Different types of vertical shaft surface
                     aerators  (schematic)
     Five typical cones are shown schematically in Fig. 5. Cone A  (which  is
in principle the old Simplex-aerator) and cone D may or may not be  used
with a draft-tube.  It is interesting to notice that one manufacturer  changed
the design of his aerator from type C to type A. The headlosses within
aerators of type C  are considerable due to the high velocities. Cones B,  E
and D (without a draft tube) may create higher turbulence  in  the aeration
tank because their  blades-dip right into the water.

     Oxygen transfer and aeration efficiency depend on numerous factors.
Measurements in clean water and under process conditions indicate  that the
value of 06 is close to unity. ATV (4) recommends the aeration efficiency  be
1.8 to 2.2 kg/kWh for good and 1.3 to 1.8 kg/kWh for fair  conditions
(3.0 - 3.6 and 2.1  - 3.0 Ib/hp-hr).
                                     146

-------
     That manufacturers tend to overestimate the capability of their pro-
ducts can be demonstrated by results from tests of three aerators. The City
of Hamburg planned to install 36 cone aerators in a treatment plant
currently under construction. The aerators were to be equipped with two-
speed motors. The maximum oxygen transfer required was 220 kg/h (485 Ib/hr),
at high speed, at low speed it should be half of that. Because of the high
investment costs the three manufacturers (out of six) which tendered the
most reasonable bids were invited to perform tests in advance, all in the
same tank. Two aerators were modified B-types and one was a C-type. The
mechanical data are as follows:

     B1: diam.  3.25 m; speed 33.9 and  25.4 per min
     B2: diam.  3.25 m; speed 39.7 and  29.7 per min
     C  : diam.  2.86 m; speed 41.06 and 30.85  per min
C

SOTR 200 -
(kg/hi

160-
•
120-
iin
I IU
80-

40-
n H
(

) 40 80 120 160 [hp]
/
REQUIRED ^-//•'^m
(HIGH SPEED) <&/ m^
*
-------
 create  shock  loads  in  the  final  clarifiers.  It  is  therefore,  favourable to
 adjust  the water  level from time to  time  only and  to  control  the  daily ^
 fluctuations  by speed  variation  using  two or three speed motors,  or as in
 the main Vienna treatment  plant, by  using fully variable-speed  motors.

      Because  of the success of circulation tanks with cage  and  mammoth
 rotors,  in the Netherlands the   Carrousel  tank with  cone aerators  was
 developed. Fig. 8 shows both a mammoth rotor and a Carrousel  tank as well.

      Other general  problems associated with surface aerators  are  spray^and
 noise emissions.  Today a splash  area should be  covered (Fig.  7).  The City
 of Hamburg for example, covers its tanks  with cone aerators completely.
               - PLASTIC
                           AIR INLET
                                                -12ft-
                                                      BAFFLE
                CONE AERATORS
MAMMOTH ROTORS
          Figure  7.   Spray protection at  surface  aerators
 Horizontal  axis  surface  aerators

      In  the sixties  the  Kessener-Brush  (diam.  0.5 m) was  very popular for
 high  rate activated  sludge plants; a modification,  the  cage-rotor (diam.
 0.7 m) was  used  in oxydation ditches. In  1963  the mammoth rotor (diam.  1.0 m)
 was invented.  It  is  used in circular and  longitudinal tanks,  with the
 circulating horizontal flow generated by  the rotor  (Fig.  8).  The tank depth
 of plants without primary clarifiers should be 2.5  m  (8 ft) and with
 primary  sedimentation the depth may be  3  to 4  m (10-13  ft). The maximum
 depth of immersion is 0.3 m (1 ft). In  order to achieve a better distri-
 bution of the  oxygen-enriched water over  the tank depth downstream,  a baffle
 is needed (Fig. 7).  This baffle reduces the water velocity, which is
necessary in tanks with more than two rotors.  Because there may still remain
 a certain gradient of the oxygen concentration over the depth it is  pre-
 ferable to keep the  separation between  rotors  at 30 m (100 ft) or more.
Measurements in clean water and under process  conditions  indicated that oe,
is close to unity. The oxygen transfer  per meter of rotor per hour is 7 to
9 kg/m-h (4.7 - 6.0  Ib/ft/hr) and the efficiency 1.8  to 2.0 kg/kWh (3.0 -
3.3 Ib/hp-hr).
                                     148

-------
                        a30m
                        (100ft)
                                               MAMMOTH ROTORS
                                  ADD. AER.
CARROUSEL-TANK
WITH CONE-AER.
          Figure 8.  Circulation tanks (schematic)
     We measured the  oxygen transfer in one tank of the Braunschweig treat-
ment plant under process  conditions. The tank volume is 8700 m3 (307,000
cu ft) and it  is equipped with 7 double rotors (2 x 7.5 m or 2 x 24.6 ft)
each driven by a 75 kW motor (100 hp).  A baffle is installed downstream of
each rotor. The results are shown in Table 4, they clearly indicate that
with an increasing number of rotors in  operation the oxygen transfer and
the efficiency decrease.  That this is due mainly to the decreasing velocity
between the rotor and the water is indicated by the decreasing power uptake
per motor. Some sort  adjustable baffle  has to be developed by which the
water velocity can be controlled. The oxygen transfer is most easily con-
trolled by switching  rotors on and off.
                 Table  4.   Results from measurements with
                           mammoth rotors under process
                           conditions
Rotors in operation
Total length of rotors (m)
<*SOTR (kg/m-h)
«
-------
Tank configurations

     The mixing and flow pattern in aeration tanks distinguish the following
tank configurations:

           - plug  flow and cascades
           - completely mixed tanks
           - circulation tanks

Circulation tanks  are similar to completely mixed tanks except that due  to
the wide separation between aerators the D.O. profile within the circulation
tank is sawlike.

     Horizontal axis surface aerators  (mammoth rotors) can only be used  in
circulation tanks. In general diffused air is employed in plug flow tanks
and cascades.  Short and wide tanks using air diffusion can be visualized as
being  completely mixed. Diffused air systems can however, also be applied
in circulation tanks if the water circulation is generated by mechanical
means, e.g. a  propeller. Cone aerators are used in cascaded or completely
mixed  tanks and in circulation tanks (refered to as Carrousel).
 SELECTION OF AERATION  SYSTEMS

 General  considerations

      Criteria  for  selecting aeration systems are adequately outlined by
 v.d.  Emde (8). A step-by-step procedure of some aspects of the decision
 making will be outlined in this chapter.

 Diffused air or surface aerators?

      The recommended tank depth for surface aerators is in the range of  2.4-
 4.6 m (8-15 ft).   In tank deeper than 4.6 m (15 ft) diffused air or other air
 aeration systems must be installed.

      Diffused air  systems seem to be preferable because of the slight
 heating  effect of  the compressed air in areas with a cold climate especially
 for small  plants with long aeration times. In larger plants however, surface
 aerators  can be covered (as shown in Fig. 7) in order to prevent ice from
 growing  on the aerator or the structure. If nitrification is required  even
 during strong winter periods diffused air systems may be superior. Con-
 sequently the headers should be submersed in order to lose as little heat
 as possible.

     Certain wastewater constituents may cause clogging of diffusers.
 It has to be considered that small filaments tend to spin together even
after the best screening and form rags which may wind around the diffusers
 if no primary sedimentation is employed. Only at type C cone aerators
 (Fig.  5)  clogging by rags was observed.
                                     150

-------
     If the exhaust air is to be treated deep tanks with diffused or other
air systems are preferable. An elegant method of air treatment is being
installed by the City of Frankfurt where a two stage activated sludge plant
(both stages with diffused air systems, tank depth 20 ft) is under con-
struction. The first stage aeration tank will be covered and the resulting
spent air and some fresh air will be used as the air supply for the blowers
of the second stage.

     The variation of the oxygen supply by switching on and off blowers or
motors is possible with diffusion in any type of tank and with surface
aerators in circulation tanks. If cone aerators are installed in tanks in
series each aerator has to be independantly adjustable (immersion and/or
speed) for economical operation. A similar problem is the distribution of
the air along aeration tanks.

     Noise and spray problems of surface aerators can be solved by an
appropriate cover of the splashing area and if necessary by a hood over the
motor.

     Smaller plants can be aerated by some means of compressed air only,
because surface aerators for appropriately small oxygen transfer capacities
are not available. In order to reduce operational problems big bubble
aeration is preferable for very small plants.

     Cost comparisons under German conditions indicate that the investment
costs for diffused air systems with a comparable aeration efficiency (evenly
distributed diffusers) are higher than for surface aeration systems in
general. The power costs should be of the same order to magnitude but
maintenance costs for the diffuser cleaning and replacement of diffused air
systems may be higher. The critical points of surface aerators are the
gears and the bearings. However, experience indicates that surface aerators
with well-designed  (safety factor 2.0 to 2.5) gears can be operated without
malfunction for 15 years or more.

Vertical shaft or horizontal axis surface aerators?

     Mechanically, the difference between the two surface aeration systems
is that the vertical shaft aerator receives axial thrusts and radial forces
while the horizontal axis aerator is forced radial only. All the resulting
forces are transfered to the gear unit of vertical shaft aerators. At
horizontal axis aerators the forces are distributed both on the gear and
the bearing. For the same power transmission it may be assumed that the
forces on the gear of horizontal axis aerators are smaller than those of
vertical shaft aerators.

     From the point of view of tank construction costs the differences are
small. The head ends of the circulation tanks for horizontal aerators are
a little more elaborate but on the other hand the cone aerators need bridges
or supports for the aerators. The areas to be covered are of comparable size
but the covers for the horizontal rotors may be prefabricated.
                                     151

-------
     High oxygen transfer rates are possible with cone aerators  (e.g. Hamburg
 220 mg/l/hr). Because of the required spacing of at least 25-30 m  (80 to
 100 ft) between horizontal rotors and an optimal water depth of 2.5 m the
 maximum oxygen transfer rate is about 100 mg/l/hr. If higher oxygen transfer
 rates are required horizontal rotors must not be used.

 Loading rates and process requirements

     In order to achieve a high quality effluent with regard to BOD and COD
 most plants  in Germany today are designed for a loading rate of
 F/M  0.15 kg BOD/kg MLSS/d, at which nitrification will be established.
 Although today ammonia removal is required only at plants in some  larger
 cities, it has to be considered that regulations may call for ammonia removal
 in every plant in future.

     Once nitrification becomes necessary or is considered it is preferable
 to remove as much nitrate as possible by denitrification without using
 external sources of carbon in order to save energy. Two successful deni-
 trification  systems are pre-denitrification (9) and simultaneous deni-
 trification. The latter was first observed in oxidation ditches  (10), and
 the first larger plant with simultaneous denitrification was Vienna
 Blumental (11).

     Simultaneous denitrification is possible in any completely mixed or
 circulation  tank. The secret of the process is a proper control of the
 oxygen transfer. It has to be high enough for nitrification and  low enough
 to establish anoxic conditions at certain times or in certain parts of the
 aeration tank. Simultaneous denitrification is superior to pre-denitrifi-
 cation because no tank separation is needed in order to distinguish between
 anoxic and oxic tankage. In addition it is more flexible because during
 periods of higher loads the denitrification rate increases and consequently
 the anoxic volume or period can be smaller thus leaving more volume for
 nitrification; the opposite holds for periods with low loads.

     Anoxic  zones are established in  circulation tanks with mammoth rotors
 or Carrousel tanks in between the aerators. The length of the anoxic zones
 can be controlled by switching aerators on and off.

     By proper control of aeration it might be possible to obtain  anoxic
 zones in completely mixed tanks (for example with cone aerators  in the bottom
 area);  another solution is to periodically reduce aeration in order to
 achieve anoxic periods. It may be necessary to install some mixing device,
 e.g.  a propeller, in order to sustain mixing during periods of reduced
 aeration at least in smaller plants.

     In order to make use of the good flexibility of simultaneous  deni-
 trification aeration control is necessary. At the Vienna-Blumental plant  the
rotors  were timer operated at first. Later a continuous respirometer was
developed (12) which was specially adapted to the process. It is still  in
operation and works successfully.
                                     152

-------
     It is thought that the nitrate concentration in the aeration tank would
be a better control parameter.  First nitrate probes were tested; the results,
however, were not promissing. Next a continuous photometric NC>3-N02-analyzer
was assembled. The basic problem was to obtain a continuous sample from the
aeration tank without suspended solids. After trying different filters
without success, the problem was solved by use of an ,.Ultra-Filtration-Tube"
from Kent-BBC. The flow diagram of the set-up is shown in Fig. 9. The
results on Fig. 10 indicate that it is possible to achieve a very low nitrate
concentration in the effluent.
AER
i
i
i
ATION TANK
t 1
I 	 v
t
crrrt

^
\
o-
c™_:s

SUBM. ULTRA FILTER r-
PUMP — i
1 CAMDI C

— juu mumin.
-30 m3/h
(150 gal/min)
[
D O O
I
i

ii
kJ
i
i
*
1 	 T
	 	 	 	 1
NOX-
ANALYZER
RECORDER
CONTROLLER
          Figure 9.  Aeration control for simultaneous denitrification
             mg/l  NOx-N

             2.0n

             1.0-

             0.5-


              0-
              mg/l  02
               .E

~"
»-

\ \
f

\Hr-ir-i

r-\ o n
I
i — inn
                  6 A.M.     12
6  P.M.     12
          Figure 10.  Nitrate concentrations and D.O.  in an aeration
                      tank with controled aeration
                                     153

-------
     Phosphorus  is another source of concern today. It is known that  the
phosphorus  concentration of domestic wastewater can be reduced considerably
solely by microbial activity. A mixing tank for wastewater and return  sludge
is needed ahead of the  full denitrifying plant.

     In many  cases operating without primary sedimentation solves the
problem of bulking sludge due to filamentous bacterial growth. Mixing  of
the wastewater and the  return sludge ahead of the operation tank^as  for
phosphorus removal also improves a settle ability of the mixed liquor.
 A TREATMENT  PLANT FOR THE FUTURE

      With  consideration for the points raised in the previous section,  the
 problem of nitrogen  (and probably phosphorus as well) must be faced. A  plant
 without primary  sedimentation is favourable in these and many other respects.
 The  only disadvantage may be the problem of sludge handling, which may  be
 circumvented in  another way.

      A plant of  the  future should be flexible. As outlined, simultaneous
 denitrification  is more flexible than pre-denitrification.

      For plants  without primary sedimentation surface aerators are superior
 to diffused  air  systems if there are no other restrictions such as the  need
 for  deep tanks due to limited available area.

      Circulation tanks are ideal for simultaneous denitrification because
 of the sawlike oxygen profile. As parameter for the control of the aeration
 the  nitrate  concentration of the mixed liquor is usefull.

      Additional  mixing tanks ahead of the aeration tank may be built later
 if phosphorus removal is required.

      There are many  other possibilities to plan a treatment plant for the
 future.  In any case  the aeration system has to be selected in accordance
 with the process requirements.
REFERENCES

1. Kayser, R. "Measurements of oxygen transfer in clean water and under
   process conditions". Prog. Wat. Techn. 11. No. 3, 23 (1979).
2. Brown, L.C. and Fisette, G.R. "Non-linear estimation for unsteady-state
   oxygen transfer". Manuscript, Workshop San Diego  (Nov.  1979).
3. ATV-Regelwerk, Arbeitsblatt A131, "Grundsatze fiir die Bemessung von ein-
   stufigen Belebungsanlageri mit AnschluBwerten iiber 10.000 Einwohnergleich-
   werten". Abwassertechn. Vereinigung, Bonn (1981).
4. Lehr- und Handbuch der Abwassertechnik, Band II,  2. Aufl. Verlag  W. Ernst
   Berlin (1975).
5. Lister, A.R. and Boon, A.G. "Aeration in deep tanks: An Evaluation of
   a fine-bubble diffused-air system". Wat.Pollut.Contr. 72, 590  (1973).
                                     154

-------
 6.  Rinke,  G.  and Holier,  U.  "EinfluB von Einblastiefe und Luftdurchsatz auf
    Sauerstoffzufuhr und Sauerstoffertrag bei feinblasiger Breitband-Be-
    luftung".  Forschungsberichte NRW Nr.  1792, Westdeutscher Verlag Koln-
    Opladen (1967).
 7.  Pasveer,  A.  and Sweeris,  S.  "A new development in diffused air aeration".
    JWCF  37,  1267 (1965).
 8.  Emde, v.d.W.  "Criteria for selecting aeration systems". Prog.Wat.Techn^ 11
    No.  3,  201 (1977).
 9.  Barnard,  J.L. "Cut  P and  N without chemicals". Water and Wastes Eng. 11>
    33;  41  (1974).
10.  Pasveer,  A.  "Beitrag iiber Stickstoffbeseitigung aus AbwSssern". Miinchener
    Beitrage  Band 12, S. 197  (1965).
11.  Matsche,  N.  "Removal of nitrogen by simultaneous nitrification-denitri-
    fication  in  an activated  sludge plant with mammoth rotor aeration".
    Prog. Wat. Techn. 8. No.  4/5, 625 (1977).
12.  Usrael, G. "Control of aeration at the treatment plant Vienna-Blumental".
    Prog.Wat.Techn. 8.  No. 6, 245 (1977).
13.  Ermel,  G.  and Kayser,  R.  "Simultane Stickstoffelimination - Beitrage zur
    praktischen  Anwendung auf Klaranlagen". Wasser und Boden 34. In print
    (1982).


 DISCLAIMER

     The work described in this  paper was  not  funded  by the U.S. Environ-
 mental Protection Agency.  The  contents  do not necessarily reflect  the views
 of  the Agency, and no  official  endorsement should be  inferred.
                                      155

-------
IMPROVEMENTS OF AIR DIFFUSION SYSTEMS APPLIED IN THE NETHERLANDS

Johannes H. Popel, Professor
Department of Civil Engineering
Delft University of Technology
2628 CN DELFT, Netherlands

ABSTRACT

     Field tests with water on improved (uniform arrangement of diffusors or
horizontal flow) air diffusion systems are reported and evaluated. Both im-
provements reduce the upward velocity of the water above the diffusors.
Depending on this reduction, 4.5 to 7.7% oxygen absorption per m diffuser
depth are obtained and 2.3 to 4.3 kg O^/kWh oxygenation efficiency. Extreme
efficiencies are mainly caused by low (40%) or high (60%) motor/blower effi-
ciencies.

     Upward velocities are measured in one system. Their influence on oxygen
transfer is modelled and applied to the other systems. Model results allow to
assess whether further improvements of oxygen transfer and efficiency by
changing the system design are possible.
 INTRODUCTION

      In  the Netherlands, roughly 70% of the municipal wastewater is treated
 by  the activated sludge process. 30% of the total treatment capacity is re*-
 presented by extremely low loaded plant (oxidation ditch principle) and the
 majority of the remaining 40% is designed for nitrification by applying low
 sludge loading rates also. The implications with respect to aeration are:
 - relatively low densities of oxygen transfer (g 0~/m3 tank volume.h) are
  required which may cause mixing problems;
 - relatively large amounts of oxygen have to be transferred per unit mass of
  BOD removal. This would mean that improvements of the areation system could
  yield  considerable savings at increasing energy prices.

      Improvements of the air diffusion technology, but also more stringent
 (legal)  requirement with respect to noise and odor problems, have stimulated
 the application of fine bubble aeration again and at the same time reduced
 further  expansion of mechanical aeration.

      The technological improvements are essentially based on the "new deve-
 lopment  in diffused air aeration" as put forward by Pasveer and Sweeris in
 1965  (1): by special means swift upward flow velocities in the aeration tank
above the diffusors have to be prevented because they significantly reduce
the retention time of the bubbles in the water and hence the total interfaci-
al area.  Practical means to achieve this goal are:
- a uniform arrangement of the diffusors all over the tank bottom  (no spiral
  flow);
- creating horizontal streaming velocities in a circular tank by a rotating
  bridge (counter-currant aeration), by propellers mounted within  the tank
  (e.g.  Roto-Flow aeration), or by a vane-column mounted in a circuit-like

                                     156

-------
                 RB
       RB
                          B
AT :  aeration tank
RB :  rotating bridge
FD(S): fixed diffusers (section)
RD :  rotating diffusers
B  :  baffles
P  :  propeller(s)
RSB:  rotating shaft with blades
     (shaft length - tank depth)
                                      \\
                                        \\
                                                 AT
'i
w*
DCD
KOD


w
^•^













































F








L








)J








3





































































































I

































\
V X ">
\ V ^^-O j
\ \ y
\ \ — S

1 • ^^
\ \
\ \ ^
\\
     A:  Counter-Current Aeration with and without Fixed  Diffusers
     B:  Roto-Flow Aeration
     C:  Landox Aeration

Fig.  1:  Horizontal Flow Air Diffusion System
                                 157

-------
  tank (see Figure 1).                                                .  .
- creating a downward flow water velocity, somewhat larger than the rising
  velocity of the bubbles (ICI Deep Shaft).

     In the Netherlands, the first two approaches are used and field  tests
on the oxygen transfer capacity and efficiency have been performed. It  seems
worthwhile to report the results of these tests and to evaluate them  tho-
roughly together. The obtained results are then compared with a theoretical
maximum rate of oxygen transfer. An attempt is made to explain deviation
between both by an average upward streaming velocity of the water in  the
tank, induced by the diffused air. Measurements performed at one of the tes-
ted aeration systems permit to check this approach. Finally, the question is
raised whether further significant improvements of bubble aeration seems
possible.
 CONVENTIONAL DESIGN PRINCIPLES

      The design of conventional air diffusion system with fine bubbles is
 generally based in the Netherlands on the following guide values, validQfor
 standard conditions (normal pressure, air volume at a temperature of 0  C,
 water temperature 10  C, zero dissolved oxygen concentration). Furthermore
 5 Wh  of energy are assumed to be required for 1 m3 of air per m diffusor sub-
 mergence .

 In  "clean" water:
 percent oxygen absorption per m bubble rise:             3 to  4%/m
 specific oxygen absorption in g 02 per m3 of air
 per m bubble rise:                                       9 to  12 g/m .m
 oxygenation efficiency                                 1.8 to 2.4 kg 02/kWh

      For operational conditions an a-factor of 0,7 ... 0,8 is applied for
 fine  bubble aeration. Hence, the following guide values are obtained:

 percent oxygen absorption                                2 to   3%/m
 specific oxygen absorption                               7 to  10 g/m3.m
 oxygenation efficiency                                 1.3 to  2.0kg 02/kWh

      As pointed out earlier, significant increases of these relatively low
 guide values are possible by reducing the upward velocity vector of the
 water. The results of field tests performed in the last years in the Nether-
 lands  can contribute to the experience to what extent improvements are possi-
 ble at present.
OXYGENATION CAPACITY DETERMINATIONS IN WATER UNDER FIELD CONDITIONS

     At a number of activated sludge treatment plants oxygenation capacity
determination have been performed prior to starting operations. Much of  the
results has been published elsewhere, but they never have been evaluated
together.
                                     158

-------
Table 1 Tested Aeration Systems
Nr
tl L, •
1
2
3
4
5
6
7
8

9


10


Plant
JL J- t» a **
Steenwijk
ii
Ams terdam
ii
M
ti
n
u

Nieuw-
Lekkerland

Hulst


Tank dimensions (m)
w *
6.75

4.40

4.40




d *
4.00
same
3.70
same
4.13
same
same
same

1 = V
6.30

33.00

33.00




Counter-Current
diam. 25 and 36
depth 3.20 m
Roto-Flow
diam. 42 m
depth 4.50 m
(mJ)
170

535

600




1750
m

5961


diffusers and arrangement
d
s
3.80

3.35

3.85
4.00
4.00
4.00

3.00


4.00


type
Act. SI. Dome
same
Degremont
domes
same
Brando 1 60
NokiaHKP500
Schuler Fie
Xolith V400
Brando 1 60
45% rotat.
55% fixed
Cellpox


density
4 . 7/m^
2.8/m2
4.0/m2
1.6/m2
1.7/m2
1.4/m2
1.4/m2
, / / 2
1 .4/m

.13/m3

.36/m2


arr .
u
u-cc
u
u-c
u-cc
u
u

u

hf
r
hf
•L


Ref.
2
2
3
3
4
4
4

4

5

6


d = depth of diffusor submergence
s
u = uniform arrangement
u - c(c) = like u, but with some (more intensive) circular motion
h_ = horizontal flow
f
Table 2 Test Specifications
Nr.
1
2
3
4
5
6
7
8
9
10
T-Range
°C
6.0 - 9.0
6.0 - 9.0
16.4 - 19.6
15.8 - 19.5
16.5 - 17.5
17.0 - 17.5
16.5 - 18.0
16.0 - 17.0
19.5 - 21.0
19.0 - 20.0
specific air flows
3.3,
m /m .h
a w
0.95 - 5.73
0.51 - 3.76
1.07 - 5.24
1.10 - 3.00
1.45 - 2.42
1.43 - 2.35
1.44 - 2.17
1.39 - 2.29
0.21 - 0.42
0.11 - 0.37
m3/h.diff .
a
0.81 - 4.88
0.72 - 5.33
1.01 - 4.85
2.56 - 6.95
3.30 - 6.00
4.00 - 6.70
4.20 - 6.30
4.00 - 6.70
5.13 - 10.2
1.34 - 4.40
159

-------
     In all cases the oxygenation capacity has been determined by the unstea-
dy state method after deoxygenation with sodium sulfite and cobalt as cata-
lyst. From the course of the oxygen concentration the mass transfer coeffi-
cienttL.a (h"1, base e) is determined by applying various techniques and
finally the OC by

OC = 0.0113.K..a.V.f .f                                        (kg 02/h)  (1)
             J_t      J.  C

with V  = volume of water aerated (m3)
     f  = temperature correction factor, generally based on exponential model
          with 9 = 1.018 to 1.02
     f  = saturation concentration correction factor = obtained supersatura-
      Q
          tion factor by test
     0.0113 =  standard saturation concentration in kg 02/m3
     1L = "liquid film" coefficient (m/h)
     a  = specific interfacial area = total bubble area A divided by water
          volume V.

     Other transfer rate parameters to be discussed later, can be derived
from the OC or 1C. a and the specific characteristics of the aeration system
and the test.
The oxygenation efficiency (kg 02/kWh) is calculated by dividing the OC (kg
02/h) by  the gross power (kW) drawn from the distribution net.

Tested Aeration Systems

     In total, tests on 10 aeration systems at 4 plants are reported. For
each system between 3 and 14 tests were performed by varying the airflow.
Details of the aeration systems are given in table 1: tank dimensions, dif-
fuser type, density and arrangement and also their depth of submergence (d )
are stated. In total, ds varies between 3.35 m and 4.0 m, the diffuser den-
sity between  1.4 and 4.7 per m2 for normal flow conditions and between only
0.13 and 0.36 per m2 for horizontal flow systems (including the rotating
diffusers).

Test Specification

     The most  important test conditions are specified in table 2 by giving
the range of water temperatures (6 to 21 °C) and information on the air flows
applied. For better comparison among the systems, only specific rates of  air
flows are reported. Firstly, the m3 of air diffused per m3 of water per hour
is stated, ranging from 1.0 to 5.7 m3/m3.h for the normal systems and amoun-
ting to only one tenth of these values for the horizontal flow systems.
Secondly,  the rate per diffuser is given (m3 of air per diffuser per hour),
ranging from 0.7 to 10 m3/h.diff for all systems.

Test Results

     The test results are compiled in table 3. The measured oxygenation capa-
cities   (eq.l) are converted into specific oxygen transfer measures. First-
ly, the range of specific oxygen absorption (g 02 absorbed per m3 of air  per
m diffuser submergence = bubble rise) is given, followed by the percent oxy-

                                     160

-------
S.
3
O

i
     90-
     80
     70-
     60-
     50
     40
E    Q.

 CM
O

01
     30-
     20
      10
       0-
     -10
              nr):test  number

                  (see  tables  1-4)

              nr :percent absorption

                  per m
                       extra points

                       test
                             3          3
                            m  air per m  water per hour
         Fig.  2:  Standardized Results of Tests 1 to 5
                                   161

-------
Table 3 Test Results
Nr.
1
2
3
4
5
6
7
8
9
10
spec, oxygen transfer
g/ma-m
16.5-19.9
13.6-21.7
16.0-23.2
14.3-19.6
14.7-15.4
18.3-20.9
18.2-19.4
17.6-18.5
17.0-20.4
18.3-22.3
%/ro
5.50-6.63
4.53-7.23
5.33-7.73
4.77-6.53
4.90-5.13
6.10-6.97
6.07-6.47
5.87-6.17
5.67-6.80
6.10-7.43
transfer
ef f ic.
kg/kWh
2.7-3.7
2.3-3.4
2.7-3.9
2.4-3.3
2.9-3.1
3.8-4.3
3.5-4.0
3.6-3.8
3.1-3.4
2,0-2.7
energy
consumpt.
Wh/m .m
a
5.31-6.73
5.25-6.84
5.97
5.97
4.69-5.41
4.65-4.89
4.83-5.23
4.60-5.16
4.85-4.96
4.79-5.27
mixing
power
W/m








0.68-0.74
2.78-2.80
sat. factor
fc
-
1.14-1.18
1.11-1.20


1.11-1.15
1.11-1.12
1.13-1.15
1.12-1.15

1.10-1.16
gen absorption per m bubble rise (based on 300 g 02/m3 of air). The following
3 columns refer to transfer efficiency. The transfer efficiencies (kg 02/kWh)
are given, also the specific energy consumption for air diffusion (Wh per m3
of air per m diffuser submergence)  and the additional power dissipation for
mixing for both horizontal flow systems. Supersaturation factors of 1.10 to
1.20 were found.

     In summarizing all test results the following ranges are given:
g 02 per m3 of air per m diffuser submergence
percent oxygen absorption per m diffuser submergence
oxygenation efficiency
mixing power dissipation

Evaluation and Comparison of Results
13.6 - 23.2 g 02/m3.m
4.53 - 7.73%/m
2.0  - 4.3 kg 02/kWh
0.7  - 2.8 W/m3
     For effectively comparing the results of the tests they are evaluated by
the following empirical relation

OC = ko + kj.Qa                                               (kg 02/h)   (2)
     k   empirical constant (kg 02/h)
     ki  empirical constant (kg 02 transferred per m3 of air)
     Q   total air flow (m3/h)
      a                                                             —

     Both constants and the calculated correlation coefficients (0.979 to
0.999) are stated in the first columns oftable 4. This approach does not  take
into consideration any differences of the specific rates of aeration nor  of
different depths of diffuser submergence. Therefore, equation 2 is divided by
the water volume V (m3) and the diffuser depth d , to obtain the specific
                                     162

-------
 S_
 OJ
 Q.

 i.
 
       10-
       5-
               nr
                test number

                (see tables  1-4)

                percent  absorption

                per m
                      0.5
                                 1.0
1.5
2.0
                              3          3
                             m  air per m  water per hour
          Fig. 3: Standardized Results of Tests 5 to 10
                                      163

-------
Table 4 Standardized Evaluation
of Tests



Nr.

1
2
3
4
5
6
7
8
9
10
k
o
kg 02
h

2.33
2.67
-14.85
-14.98
2.39
22.56
14.17
8.10
6.56
16.33
*l

kg 02
m3 air

0.0613
0.0469
0.0812
0.0745
0.0555
0.0567
0.0632
0.0646
0.0429
0.0668
r

corr.
coeff .

0.999
0.998
0.998
0.997
0.995
0.993
0.979
0.998
0.999
0.999
K K!
o
g 02
m3.h.m
w
3.61
4.13
-8.29
-8.36
1.03
9.40
5.90
3.38
1.25
0.68
g 02
m3.m
a
16.1
12.3
24.2
22.3
14.4
14.2
15.8
16.2
14.3
16.7
          m  air per m  water per hour
Fig.  4:  Basic Interpretation of Transfer Regression Line
                           164

-------
oxygenation capacity SOC  (g 02transferred per m3  of water  per hour  for  one  m
of bubble rise):
     1000.k     iOOO.k! Q       g 02

SOC " -V^ + —	-f  3   ,   „                                     (3)
          s       s        m° water.h.m
which  is rewritten as

SOC = K  + K!.q   	§_^	                                            (4)
— JL " T.
1000
V.d
1000
d
s
m3 water. h.
•ko f * °2
V.
s m3 water.
•kl g 02
( \
m3 air.m
m
ciupj.Li.c3j. consuanc
h.m
= empirical constant
      q  = specific rate of aeration  (m3 air/m3 water.h)
       3.

     The standardized constants K  and K^ are also given in table 4. The cor-
relation coefficients hold for this relation also. The results are    gra-
phically represented in figures 2  (test  1 to 5) and 3 (test 5 to 10). The
graphs contain also lines of  constant percent oxygen absorption per m diffu-
ser depth, which allows to easily  judge  the performance of the aeration sys-
tem.

Discussion of Regression Equation

     The intersection K  of equation 4, being mathematically independent of
the specific air flow q , varies over a  considerably large range (-8,4 to +
9.4). It is not possible to interprete K  as the contribution of the tank
water surface to oxygen transfer,  because (a) this would include negative
contributions (tests 3 and 4), (b) some  contributions (55 to 155 g 02/m2.h)
are unrealistically high (tests 1, 2, 6, 7, 8). Only test 5 and both hori-
zontal flow systems (9 and 10) would give reasonable values ranging from 11.8
to  16.5 g 02/m2.h.

     A more reasonable explanation of high positive values of K  can be given
via a relation between the SOC and the specific air flow q  over a wider
range than applied in the reported tests. This assumed relation is represen-
ted in figure 4A. At very low air  flows  the produced bubbles are very small,
upward velocities of the water almost negligible. As a consequence of both
effects, the interfacial bubble area is  relatively large and so is the per-
cent oxygen absorption. At increasing specific air flow q  both effects
become less favorable and the percent oxygen absorption is decreasing. After
this transition the range of  a straight  line relationship between SOC and q
is reached. Exceeding this range a smaller increase of the SOC is conceived
of than predicted by the straight  line.

     Within this reasoning a  third effect has not been taken into account,
namely the decrease of oxygen in the air bubbles during transfer, its effect
on the saturation concentration and hence on the rate of transfer. This ef-
fect is obviously greatest at high percent oxygen absorptions, i.e.  at low
specific air flows q . If this effect is of significance, then an increase
                    3.

                                     165

-------
of q  will increase the mass of oxygen transferred per m3 of air  (per m) .But
thisalatter increase will be such, that the percent oxygen absorption is de-
creasing. Thereby the oxygen content in the air bubbles is higher than at
lower rates of airflow and so are the saturation concentration and the rate
of transfer. If, with increasing q , the percent oxygen absorption remains
(almost) constant (K  relatively small) or increases even (KQ negative),_
then effects other tfian oxygen depletion in the diffused air are responsible.
Since the bubble size increases with increasing q , the only explanation for
a negative K  can be an assumption about the mixing conditions becoming more
favorable as°the air flow increases (see figure 4B). Such changes cannot
easily be conceived of and therefore also this does not explain a negative
K -value satisfactorily.

     Interestingly enough, 4 runs with the aeration system 3 (K  = - 8.31)
gave at some 2.2 m3 air per m3 water per hour exceptionally high SOC-values
of some 50 g 02/m3 water per hour per m bubble rise (7.6%/m). These 4 mea-
surements (see "extra points test 3") do not belong to the regression popu-
lation of line 3 in the statistical interpretation.

     The slope of the regression lines K^ varies from 12.3 to 24.2 g 02/m
air per m. Converted to percent oxygen absorption this would mean 4.1 to 8.1%
per m. But these values have to be interpreted together with the intersection
K  . The actual range of specific oxygen transfer can be seen in table 3 and
in figures 2 and 3 and goes from 13.6 to 23.2 g 02/m3.m and from 4.5 to 7.7%.

     There is a tendency that large values of the slope go along with low
values for the intersection (negative correlation). For all 10 test a signi-
ficant (P > 99%) correlation coefficient (r = - 0.862) is calculated,(mainly
due to large positive and negative values of K . Omitting the negative K ,
the correlation (r = - 0,209) remains negative, but becomes insignificant.

Oxygenation Efficiency

     The oxygenation efficiency OE (kg 02/kWh) of the aeration systems can be
calculated by dividing the following 2 parameters: (a) the specific oxygen
transfer rate (g 02 per m3 air per m diffuser depth) over (b) the specific
energy consumption in Wh (= 3.6 kJ) per m3 of diffused air per m diffuser
depth. The first factor varies over a considerable range (see table 3), whe-
reas the second does not change very much. Measured values for root blowers
(test 5 to 10) show a range of only 4.5 to 5.4 Wh/m3.m. The value of 6.0 of
tests 3 and 4 has been assumed from earlier measurements. The aeration sys-
tems 1 and 2 had a centrifugal blower showing low efficiency and correspon-
dingly high specific energy consumption (6.8 Wh/m3.m) when throttled to ob-
tain low air flows.  Without throttling a value of 5.3 Wh/m3.m is obtained.

     For horizontal flow systems the energy used by the seperate water pro-
pulsion device has to be added to the above estimate. The specific power dis-
sipation (W/m3)  is a reliable incidation of this element of energy comsump-
tion.  Since it is (almost) independent of the air flow, its share is rela-
tively high at low air flows and oxygen transfer rates (and vice versa).
System 9  dissipates only 0.7 W/m3 and shows OE-values above 3 kg 02/kWh,
therefore;system 10,  however, dissipating 2.8 W/m3 with the propellers, does

                                    166

-------
not exceed 2.7 kg 02/kWh.

     The above considerations allow to explain extremely low OE values of the
tests as follows:
-the lower values of tests 2 and 4 are caused by low transfer rates and rela-
 tive high specific energy consumption;
-the lower values of test 1 and 3 can be explained by high specific energy
 consumption;
-the low values of test  10 are due to the high power dissipation for mixing.

     Avoiding above drawbacks it seems possible to reach at least 3 kg 02/kWh
in water and 2.2 kg 02/kWh under operational conditions for systems with
small upward velocities. Maximum values of 4 kg 02/kWh in water and of 3 kg
02/kWh under operational conditions appear realistic.
INTERPRETATION OF RESULTS

     The uniform arrangement of diffusers in relatively deep tanks (4  m) to
prevent (double) spiral flow and the development of horizontal flow systems
aim at reducing vertical streaming velocities above the diffusers and provide
at the same  time enough mixing intensity to keep the activated sludge solids
in suspension. In this section an attempt is made to estimate on theoretical
grounds to what, extent such a reduction is accomplished in the tested
aeration systems. Measurements of the vertical velocity component above the
diffusers in one of the systems, knowledge of the bubble size of that system
and a simple estimate of the effect of oxygen decrease in the diffused air
allow to check the theoretical approach.

Theoretical  Background

     The approach is based on the "penetration theory" proposed by Higbie (7)
and worked out for bubble aeration by Pasveer (8) . According to this theory
the "liquid  film coefficient" or better liquid transfer coefficient 1C is
given by

        = 2-l                       m/s                                (5)
     D  = diffusion  coefficient of oxygen  in water
          at  10 °C   D =  1.39.10~9 m2/s = 5.00.10~6 m2/h
     v  = "slip" velocity  of bubbles  in water
      s
     d  = diameter of "spherical" bubbles
      B
     The slip velocity of  bubbles of  2 to  7 mm diameter  in water  is  constant
 (9)  amounting to v   =0.23 m/s and is independent of  temperature  (10).  In
 filtered or pure water there is a decrease  with increasing bubble  size within
 the  above range  (10):
v
 s
            /3.64.a                       /                                 ....
          y/— -; -                     m/s                                (6)
          VP'dB
         = surface tension of water  at  10  C   a  =  0.0742  N/m
                                     167

-------
     p  = density of water   _
     v  (at 10 °C) = 0.0164//d   m/s
      S                       o
     The liquid transfer coefficient can therefore be estimated for water
when the bubble size is known. For the Nokia-diffuser (test 7) the bubble
size was (11)
                                                                         *
     dR = 0.080.qd + 2.9 mm                                              <7)
     q, = air flow m3/h.diffuser

     This information allows to calculate the average liquid transfer  coeffi-
cient as a function of the air flow q , .

     In order to apply equation 1 for theoretical estimation of the OC,  the
specific interfacial area of the bubbles a = A/V has to be known. The  total
area depends on (1) the total air flow Q  (m3/h), (2) the bubble size  (dia-
meter of spherical bubbles) , (3) the retlntion time of the produced bubbles
in the water T  which equals the diffuser depth d  divided by the bubble rise
velocity in the tank v  = v  + v .
For spherical bubbles we gel
         6.Q .T,,    6.Q .d        6.Q .d
     A =    a  B       a  s          a  s
            ,,      d_.(v  + v )  d_.f .v
            B       B   s    vBvs
     f = bubble velocity factor
       = rising velocity of bubbles above diffuser
         slip velocity of bubbles in quiescent water
       = 2 to 3 in spiral flow tanks (8)

     Combining equations 1, 5, and 8 yields with a = A/V

     OC- 0.1356- i/ - 5_.Q  .d - 1  ° kg 02/h                            (9)
                  I /     j J  SL  Q  2.
                  l/TT.V .d          V
                  *    s  B
     The oxygen supplied to the system by the total air flow Q  amounts  to
0.3.Q kg 02/h. The fraction of oxygen absorbed per m of diffusor depth p'
is then obtained by dividing equation 9 by 0.3.Q .d  to give
                         , f .f                  as
-, / - D
 l/7r.v.
     p\ = 0.452-,  - —.---        n                                  (10)

                            v
     Inserting a constant v  =0.23 m/s and the stated numerical values  into
equation 10 gives for d_ in mm
          0.627. f .f   B
     p. -- _£_£                  m                                  (]1)
          W (mm)
     Assuming filtered or pure water conditions, equation  6  is  inserted  into
10 to give with the indicated numerical values for dB in mm
          ,  ,1.25,  .
          fv.dB   (mm)
                                       m
                                     168

-------
     Multiplication of  equations 11  and 12 with 100 would give  the  percent
oxygen absorption per m diffuser depth pi and p2. Maximum values  of p at 10
 C (f™ = 1) are obtained for conditions of quiescent water  (f   =  1). pj  is
plotted for a typical oversaturation factor of 1.15 for bubble  sizes of  2 to
7 mm in figure 5. For this  range, percent oxygen absorptions decreasing  from
25%/m  to 4%/m are calculated.

     The greater percentage absorption, the more the transfer limiting influ-
ence of oxygen depletion in the diffused air becomes significant. The above
theoretical approach by equations 11 and 12 should be corrected for this in-
fluence, therefore. It  can  be shown - along similar reasoning as  presented in
(12) - that the correction  factor for oxygen depletion f  can be estimated by
                  1
           .   0.531.d
           1 +        s   ,
              —3=	.p1
(13)
     The corrected  absorption p1 can then be estimated by
                  234567
                                  bubble diameter (mm)

                   Fig.  5: Maximum Percent Oxygen Absorption from Spherical Bubbles
                         at f = 1.15 in Quiescent Water
                                      169

-------
     "i.c-V!               'J,= -V»i        ""                    04)
     PljC-values  (%/m) are plotted for f  =1.15 and dg = 4 m in figure  5. A
 comparison with the not  corrected pj-values shows the significant influence
 of  oxygen depletion in the diffused air, especially at high absorption,  i.e.
 with small bubbles. The corrected percent absorptions per m decrease  in the
 cited  bubble  range from  17 (was 25) to 3,3  (was 4) %/m.

 Verification  of Theoretical Approach

     For  test 7 not only the bubble size was known as a function of  the  dif-
 fuser  loading (m3 air/h.diff .) as given by  equation 7, but also the  upward ve-
 locity v  of the water half-way above the diffusers was measured as a function
 of  the specific air flow q   (m3 air/m3 water per hour). The relation was  (11)
 generalized (r =  0.929)  by&

     v = 0.121. iTq" -  0.01           m/s                                (15)
      Hence  all  information required to theoretically calculate the oxygen
 transfer (expressed  as percent absorption per m diffuser depth) is available.
 In this  approach,  the average bubble size, the mixing conditions in the tank
 as affecting  the bubble size velocity and hence its retention time are taken
 into account  and also the influence of oxygen uptake from the diffused air
 on the rate of  transfer.

      The pertinent calculations are summarized in table 5 for 6 aeration test
 runs (system 7)  of which 3 were performed in duplicate. Only averages of every
 duplicate are reported here. In the first 5 rows the experimental data are
 given. The  4 rows thereafter refer to velocities. A constant slip velocity of
 v   =0.23 m/s is assumed. The measured upward water velocity and the bubble
 rise velocity are stated together with the bubble velocity factor f . Where-
 as  f -values of 2 to 3 are reported for spiral flow tanks (8) , the uniform
 arrangement of  diffusers brings this factor down to 1.6 to 1.7. In the last
 4  rows the  oxygen transfer calculations are given. Firstly, the percent oxy-
 gen absorption  per m p\ is calculated taking into account only the measured
 upward velocities, not oxygen decrease in the diffused air. As expected, this
 prediction  in much higher (9 to 19%) than the measured values. The oxygen
 correction  factors f  are 0.88 on the average. Finally, the theoretical pre-
 dictions differ from the measured values by not more than - 3.5 to + 4.1%.

Application of Theoretical Approach

     The small deviations between measured and predicted values encourage to
apply the theoretical approach also to the other tests. Unfortunately, how-
ever, neither bubble size (relations)  nor upward velocities are known. Hence,
this approach allows only to estimate the quantity f .d*»5 with reasonable
accuracy along the following lines.                 v

     The measured oxygen absorption per m is assumed to be equal to the cor-
rected theoretical absorption value pijC. The uncorrected theoretical absorp-
tion p\ is then calculated (inversion of eq. 13 and 14) by

                                     170

-------
        Table  5    Theoretical Transfer Calculations
        Parameter
                              Dimension
Average value of run
        Data
        diff.depth d
                               m3/m3.h
                                a  w
                               m3/h.diff
                                3.
        spec,  air flow q
        spec,  air flow q,
        sat.  factor f
        perc.  oxyg. abs, p
        bubble diam. (eq.7)
        Velocities
        slip velocity v
        upw.  veloc.v  (eq.15)
        bubble vel.v =v +v
                    r  s  v
        vel.factor f =v /v
                    v  r  s
        Transfer calculat.
        pi (eq. 1 Ouncorr.
        corr. fact.f (eq.13)
        Pl,c(e
-------
- the upward velocity factor f  for assumed bubble diameters of 3 and 3.5 mm.
For comparison, table 6 contains also the upper parameters calculated for the
conservative guide values (pi,c) of 3 and 4%/m at a diffusor depth of 4,0 m
and a supersaturation factor of 1.15.

     For the ease of further discussion the above product is named "effecti-
vity number" with the notation

     E = f .d*'5(mm)                   mm1'5                            (18>
          V  D

     A large E-number is obtained by either a large upward velocity (factor)
and/or a large bubble diameter. Both conditions hold for a poor aeration
system. Large E-numbers, therefore, indicate an inferior, small E-numbers a
superior aeration system. The E-number (in contrast to other transfer para-
meters) allows to compare the most diverse systems of air diffusion by only
one parameter. Diversity of the following parameters is included in the E-
number
- effective bubble size (definition eq. 18);
- upward velocity (factor)(definition eq. 18);
- diffusor depth (eq. 16);
- the influence of diffusor depth on the depletion of oxygen in the diffused
  air (eq. 13 and 16);
- rate of air flow and its influences on upward velocity (eq. 8) and oxygen
  depletion (used in deriving eq. 13 and 16);
- supersaturation factor of the aeration system (many equations);
- temperature correction factor.

     For effectively comparing the results of table 6, the obtained E-numbers
are classified as follows

E > 15   conservative guide values
E > 12   tests 2 and 5
E > 10   tests 1, 4, 8, and 9
E >  8   tests 3, 6, 7, and 10

     Some class differences can be explained directly by differences in the
upward velocity (factor). Conventional spiral flow designs have E-values
above 15. Both systems with appreciable circular motion  (2 and 5) have E-va-
lues above 12.

     Due to the lack of information on upward flow velocities and bubble  size,
differences   between systems "E > 10" and "E > 8" cannot be explained rigo-
rously. However, some comparative comments can be made.  Comparing tests 3
and 4 (same diffusors, same tank), the better performance  (smaller E-value)
of test 3 is evident from (a) more domes per m ,  (b) better prevention of
circular motion by more uniform arrangement of the diffusors,  (c) lower
(60 % of test 4) average diffusor loadings. Similarly, differences of both
horizontal flow systems (9 and  10) can be based on the large difference of
diffusor loadings. Systems 6, 7, and 8 differ only in the  type of diffusor
and E-differences must be assigned to bubble size differences.

     With respect to the upward velocity factors  (in table 6),  it should  be

                                     172

-------
Table 6 Product f .d*-5 and Bubble
Velocity ¥actor fn
Nr.

1
2
3
4
5
6
7
8
9
10
3%
4%
f .dl.5
v B

11.42
13.81
8.96
11.30
13.03
9.58
9.94
10.59
10.84
9.75
22.70
16.69
upw.vel.fact.f for
d =3. Omm
2.20
2.66
1.72
2.17
2.51
1.84
1.91
2.04
2.09
1.88
4.37
3.21
d = 3 . 5mm
1.74
2.11
.37
.72
.99
.46
.52
.62
.66
.49
3.47
2.55
mentioned, that the effectivity number is  (very) sensitive to  (small) bubble
diameters. If this factor is not determined from velocity measurements, but
calculated from bubble diameters, this information  (e.g. from  diffusor manu-
facturers) should be very accurate. For the conditions of test 6, f -values
from  1.59 to  1.73 (table 5) have been determined from velocity measurements.
Table 6 reveals f -values for an assumed bubble size of 3 mm between 1.8 and
2.7 and for d  = 3.5 mm between 1.5 and 2.1. The conservative  guide values
lead to f -values from 2.6 to 4.4 under above assumption, somewhat higher
than reported elsewhere (8).

     Measuring or estimating the upward velocity (factor) is of great practi-
cal importance. A large factor (e.g. f  >  2) indicates that improvements of
oxygen transfer are possible by more uniform arrangement of the diffusors, by
changing the diffusor density and the distance from each other, and by chan-
ging the tank geometry. A low value (e.g.  f  = 1.5) would indicate that such
measures would not have very much effect,  since a certain upward velocity
cannot be avoided, obviously. Information  on only oxygen transfer parameters
is not sufficient in this context, because it does not include diffusor depth
effect and oxygen depletion.  Maximum theoretical transfer values can be esti-
mated (see fig. 5) for comparison, in that case.

     The application of the effectivity number E to the tests  has shown that
it is a useful parameter to compare very diverse air diffusion systems opera-
ted under different conditions by one quantity only. The error of the appro-
ach is estimated to be less than 5%.  Disadvantages of E are seen in that it
is not dimensionless and that it decreases as the performance  of the aeration
sys tern improves.
                                    173

-------
SUMMARY AND CONCLUSIONS

     A number of full scale aeration tests has been performed in the Nether-
lands on diffused air systems which try to minimize the upward velocity v^
of the water above the diffusors by either uniform arrangement of the dif-
fusors on the tank bottom or inducing horizontal flow by a seperate water
propulsion device. The very diverse test conditions and results are reported
and evaluated.
                                                                            t
     The evaluation indicates that conservative spiral flow guide values of
3 to 4% oxygen absorption per m diffusor depth and 1.8 to 2.4 kg 02/kWh
oxygenation efficiency for water (a = 1) can be improved by these systems
to 4 to 7%/m and 3 to 4 kg 02/kWh (based on 5 Wh/m3 air.m diffusor depth for
air diffusion and 1 W/m3 for horizontal flow generation.

     Upward velocities v  were measured in one system (0.13 to 0.17 m/s) .
Together with an assumedVconstant slip velocity of small (2 to 6 mm) bubbles
in water of v  = 0,23 m/s, this gives bubble rise velocities in the tank
(v  = v  + v f of 0.36 to 0.40 m/s and "bubble velocity factors" (fy = vr/vg)
ofr1.6 ?o  1.7\ A model is developed to quantify the effect of f  on the rate
of oxygen  transfer, which predicts the measured values by more than 5% accu-
racy. It includes the effect of bubble size, diffusor depth, oxygen decrease
in diffused air during aeration.

     Since in the other aeration systems bubble size and upward velocity
 (factor) have (as usually) not been measured, both are combined in  an "effec-
 tivity number" E " f  -dl'5 which can be computed from transfer data by the
model. The E-number Xllows to differentiate and compare diverse diffused air
systems. Conventional spiral flow systems yield E =  16 to 23, whereas the
 system tests give E = 9 to 14. For an assumed bubble size of 3.5 mm this
would mean bubble velocity factors f  = 2.6 to 3.5 for spiral flow  and  1.5
to 2.1 for  the tested systems. E-number and velocity factors indicate whether
further improvement of the rate of oxygen transfer (reduction of v  and d )
is possible by design. Design options in this respect are (a) moreVuniform
arrangement of diffusors, (b) higher diffusor density (elements per m2), (c)
lower diffusor loadings, (d) change of tank geometry to further reduce spiral
movement. Maximum attainable (f  = 1; v  =0) are calculated as hypothetical
goal (fig.  5). Minimum values of 1.5 (vV =0.12 m/s) seem possible.

     Above  improvements increase at the same time the oxygenation efficiency
(kg 02/kWh) to the values proposed above. A further significant increase of
OE is limited by the low efficiency of diffusing air into water ranging from
only 40% (system 1 at 6.8 Wh/m3.m) to 60% (system 6 at 4.5 Wh/m3.m).
NOTATION

A    m2             interfacial area between bubbles  and water
a    m2/m3          A per m3of water aerated;specific interfacial  area
d    m              diffusor depth = bubble rise
d    m, (mm)        diameter of spherical bubble
E    mm1'5          effectivity number

                                     174

-------
kv   kg 02/h
K°   g 02/m3.h.m
kl   kg 02/m3 air
Kl   g 02/m3 air.m
1L   m/h
OC
OE
V
v
v
v"

P
a
T,.
kg 02/h
kg 02/kWh
m^ air/h
m3/m3.h
 m3A.diff.
%/m
     m
       'I
SOC  g 02/m3.h.m
           Wf
 3
m/s
m/s
m/s

kg/m3
N/m
s, h
oversaturation factor
oxygen depletion factor
temperature factor
upward velocity factor
empirical constant
empirical constant
empirical constant
empirical constant
liquid mass transfer coefficient, also "liquid film
coefficient"
oxygenation capacity
oxygenation efficiency
total air flow
specific air flow (m3 air/m3 water per hour)
specific air flow (m3 air/h per diffusor)
percent oxygen absorption per m bubble rise
or per m diffusor depth
fraction of oxygen absorbed per m bubble rise
or per m diffusor depth
for subscripts see context
specific oxygenation capacity: g C>2 transferred per
water per h per m diffusor depth or bubble rise
volume of water under aeration
(geometrical) rising velocity of bubble (in tank)
bubble slip velocity in water
upward velocity of water above diffusors

density of water
surface tension of water in contact with air
retention time of bubbles in water
REFERENCES

1. Pasveer, A., Sweeris, S. "A New Development in Diffused Air Aeration".
   Jour. Water Poll. Control Fed. 37, 1267 (1965).
2. Ilsink, M.W., Brandse, F.A. "Fine Bubble Diffused Air Systems (Domes)  at
   the Sewage Treatment Plant Steenwijk" (in Dutch). H20 11, 107 (1978).
3. "Research into the Oxygenation Capacity of Domes at Various Densities  and
   Various Dome-Loadings" (in Dutch). Report No. 80.039 of the Governmental
   Institute for Sewage and Wastewater Treatment (1980).
4. "Measurement of the Oxygenation Capacity of Air Diffusors of Various
   Manufacturers" (in Dutch). Report No. 77.002 of the Governmental Institute
   for Sewage and Wastewater Treatment (1977).
5. Blom, P.H.M. "Biological Treatment According to the Counter Current Aera-
   tion Process" (in Dutch). H20 II, 142 (1978).
6. Diekema, K., Verbraaken, H.A.M. "Results of Tests on a Sewage Treatment
   Plant According to the Counter Current Aeration Process (Rotoflow) at
   Hulst" (in Dutch). H20 13, 136 (1980).
7. Higbie, R. "The Rate of Absorption of a Pure Gas into a Still Liquid
   During Short Periods of Exposure". American Institute of Chemical Engi-
   neers. 365 (1935).
                                     175

-------
8. Pasveer, A. "Research on Activated Sludge:  VI.  Oxygenation of Water with
   Air Bubbles". Sewage and Industrial Wastes  27,  1130 (1955).
9. King, H.R. "Mechanics of Oxygen Absorption  in Spiral Flow Aeration Tanks
   I: Derivation of Formulas".  Sewage and Industrial Wastes 27,  894 (1955).
lO.Horler, A. "Rising Velocity  and Rising Resistance of Air Bubbles in Pure
   Water" (in German). Gas- und Wasserfach 105,  764 (1964).
11.De Kort, K.F. "Aspects of Bubble Aeration"  (in Dutch).  H20 7, 162 (1977).
12.Ewing, L. et al. "Testing and Data Analysis of  Diffused Aeration Equip-
   ment". Jour. Water Poll. Control Fed.  51, 2384  (1979).


DISCLAIMER

     The work described in this paper was not  funded by the U.S. Environ-
mental Protection Agency.  The  contents do not necessarily reflect  the views
of the Agency, and no official  endorsement should  be inferred.
                                    176

-------
 BLOWER DESIGN CONSIDERATIONS

 G. G. Powell, P. Eng., Vice President
 Gore & Storrie Limited Consulting Engineers
 Toronto, Ontario  M4G 3C2

 ABSTRACT

      A major design consideration in diffused aeration is the selection of
 the air blowers.  Over the design life of a treatment facility the organic
 loading to the aeration tanks will vary significantly creating both lower and
 higher oxygen demands.  The aeration system which can efficiently meet this
 varying air requirement over the design period of the plant will realize to
 the owner a considerable cost saving.

      To properly select an efficient blower system the designer must
 appreciate not only the organic loading condition but also the performance
 characteristics of the blower used, its relationship to other blowers when
 running in parallel and its relationship to the air system.


 INTRODUCTION

      For  secondary treatment of wastewater, activated sludge is one of the
 most  widely used and effective processes.   Its success,  however,  depends on
 an adequate and continuous supply of oxygen to the micro-organisms grown and
 maintained in the system.   Diffused aeration is one of the most proven and
 reliable  methods for transfering oxygen into the aeration tanks.   This paper
 addresses blower design considerations specific to the diffused aeration
 process.
 ORGANIC LOADING

     It is necessary  to  consider  in  design  the range of organic  loading
 conditions that  the treatment  system will experience;  not just  at  the design
 condition but at  commissioning, during weekend low strength periods, weekday
 peak periods in  other words  throughout the  plants design  life.

     If for example,  we  consider  normal  sewage, one not diluted  by  abnormal
 contributions of  infiltration  and inflow, we can expect a sewage strength of
 perhaps 200 mg/L  of BOD,.  As  flow increases throughout the day
 variations in organic loading  are experienced.  Figure 1  shows a typical
 pattern of loadings to the aeration  tanks after primary settling.

     In this example  you will  note that  the organic loading as measured in
 kg/day BOD. varied from  1500 to 7000, a  ratio of approximately 1 to 5.

     Throughout the design life of the plant the organic  loading will
normally increase due to population  or industrial growth  as shown in
Figure 2.

                                     177

-------
                             conra mm TOLLPTIOH COBTIOL nun
                          I VS. nOH(Q) i LOADIHG (BOD'S I SUSPENDED SOLIDS)
                                  nnuiY imuiifi'
  40
     MOO



     7000


     MOO


    ^ 5000
    a
 13-
  5 •
•^20.8*000
    1000,
=^=s,iffi.«w
    • MD'S
                                                          g  g  g  g  k  g  S  g
                                                          sassssss
                               IDC (bour«l
        Figure 1.   Flow and  Load Variations at  a Treatment Plant
            Z500O
            20OOO
            toco
           . K30OO-
          UJ
          ec
                              HIGH
                                 NORM
                                             LOW-
                                  TIME  - YEARS


                 Figure  2.   Air  Requirements Versus  Time
                                      178

-------
      Understanding the range of organic loading condition is therefore an
  important  first  step in the blower selection process.
OXYGEN REQUIREMENTS

     The amount  of oxygen  required  to  satisfy  the  aeration tank
micro-ogranisms  is dependent  upon the  type  and strength of the waste and the
activated sludge solids  retention time necessary to reduce the waste to the
effluent criteria.

     Stengel and Shell(l)  have  proposed  the following equation
             SET
                   [A + Ab  (SRT) + a B  (SRT)]
where     r     =   oxygen required  Ibs  per  day
          A     =   oxygen required  for  BOD  (substrate) synthesis,
                   Ibs  02/lbs BOD,
          B     =   oxygen required  for  endogenous  respiration,
                   Ibs  02/lbs  solids
          Q    =   wasterwater flow,  gal/day
        BOD    =   biochemical oxygen demand  at  5  days and
                   20°C mg/L
        SRT    =   sludge  retention time  days
          a    =   cell yield  co-efficient  Ibs cells /Ibs
                   substrate
          b    =   micro-organisms endogenous decay co-efficient
                   per  day


     Of note in the  above equation are the following:

     1.   As the activated sludge (substrate) removal rate increases so does
          the oxygen requirement.

     2.   As the activated sludge solids retention time  (SRT) increases so
          does the oxygen requirement.

     3.   As activated sludge synthesis  rate yield increases so does the
          oxygen requirement.

     An understanding  of  the  industrial  mix  of  a  community is important to
this particular item.   If there  is significant  high  strength, solubilized
wastes expect BOD  spikes  and  higher  oxygen requirements.

     Treatability  studies and knowledge  of other  industries with similar
wastes are most helpful in determining oxygen requirements.
                                     179

-------
 OXYGEN TRANSFER EFFICIENCY

      To arrive at  the  air supply  required  the oxygen  transfer  efficiency  of
 the proposed  aeration  equipment must  be determined.   Normally  this  is  done in
 the laboratory or  manufacturers plant by actual oxygen  transfer  testing under
 standard controlled  conditions.   The  usual test procedure  is the clean water
 unsteady state method.   The  efficiency is  stated as Ib  oxygen  transferred per
 hp-hr.

      The oxygen transfer efficiency arrived at under  the test  condition is
 then converted to  field  conditions using the following  equation:
                1                               CT~20
               E   =  alpha  E  (Beta Csat - DO) Q
                                   Csat


 where     E   =  mixed liquor  oxygen  transfer efficiency
          E   =  clean water oxygen transfer efficiency
        alpha  =  the oxygen  transfer  correction factor  for the
                 particular  waste usually  0.75 to 0.95

         beta  =  the correction factor for salinity-surface tension
                 usually 0.95  to  1.00

         Csat  =  oxygen  saturation in tank usually taken at mid  depth

           DO  =  minimum dissolved oxygen  in mixed liquor
                 usually taken as 2 mg/L

           Q  =  temperature correction constant, 1.024

           T  =  temperature of mixed liquor C°


      Important factors affecting  the  rate  of oxygen transfer are:

       1.   Oxygen concentration gradient of the wastewater  into and  out of the
           aeration system.
       2.   Temperature of the wastewater.
       3.   Turbulence in  and  around the air/liquid interface.
       4.   Wastewater characteristics.
       5.   Depth of wastewater.
       6.   Contact  time of the  air bubble and wastewater.
       7.   Size of air bubble.
       8.   Rate of air flow.
       9.  Type of air diffuser.
     10.   Position of the air  diffuser.
     11.  Aeration tank  geometry.

     Experience with the actual diffuser device and discussion with a
knowledgeable and reputable  manufacturer are essential  at  this stage as shown
in Figure 3.


                                      180

-------
00
                    60

                    55
•g  50
                u
                    45
 °:  40

£  35
U.
c/>  so
t£   25
H-
Z   20
UJ
O   15
>

o   10

      5

      0
                                                                                     WPRL Report 365R
                                                                                      Clean Water _
                                                                                     Row Spacing 30
                                                                                               ftT
                                                                                                         a = .75
                                                                                                         B - .95
                                                                                                       D.O. » 2.0

                                                                                                 -D.O.)  CSAT = Mid
 »r^u Report 365R
Water with 5ppm Dgtergent
  ROW Spacing 30^
  iL_»-— ~~*
                                                                                                           WPRL Report 390R
                                                                                                             Mixed Liquor
                                                                                                           Row Spacing 30°
                                   Actual  Plant
                                   Operating Data (x)
                                                      Plant Operating
                                                      Data Minimum Results
                           10   11  12   13  14   15   16   17   18   19  20   21   22  23   24  25  26   27   28  29   30

                                                      DIFFUSER SUBMERGENCE  (Feet)
                                                 Figure  3.   Aeration  Device Information

-------
     All too often claims of high oxygen transfer efficiencies have  not been
realized in the field due to the presumptuous claims by aeration equipment
manufacturers or designer naivety.

     In considering a manufacturer's oxygen transfer efficiency:

     1.   Request test data on the diffuser.

     2.   Ensure testing is carried out by an independent laboratory
          specializing in this type of testing.

     3.   Examine carefully the sampling procedures, the method of testing,
          the test tank geometry and the power-to-volume ratio.

     4.   Find out how the diffuser has performed in actual sewage treatment
          applications similar to yours'.


AIR REQUIREMENTS

     Once the designer has arrived at the field oxygen transfer efficiency
the air required can be determined as follows:

Standard  =  	oxygen required in Ibs/day	
Cubic        1440 min   0.075 Ibs   0.232 x oxygen transfer
Feet of         day      f 3 .       %       efficiency under
Air Per                             oxygen  field conditions
Minute                              in air
                                    by weigh
     In the above equation only the oxygen required per day and the diffuser
oxygen transfer efficiency under field conditions are true variables.

     Standard conditions are stated as 14.7 PSIA, 68°F and 36 percent
R.H. (relative humidity).  These conditions are rarely if ever the actual
conditions at site and standard cubic feet per minute (SCFM) must be changed
to the specific site condition.

     Corrections to temperature, altitude, and relative humidity are made to
convert SCFM to actual, or inlet, flow rate, ACFM at the air blower inlet
conditions.   The formulae used is a combination of Charles and Boyle's law
corrected for relative humidity (RH).

            VI  -  V2 x Tl x Ps - (RHs x PVs)
                        Ts   PI - (RH1 x PV1)

          The above formulae can be simplified to:

          ACFM  -  SCFM x 0.0276 pl J
                                     182

-------
where Vj  =  actual or  inlet  cubic  feet  per minute
             ACFM at  blower inlet

      \2  =  standard cubic feet per  minute at
             blower inlet  SCFM

      T.  =  Blower inlet  temperature in
             degrees  Rankine

     Ts  = Standard temperature in  deg.  R.
           (68°F + 460  - 528°R)

     Pj  = Blower inlet pressure in psia

     Ps  = Standard pressure  in psia  (14.7)

     RH  = Relative humidity  at blower inlet in percent

     RHs = Standard relative  humidity in percent (.36)

     PV. = Vapour pressure of water at T in psia

     PVs = Vapour pressure of water at Ts in psia


POWER REQUIREMENTS

     To determine power requirements  the following formulae applies:

Brake horsepower =    w  R Tl    [  P2n-l  1
                   33,000 n e [ TLJ

where  w  = weight flow of air Ib/min (acfm x .075 Ibs/ft  air)
      R   = gas constant (53.5)
      T,  = inlet temperature °R
      P   = inlet pressure, psia
      P   = discharge pressure, psia
       n  - (k-l)/k = 0.283 for air
       k  = 1.395 for air, specific heat at constant pressure
            divided by  specific heat  at  constant volume
       e  = efficiency  usual  range  for compressor is 70 to 80


     In order to better appreciate  power requirements the cost of power
should be based on utility energy escalation projections.  In Ontario one
horsepower has a projected present  worth of $3,500 (Cdn).  For a 1000 HP
blower one percent difference  in overall isentropic efficiency is worth
$35,000 (Cdn).

     Generally when making blower selection first costs are not the
determining factor,  but rather, operating costs.
                                 183

-------
 AIR BLOWER TYPES

      There are two types  of  air  blowers commonly used for aeration of sewage:

           -  Centrifugal

           -  Rotary Positive Displacement

      Selection of  which of the two  types of  blowers to use is generally based
 on efficiency and  cost, however,  with increased attention being paid to noise
 control care must  be taken to reduce  blower  noise levels  by careful design
 and selection of building and air piping materials, blower silencers, etc.

      The sound generated  by  a blower  unit  is created by all the rotating
 elements in the blower, motor and speed increaser.   The sound is transmitted
 via the blower casing and foundation,  piping system and the motor windage.

      Reduction of  noise is a major  design  consideration and can in most cases
 be reduced to acceptable  levels by:

      -  impeller design

      -  smooth blower air passages

      -  blower casing material i.e. cast iron

      -  dynamic balancing of rotating components

      -  alignment  of blower  and motor

      -  isolation  of piping  and material

      -  good foundation design

      -  isolation  of building structure

      -  gear increaser selection

      —  silencing  on blower  inlet and  discharge

      -  sound insulation  of  air piping

      -  in  acoustical  treatment of blower  room

      The blower  size and  type should  fit into what  is generally commercially
available.   In capacities from 10 CFM  to 15,000 CFM.   Positive displacement
blowers are  generally more efficient.   Above 15,000 CFM centrifugal machines
are almost always used.  The  reason for this is that centrifugal blowers have
higher efficiency at larger volumes of  air than do  positive displacement
blowers.
                                    184

-------
      Where  unit  capacities are lower than 1,000 CFM positive displacement
 blowers  are preferred and above 10,000 CFM centrifugal blowers are used.

      Certain characteristics of the centrifugal and positive displacement
 blowers  are worth noting and are discussed in the following.


 CENTRIFUGAL BLOWERS

      A centrifugal blower develops pressure by means of bladed rotating
 impellers.   Air  enters the impeller at its eye in an axial direction.   The
 flow is  then changed to a radial direction inside the impeller and
 accelerated toward the peripheral area of the blower casing.  Figures  4,  5
 and 6 show  single stage machines.   In a multistage blower  air is  returned to
 the eye  of  the next impeller thus incrementally increasing the pressure.
 Figure 7 shows a section through a multistage machine.   In sewage treatment
 centrifugal blowers of single stage and up to 3 (multi) stage impellers are
 employed.

      In  order to produce pressure in the range of 7.5 psig single stage
 blowers  must turn at speed in excess of 3600 RPM requiring speed  increases.
 Multi-stage machines generally do not have to turn at not  more than 3600  RPM
 and can  in  a well designed machine run with less vibration and noise than a
 single stage machine.

      One of the  benefits of centrifugal blowers is that they are  variable
 capacity, constant pressure machines.   Change in capacity  is achieved  by
 throttling  at the suction or with one particular manufacturer by  varying
 diffusers strategically placed on the pressure side of  the impeller.   By  the
 above methods turn down capabilities to approximately 45 percent  of  the rated
 capacity are possible with resulting savings in power.

      Figure 8 illustrates theoretical and actual centrifugal blower
 chacteristics.   Because of internal losses the actual volume output
 decreases.

      The  pressure developed within a centrifugal blower is significantly
 affected by air  temperature and density.   The highest inlet temperature
 (lowest  air density)  produces the  lowest  pressure at a  given inlet  volume
Figure 9 illustrates  the effect of temperature.

      With centrifugal  blowers care must be taken to avoid  surge conditions
and  this is the  low capacity limit for this type of machine.   Figure 10 shows
the  surge point  in a  typical blower performance curve.

      The pressure  developed by  a constant speed centrifugal blower varies
approximately with the  inlet air temperature and density.   The blower  will
seek  equilibrium between air flow  and  density.   Figure  11  shows resulting
changes in  the horsepower and pressure-volume characteristics curve.   If,  for
example,  the  average  inlet  temperature is 60°F the capacity handled  will  be
at point "b".  If  the inlet temperature drops to 0°F the air density
                                    185

-------
             DISCHARGE
   PARALLEL
   WALL
   DIFFUSER
   IMPELLER
                                     DISCHARGE
                                     NOZZLE
                                   INLET
                               VOLUTE
Figure 4.   Single Stage Centrifugal Blower
                      186

-------
                                 US EPA Hear1;-."
                                       M2''€,.,£3- !
                                 1200 Pem^yivsAu ^ •->•--
                                                  ""
COOLING
  FAN
             PARALLEL
           WALL OIFFUSER
                             CROSS SECTION
                LABYRINTH
                AIR SEALS
  LABYRINTH
   OIL SEAL
                                                  DISCHARGE
                                             VOLUTE
                                             CASING
       Figure 5.   Single Stage Blower with Speed Increaser
                            187

-------
                      CROSS SECTION
 BEARING
  STAND
 THRUST
 BEARING
          DISCHARGE
           NOZZLE
                                     GUIDE VANE LINKAGE
                                    PARALLEL
                                      WALL
                                    DIFFUSER
                                                     GUIDE
                                                     VANES
                                                      INLET
                                                 INLET NOZZLE
Figure 6.  Single  Stage Centrifugal Blower with Inlet  Throttling
                            188

-------
     CROSS SECTION
 Part No.       Part Name          Material
    1.   Inlet Nozzle            Cast Iron
    2.   Anti-Friction Bearing    Steel
    3.   Bearing Housing        Cast Iron
    4.   Shaft                  Carbon Steel
    5.   Baseplate              Steel
 6.   Labyrinth Seals
 7.   Shaft Seal
 8.   Discharge Volute
 9.   Case—Diaphragm
10.   Impellers
11.   Drive Coupling
'Recommended Spare Parts — Bearings, Shaft Seals, Coupling Seals
                    Figure 7.   Multi-Stage  Centrifugal Blower
Aluminum
Carbon Ring
Cast Iron
Cast Iron
Cast Aluminum
Steel
                                           189

-------
     UJ
     to
     cc

     ui
     tc
     =>
     w
     to
     Ul
     cc
     a.
        THEORY-
         mtuKT-7


 ___	f	u
•        ^^^^ ^^^^ ^^^^^ ^^^^ "*
                  -ACTUAL
                         VOLUME
  Figure 8.  Typical Centrifugal Blower  Characteristics
     M
       150
     U)
     tn
       125
     DC
     Q.
       100
       125
    o
    n.
    UJ
    ut
    it:
    <
    cc
    00
       100
       75
                                         .O'F
        JOO'F
                                  100*F
                 SO         75         100

                       % INLET VOLUME
Figure 9.   Temperature Effects  on Centrifugal Blowers
                             190

-------
                110
                       40   50   60   70  80   90   100  110  120
             INLET VOLUME BEFORE THROTTLE VALVE — PERCENT
          Figure 10.   Blower Volume Control By  Inlet Throttling
                                   0* F - MIN. TEMPERATURE
                            60* F — AVG. TEMPERATURE
                        100* F — MAX. TEMPERATURE
                O 21-0
                in
                  1600 (

                  1500'

                 |,400 |

                 Pl300!
                 u    '
                 ^ 1200

                 §1100
     DESIGN CONDITIONS
INLET PRESSURE      13.64 PSIA
INLET TEMP.
HEL. HUMIDITY
DISCH. PRESSURE
INLET VOLUME
SPEED
BMP AT BLOWER
         I
100* F
60%
22.24 PSIA
30.000 CFM
6100 RPM
1214
                 X1000

                 n 900

                   BOO

                   700
      60*
  10    14    IB    22    26
         INLET VOLUME — MCFM
                                                     30
                                                          34
Figure 11.   Expected Performance  of Single Stage  Centrifugal Air  Blower

                                      191

-------
increases and  the  operating point moves to "c".   If  the  inlet temperature
rises  to 100°F the blower  operating point drops  to "a".   It  is therefore
important in order to deliver the desired air flow and pressure that the most
adverse inlet  conditions are anticipated.  Design for the lowest inlet
pressure and highest air temperatures when specifiying the blower.

     Of just as great concern is the operation of the blower motor.   If the
inlet  volume of a  blower whose design point is at "a" at  100°F is left
unthrottled during the coldest period expected,  0°F, the  horsepower
requirements would rise from approximately 1200  HP to over 1600 HP
respectively.   Throttling  during cold weather periods is  possible but sizing
of  the motor for the cold  weather conditions is  prefered.

     The manufacturers can provide impellers with either  backward or radial
blades.  The backward bladed impeller is best suited for  sewage treatment
operation as pressure changes have a smaller effect  on volume delivered and
because of this, is better suited for parallel operation  ( Figures 12 and 13).
ROTARY  POSITIVE DISPLACEMENT BLOWERS

     A  rotary positive displacement blower is a fixed volume machine.   It
develops pressure by trapping a volume of air in two lobe  type impellers and
carrying the air around the compressor casing to the discharge as  shown on
Figure  14.

     The theoretical and actual performance of this type of a blower is shown
in Figure 15.

     Variations in capacity can be made but require the change of  speed of
the blower.  This is normally done by changing the pulley  size on  the  blower*

     Compared to the centrifugal blowers, variation in inlet air temperatures
do not  have a significant effect on performance.

     Figure 16 indicates the effect of temperature on blower capacity  and
horsepower.

     Figure 17 shows a typical manufactures performance curve.
SYSTEM CURVE

     The system curve determine where the blower operates on  a  particular
blower pressure volume curve.  Figure 15 shows a typical system curve with
parallel operation of centrifugal blowers.

     In calculating this curve the key factors are:
                                      192

-------









:
.
•'









«

















BACKWARD LEANING IMPELLERJ-J-
-H
ND ADJUSTABLE INLET GUIDE VANES

	 - = = :4=*.
£.
£ •*• ^ ""^
"T" " X

V ^
V

V
^
\p

t









s
4 'S
"2 2 ^
^
2 «: "" •*
y ^ **
rj*S
O
fa
J

\ ' • I :
~t PT TT w rr ro ™

^4 	 	 *•*• —~ —m ^ ^— ^ ^
"^ \
^V 1^^
\ >
^ 'v
v V H It Jf J

V 7

^






""* ^ ^
<. ^
•^ S
r A
T" s
^ **
	 i









1 ) i 3 1




































Figure 12.   Open Backward-Bladed Impeller
                   193

-------
                                                               120
Figure 13.  Comparison of Backward Curved and Radial Impellers
                            194

-------
           OPERATING   PRINCIPLE
          Two figure eight lobe Impellers mounted on parallel shafts rotate In opposite directions. As
       each Impeller lobe passes the pump Inlet, It traps a definite volume of gas and carries It around
       the case to the pump outlet, where the gas Is discharged.
          Timing gears on each Impeller shaft control the relative position of the Impellers to each
       other and maintain the clearances. With constant speed operation, the displaced volume Is es-
       sentially the same regardless of pressure, temperature or barometric pressure.
             Figure  14.   Positive  Rotating Displacement Blower
                  Ul
                  to
                  01
                  c
                  3
                  in
                  
                  LLJ
                  OC
                  a.
                               ACTUAL-
THEORY
                                     INLET VOLUME
Figure  15.   Typical Rotary Positive Displacement  Blower Characteristics
                                            195

-------
                  100
                  75
                co
                a.
                to
                ui
                c
                to
                a:
                a.
25
                   0


                 100
                  75
                o
                o.
                UJ
                V)

                O 50

                x

                UJ
               K

               CD
                                              100*F
                                                       0*F
                                            100*F
                                                  100      120

                                                 INLET VOLUME
Figure  16.   Temperature Effects  on Rotary  Positive Displacement  Blowers
                                    196

-------
600  650   700   750  800   850  900  950   1000  1050   1100  1150  1200  1250  1300
                           srilt I.P.N.
              INLET   CONDITIONS - AIR •  14.7 PSIA  a 70°F
   Figure 17.  Typical Rotary Positive  Displacement
                Blower Performance
                               197

-------
      -  static head or depth of submergence over  the  diffuser

      -  diffuser loss

      -  pipe friction loss

      -  minor losses such as bends  and constrictions

      -  control valve losses

      -  blower discharge and inlet  losses

      Maximum wet weather aeration tank level,  diffuser  clogging,  pipe
 friction,  inlet and filter clogging are considerations  that  should be
 carefully  evaluated to ensure the system curve  is representative  of the range
 of conditions the aeration system will be subjected to.


 SELECTION  OF AIR BLOWER

      The number of blowers selected is important in terms  of  being able to
 efficiently match aeration tank air requirements and  blower  capacity.  It is
 impractical to size blowers for 20  year projected condition  if  for the first
 10 years energy will be wasted.  Better to  select a greater  number of smaller
 units matched to the air requirement growth projection.  A stand-by spare
 blower of  at least the largest size should  be  provided  for process and
 operational reliability.

      Generally the fewer number of  blower units the lower  overall cost as
 larger units are more efficient,  and capital cost in  terms of $/CFM are
 lower.   Care however, must be taken to match the blower  capacities and turn
 down capabilities with the projected air requirements.

      For a projected 20 year growth and a centrifugal blower  installation air
 requirements were shown graphically in Figure 2.  The design consideration is
 to efficiently supply air over the  20  year  period within the  boundaries of
 the low and high air requirements.

      A centrifugal blower can vary  its  capacity down  to  approximately
 45  percent  but for safety the relationship  of minimum turn down capability
 and temperature  must  be carefully considered.

     As  the  blowers  will operate  in parallel constantly  rising  pressure
 curves  (backward leaning impellers)  are preferred.  With this type of
 impeller slight  pressure changes  caused by  temperature have  a smaller effect
 on  the volume  of air  produced.  Figure  18 shows parallel centrifugal blower
 operation.

     Generally it  is  desireable to  have  similarly sized  machines  with one
machine as stand-by.  For the example chosen the following  possibilities
arise.
                                    198

-------
      NO.
(includes stand-by)
      4

      5
 SIZE
  CFM

25,200
12,600


 9,200

 6,300
     REMARKS

Too large can not meet
initial low air requirement

Doubtful whether can meet
initial low air requirement

Possible

Possible
     Performance curves are  considered with particular scrutiny given to
parallel operation, turndown capability, efficiency, and temperature.

     For the example, the design point is 18,400 CFM and the best fit would
be the 9,200 CFM machines.   One of the four machines unthrottled would handle
the initial average air requirement and could be throttled to provide the
initial low air requirement.  Two machines both of which would require
throttling would meet the high air requirement.

     Projected 20 year average air requirements could be met by 2 - 9,400 CFM
machines with one machine wide open and possibly with inlet guide vanes over
rotated to meet the future low air requirement.  Over rotation of guide vanes
is possible on larger centrifugal machines and can provide slightly greater
volume.  The projected 20 year high air requirement would require three
machines in operation with slight throttling.

     Blower evaluation should consider building, and mechanical requirements.
Operating costs should consider future power costs.  Life-cycle cost
comparisons can be used and  provides the designer with a mechanism of
comparing overall project lifetime costs.  By knowing the total overall cost
of two or more blower system alternatives the designer is obviously in a
better position to make a better blower selection.
CONCLUSION

     Rising energy and maintenance costs have made the designer aware of the
importance of air blower selection.  The goal of the designer must be to
assure long term reliability and cost effective operation.
                                     199

-------
                                   2 BLOWERS
                                   IN PARALLEL(III)
                                30.000      60.000
                                INLET VOLUME. CFM
            Figure 18.  Parallel Centrifugal Blower  Operation
REFERENCE

1.   Stensel, H.D. and Shell, G.L.   "Two Methods of Biological Treatment
     Design.  Jour. Water Poll.  Control Fed., 46, 1, 271 (1974).


DISCLAIMER

     The work described  in  this  paper  was not funded by the U.S.  Environ-
mental Protection Agency.   The contents do not necessarily reflect the views
of the Agency, and no official endorsement should be inferred.
                                    200

-------
AUTOMATIC DISSOLVED OXYGEN  CONTROL  IN THE  ACTIVATED SLUDGE  PROCESS

Joe P. Stephenson  (presently  visiting the  Water  Quality  Institute, Denmark)
Wastewater Technology Centre,  Environment  Canada
Burlington, Ontario  L7R  4A6

ABSTRACT

     The importance of  control over the  dissolved  oxygen concentration in ac-
tivated sludge systems  is discussed.  Dissolved oxygen control is necessary
from the viewpoint of energy  savings and moreover  to improve the monitoring
and operational control of  these  systems.  Examples from the literature of
these benefits are given.

     A brief review of  possible control  strategy calculations is given. It is
shown that the oxygen transfer efficiency  may be estimated  on-line from dis-
solved oxygen and power input measurements alone.
 INTRODUCTION

     One approach  to more  efficient  operation of wastewater treatment systems
 is through the application of  on-line  instrumentation and automatic control.
 Control, whether automatic or  manual,  is  essential  since wastewater treatment
 systems are rarely, if  ever, at steady-state. This  is true for the activated
 sludge process and especially  for  its  aeration component. The need for con-
 trol of aeration is reflected  directly in the cost  of energy for this task
 and in the improvements in process performance and  process monitoring that
 can be accomplished. Clear reference to the advantages to be gained from aer-
 ation control has  been  made by Flanagan and Bracken in the U.S. EPA's design
 manual concerning  procedures and guidelines for the selection of control sy-
 stems  (1). This paper is directed  towards an overview of the basic process
 and computational  requirements of  dissolved oxygen  control in activated
 sludge systems. For a more detailed  review, see the U.S. EPA design manual.
 Also, an indication is  given about the additional process information which
 can be gleaned from aeration control data by making maximum use of the po-
 tential of digital microprocessor  capabilities. The proceedings of recent
 workshops by the IAWPR  attest  to the great interest in the subject of auto-
 matic control for  improved operations  of  wastewater treatment facilities
 (2, 3).

 Justification for  Automatic Aeration Control

     The justification  for aeration  control has been known for some time
 (4, 5). In 1969, Brouzes (5) indicated that energy  consumption in the activ-
 ated sludge process could  be minimized through the  automatic control of dis-
 solved oxygen concentration with a spin-off benefit accruing from the auto-
 matic control of biomass inventory by  relating biomass growth rate to the
 oxygen transfer rate. Since then,many  pilot and full scale examinations of
 aeration control have been conducted with positive  results for energy savings
 (1, 6, 7, 8, 9). A report  financed by  the Commission of the European Communi-
 ties (10) detailed the  examination of  eight full scale activated sludge sy-

                                     201

-------
 stems  in  Denmark  and  showed  energy  savings  ranging  from  2.5%  to  60%  through
 aeration  and  solids inventory  control. The  12 plants within the  U.S.  study
 (1)  showed an average 33%  improvement  in BOD removed per kWh  of  energy  sup-
 plied.  It appears on  average,  that  about 25% of energy consumption could be
 conserved by  aeration control  in  activated  sludge systems but that the  actual
 savings would be  site specific. A 1977 survey  (11)  including  30  activated
 sludge plants in  Ontario,  Canada, showed that annual utility  requirements
 comprised 21% of  the  total plant  operating  costs. Thus,  25% utility  saving
 could  be  translated to about a 5% saving in the total plant operating costs
 by  tackling only  one  major source of expense - aeration. Furthermore, several
 reports have  indicated improved sludge settleability in plants having dis-
 solved oxygen control (8,  12,  13).  For example, Wells  (8) reported that fila-
 mentous organisms were minimized  in a dissolved oxygen controlled system com-
 pared  to  a parallel manually controlled system, and that only one-half  the
 clarifier capacity was necessary  in the controlled  case. Palm et al.  (12)
 have provided clear evidence of the need for further investigation of dis-
 solved oxygen control in the light  of its apparent  important  impact  on  the
 growth of filamentous organisms in  relation to non-filamentous organisms.
 They give compelling  reasons for  the need to consider aeration control  over-
 and-above energy  savings.  Tuntoolavest et al.  (13)  gave further  justification
 for minimizing excess aeration due  to the complex interaction between aer-
 ation  shear rates and solids clarification.

     Flanagan and Bracken  (1)  have  summarized conditions under which  dis-
 solved oxygen control is justified. To avoid the pitfalls of  an  improper de-
 sign,  their conclusions should be reviewed  before deciding upon  the  selection
 of  a control  system.
 ACTIVATED SLUDGE AERATION PROCESS  INTERACTIONS

      It is known that many  interacting factors govern the design and con-
 struction of  aeration systems, particularly from a steady-state viewpoint.
 These factors include  (14):

    • Mechanical design  of the aeration system
    • Hydraulic disturbances of the influent wastewater and return  sludge
    • Recycle  from other process  streams
    • Changes  in the  feed pattern
    • Organic  load disturbances and substrate storage
    • Nitrification/denitrification
    • Mixing
    • Changes  in excess  sludge wasting practises
    • Disturbances in the oxygen  transfer efficiency

      It is  less appreciated that these interacting factors also govern  the
effectiveness of an aeration control system in a dynamic rather than steady-
state sense. To a certain degree each factor dictates the controller's  suc-
cess  or failure. Houck  and  Boon  (15) have documented many examples of the ef-
fect  of these interactions  from  a  steady-state viewpoint and  the problems
that  can result from inattention to them, especially for plants having  fine
bubble aeration. Similar examples would apply to other aeration designs.

                                     202

-------
Flanagan and Bracken  (1) provided  several  examples of design flaws which pro-
hibited the use of  installed  aeration  control  systems due to inattention to
the complexities of these  interactions in  the  design stage. A typical problem
is the oversizing of blowers  or  air  flow control  valves to accommodate future
plant loads at the  expense of current  operation efficiency- Improperly di-
mensioned equipment can  result in  uncontrollable  systems which will not re-
spond effectively to control  actions.

     Olsson and Andrews  (16,  17) have  effectively analyzed the interaction
between process variables  and aeration dynamics to gain insight into the per-
formance and control of  the activated  sludge process. The relationships be-
tween these variables  were examined  through monitoring only the dissolved
oxygen concentration profile  and oxygen transfer  rate within the aeration
tank. Further efforts  in the  area  of improving operational monitoring and
control have been discussed by Holmberg (18),  Olsson & Hanson  (19) and Cook &
Libelli  (20). In all cases the need  for effective aeration control and recur-
sive updating of the oxygen transfer coefficient  was established. In these
cases, effective use of  digital  controller capabilities must be made. Of spe-
cial interest to this  workshop is  the  possibility for the on-line estimation
of the process oxygen  transfer coefficient, otKLa, using the interactions and
relationships between  the  process  variables to advantage.

     It also is important  to  differentiate between the long term average
steady-state demand for  oxygen transfer and the actual non steady state de-
mand due to the plant  loads and  disturbances.  Clifft & Andrews  (21) and S0-
rensen  (6) have shown  that short term  variations  in plant organic load are
attenuated by the storage  and later  oxidation  of  substrate by the biomass.
This has important  implications  for  the application of process control stra-
tegies, since the magnitude of measured variations in carbonaceous and nitro-
genous loads  is not  immediately reflected in  the plant oxygen demand. Thus,
the control system  hardware  should  not be designed directly on the measured
non steady state carbonaceous and  nitrogenous  loading and removal, but on
providing sufficient oxygen transfer by manipulation of the oxygen transfer
coefficient to meet the  actual in-situ oxygen  demand. The oxygen transfer co-
efficient is a direct  function of  the  power input to the aeration system and
can be varied by manipulating the  power input.

     Figure  1 shows an example of  the  relationship between the actual oxygen
transfer rate and the  variation  in carbonaceous and nitrogenous  load ex-
pressed in oxygen equivalents for  a  DO-controlled pilot scale activated
sludge system treating municipal wastewater  (22). The DO concentration was
automatically controlled at a setpoint value of  2.0 g.m   . To maintain the DO
concentration at 2.0 g.m"3 required  a  significantly greater amplitude in
energy input  (air flow rate)  than  was  measured for the actual oxygen transfer
rate. Similar observations were  reported by S0rensen  (6).

     The relative amplitude of the energy  input  to  the actual  oxygen require-
ment  (transfer) is  directly dependent  on the relationship between  the  in-situ
oxygen transfer coefficient,  KLa,  and  the  power  input. Figure  2  depicts  the
general form of a linear relationship  between  KLa and power  input  as a  func-
tion of increasing  transfer efficiency. In a diffused air system,  the power
input is proportional  to the  air flow  rate required  to maintain  the  dissolved

                                     203

-------
       160
                                                                  450
                    + ATKN
               (OXYGEN EQUIVALENTS)
                           16       20       0

                                TIME (hours)
     FIGURE 1.  Relation between C>2 transfer, subtrate removal,
                and air flow rate under diurnal loading conditions (22) .
oxygen concentration. For surface aerators, the power draw depends on the ro-
tational velocity and submergence depth. From the relationship shown  (Figure
2), it is clear that the ability to manipulate the oxygen transfer coeffi-
cient to control the oxygen transfer rate and the desired DO concentration is
entirely dependent upon the ability to manipulate the power input. In the
simplest case, the latitude of control achievable is governed by the  slope of
the relationship shown in Figure 2 and the upper and lower limits on  the ac-
tual power which can be delivered. Regardless of the system under considerat-
ion, it is necessary to be able to mechanically operate the aeration  system
along an appropriate curve shown in Figure 2 to achieve the degree of control
desired under variable loading conditions. It is essential to consider the
load characteristics, the primary aeration device, the control elements  (e.g.
control valves, variable speed drives), and the latitude of control possible
to ensure that controllability can be achieved with ease and that real ener-
gy savings result. See reference  (1) for more detail.

     A further consideration in aeration control is the interaction between
the aerator and the clarifier  (6, 21) . Since the respiration of the biomass
is the dominant factor governing the need for aeration, the distribution of
solids between the aerator and clarifier influences the total oxygen  demand
in the aeration tank. This is true in a dynamic and in a static sense.

                                     204

-------
                                               INCREASING TRANSFER
                                               EFFICIENCY
                                                                 12
                           POWER INPUT
      FIGURE  2.   Relationship between ©2 transfer  coefficient  and
                 power input for various transfer  efficiencies as-
                 suming a linear relationship.
 If  the clarifier  capacity to store  solids is large  and  the  capacity of  return
 sludge pumping  is limited at peak clarifier  loading rates,  then  there will be
 a transfer of solids  from the aerator to the clarifier.  This will  tend  to de-
 crease the oxygenation  demand in the  aeration tank  a corresponding amount.
 S0rensen  (6) has  also discussed the impact of clarifier solids inventory
 buildup upon the  total  oxygen demand  and upon the possibilities  for nitrific-
 ation from a static viewpoint.  However,  in general  there does not  appear to
 be  a consensus  on the best operating  procedure for  controlling the distribut-
 ion of solids between the settler and the aerator.  Significant efforts  are
 underway  in Sweden to examine rational procedures for solids inventory  con-
 trol (23) .
SIGNIFICANCE OF REACTOR CONFIGURATIONS

     A variety of reactor configurations have been  developed for  the  activat-
ed sludge process.  In general,  these designs can be related  to  the  mixing
characteristics of  the reactors.  These  characteristics  range from a small
amount of dispersion to essentially completely  mixed (24). Each mixing regime
has a significant effect on the aeration dynamics and must be considered in

                                     205

-------
the design of an aeration control system.

     Olsson and Andrews  (16,  17) have discussed in detail the impact of mix-
ing, reaction and mass transfer upon aeration dynamics through the examinat-
ion of the spatial and temporal DO profile. It is clear that a spatially va-
rying oxygen demand along an  aeration tank requires a non uniform rate of
oxygen transfer for uniform dissolved oxygen concentration control. This can
be accomplished by varying the power input spatially along the tank. In the
steady-state case, this has been accomplished by using a tapered diffuser  sy-
stem, see Flanagan and Bracken  (1) and Brouzes (5). One example of an auto-
matic control system designed to regulate the distribution of air along the
aeration tank is the Rye Meads treatment plant in the United Kingdom (4, 25).
In the absence of a system to regulate the distribution of oxygen transfer
along an aeration tank,  then  it is only possible to regulate the dissolved
oxygen concentration at a single point. Olsson (26) has analyzed this situat-
ion and shown the advantage to be gained from controlling the shape of the
dissolved oxygen profile in a non-completely mixed reactor rather than a
single spatial concentration. In completely mixed reactors, where the oxygen
demand is uniformly distributed, the control of dissolved oxygen concentrat-
ion reduces to the control at a single point to ensure relatively even dis-
tribution of the DO concentration within the reactor. In the design of over-
all aeration systems, the tapering of the power input must be carefully con-
sidered in relation to dimensioning of pipes, control valves and diffuser
density to obtain controllability.
DISSOLVED OXYGEN/AERATION CONTROL STRATEGIES

     In activated  sludge process aeration control strategies, the primary ob-
jectives are to  satisfy the oxygen demand of the respiring biomass and to mi-
nimize the consumption of energy. This  is almost invariably accomplished by
controlling the  dissolved oxygen concentration to a prescribed setpoint leveL
Several studies  (1, 4-10, 27) have compared the performance of manual control
of the dissolved oxygen concentration to its automatic control, but the con-
clusions reached have shown that manual control of the oxygen transfer rate
is always inferior to some form of automatic control. Manual control leads to
about 25% increased energy consumption  and widely varying values of the dis-
solved oxygen concentration. Conversely, effective results in minimizing
energy consumption while controlling dissolved oxygen concentration to a set-
point value were reported using closed-loop automatic control to adjust the
oxygen transfer  efficiency. Stephenson  et al. reported a reduction in the
standard deviation of the dissolved 02  concentration from 0.9 mg.L~*,.to 0.2
mg.L"1 through DO  control at a 2 mg.L"1 setpoint (28).              w

     The usual form of automatic control for dissolved oxygen concentration
in the activated sludge process has been feedback PI regulation and feedback
switching control  (on/off or multi-speed) based on the signal from a dis-
solved oxygen sensor. PI regulation has been satisfactorily used to manipul-
ate variable speed drives (variable frequency type) on positive displacement
blowers and on surface aerators, and inlet guide vanes or throttling valves
on centrifugal blowers. On/off and multi-speed controllers have been used for
intermittently switching drive motors to different set speeds or on and off

                                     206

-------
as necessary. Furthermore, investigations are on-going at the Wastewater
Technology Centre and in Sweden  (23) to examine continuous self-tuning regul-
ators for the control of dissolved oxygen.

     A block diagram representing a simplified feedback control loop for con-
tinuous regulation is shown in Figure 3 for dissolved oxygen control. In
practice? in the case of diffused aeration with an analog PI regulator, a
cascade controller may be required to control the blower output volume and/or
pressure to achieve stable and responsive control  (1). In such a case the
output of the dissolved oxygen sensor is used as a basis for adjusting the
setpoint within the inner controller. With a self-tuning controller, it may
be that a single feedback loop with dead-time compensation could provide ef-
fective aeration control for diffused and surface aeration systems.
                                                               02 demand
 D0setpoint
                    ^"measured
                                 DO
     FIGURE 3.  Simplified block diagram for DO control system.
     Two types of controller hardware, analog and digital, are in use for  in-
dustrial process control. However, programmable digital controllers are be-
coming much more popular due to their decreasing price and enormous flexibi-
lity. The use of digital controllers offers the opportunity for relatively
straightforward dead-time compensation and the implementation of more ad-
vanced controllers that can be extended beyond the empirical engineering so-
lutions exemplified by PI control. In addition, raw signals from the process
sensors can be digitally filtered to remove undesirable  noise. Simple time
averaging  (7) and exponential filters  (19, 27) have been used. The first-or-
der discrete filter resulting from the continuous system differential equat-
ion:
                Tdc/dt  +  c  =  x

is used very often in digital systems.
equation becomes:
                Upon discretization this continuous
                 "n+1
                      =  ac
          (l-a)x
                                       n
where
continuous system time constant
discrete filtered signal for n-th time interval
discrete raw measurement at n-th time interval
                                     207

-------
                                                            -At/T
                a   =  equivalent filter time constant  =  e
                At  =  sample interval

Petersack and Smith (27) used this filter in a demonstration aeration control
project with a sample interval, At, of 6 seconds and a value for a of 0.97.
This meant that about the last 30 observations (I/(1-0.)) of the raw DO signal
were effectively used in the control algorithm. Wells (8) used combination
analog and digital filters to smooth out noise from raw dissolved oxygen con-
centration measurements. It can be seen that the discrete first-order filter
can be programmed easily on a digital controller and needs only two values
(a and cn) to be stored between sampling intervals. Filters should be used  •=
with care and forethought since a disadvantage of their use is the masking
of potentially useful process and system measurement information.

     For the case of direct continuous PI control for motor speed or valve
position control, the velocity form of the discrete PI controller has been
used frequently  (27, 28):

                Vut+l  =  ye^l+At/Tj-e^)

where           Vu   =  change in control signal at time t
                Kp   =  proportional gain
                T    =  integral time
                At   =  sample and control time interval
                e^-   =  current deviation of DO from setpoint at time t.

The actual signal output by the controller is simply calculated:


                Ut+l  =  ut  +  Vut+l

Acceptable values for At in activated sludge systems are between 30 - 300 se-
conds. Values of the gain and integral time must be obtained from experiment-
al tuning. The velocity form of the discrete controller can be implemented
easily on a microprocessor.

Although the PI controller is appealing in its simplicity for automatic re-
gulation, advances in process control have lead to the development of self-
tuning minimum variance controllers. These controllers have the general form
(29):





where           u    =  control signal

                V    =  difference operator, e.g. V u  = u  - u
                                                     T-    t    L™ J.
                B    =  backward shift operator, e.g. By  = y

                a,$  =  controller tuning parameters

                l,m  =  orders of polynomials in a(B), 3(B) respectively

                y    =  observed measurement signal.

                                     208
2
(oto + a.B + a B + ...
(Bo + B-B + 3 B2 + . . .
+ a Bm)
m


-------
Usually d is equal to one and 1, m - 2. If the correct order of the control-
ler is chosen, which can be established through diagnostic checks, then the
controller tuning parameters  (a, 3) can be calculated and updated recursively
as process conditions change, for example due to clogging of diffusers or
seasonal temperature fluctuations in aeration systems. Dead-time,also, can
readily be accomodated by these controllers. With proper usage, this ensures
that the continuous controller remains tuned to track slowly changing process
parameters adaptively. These adaptive controllers can be implemented for
feedforward-feedback situations and for situations where the sequential
changes in the manipulated variable must be constrained  (e.g. motor drive
changes). This can help not only to avoid process upsets due to rapid and
large changes in the control action but also to minimize wear due to exces-
sive control action on mechanical components such as drives and bearings. The
usual method of solving the control requirement is through use of a compact
recursive least squares algorithm which can be implemented on* a digital pro-
cessor using little space. See Harris  (30) for example.

     Furthermore, the use of recursive least squares or maximum likelihood
algorithms for control or estimation purposes provides an opportunity to
estimate process parameters on-line. Under the appropriate conditions, it has
been shown that the in-situ value of K^a and volumetric respiration rate (R)
can be determined recursively from dissolved oxygen concentration and power
input measurements alone  (18, 19, 20, 26, 31). For example, Holmberg  (18) has
examined several ways of estimating KLa for estimation and control purposes
in a diffused air system, using the formulation that the oxygen transfer ki-
netics in a respiring system could be defined as:

                Tdy/dt + y  =  gu

where           y  =  dissolved oxygen concentration
                T  =  time constant  (l^a  )
                t  =  time
                g  =  system gain parameter
                u  =  air flow rate.

Upon discretization, this continuous system was described by:

                Y   ==  ^v-i-_i + (JL)U4- 1

                        -K^a  • At
where           6   =  e
                    =  1 - K a ' At

                to   =  g(l-6)

                At  =  estimation time interval

                g   =  u~1(ys - R/KLa)
                y   =  saturation dissolved oxygen concentration

Proper selection of the estimation interval, At, and the presence of suffi-
cient disturbance in u  permitted the recursive estimation of the aeration


                                     209

-------
parameters, KLa and R.  In addition to the dynamics of aeration, Olsson an
Hanson  (19) included a  stochastic moving average noise term to account for
unknown process disturbances  in a recursive model used to estimate KLa a  a
full-scale activated sludge plant in Sweden. Where the parameters are expect-
ed  to vary with time  (e.g. due to diffusor clogging), then discounted recur-
sive estimation algorithms are used to reduce the impact of old data upon the
current parameter estimates.  Recursive least squares or maximum likelihood
methods can be used. See  (19, 29, 30) for example. These methods should track
changes in aerator efficiency due to ambient environmental changes. They pro-
vide tremendous possibilities for monitoring the actual performance of aerat-
ion systems and for improving the control of the activated sludge process
 (5, 18, 20).  Significantly more effort is needed to advance these possibili-
ties. For example, see  Yust et al.  (32) for the need to estimate respiration
rate on-line. To date,  no work would appear to have been done to adapt these
methods to surface aeration systems.


AERATION CONTROL IN PRACTICE

     The U.S. EPA aeration control manual (1) documented twelve case histo-
ries discussing aeration control in the United States in the 1970's. The ge-
neral conclusions reached indicated that dissolved oxygen control systems
were warranted for activated  sludge plants having at least 0.044 m3.s~  hy-
draulic capacity, if the guidelines presented by the manual were met.

     Andersson  (10) documented the performance of eight full scale systems in
Denmark ranging in hydraulic  capacity from 0.0046 to 0.30 m . s~* capacity.
Except for one of these plants, he indicated economically attractive results
through implementing instrument-based feedback aeration control.

     For the European and United States' case histories, a variety of aerat-
ion devices and control procedures were described. These included diffused
and surface aeration devices. Positive displacement and centrifugal blower
descriptions have been provided for the case of diffused air, and turbine and
brush systems for surface aeration. Control strategies have ranged from con-
tinuous PI control of blower  output and impeller submergence to on/off con-
trol procedures for all aeration equipment types.
     Still, it must be said that further development of aeration control po-
licies is necessary. Though strategies exist and have been demonstrated for
dissolved oxygen control, their implementation is not widespread in all coun-
tries even where the economics may readily warrant it. There are cases where
the measurement and control hardware has or will be installed but will only
be used for data logging purposes rather than control. Other cases have been
known where operators lack even adequate portable instrumentation for moni-
toring plant dissolved oxygen levels. In these cases maximum aeration capaci-
ty has been used at all times even though largely not needed. These examples
imply that there are often unnecessary hidden costs not accounted for in
plant operation.


                                     210

-------
     There would appear to be the need for more effort toward the control of
point source surface aeration systems. Although the hardware to control sur-
face aeration systems does exist, the control strategies for these systems
have received less attention than those directed to diffused air systems.  In
the case of multi or variable speed drives, the designer is faced with the
requirement to install these units on several surface aerators compared to a
relatively small number of units for blower systems. Efforts are needed to
evaluate the design, economy and effectiveness of such systems.
     Efforts are underway at the Wastewater Technology Centre in Burlington
to implement a feedback self-tuning minimum variance regulator for the con-
trol of dissolved oxygen concentration in a pilot scale activated sludge re-
actor. This will be adapted to a package scale system having a positive dis-
placement blower with variable frequency drive. The results will be used for
retrofitting existing larger scale facilities where warranted. Efforts are
also underway in Sweden (23) to examine advanced operating control strategies
including adaptive control.

     In Denmark  (33), where aeration control is widely practised, there are
studies underway to use the aeration control measurements to regulate the
feed distribution in a step feed system using a microprocessor. In this case,
the controller will calculate the aeration tank oxygen uptake rate from dis-
solved oxygen and air flow rate measurements alone. The uptake rate will be
used to control feed distribution in an attempt to minimize energy consumption,
sludge production and effluent quality.
SUMMARY

     The hardware for aeration control in activated sludge systems exists and
should be further exploited to improve plant operations. Although energy
savings have been the dominant force behind aeration control, it is becoming
recognized that other benefits also are derived from control, e.g. improved
settleability.

     It also is becoming recognized that other process information can be ob-
tained from aeration control measurements. These include estimation of re-
spiration rate and KLa within the process under time varying conditions. Di-
gital processors must be used to obtain these estimates. With these measure-
ments and tools it should prove to be possible to improve control over such
other variables as sludge inventory. More investigation of real-time estimat-
ion of process parameters is necessary. Advances in strategies for control of
diffused air systems need to be translated to surface aeration systems.
REFERENCES

 1.  Flanagan, M.J., and Bracken, B.D., "Design Procedures for Dissolved Oxy-
     gen Control of Activated Sludge Processes". U.S.EPA-600/2-77-032, June,
     (1977).


                                     211

-------
  2.   International Association Water Pollution Research,  Progress  in  Water
      Technology,  <3  (5-6),  (1978).
  3.   International Association Water Pollution Research,  Water  Sci. Tech., J_3
      (8-12),  (1981).
  4.   Jones, K., Briggs, R., Carr, J.G., and Potten, A.H.,  "Automatic  Control
      of Aeration  in  a  Fully Nitrifying Activated-Sludge  Plant". Paper to the
      Inst. Public Health Engineers, Carton Hall, Westminster, London,  March,
      (1969).
  5.   Brouzes,  P., "Automated Activated Sludge Plants with Respiratory Meta-
      bolism Control".  Advances in Water Pollution Research, Proc.  4th Int.
      Conf., Prague,  (1969).
  6.   S0rensen, P.E.,  "Pilot Scale Evaluation of Control  Schemes for the  Ac-
      tivated  Sludge  Process". Water Quality Institute  (Denmark), Report  No.  1
      H0rsholm, Denmark, July,  (1979).
  7.   Stephenson,  J.P., Monaghan, B.A., and Laughton, P.J.,  "Automatic Control
      of Solids Retention Time and Dissolved Oxygen in  the Activated Sludge
      Process." Wat.  Sci. Tech. JL_3  (12), 751,  (1981).
  8.   Wells, C.H., "Computer Control of Fully Nitrifying  Activated  Sludge Pro-
      cesses".  Instrumentation Technology,  (4) 32,  (1979).
  9.   Joubert,  P.J.,  and Mignone, N.A., "Speed-Controlled Waste  Treatment Aer-
      ators Conserve  Horsepower". Specifying Engineer,  (118-123), July,  (1979).
 10.   Andersson, L.G.,  "Energy Savings at Wastewater Treatment Plants". Final
      report to the Commission of the European Communities and the  Danish
      Council  of Technology, Contract No. 314-77-EE-DK, August,  (1979).
 11.   Ontario  Ministry  of the Environment,  "Water Pollution Control Projects:
      Operating Summary". Municipal and Private Section,  Pollution  Control
      Branch,  Toronto,  (1977).
 12.   Palm, J.C.,  et  al., "The Relationship between Organic Loading, Dissolved
      Oxygen Concentration and Sludge Settleability in  the Completely-Mixed
      Activated Sludge  Process". J. WPCF, J52  (10) 2484,  (1980).
 13.   Tuntoolavest, M., Miller, E., and Grady Jr., C.P.L., "Characterization
      of Final Treatment Plant Final Clarifier Performance". Purdue Universi-
      ty Water Resources Research Centre, Tech. Report  No. 129,  June,  (1980).
 14.   Olsson, G.,  "Automatic Control in Wastewater Treatment Plants".  Trib,
      Cebedeau, 436  (33) 121, (1980).
 15.   Houck, D.H., and  Boon, A., "Survey and Evaluation of Fine  Bubble Dome
      Diffusor Aeration Equipment". U.S. EPA Grant No.  R806990,  March,  (1980).
 16.   Olsson, G., and Andrews, J.F., "Estimation of Loading Changes for the
      Nitrifying Activated Sludge Process". Paper presented at the  IFAC World
      Congress, Kyoto,  Japan, (1981).
 17.   Olsson, G., and Andrews, J.F., "The Dissolved Oxygen Profile  - A Valu-
      able Tool for Control of the Activated Sludge Process". Water Research,
     _1J2 (11)  985, (1978).
 18.  Holmberg, A., "Microprocessor - Based Estimation  of Oxygen Utilization
      in the Activated  Sludge Wastewater Treatment Process". Int. J. Systems
     Sci., t2  (6)  703, (1981).
 19.  Olsson, G., and Hansson, O., "Stochastic Modelling  and Computer  Control
     of a Full Scale Wastewater Treatment Plant". Report 7636  (C), Dept. of
     Automatic Control, Lund Inst. of Tech., August,.  (1976).
20.  Cook, S., and Marsili-Libelli, S., "Estimation and  Control Problems in
     Activated Sludge  Processes". Wat. Sci. Tech., 13  (12)  737,  (1981).
21.  Clifft,  R.C., and Andrews, J.F., "Predicting the  Dynamics  of  Oxygen

                                     212

-------
     Utilization in the Activated Sludge Process". J. WPCF, 53  (7) 1219,
     (1981).
22.  Stephenson, J.P., "Automated Operating Strategies for Improved Control
     of the Activated Sludge Process". 4th Collogue Annuel sur le Traitement
     des Eaux Usees, Montreal, Canada, 16-17 November, (1981).
23.  Olsson,  G., "Research Proposal - Control of the Activated Sludge Pro-
     cess".  Lund Inst. Tech, Lund, Sweden, February,  (1982).
24.  Tomlinson, E.J., and Chambers, B., "The Effect of Longitudinal Mixing
     on the Settleability of Activated Sludge". Water Research Centre (U.K.),
     Technical Report No. 122, July,  (1979).
25.  Wakeford, A.C., "Specification for the Control System to be Used for the
     Energy Saving: Fine Bubble Aeration Optimization Experiments at Rye
     Meads STW". Internal report, Water Research Centre (U.K.), Project 1815,
     February,  (1981).
26.  Olsson,  G., "Some New Results on Activated Sludge Control Based on Dis-
     solved Oxygen Profiles". J. ACC, San Francisco, August,  (1980).
27.  Petersack, J.F., and Smith, R.G., "Advanced Automatic Control Strategies
     for the Activated Sludge Treatment Process". U.S. EPA-670/2-75-039, May,
     (1975).
28.  Stephenson, J.P., Monaghan, B.A., and Yust, L.J., "Pilot Scale Invest-
     igation of Computerized Control for the Activated Sludge Process".  Re-
     port SCAT-12, Canada Mortgage and Housing Corporation, Ottawa, Canada,
     in press,  (1982).
29.  Harris,  T.J., MacGregor, J.F., and Wright, J.D., "Self-Tuning and Adapt-
     ive Controllers: An Application to Catalytic Reactor Control". Techno-
     metrics, 2:2 (2) 153, (1980).
30.  Harris,  T.J., "Theory and Application of Self-Tuning Regulators". M. Eng.
     Thesis,  McMaster Univ., Hamilton, Ontario, Canada, May,  (1977).
31.  Mclnnis, B.C., Lin, C.Y., and Butler, P.B., "Adaptive Microcomputer Dis-
     solved Oxygen Control for Wastewater Treatment". Proc. 5th IFAC Symp.  on
     Identification and System Parameter Estimation, Darmstadt, FRG,   (1979).
32.  Yust, L.J., Stephenson, J.P., and Murphy, K.L., "Dynamic Step Feed Con-
     trol for Organic Carbon Removal in a Suspended Growth System". Wat. Sci.
     Tech.,  13  (12) 729, (1981) .
33.  Andersson, L.G., "Research Plan: Energy and Process Economy in Control
     of Purification Plant Effluent Quality". Water Quality Institute (Den-
     mark),  Assignment No.  81.782, December,  (1981).

DISCLAIMER

     The  work described  in  this  paper was not funded  by the U.S.  Environ-
mental Protection  Agency.   The contents do  not necessarily reflect the  views
of the Agency,  and no  official endorsement  should  be  inferred.
                                     213

-------
 EVALUATION OF A DISSOLVED OXYGEN FIELD TEST PROTOCOL

 Walter W. Schuk
 U.S. Environmental Protection Agency
 Municipal Environmental Research Laboratory
 Cincinnati, Ohio

 ABSTRACT

      A test protocol which was developed  for on-line dissolved oxygen meters
 was evaluated in a 60-day field test conducted  in an activated sludge
 aeration basin with thirteen meters  from  seven  cooperating manufacturers.
 The paper discusses the important features  of the test protocol and describes
 the procedures and results of the field test.   These results provide informa-
 tion on methods for testing, calibrating  and performance monitoring, informa-
 tion useful for developing specifications,  and  information eft rputine main-
 tenance for on-line dissolved oxygen meters in  a specific plant environment.

      A critical performance requirement for on-line dissolved oxygen meters
 is to accurately measure dissolved oxygen (within +0.5 ppm of Og) with a
 mean time between maintenance (cleaning,  calibration, etc.) of no less than
 30 days.  Of the meters tested by the developed protocol, nine meters from
 five manufacturers met the criteria.   Of  the four meters (two manufacturers)
 which did not meet the maintenance requirements, one manufacturer has an
 accessory component which,  if made standard, may bring the meter into
 conformance with the protocol.   The  other manufacturer has an alternative
 model which may meet the protocol performance requirements.

      The study indicated that the developed protocol is appropriate for
 acceptance testing of on-line dissolved oxygen  meters and that the state-of-
 the-art meters met the performance requirements of the protocol.


 INTRODUCTION

      Continuous monitoring  of dissolved oxygen  is necessary in wastewater
 treatment  and  particularly  in activated sludge  process control in order to
 insure  satisfactory  effluent  quality  while  at the same time conserving
 energy.  Yet only  limited systematic  information is available on the behavior
 of commercial  dissolved  oxygen  analyzers  in wastewater treatment plant
 environments.  Therefore, the program described in this paper was initiated
 in order to develop,  for a  particular set of plant conditions:  (1) Informa-
 tion useful for writing  procurement specifications for on-line dissolved
oxygen analyzers;  (2) Methods for accurate  and  fair acceptance testing,
calibration and performance monitoring; and (3) Information on routine
maintenance requirements.

     The field test was  a cooperative effort involving the National Bureau of
Standards  (Fluid Engineering  Division), the Environmental Protection Agency
 (Municipal Environmental Research Laboratory),  the Washington Suburban Sani-
tary Commission  (WSSC),  and participating manufacturers of on-line dissolved
oxygen meters.

                                     214

-------
    .The test document was divided into five sections.  Each section provided
a detailed description of a specific aspect of the test.

     Section one  stated the objectives  of the test  and  the  test protocol,
identified the  participants,  and defined the responsibilities  of each par-
ticipant during the course of the test.

     Section two described the methods  used  to solicit equipment  for the test,
and the type of equipment that would and would not be tested.

     Section three provided the location of the facility selected for the test,
and typical operating parameters  for activated sludge process at  the facility.
This section also stated that each D.O.  meter  would be assigned a specific
location in the aeration basin and that the test period was sixty days.

     Section four described the  on-site  provisions  for equipment mounting,
installation assistance,  and defined  the  fourteen  day period  allowed  for
installation and  start up.

     Section five described the procedures  that would be followed during the
sixty day  test  period,  this  included  the test starting date,  the last date
manufacturers would be allowed to adjust equipment, the method for determining
the level of maintenance that would be provided during the test, the definition
of equipment failure, the response to equipment failure, and all information
describing the development and logging of test data.
THE TEST PROTOCOL

     A  written test  protocol was  developed  to give  all  participants  an
opportunity  to understand and comment  on the proposed  test conditions and
further to insure  that  all  meter manufacturers would be treated alike.  The
final test protocol used in this  test  program reported on here is given in the
appendix.  It  incorporated comments obtained  by  circulating earlier drafts
among prospective participants.  Recommendations for  additional modifications
of the test protocol based on the  field tests  reported on here are presented in
the Conclusions and Recommendations.

Test Duration

     Because many  users have  indicated  that they would  consider  30 days a
reasonable interval between  instrument  servicings,  a test duration of  twice
that length,  60 days, was selected as adequate to provide useful information on
routine operation and maintenance requirements.  To  further  improve the data
base, each manufacturer was  required  to furnish two meters  for  the test.  A
substantially  longer test period would have  taxed the resources  of the  staff
conducting the tests and  would  have kept  the manufacturers'  demonstration
units tied up  for  an  inordinate  time, while not necessarily  furnishing more
information on instrument maintenance.
                                     215

-------
 Responsibilities of Participants

 Manufacturers

      To assure  proper installation  and start-up  of  all  units,  the  manu-
 facturers  were asked to have their own personnel install two identical meters
 at assigned locations in an aeration basin.  Two weeks  just  prior  to  the test
 were allowed for this purpose.  These representatives  could  give  written or
 oral maintenance  instructions  to  NBS  and  WSSC personnel  at the  time  of
 installation but were not to be present, unless contacted,  once the test period
 started.

 Washington Suburban Sanitary Commission

      WSSC  operators were  to log the  meter  readings  twice  per  shift (approxi-
 mately four-hour intervals) around the  clock.   WSSC operators were  also to
 perform all "routine" maintenance on the meters (day shift only) as requested
 and  logged by  the  NBS  project -supervisor.    The  definition  of   routine
 maintenance was to  be determined on a meter-by-meter basis by discussion with
 the manufacturers  at the time  of  installation.    (Membrane/electrolyte  re-
 placement  was  generally  agreed upon  at  that  time as the  upper bound  of
 maintenance to  be   provided by operators.)   Routine  maintenance,  such  as
 membrane cleaning,  was to be performed as indicated in the  user's manual or per
 oral instructions,  but  no  maintenance  requiring  the services  of  instru-
 mentation  specialists was  to be furnished.

 National Bureau of  Standards

      NBS staff was  to make reference dissolved oxygen  measurements  at least
 once a  day and be responsible for quality control of the collected  data.  NBS
 was also to furnish on-site  project  supervision with authority to  determine,
 for example, whether meters that  failed  beyond operator  repair should  be
 removed from the test or be  serviced by  the  manufacturer.
 DESCRIPTION OF  THE  TEST  SITE

 The  Aeration Tank

     The  test was  conducted at the WSSC Seneca Interim Wastewater  Treatment
 Plant  near Gaithersburg, Maryland.   This  is a 5-mgd plant using  five  1-mgd
 aeration  basins  operating  in  extended  aeration.    There  is  no  primary
 sedimentation before the activated  sludge process and  the aeration  basins
 operate with a  suspended solids  concentration of  4000-5000 mg/liter.

     As shown in Figure 1, the aeration basin comprises about seven-eighths of
 the outer annulus of the 40 m (130 ft) diameter, 5-1/2 m (18 ft) deep tank.  The
 remainder of the annulus is devoted to sludge digestion.  The central circular
 section of the tank is  the clarifier, from which samples for probe calibration
were taken at the weir overflow.  Air diffusers are located radially along the
bottom of the annulus  at intervals of about  22 degrees.  This air flow,  which
 also induces the mixing currents,  is kept  essentially  constant except  for

                                      216

-------
                  AREA
Figure 1.  Aeration basin test site.
                    217

-------
 diversions to the digester section.  A test area about 30 m (100 ft) long a^"°und
 the periphery  of the tank (Figure  1) was  selected mainly because it could be
 conveniently reached from ground level.   An overall view of the  test site is
 shown in the photograph of Figure  2.

      The test  area encompassed five diffusers (Figure  3).  Air flow from these
 diffusers induced strong,  unsteady,  lateral currents,  which caused  foam to
 collect on the  surface midway between the diffusers.   It was inferred from the
 visible currents that there would be a general downward component of velocity
 between  the diffusers with  concentrated upward currents  at and  near  the
 diffusers.

 Assigning Test Probe Locations

      Twenty test  meter  positions  (A  through  T)  were  laid  out  as  shown
 schematically  in Figure  3.  The space  approximately  6  to  7 feet on either side
 of the air risers was avoided  in order  to minimize membrane damage  by solids
 entrained in strong  upward currents.   When it was learned  later that  there
 would be  fewer meters than anticipated, the center positions midway between the
 air risers were eliminated.  Two meter positions for each  manufacturer were
 assigned so that  one would be in an upstream zone of relatively low  dissolved
 oxygen (d.o.) while  the other was down-stream in  a higher d.o. zone.   Further,
 if the manufacturer's  first meter was  to be close to an air  riser, the second
 was assigned a  location  farther  from the  riser.

 The Internal Velocities

      An  approximate  check on the  internal  velocities  in  the  area of  the
 assignment meter locations was  obtained  during  the pre-test  investigations
 using a 10-cm (4-in) diameter weighted drag sphere connected by a thin wire to
 a cantelever beam.  With the buoyant weight and the drag coefficient  (1)  known
 (the latter is  essentially constant  for  the  anticipated range of  Reynolds
 number),  it is  possible  in principle to deduce the velocity at the  sphere by
 measuring the  vertical deflection  of  the beam and the radial  and tangential
 deflections  of  the  wire.    In practice  it  was  difficult to make  these
 measurements  accurately  because of severe  eddying  in  the aeration  basin.
 Nevertheless it  was  possible to  ascertain even with rough measurements  that,
 about two  feet  below  the surface  across  the  area  designated  for  meter
 installation,  there  were no "dead" spots in the motion,  there  is a downward
 component to the average velocity,  and the estimated average velocities  range
 from about 45 to 75 cms (1.5 to 2.4 ft) per second, exceeding  the minimum values
 recommended by the manufacturers.

      The  test area was also surveyed with  a portable dissolved oxygen meter as
 part  of the pre-test  investigation.  No "dead zones"  or severe discontinuities
were  observed along the proposed  test locations.  However, it was apparent that
 there were substantial differences in d.o. within the  test area (see Figure 4
 for  later examples)  and  that  individual reference readings  would have  to be
made  next  to each mounted  probe.
                                     218

-------
             Figure  2.   Photograph of test  site
         Paired Locations
A N
B 0
D K
E L
f $
G T
I P
J Q
                  («itbdr.»r.,
*S
_j_
i
— .1
i
-i
                     F  G  H  I  j   K  L  M  N 0
Probes installed approximately j  feet  below liquid line
                 Figure  3.   Meter  locations.

                                219

-------
REFERENCE MEASUREMENT DEVELOPMENT  IN LABORATORY AND FIELD

The Reference Meter

     The use of a portable dissolved oxygen meter as a transfer standard rather
than direct Winkler tests was decided on because of the  inaccuracies  inherent
in  sampling  in  the aeration basin and the difficulty in performing  accurate
Winkler  tests  directly on  mixed liquor  samples  (2).    The combined  errors
incurred in sampling, sample handling, sample conditioning and analysis would
undoubtedly exceed those involved in the careful use of a transfer standard.  A
Yellow Springs  Instrument (YSI)  Model  54  portable dissolved oxygen meter was
selected as  the transfer instrument,  in  part  because of  its  common  use  in
treatment plants and also because it is not a direct competitor of the on-line
instruments participating in  the test.

     Before the start of the 60-day test,  investigations were conducted on the
YSI reference meter to supplement the information provided by the manufacturer
on  estimated measurement errors.   The  temperature measurement  was  checked to
see that it was within specifications.   The user's manual indicated that during
measurements the probe should be  oscillated axially in the sample at a  velocity
of  about 30 cms  (1 ft) per second to provide adequate  oxygen transfer.  It was
noted  while  experimenting with the probe that the reading  would  increase by
about  0.1 mg per liter  if the  probe motion was increased to about 60 to 90 cms
 (2  to 3  ft)  per  second.   During  the tests we attempted  to calibrate  the
reference meter  using probe velocities  close to those in the aeration basin.
Nevertheless a  potential error source  exists here.

     During  preliminary  site tests,   it  often appeared   that  the   reading
depended upon whether the probe  was held  stationary or  slowly  oscillated,  in
spite  of  the   adequacy of the   tank   currents.   More surprising  was  the
occasionally observed reduction  rather  than  increase  in  the reading  upon the
start  of  oscillations.   It was  conjectured  that  air bubbles  trapped in the
cylindrical head guard of the  YSI probe were flushed out  by  slow oscillations.
This theory was at least partially confirmed in laboratory  tests conducted in
a  small  water container through which air  was  bubbled.   Therefore,  it  is
recommended that  (at least  for the YSI  probe)  the probe  be  oscillated during
measurement regardless  of the existence of apparent strong mixing currents,
and that the head  guard be  removed during such tests.

The Calibration Sample

     Available  information  strongly suggests  that it  is preferable  to cali-
brate a reference meter using clarifier effluent so that any  effort of salts or
oxidizing agents in the liquid will be  taken into account  in the  calibration
(3,4).  A question remained  as  to whether the clarifier effluent could be used
as  drawn  or  whether it should  be aerated  to saturation.   To resolve this
question, the d.o. stability of an effluent  sample taken at  the overflow weir
was monitored  with an  air-calibrated  probe.   The  d.o.  (initially  about 4
mg/liter) was seen to rise in a very few minutes, apparently owing mainly  to the
rapid oxygen transfer across the  air-liquid interface encouraged by the probe
motion during the  readings.   When  the test was repeated on an air-saturated
sample, the  readings  were  constant  for  almost one  hour.    Therefore it was


                                   220

-------
apparent that errors introduced in handling a relatively  low d.o. sample to be
used for calibration were  potentially  larger than those  introduced  (owing to
possible non-linearity  in  the meter)  in measuring relatively  low  d.o.  values
with an instrument  calibrated at  a  saturation value.

Winkler Tests

     The pre-test period was also used  to train the involved NBS staff in the
laboratory  technique needed  for the Winkler (azide modification)  test.   Many
tests were run on distilled water which was aerated with a  magnetic stirrer.  It
was noted that the Winkler result was  invariably lower, by amounts up  to  3 to
4  percent,  than  the standard tabular  (5) saturation values.   Procedures  were
reviewed and reagents checked and replaced,  but the results remained lower; and
it was concluded that magnetic stirring for  20 to 40 minutes under the existing
ambient conditions  did  not saturate the sample to its  theoretical limit.

Error Estimates  for Reference Readings

     In order to compare fairly the  readings of the manufacturers'  meters  with
the reference readings, it is necessary to have an estimate of the  accuracy of
the latter.  Potential error sources include errors inherent in the instrument
itself, errors due to calibration  and drift, and errors unique to this  type of
test.

     Errors inherent in the instrument,  stemming mainly from non-linearity and
temperature compensation, were estimated from information furnished in  the YSI
manual to total (quadratically,  i.e.,  square root of sum of squares) about  0-12
mg per liter.

     The potential  errors  in calibration include  imprecision in the  Winkler
average  (estimated at  0.05 mg/1)  and the previously  mentioned  effort  of
velocity on the membrane.  The total quadratically summed calibration error is
estimated to be  not larger than 0.12 mg/liter.

     During the  field tests any  reference meter drift, whatever its cause, was
monitored by  checking  the  calibration after the  reference measurements  were
completed.  The  change  rarely exceeded  0.1  mg/liter and  usually was smaller.

     The error sources specific to this type of test stem from spatial  (Figure
4) and  temporal variations  of  dissolved oxygen  in  the  aeration tank.    In
practice the reference  probe had  to be a small distance,  about 30  cms  (1  ft),
away  from  the mounted  probe.    The  resulting uncertainty was assigned  an
estimated value  of  0.05 mg/liter.  Also,  at a fixed point, dissolved oxygen
fluctuations  of  varying amplitude  and period were registered on  both  the
mounted and the  reference  meter.   The magnitude  of the  resulting inaccuracy
depends on the success  of the observer in determining a  meaningful  reading for
both meters.  To attempt to quantify the error, instantaneous  readings at 10-
second intervals were  made  on  the reference  and mounted meters at  several
stations.    Figure 5   shows  one  example.     Some stations  showed  smaller
variations, but it is  clear that a single observer making visual readings  over
a  limited time introduces  an uncertainty which was estimated  to be  about 0.1
mg/liter.

                                    221

-------
0>

E
c
0)
0>
>«
X
o
a>
_>

o
40
in
4 1 	 r
po
, 1
1
21 	
1
1
r
4 »

•^ai
1 - '

r
4
i

-- t i
1
( 1








T
41
r—
L-
-, 1
3 rn
L_

< i , ,
- < i

<
< i - —

> '


1 *


	 1
O

1 1
IO/I6/8C
i
L
air-
1 -i
1
-1
i
i

i

9/2


L
I — -
A. ' < '

" O

it <

>

i| |

1
t


1-
	
1-
4!"

!


i> *


<>•


( •<

i
>

t


• n
J/80 -1
i

i
H
i

i •
9/(
i
S/80
i
• 1
1

j 1 |-j
IO/5/8C


) -i
i
1
i
AB DE      G  IJ    KL NO

           Meter  locations
                                           Q   T
      Figure 4.  Examples of  reference readings.
         •o—*• Meter B

         -+-+-YSI ot sio.B
                                     "*-*;
      01      234567


                  Time,  minutes



Figure  5.   Example of D.O.  fluctuations at a point.
                         222

-------
     The errors estimated in this section are summarized in Table 1.  The way
in which they are combined may vary in specific tests and should be agreed upon
at the start.  However, the likelihood of all errors being in the same direction
appears remote, so that an arithmetic sum may not be realistic.  A quadratic
summation, which gives a combined error of about 0.23 mg/liter, was used for
this test.

                      TABLE 1.   ERROR ESTIMATES  FOR  THE
                                REFERENCE MEASUREMENTS

Source
Instrument
Calibration
Drift
Probe Distance
Fluctuations
Error, mg/1
0.12
0.12
0.10
0.05
0.10

THE FIELD TESTS

Test Meters

     Thirteen meters were installed by seven manufacturers (one did not have a
second demonstrator available) in the  locations shown in Figure 3.  All of the
probes  used electrochemical  cells separated  from the  sample liquid  by  a
membrane permeable to oxygen.  To the  extent that descriptions were available
in the user manuals, both passive and  active cells were represented as well as
different electrode  materials and electrolytes.   Those  membranes for which
information was given were 2 mils thick, except for  one which was 0.1 mm thick.
(Membrane thickness  will affect  response  time.)    One pair  of  probes  used
"permanent" membranes.  All of the probes except for one pair were basically
cylindrical shapes (with the membrane  and cell at one end) to be immersed di-
rectly in the ambient liquid  flow.  The remaining pair  used a  continuous flow
of tap water to induce  sample flow through  a  short pipe which contained the
sensor cell in its wall.  The  participants of the following manufacturers in a
test  that  did not  involve a procurement  action  is  sincerely appreciated:
Beckman Instruments,  Incorporated  (Model 7002); Envirotech Corporation, Nat-
ional Sonics Division, Delta  Scientific  (Series 8210/8310); Fischer & Porter
(Series 17D01000); Leeds & Northrup (No. 7931); Orbisphere Laboratories (Model
2716); Rexnord Instrument Products (Model 62 Probe, 3000-5 Analyzer); Uniloc
Division of Rosemount,  Incorporated (Model 424).

Procedure for the Daily Tests

     A fresh 2000 mm sample of clarified effluent was  aerated  in a beaker for
20 to 30 minutes with a magnetic stirrer.   Three  300-ml B.O.D. bottles were
filled with this  aerated  sample  for Winkler tests, and  the remainder of the
sample was retained for  meter  calibration and the post-test calibration check.
                                   223

-------
An effort was made  to  keep  the  sample temperature constant.  The arithmetic
average of the Winkler analyses was used to calibrate the reference meter.

     The calibrated meter was carried to the nearby test site and the reference
probe was immersed next to the on-line probe mounted at station A.  A mixed
liquor temperature reading was made at this point.  Care was taken to assure
that the reference meter had attained a steady state before meter readings were
recorded.  This precaution had  to  be  observed  at  all stations, although the
stabilization time was  considerably shorter in cases where the reference probe
did  not  have to  be removed from  the liquid when changing stations.   The
reference readings at each station were made over a period of about one minute,
depending upon their steadiness.   The complete set of reference measurements
was  usually  completed within 25 minutes.   The reference probe was  then rinsed
and  the meter was  carried back  to  the laboratory, where its calibration was
checked in the original aerated effluent sample.

Auxiliary Measurement  for Quality  Assurance

     At random during  the test  period, the reference meter was carefully air
calibrated according to the manufacturer's instructions and its  reading  in the
aerated effluent sample was noted prior to adjusting the calibration setting to
conform to the Winkler  average.  Also, at  random a  one Winkler test was made on
distilled water aerated to saturation, and an air-calibrated probe reading was
then taken in the  same  sample.  These tests were  designed  to monitor the state
of  the Winkler procedures and reagents as well  as  the condition of the probe.

     Both of the foregoing sets  of auxiliary measurements  formed a body of data
on the efficacy of air calibrations. Examples of the measurements with an air-
calibrated probe in the clarifier effluent are shown in Table 2. The results of
28 measurements showed the probe readings exceeding  the Winkler values by an
average  of  4 percent +_ 2 percent  standard deviation.   Examples of the air-
calibrated probe measurements in aerated  distilled water are given in Table 3.
The  results  of  33 air  calibrations on 20 distilled water  samples showed the
probe readings exceeding the Winkler values by an average  of 1  percent +_ 1.4
percent standard deviation.   It  should be noted  that these auxiliary tests were
somewhat handicapped by  lack of a.  temperature controlled environment.


            TABLE  2. EXAMPLES  OF AIR-CALIBRATED PROBE VS.  WINKLER
                     IN  SATURATED  CLARIFIER EFFLUENT
            Test             Probe         Winkler          Ratio
            Date	(mg/1)	(mg/1)	(Pr/Wi)

          10-20-80           7.68           7.54           1.019
          10-14-80           8,90           8.67           1.027
          10-06-80           8.52           7.85           1.085
           9-26-80           8.45           8.35           1.012
           9-22-80           8.20           7.87           1.042
           9-09-80           7.12           6.64           1.072
                                     224

-------
          TABLE 3.  EXAMPLES OF READINGS  IN AERATED DISTILLED WATER
Test Winkler
Date (mg/1)
9-26-80
9-23-80
9-29-80
9-30-80

10-15-80 '

8-21-80
8.64
7.59
8.77
8.74

8.96

8.04
Std. Sat.(*)
(mg/1)
8.78
8.00
9.04
9.07

9.18

8.39
Winkler
Std. Sat.
.984
.949
.970
.964

.976

.958
Probe
Std. Sat.
.986
.974
.973
.980
.996
.967(**)
.985
.947
Probe
Winkler
1.002
1.026
1.003
1.017
1.033
0.991
1.009
0.989
 (*)
 (**)  Probe tests repeated when air calibration did not check after reading
 in distilled water
 Typical Test Data

     Figure  6  shows the  entire 60-day record  of operator readings for  one
 meter.  Also shown are the two points at which this meter was re-spanned  (re-
 calibrated) downward when its deviations from the reference readings  began to
 exceed about twice the estimated reference meter error (0.4 or 0.5 mg/liter).
 This performance typified that  of many  of  the meters.

     Figure 7 shows details of a six-day portion of the record for two meters,
 B and 0,  from the same manufacturer but  located about 17 m (55 ft) apart.   The
 expected tendency toward a  diurnal d.o. variation is visible here as well as
 the increase in average d.o. at the downstream station.  The two sets  of values
 are seen to be  in phase and,  in this  case, about  1.0 mg/1  apart.

     Figure 8 shows examples  of  the  60-day record of differences between  the
 daily  reference  readings  and   the  simultaneous meter readings,   positive
 deviations  indicating that  the meter  value  was high.  Longer term trends
 estimated from five-day averages  are also shown.  The trends  are  interesting in
 that occasionally  (see K,  for example) the deviations from reference values
 would recede without adjustment  after a period  of larger excursions.  In  the
 case of meter  L,  some of the excursions are due  in part to relatively large
 oscillations in its readings.  The user manual describes a cell modification to
 reduce these transients.

Maintenance Summary

     Each meter was  calibrated by  the manufacturer's representatives at  the
 time of  installation,  and the  project  supervisior  was asked to re-span  the
meters if necessary to agree with the reference reading on the first day of the
 test.    Oral  instructions  further  suggested  that  the  probes  need not be
routinely removed from the tank unless a need was indicated by  poor readings.
This policy  was followed,  except  for  meter  I,  for which the membrane  was
cleaned twice to see if there was any effect.   There was none.

                                     225

-------
D.O.
mg/1
10


9



8


7


6


5


4


3


2
           I     1
1      I
                               SIXTY-DAT MCOROFOR MITCH A
                          20
                                   30
                                  Days
                                                40
                           50
         Figure 6.   Example of  a 60-day  meter  record.
                 6

                 5

                |4

                o3
                o
                 2

                 \

                 0
                    Miter '8'
                                                                       60
                       Meter '0
                  12M  N  12M H   12H N 12M  N  12M N  12M  N  12M

                  | Oct 7  | Oct 8 |  I0ct 9 JQct 10 | Oct 1] | Oct 12 |


     Figure  7.   Six-day record  for two  meters of the same make.
                 + =  reference reading.
                                    226

-------
        _ METER 8


              I
                    so
                                               MEIER K
                                                 I 	I
                                                        SO
                                                                    60
 E     '
 o>
 o>
 o

 O>

 O)
»*—

 e


•O

 O
 1
to
 0)
 U
 o>
•»-
•^

O
        - METER C


              J
      ^1    ii
                    30
                                CO
       [_ METER I
                1    1
            ii
                    90
                                60
          METEM
                    SO
                  Doys
                                              METER
                                                                     60
                                                       Doys
             Figure 8.   Examples  of daily deviations.
                                     227

-------
      The maintenance records are summarized  in Table 4.  It  is seen that of the
 13  meters, the  servicing required by  nine of  them involved no  more  than
 correction of the initial calibration.  Of these nine meters, five were not re-
 spanned after the first two days, while three others were not re-spanned after
 the  first  two weeks.   Given  the  pre-test calibration methods employed  by the
 manufacturers and  the  lack of acclimation time,  some recalibration  should be
 expected.  Of the  four meters that required more than recalibration, the two
 using induced sample  flow required considerable attention and the  remaining
 pair required an intermediate level of maintenance, including a  replacement of
 membrane and electrolyte.

      It should be  noted  for  information purposes that,  prior to the start of
 the  60-day period, the submerged wire leading to one of the  probes collected a
 large ball of  stringy material  and the subsequent buffeting pulled the  wire
 loose.  The unit  was repaired by the manufacturer, who remedied  the problem by
 taping the wire  directly to  the  probe support structure.

                        TABLE 4.   MAINTENANCE SUMMARY


     Manufacturer    Meter Position                  Maintenance

         1                  A             Re-span  on day 15, 27
         1                  N             Re-span  on day 1.

         2                  b             Re-span  on day 1, 15
         2                  0             Re-span  on day 15

         3                  D             Re-span  on day 1
         3                  K             Re-span  on day 1

         4                  E             Re-span  on day 1, 15, 45, 56;
                                          membrane  and  electrolyte  replaced
         4                  L             Re-span  on day 1, 2, 51;  suspected
                                          failure  day 60 (no cell replacement)
         6                  G             Re-span  on day 15
         6                  T             Re-span  on day 1, 2

         7                   I             No servicing  required

         8                   J             Re-span  on day 1, 2.  Frequent probe
                                          removal:  days 14, 22, 27, 43,  50,
                                          51, 56,  59.   Probe moved  toward
                                          upward current on day 35, 50.
         8                   Q             Re-span  on day 1, 2.  Frequent
                                          probe  removal:   days 4, 8,  22, 27,
                                          50, 59.   Probe moved toward  upward
                                          current  on day 35, 50-


NOTE:  Manufacturer 5 withdrew prior to the  test.
                                     228

-------
CONCLUSIONS AND RECOMMENDATIONS

Calibrating the Reference Meter

     The  use  of  a  portable  transfer  standard or reference  meter that  is
calibrated  in Winkler-tested saturated  clarifier  effluent  appears to be  an
accurate  and  controllable method  for  d.o.  measurements in activated  sludge
aeration  basins.   The  results  in Table 2 show that in this  case  accuracy was
improved  by using  the Winkler method to analyze the calibration solution, and
it  is  recommended  for  acceptance tests or any  applications in which maximum
accuracy  is  important.   For  routine  periodic performance monitoring, the
simpler and faster air calibration is preferred provided preliminary investi-
gations  indicate   that  the  difference  between an  air  calibrated probe and
Winkler results on clarifier effluent  is not large.  Caution must be taken  to
assure that ingredients in the plant inflow will not intermittently upset this
difference.   Air calibrations  should be made  in moisture saturated air.  Our
experience  here  also  suggests  that  the  probe   should be rinsed  and the
calibration checked  after use.

     Another  option involves  calibrating the  reference meter  in  distilled
water  that  has  been aerated  to  apparent saturation  and using  the standard
saturation  values  for dissolved  oxygen content.   This  method  presumably
introduced  errors  similar to  those  for air  calibration (Table  2) and will
require the same preliminary test when  applied to effluent.   On the other hand,
the  auxiliary tests  also indicated  (Table  3) that aeration with  a magnetic
stirrer for 20 to 30 minutes, as  is  frequently suggested in instructions, does
not  necessarily yield  the tabular  saturation value.   This effect  should  be
investigated  if the user wishes to calibrate in this way.  It is further noted
in  regard to  both  air  and water  calibrations that  slightly different  (up  to
about  1 percent) tabular  saturation values  can be  found in  the  literature.

Using  the Reference  Meter

     As discussed  under reference measurement development,  in  order to avoid
momentary entrapment of air bubbles, it is recommended that  the  portable probe
be  slowly oscillated during the readings even if  the  sample velocities are
otherwise adequate.

     After  the  reference probe  is  first immersed in the sample,  it must  be
allowed enough time  (determined by monitoring the meter readings) to attain a
steady  state.  The  subsequent reference readings should  be continued long
enough to obtain a representative average.

     The  reference  meter  calibration  should be checked  after  the test
measurements  are completed.  This precaution  is recommended not only to check
for drift but also to guard against the results  of  inadvertent jostling of the
meter or  accidental  movement of  the  control knobs.

Calibrating the On-Line Meters

     As indicated by the patterns in Figure  8, it is usually not meaningful  to
make adjustments   based  on  the  first  day  or  two of  readings, unless the

                                     229

-------
 deviations  are substantially larger than the reference-reading  error deter-
 mined from  estimates  equivalent  to those  in Table  1.   Rather,  the deviations
 should be observed for at  least  several days while the meter equilibrates so
 that averages  or  trends can be discerned  before  recalibration.   This applies
 not  only to the initial test period but also throughout the  test.

      If the dissolved oxygen in  the aeration tank  varies during  the day in a
 quasi-periodic manner, the  pattern should be noted and the specific times of
 the  daily reference readings should be varied so as to cover as  large a d.o.
 range as possible for the  performance evaluation.

 Recommendations Related to  Probes  and Probe Mounting

      The aeration basin of  the activated  sludge  process will contain stringy
 fiberous material  (Figure  9)  as well as  biological  solids  mixing  at  high
 velocity.   The  process  and most  equipment  related  to   it  are  outside,
 unprotected from  the  weather.   Maintenance and operational personnel  must
 often  work under  unpleasant performance  limiting  conditions.   At  large
 treatment facilities   repair  shops  and  laboratories may  be  several  hundred
 meters from operating equipment.

      The following recommendations  are designed  to reduce failures  caused by
 the  above conditions,  and improve the efficiency of maintenance personnel when
 repairs are required:

      •  The probe, and the submerged support structure,  should  be  shaped to
         present a smooth, straight cylindrical outline to flow.  That is, sharp
         right  angle turns, abrupt enlargements, and other geometries conducive
         to  the collection of material should be avoided.  It is preferable that
         all wiring be  confined inside the mounting elements.

      •  The sensor should be exposed directly to  the aeration basin.  That is,
         no  sample-flow induction or agitation should be used to conduct liquid
         sample past the membrane.    This  does  not  apply where the  sample of
         interest  is in a quiescent  state.

      •  Mechanical devices  used to hold the probe assembly in place should be
         the quick release variety.   That  is, hand tools  such as wrenches or
         screwdrivers should not be required to  remove the probe assembly from
         the process.    Further,  fastening  devices  used to  hold the  probe
         assembly  in place should  be  retained mechanically  to prevent them from
         being  dropped  into  the basin during maintenance.

      •   The  design should facilitate easy,  fast  replacement  of the  membrane.

      •   Directly  exposed sensors  should  be oriented  so that the  strongest
         currents do not  impact perpendicularly to  the  membrane.

     The above listed  recommendations are  primarily developed for shallow (1-
2 meters deep)  installations.  Where greater  submergence  is required the design
engineer may not be able to  follow each point but  should at least consider them
for the  specific application.

                                     230

-------
Figure 9.  Example of fouled probe,
                 231

-------
 Operation and Maintenance  Recommendations

      The  directly  exposed sensors  in  this  test  were  all  oriented  either
 vertically downward or horizontally (or  nearly so).  Mounting these units in a
 flow with a  general  downward component proved  to be a generally  effective
 procedure,  since  only one  case  of  apparent  membrane  rupture  was  experienced.

      It is not possible to  recommend membrane/electrolyte change intervals for
 preventive maintenance, because there were not enough cell failures within the
 60-day period  to warrant  such conclusions  (see  Table 4).    In any  case,
 maintenance  frequency will be  impacted  by  the  specific site  and should  be
 designed  from data produced on-site.

      A daily check with a  calibrated portable meter is recommended.   Membrane
 and general probe cleaning is not recommended,  unless the daily  comparisons
 reveal a  need for more frequent servicing.

 Test Protocol Modifications

      A data recording or logging system would have been useful in this test and
 is recommended for future tests of the same type.  There were several instances
 in the manual logging of  the four-hour readings where it was clear that  an
 incorrect scale  was read by  the operator,  even  though the proper scale  was
 marked on each meter.

      This test considered  only  the  quasi-steady performance of  the meters.
 Development  of information on unsteady-state  response  under  field conditions
 is necessary  for  process  control  applications.    Of particular  interest  is
 information on effects of  cell and membrane  aging and of  membrane  cleaning on
 response  time.   It is recommended  that appropriate  field  test methods  be
 developed.

 General Comment

      Field tests are designed to measure the performance of a  device in a given
 environment.   Analysis of  basic design  criteria  such as  component selection,
 construction  material, workmanship, or measurement methodology is not part of
 such tests.  Therefore, no systematic effect  of sensor details (i.e., electrode
 materials, electrolyte chemical, active  vs.  passive cell) can be inferred from
 these  tests.   However, the utility  of  such tests  for  prequalification  of
 bidders is exhibited by 0  and L  in Figure 8; both are passive cells using the
 same materials, yet there  is  a  substantial  difference  in performance.

     An organization wishing to  conduct  field tests  of  on-line  dissolved
oxygen  analyzers  can  utilize the protocol given in  the  appendix  provided
cognizance is taken of the  modifications in technical details described above.
In addition, administrative details must be  changed to recognize the specific
organizations and  locality involved in  the  test.
                                     232

-------
REFERENCES
1.  Schlichting, H., "Boundary Layer Theory," MqGraw  Hill, N.Y.,  16  (1955).
2.  Mentick, A.F. et al., "Investigation  of  a Honeywell Dissolved Oxygen
    Parametric System," EPA Rep.  600/4-77-023  (April,  1977).
3.  Mancy, K.H. et al., "A Galvanic Cell  Oxygen Analyzer," J.  Electroanal.
    Chem. 4., 65 (1962).
4.  "Methods for Chemical Analysis of Water  and Wastes",  EPA  600/4-79-020,
    U.S. EPA, EMSL, 360.1-2  (March, 1979).
5.  "Standard Methods for the Examination of Water  and Wastewater,"  13th  ed.,
    APHA, AWWA, WPCF (1970).
DISCLAIMER

     This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
                                      233

-------
                                   APPENDIX

                A FIELD EVALUATION OF  ON-LINE  DISSOLVED OXYGEN
                     METERS  FOR SEWAGE  TREATMENT PLANTS
 1.       General

 1.1      This  program  is  proposed  for  the evaluation  of on-line  dissolved
         oxygen  monitors under  controlled  conditions  in  a  field  (aeration
         basin) environment.   Its  purpose  is to:

              -   Provide information for the development of specifications for
                 instrument procurement at  wastewater treatment  plants and for
                 acceptance testing;

                 Provide information for development of accurate testing meth-
                 ods  at wastewater treatment plants; and

              -   Provide information on routine maintenance requirements under
                 a  specific set of operating conditions.

 1.2      This  document briefly describes the proposed evaluation and lists the
         participants and their respective responsibilities.

 1.3      Participants

 1.3.1    The  National Bureau of Standards (NBS)  will  provide project  super-
         vision,  test procedures,  reference methods or  instruments,  and  analy-
         sis of results.

 1.3.2    NBS  is  conducting  this evaluation as  part  of a  larger-scope  inter-
         agency agreement with the  Environmental Protection Agency's  Municipal
         Environmental  Research Laboratory (EPA-MERL), dealing with perform-
         ance  specifications  and   test methods  for selected  flow and  water-
         quality measuring instruments  in wastewater treatment plants. Program
         manager  for  NBS is  Dr.  George  Mattingly, Chief, Fluid Engineering
         Division,  NBS; program  monitor  for  EPA-MERL is  Mr.  Walter  Schuk.
         Immediate  supervision for the dissolved-oxygen meter evaluation pro-
         ject will  be provided by  Dr. G. Kulin, NBS.

 1.3.3    The Washington  Suburban  Sanitary  Commission (WSSC) will furnish the
         aeration-basin  location  and specific  support  services described  in
         this document.  The WSSC contact for this program is  Dr.  Stanley Dea,
         Director,  Bureau of Design.

 1.3.4   Manufacturers of dissolved-oxygen  meters,  solicited in accordance with
         paragraph 2,  will furnish  the set up instruments for this evaluation as
        described  in paragraph 4.
2.       Test Instruments
                                     234

-------
2.1     Solicitation.  Known manufacturers of dissolved-oxygen meters for the
        wastewater  trade  have been contacted by EPA-MERL representatives to
        ascertain their interest  in participating  in  this evaluation.  Also,
        the program has been announced in the Commerce Business Daily for the
        benefit of  manufacturers  who may not be known to EPA-MERL and NBS.

2.2     Exceptions.  The agencies conducting this evaluation are committed to
        fairness to all manufacturers.   However, in  the interest of orderly
        investigation,  the project supervisor  reserves  the right to reject
        offered  instruments  which do not have  a demonstrable potential for
        activated-sludge application or which,  for  technical  reasons, cannot
        be accommodated in this evaluation.

2.3     Meters.  Manufacturers are asked to  furnish two identical instrument
        systems, including panel-meter  (or equivalent) readouts and  instruc-
        tion  manuals.   Only  one  pair of meters will be  accepted from each
        source  unless  the  manufacturer  makes  a  second  instrument of  a
        generically different type.

3.      General Test Conditions

3.1     Location.   The evaluations will be conducted  in an  aeration basin at
        the WSSC Seneca  Interim Wastewater  Reclamation Plant on Riffle Ford
        Road,  near Gaithersburg,  Maryland.    This plant uses  the  extended
        aeration process and  the basin typically has  4,000 to 5,000 mg/1 solids
        and 2  to 4 mg/1 dissolved oxygen.  The  project supervisor will assign
        each  instrument a  specific test location within the basin.  In doing
        this,  the  project supervisor will attempt  to select locations such
        that all probes will be subjected  to  conditions that are sufficiently
        similar for the purposes of this test.   Nevertheless,  the probability
        of reasonable differences  in small-scale environment is recognized and
        will  be  allowed for during the tests  by appropriate monitoring and
        probe  relocation if  necessary.

3.2     Duration.  The length of the  evaluation period will be 60 days. All of
        the submitted  instruments will be evaluated  during the same 60-day
        period.  If unforeseeable circumstances  should cause  an  interruption
        of the program during this period,  the  conditions  for  (and the length
        of) the resumed program will  be determined  by  the  project supervisor.

4.      Instrument  installation

4.1     Location.  The sensing probes will be  mounted at the location described
        in paragraph 3.1.

4.2     On-site provisions.  A standard handrail is available for mounting the
        equipment.   The   attached drawing   gives  pertinent  dimensions  and
        related  information  which enables the  manufacturers to arrange for
        attaching their probes and analyzers to  the supports. WSSC will also
        furnish two 110-volt outlets within  6 feet  of each  analyzer  station.

4.3     Installation assistance.  In general, manufacturers will be  expected
        to provide  the  materials, fittings  and  manpower  necessary  to mount
                                     235

-------
         their  equipment  on  the railing  in accordance with  paragraph 4.2  Any
         additional  material or manpower assistance  is  expected to  be  rela-
         tively minor and should  be  arranged in advance with WSSC.  To  the
         extent that  unusual requirements for modifications  or assistance  can
         be  anticipated by the vendor well  in advance, such  things  should be
         made known  to the project supervisor at the time of the  agreement to
         participate.

 4.4      Installation period.   A  period of  two weeks will be  set aside  during
         which  manufacturers may  set  up their instruments  and do all  of the
         necessary  preliminary work on  them.   Although the manufacturers may
         come in at any  time  during this two-week period,  they will be requested
         to  notify  WSSC (Russel Sharp, Seneca Plant Superintendent,  telephone
         301-428-3117) of their specific installation time in advance.  The  two-
         week period will  not  begin earlier than June 15.

 5.       The Evaluation Tests

 5.1      Start.  The 60-day  evaluation period will begin no  later than  5  days
         after  the end of  the  two-week installation period.

 5.2      General test conditions.  See paragraph 3.

 5.3      Final   adjustments.    At  the start  of the  60-day  test period,   the
         manufacturers  will be  allowed to  make  final  adjustments  on their
         instruments if  they  so  desire.    Further adjustments  will  not be
         permitted during  the test period except as  noted  in paragraphs 5.4 and
         5.7

 5.4      Maintenance.   During  the test  period,  routine maintenance will be
         provided by WSSC  operators.  The level of maintenance tasks which can be
         considered   "routine"  for each instrument  will  be  agreed upon in
         discussions  among  NBS,  EPA, WSSC  and the  individual  manufacturers
         during the  two-week setup period.  This period may  also be used  by the
         manufacturers for communicating information and  instructions on their
         instruments  to  WSSC  operators  and  NBS  personnel.   Maintenance by
         operators will be performed only during the day shift, and then only as
         recommended  and  witnessed by the   project supervisor or his  repre-
         sentative.   All  maintenance  that is  done will  be  logged  in detial.
         Maintenance  requiring the services  of instrumentation specialists  will
         not be provided during the test period.

5.5      Readings.

5.5.1   Manufacturers meters.  Readings with a reference meter (YSI Model 54)
        will be made by NBS personnel at least daily.  Calibration  method and
        calibration  frequency for this meter and/or guidelines  for  the  use of
        other  reference  methods will have  been  determined  as  described in
        paragraph 5.6.

5.6     Reference measurements.
                                    236

-------
5.6.1   NBS will make reference measurements with  a  portable dissolved-oxygen
        meter as  noted  in 5.5.2.   Methods  of calibrating this meter will be
        investigated but at this  time  it is anticipated that laboratory Winkler
        tests on clarifier effluent  will be used. Prior to  the start of  the 60-
        day test period  NBS personnel  will also evaluate the reference meter(s)
        to obtain information on error band, necessary calibration  frequency,
        methods for assuring adequate agitation, etc.

5.6.2   NBS personnel will monitor the use and calibration of the reference
        meter during  the  field  evaluation and  will be responsible  for its
        maintenance.

5.7     Instrument failure.

5.7.1   For the  purpose of this evaluation,  "failure" or  "malfunction" will
        include not only the obvious cases of  continuous  zero or continuous
        full-scale  readings but  also any  continuous  readings which  are  so
        constant as to constitute obvious cases of complete  lack of response to
        change.

5.7.2   If an instrument should fail or malfunction  during the test as described
        in  paragraph 5.7.1,  WSSC will  take  only  those  steps  which  are
        specifically agreed upon prior to the start of the testing and which can
        be accomplished by operators, and/or  those  steps   that  are normally
        taken by operators  in  these  cases,  such  as  examination  for  probe
        fouling, etc.   Repairs by  operators will be  made  only  during the day
        shift,  and  then  only  as  recommended  and witnessed  by the  project
        supervisor or his representative.  All repairs will be logged  in detail.
        Repairs requiring the services of instrumentation specialists will not
        be made during  the test period.

5.7.3   In the  event  of instrument failures or  malfunctions which cannot  be
        remedied by operators, the project supervisor will decide on a case-by-
        case basis  whether a particular  instrument  should be  withdrawn from
        further participation in the  tests or whether  the manufacturer should
        be asked to repair/replace  it.   These  decisions will attempt to take
        into account both the need  for useful test data and the commitment to
        fairness to all manufacturers.  In cases where replacement is permit-
        ted,  only an identical model of the  instrument  will be acceptable.  In
        any event, manufacturers  will  be asked to determine the cause of failure
        and notify the  project supervisor for information  purposes.

5.7.4   If the  failure   of  an instrument  is  due  to an obvious  accident  or
        incident  beyond the control  of  the  manufacturer and not  normally
        encountered in  activated sludge basin,  repair or  replacement  of the
        instrument and  continuation of its  evaluation  will be  permitted, but
        the final determination will be made by  the project supervisor.
                                    237

-------
THE CALIBRATION AND EVALUATION OF DISSOLVED OXYGEN SENSORS IN A PILOT SCALE
ACTIVATED SLUDGE PLANT

G. Speirs, D. Chapman, E. Luxon and J. Matthews
Wastewater Technology Centre, Environmental Protection Service
Burlington, Ontario  L7R 4AD
ABSTRACT

     The performance of three types of dissolved oxygen (DO) sensors was
evaluated in a pilot scale activated sludge plant.  As the instruments were
interfaced to a minicomputer, a curve fitting technique was developed to
compensate for instrument non-linearity.  Weighted regression was performed
to update initial calibrations.

     Fouling due to the microbial growth on the surface of the probe was the
principal cause of decline in sensor performance with time.  The rate of
fouling depended on the type of sensor used, the mode of operation, the fre-
quency of maintenance and the velocity of fluid past the probe.  A number of
alternatives were evaluated for reducing or compensating for the effects of
fouling and other causes of inaccuracy.

     To determine the suitability of sensors for tracking DO concentration
during non-steady state tests for determining overall mass transfer coeffi-
cients (ICa), the probes were subjected to repeated step changes in DO
concentration.  First-order time constants (T) for the sensors were estimated
based on the results of each step.
INTRODUCTION

     For aerobic wastewater treatment systems it is desirable to obtain, on a
continuous basis, an accurate indication of dissolved oxygen concentration
in the aeration compartment.  Electrochemical dissolved oxygen monitors are
readily available and used routinely to aid in the maintenance of a minimum
concentration of dissolved oxygen necessary to insure the viability of the
microbial population, to allow blowers and aerators to be controlled to
provide adequate but not excessive aeration and, finally, to permit aeration
equipment to be evaluated under field and laboratory conditions.  Experience
gained in the continuous on-line monitoring of dissolved oxygen concentration
with different makes of sensors resulted in the development and evaluation of
methods for improving sensor performance.  Factors influencing the suit-
ability and performance of DO sensors included installation, calibration,
maintenance and the dynamics of the sensor in responding to sudden changes
in DO level.
                                    238

-------
RESEARCH EQUIPMENT

Pilot Plant  Facility

     The research was carried out using an  extended aeration activated
sludge package plant modified and operated  in  the conventional mode.   Flow
to the plant,  controlled at  a rate of 200 m3/d,  consisted of degritted
municipal wastewater.  The aeration basin was  baffled into three  sections,
each with a  volume of 22.2 m3.   Air was supplied through coarse bubble
diffusers by a 5.6 kW positive displacement blower.   On-line process  instru-
ments interfaced to a real-time minicomputer provided monitoring  and  control
for the plant.

Dissolved Oxygen Sensors

     Dissolved oxygen sensors from three different manufacturers; Yellow
Springs Instrument Company  (YSI), pHOX Systems Ltd., and Zullig,  were used in
this research.  Each consisted of a probe and  an analyzer/transmitter which
displayed the indicated dissolved oxygen reading in milligrams dissolved
oxygen per liter.

     Commercial dissolved oxygen probes are electrochemical cells which
generally operate on one of  two principles.  Galvanic cells consist of a
noble metal  cathode, a base  metal anode and an electrolyte.  Chemical
reactions involving oxygen ojcur at the cathode  and are spontaneous,  causing
a measurable current flow in proportion to  the oxygen concentration present.
Polarographic cells operate  in a similar manner  but use two noble metal
metal electrodes and require a constant external voltage to polarize  the
electrodes before the necessary chemical reactions occur.  In some cases,  a
gas-permeable plastic membrane isolates the sensor elements and electrolyte
from foreign material and impurities in the sample.   Usually the  electrolyte
is potassium hydroxide, potassium chloride  or  potassium bicarbonate in
liquid or gel form.  For some makes of DO sensors the electrodes  are  in
direct contact with the sample stream which serves as the electrolyte.
Partial specifications for the three makes  used  in this research  are  listed
in Table 1 and each probe is shown in Fig.  1.
       Table 1.  List of instrument specifications
         Hake and Model
                      NuBber  lange*  Operating
                     Evaluated  («g/l)  Principle
              KBbraoe                 Temperature
              Type   Electrodea Electrolyte Coapenaatlon
                            0-20 Polarographic Teflon
       Model 5*A Trenailtter
       Model 5734 Probe

       pHOX
       Model 66 Trmnmltter
       Model 65-88 Trannltter
       Model 670 Probe

       Zullig
       Model P-76-2-1
                    Gold
                    Silver
0-20  Galvanic
0-20  Galvanic
             Polythene Sliver
                    Poroua Lead
ABalgaa
Zinc
         Liquid
         ttl
         Gel
         KHCOj
                                                                  Tea
                                      Tea
                                      Tea
       * Inatruacnte wre uaed at the 0-10 «g/l acale aettlng.
                                      239

-------
     The YSI  dissolved oxygen sensor is a portable unit  commonly used for
analytical procedures and discrete sampling at wastewater treatment plants.
A thin teflon membrane is stretched over the cell electrodes located at the
end of the probe.   The membrane and electrolyte are  easily replaced by the
user.  The unit  is ideal for comparing calibration of  other sensors and one
was used as a reference exclusively for this purpose.  A second unit was
used for continuous on-line monitoring in a dip mode,  however,  additional
electronic signal  isolation equipment was required to  produce acceptable
results.

     The pHOX probe is of robust construction and designed for  continuous
on-line monitoring in either a dip or flow-through mode.   A perforated,
cylindrical silver cathode of large surface area fully supports a tubular
polythene membrane.   The membrane is replaceable but the procedure requires
a great deal  of  care.  The cell is a plug-in unit which,  when exhausted,  is
rejuvenated by the manufacturer on an exchange basis.

     The Zullig  sensor uses a dip probe designed to  be self cleaning.   A
motor drives  an  electrode-cleaning grindstone and activates a beaker sur-
rounding the  electrodes to alternately draw and expel  a  fluid sample on a
continual basis.   Since the sample serves as the cell  electrolyte, this
action ensures that the sample is constantly renewed to  maintain conductiv-
ity.  The unit is  rugged and designed for continuous on-line use.   The
grindstone is the  only user-replaceable part and is  easily changed.
      Figure 1.  Dissolved oxygen probes:  i) YSI; a) protective shield, b) electrode
               head,  ii) pHOX; c) probe head, d) porous lead anode, c) silver, cylin-
               drical cathode,  iii) Zullig; f) electrode assembly, g) rotating grind-
               stone, h) cam-action draw and fill beaker assembly
Sensor/Computer  Interface

     All of the  DO  sensors,  with the exception of the  reference YSI were
interfaced to a  real-time minicomputer.  As shown schematically in Figure 2,
the DO concentration was  converted by the DO sensor to an analog signal

                                     240

-------
                    COMPONENTi
                      0.0. SENSOR
                                  •FRONT-END' MICROPROCESSOR
                                                       COMPUTER
ANALYZER
H
n
Ii 4-2B HA
SIGNAL HIRE
HP-2240

N> t-\t.m

INTERFACE BUS
HP-IBM
                           PROBE


                    FUNCTION^
                      •ad. MEASUREMENT
                      •a a /i CONVERSION
•MULTIPLEXING
•SIGNAL CONDITIONING
•ANALOG/DIGITAL CONVERSION
•DATA LOGGING
•DATA REDUCTION
•DATA PRESENTATION
•ALARMS
•PROCESS CONTROL
     Figure 2.  Components of computer monitoring system
 which was sampled,  conditioned and digitized by the front-end microprocessor.
 The digitized sensor  signal was converted  by computer software to appropriate
 engineering units.  As  the lowest value  for  the signal output from  the compu-
 ter was 20% of the  signal range, the computer could distinguish between an
 actual measured value of zero DO concentration - a "live zero" - and  a zero
 caused by the instrument being inadvertently switched off or by a damaged
 signal wire.  An alarm  condition (the DO concentration below a critical
 minimum) if detected  by the computer would sound a buzzer to alert  the plant
 operator and write  a  message describing  the  nature of the alarm to  a  termi-
 nal.

      Data collected from DO sensors were summarized and displayed using
 computer peripherals  including terminals (CRTs),  a line printer and a
 plotter.  For normal  data logging, the sensor signal was sampled at a five-
 second interval with  the data stored in  disc files as 15-minute averages.
 Figure 3 shows a plot of DO for three sensors,  each placed in one of  the
 compartments of the aeration basin.  For a constant airflow rate, the plot
 illustrates the effect  of diurnal loading  on DO concentration as well as the
 variation in DO for three complete mix reactors in series.
CALIBRATION PROCEDURES

Standard  Calibration Procedure

     Accurate calibration  of DO sensors is essential for reliable process
monitoring and control.  Although specific calibration procedures vary  with
make and  model of instrument, DO sensors are  generally calibrated by  ad-
justing the instrument  zero and span.
                                      241

-------
            IB

             B

             8

          ~   7


             6

             5

             4

             3

             2

             1

             0
uj
o
                     PILOT PLANT SOCMATIC
	a a SiotiOT A
	 0.0. S>o«i«n •
	 0.0. Scvtlan C
                  10   12   14  16  18  20   22

                                    TIME Chr>
      Figure 3.  Changes in DO concentration due to diurnal loadings
     Adjustment  of  the  instrument  zero,  if provided,  can be accomplished by
manipulating  the zero potentiometer with the selector switch at the zero
position.  With  the selector  switch in this position, the instrument is not
influenced by the signal  received  from the probe.   Alternatively, the instru-
ment zero is  adjusted using a DO-free solution obtained by adding excess
sodium  sulfite and  a trace of cobalt chloride to water (Standard Methods,
1975).   The instrument  span is adjusted  following  the immersion of the probe
into a  bucket of water  which  has been aerated to saturation.  The dissolved
oxygen  content of the sample  is determined according  to saturation tables or
accepted analytical methods.   For  sensors of the membrane type, some manu-
facturers suggest air calibration  whereby the electrode is allowed to reach
equilibrium in air.  The  span is adjusted to match the value obtained from
saturation tables.

     A  number of calibration  details influence the resulting sensor accuracy.
The calibration  samples must  be stirred  to eliminate  stagnant layers of
water which interfere with the transfer  of oxygen  into the electrochemical
cell.   For sensors  with electrodes in direct contact  with the water and,
therefore, without  membranes  or electrolyte, the calibration samples must be
sufficiently  conductive for the probe to register  DO.  For example, for the
Zvillig  sensor, a minimum  of 300 ysiemen  was recommended for calibration.
Interaction between zero  and  span  potentiometers requires repeated checks at
the high and  low DO levels to ensure accuracy.

Regression Techniques as  Applied to Sensor Calibration

     The calibration procedure described in the preceding section fails to
detect  or compensate for  non-linearity in DO sensors.  Sensor non-linearity,
                                    242

-------
sometimes expressed as a percentage, is the amount of deviation of the rela-
tionship between DO concentration and measurement signal from a straight line
defined by the measurement signal at zero DO and at maximum DO (Liptak, 1969).
As DO sensors were interfaced to a minicomputer, a more sophisticated cali-
bration procedure could be implemented using regression analysis to compen-
sate for detected non-linearity.

     Data pairs consisting of a reference YSI DO reading, y., and
corresponding digital value, x.^, of the test sensor were collected over the
range of the test sensor.  To accomplish this, the probes were immersed in a
sample of mixed liquor which had been aerated to saturation.  The data pairs,
xi and y., were collected at approximately 0.5 mg/1 intervals as the DO level
decreasea due to microbial respiration.

     The data pairs (x ,y ) were plotted to determine the degree of non-
linearity.  When the digital values varied linearly with the reference YSI
readings, a calibration equation of the form y = a + bx, based on simple
linear regression, was used for the software conversion of incoming signals.
For a relationship between the digital values and DO, which was non-linear,
a curve or combination of lines and/or curves was fitted to the data.

     Experience over several months indicated that, for non-linear cases, it
was necessary to fit, at most, two lines or a line and a cuuve to the ob-
served values.  For those portions of the plots which were smooth curves, a
number of functions were used to fit the data.  The most convenient functions
were those which could be transformed to a linear form of the type y* = a +
bx1 where y1 and x' are defined in Table 2.  Functions were fitted to the
curve by carrying out the appropriate transformation on the data and per-
forming linear regression on the transformed data.  A summary of a number of
functions and the appropriate transformations is given in Table 2.

     The results of the regression technique for two sensors which exhibited
non-linearity are given in Fig. 4 and Fig. 5.  The calibration relationship
plotted in Fig. 4 was represented by two straight lines.  For Fig. 5 the
calibration relationship was represented by a line and a geometric curve.

     The regression technique as applied to sensors has a number of advan-
tages when compared with less systematic approaches to calibration.  They
include:
     a) detection of and compensation for sensor non-linearity;
     b) significant reduction in the amount of adjustment required should
        the zero and span be interactive or the internal potentiometers
        require adjustment to restore sensor linearity; and
     c) ability to quantify and document changes in probe characteristics
        with time.
                                    243

-------
 Table 2.  Curvilinear functions of one variable and transformations to linear  for
           (after Hinmelblau, 1968)
    Equation
                                                   Transformation (y' " a + bx1)
                                                  _                _     __
  i/y
    y
  x/y
    y
    y
a + bx
a + b(l/x)
a + bx
ax*
abx
  i/y
    y
  x/y
log y
log y
    x

  1/x
    X

log x
    X
      10.0
          1000    2000     aooo    4000     sow     eooo    TOGO     eooo     9000

                                  SENSOR  SIGNAL  CmV)
                                                                              10000
Figure 4.  Sensor non-linearity - two straight  lines
                                        244

-------
       0)
       E
       U
       o
       U
       o
       d
       O)
          10.0
          B.O -
          8.0 -
          7.0 -
          0.0
            1000   2000   3000   4000    5000    6000    7000

                                SENSOR SIGNAL  CmV)
                                            8000
                                                  9000
                                                        10000
    Figure 5.  Sensor non-linearity straight line and geometric curve
Updating  the Calibration  Equation

     To counteract gradual  changes in instrument  performance which with time
invalidated the initial calibrations, paired reference readings were added
to the initial calibration  data equations.  As paired values obtained more
recently  were more credible than "older" calibration  data,  the updating
procedure weighted the data accordingly.  Weighted regression was used to
obtain the updated calibration equations as follows (Himmelblau, 1968):
Updated equation:
     /\       .  ,
     y =  a   + b x
          u    u
where:
      u
     x
bn - b  x
 0    u

Iw.x.

 Iw.
          y  =
                                      245

-------
          Iw (x -x)y
     b      1~J -
      u    „  ,   -N2
           Iw. (x.-x)
             J  J

     w. = weighting factor on day j
     x. = digital values as obtained by computer on day j

     y . = reference YSI DO readings on day j

     y  = updated sensor DO estimates

     a  = updated intercept

     b  = updated slope

The weighting factors were determined from an exponential weighting function
as follows :

           -(t  - t.)/k
     w . = e   c    i
where:
     w. == weighting factor at t.; 0 _< w. <_ 1.00
     t  = current day
     t. = day on which the data pair x., y. were obtained
      J                               J   J
     k  = constant

     The weighting function dictated that data obtained on the current day
counted 100% in the regression analysis performed on that day; data obtained
k days prior to the current day was counted 37%; data obtained 2k days prior
counted 14% and so on.  The constant k can therefore be thought of as a
"pseudo-time constant" which determines the influence of past data on the
updated calibration equation.  The results of the technique are illustrated
in Fig. 6 and Fig. 7.  Figure 6 shows the initial calibration relationship
of February 3 and seven subsequent values obtained to update the calibration.

     Using an arbitrarily selected k value of 7.0 days and the data of
Figure 6, the lines representing the plots of the updated calibration equa-
tions are shown in Figure 7.

     To compensate for membrane fouling, the calibration line for February 24
was shifted, as a result of the regression analysis, to a position above the
initial calibration line.  The probe was cleaned on February 25 and the
subsequent calibration lines (February 26 and March 2) were returned to
positions close to the Initial calibration line.  The weighted regression
technique updated the initial calibration with more current information,
essentially "tuning" the sensor.
                                     246

-------
         10.0
             1000
                     2000
                            aooo
                                    4000    5000

                                   SENSOR SIGNAL (mV>
                                                            7000
                                                                   WOO
                                                                           9000
                                                                                   10000
  Figure 6.  Weighted regression Initial  calibration relationship with subsequent
             data
        3.00
           4500
                  4600
                         4700
                                4800
                                              MOD   5100   5200   5300    5400   SSOO
                                  SENSOR SIGNAL (mV)

Figure 7.   Weighted regression:   updated calibration  relationships
                                             247

-------
DETERIORATION OF SENSOR PERFORMANCE

     Instrument drift can be defined as a variation or deviation from an
initial adjustment with time.  This variation can range in magnitude from
very subtle and almost undetectable to abrupt and very obvious.  There are  a
number of causes for such behaviour which can be addressed by various preven-
tative actions.

Fouling:  Causes and Prevention

     A buildup of relatively inert material such as oil or grease on the
membrane surface or electrodes can restrict oxygen diffusion into the cell  or
reduce the effective electrode area resulting in low indicated DO readings.
If  the fouling is caused by an oxygen consuming biofilm the sensor will not
measure the true oxygen content of the sample fluid.

     The rate at which the sample is displaced past the probe dictates the
scouring or cleaning action on the membrane surface and thereby influences
the extent to which fouling will likely occur, particularly in the case of
biofilm accumulation.

     Aeration systems which use coarse bubble submerged diffusers or mechani-
cal surface aerators will likely induce adequate in situ fluid velocities to
maintain the necessary cleaning action.  Under such relatively high velocity
conditions, dip probes can be exposed directly to the aeration basin.  This
installation mode is relatively simple and requires little additional equip-
ment.  Mounting hardware should be designed to allow for fast and convenient
cleaning and servicing, and flexibility with respect to probe location in the
basin.  The probe location should take into account the tank geometry and
fluid mixing characteristics to ensure a representative sampling point.
Routine maintenance involves membrane cleaning at an interval determined for
the specific application.

     pHOX sensors were installed at the pilot plant by attaching probes to
lengths of 25 mm diameter PVC pipe and immersing them in the aeration compart-
ments as shown in Figure 8.  Each probe was located a minimum of 0.5 m from
the compartment walls and approximately 1.0 m below the liquid surface.
Removal of a perforated PVC sensor guard supplied with the probe eliminated
possible air bubble impingement on the membrane surface due to turbulent
flow through the shield.  The effect of fouling on a dip probe even under
fairly high velocity conditions can be seen in Figure 9, which represents 21
days between membrane cleanings with a flow velocity in the range of 0.3 to
0.6 m/sec.

     Stephenson et al. (1981) found that dip probes used at low velocity
conditions rapidly became fouled and required high maintenance to ensure
good performance.   Two alternatives available for low velocity conditions
include the inducement of fluid flow past the probe and the use of self-
cleaning probes.   A pHOX probe was installed in a flow through cell  (see
Fig. 10)  through which mixed liquor from the aeration basin was circulated  by
a mechanical pump.   The inlet and outlet piping arrangement resulted in flow

                                     248

-------
Figure 8.   Dip mode showing mounting hardware
       10



        9



        e
    in

   £    6
   X
   o


   Ifl
BEFORE CLEANING
                     AFTER CLEANING
                                                  CORRECT

                                                  FOULED
                          10      11      12       13      14      15       16



                                        TIME  
Figure 9. The effects of probe cleaning on sensor performance 249

-------
tangential to the membrane surface.   The flow  cell was located on the dis
charge side of the pump, a configuration which resulted in unsatisfactory
performance since sensor readings were often arratic and higher than those
obtained when the same probe was  dipped directly  into the aeration basin as
shown in Figure 11.  The higher readings were  caused by air entering the
pump casing and being circulated  past the probe.   Flow cells  used on other
pilot plants have been placed  on  the suction side of the pump  and have per-
formed satisfactorily for extended periods of  time (Stephenson et^ al. , 1981).
     Figure 10.  Flow through configuration
           2. s
           2.0
 0>
 E
sx

i—i
01
>-

 I

X
           1.0
           0.5
           0.0
          -0.5
                 DIP MODE
                                    FLO* THROUGH MODE
                                                             DIP NODE
                 _]	I	I	I

             0.0  1.0   2.0   3.0   4.0   5.0  6.0   7.0   6.0  8.0   10.0   11.0  12.0

                                    TRIAL  NUMBER


     Figure 11.  Difference in sensor readings according to installation mode


                                      250

-------
     A comparison of  the two installation modes suggest that because of the
ease of inspection and  cleaning and the flexibility of location, the dip
mode be used where velocities provide sufficient cleaning action.  Flow
inducers require additional hardware and equipment (and, therefore, create
other maintenance problems) but reduce membrane fouling.

     The self-cleaning  Zullig required periodic servicing of the grinding
mechanism.  The grindstone became "blinded" with material from the electrodes
which resulted in less  effective cleaning and arratic performance.  As docu-
mented in Figure 12,  this can lead to a rapid drop in indicated dissolved
oxygen and should be  regularly checked to ensure data integrity.

     Other causes of  performance loss relate to physical deterioration of
equipment.  Electrode decomposition and electrolyte inactivity require that
the probe be returned to the manufacturer for repair and rejuvenation or
replacement.  Membranes age, gradually suffer loss of permeability, and must
be replaced.  The analyzer/transmitter, like other electronic instruments,
is full of electronic components which are subject to failure.  Trouble
shooting, adjustment  and repair can be extremely time consuming and requires
skilled and experienced technical personnel.
   10

    s

    e

    7

£   e
z
13   5
x
o
ui
o
    3

    2

    1
                                                        CORRECT PROBE
                                                        ZULLIG PROBE
                                         GRINDSTONE BLINDED
                                         ELECTRODES FOULED
            ?7   19  21  23
                          357

                           TIME  
1 1 13 15 17 Figure 12. Plot demonstrating failure of the self cleaning mechanic PROBE DYNAMICS One important application for DO sensors is the monitoring of DO concen- tration during non-steady state tests for determining overall mass transfer coefficients (K.a). The suitability of DO sensors for tracking DO concen- tration which was changing with time was determined by measuring a first- order time constant (T) using step testing. 251

-------
     An upward step in DO was  induced by transferring the DO probe under
investigation from a stirred bucket  containing DO-free water to a second
stirred bucket containing DO-saturated water.   A downward step was induced
by reversing the order of transfer.   The transfer was accomplished in  less
than one second.  During the step  testing,  the output signal from the  DO
sensor was sampled by the computer at one second intervals.

     For a first-order system  responding to a  step change, the following
relationship applies between the time elapsed  following introduction of  the
step and the percentage of  the total step completed:
              Elapsed Time
Percent of Total Step
               1/2 XT                         39.3
                     T                         63.2
                 2 x T                         86.5
                 3 x T                         95.0
                 4 x T                         98.2

where T = system first-order  time  constant.

     Using this relationship,  five estimates  of the sensor time constant
were determined for each step  test.  A typical plot of the one-second values
with the five appropriate estimates  is given  in Figure 13.
11000.
iimn
*"-N
LJ
aofYYi

=£
^" BOD.

7 7000.
0
<" am
z
z sooo.
2
1—
en 4DQQ.
3000.


2000.
imn
. 4 « TAU (BBZ>
2 « TAU O7C


-

TAU aao 	

-
1/2 « TAU SOD
1-

-
_



1 jl




/












1

fl
'


















































|




OMUTEDi T-a-STARDi TAlk
(SEC.) OEC.)
ID 41.1 111
B5> 25.2 1.4
in 14. e 7.4
en e.« 1.1
an Z.B 5.5
BW- 3B.1
TAU AVER.- 38. 1/5- 7.6 SEC.




l i
           0.0
                      1.0
                        T-itART
                                2.0
                                           310
                                                      4.0
                             ELAPSED TIME (min.)
   Figure 13.  Plot of sensor response to a step change in DO and time constant
             estimates
                                     252

-------
      The step test results  are summarized in  Table 3 which contains, for each
 of  the steps, values of the first-order time  constant (T) estimated at 63%
 of  the response, 95% of the response and as the  mean of the estimates at 39%,
 63%,  87%,  95% and 98% of  the response.  Also  included in the table is an
 estimate for a time constant for a Beckman analyzer and Model 39553 DO sensor
 as  measured by Shiba (1981).
      Table 3.  Sensor first-order time constant as measured by step testing
                   Temperature of Solution
             Time Constants (T)

pHOX (1,1)*


pHOX (1,2)*


pHOX (4,9)*



Zullig



YSI



DO Free
(°C)
18
15
17
16
13
16
16
15
16
15
15
13
15
15
21
21
23
21
DO Saturated
CO
17
15
15
13
13
13
15
14
14
15
14
13
14
15
24
23
23
24
Step
Up
Up
Down
Up
Down
Down
Up
Up
Down
Down
Up
Up
Down
Down
Up
Up
Down
Down
At 6 32
(sec)
6.8
6.6
8.5
14.6
11.9
10.5
8.1
9.9
13.9
11.3
27.9
10.9
14.4
14.5
8.5
7.6
13.0
12.8
At 952
(sec)
8.4
8.1
15.3
23.1
22.5
20.4
19.6
18.9
25.9
20.9
17.0
10.2
14.2
13.5
13.9
10.4
17.3
15.6
He an
(sec)
7.6
7. it
13.3
20.1
18.5
17.3
15.6
15.8
22.0
17.3
23.9
10.7
15.3
14.6
11.5
10.1
15.1
14.5
     Beckman
Up
                                                   15.0**
      *  The numbers refer to the slot and channel of the sensor Interfaced at the
         computer. They are used to distinguish between the three sensors of the same
         make.
     **  After  Shiba, 1981.
     For  all the sensors, with the exception of the  Zullig (the only probe
without a membrane or electrolyte), the estimates of the first-order time
constant  Increased as the percentage of the total response (upon which the
estimate  was based) increased.   Therefore, for the sensors employing mem-
branes, the dynamic response  is not strictly first-order.   The deviation
from first-order could be caused by diffusion limitations  across the
membrane.

     A Student's t-test was performed to determine if a  difference in time
constant  could be detected between responses to upward and downward steps.
The YSI was the only probe for which there was a significant  difference at
the 95% confidence level.  Based on the results of the t-test, the mean

                                      253

-------
 first-order  time  constants  are summarized in Table 4.

     As  can  be  seen  in  Table 4,  the mean time constants were less than 20
 sec. and were in  the same range as  for the YSI which would normally be used
 for DO determination during non-steady state ILa tests.  Therefore, based
 on the results  of the step  tests, the on-line probes were judged suitable
 for monitoring  DO concentration during non-steady state K^a tests.

     The dynamic  response of a DO sensor as measured by step testing is
 thought  to be influenced by such factors as the size of the forcing step,
 test conditions,  and the age and condition of the membrane and cell.  Kok
 and Zajic  (1975)  determined YSI response time for a number of downsteps and
 found that the  response time was not influenced by the magnitude of the
 downstep over the range investigated.   Additional work is required before
 this conclusion can  be  applied to other makes.   Kok and Zajic also observed
 a decrease in the YSI time  constant from one day to the next and attributed
 the change to the aging and stretching of the membrane.  These observations
 and comments by other researchers  (Kulin and Schuk, 1978) indicate that
 membrane and cell aging may exert a considerable influence on sensor response
 time and that further research is justified to expand knowledge in this area.


      Table 4.  Summary of first-order time constants

             Sensor                     Mean First-Order Time Constants
                                                (sec)

            pHOX (1,1)                               9
            pHOX (1,2)                              19
            pHOX (4,9)                              18
            Zulllg                                 16
            TSI                                   11 (up)
                                                 15 (down)
 SUMMARY AND CONCLUSIONS

      Experience with sensors  used to  continuously monitor dissolved oxygen
 concentration  in  the aeration tank of an activated sludge pilot plant
 resulted in improved calibration procedures,  evaluation of installation and
 maintenance alternatives  and  determination of first-order time constants for
 each  of the sensors.

      Standard  calibrations  which involved adjustment of instrument zero and
 span  failed to detect or  compensate for non-linearity.  Regression analysis
 on pairs  of values  from the reference and test sensors compensated for non-
 linearity,  reduced  the number of potentiometer adjustments and provided
 documentation  of  the calibration process.   Exponentially weighted regression
 analysis  used  more  recently obtained  reference values to update the initial
 calibration equation and  correct for  drift.

     Probe  fouling  due to the growth  of microbial films was the major mainte-
nance problem.  The severity  of  the problem depended on the fluid velocity

                                     254

-------
 in the aeration tanks.  For basins with high fluid velocities, probes
 immersed directly into the liquid performed well and provided flexibility
 and simplicity of installation and inspection.  With low fluid velocities
 special  self-cleaning" probes, flow-through cells which use pumps to convey
 the fluid past the probe, or frequent inspection and cleaning of probes used
 in the dip mode,  were required to maintain performance.

      Sensor inaccuracy also resulted from membrane aging and damage, the
 clogging of the anode and cathode with decomposition products, electrolyte
 exhaustion and failure of electronic components.  These problems of physical
 deterioration occurred less frequently and are a function of the design life
 of the sensor components.

      First-order time constants were determined by subjecting each of the
 sensors to repeated upward and downward steps in DO concentration.  The mean
 time constants for the probes tested were estimated to be less than 20 sec.
 in all cases.   The dynamic response of a sensor without a membrane was much
 closer to that of a first-order system than was that of sensors with mem-
 branes.  Additional research is required to determine the effect of cell and
 membrane aging on the dynamic response of DO sensors.

      It is hoped that this work will encourage other users to communicate
 their experiences with installation,  calibration and maintenance of sensors
 used to continuously monitor or control DO in the activated sludge system.


 ACKNOWLEDGEMENTS

      The authors  wish to acknowledge that this material was previously
 presented at  the  17th Canadian Symposium on Water Pollution Research,  Canada
 Centre for Inland Waters, Burlington,  Ontario on March 11,  1982.   The
 manuscript has been submitted for publication in the Water Pollution Research
 Journal of Canada,  Volume 17,  1982.
REFERENCES

APHA-AWWA-WPCF,  "Standard Methods  for  the  Examination  of Water  and Waste-
   water", 14th  ed., American  Public Health  Association, Washington, D.C.,
   pp. 1193  (1975).
Himmelblau,  D.M.,  "Process Analysis by Statistical Methods," John Wiley  and
   Sons, New York,  pp. 463 (1968).
Kok, R. and  Zajic,  J., "Dynamic Response of  a Polarographic Oxygen Probe,"
   Biotechnology and Bioengineering, 17, 527 (1975).
Kulin, G. and Schuk, W., "Evaluation of a  Dissolved Oxygen Field Test Proto-
   col", U.S. EPA,  EPA/78-D-X0024-1, pp. 21  (1978).
Liptak, E.G., "Instrument Engineers' Handbook", Vol. 1, "Process Measure-
   ment", Chilton  Book Co., New York,  pp.  1174  (1969).
Shiba, S., "Calibration of Non-Steady  DO Measurement", Jour. Eny. Eng. Diy..
   Amer. Soc. Civ.  Engr.. 107, EE1, 279 (1981).
Stephenson, J.P. et al., "Evaluation of Instruments for Continuous Activated
   Sludge Monitoring," Water Science Technology, 13, 713 (1981).

                                    255

-------
DISCLAIMER

     The work described in this paper was not funded by the U.S.  Environ-
mental Protection Agency.  The contents do not necessarily reflect the views
the the Agency, and no official endorsement should  be inferred.
                                    256

-------
 OPERATION AND  MAINTENANCE OF CERAMIC FINE BUBBLE DIFFUSERS IN ACTIVATED
 SLUDGE SYSTEMS

 Daniel H.  Houck,  P.E.
 D.  H.  Houck Associates
 8905 Garland Avenue
 Silver Spring, Maryland  20901

 ABSTRACT

      Experience with the use of ceramic fine bubble diffusers for activated
 sludge aeration in the United Kingdom,  Holland,  and the U.S.  is  discussed.
 Results of a survey of 19 treatment plants using this equipment  and  the
 author's  subsequent experience and observations  are summarized.

      The  impact of plant design on the  subsequent efficiency  and operability
 of  ceramic diffuser systems is outlined.   Causes of failure and  their  symp-
 toms are  summarized.   Design and operational strategies for optimal  system
 performance are introduced.

 INTRODUCTION AND  BACKGROUND

      It has been  recognized for some time that fine bubble aeration  systems
 for activated  sludge are inherently more  energy  efficient  than coarse  bubble
 systems.   The  use of fine bubble aeration systems for oxygenation of activa-
 ted sludge processes has been practiced with varying results  since around the
 turn of the century.   Earlier systems often used ceramic plates  which  were
 mounted in place  on the tank floor over a cast in place air distribution mani-
 fold.   Cleaning of these systems was costly and  time consuming.   The steady
 decline of energy costs in real dollars after World War II, coupled with the
 development of very low maintenance coarse bubble diffuser systems resulted
 in  the conversion of many of the earlier  fine bubble systems  in  the U.S. and
 use of coarse  bubble equipment in most  new plants.

     Historically,  energy costs in the  United Kindom and Europe  have been
 considerably higher than those in the U.S.   Consequently,  there  has been
 continued  development  and application of  more efficient aeration systems
 there,  particularly in the U.K.   Slow speed mechanical turbines, typlified by
 the Simplex and Simcar aerators,  were developed  as a more  efficient alternate
 to  coarse  bubble  systems and were widely  applied in the U.K.,  and later in
 the U.S.   In the  late  1950s,  Activated  Sludge Ltd.,  developed and patented a
 fine bubble ceramic diffuser using an air grid that was mounted  above  the
 aeration tank  floor.   The individual cylindrical diffusers were  about  6
 inches  high and 4  inches in  diameter and  were mounted on top  of  the air pipes
 in  an  even pattern.  The entire assembly  was bolted to the tank  floor.  The
 intent  of  this early design  was to greatly simplify the construction and main-
 tenance of ceramic  fine bubble diffuser systems  while retaining  whole  tank
 floor  coverage.  A second generation of the ceramic dome diffuser was  brought
 out  in  the late 1960s.   The  height of the diffuser was reduced to about 2
 inches  and the diameter was  increased to  7  inches resulting in an improved
bubble  pattern and  greater capacity.  This  design,  shown in Figure 1,  is the
predominant  type of ceramic  fine  bubble diffuser in use in England, the U.S.,
and Canada  and is manufactured by Hawker  Siddeley Ltd.  in  the U.K. and by the
                                      257

-------
                      ORIFICE 9OLT
      POROUS
      ALUNDUM
      OOMC
                                                                  © 7" Diameter Aloxite Dome Diffuser
                                                                  ©P.V.C. Dome Saddle
                                                                  ©Air Flow Control Orifice
                                                                 ' © P.VC. Pipe Retaining Strap
                                                                  ©4" Diameter RV.C. Air Header
©Adjustable P.VC. Pipe Support
© Stainless Steel Cinch Anchor
©Aeration Tank Floor
          NORTON
Figure 1.   Norton/Hawker  Siddeley and Gray  Engineering  7 inch dome  diffusers.
                        CIFPJSSS RETAINING .
                              PROPYlSIE)
                              M6EIA 9'OIAX
                                                         S ANITA I RE
                                       ORIFICE  CONTROL NOZZLE
                                       RETAINER RING         \
                                      /"
                                                  A
                                          Ol
                                   DIFFUSER ELEMENT

                                       "O"RING
                                                                                     HOLDER
                                           A/«  DISTRIBUTION
                                           HEADER
                                                  DIFFUSER  ASSEMBLY

                Figure 2.   Sanitaire  and Envirex disc diffusers.

                                          258

-------
Norton Co.  in  the U.S.  Gray  Engineering,  a Canadian firm markets a similar 7
inch dome  diffuser which features  an  offset dome saddle  which  they claim re-
duces the  dome  bolt stress  on the pipe.  Another  type  of ceramic fine bubble
aerator uses a  disc configuration.  An example of  this, shown in Figure 2, is
manufactured by the  Sanitaire Co.  Disc  diffusers  are also manufactured in
the U.S. by Envirex, Infilco-Degremont, and  EPI-Nokia.

     This  paper  relates  observations  and  experience  from systems  using  the
Norton/Hawker  Siddeley 7  inch dome diffusers.  Applicability of  the  data to
disc type units  is  discussed later herein.

BASIC DESIGN CONCEPTS  FOR DOME DIFFUSER ACTIVATED SLUDGE SYSTEMS

     Operation  and  maintenance  of  dome  diffuser  activated sludge  systems
(DDAS)  are very  much  affected  by aeration  system  design.   Inlet  design and
tank geometry  in particular play  a very large role in the operational charac-
teristics  of  the system.  Also,  the system design  should provide the operator
with  the  means  to monitor  dissolved   oxygen  levels  and  control  air flows to
portions of  the aeration tank if  he is to  have some  control over the process.

     Most  aerators  used  in  activated   sludge  systems  perform  two  functions:
(1),  aeration  and  (2),  mixing  of  the tank  contents.   Adequate  mixing is
required to  avoid settling  of mixed liquor solids and  disperse the dissolved
oxygen  in  the  aerated  volume.   Distinct   types  of   aerators  have  equally
distinct mixing characteristics.   For  example,  a  mechanical  surface turbine
concentrates  its  mixing  energy at  a  central  point  in  the  tank  and  creates
strong  surface  currents  which extend to  the  tank wall,  returning  to  the
impeller along   the  tank  floor  (Figure 3).   Spiral  roll  diffused  air systems
input  mixing energy  along  one side  of the  length  of  the  tank,  creating  a
rolling  motion  of  the  tank  contents.   Dome  diffuser  systems,  on  the other
hand, input  mixing energy in  a grid configuration,  relying on  the release of
energy  from the  rising  column of air  bubbles  to create  a  localized air  lift
pumping  effect.  As  a result  of  this  configuration,   the  dome  diffuser  grid
provides  excellent  and  efficient  mixing along  the  vertical axis.   However,
the  induced  surface currents  are  weak  and short lived.   Thus,  the horizontal
mixing  along  the length or  across the width of  the  aeration tank, especially
for  very wide  tanks,  is  much  less  pronounced  than  for surface  turbines or
coarse  bubble   spiral  roll   systems.   Optimal design of  dome  diffuser aerated
systems  requires  that  the   tank  and inlet  geometry  take  this  characteristic
into account.

     Influent distribution  configurations that work  satisfactorily for coarse
bubble  aerators may not be  adequate for ceramic diffusers.  For example, in-
fluent  was  observed to channel  along  one side of a step feed tank that uses a
series  of  gates  in the  tank wall  ,  a common  configuration in  U.S.  plants.
This  situation   should  be  avoided  in   a  new plant and  extra  diffusers and/or
baffles  should  be  placed  around   the  inlet zones where existing  plants are
being converted to  dome  diffusers.   Short  circuiting  has  also  been observed
in the  U.K.  in  the  past  and most  systems  there now  combine the return sludge
and  settled  sewage in a mixing channel prior to distributing it evenly across
the width of the  tank  at the inlet end.
                                      259

-------
                       SLOW SPEED TURBINE
                                       t   ill   t
             single roll
double roll
                 COARSE BUBBLE  SPIRAL ROLL
                   *  4
                                 '  1 '  ' ' T  1 '  ' ^ T  -»
                                 »» 3*trw
                FINE BUBBLE CERAMIC DOME/DISC
Figure  3.  Mixing profiles of some  common activated  sludge aerators.
                               260

-------
        D.O.
      mg/ I
                infl
TANK LENGTH
 Figure 4.  Rising  dissolved  oxygen profile in long, narrow plug flow tank.

     A  second  problem,   overaeration,   was  observed  in  many of  the  U.K.
treatment  plants  visited  by  the  author  in  an  earlier  study .   Excessive
aeration  of  all  or  most of  the  aeration volume  can  result  in  poor  energy
efficiency  of  the  system  and  may also  adversely  affect  sludge  settling.
Plants  that have  this problem  typically  exhibit  the dissolved oxygen curve
shown in  Figure 4.  D.O.  levels, in the proper range at the head of the tank,
rise  steadily  along   the  tank  length   until  they  are  quite elevated  at  the
effluent  end.  This   effect  was  repeatedly  observed in  U.K.  plants  and  was
most  severe  in _plug  flow systems with very  high  (over  20)  ratios  of  tank
length  to width .  For the most  part,  the plant operators  could  not  correct
for the problem by adjusting  air flows.   Operators  at  the main plant for the
City  of  London, Beckton,  have  removed  up  to  one  third of  the aerators  from
some of the tanks  in  an  attempt  to  reduce the overaeration effect.  In doing
so,  they  have  verified  that the  system  will  maintain solids  suspension at
very low  air  flow  rates per unit volume.

     A  variation  of   the  overaeration  problem occurs  in  the  system  that is
simply too  conservative in design.  Such a system  may  have considerably more
domes  than  required   to  meet  the  oxygen  demand of  the   system.    Each  dome
diffuser,  and hence  by extension  the   entire  system, has  a  minimum air  flow
rate that it  can be  operated  at.  This  minimum, 0.4-0.5 cfm for 7  inch domes,
is  required to avoid backflow  of mixed  liquor into the  diffuser.   Backflow
can cause bio-fouling  if  the system is  operated at  lower  air  flow  rates  for
extended  periods.   Seriously  overdomed  plants  will  supply  excessive  amounts
of oxygen to  the mixed liquor during much or  even most  of the daily treatment
cycle.  An  extreme example of  this was  observed by the author at  Strongford,
England,  where  plant  operators  must  run  the  system at minimum air flow  for
20-22 hours  per  day  and dissolved oxygen  can  range  from  4-5  mg/1  at  the
influent  end   of   the  tanks   to  near  saturation  at  the  effluent.   Not
surprisingly, the energy efficiency of this system is low.
                                      261

-------
      The  combination of high  length/width ratio and  the  weak horizontal  mix-
 ing  characteristics of dome  diffuser systems can lead to the  situation illus-
 trated  in  Figure  5.   In  this  example,  typical  of  many  of  the older  U.K.
 plants,  settled sewage and  return sludge is merged in an open aerated  channel
 and  distributed across one  end of a  three  pass  plug  flow  tank.   As  the mixed
 liquor  progresses   along  the  treatment  length,  two  things  are occurring.
 First,  the more degradable  soluble  BOD is metabolized and  the oxygen  deficit
 of  the  return  sludge  is  exerted mostly in  the  upper portions of  the aeration
 process.   This  results  in a high  aeration  demand in  the  first tank section.
 Because  there  is only  limited  upstream-downstream mixing by  the aerators,  the
 domes in  the  first section are  required  to supply  this  demand almost  alone.
 In  the  second  and  third  passes, the oxygen demand rate has fallen off  sharply
 as  the  less  easily  metabolized  BOD is removed.

      In  the absence of other  influences,  the diffusers  in  the upper sections
 of  this  system  already must  supply most of  the oxygen demand.  In addition,  a
 second  condition makes the  load on the upper tank sections even more extreme.
 The   alpha  factor  in  this  situation  is  also  variable!   The  projected curve
 shown in  Figure 6, from  observations in other  similar  situations  indicates
 a 100 percent  variation  in  alpha from front to  back of the  three pass  system
 may  be  possible.   In  concrete  terms, a dome diffuser in the  first part  of  the
 first pass  is  only one  half  as  efficient  as one  located  immediately  before
 the  outlet.  At the  same  time, the  first  dome  is operating  in   thea zone of
 highest  oxygen  demand,  where  D.O.  requirements  may be 2-3 times  higher  than
 in  the  lower zones  of  the system.

      Tapering  the  aeration,  placing  more of the diffusers  in the first  sec-
 tions of the tank,  has  been recommended  for quite  some  time by both the U.K.
 and  U.S.  manufacturers of dome diffusers.  However, this remedy only works up
 to  the  point where  mixing  requirements begin to control the number of  diffus-
 ers  required in the lower  portions of the  aeration  tank.   Beyond that  point,
 any  more  diffusers  needed  in  the upper  aeration zones must  be   added   to  the
 total,  and  overaeration  in the  lower zones  is  likely to  result.  It   is  the
 author's  opinion at this  point that  tanks with  length/width ratios exceeding
 16/1  cannot  reap  the  full benefit of the efficiency of ceramic  dome or disc
 aerators  due to geometric  limitations.
                                                                     2
      Many  of the  plants  visited by  the  author in an  earlier study  suffered
 from  combinations  of  poor  tank geometry  and  maldistribution of aeration.
 Even  where aeration was  tapered,  tapering was inadequate  in situations where
 length/width exceeded  10/1.   Thus,  an operator  confronted with a plant  design-
 ed  like  this  can  do  little to  optimize  its performance.  He  is  lucky if he
 can avoid  slime growth and  other operational problems which result from  poor
 tank/inlet configuration.

      If the  designer does not  provide a  reliable  dissolved oxygen monitoring
 system and adequate air  valving  and controls,  the operator  will  have  little
 incentive  for  and   great  difficulty  with optimization  of  D.O.   levels. D.O.
 levels are  most critical  in  the first section,  where very  low oxygen can  lead
 to slime  growth on diffusers and piping,  and in the finalj  section where there
may be  a tendency  to  over aerate.   At minimum,  D.O.  should  be monitored  and
 air flows  individually controlled  at  these  points.  Other  control points  may


                                     262

-------
         DO


      DEMAND
                 l^1



                I
                infl
eft I
              to-r
     ALPHA
              &•
              A-
              .2-
                J
               irifl
I
effl
Figure  5.   Oxygen  demand  profile  and  alpha  variation  in extreme  plug  flow

            systems.
                                       263

-------
be required depending  on  system geometry.

     Finally,  the  plant design needs  to  take into account the reality of main-
tenance.   When hundreds or even  thousands  of aerators with  associated piping
are  installed in  a  tank  floor, it is likely  that something will  go wrong on
occasion  and   will  need repair.   A plant  design that makes  it  difficult for
plant  operators to  dewater  and clean a  tank  and make repairs,  is an invita-
tion  to  major headaches.  The  plant  should be designed with multiple aeration
basins  such  that  any  one can be  shut  down  for  routine cleaning and repair
without  overloading the  rest  of the system.   It should be easy  and safe for
the  operator  to drain  and wash down  a  tank.  Plant  equipment should include a
portable  hoist  for  lifing buckets  or  pallets  of diffuser stones out of the
basin.   Tanks that are being  put  in  service are first checked by  covering the
diffusers  with  clear  water  to  a  depth of 1  foot  and checking  the diffuser
bubble  pattern  for irregularities.   Consequently,  an  accessible  source  of
clean water or plant effluent  should  be  provided.

COMMON OPERATIONAL PROBLEMS  AND REMEDIAL MEASURES

      Summarized  herein are some of  the  more common operational  problems that
can  occur with  DDAS  systems  and  possible  remedial measures.   These comments
are  based on the  author's  observations and  continuing  studies,   supplemented
by an  in  house evaluation conducted by CH^M-Hill Engineers .

Diffuser  Sliming

     Diffuser  sliming  results when conditions  are  right  for  biological
growths  to occur  on the  surface  of the  diffuser and piping.   In severe cases,
the  growths  penetrate  the dome to a depth of 1/8 inch or more.   The sliming
acts  to  reduce  transfer  efficiency  in  two ways.  First, the  slime causes  a
recoalescence of  the  fine  bubble  as  they  emerge  from  the  diffuser  face.
Secondly,  biogrowths  which  have  penetrated  the  diffuser surface  impede  the
flow  of  air,  raising  the backpressure  at  the  diffuser.  Sliming is readily
observed  at  the  surface  of  the  aeration  tank where  the  effervescent gentle
aeration  pattern is replaced  with  an irregular pattern characterized by large
bubbles.   All of  the  causes  of slime are not  well understood.  However, it is
most  likely  to  occur  at  the  point  where  settled sewage enters  the tank and
with  wastes  that  are  highly  degradable.   It  is  exacerbated by  low dissolved
oxygen and may also depend  partly  on temperature.  Sliming  is more likely to
occur  in  long  narrow  tanks  that  are inadequately  aerated in  the front sec-
tion.
                     2
     Work  at   Beckton   suggests that  light sliming is  reversible by shutting
off the  flow  of untreated waste  to a tank and aerating recirulated flow only
for  24-48 hours.   More  severe  cases will jrequire  draining  of  the  tank and
cleaning  of  the diffusers.    At  Beddington ,  the diffusers  are   periodically
brushed  to remove growths.    Madison  steam  cleaned  badly  slimed diffusers,
which seemed  to cure  the  problem,  although the pressure loss through the dif-
fusers remained  somewhat  elevated  after  cleaning.  Badly fouled diffusers may
require retiring in a kiln, which  necessitates their removal and  replacement.
This  approach, commonly used  in  the  U.K.,  costs about one half of the cost of
new diffusers.  At  this  time,  there is no one method  that  is  wholly satis-

                                      264

-------
factory for  dealing with sliming.   The  Sanitaire Corp.  has been developing an
in-line acid cleaning  system,  which  is currently being  installed  in  several
plants.   However,  there is  inadequate  experience with  the  system at  this
point  in  time  to conclude  that  it  will  be able to deal  with diffuser fouling,
particularly  of   the  organic  variety.   At this  point  in time,  the  most  cost
effective  means   of dealing  with diffuser  fouling is  to prevent  it  through
proper design and operation.

Diffuser Plugging

     Diffusers can  also  be plugged internally  by dirt  in the  air  supply,  or
by  rusting of metallic pipes.    In  either case, the material  is  deposited  on
the  inside  of  the  dome,  gradually  plugging up the pores of the ceramic  mater-
ial.   Iru a few cases,  inorganic  scaling on the exterior surfaces of domes has
occurred  .   Acid washing  has  been  used  to  remove both types  of  plugging.
The  best  remedy  for  internal fouling is to avoid the  problem by maintaining
air  cleaners  properly  and  by  using  non-metallic  pipe materials  where  ever
moisture  may  be  present,   particularly  below  the  water  lines in  aeration
tanks.  Scaling  may not  be  preventable  and use of ceramic fine bubble  diffus-
ers  should probably be  avoided where this is known to  be a severe problem.
   /
Failure of Plastic  Pipe and Fittings

     Out  of  6  plants  visited  by  the CH-M-HilL team,  problems  were  experi-
enced  at   5  with  plastic  piping and  hardware.    The  sixth  plant,  Toronto,
uses metal  piping  and  has  had no failure with the exception of possible leak-
age  around  the dome gasket.  The most common  failure reported is breakage of
the  plastic  dome retaining bolt.   This parallels the  British experience and
is  usually caused  by  overtightening  of  the bolt by installers.  Other  causes
of  failure include  cracks  in the  pipe  saddle, failure  in  expansion  joints,
blow off  line  breaks,  and  air  leakage around the dome gasket.   Overtightening
was  cited  as the principal  cause of most of the failures at the plants  visit-
ed  by  CH M-Hill.   There were indications at  one plant  that  assembly  of the
piping  during  hot  weather  may have caused failures as  well.   Figure 6,  from
Reference  1, illustrates the  principal failure modes.

     Interestingly, the  implied frequency  of  failure  in the  CH.M-Hill  work
may  be somewhat  greater than that  experienced in the U.K.  Possibly this  is
due  to  the fact   that  most  of  the aeration grids in the  U.K. were installed by
the  equipment  manufacturer,  Hawker Siddeley,  with experienced  crews.   Plant
operators  in  the  U.K.  note  that  proper  initial  installation is  of  primary
importance  for a relatively  troublefree  installation.    Also,  British  opera-
tors expect  and  accept  the need to periodically maintain  dome  diffuser  sys-
tems.   Dewatering  of  tanks,  cleaning  and  checking  are  routinely  performed
every  2-3  years  in  most  U.K.  plants.   With  routine maintenance of this type,
there  is   a  greater  liklihood  that  problems  will be  corrected before  they
become serious.

APPLICABILITY OF  DOME  DIFFUSER FINDINGS  TO DISC SYSTEMS

     As noted previously, several U.S. and foreign manufacturers have develop-
ed  disc  aerators  (Figure  2)  which  are  somewhat  similar to  dome  diffusers.

                                     265

-------
                         Retaining Boll
                         Breaks Here
                                                                     Gasket Leiks
                                                               0 *   Here
                                                                O
 Pipe Sirap
 Breaks Here
                       .»
                           Compr*$$ion Nut
                           Cracks Along
                           Circumference
Figure 6.   Dome  diffuser and  plastic piping failure modes  .
                                 266

-------
Often,  these  units are being £i£ as  alternates on projects specified for dome
diffusers.   Several efforts '  •*  have  been made to assess  the  substitutabil-
ity of  disc  diffusers for.domes.   Experiments at Holten-Markelo,  in Holland,
suggested  that oxygen transfer  between dissimilar sized ceramic  diffusers  is
directly  proportional to  surface area/   More extensive work at  Los  Angeles
County,  where the  8.7  inch Sanitaire  disc was  compared to the 7  inch  Norton
dome,  concluded  th^at 3  of the discs would  transfer  about  the same  amount  of
oxygen  as 4  domes .   However, mixing  has  not been addressed specifically  in
any of  these  studies.  Substitutions of this type should  be  reviewed  closely
before  being  approved as  they  might  lead  to  mixing problems  in  a  small  number
of cases where diffuser density and  air flow  are  already very low.

     Mechanically,  the  disc diffusers  may  offer several advantages  over  dome
type  units.    The  elimination  of  the center  bolt   eliminates  the  problems
caused  by bolt  stress  on  the  pipe  saddle  (Gray Engineering's offset  design
also  mitigates this  problem).   Applying hold down pressure uniformly  around
the periphery of the diffuser  is  also  more  likely  to  prevent leaks around the
base  gasket.   However,  to date  experience  with  these  units is  rather  short
and long  term problems may yet  occur.

CONTINUING RESEARCH

     Spurred  by  the U.S.  EPA,  the British Water  Research Centre,  and Environ-
ment Canada,  major research efforts  in the  last  4 years have yielded signifi-
cant  strides  in understanding  ceramic  diffuser  system design and  operation.
Current  research  is  focusing  on optimization  of  the efficiency  of  these
systems  and   in  avoiding  or  solving  problem areas.   Substantial changes  in
past  design   and  operational  practices is  already underway  and  current  and
future  research  should  result  in  further  advances.   Presently,  there  are
three major research  projects underway:

     The  Reymeads,  U.K.,  Demonstration Project-  Based on  the  initial
     investigations  detailed  in  Reference  2,  and  further developmental
     work by  the  WRC, a full scale demonstration of optimized design and
     opeation  of  a dome diffuser  system has  been set up at  the Ryemeads
     plant.   One  third  of the  plant, treating 3.2 mgd,  has been revamped
     to  improve  the diffuser  configuration  and provide  dissolved  oxygen
     monitoring  and automatic  air flow control.  Substantial  increases
     in measured  system efficiency have already  been achieved, and  fur-
     ther modifications  were made recently which should result in  over-
     all efficiency improvements  in excess of  100 percent over the exist-
     ing system.

     Comparative  Studies   of  Fine  Bubble  Aeration Systems,  Los  Angeles
     County,   California-  A  joint  project  between  the  U.S.  EPA and  Los
     Angeles   County  is  conducting detailed efficiency and  design  studies
     of  a number   of  fine  bubble  aeration  systems,  including   domes,
     discs, jet aerators,  and tubular units.

     Biofouling/Performance  Studies  at Madison, Wisconsin- Efficiency
     measurement using  off-gas  analysis  is   the  focus  of  this  project,
                                     267

-------
     which  is  also studying the  causes  and cures of  the  episode of bio-
     fouling experienced at  this  plant  in  1980.

     From  the  results  of  these  and  other work,  a design rationale  for plug
and  step  feed applications  of ceramic  grid  fine  bubble diffuser  systems is
currently  being  developed  by  the  author  of  this  paper.  Using  a  micro-com-
puter  to  handle  tedious  calculations,  the design model  treats  the  aeration
tank as a  series of completely mixed  square  reactors  and considers  the varia-
tion in alpha  and oxygen demand  with  detention  time  (tank length).   When fin-
ished, the  program will be  used  to calculate the  number of  diffusers required
at  any point  in the tank(s) with much greater  precision  than current practi-
ces.  Of course,  the output  from  this  design  model  must  be coupled with exper-
ience gained from  study of  existing systems in arriving  at the final design.
CONCLUSION

     Extensive  experience  in  the United  Kingdom with  ceramic  dome diffusers
has  shown  them to be  efficient and reliable  when applied and operated proper-
ly.  The  key  to effective  use  of this technology  is  understanding  its design
and  operational  boundaries.   Aeration basin  and  inlet  geometry is  of  prime
importance  in new  or  retrofitted  plant  design.   Operators  of  these systems
should  be   trained  to  spot  the symptoms of impending problems  and  respond to
them quickly.   Close  dissolved   oxygen  monitoring  and  control  is  a  must for
system efficiency  and  control  of  bio-fouling. Maintenance need not  be excess-
ive, but  it must be routine and  thorough  for best  results.   Ceramic diffuser
grid systems  are  not  "install it and  forget it" systems!  Even  so,  the cost
incurred  for  additional maintenance is  offset many times by  the 200-300 per-
cent improvements in  electrical  efficiency  that  can be  realized over coarse
bubble  systems.   Other  advantages include  lack of  noise and  airborne  aero-
sols,  and  ease of plant  sanitation.   The  savings  in  power costs, coupled with
other  operational advantages,  have resulted  in  fine bubble  ceramic grid sys-
tems becoming  the system of choice  in the U.K.  North  American plant design-
ers  and  operators  can expect  the  same  benefits from this  technology if they
are  willing to apply  it  correctly  and benefit  from  the more than 20 years of
overseas experience with  ceramic  fine  bubble  aeration systems.
REFERENCES

1.   Daigger,  G.,  Marotte,  R. and  Kohne, P., "Dome Diffuser Evaluation",
     Internal  CH2M-Hill Report  (February 13, 1981).
2.   Houck,  D.  and Boon,  A.  G.,  "Survey and Evaluation of Fine Bubble  Dome
     Diffuser Aeration  Equipment", Report for U.S. EPA, Grant No.  R806990
     (Sept.  15, 1980).
3.   Boon,  A. G.,  "Oxygen Transfer in  the Activated Sludge Process", In
     Proceedings:   Workshop  Toward an  Oxygen Transfer Standard, ASCE and  EPA,
     Pacific Grove,  California,  EPA-600/9-78-021 (1978).
4.   Yunt,  F.W.,  Letter to Mr.  Jerry Wren,  Sanitaire Corp. on the  results of
     research at  Los Angeles County (December 17, 1979).

DISCLAIMER

     The work described  in  this paper  was  not funded  by  the U.S   Environ-
mental  Protection Agency.   The contents do not necessarily reflect  the views
of the Agency, and no  official endorsement should be  inferred.
                                    268

-------
OPERATION AND MAINTENANCE  OF  AERATION SYSTEMS- AN OPERATOR'S PERSPECTIVE

J.W. Gerald Rupke,  P.Eng., President
Rupke & Associates  Ltd.,
Bradford, Ontario   LOG ICO


ABSTRACT

     This paper will  deal  with  the  most  commonly  found  problems encountered
in  the operation and  maintenance  of aeration systems  in Wastewater Treatment
Plants. The observations will be  made from the  Operator/Troubleshooter per-
spective and will be  based on the author's 16 years of  experience in trouble-
shooting and operating Pollution  Control Plants.  Each generic  type of equip-
ment will be discussed and both positive and negative aspects  noted.


MECHANICAL SURFACE  AERATORS

     There are currently a wide variety  of mechanical surface  aerators
available for the pollution control market. They  may  be fixed  mounted or
floating, low speed or high speed,  rotate in the  vertical  or horizontal
plane, be gear driven, chain  driven or belt driven, have individual drive
motors or line shaft  driven,  with or  without draft tubes,  and  even have
lower mixing impellers.

     This extreme variability makes it difficult  to generalize on the type of
problems associated with mechanical surface aerators. However, there are a
number of reoccurring problems  that warrent discussion.

     The inability  of surface aerators to transfer the  required or designed
oxygen capacity can frequently  be traced back to  the  submergence sensitivity
of  these devices. Both the power  draw and the* oxygenation  capacity are
directly related to the submergence of the aeration device  into the liquid.
Although some proprietary  devices claim  to be less submergence .sensitive,
all share the same  tendency..  Frequently  changes, in flow and/or downstream
level control devices leave the surface  aerators  with inadequate submer-
gence and a resultant oxygen  deficiency,

     The loss of buoyancy  through the accumulation of ice  or the waterlog-
ging of floatation  material produces  the opposite effect of excessive sub-
mergence in floating  aerators with  the resultant  power  overload and inade-
quate oxygenation capacity.

     Alteration of  rotational speed through gearing changes or motor changes
can frequently lead, to -loss of  aeration  .capacity.. Less  frequently, changes to
the aerator blades  imparts .a  .negative impact .
     Mechanical , prpblemsf with  th^  gearboxes or ; motors, were  at one-time  a
significant problem.  In  the  last decade  these problem areas ,have been rede-
signed resulting,  in",. long; term  trouble  free,, operation; from much of  the equip-
ment, being sold today. As, a  result, the problem off maintaining a  large number
(20 to 50) mechanical aerators at  a large  plant  are  not  as  onerous as once

                                     269

-------
was the case.

     Mechanical aerators have several inherent advantages, namely: relative-
high oxygenation efficiency, a great deal of operational flexibility,  low
impact of mechanical failure (i.e. if one of 20 aerators is out of service
the impact on treatment efficiency is low), generally can be used in existing
shallow tanks.

     The effect of wastewater characteristics such as low alpha values
frequently have less of a negative impact with mechanical aerators than in
many diffused air devices.
DIFFUSED AIR

     The use of diffused air aeration systems has been common for most of
this century in Water Pollution Conttol Plants. There are two basic system
types  available on the market, namely: course bubble and fine bubble systems,

Fine Bubble Diffusers
      Historically  fine bubble systems were plaqued with airside clogging
 problems  associated with air borne dirt. No satisfactory cleaning systems
 were  developed until  in the last decade when new materials of construction,
 such  as PVC  pipes, for air distribution and improved low cost air filtration
 systems were introduced to the market.

      The  advent of this new technology along with the need to improve the
 energy efficiency  of  diffused air systems has led to a proliferation of
 fine  bubble  air diffusion systems reaching the market. Many of the systems
 on  the market today have not fully realized the mechanical difficulties of
 operating a  diffused  air system in an aeration tank. The equipment tends to
 be  far to fragile  and subject to corrosion leading to premature mechanical
 failure.  It  is imperative that the equipment designers appreciate the need
 for long  term ( 10+ years) mechanical reliability and allow for this
 in  their  design. There is nothing more frustrating then needing to dewater
 aeration  tanks to  fix a substandard piece of aeration equipment and then
 having to redo it  six months later because the same component has failed
 again.

      The  solving of the air side plugging of fine bubble aeration devices
 has not,  however, totally solved the fouling problems. There still exists
 a significant potential of wet side plugging associated with either bio-
 logical slime growth or deposition of inorganic solids.

      The  growth of biological slimes on the surface of fine bubble devices
 is usually related to low oxygen concentrations and/or high organic loads.
Both  of these causes can be handled adequately in the design of the aeration
system by ensuring adequate DO levels are maintained at the head end of the
aeration  tank where the organic load and resultant oxygen demand are highest.
The demands  can be met by providing a tapered aeration system or step feeding
the organic  load down the length of the aeration tank.Both solutions have


                                    270

-------
been found to rectify problems with biological fouling.

     The blinding of fine bubble devices by the deposition of inorganic
solids, generally considered to be calcium carbonate, has continued to be
a problem in the area of industrial waste treatment where the dissolved
solids are high. The potential for inorganic blinding increases in direct
proportion to the level of dissolved inorganics.

     In municipal sewage treatment inorganic fouling has not commonly been
a problem and would not be anticipated unless the municipal water supply was
extremely high  (500* mg/L) in hardness, or significant industrial inputs
raised the raw  sewage dissolved solids to the 1500 mg/L range.

     A recent inspection of a one year old fine bubble aeration system in-
stalled in a high rate activated sludge plant (200 Ib BOD/1000 cu ft/day)
treating tannery wastes showed almost complete inorganic fouling of the dome
diffusers. The  ceramic diffusers were coated with a heavy layer up to  1/8"
thick of a smooth gray inorganic deposit that was readily dissolved in
dilute hydrochloric acid, generating C02 gas.

     The small  areas not covered by the inorganic solids were coated with
biological slime. During the year of operation the air pressures increased
by only 1.0 to  1.5 psi. Inspection showed that almost all the air was esca-
ping at the seat gasket in a coarse bubble pattern. No evidence of air side
fouling was noted.

     The domes  were restored to a visually clean situation by immersing them
in dilute hydrochloric acid overnight and then rinsing them with water.

     The tannery wastewater has a dissolved solids content of 8000 mg/L with
a major contributor being lime used in the dehairing process. It is anti-
cipated that inorganic fouling will reoccur if some means of in situ cleaning
is not installed. An H Cl gas cleaning system is currently under design
complete with a head loss monitoring system to ensure that remedial action
is taken before total clogging occurs. It is anticipated that perhaps even
weekly cleaning may be required.

     Mechanical inspection of the aeration system showed that several of the
support brackets had come loose from the floor by unscrewing the holddown
bolt, likely due to system vibration.

     No pipe, dome, or bolt breakage was evident.

     An adjacent equalization tank, agitated  using coarse bubble diffusers,
showed significant inorganic deposition on the sparger type diffusers but
not sufficient  to cause plugging of the air flow. A significant number of
diffusers were  plugged with organic debris associated with the tannery
wastewater.

Aeration Blowers

     The operation of small (up to 150 HP) positive displacement blowers

                                      271

-------
particularly when operating at rotational speeds of  1000 rptn or less has
proven to be very reliable. It is not uncommon to experience ten or more
years of trouble free operation with only normal lubrication schedules being
followed. As the rotational speed increases the rate of wear and the noise
level increase dramatically, resulting in more frequent breakdowns.

     The use of multi stage centrifugal blowers particularly those operating
at 3600 rpm has proven to be less satisfactory. In some cases major bearing
and/or vane  problems have been experienced yearly while other facilities
appear to have long term trouble free operation. It  is apparent that the
higher rotational speeds used in these blowers demand far closer alignment
tolerances than seen in the lower speed positive displacement blowers. The
results of any misalignment are premature motor bearing failure; the most
common operational problem with centrifugal blowers.

     In large plants the use of a small number of large centrifugal blowers
frequently leads to a severe lack, of flexibility in  the air supply system,
even when using the limited turn down capacity of these systems. Frequent
overaeration during low flow periods particularly before the plant reaches
its design load is a common problem. This does not reflect badly on effluent
quality but it does result in a significant loss of  energy efficiency. The
use of a large number of smaller blowers to allow the air supply to match
the organic load would be preferable.

Coarse Bubble Diffusers
     The mechanical and process reliability of coarse bubble diffusion de-
vices make them  ideally suited for use in activated sludge systems. Many
coarse bubble systems have been in service for 5 to  10 years without being
inspected or causing operational problems. When used in municipal wastewater
seldom is clogging a significant problem. Some industrial wastes containing
stringy organic  solids such as hair or fleshings may cause plugging of the
orfice holes.

     The major drawback to the use of coarse bubble diffusers is their low
energy efficiency. A wide variety of changes in header and diffuser spacing
as well as deep  tank configurations have made minor improvements to the
energy efficiency but has not been able to approach the efficiency seen in
fine bubble devices or mechanical aeration systems.

Hybrid Aeration  Devices

     In attempts to improve aeration electrical efficiency a large number of
hybrid aeration  devices, combining air blower or aspirators  with either
mechanical mixers or pumps, have been marketed. Some are intended to be
operated as surface devices while others are floor mounted subsurface systems.
A recent innovation of a rotating subsurface diffusion system also falls
into this category.

     From an operator's perspective these systems,although offering some
improvement in efficiency, also result in increased potential for mechanical
failure. Experience has shown that annual pump and/or blower failures are


                                     272

-------
seen with nozzle plugging not being uncommon particularly in systems carrying
high (5000  mg/L)  aeration solids.

     The use of these devices in long detention systems to prevent winter
icing problems on mechanical aerators has been satisfactory.  Parallel
operation of two long detention (5  days) activated sludge systems treating
brewery wastes in central Ontario,  one utilizing surface aerators and the
second using a jet aeration system, have shown a dramatic difference in win-
ter aeration tank temperatures.  The mechanically aerated system operated at
-1PC with almost a total ice cover  on the aeration tank.  The jet aeration
system operated at 4 C with little to no ice cover.  Both systems were fed
equal amounts of wastevater and are physically identical in size and depth.
The advantage of reduced heat loss  particularly in northern climates is
obvious from an operational viewpoint as well as providing increased levels
of biological activity.

     This advantage of reduced heat loss is shared by all subsurface aeration
devices and is used to advantage in aerated lagoons by the use of static
mixer type aeration systems.  There are a number of generic devices on the
market designed to be used in aerated lagoons to provide both mixing and
oxygen supply.  These devices share the principal that the air bubbles and
the moving water column are confined in a vertical tube and mixed by some
physical mixing system.  The aeration efficiency is improved above straight
coarse bubble aeration.

     Problems with poor anchoring system and plugging of the mixing devices
are common.  The use of these aerators in activated sludge plants has proven
to be less than satisfactory because of their inability to maintain the aera-
tion solids in suspension at normal aerator spacings.  If the aerators are
brought closer together to improve the mixing regime the aeration efficiency
is significantly reduced.
DISCLAIMER

     The work described in this paper was not funded by the U.S.  Environ-
mental Protection Agency.  The contents do not necessarily reflect the views
of the Agency, and no official endorsement should be inferred.
                                     273

-------
OPERATION AND MAINTENANCE/TROUBLESHOOTING

David Redmon
Ewing Engineering Company
Milwaukee, Wisconsin  53209

    The subject matter of this paper deals primarily with areas affecting the
operation and maintenance of  fine bubble aeration systems, particularly grid
systems employing domes, discs and plates.  The above areas, however, are
applicable to fine bubble diffusion systems in general.

    Due to the relatively high oxygen transfer efficiency of fine bubble,
ceramic grid equipment, many  engineers are considering these systems for
retrofits as well as new wastewater treatment facilities.  Considering recent
studies that indicate 50 to 90 percent of the net power demand for a treatment
plant lies within the aeration system (1), it is not surprising that energy
efficient aeration systems are being sought out.

    Various clearwater data of dome and disc fine bubble systems (2,3,4),
certainly substantiate the high clearwater efficiencies, and recent offgas
analyses of these systems indicate very efficient dirty water performance as
well, when compared to other  diffused systems studied (5).   A discussion of
the offgas technique and a highlight of the results will be given in Session
4 of this seminar.

    Concern, however, has been registered by some regarding the maintenance of
fine bubble systems owing at  least partly, to their suspectibility to clogging
and the mechanical integrity  of the system.

    Houck and Boon (6), who carried out an extensive survey of 19 wastewater
treatment plants, employing dome diffuser systems, in the United Kingdom,
Netherlands, and the U.S., indicated the need for system maintenance.  The
major operational problem reported was associated with the formation of
biological slime on diffusers operating in zones of high biological loading
and/or low dissolved oxygen.

    In January of 1982, Boyle and Redmon (7), presented a paper entitled,
"Biological Fouling of Fine Bubble Diffusers", which brings together much of
the information on biological fouling.

    Fouling may lead to a deterioration in efficiency both from a standpoint
of mass transfer and elevated pressure to operate the system.  In addition,
fouling may incur substantial costs for cleaning and maintenance.

    The economic impetus to operate these systems in an optimal fashion both
from an efficiency and maintenance viewpoint justifies careful specification
of the equipment, system survelance, and fouling diagnosis of the diffusion
system,  on an on-going basis.  The latter can be used to evaluate the degree
of fouling and system performance with time and can be used as a basis in the
optimization of cleaning cycles.
                                      274

-------
MECHANICAL ASPECTS

    In attempting to minimize operational problems and maximize dirty water
efficiency, attention should be paid to the mechanical aspects of the equip-
ment.  One should strive to achieve good distribution of airflow among the
units on a given grid through the use of properly sized individual control
orifices.  Since the wet pressure drop across the ceramic itself is quite
insensitive to flow rate, balancing is required to assure reasonable distri-
bution among the elements.  Poor distribution of air greatly reduces the
efficiency and may exacerbate the fouling problems.

    Regarding systems piped above the tank bottom, such as Norton, Sanitaire,
Rexnord, Carbo Grey, etc., the support system should be capable of withstand-
ing uplift, thrust due to expansion and contraction forces, as well as loss of
air pressure, and should be capable of simple elevation adjustment to within
4^ 1/4 inch.  The expansion joints should operate with minimal forces with or
without dependence on lubricants and be able to provide leak-free operation
at a positive internal pressure differential of up to 4-5 psi, while sustain-
ing shear loads and/or bending forces which the system will be subject to in
operation.

    For high temperature applications, special care should be taken regarding
the long term stability of the plastic systems.  At elevated temperature and/
or stress, plastic systems may represent a problem if not properly engineered
to operate in these environments.
DIFFUSER CHARACTERIZATION METHODS

    In specifying a ceramic system proper specifications of the ceramic
element itself is of great importance.  Some areas of interest include the
desired specific permeability of the element, it's wet pressure characteristics
and the uniformity of air release across an individual ceramic element.


    Ewing Engineering Company has been testing, developing and applying fine
bubble diffusers over the past 4-5 years.  During this period we have develop-
ed a series of test to quantitatively measure fine bubble diffuser character-
istics on new and fouled diffusers.

    These tests measure ceramic wet pressure drop versus flow rate, the
uniformity of the surface pores and uniformity of air release.  A general
description of the parameters, specific permeability, dynamic wet pressure,
bubble release pressure and air flow profile are presented below.

Specific Permeability

    The manufacturers of ceramic diffusers have used and are familiar with the
permeability test.  It has served as a quality control procedure to assure
that the units sent to a jobsite are similar with respect to their average
frictional resistance to flow, when dry, to wiAin some specified limits.
This was especially important in many older plants where several plates were

                                     275

-------
installed into a single plenum without individual flow balancing means toward
the objective of improved uniformity of air flow among the units when in
operation.

    The test generally consists of sealing the ceramic unit in a test fixture
substantially, as it is sealed in an actual aeration tank, and then passing
sufficient air through the dry element to produce a pressure differential of
2.0 inches water gauge.  The permeability is reported as the airrate required
to produce this differential.  In the U.S., the airrate is in SCFM, (standard
cubic feet per minute), where a standard cubic foot of air is considered to
occupy 1.0 cubic foot of volume at one atmosphere, (14.696 psi; 760 mm HG) ,
70  F, and 36% relative humidity.  Historically, the test was carried out on
ceramic plates 12 inches x 12 inches x 1 inch thick.

    In the way of an example, if we ran a permeability test on such a ceramic
plate, we might find it took 25 SCFM to produce a 2.0 inch wg. differential
pressure.  In this case, the permeability rating would be 25.  A plate of
identical material, but half as thick, would be expected to have a permeability
of 50 instead of 25, since the flow paths through the ceramic would be about
half as long and offering correspondingly less frictional resistance.   Had the
element been 1.0 inch thick,  and  had an area of 72 sq. inches instead of
144 sq. inches, the permeability would be approximately 12.5 since there would
have been only about half the area of the first case.

    Even though the above examples had identical pore size and configurations,
the permeabilities of the three vary from 12.5 to 50.  Thus, using permeability
to compare ceramic elements of different shape, thickness, materials of
construction and the like, is not meaningful and has been a confusing factor
in the engineering community.

    In an effort to employ permeability test results as a measure of resistance
characteristics of the material, we adopted the term "specific permeability",
which is the equivalent amount of air at standard conditions to produce a
2.0 inch differential pressure across the dry element if the element were
1.0 sq.  foot in area, (12 inch x 12 inch),  and 1.0 inch thick.

    An approximate expression to convert the permeability of any porous
structure to specific permeability is as follows:
                 S.P. = P x
                                Where:  S.P. = specific permeability, SCFM
                                          P  = permeability of the element
                                               itself, SCFM
                                          A  = area of element, sq. ft. when
                                               made to hypothetically conform
                                               to a flat surface
                                          t  = mean weighted thickness of the
                                               element, inches
                                     276

-------
Dynamic Wet Pressure

    The dynamic wet pressure, DWP, is the pressure  differential across the
diffusion element alone when operating  in a  submerged condition and is expressed
in terms of inches of water, at some specified  flow rate.   It differs sub-
stantially from the permeability test,  where the  specific  air flow rates are
typically much higher and where surface tension of  the submerging liquid is
not a factor.
OO Blower
[Flometer
II
If
PLAN
—O Pressure
— OTemperoti
VIEW SCHEN
Valve

Valve
IATIC
                                             'See Detail
                                                  Air Source
                         DETAIL
                                    Tapl
A
;
k
i


Mam


=
B
J
meters
Top 2 .
ILi; '••'•'•' •'•• J A
                                                    Tap 3
                                                      Bubbles
Air.Flow Control
   Orifice
          Figure 1.  Measurement of Air  Line  and Diffuser Pressure
    In the dynamic wet pressure  test, most  of the pressure differential is due
to the force or pressure required  to  form bubbles against the force of surface
tension and only a small fraction  of  the  total pressure gradient is required
to overcome frictional resistance.

    The DWP is normally relatively insensitive to airflow rate.  This results
since the surface tension effect is not greatly influenced by flow rate in
flow ranges normally applied and the  contribution of frictional resistance is
a small part of the total pressure gradient.
                                     277

-------
    That  the  frictional  resistance  is  small with respect to the  total,  can be
 shown  as  follows:

    Assume  the  typical airflow  rate at average design conditions  is  about
    2  SCPM  per  square foot  of diffuser surface area.  Since the  frictional
    resistance  is  only 2.0  in.  wg.  at  25 SCFM,  (specific permeability of 25),
    and this  unit  is typically  operated at 2 SCFM under average design  condit-
    ions, it's  theoretical  frictional  resistance at this rate is  roughly
    2.0/25.0  of 2  in. wg. or about  0.2 in. wg.  The total DWP at  2.0 SCFM  is
    about 6.9 in.  wg.  So the frictional resistance at this airflow  rate is
    roughly 3%  of  the total resistance.

    The DWP test is of value in analyzing diffuser fouling in at  least  two
 ways.  It is  a  much more sensitive  indicator of fouling than air  main pressure
 readings  and  the analysis of the DWP vs. flow relationship gives  some indicat-
 ion of the  nature  and type  of fouling.

 Bubble Release  Pressure

    The bubble  release pressure,  (BRP), test provides  a means of comparing
 the relative  effective pore diameter at any point on the surface  of  a ceramic
 element to  other point(s) on it's surface.  Swing Engineering Company developed
 this test procedure as a tool to assess the uniformity of pores on the  surface
 of ceramic  diffusers.

    The bubble  release pressure test,  as indicated by the name, measures the
 pressure  in inches of water guage required to emit bubbles from a localized
 point  on  the  surface of  a submerged  and thoroughly wetted porous  diffuser
 element.  This  is  accomplished  by forcing air at a very low rate  of  flow into
 the diffuser  and measuring  the  pressure of the air when bubbles are released
 from the  diffuser  at the point  in question.

    The test  procedure was  developed from a similar test found in ASTM, Part
 41, under Standard Test Method  E128-61.  This standard describes  a technique
 for measuring the  maximum pore  on a porous structure and uses the force to
 make bubbles  in a  known  fluid to calculate a capillary diameter corresponding
 to this force or pressure.

    The hardware employed in this test has been modified for this application.
 BRP observations of a statistically  significant number of points  on the element
 were obtained to ascertain  the  degree  of uniformity.  After determining the
 air flow distribution on several elements on which BRP data was available,  it
 became apparent that the BRP test data was a sensitive indicator  of  flow
 distribution.

    This test has been applied  to a  variety of fine pore diffuser elements.
 It has been found  to be a sensitive  indicator of the degree of fouling  of
porous diffusers and provides a quantitative means of assessing the  rate and
 location of plugging.
                                      278

-------
    Furthermore, analysis of partially fouled diffusers have given indication
that the BRP test is even more sensitive than the DWP test.  Consequently, in
diffuser fouling investigations, the use of these test methods permits signif-
icant shortening of the test period required to obtain definitive conclusions
about diffuser clogging.

Air Profile Test

    The airflow profile test uses quantitative techniques to evaluate the
uniformity of air release across the surface of ceramic diffusers while
operating rather than appraising uniformity by visual means.  This is accomp-
lished by testing the element at an air rate which is approximately equal to
2 SCFM/sq. ft., or at the recommended design rate, with anywhere from 2-8
inches of water over it.

    The rate of air release from small areas is measured by displacing water
from an inverted graduated  cylinder and recording the rate of displacement of
water with a stopwatch.  The same sites monitored in the BRP test are tested
in this procedure.  By combining the cylinder area and the rate of air dis-
charge, a flux rate, expressed as SCFM/sq. ft. or other convenient units, can
be caluculated.  By comparing the individual readings with one another, a
quantitative  measure or graphical representation of the profile can be
generated.

    The combination of DWP, BRP, and flow profile tests applied to new diffusers
and at various stages in their operating history, provides a very useful
diagnostic tool in evaluating the rate, the nature and the effect of fouling,
be it organic or inorganic, on fine bubble porous diffusion elements.  It is
also effective in appraising the effectiveness of various cleaning procedures.
DIFFUSER - FOULING

    Porous diffuser plates were used in activated sludge processes as early
as 1916 and they became the most popular method of aeration by the 1930's
(8,9).  Shortly after their installation, however, it became clear that clogging
could be a problem.  In a comprehensive review of fine bubble diffuser
servicing, Roe, (10), outlined a number of causes of diffuser clogging.  This
list has been expanded over the years and is presented in Table I.

    It was clear even in the early years of fine bubble diffusion that clogging
was highly site specific and quite often difficult to forecast (9,11).  Diffuser
clogging was chronic at some plants after only a few months of operation, yet,
at others, virtually no difficulties  with clogging arose.  Substantial effort
was applied in the early years to reducing internal clogging by means of
efficient air cleaning (12,13,14), a practice still performed today.  External
clogging was not as easy to analyze or control, however, and operations
research concentrated primarily on treating the symptom rather than the cause.
                                     279

-------
                   TABLE  I.  CAUSES OF DIFFUSER CLOGGING
     Air  Side
          -   Dust  and Dirt  from unfiltered air
             Oil from compressors or viscous air filters
             Rust  and scale from air pipe corrosion
          -   Wastewater  solids entering through diffuser or pipe leaks
     Liquor Side
             Fibrous material attached to sharp edges
             Inorganic  fines entering media at low or zero air pressure
             Organic solids entering media at low or zero air pressure
             Oils  or greases in wastewater
             Precipitated deposits, including iron and carbonates
             Biological growths on diffuser media
     In  today's plants  air side fouling has been all but eliminated except for
 a few cases where mixed  liquor leaks into the system through a bad joint or
 broken  pipe.  In a  few cases, however, we have observed the presence of fibers
 from air silencers  on  the internal surfaces of tubes and discs.

     By  and large, the  majority of fouling that is observed today involves
 liquor  side fouling.   This  fouling is conveniently divided into inorganic
 fouling, as a result of  the precipitation substances including carbonates and
 iron, and biological fouling as a result of the development of biological
 growths or slimes on the fine pore media.

     An  excellent reference  on inorganic clogging is a paper published by
 Morgan  (15).  He investigated and reported on clogging related to the presence
 of iron, calcium bicarbonate, fine sand in the mixed liquor and particulate
 matter  in the air supply.   Regarding iron he indicates information on iron
 clogging has confirmed that which was known from plant experience, that
 ferrous iron will result in external clogging.  For plants adding iron salts
 for phosphorus removal this may be a problem and should be considered.  Morgan
 indicates that clogging  caused by calcium bicarbonate in the water system is
 not well understood, probably because it is not commonly encountered.  Of
 these bicarbonates, only calcium is objectionable and is found in troublesome
 concentrations in few  cities.  In industrial applications, such as the pulp
 and paper industry, calcium fouling may be more likely to ocur.

     Based on our experience and that of Houck and Boon (6), it appears that
 liquor  side fouling related to the occurrence of biological growths on the
media represents  a  much more significant problem than does the inorganic
clogging.  In the Houck  and Boon survey (6), it is reported that the single
greatest operational problem encountered in the survey was the formation of
biological growths on  the diffusers operating in zones of high volumetric
loading  and/or low dissolved oxygen.  In only two cases, however, were domes
observed to contain excessive amounts of biological slime,  (Beddington, U.K.,
and Madison,  WI).  All other plants apparently exhibited signs of "coarse


                                      280

-------
bubbling" which the authors attributed to the development of biological slimes.
The authors further cited pilot plant work at the Water Research Centre at
Stevenage, U.K., which verified the causes of diffuser sliming as the result
of very high loading rates (12).  No data was presented, however.

    Boyle and Redmon (7), drawing from the literature and their experience,
reported causes of biofouling, biofouling agents, the effects of biofouling
and design for control of biofouling.  Some of the suspected factors affecting
biofouling are presented in Table 3.
             TABLE III.  SUSPECTED FACTORS AFFECTING BIOFOULING
            Low permeability of diffusers  (16, 11)
            Low air flow rate per diffuser  (12)
            High organic loading  (6, 12)
            Low dissolved oxygen concentration  (6)
            High temperatures  (12, 14)
            Uneven air distribution among diffusers  (10, 12, 16)
            Uneven air distribution within a diffuser
            Waste Characteristics
    Boyle and Redmon (7), point out that the development of a biological
growth on diffuser media could have several effects on diffuser performance.
Two possible scenarios were given. One involved a gradual build up of biolog-
ical growths where the bioslime leads to bubble coalescence and little
noticeable change in DWP is observed until the slime envelopes the entire
media surface.  Umler these conditions, the uniformity of the air released
from the diffuser may be detrimentally affected and poor distribution may
result.  The OTE may decrease as bubble coalescence occurs and there could
be visible changes in bubble patterns and the appearance of "coarse bubbling"
at the surface.  This type of diffuser cloggging may not be evident as
routinely measured by air line pressure losses or demand for  more air,  (as a
result of some loss in transfer efficiency).  The second scenario involved
biofouling simultaneous with internal clogging or external encrustation.  In
this case it was hypothesized that the larger pores would plug first resulting
in the progressive utilization of the finer pores producing,  perhaps, more
uniform distribution and finer air bubbles.  Increases in DWP could be small
at this point.  Eventually the biological mat may result in bubble entrapment
and coalescence.  The net effect of this progressive clogging might be an
increase on OTE up to a point, (as air distribution improves and bubble  size
decreases), followed by a drop in OTE as coarser bubbles begin to form.
Dynamic wet pressure would slowly rise to a point where only the finer pores
are carrying air; after this, pressures would rapidly rise.
DIFFUSER SURVEILLANCE

    Considering that the nature and magnitude of fouling is highly site
specific and difficult to forecast (9,11), it seems desirable to install a

                                     281

-------
 test  header in an existing aeration basin to  identify  the  rate of clogging
 and nature of the foulants.   Using the  procedures  previously described,  a
 rapid indication of the rate of clogging can  be  obtained.   In addition,
 amenability to cleaning and definition  of optimum  cleaning methods can be
 explored.

    Sanitaire, one of our clients, has  done a significant  amount  of this
 testing.   In some test installations the rate of pressure  rise was very  low
 and in some others, such as a potential industrial application, a 20-40  inch
 water gauge rise was observed over a period for  four to  six weeks.  The  latter
 case  would be one where the use of fine pore  media would not be a good
 selection.

    Similar diffuser monitoring for an  operating ceramic grid system can also
 be quite beneficial.   This monitoring can be  done  on several diffusers in the
 operating  system and/or in conjunction  with a small test header.   Under  this
 condition  one can characterize several  ceramic elements  for DWP vs.  flow,
 BRP and airflow uniformity prior to installation and monitor DWP  at a  fixed
 air rate over a period of time to observe the rate of pressure increase.
 Figure 1 illustrates the pressure measurements needed to do this.

    In conjuction with the observations of DWP with time, measurement  of
 oxygen transfer efficiency with time can help identify the  appropriate time
 for cleaning.   One way to accomplish this is  the measurement of local OTE and
 uniformity of air release by the offgas apprach.   This technique will be
 described  in Session 4.

    Using  data generated as  above,  the  operator  would have  sensitive objective
 information upon which to determine the extent of  fouling.   Using  a small test
 header in  conjunction with the operating system  permits  visual observation of
 the foulants by removing the small header from the  aeration  basin.  The
 degree from fouling and the  most appropriate  cleaning procedure can be
 ascertained from the test diffuser.

    One apparent problem with many of the existing  ceramic  grid systems is
 that  the existing monitoring methods of measuring  line pressure are not
 sufficiently sensitive,  and  the operating'personnel do not decide  to clean
 the system until the  system  is substantially  plugged.  In this condition, it
 may be very difficult or impossible, to restore  the system.  Careful survelance
 in  conjunction with an active maintenance program which  attempts to keep the
 system in  relatively  new operating condition  is  considered  a proceedure.
DIFFUSER CLEANING

    Depending upon the nature and extent of clogging as well as the diffuser
materials and system make up, there are various cleaning methods available.
A few methods that come to mind include sand blasting, brushing, steam cleaning,
surface cleaning followed by acid soaking, retiring with or without prior
acid soaking, ultrasonic cleaning, high pressure water sprays, and in-situ
cleaning with acidic gases.  Except for the last method, all the other methods
require process interrruption.

                                      282

-------
    The most popular techniques in present use seem to be cleaning the media
surface by brushing, hosing, or the like, followed by acid soaking.  Houck
and Boon (6), report dome cleaning at the Beckton Plant in the United Kingdom,
which includes  surface cleaning, retiring and acid washing.  The refiring is
done on site in their refiring furnace.  In the U.S., refiring has not been too
prevalent, but some manufacturers  are suggesting it in extreme cases.  Whether
this is a cost effective procedure in the U.S., where the ceramic elements
must be shipped back to the manufacturer, remains to be seen.

    A relatively recent innovative approach regarding cleaning, developed by
Sanitaire (17),  involves injection of gases which are aggressive to the
fouling agents.  In most cases, HC1 gas is introduced into the supply air which
results in an acidic reaction product when the gas contacts the water entrained
in the element.

    The philosophy of the gas cleaning approach is to carefully observe the
dynamic wet pressure of the system and clean at such frequency to maintain the
media in a like new condition.  Cleaning at this frequency should prevent the
system from ever becoming substantially clogged, thereby maintaining it's
oxygen transfer efficiency and prolonging the life of the porous media.
Several tests of this procedure have been tried on a small scale, and in the
great majority has been very effective in reducing the system back pressure
requirements and maintaining uniformity of air distribution.

    When using this method, it is important that the materials of construction
are resistant to the treating gas.  Also of importance, is the uniformity
of air release within the element and the distribution of air among the system,
since the gas acts primarily at the active sites on the diffuser element.
Thus, for a system with poor uniformity across the media and between diffusers,
gas cleaning will preferentially clean those areas of highest flow and may not
result in substantial improvement of system performance.
FUTURE RESEARCH

    The impact of long-term fine bubble diffuser operation in wastewater with
respect to maintenance requirements  and OTE is still poorly documented.  It
is clear that some biological growth will develop on diffuser surfaces but
the impact of that growth on performance is unclear.  Research should be
directed under controlled experimental conditions toward measuring the
progression of biological fouling employing sensitive BRP and air flow
distribution techniques.  These studies should include a variety of diffuser
materials and geometry.  The selection of diffuser systems is often predicated
on the performance of clean diffusers in clean water.  The long term changes
in OTE during diffuser operation in wastewater should be more clearly
delineated in order to provide the engineer with a clearer picture of what
happens to OTE over the design life of the diffuser.

    In addition, studies to determine the optimum pore size or specific
permeability, that is the pore size which optimizes the rate of fouling and
OTE, would be of real value.
                                     283

-------
    As energy costs continue to escalate, engineers will seek more effective
energy conserving systems for the transfer of oxygen to wastewater.  Until a
better understanding of the behavior of fine bubble diffusers under long term
field conditions is realized, doubts will continue to be raised regarding the
effectiveness of the fine bubble system.  Futhermore, mis-application of
these systems in certain situations may result in a greater expenditure of
manpower and energy than necessary.


REFERENCES

1.  Wesner, G.M., et al.  Energy Conservation in Municipal Wastewater
    Treatment.  EPA 430/9-77-011; Office of Water Program Operations;
    U.S. EPA: (March, 1978).
2.  Yunt, P., et al.  An Evaluation of Submerged Aeration Equipment -
    Clean Water Test Results, WWEMA Industrial Pollution Conference,
    Houston, Texas, (June 5, 1980)-
3.  Huibregtse, G.L., Rooney,  T.C., and Rasmussen, D.C., Factors
    Affecting Fine Bubble Diffused Aeration Performance, 54th WPCF
    Conference, Detroit, Michigan, (October 7, 1981).
4.  Paulson, W.L.,  Oxygen Absorption Efficiency Study - Norton Co. -
    Dome Diffusers, Morgan Laboratory, (March, 1976).
5.  Redmon, D.T., and Boyle, W.C., Preliminary Findings:  Offgas Analysis,
    a Report to the ASCE (Oxygen Transfer Standard Committee,  Detroit,
    Michigan, (October 6, 1981).
6.  Houck, D.H. and Boon, A.G.,  Survey and Evaluation of Fine Bubble Dome
    Diffuser Aeration Equipment.  EPA 600/S2-81-222;  Municipal Environ.
    Research Lab;  U.S. EPA (October, 1981).
7.  Boyle, W.C., and Redmon, D.T., Biological  Fouling of Fine Bubble
    Diffusers, EPA National Conference - Operation and Maintenance of
    POTWs  (January 12, 1982).
8.  Beck, A.J.,  Diffuser Plate Studies, Sew. Wks.  Jour., 8:22 (1936).
9.  Committee on Sewage Disposal.  The Operation and Control of Activated
    Sludge Sewage Treatment Works,  Sew. Wks.  Jour., 14:3 (1942).
10. Roe, F.C., The Installation and Servicing of Air Diffuser Mediums,
    Water and Sew. Wks., 81:115 (1934).
11. Wisley, W.H.  Summary of Experience in Diffused Air Activated Sludge
    Plant Operation, Sew. Wks.  Jour., 15:909 (1945).
12. Setter, L.R., Air Diffusion Problems at Activated Sludge Plants,
    Water and Sew. Wks.,  95:450   (1948).
13. Setter, L.R. and Edwards, G.P-,  Experimental Laundering of Air Diffuser
    Plates, Sew. Wks Jour., 17"867 (1945).
14. Anderson, N.E.  Tests and Studies on Air Diffusers for Activated Sludge,
    Sew. and Ind. Wastes,  22:461  (1950).
15. Morgan, P.F., Clogging Studies of Fine Bubble Diffuser Media,  Jour.
    Water Poll.  Cont. Fed., 31:153   (1959).
16. Lamb, M.   Designing and Maintaining Porous Tube Diffusers, Wastes
    Engineering,  25:405 (1954).
17. Schmit, F.L., In-Place Ceramic Diffuser Cleaning System,  EPA: National
    Conferences on Innovative & Alternative Technology,  Various Locations,
    (1980).
                                     284

-------
DISCLAIMER

     The work described in this paper was not funded by the U.S.  Environ-
mental Protection Agency.  The contents do not necessarily reflect the views
of the Agency, and no official endorsement should be inferred.
                                    285

-------
SELECTED EXPERIENCE WITH AERATORS USED IN
THE TREATMENT OF PAPER INDUSTRY WASTEWATERS

James J. McKeown, NCASI
Tufts University, Anderson Hall
Medford, Massachusetts  02155

INTRODUCTION

     It was mentioned in a previous paper (1) that the selection
of aerators for use in treating pulp and paper mill wastes is
based on a series of considerations, only one of which is oxygen
transfer.  Many times it is of equal or greater importance that
the machine selected perform reliably, require a minimum of
maintenance and that it be a cost effective selection.

     This paper presents the operational experience from eight
out of over 400 plants which use aerators in treating paper
industry wastewaters.  Although only a few plants are presented,
their problems and solutions are representative of those
uncovered at other plants employing similar equipment during sim-
liar periods of time.  No attempt has been made to cover all of
the types of aerators in use within the industry or all of the
manufacturers which supply the industry with aeration equipment.
Further, it is expected that improvements in machinery design,
inspection procedures and operational controls provide today's
purchasers with a more reliable machine for use in industry and
municipal wastewater treatment systems than may have been the
case in the 1970's.

GENERAL EXPERIENCE

     The paper industry has had considerable experience operating
aeration equipment for the purpose of reducing the BOD of its
wastewater.  For example, it is estimated that over 200,000
horsepower are currently employed to aerate about 4000 MGD of
wastewater in the U.S. paper industry.

     In the paper industry, it is common practice to interview
operations personnel at many sites before a choice of aeration
equipment is made for a new site or an expanded site.  These
interviews provide the industry with up to date information on
the performance of equipment as well as document problems in
need of correction.  In addition, the summaries of NCASI and
other workshops often contain an organized review of operating
and maintenance information which is useful in surveying the
experience of others.

     It should be mentioned that the paper industry employs
mostly surface mechanical aerators in its wastewater plants.
Therefore, the experience collected is largely directed toward
this type of aerator.

                              286

-------
Gear Wear

     Experience with operation of gear reducers on slow speed
large diameter aerators is quite varied within the paper indus-
try.  For example, some units have operated with the original
gear boxes for over twenty years, whereas, others require com-
plete replacement because of wear on gears and bearings after
only a short period of operation.  This problem has been the
most frequently reported difficulty encountered with aerators in
the paper industry.  Where gear systems fail, the cost of re-
placing the parts is often sufficient to consider purchasing a
new aerator.  The case histories presented later illustrate this
practice.

Fastening System Failures

     Some fastening system problems reported in the paper indus-
try have involved stationary mixers fastened to an air header
resting near or on the bottom of the basin.  The vibrations
caused by turbulence in transporting air and water have been
sufficient to cause fracture in some fastening systems and
fatigue in others, resulting in losing the aerator.

     Although not very prevalent in the paper industry, dual
blade mechanical aerators have lost blades and experienced
twisted shafts.  The cause is usually related to improper
fastening of the lower shaft resulting in uneven distribution of
forces along the shaft.

     Although only a fastening problem in the broadest sense,
the action of ice on various structures which support the aer-
ator can cause the development of uneven forces on the aerator
shaft and in a few cases has caused the capsizing of floating
aerators.  In many locations where freezing is normal, pier
corrals or cable supports have been provided to prevent floating
units from tipping.  Also, fixed mounted aerators, although not
subject to turnover, can be shielded from accumulating large
blocks of ice which, if loosened, could cause damage to an
impeller.

Seal Failures

     High speed direct drive surface aerators are usually floated
on the water surface.  There has been only an occasional report
of failure of the seal which protects the motor from contacting
water on this type of device.

Construction Material Failures

     Paper industry effluents may contain clay and fly ash which
are abrasive to certain materials of construction.  Also, cer-
tain wastewaters may be corrosive to unprotected mild steel.
Aerators may also be placed in natural basins where stumps and
                               287

-------
debris were never removed.

     These wastewater and system characteristics have caused iso-
lated problems, some of which are presented in the case histories
which follow.  In most cases, when surfaces are frequently inspec-
ted for wear and corrosion, the damage to concrete piers, steel
pontoons and the like can usually be reduced by taking some pre-
ventative action.

CASE HISTORIES

     The paper industry is used to operating high speed mechan-
ically complex equipment  in processing pulp and manufacturing
paper.  Skilled personnel, well equipped shops and maintenance
machines are available at most plants.  Furthermore, the pro-
cesses are operated continuously and personnel are familiar with
making repairs during operating days in a timely manner.  The
maintenance of a mechanical aerator or a blower is not much
different from other machinery, once its operating principles
become familiar.

     This portion of the paper assembles several case histories
of aerator operational problems taken from contact with industry
representatives.  Most of the cases reviewed cover relatively
long histories of experience with aerator operation.  Although
these situations don't cover every problem that might arise or
represent the experience of the entire paper industry, it  is
hoped that they present the user's perspective which will be
helpful to others operating these systems.

Mill A

     This bleached kraft mill designed its own aerator in the
late 1950's which consisted of a top and bottom impeller located
on a shaft which rested on the bottom of the basin.  Each of
four activated sludge basins eventually contained three 25 HP
and two 20 HP units which were powered by a motor and gear re-
ducer system.  The mill is still operating these units, although
after over 25 years of operation they are now being replaced
with larger commercial aerators.

     Mill personnel believed that these units required relative-
ly high maintenance and are hoping to reduce that maintenance
with the new units.  The shop built units required frequent re-
placement of the foot bearing due to wear. Also, occasional
shaft failures and gear reducer bearing failures were experi-
enced.   Although each gear reducer bearing performed with a dif-
ferent rate of wear, the average interval between replacements
was about three years.

Mill B

     Mill B operates an activated sludge plant with about 800
stationary bubble mixers  installed in a 2.5 acre basin which is
                               288

-------
12 feet deep.  During the first two years of operation, all of
the original fiber glass mixers were worn out due to erosion
from clay in the effluent.  The underwater piping was also sev-
erely damaged when hold down straps cut through the PVC piping.

     Replacement of the system during operation required division
of the basin by a massive concrete wall and repiping of the basin
so that each part could be operated independently.  After the
wall was in place, half of the basin was taken out of service,
pumped down, and the aerators repaired.  The PVC piping and fiber
glass mixers were replaced with duplicate equipment manufactured
from polyethylene.  Problems with breakage of the mixer support
legs were experienced and these legs have been strengthened or
replaced.

     Pluggage of the air distribution piping is a minor operat-
ing problem which increases with time.  The underwater piping
has been cleaned by divers periodically while the basin is in
operation.  Increasing air pressure due to plugging of the air
distribution piping has caused a surging condition in the centri-
fugal air blowers.  Surging can and did destroy the blowers.
Prevention was easily obtained by installation of low amperage
shutdown of the blower  motor.  Also, when down, the air piping
is cleaned to reduce the back pressure.

Mill C

     A key element in Mill C's activated sludge process is its
17-foot deep aerated basin.  Oxygen transfer is provided by 100
HP low (single) speed surface aerators.  Adjustment in aerator
performance is accomplished by adjusting a variable level over-
flow weir in the aerated basin.  The three goals in aerator oper-
ation are:

     (1)  To maintain in suspension the required level of
          biological solids to accomplish BOD removal.
          Additional concern was expressed in minimizing
          any potential deposition in the aerated basin
          due to the carryover of solids from primary
          clarification.  The activated sludge process at
          this location is treating a deink mill wastewater.

     (2)  To maintain sufficient dissolved oxygen in the
          mixed liquor of the aerated basin.

     (3)  To minimize the stress placed on the aerator
          gear box (reducer) by the varying hydraulic load
          resulting in a fluctuating submergence on the
          rotating aerator element.

     It was pointed out that the above goals could be implemented
by manipulating the adjustable overflow weir; however, the re-
sponse time was too long, particularly in minimizing the "wear
and tear" on the aerators.  The alternate solution was to remove
                               289

-------
some of the aeration blades  (scoops) to reduce amperage and gear-
box stress.  No reduction in process performance has been ob-
served as a result of this modification.  It should be noted,
however, that residual dissolved oxygen in the aerated basin
prior to the aerator modification was in the 4 mg/1 to 5 mg/l
range.  This allowed the opportunity to sacrifice some oxygen
transfer capability in an attempt to achieve goals 1 and 3 cited
above.

Mill D

     Mill D is a largely intergrated kraft pulp and paper mill
producing 550 tons per day of fine printing and specialty grades.
The mill generates approximately 17 MGD of wastewater with an
average untreated BOD of 40,000 pounds per day which is treated
by a conventional actived sludge plant.  The design flow and BOD
to the plant is 21 MGD and 50,000 pounds per day, respectively.

     Two rectangular basins each 400 ft. long by 120 ft. wide
are operated in parallel.  Aeration basin detention times are 9
to 10 hours at normal flows.  Mixed liquor suspended solids are
maintained at approximately 3500 mg/1 with 50% of forward flow
recycled back to the aeration basins.

     Eight pedestal mounted mechanical aerators, four units per
basin, provide aeration for the system.  The electrical drives
are two speed motors that have a nameplate output of 150 HP on
high speed and 75 HP on low speed.  The aerator gearbox provides
a 38 to 1 speed reduction with a final output speed to the aer-
ator impeller of 48 RPM on high speed and 32 RPM on low speed.

     In general, it had been experienced that the gear reducers
gave very reliable service, provided they were properly lubri-
cated and cooled.  Therefore, no spare gearboxes were ordered.
Some problems were anticipated with the electrical drives and
one standby unit was purchased.   However, although no major prob-
lems developed with the electrical drives, five out of eight
gearboxes failed since startup in May of 1976.  Two units failed
in January of 1978, and two more failed in May that same year.
At some points, the replacement parts had not been received
which resulted in having only 50% of design aeration capacity
available.

     An attempt was made to operate with only two aerators in
each basin,  but no matter how functional units were positioned,
there were just too many dead zones in the basins and the efflu-
ent quality,  particularly turbidity and suspended solids, became
noticeably affected.   Reducing the mixed liquor concentration
had no detectable impact upon effluent quality.

     During this period, mill maintenance crews became very
skilled at jockeying aerators around so it was decided to divert
all of the flow to one basin and switch all of the remaining
                               290

-------
aerators to that basin.  This, in effect, reduced the detention
time in the aeration basin to slightly more than four hours, but
with the aerators on high speed there were no dead zones and ef-
fluent quality improved, but not to required levels.  Operations
continued in this mode for approximately one month until replace-
ment parts for two units were received.  With repair of these
two units, the operational mode was changed to three aerators
per basin and effluent quality returned to normal levels.

     In August of 1979, just after replacement parts for the last
two gearboxes had been received, a fifth unit failed.  In all
cases, the failures have been very similar; the teeth from the
intermediate gears were stripped.  Each time the gears were re-
paired, if the teeth on one gear were damaged, the gear to which
it meshes, even though visibly undamaged, was replaced.

     Shortly after the aerator gearboxes began to fail, person-
nel became concerned about the service life of the remaining
units.  For our particular system, several measures were taken:

(1)  The aeration basin effluent weirs were lowered to their mini-
     mum attainable levels.  This reduced the effective
     water level by 2.5 inches.

(2)  The mixed liquor suspended solids were reduced by 1,000 mg/1
     and the recycled sludge rate was lowered to only 30% of for-
     ward flow.

(3)  The aerator impeller shafts were shortened by 3 inches.  This
     was the most effective measure taken, but also the most de-
     manding in terms of mill maintenance crews.

(4)  The aerator gearboxes were raised by about 1/2 inch by
     turning the adjustment nuts on the support bolts.

     Prior to the development of the aerator gearbox failures,
the aerators were set at an average submergence of five inches
below the water surface, but after the above corrective measures
were taken, the average aerator setting was one inch above the
water surface.  A recording wattmeter was placed on each aerator
for a minimum of several days to insure that the aerator motor
was drawing less than 150 HP.

     After having spent many frustrating hours dealing with the
problem of failing aerator gearboxes, mill personnel suggest the
following guidelines for future installations:

(1)  The aeration basins were equipped with DO probes connected
     to an automatic controller that would check DO levels
     every 20 minutes and change aerator speed from high to
     low or vice,versa.  This system should have a time delay
     relay to insure that when going from high speed to low speed
     the impeller shaft has stopped, or slowed down, to an RPM
                               291

-------
      less than slow speed  before  the  motor  actually  switches
      slow speed.   It is felt that the lack  of  such a system is
      probably the most important  single  factor in the piematui-e
      failure of our gearboxes.

 (2)   Each aerator should be  started up with normal flow patterns
      and the submergence set by the use  of  a recording wattmeter.
      After each unit is set, it should be left connected to the
      wattmeter for at least  48  hours.  All  procedures should  be
      clearly documented and  saved for future reference.

 (3)   Each aerator should be  equipped  with its  own ammeter.  How-
      ever, do not rely totally  on ammeter readings.

 (4)   Consider the purchase of a complete set of internal gears
      and bearings to act as  a spare.

 (5)   Be aware that the effective  water level changes with flow.
      In this case, the difference between zero flow  and  average
      flow conditions was 2.5 inches.

 (6)   Finally, if  any warranty agreement  is  reached that  is  even
      a little bit different  than  the  standard  warranty,  make
      absolutely sure that  these differences are very specific and
      in writing.

 Mill  E

      Mill E  has recently installed a  new aeration system in its
 activated sludge  plant.  After  one year of  operation, it was  dis-
 covered that the  lower  shaft has  sheared off at the  flange  below
 the surface  impeller.   These dual blade aerators contain a  small
 mixing impeller at the  base  of  the shaft.   Mill personnel did
 not notice that the  mixing impeller was no  longer attached  be-
 cause  plant  performance  was  not affected.   The shafts have  been
 straightened  or replaced and operators periodically  check to  see
 if they are  still  attached.   The  failure is believed due to im-
 proper fastening of  the  lower shaft during  construction.

 Mill F

     This mill  has operated  an  aerated stabilization basin  since
 1964.    The initial aerators  were  60 HP slow speed units  floated
 upon hollow  steel  pontoons.   Maintenence of the pontoon  surface
 to prevent corrosion was a continuous  problem.

     The  gear reducers  failed on  several of the aerators begin-
 ning in 1967.   Initially,  these units  were  replaced  with 20 HP
 high speed floating  aerators and  later with 50 HP high speed
 units.  However,  the  first lot  of 50  HP units  were subject  to
major  bearing problems  which eventually  led to the manufacturer
 installing an added  shaft  support to  the bearing system which
 solved  the problem.
                               292

-------
     When mill capacity increased in 1972, a group of 75 HP high
speed units with stainless steel floats were installed.  These
units developed problems in keeping the impeller pinned to the
shaft.  Eventually, a double pin was used which solved this prob-
lem.

     As of 1982, the mill currently is no longer using the orig-
inal slow speed units.- In their place most of the 20 HP and 50
HP units operating, and the load related to all recent increases
in production is being satisfied with new 75 HP high speed units.

Mill G

     This mill utilizes an aerated stabilization basin which in
1971 utilized nine 125 HP floating slow speed units.  The gear-
boxes started breaking down after 1.5 years operation.  Problems
with their repair by an outside contractor prompted mill per-
sonnel to make their own repairs.  Two additional slow speed
units were added in.1975.

     In 1978, five 75 HP high speed aerators were installed which
were anchored to the bottom by three or four cables.  Since that
time, these units have operated with few problems.  One bearing
burn-out occurred and occasionally an anchor cable breaks.  Later,
an additional group of 75 HP units replaced more of the original
units as gear boxes wore out.  The mill engineers believe this
policy will continue as long as it is cost effective.

Mill H

     This bleached kraft mill installed its activated sludge sys-
tem in 1960 which incorporated 12-60 HP dual speed fixed mounted
aerators.  Another six of the slow speed aerators were installed
in 1966.  The system was DO controlled with a 20-30 minute delay
built into the switching system. Because of the wear caused to
the aerator gearing system, the DO control was performed man-
ually rather than automatically.  A two hour cycle replaced the
half-hour cycle.

     The mill began replacing the old units in 1976 with high
speed floating aerators because it was less expensive than re-
placing gear sets.  Also, by this time, the motors on the old
units were showing signs of severe corrosion.

     In 1976, the mill also installed the first of 240 station-
ary bubble mixers.  Initial experience with the bubble mixers
produced fastener problems in that the bolts sheared at the
plate fastener because the bolt size was 1/16" less than the hole
in the plate.  A tighter tolerence solved the problem.  Also,
the air headers required additional bracing to hold them in
place.  Since these early problems, only 14 aerators have been
lost.

                               293

-------
SUMMARY

     A general review of the types of problems requiring aerator
replacement or overhaul has been presented with the aid of eight
case histories which illustrate the class of problems encountered
in aerating paper mill wastewater.  The problems were classified
as ones largely involving (a)  gear wear or (b) the failure in
systems which fasten the aerator to the tank,  basin or air head-
er.  Additional problems involving using incompatable materials
of construction and the failure of motor/shaft seals were occa-
sionally reported.

REFERENCES

(1)  McKeown, J.J. "Notes for  Workshop Toward  an Oxygen
     Transfer Standard", Proc.  Workshop Toward an Oxygen
     Transfer Standard, EPA-600/9-78-021,  p.  228, (April 1979).

(2)  "Operation and Maintenance of Aeration Equipment with
     Emphasis on Avoiding Mechanical Failures",  Proc. of the
     1979 NCASI Northeast Regional Meeting, 260  Madison Ave.,
     N.Y., Special Report No.  80-03 (Feb.  1980).

DISCLAIMER

     The  work  described  in this paper was  not funded by  the U.S.
Environmental  Protection Agency.  The contents do not  necessar-
ily reflect  the views  of the Agency, and no official endorsement
should be inferred.
                               294

-------
PROPOSED STANDARD FOR MEASUREMENT OF OXYGEN TRANSFER IN CLEAN WATER

C. Robert Baillod, Prof,  of Civil Engineering, Michigan Technological  Univ.,
Houghton, Michigan  49931
Wayne Paulson, Prof, of Environmental  Engineering, University of Iowa, Iowa
City, Iowa  52242


ABSTRACT


     This standard was developed to measure the rate of oxygen transfer
from diffused gas and mechanical oxygenation devices to relatively large
volumes of water (up to 1 million gallons) typical of those found in the
activated sludge wastewater treatment process.  It is intended that this
standard be referenced in engineering specifications for compliance
testing.

     This test method is based upon removal of dissolved oxygen (DO) from
the water volume by sodium sulfite followed by reoxygenation to near the
saturation level.  The DO inventory of the water volume is monitored
during the reaeration period by measuring DO concentrations at several
points selected so that each point senses an equal tank volume.  The
method specifies a minimum number, distribution, and range of DO measure-
ments at each point.

     The data obtained at each determination point are analyzed by a
simplified mass transfer model to estimate the apparent volumejric mass
transfer coefficient, K. a, and the saturation concentration, Coo.  Non-
linear regression is employed to fit the model to the DO profile measure^
at each point during reoxygenation.  In this way, estimates of K, a and C^,
as are obtained at each point.  These estimates are adjusted to standard
conditions and the standardized oxygen transfer rate (mass of oxygen dissolved
per unit time in clean water under standard conditions of temperature and
pressure at an hypothetical concentration of zero DO) is obtained^as the
product of the average adjusted K,a value, the average adjusted C^ value
and the tank volume.  A procedure based on the clean water test results
is prescribed for estimation of oxygen transfer rates in dirty water
under process conditions.

     Various components of power consumption are defined and methods
for measurement of gas rate and power consumption by the oxygenation
device are given.  Energy efficiency of the oxygenation device is
evaluated as the mass rate of oxygen transferred per unit power consumed.
                                     295

-------
       Measurement  of  Oxygen Transfer  Rate in Clean Water
      This method  covers  the measurement of the oxygen transfer rate
 OTR,as a mass  of  oxygen  per unit time dissolved in a volume of water
 by an oxygen transfer  system operating under given gas rate and power
 conditions.  Methods for measurement at gas rate and power are also
 described.  It is intended primarily to measure the rate of oxygen
 transfer from  diffused gas and mechanical oxygenation devices to
 relatively large  volumes of water  (up to 1 million gallons) typical
 of those found in the  activated sludge wastewater treatment process.
 Although the method is intended primarily for clean water, it is
 applicable to  dirty water when the rate of oxygen demand is negligible.

      The method is applicable to tests in full scale field facilities,
 in factory shop tanks  and in small laboratory tanks and is valid for
 both completely mixed  and incompletely  mixed tanks.  The primary result
 is expressed as the Standardized Oxygen Transfer Rate, (SOTR), a
 hypothetical mass of oxygen transferred per unit time at zero dissolved
 oxygen concentration,  water temperature of 20°C and barometric pressure
 of 1.00 atm, under specified gas rate and power conditions.  The
 results can  be applied to estimate oxygen transfer rates in dirty
 water at process  dissolved oxygen  (DO) concentrations.


 2.   Summary  of Method


      The test  method is  based upon removal  of dissolved oxygen (DO) from
 the  water volume  by sodium sulfite followed by reoxygenation to near the
 saturation level.  The DO inventory of the water volume is monitored
 during  the reaeration  period by measuring DO concentrations at several
 determination  points selected so that each point senses an equal tank
 volume.   These  DO concentrations may be either sensed in situ using
 membrane  probes or measured by the Winkler method applied to pumped
 samples.  The method specifies a minimum number, distribution and range
 of DO measurements at each determination point.

     The  data obtained at each determination point are then analyzed by
 a simplified mass transfer model to estimate the apparent volumetric
mass transfer coefficient, K.a, and the saturation concentration, C^.
The basic model is described in Reference 1 and is given by
                                    296

-------
          c  = c! - 
-------
      4.1.3  Delivered Shaft Power.   Shaft power delivered  to  a  turbine
 or other impeller; calculated based upon shaft torque  and  angular
 velocity measurements.

      4.1.4  Pump Delivered Power.   Water power delivered by a pump;
 calculated based on the volumetric  flow rate,  weight density  and  total
 dynamic head.

      4.1.5  Total Delivered Power.   Sum of delivered gas power  plus,
 delivered shaft power plus pump delivered power.

      4.1.6  Delivered Standard Air  Power.   Delivered gas power  calcu-
 lated based on adiabatic compression of air from  14.60 psia to  15.70
 psia plus pressure loss through the diffuser psi  plus  hydrostatic
 pressure at diffuser submergence psi.

      4.1.7  Delivered Standard Shaft Power.  Shaft power required to
 turn on impeller in water at 20UC at an angular velocity equivalent to
 the actual angular velocity.

      4.1.8  Pump Delivered Standard Power.   Pump  delivered  power  when
 the specific gravity of water is 1.00.

      4.1.9  Total Delivered Standard Power.  Sum  of delivered standard
 air power plus delivered standard shaft power  plus pump delivered
 standard power.

      4.1.10  Hire Power.   Electrical  power  drawn  by a  motor.


      4.1.11  Wire Standard  Power.   Wire power  drawn by an electrical
 motor of standard efficiency  and transmitted by a system at standard
 efficiency to achieve  a  given  level  of  total delivered standard power.
                                                                *
 4.2  Mass  Transfer Terms

      4.2.1  Oxygen  Transfer Rate (OTR).  Mass  of  oxygen per unit  time
 dissolved  in  a volume  of water  by an oxygen  transfer system operating
 under given conditions of temperature,  barometric pressure, power, gas
 rate  and dissolved  oxygen concentration.

      4.2.2  Oxygen  Transfer Rate at  Zero DO  (OTR  ).  OTR when the DO
 concentration is  equal to zero  at all points in the water volume.

     4.2.3  Oxygen  Transfer Rate in  Dirty Water (OTRf).  OTR  for  the
oxygenation system  operating in dirty water at a  specified  average DO
concentration and temperature.
                                     298

-------
     4.2.4  Standardized Oxygen Transfer Rate (SOTR).  OTR in clean
water when the DO concentration is zero at all  points in the water
volume, the water temperature is 20°C and the barometric pressure is
1.00 atm.

     4.2.5  Aeration Efficiency (AE).  OTR per unit total  power input.
Power input may be based either on delivered power or wire power.

     4.2.6  Standardized Aeration Efficiency (SAE).  SOTR per unit
standard power input; may be based either on Total Delivered Standard
Power or Wire Standard Power.

     4.2.7  Oxygen Transfer Efficiency (OTE).  Fraction of oxygen
in an injected gas stream dissolved under given conditions of temper-
ature, barometric pressure, gas rate, and DO concentration.

     4.2.8  Oxygen Transfer Efficiency at Zero DO (OTE ).  OTE when the
DO concentration is equal to zero at all points in the water volume.

     4.2.9  Standardized Oxygen Transfer Efficiency (SOTE).  OTE  when
the water temperature is 20 C and the barometric pressure is 1.00 atm.


5.  Apparatus and Methods


5.1  Tank.  A tank or vessel is required to contain the water to which
oxygen is transferred.  The geometry and size will depend on the par-
ticular oxygenation system to be tested.  Tank volumes may range from
small laboratory vessels of a few liters to large tanks of over 1 million
gallons.

5.2  Mater.  For determination of a Standardized Oxygen Transfer Rate,
the water to which oxygen is transferred should be equivalent in quality
to a potable public water supply.  Further specifications of clean water
are given in Section 6.5.  This method may also be applied to measure
an oxygen transfer rate in wastewater when the rate of oxygen demand is
negligable.

5.3  Oxygenation Device.  This method is applicable to a wide variety
of oxygenation devices installed in the tank including but not limited
to the following:
          Surface Aerators:  high speed, low speed and horizontal
                             shaft rotors.
          Subsurface Oxygenation Devices:  diffused air, static
                             tubes, submerged turbines and jet aerators.

The method can be applied to systems based on air oxygenation as well as
to those based on oxygenation by high purity oxygen.
                                     299

-------
5-4  Sampling Devices.  Submersible pumps and tubing are necessary when
DO concentrations are to be measured on pumped samples in accordance
with Section 6.10.2.

5.5  Dissolved Oxygen Measurement

     5.5.1  Wet chemical measurement of DO on pumped samples shall be in
accordance with the Azide Modification of the Winkler Method described in
Section 421B of the 15th Edition of Standard Methods (5) except when a
blank correction is required as described in Section 6.11.3 in which case
the corresponding procedure given in the llth Edition of Standard Methods
(6) should be used.

     5.5.2  Membrane Electrode Measurement of DO either on pumped samples
or lailty. sha11 be 1n agreement with Section 421F of Standard Methods (5)
 5.6  Temperature Measurement.  Water temperature measurement shall be in
 accordance with Section 212 of Standard Methods (5).

 5.7  Deoxygenation Chemicals

     5.7.1  Sodium Sulfite.  Technical Grade sodium sulfite (^503) shall
 be used for deoxygenation in accordance with Section 6.8.  It is preferable
 that the sodium sulfite be free of cobalt.  However, chemical containing
 a known concentration of cobalt may be employed provided that this cobalt is
 considered as part of the total cobalt addition discussed in Section 6.9.1.

     5.7.2  Cobalt Catalyst. Either reagent grade or technical grade cobalt
 chloride, CoClg. or cobalt sulfate, CoS04, shall be used to catalyze the
 deoxygenation reaction in accordance with Section 6.8.

 5.8  Electronic Computer.  A digital computer or microcomputer capable of
 handling one of either the Fortran or Basic languages is required for the
 recommended non linear regression method of parameter estimation described
 in Section 7.2.1.

 5.9  Gas Flow Measurement Apparatus.  For oxygenation systems based on sub-
 surface gas injection, an apparatus capable of measuring the gas flow with an
 accuracy of ± 5% in accordance with Section 6.3 is  required.

5.10  Power Measurement Apparatus.  The apparatus required for power
measurements will  depend on the specific oxygenation device but, in general,
apparatus suitable for measurement of Total Delivered Power and Total Wire
Power in accordance with Section 6.4 is required.
                                     300

-------
6.  Procedure
6.1  Advance Preparation and Responsibilities.  When this method is to be
applied, the engineer-owner-manufacturer (EOM) representatives shall agree
in advance on the specific system to be tested and the test conditions.
Agreement must be reached on

          test location:  field installation or shop tank
          tank size and geometry
          aerator placement
          aerator power and gas rates

Where field testing is to be conducted, the engineer-owner representative
should provide the manufacturer with detailed drawings and specifications
of the tank or tank section in which the test will be conducted.   Informa-
tion on the water supply source and available water chemistry data should
be provided.  Water samples should be made available to the manufacturer
for laboratory experiments regarding the chemical  additions that will  be
made.

     Once the installation of aeration equipment is completed, provision
should be made for EOM representatives to inspect the installation to  verify
placement and testing conditions.  Systems employing diffused air aeration
should be tested to eliminate leaks.  Provisions for power and air flow mea-
surement should be verified and modifications made as needed.  It may  be
necessary to install equipment such as meters for power measurement, supple-
mental air piping, orifice plates and manometers.

     Upon completion of the installation of the aeration equipment, the test
tank should be cleaned prior to filling for testing.  Once the tank is filled
with the test water, chemical and biological contamination should be avoided.
It may be necessary to dewater and refill  the test tank during the testing
and adequate pumping and discharge arrangements should be made.

6.2  Test Tank Geometry and Aerator Placement.  It is difficult to describe
a required geometry or placement for testing conducted in tanks other  than the
full-scale field facility.  Appropriate configurations for shop tests  should
as closely simulate the field conditions as possible.  Water depths should be
similar if not identical and interferences due to wall effects and any extra-
neous piping or other materials in the tank should be minimized.   The  density
of aerator placement, air flow per unit volume or area and power input per
unit volume are examples of parameters that can be used to assist in making
comparative evaluations.
                                     301

-------
     Testing of tank sections is also useful in certain situations (e.g.,
long, narrow diffused aeration tanks) where there is little water cir-
culation between adjacent sections.  In this approach, a tank is divided
into sections and each section is tested individually.  When this testing
is performed, partitions shall be installed between adjacent sections to
prevent interchange of oxygen by advection and diffusion.   This technique
can provide information on spatial variation of K[_a and SOTR in tanks de-
signed for tapered aeration.

     Consideration should be given to utilization of shop  testing or
testing of tank sections when full-scale facilities are very large (e.g.,
in excess of 1 million gallons).  Other criteria to be considered in making
this judgement are:

          Ease of distribution of deoxygenation chemicals; distribution
          may be difficult in certain tanks.

          Sampling requirements; very large tanks may be difficult to
          sample adequately.

          Bulk flow and mixing patterns; shop or section testing should
          not be done when the mixing pattern in the shop  or section tank
          would not be representative of the full scale unit.
                                      302

-------
6.3  Gas Flow Rate Measurement.  For aeration systems based on subsurface
gas injection, the gas flow rate shall be accurately and precisely deter-
mined.  Recommended references for procedural assistance are Spink, L. K.,
Principles and Practices of Flow Meter Engineering  (8), Cusick, C. F.,
Flow Meter Engineering Handbook"  (.9), and A.S.M.E^ Power Test Codes
TO

     Full-scale plant gas flow measurement systems should be used with
caution when conducting field tests in a portion of the plant.  The pre-
cision and accuracy of the measurement device may not be adequate for
test flow rates.  Estimating gas flow rates to part of the plant by volu-
metric or area served ratios could be used as a check but not for the
primary air flow test information.

     Primary gas flow measurements shall be made using calibrated orifice
plate meters, venturi tube meters, pitot traversing methods, e.g., the
Annubar method, or other meters with an accuracy of ± 5 percent.  It is
desirable to provide a back-up or supplemental measurement system as a
check.  The instantaneous gas flow rate should not vary by more than
± 5 percent during the test.

     The gas flow measurement system shall be installed to avoid any
potential pulsation effect from positive displacement blowers.  Pulsa-
tion can cause significant error in head meter readings and is evidenced
by meter oscillations, inability to zero the meter under a no flow con-
dition, or meter readings being affected by connecting tubing length.
According to Spink, it can be mitigated by additional  volume and head
loss between the blower and meter.

     The following data shall be obtained to relate the measured gas
flow to standard conditions and to calculate the oxygen mass rate:

          Measurements in flowing gas feed line:

               Pressure, Pf
               Temperature, Tf
               Gas composition for gases other than air

          Ambient Measurements:

               Pressure, Pa
               Temperature, Ta
               Relative humidity, Rn

Ambient pressure and relative humidity may be obtained from Weather Service
data.

     Standard conditions for air flow are defined as 20°C, 1.00 atm, and
36 percent relative humidity.  For gases other than air, the standard
condition is defined as 20 C, 1.00 atm and zero humidity.  For air, the
volumetric flow rate at standard conditions is related to the actual volu-
metric flow rate in the feed line by,

                                   303

-------
          Qs  =  36.2
     where:   Q^    =  air fl(JW rgte  at  standard  con(jitions, ft3/min
              Qt   =  actual  air flow rate  in  the  feed  line, ft3/min
              Pf   =  total  pressure  in  the air feed line, psia
              Tf   =  air temperature in  the feed  line, °R
              Rh   =  relative  humidity of ambient air, dimensionless
                      fraction  at T,
                                  d
              Pya   =  vapor  pressure  of water at ambient temperature, Tg, psia
              Pa   =  ambient pressure, psia
Taking air to be 23 percent oxygen by weight, the mass flow rate of oxygen
is given by,
          WQ2   =   1.034  Qs                                        3
          where:    w     =  mass  f|ow Of oxygen in air stream, Ib/hr
                    Q     =   volumetric air flow rate, scfm
6.4  Power Measurements.   Power expended during  an unsteady-state oxygen
transfer test should be determined.   Various  power terms  are defined in Sec-
tion 4 and it is important that the  specific  components of power to be mea-
sured are clearly specified.
     The power components to be measured depend  on the specific oxygenation
device but, in general, the total delivered power (power  delivered to the
oxygen transfer device) and total wire power  (electrical  power drawn by motors
driving the device) should be measured.
     Measurement of power and related quantities  should be in accordance
with the following:
          Shaft power measurement -  ASME Power Test Code  PTC 19.7 (10)
          Electrical power measurement - ASME Power Test  Code PTC 19.6 (10)
          Pressure  measurement  - ASME Power Test  Code  PTC 19.2 (10)
          Rotary speed measurement -  ASME Power  Test Code PTC 19.3 (10)
                                     304

-------
     6.4.1  Delivered Power

          6.4.1.1  Total Delivered Power is given as the sum of all the de-
livered power terms.

Total           Delivered        Delivered        Pump
Delivered   =   Gas         +    Shaft       +    Delivered       4
Power           Power            Power            Power

Specific oxygenation  devices may have only one or two delivered power terms,
e.g., a surface aerator has only shaft power whereas a submerged turbine has
gas plus shaft power.

          6.4.1.2  Delivered Gas Power shall be calculated based on adiabatic
compression by


                             WRT,
          Delivered Gas   =  -B—
             Power

     where the symbols and corresponding dimensions are given by

          w   =  weight flow rate of gas, f •  t
          R   =  gas constant, L T"1

          T,  =  absolute temperature before compression, T
                                                          -2
          P,  =  absolute pressure before compression,  f L
                                                          7
          P~  =  absolute pressure after compression, f L

          k   =  ratio of specific heats for gas,  Cp/Cv, dimensionless
          K   =
                 k-1 , dimensionless
                  k

          6.4.1.3  Delivered Shaft Power should be based on direct measure-
ment of torque and angular velocity delivered to the oxygenation device and
calculated from

          Delivered                      N
          Shaft          =    27T T        r
                                      , ,
          Horsepower                  3'3 x

     where the symbols and corresponding units are given by

          T   =  torque,  ft'lb,  measured using either a cradled dynamo-
           "     meter (cradled  motor, generator,  prony brake)  or surface
                 strain dynamometer (surface strain,  angular twist)
          N   =  rotational  speed,  revolutions per minute
                                    305

-------
When accurate conversion efficiencies are known, Delivered Shaft Power may
be calculated by multiplying the motor wire or motor shaft horsepower by
the appropriate efficiency.

          6.4.1.4  Pump Delivered Power shall be calculated as
          Pump Delivered
          Horsepower
            QW  (T.D.H.HS.G.)
                  3960
     where the symbols and corresponding units  are given by

          Q       =  liquid flow rate,  gal/min
          T.D.H.  =  total pump head,  ft
          S.G.    =  specific gravity  of water  at  test  temperature

     6.4.2  Delivered Standard Power.   For purposes  of  comparison,  it  is
convenient to base delivered power on  standard  conditions.

          6.4.2.1  Delivered Standard  Air Power is the  gas  power  consumed
by the oxygen transfer device operating on water at  68°F (20°C) when fed
standard air at a mass rate of oxygen  equivalent to  the actual air  feed rate
with a blower intake pressure loss of  0.10 psi  and a pressure loss  in  the
blower discharge line of 1.00 psi.  This shall  be  calculated by
          Delivered Standard
          Air Horsepower

     where:
         =  0.227 Q.
 2s
• i • •

'is
                                                   0.283
-1
          Q   =  air rate expressed  as  scfm
and
P2s


PP and P,
£o l j
= standard
atmos-
pheric
pressure
(14.70
psia)
                refer to  standard  discharge and  inlet pressures such  that
hydro-
static
pressure
at the
diffuser,
PL
                               +  diffuser  +  standard =  15.70 psia + P  +  P.
                                  pressure     pipe
                                  loss         pressure
                                  P,           loss
                                   a           (1.0 psia)
                                  306

-------
P,     =    standard         -    standard
            atmos-                suction
            pheric                pressure
            pressure              loss
            (14.70 psia)           (0.1 psia)

       =  14.60 psia

          6.4.2.2  Delivered Standard Shaft Power is the shaft power required
to turn the turbine in water at 20°C at an angular velocity equivalent  to  the
actual angular velocity.   This should be estimated from Equation 6  based upon
the variation of shaft torque with water temperature at constant angular ve-
locity.

          6.4.2.3  Pump Delivered Standard Power is the pump delivered  power
given by Equation 7 when  the specific gravity is 1.00.

          6.4.2.4  Total  Delivered Standard Power is given  by
Total
Delivered
Standard
Horsepower
Delivered
Standard
Air H.P.
Delivered
Standard
Shaft H.P.
Pump
Delivered
Standard
H.P.
     6.4.3  Wire Power.   Wire power is  related to  delivered  power  by


          Wire Power   =   Delivered Power                        1Q
                           ea eb ec

     where:

          e ,  e,, e   are the efficiencies of  the  various  power  conversion
                      devices (motors,  gear drives,  reducers,  blowers)  used
                      to transform electricity to  delivered  power.

Wire power should be measured using accurate recording  polyphase watt meters
capable of monitoring 10 cycle per second  peaks.   An ammeter can also be
used if the voltage and power factor are measured.   An  expression  for calcu-
lating the 3-phase power from current,  voltage and power factor  measurements
is

          Wi re                        o
          Horsepower   =   2.319 x 10   E I  F                   11

     where:

          E  =  voltage, volts

          I  =  current, amperes
          F  =  power factor


                                    307

-------
     6.4.4  Hire Standard Power.  Wire Standard Power is useful when effi-
ciency comparisons are to be made on the basis of wire power.  In certain
test situations, blower  or motor  efficiency may not be typical of the
efficiency attained in an actual system.  Consequently, actual wire power
might be representative of a typical installation.  Wire Standard Power snail
be calculated by applying the recommended standard efficiencies given in
Table 1 to the appropriate Delivered Standard Power Term.

6.5  Water Quality

     6.5.1  General and Total Dissolved Solids.  The water supplied for the
initial test shall be equivalent in quality to a potable public water supply.
Subsequent tests may be conducted in the same water and, because of the  ,
addition of deoxygenation chemicals, the Total Dissolved Solids concentration
(TDS) will increase.  Repetitive testing may be conducted in the same water
provided that the TDS does not exceed 1500 mg/1.  When the water supply contains
TDS greater than 500 mg/1, the 1500 mg/1 limit may be raised to 2000 mg/1 with
the approval of EOM representatives.

     6.5.2  Temperature.  Water temperature should be between 10°C and 30°C
and as close to 20°C as possible.  Testing outside this temperature range may
be necessary in some field situations and can be done with the approval of
EOM representatives.  Low temperatures slow the deoxygenation reaction and
this may introduce some error.  It is recommended that a standard e value of
1.024 be employed to adjust for temperature.  Appreciable error can be intro-
duced when the actual e value differs from this and the temperature difference
is more than 5 degrees Celsius.   Water temperature shall not change by more
than 2 degrees Celsius during a  single unsteady state test.

     6.5.3  Water Quality Analyses.   Initial Analyses:  Prior to beginning
the testing program, a representative sample of the water to be used in the
test tank shall be analyzed for:  TDS, alkalinity, sulfate,  iron, manganese,
residual chlorine, pH, total  organic carbon or chemical oxygen demand, cobalt,
surfactant (MBAS), and temperature.   EOM representatives should review these
data to assess possible effects  of water quality on the test results.  Based
on these data, EOM representatives should establish:  1) upper limits on the
allowable TDS or sulfate concentrations; 2) the concentration of cobalt to be
employed; 3)  possible modification to the dissolved oxygen measurement proce-
dure e.g. use of a blank in Winkler analyses; and 4) possible modifications of
the test water quality.

     Normally, no modifications  of test water quality are made for clean water
testing.  However, it may be that, due to a particular set of field conditions,
a surface water may have to be treated.  Demonstrated oxygen transfer  rate
effects due to water chemistry differences between the field conditions and
prior test data used to predict  the field performance may be considered in
agreeing upon any water quality modifications.
                                    308

-------
   Table 1.  Recommended Equipment Efficiencies to be
      Used for Standard Power Determinations

Item
Compressor
Gear box
Coupling or drive
Motor
*Efficiencies quoted
Efficiency Range Recommended Standard
Efficiency
(decimal) (decimal)
0.50 -
0.94 -
0.90 -
0.90 -
are for full
0.80
0.96
1.00
0.95
load conditions.
0.70
0.95
0.95
0.92


  Table 2.  Hater Quality Measurements During Testing	
Measurement                   Frequency of measurement
Temperature*                  every test - beginning and end
Total Dissolved Solids        every test or calculated by mass balance
Filterable Cobalt             every test, end
pH*                           every test
DO Demand*+                   every test plus early in test
                              when DO <  50 percent saturation

*measured at test site during testing.
+optional, to be performed when unreacted sulfite is suspected.
                               309

-------
     Water Quality Analyses During Testing:  When repetitive tests are
conducted in the same water, the water should be analyzed for the Cari
constituents and properties at the frequencies indicated in Table 2.  me
DO demand measurement is made by filling a BOD bottle with sample ana mea-
suring the oxygen depletion at 1 and 5 minutes.  This measurement is in-
tended to indicate residual unreacted sulfite.  A significant DO demand oy
unreacted sulfite seriously reduces the apparent transfer rate and will-in-
validate the results.

     The cobalt concentration normally should not change during testing but
slight decreases have been observed.  If the concentration of cobalt tails
below 0.1 mg/1, additional cobalt should be added to assure complete oxida-
tion of the sulfite.

     When detergent-added testing is conducted, analyses for anionic sur-
factant and surface tension should be made at intervals during each run.  It
has been reported that the concentration of surfactant decreases during a
test.  Allowing for a 20 minute delay after addition of detergent and prior
to deoxygenation may assist in stabilizing the surfactant concentration.

6.6  System Stability.  The aeration system should be operated to achieve
steady state hydraulic conditions prior to starting the oxygen transfer eval-
uation.  The hydraulic mixing regime should be established in the test tank
for each test condition prior to deoxygenation.  A steady power draw can
indicate a steady state hydraulic condition for mechanical  aerators.  Some
mechanical and diffused air systems require 30 to 40 minutes to achieve a
steady state hydraulic regime.

     For diffused air systems,  water shall be displaced from the aeration
system prior to beginning the test.  Steady manometer readings for orifice
air flow measurement and consistent air flow rate measurements for other
flow measurement devices are indicative of this displacement.  Lines with
values for purging water from the aeration system may be added for testing
purposes.

     For tests of fixed platform surface aerators and rotors, the water sur-
face elevation shall  be held constant so that the power draw is'constant
during the test.   For other systems, the volume of water under aeration shall
not vary by more than ± 2 percent during any one test.
6.7   Initial  Run.   Data anomalies  have frequently  been  reported  during  the
initial run conducted after filling  the test  tank.   Because  of this,  this
run is frequently not used as part of the oxygen transfer evaluation.   Instead,
it is often used to stabilize water  chemistry and  to overcome possible  problems
in test procedures  including sulfite dispersion and  sampling techniques.   It
also can provide an opportunity to check for  possible analytical  interferences,
to assure proper probe calibrations  and to verify  adequate cobalt residual
based on deoxygenation and reaeration patterns.  The initial run  may  be used
as part of the oxygen transfer evaluation with the approval  of EOM representa-
tives.
                                  310

-------
6.8  Deoxygenation Chemicals.  Technical grade, sodium sulfite (Na?SO-)
shall be used for deoxygenation.  The sulfite shall be essentially cobalt
free and contain no impurities that would alter the oxygen transfer rate
analysis.  Sodiam sulfite should be added in solution or slurry form.  This
is accomplished by dissolving the sulfite in a separate mixing tank prior
to its addition to the test tank.  Saturated solutions contain 2.23 Ib/gal
at 20°C and 3.00 Ib/gal at 30°C.

     Technical grade sulfite from a given batch should be evaluated for
impurities and potential interference in oxygen uptake evaluations.  This
should be done by chemical analysis for cobalt, iron and manganese" and by
comparative laboratory oxygen transfer tests in which the results using the
supplied technical grade sulfite for deoxygenation are compared against those
obtained using analytical grade sulfite or nitrogen stripping deoxygenation
methods.

     The sulfite deoxygenation reaction is  catalyzed by cobalt.   The  cobalt
source utilized should be cobalt chloride,  Cod?,  or cobalt  sulfate,  60804,
reagent or technical  grade.   The cobalt should be  dissolved  prior to  its
addition to the test  tank.   Solubilities of CoCl2  and CoS04  are  approximately
50 g and 36 g per 100 ml  of  water respectively at  20°C and increase with
temperature.   Care should be taken to  ensure  that  complete dissolution  of
the cobalt salt is obtained, especially when  it is added in  the  sulfate form.

6.9  Addition of Deoxygenation Chemicals

     6.9.1  Cobalt Addition.  A solution of cobalt salt shall be added to the
test tank to achieve a soluble cobalt concentration between  0.10 mg/1 and
0.50 mg/1 in the test water.  Cobalt concentrations greater  than 0.50 mg/1,
but less than 2.0 mg/1 are permissible when laboratory aeration experiments
demonstrate that greater concentrations are necessary to catalyze the deoxy-
genation reaction in a particular test water or when the water temperature is
less than 10°C.  If a cobalt concentration greater than 0.5  mg/1 is employed,
analytical precautions should be taken to eliminate potential interferences
in the Winkler titration.  Soluble cobalt measurements should be made at the
conclusion of each test and these results may indicate that  supplemental addi-
tions of cobalt are required.

     The cobalt solution shall be added prior to the beginning of oxygen
transfer testing with the aeration system operating.  The solution shall be
uniformly distributed into the test tank.  Pumps and distribution systems
may be required for large tanks.  The cobalt solution shall  be dispersed
throughout the tank by operating the aeration system for approximately 30
minutes.  The cobalt catalyst should normally be added once  for each test
water.
                                   311

-------
      6.9.2   Sulfite Addition.  The theoretical sodium sulfite requirement
 for  deoxygenation  is  7.88 mg/1 per 1.0 mg/1 DO concentration.  Sulfite
 additions are made in excess of stochiometric amounts.  The amount of excess
 is dependent on  the oxygen transfer rate of the aeration system and tne size
 of the  test  tank.  The amount of excess varies from 20 to 150 percent and
 increases for high transfer rate systems.

      Sufficient  sulfite solution shall be added to depress the DO level
 below 0.50 mg/1  at all points in the test water.  It should be noted that
 consistent repetitive testing results have been observed where the DO con-
 centration has reached zero at all sample points and remained at zero at
 least two minutes.  Results from the initial test run can be used to help
 establish the proper  quantity to be added.

      Sodium  sulfite shall be dissolved or slurried in mixing tanks outside
 the  test tank and distributed uniformly and rapidly (in less than 3 to 4 min)
 into the test tank.   The use of pumps and flexible piping to distribute the
 solution across  the tank surface is recommended.  Testing conditions may pre-
 vent achieving a true sulfite solution prior to the addition.  Addition in
 a slurry form is preferred over the direct addition in a crystal form.  Extreme
 care should  be exercised to assure adequate dispersion and dissolution in the
 test tank.

 6.10 Determination of Dissolved Oxygen at Various Points in the Tank During
 the  Unsteady State Test.  Dissolved oxygen concentration shall be determined
 at various points in  the tank and at various times during the unsteady state
 test.   This  determination shall  be carried out by one or both of the following
 procedures:

          Samples pumped to BOD bottles followed by measurement
          of dissolved oxygen by either the Winkler method or
          membrane probe.

          In situ measurement of dissolved oxygen in the tank by
          membrane probes.

     6.10.1   Location  of Dissolved Oxygen Determination Points.  The number
and  location of dissolved  oxygen determination points will be dictated by
the  size of the test  tank,  aerator placement and mixing pattern in the tank.
The  following criteria shall  be  used in establishing the number and location
of points.

          A  minimum of four determination points shall be used.  One
          should  be at a shallow depth; one should be at a deep loca-
          tion  and  one should  be a mid-depth.   The points should be
          at  least  2  feet  from the walls, floor and surface and no
          closer  to the  surface  than 10 percent of the minimum tank
          dimension.

          The determination points shall be located so that each
          senses  an equal  portion of the tank volume and shall  be
          distributed vertically and horizontally to best represent
          the tank contents.

                                    312

-------
          More than four determination points should be used when
          the tank is large (e.g.  > 500,000 gal),  when significant
          dissolved oxygen gradients exist (e.g. more  than 1.0 mg/1
          DO difference between points),  or when a point  source
          oxygenation device is installed in a  large tank.

For large tanks with point sources of oxygen and non-uniform dissolved
oxygen concentrations, as many as  8 determination  points  may be required.

     Occasionally, a DO probe will malfunction  during  a test and,  for this
reason, it is advisable to use backup probes near  selected determination
points so that the test data may be valid even  if  one  probe malfunctions.
Backup probes should be clearly identified as such, and the determination
points to which each backup probe  applies should be specified.


     6.10.2  Collection of Pumped  Samples.  When  pumped samples are to be
obtained for dissolved oxygen analysis, submersible pumps.(discharging
through tubing to BOD bottles) should be installed at  the determination
points.  The pump inlet should be  designed and  located to avoid air bubble
extrainment.  The pumping rate and tubing lengths  should be adjusted so that
the transport times between the pump and bottle are equal for all  determina-
tion points and not greater than 15 seconds, and  the 300 ml BOD bottles fill
in 6 to 10 seconds.  Sudden pressure drops in the  sample line will degas the
water and thus should be avoided.

     With this system, the DO measurement may be made  by a DO probe mounted
in the sample line, by a DO probe  with a stirrer placed into the sample
bottle or by Winkler titration of  the DO in the sample bottle.  The in-line
probe system must be observed continuously to prevent  line clogging or damage
to the membrane.  If individual bottle samples  are used,  they must be care-
fully stored to prevent temperature change and  degassing.  They should be
analyzed for DO as soon as possible.

     If the pumped sample-individual bottle technique  is  used, it is recom-
mended that at least one recording DO probe system also be installed.  This
system would be used to help set sampling times,  indicate when zero DO was
reached during deoxygenation and assist in determining the end of run by
noting a stable maximum DO concentration.

     6.10.3  Times of Dissolved Oxygen Determination.   A minimum of 10 to  15
dissolved oxygen values shall be measured at prescribed times at each deter-
mination point during the unsteady state test.   Measurement of up to 20 values
is desirable but more than 20 values will not usually  result in a significant
improvement in precision.  Dissolved oxygen determination shall be made at
times so that two-thirds (63 to 70 percent) of  the values are evenly distri-
buted over the period of time between the initial  data point which is above
zero and 2/K.a (0 to 86 percent saturation) and one-third (30 to 37 percent)
of the values evenly distributed over the period of time from 2/K. a to 4/K, a
(86 percent to 98 percent saturation).  In cases of rapid transfer, the
minimum interval between measurements should be 0.4 minutes for pumped samples.
No minimum interval is specified for jjn situ probes.


                                    313

-------
     It is convenient to plan a test based on either 12,  15>.18.°r ^
measurements.  Table 3 shows the determination intervals  estimated to
the required distribution of data points for these cases.   To use this
table for calculation of intervals,  one must have an a. priori  estimate or
either KLa or OTR.  Furthermore, the table assumes that sampling will  jegin
at time zero which is assumed to occur at zero dissolved  oxygen  concentration.
If zero DO is not attained or if early truncation is to be  practicea,  tne
determination interval for zero to 86 percent saturation  should  be decreased
by roughly  25 percent so that the required number of points  are oDtainea in
this region.

     6.10.4  Run Duration and DO Saturation.   DO  data should  be  obtained over
as wide a range as possible.  Truncation of data  at DO levels  less than dQ
percent of C* is allowable to avoid  lingering effects of  the  deoxygenajion
technique.  Tn no case shall values  of DO greater than 30 percent of Cro be
truncated.

     All test runs shall be continued for a period of time  approximately
equal to 4 divided by the anticipated value of KLa.   This is  equivalent to
continuing the run until the DO concentration is  98 percent of the saturation
concentration, C*, which is normally approached at a time equal  to 6/KLa.  For
purposes of comparison with the value of C^ estimated by  the  model, it is
useful that at least one run be continued a time  at 6/KLa for each testing
condition (temperature, aerator operation and geometry) to  obtain a test
saturation DO concentration.

     Measured values and tabulated values of  dissolved oxygen  surface satura-
tion concentrations shall  be used for comparative information  only and shall
not be used as model parameters for  calculation of oxygen transfer rates.

6.11   Dissolved Oxygen Measurements

      6.11.1   Measurement by In Situ and Sample Line  DO Probes.   The in  situ
DO probes shall  be  fast response probes with 1.0-mil membranes  and  should  be
equipped with agitators.   If agitators are not employed, care shall be  taken
to ensure that there  is sufficient water velocity past the probe.   The  probes
shall be calibrated using  the Winkler procedure with test  tank  water  and checked
for  linearity against Winkler procedure titrated samples.  The  calibration and
linearity shall  be established prior to the  initial  run and shall  be  verified
every 3 to 4  hours during  subsequent runs.   It is recommended that a  calibra-
tion and linearity check be performed on at  least one probe during  each run.
Probe calibration and linearity check can be conveniently  accomplished  by
comparing probe  readings with Winkler measurements  on discrete  samples  taken
at the probe  locations.  Each probe should be equipped with a strip chart or
digital recorder which will permit reading of the dissolved oxygen concentra-
tion at 0.5 minute intervals with a precision of ±  0.05 mg/1.   However, in
the absence of recording devices, it is permissible  to manually record  probe
data at the appropriate determination intervals.
                                    314

-------
  Table 3.   Estimated Dissolved Oxygen Determination Intervals

Total
Number of
Data Points
12
15
18
21
Time from
(0 to 86%
Number of
Points
8
10
12
14
0 to 2/KLa
Saturation)
Determination
Interval , min.
17.1/KLa
13.3/KLa
10.9/KLa
9.2/K,_a
Time from
(86% to
Number of
Points
4
5
6
7
2/KLa to 4/KLa
98% Saturation)


Determination
Interval , min
30/K,_a
24/KLa
19.8/KLa
17.1/KLa




    K.a = volumetric transfer coefficient,  hr"   and may be  approximated  by
    KL
a =
     10W
where:  OTR  = expected oxygen transfer rate at zero dissolved
               oxygen, Ib/hr
           U = weight of water, 106 Ib
                                     315

-------
      If the probes  are placed  in  sample  lines to measure DO, they  need  not
 be equipped with agitators.  The  velocity of flow past the probe shall  pe
 established and maintained  to  provide  an accurate response from the probe.
 Calibration, linearity check,  and recording requirements for sample line
 probes are equivalent to those for jn  situ probes.

      Dissolved oxygen probes require considerable care and attention  to
 provide continuous  reliable data.  Back-up probes and meters should be
 available for replacement as needed.

      6.11.2  Measurement by Probes in  BOD Bottles.  Probes used to measure
 the DO in 300-ml sample bottles shall  be equipped with a stirrer.  Probe
 manufacturer recommendations shall be  followed for calibration and linearity
 checks.

      6.11.3  Measurement by the Hinkler Procedure.  The modified Winkler DO
 analytical procedure  of the 15th  Edition of Standard Methods (Section 421  B)
 should be followed  for DO titrations.  Procedures using excess iodide to
 prevent the loss of iodine vapor  at high DO concentrations should be consi-
 dered if unexplained  lower readings in the higher DO ranges are experienced.

      Chemical  interferences in Winkler DO titrations have been observed with
 some testing,  especially when  higher cobalt levels are used (in the 2.0 mg/1
 range).  The analytical  procedure  presented in the llth edition of Standard
 Methods (1960)  should be used  to  correct for these interferences when cobalt
 concentrations  greater than 1 mg/1 are present.  A sample of the test water
 should be taken at  the end of  the  test run and titrated directly without the
 addition of manganous sulfate  solution, but with the addition of the alkali-
 iodide-azide reagent.   This value, called a "blank", should be subtracted
 from every DO measurement in the  test  tank to correct for chemical interfer-
 ences.

6.12  Testing with Detergent Addition.   The  unsteady state  clean  water oxygen
transfer test is occasionally performed with  clean water .containing added
detergent to roughly simulate the  performance  of  the oxygen transfer device
in a dirty water and to mask the effect of trace  contaminants  in  tap water.
The transfer rates obtained  in  the presence  of detergents  are  not necessarily
equivalent to those  in municipal wastewater  or mixed liquor.   However, the
transfer rates in the  presence  of  detergents  are  generally  closer to those
obtained in actual process water than are the  usual  clean  water transfer rates.

     A  strong solution of detergent, linear alkylate sulfonate (LAS)  or
househould  detergents, shall be prepared using hot water (80-90°C).   The
detergent solution shall be added  in an amount to achieve an average  concen-
tration of  approximately 5 mg/1 (measured as methylene-blue active substance-
MBAS) during a test run.  An initial concentration of 7 mg/1 of detergent  in
the test water has been  used.   The detergent solution shall be added  prior to
the reaeration period and shall be uniformly dispersed throughout  the tank
contents.
                                     316

-------
     The detergent concentration (MBAS) in the test water shall be analyzed
at the beginning and end of the run and at several intermediate points to
determine an average concentration for the run.  Static surface tension
measurements at these times are also recommended.

     Because cobalt may be complexed by detergents, the cobalt concentration
should be checked to verify that adequate catalyst is present.  It is
suggested that a cobalt addition equivalent to 0.10 mg/1 in the test water
be made for each run.   Provision should be made to dispose of any foam that
develops during the test.   Some difficulty with detecting end points with
the Winkler DO titration procedure has been noted.   When detergents are
present it is recommended that probes be used to measure the DO concentration.
The remaining test procedures are as noted in earlier sections.


7.  Data Analysis


7.1  Preparation of Data for Analysis.  The purposes of data preparation are:

          to obtain discrete numerical values from continuous probe
          readi ngs,

          to plot and examine the data to determine if low-end trun-
          cation is necessary,

          to truncate low dissolved oxygen values, and

          to assure that the data values to be analyzed meet the
          timing criteria outlined in Section 6.10.3.

     7.1.1  Continuous Probe Data.  When DO probes equipped with continuous
strip chart recorders are used, a continuous record of dissolved oxygen versus
time will be available at each determination point.  In this case, approxi-
mately 21 discrete data values shall be selected from the recorder output for
each determination point at the approximate time intervals given by Table 3.
If the probes are equipped with digital print out, the discrete values closest
to the interval divisions shall.be selected.  The initial data value shall be
at the beginning of record and shall be greater than zero, but less than 0.5
mg/1 dissolved oxygen.  The discrete data values shall be read directly from
the recorder graph or digital output and shall not be subject to any smoothing
or averaging procedure.  If low-end truncation is anticipated or if the record
does not begin at zero DO, the determination interval for 0 to 86 percent
saturation (0 to 2/K|_a) should be decreased from the value of 9.2/K|_a indicated
in Table 3 to about 7/K|_a so that 14 or more points are selected in the zero
to 86 percent range.  Approximately seven discrete data values shall be se-
lected in the range of 86 to 98 percent saturation (2/K|_a to 4/^3) at inter-
vals of approximately 17.1/KLa.

     Use of Table 3 requires an approximate value of Kj_a and one way of
obtainina this is based on the expected value of OTR0 as indicated in the
Table.  However, an approximate value of KLa can more easily be established

                                   317

-------
based on inspection of the DO versus time plot by noting the approximate
value of the saturation concentration, C*, approached at infinite time.  An
approximate value of K|_a is then given as the reciprocal of the time interval
between dissolved oxygen concentrations of zero and 63 percent of C^.  (For
this method to be applied, the data may have to be extrapolated to zero DO).

     The 21 or more data values selected at each determination point should
be plotted as DO versus time, and the low DO data should be examined for
lingering effects of the deoxygenation chemicals.  These effects would be
manifested by the rate of increase in DO (dC/dt) being constant or increasing
with DO (C) up to an inflection point.  If this pattern is suggested, the
DO data below 20% of the approximate C* value, low DO data may be truncated
up to a DO concentration of 1.5 times Ihe DO concentration at the inflection
point.  However, in no case shall DO values greater than 30 percent of C^ be
truncated.  DO values in the 86 to 98 percent saturation range shall not be
truncated except as required in the application of the Alternate Best Fit Log
Deficit parameter estimation method (Section 7.2.2).

     Following possible low-end truncation, the data plotted for each deter-
mination point should be checked again for compliance with the timing criteria
outlined in Section 6.10.3.  In the case of data obtained from recording
probes, 63 percent to 70 percent of the DO values (at least 13) shall be
evenly distributed over the period of time between the initial data point and
2/KLa (where the DO is 86 percent of C*) and 30 to 37 percent of the DO values
(at  least 6) shall be evenly distributed over the period of time from 2/KLa
to 4/K|_a (86 percent to 98 percent of C.J.


     Occasionally, a DO probe will malfunction and data from a malfunc-
tioning probe shall note be used to calculate the SOTR.  Frequently, a
malfunction will be obvious but, in other cases, a malfunction may be only
suspected.. The following criteria should be employed to judge whether
data from a suspected probe may be deleted.

     - Results of a probe calibration or check conducted immediately
       before, during, or immediately after the test.

     - Comparison of the K|_a2o value for the suspected test and point
       with the average of those obtained for the same point in at
       least three replicate tests (including the suspected test) as
       described in Section 8.2.2.  Variation by more than 15% of the
       average suggests probe malfunction.

Deletion of data from a malfunctioning probe may invalidate the test in
accordance with Section 8.2.2.

     7.1.2  Discrete Pumped Sample Data.  When discrete pumped samples are
analyzed either by the Winkler method or by a membrane probe, the data consist
of dissolved oxygen concentration values at the determination intervals used
for each sampling location.  In this case, it is crucial that the determination
intervals or sampling times have been well planned in accordance with Table  3.

     The data shall  be plotted as DO versus time and the low DO data should
be examined for lingering effects of the deoxygenation chemicals as  described

                                     318

-------
in Section 7.1.1 above.  These effects are normally more difficult to detect
in discrete sample data.  Based on this examination, data values up to 20
percent or 30 percent of C* may be truncated as outlined in Section 7.1.1.
DO values greater, than 30 percent of C* shall never be truncated.  DO values
in the 86 to 98 percent saturation range shall not be truncated except as
required in the application of the Alternate Best Fit Log Deficit Method
(Section 7.2.2).

     Following possible low-end truncation, the data plotted for each
sampling point should be checked for compliance with the timing criteria
outlined in Section 6.10.3.  In the case of data obtained -from discrete
samples, 63 to 70 percent of the DO values (at least 7) shall be evenly
distributed over the period of time between the initial data point and
2/KLa (where the DO is 86 percent of C*) and 30 to 37 percent of the DO
values (at least 3) shall be evenly distributed over the period of time
from 2/KLa to 4/KLa (86 percent to 98 percent of C*).  The discrete data
values to be analyzed shall not be subjected to any smoothing or averaging
procedure.

7.2  Parameter Estimation.  The purpose of the parameter estimation proce-
dure is to determine the best estimates of the three model parameters, K[_a,
C*, and CQ so that the model given by Equation 1 best describes the varia-
tion of DO with time at each determination point location in the tank.
Parameter estimates shall be performed for the data obtained at each deter-
mination point location by one of two methods, the recommended non-linear
regression method or the alternate best fit log deficit method.  The parameter
estimates given by each method will differ slightly because of the top end
truncation and logarithmic transformation required in the alternate method.
However, the values of the transfer rate (OTRg) calculated based on the
parameter estimates will generally agree within a few percent.  Other para-
meter estimation methods such as the ordinary log deficit or direct methods
can be useful in determining approximate values for K|_a and C^.  However, the
parameter estimates based on these other methods shall not be°°used for calcu-
lation of the transfer rates (OTR0 and SOTR).

     7.2.1  Recommended Non Linear Regression Method.  This method is based
on non linear regression of the model (Equation 1) through the DO versus
time data as prepared for analysis in Section 7.1.  The best estimates of
the parameters K|_a, C* and C0 are selected as the values which drive the model
equation through the prepared DO concentration versus time data points with
a minimum residual sum of squares.  That is, the parameter estimates are
selected so that the sum at the squares of the residuals is minimized.  Here,
a residual refers to the difference in concentration between a measured DO
value at a given time and the DO value predicted by the model at the same
time.

     Application of this method requires either a digital  computer or micro-
computer capable of handling one of either the Fortran or Basic Languages.
A user-oriented Fortran program which will  give the least square estimates,
and standard deviations of K^a, C* and Cg along with an output of the data,
fitted values and residuals is given in Appendix A.   A similar user-oriented
Basic program written for the Apple II Microcomputer is given in Appendix B.
To use either program,  the prepared data for a particular determination
(sampling)  point are entered along with initial approximations of the three

                                     319

-------
parameters.  The computer then adjusts these Initial approximations to mini-
mize the residual sum of squares and reports the final parameter estimates*

     The values of DO predicted by the model based on the final parameter
estimates  should be plotted along with the measured values and examined.  A
visual examination of such a plot will indicate any gross mistakes in the
non linear regression.

     7.2.2 Alternate Best Fit Log Deficit Method.  This method is acceptable
when lack  of a computing facility prevents application of the recommended
method.  The chief advantage of this approach is that it can be applied by
using an ordinary scientific calculator (a calculator such as the Texas Instru-
ments Model 55 with linear regression capability is recommended) instead of a
computer.  However, this method can be time-consuming and tedious.

     The method  is based on linear regression of a logarithmic form of the
model equation using a  logarithmic function of the DO data.  The logarithmic
form of the model equation is
           ln(C*-C)
ln(C*-C0) + KLa t0 -  KLa t
      where:
               =   time when C = C
                                o
12
 This  equation  shall be fit to the prepared data for each determination
 (sampling)  point  by performing a linear regression of ln(C*-C) versus time.
 The value of C* shall be estimated by trial and error.  The best estimate
 shall  be the value of C* giving the minimum residual sum of squares.  The
 best  estimate  of  K[_a shall be determined as the slope of the linear regression
 based on the best estimate of C^.  Because of the lack of precision in very
 small  values of (C^-C) and impossibility of taking the log of negative values
 of  (C^-C),  it  is  necessary to truncate (delete) values of C within 0.2 to 0.3
 mg/1  of C   when this parameter estimation method is applied.  To apply this
 method, tfie following steps shall be followed.

      1.  The prepared data plotted according to Section 7.1.1 or
         7.1.2 shall be examined and an initial trial value of C*
         determined by inspection of the plot.

      2.  Data  points within 0.2 to 0.3 mg/1 of the trial value of
         C* shall be truncated.  However, the highest DO value
         included in the analyses shall be at least 92 percent of
         the trial C*.  In the case of continuous probe data, at
         least 3  values shal^ be at DO values greater than 86 per-
         cent  of  the trial Cro.  For discrete sample data, at least
         2  values shall be at DO values greater than 86 percent of
         the trial C .
                    GO

 *Note:  Because of the way in which C0 is defined in the model, it is possible
 to have negative values for the best estimate of C0.

                                  320

-------
3.
    The quantity, InCC^-C) shall be computed for each remaining DO
    value and a linear°°least squares regression of this quantity
    versus time shall be made.  (See any standard statistics text
    or suitable electronic calculator instructions).  The regres-
    sion coefficient is the slope of the plot of the ln(C*-C) vs.
    time and is equal to^-Kia.  The value of Ki a corresponding to
    the trial value of C^ shall be recorded.  Also the value of
    the y intercept (point at which the regression line crosses the
    In^-C) axis at t = 0) shall be recorded.  The predicted values
    of 1n(C -C) shall be calculated for the time of each observation
           °°
    by
         ln(C*-C)
                      = y Intercept - K|_a t
                                                             13
                   predicted
    The residuals shall be calculated for each data point by
Residual   * ln(C*-C)
        n        °°
                                     - 1n(C*-C)
                              Measured
                                                             14
                                                predicted
    Finally, the resjdual sum of squares corresponding to the
    trial value of C  shall be calculated for all N data points
    as
                              \
                                                             !5
4.  Steps 2 and 3 shall be repeated for at least three more trial
    values of C* in an effort to located the value of C* which
    minimizes tn°e Residual Sum of Squares.  The minimunTresidual
    sum of squares shall be established by plotting the Residua. 1
    Sum of Squares as the ordinate versus the trial value of C^.
    A number of trial C* values shall be selected to establish"
    the value of C* which minimizes the residual sum of squares.
    Values of the residual sunrof squares cannot be compared when
    based on different numbers of data points.   Consequently, the
    truncation criterion (0.2 to 0.3 mg/1 ) shall be chosen so that
    the calculations for successive trial values of C* are based on
    the same number of data points.

5.  The best estimate of C* shall be the value which minimizes the
    residual sum of squares.   The corresponding value of K|_a shall
    be determined from the linear regression based on that value
    of C*.
                                  321

-------
8.  Interpretation and Reporting of Results
8.1  Standarized Oxygen Transfer Rate (SOTR).  By convention, the oxygen
transfer capacity of an oxygenation system is usually expressed as the
rate of oxygen transfer predicted by the model at zero dissolved oxygen
under  standard conditions of temperature and pressure, usually 1-pO atmo~
phere  and  2(PC.  This  is termed the Standardized Oxygen Transfer Rate
It  should  be noted that the SOTR is a hypothetical value based on zero
dissolved  oxygen in the oxygenation zone which is not usually desirable in
real oxygenation systems operating in process water.  The SOTR value shall
be  determined by correcting the values of KLa and C* estimated according
to  Section 7.2.1 or 7.2.2 for each determination point to standard condi-
tions  by
          KLa20
                          (20-T)
                                                 16
          "•20
=  C
                                                                  17
     where:
          K, a   =  determination  point  value  of  Kia  estimated  according
                   to Section  7.2.1  or  Section 7.2.2.

           La20 =  determination  point  value  of  KLa  corrected  to  20°C

          e     =  empirical temperature  correction  factor,  defined by
                   Equation  16; shall be  taken equal  to  1.024  unless
                   proven  to have a  different value  for  the  aeration
                   system  and  tank tested (see Section 9.3).
                =  determination  point  value  of  C*   estimated  according  to
                   Section 7.2.1  or  Section 7.2.2.
                   determination  point  value  of  C*  corrected  to  20°C  and
                   a  standard  pressure  of  1.00  atm.

                =   temperature correction  factor =
                                   'st
                                                    's20
                                     322

-------
          C .     =  tabular value of dissolved oxygen surface saturation
                    concentration, mg/1,  at the test temperature and a
                    standard total pressure of 1.00 atm, (5)

                 =  tabular value of dissolved oxygen surface saturation
                    concentration, mg/1,  at 20°C and a standard total
                    pressure of 1.00 atm, shall be taken as 9.07 mg/L (5)


                                                 Pb + Ywtde ' Pv20
          ft      =  pressure correction factor = -5 — T - :p - 5^ —
                                                 Ks   Ywsae " Hv20

          P.      =  barometric pressure during test, f/L2.

          Pv20   =  saturatecl vapor pressure of water at 20°C.
                                                                 2
          P      =  standard barometric pressure of 1.00 atm, f/L .

          Y +     =  weight density of water at the test water temperature,
           wt
          Yws    =  weight density of water at 20.0°C,  f/L3.

          P .     =  saturated vapor pressure of water at the  test temperature,
           vt       f/L2-

          d      =  effective saturation depth at infinite time,  defined by
H — .
de Y
rw
*
C*
oo
K
                                        -  Pb
     The average values of Ki_a2o and c*20 shall  be calculated by averaging
the values at each of the n determination points by

                                             1    n
                    Average K, a  =  -p-—	= —  i Ki a9n          18
                             L      K, a?n    n    T  L 20
                    Average C^ = -, - =  -  "   *             19
     The Standard Oxygen Transfer Rate (SOTR)  shall  be computed by


                    SOTR  =  F- -- j -                       ,20
                             KLa20   C     V
                                     323

-------
  where:   V  =  volume of water in the test tank

     The individual and average values of K,a?n, C*2Q,  de,  and the actual
test temperature and tank volume shall be reported along with the SUIK.
For subsurface gas injection systems, the value of SOTE should also be re-
ported (See Section 8.4).  If possible, the standard deviations of tne
&1&, C*, parameter estimates should also be reported/


8.2   Spatial  Uniformity  and  Reproducibility of  KLa and C* Values.  In some
cases,  replicate  tests are conducted  sequentially under^the same conditions
of  temperature  and pressure  and the  replicate KLa and C^ values can be com-
pared directly  without temperature and pressure adjustments.  However, when
temperature and pressure vary between replicates, comparisons should be
made  using the  K^O and COT2o values.
      8.2.1   Spatial Uniformity of KLa20 and C*20 Values Related to Mixing
 and  Data Adequacy.Uniformity of the point KLa2Q and CJ^O values between the
 various determination points is a necessary condition for theoretical complete
 mixing with  respect to dissolved oxygen concentration and oxygen transfer.
 Non-uniformity of these parameters indicates that theoretical complete mixing
 is not achieved.  Although non-uniformity does not necessarily invaldiate
 the  test, it does suggest non-uniform aeration intensity or short circuiting
 of the flow with some zones of the tank achieving higher oxygen transfer rates
 than  others.  Therefore, when non-uniformity is evident, care should be taken
 to ensure that a sufficient number of determination (sampling) points are
 adequately located to correctly sense the changes in dissolved oxygen inventory
 of the tank.

      The spatial variation of average point K^Q values determined from the
 results of at least three replicate tests can be used to judge the adequacy
 of determination point numbers and locations.  When a minimal number of deter-
 mination points (e.g. four) are employed, variation of these average point
 values should be limited so that three-fourths of the values are within ± 10%
 of the mean value for the tank.  When spatial variations greater than this
 are observed consideration should be given to using a greater number of deter-
 mination points (e.g. six to eight) or to testing by tank sections.

      8.2.2  Reproducibility of Point KI^Q Values as a Criterion of Validity.
 When  a series of atleast three replicate tests are conducted, the point
 values determined in each replicate shall not vary by more than ±15% from
 the mean value for that point.  Replicate point K|_apo values which exhibit
 greater variation shall be considered as invalid and shall not be used for
 calculation of the measured SOTR.  Deletion of data from a determination point
 shall invalidate the replicate test unless suitable data from a backup probe
 are available for that test.

8.3  Standardized Aeration Efficiency.  Frequently, the Standarized Aeration
Efficiency (SAE) or rate of oxygen transfer per unit power input is of in-
terest and shall be computed from:
          SAE =  SOTR/Power  Input                                  21

                                    324

-------
This parameter is normally expressed in units of pounds per horsepower hour
or kilograms per kilowatt hour.

8.4  Oxygen Transfer Efficiency (OTE).  Oxygen transfer efficiency (OTE)
refers to the fraction of oxygen in an injected gas stream dissolved under
given conditions.  Frequently, the Standardized Oxygen Transfer Efficiency
(SOTE) which refers to the OTE at a given gas rate.  Water temperature of
20°C, and barometric pressure of 1.00 atm is calculated.  For a given flow
rate of air, this is given by


                _  SOTR Ib/hr                                     ??
                ~   1.034 Qs                                      "


     where: Qs   =  volumetric  air flow rate, scfm

 8.5  Performance Evaluation Criteria.  Required oxygen transfer performance
 should be stated as  a required standardized oxygen transfer  rate  (SOTR)
 for specified conditions of tank volume,  placement geometry,  power  and  gas
 flow-rate.   System performance under the  specified conditions shall  be
 evaluated by a minimum of three replicate tests.   The measured SOTR for
 performance evaluation shall  be determined as  the  average  SOTR based on
 all valid replicate  tests under the specified  conditions,  and this  average
 shall equal or exceed the required value.   Furthermore,  the  individual  SOTR
 values determined from at least two-thirds of  the  valid replicate tests
 shall exceed the required value.

9.  Application of Clean Water Test Results to  Estimate Oxygen Transfer  Rates
    in Dirty Water at Process DO Levels.

         The SOTR value determined by clean water non-steady  state oxygen
transfer tests may be applied to estimate  the oxygen transfer rate,  OTR^,
for the same oxygenation system operating  in the same tank  under the same
conditions, but in dirty water at an average process level  DO concentration
and temperature.  The estimated value of OTRf shall be calculated  by
OTRf  .  a(SOTR)9'V20'
                                                             23
                    *

                   C»20
where:
       OTRf    =  oxygen transfer rate estimated for the system operating
                  in dirty water at an average DO concentration C and
                  temperature If

       T       =  temperature at the dirty water, °C
                  average               	
                  dirty water KLa       Ki af?o
                  average            ~  ^^
                  clean water K. a       K. a?n

                                     325

-------
                                            *
                                           C sf
 t.p      =   temperature correction, factor = —*	

                                           Cs20

 C   f    =   tabular value of dissolved oxygen surface saturation
  b       concentration, mg/1, in clean water at the temperature
           of  the dirty water.  See Standard Methods (5) Table
           421:1.

                      -*•
                      r
0          dirty water  °°
£5       _   	*	—	

           clean water C
                       oo

if       =   pressure correction factor
                                               Pbf + V^ - Pv20

                                               Ps + V^ ' Pv20
                  =   tabular value of DO surface saturation concentration at
                     20°C,  shall be taken as 9.07 mg/1.

          d        =   average effective depth, calculated as the average
                     of the de values measured at the clean water test
                     determination points =
                                             1
                                            —  E  d
                                             n   -I   e
          Pbf      =   barometric pressure for the dirty water application

          C        =   dissolved oxygen concentration averaged over the entire
                     dirty  water volume

          Other symbols are as defined in Section 8.2.

     Although the application of clean water SOTR values to estimate transfer
 rates in  dirty water is conceptually straightforward, the estimate of OTRf is
 subject to considerable uncertainty because of the uncertainty contained in
 the a value.  This uncertainty is magnified when the dirty water application
 is based  on tank geometry and temperature which differ from those of the clean
 water test.

     Table 4 is a guide to the application of Equation 23 and indicates the
 source_of information for the parameters needed to estimate OTR^.  Values of
 C^OQ, d  and SOTR must be known from the clean water test.  The average DO
 value    C must be determined from the dirty water conditions and should be
evaluated as the process level DO concentration averaged over the entire
aeration volume.   It should not be taken as the DO concentration in the in-
fluent to a point source aerator.  The temperature correction factor, -ty,
and pressure correction factor, &p, should be calculated based on the defini-
tions following Equation 23.
                                      326

-------
Table 4.  Guide to the Application of Equation 23
  Parameter                 Source of Information
      °20                    clean water test results
     e                      clean water test results
    SOTR                    clean water test results
    C                       Given by the dirty water
                            conditions
     f                      Given by the dirty water
                            conditions
    Tf                      Calculated based on tabulated
                            DO surface saturation values
                            Calculated based on d  and
                            barometric pressure
                            estimated based on experience
                            and on measured K.a values
                            Calculated based on total dissolved
                            solids measurements
                            Taken as 1.024 unless experimentally
                            proven to differ
                           327

-------
9.1  Alpha  Factor, a.  The alpha factor represents the ratio of the KLa in
process wastewater to the KLa in clean water, and this ratio can range rrom
approximately 0.4 to 1.2.  It is influenced by a great number of process
conditions  including surfactants, tubulence, power input per unit volume,
geometry, scale, bubble size, degree of treatment, and other wastewater
characteristics.  Ideally, the alpha factor would be measured by conducting
full scale  oxygen transfer tests with clean water and wastewater, but this
is  normally impractical.  Several studies have described small scale (less
than 50 gallons) oxygen transfer tests for measurement of the alpha factor
and a  state of the art method based on these studies has been recommended by
Stenstrom (4).  If alpha factor measurements in small scale vessels are
necessary,  the method given by Stenstrom should be followed.  However, alpha
factor measurement in small scale vessels is, at best, only an educated guess.
In  selection of an alpha factor for use in Equation 23, it should be borne
in  mind that, for a given wastewater stream, the alpha factor is normally not
constant and a possible range of alpha values should be considered.

9.2 Beta Factor, g . The beta factor is defined as the ratio of the average
saturation  concentration, IT, in wastewater to the corresponding value in
clean  water.  This ratio can vary from approximately 0.8 to 1.0 and is
generally close to 1.0 for municipal wastewaters.   Because it cannot be
measured by a membrane probe and because many wastewaters contain substances
which  interfere with the Winkler method, it is difficult to measure accurately.
For this reason, the value of beta for use in Equation 23 shall be calculated
as  the ratio of DO surface saturation concentration in the waste to the DO
surface saturation concentration in clean water.  The corresponding surface
saturation  concentrations shall  be interpolated from Table 5 based on the total
dissolved solids content of the wastewater and clean water.  Table 5 is
adapted from Table 421:1 of Standard Methods (5) by converting the chloride
to  TDS for  seawater.

9.3 Theta  Factor, e.  The theta factor is employed to correct K|_a for changes
in  temperature according to Equation 16.  Values of theta reported in the
literature  have ranged from 1.008 to 1.047 and are influenced by geometry,
turbulence  level  and type of aeration device.  There is little consensus
regarding the accurate prediction of theta values, and for this reason, clean
water  testing for the determination of SOTR values should be at temperatures
close  to 20°C.

     The value  of theta shall  be taken equal to 1.024 in Equations 16 and
23.  Use of a theta  factor differing from 1.024 shall be permissible only
when experimental  data  for the particular aeration system indicates conclu-
sively that the value of theta is significantly different from 1.024.
                                     328

-------
Table 5.  Dissolved Oxygen Surface Saturation Concentrations, mg/1 at
          Various Temperatures and Approximate Total Dissolved Solids
	Concentrations.  Adapted from Standard Methods (5).	
                  	Total Dissolved Solids mg/1	


 Temp. °C              0         9100         18,200        27,200

    10                11.27      10.65         10.05         9.49

    15                10.07       9.53          9.01         8.53

    20                 9.07       8.60          8.16         7.73

    25                 8.24       7.83          7.44         7.06

    30                 7.54       7.17      .    6.83         6.49
ACKNOWLEDGEMENTS

      This proposed standard is based on nearly four years of work by
 the ASCE Subcommittee on Oxygen Transfer Standards.  A significant por-
 tion of the Committee's work was supported by grants from the U.S.
 Environmental  Protection Agency.

      Preparation of a standard general  enough to be applied to all  clean
 water unsteady state tests and specific enough to incorporate all essen-
 tial procedures was difficult.  Users of this standard must give partic-
 ular attention to use of the mandatory "shall" and advisory "should"
 terms.   For particular applications of this standard, it may be advan-
 tageous for the user to elevate certain advisory steps to the mandatory
 level.   It is  intended that this- standard be referenced in engineering
 specification  for compliance testing.  When this is the case, the
 engineer should consider the costs of requiring extensive compliance
 testing in relation to the initial cost of the oxygen transfer system
 and present worth of future operating costs.

      Several engineers, owners and manufacturers have developed sound
 oxygen  transfer testing procedures and the procedures incorporated into
 this standard  were judged to be the most accurate and generally appli-
                                      329

-------
 cable of those available.  The substance of this standard is based on
 recommendations made in the report, "Development of Standard Procedures
 for Evaluating Oxygen Transfer Devices" by the A.S.C.E. Oxygen Transfer
 Standards Subcommittee, W. C. Boyle, Chairman.  The user is referred to
 this document which contains references 1 to 4 and to reference / tor
 background information.
  REFERENCES

  1.   Baillod,  C.R.  and  Brown, L.C., "Modelling and Data Interpretation,"  in
      Development  of Standard Procedures for Evaluating Oxygen Transfer
      Devices".  A.S.C.E. Oxygen Transfer Standards Subcommittee, Ed. W.C.  Boyle
      Final  Report to U.S.  EPA, Cooperative Agreement CR805868 0982).
  2.   Paulson,  W.L.  and  Rooney, T., "Unsteady State Clean Water Test," in
      Development  of Standard Procedures for Evaluating Oxygen Transfer
      Devices."  A.S.C.E. Oxygen Transfer Standards Subcommittee, Ed. W.C.  Boyle
      Final  Report to U.S.  EPA, Cooperative Agreement CR805868 (1982).
  3.   Yunt,  F.  and Salman,  R., "Gas Flow Measurements, Power Measurements,"
      in  Development  of  Standard Procedures for Evaluating Oxygen Transfer
      Devices",  A.S.C.E. Oxygen Transfer Standards Subcommittee, Ed. W.C.  Boyle
      Final  Report to U.S.  EPA, Cooperative Agreement CR805868 (1982).
  4.   Stenstrom, M.K.  and Gilbert, R.G., "Effects of Alpha, Beta and Theta
      Factors and  Surfactants Upon Specification, Design and Operation of
      Aeration  Systems," in Development of Standard Procedures for Evaluating
      Oxygen Transfer Devices," A.S.C.E.  Oxygen Transfer Standards Subcommittee
      Ed. W.C.  Boyle,  Final Report to U.S.  EPA, Cooperative Agreement CR805868
      (1982).
  5.   Standard Methods for the Examination  of Water and Wastewater, 15th
      Edition, A.P.M.A., W.P.C.F., A.W.W.A. (1980).u
  6.   Standard Methods for the Examination  of Water and Wastewater, llth
      Edition, A.P.H.A., W.P.C.F., A.W.W.A. (1960).
  7.   "Proceedings:   Workshop Toward an Oxygen Transfer Standard", Ed. by
      W. C. Boyle, EPA-600/9-78-021 (1979).
  8.   Spink, L.K., Principles and Practice  of Flow Meter Engineering. Ninth
      Edition, Foxboro Company, Foxboro,  MA (1967).
  9.   Cusick, C.F., Flow Meter Engineering  Handbook. Fourth Edition, Honeywell,
      Fort Washington, PA (1968).
10.  American Society of Mechanical  Engineers, Power Test Codes Instruments
     and Apparatus Supplement -
     Part 2.  Pressure Measurement,  PTC 19.2 (1964).
     Part 5.  Chapter 4, Flow Measurement, PTC 19.5;4 (1959).
     Part 6.  Electrical Measurements  in Power Circuits, PTC 19.6 0955).
     Part 7.  Measurement of Shaft Horsepower, PTC 19.7 (1961).
     Part 13.   Measurement of Rotary Speed, PTC 19.13 (1961).
     ASME,  345 East 47th Street,  New York, NY.

DISCLAIMER

     The work summerized herein  was previously reviewed and approved for
publication by the U.S. Environmental  Protection Agency (EPA-600/2-83-002).

                                     330

-------
                          Appendix A
           Fortran Program for Non Linear Regression

     The Fortran computer program which follows in this Appendix is a
non linear estimation of the parameters in the exponential  form of the
oxygen transfer model.  The program uses the Taylor series  linearization
method and is in two parts.  For greatest user ease, the computational
and output portions are written in subroutine form (subroutine KLANL).
The user supplied main program serves as an interface between the system
constraints of the user's computer and the regression computational
algorithm, in the subroutine.

     The MAIN program must perform the following tasks.  A  sample MAIN
program is shown in the Fortran listing.

     1.  Read in the data to be fitted; concentration (C) versus
         time (T).
     2.  Read in initial parameter estimates (CS, CO, XKLA).
     3.  Provide a descriptive name for the data set (INAME).
     4.  Provide the logical device number for output (NOUT).

     Subroutine KLANL performs all the computations for the nonlinear
estimation and controls the output of the results.  The notation in the
subroutine is consistent with that in this Appendix.  The estimation and
output proceeds in the following steps.

     Step 1.  Write titles, name of data set and headings for the
              iterations.
     Step 2.  Initialize internal variables and compute the fitted
              values and residual sum of squares based on the initial
              parameter estimates.
     Step 3.  Set up the normal equations using current parameter
              estimates.
     Step 4.  Solve normal equations for the corrections to the parameter
              estimates.
     Step 5.  Update parameter estimates and calculate new  fitted values
              and residual sum of squares.
     Step 6.  Test for convergence.  The convergence criteria are:
              a.  Relative change in parameters less than 0.00001.
              b.  Relative change in sum of squares less than 0.000001.
              c.  The algorithm will also exit from the iteration loop
                  if more than 10 iterations are required for conver-
                  gence.  A diagnostic message is printed.
     Step 7.  Calculate the estimated standard deviations of the model
              parameters.
     Step 8.  Write out a summary of the data, fitted values (F) and
              residuals (R).

An example estimation problem follows the Fortran listing.
                                    331

-------
00100
00200
00300
00400
00500
00600
00700
00800
00900
01000
01100
01200
01300
01400
01500
01600
01700
01800
01900
02000
02100
02200
02300
02400
02500
02600
02700
02800
02900
03000
03100
03200
03300
03400
03500
03600
03700
03800
03900
04000
04100
04200
04300
04400
04500
04600
04700
04800
04900
05000
05100
05200
05300
05400
C
C
C
C


1

2


10

15


19





C
C
C
C



C
C
C

20



21

C
C
C






50

99



 MAIN PROGRAM FOR NONLINEAR ESTIMATION OF OXYGEN
         TRANSFER PARAMETERS

 DIMENSION C(100),T(100),F(100),R(100)
 WRITE(5,1)
 FORMATC  NAME OF INPUT DATA FILE:     $)
 READ(5,2) INAME
 FORMAT(A5)                                      .  .
 OPEN(UNIT=20,MODE='ASCII',ACCESS='SEQIN',FILE=INAME)
 READ(20,10) CS,CO,XKLA
 FORMAT(3F)
 N=l
 READ(20,10,END=19) T(N),C(N)
 N = N+l
 GO TO 15
 CONTINUE
 NOB = N-l
 NOUT = 5
 CALL KLANL(C,T,F,R,NOB,CS.CO.XKLA,INAME,NOUT)
 CALL EXIT
 END

 SUBROUTINE KLANL PERFORMS  NONLINEAR ESTIMATION COMPUTATIONS
         AND CONTROLS OUTPUT SUMMARIES

 SUBROUTINE KLANL(C,T,F,R,NOB,CS,CO,XKLA,INAME,NOUT)
 DIMENSION C(100),T(100),F(100),R(100)
 REAL K1SK2,K3

 STEP 1 - WRITE TITLES

 WRITE(NOUT,20) INAME
 FORMAT(//20X,'NON LINEAR ESTIMATION'/
115X,'UNSTEADY STATE OXYGEN TRANSFER1/
223X,'DATA SET',A5)
 WRITE(NOUT,21)
 FORMAT(///'  ITERATION',29X,'KLA',8X,'SUM OF'/
12X,'NUMBER1,6X,'C-STAR1,6X,'C-ZERO1,6X,1 PRIME1,6X,'SQUARES1/)

 STEP 2 - INITIALIZATION OF VARIABLES
 K=0
 OSSQ =0.0
 DO 50 1=1, NOB
 F(I) = CS-(CS-CO)*EXP(-XKLA*T(I))
 OSSQ = OSSQ + R(I)*R(I)
 WRITE(NOUT,30) K,CS,CO,XKLA,OSSQ
 K=K+1
 All=0.0
 A12=0.0
 A13=0.0
                     332

-------
05500
05600
05700
05800
05900
06000
06100
06200
06300
06400
06500
06600
06700
06800
06900
07000
07100
07200
07300
07400
07500
07600
07700
07800
07900
08000
08100
08200
08300
08400
08500
08600
08700
08800
08900
09000
09100
09200
09300
09400
09500
09600
09700
09800
09900
10000
10100
10200
10300
10400
10500
10600
10700







C
C
C
C















1
C
C
C
C











C
C
C






i;
C
A22=0.0
A23=0.0
A33=0.0
Cl=0.0
C2=0.0
C3=0.0
SSQ=0.0

STEP 3 - SET UP NORMAL EQUATIONS FOR LINEARIZED MODEL
        USING CURRENT LEAST SQUARES ESTIMATES.

DO 100 1=1,NOB
Z2=EXP(-XKLA*T(I))
Z1-1.0-Z2
Z3 = T(I)*Z2*(CS-CO)
All = All + Z1*Z1
A12 = A12 + Z1*Z2
A13 = A13 + Z1*Z3
A22 = A22 + Z2*Z2
A23 = A23 + Z2*Z3
A33 = A33 + Z3*Z3
F(I) = CS - (CS-CO)*Z2
R(I) = C(I) - F(I)
Cl = Cl + R(I)*Z1
C2 = C2 + R(I)*Z2
C3 = C3 + R(I)*Z3
CONTINUE

STEP 4 - SOLUTION OF NORMAL EQUATIONS FOR CORRECTIONS TO
        THE PRIOR LEAST SQUARES ESTIMATES

Dl = A11*A22 - A12*A12
D2 = A11*C3 - A13*C1
D3 = A11*A23 - A13*A12
D4 = A33*A11 - A13*A13
D5 = A11*C2 - A12*C1
BN3 = D1*D2 - D3 *D5
BD3 = D1*D4 - D3*D3
B3 = BN3/BD3
BN2 = D5 - D3*B3
B2 = BN2/D1
Bl = (Cl - A12*B2 - A13*B3)/A11

STEP 5 - UPDATE ESTIMATES, SUM OF SQUARES

Kl = Bl + CS
K2 = B2 + CO
K3 = B3 + XKLA
DO 120 1=1,NOB
     = Kl -(K1-K2)*EXP(-K3*T(I))
     = C(I) - F(I)
SSQ = SSQ + R(I)*R(I)
                 333

-------
10800
10900
11000
11100
11200
11300
11400
11500
11600
11700
11800
11900
12000
12100
12200
12300
12400
12500
12600
12700
12800
12900
13000
13100
13200
13300
13400
13500
13600
13700
13800
13900
14000
14100
14200
14300
14400
14500
14600
14700
14800
14900
15000
15100
15200
15300
15400
15500
15600
15700
15800
C
C



C
C
C
150
30

C
C
C
200
C
C
C


C





300




22

23

C
C
C













 STEP 6 - TEST FOR CONVERGENCE - PARAMETERS

 IF(B1/K1.GE.0.00001) GO TO 200
 IF(B2/K2.GE.0.00001) GO TO 200
 IF(B3/K3.GE.0.00001) GO TO 200

 ITERATIONS CONVERGED

 WRITE(NOUT,30) K,K1,K2,K3,SSQ
 FORMAT(3X,I4,3X,4E12.4)
 GO TO 300

 PARAMETERS NOT CONVERGED, TEST SUM OF SQUARES

 IF(ABS((OSSQ-SSQ).LE.0.000001) GO TO 150

 SUM OF SQUARES NOT CONVERGED, TEST ITERATIONS

 IF(K.GT.IO) GO TO 350
 WRITE(NOUT,30) K,K1,K2,K3,SSQ
 NEW ESTIMATES
 CS = Kl
 CO = K2
 XKLA = K3
 OSSQ = SSQ
 GO TO 99
 CONTINUE
 XDF = NOB - 3
 RSM = SSQ/XDF
 ERROR = SQRT(RSM)
 WRITE(NOUT,22) ERROR
 FORMAT(/' ESTIMATE OF ERROR FROM RESIDUAL MEAN SQUARE  ',F6.2)
 WRITE(NOUT,23)
 FORMAT(//17X,'STANDARD DEVIATIONS1/
lieX.'OF PARAMETER ESTIMATES1//)

 STEP 7 - COMPUTE STANDARD DEVIATIONS OF THE PARAMETER ESTIMATES

 DETP = A11*A22*A33 + 2.0*A12*A13*A23
 DETN = A11*A23*A23 + A22*A13*A13 + A33*A12*A12
 DET = DETP-DETN
 Ell = A22*A33 - A23*A23
 E22 = A11*A33 - A13*A13
 E33 = A11*A22 - A12*A12
 VARK1 = (E11/DET)*RSM
 VARK2 = (E22/DET)*RSM
 VARK3 = (E33/DET)*RSM
 SIGCS = SQRT(VARKl)
 SIGCO = SQRT(VARK2)
 SIGKL = SQRT(VARKS)
 WRITE(NOUT,31) SIGCS,SIGCO,SIGKL
                       334

-------
15900
16000
16100
16200
16300
16400
16500
16600
16700
16800
16900
17000
17100
17200
17300
17400
17500
17600
17700
17800
17900
31
32
C
C
C

33
370
34

350
35
400
 FORMATC  ABSOLUTE'/aX'UNITS'.ZX.SElZ^)
 SIGCS = SIGCS/CS*100.0
 SIGCO = SIGCO/CO*100.0
 SIGKL = SIGKL/XKLA*100.0
 WRITE(NOUT532) SIGCS.SIGCO,SIGKL
 FORMAT(/'  PER CENT'/2X'OF LSE1,2X,3F12.1)

 STEP 8 - WRITE FINAL SUMMARY

 WRITE(NOUT,33)
 FORMAT(//15X,1SUMMARY OF DATA'//
110X,'TIME1,6X,'CONC',5X,'FITTED',3X,1RESIDUAL1/
230X)'VALUE1/)
 DO 370 1=1,NOB
 WRITE(NOUT,34) I,T(I),C(I),F(I),R(I)
 FORMAT(I5,4F10.2)
 GO TO 400
 WRITE(NOUT,35)
 FORMAT(//3X,'SOLUTION NOT CONVERGED IN 10 'ITERATIONS!'//)
 RETURN
 END
                                      335

-------
 FORTRAN:   KLANLN
 MAIN.
 KLANL
 LINK:     Loading
 CLNKXCT KLANLN  execution

  NAME  OF  INPUT  DATA  FILE:   COMB
                      NON LINEAR ESTIMATION
                 UNSTEADY STATE OXYGEN TRANSFER
                        DATA SET COMB
ITERATION . KLA
NUMBER C-STAR C-ZERO PRIME
0 0.1200E+01 0.5000E+00 0.1500E+00
1 0.1116E+02 0.1935E+01 0.6928E-01
2 0.1133E+02 0.1186E+01 0.8733E-01
3 0.1143E+02 0.1122E+01 0.8691E-01
4 0.1143E+02 0.1122E+01 0.8692E-01
5 0.1143E+02 0.1122E+01 0.8692E-01
ESTIMATE OF ERROR FORM RESIDUAL MEAN SQUARE
SUM OF
SQUARES
0.4893E+02
0.6109E+01
0.7795E-01
0.1617E-01
0.1617E-01
0.1617E-01
0.03
 ABSOLUTE
   UNITS

 PER  CENT
  OF  LSE
         STANDARD DEVIATIONS
        OF PARAMETER ESTIMATES


0.1822E-01  0.3769E-01   0.6443E-03
       0.2
3.4
0.7
                         SUMMARY OF DATA

          TIME       CONC           FITTED
                                     VALUE
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
18.00
20.00
22.00
25.00
30.00
35.00
40.00
45.00
50.00
55.00
2.77
4.15
5.35
6.25
7.08
7.80
8.34
8.85
9.28
9.62
9.93
10.24
10.70
11.00
11.14
11.20
11.25
11.30
                         5.
                         6.
                         7.
                         7.
      2.77
      4.15
       ,31
       .29
       .11
       .80
      8.37
      8.86
      9.27
      9.61
      9.90
     10.25
     10.67
     10.93
     11.11
     11.22
     11.29
     11.34
    336
       RESIDUAL

          0.00
          0.00
          0.04
         -0.04
         -0.03
          0.00
         -0.03
         -0.01
          0.01
          0.01
          0.03
         -0.01
          0.03
          0.07
          0.03
         -0.02
         -0.04
         -0.04

-------
                           Appendix B
             Basic Program for Non Linear Regression
     This appendix gives the BASIC computer language adaption of the
FORTRAN non linear estimation program discussed in Appendix A.   Also
given are examples of output obtained by applying the BASIC program
to typical data sets using an APPLE II Microcomputer.
                                 337

-------
10
20
30
40
50
60
70
80
90
100
110
120
130
140
150
160
170
180
190
200
210
RINT
220
230
240
250
260
270
280
290
300
310
320
330
340
350
360
370
380
390
400
410
420
430
440
450
460
470
480
490
500
510
520
530
540
D$ =  CHR$ (4)
 REM
 REM
 REM
 REM
 TEXT
 REM
 REM
 REM
  REM
  REM  ::::::::::::::::::::::
  REM
  PRINT "***************************************"
  PRINT "       NON-LINEAR ESTIMATION FOR"
  PRINT "     UNSTEADY-STATE OXYGEN TRANSFER"
  PRINT "***************************************"
  PRINT "                 BY"
  PRINT "LINFIELD C. BROWN & GEORGE R. FISETTE"
  PRINT "VERSION  1.0-NOVEMBER 11, 1979"
NON-LINEAR LEAST SQUARES PROGRAM IN APPLE II BASIC
FOR OXYGEN TRANSFER PARAMETERS
OUTPUT SETUP FOR 40 POSITION CRT/MONITOR
: CALL  - 936: REM  CLEARS SCREEN
STEP 1
 WRITE TITLES
  PRINT
  INVERSE :  PRINT "THE VALUES ARE TRUNCATED": PRINT "AND NOT ROUNDED OFF." :P
  : NORMAL
  REM
  REM  PROGRAM HAS MAXIMUM LIMIT OF 30 DATA POINTS
  REM
  DIM C(30),T(30),F(30),R(30)
  INPUT "IS DATA IN DISK FILE;Y/N?";A$
  INPUT "INPUT NAME OF DATE FILL?";N$
  IF A$ = "  " GOTO 650: REM  GET DATA FROM DISK FILL
  INPUT "DO YOU WANT INPUT DATA SAVED ON DISK,Y/N?";A$
  PRINT "INPUT DATA IN TIME,DO DATA PAIRS"
  PRINT "INPUT 999,999 AS LAST DATA PAIR"
  FOR I = 1  TO 30
  INPUT T(I),C(I)
  IF T(I) =  999.0 GOTO 360
  NEXT I
 ND = I - 1.0
  INPUT "BEST ESTIMATE FOR C-STAR OR USE 10.0 MG/L?";CS
  INPUT "BEST ESTIMATE FOR C-ZERO OR USE 0.0 MG/L?"-,CO
  INPUT "BEST ESTIMATE FOR KLA-PRIME OR USE 4.0 1/HR?";XK
 XK = XK / 60.0
  IF A$ = "N" GOTO 790
  REM
  REM  WRITE DATA TO DISK FILE
  REM  SPECIFIC FOR APPLE/MICROSOFT BASIC
  REM
  PRINT D$;"OPEN "N$;",VO,L15"
  FOR I = 1  to ND
  PRINT D$;"WRITE "N$;",BO,R";I
  PRINT T(I):  PRINT C(I)
  NEXT I
  PRINT D$;"WRITE "N$;",BO,RO"
  PRINT ND
  PRINT D$;"WRITE "N$;",BO,R";ND + 1.
  PRINT  CS
                                       338

-------
550   PRINT D$;"WRITE "N$;",BO,R";ND + 2.
560   PRINT CO
570   PRINT D$;"WRITE "N$;",BO,R";ND + 3.
580   PRINT XK
590   PRINT D$;"CLOSE "N$
600   GOTO 790
610   REM
620   REM  READ DISK FILE FOR DATA
630   REM  SPECIFIC FOR APPLE/MICROSOFT BASIC
640   REM
650   PRINT D$;"OPEN "N$;",VO,L15"
660   PRINT D$;"READ "N$;",BO,RO"
670   INPUT ND
680   FOR I = 1 TO ND
690   PRINT D$;"READ "N$;",BO,R";I
700   INPUT T(I),C(I)
710   NEXT I
720   PRINT D$:"READ "N$;">BO,R";ND + 1.
730   INPUT CS
740   PRINT D$;"READ "N$;",BO,R";ND + 2.
750   INPUT CO
760   PRINT D$;"READ "N$;",BO,R";ND + 3.
770   INPUT XK
780   PRINT D$;"CLOSE "N$
790   PRINT : FLASH : INPUT "HIT RETURN FOR ITERATIONS.";I$: NORMAL
800   CALL    936: PRINT : PRINT "          DATA SET ";N$: PRINT
810   PRINT "ITERATION" TAB( 11)"C-STAR" TAB( 18)"C-ZERO" TAB( 26)"KLA" TAB ( 33)"
SUM OF"
820   PRINT TAB( 2)"NUMBER" TAB( 26)"PRIME" TAB( 33)"SQUARES"
830   PRINT TAB ( 11)"(M6/L)" TAB( 18)"(MG/L)" TAB ( 26)"(1/HR}"
840   PRINT
850   REM
860   REM  ::::::::::::::::::::::
870   REM  STEP 2
880   REM  INITIALIZATION OF VARIABLES
890   REM  DO ITERATION CALCULATIONS
900   REM  ::::::::::::::::::::::
910   REM
920  K% - 0
930  OS = 0.0
940   FOR I = 1 TO ND
950  F(I) = CS - (CS - CO) * EXP ( - XK * T(I))
960  R(I) = C(I) - F(I)
970  OS = OS + R(I) * R(I)
980   NEXT I
990  ZZ$ = STR$ (CS) :VA = 5.: GOSUB 2900
1000  CS$ = ZZ$:ZZ$ = STR$ (CO): GOSUB 2900
1010  C0$ = ZZ$:ZZ$ = STR$ (XK * 60.): GOSUB 2900
1020  XK$ = ZZ$:ZZ$ = STR$ (OS): GOSUB 2900
1030  OS$ = ZZ$
1040   PRINT TAB( 4)K% TAB( 10)CS$ TAB( 18)CO$ TAB( 26)XK$ TAB( 33)05$
1050   GOTO 1070
1060   REM
1070   REM  CALCULATION LOOP - INITILIZE VARIABLES
                                        339

-------
1080
1090
1100
1110
1120
1130
1140
1150
1160
1170
1180
1190
1200
1210
±. Lm J.U
1220
1230
1240
1250
±L.*J\J
1260
1270
1280
1290
1300
1310
1320
1330
1340
1350
1360
1370
1380
1390
1400
1410
1420
1430
1450
1460
1470
1A8D
J.*TUU
1490
1500
1510
1520
1530
1540
1550
1560
1570
1580
1590
1600
1610
REM
K% =
Al =
A2 =
A3 =
A4 =
A5 =
A6 =
Cl =
C2 =
C3 =
SQ =
REM
RFM
r\L.ri
REM
REM
REM
RFM
r\t.i i
REM
FOR
Z2 =
Zl =
Z3 =
Al =
A2 =
A3 =
A4 =
A5 =
A6 =
F(D
R(I)
Cl =
C2 =
C3 =
K% +
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1
STEP 3
SETUP
USING


NORMAL
CURRENT
I = 1 TO ND
EXP ( - XK *
1.0 - Z2
T(I)
Al
A2
A3
A4
A5
A6
+
+
+
+
+
+
*
Zl
Zl
Zl
Z2
Z2
Z3
= CS -
-= C
Cl
C2
C3
;(I
+
+
+
)
R(
R(
R(
Z2 *
* Zl
* Z2
* Z3
* Z2
* Z3
* Z3
(CS -
- F(I
I) *
I) *
I ) *

EQUATIONS FOR LINEARIZED MODEL
LEAST SQUARE ESTIMATES
T(D)
(CS - CO)












CO) * Z2
)
Zl
Z2
Z3




NEXT I
REM
REM
REM
REM
RFM
r\ui i
REM
Dl =
D2 =
D3 =
D4 =
D5 =
XN =
XD =
X3 =
YN =
X2 =
XI =
REM






STEP 4
SOLUTION OF
TO THE


Al
Al
Al
A6
Al
Dl
Dl
XN
D5
YN
(Cl



*
*
*
*
*
*
*
/
-
/


A4
C3
A5
Al
C2
D2
D4
XD
D3
Dl
NORMAL EQUATIONS FOR CORRECTIONS
PRIOR


- A2
- A3
- A3
- A3
- A2
- D3
- D3

* X3



*
*
*
*
*
*
*



- A2 * X2




LEAST SQUARES ESTIMATES


A2
Cl
A2
A3
Cl
D5
D3



-A3 * X3) / Al

340

-------
1620   REM  ::::::::::::::::::::::
1630   REM  STEP 5
1640   REM  UPDATE ESTIMATES, SUM OF SQUARES
1650   REM  ::::::::::::::::::::::
1660   REM
1670  Tl = XI + CS
1680  T2 = X2 + CO
1690  T3 = X3 + XK
1700   FOR I = 1 TO ND
1710  F(I) = Tl - (Tl - T2) * EXP ( - T3 * T(I))
1720  R(I) = C(I) - F(I)
1730  SQ = SQ + R(I) * R(I)
1740   NEXT I
1750   REM
1760   REM  ::::::::::::::::::::::
1770   REM  STEP 6
1780   REM  TEST FOR CONVERGENCE - PARAMETERS   1 PART IN 100,000
1790   REM  :::::::::::::::::::::
1800   REM
1810   IF (XI / Tl > 0.00001) AND (X2 / T2 > 0.00001) AND (X3 / T3 > 0.00001) GOT
0 2160
1820   REM
1830   REM  PARAMETERS NOT CONVERGED,
1840   REM  TEST SUM OF SQUARES - 1 PART IN 1,000,000
1850   REM
1860   IF ABS ((OS - SQ)/SQ)£0.000001 GOTO 2160
1870   REM
1880   REM  SUM OF SQUARES NOT CONVERGED,
1890   REM  TEST NO. OF ITERATIONS
1900   REM
1910   IF (K%   10) GOTO 2090
1920  ZZ$ = STR$ (Tl) : GOSUB 2900
1930  Tl$ = ZZ$:ZZ$ =  STR$ (T2) : GOSUB 2900
1940  T2$ = ZZ$:ZZ$ =  STR$ (T3  * 60.): GOSUB 2900
1950  T3$ = ZZ$:ZZ$ =  STR$ (SQ) : GOSUB 2900
1960  SQ$ = ZZ$
1970   PRINT TAB( 4)K% TAB( 10)T1$ TAB( 18)T2$ TAB( 26)T3$ TAB( 33)SQ$
1980   REM
1990   REM  NEW ESTIMATES
2000   REM
2010  CS = Tl
2020  CO = T2
2030  XK = T3
2040  OS = SQ
2050   GOTO 1090
2060   REM
2070   REM  OUTPUTS
2080   REM
2090   PRINT
2100   PRINT "SOLUTION NOT CONVERGED IN 10 ITERATIONS'"
2110   PRINT  "CHANGE VALUE IN LINE 2670 TO TRY MORE  ITERATIONS."
2120   END
2130   REM
2140   REM  OUTPUT PARAMETER ESTIMATES
                                       341

-------
2150   REM
2160  ZZ$ =  STR$  (Tl):  GOSUB  2900
2170  Tl$ = ZZ$:ZZ$  =  STR$  (T2): GOSUB  2900
2180  T2$ = ZZ$:ZZ$  =  STR$  (T3 * 60.):  SOSUB  2900
2190  T3$ = ZZ$:ZZ$  =  STR$  (SQ): GOSUB  2900
2200  SQ$ = ZZ$                                                       „
2210   PRINT TAB(  4)K% TAB(  10)T1$ TAB(  18)T2$ TAB( 26)T3$ TAB(  33)SQ$
2220   PRINT
2230   REM
2240   REM  ::::::::::::::::::::::
2250   REM  STEP 7
2260   REM   COMPUTE STANDARD  DEVIATIONS OF PARAMETER ESTIMATES
2270   REM  ::::::::::::::::::::::
2280   REM
2290  XF = ND  - 3.0
2300  RS = SQ  / XF
2310  ER = SQR (RS)
2320   PRINT "STD  DEVIATIONS OF PARAMETER ESTIMATES"
2330   PRINT
2340  DP = Al  * A4 * A6  + 2.0  * A2 * A3 * A5
2350  DN = Al  * A5 * A5  + A4 * A3 * A3 + A6 * A2 * A2
2360  DT = DP  - DN
2370  Fl = A4  * A6 - A5  * A5
2380  F2 = Al  * A6 - A3  * A3
2390  F3 = Al  * A4 - A2  * A2
2400  VI = (Fl / DT) * RS
2410  V2 = (F2 / DT) * RS
2420  V3 = (F3 / DT) * RS
2430  SI = SQR (VI)
2440  S2 = SQR (V2)
2450  S3 = SQR (V3)
2460  ZZ$ = STR$ (S1):VA =  5.: GOSUB 2900
2470  Sl$ = ZZ$:ZZ$  =  STR$  (S2): GOSUB 2900
2480  S2$ = ZZ$:ZZ$  =  STR$  (S3 * 60.): GOSUB 2900
2490  S3$ = ZZ$
2500   PRINT " UNITS"  TAB(  10)S1$ TAB( 18)S2$ TAB( 26)S3$
2510  SI = SI  / CS * 100.0
2520  S2 = S2  / CO * 100.0
2530  S3 = S3  / XK * 100.0
2540  ZZ$ = STR$ (S1):VA =  3.: GOSUB 2900
2550  Sl$ = ZZ$:ZZ$  =  STR$  (S2): GOSUB 2900
2560  S2$ = ZZ$:ZZ$  =  STR$  (s3): GOSUB 2900
2570  S3$ = ZZ$
2580   PRINT "% OF LSE"  TAB( 10)51$ TAB( 18)S2$ TAB( 26)53$
2590   PRINT
2600  ZZ$ = STR$ (ER):VA =  4.: GOSUB 2900
2610  ER$ = ZZ$
2620   PRINT "ESTIMATE OF ERROR = ";ER$
2630   REM
2640   REM  ::::::::::::::::::::::
2650   REM  STEP 8
2660   REM  WRITE SUMMARY
2670   REM  ::::::::::::::::::::::
                                      342

-------
2680   REM
2690   PRINT
2700   FLASH :  INPUT "HIT RETURN FOR SUMMARY OF DATA.";I$: NORMAL
2710   CALL - 936: PRINT :  PRINT : REM  CLEARS SCREEN
2720   PRINT TAB( 13)"SUMMARY OF DATA"
2730   PRINT :  PRINT
2740   PRINT TAB( 8)"TIME"  TAB( 16)"CONC" TAB( 22)"FIT VALUE" TAB( 32}"RESIDUAL"
2750   PRINT TAB( 8)"(MIN)" TAB( 15)"(MG/L)" TAB( 23)"MG/L"
2760   PRINT
2770   FOR I = 1 TO ND
2780  ZZ$ = STR$ (F(I)):VA = 4.: GOSUB 2900
2790  Hl$ = ZZ$:ZZ$ = STR$ (R(I)): GOSUB 2900
2800  H2$ = ZZ$
2810   PRINT TAB( 2)1 TAB(  8)T(I) TAB( 16)C(I) TAB( 25)H1$ TAB( 33)H2$
2820   NEXT I
2830   PRINT : PRINT
2840   PRINT "***************************************"
2850   END
2860   REM
2870   REM  OUTPUT FORMATTING ROUTINES
2880   REM  SPECIFIC FOR APPLE/MICROSOFT BASIC
2890   REM
2900  LL = LEN (ZZ$)
2910   IF LL<12 THEN ZZ$ = LEFT$ (ZZ$,VA): RETURN
2920   IF  HID$  (ZZS.LL - 2,1) = "+" THEN ZZ$ =-LEFTS (ZZ$,VA - 3) + RIGHTS  (ZZ
$ 3):  RETURN
2930  CC = 2.: IF LEFT$ (ZZ$,1)   =   "-" THEN CC = 1.
2940   IF MID$ (ZZS.LL - 3,1) = "E" THEN EE = VAL (RIGHTS (ZZ$,2)):NN$ = MID$
 (ZZ$,CC,1): FOR J = 1 TO EE:NN$ = "0" + NN$:: NEXT J:ZZ$ = "." NN$ + MID$ (Z
ZS.CC + 2,LL - 4): IF CC = 2. THEN ZZ$ = "-" + ZZ$
2950  ZZS - LEFTS (ZZ$,VA): RETURN
2960   REM
2970   REM  NON-LINEAR LEAST SQUARES PROGRAM FOR
2980   REM  UNSTEADY-STATE OXYGEN TRANSFER
2990   REM  LY LINFIELD C.  BROWN & GEORGE R. FISETTE
3000   REM  VERSION  1.0-NOVEMBER 11, 1979
3010   REM  COPYRIGHT BY ASCE
                                     343

-------
RUN
************************************************************************
            NON-LINEAR ESTIMATION FOR UNSTEADY-STATE OXYGEN TRANSFER
*******************************************************************************
BY
LINFIELD C. BROWN & GEORGE R. FISETTE
VERSION  1.0-NOVEMBER 11, 1979
IS DATA IN DISK FILE; Y/NPY
INPUT NAME OF DATA FILE7TEST2
                               DATA SET TEST2
ITERATION
NUMBER
0
1
2
3
 C-STAR

 10
 10.1592458
 10-1889055
 10.1893432
C-ZERO
.1
.713872799
.694460338
.694630637
KLA
PRIME
.07
.0601645437
.0603440105
.0603354201
SUM OF
SQUARES
.869733335
.189522011
.185289531
.185289365
STANDARD DEVIATIONS OF PARAMETER ESTIMATES

ABSOLUTE UNITS  .195594007      .111803446
PERCENT OF LSE  1.91967633      16.0993277
                                 3.18746727E-03
                                 5.28216013
ESTIMATE OF ERROR = .152187945

SUMMARY OF DATA
1
2
3
4
5
6
7
8
9
10
11
TIME
.77
1.75
2.67
5.08
7.77
10.65
13.3
17.88
23.4
34.52
49.13
                                CONC
1
1
2
3
4.1
5
5
  25
  75
  05
 .25
 .85
7.1
8.1
8.9
9.65
FIT VALUE
1.12564889
1.64603691
2.10734213
3.20108643
4.24803791
5,
5,
19571448
93358056
                6.96110108
                7.87555468
                9.00645823
                9.69943839
RESIDUAL
.124351112
.103963087
-.0573421288
-.201086428
-.148037907
.0542855244
-.0835805573
.138898922
.224445321
-.106458228
-.0494383909
                                     344

-------
OXYGEN TRANSFER TESTING UNDER PROCESS  CONDITIONS

H. J. Campbell, Jr.
E. I. du Pont de Nemours & Co.,  Inc.
Wilmington, Delaware  19898

INTRODUCTION

     Adequate  test procedures must  be available to  permit  the  investigator
to  assess aeration  equipment performance  during  actual  process operation.
Several  test  methods have been  employed to  determine  oxygen transfer  under
field conditions.

     In  general,   the methods  can be  categorized  according  to  the  rate of
change of  dissolved  oxygen  (DO)  in a  given  reactor (or segment  of reactor).
Systems  in which  the rate  of  DO change  is  zero at  any  given  point are
referred to as  steady state  systems; the others are classified as non-steady
state systems.  In some  cases,  the  influent  wastewater may  be diverted  from
a  reactor being  tested.-  These are  referred  to as  batch tests.   The  term
"continuous test"  is  used  for those  cases where the influent wastewater  flow
is not diverted.

     Several  in situ  test  methods  do  not  require a  direct measure  of the
oxygen-uptake rate.   These have  been broadly categorized as the  mass balance
method,  the off-gas  method,  and the tracer  method.  The mass balance method
requires data on  the net change in  oxidation  level between  all  entering and
exiting  liquid  flows.  The off-gas method is simply a mass balance on oxygen
that includes both the  liquid and  gas  streams.  The tracer method indirectly
measures  the rate  of oxygen  transfer by  determining  the rate of transfer of
a radioactive tracer.

     The  most  commonly  used  methods  involve  the  direct measurement  of the
oxygen  uptake  rate,  R,  of  a  respiring biological  system.   Two  of   these
methods  are  carried  out with  little or  no  DO variation in  either  batch or
continuous flow systems;  these  are referred  to as  "Steady State  Batch Tests"
and  "Steady  State  Continuous Tests,"  respectively.    In  the  remaining two
tests, the DO  level  in the  reactor is adjusted at  the  beginning of the  test
to be either greater  than or  less  than the  steady state DO.   These tests are
referred  to  as  "Non-steady  State Batch Tests"  if the  influent wastewater
flow is  discontinued  for  the  test  and  "Non-steady State Continuous Tests" if
the influent wastewater flow is  continued during testing.
KEY FACTORS IN CONVENTIONAL TESTING

     The   successful   evaluation  of   aeration  equipment   in   situ  using
conventional  techniques   involves  careful  measurements   of  several  key
parameters  under  full-scale  operating  conditions.   A  discussion  of these
measurements and their importance  in field evaluations follows.
                                      345

-------
Dissolved  Oxygen

     An  accurate measure  of  the DO  concentration  in an  aeration volume  is
essential  to  any  evaluation  of aeration  equipment.   For  in situ  testing,
direct-reading DO probes  are  the only  practical  means of  measuring the  DO
concentration  in mixed  liquor  suspended  solids  (MLSS)  samples.   Following
proper  calibration  of  the  probes,  considerable  care  and  attention  are
required to  assure continuous  reliable  results under  field conditions.

     Aeration  testing may  be  carried out on an entire tank or  on  an  isolated
mixing  zone  within  a test volume.  Typically, a minimum of  three probes  are
used  in field testing; and,  depending  on  the type of  aeration system  being
tested,  placement  of  the  units  in the  test  tank  can  be  critical.    For
example, in  aeration volumes  with a relatively uniform DO concentration (eg,
well  mixed with respect  to DO), probes may  be  located in  the tank without
particular regard  to aerator mixing  pattern.  However,  for  aeration volumes
that  are  not well mixed with  respect to DO,  probes  should  be strategically
placed  around  the  flow  pattern  characteristic   of  the   aeration device.
Depending  on the specific application,  many  aeration devices will not  yield
a  uniform  DO concentration in  the  tank under test conditions.   Thus,   it  is
important  to establish a  complete DO profile on these aeration systems  prior
to testing.   The DO profile will  indicate  the actual mixing  characteristics
of the  test  tank and allow proper placement of the DO probes.

     During  field  testing of  aeration  systems,  the  DO  concentration should
not be  the limiting  factor in the biological reaction.   For instance, the  DO
can become limiting at approximately 0.5 mg/£  for  non-nitrifying  activated
sludge  systems,  and  at   approximately  1.0  to  1.5  mg/&   for  nitrifying
activated  sludge  systems.   Therefore, testing where  the minimum DO is  above
0.5 mg/£  and  1.5 mg/Ji  for non-nitrifying  and  nitrifying  activated sludge
systems, respectively, is  essential.

Oxygen Uptake  Rate

     A  significant  factor for evaluating aeration  equipment under operating
conditions  is an accurate measurement  of  the MLSS  oxygen  uptake  rate,  R.
Experience has shown that  accurate  measurement  of the  rapid oxygen uptakes
created by high organic  loads  is virtually impossible.   Ideally, the oxygen
uptake rate measurements should  be  taken jja situ or immediately at the  point
of sample  collection.  However,  as a practical matter,  a  finite  time period
elapses  prior  to  field  measurements  of this  parameter.   Since  the oxygen
uptake  rate  of  a  sample  will  vary  as the  available soluble  substrate  is
oxidized,  significant variations  in  the   uptake  rate may  be  observed  for
samples  taken  from moderately  to highly  loaded systems.   Thus,  caution  is
urged  where  oxygen  uptake  rates   are  being  measured under  actual   plant
loading situations.

     A  practical  approach  used to  minimize  the  variability  of  the oxygen
uptake rate  is testing under endogenous  respiration  conditions.  Endogenous
respiration  can be achieved  by diverting  the influent wastewater  to  other
aeration  tanks in  order  to  allow  the MLSS  to  assimilate  the  remaining
soluble  organics.    This   procedure  is  the  basis   for  batch  testing  and

                                    346

-------
establishes a  low,  relatively constant  oxygen uptake rate  that can be  more
accurately measured.

Alpha and Beta Factors

     Alpha  (a) and  beta  (3)  values  are  necessary  for describing  the  in-
fluence  of  mixed   liquor   characteristics   (eg,   dissolved   substances   and
suspended  solids)  on  the  transfer  capability of  the  aeration equipment  in
clean water.  Generally, a   and  3  measurements are made only when the  field
transfer  rates are  to  be  compared  to  standardized transfer rates developed
under clean water conditions.

     Due  to the  variability  of  influent wastewater quality,  mixed liquor
characteristics, and type  of aeration device,  the  a  level may be extremely
variable  for  a  given  application.   Testing  under  endogenous respiration
conditions  is  advantageous  because  it tends  to  minimize  the variability  of
 a  during  the test period.   However,  the  oxygen  transfer  rate   obtained
cannot  be  assumed  to  be  the  same  as  that  occurring  during  operating
conditions  under  normal  plant   loading.    A  more  detailed  review  of   the
complexities  associated  with  a   and  8   factors  has been  presented  by
Stenstrom and Gilbert (1).

Wastewater Temperature

     Wastewater  temperature  also  affects   the  evaluation  of  field oxygen
transfer  rates.    Both  oxygen  transfer  rate and  oxygen  update  rate   are
significantly  altered by  changes  in temperature.   Biological  reactions  are
thought  to  be governed by  a  doubling of  rate  for  every  10°C  increase  in
temperature  [corresponds  to  R^ =  R2Q  1.072^~20)]  within  the  practical
operating range of  10-30°C (2).   Thus, if an  oxygen  uptake rate is  measured
at a wastewater temperature  different from the actual operating temperature,
the appropriate temperature  correction for  reaction rate must be made during
data evaluation.

     At the same  time,  since aeration equipment  is typically rated   at 20°C,
but is  usually tested  at  some other  temperature,  appropriate transfer  rate
adjustments must  also be  applied  to account  for  the effects  of wastewater
temperature.   The   temperature  correction  for  oxygen  transfer  has   been
discussed in detail  by Stenstrom and  Gilbert  (1).   Where  practical,  testing
at temperatures near 20°C is  desirable,  since it  will  minimize temperature
corrections  and  allow  a  more accurate  comparison  with  transfer rates  at
standard conditions.
REVIEW OF TEST METHODS

Steady State Testing

Test Descriptions

     Steady state testing  involves  simultaneous  measurement of DO and oxygen
uptake rates in full-scale  aeration tanks.   Testing may be conducted with  or

                                     347

-------
without  influent wastewater  flow to  the aeration  tank.   For convenience,
these  testing   methods   are   referred   to  as  the  continuous  test   (with
wastewater flow) and the batch endogenous  test (without wastewater flow).

     The  continuous test  is  the classical  approach  of  determining   field
oxygen  transfer  in  an  aeration  tank  under normal  operating  conditions
(3-6).  Generally,  the  DO,  oxygen uptake rate, and temperature are monitored
in the tank  as  wastewater continues  to  flow  through  during a specified test
period.   The DO of  the  influent  and   effluent  streams  are  also  measured
throughout   the  test.    The  batch   endogenous  test   is   conducted  by
discontinuing the influent wastewater flow to the aeration tank prior to the
test  period.  The  recycle  sludge  flow  to   the  aeration  tank  may  be  held
constant  or  discontinued   during  the   test  period.    Operation   in  the
endogenous  phase should  yield  more  uniform  aeration tank  contents  (eg,  R,
DO, and a  value) and should  result in  increased  testing accuracy.

Data Evaluation

     The  data obtained from  the  steady  state  tests  may  be  evaluated   using
Equation  1 as described below:

                 System Test  Constant
          Iv, 3i .. ~
           Li
'f        C*  , - C
                        00
             f   °R                                           (1)
     The  system  test  constant  is defined as follows for batch and continuous
 testing:

     Batch Test

                    Qr
          B  = R 	—  (C  - C  )  (with recycle sludge flow)                ,  ,


          B  = R (without recycle  sludge  flow)                             ,  .


     Continuous  Test


                          'o-V                                          (4)
where:
     C*  _ = saturation DO concentration, mg/&  in mixed
            liquor

     CR    = DO  level,  mg/£,  that  most accurately  represents  the  driving
             force, C*mf - CR, in  the  transfer zone of  the  system.

     Cr    = DO level, mg/Jl, in  the  recycle  sludge flow,  Qr

     C'0   = DO level, mg/£, in  the  test volume influent  flow,  Q1
                                      348

-------
     R  =  average  oxygen uptake  rate,  mg/£/hr,  of  the  MLSS  in  the  test
           volume, V

     Qr    = recycle sludge flow, £/hr, to the test volume, V

     Q1 =  influent flow, £/hr, to the  test volume, V

     V  =  test volume, £

     The  resulting  KLaf  represents  the  field  oxygen  transfer  rate   for
the specified test temperature.

Test Limitations

     The limitations of  a specific test method  must  be  recognized to ensure
adequate data collection and  allow  an  appreciation for  the  accuracy of  the
end  result.   For   the   steady  state   testing  approaches,   the  following
discussion outlines the specific test limitations  that have been identified.

     Continuous testing - An  accurate determination  of the value of R may be
very difficult  to obtain  if  system  influent wastewater  flow  and substrate
concentration are not nearly constant  during  the test  period  (6,  7).   In
addition,  if  the  system being  tested has  a  high  organic  loading,  the value
of  R  will  be practically  impossible  to determine  accurately.   Thus,   the
accuracy  of  the  Kj^aj  determination  is  directly tied  to  the ability to
determine  a  representative  value  of  R  for  the   system.   McKinney   and
Stukenberg  (8) found  where  R values were  greater  than  60  mg/£/hr  errors in
measurements can be significant and increase with  increasing R values.

     The inability  to control  the  a   value  of the  test volume  during   the
test period may  be  a serious drawback  to continuous  testing.   The a  level
may vary substantially  (as  much as +20  percent) at a given  location within
the  tank  during the  test   duration  due  to variable   influent  wastewater
characteristics.   In  addition,  the  a  level  may  even vary spatially within
the  tank  at  a  given  time   during  the  test  depending  on  the   degree of
treatment and system mixing patterns.

     For many applications, specifically  for  surface aeration installations,
the DO concentration may vary  spatially within  the tank at  a  given  time
during   testing   under  continuous   loading   conditions.    Therefore,    the
continuous  testing  approach  may  not be  valid  for  those  applications where
such variation exists.   The  requirement of maintaining  the excess DO level,
CR,  within  a  practical  range  during  the  test  period  may  also  be  a
limitation  for  some test applications.  In  order to maintain  a reasonable
range  for  accurate  DO  determination,  the  measured  C^ level   should be
between  a  minimum  of  2  mg/£  and  a maximum of  75  percent  of  the system
C*oo £.   Limiting  the  test conditions  to  within this  range will  ensure  the
statistical validity of the DO  measurement and  provide a more meaningful  end
result.

     The  inability  to  determine   a   C*mf   value  for  submerged  aeration
systems effectively  limits  the use  of  the continuous test.   This technique

                                     349

-------
may  be used  for  submerged  aeration  systems  only where  it  is  possible  to
perform  the  test  at  a  minimum  of  three  different  oxygen  uptake  rates.
Different  oxygen  uptake  rates  may  be  established  by  varying   the  MLSS
concentration under a  constant  organic loading  rate or  by varying the organic
loading rate for a constant MLSS  concentration.
     Batch  testing -  To  achieve an  endogenous  respiration  condition,   the
influent  flow to  the  test volume  must be  discontinued.  For  systems  with
limited operational flexibility,  this  approach  may not  be pracitcal.

     When  testing  under  endogenous repiration conditions, the a  level would
not  necessarily   reflect   the  a    level  at   normal   operating  conditions.
Therefore,  if a  field transfer rate  under  operating  conditions  is  desired,
the  batch  test   may  not yield  an   accurate estimation  of  that   value.
Depending  on  the  specific application, testing under  endogenous  conditions
would  be  expected to yield  higher field  transfer rates  due to  higher a
levels.  On  the other hand,  if the test is  performed  for  comparison with a
standard  oxygen transfer  rate,  this  approach  should  provide representative
information for  that comparison.

     The  requirement  of  maintaining   the   excess  DO  level,   CR,  within  a
practical  range during the test period may prove  to be a limitation  for  some
applications  of  steady  state  batch   testing  as   with  continuous   testing.
Also,  the  inability  to  determine  a  C*oof value   for  submerged  aeration
systems  effectively further  restricts  the  use of  the batch  test   approach
(see discussion  in previous subsection on continuous test limitations).

Non-Steady State Testing

Test Descriptions

     Non-steady   state  testing  involves  reduction  or  discontinuation of
aeration  in  the  test tank.   This allows  biological  action,  through  the
oxygen  uptake  of  the  MLSS,  to  reduce  the DO  concentration  in  the  test
volume.   After  the  DO  level  has  been depressed  sufficiently, aeration is
increased  or  reintroduced  to the  test tank.  By  simultaneously  monitoring
the  increase  in   DO  level and  the  oxygen   uptake  rate,  the  field  oxygen
transfer  rate  may be determined  using  a   classical   reaeration  approach.
Testing  may  be  conducted  with  or  without   influent   flow  to  the   aeration
tank.  As  with steady state  testing,   these  two methods are  referred to as
the  continuous test  (with wastewater flow) and  the  batch  endogenous  test
(without wastewater flow).

     A  batch  endogenous  desorption   test   that   is   very  similar   to   the
non-steady  state  batch  endogenous  test has  been proposed  by Kayser  (9).
Rather than starting the  test at  a very low  DO concentration, the desorption
test starts at  a  supersaturated DO level in the  test solution.  The  premise
of the test is that the rate  of oxygen desorption  is  the same as the  rate of
oxygen transfer to the water.   The initial high DO concentration is  achieved
by  a  slug  addition  of  hydrogen  peroxide  uniformly  distributed  over   the
entire   test   volume.    Following   the  initial   peroxide   addition,   by
simultaneously monitoring  the decrease in  the DO level and the oxygen uptake


                                     350

-------
rate at the same  test  tank locations,  the field  oxygen  transfer rate may be
determined using a classical deaeration approach.

Data Evaluation

     The data obtained  from  the  non-steady  state test may be evaluated using
Equation 5 as described below:

    (C_ - C) = (C  - C  ) exp- [(K,a^ + System Time Constant)t]            (5)
      K          K    O          Li £

    where :
    C     = DO concentration, mg/£ , at any time,  t

    CR    = DO  level,   mg/£,  that  most  accurately  represents   the  driving
            force, C*ocf - CR, in the transfer zone of the system

    C0    = DO concentration at  test time, t = 0, mg/£

    System Time Constant is:

    Batch Test


         r (with recycle sludge  flow)
        V

        0  (without recycle sludge flow)


    Continuous Test
    Equation 5 is the same general  form  as  the equation used for clean water
evaluations.   Therefore,  the  same  evaluation  techniques  are  applicable,
except  that  for respiring  systems  the  term  (Ki,af  + System  Time  Constant)
is obtained where KLa would be determined for  clean water testing.

    Once  Kj^af  has  been  determined,   the  corresponding  effective   field  DO
saturation  concentration, C*oof,  can  be calculated.   This  calculation  may
be performed utilizing Equation 6 as  follows:

                     R +  (System Time Constant)(CR - C' )

         °*-f = C" +               \^                ~               (6)

where :

    C'0   = DO level, mg/& ,  in the  combined wastewater  influent  and  sludge
            recycle  flow, Q1 ,  for  continuous  testing and  DO level  in the
            sludge recycle flow, Qr,  for batch testing

                                     351

-------
Test Limitations

     For  the  non-steady state  testing  approaches,  the  following discussion
outlines  the specific test limitations that have been identified.

     Continuous  testing - An accurate  determination of  the  R value  may be
very difficult  to obtain  if  influent flow and  substrate concentration vary
substantially during  the test period.  If either  of these conditions exist,
the  reaeration  curve may  be erratic  and  the  calculated C'-^bof  value,  which
is  calculated  using  R,  may be  in error.   As  previously mentioned,  a high
organic  loading  to the  test  system (R >  60  mg/£/hr) will also  preclude an
accurate  measurement  of  R.

     Due  to the  flow-through  testing  conditions,  the influent of non-uniform
influent  wastewater  flow  and  substrate  concentration  will   cause  variable
system  R and   a   values  over   the  test   period.    In  addition,  under
flow-through conditions, the  maldistribution  of  the  influent  flow throughout
the  tank  volume  can bring  about spatial  variation of R  and a  values within
the  test volume.    Therefore,   testing   should  be  conducted  during  the
operational periods  when fluctuations in influent flow  and  organic strength
are minimal.

     For  many applications,  specifically for  surface aeration installations,
the  DO concentration  may  vary  spatially  within  the tank at a  given time
during  testing  under  continuous  loading  conditions.   Therefore,  this test
approach  may not be valid  for those applications where such variation exists.

     For  systems with  a low oxygen transfer rate  and  a  high  organic loading
rate,  the system equilibrium  DO  level,  CR,  may  be  relatively  low.   To
obtain  sufficient   test  information  for  statistical   data  analysis,  a
reasonable range  for the  test  parameter  (Cjj -  C) is essential.   Since the
practical  initial  DO  level  will  be  approximately  0.5   mg/&,  a  minimum CR
level of  4.0 mg/£  is  recommended for  reliable evaluation of test data.  This
provision may prohibit continuous  testing at some  installations.

     For  surface  aeration  installations,  the only method of  reducing the DO
concentration sufficiently may  be to  completely shut off the aerators or to
operate  them  intermittently.   This   action   would   most  likely  result  in
inadequate mixing  of  the  test  volume,  bringing  about  solids   segregation
during the deaeration phase of  the testing.   As the  units  are turned on the
the reaeration phase of  the testing,  erratic  results will be produced  during
the  initial   data   collection period  due  to  the  re-establishment  of  the
uniform mixing regime.   Depending on  the size and geometric configuration of
the  test  volume,  the  effect  of  inadequate  mixing  could lead  to  erroneous
test results.

     Batch testing -  This method has  the   same  basic  limitations  as  the
steady   state   batch   approach.    Additionally,   for  surface   aeration
installations, the only  method  of reducing the DO concentration  sufficient^y
may be to completely shut off  the aerators  or  operate   them  intermittently.
This action would  result in  inadequate mixing  of the test volume and cause
erratic results as described  for  non-steady state  continuous  testing.

                                     352

-------
      Batch  desorption testing  -  This method has  the same basic  limitations
 as  the  conventional non-steady state batch method.   In  addition,  this  method
 is  based upon the  assumption  that the  hydrogen  peroxide  immediately  and
 fully  disassociates   to   DO   and  water   when   added  to  the   test  volume.
 Furthermore,  this  approach is  based on  the premise  that  the rate of  oxygen
 desorption  from  water-to-air  is  the  same  rate  as  oxygen  transfer  from
 air-to-water.   Kayser's  work  (10)  seems  to  verify  both of  these assump-
 tions.   However,   if   the  peroxide  does  not  fully  disassociate  before  the
 start of  the  test, the resulting  oxygen  transfer  rate determination will  be
 erroneously lower  than the actual field transfer rate.

 Other Test Approaches

      In  addition   to  the  above  tests,   other  procedures  have   also  been
 employed  to  evaluate  aeration equipment  operating  in  biological treatment
 systems.  For the  most part,  these procedures have not been  extensively  used
 and,  therefore, are   not  primary  tests  for  respiring   system  field  tests.
 However,  these tests may have application  for specific treatment operations.

 Mass  Balance:   Activated Sludge Systems

      The  mass balance  approach   has  been  proposed   to  determine   oxygen
 transfer  in  operational   activated  sludge systems.   Total  oxygen balances
 must  be made  on the   influent, effluent,  and  waste  activated  sludge  flows.
 The change  in total oxygen across  the entire  activated  sludge system  equals
 the  oxygen  transferred by the aeration  system.  Total  oxygen measurements
 should  be  based   on  COD, with  correction  for  nitrification.    The  major
 problems  with the  mass balance technique  lie  in satisfactory measurement  of
 waste activated sludge volumes and  in  obtaining  representative  samples  of
 waste activated sludge for analysis.  The high suspended  solids   concentra-
 tions in waste activated  sludge make  it difficult to  obtain  valid  data
 because  of  large   errors   that  can result  from  minor variations  in  the
 aeration  volume solids inventory.  While  is is  possible to achieve accurate
 mass  balance  measurements  on small  laboratory  systems,  this  technique has
 limited practical  value in field-scale evaluation  of aeration  equipment for
 activated sludge systems.

 Mass  Balance:   Aerated Stabilization Basin  Systems

      The  mass  balance  procedure   has  been  applied  to  low-rate  aerated
 stabilization  basin (ASB)  systems because  oxygen  transfer  cannot  be  readily
measured using more  direct methods.  The accuracy  of this  method  depends  on
 the extent  to  which  all   factors  that  supply and  withdraw oxygen  from the
 system  are measured.   A number of these  factors can  be  minimized  during the
 test  period   so that  the  main  factor  that  determiaes  the  aerator   oxygen
 transfer  is  the reduction in  BOD  through the  respiring system.   The mass
balance method  is  simple and  straightforward,  especially when the  ASB  system
is  aerator  (oxygen)   limited  and  where   significant settleable  biological
solids are not produced (11).
                                     353

-------
Off-Gas Analysis

     Off-gas  analysis  has  been  used  with  full-scale,   diffused
activated  sludge  systems  and  covered  tank,  pure  oxygen  activated
systems  (12,  13).  A  mass  balance on  oxygen  in the gas  phase is require  ,
therefore,  this  method is  not applicable to surface aeration equipment.   For
systems  that  are not  enclosed,  a collector or hood  is  placed  in the zone o
interest  in the aeration  tank  so a  representative sample  can be collected
and  analyzed  for residual oxygen.  As  with the other  methods, the  validity
of  this  technique depends  on   the  accuracy  and  precision  of  the  various
parameters  that  have  to be measured.    For  low efficiency  of oxygen  transfer
(3  to  8 percent)  associated  with many  diffused  aeration  systems, obtaining
the  requisite accuracy and  precision of the  mass  balance on  the  input  and
discharge  gases  is difficult under the best of conditions.   However,  it is
recognized  that  significant advances  have  been  made in  the  recent  past
regarding   instruments  and  techniques   for  measuring   gas-phase   oxygen
concentration.    Furthermore,   aeration   systems   with    oxygen    transfer
efficiencies  of  about  10  percent or greater are becoming  more the rule  than
the  exception.    Therefore,  the  combined  effect  of  these  developments
suggests  the  desirability of additional  investigation  into the utility  and
merit of this method.

Tracer Method

     A  radioactive tracer  technique   has  been  proposed  to  measure  oxygen
transfer rate in  any aeration system,  either iti clean water or  in wastewater
(14).  The  tracer  method  requires considerable planning,  special radioactive
counting  equipment,  and a  radioactive materials  license  to  use  the  radio-
isotopes,  krypton-85,   and  tritium.   The  basic concept  of  the radioactive
tracer   technique  involves   direct  measurement  of  the  mass  transfer  of
krypton-85, which  is   related to  the  oxygen  transfer  rate.   The  tritium is
used  to  measure the dispersion  of the  tracers  in  the aeration tank.   Both
tracers  are added to  the  aeration tank  at  a  single point.   As the tracers
are  dispersed  in the  aeration tank,  the  conservative  tracer,   tritium,  mixes
with the mixed  liquor  while  the  krypton-85 is mixed and  stripped off in the
gas  phase.  The  key  to this  procedure  lies in the  fact  that  the krypton-85
stripping  rate   from  the mixed  liquor  is  directly  related  to  the oxygen
transfer  rate  from the  gas   into  the mixed liquor.   In  effect,  the tracer
method  is  an indirect  method for measuring oxygen  transfer.  A  series  of
grab  samples  are  collected  and  counted in  a scintillation  counter.   Care
must be  taken  to insure that gas  bubbles to not form  in  the  samples as the
krypton-85  would  come   to  equilibrium with  the gas  bubbles and  produce an
error  in  the radioactive  counts.   Accurate  counting  of  the  samples  is
essential for obtaining good results.   The counting efficiency is  about 30
percent  for  tritium  and   about  90   percent   for   krypton-85.   The  tracer
technique has had  limited application to date but has  considerable potential
in the oxygen transfer  measurement  field.   More studies are urged using this
technique.
                                     354

-------
FIELD TEST RESULTS

Test Conditions

    Since both  the  off-gas analysis and  tracer method have  had limited use
but  show  excellent potential,  a field  study was undertaken utilizing both
techniques in parallel.  The  purpose of the  test was  to  determine the  field
oxygen  transfer rate  using  each  test  method  in  a  full  scale  biological
treatment system and compare results.

    Figure  1  illustrates  the test  setup in  a 23M  (75   ft)  diameter,  6.1M
(20 ft)  liquid  level  tank  with a  multiple  inlet  wastewater  distribution
system.   The  aeration  system was   a  course  bubble,  bottom  cover  diffuser
system  as  described   in   Figure   1.   The   testing   was  performed   under
flow-through conditions.   The  wastewater  flow was 22  Jl/sec  (350 gpm)  with a
sludge recycle  flow of 25 £/sec  (400 gpm).

    The  operating  conditions that  prevailed  during  the -test  period,  as
summarized in Table 1,  indicate  a highly  loaded actuated  sludge system.  The
2.5 x 10" H   (0.66MG)  tank had  sufficient mixing for solids suspension but
the aeration rate was  inadequate for maintaining a positive  DO  level  in the
6100  mg/ £  MLSS  at  an operating   temperature  of  21.8°C.   Since  neither
technique to be used  in the  study  rely on a positive  DO  level, the lack of
DO in the test  tank had no impact on the test results.

    Submersible  pumps,  which  were   installed at  four different  locations,
were used to  collect radiotracer samples at  a  tank  depth of 4.2M  (14 ft).
DO probes were  installed at  three of the  sample pump  locations  to monitor DO
throughout the  test.  The  radiotracer  release point  was approximately in the
center of the tank.

    The test was  initiated by breaking four  separate  bottles containing the
radiotracer mixture (tritiated water and  krypton-85).   The four bottles were
wired together  in  a cluster and contents  were released at  the  same instant
using electrical explosive caps  wired  to  a  common battery.   [Note :   Due to
unforseen problems  with  normal tracer procurement, a  make-shift tracer unit
(four  bottles)   was   used   versus   a  single  bottle  under  normal  test
conditions.]   Samples   were   then  collected  from each  of  the  four  sample
locations at  one minute  intervals   for a period  of   fifty  minutes.   During
collection,  the samples  were preserved and  later were packed  in  a constant
temperature bath for shipment to the laboratory  for analysis.
    Immediately   prior   to  the  radiotracer   test,   a   1.6M^   (17.5
off-gas  hood had  traversed  the  left  half of  the  aeration  tank.   Twelve
separate  locations  were utilized to  approximate representative  coverage of
the aeration tank (the  left  and right halves of the  tank  were  assumed to be
identibal).  At each location,  the hood  was used to collect the tank  off-gas
for  analysis  of  its   components.   The  percent  oxygen  transfer could be
calculated by comparing  the  hood  collected  oxygen concentration with  that in
background air.
                                     355

-------
                                           Effluent
                                                     D.O. Probe No.  1

                                                           •
                                                    Sample Pump No.  4
         Diameter = 23M
           (75 ft)
         Side wall depth
           (24 ft)
         Liquid level * 6.1M
           (20 ft)
         Wastewater inlet elevation
           2.9M (9.5 ft)
         Sample pump and D.O.  probe
           elevation @ 4.2M (14.0  ft)
Influent
                                                                       Aeration System
                                                                           Details
A coarse bubble
diffuser system
consisting  of  205
stations Ltwo
13MM (1/2 inch)
diameter holes].
The stations are
spaced 1.2M (4 ft)
center to center
on laterals which
are on 1.4M (4.5
ft) centers.
Figure  1.   Test  setup for  the  radiotracer  and off-gas oxygen transfer tests.
                                            356

-------
TABLE 1.  OPERATING CONDITIONS FOR RADIOTRACER AND
          OFF-GAS OXYGEN TRANSFER TESTS
Test Parameter
                                       Level
Aeration volume

Wastewater flow

Recycle flow

Aeration stations

Air flow rate

Water temperature

MLSS

TDS

()„ uptake

D.O.
                               2.5 x 106 £(0.66MG)

                               22 £/sec (350 gpm)

                               25 £/sec (400 gm)

                               205 (7.1 H/sec/sta)

                               1450 £/sec (3075 scfm)

                               21.8°C

                               6100 rag/fc

                               5000 mg/fc (3 = 0.94)

                               198 mg/£/hr

                               0.0 mg/£
                     357

-------
Radio Tracer Results

     Each  sample  from  the  four  sample   locations  were  analyzed   in
laboratory  for  tritium  and  kryton-85.   Typical  counting  efficien
this  study were about  25  percent  for  tritium and  about  a/ pe
krypton-85.

     As  previously discussed,  this study  deviated  from  the  normal  tracer
measurement procedure  of  having  both  the  tritium and  krypton-8!> gas released
from  a  single  homogeneous  bottle.   Due  to  this  deviation,  the  basic
assumption  of  identical  concurrent  mixing for  both  tracers  is invalidated
and only  krypton:tritium ratios  measured after  complete mixing was achieved
were useful for  calculating the  krypton-85 transfer  rate.   This complete mix
criterion  gives  approximately thirty minutes  of  useful  gas  transfer data tor
Stations  I, 2  and  3  and  about fifteen minutes of useful data for  Station 4.
A   least  squares   analysis   of   the  logs   of  useable   krypton:tritium
concentration  ratio  data vs  time was  performed for each  sampling station.
The  log-slope  from  each  plot  is  the  krypton  transfer  rate coefficient,
a Kkr,   which    is   converted  to   the  equivalent   a KLa   value   by   the
relationship (Kkr  : KLa) = 0.83  (14).

     The   computed    oxygen   transfer   rate   coefficients,   along   with
corresponding  statistical  fit parameters  are shown in Table  2.  The  average
field  oxygen   transfer rate,  aKlaT,  was  determined  to be  4.85  Hrs    with
a range from 4.72  to 5.00 Hrs"1 for the four sample locations.

Off-Gas Analysis Results

     The hood  data from  the twelve aeration  tank  locations  were analyzed to
determine  the  percent  oxygen  transfer at  each location.  These results were
then  averaged,  based  upon the  weight of  the   unit  gas rate  (£/sec/M2)  to
obtain  the overall average tank  transfer  efficiency of 7.7 percent.  Since
the extrapolated air  flow collected by the off-gas hood was slightly higher
than that  measured going into the  aeration  tank,  a range  of  transfer rates
was calculated.   The  air collected  by the hood averaged  3.8  ^/sec/M^ vs
3.5    £/sec/M^ measured  (by an orifice  plate)  coming into  the tank.  Using
this air  flow  range  and  the  overall  transfer efficiency determined   for the
tank (7.5 percent), the a K^af level was determined using Equation  7 below:

                                (OTE)(C.F.)(QA)

                       aKLaT=  
-------
TABLE 2. OXYGEN TRANSFER TEST RESULTS FOR THE
              RADIOTRACER METHOD

Station3
1
2
3
4
Average
Standard
Deviation

aKLa @ T = 21.8°C
4.86 hrs'1
5.00 hrs"1
4.72 hrs"1
4.81 hrs"1
4.85 hrs-1
0.12 hrs-1
Coefficient of
Variation, r^
0.997
0.989
0.987
0.950
0.981
0.021
 aNote:   Stations  are  the  same  as  sample  pump  locations.
                     359

-------
     QA       =  Airflow Rate  Per  Unit Area  (£/sec/M2)

     BC* T    =  Field  DO Saturation  Concentration  (mg/£)
              (Note:   C*ooT  =  10.0 mg/£  determined  in  previous  clean  water
              tests;  g = 0.94  from  Table 1.)

     CR       =  Residual DO Concentration = 0 mg/£

     D        =  Depth  = 6.1M

     Table  3 shows the range  of field oxygen  transfer  rates, a  KLaT,  to  be
 4.75-5.09  Hrs-1  for  the off-gas  analysis.   Comparing  this  range  with  that
 from the radiotracer test  (4.72-5.00 Hrs"1),  there is  excellent  agreement
 between   these  two    test  methods.   Thus,   for   this  submerged   aeration
 application both  of  these testing  methods appear to be  equally appropriate.
 CONCLUSIONS

     Aeration  testing with respiring  activated  sludge systems  is  not  easily
 carried  out and  is  not  recommended  over clean  water aeration  testing  for
 verifying  aeration performance  specifications.   However, with  careful  data
 collection  and  evaluation,  it  is possible to obtain reasonably valid  results
 under  process  conditions.   After  review of  current  testing  experience  and
 the  inherent errors of each  test, the following conclusions are appropriate:

 o  Batch  endogenous  testing procedures  are more  accurate  than   continuous
   testing  techniques   for  estimating   field   transfer   rates  of   various
   aeration equipment.

 o  The main  limitation  of continuous testing is  the inability to  accurately
   measure the biological oxygen uptake rate.

 o  For many  applications, specifically  for  surface  aeration  installations,
   the DO  concentration  may  vary spatially  within  the tank  at  a   given  time
   during   testing   under   continuous   loading   conditions;    therefore,
   conventional  testing  approaches  may  not be valid for  these applications.
   Further study is required for such installations.

 o  For surface  aeration  equipment,  steady state  batch endogenous   testing  is
   preferable to non-steady  state batch endogenous  testing.

 o  For  submerged  aeration  equipment,   non-steady  state  batch   endogenous
   testing is preferable  to  steady state batch endogenous testing.

o  Batch  endogenous   desorption  testing  using  hydrogen  peroxide  has  been
   demonstrated  as   an   effective   technique   for  measuring  field   oxygen
   transfer in surface and submerged aeration systems.

o  The mass  balance  approach  for  determining  field  oxygen  transfer  rates
   requires extensive data  collection.   For  activated  sludge  systems,  the
   approach appears to have  value  in small,  well-mixed systems but is simply


                                     360

-------
   TABLE 3.  OXYGEN TRANSFER TEST RESULTS FOR THE OFF-GAS
             ANALYSIS TECHNIQUE
  Data
  Description
Mean3
OTE
Unit Air
Flowrate
                           a
  Hood Collected
  Air Flow Rate
7.-
3.8   1/sec/M2   5.09 hrs~L
  Measured
  Influent Air
  Flow Rate
7.'
3.5   l/sec/M2   4.75 hrs'l
aMean OTE:  Average oxygen transfer efficiency (%) under
 actual field operating conditions (see Table 1).
         Average oxygen transfer rate parameter (hrs~)
 under actual field operating conditions (see Table 1).
                             361

-------
   not  as  accurate as other tests  for  most  systems.  For  ASB  sys em ,
   approach  may   be   the   only  practical   method  available  for  testing.
   Therefore, with  reasonable  care  in  data  collection and analysis, tne mass
   balance   approach   can   be   useful   for  evaluating   aeration  equipment
   performance.

o  The  off-gas  analysis  method for  field  oxygen transfer measurement has had
   limited  application to  date.  One  current  application  of  this technique
   is with  covered tank, pure  oxygen activated  sludge  systems.   However, in
   view   of  advancing  technology  in  oxygen   concentration   measurement
   equipment,   and  the  progressive   increase   in  efficiency   of  aeration
   systems      generally,   this   method    deserves   further    evaluation.
   Preliminary  findings  using  this approach  indicates  high potential  for
   this method  in bottom coverage diffused aeration applications.

o  The  tracer  method   for  field  oxygen   transfer  determination  remains
   relatively  untested.    Preliminary   findings  using  this  technique  are
   encouraging.  The approach requires  and deserves further evaluation.
REFERENCES

 1.   Stenstrom,  M.  K. , and  Gilbert,  R.  G. , "Alpha,  Beta,  and Theta  Factors
     for  Oxygen  Transfer  Testing."   Presented  at  the  1980 Environmental
     Engineering  Specialty Conference, ASCE, New York,  New York, pp.  86-93,
     (July,  1980).

 2.   McKinney,  R.  E.,  "Mathematics  of  Complete Mixing  Activated  Sludge."
     Journal of  the  Sanitary Engineering  Division, ASCE,  88 (SA3):87-113,
     (May,  1962).

 3.   Eckenfelder,  Jr.,  W. W. , "Factors Affecting  the Aeration Efficiency  of
     Sewage  and  Industrial Wastes,"  Sewage  and  Industrial Wastes,  31:60-70,
     (January, 1977).

 4.   Nogaj,  R.  J.,  and Hurwitz, E.,  "Determination of  Aeration Efficiency
     Under  Process Conditions."   Proceedings of  the  18th  Industrial  Waste
     Conference,  Purdue University, pp. 674-683  (April  1963).

 5.   Kayser,   R. ,   "Comparison   of   Aeration   Efficiency   Under    Process
     Conditions."   Proceedings  of the 4th  International  Conference on  Water
     Pollution Research,  Prague,  pp.  477-496  (1969).

 6.   Kalinske, A. A.,  "Problems  Encountered in Steady State Field Testing  of
     Aerators and Aeration Systems,"  Proceedings,  Workshop Toward  an  Oxygen
     Transfer Standard, EPA-600/9-78-021,  pp.  205-209 (April  1979).

 7.   Stukenberg,  J.  R., V.  N.  Wahbeh, and  R.  E.  McKinney,   "Experiences  in
     Evaluating and  Specifying Aeration Equipment,"  Journal  WPCF,  49:66-82,
     (January, 1977).


                                     362

-------
 8.  McKinney,  R.  E.  and  Stukenberg,  J.  R.,   "On-Site  Evaluation:  Steady
     State vs.  Non-Steady  State  Testing,"  Proceedings,  Workshop Toward  an
     Oxygen Transfer Standard, EPA-600 9-78-021, pp. 195-204 (April 1979).

 9.  Kayser,  R.,  "Measurements  of Oxygen  Transfer  in Clean Water  and Under
     Process Conditions," EIPRE Conference on Aeration, Amsterdam (1978).

10.  Kayser,  R., "Testing Aeration Performance  of  the  Treatment  Plant of the
     City of Nienburg," Unpublished report in German (1970).

11.  McKeown, J. J.,  and Buckley, D. B.,  "Mixing  Characteristics  of  Aerated
     Stabilization Basins," TAPPI, 54:1664-1672 (October 1971).

12.  Mueller, J. A.,  Famularo,  J.,  and Mulligan, T. J.,  "Oxygen Transfer  in
     Closed  Systems,   Proceedings,   Workshop  Toward  an   Oxygen  Transfer
     Standard," EPA-600/9-78-021,  pp. 180-194 (April 1979).

13.  Downing, A.  L.   and  Boon,  A.  G.,  "Oxygen  Transfer  in  the  Activated-
     Sludge Process,"  In:   Advances in  Biological  Waste Treatment,  Ed.  by
     W.  W.  Eckenfelder, Jr.  and B. J. McCabe, Pergamon Press,  New York City,
     (1963).

14.  Neal,  L.   A.,  "Use  of  Tracers  for  Evaluation  of  Oxygen  Transfer,"
     Proceedings,    Workshop    Toward    an   Oxygen   Transfer    Standard,
     EPA-600-9/78-021, p. 210-227  (April  1979).
DISCLAIMER

     The work summarized herein was previously reviewed and approved for
publication by the U.S. Environmental Protection Agency (EPA-600/2-83-002).
                                      363

-------
AERATOR PERFORMANCE EVALUATION - THE ONTARIO PERSPECTIVE

Stephen A. Black, Supervisor
Pollution Control Branch
Ontario Ministry of the Environment
135  St. Clair Ave. W.
Toronto, Ontario  M4V IPS

ABSTRACT

     Aeration equipment employed in the activated sludge process is usually
the  single largest consumer of energy in a wastewater treatment plant.
Consequently, the selection of efficient aerators is an essential aspect of
treatment plant design.  Although oxygen transfer efficiency is only one of
several major selection criteria, it has a foremost impact on the long-term
energy requirements for plant operations.

     This paper discusses acceptance and performance testing procedures
adopted by the Ontario Ministry of the Environment in the specification and
selection of aeration equipment for activated sludge wastewater treatment
facilities in Ontario.
 INTRODUCTION

     The requirement of effective aeration devices to maintain an aerobic
 environment in the activated sludge process is universally recognized.  Such
 an aerobic environment not only implies a sufficient dissolved oxygen con-
 centration, but also adequate contact between wastewater, microorganisms and
 the dissolved oxygen.  Usually, the aeration device must also provide for
 this contact.

     Although there are many factors which enter into the selection of the
 aeration device for a particular project, the primary measure of effective-
 ness of an aeration device is its energy efficiency.

     As with other energy intensive industries, energy conserving designs
 and operations are receiving increased emphasis in the wastewater treatment
 field.  Aeration equipment employed in the activated sludge process is
 usually the single largest consumer of energy in a wastewater treatment
 plant, normally accounting for 60-80 percent of the total power demand.
 Effective aeration, therefore, implies maximum oxygen input with a minimum
 horsepower requirement.  Oxygen input is measured in terms of oxygen  transfer
 efficiency.

     Few, if any, unit operations in the wastewater treatment field receive
the design effort, intensity of performance testing or general operational
attention reserved for aeration devices.  The great variability in aeration
type and design, and the large number of manufacturers, ensures a keen
competition.
                                     364

-------
     Thus, aeration testing methods and results are scrutinized and
critically reviewed to an uncommonly high degree by all concerned;  the
manufacturer, the design engineer, the prospective owner and in many cases,
the regulatory agencies as well as the competition.


ACCEPTANCE OF AERATION DEVICES FOR USE IN ONTARIO

     Most aeration devices employed in municipal wastewater treatment
facilities in Ontario receive prior Ministry of the Environment (MOE)
acceptance.  Such acceptance, although not essential for marketing purposes,
is generally looked upon by consulting engineers in Ontario, and indeed
across Canada, as an informal accreditation of the equipment.

     Therefore, MOE acceptance is generally sought by the manufacturer/
supplier when a new aeration device is presented for consideration in the
Canadian market.  In such cases, manufacturer's claims of the equipment are
initially reviewed.  If the claims are considered reasonable and in line with
similar known equipment, general acceptance for use in Ontario is given.
Such general acceptance is not to be construed as final approval as each
project must receive MOE approval on its own merits for the particular condi-
tions applying.

     If the manufacturer's claims are questionable or exceed generally
accepted standards, or if the aeration device works on an unfamiliar princi-
ple, performance testing may be requested.  Such testing may be conducted
outside the province, provided that an accepted testing procedure is follow-
ed and adequately documented, or the testing can be conducted under MOE
supervision at the Ontario Experimental Facility operated by the Wastewater
Treatment Section, MOE.  The latter situation not only provides for MOE certi-
fication of the test results, but also provides an opportunity for the manu-
facturer/supplier to introduce both himself and his equipment to local con-
sultants and MOE personnel.

     Should the MOE reject the acceptance of the device, it may still be
marketed in Ontario for private and municipal projects.  However, as
mentioned above, in general, consultants in Ontario and elsewhere in Canada
look to MOE acceptance as a prerequisite in aeration device selection.


SELECTION OF AERATION DEVICES IN ONTARIO

     Generally, the selection of aeration equipment in Ontario is the
responsibility of the design consulting firm for the project.  To the
consulting engineer, the selection of aeration equipment involves the
development of a total aeration system capable of serving a variety of
needs within the wastewater treatment plant.  Other air requirements such
as for pre-aeration, aerated grit removal, channel aeration, dissolved air
flotation, etc., must be considered, and the entire biological treatment
system design will definitely influence aeration equipment type'selection.
The most efficient and economical transfer of dissolved oxygen is only one
of several major selection criteria.  Some of the other criteria specific to
the aeration device include capital and installation cost, aeration capacity,

                                     365

-------
equipment reliability and mixing performance.  Temperature loss or addition
due to the aeration may also be a factor in northern climates as well as
aerosol minimization where facilities are covered.

     The aerator is usually the only source of mixing within the aeration
tanks.  It must provide sufficient velocity within the tanks to maintain all
biological floe in suspension and thorough mixing to provide a uniform and
rapid contact between the wastewater, the introduced air (oxygen) and the
microorganisms.  Such mixing must at the same time, keep biological shear
to a practical minimum.

     In some MOE managed projects, the aeration equipment may be preselected.
Preselection may be requested for a variety of reasons including cost-
savings, time-saving, assurance of quality equipment and possibility of pre-
installation performance testing.

     The MOE has been involved in equipment selection for MOE managed projects
for many years.  As the owner and operating authority of a facility, it is
felt essential to select quality equipment from reliable suppliers in order
to minimize subsequent operating problems.  While the consulting engineer's
recommendation is largely relied upon, the MOE reserves the right to make
the final decision in the selection of major pieces of equipment.

     The MOE maintains an information list of prequalified, acceptable
suppliers and manufacturers of major equipment including aeration devices.
As only prequalified suppliers are invited to submit quotations for an item
of equipment that is to be preselected, other suppliers and manufacturers
are continuously seeking general acceptance by the Ministry of various
equipments not yet ruled as acceptable for MOE capital works.  Consequently,
procedures such as the oxygen transfer testing procedure have been establish-
ed to provide uniformity in evaluating classes of equipment.

     In cases of preselection, the consulting engineer prepares the necessary
specifications for MOE review.  The consultant then invites quotations from
prequalified suppliers, evaluates the quotes and makes appropriate recommend-
ations to the MOE, who makes the final decision on selection.

     The equipment may then be pre-ordered by purchase order directly by the
MOE if long deliveries or price increases are anticipated, arranged for
through a committing letter of intent to permit manufacturing to begin, or
the preselected equipment may be ordered by the general contractor following
successful tendering.

     Through the preselection process, certain major pieces of equipment are
selected in such a way that public funds are spent in a most responsible
manner.
PERFORMANCE EVALUATION AND COMPLIANCE TESTING

     Because of the many and varied types of aeration devices on the market
today, some means of comparing their relative efficiencies is essential.

                                    366

-------
Although oxygen transfer efficiency is only one of the several major criteria
used in the selection of an aeration device, it is extremely important in
assessing long-term energy requirements.

     A standard oxygen transfer testing procedure has been developed and
adopted by the Ontario Ministry of the Environment for use in comparing the
oxygen transfer efficiencies of different aeration devices.  The testing
procedure was developed and refined over a period of almost 20 years of
testing by the Ontario Water Resources Commission and now Ontario Ministry
of the Environment.  It has seen many modifications and refinements but is
now considered to reflect the current state-of-the-art.

     This oxygen transfer test is really a clean water test carried out under
standard, but rather ideal conditions and can be used only in determining the
oxygen transfer efficiency relative to other aeration devices.  Compliance
testing may, and indeed frequently is specified by the consulting engineer
to determine whether or not design specifications are met under field
conditions.  Compliance testing is conducted following installation of the
aeration device and may be used to determine modification requirements in a
multi-unit facility.  This is one instance in which equipment preselection
can be extremely beneficial.  Compliance testing procedures are site specific
and are generally developed to meet the needs of a particular project.
MOB OXYGEN TRANSFER TESTING PROCEDURES

     While on the surface it may seem a simple enough matter to devise a
standard oxygen transfer testing procedure, in practice the difficulties are
diverse due to the many variables which must be considered.  There are
currently over two dozen oxygen transfer test methods in use including those
of the MOE, ASCE, ASTM, EPA and of individual consulting firms.  Although
many of the methods are very similar and vary only in detail, such detail,
particularly in the method of calculation, can influence final results to
such a degree that judicious selection of method could enable one to obtain
predestined transfer results.  From past MOE experience, many manufacturers'
performance claims are exaggerated when compared to MOE field tests on the
same aeration device.  This is usually a result of the claimed performance
being based upon a method which maximizes apparent performance.

     The Ministry of the Environment oxygen transfer acceptance testing
procedure is based on the fullscale, non-steady state tapwater test.
Although the test procedure is presented in Appendix A as "Procedure for
Evaluating Aerator Performance", general comments concerning the test are
addressed below.

Site Preparation

     Testing is conducted at the Ontario Experimental Facility in a large
scale concrete tank which is described in Appendix B.

     The test tank has length, width and depth measurements of 27 ft x 24 ft x
14 ft with a capacity of about 9,000 ft3.  Small pilot-scale tests are not

                                    367

-------
acceptable in Ontario because of possible side wall effects and scale-up
problems.  The aeration device is installed in accordance with the manufac-
turer's instructions and usually run under operating conditions (submergence,
air flow rate, etc.) selected by the manufacturer to maximize performance in
that particular test tank.  A range of operating conditions may be used.

Test Medium

     Municipal tapwater is used as the medium in aerator performance testing.
Although water chemistry can be an important factor in oxygen transfer, and
such materials as iron salts and other dissolved solids, even in relatively
small amounts can enhance or depress oxygen transfer performance,  municipal
tapwater supplies are generally of relatively high and consistent quality.
For tapwaters in Ontario, « and 3 are considered to be unity, however, the
test engineer should be satisfied from past knowledge, or test water analyses,
that no impurities are present to interfere with the tests.

     Although some test procedures call for a standard test water temperature
of 20°C, this is not always practical to achieve.  Acceptance tests haveQbeen
carried out in Ontario at water and ambient air temperatures as low as 1 C
and -30°C, respectively.  As long as the temperature correction to standard
conditions is judiciously applied, the MOE testing procedure applies to any
reasonable test water temperature„

Deoxygenat ion

     In the MOE testing procedure, the aeration equipment is run at a steady-
state for at least 30 minutes prior to testing to allow establishment of the
hydraulic regime, constant temperature and pressure of air supply, steady-
state power draw, etc..   During this period, Cobalt as cobalt chloride is
added as a catalyst at a concentration of 0.25-0.5 mg/L.  Sodium sulfite is
then added, preferrably as a solution but permissibly as a slurry (cold
weather conditions), to the test tank as rapidly as possible using a sub-
mersible pump and spraying over the entire tank surface.  Sulfite additions
are made in excess of stoichiometric requirements, the magnitude of which
depends on expected oxygenation and mixing rates.

Dissolved Oxygen Measurement

     Reoxygenation of the dissolved oxygen (DO) depleted tapwater is measured
by means of several DO meters whose probes are located at selected points in
the tank to check for any significant variation in the oxygen transfer
coefficient.

     Several other testing procedures recommend or require the use of the
Winkler DO test requiring withdrawal of discrete samples or the use of
submersible pumps.  However, after exhaustive testing of most of the avail-
able DO measurement methods over a twenty-year period, the MOE has full
confidence in DO meters provided they are in good condition, accurately cal-
ibrated and show fast response times.  The DO meters used in the MOE proce-
dure, usually 4 or 5 with a minimum of four, must be accurately calibrated
each day before testing begins.

                                    368

-------
     During the test run, the tank DO is allowed to reach 90% of the satura-
tion value at the test temperature before sulfite is added for the next run.

Dissolved Salt Limitation

     As the addition of sulfite increases the dissolved salts concentration
in the tapwater, incremental addition of sulfite is allowed to a maximum of
1,500 mg/L total dissolved solids before the testwater must be changed.


POWER MEASUREMENT AND EXPRESSION OF RESULTS

     During the test runs, power and/or air flow readings are taken to
measure watts, volts, amps, power factor and scfm supplied.  Calculation of
the test results is by the log deficit method as outlined in Appendix 1.

     Performance is described in terms of capacity (Ib 0~/h), electrical
efficiency (Ib 0-/kWh) and % efficiency.  It should be emphasized that power
factor is measured in these tests.  It has been found that the power factor
is frequently as low as 0.5 rather than the normally assumed 0.85 value,
considerably affecting power draw calculations.

Mixing

     Mixing performance, although a very important consideration in aerator
selection, is a very difficult parameter to define.  Fortunately, it usually
follows that should oxygenation performance be good, then mixing will also
be satisfactory.  Standard requirements for mixing are that a velocity of
0.25 m/s is maintained at the tank bottom to suspend solid material.  In
some instances where such velocities are achieved, solids settling reportedly
still occurs.  Lately, several aeration systems have been marketed which
provide finally diffused air through a grid of diffusers covering the total
tank bottom.  Such systems do not provide a high unidirectional bottom
velocity but do provide adequate vertical shear to prevent bottom deposition.
Thus, although velocity measurements are a useful indication of mixing
performance, they are not always applicable.  As an indicator of mixing
throughout the whole test tank, the Ministry of the Environment standard
requires that the measured ILa for each sampling point shall not vary more
than ^ 10% from the mean ILa for all sampling points.  This requirement is
fairly lenient and most tests conducted to date have fallen well within these
limits; some mixing devices have, however, failed to meet the mixing require-
ments in MOE acceptance testing and were, therefore, deemed unacceptable.
Other devices have only failed to meet the mixing criteria under specific
operating conditions, and the testing has defined equipment limitations.


SUMMARY

     In summary, proper selection of aeration equipment is essential to the
successful and efficient operation of the activated sludge wastewater treat-
ment process.  The large number and diverse nature of aeration devices
available on the market necessitate a standardized approach to aerator
evaluation and selection.

                                     369

-------
     Ontario has adopted such a standardized approach including performance
evaluation and compliance testing involving a detailed oxygen transfer test
procedure.  While some criticism may be directed at the conservative nature
of the MOE method, it must be realized that all aeration devices receiving
MOE acceptance are evaluated and rated in a similar manner.


APPENDIX A - PROCEDURE FOR EVALUATING AERATOR PERFORMANCE

1)  Test Conditions and Initial Preparations

     Prior to conducting oxygen transfer testwork, discussions shall be held
between MOE personnel and the equipment supplier/consultant to ensure that
the installation conforms to required specifications and that subsequent test
results obtained will definitively reflect the performance of the aeration
device under the test conditions.

     All aerator acceptance tests will be conducted on clean tapwater; if
requested, information on the water supply chemistry shall be provided by
the MOE to the supplier /consultant.

     On the basis of all available data, the supplier/consultant will provide
the MOE with a performance claim or guarantee of transfer capacity in kg 0 /h
and kg 0^/kWh, and show the test conditions of submergence, aerator speed,
etc. relating to the guarantee.

     After aerator installation, the test tank shall be thoroughly cleaned
and filled with fresh tapwater.  From visual and/or chemical inspections, all
parties shall agree upon the suitability of the testwater for aerator
performance characterization.

2)  Deoxygenation of Testwater

     Cobalt free anhydrous sodium sulphite, Na_SO«, shall be used to deoxy-
genate the testwater.  The quantity of Na-SO  added for each run will be
based upon the equation:
which shows that 7.9 kg of Na2S03 will remove 1 kg of dissolved oxygen from
the testwater.  From an initial dissolved oxygen measurement on the testwater,
the weight of oxygen present will be calculated and the stoichiometric quantity
of Na2$0  plus an adequate excess added to achieve zero DO.  The excess
Na_SO  aaded will be based on the expected oxygen transfer rate of the
aerator.

     The Na^SO.^ will be dissolved in an external tank mixed by a propeller or
recirculating pump, and will be introduced to the testwater as rapidly as
possible by a submersible pump spraying the chemical equally over the entire
surface of the test tank.
                                     370

-------
     While not preferred, it is permissible in some instances to add the
Na_SOq to the test tank as a slurry.

3)  Cobalt Catalyst

     Cobalt shall be used to catalyze the deoxygenation reaction.  The cobalt
concentration used shall be at minimum 0.25 mg/L and at maximum 0.5 mg/L.

     Cobalt chloride, CoCl *6H20, shall normally be used as the source of
cobalt catalyst.  The cobalt salt will be completely dissolved in tapwater
and added to the testwater during steady-state aerator operation at least
30 minutes prior to the first run.

4)  Physical Parameters

     The testwater volume shall be maintained constant during each test
series.  Where several water levels are to be tested, the greatest shall
be conducted first to allow assessment of motor/blower suitability under
high load conditions.

     The aeration device shall be operated at test conditions for at least
30 minutes prior to starting any test run to allow establishment of the
hydraulic mixing regime, steady power draw on drive motors, etc..

     Under no circumstance will the aeration device be stopped from
commencement of the above 30 minute stabilization period until the end
of the test run.  A maximum of 10 runs, or concentration of 1500 mg/L
Na_SO,, whichever occurs first, shall be allowed on any batch of testwater.

5)  Sampling and DO Measurement

     A minimum of four DO measuring stations shall be selected by discussion
between all parties.  All efforts will be made to select measuring stations
which best represent the overall tank characteristics.

     The primary DO measurement method will be by fast response membrane type
probes.  These shall be capable of >90% needle deflection from DO saturation
towards zero DO in 45 seconds or less.  All meters will be accurately cali-
brated daily against the. Winkler DO method as outlined in Standard Methods.
While it is permissible to use DO recorders, it is preferred that test
calculations be based on readings taken directly from the meters.

     Although not required, measurement of oxygen transfer performance by
removal of discrete samples for Winkler analyses is acceptable.

     At least 10 DO readings will be made during each test at each measure-
ment station.  The maximum allowable time between each reading will be
1 minute.

     During each run, DO readings will be continued to at least 90% of the
saturation value.
                                     371

-------
6)  Data Interpretation

     A minimum of three runs at each test condition will be made to define
performance at that condition.  At least three consecutive runs at the same
conditions shall be within *10% of the average of the runs.  Failure to
comply will necessitate further runs at the test condition to define the
performance obtained.

     For each run, ILa values will be calculated for each DO measuring
station.  The JLa value for each station shall not vary more than *10% from
the mean value lor each test.  Greater variation will indicate incomplete
mixing and invalidate the test.

     The raw data will be analyzed using the log deficit method.  DO values
less than 20% and greater than 90% of the test DO saturation value may be
discarded in the calculations.  It is preferred that data will be depicted
graphically in addition to, or opposed to computer program analyses.

Calculation of ILa for Mechanical Surface Aerators

     In  the case of mechanical surface aerators, the overall coefficient of
oxygen transfer, K^a7_, shall be calculated by:


                 TT fl   -     1      1n   (Cs"Cl)   fl 20-T
                       -  -         ln          -   9
     where:  Kra2o = overall coefficient of oxygen transfer at 20 C

             C     = Standard Methods value for dissolved oxygen
                     saturation at test temperature, T°C

             C1'C2 = D0 concentration at times, t.. and t«

             6     = temperature correction coefficient = 1.024

             T     = temperature of testwater,  C

Calculation of lL.a for Submerged Aeration Devices

     In the case of submerged aeration devices, the overall coefficient  of
oxygen transfer, &La2Q> shall be calculated by:


                 , .   ,   _1     ,.   "WV    . 20-1
where C   is the mid-depth dissolved oxygen saturation saturation value  and
is calculated by:

                            C   = C    [P + 0.5x1 - p
                             sm    s      760 - p

                                     372

-------
     where:   C  = the Standard Methods DO saturation value at test tempera-
              s   ture,  T°C

             p  = saturated water vapour pressure at test temperature, T

             P  = barometric pressure, mm

             x  = hydrostatic head above air release point

             9  = temperature correction coefficient = 1.024

     For each test,  $La   values for all DO measuring positions shall be
averaged and used to calculate oxygenation performance.

     In the case of  mechanical surface aerators, performance shall be
expressed in terms of transfer capacity, kg 0,,/h, and electrical efficiency,
kg 02/kWh.

     In the case of  submerged aeration devices, performance shall be shown
in terms of capacity, kg 02/h, % efficiency, and also as kg 02/kWh, where
kW is blower motor wire, power.

     The calculated  mid-depth oxygen saturation value shown previously will
be used to calculate the above.

     « and 6 factors shall be assumed as unity in tapwater tests.

7)  Power Measurement

     In the case of  mechanical surface aerators, power shall be measured by
an indicating polyphase wattmeter.  Manufacturer's calibration data shall be
used to determine the motor and reducer efficiencies.  In addition, the
voltage in all three phases shall be shown, the current in at least two
phases, and the power factor determined.  Instrument accuracy shall be 1% or
better, current transformers 0.5% or better.

     A similar instrument shall be used to measure blower power draw in the
case of submerged devices.  In addition, air flow rates should be accurately
measured using an in-line orifice plate or other acceptable means.

     All power and air flow readings shall be taken when the aeration device
is at steady-state operating conditions.  Several sets of readings shall be
taken during each oxygen transfer test.

8)  Mixing Performance

     Compliance with the *10% ILa variation shown in Item 6, is acceptable as
proof of adequate mixing performance.  However, in cases where actual velo-
city measurements are made, these shall be conducted using an Ottmeter or by
a Gurley meter.

     Minimum liquid  velocity at all points in the tank shall be 0.25 m/s at
full motor load.
                                    373

-------
APPENDIX B  -  OXYGEN TRANSFER TEST TANK DESCRIPTION *•• ONTARIO EXPERIMENTAL
                                                                  FACILITY
      The  test  cell  is  an isolated bay within an existing 100 ft (-30. m)
diffused  aeration tank of the Ontario Experimental Facility, a 5,0 MIGD
activated sludge  plant maintained by the MOE for research and training
purposes.   The bottom  longitudinal corners are filleted and all construction
is  concrete including  the isolation wall installed to form the test cell,
A Y-wall  separating the test  cell from a parallel common wall tank limits
effective test depth to 14'  (4.3  m).

      Test tank specifications as  as follows;

      Dimensions;    L 27 ft (8.2 m); W 24 ft (7,3 m);  D 14 ft (4,3 m)

      Capacity;     9,072 ft3  C257 m3)
                    56,609 IG  (67,931 USG)
      Air  Supply:    Blowers -  2 Stitorbilt 30 HP,  980 cfm @ 7,5 psi each;
                            -  1 Cycloblower 50 HP, 1940 cfm @ 18 psi.
      Electrical;    575 V and  110  V outlets on-rsite (stepdown transformer
                    can be installed for 440 V),

      Water:        Tapwater available from 5 nearby hydrants.
      Laboratory:    Both permanent and mobile laboratories available on^-site.

      Hardware  &    Winkler DO chemical and bottles supplied;
      Analytical    Stream Sampler supplied;
      Equipment:    DO  meters  available (YSI and  EIL);
                    Sulfite makeup tank complete  with mixer and pump supplied;
                    Electrical test measurement kit available;
                    Truck mounted  hoist available.
DISCLAIMER

     The work described in this paper was not funded by the U.S. Environ-
mental Protection Agency.  The contents do not necessarily reflect the views
of the Agency, and no official endorsement should be inferred.
                                     374

-------
COMPARISON OF DUAL NONSTEADY STATE AND STEADY STATE TESTING OF FINE
BUBBLE AERATORS AT WHITTIER NARROWS PLANT, LOS ANGELES

James A. Mueller,  Associate Professor
Environmental Engineering and Science Program
Manhattan College
Bronx, New York 10471
ABSTRACT

     The dual rionsteady state technique has been developed to evaluate aera-
tion system oxygen transfer coefficients and saturation values under process
conditions.  This paper provides the results of application  of  the  technique
to a long  narrow  aeration tank at the Whittier Narrows plant,  Los  Angeles,
Ca.
INTRODUCTION

     During the summer of 1981 aerator testing using different techniques was
conducted at  the Whittier  Narrows plant, Los Angeles  County.   The testing
program was a cooperative effort between the ASCE Committee on  Oxygen  Trans-
fer Standards and  the Los  Angeles County Sanitation District (LACSD)  which
was conducting an EPA supported  study  on the  efficiency of various aeration
devices at  the  plant.   Three techniques were used  for testing  during  this
period; (1) the dual nonsteady state analysis,  (2)  the  steady state analysis
and (3) the off-gas analysis.

     This report presents  the  results  of  the  first two testing  techniques,
the dual nonsteady state and the steady state analyses.  The main body of the
report summarizes the study procedures and results of the two techniques.
DESCRIPTION OF TESTS

     The Whittier Narrows wastewater treatment plant is a facility containing
primary clarifiers, long narrow aeration tanks, secondary clarifiers, chlori-
nation tanks and  tertiary  filters.   It treats mainly  domestic  sewage at a
relatively constant flow rate over  the day.   The influent COD concentration
exhibits a marked  diurnal  variability  with minimum values occurring  in  the
late morning hours, increasing  during  the day to a maximum value which  re-
mains relatively constant from the early afternoon hours  through  the  evening
hours (Figure 1).  To attain a relatively  constant 0^  uptake  rate during the
tests, non steady  state data was obtained in  the evening starting at the
earliest around 5:30 P.M. with some of the tests completed after midnight.

     Figure 2  shows a schematic of  the aeration tank  layout at  Whittier
Narrows.   Fine bubble diffusers were located in  the first two bays,  one  with
full floor coverage discs and the other with double spiral rolls  tubes,  both
constructed in a tapered aeration mode with approximately 45% of the

                                     375

-------
   15
   10
o
2

O
    0
  400
13Aug.1981
  300
c
O
O
  200
  100
                                         Primary Effluent

                                         March —April, 1973
      Mid
           Noon
                                                                          Mid
    Figure 1.  Diurnal Variability of Wastewater Flow and COD Concentrations
               for Whittier-Narrows Plant
                                    376

-------
                                     •Figure 2,   ^hittier Harrows Plant - Aeration TanK Layout
co
         (Primary Effluent>Q1
         (Return Sludge)


1
30'

t>Qi
\


,

IWMveM
9tnm*^*m
-i • „ , 300* i

JET DIFFUSERS
1' off bottom
TUBE DIFFUSERS 1
2' off bottom j
Grid //I. 270 tubes Grid #2, 210 tubes [Grid 03, 120 tubes
— »• DISC DIFFUSERS j
2' off bottom |
Grid #1, 724 Units Grid 92, 594 Units iGrid #3, 352 Units


jets perpendicular
to wall
double spiral roll
full floor cover-
age
                                            Water Depth for Disc Diffusers - 14.3-14.4'

-------
diffusers in the first third of the tank, 35% in the second  third  and 20% in
the last third  of  the tank.   The third bay contained jet  diffusers located
perpendicular to the wall.  The initial configuration was jet clusters in the
center of the tank.  These were removed and the longitudinal header installed
in their place.  A number of the tube diffusers in the  second bay  had broken
off prior to the study with aeration patterns erratic down the  length of the
tank as evidenced  by  the  location  of  the standing wave on the  tank surface
due to  the double  spiral  roll.   Due to  the  poor operation of the  tube
diffuser aeration system and low DO's in the jet system during  the summer of
1981, neither system was  analyzed  by  the iionsteady state testing  technique.
All testing was concentrated  on the disc system which  exhibited  excellent
operational  characteristics  during  this period  although  unequal  flow
distribution was evident in Grid #2.

     Six dual nonsteady state tests were conducted on Grid #3 and  one test on
Grid #2 as indicated in Table #1.  The normal flow rate to the system includ-
ing both primary effluent and return sludge was 5.4 to  6.0 MGD.  Five of the
seven tests were conducted at  the  normal plant  flow rate allowing gas  flow
ratios  (high/low)  of  1.8  to 2.4.  The high  gas  flows  were limited by  the
available blower capacity at the plant and pressure at  the blower  discharge.
The low gas  flows  were limited by the requirement of maintaining  a minimum
D.O. concentration of  _ 0.6 to  1.0 mg/£  to  insure  that  oxygen concentration
does not limit  uptake  rate.  On 23 July two series of tests were conducted at
half flow rate, one on Grid #2 and the other on Grid #3.   This  allowed  lower
gas flows  to be used  providing a greater difference in the high to low  gas
flow ratios.  This is  desirable  to minimize  the variability in estimated Q
uptake  rates and saturation values calculated from the  dual  non steady  state
analysis (Mueller  and  Rysinger, 1981)..

     Additional tests  at half flow could not be conducted  due to the  lack of
plant flexibility  imposed  by  the requirement of  maintaining a  high quality
effluent at the normal flow rate for wastewater reuse downstream.  An attempt
was made prior  to  the last test to run  two blowers in  parallel instead of
just  the one used  for normal  plant operation.   However this  resulted  in
overheating and surging of the smaller of the two  blowers  which then had to
be shut down.

     Figure 3  indicates the probe  locations used in the  tests.  They were
concentrated in the latter portions  of  a grid where the  denser  diffuser
placement of the upstream grid would have a negligible effect  on transfer
efficiency of  the  grid being  analyzed.   The probes were normally  spread at
regular  intervals  to  allow analysis of  a  tank segment knowing inlet and
outlet boundary conditions or analysis of the  total grid segment  assuming a
plug flow mode.
                                     378

-------
               TABLE 1.   DUAL NONSTEADY STATE TEST CONDITIONS
                    FOR DISC DIFFUSER SYSTEM
Run
1
2
3
4
5
6
7
Date
1981
23 June *
24 June
7 July
13 July
23 July
23 July
13 Aug
Flow
Rate
(MGD)
6.0
5.4
5.4
5.7
2.9
2.9
5.6
Grid
Tested
3
3
3
3
2
3
3
Mixed
Liquor
Temp.
27.7
27.8
28.1
28.1
28.1
28.2
28.0
Gas Flow
Ratio
Gs2/Gsl
1.84
2.36
1.97
1.78
3.72
3.51
1.89

Includes Primary Effluent  and Return Sludge Flow,  Q,,/Q.
Runs 1,2,3,4,7.   Q_/Q.  = 0.15 Runs  5,6.
                  K.  1
                                                           0.23
     Difficulties were encountered with  the D.O. probes during  the  tests.
For runs 1 through  4  and run 6, five probes were available.   The probe in
position #4 was a field  "Delta"  probe  connected to  a recorder in the plant
operations building.   Three YSI probes  were normally used,  one each  in
positions #2  and #3 and one with a laboratory stirrer connected  to the  probe
used for the 0«  uptake analyses.  In position #1 an older laboratory Delta

probe was utilized.  This  older  probe  tended to lose calibration over  the
test and was  finally abandoned by run #7.  The major limitation of the probes
was the lack  of field mixers to provide a constant  level of turbulence across
the probe membrane equal to that provided during calibration and  to dislodge
any gas bubbles entrapped on the membrane surface during field testing.  Run
#1 was conducted for technique shakedown and probe calibration was question-
able.   For Run #5 only three in- tank probes were available  since the field
Delta probe could not  be moved  to  the  Grid #2 position.  For the last  run,
the two YSI probes were used at positions 1 and 5 with the field  Delta  probe
at position 7.  Weston-Stack probes were then  used  for  the other locations.
To prevent gas entrapment which  was found  to markedly affect probe readings
of these latter probes, all probes were  inverted for this  run.   However, the
Weston-Stack  probes severely  lost  calibration and  therefore  the data from
only probes 1, 5, and  7  were  utilized.   Thus the effect of probe depth and
lateral location in the  tank  could not be  ascertained.  The specific test
procedure used for a run is outlined in Appendix A.
                                     379

-------
        Figure 3.   D.O.  Probe Locations in Disc Diffuser System


                     Grid No.2               Grid No.3
Runs
 1,6
 Runs
 2,3,4
Run
  5









^
}


f
)




1
1
1
1



rroDe uepin
+ o*
T ^
• 10'






4-U
4

















2




I
I
! 4'-L 3
f






I
I
-
t

I
3 i
U lni-
1 III

14G

f
K H








4





r
y


> 3


__
3fi

52'

15

•^ . .
35
52'






t.
4
*l
5"





I
4
X









-4'








«-4'







nun s
7 .(
L_





i 	
i
i
t

Probe Depth
+ O ^
2
x 6'
• 8'
is!
4-1
•





— »
[ —




3
+





4 7
56*
x t
1 -H
KJ"
"• M
r 35'
52'


^ ,




                                    380

-------
DATA ANALYSIS TECHNIQUES

     To date,  the  nonsteady state  aeration testing techniques  have been
applied to either batch or completely mixed  systems  for  both  clean water and
respiring systems testing.  The equations  required  to  analyze these systems
are given in Appendix B for the dual non- steady state analysis.  Neither one
of these approaches applies to the long narrow aeration tank at Whittier-Nar-
rows under process conditions.  Initially  it was anticipated that  the  tank
could be divided into completely mixed segments with inlet and  outlet condi-
tions defined.  If the  change  in  concentration between the inlet and outlet
boundaries was constant with time (constant  AC analysis) during  a  non steady
state test,  then  the analysis for  K a  reduced to the batch  system.  The
                                    L
latter runs showed that this was  not  the case, therefore requiring another
technique for data analysis.

     Upon completion of Run #7, a D.O. profile was  taken down the length of
the total aeration tank as  shown  in Figure 4.   Significant 0  concentration

gradients existed throughout the tank especially in Grids 1 and 3, indicative
of a plug flow mode rather than a completely mixed one.

     To determine the best technique in which to analyze the  nonsteady  state
data in  a plug flow type system, an analytical solution of  the  equations
describing the system were obtained.  The following conditions were used for
this solution based on the Whittier Narrows data:

     (1)  The  initial tank concentrations  were set at  the  steady  state
concentrations for the previous gas flow prior  to  changing gas flow for the
nonsteady state profiles,
     (2)  The  09 uptake and transfer rates  over the region  being analyzed

were constant, and
     (3)  The change in concentration at the upstream boundary was described
by an exponential function.

A simulator of the plug flow  system was constructed and various  conditions
analyzed similar to those encountered in the Whittier Narrows runs.

     The results of  this  exercise indicated that the  points  further down-
stream from the boundary gave accurate estimates of  the  true K a  existing in
                                                              J_i
the tank when analyzed by the batch nonsteady state equations.  This was also
true of the oxygen saturation  value if  negligible  spatial gradients existed
at steady state at these  locations.  The final data analysis then used the
latter probes for the ILa and  C  values for both the  steady  and nonsteady
                        L        K
state analyses.  For most runs,  simulator  results  were also  obtained  to
insure proper data analysis.  The final  results  for  each run are summarized
below.
                                     381

-------
Figure 4.  Profiles of  D.O.  Concentration and Uptake Rate at End
           of Run #7 over Total  Tank Length
                    100                    200
                     Distance from influent, ft
300
                              382

-------
RESULTS

     The gas flow measurements, average oxygen uptake rates and 0  saturation

values estimated from clean water data are presented in Table 2 for each run.
Relatively low  gas  flows were obtained  in  Runs 5 and 6  due  to the lower
uptake rates  occurring  at half  normal  flow.  The  clean  water saturation
values were relatively  constant  since the water depth and  temperature  were
similar for all runs.  Two 3 factors, 0.95 and 0.99, were used for this data.
The 0.95 value was common practice of the LACSD-  However based on TDS values
from 480 to 570 mg/£, a  3  factor  of  0.99 would be more applicable.  Table 3
summarizes the dual nonsteady state results for each run.  No 0,,. uptake rates

or saturation values could be obtained for  Runs #2  and #3 since the low gas
flow steady state oxygen value was less  than zero and  a  reliable  estimate  of
ILa could not be attained.  In Run #7 a spatial gradient still existed at the

effluent end of the tank requiring an estimate of the steady state D.O. value
for the uptake  rate and saturation value determination.   The steady state
results are presented in Table 4.  These were calculated using the saturation
values estimated from clean water values in Table 2 with  the  steady  state  0,.

concentration in Table 3.  For the high  gas  flows the  higher  g value of 0.99
yields 6 to 10% lower calculated K a values while the  lower gas flows  show a

reduction of  3  to  4% due to the lower steady  state D.O.  concentrations and
increased driving forces.

     Figures 5 and 6 show the correlations of the field oxygen transfer coef-
ficients with gas  flow  for both nonsteady and  steady  state analyses.  The
degree of scatter for the  steady state  analysis (3 = 0.95) is significantly
greater than that for the  nonsteady  state results.   This  may  be due to poor
probe  calibration  since the K^ a values  for the steady state  analysis  are

dependent on probe calibration while for the nonsteady state analyses,  the

K, a estimate  is  ,  ,     ,      ,.    , .,           ,        . _ .
 L              independent of  calibration  as long as it is constant over a
run.  Both analyses exhibit curvature when  the  line of best fit (by eye) is
assumed to pass through the axis.

     When the average R^a  value  from two D.O. probes was used for  the  dual

nonsteady state results,  the  range  of values is presented.   The  difference
between the actual K a  values  and the mean  is generally less  than  6%.   For
                    L
Run #2 a  difference of  12% was  obtained possibly  due  to different probe
locations near the tank effluent, one at the tank center at a depth of  10 ft.
and one at the tank wall at a depth of 2 ft.  For Run #5 conducted in Grid #2
a marked spatial difference in K a values  existed,  with the probe  in  the
center of the segment yielding  significantly higher ILa values  than  the

downstream probe.   This was probably due to the unequal gas flow distribution
existing in the second  Grid,  the average K  a  "value  used to  represent the
                                           i-i  L
overall grid segment analyzed.   The  curve through the nonsteady  state  data
yields oxygen transfer  coefficients  about 7% higher  than that through the

                                    383

-------
    TABLE 2.  MEASURED AND ESTIMATED GAS FLOWS, UPTAKE RATE AND
              SATURATION VALUES FOR WHITTIER NARROWS RUNS

ESTIMATED WASTEWATER DISSOLVED OXYGEN
MEASURED VALUES
Gas Flow, scfm On Uptake
Tt -T TT • 1_ ^"
Kun .LOW uign 	
Gsl Gs2 mg/£-hr 0
1 520 960 35.4
2 500 1190 29.8
3 590 1160 34.2
4 690 1230 30.8
5 390 1440 22.0
6 330 1160 18.5
7 600 1130 31.0
SATURATION CONCENTRATION
A
C~f
= 0.95 g = 0.99
8.2 8.5
8.2 8.5
8.1 8.4
8.1 8.4
8.1 8.4
8.2 8.5
8.2 8.5

TABLE 3. DUAL NONSTEADY STATE RESULTS FOR
WHITTIER NARROWS RUNS

Steady State 02
Concentrations ,
rng/A
Run KLafl KLaf2 Si- CR2°°
1 3.7 6.1 0.7 4.1
2 - 6.8 < 0 3.9
3 - 6.8 < 0 2.9
4 5.4 7.2 1.2 3.5
5 3.0 8.1 0.7 5.2
6 3.0 7.2 0.8 5.4
7 4.0 7.7 0.5* 3.8
Calculated
Q Uptake 0 Saturation
1 L *
Rate, R, Value, C, ,
mg/£-hr mg/Jl
32.0 9.3
-
-
49.7 10.4
21.4 7.8
23.7 8.7
27.5 7.4

Estimated at 0.2 mg/£ lower than measured value at last downstream
station due to spatial 0_ gradient.

                              384

-------
TABLE 4.  STEADY STATE RESULTS FOR WHITTIER NARROWS RUNS

* -1
02 Transfer Coefficients , hr
Run
1
2
3
4
5
6
7
3 = 0.95
Vfl Vf2
4.7 8.6
6.9
6.6
4.5 6.7
3.0 7.6
2.5 6.6
4.0 7.0
B = 0.99
Vfl Vf2
4.5 8.0
6.5
-
4.3 6.2
2.9 6.9
2.4 6.0
3.9 6.6

*
f - R/(c*f - V

                           385

-------
10
              • GridNo.3


              + Grid No. 2
             250        500        750        1000        1250       1500
                                 G scfm
                                   s,
  Figure 5.   Effect of Gas Flow on K^  for Dual Unsteady State Analysis
                                 386

-------
   10
.c
 1*1
(Q
           • Grid No. 3

             Grid No. 2
             (3=0.95
               250         500
  750

Gs, scfm
1000        1250
                                                                         1500
   Figure 6.  Effect of  Gas  Flow on ILa, for Steady State Analysis
                                    387

-------
 steady  state  data.   Since  the  higher  B value of 0.99 would yield still lower
 steady  state  K a values, the g of 0.95 was used  in  further  comparisons.

     The  oxygen uptake  rates calculated  by  the dual nonsteady state analysis
 are  compared  in Figure  7 to  the  average measured values for each run.  Good
 agreement between measured and observed  values is obtained except for Figure
 4 which is  67% greater  than  the  measured value.  Figure 8  shows good  agree-
 ment between  the clean  water estimated and  field estimated oxygen saturation
 values  except for Run #4.   Figure 9 shows the differences  between  measured
 and  calculated oxygen uptake rates and saturation values as a function of  the
 high  to low  gas flow ratios used  in  the dual  nonsteady  state studies.
 Generally the higher the gas flow ratio, the  greater  the  agreement between
 observed  and  calculated values.   The  greatest difference occurred in  Run  #4
 which had the lowest  gas flow ratio with a resulting K a ratio of only 1.3.
                                                              *
 This low  ratio would  lead  to inaccurate estimations of R and C^ as  indicated
 in the  original work  of Mueller  and Rysinger,  1981.

     The  above  results  indicate  that  the measured CL uptake  rates  are good

 estimates of  the  rates actually  occurring in  the tank.  This  is  due to the
 fact  that the majority of the runs were conducted near the  tank effluent
 where the soluble organic  load 'would have been oxidized; the  uptake rate  due
 mainly  to sorbed  organics  within the  activated  sludge floe particles. The
 oxygen  saturation  values  estimated from  clean water data  also  appear to
 adequately  represent  the actual  tank values.   Due to the variability inherent
 in the  dual non steady  state estimation of C  , it cannot be used to delineate
 the  field saturation value any better than +_ 5%  at  a gas flow ratio of about
 4.0  and +_ 12% at a  gas  flow  ratio  of  about  2.0.   Oxygen uptake rates  should
 have approximately  the  same  limits.   The K a  reliability from the  dual non

 steady  state  results  is better than the above estimates, an average coeffi-
 cient of  variation  of 3.6% attained by the  three parameter estimation model
 for  the ¥L a values  obtained in  this study.   The above variability  for the

 oxygen  saturation  values,  uptake  rates and  oxygen transfer  coefficients
 should  be reduced with  use of  better  D.O.  probes and field stirrers  on the
 probes  when the probes require  a minimum velocity across the  membrane.
 Figure  10 summarizes  the oxygen  transfer efficiencies  attained  for each run
 as a function of the gas flow  rate per diffuser.  The oxygen transfer effi-
 ciencies  (OTE) are  calculated  at zero dissolved  oxygen  concentration  using
 the field 1C a values and clean water saturation values at 28°C as follows:
                          = Vf C*f 8.34V/1.04Ga
where:  IL af [ = ] hr
        V = total volume of one Grid, 0.322 MG
        G  [=] scfm
         s
                                     388

-------
  60
  50
  40
O)
  30
  20
  10
                 I
       •  Grid 3
       *  Grid 2
       1  Run No.
                          • 6
                                         • 4
                10
20
30
40
                                  meas.,
                                               •hr
50
60
    Figure 7.  Comparison of measured and calculated CL uptake rates
                                      389

-------
         Temp. = 27.8 - 28.2° C

         Submergence = 12.3ft
                        46s
                C*oo f  . estimated from clean water, mg/l
Figure 8.   Comparison of  Calculated and Estimated  (0 - 0.95) Oxygen
           Saturation Values.
                                390

-------
       80
       60
    2>
    0)
       40
    0)
    •+•«
    J3
    O
    £   20
                              • 4
                                              I

                                        •   R
                                        A   Ct»f
• 6
                               G JG ,
                                s2  s1
Figure 9.  Effect of Gas Flow Ratio on Difference  Between Measured
           and Calculated 0- Uptake Rates and  Saturation Values
                              391

-------
            20
CO
vo
ro
      UJ

      O
            1 5
10
                                        2.00 Gsd, r  = 72%
                         'Nonsteady State K L a Values"
                                                                                          = 37%
                                                 6= 0.95

                                                 0*00^8.1-8.2^
                                                              • Grid #3

                                                              + Grid #2
"Steady State K L a Values"
                            1             2             3          4,0           1            2            3

                                                   Gsd, Gas Flow/Diffuser, scfm/unit

                   O2Uptake Rate, mg/1-hr

                   Figure 10. Effect of Gas Flow/Diff user on Oxygen Transfer Efficiency at zero D.O. and 28° C for Non Steady and

                   Steady State KL a values

-------
     For both Grids 2 and 3, transfer efficiency decreases with increased gas
flow/per diffuser.  A linear  regression on the Grid  #3  data gives similar
results except a  significantly  greater correlation coefficient is  obtained
using the non steady state estimate for K  a  over  the  steady state.  For the
non steady state  estimates  only 28% of  the  variability is due to  factors
other than gas flow differences while  the  steady  state results show 61%  of
the variability in OTE  is  due to factors  other than  gas  flow change.   Al-
though the data base is  small this  does tend to indicate  probe calibration
deficiencies as discussed previously.   The lower  OTE  values for Grid #2 may
be due to possible plugging problems or the  greater diffuser  density in  this
grid compared to #3 with similar  total gas flows  giving similar OTE values.
Reduced alpha values in Grid  #2 compared to  Grid  #3 would also provide this
type of result, however  the Grid  #2 data was attained at half flow at rela-
tively low oxygen uptake rates where the alpha value  may  have been close to
that normally present in Grid #3.
CONCLUSIONS

     1.   Both the dual nonsteady state and steady state analyses are capable
of measuring oxygen transfer  coefficients  under process conditions in  long
narrow aeration tanks when constant oxygen uptake rates are obtained.

     2.   For the Whittier Narrows study,  the  average non steady state K a
                                                                         LJ I
results showed somewhat less variability than the steady state results.  This
was caused to some extent by  lack of  field stirrers on the D.O. probes  and
possible inaccuracies in probe calibration.

     3.   In the dual nonsteady state analyses, a gas  flow ratio of 2:1  pro-
vides estimates of oxygen uptake rate and saturation value within +_  12%  of
steady state uptakes and saturation values  estimated from clean water data.
This variability is reduced to +_ 5% at a gas flow ratio of 4:1.  Unless  this
variability can be further  reduced by using higher  gas flow ratios and/or
better D.O. probes,  clean water  saturation values  properly corrected  for
waste total dissolved solids should be used as field saturation values.

     4.   For the fine bubble diffuser system studied at Whittier  Narrows,
both nonsteady state and  steady  state results showed  oxygen transfer  effi-
ciencies (OTE)  to decrease with increasing gas flow per diffuser.
RECOMMENDATIONS

     1.   For  a  long narrow (plug flow type) aeration tank, dual  non  steady
state testing  should be  conducted in a section with uniform uptake  rate  and
02 transfer  rate using two sets of probes, one set n£ar the upstream  end of
the section  to define the boundary effects and the other  some minimum dis-
tance downstream approximately 2.5  to  3.0 times the ratio  of the forward
velocity  to  the   oxygen  transfer  coefficient (U/KLaf).  A minimum  of
two
                                     393

-------
probes  should be  used  at the downstream station which are calibrated at the
wastewater  temperature and  contain  field  stirrers when  required  by the
particular  probe.   Oxygen uptake rates should be run at  three  locations in
the section during  the testing to  insure minimal  temporal and spatial varia-
bility.   Steady  state  dissolved  oxygen profiles  should be obtained in the
section before  and  after each test.   Low  gas flows (G .) should  maintain

minimum dissolved oxygen  levels of at  least 0.5-0.7 mg/£ at all points in the
section being analyzed.   High gas  flows (G „) should be set as high as pos-

sible to  attain maximum accuracy for 0,, uptake rate and saturation values.

G ,JG   ratios of at least 1.8/1 should be attained.
 $/.  S.L
     2.   Steady  state testing can be  conducted in  a tank section  similar  to
that above.  Both oxygen  uptake  rates  and steady  state dissolved oxygen  pro-
files should be  obtained in the  section at a period of relatively  constant
flow and  organic loading to the tank.  Data  analysis  should be conducted
using the dissolved oxygen concentration at spatial equilibrium, C  .


REFERENCES

Mueller,  James A. and  Rysinger, J.J.   "Diffused Aerator Testing Under Process
Conditions," 36th Annual  Purdue Industrial Waste Conference, May 1981.
ACKNOWLEDGEMENTS

     The  author  wishes to acknowledge the  assistance  of the following  in
conducting the study:  The LACSD for providing the facility  and  particularly
Fred Yunt for his assistance and long work  days  during data collection; Dr.
Michael Stenstrom for  providing  the recorder,  the three UCLA graduate  stu-
dents, namely  Gail  Masutani,  Hwung J. Hwang,  and Hoa Tran;  and Dr.  John
Connolly  from Manhattan College  for his  assistance  in  solving the nonsteady
state plug flow equations.


DISCLAIMED

     The work described in this paper was not funded by the U.Sr.  Environ-
mental Protection Agency.   The contents do not necessarily reflect the views
of the Agency, and no official endorsement should be Inferred.
                                     394

-------
                              APPENDIX
A.   DUAL NONSTEADY STATE TESTING PROCEDURE

     The following procedure was  used  during the Whittier Narrows
study.

     1.    Set  up probes and recorder on aeration tank

     2.    Utilize tap water at temperature equilibrium with aeration
          tank for probe calibration.  This was accomplished by
          submerging a plastic bottle containing tap water in the
          tank for r» h hour.   Two Winkler tests were conducted and
          the  contents poured into a BOD bottle for probe  readings
          using a magnetic stirrer for agitation.   For the larger
          probes, a bucket at temperature equilibrium was  used for
          calibration.

     3.    Calibrate recorder to probe readings.

     4.    Insert probes in proper position in aeration tank.

     5.    Measure manometer readings across the orifice plate on  the
          downcoraer for each grid and tank D.O. values to  insure
          D.O. not limiting.   Also measure in line air temperature
          and  pressure.

     6.    Change valve on main header for gas flow diversion from
          the  two aeration tanks not being studied co the  one under
          study.  Open valve on downcomer of grid being analyzed  to
          attain maximum gas flow.  This procedure was accomplished
          in less than one minute.

     7.    Continually record D.O.  readings either manually or with
          the  recorder for the test duration.  Test durations were
          25 to 40 minutes for the high gas flow and 40 to 60
          minutes for the low gas flow.

     8.    During the test, sample at each probe location for 0 up-

          take rate.  Each uptake rate required approximately 10
          minutes, duplicate uptake rates obtained at each location
          during the total study.

     9.    When steady state was reached for the high gas flow, the
          valve changes were reversed.

    10.    D.O. readings were again recorded and uptake rates
          measured until steady state was attained.
                                   395

-------
    11.   Upon completion of the study, probe calibration was
          checked using the procedure in #2.
    12.   Other parameters .such as flow rate (raw and return
          sludge), tank water depth and atmospheric conditions
          (barometric pressure, temperature and relative humidity)
          were also obtained.

B.   Equations for Dual Nonsteady State Analysis

     1.   Batch System or Plug Flow System

          a.   K a :
                L £                    -K a.t
                    CR - C - (CR - Co>G    r
               where: C  = steady state concentration at a location
                           in the aeration tank a minimum distance
                           of 2.5 to 3 x U/K a
                                            ij L
                       C = probe reading at any time t at the above
                           location

                       C  = probe reading at zero tirae at the above
                            location

                     K a  = field oxygen transfer coefficient
                      Li T.
                        t = time

                        U = longitudinal velocity in aeration tank
                            - (Qi + QR)Mc

                       A  = tank cross sectional area
                        c

                   Q^» QR = raw and sludge recycle flow rates,  re-
                            spectively.

The three parameter estimation model gives both K a   C , and C  for
each non steady state test.                                    °

          b.   R:
                    R= ^--
          where:     R = Oxygen Uptake Rate

                    Subscripts 2  and 1 = high and low gas flow
                    conditions respectively.
                                   396

-------
               Subscript ° = value at spatial equilibrium
                u         r,
               where -7— = 0
                     dx
     c.   C*  = €„  +  R
           «f    R»   -—
          Either condition 1 or 2 may be utilized for this
     computation.  C*  is assumed constant for both gas flows.

2.   Completely Mixed System

     a.   K a:
                                   -(K a  + 1/t )t
               CR ~ C " (CR - V e
     where:  t  = detention time, V/(Q. + Q )
              O                       1    K

              V = volume of completely mixed tank
     b.   R:
                             r  — r     r  - r
                          i    ,•     R i     -,•    R?
              c     c   +   (                   '
               R2 "  Rl   t ^  KTa.1  "   K_a..0 '
                           o    L fl       L f2
          R — 	•	1	_
     c.   C c:
           oof
              _ c
              "
                     Vf   Vf C0
     Either condition 1 or 2 may be utilized for this
                    A
     computation.  C , is assumed constant for both gas flows.
3.   Nonsteady State Plug Flow Equations at any position in
     aeration tank:

     a:   Assumptions:
         (1)  Constant R
         (2)  Constant K a
                                   397

-------
(3)   Initial D.O.  profile at steady state in aeration
     tank at any initial K, a  value
                          LI r
                                 -K.  x/U
                                   i
     C(x,o)  = initial steady state concentration at
              any x

     K=KLaf

     subscript i = initial condition

     C  = initial D.O.  at boundary

(4)  Upstream boundary condition can be described by
     an exponential function similar to nonsteady
     state aeration equation.

                             -K t
                               o
     C  - C(o,t) = (C  - C )e
      m              mo
     C  = steady state concentration at boundary after
      ro       ,.
          gas flow change

     K  = coefficient describing rate of change of D.O.
          concentration at boundary
     b.   Solution:

(1)  For t - x/U <_ 0


     V - C(x't) = (CR- - W6    + (CR~i

     where C(x,t) = D.O. concentration at any x at any time
                    <^ x/U

                K = 1C a  after change in gas flow


(2)  For t - x/U > 0

                               -K (t - x/U) - K x/U

     CR- - C(X'C) = (CR~ - Cm)e  °

     where C(x,t) = JJ.O. concentration at any x at any time
                    >  x/U
                       -Kt                   r-(K-Ki)t
                         398

-------
(3)  At  steady  state,  t = °° and Equation (2) becomes:

                               -K x/U
    CM - C(x.t)  - (CRM - Cm)e
                                     399

-------
 NEW DIRECTIONS IN AERATOR EVALUATION

 J. S. Hovis and James J. McKeown
 NCASI, Tufts University, Anderson Hall
 Medford, Massachusetts  02155
      Aerators must operate within  a system  composed  of" tank  dimen-
 sion  and particular mixed liquor characteristics.  There  is  a  need
 to be able  to evaluate the performance of  the entire  system,  and  to
 be able to  sort out the  cause of any inefficiencies  which may  be
 found.

      This paper discusses two approaches  designed  to evaluate  total
 system performance in  terms of aerator performance.  Each of these
 approaches  is developing and their limitations are being  better
 definied.   Both involve  testing respiring systems.   The first  is
 the mass balance approach developed to assist in the evaluation  of
 aerated stabilization  basins where other  tests are impractical.
 The second  is the inert  gas tracer method which has  the capability
 to determine the oxygen  transfer in almost  any system.


 THE MASS BALANCE APPROACH

      The mass balance  procedure measures  the change  in  the oxida-
 tive  state  of the respiring biological system.  The  method can be
 applied to  high rate activated sludge systems and  low rate aerated
 stabilization basin systems.  The accuracy  of the method  depends on
 the extent  to which all  factors which supply and withdraw oxygen
 from  the system are measured.  A number of  these factors  can be
 minimized during the test period so that  the major factor which
 determines  the aerator oxygen transfer is the reduction in BOD
 through the respiring  system.

      The mass balance  procedure has been  applied to  low rate aerat-
 ed stabilization basin (ASB) systems because oxygen  transfer cannot
 be readily measured using more direct methods.  The  mass  balance
 method is simple and straightforward especially when the  ASB system
 is aerator  (oxygen) limited and where significant  settleable biolo-
 gical solids aren't produced (1).

Minimum DO Procedure

      In order to minimize or eliminate a  correction  factor for the
working DO in the basin, it is preferred  that the  system  be  oxygen
 limiting during the test period.  Tuning  the system  may be neces-
sary  (i.e. turning off or redeploying aerators) if the  system  nor-
mally carries excess dissolved oxygen.  Actually,  several tests  at
approximately 60, 80 and 100% of tuned aeration capacity  may be
performed to insure an accurate test.  If the system is oxygen
limited,  the oxygen transfer efficiency (kg/kwh) should be equal
for each test.   If the values don't agree,  other limitations may
exist, such as insufficient nutrient or the presence of toxic  ma-
terials.
                               400

-------
     Certain systems will  lend  themselves  to segregation for the
 purpose of testing.  Tanks  in series,  or  long narrow tanks,  would
 be  subject to testing for  various  groupings  of aerators.  Thus,  it
 is  possible that tests could be made on various portions of  the
 tuned capacity.  All portions of each  cluster would  have to  be
 oxygen limited.  The last  stages of such  a system might be omitted
 in  order to improve the accuracy of the test procedure  because
 these later stages are apt  tc produce  excess DO during  a portion of
 the test period.

     The major variables in the calculation  are the  incoming and
 exiting mass of BOD, COD,  TOC (converted  to  its oxygen  equivalent)
 or  ultimate BOD.  In certain cases, such  as  with many pulp and
 paper mill works, the BOD5  can  be  used because the system removes
 only a small fraction of the COD(l-2). The  calculation determines
 the difference between influent and effluent BOD during the  test
 period.  This difference is then divided  by  the operating power
 used by the aerators during the same period.   The calculation pro-
 duces the results in terms  of the  pounds  of  oxygen transferred per
 horsepower per day (or kg/kwh).

     It should be recognized that  under oxygen limited  conditions
 there is a potential to accumulate BOD in  the microbial mass which
 may result in an apparent  oxygen transfer  rate (OTR)  greater than
 the actual OTR.  However,  because  most ASB systems are  respiring at
 extremely low rates, synthesis  effects are minimal and  can be neg-
 lected.

     There are several other adjustments which can be incorporated
 into the calculation.  A discussion of each  follows:

     1.  DO Difference - If the system has been properly tuned,  DO
 returning to the aerator should be zero and  DO leaving  the basin
 will probably be zero.  Thus, the  case reduces to the situation
 where the incoming wastewater contains DO  which would have to be
 debited from the aerator transfer.  For most tuned situations, this
 factor represents less than 1%  of  the  total  transfer.

     2.  Surface Area - This factor is usually disregarded.   How-
 ever, the aerators could be debited with  a value of  50  Ibs oxygen/
 acre of surface.  This value represents BOD  removal  associated with
 anaerobic stabilization basins  which are  not mechanically aer-
 ated(3).  Unless the ASB is very large in  comparison to the  power
 combined in the aerators (i.e.  greater than  10  gal/HP), the con-
 tribution to the surface area to aeration  is not a major factor.

     3.  Benthal Oxygen Demand  - Aerated  stabilization  basins are
 not completely aerobic.However,  the  pumping rate of the mechnical
or diffused aerators recirculates  the  total  basin contents through
 the aerobic zone to an extent which impedes  the development  of
highly reduced conditions.  Thus,  appreciable amounts of oxygen  are
not required to satisfy an  immediate chemical oxygen demand.


                                401

-------
However, bottom deposits, no matter how thin, constitute a con-
tinued sink for any oxygen which reaches the bottom.  Sediment
oxygen demand studies on a variety of benthal deposits have shown
uptake rates varying from 1.0-20.0 gm/m /day.  A credit of 5.5
gm/m /day (50 Ibs/acre/day) is suggested for the benthal oxygen
demand (4,5,6).  Again, even if DO exists near the bottom of most
of the basin, the amount of oxygen removed is minimal.

     4.  Nitrification-Denitrification Factor - The oxygen equiva-
lent of the nitrogen balance will have to be considered in some
systems.  The aerators should be credited with transferring the
oxygen necessary to balance the change in the oxidative state of
the nitrogen species entering and exiting the system.  However,
nitrification is suppressed under oxygen limited conditions (e.g.
DO less than 1.0 mg/1) and also ceases at BOD:N ratios of 16 or
more(7).  Thus, nitrification and denitrification aren't signifi-
cant in the oxygen limited equilibrium test case and the high
BOD:N case.

     5.  Correcting for Temperature - One advantage of using the
mass balance approach is that no temperature correction is re-
quired.  Biological kinetics call for a doubling of the transfer
rate coefficient each 10°C rise in temperature, (i.e. 0 = 1.072).
(8-10)  Thus, an oxygen limited system operating at 10°C would
simply require less oxygen (HP) than when operated at 20°C.  The
oxygen demand would be matched only to the extent that there was
aeration capacity in the system.

     The transfer rate coefficient, K^a, is temperature dependent.
The 6 value most commonly used is 1.024.  This 6 value offsets the
temperature coefficient for the saturation concentration of DO in
water.  As an example, when the DO is limited  (i.e. D0=0) the
deficit equals the saturation concentration.  The result is that
the same mass rate of oxygen is transferred into water at 10°C as
at 20°C or 30°C under the oxygen limited condition.

Excess DO Procedure

     Some ASB systems can't be turned down to minimize DO through-
out the basin.  Several additional factors must be considered  in
order to compute the oxygen transferred by the aeration equipment
when excess DO is present in the basin.  These factors are briefly
reviewed as follows:

     1.   Driving Force Factor - If the basin  is tested when ex-
cess DO is present, the aeration equipment isn't operating at
minimum efficiency.  Thus, a mass balance based on BOD won't ac-
count for the potential transfer which would occur had the test
been performed at zero DO.  In order to compute the potential
transfer, it is necessary to_determine beta and the average DO in
the basin.  The average DO, C  , can be determined by  testing re-
presentative volumes of the basin during the test period.  The
beta will have to be measured using procedures described elsewhere.

                               402

-------
  The potential BOD removal can then be  increased by  the  ratio  of
the oxygen solubility to the oxygen deficit which exists  during
the test.  In order to calculate the credit, a  sampling program
will have to be developed based on the temporal variation in  DO at
various points in the basin.  Daily sampling at key locations in
the basin may be required, especially if BOD load to  the  basin
fluctuates markedly.  In systems where the photosynthetic pro-
duction of DO is a factor, the mass balance may have  to be de-
termined on an hourly basis in order to properly proportion the
excess DO to the aeration system.
     2.   Temperature Correction - As mentioned under the minimum
DO prodecure, it is K a which is oxygen dependent.  Because K a
isn't a factor in the mass balance calculation, a temperature
correction isn't required.  The effect of temperature will be
included in the average DO in the basin.  Again, this is offset  in
the calculation by the change in saturation provided 0 = 1.024.

     3.   Nitrification-Denitrification Factor - Because excess  DO
is present in the system, nitrification is expected to proceed in
systems treating significant quantities of nitrogen.  In activated
sludge systems treating sewage, nitrogen oxidation may account for
5-25% of the total oxygen demand.  Thus, it is reasonable to expect
that the change in the nitrogen oxygen demand (NOD) across the
long-term system may be significant where BOD:N ratios are less
than 16:1 in the influent.  In this case, this balance will have
to be computed in terms of oxygen equivalents.

Limitations of the Mass Balance Approach

     Additional attempts at using the approach have also been
somewhat successful in activated sludge plants.  More trials which
incorporate the method will undoubtedly produce more knowledge of
its limitations.  The following represent some of the circum-
stances which may limit the application of the procedure.

     First, the test period should be long enough to reduce ef-
fects of short-term fluctuation in load, performance, and weather.
This is especially important in long term systems.  In order to
overcome this limitation, it is recommended that a period equal  to
three theoretical detention times, (V/Q), be used for averaging
the data.

     Second, the procedure assumes that the system is at equili-
brium during the test period.  Thus, the system should not be
accumulating or releasing BOD which cannot be readily accounted
for.  The loss of settleable solids may be a sink for BOD in cer-
tain systems.  Thus, the method is limited to systems where the
mixed liquor contents contain less than 0.1 ml/hr settleable solids
throughout the test period, or, where the oxygen sink associated
with settleable solids can be computed.

     The last limitation presented pertains to the fact that the
mass balance method hasn't been employed in these cases where the
lagoon is photosynthetically active.
                               403

-------
THE INERT GAS TRACER METHODS


     The  inert  tracer methods of oxygen transfer measureitie^on.
allows the in situ measurement of operating wastewater treatment
system oxygen transfer equipment without interference from nyaro-
dynamic, biological or chemical processes.  The method is an out-
growth of the laboratory and field investigations of the  inert
tracer method for stream reaeration measurement by Tsivoglou, £t
ajl. (11,12,13).  The technique relies on three separate tracers,
an inert gas tracer to measure gas transfer, an inert dissolved
tracer to measure dispersion, and a fluorescent dye tracer to
track the other two tracers.  Because the gas tracer is inert, the
technique may be applied to biologically and/or chemically active
systems, like respiring wastewater treatment plants.  The use of a
dispersion tracer allows gas transfer measurements to be made
under the full range of hydraulic conditions, from completely
mixed to plug flow.  The inert  tracer method is applicable to all
forms of oxygen transfer equipment, coarse and fine bubble dif-
fused air, turbines and mechanical surface aeration.

     Because of its expense, the inert  tracer method when fully
developed is currently being viewed as a referee technique for
other, less costly methods.  Current research by NCASI and the
Amherst College Department of Physics is directed at (a)  reducing
the cost of the inert  tracer method, (b) determining the pre-
cision and accuracy of the technique in a variety of applications
and (c) developing the dosing, sampling and analytical protocols
needed to carry out the test.  The  inert tracer method is based
on several theoretical considerations in turbulent gas transfer
between air and water.  A common methematical expression used to
described gas transfer in wastewater systems is:
                     _= «
-------
Equation 1 is applicable to all sparingly soluble gases.   It  is
also applicable to transfer both into the water  (C(t) < C  ) and
out of the water (C(t) > C).                            s
                          s

     Based on an extension of Einstein's law for the diffusion of
particles within a viscous medium  (14) it is possible to demon-
strate that the rate of molecular diffusion within a single medium
is inversely proportional to the radius of the diffusing molecule
(15).  Applying the assumption that the transfer of sparingly
soluble gases into and out of turbulent water is limited by dif-
fusion in the water phase, it should be possible to relate the
relative transfer rates of various gases to their respective mole-
cular radii.  This relationship has been made both theoretically
and empirically by Tsivoglou, e_t al. for several molecular and
atomic gases, including oxygen (11,12).  Additional relationships
of this type have been demonstrated for several light hydrocarbon
gases and oxygen (16).

     Based on the above theoretical considerations it is possible
to relate the transfer of an inert gas tracer (r = 0 in Eq. 1) to
the transfer of oxygen in a respiring wastewater treatment system.
By super-saturating the system with the inert gas tracer and ob-
serving the decay (transfer out of the water) of that tracer to
equilibrium,°CKLa for the tracer gas may be determined.  Then by
means of the theoretical and empirical relationship between the
gas tracer transfer rate and the oxygen transfer rate,otKLa for
oxygen in the system may be calculated.  Because a dispersion
tracer is also used, the calculated °^KLa n\ay also be corrected for
any hydraulic dispersion in the system.

     Several gas and dispersion tracer combinations have been
proposed and used for  inert tracer measurements of gas transfer.
The original tracer combination developed by Tsivoglou, et al.
(11,12,13) for stream reaeration work was the radiotracer  krypton-
85 as the gas tracer and tritiated  (H-3) water as the dispersion
tracer.  Both of these tracers may be sensitively measured by
liquid scintillation counting  (17).  This tracer combination  has
been used in wastewater treatment  systems by Neal and Tsivoglou
(18) where it was demonstrated that the  inert tracer methodology
may be applied to both surface aerators and diffused air systems.

     The USGS has developed a non-radioactive tracer system for
use in stream reaeration studies (19,20).  This  tracer system used
low molecular weight hydrocarbons  (either ethylene or propane) as
the gas tracer.  These hydrocarbons are measured by gas chromato-
graphy (21).  In this technique the fluorescent dye is also used
as the dispersion tracer.  There has been concern voiced about the
non-conservative nature of the fluorescent dye tracer and  the
biological activity of the hydrocarbon tracer gases (22).  The
USGS tracer combination has not yet been applied to wastewater
treatment system aeration equipment.

     NCASI is currently developing a non-radioactive tracer system
for use in streams and wastewater  treatment systems.  The  gas
tracer in this system is atmospheric krypton gas.  The krypton
will be initially measured by isotope dillution mass spectrometry
                               405

-------
The conservative tracer in the NCASI research is lithium chloride
which will be measured by atomic absorption spectroscopy.  T£1fh_
tracer combination is expected to overcome the limitations or tne
USGS tracer system, and build upon the established research or
Tsivoglou with other forms of krypton gas.

     The protocol of Neal and Tsivoglou (18) for the measurement
of wastewater treatment system aeration equipment provides a good
example of the direct tracer procedures.  The treatment system was
dosed with a well-mixed volume of krypton-85, tritiated water and
fluorescent dye.  A bottle containing the tracers was submersed in
the treatment system and shattered by means of an electrically
actuated, attached blasting cap.  Just prior to dosing, samples
were collected in the basin to establish background levels for the
liquid scintillation counting.

     Sampling in the basin was conducted for about one hour after
the tracer dosing.  In the early treatment system studies, a flow-
through fluorometer was used to detect the fluorescent dye tracer.
Samples for scintillation counting were collected at the outlet
from the fluorometer flow.  Later, in the studies, it was dis-
covered that the fluorescent dye tracer was not needed in highly
mixed treatment systems.  In the later studies, sampling was ac-
complished with a manual grab sampler.

     The gas and dispersion tracers were analyzed in the labora-
tory using a three channel liquid scintillation counter.  Sample
counts were standardized against an automatic external standard.
For each field sample, two 2 ml replicate counting samples were
counted three times apiece, resulting in six replicate counts per
field sample.

     In order to calculate the gas transfer rate coefficient from
the Kr-85 data it was necessary to correct that data for hydraulic
dispersion.  This was done by using the ratio of the Kr-85 to H-3
concentrations (R) for each station, rather than the Kr-85 concen-
tration alone in the rate coefficient calculations.  This effect-
ively corrected the gas tracer data for dispersion.  By preparing
a semi-log plot of the ratio, R, versus the sampling time, t, the
krypton gas transfer rate coeff icient, 
-------
The <*KLaK  arrived at  from Equation 2 can  then be used to cal-
uculate  tne °*KLaox f°r oxygen by using  the gas transfer rate
ratios mentioned earlier.   Tsivoglou  (11,12)  has found that the
ratio between krypton  and  oxygen transfer  rates is:


                     KLaKr/ Vox = °'83 ±  °'04

     Based on the protocol outlined above  it  is therefore possible
to estimate the oxygen transfer coefficient in respiring waste-
water treatment systems.   Similar protocols would be applicable  to
the  other, non-radioactive tracer combinations like the one under
development by NCASI.  Based on the work by Neal and Tsivoglou (18)
it appears that the direct tracer method should serve as an excel-
lent referee technique for testing oxygen  transfer measurements  of
all  types of aeration  equipment.  NCASI's  efforts are directed
toward reducing costs  thus allowing the method to be more univer-
sally available and perhaps become the method of choice in testing
aeration systems.


SUMMARY

     In an effort to  indicate  emerging directions for evaluating aeration
systems, two methods have been  discussed.  The  first is inexpensive and uses
a mass balance  in the  oxidation state of the system to arrive at mass transfer.
This  method has been applied  to aerated stabilization systems treating paper
industry wastewater and is most suitable for longer detention time systems.
The second method is an inert gas tracer method which is applicable to almost
any system but  which may be expensive compared  to other methods.  The gases
used have been  radioactive krypton, ethylene and propane.  A method which
uses  stable isotope krypton is  currently under  development by NCASI working in
conjunction with the Physics  Department of Archerst College.
REFERENCES


(1)   McKeown, J.J. and Buckley, D.B.,  "Mixing Characteristics of
     Aerated Stabilization Basins, TAPPI,  54, (10), 1664  (1971)

(2)   Benedict, A.H.  and McKeown, J.J.,  "Oxidation Analysis  of
     Mill Effluents,"  Stream Improvement Bulletin, No.  256,
     NCASI, 260 Madison Ave., N.Y.,  33  pp  (May 1972)

(3)   "A Manual of Practice for Biological  Waste Treatment  in the
     Pulp and Paper  Industry," Stream  Improvement Bulletin No. 214,
     NCASI, 260 Madison Ave., N.Y.,  115 pp (April 1968)

(4)   McKeown, J.J.,  Benedict, A.H. and  Locke, G.M., "Studies on
     the Behavior of Benthal Deposits  of Wood Origin,"  WPCF, 40
     (82) R333  (1968)
                                 407

-------
 (5)   Whittemore, R.C. and McKeown,  J.J,  "Interfacial Velocity
      Effects on the Measurement of  Sediment Oxygen Demand,
      Stream Improvement Bulletin No.  317,  NCASI,  260 Madison
      Ave., N.Y., 31 pp (Nov. 1978)

 (6)   Whittemore, R.C. and-McKeown,  J.J.,  "Further Studies of Sedi-
      ment Oxygen Demand Measurement and  Its Variability,   Stream
      Improvement Bulletin,  No.  321, NCASI,  260 Madison Ave., N.Y.,
      25 pp (Mar. 1979)

 (7)   Klein, L., River Pollution, III  Control,  Butterworth,  London,
      p. 136 (1966)              "           ~~


 (8)   McKeown,  J.J., Buckley, D.B.,  and Gellman, I.,  "A Statistical
      Documentation of the Performance of Activated Sludge and
      Aerated Stabilization Basin Systems Operating in the Paper
      Industry," Proc. 29th Ind.  Waste Conf., Purdue Univ. (Pt.  2),
      p. 1090 (1974)

 (9)   Alferova,  L.A.,  Skirdov, I.V., Ponomarev,  B.M.,  Gladkov, V.A.,
      and Rogovskaga,  I.,  "Sewage Treatment  in  the Northern  Areas
      of the U.S.S.R., "Report on Int. Symp.  on Wastewater TMT.  in
      Cold Climates,"  Env.  Canada Report EPS 3-WP-74-3,  p. 64
      (Mar. 1974)

(10)   McKeown,  J.J.  and  Benedict,  A.H., "The Effect of Temperature
      on Treatment Plant Performance and Related Temperature
      Studies,"  Stream Improvement Bulletin  No.  312,  NCASI,  260
      Madison Ave.,  NY,  NY,  64 pp (May 1978)

 (11) Tsivoglou, E.G., O'Connell, R.L., Walter, C.M., God-
      sil, P.J., and Logsdon, G.S.,  "Tracer Measurements in
      Atmospheric Reaeration - 1. Laboratory Studies,"
      JWPCF 37,  ;1343, (1965).

 (12) Tsivoglou, E.G., "Tracer Measurement  of  Stream  Reaeration,"
      Federal Water Pollution Control Administration,  U.S.
      Department of the Interior, Washington,  B.C.  (June, 1967).

 (13) Tsivoglou, E.G., Cohen, J.B.,  Shearer, S.D., Godsil,  P.J.,
      "Tracer Measurement of Stream Reaeration. II.  Field
      Studies", JWPCF 40, 285 (1968).

 (14)  Einstein,  A.,  "Investigations  on the Theory  of  the
      Brownian Movement,"  Ann. Physik  1/7:549, (1905).

 (15)  Mysels, K.J.,  "Introduction to Colloid Chemistry,"
      Interscience  Publishers, Inc., New York,  (1959).
                                408

-------
(16)   Rathbun, R.E.,  Stephens, D.W., Shultz, D.J., Tai, D.Y.,
      "Laboratory Studies of Gas Tracers for Reaeration," JEED
      ASCE 104, EE2:215, (1978).

(17)   Cohen, J.B., Setze^, J.L.g5Kelley, W.D., Shearer, S.D. Jr.
      "Determination  of  H and   Kr in Aqueous Samples by Liquid
      Scintillation Techniques," Talanta 15:247, (1974).

(18)   Neal, L.A. and  Tsivoglou, E.G., "Tracer Measurement of
      Aeration Performance," JWPCF 46; 247, (1974).

(19)   Rathbun, R.E.,  Shultz, D.J., and Stephens, D.W., "Preli-
      minary Experiments with a Modified Tracer Technique for
      Measuring Stream Reaeration Coefficients," U.S.G.S.
      Openfile Report 75-256, Bay St. Louis, MI, (June 1975).

(20)   Rathbun, R.E.,  and Grant, R.S., "Comparison of the
      Radioactive and Modified Techniques for Measurement of
      Stream Reaeration Coefficients," U.S.G.S., Water-
      Resources Investigations 78-68, (May, 1978).

(21)   Shultz, D.J.,  Pankow, J.F., Tai, D.Y., Stephens.  D,W.
      and Rathbun, R.E., "Determination, Storage, and Pre-
      servation of Low Molecular Weight Hydrocarbon Gases in
      Aqueous Solution," Jourt Research U.S. Geol. Survey 4,
      2:247, (1976).

(22)   Hovis, J.S., Whittemore, R.C., Brown, L.C., and McKeown
      J.J., "An Assessment of the Measurement Uncertainty in
      the Estmation of Stream Reaeration Rate Coefficients
      Using Direct Tracer Techniques," presented at the SWMM
      and Water Quality User's Group Meeting, U.S.E.P.A.,
      Washington, D.C., (March 25-26, 1982).


DISCLAIMER

     The work described in this paper was not funded by the U.S.
Environmental Protection Agency.  The contents do not necessar-
ily reflect the views of the Agency, and no official endorse-
ment  should be inferred.
                              409

-------
NEW DIRECTIONS - OFF GAS METHODS

Lloyd Ewing
Ewing Engineering Company
Milwaukee, Wisconsin  53209


     The rate of progress in aeration technology that the agenda
of this workshop implies is indeed impressive and encouraging.
With the projected long term increases in power costs, the urgent
need for such progress is apparent.


     Some of the areas in which progress is evident, and in which
our company has been active are the following:

       1.)  Development of tests and proceedures for character-
            ization of porous diffusers prior to, and following
            use, in-situ, as well as in the laboratory.

       2.)  Development of diffuser specification requirements
            that will yield improved performance and economy.

       3.)  Development of non-process interruptive cleaning
            proceedures for ceramic diffusers.

       4.)  Development of equipment and techniques sufficiently
            sensitive to predict the rate of fouling of porous
            diffusers and permit the application of cleaning
            cycles at closer to the optimum frequencies.

       5.)  Development of analytical equipment and techniques
            directed towards identification and control of
            causes of porous diffuser fouling.

       6.)  Development and application of gas phase techniques
            to evaluate the efficiency of diffused aeration
            equipment under field conditions:  the so called
            Off-Gas approach.

       7.)  The assembly of application data for various diffuser
            systems through dirty water field testing.


     This paper is intended to serve primarily as a progress
report of our experiences in off gas work since it was reported
by Redmon and Boyle (1), at the 1981 Fall meeting of the ASCE
Oxygen Transfer Committee.  Since their report did not receive
wide circulation, it is attached as an appendix to this paper.
                              410

-------
     We continue to believe the importance of obtaining reliable
methods for dirty water testing cannot be exaggerated.  The
value lies in design as well as system evaluation.  Coupled
with solid clean water data for the system, a reliable dirty
water test provides a direct approach to the estimation of
alpha.


     Between the fall of last year and the present,  the off gas
equipment has been employed in six tanks at the Milwaukee Jones
Island Plant and three at the Milwaukee's South Shore Treatment
Plant.   In all of these tanks the ceramic diffuser  plates
are disposed more or less uniformly and at a ratio of from
.2 to .25 of the total tank area.  These tests disclosed
specific transfer efficiencies more or less comparable with
the full floor coverage ceramic data reported earlier by
Redmon and Boyle, (Appendix).

     In November of 1981, one tank at the wastewater treatment
plant at Berlin, Wisconsin, was tested.  This plant  has Sanitaire
full floor coverage ceramic discs at a ratio of diffuser to
tank area of about .08.  Specific efficiency of this sytem was
found to be comparable to the Whittier Narrows plant reported
in Appendix 1.

     Also in November of 1981, the wastewater treatment plant
of a Wisconsin pulp and paper mill, was tested.  This plant
is equipped with Sanitaire coarse bubble diffusers in a grid
configuration.  Although the specific efficiency was roughly
comparable with the Madison data in Appendix 1, conclusions
of system equivalence should not be drawn without taking  into
account the differences in configuration, air rate,  type of
waste, etc.

     In December, 1981, tests were conducted in parallel with
the radio-active tracer method, as described by Neal and Tsivoglou
(2), in one of E iDuPont's industrial wastewater treatment
facilities.  Hugh Campbell has reported this work in somewhat
greater detail in Session 3 of this workshop.  This  plant
was originally fitted with Kenics coarse bubble aerators,
the helixes and tubes of which had since been discarded.
At transfer efficiencies, (OTE), of around 7%, the offgas and
radioactive tracer method agreed within 5%, using either the
captured gas rate or the measured applied air rate.   This and
the above experiences, served to enhance our optimism regarding
the accuracy and utility of the method.

     The fact that correlation between the two methods requires
accurate knowledge of air rate suggested an interesting thought.
                               411

-------
In general, liquid pahse methods based on respiration rates,
tracers, etc., will yield K.  without knowledge of air rate.
Off gas methods, on the oth&T hand, will yield OTE without
knowledge of air rate.  In any system under fixed conditions
of submergence, temperature, etc., K,  is usually approximately
linear with air rate, whereas,'under Similar constraints, OTE
is comparatively insensitive to air rate.

     Thus, if the object of the test is to predict air requirements
for a given oxygen demand, the off gas method provides the
advantage of being far less sensitive to errors in measurement
of air rate of the tests than the liquid phase methods are, where
any lack of precision in air rate measurement results in a
corresponding error in prediction.


     During the months of May and June, 1982, eleven tests
were conducted at the Munincipal Wastewater Treatment Plant
in Brandon, Wisconsin.  This plant is a small complete mixed,
activated sludge plant, fitted with Aerocleve jets.

     The objectives of these tests included evaluation of the
precision of the method as well as the correlation between it
and the continuous steady state respiration rate method.
Results in both regards are presently considered to be inconclusive.
One unexplained difficulty, not previously encountered, was
the consistent disparity between estimated air rate applied
and measured air rate collected. Sixteen tests prior to the
Brandon work indicated a mean ratio of air supplied to air
collected of 1.05 with a standard deviation of approximately
.25.  The eleven tests at Brandon yielded comparative values
of .81 and .15, respectively.  Three different gas sampling
patterns were employed at Brandon for collection, and the flux
obtained by each were surprisingly similar.  Applied air was
estimated from the PD Blower curve, and errors of the magnitude
observed are not to be expected.


     The air rate was not the only anomaly observed.  Whereas
the transfer efficiencies measured in the first eight tests were
comparable with those obtained at Whittier Narrows by Redmon
and Boyle (1), the last three  tests yielded significantly
lower efficiencies. It is our present belief that some, if
not all of this inconsistency may have resulted from air leaks
into the analytical circuit during some of the tests.  Leak
tests involving passing Nitrogen through the instrument prior
to offgas testing are now routinely applied.
                                412

-------
     On July 21, 1982, parallel tests were run in a walker
process plant in Ridgewood, New Jersey, with the steady state
test, the dual non-steady state test, and the non-radioactive
tracer test.  This data has not been fully analyzed, however,
there is reason for encouragement in that good agreement appears
to exist between the offgas and the dual non-steady state
methods.  This encouragement is heightened by the fact that
the actual transfer rates during the offgas tests were around
3%, which has heretofore been considered outside of the practical
range of the method.

     Jeff Brochtrup, a graduate student at the University of
Wisconsin, has developed a program for theoretically  evaluating
the precision of the method from assumed estimates of variation
of the principal variables.  On the basis of our first assumption,
the estimates of precision obtained from the program have
been vary favorable.  We are presently revising our initial
estimates of variation based on the field experience to date
and plan to report these findings when complete.

     The experience with this equipment over the past eight
months has resulted in some new conclusions and some modifications
to those reported last October.

     The theoretical basis for our confidence in the precision
of the method has been expanded, but unfortunately, experimental
support therefore has not consistently been forthcoming.

     This method appears to be applicable to plug flow and
non-uniformly loaded and aerated systems with significantly
less error source than the liquid phase systems.

     In complete mix systems, the errors resulting from non-
representative gas sampling do not appear to be as great as
originally expected.

     The sampling errors that may occur can be reduced, and the
facility of the testing improved, by design modifications to the
sampling hood.

     Further experience with the Leeds and Northrup Oxygen
Sensor and one other polarographic unit, (Teledyne Analytical
Instruments, Model 320), have given indication that compensation
for temperature, pressure, and humidity may best be made through
empirically determined relationships using gases similar in
composition to offgas and ambient air.  To date,  we have not
developed all the data necessary to define the empirical
relationship required for the instruments we have used.  We
are also quite sure that other types of instruments may employ
different empirical relationships.
                                413

-------
     The practical method that we have used, and recommend to
others, is to manage the reference and offgas streams to
obtain as closely as practical, the same velocity, humidity,
temperature, and pressure.  We believe it is practical to
achieve this control within acceptable limits without unduly
complicating the equipment.

     It has also become apparent that different sensors will have
different general requirements.  For example, some may require
air at or near saturated humidity, whereas, others cannot
tolerate humid air and resultant - condensation.

     We believe periodic standardization or qualification at
reduced pressures and/or with reference gases of known composition
is desirable.

     There is preliminary indication that the method may be
suitable for reliably measuring  oxygen transfer rates at
levels of less than 5%.

     We are not, at this time, in a position to specify
the optimum sensing device or the ancillary equipment to go
with it, but we do believe the system we have used, described
by Redmon and Boyle (1), is workable, and can be further
developed into a simple, useful tool for diry water testing.
     REFERENCES

     1.  David Redmon and Wm. C.  Boyle,  "Preliminary Findings:
         Offgas Analysis."  A report to the ASCE Oxygen Transfer
         Standard Committee, Detroit, Michigan.   Presented:  Oct.
         6, 1981.  Revised:  Oct. 30, 1981.
     2.  Neal, L.A. and E.C. Tsivoglou.   "Tracer Measurement of
         Aeration Performance."  Journal Water Pollution Control
         Federation, 46, p.247, 1974.
 DISCLAIMS!:

      The work described in this paper was not funded by the U.S.
 Environmental Protection Agency.  The contents do not necessarily
 reflect the views of the Agency, and no official endorsement
 should be inferred.
                                 414

-------
    APPENDIX I- PRELIMINARY FINDINGS:  OFFGAS ANALYSIS
                 BY:   David Redmon and Wm. C. Boyle

                        A REPORT TO THE
            ASCE OXYGEN TRANSFER STANDARD COMMITTEE
                       Detroit,  Michigan
                 Presented:  October 6, 1981
                   Revised:  October 30, 1981

Background
     The ASCE work toward an Oxygen Transfer Standard has emphasized
the need for a simple, reliable  dirty water test.  Without it,
results of the considerable effort in clear water testing lose
much of their significance and utility.
     There also exists a need for evaluative dirty water procedures
that reveal aerator performance  for specific locations within an
aeration basin and that may be simply applied to tanks that are
non-uniformly loaded and/or aerated.  Information on alpha and its
variability under various conditions would also be an important
side benefit of such a method.  Hopefully, the method used could
be simple and fast and relatively unaffected by transient conditions,
as well as non-disruptive of plant operating conditions.
     Consideration of error sources and the sensitivity of various
proposed dirty water test alternatives indicated that the offgas
method was theoretically quite promising, especially for percent
oxygen absorptions in excess of  five percent in wastewater,
providing precise measurement of differences of oxygen and nitrogen
plus inert mole fractions between plant air and offgas could be made.
     The above factors led to an attempt to develop the offgas
method as a means of evaluating  dirty water performance of
diffused aeration systems.

State of Development
     Two practical problems had to be overcome to make the method
feasible.  One was the need for a comparatively large area offgas
                                 415

-------
collection device which was light and easy to handle and with
which a reasonably representative gas sample could be obtained.
A conceptual design of such a collector was presented at last
year's meeting in Las Vegas.  Subsequently, a prototype, weighing
approximately 200 Ibs., with a sample area of 8.75 ft. x 2 ft.,
was built and tested and found to be quite acceptable.
     The second problem was to obtain an oxygen sensor which could
precisely detect small differences in the partial pressure of
oxygen and be adaptable to in situ measurements.
     Our investigation of oxygen sensors, although not exhaustive,
did consider paramagnetic sensors, gas chromatographs, and polaro-
graphic probes.  For various reasons the Leeds and Northrup
polarographic probe was selected for trial.  This probe, due to
its design, requires no transport of oxygen across the membrane
when in equilibrium with surrounding conditions; therefore, the
presence of water droplets or foulants on the membrane, as well
as the gas velocity pass the probe do not affect the readings as
they can with the Clark type polarographic probes.
     A schematic diagram of the offgas analyzer including the in
situ oxygen sensor is shown in Figure 1.  In general terms, the
offgas sample is drawn through a flowmeter in order to determine
the rate of offgas collection over the hood.  A small portion of
this sample is drawn through an analytical circuit in the analyzer
which includes the oxygen sensor.  Gas temperature and pressure
are monitored and controlled to approximately constant values in
the analytical stream so that the difference in the partial pressure
of oxygen between main air and offgas may be precisely obtained.

Theory
  Calculations:
     The percentage of oxygen absorption may be obtained by
determining the mole fraction of water vapor, carbon dioxide,
oxygen and inerts (including nitrogen) in both the reference
air streairi and in the offgas stream.  The percent oxygen transfer

                                416

-------
is determined using the following expression:

OTE -  °2 ln T °2 °Ut     Gi(MRo/i) VMi - 6iCMRog/i) Mo/Mi
          °2 in        "            Gi(MRo/i} VMi
in terms of mole fractions:
            Y (R)                      Y (OG)
             o           —              o
       1-Y (R) - Y (R)       1-Y (OG) - Y0(OG) - Y (OG)
          O	W	O	CO 2	W	
                         YQ(R)
                     1-Y (R) - Y (R)
                        o       w

Derivation of these equations may be found in Appendix I.  See
the List of Abbreviations and Symbols at the end of this report
for definition of the terms used.
     In cases where the same conditions of temperature and pressure
for both the reference air and offgas are maintained, the mole
fraction of oxygen in the reference air times the ratio of the
voltage output of the probe for offgas, to the voltage output of
the reference stream is approximately equivalent to the mole
fraction of oxygen in the offgas stream.  In equation form:
                     Y (OG) =  Y (R)PV(OG)
     In using the analyzer in the field, reference air is drawn
past the oxygen sensor between each offgas measurement; thus one
has a reference check on the probe before and after each offgas
observation permitting a simple correction for probe drift, if
any is observed.  This technique reduces the errors in measuring
the difference between offgas and reference stream, and makes
the reading more precise and accurate.
     The linearity of the probe may be checked periodically by
drawing the reference air past the sensor under various levels
of reduced pressure while keeping the gas temperature constant.

                                417

-------
Under these circumstances the partial pressure of oxygen is
directly proportional to the total absolute pressure.  The criterion
that was applied was that the ratio of absolute pressures for the
two conditions divided by the ratio of the meter outputs should
fall in the range 0.995 to 1.005.
     Employing this test for linearity,it was found that the
calibration setting for probe output, in millivolts, was not
critical, and essentially the same relative change in the voltage
output occurred regardless of the setting with reference air at
ambient conditions.  To obtain the maximum sensitivity, the
reference output  should be as close to full scale as practical
since probe error is reported to be a fixed fraction of the full-
scale reading.  It should be emphasized that the above procedure
requires precise  measurement of the difference in the two gas
streams and is not necessarily an accurate determination of the
absolute value of the partial pressure of oxygen in either stream.

Sources of Error  and Assumptions
     As with all  models and methods, various assumptions are em-
ployed to simplify the analysis.  Some of the assumptions that
have been employed in this instance are as follows:
          1.  Nitrogen and inerts (principally Argon)
              are conservative; that is, that there is
              no  net absorption or desorption of the
              constituents in question.
          2.  Negligible de-nitrification is occurring
              at  the test location.
          3.  There is insignificant leakage of air into
              the analytical circuit of the analyzer.
          4.  The airflow rate to the basin is constant
              during the test.
          5.  The barometric pressure is constant  during
              the test.
          6.  The offgas is at 100% relative humidity
              at  the existing mixed liquor temperature.
                                 418

-------
     In addition to whatever errors are induced by the above
assumptions, other potential sources of error have been
identified, which include:
          1.  errors associated with the measurement
              of temperature,
          2.  errors associated with the measurement
              of pressure,
          3.  errors associated with the measurement
              of humidity in the reference stream,
          4.  errors associated with the measurement
              of dissolved oxygen in the mixed liquor,
          5.  errors associated with the measurement
              of carbon dioxide,
          6.  errors associated with the measurement
              of the rate of offgas collection,
          7.  errors associated with the lag in
              temperature compensation where
              significant temperature differences
              exist between the two streams,
          8.  oxygen absorption which could occur due
              to the presence of mixed liquor solids
              on the collector's internal surfaces
              and downstream piping, and
          9.  errors associated with neglection of
              surface re-aeration as may occur in
              uncovered tanks.
Some additional errors which can be compensated for are as follows:
          1.  D.O. meter drift,
          2.  changes in mixed liquor D.O., and
          3.  differences in submergence beneath and
              outside the collection hood, due to
              small differences in water level between
              the sample and the tank.
                                419

-------
     Experience to date has been that probe drift is minor after
the meter has warmed up a sufficient period of time.  One might
observe a 0.2 Mv drift during a test out of a 160-170 Mv output
for reference air.  Due to the short time intervals between
successive reference checks (typically 5-10 minutes) and the long
term measurement of this effect, a linear interpolation for
probe drift has been adopted.
     Differences in hydrostatic head underneath and outside of
the collection hood are at most a few inches of water, thereby
resulting in what is considered a very minor error in oxygen
absorption.

Test History and Results
     Prior to going to the Whittiex'-Narrows Plant CLACSD) in the
Los Angeles area,two trial runs of the equipment were conducted.
These initial runs, which were performed at the Milwaukee Jones
Island Plant and the Nine Springs Plant in Madison, Wisconsin,
were quite encouraging and provided an opportunity to make appro-
priate changes in technique and in operation of the analyzer.
     During the week of August 9, two full days of offgas
analysis were performed at Whittier-Narrows and, subsequently,
two additional days were spent at the Madison Nine Springs Plant.
Presently, there are no reliable estimates of the precision and
accuracy of the method, although it appears that it compares
favorably with other methods and may be within acceptable limits.
     While at Whittier-Narrows, a complete analysis of tanks 1
and 3 were conducted, as well as some additional data on tank 1
during the dual non-steady state test conducted by Jim Mueller.
Tank 1 was installed with a ceramic fine bubble grid system while
tank 3 had a jet system.  The nozzles were placed along one sidewall
and were directed across the tank facing the opposite sidewall.
Due to time limitations and operational problems associated with
tank 2, which was fitted with porous plastic tubes, no offgas .
data was obtained for this system.
                                420

-------
     At the Madison Nine Springs Plant, data on three types of
diffused aeration systems were obtained.  These systems include
ceramic fine bubble grid, a fine bubble tube system installed in
a cross-roll pattern, and a coarse bubble spiral-roll system.
During this investigation, respiration rate data were obtained in
parallel with the offgas method, the results of which will be
separately reported.
     The preliminary results of the offgas analysis for Whittier-
Narrows and Madison are presented in Table 1.  It should be
emphasized that these findings are reflective of the present
state of the art of the offgas method, and may or may not
represent the absolute performance of the aeration devices or
patterns employed.  The relative performance, however, may be
of significance.
     In order to compare the performance of percent oxygen absorbed
of the various systems tests on a more or less equivalent basis,
the parameter specific oxygen transfer efficiency, OTE  2Q, was
employed.  This value was calculated by dividing the percent
oxygen absorption by the estimated effective driving force.
These and other miscellaneous calculation procedures appear in
Appendix II.  At Whittier-Narrows, the dirty water saturation
         j.
values, C   , were estimated by correcting the clearwater values
          W/ V/
previously obtained for temperature, pressure and air rate, and
multiplying by an estimated value of beta; in this instance 0.95.
At Madison, they were obtained by similar interpolative corrections
to other clearwater data of systems comparable to those tested
at Madison.  A beta value of 0.95 was assumed in this case also.
     Since the clearwater C* value for various diffused aeration
systems are different, the specific OTE values are not always
directly comparable.  The upper portion of Table 1 indicates the
OTE     for each of the systems tested at Whittier-Narrows and
   sp20
Nine Springs.  Also shown is the range and progression of OTE
inlet to outfall.
     Knowing the specific clearwater OTE and the mean weighted
average specific oxygen transfer efficiency, OTE  2Q, based on

                               421

-------
the .offgas analysis, permits the computation of alpha.  It should
be obvious that if the estimate of the clearwater performance
is inaccurate that the resulting estimate on alpha will be
similarly inaccurate.
     As with the specific OTE, Table 1 indicates the mean weighted
estimate of alpha, its range, and its progression inlet to outfall.
It is interesting to note the values of alpha for the various
systems, as well as the relative ratios among the systems tested.
It is interesting to note that, in general, alpha seems to rise
as the waste approaches the aeration tank outfall.  It was ob-
served that a much greater apparent relative change in alpha
occurred at Madison as compared with Whittier-Narrows.  Much
additional work with this technique is required before definitive
conclusions regarding the absolute values and changes in alpha
with treatment can be established.
     The lower portion of Table 1 compares the applied air rate
in scfm per square foot of liquid surface to the measured rate
of offgas collection.  For the most part the measured offgas
collection rate agrees quite well with the estimated application
rate.  Since the offgas rates are used primarily to weight the
OTE results and are not needed for mass balance computations,
the differences that exist are not considered to be critical.

Questions and Anomalies
     As with most new approaches, various questions and anomalies
arise.  In an attempt to verify the offgas results, parallel gas
chromatograph tests were conducted at the Nine Springs Plant.
The agreement was not particularly good, and the reasons for this
have not as yet been fully determined.  Further parallel tests
are planned.  In  considering  all of the details regarding these
differences, it was felt that greater confidence should be placed
in the probe readings in this case.
     An additional question remains regarding the offgas results
obtained during the dual non-steady state test conducted at
Whittier-Narrows.  The test was conducted by observing D.O.,
                               422

-------
respiration rate, etc., at one air rate and then sequentially
another at about 2-3 times the magnitude of the first, and then
a third test at about the same air rate as the first.  Each of
these air rates were maintained until equilibrium conditions
were obtained.
     Lower than expected results were encountered by the offgas
method during the period of high airflow.  At present, it is
not clear why the anomalies in performance were found during this
test sequence.  They may be related to measurement errors, tran-
sient conditions or other unknown causes.  Further tests with
the offgas method under similar conditions will be undertaken
in the future to further explore this apparent discrepancy.

Tentative Conclusions
          1.  In the experiences with the offgas equipment
              to date, the method appears to give reasonable
              results in most instances.
          2.  The method appears to be of comparable or
              better accuracy and precision than other
              dirty water test methods currently known.
          3.  The method is simple and fast and relatively
              inexpensive with little, if any, process
              disruption.
          4.  This method is the only one known which
              measures performance locally in a tank.
          5.  The method can be used in basins with
              D.O. levels of less than 0.5 mg/1.
          6.  The offgas method can yield useful data
              to determine alpha and its variability
              throughout' the basin under typical
              operating conditions.
          7.  A few additional unanswered questions
              on the offgas method still remain.
                                423

-------
8.   The data in Table 1 is considered
    reflective of the present state of the
    art, and any conclusions drawn from
    it are considered tentative at best.
    Substantial additional work is required
    before firm conclusions should be
    drawn.
                   424

-------
               4-Way
                                                   Exhauster
                    Valve
             Thermometer
                             Bypass Circuit
                                                       &	
                             Analytical Circuit
                      6
                        0
                                                       -0	
                              Manometer
                                                         Flow-
                                                         meter
                       C02 Meter
                       Volumetric
                     0, Meter
                     Polarographic
Reference
  Air
                    Offgas to Analytical Circuit
                                                       3>	
                 H20 Meter
                 Hygrometer
                                   FIGURE 1
SCHEMATIC DIAGRAM OF OFFGAS ANALYZER
                                                             Offgas
                                425

-------
                                                               TABLE  1

                                                   SUMMARY DATA RE:   OFFGAS ANALYSIS
                                                      WHITTIER-NARROWS WWTP, CALIF.
                                                   AND MADISON NINE SPRINGS WWTP, WISC.
                                                        MADISON-NINE SPRINGS
                                                          WHITTIER-NARROWS
PARAMETER
OTE 2Q (weighted)
Range of OTEsp20
Order
Estimated Alpha
(weighted)
Range of Alpha
Order
Mean Specific Airrate
scfm/ft2
Overall Mean Ratio
Meas./Est. Applied
Range of Mean :
Ratio/Tank
V'ttn/im.Li'
GRID
FLOOR
COVERAGE
1.70
1.2/2.5
Rises in-
let to
outfall
0.64
0.42/0.98
Rises in-
let to
outfall
0.28
0.84
0.66/1.97
TUBES
SPIRAL-
ROLL
1.07
0.73/1.3
Rises in-
let to
outfall
0.62
0.46/0.85
Rises in-
let to
outfall
0.53
0.72
0.66/0.90
(JUAKOJCi
BUBBLE
SPIRAL-
ROLL
0.99
0.78/1.08
Uniform
1.07
0.83/1.19
Random
0.53
0.76
0.68/0.86
GRID
FLOOR
COVERAGE
1.11
0.91/1.36
Rises in-
let to
outfall
0.45*
0.35/0.60
Rises inlet
to
outfall
0.21
1.04
0.80/1.20
JfcT
AERATION
SPIRAL-
ROLL
0.89
0.74/1/11
Rises in-
let to
outfall
0.58*
0.48/0.72
Rises inlet
to
outfall
0.37
1.08
0.70/1.46
ro
01
                 NOTE:
                        OTE
OTE %
                           sp
            See pages 8 and 9 of test  for  further  development  of  this  term.

             *Estimated  for clean water test.

-------
                     LIST OF ABBREVIATIONS & SYMBOLS




C         Dissolved oxygen concentration in the liquid phase.



C*        Equilibrium dissolved oxygen concentration in the



          liquid phase.



C*ww      Calculated value of C* and equal to (BC*cw -C),  where



          B is estimated from the TDS of the mixed liquor.  C*
                                                              YfW


          represents a hypothetical C* presuming the mixed liquor



          has no uptake.



   Subscripts of C:



   cw   Refers to clearwater



   ww   Refers to wastewater



D.O.      Dissolved oxygen


G.        Mass rate of inerts (includes nitrogen,  argon, and



          remaining trace elements in air).


OTE       Percent oxygen absorption or oxygen transfer efficiency.



OTE       Specific oxygen transfer efficiency (OTE/deficit) .
   sp


M         Molecular weight of oxygen, O0,  32.0.
 o                                     ~


M.        Molecular weight of inerts, 28.1.


Mv        Millivolt output of oxygen sensor.


MR  ,.     Mole ratio of oxygen in reference air to inerts.



MR   ,.    Mole ratio of oxygen in offgas to inerts.
  og/i


Y         Mole fraction of oxygen.
Y         Mole fraction of carbon dioxide.
 co2


Y         Mole fraction of water vapor.
   Subscripts of YQ:



   (R)  Reference air



   (OG) Offgas
                               427

-------
                           APPENDIX II

Offgas Calculations:
Mole Fraction 02 in offgas, Y  (OG):

  Absol. Cell Press. Refer.    Absol. Temp. OG    x  Mv(OG)  x  y
  Absol. Cell Press. OG        Absol. Temp. Refer.   Mv(R)
Mole Fraction 00 in Reference Air, Y (R):
               ft                    O
                   = 0.2095(1-YW(R))               (1)
Mole Ratio of oxygen to inerts in reference or offgas stream:
                                        Y_
             = MR  ,. or MR
                                  1-Y0 -

For the reference stream, all Y     refer to reference air
                               o( R)
and similarly all Y 0/o/-\ refer to offgas mole fractions for the
offgas stream:
       0  in - 02 out
           2
OTE  =        MRo/i - MRog/i
                 MRo/i
Substituting the detailed expressions  for M.i and M   .^  and
simplifying yields the following approximate solution:
                                           Yo(OG)
OTE =	.	Cn~ '
                      1 Y        V
                         0(R)  ~  w(R)
                               420

-------
C*
 WWoo
                        APPENDIX  III





              OTHER MISCELLANEOUS  CALCULATIONS
          OTEsp20
                      WWoo
-  where clear water value in comparable array at  comparable
         air rate, at standard conditions, C*    is known:
                                            20oo
                             &     ft

                            C   - C "

                 p  + 29 92  20°°   b

                  s         ^
                 29.92(1 +  2Qo"-x b )         b
                                                 x  3
          which in most cases may be simplified with acceptable


          loss in accuracy to:
                      *     e;   sT
                     C      °     •

                      20oo 29.92 C*
                                 b
          where  3 « 1 - .05
            ft
     Where Cori  is not known, but where prior test data has
            2. U00

     indicated a depth factor as a fraction of submergence that


     corresponds to the saturation value, "d  , an estimate of


     C'    may be made:
      ww°°   •*


                   [d (12/13.5) + P ] C* x 3
          P     _    e	s   b


           ww°°  "          29.92




     where P  = station pressure;  C^ = saturation value book at


     temperature T;  d  = effective saturation depth at infinite


     time;  G  = mass ratio of 0? to be applied.





                               429

-------
APPENDIX IH(continued)

     Prediction of performance at similar air rates and the same
submergence may be obtained by computing or estimating C*   as
                                       0             °  WWco
above and correcting for mixed liquor D.O. and temperature as
follows:

     Oxygen Transfer =  OTE  on (C*  - C.JG 6Ct~20)
                           spzu   ww»   F  o

     a  may be estimated where clear water transfer in similar
configuration, air rate, and submergence is known as follows:

                             OTE  __
                       a =      sP20
                            OTECwsp20
                              430

-------
FUNDAMENTAL RESEARCH  IN BIOLOGICAL WASTE WATER  SYSTEMS FOR
ADVANCED DESIGN OF MECHANICAL AERATORS

Franz Nestmann
University  of Karlsruhe
Institute of Hydromechanics
Kaiserstrasse 12
75  Karlsruhe, West Germany

At  the  Institute of Hydromechanics an extensive hydraulic  model-system was built
for a systematic  research on  mechanical aeration  systems sponsored by the German
Ministry of Science and  Technology (BMFT).  The complexity of the flow and mass
transfer process in multiphase  systems demands comprehensive instrumentation. The
equipment  was chosen in order to measure the following quantities (see Fig.  1)
common to aeration  systems:
-   the  spatial  velocity field  and the turbulence distribution is determined by a
    Laser Doppler  measuring system.  Laser and photomultiplier are  mounted on a
    trovers  to operate in three orthogonal  planes.
-   the power input  is measured by  means of the torque  on  the shaft between aerator
    and  motor.
-   the characteristic of the  mixing time is detected  by  tracer.
-   the  local percentage of dissolved oxygen  during  the  unsteady experiment is mea-
    sured with electrodes.
-   the  local bubble size  distribution together with the local gaseous air-content are
    determined  in  a  capillary  laser  lightened  test section.
-   the rpm of  the 1  kW  motor can be selected  continuously and is measured by  a
    magnetic field  sensor.

The signal  output of all  the  above mentioned quantities is  collected,  digitized and
evaluated  in a desktop computer  (1). Up to now  the flow  field in aeration tanks
did  not find the adequate attention.  Of special interest are boundary conditions
like tank geometry  and  the  interaction  of the aerator and  its  geometric environ-
ment as well  as  scale-up effects  which are looked  into by conducting experiments
with models of different  sizes. Many of the  test series  will be compared to re-
sults of field experiments later on. Within aeration systems one differentiates bet-
ween the following single processes:
-   local energy  input by the aeroator
-   interface production and  its renewal by the  exchange of fluid  particles  on  the
    liquid side  of  the interface
-   the built-up of concentration  gradients Vc which is responsible for generating
    the mass transfer besides  the physico-chemical condition and the nature of  the
    interface.

The analysis of the complete aeration process shows  that kinetic energy is  necessary
to  induce  convective mass transfer. This energy input creates  a three-dimensional
fully turbulent flow field in  the  whole  basin.  In  order  for  the flow to remain fully
turbulent it is  necessary  to overcome the viscous  damping  by  sufficient Reynolds'

                                        431

-------
              mixing
              time
              characte-
              ristic
A: aerator
                                                                        photo multiplier
                                                           traverse operating in
                                                           three orthogonal planes
Instrumentation
View  of 5m3 basin
                               Figure 1.


                                        432

-------
stresses.  Production of turbulence is supported by continuous power input into the
system. Otherwise the turbulence level subsides  due to convection, diffusion and
dissipation.

The  continuous energy consumption  can be  described by means of an  energy spec-
trum with  the  dissipation rate £  (reduction  of kinetic energy per unit mass  and
time) defined as followed        (?
                                0
With  the help of £ the power input can be  determined by
                   PM = - E   respectively   PM= -1/2   -(u'2«V2 +w" )

PM  spatially  integrated over the basin volume is equal to the power input P  at
the aerotor shaft                                    N.  revo|ution speec|  i/s
                          P  - 2TC- N  • Mf         M^: to|.que at vertical  shaft  Nm

Laser Doppler  measurements  (2) at different cross sections in the basin (see Fig.  2)
give a  good inside view of  the  spatial distribution of the local  kinetic  energy
density. Figure 3 shows the  distribution of turbulence intensity v' for the  Karlsruhe
Aerotor  in vertical direction for a quater  basin.

The  above discussed results are typical for most  of the  mechanical aeration systems.
Three different sections can  be distinguished within the flow  field of the aeration
tank (see Figure 3):
I   The  middle part of the  basin around the aerator with a high level of turbulence
   which does not help the  aeration due  to lack of any bubbles.
II Intermediate section of  the basin  with  low  level of  turbulence intensity.
Ill Outer part close to basin walls where  bubble entrainment  due to a roller can
   be observed down  to the  bottom.  The  turbulence  intensities are smaller than  in
   I) because  there is higher damping  due to the bubble  intensity,

Figure 3 shows the multiphase flow areas contain the least amount entrained  energy.
The  results  show clearly that the high levels  of  turbulence do not help the disper-
sion  of bubbles. Figure 4  illustrates  the velocity vectors  for  the vertical  cross sec-
tions.  This  knowledge allows"to identify the basin areas liable to sedimentation  as
well as  the structure  of the  macro- vortices.                      '

To optimize an aerator system, a detailed research of the flow field  is necessary
because  the turbulent  flow structure  is the basis for  the mass  transfer. With  the
help of  the above described techniques several aeration systems have been syste-
matically looked  into  as well  as new  systems have  been developed.
                                         433

-------
' 1 ' ' •/<


roo
traverse
laser
_.\-
•d
j
j

-
•' -
l-.-M
Mr
LL
j;|if
~h
i
1

transmitter

1000
110
i
1
O
O
***


301



side







c
i
•

3
•4
I
|
-

i
j

:
P
r
basn


"Y1
L.L.
/ '
rr
mirror

1
\^
hotorriultiplier
eceiver side
>>;

'/
View of laser traverse
                                 Figure 2.
                                     434

-------
                          Karlsruher - aerator
                          operation: pushing
                          H = 0.7m  N  =1.331/s
                          h/(d-cosa)=0.5
turbulence
intensity scale
•	' 100%
center of basin         y-coordinate[cm]
Distribution of velocity fluctuations  in vertical direction(v')
                                Figure 3
                                                                        ///////,
                                      435

-------
                          Karlsruher - aerator
                          operation: pushing
                          H = 0.7m  N =1.331/s
                          h/(d-cosa)=0.5
velocity-scale
—•  =  0.2 m/s
 B
 I
/ - 1
/ V
/ :

^
X
^

—
— •-


! 31.5 58 8A.5 1
L 	 m, 	 t
 confer of basin         y-coordinate[cm]
 Velocity distribution in vertical planes

                              Figure A.
                                   436

-------
REFERENCES

(1)  Nestmann, F..et al.   Verbesserung der Technik mechanischer Belufter
                           unter besonderer Berucksichtigung des Elnsatzes In
                           Klaranlagen.  Bericht Nr. 591, Institut fur
                           Hydromechanik - Universitat Karlsruhe (1982).

(2)  Lang, C.              Turbulenzmessungen in einer Modellbelebungsanlange
                           mit Hilfe der Laser-Doppler-Anemometrie.  Diplo-
                           marbeit - Institut fur Hydromechanik - Universitat
                           Karlsruhe (1981).
DISCLAIMER

     The work described in this paper was not funded by the U.S. Environ-
mental Protection Agency.  The contents do not necessarily reflect the views
of the Agency, and no official endorsement should be inferred.
                                    437

-------
WORKSHOP ON AERATION SYSTEM DESIGN

D. W. Smith


     The discussion of the work session on aeration system design centered
around three basic themes:  the a factor, energy considerations in design,
and new aerator system developments.  A summary of each theme is presented
below.

Alpha Factor

     The   factor for adjustment of IL a from clean water to dirty water condi-
tions causes difficulty in making rational design decisions.  This is caused
by the difficulties in comparing the performance of different types of
aeration equipment in dirty water.  It is known that the degree of turbulence,
the nature of the contaminate and a number of physical components combine
together to make prediction and translation of aerator performance information
difficult.

     The question of how close must an estimate of a be for designs which are
to be at capacity in 20 years was passed.  The question., although not
answered, suggests concern with the level of credibility which must be
included in design.  It appears that the lack of or limited knowledge of a
leads to conservative design decisions so that, to the extent possible,
adequate oxygen is transferred to the system.  This type of concern leads to
a significant amount of effort being spent on specifications and in situ per-
formance testing.

     The conditions required for the occurrence of a values greater than one
were also reviewed.

Energy Considerations in Design

     The aeration system design must include the aeration chamber if the
performance of the equipment is to be maximized.  The more complete use of
control systems to provide turn down capability and optimum management of the
concentration of oxygen in the mixed liquor were noted as needed improvements.

     The primary concern is with the total consumption of energy in achieving
a desired degree of treatment.  Therefore, the control system must ultimately
include all components of the treatment facility.

     It was noted that the lack of knowledge about the control of unit pro-
cesses or the lack of ability to control them will impact more as the cost
of energy increases.  This means the municipalities and industries which
require the treatment facilities will have to pay more of their operating
funds.
                                     438

-------
New Aerator System Developments

     Three new approaches to aeration system design were presented.  A con-
tractor-activated sludge unit which was designed to use the energy in the air
carried into the tank.  The system consists of a series of cells between
disks on a horizontal shelf.  This unit may provide a low total energy cost
for the biological treatment step.

     The development work on a reverse flow turbine was described.  The unit
attempts to reduce the amount of energy wasted on limited or no value turbu-
lence.  Pilot plant studies on the unit in Germany are about to be started.
Model studies are underway.

     The third system described was a verticle oxidation ditch consisting of
a conventional plug flow tank with a horizontal baffle at mid-depth and a
mixing propeller.  It was stated that better use of the oxygen added was
expected in a prototype unit.
Participants

     D. W. Smith
     Sirikun Hopkins
     Ray Thacker
     Tony Seiler
     Michael Marozzi
     Frances D. Shepard
     Mark Moederer
     Steve Black
     Henryk Melcer
     Mary Anne Lowndes
     John Belken
     Bill Baumann
     Franz Nestmann
     Jerry Wren
     Robert Alldredge
     George Smith
     Michael Doyle
     James E. Owen
                                     439

-------
WORKSHOP ON OPERATION AND MAINTENANCE

D. H. Houck
     The operation and maintenance (O&M) workshop convened at 4:30 pm,
August 3, 1982, with 20 persons in attendance and D. H. Houck serving as
chairman.

     Initially, the group discussed the problems of icing and capsizing of
floating mechanical aerators in northern climates.  It was noted that the
spraying effect of these devices serves to cool the aerated liquid to the
point of freezing and consequently they cannot be operated during the coldest
weeks of the year.

     The discussion then turned to fine bubble dome and disc aerators and
their O&M problems.  Strong sentiments were expressed by the consulting engi-
neers and O&M specialists that the equipment lacks adequate strength in its
materials and fabrication, leading to excessive maintenance requirements.
Problems cited included failure of plastic parts, notably dome hold down bolts
and gaskets, pipe hold down straps, expansion joints, and tank mountings.  It
was noted that the equipment appears to lack adequate capability to adjust
for expansion and contraction brought about by temperature change.  The
suggestion was made that the design of the equipment follows the English
design, where temperature changes are much more moderate.  Equipment vendor
representatives pointed out that the clients were purchasing the equipment on
a low bid basis, causing the level of quality to be driven to the lowest
common denominator.  It was suggested that engineers need to specify more
precisely the degree the materials and "robustness" desired.  Also, the manu-
facturers should be required, and paid for, their direct involvement in the
installation of the equipment.  If possible, they should assume responsibi-
lity for the initial installation as is the English practice.  It was noted
that the U.K. experience is that failures tend to occur more frequently
during the first year a-ter startup,  and much less frequently thereafter.

     The session adjourned at 5:45 pm.


Participants

     Dick Brenner                 Read Warriner
     George Powell                Richard Jones
     Dave Redmon                  Bill Sturtevant
     Scott Anderson               Tim Kluge
     Rod Gross                    Bob Graf
     Bruce Jank                   Forrest Johnson
     Chay S.  Ching                Paul Cummings
     Jim Laughlin                 Bob Shevlin
     Tom Pokorsky                 Dan Houck, Chairman
     Gerry Rupke
     Glen Daigger
                                     440

-------
WORKSHOP ON AERATOR TESTING

W. L. Paulson
     This session was chaired by Dr. Robert Baillod, Dr. Hugh Campbell and
Dr. Wayne Paulson.  Twenty-six people participated in the 75 minute session.
The discussion was divided into three parts, 30 minutes for clean water
testing, 30 minutes for dirty water testing and 15 minutes for open topic
discussion.
CLEAN WATER TESTING

     The clean water discussion focused on the content of the "proposed clean
water standard" presented by Baillod and Paulson in Session 3 and included in
the printed documents for study at the Workshop.

Truncation

     Two questions related to truncation.  One question was "Why truncate the
1st 20% of the DO data and then run all the way to 98% of the DO at satura-
tion?"  The 2nd question was "Why deoxygenate all the way to zero DO when
truncating anyway?"  It was observed that the non-linear method is very
sensitive to data closer to the DO at saturation.  It was also noted that
data early in the run is desirable when evaluating residuals and data adequacy.
It has been reported that better quality testing results have been obtained
when the DO level in the test tank is depleted to zero for several minutes.
Some participants observed that they had conducted successful testing when the
DO level only reached 1 or 2 mg/L.  The "proposed standard" permits a lower
limit of 0.5 mg/L.

Location of DO Determination Points
     The location of DO determination points was discussed.  It was noted
that there is an emphasis in sensing an equal volume with each determination
point.  This is recommended due to the averaging of point ILa values in
determining the standard oxygen transfer rate.  There are spacing limitations
cited in the "standard" as well.  It was noted that additional DO determina-
tion points may be located in corners, near the floor of the tank and
elsewhere for evaluating what is occurring in those locations.  These points
would likely not be part of the oxygen transfer rate analysis.  It was
observed that if a given system has unusual mixing patterns, equipment
placement or geometry the number of DO determination points for the transfer
rate analysis should be increased.

Data Analysis

     The following question was asked regarding data analysis.  "What about
averaging the DO concentration from the several determination points at a
given time and make one calculation involving ICa and the oxygen saturation
value."  It was observed that this procedure would probably be correct.

                                     441

-------
Based on  the experience of  several participants, good data yields the same
result.   It was noted, however, that point ILa data is needed in evaluating
other aspects of the testing performance.

     Some concern was expressed regarding the calculation of the DO at satura-
tion versus measuring the value at the end of a test.  The measured value
approach  is common to many  of the participants.  It was observed that the
recommended non-linear analytical method uses a calculated value and that it
may vary  between DO determination points.
Cobalt  Source
      It was  recommended  that consideration be given to changing the stated
 requirement  of  reagent grade chemical to permit the usage of a technical
 grade.  This was  requested due to the high costs incurred in large field tests
 or extended  testing programs where reagent grade cobalt is used.  It was
 reported  that equivalent catalytic action has been observed when comparing the
 two  types.   This  change  in wording will be considered in the revision of the
 "proposed standard".
 DIRTY WATER

 Oxygen Uptake Measurement

      It was reported  that there has been as much as a 20% variation in oxygen
 uptake measurements when comparing the manometric versus the BOD bottle
 method.   There was considerable discussion in an attempt to analyze this
 discrepancy.  The significance is the impact on the dirty water procedure
 that  utilizes the oxygen uptake rate as part of the analysis for oxygen
 transfer performance.  This potential variation was noted with no resolution
 of  the cause.

 Off-Gas Technique

      A question was asked regarding the application of this method to plug-
 flow  tanks,  eg. What  DO concentration do you use?  Mr. Lloyd Ewing reviewed
 the development and application of the method to date in the ASCE study.   He
 also  observed that one can use a weighting method where each sample point
 would  have  a specific oxygen transfer rate, eg. oxygen transfer efficiency/
 unit of  driving force.

 Hydrogen Peroxide Method

     Several participants, including Dr. Kayser, shared their experience in
utilizing this procedure.  In general, there appears to be a positive reaction
to its application for all activated sludge systems.  It was observed that
the method does not create new errors when contrasted to tracer techniques.
                                     442

-------
Tracer Methods

     Several participants discussed the methodology and chemistry of the
krypton and tritium tracer technique.  Dr. Popel reported on the use of helium
in the Netherlands.

Evaluation of Existing Methods

     The writer asked the opinion of several participants who are active in
dirty water testing for their evaluation of the present methods.  They indi-
cated that in their opinion the top four procedures are the hydrogen peroxide
method, the radioactive tracer method, the off-gas method for diffused air and
the respiration rate approach.
OPEN TOPICS
Alpha Analysis

     Several participants participated in a review and discussion of the
current status of this area of study.  Dr. Boyle reported that at this time
the ASCE study is comfortable with the use of tall columns for predicting
reasonable full-scale alpha values for uniform density fine bubble diffusers.
Dr. Boyle and Dr. Kayser indicated that for other systems it is necessary to
develop a historical baseline of alpha values with full-scale testing on
generic systems in different types of wastewater conditions.

Detergent Method

     The inclusion of this method in the "proposed standard" and its inter-
pretation was discussed.  It was noted that the usage of detergent testing in
the USA is very controversial.  Experience with this approach is quite
variable.  It was observed that the procedure presented in the "proposed
standard" was developed from the method utilized by Dr. Boon and others in the
United Kingdom.

     Dr. Boon commented on the success of the approach in the United Kingdom
where it is utilized to simulate the approximate transfer performance one
might expect with domestic wastewater.  He reported that in compliance testing
one test is conducted with detergent added to clean water in addition to the
clean water testing requirement.  Dr. Goronszy reported that detergent testing
is also conducted in Australia.
Participants
     Gregory L. Huibregtse
     Bruce Hadsell
     James Marx
     George Hill
     Gordon Speirs
     J. Peter Jones
     Abigail Cantor
Mervyn C. Goronszy
David H. Fredericks
Norbert W. Schmidtke
Arthur C. Boon
Steve Reiber
Thomas E. Wilson
R. Gary Gilbert
                                     443

-------
Par ticipants (Cont inued)

     Russell Sullivan
     James Mueller
     John Masters
     Robert Arthur
     Stephen McCullough
     H. J. Popel
     William C. Boyle
     Lloyd Ewing
     Jeff Brochtrup
     Hugh Campbell
     Robert Baillod
     Wayne Paulson
                                      444

-------
WORKSHOP ON OXYGEN TRANSFER IN STREAMS

James J. McKeown
     The workgroup convened in the afternoon with Messrs. John Sulliven of
Wisconsin DNR, Edward Earnhardt of ELBA, Inc. and James J. McKeown of NCASI in
attendance.  The meeting lasted approximately two hours.

     The workgroup undertook a review of the applications where enhancement
of the oxygen concentration of receiving waters had been studied.  During the
discussion twenty-five such sites were identified (Attached A).  The workgroup
also listed the processes by which DO enhancement had been achieved
(Attached B) and discussed which processes were deployed at which sites
(Attached A).  The workgroup expects to seek additional experience on this
topic as well or validate the information assembled on the attachments.

     The workgroup also developed a list of factors which have played a role
in limiting the utilization of DO enhancement in streams (Attached C).  Also,
the workgroup knew of two recent situations where the use of oxygen enhance-
ment processes were permitted for meeting DO criteria during critical periods
in the year (Attached C).

     Finally, work was started on listing some of the technical needs which
the group believed necessary to advance the state of the art in applying DO
enhancement processes to streams (Attached D).  Although the methodology has
been in the development stages for decades, it is only now receiving more
consideration as a viable alternative to achieving water quality standards.
In many respects, its resurgence is a function of a reluctance on the part of
many communities and industries to install expensive advanced waste treatment
facilities which will only be needed infrequently.

     The workgroup also wanted to encourage readers to add to the list of
technical needs based on their own experience and perspectives.  As a
beginning, readers are encouraged to forward such information to:

     J. J. McKeown
     NCASI
     Department of Civil Engineering
     Tufts University
     Medford, MA  02155
                                     445

-------
(ATTACHMENT A)

ENHANCEMENT OF OXYGEN IN SURFACE WATERS

LIST OF RIVERS AND PROCESSES
1.  Black R.   SC
2.  Millers R.   MA
3.  Miami R.   OH
4.  Andoscoggin R.   ME
5.  Willamette R.   OR
6.  Upper Passaic R.   NJ
7.  Jackson R.   VA
8.  Chicago Canal   IL
9.  Trinity R.   TX
10. Mohawk R. Canal   NY
11. Naheola R.   AL
12. Pearl R.   LA

13. Fox R.   WI
14. Wisconsin R.   WI
15. Peripheral Canal   CA
16. Thames R.   UK
17. West Germany
18. Red Deer River, Alberta
19. Parish R.   LA
20. Neuse R.   NC
21. TVA Canals
22. Flambeau R.   WI
23. Las Colinas R.   TX
24. Nashua R.   MA
25. Penobscot R.   ME
Surface Aeration
Surface Aeration, Oxygen Diffusion
Surface Aeration
Surface Aeration, Downflow Bubble

Surface Aeration, Diffused Aeration

Diffused Aeration

Cascade
Side-Stream Oxygenation
Side-Stream Oxygenation, Turbine
  Venting, Oxygen Diffusion
Turbine Venting
Turbine Venting with Oxygen

Barge Oxygenation

U-Tube

Side-Stream Oxygenation

Diffused Aeration, Turbine Venting
Surface Aeration
Surface Aeration
Turbine Venting
 (ATTACHMENT  B)

METHODS AND  PROCESSES FOR DISSOLVING OXYGEN  IN  SURFACE WATERS


Methods

     Cascade
     Bubble  Diffusion, fine and coarse, air  or  oxygen
     Mechanical Surface Aeration
     Deep tube or U-tube bubble Contactors
     Pressurization
     Aspiration

Processes

     Dam Reaeration                     Side-Stream Oxygenation
     In-Stream Aeration, Oxygenation    Turbine  Venting
                                     446

-------
(ATTACHMENT C)

LIMITATIONS TO APPLYING PROCESSES FOR DIRECT OXYGEN ENHANCEMENT OF RIVERS


1.  Legal or Regulatory

     • prohibitions in specific states or regulations
     • Red Book (i.e. 110% saturation)

2.  Aesthetics

     • foam
     • slime
     • noise
     • undersightliness

3.  Navigational

4.  Engineering

     • protection at high flow for permanent installations

5.  Ecological

     • increased scouring or coagulation
     • alteration of biological habitats
     • promotion of slime


Permitted Applications

     1.  Chicago
     2.  American Can, Naheola
                                     447

-------
(ATTACHMENT D)

TECHNICAL NEEDS


     1.  Oxygen transfer coefficients of natural water.
     2.  Knowledge of methods to efficiently mix liquids under high pressure
         with water at atmospheric pressure without losing the oxygen
         dissolved at high pressure.
     3.  Optimize type and placement of sparge ring in turbine venting.

     4.  Information on the efficiency of introducing fine versus coarse
         bubbles into the U-tube or downflow bubble contactor.
     5.  Evaluation of embolysm impact on fish at various dissolved gas levels.

     6.  Demonstration of the cost effectiveness of the methods based on
         frequency of occurrence for selected rivers.
     7.  Investigate the procedure to allocate aeration costs among dis-
         chargers.
     8.  Document costs for Oxygen Production by PSA (est. $80/ton),
         cryogenic (est. $40/ton) and membrane separation (est. ?$/ton)
         methods.
                                     448

-------
                   APPENDIX A - SEMINAR/WORKSHOP PROGRAM

          AERATION SYSTEM DESIGN, TESTING, OPERATION, AND CONTROL

                              August 2-4, 1982

August 2, 1984
0800     Registration - The Wisconsin Center/Lakeshore Room
0850     Workshop Objectives
              W. C. Boyle
              University of Wisconsin
              Madison, Wisconsin
SESSION 1     Chairman - W. C. Boyle
0900     Introduction to Workshop - E. L. Barnhart
0945     Coffee
1000     Types and Characteristics of Aeration Systems - Panel
              Diffused Air - J. Wren
              Mechanical Systems - T. Rooke
              Oxidation Ditch Systems - G. Smith
1045     Translation of Clean Water to Dirty Water - W. C. Boyle and Michael
         Doyle
1115     System Scale Up - N. W. Schmidtke
1145     Discussion
1230     Lunch
SESSION 2:  Chairman - N. W. Schmidtke
1330     Design Protocols - A Global Perspective
              North America - D. Smith
              United Kingdom - A. Boon
              W. Germany - R. Kayser
              Netherlands - H. J. Popel
1215     Discussion
1245     Lunch
1400     Workshop Sessions
         Design - D. Smith (Chairman)
         Operation/Maintenance - D. Houck (Chairman)
         Control - J. Stephenson (Chairman)
         Testing - W. Paulson and C. R. Baillod (Chairman)
1730     Adjourn
OPTIONAL TENNIS TOURNEY

August 4, 1982
SESSION 4:  Chairman - W. C. Boyle
0800     New Directions
              J. A. Mueller
              E. L. Barnhart
              J. McKeown
              L. Ew i ng
0900     Coffee
0915     Summary of Workshop Sessions
1015     Summary of Impressions:  Design - Panel
              A. Boon
              R. Kayser
              D. Smith

                                     449

-------
1115     Discussion
1200     Wrap-Up - E. L. Barnhart
1230     Closure
Afternoon - Optional field trip to Madison Nine Springs wastewater
         plant to observe off-gas transfer test.
Subcommittee on Oxygen Transfer Standards meeting (open to conferees)
                                    450

-------
                  APPENDIX B - SEMINAR/WORKSHOP PARTICIPANTS

CONFERENCE LEADERS & PLANNING COMMITTEE
BAILLOD, C. ROBERT
PROFESSOR
MICHIGAN TECH.
DEPT. OF CIVIL
  ENGINEERING
HOUGHTON, MI
UNIV.
& ENVIRONMENTAL

          49931
BLACK, STEVE
SUPERVISOR
ONTARIO MIN. OF THE ENVIR.
WASTEWATER TREATMENT SECT.
135 ST. CLAIR AVE. WEST
TORONTO, ONTARIO       M4V 1P5
CANADA
416-248-3041

BOYLE, WILLIAM C.
PROFESSOR
UNIV. OF WISCONSIN
DEPT. OF CIVIL & ENVIRON. ENG.
1415 JOHNSON DR., RM. 2205
MADISON, WI              53706

CAMPBELL, HUGH, JR.
E.I. DU PONT DE
  NEMOURS & CO., ENG'G DEPT.
LOUVIERS BUILDING, RM. 1376
WILMINGTON, DE           19898
DOYLE, MICHAEL
REXNORD, INC.
BUILDING K
5101 W. BELOIT RD.
MILWAUKEE, WI
          53214
EWING, LLOYD
PRESIDENT
EWING ENGINEERING CO.
6200 N. 39TH ST.
MILWAUKEE, WI.
          53209
JANK, BRUCE E.
DIRECTOR
WASTEWATER TECHNOLOGY CENTRE
P.O. BOX 5050
867 LAKESHORE RD.
BURLINGTON, ONTARIO    L7R 4A6
CANADA
416-637-4740
BARNHART, EDWIN L.
PROFESSOR
CIVIL & MECHANICAL
  ENGINEERING DEPT.
SOUTHERN METHODIST UNIV.
DALLAS, TX                75275

BOON, ARTHUR C.
WATER RESEARCH CENTRE
STEVENAGE LABORATORY
ELDER WAY, STEVENAGE
HERTSFORDSHIRE          SGI 1TH
UNITED KINGDOM
                        BRENNER, RICHARD C.
                        U.S. EPA
                        MUNIC. ENVIRON. RESEARCH CENTER
                        26 W. ST. CLAIR ST.
                        CINCINNATI, OH            45268
                        513-684-7657

                        CHAMBERS, BRIAN
                        WATER RESEARCH CENTRE
                        ELDER WAY, STEVENAGE
                        HERTSFORDSHIRE          SGI 1TH
                        UNITED KINGDOM

                        ECKENFELDER, W. WESLEY, JR.
                        VANDERBILT UNIV.
                        P.O. BOX 6222
                        NASHVILLE, TN             37235
                        HOUCK, DANIEL H.
                        D.H. HOUCK ASSOCIATES
                        8905 GARLAND AVE.
                        SILVER SPRING, MD         20901
                        KAYSER, ROLF
                        TECHNICAL UNIV. OF BRAUNSCHWEIG
                        P.O. BOX 3329
                        3300 BRAUNSCHWEIG
                        WEST GERMANY
                                     451

-------
CONFERENCE LEADERS & PLANNING COMMITTEE
MC KEOWN, JAMES J.
REGIONAL MANAGER
NCASI
TUFTS UNIV.
DEPT. OF CIVIL ENG'G
MEDFORD, MA
02155
PAULSON, WAYNE
PROFESSOR
UNIV, OF IOWA
ENVIRON. ENGINEERING DEPT.
2130 ENGINEERING BLDG.
IOWA CITY,  IA            52242

POWELL, GEORGE
DIRECTOR
GORE & STORRIE LTD.
1670 BAYVIEW AVE.
TORONTO, ONTARIO       M4G  3C2
CANADA
416-487-1351

REDMON, DAVID T.
VICE PRESIDENT
EWING ENGINEERING  CO.
6200 N. 39TH ST.
MILWAUKEE,  WI            53209
RUPKE, GERRY
PRESIDENT
RUPKE & ASSOCIATES LTD.
102 HOLLAND COURT
P.O. BOX 1570
BRADFORD, ONTARIO      LOG ICO
CANADA

SCHUK, WALTER W.
U.S. EPA
MUNIC. ENVIRON. RESEARCH CENTER
26 W. ST. CLAIR ST.
CINCINNATI, OH           45268
513-684-2621
SMITH, GEORGE W.
PROCESS SECTION MANAGER
ENVIREX, INC.
1901 S. PRAIRIE AVE.
              MUELLER, JAMES A.
              PROFESSOR
              MANHATTAN COLLEGE
              4513 MANHATTAN COLLEGE PKWY.
              BRONX, NY                 10471
              POPEL, H.J.
              GEORGE WASHINGTON LAAN 19
              2285 CG
              RIJSWIJK
              THE NETHERLANDS
              QUIGLEY, JOHN T.
              ASSOCIATE PROFESSOR
              UNIV. OF WISC.-EXTENSION
              ENGINEERING & APPLIED SCIENCE
              432 N. LAKE ST.
              MADISON, WI               53706
              608-262-0820

              ROOKE, TERRY
              PRESIDENT
              R.D. ROOKE & ASSOCIATES
              105 MIRANDA AVE.
              TORONTO, ONTARIO
              CANADA
              416-789-0646
                        M6B 3W8
WAUKESHA, WI
53186
SCHMIDTKE, NORBERT W.
PRESIDENT
N.W. SCHMIDTKE & ASSOCIATES LTD.
178 LOUISA ST.
KITCHENER, ONTARIO      N2H 5M5
CANADA
519-744-5931

SMITH, DANIEL W.
PROFESSOR
UNIV. OF ALBERTA
DEPT. OF CIVIL ENGINEERING
EDMONTON, ALBERTA       T6G 2G7
CANADA
403-432-4138

STEPHENSON, JOE
ZENON ENVIRONMENTAL, INC.
845 HARRINGTON COURT
BURLINGTON, ONTARIO     L8N 3P3
CANADA
416-639-6320
                                    452

-------
CONFERENCE LEADERS & PLANNING COMMITTEE

WREN, JEROME D.
SANITAIRE WATER
  POLLUTION CONTROL CORP.
P.O. BOX 744
MILWAUKEE, WI           53201
                                    453

-------
CONFEREES
ALLDREDGE, ROBERT
PROF. ENGINEER
SERPENTIX CONVEYOR CORP.
1550 S. PEARL ST.
DENVER, CO             80210
303-744-3507
ARTHUR, ROBERT
ARTHUR TECHNOLOGY
2496 WINNEBAGO
FOND DU LAC, WI        54935
414-922-9291

BAUMANN, WILLIAM B.
ENVIRONMENTAL ENGINR.
WI DEPT. OF NATURAL RESOURCES
P.O. BOX 7921
MADISON, WI            53707
608-266-3906

BEWTRA, J. K.
PROFESSOR
UNIV. OF WINDSOR
DEPT. OF CIVIL ENGINEERING
WINDSOR, ONTARIO
CANADA    N9B EP4
519-253-4232
CANTOR, ABBY
CHEMICAL ENGINEER
WARZYN ENGINEERING
1409 AMIL STREET
MADISON, WI
53713
CUMMINGS, PAUL W.
SR. ENV'L ENGINEER
NORTON CO.
1 NEW BOND STREET
WORCESTER, MA
617-853-1000
01601
DANLY, WILLIAM
ENVIRONMENTAL ENGR.
SPICER ENGINEERING CO.
818 S. MICHIGAN AVE.
P.O. BOX 1689
SAGINAW, MI            48605
517-799-4717
          ANDERSON,  SCOTT
          ENV'L ENGINEER
          MT WATER QUAL. BUREAU
          DEPT. OF HEALTN § ENV.  SCIENCE
          COGSWELL BLDG, ROOM A206
          HELENA,  MT             59620
          406-449-2406
          BAILEY,  MIKE
          ENGINEER
          GRAUE-HAWS
          P.O.  BOX 64669
          DALLAS,  TX
                       75206
          BELKEN,  JOHN D.
          GEN'L ENGINEERING CO
          151 E. COOK STREET
          P.O.  BOX 340
          PORTAGE, WI
          608-742-2169
                       53901
          BRAUN,  EUGENE
          ENVL.  ENGR.  2
          WI DEPT.  OF  NATURAL RESOURCES
          BOX 7921
          MADISON,  WI             53707
          608-267-7617
CHING, CHAY S.
ENVIRONMENTAL ENGR.
BADISCHE CORP.
P.O. DRAWER 3025
ATTN:  ALICE BRADBERRY
ANDERSON, S.C.         29621
803-261-0868
DAIGGER, GLEN T.
ENVIR. ENGINEER
CH2M HILL
P.O. BOX 22508
DENVER, CO
303-771-0900
80222
          DE RIENZO,  DOMENICK
          PROJECT ENGINEER
          PRO CONSOER TOWNSEND
          303 E.  WACKER DRIVE
          CHICAGO, IL
          312-938-0300
                       60601
                               454

-------
CONFEREES
FOELLMI, JERRY A.
CIV./SAN. ENG.
GEN'L ENGINEERING
151 E. COOK ST.
P.O. BOX 340
PORTAGE, WI
608-742-2169
CO.
     53901
GILBERT, R. GARY
PRESIDENT
AERATION TECHNOLOGIES, INC.
11 BARTLET ST.
ANDOVER, MA            01810
617-475-6387

GRAF, ROBERT D.
ENV'L. ENGINEER
IOWA DEPT. OF ENV. QUALITY
900 EAST GRAND AVE.
DES MOINES, IA         50319
515-281-8985

HADSELL, BRUCE
PLANT ENGINEER
FRIENDSHIP DAVIES  INC.
COUNTY ROUTE #20
FRIENDSHIP, N.Y.       14739
716-973-3031
HOPKINS, SIRIKUN
ENVIRONMENTAL ENGR.
EPD, GEORGIA DNR
270 WASHINGTON ST. SW
ATLANTA, GA
404-656-7400
     30334
JOHNSON, FORREST
VICE PRESIDENT
METRO-QUIP INC.
3600 PIONEER PARKWAY WEST
ARLINGTON, TX          70613
817-265-5151
JONES, RICHARD F.
GEN'L. ENGINEERING CO.
151 EAST COOK ST.
P.O . BOX 340
PORTAGE, WI            53901
608-742-2169
                                      52556
FREDERICKS, DAVE
PROJECT ENGINEER
FRENCH RENEKER § ASSOCIATES
BOX 135, 1501 S. MAIN
FAIRFIELD, IA
515-472-5145
               GORONSZY, MERVYN C.
               PROFESSOR
               VANDERBILT UNIV.
               P.O. BOX 6222
               NASHVILLE, TN
               615-322-2720
                       37235
                                      53214
GROSS, RODNEY
PROCESS ENGINEER
ENVIREX INC.
1901 S. PRAIRIE AVE.
WAUKESHA, WI
414-547-0141

HILL, GEORGE
COORDINATOR
ENVIRONMENT CANADA
351 ST. JOSEPH BLVD.
HULL, QUEBEC
CANADA    JOX 3GO
819-997-1220
HUIBREGTSE, GREGORY L.
RESEARH ENGINEER
REXNORD INC.
5103 W. BELOIT RD.
MILWAUKEE, WI          53214
414-643-2753

JONES, PETER
ASSOC. PROFESSOR
UNIV. OF SHERBROOKE
DEPT. OF CHEM. ENGINEERING
SHERBROOKE, QUEBEC
CANADA                 J1K 2R1
819-565-4421

KALINSKE, A. A.
VICE PRESIDENT
CAMP, DRESSER $ MC KEE, INC.
710 S. BROADWAY
WALNUT CREEK, CA       94596
415-933-2900
                               455

-------
CONFEREES
KERSTEN, WILLIAM W.
LAKESIDE EQUIPMENT CORPORATION
P.O. BOX T
BARTLETT, IL           60103
312-837-5640
KLUMP, DIETER
PRESIDENT
PASSAVANT CORP.
P.O. BOX 2503
BIRMINGHAM, AL         35201
205-853-6290

LAUGHLIN, JAMES
PARTNER
SHIMEK, JACOBS $ FINKLEA
8333 DOUGLAS AVE. #820
DALLAS, TX             75225
214-361-7900

LYNCH, EDWARD K.
ENVIRONMENTAL ENGR.
WI DEPT. OF NATURAL RESOURCES
BOX 7921, 101 S. WEBSTER
MADISON, WI            53707
608-266-0234

MAROZZI, MICHAEL
ASST. SANITARY ENGR.
PORTAGE COUNTY SANITARY ENGRG.
449 S. MERDIAN ST.
RAVENNA, OH            44266
216-296-3155

MELCER, HENRYK
A/HEAD, BIO. PROC. SEC.
WASTEWATER TECH.
  CENTRE/ENVIRONMENT CANADA
P.O. BOX 5050
BURLINGTON, ONTARIO L7R 4A6
416-637-4546

NESTMANN, FRANZ
RESEARCH ENGINEER
UNIV. OF KARLSRUHE
INST. OF HYDRO MECHANICS
75 KARLSRUHE,
WEST GERMANY
KLUGE, TIM
ENGINEER
ILLINOIS EPA
4500 S. 6TH STREET
SPRINGFIELD, IL
217-786-6892
62706
KO, KWANG BAIK
RESEARCH ENGINEER
KRAMER, CHIN § MAYO, INC.
1917 FIRST AVE.
SEATTLE, WA             98101
206-447-5386

LOWNDES, MARY ANNE
ENVL. ENGINEER I
WI DEPT. OF NATURAL RESOURCES
P.O. BOX 7921
MADISON, WI             53707
608-266-7420

MAEDERER, MARK
PROJECT MANAGER
GRAEF ANHALT $ SCHLOEMER
6451 W, CAPITOL DR.
MILWAUKEE, WI           53216
414-461-6900

MARX, JAMES
ENGINEER
DONOHUE § ASSOCS.
743 NORTH WATER ST.
MILWAUKEE, WI           53202
414-276-0300

NASIR, MULAZIM H.
ENVL. ENGINEER
WI DEPT. OF NATURAL RESOURCES
P.O. BOX 7921
MADISON, WI             53707
608-267-7627
OWEN, JAMES E.
MID-STATE ASSOCIATES
1111 EIGHTH STREET
BARABOO, WI     '        53913
608-356-3996
                               456

-------
CONFEREES
PESEK, TERRY T.
VICE PRESIDENT
ASSOCIATED TECHNOLOGIES LTD.
P.O. BOX 37203
OMAHA, NE                 68137
402-334-0100
ROONEY, THOMAS C.
ASST. DIR/EETC
REXNORD INC.
5103 W. BELOIT RD.
MILWAUKEE, WI
414-643-2754
53214
SHEPARD, FRANCIS D.
CONSULTING ENGR.
FRANCIS D. SHEPARD,
  ENGINEER $ ASSOCIATES
7282 CASWELL ROAD
BYRON, NY
716-548-2227
14422
SPEIRS, GORDON
ENGINEER
WASTEWATER TECH.
 CENTRE/ENVIRONMENT CANADA
P.O. BOX 5050
BURLINGTON, ONTARIO    L7R 4A6
416-637-4741

STURTEVANT, WILLIAM
GEN'L. ENGINEERING CO
151 EAST COOK ST.
P.O. BOX 340
PORTAGE, WI              53901
608-742-2169

SULLIVAN, RUSSELL
VICE PRESIDENT
AERATION TECHNOLOGIES, INC.
11 BARTLET ST.
ANDOVER, MA              01810
617-475-6387

TSEZOS, MARIOS
PROFESSOR
MC MASTER UNIVERSITY
CHEMICAL ENGINEERING DEPARTMENT
HAMILTON, ONTARIO
CANADA
416-525-9140
REIBER, STEVE
RESEARCH ASSOCIATE
DEPT. OF CIVIL ENG'G
UNIV. OF WASHINGTON
SEATTLE, WA
206-543-2574
                              FX-10
                                 98195
SEILER, TONY
MANAGER
SALZGITTER MACHINERY INC.
7315 E. ORCHARD RD, STE .  100
ENGLEWOOD, CO            80111
303-773-3911

SHEVLIN, ROBERT
PR. MGR/GRAY FINE AIR
THE GRAY ENGRG. GROUP
633 DENISON ST.
MARKHAM, ONTARIO
CANADA    L3R IB 8
416-475-9160

STENSTROM, MICHAEL K.
ASSOC. PROFESSOR
UCLA
7619 BOELTER HALL
LOS ANGELES, CA          90024
213-825-1408
        SULLIVAN,  JOHN
        ENVL.  SPEC.  Ill
        WI DEPT.  OF NATURAL RESOURCES
        BOX 818
        RHINELANDER, WI           54501
        715-362-7616

        THACKER,  H.  RAY
        ENV.  ENGINEER
        U.S.  EPA,  OFFICE OF R § D
        100 POWHATAN CIRCLE
        CHARLOTTESVILLE, VA      22901
        WARRINER,  READ
        GROUP MANAGER
        MILWAUKEE  METRO
         SEWERAGE  DISTRICT
        735 N.  WATER STREET
        MILWAUKEE,  WI            53202
        414-278-2033
                              457

-------
CONFEREES

WILSON, THOMAS E.
ASSOCIATE
GREELEY g HANSEN
222  S.  RIVERSIDE PLAZA
CHICAGO, IL                60606
312-648-1155
                       . tUS. GOVERNMENT PRINTING OFFICE: 1985/559-111/10760


                                  458

-------