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
------- |