United States
Environmental Protection
Agency
Municipal Environmental Research EPA-600/9-78-021
Laboratory April 1979
Cincinnati OH 45268
Research and Development
&EPA
Proceedings
Workshop Toward an
Oxygen Transfer Standard
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, US Environmental
Protection Agency have been grouped into nine series These nine broad cate-
nones were established to facilitate further development and application ot en-
vironmental technology Elimination of traditional grouping was fonsaously
planned to foster technology transfer and a maximum interface in related fields
The nine series are
1 Environmental Health Effects Research
2 Environmental Protection Technology
3. Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8. 'Special" Reports
9 Miscellaneous Reports
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161
-------
EPA-600/9-78-021
April 1979
Proceedings: Workshop Toward an Oxygen
Transfer Standard
Asilomar Conference Grounds, Pacific Grove, California
April 11-14, 1978
Cosponsored by the U.S. Environmental Protection Agency and
the American Society of Civil Engineers
Edited by
William C. Boyle
University of Wisconsin
Madison Wl 53706
Grant No. R805868
Project Officer
Richard C. Brenner
Wastewater Research Division
Municipal Environmental Research Laboratory
Cincinnati OH 45268
Municipal Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati OH 45268
-------
Disclaimer
This report has been reviewed by the Municipal Environmental Research Laboratory, U.S. Environmental Protection
Agency, and approved for publication. Approval does not signify that the contents necessarily reflect the views
and policies of the U.S. Environmental Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
-------
Foreword
The Environmental Protection Agency was created because of increasing public and government concern about the
dangers of pollution to the health and welfare of the American people. Noxious air, foul water, and spoiled land
are tragic testimony to the deterioration of our natural environment. The complexity of that environment and
the interplay between its components require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem solution and it involves defining the problem,
measuring its impact, and searching for solutions. The Municipal Environmental Research Laboratory develops
new and improved technology and systems for the prevention, treatment, and management of wastewater and
solid and hazardous waste pollutant discharges from municipal and community sources, for the preservation
and treatment of public drinking water supplies, and to minimize the adverse economic, social, health, and
aesthetic effects of pollution. This publication is one of the products of that research; a most vital communica-
tions link between the researcher and the user community.
The Workshop proceedings documented herein represent a major effort to summarize historical practices and
current art in the testing and evaluation of oxygen transfer devices used in the treatment of wastewater. These
proceedings form the technical base from which the American Society of Civil Engineers' Subcommittee on
Oxygen Transfer Standards is attempting to develop a tentative interim oxygen transfer standard on this research
grant project. The activities of the Subcommittee offer an excellent forum for examining the merits and fallacies
of ex-sting and proposed aeration equipment testing and evaluation methodology as it struggles to define
potential consensus positions in this controversial field.
Francis T. Mayo, Director
Municipal Environmental Research
Laboratory
-------
Preface
As this country continues to search for more effective ways to conserve energy, greater emphasis is being placed
upon the development of energy efficient processes in pollution control technology. The evaluation of the effi-
ciency of oxygen transfer devices and systems will assume greater importance as engineers seek less energy
intensive equipment. In spite of the considerable effort devoted to oxygen transfer technology, it is evident that
unanimity of opinion has not been achieved in the development of a standard procedure for the evaluation of
oxygen transfer devices. The areas of disagreement lie not only in the details of conducting oxygen transfer
tests, but also in the methods of data evaluation.
Presently, manufacturers rely on clean water shop tests (i.e., performance tests conducted in a test tank at the
manufacturing plant) for describing the oxygen transfer capability of aeration equipment. These capabilities are
normally expressed as standard oxygen transfer rates in clean water at zero dissolved oxygen (DO) and 20°C. It
has been shown that subtle differences in the method of data interpretation alone can produce differences of as
much as 10% in the clear, water standard oxygen transfer rate. Moreover, this uncertainty is further magnified
when translating these clean water test tank transfer rates to actual plant conditions. Because of differences in
wastewater characteristics, tank geometry, temperature, etc., uncertainties of 50% or more may be introduced.
This uncertainty has had a significant impact on the Nation's wastewater treatment plant construction program.
Faced with this problem, many consulting engineers have written requirements for field performance testing
into their specifications for aeration equipment. Contactors and manufacturers recognize this liability in prepar-
ing bids and will include larger factors of safety; thus, many aeration systems will be adequate but overdesigned.
Some systems, however, may still prove to be inadequate.
There is little question that a consensus standard is needed for oxygen transfer devices. Although several
standards are in existence, these standards are concerned primarily with the methodology of experimental
measurement and do not deal adequately with the interpretation and application of data to engineering design.
Moreover, there is no general agreement among engineers and manufacturers as to which standard to use. As
a result, a wide variety of techniques are employed, resulting in substantial variations in results for the same
device in clean water tests. Even larger variations will be evident in translating these results to full-scale design.
Only when a standard is developed through consensus agreement among experts in the field will a better
degree of uniformity, accuracy, and economy result. Even then, continued updating of the standard will be
required.
In an effort to obtain a consensus standard for the evaluation of aeration devices, the American Society of Civil
Engineers (ASCE) has established a Subcommittee on Oxygen Transfer Standards under the Committee on
Environmental Standards (Technical Council on Codes and Standards). The procedures which the Subcommittee
has set as its program to achieve a proposed interim standard include:
1. Review and critically evaluate the state of the art.
2. Evaluate and critically review existing standards; identify critical areas of disagreement and uncertainty.
3. Conduct a workshop on oxygen transfer.
4. Develop documentation for recommendations for an interim standard and recommend verification
methodology
Upon completion of the above tasks, the outputs will be integrated into a proposed interim standard (or, if
thought appropriate, a recommended procedural manual). The Subcommittee will subsequently solicit from the
field suggestions and comments on the proposed interim standard, determine verification methodology, field test
the standard to verify its applicability contingent on the availability of resources, and revise and submit
the proposed interim standard to ASCE for processing using the Society's voluntary consensus standards
procedures.
IV
-------
As a part of the Subcommittee's objectives, then, this Workshop was developed. This is the first in a number of
steps which will be taken in an effort to arrive at a consensus standard which will be acceptable to the
manufacturers, the users, and the designers.
William C. Boyle, Chairman
Subcommittee on Oxygen Transfer Standards
American Society of Civil Engineers
-------
Abstract
The objectives of this Workshop were to bring together experts in the field of oxygen transfer to: (1) identify
areas of agreement and disagreement in the evaluation of oxygen transfer devices and (2) identify research
needs in the development of an effective consensus standard for oxygen transfer devices.
The first day-and-a half were devoted to the presentation and discussion of topics related to the testing and
evaluation of oxygen transfer devices. The entire Workshop roster was then divided into five working group
sessions, whereby intensive discussions took place in efforts to arrive at consensus opinions on selected topics.
Areas of agreement and disagreement were delineated. The findings of each working group were presented to
the reassembled total group on the final day of the Workshop.
The papers presented in the opening day-and-a half general session and the summary reports of the working
groups comprise the major portion of the Workshop proceedings.
These proceedings were submitted in partial fulfillment of Grant No. R805868 by the American Society of Civil
Engineers under the sponsorship of the U.S. Environmental Protection Agency.
VI
-------
Contents
Page
Foreword iii
Preface iv
Abstract vi
Symbols and Nomenclature ix
Acknowledgements xi
Section I. A Perspective on Oxygen Transfer
1. Oxygen Transfer: A Historical Perspective — The Need for a Standard
W. Wesley Eckenfelder, Jr 1
2. An Aeration Equipment Manufacturer's View of an Oxygen Transfer Testing Standard
George R. Fissette 3
3. The Development of an Oxygen Transfer Standard
Edwin L. Barnhart 7
4. Philosophy of and Perspectives on a Standard by an Owner Representative
Lawrence A. Ernest 9
5. Standards in the American Society of Civil Engineers
Robert A. Crist 10
6. Philosophy of and Perspectives on an Oxygen Transfer Standard — The EPA View
Richard C. Brenner 12
Section II. State of the Art Review
7. Review of Oxygen Transfer Model Refinements and Data Interpretation
C. Robert Baillod 17
8. Oxygen Transfer Parameter Estimation
Linfield C. Brown 27
9. Review of Test Procedures
Wayne L. Paulson 41
Section III. Clean Water Testing: Shop and Field
10. Influence of Tank Geometry on Aerator Performance
Thomas C. Rooney 50
11. Influence of Mixing in Aeration
Ronald N. Salzman and Michael B. Lakin 59
12. Sampling Considerations
Gerry L. Shell 72
13. Analytical Measurement and Saturation Values for Dissolved Oxygen in Water
Vernon T. Stack, Jr 76
14. Accounting for the Effects of Water Temperature in Aerator Test Procedures
John S. Hunter, III 85
15. Influence of pH and Iron and Manganese Concentrations on the Non-Steady State Clean Water Test for
the Evaluation of Aeration Equipment
Hussein Naimie and Steve Nelson 91
vii
-------
Contents (continued)
16. Gas Flow and Power Measurement
Fred W. Yunt
17. Oxygen Transfer Data Interpretation: Non-Steady State Clean Water Tests
David T. Redmon, Jerome D. Wren, and Mikkel G. Mandt .................................... 128
18. Measurement of Alpha and Beta Factors
R. Gary Gilbert [[[ 147
19. Surface Aeration Equipment: Field Performance Testing Vs Shop Performance Testing
John R. Stukenberg and Valery N. Wahbeh ................................................. 163
Section IV. Evaluation of Respiring Systems
20. Oxygen Transfer in Closed Systems
James A. Mueller, Jack Famularo, and Thomas J. Mulligan .................................. 180
21 . On-Site Evaluation: Steady State Vs Non-Steady State Testing
Ross E. McKinney and John R. Stukenberg ................................................. 1 95
22. Problems Encountered in Steady State Field Testing of Aerators and Aeration Systems
A. A. Kalinske [[[ 205
Section V. Tracer Methods
23. Use of Tracers for Evaluation of Oxygen Transfer
Larry A. Neal [[[ 210
Section VI. Supplementary Papers
24. Notes for Workshop Toward an Oxygen Transfer Standard
James J. McKeown [[[ 228
25. Oxygen Transfer in the Activated-Sludge Process
Arthur G. Boon [[[ 232
26. Effect of Gas Phase Temperature on Oxygen Saturation Value
James A. Mueller, Martha L Quintana and Dominic DiToro .................................. 240
Section VII. Working Group Summary Reports
27. Group A. Philosophy of the Standard [[[ 244
28. Group B. Modelling and Data Interpretation [[[ 245
29. Group C. Non-Steady State Clean Water Testing. Shop and Field ................................. 247
30. Group D. Alpha, Beta, and Temperature Corrections ............................................ 249
-------
Symbols and Nomenclature
A list of the more commonly used symbols and nomenclature are given below. An effort has been made to
standardize most of the symbols in this proceedings using the listing below:
D = impeller diameter (L)
db = bubble diameter (L)
d0 = orifice diameter (L)
g = acceleration of gravity (L/t2)
M = weight of water (m)
n = impeller speed (L/t)
p = pressure (f/L2)
pa = ambient atmospheric pressure (f/L2)
p0 = partial pressure of oxygen (f/L2)
P = power (fL/t)
Q = volumetric flow rate (L3/t)
S = concentration of substrate (BOD, COD, etc.) (m/L3)
V = volume (L3)
X = concentration of active biomass (volatile suspended solids unless stated otherwise) (m/L3)
P = mass density (m/L3)
7 = weight density (f/L3)
Mass Transfer Relationship, based on the overall liquid film coefficient
W = KLa(C*-C)
W = transfer rate per unit volume in clean water (m/L3t)
W = dC/dt (for non-steady state non-flow conditions) (m/L3t)
Wf = transfer rate per unit volume in dirty water field condition (m/L3t)
KL = overall liquid phase mass transfer coefficient, based on the liquid film (clean water) (L/t)
a = interfacial surface area per unit volume (L"1)
KLa = volumetric mass transfer coefficient, given conditions of geometry, mixing, temperature, etc, based on the
liquid film (clean water) (t"1)
(KLa)f = volumetric mass transfer coefficient under dirty water field conditions of geometry, mixing, temperature (t"1)
C = DO concentration in the liquid phase (m/L3)
C* = DO saturation (equilibrium) concentration in the liquid phase (m/L3)
C* = DO saturation concentration in clean water corresponding to a given partial pressure and
temperature (m/L3)
C* = average DO saturation concentration in a dirty field application corresponding to an average
gas phase partial pressure and temperature attained in the field application; defined so that
Wf = (KLa),(q-C)
IX
-------
C* = DO saturation concentration corresponding to air at ambient atmospheric pressure and 100% relative
humidity at a given temperature (surface saturation value) (m/L3)
C£ = DO saturation concentration attained at an infinite time in a non-steady state clean water test at
a given temperature (m/L3)
Other Symbols
a - ratio of KLa in dirty water application to KLa in clean water at equivalent conditions of temperature,
geometry, mixing, etc. (dimensionless)
13 = ratio of C* in dirty water to C* in clean water at equivalent conditions of temperature and partial pressure
(dimensionless)
0 = temperature adjustment factor defined so that
(
-------
Acknowledgements
The time and effort of the American Society of Civil Engineers' Subcommittee on Oxygen Transfer Standards in
planning and organizing this Workshop are gratefully acknowledged. The contributions of the Workshop
participants and authors during the three days of meetings at Asilomar represent an excellent starting point
toward the development of consensus procedures for the evaluation of oxygen transfer devices.
The cooperation of Mr. Robert B. Morgan of the American Society of Civil Engineers in arranging for the
Workshop site and looking after numerous logistics details before, during, and after the Workshop was invalu-
able to the success of this conference.
XI
-------
Section I. A Perspective on Oxygen Transfer
Oxygen Transfer:
A Historical Perspective-
The Need for a Standard
W. Wesley Eckenfelder. Jr.
Vanderbilt University
Nashville TN 37235
The development of an aeration standard has an interest-
ing history. Prior to the mid 1950's, aeration consisted
primarily of diffused air systems with specifications
being based on ft3 of air/gal of sewage treated. These
specifications were based on experience.
Research was conducted during this period on the
mechanism of oxygen transfer and aeration which led
to the use of oxygen transfer coefficients (KLa) and
transfer capacity in wastewaters (a). Other aeration
devices, primarily mechanical surface aerators, came into
vogue in the late 1950's and early 1960's. In order to
differentiate aeration capacity and performance for
various devices, test procedures were devised and
adapted in one form or another by the various vendors of
aeration devices. While all of these procedures involved
deaeration with cobalt catalyzed sodium sulfite in water,
the cobalt concentration varied and the method of
calculation of oxygen saturation under the test conditions
varied. During this period, the writer was involved in
aerator testing for Wells Products, Walker Process and
Mixing Equipment Co. More recently, several significant
facts have emerged, namely (a) the cobalt concentration
has a marked effect on the transfer rate calculated,
(b) residual sodium sulfite may have an effect, (c) the
oxygen saturation value used in the calculation will
influence the transfer rate and (d) the temperature
coefficient employed will influence scale-up to other
operating conditions. Considerable controversy exists on
the determination of these factors. There are other far
more significant factors which will affect successful
aeration performance. These are:
1. The definition of « under full-scale operation.
2. The effect of oxygen uptake rate on oxygen transfer.
3. The effect of basin geometry on aerator performance.
4. The effect of surface active agents on aerator
performance, particularly mixing in the basin.
First, directing our attention to standard clean water
transfer testing and specifications, it would appear that,
based on the mass of information now available, agree-
ment can be reached on cobalt catalyst concentration,
maximum sodium sufite level, a calculation procedure for
1
oxygen saturation and a temperature coefficient. Even if
the values selected do not approach theoretical perfec-
tion, all vendors and all aeration devices will be based on
a common standard for purposes of performance
comparison.
In my opinion, it is far more important to define the
aerator performance under field operating conditions.
Concerning a, the accepted test is to compute KLa in
the laboratory for water and the wastewater in question
and calculate a as the ratio of KLa in wastewater/KLa in
water. When mechanical aerators are used, a is related
to the speed of the aerator. For an air diffuser, air flow is
related to a. The difficulty is scaling this value to field
conditions. The writer recalls data developed in Great
Britain some years ago in which a varied from 0.9 to 1.4
at the same detergent concentration, in basins of different
geometry, using surface aeratiors. Investigations are
needed to establish a relationship between laboratory
and field performance, if any, and to define procedures
for determining a both in the field and in the laboratory.
There is evidence that aerator transfer rate is affected by
the oxygen uptake rate. Albertson and Digregorio (1)
showed an increasing oxygen transfer rate with in-
creasing oxygen uptake rate. The writer was involved in
an evaluation of the performance of low speed surface
aerators treating a chemical wastewater. Data were
collected over a one year period during which the
organic loading progressively increased, resulting in an
increase in oxygen uptake rate. Using the accepted
steady state test, the oxygenation efficiency (Ib C>2/
hp-hr) increased linearly with oxygen uptake rate to a
value of 6.4 Ib Oa/hp-hr. The question is: Is the test
valid or is there a significant effect of oxygen uptake rate
on oxygen transfer rate?
Kalbskopf (2), von der Emde (3) and others have identi-
fied significant effects of the presence of surface active
agents on mixing patterns in aeration basins. This in
turn, no doubt, influences the oxygen transfer rate.
Further definition of this phenomenon is needed.
To illustrate this point, studies were conducted by the
writer on an aeration basin employing jet aerators.
-------
Water tests had been conducted in this basin prior to
the activated sludge studies. Tests were conducted to
define a in the laboratory. Non-steady state tests were
conducted on the basin contents and oxygenation
capacity was evaluated from the COD removal in the
basin. Steady state tests were also conducted. The
results of these tests are summarized below.
The differences between values obtained is significant.
It should be noted that the oxygen uptake rate for the
steady state test tends to read low since the sample
withdrawn from the aeration basin is without food
during the course of the test.
Results of Basin Tests for Oxygenation Capacity Basis
20°C, 1 mg/l DO
Steady State 5220 Ib C>2/day
Clean Water Corrected for a and DO 9300 Ib 02/day
COD Removal 6300 Ib
References
1. Albertson, O.E. and D. DiGregorio, "Biologically
Mediated Inconsistencies in Aeration Equipment
Performance", Journal Water Pollution Control
Federation, 47, 976, 1975.
2. Kalbskopf, K.H., "Flow Velocities in Aeration Tanks
with Mechanical Aerators", Water Research, 6, 413,
1972.
3. von der Emde, W. "Aeration Development in Europe."
Advances in Water Quality Improvement, Vol. 1,
edited by Gloyna and Eckenfelder, University of
Texas Press, 1968.
Discussion
Arthur G. Boon
Water Research Centre, Stevenage Laboratory
England, United Kingdom
Studies at the Water Research Centre have shown that
changes in the concentration of suspended solids in the
mixed liquor of an activated-sludge plant aerated by a
fine-bubble diffused-air system had a slight effect on the
rate of oxygen transfer. The oxygen transfer coefficient
decreased by about 20% when the concentration of
suspended solids was increased from 1400 to 6600 mg/l.
-------
An Aeration Equipment
Manufacturer's View of
an Oxygen Transfer
Testing Standard
George R. Fisette
Ralph B. Carter Company
Hackensack NJ 07602
Introduction
A manufacturer must become involved with the formu-
lation of an oxygen transfer test because it affects his
goals. The present variety of "official standards" leaves
a manufacturer in a quandry as to which standard
should be used. The problem is further complicated by
the inability of the trade to agree on units of measure-
ment and their definitions. The standard that this
Workshop arrives at must be complete and universally
accepted in order that our efforts are not in vain.
Manufacturer's Goal
When one bypasses all of the rhetoric that is currently
being put forth concerning a manufacturer's commit-
ment to God, country, and society, the baseline is that a
manufacturer's prime objective is to make a profit.
Realistically, all else is subservient to this goal and any
additional goals can only be pursued because that profit
is made. As discussed below, the absence or presence
of an oxygen transfer test standard falls way down the
list of factors that influence a manufacturer's profit. But
it is a factor that must be dealt with.
To reach this goal a manufacturer must consider a large
number of factors including those listed in Table 1.
While the list is not all-inclusive by any means, it
should be obvious that if several of these factors are
negative, a manufacturer will have a poor profit margin.
This paper is restricted to the effect of oxygen transfer
testing, what it should encompass, and how testing
affects a manufacturer.
Table 1. Factors Affecting an Aeration
Manufacturer's Goal
General economy
Salesman's ability
Proprietary products
Government approval
Economical manufacturing
Labor unions
Locations
Transportation
Weather
Acceptance of equipment
Believable design
Reliable construction
Competing processes
Competing equipment
Standard performance
ratings
Manufacturer's Requirements for an
Oxygen Transfer Test
In the best interest of each manufacturer is the test that
proves his piece of equipment to be better than the
competitor's. A wide spectrum of tests, of varying
degrees of reliability, have been developed to serve this
purpose. Some of these tests are analagous to govern-
ment issued MPG ratings which show that a VW gets
better gas mileage than a Porsche, while many others
are similar to a manufacturer's self-proclaimed
superiority of "whiter-whites" in your shirts. The oxygen
transfer tests presently used fall closer to the latter
image than the former
At least the government issued MPG rating has some
believability. Many oxygen transfer tests, like the whiter
whites, generally are not acceptable to very many
people since there are too many unknown variables. An
oxygen test should be acceptable to all parties partici-
pating in an evaluation. Many times this is not so.
Furthermore, the test should be believable. All of us
unquestioningly accept the electric company's reading
and report of our electrical usage. On the other hand,
many consultants will not accept "MY" test method or
in fact any other method except their own pet
procedure.
The test manufacturers prefer it should be easy to
understand, simple to perfrom, and low in cost. It
should be both accurate and reproducible. The value 01
the answer must be known. Can it be used directly or is
there a random scale-up factor? The test method and
the results must relate to the end use design. It is not
reasonable to test in a 55-gal barrel and scale-up to
a 10 million gal basin. Equally important is a direct
relationship between variations in test results and
equipment design or manufacturing procedures. If
variations in equipment have less effect than variations
in the results obtained between test methods, then the
manufacturer does not have any need for a test. The
unknowns and the buried assumptions incorporated in
present test procedures make it difficult to determine
whether the "right" answer has been obtained or not.
Therefore, a manufacturer can in all honesty proclaim
the method he uses is the only one that correctly
evaluates his equipment and design.
-------
Requirements for an Oxygen Transfer Test Standard
A standard that would be acceptable to a manufacturer
may not be acceptable to a scientist, but it should be.
Therefore, we should develop the most thorough
standard possible and let the manufacturers adjust to it.
What is needed for oxygen standard is one having the
status of an ASTM test, required by EPA for design, and
listed in "Standard Methods". Another test method,
standard or not, will not gain any greater acceptance
unless there is a quasi-legal requirement for its use. A
major problem for a manufacturer today is which
method to use. Table 2 lists over two dozen procedures
that are currently in use. Which one should a manu-
facturer use? Which should a consultant design by? By
which one does an owner rate his system?
Table 2. Oxygen Transfer Procedures
Mine
WPCF/ASCE
PEMA
TAPPI
ASCE
WWEMA
ASTM
EPA
Kenics
Mixco
R. B. Carter Company
Union Carbide
Penberthy
Sanitaire
RAMCO
Dupont
CH2M Hill
Black & Veatch
Hazen & Sawyer
Shell
Kalinske
Downing
Kayser
Non-Steady State
Biological
Steady State Biological
Steady State Sulfite
Non-Steady State Sulfite
A number of philosophical problems must be considered
in arriving at a new standard. Table 3 lists several
philosophical problems already mentioned and many
others that must be evaluated by this Workshop.
However, before these questions are considered in
detail, the specific conditions of a test standard must be
agreed upon. There is little need to discuss the broader
philosophy of a standard when as a trade group we
can't even agree upon the definitions. The instructions
to speakers at this conference had to include a standard
list of terms for writers to use. Even more basic than
that is an agreement as to what an individual term
means or measures. For example, Table 4 lists 13
conditions specifying what a standard cubic foot is. The
term standard is used in the writer's instructions
without being defined. This is a potential source of error
of over 18% before the Workshop even starts. Even this
table is not sufficient to define what a standard cubic
foot of air is. Since air is a mixture, the constituents
and their concentrations must also be defined. Ideally,
our standard cubic foot should be defined to the point of
ridiculousness as follows: that amount of air contained
in a cubic foot of space as measured by a NBS certified
wet test meter at 68°F, 14.6960 psia, 36% RH, contain-
ing 78.08%v N2, 20.95%v O2, 0.93%v Argon, 0.034% v
C02, and 0.01 %v trace elements with an average
molecular weight of 29.100. From here on one can
make whatever simplifications are felt necessary, but at
least we all start from a common reference point.
Table 3. Philosophical Points to be Considered in
Arriving at a New Oxygen Transfer
Test Standard
1. Universally accepted.
2. Everyone required to use, either legally or by
peer pressure.
3. Will all equipment maintain the same relative
differences as they now have?
4. Will it correlate with past tests and data?
5. Capability of all manufacturers to properly test and
retest and understand.
6. Capability of all consultants and owners to properly
test and retest and understand.
7. Narrowness of input to standard. Lack of input from
other diciplines. (Chemical engineering,
Pharmaceuticals, petro-chem, etc.).
8. Applicable to all aeration equipment. What about
rotating biological contractors?
9. Simple to run.
10. Cheap to perform.
11. Accuracy and precision.
12. Reproducibility.
13. Is this change for changes sake?
The standard arrived at by this Workshop must be this
complete and all encompassing. It should apply to all
biological treatment devices. There may well be varia-
tions in procedure, depending upon the generic type,
but all types as listed in Table 5 should be covered. The
various types of oxygen tests listed in Table 6 should be
included within the overall test standard. In addition,
the more specific aspects of an oxygen standard listed
in Table 7 will have to be specified. The final standard
must address itself to these and other aspects in order
to be of value to manufacturers, consultants, and
owners.
Summary
The final published standard should be of such a form
that it can be expanded by sections to cover all methods
-------
Table 4. Standard Conditions for Gas Measurements
Source
Chemist
Engineer
Federal Power Commission
ASCE (Gas book)
European
European
European
European
European
HVAC
HVAC
Hoffman Blowers
ASCE-02 Symbol Table
Pressure
1 .0 ATM
14.7 psia
14.73 psia
14.73 psia
1 .033 kg/cm2
1.033 kg/cm2
1 .000 kg/cm2
1 .000 kg/cm2
76 cmHg
14.7 psia
14.7 psia
14. 70 psia
1 ATM
Temperature Humidity
0°C 0%
32 °F ?
60 °F ?
60°F ?
0°C ?
15°C ?
0°C ?
15°C ?
10°C ?
68°F ?
70°F ?
68°F 36%
20°C 100%
ft3'
1.0000
0.997
1 .0545
1 .0545
0.9998
1 .0547
0.9678
1.0210
1 .0366
1.0735
1 .0776
1.1725
1.1852
1 ATM = 14.6960 psia = 1.03323 kg/cm2 = 76.0 cm Hg
* Assumed % RH zero for all unknown cases.
Table 5. Aeration Equipment
Surface impeller, high and low speed
Submerged impeller
Submerged turbine, upflow and downflow
Fine bubble diffusers
Coarse bubble diffusers
Static tubes
Eductor/ejector
Inline mixers, static and dynamic
Rotating biological contactors
Trickling filters
Cooling tower style trickling filters
and equipment. In the limited time available to us, we
should restrict our efforts to producing a comprehensive
method for a pure water shop test. Once we have the
most commonly used method in printed form, the
expansion of this standard to cover clean water field
test and wastewater testing will be considerably easier.
There will always be a need for clean versus dirty water
testing, and shop versus field testing. Not every
manufacturer has access to a properly designed tank for
shop test. Nor is every field test compatible with clean
water methods. For example, how does one evenly and
quickly distribute sulfite over a basin several acres in
size? The pure water shop test should be established as
the base procedure with all other variations referenced
back to it.
In order to promote usage and general acceptance,
ASCE must publicize the method, not only in our
internal magazine Civil Engineering, but also in other
trade publications. The EPA needs to state that this
standard is the only one they approve for use in
aeration equipment evaluations and that it must be
used in all federally funded projects. Finally, the
remaining 10,000 plus consultants who couldn't attend
this Workshop must be sold the advantages of specifying
this standard.
Manufacturers can help the overall process by accepting
and using the standard to evaluate their equipment. The
most difficult problem for a manufacturer, and indeed
for many others, will be what does one do with the
large quantity of results previously measured and
calculated. For it is quite probable that our final standard
will not be identical to what any of us are presently
using. For example, the Workshop may conclude that
siphon lines or pumped samples are no longer
acceptable and that in situ measurements must be made.
Or if we conclude that the "Analysis of Variance"
method of data analysis is to be the base procedure,
what do those who have used the "Log Deficit" method
do? Not only do we have to establish a general standard,
but we must specify one of the test conditions listed in
Table 6 and define each of the individual aspects listed
in Table 7. Furthermore, we will have to establish
correlations between our prime or basic procedures
and the other closely related variations.
It is soon evident that what on the surface appears
to be a simple test is going to be very difficult to
accomplish.
-------
Table 6. Test Conditions
Table 7. Aspects of Oxygen Testing
Shop
Field
Clean (how clean)
Dirty (how dirty)
Wastewater
Sulfite methods
Steady state methods
Non-steady state methods
In conclusion, I would like to state that I anticipate that
when we are all done the "new" answer to a transfer
test will be within the accuracy of our present methods.
In other words, I see no net effect to a manufacturer of
a standard other than that there will now be one
generally accepted test procedure. The existence of one
standard will reduce the bickering and finger-pointing
that now goes on. This will make life easier for all of us.
Sampling
Water analysis
Data analysis
Chemicals, purity and type
Power input
Tank size
Test apparatus
Measuring apparatus
Sulfite distribution
Aerators per unit volume
Aerators per unit surface area
Spacing
Relationship of test to field design
-------
The Development of an
Oxygen Transfer Standard
Edwin L. Barnhart
Hydroscience, Inc.
Emerson NJ 07630
To the consulting engineer the selection of aeration
equipment involves the development of a total system
capable of serving a variety of needs within the waste
treatment plant. The most efficient and economical trans-
fer of DO is only one of several major criteria that must
be satisfied.
The aerator is also the only source of mixing within the
system and must provide sufficient velocity within the
tanks to effectively provide contact between the waste
and the microorganisms. This mixing must be provided in
a manner compatible with the sheer tolerance of the
organisms and other biotic constraints. If these criteria
are not satisfactorily accomplished, the process, even in
the presence of sufficient DO, will not serve its design
purpose.
Oxygen transfer mechanisms also significantly influence
the rate of heat loss within the system. In general,
compressed air type systems add heat to the liquid while
surface aeration systems result in heat loss. In large
volume systems the influence of oxygen transfer devices
on heat balance can have a major impact on the treat-
ment efficiency of the system. Other impacts associated
with aeration include effects on pH through stripping
and ambient noise levels.
The choice of aeration equipment influences the basic
geometry of the plant and therefore the capital costs.
This phenomenon is of course interrelated with the flexi-
bility of the system and its ability to respond to variability
in the waste flow and the ease with which future
expansions can be accomplished. For example, although
a large single aerator in a square tank might be the most
efficient oxygen transfer design, a combination of three
or four smaller aerators in a group of tanks in a series or
parallel may very well present the better overall design.
From the discussion above it can be seen that oxygen
transfer alone cannot be considered as the sole criterion
when developing an oxygen transfer standard. Rather,
any system must be viewed in terms of its overall com-
patibility with the job to be accomplished. The standard
as evolved must be sufficiently flexible to allow the
consulting engineer the wide range of latitude necessary
for the optimization of oxygen transfer within his particu-
lar design context.
Discussion
Edwin 0. Simmons
Passavant Corporation
Birmingham AL 35201
Ed Barnhart has stated in a nutshell what the significance
of an oxygenation standard is in the overall considera-
tion for the design of a waste treatment plant. To place
too much emphasis on a standard which, at best, is going
to be controversial for years to come can create a "penny
wise, pound foolish" situation in the waste treatment
profession. The aeration 'tail' can be found wagging the
treatment plant 'dog'.
This does not imply that a standard test is not desirable.
Certainly it would be a relief to know that every investi-
gator would use the same temperature correction factor
(theta value) to relate to a 20°C standard condition. If a
single value for theta cannot be agreed upon, several
values being supported by credible scientific investigation,
the logical question is whether the factor is, in fact, a
temperature one. Certainly, it would be a relief to know
that there is a reference water which can be universally
reproduced for use anywhere in the world to provide
reproducible standard results anywhere in the world.
Assuming that, in time, all of these factors can be
standardized one by one, we are still left with the
uncertainty of the alphas, betas, geometries, etc., that
exist in the real world of waste treatment. The uncer-
tainty of these factors assuredly has a greater influence
on the ultimate success of the aeration system than does
the uncertainty of standard test factors.
It would seem strong consideration must be given to
separating the research efforts into two parallel studies,
one to develop the standard tests and the other to con-
duct full-scale field tests to relate back to the standard.
The first can be the basic research progression from
laboratory to pilot plant to full scale. At the point of full
scale, to be completely objective, it would be necessary
-------
to investigate every type of aeration device in the device could a relationship, or lack thereof, be developed
minimum sized structure that truly represents an actual to predict actual treatment from mere standard tests.
treatment plant of the optimum geometry. The second
would involve planned field studies in full-scale plants Tne greatest potential pitfall ahead is to unwittingly bias
with the actual wastes to be treated, following studies studies toward or away from styles of aeration equipment
first utilizing the proven reference water. Only after a for lack of objectivity or failing to, as noted by Barnhart,
number of these studies with each style of aeration consider all aspects of a complete treatment system.
-------
Philosophy of and
Perspectives on a
Standard by an Owner
Representative
Lawrence A. Ernest
Milwaukee Sewerage Commission
Milwaukee Wl 53217
While I am an employee of the Milwaukee Sewerage
District and am responsible for the operation and
maintenance of one of the oldest activated sludge plants
in the world, I am here to speak for all owners, be they
public or private.
It makes little difference to the consumer whether he
pays for the installation of improper equipment at the
tax counter or the store counter. In the ultimate end, it
is always the consumer who pays. Therefore, I feel I
represent not only the User as indicated on the pro-
gram, but the total public who pay the ultimate cost.
A performance standard for oxygen transfer devices will
prevent the installation of equipment more efficient and
more expensive than is needed to do the job. It should
also allow selection of equipment which will do the
needed job and not require expensive replacement
because it fails to do the job.
The User's needs will vary from treating a dilute settled
municipal wastewater to treating a high strength
industrial waste.
The User is entitled to a test procedure which:
1. Will be easily understood by the Manufacturer, the
Consulting Engineer, the Installing Contractor-Bidder,
and the User
2. Will give representative results in a test situation
before installation in the treatment plant
3. Can be field tested in place prior to being placed on
the production line.
I am in no position to tell you how to provide the above
test procedures. I hope that this conference will develop
them because they have been needed for a long, long
time.
We see more stringent effluent standards being set at
all regulatory levels. The installation and operation of
better facilities will be required to meet these standards.
Development of an oxygen transfer standard that can be
used to predict performance under actual operating
conditions will make attainment of these new standards
more probable.
-------
Standards in the
American Society of
Civil Engineers
Robert A. Crist
American Society of Civil Engineers
New York NY 10017
What is a voluntary consensus standard? This is a
question which has been wrestled with for many years.
In the sense of ASCE's activities it is a resource for
voluntary use by designers, owners, politicians and
regulatory agencies, to name only a few. As a voluntary
consensus standard gains acceptance it may become
legal and binding if it is adopted by regulatory agencies
such as federal, state or local governments. It is then
often referred to as a code. We should realize that
ASCE is not in the code business but rather in the
consensus standards business, although consensus
standards provide a resource base for code development.
Standards can become synonomous with and identical
to portions of codes if a standard is a good resource, as
codes are generally an assimilation of several standards.
Inconsistant terminology often creates confusion.
Standards are sometimes referred to as codes and the
preface "model" may also be used. Model is a very
descriptive and appropriate modifier. Model building
codes such as those of Building Officials and Code
Administrators (BOCA), International Conference of
Building Officials (ICBO) and Standard Building Code
Congress (SBCC) have been in existence for many years.
Perhaps the most advanced terminology is model per-
formance standards. The performance concept encourages
innovation and has an important ingredient, a provision
for evaluating performance.
The voluntary consensus standard committee must have
special characteristics:
Balance — The committee must be made up of those
who are affected by the standard. Affected parties
should be equally represented on the committee.
Consensus — A consensus must be reached with all
negative opinion resolved. The consensus starts with the
committee and reaches to the public. ASCE's audience
for consensus is the enrolled members of the Technical
Council on Codes and Standards (TCCS). The public
consensus would come through the approval of ASCE's
standards by the American National Standards Institute
(ANSI) as American National Standards.
ASCE became an ANSI accredited standards making
organization in December 1977, which represented a
major accomplishment in the early stages of the Society's
standards program. The administration of standards
committees is centered at ASCE Headquarters under the
general charge of the Managing Director of Publications
and Technical Affairs and with the Technical Council on
Codes and Standards Executive Committee. In addition to
administrative responsibilities for the standards activities.
Headquarters and TCCS are responsible for liaison to
ANSI and other standards organizations (ASTM, ASME,
ACI, etc.), and also for following legislation and other
items that affect ASCE's standards activities. The techni-
cal focus lies within the standards committees with
appropriate contributions from the Society's technical
committees.
The need for a standard must be established. The extent
of need may be determined through answers to questions
such as:
Does a similar standard exist, i.e., will a new one be
redundant?
Is the standard necessary for protection of life, safety or
performance?
Will a standard unnecessarily interfere with the market
place or contribute to it?
These items are, of course, not exhaustive. Once the
need is established, the formulation of a standard must
result from a carefully assimilated plan beginning with
the first word and going through the standard's
implementation. The responsibility assumed when a
committee partakes of standards writing is immense.
Responsibility has broad implications. Legal obligations
fall within the jurisdiction of laws enforced by the Federal
Trade Commission and the Department of Justice and
relate to such items as the restraint of trade and boycott.
Another legal responsibility is the liability for the
standard's content. For example, does the standard, if
deemed to be met, provide the safety or performance
stated or intended? There is a responsibility to society in
general which in turn is related to the justification of the
need for a standard as previously discussed. A profes-
sional responsibility is assumed by the committee to
ensure technical quality of a standard. Technical quality
should be continually critiqued and not compromised
Proper attention to evolution is very important for the
implementation of a standard. Unfortunately, improper
evolution, i.e., transformation from state-of-the-art
to a standard, cannot be determined until it is too late.
Large changes from current practice or accepted usage
result in unused standards, and inadequate use of the
state of the art results in early obsolescence. Some
10
-------
questions can be asked and, if answered correctly, can
make up a set of guidelines that correlate to appropriate
evolution.
What exists now?
What is the state-of-the-art?
What portion of the state-of-the-art is credible?
What will the audience accept?
It should be noted that the portion of the state-of-the-art
that can be used is that which is the most credible and
acceptable, remembering that experience has shown the
state-of-the-art or technology often leads applications or
standards by 5 to 10 years or more.
Proper identification of the audience for a standard may
be deceptive. Standards can have broad or narrow
audiences, and individuals or groups are addressed di-
rectly or indirectly. The direct audience should be the
primary focus. Standards are used by the designer or
industry to meet an obligation of public protection and
quality. However, standards are also used by engineers,
politicians, attorneys, and economists in the decision
making process long before the standard is applied to the
particular product, such as a building, water treatment
facility or equipment.
Standards can become complex; however, unnecessary
complication should be avoided. There are several tools
of technology that have recently become practically
available and applicable to the formulation of standards.
Performance concepts provide and encourage innovation
while still requiring that performance is met. Probabilistic
methods allow the handling of random processes in a
formal manner to estimate risk, means, variability and
uncertainty. Sometimes probabilistic approaches can
provide a method for simplifying to a deterministic format.
Decision analysis is a useful tool for analyzing the logic
or sequence of a standard and the presence of loops,
voids and dead ends. It also can provide for the construc-
tion of a simple guide for the application of a standard.
Commentary, guidelines or manuals of practice for
standards are the necessary textbooks and roadmaps.
Keeping in mind that, both nationally and internationally,
the International System of Units (SI) will be in the
forefront in the near future, alternative metric units
should be provided in all standards.
Last but not least, no matter how good a standard is, it
won't be used if potential users are not informed. Good
standards have to be promulgated through advertising,
education, discussion and application. This process is
neither self-starting nor self-perpetuating. The Society as
well as all those involved in a standard have an obligation
to an end — the popular use of a voluntary consensus
standard — which is the only measure of success.
11
-------
Philosophy of and Perspectives
on an Oxygen Transfer
Standard—The EPA View
Richard C. Brenner
U.S. Environmental Protection Agency
Municipal Environmental Research Laboratory
Cincinnati OH 45268
Introduction
Because it is removed from the design selection process,
the U.S. Environmental Protection Agency's (EPA)
construction grants program is totally reliant on the
comprehensiveness and accuracy of the historical treat-
ment data, pilot studies, and testing procedures which
form the bases for plant designs. The design engineer is
also dependent on these data bases as well as his
technical judgement in evaluating alternative design
strategies and selecting the most cost-effective option
for each client.
Greater attention has been given in recent years to
process oriented research and demonstration studies
than to the development of techniques for measuring
equipment performance on a standardized basis. We
have specific effluent quality criteria against which
process performance is normally judged and in most
cases well established analytical and monitoring
procedures for determining compliance with those
criteria. On the other hand, our bases for selecting
equipment to implement a particular process design are
fragmented, inadequate, and often unintentionally
biased. It is virtually impossible to predict with any
degree of confidence how some types of equipment will
respond or perform in wastewaters or on wastewater
sludges of differing characteristics.
Nowhere is this dilemma more apparent and poten-
tionally cost wasteful than in the design and selection
of oxygen transfer equipment. Each manufacturer has
ilc own philosophy regarding the measurement, evalua-
tion, and prediction of oxygen transfer. Two common
complaints often voiced by engineers are the difficulties
encountered in (1) correlating manufacturers' oxygena-
tion claims with shop test or field test data and (2)
projecting equipment performance in a respiring
biological system once it is felt a representative clean
water oxygen transfer capability has been established.
The discrepancies between anticipated and actual
performance are often sufficiently large to warrant
substantial field modifications to the aeration equipment
furnished. The costs of performing such modifications
and the ill will generated offer vivid testimony of the
need for improved oxygen transfer design procedures.
General Thoughts on a Standard
The development of a consensus standard for testing
and evaluating oxygen transfer devices in clean water
and wastewater cannot be circumvented if we are to
acquire uniformity and reliability in the art of specifying
aeration equipment. EPA fully endorses the objectives
of this Workshop and the American Society of Civil
Engineers' (ASCE) Subcommittee on Oxygen Transfer
Standards (formed under the Society's Technical Council
on Codes and Standards) who devised the Workshop
program.
EPA is under no illusions that the development of a
consensus standard(s) will be an easy task or that it can
be accomplished without a great deal of anguishing
over and wrestling with the technical issues involved.
That the attempt will be made through voluntary partici-
pation by a consortium of experts in the field such as
embodied in the above designated ASCE Subcommittee
rather than under EPA direction is highly preferred. The
stakes have become too high not to begin moving
toward a unified approach to testing and evaluation. It
should also be appreciated that after 5 yr of staff
development and program administration, the construc-
tion grants program is now in a position to take a more
active role in assessing design bases and standards.
Thus, by initiating its own long-range plan to develop a
consensus standard, the oxygen transfer field is fore-
stalling potential government intervention in an area
best left to the day-by-day practitioners.
Benefits of a Standard
In addition to greatly reducing the costs involved in
correcting aeration equipment deficiencies and making
field modifications when in-process performance does
not measure up to specified performance, other benefits
to the Nation which can be expected to evolve from
utilization and wide-spread acceptance of standardized,
verified-accurate testing and evaluation procedures are:
1. Stimulation of greater deployment of more efficient
oxygen transfer systems with concommitant savings
in energy
2. Reduction of bio-reactor volume requirements and
construction costs arising from increased usage of
effective high-rate activated sludge systems, again
due to rising confidence in and greater willingness to
specify higher-efficiency oxygen transfer equipment
12
-------
3. Elimination of excessive over-design techniques now
employed by some design firms to avoid unintended
undersizing of oxygen transfer equipment.
Decreased Aeration Power Requirements
Due to the rapid escalation of the cost of energy in
recent years, there exists for the first time in the history
of our country a compelling economic incentive to move
toward broader use of high efficiency oxygen transfer
equipment in lieu of relatively maintenance free, low
efficiency coarse bubble diffuser and sparger systems.
Increased oxygen transfer capability is now available in
several forms including a variety of fine bubble diffusers
and aerators, medium-size bubble aeration devices,
submerged turbines, surface aerators, and even im-
proved coarse bubble equipment designs. Generally
speaking, the smaller the bubble, the greater the oxygen
transfer percentage achieved. Other factors such as
mixing and bubble-contacting design, wastewater
chemistry, and wastewater temperature also strongly
impact the performance of aeration and oxygenation
equipment in wastewater. In the final analysis, though,
the size of the bubble defines the upper limit of oxygen
transfer efficiency which with proper aeration design
can be approached. In the United Kingdom and con-
tinental Europe, where expensive power has been a fact
of life for several decades, fine bubble aeration systems
have been receiving a high degree of acceptance and
utility for many years.
For comparative purposes, claimed and/or commonly-
accepted clean water oxygen transfer efficiencies and
rates are summarized in Table 1 for several generic
aeration devices. The information presented is for
standard conditions of temperature and pressure and
an initial DO concentration of 0 mg/l. It is quite
evident that if the claimed oxygen transfer capabilities
of fine bubble aeration systems in particular can
be substantiated via repeated testing using valid,
standardized test and evaluation methodology and if the
maintenance requirements of these devices are not
exorbitant, substantial savings in power are possible in
contrast to the traditional, widely-used coarse bubble
options.
Decreased Bio-Reactor Volume Requirements
The accelerated development and implementation of
oxygen activated sludge systems have revealed that
United States sanitary engineers and wastewater treat-
ment technology users are ready and willing to employ
efficient high-rate activated sludge designs. Our friends
in Europe undoubtedly feel it is about time. The defini-
tion of an efficient high-rate activated sludge system as
used here infers high mixed liquor suspended solids
(MLSS) concentrations, low reactor detention times,
high volumetric organic loadings, higher than traditional
food-to-microorganism (F/M) ratios, and the production
of a high quality secondary effluent. It should not be
confused with the low detention time, low MLSS
modified aeration system popular in the 60's, which
yields only 60-70% BOD removal.
The key to efficient high-rate activated sludge treatment
is operation with high biomass concentrations. Healthy,
highly-concentrated, readily-settleable biomass can be
effectively sustained only by efficient oxygen transfer
and the absolute elimination of zones of DO deficiency.
It may not be possible to develop high-rate air systems
with nominal reactor detention times as low as those
typically specified for municipal oxygen systems (1.5-2.5
hr based on Q) because of the disparity in the driving
force of the two gases. However, if the claimed transfer
efficiencies of the newer generation air devices shown
in Table 1 can stand up to the repeated scrutiny of a
consensus testing procedure, it appears realistic to
anticipate that traditional coarse bubble air reactor
nominal detention times of 5-7 hr can be reduced to
2.5-3.5 hr in many municipal treatment situations.
Design parameter ranges typically selected for municipal
oxygen and coarse bubble air activated sludge plants
are compared in Table 2 with this writer's projection of
a potentially feasible design and operating range for
high-rate air systems employing high efficiency oxygen
transfer equipment.
Table 1. Comparative Clean Water Oxygen Transfer
Information for Air Aeration Systems at
15 ft (4.6 m) Submergence
Oxygen Oxygen
Transfer Transfer
Efficiency Rate
System (%) (Ib/wire hp-hr)J
Fine bubble diffusers,
total floor coverage
Fine bubble diffusers,
side wall installation
Jet aerators (fine bubble)
Static aerators (medium-
size bubble)
Mechanical surface
aerators
Coarse bubble diffusers,
wide band pattern
Coarse bubble diffusers,
narrow band pattern
20-32*
11-15*
22-27*
12-14*
--
6-8**
4-6**
40-6.5
2.2-3.0
4.0-5.0
2.3-2.8
2.5-3.5**
1.2-1.6
0.8-1.2
f.1 Ib/hp-hr = 0.61 kg/kW-hr
* Taken from manufacturers' company bulletins and
technical reports where available.
** Commonly accepted ranges for historically used
variations of these generic systems.
13
-------
Table 2. Design Parameter Ranges for Municipal Activated Sludge Systems
Parameter
Aeration detention time based on Q (hr)
MLSS (mg/l)
F/M loading (kg BOD/daykg MLVSS)
Volumetric loading (Ib BOD/day-1000 ft3)J
Mixed liquor DO (mg/l)
Return sludge TSS (%)
Clarifier overflow rate (gpd/ft2)*
\ 1 lb/day/1 000 ft3 = 0.01 6 kg/day • m3
* 1 gpd/ft2 = 0.041 m3/day-m2
Oxygen
(typical)
1.5-2.5
3500-6000
0.5-0.8
1 00-200
4-8
1.5-2.5
450-650
Air.
Low Oz
Transfer
Efficiency
(typical)
5-7
1 500-2500
025-0.5
30-60
0.5-2
0.75-1.5
500-750
Air,
High 02
Transfer
Efficiency
(projected)
2.5-3.5
3000-4000
0.35-0.7
75-150
2-3
1-2
450-650
Aeration Upgrading Example
In 1975, the Wastewater Research Division of EPA's
Cincinnati based Municipal Environmental Research
Laboratory (MERL) was requested to perform a desk-top
analysis of potential alternative techniques for upgrading
a 190-mgd (8.3-m3/sec) activated sludge plant on the
East coast. The plant was designed and operated as a
modified aeration system with 2.5 hr of aeration time
based on plant influent flow. Plant effluent quality was
poor, averaging 57 mg/l of BOD and 74 mg/l of
suspended solids for the period of July 1974 through
June 1975.
The major problem was diagnosed as an inadequate air
supply and delivery system. A combination of three
centrifugal and three positive displacement blowers
provided a total air supply capacity of approximately
0.64 ft3/gal (4.8 m3/m3) of influent flow. Computer
analysis of past plant operating records indicated that
the existing spiral-roll, coarse bubble diffuser system
was achieving a process oxygen transfer efficiency of
only about 2.65%. The mixed liquor was often devoid of
any measurable DO and black in appearance with a
noticeable hydrogen sulfide odor. High soluble effluent
BOD's that varied from 25-50 mg/l were further
evidence of oxygen deficiency and anomalous activated
sludge performance.
Evaluation of potential upgrading alternatives revealed
that by far the most cost-effective solution would be
replacement of the existing coarse bubble diffusers with
higher efficiency oxygen transfer devices. Two different
types of higher efficiency aeration systems were studied,
a fine bubble diffuser system and a medium-size bubble
static aerator system, both designed for total basin-floor
coverage.
Computer modeling of anticipated performance with an
upgraded aeration system predicted that an effluent
total BOD of 30 mg/l or less would be achieved provided
effluent suspended solids could be lowered through
better bioflocculation to 45 mg/l. If effluent suspended
solids remained at the pre-upgrading 75 mg/l level due
to the prevailing overloaded secondary clarifier condi-
tion, effluent total BOD's in the range of 35-40 mg/l
would be expected. It was further predicted that effluent
soluble BOD's would be decreased to 6-1 2 mg/l
depending on wastewater temperature. Based on
historical organic loading patterns, it was estimated that
the static aerator system would be unable to meet a
target mixed liquor DO level of 2 mg/l only 10-15
days/yr and that the fine bubble diffuser system would
always meet the target level.
Capital costs were estimated for both aeration systems
and projected to July 1977 dollars. Construction in both
cases would be limited to removing and replacing the
existing diffusers, lateral piping, and downcomers
within the 20 existing basins (basin dimensions are 410
ft long x 22.5 ft wide x 15 ft SWD - 125 m x 6.9 m x
4.6 m). Modifications to the existing blower complex
and the header piping system delivering air to the new
downcomers would not be required. Electrostatic pre-
cipitator/agglomerators and bag collectors would have
to be added to the plant's existing air prefilters for the
fine bubble diffuser option but not for the static aerators.
Aeration system cost estimates were solicited from
representative manufacturers of the two aeration
system types. Total capital costs, including allowances
for engineering charges; legal, fiscal, and administrative
fees; and interest during construction, were estimated
at $1.35 million for the fine bubble diffuser system and
$0.85 million for the static aerator system.
14
-------
Operating and maintenance costs were also estimated.
For the fine bubble diffuser system, diffuser replacement
once every 10-yr was factored into the estimate at a per
occurrence cost of $110,000 in the event long-term
clogging would be experienced. An annual cost of
$1500 was also included for maintenance on the
expanded air filtration set-up required by the fine bubble
diffuser system. Estimated power costs were calculated
based on a rate of 3C/kW-hr using projected oxygen
transfer efficiencies provided by the respective
manufacturers.
Total annual cost, including amortization and the above
described diffuser replacement and air filtration
maintenance costs, for the two upgraded aeration
systems are compared in Table 3 with the existing
coarse bubble diffuser system. Because of the large
potential reduction in power costs, the information in
Table 3 suggests that the subject plant could absorb the
debt retirement charge of the new equipment and still
realize substantial savings in total annual cost with
either system over that now expended for power alone
with its coarse bubble diffusers. At the same time, the
plant would be benefiting from much improved bio-
logical performance.
The above example illustrates the type of thinking and
design approach favored by EPA in dealing with the
energy-sensitive activated sludge process. The establish-
ment of a standardized test protocol would lend
credibility to the claims of high oxygen transfer effici-
encies reported for some aeration devices and would
instill increased user market confidence in these
devices.
EPA Research Program in Oxygen Transfer
Within the constraints of its budget resources, the
Wastewater Research Division of MERL is supporting
research studies directed toward improved aeration
performance and cost effectiveness. Three separate
projects are described briefly below.
A research grant has recently (March 1978) been
awarded to ASCE to develop a tentative interim standard
for evaluating oxygen transfer equipment in water and
wastewater. The implementing entity for ASCE is the
previously mentioned Subcommittee on Oxygen Transfer
Standards. This Workshop is the first activity undertaken
by the Subcommittee with grant funds. The output of
the Workshop will provide the foundation for further
Subcommittee deliberations out of which will come the
tentative interim standard. It is anticipated that a project
period of 2-yr will be necessary to complete the task.
EPA support over the 2-yr period will total $70,000. An
important part of the final project report will be the
recommendation of verification procedures for the
tentative interim standard, which if carried out and the
data incorporated would result in a final consensus
standard.
Los Angeles County Sanitation Districts (LACSD) are
currently conducting clean water oxygen transfer tests
on six generic types of submerged air aeration systems.
EPA is contributing $50,000 toward the cost of the test
program in the form of a research contract. The objec-
tive of the project is to generate comparative oxygen
transfer data on submerged aeration devices with a
wide range of bubble formation sizes and widely-varying
reported transfer efficiencies using the same test tank
and identical test procedures throughout. The devices
being evaluated include fine bubble jet aerators, dome
diffusers, and tube diffusers; medium-size bubble static
aerators; and fixed- and variable-orifice coarse bubble
diffusers.
LACSD is carrying out the clean water tests at its Joint
Water Pollution Control Plant in an outdoor tank 20 ft x
20 ft (6.1 m x 6.1 m) with a variable water depth of up
to 25 ft (7.6 m) using the sodium sulfite reaeration
procedure. Data collection is expected to be completed
by the end of 1978. Oxygen transfer will be calculated
using several popular analytical models
Table 3. Estimated Annual Cost Breakdown for Desk-Top Upgrading Study of East Coast Plant
Aeration
System
Existing coarse
bubble diffusers
Replace with fine
bubble diffusers
Replace with
static aerators
Capital
Cost
(July '77)
--
$1,350,000
$ 850,000
Aeration
Power
Load
(kW)
2790
915
2055
Annual
Power
Costj
$730,000
$240,000
$540,000
Total
Annual
Cost Including
Amortization
$730,000
$371,000'
$615,000*
JBased on 3C/kW-hr.
* Based on 6-1/8% for 20-yr.
15
-------
The test program consists of three runs at each of four
liquid depths, 10, 15, 20 and 25 ft. (3.0, 4.6, 6.1, and
7.6 m). A different power level, varying from 0.3-1.5 hp
delivered to the water/1000 ft3 (0.008-0-0.04 kW/m3),
is employed for each of the three runs at any given
depth. Reaeration is carried to DO equilibrium for each
run.
Following the above clean water test program, process
field evaluations are planned for promising submerged
aeration systems. Equipment selection will be based on
the results of the clean water studies. An activated
sludge plant or plants with available parallel trains will
be utilized to facilitate comparative assessment of
process performance. At least 2-yr of field evaluation are
planned at a tentative cost to EPA of $150,000.
In addition to noting process performance, two other
facets of operation will be given close attention. First,
attempts will be made to predict and then measure
process oxygen transfer efficiencies using laboratory
and/or field alpha determinations and the previously-
generated clean water data. Second, operating require-
ments will be scrupulously documented, particularly as
they relate to clogging, headless buildup, and cleaning
requirements (if any).
Discussion
Arthur G. Boon
Water Research Centre, Stevenage Laboratory
England, United Kingdom
I agree that the first objective would be to adopt a
standard procedure for the testing of aeration equipment
under specified conditions. The 'non-steady state' method
is generally used to determine the rate of reaeration of
deoxygenated water by a given aeration device. Such a
method has been successfully used for pilot- and full-
scale testing of aerators in the UK. Because clean water
can easily become contaminated by traces of surface-
active substances which affect the rate of oxygen
transfer, it is normal in the UK to add sufficient anionic
detergent to achieve an average concentration of 5 mg/l
during each test and so minimize the effect of possible
contamination.
16
-------
Section II. State of the Art Review
Review of Oxygen Transfer
Model Refinements and
Data Interpretation
C. Robert Baillod
Michigan Technological University
Houghton Ml 49931
Introduction
An oxygen transfer model may be viewed as an equation
or tool employed to characterize the performance of an
aeration system. The model is necessary to translate the
results of shop tests to field application and to character-
ize the performance of field systems. Figure 1 illustrates
the relationship between models, shop tests, scale-up,
and prediction of field performance.
The word "model" implies some consideration of scale
and, indeed, scale-up of parameters obtained through the
application of oxygen transfer models to laboratory or
even to shop test systems does present a problem. Most
workers hope to avoid this difficulty by employing shop
test facilities that are sufficiently "similar" to the full-
scale field installations.
This paper concentrates on a review of various oxygen
transfer model refinements. Major emphasis is placed on
modelling the non-steady state submerged aeration of
clean water and application of the results to dirty water
respiring systems. Although some consideration is given
to scale-up, this topic is not covered in detail. Application
of any model requires the use of some method of
parameter estimation; e.g., "Direct Method", "Log Deficit
Method" or Computer Estimation and, indeed, the
method of parameter estimation may influence the
results more than the choice of model (4). This topic is
addressed in the next paper entitled "Oxygen Transfer
Parameter Estimation".
Figure 1. Relationship between Models, Shop Tests Scale-up and Prediction of Field Performance
Model
i
Parameter
Estimation
KLa \s
Ci \
SOTR I
L^ J
Scale-
Up
1
a
0
e
7
Dirty
Water
Field Appl.
Similar
Cond.
Dirty
Water Field
Application
Different
Conditions
17
-------
Models of Oxygen Transfer Systems
Mass Transfer Fundamentals
The basic model for oxygen transfer in a dispersed
gas-liquid system is given by:
[tote of Mass Transfer] I" Volumetric ~| j
per Unit Volume of UjMass Transfer Driving Force!
Liquid J L Coefficient JL J
According to the " two resistance theory" (18), the
transfer rate can be expressed in terms of the overall
transfer coefficients and resistances on either side of
the interface.
(D
(2)
W = KLa (C» - C)
W = KYa (Y - Y*)p,
where the "• superscript is employed to indicate the
local nature of the quantities and:
W = transfer rate per unit volume in clean water, m/L3
-------
diameter surface aerator in a tank 20 ft in diameter
and 27 ft in depth and reported that C values at a depth
of 1 ft were on the order of 1 mg/l greater than the
bulk average concentration. Since most of the transfer
occurred near the surface, use of the bulk average
concentration overestimated the driving force and
tended to underestimate KLa by about 20%. A graph-
ical integration procedure was suggested for analyzing
situations of non-uniform C.
Submerged Aeration
In submerged aeration, mass transfer occurs through-
out the volume and C* will vary with depth because of
progressive decreases in both hydrostatic pressure and
oxygen moje fractioq as the gas phase moves upward.
Moreover, KLa and C could also vary over the tank
volume. Many studies (5)(6)(16) have modelled
submerged aeration based on the assumptions that:
1. The volumetric transfer coefficient, KLa, is constant
over the tank volume.
2. Good mixing exists so that C is uniform over the
tank volume.
3. Oxygen is the only gas transferred.
In the case of uniformly distributed submerged bubble
aeration, these parameters might not vary appreciably.
However, where air input is non-uniform or where a
Circulatory motion or turnover is induced in the tank,
KLa and C could vary significantly over the volujne. It
should be noted here that spatial variations in C
which could exist and significantly influence oxygen
transfer in respiring systems might not be manifested
in non-steady state clean water testing.
A
In some instances C* has been assumed constant and
equal to the surface saturation value so that the
surface model given by Equation 6 would apply.
Stanton and Bradley (16) have shown that this
assumption seriously underestimates the true satura-
tion concentration and yields inflated values of KLa. In
other cases, an effective average saturation concen-
tration was defined so that:
(7) WV
. .
giving:
/*
= KLa I
Jo
A
(C*-C) dZ = KLa V (C*-C)
C'dZ
A
HY(pa+YZ)dZ
where:
C* = effective average dissolved oxygen saturation
concentration, m/L3
A = horizontal area of tank, L2
Z = vertical coordinate, L. Z = 0 at water surface.
Zd - aerator submergence, assumed equal to tank
depth, L
pa = atmospheric pressure, f/L2
7 = weight density of water, f/L3
The accuracy with which the oxygen transfer rate,
W, can be predicted depends greatly on the accuracy
with which C* can be determined. For all situations
where W > 0, C* cannot be measured directly but
must be calculated based on Equation 8 and some
known or assumed function of Y with depth.
l_n 1956, Oldshue (14) presented an equation relating
C* to oxygen content of the exit gas, hydrostatic
pressure, and the surface saturation value:
(9) C* = C?[(Ye/2Yd)
where:
(Pa + 7Zd)/2Pa]
Ye = mole fraction of oxygen in the exit gas
Yd - mole fraction of oxygen in the feed gas
Using Henry's Law to evaluate C| = HYd pa
gives:
(10) C»=(H/2)[Yd
-------
which says that the effective average saturation
concentration corresponds to the arithmetic average
of the feed gas exposed to a pressure at 2/3 the
depth and exit gas exposed to a pressure at 1 /3 of
the depth.
Yet another such relationship can be developed by
applying a few convenient approximations to the
differential oxygen balance written for the rising gas
stream. An outline of this development is given in
Attachment A. The resulting expression for the mole
fraction of oxygen in the gas phase as a function of
depth and liquid phase dissolved oxygen concentration
is (Equation A. 11):
(13) Y = C/Hpt + [Yd(pd/pa)-C/Hp
,1 exp
-KLa (Zd-Z)
where:
Pt = Pa + 7Z, f/L2
Pd = Pa +>Zd , f/L2
b - correction number, r1
*b = (M0/Ma){(paQa)/[H(pa+yZd/2)V][,t-1
M0 = molecular weight of oxygen
Ma = average molecular weight of air
Qa = volumetric flow rate of air at standard
conditions, L3/t
Pa = (Pa+ Zd/2)Ma/RT, m/L3
Using Equation 13 to evaluate Y in Equation 8 and
integrating gives:
(HYdPd-C)[1-exp(-KLa/(/>b)]
(14)
A comparison of Equations 10, 11, 12, and 14 can be
made for the case where:
Zd = 10ft
H = 3.0508 mg/l-psi
Cs = 9.2mg/l
C = 2 mg/l
Yd = 0.21
Ye = 0.19
KLa = 0.15mm"1
pa Qa/A =197 g/min-m2
pa = mass density of air at standard conditions,
mL3
By Equation 10 (Oldshue), C* = 10.15 mg/l.
By Equation 11 (Downing-Boon), C* = 10.08 mg/l
(with n = 2).
By Equation 12 (Linear M.F.), C* = 10.11 mg/l.
By Equation 14, C* = 11.6 mg/l
The values of the effective average saturation concen-
tration predicted by Equations 10, 11, and 12 agree
very closely. However, the value given by Equation 14
seems unrealistically high. This high value is probably
caused by the convenient approximations made in the
development of Equation 14.
Models Applied to Non-Steady State Clean Water Tests
Here, the overall oxygen balance on the liquid phase
becomes:
(Rate Transferred) _ (Rate of Accumulation)
V(dC/dt)
(15)
Surface Aeration
Combining Equation 15 with the surface transfer model
given by Equation 6 results in:
(16)
(17)
(18)
dC/dt = KLa(C|-C)
dC/(C*-C) = KLa / dt
»/0
ln[(C*-C)/C|-C0)]=-KLat
C = q-(C»s-C0)exp(-KLat)
Equations 18 and 19 are routinely employed to estimate
KLa from non-steady state test data. Although this model
predicts the dissolved oxygen saturation concentration
attained at infinite time to be equal to C\, many surface
aerators yield higher values. This indicates that sub-
surface transfer is taking place and that a subsurface
model might be applied.
Submerged Aeration
Combining Equation 15 with the subsurface transfer
given by Equation 7 results in:
(20)
dC/dt = KLa(C*-C)
Most analyses of non-steady state submerged aeration
have applied this relationship by considering the average
equilibrium saturation concentration, C*, to be constant
and equal to the average saturation concentration
attained at infinite time C£ , even though the value of C*
clearly increases during the course of a non-steady state
test as the oxygen content of the exit gas increases.
Thus, the model usually applied to analyze non-steady
state submerged aeration is:
(21)
or:
(22)
dC/dt = KLa'(C*-C)
in [
-------
or:
(23)
where:
(29)
= KLa (2-OTE0)/OTEc
= Q-(C":-C0)exp(-KLa't)
C* = average dissolved oxygen saturation concentra-
tion attained at infinite time, m/L3
KLS' = apparent volumetric mass transfer coefficient, t"1
This practice overestimates the average driving force
during the non-steady state test and estimates an ap-
parent volumetric mass transfer coefficient, KLa',
somewhat smaller than the true K|_a. A cogent argument
for the use of this model is that it fits data very well and,
at first glance, this agreement with observed data ap-
pears to deny the fact that C* is really a function of time.
However, it can be shown that any of the four relation-
ships (Equations 10, 11, 12 or 14) between C* and the
exit gas composition discussed previously, coupled with
Equation 20 and the assumption of constant gas rate,
will produce equations identical in form with Equations
21 and 23. The following development illustrates this for
the Downing-Boon relationship (5). An overall non-steady
state oxygen balance gives:
Rate In - Rate Out= Rate of Accumulation
(24) (M0/Ma) [Pa Qa (Yd-Ye)] - V(dC/dt)
Here an approximation is introduced by the assumption
of constant gas rate; i.e., the mass gas rate, paQa,
emitted from the aerator is equal to the mass gas rate
exiting at the tank surface. A similar development based
on variable gas rate is given in Attachment B and leads
to a different mathematical form for the function of C
with time (Equation B.4).
For the conditions of t = •*>, Yd = Ye and C* - C£ , Equation
11 becomes:
(25) C^ = Yd(pa+YZd/n)H
and using this to eliminate H in Equation 11 gives:
(26) C* = [d = correction number =
The correction number can also be related to the oxygen
transfer efficiency at zero dissolved oxygen, OTE0, by:
2MoPaQa
Comparison of Equations 21 and 28 shows that the true
and apparent values of KLa are related by:
(30)
KLa' = KLa/(UKLa/0d)
The results of similar developments based on Equations
10, 12, and 13 are summarized in Table 1. It can be seen
that, although the correction number is defined in a
slightly different manner for each model, the relation-
ships between the true and apparent values of KLa are
very similar for the models resulting from the Downing-
Boon, Oldshue and Linear Mole Fraction equations. The
resulting values of C£ are also similar for these three
models as the Oldshue and Linear Mole Fraction models
predict mid-depth saturation and the Downing-Boon
model predicts 1/n th depth saturation. The model based
on the integrated non-linear mole fraction profile yields a
C£ corresponding to aerator submergence. This is
unrealistic and is a consequence of the convenient
approximations made in the development of this model.
Typical values of C* observed in clean water tests fall
between the one-quarter and mid-depth values (15) (17).
Stanton and Bradley (16) have indicated that the assump-
tions of uniform KLa and Y = Yd result in values of C£
corresponding to mid-depth. Under these assumptions.
Equation 8 becomes:
(31)
C* =
= HYd(pa+7Zd/2)
However, at the steady state saturation condition, Y will
not be uniform with depth, but will decrease from Yd at
the aerator to a minimum value at an intermediate depth
and then increase to Yd at the surface. This is induced
by the hydrostatic pressure variation which causes
oxygen absorption in the lower region of the tank and
desorption in the upper region. Thus, an argument can
be made for not removing Y from the integrand of
Equation 8 and for accepting values of C£ corresponding
to other than the mid-depth value even though the model
is based on the assumption that KLa is uniform with
depth.
From the viewpoint of parameter estimation, the
Downing-Boon model can be fit to experimental data_
by adjusting two parameters, KLa and n or KLa and C£.
The three other models are suited to only one parameter
(KLa) estimates. This means that, for values of C* other
than the mid-depth value, the two parameter Downing-
Boon model will produce a better fit to non-steady state
test data.
Typical values of the correction number are on the order
of 1.5 min"1. whereas typical values of KLa' are on the
order of 0.15 min"1. According to the relationships given
in Table 1, therefore, typical values of the true K\_a would
be about 10% greater than those of the apparent KLa.
21
-------
Table 1. Relationships Between "True" and "Apparent" Values of KLa for Various Non-Steady State
Test Models
Relationship
Between
C* and Y,
Correction
Number, t"1
c:
KLa
Downing- Boon
Equation 11
*d
2M0paQa
Ma HV (pa+/Zd/n)
HYd (pa+7Zd/n)
KLa/(1-KLa'/«d)
Oldshue
Equation 10
*o
2MoPaQa
Ma HV pa
HYd (pa+VZd/2)
KLa'/(1-KLa'/00)
Integrated
Linear M.F.
Equation 12
4>L
2MoPaQa
Ma HV (PaOZd/3)
HYd (pa+7Zd/2)
KLa'/(1-KLa'/4>L)
Integrated
Non-Linear M.F.
Equations 8 and 1 3
4>b
MoPaO-a
Ma HV(Pa+7Zd/2)
HYd (Pa+7Zd)
-4>bln(1-KLa'/
-------
be artifactual. However, the discussions could not refute
the data completely. If the results of Albertson and
DiGregono are valid, it seems that clean water testing
would be of limited usefulness.
Field Aplication to Dirty Water Respiring Systems
at Steady State
In order to relate clean water to field conditions, it is
useful to define the following:
T - Temperature Adjustment Factor for Dissolved
Oxygen Saturation
r = CiT/C£20
and it is assumed that temperature influences on C* and
C*s are identical so that,
F = Exit Gas Correction Factor = CVC*
and it is assumed that the effect of exit gas depletion is
identical for both clean and dirty water conditions
so that:
(38)
F = CVQ = q/c:f
0 - (KLa)fT/(KLa)f2o
0' = (KLa')fT/(KLa')(20
Using these relationships in the following identity:
(39) c*20 (%T/c£20) (T/JFCi20-C)V
where:
ORTf = Field Oxygen Transfer Rate = WfV, m/t
An equivalent expression can be written in terms of the
apparent KLa as:
(42) ORTf = (KLa')20a0'(T-20) ( r/JC;20-C)V
The exit gas depletionjactor is evaluated based on the
relationship between C* and C£ employed in the
determination of KLa. For the Downing-Boon relationship
a combination of Equations 24 and 26 gives, with
Wf = R = dC/dt:
(43)
Summary and Conclusion
This review has indicated that the phenomenon of
oxygen transfer in dispersed gas-liquid systems defies a
completely rigorous analysis. Attempts at such analyses
invariably require certain approximations and assump-
tions and it is not clear whether the additional complexity
is justified by any improved accuracy of the resulting
models.
In the case of submerged aeration, two approaches to
modelling the oxygen transfer process have been review-
ed. The conventional approach neglects the effect of exit
gas depletion and is based on an "apparent KLa". A modi-
fied, somewhat more realistic approach considers the
influence of gas-side oxygen depletion and is based on a
"true KLa". However, if scale-up considerations are
ignored (i.e., gas rates, submergence, and spacing are
assumed to be equivalent for test and field conditions)
both approaches can be applied to predict field oxygen
transfer rates. Field predictions based on the conventional
model involve an "apparent alpha", whereas predictions
based on the modified model involve a "true alpha" and
errors associated with the estimation of these parameters
may overshadow any real differences between the field
oxygen transfer rate predictions given by each approach.
If the modelling effort is extended to include scale-up to
allow for varying gas rates and submergences, it appears
more logical to employ the modified model based on the
"true KLa". At any rate, various versions of logical
modified models considering the influence of exit gas
depletion have appeared in the literature and a
concensus standard will have to deal with both types
of models.
References
1. Aiba, S., T. Yamada, A. Yamamoto, and S. Shimaski.
"Oxygen Transfer in the Biological Treatment of
Sewage". In: Advances in Biological Waste Treat-
ment, Ed. by W.W. Eckenfelder, Jr. and B.J. McCabe,
MacMillian Co., New York, p. 103, 1963.
2. Albertson, O.E., and D. Digregorio. "Biologically
Mediated Inconsistencies in Aeration Equipment
Performance", Journal Water Pollution Control
Federation, 47, p. 976, 1975.
3. Baillod, C. Robert, and William C. Boyle. "Mass
Transfer Limitations in Substrate Removal". Journal
of the Environmental Engineering Division, Proceed-
ing of the ASCE. Vol. 96, No. SA2, Proc. Paper 7239,
pp. 525-545, April 1970.
4. Boyle, William C., Paul M. Berthouex, and Thomas C.
Rooney. "Pitfalls in Parameter Estimation for Oxygen
Transfer Data". Journal of the Environmental
Engineering Division, Proceedings of the ASCE,
Vol. 100, No. EE2, Proc. Paper 10451, pp. 391-408,
April 1974.
23
-------
5. Downing, A.A., and A.G. Boon. "Oxygen Transfer in
the Activated Sludge Process". In: Advances in
Biological Waste Treatment, Ed. by W.W.
Eckenfelder, Jr. and B.J. McCabe, MacMillian Co.,
New York, p. 131, 1968.
6. Ewing, L, D.T. Redmon, and J.D. Wren. "Experience
in Testing and Data Analysis of Diffused Aeration
Equipment". Paper presented at the 50th Annual
Conference, Water Pollution Control Federation,
Philadelphia, October 1977.
7. Kalinske, A.A. Discussion of Reference 1. Journal of
Water Pollution Control Federation, 47, p. 2711, 1975.
8. Lakin, M.B., and R.N. Salzman. "Subsurface Aeration
Evaluation". Paper presented at the 50th Annual
Conference, Water Pollution Control Federation,
Philadelphia, October 1977.
9. Landberg, G.G., B.P. Groulich, and W.M. Kipple.
"Experimental Problems Associated with the Testing
of Surface Aeration Equipment". Water Research 3,
p. 445, 1969.
10. Lister, A.R., and A.G. Boon. "Aeration in Deep Tanks:
An Evaluation of a Fine Bubble Diffused-Air System".
Journal of the Institute of Water Pollution Control,
No. 5, pp. 3-18, 1973.
11. Mattson, J.V., G.F. Bennett, and M.L Mattson.
Discussion of Reference 1, Journal Water Pollution
Control Federation, 48, p. 966, 1976.
12. McWhirter, J.R. "Fundamental Aspects of Surface
Aerator Performance and Design". Proc. 20th Ind.
Waste Conf., Pudue Univ. Engineering Ext., Series
118, pp. 75-92, 1965.
13. IMogaj, R.J., and E. Hurwitz. "Determination of
Aerator Efficiency under Process Conditions.". Proc.
18th Ind. Waste Conf., Purdue Univ. Ext. Series, 115,
pp. 674-683, 1963.
14. Oldshue, J.Y. "Aeration of Biological Systems Using
Mixing Impellers". In: Biological Treatment of Sewage
and Industrial Wastes, Vol. 1, Ed. by B.J. McCabe
and W.W. Eckenfelder, Jr., Reinhold, New York,
p. 213, 1956.
15. Schmit, F.L., and D.T. Redmon. "Oxygen Transfer
Efficiency in Deep Tanks". Journal Water Pollution
Control Federation, 47, p. 2586, 1975.
16. Stanton, J.L., and P.R. Bradley. "Experimental
Evaluation of Sub-Surface Aeration Systems". Pro-
ceedings of the 30th Industrial Waste Conference,
Purdue University, Ann Arbor Science Publ.,
Ann Arbor, Michigan, pp. 826-840, 1975.
17. Stukenberg, J.R., V.N. Wahbeh, and R.E. McKinney.
"Experiences in Evaluating and Specifying Aeration
Equipment", Journal Water Pollution Control Federa-
tion, 49, p.66, 1977.
18. Treybal, R.E. "Mass Transfer Operations", Second
Edition, McGraw Hill, New York, p. 95, 1968.
24
-------
Attachment A: Fundamental
Relationships in Submerged Aeration
Differential Oxygen Balances on Section of
Thickness AZ
ON. Ve
Zd
(y -dy)
_L
Az
N Vd
Gas S/c/e Balance
Rate
In
Rate
Out
Rate
Transf.
Rate of
Accum.
pNQNdy - WAdZ = AdZpNfN@y/3t)
(A.1)
where:
pN = mass density of nitrogen, m/L3
QN = volumetric flow rate of nitrogen, L3/t
puO-N = mass flow rate of nitrogen, m/t
A = horizontal area, L2
fN = fraction of tank volume occupied by nitrogen
y = mass ratio of oxygen to nitrogen in gas phase
W = transfer rate per unit volume, m/L3t
Liquid Side Balance
Rate in by
Transfer
Rate of
Accumulation
AdZ
(A.2) WAdZ
Transfer Relationship
(A.3) W - KLa(C*-C) = KYaPt (Y-Y*)
Henry's Law
(A.4) C* = HPtY
Relationship between Mass Ration and Mole Fraction
(A.5) Y = y/(y+J) J = M0/MN
Combining Equations A.1, A.3, A.4, and A.5 gives:
(A.6) pNQNdy/dZ = AKLaJHpt[y/(y+J)]-C}+AfNpNay/dt
Some convenient approximations are introduced here
to facilitate the solution of Equation A.6.
Pseudo steady state conditions are assumed so that the
term containing 3y/9t can be neglected. Rough calcula-
tion based on typical non-steady state test data indicate
that this term in on the order of 5% at each of the other
terms.
(A.7) PNQN(d//dZ) = AKLa|Hpt[y/(y+J)]-C}
Stanton and Bradley (16) solved this equation for the
special case where C = 0, subject to the boundary
conditions:
y = ye at Z = 0 (surface)
y = yd at Z = Zd (at aerator)
and obtained:
(A.8) yd-yd (1-OTE0) + J ln[1/(1-OTE0)] =
(HA/pNQN) KLa Zd (pa+7Zd/2)
where OTE0 = oxygen transfer efficiency at zero DO.
An analytical solution for the general case where C>0
requires the approximation that the total mass rate of
gas, paO-a - passing upward through the tank is constant
so that:
(A.9) PNQNdys(M0/Ma)paQadY
Equation A.7 can then be written as:
(A.10) (M0/Ma) paQa (dY/dZ) = AKLa (Hp, Y-C)
This equation can be solved by rearranging to a first
order linear form and letting pa = (p, Ma)/RT, subject to
the boundary conditions:
at Z = Zd, Y = Yd and p, = pd
at Z = Z, Y = Y and p, = p,
to give:
(A.11) Y = C/HPt + [Yd(pd/pa)-C/Hpt] exp
-KLa Zd-Z
where:
b = correction number=(M0/Ma){(p3Qa)/[H(pa+>Zd/2)V]},r1
Pa = (pa^Zd/2)Ma/RT, m/L3
25
-------
Lakin and Salzman developed a similar equation from
consideration of bubble phenomena (8).
Substitution of Equation A.11 into Equation 8 and
integrating gives:
(A.12)
* = C
(HYdpd-C>[1-exp(-KLa/4.b)]
KLa/>b
Combining this with the result of the liquid side balance
and the transfer equation, W = dC/dt = KLa (C*-C), gives:
(A.13) dC/dt ^
which is of the form:
dC/dt = KLa' (C* -C)
This result is included in Table 1.
Equation A. 10 can also be solved by letting C = Hpy*
and recognizing that:
dY/(Y-Y») = d(Y-Y*)/(Y-Y*)
to give, subject to the boundary conditions that:
at Z = O, Y = Ye and at Z = Zd, Y = Yd
(A. 14) (M0/M>aQa(Yd-Ye) =
Log mean
gas phase
driving force
and recognizing that:
(M0/Ma) PaQa (Yd-Ye) = V(dC/dt)
gives:
Attachment B: Variable Gas Rate
Modification
The development of the non-steady state test model
given by Equation 28 is based on the approximation that
the mass rate of gas, paQa, is constant with depth. The
following is an outline of the parallel development which
assumes that the mass rate of nitrogen is constant with
depth.
The non-steady state oxygen balance is written as:
(B.1) (M0/Ma)(padQadYdHM0/Ma)(paeQaeYe) = V(dC/dt)
where the subscript d denotes the values at the aerator
submergence and the subscript e denotes the values at
the tank surface. Since the mole fraction of nitrogen is
1-Y, the nitrogen balance is written as:
(B.2) (MN/Ma)(padQad)(1-Yd) = (MN/Ma)(PaeQae)(1-Ye)
Using B.2 to eliminate pae Qae in B.1 gives:
(B.3) Ye = lYd-K (dC/dt)]/[1 -K (dC/dt)]
where K = V/[(M0/Ma)(padQad)].
Using B.3 to eliminate Ye in Equation 26 and substituting
the result into Equation 20 leads to a quadratic first order
differential equation which can be solved by the method
of substitution. The solution, for Yd = 0.21, is:
dC/dt = KLa H (pa^Zd/2) (Log mean gas phase
driving force)
(A. 15)
or in terms of the liquid phase driving force:
3md/Pd>
-------
Oxygen Transfer
Parameter Estimation
Linfield C. Brown
Tufts University
Medford MA 02155
Introduction
The analysis of non-steady state oxygen transfer data has
received much attention in the recent literature. The
objective of data analysis is usually to estimate the
value of the mass transfer coefficient, KLa, from experi-
mental measurement of dissolved oxygen concentration
with time. A number of graphical and numerical proce-
dures have been proposed to estimate KLa. The
differences among many of these procedures result from
the use of a variety of values for C*, the oxygen saturation
values used to compute the driving force for the mass
transfer. Other procedures allow direct estimation of C*
from the experimental data.
There are two basic reasons for such a large number of
data analysis methods. First is the fact that in many
aeration systems C* is not constant, but varies with time
and over the depth of the aeration tank. Different methods
of accounting for this variation have resulted in different
methods of data analysis. Secondly, the integrated form
of the fundamental mass transfer equation can be
expressed both in a linear form (In dissolved oxygen
deficit vs time) and in a non-linear form (dissolved
oxygen concentration vs time). Estimates of the mass
transfer coefficient often vary with the assumed form of
the model.
The purpose of this paper is threefold:
1. To review the fundamental statistical concepts and
requirements for fitting equations to data
2. To review and evaluate the procedures used to esti-
mate overall oxygen transfer coefficients and oxygen
saturation values
3. To offer some observations concerning the statistical
merits of the various methods and to suggest some
criteria for a unified method of analyzing experimental
data from non-steady state oxygen transfer testing.
Fitting Equations to Experimental Data
There are three immediate areas of concern when one is
faced with the problem of fitting a mathematical model to
a given set of data. The first is whether the equation
being proposed does in fact correctly model the system
under study. The second is how to select the "best"
estimates of the parameters in the proposed model. The
third is to determine the precision of the computed
parameter estimates.
The first question has been discussed in the previous
paper. In general the guiding principles here are to use a
8
model that is consistent with one's theoretical knowledge
of the system and yet is simple enough to be verified by
experimental data. Because the topic of this paper is
parameter estimation, it will be assumed that the model
used is correct. However, one should always maintain an
awareness of patterns or indications in the data analysis
that may suggest model inadequacy.
Fundamental Oxygen Transfer Model
There seems to be wide agreement in the literature on
the general form of the oxygen transfer rate equation:
(1) W = dC/dt = KLa (C* - C)
where W = transfer rate per unit volume, KLa = volu-
metric mass transfer coefficient, and C* = dissolved
oxygen saturation concentration in the liquid phase. In
most cases, KLa and C* are assumed to be constant (or
relatively constant) over time. By specifying the initial
condition that C = C0 at t = 0, Equation 1 can be
integrated to yield a number of forms:
Thus in terms of log deficit:
(2) ln(C*-C) = m(C'-C0)-KLat
or in terms of dissolved oxygen concentration:
(3)
C = C*-(C'-C0) exp (-KLat)
Most non-steady state aeration tests are performed in
such a manner that the bulk oxygen concentration, C, is
measured as a function of time. Generally C is assumed
to be uniform over the tank volume. These data are then
modified to the form of Equations 1, 2, or 3 so that esti-
mates of the parameters C0, KLa, and/or C* can be
obtained. While C and t are the data to which the
equations are fit, other variables are measured and/or
controlled during a non-steady transfer test. These
variables include temperature, relative humidity, and
barometric pressure which allow correction to standard
conditions. Other variables such as gas flow rate, aeration
speed, exit gas composition, power input, tank geometry.
and water/wastewater characteristics are also recorded
to determine their effect on transfer.
Method of Least Squares
The task of parameter estimation is one of selecting the
values of the parameters C0, KLa, and C* in the model so
that the calculated or "fitted" values of dissolved oxygen
concentration agree as closely as possible with the ob-
27
-------
served or measured concentrations. The method of least
squares is a computational procedure for finding those
"best" parameter values. Specifically, the least squares
criterion states that the parameter values should be
selected so that the sum of the squares of the deviations
between observed and calculated values be a minimum.
Thus, using Equation 3 as an example, the least squares
criterion says that the best estimates of the values of C0,
KLa, and C* are those which minimize
(4)
s "
' calc)2 - minimum
where S is the sum of squares function. The values of
C0, KLa, and C* obtained from the solution of Equation 4
are called the least squares estimates.
For models that are linear in the parameters, the values
of the least squares estimates can be calculated explicitly
(9). The method involves solving a set of simultaneous
linear equations that express the least square parameter
values as a function of the data. Polynomials are
examples of linear models. Equation 2 also is a linear
model when the dependent variable is the logarithm of
the dissolved oxygen deficit.
For models that are non-linear in the parameters an
explicit algebraic solution to Equation 4 is not possible.
Rather an iterative procedure that searches over many
combinations of parameter values is required to find the
minimum. The search procedure is best accomplished
numerically and efficient schemes have been program-
med and are readily available (15). An example of a
non-linear model is the classic BOD equation (4).
Also, Equation 3 is non-linear in the parameters when C,
the dissolved oxygen concentration, is the dependent
variable.
A review of linear and non-linear least squares regres-
sion techniques can be found in a number of texts (3)
(9) (16) (24).
Error Structure of the Data
Implicit in the use of the method of least squares is the
assumption that each observation has the same inherent
error. It often happens in reality that data do not behave
in this manner. When the magnitude of the experimental
error varies with the value of the dependent variable, a
weighted least squares method should be used to analyze
the data. Normally the weighting factors are assumed to
be inversely proportional to the variance of the observed
data value. Thus, observed values with large error carry a
small weight while those with small errors are weighted
more heavily.
A number of techniques are available for determining the
error structure of a set of data. One method is to make
replicate observations of the parameters at fixed values
of the independent variables and then calculate the error
variance directly. Another technique involves an
examination of the residuals. A residual is defined as the
difference between an observed and a calculated (from a
model) data value. It is what remains of the observation
after the part explained by the model is taken away. If
the model fitted to the data is correct, the residuals
should correspond to the uncontrollable factors causing
experimental error. Thus, they should be random,
normally distributed and have a constant variance. If they
do not exhibit these properties but show a trend, then
either the model or the method of analysis should be
altered to accommodate this trend. The examination of
residuals is a powerful diagnostic tool for checking model
adequacy and error structure of the data (9).
If it is determined that the experimental error does not
have constant variance, a weighted least squares
procedure is required and two procedures can be used.
First, the data may be transformed to a new variable
having constant variance. For example, data having
constant percent error can be transformed to data having
constant error by taking logarithms. A number of other
transformation schemes have been suggested by Davies
(8). A method of obtaining weighting factors has been
described by Box and Hill (5) when replicate observations
are available.
Summary — Model/Errors
From the above discussion it can be seen that the form
of the model and the error structure of the data together
determine the appropriate method of data analysis. The
model determines whether linear or non-linear least
squares is necessary, and the data determine whether a
weighted or unweighted analysis is required. This inter-
relationship has been described by Boyle et al (6) and is
summarized in Table 1.
Table 1. General Classes of Parameter Estimation
Problems and Least Squares Calculation (6)
\Error Structure
^xof Data
Form ofX.
Model \^
Linear
Non-linear
Equal Variance
Unweighted Linear
Least Squares
Unweighted Non-
linear Least
Squares
Unequal Variance
Weighted Linear
Least Squares
Weighted Non-
linear Least
Squares
Example
To illustrate the concepts discussed above, an example
will be given using the data of Gilbert and Chen (10).
First, Equation 3 was fit using non-linear least squares.
The data, fitted values, and residuals are plotted in
Figure 1. Then, using the value of C* from that fit, the
dissolved oxygen deficits, C*-C, were computed and
their logarithms fit to Equation 2 using a linear least
28
-------
Figure 1. Data of Gilbert and Chen (10) — Equation 3 Exponential Method
O)
o
0)
o
o
U
c
0)
X
O
12 r
10
£ 8
Ci = 10.39 mg/l
KLa = 12.05 hr"1
12 16
Time (min)
20
24
28
O)
u
.o
o
U
"io
p
V)
(U
oc
0.2
0.1
0
-0.1
-02 L
-D
4 n a Q i2~ann'u DD20
i
a
Time (min)
D
a
D,
24
29
-------
squares procedure. The transformed data (In deficit),
fitted values, and residuals are plotted in Figure 2.
From the plot of oxygen concentration vs time in
Figure 1, it is seen that the fitted curve appears to
describe the data well. However, examination of the
residuals shows that for times less than about 8 min
(DO < 8 mg/l), the scatter in the residuals is much
larger than in the rest of the data. This indicates that
errors in C are not constant, but are large for low C's
(less than about 8 mg/l) and small for higher values of
concentration. This observation in the data is consistent
with field experience in measuring dissolved oxygen
concentrations in a non-steady state transfer experi-
ment. Thus, low values of C should carry a smaller
weight than the rest of the data and a weighted
analysis that accounts for this error structure should
be employed.
Inspection of the fit of Equation 2 in Figure 2 shows
that the model describes the data well for the first
20 min of the experiment (for oxygen concentrations
within about 0.3 mg/l of equilibrium). After this time,
the oxygen concentration is so close to equilibrium that
the error in the deficit becomes a large percentage of
the actual deficit. This larger relative error is magnified
by the logarithmic transformation, and its effect is
graphically shown in Figure 2. Many investigators have
suggested truncating the data at about 90% of equi-
librium (in this case at t = 14 min) to avoid this problem.
The merits of truncation will be considered later in the
paper. Note finally that the residuals at times less than
20 min are well behaved. Thus, although there are error
problems with concentrations close to equilibrium, the
logarithmic deficit transformation appears to provide a
proper weighting for the data up to that point.
Methods of Data Analysis
Oxygen Transfer
Non-Steady State
A variety of graphical and numerical procedures have
been proposed to analyze non-steady state oxygen
transfer data. The objective of the analysis is to estimate
the value of the mass transfer coefficient, KLa. The
differences among many of these procedures arise from
the use of a variety of values for C*, the oxygen satura-
tion value. Some methods assume values of C*, others
use measured or observed values, and still others allow
direct estimation of C* from the experimental data.
The term "model" is often used to describe the different
methods of analysis (i.e., surface saturation model, mid-
depth model, exponential model, etc). This is unfortunate
because nearly all the methods use the same model to
describe the transfer rate process. That model is Equa-
tion 1. As stated at the beginning of this paper, the
different methods arise from two general sources. The
first is the desire to analyze the data in a linear form;
thus the In deficit methods and direct methods. The
second is from the modeller's ability or inability to relate
the observed equilibrium oxygen concentration, C^ , to
the oxygen saturation value; thus the surface saturation,
mid-depth, and measured C* methods.
A general review of the data analysis methods is pre-
sented in Table 2. Each method is categorized according
to a number of criteria for each comparison. Each will be
considered briefly.
Surface Saturation Method. This method fits Equation 2
to the data. The value of C* is assumed to be the surface
saturation value (corrected for the temperature and
pressure of the test conditions). The In deficits are
Table 2. Summary of Data Analysis Techniques — Non-Steady State Oxygen Transfer Tests
Method
Surface
Saturation
Corrected
Saturation
Measured
Saturation
Trial and
Error
Direct
Method
Exponential
Time Constant
Equation
2
2
2
2
1
3
3
C*
Surface
Saturation
Mid-depth Ef-
fective depth
C* from data
oo
Trial and Error
Estimated in
Analysis
Estimated in
Analysis
C* from data
oo
Truncation
Usually
Usually
Only if data
are noisy
Usually
Usually
No
Yes
Error Structure
Biased
OK, except as C— C*
OK, except as C-*C*
OK, except as C-»C*
Errors increase as
AC/ At increases
Errors decrease as
C increases
?
Remarks
Inadequate for Sub-
Surface Aeration
K|_a sensitive to
assumed C*
C* and KLa deter-
mined from data
C* and KLa deter-
mined from data
Magnifies noise in
data
C0, KLa, C* deter-
termined from data
Use limited, data
sensitive to Cn
References
13, 19
1. 7, 14,
1 7, 1 9,
21, 23
10, 11, 12,
14, 19, 20
2, 19
10, 11, 19
22, 23
6, 10, 11, 12
10, 19
30
-------
Figure 2. Data of Gilbert and Chen (10) — Equation 2 Measured Saturation
2.0 -
o
-1.0 -
-2.0 -
-4.0
C£ = 10.39 mg/l (assumed)
KLa = 12.11 hr"1
12 16
Time (min)
O
20 24 28
« 0.4
31
-------
computed and plotted vs time. Values of KLa can be
obtained graphically from the slope of the line or from a
numerical linear least squares analysis.
The surface saturation model has been shown to work
well in describing the data from non-steady aeration tests
using surface aeration devices provided there is little
bubble entrainment. However, for sub-surface aeration,
the method is inadequate. The observed equilibrium
values are much larger than the assumed surface satura-
tion values. As a result, estimates of KLa are too large
(6) (19) and the plotted data are noticeably non-linear (10).
In addition, the data must be truncated at high values
of C because the computed deficits are often negative.
Corrected Saturation Method. The corrected saturation
method also fits Equation 2 to the data. As in the surface
saturation method, the value of C* is assumed but in
addition it contains corrections for depth and exit gas
composition. The mid-depth correction (17), log mean
driving force (19) (22), and effective depth (19) methods
are examples of the techniques used. Once the value of
C* is computed, the In deficits are plotted versus time
and estimates of KLa are obtained either graphically or
numerically.
The corrected saturation methods have met with mod-
erate success. The basic problem is an occasional
non-linear trend in the plot of In deficit vs time at low
deficits. This trend indicates that the assumed C* is not
the same as the equilibrium concentration, C£,, of the
observed data. To avoid the problem of non-linear plots
(and negative deficits), many investigators recommend
truncation of the data at 70 to 90% of C* (13) (19) (20)
(23). As shown by Boyle et al. (6), this is a questionable
practice. Truncation can introduce considerable error
in estimating KLa. The data will appear linear, but the
slope will be sensitive to the assumed value of C*.
Measured Saturation Method. Like the surface satura-
tion and corrected saturation methods, the measured
saturation fits Equation 2 to the data. In this case, the
assumed value of C* is the experimentally measured
value of C/ when the aeration test is allowed to
proceed to equilibrium. The In deficits are then com-
puted, plotted vs time, and the K a values obtained
by either graphical or numerical techniques.
This method has been employed successfully by a
number of investigators (10) (11) (12) (20). One of its
advantages is that it recognizes the fact that the
asymptomic value of the measured concentration is
C^ and that from a data analysis point of view, C£ is
probably the best estimate of C* to use in Equation 2.
Secondly, the plot of In deficit vs time is usually
linear when C^ is used to calculate the deficits. The
main disadvantage of this method is the need to truncate
the data at about 95% of C* because positive errors in
measuring C beyond that point may result in negative
deficits which cannot be handled logarithmically.
The previous example where Equation 2 was fit to the
data of Gilbert and Chen (10) is essentially an analysis
using the measured saturation method. As was observed
in Figure 2, there are statistical problems in the error
structure as the deficits become small, i.e., as C
approaches C£ . Truncating the data at about 95% of C£
will avoid this problem.
Trial and Error Deficit Method. This is another of the In
deficit methods that fits Equation 2 to the data. It
employs a trial and error procedure to determine the
value of C* that best fits the observed data. A series of
C* values are assumed. For each C*, the In deficits are
computed, Equation 2 is fitted using linear least squares
(In deficit vs time), and the residual sum of squares
Equation 4 is calculated. The fitting procedure and
residual sums of squares calculation are then repeated
for the next value of C*. The value of C* that gives the
smallest residual sum of squares is selected as the best
value of C* and is used to estimate KLa.
Like the measured saturation method, this data analysis
technique attempts to select a value of C* that is
compatible with the data, rather than using an assumed
value based on theoretical or physical arguments. It also
suffers from the same problems of truncation and error
structure as the measured saturation method.
Direct Method. The direct method fits Equation 1 to the
data. The slopes between successive points, AC/ At, are
calculated and plotted vs the mean value of the
adjacent concentrations, C. Thus:
(5) AC/ At = {Cj+1-Cj)/{tw-tj) and C = (C; + Cl+1)/2
The slope of the line is KLa and can be estimated using
graphical or numerical procedures. Using the direct
method. Equation 1 was fit to the data of Gilbert and
Chen (10) and the results are presented in Figure 3.
The direct method has the advantage of not requiring
that a value of C* be specified. Rather its value is esti-
mated from the intercept on the abscissa. This feature is
attractive because of the sensitivity of KLa to the assumed
C* values in the In deficit methods using Equation 2. The
main problem with the direct method is that it magnifies
the noise in the data. The process of approximating the
rate of transfer, dC/dt, by taking differences in succes-
sive concentration values results in a variable, AC/At,
that has a substantially larger error than C itself. Thus,
the plot of AC/At vs C tends to have considerable
scatter, resulting in imprecise estimates of KLa. Even if
the original data are smooth, the scatter in the direct
method of analysis is noticeable (10). The scatter is
especially disturbing if the concentration data are noisy
(23). In this case it is often necessary to discard substan-
32
-------
tial portions of the data at both low and high values
of C (22).
Inspection of Figure 3 clearly demonstrates these effects.
The observed values of AC/ At are scattered about the
fitted curve, much more so than in Figures 1 and 2. In
addition, the residuals do not appear to have uniform
variance. They are large when the value of the transfer
rate is large and small when the transfer rate is small.
This pattern suggests that a weighted least squares
procedure should be used in fitting Equation 1 to
non-steady state oxygen transfer data.
Exponential Method. The exponential method fits
Equation 3 to the experimental data. The values of dis-
solved oxygen concentration are used directly and
Equation 3 is fit to the data using non-linear least
squares procedures. Boyle et al (6) use the method of
Marquardt (15) while Gilbert and Libby (11) use methods
proposed by Reed and Theriault (18) in the analysis of
BOD data.
One advantage of this method is that is is not necessary
to truncate the data near C* because In deficits are not
computed. Secondly, the computational procedure
provides least squares estimates of both C* and KLa,
thus there is no need to assume a value of C* to compute
KLa. In light of the sensitivity of data truncation and the
assumed C* value on KLa, these are desirable aspects of
the method.
The main disadvantage of the exponential method is that
the error structure in C is not uniform, but is larger for
low values of C than for high values (opposite of In deficit
methods). The fit of Equation 3 to the data of Gilbert and
Chen (10) is an example of the exponential method. The
error structure problem was discussed there and can be
remedied by appropriate weighting procedures.
Time Constant Method. This method utilizes the
exponential nature of Equation 3 to estimate KLa. It
recognizes the fact that at integer values of the time
constant product (KLat = 1,2,3), the values of C are
respectively 63, 86, and 95% of C*. The times at which
these values of C occur are estimated from a smooth plot
of C vs time,and values of KLa are computed and
averaged. Thus:
(6) KLa = 1 /t0 63 = 2/t0 86 - 3/t0 95
The method is rapid and does not require statistical
treatment of data. It does require that t = 0 at C = 0, a
time which is difficut to estimate precisely. Furthermore,
the method uses only a limited number of data values to
estimate KLa, rather than the data as a whole. It has not
gained wide acceptance.
Observations on Data Analysis
The various methods of non-steady state oxygen transfer
data analysis have been summarized. A brief evaluation
of the procedures used to manipulate the data for
statistical analysis follows.
Truncation. The In deficit methods require data trunca-
tion as C approaches C* in order to avoid negative deficits
which cannot be handled logarithmically. Truncation is
necessary in the direct method because of the difficulty
in estimating AC/ At as the rate of transfer approaches
zero. Boyle et al (6) have shown that trunction increases
the correlation between KLa and C* and that it reduces
the precision of the estimated KLa. Thus, although
necessary with these methods (especially if the data
are noisy), it is not statistically desirable. The exponential
method does not require truncation.
Saturation Concentration. All of the In deficit methods
require that a value of C* be assumed in order to
compute the deficits. Assuming an incorrect value of C*
will bias the estimate of K[_a. The direct and exponential
methods avoid this problem by providing estimates of
both KLa and C* from the data analysis. It is important
that the distinction between data analysis and data inter-
pretation be made with respect to C*.
Statistical analysis of the data says that the value of C*
should be the estimated or measured value of C^ as
the test is allowed to proceed to equilibrium. This means
that C^, is the value that "best" (in a least squares
sense) describes the observed data. The problem of what
C^ means in terms of the physical system under study
is a matter of data interpretation. This is an issue that
should be addressed by the mechanistic model one uses
to describe the oxygen transfer process. The mass
transfer model should define how C£, relates to the
values of C* as they vary with time and depth over the
duration of the experiment. The data analysis should be
evaluated on the basis of its ability to estimate the para-
meters that can be estimated from these data. The direct
and exponential methods are attractive in this respect
because they both provide estimates of C^,.
Error Structure. The In deficit, exponential, and direct
methods have problems with the error structure of the
dependent variable. In Figure 2, the In deficit method has
increasingly large errors as the deficit decreases. The
exponential method has increasingly large errors for low
values of concentration shown in Figure 1. The direct
method has overall higher errors than the other methods,
and the magnitude of the error increases as the transfer
rate increases in Figure 3.
For the In deficit methods error structure can be improved
by truncation. However, the truncation must be minimal
if the precision of the KLa estimate is not to be com-
promised. Certainly data should be carried to 90% of
saturation and preferably to 95%. The determining factor
will be the precision with which C can be measured. A
guideline for this might be to use data to within three
standard deviations of C*. For example, if the standard
33
-------
Figure 3. Data of Gilbert and Chen (10) — Equation 1 Direct Method
O)
u
(
cr
to
to
O)
X
O
2.0
1.6
1.2
0.8
0.4
O
O
O
O
C£ = 10.35 mg/l
KLa= 12.50 hr"1
468
Oxygen Concentration, C
(mg/l)
10
12
c
E
o>
3
o
b!. -0.2 -
(D
•o
-0.4 -
Oxygen Concentration C
(mg/l)
34
-------
deviation of measuring C is 0.1 mg/l, then the data
should be truncated when C comes within 0.3 mg/l of
C*. However, the problem of determining C* remains.
Consider applying this strategy to Equation 2 and the
data of Gilbert and Chen (10) considered previously. The
data were truncated at C = 10.0 mg/l or 0.4 mg/l below
C£,. Equation 2 was fit to the In deficits and the results
with residuals are plotted in Figure 4. Clearly, the equa-
tion adequately describes the data and the residuals are
well behaved. This method requires a priori knowledge of
C£,. The value used was 10.39 mg/l which was obtain-
ed from the non-linear least squares fit of Equation 3 in
Figure 1.
For the exponential method, the error structure can be
improved by transforming the data. In Figure 1, it is seen
that as C increases, the error in C decreases. Thus, an
error structure proportional to the inverse of C seems
appropriate. According to Davies(8), the required trans-
formation is to square the observed data. Thus, the
square of Equation 3 should be fit to the square of the
concentration values. The parameter estimation procedure
still involves non-linear least squares but can be
managed easily using the methods of Boyle et al (6).
Applying this strategy to the data of Gilbert and Chen (10)
results in the fit shown in Figure 5. Note that the
equation describes the transformed data adequately and
that the residuals are well behaved. No a priori assump-
tions about C* are necessary with this method.
For the direct method, the errors in the transfer rate
were observed to increase as the transfer rate increased.
The suggested transformation (8) is to take the
logarithm of Equation 1 and fit it to the logarithm of the
observed transfer rates, AC/ At. Applying this strategy to
the data of Gilbert and Chen (10) results in the fit shown
in Figure 6. The fitted curve is fairly well defined,
although there is marked scatter in the data. Examination
of the weighted residuals shows that they tend to
increase as the transfer rate decreases. Thus, the
logarithmic transformation may have resulted in over-
weighting the low transfer rate values. Determination of
the proper weighting procedure for these data requires
further investigation. It is important to note also that this
weighting technique requires that negative transfer rates
(from noisy data) be discarded because they cannot be
handled logarithmically.
Precision of Parameter Estimates. Equally important as
the value of the estimated parameter is the precision of
that parameter estimate. The precision of a parameter
estimate is usually given by its variance or standard
deviation (standard error), and techniques for computing
these values are available (3) (8) (9) (16) (24). For the
methods and data used in this paper, the precision
of the KLa and C^ estimates are given in Table 3.
Inspection of Table 3 shows a number of significant
points. First, QJ, is estimated more precisely than is
KLa. This is true whether the exponential or direct
method is used, whether the analysis is weighted, and
whether the data are truncated. The error in C£ for the
deficit method is zero because in that method it is
assumed to be a known constant. Second, the exponen-
tial and In deficit methods provide estimates of KLa
Table 3. Precision of Parameter Estimates — Data of Gilbert and Chen (10)
Method of Analysis
Exponential
Exponential (Weighted)
In Deficit
Direct
Direct (Weighted)
KLa <">
(hr1)
12.05
11.99
12.11
12.50
10.55
Standard <•>
Error
%
0.9
0.8
2.1
4.0
9.0
c»(b)
(mg/l)
10.39
10.39
10.39
10.35
10.52
Standard
Error
(%)
0.1
0.1
0
0.9
0.4
Dependent
Variable
C
C2
WC* -C)
AC/At
ln(AC/At)
Truncated Data of Gilbert and
Exponential
Exponential (Weighted)
In Deficit
Direct
Direct (Weighted)
12.13
12.02
12.13
12.67
11.98
1.9
1.7
0.4
6.6
5.3
Chen (96% of C*)
10.37
10.38
10.39
10.29
10.38
0.5
0.3
0
2.0
0.5
C
C2
WC* -C)
AC/At
ln(AC/At)
(a)Least squares estimate.
(b)Relative standard deviation.
35
-------
Figure 4. Truncated Data of Gilbert and Chen (10) — Equation 2 Measured Saturation
u
y
2.0
1.0
-0.02
-0.04
C* = 10.39 mg/l (assumed)
KLa = 11.95 hr"1
12 16
Time (min)
20
24
0.04
"5
o
o
,'s 0.02
0
c
1
o
V
* 8
u
I -1.0
•5
•o
8
oc -2.0
-
D D
-
Qr— i
n D D n
, n
n 4 DD8 12 16 20 24
Time (min)
I— I n
m i i
D D
n
n
36
-------
Figure 5. Data of Gilbert and Chen (10) — Equation 3 Exponential Method (transformed)
\
K
CM
u
o
<5
S
§
o
o
c
0)
o>
100
80
60
40
20
= 10.39 mg/l
KLa = 11.99 hr'1
12 16
Time (min)
20
24
28
2.0
^ 1.0
o>
^
u 0
1
(A
CJ°
15
"O 1 /"I
.— — I .U
tr
-2.0
- D
D G D
D n Dn n D nn0 D D
, nQ , D , D n DU ,
4Q 8 12 nnDD UU 20 24U 28
P
Time (min)
n
37
-------
Figure 6. Data of Gilbert and Chen (10) — Equation 1 Direct Method (transformed)
0.8
•? -0.8
u
o
-1
r
-1.6
-2.4
-3.2
0
0.4
0.2
0
a
o
-1
\
O
1
r -0.2
ro
TJ
'w
0
IT
-0.4
o
o
o
CJ, = 10.51 mg/l
KLa= 10.55 hr~1
468
Oxygen Concentration C (mg/l)
10
D
D
D
DD
12
12
Oxygen
Concentration,
C(mg/l)
38
-------
having errors of 2% or less, while with the direct method,
the error in KLa is from 4 to 9%. This is numerical
confirmation of the larger error inherent in the direct
method analysis.
Third, truncation of the data (in this case at 96% of C£ )
leads to larger errors in the estimates of both KLa and
C£ . In the case of the exponential method, the error in
KLa approximately doubles, while with the direct method
it increases by about 50%. The error in KLa for the In
deficit method appears to decrease markedly with
truncation. This apparent reduction is dependent
however on a priori knowledge of C£ and is caused by
discarding the portion of the data record with the highest
error.
Fourth, the weighting procedures for the exponential
method and the direct method (truncated data) appear to
work because they provide estimates of KLa and C£
with greater precision than the non-weighted methods.
The poorer performance of the weighted direct method
on the entire data record is unexplained. Finally, of all
the methods employed, the exponential method provides
estimates of K^a and C£ with the greatest precision
(smallest error).
Summary
This paper has reviewed some of the fundamental
statistical concepts in regression analysis and a variety of
the analysis methods for non-steady state oxygen
transfer data. The methods have been evaluated on the
basis of data truncation, saturation concentration, and
error structure. The exponential method, using a
weighted non-linear regression procedure, appears most
attractive for the analysis of non-steady state oxygen
transfer data. The In direct method, using a measured
saturation concentration, may be useful for approximate
solutions. The direct method has larger inherent error.
References
1. Aberley, R.C., G.B. Rattray, and P.P. Douglas. "Air
Diffusion Unit". Journal Water Pollution Control
Federation, Vol. 46(5), pp. 895-910, May 1974.
2. Ball, R.O., and H.J. Campbell, Jr. "Static Aeration
Systems Problems and Performance". Proceedings
29th Industrial Waste Conference, Purdue University,
pp. 328-337, 1974.
3. Beck, J.V., and K.J. Arnold. " Parameter Estimation
in Engineering and Science". John Wiley and Sons,
New York, 1977.
4. Berthoues, P.M., and W.G. Hunter. "Problems
Associated with Planning BOD Experiments".
Journal of the Sanitary Engineering Division,
Proceedings of the ASCE, Vol. 97(SA3), pp. 333-334,
June 1971.
5. Box, G.E.P., and W.J. Hill. "Weighted Least Squares
When the Variance and Expected Value of the
Dependent Variable are Related". Technical Report
No. 142, Department of Statistics, University of
Wisconsin, Madison, Wisconsin, 1968.
6. Boyle, W.C., P.M. Berthouex, and T.C. Rooney.
"Pitfalls in Parameter Estimation for Oxygen Transfer
Data". Journal of the Environmental Engineering
Division, Proceeding of the ASCE, Vol. 100 (EE2),
pp. 391-408, April 1974.
7. Conway, R.H., and G.W. Kumke. "Field Techniques
for Evaluating Aerators". Journal of the Sanitary
Engineering Division, Proceedings of the ASCE,
Vol. 92(SA2), pp. 21-42, 1966.
8. Davies, O.L "Design and Analysis of Industrial
Experiments". 2nd Ed., Hafner, 1967.
9. Draper, N.R., and H. Smith. "Applied Regression
Analysis". John Wiley and Sons, New York, 1966.
10. Gilbert, R.G., and S.J. Chen. "Testing for Oxygen
Transfer Efficiency in a Full-Scale Deep Tank".
Proceedings 31st Industrial Conference, Purdue
University, pp. 291-311, May 1976.
11. Gilbert, R.G., and D. Libby. "Field Testing for Oxygen
Transfer and Mixing in Static Mixer Aeration
Systems". Proceedings 32nd Industrial Waste
Conference, Purdue University, May 1977.
12. Lakin, M.B., and R.N. Salzman. "Subsurface Aeration
Evaluation". Paper presented at the 50th Water
Pollution Control Federation Conference, Philadelphia,
October 1977.
13. Landberg, G.G., B.P. Graulich, and W.H. Kipple.
"Experimental Problems Associated with the Testing
of Surface Aeration Equipment". Water Research,
Vol. 3, pp. 445-455, 1969.
14. Mandt, M.G. "Improvements in Oxygen Testing and
Performance Rating". Paper presented at the 50th
Water Pollution Control Federation Conference,
Philadelphia, October 1977.
15. Marquardt, D.W. "An Algorithm for Least Squares
Estimation of Non-Linear Parameters". Journal
Society for Industrial and Applied Mathematics,
Vol 11 (2), pp. 431 -441, June 1963.
16. Miller, I., and J.E. Freund. "Probability and Statistics
for Engineers". 2nd Ed., Prentice-Hall, 1977.
17. Olshue, J.Y. "Aeration of Biological Systems Using
Mixing Impellers". In: Biological Treatment of
Sewage and Industrial Wastes, Vol 1, Ed. by B.J.
McCabe and W.W. Eckenfelder, Jr., Reinhold,
New York, pp. 231-240, 1956.
39
-------
18. Reed, L.J., and EJ. Theriault. "The Statistical
Treatment of Reaction-Velocity Data. II. Least
Squares Treatment of the Unmolecular Expression:
Y = L(1-e~kt)". Journal of Physical Chemistry,
Vol. 35, p. 950, 1931.
19. Scaccia, C, and C.K. Lee. "Large Scale Mass Transfer
Evaluation Techniques for Aeration Systems: A
Critical Review", paper presented at the 50th Water
Pollution Control Federation Conference, Philadelphia,
October 1977.
20. Schmit, F.L., P.M. Thayer, and D.T. Redmon.
"Diffused Air in Deep Tank Aeration". Proceedings
30th Industrial Waste Conference, Purdue University,
pp. 576-589, 1975.
21. Shell G., and R. Stein. "Testing and Application of
Static Aerators". Paper presented at the 1976
TAPPI Environmental Conference, Atlanta, April
1976.
22. Stanton, J.L., and P.R. Bradley. "Experimental
Evaluation of Sub-Surface Aeration Systems".
Proceedings 30th Industrial Waste Conference,
Purdue University, pp. 826-840, 1975.
23. Stukenberg, J.R., V.N. Wahbeh, and R.E. McKinney.
"Experiences in Evaluating Specifying Aeration
Equipment". Journal Water Pollution Control
Federation, Vol. 49(1), pp. 66-82, 1977.
24. Walpole, R.E., and R.H. Myers. "Probability and
Statistics for Engineers and Scientists". 2nd Ed.,
MacMillian Co., 1978.
Discussion
John R. Stukenberg
Black & Veatch
Kansas City MO 64114
It was not the intent of Stukenberg, et at, (Ref. 23 in the
paper) to recommend truncation of data from aerator
testing, but merely to point out what appeared to be
accepted practice by many aeration equipment manu-
facturers and design engineers. We recommend and
perform all analyses with all the data collected without
truncation or exclusion of any points.
Closure
Linfield C. Brown
Performing data analysis with all the data collected and
without truncation or exclusion of any points is usually a
good practice. However, in conducting the aeration test
from which the data are obtained, care must be given to
the following. Oxygen concentration values taken early in
the test should not be affected by residual sulfite in the
water. The test should be conducted and data recorded to
within a few percent of equilibrium.
40
-------
Review of Test Procedures
Wayne L. Paulson
University of Iowa
Iowa City 10 52242
Introduction
The purpose of this paper as set forth in the Workshop
scope statement is to review and summarize the non-
steady and steady state test procedures cited in the
literature as well as those currently in use by owners,
consultants and manufacturers. Emphasis should be
placed on the significant differences between the various
procedures. The sessions following this paper will
address many of the areas of testing procedure in more
detail. The primary focus of this paper is on the method-
ology of obtaining the necessary data for oxygen transfer
rate analysis. Most of the discussion deals with clean
water testing as it is and has been the most widely
practiced procedure.
There is a considerable amount of literature, and many
approaches are currently in use in testing. There is no
commonly accepted procedure, and engineering speci-
fications outlining test procedures vary considerably. It
can be observed that there is a real need for early
communication between owner, engineer and manu-
facturer in developing appropriate, well-described test
procedures in specifications. Several of the more
frequently cited references on testing procedures have
been utilized in preparing this paper.
The following test procedures will be discussed:
Non-Steady State Clean Water — Sulfite Deoxygenation
or Nitrogen Stripping
Non-Steady State Clean Water Plus Surfactant
Steady State Mixed Liquor
Some observations on evaluation of mixing will also be
noted.
Air Flow and Power Level
Accurate measurement of air flow rate is very critical to
reliable testing results. This is more difficult in field test-
ing due to the fact that the full-scale air flow measure-
ment system may not be applicable and/or sufficiently
accurate for a test in one of several aeration tanks. It
would be helpful for design engineers to consider this
question when anticipating equipment evaluation. In large
plants, it may be desirable to more fully instrument (air
flow measurement, sampling system) one of the tanks for
use by the owner in performance and operational studies.
Several methods of air flow measurement are being used
including the plant system, orifice plates, pilot traverses,
Annubar units and other devices. It is important to obtain
proper calibration and equations for air flow where
appropriate. System temperature and pressure are
measured to adjust the flow rate to standard conditions.
The barometric pressure and relative humidity are also
observed.
Some testing requirements include measurement of
system or diffuser head loss. Pressure taps and manom-
eters are commonly used. The data are used in meeting
specifications and for calculating horsepower require-
ments by the adiabatic compression or polytropic
formulae approach for diffused air systems.
Mechanical aeration units require various measurements
depending on the type of system. Factors such as speed
and immersion are observed. Power measurements are
made using watt meters, strain gauges, torque pick-ups
and other appropriate equipment for the type of unit.
Physical Requirements — Geometry and Power Level
Eckenfelder (8), Morgan and Bewtra (18) and Bewtra and
Nicholas (4) have reported on the importance of the test
tank geometry. It is generally recognized that tank
geometry, aerator placement and power level affect
oxygen transfer performance. Normally one attempts to
closely simulate field installation conditions for specific
applications. However, it would be desirable to have an
acceptable geometry for comparisons of aerator
performance.
Berk er al (3) of the Wastewater Equipment Council,
Process Equipment Manufacturers' Association (PEMA)
recommended some criteria in 1972. They are:
Mechanical Aeration
Tank Depth: 8 to 15 ft (as low as 3 ft for horizontal
devices)
Power Level: 0.04 to 0.3 hp/1000 gal
Note: Higher power levels and/or deeper basins will be
permitted where the aeration device is being tested
for application in specific waste treatment systems
which would operate under similar or identical
conditions.
Diffused Aeration
Basin Size: A rectangular tank with minimum length =
1.5 x depth, tank width = 3 to 6 ft
Air Supply: 12 to 45 scfm/1000 ft3
Diffuser Submergence: 8 to 14 ft
Note: Higher air supply and/or more diffuser submerg-
ence and/or smaller tank length and width will be
41
-------
permitted under test conditions where the aeration
device is being tested for application in specific
waste treatment systems which would operate
under similar or identical conditions.
The Oxygen Transfer section of Standard Methods for the
Examination of Water and Wastewater (2) notes that the
value of KLa is affected by the type of aeration equipment
and tank geometry. It is stated that "there are no
universally accepted factors that can be applied to all
aerators and test conditions to establish a reproducible
KLa." It is observed that the power applied to a unit
volume ratio is widely used to compare one system to
another. An acceptable range of 0.05 to 0.20 hp/1000
gal is cited.
Test Tank Preparation
Standard Methods for the Examination of Water and
Wastewater (2) states that the aeration basin should be
cleaned thoroughly and filled with fresh tap water. Berk
et al (3), Water Pollution Control Federation Manual of
Practice No. 5 (25) and Mixing Equipment Co. (16) also
recommend cleaning. One concern is to avoid any
residual surfactant or other chemical effects on the
testing. Several references recommend maintaining the
temperature of the fresh water at about 20°C, if possible.
Water Quality
One is normally constrained by whatever water supply is
available. Most procedures do not prescribe an overall
acceptable water chemistry. Cobalt and total dissolved
solids (TDS) or sulfate concentrations are commonly cited.
Some water quality parameters of concern that have
been noted are as follows:
Parameter
Temperature
Cobalt
Total Dissolved Solids
(Na2S04)
pH, Hardness (Alkalinity),
Iron, Manganese
Dissolved Gases
Comment
DO ConcentVation, KLa
Adequate Catalyst. Affect on
Winkler DO Measurement
Affect on Winkler DO and KLa
Testing Limitation
Cobalt Interactions and Effect on
Winkler DO
Probe Readings
Scaccia and Lee (21) have cited various Winkler DO
interferences including iron salts, residual chlorine,
inorganics, etc. They recommend the use of a chemical
blank to account for these interferences. It must be
recognized that source water quality varies widely and
modifying testing dates or altering water quality for
testing are difficult alternatives. In the Midwest, for
example, water temperature may vary from 5°C to 25°C,
initial TDS values may approach 1000 mg/l and it is
possible to experience frothing of treated surface water
supplies.
System Stability
Various references recommend aeration periods of from
20 to 30 min to achieve a steady state mixing condition
in the test basin. Landberg et al (13) observed that a
75-hp aerator requires 30 to 40 min to achieve a steady
state hydraulic regime at the Lightnin test facility. They
note that power readings and mixing velocities can be
monitored to determine steady state. They observed that
steady state power readings are achieved before steady
state hydraulic conditions result. It was noted that each
test condition (aerators and basin) will have its own time
constant.
Standard Methods for the Examination of Water and
Wastewater (2) notes that the aeration device should be
able to be operated at a constant power output for the
duration of the test. Stukenberg et al (24) have reported
on difficulties and procedures in designing shop tests to
simulate power levels of field-scale surface aerators.
Berk et al (3) state that the air temperature during the
test period should not be in excess of 10°C above or
below the temperature of the test tank water.
Various references note that the water level should be
maintained at its proper elevation and flows that are
withdrawn and overflowed should return to the test
basin.
Deoxygenation
Nitrogen stripping and sodium sulfite are the two
methods employed. Nitrogen stripping is primarily used in
laboratory and some shop testing. Pentech Industries has
utilized nitrogen stripping in shop testing. They have
observed that there is an initial lag period in the uptake
curve after returning the system to aeration.
Sodium sulfite has been supplied from various sources in
testing programs. One of the concerns with the chemical
source is the amount of impurities present that may
interfere with the test. Sodium sulfite is commonly used
in power plants for deoxygenating water, and some
sources contain catalysts. NALCO 19 is one example. It
contains cobalt. Multiple additions of NALCO 19 to a test
tank would result in an increasing cobalt concentration.
Santosite (Monsanto Chemical Co.) is a commonly used
source of sodium sulfite. It has generally been found to
be free of cobalt. Scaccia and Lee (21) have reported on
"contaminated" Santosite. They have found that some
batches of the chemical yield significantly different KLa
results for identical test conditions. They have recom-
mended conducting a laboratory performance test on
each batch of sulfite and comparing it to performance
using nitrogen stripping. One could identify acceptable
sulfite chemicals in this manner.
42
-------
There are other suppliers of sodium sulfite including
Stauffer and Allied Chemical. To avoid the potential
impurities problem, a reagent grade or high quality
sodium sulfite could be used. The cost differential is
significant for large-scale testing. Present Midwest costs
for reagent grade sulfite are approximately $0.85/lb
while Santosite costs approximately $0.25/lb. For a test
at 10 mg/l DO concentration with a 50% excess sulfite
addition, you would need approximately 1 lb/1000 gal
test volume.
Cobalt is provided either as cobalt chloride or cobalt
sulfate. A wide range of cobalt concentrations have been
and are being used in testing procedures. One concern is
to have a high enough concentration to be certain to
catalyze all of the sodium sulfite added. Morgan and
Bewtra (18) have reported on the effect of cobalt concen-
tration on the rate of sulfite depletion. Mixing Equipment
Company (16) recommends a concentration of at least
2.0 mg/l. Berk et al (3) cite an upper limit of 2.0 mg/l.
The second concern is the interaction of cobalt with the
Winkler DO determination. Kalinske et a/(11) have
reported that cobalt concentrations in excess of 0.05 mg/l
interfere with the Winkler DO test. Several references
are recommending concentrations in the 0.05 to 0.10
mg/l range. The Los Angeles County Sanitation Districts'
testing program as reported by Yunt (26) is using 0.10
mg/l of cobalt.
Scaccia and Lee (21) reported that cobalt concentrations
in the range of 0.5 to 1.5 mg/l had no effect on test
results and they normally use 0.5 mg/l. Stanton and
Bradley (23) recommend the use of 2.0 mg/l to assure
adequate catalyst and the use of chemical test modifica-
tions to account for any Winkler DO test uncertanties.
Sulfite Addition
The reaction for sulfite deoxygenation is as noted below:
Na2S03 + (1/2)02
Co
•Na2S04
The theoretical requirement is 7.88 mg/l of sulfite per
1.0 mg/l DO concentration. Sulfite additions are made in
excess and are dependent on the oxygen transfer rate
and mixing rates. Standard Methods for the Examination
of Water and Wastewater (2) and Manual of Practice
No. 5 (25) suggest 0.8 lb/1000 gal which is 22% excess.
Berk et al (3) suggest 1 lb/1000 gal which is 50%
excess. Various references cite the necessity to go to 2.0
to 2.5 times the stoichiometric value to achieve a zero
DO concentration. Most procedures do not indicate any
criteria regarding reaching zero DO or the length of time
at zero DO during a run. Stukenberg et al (24) state that
"if the basin is not deaerated completely, the results will
be erratic and difficult to interpret."
The sulfite should be dissolved prior to addition to the
test tank. Warm water increases the solubility of sodium
sulfite. Saturated solutions contain 2.23 Ib/gal at 20°C
and 3.0 Ib/gal at 30°C, Manual of Practice No. 5 (25)
states that the sulfite should be released in at least two
locations. It is important that the sulfite be uniformly and
rapidly dispersed. Some investigators have used pump-
fire hose combinations to disperse the chemical. Con-
siderable effort is warranted to attain proper dispersion in
large basin testing projects.
Sulfite Reaction and Testing Limitations
Scaccia and Lee (21) have observed interferences and
lack of reproducibility with the initial test of a newly filled
test tank. This has also been observed by others. They
recommend pre-conditioning the water by reacting it with
sodium sulfite and then aerating to saturation before
starting the test program.
The following is an example of the water quality change
as a result of sulfite addition experienced by the writer.
Test tank
Theoretical
64,000 gal
For each 100 Ib Na2S03 added,
117.5 Ib of Na2S04 is formed.
The increase per 100 Ib added is
TDS-220 mg/l and S04-149 mg/l
Initial
After 500 Ib
TDS (mg/l)
Theo. Actual
— 218
1320 1250
S04(mg/l)
Theo. Actual
— 27
771 800
It has been common practice to limit the number of runs
on a given volume of water. The total dissolved solids or
sulfate concentration has been used as a basis. Landberg
et al (13) noted an effect on the Winkler DO determina-
tion at 1.0 mg/l cobalt concentration and a Na2SO4
concentration of about 1200 mg/l. The following are
examples of testing limitations:
Berk et al (4)
Yunt (8)
Scaccia and Lee (4)
Na2SO4 Cone, of 1500 mg/l
(1010 mg/lasS04)
Na2SO4 Cone, of 2000 mg/l
S04 Cone, of 1000 mg/l
Procedural references do not make any distinction on
upper limits if probe analysis for DO concentration is
utilized. Some testing has been conducted at higher TDS
concentrations without apparent reproducibility problems.
The allowable TDS level could significantly limit the
number of acceptable tests for each tank filling on waters
with high naturally occuring TDS levels.
Sampling Locations
Standard Methods for the Examination of Water and
Wastewater (2) and Manual of Practice No. 5 (25) do not
specify the number or location of sampling stations. They
indicate that it is based on the type of device, size and
geometry of the test tank. Smart of the Ontario Ministry
of the Environment (22) recommends a minimum of three
locations. Mixing Equipment Company (16) recommends
43
-------
four sampling points. Miller (15) has shown that a
sampling system that composites samples from five
locations at mid-depth results in a sample that represents
the tank contents for diffused air systems. Berk et a/(3)
recommend four sample points at two locations. Location
1 would be close to the device at mid- and 3/4-depth.
Location 2 would be near the tank periphery at 1/4- and
3/4-depth. Yunt (21) is using four sample points at two
locations. Two samples are from mid-depth, one from
2 ft below the surface and one from 2 ft above the
tank floor.
Stukenberg et a I (24) recommend a minimum of three
sample points. They note that five to seven points is
normal with more than ten points utilized for abnormal
situations. The following sketch from Stukenberg et al
(24) illustrates recommended sample point locations for
three aeration systems.
The procedures cited in the references attempt to obtain
both a vertical and horizontal representation of the DO or
oxygen uptake. They also consider locations that will
represent regions of maximum variation in DO or oxygen
uptake.
Sampling — In Situ, Transport, Timing
One approach to DO measurement and oxygen uptake
measurement is to utilize in-tank DO probes, preferably
equipped with stirrers. The probe readings can be read
directly, recorded on multichannel recorders and/or
printed with digital output systems. This procedure yields
the largest amount of data and in some cases is a con-
tinuous DO record for the duration of the run.
Several references recommend either gravity or pumped
samples. Standard Methods for the Examination of Water
and Wastewater (2) recommends the use of submersible
pumps equipped with anti-air entrainment inlets. The
sampling interval should depend on the power level with
at least six sets of samples obtained between 10 and
70% saturation. DO analysis can be by probe or Winkler
procedure.
Manual of Practice No. 5 (25) recommends the use of
pumps with a sampling interval of 1 to 2 min and the
collection of six sets of samples between 10 and 70%
saturation. DO analysis by the Winkler procedure is
recommended.
Berk et al (3) recommend the use of pumps or siphons
with the BOD bottle volume displaced at least three
times and preferably ten times between samples. The
collection of at least six sets of samples between 10 and
80% saturation using an in-tank DO probe to set the
timing of samples is recommended along with DO
analysis by the Winkler Method.
Berk et al (31 and Standard Methods for the Examination
of Water and Wastewater (2) present an equation for
determining the sampling interval.
t = (138 W/OC) x 1.024(2°-T)
where: t = sampling interval, min
W = million Ib of water in test tank
OC = oxygenation capacity expected,
Ib/hr
138 = constant for 80% saturation
termination
T = temperature, °C
Testing with rapid oxygen uptake systems require
sampling intervals of less than 1 min.
Sanitaire (20) recommends the use of rotameters or other
flow measurement methods to monitor the pumpage rate.
They recommend a maximum residence time in the
piping of 20 sec. Using DO probes to set intervals, they
obtain a minimum of ten samples per location with eight
taken between 20 and 90% saturation. The BOD bottles
should have two displacements and the DO analyzed by
Winkler titration. (Note. A recent procedural change
recommends the use of in situ probes.)
Yunt (26) is using submersible pumps with a continuous
BOD bottle flow at five bottle displacements in 15 sec. A
+1 +2
Diffused
Aeration
*t
\N
Surface
Aeration
\
/
+3
Submerged
Turbine
Aeration
44
-------
minimum of eight samples are taken between 20 and
80% saturation. Initially, DO analysis was by probes.
However, a recent change to Winkler analysis was made.
Ewing et al (9) observed that higher KLa values were
observed with in situ probes versus piped Winkler
samples. The writer has observed a similar response with
high oxygen uptake rate systems and slow BOD bottle
filling rates.
An alternate approach used by the writer and others in
some studies is to mount a DO probe in the sampling
line. A uniform velocity past the probe is obtained.
Dissolved oxygen would be recorded continuously and
discrete samples could be taken for simultaneous Winkler
analysis. The DO response curve parallels the in-tank
probe response with a time delay equal to the sample
transport time.
Conway and Kumke (6) continuously pumped a sample
into a BOD bottle equipped with a magnetic stirrer. A
probe was used to determine the DO concentration.
Dissolved Oxygen Analysis
The Winkler titration method is widely recommended as
the DO analytical procedure. Kalinske et al (11) have
reported on a cobalt interference resulting in higher DO
readings as the cobalt concentration increases. The limit-
ing concentration for no interference is 0.05 mg/l. Lakin
(12) and Stanton and Bradley (23) have recommended a
testing modification to correct for chemical interference.
They recommend titrating a sample without the addition
of MnS04. This result would be subtracted from every
DO measurement to eliminate chemical interferences.
Scacci and Lee (21) also recommend the use of a
chemical blank to account for chemical interferences
with the Winkler analysis.
Montgomery (17) reports that the loss of iodine vapor in
titrations may result in lower DO values especially at the
higher concentrations. Procedures using excess iodide to
prevent this loss are cited.
Mixco (16) and Conway and Kumke (6) recommend that
care be taken to keep the BOD bottles at or below the
temperature of the basin.
DO probes are used in many testing programs. In some
cases, the probes provide primary data and in others they
are used to control Winkler sampling procedures or
provide supplemental information. Standard Methods for
the Examination of Water and Wastewater (2) reports an
accuracy of ±0.10 mg/l and a precision of ±0.05 mg/l for
probe DO analysis.
Probes have been successfully used; however, many
references cite concerns regarding special care, extensive
calibration requirements, sensitivity and response delays.
Gilbert and Chen (10) report on the successful use of DO
probes with a multi-channel recording system. Probes
were calibrated using saturated test water. The writer
has successfully utilized in-tank and externally mounted
sample line probes using the test water for calibration.
Results have generally compared favorably with Winkler
data. Yunt (26) initially used BOD bottle probe analysis of
piped samples in the Los Angeles County Sanitation
Districts' testing program. The probes were calibrated
using a saturated air method. Scaccia and Lee (21) cite
difficulties in maintaining probes in perfect working
order. They recommend an extensive procedure of
calibration using distilled water.
Ewing et a I (6) state that "there is some indication that
more representative and reliable non-steady state test
data may be obtained by multiple dissolved oxygen probes
than by piped samples analyzed by Winkler or probes.
The assumption of careful analysis and selection of the
probe system, its application and calibration schedule is
inherent in the above conclusion." Conway and Kumke
(6) concluded in their study that reliable oxygen analyzers
make possible more rapid, accurate measurements of
oxygen transfer.
The utilization of probes with automatic data printing
would also eliminate personal error in reading meters or
strip-chart recordings. Probes also may avoid some of the
chemical interference questions although the questions
of calibration procedure and comparative data remain.
Run Length and Truncation
As noted earlier in the discussion of sampling intervals,
several references have suggested terminating a test at
from 70 to 80% of DO saturation. Boyle et al (5) have
cited the importance of obtaining data closer to saturation
to minimize errors in estimating KLa. Recent procedural
approaches are more frequently recommending that DO
data be taken to 90 or 95% of the DO at saturation.
Boyle et al (5) also noted that truncation of DO data up
to 20% of DO saturation will not affect the precision of
the estimate of K\_a.
Dissolved Oxygen Saturation
It has been common practice to utilize DO values from
references such as Standard Methods for the Examina-
tion of Water and Wastewater (2) and modify them for
test location and submergence conditions in the case of
diffused air systems. A mid-depth or mid-submergence
value has been calculated using Standard Methods for
the Examination of Water and Wastewater (2) pressure
corrections or the method of Oldshue which recognizes
the variable concentration of oxygen in the air bubble
while in transit in the tank. Modifications of these proce-
dures have been suggested by some investigators.
Boyle et al (5) have illustrated the effect of utilizing
incorrect DO saturation values and have compared test
values with calculated values. Conway and Kumke (6)
45
-------
note that the careful determination of the test oxygen
saturation level requires more emphasis.
Various references are recommending obtaining a test
DO saturation value for a given volume of water by
extending a test run until the DO concentration remains
constant for a 15 min period. Stukenberg et aI (24) note
that the length of an extended run for DO saturation can
be estimated by multiplying 1/KLa by 6. It is normally
recommended that measured DO saturation values be
determined whenever there are power level or geometry
changes in the testing program.
Various references cite concerns regarding determining
the appropriate measured saturated DO concentration for
mechanical aerators with significant entrained air or
sparge systems. Significant DO gradients at saturation
have been noted.
Repetitive Runs and Precision
Stukenberg et al (24) recommend that three replicate
tests should be run per set of aeration conditions.
Standard Methods for the Examination of Water and
Wastewater (2) notes that a single KLa value is repro-
ducible within ±15% of the mean for multiple aerators
and ±8% for a single aerator per basin. Conway and
Kumke (6) report a variation of less than 5% of the
average values in their testing program.
DO and KLa Variations at Sample Points
Standard Methods for the Examination of Water and
Wastewater (2) observes that if the DO concentration at
all points is within 0.25 mg/l of the average DO at time
t, you may make one KLa determination. If not, then
determine individual KLa values for each sample point.
Non-parallel slopes indicate incomplete mixing, and the
test results should be voided.
Smart (22) requires that the KLa values for each sample
point shall not vary more than ±10% from the average
value. Greater variation shall indicate incomplete mixing
and invalidate the test. Other references have suggested
allowable KLa variations of from ±6.0 to ±7.5%. One
approach used is to accept the test if two thirds of the
sample points are within ±10%.
Clean Water Plus Surfactant
Various references have reported on experiences in
testing with the addition of surfactant to more closely
simulate field conditions. Ewing et al (9) reported that the
LAS concentration decreased substantially during the
test and that the results were quite varied indicating poor
reproducibility. Testing problems with excessive foaming
were also observed. Ewing et al (9) note that efforts in
this area are justified to improve methods for predicting
actual operating oxygen transfer values.
Downing and Boon (7) have conducted an extensive
testing program with anionic detergents. Upon the addi-
tion of anionic detergents, they observed a decrease in
transfer efficiency with diffused aeration and an increase
in KLa for several mechanical systems.
Steaty State — Mixed Liquor
Steady state testing is being utilized to evaluate oxygen
transfer under actual operating conditions. Several of the
testing procedural concerns involved in clean water
testing also apply to the steady state method. In addition,
steady state testing requires the evaluation of the biologi-
cal oxygen uptake rate, the alpha factor and the beta
factor.
The condition of steady state must first be achieved. It is
normally defined as the condition where the biological
oxygen uptake rate is constant and there is no change in
DO concentration. One of the difficulties with the proce-
dure is achieving steady state conditions and uniform
data throughout the test tank. Stukenberg et al (24) have
suggested establishing endogenous respiration conditions
in the test tank to provide a more uniform uptake condi-
tion. Results of this approach are discussed in a later
paper (Paper No. 21).
The oxygen uptake rate is determined at several locations.
Many factors affect the reliability of measurement of this
rate determination. Standard Methods for the Examina-
tion of Water and Wastewater (2) recommends transfer-
ring a mixed liquor sample to a BOD bottle and
measuring the uptake with a DO probe equipped with a
stirrer. Conway and Kumke (6) and others have noted the
importance of conducting the oxygen uptake measure-
ment as soon as the sample is collected to obtain reliable
results. Downing and Boon (7) and others have utilized
an off-gas analysis and mass rate balance to determine
the uptake rate.
Albertson and DiGregorio (1) have reported that field KLa
values are dependent on the biological oxygen uptake
rate and increase as the biological oxygen uptake rate
increases. This finding can have a significant impact
when clean water KLa values are used in predicting
steady state KLa values.
Alpha Factor
The alpha factor enables comparisons between KLa values
in tap water and wastewater and is evaluated in steady
state testing. Standard Methods for the Examination of
Water and Wastewater (2) suggests using a bench-scale
aerator at various mixing intensities. Stukenberg et al
(24), Gilbert (10) and others recommend using a labora-
tory scale unit where the KLa in the laboratory unit is
adjusted to the same value as in the full-scale tank.
Ewing et al (9) indicated that they were unable to obtain
consistent alpha values, bench-versus full-scale using
the matching KLa approach. Most sources note that the
alpha factor is sensitive to the type of aeration device
and mixing intensity.
46
-------
Beta Factor
The beta factor relates the DO saturation under test
conditions to that in tapwater. Standard Methods for the
Examination of Water and Wastewater (2) recommends
saturating a sample by violent hand mixing in a half-full
liter jar. Stukenberg et at (24) and others utilize bench-
scale aerators Aeration for a period of six times I/KLB
is suggested to achieve saturation.
Off-Gas Analysis
An alternate approach in steady state testing is to capture
the off gases from the aeration tank, analyze for the
oxygen content and determine the oxygen transferred by
a mass balance on the system. There is significant
potential for error in this approach; however, it has been
used successfully in several studies including those by
Downing and Boon (7), Conway and Kumke (6), Novak
(19) and Leary, Ernest and Katz (14). There is some
interest in using this system for pure oxygen closed
system analysis.
Tracer Approach
The use of tracers to evaluate oxygen transfer is pro-
posed in the paper to be presented by Mr. Lawrence A.
Neal (Paper No. 23).
ASCE (1965) Study Conclusion
Conway and Kumke (6) studied various approaches in
evaluating oxygen transfer performance and concluded
that "the reoxygenation (clean water) approach probably
should be used in testing to meet purchase specifications
and the biological uptake approach (off-gas or oxygen
uptake measurements) used for plant operating
purposes."
Mixing
Specifications for evaluating aeration equipment
frequently include mixing requirements. Oxygen transfer
test validity is also evaluated based on mixing conditions
in the test tank.
Parameters that have been used include allowable ranges
of K|_a variation as noted earlier, uniformity of mixed
liquor suspended solids concentrations, oxygen uptake
rate variations, bottom velocities and tracer dispersion
evaluations.
There is some question regarding the reliability of velocity
measurements in air-water mixtures with Price or Gurley
cup meters. Dense total floor diffuser placements do not
develop horizontal bottom velocity patterns. A velocity
meter without moving parts (Marsh-McBirney) has been
used by some investigators.
Stukenberg et al (24) note that mixing is difficult to
determine and generally velocity measurements are the
only check on equipment. They indicate that the inherent
error in the suspended solids test minimizes the value of
using uniform MLSS concentration as a criterion. They
further state that a good indication of uniform mixing is
either the uniformity of dO/dt throughout the basin in
the steady state test or the uniformity of K|_a in the non-
steady state test.
Summary
It is not the intent of this summary to propose an oxygen
transfer testing procedure for consideration. Several
observations as to areas of procedural development and
approaches that may be considered are noted. The dual
concern of ability to reliably evaluate equipment for
compliance and to truly predict performance under actual
field conditions must be implicit in any procedural
development.
1. A guideline for the extent of repetitive testing re-
quired to establish compliance should be developed.
In addition, a recommended reproducibility or preci-
sion for this repetitive testing in clean water and
steady state should be established.
2. Agreement should be reached on acceptable varia-
tions in KLa for the various sample points and the
interpretation if one or more locations are not within
this range. Should the test be invalidated or should
some data points be removed from consideration in
establishing a mean or composite value?
3. A recommended tank geometry for different types of
aerators would be desirable for shop testing
comparisons.
4. Some procedural guidelines on criteria for air flow
measurement, power measurement, system stability
requirements and related data should be included.
5. The required number of sample points and recom-
mended location criteria should be included. The
duration of a run and minimum number of sets of
data per location should also be noted with due
consideration given to truncation limitations.
6. The question of what is an acceptable water
chemistry in clean water testing as well as what the
upper limit for TDS concentration should be to
terminate testing on a given batch of water should
be resolved. Procedures for appropriate quality
control on the sodium sulfite used should be
included.
7. The level of cobalt to be used needs to be resolved in
light of adequate catalyst requirements and chemical
testing interferences.
8. The use of DO probes appears to be a desirable
approach for clean water testing. They enable an
essentially continuous data base when associated
with recording equipment. Discrete value print-outs
would be a desirable feature. Probes also are not
affected by several of the water chemistry problems
cited in references. Probes with stirrers would be
placed at in-tank locations or possibly used in sample
line mountings.
47
-------
An appropriate calibration procedure will require
development and acceptance. The successful use of
probes and recording equipment does require con-
siderable attention and care.
9. The Winkler method would be used in the calibration
procedure, and it is suggested that at least one
sample location could also be analyzed by the
Winkler method. The use of the Winkler method will
require resolution of the use of chemical interference
blanks and other procedural uncertainties such as
iodine vaporization.
10. Procedural requirements for obtaining a measured C*
should be noted. Factors affecting this determination
including type of aerator and power level should be
considered.
11. The criteria for indicating an acceptable steady state
condition should be noted. Reliable methods for
evaluating biological oxygen uptake rates should be
established with guidelines for acceptable variation
in the sample point values. The question of variability
of KLa with oxygen uptake rates should be resolved.
12. A common bench-scale procedure for alpha and beta
determination should be included.
13. In that mixing requirements are associated with
oxygen transfer equipment, it would be desirable if
recommendations for evaluating mixing could also be
developed.
There are many items to consider in developing a testing
procedure as well as the additional area of data interpre-
tation and methods of calculation. When one considers
all the potential items that could or should be delineated
and observes the many testing difficulties cited in the
literature, the drafting of a standard can appear to be an
overwhelming task. One is tempted to suggest that the
standard be as follows:
"The performance of the oxygen transfer equipment will
be tested for compliance in a professional manner on a
sound scientific basis in consultation with the owner,
engineer, contractor and manufacturer."
This Workshop definitely has a significant challenge
before it in aiding in the development of an oxygen
transfer standard.
References
1. Albertson, O.E., and D. DiGregorio. "Biologically
Mediated Inconsistencies in Aeration Equipment
Performance". Journal Water Pollution Control
Federation, 47, p. 976, May 1975.
2. American Public Health Association. "Article 207
Oxygen Transfer". Standard Methods for the Examin-
ation of Water and Wastewater, 14th Edition,
pp. 82-88, 1975.
3. Berk, W.L, D.J. Lad, D.H. Houck, and J.A. Roeber.
"Recommended Practice in the Testing of Aeration
and Oxygenation Devices". Technical Committee
Report, Wastewater Equipment Council, Process
Equipment Manufacturer's Association, May 24,
1972.
4. Bewtra, J.K., and W.R. Nicholas. "Oxygenation from
Diffused Air in Aeration Tanks". Journal Water
Pollution Control Federation, 36, p. 1195, October
1964.
5. Boyle, W.C., P.M. Berthouex, and T.C. Rooney. "Pitfalls
in Parameter Estimation for Oxygen Transfer Data".
Journal of the Environmental Engineering Division,
Proceedings of the American Society of Civil
Engineers, Vol. 100, No. EE2, Paper 10451, pp. 391-
408, April 1974.
6. Conway, R.A., and G.W. Kumke. "Field Evaluation of
Commercial Aeration Equipment". Preprint, National
Symposium on Sanitary Engineering Research,
Development, and Design, Penn State University,
1965.
7. Downing, A.L., and A.G. Boon."Oxygen Transfer in
the Activated Sludge Process". In: Advances in
Biological Waste Treatment, pp. 131-48, Edited by
W.W. Eckenfelder, Jr. and W.W. McCabe, 1963.
8. Eckenfelder, Jr., W.W. "Factors Affecting the Aera-
tion Efficiency of Sewage and Industrial Wastes".
Sewage and Industrial Wastes, 31, p. 60, January
1959.
9. Ewing, L, D.T. Redmon, and J.D. Wren. "Experiences
in Testing and Data Analysis of Diffused Aeration
Equipment". Paper presented at the 50th Annual
Conference, Water Pollution Control Federation,
Philadelphia, October 1977.
10. Gilbert, R.G., and S.J. Chen. "Testing for Oxygen
Transfer Efficiency in a Full-Scale Deep Tank".
Proceedings 31st Annual Purdue Industrial Waste
Conference, May 1976.
11. Kalinske, A.A., LD. Lash, and G.L Shell. "Cobalt
Interference in the Non-Steady State Clean Water
Test". Water and Sewage Works, 120, p. 54, 1973.
12. Lakin, M.B. "Chemical Catalyst Interference in the
Winkler Titration Determination of Dissolved Oxygen-
A Method for Correction". Water Research, Vol. 10,
pp. 961-966, 1976.
13. Landberg, C.G., B.D. Graulich,and W.H. Kipple.
"Experimental Problems Associated with the Testing
of Surface Aeration Equipment". Water Research,
Vol. 3, pp. 445-455, 1969.
48
-------
14. Leary R.D., L.A. Ernest, and WJ. Katz. "Effect of
Oxygen Transfer Capabilities on Wastewater Treat-
ment Plant Performance". Journal Water Pollution
Control Federation, 40, pp. 1298-1310, July 1968.
15. Miller, A.M. "The Development and Application of a
Standard Sampling Procedure for Full-Scale Aeration
Studies". Thesis, University of Iowa, Iowa City, 1960.
16. Mixing Equipment Company. "Testing Procedural
Recommendations". Lightnin Technical Manual,
June 16, 1977.
17. Montgomery, H.A.C. Discussion of "Atmospheric
Oxygenation in a Simulated Stream". Journal of the
Sanitary Engineering Division, Proceedings of the
American Society of Civil Engineers, pp. 356-358,
April 1969.
18. Morgan, P.F., and J.K. Bewtra. "Air Diffuser Efficien-
cies". Journal of the Water Pollution Control
Federation, 32, p. 1047, October 1960.
19. Novak, R.G. "Techniques and Factors Involved in
Aerator Selection and Evaluation". Paper presented
at the 40th Annual Conference, Water Pollution
Control Federation, New York, October 1967.
20. Sanitaire. "Aeration Equipment Performance Tests".
Company Procedural Statement, March 1977.
21. Scaccia, C., and C.K. Lee. "Large Scale Mass
Transfer Evaluation Techniques for Aeration Systems:
A Critical Review". Paper presented at the 50th
Annual Conference, Water Pollution Control Federa-
tion, Philadelphia, October 1977.
22. Smart, J. "Procedure for Evaluating Aerator
Performance — Mechanical Surface Aerators".
Ontario Ministry of the Environment, Procedures
Statement, June 1977.
23. Stanton, J.L., and P.R. Bradley. "Experimental
Evaluation of Sub-Surface Aeration Systems". Pro-
ceedings 30th Annual Purdue Industrial Waste
Conference, May 1975.
24. Stukenberg, J.R., V.N. Wahbeh, and R.E. McKinney.
"Experiences in Evaluating and Specifying Aeration
Equipment". Journal of the Water Pollution Control
Federation, 49, p. 66, January 1977.
25. Water Pollution Control Federation. "Aeration in
Wastewater Treatment". Manual of Practice, No. 5,
1971.
26. Yunt, F.W. "Aeration Equipment Evaluation, Clean
Water Test Procedures". Los Angeles County Sanita-
tion Districts Document, September 1977.
49
-------
Section III. Clean Water Testing: Shop and Field
Influence of Tank Geometry
on Aerator Performance
10
Thomas C. Rooney
Rexnord Inc.
Milwaukee Wl 53214
Introduction
Some of the technical information included in this paper
is part of a presentation to be given at the Purdue
Industrial Waste Conference in May, 1978. These results
have been made available to the Asilomar Workshop to
aid in developing a realistic and practical oxygen transfer
standard for the industry. Envirex Inc. can only capitalize
on the over two million dollars invested to date in aerator
testing if such a standard gains general acceptance and
if the relationship between the "standard test" and
process design is defined.
Envirex Inc. (Rexnord) began conducting non-steady state
aeration tests on various types of aeration devices 20
years ago, and for the last 15 years, the procedure has
remained basically unchanged. This standardized Envirex
aeration test procedure allows a comparative analysis of
the influence of tank geometry on the types of aerators
tested.
Based on the findings of the Subcommittee on Oxygen
Transfer Standards, this procedure may ultimately be
modified; however, the order of magnitude of the defined
geometry influences will likely remain unchanged.
Mechanical Surface Aerators
Types of Aerator Tested
1. High Speed — Floating units with direct-coupled, axial-
flow marine style props operating at motor speeds of
900 to 1800rpm.
2. Low Speed — Floating and bridge mounted motor/
reducer drives with output speeds of 25 to 125 rpm.
Most of the data were developed with the Envirex low
speed impeller; however, other configurations were
also examined.
3. Oxidation Ditch — Aerators consisting of a series of
discs rotating on an axis mounted across the ditch.
The number, speed, submergence, and configuration
of the discs can be changed to influence oxygen
transfer capacity and input horsepower.
Specific Tank Geometries Tested
1. High and Low Speed Aerators
Basin Dimensions Volume (gal)
a. 16' square x 15' deep 29,000
b. 20' square x 18' deep 54,000
c. 40' square x 20' deep 240,000
d. 40' wide x 80' long x 20' deep 480,000
e. 80' square x 20' deep 960,000
2. Oxidation Ditch Aerators
Length (ft) Volume (gaj)
a. 65 7,000
b. 130 75,000
c. 215 110,000
d. 285 180,000
e. 335 240,000
Influence of Geometry on Oxygen Transfer
1. Transfer rate per unit volume vs power input:
(a) Figure 1 is a generalized example, but all surface
mechanical aerators tested within the stated geometry
follow the same linear relationship.
(b) All high and low speed aerator power inputs were
from single units except in the 40 ft x 80 ft and 80 ft
square basins. Multiple units tested in these basins
fell on the same line as total horsepower.
(c) For each specific geometry, aerator efficiency in
terms of pounds of oxygen per horsepower hour, was
constant from the smallest to the largest.
(d) By varying the depth it was possible to duplicate
the horsepower/volume ratio of two different
geometries using the same aerator. In all cases the
actual oxygen transfer efficiency was dictated by the
specific test geometry.
2. Oxygen transfer rate vs surface area of test basin:
(a) Figure 2 is a generalized form followed by all types
of surface aerators tested.
(b) Each horsepower size tested at a specific depth
reached a surface area where there was little if any
50
-------
Figure 1. Oxygen Transfer Rate Per Unit Volume vs Shaft
Horsepower of Mechanical Surface Aerators
O)
OTR A
OTR B
OTR C
Maximum OTR for Aerator (A)
Due to Pumpage Alone rf
Maximum OTR for Aerator (B)
Due to Pumpage Alone /
Maximum OTR for Aerator (C)
Due to Pumpage Alone /
Note: Horsepower Aerator A > B > C
j i i 1 I I I I
Surface Area-
51
-------
effect on oxygen transfer rate. The actual limit differed
depending on the type of aerator.
(c) The limit of each aerator was determined by its
actual pumping capacity. At this limit the oxygen
transfer efficiency was the same for all horsepower
sizes if the design was consistent. Any increase in
efficiency beyond this defined limit is due to the
unique mixing combinations and circulation patterns
established by a specific geometry.
Aerator Testing Significance
1. The measured oxygen transfer efficiency of a specific
aerator can be increased as much as 50% by the
influence of tank geometry.
2. This geometry effect can be measured and quantita-
tively defined for each aerator.
3. Each aerator has an oxygen transfer rate which is
independent of geometry which can also be
determined.
Aeration Process Significance
1. Multiple aerator installations in large basins cannot
achieve the oxygen transfer rate of single units in test
volumes due to the inability of reproducing the same
mixing and circulation conditions.
2. Figure 3 compares the single unit influenced by test
basin geometry with multiple units of the same size in
a large basin. The linear portion of the lower curve is
a function of the multiple unit pumping capacity while
at the upper end there is an increase in efficiency due
to more energy input sites. It is unlikely that this
increase will ever exceed more than 10% of the
efficiency due to pumpage alone.
Submerged Turbines
Types of Aerators Tested
All turbines were bridge mounted overhung shafts with-
out a bottom support and operated at speeds from 30 to
100 rpm.
1. Radial Turbines (3 basic configurations) — Mounted
above a static air sparger.
2. Down Flow Axial Turbines (2 basic configurations)
Mounted either above a static air sparger or with a
rotating shear device.
Specific Basin Geometries Tested
Basin Dimensions
a. 20' square x 18' deep
b. 40' square x 16' deep
Volume (gal)
54,000
190,000
Influence of Tank Geometry on Oxygen Transfer
1. Oxygen transfer rate per unit volume vs air flow —
Figure 4:
(a) Mixing conditions are established by the mechani-
cal turbine without air. The intercept of the aerator
response curve indicates an initial oxygen transfer
rate as soon as air is supplied. As long as this condi-
tion remains constant, there is a linear response
between oxygen transfer rate per unit volume and air
supply up to the point where the turbine is "flooded"
(Curve B).
(b) Reducing the input turbine horsepower in the
same geometry will generate a parallel response curve
at a lower level (Curves B and C).
(c) The change in slope between the 20 ft x 20 ft basin
(Curve A) and the 40 ft x 40 ft basin (Curve B and C)
indicates a somewhat different hydraulic mixing
regime at the same depth independent of Pm/V.
2. Oxygen transfer efficiency as a percent of the input
oxygen in the air supply Figure 5:
(a) The data in Figure 5 are a portion of the total
response curve and are presented as a generalized
slope only to demonstrate the relative effect of depth
and input turbine horsepower.
(b) All geometries tested decreased in percent ef-
ficiency at higher air to volume ratios; however,
increasing the turbine power at the same depth also
increased the range of efficiencies (Curves B and C).
(c) Raising the water level improved the percent
transferred even though the ratio of input turbine
horsepower to volume decreased.
(d) The net effect of maximum depth and high mixing
turbine power to volume ratios will produce optimum
efficiencies particularly at oxygen transfer rates over
150 mg/l-hr.
Aerator Testing Significance
1. The influence of geometry on submerged turbines is
more complex than surface mechanical aerators.
2. It appears that the mixing conditions established in a
specific geometry by the mechanical turbine will be
the overriding factor.
3. Increasing the depth of a specific basin increases the
oxygen transfer rate at a lower turbine horsepower to
volume ratio.
4. The performance of submerged turbine aerators can
be changed 100% by selecting the geometry and input
power/volume ratio.
Aeration Process Significance
1. Fortunately, submerged turbines are often operated as
single units in one basin where full-scale testing can
be employed. This would allow high rate systems to
be designed and verified.
2. Multiple installations are likely to have operating
characteristics which differ greatly from single unit
tests.
52
-------
Figure 3. Oxygen Transfer Rate Per Unit Volume vs Total
Shaft Horsepower of Mechanical Surface Aerators
CD
£
_D
o
Q)
Q.
nj
CC
c
nj
CD
O)
Maximum OTR at Same P/V
as Test Basin
OTR Due to Individual
Pumping Capacity Only
Changing OTR Due to
Slight Influence of Basin
if Surface Area/Unit Less than
Maximum for this Aerator
Total Shaft Horsepower-
E
_3
O
100
Figure 4. Oxygen Transfer Rate Per Unit Volume vs Air Flow
for a Submerged Turbine
Note: All Turbines Same Basic Configuration
(A)
o>_ 75
o> _•_
QC ^5
•ft 50
c
CO
5
o
25
Dimensions
20 ft2 x 16 ft
40 ft2 x 16 ft
40 ft2 x 16 ft
40 ft2 x 10 ft
- Turbine Shaft hp
V = Volume in 1000 gal
100 200 300 400 500
Air Flow (scfm)
600
700
800
53
-------
Figure 5. Percent Oxygen Transferred to Water from Air
Supply vs Air Supply/Volume
50
01
CO
c
0)
o>
1)
o
o5
o.
40
30
20
10
Basin Dimensions Pm/V
C 40 ft2 x 16 ft 0.12
A 20 ft2 x 16 ft 0.16
B 40 ft2 x 16 ft 0.20
E 20 ft2 x 18.3 ft 0.14
5.0 10.0
Air Supply/Volume (scfm/1000 ft3)
15.0
20.0
Figure 6. Oxygen Transfer Rate Per Unit Volume vs Air
Supply/Volume in 16 ft2 Tank
125
CD
_3
O
100
c
I - 75
§x
OC CO
I" 50
0)
o>
x
O
25
10 20 30 40 50 60
Air Supply/Volume (scfm/1000 ft3)
70
80
90
54
-------
Diffused Aeration
Types of Diffusers Tested
1. The influence of geometry on fine bubble or porous
media diff users is part of a current study and not
available at this time.
2. The coarse bubble diffuser data presented were
developed with two types of diffusers manufactured
by Envirex Inc. Side-by-side comparative tests of these
two diffusers and other designs have demonstrated
that the type of coarse bubble diffuser itself has little
influence on oxygen transfer.
Specific Tank Geometries Tested
Basin Dimensions Volume (gal)
a. 16'square x 15' deep 29,000
b. 20' square x 20' deep 60,000
c. 30' wide x 6' long x 21' deep 29,000
d. 30' wide x 80' long x 21' deep 380,000
Influence of Tank Geometry on Oxygen Transfer
1. Oxygen transfer rate per unit volume vs airflow per
unit volume.
(a) Figure 6 (16 ft square x 1 5 ft deep). There is a
linear response between the oxygen transfer rate and
increasing air flow per volume at each depth from 4 ft
to 13 ft.
(b) Figure 7 (30 ft x 6 ft x 21 ft deep). The 13 ft depth
response line corresponds reasonably well with the
16 ft x 16 ft geometry in Figure 6; however, at 19 ft, the
slope is not predicted. For comparison, installing sock
diffusers at the same location develops an entirely
different type of response curve.
(c) Figure 8 (20 ft square x 20 ft deep). At 15 ft, the
effect of depth follows the 13 ft and under response
of the previous two geometries, but the 20 ft curve
demonstrates a very rapid increase in transfer rate
with only a slight change in air flow.
(d) Figure 9 (30 ft x 80 ft x 21 ft deep — diffuser
pattern "A"). The linear effect of increased air flow at
each depth 13 ft and below developed in previous
geometries is continued up to 19 ft. The magnitude of
the oxygen transfer rate at 19 ft is less that that found
at 15 ft in the 20 ft square tank (Figure 8).
(e) Figure 10 (30 ft x 80 ft x 21 ft deep — diffuser
pattern "B"). The effect of diffuser pattern is demon-
strated when compared to Figure 9. For this pattern,
the 16 ft depth is slightly above the 15 ft curve in the
20 ft square tank (Figure 8) indicating good correlation
for this one depth only.
Aeration Test Significance
1. For a specific geometry there will be a linear increase
in the oxygen transfer rate per unit volume as the air
flow increases only if the established mixing condi-
tions remain the same. A non-linear response
indicates a change in these conditions which may or
may not be accomplished in a dissimilar geometry.
2. At depths less than 15 ft there is little effect of
geometry although diffuser placement can affect
results by 25%.
3. At depths over 15 ft when one or more of the tank
surface area dimensions are less than the water
depth, a potential exists for creating "unique"
geometry effects.
4. The influence of geometry and diffuser patterns can
affect results at depths over 15 ft up to 100% which
renders any definition of oxygen capacity for a unit
diffuser meaningless.
Aeration Process Significance
1. Large-scale diffused aeration systems designed on the
basis of data from aeration tests which were biased
by tank geometry may not meet process requirements.
2. The linear response of the oxygen transfer rate per unit
air supplied without the influence of tank geometry
could lead to more predictable diffused aeration systems.
General Conclusions
1. Basin geometry does affect the mixing regime estab-
lished by a specific aeration device and thereby
significantly influences its oxygen transfer rate.
2. Oxygen transfer rates based on testing in "unique"
geometry configurations may not be representative of
full-scale designs.
3. For all major aeration devices the influence of basin
geometry is a definable parameter whicli could lead to
more efficient aeration basin designs.
Recommendation to the Committee
A program to develop data on the influence of basin
geometry for all types of aeration equipment would
appear to be a venture beyond the scope of this Sub-
committee. It should remain the responsibility of the
equipment manufacturer to identify the operating
characteristics of their own equipment.
At the present time, the design engineer is faced with
the difficult task of selecting aeration equipment whose
performance may be biased by either test basin geometry,
or test procedure, or both. If this Subcommittee is suc-
cessful in establishing a standardized test procedure, the
engineer can start to compare oxygen transfer efficiency
claims and determine whether or not a manufacturer has
thoroughly defined all the operating characteristics.
It would seem more likely to fall on the EPA to confirm
that basin geometry does influence oxygen transfer for
a specific device. Without that confirmation, any cost
effective comparison could err greatly by comparing the
worst situation for one device with the biased oxygen
transfer efficiency of another.
55
-------
03
E
3
O
125
100
v 75
c
ra
50
25
Figure 7. Oxygen Transfer Rate Per Unit Volume vs Air
Supply/Volume in 30 ft x 6 ft Tank
125
E 100
"5
03 .
5-
CC 03
fc§
c
re
c
V
I
O
75
50
25
10 20 30 40 50 60
Air Supply/Volume (scfm/1000 ft3)
Figure 8. Oxygen Transfer Rate Per Unit Volume vs Air
Supply/Volume in 20 ft2 Tank
70
80
j I
j I
10 20 30 40 50
Air Supply /Volume (sdm/1000 ft3)
60
56
-------
Figure 9. Oxygen Transfer Rate Per Unit Volume vs Air Supply/Volume
in 30 ft x 80 ft Tank with Diffuser Pattern "A"
0)
E
2
o
125
100
^ 75
—
(0 X
GC 0)
- 50
re
25
Note: Data at 4, 9 and 13 ft from Figure 6
10 20 30 40 50 60
Air Supply/Volume (scfm/1000 ft3)
70
80
_
O
Figure 10. Oxygen Transfer Rate Per Unit Volume vs Air Supply/Volume
in 30 ft x 80 ft Tank with Pattern Diffuser Pattern "B"
125
100
0) .
S~ 75
CC O)
/
CD
0)
5?
X
O
50
25
' '
Note: Data at 15 ft from Figure 8
' ' ' I_J 1 l I
10 20 30 40 50 60
Air Supply/Volume (scfm/1000 ft3)
70
80
57
-------
Discussion
Jerome D. Wren
Sanitaire — Water Pollution Control Corporation
Milwaukee Wl 53201
The report by Rooney was based on work conducted by
one manufacturer on various types of aeration devices. It
is suggested that the work of others should be included,
at least by reference.
Several references on the subject, as it relates to diffused
aeration are as follows:
Bewtra, J.K. and W.B. Nicholas, "Oxygenation from
Diffused Air in Aeration Tanks", Journal Water Pollution
Control Federation, Vol. 36 (10), p. 1195, October 1964.
Schmit, F.L., J.D. Wren, and D.T. Redmon, "The Effect of
Tank Dimensions and Diffuser Placement on Oxygen
Transfer", Presented at the New England Water Pollution
Control Association Meeting, Dixville Notch, NH on June
9, 1976, (Scheduled for publication in the July 1978
Journal Water Pollution Control Federation).
Closure
Thomas C. Rooney
Mr. Wren's comment is correct in that my report to the
Subcommittee is work conducted by one manufacturer. It
was not our intent to slight other studies or investiga-
tions by omission, but neither were we presenting a
literature review on the state-of-the-art. While such a
survey would be a worthwhile effort, in this instance we
were attempting to demonstrate that basin geometry had
influenced the rate of oxygen transfer for different
generic aeration devices. The purpose was to generate
discussion at the Workshop.
It is our contention that ultimately the burden of proof in
determining the absence or presence of basin geometry
effects remains with each manufacturer of a specific
aeration device. We also concluded that comparative
tests of oxygen transfer equipment conducted in research-
scale aeration tanks, without confirmation on a scale at
least approaching full-size design, could be in error by
some magnitude.
If Mr. Wren feels the work of others would shed further
light on the subject, it would be contrary to the stated
purpose of the Subcommittee to exclude these refer-
ences. Our only reservation is that these two suggested
articles also do not represent the sum total of document-
ed studies in this field.
58
-------
Influence of Mixing
in Aeration
Ronald N. Salzman and Michael B. Lakin
Mixing Equipment Company, Inc.
Rochester NY 14603
Introduction
Mixing is a key element of aerator performance.
Whether the aeration is accomplished by a mechanical
surface device, subsurface air, or mechanical dispersion
of subsurface air, the basic mixing requirements must
be satisfied. Mixing is defined as the circulation which
convects the oxygen enriched fluid throughout the basin
and provides the degree of agitation necessary to main-
tain solids suspension.
A precise process model would separate the mass
transfer at the gas liquid interface from the ensuing
convective dispersion in the tank. However, since a
simplified first order differential is generally used to
describe the overall process, mixing effects are intrinsic
to the mass transfer model. Thus, the mass transfer
coefficient is indicative of the mixing and mass transfer
interactions.
Surface Aeration
The interaction of aerator performance and mixing is
illustrated in the first series of figures. Figure 1 is a
schematic of a typical surface aerator operation. In
analyzing the surface aerator clean water performance, a
first order mass transfer model is used to characterize
the oxygen uptake rate. The rate of oxygen absorption is
expressed in the form:
11
(D
dC/dt = KLa (C*-C)
where the mass transfer coefficient, K\_a, is determined
by analyzing the change in bulk dissolved oxygen level
in the basin as a function of time. In this form, the
equation implies that oxygen transfer is occurring
throughout the basin during the period of observation. As
illustrated in Figure 1, the aeration zone comprises only a
small percentage of the total volume. The oxygen enriched
liquid is then distributed throughout the basin by means
of the pumping action of the aeration device. Therefore,
in employing the above equation, the mass transfer
coefficient determined characterizes a combination of the
basin hydraulics and the mass transfer occurring at the
liquid-gas interface.
Other assumptions implicit in this simplified surface
model are:
1. Oxygen transfer occurs at the prevailing surface
conditions of temperature and barometric pressure,
and the oxygen concentration in the gas remains
constant.
2. Equilibrium prevails at the gas-liquid interface and is
governed by Henry's Law.
3. The liquid film comprises the major resistance to mass
transfer.
4. Environmental factors other than barometric pressure
are negligible.
5. A well-mixed condition exists in the basin such that a
meaningful average DO level can be determined from
the sample point measurements.
rigure 2 illustrates a less than ideal utilization of a
surface aerator. For reasons of tank geometry or improper
impeller sizing criteria, the primary bulk flow includes
only the upper portion of tank volume; the weak second-
ary flow pattern has been induced in the lower section. A
reaeration test would show the lower region achieving a
slower oxygen uptake rate and possibly lower bottom
velocities as well. In a waste operation, this system could
result in sludge buildup and anaerobic digestion.
Assuming that neither the aerator nor basin can be
altered, a lower mixing impeller is necessary to alleviate
this problem. Several alternatives are available. Figure 3
illustrates one approach to solving this problem. An
uppumping lower impeller is installed near the bottom
and depending on the particular operating condition, this
can break the short circuiting flow pattern, establishing a
single top to bottom flow regime.
In situations where the lower flow pattern is highly
developed, it is more efficient to reinforce the lower
pattern rather than oppose it. Hence, Figure 4 shows a
downpumping lower impeller as a possible solution to
the similar problem illustrated in Figure 2. This procedure
may be preferable in those waste streams where solids
suspension is considered a major problem. As flow is
directed strongly to the bottom of the basin, high bottom
velocities can be maintained.
Oxygen is transferred to the lower pattern by diffusion
through the contact zone of interchange. Sufficient power
must be supplied to the mixing impeller to cause vigorous
and turbulent motion in this region. For waste operation,
a dissolved oxygen gradient may exist, but positive DO
levels will be maintained at all points.
One further method by which this situation can be
remedied is through the use of a draft tube. This must be
of sufficient length and design so as to permit solids to
59
-------
Figure 1. Schematic Visualization of Ideal Surface Aerator Operation
Ideal Surface Aeration
Circulation
Figure 2. Surface Aeration with Induced Secondary Flow
Surface Aeration with
Short Circuiting
1
Aeration Zone '
Zone of Low Activity
60
-------
Figure 3. Surface Aerator with Uppumping Lower Impeller — Single Flow Pattern
Uppumping Lower Impeller
I
Figure 4. Surface Aerator with Downpumping Lower Impeller — Dual Flow Pattern
Downpumping Lower Impeller
I 1
Zone of Interchange
61
-------
be drawn off the basin bottom.
To clarify these points further, surface reaeration tests in
a 223 m2 by 9.09 m deep basin (2028 m3) have been
made comparing performance with lower uppumping and
downpumping impellers. For both tests, the surface and
submerged impellers used were exactly the same. A pro-
vision was made so that the angle of attachment of the
blades on the lower impeller could be changed so as to
convert from uppumping to downpumping modes. The
power draw was approximately 70 shaft hp for the
surface aerator and 25 shaft hp for the lower impeller.
The clean water reaeration test data corresponding to a
well-mixed single flow pattern (lower uppumping impeller,
Figure 3) are presented in Figure 5. The circles represent
data at a surface sample point while the triangles are
taken from a lower sample point. The net result is that at
each pump location, the measured rate of oxygen transfer
(the slope of the line) is the same. Displacement of the
lines with respect to one another delineates the time
delay in convection of the oxygen enriched fluid to
respective portions of the basin. The magnitude of this
shift is dependent on the relative rates of oxygen transfer
and the blending.
For a surface aerator with a dual flow pattern (Figure 4),
the reaeration data will look somewhat different from the
case just cited. This difference is depicted in Figure 6.
The slopes of the DO deficit vs time curves are different
for respective sample points near the surface and
bottom. There is also a displacement of the lines, which
is an indication of delay in convecting the oxygen
enriched fluid. This displacement is greater than for the
previously cited case because of the additional resistance
created by the zone of interchange.
A comparison of Figures 5 and 6 demonstrates how
blending is an intrinsic and integral part of the mass
transfer process. Since both configurations employ the
same surface aerator at the identical speed, both devices
should be capable of delivering the same number of
kg/hr (Ib/hr) of oxygen. Yet the unit with the dual flow
pattern yielded an oxygen transfer rate 75% of that for
the single flow pattern aerator. This provides convincing
evidence that the defined mass transfer coefficient is an
overall coefficient reflective of the combined mass
transfer and convection processes within the basin.
A second element of aerator performance is the ability to
suspend solids. The data presented in Figures 5 and 6
clearly suggest that the downpumping impeller provides
a far stronger flow field along the basin bottom. There-
fore, this unit might be preferable in waste streams with
a documented solids suspension problem.
The data presented in Figures 5 and 6 represent an
unusual application for a surface aerator as a 70-shaft-
hp surface aerator is not generally applied for mixing and
aerating a basin 9.09 m in depth.
These examples demonstrate the complex interaction in
performance criteria. While the uppumping lower mixing
impeller provides a single flow pattern and higher transfer
efficiencies, the associated bottom velocities of this unit
are on the order of one-half a foot per second. For the
downpumping lower mixing impeller, the transfer
efficiency diminishes (this unit delivered 75% of the
oxygen transferred with uppumping for the same horse-
power) but the bottom velocities are higher and suitable
for wastes with persistent solids suspension problems.
Subsurface Aeration — Theory
The simplest subsurface aeration device is a single orifice
bubbler. With this device, a string of constant diameter
bubbles are released and rise to the surface under the
effects of gravity. Figure 7 (5) shows the empirical
relationship determined to express bubble size of air in
water as a function of gas velocity from a 0.061 cm
orifice (1). The bubble diameter for gas rates less than
the critical value is a constant. The functional relation-
ship for specifying this bubble size has been determined
to be of the form:
(2)
dg =
where:
db = bubble diameter (L)
KI = a constant
d0 = orifice diameter (L)
a = surface tension (m/L)
Pi = liquid density (m/L3)
pg = gas density (m/L3)
At high gas flow rates, the formation of bubbles is hin-
hindered by the presence of the preceding bubble. This
transition point is defined by a critical velocity:
(3)
where:
VCRIT ~ critica' bubble velocity (m/t)
K = a constant
M | = liquid viscosity (f t/L2)
Other studies (2) (4) have been devoted to determining
bubble rise terminal velocity as a function of diameter
(Figure 8). Combining the information of Figures 7 and 8,
exact relationships are obtained for interfacial surface
area and bubble residence time. For low stripping
efficiencies and shallow depths (30-180 cm), Coppoch
and Meiklejohn (1) were able to determine KL independent
of a. This is a pure mass transfer coefficient untainted by
convection or turbulence. These results are presented in
Figure 9. Due to the physics of the experiment, the key
elements have been quantified within a narrow range.
62
-------
Figure 5. Reaeration Test Data — Single Flow Pattern
10
9
8
7
3
o
Q
0 2
(KLa)2 = 60 ln(6.53/2.01)/(14-2) = 5.90 hr'1
= 60 ln(5.70/1.71l/(14-2) = 6.02 hr1
= 6.05 hr1 Based on Average DO @ Each Time
Average Bottom Velocity
0.64 ft/sec
Based on Handbook Saturation
6 10 14
Time (min)
18
Figure 6. Reaeration Test Data — Dual Flow Pattern
o>
O
O
(KLa)2 = 60 ln(6.4/1.9)7(20-4) = 4.55 hr1
(KLa)! = 60 ln(5.3/1.31 )/(20-4) = 5.24 hr1
= 4.65 hr'1 Based on Average DO @ Each Time
Average Bottom Velocity
1.06 ft/sec
Based on Handbook Saturation
2 4 6 8 1012141618 20 22 24
Time (min)
63
-------
Figure 7. Variation in Bubble Diameter with Gas Velocity
0.6
0.5
0.4
0.3
E
RJ
b
_
50.2
3
03
0.1
0
Bubble Formation
Air/Water System
d0 = 0.061 cm
db = 0.258 cm
= 64 cm/sec
0.1 1.0 10 100 1,000 10,000
Gas Exit Velocity (cm/sec)
Figure 8. Variation of Terminal Velocity with Bubble Diameter
36 r
30
x 24
o
1 18
I 12
o>
0
0.4 0.8 1.2
Bubble Diameter (cm)
1.6
2.0
2.4
64
-------
These key elements are the bubble size, the relative
velocity and the residence time. Even with such well
defined conditions, there is still considerable data scatter
(or uncertainty) for defining a functional relationship
between bubble size and mass transfer. In aeration
processes, submerged impellers, pumps and jet devices
create conditions of extreme shear gradients and high
levels of turbulence. There are no technical treatises
dealing with bubble size as a function of these flow
characteristics.
Subsurface Aeration — Current Practice
In any commercial subsurface aeration system, the
controlling parameters are: 1) bubble size, 2) relative
velocity and 3) residence time. These variables are in
turn dependent upon the physical environment surround-
ing the bubble release and flow path. The combined
effects of turbulence, buoyancy and interfacial tension
control the fragmentation of the gas phase as it is
released while the size distribution of the bubbles and
the hydraulic flow patterns, within the basin, control the
residence time of the gas. Dispersion, coalescence and
convection characteristics are all dependent on the
particular device and basin geometry.
For submerged aerators, a simplified first order differen-
tial equation is used to define the reaeration process.
Practically all mass transfer models for submerged
aeration take the form:
(4)
dC/dt = KLa Df
where Df can represent any one of several driving force
relations. This driving force is defined as the difference
between the liquid film oxygen concentration in
equilibrium with the gas bubble and the surrounding
bulk liquid dissolved oxygen.
Basically, there are two categorical definitions for the
driving force. The first group of models utilize a variable
value for saturation during the reaeration process. That
is, the effective equilibrium saturation value is dependent
on a changing oxygen content in the gas stream and
dissolved oxygen in the liquid. As a result, during the
clean water reaeration process, the value of this satura-
tion changes with these variables. Examples include
"log-mean driving force", "log-mean saturation" and
others as described in Stanton and Bradley (6) and Lakin
et al. (3).
The second group of models are based on a representa-
tive constant saturation value for the entire reaeration
process. This category includes variations of pseudo-
surface aeration models.
For both cases, the driving force is dependent upon bulk
properties of the gas and liquid phases. By definition, the
KLB term must necessarily envelop all of the aerator
characteristics associated with elevating the oxygen level
in the basin. Hence, the characteristic bubble size,
relative velocity, retention time and convective flow
patterns are all lumped into this single mass transfer
parameter. Any other approach would require an
extremely complex process equation and sophisticated
test procedure which are well beyond the technology
currently available.
Mixing Measurement
The most direct indication of adequate mixing is provided
by the dissolved oxygen data. Unduly large gradients in
either the waste basin or in the reaeration clean water
test would indicate an improper balance between mixing
and aeration. To illustrate the relationship between
gradients, mixing and aeration, consider a two zone
model based on Figure 1. This model is one order of
magnitude more complex than the approach used to
obtain the differential equation for the liquid phase.
Equation 1. Oxygen transfer for this model occurs only in
the "aeration" zone. The reaeration of the basin occurs
as a result of convective mixing from the aeration zone to
the major portion of the basin.
To model this process, a differential equation for each
zone is required. These equations are interdependent as
a result of the bulk flow passing between the zones. For
the aeration zone we have:
(5) p, VA(dCA/dt) = p,QB(CB-CA)+p,VA KLa(C*-CB)
Similarity for the circulation zone:
(6) p, VB(dCB/dt) = PI QB (CA-CB)
where:
VA = liquid volume of aeration zone (L3)
VB = liquid volume of circulation zone (L3)
CA = DO concentration in liquid phase, aeration
zone (m/L3)
CB = DO concentration in liquid phase, circulation
zone (m/L3)
C* = DO saturation in liquid phase (m/L3)
QA = pumping rate in aeration zone (L3/t)
Q8 = pumping rate in circulation zone (L3/t)
Note that the inlet flow to the aerator originates in the
circulation zone. Thus, the mass transfer driving force is
the difference between saturation and the dissolved
oxygen level from that region. For the case of very large
values of the pumping rate QB, the concentration CA
approaches CQ. Equations 5 and 6 then reduce to
Equation 1.
65
-------
This formulation of the aeration process is the first step
toward a rigorous solution to reaeration performance.
The actual system will require much more compart-
mentalization to accurately model the process. For each
basin region, the bulk dissolved oxygen level and the
fluid interchange between adjacent regions must be
determined.
Recognizing that this cellular model is a simplified
attempt to interpret the balance between mixing and
aeration, the equations are solved for a specific example.
A hypothetical aeration basin which contains 3378 m3 of
water is divided as follows: 20% for the aeration zone
and 80% for the convective zone. Based on this volume
split, the bulk dissolved oxygen concentration is defined as:
(7)
C = 0.2 CA + 0.8
Equation 5 and 6 are solved for pumping rates, Qg,
varying from 189 to 1135 mVmin. The reaeration
variables one could expect to measure are determined as
a function of the pumping rate.
It is apparent that four different mass transfer coefficients
can be defined using this approach. The first, from
Equation 5, begins to resemble a "true mass transfer
coefficient" in that KLa is divorced from the convective
process and addresses only the oxygen transfer in the
aeration region. The other three values of KLa are
obtained from utilizing Equation 1 and the concentration-
time histories for each region individually as well as for
the defined bulk average.
To provide a comparison of these four mass transfer
coefficients, it is convenient to assume, for illustrative
Table 1. Mass Transfer Coefficient Comparison
purposes, that the rate of oxygen transfer provided in the
aeration zone (as defined by Equation 1) is identical for
each pumping rate chosen. Table 1 presents the results
of these computations.
Corresponding values of the oxygen transfer rates are
provided in Table 2. It is observed that at low pumping
rates, the oxygen transfer capabilities are substantially
overstated by using Equation 1 in the aeration zone.
However, as the pumping capacity is increased to
757 mVmin or greater, these differences are reduced
to within the range of experimental uncertainty. The
derivation of Equation 1 requires that the bulk dissolved
oxygen be used in application of the model. Therefore,
the best indicator of performance is given by column 4 of
Table 2. To experimentally obtain an adequate represen-
tation of the basin bulk dissolved oxygen level, it is
necessary to use a sufficient number of sample points.
In pursuing this example further, extension of the
analysis is helpful in examining the relationship between
mixing, mass transfer and dissolved oxygen gradients
during a reaeration test. A graphical presentation of the
dimensionless dissolved oxygen gradient vs time per
basin turnover is provided in Figure 10. The pumping
capacities selected embrace a significant range of turn-
over times for activated sludge aeration basins. Therefore,
the dimensionless gradient is shown to be an inverse
function of the aerator pumping capacity. The gradient is
also dependent on the mass conductance of the system
as indicated in Figure 10.
The significance of the term (CA-CB)/(C*-C) can best be
demonstrated by an example. Consider the case when this
Pumping
Rate (mVmin)
189
378
757
1135
KLa, Eqns. S&BJhr1)
Two Zone Model
2.03
2.49
2.70
1.74
Aeration
Zone
2.85
2.85
2.85
2.85
KLa, Equation 1 (hr~1)
Convective
Zone
2.16
2.62
2.77
2.79
Basin Bulk
Average
2.26
2.65
2.79
2.79
Table 2. Performance
Pumping
Rate (mVmin)
189
378
757
1135
Oxygen Transfer at Test Conditions (kg/hr)
Aeration Convective Basin Bulk
Zone Zone Average
78.9 59.9 62.6
78.9 73.0 73.9
78.9 76.7 77.6
78.9 77.6 78.0
Percent Deviation
of Aeration Zone
From Bulk Average
26
6.8
1.8
1.2
66
-------
Figure 9. Bubble Diameter vs Mass Transfer Coefficient — Oxygen/Water System
0.6
0.5
0.4
5 °3
0)
ffl 0.2
0.1
\
\ Oxygen/Water
A System
\
123456
KL * 10* [gm O2/(cm2/sec-gm 02/cm3]
Figure 10. Dimensionless Dissolved Oxygen Gradient vs Turnover Time
*
o
CO
V
(Q
o
o
o
,o
c
E
Q
0.6
0.5
0.4
0.3
0.2
0.1
HKLa (Areation Zone) = 2.85 hr'
Aeration Zone, 20% x Volume
Convective Zone, 80% x Volume
Slope Increasing
with Increasing KLa
Slope Decreasing
with Decreasing KLa
0 0.05 0.10 0.15 0.20 0.25 030
Basin Turnover Time — V/QB(hr)
0.40
67
-------
dimensionless gradient group is equal to 0.30. When the
bulk average concentration level reaches 20% of satura-
tion, the observed gradient would be 24% of the
saturation value. Similarly, as the 80% of saturation bulk
concentration level is attained, this gradient is reduced to
6% of the saturation value. This convergence of gradients
and manner of comparison is highly consistent with
actual reaeration data. Figure 11. Thus, larger gradients
are observed in early portions of a test when the oxygen
uptake rates are high (large driving force region) while
significantly smaller gradients appear as equilibrium is
approached. It should be emphasized however, that the
values presented are associated with the specific
numerical example cited.
Applying perfect mixing concepts to a reaeration test
generally requires that dissolved oxygen gradients be
nonexistent during all portions of the test. For any real
test situation, however, these gradients do and must exist
since the aeration transfer process is generally localized
and must be dispersed through the basin via the bulk
mixing mechanism. The two zone model reiterates that
dissolved oxygen gradients are inherent to the reaeration
process. For a particular system or application, it becomes
a judgement decision to define what constitutes a
reasonable or acceptable dissolved oxygen gradient for
operational purposes.
In order to observe whatever gradients exist, it is
necessary to have at least four to six sample pumps
distributed throughout the basin. At least two sample
points should be within the lower third of the basin and
two sample points in the upper third. While radial posi-
tion is somewhat arbitrary, sampling in locations in the
vicinity representing 30-40% and 70-80% of the tank
volume should provide the essential data. It is desirable,
but not mandatory, to have extra sample pumps located
at mid-depth and other locations to augment these
measurements.
Placement of the sample pumps as described in the
previous paragraph should exemplify the complete range
of DO gradients within the basin. Gradients in them-
selves are an inherent part of the reaeration process and
are not necessarily indicative of any special problems. In
order to eliminate these gradients, one would need to
exert a significantly greater amount of energy to the
mixing process. The most efficient aeration system is one
which balances the mixing energy required to maintain
acceptable DO gradients and solids suspension. The net
result would be oxygen transferred and distributed at a
minimum cost of energy.
Another indicator of mixing in a reaeration test is the
bottom velocity. Gravitational forces irrepressibly drag the
suspended solid particles to the bottom of the basin. A
strong flow along the bottom will reentrain these particles
and prevent any buildup of solids in the bottom of the
basin. Hence, both magnitude and direction are important.
It should be emphasized that bottom velocities alone are
not sufficient to describe the hydraulic process within the
tank. These measurements should be viewed in conjunc-
tion with the dissolved oxygen data.
Indirect indicators of mixing, i.e., turnover time, pumping
capacity and power per unit volume do not in themselves
assure adequate solids suspension. It is quite possible to
have a vigorous flow pattern in the upper half of the
basin with low flow rates along the bottom.
In subsurface systems, similar problems in mixing and
mass transfer can result from improper gas dispersion or
flooding conditions. (Gassed processes can also be
susceptible to another problem, viz. flotation.)
A brief review of impeller characteristics provides clues
to obtaining maximum pumping capacity at minimum
power. For geometrically similar impellers operating in
equivalent flow regimes (as indicated by Reynolds
number, Froude number, etc.), the primary pumping
capacity is given as:
(8)
Q, = NQnD3
The equation for power is expressed as:
(9)
P = K3Npn3D6
Maintaining a constant power. Equations 8 and 9 are
solved for pumping as a function of impeller diameter:
(10) Q, = K4NQ(P/Np)1/3D4/3
where:
QI = primary pumping capacity (L3/t)
NQ = dimensionless flow number
n = impeller speed (L/t)
D = impeller diameter (L)
P = power (fL/t)
Np = dimensionless power number
K3,K4 = constants
The result states that pumping is proportional to
diameter to the four-thirds power. This flow refers only
to the fluid being pumped directly through the impeller
field. The flow from the mixing impeller entrains more
fluid in the tank so that the total circulating flow can
be several times greater than the value of Q, obtained
above. These results illustrate a distinct advantage to
going to larger diameter impellers operating at lower
speeds. This conclusion is consistent with observations
from several mechanical aeration devices and are yet
another manifestation of the interrelationship of mixing
and aerator performance.
-------
Performance in Waste Stream
One of the key problems addressed at this Workshop is
the application of results from a clean water reaeration
test to performance in a waste stream. The two are
distinctly different processes each with its own unique
traits. The latter is a steady state bio-chemical process,
the bio-chemical kinetics of which are often difficult to
assess precisely, especially on a large-scale. In addition,
the characteristics of the biomass vary with time within a
specific system and can be grossly different from system
to system. The clean water reaeration test uses well
defined chemical reaction and physical absorption
processes in a non-steady state procedure.
As the goal is performance in the waste stream, the
clean water reaeration test is only meaningful when the
results are interpreted in this context. Design of an
aeration system generally includes the following goals:
1. Oxygen transfer efficiency
2. DO levels above prescribed minimum
3. Solids suspension.
In a reaeration test, the physical processes associated
with and correlated to these criteria are:
1. Mass transfer in the aeration zone
2. Convection of oxygen enriched water
3. Bottom velocities — magnitude and direction.
The correlation of these items, however, does not occur
on a one-to-one basis. The waste stream is a different
environment, and these differences require special
attention.
An observed phenomenon which highlights these differ-
ences is shown in Figure 12. In wastes with a high mixed
liquor volatile suspended solids, test results have sug-
gested a biologically enhanced performance of aeration
devices. That is, performance in the waste exceeded
predicted performance from clean water. Several possible
explanations of this phenomenon have been put forth, but
the precise nature of this effect is still elusive.
A waste characteristic that often accompanies this
enhancement is a decrease in the surface tension. For
surface aeration, this decrease in surface tension appears
to correlate with a secondary type of flow pattern as pre-
viously described in Figure 2. Therefore, it is possible for
bottom velocity measurements in the clean water test to
be acceptable and perhaps not achieve proper mixing
requirements in the actual waste. Experiences of this
nature tend to encourage some minimum testing in the
waste stream. While a clean water reaeration test is
preferred for demonstrating oxygen transfer capabilities,
a simple test in the waste stream would be a much better
indication of satisfactory solids suspension. Probing the
basin bottom after several months of operation and
obtaining grab samples at several different elevations
within the basin is currently the only accurate means of
assuring adequate solids suspension. This could also be
artificially induced, with difficulty, in a clean water test
through surface tension adjustment and inert solids
addition. Rather than specifying unduly high clean water
bottom velocities, it would be more relevant and mean-
ingful to measure solids suspension directly in waste
operation.
Full- or Small-Scale Tests
Extrapolating the velocity measurements made for one
set of geometric conditions to a different set can only be
done in an approximate fashion. The flow relations for
aeration systems are quite complex and to date have not
been analytically solved. (Some numerical evaluations
have been performed on simplified systems.) Since both
oxygen transfer and solids suspensions are dependent
upon the flow patterns, the reaeration test basin should
be geometrically similar to the waste basin.
The most critical geometric conditions for scaling are side
water depth, sparge location, surface area and degree of
baffling. A fully baffled tank is defined as one in which
the baffle surface area is sufficient to remove the basin
swirl component. This swirl component is induced by
mechanical aeration devices which have a rotational
motion. Similar types of motion can also be induced by
submerged jets and other aeration devices with a prefer-
red direction for induced motion. When possible, tank
dimensions for the reaeration tests should be within
±10% of the desired waste basin configuration. If this
condition cannot be satisfied, tests in basins beyond
these limits require a substantial understanding of the
controlling parameters. While the uncertainty in such
testing increases, it is possible to provide a meaningful
indication of aerator performance.
Model studies can be used only with the greatest of care.
The process to model contains many diverse elements
and in order to properly scale the complete process, the
contribution of each element must be examined. The
elements include:
1. Local Mass Transfer — surface splash and spray, or
bubble size, and hydrostatic head
2. Convection of Enriched Fluid — hydraulic flow
patterns, bottom velocities.
Even if the scaling parameters were known precisely, it
is difficult to satisfy all necessary requirements in a small-
scale model. Reduced scale test results probably should
be limited to visualization studies of hydraulic flow
patterns and preliminary investigations. This type of test-
ing could be useful for determining baffle configurations,
bottom velocities and approaches to rectify specific flow
patterns.
69
-------
Figure 11. Typical Reaeration Test Data
•o
0>
o
'•£
Q.
\
u
a>
CC
t_
0)
c
ro
c
cu
O)
X
O
o
o)
u>
o
c
o
CJ
c
0>
X
o
o>
"o
en
C/l
10
9
8
7
6
5
4
3
2
1
0
Dissolved Oxygen
Gradients
0 2 4 6 8 10
15
20
25
Time (min)
Figure 12. Biological Enhancement of Oxygen Transfer
100 200
Biological Uptake Rate (mg/l -hr)
300
70
-------
Conclusions
1. As defined, the measured mass transfer coefficient is
innately dependent on mixing.
2. A minimum of four to six sample points are required
for clean water test measurements.
3. DO gradients are intrinsic to the reaeration test and
do not necessarily imply problems in mixing or mass
transfer.
4. Individual sample point uptake rates should agree
closely with that based on average basin DO. Pro-
nounced rate differences might indicate adverse
gradients in the waste or improper blending.
5. Bottom velocities (magnitude and direction) are a good
indication of solids suspension capabilities.
6. Turnover time, pumping capacity and power per
volume are poor indicators of mixing and solids
suspension, per se.
7. The best test of solids suspension is solids measure-
ments in the actual waste stream.
8. Scaling of aeration processes is complicated. Tests
should be run in a similar sized basin.
References
1. Coppoch, P.O., and G.T. Meiklejohn. "Trans. Institute
of Chemical Engineers". London, 28, p. 52, 1950.
2. Datta, R.L., D.H. Napier, and D.M. Newitt. "Trans.
Institute of Chemical Engineers". London, 28, p. 14,
1950.
3. Lakin, M.B., R.N. Salzman, J.Y. Oldshue, and H. Gray.
Paper presented at International Symposium in
Chemical Engineering, Mons, Belgium, 1978.
4. O'Brien, M.P., and J.E. Goline. "Industrial Engineers
Chemistry". 27, p. 1436, 1935.
5. Soo, S.L. "Fluid Dynamics of Multiphase Systems".
Blaisdell Co., Waltham, Massachusetts, 1967.
6. Stanton, J.L, and P.P. Bradley. "Proceedings 30th
Industrial Waste Conference". Purdue University, 1975.
71
-------
Sampling Considerations
Gerry L Shell
Gerry Shell Environmental Engineers, Inc.
Brentwood TN 37027
Introduction
In the evaluation of aeration equipment, obtaining a
representative samle is basic. Each phase of the evalua-
tion, test procedure, data analysis, and data presentation,
is important, but using a proper sampling method is
essential. There are two methods commonly used for the
evaluation of aeration equipment, namely the Clean
Water Non-Steady State Method and the In-Process
Method. The Clean Water Non-Steady State Method uses
two test procedures: wet chemical analysis (Winkler) and
the dissolved oxygen probe. The sampling methods
suggested in this presentation will include both the Clean
Water Non-Steady State Method and the In-Process
Method.
Clean Water Non-Steady State Method
General Considerations
The basic mass transfer equation:
dC/dt = KLa(C*-C)
requires that mixing be complete. (Complete mixing is
defined here as the continuous movement of all liquid
in the basin.) The first consideration, therefore, in any
aeration device evaluation is to assure that mixing is
complete.
One test for complete mixing that has been suggested is
that the dissolved oxygen concentration, C, be uniform
(not vary more than ±0.25 mg/l) throughout the basin.
This is not a valid indicator of uniform or complete
mixing. Since all aeration devices are point sources of
oxygen transfer, there must be an oxygen concentration
gradient across the aeration device. If not, the device is
simply not transferring oxygen. For example, a low speed
surface aerator will have an oxygen concentration
gradient of 4 to 5 mg/l between the inlet liquid and the
aerator issue. Therefore, there must be an oxygen
gradient in the aeration basin. Using a uniform oxygen
concentration as an indication of complete or uniform
mixing is not technically sound.
Using oxygen saturation as an indication of completely
mixed conditions is also not technically sound. Due to the
varying pressure conditions in a basin, oxygen saturation
can and will vary somewhat with liquid depth. This is
true even for surface aerators where mixing is intense,
causing secondary aeration (bubble entrainment) to occur
at the walls of the basin. The representative value of
oxygen saturation for any aeration system is difficult, if
not impossible to establish. Although this problem has
been discussed at length, a satisfactory conclusion has
not been reached.
12
The author prefers to use the final factor in the mass
transfer equation to establish if a completely mixed con-
dition is achieved. That factor is the overall mass transfer
coefficient, KLa. If the aeration system is truly mixed, the
value of KLa will be constant within the range of experi-
mental error. After reviewing the results of several
hundred clean water non-steady state tests, it was
established that the value of KLa should not vary more
than ±7.5% due to testing variations. If the value of KLa
varies more than ±7.5%, the test procedure should be
checked thoroughly. If no apparent problems are found,
then the test results should be rejected on the basis of
incomplete mixing.
To establish the variation in KLa, at least four sampling
points should be used. These sampling points should be
placed randomly in the basin avoiding the areas near the
aeratior issue and near the walls. The data from each
sample point should be handled separately (do not
average the dissolved oxygen concentrations) to establish
the value of KLa. If a single point varies greatly from the
remaining points, check the location to determine if it is
in a zone of intense aeration or low turbulence (behind a
column or baffle).
Sample Location
Concerning the location of sampling points, only a single
point is needed once completely mixed conditions are
established. If only one sample point is used during
general testing, it should be located at the inlet to the
aeration device. For example, for diffuser aeration devices,
the sample point should be located at the same elevation'
as the diffusers and one foot from the nearest air release
point. For jet aeration devices, it should be located at the
inlet to the pump.
If it is desired to have a number of sample points, then
the following criteria are suggested:
1. Do not place sample points in the plume or issue of
the aeration device.
2. Do not place sample points any closer than three feet
from the walls, six feet from the corners, or two feet
from the basin floor and liquid surface.
3. Place one or more sampling points near the liquid
inlet to the aerator.
4. Place sampling points at the one-third and two-thirds
points in the basin both in relation to the diameter
and depth. The diameter points should be on opposite
sides of the basin.
72
-------
When using multiple sampling points, in no case should
the oxygen concentration data be averaged for each time
increment. It is mathematically improper to average
dissolved oxygen values at each time increment to obtain
an average K\_a value. Therefore, the data from each
sample point must be handled independently to establish
separate values of KLa. Then individual KLa values may
be averaged.
Sample Volume and Rate
For the wet chemical analysis (Winkler), the sample
volume (about 300 ml) used is based on the standard
BOD5 bottle. The sampling procedure should be to place
the sampling points (pumped or siphoned) in their proper
locations in the aeration basin. Plastic tubing (about one-
quarter inch diameter) should be connected from the
sample point to a central collection point. At timed
intervals, the plastic tubes are placed into the BOD5
bottles such that the ends of the tubes are within one-
half inch of the bottle bottom.
The rate of sample flow should be such that the bottle is
filled within 5 sec. This means that the bottle contents
will be replaced a minimum of three times in 15 sec.
The length of the sample lines (for more than one sample
point) is not critical if the retention time in the tube is
less than 7.5 sec. If the retention time is greater than
7.5 sec, then all sample lines should be of the same
length. To avoid possible data bias, the maximum sample
line retention time should be 7.5 sec or less, if possible.
Longer retention times may result in liquid temperature
changes and possible additional aeration where air
bubbles may be drawn into the sample line.
Sample Point Configuration
To avoid the inclusion of air bubbles into the sample line,
the sample point should be located away from areas of
intense turbulence. The inlet to the sample point should
be in a vertical plane with the opening at the top. The
downward inlet velocity should be such that all air
bubbles larger than 500 microns are excluded.
Sample Timing
Timed samples are collected from each sample point. At
least ten samples should be collected at each point for
dissolved oxygen concentration analysis. Seven or eight
of these samples should have an oxygen concentration
between 10 and 90% of saturation. In order to achieve
the appropriate time interval, the following expression
can be used as a first estimate.
t = 100M/OTR
where:
t = sample time interval, min
M = test basin water weight, Ib * 10"6
OTR = estimated oxygen transfer rate, Ib O2/hr
The following example illustrates use of the above
expression.
Aerator - 100-hp low speed surface
aerator
Basin Dimensions - 60' x 60' x 15' SWD
Estimated Efficiency - 3.0 Ib 02/hp-hr
OTR = 100 (3.0) = 300 Ib 02/hr
M = 60 (60) (15) (62.4) = 3.37 x 10 6 Ib
t = 100(3.37)7300 = 1.1, use 1.0 min sample times
Data Variation
A major problem sometimes associated with sampling
data is noise or scatter. It has been the author's experi-
ence that data scatter or noise increases with increasing
energy input. The first reaction noted by most observers
to data scatter for clean water non-steady state testing is
that mixing is incomplete or insufficient. However, as
noted above, data scatter becomes most severe at the
highest input energy level.
A possible explanation of the above noted observation is
the following:
1. At low input energy levels, the flow pattern in the
aeration basin is uniform, nearly laminar in nature.
2. For uniform or laminar flow conditions, the data
collected are uniform, low in scatter and high in
correlation.
3. As the energy input level reaches some critical level,
uniform or laminar flow conditions are lost and
turbulent conditions are established. With increasing
turbulence, the data collected at any point in the basin
become more scattered and noisy.
An example of the above conditions was observed
recently in a field clean water non-steady state test on
low speed aerators. In this case, four 75-hp low speed
aerators were placed in a 100-ft square basin. Six
sampling points were located randomly in the basin. The
individual log deficit plots for each sampling point cor-
related poorly at full power (75-hp for each unit).
Comparison of the six resulting K|_a values showed a
relatively wide variation. However, the mixing was very
intense as the entire basin was white with air bubbles
that had been entrained throughout the basin volume. It
was suggested that the reason for the poor correlation
and scatter was due to "poor" mixing conditions.
The above testing was completed on two speed units
having a 10-inch liquid submergence variation. The entire
range of power levels, from 75-hp down to about 25-hp,
was evaluated. It was interesting to note that as the
input power was reduced, the data tended to correlate
better with less scatter.
In the above example, data scatter or noise was associ-
ated with mixing intensity. This same observation has
been noted for bottom velocity measurements. In the
case of bottom velocity, it was observed that the bottom
73
-------
velocity increased in proportion to energy input up to
some critical value. At that value and beyond, the bottom
velocity began to decrease. The critical input energy level
appeared to be the point between uniform and turbulent
mixing conditions.
Dissolved Oxygen Probes
Dissolved oxygen probes have been used extensively for
the evaluation of aeration devices. The primary motivation
for the use of DO probes is the ease of data collection.
Certainly, when the probes are applied and function
properly, their use is acceptable. However, assurance that
the probes are used properly and are functioning properly
is difficult to achieve and can be time consuming.
The first problem with DO probes is their electronic
nature. The sensor and meter are subject to environ-
mental conditions such as temperature, pressure, and
moisture. All of these factors affect the operation of
the unit.
Secondly, being electronic, most technical people are not
competent to operate and maintain the unit so that
proper results are assured.
Thirdly, the unit is fragile, requiring care to assure that
the membrane is not punctured or broken during testing.
In many tests conducted by the author, the membrane
was found faulty during the testing.
Finally, probes do not indicate a finite value of oxygen
'concentration. Since time is required for oxygen to diffuse
through the membrane, the indicated oxygen concentra-
tion is an average value over a period of time. This is one
reason why DO probe data are more uniform and cor-
relate better than wet chemical analysis (Winkler) data.
In-Process Method
Sampling is required to establish several parameters
utilized in the evaluation of aeration equipment using the
in-process test procedure.
Alpha
Alpha is the ratio of the wastewater KLa to the clean
water KLa. It is determined in the laboratory with an
aeration device which is similar to that used in the full-
scale operation. Alpha is the correction for surfactants
that may be present in the wastewater. It is the author's
opinion that alpha values should be based on a sample of
the mixed liquor with the solids present and the micro-
organisms killed using a sulfamic acid-copper sulfate
solution. Other samples such as the influent, effluent, or
settled mixed liquor do not represent the actual operating
conditions.
Beta
Beta is the correction for salinity. A sample of the mixed
liquor should be filtered through a membrane filter. The
filtrate should be then analyzed for total dissolved solids.
The saturation value for the wastewater is then corrected
for salinity, represented by the total dissolved solids.
Residual Dissolved Oxygen, C
It is commonly held that an oxygen profile should be
completed on the aeration basin contents to establish the
value of C, the basin dissolved oxygen concentration. It
should be recalled that the purpose of the test is to
evaluate the aeration equipment and not the biological
process. What does the aerator "see"? The aerator does
not "see" the average basin oxygen concentration. The
aerator "sees" only the liquid entering the device. There-
fore, the sampling point for determining C should be at
the inlet to the aeration device and not an average basin
dissolved oxygen based on a DO profile.
Oxygen Uptake Rates, R
The oxygen uptake rate is a direct measurement of the
microbial utilization of oxygen. If it has been established
that the aeration basin is completely mixed, then three or
more samples can be taken at any point in the aeration
basin. However, since a relatively high concentration of
dissolved oxygen is required in the sample for the
determination of the uptake rate, it is suggested that
three or more samples of the aerator issue be taken for
analysis.
If complete mixing of the basin is questioned, then
samples should be taken from at least ten random loca-
tions for oxygen uptake analysis. Completely mixed
conditions are indicated if the variation of oxygen uptake
is ±10% or less.
Liquid Temperature
For completely mixed conditions, one sample point is all
that is required. That sample point should be at the inlet
to the aeration device.
In-Process Formulas
Overall Mass Transfer Coefficients:
Oxygen Transfer Rate:
SOTR = (KLa)20 (C^)M
OTE = [SOTR/Qa( 1.035)]
where:
R = oxygen uptake rate, mg/l-hr
a = (KLa)2o wastewater/(KLa)20 clean water
0 = oxygen saturation correction value for salinity
Cy = clean water oxygen saturation at the inlet to the
aeration device, mg/l
6 = temperature correction factor
74
-------
(KLa>2o = overall mass transfer coefficient for standard
conditions, hr"1
T = basin liquid temperature, °C
M = basin water weight, Ib * 10"6
SOTR = standardized oxygen transfer rate in clean
water, Ib C>2/hr
OTE = oxygen transfer efficiency, dimensionless
Qa = air flow rate, scfm
Conclusion
Sampling is a key element to be considered when
evaluating aeration equipment. The following questions
should be answered before an acceptable sampling
arrangement is approved.
What is the:
Sampling location?
Sampling volume?
Sampling rate?
Sample configuration?
Sample source?
Sample time?
It should be remembered that the basis of all aeration
testing is the mass transfer equation. The basic restraint
in the use of the mass transfer equation is that the
aeration basin must be completely mixed.
75
-------
Analytical Measurement and
Saturation Values for Dissolved
Oxygen in Water
Vernon T. Stack, Jr.
Betz • Converse • Murdoch, Inc.
Plymouth Meeting PA 19462
Introduction
Solubility of oxygen in water is a relatively straight-
forward concept which is accepted primarily on the basis
of facts which involve the composition of the gas phase,
the composition of the water phase and Henry's Law.
Technical and engineering practices utilize values for
oxygen saturation in water as a basis for driving forces in
the transfer of oxygen from gaseous to liquid phases.
Oxygen transfer calculations must begin with a known
entity. Oxygen saturation values are available from litera-
ture and references sources and are convenient for use.
The available information on oxygen saturation of water
from air was obtained under carefully controlled condi-
tions where the air temperature and the water tempera-
ture were the same and the air was saturated with water
vapor (in most cases). When the "saturation value" is
applied across a relatively broad set of conditions which
involve differences in air and water temperature and less
than saturated conditions for water vapor in the air
phase, the applicability of the information may be
questioned. The following discussion is an examination
of some of the factors involved in oxygen saturation and
some of the questions both answered and unanswered.
Oxygen Saturation of Water From Air
Definition
The examination of facts and questions should begin with
a definition of oxygen saturation of water. Saturation is
the result, measured as molecular oxygen concentration
in the water phase, of an equilibrium state in which
molecular oxygen in the air phase diffuses both into and
out of the water phase at the same rate. The most
important factor in this definition of saturation is that it
assumes an equilibrium state, and therefore, the results
of oxygen saturation of water would be shifted by any
factor which shifted the equilibrium.
Examination of Information A vail able
Whipple & Whipple (8) published in 1911 information on
solubility of oxygen in water and sea water based on
observations by Fox (3) (4) in 1907 and 1909. The
published information by Whipple & Whipple is the basis
for information which is presented in the 14th Edition of
"Standard Methods for the Examination of Water and
Wastewater" (7). In 1966, Carpenter (1) published the
results of studies of oxygen solubility in water and those
13
results differed somewhat from the information published
by Whipple & Whipple. The information developed by
Carpenter (Table 1) shows somewhat lower values of
oxygen saturation than that published by Whipple &
Whipple. The difference is greatest at high temperatures
where it reaches approximately 0.2 mg/l.
Table 1. Oxygen Solubility (mg/l) in Water,
Comparison of Available Information
Temperature Whipple & Whipple (8) Carpenter (1)
(°C) 1911 1966
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
14.6
13.8
13.1
12.5
11.9
11.3
10.8
10.4
10.0
9.5
9.2
8.8
8.5
8.2
7.9
7.6
7.4
7.2
7.0
6.8
6.6
14.6
13.8
13.1
12.4
11.8
11.3
10.8
10.3
9.9
9.5
9.1
8.7
8.4
8.1
7.8
7.5
7.3
7.1
6.8
6.6
6.4
Carpenter concluded that the work by Fox potentially
represented some degree of supersaturation at lower
temperatures, but the differences at higher temperatures
were possibly due to working from cooler to warmer
conditions with inadequate time for equilibrium to be
established. Work on the solubility of nitrogen in sea
water by Rakestraw & Emmel in 1938 indicated that the
observations by Fox appeared to be too high by about 1 %
The magnitude of differences between the observations
76
-------
of Fox and the observations of Carpenter are not alarm-
ingly large, but the differences do warn that the
equilibrium state for oxygen solubility in water is not
easily attained and that supersaturation is certainly
possible. As a matter of interest, oxygen solubility
information based on the observations of Carpenter are
presented in Tables 2 and 3.
Factors Involved in Oxygen Saturation
Beginning with Henry's Law, the concentration of oxygen
molecules in the gas phase is a key factor. Since air is a
compressible fluid, the number of oxygen molecules in a
given volume will be directly proportional to the absolute
pressure and indirectly proportional to the absolute
temperature. If we apply this concept to the air-water
interface (Figure 1), the concentration of oxygen per unit
area at the interface will be directly proportional to
pressure and indirectly proportional to absolute
temperature.
If we consider the film of air immediately adjacent to the
water interface, the concentration of oxygen present is
further adjusted by the concentration of water vapor
molecules present. Since the total pressure in the gas
phase is made up of the partial pressures of the gases
present, the water molecules are present in proportion to
the partial pressure corresponding to the vapor pressure
of the water at the water temperature.
Once the oxygen molecule has migrated into the water
phase, the assumption must be made that the oxygen
molecules are diffused and/or mixed throughout the
water phase on a relatively homogenous basis. The
tendency of the oxygen molecule to remain in the water
phase rather than escape immediately back into the gas
phase is dependent upon molecular attractions between
water molecules and oxygen molecules. The equilibrium
state of the molecular attraction depends upon the
kinetic energy of the water and oxygen molecules and
upon impurities in the water which also associate with
the water molecules and influence the overall status of
molecular attraction.
Therefore, three significant factors in oxygen saturation
of water are:
1. The concentration of oxygen in the gas phase (particu-
larly at the air-water interface)
2. Forces of molecular attraction between water and
oxygen molecules in the water phase
3. Kinetic energy of water and oxygen molecules in the
water phase.
Concentration of Oxygen Molecules in Air at the
Air-Water Interlace
The factor of pressure (and oxygen concentrations) of the
gas phase is easily understood where there is a simple
interface between air and water and when the gas
phase is under atmospheric pressure or some other
applied pressure. The situtation is more complex when
there is a mechanical generation of interfaces between
air and water such as results from bubble diffusion or
turbines. Since the air in bubble form is then mixed into
the water, pressure resulting form weight of water is
added to atmospheric pressure.
For any particular device which is employed for the
generation of air-water interfaces, knowledge of the air-
water interface in terms of area of interface and distribu-
tion of the interfaces in the water system is not readily
available or modelable; therefore, pressure and the
corresponding concentration of oxygen molecules at the
generated interfaces are unknown. It is a practical
recommendation that each oxygen transfer system must
be considered unique and that the apparent oxygen
saturation value for the system must be determined
experimentally. The system which involves air and water
interfaces in the water depth of the system will reach an
"equilibrium condition" which is supersaturated with
respect to atmospheric pressure. At "equilibrium satura-
tion" the system may transfer oxygen into water in one
zone and transfer oxygen out of water in another zone.
Superimposed over the transfers may be a significant
tendency toward supersaturation. It seems practical that
the observed equilibrium state for that system is real,
and the observed saturation value should be utilized in
defining the driving force for oxygen transfer in the
particular system.
The consideration of water vapor at the air-water inter-
face is straightforward when the air and water are at the
same temperature and the air is saturated with water
vapor. If the conditions are shifted so that the air is not
at the same temperature of the water and is not saturat-
ed, is the concentration of molecules at the air interface
altered? Is the equilibrium which represents oxygen
solubility into water shifted? If it is assumed that the
interface is at the temperature of the water, then water
temperature would control the vapor pressure in the gas
phase at the interface and the concentration of oxygen
molecules at the interface (gas film) would be controlled
by the temperature of the water. Air temperature and
relative humidities would not be a factor.
When the air is dry, evaporation of water occurs.
Evaporation requires heat energy which is drawn from
the water phase at the interface. Therefore, it is logical to
assume that the interface is being cooled and is not at
the same temperature as the body of the water. If the
water at the interface is cooled, the vapor pressure of
water at the interface would be decreased and the
concentration of water molecules in the gas layer at the
interface could be less than that predicted from the
temperature of the body of water. Additionally, if the air
film at the interface is cooled by evaporation, the density
of air at the interface could be increased and would not
be represented by Henry's Law for the gas mass. Both
77
-------
Table 2. Solubility of Oxygen (mg/l) at Various Temperatures and Elevations
(Based on Sea Level Barometric Pressure of 760 mm Hg)
Temperature
(°C)
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
Tables. Solubility
(Based on
Temperature
(°C)
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
Elevation (Feet Above Sea
0
14.6
13.8
13.1
12.4
11.8
11.3
10.8
10.3
9.9
9.5
9.1
8.7
8.4
8.1
7.8
7.5
7.3
7.1
6.8
6.6
6.4
of Oxygen
1000
14.1
13.3
12.7
12.0
11.4
10.9
10.4
9.9
9.6
9.2
8.8
8.4
8.1
7.8
7.5
7.2
7.1
6.9
6.6
6.4
6.2
(mg/l) at
Level)
2000 3000 4000
Various
13.6
12.9
12.2
11.6
11.0
10.5
10.1
9.6
9.2
8.9
8.5
8.1
7.7
7.6
7.3
7.0
6.8
6.6
6.3
6.2
6.0
Temperatures
Sea Level Barometric Pressure of
760 mm
13.2
12.4
11.9
11.2
10.6
10.2
9.7
9.3
8.9
8.6
8.2
7.8
7.6
7.3
7.0
6.8
6.6
6.4
6.1
5.9
5.8
and
Hg)
12.7
12.0
11.4
10.8
10.3
9.8
9.4
9.0
8.6
8.3
7.9
7.6
7.3
7.0
6.8
6.5
6.4
6.2
5.9
5.7
5.6
5000
12.3
11.6
11.0
10.4
9.9
9.5
9.1
8.7
8.3
8.0
7.7
7.3
7.1
6.8
6.6
6.3
6.1
6.0
5.7
5.6
5.4
6000
11.8
11.2
10.6
10.1
9.6
9.2
8.8
8.3
8.0
7.7
7.4
7.1
6.8
6.6
6.3
6.1
5.9
5.8
5.5
5.4
5.2
Chlorinity
Chlorinity (%)
0
14.6
13.8
13.1
12.4
11.8
11.3
10.8
10.3
9.9
9.5
9.1
8.7
8.4
8.1
7.8
7.5
7.3
7.1
6.8
6.6
6.4
4.0
13.9
13.2
12.5
11.8
11.3
10.8
10.3
9.9
9.4
9.1
8.7
8.3
8.1
7.8
7.5
7.2
7.0
6.8
6.6
6.4
6.2
8.0
13.2
12.5
11.9
11.3
10.8
10.3
9.8
9.4
9.0
8.7
8.3
8.0
77
75
72
70
67
65
63
6.1
5.9
12.0
12.5
11.9
11.3
10.8
10.3
9.8
9.4
9.0
8.6
8.3
8.0
7.7
7.4
7.2
6.9
6.7
6.5
6.3
6.1
5.9
5.7
16.0
11.9
11.4
10.8
10.3
9.8
9.4
9.0
8.6
8.3
8.0
7.7
7.4
7.1
6.9
6.6
6.4
6.2
6.0
5.8
5.6
5.4
20.0
11.3
10.8
10.3
9.8
9.4
9.0
8.6
8.3
8.0
7.6
7.4
7.1
6.9
6.6
6.4
6.2
6.0
5.8
5.6
5.4
5.2
78
-------
Figure 1. Air/Water Interface
Air (O2, N2, H2O, Etc.)
CD
-------
factors could result in an increased concentration of
oxygen molecules at the air-water interface, which would
result in a higher saturation value for the body of water
than predicted from body of water temperature. The
opposite could be projected for conditions of
condensation.
In a experimental study completed in 1971 (unpublished),
the relationship of water temperature, air temperature
and relative humidity were examined in the range of 5 to
35°C. Water temperatures were selected at 5, 15, 25 and
35°C, and for each water temperature, the air tempera-
ture was established at a wet bulb temperature of 5, 15,
25 and 35°C. For each wet bulb temperature, the dry
bulb temperature was established as equal to the wet
bulb temperatue or 5, 10 and 15°C greater than the wet
bulb temperature. Throughout the experimental program,
air velocity at and parallel to the interface was maintain-
ed at approximately 5 mph. Through the entire range of
temperature and relative humidity conditions, there was
no indication that the saturation value of the water phase
was significantly altered (more than 0.1 mg/l) by the air
temperature and air humidity.
If the indicated experimental observations were con-
ducted at greater air velocities ( >20 mph), the degree of
turbulence established in the air flow would be increased
significantly. If the turbulence increased the rate of
evaporation, the increased rate of removal of water vapor
and heat from the interface might significantly alter the
equilibrium of oxygen saturation. An answer to this
question is yet to be obtained in a controlled experimental
study.
Molecular Attraction Between Oxygen and
Water Molecules
The concept of molecular attraction between the
molecules in the water system and the oxygen molecule
is important since changes in the degree of attraction
would influence the equilibrium state of oxygen satura-
tion in the water system. Interestingly, the saturation of
oxygen in water correlates rather closely with water
viscosity over the temperature range of 0 to 100°C
(Figure 2). Thus, the molecular attraction between water
molecules which influences viscosity and the molecular
attraction between water molecules and oxygen molecules
are interrelated. This correlation does not establish that
an alteration of water viscosity will have an impact on
oxygen solubility, but such could be the case if decreased
or increased viscosity were the result of decreased or
increased molecular attraction.
When inorganic salts are added to water, the concentra-
tion of oxygen at saturation is decreased. Presumably
this effect occurs because the dissolved inorganic solids
are associated with the water molecules and the result is
a decreased molecular attraction between water mole-
cules and oxygen molecules. The magnitude of the
decreased oxygen concentration at saturation is shown in
Figure 3 for selected salts. The correction factor in the
plot of Figure 3 is a ratio of dissolved oxygen concentra-
tions at saturation in the salt solutions divided by dissolved
oxygen concentrations at saturation for distilled water.
Thus, the ratio is the beta factor. The data in Figure 3
illustrate that very significant concentrations of dissolved
salts are required before a pronounced impact on the
beta factor is observed. The range for the selected salts
is 0.1 to 0.3% (1,000 to 3,000 mg/l) before a 1 % reduc-
tion in the beta factor is observed.
More extensive data concerning oxygen solubility in sea
water are presented in Table 3 and are correlated with
the chlorinity of the sea water. As a matter of interest,
the solubility of oxygen in synthetic sea water, such as
that defined in ASTM Standard Specification D-1142-52
is greater than the solubility of oxygen in natural sea
water. The difference is approximately 1 mg/l at 20°C.
Presumably the difference is due to a lower level of
molecular attraction in the natural sea water.
Measurement of Oxygen in Water
Winkler Titration
The accuracy and applicability of the Winkler titration for
chemical measurement of oxygen in water is well docu-
mented. The precautions necessary in the presence of
oxidizing or reducing materials is understood, and
methodology exists which can be used to accommodate
these interferences.
The application of Winkler titration to the measurement
of oxygen in clean water during oxygen transfer studies
is definitely a practical application. An analytical problem
which has been identified is related to the use of a cobalt
catalyst in the sulfite deoxygenation procedure and the
resulting precipitation of a cobalt compound (5). Upon
acidification in the Winkler titration procedure, the cobalt
precipitate can cause the liberation of an additional
amount of iodine. Thus, the interference results in an
oxygen concentration value higher than the actual.
Work by Mixing Equipment Company has shown that the
magnitude of the cobalt compound interference may be
related to hardness of the water and is eliminated if the
pH of the water is made low enough to keep the cobalt
compound in solution.
The basic precaution is to either utilize a low concentra-
tion of cobalt so that interference in the Winkler titration
is insignificant, or if higher concentrations of cobalt are
to be utilized, determine oxygen by other procedures such
as the use of dissolved oxygen probes.
The Winkler titration procedure has adequate opportuni-
ties for error. Potential errors to be considered are:
1. Air oxidation of iodide
2. Volatilization of iodine
80
-------
Figure 2. Oxygen Solubility in Water vs Viscosity (0° to 40°C)
O>
J
5
"o
0)
O
!
15 r
14
13
12
11
10
9
8
0.6 0.8 1.0 1.2 1.4
Viscosity (centipoise)
1.6
1.8
o
t>
CO
LL.
.1
S
0>
is
o
Figure 3. Effect of Dissolved Salts on Oxygen Saturated Water (22°C)
1.00
0.98
0.96
0.94
0.92
0.90
NaCI
Na2S04
0.2 0.4 0.6
Percent by Weight
0.8
1.0
81
-------
3. Oxygen contributed by the reagent solutions
4. lodate contamination of the iodide solutions
5. Consumption or production of iodine by reagent
contaminants
6. Difference between titration end point and the
equivalency point.
Carpenter (2) found in work published in 1965 that the
more significant errors were potentially from reagent
blank errors and volatilization of iodine during titration.
To reduce volatilization of iodine, he suggested an
alkaline iodide reagent containing 600 g/l of Nal and 320
g/l of NaOH. The higher concentration of Nal would
minimize the volatility of iodine through the formation of
triiodide.
In earlier work published in 1965, Montgomery et al (6)
concluded that iodine volatility was a problem and
recommended the use of Pomeroy-Kirschman alkaline-
iodide reagent. This reagent contains 400 g/l of NaOH
and 900 g/l of Nal.
Standard Methods (7) does not identify iodine volatility to
be a problem in the iodometric methods for oxygen
determination. It does offer a choice of alkaline-iodide
reagent, but does not suggest why one or the other might
be selected. Alkaline-iodide reagent No. 1 contains 135
g/l of Nal. Alkaline-iodide reagent No 2 is the Pomeroy-
Kirschman reagent (900 g/l of Nal).
The use of Standard Methods reagent No. 1 could lead to
low results because of volatilization of iodine. The
magnitude of the error would depend upon the analyst.
Adoption of transfer procedures which reduce agitation
and air contact with the sample followed by quick titra-
tion to near the endpoint would minimize iodine loss. If a
less experienced analyst poured the sample from a BOD
bottle into a graduated cylinder, poured the sample from
the cylinder into an Erlenmeyer flask and proceeded with
a slow titration, a very significant loss of iodine would
occur. For warm samples (temp. > 20°C), the error could
be significant, and at 30 to 35°C the results for a
saturated sample could be low by as much as 0.5 mg/l.
The problem of iodine volatility can be suitably reduced
by the following actions:
1. If Standard Method's alkaline-iodide reagent No. 1 is
to be utilized, add the manganeous sulfate and
alkaline-iodide reagents, carry the procedure through
the flocculation stage and cool the sample to a
temperature in the 10-15°C range before adding the
acid reagent. Titrate the sample with due considera-
tion for the volatility of iodine.
2. Use Standard Method's alkaline-iodide reagent No. 2
so that iodine is present as triiodide and iodine
volatility is not a significant problem.
Dissolved Oxygen Probes
Membraned Probes
Dissolved oxygen probes are convenient for relatively
precise determinations of oxygen, and as stated there is
no interference from the cobalt compound which inter-
feres with Winkler titration. To insure that a membraned
probe is used correctly, it is well to understand the
principle which is involved in the detection of oxygen by
the probe.
Basically the membraned probe responds to the arrival of
molecular oxygen at the cathode where it is converted to
hydroxyl ions. Hydroxyl ions are correspondingly con-
sumed at the anode in the conversion of elemental metal
to the metallic oxide. The overall reaction establishes an
electron flow which can be measured as an electrical
current.
Precautions in using a membrane probe are not related
to the reaction within the probe, but are related to the
diffusion of oxygen through the membrane before it
reaches the cathode. When placed in a water sample,
the membrane is exposed to a concentration of oxygen
molecules at the water-membrane interface. When the
water is saturated with respect to air, the number of
oxygen molecules per unit area at the water-membrane
interface is the same as the number of oxygen molecules
per unit area at the air-water interface (Figure 4). Thus,
when the probe is in the saturated water solution, the
opportunity for oxygen molecules to diffuse through the
membrane is the same as exists when the probe is
exposed directly to air.
The basic precaution which must be understood in the
use of a membrane probe is that any contaminant in the
water which reduces the molecular attraction between
water molecules and oxygen molecules will reduce the
concentration of dissolved oxygen at saturation in the
water sample; however, at saturation the response of a
membraned probe will be identical in clean or contami-
nated water and independent of oxygen concentration. To
put it another way, a membraned probe cannot be
calibrated in clean water and then placed in a contami-
nated water sample to read an oxygen concentration to
be used in the determination of a beta value. For the
oxygen probe to be utilized, it must be calibrated in the
contaminated sample utilizing some reference technique
such as Winkler titration.
Direct Contact Probes
Probes which provide a direct electrode contact with the
sample do not have the same calibration precaution as
related to the probes which utilize a membrane. The
oxygen arrives directly at the electrode in proportion to
the concentration of molecular oxygen in the water
82
-------
Figure 4. Membraned Dissolved Oxygen Probe Response in Water at Saturation
00
CO
o o
oo oo oo
o
-------
sample. Where other problems such as fouling of the
electrode surface and electrolyte concentrations are not a
factor, a direct contact probe would be more suitable for
the determination of beta values.
Summary
1. Excellent information exists for saturation of oxygen in
water, but direct application in calculation of driving
forces in oxygen transfer is subject to question.
2. Since oxygen saturation in water is a dynamic equilib-
rium, the concentration value could be influenced by
conditions at the interface such as evaporation or
condensation.
3. Controlled experiments at air velocities of 5 mph did
not reveal any significant effect on the saturation
value of oxygen in water over a range of evaporation
and condensation conditions. Different results might
be found at increased air velocities.
4. Where air-water interfaces are purposely dispersed in
an oxygen transfer system, it is recommended that
oxygen saturation be determined experimentally.
5. In addition to the interface condition (oxygen concen-
tration), solubility of oxygen is influenced by molecular
attraction. Thus, dissolved salts or other agents which
alter molecular attraction can alter oxygen solubility.
6. Winkler titration may be affected by interference from
cobalt precipitates in oxygen transfer studies, but
iodine volatility can be an even greater source of error
if interference is not corrected by the titration tech-
nique or by utilizing an alkaline-iodide reagent
containing a high concentration of iodide.
7. Membraned probes respond to the "partial pressure"
of oxygen and require a reference calibration in con-
taminated water in order to measure the actual
concentration of oxygen. The precaution is important
in determination of beta values.
References
1. Carpenter, J.H. "The Accuracy of the Winkler Method
for Dissolved Oxygen Analysis". Limnol. Oceanog.,
10, pp. 135-140, 1965.
2. Carpenter, J.H. "New Measurements of Oxygen
Solubility in Pure and Natural Water". Limnol.
Oceanog., 11, pp. 264-277, 1966.
3. Fox, C.J.J. "On the Coefficients of Absorption of
the Atmospheric Gases in Distilled Water and Sea
Water". Part 1: Nitrogen and Oxygen. J. Conseil,
Conseil Perm. Intern. Exploration Merg Publ. No. 41
26pp., 1907.
4. Fox, C.J.J. "On the Coefficient of Absorption of
Nitrogen and Oxygen in Distilled Water and Sea Water
and Atmospheric Carbon Dioxide in Sea Water".
Transactions Faraday Society, 5, pp. 68-87, 1909.
5. Kalinske, A.A., LD. Lash, and G.L Shell. "Cobalt
Interference in the Non-Steady State Clean Water
Test". Water and Sewage Works, 120, pp. 54-59
July 1973.
6. Montgomery, H.A.C., N.S. Thorn, and A. Cockburn.
"Determination of Dissolved Oxygen by the Winkler
Method and the Solubility of Oxygen in Pure Water
and Sea Water". Journal Applied Chemistry, 14:
pp. 286-296, 1964.
7. "Standard Methods for the Examination of Water and
Waste water". 14th Edition, American Public Helath
Association, American Water Works Association,
Water Pollution Control Federation, 1975.
8. Whipple, G.C., and M.C. Whipple, "Solubility of
Oxygen in Sea Water", Journal American Chemistry
Sec., 33, pp. 362-365, 1911.
84
-------
Accounting for the Effects of
Water Temperature in Aerator
Test Procedures
14
John S. Hunter, III
3M Company
St. Paul MN 55133
At any given temperature T:
(1) WT = (
where T is expressed in °C, WT is the oxygen mass
transfer rate per unit volume in clean water in mg/(l sec),
KLa is the overall liquid phase volumetric oxygen mass
transfer coefficient in sec'1, Cf is the saturation dissolved
oxygen concentration for the water in mg/l and C is the
dissolved oxygen concentration in the water in mg/l. If
the water temperature increases from 10 to 20°C, the
value of KLa can increase by more than 50% in some
cases, while the value of C* will decrease by approxi-
mately 20%. Thus, the influence of water temperature on
aerator testing is quite significant and must be taken into
account.
Saturation Dissolved Oxygen Concentration Tables
Use of C* Tables
The value of (KLa)T is not measured directly during an
aerator test but is calculated. Some procedures for
calculating (KLa)t require that Cf be known. The value
of CT is itself sometimes calculated. This is done by
referring to a table that lists saturation dissolved oxygen
concentrations in clean water at a pressure of one
atmosphere for a wide range of water temperatures. The
CT value is calculated by using Henry's Law to adjust
the saturation dissolved oxygen concentration in the table
corresponding to the water temperature employed in the
aerator test. The actual calculation is beyond the scope
of this paper and will not be discussed here.
A partial sensitivity analysis was performed by Hunter
and Ward (1 1 } to determine how critical the value of Cf
is in determining (KLa)T. It was found that a deviation of
±0.10 mg/l from the correct value of C-f can cause an
error of approximately 5% in the calculation of (KLa)T. This
takes on real significance when one considers that
values of saturation dissolved oxygen concentration that
appear in various published tables differ from one
another by as much as 0.4 mg/l for clean water at 20°C
and a pressure of one atmosphere.
Another important use of saturation dissolved oxygen
tables is in the determinination of a value for the satura-
tion dissolved oxygen concentration that can be used to
calibrate the instrumentation for measurement of
dissolved oxygen concentrations during aerator testing.
From all of the above, it can be seen that saturation
dissolved oxygen tables are a necessity in performing an
aerator test if the effects of water temperature are to be
properly accounted for. As there are differences in the
values of saturation dissolved oxygen concentrations
between various tables in the literature and because, as
shown by Hunter and Ward (11), these differences can
cause appreciable variations in the value of (K|_a)y
calculated for any aerator test, it is imperative that an
accurate table be used. This table should become a part
of any standard to be adopted for aerator testing so that
comparable results can be obtained from tests performed
by different individuals.
Selection of C* Table
For many years, the most often used values for saturation
dissolved oxygen concentrations were those of Fox (9),
Winkler (20) (21) and Montgomery, Thorn and Cockburn
(14). Fox's values for a dry atmosphere were corrected by
Whipple (19) to a water vapor saturated atmosphere.
These are the values that have been included in the past
several editions of Standard Methods for the Examination
of Water and Wastewater. As noted in a review by
Hunter (10), evidence has gradually accumulated over the
years that shows these saturation dissolved oxygen
concentrations are not accurate. The upcoming 15th
edition of Standard Methods will contain a new table of
saturation dissolved oxygen concentrations. This table is
reproduced here as Table 1. The background behind this
table is presented by Postma, Svansson, Lacombe and
Grasshoff (16). This table is based on the 1973 Inter-
national Oceanographic Tables jointly published by the
National Institute of Oceanography and the United
Nations Educational, Scientific and Cultural Organization
under the supervision of SCOR-UNESCO-ICES-IAPSO
Joint Panel of Experts for Oceanographic Tables and
Standards. Table 1 represents the most up-to-date and
accurate saturation dissolved oxygen concentration data
currently available. It is recommended that Table 1
become a pan of any standard to be adopted for aerator
testing.
Kua Water Temperature Correction Factor
Present Practice
An aerator performance test determines a value for KLa
at the water temperature employed during the test. In
order to standardize aerator test ratings, it is usual
practice to express the aerator test results in terms of
85
-------
Table 1. Solubility of Oxygen in Water Exposed to
Water Saturated Air (mg/l)
Chloride Concentration in Water
Temperature (mg/l)
(°C)
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
0
14.60
14.19
13.81
13.44
13.09
12.75
12.43
12.12
11.83
11.55
11.27
11.01
10.76
10.52
10.29
10.07
9.85
9.65
9.45
9.26
9.07
8.90
8.72
8.56
8.40
8.24
8.09
7.95
7.81
7.67
7.54
7.41
7.28
7.16
7.05
6.93
6.82
6.71
6.61
6.51
6.41
6.31
6.22
6.13
6.04
5.95
5.86
5.78
5.70
5.62
5.54
5,000 10,000 15,000 20,000
13.72
13.35
12.99
12.65
12.33
12.02
11.72
11.43
11.16
10.90
10.65
10.40
10.17
9.95
9.73
9.53
9.33
9.14
8.95
8.77
8.60
8.44
8.28
8.12
7.97
7.83
7.69
7.55
7.42
7.30
7.17
7.05
6.94
6.82
6.71
6.61
6.51
6.40
6.31
6.21
6.12
6.03
5.94
5.85
5.77
5.69
5.61
5.53
5.45
5.38
5.31
12.90
12.56
12.23
11.91
11.61
11.32
11.05
10.78
10.53
10.29
10.05
9.83
9.61
9.41
9.21
9.01
8.83
8.65
8.48
8.32
8.16
8.00
7.85
7.71
7.57
7.44
7.31
7.18
7.06
6.94
6.83
6.71
6.61
6.50
6.40
6.30
6.20
6.11
6.02
5.93
5.84
5.76
5.68
5.60
5.52
5.44
5.37
5.29
5.22
5.15
5.08
12.13
11.81
11.51
11.22
10.94
10.67
10.41
10.17
9.93
9.71
9.49
9.28
9.08
8.89
8.71
8.53
8.36
8.19
8.03
7.88
7.73
7.59
7.45
7.32
7.19
7.06
6.94
6.83
6.71
6.60
6.49
6.39
6.29
6.19
6.10
6.01
5.92
5.83
5.74
5.66
5.58
5.50
5.42
5.35
5.27
5.20
5.13
5.06
5.00
4.93
4.87
11.41
11.11
10.83
10.56
10.30
10.05
9.82
9.59
9.37
9.16
8.96
8.77
8.58
8.41
8.24
8.07
7.91
7.76
7.61
7.47
7.33
7.20
7.07
6.95
6.83
6.71
6.60
6.49
6.38
6.28
6.18
6.08
5.99
5.90
5.81
5.72
5.64
5.56
5.48
5.40
5.33
5.25
5.18
5.11
5.04
4.98
4.91
4.85
4.78
4.72
4.66
(KLa)20- This is done by dividing (KLa)T by the
dimensionless water temperature correction factor F
which is defined as follows:
(2)
F = (KLa)T/(KLa)20
Two expressions are normally used for calculating F. The
most common expression in the literature for F is:
m F - ft (T-20)
\^l < ~ "e
where 6e is the dimensionless geometric temperature
coefficient. The other expression for F is:
(4)
F = c + 0| T
where c is a dimensionless constant which is the 0°C
intercept of the plot F vs T, and 0, is the slope of the
plot F vs T, known as the linear temperature correction
coefficient that is expressed as °C~1. Reported values
for 6e range from 1.012 to 1.047 [Lakin and Salzman
(12)] and reported values for 0| range from 0.015 to
0.022 [Hunter and Ward (11)]. All reported values of 0e
and (91 have been determined empirically. These
apparent discrepancies in reported values of 0e and 6,
are significant. For example, the temperature correction
factor F for converting (KLa)io to (KLa>2o calculated
by using a value of 1.047 for Oe is approximately
30% larger than F calculated by using a value of 1.012
for 6e. This points out the need for establishing an
accurate method for adjusting (KLa)T to (KLa)2o- This
method should become a part of any standard to be
adopted for aerator testing so that comparable results
can be obtained from tests performed by different
individuals.
Proposed Method for Determining (KLa)20
Equations 3 and 4 indicate that the KLa water tempera-
ture correction factor F is exclusively dependent on water
temperature. This is definitely not the case. The water
temperature correction factor is also dependent upon
turbulence. This is evident in empirical results reported
by Elmore and West (8), Metzger (13), and Bewtra,
Nicholas and Polkowski (1). In addition, Metzger (13),
Rathbun and Bennett (17) and Camp and Meserve (4),
starting with the film-penetration theory of gas-liquid
mass transfer, have demonstrated that theoretically the
KLa water temperature correction factor should be both a
function of water temperature and turbulence and that
the correction factor should increase with decreasing
turbulence.
By combining turbulent flow theory with the film-pene-
tration theory of gas-liquid mass transfer, it can be
shown that it should be possible to express KLa for any
aeration system as a function of G, the temporal mean
velocity gradient, and other system variables [Hunter
(10)], that is:
(5)
KLa = f (G, other system variables)
86
-------
where, in terms of measurable parameters:
(6)
in which G is expressed in sec"1, P is total power input
into the water being aerated in ergs/sec, V is the volume
of water being aerated in cm3 and n is the absolute
viscosity of the water in poise.1
It can be seen from Equation 6 that G is temperature
dependent because of the inclusion of absolute viscosity
which is temperature dependent. The G term is also
considered to be an index of turbulence. Thus, Equation 5
would inherently take both water temperature and turbu-
lence into account without recourse to the use of a
correction factor or factors.
It is important to note that the temperature correction
factor F should not be calculated as the ratio M2C/MT
nor G-T/G20. but as (KLa)T/(Ki_a)2o using KLa as
described in Equation 5.
Hunter (10) has verified Equation 5 for the case of lab-
oratory-scale submerged turbine aeration systems. For
the tank geometry used, Hunter found that:
(7) KLa = [4.04 + 0.00255G2 (D/T)4]Q° 63 '
where KLa is expressed in hr"1, G in sec'1, D is impeller
diameter in cm, T is tank diameter in cm and Q is air
discharge rate in SCFH. By substituting P/V for G2 in
Equation 7, then substituting Equation 7 into Equation 2,
one obtains:
(8)
F =
[4.04 + 0.00255 (D/T)4 (P/VMT)]Q° 63
[4.04 + 0.00255(D/T)4 (P/VM2o>]Q063
For the case where D/T is 0.35, Q is 1.0 SCFH, P/V is
2000 ergs/sec-cm3 and water temperature T is 10°C
( MIO = 0.01307 poise and M20 = 0.01002 poise)
[Weast (18)1 F is 0.841. By decreasing P/V to 500 ergs/
sec-cm3, F becomes 0.924. Thus, Equation 8 is con-
sistent with the theoretical work mentioned previously
that showed F should be a function of both water
temperature and turbulence and that F should increase
with decreasing turbulence.
Data are presented by Hunter and ward (11) that show
Equation 7 is valid over a water temperature range of
0< T< 40°C (32< T< 104°F) without the use of any
correction factors. These data are presented in Table 2.
Thus, on both empirical and theoretical grounds, it
appears that relationships for expressing KLa for aeration
systems as a function of G and other pertinent system
variables adequately take into account the effect of water
temperature on KLa, including the influence of turbu-
lence on this effect.
It is recognized that equations probably do not exist at
the present time for full-scale aeration systems that
express KLa as a function of G. Therefore, it is recom-
mended research be undertaken to attempt to develop
empirically such equations and to determine if such
equations can adequately account for the effects of water
temperature on KLa at different levels of turbulence.
In the meantime, specification of a single value for 6e or
c and 0| in an aerator testing standard would be arbi-
1. Equation 5 is developed in the attachment. The reader is
referred to Camp and Stein (3) and Camp (2) for derivation of
Equation 6.
Table 2. Comparison of Measured KLa at Water Temperatures of 0 S T < 40°C with KLa
Expressed as a Function of Temporal Mean Velocity Gradient G
Water
Temperature
(°C)
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
40.0
KLa Based on
Absolute Data from
Viscosity* Hunter & Ward (11)
(poise) (hr1)
0.01 787
0.01519
0.01 307
0.01139
0.01002
0.008904
0.007975
0.007194
0.006529
7.99
9.12
10.26
11.39
12.53
13.66
14.79
15.93
17.06
KLa Calculated
from Relationship
between Kia and G AKia
(hr'1) (hr'1)
8.80
9.64
10.55
11.51
12.53
13.60
14.71
15.87
17.08
-0.81
-0.52
-0.29
-0.12
0.00
0.06
0.08
0.06
-0.02
Percent
Difference
-10
- 6
- 3
- 1
0
<1
<1
<1
<1
See pp. 91-92 of Hunter (10) for development of this table.
•Weast [(18), p. F-36]
87
-------
trary and misleading. The effects of turbulence should
result in different values for 6eor c and 0|for different
aerators as evidenced by the wide ranges of values for
0eand 0j that have been reported in the literature.
Aeration equipment manufacturers should be required to
develop empirical values for 0e or c and 0, to be pro-
vided along with values of (KLa)2o in order to describe
more adequately the performance of the aerators they
are marketing.
Conclusions
The effect of water temperature on oxygen mass transfer
rate is significant. The physical mechanism involved is
currently only speculative. In the absence of a compre-
hensible theory, empirical data available in the literature
appear contradictory. A definite need exists for more
research on this subject.
The following recommendations are put forth for con-
sideration in the development of a standard aerator test
procedure to ensure that the influence of water tempera-
ture on aerator testing is taken into account:
1. The revised saturation dissolved oxygen table that will
appear in the upcoming 15th edition of Standard
Methods and reproduced here as Table 1 should be
incorporated into the aerator testing standard. The use
of this table should be required for all calculations that
require the use of saturation dissolved oxygen tables.
2. Research should be undertaken to attempt to empiri-
cally develop equations for full-scale aeration systems
that would express KLa as a function of the temporal
mean velocity gradient G and to determine if such
equations can adequately account for the effects of
water temperature on KLa at various levels of
turbulence.
3. Aeration equipment manufacturers should be required
to empirically develop values for 0e or c and 0|that
would be provided along with values of (KLa)2o in
order to more adequately describe the performance of
the aerators they are marketing.
It is understood that the above recommendations do not
represent a final answer to the question of how to
account for the effects of water temperature on aerator
testing. It is hoped, however, that these recommendations
will provide a basis for developing a test procedure that
incorporates the best available information on this
subject.
References
1. Bewtra, J.K., W.R. Nicholas, and LB. Polkowski.
"Effect of Temperature on Oxygen Transfer in
Water". Water Research (British) 4, pp. 115-123,1970.
2. Camp, Thomas R., and R.L. Meserve. "Water and
Its Impurities". Oowden, Hutchinson and Ross,
3. Camp, Thomas R., and P.C. Stein. 'Velocity
Gradients and Internal Work in Fluid Motion."
Journal of the Boston Society of Civil Engineers 30
pp. 219-237, 1943.
4. Camp, Thomas R., and R.L. Meserve. Water
and Its Impurities. Dowden, Hutchinson and Ross,
Inc., Stroudsburg PA, 1974.
5. Committee on Sanitary Engineering Research, Sani-
tary Engineering Division, American Society of Civil
Engineers. "Effect of Water Temperature on
Steam Reaeration." Journal of the Sanitary Engi-
neering Division, Proceedings of the American
Society of Civil Engineers 87 (SA6), pp. 59-71, 1961.
6. Dobbins, William E. "The Nature of the Oxygen
Transfer Coefficient in Aeration Systems." Section
2-1 in Brother J. McCabe and W.W. Eckenfelder, Jr.,
Editors, Biological Treatment of Sewage and Indus-
trial Wastes, Vol. 1, Aerobic Oxidation, Reinhold
Publishing Corporation, New York, 1956.
7. Dobbins, William E. "Mechanisms of Gas Absorption
by Turbulent Liquids, "pp. 61-96 in W.W. Eckenfelder,
Jr., Editor, Advances in Water Pollution Research,
Vol. 2, Proceedings of the First International Con-
ference on Water Pollution Research (London), The
MacMillian Company, New York, 1964.
8. Elmore, Harold L, and William F. West. The work of
Elmore and West is "Effect of Water Temperature on
Stream Reaeration" reported by Committee on
Sanitary Engineering Research, Sanitary Engineering
Division, American Society of Civil Engineers, 1961.
9. Fox, C.J.J. "The Coefficients of Absorption of
Nitrogen and Oxygen in Distilled Water and Sea
Water and of Atmospheric Carbonic Acid in Sea
Water." Transactions of the Faraday Society 5
pp. 68-87, 1909.
10. Hunter, John S. "A Basis for Aeration Design".
Doctor of Philosophy Dissertation, Department of
Civil Engineering, Colorado State University,
Fort Collins CO, 1977.
11. Hunter, John S., and John C. Ward. "Part I —
The Effects of Water Temperature and Elevation
upon Aeration." pp. 1-52 in The Mechanism of
Waste Treatment at Low Temperature — Part B:
Sanitary Engineering by John C. Ward, John S.
Hunter and Richard P. Johansen. Colorado State
University Environmental Resources Center Comple-
tion Report Series Number 34 as submitted to the
Office of Water Resources Research, United States
Department of the Interior, Washington DC, 1972.
12. Lakin,Michael B., and Ronald N. Salzman. "Sub-
surface Aeration Evaluation." Paper presented at
the 50th Annual Water Pollution Control Federation
Conference, Philadelphia PA, October 1977.
88
-------
13. Metzger, Ivan. "Effects of Temperature on Stream
Aeration." Journal of the Sanitary Engineering
Division, Proceedings of the American Society of
Civil Engineers 94 (SA6), pp. 1153-1159, 1978.
14. Montgomery, H.A.C., N.S. Thorn, and A. Cockburn.
"Determination of Dissolved Oxygen by the
Winkler Method and the Solubility of Oxygen in Pure
Water and Sea Water." Journal of Applied Chemistry
14, pp. 280-296, 1964.
15. O'Connor, Donald J., and William E. Dobbins. "The
Mechanism of Reaeration in Natural Streams."
Journal of the Sanitary Engineering Division, Pro-
ceedings of the American Society of Civil Engineers
82 (SA6), pp. 1 -30, 1956.
16. Postma, H., A. Svansson, H. Lacombe and
K. Grasshoff. "The International Oceanographic
Tables for the Solubility of Oxygen in Sea Water."
Jour. Cons. Int. Explor. Mer (Journal of the Inter-
national Council for the Exploration of the Sea) 36
(3), pp. 295-296, 1976.
17. Rathbun, Ronald E., and James P. Bennett. Discussion
of "Effects of Temperature on Stream Aeration"
by Metzer. Journal of the Sanitary Engineering Division,
Proceedings of the American Society of Civil Engineers
95 (SA5), pp. 985-988, 1968.
18. Weast, Robert C., Editor. "Handbook of Chemistry and
Physics, 51st Edition". The Chemical Rubber
Company, Cleveland OH, 1970.
19. Whipple, G.C., and M.C. Whipple. "Solubility of Oxygen
in Sea Water". Journal of the American Chemical
Society 33, pp. 362-365, 1911.
20. Winkler, L.W. "Ber Dtsch. Chem. Ges". 22, p.1974,
1889. (as cited by Montgomery, Thorn, and Cockburn,
p. 280, 1964).
21. Winkler, LW. "Ber Dtsch. Chem. Ges". 24, p. 3602,
1891. (as cited by Montgomery, Thorn, and Cockburn,
p. 280, 1964).
89
-------
Attachment
Relationship Between G and Oxygen Mass
Transfer Rate
The physical process of oxygen mass transfer is the same
whether it occurs in a bubble diffusion, submerged
turbine, or a mechanical surface aeration system. That is
to say, any one gas-liquid mass transfer theory (such as
the two-film theory) can be used to model the oxygen
mass transfer process that occurs in any type of aeration
system. Therefore, if an accepted model of gas-liquid
mass transfer indicates that KL (the overall liquid phase
oxygen mass transfer coefficient) can be expressed as a
function of a measurable parameter such as temporal
mean velocity gradient (G), it may be possible to express
KL or KLa for any aeration system as a function of that
same parameter. The following discussion utilizes the
film-penetration theory of gas-liquid mass transfer and
turbulent flow theory to show that KL or KLa can be
expressed as a function of temporal mean velocity
gradient (G).
The film-penetration theory was developed by Dobbins (6)
using a physical model similar to that of the two-film
theory. Dobbins envisioned the liquid film to exist in a
statistical sense with the film always present and its
contents being randomly exchanged with portions of the
bulk liquid. The film-penetration theory states that:
(A-1)
KL = (rDr
J/2
ctnh(rLL/DJ1/2
in which KL is expressed in cm/sec, r is the fraction of
gas-liquid interface being replaced for each unit of time
in sec"1, Dm is the coefficient of molecular diffusion of
the gas in the liquid expressed as cmVsec and LL is the
liquid film thickness in cm.
It can be seen that KL in Equation A-1 is a function of
two parameters, r and LL. Both parameters are charac-
teristic of the hydrodynamics of the system. When
turbulence increases, r increases and LL decreases. In
his laboratory experiments, Dobbins (7) found that a
consistent relationship exists between r and LL that is
independent of mixing conditions. Because r and LL are
not independent of each other, KL can be considered to
be expressed as a function of r in Equation A-1.
O'Connor and Dobbins (15) developed stream aeration
concepts based on the film-penetration theory of gas-
liquid mass transfer. They considered r, the rate of
gas-liquid interfacial surface renewal, to be:
(A-2)
where V v2 is the vertical root mean square velocity
fluctuation of the liquid at the liquid surface expressed in
cm/sec and I is the Prandtl mixing length in cm de-
fined as:
IA-3)
I = K (dv/ds)/(d2v/ds2)
where 0.3 < K (a constant) < 0.45 and dv/ds is the
velocity gradient expressed in sec'1.
According to the Prandtl mixing theory:
(A-4) I (dv/ds) =\/v*~
By rearranging terms:
(A-5) dv/ds =\/v2/i
Substituting Equation A-5 into Equation A-2 gives:
(A-6) r = dv/ds
By definition, at a finite point [Camp and Stein, (3)]:
(A-7) G = dv/ds
By substituting Equation A-7 into Equation A-6, it is
found that:
(A-8)
r = G
It was stated previously that the film-penetration theory
expressed overall liquid phase oxygen mass transfer
coefficient KL as a function of r. However, as shown
above, r equals G so that it is also possible to express KL
as a function of G. Furthermore, because the film-
penetration model is generally assumed to be applicable
to the process of oxygen mass transfer in any aeration
system, it may be possible to express KL for all aeration
systems as a function of G. The same is true for KLB
because KLa is simply the product of KL and a, which is
the gas-liquid interfacial surface area per unit volume.
It is known that aeration system variables such as
aeration tank geometry and (except for surface aerators)
air flow rate have a significant effect on oxygen mass
transfer rate and must be taken into account. Therefore,
for any particular aeration system:
(A-9)
KLa = f (G, other system variables).
90
-------
Influence of pH and Iron
and Manganese Concentrations
on the Non-Steady State Clean
Water Test for the Evaluation
of Aeration Equipment
Hussein Naimie and Steve Nelson
Paper presented by
Daniel A. McCarthy
Eimco Process Machinery Division
Envirotech Corporation
Salt Lake City UT 84110
Introduction
The non-steady state clean water test is the most widely
accepted procedure for the evaluation and comparison of
the oxygen transfer efficiency of various types of aeration
equipment. First, a known quantity of clean tap water is
chemically deoxygenated to a very low dissolved oxygen
(DO) concentration by the addition of sodium sulfite as a
reducing agent. Typically, about one and one-half times
the sulfite required to stoichiometically react with the DO
present is used (17). High sulfite addition is reported to
result in high (KLa)2o values (6)(13).
Since the rate of the sulfite auto-oxidation reaction is
very slow and the presence of sulfite ions interferes with
Winkler DO determinations, addition of a catalyst is
necessary to increase the rate of sulfite oxidation (12).
The accepted catalyst is colbalt (II) chloride hexahydrate
or other cobalt (III) salts (Co++). Conventionally, 2-5 mg/l
or even as high as 10 mg/l of Co4"1" has been recom-
mended and employed (9). Chemical deoxygenation is
followed by oxygenation and data collection.'During the
oxygenation (aeration) process the rate of oxygen transfer
from the gas phase into the liquid phase is determined by
measuring the dissolved oxygen concentration as a
function of time.
The presently accepted non-steady state clean water test
to evaluate aerator performance is reported to have
caused interference in the Winkler DO determinations
(9X11X13X14). A precipitate is reported to cause the
interference. This interfering precipitate is postulated to
be an unidentified cobalt complex or cobalt (III) hydroxide
(9X10).
Accordingly, various recommendations have been made
to eliminate this interference. An initial cobalt (II) ion
concentration of 0.05 mg/l is recommended by Kalinske
et al. (9), while the usage of 2 mg/l of cobalt combined
with a determination of the degree of interference for
each aeration test has been suggested by Lakin (10). No
theoretical justification, however, is submitted nor
discussed by either Kalinske or Lakin to support their
recommendations.
15
In contrast to Lakin's and Kalinske's observations, insig-
nificant variation in oxygen transfer efficiency was found
by using up to 5 mg/l of cobalt in aeration tests using
city tap water at Melbourne, Australia and Rochester,
New York (IX11).
These latter results are in agreement with the results of
the non-steady state clean water tests obtained by Eimco
Process Machinery Division (PMD) in the Northeastern
United States. No differences in oxygen transfer
efficiency were found when adding as much as 8 mg/l
cobalt and then using Winkler or polarographic methods
to determine DO concentration. On the other hand, under
identical conditions and with the same type of aerator,
significant interference with the Winkler DO determina-
tion was observed using Salt Lake City tap water.
The inconsistent results reported in the literature and
experienced by Eimco PMD prompted the initiation and
continuation of diagnostic studies at Eimco. Phases I
and II of this work were designed to define the following
aspects:
1 To identify and determine the causes and mechanisms
of so-called cobalt interference
2. To develope a predictive model for the so-called cobalt
interference
3. To explain the inconsistency of cobalt interference for
different natural water systems.
The results of these studies are reported elsewhere
(13X14).
The purpose of the present study was to quantify further
the influence of water chemistry based on recent findings
and to identify an interference-free, non-steady state
clean water method to compare and objectively determine
transfer efficiencies of various types of aeration equip-
ment using different natural waters in the test system.
91
-------
Theoretical Considerations and
Summary of Earlier Findings
The chemistry-interfering aspects of the Winkler DO, the
electrochemistry of cobalt ion species in aqueous solution,
sulfite oxidation products and formation of hydrogen
peroxide during aeration testing are reported in detail
elsewhere (13X14). The essential findings contained in
the referenced work are reported here to provide clarity
and continuity to the present paper.
It was found that the interference in the non-steady state
test was not caused by the precipitate alone, as is postu-
lated by others (9X10). Actually, up to 80-90% of the
error for DO measurements by the Winkler method can
be attributed to one or more soluble components if the
initial cobalt concentration is less than 2 mg/l. One
soluble component was determined to be hydrogen
peroxide (H202) which is a decomposition product of
peroxy monosulfuric acid (H2S05). The formation of
peroxy monosulfuric acid in the presence of oxygen,
sulfite and a catalyst in aqueous solution was known and
reported as early as 1934 by Baeckstroem (2).
At relatively high initial cobaltous ion concentrations,
most of the hydrogen peroxide is catalytically decomposed
due to the formation of suspended solids at natural tap
water pH levels. The catalytic decomposition of hydrogen
peroxide in the presence of surface active substances
such as metal oxides and other colloidal matter is well
known (7).
Electron spin resonance (ESR) spectroscopic analysis and
oxygen equivalent measurement, together with other
chemical identification and structural analysis methods,
have shown that the interfering precipitate consists
mainly of hydrated trivalent and tetravalent cobalt oxide.
This interfering precipitate was determined to be formed
at pH values of 6.9 and above. No precipitate was
formed, however, if the pH was kept below 6.9.
Furthermore, the pH and the buffering capacity of test
water were determined to be the most critical water
quality parameters which directly affected the magnitude
of the interference phenomenon. Accordingly, it was
recognized that a precise quantitative prediction could not
be made of the level of cobalt interference for various
natural water systems. This also was observed if the
wastewater was physically deoxygenated (nitrogen strip-
ped). Different water systems under uniform and constant
experimental conditions (constant power intensity,
constant rpm, constant air flow) showed different (Ki_a)20
values when using the same prototype laboratory sub-
merged turbine aerator. The data from our previous work
are summarized in Table 1 (14).
In general, the so-called cobalt interference is soft water
systems was determined to be minimal or totally absent.
Water systems with higher pH and substantial alkalinity
caused interference of unpredictable magnitude.
Re-evaluation of the work previously reported indicated
that there may be a correlation between the alkalinity of
the test water and the magnitude of the interference.
Dallas, Texas water, which has a relatively low alkalinity,
showed an unexpectedly high (KLa)2o value. On the other
hand, the Dallas, Texas test water collected from the
aeration test basin showed an unusually high iron
concentration of 25 mg/l. It was suspected that the high
iron concentration could provide some sort of enhance-
ment of oxygen transfer. This aspect, however, was not
discussed or quantified in our previous study. This aspect
will be quantified in this present report.
Another aspect of the catalyzed sulfite oxidation reaction,
which was not previously discussed, is the pH depend-
ence and the catalytic mechanism of cobalt. For example,
does the cobalt act as a homogeneous or heterogeneous
catalyst? If cobalt acts as a heterogeneous catalyst, then
the rate of sulfite oxidation at low pH should be very slow
and probably some sulfite will remain in the solution
which will interfere with Winkler DO measurements.
These and other questions are dealt with as a part of the
present paper.
The rate of the non-catalytic oxidation of sulfite with DO,
as a function of sulfite ion concentration for pH values of
Table 1. Previously Reported Oxygen Transfer Data (14)
Method of Deoxygenation
Geographical Location
Harrisburg, Virginia
Spokane, Washington
Rochester, New York
Springfield, Mass.
Dallas, Texas
Concord, New Hampshire
Initial
pH
7.4
7.0
7.1
7.5
7.5
6.7
Water
Alkalinity
(mg/l)
as CaC03
192.0
152.0
80.0
50.0
41.0
24.0
Quality Parameter
Test
Water
Temp.
(°C)
22.0
22.5
23.4
22.0
21.0
22.5
DO
Saturation
C*
(mg/l)
7.51
4.48
7.4
7.48
7.5
7.9
Iron
(mg/l)
0.55
—
0.55
0.75
25.0
0.9
N2-Strip
Probe
(KLa)2o
(hr-1)
8.32
8.27
5.77
4.09
7.3
4.45
Winkler
(KLa)20
(hr-1)
9.44
8.15
6.32
4.07
7.22
7.1
pH after
N2-Strip
8.1
7.6
7.9
8.0
7.2
7.1
One SO| Addition
Probe
(KLa)2o
(hr1)
8.49
8.57
5.66
4.40
8.3
4.36
Winkler
(KL«)2o
(hr-1)
12.84
10.17
7.16
7.76
4.45
92
-------
6.5-7.7 for a DO concentration greater than 0.8 mg/l
and at 20°C, was studied and defined by Gale (4). Under
these conditions the rate is:
dC/dt = 1.48 x 1021 (S0|)233 (H+)1 58
Furthermore, it was found that the rate of sulfite oxida-
tion by DO, over the pH range of 4.2-6.4 and under the
above mentioned conditions, was dependent upon the
first power of sulfite ion concentration and some
unidentified variable and was very slow (4).
As early as 1 897, it was observed that mixtures of
compounds susceptable to auto-oxidation often gave
results quite different from those formed when a single
compound alone was studied. In particular, certain
compounds which reacted very slowly with oxygen in
pure solutions were oxidized much more rapidly when
sulfite was presented (8). Also, the rate of sulfite auto-
oxidation was found to be markedly affected by various
materials (3)(5)(8)( 18).
The studies cited above attempted to explain these
results in terms of various hypothetical reaction mechan-
isms for relatively high sulfite concentrations in aqueous
solutions. The proposed mechanisms often conflicted in
detail, many were rather vague and all were basically
speculative in nature. Nonetheless, most studies con-
tained the postulate that some kind of inactivated
intermediate molecule, ion, and/or radical was necessary
to explain the observed results (4). Specifically, the
activation of oxygen in the presence of SO^ ion and
some heavy metals were the interpretation of the above
observed phenomenon. Baeckstroem proposed the chain
reaction mechanism which is now most generally
accepted as the true explanation for the catalytic activa-
tion of oxygen molecules by certain metals (2). A chain
reaction is caused by the presence of radicals. Such
mechanisms of oxygen activation in the presence of
transitional elements in either the heterogeneous or
homogeneous phase, with and without sulfite addition,
can also be assumed and postulated in reaeration testing
using physical or chemical deoxygenation processes
followed by oxygenation.
A hypothesis which envisions the activation of oxygen by
metals and/or metal oxides is as follows. Dissolved
oxygen in the presence of certain metals is activated
(O2). This oxygen molecule can react with water to form
OH* radicals or hydrogen peroxide and atomic oxygen
(or oxygen radicals) as follows:
02 + H20.sptO- + 20H'
Reaction of these products will result in the formation of
oxygen molecules and water as follows:
*H21/2 02
O'+O* — M/2 02
This reaction path is established and is a well known
mechanism in the fields of photochemical and ozone
chemistry. In the practice of aeration testing it is often
observed that during the oxygenation step the vicinity of
the aeration basin smells chlorine- or ozone-like. The
production of chlorine in certain natural waters with high
Cl~ concentration can only be explained by the presence
of radicals, especially the hydroxyl radical which has a
redox potential of 2.8 volts (15).
Experimental Procedures
A 4,226 liter plexiglass tank approximately 5.9 square
meters in surface area and 1.2 meters in height was
used to conduct the non-steady state reaeration tests.
Reareation was accomplished by using a geometrically
and kinematically scaled-down version of the Eimco
PMD submerged turbine aerator. The turbine was 21.8
centimeters in diameter and was submerged to a depth
of 0.5 meter below the water surface. The air flow rate,
561 l/min, was metered and controlled by an accurate
flow meter (±2% full-scale) manufactured by Fisher-
Porter. Air to the turbine was supplied via an air
compressor.
To achieve uniform mixing intensity, the mixer for all
experiments was operated at 400 rpm. The rotational
speed was determined with a Pioneer Photo Tach Model
36 (±2% full scale). The temperature and pressure of the
air flow to the submerged turbine were monitored at the
point of exit from the flow meter.
Dissolved oxygen (DO) was measured using the Azide
modification of the Winkler method (16).
To determine the magnitude of the interference occurring
with the Winkler procedure, DO was also measured and
recorded continuously during each run by the polaro-
graphic method using a YSI Model 54 DO meter and a
Soltech Model 100 recorder. A probe with an integral
mixer near the probe membrane was used to record DO
polarographically. The probe was calibrated daily using
Winkler DO determinations. The pH of the test water was
monitored by a Sargent-Welch meter.
A 0.0015 molar phosphate buffer was used to adjust and
maintain pH within 6-8. Salt Lake City tap water was
used in this work.
The average makeup of test water was as follows:
Alkalinity - CaC03:180 mg/l
pH - 8.2
Fe - less than 0.01 mg/l
Mn - less than 0.1 mg/l
To determine the effect of pH on mass transfer, the test
water was acidified to the desired pH and buffered with
0.0015 molar phosphate by adding proper mass ratios
of K2HPO4 and KH2P04 to keep the pH constant during
consecutive reareation testing.
The addition of sulfite required for one test did not
change the pH and no buffering of the test water was
93
-------
necessary. Consecutive reaeration tests were run for
various pH values across the pH range of 4 to 9. Con-
secutive reaeration tests consisted in this case of
physical deoxygenation-reoxygenation (N2Strip-reaeration)
and one chemical deoxygenation by means of excess
sulfite addition to the test water, followed by reaeration
and DO data collection. About 100 mg/l Na2SC"3 as
SOf, technical grade, was used in all experiments.
It was noticed that at the low pH, one sulfite addition did
not decrease the DO level sufficiently and a second
sulfite addition was necessary. Sulfite and cobalt were
added first and then after deoxygenation of the test water
the pH was adjusted. For each consecutive reaeration
test, fresh tap water was used. All tests were run at a
cobalt ion concentration of 2 mg/l.
To determine the effect of iron and manganese on
transfer efficiency, 2 mg/l each of iron and manganese
were added to the test system.
All parameters were determined in accordance with the
procedures specified in Standard Methods (16). The metal
concentration was determined with an atomic absorption
spectrophotometer Model 120 made by Varion Techtron.
A schematic diagram of the experimental setup is shown
in Figure 1.
Results and Discussion
The Effect of Hydrogen Ion Concentration on
Sulfite Oxidation by DO
Figure 2 shows the kinetics of sulfite oxidation with DO
during a deoxygenation run at pH 6. In order to keep the
pH constant as described earlier, a mixture of KH2PO4 and
K2HP04 was added to the test water to buffer the system.
The molarity of the test water was brought to 0.0015
molar phosphate.
At pH 6, one sulfite addition, 100 mg/l as S0| and
2 mg/l cobalt brought the DO to about 4.2 mg/l from the
initial DO concentration of 8.8 mg/l within 1 min
reaction time. After 1 min, the path of sulfite oxidation
was changed dramatically and the rate decreased
substantially. Within the next 5 min, the oxygen level
was decreased from 4.2 to only 4.15 mg/l. Accordingly,
100 mg/l of additional sulfite was necessary to bring the
DO to a desired level. As shown in Figure 2, the kinetics
of the second sulfite addition during the same run seem
to be very similar to that of the first. Reoxygenation of
this test water as measured by the probe is shown on
Figure 3, which is a typical oxygenation curve. Polaro-
graphic DO measurement is not affected by the presence
of SO^ 'on as we previously reported (4).
DO was also measured by the Winkler method ten times
during the test run. The Winkler DO measurement,
however, indicated zero dissolved oxygen concentration
in all ten samples indicating the presence of unreacted
SOf ions. In the first step of the Winkler procedure.
formation of Mn02 was evident for all ten samples which
indicates the presence of DO in the sample after reaera-
tion. However, no iodine was formed during the next step
of the Winkler DO measurement by the addition of
potassium iodide followed by the acidification of the
sample. This indicates the presence of SO^ which was
acting as a reducing agent and competing with the 1 ~ ions
The polarographic DO measurements were not affected
by the presence of SOf ions as was previously men-
tioned (4). This is in agreement with the above result.
Figure 4 shows the result of chemical deoxygenation by
addition of 100 mg/l SOf at pH 7 and 2 mg/l cobalt as
a catalyst. At pH 7 the rate of catalytic sulfite oxidation
was very fast as expected.
The preceding indicates that a water system with
relatively low alkalinity needs more sulfite addition than
the recommended amount of one and one-half times the
stoichiometrical requirement for the reduction of DO
initially present (about 100 mg/l Na2SO3 as SOf).
Furthermore, the above results indicate that:
1. Catalytic sulfite oxidation with DO is pH dependent and
the rate may be increased substantially by increasing
the pH from 6 to 7.
2. Catalytic oxidation of sulfite with DO in the presence
of cobalt probably can be explained by heterogeneous
mechanism. Hydrated cobalt (II) and (III) oxide probably
act as surface active catalysts, while at a pH value of
6.5, the cobalt (III) ion is in the form of a homogeneous
catalyst.
The results of this study are in agreement with the
results of several other studies of catalytic and non-
catalytic sulfite oxidations in aqueous systems (2K3)(5)
(8X18).
To avoid unnecessary high sulfite additions and in order
to make correct DO measurements, the test water was
deoxygenated (Salt Lake City tap water has an average
pH of 8.2 to 8.3) and the pH then adjusted to the desired
value by addition of phosphate. This technique was used
in the experiments to determine the relationship between
pH and (K|_a)2o-
The Effect of Iron and Manganese on the Overall
Mass Transfer Coefficient
Several experiments were conducted to determine the
effects of iron and magnanese on the overall mass
transfer coefficient. These two transition elements are
natural tap water constituents, and the concentrations of
both differ in tap water samples from different sources.
The results of these studies using both methods of
deoxygenation and measuring the DO with both conven-
tional methods are shown in Tables 2 and 3.
94
-------
DO Meter
Multipoint
Recorder
Figure 1. Equipment Arrangement
10"
\i
Air
or
N2
-DO Probe with Mixer
and Sample Pump
Baffles
D Sample Pump
Location
t pH Probe Location
# Cobalt and Sulfite
Addition
12"
,-—Acid
\ Addition
\Location
\
Mist Shroud
95
-------
Figure 2. Chemical Deoxygenation at pH Level of 6.0
u>
1 7
.1
c
0)
o
u
c
0)
I
O
1
o
M
b
First addition of
100 mg/l of Na2S03
Test Conditions
1. 400 rpm
2. (0) zero scfm
3. 2.0 mg/l Co++ concentration
4. 0.0015 mole/l phosphate
buffered system
-Second addition of
100 mg/l Na2S03
i i
i i i i i i i i i i i i i i
I I I I
789
Time (min)
10 11 12 13 14 15
Figure 3. Reaeration of Test Water at pH 6.0 After Two Additions of Sodium Sulfite
=• 8
o>
E
c
o
E 6
8
o
o
0>
O)
>.
X
O
•Reareation data taken
from DO probe recording
Test Conditions
1. 400 rpm
2. 8.59 scfm
3. 2.0 mg/l Co++
concentration
4. 0.0015 mole/I phosphate
buffered system
in
Q
0
i i i i
i i i i i i
6789
Time (min)
10 11 12 13 14
96
-------
Figure 4. Consecutive Deoxygenation and Reaeration at pH 7.0 in a Phosphate Buffered System
Chemical deoxygenation
data taken from DO
probe recording
Reaeration data taken
from probe recording
Reaeration data taken
from coincident Winkler
sampling
Test Conditions
1. 400 rpm
2. 8.62 scfm
3. 2.0 mg/l Co++ concentration
4. 0.0015 mole/I phosphate
buffered system
i i i i i
6789
Time (min)
10 11 12 13
14
Hgure 5. The Effect of pH and Buffering Capacity of Test Water on Overall Mass
Transfer Coefficient Using Two Conventional Deoxygenation Processes;
DO Measurements Taken with Winkler Titrations
34
32
30
28
\- 26
J 24
22
20
18
16
14
Legend
1. Chemical Deoxygenation
O Phosphate buffered
system
• Not artificially
buffered
2. Physical Deoxygenation
Q Phosphate buffered
system
| Not artificially
buffered
,-1 1 L_
10
pH
97
-------
Table 2. The Effect of Various Metals on Oxygen Transfer Efficiency at pH 7
Additive
(Catalyst)
Nons
2 mg/l Co**
2 mg/l Co** &
2 mg/l Fe++
2 mg/Co++ &
2 mg/l Fe++
2 mg/l Mn++
2 mg/l Co++ &
2 mg/l Mn++
2 mg/l Mn++ &
2 mg/l Fe++
2 mg/l Co++++ &
2 mg/l Mn &
2 mg/l Fe++
Table 3. Increase in
Method of
Deoxygenation
N2
N2
Sulfite
Sulfite
N2
N2
Sulfite
Sulfite
N2
N2
Sulfite
Sulfite
N2
N2
S03
S03
Method of
Oxygen
Determination
Polarographic
Winkler
Polarographic
Winkler
Polarographic
Winkler
Polarographic
Winkler
Polarographic
Winkler
Polarographic
Winkler
Polarographic
Winkler
Polarographic
Winkler
(KLa)20 by Addition of 2 mg/l Initial
hp
0.76
0.76
0.73
0.73
0.73
0.73
0.73
0.73
0.72
0.72
0.72
0.72
0.72
0.72
0.72
0.72
Catalyst
Air Flow
(scfm)
8.67
867
8.61
8.61
8.53
8.53
8.53
8.53
8.53
8.53
8.55
8.55
8.56
8.56
8.55
8.55
at Constant
Polarographic DO Measurement
Additive
(Catalyst)
None
2 mg/l Co"1"1-
2 mg/l Co++ &
2 mg/l Fe^
Mn++
Co++ & Mn++
Mn & Fe"1"1"
Co++ & Mn++ & Fe"1"1-
N2
(KLa)20
(hr'1)
14.26
—
20.4
18.93
—
25.37
—
Method of Oxygenation
(KLa)20
(hr-1)
14.26
19.16
13.05
17.27
20.40
16.57
19.80
17.59
18.93
22.32
18.93
26.57
25.37
18.59
33.72
34.68
pHof 7
Winkler DO
Method of
S0| N2
Relative (KLa)2o
Error (hr"1)
0.0 -
- 13.05
43.0 19.80
32.7 —
- 18.93
77.9 -
— 3372
Relative
Error
—
0.0
51.7
—
45.0
—
158.3
(KLa)20
(hr-1)
19.16
—
16.57
22.31
—
18.59
—
Relative
Error
0.0
—
-13.5
6.1
—
-3.0
—
02 Transfer
Rate
(Ib/hr)
1.19
1.57
1.09
1.42
.66
.45
.61
.45
.55
.80
1.55
2.11
2.02
1.52
2.60
2.66
Measurement
Oxygenation
S0|
(KLa)20
(hr-1)
—
17.27
17.59
—
26.57
—
34.68
OTE
(Ib/hphr)
1.56
2.06
1.50
1.95
2.26
1.97
2.13
1.97
2.15
2.49
2.93
2.15
2.81
2.11
3.70
3.61
— — — — - — — —
Relative
Error
0.0
1.9
52.0
—
100.8
98
-------
The data obtained without adding any chemicals were
used as the baseline for physical deoxygenation-reoxy-
genation data analysis while the data obtained with
2 mg/l cobalt were used as the baseline for chemical
deoxygenation-reoxygenation (Table 3). The concentration
of the metal salts was 2 mg/l as metal ion for each
component. The iron concentration of 2 mg/l and
manganese ion concentration of 2 mg/l are relatively
high and do not represent the situation for most tap
water supplies. However, these concentrations were used
in our work to quantify the effects of these transition
elements on the overall mass transfer coefficient.
The results of this study definitely demonstrate that
the addition of iron and manganese affects the overall
mass transfer and results in a range of (KLa)2o values
(Table 2 and 3). Measuring DO with a probe resulted in a
relative error in (KLa)2o as high as 158% using sulfite as
the deoxygenation method and 78% using nitrogen
stripping as the deoxygenation method. The relative error
in (KLa)20 for DO measurement with Winkler methods
for nitrogen stripping was relatively small and ranged
from -13.5% to +6.1%, while the error using sulfite
deoxygenation ranged from 2 to 101% for various rates
of sulfite addition.
It is known that various amounts of cobalt (III) ion addition
in reaeration testing affects oxygen mass transfer (9K13)
(14). Accordingly, it can be postulated that various
amounts of iron and manganese, and probably other
transitional elements, even in very low concentrations
will have different synergistic effects on oxygen mass
transfer.
The (K|_a)2o data obtained by using nitrogen stripping and
measuring DO with a probe suggest a mass transfer
enhancement by addition of iron and manganese into the
test water. Activation of DO in the presence of transi-
tional elements is the most likely explanation of this
enhancement phenomenon.
Method of DO Measurement and Deoxygenation in
Relation to the Hydrogen Ion Concentration
Gaden (6) proposed the use of nitrogen gas instead of
sodium sulfite for deoxygenation of test water. The use of
sulfite as a reducing agent is widely accepted and prac-
ticed in the field of aerator performance evaluation. No
quantitative analysis, however, is available to show the
effect of altering the pH using physical or chemical
deoxygenation methods and measuring the DO with a
probe or the Winkler method.
Figure 5 shows the relationship between pH and the
overall mass transfer for the two suggested deoxygena-
tion processes while measuring DO by the Winkler
method. At a pH value of 7 in a phosphate buffered
system, there was little difference in the (KLa)2o value
between chemical and physical deoxygenation. However,
the (KLa)20 increased substantially for nitrogen deoxy-
genation-reoxygenation below and above pH 7. (KLa>2o
for pH 5 and 9 is about 94% higher than at pH 7. The
(KLaho decreases from 33 hr'1 at pH 5 to 22 hr'1 at pH 4
(Figure 5).
Data obtained by using sulfite deoxygenation-reaeration
and measuring DO with the Winkler method indicates
that the (KLa)2o is independent of hydrogen ion concen-
tration in a pH range of 5 to 7. By increasing the pH
without phosphate addition, the value of (KLa)2o in-
creased; e.g., at a pH value of 8.4, the value of (Ki_a)2o
was 47% higher and at a pH of 9 this value was 70.5%
higher than at a pH of 7 (phosphate buffered).
If the test water was buffered by a 0.0015 molar
phosphate system to pH 8, the (K|_a)2o value remained
constant at about 17 hr"1.
Figure 6 shows the relationship of pH and overall mass
transfer using a probe for DO determination and applying
both methods of deoxygentation. The (KLa)2rj data for
both deoxygenation processes were similar. The (K|_a)20
vs pH showed a constant value of 14 hr"1 for the pH
range tested in a buffered system. The (KLa>2o value for
the non-phosphate buffered system was about 21%
higher than the (KLa)2o value for the phosphate buffered
system. This result was in agreement with the data
obtained at pH 7 in a buffered system when the system
was chemically deoxygenated and when the DO was
measured with the Winkler method. The data obtained
during this series of experiments are summarized in
Table 4.
Conclusions
An accurate quantitative prediction cannot be made of
the magnitude of cobalt interference for various natural
water systems with different water quality characteristics.
Chemical deoxygenation-reoxygenation in a phosphate
buffered system while employing the Winkler DO
measurement technique resulted in relatively precise
(KLa)2o values, especially at a pH of 6.9.
Consistent but different (KLa)2o values were obtained in
phosphate and non-phosphate buffered systems when
the systems were deoxygenated by sulfite addition or
nitrogen stripping within the pH range evaluated, when
probes were used to measure DO, and when iron and
manganese concentrations were at low background
levels.
Iron and manganese and probably other transitional
elements substantially enhance (up to 100% in this
evaluation) oxygen mass transfer from the gas phase to
the liquid phase during non-steady state clean water
testing.
The data obtained from non-steady state clean water
testing can be useful in comparison of aerator efficiencies
99
-------
Figure 6. The Effect of pH and Buffering Capacity of Test Water on Overall Mass
Transfer Coefficient Using Two Conventional Deoxygenation Processes;
DO Measurements Taken with DO Probe
o
01
22
20
18
16
14
12
10
8
6
D
o
Legend
1. Chemical Deoxygenation
O Phosphate buffered
system
• Not artifically
buffered
2. Physical Deoxygenation
D Phosphate buffered
system
| Not artifically
buffered
^M
J L
i i I i i I \ I I—L
pH
J I I I I I—I 1—I L
10
100
-------
Table 4. Summary of Consecutive Reaeration Testing at Various pH Values Using 2 mg/l Initial Co"1
Run
#
1
1
2
2
3
3
4
4
5
5
6
6
68
68
7
7
8
8
1
1
2
2
3
3
4
4
5
5
6
6
Method
of
Deoxy-
genation
N2
N2
SOg
SOf
N2
N2
SOf
SOf
N2
N2
SOf
SOf
SOf
SOf
N2
N2
SOf
SOf
N2
N2
sOg-
SOg
N2
N2
SOs
soi
N2
N2
SOf
SOf
Method of
Oxygen
Deter-
mination
Probe
Winkler
Probe
Winkler
Probe
Winkler
Probe
Winkler
Probe
Winkler
Probe
Winkler
Probe
Winkler
Probe
Winkler
Probe
Winkler
Winkler
Probe
Winkler
Probe
Winkler
Probe
Winkler
Probe
Winkler
Probe
Winkler
Probe
pH
Before
Run
7.93
7.93
8.1
8.1
6.9
6.9
6.99
6.99
5.93
5.93
6.0
6.0
6.0
6.0
4.0
4.0
4.0
4.0
9.0
9.0
8.9
8.9
8.2
8.2
8.2
8.2
5.0
5.0
5.0
5.0
pH
After
Run
8.12
8.12
812
8.12
7.11
7.11
7.04
7.04
5.99
5.99
6.01
6.01
6.09
6.09
4.0
4.0
4.0
4.0
9.0
9.0
8.9
8.9
8.2
8.2
8.4
8.4
5.2
5.2
5.2
5.2
hp
0.76
0.76
0.76
0.76
0.76
0.76
0.73
0.73
0.73
0.73
0.76
0.76
0.73
0.73
0.73
0.73
0.73
0.73
0.73
0.73
0.73
0.73
0.73
0.73
0.73
0.73
0.73
0.73
0.73
0.73
scfm
8.67
8.67
8.62
8.62
8.67
8.67
8.61
8.61
8.62
8.62
8.59
8.59
8.59
8.59
8.62
8.62
8.54
8.54
8.60
8.60
8.56
8.56
8.58
8.58
8.54
8.54
8.52
8.52
8.49
8.49
(KLa)20
(hr-1)
13.70
15.06
13.32
17.41
14.26
19.16
13.05
17.27
14.53
22.04
14.45
—
14.16
16.24
17.62
21.86
17.77
13.86
33.78
17.82
29.65
17.55
21.75
17.48
24.95
17.55
33.03
17.31
17.89
14.4
02
Transfer
Rate
(Ib/hr)
1.14
1.25
1.11
1.43
1.19
1.57
1.09
1.42
1.21
1.78
1.21
—
1.18
.34
.45
.77
.46
.16
2.60
1.46
2.32
1.44
1.76
1.44
1.99
1.44
2.55
1.42
1.47
1.20
OTE
(Ib/hphr)
1.51
1.65
1.46
1.89
1.56
2.06
1.50
1.95
1.66
2.44
1.59
—
1.62
1.84
1.98
2.42
2.00
1.58
3.54
2.00
3.16
1.97
2.40
1.96
2.71
1.97
3.47
1.94
2.00
1.64
Test
Water
Fe
Cone.
Img/l)
0.01
0.01
0.08
0.19
101
-------
only for test waters with very similar water quality
characteristics.
R ecommendations
The chemical properties of the test water should be
specified in terms of alkalinity, pH, iron and manganese.
If the Winkler DO titration method is used to determine
oxygen transfer by the non-steady state technique, test-
ing should be conducted at pH 6.9 in a 0.01 molar
phosphate buffered system (1.21 g/l NaH2PC>4 + 1.43 g/l
Na2P04) by the addition of 100 g/l Na2SC>3 as S0| and
2 mg/l cobalt chloride hexahydrate as CO"1"1". The total
iron in the test water should not exceed 0.3 mg/l, and
the total manganese should not exceed 0.05 mg/l.
Further work is necessary to quantify the enhancement
effects of iron, manganese, and other transitional
elements at concentrations less than 2 mg/l so that
appropriate corrections can be made to oxygen transfer
data obtained during testing with different natural water
systems.
Some sources of reagent grade sodium sulfite contain up
to 200 mg/l of iron and other transitional elements.
These materials should not be used to deoxygenate test
water. The iron may cause substantial oxygen transfer
enhancement after 5-10 successive deoxygenations.
Acknowledgements
The authors gratefully acknowledge Mr. 0. E. Albertson
for initiation and review of this work. We also gratefully
acknowledge Mr. Robert W. Okey, Mr. David DiGregorio
and Mr. Dan McCarthy for reviewing and for the assist-
ance in editing this paper. This paper was presented by
Mr. Dan McCarthy at the Workshop.
References
1. Aberley, R.C., G.B. Rattray, and P.P. Douglass. "Air
Diffusion Units". Journal Water Pollution Control
Federation, 46, pp. 895-910, 1974.
2. Baeckstroem, H.L.J., "Der Kettenmechanismus bie
der Autoxydation von Natriumsulfite". Z. Physik,
Chem. Abteil, B., 25. pp. 122-138, 1934.
3. Bassett, H., and W.G. Parker. 'The Oxidation of
Sulphurous Acid". Journal Chemical Society,
London, pp. 1540-1560, 1951.
4. Faust, S.D., and J.V. Hunter. "Principles and
Application of Water Chemistry". Proc. of the 4th
Rudolfs Conf., pp. 380-405, John Wiley and Sons,
Inc., New York, 1967.
5. Fuller, E.C., and R.H. Crist. "The Rate of Oxidation
of Sulfite Ions by Oxygen". Journal American
Chemical Society, 63, pp. 1644-1650, 1941.
6. Gaden, E.L. "Aeration and Oxygen Transport in
Biological Systems — Basic Considerations". In:
Biological Treatment of Sewage and Industrial
Wastes, Vol. 1, Aerobic Oxidation, Reinhold Publish-
ing Company, New Yok, 1956.
7. Holleman-Wiberg. "Lehrbuch der Anorganischen".
Chemie, Auflage, Water De Gruyter, Berlin,
pp. 71-80, 1971.
8. Jorissen, W.P. "Sauerstoffaktivierung bie der
langsamen Oxydation von Natriumsulfit". Z. Physik.
Chem., 23, pp. 667-672, 1897.
9. Kalinske, A.A., LD. Lash, and G.L Shell. "Cobalt
Interference in the Non-Steady State Clean Water
Test". Water and Sewage Works, 120, pp. 54-59,
July 1973.
10. Lakin, M.B. "Chemical Catalyst Interference in the
Winkler Titration Determination of Dissolved
Oxygen — A Method for Correction". Water Re-
search, 10, pp. 961-966, 1976.
11. Landberg, G.G., B.P. Graulich, and W.H. Kippe.
"Experimental Problems Associated with the Testing
of Surface Aeration Equipment". Water Research, 3,
pp. 445-455, 1969.
12. Morgan, P.F., and J.K. Bewtra. "Air Diffuser
Efficiencies". Journal Water Pollution Control
Federation, 32, pp. 1047-1059, 1960.
13. Naimie, H., and D. Burns. "Cobalt Interference in
the Non-Steady State Clean Water Test for the
Evaluation of Aeration Equipment-l, Causes and
Mechanisms". Water Research, 11, pp. 659-666,
1977.
14. Naimie, H., and D. Burns. "Cobalt Interference in
the Non-Steady State Clean Water Test for the
Evaluation of Aeration Equipment-ll, Occurrence and
Magnitude for Various Natural Water Systems".
Water Research, 11, pp. 667-671, 1977.
15. Peleg M. "Review on the Chemistry of Ozone in the
Treatment of Water". Water Research 10, pp. 361-
365, 1976.
16. "Standard Methods for the Examination of Water
and Wastewater". 14th Edition, American Public
Health Association, American Water Works Associa-
tion, and Water Pollution Control Federation, 1975.
102
-------
17. Technical Practice Committee-Subcommittee on
Aeration in Waste Water Treatment; "Aeration in
Waste Water Treatment", Manual of Practice No. 5,
1969. Journal Water Pollution Control Federation,
41, pp. 1863-1878, 2026-2061, 42, pp. 51-76, 1970.
18. Titoff, A. "Beitrage Zur Kenntnis der Negativen
Katalyse im Homogenen System". Z. Physik. Chem.,
45, pp. 641-683, 1930.
Discussion
Arthur G. Boon
Water Research Centre, Stevenage Laboratory
England, United Kingdom
Tests of a fine-bubble diffused-air system, conducted
the Water Research Centre's laboratory at Stevenage and
at field sites in other locations in the UK, have shown
that reproducible results can be obtained, provided that
the tests are carried out under carefully controlled
conditions (particularly with a clean tank and aerator) in
clean water with and without addition of detergent.
Changes in the hardness of the water and small changes
in concentrations of inorganic salts, including sodium
sulphate content, as a result of conducting several tests
in the same water, had little effect on the results. Addi-
tion of detergent (5 mg/l) did affect the accuracy with
which the end-point of the Winkler titration could be
detected when using starch as an indicator, but the
error was generally less than 5%.
Closure
Daniel A. McCarthy
I believe that Dr. Boon's verbal comments during the
discussion period were concerned with three primary
areas: (1) his preference for DO probes (as opposed to
Winkler titrations) as the means of monitoring dissolved
oxygen during the non-steady state sulfite reaeration test,
(2) his experience that "changes in the hardness of the
water and small changes in concentrations of inorganic
salts, including sodium sulfate content as a result of
conducting several tests in the same water, had little
effect on the results", and (3) his belief that the addition
of detergent (up to 5 mg/l) to the water used during the
non-steady state sulfite reaeration test would yield more
realistic and more useful results than are ordinarily
obtained with the standard test.
The DO probe versus Winkler titration debate has been
ongoing for several years. Advocates of probes generally
state that they yield "more precise" and "more accurate"
results than Winkler DO titrations. The "more accurate"
statement arises from the many known interferences to
the Winkler titration method. These interferences were
summarized in our paper presented at Asilomar. These
interferences can be markedly reduced by employing a
phosphate buffered system maintained at a pH of 6.9 or
less. The exact phosphate buffer requirements were
summarized in the paper.
In six replicate tests in clean water, the rate of oxygen
transfer for a fine-bubble aeration system, measured
using six dissolved-oxygen electrodes sited at different
positions in the water, varied by only 8% of the mean
value. The average value measured using the electrodes
was only 2% different from the rate measured by taking
'grab' samples from near the surface of the water and
analyzing them for dissolved oxygen using the modified
Winkler Method described by Montgomery et al. *
•Montgomery, H.A.C., N.S. Thorn, and A. Cockburn. "Determination of
Dissolved Oxygen by the Winkler Method and the Solubility of
Oxygen in Pure Water and Sea Water." Journal Applied Chemistry,
Vol. 14. p. 280. 1964.
The opponents of Winkler titration procedures do not
generally address the idiosyncratic behavior of DO probes.
During the experiments I discussed at Asilomar (Table 3)
DO probes were affected to a substantial degree by the
presence of iron and manganese contaminants in the test
water. The effects of similar iron and manganese con-
centrations on mass transfer coefficients, KLa values,
determined by Winkler techniques, were not as pro-
nounced as the effects apparent with DO probes. Results
obtained with DO probes were essentially the same as
results obtained with Winkler titrations when the liquid
system was phosphate buffered and adjusted to a pH
less than 6.9.
The precision of KLa data we have obtained with DO
probes has been similar to the precision of KLa data we
have obtained by Winkler procedures. Generally, the
standard deviation of the KLa data obtained by either
method has been equal to about 7% of the mean KLa
values obtained during bench-, pilot-, or full-scale
aeration testing. The acquisition of probe data in the
field has presented numerous difficulties. Each probe
membrane usually has to be replaced several times each
day. This procedure is time consuming and therefore
costly. Strip chart recorders are often used during field
103
-------
performance tests that require the use of probes for DO
determinations.
Failure of a probe membrane, transmission cord, DO
meter, or strip chart recorder could abort a performance
test. The probe system must be calibrated. The Winkler
procedure is employed for this purpose. In our experience
the use of probes has yielded KLa data essentially the
same as data obtained by Winkler procedures in
phosphate buffered systems maintained at a pH of 6.9 or
less. The probes have proved to be cumbersome to use in
the field. Probe system malfunctions have caused
performance tests to be aborted. These interruptions and
delays have been costly in both time and money. We
prefer to use the Winkler technique to avoid these delays.
Dr. Boon's statement concerning his experience that
"changes in the hardness of the water and small changes
in the concentrations of inorganic salts, including sodium
sulfate content, as a result of conducting several tests in
the same water, had little effect on the results" is
contrary to the experience we have had with probes and
Winkler titrations over the years. Generally, measured
values of KLa have increased as the hardness, pH, and
sulfate content of the test water have increased. An
accurate prediction of the magnitude of the increase as a
function of these factors is masked by the interactive
effects of the factors. The effects are most pronounced,
for a given test water, when five or more sulfite additions
have been introduced to the test basin.
Dr. Boon's belief that the addition of detergent (up to
5 mg/l) to the test water used during the non-steady
state sulfite reaeration test would yield more realistic and
more useful results is sure to generate debate among
persons interested in aeration test procedures. The
conventional wisdom states that the addition of detergent
to bubble aeration systems will depress measured values
of KLa while the addition of like amounts of detergent to
surface aeration systems will enhance measured values
of KLa. The effects of detergent (which one?, what
concentration?) on probe and Winkler KLa determinations
for a variety of natural water systems would have to be
quantified before I could support the inclusion of the
addition of detergent to the test water in any proposed
aeration testing standard.
104
-------
Gas Flow and
Power Measurement
Fred W. Yunt
County Sanitation Districts of Los Angeles County
Los Angeles CA 90745
Gas Flow Measurement
Gas flow measurement is a fundamental part of oxygen
transfer testing. The gas flow measurement section of
this paper will discuss some of the basic types of primary
and secondary flow elements and will attempt to show
some of the advantages and disadvantages of each.
Primary emphasis, however, will be on a complete dis-
cussion of the concentric orifice plate and its application
to air flow measurement. An attempt has been made to
draw upon as much personal experience as possible. In
particular, a special section of the paper discusses a
recent experience with blower pulsation and its effect on
air flow measurement.
Two references used in the writing of the "Gas Flow
Measurement" section of this paper deserve special
mention. L. K. Spink's Principles and Practices of Flow
Meter Engineering (13) was used very extensively in the
sections dealing with the general discussion of primary
and secondary flow elements. Also, C. F. Cusick's Flow
Meter Engineering Handbook (1) was used widely in the
sections dealing with the orifice plate. Both references
were referred to at a number of points throughout the
paper and are highly recommended for anyone interested
in making flow measurements.
Primary Flow Elements
There is a wide variety of commercially available flow
measurement devices. According to Spink (13), most of
these can be referred to as head meters, although there
are a considerable number that operate on entirely differ-
ent principles.
Head Meters
Head meters convert rate of flow into a differential
pressure, a mechanical force, or a change of head or
level (over a weir or flume). This category of meters
includes the orifice plate, Venturi tube, flow nozzle,
"Lo-Loss" measuring tube, pitot tube (and its variations),
pipe elbow tap, target meter, weir, flume, and others.
Head meters have a number of advantages (13). First of
all, they are usually simple and easily reproducible.
Secondly, flow can be accurately determined without the
need for an actual fluid flow calibration of the primary
measuring device. Flow rates can be determined by
merely knowing the dimensions of the primary device
and the properties of the flowing fluid. Furthermore, only
well established and easily maintained tolerances are
critical in head meters. For the special case of
16
orifice plates, only the pipe and orifice diameters are
critical. Other parameters such as concentricity of the
orifice in the pipe, location of the differential pressure
taps, and degree of smoothness of the pipe itself, are not
too critical. Any corrosive damage to the orifice bore can
be easily seen by removing the plate from the line. A
further advantage of head meters is that the differential
pressures they produce are very accurately reproducible
under the same set of conditions.
Head meters operate on the principle that the rate of flow
of fluid is proportional to the square root of the pressure
differential. Furthermore, for a given mass flow rate, the
reading of the head meter varies inversely as the square
root of the density. Head meters are also affected by
viscosity, relative pipe size, and velocity. The Reynolds
number, however, serves to correlate these parameters
with the proper meter coefficients.
Orifice Plates
The orifice plate is probably the most common type of
head meter. It is essentially a thin plate in the flow
through which a sharp, square-edged orifice has been
bored. The ratio of the orifice diameter to the inside pipe
diameter is known as the "diameter," "orifice," or "beta"
ratio. An orifice has a number of variations. Generally
speaking, an orifice is either concentric, eccentric, seg-
mental or annular. The concentric orifice is by far the
most widely used. The other variations were derived to
handle flows containing solids in suspension and water
vapor. Orifice plates have many advantages, including
accuracy, simplicity, cost-effectiveness, versatility, and
ruggedness. A disadvantage is that they sometimes
involve a relatively high permanent pressure loss.
Venturi Tubes
Venturi tubes are used to measure flows-when a relative-
ly low permenent pressure loss is required. Usually they
will handle about 60% more flow than an orifice plate,
but their permanent pressure loss is only 10 to 20% of
the differential pressure (1). They are also good for
measuring flows with solids in suspension. Furthermore,
for the measurement of a given air flow, they can require
less upstream straight pipe than an orifice plate. Venturi
tubes consist of a converging and diverging nozzle. The
diverging section tends to make the pressure recovery
efficient so that the permanent pressure loss is low. The
Venturi works on the basic principle that the change in
105
-------
static pressure between two regions of different cross
sectional area is proportional to flow. Static pressure
readings are taken in an upstream region of the Venturi
and at the throat. Generally speaking, Venturi meters are
relatively expensive and somewhat large but in many
situations this is more than offset by their advantages.
Flow Nozzles
Flow nozzles have advantages similar to Venturi tubes.
They consist of a converging nozzle only. While the
pressure recovery is not as efficient as with the Venturi
tubes, it is still much better than with orifice plates. They
will handle about 60% more flow than an orifice plate
with a permanent pressure loss that varies from 30 to
80% of the differential pressure, depending on the
diameter ratio (1). The principle of operation is similar to
that of the Venturi tubes. Static pressure readings are
taken in an upstream region of the flow nozzle and at
the throat. Generally speaking, flow nozzles are less
expensive than Venturi tubes but are considerably more
expensive than orifice plates.
Pilot Tubes
The pilot tube or one of its variations, the pitot-static tube
or the Annubar (a commercially available device), pro-
vides a very low permanent pressure loss. Essentially,
they are in-line pressure probes that measure velocity
head. All variations of the pitot tube measure the stagna-
tion pressure of the flow, which is equal to the static
pressure plus the velocity head. The pitot tube requires a
separate measurement of static pressure, while the pitot-
static tube has a static pressure tap in the side of the
stagnation pressure probe. In both cases, the difference
between the stagnation pressure and the static pressure
is the velocity head.
The pitot tube and the pitot-static tube are limited by the
fact that they essentially measure velocity at one point in
the flow. Normally this can be related to the average pipe
velocity if the flow conditions are steady; however, up-
stream pipe disturbances can significantly alter this
relationship. The Annubar, on the other hand, measures
an average stagnation pressure across the pipe diameter
and does a better job of estimating the average pipe
velocity. Furthermore, the Annubar differs from the pitot-
static tube in that it measures a differential pressure that
is somewhat greater than the velocity head due to a
slight suction condition created at the low pressure tap.
A disadvantage to the pitot tube, pitot-static tube, and the
Annubar is that they may sometimes require a pipe size
reduction in order to increase the velocity head since
velocity heads in normal air flow applications are usually
low.
Other Types of Meters
There are a number of other types of meters available for
flow measurement. Among these are magnetic meters,
turbine meters, variable area meters (rotameters), mass
flow meters, rotary lobe gas meters, and others. These
meters are good for specific applications and with the
exception of the rotameter, are generally more complex
and costly than the head meters. For gas flow measure-
ment, however, it would seem that only rotameters and
rotary lobe gas meters have wide application.
Rotameters
The rotameter is actually a variable orifice head meter. It
consists of a float in a vertical transparent tube. The
inside diameter of the tube is linearly tapered from top to
bottom, with the widest portion at the top. Flow enters
the tube at the bottom, and the float rises until the weight
of the float just balances the lift generated by the
buoyancy of the float and the pressure differential across
it. The headloss through the rotameter is essentially
constant and usually fairly low. Furthermore, the rota-
meter scale is nearly linear with flow. The combination of
these factors makes it exceptionally desirable for measur-
ing flows that vary over a wide range (i.e., 10 to 1).
Furthermore, the rotameter is not affected significantly by
upstream piping disturbances; no straight lengths of pipe
are generally required. It is especially suitable for pipe
sizes up to 3 in. (14).
Rotary Lobe Gas Meters
The rotary lobe gas meter is a positive displacement
device. It is highly recommended for measuring the flows
from positive displacement blowers and other pulsating
equipment because it is not affected by pulsation as are
the head meters. Volume flow rate is determined by
measuring the rpm developed as the air passes through
the meter. Rotary lobe gas meters are usually expensive,
but where high accuracy is required, they are excep-
tionally good.
In summary, there are a large number of accurate gas
flow measurement devices. Each has its own advantages
and disadvantages, and one may be better suited to a
particular application than another. Holman (8) shows a
comparison of the operating characteristics of various
flowmeters, according to Miesse and Curth. For a more
detailed analysis of the various primary flow elements,
the reader is referred to Spink (13).
Secondary Flow Elements
Secondary flow elements are used to measure the output
of the primary flow elements. For head-type primary flow
elements, this output is usually a differential pressure.
While there are a wide variety of differential pressure
measuring devices, they can be generally classified into
two basic groups according to whether or not the fluid
which exerts the differential pressure is in direct contact
with an indicating fluid (13).
Wet Meters
According to Spink (13), wet meters are devices where
106
-------
the fluid which exerts the differential pressure is in direct
contact with the indicating fluid. This group of meters
can be further divided into the liquid and liquid seal
varieties.
Liquid Meters
With liquid meters, a difference in liquid level is devel-
oped between the high and low pressure sides of the
device. Liquid meters are the oldest, simplest, and
generally speaking, the most accurate and reliable of all
differential pressure measuring devices. They are excel-
lent in applications where only visual indication is needed
and where the static pressures and the nature of the
fluid are conducive to transparent tubes. Various types
of liquid manometers are available, including the U tube,
well type, and mercury float type. Of these, the U tube
manometer is the simplest, the differential pressure
being the distance between the water columns in each
leg. Well manometers are similar to U tube manometers
except that the high pressure leg of the device is actually
a large reservoir. Since the volume of the reservoir is
very large relative to the volume of the manometer
indicator tube, it is only necessary to read one leg of the
manometer; the fluid level in the reservoir does not
change appreciably. An important variation on the well
type manometer is the inclined well-type manometer.
Inclining the manometer helps to expand the indicator
scale so that differential pressure can be read more
accurately. The mercury float type of liquid manometer is
similar to the other liquid meters except that a float
attached to an indicating needle follows the mercury
surface on the high pressure side. This type of meter is
very accurate, although relatively expensive, and for
many years it was the accepted standard meter for orifice
flow measurements.
Liquid Seal Meters
With liquid seal meters, the differential pressure is deter-
mined by the displacement of a piston-like device, across
which the differential is applied and which acts against
the tension of a spring. The meter fluid acts like a seal
between the high and low pressure sides of the device.
An indicating needle senses the position of the piston.
Liquid seal meters are well suited to the measurement of
low differential pressures. Generally speaking, they are
used where simple, self powered mechanisms with con-
siderable output power are required. Varieties of this type
of meter include the cylindrical inverted bell meter, the
Ledoux inverted bell meter, and the ring balance meter
(similar in principle to the bell-type meters).
Dry Meters
According to Spink (13), dry meters are devices where
the fluid which exerts the differential pressure is not in
contact with an indicating fluid. Dry meters are becoming
increasingly popular wherever indication, recording, inte-
gration, and/or transmission are required. They are
relatively inexpensive compared to the mercury float-type
meter. Dry meters are of two basic types: motion and
force balance.
Motion-Type Meters
The motion-type dry meter consists of a bellows system
across which the differential pressure is applied; the
resulting force deflects a spring,and the motion provides
the output. This type of meter is well suited to direct
indication, recording, and integration.
Force Balance-Type Meters
The force balance-type of meter employs a diaphragm,
against which the differential pressure is applied. An
electric current or pneumatic pressure develops a second
force, and the two forces are made to oppose each other.
A sensitive motion detector detects any imbalance in the
forces and corrects the second force until they are
balanced. The magnitude of the current or pressure of
the second force is then directly proportional to differen-
tial pressure. This type of meter is used primarily as a
pressure transmitter to a remote location.
As with primary devices, the selection of a specific
secondary device for a given application depends on a
number of factors. Generally speaking, where visual
indication is sufficient, the liquid manometers are suitable
as they are simple, low cost, rugged, and accurate. Other
needs may dictate a requirement for other types. For a
much more detailed discussion of various secondary
devices and their specific application, the reader is
referred to Spink (13).
Required Measurements for Gas Flow Determinations
A number of measurements are required for accurate gas
flow determinations. As a minimum, the following read-
ings should be taken:
1. differential pressure (with most head meters), h
2. flowing gas temperature, Tf
3. flowing gas pressure (static pressure), pf
4. ambient temperature, Ta
5. ambient pressure, pa
6. ambient relative humidity, RH
Other readings may be necessary on specific types of
flow meters.
The differential and static pressure readings should be
made with a manometer of some type. Unless flow totali-
zation or recording is required, a simple U tube or well-
type manometer works very well. Pressure gauges can be
used for static pressure determinations, but experience
has shown that unless these gauges are of the oil-filled
variety or are at least protected with a pulsation damp-
ener (snubber), they can be rendered useless in a short
period of time by vibration and pressure oscillation. A
mercury-filled manometer is more rugged and accurate.
107
-------
The taps for both the differential and static pressure
readings should be constructed in a similar fashion to the
tap shown in Figure 1 (see also the "Orifice Plate Con-
struction and Installation" section of this paper).
Flowing gas temperature can be read with a bimetallic
dial thermometer that is mounted permanently in the
pipe. These thermometers can be bought with probes of
varying lengths for different pipe sizes. Where electronic
determinations are required, as in temperature compen-
sation circuits, thermistors can be employed. In any case,
the probe should extend well into the center of the pipe.
Ambient temperature can be determined with any
number of commercially available wall-mounted ther-
mometers. Care should be taken to obtain these readings
in the shade as direct sunlight will cause erroneously
high readings due to the radiant energy of the sun.
Ambient pressure should be determined with a good
quality mercurial barometer. These readings should also
be taken in a shaded area and corrections should be
made for the thermal expansion of the brass indicating
scale, the change in mercury density with temperature,
and the effect of the earth's gravitational field. Tables for
making these corrections are usually included with the
barometer.
Relative humidity readings should be taken with a
hygrometer or a psychrometer. The sling psychrometer,
in particular, is portable and inexpensive. Usually, wet
and dry bulb thermometer readings are taken and are
converted to relative humidity with the use of tables that
accompany the device. Some hygrometers are available
which read relative humidity directly.
Standard Conditions
Since gases are compressible fluids, and in some
instances may contain varying amounts of water vapor,
the ratio of mass flow to volumetric flow is not a
constant, but depends on the temperature, pressure and
water vapor partial pressure. The need to relate mass
flow to a volumetric flow by a constant, as in incompres-
sible flow, gave rise to the term "scfm," standard cubic
feet per minute. This refers to an imaginary volumetric
flow rate; it is that flow which would exist if the same
mass flow of dry gas existed at standard conditions of
temperature, pressure, and relative humidity. This volu-
metric flow rate can be multiplied by a constant factor to
obtain the mass flow of dry gas, or by a different constant
factor to obtain the mass flow of gas at standard humidity.
Standard conditions for air flow measurement are usually
taken as 68°F (20°C, 528°R), 14.70 psia (1 atm, 760
mm Hg) and 36% relative humidity. Standard temperature
is sometimes taken as 60°F, particularly by the natural
gas industry. Blower manufacturers and others, on the
other hand, sometimes use 70°F. It is felt that 68°F is
the most commonly used temperature standard in the
oxygen transfer field, and it has been selected for use in
this paper. It should also be pointed out that for gases
other than air, the relative humidity standard is zero
rather than 36%.
Orifice Plate Discussion
It would be virtually impossible to discuss all the various
types of flow meters in detail in the context of a single
paper. With this in mind, a decision has been made to
concentrate on the application of the concentric orifice
plate. It is certainly one of the most widely used primary
flow elements due to its accuracy, simplicity, cost-effec-
tiveness, versatility, and durability. The information
contained in the sections which follow should provide
guidance for the design, installation, and operation of a
successful orifice meter. Much of this information is also
applicable to head meters in general.
Basic Orifice Plate Equation for Gas Flow
Based on Cusick (1), the basic orifice plate equation for
gas flow, without refinements of any kind, can be written
as follows:
aPProx.(b)
.44/60)S0d2(Tb/Pb) /(Pf/Tf)(h/G)
in which:
Qapprox .(b) = approximate gas flow at base conditions of
Tb and Pb, cfm
S0 = orifice coefficient
d = actual internal pipe diameter, in.
Tb = base temperature, °R
pb = base pressure, psia
Tf = flowing gas temperature, °R
Pf = flowing gas pressure, psia
h = differential pressure, in. H20 (dry)
G = dry gas specific gravity
At base conditions of 68°F and 14.70 psia, this expres-
sion reduces to:
Qapprox.(s| = 130.77 S0d2 /(Pf/Tf)(h/G)
in which:
. (s) = approximate gas flow at base conditions of
68°F and 14.70 psia, cfm
The exact form of the orifice plate equation will be dis-
cussed in a later section of this paper. For preliminary
design of an orifice meter, however, all that is usually
required is the basic orifice plate equation.
108
-------
Location of the Orifice Plate
One of the most important factors in obtaining accurate
air flow measurements with any head meter is to install
the primary element in the proper location. The most
refined head meter equation will not compensate for an
improperly located primary device. A number of refer-
ences (1) (2) (13) are available which give the recom-
mended piping requirements before and after an orifice
plate for various types of pipe and fitting configurations
and for various diameter ratios. These data are usually
appropriate for flow nozzles as well. No attempt has been
made to reproduce this material in this paper; however,
certain general comments can be made.
First of all, it is good practice to install the primary
element with as much upstream straight pipe as possible.
Upstream fittings and curved pipe sections tend to upset
the straight flow conditions that are required for good
flow measurements with head meters. In particular,
upstream throttled valves and configurations that tend to
produce swirling gas flows require a particularly long
length of upstream straight pipe ahead of the orifice
meter. An example of a type of pipe configuration that
leads to swirling is a series of 90° bends that are out of
plane. Downstream piping conditions are important but
are far less critical than upstream conditions. It is
important to note that the required length of straight pipe
for a given piping configuration always increases with
the diameter ratio. This means that where adequate
length of straight pipe is at a premium, it is a good idea
to design the orifice plate with a smaller diameter ratio.
Of particular concern, due to recent experience with this
problem, is locating the orifice plate so that it will not be
affected by pulsation. Pulsation, as used here, refers to
the high frequency pressure oscillations that are pro-
duced by positive displacement blowers, reciprocating
compressors, and other similar devices (see the "Pulsa-
tion" section of this paper). If pulsation exists at the point
of flow measurement, although it may not be directly
obvious, it can cause substantial error in the flow deter-
mination. While there is probably no definite design
criteria as yet to ensure that pulsation will not occur in a
given situation, it is a good idea to follow certain basic
guidelines. First of all, the orifice plate should be located
as far away from the source of the disturbance (positive
displacement blower, etc.) as possible. The combination
of pipe volume and headless between the source and the
primary element tends to damp out the high frequency
oscillations (13). It may be desirable to install a large
reservoir or an artificial headloss upstream of the orifice
plate for this purpose. Recently a reservoir was installed
in the blower discharge piping for the Los Angeles County
Sanitation Districts' aeration equipment evaluation program
This has corrected a very substantial pulsation problem.
It is also good practice to isolate the orifice plate as much
as possible from the low frequency pressure oscillations
that may occur as a result of water level variations in an
aeration tank due to mixing. Again, pipe volume or head-
loss between the primary element and the air release
point can help alleviate this problem.
Another point of concern when locating an orifice plate is
installing it at a point where there is no danger of water
infiltration. Water from an aeration tank or other source
can get into the manometer lines and make the differen-
tial pressure readings completely inaccurate. Adequate
precautions should be taken to ensure that this does
not occur.
Line Size Determination
One of the first requirements in selecting the proper line
size is to know the range of flow rates to be encountered.
It should be kept in mind that a given orifice installation
will usually be able to handle only a 4 or 5 to 1 range in
flow rates. This is due primarily to accuracy limitations
imposed by the secondary device as to readability at low
flows and to permanent headlosses associated with the
primary device at high flows. The Reynolds number at
low flows may also be a problem, however. In any case,
if a wide range of flow rates is to be encountered, it is a
good idea to consider using several pipe lines of different
sizes. It is also possible, however, to use different orifice
plates in the same line size, but this can be done only up
to a point. Eventually, the diameter ratio may fall outside
the limits mentioned earlier in the section on orifice plate
taps or the orifice plate and/or pipe headlosses may
become excessive. Furthermore, it is possible that the
pipe Reynolds numbers at low flows may fall out of the
turbulent flow regime. If the flow is not turbulent, the
orifice plate coefficients may vary significantly with
Reynolds number; this is undesirable.
The line size determination should be based initially on
considerations of pipe headloss. Care should be taken so
that pipe losses at the highest air flows are not too high.
A number of references describe the estimation of pipe
and fitting headloss (3) (6) (9) (10). Maximum headloss
should be estimated using the worst conditions of
temperature and pressure. It should be remembered
when determining the line size that there will be an
additional headloss due to the orifice plate. A number of
references (1)(2)(13) discuss the estimation of orifice
plate headlosses.
After a tentative selection of line size has been made, the
pipe Reynolds number at low flow should be checked.
According to Cusick (1), the pipe Reynolds number can be
calculated from the following formula:
(3)
in which:
Re = 28.50 (QdgsG)/(dM)
Re = pipe Reynolds number
Qdgs = dry gas flow at 68°F and 14.70 psia, cfm
109
-------
M = gas absolute viscosity, cps
For air, this expression reduces to:
(4)
Re = 28.50
For most Reynolds number considerations, it is not
usually necessary to consider the effects of relative
humidity. The turbulent flow regime is sometimes con-
sidered to exist above pipe Reynolds numbers of 10,000,
and it is generally desirable to exceed this number at low
flow. To aid in the calculation of Reynolds numbers, the
following expression for absolute viscosity can be
used (10):
(5)
M= (161+0.28 Tf)x10"4,cps
Selection of Orifice Plate Taps
There are a number of different types of orifice plate
taps. The most widely used are flange taps. These are
located 1 in. from the upstream face of the orifice plate
and 1 in. from the downstream face of the orifice plate.
Normally, special flanges are purchased with these taps
already in place; this makes the orifice plate installation
considerably easier. Furthermore, since the taps are
symetrically located, the orifice may be used to read
reversing flows (13). For flange taps on pipe sizes 4 in.
and above, the diameter ratio must be between 0.10 and
0.75; on pipe sizes 31/2 in. and smaller, the diameter ratio
must be between 0.20 and 0.70 (1). Flange taps should
not be used on pipe sizes smaller than 2 in. (13). The
coefficients for concentric orifice plates with flange taps
have a tolerance of ±0.55% for diameter ratios from 0.20
to 0.70 on pipe sizes 2 in. and larger (1). The term toler-
ance, as used here, is equivalent to twice the standard
deviation and is the maximum error that would be
experienced in 95% of all installations (13). Extensive
tables of flange tap coefficients can be found in refer-
ences (1) and (7).
Vena contracta taps are also widely used. These taps are
nominally located 1 pipe diameter upstream and 1/2
pipe diameter downstream. The actual downstream
location varies with the diameter ratio. The exact location
of the vena contracta is as shown by Cusick (1). Cusick
also shows the correction factors for variations from the
exact location. On 4-in. pipe sizes and over, vena con-
tracta taps can be used as long as the diameter ratio is
between 0.10 and 0.80 (1). Vena contracta taps should
not be used on pipe sizes smaller than 4 in. due to the
interference between the downstream flange and the
downstream tap. The coefficients for concentric orifice
plates with vena contracta taps have a tolerance of
±0.50% for diameter ratios between 0.20 and 0.70 (1).
Special orifice plate flanges are not required when using
vena contracta taps; instead, standard pipe flanges are
used and the taps are installed in the pipe wall. Extensive
tables of vena contracta coefficients can be found in
references (1) and (7).
Corner taps are used in place of flange taps on pipe sizes
smaller than 2 in. These taps are located in the corner
between the pipe wall and the edge of the orifice plate in
both the upstream and downstream locations. Using
flange taps on pipe sizes smaller than 2 in. with high
diameter ratios would place the downstream pressure tap
downstream of the vena contracta. The region down-
stream of the vena contracta is a very turbulent zone for
which standard coefficients do not apply. Special flanges
or orifice holding rings can be purchased when corner
taps are used. Sometimes accuracy can be improved by
modifying the corner tap method such that the upstream
tap is located one diameter upstream of the orifice
plate (13).
Radius taps are very similar to vena contracta taps and
are located at exactly 1 pipe diameter upstream and 1/2
pipe diameter downstream from the inlet face of the
orifice plate.
Pipe or full flow taps are located 2Vi pipe diameters
upstream and 8 pipe diameters downstream of the inlet
edge of the orifice plate. Flow measurement with pipe
taps is not as accurate as with flange or vena contracta
taps with the same orifice plate. The probable error of
measurement is about 50% greater (13). Furthermore,
pipe taps require greater lengths of straight pipe. They
do, however, permit the measurement of flows in pipe
sizes smaller than 2 in. using standard pipe flanges. The
diameter ratio limits for pipe tapes are the same as for
flange taps (1).
In summary, for orifice plate measurements on pipe sizes
up to 2 in., either corner taps or full flow taps would be
the method of choice. On pipe sizes between 2 in. and
4 in., flange taps should probably be used. On pipe sizes
of 4 in. and larger, either flange or vena contracta taps
should probably be used.
Selection of Differential Pressure Range
Before an orifice plate can be designed, it is first
necessary to determine the desired range of differential
pressures for the given range of gas flow rates. Generally
speaking, it is good practice to keep the differential pres-
sure from becoming either too high or too low. Very low
differential pressure readings may make the flow deter-
mination less accurate since the error in reading the
manometer may be very significant in terms of flow.
Furthermore, if there are any pulsation effects in the air
line, low differential readings will give a proportionally
greater error in flow rate determination. On the other
hand, very high differential pressure readings can result
in very high permanent headlosses associated with the
primary device. While there is a wide range of acceptable
differential pressures, unless there is a specific reason
why a particular range should be selected, it is a good
rule of thumb to let the differential in inches of H2O equal
the static pressure in psia (13). This might normally be
done at average flow conditions as long as the range of
110
-------
flows is not too great. A check should then be made to
ensure that the error involved in reading the manometer
at low flows and the permanent headless from the orifice
plate at high flows are acceptable.
Diameter Ratio Determination
Once tentative decisions have been made concerning
desired pipe size and orifice plate differentials, it is
possible to calculate an orifice plate coefficient, S0. By
rearranging the basic orifice plate equation for gas flow
at standard conditions (equation 2), the following expres-
sion for S0 is obtained:
Fvc = vena contracts correction factor (for vena
(6) S0 = (7.647 x 10-3) (Qapprox.(s)/d2)V(T,/PfHG/h)
50 can be calculated by using the average Q, the corres-
ponding h, the actual d, and estimated values for T, and
pf. The next step is to calculate the pipe Reynolds
number for these condition by using Equation 4. The
reader is referred at this point to orifice plate coefficient
tables for the type of orifice plate taps employed (1)(7). By
referring to the proper pipe size at the proper pipe
Reynolds number, it will be possible to determine the
diameter ratio that corresponds to the calculated orifice
plate coefficient.
At this point, it is important to verify that the diameter
ratio determined is within the range of diameter ratios
for which the reported tolerances apply (see the "Selec-
tion of Orifice Plate Taps" section of this paper). Further-
more, the diameter ratio so calculated should be
compatible with the existing lengths of straight pipe for
the proposed air measurement system. Also, the
permanent headloss associated with the orifice plate at
maximum flow should be acceptable. If the diameter ratio
does not satisfy all three of these conditions, then it will
be necessary to modify the differential range, the line
size, or any of the other variables over which there is
control, until it does.
Exact Orifice Plate Equation for Gas Flow
The exact orifice plate equation for gas flow is:
(7) Qeb =
in which:
approx.(b) ' ^a ^pe ' wv ^m ^vc ^pv
Qeb = exact gas flow at base conditions of T&, pb,
and (RH)b (the relative humidity at base
conditions)
Y = gas expansion factor
Fa = orifice area correction factor
Fpe = pipe expansion correction factor
FWtf = relative humidity correction factor
Fm = manometer correction factor
pv
contracta taps only)
= super-compressiblility correction factor
For the special case of air at standard base conditions of
68°F, 14.70 psia, and 36% relative humidity, Qeb = Qa =
the air flow at standard conditions. An explanation of the
factors in Equation 7 is given in the discussion which
follows:
Gas Expansion Factor
The gas expansion factor, Y, accounts for the expansion
of the gas as it passes through the orifice. It is most
significant in low pressure systems with high differential
pressures. According to Cusick (1), the expansion factor
for flange and vena contracta taps can be expressed by
the following equation:
(8) Y = 1 -[0.41 +0.35(d0/d)4](x/k)
in which:
d0 - orifice bore diameter, in.
x = (differential in psi)/(upstream flowing
pressure in psia)
k = ratio of gas specific heats, Cp/Cv
(use 1.395 for air)
Cp = specific heat at constant pressure
Cv = specific heat at constant volume
Orifice Area Correction Factor
The orifice area correction factor, Fa, compensates for the
thermal expansion of the primary element at tempera-
tures other than 68°F (1) (13). The following relationship
has been developed from data derived from Spink (13)
and is applicable to Type 304 or 316 stainless steel:
(9)
Fa = (1.699* 10'5)T, + 0.9910
This equation is valid over the range of temperature to
be encountered in normal aeration practice.
Pipe Expansion Correction Factor
The pipe expansion correction factor, Fpe, compensates
for the thermal expansion or contraction of the pipe at
temperatures other than 68°F. The following relationship
has been developed from data derived from Spink (13)
and is applicable to steel pipe:
(10)
in which:
Fpe = (1.333 *1Cr5)Tp + 0.9930
Tp = pipe temperature, °R
111
-------
For an approximate calculation, it may be reasonable to
assume that Tp = Tf. This equation is also valid over the
range of temperatures to be encountered in normal
aeration practice
Relative Humidity Correction Factor
The purpose of the relative humidity correction factor,
Fwv, is two-fold (1) (13). First, it corrects the specific
gravity in the flow equation from that of dry gas to that of
the actual wet gas. Second, it converts the actual wet
gas flow obtained into an equivalent gas flow at base
humidity. This means that the mass flow rate of dry gas
is the same in both cases. The expression for Fwv is:
(11) Fwv =[1-(pw/pf)]/|[1-(pbw/Pb)]/l+
-------
correction factor can then be read directly from this curve
for the given downstream tap location.
Super-Compressibility Correction Factor
The super-compressibility correction factor, Fpv, corrects
for deviations from the ideal gas law, particularly at high
pressures (i.e., 100 psi) (1) (13). It is usually insignificant
at the low pressures encountered in most oxygen
transfer applications.
The correction factors just discussed, together with the
basic orifice plate equation, constitute the exact orifice
plate equation for gas flows. Depending on actual condi-
tions, some of these factors may be insignificant and may
be disregarded without appreciable loss of accuracy. In
any case, after the initial orifice plate design with the
basic equation, a check should be made using the exact
equation to ensure that the differential pressures for the
various flows to be encountered are still acceptable.
Furthermore, the exact equation should be used with the
extreme flowing and ambient conditions to determine the
maximum differential pressure to be encountered. This
will help in the selection of a manometer with an
adequate differential pressure range. Where maximum
accuracy is required, air flow determinations should be
made using the exact form of the equation.
Graphs and Computerization of Flow Equations
Until recently, calculating gas flow rates from measure-
ments taken in the field could be somewhat laborious.
Now, with the advent of programmable calculators, this
task is much easier. The gas flow equation in its most
exact form can be programmed into the calculator and
readings can be converted into flows in a matter of
seconds.
A particular advantage of programming the gas flow
equation is that the effect of the S0 variation with
Reynolds number can be taken into account. Normally, it is
possible to program the S0 variation into the calculator.
When a gas flow determination is required, an initial
assumption for S0 is made which results in an approxi-
mate air flow. This flow is then used to calculate a pipe
Reynolds number, which is in turn used to calulate a
more accurate S0 Normally, this value of S0 is very
close to being exact.
It is still very useful to make graphs of air flow versus
differential pressure for various flowing temperatures and
pressures. These graphs should be nearly straight lines
when plotted on log-log paper. They find particular appli-
cation for approximate air flow determinations in the
field.
Orifice Plate Construction and Installation
The orifice plate is usually made out of Type 316 or 304
stainless steel. Its construction is basically very simple;
however, certain guidelines should be strictly adhered to.
Spink (13) has the following to say about the construction
of the sharp, square edged, thin-plate concentric orifice
plate:
"In order to use the published coefficient data within
the standard tolerances, the orifice should be made to
the following specifications in which d0 = orifice
diameter and d = pipe diameter:
1. The thickness should not exceed any of the following
limits: d0/8, d/50, or (d-d0)/8, in the cylindrical por-
tion. If a thicker plate is required for rigidity, the
outlet face may be beveled or recessed to attain the
desired dimension. This should be done in such a way
that a straightedge laid across the bevel or recess will
form an angle not less than 45° to the axis of the pipe.
2. The upstream edge should be square and as sharp as
possible. Any rounding should not exceed 0.025% of
the diameter of the orifice to assure measurement
with in 0.1%.
3. The upstream face should be at least as smooth as
good commercial rolled stock.
4. The portion of the plate which extends inside the pipe
should be flat within 0.01 in. per in. of radius.
5. The orifice plate should be centered in the pipe in
such a way that the eccentricity is less than 3% of the
pipe diameter."
The orifice plate should be equipped with a tang plate as
shown in Figure 2. This plate can be stamped with the
proper diameter ratio and line size and it also serves as a
handle during installation and removal. The orifice plate
should also be equipped with a weep hole (1). For air
flow measurement, the weep hole should be located on
the bottom of the orifice plate; this allows any condensa-
tion which forms in the line to pass the plate. The exact
weep hole location and the recommended weep hole
sizes are shown in Figure 2. Normally, the error involved
in flow measurements due to the installation of a weep
hole is extremely small; however, for small orifice bores
it may be desirable to correct the flow measurement
accordingly. Weep holes should not be used when the
orifice bore is 7/8 in. or smaller.
Care should be taken so that the orifice plate flanges are
welded perpendicular to the pipe. Nipples for vena
contracta pressure taps and static pressure taps should
be welded on the outside of the pipe. They should be
perpendicular to the pipe wall, and should be at least 2'/2
times the diameter of the tap hole in length (see Figure
1). The diameter of the tap holes may vary but should
probably not be smaller than 1/8 in. nor larger than
1 /4 in. The tap hole should be drilled with the nipple as
a guide after it has been welded in place. Any burrs
around the hole on the inside of the pipe should be
removed so that the surface is smooth.
As far as static pressure taps are concerned, they are
usually located upstream of the orifice plate. Some
113
-------
Figure 1. Pipe Wall Pressure Taps
(After Spink)
Nipple Threaded
at One End
'£
*
fM
K-
k 90° and Radial
1
Drill Pipe After-
Welding Pipe
-Remove Burrs and
Round Slightly
Figure 2. Orifice Plate Dimensions
(After Cusick)
Orifice Bore as
Calculated
Dia. "B'
A = diameter of flange bolt circle minus diameter of
flange bolt hole
L = at least 3 in. for pipe sizes up to 3 in.
L = at least 5 in. for pipe sizes above 3-1/2 in.
H = inside diameter of pipe minus "B"/2
Pipe Size (in.) Dia. "B" (in.)
1-1/2 to 4 1/16
5 & 6 3/32
8, 10 & 12 1/8
14 to 20 5/32
Drill only one hole: upper for liquid flow, lower for gas or vapor flow. If the orifice bore is 7/8 in. or smaller,
do not drill Hole "B".
114
-------
expansion factor expressions, however, are in terms of a
downstream static pressure or an average upstream and
downstream static pressure. It is important that the static
pressure tap(s) be located in accordance with the expan-
sion factor expression. In the case of upstream static
pressure taps, it may be advisable to locate the tap at
least 4 pipe diameters upstream of the orifice plate to
avoid the compression region in front of the plate. It
should not be located so far in front of the plate, how-
ever, that there are significant pipe losses between the
tap and the plate.
The orifice plate should be installed with two rubber
gaskets, and care should be taken to center the orifice
plate as much as possible. The bolts around the plate
should be tightened up evenly.
Selection of a Secondary Flow Element!s)
To help in the proper selection of a secondary flow
element, the reader is referred to the section of this
paper on "Secondary Flow Elements." Certain basic
guidelines should be followed, however. First of all, a
secondary meter should always be sized to measure the
actual differential, not the high and low pressures
separately followed by an arithmetic subtraction. The
latter method usually results in a substantial loss of
accuracy.
Furthermore, it may often be appropriate to have several
secondary elements with different differential pressure
ranges available for use with a given primary element,
since it is often the readability of the secondary element
that limits the flow range of the overall gas flow
measurement system.
Finally, it is usually appropriate with manometers to use
either water or mercury for manometer fluid when
making field differential pressure measurements. Fluids
which have a specific gravity close but not equal to that
of water should be avoided, since the possibility of con-
tamination with water always exists in the field. Mercury,
on the other hand, has a specific gravity that is so much
different from that of water, that minor contamination
with water does not affect the readings significantly.
Connections Between Primary and Secondary Flow
Elements
A number of factors should be considered when making
differential pressure connections between primary and
secondary flow elements In particular:
1. The manometer should be located in a shaded area as
close to the primary element as possible.
2. For gas flow, the manometer should be mounted above
the air line so that the tubing slopes at least 1 in. per
ft. This allows any condensation in the tubing to roll
back into the air line.
3. For distances less than 50 ft, the tubing between the
primary and secondary elements should be at least
3/16 in. inside diameter. For distances greater than
50 ft, the diameter should be increased by approxi-
mately 1/8 in. for each additional 50 ft.
4. For greatest accuracy, the tubing on both sides of the
head meter should be of the same length and should
be in the same thermal environment.
The tubing used to connect the primary and secondary
flow elements may be made of a number of commercially
available materials. Copper tubing or pipe is recommend-
ed for permanent installations. Materials such as rubber,
polyethylene and vinyl are fine for less permanent instal-
lations and are more amenable to changes in the way
the system is connected. Special care should be exercised
when using hard-walled tubing such as polyethylene in
an outside environment, since low temperatures may
cause the tubing to crack.
Valves should be placed on both primary and secondary
elements. Needle valves and other restrictive valves such
as plug and globe valves are not recommended. "Equal-
izer" arrangements and check valves should be used on
manometers where a possibility of excessive differential
pressure exists. The equalizer is merely a cross-connec-
tion between the high and low pressure sides of the
manometer with a manual shut-off valve. Lever-operated,
1 /4-turn gas valves are very convenient for this applica-
tion. Generally speaking, the equalizer valve is left open
unless a reading is taken. This protects the fluid in the
manometer against over-pressure. Some manometers,
such as those for measuring line pressure, may not
require either equalizers or check valves since the
possibility of over-pressure may not exist.
A number of tubing connectors are available. For perma-
nent installations, double-ferruled connectors are
recommended. These can be used on any hard-walled
tubing and can be used on soft-walled tubing with the
proper tubing insert. Fittings can be bought which will
adapt to various types and diameters of tubing and pipe.
pipe. For less permanent installations, barbed tube or
hose connectors are recommended for use with flexible
tubing such as rubber or vinyl. This type of fitting makes
connecting and disconnecting tubing quick and easy and
yet assures a positive seal. They are limited to low
pressure installations, however.
Troubleshooting the Flow Measurement System
After a gas flow measurement system is designed and
built, it is essential that it be checked out in detail before
it can be used with confidence. First, care should be
taken to ensure that the system has been built according
to specifications. This should be followed by a thorough
leak test. The primary element, secondary element,
pressure taps, tubing connections, valves and other
related items should all be checked with a soap solution
when the line is under relatively high pressure. In the
case of differential pressure flow systems, after the leak
test has been performed, the response of the manometers
should be checked. A manometer should respond rapidly
115
-------
to the pressure differential it measures. It if does not, this
is a sign of a possible leak. An additional leak check
should be performed by closing the valves on the primary
element when there is a differential being measured. If
the manometer reading changes significantly or drifts
gradually, this may be a sign of a leak in the tubing or
fittings between the manometer and the primary element
valves. An increasing differential may indicate a leak to
atmosphere in the low pressure tubing; a decreasing
differential may indicate a leak to atmosphere in the high
pressure tubing. A decreasing differential may also be a
sign of a leak in the equalizer valve between the high
and low pressure sides of the manometer.
It is also important to make sure that the differential
pressure manometers zero properly. This should always
be done with the system under pressure. It is also
important to make sure that the blower "sees" the
primary element during this check; in other words, no
valves should be closed between the blower and the
primary element. This is necessary since if pulsation is a
problem, the manometer zero may be significantly
affected (see the "Pulsation" section of this paper).
An additional check should be made to see if pulsation
exists in the system. The manometer differentials should
be read with two different lengths of connecting tubing,
ranging from several feet to 15 ft or more in length. If the
differential varies with tubing length then this is probably
a sign of pulsation (see the "Pulsation" section of this
paper).
Look for manometer fluid oscillations at different flow
rates and line pressures. Significant oscillations can have
an adverse effect on air flow measurement. These oscil-
lations are due to unsteady conditions in the line caused
by inadequate lengths of straight pipe, aeration tank
water level variations, or possibly to blower pulsation,
although it is not to be inferred from this that blower
pulsation is always accompanied by manometer fluid
oscillations (see the "Pulsation" section of this paper). If
there is an oscillation problem and it is related to
inadequate lengths of straight pipe, then straightening
vanes (1) (13) or a new orifice plate location may help. If
it is related to aeration tank water level variations, then
this can be at least partially overcome by introducing a
headloss such as a throttled valve between the primary
element and the air release point in the aeration tank. If
the oscillation problem is related to pulsation, then it can
usually be overcome by one of the methods listed in the
"Pulsation" section of this paper.
A check should also be made to ensure that as the air
flow is steadily increased throught the line, the corres-
ponding differential pressure increase is steady as well.
A negative differential pressure reading is a sure sign of
a problem.
It is an excellent idea to check the flow measurement
system against as many known flow rates as possible. It
should be kept in mind that agreement at just one flow
does not guarantee agreement at other flows and line
pressures. Positive displacement blowers can provide a
very good check on an air measurement system as long
as an accurate blower air flow computation is performed
(the information necessary to perform this computation is
usually available from the manufacturer). Indirect infer-
ences as to flow rate can also be used to check the flow
measurement. These include oxygen transfer measure-
ments and diffuser headlosses.
In summary, after a differential pressure flow measure-
ment system has been constructed, the following checks
should be performed as a minimum requirement:
1. Make sure that the system has been built according to
specifications.
2. Perform a thorough leak test.
3. Check the response of the manometers to changes in
differential pressure.
4. Check the manometer zero under pressure.
5. Check the differential pressure readings with several
different lengths of connecting tubing.
6. Check for manometer fluid oscillations.
7. Make sure that the differential pressure increase is
steady with a steady increase in flow.
8. Check the flow measurement system against as many
known flow rates as possible.
Pulsation
Pulsation, as used here, refers to the high frequency
pressure oscillations that are produced by positive dis-
placement blowers, reciprocating compressors, and other
devices which create dynamic disturbances. It is a well
known fact that pulsation can be a major source of error
with head meters. The pressure oscillations at the two
pressure taps of the head meter do not necessarily
cancel out, with the result that the differential produced
may be considerably different than that which would
have been produced under steady flow conditions. Differ-
ent types of head meters are affected differently by this
problem, but it is probably fair to say that all are affected
to some extent. The purpose of this section of the paper
is to relate some recent personal experience with pulsa-
tion. Symptoms of the problem will be discussed as well
as methods of correcting it.
Pulsation was experienced recently during clean water
tests conducted by the Los Angeles County Sanitation
Districts at the Joint Water Pollution Control Plant in
Carson, California. The tests are part of an aeration
equipment evaluation sponsored by the EPA. It would be
appropriate at this point to describe the basic layout of
the test system.
116
-------
The clean water test facility in Carson, California employs
a 74,800 gal aeration tank, 20 ft * 20 ft x 25 ft SWD in
dimension, It utilizes a rotary lobe positive displacement
blower for air delivery which has a maximum capacity of
approximately 900 scfm at 12 psig differential and 1750
rpm. For the present work underway at the Districts'
facility, the blower has been sheaved down until its
output is approximately 420 scfm. Different air rates to
the aeration tank are obtained by wasting excess air
through a waste valve upstream of the flow measure-
ment system. The flow measurement system consists of
two different size air lines, depending on the flow rates
to be measured. Flow readings are obtained with both an
orifice plate and an Annubar (a modification of the pitot-
static tube) in each air line. The blower has 6 in. suction
and discharge piping (including a discharge silencer)
which continues for approximately 20 ft until the start of
the flow measurement piping. At this point there is a line
size reduction to 3 in. for the smaller air line and to 4 in.
for the larger air line. These pipe sizes continue for
approximately 6 ft to the orifice plate in each line. Several
feet after the orifice plate in each line, the pipe size is
further reduced by one pipe size to accommodate the
Annubars. Approximately 5-1/2 ft separates the orifice
plates and the Annubars. After the Annubars, the pipe
expands back to 6 in. to go through the aeration tank
wall. Approximately 8-1/2 ft separates the Annubars and
the aeration tank wall. Inside the aeration tank there is
approximately 15 ft of 6 in. pipe and 130 ft of 4 in. fine
bubble diffuser header pipe.
With this information as a background, it will be possible
to discuss the pulsation phenomena that were observed.
Prior to the start of the study, the air flow measurement
system was checked out thoroughly. Since the aeration
tank was empty at the time, a throttled valve was used to
simulate various water depths at different air flow rates.
Under these conditions the orifice plates and the
Annubars agreed very well and there were no signs of
any problems.
After the tank was filled with water, however, it was
clear that the agreement was no longer good. What is
more, certain phenomena were observed which were
really difficult to explain. For one thing, at low air flow
rates through a given air line, the Annubar manometer
read negative instead of positive. Extensive leak checks
were performed to no avail, and a conclusion was reached
that there was a pressure disturbance of some type in
the line. What was even more baffling, however, was the
fact that by changing the lengths of the manometer
tubing, particularly on the Annubar, considerably different
differential pressure readings could be produced. Short
tubing lengths (approximately 3 ft) tended to produce
differential pressure readings that differed by as much as
2 in. H2O from those produced by longer tubing
lengths (approximately 12-13 ft). Again, extensive leak
checks were performed to no avail. After talking to a
number of experts in the field, it was felt that larger
tubing diameters might help (1/4 in. i.d., instead of 5/32
in. i.d.) as well as making the tubing exactly the same
length on both sides of the manometer. After making
these corrections, the problem was still experienced with
the Annubar but not with the orifice plate. A decision
was made at that time to disregard the readings from the
Annubar as it was obviously being affected by some type
of pressure disturbance in the line. It was located up-
stream from a check valve and other fittings, and it was
felt that in some way they might be interferring with the
readings. It was not known at that time that the cause of
the problem was pulsation.
A large number of clean water tests were conducted with
the orifice plates as the means of flow measurement. The
Annubar readings were taken but were not used for data
analysis. Since the results of the tests did not meet the
manufacturer's expectations, a decision was made to
check out the air measurement system in more detail. At
this time, while operating in the larger of the two air
lines, a valve was closed downstream of the Annubar
such that the blower was wasting entirely through the
waste valve. Under these conditions, the Annubar
manometer should have read zero since there was no
net flow past it. Instead, it registered approximately
0.5 in. H20 in a negative direction with short tubing
leads and approximately 2.5 in. H20 in a negative direc-
tion with long tubing leads. It was at this time that
pulsation was first suspected. A valve upstream of the
flow meter was shut off, and with this the Annubar
manometer went to zero. This essentially confirmed the
pulsation phenomena.
A decision was made soon afterwards to install a large
reservoir just past the discharge silencer of the blower
in an attempt to help dampen the pulsation. This is one
of the recommended procedures to help eliminate pulsa-
tion when it exists. The tank is cylindrical, 2-1/2 ft in
diameter and 7-1/4 ft high, with a capacity of approxi-
mately 266 gal (35.6 ft3). The air inlet is side mounted
near the top, and the outlet is bottom mounted. Since the
tank has been installed, essentially perfect agreement
has been obtained between the Annubars and the orifice
plates over the full range of flows and pressures.
Furthermore, the manometers now zero perfectly and the
manometer readings are not affected by short and long
tubing leads. Also, the Annubar manometers do not read
negative at low air flows as they did prior to the installa-
tion of the reservoir.
Although the exact amount of error is not known at this
time, it does appear that the effect of pulsation on air
flow measurement was major at the combination of low
air rates (approximately 80 scfm) and high water depths
(25 ft). This was determined by a comparison of all the
orifice plate and Annubar data that had been collected to
date. There was a definite correlation of the agreement
between the Annubars and orifice plates with the air
flow and water depth. It appears that the data collected
at high air rates (approximately 400 scfm) and low water
depths (10 ft) was very nearly correct; however, as the air
117
-------
rate decreased and the water level increased, the error
became much worse. It is interesting to note that based
on these findings, it would appear that the pulsation
problem was worse at low line velocities.
The most significant outcome of the experience with
pulsation was that some of the symptoms were recog-
nized. At this point, it would appear that pulsation may
be accompanied by the following phenomena:
1. The manometer will not zero under a "no-flow" condi-
tion with the blower operating in the line.
2. The manometer readings are affected by the length of
the connecting tubing.
3. The manometer reads negative under a positive flow
condition.
It is important to point out that the absence of these
symptoms does not necessarily mean that the pulsation
problem does not exist. If one or more symptoms are
discovered, however, pulsation may be present.
As far as corrective measures are concerned, Spink (13)
has the following to say:
"The practical approach is to reduce or eliminate the
pulsation by installing added capacity or volume
together with added pressure drop in the line between
the meter and the source of pulsation."
He also goes on to say:
"Measurement may be improved by one or a combina-
tion of the following:
1. Operating at a higher differential; i.e., in a multiple
meter run installation, shutting off one or more runs.
2. Installing a higher range differential gauge and
changing operating conditions in order to use the
increased range with the existing orifice, as by reduc-
ing the flowing pressure, etc.
3. Reducing the pipe run diameter so as to use a higher
orifice-to-pipe-diameter ratio, still operating at differ-
entials as high as practicable. Increasing the d0/d ratio
will reduce the pulsation error if the operating differ-
ential remains constant.
4. Installing mufflers, headers, restrictions, or combina-
tions of capacity and pressure drop between the
primary device and the source of pulsation to reduce
pulsation amplitude.
5. Locating the primary device at a point where the
pulsation amplitude is lower (as on the suction side
of compressors)."
In our case, the installation of the reservoir (added
capacity) between the blower and primary device cor-
rected the problem.
As of yet, there are no design criteria that will guarantee
that pulsation will not be a problem in a gas flow
measurement system. Consideration of the factors above,
however, should help in the development of a good
design. Primary emphasis should be on providing sub-
stantial pipe volume and headloss between the blower
and primary device. Once built, each system should be
checked for the symptoms mentioned earlier.
Conversion of Standard Volumetric Flow Rates of
Air to Mass Flow Rates of Oxygen.
One method of converting a volumetric air flow rate in
scfm to a mass oxygen flow rate in Ib/hr is to first
calculate the fraction of the scfm flow that is dry air. This
can be done by multiplying by the mole fraction of dry air
at standard conditions. This mole fraction is related to
partial pressure in the following manner:
Ydas = (14.70-pws)/14.70
in which:
Ydas = mole fraction of dry air at standard conditions
of 68°F, 14.70 psia, and 36% relative
humidity
pws = partial pressure of water vapor at standard
conditions of 68°F, 14.70 psia, and 36%
relative humidity, psia
Making this substitution, Ydas = 0.9917. Thus, the flow
rate of dry air at STP becomes:
(21)
in which:
Qdas = 0.9917Qa
Q-das = dry a'r portion of Qa, cfm
Qa = air flow at standard conditions of 68°F,
14.70 psia, and 36% relative humidity, scfm
The density of dry air at standard temperature and
pressure is 0.07521 Ib/ft3 (6). Thus, the weight flow of
dry air in Ib/hr is given by:
(22) wdg = (0.0752 1 X60) Qdas = 4. 5 1 3 Qdas
in which:
Wda = weight flow of dry air, Ib/hr
Combining Equations 21 and 22 yields:
(23) wda = 4.475 Qa
Dry air is variously taken as 23.0 to 23.2% oxygen by
weight (6). Using 23.1 % as an average, the weight flow
of oxygen is related to the standard volumetric flow of air
in the following manner:
(24) w0 = (4:475X0.231 ) Qa = 1 .034 Qa
118
-------
in which:
= weight flow of oxygen, Ib/hr
Conversion of Volumetric Flow Rates from Standard to
A ctual Conditions
It is often necessary to determine the actual cfm flow of
air, particularly when correlations are being made with
headless. From ideal gas considerations, the scfm and
cfm air flows are related by temperature, pressure, and
relative humidity in the following manner:
(25) Q=(14.70/pfXTf/528IHpb-pbw)/pbIpf/(pf-pJ]
in which:
Q = gas flow, cfm
For air, using standard base conditions of 68°F, 14.70
psia, and 36% relative humidity, the expression reduces to:
(26) Q = 0.0276 T, Qa/(prpw)
Power Measurement
Accurate determinations of both gas and turbine horse-
power are essential in the field of oxygen transfer
Basically, they are needed for aeration system design,
economic comparisons, and data correlations. This
section of the paper will discuss the meaning of
various horsepower terms used in the oxygen
transfer field and will attempt to point out where
confusion exists. It will also discuss various methods
commonly used to make power determinations, as well
as the limitations of certain approaches. Furthermore,
particular emphasis will be placed on the need to
standardize the way in which clean water horsepower
data are reported for advertising or comparison purposes.
General Definition of Terms
At this point it would be appropriate to present the
following general definitions of some of the basic horse-
power terms used in oxygen transfer testing. These
definitions will be elaborated on in later sections of the
paper.
1. gas horsepower — that part of the total aeration
system power that is specifically related to the gas
supply
2. turbine horsepower — that part of the total aeration
system power that is specifically related to the
operation of a mechanical device such as a turbine
or pump
3. water horsepower — the work performed on the
water in the form of either gas or turbine energy
4. total water horsepower — the sum of both the gas
and turbine water horsepower
5. delivered horsepower — in the case of gas horse-
power, the theoretical horsepower required at the
blower to deliver a given mass flow of gas through
a diffuser system operating under a given static
head; in the case of turbine horsepower, the horse-
power required at the output shaft of the gear box
to turn a shaft and impeller at a given rpm in water,
under actual conditions.
6. total delivered horsepower — the sum of both the
gas and turbine delivered horsepower
7. brake (shaft) horsepower — the horsepower
developed or required at the shaft of a rotating piece
of equipment
8. total brake horsepower — the sum of both the gas
and turbine brake horsepower
9. wire horsepower — the electrical horsepower drawn
by a motor
10. total wire horsepower — the sum of both the gas
and turbine wire horsepower
The Need for Standardization
When reviewing clean water test data, it is often unclear
as to how reported power determinations were actually
made. Confusion may exist as to the following:
—Were the power determinations reported as delivered,
brake, or wire horsepower?
—Were the power determinations measured directly or
calculated?
—If calculated, what calculation procedure was used
(adiabatic, polytropic, etc.)?
—Were actual, standard, or some other ambient condi-
tions used?
—Was the effect of line loss, as well as diffuser loss
included?
—Did the determinations include a blower suction loss?
—At what gas temperature was the diffuser headless
measured?
—At what water temperature was the turbine power
draw determined?
—What blower, motor, and drive efficiencies were used?
Questions such as the above should not exist about any
reported horsepower data.
It would seem that in certain cases, rather than just
answering questions like the above in the context of a
119
-------
report, it would be desirable to report standard horse-
power data. In other words, standard conditions and
procedures should be established for power measure-
ments such that when reference is made to standard
delivered, brake, or wire horsepower, there is no confu-
sion as to what is meant. This would apply primarily to
cases where the results of shop or field tests are to be
reported for advertising or comparison purposes. In the
case of shop and field performance tests, where power
determinations are to be made for equipment acceptance
under specified design conditions, the reporting of
standard horsepower data alone would certainly not be
appropriate. In this case, the horsepower results should
be reported in terms of specified design condtions;
however, the additional reporting of standard horsepower
data would facilitate the comparison of various shop and
field performance tests made around the country.
It would seem, therefore, that it would be appropriate to
define a set of standard conditions and procedures for
power determinations. It is not a purpose of the power
section of this paper to determine the standard conditions
and procedures to be used, but rather to show a need for
standardization where it exists. Any standard conditions
or procedures should be decided upon by representatives
from the oxygen transfer field. However, possible
standards will be suggested where appropriate in
this paper.
In the case where reporting standard horsepower data is
appropriate, there are some arguments for reporting only
standard delivered horsepower instead of standard
delivered, brake, and wire horsepower. For one thing, it
would not be necessary to define standard motor, blower,
gear box, and coupling efficiencies. It is the author's
opinion, however, that standard brake and wire horse-
power should also be reported. Comparisons made
between two component aeration systems would then be
more realistic. Due to the different efficiencies associated
with a blower compared to a turbine, reporting only
standard delivered horsepower would be very misleading
as far as the overall power requirements of the aeration
system are concerned.
It is important to point out that standard clean water
horsepower numbers should not be used directly when
going from clean water test results to aeration system
design. When applying clean water test data to design, it
is important that horsepower determinations be made
which use actual ambient and operating conditions and
horsepower formulations which apply directly to the
blower being used. The purpose of standard horsepower
determinations is only to make an operating cost com-
parison of various types of diffuser systems much easier.
Gas Horsepower
Gas horsepower is that part of the aeration system power
requirement that is specifically related to the gas supply.
It refers to the general class of gas horsepower deter-
minations, including water, delivered, brake, and wire
horsepower (see Figure 4). In many cases, shop and field
performance tests are run for equipment acceptance and
the actual blower-motor combination is used. In these
cases, it is appropriate to make direct power measure-
ments. In cases where gas horsepower determinations
must be made without reference to a specific blower or
motor, it is more appropriate to calculate the power. Most
often this calculation is based on the adiabatic compres-
sion formula.
Adiabatic Compression Formula
The adiabatic formula can be derived from basic thermo-
dynamic principles. Fair and Geyer (3) have shown the
derivation. It is important to note, however, that the
final form of the equation in their derivation makes some
assumptions as to ambient conditions. Basically, the
formula represents the work done along an adiabatic
path between two gas states. It is based on the ideal gas
equation,pV = nRT, and the adiabatic relationship,
pVCp/cv = constant, where Cp and Cv are the specific
heats at constant pressure and volume, respectively. The
work performed is calculated by integrating the pressure-
volume work between the two states:
(27)
adiabatic work =
/-P2-T2
= /pdV
•A>1 Tl
After integrating, the following general form of the
adiabatic formula can be obtained:
adiabatic
(28) wire = [wRT,/(550 Kebedem)NP2/Pl)K-i]
horsepower
in which:
w = weight flow of gas, Ib/sec
R = gas constant, ft-lbf/lbm.°R (53.5 for air)
T, - absolute inlet temperature, °R
p, = absolute inlet pressure, psia
P2 = absolute outlet pressure, psia
K = (k-1)/k (0.283 for air)
k = ratio of specific heats, CP/CV (1.395 for air)
eb = blower efficiency, decimal %
ed = drive efficiency, decimal %
em = motor efficiency, decimal %
The basic form of this equation can be found in Refer-
ence (10). The case where the motor efficiency, em, is
equal to 1 is equivalent to the motor brake horsepower.
Furthermore, the case where the motor efficiency, em,
and the drive efficiency, e^, are both equal to 1 is
equivalent to the blower brake horsepower. Finally, the
120
-------
case where all the efficiencies are equal to 1 refers to
the perfect adiabatic case where there are no losses.
Substituting the known quantities for air into Equation 28
gives:
(29)
adiabatic
air wire = [(5.729x10'3)7alr QT1/(ebedem)I(p2/p1)0-283-i]
horsepower
in which:
7air = air density at the temperature, pressure, and
relative humidity for which Q is reported,
Ib/ft3
Q = gas flow, cfm
It is important to note the limitations of the adiabatic
compression formula. Too often it is used with the idea
that it accurately describes the operation of all blowers.
While many blowers are nearly adiabatic, there are some
that may be closer to polytropic in operation (11). The
polytropic horsepower formula for air is similar to the
adiabatic horsepower formula for air (Equation 29)
except that the value of 0.283 is replaced by 0.283/ep0|y,
where ep0iy is the polytropic efficiency of the blower (a
purely thermodynamic property). Furthermore, the value
of eb in the adiabatic air horsepower formula is replaced
by the mechanicl efficiency of the blower. Both the
polytropic and adiabatic formulas have been used to
describe the operation of centrifugal blowers. The opera-
tion of positive displacement blowers, on the other hand,
has usually been described with the adiabatic formula.
With this in mind, it is probably reasonable to use the
adiabatic formula when reporting clean water data, since
it has been used to describe the operation of both types
of blowers. Furthermore, the determination of the poly-
tropic and mechanical efficiencies of a blower is usually
not easy. It should be kept in mind, however, that
comparisons made between various aeration systems on
the basis of adiabatic power may not hold true when the
comparison is made on the basis of polytropic power.
Also, comparisons made between the horsepower
requirements of a given aeration system at various water
depths may differ when polytropic power is used instead
of adiabatic power.
It should also be kept in mind that the adiabatic formula
(Equation 28 with all efficiencies = 1) should not general-
ly be used to calculate the relative internal energy level
of the gas at a particular point in the gas distribution
piping. The relative internal energy level, as used here,
refers to the increase in internal energy above that at
ambient conditions. From the First Law of Thermo-
dynamics, the increase in internal energy of a gas is the
difference between the work performed on it and any
heat loss which may occur. For an adiabatic process,
where the heat loss is zero, the increase in internal
energy of a gas is thus equal to the work performed on it.
The process of supplying air to an aeration tank, however,
is not perfectly adiabatic in nature. This is to say, the
blower is not perfectly efficient, adiabatically speaking,
and the pressure and thermal losses in the distribution
piping are not usually adiabatic. This means that the
adiabatic formula cannot be used to give the relative
internal energy level of the gas. It is important to note,
however, that the adiabatic formula can be used at any
point in the gas distribution piping to determine the
adiabatic work that would be required to compress the
gas to the line pressure at that point. This is often a very
meaningful calculation.
Furthermore, the adiabatic formula (Equation 28 with all
efficiencies = 1) cannot be used to calculate the work that
is actually performed in the water. The actual water
horsepower due to the gas is the pressure-volume work
that is done as the gas rises and expands throughout the
depth of the aeration tank. This work is probably much
closer to being isothermal in nature since the expansion
probably takes place at a temperature close to that of the
water due to the very high heat capacity of water relative
to gas. If there were no oxygen transfer in the tank, it
would be theoretically possible to calculate the work
performed in the water with the following integrations:
/• water surface
(30) isothermal work = /pdV
-'gas release point
Using the ideal gas law and assuming the temperature is
constant:
water surface
(31) isothermal work = nRT/ dV/V
"gas release point
IT/d
J aas
- nRT In (Pmlet/Psurface)
in which:
n = number of moles of gas
Piniet = absolute static head at the gas release point
Psurface - absolute pressure at the aeration tank water
surface
Since under dynamic conditions oxygen transfer does
take place, performing this integration exactly is not
possible. It would seem, however, that Equation 31 could
be used to provide an approximation to the maximum gas
work performed in the water at a time of zero oxygen
transfer. Knowing the actual percent transfer, it is
possible to place a lower limit on the actual gas work
performed in the water (based on mass considerations).
Thus, Equation 31 might be used to describe an approxi-
mate range for the actual gas work performed in the
water.
121
-------
Water horsepower is really only significant from a
theoretical standpoint. It may be of some value in data
correlations, particularly in regard to mixing. Generally
speaking, however, what is most important is the power
required to deliver a given quantity of gas through a
diffuser system under specified conditions. As mentioned
previously, this can usually be approximated by the
adiabatic compression formula.
Gas Delivered Horsepower
Gas delivered horsepower is usually considered to be the
theoretical adiabatic horsepower (Equation 28 with all
efficiencies = 1) required at the blower to supply gas
through a diffuser system operating under a given static
head. Thus, for the special case of air,
air
(32) delivered = (5.729x10~3) 7.
horsepower
After conversations with a number of people in the
oxygen transfer field, it seems there are a number of
different opinions on the way in which this formula
should be applied. First of all, many experts consider the
delivered horsepower to be the adiabatic power required
at the inlet to the diffuser, rather than at the blower. In
other words, there is some confusion as to whether line
losses should be included in the gas delivered horse-
power calculation.
It should be considered at this point that the adiabatic
formula (Equation 28) provides a closer approximation to
the actual power required when applied at the blower.
For this reason, it would seem to the author that the best
definition of gas delivered horsepower would be the
theoretical adiabatic power required at the blower. In
most cases, the horsepower so calculated would be more
realistic and practical.
The gas delivered horsepower formula (Equation 28 with
all efficiencies = 1) can be used with actual ambient and
operating conditions to approximate the gas delivered
horsepower requirement on a given day at a given instal-
lation. This is sometimes necessary from the standpoint
of aeration system design and plant operation. When
reporting clean water test data for advertising or
comparison purposes, however, the use of actual ambient
and operating conditions to calculate gas delivered horse-
power can be very misleading. It is clear from the gas
delivered horsepower formula (Equation 28 with all
efficiencies = 1) that it is possible to obtain a wide range
of delivered horsepowers for a given mass flow of gas
and gauge discharge pressure. This is because the
ambient conditions are also important in determining the
horsepower requirement. This fact may tend to cloud the
real issue when comparisons are made between various
clean water tests across the country. What is really
important from a gas standpoint is the gas flow and
gauge discharge pressure at which a given test was
conducted. Different ambient conditions for the same gas
flow and gauge discharge pressure can vary the power
requirements by over 30%. In the case of power data to
be used for advertising and comparison purposes,
calculation of a standard gas delivered horsepower could
be made from the gas flow and discharge pressure by
assuming certain standard conditions.
Also, the use of actual system line losses could very well
tend to confuse diffuser horsepower comparisons made
across the country. Since line losses for any diffuser
system are usually easily controlled by varying piping
size, it would be desirable to assume a.standard line loss
when calculating the standard gas delivered horsepower.
Furthermore, if possible, diffuser headless readings
should be corrected to a standard gas temperature before
using them to calculate standard gas delivered horse-
power.
There is a need, therefore, to standardize the adiabatic
power measurements when reporting clean water test
results for advertising or comparison purposes. One pro-
posed definition of standard air delivered horsepower
might be that power calculated from the adiabatic
compression formula (Equations 32 or 29 with all efficien-
cies = 1) when compressing an scfm flow rate of air at
standard ambient conditions of 68°F, 14.70 psia, and
36% relative humidity to a discharge pressure set by the
static head, an 0.8 psi line loss and a diffuser headless
corrected to 68°F. Furthermore, a 0.1 psi suction loss
would be assumed for the blower.
The standard ambient conditions mentioned are of course
the present,standards for air flow measurement. The
value of 0.8 psi is considered to be a reasonable average
system line loss, but certainly many other values could
be used. The standard temperature of 68°F was selected
for the diffuser headloss correction because it is felt that
in most cases the diffuser air temperature is near that of
the water, and the standard water temperature for
oxygen transfer measurements is 68°F. Finally, the
blower suction loss of 0.1 psi is felt to be fairly repre-
sentative.
Making these substitutions into the air delivered horse-
power formula (Equation 32), the following expression for
standard air delivered horsepower is obtained:
standard air
(33) delivered = 0.227 Qa|t(15.5+hL+H)/14.6]° 283_
horsepower
in which:
hL = diffuser headloss corrected to 68°F, psi
H = static head, psig
While it is realized that this is not an actual power
determination, and is definitely not to be used for design
work, it does make comparisons of clean water test data
collected on different days in different parts of the
country more meaningful.
At this point, it would be appropriate to discuss the three
measurements, Qa, H, and hL, that are required for the
122
-------
determination of standard air delivered horsepower from
Equation 33. The standard air flow, Qa, has been dis-
cussed earlier in this paper. The static head determina-
tion can be made most accurately with the use of a
bubbler device (see Figure 3). The bubbler is essentially a
vertical pipe with an outlet at the diffuser level. In prac-
tice, a small air compressor can be used to bubble air
down through the bubbler pipe to the diffuser level. The
bubbler piping should be large enough so that virtually
no friction losses are obtained. A connection is made
between a pressure tap in the bubbler piping and one leg
of a mercury manometer, so that static pressure can be
read directly. A bubbler is an accurate way of measuring
the static head because it compensates for the gas hold-
up above the diffusers.
Diffuser headloss can be determined by reading diffuser
inlet pressure against the static head with a manometer.
In order to sense diffuser inlet pressure, a pressure tap
should be installed at the inlet to a diffuser(s). A connec-
tion should be made to one side of a manometer, with
the bubbler static pressure on the other side. The result-
ing differential will be the diffuser headloss. To accom-
pany the headloss reading, measurements should also be
taken of the air flow, diffuser air temperature, and static
head. As mentioned previously, for standard horsepower
determinations, the diffuser headloss should be corrected
to a standard diffuser air temperature of 68°F.
Gas Brake Horsepower
Brake or shaft horsepower refers to the power associated
with the rotating shafts of mechanical equipment. It can
refer to either input or output horsepower, depending on
the device. With motors, for example, the brake horse-
power developed is the motor output horsepower. With
blowers, the brake horsepower required is the blower
input horsepower. In an actual application where the
motor is connected to the blower by means of a flexible
coupling, belt drive, or gear reducer, there are power
losses between the motor and the blower. This means
that the motor brake horsepower developed is greater
than the blower brake horsepower required by the
amount of loss. It is important, therefore, when talking
about gas brake horsepower to distinguish between the
motor and blower brake horsepower (see Figure 4).
In the case where the actual blower-motor combina-
tions are being tested for equipment acceptance, it is
appropriate to make direct power measurements with the
techniques discussed in the section of this paper on
"Shaft Horsepower Determinations". When estimates of
brake horsepower are required without reference to a
particular blower - motor combination, then it is usually
calculated from the theoretical adiabatic horsepower by
using the appropriate efficiency.
The blower brake horsepower is related to the theoretical
adiabatic horsepower by the blower efficiency, eb:
.„.. blower brake . ..... L ,
(34) horsepower = theoretical adiabatic horsepower |eb
in which eb is in decimal %.
The blower efficiency depends on a number of factors,
including blower type (centrifugal or positive displace-
ment), number of stages, and load (including speed,
pressure differential, and ambient conditions). A range of
blower efficiencies applicable throughout the field today
might be 0.50-0.80. Where power measurements are
made directly and theoretical adiabatic horsepower is
calculated from the adiabatic formula (Equation 28 with
all efficiencies = 1), the blower efficiency can be deter-
mined from Equation 34.
Motor brake horsepower is related to the blower brake
horsepower by the drive or coupling efficiency, e^:
(35)
gas motor brake
horsepower
= blower brake horsepower/e,j
in which ed is in decimal %.
This efficiency varies with the type of coupling or drive
and with load. Coupling or drive efficiencies are usually
pretty high, probably between 0.9 and 1.0. If brake horse-
power measurements are made directly for the blower
and motor, then the efficiency, ea can be determined from
Equation 35.
When reporting clean water test data for advertising or
comparison purposes, it would be appropriate to use a
standard gas brake horsepower. It is the author's opinion
that standard gas brake horsepower should be related to
standard gas delivered horsepower by a standard blower
efficiency, ebs, and a standard drive efficiency, edS.
Reasonable values for these standard efficiencies might
be 0.70 and 0.95, respectively, although many others
could be proposed.
Gas Wire Horsepower
Wire horsepower is the horsepower drawn by the motor.
In the case where the actual blower-motor combina-
tions are being tested for equipment acceptance, it is
appropriate to make these power measurements directly
(see the section of this paper on "Turbine Wire Horse-
power"). When estimates of wire horsepower are
required without reference to a particular blower-
motor combination, then it is usually calculated from the
brake horsepower by using the appropriate efficiency.
Gas wire horsepower is related to gas motor brake
horsepower by the motor efficiency, em (see Figure 4):
(36) j?as wire = gas motor brake horsepower/em
horsepower
in which em is in decimal %.
123
-------
Motor efficiencies are a function of the motor and the
load. Typically, full-load motor efficiencies may range from
0.90-0.95. If determinations of the gas wire horsepower
and gas motor brake horsepower are made directly, then
em can be calculated from Equation 36.
As with the other horsepower terms, it would be advis-
able to define a standard gas wire horsepower. It could
be defined as the standard gas brake horsepower (stan-
dard gas motor brake horsepower) divided by a standard
motor efficiency, ems. A value of 0.92 might be reason-
able for ems.
Turbine Horsepower
Turbine horsepower is that part of the total aeration
system power requirement that is specifically related to
the operation of a mechanical device such as a turbine or
pump. It refers to the general class of turbine horse-
power determinations, including water, delivered, brake,
and wire horsepower (see Figure 5). With the exception
of wire horsepower, these determinations are normally
made with a brake, torque cell, or other torque measur-
ing device. The following section discusses various
methods used to measure shaft horsepower.
Shaft Horsepower Determinations
Shaft horsepower can be determined with a number of
torque measuring devices. One of the oldest is the Prony
brake (8), named after G.C.F.M. Richy, Baron de Prony
(1755-1839). It measures the torque required to stop the
rotation of a moving flywheel. A lever arm is attached to
a braking arrangement that grips the flywheel. The
restraining force required at the end of the lever arm is
measured with a scale of some type. The torque is then
calculated as follows:
(37) Tq = F L
in which:
Tq = torque, ft-lb
F = force required at the end of the lever arm, Ib
L = length of the lever arm, ft
The power required is related to the torque and rpm by
the following expression:
(38)
in which:
P - 27rTqNr/33,000
P = power, hp
Nr = rpm
Using this principle, actual determinations of brake
horsepower are made with the flywheel rigidly mounted
to the casing of the unit being tested, such as a motor,
blower, or gear box. The unit is mounted on a gimbal,
which allows it to rotate without appreciable power loss,
and is operated under the desired test condtions. The
torque required to stop the casing of the unit from
revolving is the brake or shaft horsepower. As might be
imagined, this type of arrangement can be very difficult
to set up.
The torque cell is another method of measuring brake
horsepower and is probably much easier to apply in
actual practice. This method works on the principle that
the strain induced in a rod or hollow cylinder is propor-
tional to the applied torque (8). Usually, a thin-walled
metal'cylinder is used as the shaft or is connected to the
shaft of the unit being tested. Strain gauges attached at
45° angles on the surface of the cylinder measure the
strain developed under load (multiple strain gauges may
be installed and connected so that the axial and trans-
verse loads will cancel out in the final readout circuit).
The torque is related to the strain at 45° angles by:
(39)
in which:
= (1/12)irG0(r«-r«)
G0 = shear modulus of elasticity, Ib/in.2
r0 = outside radius of the cylinder, in.
r{ = inside radius of the cyliner, in.
£45 = strain at 45° angles
Power can then be calculated by Equation 38.
Turbine Delivered Horsepower
For turbine equipment such as surface aerators and
submerged turbines, delivered horsepower is the power
required at the output shaft of the gear box to turn the
shaft and impeller at the desired rpm in water under
actual conditions. This power is also referred to as
turbine shaft or turbine water horsepower. The turbine
delivered horsepower is normally measured with a brake,
torque cell, or other measuring device, although it is not
usually referred to as a brake horsepower.
As with the other horsepower terms, it would be advis-
able to define a standard turbine delivered horsepower to
be used when reporting clean water data for advertising
or comparison purposes. Since water temperature would
seem to be the most important variable affecting the
turbine power draw, it might be appropriate to define
standard turbine delivered horsepower as the turbine
shaft horsepower required at a water temperature of
68°F. While it is realized that it would be impossible to
run all power tests at 68°F, it is hoped that there are
correlations existent in the field today which might relate
power draw at a given temperature to that at 68°F.
Turbine Brake Horsepower
The brake horsepower associated with turbine equipment
is similar to the brake horsepower associated with gas
delivery equipment. The gear box brake horsepower is"
124
-------
Figure 3. Static Head Bubbler Device
Pressure^
Tan ^^>^
i a\i ^^
^
— o
o
Aeration o
Tank
o
e
D iff users \
O O O O
•'i
U Filled
Manometer
^^Air
" -*' Compressor
— Ruhhlor
Tube
Diffuser
Elevation
Figure 4. Gas Horsepower Schematic
Wire
Horsepower
Motor
(ej
Motor Brake
Horsepower
Drive
-------
generally considered to be the shaft horsepower asso-
ciated with the input shaft. Furthermore, the motor brake
horsepower may not be the same as the gear box brake
horsepower due to the drive or coupling employed
between the motor and the gear reducer (see the "Gas
Brake Horsepower" section of this paper and Figure 5).
In a manner similar to gas delivery devices, the gear box
brake horsepower is related to turbine delivered horse-
power by a gear box efficiency, eg:
(40) turbme gear box = turbine de|jvered horsepower/eg
brake horsepower 9
in which eg is in decimal %.
The gear box efficiency is a function of the gear box
design (including the number of gear reductions) and the
load. The losses in a gear box are basically of two
different types, a fairly constant churning loss and a
bearing loss which is a function of load. Generally,
speaking, these gear boxes are very efficient. Rating
curves are normally available from the manufacturer. A
range of full-load gear box efficiencies applicable
throughout the field today might be 0.94-0.96. When
measurements of the powers in Equation 40 are made
directly, the efficiency, eg, can be calculated.
The turbine motor brake horsepower is related to the
turbine gear box brake horsepower by the coupling or
drive efficiency, ed:
(41)
turbine motor
brake = turbine gear box brake horsepower/ed
horsepower
in which ed is the decimal %.
The definition of the term ed, as well as its range, is the
same as that discussed in the section on gas brake
horsepower. Again, ed can be determined if the horse-
powers in Equation 41 are measured directly.
As with the other horsepower terms, it would be appro-
priate to define a standard turbine brake horsepower. The
standard turbine brake horsepower could be related to
the standard turbine delivered horsepower by a standard
gear box efficiency, egs, and a standard drive efficiency,
e,js- Possible values for both of these efficiencies might
be 0.95, although many others could be proposed.
Actual measurements of either turbine motor brake
horsepower or turbine gear box brake horsepower can be
made with the methods in the "Shaft Horsepower
Determination" section of this paper.
Turbine Wire Horsepower
Turbine wire horsepower can be measured directly with
a recording wattmeter. It is very important, however, that
the wattmeter be zeroed and calibrated accurately before
it is used. An ammeter can be used as long as the
voltage and power factor are also measured. An expres-
sion for calculating 3-phase power from current, voltage,
and power factor measurements is:
wire
(42) horsepower = 1.73 El (PF)/746 =(2.319x1 Q-3)EI(PF)
(3 phase)
in which:
E = voltage, volts
I = current, amperes
PF = power factor
It is always appropriate to use a recorder when making
power measurements. During the operation of a mechan-
ical device, there are usually fluctuations in load which
can be most easily observed with a recorder.
Turbine wire horsepower is related to turbine motor brake
horsepower by the efficiency, em (see Figure 5):
._. turbine wire , . ...
(43) horsepower = turbme motor brake horsepower/em
in which em is in decimal percent.
As mentioned previously in the section on gas horse-
power, a range of full-load motor efficiencies might be
0.90-0.95. If the horsepowers in Equation 43 are
measured directly, the em can be calculated.
As with gas horsepower determinations it is recom-
mended that a standard efficiency ems, be used to relate
standard turbine wire horsepower to standard turbine
brake horsepower (standard turbine motor brake horse-
power). As before, a possible value of ems might be 0.92.
Total Horsepower
The summation of the gas and turbine horsepowers is
the total horsepower. Thus, references can be made to
total delivered horsepower, total brake horsepower, and
total wire horsepower. When reporting clean water
horsepower data, it would be appropriate to report the
gas and turbine horsepower data separately, as well as
the combined total. In the case of data to be used for
advertising or comparison purposes, all horsepower
results should be reported in terms of standard
conditions.
References
1. Cusic, C.F. "Flow Meter Engineering Handbook"
Fourth Edition, Honeywell, Fort Washington PA,
pp. 1-14, 57-79, 85-165, 1968.
2. "Engineering Data on Differential Pressure Flow
Meters — Bulletin F1507". Bristol Company,
Waterbury CT.
126
-------
3. Fair, G.M and Geyer, J.C. "Water Supply and
Wastewater Disposal". John Wiley & Sons, Inc.,
New York NY, pp. 753-754, 1956.
4. "Fluid Meters, Their Theory and Application". Fifth
Edition, ASME, 1959.
5. "Gas Measurement Committee Report #3". AGA,
April 1955.
6. "Handy Engineering Data". Hoffman New York NY,
pp. 16, 23-24, 1973.
7. "History of Orifice Meters and the Calibration, Con-
struction, and Operation of Orifices for Metering".
Report of the Joint AGA-ASME Committe on Orifice
Coefficients, ASME, 1935.
8. Holman, J.P. "Experimental Methods for Engineers".
McGraw-Hill, New York NY, pp. 177-228, 319-322,
1966.
9. "Hydraulics and Useful Information." FMC Chicago
Pump, Chicago IL, p. 62, 1969.
10. Metcalf and Eddy, Inc. "Wastewater Engineering".
McGraw-Hill, New York NY, pp. 510-516, 1972.
11. Neerken, R.F. "Compressor Selection for the Chemi-
cal Process Industries". Chemical Engineering, Vol.
82, No. 2, p. 81, January 20, 1975.
12. Sawyer, C.N. and McCarty, P.L "Chemistry for
Sanitary Engineers". Second Edition, McGraw-Hill,
New York NY, p. 169, 1967.
13. Spink, LK. "Principles and Practice of Flow Meter
Engineering". Ninth Edition, Foxboro Company,
Foxboro MA, pp. 3-127, 415-530. 545-564, 1967.
14. 'Variable Area Flowmeter Handbook". Fischer &
Porter Warminster PA, 1976.
127
-------
Oxygen Transfer
Data Interpretation:
Non-Steady State
Clean Water Tests
David T. Redmon
Ewing Engineering Company
Milwaukee Wl 53209
Jerome D. Wren
Sanitaire-Water Pollution Control Corporation
Milwaukee Wl 53201
Mikkel G. Mandt
Houdaille Industries, Inc.
Pentech Division
Cedar Falls IA 50613
Introduction
There appears to be reasonable concensus that the
non-steady state test, more or less as described in
WPCF Manual of Practice No. 5 (1), can be a useful aid
in predicting performance of aeration systems in clean
water, and perhaps wastewater, under field conditions.
Similar concensus probably exists that opportunity
remains for improvement in the accuracy and precision
of the predictive process.
This paper discusses experiences and inconsistencies we
have encountered with this method, applied to several
different diffused aeration devices. It also suggests
procedures in data analysis which may improve the
accuracy and precision of the results.
A number of areas which should be considered in order
to avoid possible deficiencies and potential sources of
error in the non-steady state test method are the
following:
1. The appropriate assumptions of effective interfacial
area, pressure, and gas phase oxygen concentration
relationships, their effect on the saturation value, and
their variability during the test.
2. The appropriate allowance for DO gradients under
steady state conditions and its effect on the available
driving force.
3. Discrepancies and/or validity of the methods for
measuring liquid phase oxygen concentration, sam-
pling procedures, and their effect.
4. Applicability of the data to field conditions: appropriate
corrections to the overall mass transfer coefficient for
influential wastewater constituents and other
important variables such as temperature, etc.
This paper also reports and compares results obtained by
several different non-steady state test models. Agreement
and precision of KLa as determined by the log deficit
17
method, the direct method, and a non-linear least squares
data analysis method are shown as well as the relation-
ship of saturation values obtained by the direct method,
the non-linear least squares method, and actual
measurement. Suggested criteria for acceptability of test
results are proposed, along with techniques for
improvement of the various methods.
Procedure
Test Method and Facility
A description of the test method and facility have been
previously described (8); however, some of the more
important aspects of the method will be reviewed here.
The data presented are based upon DO values of gravity
flow samples obtained using the modified Winkler
titration procedure, as outlined in Standard Methods (18).
The sampling points were positioned geometrically in the
test tank in an attempt to obtain a composite of samples
that represented the average oxygen content of the tank
at any given time.
An effort was made to reaerate the tank contents until
the equilibrium saturation value was attained. In most
tests, an equilibrium saturation value was assumed to
exist after a period of time of approximately 6/KLa, as
has been suggested by Stukenburg et al. (20).
Piped samples, in situ probes, and in situ grab samples
were evaluated. Good agreement between KLB values
determined by probe data and grab samples was
achieved, and consistently higher KLa values were
obtained for these methods than for the piped samples.
To correct KLa to 20°C, a theta ( 6) value of 1.020 was
employed. It is recognized that a variety of opinions exist
as to the proper value of 6 and that additional work
128
-------
should be undertaken to verify its value as it applies to
submerged aeration devices.
Data Analysis
All test runs were analyzed by the log apparent deficit vs
time method in accordance with the procedure of WPCF
Manual of Practice No. 5 (1), except as otherwise noted.
Apparent deficit was obtained by subtracting average
tank DO concentration at the reference time from the
measured equilibrium saturation value, as described by
Boyle et al. (4) and other investigators (9) (10). The
apparent overall transfer coefficient, K|_a', (as employed
in this work) is the value obtained by linear regression
analysis of the log apparent deficit vs time data within
the range of DO concentrations corresponding to 20-90%
of the equilibrium saturation value. As subsequently
described, the value thus obtained differs from what is
considered to be the true overall transfer coefficient, KLa.
Many of the runs were additionally analyzed and KLa'
obtained by the direct method, as described by
Stukenberg (20) and Gilbert (9). Computations were made
from the same data employed in the log deficit method.
The data were analyzed by linear regression analysis; the
X-intercept was reported as C£ and the Y-intercept as
dC/dt at C = zero, which, if divided by the X-intercept,
yields the apparent KLa'.
A few runs were also analyzed by non-linear least
squares analysis of the C vs t data in accordance
with the procedure and program suggested by Baillod
and Lee (2). This method also yields values of apparent
KLa' which differ from the true KLa.
Discussion
Potential Sources of Error in the Non-Steady State Test
Detailed consideration and evaluation have been given to
the various areas of controversy and potential sources of
error in the non-steady state test method.
Influence of Effective Interfacial Area Distribution,
Pressure and Gas Phase 02 Concentration Relationships,
and Assumptions Regarding Them
A number of investigators (3) (5) (14) have considered
these relationships simultaneously as a single subject.
General recognition exists that the assumptions made
regarding them have a significant effect upon the results
obtained. The general expression describing mass
transfer between liquid and gas phases, as expressed by
Sherwood and Pigford (17), is generally accepted:
(1]
NA = KG (P-PO) = KL (C9-C)
where:
NA = mass transfer per unit of time
KG = gas film coefficient, mass transfer per unit of time
per unit pressure difference
P = solute partial pressure in gas phase
Po = solute vapor pressure in main liquid phase
KL = liquid film coefficient, mass transfer per unit time
per unit concentration difference
Ce = concentration of solution in equilibrium with
solute at partial pressure p
C = concentration of solute in main liquid phase
There appears to be general agreement (6) (12) that in
the case of air and water, the liquid film resistance is
controlling, and the relationship may be practically
considered as:
(2)
= KL(Ce-C)
However, the applicability of the expression is specifically
confined to steady state conditions where Ce and C are
constants (17).
This relationship may be expressed in more conventional
aeration equipment testing terms as:
(3)
where:
W = dC/dt = KLa (C*-C)
W = dC/dt = transfer rate per unit volume in clean water,
m/L3t
a = interfacial surface area per unit volume, L~1
C* = DO concentration in equilibrium with gas of
composition corresponding to time t, m/L3
C = DO concentration at time t, m/L3
In non-steady state testing, a, C*, and C are changing
throughout the test, with a resultant and corresponding
change in dC/dt (8) (11) (13) (19). The reduction in
oxygen absorption rate during the test results in a
progressive increase in interfacial area and a change in
its distribution during the test, as well as an increase in
exit gas phase oxygen concentration with a corresponding
increase in the saturation value, C*. However, the
expression is considered to accurately describe the
transfer phenomenon for incremental periods of time,
provided the appropriate and changing values of a and C*
are used or accounted for.
Most methods of non-steady state data analysis treat
both C* and a as constants in the analysis of the data. A
number of investigators (3) (14) (15) have suggested
the use of C* values derived from solubility tables which
give mid-depth saturation values corrected for tempera-
ture, pressure, and an assumed distribution of interfacial
area (typically uniform). Others (4) (9) (10) (16) have
recommended use of equilibrium saturation values
as C* determined by measurement, best fit of the log
deficit time relationship, or determination of the X-inter-
cept of the dC/dt plot in the direct method.
129
-------
Oldshue (14) proposed a method of accounting for
interfacial area distribution, pressure, and gas side oxygen
concentration which he considered applicable to sub-
merged turbine aerators. It may be expressed in the
following form:
(4)
where:
C* = C»s [(Ye/2Yd) + (Pa + 7Zd)/2pJ
C* = effective average DO saturation concentration,
m/L3
CS = surface saturation value, m/L3
Yd = mole fraction of oxgyen in the feed gas
Ye = mole fraction of oxygen in the exit gas
pa = atmospheric pressure, f/L2
Zd = aerator submergence, L
7 = weight density of water, f/L3
The assumption of uniform distribution of interfacial area
with depth is implicit in the expression. Although not
specifically defined, it may be inferred that this expres-
sion is applicable to clean water testing as well as to field
situations. Other investigators (3) (15) have suggested
use of the Oldshue equation with minor modifications, or
with more specific recommendations as to method of
application. All of the above references, however, suggest
treating C* as a constant during the test period. Where
oxygen transfer efficiency is low (e.g., OTE less than
0.15), or where the ratio of field transfer rate (at a given
air rate) to the cleanwater rate is high (e.g., greater than
0.7), the error induced by so doing is of minor
consequence, provided the approximately correct value of
equilibrium saturation is used in the test and in the field
application. The importance of proper selection of the
saturation value has been suggested by many (4) (7) (10)
(16) and cannot be overemphasized. An error in its
selection typically results in a several fold error in KLa'
determined by the log deficit method.
Downing and Boon (5) recognized and reported the
variability of C* during the non-steady state test and
proposed means of accounting for it. Stanton and Bradley
(19) advanced a method which assumes uniform inter-
facial area distribution with depth and plug flow in the
aeration zone. They considered it applicable to submerged
areators with mechanical agitation.
In view of the importance of the value selected for C*
and the widespread use of the simplifying mid-depth
assumption, a short digression on that subject is con-
sidered appropriate.
Various investigators (4) (9) (16) (19) have recognized
that the equivalent saturation value, expressed in terms
of tank depth, is a variable, dependent among other
things upon the type of device used and its disposition in
the system. It is acknowledged that in the applications of
some devices under some conditions, the mid-depth
saturation value will provide adequate precision, but in
the majority of cases it will not. In the special case where
interfacial area per unit volume is constant throughout
the tank and independent of depth, where mixing
characteristics are such that the DO concentration is
substantially uniform throughout the depth, and where
the net oxygen transfer is substantially zero, the expected
equilibrium saturation value can be shown to approxi-
mate the mid-depth pressure saturation value.
In this instance, oxygen will be absorbed by the liquid
from the gas below this level and extracted by the gas
from the liquid above this level. The same kinetic laws
apply to both phenomenon, and pressure is linear with
depth; therefore, net oxygen transfer ceases (saturation)
when the oxygen concentration in the liquid phase
reaches a value, in terms of tank depth, where the inter-
facial area below that reference level just equals the area
above it. Since area/unit volume is uniform, this must be
mid-depth and, under these conditions, the effective
absorptive driving force applied in the lower half just
equals the desorptive driving force in the upper half.
In other, more typical cases where interfacial area per
unit volume is not uniformly distributed, a saturation
value corresponding to some other level (or pressure) will
be obtained. We believe that in this case it will be such a
level that the integrated product of interfacial area and
driving force below this plane will approximate that above
it. The saturation value of air and water at this equivalent
pressure is believed to approximate C£, as measured
after a long period of aeration. Systems where bubbles
subdivide as they rise should tend to have saturation
values above mid-depth. Conversely, systems where
bubbles coalesce on rising should tend toward saturation
values below mid-depth.
Downing and Boon (5) also recognized the variability of
C* during the non-steady state test and proposed a
method of compensation which appears to be applicable
to a wide variety of diffused aeration air devices. Earlier
work (8) showed that their expressions may be rewritten
as follows:
(5) dC/dt = KLa (C£ Ft-C)
(6) V(dC/dt)= (M0/M>aQa(Yd-Ye) = (OTE),Q0
(7) C* = C£ F,
(8) t = ln[C»/(C£-C]/KLaF0
where:
Ft = exit gas correction factor, Downing method, at
reference time
F0 = 1 -0.5(OTE)0 = 1 -0.5(dC/dt)0 (V/Q0)
Subscript 0 = conditions where C = zero
130
-------
Qa = volumetric flow rate of air at standard conditions,
L3/t
V = liquid volume, L3
pa = mass density of air at standard conditions,
m/L3
M0 = molecular weight of oxygen
Ma = molecular weight of air
C£ = equilibrium saturation value attained by direct
method or after a long period of aeration, m/L3
(OTE), = clean water ratio of oxygen transferred to
oxygen supplied at reference time
Q0 = mass flow rate of oxygen, m/t
Downing and Boon's derivation included two simplifying
assumptions which they consider to be of negligible
adverse consequence:
1. The rate of gas flow leaving the surface of the tank
was equal to the gas flow introduced into the tank
during the test (constant pressure-interfacial area
relationship may be inferred).
2. The effective mole fraction of oxygen in the gas
phase was considered as the arithmetic average of
the mole fraction entering and leaving the tank.
Whatever error may result from assumption 1 may be
reduced by employing the following relationships re-
garding mole fraction of oxygen in the gas phase:
(9) Yd=0.21
(10) Ye = 0.21[1-(OTE)t]/{o.79+0.21[1-(OTE)t]}
Under these conditions, the following may be obtained
for Equations 5, 7, and 8:
(11)
(12)
(13)
where:
dC/dt = KLa (C£ ~{ -C)
t = ln[C»/(Q£-C)]/KLa
F'
= [1-0.605(OTE),]/[1 -0.21 (OTE),]
= [1-0.605(OTE)0]/[1-0.21(OTE)0
= exit gas correction factor, modified Downing
method
It may be shown that under the assumption of Downing,
dC/dt is linear with respect to C. When the mole frac-
tion of exit gas is corrected for absorption of oxygen,
dC/dt vs C may be shown to depart from linearity. The
fact that this departure is of minor consequence is
indirectly suggested by a preponderance of investigators
(8) (9) (20) who have observed in clean water tests
that, in fact, dC/dt vs C is substantially, though perhaps
not precisely, linear.
It is felt that the error induced by the assumption of
linearity in the modified Downing equations is of less
consequence than the error induced by the assumption
of Downing that gas flow in equals gas flow out.
As a consequence, in the direct method of analysis, the
X-intercept of the plot represents the true saturation
value of the system under test at that air rate and at an
oxygen concentration of 21%. The Y-intercept represents
the true dC/dt at C equal to zero, but at an oxygen
concentration of less than 21%. Thus, the Y-intercept
divided by the X-intercept yields an erroneous or
apparent KLa'. The true KLa may be obtained by dividing
the Y-intercept by C* at C equal to zero, which may be
obtained from Equations 7 or 12. This is equivalent to
dividing by F0 in the Downing method and by FQ in the
modified approach.
The substantial linearity of dC/dt with C is an indication
that the rate of absorption may be considered first order
with respect to C. It follows that the relationship of the
log apparent deficit vs time is comparably linear, which
is consistent with reported observations. Apparent deficit
denotes the deficit obtained by subtracting C from the
saturation value attained after long periods of aeration,
C£,. This value should either be measured directly or
estimated from the model fit to the data.
Although Downing (5) was not specific as to the applica-
tion of clean water data to field conditions, it is inferred
that the corrected KLa may be used and that the field
saturation value may be corrected for oxygen concentra-
tion at the designed or estimated dC/dt as follows.
(14) C* = (C*20) (C*sbT/C'sb20) (Pf/29.92)/3Ftf
or:
(15)
where:
q = (C£20xc*sbT/qb20)
-------
ever, a first estimate of dC/dt which accounts for a, /j,
Cf, barometric pressure, and temperature may be close
enough for all practical purposes.
To evaluate the comparative importance of gas phase
oxygen depletion, various methods of analysis, as shown
in Appendix A, were applied to test data from four
different aeration systems operating over a wide range of
transfer efficiencies. These results were then compared
with the results obtained by the method we have
previously used and described in which no correction for
gas phase oxygen depletion is made.
These results are in no way suitable for comparing the
performance of the various devices with one another as
they are at different and non-comparable depths, transfer
rates, dispositions, etc. The different systems were
selected since they may be expected to have different
pressure-interfacial area relationships. The devices
tested included ceramic plates, fixed orifice diffusers,
static aerators, and perforated aeration tubing.
In applying the test data to the field conditions, it was
assumed that a = 0.8, /3 - 0.9, Cf = 2.0 mg/l,
T = 26°C, and p = 29.0 in. Hg. Details of the test results
are summarized in Table 1. The equations and data used
to construct Table 1 are shown in Appendices A and B,
respectively.
In the computation of equivalent saturation depth, DE, the
surface saturation value was obtained by aerating 1-liter
samples of tank water. These values were typically
8.9 mg/l at standard conditions.
A comparison of the various methods may be made by
expressing the parameters of interest as ratios. Although
none of the methods are considered to accurately
describe conditions as they exist, we believe the modified
Downing method represents a better description of exist-
ing conditions from a theoretical standpoint than do the
others. Consequently, the KLa and (OTE)f data have been
expressed as ratios of the other methods to the modified
Downing method and are shown in Appendix C.
The data indicate, and it is reasonable to expect, that a
relationship between these ratios and the clean water
oxygen transfer efficiency at standard conditions, (OTE)SC,
exists. Figure 1 is a plot of the ratios of KLa and (OTE)f
Table 1. Results Obtained by Various Methods of Analysis of Five Data Sets
Run
34C
3F
2E
1G
10HD
Parameter
C20
DE
(KLa)2o
c?
-------
Figure 1. Uncorrected and Downing KLa and (OTE)f vs Modified Downing
1.2
«
|
I
1 1
i.i
1.0
KLa Downing
(OTE)f Downing
(OTE)f Uncorrected
£ 0.9 -
"* KLa Uncorrected
0 0.10 0.20 0.30 0.40 0.50
(OTE)sc
Figure 2. Mid-Depth Exit Gas KLa and (OTE)f vs Modified Downing
1.2 -
0>
B
(0
k.
(0
Q.
in
CO
'x
iii
£
Q.
O
TJ
S 1-1
+•*
E
(Q
to
o.
o> 1.0
c
'c
5
o
a
H 0.9
i
0.8
A
O
o
A
A O KLa
A A(OTE)f
O
O
o
1 1 1 1 1 1 1 1 1 1
0 0.10 0.20 0.30 0.40 0.50
(OTE)sc
133
-------
of the uncorrected and the Downing methods to the
modified Downing method vs (OTE)Sc- As might be
expected, the relationship is well defined. In terms of
(OTE)f values, the difference between the Downing and
modified Downing method is less than 1 % up to (OTE)SC
values of 0.4.
The difference in (OTE)f values by the uncorrected
method and the modified Downing method is greater
than 6% at an (OTE)SC of 0.4 under the field conditions
assumed. This is considered significant and adequate
justification for making the correction.
Figure 2 plots the ratios obtained by the mid-depth exit
gas method. It may be seen that a well defined relation-
ship with (OTE)f and KLa does not exist, and differences
between the methods are great. Ratios of (KLa)2o obtained
by the two methods range from 0.72 to 1.10. Ratios of
(OTE)f for the two methods range from 0.86 to 1.2. This,
in itself, is not evidence of the value of one method over
the other, but it suggests that careful analysis should be
made in selecting the method to use.
It is our opinion that the mid-depth exit gas method is
defective and that the general approach of Downing (5)
or some modification thereof is preferable. The principle
defects of the mid-depth exit gas method are considered
to lie in two areas.
The first, which has been previously discussed, is that
mid-depth saturation conditions are typically not attained.
In the runs reported here, the saturation values, expres-
sed as a fraction of submerged depth, range from 0.24 to
0.61. For a 20-ft submergence this corresponds to a ratio
of saturation values of about 1.19. The corresponding
ratios of K|_a' values determined by the log deficit method
will usually be considerably greater than this. The
tendency of this defect will be to yield erroneously low
KLa values where the true saturation value is less than
mid-depth.
The other area of suggested defect is that the saturation
value is considered to be a constant at the gas phase
oxygen concentration under the hypothetical conditions
of C equal to zero. In tests where oxygen transfer
efficiency is high, erroneously low C* values are used,
which yield erroneously high KLa values. An additional
disadvantage of the method is that C* may not be
obtained without knowledge of KLa and vice versa, and so
trial and error or some more sophisticated method is
required for the solution. It is further complicated by the
fact that for each assumption a linear regression analysis
is required on a new set of deficits.
Two additional, though lesser defects, which exist in the
mid-depth exit gas method are considered to be present
in the Downing (5) or modified Downing approaches as
well. Since significant oxygen is being absorbed early in
the test, the interfacial area in the upper part of the tank
is less than it is at the time C* is measured. This results
in the C* values early in the test being slightly low,
which produces corresponding high KLa values.
The second defect results from the assumption of an
arithmetic average effective oxygen concentration, rather
than a more descriptive model. This is believed to pro-
duce results in the opposite direction, i.e., high C* and
low KLa values. Downing's opinion that the error induced
is normally negligible appears reasonable to us in the
range of submergence and transfer efficiencies normally
encountered.
Effect of Oxygen Enrichment and Oxygen
Concentration Gradients Throughout the Tank
Most of the models or methods considered suggest the
assumption of uniform oxygen concentration in the liquid
phase during clean water tests, in field applications, and
in selecting the driving force for computation of oxygen
transfer efficiency. Such conditions rarely, if ever, are
obtained, but in well-mixed systems the resultant error is
minor. Significant errors, however, may result in less
well-mixed systems. In methods or models that provide a
representative average oxygen concentration of the tank,
the error that results in KLa due to this cause is
considered to be minor. The rate of oxygen absorption is
properly accounted for since the average tank oxygen
concentration is employed. The computed driving force is
considered to be near the correct value since the average
concentration of the tank at any time probably does not
differ greatly from the liquid concentration in the aeration
zone. It is further believed, however, that significant
errors may occur when these results are applied to field
designs, or in the computation of oxygen transfer effi-
ciency at standard conditions.
In cases where significant DO concentration differences
exist, we believe the field driving force should be
computed on the estimated average DO concentration in
the aeration zone, rather than the minimum value
desired, as is typically done. Similarly, computed oxygen
transfer efficiencies which use zero DO concentration for
C are considered to be fictitious since they are based on
unobtainable conditions.
The non-steady state test does not give quantitative
insight as to the correction that should be applied to
compensate for DO gradients in the aeration zone.
Application of Clean Water Data to Field Conditions
As has been shown in clean water testing, transfer rates
may be obtained either as the product of the apparent
KLa' and apparent deficit, or as the product of the
corrected K|_a and the deficit obtained using the satura-
tion value corrected for gas side oxygen depletion. In
applications to field conditions, however, the fraction of
oxygen transferred at any given air rate is different than
in clean water. As a consequence, the apparent KLa'
determined in the clean water test is not directly
134
-------
applicable to field conditions. Rather, the corrected KLa
should be used along with the field saturation value
estimated from the clean water tests and adjusted for
other known or assumed field conditions. We consider
the appropriate starting point to be the saturation
value obtained in clean water tests at the design air rates,
submergence, disposition, and geometry. Beta, or the
ratio of wastewater to clean water oxygen solubility, may
be determined or is frequently assumed. The corrections
for barometric pressure and temperature may be made
proportionally and in accordance with acceptable
temperature solubility tables, respectively. The correction
for reduced gas phase oxygen concentration may be
made as previously suggested (Equations 14 or 15) at
estimated field transfer efficiencies or by trial and error.
This method does not rely upon absolute saturation
values, but rather upon the relative solubility at various
temperatures which is not in controversy and, if in error,
would constitute a minor difference in the predicted
results. The driving force, C*- Cf,may be obtained by
using either an assumed Cf or an estimated Cf which
accounts for an estimated or determined DO gradient in
the aeration zone, e.g., Cf = Cmin +ACf/2:
where:
Cmin. ~ minimum DO in the aeration zone under steady
state conditions, m/L3
4Cf = field DO rise in the aeration zone under steady
state condtions, m/L3
Comparative Results of Methods of Parameter
Estimation
In an attempt to evaluate the agreement of results
obtained by the direct and the log deficit methods, the
data from 43 runs involving fixed orifice diffusers were
analyzed by each method and the results compared.
Similar comparisons were made with 41 runs involving
three different diffusion systems. The results are sum-
marized in Table 2 and are plotted as Figure 3.
Table 2. Direct vs Log Deficit Analysis
Parameter
(KLa')D/(KLa')LD Mean
Std. Dev.
[(OTE)0]0/[(OTE)0]LD Mean
Std. Dev.
(Ci)D/(C*) Mean
Std. Dev.
Fixed
Orifice
0.990
0.070
0.995
0.050
1.006
0.026
Various
Devices
1.000
0.090
1.001
0.067
1.003
0.028
LD = Log Deficit Method.
D = Direct Method.
NLLS = Non-Linear Least Squares Method.
Two runs in the 41 run series were omitted from the
analysis since the direct method indicated saturation
values obviously in error. One had an apparent equivalent
saturation value of 0.75 tank depth, and the other slightly
less than surface saturation. With this exception, agree-
ment between saturation values determined by the direct
method and by measurement was excellent and repro-
ducibility appears to be acceptable. The agreement
between KLa' and (OTE)sc values determined by the
direct and log deficit methods was also excellent;
however, standard deviations indicate a substantial lack
of reproducibility. The data disclose that one or both
methods, as applied to our data, are lacking in precision.
They do not disclose which. Some data have been
presented (8) which suggest that the precision of the two
analytical methods may be comparable. (Note: Sub-
sequent analysis of other more recent data obtained
since the Asilomar Workshop gives some indication that
the log deficit method is more precise than the direct
method.)
In an effort to further understand the variability of the
two methods, the dependence of the KLa' ratios upon
several potentially sensitive variables was examined.
Figures 4, 5, and 6 show the results of these analyses.
It may be concluded from these figures that lack of
dependence is shown so far as (OTE)SC and KLa' are
concerned, but reasonably good correlation exists with
respect to the ratio of the estimated and measured values
of C£. This suggests the possibility that the variability of
calculated vs observed saturation values contributes to
the variability of K|_a' in an orderly fashion. This, we
believe may partially explain the observed phenomenon
of better reproducibility in the (OTE)o values than in the
KLa' values. It also suggests that the variability of the
saturation values by the two methods, though minor, may
contribute significantly to the variation of (KLa')T.
In an effort to further explore the relative precision of
different methods of data analysis, KLa' and Q^for five of
the six runs given in Appendix B were analyzed by the
log deficit, the direct, and the non-linear least squares
methods, as previously described. The values obtained
are given in Table 3.
To appraise the degree of agreement and relative preci-
sion of the three methods, the ratios of apparent KLa' by
each method to the other were determined and the mean
values and standard deviations of the values obtained for
the five runs were compared. These results are sum-
marized as Table 4 and plotted as Figure 7. It seems
apparent that in these five limited cases, there is good
general agreement between the methods; however, the
relatively large standard deviations (0.111) in the case of
the direct-log deficit ratios indicates a lack of precision in
one of the methods. Better precision is indicated in the
cases of the NLLS/LD and Direct/NLLS by a standard
deviation in each case of approximately half of the above.
135
-------
Figure 3. KLa, (OTE)SC, and C£ Computed by Direct Method Divided by Log Deficit Method
1.10
1.05
1.00
0.95
0.90
1.10
1.05
o
uf 1.00
0.95
0.90
43
Fixed
Orifices
41
Various
Types
^Standard
\ .X^ Deviation
1.10
1.05
1.00
0.95
0.90
136
-------
Figure 4. Correlation of (KLa')D/(KLa')LD vs (Kua')LD
1.3 -
1.2
Q 1.1
0.9
0.8
0.7
t .
•
• •
10 20
(KLa')LD, hr'1
30
Figure 5. Correlation of (KLa')D/(KLa')LD vs (OTE)SC
1.3
1.2
-3
"(0
0.9
0.8
0.7
• •
•
• • • •• .
•v.\
10
20
(OTE)SC
137
30
-------
Figure 6. Correlation of (KLa')D/(KLa')LD vs
-------
Figure 7. Ratios of KLa' of Five Sets by Various Methods of Analysis
0
"ra
Q
"to
1 If)
1 OR
1 nn
0.95
0.90
1.10
1.05
1.00
0.95
0.90
1.10
1.05
1.00
0.95
0.90
^
Mean —
mmmmt
Standard Deviation •
••••
— •
i
139
-------
Table 3. Values of KLa' and C£, of Five Data Sets by Various Methods of Analysis
Run No.
34C
3F
2E
29E
1G
NLLS' m9/'
Mean' m9/'
0.1611
0.1425
0.1476
0.1504
11.64
12.27
12.11
12.01
0.1564
0.1458
0.1487
0.1503
10.58
10.76
11.09
10.81
0.3000
0.3420
0.3174
0.3198
11.43
11.11
10.94
11.16
0.09764
0.08833
0.09509
0.09369
10.40
10.82
10.62
10.61
0.03805
0.04033
0.03843
0.03894
10.75
10.55
10.45
10.58
LD = Log Deficit Method.
D = Direct Method.
NLLS = Non-Linear Least Squares Method.
Table 4. Ratios of KLa' of Five Data Sets Determined
by Various Methods of Analysis
(KLa')D/(KLa')LD Mean 0.984
Standard Deviation 0.111
(KLa')D/(KLa')NLLS Mean 1.000
Standard Deviation 0.0651
(KLa')NLLS/(KLa')LD Mean 0.9814
Standard Deviation 0.0539
LD = Log Deficit Method.
D = Direct Method.
NLSS = Non-Linear Least Squares Method.
Table 5. Ratios of KLa' of Five Data Sets Determined
by Various Methods of Analysis and
Compared to Mean Values of the Three
Methods
LD/
-------
Figure 8. Ratios of KLa' of Five Data Sets Determined by Various Methods of Analysis and
Compared to Mean Values of the Three Methods
c
a
0
a
"(0
c
a
a>
^2
"to
vT
"to
c
10
V
xg
1 10
.05
1.00
OftC
.yo
OQ/%
.yu
1.10
1.05
1.00
0.95
0.90
1.10
1.05
1.00
0.95
0.90
Mean — ^*
I
Standard Deviation ^""^ {
141
-------
with no indication of which is more representative of the
true KLa'. It is not proposed that agreement necessarily
assures the lack of defects, but rather agreement does
provide added confidence in the results. Consequently,
we believe consideration should be given to a require-
ment for some degree of agreement between the two
methods as criteria for acceptance of the results. As an
example, an acceptable range of (KLa')D/(KLa')LD
of 0.95 -1.05 and of (C* )D/(Ci )LO of 0.975-1.025
might be appropriate.
Conclusions
1. A correction procedure for exit gas composition, which
is based on the earlier work of Downing, provides a
theoretically sound, uncomplicated method for obtain-
ing clean water KLa values in the non-steady state
test. The method appears to acceptably account for
the reduced and changing gas phase oxygen concen-
tration and consequent change in saturation value
during the progress of the test.
2. The method also, through the use of in situ saturation
values, appears to acceptably account for submer-
gence and interfacial area distribution, which though
substantially constant in any given test may vary
significantly under various conditions with the same
device and are known to vary widely with different
types of aeration equipment. The general applicability
of the method is thus enhanced.
3. Some inaccuracies which result from assumptions in
the Downing derivation may be reduced by use of a
modification which more accurately accounts for gas
phase oxygen mole fraction with significant conse-
quent additional complexity. We consider this
modified procedure to be more representative of
conditions as they exist, and consequently preferred.
4. Either of the above procedures are considered to
provide greater accuracy and applicability than other
methods in common use which treat C* as a constant
and obtain the value to be used from oxygen solubility
tables, corrected for temperature, pressure, and an
assumed interfacial area distribution profile.
5. The Downing approach is also considered to provide
greater accuracy and applicability than other methods
which employ a measured or graphically obtained
saturation value, which is treated as a constant in the
determination of KLa'.
6. When the saturation value in the log deficit method is
obtained by measurement after long periods of
aeration, the direct and log deficit methods provide
values of apparent C*, KLa, and OTE which are in very
good overall agreement.
7. One, the other, or more likely both methods of analysis
are lacking in reproducibility, particularly with respect
to apparent KLa and OTE when applied to available
data. However, there is some evidence that the
methods may yield about the same level of repro-
ducibility.
8. There is some evidence that suggests that good
agreement exists in KLa' values obtained by the non-
linear least squares and the direct and log deficit
methods.
9. There is some evidence that the non-linear least
squares method of analysis may provide greater
precision than the direct or log deficit methods.
10. The non-steady state test gives no direct information
to aid the designer in the selection of the appropriate
field DO concentration, which is necessary to compute
the available field driving force. In systems and at
loadings where DO concentration is substantially
uniform, this is of minor consequence. Where the
converse is true, this may represent a significant
source of error in the application of non-steady state
data to field use.
Acknowledgements
This work forms a part of Sanitaire — Water Pollution
Control Corporation's continuing research program in
aeration. We gratefully acknowledge the opportunity to
participate in this program. We would further like to
express our appreciation for the comments and guidance
offered by Professors Boyle and Baillod; for the data
analysis and review by Mr. John Lee; and for encourage-
ment, support, and critical review provided by Mr.
Lloyd Ewing.
References
1. "Aeration in Wastewater Treatment." Manual of
Practice Number 5, Water Pollution Control Federa-
tion, Washington DC, 1971.
2. Baillod, C.R. and J. Lee, private communications,
1977.
3. Bennett, C. and G. Shell. "Submerged Static Aera-
tors: What Are They All About?" Water and Wastes
Engineering, Vol. 13, No. 5, p. 37, 1976.
4. Boyle, W.C., P.M. Bethoux, and T.C. Rooney.
"Pitfalls in Parameter Estimation for Oxygen Transfer
Data." Journal Environmental Engineering Division,
ASCE, Vol.100, No. EE2, p. 391, 1974.
5. Downing, A.L. and A.G. Boon. "Oxygen Transfer in
the Activated Sludge Process." Advances in Biologic-
al Waste Treatment, Ed. W.W. Eckenfelder, Jr. and
B.J. McCabe, p. 131, New York, 1963.
6. Eckenfelder, Jr., W.W. and D.J. O'Connor. "Theory
and Aeration Practice." Biological Waste Treatment,
Pergamon Press, Inc., New York, p. 80, 1961.
7. Eckenfelder, Jr., W.W. and D.L. Ford. "Experimental
Procedures for Process Design." Water Pollution
Control Federation, Jenkins Publishing Co., 1970.
8. Ewing, L., D.T. Redmon, and J.D. Wren. "Experi-
ences in Testing and Data Analysis of Diffused
142
-------
Aeration Equipment." Paper presented at the 50th
Annual Conference, Water Pollution Control Federa-
tion, Philadelphia, October 1977.
9. Gilbert, R.G. and D. Libby. "Field Testing for Oxygen
Transfer and Mixing in Static Mixer Aeration
Systems." Proceedings, 32nd Industrial Waste
Conference, Purdue Univ., Lafayette IN, May 10,
1977.
10. Issacs, W. P. and A. F. Gaudy, Jr. "Atmospheric
Oxygenation in a Stimulated Stream." Journal of the
Sanitary Engineering Division, ASCE, Vol. 94,
No. SA2, Proc. Paper 5905, pp. 319-344, April 1968.
11. Lakin, M.B. and R.N. Salzman. "Subsurface Aeration
Evaluation." Paper presented at the 50th Annual
Conference, Water Pollution Control Federation,
Philadelphia, October 1977.
12. Lewis, W.K. and W.G. Whitman. "Principles of Gas
Absorption." Ind. Engr. Chem. 43, p. 1460, 1951.
13. Mandt, M.G. "Improvements in Oxygen Transfer
Testing and Performance Rating." Paper presented at
the 50th Annual Conference, Water Pollution Control
Federation, Philadelphia, October 1977.
14. Oldshue, J. "Aeration of Biological Systems Using
Mixing Impellers." Biological Treatment of Sewage
and Industrial Wastes, Vol. 1, Reinhold Publishing
Corporation, New York, 1956.
15. "Recommended Practice in the Testing of Aeration
and Oxygenation Devices." A Technical Committee
Report by The Aeration and Oxygenation Devices
Subcommittee of the Technical Committee Waste-
water Equipment Council Process Equipment
Manufacturers' Association, p. 20, February 2, 1972.
16. Schmit, F.L. and D.T. Redmon. "Oxygen Transfer
Efficiency in Deep Tanks." Journal Water Pollution
Control Federation, Vol. 47, No. 11, p. 2586, 1975.
17. Sherwood, and Pigford. "Absorption and Extraction."
McGrawHill, 1952.
18. "Standard Methods for the Examination of Water and
Waste water." 14th Ed., American Public Health
Association, Washington DC, p. 446, 1975.
19. Stanton, J.L. and P.R. Bradley. "Experimental
Evaluation of Sub-Surface Aeration Systems."
Proceedings, 30th Industrial Waste Conference,
Purdue Univ., Lafayette IN, p. 826, May 6-8, 1975.
20. Stukenberg, J.R., V.R. Wahbeh, and R.E. McKinney.
"Experiences in Evaluating and Specifying Aeration
Equipment." Journal Water Pollution Control Federa-
tion, Vol. 49, p. 66, January 1977.
Discussion
C. Robert Baillod
Michigan Technological University
Houghton Ml 49931
The purpose of this discussion is to clarify a few points
raised by the authors concerning the subsurface models
based on the "true KLa" and "apparent K|_a". For a given
situation (aerator, basin geometry, water characteristics,
temperature, etc.) the two models will give identical
values for the oxygen transfer rate since, by definition:
(1) dC/dt = KLa' (C*-C) = KLa (C*-C)
It is only when adjustments for water characteristics
and temperature are applied without distinguishing
between the true and apparent values1 of a and 6 that
the OTR values predicted by each model will begin to
diverge.
In the approach based upon the "true KLa", the impact of
gas side oxygen depletion must be considered both in the
interpretation of non-steady state test data and in the
field application. However, this approach has the
advantage of dealing with a real volumetric mass transfer
coefficient. Consequently, the alpha and theta factors are
defined as ratios of mass transfer coefficients.
The "apparent KLa" approach is simpler in that the
impact of gas side oxygen depletion is automatically
reflected in the estimate of KLa'. A drawback of this
approach is that the alpha and theta factors are not
represented as simple ratios of volumetric mass transfer
coefficients. By combining the Downing-Boon model as
given by Equation 30 1 with the definitions of a and a',
it can be shown that:
fod+KLa]/| (>df+(KLa)f]}
(2) «' =
where:
4>df = correction number evaluated at dirty water field
conditions:
If the problems of scale-up are neglected, i.e., the same
gas rates and geometry prevail in the clean and dirty
water conditions,
(3)
and:
(4)
a'=a[(4>d+KLa)/(d = 1.5 min'1, KLa -
0.15 min'1, a = 0.5, and )8 = 0.9 in Equation 4 indicates
1 For a definition of terms and reference to equation numbers, see:
Baillod. C.R., "Review of Oxygen Transfer Model Refinements and
Data Interpretation", which appears earlier in these Proceedings.
143
-------
that the ratio a '/a would be about 1.057, or that failure
to distinguish between a and a' could result in an error
of 5.7%.
A similar development shows that the true and apparent
values of 6 are related by:
(5) (0')AT
Neglecting scale-up gives:
(6)
so that:
(7) (e'/8)
Again, application of typical values (# = 1.5 min"1, r«La =
0.15 min1, r = 1.1, 0 = 1.02) indicates that:
( e'/e)5 = 0.996
or that failure to distinguish between the true and
apparent values of 6 could account for a 0.5% error
when adjusting over 5°C. The net error introduced by
failure to distinguish between true and apparent values
of both a and 6 would, therefore, amount to about 7%.
The typical values employed here correspond to an
(OTE)0 of roughly 10%. The approximate error of 5%
given by this comparison agrees very well with that
indicated by the authors' Figure 1 for KLa Downing vs
KLa Uncorrected at (OTE)SC of 10%.
However, even for a given sample of wastewater, it is
difficult to measure a within an accuracy of 5%.
Considering wastewater variability, the accuracy of a
"design" value of a might be ±15%. Likewise, the
uncertainty involved in values of 6 can result in substan-
tial errors apart from the failure to distinguish between
6 andfl '.Values of 6 or 6 ' ranging from 1.020 to
1.030 are commonly employed and, over 5°C, this
alone would lead to differences of about 5% in the
prediction of field transfer rates.
It is important that the modeller and design engineer
clearly distinguish between subsurface aeration models
based on true and apparent values of KLa, a , and 6.
Failure to recognize the differences between the true and
apparent values of these parameters can lead to small,
but systematic, errors in correcting clean water test
results to field conditions.
144
-------
(II
Appendix A
Four Methods of Data Analysis
*\
c*
Used
KLa'
(dC/dt)0
KLa
<*La)20
(OTE)SC
f
(OTE)f
Unconnected
—
Not Req'd
LRAt:ln(C*-C)vst
C* P ~l
sbT "f
--. X p
°°2.\J f*# OQ Q
LCsb20 Z99J
Mid-Depth Exit Gas
M [o.433paZd paYe
^ 29.4 (29.9X0.42)
—
Not Req'd
LRAf:ln(C*-C) vst
KLa fl .C
(KLa)20(C
[0.443 PfZd PfYe 1
C* + &
sbT|_ 29.4 (29.9XO.42)J "
«w-
Downing
LRAt:ln(C*-C)vst
(KLa') (C« )
KLa'/F0
)2(20-T,J
*XV/Q0)
[CsbT Pf 1 or
C* x BFt
CMV/Q,,)
Modified Downing
LRAt:ln((£-C)vst
(KLa') (C*)
K^
[c*bT ^ Pf ]flc(
0020 LCsb20 29.92 J
fLRA - Linear Regression Analysis
tFirst approximation rather than trial and error used in all cases.
-------
Appendix B
Comparative Results of Various Methods of C* Selection
Summary of Various Runs
Assumed Field Conditions: Cf = 2 mg/l a = 0.8 (3 = 0.9 T = 26°C p = 29.0 in Hg
Run
34C
34C
34C
34C
3F
3F
3F
3F
2E
2E
2E
2E
1G
1G
1G
1G
10HD
10HD
10HD
10HD
29E
29E
29E
29E
C* Selection /
Method *>
Uncorrected
Mid-D. Exit Gas
Downing
Mod. Downing
Uncorrected
Mid-D. Exit Gas
Downing
Mod. Downing
Uncorrected
Mid-D. Exit Gas
Downing
Mod. Downing
Uncorrected
Mid-D. Exit Gas
Downing
Mod. Downing
Uncorrected
Mid-D. Exit Gas
Downing
Mod. Downing
Uncorrected
Mid-D. Exit Gas
Downing
Mod. Downing
Qa/v
scfm
1000 ft3,
37.0
37.0
37.0
37.0
59.2
59.2
59.2
59.2
33.6
33.6
33.6
33.6
20.8
20.8
20.8
20.8
0.63
0.63
0.63
0.63
12.5
12.5
12.5
12.5
) (ft)
20.3
20.3
20.3
20.3
13.0
13.0
13.0
13.0
21.3
21.3
21.3
21.3
14.0
14.0
14.0
14.0
10.0
10.0
10.0
10.0
14.3
14.3
14.3
14.3
SOTR/V
/ Ib/day
^1000ft3y
167.0
167.0
167.0
167.0
59.2
59.2
59.2
59.2
333.0
333.0
333.0
333.0
37.6
37.6
37.6
37.6
2.93
2.93
2.93
2.93
970
97.0
97.0
97.0
\ c20
' (mg/l)
10.46
11.30
10.46
10.46
9.71
9.71
9.71
9.71
11.21
10.65
11.21
11.21
10.17
10.88
10.17
10.17
10.50
9.86
10.50
10.50
10.63
9.82
10.63
10.63
0.29
0.45
0.29
0.29
0.24
0.49
0.24
0.24
0.41
0.31
0.41
0.41
0.34
0.53
0.34
0.34
0.61
0.37
0.61
0.61
(KLa)20
(hr1)
11.12
9.10
12.27
12.05
10.65
8.06
11.27
11.15
18.81
24.02
23.15
22.41
2.47
2.00
2.56
2.54
0.186
0.222
0.205
0.201
6.06
7.05
7.10
6.92
(OTE)SC
0.187
0.166
0.187
0.187
0.105
0.086
0.105
0.105
0.375
0.455
0.375
0.375
0.0723
0.0627
0.0723
0.0723
0.1837
0.2058
0.1837
0.1837
0.2944
0.31 64
0.2944
0.2944
-------
Measurement of
Alpha and Beta Factors
R. Gary Gilbert
Kenics Corporation
North Andover MA 01845
Introduction
Many parameters influence oxygen transfer in biological
wastewater treatment processes. Specifically, waste-
water constituents, liquid temperature, DO content
(driving force), type of aeration device, basin power level,
and basin geometry, to mention a few, affect oxygen
transfer. Since these parameters are usually different for
each design application, and in order to provide for
consistency in the evaluation of aeration devices,
aeration equipment manufacturers rate their equipment
at standard conditions (5). Standard conditions are
defined as clean or "tap" water at 20°C liquid tempera-
ture, 760 mm Hg barometric pressure, and maximum
driving force (zero DO).
Design engineers compute oxygen transfer requirements
at waste or field conditions, and, therefore, must properly
correct for the difference in oxygen transfer rates
between field and standard conditions. An overall correc-
tion factor can be used to adjust from field to standard
conditions as follows:
(1) SOTR = OTR/a[(/8C*-C)/C*le(T-20)
This correction factor accounts for the following specific
differences between standard and field conditions:
—wastewater constituents
—temperature
—driving force
—barometric pressure
Of the above, wastewater constituents and temperature
influence both the overall oxygen transfer rate coefficient,
KLB, and DO saturation concentration, C* (and sub-
sequently driving force).
Temperature, barometric pressure, and operating DO
level values for Equation 1 are easily determined. The
effects of wastewater constituents as expressed by the
alpha and beta factors and the temperature correction
factor are not determined as easily. The sensitivity and
significance of these three factors are discussed in this
paper. Current measurement practices together with
results of various studies are reviewed. Recommend-
ations for development of standardized alpha and beta
measurement procedures are presented.
Significance of Variation
The alpha factor, a, is the ratio of the overall oxygen
18
transfer rate coefficient in wastewater to that in clean or
"tap" water and is expressed as:
(2)
a = KLa (wastewater)/KLa (clean water)
Alpha is affected by wastewater constituents such as
soluble BOD, COD, suspended solids concentration,
surface tension of the wastewater, and temperature and
by system parameters such as type of aeration device,
power level, and basin configuration. In general, the
influence of these variables can be reduced substantially
by conducting the alpha measurement at field design
conditions.
The beta factor, j3 , is the ratio of the system DO
saturation concentration in wastewater to that in
clean water and is expressed as:
(3)
= C* (wastewater)/C* (clean water)
Beta is influenced by wastewater constituents such as
salts, organics, and dissolved gases and by barometric
pressure and temperature. Conducting measurements at
equal barometric pressure and at field design tempera-
ture will reduce variations to those caused by wastewater
constituents alone.
Typical response curves for reaeration of clean water and
wastewater are presented in Figure 1. Each curve has a
unique overall oxygen transfer rate coefficient, KLa, and
DO saturation concentration, C*. Determination of alpha
is possible by one or more of several methods of analysis
(18). Beta is determined by direct measurement. In the
example presented in Figure 1, alpha equals 0.7 and beta
equals 0.9. These are typical values for a wide variety of
industrial wastewaters and for municipal wastewater.
The significance of departure from these values is
illustrated in Figure 2.
The magnitude of error in selection of alpha and beta
values can be substantial. Reduction in an alpha factor
from 0.7 to 0.5 will result in an oxygen deficit of 40%
between requirements at field versus standard conditions
when all other factors are held constant. Similarly, an
alpha factor of 0.9 will result in an excess standard
oxygen value of 22%. When both alpha and beta vary
from a design value, the effect is compounded. A
comparison of this effect is presented below for a design
alpha = 0.7 and beta = 0.9. Equation 1 is used to compute
the percent change in oxygen required at standard
conditions.
147
-------
Variance in a&/3 % Variance
Change in 02 Required
a only a&/3
+0.1
-0.1
+12 (excess) +24
-17 (deficit) -36
As the variance in actual versus design alpha and beta
factors increases, differences between oxygen required at
field versus standard conditions also increases, with the
deficit effect being more pronounced. This effect is
illustrated by the two curves in Figure 2. Based upon the
relationship between alpha and beta factors and resulting
oxygen requirements at field versus standard conditions,
it is no wonder that aeration systems are frequently over
or underdesigned.
Variation in the alpha factor is common, as illustrated by
the typical laboratory data for various industrial wastes
presented in Table 1 (13). Not only do alpha factors vary
more than 100% for similar raw wastes, but as bio-
oxidation proceeds and contaminants are destroyed, the
alpha factor approaches that in clean water. Alpha
factors for municipal wastewater range from 0.7 to 0.9,
whereas alpha values for combined and industrial waste-
water become more variable as the wastewater becomes
more complex. Not only do alpha factors vary from
industry to industry, but also from plant to plant within
an industry. It is not uncommon to observe changes in
the alpha factor from week-to-week, day-to-day, and even
hour-to-hour, depending upon factors such as production
scheduling and rain water dilution. The author has
observed changes in alpha from 0.8 to 0.5 in a brewery
waste over a 4-hr period. As discussed earlier, these
variations in alpha cause even greater variations in
oxygen transfer.
Temperature affects both equilibrium values for oxygen
concentration and the rate at which transfer occurs due
to changes in kinematic viscosity, which in turn affects
oxygen diffusion into the liquid. The effect of temperature
on the overall oxygen transfer rate coefficient is usually
expressed as follows:
The temperature correction factor has been reported to
vary from less than 1.01 to more than 1.05. A compari-
son of four temperature correction factor references is
presented in Figure 3. Eckenfelder's theta value of 1.02
(13) was developed from experimental data and by
comparisons of his work with that of Streeter (39),
Howland (19), and Wilke (44). O'Connor's temperature
correction equation is based on changes in viscosity at
the respective temperature (34). Shell suggests that two
temperature correction factors be used, a value of 1.03
for temperatures less than 20°C and a value of 1.01 for
temperatures greater than 20°C (4). A theta value of
1.024 is considered realistic by many (32).
Data analysis by the author of more than 180 full-scale
reaeration tests conducted in water between 14°C and
26°C revealed temperature correction factors between
1.024 and 1.026. A data summary is presented in
Table 2.
As can be seen from the curves in Figure 3, temperature
correction factors vary by approximately 15% between
10°C and 30°C. Most field conditions are within this
temperature range. Caution is advised, however, when
correcting field data to standard conditions outside the
10 to 30°C range where variances become greater.
Also, the presence of surface active agents and other
waste constituents may affect the temperature correction
factor.
Measurement and Reliability
A great deal of literature has been published regarding
measurement of alpha and beta factors and the difficul-
ties involved both with the apparatus used and liquids
being tested. The type of aeration device, mixing intensity,
and wastewater constituents present are only three
parameters that affect the outcome of the alpha and beta
measurement. Researchers are not in agreement con-
cerning such questions as the effect of suspended solids
on alpha, mixing level versus KLa, surfactants effect, type
of aeration device to be used, and probe correction
factors for measurement of beta, to mention a few. The
literature contains more controversy regarding these
questions than clarification. There is a great need for
both fundamental and applied research in the areas of
alpha and beta measurement.
Alpha Measurement
Test Apparatus and Aeration Device
Bass and Shell (4), Stukenberg, Wahbeh, and McKinney
(40), and other investigators (3) (8) (15) (16) (17) (29) (35)
have described test apparatus used to conduct alpha
measurements. Three of these test devices are shown
in Figures 4, 5, and 6. The bench-scale aeration tank
shown in Figure 4 contains both a submerged turbine
agitator and an air sparger resembling a submerged
aerator (40). The tank has a 10-liter liquid capacity. The
diffused aeration equipment shown in Figure 5 employs a
fritted cylinder diffuser in a 16-liter aeration tank (4).
Apparatus resembling a surface aerator is shown in
Figure 6 (4). The tank has a 28-liter capacity, and mixing
intensity is controlled by varying agitator motor speed.
The latter two aeration devices have been used to
conduct comparative alpha measurements.
Otoski (35) and Marotte (29) have used alpha test
apparatus employing more than one type of aeration
device in their respective studies, as follows:
148
-------
Figure 1. Response Curves Illustrating Alpha/Beta Effect
O)
O
O
Clean Water Curve
KLa = 0.333 mirT1
Wastewater Curve
KLa = 0.233 min"1
Curve form C= C*(1-e-K>-at)
From Curves:
« = 0.233/0.333 = 0.70
= 9.0/10.0 = 0.90
i- C* = 10.0 mg/l
C* = 9.0 mg/l
12345
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Time (min)
Table 1. Summary of Oxygen Transfer Characteristics of Some Industrial Wastes and
Their Bio-Oxidation Effluents (13).
Waste
BOD (mg/l)
Raw
Effluent
Raw
*Paper repulping and semi-chemical wastes mixed prior to bio-oxidation.
fBleach plant and pulp mill wastes mixed prior to bio-oxidation.
{Average values.
Table 2. Temperature Correction Factor Determinations
Effluent
*Paper repulping
•Semi-chemical machine backwater
Mixed kraft mill
fPulp and paper (bleach plant)
fPulp and paper (pulp mill)
Pharmaceutical
Domestic sewage (fresh)
Synthetic fiber
Board mill
187
1872
150-300
250$
205$
4500
180
5400
660$
50
—
37-48
30J
—
380
9
585
—
0.68
1.40
0.48-0.86
0.83-1.98
0.66-1.29
1.65-2.15
0.82
1.88-3.23
0.53-0.64
0.77
—
0.70-1.11
0.86-1.0
—
0.73-0.83
0.98
1.04-2.65
—
Tank Water Depth
(ft)
20
10
10
10
Air Flow per
Aerator (scfm)
42-45
19-22
28-30
42-48
No. of
Tests
18
46
74
43
Temperature
Range (°C)
14-25
15-26
15-25
15.25
Theta
6
1.024
1.025
1.024
1.026
149
-------
Figure 2. Relationship Between Alpha/Beta and Oxygen Requirements
+70
+60
+50
+40
H +30
0)
E
| co +20
w (/)
+10
>-- 0
Design Point ( p - 0.9)
Alpha Only
-10
s t
c «
a> c
S; .£
£ £-20
8 -30
L.
-40
-50
-60
-70
-80
-90
-100
-110
-120
-130
-140
1.0
Design Point (a =0.7)
1.5
SOTR = OTR/&[(/3C$-C)/C;] 1.024(T"20)
Combined Variances
Aa&jS
+0.1
0.0
-0.1
% Req'd
124
100
64
Variances
Aa
+0.2
+0.1
0.0
-0.1
-0.2
% of Requirement
122
112
100
83
60
A% Req'd
+10
+12
-17
-23
a
150
-------
Figure 3. Effect of Temperature on Oxygen Transfer Rate.(KLa)T = (KLa)20 0 (T"20)
0=1.02Eckenfelder(13)
O'Connor (34)
,--
0= 1.01 Shell (4)
- r-0'Connor: (K,a)1/(KLa)2=
9= 1.03 Shell
6= 1.024
9= 1.02 Eckenfelder
9= 1.024
20 30
Temperature (°C)
Figure 4. Bench-Scale Aeration Tank (40)
9-3/4"
8" P.G. Pipe
with 1/4" Wall
1/4" P.G. Basel^
-1/4" Riser to
Purgemeter
1 /4" Tee
4-1 /8" P.G. Baffle-
4 at 90°
Plan
Legend
P.G. — Plexiglass
Cu — Copper
S.S. — Stainless Steel
Purgemeter
10 liter Liquid
Capacity
3/8" S.S. Shaft
Baffles, Provide
(2)1/8"x4-1/2"
Slots at Wall
(4) 1/8" S.S. Blades
Welded to Collar
(Rotor Speed
300 rpm)
3/8" S.S. Collar
with Set Screw
1 /8" Air Opening
(Air Flow Rate
0.4 l/min)
Section
151
-------
Figure 5. Schematic Diagram of Diffused Aeration Alpha Apparatus (4)
Two-Stage
Regulator
DO and Temperature
Meter
Oil Trap
Filled with
Glass Wool
DO and Temperature
Probe
Water Trap
Half Filled
with Water
Compressed
Air
Rotameter
(0.1 to 5 SLM)
Tygon Tubing
••— Gas Dispersion Tube with 12-mm
Diameter Fritted Cylinder
Timer Stopwatch
Plexiglas
Aeration
Basin Liquid
Vol.= 16 liters
Glass Y
Connector
12" Air Stones,
Tubing and Connector
for Wastewater
12" Air
Stones for
Clean Water
152
-------
Ostoski
Marotte
—Surface agitator
with lower mixing
impeller
—Diffuser stones
—Static Mixer
—Surface agitator
—Static Mixer
—Submerged agitator with
air sparger
—Diffuser
Both investigators have observed variations in alpha with
an aeration device over a wide range in KLa. Similar
observations were made by Bass and Shell (4) from alpha
tests conducted on a complex industrial wastewater.
From these studies, it appears that variations of up to
40% can occur in alpha measurements simply as a result
of using different aeration devices.
Similitude of aeration devices seems to be a reasonable
approach for minimizing measurement effects of the
aeration device. Kalinske (23) and others (8)(33) state that
the same type of aeration device used in the field instal-
lation should be used in the alpha test vessel. The author
has had good success using an aeration device scaled
down from the prototype aerator for alpha measurements.
Turbulence
Alpha measurement is also affected by the mixing level
or turbulence in the test vessel. WPCF Manual No. 5 (1)
cautions that extreme care must be employed in the
interpretation of bench-scale alpha data, and that the
pattern of turbulence has an influence on alpha which
may be different than from full-scale aeration tanks.
Otoski (35) found a direct relationship between turbu-
lence and alpha value for all three aeration devices
studied. Stukenberg et al. (40), Weis and Lad (43), and
Bass and Shell (4) recommend that alpha tests be con-
ducted so that the alpha test vessel KLa approximates the
field value to minimize the turbulence factor. It is believed
that with the same KLa, the two systems will be similar
enough in mixing characteristics to yield the same alpha
factors within normal experimental differences.
Recent studies, conducted to determine the effect alpha
test vessel liquid depth has on alpha (29) (45) and to
compare bench- with full-scale alpha measurements (35),
indicate that matching test vessel and field KLa values
may not represent equivalent mixing conditions for sub-
merged aeration devices. KLa is a function of depth in
submerged aeration; therefore, an alpha test vessel 2 ft
in depth requires greater air flow per unit volume than
does a prototype aeration system 20 ft in depth with an
equivalent KLa. Comparative tests between a 3 ft deep
alpha test vessel and a 10 ft deep full-sized system,
using scaled down and full sized aeration devices,
revealed the following relationship between air flow rate
per unit volume and KLa:
Figure 6. Schematic Diagram of Surface Aeration Alpha Appratus (4)
Variable Speed Motor
Electronic
Tachometer
Liquid Volume = 28 liters
12"
-19" Diameter-
Stainless Steel Tank
153
-------
Bench-Scale Bench-Scale KLa, Full-Scale KLa,
Air Flow Rate* 3 ft Liquid Depth 10 ft Liquid Depth
0.08 min'1 0.24 min'1
0.13min-1 0.33 min'1
25 scfh**
40 scfh**
*Air flow rate per unit volume same for both systems.
•'Standard ftVhr.
For equal air flow per unit volume, KLa in the full-scale
aeration system was 2.5 to 3 times greater than in the
bench-scale unit. The additional depth in the full-scale
system accounted for the larger K|_a value.
Zlokarnik (45) recommends that alpha test vessels, using
submerged aeration devices, have at least a 10-ft liquid
depth to provide for adequate bubble formation and
coalescense properties characteristic of a full-scale
aeration system. Both Marotte (29) and Zlokarnik (45)
have observed variations in alpha values with test vessel
liquid depth. In light of the effects that turbulence has on
alpha measurement, determination of alpha at several
aeration intensities should be considered prior to select-
ing an appropriate values) for design purposes.
Suspended Solids
There seems to be little or no concensus concerning the
effect that mixed liquor suspended solids (MLSS) has on
alpha measurement. Stukenberg et al. (40) explain two
ways in which oxygen transfer can be affected by MLSS:
by changing viscosity and by releasing soluble organics.
It is their belief that the MLSS concentration under
aeration is not high enough to adversely affect viscosity.
On the other hand, soluble organics released by the
active microbial mass in the aeration tank could affect
oxygen transfer by releasing soluble surface active
materials in the presence of an oxygen-deficient environ-
ment. As long as the cell mass is aerobic, organics
should be consumed and there should be no affect upon
alpha. The authors further state that alpha can be
determined on the final effluent or on the supernatant
from the settled mixed liquor, rather than the mixed
liquor under aeration. Obviously, if the activated sludge is
allowed to go anaerobic during testing, release of some
organics could reduce the alpha factor below the actual
value in the system. However, since alpha approaches
unity in proportion to purification of the wastewater (13)
(26) (40), conducting alpha measurements on the final
effluent may be on questionable value.
Bass and Shell (4) recommend that alpha determinations
be preformed on waste that contains the same approxi-
mate type and concentration of biological solids as will
be present in the full-scale system. Supportive data
indicate that the presence of biological solids has a
significant detrimental effect on the oxygen transfer rate
in samples of waste containing solids, versus samples
which had been filtered prior to testing. Casey and
Karmo (22) found that activated sludge solids in concen-
trations of 2,000 mg/l affected the rate of oxygen
transfer; whereas, non-flocculated suspensions of
polyvinyl chloride (PVC) powder and fine and coarse
dried peat had a negligible effect.
Jackson et al. (6) report average alpha values of 0.63 for
groundwood paper mill wastewater mixed liquor both in
bench- and in full-scale tests. They indicate, however,
that higher alpha values were obtained with the clear
liquor (cells separated) and attribute the difference in
part to the fiber content of the wastewater.
Damhaug and Balmer (9) concluded that determination of
the alpha factor on activated sludge supernatant or
secondary effluent may give misleading results. They
summarized literature on the effect of MLSS by stating
that the primary effect of suspended solids on KLa is a
reduction of liquid hydrodynamics caused by an increase
in the liquid viscosity and consequently a reduction in the
KLa value. Conversely, Matson and Bennett (30) discuss-
ed a positive effect of suspended solids on alpha. They
presented findings that microorganisms can increase
oxygen transfer rates by disrupting the liquid surface film
surrounding the gas bubble.
Surface Active Agents
Aeration tank alpha values are dependent on the exact
level of bio-oxidation of the organic matter and deter-
gents present. As previously mentioned, partial
decomposition of organics produces various organic acids
that affect the alpha value as do any other surface active
agents present. The "surface active agent effect" is
another area where basic and applied research is
required.
The presence of materials such as proteins, detergents,
oils, and other surface active agents affect film thickness
and surface tension. Changes in film characteristics due
to dissolved organics tend to decrease oxygen transfer.
Aiba and Toda (2) suggested that the effect of surface
active agents is to make the interface more quiescent,
and, therefore, to decrease the rate of surface renewal.
Experiments with pure surfactants and other organics (2)
(10) (11) (12) (27) (28) (31) (35) (41) have typically result-
ed in curves such as those shown in Figure 7.
Eckenfelder and Ford (15) postulated the relationship
between the alpha factor and the degree of turbulence
generated by the aerator at a given surfactant concentra-
tion. The mechanism of interference is shown graphically
in Figure 8. Under quiescent conditions, the principle
resistance to oxygen transfer is diffusion through the
body of liquid so that the establishment of an interracial
film has little effect on the total transfer. At intermediate
levels of turbulence, complete mixing is achieved
throughout the liquid volume and the interracial resist-
ance of the absorbed surfactant reduces the transfer rate.
At high mixing levels, the interfacial renewal is so great
that no interfacial barrier is permitted to be established
and the value of alpha is again approximately unity.
154
-------
Figure 7. Effect of Surfactant Concentration on Mass Transfer Characteristics
t
t
(0
5
0)
S
n
3
CO
Surfactant Concentration
Surfactant Concentration
(b)
Figure 8. Relationship Between the Alpha Factor and the Degree of
Turbulence at a Given Surface Active Agent Concentration (15)
Depth of
Liquid
1.0
Surfactant
Molecules
Oxygen
Concentration
Gradient through
Depth of Liquid
©
Quiescent
Surface
Oxygen
Dispersion
High
Interfacial
Renewal
High Air
Entrainment
Intensity of Mixing or Turbulence-
155
-------
Higher degrees of turbulence result in entrainment of
fine air bubbles which increase the interfacial area and
alpha to a value greater than unity.
Since surface, submerged turbine, and diffused aeration
all produce different types of mixing regimes with respect
to the surface active agent effect, it is not unreasonable
to expect that different values of alpha might result for
the same liquid being aerated by different aeration
devices. Furthermore, bench- (35) and full-scale surface
active agent studies, reported upon here for the first time,
indicate that alpha varies significantly with the type of
diffused aeration device used.
From Otoski's research (35), it was determined that
distinct relationships exist between aeration device, sur-
factant concentration, and mixing. These relationships
are summarized as follows:
Aeration Device
Effect on
Alpha as a Function of Increasing:
(a) Mixing Leve(b) Surfactant Cone.
Surface agitator
Static mixer
Diffuser stone
a decreases
a increases
a increases
a decreases
a increases
a decreases
The studies were conducted in a 100-liter capacity
bench-scale aeration tank using the three aeration
devices previously described, three mixing levels, and
three levels of surfactant concentration. A pure anionic
surface active agent (Richonate 40B — sodium dodecyl-
benzene sulphonate [C^l^sCeSOaNa]) was used in
concentrations of 5, 10, and 20 mg/l. A summary of
alpha factors is presented in Table 3 for each aeration
device. Similar results have been reported by Marotte
(29) for pulp mill wastewater.
Similar full-scale studies, conducted by the author,
provided results which were in good agreement with the
bench scale work. The full-scale test facility has been
described elsewhere (18). The same surface active agent
was used during the testing, aerators were full-sized
prototypes, and the volume of water tested was 87,000
gal. Mixing intensities were varied from less than 10
scfm/1,000 ft3 to over 190 scfm/1,000 ft3 of aeration
volume. Alpha values ranged from unity to over 1.2,
averaging 1.1 for all tests.
Generally, alpha increased with mixing intensity as
would be expected. Surfactant concentration had little
influence on relative oxygen transfer.
A second series of tests were conducted using a coarse
bubble sparger aeration system. The type and quantity of
surfactant, aeration volume, and mixing intensities
remained unchanged. The results of these tests were
different; the alpha value averaged approximately 0.7 to
0.8 rather than 1.1 for the static aerators. Alpha values
did increase with mixing intensity, however.
Other investigators (37) (42) have conducted full-scale
alpha tests using surface active agents. Schmit et a/.
(37) described results of full-scale tests conducted with
approximately 20 ft of submergence over broad band
coarse bubble diffusers at two mixing intensities (25.4
scfm/1,000 ft3 and 92.5 scfm/1,000 ft3). The
aeration tank volume was approximately 19,000 gal.
Table 3. Summary of Bench-Scale Alpha Factor
Surfactant Test Results (35)
Mixing Level'
Surfactant Concentration,
mg/l
O2 5 10 20
Static Mixer
25 SCFH3
40 SCFH3
55 SCFH3
Diffuser Stone
5 SCFH3
10 SCFH3
15 SCFH3
Surface Agitator
370 rpm5
440 rpm5
510 rpm5
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.98
0.97
1.09
0.85
0.91
0.91
0.94
1.05
0.91
1.00
1.00
1.05
0.79
0.93
0.86
1.00
0.91
0.88
0.93
1.06
1.19
—
—
1.02
0.97
0.79
1 Mixing levels were considered equivalent in all devices
when KLa values in clean water were equal.
2 KLa values approximately equal in all devices at each
mixing level.
3 Standard ft3 air/hr.
4 Tests not conducted due to excessive foaming.
5 Revolutions per minute.
The type and concentration of surface active agent used
in the tests was not reported. Alpha values ranged from
0.65 at the low mixing level to 0.75 at the high mixing
level. Bench-scale alpha tests conducted on the same
water did not yield similar results. These tests indicated
that the presence of the surface active agent had no
effect on the bench-scale alpha.
British government research on fine bubble dome dif-
diffusers (42) indicates that water with a 5 mg/l concen-
tration of detergent produced an alpha value of less than
0.5; whereas, the alpha value in mixed liquor was
reported to be about 0.6, but ranged from 0.3 to 0.8
depending upon aeration time. The mixing intensity was
approximately 12.5 scfm/1,000 ft3 at the 20-ft
water depth condition. A summary of the three full-scale
alpha test results is contained in Table 4.
156
-------
Table 4. Summary of Full-Scale Alpha Factor Surfactant Test Results
Aeration Device
Tested
Static Aerator
Coarse Bubble Sparger
Coarse Bubble Broad
Band (37)
Fine Bubble (42)
Aeration Tank
Volume (gal)
87,000
87,000
19,000
—
Tank Depth
(ft)
10.0
10.0
22,5
5-25
Mixing Level
(scfm/ 1.000 ft3)
10-190
80-190
25-92
10-30
Alpha Factor
a
1.0-1.1
0.7-0.8
0.65-0.75
0.4-0.5
Kalinske (24) reported that at least one large consulting
firm considers the non-steady state test using clean
water as the only reliable way to evaluate and compare
aerator performance as a result of the variable nature of
the alpha value. This same consulting firm (6) does
require, however, that a surface active agent be present
in the clean water tests to approximate the "detergent
effect" wastewater has on the aeration device. The alpha
values in Table 4 seem to indicate this effect and support
justification for use of a surface active agent in clean
water tests. Schmit et a/. (37) indicate that this approach
is reasonable.
This author, however, sees no benefit in rating aerator
performance with surface active agents present. The only
accomplishment would be to introduce another variable
to an already too-confused technical area. The combina-
tion of types and concentrations of available surface active
agents are almost infinite. Another factor to be considered
is the reduction in surfactant concentration as an aeration
test proceeds. Figure 9 contains a plot of surface tension
versus reaeration time for a clean water test in which
5 mg/l of Richonate 40B was added prior to the test. The
surface tension dropped from 72 dynes/cm at the begin-
ning of the test before adding the chemical to a low value
of 53 dynes/cm 5 min after the test had begun. Surface
tension increased to about 60 dynes/cm after 20 min of
aeration and remained at that level throughout the
remainder of the test. Other test conditions have yielded
different results.
Beta Measurement
The beta factor has been defined as the ratio of the
dissolved oxygen saturation concentration for wastewater
at field conditions to the saturation value for clean or tap
water at field conditions. Beta has been commonly
referred to as the salinity correction factor because
dissolved salts reduce oxygen solubility in wastewater. In
addition, dissolved organics and gases in the wastewater
can reduce oxygen solubility as well. Unfortunately, these
constituents also affect the measurement of DO.
Several methods are available for measurement of
DO. The various methods can be divided into three
principal categories: 1) laboratory methods, 2) elec-
trochemical analysis, and 3) membrane electrode
methods. The most popular and most widely used
laboratory method is the Winkler Method (38). This
method is used as a standard for calibration and
for determining DO content in relatively clean waters.
Unfortunately, the Winkler Method is not suitable for
use with complex or colored wastewaters due to
possible test interferences. (For example, there may be
oxidizing or reducing agents present in the sample that
may either liberate or reduce iodine which will introduce
error). The electrochemical methods are not suitable for
measurement in wastewater because of interferences
and electrode contamination.
The most popular methods of DO measurement are
the membrane electrode methods (36). The success
of these methods is due to the isolation of the
electrodes and electrolyte from the sample by means of
a semi-permeable membrane. There are two types of
membrane electrodes: 1) the galvanic membrane elec-
trode and 2) the amperometric membrane electrode. The
galvanic type relies upon the development of a potential
between two dissimilar metal electrodes when contact is
made with oxygen in the sample to be analyzed. The cell
generally consists of a lead anode and a platinum or
silver cathode. As oxygen passes through the membrane
at a rate proportional to its partial pressure, a potential
develops between the anode and cathode. The resulting
current flow is proportional to the oxygen present in the
reaction.
The amperometric membrane technique is probably the
most widely used method for measurement in waste-
water. The electrodes (usually a gold cathode and a silver
anode) are driven by an externally applied polarizing
voltage. The current flow between the electrodes is
directly proportional to the amount of oxygen present in
the test solution.
Bass and Shell (4) contend that probes will yield false
saturation readings because they measure percent
saturation and not actual DO concentration, and,
therefore, cannot be used for beta measurement.
The example is given that if a probe is calibrated to read
saturation at 9.2 mg/l, it will read 9.2 mg/l when placed
in any saturated solution. This is not entirely correct.
Probes measure oxygen partial pressue. Inside the
membrane, next to the cathode, the oxygen pressure is
157
-------
zero because oxygen is consumed as it reaches the
cathode. Outside the membrane, the oxygen pressure is
the ambient pressure to be measured. The flow of oxygen
through the membrane is directly proportional to the
differential oxygen pressure across the membrane, and
since the oxygen pressure is zero on the cathode side of
the membrane, the flow is proportional to the absolute
oxygen pressure on the outside of the membrane. The
current resulting from the chemical reaction is in direct
stiochiometric relation to the amout of oxygen being
consumed (20) (21).
The apparent false saturation measurement can be
corrected on some DO analyzers (21) by making
a salinity correction adjustment to the measurement
obtained in wastes with high concentrations of
dissolved salts. This correction is to compensate for
salinity effects on the membrane which change the true
rate at which oxygen passes through the membrane
and, therefore, the oxygen partial pressure.
The effects of various constituents on oxygen saturation
measurements using two types of probes and the Winkler
analysis are summarized in Table 5. In some cases, the
manufacturer's correction procedures agree with the
Winkler Method (NaCI and Na2S04 solutions). In other
cases, there is no agreement and even the Winkler
Method results are meaningless or questionable. Because
of these interferences, Bass and Shell (4) recommend an
alternate method of determining beta by computing the
ratio of the saturation concentration for tap water at field
temperature, pressure, and salt concentration to the
saturation for tap water at field temperature, pressure,
and zero salt concentration.
Dissolved organics and dissolved gases can also interfere
with the rate at which oxygen passes through the
membrane. When using a probe to measure DO,
one must be aware of the sensitivities of the instrument
to the waste constituents present in the sample. The
probe manufacturer's literature should be referred to
prior to calibration and making measurements. If the
literature does not contain information on interferences,
then the oxygen analyzer manufacturer should be
consulted.
Considerations for a Standard
The complete lack of concensus regarding the signifi-
cance and influence of the factors affecting alpha and
beta determinations presented in this paper demon-
strates the need for standardized procedures for
determination of alpha and beta factors. Even with
standardized procedures, however, the factors may not be
valid at the worst, or accurate enough for performance
test evaluation at best. Keegan and Busch (25) judged
the determination of alpha as neither a valid nor effective
procedure. The process Equipment Manufacturer's
Figure 9. Change in Surface Tension During Reaeration Testing with Surfactants Present
80
o
o
'35
o>
70
8 60
<0
t
3
C/3
50
-5 mg/l of Richonate 40B Added at Beginning of the Test
—©
10
20 30 40
Aeration Time (min)
158
50
60
-------
Figure 10. Considerations for Alpha and Beta Measurements
Purpose for Conducting Alpha and Beta Tests
Design New
System
Augment Existing
System
I
Wastewater
Identification
I
Evaluate Existing
System
Select a Bench-Scale Aeration Device
Surface Submerged Two-Phase Statjc Diffused
Mechanical Turbine Jet
^\^ \
/ ^^^
Establish Full-Scale Clean Water KLa
Determine Bench-Scale Clean Water Mixing
Level to Produce Full-Scale Clean Water KLa
Conduct Alpha and Beta Tests
I Repeat Tests for:
I
a*nna «f Different Rannp nf Different
KRL vaues Wastewater TeSatules *«*>" Reproducibility
L Streams Device
159
-------
Table 5. Summary of Oxygen Saturation
Measurements1 (mg/l)
Measurement
Technique
Distilled
Water
Solution Strength
1% 2%
NaCI @ 20.5°C
YSI 572 9.1 9.0 (8.5)3 8.55 (7.5)3
WS3304 9.1 9.1(8.65) 8.7
Winkler 1 9.1 8.6 7.75
Winkler2 9.1 8.7 7.85
Na2S04 @ 20.5°C
YSI 57 9.1 9.4(8.7) 9.4(8.2)
WS 330 9.1 9.4(8.7) 9.4(8.2)
Winkler 1 9.1 8.7 8.3
Winkler 2 9.1 8.7 8.2
(NH4)2S04@ 19.0°C
YSI 57 9.4 9.5 (8.9) 9.1 (8.0)
WS 330 9.4 9.5 —5 9.1 —
Winkler 1 9.4 2.2 —
Winkler 2 9.4 1.2 —
C H1206 H2O (Dextrose) @ 21.5°C
YSI 57 8.9 8.8 8.7
WS 330 8.9 8.8 8.5
Winkler 1 8.9 8.0 8.0
Winkler 2 8.9 8.0 8.0
i Temperatures vary by test, but not within a test.
2 Yellow Springs instrument (21).
3 ( ) indicates corrected measurement.
4 Weston and Stack instrument (20).
5 No measurement taken.
Association (1) concluded that it is extremely difficult to
accurately determine alpha, making the steady state test
unacceptable as a standard method for rating aeration
equipment or determining compliance of an aeration
device with a performance guarantee. Similar conclu-
sions were drawn by Kalinske (24).
In spite of the problems associated with the measure-
ment of alpha and beta, several investigators (4) (16) (33)
(40) have recommended sound procedures to follow,
emphasizing those factors which affect measurement the
most. Eckenfelder and Ford (17) describe in detail a test
method for determining alpha and beta in the supernate
of mixed liquor. Nogaj and Hurwitz (33) presented a
method for determining alpha and beta in the supernatant
without determining the oxygen uptake rate of the mixed
liquor. Bass and Shell (4) describe a detailed step-by-step
alpha measurement procedure with emphasis on five
significant factors:
—Waste sample composition
—Type of aeration device used
—Test with suspended solids present
—Test at design temperature
—Test at design mixing level
Although Standard Methods (38) has a tentative pro-
cedure for determination of alpha and beta (207B5b and
c), much work remains to be accomplished in the
standardization of alpha and beta factor measurement
procedures. Several key factors concerning measurement
procedures are contained in Figure 10. Before detailed
procedures can be developed, the major questions such
as: 1) What defines equivalent mixing level? 2) What
type of aeration device should be used? 3) Should sur-
factants be used in alpha testing and why? and 4) Are
suspended solids in or out? must be addressed and
answered. Only after a concensus has been reached
regarding these major factors in alpha and beta measure-
ments, can well defined standard procedures be adopted.
References Cited
1. "Aeration in Wastewater Treatment, WPCF Manual of
Practice No. 5." Water Pollution Control Federation,
Washington DC, 1971.
2. Aiba, S. and Toda, K. "The Effect of Surface Active
Agents on Oxygen Absorption in Bubble Aeration."
General Applied Microbiology, 9, 1963.
3. Albettson, O.E. and DiGregorio, D. "Biologically
Mediated Inconsistencies in Aeration Equipment
Performance." Journal Water Pollution Control
Federation, 47, 1975.
4. Bass, S.J. and Shell, G.L "Evaluation of Oxygen
Transfer Coefficients of Complex Wastewaters".
Proceedings of the 32nd Industrial Waste Conference,
Purdue University, West Lafayette IN, 1977.
5. Berk, W.L, Lad, D.J., Houck, D.H., and Roeber, J.A.
"Recommended Practices in Testing of Aeration
and Oxygenation Devices." Process Equipment
Manufacturers Association, Technical Committee
Report, May 24, 1972.
6. Casey, T.J. and Karmo, O.T. 'The Influence of
Suspended Solids on Oxygen Transfer in Aeration
Systems." Department of Civil Engineering, Uni-
versity College, Dublin, 1973.
7. Contract Specifications, Millinocket, Maine Waste-
water Treatment Facilities, Camp, Dresser, & McKee,
Boston MA, 1976.
8. Conway, R.A. and Kumke, G.W. "Field Techniques
for Evaluating Aerators." Journal Sanitary Engineer-
ing Division, Proceedings of the American Society
of Civil Engineers, 89, SA2, 1966.
9. Damhaug, T. and Balmer, P. "On the Determination
of Alpha Factors." Vatten, 1, 1975.
10. Downing, A.L. "Aeration in the Activated Sludge
Process." Journal of the Institution of Public Health
Engineers, April 1960.
160
-------
11. Eckenfelder, Jr., W.W., Raymond, L.W., and Lauria,
D.T. "Effect of Various Organic Substances on
Oxygen Absorption Efficiency." Sewage and Indus-
trial Wastes, 28, 1956.
12. Eckenfelder, Jr., W.W. "Factors Affecting the
Aeration Efficiency of Industrial Wastes." Sewage
and Industrial Wastes, 31, (1959).
13. Eckenfelder, Jr., W.W. and O'Connor, D.J. "Biological
Waste Treatment". Pergamon Press, New York,
1961.
14. Eckenfelder, Jr., W.W. and Barnhart, E.G. "Per-
formance of a High Rate Trickling Filter Using
Selected Media." Journal Water Pollution Control
Federation, 35, 1963.
15. Eckenfelder, Jr., W.W. and Ford, D.L "Engineering
Aspects of Surface Aeration Design." Proceedings of
the 22nd Industrial Waste Conference, Purdue
University, West Lafayette IN, 1967.
16. Eckenfelder, Jr., W.W. and Ford, D.L. "New Con-
cepts in Oxygen Transfer and Aeration." Advances in
Water Quality Improvements, University of Texas
Press, 1968.
17. Eckenfelder, Jr., W.W. and Ford, D.L. "Water Pollu-
tion Control, Experimental Procedures for Process
Design". Pemberton Press, New York, 1970.
18. Gilbert, R.G. and Chen, S.J. "Testing for Oxygen
Transfer Efficiency in a Full-Scale Deep Tank".
Proceedings of the 31st Industrial Waste Conference,
Purdue University, West Lafayette IN, 1976.
19. Howland, W.E. "Effect of Temperature on Sewage
Treatment Processes." Sewage and Industrial
Wastes, 25, 1953.
20. Instructions for W & S 330 Dissolved Oxygen Probes,
Weston & Stack, Inc., Malvern PA.
21. Instructions for YSI 5700 Series Dissolved Oxygen
Probes, Yellow Springs Instrument Co., Inc. Yellow
Springs OH.
22. Jackson, M.L. et al, "Deep Tank-Flotation Biological
Treatment: Groundwood Paper Mill Wastewater."
Presented at the Pacific Northwest Pollution Control
Association Meeting, Seattle WA, October 1976.
23. Kalinske, A.A. "Economic Evaluation of Aerator
Systems." Environmental Science & Technology, 3,
1969.
24. Kalinske, A.A. Discussion, "Biologically Mediated
Inconsistencies in Aeration Equipment Performance."
Journal Water Pollution Control Federation, 47,
1975.
25. Keegan, R.T. and Busch, A.W. Discussion, "Speci-
fying and Evaluating Aeration Equipment." Journal
Sanitary Engineering Division, Proceedings of the
American Society of Civil Engineers, 94, SA4, 1968.
26. King, H.R. "Mechanics of Oxygen Absorption in
Spiral Flow Aeration Tanks, Experimental Work".
Sewage & Industrial Wastes, 27, 1955..
27. Mancy, K.H. and Okun, D.A. "The Effects of Surface
Active Agents on Bubble Aeration." Journal Water
Pollution Control Federation, 32, 1960.
28. Mancy, K.H. and Barlage, Jr., W.E. "Mechanism of
Interference of Surface Active Agents with Gas
Transfer in Aeration Systems." Advances in Water
Quality Improvement, University of Texas Press,
1968.
29. Marotte, F.K. Private Communication. CH2M-H1II,
Bellevue WA, 1977.
30. Matson, J.V., Bennett, G.F., and Matson, M.L.,
Discussion, "Biologically Mediated Inconsistencies in
Aeration Equipment Performance." Journal Water
Pollution Control Federation, 48, 1976.
31. McKeown, J.J. and Okun, D.A. "Effects of Surface
Active Agents on Oxygen Bubble Characteristics."
In Advances in Biological Wastewater Treatment,
(W.W. Eckenfelder Jr. and B.J. McCabel eds.),
Pergamon Press, New York, 1963.
32. Metcalf & Eddy, Inc. "Wastewater Engineering".
McGraw-Hill, New York NY, 1972.
33. Nogaj, RJ. and Hurwitz, E. "Determination of
Aerator Efficiency Under Process Conditions." Pro-
ceedings of the 23rd Industrial Waste
Conference, Purdue University, West Lafayette IN,
1968.
34. O'Connor, D.J. and Dobbins, W. "The Mechanics of
Reaeration in Natural Streams." Journal Sanitary
Engineering Division, Proceedings of the American
Society of Civil Engineers, 82, SA6, 1956.
35. Otoski, Robert. Master's Thesis. Tufts University
Graduate School, Medford MA, 1978.
36. Reeves, J.R. "Membrane Electrode Techniques
Measure Dissolved Oxygen." Water & Sewage
Works, February 1976.
37. Schmit, F.L., Wren, J.D., and Redmon, D.T. "The
Effect of Tank Dimensions and Diffuser Placement on
Oxygen Transfer." Presented at the New England
Water Pollution Control Association Meeting,
Dixville Notch NH, June 1976.
38. "Standard Methods for the Examination of Water and
Wastewater". 14th Ed., American Public Health
Association, Washington DC, 1975.
39. Streeter, H.W. et al. "Measures of Natural Oxidation
in Polluted Streams." Sewage Works Journal, 8,
1936.
40. Stukenberg, J.R., Wabeh, V.N., and McKinney, R.E.
"Experiences in Evaluating and Specifying Aeration
Equipment." Journal Water Pollution Control Federa-
tion, 49, 1977.
161
-------
41. Timson, W.J. and Dunn, C.G. "Mechanism of Gas Engineering Conference, University of Kansas,
Absorption from Bubble Under Shear." Journal Lawrence KA, 1971.
Industrial Engineering Chemistry, 52, 1960. ^ Wj|ke CR and Chapg p »Corre|ation of Djffusjon
42. Water Pollution Research Laboratory. "Performance Coefficients in Dilute Solutions." Journal American
of D iff users in Mixed Liquor." WRP Report 390P, Institute of Chemical Engineers, 1, 1955.
May 1972. 45. Zlokarnik, M. Private Communication. Bayer,
43. Weis, F.S. and Lad, D.J. "Development of a New Leverkusen, W. Germany.
Mechanical Aerator." Presented at the 21st Sanitary
162
-------
Surface Aeration Equipment:
Field Performance Testing Vs
Shop Performance Testing
John R. Stukenberg and Valery N. Wahbeh
Black & Veatch
Kansas City MO 64114
The recent trend in surface aerator testing has been to
conduct performance tests, i.e., oxygen transfer tests, in
the manufacturer's test tanks. Several manufacturers
have spent hundreds of thousands of dollars to develop
shop testing facilities. Although they were constructed
primarily to enable development and refinement of aera-
tion equipment, the considerable cost involved forced
equipment manufacturers to seek additional ways to
utilize these facilities. It was only logical that the
manufacturers would want to use shop facilities for the
aerator performance testing often required by engineering
specifications. It was reasoned that shop tests would
reduce the cost of testing and consequently the owner's
cost for the equipment, as well as provide controlled
conditions for the performance tests. The idea caught on
and many specifications allowed performance tests to be
conducted in the shop. However, experience has proven
that shop testing is not without its drawbacks. This paper
will be devoted to a discussion of the problems associated
with shop testing of surface aerators.
Problems Detected
Shop testing of surface aeration equipment to determine
oxygen transfer capabilities has been allowed in recent
years by the authors' firm, with the added requirement
that tests of the physical aspects of the equipment be
made after installation in the field. Typically, the specifi-
cations pertaining to the physical aspects of the aeration
equipment have included:
1. Adequacy of Baffling. Baffling shall be provided as
required to prevent interference between aerators and
to satisfy specified power tests. If the aerators fail to
satisfy performance requirements, additional baffling
shall be installed or modifications made to the
initial baffling and the tests repeated until
satisfactory results are obtained.
2. Power Testing. Aerators shall be tested for power
consumption throughout their operating range, with
frequent checks for the influence of adjacent units and
the adequacy of baffling. Power measurements shall
be made with a recording wattmeter. With all units
operating, the transient power consumption of each
aerator shall not vary more than ±5% from the average
value and shall be unaffected by the number of units
operating in the basin.
3. Mist and Spray Protection. Under all operating condi-
tions, liquid or spray shall not be thrown outside the
basins nor shall mist or fine spray be created which
19
could cause moisture or ice to collect on adjacent
equipment, walkways or other structures.
Shop performance tests are generally assumed to be
conducted under more controlled conditions than field
tests. In an attempt to simulate field conditions, shop test
tanks are supposed to be approximately equal in volume
to the aeration basin volume influenced by one aerator.
Yet, there are often considerable differences between the
results of shop tests and field tests. This has been
particularly true of power consumption, where the power
consumption of aerators as installed in the field has often
exceeded the limitations set forth in the specifications
and occasionally exceeded the rating of the motors. The
only recourse to correct power consumption in the field
is to modify the aerator impeller or adjust the impeller
submergence. If the oxygen transfer testing was con-
ducted in the manufacturer's test tank, the aerating
capability of the field modified equipment will be
unknown unless additional oxygen tests are conducted
in the field.
Field tests have also pointed out the need for added
baffling in the aeration basins, in addition to that used in
shop tests or in addition to that anticipated by the manu-
facturer. In one field test consistent and reliable oxygen
transfer data could not be taken until the baffling on the
basin was corrected. Mist and spray problems have also
been obvious in the field during non-steady state testing
where shop tests failed to indicate such a problem. The
mist and spray developed by a group of aerators in an
aeration basin cannot be duplicated by a single aerator in
a test tank. Also, the mist and spray is much more
obvious after an aeration basin is put into service than
during aerator tests with clean water.
Case Histories
Six different installations in which the field test results
differed significantly from the shop test results or from
the manufacturer's expected results based on earlier
shop tests are summarized in Tables 1 and 2. Equipment
installed in these projects was supplied by five different
manufacturers. Three different basic aerator impeller
designs were used. Draft tubes were used in all of these
installations, and all impellers were designed to be
sensitive to variations in liquid level. All performance
tests were conducted with the non-steady state method.
Installation A. Several problems were encountered with
this installation, most notably; 1) a variation in aerator
163
-------
Table 1. Characteristics of Aerator Installations
Impeller Characteristics
Number
Installation Aerators
Code per Basin
A 3
B 4
C 4
D 9
E 9
F 4
Aerator
Size,
kW
(hp)
130
(175)
45
(60)
75
(100)
75
(100)
75
(100)
130
(175)
Type
Cone
Cone
Cone
Flat(2)
Flat
(3)
Diameter,
m
(ft)
2.84
(9.33)
1.78
(5.83)
2.29
(7.50)
2.62
(8.58)
2.74
(9.00)
3.00
(9.83)
No. of
Blades
8
12
8
6
8
16
Maximum
Rotational
Speed,
rpm
37
56
47
45
47
42
Volumetric
Energy
Inputd),
kW/1000m3
(bhp/IOOOft3)
82.1
(3.12)
51.3
(1.95)
67.4
(2.56)
50.8
(1.93)
70.8
(2.69)
73.5
(2.79)
(1 (Maximum motor output per unit aeration basin volume.
(2)Vertical flat blades, curved or backswept.
(3)Partially closed cone.
Table 2. Basin Dimensions and Volumes
Shop Test Basin
Installation
Code Length,
m
(ft)
A 10.9
(35.8)
B 13.7
(45.0)
C 13.7
(45.0)
D 18.3
(60.0)
E —
F —
Width, Dia.,
m m
(ft) (ft)
10.7 —
(35.0)
13.7 -
(45.0)
13.7 -
(45.0)
15.2 —
(50.0)
— 21.3
(70.0)
— __
Depth,
m
(ft)
9.1
(30.0)
5.0
(16.3)
4.9
(16.0)
6.4
(21.0)
4.3
(14.2)
—
Volume,
m3
(ft3)
908
(32,060)
933
(32,930)
918
(32,400)
1,784
(63,000)
1,548
(54,650)
—
Length,
m
(ft)
32.0
(105.0)
25.3
(83.0)
29.0
(95.0)
39.6
(130.0)
41.1
(135.0)
32.0
(105.0)
Aeration Basin
Width.
m
(ft)
10.9
(35.8)
25.3
(83.0)
23.5
(77.0)
39.6
(130.0)
41.1
(135.0)
31.7
(104.0)
Depth,
m
(ft)
9.1
(30.0)
4.7
(15.5)
4.9
(16.0)
6.1
(20.0)
5.2
(17.2)
5.8
(19.0)
Volume per
Aerator,
m3
(ft3)
908
(32,060)
756
(26.700)
829
(29,260)
1,063
(37,550)
986
(34,830)
1,469
(51,870)
164
-------
power consumption relative to impeller submergence
from shop test conditions to field conditions and 2) the
installation of different equipment in the field from that
tested at the shop.
In the shop test, one aerator was tested in a tank with
exactly the same effective dimensions as the zone
influenced by one aerator in the aeration basin. The
proposed aeration equipment including a 4.03 m
(11.83 ft) draft tube, was installed in the tank for oxygen
transfer testing. During pretests, however, the aerator
design was found to cause excessive spray. Modification
of the aerator to correct the problem resulted in a smaller
diameter impeller. The modified impeller was tested at
46 and 34 rpm, with the original 4.03 m diameter draft
tube in the basin. Final performance tests on this
equipment configuration resulted in the data shown in
Figure 1.
The shop test results indicated that the aerator trans-
ferred oxygen in excess of that required. The
manufacturer suggested that the impeller speed be
reduced to reduce power consumption through the two
speed ranges. Estimates of the performance of the
impeller at the reduced speeds of 37 and 27.8 rpm were
made by the manufacturer based on the results of the
shop test. These curves are also shown in Figure 1.
The aerators were installed in the field to operate at the
reduced speed. The diameter of the draft tubes installed
in the field was 2.84 m (9.33 ft), which is 0.76 m (2.5 ft)
smaller than used in the shop test. The manufacturer
had routinely reduced the draft tube diameter to coincide
with the reduction in impeller diameter. The units were
tested in the field for power consumption. The results of
the first test are shown in Figure 2, along with the
power curves estimated by the manufacturer. Closer
inspection of the field-installed aerators revealed that the
impellers as installed were different from the unit tested
in the shop. Modifications were made to all the impellers
in one aeration basin, and the units were retested. The
results of this second field test are also shown in
Figure 2. The resulting values were still approximately
18% higher than predicted by the shop tests. Further
modifications to the impeller to reduce the power
consumption in the field are still being considered.
It was expected that, since the shop test tank was an
exact duplication of the zone of influence of one aerator
in the field, the shop test results would directly relate to
the field test results. This does not appear to be the
case. However, the effect of a different draft tube
diameter on the power consumption and the accuracy of
the manufacturer's estimated power curve is unknown.
It is apparent that with further modification the impeller
can be made to approximate the power consumption
measured during the shop tests. The effect of the
different draft tube diameter and a possible impeller
modification on the oxygen transfer characteristics of the
aerator will remain unknown, however, without sub-
sequent field performance testing.
Installation B. The principle problem encountered with
this installation was power consumption. The power
consumption in the field was approximately 35% more
per aerator than measured in the shop tests at the same
submergence. The power consumption curves from the
two tests are indicated in Figure 3. In order to correct the
problem, the manufacturer reduced the impeller blade
depth from 152 mm (6 in.) to 127 mm (5 in.) and raised
the aerator 51 mm (2 in.) out of the water. Subsequent
testing indicated that the modifications affected power
consumption as desired, but also caused a power varia-
tion in excess of 5% from the average. To correct this,
the draft tubes were raised 178 mm (7 in.) to provide
better baffling of the aerators.
Modification of the aerators and draft tubes resulted in a
power draw at high speed very similar to that observed in
the shop tests, as shown in Figure 3. However, the
modifications lowered the low speed power consumption
well below the shop test results. With the 229 mm (9 in.)
liquid level variation specified for the basin, it is possible
that the power consumption will be continuous through-
out the range of liquid level variation and speed change.
It is doubtful, however, whether the unit could be
operated at much less than minus 127 mm (5 in.) sub-
mergence (127 mm emergence) at low speed.
How the modifications affected the oxygen transfer
characteristics of the aerators is unknown. During the
course of the modifications, the manufacturer supplied
limited data to demonstrate the independence of oxygen
transfer from the impeller blade depth. While this
conclusion appeared possible, had field performance
testing been specified, the oxygen transfer capabilities of
the aerators would have been ascertained more definitely.
Installation C. An increase in power consumption from
shop conditions to field conditions was also encountered
in this installation. The impeller blades were trimmed in
the field from 267 mm (10.5 in.) used in the shop tests to
203 mm (8 in.) in order to simulate shop power consump-
tion. Power curves for the aerators from both shop and
field tests are shown in Figure 4. Again, trimming the
impeller had a greater effect on low speed power
consumption than on the high speed power consumption
but the power draw remained continuous over the entire
range of liquid level and speed changes.
Oxygen transfer tests were performed in the shop tests
and the field tests; the results of both are presented in
Figure 5. At high speed, the oxygen transfer character-
istics of the modified aerators installed in the field were
similar to those of the original unit tested in the shop. At
low speed, the field units exhibited a somewhat lower
transfer capability than did the original unit in the shop.
At both speeds, the aerators as installed in the field
appear to be slightly more efficient than the original unit
165
-------
Figure 1. Power Consumption by Aerator A — Shop
en
O)
bhp kW
160-T120
140-
120
100-
Q.
O 80
*
<5
60J
20-
110
100
90
- 80
70
60
- 50
- 40
40-- 30
- 20
- 10
O-1- 0
-25
h
-1
Shop
X— —— — X Estimated Field Performance
X 27.8 rpm
1
1
1
1
1
1
1
« F i
456
Impeller Submergence
i
7
46 rpm
1
25 50 75 100 125 150 175 200 225 250 275 300mm
I i i i i I i I i i i i
8
10
11
12 in.
-------
Figure 2. Power Consumption by Aerator A — Field
O)
bhp
120-
100-
80-
3
Q.
3 60-
O
0
5 40-
20-
0-
kW
-90
-80
-70
-60
-50
-40
-30
-20
-10
- 0
-7
|
-3
-
~ ^«M— «M««»0 EstirndtGd FiGlu P6rforrn3nc6 -^^"^
X X 1st Field Test ^^*~^~~
©—••—© 2nd Field Test _____.-—• — *"""" .^--0 37 rpm
*" ~ " ' «
..-©• ' '" ' ^*—
•••• ""^ • '
_® • * x
«__«.— — — — ""—"- 27.8 rpm
•' *
-
1 1 1 1 1 1 I 1 1 1 1 1
5 -50 -25 0 25 50 75 100 125 150 175 200 225 250mm
I i i i i i i 1 i i i i i
i I ' i i i i I I i i i i
-2-1 0 12 345 6 7 8 9 10 in.
Impeller Submergence
-------
Figure 3. Power Consumption by Aerator B
bhp kW
r55
70i
60-
50-
a 40-
30,
10-
-50
-45
-40
-35
^30
-25
-20
20--15
-10
\~ 5
0-L 0
-• Shop Test
X X 1st Field Test
O __o 2nd Field Test
—©
I
I
-250 -225 -200 -175 -15O -125 -1OO
-75
-50
-10
-9
i
-7
-25
•4-
o
4-
-5 -4 -3
Impeller Submergence
-2
-1
42 rpm
I
25
-H
50 75 mm
_i L
1
3 in.
-------
Figure 4. Power Consumption by Aerator C
CD
CO
bhp kW
r70
90-
80-
70-
60
Q.
O
o
40-
30-
20-
10-
-65
-60
-55
-50
45
-40
-35
-30
-25
-20
15
10
- 5
o-1- o
r
-12
•• Shop Test — 267-mm Blade
X X Field Test — 203-mm Blade
47 rpm
35 rpm
^ -*-
i——*""
J.
J.
J.
1.
-300 -275 -250 -225 -200 -175 -150 -125 -100
J J ^ rJ rl rA-
—r-1 i-1 r1 r1 rj T T
-11 -10 -9 -8 -7 -6 -5
Impeller Submergence
T
-75
J_
-50
"T
-3
-25
J_
-2
-1
25 mm
1 in.
-------
Figure 5. Oxygen Transfer Characteristics of Aerator C
Ib/hr
300-
250-
T3
| 200
c
(D
g 150
X
O
100-
50-
0 -1-
kg/hr
1 OU
-125
-100
- 75
- 50
- 25
• • •• Shop Test — 267-mm Blade
~ X X Field Test — 203-mm Blade x _
X ^ 9
.--"""TL^-1 — ' ""
<^f~*'^^** — X""*^ 47 rpm
^e**^
^-Je*f*~
•v-"""' ______—- r^--» 36r^"
•~ZZZZ-- ""
- x— "" "~
I 1 ( I 1 1 1 1 1 1 1 1
-275 -250 -225 -200 -175 -150 -125 -100 -75 -50 -25 0 25 50mm
i i i i i i i i i i i i i i
1 1 1 1 I i | i i 1 1 1 1 I
-11 -10 -9 -8 -7 -6 -5 -4-3-2-1 0 1 2 in.
Impeller Submergence
-------
in the shop tests. However, the difference may be the
result of the temperature correction factor. The shop test
was conducted at approximately 27°C liquid temperature,
whereas, the field test was conducted at approximately
17°C liquid temperature.
Installation D. Here again, field power consumption was
found to be greater than the shop power consumption at
equal impeller submergence. As a result of this problem
two field power tests were conducted, the second follow-
ing modification of the aerator installation. The differences
between the shop and initial field power test values were
greater than in any other installation known to the
authors. As much as 50% more power was consumed by
the aerators in the field (when all nine units were
operating) than was observed in the shop test at the
same impeller submergence. These data are presented in
Figure 6, as are the results from operating only the
aerator in the center of the aeration basin. Even greater
power consumption was observed in this mode of
operation.
In order to reduce power consumption to the specified
range, the submergence of the impellers was reduced by
raising the aerators approximately 150 mm (5 in.). The
draft tubes were raised a similar amount to maintain the
same relationship between the impeller and draft tube as
used in the shop test. No other modifications were made
to the aerators. Power tests were again conducted, as
were oxygen transfer tests. The results of the power tests
are shown in Figure 7. The power consumption was
reduced as desired, and the slope of the curve remained
relatively unchanged by the modification of the impeller
submergence. Although the manufacturer assured the
owner and engineer that the oxygen transfer efficiency
would be unaffected by the change in impeller sub-
mergence, a 10 to 15% decrease in transfer efficiency
from shop test results was observed, as indicated in
Figure 8. Had the specifications required a transfer
efficiency of 0.5 kg oxygen/MJ (3.0 Ib oxygen/bhp-hr),
an efficiency achievable by most aerators currently
manufactured, it is doubtful whether the equipment
would have passed the oxygen transfer test following
modification.
Installation E. One final example of change in power
draw characteristics from shop to field conditions is
provided by installation E. While the previously discussed
installations all exhibited an increase in power consump-
tion when the aerators were installed in the field, the
aerators at installation E drew less power per aerator at
a given submergence in the field than in the shop test
tank. To rectify this problem, the manufacturer lowered
the impellers 89 mm (3-1 /2 in.) into the water by placing
spacers in the drive shaft couplings and adjusting the
jacking studs on the gear boxes. As shown in Figure 9,
the power consumption characteristics of the installed
aerators as determined by two field tests were the same
as those of the unit tested in the shop, except for the
difference in impeller submergence.
The results of oxygen transfer tests conducted on the
aerator are shown in Figure 10. The transfer efficiency of
the aerators as installed in the field was slightly greater
than the shop test efficiency, particularly at low liquid
levels. However, a substantial difference in test water
temperature may have also affected these results. The
shop tests were conducted at a water temperature of
5-1/2°C, while the field tests were conducted at 25 to
27°C. In general, the oxygen transfer characteristics of
the aerators were virtually the same in the field and in
the shop.
Installation F. Specific shop tests were not conducted for
this installation. Field results were predicated on earlier
shop tests. Field testing indicated that the aerators
probably could transfer the required quantity of oxygen
into the water. However, the basin baffling provided by
the draft tubes was not sufficient to prevent surging of
the aerators, which in turn affected the oxygen transfer
characteristics of the units. The surging was probably the
result of interaction of the aerators, a condition not
simulated in shop tests.
Drifting of aerator power consumption prior to field
modifications is shown in Figure 11. The variation in
power consumption from the average was approximately
±11.4% at high speed. Under these conditions, the
aerators were unable to transfer the specified amount of
oxygen. Baffling of the basin resulted in a reduction of
the power variation of the aerators to ±3.3% at high
speed. The baffling was not as effective on low speed
power variations, although the variance was held to
approximately ±10%. With the installation of the baffles,
the aerators satisfied the specified oxygen transfer
requirements, apparently as a result of more efficient use
of power. If performance testing had been by shop tests,
it is doubtful whether the oxygen transfer capabilities of
the units in the field would have been known for certain.
In this case, performance testing was not conducted until
all modifications had been made so that the results truly
represented the capabilities of the installed aerators.
Discussion
From the six cases presented, it is evident that the
performance of a surface aerator installed in an aeration
basin with as many as eight other aerators is not the
same as when it is operated in the isolated conditions
of a shop test tank. In most of the examples cited, the
power draw of the aerators at a given impeller sub-
mergence increased when the units were installed in the
field. Installation E demonstrated that the change in
power consumption can just as easily be downward.
The reason why field power consumption is often
different from shop power consumption is not readily
apparent. It does not appear to be related to impeller
design or manufacturer. The largest change in power
consumption was exhibited by the aerators of installation
D, which also had the largest difference between test
171
-------
Figure 6. Power Consumption by Aerator D — Shop
N>
bhp
160-
140-
120-
100-
I
0 80-
2
i
60-
20-
kW
-120
-110
-100
- 90
- 80
- 70
- 60
- 50
- 40
- 3O
*J\J
- 20
- 10
-3
1
-13
©
—
© ^-"^X
© ^~~"
X ^^^"^
X ^^-"
~_ J'"'"" _^--— ~~^^
^~'~ x x ^^^^"^
— ^-*'"© ^ —
""'*'* ^*^~~
^ ^ ^ r*\ ^ ^—~+
W^^^^^^^^^^ OfH/P 1 COL
X— — ~""*X Field Test — 9 Units
© 0 Field Test — 1 Unit
i i i i i I 1 1 1 1 1 1
25 -300 -275 -250 -225 -200 -175 -150 -125 -100 -75 -50 -25 On
i i i i l I 1 i i i 1 I t
liiiii iilili
-12 -11 -10 -9-8-7 -6-5-4 -3 -2 -1 0
im
in.
Impeller Submergence
-------
Figure 7. Power Consumption by Aerator D — Field
GO
bhp
120-
100-
80-
^
1
0 60-
|
40-
20-
0-
kW
1 VA/
- 90
- 80
- 70
- 60
- 50
- 40
- 30
- 20
- 10
1 1 1 1 1 1 1 1 1 X< 1 1
X X 1 st Field Test — 9 Units ,,'
~ 0 o 2nd Field Test — 9 Units ,'
s
s
* /
/'
s
s
— S —
'' 'x x
s ..S"
^../6
0^
— —
- —
1 I 1 1 I 1 1 I 1 1 1 i
-550 -500 -450 -400 -350 -300 -250 -200 -150 -100 -50 0 50 100
1 i i i i i i i i i i i i
mm
-20 -18 -16 -14 -12 -10-8-6-4-2 0 24 in.
Impeller Submergence
-------
Figure 8. Oxygen Transfer Efficiency of Aerator D
Ib/bhp-hr kg/MJ
3.5-
3.0-
2.5-
S 2.0-
Q.
'S
o
3 1 5-
o "
1.0-
0.5-
0_
- v.o
- 0.5
-0.4
-0.3
-0.2
-0.1
r\
I 1 I I I I I
0
*•• — o— o
0
©
— —
— — .
— —
• • Shop Test
0 0 2nd Field Test -
1 1 1 1 1 1 1
III 1 1 1 1
0 10 20 30 40 50 60 70 80 kW
1 1 1 1 1 I I i
i i i i i i i i ' i i
0 1O 20 30 40 50 60 70 80 90 100 110bh
Oxygen Transfer Efficiency
-------
Figure 9. Power Consumption by Aerator E
bhp kW
r70
90-
-65
80-
70-
-60
-55
-50
60- -
o. 50-
D
O
I 40
30-
20- -
10-
45
-40
-35
4
30
25
20
15
10
5
O-1- 0
©
0 9 Shop Test
X X 1 st Field Test
© © 2nd Field Test
I
I
I
I
I
-250 -225 -200
-175
J
-150
J
-125 -100
J J-
-75
-50
L_
-25
r r r r r r r
-10 -9 -8 -7 -6 -5 -4 -3 -2
Impeller Submergence
0
-h
25
50
_L
-1
T
2
I
75 100mm
T
3
4 in.
-------
Figure 10. Oxygen Transfer Characteristics of Aerator E
o>
Ib/hr
300-
250-
200-
09
£ 150H
0)
0 100-
50-
kg/hr
r-150
-125
-100
- 75
- 50
- 25
O J- 0
-9
X
x-'
•• Shop Test
•X Combined Field Tests
I
I
I
J_
I
I
I
-225 -200 -175 -150 -125 -100 -75
-50
-25
25
-8
-7
-6
I
-5
r i i
-4 -3 -2
Impeller Submergence
50
•4-
-1
0
75 100mm
i i
4 in.
-------
Figure 11. Power Consumption by Aerator F
150
30
32
34 36 38
Approximate Time After Startup (min)
40
44
-------
tank volume and zone of influence volume in the aeration
basin. The test tank volume was 68% larger than the
zone of influence of one aerator. The power consumption
increased from shop to field conditions by 49%. Yet, there
does not appear to be a direct relationship between
volume change and power consumption change. There
was no change in the volume aerated between shop
tests and field tests of aerator A, but the power con-
sumption in the field was approximately 18% greater
than that recorded in the shop tests. A 9% reduction in
aeration volume resulted in a 33% increase in power
consumption between shop and field tests for aerator C,
while a 36% volume reduction from shop to field tests
resulted in a 22% reduction in power consumption for
aerator E.
Another possibility is that the change in power consump-
tion from shop to field conditions is a function of the
volumetric energy level of the aeration basins. The cases
reported all had a relatively high energy level, from 50.8
to 82.1 kW/1000 m3(1.93 to 3.12 bhp/1000ft3) based
on motor output. However, five other installations having
volumetric energy levels from 50.8 to 70.3 kW/1000 m3
(1.93 to 2.67 bhp/1OOO ft3) apparently performed as
expected, since no field modifications were required prior
to performance testing. Furthermore, at the time of this
writing, there were two other installation with volumetric
power levels of approximately 39.5 kW/1000 m3 (1.5
bhp/1000 ft3) known to be consuming more power in the
field than the shop tests indicated.
In addition to often failing to predict the performance of
an aerator in the field, shop tests can also be misleading.
Equipment installed in the field may not always resemble
the equipment or configurations used in the shop tests.
As was the case for most of the installations discussed,
the zone of influence for an aerator installed in the field
is seldom the same as for the configuration of the shop
test tank. This is usually understood by the engineer
before shop tests are conducted. A more subtle change
between shop and field conditions is modification of
equipment before field installation. While this is not
frequent, it has occurred. In the case of installation A, it
is not known how modification of the equipment may
have affected the performance of the equipment, or
whether it did affect it. It is known that there was differ-
ence in power consumption between the shop test and
field thest and that further tests are necessary.
Two aspects of aerator testing that cannot be determined
until the aerators are installed in the field are the inter-
action of the aerators and the mist and spray character-
istics of the aerators. In the case of installation F, it is
doubtful that the manufacturer expected power surging
problems. In most instances, the basin baffling provided
by draft tubes has been adequate to prevent surging.
However, the field performance tests on aerator F
demonstrated the need for additional baffling. Because
performance tests were conducted in the field, the
oxygen transfer characteristics of the aerators as installed
in the baffled aeration basin are known.
It is difficult to assess the mist and spray characteristics
of an aerator during shop tests, primarily because most
problems with misting result from the interaction of all
the aerators in a basin. Excessive mist and spray can
usually be detected during field tests either by moisture
on walkways and handrails, or if weather conditions
promote rapid evaporation of moisture, by the deposition
of chemicals from the test water on walkways, handrails
and aeration equipment. If necessary, conclusive albeit
messy proof of excessive mist and spray can be obtained
by operation of the aerators with mixed liquor in the
aeration basins. Regardless of the test for spray and mist.
if the performance test is conducted in the field, it is
possible to modify the aeration system to correct spray
and mist problems before the aeration tests are con-
ducted. In this manner, modifications that may affect the
oxygen transfer characteristics of the aerators, such as
lowering the mist shroud, will be included in results of
the performance test of the aerators.
Conclusions
The experiences with shop testing of surface aerators
where a single unit is tested, and field testing where all
the aerators in an aeration basin are tested have led to
the following conclusions:
1. Shop testing does not adequately predict the perform-
ance of aerators after they are installed in the field.
2. At a given impeller submergence, the power con-
sumption of surface aerators installed in the field may
vary as much as 50% from the power consumption
measured during shop tests.
3. The change in aerator power consumption character-
istics from shop tank to aeration basin is not
predictable, nor does it appear to be directly related
to the impeller design, the relationship of shop tank
volume to zone of influence volume in the aeration
basin, or the volumetric energy level in the aeration
basin.
4. Both impeller modification and alteration of impeller
submergence have been successfully used to correct
field power consumption of aerators to shop test
results.
5. There is evidence from two installations discussed
that oxygen transfer may be proportional to the power
consumption of the impeller, and slight modifications
of the impeller do not alter the oxygen transfer
efficiency of the impeller. A third installation, however,
exhibited a 15% decrease in oxygen transfer efficiency
as a result of modification of the aerator to correct
power consumption.
6. Equipment installed in the field is not always identical
to equipment tested in the shop.
178
-------
7. Surging of the power consumption of aerators and baffling and corrections to conditions resulting in
mist and spray characteristics of aerators can only be excessive mist and spray before performance testing.
determined in the field. Unlike shop tests, field performance tests represent
8. Field performance testing permits modification of the aeration equipment as installed in the aeration
impeller design or submergence, additions to basin Dasm.
179
-------
Section IV. Evaluation of Respiring Systems
Oxygen Transfer in
Closed Systems
James A. Mueller and Jack Famularo
Manhattan College
Bronx NY 10017
Thomas J. Mulligan
Hydroscience, Inc.
Emerson NJ 07630
Introduction
In aeration, two basic equations are used to describe the
oxygen transfer capabilities of a system. Equation 1,
Henry's Law, relates the equilibrium (saturation) concen-
tration of oxygen in solution, C*, to the partial pressure
of oxygen in the gas phase, p02, by:
(D C* = Hp02
where H is Henry's constant. Equation 2 relates the rate
of change of oxygen concentration in the liquid to the
driving force, the oxygen deficit, (C*-C):
20
(2)
dC/dt = KLa (C*-C)
where C is the actual dissolved oxygen concentration,
and K|_a, the oxygen transfer coefficient, is determined by
the type, size, and quantity of aeration devices.
When evaluating the transfer capabilities of open tank
surface aeration systems, the partial pressure of oxygen
in the atmosphere above the tank is constant. Thus, the
saturation value, C*, is constant and the analysis is
relatively straight forward. In diffused or turbine aeration
systems, the partial pressure of the oxygen changes as
bubbles rise from bottom to surface due both to static
pressure and 02 transfer; thus, the saturation value
differs from top to bottom of the tank. However, the
influent gas partial pressure is constant since atmos-
pheric air is utilized. Also, the buildup of other dissolved
gases in the aeration tank is relatively small since when
above saturation, they are stripped out of solution and
released to the atomosphere.
In evaluating the oxygen transfer capabilities of closed
aeration systems, whether surface or submerged aeration
equipment is used, buildup of other gases occurs in both
the liquid and gas phases, causing a reduction of the
oxygen mole fraction. The resulting oxygen partial
pressure in the gas phase cannot be predicted without
conducting mass balances on both gas and liquid phases
where consumption and production of all gases are
properly accounted for.
Model Development
Development and verification of this approach was
conducted initially for closed tank, staged, high purity
oxygen systems (1) where the above effects had signifi-
cant design implications. However, it is also applicable
to closed tank air aeration systems since the same
phenomena occur, only with a different influent gas
composition.
Three gas transfer mechanisms occur when pure oxygen
is introduced into a covered aeration tank: 1 ) oxygen is
transferred from the gas to the liquid phase, 2) nitrogen
originally present in the liquid phase is transferred to the
gas phase, and 3) carbon dioxide produced by the
biological reaction is transferred to the gas phase. In a
sealed aeration tank where gas is vented only from the
last stage, these latter two mechanisms cause the
oxygen gas phase partial pressure to decrease from stage
to stage.
Figure 1 presents a schematic of the closed system
analyzed in this paper, a staged system with a gas phase
flowing concurrently with the liquid phase. Both gas and
liquid phases are completely mixed for each stage with
gas vented only from the final stage. Two mass balance
equations are constructed for each parameter in the
system, one for the liquid phase and one for the gas
phase as follows:
Liquid Phase
(3)
dCj-n/dt =
Gas Phase
(4)
dPi n/dt =(Gn.lPj>n.1-GnPi>n)/V«
180
-------
Figure 1. Completely Mixed Series Reactor Schematic for
Concurrent Flow of Liquid and Gas Phases
Gn f
Gn-1
Gas
Phase
Liquid
Phase
a
a
"T
- Q
181
-------
where:
C,p = parameter concentration in the liquid and
partial pressure in the gas
v',Vg = liquid and gas volumes
Q,G = liquid and gas flow rates
R,T,M = gas constant, absolute temperature, and
molecular weight
i,n = species identification and reactor stage
location
R ,R9 = sources and sinks in the liquid and gas
phases
For each parameter of concern, both the liquid phase
concentration and gas phase partial pressure are
unknown, with Equations 3 and 4 available for each.
However, an additional unknown always exists in the gas
phase equation, the gas flow rate, G. The final equation,
defining a constant total pressure in the gas phase,
allows simultaneous solution of the equation set:
(5)
where:
EPi,n = P,-
p,,pv = total and vapor pressures of the gas phase
Because of the non linear nature of the gas phase equa-
tion, the equation set is normally numerically integrated
to steady state, which has proven a relatively simple,
convenient, and reliable technique. Development of a
steady state solution which requires less computer time
has been accomplished for the high purity oxygen aera-
tion system.
Table 1. Sources and Sinks for Closed System Model
Table 1 presents the source and sink terms used in
Equations 3 and 4, incorporating interphase gas transfer,
BOD oxidation, and endogenous respiration.
When significant alkalinity changes occur in solution due
to organic nitrogen hydrolysis, ammonia utilization for
growth, or nitrification, additional chemistry interactions
must be included to properly evaluate the dissolved C02
concentration (2).
The coefficients a' and b', as well as the SOD removal
coefficients, are obtained for typical bench- or pilot-scale
biological treatment studies similar to air aeration
systems
The CO2 coefficients, a2 and b2, are related to the oxygen
utilization coefficients by the respiratory quotient, RQ, as
{/•\H/\tAJC'
a2 = 1.38(RQXa')
b2 = 1.38(RQXb')
a2 = mg C02/mg BOD
a' = mg 0? mg BOD
b2 = mg C02/mg MLVSS-day
b' = mg 02/mg MLVSS-day
RQ = moles C02/mole 02
1.38 = (44mgC02/moleC02)/(32mg02/mole02)
where:
Species
02
C02
N2
BOD
Designation
Ci
C2
C3
s
Sources (+) and Sinks (-),
Interphase Gas BOO
Transfer* Oxidation
Liquid, ZR1
KLa OSH^-Ci) -a'kXyS
(KLa)2 (/3H2P2-C2) +a2kXvS
(KLa)3 3H3p3-C3)
-kX«S
(mg/l-hr)
Endogenous
Respiration
-b'Xv
+b2Xv
02
CO2
N2
Pi
P2
Pa
-KLa (/SH^-C!)
-(KLa)2(0H2p2-C2)
-KLa
DL = liquid phase diffusivity
k = first order reaction rate
Xv = volatile suspended solids concentration
182
-------
Calibration
Prior to model application, calibration of the kinetic
parameters in the model must be obtained. Calibration
of the high purity oxygen model has been obtained,
for both domestic and industrial wastewaters (1).
Figures 2 and 3, the model verification for the Batavia
Phase III data, show the typical decrease in gas flow and
oxygen partial pressure in successive stages. The be-
havior of the nitrogen concentration is interesting. In the
first stage of the aeration tank, nitrogen is stripped out of
solution into the gas phase, causing a drop in the liquid
phase nitrogen concentration. However, in the second
and third stages, the equilibrium shifts and nitrogen is
transferred back to the liquid phase, causing a rise in the
dissolved nitrogen concentration. Because of its high
solubility, the majority of the COa remains dissolved in
solution, yielding effluent concentrations greater than
250 mg/l and a resulting effluent pH of 6.3. In the
parallel air aeration system at Batavia, the pH in the
aeration tank remained at approximately 7.1, the same as
the raw wastewater, due to the continual stripping of CO
to the atmosphere.
The above model has also been applied to analysis of a
four-stage pure oxygen pilot plant treating an industrial
waste. Table 2 indicates that increasing F/M ratios
required decreasing respiratory quotients, RQ, to effec-
tively model the domestic and industrial waste data. This
is due to either changing waste characteristics, or to
incomplete oxidation at the higher loading rates.
Table 2. Effect of BOD Loading Rate on Respiratory
Quotient
F/M
Waste (Ib BOD/
Ib MLVSS-day)
RQ
(rnolesCO2/moleO2)
Domestic
Industrial
Industrial
0.55
0.87
1.38
1.00
0.85
0.60
Additional calibration of the chemical interactions was
more recently obtained on a pulp and paper mill waste-
water, as shown in Figures 4 and 5. An RQ of 0.90 was
used to fit the data. Although the ammonia nitrogen data
was scarce, a lower requirement for growth was indi-
cated by the data than typically expected for biological
systems. Due to secrecy agreements, only the relative
gas phase data are presented.
Model Application to System Design and Analysis
The above model framework is presently used for analysis
and design of closed tank high purity oxygen and air
aeration systems.
For high purity oxygen systems, the ability to evaluate
tradeoffs between mixer horsepower and pure oxygen
gas flow (02 production horsepower) is shown in Figure
6. In this case, KLa is related to N0 as follows:
(6) KLa= aN0Pv0/8.34C"20
where:
KLa = hr1
a = dimensionless
N0 = standard aeration efficiency, Ib O2/hp-hr
0(T-20) _ djmerisionless
pv = power level, hp/mil gal
For the closed system model, the effect of waste
characteristics on transfer rate, as well as temperature
effects, must be taken into account as in air aeation
systems. The saturation value must also be corrected for
depth if submerged turbine or diffused aeration systems
are used.
Equation 6 relates the required aeration horsepower to
the oxygen transfer coefficient used in the model and
provides the ability to evaluate the effect of horsepower
split (Oa generation and aeration) on total power required
for all operating conditions. To illustrate application of the
model it will be applied to treatment of a wastewater
from a chemical plant with the following conditions:
Q = 1 5.4 mgd
TOD = 1 144 mg/l at 24-hr peak load
F/M = 1.58lbTOD/lbMLVSS-dayat24hrpeakload
TOD = 608 mg/l at average daily load
F/M = 0.84 Ib TOD/lb MLVSS-day at average
daily load
Alkalinity = 500 mg/l, a' = 0.53, b' = 0.05 day1
pH = 10.3
Using a 2-1-1 tank configuration and an N0 of 3.0 Ib
02/hp-hr with 80% TOD removal, the design option of
the program was utilized to evaluate minimum horse-
power required to obtain a desired dissolved oxygen level.
Modern oxygen generation plants require between 0.88
and 1 .29 hp/scfm of oxygen.* Curve (a) in Figure 7 is a
design curve for the peak load of 1 144 mg TOD/I using
a power charge of 1 .25 hp/scfm for oxygen generation.
Portions of the curve at low aeration horsepower cor-
respond to high influent oxygen flow rate. As aeration
horsepower is increased, less oxygen flow is required to
maintain a DO of 2 mg/l and the percent oxygen utiliza-
tion represented in curve (b) increases. It is apparent that
*UNOX Brochure
183
-------
Figure 2. Gas Phase Model Calibration for Batavia, New York Pure Oxygen Treatment Plant
30
-^ 20
I
1
in
03
O
C> Phase ID Batavia Data, Average + Range
o
o
o
0>
(A
re
(A
re
-------
Figure 3. Liquid Phase Model Calibration for Batavia. New York Pure Oxygen Treatment Plant
7
a
5
300
200
~ 100
c
o
o
o
a.
0
12
Influent
Phase m Batavia Data, Average + Range
Alkalinity = 250 mg/l Raw Waste pH = 7.1
J 1
F/M = 0.55 Ib BOD/lb MLVSS-day
RQ = 1.0
Raw Waste BOD5 = 262 mg/l
Range of Average Measured Cb
Stage
185
-------
Figure 4. Pulp and Paper Pilot Study Liquid Phase Data High Purity Oxygen Model Calibration
400
_ 300.r
\
O)
Q 200
o
CO
100
400
« 200
100
15
- 10
o>
8 5
I
1234
Aeration Basin Stage
600
400
01
CM
o
o
o>
5 §
z
i
0)
u
'c
to
200
20
10
-------
Figure 5. Pulp and Paper Pilot Study Gas Phase Data, High Purity Oxygen Model Calibration
o
5
o
U
CO
in
a
in
to
O
o
S
o
o
a.
in
CO
O
0
CM
O
O
CO
O
1234
Aeration Basin Stage
1234
Aeration Basin Stage
187
-------
Figure 6. Effect of Rated Capacity and Gas Flow on Dissolved Oxygen Concentration and Oxygen Utilization
3
o
o
o
o
.2
I
5
c
I
100
90
80
70
60
50
0.75
.N0(lb02/hp-hr)
2
Stage
2
Stage
1.0 1.25
Gas Flow Ratio
Rated Capacity
Effect
Alkalinity = 100 mg/l
F/M = 0.9 day1
Gas Flow
Effect
J.
3
N0 = 3.2 Ib 02/hp-hr
F/M = 0.9 day'1
Alkalinity = 100 mg/l
1.5
188
-------
Figure 7. Treatment Power Consumption for Average and Peak Load Conditions
2,000
1,800
1,600
o
I
§ 1,400
C
CM
O
1,200
1,000
800
600
380
• Maximum Load
DO = 2 mg/l
I
(b)
«
400
Average Load
90% 02 Utilization
DO = 12-17 mg/l
I
420 440
Aeration Mixer Horsepower
460
00
90
80
70
60
50
40
30
480
189
-------
optimum operation corresponds to use of as little oxygen
as is possible, a direct consequence of the above power
charge. Total horsepower requirements for average
conditions, using the same mixer horsepower required
for the maximum conditions, are shown as curve (c) for
90% oxygen utilization.
In Figure 8, a design curve corresponding to maintenance
of 4 mg/l of dissolved oxygen is shown for the average
loading. The aeration horsepower of 250, adequate to
maintain a DO of 4 mg/l under average loading condi-
tions is unable to achieve a DO of 2 mg/l during periods
of peak loading. An examination of Figure 8 reveals that
at least 390 aeration horsepower is required to accom-
plish this.
Clearly, the peak demands of this system suggest the
need for a variable horsepower aeration system which
can be operated at 450 hp during peak load conditions
and reduced to 250 hp at average loading.
It is desirable, in certain instances, to cover conventional
air activated sludge systems. This may be necessary in
extremely cold climates to conserve heat in the system in
order to maintain good organic removal and if waste-
waters contain materials that are volatile and have the
tendency to strip from the wastes during aeration, thus,
causing potential odor problems to the surrounding
community.
As an example, covers were installed over the aeration
basins of a conventional air activated sludge system in
order to capture the odorous gases that were being
stripped from an industrial wastewater during treatment.
Two aeration basins of 0.7-mil gal capacity each had
three fixed-mounted 75 hp surface aerators. After
covering the basin, an exhaust system was installed to
collect off gases. Air now enters the basins by the slightly
negative pressure created by the exhaust system,
primarily through the openings where the aerator impel-
ler shafts pass through the cover.
The exhaust system had to be sized to provide the
required gas flow through the basins, so that the neces-
sary oxygen could be transferred to the liquid phase by
the existing aerators. The required gas flow was
determined to maintain 2 mg/l of dissolved oxygen in the
mixed liquor at the following maximum design loading
conditions:
Wastewater Flow = 0.54 mgd
COD = 32,000 Ib/day
Alkalinity = 2,500 mg/l
Amrnonia-N = 900 mg/l
Figure 9 presents the results of these analyses.
At gas flows of less than 3,000 scfm, basin
dissolved oxygen levels would be less than 1 mg/l. At air
flows of 7,500 scfm, the projected dissolved oxygen
concentrations would range from 1.5 to 2.0 mg/l. The
exhaust fan and duct work to collect the off gases were
sized for the 7,500 scfm air flow. The exhaust gases
were combined with other plant waste gas streams and
incinerated.
Sampling of the covered basin showed that relatively
uniform gas composition existed in the enclosed gas
space, with oxygen concentrations at various locations
ranging from approximately 18 to 19.2%. Basin dissolved
oxygen concentrations, at peak loadings, were approxi-
mately 1.5 mg/l.
In evaluating the desirability of covering any basin,
consideration should be given to the temperature
changes that will result from the covering. The tempera-
ture increase in many instances will be beneficial;
however, it may also cause problems. With the given
wastewaters which had relatively high temperatures
(30°C), the temperature in the aeration basin increased
from high temperatures of approximately 33°C to nearly
42°C. The impact of increased temperatures on the
biological activity is shown in Figure 10, where it can be
seen thct the reaction rate decreased significantly as
temperatures increased above 32°C.
Application to Steady State O2 Transfer Evaluation
For respiring systems, closed tank aeration can provide a
significant improvement over open tank systems in
evaluating K|_a since an additional phase, the gas phase,
is available for analysis. If accurate gas phase data are
collected, the total 62 utilization rate in the respiring
system is known, accurately providing one of the para-
meters difficult to evaluate for highly loaded systems with
high 02 uptake rates.
The steady state equations for KLa determination in each
stage of a multi-stage covered reactor are:
Liquid Phase
(7) KLa = [(Ci,n-C1.n-i)+r]/to/()8H1pi,n-C1,n)
where: r = a'kXvS + b'Xv = oxygen uptake rate
to ~ Vp /Q = detention time
Gas Phase
(8)
Kt8 = (M/vi RT)(Gn_, Pi.n-i-GnPi.rWHlPlin-C1in)
Presently, Equation 7 is used to analyze for KLa in closed
tank, high purity oxygen systems since the exit gas flow
from each stage is not measured for use in Equation 8.
Thus, the accuracy of the K|_a determination is a function
of the accuracy of the oxygen uptake measurement,
similar to the evaluation of open tank systems.
To use Equation 8 effectively, the gas flow measurement
must be made for each stage of a closed tank series
reactor system along with the oxygen partial pressure
and concentration. This is presently not done with staged
closed reactor systems. If this could be accomplished
190
-------
Figure 8. Variation of Total Horsepower with Mixer Horsepower to Achieve DO = 4 mg/l at Average Load
950
900
850
c
.3
§
6
1
a.
04
800
750
To
3
700
650
600
220
J.
230
100
90
80
70
60
w
.H
15
i
I
£
50
40
240 250
Aeration Mixer Horsepower
30
260
270
191
-------
Figure 9. Closed Tank Air Activated Sludge, Gas Flow-Oxygen Relationship
30
20
to
(D
+*
'x
LLJ
CM
o
10
\
\
\
I
5,000
10,000
i
15,000
I
30
20 S
c
X
O
•*—
10 £>
CL
20,000
0
25,000
O
Q
0
1
1
N0 = 3.5 Ib 02/hp-
Ib 02/hp-hr
•nr~7
5,000
10,000 15,000
G0 (scfm)
20,000
25,000
192
-------
Figure 10. Temperature Effects on Reaction Rate
1.0
0.8 -
0.6
0.4
e =1.05
V
<3
oc
c
o
V*
o
(0
0>
oc
0.2
Biological Removal
0.1
KT = K32(1.05)<32-T'
A —Pilot Study
• —Full-Scale System
• —Laboratory Study
.08
.06
10
20
30
40
50
60
Basin Temperature (°C)
193
-------
practically, it would provide an accurate technique for
evaluating the transfer capabilities of aeration equipment
under actual operating conditions.
Conclusions
1. Analysis and design of aeration equipment require-
ments for closed tank aeration systems is more
difficult than for open tank systems due to the varia-
tion of the oxygen partial pressure in the gas phase.
2. A model interrelating gas and liquid phases provides
a tool for understanding and properly evaluating these
systems.
3. Aeration equipment capabilities, as measured by the
oxygen transfer coefficient, K|_a, are the same for air
and oxygen systems, whether open or closed, for a
given piece of equipment at the same power level and
system geometry.
4. Closed aeration systems provide an excellent oppor-
tunity to accurately evaluate "in situ" KLa values using
the gas phase mass balance equation. Interstage gas
flow measurements are required.
References
1. Mueller, J.A., T.J. Mulligan, and D.M. DiTro. "Gas
Transfer Kinetics of Pure Oxygen System". Journal
Environmental Engineering Division of the ASCE
99, pp. 269-282, June 1973.
2. Mueller, J.A., J. Famularo, and T.J. Mulligan. "Appli-
cation of Carbonate Equilibrium to High Purity Oxygen
and Anerobic Filter Systems". Proceedings, American
Chemical Society Symposium, New Orleans,
March 1977.
194
-------
On Site Evaluation:
Steady State Vs
Non-Steady State Testing
Ross E. McKinney
University of Kansas
Lawrence KA 66045
John R. Stukenberg
Black & Veatch
Kansas City MO 64114
The overall objective of any aeration system is the
transfer of oxygen from the gaseous phase to the micro-
organisms in the activated sludge. The gaseous phase
can be either air or pure oxygen. The oxygen transfer
reaction occurs in a series of steps. The first step is
transfer from the gaseous phase to the liquid phase. The
second step is transfer from the liquid outside the
microbial cells to the liquid inside the microbial cells. The
non-steady state oxygen transfer test measures only the
oxygen transfer from the gaseous phase to the liquid
phase, while the steady state oxygen transfer test
measures both steps of oxygen transfer. Theoretically,
the steady state oxygen transfer test should be preferred
to the non-steady state test since it is the only measure
of oxygen transfer under actual operating conditions.
Unfortunately, the steady state oxygen transfer test is
subject to a number of variables which can make evalua-
tion subject to considerable error. For this reason, the
steady state test is seldom used anymore for oxygen
transfer tests, particularly in the case of manufacturer's
acceptance tests.
In response to the efforts of the ASCE to develop an
oxygen transfer standard, a study was undertaken to
determine whether the steady state test should be
included in the standard. Data from a prior test of
mechanical surface aerators, where the aerators were
tested both by the steady state method in the field and
the non-steady state method in the shop, were reviewed
and reanalyzed. Steady state tests were then conducted
on a full-scale aeration system fitted with mechanical
surface aerators, and the results were compared to
manufacturer's acceptance tests conducted previously
using the non-steady state method. The results of this
study and the recommendations to the ASCE committee
are presented herein.
Theoretical Relationships
The oxygen transfer relationships in the steady state test
differ from those of the non-steady state test in two
ways. First oxygen is not being transferred to clean
water; but rather, is being transferred to water containing
varying concentrations of contaminants, both organic and
inorganic. The contaminants in the test water can affect
both the rate of oxygen transfer and the saturation
concentration. Surface active organic compounds tend to
21
have the greatest affect on the oxygen transfer relation-
ships. Secondly, in the steady state test the activated
sludge creates a continuous oxygen uptake from the
water phase. The oxygen transfer equation for the steady
state test is shown in Equation 1.
(D
dC/dt =
Under steady state conditions, dC/dt = 0. The important
variables are alpha, beta, dissolved oxygen and microbial
oxygen uptake. Alpha and beta are used to correct the
oxygen transfer and the oxygen saturation values,
respectively, from the actual test water conditions to
clean water conditions. One of the most critical measure-
ments is the dissolved oxygen concentration. The problem
is not in the dissolved oxygen measurement, but rather
in which value to use in the evaluation equation. The
oxygen uptake rate by the activated sludge is the most
difficult parameter to measure and has been the source
of problems in a number of field tests.
Alpha and Beta Determinations
Alpha and beta determinations are made by comparing
oxygen transfer rates and oxygen saturation values in
clean water against the water in the mixed liquor. A test
apparatus such as shown in Figure 1 can be used to
collect the necessary data. This unit employs a variable
speed mixer together with a controlled flow of air or
oxygen. Since the aeration equipment is evaluated at
20°C, it is possible to adjust the test fluid to 20°C with a
minimum of effort so that temperature can be eliminated
as a variable. If the temperatures of the test fluids cannot
be easily adjusted to 20°C, they should be allowed to
adjust to the same temperature to minimize temperature
effects.
The important variables in the alpha determinations are
mixing, air flow rate and temperature. Eckenfelder and
Ford (1) indicated that alpha values varied with mixing,
making it difficult to determine alpha accurately. The
mixing effect can be minimized by operating the test unit
at essentially the same KLa that the aeration system was
designed to achieve. It will be necessary to run a series
of tests on clean water to determine the desired mixing
speed and air flow rate. Once the mixing and air flow
195
-------
Figure 1. Bench-Scale Aeration Tank
9-3/4'
cc
O)
Plan
Legend
8" P.G. Pipe
with 1/4" Wall
1 /4" P.G. Base l£
J-1/4" Riser to
Purgemeter
M/4"Tee
--1/8" P.G. Baffle-
4 at 90°
P.G. — Plexiglass
Cu — Copper
S.S. — Stainless Steel
Purgemeter
10 liter Liquid
Capacity
3/8" S.S. Shaft
Baffles, Provide
(2) 1/8" x 4-1/2'
Slots at Wall
(4) 1 /8" S.S. Blades
Welded to Collar
(Rotor Speed
300 rpm)
3/8" S.S. Collar
with Set Screw
1 /8" Air Opening
(Air Flow Rate
0.4 l/min)
Section
-------
rate have been selected, they should remain constant
throughout the rest of the study.
Deoxygenation of the test water can be accomplished by
using nitrogen gas to strip the oxygen, or by using
sodium sulfite and cobalt chloride to chemically remove
the oxygen. Dissolved oxygen measurements can easily
be made with a direct reading dissolved oxygen probe.
The reaeration data are normally collected from zero
oxygen concentration to near saturation at constant time
increments selected to give a reasonable change in
oxygen between readings. As the dissolved oxygen con-
centration approaches saturation, the rate of change in
concentration is so low that the error in reading the
meter will exceed the actual change in concentration
making additional data collection meaningless.
Oxygen saturation can be determined directly or
measured from the data itself. While long term aeration
can produce saturation, care must be taken to insure that
the temperature does not increase. Incorrect data can be
obtained if test conditions are not held constant. A plot of
dC/dt on the ordinate axis vs C on the abscissa (2)
should yield a straight line which passes through C* at
dC/dt = 0. Further, the slope of the line is KLa, making
it unnecessary to develop the semilog plot of (C*-C)
vs t. In addition, the dC/dt vs C plot clearly shows the
degree of mixing in the test vessel.
The test water from the aeration tank should be taken at
the same time that oxygen uptake measurements are
made so that the two measurements can be correlated. A
mixed liquor sample should be taken and allowed to
settle and the supernatant removed for testing. In many
systems it is possible to use the final effluent for the test
water if tests are made when the activated sludge system
is working well. There is little difference in the chemical
characteristics of the liquid in the aeration tank and in
the final sedimentation tank if complete mixing occurs in
the aeration system. In large aeration tanks, the rate of
mixing is such that a gradient can exist from the inlet to
the outlet. For this reason it may be necessary to deter-
mine alpha and beta values at several points in the
aeration system.
The activated sludge-free supernatant is deoxygenated by
the same procedure used for deoxygenating the tap
water. The oxygen transfer characteristics are determined
under the same conditions as for the tap water. Alpha is
calculated from the ratio of KLa of supernatant/KLa of tap
water. Beta is calculated from the ratio of C* of super-
natant/C* of tap water. If data are collected at different
temperatures, it will be necessary to correct the results
to a common temperature base.
Oxygen Uptake Measurement
Oxygen uptake measurements must be made with care if
meaningful data are to be obtained. It should be recog-
nized that the sample of activated sludge being analyzed
is continuously changing, making it necessary to obtain
measurements as quickly as possible at the aeration
basin. Since fresh wastes cannot enter the sample
continuously, the rate of oxygen uptake decreases and
the organics are metabolized to the point that endo-
genous respiration is the sole cause of the oxygen
uptake. Since endogenous respiration is relatively
constant, it gives an apparent steady oxygen uptake rate
that is often mistaken for the overall oxygen uptake rate
in the aeration basin.
As a result of the problems in obtaining the real oxygen
uptake rate in the aeration basin, it has been suggested
that the aeration tests be run under endogenous respira-
tion conditions (2). By stopping the waste flow to the
aeration basin being tested and continuing the sludge
recirculation system, the activated sludge will quickly
reach endogenous metabolism and will have a uniform
oxygen demand rate over the aeration system. The rate
of change in endogenous respiration is slow enough that
measurements can be made with less chance for error.
Since the oxygen respiration rate will be relatively low,
the dissolved oxygen in the aeration basin will be high,
making it very easy to take a sample and determine the
oxygen uptake rate without any additional aeration. If the
initial dissolved oxygen level of the sample is low, it will
be necessary to add oxygen prior to determining the
oxygen uptake rate.
In the activated sludge systems that are lightly loaded,
the rate of endogenous respiration will be too low for
proper evaluation. Under these conditions, it will be
necessary to test the aeration system under normal
loading conditions. The rate of oxygen uptake will usually
still be low enough that it can be measured without too
much difficulty. Experience indicates that oxygen uptake
rates under 60 mg/l-hr can be measured with a
minimum of error.
The oxygen uptake data can easily be taken with a direct
reading dissolved oxygen probe. The sample of mixed
liquor must be continuously stirred to keep the solids in
suspension and dispersed so that oxygen is available to
all the microbes. Stirring can be accomplished with a
magnetic stirrer, a mechanical mixer or simply by mixing
with the DO probe itself. As the dissolved oxygen con-
centration decreases, a point will be reached where the
rate of oxygen uptake will slow down. This slowing in
oxygen uptake generally occurs when oxygen becomes
limiting in the test system and has no relationship to the
full-scale aeration tank. A plot of the oxygen uptake data
vs time will show when oxygen becomes limiting in
the test.
Dissolved Oxygen Measurements
Aeration test data are meaningful only if the entire
aeration tank has excess dissolved oxygen at all points.
Testing during endogenous respiration generally insures
adequate dissolved oxygen, while testing during feeding
197
-------
of wastes can result in some points within the aeration
tank not having adequate dissolved oxygen. The important
dissolved oxygen concentration in the aeration tank is the
DO in the liquid approaching the aerator since that value
determines the driving force across the aerator. Since
aerators transfer oxygen primarily at a point, the dis-
solved oxygen concentrations vary around the aeration
tank, depending upon the overall pattern of fluid flow in
the aeration tank. Determining the dissolved oxygen
profile around the aeration tank can be quite helpful in
understanding the fluid flow patterns in the aeration
tank. It can easily be determined if the fluid is short-
circuiting around the aeration tank or flowing as
expected.
KLa Calculations
Equation 1 can be used to calculate KLa at the test
temperature. It will be necessary to correct the KLa to
20°C as shown by Equation 2.
(2)
(KLa)20 = (KLa)T0
(20-T)
The value of 6 has largely been determined experi-
mentally, with the value of 1.024 generally being
accepted. However, examination of KLa indicates that the
oxygen transfer to the water and the mixing of the
oxygenated water internally is controlled by the dynamic
viscosity of the liquid. It would appear that the relative
ratio of dynamic viscosities of water at 20°C and the test
temperature could be a better correction factor than an
empirical factor determined from a limited set of data.
Results from Full-Scale Testing
Midland, Texas
The specified manufacturer's acceptance test for the
Midland, Texas, aeration equipment was the steady state
method. However, when differences arose between the
manufacturer and the engineer in evaluating the data,
non-steady state tests were conducted on a similar
aerator in a shop test tank.
The Midland aeration basins are 76 ft in diameter with a
16 ft side water depth. Each basin is equipped with four,
50 hp slow speed surface aerators mounted above draft
tubes. A total of 13 steady state tests were conducted
over a 4-day period in November 1974. Dissolved oxygen
measurements were made with a direct reading dissolved
oxygen probe. Each test consisted of three oxygen uptake
measurements, 13 DO measurements in the aeration
basin and alpha and beta measurements. The 13 DO
measurements in the aeration basin consisted of 3 sets
of 4 measurements at 1.5, 4, 8 and 14.5 ft below the
surface and one measurement in the middle of the draft
tube of one of the aerators. The test specifications
permitted averaging all of the DO measurements rather
than using just the DO value in the draft tube. This
procedure would have resulted in high values for KLa if
the oxygen uptake data had been properly taken. Since
some time elapsed between the time mixed liquor
samples were taken and uptake rates were measured,
the measured uptake rates were probably lower than the
actual values in the aeration basin. Overall, the two
errors resulted in a (KLa)20 value of 7.2 hr"1, considerably
less than the specified value of 13.0 hr1. In order to meet
the specification, it was suggested by the manufacturer
that KLa varied with oxygen uptake and that extrapolation
of the data would yield the desired value. Since KLa had
not been shown to vary with oxygen uptake rates, this
concept was rejected and the aerator was subjected to a
non-steady state test.
The non-steady state test was carried out in a square
tank with the aerator in the center. The total volume of
fluid was essentially the same as the volume influenced
by a single aerator in the aeration basin. Five points were
selected for dissolved oxygen measurements with two
points 3 ft below the surface, two points 2 ft off the
bottom and one point under the draft tube. The data
showed the affect of the tank configuration, giving a
surge pattern as shown in Figure 2. The surge which
was noted at an upper sample point was damped out at
the sample point beneath the draft tube, but the relative
variations were similar. A total of three replicate tests
were made, yielding a (KLa>2o of 15.6 hr"1 as the average
of 15 data points. The K|_a values were determined from
three-point running averages (denoted by solid triangles
on the figure) using the dC/dt vs C curves. The conven-
tional semilog plots yielded an average (KLa)2u of 13.4 hr'1.
The non-steady state test clearly indicated that the
aerator was capable of meeting the specified (KLa)2o
value.
The differences in the results prompted reexamination of
the steady state data. The data were examined on the
basis of dC/dt vs C plots rather than the simple
calculations of K^a. Theory indicated that a plot of
(dC/dt)/a vs C should yield a straight line passing
through C* at (dC/dt)/a = 0. Examination of three sets of
data at different oxygen uptake rates on three different
days indicated that these data points were on a straight-
line, but that line did not intercept the abscissa at the
measured values of C*. The data indicated that the
oxygen uptake rates were low. A plot of a line through
the data and the measured C* gave a theoretical line that
should have corresponded to the actual oxygen uptake
rates divided by alpha. The (KLa)20 from this line was
16.2 hr"1, very similar to non-steady state results. These
results clearly pointed out the problems in steady state
testing.
Lawrence, Kansas
The principal problem with the Midland, Texas, steady
state test results was the apparently low oxygen uptake
values. To determine if this error could be minimized,
steady state tests were performed on all the aeration
equipment in one aeration basin at the Lawrence, Kansas,
wastewater treatment plant. This same equipment had
198
-------
been tested earlier in the aeration basin by the non-
steady state test as a part of the manufacturer's accept-
ance tests. Thus, an excellent opportunity was provided
for comparison of steady state vs non-steady state test
procedures and results.
Non-Steady State Testing
The Lawrence aeration basins are 105 ft in diameter with
a 14 ft side water depth. Each aeration basin is equipped
with four 60 hp surface mechanical aerators mounted
over draft tubes. Non-steady state testing was conducted
in May 1976 with direct reading dissolved oxygen probes
located at five points in the basin. The signal from
each probe was recorded on a continuous graph so that
it was possible to examine the surges in mixing within
the aeration basin. These data were very carefully
reevaluated for this paper to provide a base for compari-
son with the results from steady state testing. The
analysis of the non-steady state data indicated that a
(K 3)20 °f 9-4 hr"1 was obtained at 51.6 motor shaft hp.
The dC/dt vs C plot for the DO probe located below
the draft tubes is shown in Figure 3. This point was
selected since the continuous graph showed definite
surging in the dissolved oxygen data. The shifts in the
dC/dt reflect the surges caused by the irregular shape of
the tank volume affecting this aerator. Figure 3 also
shows the conventional semilog plot of these same data.
The semilog plot definitely dampens out the variations in
the data and yields a reasonable set of data that can be
easily evaluated. The dC/dt vs C plot was evaluated
using a three-point running average and linear regression
analysis. The semilog plot was evaluated by the equip-
ment manufacturer using their computer program and
yielded a (KLa)2o of 9.3 hf1. Essentially, both methods
yielded the same results, but the dC/dt plot showed the
surging that occurred in the aeration tank.
Steady State Testing
Steady state testing was conducted at Lawrence on
February 3, 1978. The testing was performed in the
same basin used for non-steady state testing. Power
measurements were made on the same aerator moni-
tored during the non-steady state tests. Two steady state
tests were conducted, one in the morning and one in the
afternoon. At the conclusion of the morning tests, it was
discovered that the DO meter used for a, /3 and dO/dt
measurements were incorrectly calibrated. However, all
measurements were off by a constant factor. Therefore,
the data are valid and have been included in this paper.
The morning tests were conducted on mixed liquor which
had been allowed to enter into endogenous respiration.
Wastewater flow to the aeration basin was discontinued
2 hr prior to testing to insure oxidation of all the re-
sidual organic matter from the wastewater. After the
morning tests had been completed, wastewater flow to
the aeration basin was reestablished. The afternoon tests
were conducted on mixed liquor that had been actively
metabolizing wastewater for approximately 1-1/2 hr. In
both tests, the sludge recycle pumps were left in opera-
tion to return sludge from the secondary clarifier to the
aeration basin. Also, in both tests, the power consump-
tion of the aerators was adjusted to approximate the
power consumption measured during the non-steady
state tests.
In the morning tests, a and p measurements were made
on final effluent using the apparatus shown in Figure 1.
Conditions were set so that the (KLa)2o °ftne test
apparatus (14.0 hr"1) was similar to the (KLa>2o °ftne
aeration basin (9.4 hr"1). Figure 4 shows the uncorrected
data used for the a and j9 determinations, plotted both
by the dC/dt vs C and by the log deficit methods.
Although both methods yield the same results, a - 0.82
and /3 = 0.98, the variations in the data are more readily
apparent from the dC/dt vs C plot. The oxygen uptake
rate of the mixed liquor, corrected for the probe calibra-
tion error, was 13.6 mg/l-hr. The mixed liquor tempera-
ture was 9.6°C at the time of testing and the MLSS were
1810 mg/l. The corrected DO concentration at the inlet
to the draft tube was 10.1 mg/l. From these values, the
estimated (KLa)2o was 9.1 hr"1, comparable to the value
determined by non-steady state testing. The measured
aerator power output was 52.2 motor shaft hp.
Testing in the afternoon was conducted with wastewater
flow through the aeration basin to increase the oxygen
uptake rate above the relatively low endogenous respira-
tion rate. Tilting weirs were adjusted with the initiation of
flow through the basin so as to keep the power con-
sumption of the aerators the same as that used in the
morning tests. The problems with the DO probes were
corrected, and clarifier effluent was sampled for a and /3
determinations. The a and P data are plotted in Figure 5
using the dC/dt vs C plot.
The long aeration period without wastewater resulted in
a change in alpha to 0.97. The beta value remained
essentially the same at 0.97. The KLa values for the alpha
and beta determinations were again somewhat higher
than the full-scale unit, but it was not felt the difference
was significant. The DO profile in the aeration tank,
shown in Figure 6, gave a mixing pattern which was
typical for this type of unit. The fluid from the aerator
appeared to move in a thin layer across the surface to
the wall and down to the bottom with normal intermix in
the middle zone. The minimum DO was estimated to be
7.3 mg/l from the data at the bottom of the tank next to
the draft tube. The range of DO values in the aeration
tank definitely shows that the average DO is not related
to the oxygen transfer, except indirectly. The (KLa)2o was
calculated as 9.1 hr"1. The (^3)20 value from the after-
noon steady state test was essentially the same as the
results from non-steady state testing, within the normal
error range for such tests.
199
-------
Figure 2. Non-Steady State Test Data, Midland, Texas
.C
\
I
cc
120
110
100
90
80
70
60
50
40
30
20
10
a. Point 3' Below Surface
3 4
j_
j
6 7 8 9 10
\
o>
E
120
110
100
90
80
70
60
50
40
30
20
10
b. Point Beneath Draft Tube
C (mg/l)
3456
C (mg/l)
8 9 10
Figure 3. Non-Steady State Test Data, Lawrence, Kansas Sample Point Under Draft Tube
a. Direct Analysis
o>
10r
b. Conventional Analysis
u
67 89 10 0123456789 10 11 12 13 14
C (mg/l)
Time (min)
200
-------
Figure 4. Morning Alpha and Beta Test Data Steady State Test at Lawrence, Kansas
tr
a. Final Effluent —
Direct Analysis
09
0
b. Tap Water —
Direct Analysis
o>
E
o
u
c. Final Effluent —
Conventional Analysis
468
Time (min)
10
o>
u
*
2 -
0
d. Tap Water —
Conventional Analysis
468
Time (min)
10 12
201
-------
Figure 5. Afternoon Alpha and Beta Test Data Steady State Test at Lawrence, Kansas
a. Final Effluent — Direct Analysis
b. Tap Water — Direct Analysis
o
fO
\
f
cc
cc
10 11 12
10 11 12
C (mg/l)
C (mg/l)
-------
Figure 6. Dissolved Oxygen Profile in Aeration Tank in Afternoon Tests at Lawrence, Kansas
to
o>
O
(A
CO
CQ
Aerator
Impeller
Aerator &
Draft Tube
94" Diameter
Draft Tube
14
-------
Conclusions
The testing at Lawrence, Kansas, and to a limited extent,
the testing at Midland, Texas, demonstrated that it is
possible to obtain comparable results by the steady state
and the non-steady state methods. The use of endo-
genous respiration rather than the fed system offers an
easier technique for evaluating aeration equipment by
the steady state method. However, at oxygen uptake
rates below 60 mg/l-hr, good results can be obtained
with either method.
The steady state test has greater potential for error than
the non-steady state test since more analyses must be
made and correlated together. Testing at the Kansas and
Texas sites demonstrated the ease in which errors in
measurement can be made, whether due to sampling
techniques or instrument problems. From a practical
point of view, the non-steady state test appears to be the
best method for evaluating aeration equipment perform-
ance, especially when tested in the actual aeration tank
with continuous data recording to permit evaluation of
surge flows as well as oxygen transfer.
Acknowledgements
The authors gratefully acknowledge the generous assist-
ance of James A. Bell, Ecodyne Corp., Smith & Loveless
Div., during the steady state testing at Lawrence, Kansas.
References
1. Eckenfelder, Jr., W.W. and D.L. Ford. "New Concepts in
Oxygen Transfer and Aeration". Advances in Water
Quality Improvement, University of Texas, pp. 215-236
1968.
2. Stukenberg, J.R., V.N., Wahbeh, and R.E., McKinney,
"Experiences in Evaluating and Specifying Aeration
Equipment", Journal Water Pollution Control
Federation, 49, p. 66, January 1977.
204
-------
Problems Encountered in
Steady State Field Testing of
Aerators and Aeration Systems
A. A. Kalinske
Camp Dresser and McKee, Inc.
Walnut Creek CA 94596
The two general methods currently in use for determining
the oxygen transfer capacity of any aerator or aeration
system are the non-steady state, clean water technique
and the steady state test method conducted under
process conditions, which, for an activated sludge aera-
tion basin, means testing after a wastewater treatment
plant is placed in operation and a suspension of microbial
solids is developed. In the latter case, the liquid in which
the aerator is tested has an oxygen uptake which must
be considered in the test measurements and the oxygen
transfer rate of the aeration system.
The general mathematical models used for both test
methods are similar and are based on the same basic
assumptions; the most important, and incidently the most
frequently ignored, is that of a completely mixed liquid
mass in which the oxygen uptake and transfer are
occurring. Therefore, theoretically, both test techniques
should give the same end results. However, from a
practical standpoint this is not readily achieved and the
basic reasons and problems will be discussed in this
paper.
The fundamental ^mathematical model that has been
adopted to describe the oxygen transfer phenomenon
when oxygen is supplied to a liquid masVby^ome type of
aeration device is: /
22
-------
common method for obtaining the value of KLa is to
determine R by taking a sample of the mixed liquor,
raising its DO level by rapid aeration, and then measur-
ing the drop in DO with respect to time by use of a fast-
response DO meter. The value of KLa is then calculated
from Equation 1, since for steady state conditions,
dC/dt = 0. The relation is as follows:
(2)
where:
KLa = R/fC'-C1)
C1 = DO concentration at point where sample is
taken for determining R
A problem regarding use of this technique is that for
localized types of aerators (surface entrainment or sub-
merged turbine type), the value of C1 will vary from point
to point spatially, thus necessitating use of an average
value or several determinations of R from different
sample points. Moreover, the externally determined value
of R is not necessarily the value in the basin at the
sample point, since it has been shown that R is depend-
ent on the turbulence level. Also, the biological organisms
in the basin are receiving food at some steady rate and
this is not true of the organisms in the test jar. In fact,
the value of R obtained in the external test jar can be
increased by supplying some food during the oxygen
uptake test.
Stukenberg et at. (10) discuss the matter of R varying
throughout a basin. Of course, the value of KLa can be
determined for several BOD loads, or different average
values of R, and such values if plotted against the related
value of C1 can be used to calculate KLa from the slope of
the resulting straight line, per Equation 2. This will
provide a check on the proper value of C* against the
calculated or assumed value since it is the value of C1
when R = 0, or the intercept on the horizontal axis of the
straight line connecting different values of R and C1.
For diffused air systems, a better and more direct method
of determining the value of oxygen transfer rate, or KLa,
can be used. This is based on making an oxygen balance
around the aeration basin by measuring the oxygen
concentration (or C02> in the off-gas. This method was
used by Leary et a/.(5) for determining air diffuser
efficiency at the Milwaukee plant and also by Lister and
Boon (6) in England. The latter compared the off-gas
measurement method for determining KLa to measuring
the uptake rate, R, externally. The latter method gave
Kua values up to 30% higher, which they attributed to
the much higher turbulence level in their respirometer
than existed in the aeration basin.
Another method that has been used successfully to
measure the oxygen transfer rate under process condi-
tions involves establishing a steady and stable condition
in the aeration basin, then shutting off the aeration and
raw sewage and recycle flows, allowing the DO to drop
to zero, and then starting up the aeration and flows and
measuring the values of C and t at several points in the
aeration basin until a stable value of C is obtained. This
value should be at least equal to 3 mg/l but below
5 mg/l for air systems. The value of KLa is determined
by plotting values of C vs t, drawing a smooth curve
through the points, then graphically (or arithmetically)
calculating dC/dt for different values of t. The latter
should give a straight line, if the system is stable, whose
slope is KLa. This comes about by transforming Equation
1 to the following form:
(3)
dC/dt = (KLaC*-R)-KLaC
In this procedure, KLa is determined without the need to
measure R externally, or even to estimate the value of C*
under the process conditions. Some data using this
procedure were recently obtained for an aerator installa-
tion where the owner required definite proof regarding
the amount of oxygen that was being supplied by the
aerators under existing process conditions. The data are
shown in Figure 1 and 2, together with the calculated
values of KLa and R. The latter can be calculated from
the intercept of the straight line at C = 0, as is apparent
from Equation 3, after determining the proper value for
C* under existing process conditions.
In all the above described test methods, in order to
determine the true oxygen transfer rate of the aeration
system, the amount of oxygen that is carried out by the
flow of wastewater and recycle should be added to that
calculated due to the uptake rate, R. Usually, this will be
under 5% of the uptake rate for normally loaded and
respiring activated sludge.
Conversion of Oxygen Transfer Rate Data to
Standard Conditions
Though steady state testing under process conditions
provides valuable information relating to the oxygen
transfer characteristics and efficiency of aerators under
actual specific operating conditions, such data are difficult
to convert to standard conditions, which are defined as
transfer into clean water at 20°C, zero DO, and sea
level pressure. In order to compare and evaluate various
designs of aerators and aeration systems, it is necessary
that their performance be determined for standard
conditions. It is then the responsibility of the design
engineer to convert such standard conditions data to the
performance that can be expected under actual specific
operating conditions for the treatment of a particular
wastewater over some temperature range.
To convert oxygen transfer data obtained under process
conditions to clean water performance, the major in-
adequately known factor is the so called a factor, which
is influenced primarily by surface-active agents present
in wastewaters and produced as various soluble organics
during biodegradation. Their physical effect on oxygen
transfer from a gas through the liquid film has been
described in several papers on aeration of wastewaters.
Two papers by Mancy et at. (7) (8) give excellent basic
physical explanations. A very significant item relating to
206
-------
Figure 1. Reaeration of Activated Sludge Basin Under Process Conditions
dC/dt = KLa(C'-C)-R
Steady State C = C = 3.6 mg/l
C* = 11.1 mg/l
0 10
40 50 60
Time (min)
80 90
Figure 2. dC/dt Values from Figure 1 Curve
0.08
•£• 0.06
0.04 -
0.02 -
KLa =A(dC/dt)/AC (slope of line) = 0.0259 min"1
= 1.554 hr1
R = KLa C* -0.091 (Y axis intercept) = 0.19 mg/l-min
= 11.40 mg/l-hr
Oxygen Transfer Rate
W = R = 11.40 mg/l-hr
also, W = KLa(C*-C1) = 1.554(11.1-3.6) = 11.65 mg/l-hr
207
-------
the factor a is that for a given wastewater or concentra-
tion of surface-active agents, a can be vastly different for
different types of aeration systems. Systems producing
high levels of turbulence and interface renewal rates
produce much higher values for a, everything else being
similar, than do low-turbulence systems. For example.
Downing and Boon (1) report that water with about
5 mg/l of detergents or similar surface-active agents can
give an a value of 0.4 for fine bubble diffusers and a
value of a up to 1.5 for surface type mechanical aerators.
Obviously, an engineer must be able to estimate the
value of a to use for any particular type of aeration
system in order that he be able to convert clean water
performance data to that to be expected under particular
process conditions. Data obtained for various aeration
systems using previously described steady state tech-
niques should be compared with clean water test data in
order to have some basis for estimating values of a to
use in sizing aerators. Sometimes it will be necessary to
make non-steady state tests in clean water to which
some detergent has been added to obtain an estimate of
how surface active agents influence the performance of
any particular aeration system. The English literature on
aerator testing indicates that it is common practice to
make the non-steady state tests in clean water to which
5 mg/l of anionic detergent have been added, since their
tests under process conditions indicate that this roughly
simulates the effect obtained in activated sludge aeration
basins treating domestic wastewater.
The other important parameter that must be considered
in converting test data for oxyen transfer rates obtained
under process conditions, and also from clean water
tests, is that of liquid temperature. In addition, if the
water temperature is considerably different from the air
temperature from which the oxygen is being obtained,
the value of C* to use should correspond to a tempera-
ture that is between that of the water and of the air.
There is inadequate information in regard to this to
permit any accurate calculation of the true temperature
of the liquid film when the air temperature is consider-
ably different from that of the water.
The influence of temperature on KLa resultes from the
effect of temperature on molecular diffusivity of oxygen
in a thin liquid film of a certain thickness. However, the
mathematical model on which Equation 1 is based does
not incorporate any parameter that would quantify this
film thickness. The model is based on the assumption
that diffusion of the oxygen outside of the interfacial film
is by turbulent diffusion which is several orders of
magnitude greater than molecular diffusion and is not
temperature dependent. Temperature affects the liquid
viscosity and it has been shown that the concentration of
suspended solids in activated sludge increases the
effective viscosity. This has been demonstrated by the
great differences in liquid pumping action and flow
pattern which some types of localized aerators exhibit in
clean water and in mixed liquor having several thousand
mg/l of suspended biological solids.
There is good evidence that the effect on KLa of liquid
temperature and of the concentration of suspended solids
will depend greatly on the intensity and type of tur-
bulence generated by the aeration system. Fine-grained
microturbulence, such as produced by the high shearing
action of mechanical aerators, will greatly reduce and
minimize the effect of-changes in liquid viscosity due to
temperature and suspended solids and, thus, the effect
on molecular diffusion. In other words, high energy
systems will not exhibit large effects of temperature on
KLa. Imhoff and Albrecht (3) very clearly show how liquid
temperature affects the relative oxygen absorption (for
comparable oxygen deficit values) for different aeration
systems. They indicate a significant correction for
temperature for low-turbulence systems, such as bubble
diffusers and certain types of mechanical aerators, while
the correction is practically non-existent for high-turbul-
ence systems. They classify systems having energy inputs
in excess of 0.75 hp/1,0003 as being high-turbulence
systems. This is very likely an over-simplified designation.
In any case, it is fairly obvious that a constant tempera-
ture correction factor for KLa for all types of aerators is
not justified. Moreover, low-turbulence systems will
exhibit greater effects on KLa of high concentrations of
mixed liquor suspended solids than will high-energy
systems, thus further complicating conversion of test
data obtained under process conditions to standard
conditions.
Summary
To summarize, testing aerators under process conditions
should be done whenever possible in order to obtain
information that will provide a broader basis for convert-
ing such test data to performance under standard, clean
water conditions. This then permits the design engineer
to have more reliable information for converting the
performance information given for aerators under
standard conditions to the wide variety of practical
operating conditions in wastewater activated sludge
aeration basins.
The problems encountered in testing aerators under
process conditions are as follows:
1. Difficulties in establishing steady state, stable condi-
tions as far as BOD loading and mixed liquor solids
are concerned in treatment plants.
2. Sampling problems when using the off-gas analysis
method for evaluating performance of diffused air
systems. Replicate testing reports predict a 95%
confidence level for a mean oxygen transfer efficiency
of about 20%.
3. Data for externally measured oxygen uptake rates for
samples from aeration basin mixed liquor are not
208
-------
necessarily representative of the actual average
uptake rate in the basin. A more preferable method is
to use a reaeration technique which permits determin-
ing KLa under process conditions without separate
measurement of R.
4. Conversion of steady state process test data to oxygen
transfer rates under standard conditions requires
knowledge of the in situ value of a, which is depend-
ent on the chemical composition of the liquid phase,
turbulence-level, and concentration of suspended
solids in the aeration basin.
References
1. Downing, A.L, and A.G. Boon. "Oxygen Transfer in
the Activated Sludge Process". Proceedings of 3rd
Manhattan College Conference on Biological Waste
Treatment, Pergamon Press, p. 123, 1963.
2. Eckenfelder, Jr., W.W. "Factors Affecting the Aera-
tion Efficiency of Sewage and Industrial Wastes".
Sewage and Industrial Wastes, Vol. 31, p. 60, 1959.
3. Imhoff, K.R.,and D. Albrecht. "Influence of Tempera-
ture and Turbulence on Oxygen Transfer in Water".
Proceedings of 6th Conference of Water Pollution
Research Association, Jerusalem, p. 23, 1972.
4. Kayser, R. "Comparison of Aeration Efficiency Under
Process Conditions". Proceedings of 4th International
Conference on Water Pollution Research,. Prauge,
p. 477, 1969.
5. Leary, R.D., LA. Ernest, and WJ. Katz. "Full-Scale
Oxygen Transfer Studies of Seven D iff user Systems".
Journal of Water Pollution Control Fedration, 41,
p. 459, 1969.
6. Lister, A.R.,and A.G. Boon. "Aeration in Deep Tanks:
An Evaluation of a Fine-Bubble Diffused-Air System".
Journal of the Institute of Water Pollution Control,
Great Britain, 72, No. 5, p. 590, 1973.
7. Mancy, K.H. et al. "Effects of Surf ace-Active Agents
on Aeration". Journal of Water Pollution Control
Federation, 37, p. 212, 1965.
8. Mancy, K.H. et al. "Mechanism of Interference of
Surface-Active Agents with Gas Transfer in Aeration
Systems". Advances in Water Quality Improvement,
University of Texas, Austin, p. 262, 1968.
9. Nogaj, R.J., and E. Hurwitz. "Determination of
Aeration Efficiency Under Process Conditions,"
Proceedings of 18th Industrial Waste Conference,
Purdue University, p. 674, 1963.
10. Stukenberg, J.R., V.N. Wahbeh, and R.E. McKinney.
"Experiences in Evaluating and Specifying Aeration
Equipment". Journal of Water Pollution Control
Federation, 49, p. 66, 1977.
209
-------
Section V. Tracer Methods
Use of Tracers for
Evaluation of Oxygen Transfer
Larry A. Neal
Law Engineering Testing Company
Marietta GA 30067
As energy costs continue to escalate, we are faced with
two difficult challenges related to the design and opera-
tion of oxygen transfer equipment:
1. Oxygen transfer equipment manufacturers must
continue to develop more energy-efficient designs.
2. Engineers designing wastewater treatment facilities
must specify oxygen transfer requirements that are
adequate to get the job done with a minimum of
wasteful overdesign.
The economic aspect of these challenges is great. Ten
years ago, a 50-hp mechanical aerator sold for about
$6,500 and required some $6,200/yr for electrical
energy. Today, the same basic aerator sells for about
$16,000 and associated electrical energy costs may
exceed $12,000/yr.
Over the design life of an oxygen transfer system, even
modest equipment efficiency improvements can represent
considerable potential cost savings. However, overly
conservative facility designs that result in too much
installed oxygen transfer capacity can burden owners
with unnecessary energy costs that might offset savings
provided by improved equipment efficiency.
If these two challenges are to be met, we must have
accurate and reliable testing methods available for
evaluating oxygen transfer. We need testing methods
that can be reliably used to evaluate individual oxygen
transfer devices under carefully controlled test
conditions, but more importantly, we need testing
methods that can be used to accurately evaluate
oxygen transfer in total systems operating in the field
under real process conditions.
Most equipment testing done today depends on either a
direct or indirect DO balance that is difficult to conduct
because oxygen is a common, sparingly-soluble gas that
is both chemically and biologically reactive over a wide
range of environmental conditions. Oxygen is susceptible
to so many influencing factors that accurate measure-
ment of simple physical transfer in an operating
treatment plant is difficult if not impossible. Conventional
clean water tests are limited to complete mix conditions,
and yet complete mix is not achieved or required in
many (most) field installations.
23
To overcome the problems associated with present con-
ventional testing methods, the gaseous tracer method is
proposed. This paper describes the tracer method for
making direct measurement of gas transfer performance
in test tanks or in operating field installations under real
process conditions. Previously reported applications of
the tracer method are discussed and potential applica-
tions are also identified. Much of the material in this
paper is taken from an earlier work by Neal and
Tsivoglou (6).
Tracer Method History
The tracer method involves the simultaneous use of an
inert radioactive gaseous tracer for oxygen (or any other
gas of interest) and a radioactive dispersion/dilution
tracer.
The tracer method for measurement of gas transfer
capacity is not new in that it was originally developed
about 12 yr ago for stream reaeration measurement by
Tsivoglou and others (7X8X9) at the U.S. Public Health
Service (USPHS) in Cincinnati OH.
Since the first field application in 1966, the tracer
method has been used to provide hundreds of accurate
stream reaeration measurements in several states as
summarized by Tsivoglou and Neal (8). The tracer method
is now widely considered to be the only reliable method
for accurate measurement of stream reaeration capacity.
From 1968 to 1970, Gordon and Etzel (2) conducted
mechanical aerator performance tests in a pilot-scale
test tank using the tracer method (gas tracer only).
Although the complete studies have not been published,
Gordon (3) did provide most of the test results in a 1976
"Communication" to the Water Pollution Control Federa-
tion Journal. (The Gordon and Etzel studies will be
discussed further in a subsequent section of this paper.)
In 1971, Neal (5) applied the tracer method to four
different aeration systems at operating wastewater treat-
ment facilities as part of a graduate study program at
Georgia Tech. In 1972, Neal and Tsivoglou (6) presented
the studies at the 45th Water Pollution Control Federation
Conference and the paper was subsequently published
in 1974.
210
-------
The Army Corps of Engineers is now sponsoring tracer
measurements of gas transfer capacity in hydraulic
models at the Waterways Experiment Station and at
Georgia Tech. The purpose of these studies is to improve
techniques for predicting gas transfer characteristics at
hydraulic structures, which can thereby lead to improved
designs.
Oxygen Transfer Considerations
In order to provide a common basis for describing the
tracer method, the following basic oxygen transfer con-
siderations are presented.
A commonly used general mathematical expression
describing aeration of wastewater is:
for oxygen at temperature T, and 6 is the temperature
correction coefficient. Tsivoglou (4) has demonstrated the
mean value of 6 , from theoretical considerations, to be:
(D
where:
cK/3C*-C)/dt =-aKLa(0C*-C)
C* = DO saturation concentration for clean
water, mg/l
C = DO concentration in water at time t, mg/l
K(_a = gas transfer rate coefficient for oxygen in
clean water, hr"1
/3 = ratio of the DO saturation concentration for
wastewater to that for clean water
a = ratio of the oxygen transfer rate coefficent
for wastewater to that for clean water
R = the rate of DO utilization by the system,
mg/l-hr
The product (<*KLa) represents the actual oxygen transfer
rate coefficient for the wastewater.
DO Saturation. The DO saturation limit in clean water,
C*, may be obtained from "Standard Methods" knowing
the water temperature, the prevailing atmospheric
pressure, and the pressure of saturated water vapor at
the specified water temperature. In diffused aeration
systems, where air is released at some depth below the
free water surface, the solubility of oxygen is influenced
by the increased partial pressure of oxygen in the
bubbles that have been compressed by the weight of the
water above and by the decreasing partial pressure as
oxygen is absorbed from the bubbles.
Gas Transfer Rate. For any specific aeration system, the
oxygen transfer rate coefficent for clean water, KLa, is
dependent on turbulent mixing, aeration equipment
characteristics, geometric configuration of the system,
and operating conditions within the system.
The effect of temperature on KLa may be described by:
(KLa)20 = (KLa)T/0(T-20)
in which (K|.a)20 is the gas transfer rate coefficient for
oxygen at 20°C, (KLa)r is the gas transfer rate coefficient
e
mean
= 1.022 ±0.004
in the temperature range 0 to 30°C. This predicted
value agrees well with the detailed observations of
Churchill et al. (1), which provided a value of 1.024. It
should be noted that these values of 0 were determined
by considering clean water conditions and should be
applied to wastewater conditions with caution.
DO Saturation in Wastewater. Pollutants can alter the
DO saturation limit. "Standard Methods" shows the
depression of oxygen solubility caused by chloride ion
concentration. Generally, however, the ratio of C* for
wastewater to C* for clean water, /3, is not readily
predicted and must be determined experimentally.
Gas Transfer in Wastewater. Various pollutants alter the
ability of water to transfer gas molecules. This alteration
causes the value of KLa to vary under identical mixing
conditions, depending on whether clean water or waste-
water is being aerated. The ratio of KLa for wastewater to
KLa for clean water, a, is related not only to the waste
constituents and concentrations, but also to the turbulent
mixing conditions within the fluid. This means that
determination of a must be conducted either in the full-
scale system or in a system that duplicates the turbulent
mixing within the real system.
DO Utilization. Oxygen utilization is the result of a
biochemical process and as such is independent of the
physical gas transfer process which is the subject of this
paper. The fact that both of these independent processes
influence the same parameter, DO concentration, is a
point of confusion for some investigators. It is empha-
sized that the actual gas transfer coefficient for oxygen
in a real wastewater system is equal to the product of a
and KLa. Similarly, the actual DO saturation concentra-
tion is equal to the product of /3 and C*. In order to
specify actual field aeration performance, the actual
oxygen transfer rate coefficient, aKLa, and DO saturation
concentration, 8C*. must be known.
Tracer Measurement of Oxygen Transfer
The tracer method depends on the simultaneous use of a
gaseous tracer for oxygen and a dispersion/dilution
tracer. In selecting a tracer gas, there are several criteria
to be considered:
1. It must be possible to relate the physical transfer
properties of the tracer gas to those of oxygen over
the expected range of environmental conditions.
2. The tracer gas must not be susceptable to biological or
chemical reactions that might be encountered in an
operating aeration basin.
211
-------
3. The tracer gas must be reasonably easy to handle and
measure.
4. The tracer gas must not be too expensive.
5. The tracer gas must be safe to use.
After much careful consideration and numerous labora-
tory studies, radioactive krypton-85 was selected by
Tsivoglou et al. (7) as the overall best gaseous tracer for
oxygen. Krypton is a monatomic "noble" gas that is
neither chemically or biologically reactive. As a radio-
active isotope, krypton-85 can be accurately measured
at very low concentrations by liquid scintillation methods.
Krypton-85 is readily available and inexpensive. Because
krypton-85 is a radioactive material, a state or federal
license is required to use it, but with proper handling,
krypton-85 can be safely used as a gas tracer.
In order to use krypton-85 gas as a tracer for the
evaluation of oxygen transfer, it is necessary to know the
relative transfer capabilities of the two gases. To experi-
mentally evaluate relative gas exchange rates, Tsivoglou
et al. (7) (8) conducted a series of laboratory tests in
which simultaneous transfer of gas pairs (e.g., oxygen
and krypton) could be observed in the same turbulent
water reactor. These experiments were conducted to
determine the relative gas exchange capabilities of
krypton-85, oxygen, radon-222, hydrogen, helium,
nitrogen, and carbon dioxide.
Krypton-85 transfer in water can be described by a
modification of the basic oxygen transfer expression,
Equation 1. The concentration of krypton-85 present in
the earth's atmosphere can be considered zero for
practical purposes. Therefore, any dissolved krypton-85 in
a turbulent water system will be steadily lost from the
water to the atmosphere according to the expression:
(3)
dCkr ,/dt - (Kkr)obs Ckr t
in which Ckr, is the concentration of dissolved krypton-85
in the water at time t and (Kkr)obs is the observed gas
transfer rate coefficient for krypton-85.
In the USPHS laboratory experiments, it was found that
the relative transfer capabilities of gas pairs, measured
as the ratio of gas transfer rate coefficients, is constant.
From 36 simultaneous krypton-85 and oxygen reactor
experiments, it was found that the ratio of gas transfer
rate coefficients, Kkr/KLa, is 0.83 with a standard
deviation of 0.04. This relationship is illustrated in
Figure 1.
A logical explanation of the constant ratio of gas transfer
rate coefficients from gas pairs was derived by Tsivoglou
(9) from consideration of Einstein's basic law of diffusion.
The derived expression relates gas exchange ratios to gas
molecule diameters as:
(4)
= Dm1/Dm2 = d2/d1 = constant
where the subscripts 1 and 2 identify the two gases
considered, K is the gas transfer rate coefficient, Dm is
the molecular diffusion coefficient, and d is the effective
diameter of the gas molecule. Unfortunately, there are
very little reliable diffusion and molecular diameter data
suitable for testing of Equation 4. However, limited
experimental verification of the relationship has been
documented for oxygen, nitrogen, and carbon dioxide.
From the 36 laboratory experiments, it was found that
the Kkr/KLa value of 0.83 is not significantly affected or
changed by temperature in the range of 10 to 30°C, by
the degree of turbulent mixing, by the direction in which
the gases are transferring, or by the presence or absence
of a broken water surface. The presence of the dissolved
solids in a natural ground water and the presence of
alkylbenzene sulfonate (ABS) in two experiments had no
significant effect on the ratio.
Effect of Pollutants. Because the ratio of K values for a
pair of gases is considered to be a fundamental molecular
property of the two gases, there is no reason to expect
that pollutants in the water would significantly affect the
ratio of K values. Various pollutants alter the ability of
gas molecules (oxygen, krypton, or others) to enter and
escape from water. This alteration causes the value of
the individual K values to vary under apparently identical
hydraulic and environmental conditions, depending on
whether clean water or polluted water is being con-
sidered. The pollutant effect on gas transfer is not only
related to the pollutant constituents and concentrations,
but also to the turbulent mixing regime within the fluid.
The tracer method has been used to measure pollutant
effects on gas transfer in a series of laboratory investiga-
tions, as described by Neal (4) and Tsivoglou and Wallace
(10). In these studies, tests were conducted to establish
"alpha" factors for LAS, NTA, and mineral oil in distilled
water and for grossly polluted stream waters. The funda-
mental reasons for pollutants influencing gas transfer
have not been well established, but it is speculated that
the effect of pollutants on physical gas transfer is
principally hydrodynamic in nature and molecular scale
hydrodynamics are probably involved.
Other Experiments. The previously mentioned work of
Gordon (3) and Etzel (2) is considered here because it
provides an independent set of pilot-scale data on
simultaneous oxygen and krypton-85 gas transfer. For
the first series of pilot-scale studies, a 15-ft diameter
pool with a plastic liner was filled with 2000 gal of well
water and a 2-hp floating mechanical aerator was
centered in the pool. Five tests were conducted in the
pool with simultaneous oxygen transfer measurement (by
the non-steady state reaeration method using cobalt and
sodium sulfite) and krypton-85 transfer measurement
with duplicate sampling points designated "left" and
"right" for each test. Table 1 summarizes the five well
water tests with reported Kkr/KLa values ranging from
0.71 to 0.89 and a mean for all ten values of 0.82 with a
212
-------
Figure 1. Relative Transfer Properties of Oxygen and Krypton-85 in Water
Ckr,0
o —
•f in
ID oo
I I
O 5.
§1
O*
81
O «o
li
D, = (C*-C) = DO deficit at time t
Kkr/KLa = 0.83
-kr,t ~
Time
213
-------
standard deviation of 0.06 (7% of the mean). Although
the scatter of these results is greater than that associated
with the experimental work at the USPHS, there is
reasonably good agreement with respect to the Kkr/KLa
values reported (0.83 + 0.04 USPHS and 0.82 + 0.06
Gordon and Etzel).
After the five well water experiments, the pool was
drained and refilled with a mixture of well water, cobalt,
instant milk, and activated sludge to create a biologically-
active condition. Eleven tests were conducted in this
mixture with simultaneous oxygen transfer evaluation (by
the non-steady state respiring system method using sodi-
um sulfite to disturb the equilibrium DO) and krypton-85
transfer measurement with duplicate sampling points as
used for the previous well water tests.
Table 2 summarizes the 11 wastewater tests with
Kkr/KLa values ranging from 0.62 to 2.07 and a mean
for all 22 values of 0.96 with a standard deviation of
0.29 (30% of the mean). With such a large range of
observed ratios there is obviously a problem with the
reliability of the test results.
Examination of the reported Kkr and KLa values shows
that while the krypton test results are quite consistant
with a mean «(K(,r)2o °f 0.27/min and a standard
deviation of about 4% of the mean, the oxygen results
are quite erratic with a mean a(KLa>2o of 0.30/min
and a standard deviation of about 20% of the mean. The
range of the 22 a(Kkrho values is 0.04/min (0.29-0.25),
which is about 15% of the mean value (0.27/min); the
range of the 22 reported ff(KLa)2o values is 0.28/min
(0.42-0.14), which is about 93% of the mean value
(0.30/min). It is clear that the oxygen test results in
wastewater are erratic. The reason for this problem
cannot be determined from the available data, but it is
suspected that the non-steady state respiring-system
method used for establishing the KLa values is inherently
subject to such problems. In comparison, the gaseous
tracer method provided consistent, reproducible results
for the entire test series.
Assuming that mixing conditions were the same for both
test series (well water and wastewater), an average alpha
value can be computed for the wastewater mixture by
assuming a = 1.0 for the well water:
a = Kkr, activated sludge/Kkr, well water = 0.27/0.29 = 0.93
This means that the net effect of the wastewater mixture
on gas transfer was a 7% reduction.
Accounting for Dispersion and Dilution. In order to make
the tracer method applicable to a full range of field
situations, it is necessary to add a dispersion/dilution
tracer to the gas. The dispersion/dilution tracer must
accurately trace the movement of water without being
subject to losses of any significance during the course of
a gas transfer study. Tritium, in the form of tritiated
water, has been selected as the dispersion/dilution tracer
for use with the gaseous tracer, krypton-85. For all
practical purposes, tritiated water is no more subject to
adsorption or other losses than ordinary water and,
because it is water, it is the ideal dispersion/dilution
tracer for use in gas transfer studies. Tritium is also
radioactive and can, therefore, be measured with good
accuracy at low concentrations by the same liquid
scintillation techniques used to measure krypton-85.
Triated water and krypton-85 are low hazard radiotracers
that can be safely used within regulatory limits.
Both the krypton-85 and the tritiated water are homo-
geneously mixed together in a single tracer-unit bottle.
The krypton-85 gas is completely dissolved in water, and
the bottle is filled with the water/tracer mixture so as to
exclude air space into which dissolved krypton-85
might transfer.
The tracer unit is released by quickly breaking the bottle
below the water surface in the operating aeration basin.
Both tracers are dispersed and diluted according to the
prevailing mixing and flow conditions in the system. The
tritiated water provides an accurate measure of the total
dispersion and dilution taking place. The dissolved
krypton-85 gas is subject to the same dispersion and
dilution as the tritiated water because both tracers are
released simultaneously in a homogeneous mixture. Also,
the dissolved krypton-85 gas is lost from the water
because of physical gas transfer but, being a noble gas,
krypton-85 is not subject to other significant losses.
The rate of change in concentration of the dissolved
krypton-85 gas relative to the rate of change in concen-
tration of tritium accurately reflects the rate of dissolved
krypton-85 gas transfer. This rate of krypton-85 transfer
can then be converted to the gas transfer rate for oxygen
from the known value of Kkr/KLa = 0.83 as previously
discussed.
Experimental Basis. Consider an aeration basin into
which a single homogeneous solution containing the two
tracers previously discussed is released at a point. By
sampling at a fixed point within the basin as a function
of time, the concentration ratio of krypton-85 to tritium
can be established for each sampling time such that:
(5) R = C /C
where Ckr t and CH,I are the concentrations of krypton-85
and tritium, respectively, in a single sample taken at time
t and R, is the concentration ratio of the tracers at
time t. Applying this ratio concept. Equation 3 may be
modified to:
(6)
dRt/dt = -"(Kkr)ob8Rt
with terms as defined previously. Solving Equation 6 and
rearranging terms, we obtain:
<7> *(Kkr)obs = ln(R2/R1)/-(t2-t1)
* aKkr is used interchangeably
214
-------
Table 1. Summary of Pilot-Scale Well Water Test Results Adjusted to 20°C (0 = 1.024) —
Data from Gordon (3)
Left Station
aKLa#
Run (min1)
1 0.39
2 0.41
3 0.35
4 0.35
5 0.36
(miif '"J
0.29
0.21
0.28
0.31
0.30
aKkr/ aKLat
0.74
0.71
0.80
0.89
0.83
-------
in which RI and R2 represent the krypton-85 to tritium
concentration ratios at times t1 and \2- respectively. Thus,
the use of tritium provides an automatic correction for
the effects of dispersion and dilution. The tracers yield
the necessary data to determine the actual krypton-85
transfer coefficient, aKkr, for the system and the
conversion to an actual oxygen transfer coefficient, oKLa,
is from the established gas exchange ratio, Kkr/KLa =0.83.
It is emphasized that the tracer method does not rely on
any theoretical assumptions that cannot be verified. The
tracer method is a direct measurement of gas transfer
under real conditions of mixing, water quality (biological
and chemical), flow, temperature, weather, etc.
Field Studies
In order to demonstrate the tracer method fully, full-scale
aeration tests were conducted in four operating waste
treatment oxygen transfer basins, two mechanical aera-
tion systems and two diffused air systems.
Kenwood Plant. The Kenwood wastewater treatment
plant is a 70,000-gpd extended aeration unit. The
aeration basin has a volume of 70,000 gal with one
7.5-hp mechanical surface aerator. The plant treats
domestic waste only. As the aeration tank is designed for
nearly complete mix, two sampling stations were suffi-
cient. Figure 2 shows the location of each sampling
station and the dose release point used for each of the
three studies conducted in the basin.
Sour/? Cobb Plant. The South Cobb wastewater treatment
plant has one mechanically aerated activated sludge
basin with a design treatment capacity of 2 mgd. The
aeration tank has a volume of 502,000 gal with four
10-hp mechanical surface aerators. The South Cobb
plant treats a combination of domestic and industrial
wastes. Figure 3 shows the location of each sampling
station. The dose location was the same for each of the
three studies conducted in this basin.
South River Plant. The South River wastewater treatment
plant has four parallel activated sludge basins equipped
with diffused air aeration. Each basin has a volume of
476,000 gal and a design treatment capacity of 3 mgd.
The air blowers provide approximately 2.1 mil ft3
air/day to each of the four units. Two of the basins use
Walker Process diffusers while the other two basins use
Water Pollution Control Corporation diffusers. The first
test at this plant was conducted in a Walker Process
basin and the second test was conducted in one of the
Water Pollution Control Corporation units. Figure 4
shows the sampling and dosing locations for both studies.
Field Results
From the eight separate field tracer studies, 23 independ-
ent observations of gas transfer capacity were obtained.
Figures 5 through 7 are plots of the observed krypton to
tritium concentration ratios versus time from typical
tracer studies. Table 3 is a summary of all field test
results.
In addition to providing accurate measurement of gas
transfer, the tracer method also provides an excellent
measure of dispersion and dilution. Figure 8 shows
tritium versus time curves for four sampling stations
used in Test No. 6 at the South Cobb plant, and Figure 9
shows the corresponding krypton to tritium ratio plots
with the ratio scale shifted to show all four stations on
one plot. If a more refined definition of the dispersion/di-
lution characteristics is needed, then more frequent
sampling would be required in the first few minutes after
the tracer release. It should be noted that while the
South Cobb plant is far from a complete mix system, as
demonstrated by the tritium curves in Figure 8, the gas
transfer rates are quite readily determined as illustrated
in Figure 9 because the tritium accurately and auto-
matically accounts for dispersion and dilution.
The average krypton transfer rate coefficient, (Kkr)obs,
obtained from the six observations at the Kenwood plant
was 2.4 hr~1 with a range of 2.21 hr~1 to 2.5 hr"1. The
average a(KLa)20 was found to be 2.6 hr"1 for the three
studies.
The Kenwood plant experienced a sludge recycle pump
failure causing mixed liquor quality to vary for each
study. For the three Kenwood studies, the gas transfer
raes increased, while the basin solids (settleable)
decreased. It is noted that there appears to be a strong
linear relationship between a(KLa)2o arid mixed liquor
settleable solids for the three test results.
Test
1
2
3
a(KLa)20
(hr1)
2.43
2.64
2.81
Solids
(ml/l)
100
50
12
Of course, it is not possible to draw any firm conclusions
from only three data points but if a linear relationship is
assumed, for every 1 % increase in settleable solids there
was a 1.5% decrease in the oxygen transfer rate
coefficient.
For each Kenwood test, the two sampling stations, A and
B, gave essentially idential oKLa values, which indicate
relative uniformity of gas transfer in the basin.
Three tests were conducted in the South Cobb plant and
a total of 11 a(KLa>2o values were obtained with a range
from 3.32 hr"1 to 2.65 hr"1. The South Cobb plant was
experiencing a solids build-up problem during the test
period and the plant operator reported mixed liquor
settleable solids ranging from 400 ml/1 to 600 ml/I. For
the South Cobb test of August 4, 1971 (Test No. 6),
settleable solids were measured at each of the four
tracer sampling stations and it was found that settleable
solids and tracer-measured «(KLa)20 values were related.
216
-------
Figure 2. Kenwood Plant Sampling Locations
Plan
Effluent
7.5-hp Aerator
Influent
T
Liquid
CO
•34''
Section
217
-------
Figure 3. South Cobb Plant Sampling Locations
Plan
M
00
1
/
Effluent
I
V
E®
_
g
c®
D®
) (
.
g
\ (
t v,
B®
g
\
J
A®
\
g
\
Dose
©
'
'
u
p
i
i
Influent
^^10-hp
rr
Aerator
n
n
Mil rrn ri'i i nrrn
t
[— f
J
Liquid — '
k /
T 1-
i
, 1
T1 — r
Cone *-
k 4
• i
/
^.
Section
-------
Figure 4. South River Plant (1 and 2) Sampling Locations
Plan
1
1
1
c
fi
<3
0
«§
§
'i
S.
0)
1
1
1
Influent
1
O
Dose
OA
Unit 2
O°
OE
1
O
Dose ^.^
O*
Unit 1
^
.^•^^
OB ^
<^^^
"^-?
^^-^
•^
>c^.
*.. ^_
*^.
— "^
-^--
^^^«
"""-
^^
.
To
5
f^~*
•*
-
^
f
T
V
s
J
» 1
IT)
1
tl
^
I
1 1
K
1
1
1
/
.X
— Header
t
L Liquid
Effluent
Section
219
-------
Figure 5. Kr85 to H3 Concentration Ratio vs Time for Kenwood Plant (Test No. 2, Sta. A)
o
S
QC
o
U.^
0.1
0.09
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.009
0.008
0.007
0006
OOO5
0.004
%
\
\
\
V
\
\
\
v*>
\
\
(Kkr)ob, = 0
r =-0
\
\
n\
\
.0386 mirf1 =
.99
V
\
• 2.32 hr'1
20
40
60
80
100
120
Time After Release (min)
220
-------
Figure 6. Kr» Concentration Ratio vs Time for South River Plant (1) (Tert No. 7. Sta. B)
0.2
.9
15
-------
Figure 7. Kr85 to H3 Concentration Ratio vs Time for South Cobb Plant (Test No. 5, Sta. B)
0.2
•2 0.1
(0
t£
O
I
+rf
0)
o
o
CJ
CO
I
0.09
0.08
0.06
0.05
0.04
0.03
(Kkr)obs = 0.0456 min"1 = 2.74 hr'1
r =-0.99
10
20
30
40
50
60
Time After Release (min)
222
-------
Figure 8. H3 va Time Curve* for South Cobb Plant (Teat No. 6}
£
a>
a
m
2200
1800
1600
1400
1200
1000
800
600
400
200
*
n
tl
i
•
3
0
o
i
T
r
~
a*
0 0
^
T
p
0 Station A
Q Station B
0 Station 0
T Station E
*,,•
HWL
— 0 i
r
>*
;g^ i
r
10 20 30 40 50
Time After Release (min)
60
70
223
-------
Figure 9. Kr85 to H3 Concentration Ratios vs Time for South Cobb Plant (Test No. 6}
c
o
c
0)
o
c
o
o
IT)
CO
Station (Kkr)obs @ 24°C
(hr1
• A 2.45
D B 2.43
0 D 2.74
T E 2.76
20 30
Time After Release (min)
224
-------
Table 3. Summary of Field Test Results — Data from Neal & Tsivoglou (6)
Plant
Kenwood*
South Cobbf
South River ^\
South River 2J
Date
6/25/71
7/2/71
7/9/71
7/15/71
7/22/71
8/4/71
8/16/71
8/24/71
Test
Number
1
1
2
2
3
3
4
4
4
5
5
5
5
6
6
6
6
7
7
7
8
8
8
Station
A
B
A
B
A
B
A
A
C
A
B
D
E
A
B
D
E
A
B
C
A
D
E
Water
Temperature
(°C)
25
25
23
23
23
23
24
24
24
24
24
24
24
24
24
24
24
25
25
25
25
25
25
(Kkr)obs**
(hr'1)
2.21
2.28
2.32
2.36
2.50
2.47
2.51
2.62
2.91
3.01
2.74
2.77
2.60
2.45
2.43
2.74
2.76
1.49
1.46
1.37
1.37
1.39
1.37
«(KLa)20***
(hr-1)
2.39
2.46
2.62
2.66
2.82
2.79
2.78
2.90
3.22
3.32
3.02
3.05
2.87
2.71
2.65
3.03
3.05
1.61
1.58
1.48
1.48
1.50
1.48
* See Figure 2 for station locations.
f See Figure 3 for station locations.
t See Figure 4 for station locations.
**Used interchangeably with aK|
-------
Sta.
A
B
D
E
o(KLa)20
(hr1)
2.71
2.65
3.03
3.05
Solids
(ml/I)
550
560
460
440
In this one test, for a 1% increase in settleable solids,
there was an approximate 0.5% decrease in the tracer-
measured a(KLa)20 value. It is concluded that the varia-
tions in transfer rates observed in the South Cobb basin
are real and probably reflect variable solids concentra-
tions within the basin. It is also possible that some of the
variations in a(KLa)2o reflect mixing patterns that are not
uniform throughout the basin.
The two South River studies were conducted in parallel
aeration basins with different diffusers in each. The
South River 1 basin was equipped with Walker Process
diffusers, and the South River 2 basin was equipped with
Water Pollution Control Corporation diffusers. The two
tests were conducted on different days, but the plant
operator reported essentially identical flow, DO, and
settleable solids (100-110 ml/I) for each test date.
The average a(KLa)2o value for the South River 1 basin
was 1.56 hr~1 with a range from 1.61 hr"1 at the
influent end of the basin (Station A) to 1.48 hr"1 at
Station C (see Figure 4). This pattern of higher gas
transfer rates at the influent end was consistent with
visual observations of greater mixing intensity at the
influent end. The average a(KLa>2o value for the South
River 2 basin was 1.49 hr"1 with little variation (1.48 hr"1
to 1.50 hr"1), and mixing intensity was visually uniform.
Because the two South River tests did not use the same
sampling locations (see Figure 4), it is not possible to
draw firm conclusions about relative transfer perform-
ance. A more direct comparison could be made by
simultaneously testing both basins and using identical
sampling stations. From the limited test results available,
it is estimated that the overall oxygen transfer rate
coefficient for the two systems was about the same.
However, the South River 1 basin had a spatial distribu-
tion of transfer (higher at influent end) while the South
River 2 basin exhibited rather uniform transfer rates.
Discussion
The field tracer technique described here was designed
to measure aeration performance in any aeration system.
The tracer method is a direct measurement of gas
transfer under existing hydrodynamic conditions and does
not depend on biological or chemical parameters as a
test basis.
The tracer method is not limited to use in complete mix
systems because the conservative tracer (tritiated water)
accurately accounts for the effects of dispersion and
dilution. As the tracer method is not limited to complete
mix systems, it can also be used to evaluate aeration
performance in large aerated lagoons widely used for
wastewater treatment.
Conclusions
The following conclusions have been reached as a result
of the studies presented here and other applications of
the tracer method.
1. The gaseous tracer method can be used to determine
the efficiency of gas transfer equipment in test tanks
or in operating wastewater treatment facilities.
2. The field tracer studies of gas transfer performance
have demonstrated that the method is practical and
safe, and that the results are precise and reproducible.
3. The gaseous tracer method can be used to study the
effect of pollutants on gas transfer (a).
4. Use of the gaseous tracer method can allow accurate
full-scale compliance testing of oxygen transfer
systems after they are installed in the field.
5. The precision of the tracer method can allow for field
tests of equipment installation patterns in typical
aeration basins.
Recommendations
The tracer method is intended for use where more
conventional testing methods are not applicable or
sufficiently accurate. The tracer method can be applied
in operating aeration systems without modification or
upset of normal plant operations to accurately measure
aeration performance under real operating conditions.
Because of its general applicablity and accuracy, the
tracer method can be used as a referee method in cases
of dispute and it can also be used as a reference
method in evaluating other testing techniques.
For new designs, engineers can specify aeration
performance and have accurate field testing conducted
after the system is in operation. For existing systems that
are not now functioning as desired, the tracer method
can be used to accurately determine the operational
capacity of existing equipment before designing
additional aeration facilities.
Tracer studies in operating aeration basins should also
include measurement of alpha and characterization of
basin contents with respect to physical, chemical, and
biological parameters. This concurrent data collection
would identify operating conditions under which the
measured aeration performance was achieved. As
accurate field test results become available, the U.S.
Environmental Protection Agency could develop and
maintain an aeration performance register that would
provide all basic data from each test. With the avail-
ability of accurate and reliable field performance data,
aeration design specifications can be more precisely
established and the goal of energy-efficient systems
can be realized.
226
-------
References
1. Churchill, M.A. et al. "Prediction of Stream
Reaeration Rates." Journal Sanitary Engineers
Division, Proceedings of the American Society of
Civil Engineers, 88, SA4, p. 1, 1962.
2. Gordon J.A. Communication. Journal Water
Pollution Control Federation, 49, p. 2495, 1977.
3. Gordon, J.A. and J.E. Etzel. "Mechanical Surface
Aerator Evaluation Using Radio-Krypton as a
Standard Indicator of Mass Transfer." Unpublished
Report, Research performed at Purdue Univ., W.
LaFayetteIN, 1968-1970.
4. Neal, LA. "Field Tracer Measurement of Aeration
Performance." M.S. Thesis, San. Eng., Ga. Inst. of
Tech., Atlanta GA, March 1972.
5. Neal, LA. "Pollutant Effects on Stream Reaeration.'
Symposium Proceedings, Direct Tracer Measure-
ment of the Reaeration Capacity of Streams and
Estuaries, U.S. EPA and Ga. Inst. of Tech.,
Atlanta GA, January, 1972.
6. Neal, LA. and E.C. Tsivoglou. "Tracer Measurement
of Aeration Performance." Journal Water Pollution
Control Federation, 46, p. 247, 1974.
7. Tsivoglou, E.C. et al. "Tracer Measurements of
Atmospheric Reaeration. I. Laboratory Studies."
Journal Water Pollution Control Federation, 37,
p. 1343, 1965.
8. Tsivoglou, E.C. "Tracer Measurement of Stream
Reaeration." FWPCA, U.S. Dept. of the Int.,
Washington DC, June 1967.
9. Tsivoglou, E.C. et al. "Tracer Measurement of
Stream Reaeration. II. Field Studies." Journal Water
Pollution Control Federation, 40, p. 285, 1968.
10. Tsivoglou, E.C. and J.R. Wallace. "Characterization
of Stream Reaeration Capacity." EPA-R3-72-012,
U.S. EPA, Washington DC, 1972.
11. Tsivoglou, E.C. and LA. Neal. "Tracer Measure-
ments of Reaeration: III. Predicting the Reaeration
Capacity of Inland Streams." Journal Water Pollution
Control Federation, 48, p. 2669, 1976.
227
-------
Section VI. Supplementary Papers
Although not planned in the original program, the following papers were presented, in part, at the Workshop.
Notes for Workshop Toward
an Oxygen Transfer Standard
James J. McKeown
NCASI
Tufts University
Medford MA 02155
Having been associated with oxygen transfer studies on
and off for over 20 years and most recently having
investigated the mixing and transfer capabilities of a
number of aeration devices functioning in paper industry
wastewater, I probably join with many others in question-
ing the need for a standard at a timo when biological
treatment is fait accompli in many industries and
municipalities.
It would appear that oxygen transfer into sewage and
industrial wastes is a simple straightforward, well-studied
function. Most responsible manufacturers have recogniz-
ed that and provide guarantees based on particular
applications. Guarantees have been provided in terms of
BOD removal which is the major end product of aeration,
albeit the one which carries the most unknowns and risk
or the bidder. It happens, however, to be the best
common denominator for the paper industry, and I
suspect for other users of the equipment as well. The key
to its use is sufficient knowledge of the wastewater
which requires treatment and the aeration system
employed.
We have represented at this conference the standard
confrontation between user and supplier which is often
resolved by price, maintenance history, and other condi-
tions which are not being directly addressed in this
Workshop. Even if successful in specifying the standard
conditions for testing and rating oxygen transfer units or
systems, it remains as just one element in the user's
evaluation of alternate systems. As power costs continue
to rise, it becomes a more important element. However,
we should ask the Society if it embarks on this venture to
also consider standards for evaluating additional
engineering features such as materials corrosion, motor
and gear reducer stress and expected life, ice, frost, fog,
foam, and odor generation. In short, the user should be
in a position to judge the equipment offered by a
manufacturer using as many standardized elements as
possible. The challenge here is to first truly evaluate the
contribution which would be made in standardizing
oxygen transfer testing (which would also allow the
development of new and improved transfer devices) and.
24
second, to consider system and other engineering
parameters which are often of equal or greater impor-
tance to the user.
I've made a few calculations relating oxygen transfer for
a 500 ton per day paper mill. I think the answers will
justify my interest in knowing if a 10% transfer difference
exists between aerators A and B, all else being equal.
First, industrial power costs vary considerably across the
USA. Hydropower in the Pacific Northwest may cost as
little as $0.005/kW-hr. Oil generated power in the North-
east could cost $0.05/kW-hr. Thus, the cost for oxygen
transfer for the example mill would vary between $018-
$1.80/ton of product or 0.23-2.3 cents/lb of BOD
removed. For the mill using aerator A or B with a 10%
transfer difference would mean a savings of between
$3,400-$34,000 per year. That's equivalent to $10-$100
daily and is certainly significant at the higher power cost.
I believe the paper industry would like to know if a
transfer difference of 10% or more exists between units
and more importantly what turndown strategies can best
be utilized in full-scale systems. Therefore, I'd like to
bring the attendees up-to-date on our studies dealing
with aeration systems and how aeration is currently
deployed in systems treating papermill effluents.
Workshop Presentation
Much of the information is contained in published articles
or reports (1X2X3). New material assembled for this
Workshop consists of reviewing the types of aeration
configurations currently being operated in the paper
industry (Table 1), the percentage of fixed and mounted
aeration (Table 2), the distribution of aerator manu-
facturers and types (Table 3) currently deployed in 50
paper industry systems, and the dimensions of some of
the industry's deepest and shallowest systems which use
conventional surface aerators (Table 4).
The performance of full-scale systems involves tank
geometry selected for any number of reasons. NCASI (1)
presented the hydraulic influence of various types and
228
-------
Table 1. Breakdown of Aeration Basin Configurations
Utilized in Biological Treatment of
Paper Mill Effluents
Table 3. Distribution of Aeration* Power Among
Manufacturers for 50 Paper Industry
Wastewater Treatment Systems
Lines of Cells
in Parallel
1
2
3
Total
No. of Cells in Series
in Each Line
1 2 3
27 12 4*
900
2 1 0
38 13 4
Total
43
9
3
55
*AII oxygen systems.
Table 2. Comparison of Fixed Mounted and
Floated Aeration
Plant Type
ASB*
ASB
AST'VAIR
AST/0 2
Total
Nameplate
Mount Horsepower
Float 36,900
Fixed 8,220
Fixed 6,310
Fixed 1,180
52,610
% Total
70
16
12
2
100
Aerator"
H3
L4
L2
H1
L3
H5
L7
L1
L6
H4
S1
L5
H2
S2
Total
Type
Low Speed (L)
High Speed (H)
Stationary (S)
'Excludes oxv
Nameplate
Total
10,040
9,950
9,650
6,520
5,060
4,535
2,070
1,040
480
450
300
150
80
20
50,345
Distribution by Type
qen systems.
Horsepower
% Total
20
20
19
13
10
9
4
2
1
1
1
—
100
% Power
56
43
1
*ASB — Aerated Stablization Basin.
••AST — Activated Sludge Treatment.
"Aerator designations listed refer to
manufacturer identification.
sizes of surface mechanical aerators operating with and
without sidewall restraints. These results are applicable
in most wastewaters. We also showed that for low-rate
systems an oxygen mass balance could be made around
individual aerators using measurements of DO increase
through the spray, velocity and direction measurement of
the resultant flow, and wire horsepower. Furthermore,
for papermill wastewaters, these transfer efficiencies
were compared to overall system efficiency using: (1)
oxygen uptake measurements and associated volumes of
basin and (2) overall BOD5 removal obtained in systems
with sufficient (and, therefore, nonlimiting) residual BOD5.
Using system performance based on BOD removal, it
appears that oxygen transferred in the high-rate activated
sludge systems is identical to that determined for the
lower-rate aerated stabilization basin (ASB) systems.
Information contained in NCASI's data bank was extract-
ed to show average aerator performance in activated
sludge and ASB systems (Figure 1) and ASB geometry
and power levels (Figure 2) which represent 1976-77-78
performance for a large number of installations.
Table 4. Physical Characteristics of Selected
ASB Systems
System
5 Deepest
5 Shallowest
Average
Area
(acres)
20
16
9
70
40
33
76
13
115
154
27
Avg. Depth
(ft)
25
21
21
20
20
6
7.5
8
8
8
13
Size Aerators
(hp)
40
75
75
150
60
100
60
30
50
100
—
229
-------
Figure 1. Average Efficiency of Activated Sludge and Aerated
Stabilization Systems as a Function of Aeration Capacity
-------
References
1. "A Study of Mixing Characteristics of Aerated Stabili-
zation Basins". Stream Improvement Technical
Bulletin No. 245, NCASI, 260 Madison Ave., New York
NY 10016, 64 pages (June 1971) — also partly in
JJ. McKeown and D.B. Buckley. "Mixing Character-
istics of Aerated Stablization Basins". TAPPI, 54, (10),
p. 1664, 1971.
2. "Results of a Cooperative Field Study of a Downflow
Bubble Contactor Aerator and a Conventional Surface
Aerator". Stream Improvement Technical Bulletin
No. 237, NCASI, 260 Madison Ave., New York NY
10016, 1970 — also partly in J.J. McKeown, G.W.
Gove, and A.M. Benedict. "Field Studies of Artificial
Aeration Using a Downflow Tube Contactor". Proceed-
ings 25th Ind. Waste Conf., Purdue University, 25,
(2), p. 865, 1970.
3. Vickerman, J.L and H.W. Gehm. "World's Largest
Deep Aerated Stabilization Basin in New Zealand".
Paper Trade Journal, p. 24, July 9, 1973.
231
-------
Oxygen Transfer in the
Activated-Sludge Process
Arthur G. Boon
Water Research Centre, Stevenage Laboratory
England, United Kingdom
Introduction
Dissolved oxygen is essential for the aerobic biological
treatment of waste waters. The activated-sludge process
is designed to increase the rate of biochemical oxidation
by bringing the waste water into contact with a high
concentration of micro-organisms in the presence of
DO. Given ideal conditions for the growth of the micro-
organisms, there is a maximum rate at which the
impurities in a waste water can be biochemically oxidized
which is related to the physical and biochemical proper-
ties of bacterial cells; it cannot be exceeded simply
because DO is readily available. It is possible to increase
the rate of treatment per unit volume of the aeration tank
by increasing the concentration of micro-organisms in
the mixed liquor, provided that the necessary quantity of
DO is made available. However, the concentration of
micro-organisms can only be increased if they can be
adequately concentrated by settlement and recycled.
The total quantity of oxygen required per unit time can be
calculated from the BOD of the waste water, its concen-
tration of ammoniacal nitrogen, and the rate of treatment
required (g BOD removed/g aerated activated sludge-day).
However, the maximum rate at which oxygen can be
dissolved (kg/m3 aeration-tank capacity-hr) depends on
the performance of the aeration system, the efficiency of
which is normally judged by the rate at which oxygen
dissolves and the electrical energy required.
Trjditional methods of aeration — largely confined to
diffused-air and mechanical surface systems — have
been used successfully for many years in the UK to
provide the oxygen required by micro-organisms in lightly
loaded treatment plants designed to achieve effluents of
high quality. Factors affecting the performance of such
systems of aeration have been investigated by many
research workers (4X5X6X8X9X10X13X14), although only
recently have the full importance of tank geometry and
the relation between aeration intensity and degree of
purification achieved been realized.
Oxygen Requirements
Total Quantity
The amount of DO required for treatment of a waste
water will depend on the total oxygen demand of the
micro-organisms oxidizing both carbonaceous and
nitrogenous matter. Data obtained for full-scale conven-
tional plants have shown (16) that the quantity of oxygen
25
required for oxidation of carbonaceous matter is approxi-
mately equal to the BOD of the waste water, assuming
nitrification has been inhibited in the BOD test. When full
nitrification is necessary, and facilities are not available
to utilize oxygen from nitrate by controlled denitrification,
the additional amount of oxygen required can be calcu-
lated by assuming that each gram of ammoniacal nitrogen
requires approximately 4.25 g of oxygen for conversion
partially to nitrate and partially to bacterial cells. To
ensure that the rate of nitrification will not be limited by
lack of DO, the concentration of DO in the mixed liquor
should be maintained near to the optimal value of
2 mg/l.
The amount of oxygen required for oxidation of carbon-
aceous matter will be affected by the rate of sludge
loading (g BOD/g sludge-day). At high loadings
(>0.6 g/g-day), more of the organic matter will be
converted into bacterial cells than when the loading is
low (<0.1 g/g-day). The oxygen requirement varies from
about 0.8 to about 1.6 times the BOD of waste water,
corresponding to sludge loadings of 1 to <0.1 g
BOD/g-day, respectively; at a conventional sludge loading,
corresponding to a sludge age of about 5 days, the
oxygen requirement is equal to the BOD of the waste
water.
Not all of the oxygen used to oxidize ammonia need be
lost as nitrate in the final effluent. In anoxic conditions,
that is in the absence of DO, micro-organisms in acti-
vated sludge will utilize nitrate for oxidation of carbon-
aceous matter; the nitrate is reduced in the process to
elementary nitrogen. Recent experiments have shown
that about 50% of the nitrate in the effluent from a plug-
flow activated-sludge plant can be utilized by providing
an anoxic region at the inlet end of the aeration tank (3).
The required anoxic conditions were achieved when
equal flows of recycled sludge and settled sewage were
mixed, but inadequately aerated, for about 40 min. By
reducing the intensity of aeration to produce a second
anoxic region, with a 30 min retention period, at about
the mid-point of a similar plug-flow aeration unit and
feeding 40% of the settled sewage to this region, denitri-
fication increased to about 70%. To obtain the benefits of
denitrification, and thus reduce the total quantity of
oxygen to be supplied by the aeration system, the rate of
input of DO and the concentration of micro-organisms in
the aerobic regions must be sufficient to achieve full
nitrification.
232
-------
Kate of Demand
The total quantity of dissolved oxygen required daily can
be calculated as described above, but the rate (kg/m3
mixed liquor-hr) at which oxygen has to be introduced by
the aerator (aeration intensity) to maintain aerobic
conditions is directly related to the rate of biochemical
oxidation in the mixed liquor. This depends on the period
for which the waste water is aerated, the input of BOD,
the concentration of suspended solids in the mixed
liquor, and the temperature of the mixed liquor.
The rate at which oxygen is consumed in an aeration
tank depends on variables such as the treatability of the
waste water (influenced by changes in pH value, the
presence of substances which may inhibit biochemical
oxidation, and the biodegradability of each constituent
substance), the population of micro-organisms, the con-
centration of DO, and the temperature. Some of these
variables may approach critical limits. A relation between
specific rate of removal of BOD (g BOD removed/g
activated sludge-day) in an activated-sludge plant and
BOD of effluent has been obtained (Figure 1) at the
Water Research Centre, using pilot plants treating a
constant flow of settled domestic sewage of average BOD
(250 mg/l) at temperatures of about 15°C. Diurnal
variations in flow, which caused variations in sludge
loading, were shown to affect effluent quality rapidly (1),
although the aeration intensity was adequate at all times.
For design of an aeration system, the maximum sludge
loading expected during the retention period of mixed
liquor in the aeration tank should be used if it is required
that the standard of effluent be maintained consistently.
Nines et al (11) have claimed that the maximum rate at
which oxygen would be required by high concentrations
of micro-organisms, treating a readily biodegradable
organic substrate under ideal conditions in a fermentation
unit, is approximately 10 kg/m3 aeration-tank volume-hr.
However, from an activated-sludge plant, it is generally
necessary to produce an effluent of low BOD (<20 mg/l);
the rate of biochemical oxidation is therefore limited by
available substrate and the concentration of micro-
organisms by the capacity to settle and recycle the
sludge. Data shown in Figure 1 indicate that, when the
BOD of the effluent was 20 mg/l, the rate of biochem-
ical oxidation was only 1 g/g-day. To dissolve oxygen at
that rate would require an aeration intensity of only
0.2 kg 02/m3-hr, assuming that the concentration of
mixed-liquor suspended solids was 5000 mg/l and no
additional oxygen was required for nitrification. This
intensity of aeration was easily achieved in the laboratory
with a fine-bubble diffused-air system, passing air at a
high rate through closely spaced diffusers at the bottom
of a shallow (0.6 m) tank, and frequently cleaning the
external surfaces of the diffusers. The maximum intensity
normally achieved with full-scale conventional aerators is
about 0.1 kg/m3-hr and this may limit the maximum rate
of treatment.
The rates at which carbonaceous and nitrogenous oxida-
tion occur are affected by changes in temperature of the
mixed liquor. Studies at the Water Research Centre (1X2)
(7) have shown that both rates increase as the tempera-
ture increases, and approximately double for an increase
from 7 to 17°C. However, long-term operation of a plant
at low temperatures may result in an increase in the
proportion of viable bacteria in the activated sludge, thus
increasing the total quantity of active enzymes present
and a gradual increase in temperature from 7 to 17°C
over a period of several months might not then result in
a doubling of the rate of oxidation. Rapid diurnal changes
in temperature of the mixed liquor will affect both rates
of oxidation, and the intensity of the aeration system
should be based on the maximum rates anticipated,
preferably as determined by using a respirometer or pilot
plant fed with sewage at the appropriate temperature.
Oxygen Supply
DO is conventionally supplied to the activated-sludge
process by either diffused-air or mechanical surface
aeration. Many different types of diffused-air systems
have been developed, using a variety of diffusers —
ceramic or compressed-plastics diffusers, spargers, drilled
pipes, or expanded plastics — to produce bubbles of
various sizes. Generally, air is diffused into the liquor
near the bottom of the tank so as to achieve adequate
mixing and the maximum contact time for aeration. There
are only two basic types of mechanical surface aerators,
differentiated by the plane of rotation which may be
vertical (cone type of aerator) or horizontal (brush type of
aerator).
Two commonly used criteria of performance of aeration
equipment are aeration capacity and aeration efficiency.
The aeration capacity is defined (13) as the rate of
absorption of oxygen during aeration of completely de-
oxygenated liquid at a specified temperature (either 10°C
or 20°C). Aeration efficiency is defined as the aeration
capacity per unit of energy supplied to the aeration
system at a given temperature. In all types of aeration
systems, the mass-transfer coefficient increases by about
2% of the value at 10°C per degree-centigrade increase
in temperature in the range 0 to 30°C. As the value of
the saturation concentration decreases with increasing
temperature by about the same amount, the effect of
changes in temperature on the aeration capacity or
efficiency can be neglected unless extreme accuracy is
required (4).
The ability of an aerator to dissolve oxygen is normally
measured indirectly by conducting tests in clean water
using the non-steady-state method described elsewhere
(4X13). Reproducible results can be obtained by these
tests, provided they are carried out under carefully con-
trolled conditions. This method is not wholly satisfactory
because clean water can easily become contaminated by
traces of surface-active substances which affect the rate
233
-------
Figure 1. Relation Between Rate of Removal of BOD of Settled Domestic Sewage Per Unit Mass of
Activated Sludge and BOD of Final Effluent
80
60
O)
*4
3
40
O
O
OQ
20
I
1 2
Rate of Removal of BOD Per Unit Mass of Sludge (g/g-day)
234
-------
of oxygen transfer (4X10X13X14). To minimize the effect
of possible contamination (particularly when testing full-
scale aerators), it has been widely accepted that aerators
are best tested in tap water to which sufficient anionic
surface-active material has been added to achieve an
average concentration of 5 mg/l during each test. Well-
purified sewage effluent or river water has been used
instead of tap water in such experiments and has been
found to be satisfactory (5). The rate of oxygen transfer in
mixing liquor will, however, be affected by changes in
such factors as degree of purification of the waste water,
configuration of the aerators in the aeration tank,
geometry of the aeration tank, and concentration of
dissolved oxygen. These will be discussed later.
The performance of an aeration system can be measured
during its operation in mixed liquor, using the steady
state method (4X13). The performance of a given system
in mixed liquor is similar to that obtained in tap water to
which 5 mg/l of anionic surface-active agent have been
added, but is substantially different from that obtained in
clean water when operated in the same tank under
similar conditions. In order to relate the rate of oxygen
transfer measured in clean water to that measured in
mixed liquor under similar conditions, a proportionality
factor, 'alpha', is used (13).
The 'Alpha' Factor and its Application
The value of 'alpha' for a given aeration system is equal
to the performance of the system (aeration efficiency or
capacity) in mixed liquor under specified conditions
divided by the corresponding performance under similar
conditions in clean water. Results of previous investiga-
tions (13) show that the performance of an aeration
system in mixed liquor varies significantly according to
the nature of the liquor, the contaminants present, and
the intensity of aeration.
The value of 'alpha' for a given aerator does not remain
constant. The efficiency of transfer of oxygen into mixed
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
(Figure 2) variations in 'alpha' 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 of these experiments (13)
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/l of added detergent, under similar
conditions of air-flow rate, temperature, tank geometry,
and diffuser configuration. The performance at the outlet
end 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 varies as the sewage
becomes purified. With such aerators, the value of 'alpha'
will decrease from about 1.2 initially, to a final value of
about 1, corresponding to a decrease in the concentration
of contaminants.
It has been postulated that a change in oxygen transfer
rate as a result of the addition of anionic detergent to
Figure 2. Variation of Alpha Factor with Degree of Purification of Sewage from Inlet to
Outlet of a Plug-Flow Aeration Tank
0.8 r
re
o.
•s °4
_
(0
Depth of water (m)
3.7 O
6.1 v
8.1
D
Inlet end
I
I
Outlet end
1
100 200 300
Oxygen Absorbed by Sewage (mg/l) During Treatment
400
235
-------
tap water is due to the physical rather than the chemical
properties of the detergent; it would be expected that the
addition of cationic or non-ionic detergents would have
similar effects (8). It is thought that the increase in
transfer rate caused by the addition of detergent to a
mechanical surface-aeration system is due to a decrease
in surface tension resulting in the formation of a very
large number of small droplets of water, the surfaces of
which are continuously reformed while they are in
contact with the atmosphere. In a diffused-air system,
the addition of detergent also decreases surface tension
and smaller bubbles of air are produced but, unlike the
effect described for surface-aeration systems where the
interfacial surfaces are continually renewed, the deter-
gent is thought also to form a stationary boundry layer at
the surface of the bubble, so decreasing the rate of
transfer of oxygen.
'Alpha' values do not indicate the relative efficiency of
aeration systems. A diffused-air system may have an
'alpha' value as low as 0.4 and a mechanical system one
as high as 1.2.
Relative efficiencies of different aerators can be judged
by comparison of their mass-transfer coefficients, aera-
tion capacities, or efficiencies, only when tested under
similar conditions in mixed liquor or tap water with added
detergent. For example, a diffused-air system may have
an aeration efficiency of 5 kg/kW-hr in clean water and
2 kg/kW-hr in mixed liquor, whereas a mechanical
surface aerator could have corresponding values of 1.5
and 1.8 kg/kW-hr, respectively.
Effect of Tank Geometry
The geometry of an aeration tank and the configuration
of diffusers will affect the performance of a fine-bubble
aeration system. The results of experiments carried out
at the Water Research Centre (13), given in Table 1,
show significant variations in performance of an aeration
system over a range of depths of immersion as a result
of using different numbers of dome diffusers. Increasing
the number of diffusers from five to ten increased the
aeration intensity and efficiency at each depth of immer-
sion, although the total flow of air, and hence the power
required, was reasonably constant. A further increase in
the number of diffusers to 25, with the same nominal
flow of air per dome, did not further increase efficiency.
Duplicate tests at a depth of 2.4 m indicate the repro-
ducibility of results obtained by careful measurement.
The significant effect of changes in tank geometry on the
performance of a fine-bubble aeration system at depths
up to 3 m has been confirmed (Figure 3) by tests carried
out with similar configurations of diffusers in aeration
tanks of different shape. The effect probably results from
changes in patterns of mixing and hence the relative
velocities of bubbles and water.
Variations in the performance of a surface aerator with
changes in its size and in the geometry of the tank have
not been studied extensively. Kormanik (12) claims that
the rate of oxygen transfer is not affected by depth of
water in the tank within the range tested (4 to 6 m),
although the rate increases as the power supplied per
unit surface area is increased. The rate of rotation was
found to affect aeration effeciency, in agreement with
earlier work by Downing et al (5). An increase in rate
from 26 to 34 rev/min increased the efficiency by about
50%. However, this has not been borne out by some
recent experiments with a full-scale mechanical surface
aerator. In these, it was observed that the efficiency of
the aerator was reduced by about 40% when the speed
of rotation was increased from 42 to 55 rev/min and
double the electrical energy was supplied. This reduction
in efficiency was thought to be caused by a deflection
Table 1. Conditions and Results of Experiments with Fine-Bubble Aeration of Water Containing
5 mg/l Anionic Detergent
Depth of
Still Water
in Tank (m)
1.2
2.4
4.9
Volume of
Water in
Tank (m3)
2.57
5.31
10.80
No. of
diffuser
Domes
in Use
5
10
5
10
10
5
10
10
25
25
Nominal
Flow of Air
Per Dome
(mVhr)
1.7
0.85
1.7
0.85
0.85
1.7
0.85
1.7
0.85
1.7
Total Flow
of Air
(m3/hr)
9.0
9.0
9.5
9.3
9.3
10.4
10.4
20.4
25.5
42.1
Aeration
Intensity
(g O2/m3-hr)
35.0
45.0
26.2
43.0
42.0
28.0
49.5
69.0
103.0
170.0
Aeration
Efficiency
(kg Oz/kW-hr)
1.97
2.53
1.46
2.38
2.33
1.57
2.87
1.94
2.31
2.34
236
-------
Figure 3. Effect of Depth of Immersion of Diffusers on Oxygen-Transfer Coefficients
in Different Shaped Tanks in Which Water Containing 5 mg/l Anionic Detergent Was Aerated
E
0.15
£ 0.1
£
'5
0.05
-O Tank with Flat Bottom
and 1 Row of 5 Diffusers
•A Modified Tank with
Furrow Bottom and
1 Row of 5 Diffusers
345
Depth of Water Above Diffusers (m)
plate attached to the top of the aerator and designed to
prevent liquor from being thrown out of the tank when
the aerator was operated at high speed. The deflection
plate prevented adequate circulation of the mixed liquor
and thus wasted power.
Poor mixing of the liquor by the surface aerator at one
treatment works resulted in inadequate aeration because
only a small proportion of the contents near the surface
of the aeration tank was recirculated and aerated. This
small proportion of the liquor contained high concentra-
tions of dissolved oxygen, which reduced the saturation
deficit and hence directly reduced the rate of oxygen
transfer.
An assessment of mixing and aeration by six surface-
aeration systems has been made by the British Carboni-
zation Research Association (15). They reported that
mixing of the contents of an aeration tank may be
inadequate with some types of surface aerator, and they
have suggested methods of improving performance.
Effect of DO Concentration
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 DO.
However, the rate of transfer of oxygen is directly propor-
tional to the DO saturation deficit. Hence, the rate at
which oxygen is transferred into mixed liquor in an
aeration tank will vary, depending on the concentration
of DO. For example, when nitrification is required, and
hence an aerator is operated in mixed liquor to maintain
consistently 20% of the saturation concentration of DO,
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 DO was present in the
mixed liquor, the rate and efficiency of oxygen transfer
would be only 0.4 of the maximum.
Performance of Different Types of Aeration Systems
Because of the variety of factors which affect the per-
formance of an aerator, comparable data for different
aeration systems can only be provided by specifying the
conditions under which a given liquid is being aerated
and by stating the range over which the performance is
likely to vary with, for example, tank geometry and
aerator configuration.
The values given in Table 2 are generally applicable to
most of the conventional systems available in the UK.
Fine-bubble diffused-air systems include both ceramic
and compressed-plastics types of d iff users, and systems
for coarse-bubble diffused air include spargers, drilled
pipes, and diffusers fabricated from expanded or woven
plastics.
237
-------
Table 2. The Aeration Efficiency of Various
Conventional Aeration Systems as Measured
in Tap Water Containing 5 mg/l Surface-
Active Agent
Aeration System
Aeration Efficiency*
(kg/kW-hr)
Diffused Air
Fine bubbles
Coarse bubbles
Mechanical Surface Aerators
Rotating vertically
Rotating horizontally
1.5-3.6
0.9-1.2
1.5-2.2
1.2-2.4
'Measured at maximum deficit of DO and
calculated from the total power supplied.
Aeration efficiencies of the various systems, as given in
Table 2, should only be used as a guide to the likely
power consumption. The selection of the best system for
a given situation should also take into account other
factors which can affect operating costs, such as flexi-
bility and reliability of operation, ability to maintain the
micro-organisms in suspension by effectively mixing the
contents of the aeration tank, and the need for regular
maintenance of equipment associated with the aeration
system (compressors, gear boxes, and motors). Differ-
ences in capital costs for each system should also be
taken into account.
Conclusions
Conventional aeration systems provide oxygen at a
relatively low rate although at a fairly high efficiency. To
increase the rate at which oxygen can be suppled by
conventional equipment would require installation of
more equipment, which may only be fully used for short
periods at times of high BOD loading. The maximum
aeration intensity achieved in practice is about 0.1
kg/m3-hr.
Selection of the most suitable aeration system to be used
in a given situation requires detailed consideration of
potential operational costs as well as capital costs.
Operational costs can be directly related to the energy
required to dissolve a given weight of oxygen at the
required rate. Nevertheless, the ability of the aeration
system to maintain the micro-organisms in suspension by
effectively mixing the contents of the aeration tank and
the flexibility and reliability of operation must also be
taken into account in the selection of a system.
Acknowledgement
This paper is an abridged version of one previously
presented to an international research symposium on
'New Process of Waste Water Treatment and Recovery'
organized by the Society of Chemical Industry and held
in London on September 6-8, 1977. It is published by
permission of the Director, Water Research Centre.
References
1. Boon, A.G. and Burgess, D.R. "Effects of Diurnal
Variations in Flow of Settled Sewage on the Perform-
ance of High-Rate Activated-Sludge Plants". Wat.
Pollut. Control, 71, p. 493, 1972.
2. Boon, A.G. and Burgess, D.R. 'Treatment of Crude
Sewage in Two High-Rate Activated-Sludge Plants
Operated in Series". Water Pollution Control, 73
p. 382, 1974.
3. Cooper, P.F., Collinson, B. and Green, M.K. "Recent
Advances in Sewage Effluent Denitrification: Part II."
Paper presented to the West Midland Branch of the
Institute of Water Pollution Control, Birmingham,
p. 28, January 1976.
4. Downing, A.L. "Aeration in the Activated-Sludge
Process". Journal Institute Public Health Engineers
59, p. 80, 1960.
5. Downing, A.L, Bayley, R.W. and Boon, A.G. 'The
Performance of Mechanical Aerators". Journal Proc
Inst. Sew. Purif., p. 231, 1960.
6. Downing, A.L, Boon, A.G., and Bayley, R.W. "Aera-
tion and Biological Oxidation in the Activated-Sludge
Process". J. Proc. Inst. Sew. Purif., p. 66, 1962.
7. Downing, A.I—lones, 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. Downing, A.L. Scragg, L.J. and Boon, A.G. "Aeration
in Biological Purification Processes". Society of
Chemical Industry, Monograph No. 12, p. 187, 1961.
9. Downing, A.L. and Wheat land, A.B. "Fundamental
Considerations in Biological Treatment of Effluents".
Trans. Instn. Chem. Engineers, 40, p. 91, 1962.
10. Eckenfelder, Jr., W.W. and Barnhart, E.L. "Effect of
Sewage Waste Characteristics on Oxygen Coeffi-
cients". USPHS Proj. No. 4694, Oct. 1962.
11. Hines, D.A., Bailey, M., Ousby, J.C. and Roesler, F.C.
"The ICI Deep Shaft Aeration Process for Effluent
Treatment". In 'The Application of Chemical Engi-
neering to the Treatment of Sewage and Industrial
Liquid Effluents". Institution of Chemical Engineers,
Symposium Series No. 41, 1975.
238
-------
12. Kormanik, R. "How Does Tank Geometry Affect the 15.
Oxygen Transfer Rate of Mechanical Surface
Aerators?" Water Sewage Works, 123, No. 1, p. 64,
1976.
13. Lister, A.R. and Boon, A.G. "Aeration in Deep Tanks:
An Evaluation of a Fine-Bubble Diffused-Air System". 16.
Water Pollution Control, 72, p. 590, 1973.
14. McCabe, J. and Eckenfelder, Jr., W.W. "Biological
Treatment of Sewage and Industrial Wastes". Vol. 1.
"Aerobic Oxidation". New York, Reinhold Publishing
Corporation, 1956.
"Mixing and Aeration in Biological-Treatment Plants:
A Study of Six Designs of the Surface-Aerated
System". Carbonization Research Report No. 1,
(British Carbonization Research Association,
Chesterfield, Derbys.), 1974.
Porter, K.S. and Boon, A.G., Trent Research
Programme. "Cost of Treatment of Waste Water with
Particular Reference to the River System of the Trent
Area". Presented to the Institute of Water Pollution
Control Symposium, Nottingham, April 1971.
239
-------
Effect of Gas Phase
Temperature on Oxygen
Saturation Value
James A. Mueller, Martha L. Quintana, and Dominic DiToro
Manhattan College
Bronx NY 10017
26
Theory
Henry's Law states that the amount of gas dissolved by
a given volume of liquid, at constant temperature, is
proportional to the pressure of the gas with which it is
in equilibrium.
This law can be expressed by the following equations:
M) C* = Hp02
H = Henry's Law constant, mg/l-atm
C* = gas concentration at saturation, mg/l
Po = partial pressure of gas, atm
(2) Po2 = y(PrPv)
y = dry weight fraction of the gas in the
mixture. For oxygen in air, y = 0.209.
p, = total pressure, mm Hg
pv = vapor pressure of liquid, mm Hg
(3) p, = pb + (h/13.6)
pb = barometric pressure, mm Hg
h = manometer reading, mm H20
This study shows the effect of the gas phase temperature
on the oxygen saturation value in water at constant liquid
temperature for a water vapor saturated gas phase.
Apparatus
The arrangement of the equipment used in this experi-
ment is depicted in Figure 1.
The experiment was performed in a closed system
consisting of a reactor with water at a constant tempera-
ture. In order to keep the temperature of water in the
reactor uniform, the reactor was placed in a large water
bath with circulating tap water. An electrical heater was
used in the bath to bring the water temperature to the
desired value. Room air was compressed, pre-saturated,
and heated prior to entry into the reactor. Gas flow was
monitored in the rotameter and kept constant throughout
the experiment.
The temperatures of gas and liquid in the reactor, the
reactor pressure, and the oxygen concentration in the
liquid phase were recorded every 5 min, as well as the
air flow rate and the air temperature in the saturation
flask. These data were recorded for 60 min during every
run, at the end of which three samples of water were
taken from the reactor to determine the final concentra-
tion using the Winkler test. This analysis provided a
check of the values for dissolved oxygen read during the
experiment from the oxygen meter (YSI). The barometric
pressure was recorded at the beginning and end of every
run. All runs were performed at constant agitator speed
and torque.
Results
Table 1 summarizes the results for two runs at two
different liquid temperatures. In each case the tempera-
ture of the gas was varied, and Henry's constant was
calculated using Equation 1 of the theory. The gas flow
and the agitator speed were kept constant at 2.6 scfm
Table 1. Results
Measured Values
Run
1a
1b
1c
2a
2b
2c
T|iq
(°C)
20
20
20
30
30
30
Tgas
(°C)
20.6
30.5
39.1
21.4
30.3
39.1
C»
(mg/l)
9.2
9.0
8.7
7.5
7.4
7.4
Pb
(mm Hg)
772.0
770.3
769.4
758.0
760.3
771.4
h
(mm H2O)
39.0
55.0
60.0
44.0
58.0
63.0
Pv*
(mm Hg)
18.2
32.7
52.7
19.1
32.4
52.7
Pt
(mm Hg)
(Eq. 3)
772.3
770.7
769.8
761.2
764.6
776.0
PrPv
(mm Hg)
754.1
738.0
717.1
742.1
732.2
723.3
Po2
(atm)
(Eq. 2)
0.207
0.203
0.197
0.204
0.201
0.199
H
(mg/l-atm)
(Eq. 1)
44.4
44.3
44.2
36.8
36.8
37.2
"From Perry's Chemical Engineers' Handbook, 4th Ed.
240
-------
Figure 1. Reactor
fl
Air Pump
Rotameter
Oxygen
Analyzer
(Beckman)
Gas Thermometer
\
i
d — r
D
i
Heater
Mag Mix
Variable Speed
Agitator
\
Speed Torque
Vent
*
D-O
,Gas Thermometer
iquid Thermometer
•Electrical Heater
Manometer
-Bath Thermometer
•Water Bath
Sewer
-------
Figure 2. Vapor Pressure, pv. Saturation Concentration, C*, and
Henry's Constant, H, vs Gas Temperature, TGAS
(1)TLIQ = 20°C
60
-P 50
o>
E
I
O)
40
30
20
15
20
25
30
TGAS <
35
40
9.0
8.0
u
7.0
6.0
60
E 50
o>
o>
E
40
Q. 30
20
15
(2)TLIQ=30°C
20
25
30
TGAS(°C)
35
40
10.0
9.0
8.0 -§
*
u
7.0
242
-------
and 300 rpm, respectively. Torque itself was not recorded
but was low due to the low speed and small diameter
propeller.
Figure 2 shows the variation of vapor pressure, satura-
tion concentration and Henry's constant with gas
temperature at two different liquid temperatures, 20°C
and 30°C.
Comparison with Published Data
From Perry's Chemical Engineers' Handbook, the Henry's
Law constants expressed in atm/mole fraction are:
T liquid (°C)
20
30
H (atm/mole fraction)
4.0 x104
4.8 x104
In order to compare the above values with those reported
in Table 1, the units should be converted into mg/l-atm
as follows:
For Tljq = 20°C
H, ] x
(4x104 total kg moles-atm/kg mole 02)
32.0x106 mg 02/kg mole O2
(18.02 kg liquid/total kg mole) x (1.0 I/kg)
H = 44.4 mg O2/l-atm
Similarly, forT,jq = 30°C, H = 37.0 mg 02/l-atm
Table 2 shows a good agreement between the experi-
mental and published values for Henry's constant.
Table 2. Henry's Law Constants
WO
20
30
Experimental H
Value (average)
(mg/l-atm)
44.3
36.9
H Value from
Perry
(mg/l-atm)
44.4
37.0
Conclusions and Recommendations
1. From the results obtained in this experiment, it is
concluded that the value of Henry's constant is
independent of the temperature of the gas phase at a
given temperature of the liquid phase when the gas
phase is saturated with water vapor.
2. The experimental values of Henry's constant are
in agreement with the values reported in the
literature.
3. It is recommended to perform a series of runs similar
to those reported in this study varying the speed of
the agitator and varying gas phase vapor pressures.
243
-------
Section VII. Working Group Summary Reports
Group A. Philosophy of the
Standard
We are proposing an interim standardized method of
testing aeration equipment with respect to ability to
transfer oxygen to a reference water. The standardized
method should include all details for determination of the
performance of the equipment for which there is a
consensus by members of the sub-groups. These details
should include the mechanics of the tests, sampling,
chemical analysis, interpretation of data, and power and
air flow measurements.
It is further proposed that this interim standard be
included in a Procedural Manual within two years in
accordance with the provisions of the EPA grant. The
Procedural Manual will reflect the present state of the art.
In implementing the intent of this interim standardized
method, the consultant is responsible for providing
manufacturers with a volume of aeration basin and a
range of oxygen requirements (mg/l-hr) which accounts
for alpha, beta, oxygen deficit, etc., when conducting the
standard aeration tests in the reference water. Details of
any constraints, such as land availability and ground
conditions, would also be provided.
The manufacturer will provide power requirements to
achieve those specified rates at a given oxygen concen-
27
tration together with details of capital costs, operation
and maintenance costs, noise, flexibility, reliability, tank
geometry, etc.
The system will be selected on the basis of cost effective
analysis. After selection of the system, bidding is opened
to all manufacturers who can meet the specified require-
ments as determined by the proposed interim standard
of testing.
It is expected that the proposed interim standard method
will be used for both shop testing and field testing of the
aeration equipment.
Group A. Participants
*P.J. Kransnoff
E.L Barnhart
A.G. Boon
W.C. Boyle
R.C. Brenner
LA. Ernest
G.R. Fissette
E.D. Simmons
'Chairman
244
-------
Group B. Modelling and
Data Interpretation
The Model
The model for the analysis of non-steady state clean
water oxygen transfer data is:
28
C = A- Be
o
§
o
O
Q
n
U
t =
time
Qualifications:
1. A given aerator in a given tank
2. Steady state hydraulic behavior
3. Tank is completely mixed
4. Temperature is approximately constant
This model is similar to the familar integrated form:
C = C*-(C£-C)e-(K'-a>t)
Group B recommends use of the A, B and D symbols to
allow for differences in the way in which these para-
meters are interpreted for mass transfer models applied
to subsurface aeration systems.
Parameter Estimation
The parameters A, B and D should be estimated by
fitting the model to concentration-time data by means of
a non-linear least squares regression technique. The
advantages of the parameter estimation techniques are:
1 It works directly with the observed concentration, C.
2. Truncation of the data as DO concentration
approaches saturation is not necessary.
3. For a given set of data, the parameter estimates given
by this technique are more precise than those given
by other commonly used techniques.
The log deficit method (applied with a measured value of
A = QJ ) and the direct method are useful for approximate
parameter estimates. However, the differentiation in-
herent in the direct method produces a variable dC/dt
having larger error than the error in C. The larger error in
dC/dt causes the error in the parameter estimates from
the direct method to be larger than that of the other
methods. When the non-linear regression analysis is
applied to analyze non-steady state test data, the
standard deviation of the parameter estimates should be
reported and the error structure should be examined. The
error structure may reveal possible surging effects,
adequacy of the model and uniformity of experimental
error.
Truncation of Data
Values of C less than 10 to 20% of the saturation value,
A, may be truncated to avoid lingering effects of deoxy-
genation techniques and because these data contribute
little to the ability to estimate A and D precisely. Trunca-
tion of the data as the DO concentration approaches
saturation is discouraged because these data significantly
influence the parameter estimates for A and D. It is
recommended that the test be continued as long as
practicable. A period of time approximately equal to six
divided by the anticipated value of D is suggested.
It is recognized that temperature variations may affect
the precision and accuracy of the test. Therefore, as a
minimum, it is suggested that temperature be measured
at the beginning and end of the test and that a desirable
variation be 2°C or less. The average and extreme values
of temperature should be reported. It is recommended
that the sensitivity of the A and D estimates with respect
to temperature be investigated.
245
-------
Experimental Design
It is recommended that the analysis be based on a
minimum of 10 to 15 data points. Approximately two-
thirds of these points should be evenly distributed over
the period of time covered by 0.5/D to 2/D. The remain-
ing one-third of the points should be evenly distributed
over the period of time covered by 2/D to 6/D. In cases
of rapid transfer, the minimum time between data points
should be 0.5 min.
When multiple sample points are used, the data from
each sample point should be analyzed separately and
values of the separate A and D estimates determined
and reported for each sample point. The average should
also be determined. The uniformity of these estimates
may provide useful information about the aeration
system.
Interpretation of Parameter Estimates
The physical significance of the parameter estimates
depends on the aeration system being tested. For a
surface aerator:
A = C*, D = KLaand B = C*-C
For a submerged aeration system, the importance of gas-
side oxygen depletion should be recognized in determin-
ing a mass transfer coefficient. It is recommended that
any manual of practice resulting from Subcommittee
efforts include a development of a relationship between
D and the volumetric mass transfer coefficient, KLa,
similar to that given by the Downing-Boon Model, e.g.:
Group B. Participants
*C.R. Baillod
*LC. Brown
B. Barrett
D.L. Bennett
R. Kapur
M.B. Lakin
J. Lee
A.D. Nardozzi
D.T. Redmon
C. Scaccia
M.K. Stenstrom
"Co-chairman
246
-------
Group C. Non-Steady State
Clean Water Testing:
Shop and Field
The charge for Group C was to consider what procedures
and measurements should be included in a non-steady
state clean water testing procedure. The group size varied
from 15 to 17 members (see listing of those present). A
specific topic was presented and discussed, and a state-
ment representing a viewpoint was drafted and voted
upon for inclusion in the Group report. The statements
reported below were approved by a majority vote.
Statements 1, 4-7, 11 and 13-31 were approved unani-
mously. The results of the other voting were as follows:
Statement Number Yes No No Opinion
2 16 1 0
3 13 3 1
8 863
9 14 2 1
10 10 3 4
12 13 1 0
The discussion of the topics was very thorough and
spirited, and the Group should be commended for its
drive and effort to reach reasonable agreement on
consensus statements.
It is recommended that the testing procedural document
include:
1. A requirement that specified conditions for reporting
air flow and observed data relating to the existing
conditions of measurements be recorded.
2. A statement recommending the use of appropriate
codes or manuals for air flow measurement.
3. Appropriate procedures for system pressure
measurement exclusive of air flow requirements.
4. A recommended calculation procedure for determin-
ing horsepower for air diffusion systems.
5. Appropriate procedures for measuring power for
mechanical devices.
6. The requirement of reporting the detailed description
of the test geometry for any test conducted.
7. Reference to cleaning the basin and the use of fresh
tapwater.
8. A statement regarding the preferred water tempera-
ture range for testing.
9. A requirement that an analysis of the physical and
chemical properties of the test water be provided to
the parties involved in the testing.
10. A specification of initial numerical limits for certain
physical and chemical water quality parameters.
29
11. A statement regarding the conditions for achieving
steady state physical conditions prior to test runs.
12. A statement indicating that nitrogen stripping is an
acceptable procedure.
13. A statement that the use of analytical grade sodium
sulfite for deoxygenation is an acceptable procedure.
14. The recognition that technical grade Na2SO3 can be
used as a deoxygenating chemical with the provision
that adequate quality control to avoid interferences
be followed.
15. A statement recommending the use of CoC^-Gh^O in
solution form as the source of cobalt catalyst for
deoxygenation.
Note: The Group altered its editorial approach at this
point and has approved the following additional
statements.
16. The testing procedural document should address the
proper cobalt concentration for testing with due
consideration given to adequate catalyst (say >0.10
Co as Co) for deoxygenation and provisions for elimi-
nation of DO testing interferences with higher
concentrations. A probable maximum concentration
of 2.0 mg/l as Co was suggested. (Comment: It was
suggested that testing for adequate cobalt catalyst
should be required.)
17. The standard procedure should require a sulfite
addition adequate to reach zero DO and to remain at
zero for a reasonable period (say 2 min).
18. The addition of sulfite in slurry form should be the
recommended procedure with an alternate procedure
for dry chemical addition delineated, and considera-
tions for multiple point addition should be addressed.
(Comment: It was suggested that the sulfite be added
in solution form.)
19. The standard procedure should specify a minimum
number of sample points (say three) with due con-
sideration to true representation of the tank contents
and regions of variable DO levels.
20. The standard procedure should include provision for
gravity or submersible pumped samples with the
utilization of anti-air entrainment inlets, consideration
of pressure changes in gas releases, minimum
sample transport times, low bottle displacement
times and an adequate number of samples obtained
between the test truncation limits.
21. The use of DO probes for in-place and external
measurement should be included as an acceptable
analytical procedure. It should be duly noted that
247
-------
appropriate detailed analytical care must be exercised
to assure accurate results with probe measurements.
(Comment: It was observed that a single sampling
and analytical procedure was not recommended and
should be considered.)
22. The testing document should address the question of
allowable number of sulfite additions per tank of
water with consideration given to setting testing
limits of sulfate or total dissolved solids concentration.
23. The testing document should provide the option of
deleting the initial run on a fresh tank of water to
account for uncertain interactions with sulfite
addition.
24. The testing document should address the question of
chemical interferences and iodine vaporization in
Winkler DO analysis.
25. The testing document should address the questions
of termination of runs in the case of repetitive testing
with consideration of length of run necessary to
reach C* in every case and obtaining appropriate
data for the recommended method of calculation
of KLa.
26. The testing document should define a common
procedure for attaining a measured oxygen satura-
tion value.
27. The testing document should address the number of
repetitive tests required and the reproducibility limits
of multiple testing.
28. The testing document should address the question of
allowable percent KLa variation in individual sample
points and the interpretation of this variation.
29. It is the considered opinion of this Group that the
testing document should focus on clean water testing
only and surfactants should not be used to modify
the clean water test liquid basis.
30. The testing document should consider promulgating
appropriate terminology for defining the performance
acceptance criteria for different generic types of
aerators.
31. Group C questions the use of variations in KLa and
bottom velocities in clean water testing as true
indicators of mixing capability for solids suspension
in aeration tanks.
Group C. Participants
»W.L Paulson
*T.C. Rooney
J.A. Bell
W.L Berk
H. Brociner
T.J. Helbing
L.W. Lestochi
J.J. Marx
D.A. McCarthy
S. Nelson
LH. Piper
R.N. Salzman
J. Smart
V.N. Wahbeh
D.J. Weis
J.D. Wren
F.W. Yunt
"Co-chairman
248
-------
Group D. Alpha, Beta, and
Temperature Corrections
General
Alpha is affected by wastewater constituents such as
soluble BOD, COD, suspended solids concentration,
surface tension of the wastewater, and temperature and by
system parameters such as type of aeration device,
power level, and basin configuration. In general, the
influence of these variables can be reduced substantially
by conducting the alpha measurement at field design
conditions, i.e., testing at waste temperature and using an
aeration device and power level similar to the actual field
condition.
Beta is influenced by wastewater constituents such as
salts, organics, and dissolved gases and by barometric
pressure and temperature. Conducting measurements at
equal barometric pressure and at field design tempera-
ture should reduce variations to those caused by
wastewater constituents only.
Theta, the temperature correction factor, is not as well
understood as alpha and beta. Several theta factor
"constants" have been reported in the literature. The
relationship between theta and variables such as aera-
tion device, power level, system geometry and wastewater
impurities is not known.
Alpha
A great deal of literature has been published regarding
measurement of the alpha factor and the difficulties
involved both with the apparatus used and liquids being
tested. The type of aeration device, mixing intensity, and
wastewater constituents present are only three para-
meters that affect the outcome of the alpha measurement.
Researchers are not in agreement concerning such
questions as the effect of suspended solids on alpha,
mixing level vs KLa, surfactants effect, and the type of
aeration device to be used, to mention a few. The litera-
ture contains more controversy regarding these questions
than reinforcement.
Of all the variables that potentially influence alpha
measurement, these four are by far most significant:
1. Test apparatus and aeration device
2. Turbulence
3. Suspended solids
4. Surface active agents.
Currently there are no standard procedures which take
these factors into consideration. As a result, alpha
measurement data can be extremely variable and
misleading.
30
Beta
The beta factor has been defined as the ratio of the
dissolved oxygen saturation concentration for wastewater
at field conditions to the saturation value for clean or tap
water at field conditions. Beta has been commonly
referred to as the salinity correction factor because
dissolved salts reduce oxygen solubility in wastewater. In
addition, dissolved organics and gases in the wastewater
can reduce oxygen solubility as well. Unfortunately, these
constituents also affect the measurement of dissolved
oxygen.
Several methods are available for measurement of
dissolved oxygen. The various methods can be divided
into three principle categories: 1) laboratory methods,
2) electrochemical analysis, and 3) membrane electrode
methods. The most popular and most widely used labora-
tory method is the Winkler Method However, the
amperometric membrane technique is probably the most-
used electrode method for measurements in wastewater.
Dissolved organics and dissolved gases can interfere with
the rate at which oxygen passes through the electrode
membrane and can cause chemical interference in the
Winkler Method.
An alternate method of determining beta is to compute
the ratio of the saturation concentration for tap water at
field temperature, pressure, and salt concentration to the
saturation concentration for tap water at field tempera-
ture and pressure and zero salt concentration.
When using a probe to measure dissolved oxygen, one
must be aware of the sensitivities of the instrument to
the waste constituents present in the sample. The probe
manufacturer's literature should be referred to prior to
calibration and making measurements. If the literature
does not contain information on interferences, then the
oxygen analyzer manufacturer should be contacted.
Theta
Temperature affects both equilibrium values for oxygen
concentration and the rate at which transfer occurs due
to changes in kinematic viscosity, which in turn affects
oxygen diffusion into the liquid. The effect of temperature
on the overall oxygen transfer rate coefficient is usually
expressed as follows:
The temperature correction factor, theta, has been report-
ed to vary from less than 1 .01 to more than 1 .05. A
249
-------
comparison of four temperature correction factor refer-
ences is presented in Figure 1. As can be seen from the
curves, temperature correction factors vary by approxi-
mately 15% between 10°C and 30°C. Most field
conditions are within this temperature range. Caution is
advised, however, when correcting field data to standard
conditions outside the 10°C to 30°C range where
variances become greater. Also, the presence of surface
active agents and other waste constituents may affect
the temperature correction factor. The potential effects of
system variables such as tank geometry, power level, and
aeration device are not known. If system variables do
influence theta, the magnitude of their influence is
probably not significant.
Future Emphasis
The complete lack of concensus regarding the signifi-
cance and influence of the factors affecting alpha and
beta determinations demonstrates the need for a
standardized procedure for measurement of alpha, beta
and theta factors. Although Standard Methods has a
tentative procedure for determination of alpha and beta
(207B5b and c), much work remains to be accomplished
in the standardization of an alpha and beta factor measure-
ment procedure. Several key factors concerning a
measurement procedure are contained in Figure 2. Before
detailed procedures can be developed, the major
questions such as: 1) What defines equivalent mixing
level? 2) What type of aeration device should be used?
3) Should surfactants be used in alpha testing and why?
and 4) Are suspended solids in or out, must be addressed
and answered? Only after a concensus has been reached
regarding these major factors in alpha and beta measure-
ment, can a well defined standard procedure be adopted.
Future emphasis of Group D shall be as follows:
1. Research literature for all known information on
alpha, beta, and theta.
2. Define the significant factors influencing alpha, beta,
and theta measurements.
3. Recommend research needs to better understand
these influences on measurements.
4. Develop a recommended procedure for alpha and beta
(and theta, if necessary) measurements.
5. Develop a standardized test apparatus and procedure
for alpha measurement.
Group D. Participants
*G.L Shell
J.K. Bewtra
D.G. Fullerton
J.S. Hunter, III
C. Matsch
R.P. Milne
J.E. Mueller
'Chairman
Figure 1. Effect of Temperatue on Oxygen Transfer Rate. (KLa)T =(KLa)2o0(T~20)
2.0 r
1.8
1.6
o 1.4
__CN
CO
CO
*
1.0
0.8
0.6
-rO'Connor: (K,a)1/(K.a)2 =
0= 1.024
9 - 1.02 Eckenfelder
O'Connor
,--
0=1.01 Shell
-0 = 1.03 Shell
•6= 1.024 I
'0= 1.02 Eckenfelder j
0
10
20 30
Temperature (°C)
40
50
250
-------
Figure 2. Considerations for Alpha and Beta Measurements
Purpose for Conducting Alpha and Beta Tests
Design New
System
Augment Existing
System
1
Wastewater
Identification
1
Evaluate Existing
System
Surface
Mechanical
Select a Bench-Scale Aeration Device
Two-Phase
Jet
Static
Establish Full-Scale Clean Water KLa
Determine Bench-Scale Clean Water Mixing
Level to Produce Full-Scale Clean Water KLa
Conduct Alpha and Beta Tests
1
Range of
KLa values
Repeat Tests for:
Different
Wastewater
Streams
Range of
Temperatures
Different
Aeration
Device
Reproducibility
251
-------
Group E. Steady State
Field Evaluations
The objective of Working Group E was to evaluate the
relative merits of steady state and non-steady state
methods of oxygen transfer and make recommendations
regarding the validity of each method. In order to facilitate
the discussion, the subject was divided according to the
areas listed below:
1. Types of system in which tests would be conducted:
a. Shop test system (normally clean water)
b. Acceptance test system (in shop or field
installations)
c. Evaluation test system (full-scale)
2. Type of test
a. Non-steady state - batch vs continuous flow
b. Steady state - batch vs continuous flow
3. Type of equipment (diffused, mechanical, oxygen
enriched)
4. Goal of test - must meet specifications expected for
oxygen transfer only (other specifications would, of
course, have to be met).
After considerable discussion, the problem area was
divided into two major categories of responsibility: one
for the equipment manufacturer, the other, for the design
engineer. Specifically the equipment manufacturer should
be provided by the design engineer the clean water
oxygen transfer rate and associated horsepower for
equipment for operation in the final system. The equip-
ment should be tested in clean water (without detergent
added) after installation. If the transfer rate is achieved at
the specified horsepower, the oxygen transfer require-
ment will be satisfied by the manufacturer. The test for
this in-place, clean water evaluation should be the same
as is established by Working Group C.
The decision was made to develop test procedures which
would be valid for a wide variety of equipment types and
goals. The goals delineated included amount of oxygen
supplied each day, horsepower, distribution of dissolved
oxygen in space and with time and instantaneous oxygen
uptake rate. Mixed liquor solids distribution was felt to be
beyond the scope of the oxygen transfer directive.
The tests considered are listed below along with a brief
description of the test and possible problems.
1. Steady State Batch Endogenous Test
What System?
For continuous flow system for which the flow to the
aeration tank can be diverted for the duration of
the test.
31
Why Batch?
No wastewater flow during test.
Why Steady State?
Parameters evaluated after dissolved oxygen concen-
tration reached maximum
Description
Aerate until maximum dissolved oxygen concentration
reached after endogenous respiration established.
Problems
Defining average dissolved oxygen for high length to
width systems. May need curtain (nylon reinforced
Hypalon) for lagoon. Saturation dissolved oxygen
difficult to obtain on diffused air system.
2. Steady State Continuous Test
What System?
For continuous flow system with no flow divergence
capability.
Why Steady State?
Sample collected from aeration tank being operated as
close to steady state as possible.
Description
Remove representative sample. Run successive batch
oxygen utilization rates to estimate rate at time of
sample collection.
Problems
Getting good estimate of oxygen uptake rate at time of
collection. Load variations. Difficulties with high
oxygen uptake rate.
3. Non-Steady State Batch Test
What System?
Same as Steady State Batch Test except aeration
equipment turned off until dissolved oxygen concen-
tration approaches zero.
Problems
Getting organisms back in suspension. Organisms
accumulating storage products thus delaying
endogenous respiration.
4. Non-Steady State Continuous Flow Test
What System?
Same as Non-Steady State Batch Test except open
inlet and outlet valve after dissolved oxygen concen-
tration reaches zero.
Problems
Same as Non-Steady State Batch Test plus need to
keep track of substrate input and utilization.
252
-------
5. Mass Balance Test
What System/'
Any continuous flow system.
Description
Rate of utilization equal to COD in minus COD out
minus oxygen equivalent of cells wasted.
Problems
Average daily dissolved oxygen concentration. Close
control over solids wasting.
Each of the tests listed above are described in Attach-
ments A through E. Tests which involve tracer methods
or off-gas analysis may also be utilized. The use of
tracers was discussed by Larry A. Neal in "The Use of
Tracers for Evaluation of Oxygen Transfer" presented at
this Workshop. Off-gas analysis may be utilized for
diffused air, submerged turbine, and closed high-purity
oxygen system. Details of this method are discussed by
A.G. Boon in "Oxygen Transfer in the Activated-Sludge
Process" presented at this Workshop.
Attachment A: Steady State Batch
Endogenous Test
Object: To outline the procedure for steady state
endogenous respiration testing for evaluating the
performance of aeration equipment in activated
sludge plants.
Fundamental Concept: Evaluation of aeration equipment
in operating activated sludge plants requires careful
measurement of the oxygen uptake rate, the alpha factor,
the beta saturation value, temperature, and the dissolved
oxygen concentration. These data are used in the follow-
ing equation to determine the KLa value at the operating
temperature which can then be corrected to 20°C.
KLa = R/a(/3C*-C)
Because of the difficulty in measuring high oxygen
uptake rates, it has been suggested that the endogenous
respiration rate of the activated sludge mixed liquor (ML)
be used. The endogenous respiration rate generally
changes at a rate of only 0.02 hr1 or less so that the
oxygen uptake rate will remain relatively constant for a
long period of time. The aeration system will allow the
the DO to rise to a reasonable level so that the DO can
easily be measured.
Procedure:
1. Turn off the influent flow but allow the sludge recycle
to continue normally.
2. Adjust the water level in the aeration tank to the level
normally anticipated at maximum flow if the aerator
performance is desired at maximum conditions.
3. Check the power measurements to insure proper
operation of the aeration equipment.
4. Allow the system to aerate for 30 min to 1 hr to use
up all excess organic matter.
5. Determine the oxygen uptake rate in the ML, and
obtain DO and temperature measurements under
the surface aerator at the same time.
6. Collect a sample of ML and allow the solids to settle
while determining the oxygen uptake rate.
7. Pour off the settled supernatant and run the alpha and
beta saturation tests on the settled supernatant.
8. Determine the oxygen transfer and oxygen saturation
values in clean water at the same temperature as the
mixed liquor at approximately the same KLa value as
in the field system to minimize errors in the results.
9. Repeat the test as desired to insure the validity of the
results obtained.
Oxygen Uptake Rate
1. Collect a sample of ML at the edge of the aeration
zone so that the sample has a high DO concentration.
2. Determine the change in DO in the ML sample at
regular time increments, 0.5-1.0 min.
3. Plot the DO against time on rectangular graph paper
and draw the straight line to best fit through the data
or use a linear regression analysis for determination
of the line of best fit. (Note: If the data do not yield
a straight line down to a residual of 0.5-1.0 mg/l DO,
the ML may not have oxidized all of the organics and
the test should be repeated after an additional
30 min.)
4. Determine the rate of oxygen uptake from the slope of
the line of best fit on a mg/l-hr basis.
Alpha and Beta Saturation
1. A laboratory apparatus such as described by McKinney
and Stukenberg in "On-Site Evaluation: Steady State
vs Non-Steady State Testing" presented at this Work-
shop can be used for determining alpha and beta
saturation values.
2. The apparatus should be tested on clean water to
determine the relative KLa for desired test liquid
volume, air flow rates, and mixing speed.
3. Once the relative KLa has been determined for the
test condition to match the KLa expected in the field,
determine the rate of oxygen transfer and oxygen
saturation in clean water using the sodium sulfite-
cobalt chloride technique or nitrogen stripping for
deoxygenation.
4. Once the KLa and the C* are determined in clean
water, the test should be repeated in the ML settled
supernatant.
5. The alpha value can be determined as follows:
a = KLa in ML supernatant/KLa in clean water
253
-------
6. The beta saturation value in the supernatant can be
used directly so that beta does not have to be
calculated. If beta is desired, the following equation
should be used:
)3 = C* in ML supernatant/C* in clean water
Dissolved Oxygen
1. All DO measurements should be measured with a
direct reading DO probe standardized against Winkler
DO in clean water.
Attachment B: Steady State
Continuous Test
Object: To outline the procedure for steady state con-
tinuous testing for evaluating the performance of
aeration equipment in activated sludge plants.
Fundamental Concept: Evaluation of aeration equip-
ment in operating activated sludge plants requires
careful measurement of all parameters the same as in
the steady state batch endogenous test. Occasionally, it
will not be possible to stop the waste load and allow the
system to come to equilibrium in endogenous respiration.
This mode will require a repeat of the same procedures
outlined in Attachment A, except that the oxygen uptake
rate will be higher.
Procedure:
Oxygen Uptake Rate
1. Oxygen uptake rates must be determined as quickly as
possible on ML samples.
2. It will be necessary to add oxygen to the sample and
collect the data quickly.
3. Repeat the oxygen addition until the endogenous
phase is reached.
4. Use the maximum rate of oxygen uptake for the
sample.
5. Project the oxygen uptake rate to zero time when the
sample was collected.
Attachment C: Non-Steady State
Batch Test
A steady and stable condition is established in the
aeration basin under process conditions. Raw sewage
and sludge return flows are then discontinued and
aeration is ceased. After the dissolved oxygen content
drops to zero, aeration is reestablished. Values of C and t
are measured at several points within the basin until a
stable value of C is obtained. This value is recommended
to be at least 3 mg/l but below 5 mg/l for air systems,
but may be modified after more experience is gained with
the test.
The value of KLa is determined by direct analysis of the
data. The slope of the line through the values of dC/dt is
K|_a. Conversion of the result to standard conditions
requires knowledge of the in situ values of a and /3,
the residual C value, as well as temperature and
barometric pressure. Transfer effiency may be determined
from corresponding power measurements.
Attachment D: Non-Steady State
Continuous Flow Test (Kalinske)
This procedure is similar to the non-steady state batch
test except that after the DO content of the mixed liquor
reaches zero, raw wastewater flow into the basin is
reestablished, as is sludge return flow and aeration. This
method was reported by A.A. Kalinske in "Problems
Encountered in Steady State Field Testing of Aerators
and Aeration Systems" presented at this Workshop.
Attachment E: Mass Balance Test
Oxygen transfer is determined by a mass balance over a
respiring system. By measuring the biodegradable organic
matter entering the system, the biodegradable matter
leaving the system, sludge solids wasted from the system
by separate sludge wasting and in the effluent and
the dissolved oxygen leaving the system in the effluent,
an oxygen balance can be made. When all measure-
ments are expressed in oxygen units, the oxygen trans-
ferred to the wastewater is equal to the reduction in
biodegradable organics minus the active sludge (02
equivalent) produced plus the oxygen leaving in the plant
effluent. If the raw wastewater contained oxygen as it
entered the aeration basin, this quantity must be sub-
tracted from the results above.
Conversion of the results to standard conditions requires
determination of a and ft values, the residual C value,
as well as temperature and barometric temperature.
Transfer efficiency may be determined from correspond-
ing power measurements.
Group E. Participants
*J.R. Stukenberg
*R. Irvine
H.H. Benjes, Jr.
G.L Huibregtse
A.A. Kalinske
M.G. Mandt
J.J. McKeown
R.E. McKinney
J.L Wight
'Co-chairman
254
-------
Discussion of Working Group
Summary Reports
The discussion below was transcribed following the
presentation of the working group summary reports.
Unfortunately, the electronic transcriptions were of poor
quality and it was necessary to delete certain discus-
sions which were unintelligible. All discussions have
been edited.
Group A. Philosophy of the Standard
Matsch: I would like to bring up a point which concerns
me on the philosophy of the development of the
standard. How much involvement will the manufacturer
have in the selection of the geometry of the aeration
basin? If the engineer specifies the geometry in advance,
it gives the manufacturer very little opportunity to
optimize his equipment in that particular geometry.
Kransnoff: That's a very important point you bring up.
The group discussed that at some length during their
deliberation. I really don't know, at the present time,
whether we should actually put that into a standard. In
practice, it would seem appropriate to obtain information
on optimal geometry for various devices when a con-
sultant comes up with a design and asks the various
manufacturers to comment on whether there is a certain
geometry which would optimize their particular
equipment.
Fissette: I would like to further amplify on that point. I
think what Dr. Matsch says is exactly correct. The way
in which this summary statement (Group A) was framed
is, the consultant would provide the manufacturer with
an oxygen requirement in mg/l-hr and a tank volume.
Nothing else would be provided unless there were
constraints on land area, soil conditions, or what have
you. The consultant would also be responsible for
alpha, beta, and theta associated with the wastewater
characteristics. The manufacturer would then base his
proposal on meeting that oxygen transfer requirement
with the best geometry he has. He would optimize his
system around the waste characteristics he has been
given. Then, the consultant would look at the different
alternatives that came in and determine, by value
engineering, or least cost or whatever, the best general-
ized system which he felt would meet his requirements.
He would then put them out for bids to the manufac-
turers. The concept of this philosophy is to have the
manufacturer bid his best suit using the geometry
which best suits his equipment. Perhaps I am not saying
it the way the group felt, but I think that is probably
correct.
Hunter: I'm a little confused, but I agree with what
everybody is saying here. It seems like we were going
to develop a standard for testing, whereas the philoso-
phy seems to be developing a standard for purchasing.
32
Krasnoff: Actually, the philosophy group was more
concerned with the discussion of how such a standard
would be used, should it be developed. I think we still
have a long way to go before we even have a consensus
agreement on a standard method for testing in clean
water. But the philosophy is laid out here merely as an
indication of how such a standard would be put
into use.
Sa/zman: I guess I was a little confused by the use of
the word "interim". I think that in the discussions I have
heard here this week, we could come out with the best
state-of-the-art recommended test that reflects good
current practice but, like any standard, would have to
be updated constantly. I think the word "interim"
sounds a little too tentative.
Krasnoff: That's exactly what we had in mind. If every-
one is bothered by the word "interim", we can cross it
out. We did have in mind something which is a current
state-of-the-art recommended procedure.
Boyle: I think the word "interim" came about by virtue
of the process of review by which a standard would
have to go through the ASCE-ANSI procedure. Any
recommended standard method would have to go through
a period of review and, during that review period, would
be called an interim standard. After having been approv-
ed by consensus, it would then be called a standard.
Group B. Modelling and Data Interpretation
Shell: How does the non-linear least squares method
compare with the log deficit method in determing the
parameters?
Baillod: The log deficit method, applied with a measured
value of saturation concentration, generally compares
well with the non-linear least squares estimates.
However, differences are evident and it depends on the
actual set of data. For certain data sets, the agreement
is very good, for other data sets, not so good. In
general, since the data themselves consist of values of C
vs t, the group felt that it was preferable to work with the
actual C and t data values rather than the use of logarith-
mic transformations required in the log deficit method.
The logarithmic transformation, inherent in the log deficit
method, changes the error structure of the data and, for
statistical reasons, the most precise estimates of the
parameters are produced by handling the raw data
themseves as is done in a non-linear least squares fit.
The other techniques, such as the log deficit or direct
method, are useful approximate methods for data
analysis, but the best or most precise method, the
group felt, would be the non-linear least squares regres-
sion analysis.
255
-------
Shell: Which parameters were to be measured or
estimated from the non-linear least squares technique?
Baillod: All three parameters. A, B, and D. The reason
being that if saturation is important, and it is, it's better
to determine it from the estimation given by the model
fit rather than from one or two measured points near
the end of the test.
Mandt: I think your group did an excellent job. The
model is really the universal model that we're looking
for, recognizing that capital D is not the real KLa for all
types of aeration equipment. I see no reason why this
model shouldn't be adopted by this Subcommittee.
Paulson: I would like to have you comment on your
value of a minimum sampling time of 30 sec.
Baillod: Referring directly to our written report under
experimental design, the statement reads, "it is recom-
mended that the analysis be based on a minimum of 10
to 15 data points. Approximately 2/3 of these points
should be evenly distributed over the period of time
covered by 0.5/D to 2/D. The remaining 1/3 of the
points should be distributed evenly over the period of
time covered by 2/D to 6/D. In cases of rapid transfer,
the minimum time between data points should be
0.5 min.
What we really meant was, using the pipe sample
technique and simply saying that in case of rapid transfer
where you could not get the minimum of 10 points
within that area covered by 0.5/D to 2/D, we would be
satisfied with samples taken every 30 sec. If you had
probe data, there is no reason why you could not take
them at intervals as short as you wished within the
response time of the probe.
Saccia: Is C infinity going to be a measured value or
text book value?
Baillod: The recommendation is that that parameter will
be estimated from the curve fit. We would not be using
a "book" value. The recommendation of the committee
was not to use a "book" value but to fit the data from
the actual test using the non-linear least squares
analysis. It was also pointed out by the committee that,
if you were going to use the log deficit method for
determination of KLa, etc., the actual measured value in
the tank should be used and not the "book" value.
Boon: If you are going to use the probe, wouldn't you
calibrate the probe against the "book" value?
Baillod: Our committee really didn't get into the
question of probe calibration. We felt that was more of a
test procedure and would be taken up by a subsequent
subgroup. I suspect that one would simply calibrate the
probe by a Winkler titration procedure.
Unknown: Could you explain exactly why the committee
decided to run the test out to 6/D as opposed to simply
running the test out until the dissolved oxygen concen-
tration was relatively constant? Where did the value of
6/D come from?
Baillod: The value of 6/D was kicked around in the
Subcommittee for some time and was primarily based,
as far as I can see, upon some of the actual experiences
of Black and Veatch in the field.
Unknown: But there is no scientific evidence to
indicate that 4/D, 5/D, or 6/D would give better informa-
tion, as far as you know, or give significant improvement
in the data? It would seem to be a very arbitrary value.
Baillod: That's true. In my own mind, I can't really see
any difference in running at a value of 6/D as compared
with running until the dissolved oxygen concentration
became relatively constant.
Mueller: During this Workshop there seemed to be some
concern about the analysis of data using the non-linear
least squares method on the computer. Wouldn't it be
viable to incorporate, within the standard, a computer
program to analyze the data?
Baillod: I think that probably would be practical. At least a
reference could be given to a standard software library.
Shell: I am somewhat reluctant to bring up this point,
but I would like to make a couple of comments on the
data analysis procedures. I collect a lot of data and
have a lot of information which I have to assimilate and,
subsequently, use a computer program for this purpose.
I have learned that without looking at the data, examining
it, or plotting it graphically, mistakes were made that
could have been avoided had graphical analysis been
performed during the test procedure. I think a major
weakness of the proposed approach presented is that
there is no graphical depiction of the data.
Baillod: I agree with you. In fact, our recommendation
intends that it be done. When we said that the error
structure in the data should be examined, we meant that
a plot should be made and the data carefully examined.
We feel it is imperative that the data be visually
observed to interpret the estimated parameters.
Group C. Non-Steady State Clean Water Testing:
Shop and Field
Brenner: Two of the major problems which have come
up during our Workshop related to clean water testing
are: how are you going to collect your samples, and how
are: How are you going to collect your samples? and How
recommendations still allow a wide variation in sampling
procedure and DO measurements. Were there any
specific recommendations with regard to either sampling
or DO measurements that came out of your group
discussions?
256
-------
Paulson: The group was not opposed to pipe
samples vs direct measurement, nor was there any
specific objection to the Winkler procedure over the
probe procedure for measuring DO. The primary focus
was on problems associated with the procedures
themselves.
Brenner: Do you think your group will come down
to one recommended procedure or another for
measuring DO?
Paulson: I think, in the case of looking at the Winkler vs
the DO probe, the group would have to go in seclusion
for 2 or 3 days to figure out which way to go. I
feel that the DO probe should be used, and also that at
least one sample point should be run by the Winkler for
additional data base. I feel that continuous collection of
data, obtained with a probe, is the best way to go. In
many cases, the use of a Winkler as a backup verifica-
tion would be the best approach. Thus, the probe would
act as a primary data base source and Winkler would be
the secondary base source. As far as sampling is con-
cerned, use of the probe answers that question, also,
since you would be using outside probes within the tank.
It's possible to use piped samples with outside probes. It
limits some problems and creates others. In summary, I
believe probes are the most valuable technique to use.
Subsequently, Winkler analysis could be used as a
backup and as a secondary check on the probe system.
Boon: I would like to open the question about clean
water or the use of detergent in clean water. I was
particularly concerned with your comment under item 9
in which you state that the physical and chemical
properties of the test water should be provided. Have
you any specific thoughts on what you're going to have
analyzed? Were you going to look at the various surfac-
tants that might be contaminating the water, or
confining it to dissolved solids, pH, alkalinity, etc.?
Paulson: We started out to try and define those
parameters which should be measured. One of the first
suggestions made was that we simply use the drinking
water standards. But it was pointed out by some that
there were things not listed in the drinking water
standards that were surfactants. Someone else had
suggested that perhaps we do an alpha and beta
analysis on the water in an effort to identify the
character of the water. Actually, we did not arrive at any
specific listing of constituents that should be measured,
nor did we arrive at any kind of decision as to what a
standard water should be. This should be addressed in
more detail within the coming months.
Boon: I have a few more comments. One is related to a
comment you made under item 16. You state that there
should be a proper cobalt concentration for testing. I
would have thought that it was also necessary to specify
the actual cobalt concentration in the test tank at the
start of the test run. This would establish that there is
still an appropriate concentration of cobalt in the water.
I think one of the reasons some find the first test in so-
called clean water produces unsatisfactory results is
because the cobalt has complexed with impurities within
the water. In fact, there is no cobalt concentration
during the test run. I think item 16 should be coupled
with the specifications of the water itself in order to
confirm that the cobalt is in the test water.
Paulson: That's true. When difficulties were found in
monitoring oxygen transfer rates during test runs,
analyses revealed cobalt concentrations were not
present. There are also situtations where difficulties are
caused by factors not directly attributable to the
presence or absence of cobalt. Our purpose in writing
item 16 was to focus on two questions: one, that there
be adequate catalyst to catalize the reaction of sulfite
and, two, to think about an upper limit of cobalt which
may cause some interference problems. That was
essentially the extent the problem of cobalt had been
discussed in our group.
Shell: When making a water analysis, we always run a
complete spectrum of materials within the water,
including cobalt concentrations. In the first test of the
studies conducted, cobalt concentration has never been
a problem. We found the problem (of unsatisfactory test
results on first test) is due to an inadequate amount of
sulfite. For some reason, the use of 1.5 excess will not
completely deaerate the water in the first test. On the
second test, it will. On the first test, however, there does
not seem to be enough sulfite to deaerate the water
and, thereby, we have anomolies. I'm quite sure, how-
ever, that it is not a cobalt problem because we have
found sufficient cobalt present in the water during the
test procedures.
Boon: On item 18, regarding the addition of sulfite in a
slurry form, I'm in favor of sulfite being dissolved before
it is added to the water. If it is not in solution before the
test begins, the undissolved sulfite in the bottom can
create difficult problems.
Paulson: I would like to make one comment regarding
that. Initially, our group suggested the use of a
solution of sulfite under item 18. However, the question
was raised as to whether we really wanted to insist on a
perfect solution of sulfite. There were experiences
where dry feed was working fine, and others where
slurry feed was working fine. I think everyone agreed it
would be nice to have the sulfite in solution, but maybe
we can't afford that in every case.
Boon: Could I go on to suggest if a slurry is to be used
in feeding the sulfite to the test tank, a complementary
test should be conducted at the start of the reaeration
test procedure to insure that no free sulfite is present
in the test tank during the reaeration run.
257
-------
Paulson: That would be an idea. But how are you going
to sample these large tanks to get a representative
sample during the reareation experiment?
Shell: I will not run a test unless I am sure the sulfite is
completely in solution and that it is added immediately
above the turbine or the aerator. Regardless of whether
the water is warm or cold, it adds a lot of confusion to
the test procedure if the sulfite is not completely in
solution prior to the test run. I know it is very difficult
when you have four, five, or six aerators in the system,
but it can be done.
Paulson: I agree. That is essentially the way I normally
conduct my test runs. However, I am speaking for the
group.
Salzman: The points which this group has tried to
cover in the working group summary are more or less a
state-of-the-art review. We do not want to be overly
restrictive. This is why we didn't say DO probes, as
opposed to Winkler, and why we left a little flexibility in
the test conditions. We tried to develop a standard
people could live with, and then we exercised reason-
able control over critical areas. I think that's what the
group tried to do rather than set up a perfect test.
Group D. Alpha, Beta, and Temperature Corrections
Brociner: Would it be reasonable to run alpha tests in
only large-scale basins so that we can avoid some of
the problems associated with scale-up?
Shell: Yes, I suppose that could be done. But one of the
problems we have is whether or not there is a sufficient
amount of wastewater to test in a particular basin size.
Because of the question of scale, we debated among
the group members for some time and found a lot
of this business on alpha determinations is frustrating.
It is misleading to conduct alpha tests in a small vessel
with a surface aerator and say that it represents 150 hp,
50 hp, or even a 5 hp aerator. It simply does not.
Brociner: But we do have other small-scale tests we
undertake in the laboratory to simulate certain unit
processes, such as the centrifuge test or the column
settling tests. Perhaps one way to avoid problems
associated with scale in submerged aeration would be
to increase the depth of the test column.
Shell: It was agreed in simulating submerged aeration,
if our wall effect was minimized, we could get a reason-
able value of alpha in this particular situation if we used
20 ft, 10 ft, or any depth column.
Scaccia: It is possible to calculate the amount of surface
entrainment you get with the small-scale alpha testing
process. I'm not sure I agree that in these small-scale
units most of the transfer occurs through surface
entrainment. In fact, I don't think it does. However, I will
agree with you that in these small-scale alpha testing
facilities, we do not have geometric and dynamic
similarity with full-scale units. My question is the follow-
ing: If we maintain geometric and dynamic similarity
between the small-scale test unit and the full-scale unit,
is it reasonable to assume that the alphas will directly
scale?
Shelf: Well, it depends on the aeration device. I can
assure you, after about 5 years of experience, you
cannot scale a surface aerator.
Scaccia: I agree that you can't scale a surface aeration
unit. You can only scale elements of the unit. You can't
take a small-scale unit and say that it would be
representative of a large-scale unit. And, if that's what
you mean, I would agree with you 100%. But I'm talking
about scaling up both the geometric and dynamic
properties of these systems.
She/1:1 don't think you can make all the laws of scaling
comply. We've tried for a number of years without
successfully being able to geometrically and dynamical-
ly scale these types of units.
Scaccia: I don't agree with that, but let's let that point
go. If alpha is nothing but a ratio between the mass
transfer in whatever scale you are running in water vs
that in mixed liquor, the same thing should happen in a
large-scale test. Again, you have a ratio between two
numbers. I don't understand why you feel so strongly
that alpha cannot be scaled if, in fact, we provide
geometric similarity and dynamic similarity in our
small- and large-scale systems.
Shell: I think the problem we have is that we're looking
at a small-scale unit that is only 12 in. deep. Once you
get into pilot-scale, there is probably very little differ-
ence between pilot- and full-scale-up of the alpha value.
But, notice that most of the alpha-scale data are
generated from the Black and Veatch type test device.
If one says that represents what is going to happen in
the field, this group does not feel that's true. We
are saying, let's continue the development to a larger-
scale unit alpha testing where we can produce both
dynamic and geometric similarities in the pilot- and full-
scale unit.
Lestochi: If we start with the premise that an alpha in
small-scale is different than an alpha in full-scale, is the
alpha in the small-scale system smaller or larger than in
the large-scale, or do we know?
Shell: We do not know.
Lestochi: All right. Then, my comment is that I think we
are wasting our time on trying to develop a standard to
test aerators. If we're going to test aerators to find out
what their oxygen transfer capabilities are when
properly applied in the field, we need to know the value
of alpha in order to apply them in the field. But, if we
don't know alpha values within ±50 or 60%, then we're
258
-------
going to be off that much in trying to apply this aeration
equipment in the field. If we can't put a handle on the
very fundamental problem, like defining alpha, why are
we trying to be exact in identifying the testing proce-
dures for oxygen transfer devices?
Shell: We've seen from Gary Gilbert's presentation the
error effect of the estimation of both alpha and beta,
and we recognize that it is enormous. What the con-
sulting engineer needs is some handle on the application
of the aeration devices to the field and we have to start
somewhere. We're not using a micrometer, but we're
saying that we want to develop a test procedure that
indicates what the surfactant effect is going to be on
that aeration equipment. Now, if the only sample I have
is raw wastewater, then I'll have to run alpha on raw
wastewater. If all we have is bench-scale facilities, I'll
have to run it on a bench scale. It may be wrong, but it
will be something of a guide. What we want to empha-
size is not to stop there. Try to get a correlation between
the bench- and full-scale.
Lestochi: But at some point I need to know what that
number is before I can scale to the field.
Shell: That's true, but in the opinion of the group
that has met here, we don't have that capability.
We don't have a procedure that is going to tell you
precisely the value that will occur in the field due to the
scaling effects, the geometry effects, the surface effects,
and etc.
Scaccia: First of all, we have to distinguish between two
things, that which is available in the literature and what
is actually known from independent tests. I personally
feel that there isn't much difference between the alpha
run at small-scale and large-scale. It's not a fact that we
don't know anything about the test or that we are
wasting our time with this measurement.
Ernest: It seems this whole alpha factor is actually
based on the wastewater and it varies from hour-to-hour
and day-to-day. Even when we test a sample, we are
going to be subjected to a great deal of variation
depending upon the day or the hour which we took the
sample for analysis. This is something we will have to
accept. We need an oxygen transfer device which is
highly flexible and will be able to transfer oxygen under
a wide variety of situations from hour-to-hour and day-
to-day.
Boon: I thought the philosophy group (Group A) very
neatly got around the problem associated with alpha in
the development of standards by throwing the ball back
to the people who knew or should know the most about
the characteristics of the wastewater, the consultant and
the user. The philosophy group had suggested that
aeration equipment manufacturers would be given a
required transfer rate and the process volume. Thus, it
doesn't matter whether the water is clean or dirty
provided that the standard is universally adopted. I think
it is extremely difficult to have a standard for a clean
water, because clean water varies just as much as
sewage varies from region to region and supplier to
supplier. If you adopt a standard for clean water, you
are going to be in a situation where you have people
contesting that the tests carried out in the shop with so-
called clean water will give one result when it is tested
in the field, and another when tested in the shop. And,
you will be forever trying to get the full-scale test facility
clean water of the same quality as that in the shop test
facility. As to the question of alpha, this is something
for the consultant to decide. What I have heard today is
that it is a procedure which the consultant should be
responsible for. The consultant must be sure that his
specifications for the aeration equipment allows for the
variability of alpha in the waste. And as Ernest has quite
rightly said, a range of alphas must be specified to the
manufacturers in order to provide sufficient flexibility to
allow for the variability that will occur in the plant. I
don't know how you build your plants in the U.S. but in
the U.K., when a plant is finished, it is designed for 5 to
10 yr hence. That is your flexibilty to a certain extent.
The flexibility is also there in the aeration devices which
are specified by the engineers. Sufficient flexibility must
be built in the plant to provide for the day-to-day and
hour-to-hour variations which are normally expected in
any type of plant.
Boyle: I would now like to focus the questions on this
group's (Group D) comments on beta and theta
determinations.
Brenner: Why didn't you prescribe the use of the Winkler
procedure for DO concentration for beta measurements?
Shell: Because there are so many interferences in
wastewater that make it very difficult to use the Winkler
procedure for accurate measurement of DO. The
Winkler is not a viable method for most wastewaters,
particularly, industrial wastewaters.
Unknown: Why were no specific recommendations
made for a value for theta by the group.
Shell: We really do not have sufficient information on
the effect of mixing intensity, hydraulic regime, and etc.,
on the temperature coefficient. As yet, it has not been
well defined and there has not been a large amount of
work done on the temperature coefficient for gas
transfer. It was not the intent of the group to avoid
recommending a table of temperature coefficients,
for example, for submerged turbines, surface aerators,
or diffused aerators. This can be done. We must have
some kind of factor to use in field application. The
group did not feel that a testing procedure for theta
was appropriate at this time. This procedure may
be developed over the years, but there's still a lot of
research that needs to be done.
Mueller: I think it should be made clear that it is the
charge of this Subcommittee to put together the best
259
-------
information that's available in the field. Five years hence,
we will evaluate that information and improve on it.
What Shell is saying is we do not have very much
information on theta, but we do have some. We should
begin to collect information on the influence of tempera-
ture on oxygen transfer for various generic devices.
Hopefully, the manufacturers will be collecting that kind
of information and they will be making it available to the
profession. It could be a significant contribution as far as
this particular problem is concerned.
Group E. Steady State Field Evaluations
Scaccia: In the procedure which you outline for the
non-steady state test as discussed yesterday by
Kalinske, there is concern related to what would happen
in the system when the aeration equipment is shut down
for deaeration. It would seem that there would be
tremendous effects on the mixing and turbulence
regimes within the basin when the aeration system is
started back up and oxygen transfer is subsequently
measured. There will certainly be mass transfer
problems. For example, the gas transfer through the floe
itself could be affected by such a non-steady state
hydraulic regime. Are you convinced that this type of
technique could be applied for field evaluation?
Stukenberg: I'm not so sure I'd want to say it can or
cannot be applied. In our discussions, we felt that the
use of submerged turbines or a jet aeration device
might produce enough pumping during the deaeration
period. These might be better suited to this test than
a surface aeration or diffused air system. I agree with
your concern, and we discussed this at some length.
We are changing a lot of things within the basin, the
dimensions of the floe, the mixing characteristics, etc.
There are certainly a lot of things that could be going on
in the basin that would make one wonder about the test.
We looked at all the tests that are currently available for
field testing, and, based on the comments which you
have made and our own feelings, we did not feel that
this would be an acceptable evaluation test procedure.
Mueller: When you are running this continuous steady
state test with the high uptake rate, your procedure in
reoxygenation of the sample seems futile since you've
already lost the initial organics. As far as I can see,
you're beginning to work in the endogenous phase,
which is not related to your initial uptake.
Stukenberg:\ou're right. What we are saying is, if the
oxygen uptake rate is high enough and the DO is
depleted before you get into endogenous, we will
continue to add data to the curve. We are actually trying
to get back to what the uptake was.
Mueller: Dr. Matsch had mentioned in our working
group discussions the possibility of getting the uptake
rate very rapidly with some new equipment that's
arriving on the market.
Paulson: Prompted by your comment (Group E
Summary), whereby you state that the equipment
should be tested in clean water and definitely specify
without detergent added, raises the rather controversial
point regarding the characteristics of a "clean water". I
think before this Workshop breaks up today, it would be
wise to briefly discuss some of the pros and cons
associated with the addition of detergent to water as a
standard fluid or standard water.
In the deliberations of Working Group C on a non-steady
state clean water test, the point was raised whether or
not a detergent should be added as a part of the clean
water test. Several questions were raised: How do we
conduct the test? (That is, what are the problems with
measuring the surfactant concentrations?) What type of
surfactants should be added? And what are the problems
of conducting the test itself regardless of any other
chemistry involved in the water? Another question was:
How are we improving the situation of the standard test
by adding a surfactant to the water?
The purpose is, of course, to give us a better answer as
far as the performance of the equipment in the field.
But, how does the addition of the detergent get away
from the same problems that we had to start with, the
variable water chemistry? It would seem that we would
be adding a further complication to the test procedure.
I would like to have the comments of the individuals in
this Workshop.
Boon: I'd like to see tests run both in the "clean water"
and in clean water with detergent added. That is the
type of system which we've adopted in the U.K. for
testing of aerators. I am sure that by adding detergents,
one will add complications, particularly with the Winkler
determinations. If one carries out tests with detergent
addition, one will have to do them with DO probes since
it is far more difficult to get an end point with a titration
when detergents are added than with clean water.
The big problem though is, "What is clean water?" Con-
taminants in "clean water" will have a substantial effect
on the oxygen transfer coefficient, particularly with fine
bubble diffused air systems. Surface aeration systems
are less affected by the presence of contaminants in the
water within ±10%. With the fine pore diffused air
systems, the effects of the contaminants could be ±20
or 30%. The addition of detergents overwhelms all these
small effects caused by contaminants. Thus, the value
obtained should not be affected by the initial contami-
nants that may be present. There isn't an accumulative
effect due to the addition of more surfactant. It is the
total effect of the 5 mg/l of the detergent which
swamps the effect of the fractional concentrations of
other surface active agents that might be present in the
water, or the trace contaminant derived from the unclean
aeration basin.
260
-------
The big problem is making the comparison between the
shop and field testing. The shop test basin, which is
normally small enough, can be cleaned, and it is
possible that the water used in these tests is quite clean.
In my experience, that is possible. But, in field testing,
the situation is much more complicated. What I'm
concerned about from your point of view is putting this
down in a standard. The aim is to get the manufacturer
to test his equipment in a given tank geometry and in
the reference fluid, which you decide, thereby specify-
ing the amount of power to transfer oxygen at a given
rate into the test water. When the equipment manu-
facturer wins the contract, the equipment is then tested
under field conditions to evaluate whether the equip-
ment actually works. It is probable that under field
conditions, different results will be obtained from the
shop conditions. In the past, this could be attributed to
the geometry of the tank. But, hopefully, based on the
acceptance of the recommendations made by Working
Group A, that can be avoided. Then the question of the
nature of the test water comes into the picture. If the
results are lower in the field test tank than in the shop
test tank, the excuse may be that the water isn't clean. I
can imagine arguments coming up by the manufacturers
claiming that the water tested in the field was not the
same as that tested under shop conditions.
Shell: In reference to your comments, Mr. Boon, with
special emphasis on the term reference liquid in the
U.S., it is the responsibility of the contractor to run the
tests. He transfers that responsibility back to the equip-
ment supplier. The contractor and the equipment supplier
must then clean the tank and prepare for the test and
make sure the reference water being used is what they
choose. They certainly can't come back later and say it
isn't the right water, and I can't think of a point in my
career when this particular point was made. It is their
responsibility. If they want to use river water, fine. If they
want to use effluent, fine. But, they're going to have to
live with those results. We have always used tap water,
as clean as available, for on-site testing. It might be river
water in order to get enough to do the testing. Again, I
want to emphasize that I don't really see any problem in
this area, as far as a reference fluid, because it is the
contractor's and equipment supplier's responsibility. It's
written in the specifications that clean water shall be
used. If they use dirty water and don't pass the test, it's
their own problem.
Brociner: We already have alpha, beta, and theta, and
now we are introducing still another complication in our
test procedure with the addition of surfactant. I feel that
we are over-complicating our whole standard
procedure by this addition. I don't see any value in
adding surfactants to clean water tests.
Brenner: Although there's been a good deal of discus-
sion about the use of a standard reference water in
developing a standard, I think the real issue here is
primarily one of not being able to adequately predict
the full-scale capacity of aeration equipment. It's very
hard to go from clean water shop tests to dirty water
field applications. We might have wastewater alphas
ranging from 0.4 to 0.8, with a wide variation from hour-
to-hour and day-to-day. Admittedly, it's quite easy to use
clean water test data and then scale-up with some alpha
value. But, if the addition of the detergent will give us a
handle on what might be expected in the actual field
installation, we might be better off in the long run with
standard tests which use the detergent. Mr. Boon has
indicated that, with a 5 mg/l detergent addition, the
value of alpha obtained is quite similar to the actual field
alpha that one would observe with municipal wastes
containing only small amounts of industrial wastes.
Therefore, we would not be getting equipment that's
undersized. If this is true, and I am not able to say
whether it is or not, since the use of the 5 mg/l deter-
gent simulates wastewater from municipalities, why
don't we think about the addition of detergent to clean
water as our reference fluid? It would put us ahead in
the Step 1 design phase since it would give us equip-
ment that would be sized more realistically for the
wastewater. It would be an advantage to be able to
better design equipment earlier in the design phase. I
think it's a good idea to look into the concept of adding
detergents to standard test water, and I hope that the
Subcommittee will at least consider it in their
deliberations.
Boon: I am sorry I didn't elaborate on what the addition
of 5 mg/l of detergent to clean water would do in
relation to the actual transfer of oxygen in wastewater.
Brenner made the point that the addition of the
detergent makes the clean water behave more nearly
like municipal wastewater. However, I want to make it
clear that you would still have an alpha. In the work
done, we've actually defined alpha prime (alpha') as the
relation between the KLa in the wastewater and the KLa
in the clean water with detergent. We found that the
variability of alpha' is not anything like the variability in
alpha (which, of course, is defined as the KLa in the
wastewater to the KLa in clean water). In a given system,
alpha varied from a value of about 0.3 to 1.0, whereas,
alpha' varied from about 0.8 to 1.2. That gives you a
feeling as to why we in the U.K. have decided to
standardize on a 5 mg/l surfactant addition to the clean
water test system. Now, if you are using a surface
aeration system, the variation in alpha from clean water
to mixed liquor is not as great as compared to the fine
bubble diffusion system mentioned earlier. Normally,
the alpha value for the fine bubble diffusers systems are
substantially lower than for surface aeration equipment.
(I want to stress that the value of alpha does not relate
to the efficiency of the gas transfer device.) Therefore, it
is possible to get a grossly underdesigned system if
clean water test data are used for fine bubble aeration
systems and extrapolated to field conditions with an
alpha which is considered to be typical for aeration
devices. (Normally, alpha values for fine bubble diffusers
may range around 0.3, 0.4, or 0.5, whereas, values for
261
-------
surface air entrainment devices may range from as low
as 0.85 to higher than 1.2.) Of course, if you have
measured alpha under the actual test conditions you
expect to have in the field, there should be no problems
with fine bubble systems, but, if you are estimating it,
you could make an error in order of magnitude of 1.0,
2.0, or 3.0 on the low side.
Fullerton: The emphasis of the Workshop was based
on measurements of oxygen transfer for aeration
systems. There was very little discussion relative to
oxygen transfer for pure oxygen systems. This emphasis
was correct. However, it is a concern that if standards
are promulgated without explicit definition of oxygen
systems as an exception, the same principles may be
applied to all systems. All this time, FMC Corporation
cannot identify the possible impact of adhering to
standards imposed for aeration devices as applicable to
the open tank pure oxygen process. But, FMC Corpora-
tion is secure in the present methods for determination
of oxygen transfer efficiency using the off-gas method
during a steady state operational mode. This procedure
is certainly indicative of the actual system operational
efficiency and has, to date, been accepted by the
engineers and owners as an acceptable method for
performance evaluation. The primary exception to the
aeration test is the clean water performance. Currently,
we do not have enough data to identify existing
problems with the test as a required method. Until such
time as FMC Corporation can evaluate the impact of
adhering to a suggested manual of procedure
(standard), we would prefer to remain as an exception
with a definition of in situ off-gas testing as an accept-
able alternative.
Mueller: I don't know enough about the FMC procedure
to ascertain its accuracy. In my paper on closed
systems, I indicated that off-gas analysis was the pre-
ferred methodology since inaccuracies in 02 uptake
measurement were bypassed. However, the technique
does require an accurate determination of off-gas flow
rate for each stage of the covered reactor. The same
requirement applies for an open tank reactor. I don't
know FMC's technique to get gas flow, but accurate
determination in an open system may be difficult.
In the closed tank aeration system, the same aeration
principles apply whether air or pure oxygen is utilized
as long as the gas phase mass balances are used to
provide 02 gas phase partial pressures. In the open
tank system, the change in Oz parital pressure and gas
flow with depth will probably be required. If done
accurately, I would think the same principles apply for
air and oxygen systems. However, if not done, then
pilot- or full-scale data in s/tu would be required. This
situation also exists with the closed tank system if a
model calculating gas phase partial pressure is not
utilized.
262
-------
Appendices
Appendix A
Announced Workshop Program
Tuesday, April 11,1978
1:00 p.m. Registration
2:00 p.m. Opening Remarks
William C. Boyle, Prof.
University of Wisconsin
Madison Wl
A Perspective on Oxygen Transfer
Session Chairman:
W. C. Boyle
2:05 p.m. Oxygen Transfer: A Historical Perspective —
The Need for a Standard
Wesley W. Eckenfelder, Prof.
Vanderbilt University
Nashville TN
2:35 p.m. Philosophy of and Perspectives on a
Standard — A Panel Discussion
Manufacturer:
George Fissette
Ralph B. Carter Co.
Hackensack NJ
Consulting Engineer:
Edwin Barnhart
Hydroscience, Inc.
Westwood NJ
Owner:
Lawrence Ernest
Milwaukee Sewerage Commission
Milwaukee Wl
ASCE:
Robert Crist
Publications and Technical Affairs
New York NY
U.S. EPA:
Richard Brenner
Municipal Environmental Research Laboratory
Cincinnati OH
3:30 p.m. Break
State-of-the-Art Review
Session Chairman:
W.C. Boyle
3:45 p.m. Review of Model Refinements and
Data Interpellation
C.R. Baillod, Prof.
Michigan Technological University
Houghton Ml
4:05 p.m. Oxygen Transfer Parameter Estimation
L.C. Brown, Prof.
Tufts University
Boston MA
4:25 p.m. Discussion
5:30 p.m. Adjourn
Social Hour and Dinner
7:30 p.m. Review of Test Procedures
Wayne Paulson, Prof.
University of Iowa
Iowa City 10
Clean Water Testing: Shop and Field
Session Chairman:
Charles Matsch
Union Carbide Corp., Linde Div.
Tonawanda NY
8:10 p.m. Influence of Tank Geometry
Thomas Rooney
Rexnord Corp.
Milwaukee Wl
8:50 p.m. Influence of Mixing
Ronald Salzman
Mixing Equipment Co.
Rochester NY
9:30 p.m. Sampling Considerations
Gerry L Shell
Gerry Shell Environmental Engineers
Brentwood TN
263
-------
10:10 p.m.Oissolved Oxygen Measurement and
Saturation Values
V.T. Stack, Jr.
Betz, Converse & Murdoch, Inc.
Plymouth Meeting PA
10:50 p.m. Adjourn
Wednesday, April 12. 1978
7:30 a.m. Breakfast
Clean Water Testing: Shop and Field (Cont'd.)
Session Chairman:
Robert Irvine, Prof.
Notre Dame University
South Bend IN
8:15 a.m. Influence of Temperature
John S. Hunter, III
3M Company
St. Paul MN
8:55 a.m. Influence of Water Chemistry
Donald Burns
Eimco PMD, Envirotech Corp.
Salt Lake City UT
9:35 a.m. Air Flow and Power Measurement
Fred Yunt
County Sanitation Districts of Los Angeles County
Carson CA
10:15 a.m.Break
10:30 a.m.Data Interpretation:
Non-Steady State Clean Water Tests
Lloyd Ewing
Ewing Engineering Co.
Milwaukee Wl
Jerry Wren
Water Pollution Control Corp.
Milwaukee Wl
Mikkel Mandt
Pentech Div.
Houdaille Industries, Inc.
Cedar Falls IO
11.20 a.m.Measurement of Alpha and Beta
Gary Gilbert
Kenics Corp.
North Andover MA
12:00 Lunch
1:00 to 6:00 p.m. — Free
6:00 p.m. Dinner
Clean Water Testing: Shop and Field (Cont'd.)
Session Chairman:
Robert Irvine
7:30 p.m. Comparison of Shop and Field Tests
V. Wahbeh and John R. Stukenberg
Black and Veatch
Kansas City MO
Evaluation of Respiring Systems
Session Chairman:
Haskel Brociner
FMC Corp.
Itasca IL
8:10 p.m. Oxygen Transfer in Closed Systems
J.A. Mueller, Prof.
Manhattan College
Bronx NY
8:50 p.m. On-Site Evaluation: Steady State vs
Non-Steady State Testing
John Stukenberg
Black and Veatch
Kansas City MO
Ross E. McKinney, Prof.
University of Kansas
Lawrence KA
9:30 a.m. Problems Encountered in Steady State
Field Testing
A. A. Kalinske
Camp, Dresser and McKee
Walnut Creek CA
Tracer Methods
Session Chairman:
Haskel Brociner
10:10 a.m.Use of Tracers for Evaluation of
Oxygen Transfer
Lawrence A. Neal
Law Engineering Testing Co.
Marietta GA
10:50 p.m. Adjourn
Thursday. April 13, 1978
7:30 a.m. Breakfast
8:30 a.m. to 12:00 Meeting of Working Groups A-E
12:00 Lunch
1:30 to 5.00 p.m. Meeting of Working Groups A-E
5:30 p.m. Social Hour and Dinner
Evening is Free
264
-------
Friday, April 14, 1978
7:30 a.m. Breakfast 9:15 a.m. Group C — Clean Water Testing:
Shop and Field
Presentation of Working Group Summaries
9:45 a.m. Group D — Alpha. Beta, and Temperature
Session Chairman: Corrections
W. C. Boyle
10:15 a.m.Group E — Steady State Field Evaluations
8:15 a.m. Group A — Philosophy of the Standard
11:00 a.m.Adjourn
8:45 a.m. Group B — Modelling and Data
Interpretation
265
-------
Appendix B
Members of the Workshop
1. Baillod, C. Robert
Michigan Technological University
Department of Civil Engineering
Houghton Ml 49931
2. Barn hart, Edwin L.
Hydroscience, Inc.
411 Old Hook Road
Emerson NJ 07630
3. Barren, Graham
Eco-Research Ltd.
45 Sheppard Avenue East
Willowdale, Ontario, M2N 2Z9 Canada
4. Bell, James A.
Ecodyne Corporation
Smith-Loveless Division
14040 Sante Fe Trail Drive
Lexena KS 66215
5. Benjes, Henry H., Jr.
Culp/Wesner/Culp
P.O. Box 40
El Dorado Hills CA 95630
6. Bennett, Douglas L.
Air Products & Chemicals, Inc.
P.O. Box 538
Allentown PA 18105
7. Berk, William L
Lakeside Equipment Corporation
1022 East Devon Avenue
Bartlett IL 60103
8. Bewtra, J. K.
University of Windsor
Department of Civil Engineering
Windsor, Ontario, N9B 3P4 Canada
9. Boon, Arthur G.
Water Resources Centre
Stevenage Laboratory
Elder Way, Stevenage
Hertfordshire SG1 1TH
England, United Kingdom
10. Boyle, William C.
University of Wisconsin — Madison
Department of Civil and Environmental Engineering
3206 Engineering Building
1415 Johnson Drive
Madison Wl 53706
11. Brenner, Richard C.
U.S. Environmental Protection Agency
Municipal Environmental Research Laboratory
26 West St. Clair
Cincinnati OH 45268
12. Brociner, Haskel
FMC Corporation
Environmental Equipment Division
1800 FMC Drive West
Itasca IL60143
13. Brown, Linfield C.
Tufts University
Department of Civil Engineering
Anderson Hall
Medford MA02155
14. Chen, S. J.
Kenics Corporation
Kenics Park
North Andover MA 01845
15. Crist, Robert A.
American Society of Civil Engineers
345 East 47th Street
New York NY 10017
16. Eckenfelder, Wesley W., Jr.
Vanderbilt University
Water Resources Engineering Department
P.O. Box 6222
Station B
Nashville TN 37235
17. Ernest, Lawrence A.
Milwaukee County Metropolitan Sewage District
P.O. Box 2079
Milwaukee Wl 53217
18. Fissette, George R.
R. B. Carter Company
Aerobic Treatment Systems
192 Atlantic Street
Hackensack NJ 07602
19. Fullerton, Donald G.
FMC Corporation
Environmental Equipment Division
3999 South Mariposa Street
Englewood CO 80110
20. Gilbert, Gary
Kenics Corporation
Kenics Park
North Andover MA 01845
266
-------
21. Helbing, Thomas J.
Reid, Quebe, Allison, Wilcox & Associates
3901 Industrial Boulevard
Indianapolis IN 40254
22. Huibregtse, Gregory L.
Rexnord Corporation
Corporate Research and Development Group
5101 West Beloit Road
P.O. Box 2022
Milwaukee Wl 53201
23. Hunter, John S., Ill
3M Company
900 Bush Avenue
P.O. Box 33331
St. Paul MN55133
24. Irvine, Robert
Notre Dame University
Department of Civil Engineering
Notre Dame IN 46566
25. Jenkins, David
Joint Editorial Board of Standard Methods
11 Yale Circle
Kensington CA 94708
26. Kalinske, A. A.
Camp, Dresser & McKee
710 South Broadway
Walnut Creek CA 94596
27. Kapur, Ramesh
Donahue & Associates, Inc.
P.O. Box 489
Sheboygan Wl 53081
28. Krasnoff, Paul J.
New York City
Department of Environmental Protection
Process Control Section
Wards Island NY 10035
29. Lakin, Michael B.
Mixing Equipment Company, Inc.
135 Mt. Read Boulevard
P.O. Box 1370
Rochester NY 14603
30. Lee, John
University of Wisconsin — Madison
Department of Civil and Environmental Engineering
3206 Engineering Building
1415 Johnson Drive
Madison Wl 53706
31. Lestochi, Louis U.
Air Products & Chemicals, Inc.
P.O. Box 538
Allentown PA 18105
32. Mandt, Mikkel G.
Houdaille Industries, Inc.
Pentech Division
219 East 4th Street
Cedar Falls 1050613
33. Marx, James J.
Chicago Bridge & Iron Company
Walker Process Division
840 North Russell Avenue
Aurora IL 60506
34. Matsch, Charles
Union Carbide Corporation
Linde Division
P.O. Box 44
Tonawanda NY 14150
35. McCarthy, Daniel A.
EIMCO - PMD - Envirotech
414 West Third South
Salt Lake City UT 84110
36. McKeown, James J.
NCASI
Tufts University
Anderson Hall
Medford MA02155
37. McKinney, Ross E.
University of Kansas
Department of Civil Engineering
Learned Hall
Lawrence KS 66045
38. Milne, Richard P.
Greeley & Hansen
6 Penn Center Plaza
Philadelphia PA 19103
39. Morgan, Robert P.
American Society of Civil Engineers
345 East 47th Street
New York NY 10017
40. Mueller, James A.
Manhattan College
Bronx NY 10017
41. Nardozzi, Anthony D.
Infilco Degremont, Inc.
P.O. Box K-7
Richmond VA 23288
42. Neal, Larry A.
Law Engineering Testing Company
2749 Delk Road
Marietta GA 30067
267
-------
43. Nelson, Steve
EIMCO - PMD - Envirotech
414 West Third South
Salt Lake City UT 84110
44. Parker, Denny S.
Brown & Caldwell
1501 North Broadway
Walnut Creek CA 94596
45. Paulson, Wayne L
University of Iowa
Environmental Engineering Water Plant
Iowa City IO 52240
46. Piper, Louis H.
Infilco Degremont, Inc.
P.O. Box K-7
Richmond VA 23288
47. Redmon, David T.
Ewing Engineering Company
6200 North 39th Street
Milwaukee Wl 53209
48. Rooney, Thomas C.
Rexnord Corporation
Environmental Process Department
5101 West Beloit Road
P.O. Box 2022
Milwaukee Wl 53214
49. Salzman, Ronald N.
Mixing Equipment Company
135 Mt. Read Boulevard
P.O. Box 1370
Rochester NY 14603
50. Scaccia, Carl
Union Carbide Corporation
Linde Division
P.O. Box 44
Tonawanda NY 14150
51. Shell, Gerry L.
Gerry Shell Environmental Engineers, Inc.
Peach Court Office Building
Brentwood TN 37027
52. Simmons, Edwin D.
Passavant Corporation
Water and Wastewater Division
P.O. Box 2503
Birmingham AL 35201
53. Smart, John
Ministry of Environment
135 St. Clair Avenue, North
Toronto, Ontario, M4V 1P5 Canada
54. Stack, Vernon T., Jr.
Betz, Converse & Murdoch, Inc.
1 Plymouth Meeting Mall
Plymouth Meeting PA 19462
55. Stenstrom, Michael K.
University of California — Los Angeles
405 Hillgard Avenue
7619 Boeder Hall
Los Angeles CA 90024
56. Stukenberg, John R.
Black & Veatch
P.O. Box 8405
Kansas City MO 64114
57. Wahbeh, Valery N.
Black & Veatch
12075 East 45th Avenue
Suite 333
Denver CO 12075
58. Weis, Ronald J.
Peabody Welles
11765 North Main
Roscoe IL 61073
59. Wight, Jeffrey L.
Aqua Aerobic Systems, Inc.
6306 North Alpine Road
Rockford IL 61111
60. Wren, Jerome D.
Sanitaire — Water Pollution Control Corporation
2320 West Camden Road
P.O. Box 744
Milwaukee Wl 53201
61. Yunt, FredW.
County Sanitation Districts of Los Angeles County
24501 South Figueroa Street
Carson CA 90745
268
-------
Appendix C
Procedures Governing
American Society of
Civil Engineers
Development of Standards
1. General. The procedures described herein refer to
the adoption of new standards, revision of standards,
reaffirmation of standards, and withdrawal of
standards, all of which are referred to as standardi-
zation activities.
The ASCE Board of Direction delegates the authority
for the development of standards and associated
activities to the Technical Council on Codes and
Standards (TCCS) and to its Executive Committee
(TCCS/EC).
2. The TCCS/EC shall furnish guidance to the Board of
Direction on policy and other matters relating to
codes and standards. TCCS/EC shall also establish
liaison with appropriate Technical Divisions and other
Technical Councils of the Society to insure technical
input from such groups when the work of such
Technical Divisions and Councils relates to a specific
standard activity under the jurisdiction of TCCS.
3. TCCS/EC will establish liaison with appropriate
government agencies, industry, and other organiza-
tions for guidance in standards activities of mutual
concern. It will have the delegated power to authorize
and supervise all standards development programs
undertaken for, or in cooperation with, other organi-
zations. This will include participation in standards
writing activities of such other organizations.
4. When the need for a standardization activity arises,
TCCS/EC shall consider a proposed standardization
action and, if appropriate, establish a preliminary set
of objectives, organize a standards writing committee,
and request approval of Management Group A (MGA)
for the establishment of either a task committee or a
committee for the specific project. It shall notify the
appropriate Technical Divisions and Councils of such
action and publish a notice in the ASCE News.
5. When continued or permanent maintenance of a
standard is required, the TCCS/EC will recommend
to MGA that such group has the status of a continu-
ing standard writing committee. If no such
maintenance is required, the group shall have the
designation of a task committee.
6. All actions of TCCS standards writing committees
shall be monitored by TCCS/EC, including make up
of the membership, expenditures of funds for travel
reimbursement for control group members, engage-
ment of consultants, etc., as provided for in an
approved budget and under established procedures.
TCCS standards writing committees may, if approved
by TCCS/EC and with concurrence from the Execu-
tive Director, seek outside funds for a standards
activity.
7. TCCS standards writing committee membership shall
be made available to thoseinterests substantially .
concerned with the scope of a proposed standard and
shall be balanced among those interests and selected
from ASCE members with competence in the field of
of the proposed standard except as noted in item 8
below. The executive committees of all concerned
Technical Divisions and Councils may have a repre-
sentative on each standard writing committee at
their option.
8. Each TCCS standards writing committee shall
organize itself and subcommittees or subgroups with
appropriate members and officers (Chairman, Vice-
Chairman, etc.), budget, scope, schedules, and pro-
cedures for approval. TCCS/EC officers and control
groups will be monitored by and appointed yearly by
the TCCS/EC. Individuals may be reappointed and
hold office for a maximum period of 4 years.
9. Non-members of the American Society of Civil
Engineers may serve on a TCCS standards writing
committee as advisory members but shall be limited
to voting on matters at the standards writing com-
mittee level. TCCS/EC will approve non-members to
standards writing committes when the standard
writing activity requires special expertise or broad
representation to create balanced interest in that
activity.
10. All TCCS standards writing committees shall give
consideration to:
a. The existence of other comparable standards
having national acceptance in a given field.
b. The potential use of a given standard.
c. The technical quality and use of established,
broadly accepted engineering practices.
269
-------
d. The results of patent rights or other factors which
might favor a product for an industry.
e. Minority or dissenting views of members of
standards writing committees. Such views, if not
resolved, must be reported in detail to TCCS/EC.
11. A simple majority of a subgroup and standards
writing committee is necessary for actions at the
operating level of such groups.
12. Every effort shall be made by a TCCS standards
writing committee to obtain a unanimous vote of its
members on a proposed standard. If a unanimous
vote cannot be obtained, the proposed standard shall
have the approval of at least three-fourths of the
total standards writing committee members. Each
negative vote shall be accompanied by a detailed
reason for such vote and shall be circulated to all
committee members. Every effort should be made to
resolve negative votes. The responses to each dis-
senting comment by the standard writing committee
shall be sent to the proponent. All dissenting votes,
if not resolved, must be reported in detail to
TCCS/EC.
13. After approval of a proposed standard by a TCCS
standards writing committee and transmittal, along
with detailed information on negative votes, to
TCCS/EC, the proposed standard and the voting shall
be reviewed by TCCS/EC. Consideration shall be
given to negative votes and procedures. If deemed
necessary, the proposed standard shall be sent back
to TCCS's standards writing committee for
clarification.
14. The TCCS/EC upon approval of a proposed standard
shall direct the ASCE staff to take the necessary
action to publish a notice to the full membership of
the TCCS. Council members will be given an
opportunity to vote on all standards and this will be
designated as the Council Ballot. Cut-off dates for
such balloting shall be 6 weeks from the notice
date (exact procedures for communicating with the
Council not yet established).
a. Council Ballots shall be canvassed by TCCS/EC as
early as possible after their receipt at ASCE
Headquarters.
b. All negative votes, to be valid, shall be accom-
panied by a written explanation based upon either
technical or procedural considerations and must
be directly related to the proposed standard.
c. Negative votes shall be sent back to the appropri-
ate TCCS standards writing committee for
consideration and resolution. The standards
writing committee shall report back to TCCS/EC
the detailed consideration and resolution of all
negative ballots. TCCS/EC shall determine that
the standard was developed in conformance with
these procedures and that due process was given
to all negatives.
d. A proposed standard must receive approval by
two-thirds of those casting valid ballots.
15. Upon final approval of a proposed standard by TCCS,
it may be submitted to the American National
Standards Institute (ANSI) for approval as an
American Standard.
16. The TCCS/EC shall establish procedures to insure
that ANSI criteria for accreditation are maintained
on a continuing basis.
17. Copies of all reports (special reports, minutes, etc.) of
of TCCS standard writing committees shall be sub-
mitted to TCCS/EC and ASCE Headquarters.
270
-------
TECHNICAL REPORT DATA
(f lease read Instructions on the reverse before completing)
REPORT NO.
EPA-600/9-78-021
2.
. RECIPIENT'S ACCESSION NO.
TITLE AND SUBTITLE
Proceedings: Workshop Toward an Oxygen Transfer Standard
Proceedings of a Workshop Cosponsored by the U.S. Environmental
Protection Agency and the American Society of Civil Engineers
REPORT DATE
April 1979
6. PERFORMING ORGANIZATION CODE
AUTHORISI
Edited by:
William C. Boyle, University of Wisconsin — Madison
i. PERFORMING ORGANIZATION REPORT NO.
PERFORMING ORGANIZATION NAME AND ADDRESS
American Society of Civil Engineers
345 West 47th Street
New York NY 10017
10. PROGRAM ELEMENT NO.
IBC822, SOS#3, Task D-2/13
11, CONTRACT/GRANT NO.
Grant No. R805868
2. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati OH 45268
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings: April 11-14. 1978
14. SPONSORING AGENCV CODE
EPA/600/14
Si SUPPLEMENTARY NOTES
Project Officer: Richard C. Brenner (513)-684-7657
6. ABSTRACT
The Workshop proceedings documented in this report represent a major effort to summarize historical
practices and current art in the testing and evaluation of oxygen transfer devices used in the treatment of
wastewater. These proceedings form the technical base from which the American Society of Civil Engineers'
Subcommittee on Oxygen Transfer Standards is attempting to develop a tentative interim oxygen transfer
standard on this research grant project.
The objectives of this Workshop were to bring together experts in the field of oxygen transfer to:
(1) identify areas of agreement and disagreement in the evaluation of oxygen transfer devices and
(2) identify research needs in the development of an effective consensus standard for oxygen
transfer devices.
The first day-and-a-half were devoted to the presentation and discussion of topics related to the
testing and evaluation of oxygen tranfer devices. The entire Workshop roster was then divided into five
working group sessions, whereby intensive discussion took place in efforts to arrive at consensus opinions
on selected topics. Areas of agreement and disagreement were delineated. The findings of each working
group were presented to the reassembled total group on the final day of the Workshop.
The papers presented in the opening day-and-a-half general session and the summary reports of the
working groups comprise the major portion of these Workshop proceedings.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
•Oxygen, transferring, Standards,
Sewage treatment
*0xygen transfer standard.
Aeration devices. Clean
water aeration test. Interim
standard
13B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS {This Report)
Unclassified
21. NO. OF PACES
283
20. SECURITY CLASS (This page I
Unclassified
22. PRICE
EPA Form 2210-1 (*-7J)
271
& U. 5. GOYHNMENT PHWIM6 OFFICE: 1979-657-060/16-39 Region No. SHI
------- |