EPA-600/2-77-192
August 1977
FEASIBILITY STUDY ON
IN-SEWER TREATMENT METHODS
by
Richard D. Pomeroy
Ronald J. Lofy
Pomeroy, Johnston and Bailey
Pasadena, California 91105
Contract No. 14-12-944
Project Officer
Gerald Stern
Wastewater Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
TECHNICAL REPORT DATA
(Please read launicnons on she reverse before completing)
1 REPORT MO
EPA-600/2-77-I92
3 REC PlENT S ACCESSIOWNO .
"PILE AND SUBTITLE
FEASIBILITY STUDY ON IN-SEKER TREATMENT METHODS
5 REPORT DATE ;
August 1977 (Issuing Date)
6 PERFORMING OH<3AN1ZATIO"N CODE
7 AUTHQR'.S)
Richard D. Pomeroy and Ronald J. Lofy
8 PERFORMING ORGANIZATION REPORT NO
1BC611
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Pomeroy, Johnston and Bailey
660 South Fair Oaks Avenue
Pasadena, California 91105
10 PROGRAM ELEMENT NO
14-12--944 (11010 FWG)
13 SPONSORING AGENCV NAME AND ADDRESS
Municipal Environmental Research Laboratory-- Cin ., OH
Office of Research and Development
U. S. Environmental Protection Agency
Cincinnati, Ohio 45265
13 TYPE OF REPORT AND PERIOD COVERED
Final
11 SPONSORING AGENCV CODE
EPA/600/14
15 SUPPLEMENTARY NOTES
Project Officer: Gerald Stern (513) 684-7654
The possibilities of in-sewer treatments of wastewaters are broadly covered, in-
cluding discussions of biological, chemical, and physical methods. The best possi-
bilities are in the direction of biological oxidation, leading to partial reduction
of the BOD, or extending to the equivalent of standard complete treatment. An
appreciable reduction of BOD occurs normally in sewers. The amount of this oxidation
can be estimated by equations developed in this and other closely related research.
The requirements for a high degree of treatment are an adequate oxygen supply, an
oxidizing culture, and tine The culture may be suspended in the form of an activated
sludge, or attached to solid surfaces. The methods for supplying atmospheric or
industrial oxygen are explored. The prospects for a useful degree of in-sewer treat-
ment, especially in force mains, are very good. Under certain limited conditions the
sewer can function as an efficient flocculation device in conjunction with chemical
treatment.
KEY VVOR3S AND DOCUMENT ANALYSIS
DESCP PTORS
b IDENTIFIERS/OPEN ENDED TERMS
c COSATI Ficld/Groyp
Biochemical oxygen demand, Biological
productivity, Chemical removal (sewage
treatment), Sewage treatment, Waste treat-
ment, Activated sludge process. Sewage,
Organic wastes, Sewage disposal-sewers,
toaste water
In-sewer treatment,
Force main sewage treat-
ment, Gravity flow sewag
treatment
13B
8 0JSTR>aUTlON STATEMENT
RELEASE TO PUBLIC
19 SECURITY CLASS (This Report)
UNCLASSIFIED
20 SECURITY CLASS (Tim page/
UNCLASSIFIED
/
I
EPA Fo-m 2Z20-1 (9-73)
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U. S. Environmental Protection Agency, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect the
views and policies of the U. S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
ii
-------
FOREWORD
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 solu-
tion 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, treat-
ment, 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 publica-
tion is one of the products of that research: a most vital communications
link between the researcher and the user community.
This report discusses the possibilities of in-sewer treatments of
wastewaters including discussions of biological, chemical, and physical
methods. The greatest opportunities are in the direction of biological
oxidation, leading to partial reduction of the BOD, or extending the
equivalent of standard complete treatment. Requirements for a high degree
of treatment are discussed, equations for oxidation are discussed, and
methods for supplying atmospheric or industrial oxygen are explored.
Francis T. Mayo, Director
Municipal Environmental Research
Laboratory
iii
-------
ABSTRACT
The possibilities of in-sewer treatments of wastewaters are broadly
covered, including discussions of biological, chemical, and physical
methods. The greatest opportunities, by far, are in the direction of
biological oxidation, leading to partial reduction of the BOD, or extend-
ing to the equivalent of standard complete treatment. An appreciable
reduction of BOD occurs normally In sewers. The amount of this oxidation
can be estimated by equations developed in this and other closely related
research. The requirements for a high degree of treatment are an adequate
oxygen supply, an oxidizing culture, and time. The culture may be sus-
pended in the form of an activated sludge, or attached to solid surfaces.
The methods for supplying atmospheric or industrial oxygen are explored.
The prospects for a useful degree of in-sewer treatment, especially in force
mains, are very good. Under certain limited conditions the sever will
function as an efficient flocculation device in conjunction with chemical
treatment.
This report was submitted in fulfillment of Contract No. 14-12-94A
by Pomeroy, Johnston, and Bailey under the sponsorship of the U. S.
Environmental Protection Agency. This report covers the period from
September 23, 1970, to March 1972, and work was completed as of June 1,
1977.
iv
-------
CONTENTS
Foreword ill
Abstract iv
Figures. vi
Tables vii
Symbols viii
Acknowledgment xii
1, Introduction 1
2. Conclusions , 3
3, Recommendations . ..... 6
4. The Oxygen Supply for Biological Oxidation in Sewers. ..... 7
5. The Oxidizing Culture for Biological Oxidation in Sewers. ... 63
6. Chemical Treatment. . 77
References , 83
Appendices
A. Self Purification in Sewers . 86
B. Calculation of Equilibrium Conditions 102
-------
FIGURES
Page
1 Reaeration rates in sewers flowing half full 12
2 Relative aeration rates in a sewer at different flows,
compared to the rate for the half filled pipe 14
3 Schematic and photograph of waterfall aeration equipment 19
4 Hypothetical design options 26
5 U-tube design, Jefferson Parish Sation 5 37
6 Oxygen absorption in a U-tube 38
7 Functional drawing of an air-lift pump station 41
*
8 Possible vays to diffuse air into a sewage stream 47
9 A Jet aeration installation 49
10 Calculations of absorption of pure oxygen in wastewater
devoid of oxygen 56
11 Installation to provide oxygen supplement for
overloaded activated sludge plant 61
12 Greater detail of oxygen-dissolving tank 61
13 Treating the stream in a sewer with oxygen 62
14A Graph for determining K (high values) 70
14B Graph for determining K (low values) 71
15 Solutions for Examples A and B 74
16 Values of velocities, u, and velocity gradients, G 80
vi
-------
TABLES
Page
1 Aeration Coefficients from Data on River Weirs 18
2 Waterfall Aeration Experiments Using Tap Water 21
3 Examination of Experimental Waterfall Aeration Data 22
4 Waterfall Aeration Experiments Using Sewage 23
f
5 Expected Oxygen Absorption in Sewage Falls 23
6 Hydraulic Properties for the Respective Options 26
7 Summary of Calculations of Oxygen Uptake 29
8 Amount of Air that can be Dissolved Under Pressure in
Water Already in Equilibrium with Air at Ambient Pressure 33
9 Experiments with Nozzle Aeration 50
10 Solubility of Oxygen in Pure Water, mg/1 55
vii
-------
SYMBOLS
Symbols used in Appendix I and Appendix II are shown In each Appendix.
C Hazen-Villiams coefficient
C factor in the predictive equation to represent the increase
in the exchange coefficient due to creation of additional
air-water interface as result of high turbulence
solubility of oxyger, mg/1
pipe diameter, m or ft or in.
depth of flow, m or ft
d mean hydraalic depth, cross section area of stream divided
"by surface width, m or ft
&, &i , &s oxygen deficit and deficits at beginning and end of test
reach or upstream and downstream of a waterfall, Kg/1
2 / 3
E rate of energy dissipation per unit mass of water, m /sec or
fta/sec3, page 78
E efficiency, mg of 0 per kw-hr of energy in the fall, page 24
viil
-------
submerged surface transfer coefficient
oxygen exchange coefficient, n/hr
du/dx, mean velocity gradient, t
2 2
acceleration of gravity, m/sec or ft/sec
H , H elevations of the hydraulic energy line upstream and downstream
from the fall
K_ reaeration coefficient
_T
K mural deaeration coefficient, hr
e
1C reaeration coefficient for a waterfall, m
1C a synonym for K,
Manning coefficient
dissolved o^gen concentration
P ambient pressure, atmospheres absolute
P, pressure of the injected air, atmospheres absolute
IX
-------
flow, m /sec or cfs
R rate of change of oxygen concentration due to its utilization
by Che slime layer on the pipe wall
rate of supply of oxygen to a stream, mg/l-hr
R rate at which oxygen is depleted by reaction in the stream
hydraulic radius, m
slope of the energy line of the stream
temperature, 'Celsius
time, seconds or hours
velocity of flow, m/sec or fps
volume of air at one standard atmosphere pressure, m
Gameson weir aeration coefficient (page 17)
Potential work of isothermal expansion (page 42)
-------
ratio of exchange coefficient ID impure water to the
coefficient in pure water at the same temperature and flow
conditions
ratio of exchange coefficient at temperature I to the
coefficient at 20*C, other conditions being the same
2
flux of oxygen from the stream to the slime layer, g/m -hr
2
flux of oxygen from the atmosphere to the stream, g/m -hr
viscosity
xi
-------
ACKNOWLEDGMENTS
This reconnaiaaance study was carried out primarily under Contract
14-12-944 of the Federal Environmental Protection Agency with the fina
of Poceroy, Johnston and Bailey. The project overlapped with the re-
search on the occurrence and control of sulfide in sewerage systens,
assisted "by a Federal Environmental Protection Agency grant to the Loa
Angeles County Sanitation Districts. Studies of reaction rates of oxygen
with sewage, and the data analysis for the absorption of oxygen in streams
constituting essential elements of the sulfide study, consumed part of
the time of Pomeroy, Johnston and Bailey assigned to this reconnaissance
study on in-line treatment, and the reconnaissance study in turn benefited
greatly by the painstaking data collecting job done by the Los Angeles
County Sanitation Districts that made possible the development of the
oxygen absorption equation, so necessary for understanding in-line
oxzdation.
Appreciation is expressed to John D. Parkhurst, Chief Engineer of the
Los Angeles County Sanitation Districts and Project Director for the
sulfide research project for making possible the cooperative aspects of
the two projects. We are also greatly indebted to Charles Swanson and
Gerald Stern of the Environmental Protection Agency for their advice
in the planning and execution of this project.
xii
-------
CHAPTER 1
INTRODUCTION
The possibility of in-line treatment of sewage urges itself upon the
attention of sanitary engineers. Sewage collection systems are large
In both cost and size, representing investments that are typically two
to four times the sums spent on treatment plants. In metropolitan
systems, the sewage may spend a longer time In transit than it does In
the plant. Quite naturally, the question arises as to whether the
severe may not serve a double purpose, providing treatment as well as
transportation, so that the sewage will reach the end of the line
partially or fully treated.
The concept is intriguing. Perhaps an overloaded treatment plant can be
relieved by effecting some of the treatment in the sewers. Perhaps an
integral design concept will economize on over-all construction by lay-
ing out the sewers in ways that will facilitate in-line treatment at
small cost, with substantial savings in treatment plant construction.
Perhaps in-sewer treatment will enhance the feasibility of withdrawing
water at up-streaai points for reclamation. Perhaps sulfide control
methods that maximize BOD reduction should be chosen in preference to
other methods of sulfide control.
The application of certain treatments to sewage in transit is not new.
Principally such treatments have been for the purpose of preventing the
generation of sulfide. For nearly a half century, chlorine and other
chemicals have been added for sulfide control, usually having the in-
cidental result of easing the load on the sewage treatment plant, but
only by a slight amount under most conditions. The application of air
to force mains, first undertaken for sulfide control by Ewald Lemcke
in 1942 in Orange County, California, (1) ie now widely practiced,
sometimes with noticeable reductions of BOD.
Even if nothing is dene toward the intentional treatment of sewage
while in transit, some degree of self-purification nevertheless does
-1-
-------
occur, because all of the oxygen absorbed at the surface of the stream
is consumed in partially oxidizing the sewage.
Some sewerage systems transport wastewater for distances approaching
100 km, and even longer systems may be built in the future. The poten-
tial for achieving significant in-sewer treatment during the course of
the time that this wastewater is in transit should not be overlooked.
A large trunk sewer of such length would probably accumulate serious
concentrations of eulfide if measures were not taken to increase the
oxygen supply. If sufficient oxygen were supplied to the stream, the
result would be not only the prevention of sulfide buildup, but also a
substantial degree of biological oxidation. The control of sulfide,
partial in-sewer oxidation, and complete biological treatment in the
sewer constitute a continuum. The objectives of sulfide control and
of complete biological oxidation are widely separated, but the tech-
nologies overlap, and an understanding of the absorption and utiliza-
tion of oxygen in the sewer is basic for the whole range.
A major part of this report is devoted to biological oxidation in the
sewer, since this process is the one that offers by far the greatest
opportunity for either partial or complete in-aewer treatment of waste-
water. However, the aim of the reconnaissance study is a comprehensive
one, and consideration has been given to two other processes — sedimen-
tation and flocculation. It is found that in-sewer treatment by these
processes is possible, but the opportunities for application are •
extremely limited.
-2-
-------
CHAPTER 2
CONCLUSIONS
1. The most promising form of in-Hne sewage treatment is biological
oxidation. The treatment may range frotn the minor BOD reduction that
occurs naturally in gravity sewers to the equivalent of normal sec-
ondary treatment.
2. The three requireaents for efficient biological treatment are an
adequate oxidizing culture, a supply of dissolved oxygen, and time.
3. If the oxygen supply is adequate, the sewage develops a capability for
using oxygen at rates from 5 mg/l-hr up to as high as 20 mg/l-hr. The
oxygen utilization rate can be increased by returning activated sludge.
Schemes to increase the area of slime-supporting surfaces beyond what
is available on the pipe wall do not look, promising.
4. Measurements were made of the rate of absorption of oxygen in sewers,
and a well-substantiated predictive equation was developed. Very small
flows can keep themselves well aerated, and, given a few hours of time,
would accomplish good biological oxidation. In large trunks of, say,
a meter or more in diameter, the rate of oxygen absorption is too saall
to accomplish any significant degree of oxidation.
5. Freshening the sewer atmosphere by ventilation will have no appreci-
able effect on the oxygen supplied to the stream by surface aeration.
Enrichment of the atmosphere with oxygen has little chance of being
practical.
-3-
-------
6, The actual oxygen resources in a sewer are not limited Co surface
aeration of a stream, but are supplemented by the turbulence of Junc-
tions and drops. The oxygen absorption at a drop is much mote than
occurs with the same loss of elevation in a quietly flowing stream.
Purposely designed sewage falls, U-tubes, and other devices in which
the potential energy of a sewage stream is used to mix the sewage with
air can be useful means for increasing in-sewer oxidation.
7. A series of low air lifts along relatively flat sewers would add to
the dissolved oxygen supply and would aid in maintaining adequate
sewer velocities.
8. Diffused air is a feasible way to increase in-line oxidation in large
trunks.
9. Streans of screened sewage or effluents jetted at high velocity into
sewage streams is a possible way to increase the oxygen supply.
10. Mechanical aerators do not appear very practical for installation in
existing sewers,
11. Injection of air into force mains accomplishes a degree of biological
oxidation that can be substantial if the detention tine is long.
12. Refined oxygen is in a very favorable position for use in force mains
or in any structure in which the sewage is placed under pressure,
including U-tubes• A good degree of biological oxidation appears
possible if the retention time is a few hours.
-A-
-------
13. Refined oxygen also can be used advantageously for treatment In
a trunk sewer, by Injecting a supersaturated solution into the
sewage stream at suitable intervals.
14. Slope-flow relationships in most sewers produce velocity gradients
suitable for flocculation of aluminum hydroxide and iron hydroxide
precipitates. Application of the chemicals at upstream locations where
there will be flow times of 10 minutes to 30 -minutes ahead of a
treatment plant will result in well flocculated sewage reaching the
plant. This will be advantageous only if the sewage can then be
transferred to the primary settling tank without floe-disrupting
turbulence.
-5-
-------
CHAPTER 3
RECOMMENDATIONS
1. Engineers should consider the oxygen "balance of sewers when they are being
designed, and should choose options that favor an increased oxygen supply
to the stream. Keeping the stream aerobic should be the preferred way to
control sulfide and linimiae other sewage odors.
2. An experimental sewer should be constructed, perhaps paralleling an exist-
ing sewer that has inadequate slope, with a series of law-head air-lift
pumps and increased sewer slopes between.
3. Experiments should be made in a force main that provides a detention time
of several hours, to measure the rate of oxygen usage by the slime layer,
and other factors influercing the over-all biological oxidation. Air as
well as oxygen should be tried. An existing force main can be found that
will be suitable for these experiments.
A. A demonstration project should be undertaken in which a supercharged
oxygen solution is injected into the stream in a large trunk sewer.
5. A demonstration project should be undertaken in which the stream in a
large trunk sewer is aerated by diffused air.
6. A few more experiments should be made to round out the work, already started, on
aeration by waterfalls and jets, using an improved apparatus. Observations
should be made of the amounts of oxygen absorbed at existing points of
high turbulence in severe.
-6-
-------
CEAJTER 4
TEE OXYGEN SUPPLY FOR BIOLOGICAL OXIDATION IN SEVERS
BASIC REQUIREMENTS
There are three basic requirements for a biological purification process!
l) an osygen supply, 2) a culture of aerobic organises, and 3) time. The
first two are largely independent and can "be treated separately. The
tine requirement is Implicit or explicit in the discussions of the first
two.
This section deals with the oxygen supply. Both spontaneous dissolution
of oxygen from the atmosphere and special aeration or oxygenation tech-
niques are considered.
SCOP3 OF THIS CHAPTER
There are numerous processes that can add dissolved osygen to the waste-
vater in a sewerage eysten. These include the processes that occur
spontaneously in existing systems, and that can be amplified in nev
systems by suitable choices of design options, and also processes or
devices that are not features of conventional systems but that can be
installed for the purpose of increasing the oxygen supply, either in
existing or new systems.
The oxygen supplied by some of these processes may be trivial under
ordinary conditions. They arust nevertheless all be examined in. a
reconnaissance of the subject, in order that they can be evaluated. The
possibilities for full secondary biological treatment in sewers are very
limited, tut substantial BOD reduction is often possible. Even a trivial
reduction of BOD is worthwhile if it is accomplished at trivial coat.
-------
More attention is being given to the dissolved oxygen balance in sewers
as the major defense against sulfide problems. Where a process is
employed to add dissolved oxygen for the purpose of sulfide control, an
inevitable consequence is that there will be a reduction of the BOD or
COD load reaching the treatment plant. The reduction is sometimes more
than trivial.
The subject matter of this chapter will be presented under the following
headingsi
—Power Efficiency as a Feature of Aeration Processes
—Spontaneous Surface Aeration
—Ventilation
—Aeration at Falls and Other Points of High Turbulence
—Air Injection Into Pressure Mains
—U-tubes, Air Lifts
—Installation of Aeration Devices in Sewers
—Use of Commercial Oxygen
POYER EFFICIENCY AS A FEATURE OF AERATION PROCESSES
The choice of a process to add dissolved oxygen to wastewater will be
based upon several considerations, including capital cost, operation and
maintenance cost, space requirements, enviro'nmental effects, etc. It is
beyond the scope of this study to discuss all of these variables for all
oxygen dissolving processes. However a major iten of operating cost in
many of these processes is energy. The amount of oxygen dissolved per
unit of energy is therefore an important characteristic, and one that
merits special attention.
The energy efficiency of an oxygen-dissolving process is usually expressed
as kilograms of oxygen dissolved per kilowatt hour of energy.
To be meaningful, it is necessary to know what" energy is referred to.
It may be the electrical energy delivered to the motor, energy trans-
-------
mitted "by the drive shaft of a motor or other priiae mover, potential
energy and/or kinetic energy of water, energy of addabatic compression of
air, energy of isothermal compression of air, lose of efficiency of a
pruap or a tur"bine generator, etc. There nay be appropriate uses for all
of these measures, but confusion often arises when the basis is not
clearly stated.
The reliable evaluation of the oxygen-dissolving efficiency of an aera-
tion process ie often a difficult and tedious undertaking. It would
require a longer treatise than can be included here to deal with the
sources of errors in such measurements. Suffice it to say that reported
efficiency should be examined critically as to methodology, and with
whatever degree of skepticism is appropriate before accepting the results.
Usually the observed rate of oxygen -uptake is adjusted to the rate for
pure water at zero dissolved oxygen, one standard atmosphere ambient
pressure, and 20 deg C, but sometimes these adjustments are not appro-
priate. It ie unlikely that a useful purpose would be served by trying
to establish a standardised efficiency rating at this time, but insofar
as possible, different systems will be here compared on the customary
basis.
NOSMAL SHIFACE AERATION
It is well established that the flux of oxygen through the surface of a
stream is proportional to the oxygen deficit, that is, the difference
between the dissolved oxygen concentration and the concentration that
would prevail if the stream were in equilibrium with the auperjacent
atmosphere. The relationship is expressed by the following equationi
tf = ffi
where 0. = oxygen flux through the stream surface, g/m -hr
-9-
-------
f, defined by the equation, ia designated the exchange coefficient,
m/hr (f is practically the saae as Kj used by some authors)
£ = oxygen deficit of the atream in respect to the superjacent
atmosphere, g/m' or mg/1
The reaeration rate, Rj, is defined as the rate of change of oxygen
concentration as a result of oxygen absorption, and is expressed in
mg/l-hr. It is related to other quantities by the equationi
vhere
d = mean hydraulic depth, m, defined as the cross-section area
of the stream divided by the surface width
The reaeration coefficient, K£, is the proportional rate of satisfying
of the oxygen deficit, expressed in reciprocal ho-ors. It is related
to other quantities by the equation!
S2 = Rf/* = f/djj.
Kp ia the same as K^a as currently used by some authors.
a is practically the same as the reciprocal of d^.
If oxygen absorption is the only process affecting- oxygen concentration,
the rate of change of oxygen concentration may be represented as follovai
• &-Z &
"f ~at ~ dt ~ V
where
[CUJ = oxygen concentration, mg/1
t = time, hours
Integration and insertion of limita yields the equation!
$. (Restriction! surface aeration
2.303 l°ejg~ = Kg (t1 - t ) is the only factor influencing
2 oxygen concentration)
-10-
-------
It has "been found (2) that the exchange coefficient for oxygen between
the sever atmosphere and the stream can "be predicted quite reliably
by the empirical equation:
f = 0.96 CAy(Bu)3//B
where C^ = factor representing the creation of new interface
in a highly turbulent stream
y = temperature function, equal to unity at 20 d.eg C
B = elope of the energy line of the streaa
u = velocity, m/sec
C. can be approximated as follows 1
c
A
where g = gravitation constant
u /gd = Proude number} it is the square of the Prou.de number
m commonly used in the United States in respect to flow
in streams .
Only in quite shallow or swiftly flowing wastewater streans does CA
exceed 1.2; mostly it is below 1.1.
The y function of tempera-bore ia influenced by the turbulence of the
stream. In very slow streams y may increase 4 percent per degree
Celsius; in turbulent streams the change is nuch less (3). Until such
time as more information is available, the best estimate is that y
increases about 1 percent per deg C under typical flow conditions in
sewere .
Figure 1 has been drawn to show values of IL in sewers flowing half full
at 20 deg C. Prom the velocity and pipe diameter, slope was calculated,
using a half-pipe Essen-Williams coefficient of 109, eq.uivalent to a
full-pipe coefficient of about 124, a typical value for functioning
-11-
-------
FIGURE I
Rl-AERA'IION RATES IN SEWERS
FLOWING HALF FULL
VELOCITY, f ps
E
r»-
H
K
o
C
UJ
O
rsi
O
UJ
5
K
UJ
<
UJ
GC
-12-
-------
sewers. If the calculations had teen made with a higher assumed Eazen-
Williaiaa coefficient, the calculated slope for the sane velocity and
flow would have "been lower, and the calculated f and E2 would have "been
lower, varying inversely as the 0.7 power of the assumed value Of C.
For a given pipeline (fixed slope) and Q, an increase of actual C vould
increase the velocity and decrease the mean depth, thus increasing R_
approximately in proportion to C.
The values of f are to be multiplied by the oxygen deficit to obtain the
rates of oxygen supply, R,,, in mg/l-hr,, The oxygen deficit is influenced
by ambient atmospheric pressure and by the oxygen content of the sewer
atmosphere, which may be aa low ae 90 percent of normal, or lees if air
flow in the sewer is blocked. Assuming 95 percent as typical, and an
atmospheric pressure of, say, 0.99 standard atmosphere, which wojld be
at an elevation of about 90 m above sea level, the solubility at 20 deg C
vould be 8.5 Tag/1. To find the rate of dissolving of oxygen when the
concentration of oxygen in solution is enough to maintain good aerobic
action, one should use a figure of about 7 mg/1 for the deficits The
scale on the right side of graph in Figure 1 shows the oxygen supply
rate under these conditions.
An increase of temperature increases f. More important, however, is
the effect of temperature on the solubility of oxygen. To maintain a
desired level of dissolved oxygen, a lower osygen deficit will be
available at the higher temperature.
The effects of relative depth of flow in a given sewer are as shown in
Figure 2. The calculation was made for a 24-in. (0.61 m) pipe that
flows at a velocity of 1.0 m/flec when half full (1.14 m/sec when full).
A curve calculated for any other pipe size and velocity within reason-
able limits wou3d differ imperceptibly from Figure 2, and then only
because of changes in C.. Factors read from Figure 2 can be used as
urultipliers for reaeration coefficients read from Figure 1. In using
Figure 1 in this way the velocity that should be used is the velocity that
-13-
-------
FIG. 2 RELATIVE AERATION RATES IN A SEWER AT DIFFERENT FLOWS,
COMPARED TO THE RATE FOR THE HALF FILLED PIPE
tr
u.
O
M
U.
O
(O
UJ
UJ
UJ
a:
\
0.2 0.4 0.6 0.8
RELATIVE DEPTH OF FLOW
1,0
-14-
-------
would prevail If the Bewer were flowing half full, not the actual vel-
ocity at the actual depth of flov.
The concept of power efficiency in the dissolution of oxygen is not very
useful in respect to surface aeration of streams. The energy dissipation
in a sewer where uniform flov conditions prevail is proportional to euA,
in which s is the slope, -u is the velocity, and A ia the cross-section
area of the stream. The amount of oxygen dissolved, however, is propor-
tional to (BU) b, in which b is the surface width of the stream. A
wide range of power efficiencies results, depending upon slope, velocity,
and size of the stream, but the efficiency is always much poorer than In
processes where air is vigorously mixed with the water.
The possible significance of normal surface aeration in a sewer can be
appraised by use of Figures 1 and 2. Consider a sewer 30 cm in diameter,
designed for a flow velocity of 1.0 m/sec when flowing half full, but
with an actual depth of 7-5 cm. The oxygen supply at low dissolved
oxygen concentration would be 18 mg/l-hr. The osygen consuming capability
of wastewater and slime would soon approach this rate. If a sewer flowed
under these conditions' for 5 to 10 km, and then passed through a clarifier,
the system would be equal to a trickling filter plant in respect to the
quality of effluent. There is probably no place where this hypothesized
condition exists. In large systems the wastewater is collected into
larger trunks with flatter slopes and deeper flows, where the osygen
supply is too small to provide a significant degree of oxidation.
It is concluded that spontaneous aeration at the surface of the atream
will seldom be sufficient to support a substantial degree of biological
purification. Nevertheless, it is a significant factor in the oxygen
balance.
VENTILATION
The calculation of reaeration rate, as shown by the right-side scale of
-15-
-------
Figure 1, is valid only for an oxygen deficit of 7 mg/1. If the oxygen
in the sewer atmosphere becomes partly depleted, the deficit, which is
the difference between the dissolved oxygen concentration and the concen-
tration that would be in equilibriun with the existing sewer atmosphere,
ia reduced. Oxygen absorption is correspondingly reduced.
The idea suggests itself that ventilation of the sewer ia a way to improve
conditions by eliminating this impoverishment of the sewer atmosphere.
The suitability of ventilation for this purpose in any given situation
can be judged by a consideration of the severity of oxygen impoverishment
and the benefit that would result.
Sewer atmospheres are in fact not greatly depleted in oxygen content
under normal conditions. Upstream from a complete blockage of air move-
ment in the sewer, the oxygen concentration may sometimes be reduced to
half of its normal value, but otherwise this does not happen. In small
sewers the oxygen content is usually between 95 and 100 percent of normal.
It is evident that ventilation would be of no significant value in auch
severs. In large trunks the concentration may sometimes be as low as
90 percent of normal, or perhaps 85 percent in extreme cases* However,
the oxygen dissolved by surface aeration in large trunks is a rather
inconsequential amount. Even if there is a reduction of 10 to 15 percent
of the normal reaeration rate because of a stale sever atmosphere, venti-
lation would have a negligible effect on the oxygen balance of the stream,
AERATION AT FOISTS OF EICJH TOH3ULENCE
In addition to surface aeration in the normally flowing streams in severs,
oxygen is added at junctions, drops, hydraulic juaps, and other places of
intensive turbulence that mixes air with the water.
From the standpoint of analysing its effects, the simplest form of a
high-turbulence process is a waterfall.
-16-
-------
falls
Measurements of aeration of water falling over river weirs were made by
(Jameson (4), and by Barrett, Gameson and Ogden (5). Gameson, Vandyke
and Ogden (6) made measurements with an experimental fall, with the
water falling into a pan 9 cm deep from heights ranging from 0.6 to 3-3 n
Gameson showed that the effect of a weir can be modeled by an equation
which may be written
fi
-IP - 1 = V(IL - H )
2
where $ and fi_ = upstream and downstream oxygen deficits
Vf = weir coefficient
H.. and E_ = elevations of the energy line upstream and downstream
from the fall
It has been shown (2) that an alternative ezpression ie also applicable:
2.303 log^ = K_(H_ - Hj
where E_ = waterfall aeration coefficient, m
a
Orer the range of conditions represented by the published results of
GameBon and coworkers, it appeared that one equation is about as good as
the other. Vhere the deficit ratio is infinitesinally greater than unity,
K_ = V, Where the ratio ie 2, that is, where half of the deficit renains
unsatisfied, K_ = 0.69 V. For the purpose of comparing efficiencies of
n
different aeration methods, it is more convenient to use the logarithmic
form,
Gameeon classified the streams where tests were made ELB "slightly
polluted," where "there appeared to be no appreciable pollution";
"moderately polluted," where the water "contained a proportion of sewage
effluent"; and "grossly polluted," where the streau was largely or en-
tirely sewage effluent. Average results vere as shown in Table 1. The
-17-
-------
averaging vas done with the uae of weighting factors advised by the
author.
Table 1. ASRATION COEFFICIENTS FROM DATA ON RIVER WEIRS
Condition
Slightly polluted
Moderately polluted
Grossly polluted
Number of
locations
9
23
5
V af1
0.81
0.58
0.35
IH. m-1
0.59
0.41
0.28
Experiments on aeration by waterfalls was undertaken as a part of this
research project. For the purpose, the equipment shown in Figure 3 was
constructed. The pump takes water from the task and returns it ty way of
a dlatoiiite filter and a rotameter to the weir box. A thermometer and
duplicate oxygen probes are inserted into the pipeline from the tank.
The water can also be returned to the tank by way of a carbon filter to
remove impurities, but not during an oxygen absorption run. The" carbon
was used from time to time between runs. The weir box is positioned on
an elevator, and the water cornea to it by way of a hoae, so the height
of the fall can easily be adjusted.
After filling the tank to the desired level and then cleaning the water
with the carbon, filter, a small amount of cobaltoua chloride and enough
reagent grade sodium sulfite were added to deoxygenate the water. Re—
circulation* over the weir was started, and as aoon as the probes showed
the return of oxygen, readings were started and continued at one-minute
intervals. Samples were also withdrawn for analysis by the tfinkler
k_
method at intervals of about 5 minutes. The experimental run was termin-
ated after the dissolved oxygen concentration reached 6 to 7 mg/1. The
logarithm of the oxygen deficit was plotted against time, and from the
slope of the line the reaeration coefficient per hour was calculated,
-18-
-------
FIGURE 3 SCHEMATIC a PHOTOGRAPH OF WATERFALL AERATION EQUIPMENT
2- DISSOLVED OXYGEN
-PROBES
DIATOMACEOUS EARTH FILTER
ACTIVATED CARBON TANK
WEIR BOX ON
ELEVATOR
-19-
-------
and was corrected to 20 deg C. Thie correction wee made on the assumption
that the weir coefficient increases by 2.4 percent of the 20 deg C value for
each degree rise of temperature, as indicated by the data of Gameson,
Van Dyke and Ogden (6). Using the rate of energy dissipation in the fall
per unit volume of water in the Byaten, K_ was, determined. Table 2 shorre
n
the results,
It developed that the greatest source of error was the lack of homogeneity
in the tank. This caused the results to be erractic, as may be seen by
examining the Table. A better equipment design would have teen
one in which water is dravn from several points in the pool.
There are four measured independent variables (including width of weir)
with rather scattered data. A multiple regression analysis to discern
the effects of the variables would not likely be reliable, so the data
were grouped and compared, as shown in Table 3. No significant differences
of Z_ can be attributed to height of fall, depth of pool, or width of
M
weir, but the results with flows of 132 and 151 liters per minute (55 to
40 gpm) ere lover than the average for all of the lower flows by an
amount 2.2 times the standard error of the difference as calculated from
the standard deviations. Indications from this type of statistical
examination should be viewed in the light of the limitations imposed by
the basic assumptions of statistics. The reliability of experimental
results is generally overestimated because the sources of error are rarely
random and independent. The data in this case suggest that efficiency was
less at the higher flows, but this is not a conclusive finding.
The average value of K_. at 20 deg C for all runs is 0.79 m~ . Prom the
n
20 deg C data of Gameson, Van Dyke and Ogden (&} for a 2-ft fall of clean
water, a value of 0.82 m~ is calculated for K_.
-20-
-------
Table 2, WATERFALL AERATION EXPERIMENTS USING TAP WATER
Independent Variables
Flow
rate,
1/min
Results for
130
57
57
130
57
130
57
130
150
130
57
130
150
130
114
95
76
57
151
130
114
95
76
151
114
76
38
Results for
130
130
57
57
130
130
95
Depth
of
water,
cm
12-in.
30
30
30
30
76
76
76
76
46
53
53
76
30
30
30
30
30
30
25
25
25
25
25
61
61
61
61
3-in. (
76
76
30
53
53
53
53
Height
of
fall,
en
{"30-cn) weir
152
152
107
107
107
107
152
152
152
107
107
62
152
152
152
152
152
152
122
122
122
122
122
122
122
122
122
1*5 cm) weir
152
107
107
107
107
107
107
Energy
of the
falling
water,
kw
0.0331
0.0142
0.0099
0.0231
0.0099
0.0231
0.0142
0.0331
0.0378
0.0231
0,0099
0.0132
0.0373
0.0331
0.0283
0.0236
0.0189
0.0142
0.0301
0.0265
0.0227
0.0189
0.0151
0.0301
0.0227
0.0151
0.0076
0.0331
0.0231
0.0099
0.0099
0.0231
0.0231
0.0165
Oxygen
trans-
ferred,
kg/hr
0.0515
0.0335
0.0254
0.0467
0.0245
0.0490
0.0306
0.0785
0.0682
0.0359
0.0212
0.0390
0.0663
0.0739
0.0691
0.0622
0,0524
0.0413
0.0518
0.0599
0.0494
0.0461
0.0440
0.0733
0.0531
0.0365
0.0231
0.0778
0.0415
0.0243
0.0264
0.0676
0.0521
0.0389
Kg, m
waterfall
coefficient
at 20°C
0.52
- 0.79
0.85
0.67
0.82
0.71
0.72
0.82
0.60
0.52
0.71
0.98
0.78
0.74
0.81
0.87
0.92
0.97
0.57
0.75
0.72
0.81
0.97
0.81
0.78
0.80
1.00
0.78
0.60
0.81
0.95
0.97
0.94
0.78
-21-
-------
Table 3. EXAMINATION OF EXPERIMENTAL WATERlAiL AEHA7ION DATA
Effect of height of falli
Height range, in.
Height range, cm
Sunber of runs
Average K,,, m
a.
Effect of de-pth of pool:
Depth range, in.
Depth range, cm
Number of runs
Average 1L., m
n
Effect of width of vein
Width of veir, in.
Vidth of weir, cm
Number of runs
Average IL., m
n
Standard deviation of KO
n
Standard error of the averages
Effect of flow:
Plow range, gpn
Plow range, 1/min
Number of runs
Average X-,, m
u —£r
Standard deviatioii of L
ri
Standard error of the averages
24-12.
61-107
13
0.79
10-12
25-30
16
0.785
3
7.5
7
0.835
0.13
0.068
10-15
38-57
9
0.85
0.84
0.12
0.028
48
122
12
0.80
18-24
46-61
11
0.81
12
30
27
0.78
0.13
0.025
20-30
76-114
9
0.8-J
60
152
9
0.78
30
76
7
0.775
35-40
132-151
16
0.74
0.14
0.035
Standard error of the difference
0.045
-------
Two tests were made with settled wastewater from the Azusa plant of the
Los Angeles County Sanitation Districts, using the apparatus shown in
Figure 3- (it is assumed that raw and settled wastewater would not
differ significantly in oxygen absorption rate.) The sewage was
retained in the tank until the dissolved oiygen concentration was low,
hypochlorite was added to stop aerobic action, and then the oxygen
increase vas measured. Results are shown in Table 4.
Table 4. WATERFALL AERATION EXPERIMENTS USING SEWAGE
Results for 3-in. (7.5 cm) Weir
Independent variables
Flow
35
25
rate
~\ fmi-n
130
95
Depth
of water
in. .cm
24 31
24 31
Height
of fall
in. cm
42 107
42 107
Energy
of the
falling
water,
kw
0.0231
0,0165
Oxygen
trans-
ferred,
ke/hr
0.0361
0.0206
waterfall
coefficient
at 20°C
0.41
The average of these results compared to 0.79 m for clean water would
indicate »*= 0.59 for settled sewage in this aeration method. Because
of the erratic variations, and in view of the findings of (Jameson (4)
and the low a value for wastewater streams indicated by the studies of
Parkliurst and Fomeroy (2), it seems prudert to use the smaller of the
two results in Table 4, that is, Z_ = 0.41 m , when, estimating the
M
effect of drops in sewers. The corresponding value of * is 0.52. Using
K_ = 0.41 m~ , Table 5 has teen calculated to show the ozygen absorption
ji
that can be expected in wastewater falls of various heights.
Table 5. EXPECTED OXYGEN ABSORPTION IN SEWAGE FALLS
Calculated with K,, = 0.41 m"1
n.
E,
meters
i
1
2
3
Percent oxygen
deficit satisfied
18
33
56
71
E. meters
4
6
8
10
Percent oiygen
deficit satisfied
81
91.4
96.3
98.3
*a is the ratio of the oxygen absorption rate in an impure water to the rate
in pure water, other conditions being equal.
-23-
-------
The energy efficiency for the dissolving of oxygen in a waterfall can
be calculated by the relationship
B = 0.367 C E_.
6 II
where B = efficiency, kg of 0_ per kw-hx of energy in the fall
C = solubility of oxygen, mg/1
0 _1
Kg >* vaterfall aeration coefficient, m
0-367 = kg-meter equivalent (x 10~6) of one kw-hr
For pure vater at 20 deg C, Ce = 9.02 and KH = 0.79 m~ ; the efficiency
is 2.62 kg/kw-hr. The efficiency of a fall for aerating vastewater for
vhich Kg ie taken to be 0.41 m~ is 1«35 kg per kv-hr. Assuming that
the vater is pumped to produce the fall, and that the combined efficiency
of motor and pump is 55 percent, about 1»4 kg of oxygen would be dissolved
in pure water per kv-hr of electrical energy supplied to the motor, and
about 0.75 kg per kw-hr in the case of sewage.
The assumed wire-to-water efficiency Is lower than ie attainable under
ideal conditions, hut is a reasonable estimate for average pumping
installations.
As in all efficiency ratings for aeration devices, the efficiencies above
stated are for zero DO, and thus vould "be strictly applicable only to a
waterfall of very email height, starting with oxygen-free water.
The energy dissipated in a waterfall is much more efficient in promoting
the dissolving of oxygen than is the energy dissipated in a smoothly
flowing stream, presumably because the energy serves to mix air vigor-
ously with the water. It follows that if part of the available elevation
difference between tvo points along a proposed sewer can be dissipated
in a fall or in a series of falls instead of being used to provide steeper
slope and higher velocity, greater oxygen absorption will thereby result.
An imaginary situation will illustrate this point.
-24-
-------
It IB assumed that there will be a sever 5,COO m long with a total
elevation loss of 8 m, carrying a flow of 0.20 m /see (7 cfe). Three
design options are consideredt
l) a 91-cm (36-in.) pipe at a elope of 0.0006 to maintain a
velocity of 0.61 m/sec (2 fps), with a 5-it fall at the end;
2) same slope but with five one-meter falls; and
3) a continuous elope of 0.0016.
In the last case the pipe would be 76 cm (30 in.) in diameter. Figure 4
and Table 6 Illustrate the. main features of the three options. Calcula-
tions of oxygen transfer to the etrean follow the Table. Kg is taken
to be 0.41 E~ . For purposes of simplification, it is assumed that the
oxygen consumption rate equals the respective reaeration rate in all
situations, and that the oxygen deficit is 7 mg/1 before each fall.
-25-
-------
FIGURE 4 HfPCmeriCAL DESIGN OPTIONS
Table 6. HYDRAULIC PROPERTIES FOR THE RESPECTIVE OPTIONS
0-ption
1
2
3
Q,
m3/eec
0,20
0.20
0.20
m/sec
0.61
0.61
0.91
D
in.
36
36
30
cm
91
91
76
d/D
0.5
0.5
0.487
8
O.OO06
0.0006
0.0016
d •
m'
m
$. 360
0.360
0.289
-26-
-------
Option 1 —
The component addition of oxygen due to spontaneous surface aeration in
a distance of 5,000 meters at a elope of 0.0006 can lie calculated by
Rf = f*/dffi = (0.96) (1+0. 17 u2/gdta)(Bu)3/8S/ata
Time in transit is
=0.98 mg/l-hr
5000m —
u.ol m/ sec
The amount of orygen traaeferred by surface aeration is
(0.9B mg/l-hr}(2028 hr) = 2.2 ng/1
The logarithmic equation for waterfall aeration for a 5-m fall, icith
Eg = 0,41 gives
£
2.303 log^ = 5(0.41) = 2.05
2
=7.86
fi2 =0.9 mg/1
Oxygen uptake due to the fall = 7 - 0.9 = 6.1 mg/1
Total osygen uptake = 2.2 + 6.1 = 8.3 ng/1
-27-
-------
Qy\ip.n 2 —
Because the slope of the pipe for Option 2 is the same as in Option 1,
the spontaneous surface aeration is alao the same, and the amount of
oxygen transferred is 2.2 mg/1. For each falli
»
logj- = 0.178
2
7/*2 = 1.51
«2 = 4.6
Oxygen uptake at each fall = 7.0 - 4.6 = 2.4 mg/1
Total oxygen uptake = 2.2 + 5 x 2.4 = 14.2 mg/1
Option 3 —
The only oxygenation occurring in this situation ia that due to spon-
taneous surface aeration.
R, = f«d ~1 = (0.96)(l+C.17 u2/gd )(su)5y/8(7.0)d ^
x a ni m
= (0.96) (l+0.17-^*;
=2.17 mg/l-hx
Time in transit is • -— = 5500 sec = 1.53 hr
O.yi m/eec
The amount of orygen transferred in 5000 meters is
(2.02 mg/l-hr)(l.53 hr) = 3.3 mg/1
-28-
-------
The results of the calculations for the three different options are
summarized in Table 7. They illustrate the greater effectiveness of
falls, and particularly a series of low falls in comparison with one
between the falls.
high fall, provided that there is time for oxygen depletion,* There is
no conclusive evidence that a cascade made up of a series of closely
spaced falls is markedly better than one large fall.
Tatle 7. SDMMRY OF CALCULATIONS 07 OXTGEN UPTAKE
Hypothetical Design Options
OpA1,?1
1
2
3
Q,
i sec
0.20
0.20
0.20
Height
u, No . of
sec
0.61
0.61
0.91
/'
m/m
0.0006
0.0006
0.0016
D,
cm
91
91
76
of
falls
1
5 •
0
fall,
m
5
1
0
Surface
aera-
tion
2.2
2.2
3.3
transferred
Vater-
fall
aeration
6.1
5(2.4)
_ ma/I
Total
9.3
14.2
3.3
The example is imaginative, and is not intended to imply that any actual
case would provide comparable alternatives, or that the amount of oxygen
dissolved would be an important factor in the choice when costs and
other factors are considered. It does throw some light, however, on what
actually happens ir sewers, and it has a "bearing on the conditions of
sulfide buildup in severs, since relatively small differences of oxygen
input determine whether or not sulfide problems will develop,
The larger the sewer, the lower is the rate of oxygen absorption during
normal flow, as illustrated by Figure 1. A fall, however, is in general
presumed to cause the same increase of oxygen concentration whether the
flow is large or arall. This ia probably not true where the flow is BO
large in proportion to the drop in elevation that there is no free fall,
but only an acceleration of the stream followed by a hydraulic jump.
There is no doubt that oxygen absorption in such cases is greater than
-29-
-------
in smooth flow through the same elevation loss, "but it may be consider-
ably less than calculated by the equation for waterfall aeration.
Junctions
Junctions are also usually places of high turbulence. Sometimes a
tributary flow drops into a larger stream, or one or both streams are
accelerated and produce a hydraulic jump or standing wave. A change of
pipe diameter is also a point of abnormal turbulence.
Junctions, falls and other turbulence-producing structures are common
in sewerage systems, and in the large trunks they are the principal source
of dissolved oxygen. They do not, however, supply enough oxygen in the
large trunks to accomplish a significant degree of biological oxidation.
In the smeller sewers both surface aeration and points of high turbulence
provide oxygen in amounts that may be significant, and in hilly terrain
may suffice for a substantial reduction of BOD. A notable example is the
system serving the south slope of the San Pedro Hills in the Los Angeles
County Sanitation Districts. It was estimated (7) that BOD reduction in
the main trunk averages 60 to 75 ng/l» in addition to an unknown amount
of reduction in the small collecting severs. This ia an unplanned bonus
for the hilly terrain.
The fact that this much aeration can occur in a sewerage system suggests
that attention should be given, during the design of a systen, to the
possibility of maximizing this effect insofar aa feasible alternatives
may allow.
Where a pump station must be used, it may sometimes be feasible to raise
the discharge elevation to allow for one or more falls downstream from
the discharge point, provided proper precautions are taken against such
unwanted effects as release of hydrogen sulfide and other odors. The
-30-
-------
additional energy for punping to a greater elevation will be approximately
offset by a reduction In air requirement at the wastewater treatment
plant.
Aeration in JPumps
Another example of a point of high turbulence is a centrifugal pump.
Sometimes a snail amount of air can be mised with the water in a pump,
using some of the energy of turbulence that ie dissipated in the pump
anyway, without materially reducing the efficiency of the pump. More
information ie needed on the dissolved oxygen-energy relationships,
and on possible harmful effects on the pump. It ia possible that this
is an inexpensive way to add to the oxygen supply. A comparable pro-
cess ia the addition of air to a hydroturbine, which has been shown to
yield a high efficiency of osygen absorption in relation to the loss
of efficiency of the turbine (e).
Ad INJ3CTION INTO PRESSURE MAINS
The injection of air into force mains and inverted aiphons ±e a well
developed technology for the prevention of sulfide buildup in such mains.
A general discussion of principles, practices and results ie presented
in the U. S. Bnvironaental Protection Agency Sulfide Control Manual (9)«
On page 5-5 of that Xarual, BOD reductions of 16 mg/1, 60 mg/1, and
10 mg/1 are cited for three aerated pressure mains on which data are
available.
Air injection into pressure nains is not always practical, especially
where the profile is irregular, or where there is very little slope BO
that energy dissipation by the moving air bubbles does not dissolve
enough oxygen to be effective. Where air injection is practical, it is
an efficient cethod of aeration. The energy efficiency is dependent
upon several variables, including rate of air injection, rate of oxygen
-31-
-------
utilization by the wastewater, and detention time in the main.
Vhites Point Force Main of the Loa Angeles County Sanitation Districts
is a 61-cm (24-in.) pipe 1180-m long, with a pressure of 8.6 atmospheres
(291 ft of water) at the pump station end. What the efficiency of the
dissolving of oxygen would be with pure water in the pipe is not known,
but measurements made under normal operating conditions at relatively
low air injection ratea showed that 4.36 kg of oxygen were dissolved
per kw-hr of potential work of isothermal expansion of the injected
air. Isothermal expansion work is here used because that is the maxi-
mum work that could be exerted hy the air ae it rises through the
water. Actually, part of the energy is lost due to dissolving of the
gas, and, if the water is moving, part of it is expended in helping to
lift the water. The very good efficiency shown despite these subtrac-
tions from the energy available to produce turbulence is explained by
the increase of solubility of the gases at high pressure.
It was shown, for the Whites Point Force Main, that the work of
isothermal expansion of the air was 34 percent of the electrical energy
applied to the motor. Thus, the efficiency was 1»48 kg of oxygen per
kw-hr of electrical energy. At high injection rates the efficiency was
1.0 kg per kv-hr of electrical energy.
In three other force mains, operating at lower pressures, the efficien-
cies vere 2.3 to 3.4 ig of 02 per kw-hr of isothermal expansion energy.
It is probable that the compressor system could be designed to convert
4C percent of the electrical energy to energy in the air, and that an
oxygen-dissolving efficiency of 0,9 to 1.3 kg per kw-hr of electrical
energy may be a reasonable expectation.
-32-
-------
IT-TUBES, AI3. LIFTS, AND OTHER PRESSURE SYSTEMS
If diffused air vere used to increase the dissolved oxygen content of a
stream of sewage, it would not be economical to raise the oxygen content
very high because of the consequent reduction of efficiency. If air is
forced into water under high pressure, both oxygen and nitrogen can
"be dissolved in relatively large amounts. The injection of air Into a
high-pressure force main illustrates this effect. It ie necessary to
examine thia process in some detail when considering the purposeful
subjecting of the water to pressure to cause air to dissolve.
Starting with sewage containing nitrogen and argon in equilibrium with
the atmosphere, but with the oxygen biologically depleted, the amounts
of air that can completely dissolve are as shown in Table 8.
Table 8. AKOUOT 01 AIR THAT CAtr BE DISSOLVED UNDER PRESSURE
I3T VA7ER ALREADY IN EQUILIBRIUM WITH AIR AT AKBIENT PRESSURE
(Volumes of air are measured dry at one atmosphere pressure
and the temperature of the water.)
Temp. ,
deg
C
15
25
35
Amount of air dissolved
t)er atmosphere ease -pressure
Volume of air
di s s o Iv e d . ml /I
20.0
19.0
18.2
Weight of
dissolved,
5.68
5-21
4.8?
oxygen
me/1
As an illustration
Of the meaning of the table, suppose that it is required that 20 mg/1 of
oxygen be dissolved in wastewater at 25 deg C. The required pressure
would be 20 4- 5.21 = 3-9 atm = 40 m of water. It should be noted that
the table does not show the amount of gases that would dissolve if a
-33-
-------
large excess of air were contacted by the water, because in that case
the gases would not dissolve in the same ratio that they are found in
air. More oxygen would be dissolved if an excess of air were diffused
through the water at any given pressure.
A possible way to take advantage of the preasure effect for dissolving
air is to construct a TMrube. Air would be introduced into the rising
leg near the bottom of the TJ-tute, or would be introduced near the top
of the descending leg in a sufficiently dispersed condition so that it
is carried down by the water. Even if the air mist be pumped in under
quite high pressure, the power efficiency for getting the oxygen into
solution is good if all of the air dissolves.
There is a drawback that limits the effectiveness of dissolving air
under pressure. It is the reversal of the process when the water ia
returned to normal pressure. The tendency will be for the partial
pressure of the dissolved gasea to return to the ambient pressure. They
will not actually reach that state by the process of bubble formation,
because substantial supersaturation is required to initiate a bubble.
If the water is agitated or aerated at atmospheric pressure, equilibrium
will of course be approached. Without agitation, it is guessed teat
supersaturation of something like 40 percent would renain. This does
not mean that the oxygen would be present at 140 percent of its solu-
bility in equilibrium with the air, because the gases will not be present
in the ratio found when water is equilibriated with air. The effects
can be estimated.
Suffice it to say that if water has no air in it and a quantity of air
is completely dissolved under a pressure of several atmospheres, or if
air is used in a large excess over the amount that can dissolve, the
release of pressure will leave a solution somewhat supersaturated with
oxygen and nitrogen in approximately the proportions found in air—saturated
water.
-34-
-------
The benefit of the press-ore would appear to be largely nullified by
regassification. However, the pressure has had the useful effect of
forcing the air to dissolve without a large expenditure of energy for
mixing the air with the water.
Whites Point Force Main, discussed in the section on force mains, IB an
illustration. The normal rate of air injection is 0.5 to 0.6 m /min
(20 cfm). Under daytime flow conditions all of this promptly dissolves,
in a region where very little of the potential energy of the compressed
air has been dissipated in turbulence. Host of the energy haa been lost,
in fact, because of the dissolving of the air. About half way up the
pipeline air begins to come out of solution, but by that time a consider-
able amount of oxygen has been consumed biologically. The energy
efficiency for the process is high as a result of this utilization of
the oxygen. High efficiency would be shown in any high-pressure aeration
system in -which the wastewater could be held under pressure for a fairly
long tine. TJnfortunately, there are few places where wastewater can be
retained under high pressure for sufficient time for a substantial
biological oxidation.
If a chamber deep in the earth were used to hold the wastewater, and
there were some way to keep the sludge suspended, then activated sludge
treatment could be accomplished with a very low aeration cost.
Two "-tubes were installed at the ends of force mains in Jefferson
Parish, Louisiana, and four were placed in pump stations at Port Arthur,
Texas. These installations were made possible by research grants from
the TJ. S. Environmental Protection Agency (lO)(ll). The purpose was
sulfide control, but since the U-tubes accomplish this function by
dissolving oxygen, they are of interest as devices that contribute to
BOD reduction in sewers.
In the existing installations, the air enters the descending leg either
-35-
-------
by way of a venturi aspirator or by a diffuser collar. Figure 5 shows
the induction of the air by an aspirator. Insofar as its effect on the
dissolving of oxygen is concerned, the effect of an aspirator is to
break the air up into small bubbles^ It has no unique virtue not shared
by other devices that use high velocity to disperse the air. Air can
also be introduced by simple entrainment if there is a fall from a
sufficient height into the descending leg of a U-tube. At a downward
velocity of 1.5 fps, well dispersed air in amounts up to 10 percent of
the volume of the water may remain dispersed; larger amounts lead to
the formation of an air pocket. In one way or another the turbulence
necessary to cause oxygen to dissolve is reflected in a higher hydraulic
head at the U-tube inlet.
Figure 6 shows a semi-logarithmic plot of the oxygen deficits observed
at various head losses through the aspirator and U-tube derived from a
report on the unit by Mitchell fl2>. The deficits are figured on the basis
of a virtual solubility of 11 mg/1. That figure represents the maxinum
amount of oxygen that would be retained by the water after it returns to
atmospheric pressure.
The points at low head losses are quite scattered, but at the higher
heads can be approximated as a straight line. The dashed line in the
figure is a rough interpretation of what the trend might be for the
results at lower head losses, obtained at lower air input rates.
The slope of the solid line corresponds to a waterfall aeration coeffi-
cient, KJJ, of 0.17 m . For the dashed line the corresponding K^ figure
is 0.47 m'1. Using 11.0 for Cg, the energy efficiencies at zero D.O.
are 0.7 and 1.9 kg of oxygen per kw-hr. These are figures for waste-
water, not clean water, and refer to the energy dissipated in the U-tube
system, not the increment of electrical energy demand.
-36-
-------
EXISTING FORCEMAIN
FIG.5 U-TUBE DESIGN, JEFFERSON PARISH STATION 5
VENTURI ASPIRATOR
'THROAT LD. 6.40"
GRADE EL 43.2
••fWvtfAtt*.
EL32.7_
3t
£
I
«
•\
•>
•i
_f.
i
*•*
v:
•f
1
T
y s
f»_ / J%_ «
-^
EL 0.0
1!
-P
M
^
_
i^
Cc
- 1
— « -
9
W^i'.w*:-
M^-^F|
^
*
*
1
i*
'i
f C
%i'
**
V
Tl *
ri /»*^.« *-f fyi
VAvfV.il
i .^^^^^^^ PI A 9 &
=4>
-------
FIGURE 6, OXYGEN ABSORPTION IN A U-TUBE
10
8
~ 6
o>
E
t 5
o
u.
LU
o
z 4
UJ ^
o
>
X
0
3
2,5
2
K
\\
\\
V v
.\
. \ •
i
\
\
\
K
•\
\
\
,KH=0.47
\
\
\H=
• \
\
0.17
\
A
. •
v
•\
10 oo -j O>cnj^tar\)~c
EFFLUENT DISSOLVED OXYGEN, mg/l
4 6 8 10
HEAD LOSS, METERS
-38-
-------
More data are needed to clarify the reason for the scatter of the points,
and to ascertain the efficiency of the process.
The purpose of the U-tubes in Jefferson Parish and Port Arthur is to
dissolve enough oxygen to accomplish sulfide control. In this they are
successful. The oxygen-dissolving efficiency le not pivotal* More
important in the choice of U-ttibeB where they vere installed was the
convenience and simplicity of the installation under certain prevailing
conditions. It is not likely that U-tubes would be suitable for supplying
successive increments of dissolved oxygen for an extensive in-sewer BOD
reduction project.
A more complete discussion of U-tube aerators is presented in the Sulfide
Control Manual (9)*
If air is injected into the rising leg of a U-tube, it would constitute
an air lift. A zero-head air lift might be constructed alongside a
sewer, having the inlet and the outlet at the same elevation. The water
elevation In the descending leg would be drawn down; water would pour into
it. This would be a eort of diffused air aerator with an arrangement
which in effect provides greater submergence of the diffusers than if
they were placed in the bottom of a sewer.
Air lifts should receive consideration for the pumping of wastewater where
there is a relatively low lift. Air lifts are less efficient than centrif-
ugal pumps. The failure of the expansion energy of the air to accomplish
the equivalent amount of work in lifting the water ie due mostly to the
"slippage" of the air past the water, that is, the tendency of the air to
rise through the water. It is precisely this dissipation of energy by air
rising through the water that causes the dissolving of oxygen. If the
effects of the air lift in dissolving oxygen as well as lifting the water
are considered, the efficiency will look much better than when it is
considered only as a pump.
-39-
-------
The efficiency would be poorer for high lifts, exceeding, say, 10 m,
because the water would be brought near to saturation with oxygen. In
flat terrain, a series of low-lift stations using air lifts would prevent
eulfide conditions and accomplish a significant degree of biological
oxidation. The stations would be less expensive than conventional
stations.
Figure 7 is a functional drawing of an air lift station. Note that the
compressor takes suction from an air by-pass connecting the wet well and
the downstrean sewer. In this way there is no exchange of air with the
outside atmosphere, and no odor release.
INSTALLATION 0? AERATION DEVICES IN SEVSHS
Many steadies of methods for dissolving oxygen in water have been made
over the last half century, since this is a major feature of the activated
sludge process for sewage treatment. The possibility that similar methods
can be used for supplying dissolved, oxygen to sewage in transit crust be
examined.
There are basically two waya of exposing water to air for facilitating
gaseous transfer. One way IB to epray the water into the air, the other
ia to diffuse air into the water. If the objective ie to bring the water
close to equilibrium with the air, that is, close to saturation with
ozygen, or close to zero concentration if the purpose is to strip a gas
from the water, then spraying the water into the air is the most efficient
process. For biological treatment it is not necessary to maintain high
oxygen concentrations. Efficient processes are those in which the trans-
fer of oxygen occurs into water with a fairly high oxygen deficit. Spray-
ing fine droplets into the air would not be efficient from an energy
standpoint, because the droplets would ba brought close to saturation.
-40-
-------
FIGURE 7 i
FUNCTIONAL DRAWING OF AN AIR-LIFT PUMP STATION
TIGHTLY SEALED
WET WELL
AIR INJECTION COLLAR
-41-
-------
Aeration. ...by Diffused Air
Aeration in activated sludge sewage treatment plants ia most commonly
accomplished by diffusing air into the water, using perforated pipes,
ceramic plates or tubes, tubes wrapped with fabric or cord, devices that
force the air out between plates or through springs, arrangements that
aim to produce a high water velocity past orifices from, which air IB
released, devices with moving parts that are intended to aid in mixing
the water with the air, and equipment that channelizes the air-water
mixture in a way that alms to increase or prolong the contact of the air
with the water. If no external energy is applied, the energy for mi^ng
is the potential work of isothermal expansion of the air when it is
released at the bottom of the tank.
This work potential can be calculated by the equation
V = 0.0645 V log P-j/Pj
where V = work, kw—hr
0.064.5 = 0.02815 kv-hr/m3-atm * log 10
e 3
V as volume of sir at one standard atmosphere pressure, m
P. and P = initial and final absolute pressures
Part of the energy of the air goes to circulate the water ness, which is
necessary to keep sludge in suspension, but a major part of it goes to
create the small-scale turbulence that leada to the dissolving of oxygen.
Attempts to channelize the air-water mixture are likely to dissipate some
of the energy in friction, but cannot materially increase the fraction of
the energy dissipated ae small-scale turbulence in the air-water mixture.
Diffusers that produce small "bubbles may lead to the dissolving of 5 to
10 percent of the oxygen in the air used; large-bubble diffusers dissolve
lees. The fine-bubble diffusers, however, require more air pressure to
-42-
-------
drive the air through the diffusers, part of which is due to plain friction.,
but part is also required for the energy expended in producing new Inter-
face against the surface tension of the vater.
Several authors have treated the subject of the efficiency of aeration
by diffused air. A complete review of this subject is beyond the scope
of this report, but a few significant studies may be mentioned as indica-
tion of the sort of results that have "been obtained.
Pasveer (l?)(l4) made an experimental and theoretical study of the energy
efficiency of bubble aeration. The energy was calculated as the displace-
ment energy of the free air volume against a pressure equal to the depth
of submergence plus 10 cm. (For comparison with figures used elsewhere
herein., the displacement work at a pressure equal to 3 m of water is
14 percent greater than the work of isothermal expansion.) He points
out that in actual operation the resistance of the dlffuser may be of
large magnitude. The importance of small bubble size is shown, but figures
that would be applicable to operating sewage treatment plants are not
available.
Bewtra and Nicholas (15) studied the efficiencies of saran tube diffusers
and spargers. Their curves indicate that-at 20 deg C and zero dissolved
oxygen the saran tubes yielded 4,0 kg of dissolved oxygen per kw-hr of
energy of adiabatic compression of air at a pressure equal to the depth
of submergence. At that submergence the work of isothermal compression
or expansion is about 95 percent of the adiabatic compression figure.
If the energy were expressed on the basis of isothermal expansion, the
efficiency would be 4.2 kg of oxygen per kv-hr.
This calculation neglects the energy required to force the air through
the saran tubes to produce the small bubbles. No data were given on the
pressure of the air delivered to the diff users <, At normal depths, 20
percent of the energy may be expended in diffusing the air, even with
clean diffusers, and a much higher proportion when the tubes have been
-43-
-------
In service for some time. Assuming BE. efficiency of 50 percent in con-
verting electrical energy into the energy of compressed air, the amount
of oxygen absorbed may be 1*7 kg per kw-br of electrical energy when the
tubes are nev, and -some smaller amount when they are old.
For spargers the energy efficiency found by Betrtra and Nicholas is
equivalent to 2.3 kg of oxygen per kw-hr of work of isothermal expansion.
These figures are for clean water. There is much uncertainty about the
a factor for diffused air aeration. Sometimes a = 0.8 is used, but the
origin of this figure is not clear. Mancy and Okun (16) showed that
surface active agents have no effect at low Reynolds numbers, but » may
be below 0.4 at higher Reynolds numbers. It has been shown herein that
a is probably between 0.4 and 0.6 for waterfall aeration. It would not
be correct to assume that the same value would apply for diffused air
aeration, but it seeing imprudent to use a value higher than 0.6.
It is concluded that in wastewater aeration, spargers probably do not
dissolve more than 0.6 kg of ©2 per kw-hr of electrical energy used, and
that fine bubble diffusers when new aay dissolve 1.0 kg/kw-hr. These
figures are applicable only at zero D.O.
In an existing sewer, poroua tubes or plates could be placed in the bottom
of a manhole, or a perforated pipe or hose could be extended down the
aewer. Short lengths of metal or plastic pipe could be eucceslvely
coupled and pushed down the sewer, a rather tedious procedure if they
would need to be removed at intervals of, Bay, once a year for maintenance.
Holes in an iron pipe tend to enlarge by corrosion when used for diffusing
air. A perforated hoae or plastic pipe of sufficient flexibility to be
inserted into the sever by way of the manhole would doubtless be more
practical. It would need to be weighted to prevent floating. Diffusera
of thia type are often used for the aeration of sewage lagoons.
-44-
-------
Pipes or hoses running down the sewer would have the advantage of dis-
tributing the air 30 that oxygen would continue to be supplied while the
sewage uses it. However, diffusers located in each manhole, served by a
compressed air pipeline paralleling the sewer, would be able to supply
enough oxygen to meet the demand in the reach to the next manhole, with
only a slight sacrifice in efficiency under most conditions. It is
Judged that diffusers in manholes would be more practical than pipes or
tubes extended down the sewer.
Considering the greater naintenance cost for small-bubble aerators, and
the fact that their efficiency is likely to be less at shallow depths
because of the relatively high pressure loss through the diffusers, it
is concluded that perforated plastic pipes would be the preferred method
if air is to be diffused into wastewater in the bottoms of manholes.
The diffused air method has the limitation that there must be enough water
depth for reasonable submergence of the diffuser. With porous tubes or
other fine-bubble devices in a shallow strean, the pressure loss across
the diffusers is likely to represent a major part of the power requirement,
with the result that power efficiency would be poorer than with deeper
submergence. A larger flow of air through a diffuser with relatively
large holes would not necessarily be more efficient, because some pressure
loss across the orifices is necessary for distribution, else the output
of each orifice will be unduly influenced by small differences in water
depth or the pressure losses in the air distribution system.
The diffuser could lie longitudinally in the sewer, connected to an air
supply line entering at an angle from the side and protected by a fillet
of mortar so that debris would not be caught by it. Just what the
shallowest depth would be for satisfactory performance is not known. It
seems likely that 30 cm (1 ft) would be about the minimum. The water
could be backed up somewhat by a stream-lined dam below the diffuser.
Perhaps a better arrangement would be to place the difuser immediately
-45-
-------
downstream from such, a dam, where high velocity and turbulence would
aid in the dissolving of the oxygen. Toe dam would be a concrete block
similar in shape to a Palmer-Bowlus slab-type weir (17) but with a
blunt downstream face, or an undercut face, with the diffuser in a
crosswise position. Figure 8 shows ways that this could be done.
Surface Aerators
Rather than to blow air through the water, a dispersion of fine bubbles
can be produced by devices that beat the surface of the water. Kessener
(18) developed the brush aerator in the 1920'g. It is a device turning
on a horizontal axis, with metal projections dipping into the water.
Pasveer (19) made careful measurements of efficiency, and optimized
physical features of the installation, finally attaining an efficiency of
2.5 kg of 0, per net kw-hr. This was at zero D.O. and 10 deg C. On the
basis of assumptions made by Pasveer, the efficiency at 20 deg C would be
2.95 kg per net kw-hr, or 2.44 kg/kw-hr of applied electrical energy.
With less than optimum efficiency of motor and speed reducer, the
efficiency might be as low as 2.0 kg/kw-hr. In wastewater it might be
1.0 to 1.2 kg of oxygen per kw-hr of electrical energy.
The greatest use of brush aeration in recent years in the United States
has been in "racetrack" sewage treatment units, in which the aerator
serves the additional function of moving the water around the racetrack
channel.
Another type of surface aerator now much used in the United States is a
device rotating on a vertical shaft, with blades that stir the water and
throw it outward as a heavy spray. For activated sludge plants, the
designs take into account the need to expend part of the power to keep
the aeration basin stirred. A draft tube under the aerator often serves
this purpose. In a sewer, the device would need only to stir the surface.
-46-
-------
FIGURE 8
POSSIBLE WAYS TO DIFFUSE AIR INTO A SEWAGE STREAM
y
/ /
DIFFUSER
SECTION
DIFFUSER
PLAN
OPTION - A-
/// /
/ /////// / // /
. ,^ .
f^- V / " .'-.^—i-'»:-v'/M'-'-*'*"'"' '
'.' -"'^
./•
AIR SUPPLY SECTION
OPTION - B-
-A7-
-------
The installation of mechanical aerators in an existing sever would be
difficult under most circumstances. Horizontal-shaft aerators would face
the severe handicaps of limited space and of maintenance problems of the
moving parts suaberged in sewage. Vertical-shaft as well as horizontal-
shaft aerators would need to be designed with consideration for varying
flow levels, 1 streamlined weir could be used to reduce the rise and fall
of the water, and might under some circumstances serve to stabilize the
water elevation sufficiently for reasonably efficient operation. There is
the possibility of using pontoons to float the aerator, or perhaps to float
the rotor, with a telescoping shaft to a fixed motor.
The sort of special structures that would be needed for installation of
conventional mechanical aeration devices are likely to be unecononical
because of the limited amount of oxygen that each such device could add to
the stream. Sufficient oxidation to justify the cost of a special struc-
ture would require the impounding of the sewage for a time, in which
case the facility would be classed as an upstrean plant rather than in-
line treatment.
_Jet Aeration
Another method for mixing air with water that could be used in a sewer is
jet aeration. A stream of water under pressure is jetted by way of a noz-
zle into a larger stream or "body of water, whereby air is entrained and
"broken up into fine bubbles. Only a plain nozzle is here considered.
Figure 9 illustrates a jet aeration installation serving to treat an
industrial wastewater. Water from the pond is pumped into the header,
which delivers it to sixteen nozzles injecting it back into the pond. The
nozzles are 2.5 cm in diameter, operating under a head of about 18 m. At
the time that the photograph was taken, the water level was below the de-
sign level. This causes rather too much fraying of the jet before it
reaches the water surface; however, it should be moderately frayed for
-48-
-------
FIGURE 9
A JET AERATION INSTALLATION
-49-
-------
entrairunent of air that is then driven into the pond. Around.
the jet the water appears milky from the fine air bubbles so produced.
A few tests of jet aeration were made with the apparatus shown in Figure
3- Results ere given in Tatle 9.
Table 9. HP1RIMEKTS VITH NOZZLE AERATION
Independent variables
now
rate
1/ni1
Results
42
42
55
1C2
12
22
.Results
42
55
Depth
of
, water,
n cm
using clean
76
76
76
76
76
81
usinc sewaee
76
76
Height
of
fall,
CE
water
10
20
20
20
20
20
20
20
Noz-
zle
dia,
cm
1.12
1.12
1.12
1-91
0.46
0.46
1.12
1.12
Pres-
sure
gage,
osie
4.5
4.5
7.25
2.90
9.4
21.25
5
7.8
Energy
of the
falling
water,
kw
0.025
0.0258
0.0489
0.046
0.0142
0.0542
C.0269
0.0531
Oxygen
trans-
ferred,
ke/hr
0.0257
0.0360
0.0617
0.0649
0.0269
0.0291
0.0243
0.0550
Oxygen
transfer
effi-
ciency,
ksr/kv-hr
1.03
1.40
1.26
1.41
1.90
0.54
0.90
1.04
The average
efficiency in clean water was 1,26 kg of oxygen per kw-ht of potential plus
kinetic energy in the jet stream, or somewhat less than half the value
found for waterfalls. It is possible that better results would be
obtained under optimized conditions. The beat result found was 1.9 kg
per kw-hr.
Even if the energy efficiency is not high, the process may sometimes be
useful because of the relatively simple installation, especially if water
under pressure is available, as at a puap station. If used for in-sewer
aeration, a supply of water under high pressure could be carried by a
-50-
-------
pipe paralleling the eewer, with branches leading to nozzles at suit-
able spacinga, not necessarily at nanholes. If vastewater from the
sewers is used for toe jets, it would tie necessary for it to be ecreened
to avoid clogging of the nozzles.
Venturi Aspirators
The high shear in the throat of a venturi that is aspirating air causes
the air to te broken up into small bubbles. The device has no unique
oxygen-dissolving potential not accounted for by the high velocity, but
sometimes it has the convenient characteristic of being able to function
also as a pump. It has been used for the purpose of dispersing air in
flotation separators, and its application ahead of a U-tube is illustrated
in Figure 5-
A venturi could be snbnerged in the sewage stream with a small pipe to
bring air to it froE above the water surface. Probably a better plan
would be to locate the venturi above the water, perhaps at a considerable
distance, with the air-in-water dispersion carried by pipeline into the
wastewater stream. Screened vastewater could be pumped to operate the
device.
The energy efficiency of a ventori as an oxygen-dissolving device is not
known.
Comparisons andPractical Instaliations for In-Sewer Aeration
As already indicated, surface aerators do not look promising as devices
to install in sewers. Since only a few mg/1 of oxygen could be added
to the wastewater stream at one time it would be necessary to install
many of then to obtain a substantial effect. Each installation would be
quite expensive.
-51-
-------
Aeration by simple jets is a process that deserves further investiga-
tion. An experimental installation should be made to secure the accessary
information. Judgment about the applicability of the process for use in
sewers needs to be deferred until more is known about the operating
characteristics,
Unless Venturis show sone advantages not presently recognized, they
appear unlikely to be favorably considered.
It appears from these considerations that if an aeration facility were
to be installed in an existing sewer it probably would be of the diffused
air type, using a perforated plastic pipe for the diffuser.
-52-
-------
THE USE OF REFINED OXYGEN
Principles of Oxygen Application
The methods considered thus far for augmenting the dissolved oxygen supply
of a sewage stream depend upon contacting the wastewater with air, whereby
oxygen is extracted from the air by the oxygen-deficient water. The use of
refined oxygen is another approach to the same objectives, but it involves
technologies which in some respects are quite different, and in some cases
it permits the attainment of results not possible by the use of air.
The term "refined oxygen" is here used to mean oxygen extracted from air
either by liquefaction and fractional distillation or by the molecular sieve
method. The distillation method is used where the quantities produced are
greater than about 25 to 50 tons per day. The purity is generally about 98
percent. It is generally produced as a gas and used near the site of the
refining facility. Alternatively it may be delivered as a liquid. It is
much more expensive in this form. The molecular sieve method does not
require liquefaction. This is the more practical method where the amounts
produced are too small for the economical installation and operation of
liquefaction equipment. The purity of the oxygen produced by this method
is generally 85 to 90 percent. It is either used at the site where it is
produced or delivered by pipeline in the form of a gas.
The relative costs of dissolved oxygen from air and from refined oxygen vary
widely. The principal cost for aeration is power. Considering basic energy
efficiencies of aeration processes, mechanical and electrical losses, the
a factor, and the fact that the oxygen dissolves into water containing some
dissolved oxygen rather than a solution devoid of oxygen, it seems probable
that most functioning systems are dissolving from 0.3 to 1.0 kg of oxygen
per kw-hr of electrical energy purchased. Prices of electrical energy vary
with location and demand, so the cost of electric power per kg of dissolved
oxygen is quite variable. Capital costs are relatively high, and operating
costs other than power are a significant factor.
-53-
-------
The prices for refined oxygen, too, vary over a wide range, depending on
location and demand as well as electric power costs. For quite large
demand rates the cost is minimized by the Installation of oxygen-producing
plants. The equipment and maintenance costs for in-sever application of
oxygen are lower than for attaining the same results by aeration.
In moat situations the over-all cost for using refined oxygen is higher
than for aeration. Oxygen has advantages, however, that is some situa-
tions may outweigh cost. This is especially true when the use is for
in-sewer treatment. The nitrogen in air interferes with the dissolving
of really large amounts of oxygen. When refined oxygen is used, much
higher concentrations can be attained. A point application of oxygen
in a sewerage system may suffice to keep the stream aerobic for one to
five miles downstrean.
The dissolved nitrogen content of sewage must be taken into account,
however, even when pure oxygen is used. It can be assumed that sewage
will usually contain dissolved nitrogen in an amount approximately in
equilibrium with the atmosphere, which means that its vapor pressure
will be about 0.79 atmosphere. If pure oxygen were added only in the
amount that would completely dissolve at equilibrium at atmospheric
pressure, the concentration would be only 8 to 9 mg/1 at ordinary sewage
temperatures, that is, it would be the same as for water in equilibrium
with the air. If a larger amount is added, some will remain undissolved,
and some nitrogen will come out of solution, so that residual gas will be
a mixture of oxygen and nitrogen. If higher concentrations in solution
are desired, nitrogen-diluted oxygen must be wasted, in amounts increas-
ing as the dissolved oxygen concentration is increased. Even though the
solubility of oxygen is about 40 mg/1, this concentration cannot be
approached in a one-step operation without wasting a substantial part of
the oxygen. For example, to obtain a concentration of 30 mg/1 at 25eC
would require the wasting of 29 percent of the applied oxygen.
-54-
-------
Appendix II presents equations for calculating the amounts of oxygen that
must be applied In a one-step operation to obtain various concentrations.
Figure 10 shows results of the calculations for 25°C and pressures of 1.0,
1.3 and 1,6 atmospheres. The higher two pressures correspond to sub-
mergences in about 3 and 6 meters of water.
In an actual case the oxygen will be somewhat less than 100 percent purity,
and the amount dissolved will accordingly be less.
If the oxygen is in contact'with the liquid for a prolonged time and
oxygen Is consumed biologically during that period, as in an activated
sludge plant using oxygen, the efficiency is of course much better.
High oxygen concentrations may be reached by dissolving oxygen under
higher pressures. Table 10 shows the solubility of pure oxygen at various
temperatures and pressures.
Table 10. SOLUBILITY OF OXYGEN IN PURE WATER, mg/1
Temp,
deg
C
10
15
20
25
30
35
1 atm
54
48
A3
39
36
33
2 atm
107
96
87
79
72
66
5 atm
268
240
217
197
179
166
10 atm
537
480
434
393
359
332
20 atm
1074
960
868
786
718
663
In water that contains the normal amount of nitrogen, the solubility of
oxygen will be reduced, but this effect is negligible at high pressures.
The impurities in sewage, other than nitrogen, do not have any significant
effect on the solubility, although they affect the rate of dissolving.
-55-
-------
FIGURE 10
CALCULATIONS OF ABSORPTION OF PURE OXYGEN
IN WASTEWATER DEVOID OF OXYGEN
10
20 30 40
Oxygen, applied, mg/1 CA)
50'
60
-56-
-------
If water that has teen charged with oxygen at high pressure is then
returned to atmospheric pressure, the water will be supersaturated. If
the supersaturation is great enough, there will be effervesence, that is,
spontaneous formation of bubbles. The tendency to effervesce is increased
by turbulence. There is not nuch practical information about oxygen losses
by effervesence, but from such evidence as is available it seems likely
that where the vapor pressures of the gases total 1.5 atmospheres, there
will not be excessive effervesence. This would be 0.8 atmosphere of
nitrogen and 0.7 atmosphere of oxygen, corresponding to about 30 mg/1 of
dissolved oxygen. There probably will not be very much loss even at 40 mg/1
of dissolved oxygen unless there is considerable turbulence. There will,
of course, be loss by transfer to the atmosphere across the phase boundary.
The saiae equation that serves to calculate oxygen absorption when there is
an oxygen deficit can be applied to the transfer in the opposite direction
when there is an oxygen surplus. The loss by this mechanism will be quite
small in large trunks.
Techniques for Using Oxygen in Sewers
One idea for using oxygen is to enrich the sewer atmosphere, so as to
increase surface oxygen absorption. It would be difficult to avoid loss
of part of the enriched atnosphere to the outside or by ingress of outside
air or the downstream air draft. It would be particularly difficult to
avoid losses froi small sewers with lateral connections, and in large trunks
the rate of aeration at the surface of the stream is so small that there
would not be very much benefit even fron an atmosphere of pure oxygen.
Another possibility is to follow the example of Unox sewage treatment plants.
The oxygen could be bubbled through the stream in the bottom of a manhole,
and then the enriched atmosphere could be used to aerate the water at the
next manhole downstream, and so on down the sewer until most of the oxygen
is used. This might be a feasible method under favorable circumstances, but
in most situations it would be difficult to avoid unacceptable oxygen losses.
-57-
-------
Oxygen Injection Into Pressure Mains
This is a method already in extensive use in the United Kingdom. A report
on one such installation at Stevenage was presented by Boon and Lister
(paper presented Sept. 13, 1974, at Seventh Conference of the International
Association of Water Pollution Research at Paris).
When air is injected into a force main, it is usually applied essentially
continuously, because the water in the main must be supplied with dissolved
oxygen by a near-continuous stream of air bubbles moving up the pipe, but
when oxygen is used it can generally be dissolved completely in sufficient
amount to last for the length of time that the water spends in the pipe.
In the Stevenage installation the oxygen is injected into the suction of
the pump while the pump is running, so that all of the wastewater receives
the same oxygen dosage.
There no doubt will be places where not enough oxygen can be supplied in
this way. In these cases a stream of oxygen may need to be allowed to
bubble up the pipe, or it may be injected at more than one point.
The dosage of oxygen applied in the Stevenage case is 75 mg/1. Of this
amount, an average of 59 mg/1 react, leaving 16 mg/1 in solution at the
end of the main. At the prevailing temperature of the vastewater, 15°C,
this much oxygen has a partial pressure of 0.33 atm. With 0.79 atm
vapor pressure of nitrogen, this gives a total of 1.12 atm, which is not
enough to cause appreciable loss. At times, however, the concentration
reaches 36 mg/1. Some may be lost at this concentration, since the sum of
the partial pressures is 1.46 atm.
Very long pressure mains, with retention times of many hours, afford the
possibility of a rather high degree of biological oxidation. It would be
difficult in such cases to supply the oxygen requirement by air injection,
but it could be done easily by repeated injections of pure oxygen. Con-
sideration may need to be given to the possible effects of low pH due to
retention of C02, but it seems unlikely that there would be any serious
problem on this account.
_58-
-------
Oxygen does not automatically go into solution just because it is held
under pressure in contact with the water. Mixing energy is required. The
slope of the pipe is therefore a critical factor. Oxygen injection into
force mains with little slope has, in two cases, failed even to provide
sulfide control.
Oxygen Injection Into U-tubes
In Port Arthur, Texas, oxygen is injected into two U-tubes. They are only
5 E deep, and part of the oxygen escapes unused. The method should be
very efficient in a deep U-tube, where the pressure will force the oxygen
into solution rapidly.
The required depth to dissolve any given amount of oxygen cannot be
stated at this time. Experimental studies are needed. Certainly the
pressure needs to be greater than the equilibrium pressure for the amount
of oxygen to be dissolved, and if the objective is to dissolve the amount
of oxygen that can be carried without loss by effervesence at the U-tube
discharge, the equilibrium pressure would be only 1.3 to 1.4 atm absolute,
which would be at a depth of 3 m to 4 m. However, the U-tube will
need to be considerably deeper to assure the complete dissolving of
the applied oxygen.
Release of a Supersaturated Oxygeri Solution Into the Stream
Oxygen can be dissolved readily at high pressure to produce a solution
containing oxygen at concentrations up to several hundred mg/1. Oxygen
-59-
-------
vould start to come out of solution upon any reduction of pressure.
However, if pressure reduction occurs only at a nozzle submerged in a
body of water, the jetted stream draws a large volume of water into
itself, diluting the oxygen concentration until the mixture is no longer
supersaturated. If any bubbles form, it is only momentarily, and they
are so minute that they quickly dissolve.
Figure 11 is a photograph of an installation of this type constructed in
1970 to provide an oxygen supplement for an overloaded activated sludge
plant. The tank, for liquid oxygen is in the background. In the foreground
ia the oxygen dissolving tank, shown in greater detail in Figure 12.
Oxygen is released under pressure into the tank, forming a gas pocket in
the top. Wastewater is pumped into the top of the tank, jetting downward
so as to dissolve oxygen. The water level seeks an equilibrium point
where the oxygen is dissolved as fast as it is fed. Any amount of oxygen
can be fed, up to a maximum dissolving rate fixed by the pressure and the
water flow. The high-pressure solution is conveyed to the bottom of the
aeration tank, where it is released by way of a nozzle. No release of
oxygen bubbles is observed in the area where the oxygen solution is
injected. The nozzle and pump must be matched for efficient operation in
an installation of this type,
A similar system could be applied for the supplying of oxygen to the
stream in the sewer. Screened sewage or any other available water could
be used for preparing the oxygen solution. Figure 13 shows how the
process might be applied in a sewer. A pressure pipeline might carry the
oxygen solution to a number of points of application. If 20 to 30 mg/1
of oxygen are supplied at each point, it would not require very many
such applications to support a substantial degree of in-sewer BOD
reduction.
-60-
-------
FIGURE II, INSTALLATION TO PROVIDE OXYGEN SUPPLEMENT
FOR OVERLOADED ACTIVATED SLUDGE PLANT
PRESSURIZED
WATER-
FIGURE 12-
GREATER DETAIL OF
OXYGEN-DISSOLVING TANK
-61-
-------
FIGURE 13.
TREATING THE STREW IN A SEWER WITH OXYGEN
TRUNK SEWER
SCREEN
= SUBMERGED
L_ NOZZLES
OXYGEN
DISSOLVING
FACILITY
-CONCENTRATED
OXYGEN SOLUTION
UNDER PRESSURE
CONTROL VALVE
OXYGEN TANK
-62-
-------
CHAPTER 5
THE OXIDIZIUS CULTURE FOR BIOLOGICAL OXIDATION IN SEWERS
An ample supply of dissolved oxygen, which can be assured by methods
described in Chapter 4, will not result in biological purification if
there is not an adequate culture of aerobic microbes. Neither will
a culture of microbes accomplish the desired result if they are not
supplied with oxygen. This chapter discusses the ways that an oxidizing
culture can be provided and indicates the sort of information that is
needed for estimating the oxygen requirement.
A biological culture capable of purifying sewage may exist either as a
suspension of microbes in the wastewater or aa a biological slime on some
supporting nediun. The two ways in which the oxidizing culture can exist
are the bases of the two principal types of secondary sewage treatment
plants, that is, activated sludge and trickling filters. Both suspended
and attached cultures may be involved in a single system of ia-sewer
biological treatment, but in most applications one type or the other will
predominate.
TEE INDIGENOUS BIOLOGICAL FLOC
A report on "Self-purification in Sewers" (Appendix I of this report)
encompasses most of what is known about aerobic processes in sewers. The
experimental and deductive work for that paper was done largely as a
part of the preparation of this report, and, in a lesser degree, as an
element of the study of sulfide control in sewers, also funded by the
U. S. Environmental Protection Agency. It was shown therein that sewage
near the source does not have a high degree of biological activity, as
measured by the rate at which it can use oxygen. By the time wastewater
from residences reaches the public sewer, it is seeded with a great array
of microbes, but seeding alone is not sufficient. For aa effective rate
-63-
-------
of 'biochemical change, there must be a high population density of species
adapted, to that environment.
If sewage is kept supplied with oxygen, suitable species will proliferate
and biological oxidation will accelerate. This change is illustrated in
Curves A and B of Figure 6 of Appendix I, To obtain a curve such as those
illustrated, a sanple of sewage is kept aerated in a bucket, and fron.
time to time a portion is taken out, tested for oxygen reaction rate, and
then returned to the bucket. The testing is done by holding the sub-sample
in a vessel where it is stirred out of contact with the air, and oxygen
concentration is monitored by use of an oxygen probe. Each point in
Figure 6 of Appendix I is the result of one sucn oxygen reaction rate
determination. The results are expressed as milligrams per liter-hour.
The culture obtained by aerating the sewage in a bucket is a non-flocculent,
non-settling suspension.
In a sewer, the oxygen reaction rate increases more rapidly than it does
when the sewage is aerated in a bucket. Generally the suspended solids
increase, the increase being in the flocculent, settleable portion. These
changes reflect the sloughing of biologically active slimes from the pipe
wall, providing a material resembling activated sludge.
If the stream of sewage remains fairly well aerated, quite high reaction
rates are reached. The most notable example of this is in the South Slope Trunk
of the Los Angeles County Sanitation Districts. The trunk receives well
aerated sewage fron tributary eewers coming off the San Pedro Hills. The
trun£ is not particularly well aerated; in fact, some sulfide is produced,
but a sufficient degree of biological activity is maintained so that the
flow reaches the Whites Point puiap station, after ac average transit time
of three hours in the trunk, with an oxygen reaction rate as high as
10 mg/l-hr in the winter when the sewage temperature averages 19 deg C,
and rises to between 20 and 25 mg/l-hr in the summer when the sewage
-64-
-------
temperature IB 26 to 27 deg C. The sewage then spends about 50 minutes
in the aerated Whites Point force main, where it oses about 16 mg/1 of
oxygen. Altogether, it -was estimated (Appendix l) that from the time of
its entrance into the South Slope trunk until its discharge from the
force main, the sewage has used 60 to 75 mg/1 of oxygen. This 10 in
addition to substantial oxidation presumably occurring in. the steep
tributary lines.
Curves D, E and P were made with sewages from major trunks tributary to
the Los Angeles City Hyperion Plant and the Joint Water Pollution Control
Plant of the Los Angeles County Sanitation Districts. These sewages had
spent about ten hours in transit. A comparison of their behavior with
that of fresh samples shows that there were major differences in their
biological condition. Samples taken near the source and aerated in the
laboratory never reached a condition reseabling the sewage from the
large trunks. Furthermore, a sample of fresh sewage held overnight out
of contact with the air did not develop a reaction rate naterially
different from the fresh sewage.
The difference between fresh sewage and that reaching the major treatment
plants is not easily explained In simple chemical terms. Sulfide con-
centrations were not over one or two mg/1, and thus could not account
for the rapid oxygen demand. Organic acid concentrations are higher in
the old sewage, and it was at first supposed that this accounted for the
high peak of oxygen demand and the sudden decline. The decline could be
due to the exhaustion of a nutrient. It was found that addition of sodium
acetate did indeed extend the rising curve, but the drop occurred long
before the acetate was exhausted. It appears that the factors affecting
the biological balance are highly complex, well deserving of further
study, but too obscure at this time for fruitful speculation,
It is evident that the slime layer is a na;or source of an oxidizing
culture suspended in the stream. There is no way to predict just how high
the oxygen consumption rate might rise in a large sewerage systeic if it
-65-
-------
were not constrained by oxygen starvation, because there has been no
opportunity to observe a large system where oxygen impoverishment does
not prevail. If adequate oxygen is supplied, and even if the reaction
rate rises only as high as observed in the South Slope Trunk (Curve C of
Figure 6 of Appendix l), a rather large degree of oxidation will occur.
EXTERNALLY GENERATED BIOLOGICAL FLOG
Rather than relying on the sloughing of biological sliae from the pipe
wall to provide a suspended aerobic culture, activated sludge or trick-
lie^ filter huEus may be added. This may be feasible where an upstream
treatment plant withdraws E portion of the flow from a trunk sewer, passes
it through a secondary treatment plant to produce an effluent for local
disposal, reuse, or ground-water recharge, and returns the waste sludge
to the sewer. There are many upstream plants operating in this mode,
but in no case has oxygen been added downstream for the purpose of
stimulating in-sewer oxidation.
The capability of an upstream plant to generate sludge for in-line.
biological treatment is limited. If only a small fraction of the sewage
flow ie taken out and treated, the amount of activated sludge produced is
not great enough to provide much enrichment of the culture in the seirerD
If most of the flow is taken out, treatment of the residual flow in the
sewer might be a minor adjunct to the main operation. If the flow taken
out and treated upstream were equal to the flow that might be subjected
to in-line treatment at some downstream point, the return sludge would be
a substantial factor, even though the sludge concentration would be quite
low in the comparison with the usual standards for mixed liquors in
activated sludge tanks.
An alternative way to supply an oxidizing culture is to pump activated
sludge from a downstrean plant to a point far up in the system by way of
an aerated pressure main^ This pressure main would be an elongated
"reaeration tank" of an activated sludge system. Ths downhill gravity
-66-
-------
sewer, if it were kept aerobic, could then function as the nixed liquor
tank.
A variation of this plan is to pump the entire flow in an aerated pressure
main that would function as the mixed liquor tank. A final clarifier at
the end of this pipeline would separate the sludge and return most of it
to the sewer. This is a practical plan if there is an upstream point of
use for the treated water. The Irvine Ranch Water District in Orange County,
California, has been developing a project based upon this concept.
BIOLOGICAL CULTURE ON SOLID SURFACES
It has been shown in previous sections that the slime layer is an important
source of an oxidizing culture in the stream. Under some conditions
oxidation on the slime layer in place on the wall of the pipe is a major
factor in the total process. The slime layer may be aerobic, anaerobic,
or a mixture, depending upon the condition of the sewage. The outer zone
of the slime layer, exposed to the sewage flow, contains aerobic organisms
if there is even an intermittent exposure to dissolved oxygen.
Measurements have been made, principally in the course of preparing this
report, of the rates at which the slime layer in sewers uses oxygen.
Results are given in Appendix I. It was concluded that the slime layer
frequently acts as an effective sink for all oxygen reaching it, in which
case the rate of reaction is equal to the rate at which oxygen can be
carried to the wall by the motion of the water.
The sline layer grown in a steam that is slow-moving but aerobic will
attain considerable thickness, with long filaments trailing in the stream.
The slime becomes progressively thinner as the velocity increases, but it
was shown that it may still be a highly active layer. Even, a slime layer
so thin as to be unnoticed by casual visual inspection, in a sewer with
a velocity as high as 2 ra/sec (7 ft/sec), was able to use oxygen as fast
as the stream supplied it.
-67-
-------
Reaction of oxygen on the slime layer is predictable on the assumption
that the slime is a sink for all oxygen reaching it. Where this condition
prevails, the flux of oxygen to the pipe wall is proportional to its
concentration, as shown by the following equation:
where
o
0 = oxygen flux to pipe wall, g/m -hr
e = submerged-surface transfer coefficient, m/hr, defined by
the equation
[0_] = dissolved oxygen concentration, mg/1 or g/m
The rate of change of oxygen concentration due to the flux to the wall,
designated R , is equal to the flux divided by the hydraulic radius:
Re = ^r'1 - e[02]r-1
where
R = rate of change of oxygen concentration due to its utilization
by the slime layer on the pipe wall
r = hydraulic radius (1/4 of pipe diameter in a filled pipe), meters
The proportional rate of loss of oxygen to the wall, designated the mural
deaeration coefficient, is represented by the symbol K . It is related
to the other quantities by the equation:
Ke - Re r°2rl = e t"1
where
K = mural deaeration coefficient, hr
The following empirical equation has been presented for estimating e
(Appendix I):
e = 5.3(su)1/2
-68-
-------
where
s = slope of the energy line of the stream
u = velocity, m/sec
Hence
„ , , . .1/2 -1
K =5.3 (su) r
The number of variables in this equation can be reduced by combining it
with an empirical equation relating s, u and r. Best for this purpose,
under turbulent flow conditions, Is the Hazen-Wi Hiatus equation, which
can be written:
n „,_ „ 0.54 0.63
u = 0.850 C s r
The Hazen-Williams coefficient for the pipe filled with sewage will be
assumed to be 125.
u- 101.2 s°-54r°-63
2 a0'93 B°-93 0.01397 u°-93
(101.2r0-63)0-92571.6r°-58~ r°'58
(su)1/2 = 0.0140 u1'43 r-°-58
- .,. 1.43 -0.58 ,,
e. = 0.074 u r , m/hr
. n., 1.43 -1.58 . -1
K = 0.074 u r , hr
R = 0.074 u1'43 r~1<5B[02], mg/l-hr
Figures 14A and 143 provide graphical solutions to the equation for K .
-69-
-------
FIGURE 14A GRAPH FOR DETERMINING Ke (HIGH VALUES)
FLOW VELOCITY, f p.s
9 r
o>
in
UJ
x
o
V)
cr
m
»-
LU
UJ
0.
FLOW VELOCITY, m/sec.
-70-
-------
FIGURE KB GRAPH FOR DETERMINING Ke (LOW VALUES)
FLOW VELOCITY, f.p.s
I 23456
o
I
1.4
1.2
1.0
0.8
0.6
04
0.2
I
FLOW VELOCITY, m/sec
-71-
-------
The. proportional rate of loas of oxygen per meter of distance, designated
K ', is equal to K divided by the velocity.
., , 0.074 0.43 -1.5S . onr,n,t 0.43 -1.58
K = ->cnn u r = 0.000026 u r
e Jbuu
where
3600 = conversion factor from seconds to hours
This equation shows that in a pipe of given length the stream will lose
more oxygen to the pipe wall when It is moving at a high velocity than
at a low velocity, despite the shorter transit time. However, because
of the shorter time the amount used by reaction in the stream will be less.
In a reach in which there is no oxygen input, as in a pressure main, the
rate of change of oxygen concentration is given by the expression
d[0 ]
-- - - = R + R
dt re
vhere
0l
- decline of oxygen concentration, mg/l-hr
dt
R m rate at which oxygen is depleted by reaction in the stream
R = rate at which oxygen is depleted by reaction on the pipe wall
Assuming R to be constant, the equation can be integrated, with the
following result:
2-303 . A + V°Z]_1
vhere
t and t. = times in. hours
[0-L and [02l2 = correBPon^inE oxygen concentrations
General curves can be drawn froE this equation for any combination of R
and R values, t^jli "-^ ^e given any appropriately high value. For
-72-
-------
any lower value of IO-_] the curve la the. same from the point representing
that initial concentration.
EXAMPLE A
Sewage will be pumped at a velocity of 0,61 m/sec (2 fps) through a force
main 1.83 m (5 ft) in diameter. The oxygen reaction rate of the sewage
is assumed to be 11 mg/1. Assume that the initial oxygen concentration
is 40 mg/1, calculate the times when it will be 30 ng/1, 20 nig/1, 15 mg/1,
10 mg/1, 5 mg/1, and zero.
EXAMPLE B
Sewage of the same composition as in Example A will be pumped through a
pressure main 0.305 m (1 ft) in diameter at a velocity of 1.28 m/sec
(4.2 fps). Calculate the tines required to reach the same oxygen concen-
trations as in Example A.
From Figures 14 A and B, R is 0.165 in Example A, and 5.7 in Example B.
Figure 15 is a graphical representation of the solutions. In Example A,
the initial R value is 6.6 mg/l-hr and (-d[0 ])-?(dt) is 6.6 + 11.0 =
17.6 mg/l-hr. Thirty-seven percent of the initial oxygen consumption rate
is due to reaction on the slime layer. In Example B the initial rate of
oxygen consumption is calculated to be 239 mg/l-hr, of which 95 percent
is due to usage by the slime layer.
The prediction of such a high rate of oxygen usage by the slime layer must
be treated with
by the equation
be treated with caution. The flux of oxygen to the pipe wall, 0 , is given
In Example B the initial flux is calculated to be
0e = 0.0762 x 5.7 x 40 = 17.4 mg/m2-hr
-73-
-------
FIGURE 15 SOLUTIONS FOR EXAMPLES ABB
I 2
TIME, HOURS
CALCULATED OXYGEN DEPLETION CURVES
-74-
-------
The highest flux that has been measured in a sewer is 2.1 mg/m -hr
CAppendtx I). Since that was in a sewer in which the velocity was
2.16 m/sec, it appears that velocity is not a limiting factor under the
conditions assumed in the examples. It is possible that the flux might
be limited by the metabolic capability of the thin biological layer that
would be supplied by oxygen diffusing into it. It seems quite likely
that flux rates as high as 20 g/m -hr could be sustained,.but it is a
question that can be answered with assurance only by experimentation.
A further limitation of the flux may be the supply of oxygen-demanding
nutrients. If these nutrients do not reach the slime layer at a rate
matching the oxygen supply, then it will be these nutrients rather than
the oxygen supply that will control the rate of oxygen usage. All things
considered, it seems likely that oxygen consumption in a small pipe with
fairly high velocity and high oxygen concentration may be sufficient to
accomplish the equivalent of secondary biological treatment in a period
of an hour or two,
In large trunks, longer times will be necessary. In some regional systems
being planned, the times may be long enough to carry the oxidation well
along toward completion, but in most cases it will be a matter of partial
oxidation, reducing the load on the sewage treatment plant. The equations
that have been presented will suffice for the development of designs that
will optimize the contribution that major sewers will make toward the
final treatment objectives.
INCREASING THE SLIME-SUPPORTING SURFACE
Since all surfaces submerged in the stream will accumulate slime, the
question naturally arises as to the feasibility of increasing the surface
area, thus augmenting the "trickling filter effect." There are several
ways that this could be done but all of them present practical difficulties.
-75-
-------
The evident greater efficiency of oxidation gained in smaller pipes
suggests minimizing pipe diameter. If there were two practical pipe
sizes for a force main, roughly equal in respect to all other determining
factors, or if the alternatives of one larger pipe vs. two snaller ones
were fairly well balanced, and if it^were considered that in-sewer
BOD reduction would be a desirable objective, then the smaller pipes
would be chosen. Such situations will be rare.
Slime-supporting area could be increased by filling one large pipe with a
number, perhaps 3, 4 or 7, of smaller pipes of light-weight plastic.
Seven small pipes would increase the slime area more than fivefold, but
at a high installation cost and with a tripling of the friction head.
A gravity (partly-filled) sewer might have the lower half lined with a
plastic sheet having vertical vanes projecting into the stream, aligned
parallel to the axis of the pipe. This arrangement, too, would be high
in cost in comparison with any benefits gained.
-76-
-------
CHAPTER 6
CHEMICAL TREATMENT
POTENTIAL APPLICATIONS
Applications of chemicals to sewage streams represent the oldest form of
in-line sewage treatment. The principal objective of these treatments
has been for the control of sulfide, although various other benefits have
been claimed, such as removal of grease, keeping the sewer clean, BOD
reduction, etc. The chemical most widely used in sewers has been chlorine,
applied since the 1920's. It destroys any sulfide and mercaptans that may
be present, depresses oxygen consumption so that an oxygen reserve accumu-
lates in the stream, and, in large doses', may inactivate sulfide-producing
slimes. Iron salts, zinc salts, calcium or sodium nitrate, and shock
doses of calcium or sodium hydroxide have also been successfully applied.
Hydrogen peroxide has been introduced recently for sulfide control, pro-
viding a source of oxygen, expensive but sometimes favored because of the
simplicity of application. Not so successful have been chlorinated
hydrocarbons, chlorinated phenols, ozone, and various "patent medicine"
products of unrevealed composition. Sulfide control, however, is not
the subject of this report. Chlorine presumably does reduce the chemical
oxygen demand by the oxygen equivalent of the chlorine added, which,
being only 0.225 of the chlorine dosage, generally does not amount to much.
Peroxide accomplishes biological oxidation to the extent of the oxygen
released, which is 47 percent of the weight of H^O-. Oxygen itself,
applied either as air or as refined oxygen, is an essential element of
biological treatment, but that subject has been covered and is not classed
as cneaiical treatment.
The principal types of chemical treatment that need to be looked into for
possible in-sewer use are those that produce precipitates capable of
entraining or adsorbing or otherwise precipitating impurities from the
water. In both water and wastewater treatment plants use is made of lime,
aluminum sulfate, and long-chain organic polymers or polymerized silicic
-77-
-------
acid (activated silica). The polymers are used nostly along with
aluminum sulfate as coagulant aids, but Bometiines acting alone as coagu-
lants. Iron salts are sometimes chosen in place of aluminum sulfate for
treatment of wastewaters. The iron and aluminum salts behave similarly,
and do not need to be treated separately for the purposes of this dis-
cussion.
The successful use of iron or aluminum salts depends upon flocculation of
the initially formed precipitate, which requires gentle mixing for a period
of time. It is this function that may be performed in the sewer, thus
saving the costs for construction and operation of flocculation equipaent.
The development of a good floe depends upon the proper intensity of stir-
ring. This intensity can be measured as a velocity gradient, which is
equivalent to the square root of the rate of energy loss per unit mass of
water divided by the dynamic viscosity of the fluid. The mean velocity
gradient, in units of sec , is given by Camp (20) as
where
-r- = rate of change of velocity (u) with distance (x)
QX
E = energy dissipation rate per unit of volume
v = dynamic viscosity
For flow in a pipeline, E is proportional to slope (s) times velocity (u).
If velocity is expressed in feet per second, G has these values:
At 10°C G = 1980
At 20°C G = 1730
At 30°C G = 1550
-78-
-------
From observations In water treatment plants, Clark (21) concluded that
suitable values of G for good flocculation are in the ranga of 20 per
second to 70 per second. Similar conclusions have been reached by
others.
Figure 18 can be used for obtaining values of G, as well as velocities,
from slopes and quantities of flow. The graph is applicable only in
partly filled pipes in which the flow is between 1 and 95 percent of
full-pipe capacity. It has been shown that within this range the velocity
is practically independent of pipe diameter (22). The same would be true
for G.
It is evident from Figure 18 that velocity gradients found in sewers are
quite often in a range considered suitable for flocculation. The require-
ment for self-cleansing velocities eliminates G values that are too low.
Values that are too high, however, may be found in large swift streams.
Ideally, the dissipation of mixing energy should be distributed uniformly
in the flocculation chamber. This ideal can never actually be attained,
and the departures from the ideal may be large. Flocculators with similar
average velocity gradients may perform differently, depending upon how
well the shear or velocity gradient is distributed. The turbulence due
to flow in a pipe will presumably come much nearer to an ideal distribu-
tion of the mixing energy than will the turbulence produced by paddles.
Good flocculation may be obtainable at higher G values than in conventional
flocculators.
It is necessary that the flocculation process be continued for a suitable
length of time. It is considered by some engineers that the dimensionless
product Gt is a useful parameter for design. Camp (20) found that good
operation has been obtained with Gt values between 20,000 and 200,000.
However, the time should not be less than ten minutes. Suitable flow
times for the attainment of good flocculation will probably be between 10
and 30 minutes in most cases.
-79-
-------
FIGURE 16
VALUES OFVELOClTIES,u,AND VELOCITY GRADIENTS ,G,
FLOW, CUBIC METERS PER SECOND
O.02 0.04 0.06 0.1 0,2 0.4 O.6 1.2
ID
a.
o
_l
to
TEMP - 20° C
PARTIALY FILLED PIPES
04 0.6 1.0 2 46 10 20
FLOW, CUBIC FEET PER SECOND
40 60 100
-60-
-------
Another requirement for thfi most efficient treatment with coagulating
chemicals is that the initial mixing of the chemical into the water be
very rapid. Addition of the chemical at a location where the stream makes
a drop of two feet or more would result in good initial mixing. Other-
wise a mixer with an energy input of one to two horsepower per cfs of flow
should be provided. The volume of the mixing chamber need be only large
enough to absorb this energy input.
The greatest obstacle to in-sewer flocculation is the requirement that
the sewage then be transferred to the settling basin without floe-
disrupting turbulence. Where the sewage must be lifted into the treatment
plant by pumps, as is true in most plants, in-sewer flocculation would be
fruitless. Even where the sewage is not punped, the floe could be pre-
served only by avoiding any hydraulic jump, high velocities in pipes, or
excessive turbulence created by valves or other fittings. It would,
therefore, be required that the level of water in the primary clarifier
be practically the same as in the sewer. Since the elevation in the
sewer varies, the level in the primary tank would need to be varied to
keep pace. A control on the discharge of the tank would serve this pur-
pose. It would not be harmful if the water were backed up somewhat in
the sewer on the lowest flows, so that the range of water surface eleva-
tion in the tank could be less than the normal range in the sewer under
uniform flow conditions. For small sewers, up to, say, 0.6 n, no
special problem would be created. With large trunks, however, it would
become difficult to operate a constant-level clarifier while at the same
time preserving a floe produced in the sewer. A 1.5-tn diameter trunk,
as an example, would require a range of at least 0.3 m and perhaps 0.6 m
feet in the clarifier. Conventional clarifiers are not built to accommo-
date such fluctuations, but they could be. It is not necessary that the
water be taken from the tank by falling over a weir. Tanks with submerged
withdrawals are used in a number of places, the level being modulated by
a float controlled valve or a float controlled pumping rate out of the
clarifier. In one case a secondary small weir is used. Such a weir could
be made adjustable. There is much to be said in favor of weirless tanks,
especially where odor release is a problem.
-81-
-------
The skimming of a variable-Level tank, is another problem that could be.
solved with a little ingenuity, as for example by a floating skimming
boom and a floating trough.
It is doubtful that a comminutor or bar screen could be made large
enough to avoid turbulence that would disrupt a flocculent precipitate.
It is possible, however, to dispense with such devices ahead of primary
settling, using relatively large pumps of suitable design to remove the
sludge to a shredder, degritter, thickener, or such other devices as may
appear necessary.
The incentive to develop designs to bring the sewage into primary tanks
without passing through a zone of rapid energy input or dissipation
should not be confined to instances of planned chemical flocculation.
In many cases the suspended solids arriving at a treatment plant are
more or less flocculent, as a result of either chemical or biological
processes occurring in the sewer. Insofar as that flocculent character
can be preserved, better clarification will result.
-82-
-------
REFERENCES
1. Pomeroy, R. D., "Generation and Control of Sulfide in Filled Pipes,"
Sewage and Industrial Wastes,, 31, 9, 1082-1095, 1959.
2. Parkhurst, J. D., and Potaeroy, R. D.t "Oxygen Absorption in Streams,"
Journal of the Sanitary Engineering Division. ASCE, Vol. 98,
No. SA1, Proc. Paper 8701, 101-124, Teb., 1972.
3. Metzger, I., "Effects of Temperature on Stream Aeration," Journal
of theSanitary Engineering Division. ASCE, 94, No. SA6, Proc.
Paper 6309, 1153-1159, Dec., 1968.
4. Gameson, A, L. H., "Weirs and Aeration of Rivers." J. Inst. of Water
Engineers. Vol. 11, No. 6, Oct., 1957.
5. Barrett, K. J., Gameson, A. L, H., and Ogden, C. G., "Aeration
Studies at Four Weir Systems," Water and Water Engineering.
Sept., 1960.
6. Gameson, A. L. H., Vandyke, K. G., and Ogden, C. G., "The Effects of
Temperature on Aeration at Weirs," Water and Water Engineering.
Nov., 1958.
7. Pomeroy, R. D., and Parkhurst, J. D., "Self-purification in Sewers,"
Proceedings 6th International Conference on Water Pollution
Control Research. Jerusalem, Pergammon Press, 1972.
-63-
-------
8. Wiley, A. J., Lueck, B. F., Scott, R. H., and Wisniewski, T. F. ,
"Coinnercial Scale Operation of Turbine Aeration on Wisconsin
Rivers," Jour. WPCF. 32:186, 1960.
9. Sulfide Control in Sanitary Sewerage Systems. Process Design Manual,
U. S. Environmental Protection Agency, Technology Transfer,
1974.
10. Mitchell, R. C., U-tube Aeration. Report to U. S. Environmental
Protection Agency, Project No. 17050 DVT, Contract 68-01-0120,
1973.
11. Sewell, R. J., Controllmg Sulfides in JLan.itary Sewers Using Air and
Oxygen, U. S. Environmental Protection Agency, Project No.
- 11010-DYO, EPA-670/2-75-060, 1975.
12. Mitchell, R. C., Preliminary Report on U-tube Aeration, Rocketdyne,
1972.
13. Pasveer, A., "Oxygenation of Water with Air Bubbles," Sewage and
Industrial Wasteis. 27:1130, 1955.
14. Pasveer, A., "Efficiency of the Diffused Air System," Sewage and
Industrial Wastes, 28:28, 1956.
15. Bewtra, J. K., and Nicholas, W. R., "Oxygenation froa Diffused Air
in Aeration Tanks." Jour.WPCF, 35, 10, 1195, Oct., 1964.
-84-
-------
16. Mancy, K, H., and D. A. Okun, "The Effects of Surface Active
Agents on Aeration," Jour. WPCF, 37:212, 1965.
17. WPCF Manualof Practice No. 9, "Design of Sanitary and Storm
Sewers," p. 157, 1969.
18. Kessener, H., "Dutch Government Institute for the Purification of
Waste Waters, Annual Report—1931," A Review by W. Rudolfs,
Sewage WorksJournal, 5:2, 346-351, 1933,
19. Pasveer, A., "Research on Activated Sludge: II. Experiments with
Brush Aeration," Sewage and Industrial Wastes, 25, 12, 1397-1404,
1953.
20. Camp, T. R., "Flocculation and Flocculation Basins," Trans. ASCE,
120:1, 1955.
21. Clark, J. W., and Viessnan, W. R,, Water Supply andPollution Control,
International Textbook Co., Scranton, Pennsylvania, 270, 1965.
22. Pomeroy, R. D., "Flow Velocities in Small Sewers," Journal WPCF.
39, 9, 1525-1548, 1967.
-85-
-------
APPENDIX I
SELF PURIFICATION IN SEWERS
RD POMERQY- and I D PARKHURST"
•Piej.dcn; Pomeroy, Johnston and Bailey, 660 South Fur Oiki Avenue,
Pwadcnfl, Cjhrornu 91105, USA " 'Chief Engineer and Genial Manager,
Los Angeles Couniy Sanitation Districts, 2020 Beverly Boulevard,
Lai Angeles, California 90057, USA
i
The oxygen balance in a stream of sewage is important in respect to the possible
generation of sulfide, and in respeci lo the degree of biological oxidation that may occur
ir a sewer
In a previous paper []], the equation for oxygen balance in a stream of sewage was
written
dE>
= Rr + R« - Rf mg/1-hr (1)
di
in which D= oxygen deficit, g/ms (mg/1)
t = time, hours ,
R, = rate of reaction of oxygen in the fluid, g/m3-hr
Rf = rate of loss of oxygen to slimes on the pipe wall, g/m3 -hi
Rj»rate of supply of oxygen to the stream by surface aeration, g/m'-hr
Rf is proportional to C.and may be replaced by
ft»
Rf = — (2)
dm
in which f is designated the "exchange coefficient," expressed in m/hr, and dm is the
mean hydraulic depth in meters
In the article cited, a predictive equation for f. applicable to sewage streams, was
developed'
f = 0.96(1 +017 F3)7(su)3s (3)
J- is the Froude number, \/ ",7 is a temperature funclion, equal to 1.00 at 20°C, s is
e"dm
the slope ratio, and u is the velocity in m/sec
If f is expressed as feet per hour and u as fps, the coefficient is changed from 0 96 to 2 0
With the aid of the predictive equation for reaeration, attention will now be given to
R, and Rg
C/21/ll/l
Reproduced by permission of the International Association of Water
Pollution Research. Presented at the 6th International Water Pollution
Research. June 18 - 23, 1972.
-86-
-------
C/21,'11/2 RD PomeroyandJD Parkhursr
Reaction Rate of Oxygen in Sewage-Laboratory Tests
In Held work, a sample is taken from a sewer in a "D O pot" of the design shown by
Parkhursi and Pomercy [ij. After brief aeration, the diaphragm, carrying the mixer,
thermometer and oxygen probe, is put in place, and oxygen concentrations are read over
a suitable interval of time, yielding data for calculating Rr In some of the laboratory
determinations, use was made of a luei beaker and a tight rubber diaphragm, with a
water powered magTetic slirrer underneath so that the whole assembly couid be placed in
a water bath Comparative tests showed that the speed of mixing does not make any
measurable difference as long as it is sufficient to keep the organic solids suspended, and
provided also 'hat the water velocity past the probes is sufficient for reliable oxygen
results when the m-.cioameter is lead Fig 1 shows typical oxygen depletion curves
Within Ihe limits of accuracy of the measurements, the reaction rate appears to be
independent of oxygen concentration until the oxygen concentration is quite low,
generally a few tenths of a milligram per liter. Kessler and Nichols (2], quoting
unpublished work by Carl H. Nordell, first reported that oxygen usage in activated sludge
is essentially independent of oxygen concentration They showed straight lines extending
into a region .of negative oxygen concentrations, construed as the development of
substances of essentially instantaneous oxygen demand
To explore this phenomenon further, the sewage in a D,0 pot was allowed to stand
anoxic for a half hour A measured amount of a solution of 35 mg/l of oxygen in water
was added by way of a syringe in a suitable amount to produce a concentration of
3 3
O
S z
c
O
!
10 20
Tstie, minutes
30
Typical oxygen consumption curves
-87-
-------
Self-Punficatu>n in Sewen C/21 /1 1 /3
DISSOLVED
- > 0 5
mg/I
Fig 2
Typical relationship between low oxygen concentration and oxygen reaction late
1.28 mg/I in ihe sewage The eaiLest dependable reading of the probe was IVi minutes after
the start of addition of oxygen solution.it showed 1 23 mg/1 The subsequent reaction rate
was 0 06 mg/l-mnule Wnhin the limits of accuracy of the measurements, U appears that
the reaction had merely come to a dead stop, going on as soon as oxygen was agam
supplied
Anaerobiosis can produce substances exerting a chemical oxygen demand, sulfide and
ferrous iron being obvious examples, but when these compounds are present in smail
amounts, then rates of utilization of oxygen are not large The results of Kessler and
Nichols may have been due not to the development of an avidity for oxygen, but to
mteiferences in Ihe VVinkler test used to determine oxygen, causing all oxygen
measurements to be low
The rate of reaction of oxygen in the high oxygen demand region is designated R'r
When R,/R'r is plotted against dissolved oxygen concentration, one obtains curves of the
general shape shown in Fig 2 It would be difficult to account for the course of the curve
on the basis of a chemical reaction mechanism The lag of the probe as oxygen reaches
zero could contribute to the curvature, but this effect is small and would not account for
the shape of the curve More likely ihe process is diffusion-controlled when oxygen
concentrations are small The biologically active material is mainly flocculent Oxygen
must reach (he interior of floe particles by molecular diffusion At low_ oxygen
concentrations, the cores of floes become starved foi oxygen, and as these anoxic neuclei
grow in size, the amount of oxygen used by ihe sewage per unit of Urre decreases This
view is supported by ihe observation that oxygen depletion rates determined on samples
with rather coarse suspended matter show curvature at higher oxygen concentrations than
do samples lacking coarse solids
Oxygen reaction rales were determined on samples from the Los Angeles County
Sanitation Districts system The standard 5 dav BOD, and even the I-day demand, bear
little relationst'ip to the 1 -hour demand, as shown by Fig 3 The rates there shown have
D
been corrected to 20°C on the assumption that = 1 07T-KO,
0>
-88-
-------
C/21/11/4 RD Pomeroy and J D Paikhurst
Results from various test sites are given in Fig 4 The designations of the sampling
sites, shown at the bottom, consist of numbers followed by letters The numbers signify
the sewer pipe diameters in inches There is a tendency for R, to be lower in small sewers
than in large ones It was found that Rr usually increases as the sewage flows along
The sewage from test site 24C has an abnormally low value It was found that the pH
of that sewage was quite high much of the time, up to 10.0, thus suppressing biological
actmly The high pH is due to the frequent discharging of hjghly alkaline industrial
wastes, a practice encouraged by the Sanitation Districts because of the beneficial effect
in suppressing sulfide generation
Two series of tests, one m September and one in January, were made along Los
Angeles County Sanitation Districts South Slope-Trunk, on the seawaid slope of the Palos
Verdes Hills (Fig 5) The sewage was limed by use of dye, and the testing paced the flow.
Row is essentially by gravity, but there are two siphons, and there is a 5,400-fl length of
above.ground steel pipe of indefinite profile, laid over a creeping landslip Two lift
stations pumping intermittently into the line were shut off during the January tests They
were not shut off during the September tests, but it appears thai they had bttle, if any,
effect, they may not have contributed to the sewage tested
(L
a
u
t
j
E *
UJ
a.
w
«t
tr
u
J
i °c
u'
a:
IACTION
• r
5 "
z
lif
O
g
'
A
.'^
^•-:-
'•a8°5
1C
One
r
.
1*
Cm
0 H
401 2C
•
'a
,>.*-'
,*
1C SC
" C BO
•
'•'
>. -•
6 " .
0 «C
0. mg/1
.-
• 1
'
0 50
.
0 'CO ZOO JOO <00
FIVE-OAT. Zo' c BOD. m«/l
Fg 3
Comparison of mvgen reaction rale with 1-day and 5-day BOD AH results corrected to XfC. BOD on setlted
samples indicated is circles BOD on whole samples indicated as dots All Rr tests were on whole seuage
-89-
-------
Self-Purification in Severs C/21 /11 /5
u
a
LJ
t-
,,8<- „,/ I
85 f ]•
IOC
101
IDE
)OA
I2H
151
IJO
«r
66
Tse
~?7J
• ^
***"ii
j» 1
/«.1 -
» *
• ,% ",
.•
<* '
^ .*
«
r- -
t
I o>» .'•" i! ".
56C.
)6H
36K
_«8S_
48B
48H
80C
606
&00
6 OF
fiCJJ^
20 Jfl
i2Ol.fi
f *
*
l"»*
*
j
i I
|
) i
.m
*
« .
*
i* *
1
"J
*
•
•
*V-r '*
*^f
7
3
a*
]
•
£
0 * 9 '2 (
DX*CEN REACTION RflTE AT 20*C MG P£R L-H«*
OXYGEN REACTION RATES
F,£ A
Oxygen Reaction Rates
Tributary sewers brought in sewage that always had lower R, rates than the mainline
flow, producing a saw-tooth profile of Rr values The circled points in Fig 5 are the only
ones actually measured on the main line. The olhe; points are calculated from the flows
and the measured or assumed values for the characteristics of the influxes. In the case of
Ihe dissolved oxygen graphs, Ihe calculated mixtures may be considerably in error,
because turbulence at the junctions doubtless carries additional oxygen into solution A
part of Ihe incremental flows is due to service connections along the route. These
conlributions. too, produce aeration as they enter the mainstream Thus, dissolved
oxygen is present throughout most of the reach, even though Rr generally exceeds the
calculated reaeration rate, tabulated at the top of Fig 5 The oxygen supply rates,
incidentally, are not greatly affected by lemperature, because the increase of the
reaeration coefficient with temperature has an effect opposite to the decrease of oxygen
solubility
After leaving the South Slope Trunk, the sewage enters the Whites Point purr>p station,
along with an additional tributary flow it spends 45 to 60 minutes in the force mam,
being kept aerobic by injection of air Fifteen pairs of offset samples at the pump station
and end of the force main showed that R, sometimes increases in the force mam and
sometimes decreases The average rate at the pump station was 164 g/m'-hr,and 17.dat
the discharge end.
The increase of R, as the sewage proceeds downstream, shown so well in tne Palos
Verdes South Slope Trunk, seems to be a normal characteristic of sewage near its source,
continuing for a considerable distance if oxygen is supplied at a sufficient rate The South
-90-
-------
C/21/1I/6 RD Pomeroy and J D Parkhurst
Slope Trunk is nol unusually well aerated, but it drains a hilly terrain, so lhat junctions
lend to keep it from becoming anoxic along most of its length Sewers in fiat areas, laid at
minimum slopes, are likely to provide adequate aeration only where the flow depth is
shallow. The deeper (lows in large; trunks become nearly anoxic
Various sewages were aerated in the laboratory, ard samples were withdrawn from
time to time for R, determinations Fig 6 shows typical curves for the changing reaction
rates. Curves A and B were made with samples of strictly residential sewage? taken near
the points of origin Curve C was obtained with sewage from the South Slope Trunk at
Whites Point pump station Even though it had spent an average of only about 3!4 hours
in the trunk, it had passed the Rr peak Curves D, E, and F are for sewages from three
metropolitan trunks entering the Los Angeles City plant at Hyperion and the Los Angeles
County Sanitation Districts plant at Harbor City These sewages probably had average
ages of about 10 hours, but for most of lhat time they had carried very little dissolved
oxygen The small secondary rise shown in curve E is not unique It was clearly shown in
several other sewages, especially as the techniques of measurement weie refined
Fig S
Oxygen balance, Palos Verdes South Slope Trunk
-91-
-------
Self-Purification m Sewers C/21/11/7
Fig 6
Changing Rr values in aera'ed sewages
The pronounced peak in the oxygen demand rate curve is intriguing On a
guess thai organic acids mighi be 3 Factor,- duplicate sample; of a residential
sewage were run m parallel, but to one sodium acetate was added la a
concentration of 44 m^,1 as acetic acid The test was run at 21 5°C Rr increased
slightly faster m the one with added acetate. The striking difference appeared
when the untreated sample reached a peak of 27 mg/l-hr at 2 S hours and Ihen
declined abruptly, but in the presence of acerate Rt climbed for ! 0 hour more.
to a peak of 47 mg/lhr The difference between the amounts of oxygen used by
the two samples up to the point when the second one declined was 22 mg/l. or
0.95 rrot of Oj per mol of acetic acid Considering only ihe slotchiometty, ore
m'ght suppose lhat the acetic acid was oxidized to giyoxylic acid. A similar
experiment wnh another sewage showed a mol ratio of 0.9
_92-
-------
C/2J/M/8 R.D PomeroyandJD Parkhurst
A sample thai had had acetate added reached an Rr peak of 38 mg/l hr at 27°C. After
aerating it for an hour more, it was allowed to stand overnight without air and then was
aerated six hours rrore, at which time R, was 6 2 Acetate was again added in the amount
of 70 mg/I Immediately thereafter R, was42 mg/]-hr at 263°
The strong evidence that acetate oxidation accounts for the peak was clouded by a lesl
run on sewage from the souae thai gave curve E in Fig 6 A sample was analysed for
volatile organic acids by distillation and tttration, yielding a result of 0 76 me/I The
sewage was aerated until ihe peak was passed, then analysed again The organic acid
content was Ihe 0 35 me/1 Gas ihromalography of the acids before and after oxidation
showed Ihe proportions by equivalents to be 63% and 6!% acetic, and 19% anc! 17%
propsomc Small amounts of the butyric, valeric and caproic acids made up the
remainders
The findings from these exploratory experiments may be sufficiently challenging so
that others will punue the subject further The results have a bearing on oxygen
utilization in sewers, and on the possibilities of m-sewer treatment, and may throw light
on the apparent deferences of treatab-lity of sewages of different ages and m different
climates, and on the effects of digester liquors relumed lo the activated sludge process.
The biological oxidizing culture develops more rapidiy in the sewer than it does in an
aerated vessel in the laboratory. In some sewers the culture tends to be (locculent tn
character, settling rapidly and leaving the sewage with a low Rr value, but in other cases it
setiled poorlv The samoles aerated in the laboratory settle very poorly It appears that
the biological floe grown n the sewer is, in part, slime that has sloughed from the pipe
wall
R, can be determined from oxygen balance data in sewers, provided Re is reduced to
zero This was done by adding sodium hydroxide to the sewage stream to raise the pH to
12 or above for about a half hour followed by sodium hypochJontc added in high
concentration to assure that the biological slime layer was not only lulled but also well
oxidized. HypochJonte treatment alone did not suffice unless continued at a high dosage
for a long une After the cessation of chlormatwn, oxygen concentration was measured
al both ends of a test reach, aJong with measurements of velocity, discharge and
temperature The test procedures were as described previously The oxygen exchange
coefficient, f, was calculated by Eq 3 Since the changes of oxygen concentration in the
reaches were usually small in companion with the average deficit, the differential
equation was usually replaced by the approximauon
D
(4)
in which (O-i ) is the oxygen concentration ;n mg/l
Results are shown in Taole 1 It is believed that the largest source of error causing
discrepancies between the values of Rt obtained by the two methods is in the
determination of A(0,)/At, especially when the change in a time as shorl as 7 minutes
must be converted to a per-hour rale There appears lo be a tendency for the sewer results
lo be lower than the poi tests, but the difference is not statistically significant
Oxygen Consumption by Slimes
Oxygen balance tests m sewers in their normal condition, together with determinations
of R, on samples withdrawn from the sewer, allow estimates of oxygen usage by
the slime layer
-93-
-------
Table 1 Oxygen Reaction Rates Determined in Sewers
I
v£>
•P-
I
Test Length Number of
Reach in runs
meter]
Slope,
s
Temperature.
T, in
degrees
centigrade
Time
in
hours
Mean
hydraulic
depth. dm.
in 'm
Velocity.
u. in
meters
per second
Average
oxygen
deficit.
QJD)in
milligrams
per liter
Grams per cubic meter-hour
(milligrams per tiler-hour)
Rate of
change of Rate of
oxygen reaeration,
concentration. Rf
Al
Rate of
reaction of
oxygen.
Col 'lt>
Col 11
Rale of
reaction of
oxygen.
Rrby
pot test
(I) (2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(ID
(12)
(13)
10A 419
15D 682
I8D 827
I8J 739
21G 797
24C 3220
2
2
2
2
5
3
00024
0 00303
00012
00097
000083
000071
219
21 6
226
234
229
22 1
0202
0434
0472
0 112
0359
1 50
0088
0088
0084
0313
0 145
0277
0576
0438
0482
1 852
0618
0653
382
642
412
6 14
639
607
-29
-02
+03
+ 13
+09
-06
39
62
34
50
2.2
13
1 0
60
37
63
3 1
07
1 4
59
25
72
2 1
03
•=?;
J
^>
5
5
Oi
-------
Table 2 Determinations of Transfer Coefficients of Oxygen to Sewer Slimes
n
i
VD
Ol
Test
Reach
(1)
IOA
!5D-a
15D-b
I8D
18F
I8H
18J-a
18J-b
2IG
24C
Length,
in meters
(2)
419
682
682
827
454
922
739
739
797
3,220
Number
of runs
(3)
7
3
4
7
8
3
4
3
3
6
Velocity, u.
in meters
per second
(4)
0646
0460
0640
0.549
1 197
1 930
1 774
2 162
0591
0698
HI,
in meter)
per second
(5)
000142
000139
000194
0 000659
000431
00154
00172
00210
0000491
0 000502
Average Oxygen
in grams
per cubic meter
(6)
403
708
627
539
623
631
709
6 15
684
642
Change of oxygen
concentration,
A02
grams
per cubic meter
(7)
-094
-105
-1 22
-1 59
-053
-079
-096
-093
-1 11
-082
Mean
Hydraulic
Depth, dm.
in meters
(8)
0090
0090
0 170
0093
0.144
0 197
0214
0 101
0 154
0361
Hydraulic
Radius, r.
in meters
(9)
0060
0069
0 101
0074
0 101
0 119
0 123
0079
0111
0174
o
JO
o
"O
o
n
V
3
CL
«— i
D
TJ
ta
[f
£
-------
Tdble 2 Determinations of Transfer Coefficients of Oxygen to Sewer Slimes (continued)
(a) (a)
Oxygen exchange Rale of oxygen
coefficient, f, supply, Rf, in
Test
Rejch
(1)
10A
15D-a
15D b
18D
I8F
IKH
18J-a
18J-b
2IG
24C
in meters
per hour
(10)
0 0937
0 0867
0 102
00705
0 154
0276
0281
0441
0059
0060
grams per
cubic meter-hour
(ID
4 2
69
38
4 1
67
88
94
269
26
1 1
Rale of change
of oxygen
concentration,
A(02>
At
grjms per
cubic mcler-hour
(12)
-5 2
-2 5
-4 1
-3 7
-4 9
-60
-82
-85
-30
-06
Rate of Reaction
of oxygen, Rr,
in grams per
cubic meter-hour
(13)
25
46
5 1
35
6 1
7 3
59
86
28
04
(b)
Rate of loss Observed
of oxygen to Oxygen transfer
slime layer, coefficient, e, e~
Re, in
grams per
cubic meter-hour
(14)
69
48
28
43
5 5
75
1 1 7
268
28
1 3
in meters (su)^
per hour
(15)
0 10
026
0 19
012
030
043
1 10
108
0 18
0 14 '
Logarithmic Average
(16)
265
70
'43
47
46
35
84
75
8 1
62
5.3
(c)
Oxygen transfer
coefficient, e.
calculated in
meters per hour
(17)
020
020
023
0 14
035
066
070
077
0 12
0 12
5?
*s
^
•5,
3
5
3
3
i
(a) Computed from predictive equation (b) Column 11 — Column 11 - Column 13 (c) By e = 5 3
O
-------
C/21/II/12 R D. Pomeroy and J D Paikhurst
Table 2 shows results of "18 test! m 8 sewers Results for each sewer were averaged,
except that wher a sewer was tes ed under two quite different flow conditions, or at two
times separated by an interval during which the hydraulic condition changed because of
cleaning or some other circumstaice, the results were divided into two groups Thus
Table 2 shows averages of ter, seis of iuns
It may be noted tlut A(0a) was negative m all cases. The reaches tested were
necessarily the maximum available distances presenting uniform conditions, wilh no
junctions In all cases there were junctions not far upstream, resulting in points of
turbulence raising dissolved oxygen so levels that could not bt sustained by the reaeration
rales in the test reaches
It seems reasonable to suppose that the net rate at which the slime layer can use
oxygen will be influenced by oxygen concentration. In view of (he rich food supply, the
slimes may develop a capacity 'o ase oxygen as fast as it arrives If this is true, then 'he
amount of oxygen used should be propomonal to the oxygen concentration in the water
The flux of oxygen to the slime layer, 0e, is equal to R^r, in which r is the hydraulic
radius. If the flux is proportional to the oxygen concentration (03),one may define a
transfer coefficient, e.
(5)
t is analogous lo the exchange coefficient from air lo water, and will be expressed in
oxygen balance calculations as m/hr
If Ihe rale at which the slime layer uses oxygen is dependent upon the rate of arnvaJ of
oxygen, then it should be proportional not only to the concentration of oxyger in
solution, but also to the rate of renewal of water adjacent to the shme layer by eddy
diffusion Thus, it may be a function of Ihe rate of energy dissipation. If it be assumed
that it is a simple power function, one may wnle
e = CE (su)q f6)
Fig 7
Effect of rate of energy dissipation on coefficient for transfer of oxygen lo the sLme layer.
-97-
-------
Self-Purification in Sewen C/21 /11 /13
Fig, 7 shows a plot of e vs. su. The regression line indicates q = 0 51 The standard
error of the exponent is about 0 11 The data are not free from internal correlations, the
uncertainly of the exponent, therefore, is greater lhan indicated by trus standard error,
Taking q to be W and determining CE, the equation becomes
e = 5 3 (su)* (e in m/hi, u in ITI/MC) (7)
In filled pipes, s vanes approximately as u2 Hence (su)rt Is proportional to u3/2. Heat
transfer to or from the wall of a pipe is generally held to vary as u5.
Eq 7 applies only where the slime is an effective oxygen sink. In a well aerated stream
of low oxygen demand, the subsl'ate may be entirely aerobic, in wtuch case the rate of
usage by the slime layer would not be proportional to the oxygen concentration.
On* should not think of the slime layer as a bulky mass. Where velocities are good, the
submerged pipe wall usually appears clean, but the sbme layer always gives the surface a
characteristically slippery feel unless it is cleaned off by rubbing. It is significant that the
highest oxygen flux to the pipe wall, 1.4 grarns/m2-hr, was in line J8J at a velocity of 7.1
fps (2.2 m/sec) The finding of a highly icactive slime layer where the stream is so swift is
in a sense an extension of the finding, previously reported [3], that a velocity of 4 fps did
not dimmish sulfide production in a force main Let no one suppose that high velocity
will strip a pipe wall free of biologically active material, unless it be at some velocity
greater than 7 fps, or unless much abrasive material is present.
Condition of Deficient Oxygen Supply
Three of the reaches tested (48D, 27D, and 27J) had very low oxygen concentrations
at the downstream ends It was calculated in all of these cases that the oxygen
concentrations had come close to the steady stats values, where d(0, )/dt contributed
little or nothing to the balance. The rates of oxygen consumption equaled the rates of
supply.
By using Eq 7 to predict slime usage and Fig. 2, together with pot tests, to estimate
R,, predictions may be made of the steady stale concentrations. Seventeen tests in the
three lines mentioned showed oxygen concentrations ranging from 0.03 to 0.27 mg/1.
Predicted concentrations were in the same range, but the standard error of the predictions
for individual runs was 0 09 mg/1. In view of the errors of measurement and the fact that
Fig 2 represents only an average trend, closer agreement would not be expected, it is
sufficient that the observed range is rationally accounted for.
Deductions
The significance of these findings for the problem of seplicity in sewers will be treated
elsewhere
Wherever the slime layer acts as in efficient oxygen sink, and with normal surface
aeration providing the only source of oxygen, concentrations will drop to low levels The
transfer coefficient to the slime layer is in general about 2W times as great as the
coefficient for absorption of oxygen from the air. If no oxygen were used in a stream of
sewage, but ihc slimc'
-------
C/21/11/14 R D Pomeray and J D. Paikhursl
actual consumption in streams more than a few inches deep in a W2rm climate, or more
lhan 2 or 3 (eel deep where sewage temperatures are low, will establish steady-state
concentrations near zero
In line 1SJ (Table 2) Ihe oxygen concentration averaged 1 95 mg/l,bu( even with a
stream velocity of 7 J fps it was losing oxygen at a rate of 8 J mg/l-hr;it would reach a
study-slate concentration of 1.2,rrg/l Seven of the ten Mis of data in Table 2 foretell
steady-state concentrations less than 0.5 mg/l.
Oxygen concentrations higher lhan Ihe steady stale levels calculated from the
prevailing flow conditions and observed R, are common, since surface aeration in
uniform-now reaches is not the only source Junctions and other points of turbulence
make significant contributions There are also many sewers in which the slime layer is not
100% effective as an oxygen sink.
Where the oxygen supply is adequate, oxidation of the sewage in transit may be a
significant factor. The Soulh Slope Trunk of the Los Angeles County Sanitation Districts
(Fig 6} is illustrative The oxygen supply is augmented by turbulence from steep
tributary lines, so that the stream is fairly well acraled aJongmuch of Us length Estimates
made on the basis of the data for January :n Fig 5 indicate that Ihe total flow reacts with
about 67 mg/1 of oxygen by the lime it reaches (he pump slation. In September, more
oxygen is used by the slieam, but use by Ihe slimes is less, the loial being about 85 mg/l
Another tiibulary flow enters just ahead of the Whites Point pump station, and then Ihe
tolaj flow is aerated by compressed air in the force mam, consuming 16 mg/l of oxygen
under average conditions The oxygen absorbed in the force main partially compensates
for the flow added near Ihe pump station Thus it appears that about 60 to 75 mg/l of
oxygen, depending upon the season, react with the sewage from the lime it enters the
(runk until it discharges from the force mam This is under daytime flow conditions The
effect would be somewhat grea'er for the 24-hour average flow.
Six samples of sewage taken at the end of the force main in March and May showed
BOD concentrations averaging ]42 mg/l An aviage BOD deiernvned on a set of samples
fion the tributary lows, taken al similar hours and seasonal conditions, showed 195
mg/l It appears well estdblished that the sewage did receive i substantial degree of
biological oxidation in the trunk and force main. Furthermore, some oxidation
undoubtedly occurs in the tributary lines, adding to Ihe tola] in-sewer BOD reduction
that occurs in this instance. The trend toward regional sewerage systems means lhal an
increasing number of large, long trunks will be built. Detention times will be long enough
in some of these trunks to permit major BOD reductions, but this does not occur in such
Irunks because of oxygen starvation If oxygen were supplied in sufficient amount,
substantial reduction could be made in the loads reaching downstream treatment plants
Summary
1 Oxygen reaction rates in sewages vary widely, being low neat the origin of Ihe sewage,
increasing for several hours if aerobic conditions prevail, up to rates as high as
20 mg/I-hr, ihen declining This behavior was demonstrated both in laboratory
experiments and ir sewers
2. By using a gredictive equation for reaeration and measuring Ihe lates of use of oxygen
by the sewage, calculations were made of the rate of oxygen utilization by Ihe slime
layer on the pipe wall 11 was hypothesized that in the sewers tested the
slime layer acted as an efficient sink for the oxygen reaching it The
-99-
-------
Self-Punficatton m Seven C/2 1 / 11 / 1 5
results were consistent with this hypothesis. A predictive equation for the
R,, the rate of loss of oxygen to the slime layer, is
3 In well oxidized streams of low oxygen demand, the slime layer may not be
completely effective as an oxygen sink
4 A substantial amount of biological oxidation occurs in some sewers, particularly where
the oxygen supply is augmented by the turbulence produced at junctions and other
structures. By designing sewers in ways to maximize aeration of the stream, or by
supplying oxygen in other ways, sufficient biological oxidation may be induced to
effect substantial reductions in the BOD loads on treatment plants
SYMBOLS
Ce general constant for the equation for transfer of oxygen to the slime layer,
m/hr
Cf geneia! constant for the predictive aeration equation, m/hr.
dm mean hydraulic depth, cross section area of stream divided by surface width, m
or ft,
ID|, IDj oxygen deficits at begjnrnng and end of a test reach, mg/1
e oxygen transfer coefficient from water to a slime layer acting as an effective
oxygen sink, m/hr.
u
F Froude number, Vdmg
f oxygen exchange coefficient, m/hr
(03). oxygen concentration, grams pet"cubic meter (mg/1)
Q flow, mj/sec or cfs
r hydraulic radius, m or ft.
Rf rate of supply of oxygen to a stream, g/m-'-ru.
R, rate of reaction of oxygen with impurities in the water, g/m'-hr
R't rate of reaction of ovygen under the condition that there is no retardation due
to a suboptimal concentration. g/m9-hr
Rj rate of loss of oxygen from stream to slime layer, g/m3-hr
s slope of the energy tine of the stream.
T temperature, °C
t time, seconds or hours
u ' velocity of flow, m/sec or ft/sec
7 ratio of exchange coefficient for oxygen absorption by a stream at temperature
T to the coefficient at 20°C, other conditions being the same
#e flux of oxygen from the stream to the slime layer, g/m2-hr
oif flux of oxygen from the atmospheie to the stream, g/m2 -hr.
-100-
-------
C/2I/1I/16 RD. Pomeroy and J.D Parkhurst
REFERENCES
1 PARKHURST, J D , «nd POMEROY, R D , "The Absorption of Oxygen in Sewers," accepted for
publication in the Journal of the Sanita/y Engineering Divmon, Proceedings of the American
Society of Civil Engineers
2 KESSLER, L H,. and NICHOLS, M S , "Oxygen Utilization by Activated Sludge," Sewage Workj
Journal, Vo! 5, No 7, Sept, 1935, pp 810-838
3 POMEROY, R D, "Generation and Control of SulTide in FiUtd Pipes," Sewage and Industrial
Wastes, VoL 31, No 9,Sepi, 1959, pp 1082-109J
ACKNOWLEDGEMENTS
This paper reports pcihons of work done under two projects (a) research on the
generation and control of sulfide in sewers, jointly sponsored by the Environmental
Protection Agency and the Los Angeles County Sanitation Districts, and (b) research on
the possibilities of partial or complete m-sewer treatment of sewage, conducted under a
grant from the Environmental Prelection Agency to the firm of Pomeroy, Johnston and
Bailey. On the sulfide control project, John Paikhuist was Project Director, Charles Carry
provided administrative oversight, Jack Livingston was in charge of data collection, and
Richard Pomeroy and Harry Bailey were consultant, Richard Pomeroy, Jay Crane, and
Ronald Lofy were the principal participants on the in-sewer treatment study
Appreciation is expressed to Charles Swanson and Jerry Stern of the Environmental
Protection Agency for their advice in the planning and execution of both projects.
-101-
-------
APPENDIX II
CALCULATION OF EQUILIBRIUM CONDITIONS WHEN OXXGEN IS MIXED WITH SEWAGE
Assume a situation in which sewage has a dissolved nitrogen concentration
in equilibrium with the atmosphere and no dissolved oxygen. If oxygen
is mixed with this sewage, oxygen will dissolve and nitrogen will enter
the gas phase. The problem is to calculate the equilibrium condition.
It is to be assumed that the sewage is not retained in contact with the
gas phase for a long enough time to allow any substantial biological
consumption of the oxygen.
Symbols
P = pressure in the gas phase, atmospheres
T = temperature, °C
d = density of dry oxygen at P, T
d = density of dry nitrogen at P, T
H = vapor pressure of oxygen, atm., over a solution of 1 mg/1 0_ at T
H = vapor pressure of nitrogen, atm., over a solution of 1 mg/1 of N. at T
V = volume of gas phase associated with one liter of water
V = partial volume of 0_ in gas phase associated with one liter of water
V = partial volume of N, in gas phase associated with one liter of water
P = vapor pressure of water, atm.
a' * initial oxygen concentration, mg/1
a = oxygen dissolved, mg/1
b = nitrogen leaving the water, mg/1
A = amount of oxygen supplied, mg/1 of water
-102-
-------
Solution
1. P = P + P + P P+P=P-P
o n w o m w
3. P = 0.79(1 - P ) - bH
n w n
(The partial pressure of nitrogen, in dry air at one atmosphere pressure
is 0.79 atm. , but in contact with water it is diluted by water vapor,
hence the factor 1 - P .)
F = aH + 0,79(1 - P ) - bH + P
o w n w
aH + 0.79(1 - P ) - P + P
, o w w
b = - _ -
n
aH + 0.79 + 0.21F - P
/ -i O ___ W __
4. b . - - -
n
The partial pressures of the gases are proportional to the partial volumes,
P P
-
_ — _
on w
follows that
V+V P+P P-P
o n o n w
v -
-------
8- v=
-------
.[P - Pw - aHo(l - ) - °
-------
Figure 10 shows the results of calculations of the a vs. A relationship
at 1 atmosphere pressure, 1.3 atmospheres (about 10 feet of water above
atmospheric), and 1.6 atmospheres (20 feet of water).
COEFFICIENTS FOR EQUATION
FOR THE
OXYGEN EQUILIBRIUM CALCULATION
d H
m H acm. oo
Temperature o, TIT" p „,,.,
a- /, a n r . aun.
C per mg/1 n n w*
10 0.01861 0,4932- 0.0168
15 C.02047 0.4993 0.0231
20 0.02251 0.5075 0.0313
25 0.02464 0.5147 0.0419
30 0.02670 0.5218 0.0555
-106-
------- |