WATER POLLUTION CONTROL RESEARCH SERIES • I6080FYAO3/7I
OXYGEN REGENERATION OF POLLUTED RIVERS •
THE PASSAIC RIVER
ENVIRONMENTAL PROTECTION AGENCY • WATER QUALITY OFFICE
-------
WATER POLLUTION CONTROL RESEARCU SERIES
The Water Pollution Control Research Reports describe the results and
progress in the control and abatement of pollution in our Nation’s waters.
They provide a central source of information on the research, development,
and demonstration activities in the Water Quality Office, in the Environ-
mental Protection Agency, through inhnuse research and grants and contracts
with Federal, State, and local agencies, research institutions, and industrial
organizations.
Inquiries pertaining to Water Pollution Control Research Reports should be
directed to the Read, Project Reports System, Planning and Resources Office,
Office of Research and Monitoring, Environmental Protection Agency, Water
Quality Office, Room 1108, Washington, D. C. 20242.
— about our cover
The cover illustration depicts a city in which man’s activities coexist in
harmony with the natural environment. The National Water Quality Control
Research Program has as its objective the development of the water quality
control technology that will make such cities possible. Previously issued
reports on the National Water Quality Control Research Program include:
Report Number Title
16080—06/69 Hydraulic and Mixing Characteristics of Suction Manifolds
160&O——lO/69 Nutrient Removal from Enriched Waste Effluent by the
Hydroponic Culture of Cool Season Grasses
16O8ODRX1O/69 Stratified Reservoir Currents
l6080—U/69 Nutrient Removal from Cannery Wastes by Spray Irrigation
of Grassland
16O80D0007/70 Optimum Mechanical Aeration Systems for Rivers and Ponds
16O8ODVFO7/70 Development of Phosphate—Free Home Laundry Detergents
16080—1O/7O Induced Hypolimnion Aeration for Water Quality Improvement
of Power Releases
16O8ODWP11/70 Induced Air Mixing of Large Bodies of Polluted Water
16O8ODUP/12/70 Oxygen Regeneration of Polluted Rivers: The Delaware River
-------
OXYGEN REGENERATION OF POLLUTED RIVERS;
THE PASSAIC RIVER
by
Water Resources Research Institute
Rutgers University
New Brunswick, New Jersey
for the
WATER QUALITY OFFICE
ENVIRONMENTAL PROTECTION AGENCY
Project #16080 FYA
March 1971
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price 65 cents
Stock Number 5501-0131
-------
EPA Review Notice
This report has been reviewed by the Water Quality
Office, EPA, and approved for publication.
Approval does not signify that the contents neces-
sarily reflect the views and policies of the
Environniental otection Agency, nor does mention
of t ’ade names or commercial products constitute
endorsement or recommendation for use.
11
-------
ABS ThACT
Field tests were made of a mechanical surface aerator and of pure oxygen
diffusers in a small polluted river, the upper Passaic. Results gener-
ally corroborated results of previous test, as to performance of surface
aerators on such rivers, in excavated pools. A somewhat higher oxygen
transfer rate was obtained with a flow concentration device, which, in a
permanent installation, would take the form of low rock spur dikes, one
extending from each bank, or flow concentration groins. Tests in shallow-
er water, about 7 feet deep, were inconclusive. Tests of oxygen diffusers
were fragmentary, due to mechanical difficulties with the equipment; but
it was demonstrated that the very fine bubbles used were very largely
absorbed in the water. A dye dispersion test gave a very high longi-
tudinal dispersion coefficient downstream of the aerator. Mathematical
modelling indicated that during the test period, parameters of biochemical
deoxygenation were not changed by the artificial aeration process.
This report was submitted in fulfillment of Project Number 16080 FYA,
under the partial sponsorship of the Water Quality Office, Environmental
Protection Agency.
iii
-------
CONTENTS
Section Page
I Conclusions i
II Recotnuiendations 3
III Introduction
IV Field Operations 7
V Oxygen Transfer of Aerators and Oxygen Diffusers 17
VI Dispersion Analysis 31
VII Matheniatical Model and Paranieters of BOD 39
VIII Acknowledge rnents 5 3
References
V
-------
FIGURES
Page
1 Test Site Depths 8
2 Mechanical Aerator in Operation 9
3 Martin Marietta Static Oxygen Diffuser 11
) 4 Martin Marietta Dynamic Oxygen Diffuser 13
Conversion Factor Vs Water Temperature and DO Level at the
Aerator 22
6 Oxygen Thansfer Rate and. Discharge 2 ) 4
7 Dye Patterns in a Stream 32
8 Dye Cloud Frontal Structure Vs Time 3
9 Differentiated Dye Cloud Frontal Structure Vs Time 36
10 Dissolved Oxygen Sag Curve Fitted by Digital Computer,
July, 1970, Aerator On
11 Dissolved Oxygen Sag Curves for Various Values of K 1 , and
August 12, 1970, Aerator On
12 Scaled Analog Computer Diagram Used in the Simulation of
BOD, NOD, and DO
13 Analog Computer Drawn Curves of Dissolved Oxygen Concen-
tration Vs Holding Time, August 12 Data
1)4 Analog Computer Drawn Curves of Dissolved Oxygen Concen-
tration Vs Holding Time, July 27 Data
l Analog Computer Drawn Curves of Dissolved Oxygen Concen-
tration Vs Holding Time, July 2)4 Data SO
vi
-------
1JthL b
No. Page
5-i Suuirnary of Steady—State Field Test Data for Mechanical
Aerator 20
5—2 Sunimary of Oxygen Transfer Rates for Mechanical Aerator
Under Steady-State Operation 21
5-3 Summary of Oxygen Transfer Data, Diffuser Plates 27
5-L 1 . Spent Gas From Oxygen Diffusers 29
7-1 Summary of Results by Digital Computer Simulation t l
7-2 Suniniary of K 1 . and Kn Values at 20°C
vii
-------
SECTION I
CONCLUS IONS
The field tests of the mechanical surface aerator reinforced previous
conclusions as to the suitability of this equipment for raising dissolved
oxygen levels in small polluted rivers. Results as to oxygen transfer
rate were generally consistent with those obtained during tests of pre-
vious years. Correlation appeared to exist between the oxygen transfer
rate and the river discharge for the observed flow range, as had been
noted in previous tests. A somewhat higher oxygen transfer rate was ob-
tained by using a flow concentration device, which represented a pair of
flow concentration groins or low spur dikes employed with surface aera-
tors in a permanent installation. Tests in shallow water (about 7 feet)
were insufficient to indicate a change in effectiveness as compared with
deeper water. Tests of oxygen diffusers were carried out only to a
limited extent due to mechanical difficulties encountered by the Martin
Marietta Corporation in the installation and initial operation. No use-
ful data were obtained for the dynamic oxygen diffuser. The static
oxygen diffuser tests, carried out with only a few diffuser plates in
operation, indicated that most of the fine-bubble oxygen produced was
being absorbed by the river. The spent gas rising to the surface had as
much as Sl% nitrogen in it, reflecting the high degree of absorption of
the oxygen and the stripping of nitrogen from the water during the process.
A dispersion test, carried out with fluorescent dye, verified the highly
dispersive action of the mechanical surface aerator in the relatively
slow, shallow Passaic River. The turbulence generated by thi% aerator
resulted in a longitudinal dispersion coefficient of S.L . x 10 ft 2 /niin,
which exceeds by a factor of five the largest “natural” coefficient yet
reported. On account of this high dispersion, the leading edge travel
distances were found to exceed the mean values by O% at the upper of the
two stations and by 30% at the lower.
The data obtained from the 1970 tests were used to recheck previously
mathematical modelling of the Passaic River. In particular, analysis was
directed at determining whether the operation of the mechanical aerator
had any effect upon deoxygenation parameters, as had appeared to be the
case in tests made during 1968. For 1970, there does not appear to have
been any such relationship. Special attention was also given to nitri-
fication effects. Based upon statistical indications only, it would
appear that appreciable nitrification was occurring during the tests;
but nitrogen balances indicated that this was not the case. These ques-
tions as to nitrogenous BOB and deoxygenation parameters in an artifi-
cially aerated stream are being studied further in connection with
continuing research projects.
1
-------
SECTION II
RECO NDATIONS
After consideration of results obtained, it is recoimnended:
That consideration be given to use of flow concentration groins or low
rock spur dikes for use with any future permanent niechanical aeration
installation in swtall polluted rivers.
That further research be conducted into the practicability and feasi-
bility of adding gaseous oxygen to polluted rivers, as an alternative
to artificial aeration.
That research be continued as to infLuences, incLuding mechanical aera-
tion, which my affect the parameters of biochemical oxygen demand.
3
-------
SECTION III
IWEPLODUCTION
The Water Resources Research Institute of Rutgers University has been
engaged upon research and testing of means of instream aeration for sev-
eral years, with financial support from the Environmental Protection
Agency, the Office of Water Resources Research and the State of New Jersey.
The first phase of this work indicated that mechanical aeration of small
rivers would provide an economical alternative to advanced waste treat-
tnent for obtaining desired dissolved oxygen standards. This work has been
fully reported upon (l)(2)(3)(L .)(S).
The second phase of testing, carried on in 1969 and 1970 on the Delaware
River, resulted in development of design considerations for instreatu
aeration on large, navigable rivers, and showed that, for the critical
area of the Delaware Estuary, instreamn aeration would provide a viable
solution to supplement secondary treatment of wastes. A report covering
this phase of the work has been forwarded to the Environmental Protection
Agency (6).
The present report covers Phase II of the project “Oxygen Regeneration of
Polluted Rivers.” This phase of work was planned (a) to test performance
of aerators with flow concentration devices, and in shallow water, (b) to
facilitate research concerning certain phenomena of deoxygenation rate
which had previously been observed downstream of river aerators, and
(c) to test surface aerators on a comparative basis with diffusers of
pure oxygen.
The testing of flow concentration devices with mechanical aerators was
suggested after tests in 1967 and l9b8, from which it was observed that
surface aerator had considerably greater oxygen transfer rates at higher
river velocities. A research project has been carried out, which has
demonstrated by hydraulic model tests that higher river velocities can
be created at mnidstreamn at an aeration site by low rock spur dikes, or
flow concentration groins, extending out from the bank at each side (7).
For the present project, the permanent rock dikes were simulated by a
long canvas strip, weighted at the bottom, and with a gap in the center.
This arrangement is referred to as a flow concentration device.
Research concerning the parameter of deoxygenation rate in artificially
aerated rivers is being conducted under Project B—027-N.J. of the Office
of Water Resources Research, “Instream Aeration and Parameters of
Nitrogenous BUD.” The operation of surface aerators under the present
project was essential to provide test facilities for the first phase of
that research. The research activity will continue for two more years,
and the aspects of parameter variation and mathematical modelling cover-
ed in this report are of an interim nature. In connection with the
analysis of effects of the aerators, dispersion tests were made, using
fluorescent dye, and a dispersion coefficient computed.
-------
Field tests were conducted during the summer of 1970 at the previously
used test site on the upper Passaic River, and the same mechanical aera-
tor was installed. Arrangements were made for the necessary water quality
observations, for the installation of the flow concentration device, and
the shifting of the aerator to shallow water. The weather was fovorable,
and some excellent results were obtained for the aerator. There was an
interruption in the tests, due to an unknown object, probably a floating
log, which entered the impeller, seriously damaging it, and causing a
delay for repairs.
Meanwhile, as planned, the Martin Marietta Corporation brought to the
test site the supplies and equipment for installing two types of oxygen
diffusers, and provided a large liquid oxygen reservoir, with a heat ex-
changer and other necessary controls. A schedule was worked out to test
the oxygen diffusers alternately with the aerators, in order to obtain
comparable results. Certain of the results are of considerable interest;
but, unfortunately, mechanical difficulties in the course of installation
of the oxygen diffusers precluded obtaining sufficient data for a mean-.
ingful comparison of transfer rate with that of aerators.
The test results for both aerators and oxygen diffusers were analyzed
by Rutgers personnel and are summarized in this report.
6
-------
SECTION I V
FIELD OPERATIONS
The location chosen for the field operations is about 100 yards below
the point where U.S. Highway )16 crosses the Passaic River, in the village
of Pine Brook, Montville Township, Norris County, New Jersey. It is the
site designated as No. 1 in the previous report (1).
At this site, two rental trailers were installed by the Rutgers research
team, an 8’ x 20’ office and laboratory trailer, and an 8’ x 0 ’ storage
trailer. The office and laboratory trailer a’e connected to electrical
and telephone lines, and had a refrigerator for the storage of BOD
samples. The storage trailer was without power but had ample space for
the storage of tools, equipment, and outboard motors. Both trailers
were used in common by Rutgers and Martin Marietta personnel.
Three different aerators were installed at the site for study and com-
parison: (a) the same Yeomans mechanical aerator used by the Rutgers
research team during the summers of 1967, 68 and 69; and (b), (c), two
different oxygen diffuser—type aerators by the Martin Marietta team.
River Bottom Conditions
The 10 foot deep basin, about lOOT x 100’ in area, excavated for the
surface aerator in 1967 and 1968 was partly filled by sediments. Most
of the excavation work done in 1970 was required for a 12 foot deep
trench for the Martin Marietta aerators, about 10’ wide (measured in
the downstream direction), dug transversely just downstream of the 100’
x 100’ basin. A contour survey of the bottom after completion of ex-
cavation is shown in Figure 1. This shows depths at a stage of 221 cfs.
The surface aerator site is shown by a cross, the diffuser site by a
dotted line.
Yeomans Aerator Installation
The Yeomans mechanical aerator was installed as nearly as possible at
the same location as in 1967 and 196b.
This aerator (Figure 2) has 3 vertical spade-like steel blades attached
to a circular impeller rotating horizontally in the water at an eleva-
tion such that the blades are partially submerged when stationary. The
impeller is supported and driven at a speed of 30 revolutions per minute
by a vertical shaft running up to a 75 horsepower speed reducer and elec-
tric meter. When the machine is operating, the continuous movement of
the blades in their circular path throws up and outwards a heavy cascade
of water, greatly increasing both the surface available for oxygen trans-
fer and the turbulence of the water.
7
-------
East bank
cx
120
80 40 0 —40 —80 —120
FIG. 1 TEST SITE DEPTHS
-------
* — .
; :; ‘;
—--.,
• - , e —
. . .‘ . -
-r.. ‘ -
;• T
FIG. 2 MECHANICAL AERATOR IN OPERATION
‘0
-------
The aerator machinery is supported by 3 floating pontoons. This keeps
the blades at a fairly constant level with respect to the water, regard-
less of the constantly fluctuating level of the river.
Constant blade submergence is extremely important. Too little submergence,
even 1” less than the correct value, would result in less oxygen input
than the design capability of the machine. Too much submergence, even 1”
excess, would overload the motor and speed reducer continuously, thus re-
ducing the number of years’ service obtainable from the investment.
The submergence of the blades can be adjusted in either of two ways,
(a) by adjusting a set of 8 large bolts connecting the impeller to the
drive shaft, or (b) by water addition to or removal from the floating
support pontoons, which have a top inlet for that purpose.
The floating aerator was energized by a 3-phase, 2)40—volt power cable
running from the public utility power supply on shore near the job site.
A disconnect switch and starting equipment were installed in a small
wooden shelter on shore; a disconnect was also installed on the aerator
for personnel safety during necessary maintenance work.
The Martin Marietta oxygen diffusers were placed horizontally on the
bottom of the stream and were supplied with gaseous oxygen through plastic
tubes from shore. The following is a technical description of one unit,
furnished by the Martin Marietta Corporation supervising engineer: “The
passive diffuser assembly consists of six (6) separate trays each contain-
ing eleven (11) full blocks of aluminum oxide ceramic plus a partial
block to complete the tray (see Fig. 3). These trays are rigidly fixed
to a sheet of 31)4” marine grade plywood and an aluminum frame constructed
of 3-inch channel section. Beneath each tray at each end is a rigid
poly-vinyl—chioride tube fitting. These fittings are interconnected with
flexible poly—vinyl—chioride tubing. The entire assembly (six trays) is
supplied nth oxygen through a single check valve and a flexible poly-
vinyl-chloride tube which connects to the oxygen header on the shore.
Oxygen permeates the cavity beneath the aluminum oxide blocks and passes
through the pores to form the desired bubble size at the block-water
interface.” Because of th ir fineness, the individual pores were in-
visible. The porous ceramic plates caused the oxygen to emerge into the
water in extremely fine bubbles, resulting in a slow rate of bubble rise,
high specific surface, and a high percentage of absorption of the oxygen
by the water.
The Martin Marietta dynamic oxygen diffusers consisted of fine—orifice
metal diffuser plates; the orifices were larger, however, than those in
the ceramic plates. The diffuser surface was vertical with respect to
the water surface. A large centrifugal pump and piping were installed
and connected to all diffusers to obtain high-velocity jets of water
sweeping horizontally past the openings on the vertical diffuser plates.
The purpose of the water jets was to reduce the average diameter of the
oxygen bubbles by shearing action. Thus a slow rate of bubble rise,
high specific surface, and high percentage of oxygen transfer should be
10
-------
1 -
• — —, —. 1111I 1 i
‘ —I
I —i
—___-- 1•
FIG. 3 MARTIN MARIETTA STATIC OXYGEN DIFFIJS
-------
obtained, just as with the porous ceratnic plates. A description of this
equipment by the Martin Marietta representative is as follows: “The
active diffuser assembly is a private invention and consists of a water
inlet pipe which feeds five (5) rectangular cavities rigidly connected
to the pipe. These cavities are each covered with eight (8) vertical
gas distribution bars which expel oxygen in a lateral direction through
capillary size tubes. Water is expelled through slots between adjacent
gas distribution bars in a manner which shears gas bubbles at the capil-
lary end before the gas bubble can reach maturity. Oxygen is supplied
to each diffuser assembly through a check valve and a flexible poly-
vinyl—chloride tube which is connected to the oxygen header on the shore.”
The dynamic diffuser is shown in Figure Li..
Oxygen Supply and Control
A liquid oxygen tank and an atmosphere-heated vaporizer with orifice
type flow recovery, flow control valves, temperature recorder, pressure
indicators, and safety devices, were provided by the Martin Marietta
Corporation to furnish substantially pure gaseous oxygen for the opera-
tion of their two diffuser aerators. The oxygen and tank were supplied
by the Burdett Oxygen Company.
The oxygen supply stream, after being vaporized and metered, was piped
to a header with multiple valves, one for each diffuser unit. From the
valve header, an individual flexible tube of 5/8” x 3/8” clear vinyl
plastic ran to each of the underwater diffuser units. Thus the oxygen
supply to each unit in the stream could be controlled or shut off from
the shore. Only the total flow of oxygen was metered, however.
Chemical for Algae Control
The growth of algae was anticipated as a source of plugging of the small
oxygen outlets in the diffuser aerators. Therefore supplies of chlorine
arid hydrogen peroxide, an a rotameter for measuring their flow rates
were incorporated into the experimental equipment.
Yeoman’s Aerator Operation and Data Collection
The Yeoman aerator was operated continuously for periods of time from
three to five days, and toward the end of each period one or more dissolv-
ed oxygen traverses of the stream were made. For comparison, the aerator
was also shut off for periods of three to five days, and identical traverses
were made. During July and August, 1970, the Passaic River was very close
to steady—state conditions, as required for the mathematical simulation
of the oxygen sag curve.
The dissolved oxygen traverses during the 1970 tests were primarily longi-
tudinal, taken by motorboat at various points from 500’ upstream of the
aerator to 50,000’ downstream. Over this reach, the river has no tribu-
taries. During these traverses, a single reading of dissolved oxygen
was taken one foot deep at the centerline of the stream; previous experience
12
-------
t
FIG. L MARTIN MAR IETTA DYNAMIC OXYGEN DIFFUSER
13
-------
has indicated that this reading is very close to the weighted mean value.
The river is three to five feet deep at centerline, and 100 to 110 feet
wide. The dissolved oxygen readings were taken with a Delta Model 8
meter, calibrated just before the traverse by Winkler titration of a
sample of river water. Also, Winkler samples were taken as a check on
1/3 to 1/2 the meter readings. when Winkler samples were taken in the
boat, they were fixed immediately using chemicals carried in the boat
and transported to the office-laboratory trailer for titration.
All readings and samples were taken frotn the boat while heading upstream,
to minimize errors due to possible local aeration of the water by the
outboard motor.
The surface aerator had operated for three previous test seasons without
serious difficulty with drift, since the strong surface current repells
any floating objects. However, on August 11, the impeller was seriously
damaged by an object believed to have been a submerged log. Eleven
blades were broken or seriously bent; and it was necessary to suspend
operations until August 21 for repairs. In addition to the delay, it
was apparent that the alignment of the straightened blades was no longer
as regular as the original. This irregularity affected efficiency of
the aerator somewhat, as described in the following section.
Flow Concentration Device Tests - Surface Aerator
An important question with surface aerators is the following. Would
concentration of the total flow of the stream directly through the aera-
tion zone increase the oxygen input in lbs 02 per horsepower hour?
To get an actual test result, a heavy canvas rectangle, 110 feet long
by 8 feet wide was fabricated, with a 20—foot-long oval opening at the
centerline. This was stretched across the stream about )40 feet upstream
from the aerator, to concentrate the total stream flow through the open-
ing and directly into the aerator’s active zone. With the aerator
operating, lateral oxygen traverses were made upstream of the device, and
downstream of the aerator. Great difficulty was encountered in weighting
the lower edge of the baffle sufficiently, and considerable flows escaped
below the canvas.
Results of this test are given in the next section.
Shallow Water Tests — Yeoman’s Aerator
To test the efrectiveness of the Yeoman’s aerator in shallow water, the
machine was towed upstream about forty feet. This was possible because
the aerator was on floating pontoons, and because the anchor cables and
power supply cable were easy to rearrange in new positions.
The aerator was then operated and lateral oxygen traverses were taken,
upstream and downstream of the aerator. In view of time which had been
lost in the accident, as described above, little time remained for tests
-------
in this location.
Results of these tests are also given in the next section.
Martin Marietta Aerator - Operation and Data Collection
The passive oxygen diffuser plates were subject to some degree of leakage
under oxygen pressure. Those plates which would operate without leaks
were placed on the bottom of the stream and each was connected to its
own oxygen supply tube. With a constant rate of flow of oxygen, steady
state was reached, two lateral oxygen traverses were then taken, one
upstream and one downstream of the diffuser plates. Oxygen contents of
the water at various depths were measured by Delta meter and Winkler
titration, and the total oxygen input was calculated by graphical inte-
gration. The oxygen flow was measured by means of the chlorine rota-
meter rather than the orifice meter, because the reduced number of passive
diffuser plates required only a small oxygen flow rate.
The percent of the pure oxygen absorbed by the stream was obtained by
collecting samples of the residual gas bubbling up in the surface from
the passive diffuser plates. The samples were analyzed for oxygen and
nitrogen on a Beckman gas chromatograph. Since the initial impurity con—
stant of the oxygen was known, the percent of 02 absorbed could be
calculated by mass balance from the analysis of the spent gas.
No data were obtained from the dynamic oxygen diffuser during the summer
of 1970 due to mechanical difficulties.
Details of the tests and of results obtained are given in the next
section.
-------
SECTION V
OXYGEN TRANSFER OF AERATORS AND OXYGEN DIE”FIJSERS
In this section, the oxygen transfer results of the surface aerator are
outlined. The oxygen transfer rates before August 11, when the impeller
was damaged, should be comparable to those previously obtained by the
same aerator on both the Passaic and the Delaware Rivers. The transfer
rates after this date are not strictly comparable with earlier periods,
since the repaired aerator was obviously operating with less than its
original smoothness; but comparisons are valid between normal operation
in a deep pool, operation with and without a flow concentration device
(on the same day), and operation in shallower water.
Oxygen Transfer Efficiency
The performance of an artificial aeration unit is characterized by the
oxygen transfer rate, or the oxygenation efficiency, which can be gen-
erally expressed as the amount of oxygen in pounds added by the aerator
per shaft horsepower-hour. In a flowing stream with relatively low
oxygen demand the oxygen transfer rate for an aerator operating under
steady conditions may be described by the following equation (1):
BQ(CdCu)
Rt=
where
= oxygen transfer rate at test conditions in lbs 0 2 /hp—br
= units conversion factor
Q = river discharge in cubic feet per second
Cd = DO concentration at downstream (of the aerator) sampling point
in rag/l
Cu = DO concentration at upstream (of the aerator) sampling point in
nig/l
P = power developed by the aerator in shaft horsepower
The oxygen transfer rate determined by Equation -l is specific for the
conditions that obtained during a particular test. For the purpose of
comparison this rate must be converted to standard conditions, which are
generally taken as 200C, 760 mm Hg atmospheric pressure, 0 mg/i DO concen-
tration, and “clean” water quality. The conversion is obtained by (3)
(c )2o
R 5 =Rt [ I
[ (Cs)T Cm] (TF)(a)
in which
= oxygen transfer rate at standard conditions
(C 0 = saturation DO concentration at 20°C
17
-------
(C 5 ) saturation DO concentration at test water temperature T
Ti = temperature correction factor - (l.025)T20 (10)
= specific DO solubility
= specific transfer rate
Cm = DO concentration at the aeratar
Rt is as defined earlier
According to laboratory test results by Hunter (8) both the specific DO
solubility, , and the specific transfer rate, ct , were found to be
1.0, and the saturation oxygen concentration values, C , were computed
with the following SED—ASCE equation.
= iJ .652 - 0.I l022T + 0.007991T 2 - 0.000077ThT 3
For a temperature of 20°C, Cs = 9.02 mg/i.
During field operations, it is rather difficult to measure the exact
value of the oxygen concentration, Cm, at the aerator due to the intensive
turbulence created by the aerator. Consequently, the driving force C 5 - Cm
could not be accurately determined, though several methods have been pro-.
posed in the literature to estimate it (9). According to previous ex—
periences on the Passaic River (1), the arithmetic average of the DO read-
ings at the upstream and the downstream sampling points provided satis-
factory oxygen transfer results. Therefore, in this study the Cm values
were estimated by
Cm = 1/2 (C + Cd)
where C and Cd are as defined earlier.
Oxygen Supplied by the Aerator
The rate of oxygen added by the aerator to the stream for steady-state
conditions can be computed by the relationship
U = 0.22)46Q(Cd - Cu) 5-5
where U = oxygen added or uptake rate due to the aerator in lbs O 2 /hr,
and Q, Cd and C are as defined earlier.
In the present study the oxygen uptake rates were determined by both
longitudinal and lateral DO traverses. For the longitudinal traverses
DO readings were taken at 200 ft upstream, and 100 ft downstream of the
aerator, at the depth of 1 foot along the centerline of the river. It
has been shown (1) that the one-foot, centerline reading could be used
satisfactorily to represent the average DO concentration over the whole
river cross-section. For the lateral traverses DO and velocity readings
were taken at various depths across the river at 200 ft upstream, and
100 ft downstream of the aerator. The mean DO for a cross-section was
computed by using averages weighted according to stream velocity readings.
It should be noted that all the DO measurements made with the Delta DO
18
-------
meter were corrected against corresponding Winkler readings, before the
mean DO values were computed.
A summary of the field tests data, together with the computed oxygen up-
take rates is given in Table -l.
Oxygen Transfer Rates
The oxygen transfer rates, Rt, under steady-state conciltions for the
mechanical aerator were computed by dividing the oxygen uptake rate, U,
by the shaft power consumed, as indicated by Equation -l. Equation —2
was then used to convert these transfer rates to standard conditions.
In order to facilitate the computation, the following relationship is
defined:
R =RtxF
S
where
F = conversion factor from test to stanciard conditions.
Since F can be expressed as a function of the following:
F f(Cm, C 5 , T, ,
for given values of C 5 , , and , F is a function of Cm and T. In this
study, F is written as
9.02
(Cs_Cm)(l•025)T_ 2 O
Thus values of F for different Cm and T can be computed ana shown graphi-.
cally. As illustrated in Figure for observed temperature and esti-
mated Cm values, F can be found easily and the computation is simplified.
The results on oxygen transfer rates at both field and standard conditions
for all the tests are listed in Table —2. Also given in Table -2 are
the dates of tests, the flow rate, and the DO concentrations upstream of
the aerator.
Comparison of Oxygen Transfer Rates
The field test results for all the experinients, excluding test with the
flow concentration baffle and with aerator in shallow water, indicate
that, over the observed flow range (107 to 268 cfs), the aerator trans-
ferred on the average l. O lbs 0 2 /hp—br under standard conditions. How-
ever, as described in Section IV, the aerator was damaged on August 11,
1970. Although repairs were made, several of the blades of the impeller
were somewhat bent. If only the data before August 11 were averaged,
the transfer rate became 1.69 lbs Op/hp-hI’, which is 9)4% of the figure
19
-------
TABLE 5-1
SUMMARY OF STEADY-STATE FIELD TEBT DATA FCR
NECHANICAL AERAT
Date
(1)
Flow Q.
in cubic
feet per
second
(2)
Water
temp. in
degrees
centigrade
(3)
Shaft
power in
horse-
power
)
Dissolved Oxygena
in milligrams
per liter
Oxygen added
in lbs of
oxygen
per hour
(7)
Upstream Downstream
(5) (6)
7/13/70 135 23.0 81.8 2.80 5.70 88.1
7/15 12I 21.0 65.6 2.30 L .2O 52.9
7/15 126 22.0 65.3 2.10 3.90 50.9
7/20 113 23.0 70.5 2.20 t .9O 68.5
7/20 113 21i.0 69.0 3.bO 5.70 58.3
7/20 115 23.5 67.0 3.L O 5.50
7/27 113 2t .3 7L .6 0.20 3.70 89.0
7/27 113 2 .5 77.0 0.80 5.10 109.2
7/28 111 25.0 65.0 2.20 5.10 71.6
7/28 111 25.3 5L .6 2.30 ) .5O 61.7
8/11 111 21.0 1 L 6 .2 2.bO IL.OO 39.9
8/12 109 2b.0 5I .7 2.60 L .5o L 6.t
8/12 109 21i.0 55.1 3.30 LL.30 2L .5
8/19 107 22.0 65.0 2.60 ). .70 50.7
8/19 107 22.0 62.5 2.90 li.80
8/20 111 23.0 63.5 3.00 li.90 147.9
8/25 268 20.0 66.0 2.814 3.814 60.3
8 / 25 b 236 22.0 66.5 3.05 14.16 58.8
8125 b 236 22.0 66.3 3.05 14.15 58.3
8/25 223 22.0 66.1 3.30 14.17 143.6
8/26 163 20.0 68.1 2.61 3.96 1 49. 1 4
8/26 153 21.0 68.8 2.81 14.50 58.1
8/26 151 21.0 68.8 2.83 1 4.51 57.0
8 / 26 b ‘145 22.0 69.14 2.98 14.62 53.14
126 23.0 72.5 2.70 14.70 56.6
126 23.0 71.0 2.70 14.70 56.6
8 / 27 C 125 23.0 72.14 2.70 14.60 53.3
a DO samples taken at 200 feet upstream and 100 feet downstream of the aerator.
b With flow concentration device installed at 100 feet upstream of the aerator.
C Aeratcr in shallow water, at 140 feet upstream of the original site.
20
-------
TABLE 5-2
SUMMARY OF OXYGEN TRft NSFER RATES F(Jt MECHANICAL AERATON.
UNDER S TE.ADY-S TATE OPERATION
averaged 0.88 1.50
a
b
C
d
Rt = oxygen transfer rate under test conditions;
R 5 = oxygen transfer rate under standard conditions.
With flow concentration baffle in place.
Aerator in shallow water.
Excluding cases when flow concentration baffle was
in place or aerator in shallow water.
- Upstrea!n DO
concentration,
Rt in lbs of
oxygen per
H 5 in lbs of
oxygen per
milligranis
per
horsepower—
horsepower-
liter
ho a
houra
(3)
(14)
(5)
Flow, Q,
in cubic
feet per
second
(2)
13S
1214
126
113
113
11 5
113
113
111
111
111
109
109
107
107
111
268
236
236
223
163
1 3
151
1 1 45
126
126
12
Date
(1)
7/13/70
7/15
7/15
7/20
7/20
7/20
7/27
7/27
7/28
7/28
8/il
8/12
8/12
8/19
8/19
8/20
8/25
8/2
8/2 5 b
8/2 5
8/26
8/26
8/26
S/ 2 6b
8/ 27 C
8/2 7 c
8/27 C
2.80
2.30
2.10
2.20
3.140
3.140
0.20
0.80
2.20
2.30
2.140
2.60
3.30
2.60
2 .93
3.00
2 .8)4
3.05
3.05
3.30
2.61
2.81
2 . 83
2.98
2.70
2.70
2.70
1.08
0.81
0.78
0.97
0.85
0.81
1.19
1.142
1.10
1.13
0.86
0.85
O .
0.78
0.72
0.76
0.92
0.89
0.88
0.66
0.73
0.8)4
0.83
0.77
0.78
0.80
0.7)4
2.12
1.27
1.18
1.62
1.81
1.71
l.S1
2.17
1.98
1.90
1.3)4
1.143
0.82
1.30
1.30
1.140
iJ 5
1.51
1.50
1.1)4
1.15
1.242
1.36
1.37
1.39
1.27
21
-------
FIG. CONVERSION FACTOR, F, VS WATER TEMPERATURE, T , AND
r\)
0
z
C
z
C
0
3.5
3.0
2. 5
2.0
1.5
1.0
0. 5
5 10 15 20 25 30 35
D.O. LEVEL AT THE AERATOR, Cm
-------
obtained during the 1968-69 experiments. (1.80 lbs 0 2 /hp-br on the aver-
age). The average after August 11 was 1.29 lbs 02/hp-br, indicating
clearly the effect of the impeller damage on the transfer efficiency
(see Figure 6). Correlation between oxygen transfer rate and river dis-
charge over the observt d flow range as shown in Figure 6 indicates slightly
higher transfer rates at the higher rates of flow. Also the figure shows
that between the flow range of 100 to 130 cfs, there is a wide range of
variation in oxygen transfer rate but the indicated variation is less if
the figures before and after August 11 are considered separately. In
addition to the effects of the damaged impeller, several possible factors
might be responsible for variation, namely, changes in the cr and fac-
tors, effects of dispersion, short time or uncorrected irregularities in
power consumption fluctuations (1), and sampling errors.
Flow Concentration Device Tests
As described in Section TV, the flow concentration tests were conducted
on August 25 and August 26, 1970. The device was placed at 100 ft up-
stream of the aerator. DO and velocity readings, as in the other tests,
were measured at 200 ft upstream and 100 ft downstream of the aerator.
Readings were taken on the same day with and without the device. The
field data are listed in Table 5-1, and the results of oxygen transfer
are given in Table 5-2.
The results indicate that, for the August 25 field test, the average oxy-
gen transfer rate with flow concentration was 1.50 lbs 0 2 /hp-br, which
was about 15 percent higher than the corresponding average transfer rate
of 1.30 lbs 02/hp-br without the device on the same day. However, for
the August 26 test, the figure of 1.36 lbs 0 2 /hp-hr with flow concentra-
tion was only slightly higher than the average of l.3L lbs 0 2 /hp—hr with-
out flow concentration. When the flow concentration device was in place,
it was observed that the stream velocity downstream of the central open-
ing showed a moderate increase. The increase in the velocity and hence
the level of turbulence should bring about an increase in the oxygen
transfer of the aerator. The fact that the device was not as tight as
was expected (water was passing through both ends and also underneath)
might be responsible for the relatively small increase noted in the trans-
fer rates. The device was installed 100 feet upstream of the aerator,
and bringing it closer would probably have increased its effectiveness.
Shallow Water Tests
The shallow water tests were conducted on August 27, 1970. As described
earlier in Section IV, the aerator was moved upstream about 0 feet from
the original site. Hydrographic surveys had shown that the mean water
depth near the aerator was approximately 7 feet during low flows at the
new site, whereas the corresponding depth at the original site was about
12 feet (see Figure 1). Three sets of lateral DO and velocity traverses
were obtained. The basic data and the oxygen transfer results are given,
respectively, in Tables 5—1 and 5-2. The average oxygen transfer rate
far the shallow water tests was l.3L lbs 0 2 /hp-br under standard condi-
23
-------
2.5
1.0
I
0
02
• Normal operation prior to Aug. 11, 1970
O Normal operation after Aug. ii, 1970
)( With flow concentration device
A Aerator in shallow water
0.5
50
100
0
150 200
Flow Rate in cfs
250
300 350
Oxygen transfer
I
rate,
standard condi-
tions,
lbs 0 2 /hp-hr.
. S
2. 0
1.5
S
S
S
r’)
x
a
— U
0
‘C
0
Mean relationship
after Aug. 11, 1970
FIG. 6
OXYGEN TRANSFER RATE AND DISCHARGE
-------
tions for discharges of 125 to 126 cfs. The efficiency was only slightly
different from the deeper water average of 1.29 lbs 0 2 /hp-hr observed
after August II for a flow range of 107 to 163 cfs.
Summary of Mechanical Aerator Results
Field tests conducted on the upper Passaic River during the summer of
1970 prior to the damage to the impeller resulted in an oxygen transfer
rate of 1.69 lbs 0 2 /hp—hr under standard conditions for the mechanical
aerator. This is about 914% of the performance observed in previous years.
Correlation appeared to exist between the oxygen transfer rate anci river
discharge for the observed flow range as had been noted in previous tests,
but the number of observations and range of flows experienced were so
limited as not to give conclusive verification.
The flow concentration device seemed to cause an increase in the stream
velocity downstream and thus the oxygen transfer efficiency of the aera—
tor. In order to be most effective, the device should be placed closer
to the aerator and should be constructed so as to avoid major leakage.
The oxygen transfer rate observed when the aerator was operating in
shallow (7 ft deep) water, was slightly greater than that with the deep-
er (12 ft) water but the differences are within the probable error of
such observations.
c cygen Diffuser Tests - Martin Marietta Corporation Equipment
Due to mechanical difficulties in the installation of the equipment, no
conclusive tests on the dynamic type diffuser were conducted, and the
static type equipment was tested in partially operative form, releasing
only a fraction of the oxygen originally contemplated. This report in.-
cludes results observed on all of the tests which produced significant
results. Equipment is described in Section IV.
For most of the tests, lateral DO traverses were taken at river sections
10 feet above and 20 feet and 30 feet below the diffuser plates. For
each section, DO and velocity readings were obtained for 10 to l points
at various positions and depths within 30 feet of the East bank. It was
observed that the effect of the operative portion of the diffuser during
the sampling times was confined to this area. On two occasions, when
approximate results were required quickly, longitudinal DO and velocity
traverses were taken at points 10 feet upstream and 80 feet downstream
of the diffuser plates, instead of the lateral traverses.
In addition, air and water temperatures were also taken during each tra-
verse. The discharge of the river was obtained from gage readings of a
U.S. Geological Survey for Pine Brook N.J. by means of a rating table.
The gage was about 100 yards upstream of the test site.
The oxygen transfer rates of the diffuser plates, in terms of pounds of
oxygen per hour, can be obtained by a mass balance computation. Under
25
-------
steady state conditions, the transfer rate of diffuser plates operating
in a stream can be evaluated by the following equation:
R = O.22) ôQ(Cd - c ) 5-9
in which
R = oxygen transfer rate in pounds of oxygen per hour
Q = river discharge in cubic feet per second
Cd = DO concentration at downstream sampling section in mg/i, and
Cu = DO concentration at upstream sampling section in mg/i.
Equation 5-9 implies that the amount of oxygen per hour supplied by the
diffuser is proportional to the difference between the concentrations of
oxygen downstream and upstream of the diffuser. In this equation the
contributions from other oxygen sources or sinks, such as atmospheric
reaeration, are considered negligible.
For the above equation to be strictly valid, it is necessary that (a) the
dissolved oxygen concentrations be uniform throughout the upstream and
downstream sampling cross-sections, and (b) the capacity of the diffuser
must be such that the entire downstream cross-section is affected by the
aerator at the time samples are taken. However, during the experiments
in question, these conditions did not hold. Not only was the distribu-
tion of oxygen uneven, the velocities of the two sections involved some-
what different quantities of flow and degrees of turbulence. Therefore,
the oxygen transfer was computed by means of the following equations:
R = C’V t A’ -ECV A 5-10
where primes note conditions downstream of the diffuser. The basic
assumption was made that ZV’&A’ = L V Pi, or that the discharge recorded
as passing through the downstream section was the same as that entering
through the upstream section. Weighted mean values of C and CT were
computed, weighting each DO reading proportionately to velocity observed
at that point. The equation then reduced to a close approximation:
R ( - )zv’ A’ 5-li
The amount of oxygen transfer was computed by Equation 5-9 for the longi-
tudinal traverse results, and by Equation 5-il for the lateral traverse
results. The results together with dates, flow rate, and water tempera-
tures are given in Table 5-3. It should be noted that in using Equations
5—10 and 5-11, the DO concentration at the upstream cross-section was
assun d to be the same as DO concentration at the downstream cross-
section before the aerator was turned on. The assumption seems reasonable
since the two cross-sections are fairly close to each other 4O ft), and
the sag curve in the vicinity was flat.
In general, the results are inconclusive. The July 9 and July 30 results
are understood to be of questionable value due to imperfectly functioning
26
-------
TABLE 5-3
SUMMARY OF OXYGEN TRANSFER DATA
D IFFUSER PLATES
NOTE: All DO data were obtained by lateral traverses,
except for August 3 and August 14, on which
longitudinal traverses were used.
Date
Time
9:15 a.m.
1:00 p.m.
1:00 p.m.
Upstream
DO
Cu
nig/l
Downstream
DO
Cd
mg/i
Oxygen
transfer
lbs/br
2.00
2.70
18.9
3.80
14.10
8.1
1.03
1.09
2.6
7/9/70
7/9
7/30
8/3
8/14
8/6
8/6
8/7
8/7
8/7
Flow
cfs
120
120
190
256
221
152
1 51
136
136
136
Water
temp.
°C
23.0
23.5
214.0
25.0
22.9
22.0
22.0
23.0
23.0
23.0
3:00
9:00
10:00
2 :145
9:15
1:00
3:00
p.m.
a. m.
a • m.
p.m.
a. rn.
p.m.
p.m.
1.75
1.16
1.95
2 .114
2 .07
2.38
2 .56
1.75
1.147
2.014
2.1414
2 .15
2.62
2.79
0.0
15.14
2.0
8.9
2.7
14.2
3.6
27
-------
diffusers. Also the longitudinal DO traverse provides only one DO read-
ing for each cross—section. Hence the results are less reliable than
those obtained with the lateral traverses.
If only the results from lateral traverses on August 6 and 7 are con-
sidered, an average oxygen transfer rate of L .28 pounds per hour was
obtained. This number seems to fall within the limits of the amount of
oxygen supplied. However, a reading of 8.9 pounds per hour was computed
far August 6, which actually exceeds the amount supplied. The accuracy
of the DO meters and current meters was such that measurement errors of
up to b lbs/br might easily occur. The measurement system was planned
to measure inputs of oxygen of the order of 7S pounds per hour, and
accuracy cannot be assured under conditions such as those reported on
here.
Proportions of Oxygen in Diffuser Bubbles and Oxygen Diffuser Efficiency
Samples were taken of the oxygen being used for the diffuser, and of the
bubbles rising to the surface from the diffuser, in order to determine
the proportion of oxygen in each. The original gas being bubbled was
99.LL% pure oxygen. Table shows percentages of nitrogen in the spent
gases arising to the surface from the oxygen diffusers. The water may
be assumed to have had approximately a saturation concentration of nitro—
gen relative to the atmosphere. As the fine bubbles of pure oxygen
entered the water, they obviously absorbed additional nitrogen from the
water, at the same time that most of the oxygen itself was entering the
water. It is apparent that these ultra-fine bubble oxygen diffusers
were operating at a very high degree of absorption.
28
-------
TAB lE
SPENT ( lAS EN.OM OXYGEN DIFFUSERS
Sample No. % % Oxygen plus Argon *
Air Atmosphere 78.3 21.7
#3 23. 76.S
#t 20.8 79.2
#S 22.0 78.0
#6 22.3 77.7
#9 3.0 97.0
#16 2 .9 7L .l
#17 26.L 73.6
#20 b3.2 56.8
#21 51.0
#2) 25.9 7L 1 ..1
#25 25.3 7L 1 .7
* By volume
29
-------
SECTION VI
D]SPERS ION AL!S]B
This section describes a dispersion test to determine qualitatively and
quantitati’v ly the downstream movement of an aerated slug of river water.
Essentially bank—to-bank aeration was effected in the Passaic River owing
to the narrowness of the river and the large mechanical aerator (see
Figure 2). Furthermore, owing to the shallowness in this particular por-
tion of the river, it can be assumed that vertical dispersion of the
dissolved oxygen occurred completely within a very short distance down-
stream of the aerator. As a result, the dispersion phenomena take place
primarily in the downstream direction. References (12) through (27) cover
the theory of dispersive phenomena.
Back ’ound Theory
Immediately after injecting a quantity of dye into a stream, the spread-
ing process that ensues is dominated by what are referred to as convec-
tive transport phenomena. The concentration distributions take on a
conspicuous skewness in this convective region. Initially quite severe,
this skewness diminishes until, at the end of the convective region, the
symmetrical concentration distributions pertaining to one-dimensional,
unsteady diffusion prevail. Once the concentration distributions attain
this symmetrical character, their behavior is predicted by the one-
dimension dispersion model. The different dispersion regions are illus-
trated as follows.
Consider a stream in which a vertical cross-sectional plane of water
oriented perpendicular to the stream direction is instantaneously colored
red. Immediately after this injection, the one-dimensional concentra-
tion distribution exhibits a decided skewness as shown in Figure 7. In
(a) the view is that of a vertical plane parallel to the stream direction;
in (b) are pictured the corresponding concentration distributions where
is the dye concentration averaged over the cross-section of the stream.
Ultimately, the concentration distribution exhibits the gaussian shapes
predicted by the simple model. However, until this occurs, the dye-cloud
spreading phenomenon is convective; following this interval, it becomes
dispersive. It is therefore important to arrange data—taking procedures
so as to record the symmetrical gaussian concentration distributions.
Several schemes have been devised through which the longitudinal dis-
persion coefficient may be determined (see Ref. 17). The one employed in
the present study is the change of moment method.
The three—dimensional, unsteady conservation of mass equation for the
dispersion of a conserved quantity will be developed by decomposing the
velocity field and the concentration of the quantity into a time averaged
and fluctuating quantity. Assuming conservation of mass for the mean
velocity field, we obtain, in Cartesian form
31
-------
C x,t)
? t x
2
2- . -
r c+
L ABL 2
=
—
(a)
FIG. 7 DYE PATTERNS IN A STREAM
+ — (u’c’) — - —(w’c’) +
2
6- 1
where , , are the usual tiTne averaged velocities (13). The quantity
is the time average concentration and is a molecular dispersion
coefficient that is assumed uniform and isotropic. Invoking the Fickian
dispersion relations:
u’c -E
x x
v’c = —E
y y
6-2
w’c’ = -E
z z
ttinltlal
t=t 2
Area A 1
Area A 2
t—t 3
Convective region
A 3
(b)
Dispersive region
32
-------
and by dropping the molecular dispersion terms, which are dominated by
their turbulent dispersion counterparts, the original conservation equa-
tion becomes:
+ = + = 1 -(E + —a- . + 1 -(E . 6-3
t x y z x x x y y y z z z
when = = 0 and the turbulent dispersion is characterized solely by
the uniform coefficient, E , we obtain our one-dimension model
&+U . c.=E 6-Lb
t X X; 2
Boundary conditions for the above equation depend upon the type of in-
jection of the conserved quantity. If, as in the case of dispersion tests
conducted in the present work, the quantity Rhodamine B dye is steadily
injected at a prescribed rate, the conditions are
c(±o,t) = 0
dM
inject — dt
M = Aj c(x,0)dx = PAJE(x,t)dx 6-5
Solutions to the one—dimensional dispersion model, satisfying these con-
ditions are given by
Q. - fx-i(t-T)J 2
= Inject t ______ kE (t- i
Ap( EJ t - T dT 6-6
Experimental determination of the coefficient, E , via the method of
moments (iS) utilizes the transformation:
67
Substituting this into the original equation and taking the partial
derivative with respect to time, we get
6-8
? t
Multiplying by 2 and integrating overall values of produces
2 d = E 2 d 6-9
x -
33
-------
for constant E . Switching the order of operation on the left-hand side
gives
d 2
d = 2Ej 6-10
Defining the variance by
2 5
— 6-11
we obtain the following relationship for E :
6-12
Equation 6-11 may be transformed
f t2
2 0 dt
cYt = dc dt — 6-13
dt
where t is the mean time:
dc
t—dt
Ut
t
61
The longitudinal dispersion coefficient, E , will be determined using the
relations of references (23) tI rough (2 ).
Experimental Results
The experimental determination of the dispersion coefficient, E , was per-
formed using the constant dye-injection rate method. With this method, dye
is injected into the turbulent crown of water in the aerator at an instant
of time and at a constant and proper rate. As this injection continues,
the sampling procedure begins at the upstream station. Here, at given
intervals of time, river samples were dipped and labelled. Spot-checking
of the dye concentration was also performed with a portable Turner fluori-
meter. After observation of the “plateau value” of the dye concentration,
which is determined by dye-injection rate and river discharge, the pro-
cedure is repeated at the downstream station. With the dye front so
monitored through two such stream stations, the finite differencing calcu-
lation determines the longitudinal dispersion coefficient.
The concentration distributions moving through two stream stations are
represented in Figure 8. The corresponding differentiated distributions
are presented in Figure 9. Using a computerized moment method 2 the value
of the longitudinal dispersion coefficient, E , is x 10 6 ft’ /min. This
314
-------
oo,oooO
0
00
0
0
0•
0
0
0
0
0
0
0
100
060
00800000
00000
A0
Ow
0
0
0
0
0
0000810
1____00I
200 300
Time alter Injection, Mm.
0
0000
400
— 20
— 15
— 10
0
a,
0
C
0
C-)
0
0
—.5
0
0
0
o0
0
0
OO I
I I
500
FIG. 8
DYE CLOUD FRONTAL STRUCTURE VS TIME
-------
0.20
00000 0
90 00
0 0
0.15 — 0
000 00 00
00 00
04 0 0 00 0
0 00 000000 0 000
0.10 * 0 0 0 00 000
o 400
0 ’ 0
0 00
0 0
o 0
0.05 0 0 0
0
0 0 0
•0
4 00 O0
0— ’ I I I I I I
100 200 300 400 500
Time after Injection, Mm.
FIG. 9 DIFFERENTIATED DYE CLOUD FRONTAL STRUCTURE
VS TIME
-------
value pertains to this mechanical aerator in this reach of the Passaic
River, as determined by the data taken from two stream stations 6,000 ft
apart. The first station was 1,000 ft (or 20 river widths) downstream of
the aerator.
This experimentally determined dispersion coefficient is slightly in-
flated owing to the effects of a motor boat that passed through the dye
front three times. The circumstance was unavoidable since the downstream
station was inaccessible by land. To minimize the effect on the coeffi-
cient, the boat was run at essentially idle speed--a speed that seldom
was exceeded on this reach of the Passaic because of the danger of strik-
ing submerged logs and other debris.
This high value for the longitudinal dispersion coefficient compares very
well with the coefficient determined downstream from a similarly turbu-
lent aerator in the Delaware River (E = 8.S x 10 6 ft 2 /min)(6). Previous
coefficient values available in the literature (27) indicate that a
“natural” value of 9.6 x l0 ft2/tnin, which is the largest yet determined,
has been measured in the Missouri River flowing at 33 000 cfs. A value
fcr the Delaware estuary of approximately 1.6 x l0Sft’ /tnin has been
determined by Paulson (22).
Photographs of the Passaic aerator in operation reveal the turbulence
imparted by it to the river water (see Fig. 2). It is this turbulence in
this relatively slow (average velocity: 22.3 ft/nun), shallow c) 4 —5 ft)
section of the Passaic that is primarily responsible for a dispersion
coefficient that is almost an order of magnitude greater than any “natu—
ral” value.
It is significant to note from Figures 8 and 9 that the concentration
levels at the dye front protrude downstream ahead of the mean stream
velocity of the cloud. Assuming that the dissolved oxygen levels in the
river behave similarly, it is concluded that an amount of river water
greater than the mean slug quantity has experienced an increased DO.
Considering travel times specifically, Figure 8 shows that the leading
edge precedes the frontal mean by approximately 0% at the upstream sta-
tion. At the downstream station, the leading edge protrudes by about
30% of the mean travel time.
Summary
The experimental results verify the highly dispersive nature of the
mechanical aerator tested in the relatively slow, shallow Passaic River.
Despite the low discharge and shallow depth, the turbulence generatgd
by this aerator causes the longitudinal dispersion of Ex = x 10 ft 2 /
mm to exceed by a factcr of five the largest “natural” coefficient yet
observed. This natural value of 9.5 x lO)ft 2 /niin was observed in the
Missouri River flowing at 33,000 cfs.
Through this excessive longitudinal dispersion, the leading edge travel-
distances are found to exceed the mean travel-distance values by 50% and
30% at the upstream and downstream stations, respectively.
37
-------
SECTION VII
MATHEMATICAL MODEL AND PARAMETERS OF BOB
It is important to determine not only how much oxygen is added to a river
by inuuced aeration, but what the subsequent reaction of the rivar will
be to the addition. Those familiar with the dynamics of biochemical oxy-
gen demand are acutely aware of the many irregularities and anomalies
which are imperfectly encompassed by the Streeter—Pheips relationships;
and of the inadequacies of the usual five-day BOB tests as indicators of
future oxygen demand. Mathematical modelling of the first two years of
aeration tests upon the Passaic River left inferences that parameters of
deoxygenation had been increased markedly coincident with the aeration
process (1). This led to a new research project (8) starting in 1970;
the first phase of which was combined with the demonstration project
herein reported on. For the field tests of 1970, a very careful check
was made, combining nitrogen balances and various techniques of labora-
tory and field BOB analysis in order to provide a sound biochemical
dimension of modelling. These studies, which will be reported on when
completed, indicate that no substantial nitrification was occurring
during the 1970 tests, despite the presence of very heavy ammonia con-
centrations and numbers of nitrifying bacteria. However, mathematical
modelling was conducted to give the best statistical indications as to
the reactions of the river. This modelling was conducted by two separate
approaches, using analog and digital curve—fitting techniques.
Basic Theory
The modified Streeter-Pheips models applicable to the Passaic River are
given by
= —K 1 ,L; L (0) L 0 7-1
= -K L; N (0) = N 0 7-2
E=Ka(C_C)KrL_KN+ B; 7-3
dT
where
T distance downstream divided by stream velocity, or time of travel.
L = ultimate BOB (carbonaceous) in mg/i.
Kr carbonaceous DOD removal constant in days (Assumed equal to Kd,
the deoxygenation constant.)
N = ultimate BOD (nitrogenous) in mg/i.
Kn = nitrogenous BOB removal constant in days
C = dissolved oxygen concentration mg/i.
Ka = coefficient of atmospheric reaeration in days
C 5 = the saturation value of atmospheric oxygen in water mg/i.
P-R = net production of photosynthetic oxygen, in mg/i/day.
B benthal oxygen demand in mg/i/day.
39
-------
In the analysis, most of the values were determined by individual obser-
vations (such as temperature, stream discharge, L, N, and C), or approxi-
mated by determinations based upon actual data and considered reasonably
characteristic of the Passaic River (C 5 , , and B). The reaeration
coefficient was determined according to the method of Owens-Edwards-Gibbs
(10). Thus the parameter search could be limited mainly to determining
the most likely values of Kr and K .
Curve Fitting Strategy by Digital Computer Method and Results
In the digital computer approach, the simulation technique was adopted
in place of the usual non—linear regression analysis because of the com-
plexity of the normal equations involved in the latter method. The pro-
cedure employed was essentially a computer search routine which would
select a particular set of parameters Kr and Kn that would give the
minimum standard error in fitting the observed DO data. In searching
for the appropriate set of parameters, a systematic sampling uniform—grid
method was used. All the computations were performed on the Rutgers
IBM 360/67 computer.
For each set of data, the curve-fitting began with setting up initial
values for both Kr and K . The following form of the Streeter-Phelps
relationships indicates the deficit at each point of observation along
the river reach below the aerator (1).
D = KrLo (e_Krt_e T) + K N 0 (e_KflT_e_I t)
I Ka-Kr Ka-K n
I B Kat P- -Kal
+ D 0 (e a ) + . (1 — e ) — (1 — e ) .... .7—tj.
a
where
DT = the oxygen deficit at holding time I , and other symbols are as
previously defined.
In the above equation all the parameters except Kr and K were either
actually n asured (11) or computed by known relationships (i). For the
cases in which both L 0 and N 0 values were not known, the averages of the
observed values were used. The computed DO deficits were then compared
with the observed values in the following manner.
E = ? (D — D )2
in which i-l
= the actually observed DO deficit at point i, and
= the DO deficit computed by Equation 7-L using assigned values of
1 Kr and Kn.
After the E value was computed for the initial values of Kr and K , a new
set of Kr and Kn values was assigned and a new value of E was obtained.
ho
-------
H
SUMMARY OF RESULTS
TABLE 7 -i
BY DIGITAL COMPUTER SIMULATION
Date Aerator
Flow
cTh
Temp.
Co
C 5
mg/i
D 0
mg/i
B
mg/i
P . . .R
mg/i
L 0
mg/i
N 0
mg/i
Ka
day—i
Kr
day -
K
day-i
ie
mg/i
7/15/70
ON
125
22.0
8.65
14.67
0.32
0.76
19.7
13.6
0.14)4
0.20
0.15
0.19
7/20
ON
1i5
23.5
6.20
2.73
0.3 ) 4
0.78
i3.8*
10.7*
0.149
o.io
o.o5
0.51
7/22
OFF
i09
21.5
8.70
5.20
0.314
0.79
13.8*
10.7*
0.146
0.25
0.15
0.1)4
7/23
OFF
109
20.7
8.88
6.38
0.35
0.76
13.8*
iO.7*
0.14)4
0.30
0.15
0.32
7/2)4
OFF
i09
22.5
8.60
5.97
0.35
0.79
13.8*
10.7*
0.146
0.60
0.05
0.38
7/27
ON
113
2 ) 4.3
8.30
6.08
0.3)4
0.75
i3.8*
10.7*
0.148
0.70
0.05
0.70
7/27
ON
i13
2)4.5
8.25
3.20
0.3)4
0.75
13.8*
10.7*
0.148
1.00
0.00
0.92
7/28
ON
111
25.0
8.20
3.37
0.3)4
0.148
13.8*
iO.7*
0.149
0.35
0.20
1.11
7/28
ON
iii
25.0
8.20
3.26
0.3)4
0.79
13.8*
10.7*
0.149
0.140
0.10
1.140
8/12
ON
109
2)4.0
8.30
3.92
0.314
0.95
17.6
7.0
0.148
0.20
0.25
0.3 ) 4
8/20
ON
lii
23.0
8.75
14.60
0.3)4
0.148
6.2
10.5
0.147
0.140
0.145
0.140
* Averages of the observed data
-------
This procedure was repeated for Kr and K values both ranging from 0.0
to 1.0, with an increment of 0.05. Thus there was a total of ) 4 ) 4l cotn-
binations of the parameter values. The computer would examine all the
)4.U1 E values and select the minimum value and the associated K 1 . and K
values. The range of 0.0 to 1.0 was chosen because it was thought that
this range would cover the values most commonly found in the literature
for such a river.
There were eight sets of DO data with the aerator “on” and three sets with
the aerator “off.” The results of the analysis are listed in Table 7—1.
The standard error of estimate Se, given in Table 7-1, was computed by
Se = n—2 z(D — Di) 2 0.5 7—6
In order to investigate the effect of the aerator operation on the
parameters K 1 . and K , the trbestn values of tt se parameters found by
the computer simulation were reduced to the standard temperature by the
following relationships: (1)
K 1 . (T) = Kr(20) X ( 10 ) 45 )T_20 77
K (T) = K (2o) x (l. 09 )T_20 7—8
The reduced K 1 . aria K values are listed in Table 7—2, according to the
aerator operating mode:
Table 7—2
SUMMARY OF K 1. AND Kn VALUES AT 20 0C
Aerator OFF
Aerator ON K
Date Kr Date K 1.
7/15/70 0.18 0.13 7/22/70 0.23 0.13
7/20 0.09 0.0)4 7/23 0.29 0.1)4
7/27 0.58 0.03 7/2)4 0.52 0.0)4
7/27 0.82 0.00 Average 0.3)4 0.10
7/28 0.28 0.13
7/28 0.32 0.06
8/12 0.17 0.18
8/20 0.35 0.35
Average 0.35 0.12
Upon examining the average values of the Kr and K parameters, it appears
that the operation of the aerator does not have any significant effect on
these parameters. The values of the parameters are a little higher when
the aerator was on than when the aerator was off, but by only very small
margins. For individual tests, the following points were observed:
1) There seem to be no consistent patterns that the variation in the
parameters would follow either when aerator was on or off. For example,
a high K value was obtained far August 12 when aerator was on. Yet on
-------
several other dates, such as July 20 and July 27, when the aerator was
also on, very low K values were obtained. As for Kr, both high and
low values were found when aerator was on while medium values were
obtained when aerator was off (see Table 7-2).
2) In general, parameter values were obtained which gave very good
fit, i.e., small standard error of estimates of the observed data. As
an exception, the July 27 and 28 data could not be fitted very well be-
cause of unusual fluctuations in the observed DO concentrations. To
illustrate this, one of the July 27 sets of data is plotted in Figu.re 10.
3) Mathematically speaking, the “best” solutions for the parameters
do not seem to be unique. One could easily change the values of the
parameters and still obtain a “near-best” fit. For example, for the
August 12 data, the best values for Kr and Kn were, respectively, 0.20
and O.2S with a standard error of estimate of O.31j. mg/i. However, if
Kr was increased to O. O and Kn reduced to O.l , the error only increased
to O.) mg/l; and even when Kr = 0.01 with a K of O.3S the error was
only O.t ij. mg/i. Strictly speaking, there is only one set of parameters
that represents actual conditions. Nevertheless, it seems that more
than one set of parameters would produce results that are practically
just as good. Also, when K was (forced) to be zero, then the best Kr
value was 1.0, which gave a poor standard error of estimate of 1.9 mg/i.
The August 12 data are plotted in Figure 11.
The Analog Computer Parameter Estimation Strategy and Results
The present state-of-the-art in curve-fitting or parameter estimation
procedures does not allow for perfect delineation between the various
parameters in the modified Streeter-Phelps models of the Passaic River.
For example, in the context of a nonlinear regression curve-fitting pro-
cedure, all of the parameters are coupled together in a nonlinear
fashion when the system equations are presented in integrated form.
This situation all too frequently leads to polymodality and multiple
candidate parameter sets which are difficult to arbitrate. It is pre-
ferable to have available, in addition to the results from a nonlinear
regression analysis based on a digital computer solution, a parameter
estimation strategy which is linked to the differential equations of
the model. In the modified Streeter-Phelps equations, the parameters
are linear when the model is in its characteristic differential equation
form. An analog computer is ideally suited for dynamic simulation and
curve-fitting strategies, which, when used systematically on the differ-
ential equations produces on-the-spot parameter estimations. These
estimations can be used as a check against estimations based on the
digital computer strategy.
In Figure 12, there is shown a fully scaled analog computer diagram
which represents the dynamic simulation of Equation 7-1, 7-2, and 7-3.
Basically, a rotational reparameterization procedure was used for
parameters Kr and Kn. A visual goodness of fit criteria was estab-
lished and the poor data points were weighted to varying degrees. All
simulations were carried out using a TR-L 8 analog coniputer. The simu-
lations were all recorded on a 1130 X-Y plotter.
113
-------
Holding time in days
FIG. 10 DiSSOLVED OXYGEN SAG CURVE FTI’TED BY DIGITAL COMPUTER
July 27, 1970, aerator on
8
6
4
2
0
0 1 2 3
-------
Holding time in days
FIG. 11 DISSOLVED OXYGEN SAG CURVES FOR VARIOUS VALUES OF Kr and K
DO
Mg/i
8
6
4
2
0
(1) Kr = 0. 20, K = 0. 25
(2) Kr = 0.50, Kn = 0.15
(3) Kr = 0.25, K = 0.25
N .
.
____
.
(3)
0
I
2
3
Aug. 12, 1970, aerator on
-------
(0. N)
Time Base
FIG. 12 SCALED ANALOG COMPUTER DIAGRAM USED ll’ THE SIMULTATION
OF BOD, NOD, AND DO
(0. L)
0 ’
-10
+
O.2 0.2B
o.1c 5
-------
Figure 13 shows some representative analog computer drawn curves for
various Kr and K parameter sets. Curve number (A) represents the set of
values of Kr and K which had been determined by purely digital computer
techniques and regression analysis. This curve was used as a base line
and reference grid for the rotational reparameterization strategy.
Curves (B) and (C) represent extreme cases where Kr and Kn are zero. It
can be surmised from the indicated trend in Figure 18 that when the aera-
tor is on, there seems to be a rather pronounced tendency for K to have
a value closer to 0.3 days 1 than to a value of zero, and this trend seems
to be little affected by Kr on the range of 0 to 0.2 days -. Figure l1
represents another set of data when the aerator was on, but the trend in
Kr and Kn is not as clear. This is due primarily to the erratic behavior
of the data in the 0.2 to 0.7S range of holding times. If these points
are weighted lightly, then curve (B) fits the remainder of the data fair-
ly well. Curve (A) represents the parameters derived from the digital
computer. The analog computer drawn curve (A) consequently agrees nicely
with what the digital computer would predict for this parameter set. If
the data points in the 0.2 or 0.7 range of holding times are weighted
more heavily than the points in the 1. or 2.0 range of holding times, we
again witness an enhanced nitrogenous coefficient, even when Kr = 0. This
is shown by curve (n). Curve (C) shows the value of Kn when the weighing
of the data points are reversed. In Figure 1 , results are shown for the
case where the aerator is off. The parameter sets for curve (A) - (D),
while all different, represent the data about equally well. For example,
curve (A), which is also the digital computer result, is drawn through the
middle of the cluster of points in an averaging fashion. Curves (B) -
(D) on the other hand are drawn so as to weigh different parts of the
data at different times. For example, curve (B) weighs the points less
than 1 day and greater than 2 days more heavily than the points between
1 and 2 days. Curve (D) does likewise except that the lower range of
points in the region greater than 2 days are weighed more heavily than
the points with higher values of DO in the range 2-3 days. Curve (C)
weighs the higher values of DO at the various values of r. An enhanced
value of Kr and a low value of Kn, however, seem to compromise the
curve-fitting anomaly and would agree with the data in Figure hi..
There are several other possible interpretations of the data, but these
would all involve increasing the number of undetermined parameters.
For example, the reaeration coefficient, Ka, is determined from the
Ownes-Edwards-Gibbs correlation, but there are several other authorities
which would predict widely varying values for Ka. Each new value of Ka
different from the one used here would alter significantly the range of
values of Kr and Kn. Any major error or rapid changes in the BOD data
would lead to much different interpretations. For example, in Figure
18, curve (C) could be interpreted as representing a situation where
= 0 and L 0 = 17.6 tng 02/1. This would correspond to a value of K 1 , =
0.39 days 1 . In Figure iIi, curve (D) could be interpreted as represent-
ing a case where K = 0 with L = 13.8 mg 02/1. This would mean that
Kr = O.7S days—l. In Figure l curve (D) would likewise represent a
situation where L 0 = 13.8 and K 1 . = 0.6 days- 1 with Kn = 0. In view of
the rapid fluctuations in BOD which have previously been shown to exist
)47
-------
Holding time - T days
= 0, Kn = 0.39
FIG. 13 ANALOG COMPUTER DRAW’N CURVES OF DISSOLVED OXYGEN
CONCENTRATION VERSUS HOLDING TIME, Aug. 12 data
Dissolved oxygen concentration,
Mg O 2 /L
7
6
5
4
3
2
1
S
.
(B) Kr 1.86, K = 0
0.2
0
1 2 3
-------
3
Holding time - T days
FIG. 1)4 ANALOG COMPUTER DRAWN CURVES OF DISSOLVED OXYGEN
CONCENTRATION VERSUS HOLDING TIME, July 27 data (p. m.)
Dissolved oxygen concentration,
Mg 0 2 /L
‘0
7
6
5
4
3
2
1
0
(B) Kr = 0.97,
= 0
(C) Kr = 0, K = 0.54
t) Kr = 0.7, K
0 1 2
K 0. 75
-------
Holding time — days 2
FIG. 1 ANALOG COMPUTER DRAWN CURVES OF D SOLVED OXYGEN
Dissolved oxygen
concentration, mg 0 2 /L
7
6
5
4
3
2
1
0
0
(C) Kr 0, K = 0.45
(B) Kr 0.78, K = 0
A) Kr = 0.6, K 0.
0 1 2 3
6
CONCENTRATION VERSUS HOLDING TIME, July 24 data
-------
(1), the possibility of such changes affecting the analysis of individual
sets of data must be taken into account.
Summary
There are various methods available for the computation of the parameters
that would best fit the data. Yet the reliability of these computed
parameters depends wholly on the accuracy of the estimation of the other
parameters in the DO equation. Hence, accurate field measurements of
certain parameters, such as Ka B, , are essential to successful model-
ling.
Since fully reliable estimates of some parameters other than K 1 . and K
cannot be obtained, it is of interest to model the oxygen dynamics regime
for other likely values, using the analog and digital computer techniques
of curve-fitting.
Based on the results of the analog and digital computer analysis, it did
not appear that the operation of the mechanical aerator during the summer
of 1970 had any significant or consistent effect on the parameters 4 and
K .
Based upon statistical indications only, it would appear that appreciable
nitrification was taking place during the test conditions. However, since
there was strong biochemical evidence to the contrary, this aspect re-
mains problematical.
-------
SECTION VIII
ACKN MIEDGEMENTS
The testing and analysis and the preparation of this report were arranged
through the Water Resources Research Institute, Rutgers University. The
several authors of the report contributed sections in line with their re-
spective interests and disciplines as follows: W. Whipple, Jr., general
coordination (Sects. I—Ill), F. W. Dittrnan, Field Operations (Sect. Tv),
S. L. Yu and F. W. Dittman, Oxygen Transfer (Sect. v), G. Mattingly,
Dispersion Analysis (Sect. VI), and S. L. Yu and B. Davidson, Mathemati—
cal Model and Parameters of BOD (Sect. VII).
The oxygen diffusers tested in this project were designed, constructed
and operated by the Martin Marietta Corporation, at the expense of the
corporation. Excavation of the basin was carried out by the Morris
County Mosquito Extermination Conirnission, at cost, through the courtesy
of Mr. Robert L. Vannote. Besides the authors, Dr. Joseph V. Hunter
assisted in technical supervision. A number of students participated,
particularly, Mr. Kenneth Hilsen, who acted as assistant field supervi-.
sor, Mr. John Cirello who assisted in the dispersion tests, and Mr.
Thomas Tuffey.
-------
SECTION IX
REFERENCES
1. Whipple, W., Jr., Hunter, J.V., Davidson, B., Ditttnan, F.W., Yu, S.L.,
“Instream Aeration of Polluted Rivers,” Water Resources Research
Institute, Rutgers University, New Brunswick, N.J. (1969).
2. Hunter, J.V., and Nhipple, W., Jr., ttEvaluating Instrearn Aerators of
Polluted Rivers,” Jour. Water Poll. Control Fed., 1 2 , No. 8,
pp 2)49-262, Pt. II (1970).
3. Whipple, W., Jr., Coughian, F.P., and Yu, S.L., “Instreani Aerators
for Polluted Rivers,” Jour. SED, Am. Soc. Civ. Engrs., 96 , SA ,
PP l153-ll6S (1970).
)4. Yu, S.L., “Aerator Performance in Natural Streams,” Jour. SED, Am.
Soc. Civ. Engrs., 96 , SA , p 1099 (1970).
. Whipple, W., Jr., Dittman, F.W., and Yu, S.L., “Induced Oxidation of
Streams and Water Quality Control Institutions,” Bulletin, Am. Water
Res. Assoc., 6 , 6, p 968 (1970).
6. Whipple, W., Jr., Hunter, J.V., Dittman, F.W., and Yu, S.L., “Oxygen
Regeneration of Polluted Rivers: The Delaware River,” Environmental
Protection Agency, Water Quality Office, Water Pollution Control
Research Series - 16080 DUP 12/70.
7. Bourodirnos, E.L, and Nichna, L., “Flow Concentration Groins for Re-
aeration in Passaic River - A Hydraulic Model Study,” Water Resources
Research Institute, Rutgers University, New Brunswick, N.J., 99 pp,
(1970).
8. OWRR Research Project B-027-N.J., t!Instream Aeration and Parameters
of Biochemical Oxygen Demand.” Results of the cr,, determination
have been summarized in Ref. (6).
9. Susag, R.H., Polta, R.C., and Schroepfer, G.J., “Mechanical Surface
Aeration of Receiving Waters,” Jour. Water Poll. Control Fed., 38 ,
1 (1966).
10. Owens, M., Edwards, R.W., and Gibbs, J.W., “Some Reaeration Studies
in Streams,” mt. Jour. Air and Water Poll, 8 , p. )469 (196)4).
11. Hunter, J.V., and Whipple, W., “Evaluating Instreain Aeration of
Polluted Rivers,” Jour. Water Poll. Control Fed., )42 , 8, Pt. II,
pp R2 ) 49—R262 (Aug. 1970).
12. Bischoff, K., and Levenspiel, 0., “Fluid Dispersion-Generalization and
Comparison of Mathematical Models,” Chetn. Engr. Sci., 17 , pp 2 ) 4S-25S
and pp 2 7-26)4 (1962).
-------
13. Daily, J., and Harleman, D.R.F., “Fluid Dynamics,” Addison—Wesley
Publishing Co. (1966).
lh. Elder, J.W., “The Dispersion of Marked Fluid in Turbulent Shear Flow,”
Jour. Fluid Mech., 5 , pp 5U -56O (1959).
15. Fischer, H., “A Note on the One Dimensional Dispersion Model,” mt.
Jour. Air and Water Poll., 10 , pp Li1i 3-L 52 (1966).
16. Fischer, H., “The Mechanics of Dispersion in Natural Streams,” Jour.
Hydraulics Div., Am. Soc. Civ. Engrs. , p 187 (Nov. 1967).
17. Fischer, H., “Dispersion Predictions in Natural Streams,” Jour. San.
Eng’g Div., Am. Soc. Civ. Engrs. , p 927 (Oct. 1968).
18. Fischer, H., “Methods far Predicting Dispersion Coefficients in
Natural Streams, with Applications to Lower Reaches of the Green and
Duwarnish Rivers, Washington,” U.S. Geological Survey Professional
Paper 582—A (1968).
19. Glover, R., “Dispersion of Dissolved or Suspended Materials in Flow-
ing Streams,” U.S. Geological Survey Professional Paper L 33—B (l96b).
20. Holley, E.R., “Some Data on Diffusion and Turbulence in Relation to
Reaeration,” University of flhinois Water Resources Report No. 21
(July, 1969).
21. Jobson, H., and Sayre, W., “Predicting Concentration Profiles in Open
Channels,” .ASCE National Water Resources Engr. Mtg., Memphis, Tenn.,
paper llL 7, Jan. 1970.
22. Paulson, R.W., “The Longitudinal Diffusion Coefficient in the Delaware
River Estuary as Determined from a Steady State Model,” Water Re-
sources Research, 5 , 1, p 59 (Feb. 1969).
23. Taylor, G.I., “Dispersion of Soluble Matter in Solvent flowing Slow-
ly Through a Tube,” Proc. Roy. Soc., Series A, 219 , p 186 (1953).
2tL. Taylor, G.I., “The Dispersion of Matter in Turbulent Flow Through a
Pipe,” Proc. Roy. Soc., Series A, 233 , p Lii 6 (l9Sb).
25. Wilson, J.F., “Time-of-Travel Measurements and Other Applications of
Dye Tracing,” mt. Assoc. Sci. Hydrol. Pub. No. 76 (1968).
26. Wilson, J.F., “An Empirical Formula for Determining the Amount of
Dye Needed far Time-of-Travel Measurements,” U.S. Geological Survey
Professional Paper 600-D, pp D51 -D56 (1968).
27. Yotsukura, N., Fischer, H., and Sayre, W., “Measurements of Mixing
Characteristics of the Missouri River Between Sioux City, Iowa,
and Plattsniouth, Nebraska,” U.S. Geological Survey Water—Supply
Paper l899 —G (1970).
56
-------
1 Acce’, ion Number 2 Subject Field & Group
SELECTED WATER RESOURCES ABSTRACTS
OS
G
INPUT TRANSACTION FORM
organization
Water Resources Research Institute, Rutgers University, New Brunswick, New Jersey
SEND TO WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERiOR
WASHINGTON. D. C 20240
J Title
OXYGEN REGNERATI OF POLTIJTEIJ RIVERS: TUE PASSAIC RIVER
1 0 1 Author(s)
Whipple, W., Jr.
Hunter, J.V.
Dittman, F.W.
Yu, S.L.
Davidson, B.
Mattingly, G.
Project Designation
16080 FYA 03/71
.
Note
22 Citation
23 Descriptors (Starred First)
* Water pollution control, Water quality control, Stream improvement, Abatement,
Aeration, Mathematical models.
25 identifiers (Starred First)
* Instreani aeration, * Induced oxygenation
27 Abstract
Field tests were made of a mechanical surface aerator and of pure oxygen
diffusers in a small polluted river, the upper Passaic. Results generally
corroborated results of previous test, as to performance of surface aerators
on such rivers, in excavated pools. A somewhat higher oxygen transfer rate
was obtained with a flow concentration device, which, in a permanent installa-
tion, would take the form of low rock spur dikes, one extending from each bank,
or flow concentration groins. Tests in shallower water, about 7 feet deep,
were inconclusive. Tests of oxygen diffusers were fragmentary, due to mechani—
cal difficulties with the equipment; but it was demonstrated that the very
fine bubbles used were very largely absorbed in the water. A dye dispersion
test gave a very high longitudinal dispersion coefficient downstream of the
aerator. Mathematical modelling indicated that during the period of test,
parameters of biochemical deoxygenation were not changed by the artificial
aeration process. (Whipple—Rutgers)
Abstractor William Whipple, Jr. institution Rutgers - The State University of New Jersey
WR.102 IREV JULY 19691
WR5I C
4 GPO ‘969—359-339
------- |