Oxygen
Aeration
Prepared by:
US Environmental
Protection Agency
Office of Research
& Development
Environmental
Research Center
Cincinnati, Ohio
Prepared for:
US Environmental
Protection Agency
Office of
Technology Transfer
Design Seminar
Program
-------
EPA EXPERIENCES IN OXYGEN-ACTIVATED SLUDGE
Prepared for the
U.S. Environmental Protection Agency
Technology Transfer Design Seminar Program
by
Richard C. Brenner
Revised
October 1974
National Environmental Research Center
Advanced Waste Treatment Research Laboratory
Office of Research and Development
Cincinnati, Ohio
-------
EPA EXPERIENCES IN OXYGEN-ACTIVATED SLUDGE
Richard C. Brenner
INTRODUCTION
Utilization of oxygen aeration for activated sludge treatment is
receiving increasing attention in wastewater treatment plant construction
in the United States. The concept, although more than 20 years old, has
received serious consideration only during the last six years with the
development of several cost-effective systems for dissolving and utilizing
oxygen gas in an aeration tank environment.
The rapid transition from the drawing boards to full-scale imple-
mentation has been possible because of intensive government and private
research and development programs. The U.S. Environmental Protection
Agency (EPA) and its predecessor organizations have contributed signi-
ficantly to the total research and development effort. The purpose of
this paper is to summarize the role of EPA during the period of 1968-1974
as the oxygen aeration process progressed to its current level of development.
As outlined in Table 1, EPA has pursued seven active projects to date.
The projects include in-house pilot plant studies to examine process
kinetics, extramural feasibility grants and contracts, extramural materials
and safety projects, and extramural demonstration grants.
The EPA contribution to the projects described in Table 1 exceeded
$3.4 million through Fiscal Year 1974 (ended June 30, 1974).
The cost breakdown by project is given in Table 2.
Test facilities, experimental plans, and results (where available)
for each of the above projects are summarized in the following sections.
THE BATAVIA PROJECTS
A research and development contract was awarded to the Union Carbide
Corporation in October 1968 to evaluate a proprietary staged, covered-
tank oxygenation system at the Batavia, New York, Water Pollution Control
Plant. Union Carbide was awarded a follow-up contract in June 1970 to
-------
TABLE 1. EPA RESEARCH AND DEVELOPMENT
PROJECTS ON OXYGEN AERATION
Project
Objective
1. Batavia I and II
(Union Carbide Corporation)
2. Newtown Creek
(New York City)
3. Las Virgenes (California)
Municipal Water District
4. FMC Corporation
5. EPA/District of Columbia
(Blue Plains) Pilot Plant
6. Bureau of Reclamation
7. Rocketdyne Division of
Rockwell International
Establish feasibility of multi-stage,
covered-tank oxygenation concept.
Scaled-up demonstration of multi-stage,
covered-tank oxygenation system.
Demonstration of single-stage, covered-
tank oxygenation system.
Establish feasibility of open-tank
oxygenation concept.
Determine process kinetics over wide
range of operating conditions.
Materials of construction corrosion
testing.
Define safety requirements and
develop safety manual and checklist.
TABLE 2. EPA RESEARCH AND DEVELOPMENT EXPENDITURES
ON OXYGEN AERATION THROUGH FY-73
Project
Cost to EPA Type of Project
Union Carbide Corporation (Batavia I and II)
New York City (Newtown Creek)
Las Virgenes (California) Municipal Water
District
FMC Corporation
EPA/District of Columbia (Blue Plains)
Pilot Plant
Bureau of Reclamation
Rocketdyne Division of Rockwell International
$
$1
$
$
$
$
$
795,000
,574,000
186,000
142,000
500,000
165,000
92,000
Contracts
Grant
Grant
Grant
Contracts and
Inhouse
Contract
Contract
TOTAL $3,454,000
-------
3
better define soluble organic removals and excess biological sludge pro-
duction and to undertake initial pilot plant studies on oxygen sludge
dewatering and stabilization. The oxygenation system was installed in
one of two existing air-activated sludge trains at Batavia. During the
first contract, the performance of the oxygen train was evaluated against
that of the intact air train. A schematic diagram of the Batavia Plant
after installation of the oxygen system is shown in Figure 1.
The oxygen system configuration evaluated at Batavia was the first
large-scale embodiment of the now well known "UNOX" process.* A typical
three-stage "UNOX" aerator is shown schematically in Figure 2. The aerator
operates as a series of completely mixed stages, thereby approximating plug
flow. Oxygen gas is fed under the aeration tank cover at the inlet end of
the tank only and flows co-currently with the liquid stream from stage to
stage. Gas is recirculated in each stage by centrifugal compressors which
force the gas down hollow shafts out through submerged rotating spargers.
Submerged turbines maintain suspension of the mixed liquor solids and
disperse the oxygen gas. A mixture of unused oxygen gas, cell respiration
by-product carbon dioxide, and inert gases is exhausted from the final stage,
typically at an oxygen composition of about 50% and a flow rate equal to
10-20% of the incoming gas flow rate. Using co-current gas and liquid flow
to match the decreasing dissolution driving force inherent in continually
decreasing oxygen gas composition with the decreasing oxygen demand of
wastewater undergoing biological treatment has proven to be a very efficient
oxygen contacting and utilization technique.
A second-generation multi-stage process has been developed and utilized
both by Union Carbide and Air Products and Chemicals, Inc. This adaptation of
the original covered-tank concept replaces the recirculating compressors and
rotating spargers with surface aerators. Oxygen transfer is accomplished by
gas entrainment and dissolution. Submerged turbines are also used optionally
where tank geometry requires additional mixing capability. As shown in Figure 3,
all other aspects of the system are unchanged. The first Air Products and
Chemicals version of the covered-tank, surface-aerator concept has been oper-
ating for approximately three yearsat *the Westgate Treatment Plant in Fairfax
County, Virginia (design flow 12 mgd). Operation commenced in July 1972
*Mention of a trade name or commercial products does not constitute Environ-
mental Protection Agency endorsement or recommendation for use.
-------
DIGESTED
SLUDGE
SLUDGE T 1
DISPOSAL
SUPERNATANT
SLUDGE
VACUUM FILTER
CONVEYOR BELT
DUMP
TRUCK
KEY
SEWAGE FLOW
SLUDGE FLOW
DESIGN POPULATION 25,000
AVG FLOW: 2.5 MIL. GAL /DAY
MAX.FLOW: 6.25 MIL GAL./DAY
U. J i
THICKENED |_»
SLUDGE
A0£
MAIN PUMP STATION
FIGURE 1. SCHEMATIC FLOW DIAGRAM FOR WATER POLLUTION CONTROL PLANT, CITY OF BATAVIA, NEW YORK
-------
AERATION
TANK COVER
OXYGEN
FEED GAS
WASTE_
LIQUOR
FEED
RECYCLE
SLUDGE
PROPELLER
DRIVE
GAS RECIRCULATION
COMPRESSORS
EXHAUST
"GAS
MIXED LIQUOR
EFFLUENT TO
CLARIFIER
PROPELLER
SPARGER
FIGURE 2. SCHEMATIC DIAGRAM OF MULTI-STAGE, COVERED-TANK OXYGENATION
SYSTEM WITH GAS RECIRCULATION COMPRESSORS AND SUBMERGED
TURBINE/SPARGERS
-------
AERATION TANK COVER
SURFACE AERATOR
OXYGEN
FEED GAS
WATER _
FEED
PCYCLE ^-
LUDGE
r
^..**~s~~~*+
\^3
d
J
* J"
?
\
J_^
V "t
1
1
1
1
1
1
1
j
1
ci
-, /
1 /
/
/ j
l**^ ^^
i
i
i
i
i
i
i
i
i
i
i
r*,
r-T
M r
i
i^-
\ cib
N.
^^ STAGE
BAFFLE
^mfm
HI
MIXER DRIVE
EXHAUST
\
MIXED LIQUOR
^5-»- EFFLUENT TO
CLARIFIER
SUBMERGED PROPELLER (OPTIONAI)
FIGURE 3. SCHEMATIC DIAGRAM OF MULTI-STAGE, COVERED-TANK OXYGENATION SYSTEM
WITH SURFACE AERATORS
-------
7
at Speedway, Indiana, the first municipal plant to utilize the Union
Carbide surface aerator system (design flow 7.5 mgd). The surface aerator
modification of the basic multi-stage process has exhibited better cost
effectiveness for tanks up to approximately 15 feet deep and is being used
increasingly in full-scale design. Market forecasts and actual experience
to date of firms selling multi-stage, covered tank oxygen systems indicate
that 80-85% of the plants that eventually utilize this oxygen system concept
will employ surface aerator designs. A report currently being prepared for
EPA by Air Products and Chemicals documenting the Fairfax County, Virginia
case history from inception through two years of operation with the oxygen
system will be available by mid-1975.
The results of the Batavia projects have been widely disseminated in
two EPA Water Pollution Control Research Series Reports (17050 DNW 05/70)
(17050 DNW 02/72). One of the conclusions expressed in these reports is
that oxygen aeration can provide equal treatment efficiency to air aeration
with only one-third as much aeration volume. This conclusion has been sub-
ject to widespread criticism. In that this generalization was reached by
comparing an efficient oxygen contacting system with a relatively inefficient
coarse-bubble air aeration system, the criticism appears to be justified.
The increasing variety of air aeration equipment being marketed offers a
wide range of oxygen transfer kinetics. Some of this equipment has been
shown to be capable of supporting the higher mixed liquor solids concen-
trations necessary for justifying smaller volume biological reactors (e.g.,
the INKA aeration system, Divet, et al., 1963). Design engineers are urged
to investigate and prepare cost estimates for both air and oxygen systems
as a basis for process selection. Process selection should be made from a
total integrated system comparison, including aeration, secondary clarifi-
cation, and excess biological sludge handling and disposal requirements.
Pertinent results of the two Batavia projects relating only to oxygen
system performance are summarized below:
1. The feasibility of achieving high oxygen gas utilization (91-95%)
was established.
2. Efficient biological performance (90-95% BOD^ and suspended solids
removals, 80-85% COD removal) was demonstrated with short aerator
detention periods (1.4-2.8 hours based on Q) and high organic
volumetric loadings (140-230 Ib BOD5/day/l,000 ft3).
3. High mixed liquor dissolved oxygen (D.O.) levels (8-12 mg/1) were
maintained at high mixed liquor suspended solids (MLSS) concentrations
(3,500-7000 mg/1).
-------
8
4. Warm weather secondary clarifier performance deteriorated above
2
an average surface loading of 1,600 gpd/ft .
5. Oxygen sludge exhibited excellent thickening properties during
secondary clarification (settled sludge of 1.5-3.0% solids).
6. Aerobic digestion of oxygen waste activated sludge with oxygen
produced comparable volatile suspended solids (VSS) reduction
rates to those given in the literature for air aerobic digestion
processes. Reductions in oxygen sludge VSS concentrations of 25
and 40% were achieved with 7 and 15 days of aerobic stabilization,
respectively.
7. Direct vacuum filtration of undigested oxygen waste activated
sludge was shown to be feasible using 10% ferric chloride for
2
conditioning. Cake yields of 3.5-4.5 Ib/hr/ft arid moisture
contents of 83-85% were achieved at a cycle time of 2.4 min/rev.
Moisture content improved to 75-80% but cake yield dropped to
2
1.5-2.5 Ib/hr/ft at a cycle time of 6.3 min/rev.
8. Vacuum filtration of aerobically digested oxygen waste activated
sludge proved to be infeasible with the chemical conditioners
tested.
THE NEUTOWN CREEK PROJECT
Results of the initial Batavia contract were judged sufficiently
encouraging to justify a scaled-up demonstration of the multi-stage oxygen
system in a large municipal plant. A research and development grant was
awarded to New York City in June 1970 to convert one of sixteen parallel
bays at its Newtown Creek facility to oxygen using the recirculating com-
pressor/submerged turbine version of the "UNOX" process. The design flow
of the test bay is 20 mgd, roughly 10 times higher than the capacity of the
Batavia oxygen system. In addition to the $1.574 minion EPA grant, New York
City provided over $1.2 million in city funds in support of the project.
The Newtown Creek plant was designed on the modified aeration principle
for 1.5 hours of aeration time (based on Q) and treatment efficiencies in
the range of 65-70%. The city is now confronted with an upgrading problem
in a land-locked neighborhood (see Figure 4), a situation common to many
-------
STREET
co
z
<
U
OXYGEN AERATION
TEST BAY
r SLUDGE
IV DIGESTION
00 O
bo o
o o
o o
SLUDGE
CONCENTRATION
MAIN
BUILDING
AERATED GRIT
CHAMBERS
AERATION
TANKS
SECONDARY
CLARIFIERS
FIGURE 4. PLANT LAYOUT FOR NEWTOWN CREEK POLLUTION CONTROL PLANT,
BROOKLYN, NEW YORK
-------
10
large urban plants in the United States. Oxygen was believed to be a good
candidate for achieving the required 90£ BOD,, removal within the confines
of the existing aeration tanks and secondary clarifiers. Future conversion
of the entire 310 mgd facility to oxygen was the ultimate objective provided
a removal up to the 90% ± BOD5 level could be consistently demonstrated in the
test bay. Two views of the Newtown Creek test bay are .shown in Figure 5.
The test bay went on stream in early June 1972. Extensive mechanical
problems and unreliable meters prevented accurate data collection during
the start-up phase (June 4, 1972-September 16, 1972) during which time the
influent flow was increased from 11 mgd to the design level of 20 mgd. Data
for this period are limited to effluent quality as summarized in Table 3.
TABLE 3. EFFLUENT QUALITY AT NEWTOWN CREEK
DURING START-UP (6/4/72-9/16/72)
Flow
(mgd)
11+2*
14 + 3b
20 + 4C
Duration
(weeks)
Total
BOD
4 10
5 8
6 15
Effluent
Soluble
BOD
4
3
4
Concentration (mg/1)
Total
COD
68
55
62
Soluble
COD
50
41
47
Suspended
Solids
18
19
18
aMLSS = 4,865 mg/1, Detention Time (based on Q) = 2.7 hr +.
MLSS = 5,920 mg/1, Detention Time (based on Q) = 2.1 hr +.
CMLSS = 4,260 mg/1, Detention Time (based on Q) = 1.5 hr +.
Metering difficulties were finally resolved by mid-September 1972 permitting
commencement of the extensive data collection program planned for this project.
From September 17, 1972 through September 1, 1973, seven phases of a ten-phase
experimental program were completed. The influent flow conditions for these seven
phases are summarized in Table 4. Diurnal peak, average, and minimum flow rates
for Phases 4 through 7 are given in Table 5. With the exception of Phase 7, the
diurnal fluctuation patterns were selected to simulate the actual influent flow
patterns of the Newtown Creek facility.
-------
FIVE GAS RECIRCULATING
PUMPS
[
RAW ^
SEWAGE
ifi
\
I
\ COMPRESSORS
D D Q D D
o
o
O
O
O
o
S
o
200' »-
X
EIGHT SUBMERGED
,/ PROPELLER MIXERS
I
55'
i
4OO'
-
SECONDARY
EFFLUENT
RAW
SEWAGE
( DRIVE
MIXER ; PROPELLER
ASSEMBLY I SPARGER
PLAN VIEW
NO SCALE
3XOC
GRIT FOUR-STAGE OXYGEN AERATOR
CHAMBER! WD=IS'
L_,
z:
SECONDARY
EFFLUENT
|
SECONDARY
CLARIFIER
WD=12'
SLUDGE
RECYCLE
ELEVATION
NO SCALE
SLUDGE
WASTING
FIGURE 5. PLAN AND ELEVATION VIEWS OF OXYGEN AERATION TEST BAY, NEWTOWN CREEK
-------
12
TABLE 4. EXPERIMENTAL SCHEDULE FOR NEWTOWN
CREEK PROJECT (9/17/72-9/1/73)
Phase
1
2
3
4
5
6
7
Dates
9/17/72 - 11/25/72
12/10/72 - 2/1/73
2/18/73 - 4/7/73
4/8/73 - 6/2/73
6/3/73 - 7/7/73
7/8/73 - 8/11/73
8/12/73 - 9/1/73
Influent
20.8 mgd
i f. .. ,., \
Avg. 17.7
8, , ,,s
AVG. 15.1
20.6 mgd
25.3 mgd
30.0 mgd
35.4 mgd
Flow Condition
(Constant)
T) .,.,.. , ,,,\ 1 "7 i
mgd (Winter U]
20 mgd
mgd (Winter Ri
(Diurnal)
(Diurnal)
(Diurnal)
(Diurnal)
mgd
pset)
estart)
TABLE 5. DIURNAL FLUCTUATION PATTERNS FOR
NEWTOWN CREEK
Phase Avg. Flow
(mgd)
4 20.6
5 25.3
6 30.0
7 35.4
(4/8/73-9/1/73)
Peak Flow
(mgd)
24
30
36
37.5*
Minimum
(mgd)
14
17
19
30
Flow
^Maximum influent pumping capacity.
-------
13
A performance summary for the oxygenation system for Phases 1 through
7 is presented in Table 6. System sludge characteristics, aerator
loadings, and secondary clarifier loadings are summarized in Tables
7, 8, and 9, respectively.
TABLE 6. PERFORMANCE SUMMARY FOR NEWTOWN CREEK
(9/17/72-9/1/73)
Total BOD5 In (mg/1)*
Total BOD Out (mg/1)
% Removed
Soluble BOD5
Soluble BOD5
% Removed
Total COD In
In (mg/1)*
Out (mg/1)
(mg/1)*
Total COD Out (mg/1)
% Removed
Soluble COD Out (mg/1)
Susp. Solids
Susp. Solids
% Removed
Sewage Temp.
In (mg/1)*
Out (mg/1)
Range (°F)
1
156
9
94
84
4
95
356
61
83
50
149
12
92
V
t
56
2
157
21
87
78
13
83
365
88
76
69
146
22
85
6,4
5*3
3
152
17
89
91
12
87
365
76
79
63
144
17
88
I
Phase
4
171
17
90
102
11
89
365
77
79
69
159
18
89
₯
v
69
5
213
22
90
113
13
88
307
70
77
58
147
24
84
6,2
*
74
6
218
21
90
99
11
89
290
64
78
49
125
17
86
8
7
212
23
89
88
15
83
308
62
80
46
131
17
87
I2
*
78
*No primary sedimentation. Concentrations shown are for raw sewage
influent to oxygen aerator.
-------
14
TABLE 7. AVERAGE SYSTEM SLUDGE CHARACTERISTICS
FOR NEWTOWN CREEK (9/17/72-9/1/73)
Phase
1
2
3
4
5
6
7
MLSS
(mg/1)
4,890
5,060
4,000
3,875
4,550
4,155
3,090
MLVSS
(mg/1)
4,110
4,150
3,200
3,110
3,640
3,340
2,485
Return
Sludge
Flow
(% of Q)
30
40
50
45
44
34
25
Return
Sludge
TSS
(rag/I)
16,260
12,835
11,370
13,420
15,975
16,330
12,685
SVI
(ml/gram)
45
59
77
53
42
43
48
SRT
(days)
_#
.*
_#
1.35
1.37
1.24
0.77
^Sludge wasting data not reliable.
TABLE 8. AVERAGE AERATOR LOADINGS FOR
NEWTOWN CREEK (9/17/72-9/1/73)
Phase
1
2
3
4
5
6
7
Detention Time
-Based on Q-
(hr)
1.43
1.68
1.96
1.44
1.17
0.99
0.84
F/M Loading
Ab BOD5/day\
^ Ib MLVSS J
0.65
0.57
0.57
0.92
1.19
1.62
2.44
Volumetric Organic
Loading
Ab BOD5/day\
^ 1,000 f t3 J
163
140
110
178
272
331
379
-------
15
TABLE 9. AVERAGE SECONDARY CLARIFIER LOADINGS
FOR NEWTOWN CREEK (9/17/72-9/1/73)
Phase Surface Overflow Mass Loading Weir
Rate Ab TSS/ft2\ Loading
i M f t / *. \ * I ._.j.
1
2
3
4
5
6
7
(gpd/ft2)
945
805
686
936
1,150
1,364
1,609
\ day )
50.1
48.2
32.6
43.7
63.0
63.3
52.0
(gpd/ lineal
129,000
110,000
93,000
127,000
157,000
186,000
219,000
From the beginning of the project, New York City officials
considered performance of the oxygen test bay during cold weather the
most critical segment of the experimental program. It was during the
prolonged severe weather period that the true upgrading potential of
oxygen for Newtown Creek would be most evident. A discussion of the
data collected through September 1, 1973 at Newtown Creek is first pre-
faced, therefore, with a summary description of the operational difficulties
encountered during the 1972-73 winter season.
During Phase 1 (autumn 1972), operation went smoothly and perfor-
mance was obviously excellent. On November 25, 1972, the test bay was
shut down temporarily to replace a bearing on the sludge recycle pump.
What was planned to be a two-day outage turned into a two-week shutdown
when the stocked spare bearing proved to be the wrong size and a new one
had to be located. During the outage, sludge in the reactor was continually
oxygenated while the sludge in the secondary clarifier was devoid
of oxygen. Due to the imminence of the upcoming cold weather, it was
decided to restart the system using the existing sludge rather than empty
the tanks and take the time necessary to generate a new biomass. The
oxygen test bay was put back into service on December 10, 1972.
-------
16
For the first several weeks of Phase 2, performance was satisfactory
as the influent flow rate was gradually increased to the design level of
20 mgd. Shortly thereafter sludge settling properties began to deteriorate
and effluent BOD , COD, and suspended solids residuals exhibited a slowly
increasing trend. Microscopic examination of the mixed liquor revealed
the appearance of filamentous organisms of both apparent bacterial and
fungus origin. Influent flow was then decreased in several increments
during the month of January 1973 in an attempt to starve or "burn out" the
filamentous culture or cultures and reestablish a "healthy" population.
Instead of eradicating the filamentous organisms, reduction of flow
and organic loading seemed to have the opposite effect of stimulating
proliferation. This proliferation was accompanied by the usual indicators
of a bulking sludge, i.e., a substantial increase in SVI, a rising sludge
blanket in the secondary clarifier, increasing suspended solids carry-over
in the final effluent, and operational difficulty in managing total system
sludge inventory.
In trying to determine the source of the filamentous intrusion, it was
postulated that the bacterial species (Sphaerotilus) may have developed in
the secondary clarifier sludge blanket during the aforementioned shutdown.
A local pharmaceutical firm is known to discharge mycelia into the Newtown
Creek sewer system and this was suspected as the source of the fungus
organisms. By the end of January 1973, with the influent flow reduced to
15 mgd, the SVI had risen from a summer background level of 45-50 to 85-100,
effluent suspended solids were exceeding 30 mg/1, effluent soluble BOD had
increased to over 20 mg/1, and the clarifier sludge blanket.was continuing
to rise. At this point a decision was made to "dump" the entire sludge
inventory, hose all settled sludge pockets out of the reactor and clarifier,
and start over. This second shutdown began on February 1, 1973.
After taking some additional time to make repairs to the sludge collection
mechanism while the clarifier was dewatered, Phase 3 commenced on February 18,
1973. Conservative loading rates were utilized initially based on the premise
that the best chance to prevent a reoccurrence of the filamentous condition
was a program of gradual and modest increases in F/M loading until the 20 mgd
design flow rate was reached. Relatively high MLSS concentrations of 5,000 mg/lt
-------
17
were maintained as a further measure to minimize F/M loading. However,
within several weeks a repeat of the experience encountered in Phase 2
became evident with the initial appearance of filaments in the oxygen
sludge. This time the organisms were definitely identified as fungus.
One possible reason for explaining the higher 1972-73 winter incidence and
enrichment pattern of these fungus organisms in the Newtown Creek oxygen
sludge as opposed to that facility's air sludges is the lower mixed
liquor pH inherent to the operation of the covered-tank oxygen system.
With the prospect of impending project failure a real possibility,
a joint decision (New York City, Union Carbide, and EPA) was reached to
accept the presence and proliferation of the fungus organisms as a cold
weather phenomenon and attempt to find an operating mode that would permit
satisfactory winter performance at design flow in spite of them. Accordingly,
a program of increased sludge wasting was initiated which eventually lowered
the MLSS concentration to less than 4,000 mg/1 and the SRT# to slightly
more than one day. At the same time influent flow was elevated in several
fairly rapid increments to 20 mgd (equivalent to an aerator detention time
of 1.5 hours based on Q). These steps yielded an F/M at the end of the
phase in the range of 0.75-0.80, considerably higher than the average of
0.57 for all of Phase 3. The altered operating philosophy proved to be
the correct decision, resulting in a controllable clarifier sludge blanket
and stable cold weather performance at design flow. Percentage removals
for the remainder of Phase 3, although not as high as Phase 1, were within
satisfactory limits. As the wastewater temperature increased during early
spring, the concentration of filamentous organisms diminished, and they
eventually disappeared in early May 1973.
During the summer of 1973,the Newtown Creek oxygen system exhibited
remarkable capability for absorbing high hydraulic and organic loadings
while still producing a high quality secondary effluent. Four diurnal
phases (Phases 4-7) conducted from April 8, 1973 through September 1, 1973
successively increased the average influent flow from 20.6 to 35.4 mgd.
During Phase 7 the average nominal aerator detention time was only 52 minutes
with corresponding average F/M, volumetric, and clarifier surface loadings of
* Defined as lb VSS under aeration/Ib VSS wasted in the waste sludge and
final effluent/day.
-------
18
2.44 Ib BOD5/day/lb MLVSS, 379 Ib BOD5/day/l,000 ft3, and 1,609 gpd/ft2,
respectively. These results confirmed and exceeded the high-rate
loading potential of oxygen-activated sludge first seen at Batavia.
At this point in the project it was possible to offer the following
interim status remarks:
1. The high-rate loading capability (nominal aeration time < one
hour) of oxygen aeration operating on Newtown Creek wastewater
during warm weather was conclusively demonstrated.
2. Prospects appeared promising that a modified method of operation
evaluated in late winter 1972-73 could circumvent the negative
effects of what may be an indigenous cold weather filamentous
condition with oxygen at Newtown Creek and permit satisfactory
performance at a flow rate at least equal to the design level
of 20 mgd (1.5 hours of nominal aeration time).
3. The operational measures employed to effect the improved
performance in late winter 1972-73, namely high F/M's and low
SRT's, occurred naturally to an even greater degree during the
high loading phases of summer 1973.
4. The above comments provide a tentative basis for speculating
that in some cases oxygen aeration may most beneficially be
employed at ultra high loading rates substantially exceeding any
which have been approved to date by State agencies.
Because of the importance attached to winter operation and performance,
the project was extended to the end of April 1974. The two major questions
which were to be addressed during the extended period were whether filamentous
organisms (particularly fungus) would again infest the oxygen sludge as
wastewater temperature dropped and, if so, would the modified method of
winter operation previously described permit continuous efficient per-
formance with a diurnal loading pattern centered around an average influent
flow rate of 20 mgd. If the first few months progressed without upset,
the flow rate was to be increased to 25 mgd and subsequently to 30 mgd
in the last 2-3 months of the winter season. The reason for holding this
latter option open was that if a year-round loading capability of 30 mgd
-------
19
could be demonstrated, the Newtown Creek Treatment Plant could conceivably
be satisfactorily upgraded by converting only 11 or 12 of the existing 16
bays from air aeration to oxygen aeration.
Data for the extended operating period (September 1973 through April
1974) are summarized along with the first seven project phases in a paper
entitled "Upgrading New York City Modified Aeration with Pure Oxygen."
This paper was prepared by New York City personnel (Nash, et al.) and
presented at the 47th Annual Conference of the Water Pollution Control
Federation. It is recommended that both this Technology Transfer report and
the New York City paper be reviewed in evaluating the Newtown Creek project.
The project will also be extensively documented and analyzed in the final
grant report now being prepared by the City. It is anticipated this report
will be available for distribution by mid-1975.
Another aspect of the project is discussed briefly below. Initially
the performance of the four-bed Pressure Swing Adsorption (PSA) Oxygen
Generator was less than satisfactory. During the 1972 summer startup
phase the unit was out of service due to mechanical problems roughly
40 percent of the time. These problems have since been largely corrected
and the generator now functions with a down-time that varies between
5 and 10 percent. During the 1972 startup difficulties one of the four
beds inadvertently became "loaded up" with water vapor. Subsequently, the
maximum achievable output of the unit was 10 tons of gas per day at
90 percent oxygen purity versus a design output of 16.7 tons of gas per
day at 90 percent oxygen purity. This necessitated an increase in
consumption of and reliance on the back-up liquid oxygen reservoir during
peak oxygen demand periods. The simplified three-bed (moving parts decreased
50 percent) PSA unit installed at Speedway, Indiana, has reportedly operated
at design output with high mechanical reliability following its installation
and startup.in mid-summer 1972.
-------
20
THE LAS VIRGENES PROJECT
A single-stage, covered-tank oxygenation system has been designed by
the Cosmodyne Division of Cordon International Corporation. The system,
given the name "SIMPLOX" and shown schematically in Figure 6, utilizes an
inflated dome-type cover to contain the oxygen-rich atmosphere over the
aerator. This concept is intended primarily for upgrading existing air
activated sludge plants with a minimum capital expenditure by utilizing
conventional air blowers and coarse-bubble air diffusers to recirculate
oxygen gas. Air blowers used in this service must be corrosion proofed
and otherwise modified to be compatible with oxygen gas. Virgin oxygen
gas is introduced to the aerator through a fine-bubble sparger located
on the tank bottom and on the opposite side wall from the conventional air
diffusers. Power required for oxygen dissolution is greater for the
"SIMPLOX" process than for the multi-stage systems because: (1) the
equipment used for transferring oxygen is modified air aeration equipment
and not specifically tailored to oxygen gas kinetics and (2) the gas phase
above the mixed liquor is completely mixed and assumes the same oxygen
composition as the exhaust gas stream; thus, the driving force for dissolving
oxygen in wastewater is less than in the lead stages of multi-stage aerators.
However, capital costs for converting an existing aerator from air to oxygen
service should be significantly less with the "SIMPLOX" approach because
staging baffles and multiple oxygen dissolution equipment assemblies are
not required. Since the gas phase is completely mixed, exhaust oxygen,
carbon dioxide, and inert gases can be bled from any point of the inflated
dome and any of several activated sludge flow configurations, including plug
flow, complete mix, and step aeration, can be used as desired.
A research and development grant was awarded to the Las Virgenes (Cali-
fornia) Municipal Water District (a suburb of Los Angeles) in June 1971 to
evaluate the "SIMPLOX" system at its Tapia Water Reclamation Facility. The
experimental program concluded on September 10, 1973. The District contributed
$62,000 in support of the project, supplementing the $186,000 EPA grant. An
empty nominal one mgd train was available for the oxygen study because of a
recent expansion at the Tapia facility. The manner in which the oxygen system
was incorporated into this existing train is shown in plan view in Figure 7.
-------
r
EXHAUST
GAS
c
LIQUID 02
STORAGE
02
VAPORIZER
PURE 02
FEED
r
I
__H i
PRIMARY
EFFLUENT
INFLATED DOME
GAS PHASE-COMPLETELY MIXED
RECYCLE
GAS
MODIFIED AERATION TANK
02 SPARGER
XLA-LJL , r 7FU,
TTTi U \ \
\
AIR
DIFFUSERS-^
RECIRCULATING
AIR COMPRESSOR
SECONDARY
EFFLUENT
RECYCLE SLUDGE
WASTE
SLUDGE
FIGURE 6. SCHEMATIC DIAGRAM OF DIFFUSED AIR AERATION SYSTEM MODIFIED TO
RECIRCULATE OXYGEN GAS, LAS VIRGENES PROJECT
-------
COVERED AERATION TANK
lS' WD
PRIMARY EFFLUENT
STEP FEED
PURE 02
GAS FEED
i i
t t
i
,\
\y QI SPARGER
AIR D1FFUSERS
???????
A A f A i
A 6 6 | 6 A
t J
hv
*
RECIRCULATING _/
AIR COMPRESSOR
--1
1
1
1
r
/ CIRCULAR \
ff CLARIFIER \
1 45' DIAxlO' I A
\^/ \
RECTANGULAR
CLARIFIER
]20'x20'xlO' WD
^SECONDARY
EFFLUENT "
-^
V^y' RECYCLE GAS 1 }
/ X 1 > EXHAUST
1 fJ GAS
FIGURE 7. FLOW DIAGRAM FOR LAS VIRGENES OXYGENATION SYSTEM
r-o
(S3
-------
23
Tha schedule followed during the experimental program for the project is
outlined in Table 10. The range of aerator loadings examined was not as
broad as at Newtown Creek due to influent flow limitations and a weaker
aerator feed (primary effluent at Las Virgenes, raw sewage at Newtown
Creek). The experimental program consisted of seven phases characterized
by increasing flow and system loadings. To effect a more pronounced
increase in aerator loading, only 45 percent of the available aerator
volume was utilized in the last five phases. This was accomplished via
the installation of a temporary bulkhead across the width of the aeration
tank after Phase 2.
TABLE 10. EXPERIMENTAL SCHEDULE FOR LAS VIRGENES
PROJECT (4/25/72-9/10/73)
Phase
1
2
3
4
5
6
7
Dates
4/25/72
9/11/72
1/22/73
3/9/73 -
4/4/73 -
5/1/73 -
5/15/73
- 7/31/72
-11/13/72
- 3/8/73
4/3/73
4/30/73
5/14/73
- 9/10/73
Influent
Flow
(mgd)
1.0
2.0
1.0
1.13
1.3
1.54
1.85
% of Aerator
In Use
100
100
45
45
45
45
45
No. of
Clarifiers
In Use
1
2
1
2
2
2
2
System performance for the Las Virgenes project is summarized in
Table 11. Tables 12, 13, and 14 summarize, respectively, system
sludge characteristics, aerator loadings, and secondary
clarifier loadings. Project data and information are presented in
more thorough fashion in the final grant report. This report, currently
being reviewed by EPA, is scheduled to be available for distribution by
the end of the first quarter of 1975
-------
24
TABLE 11. PERFORMANCE SUMMARY FOR LAS VIRGENES
(4/25/72 - 9/10/73)
Total BOD5 In (mg/l)#
Total BOD5 Out (mg/1)
% Removed
Total COD In (mg/1)*
Total COD Out (mg/1)
% Removed
Soluble COD In (mg/1)*
Soluble COD Out (mg/1)
% Removed
Susp. Solids In (mg/D*
Susp. Solids Out (mg/1)
% Removed
1
82
2
97
153
35
77
58
16
72
73
9
88
2
69
4
94
136
35
74
43
19
56
67
7
90
3
79
2
97
170
29
83
76
23
70
39
4
90
Phase
4
107
5
95
218
35
84
93
26
72
53
7
87
5
115
9
92
262
37
86
101
31
69
63
5
92
6
103
9
91
242
40
83
101
31
69
59
4
93
7
95
10
89
238
50
79
100
32
68
44
6
86
Turbidity Out (JTU) 232 3223
NH3-N In (mg/1)* 13.0 6.8 10.7 14.2 15.6 15.8 15.6
NH3-N Out (mg/1) 0.4 0.1 0.2 4.1 4.8 2.8 3.1
% Removed 97 99 98 71 69 82 80
NO3-N Out (mg/1) 16.2 15.3 8.8 6.9 5.6 7.5 8.0
*Concentrations shown are for primary effluent feed to oxygen aerator.
-------
25
TABLE 12. AVERAGE SYSTEM SLUDGE CHARACTERISTICS
FOR LAS VIRGENES (4/25/72-9/10/73)
Phase
MLSS MLVSS
(mg/1) (mg/1)
Return
Sludge
Flow
Return
Sludge
TSS
SVI
(ml /gram)
SRT
(days)
(% of Q) (mg/1)
1
2
3
4
5
6
7
3,700 2,950
3,750 3,050
3,815 2,950
3,570 2,715
3,050 2,485
2,595 2,170
2,535 2,115
30
30
32
32
40
39
40
TABLE 13. AVERAGE AERATOR
14,325
13,295
12,890
9,230
7,105
6,705
8,350
LOADINGS FOR
99
179
175
200
247
191
117
79
68
46
30
12
9
12
LAS VIRGENES (4/25/72-9/10/73)
Phase Detention
-Based on
(hr)
1 9.56
2 4.78
3 4.30
4 3.81
5 3.31
6 2.79
7 2.32
Time
F/M Loading
Ib BODs/day\
Ib MLVSS )
0.07
0.11
0.15
0.24
0.33
0.41
0.46
Volumetric Organic
Loading
/Ib BODs/dayX\
\ 1,000 ft
13
22
27
42
52
56
62
*/
-------
26
TABLE 14. AVERAGE SECONDARY CLARIFIER LOADINGS
FOR LAS VIRGENES (4/25/72-9/10/73)
Phase
1
2
3
4
5
6
7
Surface Overflow
Rate
(gpd/ft2)
417
501
417
283
326
386
464
Mass Loading
fib TSS/ft2^
\ day )
16.7
20.4
17.5
11.1
11.6
11.6
13.7
A cursory review of Table 11 reveals that effluent quality for the
entire Las Virgenes project was superb and surpassed that observed at
Newtown Creek. This can be attributed to three factors: (1) the lower
aerator organic loadings which permitted a high degree of COD insolubili-
zation, (2) the very conservative secondary clarifier surface and mass
loadings which promoted highly effective solids capture, and (3) the lack
of any significant industrial waste contributions. A major objective of
wastewater treatment in the Las Virgenes District is the production of an
ultra high quality secondary effluent after chlorination suitable for
agricultural reuse. The thrust of this project, therefore, was geared
not so much to maximizing system loadings (as was the case at Newtown Creek)
as maintaining truly superb quality effluent and determining the effect of
a relative conservative progression in system loadings on single-stage
nitrification. As shown in Table 15, virtually complete nitrification
was observed with F/M loadings between 0.07 and 0.15 Ib BOD /day/Ib MLVSS.
For F/M loadings between 0.24 and 0.46 Ib BOD5/day/lb MLVSS, nitrification
was only 69-82 percent complete. Lower wastewater temperatures may also
have played a role in the decreased nitrification of the latter four phases.
-------
27
TABLE 15. EFFECT OF ORGANIC LOADING AND
WASTEWATER TEMPERATURE ON NITRIFICATION
AT LAS VIRGENES (4/25/72-9/10/73)
Phase
1
2
3
4
5
6
7
F/M
/lb BODs/day\
^ lb MLVSS J
0.07
0.11
0.15
0.24
0.33
0.41
0.46
SRT
(days)
79
68
46
30
12
9
12
Wastewater
Temp . Range
70-77
73-79
65-67
65-67
67-70
68-71
70-75
% NH3-N
Removed
97
99
98
71
69
82
80
Fin. Eff.
N03-N
(mg/1)
16.2
15.3
8.8
6.9
5.6
7.5
8.0
Another major goal of the Las Virgenes staff was to minimize excess
biological sludge production as much as possible. This goal probably led
to the most significant problem area encountered on the project, a very
evident bulking sludge. No sludge was intentionally wasted from the
system during Phases 1 and 2. Wastage of suspended solids in the final
effluent and final clarifier skimmings was sufficient to balance net system
biomass growth at the low F/M loadings employed. Resulting SRT's were as
high as 79 days and the SVI climbed to a level near 200 ml/gram. Despite
the instigation of a scheduled wasting program in Phase 3, the sludge
continued to bulk and the SVI climbed even higher. It was not until Phase 7
at an SRT of 12 days and an F/M loading of 0.46 lb BOD5/day/lb MLVSS that a
significant drop in SVI occurred. The bulking sludge condition is attributed
here to a combination of Sphaerotilus filamentous development due to the
inordinately high SRT's and the accumulation of other poor settling debris
in the floe matrices. It was only because of the low clarifier loadings
that efficient overall performance was sustained. Sludge blanket levels
frequently rose to within a few feet of the clarifier weirs. The Las Vir-
genes experience illustrates the potential operating difficulties that can
and probably will occur at very low oxygen system loading rates.
-------
28
One definite conclusion reached during the project is that the inflated
tent (dome) concept is not suitable for permanent installation. New leaks
developed repeatedly in the polyvinyl material due to separation of the
tent/tank interface, abrasion against the tent support structure during
high winds, and bullets from pranksters' guns. The gas leak problem made
accurate oxygen consumption monitoring impossible, and during the latter
higher loading phases, the leaks became sufficiently frequent and large
that it was extremely difficult to maintain a mixed liquor D.O. above
1-2 mg/1. The rationale for using an inflated dome in lieu of a flat cover
on this research project was to permit access to the tank interior, a
procedure effectively utilized on several occasions. A permanent install-
ation would probably require a flat, more rigid cover for longevity and
minimization of leaks.
The Cosmodyne Division of Cordon International has not attempted to
establish a proprietary position with respect to the "SIMPLOX" system.
Notwithstanding the attractive capital cost features of this oxygen
dissolution concept for upgrading existing air-activated sludge plants,
without the support of a proprietary interest and an aggressive marketing
effort, utilization of this process in treatment plant construction will
most likely proceed at a much slower rate than with other oxygen processes.
THE FMC PROJECT
The FMC Corporation has developed a unique fine-bubble diffuser capable
of producing uniform oxygen bubbles of less than 0.2 mm in diameter. The
diffuser works on the shear principle of passing a high velocity liquid
stream at right angles to oxygen bubbles discharging into a vertical slot
from capillary tubes. Oxygen gas is introduced to the capillary tubes at
30 psi pressure. A graph provided by FMC showing water depth required for
complete dissolution of varying size oxygen gas bubbles is reprinted in Figure
8.* The large effect of a relatively small change in bubble size on the water
depth required for 100 percent dissolution is readily evident. For a bubble
diameter of 0.20 mm, a 17.5 foot deep tank would be required. The required
depth decreases to 8.5 feet for a 0.15 ram diameter bubble.
One of the many potential applications for this diffuser is in an open-
*This graph was prepared using tap water. Dissolution characteristics for
various size oxygen gas bubbles may and probably do differ for a wastewater
undergoing biological treatment.
-------
0.25
;
i
0.20
0.15
0.10
0.05
PARTIAL ESCAPE OF BUBBLES
BUBBLES COMPLETELY DISSOLVED
2 3 4 56789 10 15
DEPTH OF WATER REQUIRED FOR 100% DISSOLUTION - FEET
l
<
:
FIGURE 8. OXYGEN GAS BUBBLE DIAMETER VS. WATER DEPTH FOR COMPLETE DISSOLUTION
-------
30
tank oxygen-activated sludge process. To evaluate the feasibility of an
open-tank oxygenation approach, a research and development grant for
$142,000 was awarded to FMC in September 1972 for a nominal 30 gpm pilot
plant study. The firm is contributing over $75,000 of their funds to the
project. The pilot plant has been installed on the grounds of the Engle-
wood, Colorado (suburb of Denver), trickling filter plant and receives a
feed stream of primary effluent from that plant. Pilot plant configuration
and dimensions are shown in Figure 9. The aeration tank is provided with
two baffles to approximate a plug flow (three-stage) condition. Diffusers
are located in each of the stages. Mixed liquor is recirculated through
the diffusers by low head centrifugal pumps. Pump suction is taken near
the liquid surface to promote mixing and tank turnover. Throttling of the
oxygen feed is accomplished automatically by D.O. sensing and control.
Major points of research interest in the project are: (1) oxygen
utilization efficiency in an open-tank setting, (2) oxygen feed control
response based on a D.O. monitoring approach, (3) mixed liquor recirculation
rates and power requirements, (4) diffuser self cleansing (non-clogging)
capabilities, and (5) shearing effect, if any, on mixed liquor particles
caused by continuous recirculation through the pumps and diffusers. In
the event that floe disruption did occur, » short detention biological
reflocculation tank (gentle mixing, no chemicals) was interposed
between the aerator and secondary clarifier. Two aspects of system design
which cannot be adequately defined at the scale of this pilot plant study
are diffuser mixing characteristics and additional mixing requirements,
if any, for large aeration tanks. This task is being addressed by FMC in deep
tank tests using tap water at both the firm's Englewood and Santa Clara
laboratories.
Pilot plant fabrication was completed in late June 1973. System
startup required the first 20 days of July. The experimental program which
followed was divided into eight phases and is outlined in Table 16. Perform-
ance data for the four highest flow phases (Phase 4 through 7) are summarized
in Table 17. Sludge characteristics, aerator loadings, and secondary clarifier
loadings for the same four phases are presented in Tables 18, 19 and 20
respectively. Data for the first three conservative load phases as well as
Phase 8 which was still in progress at the date of this writing will be included
by FMC in the final project report. Availability of this report is expected
by mid-1975.
-------
MIXED LIQUOR
FLOCCULATION TANK
(IF NECESSARY)
CIRCULAR CLARIFIER
CENTER-FEED, RIM TAKEOFF
10' DIA x 10' WD
SECONDARY
EFFLUENT
MIXED LIQUOR
RECIRCULATING PUMPS
PRIMARY
EFFLUENT
FINE BUBBLE
DIFFUSER
RECYCLE SLUDGE
FROM GASEOUS
OXYGEN SUPPLY
THREE-STAGE
OPEN TANK AERATOR
8' LONG x 4' WIDE x IV WD
WASTE
"SLUDGE
FIGURE 9. FMC OPEN-TANK OXYGENATION PILOT SYSTEM
-------
32
TABLE 16. PLANNED EXPERIMENTAL PROGRAM FOR FMC PROJECT
Phase Dates
1 7/21/73 - 9/5/73
2 9/6/73 - 9/30/73
3 12/6/73 - 1/28/74
4 4/8/74 - 4/30/74
5 5/1/74 - 5/31/74
6 6/1/74 - 6/30/74
7 7/1/74 - 7/31/74
8 10/1/74 -10/31/74
TABLE 17. PERFORMANCE
(4/8/74
Total BOD5 In (mg/1)*
Total BOD5 Out (mg/1)
% Removed
Total COD In (mg/1)*
Total COD Out (mg/1)
% Removed
Suspended Solids In (mg/1)*
Suspended Solids Out (mg/1)
% Removed
Turbidity Out (JTU)
Sewage Temperature (°F)
Influent Flow No. of
Condition Clarifiers
in use
10 gpm
10 gpm
15 gpm
25 gpm
35 gpm
30 gpm
20 gpm
15 gpm
(Constant)
(Diurnal)
(Constant)
(Constant)
(Constant
(Diurnal)
(Diurnal)
(Constant)
1
1
1
2
2
2
1
1
SUMMARY FOR FMC PROJECT
- 7/31/74)
4
153
13
92
332
95
71
110
13
88
6
57
Phase
5 6
159 180
18 16
89 91
315 259
74 57
77 78
85 85
15 12
82 86
6 4
62 67
7
208
16
92
322
61
81
115
15
87
5
71
* Concentrations shown are for primary effluent feed to oxygen aerator.
-------
33
TABLE 18. AVERAGE SYSTEM SLUDGE CHARACTERISTICS
FOR. FMC PROJECT (4/8/74 - 7/31/74)
Phase
4
5
6
7
MLSS MLVSS Return
(mg/1) (.mg/1) Sludge
Flow
(% of Q)
5,120 4,010 11,585
4,030 3,435 10,485
4,745 3,860 12,220
3,960 3,365 10,850
Return SVI
Sludge (ml /gram)
TSS
(mg/1)
60 71
52 70
57 67
50 73
SRT
(days)
2.0
1.5
2.1
2.6
TABLE 19. AVERAGE AERATOR LOADINGS FOR
Phase
4
5
6
7
FMC PROJECT
Detention Time
-Based on Q-
(hr)
1.32
0.94
1.10
1.65
(4/8/74-7/31/74)
F/M Loading Volumetric Organic
/lb BOD5/day\
\ lb MLVSS J f
0.69
1.17
1.01
0.91
Loading
lb BODs/day^
1,000 ft3 J
173
253
244
190
TABLE 20. AVERAGE SECONDARY CLARIFIER LOADINGS
FOR FMC PROJECT (4/8/74-7/31/74)
Phase
4
5
6
7
Surface Overflow
Rate
(gpd/ft2)*
514
720
617
823
Mass Loadin
/ lb TSS/ft2
\ day
35
37
38
41
-V
Excludes influent center-well annular area which = 9.1% of total
clarifier surface area.
-------
34
As shown in Table 17, BOD^ and suspended solids removals during the
high loading conditions of Phases 4, 5, 6, and 7 were excellent. One of
the significant observations forthcoming from the project was that the
feared disruption of sludge settling properties due to floe shearing as the
mixed liquor was continually recirculated through the centrifugal pumps
and diffusers did not materialize. SVI for the above four phases averaged
a very acceptable 70 ml/gram. A highly concentrated float (4-6 percent TSS)
approximately six inches thick quickly developed on the surface of the
aerator after startup. This combination aeration/flotation effect was
anticipated in light of the fine bubbles created by the oxygen diffusers.
It was found that the thickness of the float can be controlled by adjusting
the elevation at which the mixed liquor recirculation suction is taken.
FMC personnel believe this feature offers a potential economically attractive
alternative location for extracting waste sludge from an activated sludge
system.
In a full-scale embodiment of this open-tank oxygen concept, mixed
liquor recirculation would not be accomplished by centrifugal pumps.
Instead, FMC envisions a propeller-type pump mounted inside a downcomer
draft tube. The draft tube in turn is to be connected to a pipe header
containing many gas bar diffusers. The whole assembly will rest on the
aeration tank floor and will be prevented from moving sideways by lateral
catwalks which are tied into the top of the draft tubes. Elevation, plan,
and side views of a full-scale aerator assembly as currently proposed are
pictured in Figure 10. A perspective view of a typical aeration tank
containing several of these assemblies and the resulting fluid mixing
pattern are shown in Figure 11. This system appears to possess the
essential ingredients for significantly impacting the wastewater treatment
construction industry.
-------
MOTOR PEDESTAL
THRUST BEARING
TOP OF COPING -
1
WATER SURFACE
^^^^-^b^vM T^S*rf^^ ^^^ ~~^J^^
'V ' l. '.'" ». *'* I .''' v V,' ' °« " .".»' * ".
SUPPORT
SECTION A-A
FIGURE 10. ELEVATION, PLAN, AND SIDE VIEWS OF ENVISIONED FULL-SCALE
EMBODIMENT OF FMC "MAROX" OPEN-TANK OXYGENATION SYSTEM
CO
Ln
-------
0
FIGURE 11. PERSPECTIVE VIEW OF ENVISIONED
FULL-SCALE "MAROX" OPEN-TANK OXYGEN SYSTEM
-------
37
THE BLUE PLAINS PROJECT
A multi-stage, covered tank oxygenation pilot system of Union Carbide
design (see Figure 12) was operated continuously from June 1970 through
September 1972 at the Joint EPA/District of Columbia (Blue Plains) Pilot
Plant. Nominal design throughput for the system was 70 gpm (100,000 gpd).
The results generated in over two years of work (believed to be the single
longest continuous oxygenation pilot plant study on record) were extensively
reported at the 1972 Water Pollution Control Federation Conference (Stamberg,
et al., 1972). For a detailed summary of monthly operating data, the reader
is referred to the upcoming publication of this paper in the Federation
Journal. Discussion of the project here is limited to generalized results
and observations.
The oxygen system was operated over a wide range of SRT's from 1.3
to 13.0 days. However, on District of Columbia (D.C.) primary effluent,
filamentous organisms propagate rapidly with either oxygen or air if the
SRT is held below approximately five days for any extended period of time,
producing a bulking sludge with greatly retarded settling rates. Conse-
quently, the majority of the Blue Plains operation has been intentionally
restricted to SRT's greater than five days. A technique devised by pro-
ject staff personnel of reducing the incoming flow and twin dosing the
sludge recycle stream with 200 mg/1 of hydrogen peroxide (based on influent
flow) for 24-hour periods at a one-week interval proved to be an effective
method for purging entrenched filamentous bacterial growths from an acti-
vated sludge system. The technique provides lasting benefit only if
subsequent F/M loadings are adjusted to maintain an SRT outside the critical
filamentous growth range. The conditions under which filamentous cultures
propagate and flourish are unique to each wastewater and location. Some
plants can operate in any desired loading range without encountering fila-
mentous problems. Oxygen mixed liquor at Blue Plains was normally well
bioflocculated and essentially free of fragmented debris between discrete
particles.
Above an SRT of five days, average system F/M loadings remained in
the range of 0.27-0.50 Ib BOD5/day/lb MLVSS. On those few occasions when
-------
02 RECYCLE
1
1
-r*h
INFLUENT
i
ill
£°
O
f
r
x -
b_
»
JL
*
"=
^
r
=
r
c
M
nK
0
io S
JL
ta
l^v_
0
«
^*
=1
hr
uTiTTl
EFFLUENT
WASTE
SLUDGE
FIGURE 12. SCHEMATIC DIAGRAM OF BLUE PLAINS OXYGENATION SYSTEM*
Reprinted with permission (Stamberg, et al., 1972)
u>
00
-------
39
the system was operated at an SRT less than five days, F/M loadings rose
to levels as high as 1.0 Ib BOD /day/Ib MLVSS. Corresponding average
volumetric organic loadings at an SRT above five days ranged from 57-185 Ib
3
BOD_/day/1000 ft . Aerator detention times (based on Q) were varied between
1.5 and 2.8 hours throughout the two-year+ period. For all loadings
investigated, BOD insolubilization was virtually complete. Effluent
soluble BOD,, residuals were never greater than 5 mg/1 and consistently
averaged 2-3 mg/1. Total BOD,, and suspended solids removal were a direct
function of clarifier performance. Effluent COD and TOC concentrations
typically ranged from 35-60 and 15-20 mg/1, respectively.
During the spring periods of rising wastewater temperature, nitrifi-
cation was established more slowly in the Blue Plains single sludge oxygen
system than in a parallel conventional single sludge air aerated pilot
system probably due to the lower mixed liquor pH inherent in operation of a
covered biological reactor. Once established, however, substantial nitri-
fication was exhibited by the oxygen system during warm weather. With
decreasing wastewater temperature in the fall, deterioration of nitri-
cation was directly related to SRT. At an SRT of 9.0 days and a wastewater
temperature of 63°F, at least partial nitrification was sustained. However,
once the wastewater temperature decreased to about 60°F, no nitrification
was observed in the Blue Plains oxygen system up to an SRT of 13.0 days.
Phosphorus removal experiments were conducted by adding aluminum
i » t
sulfate (alum) directly to the oxygen mixed liquor. At an Al /P weight
ratio of 1.4/1.0, phosphorus removal averaged 80% with total and soluble
phosphorus residuals of 1.8 and 1.6 mg/1 (as P), respectively. Increasing
ii i
the alum dose to an Al /P weight ratio of 1.8/1.0 decreased total and
soluble residuals to 0.62 and 0.53 mg/1 (as P), respectively, but it also
lowered mixed liquor pH from 6.5 to 6.0. At the lowered pH, the oxygen
biota eventually dispersed and the experiments were discontinued. For low
alkalinity wastewaters such as the District of Columbia's, pH control may
be necessary to achieve efficient (90% or greater) phosphorus removal when
acidic metallic salts are added directly to oxygen-activated sludge mixed
liquor.
-------
40
Oxygen clarifier performance at Blue Plains and its effect on total
system operation are addressed in a later section. Continued experiments
only recently completed at Blue Plains included evaluation of oxygen in a
step aeration flow regime and examination of the nitrification kinetics of a
second-stage oxygen system operating on full-scale D.C. modified aeration
effluent feed. Reports on these activities are in preparation.
THE BUREAU OF RECLAMATION PROJECT
The Bureau of Reclamation's Engineering and Research Center in Denver,
under an interagency agreement with EPA, has recently completed the second year
of a three-year project to test many different materials of construction to
evaluate their suitability for use with oxygen aeration wastewater treatment
systems. The materials being tested include three different types of con-
crete, twelve different metals, and eleven protective coatings, linings,
joint sealers, and gaskets.
The materials are being exposed for varying lengths of time to oxygen-
rich mixed liquor, oxygen-rich vapor above the mixed liquor, and to the
interface between the two phases and then withdrawn for examination. Oxygen
reactors being utilized for these tests include Las Virgenes; Speedway,
Indiana; and Fairfax County, Virginia. Interim results are available by writing
to EPA, Office of Environmental Engineering, Washington, D.C. (20460).
CRITICAL PROCESS PARAMETERS
Certain process parameters are vital to the successful operation and
economic attractiveness of all waste treatment processes. For oxygen
aeration systems, four of these process parameters are oxygen utilization
and consumption, sludge production, power consumption, and biological
performance versus biomass loading. Available data for the projects des-
cribed above are summarized below for each of the four parameters.
Oxygen Utilization and Consumption
A misconception which seems to have accompanied the development of the
oxygenation processes is that oxygen gas possesses mystical qualities and
can oxidize organics and ammonia nitrogen with less oxygen consumption than
air systems. In reality, of course, the same amount of oxygen is required
-------
41
to oxidize a given amount of organic carbon to carbon dioxide and water
or a given amount of ammonia nitrogen to nitrate nitrogen regardless of
the source of oxygen or the method in which it is delivered to a biologi-
cal system.
The conventional method of calculating oxygen consumption in a
covered-tank oxygen system is to monitor inlet and exhaust gas flows
and effluent D.O, and assume that all oxygen not accounted for was con-
sumed. This method will not detect any gas leaks which may develop in
and at the joints of the reactor cover. A second method which is sensitive
to detecting sizeable leaks and can be used to check the gross accuracy of
the oxygen metering equipment is an oxygen balance technique recommended
by the Blue Plains staff {Stamberg, et al., 1972) and shown in Table 21.
The method assumes that one pound of oxygen is consumed for every pound
of COD destroyed (not to be confused with COD removed from the substrate)
and that 4.57 pounds of oxygen are consumed for every pound of ammonia
nitrogen converted to nitrate nitrogen. The method is reasonably accurate
provided the wastewater does not contain certain industrial components
which do not consume oxygen in a COD determination but will utilize oxygen
in a biological system.
Oxygen utilization and supply data for Newtown Creek, Batavia,
Blue Plains, and the FMC project are summarized in Table 22. Accurate
measurement of oxygen utilization at Las Virgenes was hampered due to
excessive gas leaks in the tent cover previously described. The table
indicates a lack of informity for all three methods selected for
indicating specific oxygen supply requirements. Generally, in the
absence of nitrification, slightly more than one pound of oxygen should
theoretically be supplied for each pound of COD destroyed. The Newtown
Creek value of 0.8 for the period of September 17, 1972 through October 14,
1972 cannot possibly be correct and indicates probably either low inlet
gas measurements or low waste sludge COD determinations. Of the three
.methods shown, designing oxygen supply requirements on the basis of antic-
ipated BOD5 removal is the least reliable. Since COD destroyed usually
-------
42
TABLE 21. OXYGEN BALANCE METHODS
Method 1: Ib/mil gal Oxygen Supplied
(-) Ib/mil gal Exhaust Oxygen
(=) Ib/mil gal Oxygen Utilized
(-) Ib/mil gal Secondary Effluent D.O.
(=) Ib/mil gal Oxygen Consumed
Method 2: Ib/mil gal Aerator Influent COD
(-) Ib/mil gal Secondary Effluent COD
(-) Ib/mil gal Waste Sludge COD
(=) Ib/mil gal COD Destroyed
(+) Ib/mil gal Nitrate Nitrogen Oxygen Demand
(+) Ib/mil gal Exhaust Oxygen
(+) Ib/mil gal Secondary Effluent D.O.
(=) Ib/mil gal Oxygen Supplied (Theoretically)
f\
Assumes 1 Ib COD destroyed consumes 1 Ib 0~.
u ^
Assumes 1 Ib NH_-N converted to 1 Ib NO_-N consumes 4.57 Ib 0^.
cannot be accurately predicted in advance, using an oxygen required/
anticipated COD removal weight ratio of 0.60-0.75 and adding this to anticipated
nitrification oxygen demand, if any, is probably the best technique avail-
able for sizing oxygen supply equipment. The effect of nitrification on
oxygen supply and consumption is readily apparent at Blue Plains in May
1972. Generally, but not always, less oxygen was consumed per pound of
BOD removed as the F/M loading increased. A similar pattern was not evi-
dent for oxygen consumed per pound of COD removed.
-------
TABLE 22. SUMMARY OF OXYGEN UTILIZATION AND SUPPLY
Plant
°2
Newt own Creek
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
Phase 6
Phase 7
Batavia
5/12/69-11/10/69
9/1/70-11/30/70
Blue Plains
May 1971°
May 197 2d
FMC Proiect
Phase 4
Phase 5
Phase 6
Phase 7
Metered
Utilization
>90
>90
>90
>90
>90
>90
>90
93
92
97
97
-
F/M
/Ib BOD^/day\
^ Ib MLVSS }
0.65
0.57
0.57
0.92
1.19
1.62
2.44
0.59
0.87
0.97
0.36
0.69
1.17
1.01
0.91
Ib 02 Supplied
Ib BOD Removed
1.09
1.20
1.39
1.02
0.88
0.83
0.79
0.94
1.36b
1.04
2.09
1.22
1.11
0.92
1.08
Ib 02 Supplied
Ib COD Removed
0.55
0.59
0.65
0.55
0.72
0.73
0.61
0.60
1.13b
0.60
1.03
0.72
0.65
0.75
0.79
Ib 02 Supplied
Ib COD Destroyed
0.80a
-
-
-
-
-
-
-
1.09
1.23
w
acovers the segment of Phase 1 from 9/17/72-10/14/72 only.
bValues are high due to high clarifier loadings resulting in significant solids carryover and
lowered BOD5 and COD removals.
<=No nitrification.
Substantial nitrification.
-------
44
Sludge Production
Available data indicate that oxygen systems may produce less excess
biological sludge than air systems at comparable F/M loadings. The first
indication was provided by the two Batavia projects as shown in Figure 13.
In this figure, BOD5 removed per day per unit of MLVSS is plotted in the
conventional method against the inverse of SRT for both the air and oxygen
trains. The curves reveal an approximate 50% reduction in favor of oxygen.
Although both of these trains were operated in plug flow configuration, the
comparison is most likely heavily biased toward oxygen because of the severe
D.O. limitations under which the Batavia air reactor operated during these projects.
If oxygen does produce less sludge, it is probably due to the high mixed
liquor D.O. concentration maintained and the additional driving force it
provides for increasing oxygen penetration into and stimulating aerobic
activity within floe particle interiors. Air system sludge production
data in which the mixed liquor is not devoid of D.O. for a lengthy section
at the head of the aerator (in contrast to Batavia) will provide more
meaningful future comparisons with oxygen sludge production data.
Reliable sludge production data are available from Newtown Creek for
Phases 4-7. The data have been averaged for each phase and superimposed
on the Batavia sludge production curve in Figure 13. Three of the four
points plot reasonably close to the line of best fit (or that line extended)
for the Batavia excess sludge production data, lending credibility to this
projection of oxygen sludge production for raw wastewater feed. The fourth
point (Phase 4), which represents the least loaded condition of the four
phases plotted, falls considerably above the Batavia oxygen sludge pro-
duction curve. Additional sludge production data generated by the Newtown
Creek oxygen system during the extended operating period of September 1973
through April 1974 are presented in the previously mentioned New York City
authored paper "Upgrading New York City Modified Aeration with Pure
Oxygen." These additional data along with the data summarised in this report
provide an accurate representation of oxygen system sludge production at
Newtown Creek over a substantially broad range of loadings.
-------
45
K
e
co
CO
-------
46
Sludge production data for the Blue Plains oxygen system are compared
with data from a parallel step aeration air pilot system operated on the
same primary effluent feed in Figures 14 and 15 (Stamberg, et al., 1972).
Figure 14 plots SRT versus F/M loading and indicates that less volatile
mass under aeration was required with oxygen to reach any given SRT above
six days. Blue Plains personnel attribute the increased activity of the
oxygen volatile mass to maintaining mixed liquor D.O. between 4 and 8 mg/1
and the minimization of sludge pockets and dead spots afforded by independ-
ently controlled mixing. Figure 14 can be manipulated to produce Figure 15
by multiplying F/M values by the corresponding SRT values and inverting
the product to yield volatile solids produced per unit of BOD5 applied and
then replotting these new values against SRT. Figure 15 shows substantial
reduction in sludge production with oxygen again above an SRT of six days.
Below an SRT of about eight days, the step aeration system experienced
soluble BOD5 breakthrough and overall effluent quality was poorer than
that of the oxygen system. Because of the different flow configurations
utilized, sludge production information generated by these two systems
cannot be used directly to derive conclusions regarding the relative sludge
production rates of oxygen and air at comparable SRT's. The kinetics of
step aeration dictate that it will experience maximum sludge production
at a much higher SRT than a plug flow regime.
In an attempt to compare sludge production data from air and oxygen
systems with similar operating conditions and with the further restriction
that the air system is not oxygen transfer limited, data from the Hyperion
Plant in Los Angeles (Smith, 1969) are plotted against the Blue Plains data
in Figure 16 on a BOD,, removed basis and Figure 17 on a COD removed basis.
Hyperion is an air activated sludge plant with a consistent record of excellent
performance and adequate mixed liquor D.O. Both Hyperion and Blue Plains data
were collected on systems operated in a conventional plug flow mode on primary
effluent feed. In Figure 16 a majority of the oxygen points fall below the
Hyperion regression curve and Smith's computer program curve for Hyperion.
In Figure 17 all but two of the oxygen points fall below the air curve, many
of them well below. These two figures lend additional support to the position
-------
Ur-
0.8
0.6
0.
0-
0.4
0.2
0 2
STEP AERATION
I
I
I
I
4 6 8 10 12
SLUDGE RETENTION TIME-SRT (days)
14 16
FIGURE 14. BIOLOGICAL ACTIVITY RELATIONSHIPS - BLUE PLAINS
aReprinted with permission (Stamberg, et al., 1972)
a
-------
Ur-
0.8
OQ
- 0.6
0.4
0.2
I
I
I
i
46 8 10 12
SLUDGE RETENTION TIME-SRT (days)
14 16
FIGURE 15. EXCESS BIOLOGICAL SLUDGE PRODUCTION - BLUE PLAINS
SReprinted with permission (Stamberg, et al., 1972)
a
c»
-------
1.0 I
Of
O
UJ
to
oo
_j
O
\
0
LU
D
O
O
Qi
O.
U)
CD
_j
II
0.5
=HYPERION (AIR)
A =BLUE PLAINS
(OXYGEN )
1
1
±
0.2 0.4 0.6 0.8 1.0
LB BOD5 REMOVED/DAY/LB MLVSS IN AERATOR
1.2
FIGURE 16 . COMPARISON OF SLUDGE PRODUCTION FOR AIR AND OXYGEN SYSTEMS
ON PRIMARY EFFLUENT FEED (BOD5 REMOVED BASIS)
-------
1.0 r
to
in
CO
0.5
0
LU
u
D
Q
O
ci
QL
CO
t/)
>
co
_J
II
on
to
O
O
=HYPERION (AIR)
A=BLUE PLAINS
(OXYGEN)
J.
1
JL
1
_L
_L
L
0.4 0.8 1.2 1.6 2.0
LB COD REMOVED/DAY/LB MLVSS IN AERATOR
O -
2.4
Oi
O
FIGURE 17. COMPARISON OF SLUDGE PRODUCTION FOR AIR AND OXYGEN SYSTEMS ON
PRIMARY EFFLUENT FEED (COD REMOVED BASIS)
-------
51
that reduced sludge production is probable with oxygen, but not to the
degree predicted from the Batavia projects. Additional comparative data
under similar operating conditions and on the same wastewater feed are
needed. The current Detroit expansion is a good cand: date for supplying
such information when large-scale air and oxygen modules now in the
final stages of construction and/or startup are ready for parallel
operation.
Available sludge production data for Newtown Creek, Batavia, Las
Virgenes, and the FMC project and two typical months of data for Blue
Plains are summarized in three forms in Table 23 along with the corres-
ponding F/M loadings. This table illustrates the general futility of
attempting to correlate sludge production data from one location to
another by any of the three methods shown. Many local factors including
wastewater composition, wastewater temperature, mixed liquor D.O., and the
biodegradability rate of various organic constituents combine to influence
the amount of excess sludge that will be formed at different plants under
similar loading conditions. Table 23 also shows that even at a single
plant, unit sludge production on a BOD,, or COD removed basis does not
necessarily increase with increasing F/M loading, or vice versa. If any
rational method exists for representing sludge production at a single
location or comparing sludge production among locations, it is probably
the basic inverse SRT method utilized in Figures 13, 16, and 17, or one
of several published modifications of this basic method.
Power Consumption
One of the more attractive economic aspects projected for oxygenation
systems from the Batavia work is greatly reduced power requirements for
oxygen supply (generation) and transfer (dissolution) compared with the
power requirements of air blowers. Estimated installed HP load require-
ments as projected from the Batavia work are plotted against plant size
and BOD aerator loading for both oxygen and air systems in Figure 18.
The oxygen system band represents the additional power requirement for
mixing as aerator volume is increased from one hour detention to two hours
detention (based on Q). The air system band encompasses blower supply
rates from 0.8 to 2.4 cubic feet of air supplied per gallon of wastewater
-------
TABLE 23. SUMMARY OF OXYGEN SYSTEM SLUDGE PRODUCTION
Plant
Newtown Creekc
Phse 4
Phase 5
Phase 6
Phase 7
Bataviac
7/21/69-9/7/69
8/30/70-10/25/70
Las Virgenes^
Phase 1
Phase 2
Phase 3
Phase 4
Phase 5
Phase 6
Phase 7
FMC Project*3
Phase 4
Phase 5
Phase 6
Phase 7
Blue Plainsd
Sept. 1971
Feb. 1972
F/M Loading Ib
fib BOD5/day\
\ Ib MLVSS )
0.92
1.19
1.62
2.44
0.79
0.52
0.07
0.11
0.15
0.24
0.33
0.41
0.46
0.69
1.17
1.01
0.91
0.39
0.30
Waste Sludge TSSa
mil gal
1,300
1,185
1,055
1,025
970
1,250
-
91
103
250
250
123
1,086
730
752
878
620
430
Ib VSS Produced5
Ib BOD_ Removed
0.94
0.73
0.59
0.60
0.41
0.52
0.19
0.14
0.15
0.14
0.27
0.31
0.21
0.80
0.62
0.50
0.53
0.38
0.47
Ib VSS Produced5
Ib COD Removed
0.49
0.59
0.51
0.46
0.35
0.29
0.13
0.09
0.08
0.05
0.13
0.15
0.09
0.47
0.36
0.41
0.39
0.25
0.23
N>
^Includes waste sludge TSS only. cRaw sewage feed.
blncludes waste sludge and final effluent VSS. ^Primary effluent feed,
-------
53
LB BOD5 APPLIED TO AERATOR PER DAY
1,000 10,000 100,000
10,000
O
Z
ex.
a
z
111
13
x
G
tt
0
a.
I
CO
Z
1,000
100
10
i
HQA1R SYSTEM BLOWERS
OXYGEN SYSTEM
DISSOLUTION AND
GENERATION
EQUIPMENT
SYSTEM POWER ESTIMATES
BASED ON BATAVIA DATA
0.8 FT3/GAL
1.6 FT 3/GAL
2.4 FT3/GAL /
1,005 HP
1,041 HP
905 HP
1 HR DETENTION
2 MRS DETENTION
0NEWTOWN CREEK-
DESIGN CONDITION
NEWTOWN CREEK-
OPERATING CONDITIONS
A PHASE 1 AVG.
_ PHASE 7 AVG.
' l l l l I I
l ' I I I I I
1 10
PLANT SIZE-MGD
(BASED ON AVG. AERATOR INFLUENT BOD5 OF 130 mg/l)
FIGURE 18. NEWTOWN CREEK POWER CONSUMPTION SUPERIMPOSED
ON BATAVIA POWER PROJECTION CURVES
-------
54
treated. Using the median of the bands, 507, and 65% reductions in aeration
power requirements are projected for oxygen over air at plant sizes of 1
and 100 mgd, respectively.
Plotting the installed HP load for oxygen supply and dissolution at
Newtown Creek on this curve indicates oxygen system power requirements
estimated from Batavia may be somewhat optimistic. The installed HP load
at Newtown Creek is 1041 HP, as broken down in Table 24, for a design load
of 41,700 pounds of BOD per day.(calculated using a BOD5 of 250 mg/1 at
20 mgd). The Batavia curve predicts an installed nameplate requirement of
only 550-800 HP for the same design BOD- load.
TABLE 24. INSTALLED HP AT NEWTOWN CREEK FOR
OXYGEN GENERATION AND DISSOLUTION
Item Nameplate HP
1. PSA Compressor 450
2. Liquid Oxygen Vaporizer 96
3. 1st Stage Compressors at 40 HP ea 80
4. 2nd, 3rd, and 4th Stage Compressors at 40 HP ea 120
5. 1st Stage Mixers at 50 HP ea 100
6. 2nd, 3rd, and 4th Stage Mixers at 30 HP ea 180
7. PSA Cooling Tower Pumps 8
8. Instrument Air Compressor 7
Total 1,041
The strength of the Newtown Creek wastewater during the project
was somewhat weaker than the design projection. The actual BOD- load
for Phase 1 averaged only 29,650 pounds per day (29 percent less than
the design load) even though the average flow of 20.8 mgd was slightly
higher than the design flow of 20 mgd. However, actual power consumption
for the same period averaged 905 HP, a decrease of only 13 percent from
the installed load. This demonstrates a well-known fact that non-variable
speed drives operating below design load will consume almost as much power
-------
55
as when operating at design conditions. Consequently, power consumption
for Phase 1 when superimposed on Figure 18 falls well above the oxygen
band and up near the median of the air band.
The converse situation is illustrated in Phase 7 when, due to the
increased average influent flow of 35.4 mgd, the BOD load to the aerator
averaged 62,645 pounds per day. Oxygen supply and dissolution requirements
consumed an average power draw of 750 kilowatts (equivalent to 1,005 HP).
Thus, with an actual power consumption 3.5 percent less the installed load,
the system satisfactorily treated a BOD,, loading 50 percent greater than the
design load. For this phase, power consumption plots near the median of the
Batavia oxygen band in Figure 18.
From the above data it is apparent that actual unit power consumption
for oxygenation systems will approach installed unit power consumption
only when operating at or near design organic load. Another conclusion which
can be drawn from Newtown Creek experiences is that the oxygen module' s oxygen
transfer equipment was substantially overdesigned for the projected
BOD load.
Biological Performance Versus Biomass Loading
A strong point of oxygenation system performance noted wherever oxygen
has been tested is the relative insensitiveness of effluent quality to
changes in F/M loading. Data accumulated from Batavia, Newtown Creek, Blue
Plains, and Las Virgenes are plotted in Figure 19. These data indicate
plateaus for soluble BOD,, and soluble COD breakthrough of only about 15 and
60 mg/1, respectively, up to F/M loadings of 2.4 Ib total BOD5 applied/day/Ib
MLVSS. This denotes consistent and impressive performance under stressed
conditions. At the lower loadings employed at Blue Plains, essentially
complete insolubilization of BOD is evident. For all F/M loading rates,
however, total system efficiency for oxygen processes will be more directly
dependent on solids capture efficiency (clarifier performance) than on
biological performance deterioration.
-------
80
40
A
80 £
O
40
0
16
w
CQ
ft,
bj
W
a
O =BATAVIA
A=NEWTO\VN CREEK
Q =BLUE PLAINS
X =LAS VIRGENES
O
«
W 0
i i i
lilt
I
0.5
1.0 1.5
F/M LOADING - LB BOD5 APPLIED/DAY/LB MLVSS
2.0
2.5
FIGURE 19. EFFECT OF F/M LOADING ON OXYGEN
SYSTEM EFFLUENT SOLUBLE BOD5 AND COD
Ul
-------
57
SLUDGE SETTLING AND SYSTEM DESIGN
Perhaps the most important information generated by the Blue Plains
project has been a delineation of some of the factors affecting sludge
settling at that site and its resultant effect on system design (Stamberg,
et al., 1972). In addition to the retardent effect on sludge settling
previously mentioned due to filamentous infection of mixed liquor, other
factors which affected oxygen sludge settling rates at Blue Plains included
solids concentration, bulk sludge density (volatile solids fraction), and
wastewater temperature.
In the range of MLSS concentrations at which hindered or zone
settling occurs, it has been found that an equation in the form of v.= aC.
where v. = initial settling velocity
C. = initial solids concentration
a = intercept constant
n = slope constant
when plotted on log-log paper yields a straight line. Further, it has been
shown such a relationship exists for each of the three types of settling,
discrete particle, hindered, and consolidation settling (Dick, 1970)(Duncan,
et al., 1968). The change in slope between discreet particle settling and
hindered settling normally occurs at a C. between 2000 and 3000 mg/1. The
hindered settling zone is characterized by a discrete subsiding interface and
a zone of homogenously mixed settling particles. Clarifiers operating with
initial hindered settling are in reality operating as sludge thickeners.
It is essential that both hydraulic and mass loadings be considered in the
-------
58
design of secondary clarifiers for high solids systems. In many cases,
thickening (mass loading) requirements will control the -design. The
best available approach for evaluating thickening requirements appears
to be the batch flux (mass x settling velocity) method (Dick, 1970).
Bulk sludge density is a function of volatile solids fraction, i.e.,
density increases with decreasing volatile fraction. The incorporation
of denser inerts into the sludge mass is the primary reason why biomasses
developed on raw wastewater will generally settle better than those
developed on primary effluent. Another manner in which sludge density
is temporarily affected is the washing of accumulated inerts into a plant
from its sewer system during rain storms. This point was vividly illus-
trated at Blue Plains during a tropical storm the summer of 1972 as shown
in Figure 20 .
The least recognized parameter prior to plant startup that eventually
strongly affected oxygen sludge settling rates at Blue Plains was waste-
water temperature. Settling rates decreased significantly from summer to
winter operation. For example, during September and October 1971 (a
period when the oxygen clarifier was operated with a deep center feed
below the sludge blanket to capture unsettleable particles) as wastewater
temperature dropped from 81° to 71°F, the initial settling rate (ISR)
decreased from 10 ft/hr to 7 ft/hr in a 1-liter graduated cylinder test
at an MLSS concentration of 6000 mg/1 (see Figure 21). In November of
the same year the center feed was raised above the blanket in an attempt
to purge unsettleable particles from the system and increase bulk sludge
density. While this technique did increase the sludge density and tem-
porarily the ISR, a similar temperature effect was noted over the two-
month period of November and December 1971. As wastewater temperature
dropped from 70° to 63°F, the ISR decreased from 14 ft/hr to 9 ft/hr at
an MLSS concentration of 4500 mg/1 (see Figure 22). Conversely, as Blue
Plains wastewater temperature increased in the spring and summer of 1972,
substantial increasing settling rates were observed as illustrated in
Figure 23. The net result of this phenomenon was that a peak oxygen
2
clarifier overflow rate of 1940 gpd/ft was possible in the summer o£ 1970
with an MLSS of 8000 mg/1, while the peak overflow rate that could be
-------
59
100
80
60
50
40
~ 30
.£=
5± 20
Zl 10
LU
> 8
=1 6
£ 5
5 4
!!! 3
ec
H-
=£ 2
JUNE 22-JULY 3, 1972
(TROPICAL STORM)
10-20, 1972
3 4567 8910
20 30 40 60 80 100
INITIAL MLSS CONCENTRATION (gm/l)
FIGURE 20. ILLUSTRATION OF INCREASED SLUDGE DENSITY CAUSED BY RAIN
STORM AND ITS EFFECT ON INITIAL SLUDGE SETTLING VELOCITY
Reprinted with permission (Stamberg, et al., 1972)
-------
60
30
20
IT 15
10
8
CO
03
CD
CO
6
4
3
S 2
D.C.-Sept 1971-[78-81°F)
D.C.-Oct 1971-(71-73°F)
I II III I I
2 34 6 8 10 15 20
INITIAL MIXED LIQUOR CONCENTRATION (gm/l)
30
FIGURE 21. EFFECT OF DECREASING WASTEWATER TEMPERATURE ON INITIAL SLUDGE
SETTLING VELOCITY (SEPTEMBER-OCTOBER, 1971)a
Reprinted with permission (Stamberg, et al., 1972)
-------
61
30
20
I[ 15
**_
£ 10
C_3 0
£ 6
4
3
C9
c/?
S5 2
1
I
D.C. -Nov 1971-(68-70°F)
B.C.-Dec 1971-(63-64°F)
I I
I i I
I I
I
1 234 6 8 10 15 20 30
INITIAL MIXED LIQUOR CONCENTRATION (gm/l)
FIGURE 22. EFFECT OF DECREASING WASTEWATER TEMPERATURF^ON INITIAL SLUDGE
SETTLING VELOCITY (NOVEMBER-DECEMBER, 1971)
Reprinted with permission (Stamberg, et al., 1972)
-------
62
40
30
20
> MAY 20-31
1972
JUNE 10-
(70°F-74
~~i 1iir
JUNE 22-JULY 3,
(72°F-76°F)
10
8
7
6
5
2= 2
MAY 1-
(63°F-65
JULY
10-25,
F-79°F)
10
§F)
456 8 10
INITIAL MLSS CONCENTRATION (gm/l)
FIGURE 23. EFFECT OF RISING WASTEWATER TEMPERATURE ON INITIAL
SLUDGE SETTLING VELOCITY (SPRING-SUMMER, 1972)a
Reprinted with permission (Stamberg, et al., 1972)
-------
63
sustained in either the 1970-71 or 1971-72 winters without clarifier
2
failure was 975 gpd/ft at MLSS levels that varied from 3900-5300 mg/1.
Undoubtedly all of the factors discussed above contributed to the reduced
overflow rates necessary at Blue Plains to maintain satisfactory winter clarifier
operation. However, it appears that wastewater temperature played a
major role at this site. It is strongly emphasized though that the con-
clusions drawn from the Blue Plains project regarding wastewater tempera-
ture and sludge settling are not intended to imply that a similar effect
will be noted universally. Much additional data are needed to reach a
more definitive conclusion. Some additional data were collected over a
period of about 20 months at Newtown Creek (raw wastewater feed) and
Speedway, Indiana (primary effluent feed). Batch flux settling tests
were conducted periodically at both sites using slowly stirred
six-inch diameter, eight-foot long settling columns. Settling
velocity profiles as a function of initial MLSS concentration are plotted in
Fig. 24 for three runs conducted at Newtown Creek in December 1972 and June and
August 1973. These plots tend to verify the temperature effect observed
at Blue Plains. The decreased settling rates noted in the winter at Newtown
Creek are probably due to a combination of increased viscosity and drag of
the wastewater and alteration of biomass characteristics (proliferation of
filamentous organisms) at the colder water temperature.
Results of the long-term Blue Plains work illustrate clearly that
oxygen system design should be thought of as an integrated package con-
sisting of a biological reactor, a secondary clarifier, and sludge handling
facilities. The system should be designed for the worst anticipated climatic
conditions at a given site. Clarifier sizing should be specifically tailored
to the design and anticipated operating conditions of the reactor. There
are two basic ways of achieving a desired F/M loading: (1) a small reactor
and high MLSS or (2) a larger reactor and lower MLSS. If the first method
is selected to save on reactor costs, a larger clarifier will be necessary.
Both a small reactor and a small clarifier cannot be successfully mated in
a design unless greatly reduced MLSS concentrations are utilized. However,
opting for this selection will increase F/M loading, excess biological
sludge production, and required sludge handling capacity and costs.
-------
64
oi
sc
U
s
40
30
20
15
10
June 20, 1973 (22°C)
Aug. 8, 1973 (26°C)
Dec. 10, 1972 (17°C)
CO
H
H
2 34 6 8 10 15 20
INITIAL MIXED LIQUOR CONCENTRATION (GM/L)
30
FIGURE 24. SETTLING VELOCITY PROFILES FOR BATCH
FLUX SETTLING TESTS CONDUCTED AT NEWT OWN CREEK
-------
65
PROCESS ECONOMICS
No updating of the comparative cost estimates presented for air and
oxygen systems in the Batavia II final project report (17050 DNW 02/72)
has been attempted in this paper. Figure 25 summarizes estimated total
treatment costs (including amortization, operation, and maintenance) for
air and oxygen systems of 1-100 mgd capacity as taken from the Batavia II
final report. Interest was figured at 5-1/2% over 25 years. Oxygen
supply costs are based on on-site generation plant purchase by the
municipality. Projected savings in the cost of oxygen by buying and
operating your own plant as opposed to commodity across-the-fence purchase
current at the time of printing of the Batavia I final project report
(17050 DNW 05/70) are shown in Figure 26.
Figure 25 projects average savings in total treatment costs of about
20% with oxygen for plants of 20-100 mgd. It must be remembered, however,
that built into these curves are: (1) the assumption questioned by many
observers that oxygen reactors will universally be one-third as large as
air reactors for equal treatment and (2) what is believed to be overly
optimistic estimates of the difference in sludge production rates between
air and oxygen processes. The author concludes the one area in which
oxygenation may have a very decided economic advantage is in the upgrading
of existing overloaded secondary plants, such as Newtown Creek. Also it
is likely that many decisions to install oxygen are not made so much on
the basis of economics as on the basis of the high process reliability
and stability and the rapid recovery following toxic upsets afforded by
an enriched oxygen biological system.
-------
66
CO
CO
h-
CO
17
16
15
14
13
12
11
10
8
OXYGEN AERATION
I
20
AIR AERATION
I
40 60
PLANT SIZE-MGD
80
100
FIGURE 25. TYPICAL RANGES FOR TOTAL TREATMENT COSTS FOR NEW PLANTS
WITH PRIMARY SEDIMENTATION PROJECTED FROM BATAVIA STUDIES
-------
120
100
z
o
80
Ul
y 60
CE
CL
Z
LJ
x 40
O
20
/LIQUID PURCHASE
ON-SITE
PLANT PURCHASE
j i
ON-SITE GAS PURCHASE
_i i
0.2 0.5 I 2 5 10 20
OXYGEN USAGE RATE TONS / DAY
50 100 200 500 1000
FIGURE 26. OXYGEN COSTS AS A FUNCTION OF USAGE RATE
-------
68
CONTINUING DEVELOPMENTS
Continuing Research and Development Projects
Continuing EPA research and development efforts are underway or recently
completed in the following areas: (1) evaluation of second-generation oxygen
dissolution approaches, (2) examination of oxygen nitrification kinetics both
in single-stage and two-stage systems, (3) definition of viable alternatives
for combining chemical phosphorus removal with oxygen aeration, (4) determin-
ation of the most cost-effective sludge handling and dewatering techniques for
taking advantage of the excellent thickening properties of oxygen sludge,
(5) examination of sludge settling characteristics, (6) investigation of
aerobic sludge digestion with oxygen gas, and (7) a study of the safety
aspects of using oxygen in a wastewater treatment plant environment.
The development and maturation of new wastewater treatment processes are
usually accelerated by the parallel development of several proprietary systems,
However, because of this more rapid development, certain process details and
aspects not directly associated with the treatment of wastewater often do
not receive as thorough an evaluation as may be desirable and prudent.
The safety aspects and requirements of utilizing oxygen in activated sludge
treatment (No. 7 on the above list) are believed to represent one such aspect.
Although each firm marketing an oxygen aeration system has undoubtedly
considered safety features and requirements for its particular system,
no comprehensive generalized treatment of the subject has been undertaken.
until recently. Of particular concern is the processing of wastewaters
which periodically contain hydrocarbons and other volatile substances in
covered aeration systems with oxygen atmospheres ranging anywhere from 50
to 95%. The fundamental safety ramifications of using oxygen in this type
of duty have needed an in-depth review and evaluation by an independent
investigative team. A standard safety manual has also been urgently needed
to instruct waste treatment plant designers and operators in the safe and
proper handling of oxygen and to identify essential safety equipment and
instrumentation. Such a manual must be sufficiently broad and comprehensive
to apply to any rational concept for dissolving oxygen in wastewater.
-------
69
Due to lack of funds, this project was delayed for more than
a year. In June 1974, however, a contract was awarded to the Rocketdyne
Division of Rockwell International to undertake this study. The
purpose of the project is twofold:
(1) To evaluate the fundamental ramifications and implications from
a safety standpoint of using oxygen gas or oxygen enriched air
for aeration of activated sludge systems and based on this
evaluation to recommend an implementive course of action which
will ensure the safety and security of wastewater treatment plant
personnel and facilities.
(2) To develop a standard safety manual and safety checklist for
the safe and proper handling of oxygen in a wastewater treat-
ment plant environment that will apply in principle to any
rational oxygen dissolution concept.
Oxygen Process Implementation
Oxygenation systems are being designed and constructed for many
treatment plants across the country to meet a variety of new plant con-
struction and plant upgrading needs. At the time of this writing, 48
known municipal oxygen systems are in various stages of design, construction,
startup, or operation. As summarized in order of decreasing size
in Table 25, the total design flow of these 48 plants is 2,714.8 mgd ranging
in capacity from 0.9 to 600 mgd. Of the 48 plants,37 are designed using
surface aerators for oxygen dissolution,eight with submerged turbines, one
with fine-bubble diffusers (open-tank), one (Las Virgenes) employed a
converted air blower and air diffusers, and one is still undetermined.
Oxygenation is also beginning to make inroads into the industrial
wastewater treatment picture. As indicated in Table 26, eight systems
are currently being constructed and/or operated to treat a variety of
industrial wastes. The total design flow of these_eight systems is 62.9
mgd ranging in siz;e from 1 to 25 mgd. Six Oxygenation systems are now also
on-line in Japan.
Preliminary oxygen designs are being prepared for 30-40 additional
plants still in the negotiating phase. It appears that oxygen aeration
is definitely here to stay.
-------
70
TABLE 25. MUNICIPAL WASTEWATER TREATMENT PLANTS UTILIZING OXYGEN
Plant
1.
2.
3.
Detroit, Mich.
Detroit, Mich.
Philadelphia, Pa.
Design
Flow
(mgd)
600
300
210
Type of
Dissolution
System
Submerged Turbines
Submerged Turbines
Surface Aerators
Status
(Oct. 1974)
Design
Startup
Design
(Southwest)
4. Philadelphia, Pa. 150
(Northeast)
5. New Orleans, La. 122
6. Middlesex County, N.J. 120
7. Oakland, Calif. 120
(East Bay M.U.D.)
8. Dade County, Fla. 120
9. Louisville, Ky. 105
10. Wyandotte, Mich. 100
11. Denver, Colo. 97
12. Baltimore, HH. 73
13. Tampa, Fla.a 60
14. Miami, Fla. 55
15. Duluth, Minn. 42
16. Hollywood, Fla. 36
17. Cedar Rapids, Iowa 32.9
18. Harrisburg, Pa. 31
19. Springfield, Jfo. 30
20. Salem, Ore. 26.5
21. Danville, Va. 24
22. Euclid, Ohio 22
23. Ft. Lauderdale, Fla. 22
24. Littleton/Englewood, Colo. 20
25. New York, N.Y. (Newtown Creek) 20
26. Decatur, 111. 17.7
Surface Aerators Design
Surface Aerators Const.
Submerged Turbines Const.
Submerged Turbines Const.
Surface Aerators Design
Submerged Turbines Const.
Submerged Turbines Const.
Surface Aerators Const.
Surface Aerators Design
Surface Aerators Design
Surface Aerators Design
Surface Aerators Design
Surface Aerators Const.
Surface Aerators Design
Surface Aerators Design
Surface Aerators Design
Surface Aerators Const.
Surface Aerators Const.
Surface Aerators Const.
Surface Aerators Design
Surface Aerators Design
Submerged Turbines Oper.
Surface Aerators Const.
aTwo-stage oxygenation system.
(CONTINUED)
-------
(CONTINUED) 71
TABLE 25. MUNICIPAL WASTEWATER TREATMENT PLANTS UTILIZING OXYGEN
Plant
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
Fayetteville, N.C.
Chicopee, Mass.
New Rochelle, N.Y.
Muscatine, Iowa
Winnipeg, Manitoba
Fairfax County, Va.
Brunswick, Ga.
Tauton, Mass.
Morgantown, N.C.
Lebanon, Pa.
Fairbanks, Alaska
Speedway, Ind.
Deer Park, Tex.
Lewisville, Tex.
Mahoning County, Ohio
Jacksonville, Fla.
Calabasas, Calif.
(Las Virgenes, M.W.D.)
Littleton, Colo.
Cincinnati, Ohio
Hamburg, N.Y.
Minneapolis, Minn.
Chaska, Minn.
Design
Flow
(mgd)
16
15
14
13
12
12
10
8.4
8
8
8
7.5
6
6
4
3.4
1.8
1.5
1.2
1
1
0.9
2,714.8
Type of Status
Dissolution (Oct. 1974)
System
Surface Aerators
Surface Aerators
Submerged Turbines
Surface Aerators
Surface Aerators
Surface Aerators
Surface Aerators
Surface Aerators
Surface Aerators
Surface Aerators
Undetermined
Surface Aerators
Surface Aerators
Surface Aerators
Surface Aerators
Surface Aerators
Converted Air
Diffusers
Surface Aerators
Surface Aerators
Surface Aerators
Fine- Bubble
Diffusers (Open Tank)
Surface Aerators
Const .
Design
Design
Const.
Oper.
Oper.
Oper.
Design
Const .
Design
Rebid
Oper.
Oper.
Const .
Const .
Startup
Oper.
Discontinued
Oper.
Const .
Oper.
Const .
Oper.
Treatment of Zimpro supernatant.
-------
TABLE 26. INDUSTRIAL WASTEWATER TREATMENT
PLANTS UTILIZING OXYGEN
72
Plant
Design
Flow
(mgd)
Type of
Dissolution
System
Status
(Oct. 1973)
1. Container Corp.
(Fernandina Beach, Fla.)
Q
2. Chesapeake Corp.
(West Point, Va.)
a
3. Gulf States Paper, Inc.
(Tuscaloosa, Ala.)
4. Union Carbide Corp.
(Sistersville, W. Va.)
b
5. Union Carbide Corp.
(Taft, La.)
Q
6. American Cyanamid Co.
(Pearl River, N.Y.)
7. Standard Brands
(Peeksville, N.Y.
b
8. Hercules, Inc.
(Wilmington, N.C.
25
Surface Aerators Const.
16.3 Surface Aerators Oper.
10
Surface Aerators Startup
4.3 Surface Aerators Oper.
3.8 Submerged Turbines Const.
1.5 Surface Aerators Oper.
Surface Aerators Oper.
Surface Aerators Oper.
62.9
eulp and Paper
Petrochemical
c
Pharmaceut ica1
Distillery
-------
73
SUMMARY
1. Oxygenation systems are equally applicable to new plant construction
and upgrading of existing overloaded secondary treatment plants.
2. Oxygenation systems should be designed as integrated packages consisting
of a biological reactor, a secondary clarifier, and sludge handling
facilities.
3. There are genuine indications that reduced excess biological sludge
production is possible with Oxygenation; however, additional verifying
data are needed.
4. Research and development will continue on many areas of the total
Oxygenation process to fully exploit the potential of the process.
-------
74
REFERENCES
"Continued Evaluation of Oxygen Use in Conventional Activated Sludge
Processing", U. S. Environmental Protection Agency, Water Pollution
Control Research Series Report Number 17050 DNW 02/72, February 1972.
Dick, R. I., "Thickening" in Water Quality Improvement by Physical and
Chemical Processes, Volume II, University of Texas Press, 1970.
Divet, L., P. Brouzes, and P. Pelzer, "Short Period Aeration Studies at
Paris", presented at Annual Meeting of New York Water Pollution Control
Association, New York City, January 1963.
Duncan, J. and K. Hawata, "Evaluation of Sludge Thickening Theories",
Journal Sanitary Engineering Division, ASCE, 94, Number SA2, April
1968.
"Investigation of the Use of High Purity Oxygen Aeration in the Conventional
Activated Sludge Process", U. S. Department of the Interior, Federal
WaterQjality Administration, Water Pollution Control Research Series
Report Number 17050 DNW 05/70, May 1970.
Smith, R. and R. G. Eilers, "A Generalized Computer Model for Steady-State
Performance of the Activated Sludge Process", U. S. Environmental Pro-
tection Agency, Water Pollution Control Research Series Report Number
17090 ... 10/69, October 1969.
Stamberg, J. B., D. F. Bishop, A. B. Hais, and S. M. Bennet, "System Al-
ternatives in Oxygen Activated Sludge", presented at 45th Annual Water
Pollution Control Federation Conference, Atlanta, October 1972.
it US. GOVBillMEKT PR1KT1N6 OFFICE 1974 657-051/1058
------- |