EPA RESEARCH AND DEVELOPMENT
ACTIVITIES WITH OXYGEN AERATION
TECHNOLOGY TRANSFER DESIGN SEMINAR
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND MONITORING
NATIONAL ENVIRONMENTAL RESEARCH CENTER
CINCINNATI, OHIO
-------
EPA RESEARCH & DEVELOPMENT
ACTIVITIES WITH OXYGEN AERATION
Prepared for the
U. S. Environmental Protection Agency
Technology Transfer Design Seminar
Presented at
Pittsburgh, Pa., August 29-31, 1972
National Environmental Research Center
Advanced Waste Treatment Research Laboratory
Office of Research & Monitoring
Cincinnati, Ohio
-------
EPA Research & Development
Activities with Oxygen Aeration
by
John B. Stamberg*
INTRODUCTION
The concept of using oxygen enriched air or pure oxygen as the aeration gas
in the activated sludge process dates back over twenty years to Pirnie and Okun's
"bio-precipitation" model. Since then, many investigators have acknowledged
that oxygen gas has inherent characteristics which could be used to advantage
in activated sludge processing. On the other hand, during this same period,
oxygen aeration was not considered economically practical because existing gas
contacting mechanisms were not capable of effectively dissolving and utilizing
the relatively expensive oxygen commodity.
In 1968, a $528,000 research contract was awarded by EPA's forerunner,
FWPCA, to the Linde Division of the Union Carbide Corporation to evaluate an
oxygen aeration system which offered promise for overcoming the poor utilization
factor. This project (herein after referred to as Batavia I) was carried out
in 1969 at the Batavia, New York Municipal Pollution Control Plant (nominal
capacity - 2.5 mgd) and served as the foundation which stimulated the exceptional
growth and development of this process. In the short span of three years, oxygen
aeration has come full cycle from a research undertaking to the threshold of
widespread acceptance and utilization by the municipal and industrial waste
treatment fields.
* Sanitary Engineer, Municipal Treatment Research Program, Advanced Waste
Treatment Research Laboratory, National Environmental Research Center,
Cincinnati, Ohio.
-------
-2-
In addition to Batavia I, EPA has sponsored four other projects involving
various aspects of oxygen aeration. They include the following:
1. A follow-on research contract to Union Carbide for $209,000 primarily
to investigate handling of oxygen sludges (carried out at Batavia in
1970 and hereinafter referred to as Batavia II).
2. A $1.5 million demonstration grant to New York City to demonstrate
and evaluate the Union Carbide oxygen aeration system on a 20 mgd
scale at the Newtown Creek Treatment Plant.
3. A $160,000 R&D grant to the City of Las Virgenes, California to evaluate
a different oxygen dissolution concept developed by Cordon International
(formerly Cosmodyne Corporation).
4. An in-house research project conducted at the EPA/DC Pilot Plant in
Washington, D.C. for the past two years (hereinafter referred to as
the Blue Plains Project).
The results of the two Batavia projects are discussed briefly in the
following pages. Union Carbide will also refer to these projects in their
presentation. A description of the Newtown Creek demonstration project which
is just now in the final equipment checkout and system start-up phases will be
given prior to the plant tour on Thursday, March 2. The Las Virgenes project
is also just beginning system checkout and, not yet having generated any
operating data, will only be explained briefly at this time. Major emphasis
will be devoted in this text to a general discussion of the on-going 100,000 gpd
Blue Plains project. This is the longest continuous oxygen aeration study
undertaken at one site to examine design parameters of the process. The Blue
Plains data accumulated and refined over the two-year test period represents
-------
-3-
the major contribution which the Agency can add to this design seminar.
GENERAL DISCUSSION - BATAVIA I and II
A flow diagram of the Batavia, New York Municipal Pollution Control Plant
is shown in Figure 1. The plant consists of two mirror-image treatment trains
(nominal capacity - 1.25 mgd each) with separate secondary clarification and
sludge recycle facilities. The Batavia Plant does not utilize primary settling
of wastewater. A six-stage oxygen aeration system was installed in one of the
two aeration tanks. Liquid oxygen was stored on site and vaporized prior to
introduction to the first stage. The performance of the oxygen aeration system
was evaluated and compared to that of the parallel conventional plug flow diffused
air aeration system over a seven month period during Batavia I.
The oxygen aeration system which was installed at Batavia is illustrated
schematically in Figure 2. A gas tight cover is utilized to prevent venting
to the atmosphere. The aeration tank is segmented into stages by vertical baffle
walls. Each stage is equipped with a recirculating compressor and a combination
submerged turbine-rotating sparger. Oxygen gas is fed to the first stage and
along with gaseous decomposition products and inert gases recirculated in each
succeeding stage. Oxygen feed rates to the first stage and gas exhaust rates
from the last stage are automatically controlled.
The pertinent results of the Batavia I project are summarized below:
1. The oxygen aeration system exhibited excellent oxygen aeration transfer
capabilities with overall utilization efficiencies in excess of 90%.
2. The power required for oxygen gas dissolution was 1/5 to 1/6 of that
required by the air blowers for the parallel air system.
-------
-4-
3. The oxygen system was able to sustain high mixed liquor suspended
solids (MLSS) concentrations in the order of 6000-7000 mg/1 in a
highly aerobic environment (8-10 mg/1 mixed liquor dissolved oxygen)
with relative ease. The highest MLSS concentration that could be
maintained by the coarse bubble diffused air system without causing
anoxic conditions was about 3600 mg/1.
4. The high solids carrying capabilities of the oxygen system permitted
effective treatment of the incoming raw wastewater (average BOD,
160-260 mg/1) with nominal detention times of 1 1/2 hours (based on
raw flow) and volumetric organic loadings greater than 200 Ib BOD /day/
3
1000 ft mixed liquor, conditions which would constitute an excessive
overload to typical air system reactors.
5. Equivalent treatment was provided by the oxygen system to that of the
air system in 1/3 as much aerator volume.
6. The oxygen system produced only 50-60% as much excess biological
sludge as the air system under the conditions tested.
7. Both systems exhibited good sludge settling and compaction character-
2
istics. Average daily clarifiers loadings up to 1500 gpd/ft could
be sustained without effluent deterioration. This is at least
partially attributed to the absence of primary clarification and the
resulting denser sludge. Neither Batavia project was conducted during
the winter when sludge settling characteristics for any biological
system can be expected to be their poorest.
Oxygen sludge handling characteristics were evaluated on a pilot-scale in
the Batavia II project. The most significant result of this follow-on project
-------
WATER POLLUTION CONTROL PLANT, CITY OF BATAVIA, N.Y.
SCHEMATIC FLOW DIAGRAM
SLUDGE
DIGESTER
No. 2
FIGURE 1
PLANT
EFFLUENT
CHLORINE
CONTACT
TANKS
THICKENED
SLUDGE
DIGESTED
SLUDGE
SLUDGE
DISPOSAL
02
STORAGE
KEY
SEWAGE FLOW
SLUDGE FLOW
DESIGN POPULATION 25,000
AVG FLOW: 2.5 MIL. GAL /DAY
MAX.FLOW: 6.25 MIL GAL./DAY
-------
AERATION
TANK COVER
OXYGEN
FEED~GAF
WASTE
LIQUOR*
FEED
RECYCLED
SLUDGE
r
FIGURE 2
SCHEMATIC DIAGRAM OF MULTI-STAGE
OXYGENATION SYSTEM
UTILIZING TURBINE-SPARGERS
& RECIRCULATING COMPRESSORS
GAS RECIRCULATION
COMPRESSOR
r \
STAGE
BAFFLE
r \
EXHAUST
GAS
MIXED LIQUOR
"EFFLUENT TO
CLARIFIER
.PROPELLER
SPARGER
-------
-5-
was discovering that settled oxygen mixed liquor withdrawn from the clarifier
underflow could be satisfactorily dewatered on a vacuum filter without prior
ty
thickening or digestion. Cake yields up to 4.5 Ib/hr/ft were obtained on a
2% solids feed at a cycle time of about 2.5 minutes/revolution. The best
chemical conditioner was ferric chloride at a dosage rate of approximately
200 Ib/ton of feed solids. Under these conditions, the solids content of the
cake averaged around 15%. Data obtained through the courtesy of the City of
Milwaukee indicated that these cake yields are 40-45% higher than Milwaukee
achieves on its gravity thickened air generated waste activated sludge, with
the same filter parameters and chemical conditioner.
Data from the two Batavia projects were used to develop estimated total
treatment costs for an oxygen aeration system and a conventional diffused air
aeration system. The costs shown in Figure 3 in cents/thousand gallons treated
were estimated for new plants and include operation, maintenance and capital
amortization (5-1/2%, 25 yrs.) costs for primary treatment, secondary treatment,
and sludge handling and disposal facilities. In the 1 mgd range, there was
projected to be no cost advantage with either system. Based on the given design
assumptions, as plant size increases, cost savings begin to accrue to the oxygen
system and are projected to amount to 15-20% at the 100 mgd plant size.
GENERAL DISCUSSION - LAS VIRGENES PROJECT
The oxygen aeration system being installed at a 2 mgd scale in the Las
Virgenes, California Treatment Plant and shown schematically in Figure 4 is
intended for application to existing activated sludge plants equipped with
centrifugal air compressors and air diffusers. Oxygen gas is introduced through
a sparger to the head of the aeration tank. Gas is continuously recycled with
-------
-6-
the converted air aeration equipment. Exhaust gas is bled from the system
at the exit end of the aeration tank.
The objective of this project is to demonstrate that oxygen aeration
can be used successfully to increase organic loading capacity of existing plants
with minimum capital expenditure. The additional investment consists of a tank
cover (which at Las Virgenes will be an inflated dome type structure of material
similar to hospital oxygen tents), corrosion-proofing vital elements of the
centrifugal air compressors (positive displacement blowers are not suitable),
replacing the air compressor seals with oxygen compatible materials, gas piping,
instrumentation, and an oxygen supply source. Stage baffles and multiple stage
oxygen compressors and turbine-spargers are not used in this concept.
However, the utilization of existing equipment initially designed for
conventional air aeration will result in less efficient gas transfer and higher
specific power consumption than the multi-stage Union Carbide system. Because
the Las Virgenes system is a single gas stage system, vapor phase gas will become
completely mixed and assume a uniform composition identical to that of the
exhaust gas. The available driving force for dissolving oxygen will be less
than that available in the lead stages of the multi-stage concept.
At the conclusion of the project, the trade-offs (reduced capital costs
at the sacrifice of increased operating costs) will be compared and their
significance analyzed. Cordon International anticipates that its single stage
system will produce equivalent treatment to the multi-stage system with volumetric
organic loadings up to about 80% of those which can be handled by the multi-stage
system.
-------
18
17
LU
Lkl
£ 16
2 15
5 14
CO
5 13
kU
X
2 12
t—
h-
2 11
10
I I
TYPICAL RANGES
TOTAL TREATMENT COSTS
NEW PLANTS WM PRIMARY SEDIMENTATION
FIGURE 3
OXYGEN AERATION
1
20
AIR AERATION
I
40 60
PLANT SIZE M60
80
100
-------
OXYGEN SUPPLY
I
FIGURE 4
SCHEMATIC DIAGRAM OF DIFFUSED AIR
AERATION SYSTEM MODIFIED TO RECIRCULATE OXYGEN GAS
EXHAUST GAS
RECIRCULATING '
AIR COMPRESSOR '
—f-
OXYGEN *_o
ft i c rrrn
GAS FEEDsfe*,
I AERATION TANK COVER
WASTE-
WATER
I
I
GAS PHASE-COMPLETELY MIXED
LIQUID FLOW
AIR DIFFUSERS
jL? T ? ? ? ? ? O
FINAL
EFFLUENT
SECONDARY CLARIFIER
I
SLUDGE RECYCLE
-------
-7-
GENERAL DISCUSSION - BLUE PLAINS
As seen in the on-going research projects, there are numerous applications
for oxygen aeration processes, whether the feed is raw wastewater as at Batavia
or primary effluent as in the Blue Plains study in Washington, D.C. Likewise,
oxygen can be used to upgrade existing secondary treatment facilities as will
be demonstrated in the Las Virgenes and Newtown Creek projects. With the
different characteristics in operation and performance, the oxygen activated
sludge system should be viewed and compared as an entire system composed of three
interrelated subsystems; a biological reactor, a clarifier, and a solids handling
system.
REACTOR
The first and most unique aspect of the system is the gas tight
biological reactor shown in Figure 5. In the EPA/DC Pilot Plant, primary
effluent from the District of Columbia's plant is fed to the oxygen reactor
either on steady state flow or on a predetermined daily cycle (diurnal variation),
normally with a 2.3:1 (45-105 gpm) daily flow variation.
Using all four available stages, the 8,100-gallon Blue Plains oxygen
reactor provides 1.95 hours or detention time at the nominal influent flow of
100,000 gpd. At the peak daily flow, the detention time is 1.29 hours. Using
three of the available four stages, the detention times are reduced to 1.50
hours and 1.00 hours, respectively, at the nominal and peak daily flows.
The reactor is sealed to prevent loss of oxygen and includes submerged
hydraulic entrances and exits as well as simple water-sealed mixing equipment.
Internal spray equipment using tap water is provided to suppress foam. Also,
a partially submerged baffle plate before the internal exit trough retains
-------
FIGURE 5
SCHEMATIC DIAGRAM OF BLUE PLAINS OXYGENATION SYSTEM
02 RECYCLE
Oi2
n n n n
INFLUENT
n
i i
e
^M
^m
^m
^m
r
•I
0
n
•»
M»
•»
^M
~|
III
X
Q*
^
r
]
w
J
^^^^
^
^
n
i
»••••!
^^
^•M
•^H
^^
r
•I
0
^M
^*
^v
^M
-]
III
0
r
-------
-8-
the foam until the baffle plate is raised to allow the foam build-up to escape.
The reactor is staged to provide the proper tank geometry for efficient mixing
and oxygen usage.
Efficient oxygen usage is achieved by co-current contacting of the mixed
liquor and oxygen gas through the various stages. The addition of pure oxygen
to the reactor is controlled by a pressure regulator„ An inlet oxygen control
valve actuated by a pressure regulator maintains the overhead gas at a selected
pressure usually between 1" and 4" of water. Even with large instantaneous
fluctuations in oxygen consumption, the oxygen control valve maintains the
selected pressure. The overhead gas pressure is normally selected to maintain
the oxygen concentration at approximately 50% in the exhaust gas from the last
reactor stage. Pure oxygen is introduced to the first stage where the peak
oxygen .demand occurs. As the oxygen is used in biological metabolism, respirated
carbon dioxide and stripped inert gases reduce the oxygen concentration in the
overhead gas flowing co-currently with the mixed liquor through the succeeding
stages. The successive decrease of both oxygen availability and oxygen demand
produces efficient oxygen use before the residual gas is exhausted from the
reactor.
Mixed liquor dissolved oxygen levels in the Blue Plains oxygen reactor . are
held between 4.0 and 8.0 mg/1 by adjusting the recirculation rate of the oxygen
gas within the individual stages. The compressor in each stage pumps the
overhead gas through the rotating submerged turbine-sparger to provide efficient
dispersion and mixing of the recirculated gas. The recirculation rate in each
stage may be set either manually on the basis of the dissolved oxygen analysis or
automatically using a control system with a dissolved oxygen sensor. The gas
-------
-9-
recirculation rate in the first Blue Plains stage typically is 3-7 cfm and 1-2
cfm in each of the last three stages. Total recirculation requirements vary
3
between 0.10 and 0.20 ft /gal. of flow.
With its high oxygen transfer capabilities (which are essentially independent
of turbine mixing rates), the oxygen system is able to operate at higher MLSS
concentrations. These factors enable the system to readily adsorb shock organic
loads. Also, toxic shock loads can be better handled, much as in a totally mixed
activated sludge system. Both types of systems initially expose the toxic substrate
to a large mass of active solids and the resulting "biological inertia" buffers
the toxicity.
On the District of Columbia wastewater, as seen in Figure 6, the volatile
portion of the oxygen solids exhibit a much higher activity for the SRT range
above 6 days than the District's step air aeration pilot process. The F/M
ratio is the ratio of BOD applied to the mixed liquor volatile suspended solids
(MLVSS) under aeration. Figure 6 indicates that a lower total volatile mass
under aeration is required with oxygen than with air to obtain any given SRT
above 6 days for a similar influent BOD. Thus, shorter detention times are
possible with oxygen than with step aeration for similar MLSS concentrations to
achieve any given SRT above 6 days. Further, at identical SRT's above 6 days,
the oxygen system will produce less excess biological solids. The most probable
reason for the increased activity is attributed to maintaining the mixed liquor
dissolved oxygen between 4 and 8 mg/1. The independently controlled mixing also
minimized sludge pockets, dead spots, and shearing of the floe particles. Mixed
liquor entering the clarifier has a high dissolved oxygen content which permits a
certain amount of aerobic metabolism in the clarifier and greatly reduces the time
that the bio-mass is in an anoxic condition.
-------
1.0
~ 0.8
•s 0.6
0.4
0.2
0 2
STEP AERATION
OXYGEN AREATION
I I I I I I | I
6 8 10
SRT (days)
12 14 16
FIGURE 6
BIOLOGICAL ACTIVITY RELATIONSHIPS • BLUE PLAINS
-------
-10-
The total production of solids in the oxygen system (Figure 7) per pound
of BOD added, including underflow waste and effluent solids, is inversely related
to the solids retention time (SRT) above an SRT of 1.3 days. The solids pro-
duction with the oxygen aeration system was significantly lower than in the
conventional step aeration pilot system above an SRT of 6 days (or than in a high
rate modified air aeration pilot process tested by the District). Indeed, the
total solids production decreased from 0.65 pounds of excess solids per pound
of BOD added at an SRT of 6 days to 0.35 pounds of excess solids per pound of BOD
added at an SRT of 13 days with only a 33% increase in volatile solids concentra-
tion at the higher SRT. The parallel conventional system, operated as step
aeration or contact stabilization, exhibited increased solids production through
an SRT of 9.5 days with a peak solids production of approximately 1 pound of
excess solids per pound of BOD added. However, an approximate four-fold
increase in volatile solids was required to raise the SRT from 6 to 13 days in
this system and to begin to achieve reduced solids production. The modified
aeration system, in log growth rate biology, produced solids at the rate of
1.0 to 2.0 pounds of excess solids per pound of BOD added at its operating
SRT of less than 1 day (115 mg/1 of alum added for P removal). .
The reduction of BOD in the reactor was excellent. With an influent BOD
up to 130 mg/1, a wide range of detention times from 1.5 to 2.5 hours, and SRT's
that varied from 13 to as low as 2 days, the effluent soluble BOD was consistently
less than 5 mg/1 as described in Table 1. This indicates virtually complete
insolubilization of the BOD in the primary effluent. Thus, BOD removal on the
D.C. oxygen system is a function of clarification.
The oxygen mixed liquor was similar visually to the micro-organisms in
conventional activated sludge (Figure 8). The mixed liquor biota was normally
-------
1.0
0.8
s 0.6
0.4
OXYGEN AERATION
0.2
0
I
I
0246 8 10
SRT (days)
FIGURE 7
EXCESS BIOLOGICAL SLUDGE PRODUCTION
12
14 16
BLUE PLAINS
-------
TABLE
ORGANIC REMOVAL - BLUE PLAINS
Operating Period
Primary Effluent
(mg/1)
Final Effluent
Final Effluent
BOD (mg/1)
Month
Dates
1
June
12-30
2
July
3
August
1-25
4
September
5
October
3-11
6
November
10-30
7
January
1-16
8
January
17-31
BOD
BOD
(mg/1)
89
18
87
19
89
12
106
13
116
14
131
27
124
11
134
32
Soluble
Primary Effluent
Final Effluent
COD (mg/1)
COD
Primary Effluent
Final Effluent
(mg/1)
TOC (mg/1)
TOC
Primary Effluent
Solids (mg/1)
Final Effluent
Solids (mg/1)
(mg/1)
-
250
45
75
14
-
244
70
65
24
2
245
49
77
15
2
252
51
100
15
3
284
51
106
15
3
275
63
91
21
3
250
59
83
21
3
256
99
87
26
Suspended
113
101
102
107
120
92
98
100
Suspended
36
53
28
24
35
56
24
58
-------
TABLE 1 (CONTINUED)
ORGANIC REMOVAL - BLUE PLAINS
9
March
1-18
121
27
_
251
76
88
22
104
49
10
April
107
10
4
267
48
81
17
83
18
11
May
140
7
4
278
51
92
18
120
12
12
June
110
8
_
238
45
74
18
100
13
13
July
129
14
_
235
35
78
14
103
11
14
August
110
15
_
219
32
69
13
97
16
15
September
149
15
4
239
35
79
15
95
15
16
October
120
14
4
224
37
69
14
81
15
17
November
125
20
5
244
54
75
17
90
23
18
December
1-21
125
20
5
238
59
76
19
95
23
19
December
22-31
135
18
5
236
53
91
19
95
18
-------
FIGURE 8
-------
-li-
very well bioflocculated with active stalked cilates growing on the bacterial
mass. Zooflagellates and free swimming cilates, although few in number, remained
adjacent to or within the flocculated particles. Several varieties of large
active rotifers were present in abundance. A few nematodes existed in the sludge.
Normally, filamentous growth was not apparent. There was almost complete absence
of fragmented debris or unflocculated bacteria between the discrete particles.
In the SRT range less than 5 days, both the oxygen activated sludge and the
conventional aeration systems exhibited filamentous growth on the District of
Columbia wastewater. Filamentous growth did not occur during operation above
an SRT of 5 days. Normally, when encountering filamentous growth for a few days,
reducing system influent flow to increase the SRT reestablished a filamentous
free sludge in several days. However, after extended periods of operation with
filamentous growth, the Sphaerotilis became firmly entrenched and could not be
quickly purged from the system by flow reduction techniques. Hydrogen peroxide
added to the recycle in two 24-hour periods approximately a week apart at
dosages of 200 rag/1 (based on influent flow) was then required to purge the system
of filamentous growth.
The recirculation of respirated carbon dioxide within the oxygen reactor
stages lowers the wastewater pH from 7.0 - 6.8 in the first stage and to 6.4
- 6.1 in the final stage. With an average system pH of approximately 6.5, the
oxygen process more slowly establishes a nitrifying population than the step
aeration activated sludge system operated at a pH of 7.0 to 7.4. However, during
the warmer months when the solids wasting is reduced to a level where the
nitrifying organisms propagate faster than they are removed, the Nitrosomonas
and Nitrobacter populations increase and substantial nitrification occurrs in
the oxygen system. Nitrogen removal across the oxygen system during periods
-------
-12-
of high nitrification and partial denitrification is as high as 39-40%. Nitrogen
removal decreases to a low of 9-10% during periods without nitrification.
As in the parallel step aeration process, nitrification in the oxygen
system begins to decrease in the Fall and becomes virtually nil during the
Winter. At wastewater temperatures of about 63°F, 5 mg/1 of NO^-N is still
produced with an SRT of 9.0 days.
Without alum addition, phosphorus is removed from the oxygen system
through metabolic uptake and by wasting of excess solids; thus, the removals
vary with the metabolism of the mixed liquor. At high SRT's (highly endogenous
metabolism), total phosphorus removal averages only about 15%. At lower SRT's
with less endogenous respiration, phosphorus removals increase to 20%.
With alum addition, phosphorus removal in the oxygen system increases as
the alum weight ratio (Al /P) increases. During experiments conducted in the
Fall of 1970, for a dosage equal to an Al /P ratio of 1.4/1, 80% of the
phosphorus was removed to an average residual of 1.8 mg/1 as P and only a slight
decrease in wastewater alkalinity and pH occurred. The filtered effluent
(though 0.45/<) contained an average of 1.6 mg/1 of soluble P. When the dosage
was increased to a ratio of 1.85/1 (Al /P), the residual total and soluble
phosphorus decreased to 0.62 and 0.53 mg/1 as P, respectively. At this higher
dosage, however, the buffering capacity of the oxygen mixed liquor was further
reduced and the average pH decreased from 6.5 to 6.0. The oxygen biomasa dis-
persed, necessitating termination of the alum addition to allow the mixed liquor
to recover. In areas with low alkalinity wastewaters, additional alkalinity in
the form of lime or caustic may be required to control pH at a level which will
prevent floe dispersion. This pH adjustment may be necessary in either air or
oxygen systems but is more likely in an oxygen system because of the increased
dissolved C0£ content of the mixed liquor. The addition of alum and precipitation
of A1(PO^) and Al(OH)rj increases the inert solids carried in the system and
adequate clarification for the higher solids concentration must be provided.
-------
-13-
Consistently throughout the operation, vented gas from the fourth oxygen
reactor stage has been less than 10% of the input oxygen volume. The vented
stream is roughly 50% oxygen. Based upon the influent and exhausted oxygen
concentrations, the net utilization of oxygen in the process is about 95%.
The accountable oxygen consumption consisting of COD removal, nitrification
demand, exhaust gas, and effluent dissolved oxygen is summarized in Table 2.
The COD removed was calculated by substracting the COD in the underflow waste
solids and that in the process effluent from the primary effluent COD. With
increasing SRT, additional oxygen is required for endogneous respiration.
Likewise during periods with nitrification, additional oxygen is required.
Pertinent reactor variables and operating conditions are summarized in
Table 3 for approximately 1-1/2 years of operation on the Blue Plains oxygen
aeration pilot plant.
CLARIFICATION
The second important aspect in the oxygen aeration system is liquid/solids
separation. At the EPA/DC Pilot Plant, as at Batavia, gravity clarification is
employed. As mentioned before, soluble residual BOD in the effluent averaged
less than 5 rag/1 in the test periods indicating virtually complete BOD in-
solubilization. Thus, most of the residual BOD in the Blue Plains oxygen
system effluent is associated with suspended solids. Overall removal of
suspended solids and BOD is a function of clarification efficiency.
Clarifier efficiency is in turn a function of the basic settling
characteristics of the solids as well as of the actual design and operation of
the clarifier. With the normally higher mixed liquor concentrations used in
the oxygen aeration process, design criteria for both clarifiers (i.e., overflow
2
rates and volume) and thickeners (i.e. solids loading - Ib/ft /day) should be
considered. The Ten State Standards suggest that conventional activated sludge
-------
TABLE 2 - OXYGEN USAGE - BLUE PLAINS
Operating Period
Month
Dates
Primary Effluent COD
(Ib/million gal.)
Final Effluent COD
(Ib/million gal.)
Waste Sludge COD
(Ib/million gal.)
COD Removed from System
(Ib/million gal.)
Nitrate Nitrogen Demand
(Ib/million gal.)
Exhaust Oxygen
(Ib/million gal.)
Final Effluent D.O.
(Ib/million gal.)
Total
Oxygen Supplied
(Ib/million gal.)
1 2
June July
12-30
2080 2030
375 584
188+ 42+
1517 1404
14 69
85 54
10* 10*
1626 1537
1750 1825
3
August
1-25
2040
408
150+
1482
128
75
10
169 5
1775
4
September
2090
430
217+
1443
160
85*
10
1698
1900
5
October
3-21
2370
425
630+
1315
100
87
10
1512
1650
6
November
10-30
2290
525
247
1518
18
^
85*
10*
1631
+
7
January
1-16
2080
488
233
1359
9
65
25
1458
1700**
8
January
17-31
2140
826
336
978
14
65
25
1082
1800**
+ COD = 1.4 volatile solids
* Estimate
± Inlet meter malfunctioned
** Increase sampling and greater losses of 02 through sample ports
-------
TABLE 2 - OXYGEN USAGE - BLUE PLAINS (CONT'D)
9
March
1-18
2050
630
160
1260
0
160
30
1450
1450
10
April
2180
380
700
1100
0
130
40
1270
1300
11
May
2500
450
890
1160
0
40
50
1250
1260
12
June
1750
340
80
1330
0
80
60
1470
1600
13
July
1940
270
270
1400
60
300
60
1820
2200
14
August
1830
260
270
1300
70
260
40
1670
1690
15
September
1920
280
400
1240
270
200*
40
1750
+
16
October
1880
310
390
1180
200
200*
30
1610
+
17
November
1950
440
220
1290
220
200*
40
1750
1740
18
December
1-21
1990
490
230
1270
190
200*
40
1700
2000
19
December
22-31
1970
440
300
1230
70
150*
40
1490
1450
-------
TABLE 3
REACTOR VARIABLES AND OPERATING CONDITIONS
BLUE PLAINS
Operating Period
Month
Dates
Flow Rate (gpm)
Aeration Time (hr)
Recycle Rate
MLSS (mg/1)
MLVSS (%)
SRT (days)
F/M (Ib BOD/day/lb MLVSS)
Volumetric Loading (Ib BOD/
day/1,000 ft3)
/k.w.-hr \ **
Mixer Power ^1,000 gal.J
1
June
12-30
50-55
2.00
50%
4140
74%
7.7
0.333
57
1.27
2
July
80
1.66
50%
5180
70%
7.3
0.342
80
1.19
3
August
1-25
80
1.66
42%
5250
73%
11.8
0.296
80
0.98
4
September
70+
1.95
32%
6000
78%
10.7
0.304
96
0.92
5
October
3-21
70+*
1.95
38%
8120
67%
5.5
0.283
106
1.00
6
November
10-30
70+
1.95
37%
6350
73%
5.5
0.355
108
1.00
7
January
1-16
53
2.50
77%
5300
80%
13.0
0.275
89
1.18
8
January
17-31
53
2.50
80%
3940
81%
4.7
0.392
80
1.42
/k.w.-hr \**
Compressor Power ^1,000 gal.j
Temperature (°F)
0.39
74-80
0.28
78-84
0.39
82-85
0.41
79-83
0.35
70-79
0.26
66-69
0.34
58-60
0.32
58-60
* Alum addition
+ 2.3:1 diurnal variation
** Pilot plant equipment efficiency was not determined
-------
90
TABLE 3 (CONTINUED)
REACTOR VARIABLES AND OPERATING CONDITIONS - BLUE PLAINS
9
March
1-18
60-70
2.15
50-60%
3070
77%
3.7
0.580
10
April
31-67
3.30-1.55
90- 60%
2710
8 in
1.3-4.0
0.30-1.00
11
May
60
1.70
65%
2750
78%
2.0
0.970
12
June
30-70
3.70-1.50
100-50%
4000
73%
13.0
0.400
13
Jiily
70+
1.5
50%
6600
70%
12.6
0.430
14
August
70+
1.5
50%
7500
70%
10.0
0.32
15
September
70+
1.5
46%
7400
72%
7.5
0.39
16
October
70+
1.5
36%
6000
73%
9.5
0.31
17
November
70+
2.0
30%
4600
78%
9.8
0.39
18
December
1-21
70+
2.0
25%
4400
SO'',
9.0
0.40
19
December
22-31
70+
2.0
25%
4200
81%
6.5
0. 50
98
157
95
160
131
185
146
111
111
122
60-62
62-65
65-71
70-77
77-80
77-81
76-81
76-71
71-65
65-C3
63-61
-------
-14-
2
clarifiers be designed for average overflow rates of 800 gpd/ft . The Water
Pollution Control Federation Manual of Practice (1959) suggests that the
2
solids loading be held below a peak of 30 Ib/ft /day. Overflow rates and
solids loading criteria should be better defined for high solids systems such
as oxygen aeration. An internal EPA study is presently underway on several
oxygen aeration pilot facilities (with the cooperation of industry) to further
investigate sludge settling parameters.
At this point, what we have learned to date will be discussed. The basic
settling characteristicsof the mixed liquor typically have been found to be a
function of:
1. Concentration of the mixed liquor
2. Particle shape
3. Particle density
4. Seasonal variation
a. Physical changes in water density and viscosity with temperature
b. Metabolic changes with temperature
c. Seasonal loading variation
As seen in Figure 9, the log of the initial settling rate is a function of
the log of the solids concentration. This sample curve illustrates that two
relationships exist. The first at lower MLSS levels corresponds to free particle
settling and is characterized by the absence of an initial discrete subsiding
interface and a zone of homogenous settling solids. The second at higher MLSS
levels has both an initial discrete interface and a zone of homogenous settling
particles (zone settling). Thus, the sizing of a clarifier is a function of the
MLSS concentration and must be coordinated with the reactor (and the sludge
handling facilities) to achieve the desired biological capabilities of the system.
-------
30
i 2°
sT 15
i
i 10
uj 0
£ 6
t 4
00
S 3
" 2
(D.C.-DEC.)
\
I I
Vi=ACi
•n
Yj ^Initial Velocity
C| ^Initial Cone.
A=lntercept Constant
n=Slope Constant
I II
I I
I
1 2 3 4 6 8 10 15 20 30
INITIAL MIXED LIQUOR CONCENTRATION-C j (gm/l)
FIGURE 9
INITIAL SLUDGE SETTLING VELOCITY PROFILE
FOR BROAD MIXED LIQUOR CONCENTRATION RANGE - BLUE PLAINS
-------
-15-
Another important factor is the particle shape. Normally, as shown in
Figure 8, the oxygen mixed liquor particles have rounded shapes. However,
if filamentous growth exists, as experienced below an SRT of 5 days in D.C.
(air and oxygen), both settling rates and compaction deteriorate. The range, if
any, that filamentous growth appears is unique to each location and should be
defined for that location. The presence of industrial fibers would have an effect
similar to filamentous growth on mixed liquor settling characteristics.
Still another important factor in the basic settling characteristics is
the density of the particles in relationship to that of the water. It is the
difference in density that is the driving force for settling. The VSS/TSS ratio
(or volatile %) is one relative indication of density. There are several ways
to improve the particle density. One is to feed raw wastewater instead of primary
effluent to the oxygen aeration system, thus incorporating the normally denser
particles captured in primary sedimentation into the biomass, such as occurred
at Batavia. Again, the sizing of the reactor oxygen supply, etc. must be
compatible with the increased organic loading. In Washington, heavy rains and
unusually high flows wash silt and clay into the sewer system. These materials
subsequently become incorporated in the mixed liquor solids and have increased
sludge settling rates 30% to 60%. In like fashions, operation under different
biological conditions can alter sludge settling characteristics.
Another unique method of increasing the density of the sludge was
evaluated at the EPA/DC Pilot Plant by altering the method of clarifier operation.
Two major methods of clarifier operation are possible. One is to use the
blanket as a filter and the other is to permit classification of the settling
solids. The first method can be accomplished in two differentials: (1) By
providing sufficient depth to the clarifier such that the mixed liquor passes
up through the clarifier blanket (which acts as a filter). (2) By carrying high MLSS
-------
-16-
concentrations (usually above 4500 mg/1 in D.C.) such that the particles settle
in a subsidence (zone) settling pattern with discrete interfaces existing between
the homogeneous subsiding particles and the decant. In the subsidence zone, the
relatively uniform concentration of particles are nearly homogeneously mixed by
the countercurrent turbulence produced by water passing around the solids. The
homogeneous subsiding blanket does not allow classification of individual particles
because the settling blanket acts as a filter.
At Blue Plains, with MLSS concentrations below 4500 mg/1, subsidence (zone)
settling does not occur during the initial portion of settling. This provides for
a second method of clarifier operation where classification of the discrete settling
particles can occur if the mixed liquor is fed above the clarifier blanket level.
The lighter or unsettlable particles, thus, can be purged from the system. The
effluent suspended solids accordingly increased from 15 mg/1 to 25 mg/1 during
this method of operation with a corresponding increase in effluent BOD.
Seasonal variations also affect sludge settling characteristics in oxygen
as well as in air systems. These variations become critical as the MLSS of the
mixed liquor increases. The pure physical changes in the wastewater density and
viscosity contribute to slower settling rates as the wastewater temperature de-
creases. As the density of the water increases, the driving force for settling
(which is the difference in density between water and the settling particles) de-
creases for a similar particle density. The drag force, viscosity, also increases
with decreasing temperature ("25% from 80°F to 55°F) again contributing to slower
settling rates in colder waters. Figure 10 shows a series of liter batch settling
tests conducted in June, 1971 by only altering the temperature of the mixed liquor.
As expected, the colder samples settled slower. In Figure 11, the batch flux
(concentration multiplied by settling velocity) or the solids loading in
-------
30
20
15
S 10
£ 8
S 6
UJ
C9 4
I 3
0.5
I
85°F (Adjusted) )
50°F (Adjusted) }D-C'June1971
I I
I I
1 2 3 4 6 8 10 15 20
INITIAL MIXED LIQUOR CONCENTRATION (gm/l)
FIGURE 10
EFFECT OF ADJUSTED VS.
ACCLIMATED WASTEWATER TEMPERATURES
ON SLUDGE SETTLING RATES - BLUE PLAINS
30
-------
100
80
£2 60
50°F (Adjusted)
D.C. June 1971
85°F (Adjusted)
D.C. June 1971
40
s= 20
0
0
5000 10,000
INITIAL MIXED LIQUOR CENTRATION (mg/l)
FIGURE 11
EFFECT OF WASTEWATER TEMPERATURE
ON INITIAL BATCH FLUX - BLUE PLAINS
15,000
-------
-17-
2
Ib/ft /day is shown for the previous tests. Solids loading is often used in
thickener design. Again, the effect of wastewater temperature is evident.
Besides the physical changes caused by seasonal variation, another factor
which must be considered is the metabolic change brought about by changing
wastewater temperature. Figure 12 shows that the settling characteristics of
oxygen mixed liquor change seasonably at D.C. At similar SRT's, the initial
settling rate in a 1 liter graduated cylinder test decreased from approximately
10 ft/hr to 7 ft/hr at a concentration of 6000 mg/1 as the temperature changed
from 81°F to 71°F. The clarifier was being operated to capture unsettleable
particles at this time. In Figure 13, the clarifier was operated to purge the
unsettleable particles; but, again the solids showed a decreasing initial settling
rate with decreasing wastewater temperature for a similar biology. The initial
settling rate in the 1 liter test decreased from 14 ft/hr to 9 ft/hr at 4500 rag/1
as the temperature decreased from 70°F to 63°F. Similar patterns of decreasing
settling rates with decreasing wastewater temperature have been observed in
nitrifying and denitrifying mixed liquors also.
Clarifier operation and design are equally important to the basic settling
characteristics of the solids in gravity clarification. Besides selecting an
overflow rate compatible with reactor sizing, the depth and method of clarifier
feed are important as discussed earlier for either high solids capture or solids
classification. Other important design considerations are the volume or detention
time of the clarifier and recycle rate. At Blue Plains, the oxygen system under-
flow solids concentration varies between 1.0% and 1.4% with an average clarifier
detention time of 1.9 hours. With 2.8 hours average detention time, the underflow
solids concentration rises to 2.0%-2.4% with similar recycle rates. The sludge
recycle rate is then determined after an F/M ratio is established for the reactor
-------
30
20
J£ 15
^ 10
I 8
B 6
t 4
LU
" 3
jP
^ 2
1
1
D.C.-Sept 1971-(78-81°F)
D.C.-Oct 1971-(71-73°F)
I II III II
I
30
2 34 6 8 10 15 20
INITIAL MIXED LIQUOR CONCENTRATION (gm/l)
FIGURE 12
EFFECT OF WASTEWATER TEMPERATURE ON
OXYGENATED SLUDGE SETTLING RATES - BLUE PLAINS
(SEPL-OCT.,1971)
-------
30
20
^ 15
£ 10
2 Q
£ 6
£ 4
LU
^ 3
i 2
1
D.C. -Nov 1971-(68-70°F)
D.C.-Dec 1971-(63-64°F)
I I
I I I
I I
3
6 8 10 15 20
30
INITIAL MIXED LIQUOR CONCENTRATION (gm/l)
FIGURE 13
EFFECT OF WASTEWATER TEMPERATURE ON
OXYGENATED SLUDGE SETTLING RATES - BLUE PLAINS
(NOV.-DEC.J971)
-------
-18-
and the underflow concentration from the clarifier is likewise established.
Inventory solids are another consideration in clarifier operation. The
total solids inventory is a result of both the build-up of solids in the blanket
level and the solids actually in the transition (or settling) process. A simple
increase in the wasting rate will reduce the blanket level if the increase is
caused by the build-up of solids. However, as the initial settling rates de-
creased, as in Washington, for a given MLSS concentration, the sludge inventory
also increases in the clarifier as more solids are in transition (or settling).
In this case, increased wasting rates will not reduce the solids inventory in the
clarifier without thinning the MLSS. It appears that the rising blanket can be
a result of slower settling solids in the transition zone rather than a backlog
of solids due to inadequate wasting.
In the Summer of 1970 at an MLSS concentration in excess of 8000 mg/1,
2
peak clarifier overflow rates of 1940 gpd/ft were observed on the Blue Plains
oxygen system as shown in Table 4. During the 1970-71 Winter, the peak sustained
overflow rates which could be maintained without the blanket coming over the weirs
o
was 975 gpd/ft at MLSS concentrations that varied from 3900 to 5300 mg/1. The
causative agents which reduced the allowable overflow rates for satisfactory
operation from Summer to Winter were undoubtedly a combination of all the above
mentioned factors, not the least of which was the decreased wastewater temperature.
-------
TABLE 4
CLARIFIER VARIABLES AND OPERATING CONDITIONS
BLUE PLAINS
Operating Period
Month
Dates
Average Overflow Rate (gpd/ft
At Surface
Above Feed Skirt
Below Feed Skirt
Peak Overflow Rate (gpd/ft2)
At Surface
Above Feed Skirt
Below Feed Skirt
Average Solids Loading
(lb/day/ft2)
SVI
Underflow Solids
% Dry Solids
% Volatile
Waste Solids (Ib/million gal.)
Volatile (Ib/million gal.)
Effluent Solids
(Ib/million gal.)
Volatile (Ib/million gal.)
1 2
June July
12-30
— + - +•
750 1210
670 1075
~ + ~ +
750 1210
670 1075
37 75
80 48
1.16 1.34
75 70
161 40
121 28
296 445
198 245
3
August
1-25
- +
1210
1075
t
1210
1075
58
50
1.27
75
144
108
166
113
4
September
-f.
1280-
1050
940
1940
1580
1410
61
42
1.40
80
250
200
204
141
5
October
3-21
4.
1280-
1050
940
1940
1580
1410
88
33
2.14
65
680
441
290
189
6
November
10-30
4-
12801
1050
940
1940
1580
• 1410
68
48
1.40
81
230
202
470
342
7
January
1-16
4
975-
800
710
975
800
710
55
60
1.08
90
193
174
197
118
8
January
17-31
975-
800
710
975
800
710
42
73
1.00
80
253
202
483
400
+ Peripheral feed - no canter feed section
Area at surface 96 ft X 6 ft. deep
Area below feed skirt 107 ft2 X 5 ft. deep
Total depth 11 ft. deep
- Center feed section area at surface 78 ft.2 X 4 ft.deep
area above feed skirt 96 ft'2 X 2 ft. deep
area below feed skirt 107 ft2X 5 ft.deep
-------
TABLE 4 (CONTINUED)
CLARIFIER VARIABLES AND OPERATING CONDITIONS
BLUE PLAINS
9
March
1-18
950 i
780
700
950
780
700
35
81
0.79
77
130
100
410
375
10
April
- @
(290-620)
(290-620)
17
120-190
0.85
80
-
-
11
May
- @
560
560
21
265
0.78
77
720
550
100
12
June
- @
(280-650)
(280-650)
24
173
1.28
73
142
104
88
13
July
- @
975
650
54
50
1.92
70
168
118
92
14
August
- @
975
650
61
30-35
2.22
70
253
178
133
100
15
September
- @
975
650
58
33
2.38
70
460
323
160
104
16
October
- @
975
650
47
42
2.26
71
340
242
130
99
17
November
- @
975
650
36
36
2.05
79
178
140
194
130
18
December
1-21
- @
975
650
33
40
2.41
81
200
160
190
123
19
December
22-31
_ @
975
650
32
56
2.30
82
310
254
140
88
Two center feed clarifiers at 78 ft2 each X 11 feet deep
feed sections at 3 ft2 each X 3.5 ft deep
-------
-19-
As expected with filamentous growth In late Spring 1971, allowable clarifier overflow
rates were markedly decreased,as shown in Table 4. Again, during the current 1971-
2
72 Winter, maximum overflow rates of 975 gpd/ft have been demonstrated at Blue Plains.
SOLIDS HANDLING
The other integral part of the oxygen system is the excess solids handling
equipment. Of utmost importance is the relative ease with which the oxygen activated
sludge process can be operated in endogenous respiration, thereby substantially
reducing the quantity of excess sludge to be handled. This factor will reduce
the number and/or size of the selected sludge handling and disposal facilities.
However, the increased operating costs resulting from the increased oxygen necessary
to oxidize ("burn-up") the excess sludge and the larger reactor/clarlfier capabilities
needed to hold the increased solids inventory required for endogenous respiration
must be balanced economically with the reduction in size of the solids handling
and disposal units.
Another factor to be considered is that the larger the-clarifier volume,
the thicker the underflow solids concentration. It may be economically feasible
to properly size the reactor/clarifler combination to yield underflow solids
sufficiently thick to be dewatered directly without prior thickening or digestion.
This would be accomplished by selecting a small reactor and large clarifier.
.- 1«
An alternative to the above is to select a large reactor and a small clarifier
and provide additional thickener capabilities. In the EPA/DC Pilot Plant, the
excess solids are thickened separately by air flotation or gravity thickening.
These solids have been thickened to over 4.5% without chemical additives by both
gravity and air flotation thickening.
-------
-20-
CONCLUSIONS
1. A gas-tight biological oxygen reactor with independent control of
dissolved oxygen and mixing, coupled with an aerobic clarifier, produces a good
quality secondary effluent on District of Columbia primary effluent with 1.5 to
2.5 hours average detention time .(based .on raw flow) with MLSS concentrations
between 4000 and 8000 mg/1.
2. Biodegradable organics in the Blue Plains primary effluent are essentially
completely insolubilized by the oxygen process (less than 5 mg/1 of soluble BOD).
Total carbonaceous BOD removal depends upon the amount of suspended solids in the
effluent and, therefore, on the ability to clarify.
3. Oxygen micro-organisms are visually the same as those in a typical
conventional system; however, the rate of activity of the oxygen volatile solids
is greater above an SRT of 6 days.
4. Oxygen activated sludge is subject to filamentous (Sphaerotilis) growth
as similarly observed in the air systems when operated below an SRT of 5 days
on DC primary effluent.
5. Sludge in the oxygen system underflow settles to approximately 1.0-
1.4% solids in a clarifier with 1.9 hours of hydraulic retention time and 2.0%
to 2.4% in a clarifier with 2.8 hours of hydraulic retention time.
6.' Total production of excess biological solids is significantly lower
in the Blue Plains oxygen system than in a parallel step aeration system at
SRT's above 6 days with as little as 0.35 Ib of excess solids produced/lb BOD
added at an SRT of 13 days.
7. When the oxygen clarifier is operated with a deep feed well or with
the mixed liquor sufficiently concentrated to settle in a subsidence (zone)
settling pattern, the blanket acted as a filter and produced high quality effluent.
-------
-21-
2
In the Summer and Fall, 1970, the clarifier operated at a peak rate of 1940 gpd/ft ,
In the 1970-71 Winter, oxygen clarifier rates could not exceed a sustained
2
975 gpd/ft . A larger clarification area is required in the Winter than in
the Summer on District of Columbia wastewater for a given MLSS concentration.
8. With a shallow center feed well and with the mixed liquor concentrations
low enough (under 4500 mg/1) to permit discrete particle settling, better settling
rates are observed in the oxygen clarifier than with the method of clarifier
operation described in No. 7 above. Only moderate decreases in effluent quality
(increase in SS from 15 to 25 mg/1) are observed with this type of clarifier
operation at Blue Plains.
9. Nitrification is achieved in the oxygen aeration system in the Summer
and Fall at Blue Plains.
10. Average effluent phosphorus residuals of 1.8 mg/1 as P with an alum
I I i
dosage of 1.4 Al to 1.0 P, by weight, were achieved in the oxygen system.
Higher phosphorus removals are possible with higher alum dosages, but in areas with
moderate wastewater alkalinity, pH control may be required to prevent the depletion
of alkalinity reserves in the oxygen system.
11. Based upon the influent and exhaust gas flows, over 95% of the input
oxygen is consistently utilized in the Blue Plains oxygen reactor.
SUMMARY
Again, the oxygen activated sludge system must be viewed as an entirely
unique approach, and compared on a total system basis with other alternative
systems. Reactor and clarifier sizing must be coordinated. As the reactor size
is increased, a lower MLSS concentration is required for a given biological state
(F/M ratio). The lower the MLSS concentration, the smaller the required secondary
clarification area. The solids handling requirement of an oxygen system depend
-------
-22-
on the biology established in the reactor/clarifier combination. Thus, if
minimum excess biological sludge production is required, then more capacity
is required in the reactor/clarifier combination. Further, the concentration of
the clarifier underflow solids is dependent on clarifier volume.
U.S. GOVERNMENT fltlKTlNG OFFICE: 1972- 759-546/1008
------- |