O Batteiie
Pacific Northwest Laboratories
Batteiie Boulevard / Richland, Washington 99352
POWDERED ACTIVATED CARBON
TREATMENT OF COMBINED
AND MUNICIPAL SEWAGE
¦ •-
k
W 1#
1
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes the
results and progress in the control and abatement of pollu-
tion in our Nation's waters. They provide a central source
of information on the research, development, and demonstra-
tion activities in the water research program of the Environ-
mental Protection Agency, through inhouse research and grants
and contracts with Federal, State, and local agencies,
research institutions, and industrial organizations.
Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Chief, Publications Branch
(Water), Research Information Division, R&M, Environmental
Protection Agency, Washington, D. C. 20460.
-------
POWDERED ACTIVATED CARBON TREATMENT OF
COMBINED AND MUNICIPAL SEWAGE
by
Alan J. Shuckrow
Gaynor W. Dawson
William F. Bonner
PACIFIC NORTHWEST LABORATORIES
a division of
BATTELLE MEMORIAL INSTITUTE
Battelle Boulevard
Richland, Washington 99352
for the
Office of Research and Monitoring
ENVIRONMENTAL PROTECTION AGENCY
Project #11020 DSQ
Contract #14-12-519
November 197 2
-------
EPA Review Notice
This report has been reviewed by the Environmental Protec-
tion Agency and approved for publication. Approval does
not signify that the contents necessarily reflect the views
and policies of the Environmental Protection Agency, nor
does mention of trade names or commercial products consti-
tute endorsement or recommendation fcr use.
ii
-------
ABSTRACT
A unique physical-chemical wastewater treatment system
utilizing powdered activated carbon was developed and demon-
strated by Battelle-Northwest under contract to the EPA.
The research program included laboratory process development
followed by design, construction, and field demonstration of
a 100,000 gpd mobile pilot plant.
In the treatment process, raw wastewater is contacted with
powdered carbon, coagulated with alum, settled with poly-
electrolyte addition and, in some cases, passed through a
tri-media filter. The solids from the clarifier, composed
of raw sewage solids, powdered carbon, and aluminum hydrox-
ide floe, are readily dewaterable to 20-25 percent solids
by direct centrifugation with the powdered carbon acting as
a substantial aid to dewatering. The dewatered solids are
passed through a fluidized bed furnace developed specifi-
cally for powdered carbon regeneration. Alum is recovered
by acidifying the regenerated carbon slurry from the furnace
to a pH of 2. The recovered carbon and alum are recycled
as an acidified slurry and added to the raw sewage with the
makeup carbon.
The program demonstrated the ability of the treatment pro-
cess to consistently produce high-quality effluent from raw
wastewater.
Powdered carbon regeneration was highly successful on the
pilot scale. Full capacity recovery was achieved with less
than two percent carbon loss per regeneration cycle. Alum
recovery was also greater than ninety percent.
Initial cost estimates, including both operation and capital
amortization, are 16.8C/1000 gal. for combined sewage treat-
ment and 22-23C/1000 gal. for raw municipal wastewater.
This report was submitted in fulfillment of Project # 11020
DSQ and Contract # 14-12-519 under the sponsorship of the
U. S. Environmental Protection Agency.
-------
CONTENTS
SECTION PAGE
I CONCLUSIONS 1
II RECOMMENDATIONS 7
III INTRODUCTION 9
IV PILOT PLANT DESCRIPTION 15
V DEMONSTRATION SITE 25
VI TREATMENT SYSTEM PERFORMANCE 33
VII REGENERATION 99
VIII DESIGN AND ECONOMIC CONSIDERATIONS 133
IX ACKNOWLEDGMENTS 141
X REFERENCES 14 3
XI APPENDIX A 14 5
v
-------
NO
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
FIGURES
PAGE
PROCESS FLOW SHEET 11
MOBILE PILOT PLANT 12
SCHEMATIC FLOWSHEET OF MOBILE PILOT PLANT 16
CHEMICAL INJECTION POINT IN PIPE REACTOR 17
FLUIDIZED BED REGENERATION UNIT 19
REGENERATION SYSTEM SCHEMATIC FLOWSHEET 21
PILOT PLANT IN OPERATION AT ALBANY SITE 22
FLUIDIZED BED CONTROL PANEL 23
ISLAND CREEK DRAINAGE AREA AND
DEMONSTRATION SITE LOCATION 26
TYPICAL DIURNAL DRY WEATHER FLOW
VARIATION AT ALBANY SITE 27
AVERAGE DIURNAL BOD AND COD
FLUCTUATION AT ALBANY SITE 29
AVERAGE DIURNAL SUSPENDED SOLIDS AND
TURBIDITY FLUCTUATION AT ALBANY SITE 30
AVERAGE DIURNAL PHOSPHATE FLUCTUATIONS
AT ALBANY SITE 31
STORM EFFECT ON TURBIDITY
AND SUSPENDED SOLIDS - 7/13/71 38
STORM EFFECT ON COD - 7/13/71 39
STORM EFFECT ON TURBIDITY
AND SUSPENDED SOLIDS - 7/19/71 40
STORM EFFECT ON COD - 7/19/71 41
STORM EFFECT ON TURBIDITY
AND SUSPENDED SOLIDS - 7/29/71 42
STORM EFFECT ON SETTLEABLE SOLIDS - 7/29/71 43
STORM EFFECT ON BOD AND COD - 7/29/71 44
vi
-------
(FIGURES Continued)
21 STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 9/13-14/71 45
22 STORM EFFECT ON BOD AND COD - 9/13-14/71 46
23 STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 9/16-17/71 4 7
24 STORM EFFECT ON BOD AND COD - 9/16-17/71 4 8
2 5 STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 5/2/72 49
26 STORM EFFECT ON BOD AND COD - 5/2/72 50
2 7 STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 5/16/72 51
2 8 STORM EFFECT ON COD AND BOD - 5/16/72 52
29 STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 5/16/72 53
30 STORM EFFECT ON COD AND BOD - 5/16/72 54
31 STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 5/16-17/72 55
32 STORM EFFECT ON COD AND BOD - 5/16-17/72 56
33 JAR TEST DETERMINATION OF CARBON REQUIRE-
MENTS DURING INCREASING AND PEAK LOADING
OF 7/29/71 STORM 5 8
34 JAR TEST DETERMINATION OF CARBON REQUIRE-
DURING DECREASING AND MINIMUM LOADING OF
7/29/71 STORM 59
35 PLANT SUSPENDED SOLIDS AND
TURBIDITY REMOVAL - 7/12-13/71 64
36 PLANT COD REMOVAL - 7/12-13/71 6 5
3 7 PLANT SUSPENDED SOLIDS AND
TURBIDITY REMOVAL - 7/14-15/71 66
3 8 PLANT COD REMOVAL - 7/14-15/71 6 7
39 PLANT COD AND BOD REMOVAL - 8/17-18/71 6 8
40 PLANT COD AND BOD REMOVAL - 9/28-29/71 69
vii
-------
(FIGURES Continued)
41 PLANT SUSPENDED SOLIDS REMOVAL - 10/4-5/71 70
42 PLANT COD AND BOD REMOVAL - 10/4-5/71 71
43 PLANT SUSPENDED SOLIDS REMOVAL - 10/7-8/71 72
44 PLANT COD AND BOD REMOVAL - 10/7-8/71 73
45 PLANT SUSPENDED SOLIDS REMOVAL -
10/13-14/71 74
46 PLANT COD AND BOD REMOVAL - 10/13-14/71 75
4 7 PLANT COD AND BOD REMOVAL - 10/25-26/71 76
4 8 PLANT SUSPENDED SOLIDS AND TURBIDITY
REMOVAL - 4/14 & 4/17-19/72 78
49 PLANT COD AND BOD REMOVAL - 4/14
& 4/17-19/72 79
50 PLANT SUSPENDED SOLIDS AND
TURBIDITY REMOVAL - 4/24-26/72 80
51 PLANT COD AND BOD REMOVAL - 4/24-26/72 81
52 PLANT SUSPENDED SOLIDS AND TURBIDITY
REMOVAL - 5/1-3/72 & 5/9-10/72 82
53 PLANT COD AND BOD REMOVAL -
5/1-3/72 & 5/9-10/72 83
54 PLANT TURBIDITY AND SUSPENDED SOLIDS
REMOVAL - 5/16-17/72 & 5/23/72 84
55 PLANT COD AND BOD REMOVAL -
5/16-17/72 & 5/23/72 85
56 PLANT TURBIDITY AND SUSPENDED SOLIDS
REMOVAL - 6/7-8/72 & 6/12/72 86
57 PLANT COD AND BOD REMOVAL -
6/7-8/72 & 6/12/72 87
5 8 PLANT TURBIDITY AND SUSPENDED SOLIDS
REMOVAL - 6/13-15/72 88
59 PLANT COD AND BOD REMOVAL - 6/13-15/72 89
60 FREQUENCY DISTRIBUTION OF PILOT
PLANT EFFLUENT QUALITY 90
viii
-------
(FIGURES Continued)
61 CUMULATIVE PROBABILITY DISTRIBUTION OF
ACHIEVING REMOVAL PERCENTAGES GREATER
THAN OR EQUAL TO A GIVEN VALUE 91
6 2 CHLORINE DEMAND OF PLANT EFFLUENT 95
6 3 LIME USAGE DURING REGENERATION STUDIES 96
64 IRON INVENTORY IN REGENERATED
CARBON SLURRY 10 7
6 5 INERTS BUILDUP IN REGENERATED CARBON 109
66 JAR TESTS ON ONCE REGENERATED
CARBON - BATCH A 112
6 7 JAR TESTS ON TWICE REGENERATED
CARBON - BATCH A 113
6 8 JAR TESTS ON THREE TIMES
REGENERATED CARBON - BATCH A 114
69 JAR TESTS ON FOUR TIMES
REGENERATED CARBON - BATCH A 115
70 JAR TESTS ON ONCE AND TWICE
REGENERATED CARBON - BATCH C 116
71 JAR TESTS ON THREE TIMES
REGENERATED CARBON - BATCH C 117
72 JAR TESTS ON FOUR TIMES
REGENERATED CARBON - BATCH C 118
73 JAR TESTS ON FIVE AND SIX TIMES
REGENERATED CARBON - BATCH C 119
74 JAR TESTS ON SEVEN TIMES
REGENERATED CARBON - BATCH C 120
75 ALUMINUM BUILDUP IN REGENERATED CARBON 12 3
76 ACID USAGE DURING REGENERATION STUDIES 124
7 7 EFFECT OF BLOWDOWN RATE ON TREATMENT
COSTS FOR A 10 MGD PLANT 132
IX
-------
NO.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
TABLES
PAGE
DIURNAL WASTEWATER CHARACTERISTICS AT
ALBANY SITE DURING DRY WEATHER 28
AVERAGE WASTEWATER CHARACTERISTICS AT
ALBANY SITE DURING 1971 PILOT OPERATIONS 32
OPERATIONAL DATA DURING STORM FLOWS 37
PLANT OPERATIONAL DATA FOR 1971
STUDIES USING VIRGIN CARBON 61
PLANT OPERATIONAL DATA FOR DIURNAL STUDIES 62
PLANT PERFORMANCE DATA DURING 1971
OPERATIONS USING VIRGIN CARBON 63
PILOT PLANT PERFORMANCE AT LOW CARBON DOSES 93
REGENERATED CARBON PERFORMANCE COMPARISON 98
FLUIDIZED BED-FURNACE OPERATING CONDITIONS 102
SIEVE ANALYSES OF FLUIDIZED BED SAND 105
ALUMINUM MASS BALANCE SUMMARY 122
STACK SAMPLING DATA 126
CARBON MASS BALANCE SUMMARY 128
SIEVE ANALYSIS OF FLUIDIZED BED SAND
COLLECTED WITH REGENERATED CARBON 13 0
SYSTEM DESIGN PARAMETERS 135
CAPITAL COST ESTIMATES FOR 10 MGD
MUNICIPAL WASTEWATER AND STORM WATER
TREATMENT PLANTS 136
SYSTEM OPERATING PARAMETERS 137
OPERATING COST ESTIMATES FOR 10 MGD
MUNICIPAL WASTEWATER AND STORM WATER
TREATMENT PLANTS 138
TOTAL COSTS FOR 10 MGD MUNICIPAL WASTE-
WATER AND STORM WATER TREATMENT PLANTS 139
x
-------
SECTION XI
APPENDIX A
PAGE
LABORATORY STUDIES 145
GENERAL 145
CARBON STUDIES 145
TABLE A-l: COMPOSITION OF RICHLAND SEWAGE 146-147
FIGURE A-l: BENCH SCALE SYSTEM 148
FIGURE A-2: EFFECT OF CARBON CONCENTRATION
ON EQUILIBRIUM COD 150
FIGURE A-3: EFFECT OF CARBON CONCENTRATION
ON EQUILIBRIUM TOC 151
FIGURE A-4: EFFECT OF CONTACT TIME
ON TOC REMOVAL 152
TABLE A-2: COMPARISON OF VARIOUS
CARBONS - CONTACT TESTS 153
BENTONITE PROCESS DEVELOPMENT 154
TABLE A-3: COMPARISON OF VARIOUS
CARBONS - JAR TESTS 155
FIGURE A-5: EFFECT OF BENTONITE CONCEN-
TRATION ON EFFLUENT QUALITY 156
TABLE A-4: BENCH SCALE SYSTEM OPERATIONAL
DATA - TUBE SETTLER 157
TABLE A-5: UPFLOW CLARIFIER OPERATING DATA 159
ALUM PROCESS PRELIMINARY INVESTIGATIONS 160
CARBON REGENERATION 160
FIGURE A-6: EFFECT OF ALUM DOES ON COD
AND TURBIDITY REMOVAL 161
TABLE A-6: UPFLOW CLARIFIER SYSTEM -
ALUM PROCESS PERFORMANCE 162
xi
-------
SECTION I
CONCLUSIONS
A physical-chemical process utilizing powdered activated
carbon for the treatment of sanitary and combined sewage has
been successfully demonstrated on the 100,000 gal/day scale
at Albany, New York. Carbon regeneration in a fluidized
bed furnace and alum recovery from the calcined sludge have
also been demonstrated as has been reuse of the reclaimed
chemicals. This demonstration project has established the
technical and economic feasibility of the process for both
sanitary and combined sewage treatment. On the basis of
laboratory studies and the pilot plant demonstration, the
major conclusions listed below were drawn.
LIQUID TREATMENT PROCESS
• A process using powdered activated carbon, alum and
a high molecular weight anionic polymer is highly
effective in treating both sanitary and combined
sewage.
• A carbon contact time of 5-10 minutes prior to
hydrous aluminum oxide precipitation is required
in order to insure consistently high treatment
efficiency.
• A total carbon contact time of less than 15
minutes is required for equilibrium removal of
sorbable organics.
• The carbon dose can be adjusted to effect the
degree of sorbable organic removal required.
• A residual, nonadsorbable fraction ranging from
10-20 mg/1 BOD and 20-50 mg/1 COD existed at times
in the Albany sewage. This fraction could not be
removed at activated carbon doses as high as
10 0 0 mg/1.
• A carbon dose of 500-600 mg/1 was required to
produce a high quality effluent from the municipal
sewage treated.
• It was possible to reduce the carbon dose to 200
mg/1 during the nighttime hours at the Albany site
without affecting effluent quality. Decreasing the
-------
carbon dose in this manner during that portion of
the day when the sewage has a low soluble organic
content can greatly reduce operating costs without
sacrificing treatment efficiency.
® An average carbon dose of 500 mg/1 achieved
tertiary levels of treatment in the case of com-
bined sewage during pilot operations. Jar test
data indicate that the reguired carbon dose to
achieve these levels is less than 400 mg/1.
® Alum [AI2(SO4)3'I8H2O] and polyelectrolyte require-
ments were 200 mg/1 and 2.0 mg/1, respectively for
both sanitary and combined sewage treatment.
• Lime [Ca(OH)2l requirements when reclaimed alum was
used in the treatment process averaged 190 mg/1
during the Albany, New York demonstration. It should
be possible to reduce the lime requirement to 150
mg/1 or less in an actual operating plant.
® Tube settling contributes to minimizing process
detention time. However, conventional sedimentation
can be employed in the system if a longer process
detention time is acceptable.
® The tube settler operated effectively at a hydraulic
loading of 2880 gpd/ft^.
e Efficient filtration was accomplished at filter
loading rates in excess of 4 gpm/ft2.
« Filtration provides an added degree of reliability
which is essential for a municipal waste treatment
plant and therefore, should be included if the
process is to be employed in this manner. On the
other hand, if a system is designed to operate
during periods of combined overflow only and a
certain amount of solids carry-over in the effluent
is acceptable, filtration may not be necessary.
« The powdered activated carbon treatment process
can accommodate wide fluctuations in effluent
composition.
• The treatment system lends itself to a high degree
of automation.
® During periods of storm flows it should be possible
to adjust chemical feed rates automatically on the
basis of influent flow rate.
-2-
-------
® Turbidity can be used as an index to predict
changes in other influent parameters during periods
of storm flows. Thus, it may be feasible to adjust
the carbon dose during a storm on the basis of
turbidity.
® Treatment of raw and combined sewage can be
accomplished in a total time of 50 minutes or less.
This short detention time leads to a small land
area requirement for the treatment system.
® The powdered activated carbon treatment process
described in this report is free of the hydrogen
sulfide problem associated with physical-chemical
systems utilizing granular activated carbon.
» The process can be operated on an intermittent
basis with a negligible time requirement for
startup if chemical feed stocks are maintained.
• The treatment process is highly reliable.
® Average removals in excess of 94 percent COD, 94
percent BOD, and 99 percent suspended solids were
consistently achieved in treating combined sewage.
» Average BOD, COD, suspended solids and turbidity
levels in the pilot plant effluent from the
sanitary sewage treatment operations were 17 mg/1,
36 mg/1, 5 mg/1, and 0.6 JTU, respectively.
® In characterizing a waste stream or in shaking down
a new plant, frequent sampling must be carried out
in order to detect rapid fluctuations in waste
stream composition.
® Physical-chemical treatment of raw municipal waste
streams in some instances will produce tertiary
levels of treatment while in others only secondary
levels can be achieved. Each waste stream must be
examined on a case-by-case basis to determine the
level of BOD removal which can be achieved.
SLUDGE HANDLING
• Carbon sludge should represent 1-2 percent of the
plant flow in a full-scale facility.
-3-
-------
« A product containing 22 percent solids can be
obtained from direct centrifugation of the
carbon sludge.
¦• A conditioning polymer is required to achieve 95
percent capture in the centrifuge.
• The required conditioning polymer dose for fresh
sludge is 1-2 lbs/ton dry solids. Sludge aged
for 2-3 days required a conditioning polymer dose
of up to 4 lbs/ton dry solids to achieve 95 percent
capture in the centrifuge.
CARBON REGENERATION
® Powdered activated carbon can be successfully
regenerated in a fluidized bed furnace.
• Satisfactory regeneration can be achieved at a
temperature of 1250°F with a stack gas oxygen
concentration of less than 0.5 percent.
® After 6.7 regenerations, the regenerated carbon is
as effective as virgin carbon in removing organic
matter from raw sewage.
® Average carbon losses per regeneration cycle were
9.7 percent.
• Hearth plugging problems during the pilot plant
operations resulted from corrosion of the recycle
gas system. Such corrosion problems can be pre-
cluded easily in design of a full scale system.
• A high initial buildup of inert materials in the
regenerated carbon during the first cycle regenera-
tions is believed to have resulted from causes
external to the regeneration system. Installation
of a grit chamber in the treatment system should
guard against high fluctuations in inert material
buildup in the regenerated product.
® Inert material buildup averaged 2.9 percent per
cycle during the pilot plant operations.
® Sand carryover from the fluidized bed furnace is
believed to represent the most significant fraction
of this buildup.
® Minimum operating costs are achieved with a five
percent blowdown of carbon and inerts.
-4-
-------
® Classification of the carbon and inerts would result
in less carbon lost to blowdown and thus a reduction
in operating costs.
® Stack gases from the regeneration furnace should not
present significant air pollution problems.
ALUM RECOVERY
o Approximately 91 percent of the aluminum can be
recovered by acidification of the carbon-alumina
slurry to pH 2 with sulfuric acid after thermal
regeneration of the carbon sludge.
® Acidification of the carbon-alumina slurry dissolves
inerts in addition to alumina. These dissolved
solids are discharged in the plant effluent and thus
the solids buildup in the reclaimed chemicals is
reduced.
® Sulfuric acid requirements for alum recovery were
0.6 lbs H2SC>4/lb of carbon for the pilot operations.
A reduction to 0.5 lbs H2S04/lb of carbon appears
feasible.
-5-
-------
SECTION II
RE COMMEN DATION S
• As a result of the successful laboratory studies
and pilot plant demonstration, efforts should
proceed toward a full-scale demonstration of the
powdered activated carbon treatment process.
o The powdered carbon treatment process should be
demonstrated on a large scale basis for treatment
of both raw and combined sewage.
• Design parameters for conventional sedimentation
should be developed.
® An effort should be initiated to define the upper
limit of loading of the tube settler without solids
carryover.
® The necessity of stack gas recycle should be
examined since elimination of the stack gas recycle
stream would reduce capital costs.
• A reliable and accurate analytical method for
determining the spent carbon content of sludge
should be developed.
• If tne powdered activated carcon process is to be
used as an advanced waste treatment process for the
treatment of raw sewage, it would be desirable to
develop a method to remove phosphorus from the
reclaimed alum. This would provide the added
dimension of phosphate removal in the treatment
process.
• The composition of the nonadsorbable organic
fraction should be determined and methods for its
removal explored.
• An efficient classification method for separating
sand from regenerated carbon should be identified
to reduce total blowdown and carbon losses.
® Further investigation of the usefulness of turbidity
as an index of solids and organic loadings during
storms should be pursued with a view towards auto-
mation of carbon and chemical feed rates in response
to turbidity monitoring signals.
-7-
-------
SECTION III
INTRODUCTION
The problem posed by combined sewer overflows is
well documented. In 1967, it was estimated that $48
billion would be required to eliminate these overflows by
installing separate collection systems for sanitary wastes
and storm waterd). This estimate did not include the
monetary loss which would be experienced by commerce and
industry as the streets through the centers of major cities
were torn up for installation of the separated systems. In
addition to economics, the possibility of separation of all
sewers, particularly in areas of high population density,
was remote. Even if complete separation were accomplished,
this measure would not be entirely satisfactory since storm
water runoff is, itself, often severely polluted, particu-
larly in highly urbanized areas(2,3,4). jn fact, the
pollutional effect of surface drainage water can be so
significant that it will, in many cases, be necessary to
treat storm runoff before it is allowed to reach receiving
waters(4).
Recognizing the tremendous problems associated with sewer
separation, the Storm and Combined Sewer Pollution Control
Branch of the Water Quality Office embarked on a program
to seek alternatives to this measure. They estimated that
the development of alternative means of treatment could
conceivably reduce this cost by two-thirds.
Therefore, as part of the Water Quality Office program,
Contract No. 14-12-519 was negotiated with Battelle-Northwest
to develop, through laboratory experimentation and pilot
plant demonstration, a novel physical-chemical process for
treatment of combined storm and sanitary sewage.
The treatment process was developed with the following
goals:
® The quality of the effluent should be comparable
to that routinely discharged from a secondary
sewage treatment plant.
® Short detention times are mandatory due to the
high flow rates likely to be encountered and the
undesirability of allocating large land areas
for treatment of combined sewage.
• The process must be amenable to intermittent
operation with a minimum time requirement for
startup.
-9-
-------
® Treatment of combined sewage by such a process
must be economically feasible.
The developed process, as outlined in Figure 1, involves
contacting raw or combined sewage with powdered activated
carbon to effect removal of dissolved organic matter. An
inorganic coagulant, alum, is ther. used to aid in subsequent
clarification. Addition of polyelectrolyte is followed by
a short flocculation period. Solids are separated from the
liquid stream by gravity settling, and the effluent is then
disinfected and discharged or can be filtered prior to dis-
infection.
Bench scale laboratory experiments and pilot studies indi-
cated that a 30-45 minute overall detention time is required.
A ten-minute contact time prior to flocculation is necessary
for good organic removal. Floe with excellent settling
characteristics is consistently produced. This, coupled
with highly efficient tube settling, leads to the very short
process detention time.
Carbon sludge from the treatment process is thermally re-
generated by a fluidized bed process. Alum is recovered by
acidifying the regenerated carbon-aluminum oxide mixture to
pH 2 with sulfuric acid. This reclaimed alum is then reused
in the treatment process. A pH adjustment, accomplished
with a lime slurry, is required to raise the pH to 6.5-7.0
for aluminum hydroxide precipitation when reclaimed alum
is recycled.
A nine-month laboratory study which is described in Appendix
A and elsewhere(5) indicated that the process goals could be
met. In addition, the laboratory phase of the program
demonstrated that the treatment process could be highly
effective for raw sewage. Consequently, a 100,000 gpd
mobile treatment plant (Figure 2) was designed and con-
structed .
Following construction of the mobile pilot plant, it was
operated for a one-month shakedown at Richland, Washington.
After minor equipment alterations were made, the process
performed as well or better than anticipated from the
laboratory study. Although this operation at Richland was
primarily for shakedown purposes, a limited amount of per-
formance data was obtained. Treating raw municipal wastes,
TOC removals averaged better than 90 percent while suspended
solids removals averaged better than 95 percent. Product
water turbidities were consistently below 2 JTU. The resul-
tant sludges were dewatered to 20 percent solids in a contin-
uous solid bowl centrifuge without any preconditioning or
supplemental thickening.
-10-
-------
MAKEUP pH
ALUM & ADJUSTMENT
POWDERED (LIME)
ACTIVATED
CARBON
POLYMER
RAW
SEWAGE
COMBINED
SEWER
OVERFLOW
RECYCLED
ALUM &
POWDERED
CARBON
RAPID
MIX
FLOCCU-
LATION
MINUTE
CONTACT
TUBE
SETTLING
FLUIDIZED
BED
FURNACE
SLUDGE
DEWATERING
FILTRATION
INERT
ASH
BLOWDOWN
FIGURE 1. PROCESS FLOW SHEET
-------
to
I
FIGURE 2. MOBILE PILOT PLANT
-------
Upon completion of the shakedown operation, the mobile pilot
plant was shipped to Albany, New York, where it was operated
from mid-June through October 1971. Since the pilot plant
was not designed for cold weather operation, it was shut down
for the winter months. Operations were resumed in early
April of 197 2 and continued through the month of June of that
year. This report describes the results of the pilot plant
demonstration in Albany.
-13-
-------
SECTION IV
PILOT PLANT DESCRIPTION
The pilot plant is composed of two major systems: a liquid
treatment system and a carbon regeneration facility.
TREATMENT SYSTEM
The liquid treatment system is housed, almost entirely, in
a forty-foot mobile trailer van. A schematic diagram of the
process flowsheet is given in Figure 3. The major compon-
ents are
a Surge Tank
c Pipe Reactor and Static Mixers
® Chemical Addition Equipment
® Flocculation Chambers
o Tube Settler
a Tri-media Filter
® Centrifuge
It is designed for a nominal capacity of 100,000 gal/day.
Carbon, alum, and polyelectrolyte are added in a pipe reactor,
providing rapid mixing of the chemicals, which preceeds
flocculation followed by separation via a tube settler.
Clarified effluent is chlorinated and released with the
option of routing through a gravity filter prior to chlorin-
ation. Sludge is dewatered in the centrifuge.
The system is designed for maximum operational flexibility
and includes turbidity, pH and flow monitoring instruments.
Sanitary or combined sewage is pumped from a sewer to a surge
tank, screened and then pumped to a six-inch diameter
stainless steel pipe reactor. The pipe reactor consists
of 62 sections of 7'4" pipe arranged in an eight by eight
array. The pipe centers are located at the apexes of equi-
lateral triangles. This arrangement allows for one pipe
elbow section to be connected at six locations while rotating
through a 360° circle. Connections between straight pipe
lengths and 180° returns are made with quick disconnect
couplings.
The total pipe reactor length of 560 feet will allow a
detention time of ten minutes at a flow rate of 75 gpm.
Chemical feed connections are provided at various locations
along the pipe reactor as illustrated in Figure 4 with
-15-
-------
I
M
G\
POLYELECTROLYTE
Q P 0
HYPOCHLORITE
LIME
TUBE
SETTLER
FILTER
FLOCCULATOR
i I
i I
SURGE
TANK
OVERFLOW
SLUDGE
PIPE REACTOR
FROM
SEWER
EFFLUENT
CENTRATE
CENTRIFUGE
ACIDIFIED ALUM-
CARBON SLURRY
FROM FURNACE
DEWATERED 4^
SLUDGE *=*
TO FURNACE
SLUDGE STORAGE
FILTERED
WATER
STORAGE
STORAGE TANK
FIGURE 3. SCHEMATIC FLOWSHEET OF MOBILE PILOT PLANT
-------
FIGURE 4. CHEMICAL INJECTION POINT IN PIPE REACTOR
-------
helical static mixers located after each chemical injection
point. These static mixers assure rapid, effective chemical
mixing. Powdered activated carbon and alum are added at the
head end of the pipe reactor, lime is added to adjust the pH
approximately 280 feet downstream and finally, polyelectro-
lyte is added at the end of the pipe reactor.
After polyelectrolyte addition, the treated influent can be
divided into one, two or three parallel streams. Each
stream enters a 400 gallon flocculator. At 25 gpm per
flocculator, nominal detention time is sixteen minutes.
Flocculator effluent leaves through a six-inch valve and
enters the tube settler via an increasing cross-sectional
area channel which keeps the flow velocity below one-foot
per second to minimize break up of floe. The tube settler
contains 25 sq. ft. of steeply (60°) inclined tubes. Separated
sludge is pumped to a 3200 gallon storage tank prior to
dewatering. Tube settler effluent can be chlorinated and
discharged directly or can be filtered prior to disinfection.
Filtration is accomplished in a 16 sq. ft. tri-media filter.
The filter contains 5 in. of 40 x 80 mesh garnet sand, 9 in.
of 20 x 40 mesh quartz sand, and 16.5 in. of 10 x 40 mesh
anthrafilt.
A six-inch solid bowl centrifuge is used to dewater sludge
which is then stored in a holding tank and subsequently
pumped to the carbon regeneration facility. Centrate from
the dewatering operation and filter backwash water are
returned to the surge tank for recycle through the treatment
system.
REGENERATION FACILITY
The fluidized inert sand bed unit of the regeneration facility
is 36 in. I.D., refractory lined, and self supported. As
illustrated in Figure 5, this unit consists of three main
sections: a firebox housing the burner 30 in. I.D. x 20 in.
high, a bed section containing inert sand 27 in. I.D.
bottom, 36 in. I.D. top x 60 in. high and a freeboard 36 in.
I.D. x 72 in. high.
Combustion of propane gas takes place in the firebox which
is also the point of injection of recycling gases to main-
tain a 2,000°F atmosphere. The hot gases pass through ver-
tical holes in a brick hearth fluidizing the inert sand
bed. Carbon sludge at approximately 78 percent moisture
is injected into the 1250°F turbulent bed where rapid heat
transfer is obtained between gases and materials. The
mixture of steam, combustion products, and regenerated
-18-
-------
WATER OR
RECYCLE GAS
OBSERVATION PORT
AND SAND FEED BIN
\
DOORS-
SAND CLEANOUTs
GAS OUTLET
WATER
''SPRAY
SAND L FVFL
FLU I D I Z ED
SPRAY
36" ID
27 1 D-»-
VXWXW//^
30" ID
3'
RECYCLE GAS INLET
T
20"
11'6"
SAND LEVEL
STATIC
/
SLURRY '
INJECTION
PORT
BURNER
3 '0"
FIGURE 5. FLUIDIZED BED REGENERATION UNIT
-19-
-------
carbon rises to the freeboard area and exits near the fur-
nace top on the side at 1200°F to a venturi scrubber and
cyclonic type separator. Exhaust gases at 150-200°F are
vented to the atmosphere through a scrubber stack with a
portion being recycled to the firebox. The regenerated
carbon slurry from the scrubber is passed through a 60 mesh
screen to remove large sand particles carried over from the
fluidized bed. During the 1972 operation, a small settling
chamber, three inches deep by one foot square, was installed
ahead of the screen. Most of the larger sand particles were
captured in this chamber, thus reducing the load on the screen
and eliminating the need for frequent cleaning of the screen.
After passing through the 60 mesh screen, the slurry was
collected in storage tanks. Water for the venturi scrubber
was continuously decanted from the storage tanks at 6-12 gpm
and recirculated.
The fluidized bed is suitably lined to produce a maximum skin
temperature of 150-200°F, while a bed temperature of approxi-
mately 1500°F is maintained. The shell itself is of steel
construction. The whole system is pressurized (windbox,
bed, freeboard) such that the summation of pressure drop
through the fluidized bed and scrubbing system will be less
than the pressure developed by the turbo-blowers. The
fluidized bed is provided with necessary access port, obser-
vation port, sand inlet, sampling ports, sand clean-out,
thermocouples, pressure tap connections, feed inlets and
auxiliary propane gas gun.
Once furnace operation is begun, its control is automatic.
Combustion air, recycle gas, propane and carbon sludge flow
rates are initially set manually to achieve the desired
temperatures, O2 level, and bed velocity. Once set, the bed
temperature is maintained automatically by varying the com-
bustion air flowrate. (The propane flow changes proportion-
ately with the combustion air flow to maintain a relatively
constant O2 level.)
After collection, the regenerated carbon slurry is acidified
to pH 2 with sulfuric acid in order to reclaim the alum
before reuse in the system.
A schematic diagram of the carbon regeneration system is
presented in Figure 6. Figures 7 and 8 are two views of the
regeneration furnace in place at the Albany site.
-20-
-------
SCRUBBED
STACK GAS
]
to
DEWATERED
SLUDGE
PROPANE
COMBUSTION
AIR
V ENTURI
SCRUBBER
CYCLONE
SEPARATOR
TRAILER
FLU I D I Z ED" :
. INERT".-':1-:
BED
CARBON
&
ALUM
RECYCLE
GAS
REGENERATION
FURNACE
MAKEUP ALUM
MAKEUP CARBON
h2so4
FIGURE
6. REGENERATION
SYSTEM SCHEMATIC FLOWSHEET
-------
,..ATCP pollution control research
WATER ^ r—3
elu®*»
FIGURE 7. PILOT PLANT IN OPERATION AT ALBANY SITE
-------
FIGURE 8. FLUIDIZED BED CONTROL PANEL
-------
SECTION V
DEMONSTRATION SITE
SITE DESCRIPTION
The demonstration site was located in the 400 block of South
Pearl Street in Albany, New York. Raw or combined sewage
was drawn from a 91 inch combined sewer, which serves the
Island Creek District in Albany. A site location map showing
the demonstration site and drainage area is given in Figure 9.
This drainage area is about 550 acres in size with a popula-
tion of approximately 10,000. Perhaps 10-15 percent of the
area is comprised of small commercial businesses with the
remainder being mainly residential in nature. No light or
heavy industry is located in the drainage area.
The average annual precipitation in Albany varies between
35 and 37 inches with the heaviest rainfall usually occurring
in June, July, and August.
WASTEWATER FLOW AND CHARACTERISTICS
Average dry weather flow (DWF) in the 91 inch trunk sewer
at the demonstration site was measured as 720 gpm. Average
minimum daily flow which occurred at approximately 0500 was
610 gpm and the average maximum daily flow of 820 gpm
occurred between 1000-12C0. Figure 10 contains typical
diurnal flow data for the Albany site.
The physical and chemical characteristics of the raw waste-
water at the site were observed to be highly variable.
Initially, grab samples were obtained at two-hour intervals
for several twenty-four hour periods and analyzed for various
constituents. During the 197 2 operation, samplers were in-
stalled to collect hourly composite samples. These data are
presented in Table 1 and Figures 11-13. Daily averages for
the maximum, minimum, and average day during the pilot
operations are presented in Table 2.
An unusually high COD peak was observed to occur frequently
at about 1400 as indicated in Figure 11. Such a high COD
peak was unexpected for a purely residential area. Moreover,
the time of occurrence is also somewhat unusual. The source
of this frequent high COD slug was not identified.
-25-
-------
\
to
<3>
V
,s»»- tq)
arm
a it //
NEW YORK STAT £ THRiiWA V
*1613*® 9 *
XSX^
¦0&
t)W
si^
T;lO^
-------
800-
s:
S 700
LU
I—
g 600
I
Ll_
UJ
= 500
9/28/71
0600
1200
1800
2400
0600
1200
1800
2400
TIME OF DAY
FIGURE 10. TYPICAL DIURNAL DRY WEATHER FLOW VARIATION AT ALBANY SITE
-------
Tim
of D
0800
1000
1200
1400
1600
1800
2000
2200
2400
0200
0400
0600
NOTE
TABLE 1
DIURNAL
WASTEWATER
CHARACTERISTICS
AT ALBANY
SITE
DURING DRY WEATHER
Suspended
Solids
COD
BOD
Total P
Turbidity
(mg/1)
(mg/1)
(mg/1)
(mg/1)
(JTU)
370-63
778-138
108-54
42.6-10.0
150-34
188-13
288-18
83-11
27.1-13
80-13
273-104
616-175
390-80
60.0-20.0
220-61
179-15
381-19
162-14
35.8-16
100-13
287-27
472-206
230-62
49.8-18.0
148-43
149-13
362-18
123-10
29 .4-13
86-14
500-47
5200-165
678-58
64.5-12.6
200-31
159-16
713-20
162-11
28.7-16
82-14
304-34
575-95
300-38
137-10.0
150-31
135-17
324-22
134-15
27.3-17
85-16
420-76
572-208
200-74
31.0-16
274-25
172-17
325-20
129-16
22.3-16
105-17
247-54
390-106
151-57
36.6-6.6
144-33
130-16
251-20
100-14
20.0-17
77-13
322-10
460-82
160-48
26.4-4.0
123-27
123-14
218-19
93-12
15.0-14
51-11
198-33
215-13
120-24
17.3-2.7
64-13
89-18
132-22
75-12
11.6-17
32-13
246-14
292-32
77-12
15 .6-2.7
78-9
65-17
90-24
38-13
7.4-13
21-12
62-12
305-25
111-12
18.0-3.3
42-1
29-14
67-22
27-13
8.5-12
17-12
206-11
979-25
110-16
29.9-3.9
150-3
68-14
162-21
41-12
12.6-14
44-12
Numbers for each time period represent: I high value - low value
l average - no. of samples
-28-
-------
150
100
50 -
® BOD
O COD
0800 1200
_L
_L
_L
_L
1600 2000 2400 0400
TIME OF DAY
_L
800
600
cn
£
400 §
C_>
200
0800
FIGURE 11. AVERAGE DIURNAL BOD AND COD FLUCTUATION AT ALBANY SITE
-------
I
GO
O
I
100
CO
ai
_L
O TURBIDITY
A SUSPENDED SOLIDS
J.
_L
_L
0800 1200 1600 2000 2400
TIME OF DAY
0400
_L
0800
FIGURE 12. AVERAGE DIURNAL SUSPENDED SOLIDS AND TURBIDITY FLUCTUATION AT ALBANY SITE
-------
J I I I I I I 1 I I I I I
0800 1200 1600 2000 2400 0400 0800
TIME OF DAY
FIGURE 13. AVERAGE DIURNAL PHOSPHATE FLUCTUATIONS AT ALBANY SITE
-------
TABLE 2
AVERAGE WASTEWATER CHARACTERISTICS AT ALBANY SITE1
DURING 1971 PILOT OPERATIONS
COD
BOD
2
Turbidity
Total
Solids
Total
Volatile
Solids
Suspended
Sol ids
Volatile
Suspended
Sol ids
Settleable^
Solids
NH-j-N
Organic
Nitroqen
N02-N
NO3-N
Total
P04
Maximum
483
159
315
533
268
668
150
4.0
55
25
0.2
1.0
57
Minimum
60
47
24
320
144
43
11
1.0
12
7.6
0.002
<0.1
12.6
Average
276
104
54
419
193
130
69
2.8
22
11.3
0.06
0 . 3
24.9
No. Days
42
37
43
18
16
45
16
11
12
12
12
9
13
^"All units mg/1 except as noted
2
Units are JTU
"*Units are ml/1 after 1/2 hr. settling
-------
SECTION VI
TREATMENT SYSTEM PERFORMANCE
GENERAL
All analytical procedures were carried out in accordance
with Standard Methods^ unless otherwise noted. During
the course of the field work, most of the analytical work
was subcontracted to Environment One Corporation of Schenec-
tady, New York. The New York State Department of Environ-
mental Conservation provided significant analytical support
to this program, especially in characterizing the waste stream
at the demonstration site.
Jar tests were carried out routinely in support of the pilot
plant activities. The general procedure involved addition
of the desired quantity of carbon slurry to one liter of
sewage and then adjustment of the pH to 4 with sulfuric acid
prior to addition of 200 mg/1 of alum. Samples were then
subjected to a ten minute rapid mix whereupon the pH was
adjusted to 7 with Ca(OH)2- Rapid mixing was continued for
an additional five minutes before addition of 2 mg/1 of
polyelectrolyte. Following an additional half minute of rapid
mixing, the sample was flocculated at 10 rpm for five minutes
and was then allowed to settle for twenty minutes. Super-
natant samples were decanted for analytical determinations.
In most instances, these samples were analyzed for BOD and
COD directly. Those cases in which samples were filtered
through 0.45u membrane filters are identified in the text
as "soluble" BOD and COD determinations.
PLANT OPERATION
During the course of the pilot plant studies , both virgin
carbon and regenerated carbon were used in the treatment
system. When virgin carbon (first cycle) was employed for
any run in the pilot plant, alum was added to the carbon
slurry (6 percent carbon by weight) in a chemical feed tank
and the mixture was acidified to pH 2 with sulfuric acid
prior to use. This simulated the conditions of a regenerated
carbon slurry which had been acidified to reclaim alum.
Injection of this acidified carbon-alum slurry into the raw
wastewater resulted in a pH of 3 -5-4 =
Careful control of the pH within the system was critical to
proper process performance. The laboratory studies showed
that, in order to consistently achieve good flocculation,
it was essential to provide several minutes of carbon contact
-33-
-------
with the wastewater prior to formation of the hydrous
aluminum oxide. In the pilot plant, this v/as achieved by
maintaining the pH below 4 for the first five minutes of
contact and then adjusting the pH to 6.5-7 with a lime
slurry. It was observed, in the course of the pilot
studies, that if the pH was allowed to rise above pH 4 within
the first few minutes, fine carbon particles were carried
over from the tube settler causing rapid filter plugging.
Moreover, under these conditions, the turbidity of the
filter effluent increased considerably. A three percent
lime slurry was used to adjust the pH in the system. Lime
feed was controlled automatically on the basis of pH.
Throughout the pilot operations, the alum [AI2(SO4)3-lSl^O]
dose was held constant at approximately 200 mg/1. It was
determined that the flocculation-sedimentation process deteri-
orated considerably if the alum dose was reduced much below
the 200 mg/1 level. On the other hand, the carbon dose
could be varied from 0-1100 mg/1 while maintaining the alum dose
at 200 mg/1 with no serious effect on the flocculation-
sedimentation efficiency.
Two types of powdered activated carbon were used in the study:
Aqua Nuchar (product of WESTVACO) and Darco XPC (product of
ICI America, Inc.). Analyses showed Aqua Nuchar A to be
approximately 90 percent fixed carbon while the Darco product
contained only 70-80 percent fixed carbon. Both carbons
performed comparably in the pilot studies at equal fixed
carbon doses. The relatively low fixed carbon content of
the Darco XPC is not surprising since this is an unwashed
grade of lignite carbon. After several adsorption/
regeneration cycles, the differences between this and a
washed grade of carbon may not be significant. However, it
was decided to use the higher grade Aqua Nuchar A for the
bulk of the pilot operations.
Three different high molecular weight anionic polyelectrolytes
were used in the pilot study: Atlasep 2A2 (product of ICI
America, Inc.), Decolyte 930 (product of Diamond Shamrock
Chemical Company), and Purifloc A-23 (product of Dow Chemical
Company). All of these polymers were observed to produce
large, rapidly settling floe particles. Each of these poly-
electrolytes performed satisfactorily at a dose of 2 mg/1.
Initially, high solids carryover from the tube settler was
observed when the system was operated at a flow rate greater
than 50 gpm. It was determined that this problem was the
result of poor flow distribution within the unit. Subse-
quently, a new influent distributor was installed in the
tube settler and the unit performed well at the design flow
of 70 gpm. This represents an overflow rate of 2880 gpd/ft .
-34-
-------
Under these conditions, the turbidity of the effluent from
the tube settler was consistently <2 JTU. Filter runs averaged
ten hours at a filter loading rate of 4.4 gpm/ft^. Backwash
was initiated at a terminal head loss of approximately 10 psi.
Sludge withdrawal was accomplished by pumping the sludge
from the base of the two hopper bottoms of the pilot tube
settler. Since there was no mechanical collection system in
the tube settler, it was necessary to withdraw the sludge
at a rate of approximately nine percent of the plant flow.
This sludge rapidly settled to 10-20 percent of its original
volume in the sludge storage tank. Therefore, it is reason-
able to expect that in a large clarifier with a mechanical
scraper, sludge volume would be 1-2 percent of the plant flow.
However, the absence of a mechanical scraper in the tube
settler did cause an additional problem. Channeling tended
to occur during sludge withdrawal even though the sludge
pump was operated on a 90 seconds on, 30 seconds off cycle
to minimize this problem. Consequently, efficient withdrawal
of sludge was not accomplished and the tube settler tended to
fill with sludge after about 48 hours operation. When this
occurred, solids tended to overflow from the tube settler and
onto the filter. The only feasible means found to correct
this condition was to completely drain the tube settler prior
to continuing operations. Once again, this problem should
not occur in a large clarifier or tube settler with a
mechanical scraper.
At no time during the operations was hydrogen sulfide de-
tected in the plant effluent. Even after the system had been
idle for several days an H2S odor was generally not detected
in the closed process trailer. However, in warm weather if
the system was not in operation and sludge was allowed to
remain in the tube settler for 2-3 days time, H2S was ob-
served to form. In actual plant operation, the sludge age
should be nowhere near two days and thus hydrogen sulfide
should not be a problem.
The carbon sludge was readily dewaterable in a Bird six-inch
solid bowl centrifuge. Dewatered sludge ranged from 20-35
percent solids at 70 percent capture with no conditioning
polymer.
Initial operation at a pool depth of 0.35 inch produced sludge
containing 26 percent solids. This sludge was very viscous and
extremely difficult to pump to the regeneration facility. In-
creasing the pool depth to 0.5 inch produced a much more
pumpable, 22 percent solids sludge. It was found that rapid
mixing of the dewatered sludge reduced its viscosity render-
ing it much more easily pumped.
-35-
-------
Solids capture was found to be improved by treating the
sludge with the same polymer used in the waste treatment
process. A polymer dose of 2 lbs/ton dry solids increased
solids capture to greater than 95 percent. Since no polymer
screening tests were conducted, it is expected that the 2
lbs/ton dose can be reduced substantially by selection of
the proper conditioning agent.
It was found that as the sludge aged, the polymer dose required
for conditioning increased. Sludge more than 2-3 days old
required polymer doses as high as 4 lbs/ton to achieve the
same solids capture in the centrifuge.
COMBINED SEWAGE TREATMENT
The treatment system performed well during the course of
nine storm events which occurred during the summer and fall of
1971 and in the spring of 197 2 at the Albany site. These
storm events ranged in duration from 2 to 7 hours with the
total rainfall during a single event ranging from 0.05 to
1.13 inches. Thus the combined sewage flows handled by the
treatment system are representative of a range of conditions
typical of the Albany area.
Operational data for the pilot treatment system during these
storms are given in Table 3. Turbidity, suspended solids, COD,
and BOD data for the storms are presented in Figures 14-32.
Plant detention times (listed in Table 3) should be taken into
account when comparing effluent quality data with influent
waste composition in these figures. Throughout the course
of all of these storms, plant effluent turbidity rarely ex-
ceeded 1 JTU and effluent suspended solids ranged from <1-18
mg/1, while the influent suspended solids exceeded 8800 mg/1
at the peak of one storm. Effluent COD, BOD, and suspended
solids averaged 23, 6.0, and 4.2 mg/1, respectively. This
represents average removals of 94 percent COD, 94 percent
BOD, and 99 percent suspended solids. During the peak pollutant
loadings of these storms the effluent quality remained essentially
unaffected resulting in removals as high as 99 percent COD, 99
percent BOD, and 99.9 percent suspended solids.
Inspection of the data in Figures 20, 28, 30, and 32 reveals
that the greatest portion of the COD in the combined sewage at
the Albany site was insoluble. However, these data also show
that the soluble fraction of the COD is of significant magni-
tude. This suggests that although solids removal is perhaps
the most important factor in combined sewage treatment, it is
not sufficient to produce a high quality effluent.
-36-
-------
TABLE 3
OPERATIONAL DATA DURING STORM FLOWS
Date
Total
Rainfall
(inches)
Combined
Flow
Duration
Plant
Flow
(gpm)
Detention
Time
(minutes)
Carbon
Dose
(mg/1)
7/13/71
0 .39
2030-0200
42
85
1300
7/19/71
0.55
1100-1600
70
50
625
7/29/71
0.30
1545-2000
50
70
800
9/16-17/71
0.46
2100-2215
40
88
800
2215-0100
75
47
800
5/2/72
0.20
0330-0730
40
88
570
5/16/72
0.05
0215-0430
40
88
500
5/16/72
0.15
0530-0730
40
88
500
5/16-17/72
0.45
2000-0230
40
88
500
-37-
-------
240 -
160 -
80 —
T
CONDITIONS
DARCO XPC @ 1300 mg/2,
DETENTION TIME: 85 MIN
O PLANT INFLUENT
• PLANT EFFLUENT
PLANT EFFLUENT
<1 JTU
A.
200
CD
e 160
120
80
40
2030
2100
2200
TIME OF DAY
2300
2400
FIGURE 14. STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 7/13/71
-38-
-------
1000
800
en 600
O
o
400
200
t 1 1 1 1 1 i r
CONDITIONS
DARCO XPC @ 1300 mg/K,
DETENTION TIME: 85 MIN
o PLANT INFLUENT
• PLANT EFFLUENT
2030 2100 2130 2200 2230 2300 2330 2400
TIME OF DAY
FIGURE 15. STORM EFFECT ON COD - 7/13/71
-------
4000
3000
2000
1000
8000
6000
4000
" 2000
1 1 1
1 I 1
CONDITIONS
DARC0 XPC 0 625 mg/£
DETENTION TIME: 50 MIN
-
0 O PLANT INFLUENT
-
1 MAXIMUM EFFLUENT
I
h TURBIDITY <1 JTU
i Vwwi 1
[ MAXIMUM EFFLUENT
-
1 S.S. <13 mg/£
1 | |
1000 1200
1400 1600 1800
TIME OF DAY
2000
FIGURE 16. STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 7/19/71
-40-
-------
T
CONDITIONS
DARCO XPC @ 625 mg/J.
DETENTION TIME: 50 MIN
O PLANT INFLUENT
• PLANT EFFLUENT
lA-
1100 1200 1300 1400 1500
TIME OF DAY
1600
1700
FIGURE 17. STORM EFFECT ON COD - 7/19/71
-41-
-------
500
CONDITIONS
DARCO XPC @ 800 mg/£
DETENTION TIME: 70 M
400
PLANT INFLUENT
PLANT EFFLUENT
300
i—
I—i
O
200
ZD
I—
100
PLANT EFFLUENT
JTU
1600
£
E
° 1200
_l
£ 800
Z
LU
400
1830
1900
1800
1600
1700
1730
1630
TIME OF DAY
FIGURE 18. STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 7/29/71
-42-
-------
CONDITIONS
DARCO XPC & 800 mg/Z
DETENTION TIME: 70 MIN
O PLANT INFLUENT
25
20
CD
B
LT)
Q
° 15
t/*)
LxJ
J
CO
C
LU
f—
I—
LlJ
1800
1900
2000
1700
1600
TIME OF DAY
FIGURE 19. STORM EFFECT ON SETTLEABLE SOLIDS - 7/29/71
-------
400
CONDITIONS
DARCO XPC @ 800 mg/S,
DETENTION TIME: 70 M
300
PLANT INFLUENT
PLANT EFFLUENT
o
o
qa
100
O PLANT INFLUENT, TOTAL
& PLANT INFLUENT, SOLUBLE
• PLANT EFFLUENT
1200
1000
800
5 600
o
o
CJ
400
200
2000
2100
1800
1900
1600
TIME OF DAY
FIGURE 20. STORM EFFECT ON BOD AND COD - 7/29/71
-44-
-------
CONDITIONS
4000
AQUA NUCHAR @ 800 mg/£
DETENTION TIME: 115 MI
O PLANT INFLUENT
^ 3000
EFFLUENT TURBI
>-
i—
t—i
Q 2000
1—4
CO
CL
1—
1000
mg/1
EFFLUENT S.S.
^ 8000
cr
E
go
2 6000
o
C/O
Q
S 4000
UJ
ZD
GO
2000
0600
0500
0400
0300
0200
2400
2300
0100
TIME OF DAY
FIGURE 21. STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 9/13-14/71
-45-
-------
160
120
80
40
600
500
400
300
200
100
FI
CONDITIONS
AQUA NUCHAR P 800 mg/S>
O PLANT INFLUENT
® PLANT EFFLUENT
I
_L
«.
• —a—
300 2400
0100 0200 0300
TIME OF DAY
0400 0500 0600
URE 22. STORM EFFECT ON BOD AND COD - 9/13-14/71
-46-
-------
T
160
120
O PLANT INFLUENT
« PLANT EFFLUENT
CO
80 -
40
CONDITIONS
AQUA NUCHAR P 800 mqJl
DETENTION TIME:
BEFORE 2215 - 88 Mil
AFTER 2215 - 47 Mil
q 400
2200
2300 2400
TIME OF DAY
0100
FIGURE 23. STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 9/16-17/71
-47-
-------
160
120
CT>
E
o
o
CO
80
40
1
1 1 1 1 1
CONDITIONS
A AQUA NUCHAR @ 800 mq/l ~
SX DETENTION TIME:
P \ BEFORE 2215, 88 MIN
- 1 \ /
( \ AFTER' 2215, 47 MIN
- o
\ JO O PLANT INFLUENT
® PLANT EFFLUENT
1
400
300
o>
£
O
o
200
100
e—© . ®—•—-®—<
_L
2200
2300 2400
TIME OF DAY
0100
FIGURE 24. STORM EFFECT ON BOD AND COD - 9/16-17/71
-48-
-------
500
CONDITIONS
EGENERATED CARBON @ 500
ETENT ION TIME: 88 MIN
O PLANT INFLUENT
O PLANT EFFLUENT
400
ZD
o 300
i—
o
200
100
2500
LT)
O
~ 1500
LlJ
= luuu
LU
=>
500
0700
0600
0400
0500
TIME OF DAY
FIGURE 25. STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 5/2/72
-49-
-------
1 1
I | I |
CONDITIONS
REGENERATED CARBON 1? 500 mg/2.
300
-
DETENTION TIME: 88 MIN
A 0 PLANT INFLUENT
/ \ ® PLANT EFFLUENT
200
-
/ \
100
.
_
0
1 ?
• , ?' "• ®—9
1000
800
600 -
400
200
J i I i I i x—
0400 0500 0600 0700
TIME OF DAY
FIGURE 26. STORM EFFECT ON BOD AND COD - 5/2/72
-50-
-------
CONDITIONS
REGENERATED CARBON G> 500 mg/I
DETENTION TIME: 88 MIN
o PLANT INFLUENT
e PLANT EFFLUENT «
= 300
a 200
t—i
CO
en
100
_lAa
lAonAa
0200
0300
TIME OF DAY
0400
0500
FIGURE 27. STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 5/16/72
-51-
-------
600
CONDITIONS
REGENERATED CARBON
500
DETENTION TIME:
O PLANT INFLUENT,
TOTAL
A PLANT INFLUENT,
SOLUBLE
• PLANT EFFLUENT,
TOTAL
400
5 300
o
o
o
200
80
40
0500
0200
0400
0300
TIME OF DAY
FIGURE 28. STORM EFFECT ON COD AND BOD - 5/16/72
-52-
-------
CONDITIONS
800
REGENERATED CARBON @ 500 mg/2.
O PLANT INFLUENT
Q PLANT EFFLUENT
600
400
200
1600
^ 1200
Q
o
o
800
Q
on
At
0500 0600 0700 0800
TIME OF DAY
FIGURE 29. STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 5/16/7 2
-53-
-------
CONDITIONS
1000
REGENERATED CARBON 0 500 mg/A
O PLANT INFLUENT, TOTAL
A PLANT INFLUENT, SOLUBLE
® PLANT EFFLUENT, TOTAL
800
o
o
<_>
400
200
140
120
E
80
0800
0500
0700
TIME OF DAY
FIGURE 30. STORM EFFECT ON COD AND BOD - 5/16/72
-54-
-------
6000
REGENERATED CARBON
3 500 mg/fl.
DETENTION TIME:
5000
LANT
LANT
FLUENT
4000
ID
l—
>. 3000
i—
o
CO
? 2000
1000 _
8000
ex.
E
6000
O0
o
_1
o
1/1 4000
~
UJ
Q
LiJ
= 2000
A.
0200
2400
2200
2000
TIME OF DAY
FIGURE 31. STORM EFFECT ON TURBIDITY AND
SUSPENDED SOLIDS - 5/16-17/72
-55-
-------
CONDITIONS
1000
REGENERATED CARBON 0 500 mq/J,
DETENTION TIME: 88 M
O PLANT INFLUENT,
A PLANT INFLUENT,
• PLANT EFFLUENT,
TOTAL
SOLUBLE
TOTAL
800
600
~
o
400
200
160
cn
E
o
o
CD
2000
2200
0200
TIME OF DAY
FIGURE 32. STORM EFFECT ON COD AND BOD - 5/16-17/72
-56-
-------
These observations are supported by a series of jar tests
which were run during the course of the 7/29/71 storm. Four
samples collected during increasing, peak, decreasing and minimum
solids loading periods were examined with the results presented
in Figures 33 and 34. The data in these figures emphasize
the important role of effective plant solids removal during a
storm. The points corresponding to 0 carbon dose represent
the unfiltered supernatant liquid after alum and polyelectro-
lyte treatment. A large portion of the total BOD and COD was
removed without the use of carbon; in fact, at peak solids
loading, over 90 percent of both BOD and COD were removed.
However, even at this level of removal, a residual COD of
9 2 mg/1 remained at the storm peak solids loading. Thus, alum
coagulation alone is not sufficient to produce a low COD
effluent. These jar test data suggest that as the storm
progressed and the COD was increasing, a carbon dose of 600
mg/1 was warranted to effect good COD removal. Once the storm
peak was passed, however, the carbon dose could be reduced to
the 200-300 mg/1 level with virtually complete removal of all
sorbable COD.
During the course of each of the storms, the suspended solids,
COD, and BOD profiles of the wastewater followed the turbidity
profile with the peaks of the curves usually occurring at the
same time. In some instances, the BOD and/or COD tended to
peak slightly ahead of the turbidity. It was noted that the
soluble COD followed the same general pattern as the total
COD. These observations suggest that turbidity can be used to
monitor trend variations in other influent parameters as a storm
progresses. A decrease in turbidity should be indicative of
a corresponding decrease in soluble organics. Therefore,
after the turbidity has peaked, it should be possible to
effect virtually complete removal of sorbable organics at a
substantially reduced carbon dose until the storm has sub-
sided at which time normal operation could be resumed.
Depending upon the effluent quality required and upon the
wastewater characteristics at a particular site, it would
be possible to operate the treatment process throughout storm
periods at a much lower carbon dose than the 600 mg/1 indi-
cated for raw sewage treatment. Thus, operating economies
could be introduced by operating at a low carbon dose for
an entire storm period or by reducing the carbon dose after
the peak organic loading had occurred. The second of these
alternatives assures the highest effluent quality at the
lowest cost.
Prior to the 9/13/71 and 9/16/71 storms, the New York State
Department of Environmental Conservation installed a flow-
meter in the sewer from which the plant influent was drawn.
Therefore, it was decided to attempt to adjust the pilot
plant flow rate to parallel the flow in the sewer during
-57-
-------
120
100
80
60
40
20
100
80
60
40
20
1 1 r
SOLIDS INCREASING
1630 HOURS
RAW SAMPLE
COD 312 mq/i
BOD 83 mg/S,
S.S.128 mg/£
STORM PEAK
1715 HOURS
RAW SAMPLE
COD 1216 mg/£
BOD
S.S
37 5 ,Ttg/ 2.
446 mg/£
-O—
-®-.
_L
_L
_L
200
400 600 800 1000 1200 1400 1600
FIXED CARBON DOSE (mq/0
FIGURE 33. JAR TEST DETERMINATION OF CARBON
REQUIREMENTS DURING INCREASING
AND PEAK LOADING OF 7/29/71 STORM
-58-
-------
80
60
40
20
cr.
E
~
o
g 100
~
o
"i r
SOLIDS DECREASING
1745 HOURS
RAW SAMPLE
COD
BOD
752
188
mg /1
ma/ 9.
80
60
MINIMUM LOADING
2000 HOURS
RAW SAMPLE
COD 208 mg/e,
BOD 69 mq/Sl
S . S . 35i? mq/ 8.
0 200 400 600 800 1000 1200
FIXED CARBON DOSE (mq/2)
1400
1600
FIGURE 34. JAR TEST DETERMINATION OF CARBON REQUIRE-
MENTS DURING DECREASING AND MINIMUM
LOADING OF 7/29/71 STORM
-59-
-------
subsequent storms. Attempts to accomplish this goal during
the 9/13/71 and 9/16/71 storms were not completely successful.
However, during the latter storm, the pilot plant flow rate,
initially at 35 gpm, was increased to 75 gpm in less than two
minutes during peak storm loading with no observable effluent
quality deterioration or operational upsets. Thus, it appears
that plant performance is highly insensitive to rapid changes
in flow rate if chemical doses are rapidly adjusted to
correspond to the increased flow.
Pilot plant operations have demonstrated that it is possible
to produce a high quality product water from combined
sewage in a total treatment time of 50 minutes. Moreover,
the system can accommodate rapid changes in flow and compo-
sition. It appears feasible that a system based on this
process could be highly automated. Carbon and chemical
doses could be controlled on the basis of flow and/or
turbidity monitoring to assure maximum system reliability
with minimum chemical usage.
MUNICIPAL SEWAGE TREATMENT
During most of the 1971 campaign, grab samples of the plant
influent and effluent were routinely collected at two hour
intervals and composited over a twenty-four hour period of
plant operation. Several determinations of the diurnal
variations in influent and plant effluent quality were made
during this portion of the program. Particular emphasis was
placed on this approach during the period when regenerated
carbon and reclaimed alum were in use. Plant operational
data for the 1971 portion of the program are given in Tables
4 and 5 and the results are presented in Table 6 and in
Figures 35-47.
The diurnal data showed BOD and COD peaks in the plant
effluent which could not be explained by variations in
influent quality or by operational upsets. These peaks are
evident in the curve presented in Figure 37. Later runs in
which the influent waste stream was sampled on an hourly
basis, as in Figures 40, 42, 44, and 46, indicated rapid
fluctuation in influent quality. High BOD and COD peaks of
perhaps an hour's duration frequently occurred. These
inordinately high COD peaks (as great as 5200 ir.g/1) appeared
to occur on a fairly regular basis and were totally unexpected
in a predominantly residential area presumably free of
industrial wastes.
Observation of the highly variable nature of the influent
quality at the Albany site suggested that the two hour samples
composited over a twenty-four hour period might not be com-
pletely representative of the waste stream. Reexamination of
-60-
-------
TABLE 4
PLANT
OPERATIONAL DATA FOR 1971
STUDIES
USING
VIRGIN CARBON
Wastewater
Detention
Polyelectrolyte
Carbon
Flowrate
Time
Dose
Dose
Carbon Type
Date
(gpm)
(min)
(mg/1)
(mg/1)
-7-71
40
88
6.0
800
Aqua
Nuchar
8
40
88
6.0
800
Aqua
Nuchar
9
40
88
4.0
1000
Aqua
Nuchar
10
40
88
4.0
1000
Aqua
Nuchar
14
40
88
2.0
1000
Aqua
Nuchar
15
40
88
2.0
1000
Aqua
Nuchar
16
40
88
2.0
1000
Aqua
Nuchar
17
40
88
2.0
800
Aqua
Nuchar
21
40
88
2.0
400
Darco
XPC
28
40
88
2.0
800
Darco
XPC
-21-71
40
88
2.0
1170
Darco
XPC
13
40
88
2.0
1330
Darco
XPC
14
40
88
2.0
1180
Darco
XPC
15
72
44
2.2
865
Darco
XPC
16
72
44
2.0
800
Darco
XPC
19
70
45
2.0
625
Darco
XPC
21
70
45
1.5
800
Darco
XPC
22
73
44
1.7
800
Darco
XPC
26
69
45
1.2
630
Darco
XPC
27
70
45
2.0
790
Darco
XPC
-11-71
41
86
3.5
550
Aqua
Nuchar
25
43
82
2.6
590
Aqua
Nuchar
26
45
78
2.7
486
Aqua
Nuchar
30
72
44
2 . 8
573
Aqua
Nuchar
-1-71
64
55
2.5
950
Aqua
Nuchar
2
62
57
2.7
880
Aqua
Nuchar
9
70
45
2.9
950
Aqua
Nuchar
-------
TABLE 5
PLANT OPERATIONAL DATA FOR DIURNAL STUDIES
Detention
Carbon
Lime
Acid
Time
Dose
Dose
Usage
Date
(min)
(mg/1)
(mg/1)
(lb/lbC)
Carbon Type
7/12-13/71
88
1170
_ —
Darco XPC
7/14-15/71
88
1190
Darco XPC
8/17-18/71
112
550
119
0.32
Aqua Nuchar
9/28-29/71
50
656
165
0 .64
Regenerated
10/4-5/71
50
482
220
0 .91
Regenerated
10/7-8/71
50
357
178
0.57
Regenerated
10/13-14/71
50
440
248
0 . 66
Regenerated
10/18-19/71
50
440
194
0 .43
Aqua Nuchar
10/25-26/71
50
537
171
0 .67
Regenerated
4/14-19/72
50
621
237
0.48
Aqua Nuchar
4/24-26/72
88
657
140
0.59
Regenerated
5/1-10/72
88
601
165
0 . 60
Regenerated
5/15/23/72
88
506
213
0 .76
Regenerated
6/7-12/72
88
570
226
0 .62
Regenerated
6/13-15/72
88
632
203
0 .74
Regenerated
6/16/72
88
640
182
0.65
Regenerated
-------
TABLE 6
plant Performance Data During
1971 Operations Using Virgin Carbon
COD
(rng/1)
BOD
(mg/1)
SS
(mg/1
Date
Influent
Effluent
Influent
Effluent
Influent
Efflu.
6-7-71
400
35
113
32
192
21
8
238
28
90
27
62
21
9
442
46
96
17
73
5
10
365
28
--
—
135
10
14
280
52
--
__
86
8
15
270
50
89
6
92
3
16
355
40
99
13
102
7
17
228
34
91
17
108
9
21
448
68
150
25
152
4
28
114
20
50
6
128
3
7-12-71
340
36
120
20
206
5
13
265
35
96
20
73
10
14
220
30
78
19
72
3
15
258
40
47
7
70
4
16
399
25
—
—
94
4
19
254
12
103
4
416
3
21
96
12
32
7
—
—
22
346
31
101
14
1C 2
9
26
119
24
46
5
63
4
27
268
20
107
15
106
3
8-11-71
195
55
72
29
43
6
25
235
45
105
18
106
4
26
233
19
122
25
98
7
30
269
38
129
15
78
24
9-1-71
210
45
105
28
104
29
2
217
27
82
20
78
12
9
242
50
107
25
99
6
Note: Samples were composited over 24 hr periods from grab samples at
2 hr intervals
-------
CONDITIONS
DETENT
160
o plant
FLUENT
140'
® PLANT EFFLUENT
120
100
400
300
200
100
0800
2000
0400
2400
1600
1200
0800
TIME OF DAY
FIGURE 35. PLANT SUSPENDED SOLIDS AND
TURBIDITY REMOVAL 7/12-13/71
-64-
-------
I
171
I
800
600
CD
Q
O
<_>
0800
400 —
200 —
1 200
FIGURE 36,
CONDITIONS
VIRGIN DARCO XPC 0 1170 mg/Si
DETENTION TIME: 88 MIN
O PLANT INFLUENT
• PLANT EFFLUENT
1600 2000 2400 0400 0800
TIME OF DAY
PLANT COD REMOVAL 7/12-13/71
-------
CONDITIONS
VIRGIN DARCO XPC 0 1190 mg/I
DETENTION TIME: 88 MIN
PLANT INFLUENT
PLANT EFFLUENT
1000 1400 1800
FIGURE 37.
2200 0200
TIME OF DAY
0600
1000
PLANT SUSPENDED SOLIDS AND
TURBIDITY REMOVAL - 7/14-15/71
-66-
-------
500
CONDITIONS
VIRGIN DARCO XPC @1190 mg/£
DETENTION TIME: 88 MIN
OPLANT INFLUENT
• PLANT EFFLUENT
400
300
cn
200
100
1 200
1 600
2000
2400
0400
0800
1 200
TIME OF DAY
FIGURE 38. PLANT COD REMOVAL 7/14-15/71
-------
bUU
1 1
1 1 1 1 1
CONDITIONS
400
— /
L Q VIRGIN AQUA NUCHAR @ 550 mg/£
\ f\ DETENTION TIME: 112 MIN —
\ / \ ° plant INFLUENT
V \ ® PLANT EFFLUENT
300
— /
\
200
— 1
a x l\ —
100
V* i
1 V-* 1
200
150
B
o 100
CO
0800
1200
1600 2000
2400
0400
0800
TIME OF DAY
FIGURE 39. PLANT COD AND BOD REMOVAL 8/17-18/71
-68-
-------
1000
CONDITIONS
ONCE REGENERATED CARBON
0 656 mg/I
DETENTION TIME: 50 MIN
O PLANT INFLUENT
® PLANT EFFLUENT
0200 0600 1000 1400 1800 2200 0200 0400
TIME OF DAY
FIGURE 40. PLANT COD AND BOD REMOVAL 9/28-29/71
-69-
-------
CONDITIONS
TWICE REGENERATED CARBON @ 482 mg/£
OPLANT INFLUENT
• PLANT EFFLUENT
350
300
250
CTl
E
200
I
° 150
Q
LU
Q
LU
^ 100
CO
0400
0800
1200
1600
2000
2400
0400
TIME OF DAY
FIGURE 41. PLANT SUSPENDED SOLIDS REMOVAL 10/4-5/71
-70-
-------
2500
2000
1500
CONDITIONS
TWICE REGENERATED CARBON
@ 482 mg/£
DETENTION TIME: 50 MIN
O PLANT INFLUENT
© PLANT EFFLUENT
a
° 1000
500
CD
E
o
O
CQ
0200 0600
2200 0200
FIGURE 42.
1000 1400 1800
TIME OF DAY
PLANT COD AND BOD REMOVAL 10/4-5/71
-71-
-------
-J
KJ
I
200
cn
E
m
ZD
oo
100
CONDITIONS
CARBON
DET
O PLANT INFLUENT
• PLANT EFFLUENT
0800 1600 2400 0800
FIGURE 43. PLANT SUSPENDED SOLIDS REMOVAL 10/7-8/71
-------
5200
400
300
200
100
CONDITIONS
THREE TIMES REGENERATED CARBON @ 357 mg/S>
DETENTION TIME: 50 MIN
200
PLANT INFLUENT
PLANT EFFLUENT
150
E
o
o
CO
100
1000
0200
0600
2200
1800
1000
1400
TIME OF DAY
FIGURE 44. PLANT COD AND BOD REMOVAL 10/7-8/71
-73-
-------
iOO
cn
400
CONDITIONS
FOUR TIMES REGENERATED CARBON
0 440 mg/&
DETENTION TIME: 50 MIN
O PLANT
• PLANT
INFLUENT
EFFLUENT
—j
1
on
Q
o
UJ
o
300
200
2200
100 —
-#=®-
0200
0600
1 000
1400
1800
2200
FIGURE 45.
TIME OF DAY
PLANT SUSPENDED SOLIDS REMOVAL 10/13-14/71
-------
2500
2000
1 500
1000
500
800
600
400
200
100
CONDITIONS
FOUR TIMES REGENERATED
CARBON @ 440 mg/£
DETENTION TIME: 50 MIN
0 PLANT INFLUENT
• PLANT EFFLUENT
4^
2400
0800 1600
TIME OF DAY
240
46. PLANT COD AND BOD REMOVAL 10/13-14/71
-75-
-------
CONDITIONS
400
ONCE REGENERATED CARBON
I? 537 mg/£
300
PLANT INFLUENT -
PLANT EFFLUENT
200
100
400
300
200
100
0800
2400
1600
0800
TIME OF DAY
FIGURE 47. PLANT COD AND BOD REMOVAL 10/25-26/71
-76-
-------
the data compiled on the two hour sampling schedule indicated
that it was highly probable that the high COD peaks were
frequently missed in the influent samples. On the other hand,
these peak loadings tended to spread out over a longer period
of time in the pilot plant and had an influence on effluent
quality for a 2 to 3 hour interval. Therefore, they were
detected in the effluent samples. Thus, the net result of
missing the high influent BOD and COD peaks should have been
to make the plant performance (on a percent removal basis)
appear somewhat poorer than was actually the case.
Average plant effluent BOD, COD, and suspended solids concen-
trations for the 1971 studies were 17.8, 35, and 7.7 mg/1
respectively. This represents removals of 32.3 percent BOD,
87.3 percent COD, and 94 percent suspended solids.
Prior to the start of the 197 2 operations, an automatic
sampling system which continuously composited influent and
effluent samples for one or two hour sampling periods was
installed. Data collected in this manner confirmed the
high variability of the sewage strength. Although the COD
peaks observed in the 1971 data were much subdued due to
the averaging quality of the composite samples they were
nevertheless present. Examination of the data shows that
the plant effluent quality significantly deteriorates
following these peak COD loadings.
Plant operational data for the 1972 studies are given in
Table 5 and performance data are presented in Figures 48-59.
In general, results were comparable to those observed in the
1971 portion of the program. During the 197 2 operations the
average effluent turbidity, suspended solids, COD, and BOD
concentrations were 0.67 JTU, 3.1 mg/1, 3 9 mg/1, and 17 mg/1,
respectively. This represents average removals of 98.1 per-
cent suspended solids, 82.6 percent COD, and 81.3 percent
BOD.
Effluent quality frequency distribution curves for BOD, COD,
and suspended solids for the diurnal pilot plant operation in
Albany are given in Figure 60. Figure 61 contains removal
probability curves based on the pilot plant data for these
same parameters.
Data from the pilot studies indicate that at times there was
a significant non-adsorbable organic component present in the
wastewater at the Albany site. When present, this non-
adsorbable fraction represented BOD and COD residual of
10-20 mg/1 and 20-50 mg/1, respectively, which could not be
removed even at carbon doses as high as 1000 mg/1.
A significant non-adsorbable fraction was not detected in
the Richland, Washington studies with powdered carbon nor in
-77-
-------
CONDITIONS
VIRGIN AQUA NUCHAR 0 621 mg/P.
DETENTION TIME: 50 MIN
O PLANT INFLUENT
A PLANT EFFLUENT
4/1 4/72
4/17-19/72
2400
2400 1200
TIME OF DAY
2400
1200
FIGURE 48. PLANT SUSPENDED SOLIDS AND
TURBIDITY REMOVAL - 4/14 &
4/17-19/72
-78-
-------
CONDITIONS
VIRGIN AQUA NUCHAR
DETENTION TIME: 50
400
300
200
TOO
4/14/72
4/17-19/72
O PLANT INFLUENT
A PLANT EFFLUENT
1 50
00
50
0 •-
1 200
2400
2400
1200
2400
1 200
TIME OF DAY
FIGURE 49. PLANT COD AND BOD REMOVAL
4/14 and 4/17-19/72
-79-
-------
cr>
E
in
o
Q
2:
a.
00
400 -
300 -
200 -
100 -
CONDITIONS
ONCE REGENERATED CARBON 0 657 mg/J>
DETENTION TIME: 88 MIN
O PLANT INFLUENT
A PLANT EFFLUENT
4/24-26/72
150
ca
Q£
100
50
I^-A ^wwj
2400 2400 1200 2400
TIME OF DAY
FIGURE 50. PLANT SUSPENDED SOLIDS AND
TURBIDITY REMOVAL 4/24-26/72
-80-
-------
CONDITIONS
ONCE REGENERATED
CARBON 0 657 mg/*,
DETENTION TIME:
_88 MIN
PLANT INFLUENT
A PLANT EFFLUENT
& 100
1200 2400 1200 2400
TIME OF DAY
FIGURE 51. PLANT COD AND BOD REMOVAL 4/24-2 6/72
¦81-
-------
5/9-10
CONDITIONS
150
TWO TIMES REGENERAT
CARBON 0 601 mg/£
DETENTION TIME: 88
Ol
E
O PLANT INFLUENT
A PLANT EFFLUENT
100
I
o
GO
Q
LlJ
Q
LU
ZD
CO
Q
200
150
>*
« 100
~
CQ
CC
13
I—
2400
1200
2400
1200
1200
2400
TIME OF DAY
FIGURE 52. PLANT SUSPENDED SOLIDS AND
TURBIDITY REMOVAL
5/1-3/72 & 5/9-10/72
-82-
-------
CONDITIONS
TWO TIMES REGENERATED
_ CARBON @ 601 mg/£
O PLANT INFLUENT
A PLANT EFFLUENT-
400
300
CJ>
E
g 200
o
100
5/9-10/72
5/1-3/72
150
5 100
CI
o
50
1200
2400
1200
2400
2400
1200
TIME OF DAY
FIGURE 53. PLANT COD AND BOD REMOVAL
5/1-3/72 and 5/9-10/72
-83-
-------
5/16-17/72
5/23/72—»-
CONDITIONS
THREE TIMES
REGENERATED CARBON —
@ 506 mg/fc
DETENTION TIME: 88 MIN
O PLANT INFLUENT
APLANT EFFLUENT
1200
FIGURE 54.
2400 1200
TIME OF DAY
1200
PLANT TURBIDITY AND SUSPENDED
SOLIDS REMOVAL
5/16-17/72 & 5/23/72
-84-
-------
5/2 3/7
CONDITIONS
THREE TIMES REGENERATED
CARBON 0 506 mg/&
DETENTION TIME: 88 MIN
O PLANT INFLUENT
A PLANT EFFLUENT
600
500
400
g 300
CJ
200
100
200
CT)
E
100
o
o
CQ
1200
1200
OF DAY
2400
TIME
FIGURE 55. PLANT COD AND BOD REMOVAL
5/16-17/72 and 5/23/72
-85-
-------
300 -
200 -
100 -
200 -
150 -
100 -
CONDITIONS
FOUR TIMES REGENERATED CARBON
@ 570 mg/Jl
DETENTION TIME: 88 MIN
OPLANT INFLUENT
APLANT EFFLUENT
6/12/72-
I A/rTftAK,!^wvl
2400
1200 2000 1200 2000
TIME OF DAY
FIGURE 56. PLANT TURBIDITY AND SUSPENDED
SOLIDS REMOVAL
6/7-8/72 & 6/12/72
-86-
-------
CONDITIONS
800
FOUR TIMES REGENERATED CARBON
600
400
200
O plant INFLUENT
A PLANT EFFLUENT
200
1 50
100
50
1200 2000 1200
2400
2000
TIME OF DAY
FIGURE 57. PLANT COD AND BOD REMOVAL
6/7-8/72 and 6/12/72
-87-
-------
CT>
E
LO
Q
O
CO
O
UJ
O
z
LU
a.
(/>
150
100
50
CONDITIONS
FIVE TIMES REGENERATED CARBON
DETENTION TIME: 88 MIN
O PLANT INFLUENT
A PLANT EFFLUENT
632 mg/£
-6/1 3-15/72-
150 -
CO
a:
100 -
50 —
1600
2400 0800 1600 2400
TIME OF DAY
0800
FIGURE 58. PLANT TURBIDITY AND SUSPENDED
SOLIDS REMOVAL
6/13-15/72
-88-
-------
600
CONDITIONS
FIVE T I!'. E S
REGENERATED CARBON
0 632 mg/£
DETENTION TIME: 88 MIN
O PLANT INFLUENT
A PLANT EFFLUENT
500
400
e 300
200
100
150
o
O
CO
1600
2400
0800
1600
2400
0800
TIME OF DAY
FIGURE 59. PLANT COD AND BOD REMOVAL
6/13-15/72
-89-
-------
SUSPENDED SOLIDS
in
r-i i
12 16
EFFLUENT SS (mq/O
20
24
50 75 100
EFFLUENT COD (mg/£)
125
1 50
20 30 40
EFFLUENT BOD (mg/£)
FIGURE 60. FREQUENCY DISTRIBUTION OF PILOT
PLANT EFFLUENT QUALITY
-90-
-------
1 .0
0.8
2 0-6
d.
o. 0.4
0.2
SUSPENDED SOLIDS
J I I I I 1.
i
1 -0 r-O^,
1 .0
0.8
0.6
0.4
0.2
—°o
BOD
I
40 50 60 70 80 90 100
REMOVAL
FIGURE 61.
CUMULATIVE PROBABILITY DISTRIBUTION OF
ACHIEVING REMOVAL PERCENTAGES GREATER
THAN OR EQUAL TO A GIVEN VALUE
-91-
-------
pilot studies with granular carbon by others^' ' , but was
observed in a recently completed pilot study in Cleveland,
Ohio^ '. This suggests that in some cases, physical-chemical
treatment of raw municipal wastes may not achieve greater
than secondary levels of biodegradable organic removal while
in other instances, tertiary levels can be realized. Each
waste stream must be examined on a case by case basis to
determine the level of BOD removal which can be achieved.
Although, for a significant portion of the time, tertiary
levels of BOD and COD removal were achieved at the Albany
site, on an average basis tertiary levels of BOD and COD
removal were not achieved. Tertiary turbidity and suspended
solids effluent quality v/ere consistently realized in the
pilot operations.
One of the attractive features of the powdered carbon treat-
ment process is the ability to vary the carbon dose with the
strength of the influent waste stream. Although any attempt
at continuous variation in carbon dose would probably present
operational problems which would render this approach im-
practical, once a waste stream has been characterized, it
should be possible to operate at two or three predetermined
carbon doses during the course of a day. In fact, this
concept was tested during the night hours at the Albany
site. After it had been established that the influent
waste strength was fairly weak between the hours of 2200
and 0600, the carbon dose was reduced to 0-400 mg/1 during
this time period for several days with the results given
in Table 7. These tests established that the carbon dose
could be drastically reduced during the nighttime hours
while maintaining a high quality effluent. The data of
Table 7 indicate that suspended solids removal is independent
of carbon dose as would be expected. Moreover, in the 2 00
to 4 00 mg/1 range, BOD and COD removal is independent of
carbon dose during this portion of the day. Subsequent to
these findings, during the months of September and October
1971, the carbon dose was routinely reduced to the 200 mg/1
level between the hours of 2200 and 0600 with no detectable
decrease in plant effluent quality. Decreasing the carbon
dose to a low level during the portion of the day when the
wastewater has a low soluble organic content can signifi-
cantly reduce operating costs without sacrificing treatment
2 f f 1C 2. S nC y • T7 q v ov rn v-n o ^ *i £ 2, pi srif V 0 -> +¦
carbon dose of 600 mg/1 for sixteen hours a day and at 2 00
mg/1 for eight hours, the average carbon dose would be
substantially below 600 mg/1 (the actual dose would depend
upon flow variations). Since carbon is the single most
important operating cost item in the treatment process, any
significant reduction in carbon dose represents a major
economy in operating costs.
-92-
-------
TABLE 7
PILOT PLANT PERFORMANCE AT LOW CARBON DOSES
Plant Influent
Plant Effluent
Time
of Day
Suspended
Solids
(mg/1)
COD
(mg/1)
BOD
(mg/1)
Suspended
Solids
(mg/1)
COD
(mg/1)
BOD
(mg/1)
4 00 mg
Carbon/1 8/19/71
2400
33
162
58
5
17
9
0200
14
116
37
7
21
10
0400
18
58
21
7
12
6
0600
21
37
29
5
12
3
2 00 mg
Carbon/1 8/25/71
2400
64
204
58
1
12
8
0200
32
63
24
31
8
3
0400
48
35
12
1
<1
2
0600
37
55
28
2
4
3
0 Carbon 8/27/71
2400
40
157
76
3
54
19
0200
86
134
45
<2
23
12
0400
18
100
26
4
37
8
0600
54
97 9
41
4
54
14
-93-
-------
Although a high degree of soluble phosphorus removal was
achieved in the pilot operations, such a result is not ex-
pected in an operating plant. In the pilot plant, excess
alum was added at times as discussed in Section VII of this
report. However, in actual plant operations, phosphorus
removed by the alum in early cycles would be recycled with
the regenerated carbon and recovered alum in later cycles.
An equilibrium condition would be established with no net
soluble phosphorus removal if there were no blowdown of
regenerated carbon. However, since the phosphorus concen-
tration in the regenerated carbon-alum stream is high and
blowdown will be required, some phosphorus will be removed
from the system. The makeup alum required by a five percent
blowdown will provide an anticipated equilibrium soluble
phosphorus removal of 31 percent.
The chlorine demand of the pilot plant effluent was investi-
gated. Two samples drawn at different times of the day (1400
and 2200) were treated with various chlorine doses and allowed
contact times of 15 and 30 minutes as shown in Figure 62. The
chlorine residual after 30 minutes contact averaged 0.2 mg/1
less than with 15 minutes contact.
Coliform analyses were run at various times of the day. The
results in colonies/100 ml sample are presented below.
Time Sample Taken: 0600 0900 1600
Raw Sewage 1,150,000 4,700,000 15,900,000
Tube Settler Effluent 872 18,000 7,500
Filter Effluent 584 14,100 7,000
Removal of coliforms in the plant was found to average 99.9
percent without disinfecting the effluent. It was observed
that filtering the tube settler effluent removed an additional
ten percent of the clarifier effluent coliforms.
Filtration of the tube settler effluent at the Albany site was
also observed to significantly improve pollutant removal per-
formance. The filtered effluent averaged 7 percent lower in
suspended solids and 20 percent lower in COD and BOD than the
tube settler effluent. This may have been due in large part
to the previously discussed operational problems associated
with the pilot tube settler.
The average neutralization lime usage was 190 mg/1. As seen
from Figure 63, the dose varied substantially from run to run
-94-
-------
A SAMPLED 1400 6-12-72 15 MIN CONTACT
~ SAMPLED 1400 6-12-72 30 MIN CONTACT
O SAMPLED 2200 6-13-72 15 MIN CONTACT
• SAMPLED 2200 6-13-72 30 MIN CONTACT
A
O
2
4
6
8
CHLORINE APPL I E D (nig A )
FIGURE 62. CHLORINE DEMAND OF PLANT EFFLUENT
-------
250
X
200 —
o
CD
150
LlJ
O
-------
such that no significant trend is observed. It is anticipated
that a reduction in lime dose below this level can be achieved
since, as discussed in Section VII of this report, it is believed
that the acid requirement for alum recovery can be substantially
reduced below that employed in the pilot studies.
Discounting equipment failures in the pilot system, the treat-
ment process proved to be highly reliable and was capable of
consistently producing a high quality effluent.
REGENERATED CARBON PERFORMANCE
Regenerated carbon and reclaimed alum performed as well as the
virgin substances in the pilot system. In all of the runs with
the recovered products, coagulation and sedimentation proceeded
normally. The removal of organic matter from the raw waste-
water was good during the runs with regenerated carbon. Data
on the system performance with virgin carbon and with regener-
ated carbon are presented in Table 8 and Figures 40-47 and
50-59. It is evident from the data of Table 8 that even with
regenerated carbon at fixed carbon doses as low as 370 mg/1,
the pilot plant COD, BOD, and SS removals are comparable to
a virgin carbon dose of 6 00 mg/1. After seven regeneration
cycles, the sorption performance of the carbon was essentially
that of virgin carbon as evidenced by the pilot plant data
and the jar test data presented in Figures 66-74 (Section VII).
-97-
-------
TABLE 6
REGENERATED CARBON PERFORMANCE COMPARISON
Plant Influent Plant Effluent Plant Removal
Fixed
Carbon
Carbon Dose COD BOD SS Turbidity COD BOD SS Turbidity COD BOD SS Turbidity
Batch
(mq/1) (mcj/1) (mq/1) (mq/1)
(JTU)
(mq/1) (mq/1) (mq/1) (mq/1)
%
%
%
%
Virgin*
600
276
104
130
—
35
17.8
7.7
—
87
83
94
—
1.0 Regenerations
A
650
278
—
—
—
27
14
3.7
--
90
—
--
—
1.9
A
580
411
--
135
--
56
23
7.0
—
86
--
95
—
2.8
A
370
690
--
248
—
45
36
18.2
—
94
--
93
--
3.7
A
470
337
160
85
—
44
14
3.1
—
87
91
96
—
0.9
B
540
246
142
--
—
43
27
—
—
82
81
—
—
Virgin
C
601
187
69
139
101
29
12
2.5
0.55
85
82
98
>99
1.0
C
633
323
133
190
90
47
24
2.5
0.84
85
82
99
>99
1.9
c
570
286
106
206
102
46
18
4.4
0.99
84
83
98
99
2.9
c
506
192
80
214
294
41
17
4.2
0.66
78
79
98
>99
3.8
c
570
2 91
106
139
119
56
26
3.1
0.47
81
75
98
>99
4.8
c
632
224
99
133
70
31
13
3.9
—
86
87
97
—
~Data from two months operation, daily composite samples
-------
SECTION VII
REGENERATION
OPERATIONAL PROCEDURES
During that portion of the 1971 pilot studies that the carbon
regeneration facility was operational, the liquid treatment
system and carbon regeneration system were operated alter-
nately on a campaign basis. The treatment system was
operated for a twenty-four hour period or until approximately
500 pounds of activated carbon was consumed. Sludge was
stored until the treatment operation was completed and then
was dewatered in the centrifuge and fed into the fluidized
bed furnace.
In the 1972 portion of the program, the liquid treatment
system and the carbon regeneration system were operated
simultaneously. For any given cycle, the treatment portion
of the pilot plant was operated alone until sufficient carbon
had been accumulated for the regeneration operation to
commence. Both systems were then operated until approximately
1000 pounds of carbon was exhausted in the treatment process.
The treatment system was then shut down until the next cycle
was begun.
Regenerated carbon-aluminum oxide slurry from the off gas
scrubber was collected in holding tanks during furnace
operation. Supernatant water from these storage tanks was
continuously decanted and recycled through the venturi
scrubber until the slurry concentration in the tanks built
up to approximately six percent carbon, which was suitable
for use in the treatment operations. Cooling of the recycled
slurry from the venturi scrubber was not necessary since it
left the disengaging vessel at 60°C. When the slurry con-
centration in the carbon tanks had reached the desired
concentration, a pH adjustment to pH 2 was then effected with
sulfuric acid and the necessary makeup carbon and alum were
added to the storage tanks.
Samples of virgin and regenerated carbon slurries were
analyzed for fixed carbon content. Analyses were obtained by
combusting a sample at 1200°C in a pure oxygen atmosphere
and then chromatographically determining the carbon dioxide
produced. Reproducibility using this technique was con-
sistently within five percent. Attempts were made to
analyze the fixed carbon content of spent carbon sludge by
heating at selected temperatures to drive off organics prior
-99-
-------
to combusting a sample. However, no reproducible results
could be obtained. Since the fixed carbon content of the
sludge feed to the regeneration furnace could not be de-
termined, it was necessary to determine carbon losses on
the basis of a mass balance across the entire system. Virgin
and regenerated carbons did not contain the organic component
which created analytical errors in carbon analysis of carbon
sludge samples.
Alum recovery was also determined on a mass balance basis.
Samples were drawn from each of the acidified carbon slurry
tanks, filtered through 0.45 micron membrane filter and the
filtrate was then analyzed for aluminum as outlined in
Standard Methods(10).
Due to the fact that the regeneration furnace became opera-
tional so late in the 1971 portion of the program, there was
only a limited time available for regeneration studies. Thus
it was deemed necessary to complete each regeneration-reuse
cycle in the minimum time possible in order to achieve a
maximum number of cycles. Therefore, there was insufficient
time to obtain analytical results on carbon recovery follow-
ing a regeneration cycle prior to commencing the next treat-
ment cycle of the operation. As a result, 1971 carbon doses
in the treatment system were lower at times than desired.
Makeup activated carbon was not added in the first two 1971
regeneration cycles but was added routinely after the second
cycle in order to maintain a total quantity of 500 pounds of
fixed carbon. Sufficient alum was added after each regenera-
tion to maintain the aluminum concentration in the feed
solution at 1.6 g/1.
In the 1972 operation, the carbon inventory was initially
maintained at 1000 pounds of fixed carbon. Following the
fourth regeneration cycle, the inventory was reduced to 500
lbs to decrease the time required per regeneration cycle.
Virgin carbon and alum were added as necessary to compensate
for losses.
At times it became necessary to begin using a tankful of
regenerated carbon before the results of the aluminum analysis
became available. To be certain of a minimum feed strength
of 1.6 g/1 aluminum, alum was added beyond what was actually
required. Thus the actual alum concentration in the treatment
system was sometimes higher than 200 mg/1.
SYSTEM STARTUP
Numerous mechanical problems with the furnace equipment were
encountered during the startup phase of the operation. Most of
-100-
-------
these problems were minor in nature and were easily corrected.
However, several major difficulties of a more persistent nature
caused a significant delay in actual furnace operations. These
major problems involved heat leakage around the hearth, lower
door, and burner. A great deal of time and effort was expended
in resolution of these problems. Final sealing of these leaks
involved replacement of the burner assembly and redesign and
replacement of the hearth and lower door. Although the initial
construction of the furnace was complete on May 27, 1971, the
regeneration system did not become fully operational until
September 17, 1971.
During the 1971 operation, the extra strength castable hearth
developed deep cracks necessitating replacement twice during
the campaign. This was corrected prior to commencement of
the 1972 operation by replacing the castable hearth with a
brick hearth and by modifying the controls such that the
combustion air system was automatically shut down in the
event of a flame failure. Previously, following flame
failure, the combustion air at 150°F continued to blow into
the 1900°F firebox, thus not allowing the fluidized bed sand
to percolate through the hearth. It is believed that the
thermal shock thus produced played a major role in previous
hearth failures. Following these modifications, no major
startup problems were encountered in the spring of 1972.
FURNACE OPERATIONS
Due to the fact that the fluidized bed furnace became
operational so late in the 1971 portion of the pilot program,
there was little latitude for experimentation with the
operating conditions. Using a ten inch diameter furnace, it
had previously been established that good carbon regeneration
and recovery could be obtained at a bed temperature of 1250°F
and an oxygen level of less than one percent in the stack
gas^-*-'^-^'. Therefore, these conditions were followed in the
pilot runs at Albany. Operating conditions for the fluidized
bed furnace are given in Table 9. Bed velocities reported in
Table 9 are calculated at a point midway up the bed using the
superficial area at that point. The gas velocity in the
freeboard zone is 0.78 5 times the reported bed velocity.
A total of six carbon regenerations were accomplished in
the pilot operations in 1971. Runs 1-4 represent a single
batch (Batch A) of powdered carbon followed through consecu-
tive cycles of use and regeneration. At the end of the fourth
regeneration, a substantial quantity of the carbon was lost
due to an operational problem; therefore, it was necessary
to begin again with virgin carbon at this point. Runs 5 and 6
represent the first and second regenerations of the second
batch (Batch B) of carbon. Runs 7-13 represent a single batch
(Batch C) of carbon regenerated seven times.
-101-
-------
TABLE 9
FLUIDIZED BED-FURNACE OPERATING CONDITIONS
Regeneration Ho.
1
2
3
4
5
6
7
8
9
10
11
12
13
Carbon Batch
A
A
A
A
B
B
C
C
C
C
C
C
C
Bed Temperature (°F)
1290
1260
1250
1250
1250
1260
1280
1270
1260
1250
1260
12 4 0
1220
Firebox Temperature (°F)
1900
1880
1910
2030
1980
1905
1840
1900
1360
1860
1860
1860
1880
Bed Velocity (ft/sec)
1.4
1.6
1.6
1.4
1.3
1.1
1.7
1.5
1.7
1.9
2.1
2.2
2.2
Recycle Gas Flow (SCFM)
83
89
83
60
60
63
61
62
78
107
123
126
138
Combustion Air Flow (SCFM)
71
89
86
82
68
63
80
68
66
64
59
67
60
Propane Flow (SCFM)
3.0
3.7
3.6
3 .4
2.8
2.6
3 . 5
3 . 0
2.9
3.1
3.3
3.7
3.6
Sludge Feed Rate (lbs/hr)
120
125
120
120
115
80
119
105
96
64
94
106
117
Solids Concentration in Feed (%)
21.7
24.1
22 . 9
22.6
23.5
22.9
22.0
21.1
21.9
21.9
23.5
25.7
_
-------
Since it was impractical to completely drain all of the
carbon from the treatment system prior to each regeneration
cycle, a fraction of the carbon was not regenerated each cycle.
In the case of the 1971 runs, this represented 4-10 percent
of the carbon per cycle. The result was that even though four
cycles were completed in the case of the first batch of carbon,
the actual number of regenerations was only 3.7.
Due to the larger carbon inventory of the 1972 runs, the
system holdup represented a smaller fraction of the total
carbon. Runs 7-13, a total of seven cycles, provided 6.7
actual carbon regenerations.
The temperature of the fluidized sand bed was monitored by
four thermocouples evenly spaced around the bed at various
levels. It was observed that temperatures at all four points
could be maintained within a 50°F range when the bed was
properly fluidized. At low gas flow rates through the bed,
the thermocouple opposite the carbon feed point indicated
temperatures higher than the thermocouple adjacent to the
entering slurry, a result of poor bed fluidization. Tem-
peratures in the freeboard zone measured near the furnace
gas outlet consistently registered 50-100oF lower than the bed
temperature.
The firebox temperature was maintained below 2100°F to avoid
damage to the furnace material. Temperature was regulated by
controlling the ratio of combustion air to recycle gas. Re-
cycle gas entering at 150-200°F cooled the firebox while
keeping the oxygen level low and providing the proper
fluidization gas flow.
An effort was made to maintain the bed velocity at as low a
level as possible in order to minimize attrition of the sand
and also to prevent sand carryover with the carbon. However,
it was observed that at fluidizing velocities of about 1
ft/sec or less, heat transfer from firebox to bed was much
poorer than at higher velocities. During Run 3, at a velocity
of 1.6 ft/sec, 120 lbs/hr of sludge could be fed while keeping
the bed at 1250°F and the firebox about 1900°F. However,
when the fluidizing velocity was decreased, it was necessary
to substantially decrease the sludge feed rate to the furnace
in order to maintain the desired temperature. For example,
in Run 6, when the fluidizing velocity was decreased to
1.1 ft/sec, the sludge feed rate had to be correspondingly
reduced to 80 lbs/hr in order to maintain a bed temperature
of 1260°F.
It was necessary to maintain the combustion air blower in
continuous operation even when the regeneration furnace was
not in use in order to prevent the sand from flowing through
the holes in the hearth and into the firebox. This contributed
-103-
-------
to attrition of the sand in the furnace which was observed to
be substantial. Sieve analyses of the sand both before use
and after 21, 35, and 41 days of fluidization in the furnace
are presented in Table 10. If attrition cannot be substantially
reduced, periodic withdrawal and replacement of the sand appears
indicated. It is suggested that provisions for keeping the
sand above the hearth during blower shutdown be included in the
design of the hearth. Proper installation of bubble caps may
provide good horizontal distribution of the gas while checking
sand flow into the firebox during blower shutdown.
The oxygen content of the fluidized bed off gas was determined
by Orsat Analysis. Once the proper C>2 level was attained, a
Hayes Model 635 oxygen analyzer continuously monitored the
level. Periodic adjustments of the fuel-air ratio were made
as necessary to maintain the oxygen content of the stack gas
at the 0-0.5 percent level.
One of the major operational problems which occurred during
the course of the program was plugging of the holes in the
hearth. Toward the end of the 1971 operations in Albany, the
pressure drop across the hearth rather suddenly increased to
the point where shutdown was necessary. Inspection of the
system revealed that sand was caking inside the holes in the
hearth, thus restricting the gas flow.
Prior to beginning the 197 2 operation, a 6 inch layer of
3/4 inch diameter high temperature gravel was placed on top
of the hearth. It was felt that this would provide an addi-
tional buffer layer between the 1250°F bed and the 2000°F
firebox as well as reduce the amount of sand that percolates
through the hearth during furnace shutdowns. About two
weeks after regeneration began during the 1972 campaign, the
hearth pressure drop again became excessive. Inspection of
the hearth revealed that the 6 inch layer of gravel on the
hearth had become cemented together into one large mass.
X-ray analysis showed the "cement" to contain a large amount
of magnetite. (The presence of a mixture of FeO and Fe203 is
understandable since maintaining an oxygen level in the
fluidized bed slightly above zero required frequent adjust-
ments in operating conditions. Although the furnace atmos-
phere was normally slightly oxidizing, at times, reducing
conditions probably existed.) The caked sand removed from
the inside of the hearth holes also was found to contain
large amounts of magnetite. Plugging of the holes in the
hearth continued to be a problem throughout the 1972 portion
of the program.
In addition to causing periodic system shutdown, plugging
presents other potential problems. It is believed that as some
of the holes plugged with sand and magnetite, nonuniform
fluidization occurred. In this case, spouting of the bed
-104-
-------
TABLE 10
SIEVE ANALYSES OF FLUIDIZED BED SAND
Sand Retained
on Screen #
Quantity
Retained
(g)
Virgin Ottawa Flint Shot Sand
20
30
40
50
Pan
8.6
325.0
164.0
16.0
1.9
5IS7T
Furnace Sand Collected After 21 Days Fluidization
Fraction
Quantity
Retained
(%)
1.67
63.04
31.81
3.10
0.37
20
0.89
0.38
30
99.18
41.97
40
116.46
49.28
50
12.79
5.41
60
3.66
1.55
70
0.88.
0.37
100
1.59
C. 67
Pan
0.86
0.36
236.31
99.9$
Furnace Sand Collected After 35 Days Fluidization
20
0.88
0.21
30
121.68
28.78
40
139.83
33.07
50
63.31
14.97
60
45.14
10.68
70
24.67
5.83
100
26.33
6.23
Pan
0.97
0.23
422.81
100.00
Furnace Sand Collected After 41 Days Fluidization
20
2.4l
0.56
30
112.8
26.58
40
140.4
33.09
50
77.4
18.24
60
42.0
9.90
70
23.3
5.49
100
20.8
4.90
Pan
5.2
1.23
424.3
99.99
Contained about 1 g of chunks of fused material from the furnace.
-105-
-------
material could occur causing increased sand carryover into the
product. Uneven fluidization would also result in nonuniform
carbon retention times thus reducing the carbon regeneration
efficiency.
Identification of the source of iron became of major importance
since iron appeared to be the cause of the biggest operational
problem in the regeneration system. An iron inventory for each
regeneration cycle was determined with the results as shown
in Figure 64.
It was noted that the actual iron inventory increased initially
and then leveled off at about 13 pounds. However, when
physical losses were taken into account, the iron inventory
increased more or less constantly for the first five cycles at
about six pounds per cycle. An identification of the possible
sources could not account for anywhere near this quantity of
iron. The maximum possible contributions of iron from various
sources were identified as follows
Fe from sewage treated 0.3 9 lbs per cycle
Fe from lime impurities 0.3 5
Fe from alum impurities 0.64
Fe from H2SO4 impurities 0.01
Total Fe 1.3 9 lbs per cycle
Thus a maximum of 1.4 pounds of iron per cycle could be
identified as possibly entering the system from known
sources. This left 4.6 pounds of iron per cycle which could
not be attributed readily to any source.
When the holes in the hearth plugged, the caked material filled
the entire six inch length of the holes. This suggested that
the iron was entering the system below the hearth since it
appeared unlikely that material could work down from the top
of the hearth and plug the holes in depth while the high
velocity gas stream was blowing upward through the holes. In
addition, the absence of any significant quantity of aluminum
in the materials which caked in the hearth holes indicated
that a special source of iron introduced below the hearth was
responsible for the plugging. Neither the fresh air stream
nor the propane stream were probable sources of large quanti-
ties of iron. Therefore, it was deduced that the iron was
probably entering the firebox in the recycle gas stream.
Shortly after the iron buildup was discovered, the recycle
gas blower failed. It was discovered that two of the
thirteen blower fans were severely corroded and consequently
were out of balance thus causing the blower failure. Further
inspection of the remainder of the recycle gas system revealed
-106-
-------
20
10
6 7 8 9 10 11 12
REGENERATION NUMBER
FIGURE 64. IRON INVENTORY IN REGENERATED CARBON SLURRY
O ACTUAL Fe INVENTORY
A Fe INVENTORY CORRECTED
FOR PHYSICAL LOSSES
-------
that the inside of the six inch diameter black iron piping
was covered with a rust scale that was easily dislodged.
Following repair of the blower, the recycle gas stream was
closely observed and it was noted that large clouds of red
dust (presumably rust) were often present in the recycle
system at blower startup. Thus, it was concluded that the
major source of iron which caused plugging of the hearth
derived from corrosion of the recycle gas system.
Design of a corrosion resistant recycle gas system should
eliminate the introduction of iron into the firebox and thus
should eliminate the hearth plugging problem. This would
reduce the total iron buildup to a maximum of 1.4 pounds per
cycle or 0.0018 pounds of iron per pound of carbon used.
However, this residual iron should not cause plugging problems
in the hearth since it would be introduced into the fluidized
bed zone of the furnace and thus should not work down to plug
the hearth. Installation of bubble caps should provide addi-
tional insurance against the hearth plugging from the top
down. Therefore, it is concluded that the serious plugging
problems encountered in the Albany pilot plant operations
can be easily precluded by proper use of corrosion resistant
materials of construction.
The buildup of inert material in the carbon following each
regeneration is plotted in Figure 65. Batch A regenera-
tions are characterized by a high initial increase in inert
material in the first regeneration cycle and then a gradual
increase in inert content at the average rate of 3.2 percent
per cycle for successive regenerations. In the case of
carbon Batch B, insufficient data are available to establish
a trend. Inert material buildup in the third batch of
carbon (Batch C) followed the same general trend as that
established for the first carbon batch -- a high initial
buildup of inert material in the first regeneration cycle
and then a gradual increase in subsequent cycles, in this
instance at the rate of 2.8 percent per cycle. During the
seventh regeneration cycle, the sand trap on the product
line failed for a portion of the run thus permitting a
higher than usual sand carryover into the product and conse-
quently, a higher inert buildup.
Average inert material buildup in the carbon for cycles
subsequent to the first cycle was 2.9 percent per cycle
for the entire pilot plant operation. The average fixed
suspended matter of 23 daily composite sewage samples was
41 mg/1. Considering that 94 percent of the suspended solids
are removed in treatment at a carbon dose of 600 mg/1, the
regenerated carbon should build up ash at the rate of 6.4
percent per cycle. Actual pilot plant data show a 2.9 percent
buildup. It is likely that acidification of the regenerated
carbon to pH 2 dissolves a significant quantity of the ash.
-108-
-------
CARBON BATCH
o
CD
DC
«=£
CARBON BATCH A
Li_
O
h-
LlJ
I—
o
o
20
CARBON BATCH
5
2
3
4
6
7
0
NUMBER OF REGENERATION CYCLES
FIGURE 65. INERTS BUILDUP IN REGENERATED CARBON
-------
Most of this dissolved ash is probably discharged in the
product water from the liquid waste treatment system.
Microprobe analyses of the inert material for cations of
molecular weight greater than Na revealed Si to be the over-
whelmingly predominant species. This result was confirmed
by X-ray analyses which indicated Si02 was the main constituent
of the inert material. These observations suggest that the
major component of the inert material in the product from the
regeneration furnace was sand.
Since the inert material in the regeneration carbon stream
built up at a regular rate for all but the first cycle, it is
apparent that special circumstances unique to the first cycle
in the cases of carbon Batches A and C must have been respon-
sible for the high initial jump in inerts. Considering carbon
Batch A, the 15.5 percent increase in inerts in the first
regeneration cycle represents 122 pounds of material. In the
case of carbon Batch C, 303 pounds of inert material were
required to cause the 30 percent jump in the first cycle.
It appears that a major portion of the inert buildup during
the first regeneration of carbon Batch C resulted as a conse-
quence of improper positioning of the plant influent hose in
the sewer. Plant influent was drawn from a flat bottom area of
the sewer 7.5 feet wide used as a metering chamber. The rear
edge of a 12 inch wide by 12 inch deep trench extending across
the chamber formed a weir with the depth of overflow being used
to determine flowrate. The pilot plant influent line was
wedged into this trench to prevent its becoming dislodged dur-
ing high flows in the sewer. The trench, however, contained a
large quantity of sediment and gravel which was drawn into
the pilot plant during the first few hours of operation. In
fact, the vertical influent hose soon became plugged with
gravel.
During the 1971 operations the liquid treatment portion of
the plant was operated for several months before the regen-
eration facility became operational, allowing ample time to
clear the sand and gravel from the trough• On the other
hand, during the 197 2 operation, sludge was collected and
regenerated from the start of the treatment operations. Thus,
the material sucked from the trough was fed to the furnace
with the sludge where it could be abraded and/or thermally
cracked in the fluidized bed and then find its way into the
product carbon stream.
It is believed that the major portion of the 308 pounds of
inert material buildup in the first regeneration of carbon
Batch C resulted from this cause. Since a full scale plant
would have a grit chamber, a problem of this nature should
not occur. In fact, a grit chamber should effect some
-110-
-------
reduction in inert material buildup in each regeneration
cycle and therefore, one would expect a lower average buildup
than observed in the pilot runs.
No sufficient reason for the high initial buildup of inert
material in the first regeneration cycle of carbon Batch A is
evident. The cause cannot be attributed to a high initial
attrition of virgin sand and carryover in the product since
the furnace sand was completely replaced several times during
the pilot program and no high increase in inerts was evident
in subsequent regenerations. Thus it appears that the source
of the inert material buildup in the first cycle regenerations
of both carbon Batches A and C was external to the regeneration
system and most probably consisted of material present in the
influent.waste stream. A grit chamber should eliminate this
contribution of inerts to the regenerated carbon.
Following each regeneration cycle, jar tests were run with
samples of the regenerated carbon, duplicating as nearly as
possible the pilot plant operating conditions. Data from
these jar tests are presented in Figures 66-74. Examination
of this data indicates that the regenerated and virgin carbons
were virtually identical in sorptive capacity. During the
first part of the first regeneration run (Batch A), highly
unstable temperature and sludge feed conditions persisted in
the regeneration furnace. Figure 66 illustrates the adverse
effect that these unstable conditions had on recovery of the
sorptive capacity of the carbon. Later in the run when the
system had stabilized, regeneration was much better as indi-
cated by the sorption curve in Figure 66. The pilot plant
data for the regenerated carbon given in Table 8 substantiate
the observation that full capacity recovery could be achieved
after numerous regenerations.
ALUM RECOVERY
It is well known that aluminum hydroxide goes through the
conversion
A1(0H)3°°°C •y-Al2O310^0°C ci-Al^
The y-oxide readily dissolves in sulfuric acid to yield
aluminum ions while the a-oxide is highly acid resistant.
Since carbon regeneration is effected in the fluidized bed
furnace below 1000°C, the majority of the hydrous aluminum
oxide present in the carbon sludge should be converted to
the soluble alumina, Y-AI2O3, in the regeneration process. The
laboratory studies described in Appendix A indicated that it
should be possible to recover 8 0-100 percent of the aluminum
following regeneration.
-Ill-
-------
I
80 -
60 -
40 -
VIRGIN NUCHAR
FIRST 1/3 OF RUN
LATTER 2/3 OF RUN
RAW SEWAGE COD 436 mg/2,
200 400 600
FIXED CARBON DOSE (mg/£)
800
FIGURE 66. JAR TESTS ON ONCE REGENERATED CARBON - BATCH A
-------
VIRGIN
TWICE REGENERATED CARBON
200
E
~
o
I
= 100
LU
500
600
700
800
900
100
200
300
400
FIXED CARBON DOSE (mg/£)
FIGURE 67. JAR TESTS ON TWICE REGENERATED CARBON - BATCH A
-------
80
40
•THREE TIMES REGENERATED CARBON
20
400
600
800
200
FIXED CARBON DOSE (mg/£)
FIGURE 68. JAR TESTS ON THREE TIMES REGENERATED CARBON - BATCH A
-------
O VIRGIN CARBON
• FOUR TIMES REGENERATED CARBON
40
RAW SEWAGE COD 357 mg/£
80
60
40
200
600
400
800 1000
1200 1400 1600
FIXED CARBON DOSE (mg/A)
FIGURE 69. JAR TESTS ON FOUR TIMES REGENERATED CARBON - BATCH A
-------
140
VIRGIN CARBON
ONCE REGENERATED CARBON
TWICE REGENERATED CARBON
1 20
100
RAW SEWAGE COD 375 mg/£
80
60
40
20
800
600
400
200
FIXED CARBON DOSE (mg/J.)
FIGURE 70. JAR TESTS ON ONCE AND TWICE REGENERATED CARBON - BATCH C
-------
O VIRGIN CARBON
A THREE TIMES REGENERATED CARBON
60
RAW SEWAGE COD 424 mg/£
30
20
600
800
200
400
FIXED CARBON DOSE (mg/£)
EIGURE 71. JAR, TESTS ON THREE TJM£S REGENERATED CARBON t BATCH C
-------
O VIRGIN CARBON
1
1
80
40
400
600
200
800
FIXED CARBON DOSE (mg/£)
FIGURE 72. JAR TESTS ON FOUR TIMES REGENERATED CARBON - BATCH C
-------
lO
1
o>
Q
O
o
c
rD
Q
y—i
on
LU
120 -
100
80
60 —
40 -
20 -
T
0 VIRGIN CARBON
A FIVE TIMES REGENERATED
~ SIX TIMES REGENERATED
RAW SEWAGE COD 323 mq/l
200 400 600 800 1000 1200 1400
FIXED CARBON DOSE (mg/2.)
FIGURE 73. JAR TESTS ON FIVE AND SIX TIMES REGENERATED CARBON - BATCH C
-------
OVIRGTN CARBON
A SEVEN TIMES REGENERATED CARBON
RAW SEWAGE COD 53.5
50
to
0
1
o?
CD
O
O
C
ZD
Q
»—i
m
LlJ
en
40
30
20
10
¦xr
200 400 600
FIXED CARBON DOSE (mg/J>)
FIGURE 74, JAR TESTS ON SEVEN TIMES REGENERATED CARBON
800
- BATCH C
-------
In the pilot operations, alumina was conveyed from the furnace
with the carbon and deposited in the storage tanks. Following
the carbon regeneration operation, the carbon slurry was
acidified to pH 2 with sulfuric acid in order to dissolve the
alumina and free aluminum ion. The average recovery of
aluminum was determined to be 91.3 percent per regeneration
cycle in the 1972 pilot runs.
The aluminum mass balances for each of the Batch C regenera-
tions are summarized in Table 11. Aluminum losses ranged
from 2.8 percent to 21.4 percent per cycle and averaged 8.7
percent during this series of runs. Percentage loss was
calculated from the carbon loss equation presented earlier.
As can be seen from Figure 75, the aluminum/carbon ratio of
the regenerated slurry built up far beyond 0.027, the level
necessary for proper flocculation in the treatment process.
Excess alum was accidently added following the first and third
regeneration cycles accounting for a major part of the Al/C
ratio increases following these points. Another source of
alum was makeup carbon to which was added 0.027 lbs A1 per
lb carbon.
During the pilot runs, acidification of the regenerated carbon-
aluminum oxide slurry was carried out in 500 gallon tanks.
Sulfuric acid was added to the slurry which was rapidly mixed
and then a sample was withdrawn for a pH determination.
Additional acid was then added if required. Using this method,
sulfuric acid requirements averaged 0.63 lbs H2SO4/ lb carbon.
The acid usage during each regeneration cycle is shown in
Table 5 and Figure 76. Acid requirements for regenerated
carbon are higher than for virgin carbon as is expected since
more acid soluble species exist in the regenerated slurry.
It appears that after one or two cycles, the acid requirement
is basically constant.
It was apparent that an excess of sulfuric acid was used in
the pilot studies because the method of acidification did
not lend itself to exact control of chemical dose. If an
acidification technique utilizing more exact pH control were
employed, it should be possible to substantially reduce the
acid requirement. Laboratory titrations of the regenerated
carbon slurries from the pilot operations indicate that it
should be possible to reduce the sulfuric acid usage to 0.5
lbs H2S04/lb carbon.
STACK GAS SAMPLING
A stack sampling program was undertaken by the New York
State Department of Environmental Conservation, Division of
Air Resources. All sampling was performed according to the
-121-
-------
TABLE 11
ALUMINUM MASS BALANCE SUMMARY
Regeneration Number 7 8 9 10 11 12
Aluminum at End of
Previous Regeneration
(lbs) 0 20.5 19.7 31.3 31.1 25.6
Virgin Aluminum Added at
End of Previous Regeneration
(lbs) 28.0 7.5 16.8 13.5 0 0
Physical Losses in
Treatment System
(lbs) 1.9 7.7 2.3 12.8 1.8 3.1
Aluminum Fed to
Regeneration Furnace
(lbs) 26.1 20.3 34.2 32.0 29.3 22.5
Aluminum Recovered at
End of Regeneration 20.5 19.7 31.3 31.1 25.6 18.3
(lbs)
% Aluminum Loss
in Furnace 21.4 3 .0 8 . 5 2 .8 12.6, 18.7
-122-
-------
0.08 —
0.06 —
o
i—i
t—
en
C_J
1
>-»
,—
ro
u>
-------
0
<_>
CO
0 .
CO
o
oo
0.
1971 OPERATIONS
1972 OPERATIONS
5
NUMBER OF REGENERATION CYCLES
FIGURE 76. ACID USAGE DURING REGENERATION STUDIES
-------
EPA sampling method published in the December 23, 1971
Federal Register. Tests were performed for nitrogen oxides,
sulfur oxides, water vapor, particulates, and trace metals.
Nitrogen oxide analyses were conducted on 5/9/72 using a
chemiluminescent analyzer. During the test the firebox and
bed temperatures were 1780-1810°F and 1200-1300°F, respec-
tively. The NO stack gas content was 48-50 ppm and the NO2
concentration was 1-3 ppm.
The stack gas opacity was estimated to be 10-20 percent.
This was due to a large amount of steam and powdered carbon
in the stack gas at the time.
Samples for SO2 analysis were collected on 6/13/72. No SO2
was detected in the samples.
Stack samples for particulates, moisture content and trace
metals were collected 6/12-13/72. The results are presented
in Table 12. Particulate emissions are of major concern not
only because of the air pollution potential but because of
the loss of regenerated carbon which they represent. The
relatively high particulate emissions in the stack gas
occurred as a consequence of the wrong size venturi being
supplied in the scrubber system. Particle capture is a
function of the velocity through the venturi throat. Since
the venturi throat was oversized in the pilot system,
particle capture was not highly efficient. As can be seen
from inspection of the data of Table 12, during sampling runs
1 & 2 when the pressure drop across the venturi was 3.5-4 inches
of water, 1.8 to 1.5 percent of the carbon product was lost.
However, during run 3 when the pressure drop was increased
to 8.5 inches of water, the carbon loss was correspondingly
decreased to 0.92 percent. Venturi manufacturers guarantee
that with a pressure drop of 20 inches of water, the maximum
particulate loss will be 0.07 8 lb/hr, which corresponds to
0.67 and 0.54 percent carbon losses for runs 1 and 3, respec-
tively. Since most of the furnace operation time was at the
conditions of run #1, the recoverable loss from the stack was
approximately 1 percent of the total carbon.
The carbon fraction in the stack gas particulates could not
be determined and therefore, the carbon/inert ratio in the
regenerated slurry was used to calculate stack carbon losses.
This will obviously underestimate the carbon loss since the
ash was mostly fine sand from the fluidized bed which is much
more readily captured by the cyclone type separator. Since
the ash content of the stack particulates should be much
lower than the recovered carbon, the amount of carbon re-
coverable with a properly designed scrubber system is esti-
mated conservatively.
-125-
-------
TABLE 12
STACK SAMPLING DATA
Sample Number
Bed Temperature, °F
1230
1250
1250
Firebox Temperature, °F
1830
1860
1840
Bed Velocity, ft/stc
1.7
1.7
2.2
Sludge Feed rate, lbs/hr
80
90
100
Volume Sampled, scf
43.3
56.2
66.0
Moisture Content, %
37.2
44.1
51.4
Total Particulates, lb/hr
0.213
0.195
0.189
Carbon, % of carbon feed
to furnace
Iron, lbs/hr
1.84 1.50
0.00402 0.00313
0.92
0.000725
-126-
-------
Thus installation of the proper sized venturi should sub-
stantially reduce carbon losses in the stack gases below the
measured levels.
CARBON REGENERATION STUDIES
The recovery of carbon from the regeneration furnace was
calculated on the basis of a fixed carbon mass balance over
the whole pilot plant. The average carbon recovery over the
first four cycles was 8 9 percent per cycle.
Table 13 shows the breakdown of the carbon mass balance during
the 1972 portion of the study. The carbon loss during the
period ranged from 1.3 percent to 22.3 percent. The average
loss per cycle calculated on an overall basis was 9.7 percent.
Percentage loss was calculated by the equation
Fixed Carbon Consumer - Physical Losses
% Loss = —i - Fixed Carbon Recovered
Fixed Carbon Consumer - Physical Losses
Overall loss per cycle was determined by dividing the percentage
loss determined as in the above equation by the total number
of regeneration cycles.
This method of loss calculation assumes that all physical
losses occurred during the sewage treatment portion of the
pilot plant operation, prior to regeneration. Since small
losses did occur after regeneration, the percentage loss
figures calculated for each cycle are slightly higher than
actual.
Carbon loss in the stack gas could not be routinely determined
and therefore, is not included in Table 13. As previously
discussed, the recoverable loss from the stack at typical
regeneration conditions was, found to be approximately 1 percent
of the carbon entering the furnace. The remainder of the
carbon loss is attributed to burning in the fluidized bed.
The carbon burned during regeneration is thus calculated
to be an average of 8.7 percent of the carbon entering the
furnace.
Determination of ash buildup as the carbon was repeatedly
regenerated was complicated by the fact that some sand
was carried over from the furnace and contributed to the
apparent "ash" content. Sand carryover from the furnace
was collected with the regenerated carbon in the scrubber
stream. A portion of this sand was separated from the
carbon slurry in the small settling chamber ahead of the
carbon storage tanks. During runs 11 and 12 the average
amount of sand collected in this manner was 0.27 pounds per
pound of carbon recovered. A significant quantity of sand
-127-
-------
TABLE 13
CARBON MASS BALANCE SUMMARY
Ifegeneration Number 7 8 9 10 11 12
Fixed Carbon at
End of Previous Regeneration,
(lbs) — 687 499 828 504 468
Fixed Virgin Carbon Added
at End of Previous Regeneration
(lbs) 787 131 498 157 0 0
Physical Losses in Fixed Carbon
in Treatment System
(lbs) " 73 227 55 *336 30 57
Fixed Carbon Fed to
Regeneration Furnace
(lbs) 714 591 942 649 474 411
Fixed Carbon
Recovered at End of
Regeneration Cycle
(lbs) 687 499 828 504 468 356
% Carbon Loss in Furnace 3.8 15.6 12.1 22.3 1.3 13.4
~Approximately 3 00 pounds purposely wasted to reduce carbon
inventory
-128-
-------
was not trapped in this manner and was cycled through the
system with the carbon. Sieve analysis of the sand collected
in the settling chamber is presented in Table 14. As would
be expected the data show the sand to be very fine.
An efficient method of separating the fine sand and ash
from the powdered activated carbon is desirable in order to
reduce the amount of mass carried through the treatment and
regeneration systems. A brief laboratory investigation was
conducted to study the feasibility of classification of the
carbon-inerts in the regeneration stream. Fifty milliter
samples of the slurry from the seventh regeneration (Batch C)
were centrifuged for five minutes at various speeds corres-
ponding to radial accelerations up to 1350 gravities. The
lower ten percent of the cake was then analyzed for inert
content with results as follows:
Radial acceleration, gravities
0
50
330
750
1350
Solids content of lower tenth, %
34.4
40.2
46.8
45.3
47.0
Inert content of lower tenth, %
52.2
70.3
78.5
78.3
77.7
The data show that at accelerations greater than 33 0 gravities
a cake is attainable which contains approximately 46 percent
solids of which 78 percent is inert material. If this fraction
were wasted for blowdown, less than 22 percent of the blowdown
would be carbon thus effecting a substantial savings in
carbon loss. In addition, if the slurry were acidified prior
to the inert carbon classification, the loss of alum in the
blowdown would be reduced to only that portion dissolved in
the wasted slurry.
Based on these results, it appears that classification would
be feasible and could lead to reductions in operating costs
by reducing the carbon and aluminum lost to blowdown.
A number of factors must be considered in selecting the rate
of blowdown. At a particular blowdown rate, the inert
material will eventually reach an equilibrium concentration.
This inert fraction will be cycled through the treatment
system, dewatered, and passed through the regeneration
system. The costs associated with these operations increase
in proportion to the quantity of inert material carried.
However, a high rate of blowdown which would minimize the
quantity of inerts would increase carbon losses. Therefore,
blowdown should be selected to minimize operating costs.
Based on the design and operating parameters presented in
Section VIII of this report, the effect of blowdown rate on
total treatment costs was calculated. Two cases were assumed:
no classification and classification to achieve a blowdown
-129-
-------
TABLE 14
SIEVE ANALYSIS OF FLUIDIZED BED SAND
COLLECTED WITH REGENERATED CARBON
Weight Fraction
Retained Retained
Mesh (g) (%)
40 0.73 1.66
50 .42 0.96
60 .45 1.02
70 .40 .91
100 5.20 11.83
150 7.84 17.84
200 20.22 46.01
Pan 8.69 19.77
43.95 100.00
-130-
-------
stream containing 78 percent inerts. The results are pre-
sented in Figure 77. It is evident that classification would
significantly reduce treatment costs at optimum blowdown
rates. In the case of no classification, optimum blowdown
would be 5 percent while with classification optimum blow-
down would be 6.5 percent.
-131-
-------
o NO CLASSIFICATION
A WITH CLASSIFICATION
6 8 10
BLOWDOWN RATE (%)
12
FIGURE 77. EFFECT OF BLOWDOWN RATE ON TREAT-
MENT COSTS FOR A 10 MGD PLANT
-132-
-------
SECTION VIII
DESIGN AND ECONOMIC CONSIDERATIONS
Based upon the data evolved in the Albany pilot plant
studies, preliminary estimates relating to design parameters
and the corresponding capital and operating costs for a 10
mgd treatment plant have been developed. A conservative
approach was adopted in the development of these estimates
and therefore, they should represent a "worst case" — the
maximum probable costs — for a 10 mgd plant utilizing
the powdered carbon treatment process described in this
report. For example, the tube settler loading rate utilized
in developing the capital cost estimates was the loading
rate actually used in the pilot studies. The pumping
capacity of the pilot system and the operational problems
associated with the pilot tube settler prohibited investi-
gation of any higher loading rates. It is probable that in
a commercial unit the loading rate could be increased sub-
stantially which would result in a reduction in capital
costs. Therefore, further process development should lead
to a downward adjustment of capital and operating cost
projections.
Several major questions with regard to design of a commercial
fluidized bed carbon regeneration furnace require resolution
before an optimum, minimum cost system can be designed. If
combustion could be carried out directly within the fluidized
bed chamber, the required unit area would be halved, the size
of the scrubber would be halved, and there would be no
requirement for a recycle blower. However, this method of
operation might result in greatly increased carbon losses
through combustion. At Albany, combustion was accomplished
in a chamber underneath the hearth and temperatures were
reduced to a practical limit by off gas recycle. This latter
approach should be feasible for a commercial unit. However,
there is a practical upper limit to the size of a brick
hearth which could be constructed. Moreover, the combustion
chamber temperature cannot be raised above an upper limit
of 2000°F.
A substantial savings in capital and operating costs can be
realized if off gases are not recycled. The pilot program
utilized recycle, but earlier work by Battelle-Columbus(12)
indicated it may not be required.
Other potential cost savings lie in the area of waste heat
utilization. Conceivably, waste heat from the regeneration
-133-
-------
furnace could be used to dry the sludge before regeneration.
This would result in capital cost savings since a smaller
size furnace would be required. In addition, fuel require-
ments would be reduced thus resulting in reduced operating
costs.
Classification of the carbon and inert fractions in the
regenerated stream prior to blowdown would reduce makeup
carbon and alum requirements below those utilized for cost
projections in the current analysis.
For purposes of this report, two sets of conditions (off
gas recycle and no recycle) have been assumed and capital and
operating costs have been calculated for each case. Costs
have been developed both for operation as municipal waste-
water and combined sewage treatment plants with two
configurations for each plant type.
SYSTEM DESIGN PARAMETERS
The design parameters used in developing the treatment system
capital cost estimates are given in Table 15.
SYSTEM CAPITAL COSTS
Capital costs are based on vendor information, published cost
data (12 ,13,14) ^ an{j
engineering estimates of the required
equipment sizes. Capital cost estimates for municipal waste-
water and combined sewage treatment are given in Table 16.
SYSTEM OPERATING PARAMETERS
Table 18 presents the system operating costs based on the
parameters presented in Table 17.
TOTAL SYSTEM COSTS
Total system costs include the operating costs given in
Table 18 plus amortization of the capital costs presented
in Table 16. Amortization costs were calculated at 6 percent
over 25 years. The total system costs are presented in
Table 19.
-134-
-------
TABLE 15
SYSTEM DESIGN PARAMETERS
Treatment System
Carbon Contact
Time at pH 4
Time at pH 7
Flocculation
Velocity Gradient
Time
Tube Settler Loading Rate
Filter
Length of Filter Run
Loading Rate
Chlorine Contact Time
Chemical Storage Capacity
Sludge Storage
Regeneration System
10 minutes
5 minutes
75 fps/ft
10 minutes
2880 gpd/ft^
12 hours
5 gpm/ft^
10 minutes
12 hours
1 day
With gas Without
recycle gas recycle
Combustion Chamber Temperature, °F
2000
2000
Bed Temperature, °F
1250
1250
Fluidizing Gas Velocity, ft/sec
1.3
1.0
Heat Requirement, million BTU/hr
30.7
24.5
Bed Diameter, 2 units, each, ft
21.7
19.8
Blowdown, %
5
5
-135-
-------
TABLE 16
CAPITAL COST ESTIMATES FOR 10 MGD MUNICIPAL
WASTEWATER AND STORM WATER TREATMENT PLANTS
Municipal Plant Storm Water Plant
Item Installed Costs, $ Installed Costs, $
Screens, grit chamber,
overflow 10,000 10,000
Reaction vessels 27,000 27,000
Chemical storage tanks 32,000 32,000
Carbon slurry tanks 40,000 40,000
Sludge storage 17,800 17,800
Pumps 35,000 35,000
Agitators 45,700 45,700
Flocculation, sedimentation 475,000 475,000
Filtration 300,000
Chlorination 14,700 14,700
Centrifuge 80,000 80,000
Sludge pumps 53,700 53,700
Regeneration facility
With gas recycle 1,104,000
Without gas recycle 794 ,800 794 ,800
Subtotals
With gas recycle 2,234,900
Without gas recycle 1,925,700 1,625,700
Without regeneration 830,900
Total Capital Costs*
With gas recycle 2,462,390
Without gas recycle 2,121,270 1,791,270
Without regeneration 966,990
~Contingencies 10%
Land, 1.5 acres @ $2000/acre
-136-
-------
TABLE 17
SYSTEM OPERATING PARAMETERS
ITEM
Treatment System
Carbon dose, mg/1
Alum dose, mg/1
Polyelectrolyte dose, mg/1
Lime dose, mg/1
Sulfuric acid, #/# carbon
Sludge dewatering polyelectrolyte
dose, #/ton dry solids
Regeneration System
Carbon recovery, %
Alum recovery, %
Blowdown, %
Carbon feed rate, #/hr
Sludge feed rate, #/hr
Sludge solids content, %
Sludge inerts content
% on dry basis
Fuel, C/MBTU
Power, C/kwhr
Blowdown disposal cost, C/lb solids
MUNICIPAL
PLANT
600
200
2.0
150
0.5
STORM WATER PLANT
91
91
5
2080
21,800
22
60
25
0.7
0.4
With Regeneration Without Regeneration
400
200
2.0
150
0.75
91
91
5
1390
21,800
22
60
25
0.7
0.4
400
200
2.0
0
0
100
25
0.7
0.4
-------
TABLE 18
OPERATING COST ESTIMATES FOR 10 MGD MUNICIPAL
WASTEWATER AND STORM WATER TREATMENT PLANTS
ITEM
Sulfuric acid
Lime
Make-up alum
Polyelectrolyte1
Make-up carbon
2
Carbon regeneration
With gas recycle
Without gas recycle
Blowdown Disposal
Chlorination
Power
Labor, 60 hr/day @ $4.00/hr
TOTAL
Regeneration with gas recycle
Regeneration without gas recycle
Without regeneration
1 Includes dewatering dose
2
Includes fuel costs
3 4 0 hr/day
MUNICIPAL
With
Regeneration
C/1000 gal
4.3
0.8
0.7
2.9
6.3
3.5
2.9
0.2
0.2
0.2
2.4
21.5
20.9
STORM WATER
With
Regeneration
C/1000 gal
4.3
0.8
0.7
2.9
4.2
2.9
0.2
0.2
0.2
2.4
18.8
Without
Regeneration
C/1000 gal
0
0
5.0
2.9
30.1
0
0
0
0.2
0.2
1. 6 3
42.9
-------
TABLE 19
]
M
u>
-------
SECTION IX
ACKNOWLEDGMENTS
The authors wish to extend their gratitude to the many
people whose assistance was necessary to the timely and
successful completion of this program. A. T. Brix, J. J.
Dorgan, and J. Green made significant contributions to the
pilot plant design. R. G. Parkhurst and M. J. Mason
contributed significantly of their own time during the
pilot plant shakedown at Richland. Battelle's operating crew
at Albany, J. A. Coates, M. J. Mason, R. G. Swank, and R. G.
Upchurch invested many long hours in keeping the plant
operating and contributed in large part to the success of
the Albany demonstration. Other Battelle personnel who
contributed advice, suggestions, time, and moral support
throughout the course of this program include B. W. Mercer
and D. E. Olesen. Former Battelle personnel who also fall
in this category are Dr. C. J. Touhill and Mr. G. L. Culp.
Mr. Danforth Bodien and Mr. Frank Condon of the Environ-
mental Protection Agency provided support and assistance
throughout the course of this program.
Nichols Engineering and Research Corporation designed and
supplied the fluidized bed pilot regeneration furnace used
in the study. Mr. Charles von Dreusche, Jr. provided much
assistance in overcoming problems with the unit and
developed much of the furnace cost data used in this report.
Mr. George Nowowiejski, also of Nichols, invested much time
and effort during the shakedown of the furnace.
Dr. Leo Hetling of the New York State Department of Environ-
mental Conservation provided assistance in locating and
obtaining the Albany site. He also provided help in numerous
ways throughout the operations at Albany and has initiated
studies by the New York State Department of Environmental
Conservation to characterize the Island Creek Drainage Area.
Mr. Walter Connley of Neptune Microfloc, Inc., provided
cost information used in this report.
-141-
-------
SECTION X
REFERENCES
1. "Problems of Combined Sewer Facilities and Overflows."
Federal Water Pollution Control Administration, U. S.
Department of Interior, WP-20-11. 1967.
2. Weibel, S. R., Anderson, R. J., and Woodward, R. L.
"Urban Land Runoff as a Factor in Stream Pollution."
J. Water Pollution Control Federation, Vol. 36, pp.
914-924. I9&T.
3. Burns, R. J., Krawczyk, D. F., and Harlow, G. L.
"Chemical and Physical Comparison of Combined and
Separate Sewer Discharges." J. Water Pollution Control
Federation, Vol. 40, pp. 112-126. 1968.
4. "Water Pollution Aspects of Urban Runoff." Federal Water
Pollution Control Administration, U. S. Department of
Interior, WP-20-15. 1969.
5. Shuckrow, A. J., Dawson, G. W., and Olesen, D. E. "Treatment
of Raw and Combined Sewage." Water and Sewage Works,
pp. 104-111. April 1971.
6. Rizzo, J. L. and Schade, R. E. "Secondary Treatment with
Granular Activated Carbon." Water and Sewage Works,
p. 307. 1969.
7. Weber, W., Hopkins, C. B. and Bloom, R. "Physiochemical
Treatment of Wastewater." Journal Water Pollution Control
Federation, p. 83. 1970.
8. Zuckerman, M. M. and Molof, A. H. "High Quality Reuse
Water by Chemical-Physical Wastewater Treatment." Journal
Water Pollution Control Federation, p. 437. 1970.
9. Shuckrow, A. J., et al_. "A Pilot Study of Physical-
Chemical Treatment of the Raw Wastewater at the Westerly
Plant in Cleveland, Ohio." A paper presented at the
International Association on Water Pollution Research
Workshop, Vienna, Austria. September 1971.
10. "Standard Methods for the Examination of Water and
Wastewater." Twelfth and Thirteenth Editions, APAA,
AWWA, WPCF. 1967 and 1971.
-143-
-------
11. Berg, E. L., et al. "Thermal Regeneration of Spent
Powdered Carbon Using Fluidized-Bed and Transport Reac-
tors." Chemical Engineering Progress Symposium Series
No. 107, Vol. 67. 1971.
12. "The Development of A Fluidized Bed Technique for the
Regeneration of Powdered Activated Carbon." Water
Pollution Control Research Series, ORD-17070FBD03/70,
Federal Water Quality Administration. 1970.
13. Culp, R. L. and G. L. Culp. "Advanced Wastewater
Treatment." Van Nostrand Reinhold, 1971.
14. Guthrie, K. M. "Capital Cost Estimating." Chem. Engr.,
Vol. 76, No. 6. March 24, 1969.
-144-
-------
SECTION XI
APPENDIX A
LABORATORY STUDIES
An extensive laboratory program to develop and establish
the feasibility of the physical-chemical treatment process
described in this report was carried out from July 1969
through March 1970. The results of this laboratory program
are described below.
GENERAL
Influent to the Richland, Washington, sewage treatment plant
was used in all of the laboratory work. At the outset of the
investigation, an extensive analytical study to characterize
Richland sewage was carried out in order to investigate the
suitability of this influent for use in the laboratory
studies on the treatment process. Grab samples were obtained
at various times of the day over a 3-1/2 day period and were
analyzed for various constituents in accordance with the
procedures outlined in "Standard Methods. Ca and Mg
concentrations were determined by atomic absorption tech-
niques. The results of the characterization are presented
in Table A-l.
During the course of the research, a Beckman Model 915 Total
Organic Carbon Analyzer was obtained and when this instru-
ment became operational, TOC measurements were substituted
for COD determinations.
A bench scale continuous flow system was constructed early
in the program. As the work progressed, this system was
modified several times to incorporate features such as in-
line mixing and pH control. Two basic types of clarifiers
were employed in this system: a tube settler and an upflow
clarifier. Upon selection of tube settlers for the pilot
plant,, all further bench scale work was conducted with the
laboratory tube settler. Figure A-l is a schematic diagram
of the final laboratory system.
CARBON STUDIES
Aqua Nuchar A was selected for use in initial studies based
on its relatively low cost and satisfactory performance in
prior studies. Several sets of experiments were conducted
to examine the sorption characteristics of the powdered
carbon. In order to accomplish th5.s, 300 ml aliquots of
fresh Richland influent were placed in 8 oz polyethylene
-145-
-------
TABLE A-1
COMPOSITION OF RICHLAND SEWAGE
Time of
Day
0700
1130
1620
1930
2340
Conductivity
(ymhos/cm)
pH
Temp °C
Turb (JTU)
Alkalinity
(mg/1 as CaCO^)
Ca (mg/1)
Mg (mg/1)
COD (mg/1)
NH4-N (mg/1)
N02-N (mg/1)
950-620
750-6
7.68-7.47
7.56-6
22-20
21-2
28-10
19-2
252-199
239-6
43-36
40-4
13-10
11.5-4
160-130
142-6
20.7-10.8
15.3-6
.22-.01
.067-5
1600-820
1080-4
7/72-7.60
7.69-4
26-25
25.5-2
105-68
87-2
325-297
310-4
45-38
41-3
15-10
12-3
544-436
503-4
44.8-32.7
39.4-4
.025-.01
.019-4
460-400
430-3
7.48-7.
7.42-3
26
26-2
130-90
110-2
264-224
248-3
41-38
40-3
12-9
10-3
525-267
351-3
19 .9-14
17-3
.01-0
.003-3
36
532-460
490-3
7.36-7.30
7.33-3
24-22
23-2
150-55
103-2
254-229
248-3
44-36
40-3
12-8
11-3
525-303
401-3
17.5-13.2
15.5-3
.025-0
.012-3
590-540
567-3
7.38-7.31
7.34-2
24-23
23.5-2
160-58
109-2
277-253
267-3
45-43
44-2
12
12-2
450-400
431-3
21.8-18.3
20.1-3
.025-0
.012-3
(continued)
-------
TABLE A-1 (continued)
COMPOSITION OP RICHLAND SEWAGE
Time of
Day
0700
1130
1620
1930
2340
N03-N (mg/1)
Org. N
Sol PO.
(mg/1)
TS (mg/1)
TVS (mg/1)
SS (mg/1)
VSS (mg/1)
Settleable Solids
(ml/1)
Settleable Solids
(mg/1)
3.2-0
.66-6
17.1-6.7
11.6-6
41-20
30.1-6
1080-510
705-6
261-168
204-6
124-51
93-5
93-5
92-30
60-5
4.0-3.5
3.7-3
145-82
114-2
.15-.10
. 11-4
31.7-12.
21-4
34-27.5
31-4
894-694
764-4
474-329
400-4
320-228
274-4
274-4
208-176
188-4
15-11.5
13-3
315-148
232-2
.02-0
.007-3
12.6-10 .2
11.8-3
50-26 .5
38.8-3
735-636
685-2
303-234
294-3
286-198
219
219-3
164-122
139-3
9-5
7.1-3
180-52
116-2
.2-0
.13-3
13.9-9.4
11.6-3
34.5-30.5
33.3-3
724-672
701-3
454-319
375-3
315-168
232-3
182-103
151-3
17-5.5
9.8-3
278-112
204-3
.2-0
.1-3
12.5-11.7
12.2-3
37.5-31.5
35.1-3
770-685
721-3
450-363
418-3
275-165
211-3
203-112
173-3
8-7
7.3-3
191-164
178-2
NOTE: Numbers for each time period represent:
/
\high value - low value
average - no. of samples
}
-------
LIME SLURRY
CARBON-ALUM FEED
POL
ECTROLYTE
PUM
CONTACT
TANK
TUBE
SETTLER
METER
Ttocculator
PRODUCT
SEWAGE
FEED
TANK
SLUDGE WITHDRAWAL
FIGURE A-l, BENCH SCALE SYSTEM
-------
bottles and a quantity of powdered carbon was added to each
bottle. The set of samples was then placed on a mechanical
shaker and was agitated overnight. This was believed to be
a sufficient period in which to establish equilibrium con-
ditions. Each carbon-sewage mixture was then filtered
through a 0.4 5 p membrane filter and the COD of the filtrate
was determined. Figure A-2 contains the results for seven
different sewage samples. It appears that sorption of COD is
essentially complete at a carbon concentration of 800-1000 mg/1.
Increasing the carbon concentration beyond this point has little
or no effect on the residual COD. Two parallel sets were run in
one instance in order to determine if the presence of bentonite
had any effect on the sorption equilibrium. As seen from in-
spection of Figure A-2, 300 mg/1 of bentonite had no noticeable
effect on the equilibrium solution concentration of COD.
Sewage samples were contacted with 1000 mg/1 of Aqua Nuchar A
for various time intervals and were then centrifuged and finally
filtered through 0.45 y filters. COD analyses of the filtrate
indicated that sorption was essentially complete after a con-
tact period of ten minutes. Subsequent to these findings, a
carbon concentration of 1000 mg/1 and a contact time of ten
minutes were adopted for experimental use.
Subsequent tests, run with 500 ml sewage aliquots and one
hour contact times, suggested that TOC removal is virtually
complete at a carbon dose of 500-600 mg/1. Data from these
tests are presented in Figure A-3. In view of these findings,
another series of tests to examine required detention time
at the lower carbon dose was conducted. As can be seen from
inspection of the data in Figure A-4, a ten minute detention
time should be adequate even at this reduced carbon dose.
Subsequently, samples of fifteen different commercial grade
powdered carbons were evaluated in a series of tests. Earlier
comparisons showed little or no correlation between methylene
blue adsorption and organic carbon removal. Hence, tests were
based on residual organic carbon concentrations after both
contact tests and jar tests.
Contact tests were conducted using 500 ml sewage samples
dosed to 1000 mg C/l. Duplicate samples were prepared, one
of which was shaken for an hour and the second for 21 hours.
The samples were then filtered through 0.45 p membrane
filters and analyzed for total organic carbon. This pro-
cedure was repeated for seven of the carbons with a sewage
sample obtained on a different day. The results of these
runs are given in Table A-2.
Based on the results of the contact tests, selected carbon
types from the different companies were used in jar tests.
A carbon dose of 1000 mg/1 was used with 3 50 mg/1 alum
-149-
-------
x INF A SET #1
v INF A SET #2
O INF B
A INF C
A INF C*
~ INF D
B INF E
• INF F
CONTACT TIME = 18 HOURS
*WITH 300 mg/J£ BENTONITE
ADDED
10 -
h
-X-
_l
1000 2000 3000 4000
CARBON CONCENTRATION (mg/S>)
5000
FIGURE A-2. EFFECT OF CARBON CONCENTRATION
ON EQUILIBRIUM COD
-150-
-------
20
08
96
84
72
60
48
36
24
12
0
F
_ A
O INF 1
~ INF 2
A INF 3
CONTACT TIME = 1 HOUR
0 400 800 1200 1600
CARBON CONCENTRATION (mg/Z)
IGURE A-3. EFFECT OF CARBON CONCENTRATION
ON EQUILIBRIUM TOC
-151-
-------
120
O 600 mg/ Jt AQUA NUCHAR A
A 1000 mg/Ji, AQUA NUCHAR A
100
E
C
o
o
30
25
20
15
10
5
0
CONTACT TIME (MIN)
FIGURE A-4. EFFECT OF CONTACT TIME ON TOC REMOVAL
-152-
-------
TABLE A-2
COMPARISON OF VARIOUS CARBONS - CONTACT TESTS
Sewage A Sewage B Bulk
TOC (mg/1) TOC (mg/1) TOC (mg/1) Cost
Sample
1
hr contact
21 hr contact
1 hr contact
(«/#)
Nuchar
CEEN
7
7
7.5
14
Darco
S51
8.5
12
14.5
13
Nuchar
WAN
8.5
9
9.5
10.5
Nuchar
C190A
9
7
10.5
15.5
Nuchar
C115N
10
10.5
—
15.5
Nuchar
C115A
10
9.5
12.5
13.5
Nuchar
Aqua
10.5
7
10.5
9
Nuchar
C190N
11
12
—
15.5
Pittsburgh GW
11.5
7
10.5
15.5
Nuchar
CEEA
13 .5
8
—
14
Nor it
FQA
15
9.5
—
12.5
Darco
GFP
16
10
—
10.5
Whitco
517
16.5
11
—
—
Darco
KB
16.5
11
—
29
Nor it
F
17
12
_ _
11.5
Sewage A Soluble TOC = 3 5.5 mg/1
Sewage B Soluble TOC = 47.5 mg/1
Carbon Dose = 1000 mg/1
-153-
-------
added after 5 minutes contact time and 10 mg/1 of Magnifloc
98 5 N added after an additional 5 minutes. Samples were
then flocculated, allowed to settle, and the unfiltered
supernatant was analyzed for organic carbon as before.
Results for two different sewage samples are given in
Table A-3, together with several qualitative observations
on wettability and floe characteristics.
Based on these findings, the decision was made to continue
using Aqua Nuchar A since it appeared to offer the best
benefit/cost ratio.
BENTONITE PROCESS DEVELOPMENT
Preliminary investigations prior to the start of the project
indicated that bentonite clay, in conjunction with a poly-
electrolyte, could effectively coagulate powdered carbon.
Based on these preproject findings, Magnifloc 985 N, a high
molecular weight, nonionic, polyacrylamide was selected for
use in the early work.
A series of jar tests was conducted to determine the effect
of bentonite concentration on the flocculation-sedimentation
characteristics of the system with the results given in
Figure A-5. When no bentonite was added to the system, large
quantities of carbon remained in suspension at the end of the
five minute settling period. Although it is not obvious
from Figure A-5, the flocculation-sedimentation character-
istics of the system noticeably improved up to a bentonite
concentration of 3 00 mg/1. However, the five minute settling
period was sufficient to effect good phase separation at a
bentonite dose of 100 mg/1. This result was also observed
when the polyelectrolyte dose was reduced to 5 mg/1. Simul-
taneous addition of bentonite and carbon to the system had
no detectable effect on COD removal or flocculation. The pH
of the system was varied from 5.5-8.1 and no significant
change in system performance was observed. Turbidity of
the settled effluent consistently ranged from 5-13 JTU.
Upon filtration of these effluent samples, the residue on
the membrane filter appeared to be clay rather than carbon.
These promising results prompted the initiation of an investi-
gation of the effectiveness of the process in a continuous
flow system. Table A-4 summarizes the data of five runs with
the tube settler system. Carbon, bentonite and polyelectro-
lyte concentrations were set at 1000 mg/1, 3 00 mg/1, and
10 mg/1, respectively, in the initial runs to study the effect
of variable flow rates. It was realized that these concentra-
tions represented an excess of polyelectrolyte and probably
bentonite as well.
It appeared that the bentonite dose could be reduced to 100-
200 mg/1. In general, the bentonite process seemed to produce
-154-
-------
TABLE A-3
COMPARISON OF VARIOUS CARBONS - JAR TESTS
Sample
TOC
(mg/1)
TOC
(mg/1)
Floe
Wettability
Whitco 517
9.5
14.5
Good
Good
Pittsburgh GW
9
20.5
Good
Average
Darco S51
11.5
16 .5
Good
Good
Nuchar CEEN
12
35.5
Fair
Average
Nuchar Aqua
15.5
32.5
Good
Average
Nuchar WAN
11.5
17.5
Fairly Good
Average
Influent
47.5
75.5
--
—
Carbon Dose = 1000 rag/1
-155-
-------
T
o
©¦
O BENTON IT E ADDED
FOLLOWING 10 MINUTE
CARBON CONTACT
a CARBON AND BENTONITE
ADDED SIMULTANEOUSLY
CARBON CO N C. = 1000 mgfl
MAG INIFLOC 985 =10 mg/£
UNFILTERED SAMPLES
_L
100 200 300
BENTONITE CONCENTRATION {mq/l'.
400
FIGURE A-5. EFFECT OF BENTONITE CONCENTRATION
ON EFFLUENT QUALITY
-156-
-------
TABLE A-4
BENCH SCALE SYSTEM OPERATIONAL DATA - TUBE SETTLER
Date
Influent
Influent
Flow
Carbon
Bentonite
Polyelectrolyte
Effluent
Effluent
COD
COD
S.S.
Rate*
Cone.
Cone.
Cone.
S.S.
COD
Removal
(mg/1)
(mg/1)
(ml/min)
(mg/1)
(mg/1)
(mg/1)
(mg/1)
mg/1)
(%)
8/25/69
84
71
195
1000
300
10
18
18
6 6
195
1000
300
10
20
12
77
100
1000
300
10
18
12
77
100
1000
300
10
16
11
79
48
1000
300
10
18
13
75
48
1000
300
10
14,
14
73
8/26/69
169
195
100
1000
300
10
44
46
73
150
1000
300
10
18
38
77
200
1000
300
10
50
56
67
200
1000
300
10
50
42
75
285
1000
300
10
45
46
73
285
1000
300
6.7
30
46
73
8/27/69
184
134
300
1000
300
0 . 5
290
96
48
300
1000
300
2.5
57
37
80
300
1000
300
5.0
35
20
89
300
1000
300
7 . 5
43
23
88
100
1000
300
0.75
51
18
90
100
1000
300
3.75
43
17
90
100
1000
300
7 . 5
35
11
94
8/28/69
202
192
200
1000
0
10
117
174
14
200
1000
50
10
51
70
65
200
1000
100
10
51
49
76
200
1000
150
10
44
57
72
200
1000
200
10
26
45
78
200
1000
300
10
28
28
86
8/29/69
400
239
200
1000
300
5
13
31
92
200
1000
100
5
11
42
89
300
1000
100
5
—
24
94
~Flow rates are 1
isted in the
order in
which they were var
ied during a particular run.
A flow
rate
of 100
ml/minute
represents
a loading
rate of
4.84 gpm/ft
of tube surface
.
-------
a higher quality effluent in the continuous flow system than
in the jar tests. Some further increase in flow rate may have
been possible, but it was believed that the 3 00 ml/min flow
was near the limit of the system without carbon carryover.
This limitation was due to the sizing of the tube settler.
The one inch diameter tube settler was a standard laboratory
model purchased from Neptune MicroFloc Inc. At a system
flow rate of 100 ml/min, the throughput rate of the tube
settler was 4.84 gpm/ft^.
An additional six runs were conducted with an upflow clarifier
system. Flow rates were varied from 300-1500 ml/min, giving
a range of system detention time from 62-12 minutes. Chemical
doses were held constant at 300 mg/1 bentonite, 1000 mg/1
carbon, and 10 mg/1 Magnifloc 98 5 throughout the runs. Results
of these runs are summarized in Table A-5. Turbidity of the
effluent ranged from 3.1 JTU at a flow rate of 3 00 ml/min,
to 6.2 JTU at a flow rate of 1500 ml/min. As described later
in this section, carbon-bentonite sludges were regenerated by
Battelle-Columbus and by FMC Corporation. Jar test evaluations
of the regenerated products revealed that the coagulating
ability of the bentonite was destroyed to a considerable
degree in the regeneration process. Apparently, thermal
treatment causes the bentonite structure to collapse and
this alteration renders it inoperable as a coagulant.
It was found necessary to add a full 300 mg/1 of fresh
bentonite to the regenerated carbon-bentonite mixtures in
order to achieve good coagulation.
Even with the addition of new bentonite, the effluent had
poor turbidity characteristics in comparison to the effluent
produced using virgin carbon and bentonite. It is postu-
lated that this was caused by the colloidal suspension of
spent bentonite ash. Good clarification requires the addition
of other flocculant aids.
A screening study was conducted in an attempt to find a
suitable substitute for bentonite. Materials which were
investigated and rejected include aluminum silicate, syn-
thetic zeolite, diatomaceous earth, powdered silica, and
asbestos.
Bentonite does not appear to offer much promise if thermal
regeneration is to be utilized. It would probably be
possible to recycle the carbon-bentonite mixture several
times. However, buildup of the inorganic content will be
severe, and frequent wasting of large quantities of carbon-
bentonite would probably be required.
-158-
-------
TABLE A-5
UPFLOW CLARIFIER OPERATING DATA
Flow Overflow Number Influent COD Effluent COD COD Removal
Rate Rate of (mg/1) (mg/1) %
(ml/min) (gpm/ft2) runs Range Avg Range Avg Range Avg
300 0.189 5 213-259 206 7-33 17 85-94 92
350 0.220 1 247 10-16 13 93-96 95
600 0.378 5 213-259 209 9-46 19 82-96 91
700 0.441 1 247 41-45 43 82-84 83
1500 0.945 3 213-259 196 22-38 28 81-87 84
-------
ALUM PROCESS PRELIMINARY INVESTIGATIONS
Magnifloc 985 N was used throughout these initial investiga-
tions as was a carbon dose of 1000 mg/1. The alum [Al2(S04)3
•I8H2O] dose was varied in a series of jar tests. From in-
spection of the data of Figure A-6 it is apparent that adequate
turbidity removal was achieved at an alum dose of 15 0 mg/1.
However, objectionable quantities of powdered carbon remained
in the effluent and good flocculation was not achieved until
the alum dose had been increased to 3 50 mg/1. Additional
jar tests at this alum concentration and at polyelectrolyte
concentrations of 5-10 mg/1 indicated that COD removals of
>90% could be obtained consistently. It was observed that
coagulation was impaired in some instances if alum and carbon
were added simultaneously. A carbon contact time of at least
five minutes prior to alum addition was required in order to
insure consistently good floe formation.
Bench scale experiments with the alum process were then
conducted. As indicated by the jar tests, a definite
carbon contact time of at least five minutes was required
before addition of the alum for coagulation to proceed
normally. Data from runs with the upflow clarifier and the
tube settler are presented in Tables A-6 and A-7.
In the first two of the runs with the tube settler, numerous
mechanical difficulties with pumps, mixers, etc., arose. Due
to the low flow rates employed, the system was slow to reach
equilibrium. Therefore, the data reported for these first
two runs does not represent optimum system performance but
does indicate that even during minor upsets, a reasonable
degree of treatment can be achieved. Tube settler performance
was similar to that observed for the bentonite system:
settling occurred primarily in the inlet chamber. The floe
formed in this process was not as dense nor as tough as that
formed with bentonite. Therefore, it was necessary to exer-
cise greater care to prevent breakup of the floe and redisper-
sion of the carbon after polyelectrolyte addition.
CARBON REGENERATION
An arrangement was made with Battelle-Columbus (BCL) and with
FMC Corporation to regenerate a number of samples in their
respective laboratory carbon regeneration systems. The initial
carbon-bentonite sludge sample sent to FMC was generated in
the laboratory bench scale unit. All of the sludge samples
sent to BCL and the second set of FMC samples were generated
by treating a larger quantity of sewage on a batch basis.
Each set of sludge samples was generated by treating six
hundred gallons of sewage in batches of 50 gallons. Half
was treated using the bentonite process and half by the alum
process. After initial separation, the sludge was allowed to
-160-
-------
4C
O TURBIDITY
A COD
35
CARBON DOSE =
MAGNIFLOC 985
1000 mg/Jl
= 10 mg/1
30
25
20
LU
Q
O
° 15
10
5
0
500
400
300
200
100
ALUM DOSE (mg/A)
FIGURE A-6. EFFECT OF ALUM DOSE ON COD
AND TURBIDITY REMOVAL
-161-
-------
TABLE A-6
UPFLOW CLARIFIER SYSTEM - ALUM PROCESS PERFORMANCE
Influent
Flow
Overflow
Alum*
Polyelectrolyte
Effluent
COD
COD
Rate
Rate _
Cone.
Concentration
SS
Turbidity
COD
Removal
(mg/1)
(ml/min)
(gpm/ft )
(mg/1)
(mg/1)
(mg/1)
(JTU)
(mg/1)
%
180
600
0 .378
400
10
13
2
12
93.4
600
0 .378
350
10
19
1. 5
8
95. 6
300
0.189
350
10
13
111
4
97. 5
300
0 .189
350
5
16
1
13
93
324
600
0 . 378
350
4
36
5 .1
38
88
300
0 .189
350
5
18
2.5
26
92
*Alum - Al2 (S04)3.18H20
Carbon Dose = 1000 mg/1
-------
TABLE A-7
TUBE SETTLER SYSTEM - ALUM PROCESS
Influent
Flow Rate
(ml/min)
Tube
Loading Rate
(gpm/ft2)
Effluent
Removal
COD
(mg/1)
TOC
(mg/1)
S.S.
(mg/1)
Turbidity
(JTU)
COD
(mg/1)
TOC
(mg/1)
COD
(%)
TOC
<%)
280
46
150
7.26
14
2.1
70
8.0
76
81.5
200
9.68
9
1.6
36
8
87
82.6
188
73
150
7.26
11
3.0
28
4.5
85
93 .8
200
9.68
16
2.5
53
-
72
--
>800
86
150
7.26
20
2.0
53
11.5
>93
86.6
250
12.10
9
1.5
57
-
>93
—
186
66
250
12.10
13
3.0
32
9
83
86.3
300
14.52
13
4.0
19
21
90
46
Carbon dose - 1000 mg/1
Alum dose = 350 mg/1
Polelectrolyte dose = 10 mg/1
-------
concentrate for several hours, the supernatant liquid was
poured off, and this procedure was repeated a second time.
In most cases, the dewatered sludge was oven dried before
shipment. Physical losses in handling these small quantities
of carbon were significant in both cases and therefore it is
difficult to estimate the processing losses involved in
either system. In fact, these handling losses were so high
that large quantities of makeup carbon had to be added after
each regeneration cycle. Therefore, an exact picture of
several cycles of use and reuse could not be obtained.
Carbon was followed through three regeneration cycles for
the Battelle-Columbus process. The carbon composition of the
sludge regenerated in each cycle was as follows:
First Cycle - Bentonite Process
Alum Process
Second Cycle- Bentonite Process
Alum Process
Third Cycle - Bentonite Process
Alum Process
100% Virgin carbon
100% Virgin carbon
75% First cycle regenerated
carbon
25% Virgin carbon
50% First cycle regenerated
carbon
50% Virgin carbon
66% Second cycle
regenerated carbon
55% Second cycle
regenerated carbon
45% Virgin carbon
Samples of the various regenerated carbon mixtures were analyzed
for carbon, water, and ash content with the results given in
Table A-8. It should be noted that the ash content reading
includes the alum and bentonite residue except in the case of
the acid extracted carbon. These data show how the bentonite
content of the sludge from the bentonite process increases
with each cycle. An ash content of 4 4 percent was measured
after the third regeneration. On the other hand, the compo-
sition of the acid extracted carbon-alum mixture is comparable
for the second and third cycles. Definite conclusions con-
cerning physical losses and ash buildup in the alum process
could not be drawn on the basis of the laboratory studies.
Carbon contact tests were employed to determine the capacity
recovery of the carbon after regeneration. The aluminum-
containing mixtures were slurried overnight at a pH of 0.5
to dissolve all soluble aluminum. The dried powders, re-
generated bentonite-carbon, and new Aqua Nuchar were then
measured out at 0.5, 0.75, 1.0 and 1.5 gm/liter for contact
tests. Samples were shaken for one hour, filtered through
-164-
-------
TABLE A-8
ANALYSIS OF REGENERATED CARBON MIXTURES
Carbon Content
(%)
Water Content
(%)
Ash Content
(%)
Virgin Aqua Nuchar
1
84.1
9.0
2.4
2
.86.6
4.5
2.3
3
83.6
10 .8
1.7
BCLA
1st Cycle
81.6
6.8
12.5
2nd Cycle
75.2
7.2
17.5
3rd Cycle
73.0
6.1
19.8
ABCLA
1st Cycle
—
—
—
2nd Cycle
68.3
13.5
8.2
3rd Cycle
67.6
16.2
7.4
BCLB
1st Cycle
69.7
5.7
25. 2
2nd Cycle
60.0
1.8
37. 6
3rd Cycle
53.5
4.5
44.2
FMCA
1st Cycle
66. 6
2.7
29.3
AFMCA
1st Cycle
60.7
11.7
22.4
FMCB
Batch Scale
65.1
3.6
32.1
1st Cycle
81.6
2.3
15.5
BCLB - Regenerated carbon-bentonite mixture - Battelle-Columbus
process
BCLA - Regenerated carbon-alum mixture - Battelle-Columbus process
ABCLA - Acid extracted regenerated carbon-alum mixture - Battelle-
Columbus process
FMCB - Regenerated carbon-bentonite mixture - FMC process
FMCA - Regenerated carbon-alum mixture - FMC process
AFMCA - Acid extracted regenerated carbon-alum mixture - FMC process
-165-
-------
0.4 5 p membrane filters, and analyzed for organic carbon.
The data obtained are presented in Figures A-7, A-8, and
A-9. These data indicate that capacity recovery lies some-
where between 90-100 percent.
A series of jar tests was run on sewage using virgin and
the various regenerated carbons. Analyses performed on the
settled supernatants are presented in Table A-9. The nutrient
results were predictable except for the apparent removal of
NO^ which cannot be explained. Acid treatment of the re-
covered carbon-alum mixture produces a carbon which is
apparently as good as or better than the fresh product.
ALUM RECOVERY
An investigation was initiated to study the possibilities of
aluminum recovery from the regenerated carbon-alum mixture.
It is known that aluminum hydroxide goes through the conversion
m /r,m 500°C _ 1000°C
A1 (OH) , -+ y-Al.,0, n ^
3 2 3 ¦+¦ ct-A^O^
The y-oxide readily dissolves in H2SO4 to reform A1 ions
while the a-oxide is insoluble at reasonable acid levels.
Since carbon regeneration takes place below 1000°C, the
majority of the aluminum should be recoverable. Two gram
samples of regenerated sludge were slurried and acidified
over a range of pH values and the solution phase was then
analyzed for aluminum content colorimetrically. This pro-
cedure was followed for both the first cycle and third cycle
BCL regenerated carbon-aluminum mixtures. Results of these
experiments are summarized in Figures A-10 and A-ll. Recovery
of approximately 8 6 percent of the aluminum at a pH of 1.7 5
was observed with the first cycle mixture. Aluminum recovery
approached 100 percent for the third cycle case. The acid
requirement was virtually identical in both cases. At
this time, alum recovery appeared highly feasible.
Reuse of this recovered alum is discussed in the section on
the alum process optimization.
A similar line of investigation was pursued for a Fe2(504)3
coagulant. However, a red powder, believed to be Fe203 (which
should form at 200°C), formed in the thermal regeneration
step. This iron oxide was very acid resistant and a good
dissolution to recover ferric ion could not be achieved at
reasonable acid levels. It was concluded that the recovery
of Fe2S04, in this manner, was not feasible.
POLYELECTROLYTE SCREENING STUDY
Jar tests were run on a spectrum of commercial polyelectro-
lytes to identify those best suited for the operation.
-166-
-------
100 <
80
60
o
o
I
T
SEWAGE
SEWAGE
SEWAGE
T
T
VIRGIN CARBON
1st CYCLE ABCLA
VIRGIN CARBON
2nd CYCLE ABCLA
VIRGIN CARBON .
3rd CYCLE ABCLA
REGENERATED CARBON DOSE
CORRECTED FOR ASH CONTENT
<
O
l—i
to
LU
cn
40
20
500 1000
CARBON DOSE (mg/S,)
1500
FIGURE A-7. CAPACITY RECOVERY OF BATTELLE-COLUMBUS
REGENERATED CARBON-ALUM PROCESS
-167-
-------
100
SEWAGE
VIRGIN CARBON
SEWAGE
VIRGIN CARBON
2nd CYCLE BCLB
VIRGIN CARBON
SEWAGE
REGENERATED CARBON DOSE
CORRECTED FOR ASH CONTENT
60
o
40
U~)
0
500
1000
1500
CARBON DOSE (mg/Ji)
FIGURE A-8. CAPACITY RECOVERY OF BATTELLE-COLUMBUS
REGENERATED CARBON-BENTONITE PROCESS
-168-
-------
100
¦A— VIRGIN CARBON
a AFMCA
O FMCB
REGENERATED CARBON DOSE
CORRECTED FOR ASH CONTENT
O
o
—I
<
Q
LO
LlI
1500
500
1000
CARBON DOSE (mg/£)
FIGURE A-9. CAPACITY RECOVERY OF FMC REGENERATED CARBONS
-169-
-------
TABLE A-9
JAR TEST RESULTS WITH REGENERATED CARBONS
NO3-N NH3-N P04 TOC
(mg/1) (mg/1) (mg/1) (mg/1)
Influent
2.2
31
31.2
41
VC + Alum
0.22
28
3
9
VC + bentonite
<0.1
27
26.8
11.5
BCLB + 150 mg/1
bentonite
<0.1
26
28.8
13.5
BCLB + 300 mg/1
bentonite
<0.1
27
28.8
16.5
FMCB + 150 mg/1
bentonite
<0.1
27
26.8
12.5
FMCB + 3 00 mg/1
bentonite
<0.1
26
28.8
13
ABCLA + recovered alum
<0.1
26
25.4
8.5
ABCLA + virgin ;
alum
<0.1
26
0.8
9.5
VC - Virgin carbon
BCLB - Regenerated carbon-bentonite mixture - Battelle-Columbus
process
BCLA - Regenerated carbon-alum mixture - Battelle-Columbus process
ABCLA - Acid treated regenerated carbon-alum mixture - Battelle-
Columbus process
FMCB - Regenerated carbon-bentonite mixture - FMC process
Carbon Dose = 1000 mg/1
-170-
-------
• 1st CYCLE
o 3rd CYCLE
2 gm OF REGENERANT
MIXTURE WITH A CALCULATED
CONTENT OF 56.5 mg
pH
FIGURE A-10. EFFECT OF pH ON ALUMINUM RECOVERY
-171-
-------
10
O CYCLE §2
8
2 gms REGENERANT MIXTURE
6
4
2
0
30
6 12
VOLUME OF CONC.
18
FIGURE A-ll. ACID REQUIREMENT VERSUS pH-BCL
CARBON-ALUM MIXTURE
-172-
-------
TABLE A-10
FLOCCULATION PERFORMANCE OF VARIOUS POLYELECTROLYTES
Polyelectrolyte
0 10 mg/1
Performance in
Alum System
Performance in
Bentonite System
Nalco 60 0
Negligible
Negligible
Nalco 672
Good
Good
Magnifloc 837 A*
Good
—
Magnifloc 905 N
Fair-Good
Good
Atlas 105-659
Good
Fair
Atlas 300-400
Good
Good
Polyhall M-19
Fair
Fair
Purifloc N-17
Fair
Fair
Purifloc A-23
Very Good
Negligible
Purifloc C-23
Negligible
Negligible
Zetafloc WA
Poor
Negligible
*This is a commercial grade of Magnifloc 985 N
-173-
-------
TABLE A-11
EFFECT OF DOSE ON POLYELECTROLYTE PERFORMANCE - ALUM SYSTEM
Performance
Dosage (mg/1)
1
2.1
5
7.5
10
Polyelectrolyte
Nalco 672
Poor
Poor-Fair
Good
Good
Good
Atlas 300-400
Fair
Good
Very Good
Very Good
Very Good
Purifloc A-23
Fair
Very Good
Very Good
Very Good
Very Good
Magnifloc 837 A
Poor
Fair
Fair
Good
Good
-174-
-------
Qualitative observations on the performance of these poly-
electrolytes appear in Table A-10. As a result of this
series, four brands were selected for more extensive testing
in which doses were varied from 1-10 mg/1. Flocculating
characteristics observed in these tests are given in Table A-ll.
ALUM PROCESS OPTIMIZATION
It was decided to re-examine earlier results, which indicated
that a definite carbon contact time was required before alum
addition. This point was critical, since the process would
be highly simplified if it was not necessary to separate the
reclaimed alum from the carbon before reuse. Jar tests were
devised whereby the recovered alum was added simultaneously
with the carbon. After a five minute contact time, caustic
was added to raise the pH and then, after an additional five
minutes, polyelectrolyte was added. Flocculation proceeded
normally as long as the caustic addition was regulated to
achieve the desired effluent pH of approximately 7. This indi-
cated that the recovered alum would not have to be separated
from the carbon as thought earlier. Consequently, similar tests
were run on daily sewage samples for a period of two weeks in
conjunction with jar tests in which fresh alum was added simul-
taneously with the carbon. The recovered alum solutions
exhibited good flocculation every time while several samples
in which fresh alum was employed would not flocculate. This
result suggested that the carbon could remove the interfering
substance if the solution was maintained in an acidified state
for the first few minutes of contact. All subsequent observa-
tions reinforced this conclusion.
In order to further investigate the feasibility of using re-
claimed alum, the bench scale system was set up with a chemical
addition line located downstream from the carbon contact tank
and pH probe located slightly further downstream. Throughout
all of the runs, a lime slurry was pumped continuously in
sufficient quantity to maintain the pH in the range of 6.5-7.0
at the downstream point. In the initial run, a slurry of fresh
carbon and reclaimed alum was prepared and added in the same
manner as the carbon slurry in previous runs. Doses were
1000 mg/1 C, 10 mg/1 Purifloc A-23 and 350 mg/1 alum. A TOC
removal of 90 percent with a residual TOC of 9 mg/1 was observed.
Subsequent operation with a slurry of FMC regenerated carbon-
reclaimed alum resulted in an effluent with a residual TOC of
8.5 mg/1.
Additional bench scale experiments were conducted to study the
dose requirements of the polyelectrolytes which showed promise
in the beaker tests. FMC regenerated carbon and reclaimed alum
were employed in these runs with the results given in Table A-12.
Effluents of high quality were produced consistently in these
runs. This process employing regenerated carbon and reclaimed
-175-
-------
TABLE A-12
EFFECT OF POLYELECTROLYTE DOSE ON BENCH SCALE PERFORMANCE
Influent
TOC
(mg/1)
Polyelectrolyte
Dose
(mb/1)
Effluent
TOC
(mg/1)
TOC
Removal
129
Magnifloc 837-A
10
16
87 .7
5
14.5
88.9
2.5
14
89 .2
Atlas 300-400
5
12
90 .8
2.5
10.5
91.9
Purifloc A-23
2.5
10.5
91.9
Nalco 672
2.5
0.0
>99
5.0
0.0
>99
10.0
4.0
95
Purifloc A-23
1.0
0.0
>99
1.25
0.0
>99
2.0
2.5
97
2.5
0.0
>99
Atlas 300-400
1.25
0.0
>99
2.5
0.0
>99
76
-176-
-------
alum with pH control proved to be highly reliable in the con-
tinuous flow laboratory system.
Based on these results, it was concluded that the Atlas 3 00-4 00
and Purifloc A-23 were the two polyelectrolytes best suited for
alum flocculation.
Data obtained in the carbon contact tests suggested that the
carbon dose could be reduced substantially below 1000 mg/1.
Subsequent jar tests confirmed that there was little difference
in effluent quality if the carbon dose was reduced from 1000
mg/1 to 600 mg/1. Consequently, the bench scale system was run
at 600 mg/1 carbon. Earlier jar tests also indicated that with
the Purifloc A-23, the alum dose could be reduced to as low as
150 mg/1. The two new doses were checked simultaneously at a
Purifloc A-23 dose of 2.5 mg/1, with the results given below:
Alum Influent Effluent
Dose
(mg/1)
TOC
(mg/1)
TOC
(mg/1)
Removal
(%)
Inf. A-016
g/1 ABCLA carbon
150
105
3
97
Inf. B-016
g/1 Aqua Nuchar
200
88
11
88
Inf. A-016
ABCLA carbon
250
105
4
96
A small amount of carbon carryover was evident at the low alum
dose, but disappeared when the alum dose was increased to 200
mg/1. It was concluded that satisfactory process performance
could be achieved with a carbon dose of 600 mg/1 and an alum
dose of 200 mg/1. Further reduction in the carbon dose may
be possible with low TOC waste streams.
In order to investigate the effect of high solids and organic
loading on process performance, a special influent was pre-
pared by adding aged (60 days) primary sludge to Richland raw
sewage. The resulting mixture contained 2 68 0 mg/1 total
solids and 1400 mg/1 TOC. This waste was then treated in the
bench scale system with chemical doses at 600 mg/1 Aqua Nuchar
and 200 mg/1 alum. Initially, the system was operated at a
polyelectrolyte (Purifloc A-23) dose of 10 mg/1 with a result-
ing residual TOC of 47 mg/1 or 97 percent TOC removal. Effluent
TOC declined after startup and subsequent operation at a Puri-
floc A-23 dose of 2.5 mg/1 produced a product with a TOC resi-
dual of 17.5 mg/1, which represents 99 percent TOC removal.
Throughout the course of the run, effluent turbidities never
exceeded 1 JTU. These results indicate that the treatment
process can easily handle waste streams with high solids
contents.
-177-
-------
The system, at the conclusion of the laboratory studies, had
shown a high degree of stability with little or no upset at
startup and rapid recovery from pH disturbances. Carbon
carryover was unnoticeable, turbidity consistently less than
1 JTU, and TOC removals greater than 90 percent.
-178-
------- |