EPA-R2-73-183 _
Environmental Protection Technology Series
MARCH 1973
Activated Carbon Treatment
of Raw Sewage in
Solids-Contact Clarifiers
Office of Research and Monitoring
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
-------
-------
EPA REVIEW NOTICE
This report has been reviewed by the Environmental
Protection 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 constitute endorsement or recommenda-
tion for use.
ii
-------
ABSTRACT
Degritted raw municipal sewage was treated with powdered activated
carbon in a 28,000-gpd pilot plant. Two high-rate recirculating-
slurry solids-contact clarifiers operating in series with counter-
current carbon advance, followed by a gravity polishing filter, pro-
duced effluent equal to or better than that produced in a parallel
activated sludge plant.
TOG and COD removals averaged 88.1 and 88.7 percent, respectively,
with higher removals hindered by the concentration of adsorptive-
resistant materials present. Filtrable-TOC and -COD removals were
68.0 and 69.9 percent, respectively.
Alum and polyelectrolyte flocculated the powdered activated carbon
and raw sewage suspended solids into a fast settling floe. Subsid-
ence tests conducted on the solids slurry from the ACCELATOR® draft
tube indicated ACCELATOR overflow rates equivalent to or greater
than 2.5 gpm/ft2-
The maximum carbon adsorptive capacity for filtrable COD was 0.50
to 0.55 g COD/g carbon. This capacity was achieved whenever the
concentrations of influent COD and carbon matched or exceeded that
ratio (adsorptive-resistant COD excluded). Carbon requirements
were 55 to 60 percent of theoretical two-stage countercurrent adsorp-
tion system requirements.
Assuming regeneration recycles 85 percent of the carbon feed, respec-
tive treatment cost estimates for 10-mgd and 100-mgd plants were
13.9
-------
CONTENTS
Page
CONCLUSIONS 1
RECOMMENDATIONS 5
INTRODUCTION 7
Background 7
Process Description 8
Objectives of this Investigation 8
Experimental Approach 8
Analytical Procedures 9
PREPARATORY LABORATORY INVESTIGATION 11
Objectives 11
Screened and Degritted Raw Sewage Quality 11
Adsorption Isotherm Study 13
Coagulation Studies 16
PILOT PLANT PROGRAM 21
Apparatus and Procedures 21
Experimental Results 27
Phase 1 27
Phase 2 32
Phase 3 45
Phase 4 47
Phase 5 50
Phase 6 54
Phase 7 54
-------
CONTENTS Page
Supplemental Investigations 63
Adsorption Isotherms 63
Temperature 70
pH 74
Slowdown Suspended Solids 74
Color 74
Nutrients 74
Turbidity Measurement 75
Filter Operation 79
DISCUSSION 81
General Considerations 81
Process Economics 85
ACKNOWLEDGEMENTS 91
REFERENCES 93
APPENDIX 95
VI
-------
FIGURES
Page
1 Adsorption Isotherm Carbon Comparison, Example A 14
2 Adsorption Isotherm Carbon Comparison, Example B 15
3 Extended Range Isotherm 17
4 Schematic, Pilot Plant System 22
5 Photograph of 28,000 gpd Pilot Plant 24
6 Mean Values of Filtrable COD 30
7 Diurnal Variation, Clarifier Slurry Settling Tests,
Phase I 31
8 Mean Values of Suspended Solids 33
9 Mean Values of Filtrable TOC 35
10 Mean Values of Filtrable TOC Removal 36
11 Mean Values of Filtrable COD Removal 37
12 Mean Values of BOD5 38
13 Mean Values of Clarifier Slowdown 39
14 Recirculating Carbon Slurry Settling Rate Curves,
Part 1 41
15 Recirculating Carbon Slurry Settling Rate Curves,
Part 2 42
16 Mean Values of Spectrophotometer Turbidity 43
17 Typical Gravity Filter Headless Curves 44
18 Solids Content of Recirculating Slurry, Unit //I 48
19 Solids Content of Recirculating Slurry, Unit #2 49
20 Diurnal Variation, Clarifier Slurry Settling Tests,
Phase 5 55
vix
-------
FIGURES Page
21 Mean Values of TOG 57
22 Mean Values of COD 58
23 Mean Values of COD Removal 59
24 Mean Values of BOD5 Removal 60
25 Mean Values of TOG Removal 64
26 Daily Variation in COD, Period 15 65
27 Mean Values of Color Removal 66
28 Freundlich Adsorption Isotherm Pilot Plant Analysis 68
29 Mean Value Isotherms, TOG 71
30 Clarifier Performance, Applied TOG Load vs. Carbon TOG
Loading 72
31 Clarifier Performance, Applied COD Load vs. Carbon COD
Loading 73
32 Ratio of Spectrophotometer JTU and Hach JTU 76
33 Ratio of Spectrophotometer Turbidity to Suspended Solids 77
34 Diurnal Variation of JTU in System 78
viii
-------
TABLES
Page
1 Screened and Degritted Sewage Characteristics 12
2 Phase 1 Pilot Plant Performance 28
3 Phase 2 Pilot Plant Performance 34
4 Phase 3 & 4 Pilot Plant Performance 46
5 Phase 5 Pilot Plant Performance 51
6 Phase 6 Pilot Plant Performance 56
7 Phase 7 Pilot Plant Performance 62
8 Mean Treatment Plant Performance 83
IX
-------
CONCLUSIONS
Degritted raw municipal sewage can be treated efficiently and economi-
cally with powdered activated carbon and coagulants in a continuous
high-rate process.
The process effluent is equal to or better in quality than that pro-
duced in a parallel activated sludge plant.
Activated carbon utilization is maximized by contacting the carbon
with the waste stream in two series-connected high-rate recirculating-
slurry solids-contact clarifiers. Each clarifier is capable of exceed-
ing single-stage carbon adsorption efficiency because the carbon flows
countercurrent to the waste stream during a portion of its circulation.
Fresh carbon is fed to the recirculation zone of the downstream clari-
fier and is then transferred countercurrent to the waste stream between
the clarifiers. This equipment and flow arrangement provides an
effective two-stage countercurrent contacting system.
Under actual test conditions, the powdered activated carbon requirement
was conservatively calculated to be not more than 20 to 25 percent of
single-stage batch contacting or 55 to 60 percent of two-stage counter-
current adsorption requirements.
The unique compound-contacting system makes possible nearly complete
removal of carbon-adsorbable COD-producing compounds at peak carbon
loadings. The test sewage stream contained a significant amount of
adsorption-resistant compounds which limited the removal of filtrable
TOG and COD to 68.0 and 69.9 percent, respectively, of that in the
feed stream. The activated carbon adsorption capacity for filtrable
COD was between 0.50 and 0.55 g COD per g of carbon. This loading
range was achieved whenever the adsorbable COD and carbon were present
at this or higher ratios. The carbon adsorption capacity for filtrable
TOC was about 0.17 g per g of carbon.
The optimum carbon dosage can be calculated from the concentration of
available adsorbate and the carbon adsorption capacity. In this study,
this was found to be 132 percent of the feed stream filtrable COD or
about 1 lb/1000 gal. Higher carbon dosages removed very little addi-
tional adsorbate resulting in reduced adsorbate loading per unit weight.
Lower carbon dosages resulted in capacity carbon loading, but reduced
the degree of treatment.
The process removes the major portion of the total TOC and COD from the
sewage by coagulation and sedimentation of the raw sewage suspended
solids. The powdered activated carbon and its adsorption process are
combined with and proceed simultaneously with the solids removal.
The combined systems, along with effluent filtration resulted in mean
total-TOC and -COD removals of 88.1 and 88.7 percent as compared with
the activated sludge plant removals of 86.6 and 88.2 percent, respec-
-------
tively. BOD removals were 90 percent for the carbon adsorption system
and 88 percent for the activated sludge system.
Laboratory studies are particularly useful in determining the more
effective and economical treating chemicals. This study determined
that Aqua Nuchar A and Darco S-51 powdered activated carbons were
equally effective adsorbents for this sewage and they were effectively
coagulated, in conjunction with raw sewage solids by 6 to 8 mg/1 of
C-31 polyelectrolyte and 10 to 15 mg/1 of filter alum. Operation of
a 28,000-gpd pilot plant substantially confirmed these findings.
Unit chemical costs are relatively high. Aqua Nuchar A is the more
economical product and was used during most of the pilot plant tests.
Darco S-51 was estimated to cost 0.6^/1000 gal. more than Aqua Nuchar
A but its physical characteristics, i.e., higher bulk density and
greater particle size, offer advantages in floe density, floe dewater-
ing, housekeeping and reduced storage volume requirements.
Neither C-31 polyelectrolyte nor filter alum, when used singly, was
suitable for the needs of the process. Without filter alum the re-
quired polymer dosage was undesirably costly. Alum in quantities
needed for good clarification produced excessive volumes of sludge.
When combined, the polymer was observed to coagulate the carbon effec-
tively while the alum was instrumental in controlling turbidity causing
materials present in the sewage.
During the pilot plant study the limited pumping capacity permitted a
maximum ACCELATOR® clarifier overflow rate of 1.5 gpm/ft2. However,
subsidence tests on the solids slurry taken from the draft tube of the
ACCELATOR mechanism indicated that overflow rates equivalent to or
greater than 2.25 gpm/ft2 were easily achievable. Therefore, based
upon these subsidence tests and other in-house experience with ACCELA-
TOR units, a conservative overflow rate of 2.25 gpm/ft^ was used as
a design parameter to establish the cost of 10- and 100-mgd treatment
facilities. At this overflow rate the experimental system, including
final gravity filtration, would have a retention time of 63 minutes.
As discharged from the experimental system the solids content of the
blowdown mixture of raw sewage solids and carbon was 2.5 percent.
External thickening would be required for economical carbon regenera-
tion. Regeneration and reuse of the spent carbon were not studied.
Filtration of the treated stream is practical. Although it may not be
required for some potential applications of the process, the existence
of the polishing filter permits some reduction in coagulant dose by
confining the resulting increase in carbon-floe carryover to the plant
site. The filter assures carbon-free effluent when, necessary. No
carbon passed through the pilot plant filter except at times of ex-
perimental extremes. The filter was operated at surface loadings
up to 2.8 gpm/ft2 with a maximum headless limit of 4.5 ft. Nominal
-------
wash water consumption was 3.0 percent of that filtered during filter
runs of 25 hours, although these figures varied widely with the experi-
mental program. Modest biological growth in the underdrain system was
controlled but not eliminated by chlorinated water backwashing.
Despite varied operating conditions, the system exhibits remarkably
uniform adsorption process performance. The clarifier carbon inventory
was capable of maintaining effluent quality even with up to 8 hours
loss of carbon feed.
Coagulant consumption for the complete process is double the average
single treatment dosage previously stated because there are two clarifi-
cation steps. On a 10-mgd scale, assuming carbon, polyelectrolyte and
alum consumption of 118, 14 and 25 mg/1, respectively, the total esti-
mated cost of treatment by the pilot plant process is 13.90/1000 gal.
A 100-mgd plant of similar design would produce treated water at a cost
of 11.2C/1000 gal. Both cost figures incorporate a 1.9C/lb rate for
regeneration of spent carbon slurry assuming the raw sewage solids
contained in the slurry present no special cost problems. Both plants
feature 85 percent carbon recovery with 15 percent new carbon makeup.
Solid and liquid waste handling and disposal problems are minimal for
either plant.
-------
RECOMMENDATIONS
This pilot plant investigation demonstrated the feasibility of treat-
ing raw sewage in a short-retention-time carbon-adsorption/flocculation
system using new activated carbon. Process economics depend upon re-
covery and reuse of regenerated carbon from the mixture of raw sewage
solids and spent carbon in the blowdown slurry. Effort should be
directed at determining if such regeneration is possible. Means of
concentrating the blowdown slurry solids to optimum concentrations
for economical regeneration should be explored simultaneously. Work
to date has generally been restricted to thickening and regenerating
spent carbon from tertiary treatment processes.
Adsorption-resistant organic materials in the process effluent will
vary in concentration and character between plants. Methods for con-
trolling or removing this fraction should be investigated, possibly
with present-day processes for nutrient removal which may be desirable
or necessary.
No biological growth inhibiters were used in the carbon contacting
portion of the pilot system during this study. Potential process
interferences as the result of biological growth were not encountered;
however, this may not be typical for all applications. The effective-
ness and cost of different means of controlling biological growth
throughout the system warrant attention.
The presence of filtrable solids in the sewage being processed inter-
fered with making a precise theoretical evaluation of the pilot plant
adsorption capabilities in terms of accepted laboratory control pro-
cedures. Basic process efficiency should be determined and equated
to known adsorption theory by operating a reduced scale plant on a
laboratory prepared feed stream of adsorbate, monitored with suitable
adsorption isotherm studies. The relationships established would
simplify the problems of predicting performance for suggested practi-
cal applications.
-------
INTRODUCTION
Background
Adsorption is an established process for separating relatively small
quantities of organic impurities or valuable products from dilute
solutions. Activated carbon is one of the more efficient and widely
used adsorbents.
Activated carbon adsorption has been used in water purification for
more than 50 years.1 It has successfully competed with other purifi-
cation or recovery processes in industry, and more recently, has been
applied to advanced wastewater treatment.2 The need to control the
amount and types of waste products fouling the environment has been
the driving force for advanced wastewater treatment research. A
secondary impetus is the benefit derived from recovering wastewater
for reuse, thus preserving existing supplies of high quality raw waters.
The amount of carbon required to improve the quality of sewage plant
effluents to acceptable levels may be very high because adsorption is
controlled by aspects of kinetics and adsorption equilibrium.3 A
traditional single-stage powdered activated carbon contacting system
is relatively inefficient and costly to operate because of the equi-
librium adsorption phenomena. Continuous multistage countercurrent
systems offer improved carbon economy at the expense of greater capital
costs for the more extensive equipment system. Cost factors have
favored the use of granular activated carbon in columnar systems with
regeneration and reuse of the carbon. Recently compound systems of
contacting powdered carbon have been tested that operate with improved
economy, approaching the cost of granular carbon systems provided
regenerated carbon is used.
Two studies '" utilized powdered activated carbon in coagulated slurry
form recirculating in ACCELATOR clarifiers which are high-rate solids-
contact treating units. These early investigations were confined to
tertiary applications with the inherent disadvantage of working with
low potential adsorption equilibrium conditions. With a low initial
concentration of adsorbate, the carbon utilization was limited. Both
studies, however, produced carbon utilization levels equal or superior
to theoretical systems based upon isotherms derived from laboratory
jar tests.
Providing a higher concentration of adsorbate along with refinements
in process design would substantially improve the amount of impurities
adsorbed upon each measure of activated carbon. The highest concentra-
tion of adsorbate of interest occurs in the raw sewage entering the
treatment plant and remains little changed if not unchanged as the
sewage passes thru the primary clarifier. While substantial amounts
of suspended matter are removed by screening, grit removal and primary
-------
sedimentation the soluble material flows on to the biological or
chemical process units.
If a compound-contacting powdered activated carbon system could be
operated on primary clarifier effluent or even screened and degritted
raw sewage, more efficient use could be made of the carbon's adsorptive
properties. A substantial amount of the operating cost would be offset
by eliminating the need for traditional secondary treatment facilities
and perhaps primary clarification and sludge disposal also. Certainly
coarse screening (or comminution) and grit removal would be retained
to protect equipment. A compact chemical-carbon slurry treating plant
would be free of the biological problems associated with traditional
secondary treatment facilities.
Process Description
A two-contactor countercurrent system was tested. Application of the
process involved series operation of two solids-contact clarifier units
of a type used widely for water treatment. New carbon was fed to the
second unit only. The carbon feed for the first unit was the carbon
slurry wasted from the second unit. Spent carbon was withdrawn from
the system by blowdown from the first unit. The carbon treated and
clarified liquid was filtered through a dual media gravity filter prior
to discharge.
Objectives of This Investigation
The primary objective of the present investigation was a performance
and cost evaluation of such a system for treatment of screened and
degritted raw municipal sewage. Provisions made for processing primary
clarifier effluent, if the treatment of screened and degritted raw
sewage proved impractical, were not used.
Experimental Approach
A commercially available powdered activated carbon and flocculation
agents to improve slurry settleability were selected on the basis of
laboratory study.
These materials were then used during 27 consecutive pilot plant
operating periods under various conditions. These operating periods
have been grouped for reporting into seven phases, each grouping con-
sisting of operating periods of related interest.
A high proportion of total pilot plant operating time was expended in
investigating coagulant cost reduction possibilities.
First-phase operating periods established physical and hydraulic capa-
bilities of the system and determined the treatability of the screened
-------
and degritted raw sewage. Second-phase periods monitored the effects
of reducing and combining coagulant feeds. The third phase demon-
strated performance of a different powdered activated carbon followed
by the fourth phase, a reference operating period. Fifth-phase oper-
ating periods further pursued the coagulant cost reduction objective
using greater deviations in feed rates than previously used. Sixth-
phase periods were intended to establish operating parameters using
only an inorganic chemical coagulant. Seventh-phase periods compared
results achieved with different dosages of powdered activated carbon.
During all operating periods, data were collected from the pilot plant
and the adjacent municipal sewage treatment plant. Single-stage adsorp-
tion isotherms were developed at intervals for the feed stream and for
the upstream clarifier effluent. Precise comparison of these data by
application of adsorption theory'' was not possible. Essential labora-
tory procedures introduced an irreconcilable datum shift between pilot
plant results and isotherm results.
Peripheral laboratory analyses of 24-hour composite samples plus
auxiliary grab samples were used to disclose operational parameters
other than those directly associated with the adsorption process.
These data, together with the pilot plant control records, reveal
important practical operating requirements and results in other areas
of interest.
Analytical Procedures
With certain exceptions, methods used in this study conformed to those
in "Standard Methods."8
A modification to "Standard Methods" developed especially for low-level
COD values was used. The complete procedure appears in the Appendix.
Difficulty in obtaining reproducible results in 1C (inorganic carbon)
determinations with the total organic carbon analyzer coupled with the
presence of 65 to 90 mg/1 of 1C in all samples prompted the abandon-
ment of 1C determinations. The deviation in 1C determinations per
sample approached the level of TOG to be measured. Inorganic carbon
fractions of all samples for TOC determinations were removed by acidify-
ing a 200-ml sample with 1 ml HC1 and degassing with caustic-solution-
scrubbed and filtered compressed air for 15 minutes. Replicate 20-
microliter prepared samples were then manually injected into the TC
(total carbon) combustion chamber of the furnace module. Laboratory
trials revealed that this procedure reduced the indicated inorganic
carbon residual to 1.0 mg/1 C (the same as ^-scrubbed samples) near
the zero end of the analytical range (0-100 mg/1 C) and to 1.2 mg/1 C
at the 35-mg/l TOC level. Filtered sample TOC determinations usually
were less than 35 mg/1. Instrument calibration curves prepared in
accordance with the manufacturer's directions were compensated to elimi-
nate 1 mg/1 of inorganic carbon residual.
-------
Sets of samples prepared and analyzed for TOG in the laboratory were
submitted to the EPA Advanced Waste Treatment Research Laboratory in
Cincinnati, Ohio for comparison of results. The comparison determina-
tions differed by 1 mg/1 or less on a majority of the samples. Vari-
ous operational difficulties with the several components of the TOC
analyzer were detected and correct operation verified by frequent
standardization. Standardization always preceded sample analyses.
All samples for which filtrable (filtrate) COD and TOC are reported
were vacuum filtered through 0.45-micron Millipore filter discs prior
to analysis. The filtrates were observed at times to contain discern-
able amounts of filtrable solids. Filtration was necessary in the
case of powdered activated carbon treated samples as residual carbon
particles contribute to the COD and TOC values. In order to provide
a usable relationship between carbon treated samples and untreated
samples, most of the untreated samples were filtered and reported in
a like manner. Recently, vacuum filtering and also acidification and
degassing procedures are recognized as altering the COD and TOC test
results." Volatile organics are lost in both procedures and this was
verified during this study. A detailed investigation was not insti-
tuted and corrective procedures and equipment were not added to the
program. The resulting error is not considered to alter the study
conclusions significantly and permits direct data comparison with
other studies using similar techniques.
Suspended solids determinations on the composite samples of screened
and degritted raw sewage incorporated a preliminary coagulation step
to facilitate filtering. Filter alum was used to treat the sample
aliquot and a tap water blank. Sample results were corrected by the
amount of the suspended solids produced in the blank. Direct vacuum
filtration of the screened and degritted raw sewage aliquot was aban-
doned after encountering filter mat clogging with aliquots as small
as 25 ml.
Certain ACCELATOR clarifier control tests performed at the pilot
plant site are discussed hereinafter.
10
-------
PREPARATORY LABORATORY INVESTIGATION
Objectives
An adsorbent-coagulant system was to be developed which would be
physically manageable in the pilot plant equipment as well as produce
process data. The initial feed stream was to be screened and degritted
raw domestic sewage. The characteristics of the feed stream had to be
determined. The adsorbent, activated carbon in powdered form, is avail-
able from different sources. It is manufactured from various raw ma-
terials by a number of different processes and may be specialized in
its application.10 Activated carbons were tested with aliquots of feed
stream to assure selection of a material that would perform within the
time limitations and hydraulic conditions imposed by the pilot plant.
A flocculation system using this carbon in the feed stream had to be
found. Inorganic coagulating agents were to be avoided if possible as
they could be cumulative in the carbon when reactivation and reuse are
concerned. The flocculated slurry must settle rapidly, leaving little
if any residual carbon particles in the supernatant; and it would be
advantageous if the slurry dewatered well on standing.
For the purposes of this study, the filtrable TOC of the feed is the
primary adsorbate concentration indicator. Secondary indication was
accomplished with filtrable COD determinations. Some impurities pres-
ent may be adsorbed which do not contain appreciable TOC or exhibit COD
while others with these characteristics may not be readily adsorbed.
Sewage contains a varying combination of these materials, many of them
not identified; so the TOC and COD determinations are a useful but in-
complete pollution indicator.
Screened and Degritted Raw Sewage Quality
The City of Tucson, Arizona operates a 12.9 mgd step aeration activated
sludge sewage treatment plant constructed in 1968. Raw sewage is passed
through a mechanically-cleaned coarse bar screen and a grit-removal
system prior to entering a flow-dividing weir box situated between the
two primary clarifiers. Influent enters the bottom of the box, flows
vertically upward and discharges to each of the clarifiers over two
adjustable weir gates. Neither sediment nor floatable material accumu-
lates in the space between weirs. This space was selected as the take-
off point that would best provide a continuous representative supply to
the pilot plant.
The sewage plant flow varies around design capacity at which the reten-
tion times are 2.0 hours in the primary clarifiers, 4.2 hours in the
aeration basins and 2.1 hours in the final clarifiers. Physical and
chemical characteristics of several raw sewage grab samples collected
at random from the flow-dividing weir box are presented in Table 1.
11
-------
TABLE 1
Screened and Degritted Raw Sewage Characteristics, mg/1
Tucson Sewage Treatment Plant
Date
Time
Calcium (as Ca)
Magnesium (as Mg)
Sodium (as Na)
Potassium (as K)
Phosphorus (as P)
Ammonia Nitrogen (as N)
Silica (as Si02)
Non-Carbonate Solids (as CaCO^)
Total Cations (as CaC03)
Alkalinity (as CaCO^)
Hardness (as CaCO^)
Color (SU)
Turbidity (JTU)
Suspended Solids
TOG (unfiltered)
TOG
COD (unfiltered)
COD
BOD5 (unfiltered)
PH
11/12/69
8:30 A.M.
64
14.6
132
10.1
3.5
26.7
25.2
320
600
280
220
22
115
202
-
33.0
-
92
126
8.1
11/14/69
8:30 A.M.
59
14
119
8.6
5.7
29.5
34
262
546
298
198
18
115
206
124
25.0
-
88
134
7.9
11/18/69
4:00 P.M.
60
16.5
133
10.9
7.1
26.8
32
298
618
306
218
28
250
374
169
61.0
-
124
242
7.9
12/3/69 1/14/70
7:00 A.M. 10:30 A.M.
55
17.0
140
13.0
4.4
19.5
42
266
588
"304
203
32
130
165
85.0 180
23.0 47.5
740
106 155
162 274
7.9
Samples filtered thru 0.45-micron filter except as noted
-------
Adsorption Isotherm Study
Two powdered activated carbon products were selected for evaluation
against project requirements. These were Darco S-51, a product of
Atlas Chemical Industries, Inc. and Aqua Nuchar A, a product of the
Westvaco Corporation. A third product, Hydrodarco B, also manufac-
tured by Atlas Chemical Industries was investigated briefly. It is
marketed on a restricted basis.
Darco S-51 was used in our earlier experiments and Aqua Nuchar A was
utilized in an EPA pilot plant at Lebanon, Ohio. The rate of adsorbate
uptake for both of these products had been found acceptable in the pre-
vious studies with secondary effluent. Adsorption efficiency, however,
would need reevaluation with raw sewage.
Sufficient raw sewage was prefiltered through 0.45-micron filter discs
to perform each carbon comparison study. A stock solution of each grade
of carbon was prepared with special low-COD distilled water (appendix).
The activated carbons were added to sets of five 250-ml aliquots of the
prepared sewage at concentrations of 40, 160, 320, 800 and 1600 mg/1.
A sixth aliquot in each set was left untreated as a control. The jars
were stirred at 100 rpm for 60 minutes. Their contents were then fil-
tered thru 0.45-micron filter discs, and filtrate TOC and/or COD
determined.
Previous studies had shown the feasibility of handling the carbons as
stock suspensions. To avoid performance differentials resulting from
differences in wetting characteristics the suspensions were prepared
not less than 24 hours prior to use and subjected to frequent manual
shaking during use. The stirring speed of 100 rpm and the 60-minute
stirring time also resulted from the previous experiments.
Freundlich curves of the data (Figures 1 and 2) were developed. The
empirical Freundlich equation is generally written: X/M = KC|- 'n,
where X is the mg/1 of adsorbate taken up by M mg/1 of adsorbent
applied. Ct is the mg/1 of adsorbate remaining unadsorbed at equi-
librium and K and 1/n are constants specific for the materials and
conditions of test. When X/M is plotted against C on logarithmic
paper, a straight line is commonly obtained which Has a slope of 1/n
and an intercept K at Ct = 1.
Slight variations in the original concentrations (Co) of TOC and COD,
perhaps caused by the passage of time, between the individual carbon
test sets were retained in plotting the Freundlich curves. However,
for positioning the curves of constant carbon dosage, an average Co
value was utilized.
With rare exception, the data from 19 separate isotherm experiments
failed to plot as a straight line. This was true for all three carbons
studied and for TOC- and COD-generated curves. Curves from TOC and COD
13
-------
100
80
60
40
20
10
60
c
•H
T)
cd
o
d
o
-a
cu
•H
Sewage Sample 26798-T
60 min contact
23°C
100 rpm
Filtrable CODO =87.3 mg/1
Filtrable TOCO =28.0 mg/1
O Darco S-51
D Aqua Nuchar A
10
Carbon Dose, mg/1
40
60
20 30 40
Filtrable COD, mg/1
Concentration Remaining (Ct)
80 10 20 30
Filtrable TOC, mg/1
FIGURE 1: Adsorption Isotherm Carbon Comparison, Example A
14
-------
100
80
60
40
20
10
60
a
•H
T3
n)
o
c
o
"fi 5
cd
o
0)
4-J
•H
4-1
a
Sewage Sample 26839-T
60 min contact
23°C
100 rpm
Filtrable COD
Filtrable TOCO =
88.1 mg/1
29.5 mg/1
O Darco S-51
O Hydrodarco B
Carbon Dose, mg/1
40
10 20 30 40 60 80 10 20 30
Filtrable COD, mg/1 Filtrable TOC, mg/1
Concentration Remaining (Ct)
FIGURE 2: Adsorption Isotherm Carbon Comparison, Example B
15
-------
data from an individual jar set differed in configuration as the ratio
of COD to TOG deviated with carbon concentration. Utmost attention to
technique did not resolve these questions. Each sewage sample produced
a differently located and shaped isotherm. Figures 1 and 2 are repre-
sentative of these variations. In general, a single-stage carbon load-
ing limit was implied. At middle-range loadings, a relatively straight
curve sometimes occurred. At lower loadings there was consistent evi-
dence that residual adsorbate resisted adsorption regardless of the
carbon under test. Figure 3, an extended-range isotherm illustrates
these phenomena. There were small differences in the performances of
the carbons studied when all isotherms are considered collectively.
Darco S-51 usually exhibited a slightly higher capacity than Aqua
Nuchar A which was slightly superior to Hydrodarco B. Selection of
Aqua Nuchar A for the major portion of the program was finally made
as a result of price differential.
The manufacturer states Nuchar active carbons are made from a residual
organic material which is recovered during the manufacture of wood pulp.
The organic material is carbonized and then activated under controlled
conditions to obtain active carbon grades of the desired form and ad-
sorptive capacity. Aqua Nuchar A is recommended for the removal of
tastes and odors from potable water supplies, to control odors at sewage
plants and to correct industrial pollution. It has a phenol value of
20 mg/1 (plus or minus 10 percent) or less as measured by the phenol
test of the American Water Works Association. Available only in pow-
dered form, a typical wet-sieve analysis gives the following minimum
percentages passing each sieve: 100 mesh, 99 percent; 200 mesh, 98
percent; and 325 mesh, 95 percent. It has a surface area of 600 m /g
and an apparent bulk density of 13.5 - 15 lb/ft^. Ash content is
6 percent and moisture content is 3 percent.
Coagulation Studies
For successful operation of the high-rate clarification units compris-
ing the pilot plant, it was essential that an efficient and effective
coagulation system be developed. The floe particles had to settle
rapidly, be strong enough to withstand the repeated cycling through
the flocculation paddles, entrap a high percentage of the carbon and
raw sewage particulate matter, and dewater readily. In addition, it
was suggested inorganic chemicals should be avoided, if possible, to
minimize the accumulation of inorganic matter in the carbon. Thermal
regeneration and reuse of the carbon were not included in this study,
but the process should minimize carbon contamination with inert solids.
Organic polyelectrolytes were tested as primary flocculating agents,
augmented as indicated with various amounts of inorganic materials.
A few experiments established the dosages of inorganic flocculants to
support the balance of experiments.
16
-------
•H
T3
cfl
C
C
o
,0
VJ
ni
U
cfl
•H
+J
a
100
80
60
40
20
10
8
6
4
T
I I I
I I I
l
0.8
0.6
0.4
0.2
0.1
Sewage Sample 26804-T
60 min contact, Aqua Nuchar A
24°C
100 rpm
Filtrable CODO = 51.0 mg/1
Filtrable TOCO = 16.0 mg/1
Carbon Dose, mg/1
20
O 40
COD
TOG
3200
4000
I
4 6 8 10 20 40
Concentration Remaining (Ct), mg/1
Figure 3: Extended Range Isotherm
60 80 100
17
-------
The emphasis for much of the coagulation investigation was directed at
cationic polyelectrolytes. A wide assortment of materials was avail"
able initially and promising ones were retested with factory-fresh
supplies to eliminate effects of limited shelf life.
Products utilized and their respective manufacturers are as follows:
1.
2.
3.
4.
5.
6.
7-
8.
9.
10.
11.
12.
13.
14.
15.
Aquaf loc
CAT-FLOG
Magnifloc
Magnif loc
Primaf loc
Primaf loc
Primaf loc
Primaf loc
Purif loc
Purif loc
Purif loc
Purif loc
Purif loc
Triton
Triton
412
521-C
560-C
C-3
C-5
C-6
C-7
A-21
A-22
C-31
C-32
N-12
CF-32
QR-419
Dearborn Chemical Division
Calgon Corporation
American Cyanamid Company
American Cyanamid Company
Rohm & Haas
Rohm & Haas
Rohm & Haas
Rohm & Haas
The Dow Chemical Company
The Dow Chemical Company
The Dow Chemical Company
The Dow Chemical Company
The Dow Chemical Company
Rohm & Haas
Rohm & Haas
Filter alum, ferric sulfate, ferrous sulfate, ferric chloride, acti-
vated silica, sodium hydroxide and lime were the inorganic materials
utilized.
Over 800 individual jar tests were conducted on unfiltered screened
and degritted raw sewage grab samples with variations in mixing speed,
mixing time, carbon dosage, coagulating agent(s), order of coagulating
agent and carbon addition, and the use of floe from previous tests.
Aliquots, generally 200 ml of sample, were treated first with activated
carbon stock solution to either 200-mg/l or 250-mg/l dosages. While
being stirred at test speed, the coagulating agent(s) were added,
generally in the range of 2 to 20 mg/1. Stirring time was normally
10 minutes; however, this was shortened to 5 minutes if floe breakup
was observed during the longer period. Stirring speed was 50 to 70 rpm
with no flash mixing practiced. Supernatant turbidities were deter-
mined on most jars showing promise. A 5-minute settling period was
allowed before sampling the supernatant via pipetting from beneath the
liquid surface for turbidity measurement. Notes were made on the time
required for the prominent floe particles to settle the 3.5-in. liquid
depth to the jar bottom.
A substantial portion of these studies were undertaken before a carbon
brand selection was finalized, so both Aqua Nuchar A and Darco S-51
were tested. Better floe formation, settling rate and completeness of
clarification, perhaps because of higher bulk density (27-33 Ib/ft )
18
-------
and larger particle size (70% passing a 325-mesh sieve) were reported
for the Darco S-51.
Each clarifier of the pilot plant was expected to exhibit individual
coagulant requirements. A jar-test procedure simulating the two stages
of treatment was utilized in about 20 percent of the aforementioned
studies. Briefly, the procedure involved paired sets of jars. One of
each pair representing the 1st stage or upstream clarifier and the other
the 2nd stage or downstream clarifier. Treatment was tested in consecu-
tive cycles. Each cycle for the 1st stage involved decanting the clari-
fied supernatant, discarding a portion of the existing carbon slurry
sediment, replacing this with a similar portion from the second-stage
jar, adding raw sewage, mixing and coagulating. Each cycle of the 2nd
stage involved decanting the previously clarified supernatant (sometimes
for test, otherwise to waste) removing a portion of the previously
formed carbon sediment (to Ist-stage jar), adding the Ist-stage decanted
supernatant, mixing, adding the carbon and coagulating. In this fashion
the requirements for flocculating the raw sewage suspended solids in the
presence of already flocculated carbon were confined to the Ist-stage
jar. Requirements for flocculating new carbon were confined to the 2nd-
stage jar. Both jars retained and reused in subsequent cycles a quantity
of previously-formed carbon-bearing floe for as many as nine cycles,
during which floe characteristics change and coagulant dosage changes
may be evaluated. The solids-recycle procedure is particularly appli-
cable to the type of solids contact units used in the pilot plant.
Typical data reporting for jar studies of this type indicates the
amounts and types of materials used, as well as notes on floe formation,
quiescent settling ability, and clarification results as JTU or, as be-
ing poor, fair or good. The resulting slurry volume existing at the end
of the settling period, as well as the number of previously formed slurry
volumes returned to each jar are also recorded.
None of the polyelectrolytes when used individually were consistently
effective in reducing supernatant turbidity to the desired range of
10 JTU except in dosages of 15 mg/1 per stage or greater. Four products,
C-31, C-32, C-7 and CAT-FLOG were found to be superior to others tested.
They were equally effective at reduced feed levels when Ist-stage coagu-
lation was aided by 10 mg/1 filter alum. The polymers appeared to be
somewhat selective in coagulating activated carbon well while rejecting
fractions of the raw sewage suspended solids which contributed the bulk
of residual turbidities in the jar supernatants. Modest amounts of
alum used in the Ist-stage jars improved supernatant turbidities sub-
stantially. Calculations show filter alum produces less precipitate
weight per mg/1 fed than the other commonly used inorganic coagulants.
This possibly could be an important factor when regenerating spent
carbon.
19
-------
Based upon price and ease of handling, C-31 was selected as the primary
flocculating agent. Later, studies with the pilot plant would incorpo-
rate auxiliary alum feeds if found desirable.
Having completed selection of Aqua Nuchar A activated carbon and C-31
polyelectrolytic coagulant, operation of the pilot plant was initiated.
20
-------
PILOT PLANT PROGRAM
Apparatus and Procedures
An existing constant-rate pilot plant was modified and applied to this
study. The pilot plant was constructed in 1968 at the INFILCO Division
Test Facility adjacent to the Municipal Water Pollution Control Plant
serving the City of Tucson, Arizona.
The plant included two series-connected JBAS® ACCELATOR clarifiers,
a type of recirculating-slurry solids-contact unit which is the princi-
pal component of a pre-engineered water treatment system used extensive-
ly in the beverage industry. As depicted in Figure 4, the pilot plant
schematic flowsheet, the JBAS ACCELATOR clarifier contains an inlet-
mixing chamber, flocculators and a slurry-recycle impeller on a common
shaft, a slurry concentrator, and a clarified water zone with effluent
take-off system.
Certain features of the clarifiers were modified at the time of instal-
lation for a previous study. A submerged-orifice peripheral launder
designed for 30,000 gpd was substituted for the 6,000-gpd original
effluent-collection system. The concentrator rim was modified by add-
ing a spout to extract solids directly from the recirculating carbon
slurry. This made solids collection independent of the slurry pool
level. These alterations were retained for this study. New modifica-
tions included reducing the concentrator volume from 17.4 gal. to 6.9
gal. by means of a liner and the addition of a peripheral aluminum
scum baffle to the upstream unit.
Operation at throughput rates as high as 28,000 gpd was realized, re-
sulting in an overflow rate of 1.5 gpm/ft2 of clarification area based
on the 12.9-ft cross-section of the annular clarification zone.
Each of the 730-gal. clarifiers (4.5 ft diameter by 7.5 ft sidesheet)
included a recirculation zone volume of 225 gal. which provided carbon
contact times of 35, 23 and 12 minutes at proposed operating rates of
0.5, 1.0 and 1.5 gpm/ft^, respectively.
In a solids-contact unit of this type, the throughput rate cannot
exceed the pumping capacity of the recirculation system if short-
circuiting to the solids separation zone is to be avoided. The esti-
mated pumping capacity of the slurry recirculation impeller in the
pilot units was five times the maximum operating rate. The recircula-
tion flow also prevents deposition of solids on the tank floor. Except
for material isolated in the slurry concentrator, all of the carbon
inventory was maintained in suspension during operation.
21
-------
NO
t-0
System Sample Taps
Direction of Flow
Coagulant Pumps
Magnetic
Flow Meter
JBAS Clarifier Features (Modified)
A Inlet Chamber
B Mixing & Flocculation Zone
C Secondary Flocculation Zone
D Recirculating Slurry Pool
E Clarified Water Zone
F Submerged-Orifice Launder
G Slurry Concentrator
H Drive
Rate Set
Valves
Electric
Valve
Rotary
Screen i
Sewage Plant
Weir Box
1
Gravity Filter
Wash
Inlet
Effluent
Control
Slowdown Surge Tank
Electric Valve
FIGURE 4: Schematic, Pilot Plant System
-------
The pilot plant (Figure 5) was erected out-of-doors in an orientation
which differed only slightly from that shown in Figure 4. The feed
stream was pumped from the interior of a l/4"-mesh screened revolving
basket at the sewage plant weir box some 200 ft distant. This basket
rejected raw sewage solids of a size and type that otherwise would foul
the centrifugal feed pump, influent rate control valves and slurry dis-
charge valves. Center-driven at 7 rpm for self-cleaning, the basket
had 2.5 ft of submerged screen area and a 6-in. freeboard at normal
liquid level.
Two manually operated weir-type diaphragm valves in the feed line were
throttled to set the throughput rate. The upstream valve was throttled
to reduce the line pressure at the downstream valve which was adjusted
to give the desired rate as indicated by a magnetic flowmeter. Two
valves in series provided greater valve openings for sewage solids than
a single one. The resilient diaphragm construction of these valves also
provided a measure of self-clearing. Solids would pass that otherwise
would clog a conventional valve. Occasionally one or both valves would
require manipulation to dislodge solids.
The feed stream entered the inlet chamber of the upstream clarifier
(Unit #1) discharging below liquid level to avoid foaming. Unit #1
effluent flowed by gravity to the inlet chamber of the downstream
clarifier (Unit #2) and thence by gravity to the gravity anthracite-
sand filter for final treatment.
Carbon feed was countercurrent to the throughput flow. Fresh powdered
activated carbon was gravity fed by a volumetric slurry feeder to the
bottom end of the inner draft tube of Unit #2 along with a separate
feed of polymer flocculant. A second carbon feeder was used during
initial start-up to permit simultaneous development of slurry in both
units and as a standby for the other feeder.
Flocculated carbon slurry recirculating within Unit #2 was permitted to
accumulate to a manageable concentration which was controlled by timer-
actuated blowdown from the internal concentrator. The repeating-cycle
timer used afforded independent adjustment of both the valve-open and
valve-closed portions of the blowdown cycle from a fraction of a minute
each up to a 2-hr total cycle time.
Used-carbon blowdown from this unit was collected in the 10-gal.
polyethylene blowdown surge tank and, together with makeup flow of
semi-treated wastewater from mid-depth in Unit #1, transferred con-
tinuously to the bottom of the inner draft tube of Unit #1. A
stainless steel Jabsco Pump operating at low speed was used for the
slurry transfer after a centrifugal pump and hydraulic ejector system
was found to break up the floe. Polymer was also fed separately to
this region.
23
-------
FIGURE 5: Photograph of 28,000 gpd Pilot Plant
-------
Slowdown from Unit #1, also timer controlled, was to waste when not
being measured or sampled.
Unit #2 effluent could be divided so that any portion or all of it
could be wasted or directed to a 2-ft square by 6-ft high gravity
filter. The filter was constructed of Transite and Lucite with two
opposing transparent sidewalls allowing visual observation of the
contents. It included an extruded asbestos-cement underdrain system
supporting 11 inches of graded gravel, 16 inches of filter sand and
6 inches of anthracite coal. The filter sand had a specified effec-
tive size between 0.45 and 0.55 mm and a maximum uniformity coefficient
of 1.75. The anthracite coal had a specified effective size between
0.90 and 1.20 mm and a maximum uniformity coefficient of 1.7.
A float-operated effluent control valve was added to the filter to pace
filtration with plant flows up to 12 gpm. During higher plant operating
rates, effluent from the #2 Unit was wasted to limit the filtration rate
to 12 gpm (3 gpm/ft2 of filter area) or less. A transparent plastic
hose, tapped into the underdrain and extended to the top of the filter,
provided head loss indication during operation.
As a matter of convenience, locally available well water pumped from
a 2000-gal. storage tank was used for filter backwashing. The wash
water was batch-chlorinated in the storage tank with commercial bleach
or calcium hypochlorite prior to backwashing. The filter was back-
washed with chlorinated water every day for biological growth control
for the first 97 days. Backwashing at maximum head loss of 54 inches
was adopted for the remaining 186 days of operation. The wash rate
was set manually while observing the extent of sand-coal expansion.
Wash water consumption was computed from tank draw-down. Wash rate
information is presented elsewhere in this report as surface-wash
procedures varied during the program.
Sample taps were provided on each contactor-clarifier unit for slurry
sampling and determination of the slurry-clear water interface eleva-
tion. A specially designed multiple-dipper sampler, supplied with
continuous streams of Unit #1, Unit #2 and filter effluents auto-
matically composited samples of these streams. The dippers collected
aliquots at 1-minute intervals throughout the 24-hour compositing
periods. Samples of the feed stream were collected manually each hour
and sewage plant effluent samples were collected manually every two
hours for these 24-hour composites. The samples were accumulated in
polyethylene containers refrigerated to a temperature between 1° and
3° C.
The pilot plant program was conducted in a flexible fashion such that
sequential operating period objectives could be adjusted as results
became available. New data were reviewed frequently in consultation
with the Project Officer. The program has been divided into seven
consecutive phases:
25
-------
1. Brief operation of the system at 6.5 gpm (0.5 gpm/ft2) and a
carbon dosage of 145 mg/1 for equipment shakedown and operator
training. This flow was selected as a practical minimum dic-
tated by previous experience as well as difficulties associated
with controlling the flow rate of the solids-bearing feed
stream. Sampling and analytical work were phased in simul-
taneously. Operation continued through Operating Periods 1,
2, 3 and 4 during which time the flow rate was increased
incrementally to the maximum of 19.4 gpm (1.5 gpm/ft ) while
holding all other controlled variables constant.
2. Operation at 19.4 gpm and 145 mg/1 carbon feed during sequen-
tial reduction in polymer dosage to Unit #1 from 20 mg/1 to
6 mg/1, reduction in polymer dosage to Unit #2 from 15 mg/1
to 6 mg/1 and initiation of 10-mg/l alum dosage to Unit #1.
This phase consisted of Operating Periods 5A, 5B, 5C, 6A and
6B.
3. Brief operation duplicating Operating Period 6B except sub-
stituting Darco S-51 activated carbon for Aqua Nuchar A. The
initial Operating Period 7 was halted to modify the plant to
better manage the type of blowdown solids produced. Operating
Period 7A was with the modified system.
4. Operation again duplicating Operating Period 6B with Aqua
Nuchar A to bracket the S-51 run and monitor any significant
datum shift. The modifications added for Period 7A were re-
moved. This was Operating Period 8.
5. Sequential Operating Periods 9 through 11C during which the
flow rate was reset to 13.0 gpm and polymer dosages to both
units was varied from 4 mg/1 per unit to 6 mg/1 per unit as
alum dosage to Unit #1 was varied from 0 to 25 mg/1 and alum
dosage to Unit #2 was studied at 0, 25 and 20 mg/1.
6. Operation at 13.0 gpm (1.0 gpm/ft2) through Periods 12A, 12B
and 12C attempted to avoid the use of polymer while feeding
alum at dosages of 20 mg/1 to 50 mg/1 per unit. Polymer at
dosages of 2 mg/1 to 3 mg/1 per unit was required.
7. Operation at 13.0 gpm with carbon dosages of 200 mg/1, 250 mg/1
and 100 mg/1 was studied during Operating Periods 13, 14 and
15. Polymer and alum were fed to both units in amounts found
necessary for acceptable clarification.
During each operating period, composite and grab samples were collected
for development of laboratory isotherms and analysis sufficient to de-
fine process performance, i.e., pH, color, turbidity, suspended solids,
26
-------
filtrable and unfiltered COD, filtrable and unfiltered TOC, unfiltered
BODc; and slurry blowdown solids concentrations. Other data recorded
included quantities and types of treatment chemicals; plant flow; rate
and volume of blowdown; slurry volume after 5 minutes of settling;
slurry level; filter head loss; and the filter backwash frequency, rate
and duration. Still other data were accumulated on a lesser scale and
are discussed separately. Records were also made of events which inter-
fered with or otherwise affected plant operation or performance.
Fourteen-day runs at optimum conditions for the situation under test
were programmed to avoid biasing data because of weekly cycling of the
feed characteristics. In fact, few of the periods went 14 days or
longer as physical treatment criteria under study were frequently demon-
strated in less time. Other periods were halted prematurely by mishap
or the number of days of data originally collected during a run was re-
duced later by discarding after reviewing records of operation and
events. Continuous plant operation was undertaken to parallel the
reference sewage treatment plant situation and also to avoid potential
biological activity interferences that might occur with intermittent
operation.
Experimental Results
A sizeable quantity of raw analytical and operational information was
accumulated during the pilot plant operation. Except when presented
in example Figures and Tables, raw data are omitted from this report
for the sake of brevity. Mean-data values were computed for individual
operating periods and are the basis for experimental results reported.
Probability analyses were not attempted as too few bits of information
were retained from some periods to justify this approach.
Evaluation of the period-mean results is aided by graphical presenta-
tion of the measured variables vs. operating period from beginning to
end of the field program.
2
Phase 1 - Initial operation of the pilot plant at 0.5 gpm/ft clarifier
overflow rate continued for 28 days. Operator training, equipment
shakedown, initiation of sampling and analytical laboratory programming
consumed the early portion of this period. Valid information was col-
lected during the later portion. Table 2 presents the program followed
and results obtained for this and the other operation periods compris-
ing Phase 1. A high polymer feed was used to assure flocculation and
145 mg/1 Aqua Nuchar A (ANA) activated carbon feed was the computed
dosage required to achieve 50 percent filtrable COD removal by two-stage
countercurrent adsorption with assumed values of Co = 100 mg/1 COD and
1/n = 1.52. These assumptions originated from the many grab-sample
isotherms developed earlier.
27
-------
TABLE 2
Phase 1, Pilot Plant Performance by Operating Period, Mean Values
Operating Period
CONTROLLED VARIABLES
Overflow Rate, gpm/ft^
Carbon Feed, mg/1 ANA
C-31 Feed,- mg/1 Unit #1
Unit #2
Alum Feed, mg/1 Unit #1
Unit #2
No. of Days Sampled
No. of Operating Days
PILOT PLANT REMOVALS
TOC unfiltered, %
TOC filtrable, %
COD unfiltered, %
COD filtrable, %
BOD5 unfiltered, %
Sus. Solids, %
SEWAGE PLANT REMOVALS
TOC unfiltered, %
COD unfiltered, %
BODs unfiltered, %
Sus. Solids, %
1
0.5
145
20
20
0
0
17
28
83.8
63.1
90.6
65.8
92
98.6
82.9
87.3
81
95.1
2
0.75
145
20
15
0
0
10
13
86.1
64.8
91.0
69.6
89
96.7
72.5
78.0
66
84.0
3
1.
145
20
15
0
0
7
7
83.
55.
90.
61.
87
96.
84.
88.
92
91.
0
1
5
0
0
2
5
8
6
4
1.50
145
20
15
0
0
18
27
86.6
66.1
89.6
69.5
90
97.8
90.1
91.0
88
95.2
-------
Filtrable COD of the feed stream exceeded 100 mg/1 (Figure 6) during
the entire Phase; however, removals between 61.0 and 69.6 percent were
recorded. Pilot plant effluent COD was less than the 50-mg/l level
suggested as the original target value. Filtrable TOG removals were
slightly lower, ranging from 55.5 to 66.1 percent. Filtrable COD
carbon loading (X/M) reached 54 percent during Periods 2 and 4.
Blowdown from Unit #1 was excessive at start-up amounting to 13 percent
of the feed stream. Hydraulic currents were scouring the slurry con-
centrators so the concentrator spouts were eliminated by filling them
with grout. Slurry would now feed to the concentrators by decanting
from the upper reaches of the slurry pool and by outer draft tube dis-
charge impinging upon the filled spout and flowing horizontally over it.
Blowdown from Unit #1 decreased to 2.5 percent after restarting the
plant. Unit #2 blowdown decreased from 3.2 to 1.0 percent of throughput
as a result of the alteration.
From the onset of operations it was obvious the raw sewage solids cap-
tured in Unit #1 slurry would require discharging 2 to 3 times more
blowdown from this unit than from Unit #2. As the operating rate was
increased, dewatering of the slurry in the concentrator diminished
because of the shortened retention times. This situation was aggra-
vated by slurry deflocculation originating in the centrifugal pump and
ejector carbon-advance system such that, in Operating Period 4 at the
1-5-gpm/ft overflow rate, slurry could not be conditioned well enough
for control. The slurry interface rose well above the tap in Unit #1
that supplied water to the centrifugal pump suction and as make-up
water to the carbon advance surge tank. In effect, the system was re-
cycling up to 7.5 gpm of damaged floe through Unit #1. This system was
replaced with the resilient-impeller pump described earlier, for the
data collecting portion of Period 4.
The diurnal variation in feed suspended solids caused the slurry inven-
tory to cycle widely each day. Draft tube slurry concentration
(Vol %/5 min settling time) would build steadily through the afternoon
and evening hours and decrease quite rapidly in the morning hours. The
program to maintain approximately 10 to 15 Vol %/5 min throughout the
day required frequent adjustment of the blowdown-system timers. As
experience was accumulated, the program shifted to permitting the
normal cycle to occur with timer adjustments made only if the tests
indicated significant deviation from the expected value for that time
of day. Generally the cycle range for Unit #1 was between 5 and 15
Vol %/5 min. The cycle range in both units could be varied, however,
as the proportions of, and types of, coagulants were varied. During
Phase 1, Unit #2 draft tube solids varied only 1 to 2 Vol %/5 min per
day which approaches the reproducibility limits of the test. There
was a tendency to ignore a gradually increasing or decreasing trend
overly long and then overcorrect timer adjustment. Figure 7 is repre-
sentative of draft tube slurry data field records for this period.
29
-------
120 -
in
Operating Period
FIGURE 6: Mean Values of Filtrable COD
-------
a
•H
30
25
20
1-1
H
0)
,0
3
10
Phase 1, Period 3, Unit #1 Coagulant dosage ample for feed-solids cycles
Q = 13.0 gpm
Unit //I, 20 mg/1 C-31
Unit #2, 15 mg/1 C-31
Fri, 4/17 Sat, 4/18
Sun, 4/19
Mon, 4/20 Tues, 4/21
FIGURE 7: Diurnal Variation, Clarifier Slurry Settling Tests, Phase 1
-------
Having established pilot plant operation to the hydraulic capacity
limits of the design while simultaneously delivering an effluent com-
parable to the parallel sewage treatment plant this portion of the
program was terminated.
Phase 2 - At the quoted price of 3l£/lb for C-31 in tankcar quantities,
Phase 1 treatment cost 90/1000 gal. of sewage treated for polymer alone.
Performance of the pilot plant at reduced polymer dosages, augmented
with a minimum effective alum dosage, was investigated over the next
35 days during Operating Periods 5A, 5B, 5C, 6A and 6B. Overflow rate
was held at 1.5 gpm/ft^ and activated carbon feed was maintained at
145 mg/1 ANA.
Physical performance of the clarifiers had been marginal. Adversely
affected by factors already reported, plus several power outages,
storm flooding of the inlet screen and other incidents normal to plant
shakedown, the pilot plant had demonstrated that relatively high sus-
pended solids carryover could be tolerated. Clarifier effluent quality,
in terms of carryover, could be sacrificed if cost reduction could be
demonstrated. Arbitrary acceptable carryover levels of 50 mg/1 sus-
pended solids for Unit #1 and 25 mg/1 suspended solids for Unit #2 were
adopted as guide lines. Figure 8 shows the suspended solids relation-
ships for the various sample streams for the program.
The stepwise reduction in polymer to 6 mg/1 per unit while feeding
10 mg/1 alum to Unit #1 is listed in Table 3 along with mean performance
data for these five periods. Figure 8 shows solids carryover from the
clarifiers was within limits; however, feed solids also decreased during
this time (May and June) such that coagulant requirements would be
naturally lower. Filtrable TOG and COD of the feed dropped signifi-
cantly during the same period (Figures 6 and 9); however, removal per-
centages tended to improve slightly (Figures 10 and 11). These Figures
also show the dramatic shift in removal capability that occurred in the
previous phase as the result of the improved carbon advance system.
Unit #1 was now responsible for approximately 75 percent of the adsorb-
ate removal as compared to 55 percent previously. As the feed stream
TOC and COD diminished during Phase 2, lowering the adsorption poten-
tial in the upstream clarifier, the adsorbate removal division equalized
substantially.
Five-day BOD removals in the tables represent order-of-magnitude values
only. Insufficient determinations were made to establish accurate mean
information. Five-day BOD determinations (Figure 12) were usually done
twice weekly.
Clarifier blowdown remained reasonably constant at 6.5 to 7.0 percent
for Unit #1 and 3.0 percent for Unit #2 throughout the period (Figure
13). Blowdown samples contained twice the suspended solids measured in
the recirculating slurry. Unit #1 slurry averaged 2.63 g/1 vs. blowdown
32
-------
300
250 -
oo
e
.200
05
TJ
•H
tH
O
to
giOO
a
CO
to
50
Unit #1 Effluent
Sewage Plant Effluent
frjHHP:y:-»^iP*f"TBfr*
-------
TABLE 3
Phase 2, Pilot Plant Performance by Operating Period5 Mean Values
CONTROLLED VARIABLES
Overflow Rate, gpm/ft2
Carbon Feed, mg/1 ANA
C-31 Feed, mg/1 Unit #1
Unit #2
Alum Feed, mg/1 Unit #1
Unit #2
No. of Days Sampled
No. of Operating Days
PILOT PLANT REMOVALS
TOG unfiltered, %
TOG filtrable, %
COD unfiltered, %
COD filtrable, %
BOD5 unfiltered, %
Sus. Solids, %
SEWAGE PLANT REMOVALS
TOG unfiltered, %
COD unfiltered, %
BOD5 unfiltered, %
Sus. Solids, %
5A
1.5
145
20
15
10
0
5
5
82.2
60.5
89.6
62.7
91
97.6
84.8
90.0
91
96.6
5B
1
145
15
15
10
0
4
4
86
66
90
67
94
98
84
86
94
93
Operatin;
5C
.5
.5
.8
.3
.1
.3
.1
.1
.2
1
145
12
12
10
0
16
16
88
61
88
66
89
97
89
90
89
96
g Period
6A
.5
.5
.5
.5
.8
.7
.6
.4
.0
.0
.6
1
145
12
6
10
0
4
4
91
62
89
67
-
98
90
89
-
96
.5
.5
.8
.3
.7
.2
.3
.7
.6
.5
6B
1
145
6
6
10
0
6
6
88
64
85
62
91
95
88
88
90
95
.5
.5
.0
.2
.8
.4
.8
.5
.4
34
-------
50
1 I ^ I
Feed
Ui
40
30
o
o
H
.0
n)
20
10
y—Filter Effluent
Unit #2 Effluent
ioO CT> O OO O i—l rH
rH r-< rH rH iH rH
Operating Period
^ PQU
CNl CNlCsl
FIGURE 9: Mean Values of Filtrable TOC
-------
LO
ON
100
90
80
70
60
0)
§ 50
01
o
H 40
0)
0)
M-l
O
,0
Cfl
30
20
10
-------
Combined Clarifier Units Only
Complete Pilot Plant
Unit #1
Carbon
Advance
Modified
Operating Period
FIGURE 11: Mean Values of Filtrable COD Removal
-------
250
Feed
200
150
LO
00
M
B
o
o
FQ
100
Unit //I Effluent
Sewage Plant Effluent
Unit #2 Effluent
50
Filter Effluent
Operating Period
FIGURE 12: Mean Values of BODC
-------
u>
VO
0)
01
En
O
>
O
1
O
14
12
10
Clarlfier Overflow,
-------
of 6.15 g/1. Unit #2 slurry averaged 4.29 g/1 vs. blowdown of 8.52 g/1.
Independent dewatering tests (Figures 14 and 15) had already confirmed
slurry solids content could be increased by a factor of 10 with 10 min-
utes settling time. The pilot plant concentrators were undersized for
the quality and quantity of floe produced but physical alterations were
deferred in anticipation of future operations at reduced flow rates.
During Operating Period 5C gravity filter operation was revised. Hereto-
fore, filter runs had been purposely terminated at 24 hours for back-
washing with growth-inhibiting chlorinated wash water. Modest amounts
of algae and slime could be seen on the supporting gravel visible through
the Lucite side walls. Covering these areas with opaque aluminum sheet
had limited algae growth to a static condition. Filter effluent turbid-
ities had persisted at 10 to 13 JTU (Figure 16). Examination, as in
previous studies^, indicated the material was not powdered carbon but
the products of biological growth. The mat left on filter discs was
green-brown colored.
n
Filter head loss at a surface loading of 2.6 gpm/ft increased from a
clean filter reading of 5 in. to 33 in. at the end of 24 hours. Volume
filtered per day was 15,000 gal. with the balance of clarified liquid
lost through blowdown and the filter bypass system.
After the procedure was revised to extend filter runs to the maximum
head loss of 54 in. (operator duties permitting), runs averaged 28, 31
and 30 hours for Operating Periods 5C, 6A and 6B, respectively. Head
loss was recorded hourly throughout the program (Figure 17). Mean
filter effluent turbidity increased to 19 JTU for Period 6B. Night-
time backwashing (during which it was difficult to distinguish the
coal-wastewater interface), extended periods between chlorination and
higher bed differentials all are assumed to contribute to the solids
discharge.
The mean washwater consumption was 647 gal. per wash or 4.3 percent of
the volume filtered at this filter rate. In backwashing, the operators
were instructed to brush down the filter freeboard as the unit was de-
canting to wash level. A low upflow rate, sufficient to barely expand
the 6-in. coal bed, was established and the coal raked vigorously to
dislodge the accumulated carbon-polymer floe. This was followed by a
5-minute or longer backwash at 36 percent (8 in.) bed expansion. The
stored wash water was warm, up to 90°F at times, and instantaneous
wash rates of 20 gpm/ft were recorded.
Raking the filter media simulated surface washing, a step deemed essen-
tial to eliminate carbon-polymer-filter media mud balls noted in earlier
studies. No mud balls developed.
40
-------
o
>
Q)
(-J
S-l
3
4J
4-1
0)
Phase 1, Period 4
5/11/70
Phase 2, Period 5C
6/3/70
Phase 2, Period 5C
6/10/70
I
Unit #1
I
•Unit #1
Jnit #2
0 5 10 15 20 25 30
Settling Time, Min
0 5 10 15 20 25 30
Settling Time, Min
0 5 10 15 20 25 30
Settling Time, Min
FIGURE 14: Recirculating Carbon Slurry Settling Rate Curves, Part 1
-------
O
>
rH
O
S-l
3
iH
CO
13
0)
100
90
80
70
60
50
40
30
20
10
Phase 3, Period 7A
7/14/70
0
Phase 5, Period 11B
9/15/70
I
0 5 10 15 20 25 30
Settling Time, Min
Unit #1
Phase 7, Period 13
11/4/70
T
T T
0 5 10 15 20 25 30
Settling Time, Min
I
Unit #1
-Unit #2
I
0 5 10 15 20 25 30
Settling Time, Min
FIGURE 15: Recirculating Carbon Slurry Settling Rate Curves, Part 2
-------
250
200
3 150
JD
3
H
01
6
o
a.
o
O
0)
ex.
100
50
Unit //I Effluent
Sewage Plant Effluent
~_.-v /;
Filter
^\
u < M
< U Q W < PQ
O O O O r-l ^H
< PQ O
Operating Period
FIGURE 16: Mean Values of Spectrophotometer Turbidity
-------
CO
0)
o
G
CO
co
o
J-J
TD
cfl
CO
J-i
O
K 30 -
Wed, 6/17
Thur, 6/18
Fri, 6/19
Sat, 6/20
FIGURE 17: Typical Gravity Filter Head Loss Curves
-------
Phase 3 - The pilot plant was operated for 10 days at conditions estab-
lished in the previous period except Darco S-51 powdered activated carbon
was substituted for Aqua Nuchar A. Operating Period 7 began by changing
the carbon in the carbon feeder and allowing the existing carbon slurry
to work out of the system in a normal fashion. As anticipated, the floe
settleability changed radically. Mean values for blowdown given in
Figure 13 do not reflect the progressive reduction in blowdown volume.
At the end of Period 7 blowdown rates were 2.7 and 0.9 percent, respec-
tively, for Units #1 and #2.
Product water quality was unchanged (Table 4) but the S-51 carbon
altered the slurry produced in Unit #2. The slurry, as dewatered in
the concentrator, became so dense it would not discharge consistently
through the 1/2" blowdown valve and piping under the 5 ft of head
available. Control of blowdown became erratic and by the middle of
the third day the recirculating slurry solids had built up to 18 Vol
%/5 min. Unit #2 was manually purged resulting in the loss of most of
the slurry. Some blowdown must be maintained to sustain carbon advance
so it required another three days, with a blowdown of 0.3% of through-
put, to refill the concentrator and reestablish the recirculating slurry
inventory. A 15-hour power failure precipitated another cycle of the
blowdown problem so the run was terminated. The high solids blowdown
is desirable, but it was not compatible with the scale of the pilot
plant blowdown system.
Gravity filter runs averaged 35 hours for the period.
Operating Period 7A was a successful 12-day test using Darco S-51 and
feeds of 6 mg/1 C-31/unit and 10 mg/1 alum to Unit #1 at the 1.5 gpm/ft2
overflow rate. The concentrator in Unit #2 was bypassed by extending
the internal piping to draw blowdown directly from the recirculating
slurry. Bypassing the downstream concentrator required increasing the
blowdown to 5.2 percent from the unit. Unit #1 blowdown was constant
at 3.5 percent, the lowest mean rate recorded during the program except
when treating less than 13.0 gpm (1.0 gpm/ft2 overflow rate).
Filtrable-TOC and -COD removals were 59.5 and 68.0 percent, respectively,
somewhat lower than for Period 7. Removals are higher if based upon
Unit #2 effluent rather than filter effluent as filtrable-TOC and -COD
levels increased measurably through the sand filter. This is especially
evident after filter operation was changed in Period 5C. Determina-
tions on filtered samples do not measure all of the adsorbate-indicator
material going into the filter. A portion is particulate matter or is
adsorbed upon the carbon carryover removed by the laboratory filtration.
Some of this material appears to be solublized or desorbed in the filter
and is measured in the filter effluent determinations. Filter runs
averaged 34 hours for the period.
45
-------
TABLE 4
Phase 3 & 4, Pilot Plant Performance by Operating Period, Mean Values
CONTROLLED VARIABLES
Overflow Rate, gpm/ft^
Carbon Feed, mg/1
C-31 Feed, mg/1 Unit #1
Unit #2
Alum Feed, mg/1 Unit #1
Unit #2
No. of Days Sampled
No. of Operating Days
PILOT PLANT REMOVALS
TOC unfiltered, %
TOC filtrable, %
COD unfiltered, %
COD filtrable, %
BOD5 unfiltered, %
Sus. Solids, %
SEWAGE PLANT REMOVALS
TOC unfiltered
COD unfiltered, %
BOD5 unfiltered, %
Sus. Solids, %
7
1
S-51,145
6
6
10
0
7
10
90
67
87
71
91
96
88
88
92
97
Phase 3
Operating Period
7A
.5 1.5
S-51,145
6
6
10
0
10
12
.7 89.0
.7 59.5
.7 87.4
.1 68.0
89
.6 96.1
.8 87.8
.8 89.0
90
.0 97.6
Phase 4
8
1.5
ANA, 145
6
6
10
0
12
13
90.3
70.7
87.1
71.1
89
96.5
87.5
88.7
93
95.7
46
-------
The division in carbon loading between units was substantial during
Period 7 when the carbon advance stream averaged 0.2 gpm. Loading
was equal in the units during Period 7A with the carbon advance stream
increased to 1.0 gpm. The volume of carrier liquid in the downstream
blowdown constitutes a recirculating stream that dilutes the applied
adsorbate concentration in the upstream contactor. A higher adsorbate
concentration overflows to Unit #2. The greater the recycle the lower
the loading potential becomes in Unit #1 and vice versa.
Phase 4 - The plant was operated for 13 days for Operating Period 8
repeating the conditions of the last three runs but again using Aqua
Nuchar A as the adsorbent. The Unit #2 concentrator was returned to
service.
In comparing the performance of the two carbons as given in Table 4,
Aqua Nuchar A produced the higher filtrable-TOC and -COD removals,
70.7 and 71.1 percent, respectively. Examination of any of several
figures, say Figure 11, however, reveals the performance differential
was in the gravity filter. Filtrable-COD removal in the carbon con-
tacting units, filter excluded, was 71.1 percent for both Periods 7A
and 8. The same samples indicated filtrable-TOC removals of 65.7 and
70.4 percent, but these differentials are based upon analytical differ-
entials approaching the limits of test accuracy.
Perhaps the 28-hour average filter runs for Period 8 contributed to
improved filter performance. Wash water consumption was 5.0 percent
primarily because operators did not work the wash cycle in a timely
fashion.
Slurry blowdown rates with Aqua Nuchar A carbon were reestablished at
4.9 and 1.5 percent for Units #1 and #2, respectively.
As indicated in the earlier laboratory isotherm studies, there appeared
to be little or no difference in adsorptive capacity between the carbons
under the test conditions.
The relationship between the suspended solids content and the 5-minute
settled volume of the recirculating slurries varies with the types and
quantities of chemicals fed as well as with the character of the feed
stream. The 5-minute settled volume test is an important clarifier
operating parameter used to judge blowdown requirements. The suspended
solids determinations can be used to estimate the carbon inventory in
the system. Figures 18 and 19 show these relationships for slurries
from the two clarifiers. The curves on these figures illustrate the
effect of different chemical feed situations and indirectly imply that
a wide range in possible carbon inventory was experienced during the
program. In general, Unit #1 slurry was composed of about 2 parts raw
sewage solids to 1 part of carbon based upon feed solids and carbon feed
figures. This minimized the effect of carbon type in Unit #1 (Figure 18)
47
-------
-p-
00
0)
a
•H
H
a
•H
Q)
CO
o
>
s-l
n
rH
CO
0)
,0
H
ca
Q
i-H
•H
!=>
20
15
10
Maximum Alum Feed
40 mg/1 Alum
No C-31
Darco S-51 Trend
10 mg/1 Alum
6 mg/1 C-31
Circled Numbers Indicate Operating Period
I
2345
Unit #1 Draft Tube Slurry Suspended Solids, g/1
FIGURE 18: Solids Content of Recirculating Slurry, Unit #1
-------
25
0)
a
H
GO
•5 20
0)
CO
•H
s
iC 15
O
>
10
H
4-J
-------
as indicated by the points for Periods 7A (S-51) and 8 (ANA). In
Unit #2 (Figure 19), the Period 7A points stand apart from the others,
possibly resulting from the atypical blowdown situation for this period
only. This altered the dilution and recirculation ratios existing in
the mixing and flocculation zones of the unit.
For much of the program the activated carbon inventory was estimated
to be 7 to 9 Ib with the countercurrent flow-through time being about
6 hours while operating at 1.5 gpm/ft2. Although 5-minute settling
test values averaged about 10 percent volume in both units and both
contained about 225 gal. of slurry, the carbon was not divided evenly
between units. Unit #1 slurry had a lower suspended solids content,
say 2.5 g/1, and only 37 percent of it could be considered carbon.
In Unit #2 with about 4 g/1 suspended solids content, close to 78 per-
cent of it could be carbon. Including concentrator contents Unit #1
probably operated with 2 Ib and Unit #2 with 6 Ib of the inventory.
Settling rate curves (Figure 15) were not too helpful in discriminating
between the ANA and S-51 carbons. Both slurries settled so rapidly
that points could seldom be obtained before hindered settling commenced
at about 1.5 to 2.0 minutes into the test. The settling solids set up
density currents which, along with residual currents from the initial
sample mixing, obscured the liquid-solids interface during the crucial
1st minute of the test. In the 250-ml graduated cylinder (10.25 in.
calibrated depth) used for the tests, Unit #1 slurry was estimated to
have a free subsidence rate in excess of 4 in./min or greater than
2.5 gpm/ft overflow rate. Unit #2 slurry was estimated to settle at
twice that velocity.
Phase 5 - The program proceeded with 53 days of operation divided into
9 separate Operating Periods, each utilizing a variation in coagulant
dosage. Operating Period 10B was a 1-day transition period and data
for that day were discarded. Operational information for the other
8 periods is presented in Table 5.
Period 9 was a test of operation using 4 mg/1 C-31 per unit to estab-
lish the minimum dose for effective flocculation of the carbon. During
Period 9, the suspended solids carry-over from both units nearly doubled
and the amount of uncoagulated raw-sewage colloidal material passing
the filter more than tripled the values reported during the preceding
work with 6 mg/1 C-31 per unit. This is shown clearly in Figure 8.
Other parameters were less affected as adsorption undoubtedly con-
tinued at established efficiency whereas analytical determinations
involving 0.45-micron filtration reflect the passage of uncoagulated
particulate matter. Lab-filtered samples contained discernible
turbidity.
50
-------
TABLE 5
Phase 5, Pilot Plant Performance by Operating Period, Mean Values
CONTROLLED VARIABLES
Overflow Rate, gpm/ft^
Carbon Feed, mg/1 ANA
C-31 Feed, mg/1 Unit #1
Unit #2
Alum Feed, mg/1 Unit #1
Unit #2
No . of Days Sampled
No. of Operating Days
PILOT PLANT REMOVALS
TOC unfiltered, %
TOC filtrable, %
COD unfiltered, %
COD filtrable, %
BOD5 unfiltered, %
Sus. Solids, %
SEWAGE PLANT REMOVALS
TOC unfiltered, %
COD unfiltered, %
BOD5 unfiltered, %
Sus. Solids, %
9
1
145
4
4
10
0
5
5
87
60
83
59
-
88
89
90
-
93
10A
.5
.4
.6
.0
.7
.5
.6
.6
.9
1
145
4
4
20
0
6
6
90
66
85
69
80
95
90
90
86
97
.0
.0
.7
.3
.3
.1
.8
.8
.9
Operating Period
IOC 10D 10E 11A
1.0
145
4
6
20
0
4
5
88.8
60.8
83.2
64.4
89
93.3
90.1
88.1
87
97.5
1.0
145
4
6
25
0
7
8
88.8
52.3
81.8
54.3
83
92.7
86.3
84.7
93
93.7
1.0
145
6
6
0
25
5
7
89.5
67.3
86.7
65.5
90
96.6
87.4
86.5
91
97.5
1
145
6
6
10
25
6
7
91
73
90
70
95
97
88
88
91
97
.0
.3
.6
.2
.9
.1
.2
.9
.1
11B
1.0
145
6
6
10
20
6
7
92.1
71.1
90.0
69.0
93
96.0
89.0
89.5
90
97.4
11C
1.0
145
6
6
15
20
6
7
92.2
70.1
90.6
70.2
91
98.8
90.2
89.5
86
98.3
1 day period 10B was transitional and was discarded
51
-------
2
The overflow rate was decreased to 1.0 gpm/ft for the balance of the
program at this time. This reduced the rate of chemical consumption,
eliminated nuisances associated with hydraulics at the no-safety-factor
plant design capacity and brought solids production more in line with
slurry concentrator capacity.
Filter alum precipitates 0.263 parts of inert solids (aluminum hydroxide)
for each part fed. If alum equal to 10 percent of the carbon dose was
fed to a system utilizing carbon regeneration and reuse with a 15 per-
cent carbon loss per cycle, the inert-solids fraction of the mixture
would build up to 14.9 percent. (If the 15 percent loss is defined as
solids mixture rather than only carbon, the inert-solids fraction would
approach 17.1 percent.) Should the inert aluminum hydroxide be detri-
mental to regeneration its use might better be avoided. This would be
true only if the alum interference was specific because the screened
and degritted sewage feed stream was found to contain 65, 52 and 80 mg/1
fixed suspended solids on composite samples having 204, 204 and 272 mg/1
total suspended matter, respectively. These quantities, with the 145
mg/1 carbon feed of the pilot plant, produce a mixture that would be
31.0, 26.4 and 35.6 percent inert solids in one pass through the plant.
The slurry for a regeneration-reuse system with 15 percent carbon loss
per cycle would equilibrate at 75.0 percent inert material for the
65-mg/l sample. A modest alum feed would not be the predominant source
of inert solids in such a system. A decision was made to study higher
alum dosages supplanting portions of the polymer dosage.
For Operating Period 10A, the alum dosage to Unit #1 was increased from
10 mg/1 to 20 mg/1 with other feeds held constant. Filtrable-TOC and
-COD removals improved as the filtrable suspended matter was reduced.
Suspended solids carryover decreased to acceptable levels for both units
but solids (12 mg/1) and turbidity (18 JTU) passing the gravity filter
remained undesirably high (Figures 8 and 16). The slurry in Unit #2
was gradually deteriorating so the polymer feed was increased from
4 mg/1 to 5 mg/1 (Period 10B). After 24 hours little improvement was
noted so the polymer feed was increased again to 6 mg/1 (Period IOC) .
This was a time (August) of weak raw sewage and although pilot plant
effluent characteristics were little changed, percent removals decreased.
Carbon loading in the clarifiers dropped to 8.4 percent as TOC and 31.0
percent as COD.
Slowdown was 4.2 percent for Unit #1 and 0.8 percent for Unit #2 re-
flecting the influence of the reduced treating rate.
Gravity filter runs averaged 34 hours during Period 10A prior to floe
deterioration then deteriorated to 20 hours. They lengthened to
29 hours for Period IOC. Wash water consumption was 4.3 percent with
a filter surface loading of 2.75 gpm/ft2 for this later period.
52
-------
The alum dosage was set at 25 mg/1 to Unit #1 for Period 10D. Operation
continued at a submarginal condition with Unit //I carryover increasing
beyond the 50 mg/1 suspended solids limit. Solids and turbidity passing
the filter remained undesirably high. Carbon loading, as a result of
low feed stream strength and filtrable solids problems was 7.0 percent
as filtrable TOG and 28.1 percent as filtrable COD, the lowest en-
countered at this carbon dosage.
Slowdown and gravity filter operation remained essentially unchanged
from the previous period.
The 25 mg/1 alum feed was switched to Unit #2 and the polymer feeds
reset to 6 mg/1 per unit for the 7-day Operating Period 10E. All re-
moval figures improved considerably as filter effluent suspended solids
and turbidity decreased to 7 mg/1 and 13 JTU, respectively, and feed
stream strength began recovering.
Carbon loadings in the combined clarifiers went to 9.5 percent as
filtrable TOG and 31.5 percent as filtrable COD.
Slowdown requirements shifted substantially to 6.1 percent for Unit #1
and 2.7 percent for Unit #2 with its alum-floe load now imposed upon
both clarifiers.
The 20 mg/1 of alum-coagulated solids carryover to the gravity filter
was within limits but the filter runs decreased to 19 hours. With the
shift in flocculation potential the slurry inventory in Unit #1 re-
mained more uniform. Solids escaping the 6 mg/1 polymer in this unit
during peak solids influx periods of the day were captured in Unit #2
causing its slurry inventory to cycle diurnally.
The phase was completed with 3 periods, 11A, 11B and 11C, of one-week
duration each, studying the use of alum in both clarifiers. Table 5
and the several figures show this to be a highly satisfactory operation
with all removals as high or higher than for any previous period. Feed
stream strength was increasing.
Filtrable-TOC and -COD loadings in the clarifiers increased to 11.4 and
37.2 percent, respectively.
The greater total alum dosage (30 to 35 mg/1) doubled the blowdown rate
to 12.4 percent for Unit #1. Unit #2 blowdown decreased to 1.5 percent
from a high of 2.8 percent as a result of the greater solids capture in
Unit #1.
Gravity filter operation improved. Filter runs lengthened to 36-hour
average duration and wash water consumption decreased to 2.5 percent
of the volume filtered.
AWbt_KC LIBRARY U.S. EPA
53
-------
The slurry solids inventory relationships for part of Period 11A are
reflected in the field record of draft tube slurry volume tests, Figure
20. During the low solids influx period of Saturday and Sunday; the
inventory in the units cycled about equally; however, by Sunday evening
the increasing load plus changes in the character of the solids reached
a state wherein the potential for solids capture, i.e., the coagulation
system capacity, for Unit #1 was exceeded. The inventory equilibrated.
Uncaptured solids passed on to Unit #2 causing wide diurnal cycles
Monday and Tuesday in spite of numerous blowdown timer changes to limit
the extremes.
Phase 6 - The pilot plant was operated for 29 days with alum as the
primary coagulant. Dosages of 20 mg/1 per unit were inadequate so they
were increased stepwise per Table 6 to 40 mg/1 to Unit #1 and 50 mg/1
to Unit #2 followed by readjustment finally to 50 mg/1 to Unit #1 and
25 mg/1 to Unit #2.
Performance for the initial two days was acceptable; however, these
data were discarded as the existing polymer-bearing floe was being purged
from the system. As this occurred, uncoagulated carbon began discharging
in the clarifier effluents. Filter runs became as short as 3 hours, with
carbon penetrating the media. Filter operation was suspended intermit-
tently while alum feeds were increased and, ultimately, polymer was added
to restore floe condition.
r\
The plant could not be operated successfully at the 1.0 gpm/ft overflow
rate without polymer. The bulky alum floe overwhelmed blowdown capa-
bilities, filled the clarifiers to the launder elevation and decanted
downstream in large amounts. Polymer restored slurry control to a de-
gree allowing some reduction in blowdown.
The concentrator in Unit #1 was enlarged to 15 gal. capacity for the
final operating period of this phase.
The data are shown in the several Figures; however, they are presented
in a qualified sense that not enough determinations could be retained
to present a representative picture. In addition, the excessive and
variable blowdown altered the carbon/feed-stream ratio beyond reasonable
reconciliation. This later condition is believed responsible for the
higher indicated removals in this phase.
Phase 7 - The pilot plant had performed through most of the previous
operating periods with 145 mg/1 activated carbon feed at efficiencies
closely approximating those for the sewage treatment plant. Figures
21, 22, 23 and 24 suggest, aside from problem periods, that both systems
were governed by a common denominator despite their process differences.
The pilot plant was operated at increased carbon dosages of 200 mg/1
for 21 days and 250 mg/1 for 13 days comprising Operating Periods 13
and 14 during October and November. This was followed 11 weeks later
54
-------
30
25
20
c
•H
a
o
>
1 *
M 15
5-1
0)
,0
H 10
rt
M
Q
Phase 5, Period 11A, Unit #1 Coagulant dosage inadequate for feed-solids cycle
Unit #2. 25 mg/1 Alum and 6 mg/1 C-31
Unit #1, 10 mg/1 Alum and 6 mg/1 C-31
Fri, 9/4 Sat, 9/5
Sun, 9/6
Mon, 9/7 Tues, 9/8
FIGURE 20: Diurnal Variation, Clarifier Slurry Tests, Phase 5
-------
TABLE 6
Phase 6, Pilot Plant Performance by Operating Period, Mean Values
Operating Period
CONTROLLED VARIABLES
Overflow Rate, gpm/ft^
Carbon Feed, mg/1 ANA
C-31 Feed, mg/1 Unit #1
Unit #2
Alum Feed, mg/1 Unit #1
Unit #2
No. of Days Sampled
No. of Operating Days
PILOT PLANT REMOVALS
TOG unfiltered, %
TOC filtrable, %
COD unfiltered, %
COD filtrable, %
BOD5 unfiltered, %
Sus . Solids, %
SEWAGE PLANT REMOVALS
TOC unfiltered, %
COD unfiltered, %
BOD5 unfiltered, %
Sus. Solids, %
12A
1
145
0
0
20-30-40
20-30-40-50
2
7
93
76
90
70
88
93
88
87
65
93
.0
.6
.9
.7
.2
.5
.4
.8
.0
12B
1.0
145
0
0-3-0-3
40
50
3
7
94.0
79.4
91.7
80.3
90
98.7
87.0
85.7
87
97.9
12C
1.
145
0-1-2
3-2-3
40-50
50-25
5
15
93.
80.
90.
78.
96
97.
85.
84.
94
89.
0
6
2
2
1
0
3
0
3
Hyphenated chemical dosage data indicate changes instituted during period
56
-------
110
100 —
30
8
•Sewage Plant Effluent
o
H
20
Pilot Plant Effluent
10
V^SeHOc>kX^-^^'l™~'X'li-*'^
I I I I I I I I I I
Operating Period
FIGURE 21: Mean Values of TOC
-------
500
400
300
Ln
00
§
CJ
200
100
Sewage Plant Effluent
Pilot Plant Effluent
Operating Period
FIGURE 22: Mean Values of COD
-------
100
I I T
1 T
Pilot Plant
90
0)
-------
ON
O
0)
PP
oo
•
-------
by the final 13-day Period 15 with 100 mg/1 carbon. The objective was
to deviate from the performance level of the activated sludge plant by
exploring the limits of treatability and control of treatment.
Table 7 shows all operations were conducted at moderate levels of poly-
mer and alum dosages. The only data discarded were for transitional
days and for restarting the plant from a drained condition in the case
of Period 15.
Removal of filtrable TOG and COD was 78.0 and 78.2 percent, respectively,
for Period 13 and 73.7 and 74.6 percent, respectively, for Period 14.
These values plus those for total TOG and COD removal were the highest
obtained for the field program (excepting qualified Phase 6). The feed
stream TOG and COD levels were increased during this portion of the year
but not high enough to offset the increased carbon dosage effect. The
process did not respond in accord with isotherm formulae but was limited
by the fraction of TOG- and COD-exhibiting materials not readily adsorbed.
Activated carbon loadings for Periods 13 and 14 averaged 9.7 and 8.0
percent as filtrable TOG with applied loads of 13.0 and 10.3 .percent,
respectively. The applied loading, i.e., the feed stream filtrable
TOG expressed as a weight percentage of the carbon dose, having de-
creased as a result of the higher carbon feeds. For the same periods,
loadings of 36.8 and 30.1 percent as filtrable COD resulted from applied
loads of 47.3 and 38.5 percent, respectively. All loading data refer
only to the clarifiers, i.e., gravity filter influence is excluded.
Clarifier blowdown (Figure 13) reflected the use of alum in both clari-
fiers, the higher solids load contributed by the feed stream and carbon
feeds and, during Period 13, the increase in Unit #1 concentrator volume.
This later effect was counteracted during Period 14 by still higher
feed solids and carbon feed.
Filter runs averaged 29 hours and 25 hours for the two periods with
wash water consumption reported as 3.2 and 3.8 percent of the filtered
volume. Examination of accumulated data implied pilot plant carbon
dosage exceeded that required to place performance of the adsorption
system in the linear portion of the isotherm curve. Carbon dosage for
Period 15 was selected to be equal to the applied filtrable COD, i.e.,
100 mg/1 'estimated; such that a carbon loading of 70 percent as fil-
trable COD would have to be achieved to approximate the degree of removal
reported for many of the previous periods. Loadings of this amplitude
had been indicated in two grab sample isotherms but only at applied
COD/carbon ratios .exceeding 2.5 and were never achieved with composite
sample isotherms.
Mean applied filtrable COD was 102.8 mg/1 for Period 15, which was
scheduled in February. Filtrable-TOC and -COD removals were 47.0 and
61
-------
TABLE 7
Phase 7, Pilot Plant Performance by Operating Period, Mean Values
CONTROLLED VARIABLES
Overflow Rate, gpm/ft^
Carbon Feed, mg/1 ANA
C-31 Feed, mg/1 Unit #1
Unit #2
Alum Feed, mg/1 Unit #1
Unit #2
No. of Days Sampled
No. of Operating Days
PILOT PLANT REMOVALS
TOG unfiltered, %
TOG filtrable, %
COD unfiltered, %
COD filtrable, %
BOD5 unfiltered, %
Sus. Solids, %
SEWAGE PLANT REMOVALS
TOG unfiltered, %
COD unfiltered, %
BOD5 unfiltered, %
Sus. Solids, %
13
1
200
6
6
10
25
19
21
93
78
92
78
94
96
90
89
87
95
Operating Period
14 15
.0
.3
.0
.2
.2
.8
.3
.1
.3
1.0
250
6
6-8
10
25
12
13
91.5
73.7
91.6
74.6
93
98.2
89.6
89.0
80
96.1
1.0
100
6-8
6
15
15-20
10
13
79.5
47.0
82.4
47.8
86
95.3
81.4
81.0
90
89.8
Hyphenated chemical dosage data indicate changes instituted during period
62
-------
47.8 percent, respectively. These data increased to 50.3 and 49.8
percent for the contactor-clarifier portion of the system only. Carbon
loadings, also excluding the filter, were 15.9 percent as TOG and 51.1
percent as COD.
Total COD and TOG removal Figures 23 and 25 show sewage plant perform-
ance was reduced by nearly the same amount as the pilot plant for this
time. Mean data for the sewage plant, however, fail to show that a
process upset was discharging activated sludge in the effluent early
in the period. This was followed by a recovery trend during the final
7 sampling days. Figure 26 shows this trend and, that prior to shut-
down, the pilot plant effluent contained 14 mg/1 more COD than that of
the sewage plant. The magnitude of the loss in COD removal resulting
from the reduced carbon feed was 20.7 mg/1 COD. Color removal data
(Figure 27) show a marked decrease in color reduction for the pilot
plant as compared with the sewage plant. Color was the only solids-
free adsorbate indicator monitored for both plants.
Suspended solids in the feed were high for the period and the pre-
selected coagulation system was marginal. Solids carryover from both
clarifiers exceeded the arbitrary limits mentioned earlier. Both
polymer and alum feed were adjusted upwards during the run. Blowdown
requirements increased to 7.5 percent for Unit #1 and to 5 percent
for Unit #2.
The gravity sand-coal filter media had been removed, inspected and
returned to the filter prior to operation. No evidence of channeling,
mud balls or unexpected biological activity was discovered. With
suspended' solids carryover of 25 to 45 mg/1 going onto the filter
the 22-hour filter runs are reasonable. Wash water consumption was
3.8 percent of the volume filtered.
Supplemental Investigations
Additional data collected during the pilot plant program included the
adsorption isotherm data, limited temperature measurements, pH monitor-
ing, blowdown slurry solids content, color monitoring, limited nutrient
measurements, turbidity measurement with comparisons and gravity filter
operation.
Adsorption Isotherms - Adsorption isotherms produced from 24-hour com-
posite samples during the field program were to have been the principal
tool in evaluating process efficiency. Because of the successful use
of isotherms in earlier programs involving tertiary treatment, it was
unexpected when necessary laboratory procedures introduced factors
negating the relationships between isotherm and plant data.
63
-------
100
90
-a
0)
0)
80
0)
I 70
o
o
H
60
50
r~r
Pilot Plant
Sewage Plant
.Activated Carbon Feed, mg/1
145 .—
Operating Phase
1 »- 2
-*- 6
-*-200 250 100
I I 1
r^r-^ooc^csooo
Operating Period
FIGURE 25: Mean Values of TOG Removal
-------
500
400
300
200
§
0 150
100
50
0
M
cfl
O
CD
•H
Q
Feed
p
.fl
CO
Sewage Plant Effluent
Pilot Plant Effluent
13 14 15 16 17 18 19 20 21 22 23 24 25 26
24-Hour Composite Sample Dates, Feb, 1971
FIGURE 26: Daily Variation in COD, Period 15
65
-------
100
90
80
70
60
50
o
I 40
S-l
o
3 30
01
OJ
20
10
TI\\IIIIiT
Activated Carbon Feed, mg/1
145
Operating Phase
1 »- 2
1 [
T~1 i i i T
Pilot Plant
Average of All Determinations, 73.0%
Sewage Plant
V
.-*
~^— X-*
^
x—
\
\ -
/
V
/ Average of All Determinations, 32.0%
-X
Mean Color Values,
Feed Stream,
Unit #1 Effluent,
Unit #2 Effluent,
Filter Effluent,
Sewage Plant Effluent,
J I I I I I I
166 Samples
27 Units
15 Units
6 Units
7 Units
18 Units
I I I
<
LD
ffl
o
10
-------
The difficulties were related to filtrable suspended solids and volatile
components in the samples. To produce an adsorption isotherm, a unit
influent composite sample was vacuum filtered through a 0.45-micron fil-
ter. A portion of the filtrate was analyzed directly producing the 24-
hour sampling period results. The balance was used in the set of six
jars for the isotherm. The isotherm blank was always treated with the
coagulant dosage in use in the pilot plant unit during the sampling
period and it underwent a final vacuum filtration to duplicate the
handling of the carbon bearing jars. The initial filtrate usually had
a higher organic carbon content and COD than the resulting isotherm-
blank filtrate. Some organic filtrable suspended solids of less than
0.45 micron apparently were flocculated and retained on the second fil-
ter. The additional vacuum application possibly removed volatile mate-
rial that otherwise might exhibit COD or contain organic carbon. The
resulting adsorption isotherm, while perfectly acceptable when taken by
itself, represented a system operating on a lower CQ datum than the
pilot plant with insufficient information available to adjust for the
differential.
The differences generally were significant. For example, a feed stream
composite was found to contain 114.0 mg/1 filtrable COD with the result-
ing isotherm blank testing 81.5 mg/1. Another tested 109.2 mg/1 and
63.6 mg/1, respectively. Unit #2-influent isotherms were in better
agreement with 60.0 mg/1 and 59.1 mg/1 reported for the first example
above and 22.5 mg/1 vs. 17.5 mg/1 for the second. Better agreement
would be expected because Unit #2 influent had been subjected to floc-
culation in Unit #1, 20 mg/1 polymer in the first example and 6 mg/1
polymer and 10 mg/1 alum in the second.
The first Unit #2-influent isotherm example given above is one of the
few considered suitable for evaluating the pilot plant operation. It
is reproduced in Figure 28 along with actual pilot plant operating points
for the same date. The points lie on curves labeled performance paths
which represent the variation in carbon loading across a range of Ct
values while the carbon dosage of 145 mg/1 and the CQ values are constant
The isotherm represents single-stage adsorption performance with carbon
dosage varied so that the intercept of the Unit #2 performance path and
the isotherm, at a loading of 12.1 percent and Ct = 42.3 mg/1 COD,
establishes single-stage adsorption efficiency for 145 mg/1 carbon.
Unit #2, however, demonstrated process superiority by producing a load-
ing of 19.3 percent at Ct = 32.0 mg/1 COD. COD removal in Unit #2 was
47 percent, i.e., reduced from 60.0 mg/1 to 32.0 mg/1. The single-
stage carbon loading indicated by the isotherm at Ct = 32.0 mg/1 is
8.0 percent establishing the single-stage carbon requirement, to reach
a Ct of 32.0 mg/1, to be 350 mg/1.
67
-------
100
80
60
40
20
oo
c
tfl
o
a
o
tO
O
id
a)
4-1
to
•H
10
Performance
Paths for
Total Pilot Plant
Unit #1
Hypothetical Example:
Loadings for Total Pilot
Plant and Unit #1 at
Reduced CQ = 81.5 mg/1
1
10
Phase 1, Period il2
24 Hr Sample, 4/9-10/1970
Feed (Co) = 114.0 mg/1 COD
Feed Blank = 81.5 mg/1 COD
Unit #1 Effluent (C^ =60.0 mg/1 COD
Unit #1 Effluent Blank =59.1 mg/1 COD
M = 145 mg/1 ANA
I I Mil I I
20
30 40
60
80 100
200
300 400
Filtrable COD Remaining (Ct), mg/1
FIGURE 28: Freundlich Adsorption Isotherm Pilot Plant Analysis
68
-------
Published information for predicting two-stage countercurrent systems
did not cover the entire range of adsorbate-residual values of interest
so available data were extrapolated graphically. Assuming the principal
slope of the isotherm (1/n =2.3) to apply, two-stage countercurrent
adsorption requires 42 percent of the 350 mg/1 single-stage carbon dose
for the removal achieved. This is 147 mg/1 carbon or essentially that
actually used in Unit #2. As suggested in the original work^, the single
contactor-clarifier was producing two-stage countercurrent adsorption
performance.
In the pilot plant the once-used carbon was then advanced to Unit #1
where an additional loading of 37.2 percent was achieved. In so doing,
the feed COD of 114.0 mg/1 was reduced to the intermediate concentration
(C-j^) of 60.0 mg/1. On the downstream end of the pilot plant the gravity
filter removed an additional 1.6 mg/1 COD delivering a final effluent of
30.4 mg/1 filtrable COD. The total plant removed 83.6 mg/1 filtrable
COD resulting in a carbon loading of 57.6 percent and 73.3 percent fil-
trable COD removal.
If the reduction in filtrable COD between the feed stream determination
of 114.0 mg/1 and the Unit #1 influent isotherm blank determination of
81.5 mg/1 is considered physical removal of suspended matter, a hypo-
thetical correction in loadings can be made. A hypothetical performance
path for CQ = 81.5 mg/1 COD and 145 mg/1 carbon is included on Figure 28.
The total pilot plant removal becomes 51.1 mg/1 COD with a total loading
of 35.2 percent. Unit #1 loading is reduced to 14.8 percent; however,
this is still load added to the once-used carbon which brought an un-
changed 19.3 percent load with it from Unit #2. By comparison, the
Unit #1 influent isotherm (not shown) prepared with virgin carbon and
with C0 = 81.5 mg/1 COD gave the loading for 145 mg/1 carbon as 17.5
percent at C^- = 56.0 mg/1. The curve broke sharply in this region
never exceeding 18.4 percent loading.
In the hypothetical analysis, adsorbate removal was 62.7 percent. Single-
stage adsorption at the pilot plant Ct = 30.4 would result in a loading
of 7.0 percent and require 730 mg/1 carbon. The computed two-stage
countercurrent carbon requirement to deliver the same quality effluent
would be 34 percent of 730 or 248 mg/1 carbon. The pilot plant dosage
of 145 mg/1 was only 20 percent of single-stage or 58 percent of two-
stage countercurrent adsorption system requirements.
This high-efficiency carbon utilization existed in the pilot plant
process despite making the aforementioned allowance for the possible
physical removal of filtrable suspended solids. Activated carbon,
however, has been reported to have affinity for certain colloidal
solids that produce haze or color and differences in performances of
some carbons seemingly are related to their ability to remove a portion
of the organic matter in suspension.
6'9
-------
The pilot plant data supported earlier evidence (Figure 3) of an
adsorption-resistant fraction of TOG and COD. A mean isotherm curve
was prepared from the data of 35 individual composite-sample TOG iso-
therms (Figure 29). Adsorption proceeded reasonably similar to Freund-
lich theory up to carbon dosages in the range of 800 mg/1. At higher
dosages TOG removals were less than theoretical. In Figure 29, adsorp-
tion at 800 mg/1 carbon had removed 65 percent of the feed TOG. At
1600 mg/1 removal went to 72.2 percent at a Ct of 6.5 mg/1 whereas ex-
tending the curve in a straight line through the first three points
would have resulted in about 79 percent removed at a Ct of 4.9 mg/1.
Adsorbate removal would proceed in this range but at much reduced
efficiency and Figure 30, a plot of applied TOG vs. TOG loading, shows
this clearly. With the exception of three operating periods, all field
results were with 145 mg/1 carbon feed. The feed stream filtrable TOG
ranged widely causing loading to range similarly however the removals
persisted near a mean weighted value of 68.0 percent. The primary ex-
ceptions being the periods of higher or lower carbon dosage and the
qualified alum study data. Higher dosages forced removal increases of
only a few percentage points, thus indicating adsorptive capacity of
the carbon was not the process limiting variable. Removals were limited
to that fraction of the adsorbate indicators that could be adsorbed.
Filtrable COD removal by the pilot plant for all applicable periods
averaged 69.9 percent (Figure 31). Considering pilot plant performance
could not exceed the limit but rather approach it, it develops that
carbon feeds other than about 130 percent of the feed stream filtrable
COD are less than optimum. Three Operating Periods, 2, 4 and 15 pro-
duced mean COD/carbon loads of 53.6, 54.0 and 51.1 percent, respectively,
placing the ultimate adsorptive capacity of the carbon in this range.
Operating Period 15 was especially designed to avoid encountering the
adsorbate limitation, whereas 2 and 4 were optimized coincidentally by
the strength of the feed stream at those times. The adsorptive capacity
must match the 70 percent of COD available for adsorption for optimum
utilization, i.e., 100 x 0.70 COD/0.53 LOAD = 132% COD = mg/1 carbon
feed for the sewage studied. The isotherm in Figure 28 is from Period 2
with the carbon feed 127 percent of the feed filtrable COD for that date,
making the isotherm evaluation consistent with the subsequent discussion.
Temperature - No regular program of temperature measurement existed.
Random readings of the feed stream temperature ranged from a high of
91°F in July to a low of 75°F in February. A 1.5°F rise in temperature
was recorded in the stream as it passed through the pilot plant on hot
clear days. Although the adsorption process is temperature sensitive,
the observed temperature range does not significantly affect the results
and no adjustment in adsorption data was indicated.
70
-------
10.0
8.0
6.0
4.0
I I I
- 2.0
C
•H
o 1.0
C
O
cd
U
2 °'5
•H
Unit #2 Influent
Mean Co - 17.2 mg/1 TOG
800
1600
I
Carbon Dose, mg/1
'40
Feed
Mean C =23.6 mg/1 TOG
0.1
I
4 6 8 10 20
Filtrable TOG Remaining (Ct), mg/1
30 40
FIGURE 29: Mean Value Isotherms, TOG
71
-------
40
cfl
O
O
H
4-1
rH
•H
QJ
•H
35
30
25
20
15
10
0
Feed Isotherm, 4/6-7/1970
Feed Isotherm, 10/12-13/1970
mg/1
Feed Isotherm
6/22-23/1970
Unit #2 Influent
Isotherm
10/12-13/1970 °
S-51
-250 mg/1
200 mg/1
X Periods 1, 2 & 3
O Periods 4 through 11C
• Periods 12A through 12C
OPeriods 13, 14 & 15
Line of Mean Plant Performance
for 145 mg/1 Carbon,
Mean Value = 68.0% Removed
0 5 10 15 20 25
Filtrable TOG Activated Carbon Loading (X/M), Wt %
FIGURE 30: Clarifier Performance, Applied TOG Load vs.
Carbon TOG Loading
72
-------
120
110
100
90
80
70
60
o
cfl
O
O
O
u
0)
1-1
a 50
•a 40
0)
•H
H
S 30
<
20
10
COD Isotherms
from Composite
Samples
100
60% Removal
80% Removal
200 mg/1
250 mg/1
69.9% Mean Removal
for 145 mg/1
X Periods 1, 2, 3
O Periods 4 through 11C
• Periods 12A, 12B, 12C
Periods 13, 14, 15
I
10 20 30 40 50 60 70 80 90
Filtrable COD Activated Carbon Loading (X/M), Wt %
100
FIGURE 31: Clarifier Performance, Applied TOC Load vs.
Carbon TOC Loading
73
-------
pH - Composite samples were tested for pH during much of the program.
The testing was halted after variations were found to be small. The
feed stream pH ranged from 7.6 to 8.1, averaging 7.95. Unit #1 effluent
ranged from 7.6 to 8.2, averaging 8.05. Unit #2 effluent ranged from
7.7 to 8.25, averaging 8.1. Gravity filter effluent ranged from 7.8 to
8.2, averaging 8.1. The sewage treatment plant effluent ranged from 7.4
to 8.1, averaging 7.9. The high alum dosages may have been responsible
for lowering pilot plant effluent readings about 0.1 units during
Phase 6.
Slowdown Suspended Solids - The suspended solids content of the slurry
blowdown was determined on certain grab samples. The various configura-
tions of the slurry concentrators and suspended solids influx rates gen-
erally resulted in unsuitable internal slurry concentrating conditions.
During Operating Period 2 at a low overflow rate, however, conditions
were suitable. Blowdown samples settled less than 5 Vol % upon 5-minute
settling and were determined to contain 16.5 g/1 and 24.6 g/1 suspended
solids for Unit #1 and Unit #2, respectively. Other data for Unit #1,
adjusted by results of 5-minute settling tests made after collection,
show 20 to 26 g/1 at the higher carbon feed rates and 16 g/1 at 100 mg/1
carbon feed rate. Unit #2 averaged 22 g/1 at the high feed rates dropping
to 16 g/1 for the 100 mg/1 carbon test, possibly as the result of the
high solids carryover noted earlier. Incomplete data were obtained from
the Darco S-51 study but unadjusted results were 14 g/1 for Unit #1 and
45 g/1 for Unit #2. The suspended solids content of slurry to be regen-
erated influences regeneration costs.12
Color - The pilot plant process was substantially superior to the sewage
plant process in terms of color removal. Figure 27 gives the mean color
removal observed for both systems for comparable periods.
Nutrients - Nutrient removal was not an objective of this study but three
sets of composite samples were analyzed for ammonia-nitrogen and phos-
phorus to assess any effect that might exist. The results on laboratory-
filtered samples were as follows:
Unit #1 Unit #2 Filter Sewage
Sample Date Feed Effluent Effluent Effluent plant
11/8-9/70
Ammonia-Nitrogen as N 26 24 23 22 16
Phosphorus asP 10.0 9.1 7.8 5.8 9.9
11/9-10/70
Ammonia-Nitrogen as N 25 23 23 21 11
Phosphorus as P 11.7 10.4 8.1 5.7 10.2
11/10-11/70
Ammonia-Nitrogen as N 25 24 23 23 9.5
Phosphorus asP 10.9 8.8 4.6 4.6 10.6
74
-------
An ammonia-nitrogen removal trend exists but not to a significant degree.
Phosphorus removal, however, is significant. It is not known if this is
associated with carbon adsorption, alum flocculation or phosphorus take-
up by microorganisms.
Turbidity Measurement - Through most of the program, turbidity measure-
ments were made on grab samples from the pilot plant at 2-hour intervals
with a Each Model 2100 Laboratory Turbidimeter because this unit has
been used during other EPA powdered carbon treatment studies. Measure-
ments were also made on the composite samples daily prior to the routine
laboratory determinations with a Coleman #14 Spectrophotometer.
The Hach instrument measures scattered light and is standardized against
a solidified Formazin suspension of known turbidity; hence its readings
may be quite different from those obtained by light transmittance or
absorption as in the spectrophotometer. The carbon content of the sample
among other things affects the difference in readings between the two
systems. Figure 32 compares the two readings for the composite samples.
The feed stream ratio is the most consistent being devoid of activated
carbon and other introduced variables. The clarifier effluents vary
substantially as operation was varied. Unit #1 effluent was least
affected of the two as a substantial portion of its turbidity was raw
sewage suspended solids. Unit #2 effluent was highly variable as the
sewage solids-to-carbon ratio was less. The filter ratio variation is
not understood with the possible exception of Period 12 when carbon was
observed penetrating the filter media. If the filter was discharging
biological material produced therein, it would be unlike the sewage
suspended solids or carbon solids monitored in the other samples.
Spectrophotometer determinations more closely approximated visual esti-
mates of turbidity when activated carbon, either as an uncoagulated
suspension or pinpoint floe, was present. This is substantiated by
Figure 33. The ratios of spectrophotometer JTU to mg/1 suspended solids
vs. operating period curves center about the ratio of 1.0. The filter
effluent curve again indicates a peculiarity exists in this sample
stream.
The 2-hour field turbidity determinations provide the only directly
measured information on diurnal feed stream variation and its effect
upon the pilot plant. A 4-day period of these measurements is given
in Figure 34. The cycling noted in the feed stream JTU is readily
apparent in the clarifier effluents. Also, the inferred solids load
on the pilot plant is relatively constant at a high level for 18 to
20 hours per day but drops off for 4 to 6 hours every morning. The
curves also show the result of an operating accident. Loss of carbon
feed for 8 hours had little effect on Unit #1 effluent but produced a
measurable Increase in Unit #2 effluent turbidity and loss of filtra-
tion efficiency for a number of hours.
75
-------
6.5
6.0
e 5-5
§ 5.0
o
cd
Pd
4.5
4.0
3.5
cu
I 3.0
§• 2-5
£ 2.0
ft
en
o' 1-5
•rl
p? 1.0
0.5
0
Filter Effluent
Unit //I Effluent
Operating Phase
1 »- 2
I I I
II I i I
- 6
I I
I I
rH CM ro -o-
5 <. O Q W <|
CO 0\ O O O O rH
rH rH iH rH i-H
Operating Period
rH CM
pq o
CM CM
rH rH
FIGURE 32: Ratio of Spectrophotoineter JTU and Hach JTU
-------
4.0
CO
TJ
•H
c
0)
a.
W
3
to
2.0
0)
o
4J
O
f
a.
o
5 1.0
0)
ex
CO
o
•H
4-J
CO
Pi
T I
I I I I I I \ \ \ }
I I I
Unit #2 Effluent
Unit #1 Effluent
J_ I L__l I I J L
J I I I I 1 I I I L_J L
i I
PP O
-------
00
•H
13
•H
,0
t-4
O
O
0)
T)
,
O
&
120
100
80
60
40
20
T
- Phase 5, Period 11C
,Feed
1
* Filter Backwashed
Carbon Feed Off — *V* H
Unit #2 Slurry Lost
•*»
Unit #2 Effluent \ /Filter Effluent
Wed, 9/16
Thurs, 9/17
Fri, 9/18
Sat, 9/19
FIGURE 34: Diurnal Variation of JTU in System
-------
Filter Operation - The duration of filter runs and the quality of filter
effluent have been reported earlier in this report and are independent
of the pilot plant operators. Wash water consumption, as a percentage
of water filtered, has also been reported but this parameter was found
to be highly dependent upon the pilot plant operators rather than design
or process.
Backwashing procedure required establishing a low upflow rate for the
duration of the surface raking followed by 5 minutes or more, as needed,
of the high upflow wash rate. Without a flow-rate'indicator the initial
flow varied widely, as did the duration of the raking. The principal
wash rate also varied with water temperature and the duration varied
with the attention and judgment of the operator. Throughout the field
program 12 individuals worked as operators at one interval or another,
with the records showing water consumption ranging from a low of 325 gal.
to a high of 1205 gal. per wash, both being extreme and unsatisfactory.
Average consumption per wash by individual ranged from 481 to 809 gal.
with the average for all backwashes being 647 gal. The median volume
for all backwashes was 620 gal, however one experienced operator on a
supervised shift averaged only 515 gal. with good results. Surface
washing used 40 percent of this volume at a 10 gpm/ft rate followed
by the balance being consumed at 15 gpm/ft . Proper surface wash equip-
ment would reduce the volume used prior to backwash to an insignificant
figure; however, without the high rate manual surface wash period, the
main backwash period would have to be extended with little change in the
total water used. Based upon 515 gal. per wash, the wash water consump-
tion would be in the range of 3.0 percent with filter surface loadings
of 2.6 to 2.8 gpm/ft^. This percentage is higher than would exist if a
conventional filter, permitting head losses greater than the 4.5-ft limit
of pilot plant equipment, were used.
79
-------
DISCUSSION
General Considerations
Anthracite coal was selected for the top layer of filter media in the
gravity filter to eliminate mud balls having a greater settling rate
than the filter sand. In previous studies the polymer-flocculated
carbon conglomerated surface filter sand into mud balls which settled
to the top of the supporting gravel during backwashing. This deposited
quantities of loaded carbon below the filtering zone, leading to pro-
gressive deterioration of effluent quality. Efforts to eliminate or
control the problem were generally ineffectual. In the present study
mud balls did not develop because of the surface raking which preceded
each wash and any conglomerate would incorporate anthracite coal which
would come to rest at the sand-coal interface effectively within reach
of the surface raking. There was no danger of mechanically disturbing
the supporting gravel which hindered the work in the previous study.
The final effluent was not completely satisfactory from an aesthetic
viewpoint. It had a perceptible and distinctive sour odor (not hydrogen
sulfide) quite disagreeable as compared to that of an activated sludge
plant effluent. This odor could be detected in the clarifier effluents
suggesting it was produced from some unstabilized and adsorption-
resistant fractions present in the raw sewage, If it was present in
the raw sewage it was masked by the stronger odors present prior to
activated carbon contacting.
Since the materials contributing much of the TOG to the final effluent
appeared to be adsorption resistant, some procedure other than direct
carbon adsorption would be needed if further reduction in TOC was de-
sired. The adsorption-resistant fraction probably varies between raw
sewages and the concentration should be determined for prospective
applications of the process. The raw sewage in one study-"--* apparently
contained less than 6 percent adsorption-resistant filtrable TOC per-
mitting total TOC removals of 97 percent by adsorption. Tucson raw
sewage with about 30 percent adsorption-resistant filtrable TOC was
limited to about 90 percent total TOC removal. Both studies used the
same carbon in evaluating the raw sewages. Earlier^, the Tucson acti-
vated sludge plant effluent was successfully treated further by direct
carbon adsorption indicating adsorption-resistant materials are either
altered or assimilated by biological treatment.
Biological growth was evident on surface apparatus of both clarifiers
and on the gravity filter sidewalls and underdrain system. The units
were devoid of dissolved oxygen but slimes and facultative bacteria
of various colors developed in the launders and. on other interface
apparatus. Green algae grew in the filter. Brushing loosened accumu-
lations every few days. There was no observed manifestation of sub-
81
-------
surface biological activity in the clarifiers or in the slurry concen-
trators. There was no noticeable gas formation. Slowdown slurry odor
was the same as noted previously for the effluent. During rare periods
of inlet-screen bypassing or failure, raw solids settling with carbon
floe on internal bracing in the clarification zone of Unit #1 would gas
and pop to the surface. These clumps and other floatable material were
manually skimmed away. Grease balls did not form in the system but
some small ones entered with the influent. The scum baffle was very
effective in eliminating floating material downstream. Flotation of
carbon floe occurred in Unit #1 at times of carbon advance system mal-
function. Entrainment of air at the carbon advance system surge tank
due to accidental low liquid level approximated a dissolved air flota-
tion system. The larger floatable solids were normally eliminated
upstream of the pilot plant at the screened inlet system. A larger
plant would, of necessity, incorporate a mechanical skimming apparatus.
Mean pilot plant performance in terms of TOC and COD removals are sum-
marized in Table 8 along with comparable performance data for the acti-
vated sludge plant. Composite sample results for each of the powdered
activated carbon dosages have been averaged without regard to coagula-
tion system, slurry advance system or overflow rate. The data are
probably biased in favor of the sewage plant. It was operated in a
routine manner aside from some brief physical problem periods. In con-
trast, the pilot plant was purposely subjected to widely ranging experi-
mental situations. Nevertheless, the pilot plant system out-performed
the sewage plant most of the time.
Throughout the field program, carbon was fed into the returning slurry
streams of each contactor-clarifier. The liquid phase of the slurry
stream has the lowest adsorbate concentration in the unit. It is this
liquid phase that becomes the unit effluent. The slurry advance system
make-up liquid was taken from the clarification zone of Unit #1 to pre-
serve this relationship. Countercurrent adsorption principles, i.e.,
always introduce the least loaded carbon into the most purified stream,
were preserved in this study. Conversely, because the system was con-
tinuous, carbon seldom encountered feed stream adsorbate concentration
because the feed stream was quickly diluted many-fold with recirculating
slurry in the inner draft tube of the mechanism. This did not prevent
loading the carbon to a high degree with adsorbate.
The total liquid volume of the pilot plant treating units was 1615 gal.,
excluding interconnecting piping. The retention time at 1.5 gpm/ft^
overflow rate and 3 gpm/ft- filter rate would be 88 minutes from inlet
chamber to filter effluent. An overflow rate of 2.25 gpm/ft2 is con-
sidered a nominal maximum depending upon sewage/carbon characteristics
at which the retention time would decrease to 63 minutes. At these
overflow rates, the carbon contacting times would be 11.6 and 7.8
minutes, respectively, per contact unit. These times and overflow rates
are more conservative than tested elsewhere.13
82
-------
Carbon
Dosage
mg/1
TABLE 8
Mean Treatment Plant Performance
TOG
Pilot Plant Sewage Plant
Red' a
Inf Eff Red'n
mg/1 mg/1 %
100 (ANA) 93.5 19.4 79.3 81.4
145 (ANA) 91.8 10.9 88.1 86.6
145 (S-51) 89.8 9.2 89.8 88.2
200 (ANA) 102.4 6.7 93.3 90.3
250 (ANA) 97.9 8.3 91.5 89.6
COD
Pilot Plant
Inf Eff Red'n
mg/1 mg/1 %
392.0 69.0 82.4
364.0 41.0 88.7
344.5 42.8 87.6
357.0 27.8 92.2
368.0 31.3 91.6
Sewage Plant
Red'n
81.0
88.2
88.9
89.1
89.0
Samples not filtered in laboratory prior to analysis.
83
-------
The solids-contact clarifier offers certain simplifications and economy
for powdered activated carbon adsorption system design by combining in
one vessel the functions of mixing, flocculation, and solids separation.
Power use per stage is minimized and multiple tankage within each stage
is eliminated, thus reducing space requirements and piping complexity.
The advantages of automated proportional chemical feed controls are not
clear cut. For example, during the field studies there were occasions
when carbon feed was nonexistent for up to 8 consecutive hours. On one
occasion the carbon feeder was down 5.5 hours in the afternoon and the
drive on Unit #2 was left turned off for 2 additional hours, all during
the peak sewage strength period of the day. The composite samples for
the 24-hour period showed the mean filtrable COD removal was 72.7 mg/1.
The previous 24-hour period which was uneventful removed 67.6 mg/1. The
feed streams tested 99.6 mg/1 and 101.0 mg/1 filtrable COD, respectively,
while operating the plant at the maximum capacity of 1.5 gpm/ft overflow
rate. The carbon feed was off for nearly 9 times the single clarifier
retention time. Originally, data for days with underfeed were auto-
matically discarded but later, because the effect could not be distin-
guished from the normal daily deviation, these data were retained.
The estimated 7 to 9 Ib of carbon inventory in the recirculating slurry
had adequate adsorption capacity to span long periods without feed.
Difficulties were associated with maintaining slurry inventory, effluent
clarity and floe settleability but not adsorption. If carbon feed need
not be varied as the sewage strength changes then coagulant feeds may
not be proportioned to sewage strength either as carbon flocculation
must be maintained. Proportional feed based upon flow rate for a variable
rate plant is, however, highly practical and would result in chemical
economy as compared to feeding the minimum marginal dosages for peak
strength and flow rates continuously.
The pilot plant was a constant rate system so no adjustment in the chemi-
cal consumption data is justified for flow rate proportioning.
Periods of weak feed sewage were 4 to 5 hours per day or 17 to 21 percent
of time; however, chemical feeds were marginal for peak feed strength
periods so proportioning by feed stream turbidity measurement would
simply shift consumption from weak to strong periods. Again, no adjust-
ment in overall chemical economy is thought justified.
Regeneration and reuse of powdered activated carbon combined with raw
sewage solids and alum content is presumed possible. ' Development of
various processes is proceeding. At least one investigation-"--^ shows
promise, although the ratio of carbon to raw sewage solids was much
higher than in this study. Alum recovery and reuse were also con-
sidered in that investigation.
84
-------
Process Economics
Based upon the design approach noted below, cost estimates have been
developed for 10-mgd and 100-mgd treatment systems of the type studied.
Plant Design - Each plant incorporates complete water recovery, elimi-
nating any requirement for separate disposal. All process water needs
are met with settled filter wash water which is pumped from storage to
fill slurry makeup tanks and to dilute polymer in day tanks. Excess
backwash water is pumped to the downstream adsorption units for
recovery.
Chemical feed systems are sized to provide up to 300 mg/1, 30 mg/1 and
100 mg/1 of carbon, polyelectrolyte and alum, respectively. Carbon
recovery via regeneration is visualized and 15 percent loss per cycle
is assumed.
The site requirement for the 10-mgd plant is an area 110 ft x 150 ft.
Two ACCELATOR mechanisms in basins 62 ft square x 17.5 ft deep handle
the adsorption duties of this plant and three 4-cell gravity sand-
anthracite filters with GREENLEAF ® Filter Controls (a unique siphon
valve system for controlling the functions of multiple rapid-sand
gravity filters) provide effluent polishing at 3 gpm/ft . These con-
crete structures, as well as a two-story metal building housing the
carbon regeneration system, polyelectrolyte feed tanks and pumps, alum
feed tanks and pumps, carbon makeup and polyelectrolyte storage, office
and laboratory are constructed above grade.
Agitated covered concrete tanks located below grade between the ACCELATOR
basins and filters include three 4000-ft^ carbon-slurry day tanks. During
operation, one tank is in use, a second is being readied with water plus
carbon from regeneration and/or storage, and the third is ready and on
standby for use the following day. Two 6000-ft-^ mixed tanks provide up
to 5.4 hr of holding time for blowdown from the respective clarifiers,
one supplying partially spent carbon to the first stage unit and the
other furnishing fully spent carbon to regeneration.
The type of filter visualized requires no wash water storage. A
21,000-ft unmixed basin for backwash wastewater catchment completes
the below-grade tankage for this plant.
At a feed rate of 140 mg/1 (1.17 lb/1000 gal.), the daily carbon require-
ment is 11,700 pounds. The plant requires an in-service carbon inventory
of 31,300 Ib and storage for one carload of bagged carbon. A 20-ton car-
load will meet maximum makeup requirements for 23 days. Drum delivery of
polyelectrolyte is assumed. Consumption at 14 mg/1 is about 16 drums
per week. At a feed rate of 30 mg/1, liquid alum (50% alum dry wt.)
would be consumed at 5,000 Ib per day. A tank car load would last
16 days.
85
-------
A 100-mgd plant of similar general design occupies an area 450 ft x
450 ft. To provide appropriate flexibility and emergency operating
capability, five 20-mgd ACCELATOR units operate in parallel as the
upstream contact stage, with five similar parallel units as the down-
stream stage. These mechanisms are installed in common-wall concrete
basins, each 88 ft square x 21 ft deep. Fourteen 43-ft square, 4-cell
filters with GREENLEAF Filter Controls are required.
A two-story building covers two below-grade agitated blowdown and
three carbon slurry day tanks and houses the carbon regeneration system,
two polyelectrolyte day tanks, carbon feed systems, alum feed systems,
service pumps, offices and laboratory. Unit capacities of the blowdown
and carbon slurry tanks are 48,000 ft3 and 40,000 ft3, respectively.
In this plant, a conventional 67-ft diameter clarifier is used for
filter wash water recovery. The overflow furnishes all plant process
water and underflow solids are transferred by gravity to the first
stage blowdown tank enroute to regeneration. As before, excess filter
backwash wastewater goes to the downstream clarifiers for recovery.
The in-service carbon inventory for this plant is about 300,000 pounds.
At 140 mg/1, daily carbon use is 117,000 Ib and a make-up requirement
of 17,550 Ib/day is indicated. Bulk storage is not included. Rail
delivery of bulk carbon is visualized with cars unloading make-up
quantities into day tanks as required.
Polyelectrolyte coagulant is delivered by tank car to a 20,000-gal.
storage tank from where it is pumped as required to either of two day
tanks for dilution to feed strength. The dilute material is then pumped
to each contact-clarifier through variable-orifice rate control devices
installed along a distribution header serving all units.
Liquid alum would be pumped directly to the application points from
tank cars through a similar distribution system.
Both plant designs assume a mild climate location with weatherproof
equipment installed in the open. Carbon handling equipment is either
of corrosion resistant material or is protected with suitable coatings.
Alum handling equipment is similarly constructed. Essential instrumen-
tation for flow metering and control and low-lift process and service
pumps are included. A conventional intake, comminutor and grit removal
structure precedes the low lift pumps.
Treatment Cost - The estimated construction cost of the 10-mgd plant,
including land but exclusive of. the capital cost of a carbon regenera-
tion system, is as follows:
86
-------
Equipment delivered ($295,000) and installed $366,500
Plant structures 236,700
TOTAL PLANT COST $603,200
Engineering (10% of total plant cost) 60,300
Contractor's fee (10% of total plant cost) 60,300
Land acquisition (2% of total plant cost) 12,100
Contingencies and omissions (15% of total
plant cost) 90,500
TOTAL CAPITAL COST $826,400
.; Annual operating cost breaks down as follows:
Capital ($826,400 for 20 years @ 6%) $ 72,000
Maintenance (3% of equipment + 1% of
structures) 13,400
Labor 79,900
Power (1C/KWH) 21,400
TOTAL $186,700
This is equivalent to 5.10/1000 gal. treated, exclusive of chemical costs.
The latter are influenced by the feed-stream quality and treatment objec-
tives. For this study the feed stream is assumed similar to that en-
countered during the field period and optimum carbon dosage of 132% of
mean filtrable COD is used. Mean filtrable feed COD was 89.3 mg/1 for
all composite samples analyzed so the carbon feed is 118 mg/1. A poly-
mer feed of 14 mg/1 and alum feed of 25 mg/1 are the estimated quantities
to control the process.
Carbon regeneration will supply 85 percent of the carbon feed with the
residual 15 percent being new carbon makeup. Carbon regeneration costs
of Berg, et al., were modified to reflect 15 percent solids rather
than 25 percent solids feed to the thermal reactor. Centrifugation
costs of 0.2(?/lb are included in the regenerated-carbon unit cost.
Estimated chemical costs are:
C/1000 gal.
Regenerated carbon
(0.983 lb/1000 gal.) (1.9c/lb) (0.85) 1.6
Makeup carbon
(0.983 lb/1000 gal.) (9.5e/lb + 3c/lb freight)
(0.15) 1.8
C-31 polyelectrolyte
(0.117 lb/1000 gal.) (3l£/lb + 3
-------
On this basis, the total cost of plant operation on this scale is
13.9C/1000 gal.
Using the same approach, the estimated construction cost of the 100-mgd
installation is as follows:
Equipment delivered ($2,120,400) and installed $2,374,600
Plant structures 1,270,600
TOTAL PLANT COST $3,645,200
Engineering (10% of total plant cost) 364,500
Contractor's fee (10% of total plant cost) 364,500
Land acquisition (2% of total plant cost) 72,900
Contingencies and omissions (15% of total
plant cost) 546,800
TOTAL CAPITAL COST $4,993,900
The breakdown for annual operating cost exclusive of chemical cost is:
Capital ($4,993,900 for 20 years @ 6%) $ 435,400
Maintenance (3% of equipment + 1% of
structures) 83,900
Labor 239,800
Power (1C/KWH) 122,500
TOTAL $ 881,600
This is equivalent to 2.4
-------
A 10-mgd high-rate powdered activated carbon treatment plant^3 not util-
izing countercurrent contacting was estimated to cost 21.50/1000 gal.
Carbon regeneration was 1.50/lb; however, the single-stage process re-
quired 600 mg/1 carbon feed with high alum feeds and polymer to control
same. Chemical operating costs totaled 17.40/1000 gal. with alum re-
covery and reuse. With the greater carbon feed rate, i.e., 5 lb/1000 gal.,
cost of regeneration was 6.750/1000 gal. and makeup carbon cost 4.50/
1000 gal. for a total of 11.250/1000 gal. This approximates the cost of
using virgin carbon, freight included, on a throw-away basis in the two-
unit countercurrent process.
The projected total cost, some years ago, of a 10-mgd granular activated
carbon adsorption system using multiple downflow contact columns was
8.30/1000 gal.^ A plant of this type might require as much as 133,000
Ib of in-service carbon inventory for each 1000 gpm of system capacity.
The powdered activated carbon system described is conservatively esti-
mated to require only 4,500 Ib of in-service carbon inventory for each
1000 gpm of capacity. These inventories represent an initial carbon
investment of $240,000 (26o/lb) for a 10-mgd granular carbon plant as
compared to only $3,100 (9.50/lb) for a powdered carbon plant. Both
systems require facilities to warehouse and handle 10 - 20 percent
makeup carbon.
At 50 to 55 percent carbon loading, powdered carbon utilization would
be about 80 to 100 percent of the top utilization reported for granular
systems depending upon the extent of biological degradation assisting
the granular system. More frequent regeneration might be required for
the powdered carbon, but this factor is more than offset by the normal
price differential between powdered and granular carbons. There would
be problems applying raw sewage to carbon columns.
The present cost of activated sludge treatment is estimated to be 170/
1000 gal. for a 10-mgd plant and 8.20/1000 gal. for a 100-mgd plant,
1967 cost information^ was adjusted to 1971 by applying the EPA con-
struction cost index and updated amortization (20 years @ 6%) to plant
cost and 5 percent per year increase in operating and maintenance costs.
The pilot plant process is less costly than the activated sludge process
at the 10-mgd plant size. As unit chemical costs do not diminish
greatly with increasing plant size the activated sludge process becomes
less expensive for large plants.
89
-------
ACKNOWLEDGEMENTS
The investigators at INFILCO Division, Westinghouse Electric
Corporation, gratefully acknowledge the cooperation extended by
the City of Tucson, Arizona, Mr. F, E. Brooks, Director of Water
and Sewer Systems and Mr. E. 0. Dye, Superintendent, Sewerage
Division, in providing locations for essential equipment on municipal
property.
The design, construction and supervision of the pilot plant opera-
tion and report preparation were performed by Mr. R. L. Beebe of
the INFILCO Division with Dr. C. F. Garland, Division Manager of
Research and Test, consulting and administrating. Analytical work
was performed in the INFILCO laboratories by Mr. J. Santillan
under the direction of Mr. E. P. Young, Laboratory Supervisor.
The support and guidance contributed by Mr. E. L. Berg, Project
Officer, Environmental Protection Agency, is acknowledged with
sincere appreciation.
91
-------
REFERENCES
1. Anon., "Taste and Odor Control in Water Purification," West
Virginia Pulp and Paper, Chemical Division, 1966.
2. Sebastion, F. P., "Wastewater Reclamation and Reuse," Water and
Wastes Engineering, Vol. 7, No. 7, pp 46 - 47, (July 1970).
3. Morris, J. C. and Weber, W. J., Jr., "Removal of Biologically
Resistant Compounds by Adsorption," U. S. Department of Health,
Education and Welfare, Public Health Service Publication No.
999-WP-ll, AWTR-9.
4. "Summary Report, Advanced Waste Treatment Research Program, July
1964 - July 1967," U. S. Department of the Interior, Water
Pollution Control Research Series Publication No. WP-20-AWTR-19.
5. Beebe, R. L. and Stevens, J. I., "Activated Carbon System for
Wastewater Renovation," Water and Wastes Engineering, Vol. 4,
No. 1, pp 43 - 45, (January 1967).
6. Garland, C. F. and Beebe, R. L., "Advanced Wastewater Treatment
Using Powdered Activated Carbon in Recirculating Slurry Contactor-
Clarifiers," U. S. Department of the Interior, Water Pollution
Control Research Series Publication No. ORD-17020FKB07/70.
7. "Measuring Adsorptive Capacity of Activated Carbons for Liquid
Purification," Bulletin D-87, Atlas Chemical Industries, Inc.
1966.
8. "Standard Methods for the Examination of Water and Wastewater,"
12th Ed., American Public Health Assn., Inc., 1965.
9. Hiser, L. L., "A New Approach to Controlling Biological Processes,"
Environmental Science and Technology, Vol. 4, No. 8, pp 648
(August 1970).
10. Rinehart, T. M., Scheffler, G. H., Helbig, W. A., and Truemper,
J. T., "A Symposium on Activated Carbon," Atlas Chemical Industries,
Inc., 1968.
11. Private communication, Rinehart, T. N., 1970 (Letter 2-17-70).
12. Berg, E. L. , Villiers, R. V., Masse, A. N., and Winslow, L. A.,
"Thermal Regeneration of Spent Powdered Carbon Using Fluidized-
Bed and Transport Reactors," Chemical Engineering Progress,
Symposium Series, Vol. 67, No. 107, pp 154-164 (1970).
93
-------
13. Shuckrow, A. J., Dawson, G. W., and Olesen, D. E., "Treatment of
Raw and Combined Sewage," Water and Sewage Works, Vol. 118, No. 4,
pp 104 - 111 (April 1971).
14. Smith, R., "A Compilation of Cost Information for Conventional and
Advanced Wastewater Treatment Plants and Processes," U. S.
Department of the Interior, Advanced Waste Treatment Branch,
December 1967.
94
-------
APPENDIX
Low-Level COD Determination
Discussion - The procedure is fundamentally that given for dilute
samples in "Standard Methods for the Examination of Water and Waste-
water," 12th Edition, American Public Health Association, p. 513 (1965),
amended to further minimize the introduction of extraneous-material
error.
The published precaution concerning contaminated glassware and atmos-
phere is reemphasized. In addition, laboratory grade distilled water
or demineralized water used for reagent preparation, pre-reflux dilu-
tion, blanks, rinsing, or final dilution may cause erratic results.
Water used for these functions should be especially prepared by the
distillation procedure described fully in the "Reagents" section for
blank water.
Contamination of the sulfuric acid reagent, even though carefully
prepared, can introduce an error. The error is cancelled by running
all samples and blanks to a set with the same batch of Ag2SO^-H2SO^
solution.
No significant error has been attached to the use of a polyethylene
wash bottle for rinsing down condenser tubes and tips with the blank
water.
The apparatus described is effective in preventing dust contamination
and in protecting the analyst from injury by bumping which may occa-
sionally occur during refluxing. It differs slightly from that de-
scribed in "Standard Methods."
Apparatus -
500-ml short-necked round-bottom boiling flasks. ST joint 24/40,
Corning #4320 or equal.
Allihn type condensers, 500-ml jacket, drop tip inner ST joint
24/40, top outer ST joint 24/40, Corning #2480 or equal.
Connecting tubes 75°, both ends inner ST joints 24/40, Corning
#8920 or equal.
Tube stoppers, cap type, full length outer ST joint 24/40, 2
required per condenser.
Heaters, Precision Catalog #61560 or equal.
Running-water cooling bath large enough for the required number
of flasks.
95
-------
The apparatus is mounted such that the complete glass assembly consist-
ing of flask, condenser, and upper connecting tube may be raised from
or lowered to the heater as required. The connecting tube is placed
on the upper end of the condenser pointed away from work or chemical
storage areas. The cap-type stoppers are placed on the condenser tip
and connecting-tube tip to prevent entrance of dust particles when the
apparatus is not in use. Both caps are removed when the apparatus is
in use.
Reagents -
Blank water: Redistill a quality distilled or demineralized water
that has been passed through a bed of activated carbon (from which
most of the chloride has been leached) in the following manner.
Place in a 3-liter round bottom distilling flask containing several
boiling chips or beads, 500 ml pretreated water, 200 ml 0.25N
(approximate) potassium dichromate, 200 ml sulfuric acid reagent
and about 1 g HgS04. Swirl to mix and dissolve the HgSC>4. Mark
the level of this mixture on flask. Add about 1.5 liter of pre-
treated water and mix. Bring to a boil on the electric heater.
If a short-necked ST flask is used, a long glass stirring rod may
be kept in the flask while mixture is heating to prevent bumping.
Stir occasionally until boiling starts. Allow to boil for 10
minutes wasting steam to the atmosphere. Place the connecting
tube and condenser, and waste steam through the uncooled condenser
for 2 or 3 minutes. Start the cooling water and waste enough
condensate to rinse condenser.
Collect the distillate in a glass bottle rinsed several times with
distillate. A 9-lb acid bottle, well cleaned with chromic acid,
makes a good container. Use an adapter from the condenser tip into
bottle. The adapter tip should enter the receiver bottle through
a hole in a loose fitting foil dust cap. Distill until the level
in the flask drops to the mark.
Cap the water storage bottles when not in use.
The oxidizing mixture in the flask may be used repeatedly until
there is obvious discoloration.
Standard potassium dichromate solution, 0.025N: Prepare per
"Standard Methods" except prepare with blank water.
Sulfuric acid reagent: Place 21 g reagent grade silver sulfate in
a newly opened standard 9-lb bottle of concentrated reagent-grade
H2S04. With a flat tipped large diameter glass stirring rod, care-
fully crush any lumps of Ag2SC>4 and stir thoroughly, repeating as
often as necessary. Complete dissolution may be obtained in 20
minutes or less. Cap securely and mix thoroughly by inversion.
96
-------
Standard ferrous ammonium sulfate titrant, 0.01N: Prepare per
"Standard Methods," or preferably, 0.025N solution may be substi-
tuted and has been found to give reproducible results.
Ferroin indicator solution: Prepare per "Standard Methods."
Silver sulfate, reagent powder.
Mercuric sulfate, analytical grade crystals.
Procedure -
Set up two blanks for each set of samples.
Place 5 ml concentrated H2S04 in identified 500-ml boiling flasks
containing about 7 glass beads. Keep flasks capped with small
glass beakers at all times except when introducing samples and
reagents .
Add approximately 0.5 g HgSO^ and swirl to mix until HgSO^ is
dissolved.
Add 50.0-ml blanks and samples, or aliquots diluted to 50.0 ml
with blank water.
Add 25.0 ml 0.025N potassium dichromate solution swirling to mix
during addition.
Remove caps from reflux condensers arid top connecting tubes.
Start the cooling water through the condenser jackets. Turn
heaters on and set at high heat .
To each flask, add without mixing 70 ml Sulfuric acid reagent
solution) .
Taking each flask in turn swirl to mix contents quickly but
thoroughly and connect immediately to the condenser. Clamp and
lower the assembly to the heater. Boiling should start quickly
without bumping. When vigorous boiling is established, turn the
heater down to a setting to maintain boiling. Reflux for 2 hours.
At the end of the reflux time, taking each unit in turn, in the
same order as before, turn off the heater and raise the apparatus
several inches above the heater. Place an asbestos mat over the
heater. Allow to cool for 15 minutes.
Remove the connecting tube and. rinse down the condenser with a
little blank wa'ter. Allow about 10 - 15 seconds for draining then
lower the flask from the condenser. Rinse the condenser tip into
the flask with a little blank water and cap. Place the capped
flask in the cooling bath immediately. Replace the condenser and
connecting tube caps.
97
-------
While the flasks are cooling, prepare a ferrous ammonium sulfate
standardization. Place 25.0 ml of 0.025N potassium dichromate
solution and about 200 ml blank water in a 500-ml Erlenmeyer flask.
Add with swirling, 30 - 35 ml concentrated 1^804. Cover and place
in the water bath to cool. NOTE: If using 0.01N ferrous ammonium
sulfate, use 10.0 ml of 0.02.5 potassium dichromate solution for
the standard.
When the reflux flasks have cooled (cool to the touch) , carefully
add 200 ml blank water to each. Mix and return them to water bath.
When all flasks, including the standard, have cooled, add 2-3
drops of ferroin indicator to each. Titrate with ferrous
ammonium sulfate directly in the boiling flask. Do not transfer.
This is essential for consistent results.
Empty the flasks without removing the beads. Rinse the flasks
and beads thoroughly 3 times with distilled water and 3 times with
small portions (about 10 ml) of blank water. Replace the caps
immediately.
General Notes - The blanks should titrate to within 0.2 ml of the
standard if there is no contamination. Different lots of
solution will produce different blank titrations. With careful atten-
tion to cleanliness there will be less than 0.2 ml difference between
blanks. Within this limit use the higher blank titration result, not
the average. For very precise work (non-routine) the two blank titra-
tions should agree.
All glassware must be kept clean and dust-free. Use cleaning acid
followed by distilled water and finally blank water. To illustrate
the relevance of contamination, calculations indicate that in a 50-ml
sample cellulose dust weighing 0.05 mg would contribute about 1 mg/1
COD.
Calculations - Calculate mg/1 COD per "Standard Methods."
98 «J.S. GOVERNMENT PRINTING OFFICE:1973 514-153/233
-------
SELECTED WATER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
w
•5. Report Da^'-e
Activated Carbon Treatment of Raw Sewage in Solids-Contact 5.
Clarifiers „ „ , , ,
•8. PurfoTtnittg i
• Report tfo.
Beebe, Richard L.
.
Westinghouse Electric Corporation
INFILCO Division
Richmond,1 Virginia
17050 EGI
14-12-586
13. 'Type ce'Re port and
Period Cove-Ted •
.,!-/ -n v ••'!•-
U.S. Environmental Protection Agency, Environmental Protection Technology Series
Report EPA-R2T73-183. March 1973.
tfi. «?Mt/.,et Degritted-raw municipal sewage was treated with powdered activated carbon in
a 28,000-gpd pilot plant. Two high-rate recirculating-slurry solids-contact clarifiers
operating in series with countercurrent carbon advance, followed by a gravity polishing
filter, produced effluent, equal to or better than that produced in a parallel activated
sludge plant.
TOG and COD removals averaged 88,1 and 88.7 percent, respectively, with higher
removals hindered by the concentration of adsorptive-resistant materials present. ,
Filtrable-TOC and -COD removals were 68.0 and 69.9 percent, respectively.
Alum and polyelectrolyte flocculated the powdered activated carbon and raw -
sewage suspended solids into a fast settling floe. Subsidence tests conducted on the
solids slurry from the ACCELATOR® draft tube indicated ACCELATOR overflow rates equiva-
lent to or greater than 2.5 gpm/ft .
The maximum: carbon adsorptive capacity for filtrable COD was 0.50 to 0.55 g
COD/g carbon. This capacity was achieved whenever the concentrations of influent COD
and carbon matched or exceeded that ratio (adsorptive-resistant COD excluded). Carbon
requirements were 55 to 60 percent of theoretical two-stage countercurrent adsorption
system requirements.
Assuming regeneration cycles 85 percent of the carbon feed, respective treat-
ment cost estimates for 10-mgd and 100-mgd plants were 13.9c and 11.2c per thousand
gallons.
17a. Descriptors
*Adsorption, *Activated Carbon, *Sewage Treatment, Flocculation, Filtration,
Chemical Analysis, Waste Treatment, Cost, Settling Rates
17b. Identifiers
*Series clarifier countercurrent adsorption
19, S&a&rity Cfass.
('Report)
20. Security Cinss.
(Page)
21. tfo,-of
Pages
22. Pries
Send To:
WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON, D. C. 2O24O
A 6:
;-r, ; Richard L. Beebe
Mios Westinghouse/INFILCO, Richmond, Virginia
------- |