Environmental Protection Technology Series
FLOCCULATION-FLOTATION AIDS FOR
TREATMENT OF COMBINED SEWER OVERFLOWS
Municipal Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
EPA-600/2-77-140
August 1977
FLOCCULATION-FLOTATION AIDS FOR TREATMENT OF
COMBINED SEWER OVERFLOWS
by
N. F. Stanley and P. R. Evans
Hercules Incorporated
Allegany Ballistics Laboratory
Cumberland, Maryland 21502
Contract No. 14-12-855
Project Officers
Clifford Risley, Jr.
U.S. Environmental Protection Agency
Region V
Chicago, Illinois 60606
Richard Field
Storm and Combined Sewer Section
Wastewater Research Division
Municipal Environmental Research Laboratory (Cincinnati)
Edison, New Jersey 08817
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U. S. Environmental Protection Agency, and approved for publica-
tion. Approval does not signify that the contents necessarily reflect the
views and policies of the U. S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or recom-
mendation for use.
ii
-------
FOREWORD
The Environmental Protection Agency was created because of increasing
public and government concern about the dangers of pollution to the health
and welfare of the American people. Noxious air, foul water, and spoiled
land are tragic testimony to the deterioration of our natural environment.
The complexity of that environment and the interplay between its components
require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem solution
and it involves defining the problem, measuring its impact, and searching for
solutions. The Municipal Environmental Research Laboratory develops new and
improved technology and systems for the prevention, treatment, and manage-
ment of wastewater and solid and hazardous waste pollutant discharges from
municipal and community sources, for the preservation and treatment of public
drinking water supplies, and to minimize the adverse economic, social, health,
and aesthetic effects of pollution. This publication is one of the products
of that research; a most vital communications link between the researcher and
the user community.
Many metropolitan areas of the United States are served by combined
storm and sanitary sewer systems, which result in stream pollution when their
capacity is exceeded by severe runoff. This study reports the investigation
of one approach to the problem of economically retrofitting existing facili-
ties with an adequate pollutant removal process for peak flow periods.
Francis T. Mayo
Director
Municipal Environmental Research
Laboratory
iii
-------
ABSTRACT
The objectives of this study were to investigate the flocculation/flota-
tion characteristics of combined sewer overflow through laboratory and field
testing. The concept involves the introduction of chemicals and buoyant
flotation aids into the overflow and the subsequent coflocculation of the
suspended sewage solids about the aids which rise to the surface from where
they may be removed by skimming.
Laboratory parametric studies of three flotation aids and numerous co-
agulants and flocculants resulted in the batch demonstration of 70 to 100
percent suspended solids removal with 100 mg/1 of the polystrene flotation
aid Dylex KCD-340, 100 mg/1 of the coagulant .FeCl3 and 1 mg/1 of the floc-
culant Hercofloc 810.
Subsequent field evaluation in the Rexnord, Inc.-USEPA pilot flotation
facility in Milwaukee, Wisconsin, of the combination of chemicals and aids
selected in the laboratory tests was beset with difficulties including de-
creased flocculant strength due to long storage times and flotation aids with
poor buoyancy characteristics. Three field tests were devoted to solving
these problems. One final field test was achieved with most of the identi-
fied problems rectified and resulted in suspended sewage solids removal of
67 to 87 percent. An air flotation test conducted in the same facility for
comparison resulted in 50 to 77 percent suspended solids removal.
Recovery, cleaning and reuse of the flotation aids was judged essential
to economic feasibility for the flocculation/flotation process. Subsequent
testing of recovered used aids indicated a disparity in requirements. When
the bonding, or coflocculation, of suspended sewage solids to the fresh aid
surface was sufficiently strong to provide a high degree of suspended solids
removal, subsequent efforts to clean or remove the solids in anticipation of
aid reuse were unsuccessful. This deficiency, coupled with the oveaall me-
chanical complexity of the process, resulted in the investigation concluding
that the flocculation/flotation process is not as promising for broad field
application to the storm overflow problem as the similar dissolved air flo-
tation process. No further investigation of the flocculation/flotation
process is recommended.
This report was submitted in fulfillment of Contract No. 14-12-855,
under the sponsorship of the Office of Research and Development, US
Environmental Protection Agency. Work was completed as of January 1972.
IV
-------
CONTENTS
Foreword iii
Abstract iv
Figures vii
Tables viii
List of Abbreviations and Symbols x
Acknowledgements xii
1. Conclusions 1
2. Recommendation 3
3. Introduction 4
Background 4
Concept Description 4
Objectives 5
Alteration of Scope 5
4. Laboratory Tests 7
Introduction 7
Approach 7
Experimental Methods . . 7
Flocculant Preparation 8
Batch Flotation Tests 8
Suspended Solids, COD and pH Measurements 8
Phosphate Determination 8
Laboratory Scale Results 10
Qualitative Batch Tests 10
Quantitative Batch Tests 14
Screening 17
Batch Test Flotation Aid Recovery Tests 17
Open Tank Tests 22
Float Cake Tests 32
In-Line Tests 32
Dissolved Air Flotation Tests 41
5. Field Tests . 46
Design and Modification of the Demonstration
Facility ............. 46
Original Demonstration Facility
Modification of the Original Demonstration
Facility 49
Design of the Turbine-Flocculator 51
Flotation Aid Expansion and Feeding Procedures 51
Expansion of Flotation Aid 5].
Mixing of the Flotation Aids 51
-------
CONTENTS (Continued)
Operational Methods and Test Plan ........ 56
Operational Procedures 56
Sampling Procedures 56
Test Plan 56
Results and Discussion. ..... 59
Raw Wastewater Characterization and
Screening Efficiency .... 59
Operation of the Flotation System 59
Reevaluation of the Field Test Program Plan 64
Solution of the Major Technical Problems ....... 67
Operation & Evaluation of the Final Bead
Flotation Test ............ 69
Flotation Aid Recovery Reuse Tests .... 72
Conclusions 77
6. References. 79
VI
-------
FIGURES
Number Page
1 Concept schematic ................ 6
2 Photograph of Phipps and Bird mixer test .......... 9
3 Photomicrographs of flotation aids ............. 13
4 Laboratory open tank flow test apparatus .......... 23
5 Photograph of open tank flow test apparatus ......... 24
6 Photograph of open tank flow test apparatus ......... 25
7 Sketch of first in-line concentrations ........... 34
8 Sketch of second in-line test module ............ 35
9 In-line model design .................... 36
10 Original Hawley Road treatment facility ........... 47
11 Photographic view of Hawley Road site ............ 48
12 Flow schematic of modified demonstration facility ...... 50
13 Turbine flocculator ..................... 52
14 Bead expansion system ..................... 53
15 Schematic of bead mixing system ............... 55
16 Process schematic for head flotation test no. 5 ....... 71
17 Schematic diagram of Dylex recovery test .......... 75
vii
-------
TABLES
jSTumber Page
1 Flotation Aid Screening Test Results 11
2 Flotation Aid Properties and Costs . 12
3 Results of Quantitative Preliminary Laboratory Flotation
and Tests ............... • 15
4 Flocculant Screening Tests .... ... 16
5 Flocculant Dosage Tests (High Chemical Dosages) 18
6 Flocculant Dosage Tests (High Chemical Dosages) ...... 19
7 Laboratory Bench Scale Flocculation/Flotation Tests on
Milwaukee, Wisconsin, Combined Sewage Overflow 20
8 Effects of Screening Laboratory Batch Treatment Tests ... 21
9 Preliminary Open Tank Flow Tests, Allegany Ballistics
Laboratory Sewage ...... 26
10 Preliminary Open Tank Flow Tests, Bowling Green, Maryland,
Municipal Sewage ....... 27
11 Open Tank Flow Test Results On Line at Bowling Green,
Maryland, Municipal Treatment Plant . 29
12 Open Tank Flow Test Results, Bowling Green, Maryland,
Storm Overflow » 30
13 Flow Test Reproducibility 31
14 Float Cake Test Results 33
15 Results of In-Line Tests— First Test Module 37
16 Additional Results of In-Line Tests - First Test Module . . 38
17 Results of In-Line Tests - Second Test Module ....... 39
vnx
-------
TABLES (Continued)
Number Page
18 Dissolved Air Flotation - Bench Chemical Treatment Tests ... 42
19 Dissolved Air Flotation - Chemical Addition Technique Effects 43
20 Dissolved Air Flotation - Flocculation Time Effect ...... 44
21 Dissolved Air Flotation - Flocculant Type Effect ....... 45
22 Field Test Plan ....................... 57
23 Raw Combined Sewer Overflow Quality ............. 60
24 Screened Combined Sewer Overflow Quality ........... 61
25 Bead Flotation Effluent Quality ............... 62
26 Results of Run No. 71-4 ................... 65
27 Viscosity Data for Various Hereofloc Solutions With Time ... 70
28 Bead Flotation Effluent Quality ............... 73
29 Flotation-Aid Recovery Results ................ 76
IX
-------
LIST OF ABBREVIATIONS AND SYMBOLS
BGWTP
BOD
COD
D
Effluent
f pm
f ps
gpm
Influent
MGD
ppm
Re
SS
TOG
TS
TVS
V
VSS
P
Bowling Green Waste Treatment Plant
Biochemical Oxygen Demand
Chemical Oxygen Demand
Diameter (feet)
Treated flow
feet per minute
feet per second
gallons per minute
Raw (untreated) flow
Million gallons per day
parts per million
Reynolds Number (DVP/P)
Suspended Solids
Total Organic Carbon
Total Solids
Total Volatile Solids
Velocity (ft/sec)
Volatile Suspended Solids
Density, Ib/ft^
Viscosity, Ib/ft-sec
-------
LIST OF ABBREVIATIONS AND SYMBOLS (CONT'D)
cp centipoise
% percent
greater than approximately
> greater than or equal
^ less than or equal
~ approximately equal to
mg/1 milligrams per liter
= equal to
@ at
min minutes
hp horsepower
XI
-------
ACKNOWLEDGMENTS
Technical specialists within Hercules, Rexnord, Inc., and Koppers
Company have been consulted during the program.
Initially, Hercules Research Center, Wilmington, Delaware, was visited
to discuss flocculation, flotation, and waste water treatment techniques.
with Dr. D. Monagle, Dr. C. Smith, Dr. S. Pearson, and Mrs. P. Gilby.
Discussions concentrated on: (1) coagulant and flocculant candidates to be
evaluated in the program, (2) laboratory flocculation and flotation test
techniques, (3) probable coagulant and flocculant dosage effects, (4) com-
parative laboratory data on dissolved-air flotation and (5) the effect of
phosphate content on floe formation. Polyionic flocculants (Hercules,
Dow, and Calgon products) were supplied for evaluation.
Messrs. D. Mason and M. Guptu of Rexnord, Inc., Envirex Division,
Milwaukee, Wisconsin,were consulted to review their subcontract, the
pilot-scale facility modifications and laboratory results. Equipment modi-
fications required for the field tests were defined in these consultations.
Dylex beads are under development at Koppers Technical Center, Monroe-
ville, Pennsylvania. This facility was visited to review processing and
supply of the material for field tests. Dr. A. Ingram, head of the Kopper's
Expandable Polymers Group, supplied samples for evaluation. Mr. T. Altares,
Senior Research Scientist, provided valuable consultation services during
the field testing phase.
The support and guidance of the following United States Environmental
Protection Agency personnel are acknowledged:
Richard Field, Chief
Richard P. Traver, Staff Engineer
US EPA
Storm and Combined Sewer Section
Municipal Environmental Research Laboratory
Edison, NJ 08817
Stephen Poloncik
US EPA
REGION V
Chicago, IL 60604
XII
-------
SECTION 1
CONCLUSIONS
1. Laboratory testing of the flocculation/flotation process was suffi-
ciently successful to demonstrate the technical feasibility of the concept,
but field testing conducted by the Envirex Div. of Rexnord, Inc., at the
modified Rexnord-USEPA pilot-scale flotation facility located at Hawley Road,
Milwaukee, Wisconsin, revealed numerous operating difficulties of sufficient
severity to render that particular embodiment of the process impractical.
2. Laboratory open tank tests of dry weather ordinary sewage from four
different sources exhibiting influent suspended solids loadings from 8 mg/1
to 1200 mg/1 yielded highly variable removal rates, 0 to 100 percent, in
parametric tests of three flotation aids and numerous coagulants and floccu-
lants. Consistently good suspended solids removal rates, 70 to 100 percent,
were achieved with Koppers Co. Dylex KCD-340 heat-expanded polystyrene flo-
tation aid at a dose rate of 100 mg/1 coupled with 100 mg/1 of FeCl3 coagu-
lant and ] mg/] of Hercules Incorporated Hereofloc 8]0 flocculant.
3. A suspended solids removal level of 67 to 87 percent was demon-
strated in the last of five field tests at the Rexnord-USEPA pilot-scale flo-
tation facility. Overall performance of the process in the initial three
field tests was poor; numerous process startup difficulties were experienced
by Rexnord including mixed flocculant strength degradation under storage, and
poor flotation aid expansion/flotation. The suspended solids removal levels
achieved in the last field test are comparable to those achieved in test num-
ber 71-4 with air flotation, 50-77 percent.
4. The economics of the flocculation/flotation process dictate recovery
and reuse of the flotation aids (4 to li. per 1000 gallons of overflow with
90 percent flotation aid reuse versus 20 to 35^ per 1000 gallons, for 100
percent new aids) . The key element in flotation aid reusability is the re-
moval of coflocculated sewage solids from the aid surface so that a rela-
tively clean aid is available for subsequent reuse. Laboratory tests quali-
tatively indicated that suspended solids could be readily separated from the
flotation aid by a high shear mixing operation. However, suspended solids
removal tests on float cake (used flotation aids with coflocculated sewage
solids) from the last field test showed a high degree of resistance to re-
moval, with 70 to 90 percent solids retention being typical after vigorous
agitation. It was concluded that coflocculation adequate to ensure high
rates of suspended solids removal is probably inconsistent with flotation
aid recovery, placing the economics of the flocculation/flotation process in
serious question.
1
-------
5. The mechanical flocculation/flotation process is considerably more
cumbersome and complex than air flotation and is judged to have less poten-
tial for broad application.
6. The concept of introducing a mass of highly attracting artificial
coflocculation sites into storm overflow to enhance a rapid separation of
suspended sewage solids remains intriguing, particularly if the operation can
be conducted in situ in the sewer lines. However, many technological break-
throughs in flotation aids, aid recovery and processing will be required to
achieve worthwhile results.
-------
SECTION 2
RECOMMENDATION
No further experimental evaluation of the flocculation/flotation con-
cept discussed in this study is recommended.
-------
SECTION 3
INTRODUCTION
BACKGROUND
Pollution from storm and combined sewers is recognized as a significant
pollution problem. The magnitude of this source of pollution is indicated
in a recent study by the U. S. Environmental Protection Agency (1) which're-
ported that combined sewerage systems serve 29 percent of the total sewered
population of the United States.
Polluting substances discharge from combined storm and sanitary sewers
during wet-weather periods when the carrying capacity of the sewers is ex-
ceeded because of the excessive amounts of storm water entering them. The
capacity of the interceptor sewers largely determines the maximum flow to a
sewage treatment plant. Since the wet-weather flow from a combined sewer
system may be as much as 50 times the normal dry-weather flow, the cost of
providing treatment facilities and interceptor sewers capable of handling
the entire flow from combined collection sewers is prohibitive. Economical
methods for treatment of excess combined sewer flow are being sought. One
significant aspect of the problem is to rapidly separate dilute slurries,
suspensions, and colloidal dispersions of solids from large volumes of water
in an efficient, practical and inexpensive manner.
Hercules Incorporated offered to investigate a process believed to have
significant potential for the removal of suspended solids from combined
sewer overflow. The process involves the use of chemical flocculants to
agglomerate sewage solids about buoyant solid flotation aids which rise to
the surface of the waste stream where they can be removed.
CONCEPT DESCRIPTION
In conventional sewage treatment processes, chemical coagulants and
flocculants are commonly used to promote rapid solids separation. Both ef-
ficiency and degree of separation generally must be defined for each waste
and treatment process because of the diversity of waste characteristics.
Regardless of the application, however, the function of the chemicals is to
condition the solids suspension so that the particles attach to each other.
The attachment is achieved by reducing coulombic (van der Walls) forces be-
tween particles, by modifying surface properties of the solids or by adding
agents which bridge the solids. In the concept studied in this program, a
low-density solid flotation agent and chemicals were used to attach waste
-------
solids to the flotation agent. The resulting low-density floes float to the
surface for separation.
In the concept (Figure 1), a coagulant, a flocculant and a solid flota-
tion aid are admixed with combined sewer overflow to effect a high-rate flo-
tation of the suspended solids. Economic feasibility of the process clearly
depends upon the chemical dosages required plus recovery and reuse of the
low-density solids employed as flotation aids. The essence of this program
was an assessment of the concept by laboratory investigation of the process
parameters and pilot-scale investigation and evaluation under actual com-
bined sewer overflow conditions.
Although various aspects of flocculation have been studied (2), flow ef-
fects are not well defined for flocculation/flotation processes. Therefore,
to develop the concept for combined sewer overflow,the effect of flow tur-
bulence on flocculation and flotation was studied. In these studies, an in-
line flow turbulence concept was evaluated.
OBJECTIVES
The objectives of this investigation were (1) Phase I, to evaluate the
flocculation/flotation properties of combined sewage overflow, (2) Phase II,
to demonstrate a high-rate flocculation/flotation concept for suspended
solids separation under field conditions in combined sewer systems by uti-
lizing existing USEPA-Rexnord pilot-scale flotation equipment installed at
Hawley Road, Milwaukee, Wisconsin, and (3) Phase III, to conduct an engineer-
ing analysis and evaluation with a preliminary plant design.
SpecifLc objectives of Phase I were to evaluate selected chemical co-
agulants, flocculants, flotation aids and experimental techniques; to estab-
lish equipment modifications required for field testing in the USEPA-Rexnord
pilot-scale facility; and to assess a nonmechanical flow-turbulence, floccu-
lation/flotation concept as an alternative to the conventional mechanical
process.
The specific objectives of Phase II were to evaluate the best combina-
tion of coagulant, flocculant, and flotation aid discovered in Phase I in
the USEPA-Rexnord pilot flotation facility and to assess the feasibility of
recovering and reusing flotation aids in the field.
Alteration of Scope
During the startup and early test run portions of the field test pro-
gram, several significant and costly alterations in testing technique were
required. These were primarily associated with the proper preparation and
addition of flocculant and flotation aid. As a result, the work scope was
altered to maximize the amount of field test work possible without increas-
ing the overall contract cost. A planned Phase III, engineering evaluation
and preliminary plant design, was deleted, and the field test scope was re-
duced from a planned 20 storm events to 5 storm events. The sixth scheduled
field test was not completed because of the arrival of winter and shutdown
of the test facility.
-------
Inflow
Flotation Aid
Recovery
Flocculation
-»- Float Cake
Flotation
Treated
Effluent
Figure 1. Concept schematic.
-------
SECTION 4
LABORATORY TESTS
INTRODUCTION
The experimental approach, methods, laboratory-scale results and discus-
sion are presented in this section.
APPROACH
The experimental approach consisted of conducting preliminary experi-
ments at the Hercules/Allegany Ballistics Laboratory (ABL) facility, Cum-
berland, Maryland, in preparation for pilot-scale tests in the Rexnord, Inc.,
Envirex Division, flotation pilot plant in Milwaukee, Wisconsin. Prelimi-
nary experiments were conducted on dry-weather raw sewage at both the Hercu-
les facility and at the nearby Bowling Green, Maryland, domestic sewage
treatment plant. Actual combined sewer overflows were subsequently examined
only in tests at the Rexnord facility in Milwaukee.
In this discussion, the laboratory scale tests in Maryland are described
first. Experimental results and interpretation of results are then con-
sidered, following a brief review of the experimental methods. The following
review of experimental methods illustrates the types of tests performed to
identify flotation-aid and chemical dosages for actual storm flow testing in
Milwaukee, Wisconsin.
EXPERIMENTAL METHODS
Domestic sewage and, especially, combined sewer overflows vary signifi-
cantly in the nature of solids, solids concentrations and chemical charac-
teristics. Experimental investigations were therefore required to select
treatment chemicals, dosages and the sequence of addition to wastewater in a
treatment system. Suspended solids, chemical oxygen demand (COD), pH, phos-
phate and total solids were measured to establish initial field test condi-
tions. In addition to the above analytical tests, laboratory tests were de-
veloped to screen chemical coagulants, flocculants and flotation aids for the
field tests.
The tests employed were those given in Standard Meth_od_g__for the Examina-
tion of Water and Waste Water, 12 Ed. except as described below. Visual observa-
tion of the flocculation and flotation characteristics in batch tests pro-
vided the basis for laboratory-scale flow test techniques which were developed
-------
to establish the initial flow effects. These techniques are also considered
in this section.
Flocculant Preparation
To prepare stock flocculant solutions for the laboratory tests, % per-
cent anionic and 1 percent cationic solutions were prepared. Solution prepa-
ration consisted of stirring 0,25 or 0.50 g dry flocculant into 500 cc of
distilled water with sufficient agitation to create a vortex. Dry floccu-
lant was added slowly to the shoulder of the vortex, and the stirring was
continued for 45-60 minutes or longer, when required, to effect solution of
the flocculant.
Batch Flotation Tests
In the batch flotation tests, chemical coagulants, flocculants, and flo-
tation aids were added to 500-ml sewage samples contained in 1000 ml beakers
in the commercial six-station Phipps and Bird stirrer shown in Figure 2.
The mixing operation consisted of (1) agitation at 60-80 rpm for 1-2 rain to
suspend the solids, (2) addition of 1 percent chemical coagulant solution
while agitating at 5 rpm for 2 min, (3) addition of % or 1 percent floccu-
lant solution while agitating at 5-10 rpm for 2 min, and (4) addition of
flotation aid. Following addition of all chemicals, the mixture was either
tumbled in a 500-ml graduated cylinder or stirred vigorously to effect co-
flocculation of solids.
The flotation rate was estimated by timing separation of the floes and
clarified liquid. The clarified phase was then sampled for chemical analy-
sis .
Suspended Solids, COD and pH Measurements
Suspended solids were determined initially by the Gooch crucible asbes-
tos filter method and subsequently by the Millipore filter method. In the
latter tests, 100-ml samples were filtered through fiberglass filter paper
(934 AH which retains solids greater than 0.45 micrometer) and dried at
103°C for one hour.
The dichromate reflux method was used for COD analyses as specified
in Standard Methods for Examination of Water and Waste Water, 12 Ed. (1965).
The glass electrode method was used to measure pH. A commercial Beck-
man pH meter was employed.
Phosphate Determination
Essentially, the phosphate determination method used was that described
in Standard Methods for Examination of Water and Waste Water, 12 Ed. (1965).
The method selected was the aminonaphtholsulfonic acid method for ortho-
phosphate by colorimetric comparison of samples with standard solutions of
known phosphate content. A colorimetric with a red filter was used for the
comparisons .
-------
Figure 2. Photograph of Phipps and Bird mixer test.
-------
LABORATORY SCALE RESULTS
Qualitative Batch Tests
Flotation Aid Tests
Screening studies were conducted on a variety of flotation aid candi-
dates (Table 1). Materials of composition included glass, phenolic and
polystyrene. The test medium was a sample of Bowling Green, Maryland, do-
mestic dry-weather sewage. Candidates were accepted for further testing,
or rejected, based on visual inspection of flotation rate and'the cofloc-
culation* propensity. Coagulant, flocculant and flotation-aid dose rates
covered the ranges of 25-100 mg/1, 0.4-1.0 mg/1 and 10-100 mg/1, respectively.
Selected for further evaluation using quantitative tests were the heat-
expandable polystyrene Dylex KCD-340, glass Ecjospheres IG-101 and FT-102 and
phenolic microballoons BJO-0840 and BJO-0930, the other materials listed in
Table 1 giving markedly poorer performance. Physical property and cost data
for the acceptable flotation aids are given in Table 2. Photomicrographs of
these materials (Figure 3) are indicative of their geometry, size distribu-
tion, and surface textures.
Coagulant Tests
In addition to the tests with ferric chloride, comparative tests employ-
ing alum (Al£(304)3 • 18 H20) as the coagulant were conducted. Both Dylex
KCD-340 beads and glass microballoons were tested as flotation aids with a
variety of flocculants. Flocculant dosages tested were 1 mg/1 with alum
dosages of 25 mg/1 and 100 mg/1 for all flocculants. Although the tests
showed good visual flocculation for all chemical combinations, there was
only limited visual suspended solids coflocculation on the flotation aids.
Because of this difficulty, floes settled while the relatively clean flota-
tion aid floated. The tests with alum were therefore discontinued.
Experiments with lime as coagulant gave similar results; i.e., floe
growth without significant coflocculation. Consequently, tests with lime
were also discontinued and FeCl3 was selected as the coagulant for all sub-
sequent testing.
Treatment Chemical Sequencing
The batch tests included qualitative investigation of the order of coag-
ulant and flocculant additions. Addition of flocculant before the coagulant
appeared to have an adverse effect on floe formation, whereas additicti of
coagulant before the flocculant promoted floe formation. The order of flo-
tation aid addition, on the contrary., did not significantly alter the floc-
culation or subsequent flotation rate of resultant floes. The flotation aid
sequence tests included tests with glass microballoons, phenolic microbal-
loons and Dylex beads. All materials could be added before, after, or with
the ferric chloride coagulant or flocculant.
'-'Coflocculation in the context of this study is defined as the formation of
flocculating sewage solids about the flotation aid.
10
-------
TABLE 1. FLOTATION AID SCREENING TEST RESULTS
Flotation Aid
Results
a,b
Dylex KCD-340C
Dylex Superfine0
Dylex (> 400 micron)c
Polystyrene PYF-2921C
Eccospheres IG-101d
Eccospheres FT-102d
Bakelite BJO-08406
Bakelite BJO-09306
Very fast flotation rate and good cofloccu-
lation.
Unacceptably slow flotation rate.
Unacceptable coflocculation.
Unacceptably slow flotation rate.
Fast flotation rate and good coflocculation.
Fast flotation rate and good coflocculation.
Fast flotation rate and good coflocculation.
Fast flocculation and good coflocculation.
NOTES:
Results were visual in nature and qualitative only, with no flo-
tation aid rise velocities or sewage solids removal tests made.
Coagulants, flocculants and flotation aid dosages were started
low and progressively increased for each aid investigated.
Coagulants employed were FeCl3, alum and lime in doses of 25 and
100 mg/1. Flocculants were polyionic Hercules Incorporated
products used in doses of 0.5 and 1.0 rag/I. Flotation aid doses
were 10 and 100 mg/1.
GPolystyrene microballoon, product of Koppers Co., Inc., Pitts-
burgh, Pennsylvania.
'-'class microballoon, product of Emerson and Cumings, Inc., Canton,
Massachusetts.
ePhenolic microballoon, product of Union Carbide Corp., New York,
New York.
fSample was domestic dry weather sewage from Bowling Green, Mary-
land with a suspended solids level of about 100 mg/1.
11
-------
TABLE 2. FLOTATION AID PROPERTIES AND COSTS
Bouyant Density Nominal Diameter
Flotation Aid gm/cc Microns $/lb
Eccospheres (glass)
IG-101
FT-102
Bakelite Spheres
(phenolic)
BJO-0840
BJO-0930
0.
0.
0.25
0.
34
26
-0.35
25
250
10-250
10-150
10-150
0.69
2.00
0.86
0.95
Dylex Beads (heat
expandable poly-
styrene)0
KCD-340 0.02 200-300 0.40
Emerson and Cumings, Inc., Canton, Massachusetts.
°Union Carbide Corp. New York, New York.
GKoppers Co., Inc., Pittsburgh, Pennsylvania.
"Prices based on 100 ton quantities.
12
-------
Dylex KCD 340 Beads ( 50X)
Lot 296.01
Glass Microbal loons ( 100X)
Lot IG 101
Bakelite Microbal loons
(100X)
Lot BJO 0930
Figure 3. Photomicrographs of flotation aids,
13
-------
Mixing Condition Effects
Qualitative tests were also conducted to assess the effects of coagu-
lant and flocculant mixing times with evaluation of effects attributable to
changes in sample properties such as pH, suspended solids and chemical
oxygen demand. These experiments were qualitative rather than quantitative
because observation of floe growth revealed that visual appearance of floe
was prerequisite for coflocculation of sewage solids and flotation aid. In-
itial vigorous agitation prevented floe growth whereas gentle agitation pro-
moted floe growth without affecting the coflocculation of floe and flotation
aid. Accordingly, it was necessary to gently agitate the system to promote
floe growth and then increase the agitation (or shaking) to effect cofloccu-
lation of sewage solids and solid flotation aid. Further, these tests sug-
gested (qualitatively) that the coagulation/flocculation operation is a ki-
netic process affected by both the coagulant and flocculant concentrations;
that is, higher chemical concentrations enhanced the visual coagulation and
flocculation rates. Since it was not the purpose of this study to define
such kinetic effects in a simulated system, mixing time and action which
seemed adequate for the desired coflocculation were selected for subsequent
testing.
Quantitative Batch Tests
Flotation Aid Tests
A series of tests were conducted to determine the quantitative effects
on suspended solids removal of flotation aid dose, flocculant type, coagu-
lant and flocculant dosages and raw sewage screening. All tests employed
batch samples less than 8 hours old to minimize aging effects.
Table 3 reports the suspended solids removal efficiencies of the vari-
ous candidate flotation aids using 100 mg/1 of FeCl^ and 1 mg/1 of Herco-
floc 810, a cationic flocculant. Results were quite variable, with re-
moval percentages ranging from 0 to 100 percent suspended solids. Dylex
polystyrene and Eccosphere glass flotation aids performed best with sus-
pended solid removals ranging from 52 to 100 percent. Phenolic/Bakelite
flotation aids resulted in markedly poorer performance. The low suspended
solids content, 8-24 mg/1, of the Bowling Green domestic dry-weather sani-
tary sewage was considered to be unrepresentative of highly solids laden
storm overflow material. Its use was discontinued in favor of grab samples
jrom the Frostburg, Mary land, municipa1 treatment plant influent for further
testing. „
Flocculants
A broad range of commercially available flocculants was screened using
Koppers Co. polystyrene KCD-340 flotation aid, 100 mg/1 FeCl3 coagulant and
ordinary dr^-weather sanitary sewage influent from the Frostburg, Maryland,
sewage treatment plant. Flocculant dosages were 1 mg/1 in all cases. Re-
sults are presented in Table 4. In these tests, removal efficiencies were
similar for many cationic and anionic flocculants but some of the anionic
candidates clearly showed reduced suspended solids removal. Two of the ani-
onic flocculants showed essentially no suspended solids removal (Calgon
ST2b9 and Nalco 2066). On the basis of these results, the cationic floccu-
lants, as a group, were better than the anionic flocculants. However,
14
-------
TABLE 3. RESULTS OF QUANTITATIVE PRELIMINARY
LABORATORY FLOTATION AND TESTS^
Flotation
Aid
Dylex (KCD-340)
Sccospheres
(IG-101)
Bakelite Spheres
(BJO-0930)
Dosage
mg/1
10
10
20
100
100
10
100
10
100
K
S/Fb
2/1
1/1
1/1
1/5
1/4
1/1
1/10
1/1
1/10
c
Raw Sewage
SS mg/1
21
11
21
21
24
21
8
8
8
Treated Sewage
SS mg/1
9
0
10
6
7
1
3
8
4
Removal
%
57
100
52
71
71
95
62
0
50
NOTES:
Tested with 100 mg/1 FeCl3 and 1 mg/1 Hercofloc 810.
bS/F - Suspended Solids/Flotation Aid Ratio.
cBowling Green Waste Treatment Plant dr>-weather domestic sani-
tar> influent.
15
-------
TABLE 4. FLOCCULANT SCREENING TESTS3>b
Post Treatment
Suspended Solids
Flocculant mg/1
Ca tionic
Hercofloc 810C
Dow 1629. ld
Nalco 610e
Betz 1260f
An ionic
Hercofloc 816°
Hercofloc 822C
Dow A- 23d
Calgon ST2698
Nalco 2066e
Hercofloc X8-lc
Hercofloc X8»2S
Hercofloc X8-3C
Hercofloc X8-4C
34
77
38
58
491
387
54
1205
1164
5b
72
45
30
Percent
Removal
97
93
97
95
57
66
95
0
0
95
94
96
98
Post Treatment
COD Removal
mg/1 %
224 80
291 74
« «
267 76
300 73
333 70
275 75
262 76
NOTES:
aAll samples treated with 100 mg/1 FeCl3, 100 mg/1 Koppers Co. Dylex
KCD-340 flotation aid and 1 mg/1 flocculant.
Test medium was a grab sample of dry-weather ordinary domestic
sewage taken from the influent lines of the Frostburg, Maryland,
municipal sewage treatment plant. The suspended solids level
was 1134 mg/1 and COD was 1104 mg/1.
*
GProduct of Hercules Incorporated, Wilmington, Delaware.
Product of Dow Chemical Co., Midland, Michigan.
ePrcduct of Nalco Chemical Co., Oak Brook, Illinois.
Product of Betz Laboratories, Trevose, Pennsylvania.
gProduct of Calgon Corp., Pittsburgh, Pennsylvania.
16
-------
because of the inconsistency in the results, additional tests were conducted
with the apparently best candidates. In these tests, reported in Tables 5
and 6, all three candidate flotation aid materials were utilized, and both
high and low coagulant and flocculant dosages were investigated. Comparison
of Table 5 with Table 6 shows (1) that the lower chemical dosages were as
effective as the higher dosages and (2) that the cationic flocculants were
generally better than the anionic flocculants, particularly for the Dylex
bead flotation system,
ivr Lflo-'.: C-31, a cationic flocculant, has been employed successfully in
dissolved-air flotation tests to promote solids removal efficiency so this
flocculant was also tested here. Purifloc C-31 and Nalco 607-C were substi-
tuted for Hercofloc 810 during one phase of the bench Dylex flotation tests
on Milwaukee combined sewage to assess the sensitivity of the Milwaukee sew-
age to coflocculation with different flocculants. Results were visual quali-
tative only and are shown in Table 7. In all cases, an excellent floe was
fomed but no coflocculation occurred with the Nalco 607-C or the Purifloc
C-31 while good coflocculation occurred with the Hercofloc at 1 mg/1. The
floe was Less lightly coflocculated to the flotation aid for this sewage than
had been the case with the dry-weather Frostburg, Mary land, sewage.
Based on these initial laboratory tests, Hercofloc 810 was selected as
the field test flocculant.
Screening
The effects of screening the raw sewage influent were briefly investi-
gated using Frostburg, Mary land, dry-weather domestic sewage. Tests repre-
sentative of both high and low suspended solids concentrations were con-
ducted. R-esults, reported in Table 8, indicate that removal efficiency by
the flocculation/flocation method is relatively insensitive to screening.
Batch Test Flotation Aid Recovery Tests
From both sludge handling and chemical cost viewpoints, recovery of
flotation aid is desirable and probably essential. Preliminary qualitative
tests showed that flotation aid can be recovered by high-shear agitation of
the float cake. In these tests, a commercial Waring blender was used to
promote the separation. Flotation aid separated from the coflocculates sew-
age solids aad floated whereas most of the sewage solids settled.
Float cake samples for the Waring Mender experiments were prepared by
treating 500 cc of Bowling Green, Maryland,dry-weather municipal waste with
(0.01 gm) 20 mg/1 Dylex KCD-340 flotation aid, 100 mg/1 FeCl3 (0.05 grn) and
1 mg/1 Hercofloc 810 flocculant. The "recovered" flotation aid plus resid-
ual solids were dried to a constant weight of 0.017 gm which confirmed that
there were residual solids which did not separate from the flotation aid.
The settled solids were also filtered and dried. However, a material balance
on these small sample tests could not be achieved so tests to define recov-
ery efficiency were deferred until larger open tank flow test samples could
be obtained.
17
-------
TABLE 5. FLOCCULANT DOSAGE TESTS (LOW CHEMICAL DOSAGES)3 >b
Suspended Solids in Effluent
Flocculant
Dylex KCD-340
SS Removal
mg/ 1 %
Glass IG-101
SS Remo\al
mg/ 1 %
Phenolic BJO-0930
SS Removal
mg/1 %
Cationic
Hercofloc 810 25 91 41 85 29 89
Nalco 610 37 86 49 82 42 85
Betz 1260 38 86 58 79 52 81
Anionic
Dow A23
Hercoflox X8-4
Hercofloc 822
49
41
-
82
85
Neg.
228
71
-
16
74
Neg.
80
65
101
71
76
63
NOTES: aA11 sampies tested with 100 mg/1 flotation aid, 25 mg/1 FeCl3 and
0.5 mg/1 flocculant.
^Treatment medium was a grab sample of dry-weather ordinary sani-
tary sewage from the Frostburg, Maryland, municipal treatment
plant with a raw sewage suspended solids level of 271 mg/1. *
18
-------
TABLE 6. FLOCCULANT DOSAGE TESTS (HIGH CHEMICAL DOSAGES)a>b
Suspended Solids in Effluent
Eccospheres
Flocculant
Cationic
Hercofloc 810
Nalco 610
Betz 12oO
Anionic
Dow A23
Hercofloc X8-4
Hercofloc 822
Dylex KCD-340
SS Removal
rag/1 %
3 87
5 79
5 79
No Cofloccula-
tion
11 ii
TG-101
SS
mg/1
8
6
8
3
2
0
Removal
70
67
75
67
87
92
100
Bakelite
BJO-
SS
mg/1
4
5
8
3
5
5
0930
Removal
7,
83
79
67
87
79
79
NOTES:
aAll samples tested with 100 mg/1 flotation aid, 100 mg/1
and 1.0 nig/I flocculant.
Treatment medium was a grab sample of dry weather ordinary sani-?
tary sewage from the Frostburg, Mary land, municipal treatment plant
with a raw sewage suspended solids level of 241 mg/1.
19
-------
TABLE 7. LABORATORY BENCH SCALE FLOCCULATION/
FLOTATION TESTS ON MILWAUKEE, WISCONSIN,
COMBINED SEWAGE OVERFLOW3
'^'locculant
Flocculant Concentration
(mg/1)
Results
Puri floe C-31
Nalco 607-C
Hercofloc 810
1.0
2.0
4.0
4.0
c,.0
0.5
1.0
2.0
Good flocculation - no co-
flocculation
Good flocculation - no co-
flocculation
Excellent flocculation - no
coflocculation
Good flocculation - no co-
flocculation
Good flocculation - no co-
flocculation
Good flocculation - fair co-
flocculation
Excellent flocculation -
good coflocculation0
Excellent flocculation -
good coflocculation
NOTES:
aTests conducted on a raw Milwaukee, Wisconsin, combined sewer over-
flow sample collected on 6/8/71 and stored under refrigeration un-
til testing on 7/8/71. FeC13 was added to each test sample (40
mg/1) and agitated before addition of the flocculant and Dylex flo-
tation aid (100 mg/1).
"Good flocculation" in the context of this investigation mean's that
suspended solids formed quickly into large, tightly knitted floes
capable of rapid settlement.
c"Good coflocculation" in the context of this investigation means
that flocculation occurred about the flotation aid as a nucleation
site and was sufficiently tightly bound to the aid to rise to the
liquid surface with it while the liquid was being vigorously agi-
tated.
20
-------
TABLE 8, EFFECTS OF SCREENING LABORATORY
BATCH TREATMENT TESTSa
High Suspended Solids Testb
Dylex KCD-340 Screened Sewage0 Raw_Sewage
Dosage (mg/1)
100
50
10
100
50
10
SS (mg/1)
^9
83
167
T _ T ,
J_rOU
40
33
58
Removal (%) SS
97
95
12
Avg . 94
Suspended Solids Testd
72
77
60
Avg, 70
(mg/D
133
116
74
43
39
45
Removal (%)
94
95
22.
Avg. 95
80
82
_79
Avg. 80
AOTES: ^chemical dosages were Hereofloc 810 at 0,5 mg/1 and FeCl3 at 25
^Drv-weather ordinary domestic raw sewage from the Frostburg,
Marvland, treatment plant with an unscreened raw sewage suspended
solids level of 2345 mg/1 and a screened suspended solids level
of 1584 mg/1.
C3creened sewage was passed through a U.S. Standard No. 50 screen
297 micron openings).
^Dry-weather ordinary domestic raw sewage from the Frostburg, Mary
land, treatment plant with an unscreened raw sewage suspended
solids sewage level of 218 mg/1 and a screened suspended solids
level of 145 mg/1.
21
-------
Open Tank Tests
A series of laboratory-scale open tank flow tests were performed to
evaluate flow effects on the mechanical flocculation/flotation process and
to relate laboratory batch test results to field condition.
Test Apparatus
The open tank laboratory test apparatus is described by a schematic,
(Figure 4) and photographs (Figures 5 and 6).
Experimental Plan
The experimental plan consisted of seeking a chemical dosage optimiza-
tion by varying the flotation retention (by varying the flow rate through the
apparatus) from 2.5 to 10 min^ the FeCl3, coagulant dosage from 0 to 125 mg/1,
and the suspended solids influent loading. The baseline flotation aid, Dylex
KCD-340, was slated for all tests except a check using one other aid. Flota-
tion aid dosage was 100 mg/1 for all tests. The effects of screening versus
nonscreening were to be evaluated.
Test Results
Preliminary laboratory-scale open tank flow tests of the flocculation/
flotation concept and laboratory apparatus were conducted on samples of dry-
weather ordinary domestic sewage from the influent to the sewage treatment:
plant for Hercules Incorporated/Allegany Ballistics Laboratory (ABL) located
near Cumberland, Maryland, and from the influent line for the Bowling Green,
Maryland, municipal treatment facility. The unit was tested first at Alle-
gany Ballistics Laboratory (ABL) in 1-hour tests of about 50-gallon sewage
samples. Table 9 gives the preliminary flow test results for the ABL experi-
ments. Excellent removal efficiencies (equal to or greater than 95 percent)
were obtained in these short tests.
Results of the tests on Bowling Green sewage (Table 10) were less suc-
cessful with suspended solids removals of 39-60 percent. Ranges of FeCl3 co-
agulant dosages (0-50 mg/1) and retention times (0.1 to 0.4 min) were tested
on the low suspended solids laden samples (91 to 183 mg/SS). In Tests 4-8,
an estimated 25 percent of the inflow was diverted and treated with the chemi-
cals before combining this side stream/fraction with the balance of the in-
flow to simulate anticipated operations at the USEPA-Rexnord Milwaukee pilot
facility installation, The split flow operation did not markedly affect sus-
pended solids removal (based on comparison of Test 3 with Tests 6 and 7).«
Representative flow proportioning was complicated by rapid changes in
the character of the flow during the tests. Some inflow samples contained
large amounts of grit and particulate solids, whereas other samples were
relatively clean or contained significantly more fibrous material. The re-
ported results are for composite - samples obtained by mixing discrete samples
taken about every 15 minutes during a flow test and, therefore, may not pre-
cisely reflect changes in flow or waste characteristics during a test.
A series of open tank flow tests were conducted on line at the Bowling
Green, Maryland, sewage plant, including two tests on combined sewer overflow. The
22
-------
COAGULANT
PERISTALTIC PUMP
(VARIABLE SPEED)
OJ
100-GAL SAMPLE RESERVOIR
FLOCCULANT AND FLOTATION
AID
FLOAT CAKE
o
0°°°°
CLARIFIED EFFLUENT
15 CAL FLOTATION
CHAMBER
OPERATING CHARACTERISTICS
FLOW RATE
SURFACE LOADING
HORIZONTAL VELOCITY
RISE RATE
RETENTION
1-3 GPM
0.5-1.5 GPM/FT2
0. 1-0.4 FPM
~ 5 FPM
1-10 MIN
Figure 4. Laboratory open tank flow test apparatus.
-------
FLOTATiON CHAIV1BER|
FLOTATION AiD|
\ PERISTALTIC PUMP
,ICONTROL PANEL
INFLOW SCREEN
Figure 5. Photograph of open tank flow test apparatus.
-------
Figure 6. Photograph of open tank flow test apparatus,
-------
TABLE 9. PRELIMINARY OPEN TANK FLOW TESTS, ALLEGANY
BALLISTICS LABORATORY SEWAGEa>b
FeCl3 Flotation I
Test No. (mg/1) tionc (n
1 125
2 45
3 50
4 50
NOTES: aTest medium was
5
10
5
10
grab
leten- Suspended Solids
iin) Raw (mg/1) Treated (mg/1) Removal (7»)
378
378
1170
1170
samples of dry weather
18
12
53
55
ordinary
95
97
95
95
sanitary sew-
age taken from the influent line of the treatment plant for Hercu-
les Incorporated/Allegany Ballistics Laboratory, near Cumberland,
Maryland.
bDosages were 100 mg/1 of Dylex KCD-340 flotation aids and 1 mg/1
of Hercofloc 810 flocculant. Flow rate was 1 gal/min.
Q
Flotation retention is defined as average residence time after
addition of flocculant. *
26
-------
TABLE 10. PRELIMINARY OPEN TANK FLOW TESTS,
BOWLING GREEK, MARYLAND, MUNICIPAL SEWAGEa>b
Ferric Chloride
Test No.
1
2
3
4
5
6
7
8
Flow
(gpm)
1
1
2
2
(split)
2
(split)
2
(split)
2
(split)
2
(split)
Dosage
(mg/1)
25
50
25
25
50
25
25
0
Reten-
tion0
(min)
0.2
0.2
0.1
0.4
0.4
0.4
0.4
-
Flotation
Reten-
tiond
(rain)
5
5
5
2.5
2.5
5
5
5
Suspended Solids
Raw
(mg/1)
91
155
155
172
172
123
183
124
Treated
(mg/1)
55
61
75
105
79
72
100
83
Removal
(7o)
40
60
52
39
54
41
45
33
NOTES: a
Test medium was grab samples of dry weather ordinary domestic raw
sewage from the Bowling Green, Maryland, municipal treatment facil-
ity.
bDosages were 100 mg/1 of Dylex KCD-340 flotation aids and 1 mg/1 of
Hercofloc 810.
GFeCl3 retention is defined as average retention time before floccu-
lant addition.
flotation retention is defined as average retention time after
flocculant addition.
27
-------
results of the on-line, dry-weather flow are summarized in Table 11. In
these tests, removal efficiencies varied between 0 and 88 percent. Lower
suspended solids removals were obtained in tests without coagulant. Se-
lected tests (Tests 4-8) were conducted with an estimated 25 percent of the
inflow treated with chemicals before combining this "slip stream" fraction
with the balance of the inflow to simulate the Milwaukee pilot facility in-
stallation. The split flow operation did not adversely affect suspended
solids removal.
The results obtained suggest that flow effects are significant. Al-
though a coagulant was generally required to effect coflocculation of sewage
solids with the flotation aids in the batch tests, significant amounts of
suspended solids were removed with and without ferric chloride as coagulant
in the flow tests.
Chemical dosages appeared to affect suspended solids removal efficiency.
When the flocculant dosage was reduced to less than 1 ppm (Tests 22 and 23),
solids removal decreased. In Test 21, with 4 ppm flocculant, suspended
solids removal improved slightly. In Test 26, solids removal increased to
76 percent with relatively large dosages of both coagulant (100 mg/1) and
flocculant (7 mg/1). In Test 27, flocculant dosage was halved to 2 mg/1
with the same coagulant dosage and solids removal decreased to 65 percent.
These results indicate that relatively large chemical dosages, particularly
coagulant dosages, are probably required to achieve high removal efficien-
cies for municipal sewage.
The results of two tests conducted on combined sewer overflow are sum-
marized in Table 12. Raw sewage was screened through a U.S. No. 50 screen
in both tests. Suspended solids removal was 79 percent in the first test
using Bakelite/phenolic microballoons as the flotation aid. Suspended solids
removal on the second test decreased 28 percent. Dylex KCD-340 polystyrene
flotation aids were used for this test. The reduced suspended solids removal
efficiency is attributed in part to the change in influent solids rather
than to the flotation aid replacement. These results are interpreted as
suggesting that the initial flush is more easily flocculated and more easily
treated than the subsequent flow because of higher suspended solid concen-
trations.
Reproducibility of the Test Results
Reproducibility of the flow test results was determined (Tests 17*-20,
Table 13) to provide a basis for statistical interpretation of results. An-
alysis of the data reproducibility permits interpretation of process varia-
ble relationships in terms of statistical significance. Significance of in-
dicated effects was determined by comparing the estimated variance of
grouped data with the estimated experimental variance via the statistical
"F" tests (3-4). These data show a variance of 260 (standard deviation =
16.1).
Another useful statistical parameter for characterizing test reproduci-
bility is the statistical coefficient of variation (Cv = standard deviation/
mean x 100). For these experiments, Cv is about 11 percent which means that
28
-------
TABLE 11. OPEN TANK FLOW TEST RESULTS ON LINE AT
BOWLING GREEN, MARYLAND, MUNICIPAL TREATMENT PLANT
Ferric Chloride Suspended Solids
Test No.
16
17
18
19
20
21
25
26
27
11
12
13
14
22
23
Flotation Aid
Dylex,KCD-340
Dylex,KCD-340
Dylex,KCD-340
Dylex,KCD-340
Dylex,KCD-340
Dylex,KCD-340
Dylex,KCD-340
Dylex,KCD-340
Dylex,KCD-340
Phenolic, 0930
Phenolic, 0930
Phenolic, 0930
Phenolic, 0930
Phenolic, 0930
Phenolic, 0930
Hercofloc 810
Dosage (mg/1)
1
1
1
1
1
4
1
4
2
1
1
1
1
0.5
0.25
Dosage
50
0
0
0
0
0
25
100
100
25
25
50
0
0
0
Raw Screened Removal
(mg/1) (mg/1) (%)
295
295
295
295
295
295
199
199
199
123
259
212
227
98
98
Treated
36
123
152
159
134
112
146
48
70
90
215
151
153
86
98
Removal
88
58
48
46
55
62
27
76
65
29
17
29
33
10
0
NOTES:
Test medium was in-line flow of dry weather, ordinary domestic sewage from the Bowling
Green, Maryland, municipal treatment plant.
Flotation aid dosage was 100 mg/1.
cFlotation tank retention time was 5 min.
Hydraulic loading was 1 gpm/ft^.
-------
LO
O
TABLE 12. OPEN TANK FLOW TEST RESULTS,
BOWLING GREEN, MARYLAND, STORM OVERFLOW
Hercofloc 810
Dosage
Test No. Flotation Aid (mg/1)
1 Phenolic, 0930 1
2 Dylex, KCD-340 1
Ferric Chloride
Dosage
(mg/1)
25
25
Raw
(mg/1)
282
112
Suspended Solids
Screened
(mg/1)
177
79
Removal Treated
(%) (mg/1)
37 37
29 57
Removal
(%)
79
28
NOTES: flotation aid dosage was 100 mg/1.
flotation tank retention time was 5 min.
cHydraulic loading was 1 gpm/ft^.
-------
TABLE 13. FLOW TEST REPRODUCIBILITY
Test No.
17
18
19
20
Effluent
1
120
154
168
135
Suspended
2
125
149
149
133
Avg.
Solids
Avg.
123
152
159
134
142
aComposite influent sample for all four tests
contained 295 mg/1 suspended solids.
Suspended solids tests were run in duplicate.
Experimental conditions were 100 mg/1 Dylex
KCD-340 flotation aid, 1 mg/1 Hercofloc 810
flocculant and no coagulant. Flow was 1 gpm
with flotation chamber retention of 5 min.
31
-------
the experimental values presented are reproducible within about 11 percent.
Float Cake Tests
Float cake from the Bowling Green storm overflow tests was character-
ized. Results of the float cake tests are given in Table 14. The samples
tested were skimmed from the flotation tank, weighed, filtered and dried.
Float cake and filter cake densities were similar for both Dylex and pheno-
lic flotation-aid systems, but the former dewatered in less than one half
the time.
In-Line Tests
In the flow-turbulence concept for in-line concentration of suspended
solids sewage, flocculation and flotation unit operations are performed in
a pipeline by flow turbulence. As in the mechanical process, the concen-
trated flow is then passed to the treatment plant while the clarified
effluent flows to the river.
Test Apparatus
Two flow-turbulence modules were built and tested. Essentially, inves-
tigation of the concept consisted of operating the units with various flow
conditions and chemical treatment combinations. The flow-turbulence con-
centrator design first tested is shown in Figure 7. The unit, made of
cellulose acetate, was designed to permit visual observation of flocculation
and flotation at various levels of flow turbulence. The unit was designed
to allow both laminar and turbulent flow conditions to be investigated.
The second flow turbulence module (Figure 8) was designed with a smooth
surface and tapered chamber to promote float cake flow and separation.
Also, the module design was selected on the basis of geometric similitude
to keep similar Reynolds numbers for use in scaling the laboratory module
from 30 gpm to a reasonable demonstration scale module (0.5 MGD capacity).
In scale-up operations of this nature, it is of course desirable to main-
tain complete geometric similarity but this is not possible here since
retention time may also be important. For a fixed liquid flow system (i.e.,
constant liquid density and viscosity), Reynolds number varies inversely
with the module diameter and residence time varies directly with the module
volume. In a truncated conical module, volume varies with length and
diameter squared; volume = ir/12 (Di2 + ^>i^>2 + ®22) • Therefore, as a diameter
is increased to maintain the Reynolds number in higher capacity modules, the
length similitude cannot be maintained (Figure 9).
Test Results and Discussion
The results of the in-line flow-turbulence tests are summarized in
Tables 15, 16, and 17.
Table 15 shows that both solids and phosphate removal efficiencies
improve as the coagulant dosage increases. From a practical standpoint,
optimum removals were obtained with the 100 mg/1 ferric chloride dosage.
About 90 percent removal of suspended solids and 70 percent phosphate
removal were obtained at this dosage level. As expected, effluent pH
32
-------
TABLE 14. FLOAT CAKE TEST RESULTS
Parameter
Flotation Aid
Value
Dylex KCD-340 Bakelite 0930
Dosage, mg/1
Coagulant
Test Flow, gpm
Float Cake Bulk Density, gm/cc
Float Cake Moisture (%)
Filtration Time,a min.
Filter Cake Bulk Density, b gm/cc
Filter Cake Moisture (%)
100
FeCl3
2
0.9
95
2
0.7
86
100
FeCl3
2
0.9
96
5
0.8
87
NOTES: a.
Filtration time to dewater 100 cc float cake through a Watman
No. 5 filter paper in a standard Millipore suspended solids
apparatus.
•'Difficult measurement with small sample, tests repeated once
and averaged.
33
-------
2 IN. X 3/8 IN. ID NOZZLE
2 IN. X 3/4 IN. I.D. NOZZLE
2 IN. X 1/4 IN. BOLTS (4) / 3_3/4 IN> x 4_1/4 IN> OD
30 IN.
30-1/2 IN.
1 IN.
2 IN. X 3/8 IN. I.D. NOZZLE
NOTE: MATERIAL - 1/4 IN. THICK CELLULOSE ACETATE.
Figure 7. Sketch of first in-line concentrations.
-------
SIGHT PORT
Ln
-FeCl3
-DYLEX/HERCOFLOC SLURRY
SEWAGE
INFLOW
OPERATING CHARACTERISTICS
MAXIMUM FLOW (GPM) 20
REYNOLDS NUMBER 20,000—»-10,400
MAXIMUM VELOCITY (FPM) 54
FLOAT CAKE
EFFLUENT
Figure 8. Sketch of second in-line test module,
-------
3 IN,
6 IN.
INFLOW
FLOAT CAKE
EFFLUENT
OPERATING CHARACTERISTICS
LABORATORY MODULE DEMONSTRATION MODULE
FLOW (GPM)
REYNOLDS NO. (MIN. )
MAJOR DIAMETER
MINOR DIAMETER
LENGTH (FT)
RETENTION (MIN.)
30
15,000
6 IN.
3 IN.
15
0.4
350 (0.5 MGD)
15,000
6 FT
3 FT
20
7
Figure 9. In-line model design.
-------
TABLE 15. RESULTS OF IN-LINE TESTS - FIRST TEST MODULE
Ferric Chloride Suspended Solids
Test No.
1
2
3
4
5
6
7
Dosage (mg/1)
0
25
50
100
100
200
100
Effluent
164
275
179
33
29
23
256
Removal 7»
41
-
36
88
90
92
8
Phosphate Removal
mg/1
14
9
10
5
4
3
6
7=
7
40
33
67
73
80
60
7
7
6
6
6
6
7
PH
.3
.0
.8
.7
.5
.3
.1
Flow = 1 gpm, Re = 800, V = 1.75 fpm, Retention = 17 min.
bFlocculant - Hercofloc 810 at 1 mg/1.
cFlotation aid - Dylex KCD-340 at 100 mg/1.
Raw sewage properties: SS = 552 mg/1, Phosphate = 19 mg/1,
ph = 7.2.
eScreened sewage properties: SS = 279 mg/1, Phosphate =
15 mg/1, PH = 7.3.
Test No. 5 employed 30 minutes batch retention for influ-
ent sample in Test No. 4.
§Test No. 7 employed batch remix of Test 5 float cake with
fresh sewage.
37
-------
TABLE 16. ADDITIONAL RESULTS OF IN-LINE TESTS - FIRST TEST MODULE
Test No. Flow (gpm)
8 10
9 7%
llb 7%
12C 7%
10 5
Reynolds
Number
8,400
6,300
6,300
6,300
4,200
Velocity
(fpm)
17
13
13
13
9
Suspended Solids3
Effluent
(mg/1) Removal (%)
No separa-
tion
89 45
112 31
38 77
86 47
NOTES:
Except as noted,
Inflow suspended solids = 163 mg/1.
Flotation aid = Dylex KCD-340 at 100 mg/1.
Ferric chloride at 100 mg/1.
Hereofloc at 1 mg/1.
Phenolic microballoons as flotation aid at 100 mg/1.
;Flocculant-Hercofloc 810 at 4 mg/1.
38
-------
TABLE 17. RESULTS OF IN-LINE TESTS - SECOND TEST MODULE
Flow (gpm)
10
10
10
10
(Batch)
(Batch)
20
20
Flocculant
Dosage
Og/1)
1
1
2
3
3
3
2
2
Reynolds
Number
10,400 — »
5,200
10,400 — >.
5,200
10,400 — »-
5,200
10,400 — >
5,200
10,400— »
5,200
10,400 — *
5,200
20,800 — »
10,400
20,800 — *
10,400
Maximum
Velocity
(fpm)
27
27
27
27
_
_
54
54
Suspended
Effluent
(mg/1)
337
256
288
205
38
16
235
288
Solids
Removal (%)
20
39
32
51
91
96
36
21
NOTES: a
Inflow suspended solids = 422 mg/1 for 10 gpm tests, 366 mg/1 for
20 gpm tests.
flotation aid - Dylex KCD-340 at 100 mg/1.
cFerric chloride at 100 mg/1.
dFlocculant - Hercofloc 810.
39
-------
decreased as ferric chloride dosage increased. The maximum pH change was
about 1 with 200 mg/1 ferric chloride.
Tests 5 and 7 were included to determine (1) the effect of increased
retention time and (2) float cake reuse potential, respectively. The
increased (batch) retention time in Test No. 5 did not significantly affect
suspended solids removal (92 percent suspended solids removal versus 90
percent at 17 min. retention). Reuse of float cake (Test No. 7) in a batch
experiment resulted in poor suspended solids removal (8 percent).
Visual float cake flow properties and settleable material in laminar
flow (Re~800, V = 1.75 fpm) were of interest. At this flow condition,
float cake did not flow readily and there was a considerable amount of
settleable solids which did not scour out of the unit. (Scouring veloci-
ties in combined sewer lines are generally 2-3 fps) (5).
Tests were then conducted (Table 16) to define the upper flow limit at
which flocculation and flotation could be achieved. Tests at 10 gpm showed
visual flocculation but no significant flotation of the resulting floes.
Subsequent tests at 1\ gpm and at 5 gpm showed that flocculation and
flotation could be achieved at these flow levels. As in the low flow tests,
the degree of suspended solids removal was sensitive to chemical dosage.
In Test No. 12, the solids removal efficiency was increased to 77 percent
by increasing flocculant dosage to 4 mg/1. Although satisfactory floccula-
tion and flotation were achieved, float cake would not flow significantly
even at this flow condition. With a 100 mg/1 flotation aid dosage, the
float cake appears to be relatively nonflowable.
Since the nominal inlet velocity in the flow-turbulence module was
quite high (100 x module velocity), dissipation of this velocity (energy)
in the initial portion of the module promoted turbulence not considered in
the calculated Reynolds number. The adverse effect of high inlet velocity
on flocculation was evident, and module design must consider gross changes
in velocity which could affect flocculation and flotation. Reynolds number,
flow velocity and retention time appear to be important design criteria.
The second in-line module was designed to alleviate some of the short-
comings of module one. Table 17 gives the test results for the second
module. Comparison of these data with the data from the first module shows
a marked improvement (20 to 50 percent versus none) in solids removal at
the 10 gpm flow rate. This improvement is attributed to an increased
residence time in the modified design. Indeed, the best results were
obtained for samples which received batch retention following initial floc-
culation/flotation in the flow system. The excellent suspended solids
removal (greater than 90 percent visual clarity) was obtained by an addi-
tional 30 min, retention to promote separation of solids from the clarified
stream.
Float cake flow difficulties were encountered in these tests as in the
earlier tests, but some float cake flow was obtained. Separation of the
float cake showed a significant concentration of suspended solids in the
40
-------
float cake stream: 80,000 mg/1 vs 422 mg/1 in the influent. Both float
cake and effluent streams varied significantly during the tests, but the
separation was obvious from visual observation through a sight port in the
module.
Discussion and Recommendations
Since concentration of sewage suspended solids was achieved at a
relatively high velocity (about 1 fps), the concept may have general appli-
cation for storm flow treatment in sewer lines where long retention time
can be achieved by simply adding the treatment chemicals to the system a
considerable distance before attempting separation of the concentrated
stream. Such a technique is, of course, more general than the present
concept in that the flotation could be effected by air as well as by low
density solids or separation could be by sedimentation rather than by
flotation. As a result of flow control difficulties in this investigation,
however, further examination of such a concept should be in an actual com-
bined flow sewer system rather than at the laboratory scale.
Dissolved Air Flotation Tests
Bench-scale batch tests were conducted with Milwaukee combined sewer
overflow samples to compare the solid flotation aid system with the
dissolved-air flotation system which has been under development (6). Tables
18 to 21 give the results of these tests. The results confirm that a
combination of ferric chloride and polyionic flocculant provides excellent
flocculation. Chemical dosages tested were 25 mg/1 ferric chloride and
0.5-2.0 mg/1 flocculant to produce suspended solids removals of greater
than 90 percent.
Table 18 gives the results of tests comparing ferric chloride and
alum as coagulants. Suspended solids, total COD and soluble COD were
measured before and after treatment. Suspended solids removals were
similar for both coagulants.
Table 19 gives the results of varying chemical addition methods. In
Test No. 1, without a flocculation period, a reduced suspended solids
removal efficiency is indicated. Table 20 gives additional data on the
effect of varying flocculation time. Although floe rise rates were too
fast to measure, the results show that a short flocculation period is
beneficial.
Table 21 gives data comparing Hereofloc 810 with Purifloc C-31, the
polyionic flocculant which has been utilized previously in dissolved-air
flotation field tests. The dosages were selected for comparison on an
equal cost basis. Similar suspended solids removal efficiencies were
obtained.
In addition to the dissolved-air flotation experiments, a bench test
was conducted with the Dylex flotation aid. The results showed a reduction
in combined sewer overflow suspended solids from 337 mg/1 to 2 mg/1 with
a Dylex dosage of 75 mg/1 (Hercofloc dosage at 1 mg/1 and ferric chloride
41
-------
at 25 mg/1). The flotation aid was difficult to disperse in the test
beaker, but suspended solids removal was excellent. Since the solid flota-
tion aid is difficult to disperse in batch tests, flow tests are probably
required for meaningful comparisons.
TABLE 18. DISSOLVED AIR FLOTATION - BENCH CHEMICAL TREATMENT TESTS
Test No.
Chemical and Dosage 25 mg/1 FeCl3 25 mg/1 A12(S04)3 • 18 H20
2 mg/1 Purifloc C31 2 mg/1 Purifloc C31
Floe Time (min)
Recycle Rate (%)
Retention Time (min)
Scum Volume (gal/1000 gal)
Sludge Volume (gal/ 1000 gal)
Effluent:
Suspended Solids (mg/1)
Total COD (mg/1)
Soluble COD (mg/1)
Raw Waste:
Suspended Solids (mg/1)
Total COD (mg/1)
Soluble COD (mg/1)
1.5
15
5
15
Trace
6
37
36
228
211
73
1.5
15
5
15
Trace
10
50
42
203
193
59
SS Removal (%) 97 95
42
-------
TABLE 19. DISSOLVED AIR FLOTATION -
CHEMICAL ADDITION TECHNIQUE EFFECTS
Test No.
Fed Dose (mg/1)
Floe Time (min)
Purifloc C31 Dose (mg/1)
Floe Time (min)
Recycle Rate (7o)
Recycle Pressure (psig)
Detention Time (min)
Scum Volume (gal/1000 gal)
Sludge Volume (gal/1000 gal)
Effluent:
Suspended Solids (mg/1)
Total COD (mg/1)
Soluble COD (mg/1)
Suspended Solids Removal (7°)
1
25
0
2b
0
15
40
5
15
Trace
54
138
70
87
2
25
Oa
2C
0
15
50
5
15
Trace
21
93
68
95
3
25
0
2b
0
15
50
5
15
Trace
33
119
63
92
4
25
4
2b
0
15
50
5
12
Trace
59
112
70
89
5
25
3
2c
1
15
50
5
20
Trace
15
62
66
96
NOTES: a^Q £locculation period; however, there was a 20-sec time lapse
between chemical additions.
^Injected into pressurized flow while adding to cylinder.
cAdded by pipette to cylinder.
^Tests conducted on raw waste. Recycle was raw waste after
screening through a 50 mesh screen. SS = 431 mg/1
Total COD = 512 mg/1, Soluble COD = 102 mg/1
43
-------
TABLE 20. DISSOLVED AIR FLOTATION - FLOCCULATION TIME EFFECT
Test No.
FeCl3 Dose (mg/1)
Floe Time
Purifloc C31 Dose (mg/1)
Floe Time (min)
Recycle Rate (%)
Rise Rate (fpm)
Detention Time (min)
Scum Volume (gal/1000 gal)
Sludge Volume (gal/1000 gal)
Effluent:
SS (mg/1)
Total COD (mg/1)
Soluble COD (mg/1)
Suspended Solids Removal (7o)
1
25
2
2
0
15
5
15
None
13
24
20
91
2
25
0
2
2
15
Too Fast
5
15
Trace
12
24
18
92
3
25
1
2
1
15
5
15
Trace
8
23
20
95
w
aRaw waste sample stored 3 days in refrigerator
SS = 148 mg/1. Total COD = 168 mg/1, Soluble COD = mg/1.
Pressurized flow - raw waste that had been screened through
50 mesh screen. Pressurized at 50 psig.
44
-------
TABLE 21. DISSOLVED AIR FLOTATION -
FLOCCULANT TYPE EFFECT
Test No.
Chemical Treatment
FeCl3 (mg/1)
Hereofloc 810 (mg/1)
Purifloc C31 (mg/1)
Floe Time (min)
Flotation Data
25
0.5
0
1
25
0
2
Recycle Rate
Rise Rate (fpm)
Detention Time (min)
Scum Volume (mg/1)
Sludge Volume (mg/1)
Effluent Quality
PH
Suspended Solids (mg/1)
Total COD (mg/1)
Soluble COD (mg/1)
Suspended Solids Removal
Raw Waste Properties
PH
Suspended Solids (mg/1)
Total COD (mg/1)
Soluble COD (mg/1)
15
10
Trace
6.6
24
40
36
15
15
Trace
6.6
14
34
26
94
7.1
225
169
36
45
-------
SECTION 5
FIELD TESTS
DESIGN AND MODIFICATION OF THE DEMONSTRATION FACILITY
Original Demonstration Facility
The original treatment facility available at the Hawley Road site con-
sisted primarily of a screening chamber and a dissolved air flotation basin.
Figure 10 shows a schematic flow sheet for the treatment system. A photo-
graph of the overall system is shown in .Figure 11.
The screen is an open-end drum into which the raw waste flows after pass-
ing through a % in. bar rack. The water passes through the screen media into
a chamber directly below the drum. The drum rotates and carries the screened
solids to the spray water cleaning system where they are flushed from the
screen with previously screened water. A 50 mesh screen (297 micrometer
openings) is provided on the drum screen. This provides approximately 20-30
percent removal of pollutants and allows high throughput rates (50 gpm/sq ft)
at reasonable head losses (up to 12 in. water). The drum screen as installed
is an eight-sided drum with a cross-sectional area equivalent to a 7.5 ft di-
ameter circle. Drum length is 6 feet. Total screen area is 144 sq ft. The
wetted screen area ranges between 72 and 90 sq ft depending upon the head
loss across the screen.
The flotation basin is a rectangular chamber with a surface skimming
system to remove floated scum. Screened water is pressurized and saturated
with air under pressure in an air solution tank. When the pressure is re-
duced across a weir-type diaphragm valve, minute air bubbles (less than 100
micrometers) are formed. This air-charged stream is then blended with the
remaining screened water flow. The bubbles become attached to the garticu-
late matter in the mixing zone (detention time: approximately 60 sec) shown
in Figure 10 and rise to the surface for subsequent removal by skimmer
flights. Chemical flocculants may be added to enhance the removal effi-
ciency of finely divided particulates.
The design of the system was such that a wide range of selected vari-
ables could be evaluated. The following range of variables was possible with
the demonstration system:
46
-------
AIR TANK
SOLUTION
COMPRESSED
AIR
CHEMICAL FLOCCULANT ADDITIONS
MIXING ZONE
SCREEN BACKWASH
SIDEWATER DEPTH 8.5 FT
FLOTATION ZONE
61 FT
18 FT
EFFLUENT
FEED PUMP
Figure 10. Original Hawley Road treatment facility.
-------
Figure 11. Photographic view of Hawley Road site.
-------
Flow rate 1500 - 4400 gpm
Surface loading 2-10 gpm/sq ft
Horizontal velocity 1.30 - 3.75 ft/min
Pressurized flow rate 300 - 1100 gpm
Operating pressure 40 - 70 psig
Detention time 7-44 min
As seen in Figure 11, the system is located under an existing highway
bridge which provides overhead protection. A dam is provided in the com-
bined sewer to allow impoundment of a limited quantity of overflow for subse-
quent treatment. Flow metering equipment is provided to measure the influ-
ent flow rate, the volume of screen backwash water and the volume of floated
scum collected. The raw and screened backwash flows are measured via venturi
meters connected to differential pressure gauges which both record and to-
talize the flows. The floated scum is measured with an open-channel, float-
type meter which records and totalizes the flow of floated scum. The elec-
trical control panel provides all necessary controls for 100 percent auto-
matic operation with manual overrides on all systems. A Merchants Police
alarm is connected to the system so that personnel will be alerted when the
system goes into operation. The system is always (24 hours per day, 7 days
per week) ready to operate to allow monitoring of maximum number of overflows.
Modification of j:he Original Demonstration Facility
Essentially, the modification of the original demonstration system con-
sisted of the following:
(1) Bypassing the existing dissolved air pressure tank.
(2) Installation of a mechanical flocculation capability to
the flotation basin.
(3) Addition, mixing, storage and feeding capabilities of the
desired amounts of flotation-aid and flocculant chemicals.
The following design criteria were selected for the design of the
demonstration system modifications:
(1) Chemical dosages 100 mg/1 FeCl3
5 mg/1 cationic polyelectrolyte
100 mg/1 flotation-aid 5% slurry
(2) Retention time prior 60 sec
to addition of beads
(3) Flocculation time 5 min
(4) Rise rate 2 to 5 ft/min
Figure 12 shows a flow schematic of the modified treatment facility
utilized for the evaluation of the flocculation/flotation aid process.
49
-------
BEAD STORAGE
POLYMER
STORAGE
FeCL3
STORAGE
L71
o
AIR SOLU-
TION
TANK
MOYNO PUMP
BEAD SLURRY
9 HERCOFLOC 810
CHEMICAL
FEED PUMP
-EX—n
SPLIT
FLOW PUM
FeCl
1'FLOCCULATION ZONE
FLOTATION ZONE
o+o
TURBINE
FLOCCU-
LATOR
D-CD
HIGH OVERFLOW
RATE
FLOATED SCUM
SCREEN BACKWASH
FEED PUMP
COMBINED SEWER
EFFLUENT
TO
SANITARY
SEWER
Figure 12. Flow schematic of modified demonstration facility
-------
Bypassing of the existing dissolved air pressure tank was accomplished by the
addition of four butterfly valves. Stock solutions of the chemical floccu-
lants were stored in 400-gallon capacity fiberglass tanks, and the chemicals
were injected at the desired locations with a calibrated Triplex chemical
feed pump.
Design of the Turbine-Flocculator
It was decided to utilize a turbine-flocculator for the promotion of the
coflocculation of the flotation aid and the particulate matter in the com-
bined sewer overflows. The input horsepower requirement and the maintenance
of a suitable tip velocity of the turbine blades are the most important de-
sign creiteria of a turbine flocculator. A tip velocity of 5 ft/sec was se-
lected, and a power requirement of 3 hp was calculated for the present de-
sign. It was decided to utilize two turbines with six blades each. A sketch
of the turbine flocculator is shown in Figure 13. A mechanical variable-
speed drive with a flexible shaft coupling was provided on the flocculator
which permitted variation in the flocculator speeds in the range of 3 to 27
rpm.
Flotation Aid Expansion and Feeding Procedures
It was decided to expand the flotation aids on a batch basis since an
on-stream expansion of these aids was beyond the scope of the present test
program. It was also decided to feed the flotation aids in the form of a
uniform slurry with Milwaukee tap water in a portion of the screened raw
waste. This portion of the raw waste and flotation aid slurry were then
mixed with the main waste stream and coflocculated with a flocculant.
Expansion of Flotation Aid
Initially, twenty 55-gallons of expanded polystyrene beads (Dylex KCD-
340) prepared in an agitated boiling-water batch expander were shipped from
Koppers Co. to Rexnord, Inc. After the exhaustion of this shipment, all fu-
ture beads were expanded in batch systems within the R.exnord premises. The
final bead expansion system utilized, as recommended by Koppers, is shown in
Figure 14. Initially, the system consisted of putting approximately % gal of
wet bead cake in approximately 30 gal of boiling water in a 55 gal insulated
tank. Then steam was injected through a sparger at the bottom of the tank,
and a heating time of 30 to 60 sec was allowed for the expansion of beads.
The steam was produced at a pressure of 40 to 70 psi and was fed through a
3/8 in. to 1/2 in. pipe. The expanded beads were taken out of the 55 gal
tank and stored separately. There were several changes made in this system
as the project progressed. These changes and details of expansion technique
will be discussed in later sections of this report.
Mixing of the Flotation Aids
An integral part of the process was to whip the beads and water into a
uniformly mixed slurry so that the beads could be metered at a known concen-
tration into the flocculation/flotation chamber. Initially, only one
51
-------
VARIABLE
SPEED DRIVE
MOTOR
10.5 IN.
14.23 IN.
TURBINE BLADES
TURBINE
DISC
SHAFT
'42 IN.
HUB
20 FT
SIDE VIEW
TURBINE
18 FT
Figure 13. Turbine flocculator,
-------
STEAM SPARGER
INSULATED 55 GAL TANK
Figure 14. Bead expansion system.
53
-------
600 gal capacity fiberglass tank was provided for mixing a 5 percent bead
slurry with the expanded hollow beads supplied by Koppers Co. However, it
was immediately apparent that the 0.5 hp propeller mixer provided for the
mixing of the bead slurry was not adequate to keep the beads in suspension.
Therefore, the existing mixer was replaced with a 1.5 hp agitator mixer*
having two 12-in.-diameter propellers; Although this mixer was almost ade-
quate, it was apparent that it still would not keep a 5 percent bead slurry
completely in solution.
To overcome the bead mixing problem, consultations were held with the
MIXCO Corp. Samples of beads and Hercofloc 810 were shipped to MIXCO for
evaluation of the mixing requirements for the bead slurry. The recommended
mixer consisted of three turbine blade propellers and was driven by a 15 hp
motor. Because almost adequate suspension could be achieved with only a
1.5 hp mixer, it was felt that the MIXCO recommendation was extremely con-
servative and expensive.
Additional laboratory experiments were conducted by Rexnord to provide
design information leading to a solution of *che bead mixing problem. It was
determined that a combination of mixing and recirculation pumping would pro-
vide sufficient energy to keep a 3 to 6 percent bead slurry in suspension.
This was considered sufficient for the present test program and the proposed
combination of mixing and recirculation was recommended for field use.
A sketch of the mixing system utilized is shown in Figure 15. The sys-
tem consisted of two interconnected 600-gal fiberglass tanks, one centri-
fugal recirculation pump, two 1.5 hp low speed propeller mixers** and two
specially designed twin bronze eductors for promoting a vortex in each tank.
The twin eductors were submerged beneath the bead-water interface, and were
fed by pumping from the bottom of each tank using a 375 gal centrifugal pump.
In the process, water and eventually beads were pumped out of the tank bottom
and recirculated through the eductor nozzle where the very high velocity exit
stream passed through a collar and violent mixing occurred. The continuous
mixing by the low speed mixers and the eductors produced a homogeneous slurry
which could be maintained homogeneous as long as the mixers and the eductor
system were on. The mixed slurry was then pumped into a portion of the raw
waste via a Moyno pump equipped with a variable speed drive motor to adjust
the required concentration of the flotation aids. The flocculant (Herco-
floc 810) was added into the bead slurry before it mixed with the screened
split flow stream. The whole setup was automated and the bead mixing*system
started automatically along with the priming pump when an overflow signal was
actuated in the combined sewer.
*1.5 hp open tank, top entering agitator with steel shaft, LIGHTNIN Mixers
and Aerators, Milwaukee, Wisconsin.
**LIGHTNIN Mixers and Aerators, Milwaukee, Wisconsin.
54
-------
LIGHTNIN1
MIXERS
RECIRCULATION
PUMP
TO METERING
**" PUMP
TO METERING
PUMP
RECIRCULATION PUMP
Figure 15. Schematic of bead mixing system.
-------
OPERATION METHODS AND TEST PLAN
Operational Procedures
The demonstration system was put into operation automatically when a
float switch in the sewer sensed an overflow. An alarm signal was immedi-
ately transmitted to the Merchants Police;, the raw feed pump began to prime;
and the drum screen, the flocculator and the split flow pump were put into
operation.
Along with the priming pump, the bead recycle and mixing system was
also started to provide a uniform bead slurry. All runs except No. 71-2
were started with the tank approximately 80 percent full of water. The water
in the tank was mostly clean ground water which had been added after drain-
ing and flushing out the water from the previous run. The raw feed-pump gen-
erally primed in about 12-15 min. When primed, the raw pump was activated
and the flow meters, bead slurry pump, chemical feeder, skimmers and all
other auxiliary equipment were put into operation. At the end of the over-
flow, the system shut down automatically. All variables were then selected
for the next run and the controls positioned. Variables associated with
system operation included: raw flow rate, split flow rate and flotation aid
and flocculant dosages.
jampling Procedures
Sampling began when the raw pump primed. Raw waste and screened water
sample collection was started immediately. Effluent sample collection was
delayed from 20 to 45 min (depending upon raw flow rate) to allow purging
of water in the tank. This procedure insured collection of representative
effluent samples. Screen backwash and floated scum samples were taken man-
ually during screen backwash and scum removal periods. All other samples
were collected via an automatic sampling system. This system consisted of
two timers connected through the proper valving to automatically composite
the various samples. The first timer controlled the sample-taking frequency
(0-30 min). The second timer controlled duration of the sampling time (0-60
sec). The sampling valves are air-operated weir valves. Samples were com-
posited every 5 min with the automatic system, and approximately 2-5 gal of
sample was collected per hour. The samples were refrigerated immediately
after the run. Analyses were then started within 8 hr.
Test Plan
Process variables associated with the evaluation of the flocculation/
flotation aid concept include: hydraulic overflow rate (gpm/sq ft of tank
area), split flow rate, flotation aid type and dosage, and flocculant chemi-
cals type and dosage. Based on the promising results shown in the bench
scale tests, it was proposed initially to evaluate 20 overflow events (7).
A detailed field test plan based on a 2 x 2 factorial design is outlined in
Table 22. However, the total field evaluation was limited to only five
overflow events after a series of start-up difficulties depleted funds.
56
-------
TABLE 22. FIELD TEST PLAN
Process Variable Level No. of Storms
1) Flotation-Aid Type Dylex, Glass Microballoons 4-6
2) Flotation-Aid Level 10-100 rag/1 4-6
3) Overflow Rate 3-9 gpm/ft2 20
4) Screening With, Without 2
5) Air Flotation Comparison 3
6) Flocculation Time Optimize
7) Flocculation Speed Optimize
8) Ferric Chloride Dosage Optimize
9) Flocculant Dosage Optimize
Analytical Test Summary
Samples
Analysis
pH
Total Solids
Total Solids, Volatile
Suspended Solids
Suspended Solids, Volatile
TOG
COD
Nitrogen
Phosphates
Chlorine Demand
Total Coliform
Raw
1
1
1
1
1
1
1
1
1
1
1
Wash Screened
1 1
1 1
1 1
1
1
1
1
Scum Effluent
1 2
1 2
1 2
2
2
2
2
2
2
2
2
Total/*
Storm
6
6
6
4
4
4
4
3
O
_>
3
3
*Total/storm doubles for storms tested to identify treatability of first
flush and subsequent flow.
Flotation Aid Recovery Tests
No. of Storms
Hydrocyclone Experiments 2
High-Shear Mixing Experiments 2
Float Cake Properties 4
57
-------
The test results at the end of four storms had pointed out two impor-
tant problem areas which affected the technical feasibility evaluation of the
proposed concept. These problem areas and changes in program plan will -be
discussed In detail in later sections of this report. However, the general
range of variables investigated during this field test program was as fol-
lows:
Overflow rates 1.9 - 5.6 gpm/sq ft
Split flow rates 600-1100 gpm
Flotation aid type Dylex beads (polystyrene) KCD-340)
Flotation aid dosage 81-.113 mg/1
Flocculant chemicals:
Ferric chloride 30 - 34 mg/1
Hercofloc 810 (cationic) 2.6 -6.0 gm/1
58
-------
RESULTS AND DISCUSSION
Raw Wastewater Characterization and Screening Efficiency
A total of five overflow events were monitored during the field test
phase. The quality of the raw wastes for these overflows is shown in
Table 23. There was a wide variation in the wastewater quality. The
suspended solids varied between 179 and 916 mg/1. Four of these overflows
(71-1, 71-2, 71-3, 71-5) were float treated with polystyrene beads (Dylex
KCD-340) and one overflow (71-4) was float treated with dissolved air. All
the five overflows were passed through a 50 mesh (297 micron openings)
screen. The screened water quality for these storms is shown in Table 24.
Taken as a group, data indicates erratic and poor results. However, bead
flotation tests 1, 2 and 3 were severely hampered by start-up difficulties
and in no way represent the technical potential of the concept. As a
result of these difficulties, many data showed negative removals when
compared with the corresponding raw water quality. The reasons for the
poor screened water quality may be attributed to possible leakage of the
influent raw wastewater through the worn drum seal. However, a lack of
efficient screening removals did not influence the evaluation of the
flocculation/flotation aid concept, and therefore it will not be considered
any further.
Operation of the Flotation System
The bead flotation effluent quality data and overall pollutant removals
for Storm Nos. 1, 2 and 3 are shown in Table 25. Results of these tests
are reported as a guide to the start-up difficulties encountered and to
remedial actions taken, not as being indicative of the technical potential
of the concept.
Pollutant removal efficiencies were better for low overflow rates
compared with high overflow rates . The removal efficiencies as shown in
Table 25 varied over a wide range. Significantly better pollutant removals
were obtained for Run 71-1 compared with Runs 71-2 and 71-3. The percent
pollutant removals ranged between 65 and 75 percent for Storm No. 71-1 and
between 0 and 50 percent for the other two tests. The high removal effi-
ciencies for Run 71-1 would appear to indicate fairly good operation.
However, much of the removal accomplished during this test was attributable
to settling and not flotation. Upon draining the tank, it was found that
approximately 8 in. of settled sludge had resulted from this run. Thus the
first startup problem, flotation aids which sink, became apparent. The
influent pollutional load for this test was significantly high (Table 23),
and this might have facilitated the sedimentation of the particulate matter.
Visual observations in the field during the test showed many beads in the
flotation effluent which obviously were not effecting any removal of solids.
In Run 71-2, the effluent concentrations of most parameters were found
to be higher than the influent concentrations. The reason for these apparent
negative removal efficiencies was probably the resuspension of the particu-
late matter that had settled in the tank during Run 71-1, which was not
59
-------
TABLE 23. RAW COMBINED SEWER OVERFLOW QUALITY
Overflow
No.
71-1
71-2
71-3
71-4
71-5
Average
Range
pH Total
Units Solids
6.9 1167
7.3 365
7.2 753
7.4
6.4 351
7.0 659
6.4- 351-
7-4 1167
Total
Volatile
Solids
440
242
178
-
419
252
149-
440
Suspended
Solids
916
187
625
179
289
439
179-
916
Volatile
Suspended
Solids
360
50
200
57
101
154
50-
360
COD
646
120
65
69
123
205
65-
646
Ortho
Soluble Phosphate Kjeldahl
BOD TOG TOG as P Nitrogen
290 70 1.26 19.0
111 18 0.45 3.9
41 22 16 0.03 5.0
25 21 16 -
25 37 7 1.46* 3.1
30 96 25 0.58 7.8
25- 21- 7-70 0.03-1.26 3.1-19.0
41 290
NOTE: All results in mg/1 except where noted otherwise.
*Measured as total phosphates.
-------
TABLE 24. SCREENED COMBINED SEWER OVERFLOW QUALITY
Overflow
No. pH Units
71-1 6.9
71-2 6.7
71-3 6.8
71-4*
71-5 6.4
Suspended
Solids
1002
323
578
215
Volatile
Suspended
Solids COD BOD TOG
216 368 - 160
151 131 - 50
86 140 35 34
80 96 19 35
Soluble
TOG
32
15
-
6
NOTE: All results in mg/1 except where noted otherwise.
*Sample lost during handling.
-------
TABLE 25. BEAD FLOTATION EFFLUENT QUALITY
Vo1ume
Overflow Suspended Solids Suspended Solids COD BOD
Overflow Rate Quality Removal Quality Removal Quality Removal Quality Removal
No. gpm/sq ft mg/1 % mg/1 % mg/1 % mg/1 %
Low Overflow Rates
71-1
71-2
71-3
71-1
71-2
71-3
2.25
2.7
2.8
4.5
5.4
5.6
329
251
-
414
261
549
64.1 97
1.48
- —
High Overflow
54,8 148
165
12.2 154
73.1
-
-
Rates
58.9
-
23.0
165
110
61
328
143
177
74.5
8.3
6.2
49.2
_
-
_ _
_
31 24.4
_ _
.
38 7.3
-------
TABLE 25 (continued)
Overflow
No.
71-1
71-2
71-3
71-1
71-2
71-3
Overflow
Rate
gpra/sq ft
2.25
2.7
2,8
4.5
5.4
5.6
TOC Soluble TOG
Quality
rag /I
68
56
25
116
69
50
Removal Quality
% mg/1
Low Overflow
76.6 36
49.1 18
16
High Overflow
60.0 46
37.5 19
17
Removal
Cri
Rates
48.6
0
0
Rates
34.3
-
_
Ortho Phosphates
Quality Removal
mg/1 %
0.05
0.10
0.13
0.08
0.10
0.04
96.0
77.8
•*
93.7
77.8
-
Kiehdahl
Quality
mg/1
8.2
5.9
9.1
13.2
5.6
9.6
Nitrogen
Removal
°j
56.8
-
-
30.5
_
-
-------
removed. In an effort to improve upon the efficiency of the operation,
several changes were made in the operating conditions. The actions taken
to improve the operating results were as follows:
1. Polyelectrolyte (Hereofloc 810) was added to the bead slurry
in an attempt to effectively coat the beads and thereby
provide more effective flotation.
2. The amount of split flow volume was increased to the
achievable maximum of 1100 gpm.
3. The flocculator speed was increased to maximum to improve
upon the coflocculation of beads and solids.
However, these corrective actions did not affect the bead flotation removal
efficiencies as shown by the effluent quality for Run 71-3 (Table 25). The
removal efficiencies were found to be poor and ranged between 0 and 25 per-
cent for various parameters. The effluent quality again varied over a wide
range, and in many cases the data was erratic because of the presence of
beads in the effluent samples. The apparent increases in effluent COD, TOG
and nitrogen values over the influent concentrations were attributed to the
presence of beads. On the other hand, the BOD analysis would not be affected
by the presence of beads and did not show an increase in values as did COD,
TOG and nitrogen.
The overflow 71-3 had actually been divided in two operational periods.
The initial 60 min of operation was conducted with beads and was designated
as Run 71-3. The remaining 40 min of operation on the tail end of the same
overflow was conducted with dissolved air and designated as Run 71-4. A
tabulation of the effluent qualities and the removal efficiencies for that
test is shown in Table 26. The results of the air flotation run indicated
significantly better pollutant removal efficiencies than the bead flotation
results. The range of pollutant removals was 35 to 75 percent for air
flotation compared with 0 to 25 percent for bead flotation. It is empha-
sized that a direct comparison between the two techniques on the basis of
these two tests alone is unfair since the air flotation technique is well
developed while the bead flotation technique was barely started and was
suffering the usual startup difficulties associated with any large pilot
plant.
Reevaluation of the Field Test, Program Plan
Based on the field data obtained from the above-described four experi-
ments, a meeting was held between USEPA, Hercules and Rexnord to review
the field test progress and future program plan of the project0 Two major
technical problems were indicated to have limited the performance of the
flocculation/flotation concept effectiveness. These problems were:
1. Settling of some of the flotation aids instead of flotation.
2. Ineffective coflocculation of the coagulated particulate
matter onto the flotation aid.
64
-------
TABLE 26, RESULTS OF RUN NO. 71-4
Overflow Rates ,a gpm/sq ft
Analysis
Suspended Solids
Vol. Suspended
Solids
COD
BOD
TOG
Soluble TOG
Raw
Waste Water
Influent
Quality
179
-
69
25
21
16
1
Effluent
Quality
mg/1
42
18
46
12
14
13
.9
Removal
76.5
68.4
33.3
52.0
33.3
18.8
2
Effluent
Quality
mg/1
88
40
38
12
13
13
.8
Remova 1
50.3
29.8
44.9
52.0
38.1
18.8
NOTE:
Estimated overflow rates.
65
-------
The presence of the polystyrene beads had been noticed both in the field
test effluents as well as at the bottom of the bead slurry storage tanks.
When viewed under a microscope,, it was found that the size of most of the
expanded beads was in the range of 200 to 300 micrometers. The size of
the unexpanded beads ranged between 75 and 100. Several partially expanded
beads were observed in the grab samples of the expanded beads under a
microscope. These observations indicated that the reason for the sinking
of some of the beads might be non-expansion or insufficient expansion due
to the short heating time (20 to 60 sec) employed during the expansion
technique.
It was also indicated that the principal reason for the inadequate
coflocculation of the beads and suspended matter was the deterioration of
the flocculant (Hercofloc 810) strength due to its long storage time during
dry weather periods. Spot viscosity checks of the field-prepared Hercofloc
solution showed a value of 15-20 cp for a 0.5 percent solution in Milwaukee
tap water compared with 290 cp for a laboratory solution in distilled water
at corresponding strength. The shelf life of a 0.5 percent Hercofloc
solution was indicated to be of the order of 2 days only.
The use of a different liquid polyelectrolyte would have greatly
simplified matters, but a series of laboratory tests with Milwaukee sewage
and such liquid flocculants as Dow D-31 and Nalco 607 as well as Hercofloc
810 (all cationic polyelectrolytes) showed that all would build a good floe
but only the Hercofloc would promote the coflocculation of this floe to the
Dylex flotation aid.
A preliminary comparison of the operating costs for the bead and
dissolved air flotation concepts showed that the cost of treatment with beads
was high even when a bead recovery factor of 90 percent was utilized. The
chemical costs alone were calculated to be 6.7
-------
Solution of the Major Technical Problems
Bead Settling Problem
Since it has been indicated that a shorter time (approximately 30 to
60 sec) was insufficient for the proper expansion of beads, the heating
time was increased to 2 min. A quality control check was also instituted
during the expansion of beads to insure proper expansion. Grab samples
of expanded beads were viewed under the microscope and were added to water
to observe their floatability. No sinking characteristics were shown in
the expanded beads immediately after expansion. However, after the beads
were mixed at the site and stored during dry weather, the beads started to
exhibit some settling characteristics. When viewed under a microscope,
these settled beads were found to be properly expanded (200 to 300 micro-
meter). A sample of such beads was also put in boiling water for 10 min
in an attempt to refloat them, but the results were negative.
The problem of sinker beads was then conveyed to Koppers Co., the
manufacturer of the beads, for further evaluation. Samples of unexpanded,
expanded but settling, and expanded and floatable beads were shipped to
the manufacturer. At the same time, bench scale tests were undertaken by
Hercules to find the cause of the problem by reproducing the bead settling
characteristics in the laboratory.
It was thought that the method of expansion of the beads might be
responsible for the sinkers. The improper expansion of beads could impart
excessive heat to the bead cells, causing the cell walls to blow out and
leaving the beads with an open network of capillaries which would gradually
fill with water and sink. Based on suggestions by Koppers Co., it was
decided that the following modifications in the bead expansion technique
might be desirable.
1. An increase in the heating time for bead expansion from
2 min to 5 min.
2. Installation of a mixer on the bead expansion tank to
provide uniform heating.
3. Use of a covered expansion tank to avoid a sudden
exposure of the heated beads to the ambient atmosphere.
Bench scale tests conducted both at Koppers and Hercules did not pro-
duce any sinkers in the laboratory by soaking the expanded beads in water
for prolonged durations up to 3-4 weeks. Any effects of long storage on
the expandability potential of the Dylex beads were discounted by Koppers
after examining the density of the wet cake. Short time (approximately
30 to 60 sec) mixing of the beads with water in a Waring blender did not
produce any sinkers in the laboratory. However, when a 5 min blending
period was provided to the bead water mixture, some sinker beads were
noticed. After two 5-min blending periods, more sinkers were observed.
After three 5-min blending periods and an overnight settling period, 60-
65 percent of the beads were on the bottom of the jar as sinkers. Micro-
scopic examination showed that the blender-treated beads had been physically
67
-------
sheared or chopped. Although the samples of sinker beads from Hawley Road
site did not exhibit strong shearing under microscopic examination, it was
felt that the essential source of sinker beads in field was due to their
exposure to the high shear fields in the eductor system which could have
caused bead damage. Therefore, the most optimum solution under the present
circumstances would be to expand a fresh batch of Dylex beads under best
possible conditions and use the bead mixing system as sparingly as possible
prior to the actual field operation.
Preparation of Hercofloc 810 Solution
Since the reason for the deterioration of the Hercofloc solution was
concluded to be its aging, it was decided to check the viscosity of a newly
prepared batch at various time intervals and feed the polymer at a propor-
tionately higher rate according to its strength. (Note: This polyer is
shipped in a dry powdered form and is ordinarily mixed and immediately used
via chemical metering systems.) A 0.5 percent solution of the Hercofloc
810 was prepared in Milwaukee tap water available at the site. It was
found that a large number of fisheyes, globules of undissolved flocculant,
remained in the 400 gal tank. To get the fisheyes in solution, the contents
were mixed over night. A viscosity check next morning revealed a very low
viscosity of the solution compared with the standard viscosity vs solution
strength curves available from Hercules Incorporated. The viscosity of
the field prepared solution was only 33 cp compared with the 290 cp for a
standard 0.5 percent solution in distilled water. Solutions made in the
laboratory with distilled water confirmed the correctness of the standard
curves.
Assistance was sought from Hercules sales-services field personnel
associated with the application of Hercofloc solutions. It was indicated
that the water flow rate and pressures available at the site were not
sufficient to provide the required vacuum needed through the eductor which
was utilized for mixing the polymer. Alternate solutions such as the use
of a smaller eductor, higher flow rate and substitution of another polymer
were considered.
Smaller ejectors* were obtained and tried in the laboratory. None of
the above models produced any encouraging results.
Alternatives open for providing higher flow rates at the site were:
either to tap a 1 in. to \\ in. line into the municipal main or to*tempor-
arily rent the fire hydrant across the road. Both of these solutions
involved extra expense and neither of them provided a sure solution to the
problem. Therefore, a centrifugal pump was borrowed from Rexnord stock
supplies. Using a throttling valve and the pump, sufficient water pressure
was generated and polymer was mixed in solution. However, several smaller
fisheyes still remained in solution and an additional one-hour mixing with
a lightning mixer was needed in the 400 gal tank. The viscosity of this
solution was found to be 150 cp which can be considered to be quite close
to the desired value because of the lowering effect on viscosity due to the
increased pH of the tap water (7.4) as compared with distilled water (5.5).
*Models 62A and 63A, Pen-Berthy Mfg. Co.
68
-------
It was found that the viscosity of the Hercofloc solution was very
sensitive to increases in mixing time as shown below:
Mixing time, hrs Viscosity of 0.5% Hercofloc Solution - cp
1 150
4 120
24 30
A new version of Hercofloc 810, termed 810.3, having the same polymeric
properties as 810 but a different powder texture, was easier to handle.
Therefore, in order to document the effect of pH and Hercofloc type and their
respective deterioration with time, solutions of 810 and 810.3 were made in
tap water at a pH of 5.5 and 7.5. The pH of the tap water (normally 7.4)
was lowered to 5.5 with sulfuric acid for the appropriate solutions. Another
solution of Hercofloc 810 was made in distilled water for comparison. A
minimum of mixing time ranging from 25 to 60 min was allowed in the labora-
tory, and the initial viscosities of the various solutions were recorded.
Table 27 shows the viscosity data at various time intervals for the various
solutions.
It can be seen from this table that the viscosity decreased with time
for all solutions. Although the solution was most stable in distilled water,
it would have been impractical to make a 400 gal batch in the field. The
Hercofloc 810.3 did not show any significant ease of mixing and handling and
its viscosity decreased more rapidly than that of 810 at a solution pH of
5.5. Therefore, it was decided that the best solution to the problem was
to make a Hercofloc 810 solution at a pH of 5.5, which could be stored up to
7 to 10 days and polymer pumpage be controlled according to its strength at
the time of use. (For actual plant use, of course, an automated system to
mix the polymer on call would be installed.)
Operation & Evaluation of the Final Bead_Flotation Test
A fully satisfactory solution of the bead settling problem would have
been to install an on-stream continuous bead expansion facility and feed
them dry to the raw waste stream. Since such an installation was obviously
beyond the scope of the present project, it was mutually decided between
Hercules, USEPA and Rexnord that at least one additional field test be
conducted with most optimum operation conditions to demonstrate the technical
feasibility of the flocculation-flotation concept. The suitable conditions
for bead expansion were a heating time of 5 min, and the use of a mixer and
insulated lid over the 55 gal. tank as suggested by Koppers Co. Sufficient
beads were expanded under these conditions for one test realizing that some
bead damage would still result from the slurrying system. A fresh batch
of 400 gal of Hercofloc 810 at a pH of 5.5 was prepared in field with
Milwaukee tap water under conditions recommended previously, and its vis-
cosity was monitored closely with time. The flow schematic utilized during
this test is shown in Figure 16. Nearly all of the test flow was routed
through the split flow system and through a relatively long pipe run after
the injection of the flocculant and flotation aid. This was done to provide
the maximum retention time possible in the test unit for the turbulent
69
-------
TABLE 27. VISCOSITY DATA FOR VARIOUS
HERCOFLOC SOLUTIONS WITH TIME
Hercofloc Solution Viscosities -
Hercofloc 810
Storage Time
Hours
0
24
53
148
200
in Tap
pH 7.4
111
18
16
18
18
Water
pH 5.5
132
123
112
93
93
Hercofloc 810 in
Distilled Water
pH 5.4
295
280
268
240
240
Centipolsea
Hercofloc 810.3
in Tap
pH 7.4
122
23
14
15
15
Water
pH 5.5
160
127
91
38
38
aAll viscosities taken with a Brookfield Viscometer at 60 rpm and at room
temperature.
70
-------
AIR-SOLUTION
TANK (BY-PASSED)
SPLIT
FLOW
PUMP
MOYNO
PUMP
BEAD SLURRY
RAW FLOW-
~HERCOFLOC 810
SPLIT FLOW
FLOCCULATION ZONE '
DRUM SCREEN
FeCl,
TURBINE
FLOCCU-
LATOR
FLOTATION ZONE
LOW OVERFLOW RATE
HIGH OVERFLOW
RATE
Figure 16. Process schematic for head flotation test no. 5.
-------
mixing of coagulated sewage, flocculant and flotation aid, it having been
noted in the Phase I laboratory tests that such extended agitation greatly
promoted cof locculation. The overflow rates were low since only a limited
flow could be so routed. The operating conditions utilized during this
test were as follows:
Raw flow rate, gpm 1500
Split flow rate, gpm 930 - 950
Overflow rates, gpm/sq ft High 3.8: Low 1.9
Bead dosage, mg/1 110
Hereof loc 810 dosage, mg/1 6=0 @ 80 cp for 0.5%
Solution (equivalent to 1.65
ing /I at 290 cp)
Ferric chloride dosage, mg/1 30
Note that the viscosity of the flocculant and hence its strength had
decayed significantly and the amount added was increased proportionately.
On the basis of 290 cp representing full strength:
= 3.62 times as weak
80
= 1.65 -mg/1 equivalent
3 .62
The effluent qualities and pollutant removal efficiencies for this test
are shown in Table 28. The overall pollutant removal efficiencies for most
parameters ranged between 50 and 85 percent. The removals were generally
higher by 10-20 percent for the low overflow rate of 1.9 gpm/sq ft compared
with the higher overflow rate of 3.8 gpm/sq ft, reflecting the process
sensitivity to cof locculation time. Particulate removals as shown by
suspended solids and volatile suspended solids were relatively higher (15
to 25 percent) compared with the removals for BOD, COD and TOG. Removals of
75 and 84 percent were achieved for total phosphates. Nitrogen removals
were 12 and 23 percent.
Mechanically, the system worked very smoothly and no problems were
encountered during the testing period. The amount of the sludge build-up
at the bottom of the flotation tank was estimated to be insignificant as
observed by draining out the tank at the end of the run. The pollutant
removals achieved during this test were generally comparable to that
expected via dissolved air flotation although the overflow rates utilized
were considerably lower for this test. Some sinker beads were noticed in
the composite as well as the grab samples obtained during this test,
indicating that the flotation aid slurrying system was causing some bead
damage even when operated for a minimum period of time.
Flotation Aid Recovery Reuse Tests
The objective of this experiment was to define the separability char-
acteristics of the cof locculated sewage solids and flotation aids. No
efforts could be made to define the quantities of cof locculated flotation
72
-------
TABLE 28. BEAD FLOTATION EFFLUENT QUALITY
Overflow Rates, gpm/sq
Analysis
pH, units
Suspended Solids
Vol. Suspended Solids
COD
BOD
TOG
Soluble TOG
Total Phosphate as P
Kjeldahl Nitrogen
Quality
mg/1
6.1
39
15
42
10
14
8
0.23
2.38
1.9
Remova 1
-
86.5
85.1
65.9
60.0
62.2
-
84.2
23.2
Quality
mg/1
6.2
96
43
56
12
17
7
0.37
2.72
ft
3.8
Remova 1
-
66.8
57.4
54.5
52.0
54.1
0
74.7
12.3
73
-------
aid which were recovered as a percentage of those introduced into the over-
flow. Following run 71-1, six batch flotation aid recovery experiments
were conducted with 60 gal of scum collected during the storm. The tests
employed a high-shear Cowles mixer and the hydrocyclone shown in Figure 17.
Analysis of solids in the scum (control sample) and a material balance on
the test setup showed that most of the sewage solids in the scum could be
readily separated from the flotation aid. Since coflocculation of flota-
tion aid and sewage solids was inadequate, separation and recovery of the
flotation aid was probably less difficult (required less energy) than would
be the case for better coflocculated solids. Although not reflected in the
results, most of the sewage solids in the scum were so loosely attached to
the flotation aid that simply handling the scum caused significant separation
and sedimentation of the sewage solids.
Results summarized in Table 29 show that a cyclone can be employed to
separate loosely attached sewage solids from flotation aid. The higher
flow rate in Bead Recovery Test 2 improved sewage solids removal slightly,
from 15 percent to 25 percent. The scum sample used in Test 2 was subjected
to high-shear agitation in bead recovery Tests 3 and 4 to determine if
additional solids could be removed from the flotation aid. The results show
only 0-1 percent additional solids removal.
Bead recovery Tests 5 and 6 were conducted to determine separation
capability of the mixer-cyclone combination. Results of these two tests
(11 percent and 28 percent removal of sewage solids) were not significantly
different from the results in Tests 1 and 2 without high-shear agitation.
Recirculation of the water plus flotation aid stream (Test No. 6) improved
solids removal but overall removal remained poor.
In addition to the flotation aid recovery experiments, samples of scum
were tested for viscosity, density and filtration characteristics, and
recovered flotation aid was tested in batch flotation tests to assess flota-
tion aid recycle potential.
Scum viscosity was 708 cp (Brookfield No. 1 spindle at 30 rpm) and
scum density approached that of water (samples were approximately 98.5
percent water). Scum filtration time (No. 5 Watman filter paper) was quite
variable but was generally longer than 5 min for a 250 cc sample.
Some tests for reusability of the partially cleaned flotation *ids for
Test 2 were conducted. Recovered flotation aid remixed with ferric chloride,
Hercofloc 810 and Milwaukee overflow sewage verified preliminary laboratory
results with Bowling Green, Maryland municipal (domestic sanitary dry-
weather) sewage which showed that recovered flotation aid is reusable.
Comparative batch flotation (clarification) tests with fresh and recovered
flotation aid showed: (1) flotation rate is sensitive to solids cofloccula-
tion, and (2) flotation rate of the recovered material is lower than that
of unused flotation aid. Clarification time for a 250 cc sample, although
difficult to measure quantitatively, was about 1/3 min for unused flotation
aid versus 2/3 min for the recovered material. When chemical dosage was
increased to promote solids coflocculation, clarification time increased
to 1-1/2 minutes.
74
-------
LL
2.2 LG/MIN SUSPENDED SOLIDS
COWLES MIXER
2.6 LB/MIN SUSPENDED SOLIDS
o
30 GAL DRUM
CENTRIFUGAL
PUMP
(1 HP)
HYDROCYCLONE
(10-20 GPM)
SEWAGE SOLIDS
0.4 LB/MIN
SUSPENDED SOLIDS
Figure 17. Schematic diagram of Dylex recovery test.
75
-------
TABLE 29. FLOTATION-AID RECOVERY RESULTS
Test No. Method
1 Cyclone
2 Cyclone
3 Mixer
4 Mixer
5 Mixer + Cyclone
6 Mixer + Cyclone With
Time
(min)
5
15
5
15
< 5
15
Flow
(gpm)
10
10-20
„
-
10
10
Solids Removal
(%)
15
25
1
0
11
28
Recirculation
-------
CONCLUSIONS
Based on the results from storm event No. 71-5, it was demonstrated
that coflocculation of beads and suspended solids found in combined sewer
overflows was possible. Experiments on storm events 71-1, 71-2, and 71-3
were of very limited value with respect to data production. They served
to pinpoint plant startup problems which had to be solved before any
meaningful data could be generated. Unfortunately project funds were
depleted with only one storm event generating useful data.
Although operating conditions were certainly not optimal for storm
event 71-5, good effluent quality resulted. Overflow rates were low,
about 2 gpm/sq ft , when compared with those usual to the test facility
for air flotation, about 3-4 gpm/sq ft. It is not known whether solids
separation efficiency would be significantly affected at higher overflow
rates. The low rates for event 71-5 were necessitated by the operating
requirement for a longer retention time for mixing flocculant, flotation
aid and sewage solids. This requirement was met by maximizing the length
of pipe through which coagulated sewage solids, flotation aid, and floc-
culant could flow while simultaneously mixing. This required routing
essentially the entire test flow through the split flow loop, and the ca-
pacity of this loop was limited to approximately 1000 gpm.
Several technical problems arose during the field testing phase which
limit the practical applicability of the concept. These may be summarized
as follows:
1. Maintenance problems in the handling of flotation aids
and the high energy requirements for keeping these in
suspension.
2. Sensitivity of the successful coflocculation of flotation
aids and suspended matter to the type of flocculant.
Problem No. 1 (above) may be mitigated to a great extent by the
installation of an on-stream bead expansion and dry feed facility but this
would also add additional cost to the total cost of treatment. Storage
and mixing of the flocculant (Hercofloc 810.) was troublesome during the
field tests. However, when this polymer is used in sufficient quantity to
justify the installation of proper equipment for dry storage, rapid mixing
and immediate use, as would be the case in a plant installation, these
problems would disappear.
The logistics of using beads for a flotation aid are very cumbersome
when compared with air bubbles. The mechanism of attaching coagulated
sewage solids to the bead flotation aid is a somewhat complex surface
chemistry phenomenon. Radical changes in the nature of the sewage would
probably necessitate an equally radical change in flocculant, whereas with
air flotation the flocculation mechanism is not so sensitive. The recovery
and recycle of flotation aids have been demonstrated only on a laboratory
77
-------
scale, although the used aids were collected from an actual storm overflow
test run. Indications have been that the recovered beads may exhibit some
loss of efficiency. Combined sewer overflows occur at unpredictable and
inconvenient times from the point of view of system maintenance and super-
vision. With all the above considerations in mind, it is concluded that the
operational complexities of the mechanical version of the flocculation/
flotation process are such that although the process is technically workable,
it is not practical and should not be utilized in the treatment of combined
sewer overflows. Any future field testing should be terminated.
78
-------
SECTION VI
REFERENCES
1. APWA. "Problems of Combined Sewer Facilities and Overflows, 1967".
EPA Report No. 11020 12/67, U.S. Environmental Protection Agency,
(PB NTIS 214-469), December, 1967.
2. Argaman, Y. and Kaufman, W. J., "Turbulence and Flocculation," Journal
of the Sanitary Engineering Division, ASCE, Vol. 96, No. SA2, April
1970.
3. Volk, W., Applied Statistics for Engineers, McGraw-Hill Book Company
(1958).
4. Dixon, W. J., and Massey, F. J., Introduction to Statistical Analysis,
McGraw-Hill Book Company (1957).
5. Hardenbergh, W. A. and Rodie, E. B., Sewerage, ICS Course Book,
International Textbook Company (1966).
6. Mason, D. G., "The Use of Screening/Dissolved-Air Flotation for
Treating Combined Sewer Overflow", Seminar on Storm and Combined
Sewer Pollution Problems, Edison, New Jersey, November 1969.
7. "Development of a Flocculation/Flotation Concept for Solids Separation
in Storm Sewer Systems", Interim Report by Hercules Incorporated, Con-
tract No. 14-12-855 between EPA and Hercules Incorporated, Cumberland,
Maryland, April 1971.
79
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
EPA-600/2-77-140
3. RECIPIENT'S ACCESSI Of* NO.
4. TITLE AND SUBTITLE
FLOGCULATION-FLOTATION AIDS FOR TREATMENT OF COMBINED
SEWER OVERFLOWS
5. REPORT DATE
August 1977_(Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO
N. F. Stanley and P. R. Evans
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Hercules Incorporated
Allegany Ballistics Laboratory
P.O. Box 210
Cumberland, MD 21502
10. PROGRAM ELEMENT NO.
1BC611
11. CONTRACT/CHANT NO.
14-12-855
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory—Cin.
Office of Research and Development
U. S. Environmental Protection Agency
Cincinnati, OH, 45268
OH
13. TYPE OF REPORT AND PERIOD COVERED
FINAL
14. SPONSORING AGENCY-CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
Project Officers: Clifford Risley, 312-353-2200
Richard Field, 201-321-6674, (8-340-6674)
16. ABSTRACT
The objectives of this study were to investigate the flocculation/flotation char'
acteristics of combined sewer overflow through laboratory and field testing. The con'
cept involves the introduction of chemicals and buoyant flotation aids into the over-
flow and the subsequent coflocculation of the suspended sewage solids about the aids
which rise to the surface from where they may be removed by skimming.
Recovery, cleaning and reuse of the flotation aids was judged essential to
economic feasibility for the flocculation/flotation process. Field testing of
recovered used aids indicated a disparity in requirements. When the bonding, or co-
flocculation, of suspended sewage solids to the fresh aid surface was sufficiently
strong to provide a high degree of suspended solids removal, subsequent efforts to
clean or remove the solids in anticipation of aid reuse were unsuccessful. This
deficiency, coupled with the overall mechanical complexity of the process, resulted
in the investigation concluding that the flocculation/flotation process is not as
promising for broad field application to the storm overflow problem as the similar
dissolved air flotation process. No further investigation of the flocculation/
flotation process is recommended.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
-Coagulation, *Combined sewers, Flocculants,
^Flocculating, ^Flotation, Flotation
conditioning, Flotation Reagents, Sewage,
Water pollution, Sewage treatment,
b.IDENTIFIERS/OPEN ENDED TERMS
Combined sewer overflows,
Treatment screening/
dissolved air flotation,
Sewage effluents
c. COSATI Field/Group
13 B
3. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO..OF PAGES
92
20. SECURITY CLASS (This page)
. UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
80
iT U.S. GOVERNMENT PRINTING OFFICE. 1977—757-056/6493
------- |