EPA
United States
Environmental Protection
Agency
Municipal Environmental Research EPA-600/2-78-122
Laboratory August 1978
Cincinnati OH 45268 ,
Research and Development
The Swirl Primary
Separator
Development and Pilot
Demonstration
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are: ',
1. Environmental Health Effects Research
2, Environmental Protection Technology
3. Ecological Research :
4. Environmental Monitoring
5, Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7, Interagency Energy-Eni/ironment Research and Development
8. "Special" Reports
9. Miscellaneous Reports!
This report has been assigned tt> the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipmbnt, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the rjublic through the National Technical Informa-
tion Service, Springfield, Virginia 122161.
-------
EPA-600/2-78-122
August 1978
THE SWIRL PRIMARY SEPARATOR:
DEVELOPMENT AND PILOT DEMONSTRATION
Richard H. Sullivan
Morris M. Cohn
James E. lire
Fred Parkinson
G. Galiana
Ralph R. Boericke
Carl Koch
Paul Zielinski
American Public Works Association
Chicago, Illinois 60637
Contract No. 68-03-0272
Grant No. S-803157
Project Officers
Richard Field
Hugh Masters
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
Tills report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for
publication. 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 recommendation
for use.
11
-------
FOREWORD
The U.S. Environmental Protection Agency was created because of
increasing public and governmental 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 Labora-
tory develops new and improved technology and systems for the
prevention, treatment, and management 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 aes-
thetic 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.
The study describes what has been learned from both laboratory
and prototype testing of a new type of primary settling device. This
device is to be used with urban stormwater runoff and combined sewer
overflows, as well as domestic sanitary sewage, to provide primary
treatment.
Francis T. Mayo
Director
Municipal Environmental Research
Laboratory
111
-------
ABSTRACT
A study was conducted to determine if the swirl concentrator principle
could be used to provide primary treatment to combined sewer overflows and
municipal wastewater. A hydraulic model with synthetic wastewater and a
mathematical model were both used to arrive at an optimized configuration
and a design basis. The design was then tested under actual wet- and
dry-weather flow conditions using a large scale, 1,137 cu m/d (0.3 mgd) pilot
Constructed in Toronto, Canada. The Toronto pilot evaluations confirmed the
accuracy of the design (and associated design curves) developed under the
model studies.
The model and pilot studies indicated that the device could achieve 30 to
50 percent settleable solids removal efficiency for flows of less than 22 I/sec
(d.5 mgd) at costs comparable to, or less than, conventional treatment units.
Overflow rates of two to three times that of conventional units make possible
the saving.
Testing of the model and prototype was based upon the need to treat both
domestic sanitary sewage and combined sewer overflows. Extensive laboratory
work was conducted to determine the settling characteristics of solids to
provide laboratory control and provide a correlation between the laboratory and
prototype testing programs.
The swirl's height and diameter are equal, providing a relatively deep
structure which enhances sludge thickening.
The Toronto pilot evaluations of the prototype unit constructed in
Toronto Ontario — 1,137 cu m/d (0.3 mgd) - to determine operating
efficiencies confirmed the accuracy of the design and associated curves
developed under the model studies.
The report contains thorough descriptions of the hydraulic/mathematical
and pilot studies, and most importantly, the detailed design methodology.
This report is in partial fulfillment of U.S. Environmental Protection
Agency (EPA) contract 68-03-0272 (for hydraulic and mathematical modeling)
jointly sponsored by the American Public Works Association Research
Foundation (APWA), and EPA demonstration grant S-803157 (for pilot
evaluation) jointly sponsored by APWA and the City of Toronto, Canada. Work
was completed in October 1976.
IV
-------
TABLE OF CONTENTS
Page
Forword . ....................................... iii
Abstract ............................ ........... iv
Acknowledgements .................................... x
Section 1 . Conclusions, Recommendations, and Overview of the Studies .......... 1
Section 2. The Study .................................. 5
Section 3. Design Guidelines and Construction Costs .................. 23
Section 4. Pilot Treatability Evaluation With Sewage .................. 36
Section 5. Glossary .................................. 58
References ................................. 58
Appendices
A. Hydraulic Model Study ........................... 59
B. Mathematical Model Study ......................... Ill
References .............................. 153
C. Settleability Tests and Hydrualic Characterization of a Pilot (with. real sewage)
Swirl Separator as a Primary Treatment Facility ............... 154
D. Pilot Test Results .................... ....... 166
FIGURES
Page
1. Isometric: Swirl Primary Separator .......................... 3
2. Model Layout ... ................................ 7
3. Swirl Primary Separator: First Layout Tested on Model ................ 9
4. Swirl Primary Separator: Model Layout for Tests 161 to 184 —
Modification 10 ............................... ... 10
5. Swirl Primary Separator: Model Layout for Tests 188 to 194
(3 to 9) - Modification 1 1 ............................ 11,12
6. Gradation Curve for Petrothene Used in Model ................... 14
7. Settling Velocity vs Particle Size for IRA-93 Anion Exchange Resin .......... 15
8. Swirl Primary Separator: Sanitary Sewage in
Prototypes Represented by IRA-93 in Model ........... ........ ._ 16
9. Diagram of Swirl Primary Separator Chamber as Represented by
Mathematical Model ................................ 19
10. Pilot Facility - Metro Toronto ........................ 22
11. Predicted Prototype Solids Removal Efficiency for Sanitary Sewage ........ 24,25
1 2. Detention Times ................................. 26
13a. General Design Dimensions ............................ 27
1 Sb.General Design Dimensions ................ ............ 28
14. Swirl Primary Separator .............................. 31
15. Conventional Primary Settling Tank ........................ 33
16. Cost vs. Diameter, Swirl and Conventional Primary Treatment Units ......... 34
17. Test Layout - Humber Plant - Toronto, Canada 1975 ............... 37
18. Diurnal Flow and Suspended Solids (mg/1) Average of 5 Dry-Weather
Days, Humber Plant -Toronto, April 1975 .................... .39
1 9. Section of Swirl Primary Separator ......................... 40
20. Plan of Swirl Primary Separator ...... .................... 41
-------
FIGURES (continued)
21. Dry-Weather Removal of Total Suspended Solids-Toronto, September 1975 46
22. Dry-Weather Removal of Fixed Suspended Solids —Toronto, September 1975 .... 47
23. Dry-Weather Removal of Volatile Suspended Solids -Toronto, September 1975 ... 48
24. Dry-Weather Removal of Settleable Solids by Volume - Toronto, September 1975 . . 49
25. Dry-Weather Removal of Settleable Solids by Weight-Toronto, September 1975 ... 50
26. Predicted Versus Actual Solids Removal 54
27. Comparison of Time to Achieve Treatment 55
28. Swirl Primary Separator: Conical Bottom and Low Inlet 60
29. Swirl Primary Separator — Modification 2, (Raised Inlet 61 cm [24 in]
Diameter Weir and 71 cm [28 in] Diameter Skirt) 60, 61
30. Swirl Primary Separator: Operating Efficiencies for Modifications 2 and 3 65
31. Swirl Primary Separator: Model Layouts for Tests 20 to 45 — Modification 3 66
32. Prototype Particle Sizes Simulated by Shredded Petrothene 67
33. Swirl Primary Separator: Model Layouts for Tests 27, 28, 34 and 40 -
Modifications 3 and 4 68
34. Swirl Primary Separator: Operating Efficiencies for Modifications 3 and 4 .... 69
35. Swirl Primary Separator: Model Layout for Tests 46 to 68 — Modification 5 .... 70
36. Swirl Primary Separator: Operating Efficiencies for Modification 5 .... 71
37. Swirl Primary Separator: Model Layout for Tests 69 to 91 for Modification 6 .... 72
38. Swirl Primary Separator: Operating Efficiencies for Modification 6 73
39. Swirl Primary Separator: Model Layout for Tests 113-136-Modifications 75
40. Swirl Primary Separator: Operating Efficiencies for Modification 8 76
41. Swirl Primary Separator: Location of Measuring Points for Velocity Contour
Tests 10.98m (36 ft) Chamber Prototype Scale 1/12-Modification 6 77
42. Swirl Primary Separator: Tangential Velocity Contours at 0° Position 78
43. Swirl Primary Separator: Tangential Velocity Contours at 90° Position 78'
44. Swirl Primary Separator: Tangential Velocity Contours at 180° Position .... 79
45. Swirl Primary Separator: Tangential Velocity Contours at 270° Position .... 79
46. Swirl Primary Separator: Model Layout for Tests 92 to 112 - Modification 7 .... 80
47. Swirl Primary Separator: Operating Efficiencies of Closed Bell with
Four Orifices - 3.79 cm (1.5 in) 6, 22,9 cm (9 in) High, Skirt
0.61m (24 in) 6 for Modification 7 \ 81
48. Swirl Primary Settler: Location of Measuring Points for Velocity Contour
Tests 10.98 m (36 ft) Chamber, Closed Bell with Four Orifices - 0.457 m
(1.5 It) 6, 2.74 m (9 ft) High, Skirt 7.32 m (24 ft) 6 Prototype
Scale 1/12-Modification 7 .'...; 82
49. Swirl Primary Separator: Tangential Velocity Contours at 0° Position 83
50. Swirl Primary Separator: Tangential Velocity Contours at 90° Position 84
51. Swirl Primary Separator: Tangential Velocity Contours at 180° Position 85
52. Swirl Primary Separator: Tangential Velocity Contours at 270° Position 86
53. Swirl Primary Separator: Model Layout for Tests 137 to 160 — Modification 9 . ... 88
54. Swirl Primary Separator: Operating Efficiencies for Modification 9 89
55. Swirl Primary Separator: Model Layout for Tests 161 to 184 — Modification 10 ... 90
56. Swirl Primary Separator: Operating Efficiencies for Modification 10 —
Increase of the Skirt Diameter . . 91
57. Swirl Primary Separator: Operating Efficiencies for Modification 10 —
Recovery Rate vs Slot Height . . 92
58. Modification 10, Petrothene Grains Reaching End of Inlet, and Petrothene
Grains Slidin Down Against the Cone Wall Near Inlet 93
59. APWA Swirl Primary Settler: Retention Time vs Discharge With Scale 1/12 94
VI
-------
FIGURES (continued)
60. Swirl Primary Separator: Model Layout for Modification 11 98, 99
61. Swirl Primary Separator: Model Layout for Modification 12 100
62. Swirl Primary Separator: Model Layout With Modification 12 and
Inlet Baffle Type 3 a 102
63. Swirl Primary Separator: Model Layout with Modification 12 and
Inlet Baffle Type 3b ' 103
64. Swirl Primary Separator: Model Layout with Modification 12 and
Inlet Baffle Type 3 c 104
65. Swirl Primary Separator: Model Layout with Modification 12 and
Inlet Baffle Type 3d 105
66. Swirl Primary Separator: Suggested Recovery Curve for Anion Exchange
Resin IRA-93 in Model 149 ju> d> 75 M 106
67. Swirl Primary Separator: Predicted Prototype Solids Recover from
Sanitary Sewage 108
68. Recovery Rates on Model as Function of Particle Settling Velocity 109
69. Recovery Rates with Different Particle Settling Velocities for:
a. 0.91 m (3 ft) Diameter Chamber 110
b. 1.83m (6 ft) Diameter Chamber HO
c. 3.66m (12 ft) Diameter Chamber 112
d. 7.33m (24 ft) Diameter Chamber 112
e. 10.98m (36 ft) Diameter Chamber 113
70. Diagram of Laboratory Swirl Chamber Configuration 115
71. Diagram of Swirl Chamber as Represented by Mathematical Model 116
72 Settling Velocity Distribution at 30.5 cm (1 ft) Column Sample Ports 119
73. Settling Velocity Distribution at 61 and 91 cm (2 and 3 ft) Column Sample Ports . . 119
74. Settling Velocity Distribution for all Sample Ports 120
75. Settling Velocity Distribution at Various Column Depths 120
76. Settling Column Data Interpretation: Flocculation vs Nonflocculating Particles ... 122
77. Streamlines for Initial Lab Configuration with Circular Weir 136
78. Tangential Velocity Contours for Initial Lab Configuration with Circular Weir .... 137
79. Vertical Velocity Profiles for Initial Lab Configuration with Circular Weir 138
80. Streamlines for Modified Lab Configuration with Radial Gutters and
Standpipe Removed 140
81. Tangential Velocity Contours for Modified Lab Configuration with Radial
Gutters and Standpipe Removed 141
82. Vertical Velocity Profiles for Modified Lab Configuration with Radial
Gutters and Standpipe Removed 142
83. Particle Paths for Initial Lab Configuration with Circular Weir and Particle
Settling Velocity of 0.05 cm/sec (0.0016 ft/sec) 143
84. Particle Paths for Modified Lab Configuration with Radial Gutters and Standpipe
Removal at Particle Settling Velocity of 0.05 cm/sec (0.0016 ft/sec) 145
85. Particle Paths for Modified Lab Configuration with Radial Gutters and Standpipe
Removal at Particle Settling Velocity Equal to Average Upflow Velocity of
0.25 cm/sec (0.0083 ft/sec) 146
86. Particle Paths for Modified Lab Configuration with Radial Gutters and Standpipe
Removal at Particle Settling Velocity of 0.3 cm/sec (0.0098 ft/sec) 147
87. Comparison of Smisson 3 m (10 ft) Prototype Data with Theoretical Performance . . 149
88. Correlation of Observed and Theoretical Model Performance 150
89. Settling Velocity Characteristics: Humber Treatment Plant—
Storm Flow June 5, 1975 ,159
vu
-------
FIGURES (continued)
90. Settling Velocity Characteristics: Humber Treatment Plant —
Storm Flow June 16, 1975 159
91. Residence Time Distribution 165
92. Residence Time Distributions 165
93. Fitted Responses , 166
94. Fitted Responses 166
95. Fitted Responses 167
96. Fitted Responses 167
97. Percent Dead Volume Versus Flow Swirl Separator 168
TABLES
Page
1. Solids Removal Efficiencies of IRA-93 at Various Discharge Rates 17
2. Comparison of Diameter, Detention Time, and Suspended Solids Removal for
Swirl Primary Separator and Detention Time for Conventional Settling for
Various Overflow Rates 29
3. Construction Cost of Swirl Primary Separator 32
4. Construction Cost of Conventional Primary Settling Tank 34
5. Comparison of Operation and Maintenance Costs for Primary Treatment Units ... 35
6. Present Worth, Swirl Separator Primary Treatment Units 35
7. Removal of Total Suspended Solids: Wet-Weather Flow -May 4- July 12, 1975 .... 37
8. Removal of Total Suspended Solids-June 23-27, 1975 42
9. Removal of Total Suspended Solids-July 2-8, 1975 . 43
10. Removal of Total Suspended Solids: Dry-Weather Flow-May 1-June 10, 1975 ... 44
11. Summary of Tests-September, 1975 51
12. Detention Time and Overflow Rate: Primary Tanks — September 1975 and June 1976 52
13. Total Suspended Solids 52
14. Removal of BOD-June 23-July 8, 1975 55
15. Modifications Tested on the Model 64
16. Successive Modifications of the Model and Recovery Results with Anion Exchange
Resin IRA-93 149 v>d>14n 95, 96
17. Recovery Rate of IRA-93 at Various Flowrates 101
18. Representation of IRA-93 to Sewage, Based on Scale Factor 107
19. Collision Rates for Various Mechanisms 123
20. Particle Size and Number Density for Various Settling Velocities 126
21. Estimated Collision Rates 128
22. Applicable Boundary Conditions for Particle Continuity Equation 131
23. Particle Settling Velocity Distribution for 100-200 Mesh IRA-93 Resin 148
24. Predicted Removal Efficiency for 100-200 Mesh IRA-93 Resin (0.5 I/sec [0.02 cfs]) . 148
25. Predicted Removal Efficiency for Hypothetical Prototype Unit Using
Mathematical Model . 153
26. Predicted Removal Efficiency for Hypothetical Prototype Unit Using
Laboratory Data 154
27. Model Fitted Parameter Estimates 162
28. The Axial Dispersion Model 163
29. The Equal Tanks-In-Series Model 164
30. Dry-Weather Test Schedule 169
31. Wet-Weather Test Schedule 169
32. Wet Weather Test Data, Swirl Separator -May 4-7, 1975 170
Vlll
-------
TABLES (continued)
33. Wet-Weather Test Data, Swirl Separator-May 15-June 12, 1975 171
34. Wet-Weather Test Data, Primary Tanks -May 4-7, 1975 172
35. Heavy Metal Test Data 173
36. Dry-Weather Test Data-Swirl Primary Separator-May 1-11, 1975 174
37. Dry-Weather Test Data - Swirl Primary Separator -May 12-18, 1975 175
38. Dry-Weather Test Data-Swirl Primary Separator-May 30-June 10, 1975 176
39. Dry-Weather Test Data - Swirl Primary Separator -June 16-21, 1976 177
40. Dry-Weather Test Data - Swirl Primary Separator- June 22-25, 1976 177
41. Dry-Weather Test Data-Primary Tanks-May 1-11, 1975 178
42. Dry-Weather Test Data-Primary Tanks-May 12-18, 1975 179
43. Dry-Weather Test Data - Primary Tanks - May 30-June 10, 1975 180
44. Dry-Weather Test Data-Primary Settling Tank-June 16-21, 1976 181
45. Dry-Weather Test Data-Primary Settling Tank-June 22-25, 1976 181
46. Proposed Tests Second Series 132
47. Dry-Weather Test Data - Swirl Primary Separator- June 12-17, 1975 182
48. Dry-Weather Test Data-Primary Tanks-June 23-27, 1975 183
49. Dry-Weather Test Data - Swirl Primary Separator- July 2-8, 1975 184
50. Dry-Weather Test Data - Primary Tanks - July 2-8, 1975 '. 185
51. Proposed Tests Final Series ^.86
52. Dry-Weather Test Data-September 2, 1975 137
53. Dry-Weather Test Data September 3, 1975 188
54. Dry-Weather Test Data September 4, 1975 189
55. Dry-Weather Test Data September 5, 1975 ' ', 190
56. Dry-Weather Test Data September 8, 1975 '.'.', 191
57. Dry-Weather Test Data September 9, 1975 192
58. Dry-Weather Test Data September 10, 1975 193
59. Dry-Weather Test Data September 11, 1975 ' 194
60. Dry-Weather Test Data September 12, 1975 ' ' 195
61. Dry-Weather Test Data September 15, 1975 '. 196
62. Dry-Weather Removal of Settleable Solids By Volume 197
63. Dry-Weather Removal of Settleable Solids By Weight 198
64. Dry-Weather Removal of Total Suspended Solids 199
65. Dry-Weather Removal of Volatile Suspended Solids 200
66. Dry-Weather Removal of Fixed Suspended Solids 201
67. Dry-Weather Removal of Settleable Solids By Volume 202
68. Dry-Weather Removal of Settleable Solids By Weight 203
69. Dry-Weather Removal of Total Suspended Solids 204
70. Dry-Weather Removal of Volatile Suspended Solids 205
71. Dry-Weather Removal of Fixed Suspended Solids 206
EXHIBITS Page
1. Settling Column Test Methods 158
2. Description of Hydraulic Mixing Models . . '. '. '.'.'..'... 162, 163 164
IX
-------
ACKNOWLEDGEMENTS
The American Public Works Association is deeply indebted to the following
persons and their organizations for the services they have rendered to the APWA
Research Foundation in carrying out this study for the U. S. Environmental
Protection Agency.
PROJECT DIRECTOR
Richard H. Sullivan
CONSULTANTS
Morris M. Cohn, P.E.
Paul B. Zielinski, P.E.
Bernard S. Smisson
ALEXANDER POTTER ASSOCIATES, CONSULTING ENGINEERS
Morris H. Klegerman, P.E.
James E. Ure, P.E.
GENERAL ELECTRIC COMPANY
Ralph R. Boericke
: Carl Koch
LA SALLE HYDRAULIC LABORATORY, LTD.
F. E. Parkinson
George Galiana
METROPOLITAN TORONTO
Ross L. Clark
Earl Baldock
T. W. BEAK CONSULTANTS
Stephen L. Hodd
AMERICAN PUBLIC WORKS ASSOCIATION STAFF
Ronald H. Ball
Lois V. Borton
Jan A. T. Harvey
John R. Moy
Rose M. Ohlman
Shirley M. Olinger
Cecelia E. Smith
Oleta M. Ward
U. S. ENVIRONMENTAL PROTECTION AGENCY
Richard Field, Chief,
Storm & Combined Sewer Section, (Edison, New Jersey)
Hugh E. Master, Staff Engineer
Storm & Combined Sewer Section (Edison, New Jersey)
-------
SECTION 1
CONCLUSIONS, RECOMMENDATIONS, AND OVERVIEW OF THE STUDIES
CONCLUSIONS
The studies indicate that the swirl flow
principle can be a substitute for the function
of primary, settling. Previously the swirl flow
principle was shown to be applicable for the
simultaneous flow regulation and rough
clarification of combined sewer overflow
liquids and for the selective separation of grit
in sanitary sewage and admixed sewage and
storm runoff.
1. As a result of hydraulic and
mathematical model studies, it has been
concluded that the swirl separator principle
can be utilized to obtain clarification of
wastewaters, such as sewage, combined sewer
overflows, and storm water discharges. Such
clarification is competitive with conventional
settling units for low flow rates and at less
than conventional primary settling
removal requirements.
2. The; swirl separator principle applied
to primary separation involves the
establishment of long-path circular separation
within a cylindrical tank. The opportunity for
natural flocculation and the lessened
possibility of short circuiting in the chamber
appear to improve performance.
3. The absence of moving parts in the
swirl separator reduces maintenance and
operating costs of the unit.
4. The relatively high overflow rates
which may be used with the swirl separator at
various levels of suspended solids removal
requires less space, thus enhancing its use in
wastewater plant expansion and combined
sewer overflow and stormwater treatment.
5. Tests covering the settling velocities of
sewage and flocculation mechanisms indicate
a good relationship between the synthetic
solids used in model testing and real sewage.
6. Testing of the pilot unit was
conducted under the adverse conditions of
wide variations in flowrates of industrial
wastes. Thus the fact that the unit performed
as well as the existing units indicates that
performance may be enhanced under more
uniform operating conditions.
7. The, cost of constructing a swirl
separator for flow of 7.5 I/sec (0.17 mgd) for
a suspended solids removal efficiency of 45
percent was less than for a conventional unit,
$55,200 to $76,000. However, above this
flowrate and removal efficiency, the conven-
tional unit becomes less expensive.
8. With our present knowledge of fine
particle scale-up laws, the swirl separator does
not appear to be economically competitive
for diameters greater than 5.5 m (18 ft). How-
ever, model and limited field test results are
sufficiently encouraging to suggest construc-
tion of a larger experimental unit to provide
full scale data for actual sewage.
RECOMMENDATIONS
1. The hydraulic and mathematical
model investigations and the prototype tests
indicated good removal of sewage solids. The
studies, however, may have under-predicted
the swirl primary separator efficiencies which
can be accomplished. It is recommended that
the prototype installation constructed by
Metropolitan Toronto be tested at another
facility not subject to large variations in
industrial waste loads. The potential benefits
of low construction cost, applicability of
units to augment presently overloaded
primary facilities, freedom from mechanical
parts or devices to handle sludge residues and
ability of the deep conical sludge hopper con-
figuration warrant the expenditure of funds
for such an evaluation.
2. Further testing and development of
the swirl separator principle should be
conducted utilizing mechanical sludge rakes
and the modified design basis of Mr. Bernard
Smisson.
OVERVIEW OF THE STUDY
The swirl separator can prove to be a
highly valuable and innovative tool in the
nation's efforts to clean up pollution
conditions in its water resources. Previous
studies carried out at Bristol, England, but
never before duplicated on the American
continent, gave indication that the swirl
-------
separator principle could be applied to the
separation of liquid and solid phases in
wastewaters of various types.
The current studies of the swirl primary
separator is, in fact, a "third-generation"
investigation of the swirl separator principle.
In July 1972, the APWA Research
Foundation completed development and
evaluation studies of the swirl separator as a
combined sewer overflow-regulator facility
jointly sponsored by the City of Lancaster,
Pennsylvania, and the U.S. Environmental
Protection Agency (EPA).1 The investigations
demonstrated the practicability of utilizing
the swirl principle for removing settleable and
floatable solids from combined sewer
overflow and for simultaneously diverting
predetermined amounts of so-called foul
sewage (concentrated liquid-solids) to
downstream wastewater treatment facilities or
to holding (storage) chambers.
Encouraged by the results of the swirl
separator as a combined sewer overflow
regul a t o r/s eparatorJ device a
"second-generation" study was conducted on
the development and evaluation of the swirl
principle for the selective separation of grit or
inorganic solids from wastewater flows.2 The
swirl degritter development was conducted by
the American Public Works Association under
the sponsorship of the USEPA in 1973.
In 1974 a study of the use of the swirl prin-
ciple for removal of erosion products from
stormwater runoff was conducted.6
The present studies of the development
of a swirl separator chamber as a primary
clarifier of sanitary sewage and combined
sewer wastewaters involved the development
and evaluation of a hydraulic model at the
La Salle Hydraulic Laboratory, La Salle,
Quebec, Canada, the verification :and
validation of hydraulic model findings by
mathematical-computer model techniques
performed by the General Electric Company,
Re-Entry and Environmental Systems
Division, at Philadelphia, Pennsylvania, and
the evaluation of a prototype unit with san-
tary sewage in Toronto, Ontario.
In addition to the aforementioned mpdel
investigations, a supplemental investigation
was carried out by Beak Consultants Limited,
Rexdale, Ontario, Canada, to determine the
settling velocities of combined sewer overflow
and stormwater runoff with respect to the
application of swirl separator devices as
primary separators.3 These studies were
interrelated with the hydraulic and
mathematical model investigations because
they dictated the nature of the synthetic
model-scale solids which were tested and
evaluated as representative of actual
wastewater solids that would be treated by
swirl devices in the field. In conjunction with
the prototype field test in Toronto, addition-
al tests of settling characteristics of solids in
combined sewer flows and flow patterns
within the swirl separator were conducted.
The hydraulic studies covered 12
modifications of a swirl separator chamber
originally patterned after the unit previously
used in the earlier overflow regulator
investigations.1'2'3 A total series of 194 test
runs were carried out, covering
determinations of solids settling velocities,
flow patterns, solids removal efficiencies, and
other pertinent factors. Out of these 'Studies
came firm decisions on the most effective size
and location of inlet line, shape of bottom
cone solids hopper, baffle and overflow weir
details, and other internal appurtenances.
Figure 1 shows the various parts of the unit.
The mathematical model studies were
designed to verify all hydraulic patterns and
performances.
The studies provided proof of the
applicability of the swirl principle to the
function of primary clarification of suspended
solids-bearing wastewaters. In a short
detention period solids are deposited by
inertial and gravity action and agglomeration
mechanisms. Removal efficiencies matched
actual performance of conventional primary
settling facilities in shorter periods of time at
the Metropolitan Toronto facility. With the
design configuration, without chemical add-
itives, the swirl was not found useful above
50 percent suspended solids removal or for
flows in excess of 40 I/sec (0.9 mgd) treated
to 30 percent solids removal.
The hydraulic model studies were based
on synthetic solids made of Amberlite® anion
exchange resin IRA-93, which was considered
-------
Sludge Discharge
A inlet
B baffle
C skirt
D gutters
E clear effluent outlet
F baffle
G sludge discharge
FIGURE 1 ISOMETRIC: SWIRL PRIMARY SEPARATOR
-------
to properly simulate actual solids in sanitary
sewage flows.3 The design criteria are based
on Froude Law scale up.
The hydraulic model studies with
synthetic solids led to the construction of a
3.7 in- (12 ft) diameter pilot with real sewage
for further hydraulic evaluation of swirl,flow
phenomena, size and geometric configuration
of basic chambers and internal structure
features, performance in terms of solids
removal and the effects of solids coagulation
induced by the swirl patterns.
The pilot unit was tested at a design flow
of 1,137 m3/day (0.3 mgd) and at 1,700
m3/day (0.45 mgd). The results of the tests
indicated that the unit performed as
effectively, 40 percent suspended solids
removal, as conventional basins at the
Humber wastewater treatment plant operating
at an overflow rate of 81.46 m3/day/m2
(2,000 gal/day/ft2) with 1.06 hour detention
time. The detention time in the swirl
separator was 20.4 minutes and an overflow
rate of 108 m3/day/m2 (2,650 gal/day/ft2)
and 14 minutes with an overflow rate of 162
m3/day/m2 (3,980 gal/day/ft2), respectively
for the two above mentioned flowrates.
However, when the size of the unit is
calculated for a full 60 percent suspended
solids removal, the size and retention time
becomes equal to that of the conventional
unit.
General Comments
The hydraulic model design was started in
the summer of 1974. At that time the work
of Mr. Bernard Smisson was closely evaluated
because of his extensive research in the field
of secondary motions in flow fields. Mr.
Smisson had experimented with both
mechanical and nonmechanical
configurations. At the time of the study, Mr.
Smisson had decided to proceed on the basis
of a mechanical scraper to minimize the depth
of the unit and improve operating efficiency.
Inasmuch as a primary goal of the current
study was to develop a unit for treating
combined sewer overflows and urban
stormwater runoff, the decision was made to
develop a nonmechanical device. With this
limiting design feature several major changes
exist in the configurations which are different
from the current designs being evaluated by
Mr. Smisson. Among the major features are
the following:
1. Shape
This Study the height equals diameter
with 60° sloped conical bottom
Smisson a conventional shaped circular tank
2. Draw-off clarifie'd flow
This Study conventional gutters with
gravity discharge through side of tank
Smisson circular weir and
central shaft discharge
3. Draw-off of solids
This Study
Smisson
4. Deflector
This Study
Smisson
5. Dip plate
This Study
tests indicate intermittent
draw-off best
continuous draw-off
and uses mechanical rake
found no advantage
to inlet baffle or deflector
uses skirt deflector to
create distinct zones
found no hindrance
to settling from surface interferences,
i.e., pipe across outer flow field
Smisson does not allow a dip plate
or interference in outer flow field
Smisson reports that his configuration is
performing better than that found by th'e
APWA under this study.
4
-------
SECTION 2
THE STUDY
Primary separation of solids is the
common denominator of practically all
treatment systems whether for sanitary or
combined wastewaters plant processes. While
present federal laws and the policies of all
states require the use of higher degrees of
treatment than simple primary clarification,
the removal of settleable solids and floatable
suspended material from wastewaters by some
form of clarification facility is universally
practiced. Some secondary treatment systems
may forego primary removal of suspended
solids prior to their secondary stages, such as
in the case of so-called package activated
sludge or extended aeration plants, but initial
clarification is normally provided in
conventional treatment plant design.
The basic conventional methodology of
primary clarification usually involves the
gravimetric separation of settleable solids,
together with floatable materials, in
flow-through chambers which provide
detention periods and lowered hydraulic
velocities that produce deposition and rise of
solids and the resultant discharge of a clarified
effluent. Decreased flow velocities to provide
this gravity clarification phenomenon
usually involve relatively large settling
chambers and relatively long detention
periods.
The swirl separator achieves clarification
of solids-laden liquid flows, not by simple
gravimetric separation under quiescent
settling conditions, but by introducing
secondary motion flows. The purpose of this
study has been to investigate the application
of the swirl separator principle to the process
of primary separation of raw wet and dry
weather sewage solids and to evolve practical
design parameters for prototype installations.
The successful achievement of the application
of the swirl separator principle for this
universal waste water treatment process could
provide relatively speedy, or "flash," solids
separation in relatively small chambers
without use of mechanical devices for sludge
and scum removal, thus providing economy
of construction and operation.
HYDRAULIC MODEL STUDY
The hydraulic model studies were carried
out by LaSalle Hydraulic Laboratory where
the previous investigations of swirl separator
for combined sewer overflow regulator
purposes and for grit removal by swirl action
were undertaken. The experience of this
hydraulic laboratory and its knowledge and
availability of previous swirl chamber
components greatly benefited and expedited
the current research work.
The hydraulic model studies, augmented
and supported by the mathematical and
computer simulation of liquid and particle
flowfield conditions were programmed to
produce optimum swirl separator chamber
dimensions and configurations and structural
details of internal appurtenant unit
components; to evaluate performance in
terms of most effective solids removal
efficiencies; and to develop design procedures
and parameters that translate the scale model
features into full-scale prototype swirl
primary separator installations in the field.
This section applies to the hydraulic model
investigations performed at the LaSalle
Hydraulic Laboratory.
In order to initiate the swirl studies, the
model chamber previously used as a combined
sewer overflow regulator was modified to
serve as a primary separator device. Changes
were made in the geometries and internal
details of the original unit, including
modifications to the chamber depth; floor
configuration; inlet shape and location; wen-
diameter, configuration, and structural
details; skirt location, size, and immersion;
and slot opening dimension under the skirt.
The first model configuration and all
subsequent changes in the main chamber and
internal appurtenant units were intended to
provide a controlled combination of swirl
-------
flow patterns and gravimetric settling and a
bottom hopper shape to expedite the
deposition and compaction of deposited
sludge solids to a prearranged point of
discharge without need for mechanical
collectors of any type. An outlet for the
hopper sludge was provided in the original
model but no routine use of this sludge line
was made in the studies. However,
intermittent draw-off under flow influence
was included in the testing program.
The first model layout was based on
various sources of information and experience
with units of similar or related intent and
design. The general arrangement with a deep
skirt for separation of the external and
internal sectors of the swirl chamber, as
developed by Mr. Smisson in model studies at
Bristol, England, was first utilized. The
bottom cone of the sludge collecting hopper
of the chamber was chosen as a 60° slope for
easy movement of sludge to the outlet using
gravity force only. Larger inlet line diameter,
weir diameter, and skirt diameter than used
previously in swirl separator studies of
combined sewer overflow regulators * and of
grit removal facilities2 were provided.
In order to establish practicable
prototype values for the hydraulic model
studies the scale of 1 : 12 previously used in
regulator overflow and grit separation studies
was again utilized. In prototype scale up
dimensions, this dictated the use of a 1.22 m
(4 ft) square influent sewer, entering a 11 m
(36 ft) diameter swirl chamber with overflow
weir crest 2.75 m (9 ft) above the upper
cylindrical portion of the chamber during the
first stage portion of the overall hydraulic
model studies. Various changes were
subsequently made to improve operational
performance but the diameter of the chamber
itself was retained throughout.
This size chamber provided a 21 minute
detention time in the prototype for a flow
rate of 0.31 m3/sec (11 cfs), corresponding to
0.632 I/sec (0.022 cfs) in the model.
The discharge Q and time to flow relation-
ship can be developed from the continuity
equation where Q prototype = L3
prototype/time prototype or:
QP=Lp3/t (.1)
and for the model: Qm = Lm 3 /tm.
Combining the continuity equation with
the Froude Model.Law (V/^/Lg) leads to the
relationship involving the time scale and
discharge scale between model and prototype
as
t fc-r,
*P -\QrnJ *» (2)
The relation between the discharge Q and the
length scale is.
OIL. = f lx^\ 2'5
Qm \Lm ) (3)
To simplify the testing procedure, four
model discharge rates were selected to provide
workable prototype equivalents, as follows:
0.9m (3 ft)
Diameter
Model
I/sac
0.5
0.632
1.0
1.5
Detention
Time
minutes
7.67
6.06
3.83
2.56
m3/sec
0.25
0.31
0.5O
0.75
1 1 m (36 ft)
Diameter
Prototype
cfs
8.8
11.0
17.6
26.4
mgd
5.7
7.1
11.4
17.1
Detention
Timo
minutes
26
21
13
8.7
Similar calculations for the 3.7 m (12 ft)
diameter swirl unit used in pilot scale testing
at Toronto are:
Flow Rate
cfs
0.232
0.309
0.463
0.695
mgd
0.15
0.2
0.3
0.45
Detention Time Overflow Rate
m'/sec
0.00657
0.00876
0.013
0.02
minutes*
40
30
20
13
gpd/ft2
1,326
1,768
2,650
3.980
m3/day/mz
0.466
0.621
0.932
1.400
* allows 0.9 m (3 ft) for sludge accumulation
Description of Hydraulic Model
Figure 2 shows the swirl chamber and
inlet supply line; a solids hopper and vibrator
for introducing test solids into the swirl
device via the inlet sewer line; a clear overflow
settling basin equipped with a calibrated
V-notch weir for gauging discharge flows; a
foul outlet connection to the bottom of the
swirl chamber, leading to a foul overflow
settling tower and recovery screen for
recapturing the introduced solids; together
with necessary control equipment.
The central unit of the model system was
the swirl separation chamber. It consisted of a
vertical concrete conical hopper at the bottom
with a 60° side slope, topped by a vertical
cylinder chamber 91.5 crn (3 ft) in diameter,
made of 1.3 cm (0.5 in.) Plexiglas.® In the
-------
Foul outflow Foul solids
settling tower recovery screen
Chamber cylinder - 13 mm (0.5 in).
plexiglass - 914 mm (36 in. dia)
Clear outflow settling bosin
\
r I-K A \/
, /
'
Clear water overflow
outlet pipe - 102 mm (4 in) plexiglass
- Discharge returned
to pumping station
1
— ;=-- 1
f i /'
/ /
V
l_ J.
1 \
' / / / / /11
Clear outflow settling bosin ,
Small water supply
for solids injection
Water supply from
pumping station
Chamber cylinder - 13 mm (0.5 in)
plexiglass - 914 mm (36 in. dia)
Clear water overflow pipe -
102 mm (4 in) plexiglass
Foul solids
recovery screen
ELEVATION
Section A-A
FIGURE: 2 MODEL LAYOUT
-------
first phase of the model tests, the bottom
cone was adjusted to provide a 10 cm (4 in)
wide annular flat bench between the top of
the conical hopper and the bottom of the
swirl separator cylindrical chamber to serve as
a partial chamber floor as shown in Figure 3.
For subsequent series of tests, designed to
improve the characteristics of the swirl
pattern and the solids separation efficiency,
the concrete hopper cone was extended
upward to the periphery of the chamber wall
and the flat bench surface was eliminated, as
shown in Figure 4.
A Plexiglas skirt, supported by the
chamber wall and concentric to the chamber's
vertical axis, divided the chamber into two
concentric sectors — inner and outer. The
vertical distance from the bottom of the skirt
to the bottom of the cylindrical chamber, or
the top of the conical hopper represented the
inter-connecting slot between the two
concentric chamber sectors. The slot opening
was readily changeable by means of calibrated
supporting blocks.
A 15.2 cm (6 in ) PVC effluent discharge
downshaft installed concentrically around the
chamber's vertical axis also provided
structural support for the overflow weirs. The
height of the weir could be easily changed by
adding or removing custom-cut pieces of
downshaft.
The vertical downshaft was later removed
for subsequent studies to leave the inner
chamber unencumbered and the weir was
attached to and supported by the skirt. The
bottom of the conical hopper was closed and
the clear effluent overflowed through a
circular gutter fixed around the skirt in the
outer chamber, connected to an outside drain
with two 2.5 cm (1 in) diameter Tygon®
pipes which passed through the chamber wall
as shown by Figure 5.
The chamber inflow pipe was 10 cm (4
in.) diameter polyvinyl chloride (PVC) set at a
slope of 1/1000. A vibrating solids injection
system was placed on the inflow line 2.14 m
(9 ft) upstream of the chamber. Water was
supplied directly from the constant level tank
in one of the laboratory's permanent pumping
stations. This supply device was used as long
as the recovery of large size grain material
injected presented no problem. When smaller
grain sizes were used, a closed circuit device
was built. PVC pipe was removed and
replaced by a 76.2 cm (30 in) long Plexiglas
pipe while the solid injection system was
reduced to a mere hopper attached on the top
of the inlet duct as shown by Figure 5,
through which diluted material was
introduced into the model at a constant rate.
Outflow from the central pipe, and later
from the collecting annular gutter, went to a
large settling basin equipped with a calibrated
V-notch weir.
A point gauge in a manometer pot read
the level in the basin, determining the
discharge going over the V-notch weir. This
represented the total discharge passing
through the separation chamber.
The foregoing descriptions, supplemented
by the figures in Appendix A showing the
original structural format of the swirl
chamber and the subsequent changes
involving the elimination of the flat bench at
the top of the conical bottom hopper,
indicate that variances were made in the
model for the purpose of initiating added
studies of swirl patterns and performances.
These changes were only part of the many
structural modifications introduced into the
model, and the numerous test runs carried out
with each of such physical changes.
In all, twelve modifications were made,
involving the shaping of the conical hopper to
provide a bench flat floor between it and the
cylindrical swirl chamber, and to eliminate
this plateau; changes in the elevation of the
inlet sewer entering the swirl chamber; shapes
of the overflow weir; diameter of the skirt;
depth of the skirt; diameter of the weir; use
of the overflow weir as a sealed unit, with
concentric orifice ports for the upflow of
effluent from the bottom chamber to the
upper or discharge chamber; variations in slot
height under the skirt; elimination of the
• circular effluent weir and-replacement with
four radial weirs discharging into the central
overflow outlet downdraft pipe; variations in
the length of such radial overflow weirs;
installation of eight radial effluent weirs with
flat-surface crests; conversion of these eight
weirs into "saw-tooth" crested weirs; and
removal of the central outlet downdraft pipe
and discharge of effluent from the collecting
-------
0.91 m 0 Chamber
(36 in)
0.71 m 0 Skirt
(26 in)
0.61 m 0 Weir
(24 in)
320° FOUL OUTLET
PLAN
;i:::; /-First modification to inlet
'?'':' -Raised so crown at same
level as weir
XI 0.16 cm
I (4in)
Section A-A
FIGURES SWIRL PRIMARY SEPARATOR
FlRST LAYOUT TESTED ON MODEL
-------
270°
.0.9! m 0 Chamber
(36 in)
0.71m 0 SUirt
(28 in)
10.16x10.16 cm INLET
(4x4 in)
2.21 cm (0.87 in)
Former Chamber Floor
ELEVATION
FIGURE 4 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT FOR TESTS 161 to 184 - MODIFICATION 10
10
-------
Skirl
3cm (1.39 in) Wide
Circular Gutter
8 cm (3.13 in) High
Circular Gutter
6cm 0 Pip«
(2.34 in)
6 x 6 cm Inlet
(2.34 x 2.34 in)
Skirt 0.71 m 0
6 x 6 cm Inlet
(2.34 x 2.34 in)
6.35 cm (2.5 in)
T~
11.45 cm (4.5 in)
5.08 cm (2 in)
T
SECTION A-A
FIGURE 5a SWIRL PRIMARY SEPARATOR
MODEL LAYOUT FOR TESTS 188 to 194 (3 to 9)
MODIFICATION 11
11
-------
Overflow
0.91 m (36m)
0.71 m (28 in)
Skirt O.7lm 0
(28 in)
.Overflow
SECTION B-B
FIGURE 5b SWIRL PRIMARY SEPARATOR
MODEL LAYOUT FOR TESTS 188 to 194 (3 to 9)
MODIFICATION 11
weirs via a concentric channel connected by
outlet pipes through the swirl chamber wall.
Other internal details were involved in the
series of model modifications.
Each structural modification was
accompanied by a series of runs, all controlled
in terms of types of materials added to the
flow, rates of flow, removal efficiency studies,
tracing of velocity contours and :flow
patterns; determinations of settling velopities
vs. particle size for Petrothene® and IRA-93
anion exchange resin — the two basic
materials used to simulate sewage solids of
settleable and colloidal characterisitcs; effects
of the installation of a baffle deflector at the
inlet to the swirl chamber to induce greater
secondary flow action. A total of 194 test
cycles were involved in all of the structural
modifications and the analytical
investigations.
Definitive findings on model efficiencies
of solids removal from the influent flow
depended on the representativeness of the
simulated model solids. During the period
while the Beak studies3 were underway to
characterize actual sanitary sewage solids
components and to develop a synthetic
material to represent these material
characteristics in model scale, the LaSalle
studies were carried out with shredded
Petrothene, which was adequate to guide
basic determinations of chamber sizes and
12
-------
geometric internal shapes and dimensions.
When the results of the Beak studies became
available, a new material — an anion exchange
resin designated as IRA-93 manufactured by
Rohm and Haas — was utilized in the model
studies.
The Petrothene material had been
shredded and possessed a gradation curve
shown in Figure 6. Settling velocities for this
material were plotted vs. particle size, as
measured by Beak, and theoretical settling
velocities were deduced for material having a
specific gravity of 1.01. Subsequently, a
family of curves was devised for different
specific gravities for the purpose of
determining the particle size for each selected
specific gravity and ascertaining the
comparable settling velocity.
Before the use of the anion exchange
resin in the swirl studies was initiated, it was
found that relatively uniform solids recoveries
of from 95 to 100 percent of Petrothene were
being achieved. Thus, the Petrothene did not
offer enough range to differentiate between
the chamber effectiveness levels covering the
modifications to be studied, and it precluded
decisions on the swirl geometries that would
produce optimum clarification results.
The use of IRA-93 anion exchange resin
provided the flexibility and decision-making
characteristics desired in the studies. A
nominal specific gravity of 1.04 was given for
the new test material and the procedures used
for Petrothene in ascertaining settling
velocities vs. particle size were reproduced for
the more serviceable simulated solids, as
shown in Figure 7. The same procedure
described earlier for Petrothene was followed
in scaling up this curve to different scales to
define the settling velocities of particles
simulated by the IRA-93. The choice of solids
fractions followed the findings of the Beak
studies on representative sewage solids
characteristics and the recommendation for
simulated model particle sizes.
The approach used in determining the
actions of various particle sizes was adequate
for each such characteristic, but it did not
provide information on the distribution of
grain sizes in the simulated loading on the
swirl model. Subsequent settling column
studies by Beak on representative sewage
samples at Philadelphia, Pennsylvania,
precisely defined the recommended fraction
of the crushed IRA-93 material as passing a
No. 100 sieve and retained on a No. 200 sieve.
A model curve of settling velocities vs.
percentages less than or equal to the presence
of various particle sizes was, evolved for
sanitary sewage by the Beak studies and a
family of curves covering various particle sizes
was developed therefrom, at scales transposed
according to Froude's Law. These data
permitted calculation of the percentages of
sanitary sewage in the differently scaled
prototypes that were simulated by the
IRA-93 used in the model.
Out of these scalings, Figure 8 was
developed. This figure is the key in finding
the portions of the prototype sewage
simulated by the model material; it was used
in the final prediction of prototype recovery
rates. While the IRA-93 did not simulate as
complete a range of prototype sewage
particles as would be desired, it was
recognized that sewage composition is highly
variable from location to location, and even
from time to time in the same system.
Operation of the swirl concentrator as a
primary separator facility to replace
conventional sedimentation tanks would
normally involve a continuous steady state
discharge — the design discharge. Since it was
the purpose of the model studies to adjust the
chamber dimensions and internal appurtenant
facilities to the flowrate to be treated, the
range of the study was extended to cover the
three different flows previously described
(0.5, 1.0, 1.5 I/sec). This allowed observations
on the behavior of the separation chamber
and the variation of the solids recovery rate
when conditions of operation were changed.
The three flowrates demonstrated the
combined influence of the inlet and slot
velocities on the solids separation process.
The studies covered, in addition, evaluations
of the impact of different combinations of
weir and skirt dimensions and locations on
settling efficiencies.
For each individual test in the overall
series, steady-state flow was established and
equilibrium attained. One 1 (17 oz) of wet
Petrothene or IRA-93 was injected into the
incoming flow over a period of five minutes
13
-------
U.S. STANDARD SIEVE NUMBERS
34 6 8 IO 16 2O 3O 4O SO 7O IOO I4O
\ \
ill
64 2 1 0.6 O.4 0.2 0.1
(0.234) (0.156) (0.078) (0.04) (0.02) (0.015) (0.008) (0.004)
IOO
•90
•80
70
•60 *-
50
CD
•40
•30
20
•IO
FINE
GRAVEL
-
COARSE
| MEDIUM
FINE
SAND
FIGURE 6 GRADATION CURVE FOR PETROTHENE USED \H MODEL
14
-------
10 (.33)
5 (.16)
o
0)
.8
u
o
O
O)
c
1.0 (.03)
0.5 (.02)
0.1 (.003)
0.05 (.002)
0.01 (.0004)
W-
•f-—
JQ_
«
f
I
l-A
S.G. 1.06
0.1
(.0004)
1
(.004)
Particle Size, mm (in)
1.05 S.G.
1.05 S.G.
1.04 S.G.
10
(.04)
FIGURE 7 SETTLING VELOCITY VS PARTICLE SIZE FOR IRA-93
ANION EXCHANGE RESIN
15
-------
o
0}
s
in
I
1.0 (7.4)-j
0.9 (0.35)-
0.8 (0.31)
0.7 (0.3)-
0.6 (0.23)-
0.5 (0.2)-
0.4 (0.16)-
0.3 (0.12).
0.2 (0.078)-
0>
OJ
c
d > 74/r
FIGURES SWIRL PRIMARY SEPARATOR
SANITARY SEWAGE IN PROTOTYPES REPRESENTED BY
IRA-93 IN MODEL
and the model was allowed to run ten minutes
after the end of the solids application period.
The amounts of the applied solids found in
the bottom of the hopper cone and the
floating material in the outer sector of; the
swirl unit were measured separately. The
material floating in the settling basin was ;also
measured.
The remaining portion deposited in, the
settling basin was determined by subtraction,
based on the assumption that no material; was
lost. The swirl solids separation efficiency;was
taken as the percentage represented by the
amount measured on the bottom of the cone
as compared to the total found in the cone
and on the bottom of the settling basin.
The details of the removal efficiencies
achieved with shredded polythene, Petrothene
X® and anion exchange resin IRA-93 under
varying conditions in the 12 modifications of
the swirl chamber outlined above are
contained in Appendix A. The findings
highlights are listed here:
• Using shredded Petrothene in the
model with circular weir, the recovery rate
was 84 percent at a rate of 1 I/sec (11.38
16
-------
mgd) and a 1.2 cm (0.5 in.) slot height.
However, with the inlet sewer at the same
level as the outer chamber floor, undesirable
turbulences were experienced under the weir
in both the inner and outer chambers.
• With the crown of the inlet line raised
to the level of the weir lip, the recovery rate
remained at approximately the same — 78
percent — but it was demonstrated that the
turbulence was mainly due to the location of
the bottom of the skirt just above the
horizontal floor. Recoveries improved as
flowrates decreased and detention time
increased. Slot height tests were inconclusive.
Removal of the skirt resulted in marked
reduction of recovery.
• Using ground Petrothene, solids
removals paralleled those experienced with
shredded Petrothene. High recoveries were
produced at 1.0 and 1.5 I/sec (11.38 and
17.07 mgd) rates but they deteriorated when
the skirt was removed. At high flows of 1.5
I/sec (17.07 mgd), recovery started at 50
percent but increased progressively as the slot
height was increased.
• With a 16.9 cm (6.68 in.) diameter
weir, a fairly constant recovery rate of 98
percent was obtained for a flowrate of 0.5
I/sec (5.69 mgd). Recovery rates dropped as
the slot height increased and very low
performance was recorded for higher slot
openings.
• Using anion exchange resin IRA-93 as
representative of the fine-grain recovery that
would simulate in the model the actual
sewage conditions to be encountered in a
prototype installation, recovery efficiencies
were determined for the successive
modifications of the model. Because of
changed flow patterns and other factors
induced by these changes, solids recoveries
ranged from 91 to 81 percent at rates of 0.1
I/sec (1.14 mgd); from 65.3 to 44 percent at
rates of 0.3 I/sec (3.42 mgd); from 52 to 23
percent for flows at 0.5 I/sec (5.69 mgd);
from 68.8 to 23 percent for flows at 0.75
I/sec (8.54 mgd); and from 33.7 to 27.5
percent at rates of 1.0 I/sec (11.38 mgd).
The predicted solids recovery
performances in prototype units, based on the
final use of IRA-93 anion exchange resin was
in a sense, the culmination of all the studies.
The corresponding solids removal efficiencies
for the five model flowrates used in .the
various tests are tabulated in Table 1.
The median settling velocity for sanitary
sewage solids was selected as 0.054 cm/sec
(0.00177 ft/sec); this was assumed to be the
representative point for all sewage to be
handled in any prototype. Curves were
deduced that would enable a designer to
determine removal efficiencies for a given
sewage flow as a function of the diameter of
the main chamber.
MATHEMATICAL MODEL STUDY
The purpose of the mathematical model
study of the swirl primary separator was to
validate and verify the findings of the
hydraulic model investigation. The
applicability of the swirl flow patterns for the
purpose of separating solids and liquid phases
had been explored previously1 and verified as
a feasible, workable hydraulic principle for
use in combined sewer overflow-regulator
conditions and for selective removal of
TABLE 1
SOLIDS REMOVAL EFFICIENCIES OF IRA-93
AT VARIOUS DISCHARGE RATES
Model Discharge:
I/sec 0.1 0.3 0.5 0.75 1.0
cfs 0.0035 0.0106 0.0176 0.0264 0.0352
Recovery Rate:
Percent 91 58 41 31 27
17
-------
inorganic grit from organic, lighter solids
materials in sanitary and combined sewage.
The approach to mathematical modeling
followed the procedures used in the previous
study of combined sewer overflow regulation.
The flows within the chamber were assumed
to be axisymmetric about the vertical axis. The
chamber was then overlaid with a
computational grid and the flowfield
velocities were computed at each grid point
by solving the liquid continuity equation and
equations of motions. An eddy viscosity
condition was utilized to represent the
turbulent shear stress. Plots of streamlines and
velocity profiles were prepared to portray the
liquid flowfield as predicted by the
mathematical model.
Particle paths were calculated by
superimposing particle settling velocities on
the liquid flowfield. The studies demonstrated
that a simplified solution could be used to
predict a theoretical upper limit on solids
removal efficiencies in the swirl chamber.
Good agreement was found between the
theoretical upper limit of solids recovery and
actual removal efficiencies observed in the
hydraulic model studies.
The mathematical computer study,
substantiated by the hydraulic model
findings, provided data upon which to base
predicted solids removal efficiencies for
prototype installations of swirl settling
facilities. The mathematical model work
emphasized the importance of exact
information on sewage solids settling
properties before definitive chamber design
can be undertaken.
In mathematical evaluation of the
flowfield, a simplified configuration of the
swirl settler unit was utilized, as shown in
Figure 9. The region between the skirt and
the outer wall of the chamber was not
included in the mathematical model because
most of the particle settling occurs in the
main body of the chamber. This simplified
the specification of the boundary conditions
for the mathematical-computer studies. The
liquid flow was assumed to axisymmetric,
and the flowfield was assumed to be identical
at every radial cross section and independent
of the angular position.
The assumption ,of axisymmetry is
appropriate because the flow enters the main
portion of the chamber at all points along the
circumference of th'e skirt. As a consequence,
the mathematical representation of the liquid
flowfield was very close to the actual
conditions in the hydraulic model. The
eddy-viscosity mixing-length constant and the
skin friction coefficient were assumed to be
equal to the values obtained in the previous
mathematical model studies for combined
sewer overflow swirl treatment. A closer
agreement between the velocity flowfield
predicted by the mathematical model and the
hydraulic model could be obtained by
refining these values but particle removal
efficiencies would be relatively unchanged.
Actual prediction of solids removal
efficiencies in any prototype unit, based on
mathematical model computations, would
require firm information on the properties of
actual sewage and the simulated sewage solids
used in the hydraulic model. This information
was provided by the Beak test program on
column settling characteristics. These column
test data have been utilized to determine a
frequency distribution of the particle settling
velocities which can be used to compare
settling properties of different materials.
Three types of material were considered in
the hydraulic and mathematical model
studies: Shredded Petrothene, IRA-93 anion
exchange resin, and so-called Arizona Road
Dust. The latter material simulates the fines
and colloidal material in actual sewage but
was too costly and unavailable in adequate
quantities for extended laboratory studies.
The mathematical model studies
examined the phenomena of solids settling in
swirl chamber conditions, covering
single-settling velocity distribution without
any flocculation reactions; hindered
settlement in which solids impinge upon each
other and thus inhibit simple gravity
subsidence; and flocculation or agglomeration
mechanisms which result in the coalescence of
solids and enhancement of settling
characteristics. The agglomeration
mechanisms involve the forces of gravity,
shear flow, turbulent acceleration, turbulent
entrainment, and Brownian motion.
18
-------
OVERFLOW WEIR
STANDPIPE'
1
1
COM
,
PI
4
OVERFLOW VELOCITY
PROFILE
CIRCULAR SKIRT
FOUL OUTLET
(WHERE APPLICABLE)
PLAN
ENTRANCE
VELOCITY
PROFILE
FIGURE 9 DIAGRAM OF SWIRL PRIMARY SEPARATOR CHAMBER
AS REPRESENTED BY MATHEMATICAL MODEL
19
-------
From settling column tests carried out by
Beak Consultants, it was concluded that test
solids settle as discrete particles while actual
sewage can exhibit flocculation properties
even at low concentrations. The mathematical
model studies involved detailed analyses of
the kinetics of flocculation. It was found that
the degree of flocculation or agglomeration is
dependent on particle collisions and estimates
were made of the number of collisions which
result from the mechanisms of gravity action,
shear action, turbulent acceleration, turbulent
entrainment, and Brownian motion. The
turbulent entrainment mechanism was found
to be 'the most important in the outer .annulus
section of the swirl separator where most of
the inlet energy is dissipated.,
It was found numerically impractical to
carry out exact mathematical modeling of the
flocculation effects on particle settling rates.
Therefore, to accommodate the flocculation
phenomenon in actual sewage a settling
velocity distribution was assumed, indicative
of sewage after flocculation in the inlet sewer
and outer annulus of the swirl chamber. This
condition will produce a settling velocity
distribution equivalent to that obtained by
the gravity mechanism in the Beak settling
column test procedures. This was deemed to
be a conservative assumption.
The full report on the mathematical
model studies is included as Appendix B. It
describes in detail the intricate methodologies
utilized in developing calculations of particle
flow, in terms of particle paths; boundary
conditions; numerical methods; scaling
techniques of the liquid flowfield; and scaling
of the particle flows. In order to maintain
similar liquid flows in different size swirl
separation facilities it is necessary to Use
Froude scaling, relating liquid flow velocities
by the square root of the scale factor and
flowrates by the 5/2 power of the scale
factor. A simplified equation was evolved for
obtaining a first estimate of remqval
efficiency for a particle having a specified
settling velocity, given the size and design
flowrate of any prototype unit. A more
accurate estimate of the removal efficiency
can be obtained by scaling back the prototype
flowrate and particle settling velocity to the
laboratory model.
The mathematical model of the liquid
flowfield was varied concurrently with the
various modifications made in the hydraulic
laboratory model to achieve improved solids
removal efficiencies. Quantitative
comparisons showed close concurrence
between hydraulic and mathematical findings.
Tangential velocity contours and streamlines
for modified laboratory configurations were
mathematically plotted. Removal efficiencies
were projected mathematically for what was
assumed to be the most typical for the
materials used in the hydraulic model tests.
The mathematical model predicted a 33.2
/ percent removal as the maximum limit on
swirl chamber performance at a 0.5 I/sec
(5.69 mgd) flow for IRA-93 resin having a
size range from 100 (0.149 mm or 0.00049
in.) to 100 mesh (0.074 mm or 0.00024 in.).
The actual IRA-93 resin solids removal
observed in the hydraulic model was in the
general range of 50 percent.
Initially four possible explanations were
offered for this apparent discrepancy:
swelling of resin material on contact with
water, with consequent variations in settling
velocity; electrostatic attractions of particles
to the chamber walls or to other particles;
failure of particles to completely disperse as
single particles; and stratification of solids in
the region under the skirt, producing higher
concentrations along particle paths and
improvement in settling. Subsequent
investigations ruled out the first three
phenomena but the stratification of solids
material under the skirt was observed in the
hydraulic model.
Appendix B also describes the effect of
scaling on chamber performance and the
influence of geometric variables on the
efficiency of solids removal. The complexity
of the data required to use the mathematical
model makes it generally impractical to use
for general design. The appendix describing
the model has been included to assist future
research, work and to provide the theoretical
basis for what was observed in the hydraulic
laboratory.
20
-------
THE PILOT STUDY
In August, 1974, the Department of
Public Works (DPW) of Metropolitan Toronto
constructed a 3.7 m (12 ft) diameter pilot
unit. The. unit followed the recommended
design with two exceptions: a mechanical
scum rake and the use of a slender central
support column. The unit was fabricated of
steel and installed by DPW employees at the
Humber Wastewater Treatment Plant.
The Humber plant treats combined sewer
flow from a large industrial district. The plant
has experienced severe overloading problems
and during the period of the testing, a major
plant expansion was being undertaken. The
total flow to the plant averages about
227,400 m3/day (60 mgd) and can more than
double during precipitation events.
Flow to the swirl concentrator was
pumped from the influent flume to the
primary tanks through a Parshall flume. Flow
was varied to follow the rate of flow to the
plant in the May-July tests but was kept
constant in all subsequent tests.
Figure 10 contains photographs of the
facility.
The influent flume leads to six primary
tanks. Each primary tank is 10.4 m (34 ft)
wide and 72.7 m (237 ft) long, with average
depth of 3.2 m (10.6 ft). Total volume of the
primary tanks is 14,540 m3 (513,400 ft3).
Total surface area is 449 m2 (4,833.2 ft2).
The details of the swirl separator are as
follows: diameter is 3.7 m (12 ft); distance
from overflow weir to bottom of straight
sides is 0.53 m (1.75ft); the 60-degree
cone-shaped bottom is 2.9 m (915 ft) deep.
The surface area of the tank is 10.5 m2 (113
ft2). The volume of the tank below the weir
elevation is 16.3 m3 (577 ft3J. If the bottom
1.2 m (4 ft) of the cone, with volume of 0.9
m3 (32 ft3), is designated as sludge storage
then the net volume of the cone is 15.9 m3
(560 ft3). Prior to testing of the unit, detailed
testing was conducted to characterize the
flow and the flow rate. Beak Consultants
Ltd., conducted tests to determine the
settling rate characteristic of the solids and
compared the results to the IRA-93 particles
used in the hydraulic model study. Beak also
studied the flow characteristics within the
swirl unit and concluded that plug flow
conditions prevailed.
Initial tests were conducted from April
29 to June 12 with disappointing results.
Analyses of the data indicated that sludge was
not being drawn off at frequent enough
intervals.
A second series of tests was conducted at
the design capacity of 1,137 m3/day (0.3
mgd) and 1,700 m3/day (0.45 mgd) from
June 23 to July 8. This corresponds to
overflow rates of 108 m3/day m2 (2,650
gal/day/ft2) and 162 m3/day/m2 (3,980
gal/day/ft2), and detention times of 20
minutes and 13 minutes, respectively.
The tests were run during week days for
four hours a day. Tests were started at 11:00
a.m. in order to allow a minimum of three
hours after wasted activated sludge had been
returned to the system. Samples were taken
each one-half hour and hourly composited
samples were prepared for analysis. Both
settleable and suspended solids were
determined.
A final series of tests were run from
September 2 to 15 with slightly revised
procedures.
The results of the test program are
presented in Section 4, Test of a Pilot
Unit, and the Beak study results are contained
in Section 8, Appendix C.
21
-------
FIGURE 10 PILOT FACILITY - METRO TORONTO
Sludge Drawoff
Collection and Valve
Interior View of Internal
Support for Wires
22
-------
SECTIONS
DESIGN GUIDELINES AND CONSTRUCTION COSTS
Conventional primary sedimentation
tanks are generally designed on the basis of
overflow rate and, to a lesser extent, on
detention time. The term "overflow rate" or
"surface settling velocity" is the unit volume
of flow per unit of time divided by the unit of
tank area. In U.S. customary units this is
expressed as gallons per day per square foot
(gal/day/ft2) and in metric units' may be
expressed as cubic meters per day per square
meter (m3/day/m2). The American Society of
Civil Engineers (ASCE) Manual of Engineering
Practice Number 364 lists data on various
primary settling tanks which indicate removal
of suspended solids ranging from 20 to 80
percent. Figure 6 of that publication indicates
the relation between removal of suspended
solids and overflow rate. Many tanks fall in
the range of 60 to 70 percent removal of
suspended solids. If we accept 60 percent
removal of suspended solids as a desirable
objective then Figure 6 indicates the
necessary overflow rate is 36.67 m3/day/m2
(900 gal/day/ft2). The ASCE manual's curve
in this range of suspended solids removal has
been verified by more analyses of field data
by Smith.5 Detention time is no longer
considered as the only factor in design of
primary settling tanks. However, the use of
tanks with liquid depths of 2.13 to 3.66 m (7
to 12 ft) combined with accepted overflow
rates will result in nominal detention times of
1 to 2 hours. For instance, the use of a 3.05
m (10 ft) liquid depth with an overflow rate
of 36.67 m3/day/m2 (900 gal/day/ft2) will
result in a detention time of 2 hours.
The equation developed by Smith5 from
the analyses of field data can be used to
estimate the removal efficiency (percent) of
suspended solids, 17 as a function of overflow
rate (OVFRA) in gpd/ft2 as follows:
7? = 0.82e -OVFRA/2,780
For OVFRA = 36,668 1/d/m2 (900 gpd/ft2),
1? = 59.3 percent. This value is in reasonable
agreement with the 60 percent removal
estimated by the use of the ASCE figure for a
36,668 1/d/m2 (900 gpd/ft2) overflow rate.
The swirl separator was not found
suitable for use where 60 percent removal of
suspended solids is the desired objective due
to the size and resulting cost of the unit
compared to conventional units. It appears
that after the heavier, more easily handled
solids are removed from the flow field, only
quiescent settling will accomplish the balance
of the solids removal. From the tests which
were conducted, the low efficiency of solids
removal and long detention times make large
units impractical with the present design.
Table 2 gives the flow and detention times for
several size units as taken from Figures 11 and
12. The table indicates that the swirl
separator has less detention time than
conventional settling tanks over a small range
of flows. At a diameter of 5.5 m (18 ft) the
detention time necessary to achieve 40 to 50
percent suspended solids removal is
approximately that of conventional units.
The design of the swirl separator is based
neither on overflow rate nor detention time,
but on the results of model tests. However,
these two parameters are useful in comparing
the size of the swirl separator with a
conventional tank.
DESIGN PROCEDURE
Figures 11, 12, and 13 are used for
design.
As indicated, the swirl separator cannot
economically achieve conventional suspended
solids removal of 60 percent. Therefore, the
following design example is based upon 45
percent suspended solids removal, which is
near the upper level of its efficiency for san-
itary flow. Settling characteristics of com-
bined sewer overflow solids are usually better
than for sanitary sewage.
23
-------
§
2L
V)
tr
in
o
Ul
cc
IOO-*
90-
80-
0.5
0.02
30 40 50
1
DISCHARGE - I/sec
1 1 — L — i i i i i i i
• .1 . . . . 1
J_
0.05 O.I 0.2 0.3 0.40.5
DISCHARGE- mgd
1
' «
0.05^ O.I 0.15 02 O.3 0.* 0.5
DISCHARGE - cfs
1 - LJ I I
i.o
J- L
2.0
FIGURE 11a PREDICTED PROTOTYPE SOLIDS REMOVAL EFFICIENCY
FOR SANITARY SEWAGE
24
-------
100-
90-
J I
2'0 3O 40 50 100
DISCHARGE - I/sec
i I . i . . I
200 300 400 500
I .... I
0.5
1.0
2.0
I
DISCHARGE - Mgd
I . I I
3.0 4.0 5.0
i I i I il
10
0.2 0.3 0.4 0.5
I 2.0 3.0 4.0 5.0
DISCHARGE - Cfs
10 15
FIGURE 11b PREDICTED PROTOTYPE SOLIDS REMOVAL EFFICIENCY
FOR SANITARY SEWAGE
25
-------
00
Ul
Z
i
I
Ul
S
H
2
O
z
ui
ui
Q
34 567 9,10
100
500' I/sec
1 1 1
0.05 0.1
iii i
til I ii
0,5 1.0
i i i i I i
• i I
5 10
, , , j^
0.05 0.1
FLOW RATE
cfs
10 mgd
FIGURE 12 DETENTION TIMES
26
-------
FIGURE 13a GENERAL DESIGN DIMENSIONS
27
-------
Overflow
outlet -*—
}
H7
1
•Skirt
SECTION B-B
D = Diameter of Chamber (from Figure)
D, = 0.066 D INLET
D2
D4
D5
D6
EI
E2
H
= 0.67 D
- 0.58 D
= 0.056 D
= 0.042 D
= 0.028 D
= 0.028 D
= 0.056 D
SKIRT
GUTTER
SLUDGE
DRAW
OFF
OUTLET
WEIR GUTTER
WIDTH
SLOT WIDTH
SLOT HEIGHT
Note: Elevations are referred to Top of Cone.
H2 = 0.07 D
H3 = 0.125 D
H4 = 0.2 D
H5 = 0.04 D
H6 = 0.04 D
H7 = 0.19 D
Hg =0.8 D
Overflow
-»-outlet
INVERT ELEV.
CIRCULAR
GUTTER
HEIGHT
CIRCULAR
GUTTER
TOP ELEV.
GUTTER TOP
ABOVE WEIR
LIP
WEIR GUTTER
DEPTH
GUTTER
DEPTH AT
OUTLET
CONE HEIGHT
FIGURE 13b GENERAL DESIGN DIMENSIONS
28
-------
Normal practice is to provide a minimum
of two plant units of each type in a plant.
Thus the initial construction phase would
include at least two primary settling tanks.
From Table 2 it is obvious that if the
detention time is to be less than that of a
conventional unit, the diameter will be less
than 5.5 m (18 ft). Thus, for 40 percent
suspended solids removal, the maximum flow
would be 10 I/sec (0.22 mgd). Since in
conventional practice two tanks are used, the
maximum plant design capacity would be 20
I/sec (0.44 mgd) or less.
The design of a swirl primary separator
follows:
1. Plant design average daily flow is 15
I/sec (0.34 mgd)
2. Removal efficiency of suspended
solids desired is 45 percent
3. Use two swirl primary separators.
Design flowrate per unit is 7.5 I/sec
{0.17 mgd). Peak flowrate is 11.2
I/sec (0.26 mgd).
4. Enter Figure 1 la or 1 Ib with design
flowrate. For 45 percent efficiency,
select 77 = 3.7 m (12 ft). Surface area
is 10.5 m2 (113 ft2). Overflow rate is
61,295 1/day/m2 (1,505 gal/day/ft2).
5. Enter Figure 12 with design flowrate
of 7.5 I/sec (0.17 mgd) and D of 3.7
m (12 ft). Detention time is 37
minutes.
Note: For conventional settling units, the
detention time would be 51 to 63 minutes.
6. Enter Figure 1 la with peak flow of
11.2 I/sec (0.26 mgd) and D of 3.7 m
(12 ft). Read recovery is 38 percent.
7. Enter Figure 12 with peak flow of
11.2 I/sec (0.26 mgd) and D of 3.7 m
(12 ft). Read detention time is 25
minutes.
8. Determine dimensions of structure
from Figure 13, as follows:
D =3.7m(12ft) inside diameter of tank
D! = 0.24 m (0.8 ft) inlet (side of square)
D2 = 2.4 m (8 ft) skirt diameter
£>4 = 2.1 m (7 ft) gutter diameter
From the values given for D2 and Z)4 the
circular gutter width is 0.3 m (1 ft). D4 does
not appear to be a critical dimension insofar
as the tank performance is concerned and
therefore we assume the gutter width could
TABLE 2
COMPARISON OF DIAMETER, DETENTION TIME, AND
SUSPENDED SOLIDS REMOVAL FOR SWIRL PRIMARY SEPARATOR AND DETENTION
TIME FOR CONVENTIONAL SETTLING FOR VARIOUS OVERFLOW RATES
Swirl %S.S. Removal
30 40 50 60
Diameter
m
1.8
3.6
5.5
6.1
(ft)
I/sec
6 4.5
12 15
18, 27
20 28
Flow
(mgd)
0.1
0.34
0.60
0.62
Detention
Time
(min)
Flow
I/sec (mgd)
8 2.8 0.06
19 9.8 0.22
30. 15 0.33
35
Detention
Time
(min)
Flow
I/sec
13 2
30 6.5
54 10
(mgd)
0.05
0.15
0.22
Detention
Time
(min)
Flow
I/sec, (mgd)
18 1.6 0.04
45
75
Detention
Time
(min)
24
CONVENTIONAL SETTLING TANKS
% S.S. Removal
Overflow Rate
l/day/m2
Detention Time
(min)
gal/day/ft2 (10 ft depth or over)
60
50
40
30
36,653
57,017
51,453
114,034
900
1,400
2,000
2,800
120
77
54
38
29
-------
Table 2 (continued)
OVERFLOW RATE COMPARISON FOR SWIRL SEPARATOR
Flow Overflow Rate
I/sec mgd l/day/m2 gal/day/ft2
Diameter 1.8
Diameter 3.6
Diameter 5.5
Diameter 6.1
.m (6 ft)
4.5
2.8
2
1.6
m(12ft)
15"
9.8
6.5
m(18ft)
27
15
10
m (20; ft)
28
0.1
0.06
0.05
0.04
0.34
0.22
6.15
0.6
0.33
0.22
0.62
144,170
86,340
72,085
57,625
122,585
79,415
54,165
96,110
52,945
35,230
80,435
3,540
2/120
1,770
1,415
3,010
1,950
1,330
2,360
. 1,300
865
1,975
be changed if greater width is necessary to
carry off the weir discharge.
Ds is not a critical dimension. Suggest Ds
= 0.2 m (0.67 ft).
D6 is not a critical dimension and
designer may select size depending on
hydraulics. Suggest D6 = 0.2 m (0.67 ft).
HI = 0.2 m (0.67 ft) slot height.
H2 = 0.25 m (0.84 ft) vertical distance
from invert to junction of tank slope and tank
side.
H3 = 0.45 m (1.5 ft) height of circular
gutter.
HH = 0.73 m (2.4 ft) vertical distance
from top of circular gutter to junction of tank
slope and tank side.
Hs = 0.15 m (0.48 ft) vertical distance
from circular gutter top to overflow weir.
H6 = 0.15 m (0.48 ft) depth of weir
gutter. Designer should check to make sure,
this depth is adequate. ;
HI = 0.69 m (2.28 ft) vertical distance;
from gutter top to invert of outlet pipe.
Hs= 2.92 m (9.6 ft) depth of chamber
with sloping sides. The horizontal dimensions
of sludge hopper bottoms are usually no
larger than 0.61 m (2.0 ft). If the bottom is
given this width then H8 will be reduced by
0.53 m (1.73 ft). Hence Hs = 2.4, m (7.9 ft).
EI = 0.1 m (0.33 ft) weir gutter-width.
E2 = 0.1 m (0.33 ft) slot width at right
angles to slope.
The size of the resultant structure for an
average design flow of 7.5 I/sec (0.17 mgd) is
shown in Figure 14.
The design and size of the overflow weirs
and effluent gutters should be based on
principles used in conventional primary tanks
and should be revised from the values derived
from Figure 13 as required.
CONSTRUCTION COSTS
A conventional round tank designed to
handle the same flow and at the same
suspended removal efficiency 7.5 I/sec (0.17
mgd), 45 percent suspended solids, would
have essentially the same diameter, but less
depth.
Cost estimates of the swirl primary
separator were made for two purposes: 1) to
indicate the probable construction cost of the
facility; and 2) to compare its costs with that
of a conventional primary settling tank-
designed for the same efficiency.
30
-------
CAPACITY
7.5l/sec<0.17mgd)
Inlet 0.25m (10 in.)
NOTE: Provide surface skimming
device for floatables
SECTION A-A
FIGURE 14 SWIRL PRIMARY SEPARATOR
31
-------
The cost estimates are considered to be
reasonable engineer's estimates. However,
during periods of economic inflation, it is not
unusual for contractors' bids to materially
exceed engineers' estimates.
Cost Basis
The costs are based on the following::
a. Engineering News Record
Construction Cost Index average for
U.S. is 2,100.
b. Unit prices as follows:
Steel Sheet Piling $86/m2 $ 8/ft2
(for temporary use during construction)
Excavation $16/m3 $ 12/yd3
Reinforced concrete (swirl). $ 195/m3 $ 150/yd3
Reinforced concrete (conventional
$326/m3 $250/yd3
Note: The concrete for the swirl unit will
require less reinforcing steel, thus a lower
cost. i
c. Contingency and engineering costs 25
percent of the foregoing items. ;
The estimated quantities of materials are
based on the dimensions shown in Figures 14
and 15.
The swirl separator dimensions are
derived in the previous section. It is assumed
that the ground surface is 0.6 m (2 ft) above
the crown of the inlet pipe and the top of
tank is 0.3 m (1 ft) above ground surface.
Since the top of overflow weir is 0.2 m (0.7
ft) above crown of inlet pipe, this provides
0.7 m (2.3 ft) of freeboard above the weir.
The conventional primary settling tank
dimensions are inside diameter of 3.6 m (12
ft) and side water depth of 2.44 m (8 ft).
These dimensions provide an overflow rate of
61,260 1/day/m2 (1,500 gal/day/ft2) and a
detention time of. 57 minutes. The tank is set
to provide a freeboard of 0.7 m (2.3 ft) with
top of wall 0.3 m (1 ft) above ground surface.
The following assumptions are made for
both structures:
a. Excavation is all in earth. The unit
price includes cost of backfilling.
b. Temporary steel sheet piling is
required 0.61 m (2 ft) outside
exterior walls of structure.
c. Equipment cost for conventional
settling tank includes cost of
rake-type sludge collector with fixed
bridge and center drive, scum
collector, weir plates, telescopic
valve, and electrical work.
d. Miscellaneous costs for swirl
separator includes cost of skirt, weirs,
gutters, telescopic valve, center
support for weir gutters, piping, and
railing around tank.
e. Miscellaneous costs for conventional
settling tank includes piping within
limits of structure, gratings, and
railing around periphery of tank.
Cost of Swirl Separator as
a Primary Separator
The estimated construction cost of a swirl
separator with a capacity of 7.5 I/sec (0.17
mgd) is $55,250. The breakdown of this cost
is shown in Table 3.
TABLE 3
CONSTRUCTION COST OF SWIRL
PRIMARY SEPARATOR
Capacity 7.5 I/sec (0.17 mgd)
Item Quantity Amount
Sheet Piling 128m2 $ 11,000
(1,380ft2)
Excavation 340m3 5,450
(440 yd3)
Reinforced Concrete 125m3 18,750
(160yd3)
Miscellaneous Costs Job 9,000
Subtotal
Contingents
Engineering Costs
25%+
Total"
$ 44,200
11,050
$ 55,250
Cost of Conventional Primary Settling Tank
The estimated construction costs of a
conventional primary settling tank with a
capacity of 7.5 I/sec (0.17 mgd), based on the
dimensions shown in Table 4, is $75,940. The
breakdown of this cost is also shown in Table
4. . .
As the capacity of the swirl unit
increases, there is a rapid increase in cost as
compared to the cost of conventional units,
due mostly to the increased excavation
sheeting and amount of reinforced concrete.
Similar construction calculations were made
for comparable units having a capacity of
32
-------
CAPACITY
7.5 I/sec (0.17 mgd)
Inlet
0.3 m (1
\
I
\
\
\
j~ir:
f
/ |
(}
Outlet
0.3m
, 9
i
(1ft)
Telescopic
Valve
PLAN
Railing
0.64m (1.3 ft)
0.31 m (1 ft)
NOTE: Drawing not to scale
0.31 m (1 ft)
SECTION A-A
FIGURE 15 CONVENTIONAL PRIMARY SETTLING TANK
33
-------
TABLE 4
CONSTRUCTION COST
OF CONVENTIONAL
PRIMARY SETTLING TANK
Capacity 7.5 I/sec (0.17 mgd)
Item
Sheet Piling
Excavation
Reinforced Concrete
Equipment
Miscellaneous
Subtotal
Contingent &
Engineering Costs
.96m2
(1,050ft2)
345m2
(450yd3)
40m3
(51yd3)
Job
Job
25%+
Total'
$ 8,230
5,520
10,000
35,000
2,000
$60,750
15,190
$75,940
21.9 1/sec (0.5 mgd) with a suspended solids
efficiency of 65 percent. The construction
cost of the swirl unit was estimated to be
$55,250 and the conventional unit $75,-940.
Figure 16 is a plot of the cost comparisons
made.
300
250
§
ti200
100
50
Key
• -Conventional
• -Swirl
$55,250
12ft,
$138,000
38 ft, 65% ss
$75,940
20 ft 45% ss
3 6 9 12
(10) (20) (30) (40)
Diameter - m (ft)
FIGURE 16 COST VS DIAMETER, SWIRL
AND CONVENTIONAL PRIMARY
TREATMENT UNITS
Operating and maintenance costs for a
43.8 I/sec (1 mgd) unit, the smallest size for
which USEPA data is available, can be
estimated as shown in Table 5.
Comparison of Costs
From the foregoing it is seen that the
construction cost of the swirl separator will
be $55,250 compared to $75,940 for a
conventional settling tank. Annual O & M
costs may be $3,000 less with the swirl unit.
The surface area required for units of this low
volume does not appear to warrant a
comparison of land cost savings.
This comparison assumes that the two
structures will produce equal results in
removal of suspended solids in raw sewage.
The sizing of the conventional primary
settling tank is based on standard design
criteria. The sizing of the swirl primary
separator is based on model results in the
laboratory using IRA-93 resin as
representative of suspended solids in raw
sewage.
Cost comparison of large size units do not
appear warranted at this time. A different
configuration is obviously needed for large
units to avoid the adverse construction costs
of such a deep structure. A flat bottom with
scrapers as suggested by Smisson, sacrificing
the principal of no moving parts appears
reasonable.
PRESENT WORTH
The present worth of the swirl separator
units is shown in Table 6. The present worth
is based on a life of 20 years and an interest
rate of 6-1/8 percent. Hence the present
worth of the operation and maintenance costs
for a 20-year period is 11.35 times the annual
cost.
For the unit with capacity of 7.5 I/sec
(0.17 mgd) at 45 percent ss removal the
present worth of the conventional unit is
$117,500 and the swirl separator is $62,500.
Thus the present worth of the conventional
unit is 88 percent greater than that of the
swirl separator.
For the unit with capacity of 21.9 I/sec
(0.5 mgd) and 60 percent ss removal the
present worth of the conventional unit is
$179,500 compared to $282,200 for the swirl
separator. Thus the present worth of the
conventional unit is 57 percent less than that
of the swirl separator.
34
-------
TABLE 5
COMPARISON OF OPERATION AND MAINTENANCE COSTS
FOR PRIMARY TREATMENT UNITS
1. Labor: Operation, 1 hr/day @ $7/hr
Maintenance, 0.54 hr/day @ $7/hr
2. Materials and Supplies:
$300/yrx325 (Ratio 1975/1971 EPA
150.6 Cost Index)
3. Power: 2 pumps @ 1/2 hp, 1 hr/day $0.03/kwh
4. Annual Maintenance @ 3% of capital cost
Primary tank sludge collections
Raw sludge plunger pumps
Total Annual O&M
Conventional
$1,820
910
650
100
100
80
$3,660
Swirl
$455
100
80
$635
TABLE 6
PRESENT WORTH
SWIRL SEPARATOR PRIMARY
TREATMENT UNITS
Conventional Swirl
Tank Separator
Capacity 7.5 I/sec (0.17 mgd)
(45% ss removal)
Construction Cost $75,940 $55,250
Operation and
Maintenance Cost 41,540 7,200
Cost Total PresentWorth $117,480 $62,450
Capacity 269 I/sec (0.5 mgd)
60% ss removal
Construction Cost $138,000 $275,000
Operation and
Maintenance Cost 41,540 7,200
Cost Total PresentWorth $179,540 $282,200
35
-------
SECTION 4
PILOT TREATABILITY EVALUATIONS WITH SEWAGE
The opportunity to investigate the swirl
separation phenomenon in the field was made
available by the installation of a pilot unit at
the Humber Wastewater Treatment Plant of
the Municipality of Metropolitan Toronto,
Canada.
The findings described in this section are
based on conditions at the Toronto
installation and the comparison between the
swirl pilot performance and that of the
conventional settling tanks in use at the plant.
While specific conclusions have been drawn as
a result of the swirl pilot study they must be
characterized as applying only to these
conditions. No single field investigation of a
previously unevaluated treatment method —
whether it demonstrated highly favorable or
unfavorable performance — can be taken as
the final proof of applicability. Further
investigations, under different conditions and
for different applications and purposes, would
be desirable.
The Humber Wastewater Treatment Plant
has six rectangular-shaped primary settling
tanks, each with a width of 10.36 m (34 ft), a
length of 72.24 m (237 ft), and average depth
of 3.24 m (10.62 ft). The six tanks have total
surface area of 449 m2 (48,350 ft2) and a
volume of 14,540 m3 (513,400 ft3). The
plant, during the period of the 1975 tests, was
operating far in excess of design capacity.
Plant additions were under construction to
add treatment and sludge handling facilities.
Pilot Test Results :
The swirl separator was installed initially
near the grit removal chamber in September
1974. Several preliminary test runs were made
at this location on sewage pumped out of the
effluent end of the grit chamber.
These tests indicated removal of total
suspended solids ranging from 0 percent to
64.7 percent for various flows. No data were
available on test procedures for this
preliminary tryout run.
Following these preliminary test runs, the
swirl unit was moved to the influent end of
the primary settling tanks and .installed as.
shown in Figure 17. The initial evaluation was
made with two test runs, each of seven days
duration, with average flows of 13.1 I/sec
(0.30 mgd) and 19.7 I/sec (0.45 mgd),
respectively, in the swirl separator. The
overflow rates for these flows would be 0.932
and 1.4 m3/day/m2 (2,650 and 3,980 gal/
day/ft2) with corresponding detention times
of 20 and 13. minutes. The flow in the swirl
was varied in accordance with the diurnal flow
to the plant.
In order to gain information concerning
conditions during periods of stormflow, the
swirl separator testing was modified during
actual storm periods.
If rainfall occurred during the tests and
the plant flow exceeded the diurnal flow by
788 I/sec (18 mgd) the_swirl was operated at a
constant rate of 18.8 I/sec (0.43 mgd). The
detailed test schedule for dry-weather periods
is shown in Table 30, and the tests for
wet-weather periods in Table 31 of Appendix
D. Limited tests for heavy metals on influent
and effluent samples to the primary tank
during dry-weather flow periods were also
conducted to allow an evaluation of
treatability. These tests were scheduled on a
weekly basis.
The first series of tests was carried out
between April 29th and June 12, 1975. There
were eight separate storm events in this
period. The results for the swirl separator
with a flow of 18.8 I/sec (0.43 mgd) are given
in Tables 32 and 33 . Tables 34
and 35 show the comparable results for
the primary settling' tanks, including the .
test results for heavy metals. The heavy
metal testing indicates considerable concen
trations of the metals in the sludge with
greater uniformity of concentration in
the sludge than in the effluent.
For the eight storm events, the removal
of total suspended solids in the swirl separator
ranged from zero to 34.4 percent, with an
average of 10.4 percent. The removals for the
primary tanks ranged from zero to 79.7
percent, with an average of 32.8 percent.
These results are shown in Table 7.
36
-------
PILOT
SWIRL
PRIMARY SEPARATOR
PARSHALL
FLUME
EFFLUENT
STEEL DRUM
FOR SLUDGE
VARIABLE
SPEED PUMP
AERATED
INFLUENT FLUME
CONVENTIONAL
PRIMARY
TANKS
1
1 '
FIGURE 17
1 INFLUENT GRAB SAMPLE
2 SWIRL EFFLUENT GRAB SAMPLE
3 PRIMARY SETTLING TANK GRAB SAMPLE
TEST LAYOUT - HUMBER PLANT - TORONTO, CANADA 1975
TABLE 7
REMOVAL OF TOTAL SUSPENDED SOLIDS - WET-WEATHER FLOW,
MAY4-JULY 12, 1975
Percent Removal
Swirl Primary
Date Concentrator Tanks
Average Flow
May 4
May 5
May 6
May 7
May 15
May 31
June 4-5
June 12
Total
Average
18.8 I/sec
(0.43 mgd)1 -2
0
0
0
17.7
34.4
4.0
0
27.1
83.2
10.4
5,688 I/sec
(130 mgd)3
0
37.5
19.0
79.7
70.4
0
10.2
45.6
262.4
32.8
Constant Flow, overflow rate: 155 ,
gal/day/ft2); detention time: 14 minutes.
Sludge drawoff interval — 4 hours.
rr>3/day/m2
(3,805 Average for days during which storm occurred, overflow
rate: 110 m3/day/m2 (2,688 gal/day/ft2); detention time:
42 minutes.
37
-------
The diurnal flow of the Humber Plant,
based on hourly measurements taken in April
1975, is shown in Figure 18. The total flow to
the primary tanks is shown with and without
waste activated sludge for the period.
Waste activated sludge is normally
pumped to the primary tank influent only
during the night hours but some waste
activated sludge may return to the primary
tanks by gravity during the daytime hours.
Figure 18 also shows the variation in total
suspended solids and volatile suspended
solids. The concentration of total and volatile
suspended solids is highest between the hours
of 2:00 a.m. and 8:00 a.m. when flow is
lowest due to the return of waste activated
sludge to the plant influent. The variation in
suspended solids during the daytime hours
may be due to discharge of industrial wastes
which enter the metropolitan sewers. It is
recognized that unpredictable slugs of such
wastes may affect the plant influent at any
time of any day.
The return of waste activated sludge to
the primary tanks will be discontinued in the
future by modification in design; and
operation. To avoid the presence of waste
activated sludge in the influent and its effect
on effluent quality to the greatest extent
possible during the study, the final tests on
the swirl unit and on the plant primary
tanks during runs in September 1975 were
conducted from 1100 hours (11:00 a.m.) to
1500 hours (3:00 p.m.). During this period
the total plant flow, based on Figure 18 data,
is fairly constant and decreases only about
five percent.
Based on the developmental work at the
LaSalle Hydraulic Laboratory it was decided
to use a swirl separator with a diameter of
3.66 m (12 ft) for parallel tests with the
Humber Plant conventional primary tanks.
Based on data given in Figure 13b, the
dimensions selected for the swirl separator are
shown in Figure 19 and 20.
The Toronto pilot unit was constructed
with a mechanized skimmer for scum
collection in the outer ring instead of using a
scum pipe as recommended by the hydraulic
laboratory and as shown in 'Figure 19 and 20.
A washing spray was provided to convey the
scum out of the swirl chamber. The chamber
was sized for a minimum flow of 7.88 I/sec
(0.18 mgd) and a design flow of 16.2 I/sec
(0.37 mgd). However, in the final tests the
flows were 13.1 I/sec (0.3 mgd) and 19.7 I/sec
(0.45 mgd). The surface area of the swirl
concentrator is 10.49 m2 (113 ft2). The total
volume below the weir elevation is 16.34 m3
(577 ft3). When the bottom of the cone is
filled 0.91 m (3 ft) deep with settled solids,
the liquid volume is reduced to 15.86 m3
(560 ft3). Thus, the^verage volume is 16.1 m3
(568.5 ft3) and at a flow rate of 13.1 I/sec
(0.3 mgd) the detention time is 20.4 minutes.
A variable speed suction pump was installed in
the aerated influent flume to the primary tanks
for the purpose of providing the swirl unit with
the same raw wastewater as was being handled
by the plant tanks. The flow was lifted about
3.6 m (12 ft) and conveyed through a Parshall
flume to the swirl separator. The swirl effluent
was returned to the primary tanks. Sludge was
withdrawn from the swirl separator by gravity
into a 0.18 m3 (50 gal) steel drum and, after
measurement and sampling, was discharged
into the primary tanks. A line diagram of the
layout is shown in Figure 17. The sampling
points to compare performance of the swirl
chamber and the plant primary tanks are
shown.
Series of Tests
Relatively poor performance of the swirl
separator during early test periods was
observed. It was thought that the causes of
the poor performance were the return of
waste activated sludge during the night hours
and infrequent sludge drawoff. The sludge
removed from the swirl separator at four-hour
intervals had a solids content of about 4.5
percent. This is similar to the solids content
of the sludge removed from the conventional
tank. The separator sludge volume was only
about 12 percent of the sludge volume
removed from the primary tanks based upon
flow.
It was concluded that the swirl separator
hopper lacked the storage volume to permit
withdrawals at four-hour intervals and that
the relatively large volume occupied by sludge
in the upper cylindrical portion of the swirl
unit was consequently carried out by the
dynamic swirling flow in the effluent
38
-------
I/Bui 'uojieajuaouoQ spi|os papuadsng
s
in
CN
to
< oo
in
r^
CO
J.
C/5
<
D
CC
HI
LU
CC
o
in
u.
O
ui
cc
LU
> ~a
2 1
52 5 W,
-------
3.66m(12ft)Dia.
3 Slots in Pipe
for Sludge Removal
24 cm (9.5 in)
0.10m (0.33 ft) Plug
Valve for
Sludge Drawoff
Theoretical Apex
of Cone
SECTION A-A
FIGURE 19 SECTION OF SWJRL PRIMARY SEPARATOR
40
-------
Inlet
0.2m (0.67ft)
Scum Pipe
PLAN
FIGURE 20 PLAN OF SWIRL PRIMARY SEPARATOR
0.15m (0.5 ft)
- 0.24m
(0.8 ft) Sq.
41
-------
discharge. Therefore, it was decided to limit
the tests to four hours during the peak
daytime hours when it was assumed that there
was little or no return of waste activated
sludge, and to reduce the number of tests as
shown in Table 46 of Appendix D.. To
improve sludge withdrawal from the swirl
separator a horizontal disc was installed as a
baffle immediately above the sludge opening
in the vertical riser pipe.
The second series of tests (June 23 to 27,
1975) were run with an average flow of; 13.1
I/sec (0.30 mgd); and from July 2 to 8, 1975,
with an average flow of 19.7 I/sec (0.45 mgd).
These flowrates were varied to agree with the
diurnal flow to the plant. The results of these
tests are given in Tables '47,48,49. and 50 of
Appendix D and all results are for dry weather
conditions. With an average flow of 13.1: I/sec
(0.30 mgd) to the swirl separator, and hourly
removal of sludge, the removal of total
suspended solids ranged from 9.8 to 49.5
percent, with an average of 36.6 percent. The
removals in the primary tanks ranged from
14.1 to 57.3 percent, with an average of 43.8
percent. Thus, the removal in the primary
tanks was about 20 percent greater than the
removal in the swirl. These results are shown
in Table 8.
During the second series of tests, the
average flow to the swirl separator was
increased to 19.7 I/sec (0.45 mgd), and sludge
withdrawal was increased to half-hour
intervals to prevent excessive build-up in the
swirl hopper. The removal of total suspended
solids in the swirl separator ranged from 2.8
to 27.2 percent with an average of 17.7
percent. In the primary tanks the removals
ranged from 0 to 50 percent with an average
of 16.6 percent. These results are shown in
TABLE 8
REMOVAL OF TOTAL SUSPENDED SOLIDS, JUNE 23-27, 1975
(DRY-WEATHER)
PERCENT REMOVAL
Average
Influent
Concentration Swirl
(mg/l) Separator1'2
Date
Average Flow
Primary
Settling Tanks3
13.1 I/sec (0.3 mgd) . 4,239 I/sec (96.9 mgd)
June 23
24
25
26
27
380
525
488
904
368
49.5
36.8
48.4
38.5
9.8
42.1
57.3
42.6
62.8
14.1
Average
36.6
ADDITION OF WASTE ACTIVATED SLUDGE
43.8
Time Pumping
Ceased
9:00
8:00
10:00
9:30
9:30
Duration of
Pumping (Hrs)
9
8
10
9.5
9.5
Date
June 23
24
25
26
27
1Overflow rate: 108 m3/day/m2 (2,650 gal/day/ft2); detention time: 20 min.
2Sludge drawoff at 1 hr intervals
3Overflow rate: 81.8 m3/day/m2 (2,004 gal/day/ft2); detention time: 57.1 min.
42
-------
Table 9. These removals were less than 50
percent of those obtained in the initial second
series tests at the lower flow of 13.1 I/sec (0.3
mgd). Therefore, it was decided to carry out
additional tests using slightly revised
procedures which were conducted in
September 1975.
The detailed data for the dry-weather
tests made from May 1 to June 10 are given in
Tables 36,37 and 38for the swirl separator,
and for the primary settling tanks in Tables
41,42 and 43, Appendix D. With a flow of
13.1 I/sec (0.30 mgd) in the swirl separator
the removal of total suspended solids ranged
from zero to 50.3 percent, with an average of
23.5 percent for the 14 days. When the flow
through the swirl separator was increased to
19.7 I/sec (0.45 mgd), the removal ranged
from zero to 16.9 percent, with an average of
7.2 percent for the seven days. By
comparison, the removals in the primary
tanks ranged from 48.8 to 88.3 percent with
an average of 68.8 percent for the first period
of 14 days. For the second period of seven
days, the removals in the primary tanks
ranged from 30.4 to 83.8 percent with an
average of 62.1 percent. These results are
shown in Table 10.
The sampling and analytical program for
the third and fourth series of tests is shown in
Table 51 of Appendix D. These tests are
similar to the previous series except that
analyses were limited to determination of
solids and were taken hourly on a composite of
two samples rather than once in four hours on a
composite of four samples. The flowrates to
the swirl concentrator were kept constant
during the four-hour test period because the
variation in plant flow in that period was only
10 percent. Tests for BOD and COD were
TABLE 9
REMOVAL OF TOTAL SUSPENDED SOLIDS, JULY 2-8,1975
(DRY-WEATHER)
PERCENT REMOVED
Average
Influent
Concentration Swirl Primary
Date (rng/D Separator1 <2 Settling Tanks3
Average Flow
19.7 I/sec (0.45 mgd) 4,244 I/sec (97.0 mgd)
July 2
3
4
7
8
464
436
512
500
440
25.8
2.8
16.4
27.2
16.4
17.2
0
15.6
0
50
Average
Date
July 2
, 3
4
7
8
17.7
ADDITION OF WASTE ACTIVATED SLUDGE
16.6
Time Pumping
Ceased
11:00
9:00
9:00
9:00
9:00
Duration of
Pumping (Mrs)
11
9
9
10verflow rate: 162 m3/day/m2 (3,980 gal/day/ft2); detention time: 14 min.
2Sludge drawoff at 0.5 hr intervals
3Overflow rate: 82 m3/day/m2 (2,005 gal/day/ft2); detention time 42.6 min.
43
-------
TABLE 10
REMOVAL OF TOTAL SUSPENDED SOLIDS - DRY WEATHER FLOW,
May 1 -June 10,1975
Average
Influent
Date
Average Flow
May 1
2
3
8
g
10
11
30
June 2
4
7
8
g
10
Concentration ;
(mg/l)
576
412
432
456
460
512
476 •
660
620
668
632
572
712
472
Swirl
Separator
13.1 I/sec (0.3 mgd)
32.6
13.6
12.0
21.0
21.7
10.1
0.0
30.9
25.2
50.3
44.3
20.3
37.6
g.s
Primary
Settling Tanks
4,01 2 I /sec (91 .7 mgd)
66.5
55.7
85.3
74.2
58.5
77 .g
67.7
62.3
48.8
65.g
88.2
88.3
60.8
63.7
Average
Average Flow
May 12
13
14
15
16
17
18
Average
Flow
444
428
476
604
476
428
568
23.5
19.7 I/sec (0.45 mgd)
8.1
0.0
7.6
g.s
8.4
0.0
16.g
7.2
Swirl Separator
13.1 I/sec 19.7 I/sec
(0.3 mgd) (0.45 mgd)
Overflow Rate
m3/day/m2 ' 108 162
(gal/day/ft2) (2,650) (3,080)
Detention Time (min) 20.4 14
deleted. Sludge was removed from the swirl
separator at one-half hour intervals. Effluent
samples were collected at time intervals;after
the influent samples, more or less equivalent
to the detention time in the unit. Effluent
68.8
3,924 I/sec (89.7 mgd)
52.0
30.4
56.5
65.4
68.5
83.8
78.2
62.1
Primary Settling Tanks
4,012 I/sec 3,924 I/sec
(91.7 mgd) (89.7 mgd)
79.2
(1,896)
60.3
73.7
(1,855)
61.7
samples from the primary tanks were taken
one hour after the influent samples, to
conform with the theoretical detention time
of the units.
The data for the tests made from
September 2 to 15, 1975, are given in Tables
44
-------
52' to 61' in Appendix D. Data for tests made
from June 16 to 25, 1976, are given in Tables
39, 40, 44 and 45 of Appendix D.
The April 1976 tests were conducted to
determine solids removal efficiencies at 6.6
I/sec (0.15 mgd) and 8.7 I/sec (0.2 mgd). The
overflow rates were thus 54.1 m3/day/m2
(1,376 gal/day/ft2) and 72 m3/day/m2
(1,766 gal/day/ft2).
The percent removal of the various solids
parameters for the September 1975 tests are
shown in Tables 62 to71 in Appendix D. The
removals are shown for each hourly test, and
averaged for each day and for the five days'
run at each flow. The data show that on some
days there was considerable variation in the
percent removal for the hourly tests. This
may have been due to three factors: (1) Grab
samples were used for sampling. Tests were
made on composites of two grab samples
taken one-half hour apart. Grab samples of
raw sewage may vary considerably in solids
content. More consistent results may have
been obtained if automatic samplers had been
used to provide a composite over a four-hour
period with samples taken at intervals of a
few minutes. (2) The actual detention times
may have differed from the theoretical
detention times since the flow in both the
swirl separator and the primary tanks may
have been subject to "short circuiting." The
actual detention time may have been as much
as 50 percent of the theoretical detention
time. For this reason the effluent samples
. may not have reflected the true change in the
influent samples. (3) There were industrial
wastes in the raw influent flow. Further, the
discharge of industrial wastes into the sewer
system tributary to the Humber plant may
have caused temporary but abrupt changes in
the percent of solids in the plant influent or
the settling velocities of the industrial
discharged solids.
The foregoing reasons may account foi|
the erratic results obtained in the hourly tests]
Therefore, it is considered appropriate to
disregard the hourly results and use the
averages of the daily tests and the average of
the five-day test runs. These results are shown
graphically in Figures 21 through 25.
During the five-day test period, with a,
flow of 13.1 I/sec (0.3 mgd) and an overflow
rate of 108 m3/day/m2 (2,650 gal/day/ft2)
and a detention time of 20.4 minutes in the
swirl separator, the removal percentages of all
five of the tests were greater in the swirl
separator than in the primary tank. The
removal of total suspended solids for all five
days was 43.1 percent for the swirl separator
compared to 41.9 percent for the primary
tanks.
With a flow of 19.7 I/sec (0.45 mgd) and
an overflow rate of 155 m3/day/m2 p,B05
gal/day/ft2) and a detention time of 14
minutes through the swirl separator, it
outperformed the primary tank in removal of
settleable solids — 47.7 percent for the former
versus 36.3 percent for the latter. In terms of
suspended solids, the swirl chamber removed
25.3 percent versus 36.0 percent for the
primary tank.
The average of the five four-hour test
runs is shown in Table 11.
EVALUATION OF SOLIDS TESTS
The Recommended Standards for Sewage
Works by the Great Lakes-Upper Mississippi
River Board of State Sanitary Engineers
(commonly referred to as the 10-State
Standards) contain criteria for design of
primary settling tanks for use in secondary
treatment plants. These stipulate depths of
not less than 2.13 m (7 ft) but no mention is
made of criteria for removal of settleable or
suspended solids. A curve is given for BOD
removals based on surface settling rate — i.e.,
overflow rate. The percent BOD removal
ranges from 22 percent at overflow rate of
81.46 m3/day/m2, (2,000 gal/day/ft2), to 32
percent at 36.66 m3/day/m2 (900
gal/day/ft2), and to 37 percent at 16.29
m3 /day/m2 (400 gal/day/ft2 ).
The ASCE Manual of Engineering
Practice No. 36* is more specific. This manual
states that the object of settling tanks is to
remove settleable solids and to reduce the
suspended solids content of the sewage. It
states that both displacement time (detention
time) and surface area should be considered in
design and that detention periods of 60 to
120 minutes should, in theory remove 50 to
70 percent of the influent suspended solids.
Figure 6' of ASCE Manual No. 36'4.indicates
the relationship between overflow rate and
45
-------
9-
•2
-75
• •
9-
3
-75
M
1 ""•
-75
:
[. h
_1____
9-
5
-75
9-
8
t*
I:
0
0
I
C
»
3
A
QL
<
I-
"-
k
i "! -
-L
80
_i
40 o
5
UJ
5?
0
Swirl 13.1 I/sec (0.3 mgd>
Primary 4,121 l/sec(94.2 mgd)
AVERAGE FLOWRATE
Overflow rate - 108 m3/day/m2(2,650 gal/day/ft2)
Overflow rate - 79.42 ma /day/rr>2( 1.950
mr
III
9-9)
-75
9-IC
l-7£
'
Ml
-75
AVI
>
ERA
GE I
9-12
=LO
-7J
\NftA
fTE
—
Ml-Td
_J
fK
CO
AL
>-
•£
Q.
L
80
_,
40 i
UJ
S?
o
Swirl'lfi .7k|/sec (0.45 mgd) Overflow rate - 108 m3/day/m2(2.650 gal/day/ft2)
Primary 4,340 l/seciaa.2 mgd) Overflow rate - 83.5 m3/day/m2(2,050 gall/day/ft2)
§ —
DAU.Y AVERAGE OF 4 TESTS
Atrcn Ar»c r»c on TCOTO
«
— fr
'4 —
^
if —
fc
FIGURE 21 DRY-WEATHER REMOVAL OF TOTAL SUSPENDED SOLIDS
TORONTO, September 1975
46
-------
'•
r
i
fTTTi
9-2-75
; :
i
I
9-3-75
1 • • i '••
mil — i
9-4-75
1
9-
MMf
5-75
$-8-75
t£
ft
U) a.
AUL
80
40
'
0
111
AVERAGEFLOWRATE
Swirl 13.1 I/sec (0.3 mgd) Overflow rate - 108 m3/day/ma(2,650 gal/day/ft2)
Primary 4,121 l/sec(94.2 mgd) Overflow rate - 79.42 m3/day/m2 (1,950 gal/day/ft2)
j
9-
•S
j -t
.
-re
- r s.
'f
3-
-I
i1
M
i
i
i
\
J-
II
*
-7S
3-
12
;
-7J
_i
IB
MJ
^71
~f V
1
f
C
I
J
f
i
0
a
>-
rr
; <
' E
; °-
,L
1
art
40
0
I
IU
oc
AVERAGE FLOWRATE
Swirl 19..7 I/sec (0.45 mgd) Overflow rate — 108 m3/day/m2 (2,650 gal/day/ft2)
Primary 4.340 l/sec(99.2 mgd) Overflow rate - 83.5 m3/day/m2(2,050 gal/day/ft2)
'
Dt
I
• I • '
r
ULY AVERAGE OF 4 TESTS '
1 .
.,.
i
i • i
i
!
j
^
, „
fc
j
i
FIGURE 22 DRY-WEATHER REMOVAL OF FIXED SUSPENDED SOLIDS
TORONTO, September 1975
47.
-------
-- — —
— - -•
9-2-75
9-3-75
• !
. 9-4-75
..
9-5-75
>
.
u_
^ ' r~p
CO Q-
ff*
'•" |~
f
9-8-75
ALL
BO
40 jo
'• cc
is?
1
1
o
AVERAGE FLOWRATE
Swirl 13.1 I/sec (0.3 mgd) Overflow rate - 108 m3/day/m2(2.650 gal/day/ft2)
Primary 4,121 l/sec(94.2 mgd) Overflow rate - 79.42 m3/dayAn2( 1.950 gal/day/ft2)
™-_.__^__^__
— -
-—
...
9-9-75
— _i_
_____
9-10-75
'
•
9-IIJ-75
"
I
i
9
-1^-75
* :
-
9-15-75
•
or
^
CO
f|
i —
*>* !
1 U_
Q.
ALL
80
>
4qi
f
1
0
AVERAGE FLOWRATE
Swirl 19.J I/sec (0.45 mgd) Overflow rate - 108 m3/day/m2(2,650 gal/day/ft2)
Primary 4^40 l/sec(99.2 mgd) Overflow rate - 83.5 m3/day/m2(2,050 gal/day/ft2)
: f
« 4 I
i,
DAILY AVERAGE OF
i 1 1 1
AVERAGE
4 TESTS
!b?
' ' ^
20 TESTS "^
j
> 1
• ! i *^
! '|
«j 1 ft
• A
| X^ !
. r . .;, .
f
FIGURE 23 DRY-WEATHER REMOVAL OF VOLATILE SUSPENDED SOLIDS
TORONTO, September 1975
48
-------
1
9
I ' '
-2
-75
t
3-
:
3-75:
! ' •
9-4^75
- . :
9-S-75
9-8-75
>.
_i <
^ S
CO CL
" "
HOBM
; ALL
80
VAO1A13
; LL
1
1
0
AVERAGE FLOWRATE
Swirl 13.1 I/sec (0.3 mgd) Overflow rate - 108 m3/day/m2(2,650 gal/day/ft2)
Primary 4,121 l/sec(94.2 mgd) Overflow rate - 79.42 m3/day/m2(1,950 gal/day/ft2)
— I
| }
M
i
-75
I 'il '
; I '•
I
i
i
M0-75
! I
'" '
— |—
9- 1
(-75 I
:
£K
\-7i
r
'. i . i-.
< i'
3-I4-75
»v*
_j S
ce 2
=£ a:
cb Q-
/
ML '
80
^
o
40,^
cc
!5?
0 |
I
i
AVERAGE FLOWRATE
Swirl 19.7 I/sec (0.45 mgd) Overflow rate - 108 m3/day/m2(2,650 gal/day/ft2)
Primary 4.340 l/sec(99.2 mgd) Overflow rate - 83.5 m3/day/m2(2.050 gal/day/ft2)
4—
L
" -
)AILY AVERAGE OF 4 TESTS
J 1 1 M 1 1 1— • 1 !
,AVLKAub Ur £\j 1 to 1 o
.
-f>
^
S
,
!
S*-TT—
. i . '.
'
— 1»
!
i
i
FIGURE 24 DRY-WEATHER REMOVAL OF SETTLEABLE SOLIDS BY VOLUME
TORONTO, September 1975
49
-------
9-Z-75
.
,.
"•""
k-3-75
;•
i
—
9-4-75
-
i ; " I ':
p
-G
^75
9-8-75
IV
1™
_j 4^
' ^
4
c
i o:
0 Q-
ALL i
80
!_i
400
re
o
AVERAGE FLOW/RATE
Swirl 13.1 I/sec (0.3 mgd) Overflow rate - 108 m3/day/m2(2,650 gal/day/ft2)
Primary 4,121 l/sec(94.2 mgd) Overflow rate - 79.42 m3/day/m2 (1,950 gall/day/ft2)
••L
9-9-75
••
•
•
9-IO-75
9-
•
i
11^75
?
-1^-75
\
£-
IS-75
,.,.,. , , .
ft"
_l! <
\ O£i ^
. ^
'; <*>
JE.
Q.
AliL
80
<
402
in
1°=
ss
0
AVERAGE FLOWRATE
Swirl 19.7 I/sec (0.45 mgd) Overflow rate - 108 m3/day/m2(2.650 gal/day/ft2)
Primary 4.340 l/sec(99.2 mgd) Overflow rate - 83.5 m3/day/m2(2.050 gal/day/ft2)
1 t f
f — . 1 J
^-r-r i
i
5 .BJ,t.J=ss!!a-s,
DAILY AVERAGE
1 1 1 1 1 r|
, , _J 1 1_^
! AVERAGE
, , i 1 • i
OF 4 TESTS
OF 20 TESTJ
1
i
1 h
« <
' 1
— i — i — | — n
— . — i — , —
i ; . •
4 —
i L_
\—#—
— »
FIGURE 25 DRY-WEATHER REMOVAL OF SETTLEABLE SOLIDS BY WEIGHT
TORONTO, Septemb&r 1975
50
-------
Swirl Flow
TABLE 11
SUMMARY OF TESTS -
SEPTEMBER, 1975
(DRY-WEATHER)
PERCENT REMOVAL
(only samples where
(all samples) positive removal)
'Swirl i Primary Swirl Primary
0.3
0.3
0.3
0.3
0.3
0.45
0.45
0.45
0.45
0.45
Settleable solids, ml/I
Settleable solids mg/l
Total suspended solids, mg/l
Volatile suspended solids, mg/l
Fixed suspended solids, mg/l
Settleable solids, ml/I
Settleable solids mg/l
Total suspended solids, mg/l
Volatile suspended solids, mg/l
Fixed suspended solids, mg/l
Swirl Constant Flow
I/sec mgd •
13.1 0.3
19.7 0.45
75.2
60.4
43.1
45.8
26.5
47.7
42.0
25.3
25.9
22.0
Primary Average
I/sec mgd
4,121 94.2
4,340 99.2
49.1
48.8
41.9
44.8
24.2
36.3
49.7
36.0
35.8
37.1
Flow
75.2
67.3
43.1
45.8
38.3
47.7
43.7
25.3
25.9
29.8
49.1
55.3
• 46.3
47.2
45.4
43.4
49.7
36.0
35.8
57.4
Note: Each result is average of five four-hour tests
suspended solids removal, based on actual.
plant results. The median line on this graph is
an indication of expected performance
although actual plant results vary widely from
this line. Based on the foregoing criteria, a
primary settling tank with an overflow rate of
36.66 m3/day/m2 (900 gal/day/ft2) would
remove 60 percent of suspended solids. Such
a tank, with an average depth of 3.05 m (10
ft) would have a detention time of two hours.
Work by Smith5 confirms these general
guidelines.
The detention times and overflow rates
for the 10 four-hour tests made in September
1975 are shown in Table 12. During the first
five tests, the average detention time was 0.93
hour and the overflow rate was 83.50
m3/day/m2 (2,050 gal/day/ft2). The average
removal of suspended solids was 36.0 percent
compared to 40 percent indicated in Figure 6
of ASCB Manual No. 364. Thus, it can be
concluded that the primary tanks were
removing on the average a reasonable amount
of suspended solids, especially if it is assumed
that some flow of waste activated sludge
continued to circulate through the primary
tanks throughout the day hours while the
tests were underway.
Additional tests were conducted on the
primary treatment units in June of 1976, and
these data are shown at the bottom of Table
13. These results are more varied than those
taken in the September 1975 tests, regardless
of the fact that the detention times and
overflow rates are similar.
The pilot swirl separator, as constructed
at the Humber Plant, is shown in Figures 19
and 20. The gross volume below the weir
elevation is 16.28 m3 (575 ft3). During the
September 1975 series of tests, sludge was
withdrawn at one-half hour intervals. The
maximum sludge volume collected during the
four-hour test run was 3.54 m3 (935 gal). With
seven drawoffs, the average amount stored in
each half hour period between pumping
periods was 0.44 m3-(67 gal). This volume
of sludge would occupy about 0.62 m (2 ft)
of depth in the bottom hopper and would
reduce the volume available for sewage to
16.03 m3 (566 ft3). Using this volume, the
detention time for a flow of 13.1 I/sec (0.3
mgd) would be 20.4 minutes; for a flow of
51
-------
Date
TABLE 12
DETENTION TIME AND OVERFLOW RATE -
PRIMARY TANKS, SEPTEMBER 1975 AND JUNE 1976
(DRY-WEATHER)
I/sec Flow Detention Overflow Rate
Percent Removal
mgd
Time Hours m /day/m2 gal/day/ft2 Suspended Solids
1975
9-2 4,218 4.22
9-3 4,139 4.14
9-4 4,169 4.17
9-5 3,373 4.05
9-8 3,369 4.04
Average
Maximum Deviation
9-9 3,491 4.19
9-10 3,426 4.11
9-11 3,964 4.76
9-12 3,933 4.73
9-15 3,281 394
Average
Maximum Deviation
1976
6-16 4,350 4.35
6-17 4,150 4.15
6-18 4,170 4.17
6-19 3,870 3.87
6-22 4,060 4.06
6-23 4,170 4.17
6-24 4,200 4.20
6-25 4,490 4.49
96.4
94.6
95.3
92.5
92.4
94.2
95.7
93.9
108.7
.107.9
90.0
99.2
99.4
94.7
95.3
88.4
92.6
95.2
96.0
102.5
0.99
0.97
0.97
1.00
1.00
0.98
2%
0.96
0.98
0.85
0.85
1.02
0.93
10%
0.93
0.97
0.97
1.04
0.99
0.97
0.96
0.90
0.678
0.689
0.692
0.671
0.671
0.696
0.682
0.787
0.783
0.653
0.722
0.688
0.692
0.642
0.673
0.692
0.698
0.745
1,930
1,960
1,970
1,910
1,910
1,940
2%
1,980
1,940
2,240
2,230
1,860
2,050
10%
2,055
1,957
1,970
1,827
1,914
1,968
1,984
2,119
14.0
57.5
48.2
43.2
42.7
41.9
30.9
29.6
22.9
49.8
46.0
36.0
66.0
75.0
0
61.0
3.0
15.0
6.0
72.0
TABLE 13
TOTAL SUSPENDED SOLIDS
Influent
Concentration %
REMOVAL
Removal
(mg/l) in Swirl Separator
Swirl Primary Flow:
• 8.7 I/sec (0.2 mgd)
June 16, 1976
June 17, 1976
June 18, 1976
June 21, 1976
Swirl Primary Flow:
6.8 I/sec (0.1 5 mgd)
June 22, 1976
June 23, 1976
June 24, 1976
June 25, 1976
133
164
162
175
Average —
i
187
169
138 !
158
62
48
63
72
61.25
72
72
72
67
(DRY-WEATHER)
I/sec
4,354
4,148
4,174
3,872
4,056
4,170
4,205
4,490
Flow in
Primary
mgd -
99.4
94.7
95.3
88.4
92.6
95.2
96.0
102.5
% Removal
in Primary
66
75
0
61
3
15
6
72
Average - 70.75
Swirl 0.15 mgd corresponds to 1,326 gpd/ft2 (54.1) m3/day/m2 — 40 minute detention time
Swirl 0.2 mgd corresponds to 1,768 gpd/ft2 (72.2) m3/day/m2 - 30 minute detention time
52
-------
19.7 I/sec (0.45 mgd) it would be 13.6
minutes. These calculations are based upon
the volume available for the liquid. In reality,
much of the cone hopper volume is not used
as flow-through space. If only the volume of
the cylindrical portion of the swirl were
considered, its volume would be 9.8 m3 (349
ft3) resulting in a detention time of 12.4
minutes at a flow of 13.1 I/sec (0.3 mgd) and
8.3 minutes at a flow of 19.7 I/sec (0.45
mgd). Thus the real detention time is between
these values. For economic comparison with
conventional primary settling tanks, the most
adverse condition was chosen.
The surface area of the swirl separator is
10.5 m2 (113.1 ft2). With a flow of 13.1
I/sec (0.3 mgd) the overflow rate is 107.9
m3/day/m2. (2,65.0 gal/day/ft2), and with a
flowrate of 19.7 I/sec (0.45 mgd) the
overflow rate is 162.1 m3/day/m2 (3,980
gal/day/ft2). In Table 64 the removal of total
suspended solids in the swirl separator with a
flow of 13.1 I/sec (0.3 mgd) is 43.1 percent
compared to 41.9 percent in the primary
tanks. Therefore, the performance of the swirl
separator with a flow of 13.1 I/sec (0.3 mgd)
and an overflow rate of 108.3 m3/day/m2
(2,650 gal/day/ft2) may be considered to be
equivalent to the primary tank with an
overflow rate of 81.46 m3/day/m2 (2,000
gal/day/ft2) and a one-hour detention time.
In Table 69 the removal of total
suspended solids in the swirl separator with a
flow of 19.7 I/sec (0.45 mgd) is 25.3 percent.
This percent removal is below design
standards and hence it is concluded that the
swirl separator of the size used was not.
suitable for this high rate of flow.
Additional tests conducted on the swirl
pilot unit during June 1976 at the lower flow
rates of 6.57 I/sec (0.15 mgd) and 8.76 I/sec
(0.2 mgd) yielded total suspended solids
removals in the swirl unit which were
significantly better than those discussed
earlier. The averages for the two tests are
61.25 and 70.75 percent respectively for the
two flowrates. The calculated "test data are
shown in Table 13 and the actual data are
presented in Tables 39,49,44 and 45 in
Appendix D. All test procedures were similar
to those used in the September, 1975, tests.
The increased detention times resulted in
significant improvement in removal
efficiencies.
One of the purposes of these tests was to
confirm the laboratory results obtained at the
LaSalle Hydraulic Laboratory where synthetic
sewage solids were used. The resultant design
curves were given in Figure 11.
The results obtained in the Toronto tests
for a swirl separator with a diameter of 3.66
m (12 ft) are compared in Figure 26 with the
results predicted from the laboratory model
tests. Thus with a flow of 13.1 I/sec (0.30
mgd) the laboratory predicted 37 percent
removal, versus 43.1 percent obtained at
Toronto. For a flow of 19.7 I/sec (0.45 mgd)
the laboratory predicted a removal of 28
percent, versus 25.3 percent obtained at
Toronto. The two results obtained in the field
do not vary enough from the laboratory
findings to cause any change in the curves
developed from the LaSalle model study. Thus,
it can be concluded that the field work has
confirmed the laboratory work, based on the
tests to date.
Figure 27 compares the time to achieve
treatment for the swirl primary and the
conventional unit at Toronto. The high initial
efficiency of the 3.7 m (12 ft) unit is evident.
For the June 1976 tests, the 6.57 I/sec
(0.15 mgd) flow shows a predicted removal
efficiency of 57 percent in Figure 1 la, while
the 8.76 I/sec (0.2 mgd) flow shows a
predicted removal efficiency of 48 percent. In
both cases, the actual removals exceeded the
predicted performance.
EVALUATION OF BOD TESTS
The Recommended Standards for Sewage
Works by the Great Lakes-Upper Mississippi
River Board of State Sanitary Engineers
(10-States Standards) includes a figure
relating BOD removals to overflow rates in
primary settling tanks. It also states:
"However, BOD removals for sewage
containing appreciable quantities of industrial
wastes should be determined by laboratory
tests and consideration of the quantity and
characteristics of the waste."
In the second series of tests conducted
for four-hour periods from June 23 to July 8,
53
-------
100-
90-
TORONTO PILOT TEST SEPTEMBER 1975
I
I
||
30 40 50 100
FLOWRATE - I/sec
i i I i I i
200 300 500
t t i i L
0.2 0.3 0.40.5 1 2;0 3.0 4.0 5JQ
FLOWRATE - mgd
10
FIGURE 26 PREDICTED VERSUS ACTUAL SUSPENDED SOLIDS REMOVAL
54
-------
conventional primary
swirl separator
I
60
120
Time (minutes)
FIGURE 27 COMPARISON OF TIME TO ACHIEVE
TREATMENT
BODS tests were carried out; 'the results are
shown in Tables47, 48, 49 and 50. The
percent removal of BODS derived from these
tests is shown in Table 14., During the test
period, the average percent removal in the
primary tanks was 13.1 percent. For the
overflow rate of 81.46 m3/day/m2 (2,000
gal/day/ft2) in the primary tank the 10-States
Standards indicates a removal of 22 percent.
Hence, the removal was about 60 percent of
the anticipated amount. The removal
percentage for the swirl separator with a flow
of 13.1 I/sec (0.3 mgd) and an overflow rate
of 108 m3/day/m2 (2,650 gal/day/ft2) was
slightly less than for the primary tank.
In the tests made from July 2nd to 8th,
the removal percent in the primary tanks was
5.2 percent, versus 22 percent shown in the
10-States Standards. Thus, the removal was
about 24 percent of expectation. Again, the
TABLE 14
REMOVAL OF BODs-
JUNE 23-JULY 8, 1975
Date
1975
6-23
6-24
6-25
6-26
6-27
Average
Influent
(mg/l)
284
301
300
382
339
321
Swirl
248
278
247
361
289
285
Effluent (mg/l)
Primary
251
227
277
343
295
279
% Removal
Swirl ' Primary
12.6 11.6
7.6 24.6
17.6 7.7
5.5 10.2
14.7 13.0
11.2 13.1
Data for above
Flow: SwirrSeparator = 13.1 I/sec (0.3 mgd)i
Primary Tanks = 4,245 I/sec (96.9 mgd)
Overflow Rate:
Swirl = 108 m3/day/m2 (2,650 gal/day/ft2)
Primary Tanks = 81.46 m3 /day/m2 (2,000 gal/day/ft2)
332 326
7-2
7-3
7-4
7-7
7-8
296
361
307
347
Average
295
309
307
343
316
329
Data for above
Flow: Swirl Separator = 19.7 I/sec (0.45 mgd)
Primary tanks = 4,254 I/sec (97.1 mgd)
Overflow Rate:
Swirl = 155 m3/day/m2 (3,805 gal/day/ft2)
Primary Tanks = 81.87 m3/day/m2 (2,010gal/day/ft2)
321
290
314
332
302
312
1.8
0.0
14.4
0.0
1.1
4.0
3.3
2.0
13.0
0.0
13.0
5.2
55
-------
removal in the swirl separator with a f!0w of
19.7 I/sec (0.45 mgd) was slightly less than
that in the primary tanks — 4.0 versus 5.2
percent. Basically, the function of primary
clarification is removal of suspended and
settleable solids rather than BOD.
SETTLEABILITY OF COMBINED
SEWAGE FLOW SOLIDS
A report was prepared by :Beak
Consultants Ltd. on the settling velocity
characteristics of Toronto wastewater solids
in storm flow periods and on the hydraulic
character of the swirl separator and the
primary settling tanks. This report is
contained as an Appendix C to this report.
Samples were collected of the sewage
influent to the plant during two storm events.
Samples were collected from the influent
channel to the primary tanks during the
rising, peaking, and falling phases of the storm
flows. Settling column tests performed on
these samples indicated that the settling
characteristics of the combined sewage solids
were similar to the Amberlite IRA-93 anion
exchange resin used to simulate sewage solids
in the previous laboratory work by LaSalle.
The settling column tests also showed
increased settleability of the solids during the
storm event. Thus, in the first storm, for a
constant overflow rate of 173 m3/day/m2
(4,240 gal/day/ft2), the predicted solids
removals were 74, 76, and 81 percent for
rising, peaking, and falling storm conditions,
respectively. In the second storm event for
the same overflow rate, the predicted solids
removals were 60, 80, and 83 percent for
rising, peaking, and falling storm conditions,
respectively. The actual removal of total
suspended solids during the first storm was
zero percent in the swirl separator and 10.2
percent in the primary tanks. The overflow
rate in the swirl separator was1 155
m3/day/m2 (3,800 gal/day/ft2). The average
overflow rate for the day in the primary tanks
was 59 m3/day/m2 (1,449 gal/day /ft2 i) but
during the storm period it was as high as
105.4 m3/day/m2 (2,575 gal/day/ft2).
In the second storm event no records
were provided for solids removals in the swirl
separator and the primary tanks.
Tracer studies, using fluorescent dye,
were conducted on the swirl and the primary
tanks to determine the character of the
hydraulic flow. The results of the tracer study
showed that the hydrodynamics of the swirl
separator lie more towards the plug flow
regime of mixing than they do in the primary
clarifiers. The apparent decrease in active
volume and shift towards plug flow which
occurs as the flow increases in the swirl
separator would indicate that the quiescent
cone of fluid along the axis of the swirl may
increase with increasing flow.
Removal efficiency decreases as the
mixing regime shifts from plug flow to
well-mixed' conditions. The results of the
tracer study would predict that the swirl,
being closer to plug flow in all cases tested,
would accomplish the same removal as the
primary clarifiers but at a significantly higher
upflow velocity or surface overflow rate.
INTERPRETATION AND OTHER
APPLICATIONS
The statement was previously made that
the results should be interpreted as
specifically relating to the flow conditions,
wastewater character, and treatment plant
operating procedures indigenous to the
demonstration installation. This statement
was intended to limit the findings of this
single investigative application to this
application only, and to avoid the hazard of
extrapolating these findings to represent what
a similar swirl separator might accomplish
under other test circumstances.
With this disclaimer, it is possible to
interpret the performance of the swirl unit at
the Humber plant, vis-a-vis the treatment
efficiencies of the primary settling tanks at
this location. Recognition must be made of
the possible deleterious effect of "hang-over"
waste activated sludge in the influent on the
performance of the swirl unit which would,
necessarily, be relatively incapable of
retaining light, semi-colloidal solids under the
flow pattern conditions in such a settling
device. In addition, the unpredictable
presence of industrial wastes of some types
could have had an adverse effect on
clarification under swirl flow conditions.
56
-------
Further, indeterminate phenomena may have
failed to be disclosed because of grab
sampling and compositing practices during the
test runs. These, and other factors, have been
referred to in the preceding portions of this
pilot report.
However, the following interpretations of
the study data are valid for the Toronto pilot
installations:
• The Toronto swirl pilot, in general,
confirmed the basic findings of the LaSalle
Laboratory research study, as described in Ap-
pendix C. Some phases of performance were
better than anticipated; others were less
effective.
• The Toronto swirl unit, thus,
demonstrated the ability of this type of
hydraulic flow to achieve removals of
wastewater solids under both dry-weather and
wet-weather flow conditions.
• The Toronto swirl separator offered no
advantages over the primary settling tanks at
the treatment plant, based on standard design
criteria, in terms of solids removal. For small
treatment facilities if 60 percent suspended
solids removal was not required, the swirl
might be cost effective.
• With heavy solids which are a character-
istic of wet-weather flow conditions, the
efficiency of solids removal is improved as
compared to dry-weather flow conditions.
• • The Toronto swirl separator, due to
the nature of the wastewater flow, achieved
lower BOD removals than the conventional
primary settling tanks, and less removal
efficiency than would normally be expected.
Later tests performed at lower flow rates
yielded total suspended solids removals which
exceeded predicted performance.
• The basic advantage of the swirl
clarification principle is that it requires no
mechanical devices or equipment for the
removal of settled solids from the retention
hopper. This advantage is achieved by
providing greater depth than required by
so-called conventional mechanical primary
settling tanks, thus, imposing additional cost
and construction problems for the former
type of primary facility.
As stated, a swirl pilot unit under
operational conditions other than those
imposed at the Toronto plant may show
primary settling efficiencies that offer
advantages over regular clarification facilities,
but this can be demonstrated only by other
test installations. Such further applications
should be encouraged to avoid the drawing of
overall conclusions that are based on only one
installation at Toronto. Further investigations
must be proposed and made with full
recognition of the comparative performances
described in this report.
The absence of mechanical solids removal
devices in the swirl separator may make it
particularly applicable and advantageous in
some locations and for some purposes where
clarification is carried out at other than
treatment plant sites that are adequately
manned by operation personnel. Such an
application could be one at a combined sewer
overflow location upstream of the treatment
plant or adjacent to an automatic pumping
station.
The handling of sludge from the swirl
hopper would require periodic removal of this
material to some type of storage and disposal
facility. In addition, scum removal equipment
will be needed. Both the scum and sludge
withdrawal problems are common to all types
of primary treatment units.
57
-------
SECTION 5
GLOSSARY
Long-Flow Pattern - The path of the
swirl flow pattern through the swirl separator,
induced by proper baffling which causes the
liquid to traverse the circular chamber more
than once and prevents the incoming flow
from being diverted or short-circuited directly
to the overflow weir, thereby inducing the
solids to discharge into the foul sewer channel
and outlet.
Settling Velocity — Downward velocity
of a particle in sewage.
Settledble Solids — That portion of the
solids contained in the wastewater flow into a
swirl chamber which will subside and be
collected in the chamber due to gravity and
other liquid-solids kinetic conditions induced
by the controlled swirl flow pattern. (Note:
Not all suspended solids are settleable solids,
such as colloidal or other finely dispersed
solids.)
Suspended Solids — 1) The quantity of
material deposited when a quantity of water,
sewage, or other liquid is filtered through an
asbestos mat in a Gooch Crucible or a 0.35
m-0.45 m millipore fiberglas filter.6 2) Solids
that either float on the surface of, or are in
suspension, on water, wastewater, or other
liquids, and which are largely removable by
laboratory filtering.
REGISTERED TRADEMARKS
Plexiglas
Amberlite
Tygon
Petrothene
Petrothene X
Polythene
REFERENCES
1. American Public Works Association —
The Swirl Concentrator as a Combined
Sewer Overflow Regulator Facility; EPA
Report No. EPA-R2-72-008, NTIS No.
PB 214 134, September, 1972.
2. Sullivan, R.H., et al - The Swirl
Concentrator as a Grit Separator Device,
EPA Report No. EPA-670/2-74-026;
NTIS No. PB-233 964. June, 1974.
3. Dalrymple, R.J., et al — Physical and
Settling Characteristics of Particulars in
Storm and Sanitary Wastewaters — ^EPA
Report No. EPA-670/2-7 5-011, NTIS No.
PB-242 001, April, 1975.
4. Society of Civil Engineers — Sewage
Treatment Plant Design; Manual No. 36,
ASCE-WPCF, New York, 1959.
5. Smith, R. — Preliminary Design of
Simulation of Conventional Wastewater
Renovation Systems Using the Digital
Computer — U.S. Department of the
Interior, FWPCA, 1968.
6. Standard Method for the Examination of
Water and Wastewater — American Water
Works Association, Water Pollution Con-
trol Federation, American Public Health
Association, 14th edition, 1975. p. 94.
58
-------
APPENDIX A
HYDRAULIC MODEL STUDY
Previous studies of the swirl concentrator
principle have investigated its application as a
combined sewer overflow regulator and as a
grit chamber. Hydraulic model research
resulted in the development of swirl chamber
configurations and dimensions that will
produce effective solids-liquid separations and
design criteria for prototype installations
were evolved. The present study was
undertaken to determine the application of
the swirl chamber principle to the primary
clarification of wastewater flows.
The same model used in the previous
studies was retained and modified to meet the
new criteria of primary clarification facilities.
These criteria involved differences, in solids
content and characteristics, and reduced
flows, with slower, gentler rotation in the
chamber. Changes in solids content of
influent flows involved the use of lighter and
finer solids material in the model to simulate
actual conditions in field prototype
installations.
Utilization of such finer material
introduced the need for laboratory measuring
techniques not available through the LaSalle
Hydraulic . Laboratory staff, so model test
monitoring analyses were performed by Beak
Consultants, Ltd.*
Principle and Scope of the Study
The principle used in the primary settling
structure is a controlled combination of solids
settling, rotational velocity, flow equalization
of spill of clarified liquid over the overflow
weir, and the slope and shape of the chamber
floor in order to produce the best possible
removal of. the solids.
Although a. foul outlet was provided
during fabrication of the hydraulic models, no
regular use of it was made to test its collecting
efficiency. Intermittent draw-off during flow
was assumed to be the means of removing the
settled solids, or sludge, in actual prototype
units.
The basic layout was evolved from several
* See Section 11, page 6.
different sources: Mr. Smisson's general
arrangement with deep skirt; 60 degree
bottom chamber cone from the experience of
Sogreah Hydraulic Laboratory, Grenoble,
France; larger diameter inlet, weir and skirt
diameters from. LaSalle's previous experience.
The initial layout is shown in Figures 2 and 28.
In order to establish practical prototype
values for the study, the same scale of 1:12
already used for the earlier studies, was
retained. This means the model would be
operating with a 1.22 m (4 ft) square sewer
coming into a 10.98 m (36 ft) diameter
chamber during the first stage of the study.
Later, the square sewer was reduced to 0.72
m (2.36 ft), while the chamber remained
unchanged.
The size chamber, with the weir crest 2.75
m (9 ft) above the chamber floor, would give
21 minutes of retention to 0.31 m3/sec (11
cfs), corresponding to 0.632 lit/sec (0.022
cfs) on the model.
To simplify the testing, three model
discharges were selected whose prototype
equivalents would be as follows:
Model
I/sec •
0.5
1.0
1.5
0.25
0.50
0.75
Prototype
cfs
8.8
17.6
26.4
mgd
5.7
11.4
17.1
MODEL DESCRIPTION
The central feature of the model was the
separation chamber itself, which took the
form of a vertical concrete cone with 60
degree side slope, topped concentrically by a
vertical cylinder 91.5 cm (36 in) diameter
made of 13 mm (0.5 in) Plexiglas®. In the
first stage of the tests, the height of the cone
was adjusted so as to leave between the top of
the cone and the cylinder a 10 cm (4 in) wide
annular flat bench called the chamber floor,
as shown in Figure 3. Later, for subsequent
series of tests, the concrete cone was extended
59
-------
Empty Chamber Showing
Conical Bottom and Low
Inlet — Invert Level with
Outside Floor
Original Layout
FIGURE 28 SWIRL PRIMARY SEPARATOR
(CONICAL BOTTOM AND LOW INLET)
A.
Raised Inlet — Crown 22.86 cm
(9 in) Above Outside Floor
Modification 2
FIGURE 29 SWIRL PRIMARY SEPARATOR, MODIFICATION 2,
(RAISED INLET 61cm [24 in] DIAMETER WEIR AND 71 cm [28 in]
DIAMETER SKIRT)
60
-------
B.
Conical Bottom and Raised Inlet
with 61 cm (24 in) Diameter Weir
Inlet Crown Level with Weir Lip
Modification 2
C.
Conical Bottom and Raised
Inlet Chamber with 61 cm
(24 in) Diameter Weir and
71 cm (28 in) Diameter Skirt
Modification 2
FIGURE 29 SWIRL PRIMARY SEPARATOR, MODIFICATION 2,
(RAISED INLET 61 cm [24 in] DIAMETER WEIR AND 71 cm [28 in]
DIAMETER SKIRT)
61
-------
upward to the periphery of the chamber;wall
so as to eliminate the horizontal bench floor.
A Plexiglas skirt, supported by the
chamber wall and concentric to the cone jaxis,
divided the chamber into two parts; the inner
and outer chambers. The vertical distance
from the bottom of the skirt to the surface of
the cone determined the slot height below the
skirt. The distance could be changed at will
by means of calibrated supporting blocks.
In the center of the cone, an embedded
PVC pipe, 15.2 cm (6 in) inside diamfeter,
provided the support for the circular weir and.
the outlet to evacuate the clear overflow
spilling over the weir. The level of the jweir
could be raised or lowered at will by adding
or removing custom-cut pieces of, the
supporting pipe.
The overflow pipe was later removed to
leave the inner chamber unencumbered; and
the weir was attached to the chamber skirt.
The bottom of the cone was closed and the
clear overflow was evacuated by means of a
circular gutter fixed around the skirt in the
outer chamber connecting with two 2.5 cm (1
in) diameter Tygon® hoses passing through
the chamber wall. I
Inflow to the chamber was supplied
through a poly vinyl chloride (PVC) pipe 10
cm (4 in) diameter set at a slope of 1/ljOOO.
A vibrating solids injection system was placed
on this supply line 2.14 m (9 ft) upstream of
the chamber. Water supply to the model
through the pipe was taken directly from the
constant level tank in one of the laboratory's
permanent pumping stations. This supply
device was used as long as the recovery of the
large-size grain material injected presented no
problems. When smaller grain sizes were used,
a closed circuit device was built. The PVC
supply pipe was removed and replaced by a
76.2 cm (30 in) long Plexiglas pipe while the
solids injection system was changed to
provide just a hopper attached to the top of
the inlet duct, by means of which diluted
material was introduced into the model at a
constant rate. t
Overflow from the central pipe, and later
from the collecting annular gutter, was
delivered to a large settling basin equipped
with a calibrated V-notch weir.
A point gauge in a manometer pot read
the level in the basin, thus determining the
discharge passing over the V-notch weir, and,
hence, the total discharge passing through the
separation chamber.
Solids Simulation
The swirl settler development tests were
being carried out at the same time that Beak
Consultants were conducting studies to
characterize the suspended solids in actual
sanitary sewage, and to identify a suitable
material to be used in the laboratory to
simulate this material. While Beak's work was
underway, first-stage testing was performed,
using shredded Petrothene®. This was
adequate to guide basic chamber shape
choices. When the result of Beak's
investigations became available, a new
material, an Anion Exchange Resin designated
as IRA-93, manufactured by Rohm and Haas,
was used as described in Section II.
Testing Procedure
Operation of the swirl concentrator as a
primary settler would normally involve a
continuously varying discharge, which can be
characterized by the average daily flow taken
as the design discharge. Since it was the
purpose of the model to adjust the chamber
dimensions to the discharge to be treated, the
range of the study was extended to cover
three different flows.
This process allowed observation of the
behavior of the separation chamber and the
variation of the recovery rate when conditions
of operation changed. It has already been
mentioned that the selected discharges used
for each layout were 0.5 I/sec (0.017 cfs),
1.0 I/sec (0.035 cfs) and 1.5 I/sec (0.053
cfs). These three series of results emphasized
the combined influence of the inlet and slot
velocities on the settleable separating process.
Different combinations of weir and skirt were
also tested.
For each individual test, the steady-state
discharge was established in the model and
equilibrium conditions established. One 1 (34
oz) of wet Petrothene, sg 1.01, grain size
ranging from 1 to 3 mm (0.039 to 0.118 in),
was injected into the supply pipe using the
same vibrating rate for all tests. The full liter
was added over a period of 5 minutes and the
model was allowed to run 10 minutes after
the end of the injection.
62
-------
The amount of Petrothene found on the
bottom of the cone or floating in the outer
chamber was measured separately. The
Petrothene floating in the settling basin was
also measured. The remaining portion
. deposited in the settling basin was found by
subtraction, assuming no material was lost.
The recovery rate was taken as the
percentage represented by the amount
measured on the bottom of the cone as
compared to the total found in the cone and
on the bottom of the settling basin.
SETTLEABLE SOLIDS
RECOVERY RESULTS
In discussing settleable solids for the
purpose of this section of the report,
reference is made to the recovery rates for
shredded Polythene® and Petrothene X. As
described in Section n, the shredded
Polythene and Petrothene X was considered
as representing organic materials over the
ranges as defined in the discussion. The
different steps followed in the model testing
are recorded in Table 15.
1. Tests Carried Out With
Circular Weir
a) Material Used for Testing: }Shredded
Polythene: Although the recovery rate was
encouraging (84 percent for 1 I/sec and a 1.2
cm [0.5 in] slot height), the first tests carried
out with the inlet sewer at the same level as
the outside floor in the chamber revealed the
existence of undesirable turbulence and under
the weir in both the outer and inner chambers.
Subsequent tests performed with the inlet
crown raised to the level of the weir lip, 22.9
cm (9 in) above the floor — and with other
conditions unchanged, as shown in Figure 29,
showed no. improvement in the recovery rate
(78%), as shown in Figure 3(X However, it
appeared evident that the turbulence created
in the chamber was mainly due to the bottom
of the skirt being just above the horizontal
floor or bench.
The flb^y tended to descend outside the
skirt, turn under the lip quickly, and roll up
inside, creating a degree of turbulence. This
turbulence entrained more suspended
particles in the effluent passing over the weir.
This led to the use of a smaller diameter
skirt which would sit lower in the cone. A 61
cm (24 in) diameter skirt, together with a
50.8 cm (20 in) diameter disc weir, as
portrayed in Figure 31, were selected and
tested. The whole series of tests and the results
are shown in Figure 32. The curves appear
and the results are shown in Figure 31,
Prototype Particle Sizes Simulated by
Shredded Petrothene. The curves appear
relatively uniform in that the recoveries
improved as the discharge decreased or the
retention time increased. On the other hand,
slot height tests were inconclusive at this stage
of the test program. The smaller diameter
skirt, sitting deeper inside the cone,
noticeably improved the recovery rate.
Test 27 carried out without a skirt, as
shown in Figure 33, showed a sharp drop in
the recovery rate (less than 50 percent) and
proved that the skirt was an essential part of
the separation chamber.
b) Tests Carried Out with Ground
Petrothene X: The results of tests 28 and 33,
shown in Figure 34, which were repeats of
tests 20 to 27, proved the consistency of the
results already obtained with shredded
Polythene, as shown in Figure 30.
Further tests carried out with 1.0 and 1.5
I/sec (0.035 and 0.053 cfs) discharges
showed that the recovery rate remained high
for small and intermediate flows, then
dropped slightly when the skirt was removed.
For the high flow 1.5 I/sec, the recovery rate
started in the 50 percent range, but showed a
progressive improvement as the slot opening
was increased.
Removal of the skirt proved to be less
significant in this case, as shown on the
right-hand side of Figure 34.
Visual observations of the flow pattern
obtained with a 50.8 cm (20 in) diameter
weir showed the existence of a central mass of
water around the central overflow pipe, under
the weir, representing a zone 5 cm (2 in) to
7.65 cm (3 in) wide, which apparently had
no flow-through discharge, was not rotating,
and in which very few sediment particles
could be detected. This observation led to
reducing the weir diameter. The extreme case
was selected first: that is, the weir was
63
-------
TABLE 15
MODIFICATIONS TESTED ON THE MODEL
Mod Ref.
Nbr. Figure
1 2
2 12
3 14
4 15
5 17
6 19
7 28
8 21
9 35
Tests Skirt Height Inlet Recovery
Number 0 0 Above Elevation Fig. Ref.
cm (in) cm (in) Chamber Floor
Material Used Shredded Polythene (up to Test 27) and Petrothene X
10 x 10 cm (4 x 4 in) Inlet Duct, 10 cm (4 in) flat chamber floor
11.0 [ 61.0 23.0 Invert at bottom
(28) I (24) (9) Elevation
1to19 71.0 : 61.0 23.0 Crown at 13
(28) i (24) (9) WeirElev.
20to45 61.0 51.0 23.0 Crown at 13 and 16
(24) (20) (9) Weir Elev.
27,28,34 61.0 51.0 23.0 Crown at 16
and 40 (24) (20) (9) WeirElev.
46 to 68 61.0 17.0 23.0 Crown at 18
(24) (6.625) (9) WeirElev.
69 to 91 61.0 38.0 23.0 Crown at 20
(24) ' (15) (9) WeirElev.
92to112 61.0 closed 23.0 Crown at 29
(24) bell (9) WeirElev.
113 to 136 61.0 38.0 33.0 Crown 10 cm 22
(24) (15) (13) (4 in) under
weir level
137 to 160 61.0 4 radial 23.0 Crown at 36
(24) gutters (9) Weir level
10 x 10 cm (4 x 4 in) Inlet Duct, Cone Extended to Chamber 0
10 37
10 37
161 to 184 71.0 4 radial 12.4cm Crown at 38
(28) gutters (4.875 in) Weir level
; above top
of cone
Material Used Anion Exchange Resin IRA-93
149 u d 74 u
6 x 6 cm (2.4 x 2.4 in) Inlet Duct, Cone Extended to Chamber 0
185 to 195 71.0 4 radial 12.4cm Crown at 39
(0) (10) (28) gutters (4.875 in) Weir level
above top
of cone ^
Noto: Test numbers In parentheses refer to second series started with fine material.
64
-------
TABLE 15 (continued)
Mod
Nbr.
Ref.
Figure
Tests
Number
Skirt
0
cm (in)
0
cm (in)
Height
Above
Chamber Floor
Inlet
Elevation
Recovery
Fig. Ref.
Central pipe overflow removed
6 x 6 cm (2.4 x 2.4 in) Inlet Duct, Cone Extended to Chamber 0
11 41 196 to 204 7.1.0 8 radial Weir lip 2 cm Crown 12.4 cm 47 and 49
(11) (19) (28) gutters (0.78 in) above (4.875 in) above
232 to 235 top of duct top of cone
(48) (51)
12 42 205 to 231 61.0 8 radial Weir lip 2 cm Crown 12.4 cm 47 and 49
(20) (47) (24) gutters (0.78 in) above (4.875 in) above
236 to 243 top of duct top of duct
(52) (59)
lOO-
se
I 90-
cc
HI
O 80-
0
HI
cc
70-
6O-
i
4
— —
<
/
i"""
»
— !
> <
[
i
""" "~~ i
-».
~~~ "H
)~*— --— .'
3 .
t
^-""'c
^_ '
____ t
!t«.
^
\
]
-
]
0.63cm 1.27cm 2.54cm 3.79cm 5.08cm
(0.25 in) (0.5 in) (1 in) (1.5 in) (2 in)
SLOT HEIGHT UNDER SKIRT
TESTS 5-9:QM= 1 I/sec (0.035 cfs), SKIRT 0.71 m (28 in) 0 • ®
TESTS 10-14: QM = 1.5 I/sec (0.052 cfs), SKIRT 0.71 m (28 in) 0
TESTS 15-19: QM = 0.5 I/sec (0.017 cfs), SKIRT 0.71 m (28 in) 0
O (2)
TESTS 20-27: QM = 0.5 I/sec (0.017 cfs), SKIRT 0.61 m (24 in) 0
MATERIAL USED: shredded polythene - nominal SG 1.01
— grain sizes 1-3 mm (0.039-0.118 in)
irregular shapes
A (3)
MOD
FIGURE 30 SWIRL PRIMARY SEPARATOR
OPERATING EFFICIENCIES FOR MODIFICATION 2 AND 3
65
-------
90°
0.91 m 0
(36 in)
°-61m0 Skirt
(24 in)
0.50m 0
(20 in)
Chamber
Weir
320° FOUL OUTLET
DISC WEIR
0
5.08cm (2 in)
L79 cm (1.5 in)
2.54 cm (1 in)
1.27cm (0.5 in)
/ 0.63 cm (0.25 in)
Section A-A
FIGURE 31 SWIRL PRIMARY SEPARATOR
MODEL LAYOUTS FOR TESTS 20 TO 45
MODIFICATION 3
66
-------
0.0
(0.0004)
PARTICLE DIAMETER, mm (in)
FIGURE 32 PROTOTYPE PARTICLE SIZES SIMULATED BY
SHREDDED PETROTHENE
67
-------
0.91 m 0 Chamber
(36 in)
0.50m 0 weir
(20 in)
320° FOUL OUTLET
-«^-
^->*--" _— -i?^
•
°° , A
PLAN
ELEVATION
Section A-A
FIGURE 33 SWIRL PRIMARY SEPARATOR
MODEL LAYOUTS FOR TESTS 27, 28, 34, AND 40
MODIFICATION 3 AND 4
68
-------
100
0.63 cm
(0.25 in)
1.27cm
(0.5 in)
WEIR 0.50m (20 in)
2.54cm 3.79cm 15.08 cm 12.70cm
(1 in) (.1.5 in) (2 in) (5 jn)
SLOT HEIGHT UNDER SKIRT
22.9 cm (9 in) HIGH, SKIRT 0.61 m (24 in)
TEST 28-33 : Qm = 0.5 I/sec (0.017 cfs)
TEST 34-39 : QM = 1 I/sec (0.035 cfs)
TEST 40-45 : Qm = 1.5 I/sec (0.052 cfs)
Q.-.
MATERIAL USED:
Ground Petrothene X - nominal sg 1.01
Grain sizes - 1-3 mm (0.039-0.118 in)
irregular shapes
FIGURE 34 SWIR L PR JIM ARY SEPARATOR
OPERATING EFFICIENCIES FOR MODIFICATION 3 AND 4
completely removed, leaving just the central
pipe opening, 16.8 cm (6.63 in) in diameter,
as shown in Figure 35. Furthermore, in order
to pinpoint the recovery fall-off point, a 12.7
cm (5 in) slot height was added to the five
already used.
Results obtained with the 16.8 cm (6.63
in) diameter weir showed a fairly constant 98
percent recovery rate for the 0.5 I/sec (0.15
gal/sec) flow. Greater discharges provided less
successful recoveries. The recovery rate
decreased as the slot height increased,
achieving very low values for higher openings,
as shown in Figure 36.
Observations showed that this reduced
recovery was caused by a rising helical flow
generated inside the conical chamber around
the central overflow pipe. As the flow
increased, the current stirred up sludge
sediment which had already deposited on the
bottom after sliding down the conical hopper
wall. The bottom deposit was caught by the
ascending current and carried up to the weir
lip and discharged with the overflow.
An intermediate solution was then tested,
using a 38.1 cm (15 in) diameter weir disc as
portrayed in Figure 37. No helical flow pattern
was observed. As shown in Figure 38, no
69
-------
0.89m, 0 Chamber
(35 In)
0.61 m 0 skirl
(24 in)
16.82cm 0 Weir
(6.67 in}'
90°
320° FOUL OUTLET
t.'1
L 22.9cm J
' P"
0.89 m
•f^^^MHV
0.61 m
(6.67 in)<
k
i
-12.70cm (5 in)
5.08cm (2 in)
3.79cm (1.5 in)
2.54 cm (1 in)
,1.27cm (0.5 in)
0.63cm (0.25 in)
i
ELEVATION
Section A-A
FIGURE 35 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT FOR TESTS 46 TO 68
MODIFICATION 5
70
-------
100
90
80
70
5?
CC
Ul
o
111
CC
60
40
30
2O
10
NO SKIRT CONDITIONS
iiiiiiiiiilimiiiniiiiiiiiHiitminuiiii!!
0.63cm 1.27cm
(0.25 in) (0.5 in)
WEIR 16.82 cm (6.67 in)
i i i i 11 i rTTniiimmmiiuiiiiiiiinminimiiiiiim
2.54cm 3.79 cm 5.08 cm 12.70cm
(lin) (1.5 in) (2 in) (Sin)
SLOT HEIGHT UNDER SKIRT
22.9 cm (9 in) HIGH, SKIRT 0.61 cm (24 in)
TESTS 46-52 : Qm = 0.5 I/sec (0.017 cfs)
TESTS 53-59 : Qm = 1.0 I/sec (0.035 cfs)
TESTS 60-68 : Qm = 1.5 I/sec (0.052 cfs)
MATERIAL USED: Ground Petrothene X - nominal sg 1.01
Grain sizes - 1-3 mm (0.039-0.118 in)
irregular shapes
FIGURE 36 SWIRL PRIMARY SEPARATOR
OPERATING EFFICIENCIES FOR MODIFICATION 5
71
-------
90°
0.91 m 0
(36 in)
0.61 m
(24 in)
0.31 m
(15 in)
Chamber
0 Skirt
0 Wetr
320° FOUL OUTLET
PLAN
FLAT DISC. WEIR
0.31m US")
12.70cm (Bin)
5.08cm (2 in)
3.79cm (1.5 in)
2.54 cm (1 in)
1.27cm (0.5 in)
0.63 cm (0.25 in)
ELEVATION
Section A-A
FIGURE 37 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT FOR TESTS 69 TO 91
MODIFICATION 6
72
-------
100
90
80
65 70
>•
oc
LU
0 60
HI
oc
50
40
D
No Skirt Conditions
i i i 11111 minim
0.63cm 1.27cm, 2.54cm 3.79 cm, 5.08 cm
(0.25 in) (0.5 in) (1 in) (1.5 in) (2 in)
SLOT HEIGHT UNDER SKIRT
12.70cm
(Sin)
FLAT DISK WEIR 0.31 m (15 in) 22.9 cm (9 in) HIGH, SKIRT 0.61 m (24 in)
TESTS 69-76 : QM = 0.5 I/sec (0.017 cfs) O-
TESTS 77-84 : QM = 1.0 I/sec (0.035 cfs) •-
TESTS 85-91 : Qm = 1.5 I/sec (0.052 cfs) D
MATERIAL USED: Ground Petrothene X - nominal sg 1.01
Grain sizes 1-3 mm (0.039-0.118 in)
irregular shapes
FIGURE 38 SWIRL PRIMARY SEPARATOR
OPERATING EFFICIENCIES FOR MODIFICATION 6
73
-------
change in the recovery curve occurred for the
small discharge of 0.5 I/sec (0.0187 cfs).
However, the two other curves showed a Very
interesting phenomenon, with the slot between
3.7 and 5.4 cm (1.5 and 2 in). High recovery
rates were not less than 94 percent for
the three discharges, then dropped off with
greater or smaller slots. (
c) Recovery Rate Versus Height of Weir
Above the Chamber Floor: In ordejr to
pinpoint the influence of weir height above
the chamber floor, results were determined
for the 38.1 cm (15 in) weir disc. At 23 cm
(9 in) above the floor, as shown in Figure 38,
the recovery curves diverged for the three
discharges considered except in the slot range
3.7 to 5.4 cm (1.5 to 2 in), where recovery
rates were not less than 94 percent. ;
Raising the weir lip to 33 cm (13 in)
from the floor, as shown in Figure 39,
resulted in flattening the recovery curves and
brought them closer to the 0.5 I/sec (O.pl87
cfs) flow efficiency which stayed constant at
98 percent, as shown in Figure 40. Hence, for
slots of 5 to 12.7 cm (2 to 5 in) high,
recovery rates were not less than 97 percent.
Recovery dropped off for greater or smaller
slots.
' d) Flow Velocities Inside the Chamber: A
series of measurements was carried out with
the 38.1 cm (15 in.) weir disc, with
configuration as shown in Figure 29. Flow
velocity was taken inside the chamber by
means of a midget current meter successively
set up at positions shown on Figure 41.
Measurements were made with a model flow
discharge of 1 I/sec (0.035 cfs), a 2.5 cm
(1 in) slot, and a 61 cm (24 in) skirt.
Results plotted are the horizontal
tangential component of the flow velocity.
Direction of flow is assumed going into the
paper. Interpolation between measured
positions yielded the velocity contours shown
on Figures 42 through 45.
e) 38 cm (15 in) Weir Disk-.
At 0° position, close to the sewer
inlet outside the skirt, flow velocity ranged
from 25.9 cm/sec (0.85 ft/sec) at the bottom,
to 36.6 cm/sec (1.2 ft/sec) at the surface. As
the flow progressed to the different positions,
it showed a slight decrease of the surface
velocity accompanied by a slight increase
at the bottom.
Momentum inside the skirt progressively
lessened from the slot inlet to the free
surface. Under the weir itself, velocity contours
were featured by flow patterns simulating
rotating concentric cylinders whose velocity
increased with the distance from the central
downdraft pipe.
2. Tests Carried Out with Closed
Bell and Four Orifices
In order to avoid concentration of flow in
the area extending between the weir lip and
skirt, a new weir was tested consisting of a
closed bell with four orifices of 3.8 cm (1.5
in) diameter, as shown in Figure 46 for Tests
92 to 112. Results obtained are presented in
Figure 47. They showed a fairly constant
recovery rate for the three discharges
considered over the 5 cm to 33 cm (2 to 13
in) slot height range. Recovery efficiency for
the largest discharge of 1.5 I/sec (0.053 cfs)
proved to be only two percent below the
intermediate and low flow performances.
a) Flow Velocity Inside the Closed Bell
Chamber with Four Orifices 3.8 cm (1.5
in): Flow velocities were taken inside the
chamber at positions shown in Figure 48.
Flow contour patterns are shown in Figures
49 to 52. The same phenomenon as described
above occurred in the chamber outside the
skirt although it was less pronounced. Within
the skirt, high velocities concentrated under
the orifices with a marked tendency to
increase the rptating effect 24.4 cm/sec (0.8
ft/sec) at 180 degrees. The same concentric
layers with increasing velocity outward from
the central downdraft pipe existed, with
maximum velocities present under the
orifices.
Since the closed bell with orifices
improved the recovery rate, and flattened the
curves at the highest values, it was expected
that better results could be achieved if the'
orifices were placed so as to produce a better /
radial flow distribution. That was to be the
next step in the study approach, but a radial
weir gutter was suggested during discussions
in a team meeting as more likely to provide
better upward flow distribution inside the
skirt. If successful, the new device would
avoid both an undesirable concentration of
flow on the one hand, and helical ascending
currents on the other. It would, however,
74
-------
*-A
90°
0.91 m 0 Chamber
(36 in)
°-61 m 0 Skirt
(24 in)
0.38 m 0 Weir
(15 in)
320° FOUL OUTLET
;
15
rr>
to
ti
i
1 -•*;
ll
1
\
60°)i
r
^-^
0.91 m
;/
.^--^^
[o°(36in)_^A
0.61m (24") _
0.38m
-^
i
i
i
fk
\
16.8
(6.6'
I
(15 in) I
ljr/"
f
cm
Mn)
I
*—
^^
^^ )
^^ m
^
~
' \
&&::?
PLAN
^FLAT DISC WEIR
^ 0.38 m (15 in)
'.-s
u
::2g 1^20.32 cm (Sin)
?•— ° * 12 70 cm (Bin)
1 f j ^______ — -5.08 cm (2 in)
i: p ^3.79 cm (1.5 in)
^- T i ~~ — 2.54 cm (1 in)
10.161 ^XSi^? cm (0.5 in)
cm r* ^-0.63 cm (0.25 in)
(4 in)
ELEVATION
Section A-A
FIGURE 39 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT FOR TESTS 113 TO 136
MODIFICATION 8
75
-------
69
CC
LU
O
111
CC
100
90
80
70
60
50
40
30
nzrrf^-ffl .fr1^
-•— •• *>->-.—
No Skirt Conditions
niiiiiiiiiiiniiiniiiiiiiwuTrTTrninTnnniimi'
0.63 cm 1.27 cm 2.54 cm 3.79 cm5.08 cm
(0.25 in) (0.5 in) (1 in) (1.5 in) (2 in)
SLOT HEIGHT UNDER SKIRT
rnTiT
12.70cm 20.32cm 43.18cm
(5 in) (Sin) (17 in)
FLAT DISC WEIR 0.38 m (15 in) 0.33 m (13 in) HIGH, SKIRT 0.61 m (24 in)
TESTS 113-119,135 : QM = 0.5 I/sec (0.017 cfs) O
TESTS 120-126,136 : QM = 1.0 I/sec (0.035 cfs) •
TESTS 127-134 : Qm = 1.5 I/sec (0.052 cfs) O
MATERIAL USED: Ground Petrothene X - nominal sg 1.01
i Grain sizes — 1-3 mm (0.039-0.118 in)
1 irregular shapes
FIGURE 40 SWIRL PRIMARY SEPARATOR
OPERATING EFFICIENCIES FOR MODIFICATION 8
76
-------
2.74m
Oft)
SCALE: 1 cm = 0.48 m
(1in = 4 ft)
FIGURE 41 SWIRL PRIMARY SEPARATOR LOCATION OF
LOCATION OF MEASURING POINTS FOR VELOCITY
CONTOUR TESTS 10.98 m (36 ft) CHAMBER
PROTOTYPE SCALE 1/12
MODIFICATIONS
77
-------
NOTE: Velocities in ft/sec (to get cm/sec, multiply by 30.5)
FIGURE 42 SWIRL PRIMARY SEPARATOR
TANGENTIAL VELOCITY CONTOURS AT 0° POSITION
10.98 m (36 ft) CHAMBER
DISCHARGE = 0.49 m3/sec (17.6 cfs)
PROTOTYPE SCALE 1/12
MODIFICATION 6
NOTE: Velocities in ft/sec (to get cm/sec, multiply by 30.5)
FIGURE 43 SWIRL PRIMARY SEPARATOR
TANGENTIAL VELOCITY CONTOURS AT 90° POSITION
10.98m (36:ft) CHAMBER
DISCHARGE = 0.49 m3/sec (17.6 cfs)
PROTOTYPE SCALE 1/12
MODIFICATION 6
78
-------
i
NOTE: Velocities in ft/sec (to get cm/sec, multiply by 30.5)
FIGURE 44 SWIRL PRIMARY SEPARATOR
TANGENTIAL VELOCITY CONTOURS AT 180° POSITION
10.98m (36 ft) CHAMBER
DISCHARGE-0.49m3/sec (17.6 cfs)
PROTOTYPE SCALE 1/12
MODIFICATIONS
T.
NOTE: Velocities in ft/sec (to get cm/sec, multiply by 30.5)
FIGURE 45 SWIRL PR I MARY SEPARATOR
TANGENTIAL VELOCITY CONTOURS AT 270° POSITION
10.98 m (36 ft) CHAMBER
DISCHARGE = 0.49 m3/sec (17.6 cfs)
PROTOTYPE SCALE 1/12
MODIFICATION 6
79
-------
0.91 m 0 Chamber
(36 ft)
0.61 m 0 Skirt
(24 in)
4-3.79 cm (1.57-1.49 in) 0
Weir Orifices
12.7cm (5 in)
5.08cm (2 in)
3.79cm (1.5 in)
2.54cm (1 in)
1.27cm (0.5 in)
0.63cm (0.25 in)
ELEVATION
Section A-A
FIGURE 46 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT FOR TESTS 92 TO 112
MODIFICATION 7
80
-------
cc
o
LU
cc
100
90
80
70
60
50
40
30
O
-a"
=§=-
—~ D —
No Skirt Conditions
. D
•imiiiiiiiiiiiiiiiiiiiiiMiiiimiiiiiMiiiMiiiiiiiii
_LU
t«*MMWMMMMMMmaM
0.63cm 1.27cm 2.54cm 3.79 cm 5.08 cm 12.70cm 33.02cm
(0.25 in) (0.5 in) (1 in) (1.5 in) (2 in) (5 in) (13 in)
SLOT HEIGHT UNDER SKIRT
TESTS 92-98 : Qm = 0.15 I/sec (0.005 cfs)
TESTS 99-105 : Qm = 1.0 I/sec (0.035 cfs)
TESTS 106-112 : Qm = 1.5 I/sec (0.052 cfs)
O-
FIGURE 47 SWIRL PRIMARY SEPARATOR
OPERATING EFFICIENCIES OF CLOSED BELL
WITH FOUR ORIFICES - 3.79 cm (1.5 in) 0,
22.9 cm (9 in) HIGH, SKIRT 0.61 m (24 in) 0
FOR MODIFICATION 7
-------
2.74m
Oft)
(1ft) (2 ft) (2ft) 1ft
FIGURE 48 SWIRL PRIMARY SEPARATOR
LOCATION OF MEASURING POINTS FOR VELOCITY
CONTOUR TESTS 10.9? m (36 ft) CHAMBER
CLOSED BELL WITH FdUR ORIFICES - 0.457 m (1.5 ft) 0,
2.74 m (9 ft) HIGH, SKIRT 7.32 m (24 ft) 0
PROTOTYPE SCALE 1/12
MODIFICATION?
82
-------
NOTE: Velocities in ft/sec (to get cm/sec, multiply by 30.5)
FIGURE 49 SWIRL PRIMARY SEPARATOR
TANGENTIAL VELOCITY CONTOURS AT 0° POSITION
10.98 m (36 ft) CHAMBER
DISCHARGE = 0.49 m3/sec (17.6 cfs)
PROTOTYPE SCALE 1/12
MODIFICATION?
83
-------
NOTE: Velocities in ft/sec (to get cm/sec, multiply by 30.5)
FIGURE 50 SWIRL PRIMARY SEPARATOR
TANGENTIAL VELOCITY CONTOURS AT 90° POSITION
10.98m (36 ft) CHAMBER
DISCHARGE = 0.49 m3/sec (17.6 cfs)
PROTOTYPE SCALE 1/12
MODIFICATION?
84
-------
NOTE: Velocities in ft/sec (for cm/sec, multiply by 30.5)
FIGURE 51 SWIRL PRIMARY SEPARATOR
TANGENTIAL VELOCITY CONTOURS AT 180° POSITION
10.98 m (36 ft) CHAMBER
DISCHARGE = 0.49 m3/sec (17.6 cfs)
PROTOTYPE SCALE 1/12
MODIFICATION 7
85
-------
NOTE: Velocities in ft/sec (for cm/sec, multiply by 30.5}
FIGURE 52 SWIRL PRIMARY SEPARATOR
TANGENTIAL VELOCITY CONTOURS AT 270° POSITION
10.98 m (36 ft) CHAMBER
DISCHARGE = 0.49 m3/sec {17.6 cfs)
PROTOTYPE SCALE 1/12
MODIFICATION 7
86
-------
introduce obstacles into the free motion flow
at least at the surface.
Tests Carried Out with Radial Weir Gutters
a) Flat Chamber'Flo'of 10'cmT4 in) Wide
with Four Weir Gutters: Gutters were 7 cm
wide by 5 cm high (2.75 in by 2 in), inside
dimension. They extended radially 29.2 cm
(11.5 in) from the center line of the chamber.
The gutter crests were _set horizontally 23 cm
(9 in) above the chamber floor, as depicted
in Figure 53.
b) Recovery Rate Vs Slot Height: Figure
54 presents the results of the study of recovery
rates vs slot height. Results were disappointing,
considering previous recovery rates obtained
with the closed bell and 4 orifices 3.8 cm (1.5
in) diameter. The curves converged at the 5
cm (2 in) slot where the recovery rate was 98
percent for the 0.5 and 1.0 I/sec (0.0187 and
0.0375 cfs) discharges, and 95 percent for the
1.5-1/sec (0.0562 cfs) discharge. The efficiency
dropped off for greater or smaller slots.
c) Conical Bottom Extended up to
Chamber Wall: The 10 cm (4 in) flat chamber
floor utilized in the studies up to test 160
delayed the fall of particles into the conical
bottom zone as they entered the chamber.
The model showed that Petrothene grains
described an angular trajectory of about
90 degrees, after leaving the inlet duct before
crossing the edge of the flat floor. This
phenomenon produced a substantial
concentration of grains in the 90 degree to
270 degree area, which hampered the free
falling motion of grains. This was particularly
true for small slot heights.
In addition to this, some deposits formed
on the flat chamber floor, about 45 degrees
from the inlet when small discharges 0.5
I/sec (0.0187 cfs) were tested. To avoid .
these difficulties the flat portion of the
chamber: floor was eliminated and the conical
section was extended up to meet the full
chamber diameter. The inlet bottom was then
set 2.2 cm (0.875 in) above the cone top, as
shown in Figure 55.
To remain consistent with the previous
test approach, the crests of the four gutter
weirs were set at the same elevation as the
inlet top, as shown in Figure 53.
d) Increase of the Skirt Diameter: By
taking advantage of the new chamber floor, a
larger skirt could be built in order to increase
the size of the internal chamber. Subsequent
tests were performed with a 71 cm (28 in)
diameter skirt, thus reducing the width of the
external channel to 10 cm (4 in).
The slot height was measured as usual
with respect to the level of cone
corresponding to 71 cm (28 in) diameter, as
shown in Figure 56.
e) Recovery Rate Vs Slot Height: These
new conditions gave fairly good results, since
flattening of the recovery curves was greatly
improved. As shown in Figure 57, the
efficiency corresponding to small (0.5 I/sec
[0.0187 cfs]) and intermediate (1 I/sec
[0.0375 cfs]) discharges reached to the 99
percent level for the whole range of slot
heights, including no-skirt conditions. The
recovery rate for high discharge (1.5 I/sec
[0.0562 cfs]) dropped for slot heights under
2.5 cm (1 in), then remained constant at the
98 percent level up to the 20.3 cm (8 in)
slot height. Removing the skirt caused a drop
of two percent, bringing the efficiency rate
down to 96 percent. Figure 58 shows the
conditions involved.
FINE GRAIN RECOVERY
TEST PROCEDURE
Anion Exchange Resin IRA-93 with grain
size \49n>d>74iJ. (passing No. 100 sieve and
retained on No. 200) was used to represent
settleable solids. An improved measuring
procedure was developed with the assistance
of Beak Consultants, Ltd., for the subsequent
tests.
The principle involved in this phase of the
model study was to treat the same amount of
water (1,800 1 [478 gal]) with a constant
influent concentration of 38 ppm, at different
discharges and to analyze samples of the
overflow liquor.
Selected discharges were, respectively,
0.1, 0.3, 0.5, 0.75 and 1 I/sec (0.0035,
0.010, 0.018, 0.026, and 0.035 cfs). Constant
solids concentration was obtained by injecting
3 cc of material at equally spaced intervals of
time depending upon the discharges. 100-cc
samples of overflow and pump inlet were
regularly taken so as to provide two
composite samples. Sampling operations
began after four retention times had elapsed,
as depicted in Figure 59, and made at the
87
-------
0.91m 0 Chamber
(36 in)
0.61 m0 Skirt
(24 in)
10.16x10.16 cm INLET
(4x4 in)
20.32 cm (8 in), 12.70 cm (5 in)
r5.08 cm ( 2 in), 3.79 cm (1.5 in)
2.
54 cm (1 in), 1.27 cm (0.5 in)
0.63 cm (0.25 in)
ELEVATION
FIGURE 53 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT FOR TESTS 137 TO 160
MODIFICATION 9
88
-------
ITTTIIIIIIIIIIIHLUIIIIIIlIiniDMIillillllllllllll I I I I I I I III
65
>•
EC
LU
8'
in
QC
100
90
80
70
60
50
40
30
cr
No Skirt Conditions
0.63 cm
(0.25 in)
1.27 cm 2.54 cm,3.79 cm 5.08 cm
(0.5 in) (1in) (1.5 in) / (2 in)
12.70 cm ,20.32 cmi 33.02 cm
(5 in) (8 in) (13 in)
SLOT HEIGHT UNDER SKIRT
4 RADIAL GUTTER WEIRS 7.62 cm (3 in) x 5.08 cm (2 in) - 22.9 cm (9 in) HIGH
SKIRT 0.61 cm (24 in)
TESTS 137-144 : Qm = 0.5 I/sec (0.017 cfs) O
TESTS 145-152 : Qm = 1.0 I/sec (0.035 cfs) «
TESTS 153-160 : Qm = 1.5 I/sec (0.052 cfs) b
MATERIAL USED: Ground Petrothene X - nominal sg 1.01
Grain sizes — 1-3 mm irregular shapes
FIGURE 54 SWIRL PRIMARY SEPARATOR
OPERATING EFFICIENCIES FOR MODIFICATION 9
89
-------
270°
0.91m 0 Chamber
(36 in)
0.71m 0 Skirt
(28 in)
I 10.16 x 10.16 cm INLET
V (4x4 in)
2.21 cm (0.87 in)
Former Chamber Floor
ELEVATION
FIGURE 55 SWIRL PRIIVIARY SEPARATOR
MODEL LAYOUT FOR TESTS 161 TO 184
MODIFICATION 10
90
-------
90-
80-
70-
60-
>- so-
ar
UJ
8 40-
UJ
ar
30-
20-
10-
QDm
Qn CfS
O.I
0.2
0.3 0.4 . 0.5
I |
0.6
0.7
0.8
0.9
1.0
I
10
15 20
Discharge
25
30
FOR MODEL CONDITIONS SEE FIGURE 37
SLOT HEIGHT 5.08 cm (2 in)
Material Used = Anion Exchange Resin IRA-93
Grain Size = !49/£>cl > 74/i
• • Results obtained by volumetric measurements of recovered material
A----A Results obtained by sample analysis done by Beak
FIGURE 56 SWIRL PRIMARY SEPARATOR
OPERATING EFFICIENCIES FOR MODIFICATION 10
Increase of Skirt Diameter
91
-------
oc
o
o
Ul
100
90
80
70
60
50
40
30
ruTiiiiinrrrnTTTrnniiminmn—
IKIMUI i i i mTTrmTm'ii
D
No Skirt Conditions
I i i i i i 111
0.63 cm
(0.25 in)
1.27 cm 2.54; cm 3.79 cm 5.08 cm
(0.5 in) (1in) (1.5 in) ! (2 in)
12.70 cm j20.32 cm 30.48 cm
(Sin) (Sin) (12 in)
SLOT HEIGHT UNDER SKIRT
4 RADIAL GUTTER WEIRS 7,62 cm (3 in) x 5.08 cm (2 in) LIP AT TOP OF
INLET - SKIRT 0.71 m (28 in) (£> (SEE FIGURE )
TESTS 161-168 : Qm = 0.5 I/sec (0.017 cfs) O-
TESTS 169-176 : Qm = 1.0 I/sec (0.035 cfs) •-
TESTS 177-184 : Qm = 1.5 I/sec (0.052 cfs) a-
MATERIAL USED: Ground Petrothene X - nominal sg 1.01
'Grain sizes — 1-3 mm (0.039-0.118 in)
irregular shapes
i
FIGURE 57 SWIRL PRIMARY SEPARATOR
OPERATING EFFfclENClES FOR MODIFICATION 10
Recovery Rate versus Slot Height
92
-------
Petrothene grains reaching end
of 10 x 10 cm (4 x 4 in) inlet
Modification 10
Petrothene grains sliding down
against the cone wall near inlet
Modification 10
FIGURE 58 MODIFICATION 10, PETROTHENE GRAINS REACHING END OF INLET,
AND PETROTHENE GRAINS SLIDING DOWN AGAINST THE CONE WALL
NEAR INLET
93
-------
e
E
*-
12-
II-
10-
9-
8-
7-
6-
5-
4-
3-
2-
I-
E
a.
40-
30-
20-
10-
rr~ Inside 71 cm (28 in) Skirt
Inside 0.61 m (24 in) Skirt
Discharge Model
QpCfs
Q0mgd
0.5 1.0 1.5
DISCHARGE PROTOTYPE SCALE 1/12
O.I 0.2 0.3 : 0.4 0.5 0.6 0.7 0.8
10
15
20
25
30
I
i 1 I
5 i 10 15
RETENTION TIME WITH CENTRAL OVERFLOW PIPE
RETENTION TIME WITHOUT CENTRAL OVERFLOW PIPE
20
FIGURE 59 SWIRL PRIMARY SEPARATOR
RETENTION TIME Vs DISCHARGE WITH SCALE 1/12
94
-------
rate of one 100-cc sample every two
injections.
The model was left running without
injection for one more retention time, then
stopped. In addition to collecting samples
which were sent to Beak for concentration
analyses, volumetric measurements of the
material respectively deposited in the
chamber and overflow tank were also carried
out. The volumetric measurements proved to
be too time-consuming and they were later
discontinued.
A new test numbering series was started
for studies with the fine material. Test 1 in
the new series corresponded to Test 185 in
the old series. Both numbers are given in the
resume in Table 15, whereas only the new
series is referred to in Table 16.
Original Model Conditions
The first series of tests with fine material
was carried out with the layout shown on
Figure 55. This configuration provided the
following details:
• The conical section extended up to
the chamber diameter, eliminating the
flat floor.
• Four radial weir gutters, 7 cm (2.75
in) wide by 5 cm (2 in) high - inside
dimensions — extended radially 29.2
cm (11.5 in) from the chamber axis.
« The weir lip was located at inlet
crown elevation.
• The skirt was 71.1 cm (28 in) in
diameter.
• The central overflow cylinder had an
outside diameter of 16.8 cm (6.87
in).
To provide uniform comparable results,
the slot height under the skirt for this series
of tests was set at 5.1 cm (2 in), the value
which gave the best recovery rate for
Petrothene with discharges ranging from 0.5
to 1.5 I/sec (0.98 to 0.0525 cfs) as shown in
Figure 54.
The results of this second series of tests
are recorded in Table 16.
Fine Grain Recovery
Results and Comments
As shown on Figure56 , the recovery of
fine settleable solids dropped progressively
with increasing discharge rates. Such results
were expected since the time allowed for a
TABLE 16
SUCCESSIVE MODIFICATIONS OF THE MODEL AND
RECOVERY RESULTS WITH ANION EXCHANGE RESIN
IRA 93 149M>d>74M
Stage No. Mod.
Skirt
Inlet Size
1 Mod. 10
71. 1cm (28 in)
10x 10cm
(4x4 in) Inlet
2 Mod. 11
71. 1cm (28 in)
6 x 6 cm
(2.37 x 2.37 in)
Inlet
Other
Model
Features
4 radial wiers
Central column
No deflector
8 radial weirs
gutter in outer ch
No central column
No deflector
Discharge Qm l/s and O^ mgd
Qm 0.1 0.3
Op 1.14 3.42
91 62
(D ©
96 52
rtfft (2^
0.5 0.75 1.0
5.69 8.54 11.38
52 32 30
0(D ® (D
40 31 27.5
vJ vJ J)
95
-------
TABLE 16 (continued)
Discharge Qm l/s and Op nngd
Stage No. Mod.
Skirt
inlet Size
2aMod. 11
71. 1cm (28 in)
6x6 cm
(2.37 x 2.37 in)
Inlet
3 Mod. 12
61 cm (24 in)
6x 6cm
(2.37 x 2.37 in)
Inlet
3aMod. 12
61 cm (24 in)
6x 6cm
(2.37 x 2.37 in)
Inlet
SbMod. 12
61 cm (24 in)
6x 6cm
(2.37 x 2.37 in)
Inlet
3cMod. 12
61 cm (24 in)
6x 6cm
(2.37 x 2.37 in)
Inlet
3d Mod. 12
61 cm (24 in)
6x 6cm
(2.37 x 2.37 in)
Inlet
3iMod. 12
61 cm (24 in)
6x 6cm
(2.37 x 2.37 in)
Inlet
other
Model
Features
Same as 2
22.9 cm (9 in)
emerging
deflector
8 radial weirs
gutter in outer ch
no central column
no deflector
Same as 3
22.9 cm (9 in)
Emerging def.
to bottom of duct
Same as 3
radial deflector
Top of duct
to top of cone
Same as 3
24.1 cm (9.5 in)
long deflector
Top of duct to
top of cone
Same as 3
24.1 cm (9.5 in)
long deflector
Top of duct to
bottom of duct
Same as 3
Draw off
underflow 5%
Qm 0.1 0.3 0.5 0.75 1.0
Qp 1.14 3142 5.69 8.54 11.38
87 54 37
! H) ty ©
33 © 23 @>
81 (26- 64 (fjl) 30 @> 66.8 © 33.7 (f§)
65.3 @ 41 @ 40 @
90.5 36 @
i (58 53 ©
40 ®l
56 42 @
@) 52 @
42 @
55 35 27
® © HH
49.5 38 37.4
@ @* @
51.6 43.7 30T.5
-*• y — v /"" — S / — "\
,44' @ @ (59)
i 44 23
; Qy ^y
Note: Porcantnge recoveries given in plain typed figures'in body of table; i.e. 44.
Tost numbers in second series shown in circles; i.e. E24)
96
-------
particle to settle shortened progressively as
the discharge increased, as portrayed in Figure
59.
Efficiency rates determined from
volumetric measurements decreased from 91
percent for 0.1 I/sec (0.0035 cfs), to 24
percent for 1 I/sec (0.035 cfs). Recovery
values derived by Beak from the sample
analyses were slightly higher, as shown by the
dotted line .in Figure 59.
This can be explained by the difficulties
encountered in recovering the finer material
for the volumetric measurements. Part of the
solids material was lost during each test
despite all precautions taken. The loss
probably represents the portion of the
material finer than No. 200 sieve (74
microns), but which was retained as dust in
the sample during the original dry sieving.
The results of the concentration analyses
performed by Beak must be accepted as
providing the better representation of the
recoveries.
MODIFICATION OF THE MODEL
FOR FINE RECOVERY - STAGE 2
The next steps taken in the attempt to
increase the efficiency of the swirl chamber
consisted of:
1) Removing the central cylinder used for
overflow evacuation.
2) Increasing the number of radial weir
gutters to 8, so as to better distribute the
flow inside the chamber.
3) Increasing the inlet velocity to avoid
solids deposits in the inlet.
The new structural details of the model
are shown in Figure 60. They include:
• 8 Radial Weir Gutters were provided,
4 cm (1.6 in) high by 2 cm (0.8 in)
wide with crest at 2 cm (0.8 in) above
the inlet crown elevation.
• The inlet was reduced to a 6 x 6 cm
(2.35 x 2.35 in) square duct, with top
elevation at the same level with
respect to the end of cone wall as in
previous conditions.
• The .skirt was 71 cm (28 in) in
diameter.
• The slot height was 5 cm (2 in).
• The overflow was discharged over a
circular gutter 8 cm (3.5 in), high by
3 cm (1.2 in) wide fixed around the
outside of the skirt and
communicating with the overflow
tank by two .2.5 cm (1 in) diameter
tygon tubes were passing through the
chamber wall.
TEST RESULTS AND COMMENTS
(MODIFICATION 11) - STAGE 2
Tests carried out under the new
conditions with the same material proved to
be disappointing. The recovery rate varies
unevenly, showing some gain for the small
discharge of 0.1 I/sec (0.0035 cfs) but a loss
under all other operating conditions.
This was probably due to the velocity
increase which occurred in the outer chamber,
resulting from the simultaneous reduction of
the inlet duct cross section and of the outer
chamber volume due to the presence of the
circular gutter for collecting the overflow.
A similar test run with the 0.5 I/sec
l(0.018 cfs) discharge, and Arizona Road Dust.
with grain size 20/jL^d^ 10;rarid SG-2.65,
showed a relatively minor decline of the
recovery rate — 36 percent instead of the
previous 40 percent obtained with IRA-93
Anion Exchange Resin.
Additional tests carried out with a 22.86
cm (9 in) long, deflector at the inlet end
(Stage 2a) did not prove to be conclusive in
the improvement of the recovery rate.
MODIFICATION 12 - (STAGE 3)
In order to decrease the velocity
prevailing in the outer chamber, the skirt
diameter was reduced to 61 cm (24 in),
giving a ratio D^/D (skirt to chamber
diameter) equal to 2/3. The radial cross
section of the outer chamber was, thus,
increased by about 50 percent, as shown in
Figure 61.
Results obtained when tests were repeated
did not always yield the same values. Plotted
points were widely scattered, especially for
the 0.5 I/sec (0,18 cfs) dishcarge. In this
zone, the repeated values ranged from 30 to
50 percent. Conversely, at the 0.3 I/sec
97
-------
B
Skirt
3 cm (1.39 in) Wide
Circular Gutter
8 cm (3.13 in) High
Circular Gutter
6cm 0 Pipe
(2.34 in)
6 x 6 cm Inlet
(2.34 x 2.34 in)'
Skirt 0.7I m 0
«28")
6x6cm Inlet
(2.34 x 2.34 in)
6.35 cm 12.5 in)
11.45 cm (4.5 in)
FIGURE 60 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT FOR MODIFICATION II
98
-------
Overflow
_0.91_m.j36jn).
0.71 m (28 in)
Skirt 0.71m I
(28 in)
.^.Overflow
SECTION B-B
FIGURE 60 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT FOR MODIFICATION II
99
-------
3cm (13x16 in)
WIDE
CIRCULAR
GUTTER
SKIRT
6 x 6 cm Inlet Duct
(2.34 x 2.34 in)
8 cm
(3.13 in) HIGH
CIRCULAR
GUTTER
FIGURE 61 SWIRL PR I MARY SEPARATOR
MODEL LAYOUT FOR MODIFICATION 12
100
-------
(0.011 cfs) level, results showed a relative
consistency at 65 percent.
In an attempt to improve the recovery
rates by slowing down the flow velocity in the
outer chamber, various baffles were tested at
the inlet. The first baffle tested was set in
between the outside chamber wall and the
circular collector gutter on the skirt, as shown
in Figure 62. As listed in Table 16, this
produced a slight improvement in recoveries.
However, it was found that having the baffle
emerge from the water, stopping the surface
flow, also impeded the floatables movement,
leaving them randomly scattered around the
outer chamber.
A submerged baffle, as shown in Figure
63, was tried next. This radial arrangement
decreased the recoveries for all discharges. A
layout with the same cross section as 3b, but
with the baffle tangential to the skirt was
next tested, as shown in Figure 64. This
provided 'an improvement, but still did not
perform the recovery achieved with Baffle 3a.
Finally, the configuration shown in Figure
65 was tested. Results from these tests were
inconclusive, showing slightly less recovery
for 0.3 I/sec (0.011 cfs) and slightly more for
0.5 I/sec (0.018 cfs) than for baffle 3a.
However, since baffle 3d combined fairly
satisfactory recoveries and the advantage of
leaving the surface uninterrupted, it is the
recommended form.
INFLUENCE OF CONTINUOUS
UNDERFLOW DRAW-OFF
Only two tests were carried out with a
continuous discharge of deposited slurry
being drawn off through the foul outlet.
These tests were performed with total inflow
discharges of 0.3 and 0.5 I/sec (0.011 and
0.018 cfs), respectively. In each case, 5
percent of the inflow volume was withdrawn
through the foul outlet. The model layout
used was that for Modification 12, without
any baffles, as shown in Figure 61.
The results, as outlined in Table 16, for
Tests 24 and 25, showed a distinct drop in
recoveries' when compared with the tests
achieved without any foul draw-off.
Observations on the model indicated that the
discharge going to the foul outlet catalyzed
increased velocities in the lower regions of the
conical bottom. These, in turn, were
sufficient to resuspend particles which had
been settled out along the floor higher up in
the cone. Once in suspension for the second
time, a greater portion of the particles was
intrained upward and carried over the gutter
weirs with the chamber overflow liquid.
Comments
Analysis of the recovery values for the
various layouts tested in the fine grain
materials series given in Table 16 shows
considerable random scatter. Some of this was
obviously due to experimental error, but the
time was not available to carry out repeat
tests in many cases.
It was necessary, therefore, to decide
subjectively on the recovery curve which best
described the structure's operation in the
laboratory. The one selected is shown on
Figure 66. This curve was developed graphically
as representing the best mean of the data
points for the tests performed with baffles.
In fact, the test results were not decisive
enough to allow an absolute selection of the
best baffle configuration. However, from
observations on the model it appeared that
the Type 3d layout presented certain
advantages, so this one is recommended for
prototype testing.
PREDICTED PROTOTYPE
SOLIDS RECOVERY
As discussed in the preceding section of
this report, recovery curve evolved for the
IRA-93 in the model studies is as shown in
Figure 66. For the five model discharges used
iin the tests, the corresponding recovery rates
are given in Table 17.
TABLE 17
RECOVERY RATE OF IRA-93
AT VARIOUS FLOWRATES
Model Discharge:
I/sec ' 0.1 0.3 0.5 0.75 1.0
cfs 0.0035 0.0106 0.0176 0.0264 0.0352
Recovery Rate:
% 91 58 41 31 27
In order to transpose this model curve to
prototype, reference was then made to Figure
101
-------
6x 6cm
(2.34 x 2.34 in)
Inlet Duct
3 cm wide
circular
gutter
(1.18 in)
Skirt
8 cm high
circular
gutter
(3.15 in)
DEFLECTOR
type 3 a
FIGURE 62 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT WITH MODIFICATION 12 AND INLET
! BAFFLE TYPE 3a
102
-------
6x 6cm
(2.34x2.34 in)'
Inlet Duct
t
3 cm wide
circular
gutter
(1.18 in'
Skirt
DEFLECTOR
PLAN / type 3b
8 cm high
circular
gutter
FIGURE 63 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT WITH MODIFICATION 12 AND INLET
BAFFLE TYPE 3b
103
-------
B-*
6 x 6 cm
(2.34x2.34 in)
Inlet Duct
3 cm wide
circular
gutter
(1.18 in)
Skirt
8 cm high
circular
gutter /
(3.15 in)
FIGURE 64 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT WITH MODIFICATION 12 AND INLET
BAFFLE TYPE 3c
104
-------
B -*
6x 6cm
(2.34x2.34 in)
Inlet Duct
t
3 cm wide
circular
gutter
(1.18 in)
Skirt
8 cm high
circular
gutter
(3.15 in)
DEFLECTOR
PLAN type 3
6cm
(2.34 in)
FIGURE 65 SWIRL PRIMARY SEPARATOR
MODEL LAYOUT WITH MODIFICATION 12 AND INLET
BAFFLE TYPE 3d
105
-------
CC
UJ
>
o
o
UJ
o:
100-
90.
80-
70-
60-
50-
40-
30-
20-
• •- t :••}:•.-; -.;:)••—:; ::;..:;-:••;;
"•.::"•'.-I - -A-—..._<•...:
Il'I^Vj^.'^T'TAl" r'--T^-'r f~TT-T--'"
;'.,"i"
i:-.:-":;::1::--.":*-.-:-!:
H"H-iTrmr:T"'~;":Tr"
,;i. ;:;}'•:.. ;,;•:!::;•;.
:.rr.~4:in:.:>CT~-.,. :+.. r.::,:.: IT:.™
., ,.,.::.*..... .;J.TS^.I-.... T^_-^.:r-i..---
'._:.; ;,t^T:"^'rrj;~iLr Mr;~:; i :.iinT™.r:"
. .I- •-~'""
to-
Qml/s 0 0.1
0.2 0.3 0.4 0.5 0.6 0.7 0.8
DISCHARGE PROTOTYPE SCALE 1/12
_l_
J_
_i_
_L
_L
J_
_1_
0.9
-U
1.0
0.05 0.10 0.15 0<20 0.25 0.30 0.35 0.40 0.45 0.50
L i i i i i i t • i i i i i i—i—i—i 1—
Q.cfs 0 I 234 5 6 7 8 9 10 II 12 13 14 15 16 17
r i i
Qpmgd o
10
61 cm (24 in) Skirt Slot Height 5 cm (2 in) Inlet Deflector
FIGURE 66 SWIRL PRIMARY SEPARATOR
SUGGESTED RECOVERY CURVE FOR ANION
EXCHANGE RESIN IRA-93 IN MODEL
149M>d>75ju
106
-------
36, which charted the portions of prototype
sewage represented by the IRA-93 in the
model. Values were selected from the
recommended curve, Figure 36, for each of
the scales being considered, as shown in Table
18'.
TABLE 18
REPRESENTATION OF IRA-93 TO
SEWAGE, BASED ON SCALE FACTOR
SCALE
1:12 1:8 1:4 1:2
Model
Percent
Represented, 56 64 79 100 100
Prototype
Chamber (m) 10.98 7.33 3.66 1.83
0.915
Diameter (ft) 36
24 12
Finally, each "Percent Represented" value
from Table 18 was .multiplied by the
"Recovery Rates" from Table 16, and plotted
in Figure 67 as a function of the appropriate
scaled-up discharge, and chamber diameter.
It is therefore possible to use a given
discharge in Figure 67, to determine what
solids recovery from sanitary sewage could be
obtained with various sized chambers. It must
be remembered that these recovery rates were
based on the sewage samples as defined by the
settling velocity distribution curve in Figure
35.
Since it is possible that other uses of the
swirl concentrator as a primary settler may- be
considered, it was necessary to express the
recovery rates for a range of different particle
settling velocities. The first step in
determining this was to refer to tests carried
out for the 0.5 I/sec (0.018 cfs) case on the
model using materials covering a range of
settling velocities. The data for these tests are
shown in Figure 68.
Next, the model recovery curve in Figure
66 was re-drawn as the bold line on
semi-logarithmic paper in Figure 69a. This
curve corresponds to the IRA-93 with particle
sizes 74/u< d < 149/u. The point indicated as a
star is for the 0.5 I/sec (0.018 cfs) discharge
for this material. Vertically above this point,
in circles, were plotted even settling velocities
at their corresponding recovery rates taken
from Figure 68.
The final step is based on the rationale
that, although these other particle settling
velocity materials were tested only for one
discharge, it is assumed that they would have
the same form of recovery rate curve as the
IRA-93, 74ju< d< 14% over the range of
discharges for which it was tested. Following
this approach, the curve of the bold line on
Figure 69a was re-drawn to form a family of
curves, each one passing through .the control
point on the 0.5 I/sec (0.018 cfs) line.
With these curves established for the
model a straightforward application of the
Froude.Law transposed the original bold line
to Figures 69b, c, d, and e for the larger
prototypes. The same procedure was followed
in laying out the different settling velocity
control points above the working point
scaled-up from the 0.5 I/sec (0.018 cfs), and
then drawing the family of parallel recovery
rate curves.
• After the ranges of discharges of normal
operation for the different sized chambers
had been determined, computations were
carried out to ascertain the corresponding
retention times. The approach followed was
to -consider the whole volume of the chamber,
both inside and outside the skirt, up to the
radial gutter crest level. The results of this
work are presented in Figure 12.
The design procedure is given in Section
HI.
Conclusions
1. With an inlet pipe to chamber diameter
ratio of 0.11, the solids recovery in the
chamber was relatively high. However, the
velocities in the inlet were so low that
solids deposition would certainly occur
there.
2. The inlet was reduced to a ratio of 0.07,
meaning that prototype velocities in the
inlet would be in the range of 30.5 cm/sec
(1 ft/sec), thus eliminating the possibility
of serious deposition. The chamber
recovery rate dropped slightly.
3. Tests with the inlet below the water level
in the chamber gave poor recoveries. Later
testing with the inlet crown at the same
level as the overflow weirs proved that
this was the best setting.
4. The flat floor around the outside
perimeter of the chamber was eliminated
107
-------
009Ol OS OS O4O9
DISCHARGE - mgd
005 01 OIS 01 OS 0« 0,8
DISCHARGE -eft i
09 10 20 VO 40 90
I DISCHARGE mgd I
I. . . I ... .1 . I I I I i i I il —U
02 030409 I, 20 3040 90 K>I9
DISCHARGE cfs
FIGURE 67 SWIRL PRIMARY SEPARATOR
PREDICTED PROTOTYPE SOLIDS RECOVERY FROM
SANITARY SEWAGE
108
-------
100
90-
80-
70-
60 H
o
o
LU
o: 40 -
30 -
2O -
10 -
O.OI
IRA- 93, 297ju «£ d ^ 590/u
' x
'SHREDDED
PETROTHENE
0.5-0.3 mm
(0.02-0.118 in)
O IRA-93, I49JU ^ d ^ 297At
IRA v 93, 74/j;Sd $149/1
« ARIZONA ROAD tttJjST
O.05 O.I O.2 0.3 0.4 0.5
PARTICLE SETTLING VELOCITY - cm/sec
1.0
NOTE: Tests with different materials done for discharge
0.5 l/s (0.02 cfs)
FIGURE 68 RECOVERY RATES ON MODEL AS FUNCTION OF
PARTICLE SETTLING VELOCITY
109
-------
FIGURE 69a RECOVERY RATES FOR 0.91 m (3 ft) DIAMETER CHAMBER
WITH DIFFERENT PARTICLE SETTLING VELOCITIES
001
1 .
005
... 1
0
1 .
005;
f
. 1
01
DISCHARGE
1 .... 1
05 10
. 1 , ,
05
. 1
1 0
mgd
FIGURE 69b RECOVERY RATES FOR 1.83m (6ft) DIAMETER CHAMBER
WITH DIFFERENT PARTICLE SETTLING VELOCITIES
110
-------
. , 1
O.I
1 .... 1
. , 1 . , .
o s
: 1
. 1
1.0
.... 1
. 1 ....
di
0.5
DISCHARGE
FIGURE 69c RECOVERY RATES FOR 3.66m (12 ft) DIAMETER CHAMBER
WITH DIFFERENT PARTICLE SETTLING VELOCITIES
1000
1 .
o.s
, ' .
. . . 1
1.0
1 .... 1
1 1 , , , . 1
5 10
. 1 .... 1
1
20
1 i
30 40
I i
FIGURE 69d RECOVERY RATES FOR 7.33m (24 ft) DIAMETER CHAMBER
WITH DIFFERENT PARTICLE SETTLING VELOCITIES
111
-------
1 . . . 1 .
5
... 1
10
1 .' . .
1
20
. 1
10
1
30
20
1 ,
SO
. . . 1
100
1 , ,
50
' -
ISO
, 1
100
mgd
FIGURE 69e RECOVERY RATES FOR 10.98m (36 ft) DIAMETER CHAMBER
WITH DIFFERENT PARTICLE SETTLING VELOCITIES
on two grounds; first, it caused-a flow
disturbance as the flow went over the
edge down into the cone section of the
chamber; and second, deposits formed on
it and remained entrained there.
5. The 60 degree conical floor angle,
continuing up to the chamber perimeter
diameter provided the best geometrical
pattern. ;
6. A skirt-chamber diameter ratio of 2:3 was
found to be the most efficient. This
arrangement left a sufficiently large
annular chamber to allow first-stage
sedimentation for the flow. '
7. The skirt was lowered until the slot
.between its lower edge and the conical
floor was 0.08D. At this position, the slot
entrance to the inner chamber was
sufficiently far enough below the inlet to
eliminate any inflow disturbance. ;
8. Circular overflow weirs inside the skirt
were eliminated, because they left a large
dead water zone around the central
downpipe under the weirs.
9. Closing the top of the skirt to forrri a
closed bell with orifices in the top for
clear overflow showed some promise, but
this pattern involved problems with dead
water zones at the center under the bell.
10. Radial overflow gutters proved to be the
most efficient arrangements to create
upflow over the whole cross section of the
inner chamber.
11. Tests carried out with fine grain material
were difficult and time-consuming, and the
results were indeterminate.
12. Analysis of these data presented a good
indication of the structure's general
capabilities at the model scale. However,
it was not possible to obtain specific
detailed measurements to make a
selection among the various forms of inlet
baffles that were tested.
13. Construction of the first prototype unit
should make provision for testing
different baffle arrangements.
14. A first design procedure was developed,
based on the "typical" sewage used in this
project.
15. A second design procedure was then
developed for use over a range of
prototype particle settling velocities.
112
-------
APPENDIX B
MATHEMATICAL MODEL STUDY
INTRODUCTION
Previous studies have shown that a swirl
concentrator can be an effective device for
the separation of grit and solids from
storm water flows.1 The objective of this
study was to evaluate whether these same
principles could be applied to the primary
clarification of sewage and combined sewer
discharges. The role of the General Electric
Company toward the fulfillment of this
objective was to develop a mathematical
model of the particle and liquid flowfields.
The general geometry- of the swirl
concentrator is shown in Figure 70,
Laboratory Swirl Concentrator Configuration.
The inflow enters the device tangentially in an
annular region between the circular skirt and
outer wall. The flow is then directed below
the skirt and into the central region. The
liquid then flows upward and leaves the
chamber at the top. Although Figure 70
illustrates a circular overflow weir, this weir
was later replaced with radial gutters. The
central standpipe was also removed and
replaced with a discharge structure external to
the device.
The approach to the mathematical
modeling of the swirl concentrator closely
follows the techniques used in previous
studies of secondary motion flow devices. The
flows within the chamber are assumed to be
axisymmetric about the vertical axis. The
chamber is then .overlaid with a computation
grid and the liquid flow velocities are
computed at each grid point by solving the
liquid continuity equation and equations of
motion. An eddy viscosity was utilized to
represent the turbulent shear stress. Plots of
streamlines and velocity profiles have been
prepared to depict the liquid flowfield as
predicted by the mathematical model.
Particle paths have been calculated by
superimposing particle settling velocities on
the liquid flowfield. Under certain
assumptions, it was demonstrated that a
simplified solution could be used to predict a
theoretical upper limit on the removal
efficiencies. Good agreement was observed
between the theoretical upper limit and actual
removal efficiencies observed by LaSalle
Hydraulic Laboratory in its hydraulic model
studies.
A design example is presented, illustrating
how both the mathematical model and
laboratory data can be used to predict
removal efficiencies for a prototype chamber.
Attention is called to the need for good
characterization of the sewage settling
properties for proper chamber design.
LIQUID FLOW CALCULATIONS
Equations of Motion
The ultimate configuration of the primary
swirl concentrator was derived from meetings
of the project team and from test data
obtained by LaSalle Hydraulic Laboratory.
Figure 70 depicts the geometry of the
physical laboratory model utilizing a circular
overflow weir and a central standpipe outlet
arrangement. Certain assumptions were
required in order to reduce the configuration
to a form amenable to mathematical
modeling. Figure 71 illustrates the simplified
configuration used in the mathematical model.
It is pointed out that the region over which
the mathematical model has been applied
does not include the annular region between
the skirt and outer wall. This region was
disregarded since most particle settling will
occur in the main body of the chamber. It
also greatly simplified the specification of
the boundary conditions for the mathematical
model.
The basic assumptions and methods used
in developing the mathematical model are
essentially the same as those used in the
previous study of swirl chambers as applied to
storm water overflows.2 The liquid flow has
been assumed to be axisymmetric. This
implies that the flowfield is identical at every
radial cross section and independent of the
angular position. This allows the use of only
two independent variables (r, the radial
113
-------
OVERFLOW WEIR
STANDPIPE
CIRCULAR SKIRT
OVERFLOW WEIR
, INLET
ir
N^CIRCULAR SKIRT
INLET
S^-REGION REPRESENTED
BY MATH MODEL
60
ELEVATION
OVERFLOW - - FOUL OUTLET
FIGURE 70 DIAGRAM OF LABORATORY SWIRL CHAMBER CONFIGURATION
114
-------
OVERFLOW WEIR
STANDPIPE
COMPUTATIONAL MESH
OVERFLOW VELOCITY
PROFILE
CIRCULAR SKIRT
ENTRANCE
VELOCITY
PROFILE
• FOUL OUTLET
(WHERE APPLICABLE)
PLAN
FIGURE 71 DIAGRAM OF SWIRL CHAMBER AS REPRESENTED BY
MATHEMATICAL MODEL
115
-------
position; and z, the depth) to completely
define the flowfield. In addition, the liquid
inflow is assumed to occur uniformly through
an annular band. This assumption corresponds
to the passage of the entering liquid beloV the
skirt of the laboratory device into the inner
region described by the mathematical model.
The axisymmetric assumption is more valid in
this study than the previous one3 as a result
of this inflow
along
p
all
points on the
circumference of the mathematical
representation of the chamber.
The equations used in the mathematical
model written in cylindrical coordinates are as
follows: i
1. The Continuity Equation
du
W
+ 9w = 0
9z
(1)
2. The Momentum Equations
»x ~*\
9r oz
9w - du) 9e ;
9r 9z/ 9z (2)
9r
*r ~ r)
9e
(3)
IT
9»
9z
9e
9r
9z 9z
• (4)
where
n — radial component of liquid velocity
v = tangential component of liquid
velocity [
w = vertical component of liquid velocity
r = radial coordinate
z = vertical coordinate
e = eddy viscosity
v = kinematic viscosity
p = liquid density
P = pressure
Boundary Conditions
The boundary conditions for the model
are derived from the chamber geometry and
flowrates. As mentioned previously, inflow to
the chamber is assumed to enter uniformly
around the circumference of the skirt as
shown in Figure 71. Along solid boundaries
the two velocity components parallel to the
surface are obtained from the skin friction
coefficient' and dynamic pressure. The skin
friction coefficient utilized was set equal to
0.0025 based on the previous study.4
The velocity profile for the inflow has
been assumed to be uniform, with a direction
parallel to the sloping floor of the tank
corresponding to the flow beneath the skirt of
the laboratory device. Two different overflow
velocity profiles were assumed. For the cases
with the flat circular overflow .weir a skewed
velocity profile similar to that used in the
previous study was assumed. (See Figure 71)
For the case where the overflow weir was
replaced with radial collection gutters a
uniform upflow velocity was assumed across
the tank surface.
Numerical Method
The approach for solving the liquid
flowfield equations closely parallels that used
in the previous study.5 The continuity
equation and equations of motion were
rewritten in terms of the non-dimensional
functions/, fi , and G defined by the
expressions
r2 (5)
o =
G =L
=L2 f
~~
(6)
(7)
Summary
The numerical techniques and system of
equations used to define the liquid flowfield
are the same as those used in the previous
116
-------
study. 6 The boundary condition and
flowrates have been modified to represent the
configuration of the primary swirl
concentrator developed by LaSalle Hydraulic
Laboratory.
The mathematical representation of the
device excludes the annular region between
the skirt and wall. The assumption of
axisymmetry is an appropriate one, for this
device since the flow enters the main portion
of the. chamber at all points along the
circumference of the skirt. As a consequence,
the mathematical representation of the liquid
flowfield is very close to the actual laboratory
device.
The eddy-viscosity mixing length constant
and the skin friction coefficient have been
assumed equal to the values obtained in the
previous study.7 A closer agreement between
the velocity field predicted by the model and
the laboratory data may be obtained by
refinement of these values. However, particle
removal; efficiencies will remain essentially the
same for reasons to be explained later.
Therefore, the existing values are satisfactory
for the purposes of this study.
PARTICLE SETTLING VELOCITY
Settling Data Analysis
a. Experimental Approach: In order to make
an accurate prediction of the prototype
removal' efficiencies of the sewage solids in
the laboratory model, it is necessary to
precisely define the settling properties of
actual sewage and the simulated sewage used
in the laboratory model. Towards this end,
extensive column settling tests were
performed by Beak Consultants, Ltd. on both
actual sewage and the simulated materials. The
report prepared by Beak, Physical and
Settling Characteristics of Particulates in
Combined Sewer Overflow, Sanitary
Wastewater and Urban Stormwater, describes
the details of the test procedures. The column
test data can be used to prepare a frequency
distribution of the particle settling velocities
which can be utilized to compare settling
properties of different materials.
where
\jj = stream function
7? = | vorticity
F = nondimensional stream function
M - nondimensional vorticity function
N = non-dimensional tangential velocity
function
r = radial coordinate
z = vertical coordinate
^ref = reference velocity
L = reference length
The stream function, \p, is defined so that
(8)
__
r bz
w = -
r ^T > (9)
thereby automatically satisfying the
continuity equation
_
dr
(10)
The equations of motion are then solved,
using relaxation techniques to obtain a steady
state solution. Once the numerical values of/
Si; and G are obtained, the velocity
components (u, v, w) in the axisymmetric
coordinate system (r, 6, z) can be found from
u -
U
ref
ref
V -
T
U
ref
(ID
(12)
(13)
b. Test Results for Actual Sewage: Figures 72
through 75 depict the composite results of
column settling tests performed by Beak on
Philadelphia sewage. The data points occur in
triplicate representing three separate test
runs. The data have also been grouped by
sampling port location to identify the extent
to which flocculation is occurring.
The method for detecting flocculation
effects from settling column tests has been
described in a paper by Camp.8 It involves the
assumption that a settling column is filled
with a uniformly mixed dispersion of particles
at time zero. Samples are then withdrawn at
several column depths at different times.
Assuming no flocculation or particles
interactions, each particle will move at its free
fall settling velocity. At time zero, samples
117
-------
100 f-
o.ooi
(0.00003)
0.01
(0.0003)
SETTLING VELOCITY cm/we (ft/iec)
0.1
(0.003)
FIGURE 72
100
SETTLING VELOCITY DISTRIBUTION AT 30.5cm (1 ft) COLUMN
SAMPLE PORT
0.001
(0.00003)
0.01
(0.0003)
SETTLING VELOCITY cm/sec (ft/Me)
FIGURE 73 SETTLING VELOCITY [DISTRIBUTION AT 61 and 91cm (2 and 3 ft)
COLUMN SAMPLE PORT
118
-------
100 —
80
cc
I-
z
UJ
o
o
4
<
p
60
40
20
8
COLUMN DATA AT 30.5 cm (1 ft)
O COLUMN DATA AT 61 cm (2 ft) and 91.5 cm (3 ft)
V COLUMN DATA AT 122 cm (4 ft), 152.5 cm (5 ft) and 167.7 cm (5.5 ft)
;-'..,. ... i i I • i 11
0.001
(0.00003)
0.01
(0.0003)
0.1
(0.003)
SETTLING VELOCITY cm/»ec (in/sec)
FIGURE 74 SETTLING VELOCITY DISTRIBUTION FOR ALL SAMPLE PORTS
100 _
122 cm (4 ft). 152.5 cm (5 ft) gnd 167.7 cm (S.S ft) DEPTH
0.1
(0.0.0031
SETTLING VELOCITY cm/»c (ft/nc)
0.1
(0.03)
FIGURE 75 SETTLING VELOCITY DISTRIBUTION AT VARIOUS COLUMN DEPTHS
119
-------
withdrawn from each sample port will; thus
have the same particle concentration.
Inasmuch as particles passing by the sample
port will be exactly replaced by particles
above the sampling point, the concentration
should remain uniform immediately i after
time zero. If it is then assumed thai: the
particles with the highest settling velocity
move at velocity Vl,. the last fractiqn of
particles with settling velocity Vi will1 pass
the first sample point located at dept|i Zr
when ti = Zi/V1 as shown in Figure 76.
Thus, at time ^ the particle concentration of
sample point Zl will change to reflect the
absence of particles having a settling velocity
Vi. If this principle is followed for all
particles, a time versus concentration plot is
obtained at each sample point. The same
concentration change and distribution should
be observed at the second sample located at
depth Z2 beginning at time t2 = Z2 / P"i. Thus,
identical curves should be obtained at
subsequent sample points except for a shift in
time. This time shift can be accommodated
by plotting C versus Z/r for each sample iport.
In actuality, C has been non-dimensionalized
by dividing the initial concentration C0 and
plotted versus Z(sample port)/? to provide a
distribution of settling velocities. ;
If flocculation occurs in the column,
particles will interact, agglomerate/ and
accelerate. This will result in an alteration of
the time versus concentration relationship.
Proceeding down the column, higher
concentrations will be obtained in a relatively
shorter time period, and the plot of ;C/C0
versus Zft will be shifted to reflect the
increased settling velocity as shown in Fiigure
75. Thus, instead of having one settling
velocity distribution for all sampling points,
separate curves can be drawn at each sampling
point.
It is apparent from the data shown on
Figure 75, that different settling velocity
distributions are occurring at the lower
depths, as a result of flocculation. Figure 75
illustrates this relative effect by overlaying
curves fitted to data at different sample
depths. At the 30.5 cm (1 ft) depth, 50
percent of the initial particles had a settling
velocity less than 0.012 cm/sec (0.00039
ft/sec). At the 61 cm (2 ft) and 91.5 cm (3 ft)
depth, 50 percent of the particles had a
settling, velocity less than 0.034 cm/sec
(0.0011 ft/sec) and only about 33 percent
•had a settling velocity less than 0.012 cm/sec
(0.00039 ft/sec). Finally, from 1.22 m (4 ft),
1.52 m (5 ft) and 1.67 m (5.5 ft) depth data,
50 percent of the particles had a settling
velocity less than 0.055 cm/sec (0.0018
ft/sec). Thus, the median settling velocity
increased from 0.012 (0.00039 ft/sec) to
0.055 cm/sec (0.0018 ft/sec) as the particles
settled through the column. This indicates an
increase by a factor of 4 to 5 of the median
settling velocity attributable to flocculation.
It can be assumed that the flocculation is
more effective in increasing the settling
velocity of the larger particles since the largest
difference in the three curves occurs at the
higher velocities.
Test Results for Simulated Sewage
Although many potentially applicable
materials were examined by Beak
Consultants, only three materials were found
suitable for the performance testing of the
laboratory hydraulic model. These materials
were shredded petrothene, 100-200 mesh
IRA-93 resin, and Arizona Road Dust. The
settling velocity distributions for these
materials appear as Figures 10 through 12 in
the Beak report. No flocculation effects were
reported by Beak and a single settling velocity
distribution was fitted to all the data points
independent of sampling depth. The
discrepancies in the 'three runs for IRA-93 are
a result of sieving techniques and will be
discussed in detail in a later section.
Settling Theory
A literature search was conducted to
identify the mathematical expression for
describing the relationship between particle
concentration and particle settling velocity.
At low particle concentrations, particle
interaction is minimal and each particle settles
at its free fall velocity. In this region particle
settling velocity can be considered
independent of depth and concentrations. As
the particle concentration rises the
probability of particle interaction increases
and two phenomena can occur.
1.20
-------
SETTLING
COLUMN
CONCENTRATION
VS. TIME'PLOTS
AT EACH SAMPLE
PORT
PLOTS OF C/CC
VERSUS Z/t
SAMPLING
PORTS
c
c_
z/t
• NON-FLOCCULATING PARTICLES
FLOCCULATING PARTICLES
F.GURE 76
FLOCCULATION vs
121
-------
One of these phenomena is called
"hindered settlement." In the regime of
"hindered settlement" the particles
continuously collide with one another,
resulting in a decrease of the individual
particle settling velocities. Camp8 states that
for raw sewage, a solids concentration of
greater than 1,000 ppm must be reached
before hindered settlement will be
encountered. Concentrations of solids within
the swirl concentrator will only approach
1,000 ppm towards the bottom of the tank.
Since most of particle settling will occur in
the main body of the tank, hindered settling
should not be a major factor in determining
removal efficiency.
The second phenomenon which can alter
the particle settling velocity is flocculation.
Flocculation causes an increase in settling
velocity by the agglomeration, or clumping
together of small particles to form larger and
heavier particles. The ability of a suspension
to exhibit flocculation properties depends
primarily on particle chemistry and
electrostatic charge. Assuming that a material
has flocculating properties, agglomeration can
occur at relatively low concentrations and it
varies directly with particle concentration.
Raw sewage can exhibit flocculation at
relatively low concentrations (230 ppm) as
the major mechanisms responsible for
flocculations and agglomeration follows.
Agglomeration Mechanisms
Agglomeration results from particle
collisions induced by gravity, shear, and
turbulence, as will be discussed in greater
detail. The agglomeration of small particles
tends to increase the average particle size, and
thus improve the settling properties of the
suspension.
The efficiency of the agglomeration (or
flocculation) process depends on the number
of particle collisions. Of course, not all
particle collisions result in agglomeration. The
particles may slide past each other or, in the
case of a stable colloid, they may repel each
other electrostatically. Nevertheless, particle
collisions are a necessary condition for
agglomeration, and the number of collisions
per unit volume per time can be used to
estimate the maximum rate at which
agglomeration may occur. A review of the
rather extensive literature on flocculation and •
coagulation processes has revealed five basic
mechanisms which result in particle collisions.
These mechanisms are discussed briefly
below, and the appropriate collision rate
equations for the case of spherical particles
are summarized in Table 19.
noted by the Beak data. A detailed review of
TABLE 19
COLLISION RATES FOR VARIOUS MECHANISMS
Collision
Equation
(14)
(15)
MfiHianism Collision Rate Eauation
Gravity JV.. =\-^- (r. + r. )2
r]-r^(
Shear Flow N.. = -|- (r. + rf)3 n. n. G'
0 \ 1/2
p~l) gn'ni
Ref.
3
4
(16)
Turbulent
Acceleration
N =
*
(17)
Turbulent N _
Entrainment */
r{
r.
p
Brownian Motion
= kinematic viscosity
- diameter of particle /
= diameter of particle /
= density of particle
= density of liquid
e = energy dissipation per unit mass
n. = number density of particle /
n. = number density of particle /
jv7.. = number of. collision of t particles
with / particles
G ' = shear rate of flow
122
-------
1. Gravity: Particles settling under the
influence of gravity settle at different rates,
depending on their size and specific gravity.
Faster settling particles tend to overtake
slower .particles, resulting in collisions. For
small particles settling in the Stokes flow
regime, the collision rate is given by equation
(17) of Table 19. If the particles are of the
same size (rt = r;-), no collision will occur.
Acceleration of the ambient fluid will also
cause collisions. The collision rate in an
accelerated fluid can be obtained by replacing
g in equation (14) with the fluid acceleration
Du/Dt. For the present application, the fluid
acceleration is several orders of magnitude
smaller than the gravitational acceleration,
and may be neglected.
2. Shear Flow: In a shear flow pattern,
adjacent streamlines move with different
velocities. A particle moving along one
streamline may, therefore, overtake and
collide with a particle on an adjacent
streamline. The collision rate depends upon
the collision cross section fa + 77), and upon
the shear rate of the flow G' (ft/sec/ft) as
given by equation (15).
3. Turbulent Acceleration: In a turbulent
flow pattern, two or more large eddies is
generally small (compared with gravity), but
the small eddies in the viscous dissipation
range can have larger accelerations, given by*
•
in which e' is the energy dissipation per unit
mass, and v is the kinematic viscosity. This
mechanism is only applicable where the
particles are smaller than the scale size, X, of
the energy dissipating eddies,
where4
X =
(20)
As in the case of gravity, no collisions occur
tor particles of the same size, since both
experience the same acceleration.
4. Turbulent Entrainment: If the
particles are comparable in size to X, collisions
will.result when a particle entrained in one
turbulent eddy collides with a particle entrained
m another turbulent eddy. This mechanism
provides for collisions of particles of the same
size, m contrast to the acceleration effect of
turbulence discussed above.
5. Brownian Motion: Particles undergoing
random Brownian motion due to collisions
with molecules of the ambient fluid will
collide. However this mechanism is only
significant for very small particles (less than 1
M)-5
The collision rates given in Table 17 are
not additive. Saffman and Turner have
shown9 that the combined effect of turbulent
acceleration, turbulent entrainment, and
gravity can be approximated by
root-sum-squaring the individual expressions
given in Table 17.
Agglomeration Kinetics
If the particles in suspension are discrete
sizes, a system of equations may be derived to
describe the rate of change of the number
density of the ith species:5
dnt =
dt
N.
/=.!
'-1 .(21)
In equation (21) the number density of the
ith size particle is «., and Nt) are the collision
rates between particles of size / and size/. It is
assumed that the collision of an / particle and
a / particle results in coalescence to form a
particle of size / + / with radius
TI+I = (r? + rf )J/3 . (22)
The first term on the right hand side of
equation (21) is the sum over all collision
which results in the formation of an /-size
particle. For the collision rate Ni>k between
particles of size / and size k, /-size particles are
produced only when f + k = /. The second
term on the right of equation (21) is the rate
ot depletion of/-size particles, due to collision
with all other possible particle sizes.
Equation (21) represents a system of
equations; one equation for each discrete
particle size. Numerical solution of these
equations is possible, but very
time-consuming. Such solutions have been
performed by Gemmel,1 ° Fair and Gemmel 12
Ives13andHidy,14 for various collision
mechanisms.
Where the particle size distribution is
continuous rather than discrete, the
123
-------
coagulation process can be described with an
integro-differential equation for the
continuous distribution function n(r,t) which
is defined such that n(r,t)-dr is the number of
particles per unit volume of size range from r
to r + dr. The total number of particles per
cubic centimeter, 7Voo is then
00
= /
r,t) dr-
(23)
The equation for the continuous distribution
function n(r,t) is analagous to equation (21)
for the discrete case, and is given by
d n (r,t) _
dt
N(r,(r3 -r3)1/3) dr
N(r,r)df
(24)
In equation (24), N(r^r) is Jfte collision rate
between particles size r and'7: As in equation
(21), the first integral on the right represents
the rate of production of size r particles,, and
the second term represents their depletion
due to collisions with all other sizes. ;
Friedlander, has shown that for a
settling aerosol undergoing coagulation by
Brownian motion, a similarity transformation
for equation (24) is possible. The I size
distribution function, expressed in similarity
variables, does not change with time. In
particular, n(v,t) is expressed as
N
"a
vN0
(25)
where v is the particle volume, 4>, ;is a
non-dimensional distribution function and 8,
is the total volume fraction of particles.
v n (v) dv .
(26)
Using equation (25) in equation (24) results
in an ordinary differential equation for»//', in
terms of the similarity variable i? = vNool'e.
The same similarity transformation was
applied to coagulation by laminar shear flow
by Swift and Friedlander, However for the
case where both mechanisms (Brownian
motion and laminar shear) are operative, the
self-preservation hypothesis does not work.
In the present application to a
nonhomogenous turbulent flowing system,
none of the solution techniques described
above are applicable. Equations (21) and (24)
are both derived for homogenous systems. To
include the spatial variations in the
distribution function, n(r,t), the time
derivative in equation (24) must be replaced
by the substantive derivative along streamlines
dn _ to + v . y n . (27)
In addition, because of turbulence, an
additional source term appears on the right
hand side of equation (24) to account for
dispersion of particles across streamlines, due
to turbulence. The appropriate equation
therefore is
dn
dt
+ V • A n =
N (r,
dr
00
-/
N (r,r)dr + V • (ep Vn) (28)
where ep is the particulate eddy diffusivity. In
equation (28), n is now ajfunction of both
space and time, i.e., n = n(r,x, t).
The solution to equation (28) is
extremely complicated. An approximate
method can be developed in which the local
distribution function n(r,t) is approximated
with a Gaussian profile
n (r, x', t) =
exp I-
(29)
in which the mean radius a, standard
deviation, a, and total number density of
particles N, are each functions of space and
time. Equations for Noo, a, and a can be
derived by taking moments of equation (28).
Using equation (29), the second of the
integrals which appear in (28) can be
evaluated in closed form for any of the
individual collision mechanisms of Table 19.
However, closed form integration is not
possible with more than one mechanism
124
-------
operative. Also, the first integral in equation
(28) cannot be evaluated in closed form f6r
even a single collision mechanism.
Consequently, it appears that the solution
to the coagulation kinetics equation is
impractical in the present application.
Instead, an order of magnitude estimate has
been made of the collision frequencies for
each of the mechanisms in Table 19. From
these numerical estimates, several important
conclusions can be drawn regarding the design
of the primary settler.
NUMERICAL ESTIMATES FOR
COLLISION FREQUENCIES BY
VARIOUS MECHANISMS
The settling column data by Beak can be
used to determine the probable size range of
p articulates in sanitary sewage. The GE
analysis of the data from the 1.22, 1.52 and
1.67 m (4, 5 and 5.5 ft) depths (Figure 74)
shows that 25 percent of the particles settle
slower than 0.003 cm/sec (0.00098 ft/sec), 50
percent settle slower than 0.055 cm/sec
(0.0018 ft/sec) and 75 percent settle slower
than about 0.13 cm/sec (0.043 ft/sec). The
settling velocity of a spherical particle in the
Stokes flow regime is
w, = -i- — (X- Or2 • (30)
s 9 v ^ g ' <•'
Equation (30) can be used to estimate the
particle size from the measured settling
velocity data if the specific gravity is known.
Since the specific gravity is not known, values
of 1.01 and 1.05 were assumed in order to
obtain the particle sizes corresponding to the
25 percent, 50 percent, and 75 percent
settling velocities given above. The results of
this calculation are summarized in Table 20.
For each of the particle sizes in Table 20,
the number of particles per cc required to give
a total concentration C, of 100 mg/1 solids
was calculated from
„ =(lQOm_g} ( II
' / / UOOOcc.
/ 3 particles\
u^r
cc
lOQOmgJ
= 3xlO'4
4-rrr3 n
(31)
The resulting number densities are also
tabulated in Table 20.
An estimate of the collision rate1 from
each of the mechanisms in Table 18 was next
obtained under two assumptions.
a. Assume 100 mg/1 of particles at the
25 percent settling rate, and 100 mg/1
at the 75 percent settling rate.
b. Assume 200 mg/1 of particles at the
50 percent settling rate.
Assumption (a) is a bimodal distribution, with
two discrete particle sizes. Assumption (b)
corresponds to a mono-disperse system of a
single particle size. These calculations were
performed for each of the two specific
gravities 1.01 and 1.05. The calculations are
outlined below for assumption (a) at 1.01
specific gravity.
(1) Gravity
_ 27T(0 Q246 + 0.00373)2
9(0.011,
9 (0.011 cm?/sec) ' ' I
/ \
[ ^~~ 1 I (980'7 cmlse°2> x( 1.59 cm'3) (4550cm"3)
— .219 collisions/cc/sec
TABLE 20
PARTICLE SIZE AND NUMBER DENSITY FOR
VARIOUS SETTLING VELOCITIES
Specific
Gravity
(gm/cc)
1.01
1.05
Settling Velocity
cm/sec in/sec
0.13
0.055
0.003
0.13
0.055
0.003
(0.05)
(0.02)
(0.001)
(0.05)
(0.02)
(0.001)
Equivalent Spherical
Radius
cm in
0.0246 .
0.0160
0.00373
0.0112
0.00731
0.00171
(0.009)
(0.006)
(0.0015)
(0.044)
(0.003)
(0.0007)
Number Density for
100 mg/1
cm"3
1.59
5.76
455.
16.2
58.0
4,540.
125
-------
(2) Shear Flow
NtJ =|- (rf + r,.)3 nin)G'
The mean shear rate is obtained from the
swirl separator. Since this function is used in
the eddy viscosity formulation, it was a
simple matter to average its value over the
cross section. The resulting value of G' for the
prototype scale separator is 0.209 sec"1.
Consequently, for sg = 1.01 and rz-•¥= /•/
Ni] = A. (0.0246 + 0.00373)3 (1.59)
(455) (0.209) = 0.00458
(3) Turbulent Acceleration
The turbulent energy dissipation is
estimated, assuming that the energy inflow is
entirely dissipated in the outer annulus.
The energy dissipation per unit mass is
(32)
JT
The volume flowrate Q = 0.52 m?/sec (17-65
cfs), prototype (1 I/sec model scale). The
head loss, AP is
Ap =-1— pv2 (33)
where the entrance velocity of the 1.22 m (4
ft) entrance sewer is
= 1.4/ps(0.43 m/sec)
so that AP = 1/2 (10 4 kg/m3) (0.43)2 =;9.6
kg/m3 [AP = 1/2 (2 slugs/ft2 = 1.9 lb/ft2.]
(9 x 10~4 atm). The volume of the outer
annulus is about 60.88 m3 (2,150 ft3). Then
from equation (32)
• - (17.65) f 1.9)
6 (2150) (2)
= . 00782 ft*lsec3-
= 7.26 cm2/sec3 (1.13 foZ/sec3)
This energy dissipation rate corresponds to an
equivalent shear rate G' =-^/e'/v of 26 sec'1.
The equivalent acceleration, a, is equation
(19)
1/4
= 23.7 cm/sec2 (9.33 in/sec2 )
Since equations (14) and (15) for the collision
rate due to gravity and turbulent acceleration
are identical, except for the acceleration rate
used, the collision rate for the former can be
found as the ratio
turb accel
J gravity \
= (.219) 23.7 = 00528 collisions/cc/sec
(4) Turbulent Entrainment
{rt + rf)3 nt nf Jjj-. (17)
(appears previously in Table 1)
Using the previous result for the
dissipation per unit mass, e',
energy
Ntj =
(0.0246 + 0.00373)2 (1.59) (455)
7.26 I
LoxmJ
= 8.62 collisions/cc/sec
The results of these order of magnitude
calculations are summarized in Table 21. It
can be concluded from this table that turbulent
entrainment and gravity are the principal
mechanisms for promulgation of flocculation
in the swirl concentrator.
Summary
In order to extrapolate the results of the
hydraulic model testing to prototype design,
it is necessary to understand the differences in
settling properties between the laboratory
simulated sewage and actual sewage. From
column test results performed by Beak, it can
be concluded that the materials in the
laboratory model settle as discrete particles,
while actual sewage exhibits flocculation
properties even at low concentrations.
A detailed analysis of the kinetics of
flocculation was performed. It was found that
for flocculant materials, the degree to which
flocculation and agglomeration occur depends
largely on the number of particle collisions.
Theoretical expressions were developed for
the five basic mechanisms which can cause
particle collision. These mechanisms are:
126
-------
TABLE 21
ESTIMATED COLLISION RATES
Size
Assumption
Sg = 1.01 fj ^r2
'!='!
= 1.05r1^r2
rl/=r2
Collision Rates - Collisions per cc per sec
Turbulent Turbulent
Gravity Shear Acceleration Fnrrainmpnt
0.22
0,
4.8
o;
0.0046 0.0053
0.00015
0.044 0.12
0.0015
8.6
0.29
83.
2.8
1. Gravity: Differential settling rates
2. Shear: Differential flow velocities
along adjacent streamlines
3. Turbulent Accelerations: Entrainment
of two or more particles in a single
accelerating turbulent eddy
4. Turbulent Entrainment: Entrainment
of particles in separate turbulent eddies which
collide ;
5. Brownian Motion: Random motion
caused by thermal energy of water molecules.
These expressions were used to obtain
order of magnitude estimates for the number
of collisions which can result from each
mechanism at the flow conditions existing in
the swirl concentrator. Of the five
mechanisrris, only gravity and turbulent
entrainment appear to be significant. The
turbulent entrainment mechanism is most
important in the outer annulus of the
concentrator where most of the inlet energy is
dissipated.
It is most important to recognize that the
theoretical analysis of flocculation only
predicts .the number of collisions or the
maximum tendency toward flocculation. The
actual adherence of particles to each other
depends on the chemistry as well. It may well
be that there is a limiting amount of
flocculation which can occur as a result of
other conditions. If this is the case, an
increase in collisions beyond the number
required to reach this threshold will not result
in additional improvement in settling rates.
The column tests performed by Beak on
sanitary sewage showed that a fourfold
increase in the median settling velocity can be
observed as a result of the gravity flocculation
mechanism alone. Whether further
improvement can be attained via the
turbulent entrainment mechanism in the
outer annulus of the swirl concentrator is
unknown.
Exact mathematical modeling of the
flocculation effects on the particle settling
rate was deemed numerically impractical for
this study. Therefore, to accommodate the
flocculation of actual sewage, a settling
velocity distribution indicative of the sewage
after, flocculation was assumed. This approach
assumed after flocculation occurring in the
inlet sewer and outer annulus of the swirl
concentrator via gravity and turbulent
entrainment mechanisms will produce a
settling velocity distribution equivalent to
that obtained by the gravity mechanism in the
settling column test. This assumption would
appear to .be conservative since further
improvements in settling properties might
occur in the swirl concentrator.
PARTICLE FLOW CALCULATIONS
Particle Paths
Once the liquid flow has been determined
the particle velocities can be determined by
superimposing the particle settling rate, ws, at
each mesh point. Thus,
UP = UL (34)
(35)
w =
w
(36)
where ws = particle settling velocity the p and
L subscripts designate the particle and liquid
velocity components.-
127
-------
The particle paths can be found by
superimposing a stream function due to
settling, t//, on the liquid flow. Thus, the
particle stream function $ p is
where \jj ' (r) is defined such that
= -1
w= -
or
-w.
(37)
(38)
(39)
The particle paths for a given settling velocity,
\vs, are then determined by plotting lines of
constant. !
The particle paths obtained provide a
method for visualizing the mean trajectories
of the particles in the chamber. In actuality,
turbulence will cause particles of a given
settling path. The particle paths do not
provide any information as to the
concentration field. To define the
concentration field, it is necessary to utilize
the particle mass continuity equation as
defined in the following section.
Particle Flow Equations
Particle concentrations throughout the
swirl concentrator are determined by
substituting the particle velocities at each
mesh point, as defined by equations (34)
through (36), into the particle continuity
equation (40).
_9c
dt
+ V.
C) = V (e- V C)
(40)
where
vP - particle velocity vector
C — particle concentration ;
t = time
ep = turbulent eddy diffusity
The use of Equations 34-36 for Vp assumes
that inertial acceleration terms are negligible
when compared to the gravity acceleration, so
that particles always move at t!heir
equilibrium settling velocity with respect to
the liquid flowfield.
The values of the velocities at each grid
point are obtained directly from the liquid
flowfield solution. The boundary conditions
for the particle concentrations along the
chamber wall are specified and a relaxation
technique is applied to reach a steady state
concentration. Removal efficiencies are then
computed from mass balances utilizing the
computed concentration field and specified
flowrates^
Boundary Conditions
The boundary conditions for equation
(40) are as follows: at the inlet, the
concentration is a specified constant value,
C0. At a solid surface the total mass flux must
be zero. Mathematically this can be expressed
by
+ ep V C) • n =0
(41)
where n is the unit inward normal. The first
term in equation (41) is the convective mass
flux, and the second term is the mass flux due
to diffusion. Equation (41) can be rewritten
-'(vo+'p-U-
(42)
velocity component normal to a solid surface
the term Vp • n must equal the component of
inward normal, n.
where d/dn denotes the directional derivative
normal to the wall. Since the liquid has no
velocity component normal to a solid surface,
the term V-ri must equal the component of
the settling velocity normal to the wall, i.e.
-ep|^0 (43)
in which nz is the z-component of the unit
inward normal, 7i,
The derivative in equation (43) can be
replaced with the first order finite difference
approximation
If = • (Cfc+1-CL)/A»
(44)
in which Ch denotes the 'boundary value,
b+1
is the value at the interior point,
is the distance from
Cb+1'to
and
The
result is an equation for the boundary value
in terms of the interior point.
= C,...
(45)
128
-------
where
R = ws A n n2
(46)'
For the special case of a vertical wall, nz =
0 so that R = 0 and the boundary condition
becomes
Cb = Cb+1 (47)
Underneath the horizontal weir, nz = 1 (z
is positive downward so that ws is positive),
and equation (45) is the correct boundary
condition, with R > 0. If the settling velocity
and mesh spacing are sufficiently small that R
« 1, equation (45) gives a wall
concentration Cb, nearly equal to Cb+1.
However if R becomes large, the
concentration at the boundary becomes much
smaller than at the interior point. Provided
the settling velocity, ws, is larger or at least
comparable to the mean liquid upflow
velocity, ws, then Cb + 1 will itself be small
(because few particles can reach the upper
region of the separator), and no numerical
difficulties occur. This situation corresponds
to the case studied in the earlier application
of the swirl concentrator to stormwater
overflows.1
A problem arises when R is large but the
settling velocity is small compared with w. In
this case, a large number of particles reach the
upper region of the concentrator, and Cb + 1 is
nearly equal to the inlet concentration, C0.
Then equation (45) requires Cb« Cb+1,
which appears as a discontinuity in the
concentration field, causing numerical
problems. Physically, a large R and small ws
means that the concentration is high except
for a thin layer (smaller than one mesh space)
immediately under the weir. This is exactly
the situation for the proposed design, for
which ws is typically 0.05 cm/sec (0.00164
ft/sec) while the prototype mean upflow
velocity (at 0.52 m3/sec [17.1 cfs]) is
approximately 1.3 cm/sec (0.0043 ft/sec).
The R value for this case (with mesh spacing
of 15cm [0.5 ft]) is
R =
(0.00164) (0.5)
= 68
1.2xlO's
Thus, the large gradient in concentration
under the weir occurs in a distance much
smaller than the 15 cm (0.5 ft) mesh spacing.
Under these circumstances, the proper
boundary condition is to neglect this very
small region and assign Cb = Cb + 1 .
A similar difficulty occurs at the bottom
of the separator. In this case, the
z-component of the inward normal, nz, is
negative (-1 along the horizontal portion, and
- cos 60° along the sloping wall). Since R
(from equation (45)) is then negative, the
boundary condition calculated from equation
(45) can only be used if R=l. Where this
condition is satisfied (either because a very
fine mesh is used, or because the settling
velocity is small), the concentration gradient
at the wall can be included within the
computational mesh by applying equation
(47). .
When R is large (as in the present case),
the concentration buildup at the wall due to
sedimentation occurs in a region too small to
be resolved with the computational mesh. The
sediment layer must therefore be excluded
from the computational region. Once this is
done, some other criteria must be applied in
order to specify the mass flux normal to the
wall, Qn , where
Qn = ~
dn
(48)
For example, Qn can be positive or negative
depending on whether deposition or scouring
of the sediment layer is occurring.2 If these
mechanisms are in balance (no deposition or
scour) then Qn = 0.
The specification of the mass flux for the
general case is quite difficult. As discussed bv
Chen,1 r Qn is usually approximated by an
empirical formula as a function of the local
mean velocity of the main flow,12 a function
of the local velocity at the bottom,13'14 a
function of the boundary shear and critical
shear stresses of the Yalin type,15'16 or a
function of the local mean velocity and depth
of the main flow such as the Meyer-Peter and
Muller type17.
For a properly designed settling tank, no
scouring should occur, and Qn will always be
positive along the bottom. In the present
design, the inlet velocity may be high enough
to scour the bottom near the inlet annulus.
Where scouring does occur, the sediment
deposit will eventually be removed, so that in
the steady state Qn = 0 for these regions.
129
-------
The simplest hypothesis for the bottom
boundary condition in the settling chamber is
to assume the bottom boundary points can
then be determined by treating them as
interior points, except that one-sided
difference approximations must be used.
Particles reaching the bottom of the separator
are thus assumed to pass outside the
computational mesh into a thin sediment
layer from which no re-suspension occurs.
This procedure will tend to over-predict the
efficiency of the chamber, because
re-suspension of the bottom sediment is
neglected.
To summarize these findings. Table 22
has been prepared to indicate under what
conditions each type boundary treatment
is applicable. ,
Numerical Methods ;
A numerical solution of the particle
continuity equation (40) can now be obtained
by applying the boundary conditions of the
previous section. As noted, for large settling
velocities where ws » w, a direct numerical
solution of equation (40) can be obtained
using equation (45) for the boundary
condition under the closed top. This solution
'and boundary condition are essentially
identical to that used in the previous stuliy.1.
In addition, a theoretical upper limit on the
removal efficiency can be_readily defined for
this case. When ws» w, the mean particle
flow path of all the particles will intersect the
chamber floor. In the absence of turbulence, a
theoretical removal efficiency of 100 percent
should be obtained. In actuality, the
mechanisms of particle scour, re-entrainment
and turbulent diffusion will cause particles to
be dispersed throughout the chamber. As a
result, an actual removal efficiency less than
100 percent will occur, and concentration
gradients will exist throughout the chamber.
For the purpose of this study it is sufficient
to state that when ws is comparable to or
greater than w, the performance of the unit
will deviate from the theoretical upper limit
of 100 percent removal, due to scour,
re-entrainment and diffusion as noted above.
The second case, where1. ws^w,-. is more
representative of the conditions projected for
the prototype swirl concentrator as a primary
treatment device. For this case, the large
concentrations of particles below the closed
top cause numerical difficulties, as noted
previously, which cannot be handled with a
reasonable mesh spacing. To resolve this
problem, the particle concentration below the
weir has been assumed to be equal to the
concentration at an interior point.
A second assumption was employed to
define the particle boundary condition at the
floor. Since the definition of particle
deposition, scour and entrainment mechanism
is beyond the state of the art, it was assumed
that all particles which reach the floor are
removed from the device and the
concentration at the floor equals the
concentration at the adjacent interior point.
Thus, the large concentration gradients
immediately adjacent to the weir and to the
floor have been neglected.
The net result of these two assumptions is
that, for the case where ws
-------
concentration everywhere within the chamber
becomes uniform and equal to the inlet
concentration, C0. To see why this is so,
assume that the concentration is in fact
everywhere constant. Then the turbulent
diffusion term on the right side of equation
(40) becomes zero', sincere = 0, and equation
(44) can be reduced to
-|£ + f£.VC+VC-
As previously noted,
Vn = VT + u> e.
(49)
(50)
Where VL is the liquid velocity, and ^Tis a unit
vector in the z-direction. z
Since the divergence of the liquid velocity
(A- VL ) is zero, and since ws is a constant
V . Vp =0. (51)
Equation (49) can therefore be written as a
total derivative
-DT^°> (52)
where D/Dt is the total derivative along the
particle streamlines.
Equation (52) states that in the absence of
turbulent diffusion, the concentration
remains constant along particle paths.
All regions of the chamber cross section
which are connected to the inlet region by a
particle path must have "a concentration of
C0. The remainder of the cross section (such
as the region underneath the weir) contains
particle paths which originate at the weir or at
the free surface. Since there is no particle flux
through these surfaces in the absence of
turbulent diffusion, the concentration must
be zero in these regions.
The effect of turbulence is to scatter
particles from regions of high concentration
to regions of low concentration. Initially,
therefore, turbulence will tend to decrease the
concentration of regions with C = C0, and to
raise the concentration in the region with C =
0. However particles which are scattered into
the closed region under the weir tend to
remain there. Consequently, in the steady
state, this region will fill up with particles
until its concentration is also equal to C0.
This effect is readily apparent in the
laboratory tests. As the particle flow is
initiated, the particle cloud tends to follow
the streamlines from the inlet up to the
overflow, and the closed region under the
weir remains clear. However, the turbidity of
the closed region gradually increases, and
after about 10 minutes, no concentration
gradients are discernable.
Once the concentration in the chamber
becomes uniform, turbulence has no further
effect because the concentration gradient, AC,
becomes negligible, and the right hand side of
equation (49) can be neglected. Under these
conditions equation (52) is applicable even in
the presence of turbulence.
The efficiency of the concentrator can be
determined as the ratio of the particle flux
onto the bottom and the influent particle
flux. As shown (equation 52), the
concentration along the streamlines reaching
the bottom will be equal to the inlet
concentration, because DC/Dt = 0.
Consequently, the particle flux reaching the
bottom (from equation (48) with DC/Dt = 0,
and.C=C0)is
' z
f
ds
(53)
where (2 " rnzds) is an element of projected
surface area along the bottom. Since nz =
-dr/ds, equation (53) gives
FB = C0 w,Ab (54)
where AB is the projected area of the bottom.
For the case where some fraction, r?, of the
liquid flow is continuously withdrawn from.
the underflow, an additional mass flux of T? Q
C0 must be added to equation (54). The inlet
mass flux is
where Q is the rate of flow of the liquid.
Consequently, the separation efficiency of the
chamber is
* = FBlFin
If the ratio Q I-^-B is viewed as an overflow
rate for the chamber (the usual overflow rate
is defined by the ratio of the flowrate to the
surface area), then equation (55), with T? = 0,
is identical to the theoretical efficiency of an
ideal settling basin as derived by Camp2 in
1945. This is a somewhat surprising result,
131
-------
inasmuch as equation (55) was derived ijnder
considerably different assumptions and for an
entirely different geometry than analyzed by
Camp. !
Summary
Particle paths in the swirl concentrator
can be determined by superimposing a stream
function due to the settling, V, on the liquid
flow stream function.
Under certain conditions of particle
settling rate and computational grid size,
modification to the boundary conditions are
required to solve the particle continuity
equation. These modifications involve
neglecting regions of high concentration
gradients, where the gradient occurs over an
interval smaller than one mesh spacing. The
approximate boundary conditions are
summarized in Table 21. For the laboratory
model, where particle settling velocities are
much smaller than the mean upflow velocity,
swirl concentrator separation efficiency (ratio
of gm/sec particulates reaching bottom to
gm/sec particulates entering the chamber) can
be approximated by equation (55).
+ 7? (55)
even for cases where the settling velocity is
comparable to or greater than the overflow
rate, equation (55) represents an upper limit
to the performance of the swirl concentrator.
It is of interest to note that this equation
is equivalent to expressions used for
computing removal efficiencies for ideal
settling basins. i
SCALING TECHNIQUES
Scaling of the Liquid Flowfield
Scaling of the liquid flowfield can be
accomplished via the same Froude scaling
techniques used in the previous study.1
Essentially the following equations can
be applied to compute dimensions, flowrates,
and velocities for a prototype scale factor bf S.
Lprototype = s' ^model (56)
^prototype = s" V "model
(57)
where
S
Qmod
V
mod
-^mod
£"
' pro
scale factor = Z,pro/Lpro
discharge rate in model
velocity in model
length dimension in model
discharge rate in prototype
velocity in prototype
length dimension iri prototype
To use equations (56) through (58) it is
necessary to assume a scale factor, S. For
example, assume the laboratory model
constructed by the LaSalle Hydraulic
Laboratory was designed to represent a
prototype unit 12 times larger, i.e. S = 12
(^prototy pe/^model = 12>" • The
corresponding velocities and flowrates for the
prototype, therefore, are larger than those of
the model by a factor of
Qnro-(12)s«-.-
= 3.47Fmod
In similar fashion, the size of a device
designed to handle a specified flow of Q = Q
design at a removal efficiency equivalent to
that of the LaSalle Hydraulic Laboratory
model can be obtained from equation (58)
Q design
Q model
2/5
The corresponding S can then be used in
equation (56) and (57) to compute the
required dimensions and flow velocities.
Certain assumptions are implied when
applying Froude scaling to the laboratory
model. First the liquid flowfield- boundary
conditions as previously described must be
maintained in the prototype unit. This can be
accomplished by maintaining geometric
similarity to the laboratory model.
The second assumption requires that the
degree of turbulence, as defined by the eddy
viscosity e be maintained at the same value as
the model. The nondimensional viscosity is
represented by-(l)
(59)
^prototype
= s
2model
where
0
/
v
L
W
local dissipation function
mixing length constant
kinematic viscosity
reference length
reference frequency (Uref /L)
132
-------
The first term on the right is the eddy
viscosity arising 'from "the Reynolds Number
stresses while the second term represents the
molecular viscosity. The eddy viscosity is
independent of scale size and flowrate since
neither W nor L appear explicitly in the first
term. Thus, as the size of the chamber is
increased or the discharge rate is increased,
the turbulence level increases in just such a
way that the same nondimensional eddy
viscosity results. The second term on the
right, however, depends on both the flowrate
and size. This term is the reverse of a
Reynolds Number based on reference length,
L; reference velocity, WL; and liquid
kinematic viscosity, v.
In the previous study1 the molecular
viscosity term was very much smaller than the
eddy viscosity term and could be neglected
for practical purposes, permitting scaling of
the liquid flow. In the present study,
however, the flow regime observed in the
laboratory model cannot be classified as
completely turbulent. Although flows in the
annular region, at and below the skirt, along
the inside wall of the skirt and at the overflow
region are probably turbulent, the large
central region below the circular weir appears
to be laminar. Thus, the laboratory model can
be best characterized as transitional,
containing both laminar and turbulent
regions.
By contrast, the flow regime in the
prototype will be fully turbulent. For
example, a prototype unit represented by a
1:12 scale factor will have a Reynolds
Number 41.5 times larger than that of the
model, which should ensure fully turbulent
flow. Consequently, some differences in the
flow patterns of the laboratory model and the
prototype are to be anticipated.
Scaling of the Particle Flows
Successful scaling of the particle flowfield
hinges on having accurate information
concerning the settling characteristics of the
prototype sewage. Since both the particle
path and concentration field equations can
only be applied to discrete particle settling
velocities, it is necessary to translate settling
column data into this format. This can be
accomplished by dividing frequency
distribution diagrams, such as Figures 72
through 75 into several ranges of settling
velocities. A first approximation of the
removal efficiencies for each range can then
be predicted for a chamber of a given
dimension and flowrate for Equation (55).
E =
Q
+
The computed removal efficiency for
each settling velocity range can be multiplied
by the percentage of particles represented by
that settling velocity and totalled for all
ranges to determine the overall predicted
, removal efficiency. A similar technique can be
utilized to predict the mean particle paths
from settling velocity data.
For particles having settling velocities
much lower than the overflow rate, i.e.,
ws«Q/AB, equation (55) will closely predict
observed removal efficiency. For particles
having settling velocities comparable to or
greater than Q/AB, 100 percent removal
efficiency should be used in lieu of'equation
(55). For these cases, the mathematical model
will substantially overpredict the obtainable
removal efficiency, since it neglects particle
re-entrainment and turbulent diffusion.
Since it was not possible to incorporate
flocculation effects into the mathematical
model, it is necessary to include the effect of
flocculation in the assumed frequency
distribution of particle settling velocities. This'
can be accomplished by increasing the particle
settling velocities based on column settling
.test as described previously.
Equation (55) will always place an upper
limit on the attainable removal efficiencies
and should only be used as a first estimate. To
obtain a more realistic value for the removal
.efficiency, the removal efficiency curves
obtained by the LaSalle Hydraulic Laboratory
can be used. Scaling the laboratory results will
also tend to overpredict the prototype
performance because the increased turbulence
in the prototype will act to resuspend some of
the bottom sediments.
An outline of how the particle and liquid
scaling relationships can be used for design of
a prototype unit will be described later in this
report.
133
-------
Summary
In order to maintain similar liquid flow
conditions in different size swirl
concentrators it is necessary to use Froude
scaling. This will insure that the ratio of the
gravitational to inertial forces will be
maintained constant. Froude scaling requires
that liquid flow velocities be related by the
square root of the scale factor and flowrates
by the 5/2 power of the scale factor per
equations (57) and (58). ;
A simplified equation (55) can |be
employed for obtaining a first estimate |of
removal efficiency for a particle having a
specified settling velocity, given the size and
design flowrate of the prototype unit. To
obtain a more accurate estimate of the
removal efficiency the prototype flowrate and
particle settling velocity can be scaled back to
the laboratory model via equations (57) and
(58). A more accurate removal efficiency can
be obtained from the laboratory data.
COMPARISON OF PREDICTED LIQUID
FLOWFIELD WITH LABORATORY DATA
The mathematical model of the liquid
flowfield was exercised concurrently ; as
modifications to the original chamber
configuration were implemented on the
laboratory model. Since the objective of the
LaSalle Laboratory was to try several
different chamber configurations to achieve
improved performance, no attempt was made
to calibrate the mathematical model for the
earlier configurations. However, certain
quantitative comparisons were made which
closely concurred with the flow patterns
observed by the laboratory. ,
At the beginning of the laboratory mddel
development, the swirl concentrator was
configured as shown in Figure 70. Particmlar
attention is called to the existence of the
central standpipe and the circular overflow
weir. Figures, 77, 78, and 79 show the
streamlines, tangential and vertical velocity
plots of this configuration as predicted by the
mathematical model. A weir diameter of 46 m
(15 ft) prototype scale and flowrate of 500
I/sec (17.5 cfs) prototype scale, were assumed
in these calculations.
Figure 77 illustrates the cross flow
streamlines within the chamber. The bottom
streamline represents one percent of the flow
which is withdrawn through the foul outlet at
the base of the chamber. The remaining
streamlines which, enter beneath the skirt are
in even increments of 20 percent of the flow.
The circular pattern in the upper right hand
portion of the cross section indicates a
clockwise recirculation flow. The streamline
which outlines the large empty region under
the weir defines a counterclockwise
recirculation region. The existence of this
"dead" zone was confirmed in the laboratory
dye studies. The majority of the flow in
Figure 77 is seen to penetrate well into the
base of the separator before turning and rising
to the surface. However, during its upward
passage, the flow tends to hug the outer walls,
creating (undesirable) high upflow velocities.
Furthermore, a significant fraction of the
flow turns immediately upward after passing
under the skirt, and rises directly to the
surface, constituting a hydraulic
"short-circuit." The existence of this
short-circuit was also confirmed in the lab
experiments.
The upflow velocities at various vertical
elevations in the chamber are plotted in
Figure 79 as a function of chamber radius.
Near the top of the chamber (elevation 67 m
[22 ft] prototype scale), the upflow velocity
is zero underneath the weir, and then rises to
a peak of about 1.5 cm/sec (0.038 ft/sec).
The existence of the counterclockwise
recirculation region under the weir is
evidenced by the negative values of upflow
velocity near the inner standpipe. At an
elevation of 51 m (16.7 ft), the upflow
velocity shown in Figure 79 is very high, 2.3
cm/sec (0.075 ft/sec), because the rising fluid
is confirmed to a small annular region
between the two recirculation zones shown in
Figure 70.
This large "dead" region under the weir is
undesirable on two counts. First, it represents
wasted volume, since no flow enters or leaves
this region, the dividing streamline could be
replaced by a solid wall. Second, by reducing
the effective upflow area, it results in a high
upflow velocity which tends to entrain slower
settling particles in the clean overflow.
The calculated tangential contours are
illustrated in Figure 78. These can be
134
-------
100(3)
(10)
200 (7)
RADIUS cm (ft)
FIGURE 77 STREAMLINES FOR INITIAL LAB CONFIGURATION WITH CIRCULAR WEIR
135
-------
I
100
(3)
200 300
(7) (10)
RADIUS cm (ft)
= 0.8
= 0.5
0.3
0.2
= 0.1
= 0.05
400
(13)
FIGURE 78 TANGENTIAL VELOCITV CONTOURS FOR INITIAL LAB CONFIGURATION
WITH CIRCULAR WEIR
136
-------
CM
oas/uiD AJLI0013A
137
-------
compared with the velocities measured in \he
laboratories at various radial cross sectiqns.
The measured values are shown in Figures|49
to 56. Since the mathematical model is
axisymmetric, the calculated velocity contours
represent the average of the hydraulic model
results over the four measurement cross
sections. The calculated values are presented
in terms of V/V0, where V0 is the tangential
velocity under the skirt. Since the vetocity
there is very nearly 30.5 cm/sec (1 ft/sec) in
the LaSalle results, the specified ratios can also
be interpreted directly as velocities in cm/sec.
A comparison of Figure 78' with ; the
hydraulic laboratory results indicates that the
magnitude of the velocity under the weir is
correctly reproduced 0.00033 to 0.00066
cm/sec (0.1 to 0.2 ft/sec), but the velocity in
the outer region near the skirt is too low. The
calculated velocities decay too quickly with
distance from the annular entrance under, the
skirt. As a consequence, the shape of the
velocity profiles do not match the laboratory
results very well. The reason for jthis
mismatch is that no calibration of the model
parameters was attempted. The calculations
were based on the parameters used in the
earlier study of the swirl concentrator as a
combined sewer overflow regulator.1 The
mismatch of the calculated tangential
velocities with laboratory results does not
appear to affect the calculated streamline
patterns, which were verified in the
laboratory, or the conclusions of the study.
In order to eliminate the undesirable
recirculation region under the weir, the design
was modified in two ways: the circular weir
was replaced with a series of eight radial
gutters, and the central standpipe was
removed. The removal of the standpipe
increases the available cross-sectional area,
thereby lowering the average upflow velocity.
The radial gutter provides a more uniform
withdrawal of the flow across the surface.
The mathematical model results for the
revised configuration are illustrated in Figures
80, 81, and 82. The crossflow streamlines
shown in Figure 80 demonstrate that the
recirculation region has been eliminated by
the radial gutters providing a more uniform
upflow. However, contrary to expectation, a
substantial fraction of the flow underneath
the skirt turns sharply and rises up the inner
wall of the skirt, creating a partial short-circuit.'
The existence of this unexpected flow pattern
was also confirmed in laboratory studies.
The clustering of the streamlines near the
skirt results in relatively high upflow velocities
in that region, as shown in Figure 80. Note
the sharp rise in velocities near the skirt at all
elevations greater than 400 cm (13.1 ft).
Nevertheless, the maximum upflow velocity
is less than half of the values occurring in the
earlier configuration, Figure 77, and over the
majority of the cross section, the upflow
velocity is close to the theoretical lower limit
for uniform upflow.
The tangential velocity contours for the
revised configuration are shown in Figure
81. Laboratory measurements for this
configuration are not available, so that direct
comparison is not possible. However, the
vertical orientation of the velocity contours is
in better qualitative agreement with the
hydraulic model results for the earlier
configuration.
Comparison of Math Model Particle Flow
with Test Data
a. Particle Paths: The motion of par-::
tides within the swirl concentrator can be
delineated by superimposing the particle
settling velocity on the liquid flow
streamlines. The resulting particle paths in a
radial cross section of the separator are then
obtained by plotting the resulting particle
stream function. The results of this
calculation for the initial configuration are
shown in Figure 83. Only small differences
between the particle paths, Figure 83, and the
corresponding liquid flow streamlines, Figure
80, are apparent. The particles appear to
penetrate further into the base of the separator
as one would expect, and some of the particle
paths terminate on the bottom, indicating
settlement. The dividing streamline between
the material which settles out, and that which
is entrained in the overflow is marked by the
symbol (i*).
Figure 83 also shows the existence of a
particle path (symbol 0) within the closed
recirculation region under the weir. This
• particle path originates at the underside of the
weir, and is a consequence of assuming a
138
-------
600
(20)
100(3)
300(10)
400(13)
FIGURE 80 STREAMLINES FOR MODIFIED LAB CONFIGURATION
WITH RADIAL GUTTERS AND STANDPIPE 'REMOVED
139
-------
100
(3)
200 300
(7) RADIUS cm (ft) <10>
400
(13)
FIGURE 81 TANGENTIAL VELOCITY CONTOURS FOR MODIFIED LAB CONFIGURATION
WITH RADIAL GUTTERS AND STANDPIPE REMOVED
140
-------
141
-------
100(3)
^RADIUS cm (ft)300 «°>
400(13)
FIGURE 83 PARTICLE PATHS FOR INITIAL LAB CONFIGURATION WITH CIRCULAR
WEIR AND PARTICLE SETTLING VELOCITY OF 0.05 cm/scsc (0.0016 ft/sec)
142
-------
settling velocity of 0.05 cm/sec (0.0016
ft/sec) everywhere so that particles appear to
fall through the weir. Because the weir is
actually a solid plate, the recirculation region
under the weir (defined by the symbol t)
would generally be devoid of particles, except
for those scattered into this zone by
turbulence. This phenomenon is observed
when operating the laboratory model with
IRA-93 resin. As the material enters the skirt,
it initially flows down toward the base of the
separator, but then tends to concentrate in
the outer area near the skirt as it rises upward.
For a considerable period of time, the region
under the weir remains void of particles.
Particle patfis for the revised chamber
configuration are shown in Figures 84, 85,
and 86. The total flowrate is 1 I/sec (0.034,
cfs) (model scale) as before, but no flow is
withdrawn from the bottom.
In Figure 84, the settling velocity is 0.05
cm/sec (0.0016 ft/sec), as in Figure 83 for the
earlier configuration. Even with the more
uniform upflow velocity the particles in
Figure 84' still concentrate near the skirt as
they rise. Figure 85 also shows an interesting
effect: as the particles rising near the skirt
approach the surface they are transported
radially inward. They then tend to sink for a
time before finally being re-entrained in the
overflow, (see particle path with symbol 0 in
Figure 84). Again, the rising and falling
motion of individual particles near the surface
has been observed in the lab studies.
Figures 85 and 86 show the r particle
behavior at progressively larger settling
velocities. In Figure 85, the settling velocity
of 0.252 cm/sec (0.0083 ft/sec) is exactly
equal to the uniform upflow velocity
specified at the free surface. Consequently,
none of the particles can escape through the
top. In Figure 86 the settling velocity of 0.3
cm/sec (0.0098 ft/sec) exceeds the upflow
velocity, and most of the particles drop like
rocks to the bottom. A few, however, are
entrained in the short-circuit hydraulic path
up the inside of the skirt and tend to circulate
within a confined region near the outer wall.
b. Removal Efficiency: Removal
efficiencies were computed for the final
laboratory configuration utilizing Equation
(55).
Since there was no underflow for the
laboratory model runs, 17 was set equal to zero
in Equation (55). In order to identify the
value ofws to associate with the laboratory
material, the settling velocity distribution
curves prepared by Beak were utilized. A wide
variation in the settling velocity of the resin
was observed. Beak reported that these
differences are attributable to different
sieving techniques, and that Run 3 is
representative of the materials employed by
th? laboratory for the model testing. The data
by Beak were used to tabulate a step
distribution for use with equation (55), as
presented in Table 23. Table 23 was prepared
by dividing the abscissa axis of Beak's Figure
10 into 10-percent segments, using the median
settling velocity as representative of each
particle fraction.
Using the settling velocity distributions of
Table 23, an upper limit on removal
efficiency can be predicted from equation
(55) by computing a removal efficiency for
each 10 percent fraction, totalling the results
and dividing by 10. The results of this
computation are shown in the second column
of Table 23. If Run 3 is considered as most
typical for the material used in the laboratory
model tests, the mathematical model predicts
33.2 percent removal as the maximum limit
on the performance at the 0.5 I/sec flowrate
for 100-200 mesh resin.
The actual removal observed by LaSalle/
Hydraulic Laboratory was about 50 percent.
Four possible explanations for this apparent
discrepancy between the mathematical model
predictions and laboratory data were offered,
1. Resin material may swell on contact
with water, with consequent
variations in settling velocities.
2. Particles may experience electrostatic
attraction to the tank walls
3. Particles may not be completely
dispersed as single particles.
4. Stratification may occur in the region
under the skirt, giving higher
concentration along particle paths
which reach the bottom.
Subsequent investigations have ruled out
possibilities 1-3. However, measurements at
LaSalle have shown that stratification of the
material does occur under the skirt. The
143
-------
D
19
CD
600 ID
(20) rD
CD
CD
§
CD
e
500
(16)
*
~ 400
Z (13)
O
i
CD
CD
CD
"s
CD
CD
— CD
CD
CD
S -•
" •
300
(10)
CD
a
CD
200
(7)
100
(3)
CD
a
GO
§
o
-------
600
(20
500
(16
C 400
1 (13)
IU
300
(10)
200
(7)
100
(3)
mm
m
I
•
m
m
m
m
I
* A +
» A 4-
1 * +
* * •*•
* * m"
I I
100(3)
200(7) 300(10)
RADIUS cm (ft)
400(13)
FIGURE 85 PARTICLE PATHS FOR MODIFIED LAB CONFIGURATION WITH RADIAL
GUTTERS AND STANDPIPE REMOVAL AT PARTICLE SETTLING VELOCITY
EQUAL TO AVERAGE UP FLOW VELOCITY OF 0.252 cm/sec (0.0083 ft/sec)
145
-------
I
LU
100 (3!
400(13)
200(7) 300(10)
RADIUS cm (ft)
FIGURE 86 PARTICLE PATHS FOR CODIFIED LAB CONFIGURATION WITH RADIAL
GUTTERS AND STANDPIPE REMOVED AT PARTICLE SETTLING VELOCITY
OF 0.3 cm/sec (0.0098 ft/sec)
146
-------
TABLE 23
PARTICLE SETTLING VELOCITY
DISTRIBUTION FOR 100-200 MESH IRA-93 RESIN
(From Beak Report)
Particle
Class
90-100
80-90
70-80
60-70
50-60
40-50
30-40
20-30
10-20
0-10
median
settling
velocity
Particle Settling Velocity cm/sec (ft/sec)
Run 1 Run 2 Run 3
cm/sec ft/sec cm/sec ft/sec cm/sec ft/sec
0.051
0.037
0.031
0.0262
0.023
0.020
0.0165
0.011
X
X
0.0215
0.002
0.0012
0.001
0.0008
0.0007
0.0006
0.0005
0.0004
X
X
0.0007
0.130
0.083
0.050
0.043
0.032
0.022
0.01
X
X
X
0.026
0.004
0.003
0.002
0.0014
0.001
0.0007
0.0003
X
X
X
0.0009
0.102
0.079
0.052
0.050
0.042
0.034
0.028
0.020
0.012
X
0.038
0.003
0.0025
0.0017
0.0016
0.0014
0.0011
0.0009
0.0007
0.0004
X
0.0012
TABLE 24
PREDICTED REMOVAL EFFICIENCY FOR 100-200 MESH IRA-93 RESIN
(0.5 I/sec [0.02 cfs])
Efficiency
(Excluding annular area)
Settling Velocity
Distribution^
Revised Efficiency
(Including annular area)
Runl
Run 2
Run 3
Percent
17.1
29.4
33.2
Percent
28.4
48.4
55.1
significance of this finding is that the outer
annulus also acts as a settling chamber, so that
the projected base area, AB in equation (55),
should include the area of the annulus.
However, the stratification in the outer
annulus will be affected by the rotational
velocity in the annulus. Increasing the inlet
velocity (to prevent sedimentation in the inlet
sewer) may disrupt this stratification and
result in poorer performance.
If the projected base area of the annulus
is included in the AB term in equation (55),
substantially greater removal efficiencies
result, as given in the third column in Table
24. The revised result from Run 3 is 55.1
percent which is in close agreement with the
observed removal efficiency of 50 percent.
To test the validity of the mathematical
model further, the mathematical model was
applied to the work of Bernard Smisson at
Bristol, England. Mr. Smisson pioneered the
application of swirl concentration in
Europe18 and has amassed data on the use of
swirl concentrators for primary
sedimentation. Through correspondence with
Smisson, a summary of his results using'a 3.02
m (10 ft) diameter prototype chamber to
settle activated sludge was obtained. A scatter
diagram showing his measured efficiencies
Versus those calculated from equation (55) is
shown in Figure 87. Although substantial
numbers of his data points fall above the
theoretical upper limit, it is believed that this
is due to the flocculent nature of the activated
147
-------
2
Q
III
u o> f
s < *
111
« CM O
sift
!;.*!
S > S2
o o
oo
(&> Q)
ooooooV oooo
00
00
to
o
LU
a.
I
o
cc
Q.
EO
co z
u. O S
LU co CC
Q £20
O u" Q-
ui O
°" o o
-H
CC LU
tK
S LU
OX
O I-
00
LU
DC
O
(0
O
in
o
o
AON3IOIJJ3
CO
o
148
-------
sludge. The calculations are based on a settling
velocity of 0.00182 cm/sec (0.0007 ft/sec),
measured by Smisson in a dilute suspension.
In the actual unit, the high inlet concentration
will result in substantial increases in the
settling velocity due to agglomeration of the
primary particles. This possibly was noted by
Smisson in his correspondence.
Additional comparisons were made with
Smisson's results, using a 78 cm (30 in)
diameter conical bottom tank. In this work,
Smisson employed polystyrene molding beads
with settling velocity of 0.038 cm/sec (0.0146
ft/sec). .Measured separation efficiencies
ranged from 50 to 98 percent over a range of
flowrates from 0.45 to 01.25 m3/sec (0.016 to
0.044 ft3/sec). Equation (55) predicts a
separation efficiency of 100 percent for this
material over the entire flowrate range.
Consequently, this model configuration
appears to be less efficient than the version
developed at LaSalle Hydraulic Laboratory,
since the latter gives separation efficiencies
which are closer to the theoretical limit.
The adjustment of AB in equation (55) to
include the annulus area was retained for all
subsequent comparisons with laboratory data,
and the modified equation was used to
conclude these data, as shown in Figure 88,
which illustrates the observed removal
efficiencies as a function of the nondimensional
settling parameter (wsAB/Q) for several
modifications to the chamber geometry. The
theoretical efficiency as predicted by equation
(55) is also shown. Data points plotted as
circles in Figure 88 correspond to an early
configuration using four radial gutters, with
the central standpipe in place, and with the
original 10 x 10 cm (4 x 4 in) inlet sewer.
Subsequently, the four radial gutters were
replaced with the 8-gutter configuration, with
no standpipe. At the same time, the inlet
dimensions were reduced to 6 x 6 cm (2.3 x
2.3 in) to raise the inlet velocity in order to
prevent settling in the inlet sewer. Data points
corresponding to this configuration are plotted
with triangles in Figure 88. The recovery
efficiency with thismodificationwassomewhat
lower, probably due to the higher velocity
which tends to resuspend some of the sediment
which reaches the bottom. In the final
configurations, the skirt diameterwas decreased
from 71 to 61 cm (2.33 to 2 ft), resulting
in somewhat better performance, 41 percent
1001—
THEORY
LEGEND
0-4 GUTTERS WITH STANOPIPE. 10 cm (4 in) INLET
71 cm (28 in) SKIRT
A - 8 GUTTERS, NO STANDPIPE, 6 cm (2.4 in) INLET
71 cm (28 in) SKIRT
Q - 8 GUTTERS, NO STANDPIPE. 6 cm (2.4 in) INLET
61 cm (24 in) SKIRT
1 L 1 1 1 1 1 1 I I | I | | | ,
, 1.0 "'
SETTLING PARAMETER
ws AB'
3.0
4.0
FIGURE 88 CORRELATION OF OBSERVED AND
THEORETICAL MODEL PERFORMANCE
149
-------
versus 37 percent, as determined with
IRP resin at 0.5 I/sec (0.018 cfs). However,
somewhat poorer performance with this
configuration was obtained using the Arizona
Road Dust, 28 percent versus 36 percent: In
all cases, the laboratory data fall on or below
the theoretical upper limit shown as a solid
line in Figure 87. At high flowrates, where the
settling velocity is less than the overflow rate,
the laboratory model closely approaches the
theoretical upper limit. At the lower flowrates,
larger values of wsAB/Q, the laboratory
results fall substantially below the upper limit
due to increased turbulent diffusion and
resuspension of bottom sediments.
The data shown in Figure 88 include
results with IRA-93 resin with sieve sizes
100-200 (standard fraction) and 50-100
(coarser fraction). Results with the Arizpna
Road Dust are also included. 1
Since the Arizona Road Dust has a Very
narrow range of particle sizes, the median
settling velocity can be used to arrive at an
approximate overall removal efficiency | for
the entire distribution. For example, using the
median, or 50 percent, settling velocity from
Figure 10 of the Beak report, a rempval
efficiency of 36.6 percent is obtained as
compared to the 36.3 percent calculated by
dividing the distribution of settling velocities
into 10 percent sections. ;
For the Run 3, 100-200 mesh IRA resin
at a flowrate of 0.5 I/sec (0.018 cfs), the
medial settling velocity of 0.037 cm/sec
(0.0012 ft/sec) yields a calculated rempval
efficiency of 56.8 percent, as compared to the
55.1 percent calculated for the 10 percent
sections. Thus, for the 100-200 mesh IRA-93
resin, which also has a relatively narrow size
range of 74 u to 149w, using the median
settling velocity provides a good estimate of
the removal efficiency. This is an important
point, since it will facilitate relating | the
removal efficiency observed in the laboratory
to the prototype design.
Effect of Scale on Chamber Performance
The objective of the modeling work
has been to provide an indication of hoW a
prototype swirl concentrator can be expected
to perform on actual sewage. To this end, it is
necessary to understand the limitations [and
validity of the scaling laws. As has been
indicated, chamber sizes, flows and particle
settling velocities are scaled to represent
prototype units utilizing Froude number
scaling. This scaling procedure works quite
well for fully turbulent flows, which tend to
be independent of Reynolds Number.
However, it was noted that while fully
turbulent flows can be anticipated in the
prototype units, the flow observed in the
laboratory model at the lower flowrates was
partly laminar. Thus, some can be anticipated
in the flowfields of the model and prototpye,
especially in those portions of the chamber
where laminar flow was observed in the
model.
Two approaches can be used to scale the
removal efficiency. The most straightforward
approach is to use equation (55) directly to
predict the removal efficiency for a given
overflow rate and prototype settling velocity.
However, this equation neglects turbulent
resuspension of bottom sediment and,
therefore, will always overpredict the actual
removal efficiency when the particle settling
velocity is lower than the overflow rate. When
the particle settling velocity is of the same
magnitude or larger than the overflow rate,
equation (55) will overpredict the removal by
a considerable amount. Under these
conditions, the second approach, using the
laboratory data, should provide a more
realistic estimate of removal efficiency.
The second approach is based on the
condition that the prototype unit will give the
same removal efficiency as the laboratory
model if dimensions, flows, and velocities are
related by Froude number scaling. However,
Froude number scaling also preserves the
equality of wsAB/Q in the model and
prototype, so that
prototype
Therefore if the non-dimensional settling
parameter (wsAB/Q) is calculated for the
prototype, the corresponding efficiency can
be determined for the correlation of lab
data in Figure 26 •
This approach also overpredicts the
prototype removal efficiency. The higher
Reynolds Number in the prototype results in
150
-------
greater turbulent dispersion and particle re-
entrainment, particularly at model flowiates
less than 0.5 I/sec (0.018 cfs) where laminar
flowoccursin the laboratory model. However,
the removal efficiency as scaled from the
laboratory model should be closer to the
prototype performance than that obtained
from equation (55).
The importance of characterising the
settling properties of the sewage should be
emphasized. The data in the Beak report
shows large variations in settling velocity
distributions. Since the actual sewage win
have a wide band of settling velocities, the
settling data should be divided into
percentage bands and removal efficiencies
computed separately for each band.
Since particle re-entrainment causes a
lower removal efficiency, excessive buildup of
sludge on the bottom, if allowed, wM also
cause reductions in removal efficiency.
Flocculation effects are difficult to assess
and can only be represented by obtaining the
settling velocity distribution after flocculation
has occurred, as described in earlier sections.
influence of Geometric Variables
on Chamber Performance
The use of the radial gutter configuration,
in lieu of the circular weir, produced a
marked improvement in the performance of
the device. This can be explained from the
mathematical model results by comparing the
vertical velocity profiles for the two
configurations (Figures79 and 82).The radial
gutters resulted in more uniform upflow
velocities near the surface of the tank than
with the circular weir. As a result, the average
upflow velocity used in equation (55) is closer
to the actual value occurring at every point at
the chamber surface, and the removal
efficiency more closely approaches the
theoretical upper limit predicted by equation
(55). It is desirable to maintain this surface
velocity as uniform as possible to minimize
particle entrainment and short-circuiting.
A second very important geometric
variable is the size of the annular region. At
the flow velocities encountered in the
laboratory model, this annular region served
as an additional area in which particle settling
could occur. In order to insure that the flow
conditions in the prototype unit will also
allow the annular region to participate in
particle settling it is necessary to insure that
significant particle resuspension and
entrainment does not occur. Since the
mechanism of particle entrainment and scour
can be shown to depend on the ratio of the
particle settling velocity to the eddy viscosity,
it is necessary to maintain similarity of this
ratio between the model and the prototype.
For fully turbulent flows which are
independent of Reynolds Number, Froude
Number scaling wiH preserve this ratio.
However, some Reynolds Number, effects
may occur in this annular region. The
tangential flow velocity in the annulus
depends upon a balance of the angular
momentum of the entering fluid and the drag
losses on the wafts of the annulus. Thus, if the
inlet sewer velocities in the model and
prototype are related via the Froude Number,
the annulus velocity may not be so related
because of differences in the drag losses at the
different Reynold Numbers. The fact that the
model Reynolds Number is smaller is
counterbalanced by its smoother walls, so
that, in general, tne Froude. Number
differences between the model and prototype
should be small.
Since the tangential velocities in the
annular region are higher than in the main
chamber, its efficiency as a settling device is
probably less than that of the inner region.
Consequently, increasing the size of this
region at the expense of the inner region
could be expected to result in somewhat
poorer performance. Enlargement of the
annulus should also result in somewhat
greater tangential velocities for a given inlet
sewer velocity, because the smaller surface
area will cause lower drag losses. The larger
tangential velocity will also tend to degrade
the performance because of increased
resuspension of bottom sediment.
However, as shown by the data in Figure
88, somewhat better performance is actually
achieved with the 61 cm (2 ft) skirt than with
the 71 cm (2.3 ft) skirt. This may be the
result of the complex flow patterns which
develop in the annulus with the smaller skirt.
The annulus tangential velocity is not
constant across the width of the annulus, but
151
-------
contains an outer high velocity region and an
inner low velocity region. This fact tends to
invalidate the heuristic arguments for poorer
performance noted above.
DESIGN TECHNIQUES
An example will be given-to illustrate
how the results of the mathematical model
can be applied to estimate the rernoval
efficiency of an actual prototype unit. The
following dimensions and design flowrate
have been assumed for a hypothetical
example. ;
The sewage settling velocity distribution
will be assumed to be that shown by the curve
in Figure 75. This curve has been reduced to
the step distribution shown in Table 25. The
theoretical upper limit of removal efficiency
of 46 percent was calculated from equation
(55), with AB computed for the full 3.65 m
(12 ft) diameter, thus including the annular
region between the skirt and wall of the
effective settling zone. For settling velocities
greater than the up flow velocity a 100
percent removal efficiency was assumed. It
should be emphasized that the 46 percent is
a theoretical upper limit on the removal
efficiency that can be obtained with the
prototype unit. If median particle settling
velocity of 0.055 cm/sec (0.0018 ft/sec) had
been used instead of the assumed distribution
function, equation (55) would predict 36.2
percent removal. This emphasizes the need to
have a good definition of the settling
properties of the sewage.
A second estimate of the prototype
removal efficiency can be obtained from the
performance curves for the laboratory model
as shown in Figure 88. This figure shows the
removal efficiency measured by the
laboratory for different material settling
velocities and flowrates. To provide a
convenient mechanism for predicting
prototype removals, the data have been
nondimensionlized and plotted as the removal
efficiency versus the settling parameter
wsAB/Q, where ws equals the particle median
settling velocity AB equals the projected
cross-sectional area, and Q is the volumetric
flowrate. Smooth curves (dash line) have been
drawn through the data points approaching a
horizontal asymptote of 92.6. The 92.6
percent removal represents the value observed
by the laboratory for the 150-300 mesh IRA
resin having a settling parameter value of
6.85. This point is off the scale of Figure 88,
and is not shown.
TABLE 25
PREDICTED REMOVAL EFFICIENCY FOR HYPOTHETICAL
PROTOTYPE UNIT USING MATHEMATICAL MODEL
Percent Particles
of Settling Velosity
less than vs
95
85
75
65
55
45
35
25
15
5
Percent
Percent Particles Settling Velocity Removal as
of Settling Velocity vs Computed from
vs cm/sec (ft/sec) Equation (55)
10 ;
10
10 ;
10 i
10
10
10 :
10
10
10
0.31
0.22
0.16
0.11
0.07
0.04
0.02
0.003
(0.01)
(0.007)
(0.005)
(0.004)
(0.002)
(0.001)
(0.0007)
(0.00009)
Non-Settleable
Non-Settleable
10.0
10.0
10.0
7.2
4.6
2.6
1.3
0.1
0
0
Total Predicted Removal
46.0%
Prototype ?= °-57 cfs = 0.004 fps (0.152 cm/sec)
Overflow Rage , 00(6ft)2
152
-------
The solid line shown in Figure 88
represents the theoretical upper limit on
removal efficiency, as calculated from
equation (55). As noted earlier and confirmed
in this figure, there is close agreement
between the laboratory and mathematical
model when the settling velocity is
substantially less than the overflow rate, i.e.
values of the settling parameter less than 0.5.
Since the settling parameter used in
preparing Figure 88 is dimensionless, the
curve representing the laboratory data in
Figure 88 is independent of scale. Therefore,
removal efficiencies as predicted by the
laboratory data can be readily obtained from
Figure 88 by entering the curve with the
appropriate value of wsAB and Q. Using this
approach, the removal efficiencies for the
hypothetical example were recalculated using
the lower laboratory curve in Figure 88. The
results are tabulated in Table 26, Predicted
Removal Efficiency for Hypothetical
Prototype Unit Using Laboratory Data. The
total removal efficiency obtained from the
laboratory curve is 34.7 percent, as compared
to the 46 percent theoretical upper limit.
As noted earlier, both the mathematical
model and the laboratory data will tend to
overpredict the prototype removal efficiency.
Consequantly, it may be wise to make
allowance for this by increasing the assumed
design flowrate by a safety factor when
calculating predicted removal efficiencies.
TABLE 26
PREDICTED REMOVAL EFFICIENCY FOR
HYPOTHETICAL PROTOTYPE UNIT USING
LABORATORY DATA
% Particles
of Settling
Velocity
Vs
10
10
10
10
10
10
10
10
10
10
Prototype Settling
Velocity
cm/sec (ft/sec)
0.31 (0.01)
0.22 (0.007)
0.16 (0.005)
0.11 (0.004)
0.07 (0.002)
0.04 (0.001)
0.02 " (0.0007)
0.003 (0.00009)
Non-Settleable
Non-Settleable
Settling
Parameter
VsAB
Q
2.04
1.45
1.05
0.72
0.46
0.26
0.13
0.01
Removal
Weighted
From
Figure
8.2
7.4
6.3
5.1
3.8
2.5
1.3
0.1
0
0
Total Predicted Removal 34.7
153
-------
a = mean radius, acceleration
A = area
GO = initial particle concentration
C = particle concentration
E = efficiency
•g = unit vector ;
/ = nondimensional stream function
F = particle mass flux ]
g = gravitational acceleration :
G - nondimensional tangential
velocity function
G' = shear rate of flow ;
/ = mixing length
L = reference length
JVop = total particle concentration
My = particle collision rate
n = particle number density; particle
distribution function
n = unit normal vector
p = pressure
Q = volume flowrate
R = ws&nnjep
Re = Reynolds number
r = radial coordinate; particle radius
S = scaling factor
Sg = specific gravity
t = time
"re/ = reference velocity
u = radial velocity component
v = particle volume
v - tangential velocity component
w = vertical velocity component
ws = particle settling velocity
w = average upflow velocity
Z = depth location of column
sampling port ,
z
e
1
r
V
I
p
a
e
CO
Subscripts
in,
axial coordinate
eddy viscosity
particle eddy diffusivity
energy dissipation per unit mass
nondimensional axial coordinate
vorticity; underflow/overflow
ratio
molecular viscosity
kinematic viscosity
nondimensional radial coordinate
(r/L)
liquid density
standard deviation
volume fraction of particles
scale size
stream function
nondimensional distribution
function
reference frequency
nondimensional vorticity
function
boundary value, or bottom
value at point adjacent to
boundary
inflow
liquid
particle
Superscripts
vector quantity
154
-------
REFERENCES
1. American Public Works Association, "The
Swirl Concentrator as a Combined Sewer
Overflow Regulator." EPA Report No.
EPA£R 2- 72-008, NTIS No. PB 214134,
Sept. 1972
2. Camp, Thomas R., "Sedimentation and
the Design of Settling Tanks," Trans
ASCE; Vol HI, 1946 (pp 895-958)
3. Saffman, P.O., and Turner, J.S., "On the
Collision of Drops in Turbulent Clouds,"
/. Fluid Mechanics, Vol. 1, 1956 (pp
16-30)
4. Swift, D.L., and Friedlander, S.K., "The
Coagulation of Hydrosols by Brownian
Motion and Laminar Shear Flow," /.
Colloid Set, Vol. 19, 1964 (pp 621-647)
5. Levich, V.G., Psycho chemical
Hydrodynamics, Prentice-Hall, Inc.,
Englewood Cliffs, N.J., 1962 (pp 175,
219)
6. Ives, K.J., and Bhole, A.G., "Theory of
Flocculation for Continuous Flow
System," Proc. ASCE, J. Env. Eng. Div.,
Bol 99, No. EE1, Feb. 197.3 (pp 17-34)
7. Hidy, G.M., "On the Theory of the
Coagulation of Noninteracting Particles in
Brownian Motion," /. Colloid ScL, Vol.
20, 1965 (pp 123-144)
8. Gemmel, R.S., "Some Aspects of
Orthokinetic Flocculation," thesis
presented to .Harvard University, at
Cambridge, Mass, in 1963, in partial
fulfillment of the requirement for the
degree of Doctor of Philosphy.
9. Fair, G.M., and Gemmel, R.S., "A
Mathematical Model of Coagulation," /.
Colloid ScL, Vol. 19, 1964 (p 360)
10. Friedlander, S.K., "Similarity
Consideration for the Particle-Size
Spectrum of a Coagulating, Sedimenting
Aerosol,"/. Meteorology, Vol. 17, No. 5,
Oct. 1960 (pp 80-84)
11. Chen, Cheng-lung, "Sediment Dispersion
in Flow with Moving Boundaries," /.
Hyd. Div., Proc. ASCE, Vol. 97, No.
HY8,Aug. 1971 (pp 1181-1201)
12. Reynolds, A.J., "Waves on the Erodible
Bed of an Open Channel," /. Fluid Mech.,
Vol. 22, Part 1, 1965 (pp 113-133)
13. Hyashi, T., "Formation of Dunes and
Antidunes in Open Channels," /. Hyd.
Div., Proc. ASCE, Vol. 96, No. HY2, Feb.
1970 (pp 357-366)
14. Kennedy, J.F., "The Mechanics of Dunes
and Antidunes in Erodible - Bed
Channels,"/. Fluid Mech., Vol 16, Part 4,
1963 (pp 521-544)
15. Engelund, F., "Instability of Erodible
Bed," /. Fluid Mech., Vol. 42, Part 2,
1970 (pp 225-244)
16. Smith, J.D., "Stability of a Sand Bed
Subjected to a Shear Flow of a Low
Froude Dumber," /. Geophy. Res., Vol.
75, No. 30, Oct. 1970 (pp 5928-5940)
17. Gradowczyk, M.H., "Wave Propagation
and Boundary Instability in Erodible Bed
Channels," /. Fluid Mech., Vol. 33, Part
1, 1968 (pp 93-112)
18. Smisson, B., "Design, Construction, and
Performance of Vortex Overflow,"
Symposium on Storm Sewage Overflows;
Institute of Civil Engineers, 1967.
155
-------
APPENDIX C
SETTLEABILITY TESTS AND HYDRAULIC CHARACTERIZATION OF A PILOT
(WITH REAL SEWAGE) SWIRL SEPARATOR AS A PRIMARY TREATMENT FACILITY
This appendix reports the results of the
characterization of the settling properties of
combined sewer overflow currently being
used in studies designed to evaluate the
efficiency of the pilot swirl as a primary
separator. During two storm events in jFune
1975, the suspended solids concentration in
the influent channel of the Humber
Treatment Plant (Metropolitan Toronto) was
found to increase to about 600 mg/1. A
distinct improvement in the settleability was
found even after peak flows had occurred.
The settling characteristics of the combined
overflow was very similar to the Amberlite
IRA-93 anion exchange resin developed for
laboratory scale studies conducted in an
earlier phase of this project.1
Hydraulic characterization of the swirl
separator and the primary clarifier used in this
study were conducted using a dye tracer
response technique. The hydraulics of the
swirl separator was found to be more closely
related to plug flow regime than the primary
clarifier. Theoretically, the closer to a plug
flow mixing regime, the better should be the
solids removal efficiency.
The apparent decrease in active vojume
and the shift toward plug flow which occurs
as flow to the swirl increases, suggests that the
swirl separator should respond well as a solids
separator for combined sewer storm flow in a
sewage treatment plant.
SETTLEABILITY TESTS
A pilot, swirl separator has been
constructed at the Humber Treatment Plant
operated by Metropolitan Toronto.
Comparison studies of the removal
efficiencies of suspended solids in the pilot
swirl separator and primary clarifier were
conducted during periods of normal flow and
also during periods of storm flow. As part of
these studies, Beak Consultants 'Ltd.,
conducted settleability tests of the combined
storm flow and compared the hydraulic
characteristics of the two settling devices.
Methods for Characterization of Settleability
The settling characteristics of the storm
flow solids were measured on two separate
occasions during the month of June 1975.
Samples were collected in the influent
channel of the primary clarifier. Samples were
collected by the plant operators during
periods of increased flow as a result of storm
flow entering the plant. Arrangements were
made to sample the influent during rising,
peaking, and falling flows. The collected
samples were picked up from the plant by
Beak and all settling tests were conducted
within 18 hours of sampling.
Both storm events in June occurred in the
late evening when the plant flow was about
1,748 I/sec (40 mgd). Daily average plant flow
is approximately 3,058 I/sec (70 mgd). The
first event studied occurred at 2300 hours on
June 4th, continued three hours until about
0200 hours on June 5th. The peak flow was
5,462 I/sec (125 mgd) during this event.
On June 16, a shorter, more intense
storm increased plant flows from 2200 to
2300 hours producing a peak of 6,816 I/sec
(156 mgd). In addition to the collection of
rising, peaking, and falling stages in the
primary influent channel, the influent to the
swirl separator was tested for settleability at
peak flows. The swirl influent and the
primary clarifier influent differ only by the
fact that the swirl influent is pumped out of
the primary influent channel. This pumping
step is the only difference between the waste
received by each settling device.
The samples were tested for settleability
using the same method used by Beak in its
previous study.1 This method is described in
Exhibit 1 of this appendix.
Discussion of Settling Results
The results of the settleability tests are
presented in Figures 89 and 90. The solids
settling characteristic of the combined sewer
flow is expressed as the percent removal of
suspended solids as a function of the upflow
156
-------
EXHIBIT 1
SETTLING COLUMN TEST METHODS
Procedure
The test column consists of a 2.03 m (6.66 ft) high, 20 cm (8 in) diameter, Plexiglas® cylinder
with sampling ports as 1 ft increments. The bottom of the cylinder is fitted with a water-tight base
.30.5 cm (1 ft) diameter, to give a stable base during the test run.
Where possible, a 56.7 1 (15 gal) sample should be collected and the settling test run
immediately to prevent any changes in the sample. The most important variable is temperature and
where possible, the test should be performed before any great change occurs. In most cases it is not
practical (or meaningful) to attempt to adjust the sample temperature to that of the ambient
temperature where the test is being performed. The temperatures of the sample in the column
should be recorded at the start and the finish of the test run. The samples should be mixed
thoroughly and dumped into the test column as quickly as possible. To make sure of thorough
mixing m the column, a hand-made plunger was used to agitate the contents throughout the depth
of the column The timer is then started and the column is sampled at each port immediately
Starting irom the top of the column, the ports sampled are at the 30.5, 61, 91, 122 152 5 and
167.8 cm (1,2, 3, 4, 5, and 5.5 ft) levels.
These "time zero" samples are averaged to provide the initial suspended solids of the sample in
the column. The column is then sampled from each port at convenient time intervals The suggested
intervals are: 10, 20, 40, 60, 80, 100, and 120 minutes. '
The samples withdrawn from each sample port (except the bottom one) should be collected in
small containers (approximately 500 ml) to be analyzed for suspended solids. Care must be taken to
flush out each sample port before the sample is taken. The filter paper used for this analysis should
be Whatman GF-C or equivalent.
The depth of liquid in the column should be recorded initially and after each set of samples has
been removed. It is more convenient to measure all depths from the top of the column
* . T£f. per
-------
100-
90-
o
V)
•a
03
•o
I
W
80-
70-
EC 60-
50-
LEGEND
© Rising Storm
Q Peaking before Pump
& Falling Storm
INITIAL SUSP. SOLIDS
538 mg/l
516 mg/l
627mg/l
0.002
0.00079
0.004
0.0016
0.006
0.0024
1 1
0.008 ;0.01
0.0031 0.0039
0.02
0.0079
0.03
0.012
0.04
0.016
o.os
0.024
0.08
0.032
0.1
0.039
Overflow Rate cm/sec (in/sec)
FIGURE 89 SETTLING VELOCITY CHARACTERISTICS - HUMBER TREATMENT PLANT
Storm Flow June 5,1975
90-
•s
w
•O 80-
*
I
3
s™
•ra
I
CC
60-
50-
LEGEND
o Rising Storm
n Peaking after Pump
13 Peaking before Pump
& Falling Storm
INITIAL SUSP,SOLIDS
1 327 mg/l
287 mg/l
494 mg/l
605 mg/l
0.002
000079
0.02
0.0079
0.03
0.012
0.04
0.016
0.06
0.024
0.08
0.032
0.1
0.039
1 1 1 1
0.004 0.006 0.008 0.01
0.0016 0.0024 0.0031 0.0039
Overflow Rate cm/sec (in/sec)
FIGURE 90 ..SETTLING VELOCITY CHARACTERISTICS - HUMBER TREATMENT PLANT
Storm Flow June 16,1975
158
-------
velocity or overflow rate in a settling device.
During the storm of June 5, 1975, suspended
solids were found to increase from 538 mg/1
to 627 mg/1. The settleability of the
combined sewage flow improved as the storm
progressed. In fact, the predicted percent
removal for a given overflow rate is greater for
the combined sewage during the falling storm
hydrograph compared to the peak flow
condition. For a constant overflow rate of
20 I/sec (0.45 mgd) or 162 m3/day/m2 (4,000
gpd/ft2), the predicted percent solids removals
are 74, 76, and 81 percent respectively for
rising, peaking, and falling storm events.
The second storm monitored was shorter;
however, the storm was more intense,
producing a larger peak flow. The suspended
solids concentration increased from 327 mg/1
to 494 mg/1 to 605 mg/1 as the storm
progressed through the rising, peaking, and
falling stages. For a constant overflow rate of
20 I/sec (0.45 mgd), the predicted percent
solids removal is 60, 80, and 83 percent
respectively for rising, peaking, and falling
storm conditions.
A sample of the peaking flow after the
lift pump to the swirl separator had lower
suspended solids and poorer settling
characteristics than the sample collected
before the pump. In fact, the settling
properties of the sample taken after the pump
appears to be similar to the rising flow sample
prior to the pump. Although this indicates
that the pump changed the characteristics of
the effluent, we do not believe that this is the
correct conclusion. First, the suspended solids
concentration should not change as a result of
pumping. The fact that samples taken before
and after the lift pump were different in
suspended solids, suggests that the samples
were not of the same material.
In addition, during this short, intense
storm, lasting only one hour, the influent
varied in quality and quantity over a brief
period. The influent appeared to change in
character significantly between the time the
samples were collected, before and after
pumping. As the solids concentration in the
after pump sample is similar to the rising storm
condition sample, and the settleability charac-
teristics are identical, the pump did not appear
to affect the settling properties of the solids.
Conclusions on Settleability
The combined sewer overflow events at
the Humber Treatment Plant resulted in an
increase in the suspended solids concentration
and an improvement in the settling velocity
characteristics of the waste. The solids in the
influent were most settleable as the storm
progressed through the rising, peaking, and
i falling stages. For the seven tests conducted,
between 60 to 83 percent of the solids had a
settling velocity of 0.2 cm/sec (0.00656
ft/sec), or more. In comparison,- the IRA 93
resin simulated sewage developed by Beak for
laboratory scale studies of the swirl separator,
gave a range of removals from 56 to 75
percent for overflow rates of 0.2 cm/sec
(0.00656 ft/sec). Thus, the simulated material
used in the laboratory scale studies appears to
have very closely approximated the actual
solids in combined sewer overflows.
HYDRAULIC CHARACTERISTICS
Removal of "suspended" material from a
water, solids mixture may be accomplished in
many ways familiar to sanitary engineers. The
swirl separator, like the traditional primary
clarifier, accomplishes this using gravitational
force to help remove the fraction of influent
solids which are settleable. However, the
hydrodynamic forces, due to the movement
of the fluid in the vessel, are used to enhance
the removal in the case of the swirl separator,
whereas in the conventional clarifier, any
hydrodynamic forces are "designed out" as
the resulting turbulence tends to re-suspend
settled material. To more fully understand the
nature of the hydraulic regime governing the
hydrodynamic forces present and to
"characterize" the vessels into a sequence of
interconnected compartments, each governed
by well understood flow patterns, tracer
studies using fluorescent dye were conducted
on the swirl at various flowrates and on the
battery of parallel primary clarifiers. The
technique used is standard to chemical
engineering and involves the injection of a dye
into the feed stream of the vessel, and the
subsequent monitoring of the dye in the
effluent as a function of time. Starting with
an initial pulse of dye, the degree of mixing in
the vessel may be deduced by matching the
resultant output dye trace to the pattern that
159
-------
would result from a model comprised, of
segments or compartments, in which ithe
mixing is well understood and the output dye
trace well known. Once the vessel has been
characterized into a series of ideal tanks or
reactors, it becomes possible to compare ;the
behavior of the vessel to that expected using
classical settling theory.
Methods for Hydraulic Characterization
The flow into the vessel under
investigation is maintained as uniform as
possible to ensure that the tank hydraulics are
in steady state. !
A pulse of conservative dye is injected
into the inlet stream to the vessel in such a
way that the dye becomes as well mixed as
possible across the influent stream. ;
A time base is maintained throughout the
investigation with zero time being the instant
the dye is injected. To ensure that 'the
effluent dye trace is measuring the hydraulics
of the vessel and not the inlet/outlet channels,
effects due to the latter are eliminated by
injecting and monitoring the dye as close to
the vessel as practical. As the pulse of dye
moves through the vessel, the hydrodynamic
regime alters the profile of the dye pulse and
the monitoring station downstream measures
the shape of the dye trace or distribution as a
function of time. The alteration of the dye
trace from a pulse having essentially ,no
distribution into the final shape at the outiet
monitoring point may be interpreted
mathematically to yield some hypothetical
network of ideal tanks. In the present stujdy,
the mathematical analysis took the form of
variance analysis rather than the more
involved minimization of square error.
Rhodamine WT® was used as the conservative
tracer and was analyzed in the form of
discrete samples, each taken at a
predetermined time, using fluourometric
methods. To remove interferences due to the
sewage, a blank sample was run through the
fluorometer to establish base line calibration. •
Discussion of Hydraulic Results
Details of the statistical techniques used
to fit the models to the data are contained in
Exhibit 2. The parameter estimates are in
Table 26. Plots of the dye traces obtained and
the fitted models are shown in Figures 91
through 96. Examination of the experimental
residence times shows that in each case the
theoretical residence time of the swirl
separator, based on the calculated volume,
was greater than that exhibited by the tracer
and that the difference decreased with flow
based on three data points as shown in Figure
97. This would indicate that as flow increases,
the degree of short-circuiting increases or
conversely that as flow increases, the portion
of the total volume available for flow through
decreases. The battery of six parallel, primary
clarifiers showed 60 percent increase in the
measured detention time indicating a
significant amount of recirculation within the
tanks, and/or an imbalance in the splitting of
the flow between tanks.
The fitted parameters showed that in all
cases, the swirl separator behaved more as
plug flow than did the primary clarifier and
that as flow increased, the swirl separator
• shifted more toward the plug flow regime, but
even at high flow, remained less well mixed
than the primary tanks. The model fitting
indicated the general trend of mixing within
the vessels and the fittings procedure used was
not intended to provide a statistical criteria
through which the best fit could be obtained.
The variation in number of tanks in sewer (AT)
and axial dispersion coefficient CD) are indic-
ative of the same trends in mixing and are
equally useful in generalizing the hydraulics of
the vessels.
Conclusions on Hydraulic Characteristics
The results of the tracer study show that
the hydrodynamics of the swirl separator lie
more toward the plug flow regime of mixing
than do the primary clarifiers. The apparent
decrease in active volume and shift toward
plug flow which occurs as the flow increases
in the swirl concentrator would indicate that
the quiescent cone of fluid along the axis of
the swirl may increase with increasing flow.
Removal efficiency decreases as the
mixing regime shifts from plug flow to well
mixed. The results of the tracer study would
predict that the swirl, being closer to plug
flow in all cases tested, would accomplish the
same removal as the primary clarifiers but at a
significantly higher upflow velocity or surface
160
-------
TABLE 27
MODEL FITTED PARAMETER ESTIMATES
Axial Dispersion
Tanks in Series
Dimensionless
Settling Device
Swirl Primary
Swirl Primary
Swirl Primary
I/sec
22.2
16.5
10.4
Flow
mgd
0.51
0.38
0.24
Variance
T 2
Te
0.28
0.37
0.36
Dispersion
Number
D/uL1
0.17
0.24
0.23
Number of
Tanks in Series
N2
3.6
2.7
2.8
Actual Mean3
Residence Time
Tact (min)
6.4
13.8
21
Theoretical4
Residence Time
Tth (min)
13.4
18.0
28.5
Primary Clarifier 4.785 110
0.52
0.43
1.9
75.8
47.3
the lower the dimensionless dispersion number, the lesser the mixing in the vessel. Good settling practice minimizes the axial
dispersion.
the more tanks in series, the less short-circuiting and mixing in the vessel. Good settling practice minimizes short-circuiting and
mixing.
"^actless tnan ^tfiis indicative of a dead volume zone
4Tfft less than tact is indicative of recirculation within the vessel
EXHIBIT 2
DESCRIPTION OF HYDRAULIC MIXING MODELS
IP this study, the swirl separator and the primary clarifier hydraulic characteristic have been
fitted to two models. These are the axial dispersion model and the equal tank-in-series model.
Tables 28 and 29 describe these models.
The degree of mixing in the axial dispersion model is described by the axial dispersion
coefficient or the axial dispersion number (D/uL). Where D = the axial dispersion coefficient
(m2/sec); u = fluid velocity in the vessel; L = length of vessel. For best settling conditions in a
vessel, this number should be minimized.
Mixing in tanks-in-series model is described by the number of tanks in the series (N). For best
settling conditions in a vessel, N should be maximized.
161
-------
r
Exhibit 2 (Cont'd)
TABLE 28
THE AXIAL DISPERSION MODEL
THE
MODEL
I
L = length of vessel (m)
u = avg. fluid velocity in vessel (m sec"1 )
D = axial dispersion coefficient (m sec )
Pe = Peclet number = uL/D
RESIDENCE
TIME
DISTRIBUTION
E(0) = rE(t)
E(t) =JPe_ expF Pe (2- tjl^
+ exp —
[T«-<>
Xksin (2Xk>
\2 + Pe2 + Pe
4 4
exp
Pe
where A are the positive nontrivial roots of the transcendental equation:
•K i
tan
Pe
These may be obtained by defining j3n = X2n_i
and Tn = X2n where the ]3n and Tn are the positive roots of the
transcendental equations:
tan /J = Pe
4/3 ,
and
cot X = P!_
4X
VARIANCE
= 2(2.) - 2(2.)2 1_e
uL uL |_
_(uL,
162
-------
Exhibit 2 (ContU)
TABLE 29
THE EQUAL TANKS-tlM-SERIES MODEL
THE
MODEL
RESIDENCE
TIME
DISTRIBUTION
Q
00 V
ixr-1
~t
oo V
N
i
oo V
Q= flow/rate (mgd)
Number of tanks = N
Volume of tanks = V
Individual vessel residence time = T = VV'Q min
V = total active volume
= N. V (gal)
overall mean residence time =TT = N min T
E(fl) = rE(t)
E(t) =
(N-t)l
exp
[-Nt
^T
TRANSFER
FUNCTION
N
VARIANCE
163
-------
1,2-
1.0-
S
"us
0.8-
.0
1
£ 0.0-
o
g>
D) __
in 0,ij —
-
0.2-
0
1.2-
1.0-
UJ
I 0.8-
O
1 -
|0.6-
&
.t;
X
m 0.4-
0.2-
0
a
f
r
\
\ !
\
\
\
\ v Primary Clarifier QT = 481 I/sec (110 mgd)
\
\
X<
' iiY\ A Sw'rl Primary Separator — 22.3 I/sec (0.51 mgd)
/ / V1 fiv ; ° Swirl Primary Separator - 16.6 I/sec (0.38 mgd)
J / »V\ B Swirl Primary Separator - 10.5 I/sec (0.24 mgd)
f / \\ V
/ A ^
I \ V. \
i \
r \\
II Vx
In ^>-Os.
II ~^f***°^=?-—
ju a^^ Q ^
L ' i —
' \ 2 3
Dimensionless Time (9 )
FIGURE 92 RESIDENCE TIME DISTRIBUTION
: 164
-------
1.0-
« 0.8-]
_O
"•P
|
S 0.6 H
0.4-
0.2-
\\
0 Swirl Primary Separator - 22.3 I/sec (0.51 mgd)
Tanks in Series
Dispersed Plug Flow
\
\
\
.\
*
\
\
\
\°
Dimensionless Time' (6 )
FIGURE 93 FITTED RESPONSES
C
O
a
I
*j
'5<
ui
1.0-
0.8-
0.6-
0.4-
0.2-
/\
Swirl Primary Separator -
__ Tanks in Series
Dispersed Plug Flow
16.6 I/sec (0.38 mgd)
\\
\\
^
\
Dimensionless Time (d)
FIGURE 94 FITTED RESPONSES
165
-------
1.0-
ui
I
& 0.6-
S
Q
.*• 0.4
X
UJ
0.2-
/
Jl
II
1 I
l!
1 1
il
ii
i
>'
— •&
Y \ o Swirl Primary Separator ,-10.5 I/sec (0.24 mgd)
\ \ . Tanks in Series
\ v . Dispersed Plug Flow
\ \
\ V
\ \
\ \
\ \
\ \
\ X.
•^ ^^*^^^ 0
*** *"*•••. ^*"** ****••••, •
1 2
Dimensionless Time (6).
FIGURE 95 FITTED RESPONSES
1.2-
1.0-
/— *
S>
UJ
I
n 0.8-
1
5 0.6-
%
<
+*
'5<
UI
0.4-
0.2-
0
^
J
/
^/^^v
/ \ ? ; PRIMARY CLARIFIERQ-,. =481 I/sec (110 mgd)
/ \ -—- Tanks in Series
/ \ ____ ' Dispersed Plug Flow
I ^7 \.
/ \
I f "^^v \
| / V \ \
1 ' \ \
1 I ^ \ \
1 •. \\
1 \ \
\ \
* \ \
/ \ \
\ \
I : \ * \
S ^v.
S ^^"Sw.
\^ ^«v>^
**>"-^ ">*^-^.
"*""—--«. "^~~-— "«,'
II '
1 2 3
Dimensionless Time (0)
FIGURE 96 FITTED RESPONSES
166
-------
60-
y =-0.39+ 90.76 x
Q.1
0.2
0.3
FLOW I/sec (mgd)
o.5
FIGURE 97 PERCENT DEAD VOLUME VERSUS FLOW, SWIRL SEPARATOR i
overflow rate. Using the Froude Number in a
dynamic similarity projection, there is a limit
to the size of prototype swirl unit which will
perform as well as the scale model tests. To
compensate for the scale effects as the size is
increased, it may be necessary to change!the
dimensions of the chamber, such as ; the
diameter of the inner skirt baffle and ;the
baffle clearance, and to control the weir
overflow rate.
REFERENCES
1. Dalrymple, R.J., etal. Physical and Settling
Characteristics of Particulates in Storm
and Sanitary Waste Waters, EPA Report
No. EPA 670/2-75-011
242001, April, 1975
NTIS No. PB
167
-------
APPENDIX D
'PILOT TEST RESULTS
TABLE 30
DRY-WEATHER TEST SCHEDULE
Test No. Test No.
Swirl Cone. Primary Tank
Test Inlet Outlet Inlet Outlet
Settleable Solids
Total Suspended Solids
Volatile Suspended Solids
Fixed Suspended Solids
Temperature
PH
BODS :
COD
1
3
5
7
—
—
g
11
2
4
6
8
.-
—
10
12
16
18
20
22
24
25
26
28
17
19
21
23
—
_
27
29
Sludge Solids Concentration
Sludge Volume
Sludge
13
14
Sludge
30
31
Total Daily Flow
Sewage
15
Sewage
32
Tests 1, 2, 16, 17
Tests 3 thru 8
18 thru 23
Test 24
Test 25
Tests 9, 10, 26, 27
Tests 11, 12, 28, 29
Tests 13, 14, 30, 31
Tests 15, 32
Every four hours on grab samples.
Daily on composite grab samples taken every
two hours proportional to flow.
liun tests 16 thru 23 for one week and if little
or no variation then discontinue for dry weather
flow.
Every eight hours.
Every eight hours.
Daily on composite. Do on unfiltered sample
and filtrate.
Daily on composite.
Daily on composite.
Daily.
TABLE 31
WET-WEATHER TEST SCHEDULE
Test
Settleable Solids
Total Suspended Solids
Settling Column
Heavy Metal
Test No.
Swirl Cone.
Inlet Outlet
50 51
54 55
—
— —
Test No.
Primary Tank
Inlet Outlet
52 53
56 57
60
61
Heavy Metal
Tests 50, 51, 52, 53
Tests 54, 55, 56, 57
Test 60
Tests 61, 62
Sludge
62
Hourly on grab samples.
One test for each storm event (or daily if storm
exceeds one day) on composite grab samples
taken hourly proportional to flow.
Three tests each during two storm events, at
beginning, peak, and end of event.
One test each.
168
-------
TABLE 32
WET-WEATHER TEST DATA, SWIRL SEPARATOR
MAY 4-7, 1975
Date
Flow I/sec
mgd
INFLUENT
30 Min. Settleable
Solids (ml)
Total Suspended
Solids (m|/l)
EFFLUENT
30 Min. Settleable
Solids (ml)
Total Suspended
Solids (ml/1)
May 4
18.13
0.43
7.0
260
7.0
284
w
Z3
*— '
|2
O
c
JO
H
O
o
CO
o
o
1
5^
o
0
I
en
May 5
18.8
0.43
7.0
360
8.0
364
V)
"n
O
DO
O
C
33
H
O
"2.
O
o
o
1
o
o
o
I
May 6
18.8
0.43
6.0
304
5.0
320
rH
O
DO
D
C
DO
H
0
O
VJ
o
o
I
Its!
8
T.
e/i
May 7
18.8
0.43
30.0
564
28.0
464
CO
H
O
DO
D
C
33
H
O
o
8
i
s
o
o
I
»
169
-------
TABLE 33
WET-WEATHER TEST DATA SWIRL SEPARATOR
MAY15-JUNE 12,1975
Date
Flow I/sec
mgd
INFLUENT
30 Min. Settleable
Solids ml/I
Total Suspended
Solids mg/l
EFFLUENT
30 Min. Settleable
Solids ml/1
Total Suspended
Solids mg/l
T
May 15
18.8
0.43
15.0
524 ;
8.0
344
H
3
O
C
3J
H
O
Z
10
ro
8 ;
,
8
i
co
May 31
18.8
0.43
11.0
392
7.0
376
en
O
:D
O
c
DO
H
O
o
10
8
I'
3
o
o
X
CO
June 4 & 5
18.8
0.43
18.0
540
14.5
596
3
30
0
o
c
DO
H
O
e_
CD
is
8
I
•
I
c
3
CD
01
O
O
I
June 12
18.8
0.43
7.5
280
4.0
204
V)
O
33
0
C
3J
H
0
o
8
I
1
8
8
in
CO
170
-------
TABLE 34
WET-WEATHER TEST DATA PRIMARY TANKS
MAY 15-JUNE 12, 1975
Date
Flow I/sec
Flow mgd
INFLUENT
30 Min. Settleable
Solids ml/1
Total Suspended
Solids mg/l
EFFLUENT
30 Min. Settleable
Solids ml/1
Total Suspended
Solids mg/l
Date
6 tanks — Flow I/sec
Flow mgd
INFLUENT
30 Min. Settleable
Solids ml/1
Total Suspended
Solids mg/l
May 15
6,658
152.4
10.0
188
9.0
324
May 15
5,828
133.4
12.0
744
May 31
4,456
102.0
13.0
224
1.0
140
May 31
4,448
101.8
15.0
416
June 4 & 5
7,445
170.4
7.0
336
7.0
272
June 4 & 5
5,243
120.0
14.0
432
June 12
4,823
110.4
24.0
632
.5
128
June 12
6,575
150.5
7.5
272
EFFLUENT
30 Min. Settleable
Solids ml/I
Total Suspended
Solids mg/l
1.0
220
14.0
544
11.0
388
1.0
148
Note: For storm duration, see Table 31
171
-------
TABLE 35
HEAVY METAL TEST DATA
(mg/l)
A. INFLUENT - PRIMARY TANKS
Date
Flow I/sec
Flow mgd
Heavy Metals
Cd
Cr
Cu
Fe
Pb
Hg
Ni
i-t
Sn
Zn
SLUDGE
Heavy Metals
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Sn
Zn
mg/l (ppm)
(mg/kg)
May 4
6,658
152.4
0.04
0.64
0.49
4.67
0.55 :
0.0018
1.12
2.75
68.2;
1,827
626
11,420 ,
985
3.48
193
55.9
6,240
May 5
4,456
102.0
0.026
1.32
0.25
3.27
0.42
0.0013
0.10
1 nn» +1-1 1*1 PI Q
— Less tnan u.o—
2.35
73.5
1,861
660
11,450
998
3.82
190
91.7
6,206
May 6
7,445
170.4
0.040
0.71
0.31
4.96
0.64
0.0009
0.26
2.90
74.7
1,626
606
10,575
912
4.31
213
70.6
5,862
May?
4,823
110.4
0.047
1.17
0.50
5.92
0.96
0.31
0.31
3.13
82.1
1,675
651
11,640
1,005
4.51
224
39.0
6,313
B. INFLUENT -DRY WEATHER
Date
Flow I/sec
mgd
Cd
Cr
Cu
Fe
Pb
Hg
Ni
Sn
Zn
May!
4,234
96.9
0.062
1.43
0.48
6.95
1.65
0.0031
0.28
2.98
1975
May 2 May 3
4,225 3,705
96.7 84.8
0.089 0.072
1.60 1.24
0.43 0.44
6.60 6.70
1.33 0.98
0.0049 0.0033
0.48 0.42
1 nrr- thin C\ P
3.20 2.58
May 8
4,631
106.0
0.049
1.03
0.38
4.96
0.92
0.0029
0.46
2.25
1976
June 16, 17, 18
, and 21
Composite Sample
0.04
0.95
0.47
9.78
0.32
O.Q02
0.34
<0.05
3.22
4
Note: For storm duration see Table 30.
172
-------
TABLE 36
DRY-WEATHER TEST DATA -SWIRL PRIMARY SEPARATOR
MAY 1-11, 1975
May 1
{Thurs. )
13
0.3
17.25
576
424
152
407
122
947
14.75
388
292
96
386
122
815
May 2
(Fri.)
1 3
0.3
20.0
412
344
68
260
103
776
18.0
356
288
68
269
105
677
May 3
(Sat.)
1 3
0.3
21.0
432
360
72
253
60
733
20.0
380
340
40
243
64
683
May 8
{Thurs.)
1 3
0.3
11.0
456
356
100
288
114
636
11.25
360
264
96
308
117
714
May 9
(Fri.)
1 3
0.3
11.75
460
408
52
316
76
821
9
360
320
40
264
88
706
May 10
(Sat.)
1 3
0.3
15.0
512
396
116
270
85
743
11.0
460
348
112
240
83
658
May 11
(Sun.)
13
0.3
18.0
472
344
128
223
53
654
12.0
512
392
110
262
57
670
117
4.3
72.4
0.2
358
4.9
72.5
0.7
224
5.2
73.8
0.4
311
4.4
74.7
0.6
386
4.5
72.7
0.7
225
5.0
72.8
0.5
510
4.1
72.7
1.0
m
I /sec 5"
mgd
SOMin.Settleable
Solids ml/I
Total Suspended
Solids mg/l
Volatile Suspended
Solids mg/l
Fixed Suspended
Solids mg/l
Unfiltered - BOD
Filtrate — mg/l
Unfiltered - COD
Filtrate — mg/l
SOMin.Settleable
Solids ml/I
Total Suspended
Solids mg/l
Volatile Suspended
Solids mg/l (ppm)
Fixed Suspended
Solids mg/l
Unfiltered - BOD
Filtrate — mg/l
Unfiltered - COD
Filtrate — mg/l
Volume (gal)
Solids Concentration (%)
Total Volatile Solids (%)
TVS
Ts
Volume to be
Drawn off (%)
INFLUENT
EFFLUENT
2
7)
m
173
-------
DRY-WEATHER TEST
TABLE 37
DATA - SWIRL PRIMARY SEPARATOR
MAY 12-18,1975
May 12
(Mon.)
19.7
0.45
13.25
444
336
108
311
102
768
12.75
408
308
100
313
104
711
May 13
(Tues.)
19.7
0.45
17.0
428
332
96
331
136
793
12.0
576
472
104
347
119
895
May 14
(Wed.)
19.7
0.45
18
476
368
108
339
132
839
14
440
348
92
358
125
898
May 15
(Thurs.)
19.7
0.45
15
604
452
152
398
124
997
14.6
548
416
132
380
111
911
May 16
(Fri.)
19.7
0.45
17
476
404
72
359
119
896
13.25
436
364
72
347
121
767
May 17
(Sat.)
19.7
0.45
24
428
344
84
300
110
lie,
20.25
536
436
100
274
89
759
May 18
(Sun.)
19.7
0.45
26
568
480
88
325
88
762
23.25
472
392
80
307
71
697
768
4.3
72.8
1.0
803
5.2
73.1
1.1
731
5.1
73.7
1.0
379
4.7
74.3
0.5
543
4.1
74.3
0.7
514
3.9
75.5
0.7
570
4.5
75.2
0.8
5
m
I/sec ET
mgd S
SOMin.Settleable
Solids ml/1
Total Suspended
Solids mg/l
Volatile Suspended
Solids mg/l
Fixed Suspended
Solids mg/l
Unfiltered - BOD
Filtrate - mg/f
Unfiltered - COD
Filtrate — mg/l
30 Min. Settleable
Solids ml/1
Total Suspended
Solids mg/l
Volatile Suspended
Solids mg/l
Fixed Suspended
Solids mg/l
Unfiltered - BOD
Filtrate - mg/l
Unfiltered - COD
Filtrate — mg/l
Volume (gal)
Solids Concentration (%
Total Volatile Solids (%
*
i
n
»
m
•n
•n
n
2»
H
SLUDGE
174
-------
TABLE 38
DRY-WEATHER TEST DATA SWIRL PRIMARY SEPARATOR
MAY 30 -JUNE 10, 1975
May 30
(Fri.)
1 3
0.3
22.0
660
512
148
410
180
1,110
15
456
368
88
370
170
866
643
4.9
73.3
1.3
June 2
(Mon.)
13
0.3
17.0
620
456
164
335
136
965
12.5
464
340
124
290
133
788
824
3.8
72.6
1.6
June 4
(Wed.)
13
0.3
i
20.0
668
496
172
354
152
1,008
16
332
252
80
310
134
788
44
4.3
72.2
—
June 7
(Sat.)
13
0.3
2
632
476
156
310
121
1,070
2
352
276
76
350
83
818
489
4.3
71.1
1.0
June 8
(Sun.)
13
0.3
2
572
452
120
359
145
928
2
456
372
84
357
144
795
634
4.3
71.3
1.3
June 9
(Mon.)
13
0.3
17.25
712
576
136
325
125
1,262
11.75
444
372
72
320
115
894
609
4.3
71.9
1.2
June 10
(Tues.)
13
0.3
16.75
472
360
112
310
137
892
10.75
428
320
108
289
122
718
644
5.2
71.7
1.3
m
I/sec
mgd
30 Min. Settleable
Solids ml/l
Total Suspended
Solids mg/l
Volatile Suspended
Solids (mg/l)
Fixed Suspended
Solids (mg/l)
Unfiltered - BOD
Filtrate - (mg/l)
Unfiltered - COD
30 Min. Settleable
Solids ml/l
Total Suspended
Solids mg/l
Volatile Suspended
Solids (mg/l)
Fixed Suspended
Solids (mg/l)
Unfiltered - BOD
Filtrate — mg/l
Unfiltered - COD
Filtrate — mg/l
Volume (gal)
Solids Concentration(%]
Total Volatile Solids
Concentration (%)
•n
1
INFLUENT
EFFLUENT
SLUDGE |
' Plant shutdown on June 3 for 3 hours so swirl was shut down and cleaned put back into service 0001 hrs. — June 4.
Over 40 ml
175
-------
TABLE 39
DRY-WEATHER TEST DATA SWIRL PRIMARY SEPARATOR
JUNE 16-21, 1976
6-19
6-17
*18
6-21
1
0.2
02
02
WhoUSvnote
||
436
483
470
Volltil>Si»p«Hlld
Solidi Ippm)
351
386
387
FixidSulptodid
Solidi Ippm)
35
97
103
30-minSlttlllMl
Solidt
20
19
16
Efflmnt
Stp.nn.unt I
!}
{i
133
164
162
"i
}j
1
111
155
140
Fixed SuiptiKted
'
I
8
2
22
I — X —
Whole Sample
Total Suipended
Solidt (ppm)
166
249
173
1—ffi.J
Volatile Sutpended
Solidi (ppm)
146
129
143
1 14B
Fixed Suipended
Solidi (ppm)
20
120
30
r30-min Settleeble
Solidt
1.0
Leu
than
0.1
0.1
Supernatant
Tout Suipended
Solid* {ppml
115
132
148
Volatile. Suipended
Solidi (ppml
105
132
r Fixed Suipended
Solidi (ppm)
16
Sludge
11
13
646
%
«P!!°S moj.
2.40
2.36
1 Total Volatile
| . Solidi %
72.5
72.3
l -_J r_j L_i. 1 — _
VVhok Svnpto. Sample b«(or« 30 min, unleable lolids
Supcrrutant: Liquid portion ofumpl* aher 30 min. lenteabte solidi
TABLE 40
DRY-WEATHER TEST DATA SWIRL PRIMARY SEPARATOR
JUNE 22-25, 1976
DM
me
6-72
MJ
6-24
6-25
Flow
e
015
0.15
0,15
0.15
InfliMnt '
Vftol«S«npli
1
ff
586
528
493
457
it'll SinpiiKM
Solidl Input)
£
477
426
401,
383
xidSuiplnd«l
Solid. Ippm)
i^
109
102
92
74
1
H
24
20
19
12
SupinntMt
ll
£«
|2
187
169
138
153
|
f!
> j
162
148
,2, |
1
150
xid Suipindid
Solidl (ppm)
25
21
17
8
Effluent
Who!* Siinplii [ S
nil Suipindld
Solidt (ppm)
165
146
136
152
' 1 .1 "
atile Suipcnded
Solidt (ppm)
136
130
120
133
le
!&
S-8
Is
u.
29
16
16
19
•min Settleible
Solidi
Leis
than
0.1
0.1
Lets
than
0.1
Lets
than
0.1
atal Suipended
Solidi (ppm)
upematar
<—.-
•tile Suipendod
Solidt (ppml
157
128
113
144
136
117
103
142
j
t
•
xad Suipended
Solidi (ppm)
„ "!"!',
ii
13
2, ' 527
!"
.,
2.16
11
476
• -,d
2,33
L_ .
10 1 493 | 1.43
L..."L :':••. '.
p -j . |,....y.
2 ] 306 246
.' L ' ..L .:
?t>!d Voiatiis
Solidi X
72,7
73.0
73.7
71,4,
T!1":; I
Wtiote Sample Sample before 30 min. unleabte Jolidi
Sop«m«.»lt. Liquid portion of urnple after 3O min. mtleibla i<
176
-------
May 1
(Thurs.)
4,234
96.9
27.5
692
520
172
348
118
1,021
65.0
18.3
7.4
10.5
232
176
56
296
113
452
1,332
352
4.2
2.2
May 2
(Fri.)
4.225
96.7
26.0
664
536
128
286
117
1,010
65.0
18.3
7.7
14.25
294
244
50
238
100
507
1,347
356
4.1
2.2
DRY-WEATHER
May 3 May 8
(Sat.) (Thurs.)
3,705
84.8
25.0
588
488
100
275
82
922
64.0
17.7
-
Less than
0.5
86
80
6
135
57
289
1,378
364
4.3
2.8
4,631
106.0
12.75
504
392
112
352
137
796
64.0
17.7
7.5
TABLE 41
TEST DATA - PRIMARY T/
MAY 1-11, 1975
May 9 May 10 May 11
(Fri.) (Sat.) (Sun.)
4,238
97.0
14.25
540
484
56
322
88
883
65.0
18.3
—
Less than
0.5 8.0
130
96
34
218
135
428
1,416
374
4.2
2.1
224
208
16
230
137
489
1,378 1
364
4.2
2.2
3,853
88.2
18.5
680
452
148
295
79
875
64.0
17.7
—
5.0
150
118
32
148
58
325
,355
358
4.4
2.4
3,478
79.6
19.0
496
380
116
250
66
693
63.0
17.2
7.6
2.0
160
134
26
147
49
268
1,321
349
4.5
2.9
UMKS
D
m
I /sec
mgd
30 Min. Settleable
Solids ml/l
Total Suspended
Solids mg/l
Volatile Suspended
Solids mg/l
Fixed Suspended
Solids mg/l
Unfiltered - BOD
Filtrate — mg/l
Unfiltered - COD
Filtrate — mg/l
Temperature (°F)
Temperature (°C)
pH
30 Min. Settleable
Solids ml/l (ppm)
Total Suspended
Solids mg/l
Volatile Suspended
Solids mg/l
Fixed Suspended
Solids mg/l
Unfiltered - BOD
Filtrate — mg/l
Unfiltered - COD
Volume (m3)
Volume (galx 103)
Solids Concentration {%)
Tl
o
Z
FLUENT
EFFLUENT
SLUDGE
177
-------
TABLE 42
DRY-WEATHER TEST DATA - PRIMARY TANKS
MAY 12-18, 1975
May 12
(Mon)
4,194
96.0
18.25
608
472
136
327
105
662
65
18.3
7.5
17.0
292
224
68
236
103
523
1.389
367
4.3
2.3
May 13
(Tues)
4,242
97.1
28
440
340
100
351
140
855
66
18.8
7.6
7.75
306
242
64
255
133
712
1.325
350
4.3
2.2
May 14
(Wed)
4,277
97.9
27
520
420
100
356
138
874
66
18.8
7.5
11.0
226
- 180
40
260
128
574
1.363
360
4.4
2.2
May 15
(Thur)
4,321
98.9
18
700
528
172
416
129
975
66
18.8
7.5
14.5
242
184
58
277
140
650
1.336
353
4.2
2.1
May 16
(Fri)
4,063
93.0
16
552
440
1112
326
113
;838
64
17.7
7.8
2
174
134
40
200
128
361
1.347
356
4.4
2.3
May 17
(Sat)
3,303
75.6
24
668
560
108
375
123
727
65
18.3
_
1
108
84
24
163
98
266
1.347
356
4.6
2.8
May 18
(Sun)
3,032
69.4
26
432
340
92
296
66
680
67
19.4
_
1
94
84
10
169
81
233
1.347
356
4.7
3.1
D
5
m
I/sec
mqd
30 Min. Settleable
Solids ml/I .
Total Suspended
Solids ma/I .
Volatile Suspended
Solids ma/I .
Fixed Suspended
Solids mg/l
Unfiltered — BOD
Filtrate — ma/I .
Unfiltered - COD
Filtrate mq/l
Temperature (°F)
Temperature (°C)
pH
30 Min. Settleable
Solids ml/l
Total Suspended
Solids mq/l ,.
Volatile Suspended
Solids ma/I
Fixed Suspended
Solids mq/l .
Unfiltered - BOD
Filtrate — rnq/l —
Unfiltered - COD
Volume (m3)
Volume (qal/1, 000)
Solids Concentration
•n
1
INFLUENT 1
m
•n
-n
r
oo
O
G)
m
178
-------
TABLE 43
DRY-WEATHER TEST DATA - PRIMARY TANKS
MAY 30-JUNE 10, 1975
Date
6 tanks — Flow I/sec
mgd
INFLUENT
30 Min. Settleable
Solids ml/I
Total Suspended
Solids mg/l
Volatile Suspended
Solids (mg/l)
Fixed Suspended
Solids (mg/l)
Unfiltered - BOD
Filtrate — mg/l
Unfiltered - COD
Temperature (°C)
Temperature (°F)
EFFLUENT
30 Min. Settleable
Solids ml/I
Total Suspended
Solids rng/l
Volatile Suspended
Solids (mg/l)
Fixed Suspended
Solids (mg/l)
Unfiltered - BOD
Filtrate — mg/l
Unfiltered - COD
SLUDGE
Volume m3
Volume (gal x 103)
Solids Concentration (%)
May 30
(Fri.)
4,107
94.0
24.0
780
624
156
450
170
1,214
20.5
69
11.0
294
238
56
375
210
605
4.6
June 2
(Mon.)
4,098
93.8
21.75
648
476
172
320
128
995
19.4
67
9.25
332
258
74
267
122
621
1,525
403
4.4
2.6
June 4
(Wed.)
4,395
100.6
20.0
592
456
136
328
142
967
21.1
70
11.0
202
168
34
212
121
497
1,514
400
4.8
2.4
June 7
(Sat.)
3,469
79.4
i
610
464
146
360
122
991
20
68
i
72
60
12
146
52
288
1,718
454
4.4
3.4
June 8
(Sun.)
2,984
68.3
i
772
596
176
366
132
1,210'
19.4
67
i
90
88
2
150
131
253
1,798
454
4.3
4.0
June 9
(Mon.)
4,063
93.0
17.5
884
732
152
317
119
1,090
20.5
69
7.0
346
298
48
216
94
643
1,798
475
4.8
3.1
June 10
(Tues.)
4,596
105.2
16.0
496
372
124
342
132
853
20.5
69
6.9
180
138
42
213
115
537
1,786
472
4.8
2.7
Over 40 ml/1
179
-------
TABLE 44
DRY-WEATHER TEST DATA PRIMARY SETTLING TANK
JUNE 16-21,1976
Dtu
me
6.16
617
618
621
Flow
1
829
(99.431
194,691
794
195,281
73.7
188441
lnHu.it
Toul SuipiiKWd
Soli* Ippm)
WhoUSanipto
1
(i
I3
1
e M
SuptRMUnt
ToUISweilxM
Solldi Ippml
VoliUh SwpmM
Solhklwn)
jl
I?
|
('
Efflu.it
WhoteSMipto
jl
147
122
589
211
VolrtkSuvinM
Solkli(K>nil
131
100
465
196
FJMd SuiptfKkd
Soil* torn)
16
22
124
15
30-nHii Smlubk
tolkk
0.1
0.1
28
2.5
Cupmuunt
Toul&wnddl
SoUdilppm)
107
107
121
126
VdteittmtmM
SolkHlppml
80
89
100
119
ll
fi
27
18
21
7
tlwt't»'
ii
1-
3.94
2.04
3.93
3.63
ToulVolitito
to»d>%
68.7
56.6
67.7
71.4
Who.'. Sjmcte Simple before 30 min, wnhabta ulidt
Supnnttint; Liquid portion of wrnpit ifter 30 min. Mltf»U« (otidt
TABLE 45
DRY-WEATHER TEST DATA PRIMARY SETTLING TANK
JUNE 22-25,1976
Supcmitxnt
Whoto SwnpK. SMTP!* btfoo 30 min, unttBbfe tolidt
Liquidportwnof Mmpta afMr30min,nnteJt
180
-------
TABLE 46
PROPOSED TESTS SECOND SERIES
Sludge Sampling Sampling
Withdrawal (influent to swirl (swirl influent)
&
primary effluent)
(hrs) (hrs) (hrs)
1-100 1,130 1,115-1,145
1,200 1,230 1,215-1,245
1,300 1,330 1,315-1,345
1-400 1,430 1,415-1,445
Required Analysis: .
Influent Effluent Effluent
to Swirl of primary clarifier .of Swirl
(composite of 4) (composite of 4) (composite of 8)
BOD5 (total) BODS (total) BODS (total)
ss, vss ss, yss ss, vss
COD
Once during the 4-hour testing period, a settleable
solids test should be done from each location.
TABLE 47
,
DRY-WEATHER TEST DATA - SWIRL PRIMARY SEPARATOR
JUNE 23 -27, 1975
Date June 23 June 24 June 25
(Mon) (Tues) (Wed)
Flow I/sec 13 13 13
Flow mad 0.3 0.3 0.3
INFLUENT
30 Min Settleable Solids ml/l (ppm) 11 15 -\-\
Total Suspended Solids mg/l (ppm) 380 525 488
Volatile Suspended Solids mg/l (ppm) 284 380 384
Fixed Suspended Solids mg/l (ppm) ' 96 144 104
Unfiltered - BOD 284 301 300
Unfiltered - COD 710 735 757
EFFLUENT
30 Min. Settleable Solids ml/l (ppm) 5 11 6
Total Suspended Solids mg/l (ppm) 192 332 252
Volatile Suspended Solids mg/l (ppm) 144 232 212
Fixed Suspended Solids mg/l (ppm) 48 100 40
Unfiltered - BOD 248 278 247
Unfiltered - COD
SLUDGE " " : ~
Volume m3 ., -,» „ _..
.. . ... 1-74 2.61 3.58
v°l"me (gal) 459 690 947
Solids Concentration 26 38 30
Total Volatile Solids Concentration 66.8 692 7m
'Over 40 ml/1 0.9 1.4 1.9
June 26
(Thur)
1 3
0.3
i
904
692
212
382
1,320
i
556
432
124
361
1.78
469
2.9
75.4
O9
June 27
(Fri)
1 3
0.3
22
368
296
72
339
830
8
332
260
72
289
2.01
530
2.9
72.9
1.0
181
-------
TABLE 48
DRY-WEATHER TEST DATA PRIMARY TANKS
JUNE 23 - 27, 1975
Date
INFLUENT
(6 Tanks Operating) I/sec
Flow — mgd
30 Min Settleable Solids ml/I
Total Suspended Solids mg/l
Volatile Suspended Solids mg/l
Fixed Suspended Solids mg/l
Unfiltered - BOD
Unfiltered - COD
EFFLUENT
30 Min Settleable Solids ml/I
Total Suspended Solids mg/l
Volatile Suspended Solids mg/l
Fixed Suspended Solids mg/l
Unfiltered - BOD
SLUDGE
Solids Concentration (%)
Total Volatile Solids Concentration (%)
June 23
(Won)
4,264
97.6
11
380
284"
96
284
710
3
220
164
56
251
3.9
67.7
June 24
(Tues)
4,238
97.0
15
525
380
144
301
785
3
224
168
56
227
4.1
67.9
June 25
(Wed)
4,142
94.8
11
488
384
104
300
757
8
280
244
36
277
4.5
69.0
June 26
(Thur)
4,142
94.8
i
904
692
212
382
1,320
11
336
268
68
343
4.2
68.6
June 27
(Fri)
4,382
100.3
22
368
296
72
339
830
5
316
256
60
295
4.5
70.0
Over 40 ml/1
182
-------
TABLE 49
DRY-WEATHER TEST DATA SWIRL PRIMARY SEPARATOR
JULY 2-8, 1975
Date
Flow I/sec
Flow mqd
INFLUENT
30 Min Settleable Solids ml/I (ppm)
Total Suspended Solids mg/l (ppm)
Volatile Suspended Solids mg/l (ppm)
Fixed Suspended Solids mg/l (ppm)
Unfiltered - BOD (mg/l)
Unfiltered - COD (mg/l)
EFFLUENT
30 Min Settleable Solids ml/I (ppm)
Total Suspended Solids mg/l (ppm)
Volatile Suspended Solids mg/l (ppm)
Fixed Suspended Solids mg/l (ppm)
Unfiltered - BOD (mg/l)
SLUDGE
Volume m3
Volume
Solids Concentration (%)
Volatile Solids Concentration (%)
July 2
19.7
0.45
10.0
464
388
76
332
1,250
8.0
344
272
72
326
1.03
272
2.8
76.0
JulyS
19.7
0.45
10.0
436
340
96
296
786
5.0
424
328
96
295
0.61
161
2.6
71.9
July 4
19.7
0.45
10.0
512
400
112
361
806
9.0
428
316
112
309
0.56
147
2.7
72.3
July 7
19.7
0.45
12.0
500
380
120
307
765
10
364
256
108
307
0.64
170
3.5
70.9
JulyS
19.7
0.45
8.0
440
340
100
347
783
7.0
368
268
100
343
0.75
199
2.8
76.5
183
-------
TABLE 50
DRY-WEATHER TEST DATA PRIMARY TANKS
JULY 2-8, 1975
Date
July 2
July 3
July 4
July 7
JulyS
INFLUENT
I/sec (6 tanks in service)
Flow mgd
30 Min Settleable Solids ml/1
Total Suspended Solids mg/l
Volatile Suspended Solids mg/l
Fixed Suspended Solids mg/l
Unfiltered -BOD (mg/l)
Unfiltered - COD (mg/l)
EFFLUENT
30 Min Settleable Solids ml/1 (ppm)
Total Suspended Solids mg/l (ppm)
Volatile Suspended Solids mg/l (ppm)
Fixed Suspended Solids mg/l (ppm)
Unfiltered - BOD (mg/l)
SLUDGE
Solids Concentration (%)
Volatile Solids Concentration (%)
3,698
97.7
10.0
464
388
76
332
1,250
21.0
384
300
84
321
3.8
68.4
3,819
100.9
10.0
436
340
96
296
786
8.0
436
344
92
290
3.7
70.0
3,687
97.4
10.0
512
400
112
361
806
15.0
432
328
104
314
3.9
68.4
3,516
92.9
12.0
500
380
120
307
765
40.0
672
484
188
332
3.9
69.2
3,649
96.4
8.0
440
340
100
347
783
5.0
220
136
84
302
4.0
69.4
184
-------
TABLE 51
PROPOSED TESTS FINAL SERIES
Sludge Sewage
Withdrawal «,„,„,:„„
(hrs)
11 00 (No sample)
1130
1200
1230
1300
1330
1400
1430
1500
\ Time when flow
2 Time when flow
Influent
(hrs)
1100
1130
1200
1230
1300
1330
1400
1430
is 1 .3 I/sec
Swirl
Effluent
1 '(hrs) 2
1130 1115
1200 1145
1230 1215
1300 1245
1330 1315
1400 1345
1430 1415
1500 1445
(0.3 mgd)
Primary
Effluent
(hrs)
1200
1230
1300
1330
1400
1430
1500
1530
is 19.7 I/sec (0.45 mgd)
TEST
TEST No.
Sewage
Swirl Primary
Total Suspended Solids
Volatile Suspended Solids
Fixed Suspended:Solids
Settleable Solids by Wt.
Settleable Solids by VoL
1
2
3
4
13
5
6
7
8
14
i^iiiueni
g
10
11
12
15
1 thru 15
16 thru 20
Sludge
Volume
Solids Concentration
Volatile Solids %
4 tests daily on composite of two grab
samples at 0.5 hour intervals
1 test daily on composite taken at 0.5
hour intervals beginning 1/2 hour after
start of test
Swirl Primarv
16
17
18
19
20
185
-------
L
c
•&
~
to
SI
S
IS) O)
CD CD
*"" CO
CM "
£ 5
1 S
ui a:
H«
_ a.
a.
Ul
if) <£
£&
UJ
H
CC —
Ul "§j
H E
< w.
UJ O
S» o
5" «n
cc —
D ro
DC
1
8
u.
c*
3
* 1
ill
t Ul
Ul
c
111
w a. 11
ui 2
H
Ul
_i
u.
NOIJ.VU1N30NOO
sanos 31U.V1OA
NOIJ.VH1N33NOO SOIIOS
(ivjLOD awmoA
NOI1VU1N30NOO
NOI1VH1N30NOO SOIIOS
m.M^n.«y
sanos aaaNadshs SIIIVIOA
sanos asaNadsns nvioj.
sanos
310NIW OE
sanos a3ON3dsns asxid
sanos a30N3dsns annviOA
sanos a3OM3dsns nvioi
31SV31J-L3S
31HNIW 06
sanos a3GN3dsns asxu
sanos aaaN3. 9.
CO CO
» s
00 CM •
m »—
CO CO
1 1
CM CM
CO O
in in
CM a>
8888
«- CM CM «-
in
in
CO
iS
p
in
CM
S $
s i
CO CO
CM <"•
in o
ai ad
8 52
5 1
(O (O
CM CM
8 9
tf.t in
oi CM'
S *
N CM
S CM
CN S
in o
r^ cd
88888
^ IN CM CO CO
186
-------
LO
CO
cc
LU
CO
LU
s £
to eo
LU <
CO
.1 Q
' co
LU
h-
cc
LU
LU
CC
Q
•o
E
CO
s
u
<
CO
LO
co"
CC
Q.
•n
O)
CO
o
•u
CO
CC
CO
I
o
u.
UJ
s
LJ
V
oc
<
5:
a.
_j
oc
S
ec x
< z
1-
a.
_j
DC
•3.
!*S
1— nj
12
d "•
j_UJ
in
w
ei
< ^
oc 3
s^
"• UL
S"
»-
UJ
_l
u.
z
NOI1VH1N30NOO
sanos ai uyiOA
NOIJ.VU1N30NOO SQIIOS
nvj.01) 3i/\imoA
NO1J.VUJ.N30NOO
sanos3iu.\noA
NOI1VU1N30NOO SQIIOS
sanos Q3QN3dsns aaxu
sanos a3aiM3dsn.s STIJLVIOA
sanos aaaN3dsns nvioi
sanos
anavannas
sinNiwoe
sanos a3QN3dsns daxid
sanos aaaN3dsns 3TiJ.\noA
sanos aaaN3dsns nva.oi
sanos
an8van.LJ.3S
3inNin os
sanos oaaNadsns aaxu
sanos a30N3dsns an nxnoA
sanos Q3aW3dsns nvxo±
sanos
anav3nnas
ainwiwoE
f - <
fi 0 |
%
%
S|e6
%
%
I/Bui
I/Bun
I/Bui
1/6 iu
I/IUJ
I/Bui
I/Bui
I/Bui
I/Bui
l/l«"
I/Bui
I/Bui
I/Bui
I/Bui
l/|l"
§
S
e«j
1
I
o
O)
8
*~
8
S
CM
a
CM
O
co
g
CM
8
i
(O
CM
s
in
CM'
s
8"
CM
r*
CO
g
in
O)
8
CM
S
(O
1^
r-
1
oo
v
0
CM'
8
S
§
i
%
in
CM
(O
CO
1
(0
9
to
.<*>
in
o>
§
o
CM
to
r~
S
s
p
8
CM
^
I
S
§
o.
CM
00
V
i
CO
CO
CM
CO
in
00
o
CO
CM
g
CM
CO
§
C^
p
8
CO
in
»—
h-
00
CO
s
o
iri
r*.
o
CM'
§
CM
O)
8
•r-
-------
•a
CO
$
CO
J"
LO O
Is* CO
« CO
"*" >
IT OC
% i
LU
fc
LU
UJH
CD D
LU
h-
LU
oc
I °°.
•" S
m °
i"
2 fe <
p 0 Q
%
%
l«*
%
'%
I/Bui
l/Bm
l/6iu
l/8iu
l/iuu
I/BUI
I/BUI
I/BUI
I/BUI
l/|l"
I/BUI
l/Biu
I/BUI
I/8UJ
I/I"J
CM
r»
s
CO
§
|
ID
O
8
^
•8
CD
O
CM
in
CM
§
o
CO
8
CM
ig
S
1
CM
If)
If)
S
1
1
8
o
0
§
CM
r-
§
§
§
3
o
CM
§
S
§
1
R
o
(O
r-
§
CO
$
co
1
in
r»'
8
°
§
*—
1
g
o
8
-------
•&
U)
CM'
O)
LO O
r-. in
en (f)
•« ' *
cc.
UJ
00
LLI
s<
03 Q
I-'
lil —
H "i,
oc I
UJ =
X n.
i- S
< H
in S
3? 5
> n
§ 7
_r
DC
UJ
CD
>
cc
<
E
a.
_j
cc
%
Z2
< I
5 <
E »-
a.
_i
CC
CO
*
< 1-
13
d £
r w
UJ
CO
e£
S
CM
?
(O
CO
CM
8
CM
S
o
*t
§
o
CM
S
§
§
o
Tl-'
S
CM
CO
co
§
S
in
CO
S
CM
CO
S
S
3
q
CM'
8
CM
CM
co
§
CN
CM
8
O
CM'
3
§
CM
8
s
»~
in
CO
S
CM
to
2
O
J±
CM
O
CM'
8
co
CO
r^
<*
CO
i^
TST
•*
r^-
in
CM
%
co
cs
S
§
o
CM'
S
CO
189
-------
•a
O)
CM
£.
o
in %
t»» li
*"- 2
CO **
cc co*
LU ••
CO >
S cc
LU <
fc 1
LU CC
LO <
CQ D
LU
CC
LU =•
cc 8
D ^
cc
I
o
SLUDGE |
PRIMARY 1
_j
cc
£<«
5*
JU
l/Biu
l/Biu
l/Biu
|/6ui
l/l"i
ui >
183
00
(O
.1
(0
5
§
u>
CO
«••
«
CM
CO
s
CM
3
CM
s
O
CM
8
cV
%
CM
a
S
CM
g
O
r~'
8
8
to
in
CO
Q
<
c
c»
1
O
S
1
1
SJ
o
<«!
O
(
§
i
N
S
o
c
3
CM
c
CO
§
o
a
c*
o
-------
•a
CO
LT)
05
I
«-
cc
111
03
2
UJ
Q_
UJ
r» co
LU H-
-J <
CQ Q
et
cc m
Ul <4-
X o
< o
ui S
> s
i s
< I
E •"
_, Q.
CC
« c
>• < *-
CC h- Z
us
CC _l "-
£L C. U.
H ^
til
CO
cc .
I^J ^T
^J "^ yj
SE S 3
§£
Ul
_l
u.
Z
MOIJ.VU1N30NOO
sanos 3iixviOA
NOI1VUJ.N30NOO SQHOS
( 1VXOJ.) 3WniOA
NOI1VU1N30NOO
sanos 31 U.VIOA
NOI1VUJ.N33NO3 SOIIOS
Knxwiwmonid
««.«.»«.».
sanos a30N3dsns nvioi
sanos
318V31113S
3inNIWO£
sanos a-iaN3dsns asxid
sanos asaNadsns 3iu\noA
sanos a30N3dsns iviox
sanos
au""«-
sanos anawsdsns asxid
sanos aaaNiidsns STIXVIOA
sanos a30N3dsns ivxox
saiios
axnNiw cc
i&i
%
%
|B6
%
%
l/Biu
I/BUI
l/Biu
I/Bui
I/I""
l/Biu
l/Biu
I/Bui
|/6tu
„,.
I/Bui
I/Bui
I/Bui
l/Buj
,/,-
s
CM
1
CD
CO
ri
9
T-
s
1
1
1
o
in
to
'•
§
CM
1
CM
q
S
CM
S
s
o
CM'
o
CO
CM
S
§
1
I
in
«,
CM
CM
?
i
S
in
CO
(O
R
8
S
o
CO
0
s
s
CO
5
o
CO
.
(O
a>
1
8
s
p
s
i
00
§
o
0
o
CO
CM
g
i
g
«
p
in
in
CM
co
R
CO
CO
o
in
O)
o
oi
a
i
i
Si
T—
in
-
0
CO
191
-------
•a
CN
00
O
03
j/J,
00
10 S
S; co
LJU
00 CC
Ul
CL
00 LU
in co
III
!§
cc
CO
I
I
u.
Ul
s
to
tt
1
_l
K
n
5*
CM
IS
o
(>.'
in
«c
CM
O
CM
O
S
8
CM
$
ID
to
CO
CM
§
CO
a
s
P
8
s
CO
o
CO
00
s
1
P
!•>•'
in
t
0
§
CM
§
CM
5
in
00
s
i
§
s
tn
o
s
CM
S
00
s
CM
t—
CO
to
^
in
(D
in
'j-
CM
0
s
en
CO
8
in
CO
CO
t~
00
3
o
*r
1
S,
o
(D
O
CO
CO
192
-------
•g.
CD
{/ ^
CO H
LU CL
h-
0.
LU O)
I- S
oc 12
UU "*.
pc o
LU
™ 01
oc ?:
O ••
eo
UJ
O
o
CO
>
cc
1
E
°"
i
<
_i i
oc <
fi!
1
cc*
<1
5 rf
— 2
OCH
0.
_l
oc
i
12
1*
H 2
»D
= £
ts
u
/>
*5!-
51
= 3
UL
z
NOIJ.VU1N30NOO
sanos 31U.V1OA
NOIJ.VUXN33NOO SOROS
CIV101) 3WmOA
NOU.VUJ.N33NO3
sanos 31UVIOA
NOiivaiNaoNoo sanos
sanos Q3QN3dsns aaxid
sanos a3ON3dsns SIUVIOA
sanos aaoNSdsns ivioi
sanos
?ii(i\m 1 1 ^c
31OMIW OE
sanos a30N3dsns aaxi d
sanos aiaNidsns BIU.VIOA
sanos aa(iN3dsns ivioj.
sanos
3n8V3nj_L3S
3JLONIW 06
sanos a3aN3dsns
^ UL ^
t*2
%
%
|B6
%
%
l/Buj
l/Bm
I/Bui
i/eiu
l/|iu
I/BUI
I/BUI
l/Biu
l/6iu
l/|l"
/BUI
/BUI
/BUI
/fku
I/|W
C
«~
§
1
1
s
CM
»
8
D.
•>
(O
CO
•«»•
p
M
CM
U>
CO
O
CO
CJ
S
8
CO
i
N
q
u>
1-
CM
S
§
i
N
Ifl
0
§
1
s
in
8
N
i
1
i
o
(O
in
«•
CM
C"
§
1
5
in
o>
S
o
N
S
g
in
oi
8
CM
S
8
CM
S
CM
00
If
O
•»
in
*t
CM
00
p
r-
(O
CO
t«
in
T'
r^
p
CM
8
CM
CM
CM
r-*
CM
S
in
o
CO
CO
193
-------
•&
LO 5£
s 1
CO S
LU Q.
Q.
LU
LU H
-1 <
OQ D
LU
H
CC
LU
LU
CC
D
o>
CC
O
u.
Ul
O
Q
i
.
OC
<
Q.
_J
OC
_
>•
5*
< z
-H
E H
o.
CC
i
W
52
§ H
£g
S3
"^ u.
d it
PUI
Ul
o>
n »-
2 2
5 £
§1
S*
t-
2
_i
u.
2
NOI1VU1N33NO3
sanos amvioA
NOI1VH1N30NOO SQI1OS
(1V1O1) 3WmOA
NOI1VH1N30NOO
sanos aiiivlOA
NOI1VUXN30NOO SQIIOS
sanos aaaN3dsns aaxid
sanos aaaNBdsns aiuxnoA
sanos asoMSdsns nviox
sanos
318V31J-L3S
31ONIW OE
sanos aaaNBdsns aaxu
sanos a3QN3dsns snuxnoA
sanos aatiN3dsns nvioi
: sanos
aiawai i I3<5
ainNiw oe
sanos a=iaN3dsns O3xu
sanos a3QN3dsns BIUVHOA
sanos aaoNBdsns nvioi
sanos
31SV31JJ.3S
3JJ1NIW OE
Nl
%
%
|BB,
%
%
/BUI
/Bua
/BUI
/6iu
I/It"
I/BUI
I/Bui
I/Bui
I/6UJ
I/IH«
I/BUI
I/BUI
I/BUI
l/6ui
I/I"J
J5
S
co
s
^
g
1C
c
8
g
I
S
w
5
o
Q
p
£
co
y
^
CO
«c
•p
u
1
j
s
194
-------
T3
q
g
CA
K 5
o> p
•
UJ <
i i
pi
UJ
co
UJ <
cc
LU in
a: •*.
Q ?!
DC
I
I
s
OC*
<2
LU CC
Sa-
" c
2
to
>JH
5 K 5
|S3
£ts
UJ
CO
K j_
OC tr *"
is?
H
UJ
_l
u.
z
NOI1VU1N33NO3
sanos 31 IIVIOA
(1VJ.OJ.) 3WmOA
NOIJ.VUJ.N33NO3
sanos 31 IIVIOA
NOIJ.VU1N33NO3 SQI1OS
sanos a3aw3.jsns aaxi j
««..«»«»»»'
sanos Q3aN3
8
*-
i
8
S
s
S
U)
•-
K
1
I
§
§
§
§
(0
8
q
8
CM
CM
1
i
PI
q
u>
CM
00
8
CO
i
1
o
d
3
<•>
R
CO
00
o
CN
8
00
00
(O
CM
R
CM
1
o
CO
It)
CM
co
1
s
s
o
CO
00
1
CM
(O
o
CD
8
CM
00
$
1
8
q
in
CM
1
1
CM
CO
i
CD
o
CN
.to
CM
g
O
11)
O
CO
CO
195
-------
TABLE 62
DRY-WEATHER REMOVAL OF SETTLEABLE SOLIDS BY VOLUME
SWIRL FLOW 1.3 I/sec (0.3 mgd)
Date Time Settleable Solids ml/I % Removal
1975
9-2
9-3
9-4
9-5
9-8
11.30
12:30
1.30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
Total
Influent
12.5
9.5
7.5
8.0
37.5
9.0
9.5
7.5
8.5
36.5
10.5
10.0
7.5
7.5
35.5
11.5
9.0
8.5
8.5
37.5
13.5
12.0
11.0
8.0
44.5
191.5
Swirl
Effluent
3.0
3.0
2.5
2.5
11.0
3.0
2.0
1.0
1.0
7.0
3.0
2.0
1.0
1.0
7.0
4.0
4.0
2.0
2.0
12.0
2.0
4.0
2.5
2.0
10.5
47.5
Primary
Effluent
10.5
7.0
6.5
8.0
32.0 avg.
2.5
2.5
2.0
2.0
9.0 avg.
5.5
1.0
2.0
2.0
12.5 avg.
7.0
4.0
2.0
2.0
15.0 avg.
7.0
10.0
10.0
2.0
29.0 avg.
97.5 avg.
Swirl
76.0
68.4
66.7
68.7
70.7
66.6
78.9
89.4
88.2
80.8
71.4
80.0
86.7
86.7
80.3
65.2
55.6
76.5
76.5
68.0
85.2
66.7
77.3
75.0
76.4
75.2
Primary
16.0
26.3
13.4
0.0
14.7
72.2
73.7
78.9
76.4
75.3
47.6
90.0
46.7
73.3
64.8
39.1
55.6
76.6
76.5
60 uO
48.1
16.7
9.1
75.0
34.8
49.1
196
-------
TABLE 63
DRY-WEATHER REMOVAL OF SETTLEABLE_SOLIDS BY WEIGHT
SWIRL FLOW 1 3 I/sec (0.3 mgti)
Date Time
1975
9-2 11 :30
12:30
1:30
2:30
Subtotal
9-3 11:30
12:30
1:30
2:30
Subtotal
9-4 11:30
12:30
1:30
2:30
Subtotal
9-5 11 :30
12:30
1:30
2:30
Subtotal
9-8 1 1 :30
12:30
1:30
2:30
Subtotal
Total
Settleable Solids mg/l
Influent
32
192
220
64
508
240
204
316
132
892
144
80
128
216
568
164
236
128
160
688
224
180
220
140
764
3,420
Swirl
Effluent
76
80
100
40
296
12
48
56
36
152
48
60
68
28
204
192
72
68
20
352
164
92
76
20
352
1,356
Primary
Effluent
124
96
200
144
564 avg
52
36
48
16
152 avg.
112
72
104
96
384 avg
84
48
120
48
300 avg
80
84
120
68
352 avg
1,752 avg
% Removal
Swirl
+137.5
58.3
54.5
37.5
41.7
95.0
76.5
82.3
72.7
82.9
66.7
25.0
46.9
87.5
64.1
+ 17.1
69.5
46.9
87.5
48.8
26.8
48.9
65.5
85.7
53.9
60.4
Primary
+ 287.5
50.0
9.1
+ 125.0
+ 11.0
78.3
82.4
84.8
87.9
82.9
22.2
10.0
18.8
55.6
32.4
48.8
79.7
6.3
70.0
56.4
64.3
53.3
45.5
51.4
53.§
48.8
Total (disregarding samples when erosion caused effluent to be higher than influent)
Swirl 3,222 1,088 67.3
Primary 3,324 1,484 avg
55.3
197
-------
TABLE 64
DRY-WEATHER REMOVAL OF TOTAL SUSPENDED SOLIDS
SWIRL FLOW 1 3 I/sec (0.3 mgd)
Date
1975
9-2
9-3
94
9-5
9-8
Time
11:30
12:30
1:30
2'30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2-30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
Total
Total Suspended
Influent
396
380
348
244
1,368
456
372
456
316
1.600
424
492
396
348 ,
1,660
392
388
360
332
1,472
416
356
364
280
1,416
7,516
Solids mg/l
Swirl
Effluent
248
228
216
216
908
232
200
196
188
816
252
200
200
136
788
300
280
196
140
916
252
240
212
148
852
4,280
Primary
Effluent
316
220
312
328
1,176avg
216
180
152
132
680 avg
260
200
200
200
860 avg
264
196
212
164
836 avg
208
204
228
172
81 2 avg
4,364 avg
% Removal
Swirl Primary
37.4
40.0
37.9
11.5 +
33.6
49.1
46.2
57.0
40.5
49.0
40.6
59.3
49.5
60.9
52.6
23.5
27.8
45.6
57.8
37.8
39.4
32.6
41.8
47.1
39.8
43.1
20.2
42.1
10.3
34.4
14.0
52.6
51.6
66.7
58.2
57.5
38.7
59.3
49.5
42.6
48.2
32.6
49.5
41.1
50.6
43.2
50.0
42.7
37.4
38.6
42.7
41.9
Total {disregarding samples when erosion caused effluent to be higher than influent)
Primary 7,272 4,036 avg
46.3
198
-------
TABLE 65
DRY-WEATHER REMOVAL OF VOLATILE SUSPENDED SOLIDS
SWIRL FLOW 1 3 I/sec (0.3 mgd)
Date Time Volatile Suspended Solids mg/l % Removal
1975
9-2 11:30
12:30
1:30
2:30
Subtotal
9-3 11:30
12:30
1:30
2:30
Subtotal
9-4 11:30
12:30
1:30
2:30
Subtotal
9-5 11:30
12:30
1:30
2:30
Subtotal
9-8 11:30
12:30
1:30
2:30
Subtotal
Total
Influent
388
312
252
236
1,188
376
308
420
268
1,372
352
408
380
316
1,456
344
320
332
280
1,276
348
296
272
252
1,168
6,460
Swirl
Effluent
208
192
136
200
736
196
176
176
132
680
216
120
180
92
608
272
236
164
124
796
220
208
152
104
684
" 3,504
Primary
Effluent
248
200
232
280
960 avg
188
156
148
92
584 avg
172
136
192
164
664 avg
212
176
180
128
696 avg
152
164
212
132
660 avg
3,564 avg
Swirl
46.4
38.5
46.0
15.3
38.0
47.9
42.9
58.1
50.7
50.4
38.6
70.6
52.6
70.9
58.2
20.9
26.2
50.6
55.7
37.6
36.8
29.7
44.1
58.7
41.4
45.8
Primary
36.1
35.9
7.9
+ 18.6
19.2
50.0
49.4
64.8
65.7
57.4
51.1
66.7
49.5
48.1
54.4
38.4
45.0
45.8
54.3
45.5
56.3
44.6
22.0
47.6
43.5
44.8
Total (disregarding samples when erosion caused effluent to be higher than influent)
Primary 6,224 3,284 avg
47.2
199
-------
TABLE 66
DRY-WEATHER REMOVAL OF FIXED SUSPENDED SOLIDS
SWIRL FLOW 1 3 I/sec (0.3 mgd)
Date
1975
9-2
9-3
9-4
9-5
9-8
Time
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
Total
Fixed
Influent
8
68 :
96 :
8
180
80
64
36
48
228
72
84
16
32
204
48
68
28
52
196
68
60
92
28
248
1,056
Suspended Solids mg
Swirl
Effluent
40
36
80
16
172
36
24
20
56
136
36
80
20
44
180
28
44
32
16
120
32
32
60
44
168
776
/I
Primary
Effluent
68
20
80
48
216avg
28
24 '
4
40
96avg
88
64
8
36
196 avg
52
20
32
36
140avg
56
40
16
40
1 52 avg
800 avg
% Rerr
Swirl
+400.0
47.0
16.7
+ 100.0
4.4
55.0
62.5
44.4
+ 16.7
40.3
50.0
4.8
+ 25.0
37.5
11.8
41.7
35.3
+ 14.3
69.2
38.8
52.9
46.7
34.8
+ 57.1
32.3
26.5
loval
Primary
+ 750.0
70.0
16.7
• + 500.0
+ 20.0
65.0
62.5
88.9
16.7
57.9
+ 22.2
23.8
50.0
12.5
3.9
+ 8.3
70.6
+ 14.3
30.8
28.6
17.6
33.3
82.6
+ 42.9
38.7
24.2
Total (disregarding samples when erosion caused effluent to be higher than influent)
Swirl 920 568 38.3
Primary 864 472 avg
45.4
200
-------
TABLE 67
DRY-WEATHER REMOVAL OF SETTLEABLE SOLIDS BY VOLUME
SWIRL FLOW 19.7 I/sec (0.45 mgd)
Date Time
1975
9-9 1 1 :30
12:30
1:30
2:30
Subtotal
9-10 11:30
12:30
1:30
2:30
Subtotal
9-11 11:30
12:30
1:30
2:30
Subtotal
9-12 11:30
12:30
1:30
2:30
Subtotal
9-15 11:30
12:30
1:30
2:30
Subtotal
Total
Settleable Solids ml /I
Influent
13.0
12.0
8.0
10.0
43.0
13.0
12.0
11.0
10.5
46.5
14.0
13.5
11.5
9.5
48.5
12.5
11.0
9.5
9.0
42.0
12.5
14.0
12.0
9.0
47.5
227.5
Swirl
Effluent
5.0
4.5
3.0
.5.0
17.5
7.0
7.0
7.0
6.5
27.5
8.0
7.0
6.0
4.0
25.0
6.0
6.5
6.0
5.0
23.5
8.5
7.0
6.0
4.0
25.5
119.0
Primary
Effluent
7.0
3.5
4.0
4.5
19.0avg
13.5
6.5
8.5
6.0
34.5 avg
18.0
13.5
9.5
7.5
48.5 avg
2.0
3.0
3.0
7.0
10.0 avg
12.0
10.0
6.0
5.0
33.0 avg
145.0 avg
% Removal
Swirl
61.5
62.5
62.5
50.0
59.3
46.2
41.7
36.4
38.1
40.9
42.9
48.1
47.8
57.9
48.5
52.0
40.9
36.8
44.4
44.0
32.0
50.0
50.0
55.6
46.3
47.7
Primary
46.2
70.8
50.0
. 55.0
55.8
+ 3.8
45.8
22.7
42.9
25.8
+ 28.6
0.0
17.4
21.1
0.0
84.0
72.7
68.4
77.8
76.2
4.0
28.6
50.0
44.4
30.5
36.3
Total (disregarding samples when erosion caused effluent to be higher than influent)
Primary 200.5 113.5 avg
43.4
201
-------
TABLE 68
DRY-WEATHER REMOVAL OF SETTLEABLE SOLIDS BY WEIGHT
SWIRL FLOW 19.7 I/sec (0.45 mgd)
Date
1975
9-9
9-10
9-11
9-12
9-15
Time
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
Total
Settleable Solids mg/l % Removal
Influent
136
192
128
380
836
252
248
152
128
780
284
220
280
132
916
228
160
256
196
840
608 :
100
148
116 :
972
4,344
Swirl
Effluent
104
108
124
132
468
160
76
160
116
512
172
172
180
48
572
144
136
132
148
560
136
72
140
60
, 408
2,520
Primary
Effluent
112
84
84
156
436 avg
164
96
124
84
468 avg
284
172
144
56
656 avg
47
88
—
52
187 avg
, 120
144
92
80
436 avg
2,1 83 avg
Swirl
23.5
43.8
3.1
65.3
44.0
36.5
69.4
+ 5.3
9.4
34.4
39.4
21.8
35.7
63.6
37.6
36.8
15.0
48.4
24.5
33.3
77.6
28.0
5,4
48.3
58.0
42.0
Primary
17.6
56.3
34.4
58.9
47.8
34.9
61.3
18.4
34.4
40.0
0.0
21.8
48.6
57.6
28.4
79.4
45.0
—
73.5
77.7
80.3
+ 44.0
37.8
31.0
55.1
49.7
Total (disregarding samples when erosion caused effluent to be higher than influent)
Swirl 4,192 2,360 43.7
202
-------
TABLE 69
DRY-WEATHER REMOVAL OF TOTAL SUSPENDED SOLIDS
SWIRL FLOW 19.7 I/sec (0.45 mgd)
Date Time
1975
9-9 1 1 :30
12:30
1:30
2:30
Subtotal
9-10 11:30
12:30
1:30
2:30
Subtotal
9-11 11:30
12:30
1:30
2:30
Subtotal
9-12 11:30
12:30
1:30
2:30
Subtotal
9-15 11:30
12:30
1:30
2:30
Subtotal
Total
Total Suspended Solids
Influent
336
376
352
528
1,592
528
356
348
360
1,592
472
436
636
344
1,888
432
364
524
400
1,720
840
416
376
296
1,928
8,720
Swirl
Effluent
300
308
328
372
1,308
328
292
328
312
1,260
344
384
436
252
1,416
324
344
308
372
1,348
364
340
276
204
1,184
6,516
mg/l
Primary
Effluent
304
264
252
280
1,100
376
260
280
204
1,120
460
384
340
272
1,456
208
208
280
168
864
340
308
220
172
1,040
5,580
% Removal
Swirl
10.7
18.1
6.8
29.5
17.8
37.9
18.0
5.7
13.3
20.8
27.1
11.9
31;4:
26.7
25.0
25.0
5.5
41.2
7.0
21.6
56.7
18.3
26.6
31.1
38.6
25.3
Primary
9.5
29.8
28.4
47.0
30.9
28.8
27.0
19.5
43.3
29.6
2.5
11.9
46.5
20.9
22.9
51.8
42.9
46.6
58.0
49.8
59.5
26.0
41.5
41.9
46.0
36.0
203
-------
TABLE 70
DRY-WEATHER REMOVAL OF VOLATILE SUSPENDED SOLIDS
SWIRL FLOW 19.7 I/sec (0.45 mgd)
Date
1975
9-9
9-10
9-11
9-12
9-15
Time
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
11:30
12:30
1:30
2:30
Subtotal
Total
Volatile Suspended Solids mg/l
Influent
276
272
276
404
1,228
470
316
320
304
1,410
420
376
560
276
1,632
356
296
392
332
1,376 !
742
336
332
288
1,698
7,344
Swirl
Effluent
240
224
244
304
1,012
316
260
308
248
1,132
304
332
364
220
1,220
264
244
240
316
1,064
284
268
268
196
1,016
5,444
Primary
Effluent
224
192
156
260
832
336
240
280
140
996
392
320
308
212
1,232
196
160
276
136
768
268
300
188
132
888
4,716
% Removal
Swifl
13.0
17.6
11.6
24.7
17.6
32.8
17.7
3.8
18.4
19.7
27.S
11.7
35.0
20.3
25.2
25.8
17.6
38.8
4,8
22.7
61.7
20.2
19.3
31.9
40.2
25.9
Primary
18.8
29.4
43.5
35.6
32.2
28.5
24.0
12.5
53.9
29.4
6.7
14.9
45.0
23.2
24.5
44.9
45.9
29.6
59.0
44.2
63.9
10.7
43.4
54.2
47.7
35.8
204
-------
TABLE 71
DRY-WEATHER REMOVAL OF FIXED SUSPENDED SOLIDS
SWIRL FLOW 19.7 I/sec (0.45 mgd)
Date Time
1975
9-9 11:30
12:30
1:30
2:30
Subtotal
9-10 11:30
12:30
1:30
2:30
Subtotal
9-11 11:30
12:30
1:30
2:30
Subtotal
9-12 11:30
12:30
1:30
2:30
Subtotal
9-15 11:30
12:30
1:30
2:30
Subtotal
Total
Fixed Suspended Solids
Influent
60
104
76
124
364
56
40
28
56
180
52
60
76
68
256
76
68
132
68
344
98
80
44
8
230
1,374
Swirl
Effluent
60
84
84
68
296
12
32
20
64
128
40
52
72
32
196
60
100
68
56
284
80
72
8
8
168
1,072
ppm
Primary
Effluent
80
72
96
20
268 avg
40
20
0
64
124 avg
68
64
32
60
224 avg
12
48
4
32
96 avg
72
8
32
40
1 52 avg
864 avg
% Removal
Swirl
00.0
19.2
+ 10.5 ,
45.2
18.7
78.6
20.0
28.6
+ 14.3
28.9
23.0
13.3
5.3
52.9
23.4
21.0
+ 47.0
48.5
17.6
17.4
18.4
10.0
81.8
0.0
26.9
22.0
Primary
+•-33.3
30.8
+ 26.3
83.9
26.4
28.6
50.0
100.0
+ 14.3
31.1
+ 30.8
+ 6.7
57.9
11.8
12.5
84.2
29.4
97.0
52.9
72.1
26.5
90.0
27.3
+ 40.0
33.9
37.1
Total (disregarding samples when erosion caused effluent to be higher than influent)
Swirl 1,174 824 29.8
Primary 1,062 452 avg
57.4
205
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-78-122
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
THE SWIRL PRIMARY SEPARATOR:
DEMONSTRATION
DEVELOPMENT AND, PILOT
5. REPORT DATE
August 1978 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
APWA 73-3
Richard H. Sullivan, Morris M. Cohn, James t.
lire, Fred Parkinson, 6. Galiana, Ralph R. Boericke,
Carl Koch. Paul Zielinski
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
American Public Works Association
1313 East 60th Street
Chicago, Illinois 60637
10. PROGRAM ELEMENT NO.
1BC611
11. CONTRACT/GRANT NO.
68-03-0272
S803157
12. SPONSORING AGENCY NAME AND ADDRESS . „. -,.
Municipal Environmental Research Laboratory—Cm.,OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/14
is.SUPPLEMENTARY NOTES This report supplements EPA-R2-72-008, EPA-670/2-74-039, and
P.O. Richard Field (201) 321-6674 FT$ 340-6674 EPA-670/2-75-011.
Hugh Masters (201) 321-6678 FTS 340-6678
stud ^ concjucted to determine if the swirl concentrator principal could
be used to provide primary treatment to combined sewer overflows and municipal waste-
water. A hydraulic model with synthetic wastewater and a mathematical model were both
used to arrive at an optimized configuration and a design basis. The design was then
tested under actual wet- and dry-weather flow conditions using a large scale, 1,137 cu
m/d (0.3 mgd) pilot constructed in Toronto, Canada. The Toronto pilot evaluations con
firmed the.accuracy of the design (and associated design curves) developed under the
The model and pilot studies indicated that the device could achieve 30 to 50 percent.
settleable solids removal efficiency for flows of less than 22 I/sec (0.5 mgd) at cost:
comparable to, or less than, conventional treatment units. Overflow rates of two to
three times that of conventional units make possible the saving.
Testing of the model and prototype was based upon the need to treat both domestic
sanitary sewage and combined sewer overflows. Extensive laboratory work was conducted
to determine the settling characteristics of solids to provide laboratory control and
provide a correlation between the laboratory and prototype testing programs.
The swirl's height and diameter are equal, providing a relatively deep structure
which enhances sludge thickening. . . .
The Toronto pilot evaluations of the prototype unit constructed in Toronto Ontario -
1,137 cu m/d (0.3 mgd) - to determine operating efficiencies confirmed the accuracy of
design and associated curves developed under the model studies.-
The report contains thorough descriptions of the hydraulic/mathematical and pilot
studies, and most importantly, the detailed design methodology.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Waste treatment, Design criteria,
Hydraulic structures, Combined sewers,
Overflows, Prototypes
Swirl primary separator,
Urban stormwater runoff,
Combined sewer overflow,
Sanitary sewage primary
.tr.eatments .Prototype..
tests
13B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)'
Unclassified
21. NO. OF PAGES
216
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
206
* U.S. GOVERNMENT PRIMING OFFICE; 1978— 757 -140 /138 0
-------
-------
Do s§
*. Q CD CD
S _ c"i
•oo
ro
ro
5*
Z TJ
m
1
n
O m
-• 3 5
o ° -
O
-n
O
o
^??
?^S
H
m
m
m
r1
r w
O co
m
U)
§ I '
|. ^ ^ mi
CD
C/3
05
D) c/5 O)
:± CD
QJ
o 9-
^ I^Sl
s i ° ml
K3 CD <
o
» 2 s- O
O g
§ 1
£T ^
CD •-*
Ql
ml
PP!
ol
0)
T3
CD o
J > >
> r z
CO "0 D
O
H rn
m 0)
O T,
O Fi
O
------- |