EPA 625/l-75-003a
PROCESS DESIGN MANUAL
FOR
SUSPENDED SOLIDS REMOVAL
U.S. ENVIRONMENTAL PROTECTION AGENCY
Technology Transfer
January 1975
-------
ACKNOWLEDGEMENTS
This manual was prepared for the Technology Transfer Office of the U.S. Environmental
Protection Agency by Hazen and Sawyer, Engineers. The previous edition of this manual
(October, 1971) was prepared for Technology Transfer by Burns and Roe, Inc. Major
U.S.VEjPA. contributors and ; reviewers were 'J.F. Kreissl and J.M. Smith : of the
U.S.EPA National Environmental Research Center, Cincinnati, Ohio, and D.J. Lussier of
Technology Transfer, Washington, D.C.
NOTICE
The mention of trade names or commercial products in this publication is for illustration
purposes and does not constitute endorsement or recommendation for use by the U.S.
Environmental Protection Agency.
-------
ABSTRACT
This manual surveys current practice in the removal of suspended solids in both traditional
and advanced treatment of municipal wastewater. Specific processes are described, design
considerations are discussed and results are illustrated by data from actual installations.
Included are processes such as sedimentation, straining and granular media filtration
which affect physical separation of solids as well as coagulation and flocculation processes
which alter solids characteristics to facilitate such separation. Detailed information is also
provided concerning handling and application of coagulant chemicals.
Aspects of operation and maintenance pertinent to design are discussed and estimated
costs of construction and operation are provided for particular processes.
in
-------
TABLEOFCONTENTS
CHAPTER PAGE
ACKNOWLEDGMENTS jj
ABSTRACT jjj
TABLEOFCONTENTS v
LIST OF FIGURES vii
LIST OF TABLES xjji
FOREWARD xv
1 INTRODUCTION
1.1 Purpose 1-1
1.2 Wastewater Solids 1-1
1.3 References 1-3
2 GENERAL DESIGN CONSIDERATIONS
2.1 Applications of Suspended 2-1
Solids Separation Processes
2.2 Process Selection 2-1
3 FLOW VARIATIONS AND EQUALIZATION
3.1 Flow Variation 3-1
3.2 Performance vs. Flow Variation 3-1
3.3 Flow Equalization 3-2
3.4 References 3-2
4 PRINCIPLES OF CHEMICAL TREATMENT
4.1 Introduction 4-1
4.2 Destabilization Mechanisms 4-1
4.3 Selection of Chemical Coagulants 4-2
4.4 Coagulation Control 4-6
4.5 References 4-16
5 STORAGE AND FEEDING OF CHEMICALS
5.1 General 5-1
5.2 Aluminum Compounds 5-1
5.3 Iron Compounds 5-16
5.4 Lime 5-24
5.5 Other Inorganic Chemicals 5-34
5.6 Polymers 5-50
5.7 Chemical Feeders 5-59
5.8 References 5-66
6 CHEMICAL MIXING, FLOCCULATION AND
SOLIDS-CONTACT PROCESSES
6.1 Introduction 6-1
6.2 Chemical Mixing 6-3
6.3 Flocculation 6-5
6.4 Solids-Contact Processes 6-8
6.5 References 6-12
-------
TABLE OF CONTENTS - Continued
CHAPTER PAGE
7 GRAVITY SEPARATION
10
7.1 Introduction
7.2 Configuration of Sedimentation units
7.3 Basic Factors Affecting Settling Tank Design
7.4 Clarifier Design Considerations
7.5 Primary Sedimentation
7.6 Secondary Sedimentation
7.7 Chemical Sedimentation
7.8 Flotation
7.9 Shallow Settling Devices
7. 1 0 Wedge-Wire Settlers
7.1 1 References
PHYSICAL STRAINING PROCESSES
8.1 General
8.2 Wedge-Wire Screens
8.3 Microscreening
8.4 Other Screening Devices
8.5 Diatomaceous Earth Filters
8.6 Ultrafiltration
8.7 References
GRANULAR MEDIA FILTRATION
9.1 Introduction
9.2 Process Alternatives
9.3 Process Variables
9.4 Selection of Filtration Rate
and Terminal Headless
9.5 Filtration Media
9.6 Filter Control Systems
9.7 Filter Cleaning Systems
9.8 Filter Structures and General
Arrangement
9.9 Pilot Studies
9.10 Special Designs
9.11 Slow Sand Filters
9.12 References
COST ESTIMATES
10.1 Introduction
10.2 Curve Content
10.3 Operation and Maintenance Costs
10.4 Freight
10.5 How to Use Cost Curves
10.6 Curve Descriptions
10.7 References
7-1
7-1 •
7-2
7-9
7-14
7-15
7-22
7-23
7-27
7-32
7-35
8-1
8-1
8-9
8-31
8-33
8-36
8-43
9-1
9-3
9-9
9-20
9-26
9-31
9-35
9-45
9-46
9-48
9-52
9-54
10-1
10-1
10-1
10-2
10-2
10-2
10-16
VI
-------
LIST OF FIGURES
Figure No. Page
4-1 Jar Test Units With Mechanical (Top) 4-8 .
and Magnetic (Bottom) Stirrers
4-2 Six-Position Sampler 4-9
4-3 Settling Curves Frequently Obtained 4-10
4-4 Jar Test Results 4-12
4-5 Zeta Potential Apparatus 4-13
4-6 Coagulation of Raw Sewage With Alum 4-15
5-1 Typical Dry Feed System 5-7
5-2 Crystallization of Alum Solutions 5-11
5-3 Viscosity of Alum Solutions " 5-12
5-4 Alternative Liquid Feed Systems 5-14
For Overhead Storage
5-5 Alternative Liquid Feed Systems 5-14
For Ground Storage
5-6 Freezing Point Curves For Commercial 5-17
Ferric Chloride Solutions
5-7 Viscosity vs. Composition of Ferric 5-18
Chloride Solutions at Various Temperatures
5-8 Typical Lime Feed System 5-31
5-9 Lime Requirement For^pH > 11.0 as a 5-35
Function of the Wastewater Alkalinity
5-10 Viscosity of Soda Solutions 5-37
5-11 Viscosity of Caustic Soda Solutions 5_40
5-12 Typical Caustic Soda Feed System 5-45
vi i
-------
LIST OF FIGURES (continued)
Figure No. Page
5-13 Typical Schematic of a Dry Polymer Feed System 5-57
5-14 Typical Automatic Polymer Feed 5-58
System for Large Plants
5-15 Positive Displacement Pumps 5-61
5-16 Screw Feeder 5-63
5-17 Positive Displacement Solid Feeder-Rotary 5-63
6-1 Impeller Mixer 6-4
6-2 Mechanical Flocculation Basin 6-7
Horizontal Shaft-Reel Type
6-3 Mechanical Flocculator Vertical 6-7
Shaft-Paddle Type
6-4 Solids Contact Clarifier Without 6-9
Sludge Blanket Filtration
6-5 Solids Contact Clarifier With Sludge 6-10
Blanket Filtration
7-1 Rectangular Settling Tanks 7-3
7-2 Typical Clarifier Configurations 7-4
7-3 Results of Salt-Injection Tests With 7-6
Different Types of Sedimentation Tanks
7-4 Schematic Representation of Settling Zones 7-10
7-5 Sedimentation In a Secondary Settling Tank 7-10
7-6 Dependence of MLSS Concentration on Secondary 7-17
Settling Tank Underflow Concentration
7-7 Typical Flotation Unit 7-25
7-8 Module of Steeply Inclined Tubes 7-28
Vlll
-------
LIST OF FIGURES (continued)
Figure No. Page
7-9 Tube Settlers in Existing Clarifier 7-29
7-10 Plan View of Modified Clarifier 7-29
7-11 Tube Settler-Flow Pattern 7-30
7-12 Simple Wedge Wire Clarifier 7-33
7-13 Installation of Wedge Wire Panels in a Clarifier 7-33
8-1 Hydrasieve Unit 8-3
8-2 Schematic 8-3
8-3 Screen Detail 8-4
8-4 Curved Screen Bars 8-4
8-5 Rotating Wedge Wire Screen at North 8-8
Chicago ST.P.
8-6 Typical Microscreen Unit 8-10
8-7 Micro-Matic® Strainer 8-11
8-8 Microscreen Capacity Chart 8-16
8-9 Microscreen Removal at Various Flow Rates 8-23
8-10 Microscreening of Trickling Filter Plant Effluent 8-24
8-11 Microscreen Unit With Pleated Outer Surface 8-27
8-12 The Sweco Concentrator 8-32
8-13 Vertical Leaf Vacuum Filter 8-34
8-14 Vertical Leaf Pressure Filter 8-35
8-15 Schematic Flow Diagram of the Pikes Peak 8-37
Treatment & Reuse System
IX
-------
LIST OF FIGURES (continued)
Figure No. Page
8-16 "Storage Battery" Membrane Modules 8-41
9-1 Typical Rapid Sand Filter 9-4
9-2 Filter Configurations 9-5
9-3 Cross Section of Upflow Filter 9-6
9-4 Typical Pressure Filter 9-8
9-5 Run Length vs. Filter Rate for Various 9-21
Terminal Headlosses
9-6 Net Production Rate vs. Filter Rate 9-24
for Various Run Lengths
9-7 Grain Size Curve 9-27
9-8 Flow Control Systems 9-33
9-9 Automatic Gravity Filter, Single Compartment 9-36
9-10 Minimum Fluidization Velocity, Vmf, to Achieve 9-37
10 Percent Bed Expansion at 25 °C
9-11 Effect of Temperature on Vmf for Sand and Coal 9-38
and on Absolute Viscosity of Water
9-12 Underdrains 9-40
9-13 Simater Filter 9-49
9-14 Hydromation In-Depth Filter 9-50
9-15 Hardinge Travelling Backwash Filter 9-51
10-1 Flocculators-Flash Mixers 10-3
10-2 Chemical Feed Systems 10-5
10-3 Sedimentation 10-7
-------
LIST OF FIGURES (continued)
Figure No. Page
10-4 Solids Contact 10-8
10-5 . Flotation 10-10
10-6 Wire Septums and Settling Tubes 10-11
10-7 Wedge Wire Screens: Rotating and Stationary 10-13
10-8 Microscreens 10-14
10-9 Media Filters 10-15
XI
-------
LIST OF TABLES
Table No. page
2-1 Selected SS Separation Process Applications 2-2
4-1 SS Removal Performance For Chemical Coagulation 4-4
Applications To Phosphate Removal
5-1 Partial List of Alum Manufacturing Plants 5-2
5-2 Reactions of Aluminum Sulfate 5-15
5-3 Reactions of Ferric Sulfate 5-23
5-4 Reactions of Ferrous Sulfate 5-24
5-5 Partial List of Lime Manufacturing Plants 5-25
5-6 Reactions of Lime 5-34
5-7 Partial List of Caustic Soda Manufacturing Plants 5-39
5-8 COaYields of Common Fuels 5-46
5-9 Partial List of Carbon Dioxide Manufacturing Plants 5-47
5-12 Partial List of Polymer Sources and Trade Names 5-51
5-11 Types of Chemical Feeders 5-65
7-1 Performance of Special Settling Tank Inlets 7-13
7-2 Typical Design Parameters For Primary Clarifiers 7-14
7-3 Typical Design Parameters For Secondary Clarifiers 7-16
7-4 Dissolved-Air Flotation Applications 7-26
7-5 Tube Settler Installations 7-31
xin
-------
LIST OF TABLES (continued)
Table No. Page
8-1 Physical Straining Processes 8-2
8-2 Specifications of Hydrasieves 8-5
8-3 Wedge-Wire Screens: Municipal Treatment Installations 8-7
8-4 Data Sheet-Wedge Wire Screens 8-9
8-5 Microscreen Design Parameters 8-13
8-6 Microscreen Installations 8-17
8-7 Municipal Microscreen Installations 8-18
8-8 Typical Microscreen Power and Space Requirements 8-28
8-9 Diatomaceous Earth Filtration of Secondary Effluent 8-36
8-10 Results of Ultrafiltration Installations 8-38
8-11 Summary of Pikes Peak Data 8-39
8-12 Typical Membrane Specifications 8-42
9-1 Results of Studies of Filtration of Effluents From 9-12
Secondary Biological Treatment
9-2 Results of Studies of Filtration of Chemically 9-15
Treated Secondary Effluent
9-3 Results of Studies of Filtration Following Chemical 9-16
Treatment of Primary or Raw Wastewater
9-4 Expected Effluent Suspended Solids From Multimedia 9-17
Filtration of Secondary Effluent
9-5 Typical Media Designs For Filters 9-30
9-6 Filter Gravel Design 9-43
xiv
-------
FOREWARD
The formation of the United States Environmental Protection Agency marked a new era
of environmental awareness in America. This Agency's goals are national in scope and
encompass broad responsibility in the areas of air and water pollution, solid wastes,
pesticides, and radiation. A vital part of EPA's national water pollution control effort is
the constant development and dissemination of new technology for wastewater treatment.
It is now clear that only the most effective design and operation of wastewater treatment
facilities, using the latest avaiable techniques, will be adequate to meet the future water
quality objectives and to ensure continued protection of the nation's waters. It is
essential that this new technology be incorporated into the contemporary design of waste
treatment facilities to achieve maximum benefit of our pollution control expenditures.
The purpose of this manual is to provide the engineering community and related industry
a new source of information to be used in the planning, design and operation of present
and future wastewater treatment facilities. It is recognized that there are a number of
design manuals, manuals of standard practice, and design guidelines currently available in
the field that adequately describe and interpret current engineering practices as related to
traditional plant design. It is the intent of this manual to supplement this existing body
of knowledge by describing new treatment methods, and by discussing the application of
new techniques for more effectively removing a broad spectrum of contaminants from
wastewater.
Much of the information presented is based on the evaluation and operation of pilot,
demonstration and full-scale plants. The design criteria thus generated represent typical
values. These values should be used as a guide and should be tempered with sound
engineering judgment based on a complete analysis of the specific application.
This manual is one of several available through the Technology Transfer Office of EPA to
describe recent technological advances and new information. This particular manual was
initially issued in October of 1971 and this edition represents the first revision to the
basic text. Future editions will be issued as warranted by advancing state-of-the-art to
include new data as it becomes available, and to revise design criteria as additional
full-scale operational information is generated.
-------
CHAPTER 1
INTRODUCTION
1.1 Purpose
This manual is intended to provide:
1. A basis for selection of processes to meet specific suspended solids (SS) removal
^requirements
2. A basis for design of particular processes
3. A basis for selection of particular equipment configurations for a given process.
Since the emphasis is on information applicable to design or modifications of solids remov-
al facilities, only those processes are included for which reliable data from actual appli-
cations are available.
1.2 Wastewater Solids
The total solids in wastewater exist in a distribution of sizes from individual ions up to vis-
ible particles. Specific analytical procedures (1) have been established to distinguish the sus-
pended fraction of the total solids and to further distinguish the settleable fraction within
the SS. A typical concentration of SS for raw domestic wastewaters is 200 mg/ 1, but this
can vary substantially from system to system (see below). The lower limiting size for the SS
fraction (about 1.5 microns) is arbitrarily defined by the test procedures and it should be
noted that variations in test procedures themselves can also lead'to widely varying results, es-
pecially at the low solids levels characteristic of treated effluents.
Other workers (2) (3) (4) have applied procedures which distinguish four solids fractions,
and determine proportions of other wastewater characteristics such as COD, Nitrogen, Vol-
atile (organic) matter in each fraction. For a New Jersey municipal raw wastewater, solids
distribution in terms of these fractions was found to be as follows (2):
Fraction
Soluble
Colloidal
Supra-
Colloidal
Settleable
<0.001
0.001-1
1-100
>100
351
31
57
74
Raw Wastewater
Volatile
Matter
>mg/l
116
Secondary Effluent
Volatile
Matter
mg/1
23
43
59
312
8
28
0
62
6
24
0
1-1
-------
The settleable and supracolloidal fractions together are essentially equivalent to the sus-
pended fraction referred to above. Dividing lines between fractions again are somewhat ar-
bitrary depending on tests applied, and overall concentrations in different fractions can vary
substantially between systems depending on factors such as water use, travel time in sewers,
ground-water infiltration, and prevalence of home garbage grinding. Contributions of dis-
solved, colloidal and suspended solids from individual homes, multi-family dwellings or oth-
er point sources often have concentrations two or more times the average for a whole sys-
tem (5).
In addition to particle size, specific gravity and strength or shear resistance of wastewater
solids may affect solids separation performance. The three basic types of solids separation
processes—gravity separation, physical straining, and granular media filtration are dis-
cussed in Chapters 7, 8 and 9, respectively. Wastewater solids characteristics can be altered
to enhance performance of the above separation processes. Chapters 4 and 6 discuss chem-
ical treatment (precipitation and/ or coagulation) and physical treatment (flocculation)
aimed at alteration of solids characteristics. In addition, during the separation processes
themselves, agglomeration and compaction of solids generally continues, increasing separa-
tion efficiency and reducing the volume of separated solids.
Biological wastewater treatment processes also affect solids characteristics and hence solids
separation. Activated sludge solids have been found (6) to have a distinct bimodal dis-
tribution with one mode in the supracolloidal to settleable range and another near the bor-
der between the colloidal and supracolloidal fractions. The concentrations and size limits in
each range are affected by conditions in the biological reactor (Chapter 6). Dean (7) has
noted that bacteria, cellular debris, etc. fall into the finer (colloidal-supracolloidal) range.
Agglomeration of these finer solids generally increases the efficiency of subsequent separa-
tion processes.
1-2
-------
1.3 References
1. Standard Methods for the Examination of Water and Waste-water, 13th Edition,
American Public Health Association, New York (1971).
-i
2. Rickert, David A. and Hunter, Joseph V., General Nature oj Soluble and Paniculate
Organics in Sewage and Secondary Effluent, Water Research, 5, 421 (1971).
3. Hunter, J. V., and Heukelekian, H. The Composition of Domestic Sewage Fractions,
JWPCF, 37, 1142 (Aug. 1965).
4. Helfgott, T., Hunter, J. V. and Rickert, D., Analytic and Process Classification of Ef-
fluents, Jour. SED, ASCE, 96, 779 (June 1970).
5. Rawn, A. M., Some Effects of Home Garbage Grinding Upon Domestic Sewage, The
American City, 66, 110 (Mar. 1951)
6. Tchobanoglous, G., and Eliassen, R., The Filtration of Treated Sewage Ejjluent, Pro-
ceedings of the 24th Industrial Waste Conference, Purdue University Engineering
Bulletin, Extension Series No. 135, 1323 (May, 1969).
7. Dean, Robert B., Colloids Complicate Treatment Process, Env. Sci. and Tech., 3, 820
(Sept. 1969).
1-3
-------
Meaningful cost comparisons usually involve practically the entire process configuration of
the treatment facility, including processes for disposal of solid residues, and reflect how the
individual unit processes affect one another.
Cost data on individual processes for suspended solids removal are given in Chapter 10.
Outlined below are some additional factors not reflected in the unit process cost figures, but
which may warrant consideration in overall comparisons.
1. Sludge Handling. Where chemical treatment is used to remove BOD or phos-
phates or improve SS removals, significant quantities of chemical sludge are pro-
duced. The cost of disposal of this sludge must be considered in process selection
unless configurations being compared involve similar chemical treatment. The ac-
tual cost involved will depend greatly on the particular method of sludge disposal to
be used. For general guidance, in a 10 mgd plant using thickening, digestion, vacu-
um filtration and landfill for sludge disposal, chemical addition of 200 mg/1 alum
would increase the sludge disposal costs by almost 20 percent (from 9.7 cents/1000
gal to 11.5 cents/ 1000 gal of plant flow on a 1972 cost basis) (Chapter 10, Refer-
ence 14).
Where this difference appears significant in the comparison of alternatives for SS
removal, specific sludge disposal figures should be included in process comparisons.
Information on expected sludge quantities from particular chemical treatment pro-
cesses is provided in Chapter 4.
2. Buildings. The need for housing specific unit process varies with climate and other
local conditions. Where the housing requirements of alternative processes obviously
differ widely under particular local conditions, building cost should be considered
in the selection.
*,
3. Land Requirements..Generally, land requirements are a small enough factor in
overall cost that the differences for various process alternatives are not significant.
Where adequate land is unavailable or very costly, however, area requirements of
alternative processes should be compared in detail. Minimum land requirements
may be estimated, at between 1.20 (large plants) and 2.0 (small plants) times the
area of the process units themselves.
(
4. Head requirements. .Some of the processes employed for SS separation (sedimen-
tation, microscreens, etc.) require relatively small head (only 2 to 3 ft. to overcome
losses at inlet and effluent controls and in connecting piping). Others, such as gran-
ular-media filters, and wedge-wire screens, require greater differential head (10 ft
or more). Differences in head requirements are most significant where they necessi-
tate capital outlay for an extra pumping step. The costs for pumping, however, even
with lifts above 10 ft. are usually not large in relation to the overall costs for treat-
ment facilities.
2-3
-------
CHAPTER 2
GENERAL DESIGN CONSIDERATIONS
2.1 Applications of Suspended Solids Separation Processes
Processes for SS separation may fill three distinct functions in wastewater treatment.
1. Pretreatment to protect subsequent processes and reduce their loadings to re-
quired levels
2. Treatment to reduce effluent concentrations to required standards
3. Separation of solids to produce concentrated recycle streams required to maintain
other processes.
In the first two functions effluent quality is the prime consideration, but where the third
function must be fulfilled along with one of the others, design attention must be given to
conditions for both the separated solids (sludge) and the process effluent.
Table 2-1 compares several SS separation process applications selected to illustrate how
their performance and their loading requirements are functions of their applications.
Wedge-wire screens can operate at very, high hydraulic and solids loadings, but do not great-
ly reduce SS. Hence, wedge wire screens are limited to pretreatment applications where
subsequent processes will assure production of a satisfactory final effluent. They can be
considered as an adjunct to primary sedimentation or, where conditions prescribe, as an
alternative.
Sedimentation units must operate at relatively low hydraulic loadings (large space require-
ments), but can accept high solids loadings. With proper chemical or biological pre-
treatment and design, they can produce good quality effluents.
Microscreens and granular-media filters, operating at significantly higher hydraulic loads
than sedimentation units, can produce an effluent with lower SS than is possible with sedi-
mentation alone. In general they are not designed to accept high solids loadings, and are
normally used following other processes which put out relatively low effluent SS concentra-
tions.
2.2 Process Selection
Selection of one of the alternative processes can be based on cost only where all factors not
reflected in cost are equivalent. Direct cost comparison of individual solids removal pro-
cesses usually proves impossible because of differences in factors such as: 1) effluent quali-
ty, 2) pretreatment requirements, 3) effects on sludge processing, 4) housing, space and
head requirements.
2-1
-------
Type of Separation
Process
Straining
Wedge Wire Screens
TABLE 2-1
SELECTED SS SEPARATION PROCESS APPLICATIONS
Application
Preliminary Treatment
of Raw Wastewater
Hydraulic
gpm/sq ft
10-30
Typical Loading Ranges
InH. Solids
mg/1 Ib/day/sq ft
200
25-75
Expected Effluent
SS(a)(b)
mgTl
150-190
Remarks
K)
Microscreens
Gravity Separation
Plain Sedimentation
Chemical Coagula-
tion and
Settling
Plain Sedimentation
(secondary)
Polishing of Bio- 3-10 30 1-2
logical System Effluent
Primary Treatment 0.4-1.6 200 0.5-2
Chemical Treatment 0.3-1.0 200(c) 1-6
of Raw Sewage
(Phosphate Removal
Levels)
Separation of Solids 0.25- 2000- 4-40
after Activated 0.75 5000
Sludge Treatment
5-15
120-80
20-60
10-50
Chemical treatment for
90% + phosphorus removal.
Upper effluent quality
Limit may increase with poor
biological treatment.
Allowable solids loadings de-
pends on solids characteristics.
Granular Media
Filtration
Polishing of Bio- 4-8
logical Effluent or
Filtration of Chemi-
cally-Coagulated and
Settled Raw 3-5
Wastewater or
Secondary Effluent 3-5
30
40
5-10
1-2
1-2
1-2
5-15
10-20
1-3
Secondary treatment may be
biological or by activated
carbon.
(a) Based on raw wastewater SS of 200 mg/1.
Xb) Performance is highly dependent on character of solids appli
|) Influent solids do not include chemical solids.
id hence on conditions in prior treatment.
-------
CHAPTER 3
FLOW VARIATIONS AND EQUALIZATION
3.1 Flow Variation
Both the rate and characteristics of the inflow to most treatment plants vary significantly
with time. Diurnal cycles are found in all domestic discharges. Weekly and seasonal cycles
are common in municipal systems as are variations between wet and dry weather.
Even where only domestic flows are involved, the magnitude of variations can differ widely
between different systems depending on system configuration, water use habits of the
population and opportunities for groundwater infiltration or direct inflow of surface or
subsurface drainage. Industrial and institutional flows where significant, can further alter
domestic patterns.
Because of these wide differences, design of treatment facilities should be based, whenever
possible, on measurements of actual flow variations in existing systems. In projects being
submitted for federal construction grants, analysis of existing flows is required in any case
to identify "excessive" infiltration/ inflow. Flows are considered excessive if they can be
eliminated more cheaply than they can be treated. Projected flow variations from existing
systems should reflect elimination of excessive flows.
Flows tend to be less variable in larger systems, due chiefly to differing times of travel from
different sections and to damping effects of flow storage in large sewers. Widely varying
relations have been reported between peak-to-average or minimum-to-average flow ratios
and system size (i.e. average flow or tributary population) (1) (2) (3) (4). Care should be
taken in using any of these relations for estimating flow variations in new systems or system
additions. In terms of the factors which affect flow variations, applications should be
limited to systems similar to those for which the relation was originally developed.
Relations for which the basis is unclear should be disregarded.
3.2 Performance vs. Flow Variation
Variations in influent flow rate and characteristics affect performance of all suspended
solids removal processes to some degree. Relations between performance and hydraulic or
solids loadings are discussed for individual processes in succeeding chapters. Magnitude and
character of significant recycled flows resulting from specific processes are also indicated.
Relations between performance and loadings are frequently developed in pilot units run
under steady flow conditions, or from data from actual plants compiled without close
attention to short-term peaks. In using such relations for design decisions, care must be
taken to allow for the effects of short term flow variations on performance. Short term
would include any time span less than that for which performance requirements are stated.
Typically requirements are on a monthly average basis, often with a less stringent
requirement for the worst week or worst day within the month.
3-1
-------
Designs based on maximum 24-hour flow, with allowance for diurnal peaks, provide some
margin so that weekly or monthly requirements can be met even when other factors
affecting process performance are not optimum.
3.3 Flow Equalization
Equalization storage can be used to reduce diurnal variations in flow and in concentration
of SS or other wastewater characteristics. Storage may also be used to handle peaks caused
by direct inflow to the sewers during wet weather. Assuming that equivalent performance
can be obtained either by increasing the size of treatment facilities or by providing
equalizing basins, selection between these approaches can be based on their relative costs
and environmental impacts. In plants using processes involving large, short-term recycle
flows—such as for backwashing granular media filters—equalization is almost always
justified.
The EPA Process Design Manual for Upgrading Existing Wastewater Treatment Plants
provides a basis for design of equalization facilities to achieve any given degree of equaliza-
tion of either peak flows or peak flows and solid loadings (5). Material from the Design
Manual relevant to flow equalization only is also available in a separate publication (6).
3.4 References
1. Smith, R., and Eiler, R. G., Simulation of the Time-Dependent Performance of the Ac-
tivated Sludge Process Using the Digital Computer, U.S. EPA, National Environmental
Research Center, Cincinnati, Ohio (October, 1970).
2. Duttweiler, D. W. & Purcell, L. T., Character and Quantity of Wastewater from Small
Populations, Jour. WPCF, Vol. 34, pg. 63 (1962).
3. Boyle Engineering and Lowry and Associates, Master Plan Trunk Sewer Facilities for
County Sanitation District No. 3 of Orange County, California, (June, 1968).
4. Design and Construction of Sanitary and Storm Sewers, ASCE Manual of Engineering
Practice No. 37, WPCF Manual of Practice No. 9 (1970).
5. Process Design Manual for Upgrading Existing Wastewater Treatment Plants. U.S.
Environmental Protection Agency, Technology Transfer, Washington, D.C. 20460
(revised 1974).
6. Flow Equalization, Technology Transfer Seminar Publication, U.S. Environmental Pro-
tection Agency, Washington, D.C. 20460 (May 1974).
3-2
-------
CHAPTER 4
PRINCIPLES OF CHEMICAL TREATMENT
4.1 Introduction
Chemical coagulation and decollation are accomplished by a combination of physical and
chemical processes which thoroughly mix the chemicals with the wastewater and promote
' the aggregation of wastewater solids into particles large enough to be separated by sedimen-
tation, flotation, media filtration or straining. The strength of the aggregated particles de-
termines their limiting size and their resistance to shear in subsequent processes.
For particles in the colloidal and fine supracolloidal size ranges (<1 to 2 microns) natural
stabilizing forces (electrostatic repulsion, physical separation by absorbed water layers) pre-
dominate over the natural aggregating forces (van der Waals) and the natural mechanism
(Brownian movement) which tends to cause particle contact. Coagulation of these fine par-
ticles involves both destabilization and physical processes which disperse coagulants and in-
crease the opportunities for particle contact. Destabilization, the action of chemical coag-
ulants, is discussed in this chapter. Physical processes, including chemical mixing, floccula-
tion, and solids contact processes, are discussed in Chapter 6.
Chemical coagulants used in wastewater treatment are generally the same as those used in
potable water treatment and include: alum, ferric chloride, ferric sulfate, ferrous chloride,
ferrous sulfate and lime. The effectiveness of a particular coagulant varies in different appli-
cations, and in a given application each coagulant has both an optimum concentration and
an optimum pH range.
In addition io coagulants themselves, certain chemicals may be applied for pH or alkalinity
adjustment (lime, soda ash) or as flocculating agents (organic polymers). For full effective-
ness chemical coagulation requires initial rapid mixing (Chapter 6) to thoroughly disperse
the applied chemicals so that they can react with suspended and colloidal solids uniformly.
4.2 Destabilization Mechanisms
The destabilizing action of chemical coagulants in wastewater may involve any of the fol-
lowing mechanisms:
1. Electrostatic charge reduction by adsorption of counter ions
2. Inter-particle bridging by adsorption of specific chemical groups in polymer chains
3. Physical enmeshment of fine solids in gelatinous hydrolysis products of the coag-
ulants.
The significance of these mechanisms in design is considered briefly below. Extensive dis-
cussion of the mechanisms can be found in the literature (1) (2) (3) (4).
4-1
-------
4.2.1 Electrostatic Charge Reduction
Finely dispersed wastewater solids generally have a negative charge. Adsorption of cations
from metal salt coagulants (in the case of iron and aluminum from their hydrolysis prod-
ucts), or from cationic polymers can reduce or reverse this charge.
Where electrostatic charge reduction is a significant destabilization mechanism, care must
be taken not to overdose with coagulant. This can cause complete charge reversal with res-
tabilization of the oppositely charged coagulant-colloid complex.
4.2.2 Interparticle Bridging
When polymeric coagulants contain specific chemical groups which can interact with sites
on the surfaces of colloid particles, the polymer may adsorb to and serve as a bridge be-
tween the particles. Coagulation using polyelectrolytes of the same charge as the colloids or
non-ionic polymers depends on this mechanism. Restabilization may occur if excessive do-
sages of polymer are used. In this case all sites on the colloids may adsorb polymer mole-
cules without any bridging. Excessive mixing can also cause restabilization by fracture or
displacement of polymer chains.
4.2.3 Enmeshment in Preciptated Hydrolysis Products
Hydroxides of iron, aluminum or, at high pH, magnesium form gelatinous hydrolysis prod-
ucts which are extremely effective in enmeshing fine particles of other material. These
hydroxides are formed by reaction of metal salt coagulants with hydroxyl ions from the nat-
ural alkalinity in the water or from added alkaline chemicals such as lime or soda ash. Suf-
ficient natural magnesium is frequently present in wastewaters so that effective coagulation
is obtained merely by raising the pH with lime. Organic polymers do not form hydrolysis
products of significance in this mechanism. At a pH value lower than that required to pre-
cipitate magnesium, the precipitates produced by lime treatment are frequently ineffective
in enmeshing the colloidal matter in wastewater. The remedy for this condition generally in-
volved addition of low dosage of iron salts or polymers as coagulant aids both to destabilize
and to increase the probability of enmeshment of colloids.
Coagulants may also react with other constituents of the wastewater, particularly anions
such as phosphate and sulfate, forming hydroloysis products containing various mixtures of
ions. The chemistry of the reactions is extremely complex and highly dependent on pH and
alkalinity. The presence of high concentrations of these anions may require increased doses
of coagulants or pH adjustment to achieve effective removals of SS.
4.3 Selection of Chemical Coagulants
Design of chemical treatment facilities for SS removal must take into account: 1) the types
and quantities of chemicals to be applied as coagulants, coagulant aids and for pH control
and 2) the associated requirements for chemical handling and feeding (Chapter 5) and for
mixing and flocculation after chemical addition (Chapter 6). Reactions of specific coag-
4-2
-------
ulant chemicals are detailed in Chapter 5.
Selection of coagulants should be based on jar testing of the actual wastewater (Section 4.5)
to determine dosages and effectiveness, and on consideration of the cost and availability of
different coagulants. Where expected changes in waste characteristics or market conditions
may favor different coagulants at different times, chemical feed and handling should be set
up to permit a switchover. In developing a testing program general information on ex-
perience at other locations and on costs should be considered to aid in selection of processes
and coagulants to be tested.
Experience to date with improved SS removal from chemical coagulation has been almost
solely in systems designed to remove phosphorus. Guidelines for design and coagulant selec-
tion for such systems are available in another manual (5). Descriptive data and SS removal
performance for several existing phosphorus removal installations are summarized in Table
4-1.
Few cases have been reported involving chemical coagulation aimed at SS removal alone
without phosphorous removal requirements. Anionic polymers have been used to increase
SS removal in primary treatment at Rocky River, Ohio (6). Doses of 0.3 mg/1 reduced SS
from 107 mg/1 to 65 mg/1. Mogelnicki (7) reported use of anionic polymer at a dosage of 1
mg/1 to improve primary clarifier SS removal from 43 percent to 76 percent.
In discussing the favorable results sometimes obtained in polymer applications O'Melia (4)
warns that it can be a time-consuming task to find the specific conditions (pH, ionic
strength, polymer type, molecular weight, degree of hydrolysis, etc.) which will provide
economy and effectiveness.
Pilot work at Denver, Colorado (8) on coagulation of effluent from an activated sludge nit-
rification system showed substantial reductions in SS, turbidity, BOD and other pollution
parameters, using lime and alum doses well below those needed for effective phosphate re-
moval. Lime dosages of 100 mg/ 1 were sufficient to reduce SS to below 15 mg/1 after settl-
ing and 5 mg/ 1 after filtration. Phosphate reduction was less than 80 percent. Alum do-
sages of about 50 mg/ 1 were sufficient to reduce suspended solids and phosphorus concen-
trations to similar levels. In direct filtration of alum-coagulated nitrified effluent, SS were
reduced to less than 2 mg/ 1 with an alum dosage of 60 mg/ 1. Phosphate reductions at this
alum concentration were only about 65 percent (6-7 mg/ 1 residual). This latter practice is
accompanied by shorter filter runs due to significant increases in solids loading.
Calgon Corporation investigated the use of ferric chloride with polymer addition for a small
municipal wastewater treatment plant at Leetsdale, Pennsylvania (9). Ferric chloride do-
sages were less than those necessary for 80 percent phosphate removal. Dosages and SS re-
ductions are shown below:
4-3
-------
TABLE 4.1
SS REMOVAL PERFORMANCE FOR
CHEMICAL COAGULATION APPLICATIONS TO PHOSPHATE REMOVAL
SUSPENDED SOLIDS
LOCATION PROCESS
Lebanon, Ohio IPC
EPA, Blue Plains IPC
Plant, Wash-
ington, D.C.
Ely, Minn. Tertiary
S.LakeTahoe, Tertiary
California
Lebanon, Ohio Tertiary
Nassau County, Tertiary
New York
Salt Lake City, IPC
Utah
Leetsdale, Pa.
Key: FM - Flash Mix Unit
PLANT
SIZE
mgd
0.1
0.1
l.S
7.5
0.1
0.6
0.04-
0.1
0.05-
0.09
0.03-
0.18
0.6
AVERAGE
CHEMICAL FEED
mg/1
Lime ~ 250
Lime 460 ^
+FeCl3 sCb)
Lime 250-350 (a)
+Polymer .2(a)
+FeCl3 6d>)
Lime 400
275
Lime 220-270
Alum 20
Iron 34-41
Polymer 0-1.5
Alum 14
Polymer 0-0.25
Lime 270-586
FeCl3 100
+Polymer 0.5
BASIC
pH EQUIPMENT
9.5 1-stage SC
11.5(a) 2-stage
10.0(b) sc(a)
FM,FL,S(b)
ll.sCa) 2-stage
9.5(b) SC
11.3 FM.FL, S
10.5 SR
9.5 SC
SC
SC
SC
9.8- SC
11.0
6.5- S
.6.7
Inf.
mg/1
109
158
75
38
38
43.5
22.5
90
95
101
280
Settled
Eff.
mg/1
30
14.4
10
10
25
16.5
2.5
21.9
26.9
11.6
38
DATA OVERFLOW
PERIOD RATE COMMENTS
Months gpd/sq ft
45 1440 Acid pH adjustment.
Reference 19
6 500-1800 Two stages; with inter-
mediate recarbonation.
Reference 20
(a)
5 570 System designed for nearly
66o(b' complete (eff.SO.05 mg/1)
phosphorus removal.
Reference 21
21 400-600 Recarbonation
Reference 22
10 1440 Acid pH adjustment made.
Reference 23
72 860 Bulking 2-3 times a year.
Reference 24
2.5 360-1080 Reference 11
1 500-870
5 290-1800 H-S04; pH adjustment.
Primary Treatment, Cons-
tant Feed - Reference 9
Notes: Lime as Ca(OH)2
FL - Flocculator
SC - Solids Contact (Sludge Blanket)
SR - Solids Recirculation
S - Settling
IPC - Independent Physical Chemical
Alum as A1++*
Iron Salts as Fe++*
(a) First Stage
(b) Second Stage
Overflow rates are minimum and maximum where range appears.
-------
Dosage SS
Polymer
mg/1
74 0.5
50 0.6
37 0.08
Influent
mg/1
160
71
85
Effluent
mg/1
23
27
33
Reduction
Percent
86
62
61
Aluminum or iron salts tend to react with soluble phosphate preferentially so that substan-
tial phosphorus removal must be involved before organic colloids can be destabilized (10).
Required dosages will be affected by phosphorus content. Similarly lime treatment to a pH
at which coagulation is effective precipitates substantial phosphorus. Because chemical do-
sage and pH range for optimum SS removal may differ somewhat from those for optimum
phosphorus removal, coagulant requirements may be determined by the effluent criteria for
either pollutant, depending on wastewater characteristics and the choice of chemical.
4.3.1 Sludge Production
Chemical coagulation increases sludge production in sedimentation units due both to great-
er removal of influent suspended solids and to insoluble reaction products of the coag-
ulation itself. For phosphorus removal, data on sludge production and sludge characteristics
and sample procedures for estimating sludge quantities are presented elsewhere (5).
The weight of sludge solids can be estimated by calculation of the sum of the expected SS
removal and of the precipitation products expected from the coagulant dosages applied.
Usually jar tests can be employed to obtain the necessary information for this calculation.
4.3.2 pH Control and Alkalinity
The critical factor in the control of lime reactions is pH. The pH for optimum effectiveness
of lime coagulation, determined from jar testing and process operating experience, can be
used as a set point for a pH control of lime dosing.
Alum and iron salt coagulation are much less sensitive to pH. Testing can determine opti-
mum dosages for coagulation and whether natural alkalinity is adequate for the reactions
(see Ch. 5). If supplemental alkalinity is needed either regularly or on an intermittent basis
(e.g. during high wet weather flows) provisions should be included for feeding necessary
amounts of lime or soda ash.
4-5
-------
4.3.3 Points of Chemical Addition
In independent physical-chemical treatment or in phosphate removal in the primary clari-
fier ahead of biological treatment, chemicals are added to raw sewage. In tertiary treatment
for phosphate removal and SS reduction, they are added to secondary effluent. In both
cases, proper mixing and flocculation units are needed. For phosphate removal or improve-
ment of SS capture in biological secondary treatment, chemicals are often added directly to
aeration units or prior to secondary settling units, without separate mixing and flocculation.
In some phosphate removal applications coagulants have been added at multiple points, e.g.
prior to primary settling and as part of a secondary or tertiary treatment step.
4.3.4 Supplementary Coagulants
Addition of the hydrolyzing metal coagulants to wastewater often results in a small
slow-settling floe or precipitate of phosphorus. Additional treatment is required to produce
a water with low residual suspended solids. Polymeric coagulants have proved to be quite
beneficial in aggregating the precipitation products to a settleable size and increasing the
shear strength of the floe against hydraulic breakup (11). Data on particular applications
appear in Table 4-1.
4.4 Coagulation Control
Because coagulation represents a group of complex reactions, laboratory experimentation
is essential to establish and maintain the optimum coagulant dosage and to determine the
effects of important variables on the quality of coagulation of the wastewater under in-
vestigation. With alum and iron coagulants two procedures are generally followed for this
purpose: the jar test and measurement of zeta potential. Proper control of lime coagulation
may be maintained by measuring the pH or automatically titrating alkalinity after lime ad-
dition.
4.4.1 Jar Test
The single, most widely used test to determine coagulant dosage and other parameters is the
jar test. The equipment for this test and the directions for its proper performance have been
published (12) (13) (14) (15). The jar test attempts to simulate the full scale coagula-
tion-flocculation process and has remained the most common control test in the laboratory
since its introduction in 1918. Since the intent is to simulate an individual plant's condi-
tions, it is not surprising that procedures may vary but generally have certain common ele-
ments. The jar test apparatus consists of a series of sample containers, usually six, the con-
tents of which can be stirred by individual mechanically-operated stirrers. Wastewater to be
treated is placed in the containers and treatment chemicals are added while the contents are
being stirred. The range of conditions, for example, coagulant dosages and pH, are selected
to bracket the anticipated optima. After a short, 1 to 5 minute period of rapid stirring to en-
sure complete dispersion of coagulant, the stirring rate is decreased and flocculation is
allowed to continue for a variable period, 10 to 20 minutes or more, depending on the simu-
lation. The stirring is then stopped and the floes are allowed to settle for a selected time.
The supernatant is then analyzed for the desired parameters. With wastewater the usual
4-6
-------
analyses are for turbidity or suspended solids, pH, residual phosphorus and residual coag-
ulant.
If desired, a number of supernatant samples may be taken at intervals during the settling
period to permit construction of a set of settling curves which provide more information on
the settling characteristics of floe than a single sample taken after a fixed settling period. A
dynamic settling test may also be used in which the paddles are operated at 2 to 5 rpm dur-
ing the settling period. This type of operation more closely represents settling conditions in
a large horizontal basin with continuous flow.
It should be noted that simple jar tests cannot simulate the conditions in solids contact reac-
tors (Chapter 6) and may indicate somewhat higher coagulant dosages than are actually
necessary when using these units for coagulation.
Several six-position stirrers are available commercially for running jar tests; one from
Phipps and Bird, (Phipps and Bird, Inc., Richmond, Va.), another from Coffman In-
dustries, (Coffman Industries, Inc., Kansas City, Ka.), are shown in Fig. 4-1. Standard
laboratory mixers have also been used; however, it is difficult to obtain reproducible mixing
conditions using different pieces of equipment. Various types of containers, usually beakers
or jars, are used to hold the samples. Improved mixing may be obtained by adding stationa-
ry plates in the containers as described by Camp and Conklin (15). The Coffman stirrer has
an attachment which makes it possible to add coagulant to all containers simultaneously.
Good results, however, can be obtained by rapidly adding coagulant from a large graduated
pipette to each jar in sequence.
A simple apparatus, shown in Fig. 4-2, can be constructed from tubing, rubber stoppers
and small aquarium valves to permit rapid sampling of supernatant. The unit is placed next
to the sample jars at the beginning of the settling period with the curved stainless steel tubes_
dipping into the jars. At desired intervals the vent valve is covered with a finger, permitting
vacuum to draw samples into the small sample bottles. The needle valves are adjusted so
that supernatant is drawn into all the bottles at the same rate. When sufficient sample is ob-
tained, the vent is uncovered and the bottles are replaced with empties. The maximum
sampling rate is about once per minute.
Fig. 4-3 shows characteristic types of settling curves which may be obtained. Curve A in-
dicates a coagulation which produced a uniformly fine floe so small that at the end of 1 to 2
minutes settling, the supernatant had a turbidity equal to that of the starting water due, in
part, to the fine floe which resisted settling. Settling was slow and the final turbidity was not
satisfactory. Curve B represents the most common type of settling rate obtained. During the
first 5 minutes, the settling rate was practically a straight line on a semilog plot. Settling
was rapid and clarification was satisfactory. The coagulation represented by curve C shows
that a mixture of large rapid settling floe and small, slow-settling particles was obtained.
Settling was rapid for the first two minutes, but with little further clarification after that.
High residual turbidity may also have resulted from incomplete coagulation. Curve D rep-
resents the ultimate in coagulation. Practically all of the floe particles were so large and
dense that 97 percent settled within three minutes. Sedimentation was essentially complete
within that time since only 0.5 percent additional floe settled in the next 27 minutes. Final
4-7
-------
FIGURE 4-1
JAR TEST UNITS WITH MECHANICAL (TOP)
AND MAGNETIC (BOTTOM) STIRRERS
4-8
-------
TO VACUUM SOURCE
VENT
TWO-HOLE
STOPPER
SAMPLE
BOTTLE
SAMPLE
TUBE
U
MAIN CONTROL VALVE
PLUG
PLUG
REGULATING VALVES
U.S EPA Headquarters Library
Mai! code 3404T
1200 Pennsylvania Avenue NW
Washington, DC 20460
202-566-0556
FIGURE 4-2
SIX-POSITION SAMPLER
4-9
-------
100
SETTLING TIME-MIN
FIGURE 4-3
SETTLING CURVES FREQUENTLY" OBTAINED
4-10
-------
clarity of the supernatant was entirely satisfactory.
Measurement of turbidity provides the most rapid indication of the degree of solids removal
obtained. The recommended procedure for turbidity measurement by light scattering is giv-
en in the 13th edition of Standard Methods for Examination of Water and Wastewater;
however, other methods varying from simple visual evaluation to measurement of light
transmitted on a laboratory spectrophotometer can be used for purposes of comparison.
Measurement of residual suspended solids is the only procedure which gives the actual
weight concentration of solids remaining, but the procedure is too slow for purposes of pro-
cess control. Where the character of the solids does not vary widely, their concentration
generally correlates well with measured turbidity.
A typical jar test might be run as follows:
Wastewater samples are placed in containers and rapid mix is started at 100 rpm. Selected
dosages of coagulant covering the expected range of the optimum concentration are rapidly
added to the containers and mixed for approximately 1 minute. If a polymer is to be used as
a coagulant aid, it is usually added to each jar at or just before the end of the rapid mix. The
paddles are then slowed to 30 rpm and mixing continues for 20 minutes. The paddles are
then stopped and the sampling apparatus previously-described is placed in position. At
settling times of 1,3,5,10 and possibly 20 minutes samples of supernatant are drawn for tur-
bidity measurement. After the final turbidity sample is drawn, a larger volume of super-
natant may be decanted for more complete analysis. Results are plotted as in Fig. 4-4 for
judgment as to the desired coagulant dosage.
If additional alkalinity is required to hold the coagulation in the optimum pH range, this
should be added to the samples ahead of the coagulant unless automatic titrators are set up
for pH control.
Once an approximate optimum coagulant concentration has been determined, it may be de-
sirable to repeat the jar test using that optimum with varying quantities of added alkalinity
to give different pH values. Experience in coagulating a given wastewater provides the best
guide as to methods for controlling the process.
4.4.2 Zeta Potential
Measurement of particle charge is another procedure which may be useful for control of the
coagulation process (16) (17) (18). The total particle charge is distributed over two con-
centric layers of water surrounding the particle: an inner layer of water and ions which is
tightly bound to the particle and moves with it through the solution, and an outer layer
which is a part of the bulk water phase and moves independently of the particle. Charges of
these layers are not directly measureable, but the zeta potential, which is the residual charge
at the interface between the layer of bound water and the mobile water phase, can be deter-
mined indirectly with commercially-available instruments.
In the zeta potential measurement procedure, a sample of treated water containing floe is
placed in a special plastic cell under a microscope as shown in Fig. 4-5. Under the in-
4-11
-------
100
g
H
3
S
Q— 26 mg/1
I I
A— 28mg/l
SETTLING TIME—MIN
FIGURE 4-4
JAR TEST RESULTS
4-12
-------
FIGURE 4-5
ZETA POTENTIAL APPARATUS
4-13
-------
fluence of a voltage applied to electrodes at the ends of the cell, the charged particles will
migrate to the electrode having a polarity opposite that of the particle. The velocity of mi-
gration will be proportional to the particle charge and to the applied voltage. The particle
velocity can be calculated by observing the time it takes a particle to travel a given distance
across an ocular micrometer. The zeta potential can then be obtained from a chart which
combines the particle velocity with instrumental parameters. Detailed operating instruc-
tions are supplied with the instruments. Because of uncertainties in the constants relating
charge and particle mobility, many test results are reported directly in terms of particle mo-
bility.
To control the coagulation by zeta potential, samples of water while being mixed are dosed
with different concentrations of coagulant. Zeta potentials are then measured and recorded
for floe in each sample. The dosage which produces the desired zeta potential value is ap-
plied to the treatment plant. Zeta potentials of floe produced in the plant may also be mea-
sured as a means of control. The zeta potential value for optimum coagulation must be de-
termined for a given wastewater by actual correlation with jar tests or with plant perform-
ance as in Fig. 4-6. The control point is generally in the range of 0 to 10 millivolts. If
good correlations can be obtained between some zeta potential values and optimum plant
performance, then it is possible to make rapid measurements of particle charge to com-
pensate for major variations in wastewater composition due to storm flows or other causes.
Short term variations such as those due to sudden industrial waste dumps are still beyond
control with any present techniques because of the time lag between recognition of a prob-
lem with coagulation and adoption of a satisfactory change of coagulation conditions.
4-14
-------
+10
0
0
100
400
200 300
ALUM DOSAGE, (mg/1)
FIGURE 4-6
COAGULATION OF RAW SEWAGE WITH ALUM
500
4-15
-------
4.5 References
1. Stumm W., and Morgan, J. J., Chemical Aspects of Coagulation, Jour, AWWA, 54,
971 (1962).
2. Black, A. P., Basic Mechanism of Coagulation, Jour. AWWA, 52, 492 (1960).
3. O'Melia, C. R., A Review of the Coagulation Process, Public Works, 100, 87 (May
;1969).
4. O'Melia, C. R., Coagulation and Flocculation, Chapter in Physicochemical Pro-
cesses for Water Quality Control (Walter J. Weber, Jr.) John Wiley and Sons (1972).
5. Process Design Manual for Phosphorus Removal, U.S. EPA, Technology Transfer,
Washington, D.C. (revised 1974).
6. Rizzo, J. L. and Schade, R.E., Secondary Treatment -with Granular Activated Car-
bon, Water and Sewage Works, 116, 307, (August 1969).
7. Mogelnicki, S.; Experiences in Polymer Applications to Several Solids-Liquids Sepa-
ration Process, Proceedings, Tenth Sanitary Engineering Conference-Waste Disposal
from Water and Wastewater Treatment Processes, Univ. of Illinois (February 6-7,
1968).
8. Linstedt, K.iD. and Bennett, E. R. Evaluation of Treatment for Urban Wastewater
Reuse, U.S. EPA Office of Research and Monitoring, Publication U.S. EPA
R2-73-122 (July, 1973).
9. Bernardin, F. E., Jr., Kusnirak, R., Chemical Treatment For Municipal Wastewater,
WPCF Deeds and Data (March, 1974).
10. Tenney, M. W. and Stumm, W. Chemical Flocculation of Micro-organisms in Biolog-
ical Water Treatment, Jour. WPCF 37, 1370 (1965).
11. Burns, D. E. and Shell, G.L. A New Approach to Phosphorus Removal by Chemical
Treatment, Paper presented at 45th Annual WPCF Conference, Atlanta. Georgia(Oc-
tober 9, 1972).
12. Cohen, J. M., Improved Jar Test Procedure, Jour. AWWA, 49, 1425 (1957).
13. Black, A. P., Buswell, A. M., Eidsness, F. A., and Black, A. L., Review of the Jar
Test, Jour, AWWA, 39, 1414 (1957).
14. Black, A. P., and Harris, R. J., New Dimensions for the Old Jar Test, Water &
Wastes Engrg., 6. 49 (Dec. 1969).
4-16
-------
15. Camp, T. R., and Conklin, G. F., Towards a RationalJar Test for Coagulation, Jour.
AWWA, 84. 325(1970).
16. Black, A. P. & Chen, C., Electrophoretic Studies of Coagulation and Flocculation
of River Sediment Suspension with Aluminum Sulfate, Jour. AWWA, 57, 354 (1965).
17. Riddick, T. M., Role of Zeta Potential in Coagulation Involving Hydrous Oxides,
TAPPI, 47, 17A(1964).
18. Riddick, T. M., Control of Colloid Stability Through Zeta Potential, Vol. 1,
. Zeta-Meter, Inc., 1720 First Avenue, New York, New York 10028.
19. Villiers, R. V., Berg, E. L. Brunner, C. A. and Masse, A. N., Municipal Wastewater
Treatment, Paper presented at 45th Annual WPCF Conference, Atlanta. Georgia
(October 9, 1972).
20. Bishop, D. F., O'Farrell, T. P., and Stamberg, J. B.; Physical-Chemical Treatment of
Municipal Wastewater, Jour. WPCF 44, 361 (March 1972).
21. Brice, R. M., Shagawa Lake Project, Ely, Minnesota, Personal Communication
(August 1973).
22. Evans, Wilson, Gulp, Suhr and Mover; A Summary of Plant Scale Advanced Waste
Treatment Research at South Lake Tahoe; work for partial fulfillment of an
U.S. EPA demonstration grant WPRD-52-01-67.
23. Berg. E. L., Brunner, C. A. and William, R. T.; Single-Stage Lime Clarification, Wa-
ter and Waste Engineering Vol. 7, No. 3, pg. 42 (March 1970).
24. Oliva, J. A., Department of Public Health, County of Nassau, Personal Commu-
nication (March 1973).
4-17
-------
CHAPTER 5
STORAGE AND FEEDING OF CHEMICALS
5.1 General
This chapter surveys the chemicals most commonly used for suspended solids removal, with
respect to their properties, availability, storage, transport, reactions and feeding. All
chemical costs quoted in this chapter were obtained from the latest issues of
"Chemical Marketing Reporter" (Schnell Publishing Co., Inc., New York, N. Y.)
available during preparation of this manual. Wide ranges in bagging costs primarily
reflect bag sizes that may be ordered. All chemical costs presented are for guidance
only and are subject to significant variations due to time and current market con-
ditions. Actually costs for the chemicals being considered should be carefully
checked prior to selection.
5.2 Aluminum Compounds
The principal aluminum compounds that are commercially available and suitable for
suspended solids removal are dry and liquid alum. Sodium aluminate has been used in
activated sludge plants, but for phosphorus removal, and its applicability for suspended
solids removal is limited.
5.2.1 Dry Alum
5.2.1.1 Properties and Availability
The commercial dry alum most often used in wastewater treatment is known as "filter
alum." and has the approximate chemical formula Al2(SO4)3*14H2Oand a molecular
weight of about 600. Alum, is white to cream in color and a 1 percent solution has a pH of
about 3.5. The commercially available grades of alum and their corresponding bulk den-
sities and angles of repose are:
GRADE ANGLE OF REPOSE BULK DENSITY
Ib./cubic feet
Lump 62 to 68
Ground 43 60 to 71
Rice 38 57 to 71
Powdered 65 38 to 45
Each of these grades has a minimum alumium content of 27 percent, expressed as AlzOa,
and maximum Fe20a and soluble contents of 0.75 and 0.5 percent, respectively. Visosity
and solution crystallation temperatures are included in the subsequent section on liquid
alum.
Since dry alum is only partially hydrated, it is slightly hygroscopic. However, it is relatively
stable when stored under the extremes of temperature and humidity encountered in the
United States.
5-1
-------
The solubility of commercial dry alum at various temperatures is as follows:
Temperature'
32
50
68
86
104
Solubility
Ib/gal
6.03
6.56
7.28
8.45
10.16
Dry alum is not corrosive unless it absorbs moisture from the air, such as during prolonged
exposure to humid atmospheres. Therefore, precautions should be taken to ensure that the
storage space is free of moisture.
Alum is shipped in 100 Ib bags, drums, or in bulk (minimum of 40,000 Ib) by truck or rail.
Bag shipments may be ordered on wood pallets if desired. Locations of the major
production plants are listed in Table 5-1. At present, the price range for dry alum in bulk
quantities is $58 to $63/ton. F.O.B. the point of manufacture. Freight costs to the point of
usage must be added to this. Bagging adds approximately $4 to 5/ton to the cost.
TABLE 5-,l
PARTIAL LIST OF ALUM MANUFACTURING PLANTS
Location
Manufacturer
Form of Alum Available
ALABAMA
Coosa Pines
Demopolis
Mobile
Naheola
ARKANSAS
Pine Bluff
CALIFORNIA
Bay Point (San Francisco)
El Segundo (Los Angeles)
Richmond (San Francisco)
Vernon (Los Angeles)
COLORADO
Denver
American Cyanamid
American Cyanamid
American Cyanamid
Stauffer
Allied
Allied
Allied
Stauffer
Stauffer
Allied
Liquid
Liquid
Liquid and Dry
Liquid
Liquid
Liquid and Dry
Liquid
Liquid
Liquid
Liquid and Dry
5-2
-------
Location
FLORIDA
Fernandina Beach
Jacksonville
Port St. Joe
GEORGIA
Atlanta (2 plants)
Augusta
Cedar Springs
Macon
Savannah
ILLINOIS
E. St. Louis
Joliet
LOUISIANA
Bastrop
Baton Rouge
Monroe
New Orleans
Springhill
MAINE
Searsport
MARYLAND
Baltimore
MASSACHUSETTS
Adams
Salem
TABLE 5-1 (continued)
Manufacturer
Tennessee Corp.
Allied
Allied
Burris, Allied
Tennessee Corp.
Tennessee Corp.
Allied
Allied
Allied
American Cyanamid
Stauffer
Stauffer
Allied
Allied
Stauffer
Northern
Olin
Holland
Hamblet & Hayes
Form of Alum
Available
Liquid
Liquid
Liquid
Liquid and Dry
Liquid and Dry
Liquid
Liquid
Liquid
Liquid and Dry
Liquid and Dry
Liquid
Liquid
Liquid
Liquid and Dry
Liquid
Liquid and Dry
Dry
Liquid
Liquid and Dry
MICHIGAN
Detroit
Escanaba
Kalamazoo (2 plants)
MINNESOTA
Cloquet
Pine Bend
Allied
American Cyanamid
Allied, American Cyanamid
American Cyanamid
North Star
Liquid
Liquid
Liquid
Liquid
Liquid and Dry
5-3
-------
TABLE 5-1 (continued)
Location
MISSISSIPPI
Monticello
Vicksburg
NEW JERSEY
Newark
Warners
NORTH CAROLINA
Acme
Plymouth
OHIO
Chill icothe
Cleveland
Hamilton
Middletown
OREGON
North Portland
PENNSYLVANIA
Johnsonburg
Marcus Hook
Newell
SOUTH CAROLINA
Catawba
Georgetown
TENNESSEE
Chattanooga
Counce
Springfield
Manufacturer
American Cyanamid
Allied
Essex
American Cyanamid
Wright
American Cyanamid
Allied
Allied
American Cyanamid
Allied
Stauffer
Allied
Allied
Allied
Burris
American Cyanamid
American Cyanamid
Stauffer
Burris
Form of Alum
Available
Liquid
Liquid
Liquid and Dry
Liquid and Dry
Liquid
Liquid
Liquid
Liquid and Dry
Liquid and Dry
Liquid
Liquid and Dry
Liquid
Liquid and Dry
Liquid
Liquid
Liquid and Dry
Liquid and Dry
Liquid
Liquid
TEXAS
Houston (2 plants)
Stauffer, Ethyl
Liquid and Dry
5-4
-------
TABLE 5-1 (continued)
VIRGINIA
Covington
Hope well
Norfolk
WASHINGTON
Kennewick
Tacoma (2 plants)
Vancouver
WISCONSIN
Menasha
Wisconsin Rapids
Manufacturers and Addresses
Allied
Allied
Howerton Gowen
Allied
Stauffer, Allied
Allied
Allied
Allied
Liquid
Liquid
Liquid
Liquid
Liquid
Liquid and Dry
Liquid
Liquid
Allied Chemical Corporation
Industrial Chemicals Division
P.O. Box 1139R
Morristown, New Jersey 07960
American Cyanamid Company
Ind. Chem. Div.
P.O. Box 66189
Chicago, Illinois 60666
Burris Chemical Company
Charleston, South Carolina
Essex Chemical Corporation
1402 Broad Street
Clifton, New Jersey 07015
Ethyl Corporation
Houston, Texas
Hamblet & Hayes Company
Colonial Road
Salem, Massachusetts 01970
Holland Chemical Company
Adams, Massachusetts 01220
Howerton Gowen Company, Inc.
Norfolk, Virginia
Northern Chemical Industries, Inc.
Searsport, Maine 04974
North Star Chemicals Inc.
P.O. Box 28-T
South St. Paul, Minnesota
Olin Corporation,
Chemicals Division
745 Fifth Avenue
New, York, New York 10022
Stauffer Chemical Company
299 Park Avenue
New York, New York 10017
Tennessee Corporation
Cities Service Company
Industrial Chemicals Division
P.O. Box 50360
Atlanta, Georgia
Wright Chemical Co.
Acme, North Carolina
5-5
-------
5.2.1.2 General Design Considerations
Ground and rice alum are the grades most commonly used by utilities because of their
superior flow characteristics. These grades have less tendency to lump or arch in storage
and therefore provide more consistent feeding qualities. Hopper agitation is seldom
required with these grades, and in fact may be detrimental to feeding because of the
possibility of packing the bin.
Alum dust is present in the ground grade and will cause minor irritation of the eyes and
nose on breathing. A respirator may be worn for protection against alum dust. Gloves may
be worn to protect the hands. Because of minor irritation in handling and the possibility of
alum dust causing rusting of adjacent machinery, dust removal equipment is desirable.
Alum dust should be thoroughly flushed from the eyes immediately and washed from the
skin with water.
5.2.1.3 Storage
A typical storage and feeding system for dry alum is shown in Figure 5-1. Bulk alum can be
stored in mild steel or concrete bins with dust collector vents located in, above, or adjacent
to the equipment room. Recommended storage capacity is about 30 days. Dry alum in bulk
can be transferred with screw conveyors, pneumatic conveyors, or bucket elevators made of
mild steel. Pneumatic conveyor elbows should have a reinforced backing as the alum can
contain abrasive impurities.
Bags and drums of alum should be stored in a dry location to avoid caking. Bag or drum
loaded hoppers should have a nominal storage capacity for eight hours at the nominal
maximum feed rate so that personnel are not required to charge the hopper more than once
per shift. Converging hopper sections should have a minimum slope of 60 degree to prevent
arching.
Bulk storage hoppers should terminate at a bin gate so that the feeding equipment may be
isolated for servicing. The bin gate should be followed by a flexible connection, and
transition hopper chute or hopper which acts as a conditioning chamber over the feeder.
5.2.1.4 Feeding Equipment
The feed system includes all of the components required for the proper preparation of the
chemical solution. Capacities and assemblies should be selected to fulfill individual system
requirements. Three basic types of chemical feed equipment are used: volumetric, belt
gravimetric, and loss-in-weight gravimetric. Volumetric feeders are usually used where
initial low cost and usually lower capacities are the basis of selection. Volumetric feeder
mechanisms are usually exposed to the corrosive dissolving chamber vapors which can
cause corrosion of discharge areas. Manufacturers usually control this problem by use of an
electric heater to keep the feeder housing dry or by using plastic components in the exposed
areas.
5-6
-------
,-DUST COLLECTOR
XA-FILL PIPE (PNEUMATIC)
BULK STORAGE
BJN
DAY HOPPER
FOR DRY CHEMICAL
FROM BAGS OR DRUMS
BIN .GATE
FLEXIBLE
CONNECTION
ALTERNATE SUPPLIES DEPENDING
ON STORAGE
DUST COLLECTOR
XBAG FILL
-SCREEN
WITH BREAKER
~\
SCALE OR SAMPLE CHUTE
DUST AND VAPOR REMOVER
WATER
f"1-!
DRAIN
SOLENOID VALVE
CONTROL
VALVE -"ROTAMETER
PRESSURE REDUCING
VALVE
GRAVITY TO
APPLICATION
PUMP
TO APPLICATION
FIGURE 5-1
TYPICAL DRY FEED SYSTEM
5-7
-------
Volumetric dry feeders presently in general use are of the screw type. Two designs of screw
feed mechanism are available. Both allow even withdrawal across the bottom of the feeder
hopper to prevent hopper dead zones. One screw design is the variable pitch type with the
pitch'expanding unevenly to the discharge point. The second screw design is the constant pitch.
type expanding evenly to the discharge point. This type of screw design is the constant
pitch-reciprocating type. This type has each half of the screw turned in opposite directions
so that the turning and reciprocating motion alternately fills one half of the screw while the
other half of the screw is discharging. The variable pitch screw has one point of discharge,
while the constant pitch-reciprocating screw has two points of discharge, one at each end of
the screw. The accuracy of volumetric feeders is influenced by the character of the material
being fed and ranges between ±;l;percenf for free-flowing materials and =fc 7 percent for
cohesive materials. This accuracy is volumetric and should not be related to accuracy by
weight (gravimetric).
Where the greatest accuracy and the most economical use of chemical is desired, the
loss-in-weight type feeder should be selected. This feeder is limited to the low and
intermediate \feed rates up to a maximum rate of approximately 4,000 lb/ hr. The
loss-in-weight type feeder consists of a material hopper and feeding mechanism mounted on
enclosed scales. The feed rate controller retracts the scale poise weight to deliver the dry
chemical at the desired rate. The feeding mechanism must feed at this rate to maintain the
balance of the scale. Any unbalance of the scale beam causes a corrective change in the
output of the feeding mechanism. Continuous comparison of actual hopper weight with set
hopper weight prevents cumulative errors. Accuracy of the loss-in-weight feeder is ± 1 per-
cent by weight of the set rate.
Belt type gravimetric feeders span the capacity ranges of volumetric and loss-in-weight
' feeders and can usually be sized for all applications encountered in wastewater treatment
applications. Initial expense is greater than for the volumetric feeder and slightly less than
for the loss-in-weight feeder. Belt type gravimetric feeders consist of a basic belt feeder
incorporating a weighing and control system. Feed rates can be varied by changing either
the weight per foot of belt, or the belt speed, or both. Controllers in general use are
mechanical, pneumatic, electric, and mechanical-vibrating. Accuracy specified for belt type
gravimetric feeders should be within ± 1 percent of set rate. Materials of construction of
feed equipment normally include mild steel hoppers, stainless steel mechanism components,
and rubber surfaced feed belts.
Because alum solution is corrosive, dissolving or solution chambers should be constructed
of type 316 stainless steel, fiberglass reinforced plastic (FRP), or plastics. Dissolvers should
be sized for preparation of the desired solution strength. The solution strength usually
recommended is 0.5 lb of alum to 1 gal. of water, or a 6 percent solution. The dissolving
chamber is designed for a minimum detention time of 5 minutes at the maximum feed rate.
Because excessive dilution may be detrimental to coagulation, eductors, or float valves that
would ordinarily be used ahead of centrifugal pumps, are not recommended. Dissolvers
should be equipped with water meters and mechanical mixers so that the water to alum ra-
tio may be properly established and controlled.
5-8
-------
5.2.1.5 Piping and Accessories
FRP, plastics (polyvinyl chloride, polyethylene, polypropylene, and other similar
materials), • and rubber are general use and are recommended for alum solutions. Care
must be taken to provide adequate support for these piping systems, with close attention
given to spans between supports so that objectionable deflection will not be experienced.
The alum solution should be injected into a zone of rapid mixing or turbulent flow.
Solution flow by gravity to the point of discharge is desirable. When gravity flow is not
possible, transfer components should be selected that require little or no dilution. When
metering pumps or proportioning weir tanks are used, return of excess flow to a holding
tank should be considered. Metering pumps are discussed further in the section on liquid
alum.
Valves used in solution lines should be plastic, type 316 stainless steel or rubber-lined iron or
steel.
5.2.1.6 Pacing and Control
Standard instrument control and pacing signals are generally acceptable for common feeder
system operation. Volumetric and gravimetric feeders are usually adaptable to operation
from any standard instrument signals.
When solution must be pumped, consideration should be given to use of holding tanks
between the dry feed system and feed pumps, and the solution water supply should be
controlled to prevent excessive dilution. The dry feeders may be started and stopped by tank
level probes. Variable control metering pumps can then transfer the alum stock solution to
the point of application without further dilution.
Means should be provided for calibration of the chemical feeders. Volumetric feeders may
be mounted on platform scales. Belt feeders should include a sample shute and box to catch
samples for checking actual delivery with set delivery.
Gravimetric feeders are usually furnished with totalizers only. Remote instrumentation is
frequently used with gravimetric equipment, but seldom used with volumetric equipment.
5.2.2. Liquid Alum
5.2.2.1 Properties and Availability
Liquid alum is shipped in rubber-lined or stainless steel, insulated tank cars or trucks. Alum
shipped during the winter is heated prior to shipment so that crystallization will not occur
during transit. Liquid Alum is shipped at a solution strength of about 8.3 percent as AhOa
or about 49 percent as Al2(SO4>3 • 14H2O. The latter solution weighs about 11 Ib/gal at
60 °F and contains about 5.4 Ib dry aium (17 percent AhOa) per gal of liquid. This solution
will begin to crystallize at 30 °F and freezes at about 18 °F.
5-9
-------
Crystallization temperatures of various solution strengths are shown in Figure 5-2.
The viscosity of various alum solutions is given in Figure 5-3.
Tank truck lots of 3,000 to 5,000 gal are available. Tank car lots are available in quantities
. of 7,000 to 18,000 gal. Production locations of liquid alum are listed in Table 5-1. The cur-
rent price range of liquid alum on an equivalent dry alum (17 percent AbOa) basis is about
$45 to $507 ton, F.O:B. the point of manufacture. Liquid alum will generally be more eco-
nomical than dry alum if the point of use is within a 50 to 100 mile radius of the manufac-
turing plant.
5.2.2.2 General Design Considerations
Bulk unloading facilities usually must be provided at the treatment plant. Rail cars are
constructed for .top unloading and therefore require an air supply system and flexible
connectors to pneumatically displace the alum from the car. U.S. Department of
Transportation regulations concerning chemical tank car unloading should be observed.
Tank truck unloading is usually accomplished by gravity or by a truck mounted pump.
Established practice in the treatment field has been to dilute liquid alum prior to application.
However, recent studies have shown that feeding undiluted liquid alum results in better
coagulation and settling. This is reportedly due to prevention of hydrolysis of the alum.
No particular industrial hazards are encountered in handling liquid alum. However, a face
shield and gloves should be worn around leaking equipment. The eyes or skin should be
flushed and washed upon contact with liquid alum. Liquid alum becomes very slick upon
evaporation and therefore spillage should be avoided.
5.2.2.3 Storage
Liquid alum is stored without dilution at the shipping concentration. Storage tanks may be
open if indoors but must be closed and vented if outdoors. Outdoor tanks should also be
heated, if necessary, to keep the temperature above 45°F to prevent crystallization. Storage
tanks should be constructed of type 316 stainless steel; FRP; steel lined with rubber,
polyvinyl chloride, or lead. Liquid alum can be stored indefinitely without deterioration.
Storage tanks should be sized according to maximum feed rate, shipping time required, and
quantity of shipment. Tanks should generally be sized for P/2 times the quantity of
shipments. A 10-day to 2-week supply should be provided to allow for unforeseen shipping
delays.
5-10
-------
10 19 20 25 30 35 40
DEGREE BAUME AT 60°FARENHEIT
45 50
FIGURE 5-2
CRYSTALLIZATION TEMPERATURES OF ALUM SOLUTIONS
(Courtesy of American Cyanamid Co.)
5-11
-------
£
cs>
60
40
20
10
8.0
6.0
4.0
2.0
1.0
0.8
0.6
O.U
0.2
0.1
I I
I I I I I I I
I I I
30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220
TEMPERATURE,°F
FIGURE 5-3
VISCOSITY OF ALUM SOLUTIONS
(Courtesy of Allied Chemical Co.)
5-12
-------
5.2.2.4 Feeding Equipment
Various types of gravity or pressure feeding and metering units are available. Figures 5-4
and 5-5 illustrate commonly used feed systems. The rotodip-type feeder or rotameter is
often used for gravity feed and the metering pump for pressure feed systems.
The pressure or head available at the point of application frequently determines the feeding
system to be used. The rotodip feeder can be supplied from overhead storage by gravity
with the use of an internal level control valve, as shown by Figure 5-4. It may also be
supplied by a centrifugal pump. The latter arrangement requires an excess flow return line
to the storage tank, as shown by Figure 5-5. Centrifugal pumps should be direct-connected
but not close-coupled because of possible leakage into the motor, and should be constructed
of type 316 stainless steel, FRP, and plastics.
Metering pumps, currently available, allow a wide range of capacity compared with the
rotodip and rotameter systems. Hydraulic diaphragm type pumps are preferable to other
type pumps and should be protected with an internal or external relief valve. A back
pressure valve is usually required in the pump discharge to provide efficient check valve
action. Materials of construction for feeding equipment should be as recommended by the
manufacturer for the service, but depending on the type of system, will generally include
type 316 stainless steel, FRP, plastics, and rubber.
5.2.2.5 Piping and Accessories
Piping systems for alum should be FRP, plastics (subject to temperature limits), type 316
stainless steel, or lead. Piping and valves used for alum solutions are also discussed in the
preceding section on dry alum.
5.2.2.6 Pacing and Control
The feeding systems described above are volumetric, and the feeders generally available can
be adapted to receive standard instrument pacing signals. The signals can be used to vary
motor speed, variable-speed transmission setting, stroke speed and stroke length where
applicable. A totalizer is usually furnished with a rotodip-type feeder, and remote
instruments are available. Instrumentation is rarely used with rotameters and metering
pumps.
5.2.3 Reactions of Aluminum Sulfate
Reactions between alum and the normal constituents of wastewaters are influenced by
many factors, hence it is impossible to predict accurately the amount of alum that will react
with a given amount of alkalinity, lime or soda ash which may have been added to the
wastewater. Theoretical reactions can be written which will serve as a general guide, but, in
general, the optimum dosage in'each case must be determined by laboratory jar tests.
5-13
-------
OVERHEAD
STORAGE
TANK
FLOAT
VALVE
tn
V)
a.
^
o
CJ
ROTAMETER
ROTODIP-TYPE FEEDER
GRAVITY FEED GRAVITY FEED
METERING PUMP
GRAVITY FEED
FIGURE 5-4
ALTERNATIVE LIQUID FEED SYSTEMS
FOR OVERHEAD STORAGE
PRESSURE FEED
-ROTODIP-TYPE FEEDER
^CONTROL VALVE
^ROTAMETER
\,
^~
r
YTVT^7
^-r |
1
1
f
VITY FEED 'i
TRANSFER PUMF
\
"h
xREC
••>
IRCULATION
GROUND
STORAGE
TANK
«
L
^ -
•< Q- S
s§ 2 ;
LL. Q. 0
_ o
ac. o z
^ a «
° UJ
^ * • 1
H * '
1 o
I s
<- >
- s
^ = PRESSURE
e a FE|
oe
Q.
L o
^ 2
^)
FIGURE 5-5
ALTERNATIVE LIQUID FEED SYSTEMS
FOR GROUND STORAGE
5-14
-------
The simplest case is the reaction of Al + with OH ions made available by the ionization of
water or by the alkalinity of the water.
Solution of alum in water produces:
:A12(SO4)3^±2A13+' +3(SO4)2~
Hydroxyl ions become available from ionization of water:
The aluminum ions (Al|3^) then react:
2A13+ +6OH':^±2 A1(OH)3
Consumption of hydroxyl ions will result in a decrease in the alkalinity. Where the
alkalinity of the wastewater is inadequate for the alum dosage, the pH must be increased by
the addition of hydrated lime, soda ash or caustic soda. The reactions of alum with the
common alkaline reagents are shown in Table 5-2. While these reactions are an
oversimplification of what actually takes place, they do serve to indicate orders of
magnitude and some by-products of alum treatment.
TABLE 5-2
REACTIONS OF ALUMINUM SULFATE
A12 (SO4)3 + 3 Ca (HCO3)2— ^2 Al (OH)3J + 3 CaSO4 + 6 CO2f
Ak (SO4)3 + 3 Na2CO3 + 3 H2O— ^2 AL (OH) 3\ + 3 CO2f
Ah (SO4)3 + 3 Ca (OH)2— *-2 Al (OH)3| + 3 CaSO4
In terms of quantities, the reactions in Table 5-2 can be expressed as follows:
f •
1 mg/ I of alum reacts with:
0.50 mgV 1 alkalinity, expressed as CaCOs
0.39 mg/ 1 95 percent hydrated lime as Ca (OH)2
0.54 mg/ 1 soda ash as Na2COs
These approximate amounts of alkali when added to wastewater will maintain the alkalinity
of the water unchanged when 1 mg/ 1 of alum is added. For example, if no alkalinity is
added, 1 mg/ 1 of alum will reduce the alkalinity of 0.50 mg/ 1 as CaCOs but alkalinity can
be maintained unchanged if 0.39 mg/ 1 of hydrated lime is added. This lowering of natural
alkalinity is desirable in many cases to attain the pH range for optimum coagulation.
For each mg/ 1 of alum dosage, the sulfate (SO4) content of the water will be increased
approximately 0.49 mg/ 1 and the CO2 content of the water will be increased approximately
0.44 mg/1.
5-15
-------
5.3 Iron Compounds
Iron compounds have pH coagulation ranges and floe characteristics similar to aluminum
sulfate. The cost of iron compounds may often be less than the cost of alum. However, the
iron compounds are generally corrosive and often present difficulties in dissolving, and their
use may result in high soluble iron concentrations in process effluents.
5.3.1 Liquid Ferric Chloride
5.3.1.1 Properties and Availability
Liquid ferric chloride is a corrosive, dark brown oily-appearing solution having a weight as
shipped and stored of 11.2 to 12.4 Ib/gal (35 to 45 percent FeCb) (1). The ferric chloride
content of these solutions, as FeCh, is 3.95 to 5.58 Ib/gal. Shipping concentrations vary
from summer to winter due to the relatively high crystallization temperature of the more
concentrated solutions as shown by Figure 5-6. The pH of a 1 percent solution is 2.0.
The molecular weight of ferric chloride is 162.22. Viscosities of ferric chloride solutions at
various temperatures are presented in Figure 5-7.
Liquid ferric chloride is shipped in 3,000 to 4,000 gal bulk truckload lots, in 4,000 to 10,000
gal bulk carload lots, and in 5 and 13 gal carboys. Liquid ferric chloride is produced at the
following locations (2):
Dow Chemical Co.
Midland, Michigan
Pennwalt Corp.
Philadelphia, Pa. (Plant at Wyandotte, Mich.)
The current price of liquid ferric chloride in bulk quantities is about $0.04 to $0.05/lb (as
FeCb), F.O.B. the point of manufacture.
Tank trucks and cars are normally unloaded pneumatically, and operating procedures must
be closely followed to avoid spills and accidents. The safety vent cap and assembly (painted
red) should be removed prior to opening the unloading connection to depressurize the tank
car or truck, prior to unloading.
5.3.1.2 General Design Considerations
Ferric chloride solutions are corrosive to many common materials and cause stains which
are difficult to remove. Areas which are subject to staining should be protected with
resistant paint or rubber mats.
5-16
-------
70
60
50
40
30
o"- 20
ul
| 10
oe
LU 0
I
--10
-20
-30
-40
-50
I I I I I I I I I I I I I I I I
A Agitated Solutions may Degin to develop crystals
below th i s 1i ne
B Unagitated Solutions Degin to develop crystals when
the Dulk solution temperature drops to about this line.
Ice crystals form Delow 33% FeCl, &
Crystals form aoove 3$ Fed,. d
-60
I I I I I I I I
I I I I I I
10
18
22
26
30
Fed
34
38
42
FIGURE 5-6
FREEZING POINT CURVES FOR
COMMERCIAL FERRIC CHLORIDE SOLUTIONS
(Courtesy of Dow Chemical Co.)
46
so
5-17
-------
100
80
60
50
40
oo 30
LU
ae
£
00 20
I—
S 15
o
CO
10
8
6
5
4
I I I I I I I I I I I
(Aosolute Viscosity)=(Kinematic Viscosity)(Density)
Centipoises = Cent i stokes x Q"1
cc
10
20
40
50
% Fed,
FIGURES-?
VISCOSITY VS COMPOSITION OF FERRIC
CHLORIDE SOLUTIONS AT VARIOUS
TEMPERATURES
(Courtesy of Dow Chemical Co.)
5-18
-------
Normal precautions should be employed when cleaning ferric chloride handling equipment.
Workmen should wear rubber gloves, rubber apron, and goggles or a face shield. If ferric
chloride comes in contact with the eyes or skin, flush with copious quantities of running
water and call a physician. If ferric chloride is ingested, induce vomiting and call a
physician.
5.3.1.3 Storage
Ferric chloride solution can be stored as shipped. Storage tanks should have a free vent or
vacuum relief valve. Tanks may be constructed of FRP, rubber lined steel, of plastic lined
steel. Resin-impregnated carbon or graphite are also suitable materials for storage
containers.
It may be necessary in most instances to house liquid ferric chloride tanks in heated areas
or provide tank heaters or insulation to prevent crystallization. Ferric chloride can be stored
for long periods of time without deterioration. The total storage capacity should be 1 l/2
times the largest anticipated shipment, and should provide at least a 10-day to 2-week
supply of the chemical at the design average dosage.
5.3.1.4 Feeding Equipment
Feeding equipment and systems described for liquid alum generally apply to ferric chloride
except for materials of construction and the use of glass tube rotameters.
It may not be desirable to dilute the ferric chloride solution from its shipping concentration
to a weaker feed solution because of possible hydrolysis. Ferric chloride solutions may be
transferred from underground storage to day tanks with impervious graphite or rubber
lined self-priming centrifugal pumps having teflon rotary and stationary seals. Because of the
tendency for liquid ferric chloride to stain or deposit, glass-tube rotameters should not be
used for metering this solution. Rotodip feeders and diaphragm metering pumps are often
used for ferric chloride, and should be constructed of materials such as rubber-lined steel
and plastics.
5.3.1.5 Piping and Accessories
Materials for piping and transporting ferric chloride should be rubber or Saran-lined steel,
hard rubber, FRP, or plastics. Valving should consist of rubber or resin-lined diaphragm
valves, Saran-lined valves with teflon diaphragms, rubber-sleeved pinch-type valves, or
plastic ball valves. Gasket material for large openings such as manholes in storage tanks
should be soft rubber; all other gaskets should be graphite-impregnated blue asbestos,
teflon, or vinyl.
5.3.1.6 Pacing and Control
System pacing and control requirements are similar to those discussed previously for liquid
alum.
5-19
-------
5.3.2 Ferrous Chloride (Waste Pickle Liquor)
5.3.2.1 Properties and Availability
Ferrous chloride, FeCh, as a liquid is available in the form of waste pickle liquor from steel
processing. The liquor weighs between 9.9 and 10.4 Ib/gal and contains 20 to 25 percent
FeCb or about 10 percent available Fe2+. A 22 percent solution of FeCb will crystallize at
a temperature of-4 °F. The molecular weight of FeCb is 126.76. Free acid in waste pickle
liquor can vary from 1 to 10 percent and usually averages about 1.5 to 2.0 percent. Ferrous
chloride is slightly less corrosive than ferric chloride.
Waste pickle liquor is available in 4,000 gal truckload lots and a variety of carload lots. In
most instances the availability of waste pickle liquor will depend on the proximity to steel
processing plants. Dow Chemical Company produces a waste pickle liquor, having an FeCb
content of about 22 percent at a price of $0.04/lb of FeCh in bulk car or truckload quan-
tities. F.O.B. Midland, Michigan.
5.3.2.2 General Design Considerations
Since ferrous chloride or waste pickle liquor may not be available on a continuous basis,
storage and feeding equipment should be suitable for handling ferric chloride. Therefore, the
ferric chloride section should be referred to for storage and handling details.
5.3.3 Ferric Sulfate
5.3.3.1 Properties and Availability
Ferric sulfate is marketed as dry, partially-hydrated granules with the formula Fe2(SO4)3 •
X HaO, where X is approximately 7. Typical properties of one commercial product (2) are
presented below:
Molecular Weight 526
Bulk Density 56-60 Ib/cu ft
Water Soluble Iron Expressed as Fe 21.5 percent
Water Soluble Fe+3 19.5 percent
Water Soluble Fe+2 2.0 percent
Insolubles Total 4.0 percent
Free Acid 2.5 percent
Moisture @ 105°C. 2.0 percent
Ferric sulfate is shipped in car and truck load lots of 50 Ib and 100 Ib moistureproof paper
bags and 200 Ib and 400 Ib fiber drums. Bulk carload shipments in box and closed hopper
cars are available. The major producer is Cities Service Company, with a plant located at
Copper Hill, Tennessee.
5-20
-------
The current price of ferric sulfate (21.8 percent Fe) is about $397ton, F.O.B. Copper Hill,
Tennessee. The bagged form costs from $6 to $11/ton more than the bulk.
General precautions should be observed when handling ferric sulfate, such as wearing
goggles and dust masks, and areas of the body that come in contact with the dust or vapor
should be washed promptly.
5.3.3.2 General Design Considerations
Aeration of ferric sulfate should be held to a minimum because of the hygroscopic nature of
the material, particularly in damp atmospheres. Mixing of ferric sulfate and quicklime in
conveying and dust vent systems should be avoided as caking and excessive heating can
result. The presence of ferric sulfate and lime in combination has been known to destroy
cloth bags in pneumatic unloading devices (3). Because ferric sulfate in the presence of
moisture will stain, precautions similar to those discussed for ferric chloride should be
observed.
5.3.3.3 Storage
Ferric sulfate is usually stored in the dry state either in the shipping bags or in bulk in
concrete or steel bins. Bulk storage bins should be as tight as possible to avoid moisture
absorption, but dust collector vents are permissible and desirable. Hoppers on bulk storage
bins should have a minimum slope of 36° however, a greater angle is prefered.
Bins may be located inside or outside and the material transferred by bucket elevator, screw
or air conveyors. Ferric sulfate stored in bins usually absorbs some moisture and forms a
thin protective crust which retards further absorption until the crust is broken.
5.3.3.4 Feeding Equipment
Feed solutions are usually made up at a water to chemical ratio of 2:1 to 8:1 (on a weight
basis) with the usual ratio being 4:1 with a 20-minute detention time. Care must be taken
not to dilute ferric sulfate solutions to less than 1 percent to prevent hydrolysis and depos-
ition of ferric hydroxide. Ferric sulfate is actively corrosive in solution, and dissolving and
transporting equipment should be fabricated of type 316 stainless steel, rubber, plastics, ce-
ramics or lead.
Dry feeding requirements are similar to those for dry alum except that belt type feeders are
rarely used because of their open type of construction. Closed construction, as found in the
volumetric and loss-in-weight-type feeders, generally exposes a minimum of operating
components to the vapor, and thereby minimizes maintenance. A water jet vapor remover
should be provided at the dissolver to protect both the machinery and operator.
5-21
-------
5.3.3.5 Piping and Accessories
Piping systems for ferric sulfate should be FRP, plastics, type 316 stainless steel, rubber or
glass.
5.3.3.6 Pacing and Control
System pacing and control are the same as discussed for dry alum.
5.3.4 Ferrous Sulfate
5.3.4.1 Properties and Availability
Ferrous sulfate or copperas is a byproduct of pickling steel and is produced as granules,
crystals, powder; and lumps. The most common commercial form of ferrous sulfate is
FeSO4- VHaO, with a molecular weight of 278, and containing 55 to 58 percent FeSO4 and
20 to 21 percent Fe. The product has a bulk density of 62 to 66 Ib/cu ft. When dissolved,
ferrous sulfate is acidic. The composition of ferrous sulfate may be quite variable and
should be established by consulting the nearest manufacturers.
Bulk, drum (400 Ib) and bag (50 and 100 Ib) shipments are available from producers at the
following locations:
American Cyanamid Co. Savannah, Georgia
Byproducts Processing Co., Inc. Baltimore, Maryland
Glidden Co. Baltimore, Maryland
Cosmin Corp. Baltimore, Maryland
NL Industries St. Louis, Missouri
NL Industries Sayreville, New Jersey
The current price of ferrous sulfate in bulk carload and truckload quantities is about $187
ton (21 nercent FeV The happed cost is $247 ton
ton (21 percent Fe). The bagged cost is $247 ton.
Ferrous sulfate is also available in a wet state in bulk form from some plants. This form is
likely to be difficult to handle and the manufacturer should be consulted for specific
information and instructions.
Dry ferrous sulfate cakes at storage temperatures above 68°F, is efflorescent in dry air, and
oxidizes and hydrates further in moist air.
General precautions similar to those for ferric sulfate, with respect to dust and handling
acidic solutions, should be observed when working with ferrous sulfate. Mixing quicklime
and ferrous sulfate produces high temperatures and the possibility of fire.
5-22
-------
5.3.4.2 General Design Considerations
The granular form of ferrous sulfate has the best feeding characteristics and gravimetric or
volumetric feeding equipment may be used.
The optimum chemical to water ratio for continuous dissolving is 0.5 Ib/gal. of 6 percent
with a detention time of 5 minutes in the dissolves Mechanical agitation should be provided
in the dissolver to assure complete solution. Lead, rubber, iron, plastics, and type 304 stain-
less steel can be used as construction materials for handling solutions of ferrous sulfate.
Storage, feeding and transporting systems probably should be suitable for handling ferric
sulfate as an alternative to ferrous sulfate.
5.3.5 "Reactions of Iron Compounds
Ferric sulfate and ferric chloride react with the alkalinity of wastewater or with the added
alkaline materials such as lime or soda ash. The reactions may be written to show
precipitation of ferric hydroxide, although in practice, as with alum, the reactions are more
complicated than this. The reactions are shown in Table 5-3, using ferric sulfate.
TABLE 5-3
REACTIONS OF FERRIC SULFATE
Fe2(SO4)3 + 3 Ca(HCO3)2—»-2 Fe(OH)3{ + 3 CaSO4 + 6 CO2f
Fe2(SO4)3 + 3 Na2CO3 + 3 H2O—*-2 Fe(OH)3J + 3 Na2SO4 + 3 CO2f
Fe2(SO4)3 + 3 Ca(OH)2—»-2 Fe(OH)3{ + 3 CaSO4
Ferric chloride can be substituted in these reactions.
In terms of useful quantities, the reactions of Table 5-3 can be expressed as follows:
1. 1 mg/ 1 of Fe2(SO4>3 • 7H2O reacts with:
0.57 mg/ 1 alkalinity, expressed as CaCO3
0.44 mg/ 1 95 percent hydrated lime as Ca(OH)2
0.62 mg/ 1 soda ash as Na2CO3
2. 1 mg/ 1 of anhydrous FeCL3 reacts with:
0.92 mg/ 1 alkalinity expressed as CaCO3
0.72 mg/ 1 95 percent hydrated lime as Ca(OH)2
1.00 mg/ 1 soda ash as Na2CO3
5-23
-------
Ferrous sulfate and ferrous chloride react with the alkalinity of wastewater or with the
added alkaline materials such as lime to precipitate ferrous hydroxide. The ferrous
hydroxide is oxidized to ferric hydroxide by dissolved oxygen in wastewater. Typical
reactions are shown in Table 5-4, using ferrous sulfate.
TABLE 5-4
REACTIONS OF FERROUS SULFATE
FeSO4 + Ca(HCOa)2—^Fe(OH)2| + Ca SO4
FeSO4 + Ca(OH)2—*~Fe(OH)2j + Ca SO4
4 Fe(OH)2 + O2 + 2H2—*- 4 Fe (OH)3 (
Ferrous hydroxide is rather soluble and oxidation to the more insoluble ferric hydroxide is
necessary if high iron residuals in effluents are to be avoided. Flocculation with ferrous iron
is improved by addition of lime or caustic soda at a rate of 1 to 2 mg/mg Fe to serve as a
floe conditioning agent. Polymers are also generally required to produce a clear effluent.
5.4 Lime
The term "lime" applies to a variety of chemicals which are alkaline in nature and contain
principally calcium, oxygen and, in some cases, magnesium. In this grouping are included
quicklime, dolomitic lime, hydrated lime, dolomitic hydrated lime, limestone, and
dolomite. This section is restricted to discussion of quicklime and hydrated lime, but the
dolomitic counterparts of these chemicals, i.e., the high-magnesium forms, are quite
applicable for wastewater treatment and are generally similar in physical requirements.
5.4.1 Quicklime
5.4.1.1 Properties and Availability
Quicklime, CaO, has a density range of approximately 55 to 75 Ib/cu ft, and a molecular
weight of 56.08. A slurry for feeding, called milk of lime, can be prepared with up to 45 per-
cent solids. Lime is only slightly soluble, and both lime dust and slurries are caustic in na-
ture. A saturated solution of lime has a pH of about 12.4.
Lime can be purchased in bulk in both car and truck load lots. It is also shipped in 80 and
100 Ib multiwall "moistureproof" paper bags. Lime is produced at the locations indicated
by Table 5-5.
5-24
-------
TABLE 5-5
Location
PARTIAL LIST OF LIME MANUFACTURING PLANTS (4)
Form of
Manufacturer Lime Available
ALABAMA
Allgood
Keystone
Landmark
Montevallo
Roberta
Saginaw
Siluria
ARIZONA
Douglas
Globe
Nelson
Cheney Lime & Cement Co.
Southern Cement Co.
Div. Martin Marietta Corp.
Cheney Lime & Cement Co.
U.S. Gypsum Co.
Southern Cement Co.
Div. Martin Marietta Corp.
Longview Lime Co., Div.
Woodward Co., Div. Mead Corp.
Alabaster Lime Co.
Paul Lime Plant, Inc.
Hoopes & Co.
U.S. Lime Div., The Flintkote Co.
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
Hish Calcium
ARKANSAS
Batesville
CALIFORNIA
City of Industry
Diamond Springs
Lucerne Valley
Richmond
Salinas
Sonora
Westend
COLORADO
Ft. Morgan
CONNECTICUT
Canaan
Batesville White Lime Co.,
Div. Rangaire Corp.
U.S. Lime Div., The Flintkote Co.
Diamond Springs Lime Co.
Pfizer, Inc., Minerals, Pigments
and Metals Div.
U.S. Lime Div., The Flintkote Co.
Kaiser Aluminum & Chemical Corp.
(currently captive lime)
U.S. Lime Div., The Flintkote Co.
Stauffer Chemical Co.
Great Western Sugar Co.
Pfizer, Inc., Minerals, Pigments
and Metals Div.
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
Dolomitic
Dolomitic
High Calcium
High Calcium
Dolomitic
5-25
-------
^Location
FLORIDA
Brooksville
Sumterville
ILLINOIS
Marblehead
McCook
Quincy
So. Chicago
Thornton
INDIANA
Buffington
IOWA
Davenport
KENTUCKY
Carntown
LOUISIANA
Morgan City
New Orleans
MARYLAND
LeGore
Woodsboro
MASSACHUSETTS
Adams
Lee
MICHIGAN
Detroit
Ludington
Menominee
River Rouge
MINNESOTA
Duluth
TABLE 5-5 (continued)
Manufacturer
Chemical Lime Co.
Dixie Lime and Stone Co.
Marblehead Lime Co.
Standard Lime & Refractories
Div., Martin Marietta Corp.
Marblehead Lime Co.
Marblehead Lime Co.
Marblehead Lime Co.
Marblehead Lime Co.
Linwood Stone Products Co., Inc.
Black River Mining Co.
Pelican State Lime Corp.
U.S. Gypsum Co.
LeGore Lime Co.
S.W. Barrick & Sons, Inc.
Pfizer, Inc., Minerals, Pigments
and Metals Diy.
Lee Lime Corp.
Detroit Lime Co.
Dow Chemical Co. (currently captive lime)
Limestone Products Co., Div.
C. Reiss Coal Co.
Marblehead Lime Co.
Cutler Magner Co.
Form of
Lime Available
High Calcium
High Calcium
High Calcium
Dolomitic
High Calcium
High Calcium
Dolomitic
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
Dolomitic
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
5-26
-------
Location
TABLE 5-5 (continued)
Manufacturer
Form of
Lime Available
MISSOURI
Bonne Terre
Hannibal
Ste. Genevieve
Springfield
NEVAD\
Apex
Henderson
McGill
Sloan
Valley Dolomite Co.
Marblehead Lime Co.
Mississippi Lime Co.
Ash Grove Cement Co.
U.S. Lime Div., The Flintkote Co.
U.S. Lime Div., The Flintkote Co.
Morrison-Weatherly Corp.
U.S. Lime Div., The Flintkote Co.
Dolomitic
High Calcium
High Calcium
High Calcium
High Calcium
Dolomitic &
High Calcium
High Calcium
Dolomitic &
High Calcium
NEW JERSEY
Newton
OHIO
Ashtabula
Carey
Cleveland
Delaware
Geona
Limestone Products Corp. of America
Union Carbide Olefins Co.
National Lime & Stone Co.
Cuyahoga Lime Co.
Marble Cliff Quarries Co.
U.S. Gypsum Co.
Gibsonburg (2 plants) Pfizer, Inc., Minerals, Pigments
and Metal Div., National Gypsum Co.
Huron Huron Lime Co.
High Calcium
High Calcium
Dolomitic
High Calcium
High Calcium
Dolomitic
Dolomitic
High Calcium
Marble Cliff
Millersville
Woodville
OKLAHOMA
Marble City
Sallisaw
OREGON
Baker
Portland
Marble Cliff Quarries Co.
J. E. Baker Co.
Ohio Lime Co., Standard Lime &
Refractories Div., Martin Marietta Corp.
St. ClairLimeCo.
St. Clair Lime Co.
Chemical Lime Co. of Oregon
Ash Grove Cement Co.
High Calcium
Dolomitic
Dolomitic
High Calcium
High Calcium
High Calcium
High Calcium
5-27
-------
TABLE 5-5 (continued)
Location
PENNSYLVANIA
Annville
Bellefonte (2 plants)
Branchton
Devault
Everett
Pleasant Gap
Plymouth Meeting
SOUTH DAKOTA
Rapid City
TENNESSEE
Knoxville (2 plants)
TEXAS
Blum
Cleburne
Clifton
Houston
McNeil
New Braunfels
Round Rock
San Antonio
UTAH
Grantsville
Lehi
Manufacturer
Bethlehem Mines Corp.
National Gypsum Co., Warner Co.
Mercer Lime & Stone Co.
Warner Co.
New Enterprise Stone & Lime Co.
Standard Lime & Refractories Div.,
Martin Marietta Corp.
G. & W. H. Corson, Inc.
Pete Lien & Sons, Inc
Foote Mineral Co., Williams Lime
Manufacturing Co.
Round Rock Lime Companies
Texas Lime Co., Div. Rangaire Corp.
Clifstone Lime Co.
U. S. Gypsum Co.
Austin White Lime Co.
U. S. Gypsum Co.
Round Rock Lime Companies
McDonough Bros., Inc.
U. S. Lime Div., The Flintkote Co.
Rollins Mining Supplies Co.
Form of
Lime Available
High Calcium
High Calcium
High Calcium
Dolomitic
High Calcium
High Calcium
Dolomitic
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
High Calcium
Dolomitic &
High Calcium
High Calcium
VERMONT
Winooski
VIRGINIA
Clearbrook
Kimballton (2 plants)
Vermont Assoc. Lime Industries, Inc.
W.S. Frey Co., Inc.
Foote Mineral Co., National Gypsum
Company
High Calcium
High Calcium
High Calcium
5-28
-------
Location
Stephens City
Strasburg
WASHINGTON
Tacoma
TABLE 5-5 (continued)
Manufacturer
M.J. Grove Lime Co., Div. The
Flintkote Co.
Chemstone Corp.
Domtar Chemicals Inc.
Form of
Lime Available
High Calcium
High Calcium
High Calcium
WEST VIRGINIA
Millville
Riverton
WISCONSIN
Eden
Green Bay
Knowles
Manitowoc
Superior
Standard Lime & Refractories Div.,
Martin Marietta Corp.
Germany Valley Limestone Div.,
Greer Steel Co.
Western Lime'& Cement Co.
Western Lime & Cement Co.
Western Lime & Cement Co.
Rockwell Lime Co.
Cutler-LaLiberte-McDougall Corp.
Dolomitic
High Calcium
Dolomitic
High Calcium
Dolomitic
Dolomitic
High Calcium
Current prices for bulk pebble quicklime range from $187ton to $21/ton with the higher
prices generally in the far west, and higher than average in the north. Bagging adds
approximately $47 ton to the cost.
The CaO content of commercially available quicklime can vary quite widely over an ap-
proximate range of 70 to 96 percent. Content below 88 percent is generally considered be-
low standard in the municipal use field (5). Purchase contracts are often based on 90 per-
cent CaO content with provisions for payment of a bonus for each 1 percent over and a pen-
alty for each 1 percent under the standard. A CaO content less than 75 percent probably
should be rejected because of excessive grit and difficulties in slaking.
Workmen should wear protective clothing and goggles to protect the skin and eyes, as lime
dust and hot slurry can cause severe burns. Areas contacted by lime should be washed
immediately. Lime should not be mixed with chemicals which have water of hydration. The
lime will be slaked by the water of hydration causing excessive temperature rise and
possibly explosive conditions. Conveyors and bins used for more than one chemical should
be thoroughly cleaned before switching chemicals.
5.4.1.2 General Design Considerations
Pebble quicklime, all passing a 3/4 in.iscreen and not more than 5 percent passingia. No.1100
screen, is normally specified because of easier handling and less dust. Hopper agitation is
5-29
-------
generally not required with the pebble form. Published slaker capacity ratings require "soft
or normally burned" limes which provide fast slaking and temperature rise, but poorer
grades of limes may also be satisfactorily slaked by selection of the appropriate slaker
retention time and capacity.
5.4.1.3 Storage
Storage of bagged lime should be in a dry place, and preferably elevated on pallets to avoid
absorption of moisture. System capacities often make the use of bagged quicklime
impractical. Maximum storage period is about 60 days.
Bulk lime is stored in air-tight concrete or steej bins having a 55 to 60 deg slope on the bin
outlet. Bulk lime can be conveyed by conventional bucket elevators and screw, belt, apron,
drag-chain, and bulk conveyors of mild steel construction. Pneumatic conveyors subject the
lime to air-slaking and particle sizes may be reduced by attrition. Dust collectors should be
provided on manually and pneumatically-filled bins.
5.4.1.4 Feeding Equipment
A typical lime storage and feed system is illustrated in Figure 5-8. Quicklime feeders are
usually limited to the belt or loss-in-weight gravimetric types because of the wide variation
of the bulk density. Feed equipment should have an adjustable feed range of at least 20:1 to
match the operating range of the associated slaker. The feeders should have an over-under
feed rate alarm to immediately warn of operation beyond set limits of control. The feeder
- * . . •
drive should be instrumented to be interrupted in the event of excessive temperature in the
slaker compartment.
Lime slakers for wastewater treatment should be of the continuous type, and the major
components should include one or more slaking compartments, a dilution compartment, a
grit separation compartment and a continuous grit remover. Commercial designs vary in
regard to the combination of water to lime, slaking temperature, and slaking time, in
obtaining the "milk of lime" suspensions.
The "paste-type" slaker admits water as required to maintain a desired mixing viscosity.
This viscosity therefore sets the operating retention time of the slaker. The paste slaker
usually operates with a low water to lime ratio (approximately 2:1 by weight), elevated
temperature, and five-minute slaking time at maximum capacity.
The "detention" type slaker admits water to maintain a desired ratio with the lime, and
therefore the lime feed rate sets the retention time of the slaker. The detention slaker
operates with a wide range of water to lime ratios (2.5:1 and 6:1), moderate temperature,
and a 10 minute slaking time at maximum capacity. A water to lime ratio of from 3.5:1 to
4:1 is most often used. The operating temperature in lime slakers is a function of the water
to lime ratio, lime quality, heat transfer, and water temperature. Lime slaking evolves heat
in hydrating the CaO to Ga(OH)2 and therefore, vapor removers are required for feeder
protection.
5-30
-------
NOTE: VAPOR REMOVER
HOT SHOWN FOR CLARITY
COLLECTOR
FILL PIPE (PNEUMATIC)
BULK STORAGE
BIN
\
SCALE
FEEDER
SOLENOID
VALVE,,
OR SAMPLE CHUTE
ROTAMETERS
SLAKING WATER
DILUTION WATER
BIN GATE
FLEX IBLE
CONNECTION
FLOW RECORDER
WITH PACING
TRANSMITTER^
pH RECORDER
CONTROLLER
ROTOD IP-TYPE
FEEDER
GRAVITY FEED
RECIRCULATION
HOLDING
TANK
METERING
PUMP-
'BACK
PRESSURE
VALVE
FIGURE 5-8
TYPICAL LIME FEED SYSTEM
5-31
-------
5.4.1.5 Piping and Accessories
Lime slurry should be transported by gravity in open channels wherever possible. Piping
channels, and accessories may be rubber, iron, steel, concrete, and plastics. Glass tubing,.
such as that in rotameters, will cloud rapidly and therefore should not be used. Any abrupt
directional changes in piping should include plugged tees or crosses to allow rodding-out of
deposits. Long sweep elbows should be provided to allow the piping to be cleaned by the use
of a cleaning "pig". Daily cleaning is desirable.
Milk of lime transfer pumps should be of the open impeller centrifugal type. Pumps having
an iron body and impeller with bronze trim are suitable for this purpose. Rubber-lined
pumps with rubber-covered impellers are also frequently used. Make-up tanks are usually
provided ahead of centrifugal pumps to ensure a flooded suction at all times. "Plating-out"
of lime is minimized by the use of soft water in the make-up tank and slurry recirculation.
Turbine pumps and eductors should be avoided in transferring milk of lime because of
scaling problems.
5.4.1.6 Pacing and Control
Lime slaker water proportioning is integrally-controlled or paced from the feeder.
Therefore, the feeder-slaker system will follow pacing controls applied to the feeder only.
As discussed previously, gravimetric feeders are adaptable to receive most standard
instrumentation pacing signals. Systems can be instrumented to allow remote pacing with
telemetering of temperature and feed rate to a central panel for control purposes.
The lime feeding system may be controlled by an instrumentation system integrating both
plant flow and pH of the wastewater after lime addition. However, it should be recognized
that pH probes require daily maintenance in this application to monitor the pH accurately.
Deposits tend to build up on the probe and necessitate frequent maintenance. The low pH
lime treatment systems (pH 9.5 to 10.0) can be more readily adapted to this method of
control than high-lime treatment systems (pH 11.0 or greater) because less maintenance of
the pH equipment is required. In a closed-loop pH-flow control system, milk of lime is
prepared on a batch basis and transferred to a holding tank with variable output feeders
set by the flow and pH meters to proportion the feed rate. Figure 5-8 illustrates such a
control system.
5.4.2 Hydrated Lime
5.4.2.1 Properties and Availability
Hydrated lime, Ca(OH)2, is usually a white powder (200 to 400 mesh); has a bulk density of
20 to 50 Ib/cu ft; contains 82 to 98 percent Ca(OH)2; is slightly hydroscopic; tends to flood
the feeder, and will arch in storage bins if packed. The molecular weight is 74.08. The dust
.and slurry of hydrated lime are caustic in nature. The cost of bulk hydrated lime varies
from $18 to $22/ton. Bagged lime is available but increases the cost from $4 to $167ton.
The availability of hydrated lime may be determined by contacting manufacturers listed in
5-32
-------
Table 5-5. The pH of a saturated, hydrated lime solution is the same as that given for quick-
lime.
5.4.2.2 General Design Considerations
Hydrated lime is slaked lime and needs only enough water added to form milk of lime.
Wetting or dissolving chambers are usually designed to provide 5 minutes detention with a
ratio of 0.5 Ib/gal of water or 6 percent slurry at the maximum feed rate. Hydrated lime
is usually used where maximum feed rates do not exceed 250 lb/hr., i.e., smaller plants.
Hydrated lime and milk of lime will irritate the eyes, nose, and respiratory system and will
dry the skin. Affected areas should be washed with water.
5.4.2.3 Storage
Information given for quicklime also applies to hydrated lime except that bin agitation
must be provided. Bulk bin outlets should be provided with non-flooding rotary feeders.
Hopper slopes vary from 60 to 66 deg.
5.4.2.4 Feed Equipment
Volumetric or gravimetric feeders may be used, but volumetric feeders are usually selected
only for installations where comparatively low feed rates are required. Dilution does not
appear to be important, therefore, control of the amount of water used in the feeding
operation is not considered necessary. Inexpensive hydraulic jet agitation may be furnished
in the wetting chamber of the feeder as an alternative to mechanical agitation. The jets
should be sized for the available water supply pressure to obtain proper mixing.
5.4.2.5 Piping and Accessories
Piping and accessories as described for quicklime are also appropriate for hydrated lime.
5.4.2.6 Pacing and Controls
Controls as listed for dry alum apply to hydrated lime. Hydraulic jets should operate
continuously and only shut off when the feeder is taken out of service. Control of the feed
rate with pH as well as pacing with the plant flow may be used with hydrated lime as well as
quicklime.
5.4.3 Reactions of Lime
Lime is somewhat different from the hydrolyzing coagulants. When added to wastewater it
increases pH and reacts with the carbonate alkalinity to precipitate calcium carbonate.. If
sufficient lime is added to reach a high pH, approximately 10.5, magnesium hydroxide is also
precipitated. This latter precipitation enhances clarification due to the flocculant nature of
the Mg(OH)2. Excess calcium ions at high pH levels may be precipitated by the addition of
soda ash. The above reactions are shown in Table 5-6.
5-33
-------
TABLE 5-6
REACTIONS OF LIME
Ca(OH)2 + Ca(HCO3)2— ^2 CaCOaf + 2H2O
2 Ca(OH)2 + Mg(HCO3)2— +- 2 CaCOaj + Mg(OH)2) + 2H2O
Ca(OH)2 + Na2CO3— *• CaCOa { + 2 NaOH
Reduction of the resulting high pH levels may be accomplished in one or two stages. The
first stage of the two-stage method results in the precipitation of calcium carbonate through
the addition of carbon dioxide according to the following reaction:
Ca(OH)2 + COa— »-CaCO3 + H2O
Single-stage pH reduction is generally accomplished by the addition of carbon dioxide,
although acids have been employed. This reaction, which also represents the second stage of
the two-stage method, is as follows:
Ca(OH)2 + 2 CO2 — ^Ca(HCO3)2
As noted for the other chemicals, the above reactions are merely approximations to the
more complex interactions which actually occur in wastewaters.
The lime demand of a given wastewater is a function of the buffer capacity or alkalinity of
the wastewater. Figure 5-9 shows this relationship for a number of different wastewaters
(6).
5.5 Other Inorganic Chemicals
In addition to aluminum and iron salts and lime a number of other inorganic chemicals
have been used in wastewater treatment. Only three are discussed in this section, i.e., soda
ash, caustic soda, and carbon dioxide, but others have been and will be employed. Mineral
and other acids are prime examples. For information on any of these chemicals, the local
supplier or manufacturer should be contacted.
5.5.1 Soda Ash
5.5.1.1 Properties and Availability.
Soda ash, Na2COa, is available in two forms. Light soda ash has a bulk density range of 35
to 50 Ib/cu ft and a working density of 41 Ib/cu ft. Dense soda ash has.a density range of
60 to 76 Ib/cu ft and a working density of 63 Ib/cu ft. The pH of a 1 percent solution of
soda ash is 11.2. It is used for pH control and in lime treatment.
5-34
-------
500
5
o
§
tu,
Q
W
«
I
w
a:
w
400
300
200
w
100
0
100 200 300 400
WASTEWATER ALKALINITY mg/1 - CaCO3
500
FIGURE 5-9
LIME REQUIREMENT FOR pH£ 11.0 AS A FUNCTION OF THE
WASTEWATER ALKALINITY
U S EPA Headquarters Library
Mail code 3404T
1200 Pennsylvania Avenue NW
Washinaton, DC 20460
202-566-0556
5-35
-------
The molecular weight of soda ash is 106. Commercial purity ranges from 98 to greater than
99 percent NaaCOs. The viscosities of sodium carbonate solutions are given in Figure 5-10.
Soda ash by itself is not particularly corrosive, but in the presence of lime and water, caus-
tic soda is formed which is quite corrosive.
Soda ash is available in bulk by truck, box car and hopper car, and in 100 Ib bags from the
following partial list of manufacturers.
Location
Manufacturer
CALIFORNIA
Bartlett
Trona
Westend
PPG Industries, Inc.
American Potash and Chemical
Corp.
Stauffer Chemical Co.
GEORGIA
Brunswick
Allied Chemical Co.
LOUISIANA
Baton Rouge
Lake Charles
Allied Chemical Co.
Olin Chemicals
MICHIGAN
Wyandotte
NEW YORK
Solvay
OHIO
Barberton
Painesville
Wyandotte Chemicals Corp.
Allied Chemical Co.
PPG Industries, Inc.
Diamond Shamrock Chemical Co.
TEXAS
Corpus Christi
WEST VIRGINIA
Moundsville
PPG Industries, Inc.
Allied Chemical Co.
WYOMING
Green River (3 plants)
Allied Chemical Co, FMC Corp., and
Stauffer Chemical Corp.
5-36
-------
(COURTESY PPG INDUSTRIES INC-, CHEMICAL DIV.)
CO
LU
co
p
>-
6.0
5.0
4.0
3-0
2.0
1.0
10
15
20
25
30
FIGURE 5-10
VISCOSITY OF SODA ASH SOLUTIONS
5-37
-------
The current price for soda ash ranges from $40 to $50/ ton, F.O.B. the point of
manufacture, however, prices vary substantially between manufacturers and should be
obtained from the closest manufacturers or local distributors. Bagging may add
substantially to the cost of the chemical.
5.5.1.2 General Design Considerations
Dense soda ash is generally used in municipal applications because of superior handling
characteristics. It has little dust, good flow characteristics, and will not arch in the bin or
flood the feeder. It is relatively hard to dissolve and ample dissolver capacity must be
provided. Normal practice calls for 0.5 Ib of dense soda ash per gal. of water or a 6 percent
solution retained for 20 min in the dissolver.
The dust and solution are irritating to the eyes, nose, lungs and skin and therefore general
precautions should be observed and the affected areas should be washed promptly with
water.
5.5.1.3 Storage
Soda ash is usually stored in steel bins and where pneumatic filling equipment is used, bins
should be provided with dust collectors. Bulk and bagged soda ash tend to absorb
atmospheric CO2 and water, forming the less active sodium bicarbonate (NaHCOa).
Material recommended for unloading facilities is steel.
5.5.1.4 Feeding Equipment
Feed equipment as described for dry alum is suitable for soda ash. Dissolving of soda ash
may be hastened by the use of warm dissolving water. Mechanical or hydraulic jet mixing
should be provided in the dissolver.
5.5.1.5 Piping and Accessories
Materials of construction for piping and accessories should be iron, steel, rubber, and
plastics.
5.5.1.6 Pacing and Control
Controls as discussed for dry alum apply also to soda ash equipment.
5.5.2 Liquid Caustic Soda
Anhydrous caustic soda (NaOH) is available but its use is generally not considered
practical in water and wastewater treatment applications. Consequently, only liquid caustic
soda is discussed below.
5-38
-------
5.5.2.1 Properties and Availability.
Liquid caustic soda is shipped at two concentrations, 50 percent and 73 percent NaOH. The
densities of the solutions as shipped are 12.76 Ib/gal for the 50 percent solution and 14.18
Ib/gal for the 73 percent solution. These solutions contain 6.38 Ib/gal NaOH and 10.34 lb/
gal. NaOH, respectively. The crystallization temperature is 53°F for the 50 percent solution
and 165°F for the 73 percent solution. The molecular weight of NaOH is 40. Viscosities of
various caustic soda solutions are presented in Figure 5-11. The pH of a 1 percent solution
of caustic soda is 12.9.
Truck load lots of 1,000 to 4,000 gal are available in the 50 percent concentration only.
Both shipping concentrations can be obtained in 8,000, 10,000 and 16,000 gal car load lots.
Tank cars can be unloaded through the dome eduction pipe using air pressure or through
the bottom valve by gravity or.by using air pressure or a pump. Trucks are usually unloaded
by gravity or with air pressure or a truck mounted pump.
Major producers of caustic soda and their respective plantlocations are listed in Table 5-7.
The current price for liquid caustic soda ranges from $76/ton @ 50 percent and $81/ton @
73 percent, (NaOH), F.O.B. the point of manufacture.
TABLE 5-7
PARTIAL LIST OF CAUSTIC SODA MANUFACTURING PLANTS
Location Manufacturer
ALABAMA
Lemoyne (Mobile) Stauffer
Mclntosh Olin
Muscle Shoals Diamond Shamrock
CALIFORNIA
Pittsburg Dow
DELAWARE
Delaware City Diamond Shamrock
GEORGIA
Augusta Olin
Brunswick , Allied
KANSAS
Wichita Vulcan
KENTUCKY
Calvert City (2 plants) Pennwalt, Goodrich
5-39
-------
CO
LU
CO
o
200
iOO
80
60
40
20
(COURTESY OF HOOKER CHEMICAL Co-)
;= 10
Z 8
o ,
!2 2
i
0.8
0.6
0-4
0.2
I I I
T I I I I I
I I
I
J I
I I I I
I
60 70 80 90 100 110 120 130 140 150 160 170 180 190 200
TEMPERATURE, °F
FIGURE 5-11
VISCOSITY OF CAUSTIC SODA SOLUTIONS
5-40
-------
Location
Manufacturer
LOUISIANA
Baton Rouge
Geismar
Lake Charles (2 plants)
Plaquemine
St. Gabriel
Taft
Allied
Wyandotte
PPG, Olin
Dow
Stauffer
Hooker
MICHIGAN
Midland
Montague
Wyandotte (2 plants)
NEVADA
Henderson
Dow
Hooker
Pennwalt, Wyandotte
Stauffer
NEW JERSEY
Linden
GAP
NEW YORK
Niagara Falls (3 plants)
Solvay
NORTH CAROLINA
Acme
Hooker, Olin, Stauffer
Allied
Allied
OHIO
Barberton
Cleveland
Painesville
PPG
Harshaw
Diamond Shamrock
OREGON
Portland
Pennwalt
TENNESSEE
Charleston
Olin
TEXAS
Corpus Christi
Deer Park (Houston)
Freeport
Port Neches
PPG
Diamond Shamrock
Dow
Jefferson
5-41
-------
Location
Manufacturer
VIRGINIA
Saltville
WASHINGTON
Tacoma (2 plants)
Olin
Hooker, Pennwalt
WEST VIRGINIA
Moundsville
Natrium
South Charleston
Allied
PPG
FMC
Manufacturers and addresses
Allied Chemical
Solvay Process Division
40 Rector Street
New York, New York 10006
Diamond Shamrock Chemical Co.
300 Union Commerce Building
Cleveland, Ohio 44115
Dow Chemical Co.
Abbott Road
Midland, Michigan 48640
GAF Corp. Chemical Division
140 West 51st Street
New York, New York 10019
FMC Corporation
Inorganic Chemicals Div.
633 Third Avenue
New York, New York 10017
B.F. Goodrich Chemical Co.
3135 Euclid Avenue
Cleveland, Ohio 44115
Harshaw Chemical Co.
1945 East 97th Street
Cleveland, Ohio 444106
Hooker Chemical Corp.
P.O. Box 344
Niagara Falls, New York 14302
Jefferson Chemical Co., Inc.
3336 Richmond Avenue
Houston, Texas 77006
Olin Corporation
Chemicals Division
745 Fifth Avenue
New York, New York 10022
5-42
-------
Pennwalt Corporation
Pennwalt Building
Three Penn Center
Philadelphia, Pa. 19102
PPG Industries, Inc.
Chemical Division
1 Gateway Center
Pittsburgh, Pa. 15222
Stauffer Chemical Co.
Industrial Chemical Div.
299 Park Avenue
New York, New York 10017
Vulcan Materials Co.
Chemicals Division
P.O. Box 545-T
Wichita, Kansas 67201
Wyandotte Chemicals Corp.
Michigan Alkali Division
Wyandotte, Michigan 48192
5.5.2.2 General Design Considerations
Liquid caustic soda is received in bulk shipments, transferred to storage and diluted as
necessary for feeding to the points of application. Caustic soda is poisonous and is
dangerous to handle. U.S. Department of Transportation Regulations for "White Label"
materials must be observed. However, if handled properly caustic soda poses no particular
industrial hazard. To avoid accidental spills, all pumps, valves, and lines should be checked
regularly for leaks. Workmen should be thoroughly instructed in the precautions related to
the handling of caustic soda. The eyes should be protected by goggles at all times when
exposure to mist or splashing is possible. Other parts of the body should be protected as
necessary to prevent alkali burns. Areas exposed to caustic soda should be washed with
copious amounts of water for 15 min to 2 hr. A physician should be called when
exposure is severe. Caustic soda taken internally should be diluted with water or milk and
then neutralized with dilute vinegar or fruit juice. Vomiting may occur spontaneously but
should not be induced except on the advice of a physician.
5.5.2.3 Storage
Liquid caustic soda may be stored at the 50 percent concentration. However, at this solu-
tion strength, it crystallizes at 53°F. Therefore, storage tanks must be located indoors or
provided with heating and suitable insulation if outdoors. Because of its relatively high crys-
tallization temperature, liquid caustic soda is often dilluted to a concentration of about 20
percent NaOH for storage. A 20 percent solution of NaOH has a crystallization tempera-
ture of about -20°F. Recommendations for dilution of both 73 percent and 50 percent solu-
tions should be obtained from the manufacturer, because special considerations are neces-
sary.
Storage tanks for liquid caustic soda should be provided with an air vent for gravity flow.
The storage capacity should be equal to 1 l/2 times the largest expected delivery, with an
allowance for dilution water, if used, or 2-weeks supply at the anticipated feed rate,
whichever is greater. Tanks for storing 50 % solution at a temperature between 75°F and
5-43
-------
140°F may be constructed of mild steel. Storage temperatures above 140°F require more
elaborate materials selection and are not recommended. Caustic soda will tend to pick up
iron when stored in steel vessels for extended periods. Subject to temperature and solution
strength limitations, rubber, 316 stainless steel, nickel, nickel alloys, or plastics may be used
when iron contamination must be avoided.
5.5.2.4 Feeding Equipment
Further dilution of liquid caustic soda below the storage strength may be desirable for
feeding by volumetric feeders. Feeding systems as described for liquid alum generally apply
to caustic soda with appropriate selection of materials of construction. A typical system
•schematic is shown in Figure 5-12. Feeders will usually include materials such as ductile
iron, stainless steels, rubber, and plastics.
5.5.2.5 Piping and Accessories
Transfer lines from the shipping unit to the storage tank should be spiral-wire-bound
neoprene or rubber hose, solid steel pipe with swivel joints, or steel hose. Because caustic
soda attacks glass, use of glass materials should be avoided. Other miscellaneous materials
for use with liquid caustic soda feeding and handling equipment are listed below (7):
Components
Rigid Pipe
Flexible Connections
Diluting Tees
Fittings
Permanent Joints
Unions
Valves—Non-leaking (Plug)
Body
Plug
Pumps (Centrifugal)
Body
Impeller
Packing
Storage Tanks
Recommended Materials
for Use With 50 % NaOH
Up to 14QQF
Standard Weight Black Iron
Rigid Pipe with Ells or Swing Joints,
Stainless Steel or Rubber Hose
Type 304 Stainless Steel
Steel
Welded or Screwed Fittings
Screwed Steel
Steel
Type 304 Stainless Steel
Steel
Ni-Resist
Blue Asbestos
Steel
5-44
-------
-TRUCK FILL LINE
DILUTION
WATER
SODIUM HYDROXIDE
STORAGE TANK
VENT, OVERFLOW
AND DRAIN
VENT, OVERFLOW
AND DRAIN
MIXER
SAMPLE TAP
SODIUM HYDROXIDE
FEEDER
POINT OF
APPLICATION
FIGURE 5-12
TYPICAL CAUSTIC SODA FEED SYSTEM
5-45
-------
5.5.2.6 Pacing and Control
Controls as listed for liquid alum also apply to liquid caustic soda equipment.
5.5.3 Carbon Dioxide
5.5.3.1 -Properties and Availability
Carbon dioxide, ,CO2, is available for use in wastewater treatment plants in gas and liquid
form. The molecular weight of CO2 is 44. Dry COa is not chemically active at normal
temperatures and is a non-toxic safe chemical; however, the gas displaces oxygen and
adequate ventilation of closed areas should be provided. Solutions of CO? in water are very
. reactive chemically and form carbonic acid. Saturated solutions of CO2 have a pH of 4.0 at
68°F.
The gas form may be produced on the treatment plant site by scrubbing and compressing
the combustion product of lime recalcining furnaces, sludge furnaces, or generators used
principally for the production of COa gas only. These generators are usually fired with
combustible'gases, fuel oil, or coke and have COa yields as shown in Table 5-8 (8).
TABLE 5-8
COa YIELDS OF COMMON FUELS
Fuel Quantity . COa Yield
I ! Ib
Natural Gas 1,000 cu ft 115
Coke 1 Ib 3
Kerosene 1 gal. 20
Fuel Oil (No. 2) 1 gal. ' 23
Propane l.,000 cu ft 141
Butane l.OOOcuft 142
The gas forms, as generated at the plant site, usually have a COa content of between 6 per-
cent and 18 percent depending on the source and efficiency of the producing system.
The liquid form is available from commercial suppliers in 20 to 50 Ib cylinders, 10 to 20 ton
trucks and 30 to 50 ton rail cars. The commercial liquid form has a minimum COa content
of 99.5 percent.
Current prices range from $307 ton for 3,000 tons per year and over, to $687 ton for a
quantity of 150 tons/year. These prices include an allowance for freight within a 100 mile
radius of the point of manufacture. Another $6/ton may be added for each additional 100
miles to the point of destination. Major producers of commercial COa are listed in Table
5-9.
5-46
-------
TABLE 5-9
PARTIAL LIST OF CARBON DIOXIDE MANUFACTURING PLANTS
Location
CALIFORNIA
Watson (Los Angeles)
Oakland
Brea
Lathrop
Ventura ,
Taft
GEORGIA
Augusta
ILLINOIS
Morris
Chicago
INDIANA
Jeffersbnville
IOWA
Clinton
Ft. Madison
Ft. Dodge
Muscatine
KANSAS
Dodge City
Lawrence
Lawrence
KENTUCKY
Doerun (Brandenburg)
LOUISIANA
New Orleans
Luling
MASSACTUSETTS
Tewksbury
MISSISSIPPI
Yazoo City
MISSOURI
Kansas City
Le May (St. Louis)
Manufacturer
Liquid Carbonic
Liquid Carbonic
Airco
Airco
Cardox
Standard Oil
Liquid Carbonic
Cardox
Airco
Cardox
Airco
Liquid Carbonic
Liquic Carbonic
Publicker
Liquid Carbonic
Airco (1972)
Cardox
Olin
Liquid Carbonic
Airco
Liquid Carbonic
Airco
Airco
Cardox
5-47
-------
Location
Manufacturer
NEW JERSEY
Paulsboro
Belleville
Deepwater
NEW MEXICO
Bueyeros
Solana
Mosquero
NEW YORK
Olean
OHIO
Toledo
Oregon (Toledo)
Lima
Huron
PENNSYLVANIA
Philadelphia
Thermice (Philadelphia)
TENNESSEE
Woodstock (Memphis)
TEXAS
Texas City
Dallas
Dumas
VIRGINIA
Hopewell
Saltville
WASHINGTON
Finley
Olin
Liquid Carbonic
Airco
SEC
SEC
SEC
Airco
Cardox
Liquid Carbonic
Airco
Cardox
Liquid Carbonic
Publicker
Cardox
Liquid Carbonic
Cardox
Diamond Shamrock
Airco
Olin
Airco
Manufacturers and Addresses
Cardox Div. of Chemetron Corp.
Dept. TR
840 N. Michigan Avenue
Chicago, Illinois 60611
Airco Industrial Gases Div.of
Air Reduction Co.
575 Mountain Avenue
Murray Hill, N.J. 07974
Olin Corporation
Chemicals Division
745 Fifth Avenue
New York, New York 10022
Publicker Ind., Inc.
Walnut & Thomas
Philadelphia, Pa.
5-48
-------
Liquid Carbonic Corporation Diamond Shamrock Chemical Co.
Dept. TR 300 Union Commerce Building
135 S. LaSalle Cleveland, Ohio 44115
Chicago, Illinois 60603
Standard Oil Company of California
1033 Humble Place or • /^ IT n*,nA
„. „ ~ —oc-, San Francisco, California 94104
El Paso, Texas 79987
5.5.3.2 General Design Considerations
Recovery of CO2 from recalcining furnaces or incinerators is the least expensive source, but
maintenance of stack gas systems is likely to be extensive because of the corrosive nature of
the wet gas and the presence of particulate matter. Scrubber systems are required to clean
the stack gas and specially designed gas compressors are necessary to provide the process
injection pressure.
Pressure generators and submerged burners require less maintenance because the system
pressure is established by compressors or blowers handling dry air or gas. On-site
generating units have a limited range of CO2 production as compared with the liquid
storage and feed system, and therefore may require multiple units.
The liquid COa storage and feed system generally includes a temperature-pressure
controlled, bulk storage tank, an evaporation unit, and a gas feeder to meter the gas.
Solution feeders, similar in construction to chlorinators, may also be used to feed COa.
5.5.3.3 Storage
This section applies only to use of commercial liquid COa. Liquid system capacities
encountered in wastewater treatment usually require on-site bulk storage units. Standard
pre-packaged units are available, ranging in size from 3/s to 50 tons capacity, and are
furnished with temperature-pressure controls to maintain approximately 300 psi at 0°F
conditions. The typical package unit contains refrigeration, vaporization, safety and control
equipment. The units are well insulated and protected for outdoor location. The gas from
the evaporation unit usually passes through two stages of pressure reduction before entering
the gas feeder to prevent the formation of dry ice.
5.5.3.4 Feeding Equipment
Feeding systems for the stack gas source of COa consist of simple valving arrangements, for
admitting varying quantities of make-up air to the suction side of the constant volume
compressors, or for venting excess gas on the compressor discharge. A typical system is
described elsewhere (9).
Pressure generators and submerged burners are regulated by valving arrangements on the
fuel and air supply. Generation of COz by combustion is usually difficult to control,
requires frequent operator attention and demands considerable maintenance over the life of
5-49
-------
the equipment, when compared with liquid CO2 systems.
Commercial liquid carbon dioxide is becoming more generally used because of its high
purity, the simplicity and range of feeding equipment, ease of control, and smaller, less
expensive piping systems. After vaporization, the COa with suitable metering and pressure
reduction may be fed directly to the point of application as a gas. However, vacuum
operated,-solution type gas feeders are often preferred. Such feeders generally include safety
devices and operating controls in a compact panel housing, with materials of construction
suitable for CO2 service. Absorption of CO2 in the injector water supply approaches 100%
when a ratio of 1.0 Ib of gas to 60 gal of water is maintained.
5.5.3.5 Piping and Accessories
Mild steel piping and accessories are suitable for use with cool, dry, carbon dioxide. Hot,
moist gases, however, require the use of type 316 stainless steel or plastic materials. Plastics
or FRP pipe are generally used for solution piping and diffusers. Diffusers should be
submerged at least 8 ft, and preferably deeper, to assure complete absorption of the gas.
. 5.5.3.6 Pacing and Control
Standard instrument signals and control components can be used to pace or control carbon
dioxide feed systems.
Using stack gas as the source of COa, thejfeed rate can be controlled by proper selection^and
operation of compressors, by manual control of vent or bleed valves, or by automatic
control of such valves by a pH meter-controller system.
In commercial CO2 feed systems, solution feeders may function as controllers and can be
paced by instrument signals from pH monitors and plant flow meters.
In feeding commercial CO2 directly to the point of application as a gas, a differential
pressure transmitter and a control valve may function as the primary elements of a control
system. Standard instrument signals may be used to pace or control the rate of CO2 feed.
CO2 generators are difficult to pace or control other than by manual or automatic
operation of vent or bleed valves that waste a portion of the produced gas according to the
plant requirements.
5.6. Polymers
Polymeric flocculants are high molecular weight organic chains with ionic or other
functional groups incorporated at intervals along the chains. Because these compounds have
characteristics of both polymers and electrolytes, they are frequently called
polyelectrolytes. They may be of natural or synthetic origin.
All synthetic polyelectrolytes can be classified "on the basis of the type of charge on the
5-50
-------
polymer chain. Thus polymers possessing negative charges are called anionic while those
carrying positive charges are cationic. Certain compounds carry no electrical charge and
are called nonionic polyelectrolytes.
Because of the great variety of monomers available as starting material and the additional
variety that can be obtained by varying the molecular weight, charge density and ionizable
groups, it is not surprising that a great assortment of polyelectrolytes are available to the
wastewater plant operator. A partial listing of manufacturers is shown in Table 5-10. This
list is based mainly on three major sources (10) (11) (12) and does not purport to be a
complete list.
Extensive use of any specific polymer as a flocculant is of necessity determined by the size,
density and ionic charge of the colloids to be coagulated. As other factors need to be
considered, i.e. coagulants used, pH of the system, techniques and equipment for
dissolution of the polyelectroyte, etc.; it is mandatory that extensive jar testing be
performed to determine the specific polymer that will perform its function most efficiently.
These results should be verified by plant-scale testing.
5.6.1 Dry Polymers
5.6.1.1 Properties and Availability
Types of polymers vary widely in characteristics. Manufacturers should be consulted for
properties, availability, and cost of the polymer being considered. References are available
that indicate the types and characteristics of polymers available (10) (11) (12). Bulk
shipments are generally not desirable. Polymers are available in a variety of container or
package sizes.
TABLE 5-10
PARTIAL LIST OF
POLYMER SOURCES AND TRADE NAMES
Source Trade Name (s)
Allied Colloids, Inc. Percol
One Robinson Lane
Ridgewood, N.J. 07450
Allstate Chemical Co. Allstate
Box 3040
Euclid, Ohio 44117
5-51
-------
TABLE 5-10 (continued)
Source
Allyn Chemical Co.
2224 Fairhill Rd.
Cleveland, Ohio 44106
American Cyanamid Co.
Berdan Ave.
Wayne, N.J. 07470
Atlas Chemical Ind., Inc.
Wilmington, Dela. 19899
Berdell Industries
28-01 Thomson Ave.
Long Island City, NY 11101
Betz Laboratories, Inc.
Somerton Red.
Trevose, Pa. 19047
Bond Chemicals, Inc.
1500 Brookpark Rd.
Cleveland, Ohio 44109
Brennan Chemical Co.
704 N. First St.
St. Louis, Mo. 63102
The Burtonite Company
Nutley, N.J. 07110
Trade Name (s)
Claron
Superfloc
Magnifloc
Sorbo
Atlasep
Berdell
Betz
Polyfloc
Bondfloc
Brenco
Burtonite
Calgon Corporation
P.O. Box 1346
Pittsburgh, Pa. 15222
Commercial Chemical
11 Paterson Ave.
Midland Park, N.J. 07432
Cat-Floe
Speedifloc
5-52
-------
TABLE 5-10 (continued)
Source
Dearborn Chemical Div.
W.R. Grace & Co.
Merchandise Mart Plaza
Chicago, 111. 60654
Dow Chemical USA
Barstow Building
2020 Dow Center
Midland, MI. 48640
Drew Chemical Corp.
701 Jefferson Rd.
Parisippany, N.J. 07054
Du Bois Chemicals Div.
W.R. Grace & Co.
3630 E. Kemper Rd.
Sharonville, Ohio 45241
E.I. DuPont de Nemours & Co.
Eastern Laboratory
Gibbstown, N.J. 08027
Environmental Pollution Investigation
& Control, Inc.
9221 Bond St.
Overland Park, KS. 66214
Fabcon International
1275 Columbus Ave.
San Francisco, Calif. 94133
j'rade Name (s)
Aquafloc
Dowell
PEI
Purifloc
Separan
XD
Drewfloc
Amerfloc
Flocculite
Du Pont
Dynafloc
Zuclar
Fabcon
Henry W. Fink & Co.
6900 Silverton Avenue
Cincinnati, Ohio 45236
Gamlen Sybron
321 Victory Avenue
S. San Francisco, Calif. 94080
Kleer-Floc
Gamafloc
Gamlose
Gamlen
5-53
-------
TABLE 5-10 (continued)
Source^
Garrett-Callahan
111 Rollins Rd.
Millbrae, Calif. 94030
General Mills Chemicals
4620 N. 77th Street
Minneapolis, Min. 55435
Hercules, Inc.
910 Market St.
Wilmington, Dela. 19899
Frank Herzl Corp.
299 Madison Avenue
New York, N.Y. 10017
ICI America, Inc.
Wilmington, Dela. 19899
Illinois Water Treatment Co.
840 Cedar St.
Rock ford, 111. 61102
Kelco Company
8225 Aero Dr.
San Diego, Calif. 92123
Key Chemicals
4346 Tacony
Philadelphia, Pa. 19124
Metalene Chemical Co.
Bedford, Ohio 44014
Trade Name(s)
Garrett-Callahan
Supercol
Guartec
Hercofloc
Perfectamyl
Atlasep
Illco
Kelgin
Kelcosol
Key-Floe
Metalene
The Mogul Corporation
20600 Chagrin Blvd.
Cleveland, Ohio 44122
Nalco Chemical Co.
6216 W. 66th Street
Chicago, 111. 60638
Mogul
Nalcolyte
5-54
-------
TABLE 5.-10 (continued)
Source
Trade Name (s)
Narvon Mining & Chemical Co.
Keller Ave. & Fruitville Pike
Lancaster, Pa. 17604
National Starch & Chemical Corp.
1700 W. Frront St.
Plainfield, N.J. 07063
O'Brien Industries, Inc.
95 Dorsa Avenue
Livingston, N.J. 07039
Oxford Chemical Div.
Consolidated Foods Corp.
P.O. Box 80202
Atlanta, Ga. 30341
Reichhold Chemicals, Inc.
RCI Building
White Plains, N.Y. 10602
Standard Brands Chem. Ind., Inc.
P.O. Drawer K
Dover, Dela. 19901
A.E. Staley Mfg. Co.
P.O. Box 151
Decatur, 111. 62525
Stein, Hall & Co., Inc.
605 Third Avenue
New York, N.Y. 10016
Swift & Company
Oakbrook, 111.60521
James Varley & Sons, Inc.
1200 Switzer Ave.
St. Louis, Mo. 63147
W.E. Zimmie, Inc.
810 Sharon Dr.
Westlake, Ohio 44145
Sink-Floe
Zeta-Floc
Floe-Aid
Natron
O'B Floe
Oxford-Hydro-Floc
Aquarid
Tychen
Hamaco
Hallmark
Jaguar
Polyhall
Swift
Varco-Floc
Zimmite
5-55
-------
5.6.1.2 General Design Considerations
Dry Polymer and water must be blended and mixed to obtain a recommended solution for
efficient action. Solution concentrations vary from fractions of a percent up. Preparation
of the stock solution involves wetting of the dry material and usually an aging period prior
to application. Solutions can be very viscous, and close attention should be paid to piping
size and length and pump selections. Metered solution is usually diluted just prior to
injection to the process to obtain better dispersion at the point of application.
5.6.1.3 Storage
General practice for storage of bagged dry chemicals should be observed. The bags should
be stored in a dry, cool, low humidity area and used in proper rotation, i.e., first in, first
out.
Solutions are generally stored in type 316 stainless steel, FRP, or plastic lined tanks.
5.6.1.4 Feed Equipment
Two types of systems are frequently combined to feed polymers. The solution preparation
system includes a manual or automatic blending system with the polymer dispensed by
hand or by a dry feeder to a wetting jet and then to a mixing-aging tank at a controlled
ratio. The aged polymer is transported to a holding tank where metering pumps or rotodip
feeders dispense the polymer to the process. A schematic of such a system is shown by
Figure 5-13. It is generally advisable to keep the holding or storage time of polymer
solutions to a minimum, 1 to 3 days or less, to prevent deterioration of the product.
5.6.1.5 Piping and Accessories
Selection must be made after determination of the polymer, however, type 316 stainless
steel or plastics are generally used.
5.6.1.6 Pacing and Controls
Controls as listed for liquid alum apply to the control of liquid dispersing feeders.
The solution preparation system may be an automatic batching system, as shown by the
schematic on Figure 5-14, that fills the holding tank with aged polymer as required by level
probes. Such a system is usually provided only at large plants. Prepackaged solution
preparation units are available, but have a limited capacity.
5-56
-------
WATER SUPPLY-
— DRY
FEEDER
DISPERSER
MIXER
DISSOLVING-AGING
TANK
-HOLDING TANK
-SOLUTION FEEDER
POINT OF
APPLICATION
FIGURE 5-13
TYPICAL SCHEMATIC OF A DRY POLYMER
FEED SYSTEM
5-57
-------
c/1
I
oo
HOT
WATER
SOLENOID
/ VALVE
.SCALE
COLD,
WATER
BLENDER
NOTE: CONTROL & INSTRUMENTATION
WIRING IS NOT SHOWN
SOLUTI ON
FEEDERS
POINT OF APPLICATION
MIXER
DISPERSER
LEVEL
PROBE
LEVEL PROBE
MIXER
MIXING-AGING
TANK
MIXING-AGING
TANK
-oo-
TRANSFER PUMP
O—LEVEL PROBE
HOLDING TANK
FIGURE 5-14
TYPICAL AUTOMATIC DRY POLYMER FEED
SYSTEM
-------
U.S EPA Headquarters Librai.
Mail code 3404T
5.6.2 Liquid Polymers 1200 Pennsylvania Avenue NW
Washington, DC 20460
5.6.2.1 Properties and Availability 202-566-0556
As with dry polymers, there is a wide variety of products, and manufacturers should be
consulted for specific information.
5.6.2.2 General Design Considerations
Liquid systems differ from the dry systems only in the equipment used to blend the polymer
with water to prepare the solution. Liquid solution preparation is usually a ha.nd batching
operation with manual filling of a mixing-aging tank with water and polymer.
5.6.2.3. Feed Equipment
Liquid Polymers need no aging and simple dilution is the only requirement for feeding. The
dosage of liquid polymers may be accurately controlled by metering pumps or rotodip
feeders.
The balance of the process is generally the same as described for dry polymers.
5.7 Chemical Feeders
Chemical feed systems must be flexibly designed to provide for a high degree of reliability
in light of the many contingencies which may affect their operation. Thorough waste
characterization in terms.of flow extremes and chemical requirements should precede the
design of the chemical feed system. The design of the chemical feed system must take into
account the form of each chemical desired for feeding, the particular physical and chemical
characteristics of the chemical, maximum waste flows and the reliability of the feeding
devices.
In suspended and colloidal solids removal from wastewaters the chemicals employed are
generally in liquid or solid form. Those in solid form are generally converted to solution or
slurry form prior to introduction to the wastewater stream; however, some chemicals are
fed in a dry form. In any case, some type of solids feeder is usually required. This type of
feeder has numerous different forms due to wide ranges in chemical characteristics, feed
rates and degree of accuracy required. Liquid feeding is somewhat more restrictive,
depending mainly on liquid volume and viscosity.
The capacity of a chemical feed system is an important consideration in both storage and
feeding. Storage capacity design must take into account the advantage of quantity purchase
versus the disadvantage of construction cost and chemical deterioration with time (13).
Potential delivery delays and chemical use rates are necessary factors in the total picture.
Storage tanks or bins for solid chemicals must be designed with proper consideration of the
angle of repose of the chemical and its necessary environmental requirements, such as
temperature and humidity. Size and slope of feeding lines are important along with their
materials of construction with respect to the corrosiveness of the chemicals.
5-59
-------
Chemical feeders must accommodate the minimum and maximum feeding rates required.
Baker (13) indicates that manually controlled feeders have a common range of 20:1, but this
range can be increased to about 100:1 with dual control systems. Chemical feeder control
can be manual, automatically proportioned to flow, dependent on some form of process
feedback or a combination of any two of these. More sophisticated control systems are
feasible if proper sensors are available. If manual control systems are specified with the
possibility of future-automation, the feeders selected should be amenable to this conversion'
with a minimum of expense. An example would be a feeder with an external motor which
could easily be replaced with a variable speed motor or drive when automation is installed!
(13). Standby or backup units should be included for each type of feeder used. Reliability
calculations will be necessary in larger plants with a greater multiplicity of these units.
Points of chemical addition and piping to them should be capable of handling all possible
.changes in dosing patterns in order to have proper flexibility of operation. Designed
flexibility in hoppers, tanks, chemical feeders and solution lines is the key to maximum
benefits at least cost (14).
Liquid feeders are generally in the form of metering pumps or orifices. Usually these
metering pumps are of the positive-displacement variety, plunger or diaphragm type. The
choice of liquid feeder is highly dependent on the viscosity, corrosivity, solubility, suction
and discharge heads, and internal pressure-relief requirements (10). Some examples are
shown in Figure 5-15. In some cases control valves and rotameters may be all that is
required. In other cases, such as lime slurry feeding, centrifugal pumps with open impellers
are used with appropriate controls (9). More complete descriptions of liquid feeder
requirements can be found in the literature and elsewhere (14).
Solids characteristics vary to a great degree and the choice of feeder must be considered
carefully, particularly in the smaller-sized facility where a single feeder may be used for
more than one chemical. Generally, provisions should be made to keep all chemicals cool
and dry. Dryness is very important, as hygroscopic (water absorbing) chemicals may
become lumpy, viscous or even rock hard; other chemicals with less affinity for water may
become sticky from moisture on the particulate surfaces, causing increased arching in
hoppers. In either case, moisture will affect the density of the chemical and may result in
under-feed. Dust removal equipment should be used at shoveling locations, bucket
elevators, hoppers and feeders for neatness, corrosion prevention and safety reasons.
Collected chemical dust may often be used.
The simplest method for feeding solid chemicals is by hand. Chemicals may be preweighed
or simply shoveled or poured by the bagful into a dissolving tank. This method is of
economic necessity limited to very small operations, or to chemicals used in very weak
solutions.
Because of the many factors, such as moisture content, different grades and compressibility,
which can affect chemical density (weight to volume ratio), volumetric feeding of solids is
normally restricted to smaller plants, specific types of chemicals which are reliably constant
in composition and low rates of feed. Within these restrictions several volumetric types are
5-60
-------
DISCHARGE VALVE
PLUNGER
PLUNGER PUMP (Courtesy of Wallace & Tiernan)
.SUCTION VALVE
DISCHARGE VALVE
SUCTION VALVE
-DIAPHRAGM
DIAPHRAGM PUMP (Courtesy of Wallace & Tiernan)
FIGURE 5-15
POSITIVE DISPLACEMENT PUMPS
5-61
-------
available. Accuracy of feed is usually limited to ± 2 percent by weight but may be as high
as ±15 percent.
One type of volumetric dry feeder uses a continuous belt of specific width moving from
under the hopper to the dissolving tank. A mechanical gate mechanism regulates the depth
of material on the belt, and the rate of feed is governed by the speed of the belt and /or the
height of the gate opening. The hopper normally is equipped with a vibratory mechanism to
reduce arching. This type of feeder is not suited for easily fluidized materials.
Another type employs a screw or helix from the bottom of the hopper through a tube
opening slightly larger than the diameter of the screw or helix. Rate of feed is governed by
the speed of screw or helix rotation. Some screw-type designs are self-cleaning, while others
are subject to clogging. Figure 5-16 shows a typical screw-feeder.
Most remaining types of volumetric feeders generally fall into the positive-displacement
category. All designs of this type incorporate some form of moving cavity of a specific or
variable size. In operation, the chemical falls by gravity into the cavity and is more or less
fully enclosed and separated from the hopper's feed. The size of the cavity, and the rate at
which the cavity moves and is discharged, governs the amount of material fed."The positive
control of the chemical may place a low limit on rates of feed. One unique design is the
progressive cavity metering pump, a non-reciprocating type. Positive-displacement feeders
often utilize air injection to improve the flow of the material. Some examples of
positive-displacement units are illustrated in Figure 5-17.
The basic drawback of volumetric feeder design, i.e., its inability to compensate for changes
in the density of materials, is overcome by modifying the volumetric design to include a
gravimetric or loss-in-weight controller. This modification allows for weighing of the
material as it is fed. The beam balance type measures the actual mass of material. This is,
considerably more accurate, particularly over a long period of time, than the less common
spring-loaded gravimetric designs. Gravimetric feeders are used where feed accuracy of
about 1 % is required for economy, as in large scale operations and for materials which are
used in small, precise quantities. It should be noted, however, that even gravimetric feeders
cannot compensate for weight added to the chemical by excess moisture. Many volumetric
feeders may be converted to loss-in-weight function by placing the entire feeder on a
platform scale which is tared to neutralize the weight of the feeder.
Good housekeeping and need for accurate feed rates dictate that the gravimetric feeder be
shut down and thoroughly cleaned on a regular basis. Although many of these feeders have
automatic or semi-automatic devices which compensate to some degree for accumulated
solids on the weighing mechanism, accuracy is affected, particularly on humid days when
hygroscopic materials are fed. In some cases, built-up chemicals can actually jam the
equipment.
No discussion of feeders is complete without at least passing reference to dissolvers, as any
metered material must be mixed with water to provide a chemical solution of desired
strength. Most feeders, regardless of type, discharge their material to a small dissolving
5-62
-------
FIGURED-1#
SCREW FEEDER
MOTOR AND
GEAR REDUCER
SOLUTION
CHAMBER
HOPPER
ROTATING &
RECIPROCATING
FEED SCREW
SOLUTION
LEVEL
JET MIXER
' FIGURE 5-17
POSITIVE DISPLACEMENT SOLID
FEEDER—ROTARY (15)
5-63
-------
tank which is equipped with a nozzle system and/or mechanical agitator depending on the
solubility of the chemical being fed. Solid materials, such as polyelectrolytes, may be
carefully spread into a vortex spray or washdown jet of water immediately before entering.
the dissolver. It is essential that the surface of each particle become thoroughly wetted
before entering the feed tank to ensure accurate dispersal and to avoid clumping, settling or
floating.
A dissolver for a dry chemical feeder is unlike a chemical feeding mechanism, which by
simple adjustment and change of speed can vary its output tenfold. The dissolver must be
designed for the job to be done. A dissolver suitable for a rate of 10 lb/ hr may not be
suitable for dissolving at a rate of 100 Ib/hr. As a general rule, dissolvers may be oversized,
but dissolvers for commercial ferric sulfate or lime slakers do not perform well if greatly
oversized.
It is essential that specifications for dry chemical feeders include specifications on dissolver
capacity. A number of factors need to be considered in designing dissolvers of proper
capacity. These include detention times and water requirements, as well as other factors
specific to individual chemicals.
The capacity of a dissolver is based on detention time, which is directly related to the
wettability or rate of solution of the chemical. Therefore, the dissolver must be large enough
to provide the necessary detention for both the chemical and the water at the maximum rate
of feed. At lower rates of feed, the strength of solution or suspension leaving the dissolver
will be less, but the detention time will be approximately the same unless the water supply
to the dissolver is reduced. When the water supply to any dissolver is controlled for the
purpose of forming a constant strength solution, mixing within the dissolver must be
accomplished by mechanical means, because sufficient power will not be available from the
mixing jets at low rates of flow. Hot water dissolvers are also available in order to
mimimize the required tankage.
The foregoing descriptions give some indication of the wide variety of materials which may
be handled. Because of this variety, a modern facility may contain any number of a variety
of feeders with combined or multiple materials capability. Ancillary equipment to the
feeder also varies according to the material to be handled. Liquid feeders encompass a
limited number of design principles which account for density and viscosity ranges. Solids
feeders, relatively speaking, vary considerably due to the wide range of physical and
chemical characteristics, feed rates and the degree of precision and repeatability required.
Table 5-11 describes several types of chemical feeders commonly used in wastewater
treatment.
5-64
-------
Type of Feeder
Dry feeder:
Volumetric:
Oscillating plate .
Oscillating throat (universal)
Rotating disc
Rotating cylinder (star) ....
Screw. .
Ribbon.
Belt . . .
Gravimetric:
Continuous—belt and scale
Loss in weight
Solution feeder:
Nonpositive displacement:
Decanter (lowering pipe) .. .
Orifice
Rotameter (calibrated valve)
Loss in weight (tank with
control valve).
Positive displacement:
Rotating dipper
Proportioning pump:
Diaphragm
Piston
Gas feeders:
Solution feed .
Direct feed.
TABLE 5-11 "
TYPES OF CHEMICAL FEEDERS
Use
Any material, granules or
powder.
Any material, any particle
size.
Most materials including NaF,
granules or powder.
Any material, granules or
powder.
Dry, free flowing material,
powder or granular.
Dry, free flowing material,
powder, granular, or lumps.
Dry, free flowing material up
to I'/z-inch size, powder or
granular.
Dry, free flowing, granular
material, or floodable
material.
Most materials, powder,
granular or lumps.
Most solutions or light slurries
Most solutions
Clear solutions
Most solutions
Most solutions or slurries ...
Most solutions. Special unit
for 5% slurries.1
Most solutions, light slurries.
Chlorine
Ammonia
Sulfur dioxide .
Carbon dioxide
Chlorine
Ammonia
Carbon dioxide
General
loader for
arching.
agitator to
maintain
constant
density.
No slurries . .
No slurries . .
Limitations
Capacity
cu ft/hr
001 to 35
002 to 100
001 to 1 0
8 to 2,000
or
7.2 to 300
0.05 to 18
0.002 to 0.16. . . .
0 1 to 3 000 ... .
0 02 to 2
0.02 to 80
0.01 to 10
0 1 6 to 5
0.005 to 0.1 6
or
0.01 to 20
0.002 to 0.20
0.1 to 30
0.004 to 0.15
0.01 to 170
8000 Ib/day max
2000 Ib/day max
7600 Ib/day max
6000 Ib/day max
300 Ib/day max
1 20 Ib/day max
10,000 Ib/day max
Range
40 to 1
40 to 1
20 to 1
10 to 1
or
100 to 1
20 to 1
10 to 1
10 to 1
or
100 to 1
100 to 1
100 to 1
100 to 1
10 to 1
10 to 1
30 to 1
100 to 1
100 to 1
20 to 1
20 to 1
20 to
20 to
20 to
10 to
7 to
20 to
1 Use special heads and valves for slurries.
5-65
-------
5.8 References
1. BIF, data sheets, Chemicals Used in Water and Wastewater Treatment-Ferric
Chloride, Wat. and Wastes Eng., Vol. 7, No. 3, pg. 65 (1970).
2. Ferric-Floe Jor wastewater treatment. Cities Service Co., Industrial Chemical
'Division, Atlanta, Ga. (1972).
3. Schworm, W.B., Iron Salts Jor Water and Waste Treatment, Public Works, Vol. 94,
No. 10, pg. 118(1963).
4. National Lime Association, Personal Communication (May, 1971).
5. Lime for Water and Waste-water Treatment, BIF Reference No. 1.21-24, BIF
Industries, Providence, Rhode Island (June, 1969).
6. Mulbarger, M.C., Grossman, E., Dean, R.B., and Grants O.L., Lime Clarification,
Recovery, Re-use and Sludge Dewatering Characteristics, JWPCF, Vol. 41, pg. 2070C
'(1969).
7. Caustic Soda, PPG Industries, Inc., Chemical Division, Pittsburgh, Pa. (1969).
8. Haney, P.D. and Hamann, C.L., Recarbonation and Liquid Carbon Dioxide,
JAWWA, Vol. 61, No. 10, pg. 512 (1969).
9. Gulp, R.L., and Gulp, G.L., Advanced Wastewater Treatment, Van
Nostrand-Reinhold Company, New York (1971).
10. Russo, F. and Carr, R.L., Polyelectrolyte Coagulant aids and Flocculants: Dry and
Liquid, Handling and Application, Wat. and Sew. Works Vol. 117, pg. R-72 (1970).
11. U.S. EPA OAWP, Report on Coagulant Aids for Water Treatment (July 1973).
12. Carr, R.L., Polyelectrolyte Coagulant Aids-Dry and Liquid Handling and
Application, Wat. and Sew. Works—Reference No., 114: R.N., p 4-64, (1967).
13. Baker, R.J., Chemical Feed Systems Determine Plant Efficiency and Reliability,
Water & Sew. Wks., Vol. 116, pg. R-21 (November 1969).
14. R.P. Lowe, Chemical Feed Systems, 10th Annual Water Conference of Eng Soc. of
W. Penna. (Oct. 17-19, 1949).
15. Design manual, civil engineering, Navdocks DM-5; Department of the Navy, Bureau
of Yards and Docks, Washington, D.C. (1972).
5-66
-------
CHAPTER 6
CHEMICAL MIXING, FLOCCULATION AND SOLIDS-
CONTACT PROCESSES
6.1 Introduction
Chemical mixing and flocculation or solids-contact are important mechanical steps in the
overall coagulation process described in Chapter 4. Application of the processes to waste-
water generally follows standard practices and employs basic equipment used for years in
the water-treatment field. Chemical mixing thoroughly disperses coagulants or their hy-
drolysis products so the maximum possible portion of influent colloidal and fine supracol-
loidal solids are absorbed and destabilized. Flocculation or solids contact processes increase
the natural rate of contacts between particles. This makes it possible, within reasonable de-
tention periods, for destabilized colloidal and fine supracolloidal solids to aggregate into
particles large enough for effective separation by gravity processes or media filtration.
All the processes discussed in this chapter depend on fluid shear for coagulant dispersal and
for promoting particle contacts. Shear is most commonly introduced by mechanical mixing
equipment. In certain solids-contact processes shear results from fluid passage upward
through a blanket of previously settled particles. Some designs have utilized shear resulting
from energy losses in pumps or at ports and baffles.
Numerous theoretical descriptions of the flocculation process have been developed (1) (2)
(3) (4) (5) (6). Almost all relate to experience in water treatment but can be applied to
wastewater coagulation with proper attention to significant differences in the nature of sol-
ids.
All theoretical approaches recognize the importance of time (t in sec.) and velocity gradient
(G, a measure of shear intensity in fps/ft or secrl) as controlling parameters in determining
performance of mixing and flocculation processes. It should be noted that chemical mixing
and flocculation differ only in intensity and duration and that some aggregation takes place
in the chemical mixing stage.
In addition to velocity gradient and time, expressions for the rate of aggregation in floccula-
tion or solids contact processes involve parameters reflecting the total volume and the size
and number of floe particles. When destabilized, particles in the fine colloidal range rapidly
aggregate under natural conditions to form small floes of fine supracolloidal size, about 1
micron diameter (5), often termed primary sized particles. In developing mathematical rela-
tions this is generally the assumed initial size of particles to be further aggregated.
The rate of aggregation is commonly taken as a function of the dimensionless product GCt
where C is the ratio of the volume of floe to total volume of suspension and G and t are as
defined above.
The floe volume concentration resulting from a given coagulant dosage depends, among
6-1
-------
other things, on the amount of water entrained in the floe. Hudson (7) and Camp (1) have
shown that more water is entrained and higher floe volumes result when flocculation takes
place at lower values of G.
The value of C may be increased greatly by recirculation of settled solids. This is used in
certain types of solids contact reactors (Section 6.4) and has been applied at Lake Tahoe as
part of a conventional coagulation system with separate rapid mix, flocculation and sedi-
mentation basins (8).
Design of rapid mix and flocculation units generally involves the choice of detention and G
value and selection of configurations, of mixing equipment, tanks, piping, etc. Unless the
designer provides for direct control of floe volume concentration through solids recircula-
tion, operating values of this parameter are determined indirectly through the chemical do-
sage and choice of G value and detention. Special attention should be given to avoiding ex-
cessive focalized shear and reducing short circuiting. Pretreatment should assure that waste-
water is free of debris (rags, sticks, etc.) which could damage mixing equipment. Special
considerations in design of solids-contact units are presented in Section 6.4.
G represents the root mean square velocity gradient (fps/ft) over the mixing basin. For me-
chanically-stirred basins it can be calculated from the relation:
P \^
G=' — 1 2
Where: P = power applied to stirring (ft-lb/sec = HP x 550)
V = reactor volume (cu ft)
•u = viscosity of fluid (Ib-sec/sq ft)
Viscosity varies with temperature as follows:
T
°C Ib-sec/sqft
1 0.361 x 10-4
5 0.316 x 10-4
10 0.273 x 10-4
15 0.239 x 10-4
20 0.210 x 10-4
25 0.187x 10-4
30 0.167x lO'4
Formulas for calculating G from head losses in baffled basins or in conduits are given by
Camp (9).
6-2
-------
6.2 Chemical Mixing
Chemical mixing facilities should be designed to provide a thorough and complete dispersal
of chemical throughout the wastewater being treated to insure uniform exposure to pollu-
tants which are to be removed.
The intensity and duration of mixing of coagulants with wastewater must be controlled to
avoid overmixing or undermixing.
Overmixing excessively disperses newly-formed floe and may rupture existing wastewater
solids. Excessive floe dispersal retards effective flocculation and may significantly increase
the flocculation period needed to obtain good settling properties. The rupture of incoming
wastewater solids may result in less efficient removals of pollutants associated with those
solids (2) (4).
Undermixing inadequately disperses coagulants resulting in uneven dosing. This in turn
may reduce efficiency of solids removal while requiring unnecessarily high coagulant do-
sages.
In water treatment practice several types of chemical mixing units have been used. These in-
clude high-speed mixers, in-line blenders and pumps, and baffled mixing compartments or
static in-line mixers (baffled piping sections). The high-speed mixer, as shown in Figure 6-1,
has been the most common choice for water treatment. Designs usually call for a 10 to 30
second detention time and approximately 300 fps/ft velocity gradiant (10). Hudson and
Wolfner (11) recommend variable-speed mixers to allow for varying requirements for opti-
mum mixing. In solids-contact reactors the G values in the immediate mixing zone approxi-
mate those for high-speed mixing (See section 6.4).
High speed mixers designed on the above basis should be equally satisfactory for waste-
water applications. Gulp, et al, (12) recommend providing two parallel units with a some-
what larger detention: 2 minutes at total design flow with both units. It has been question-
ed, however, whether in-line blenders (with G values as high as 5000 fps/ft) should be used
for wastewater in view of the possibility of rupturing organic solids (4). Where flows must
be pumped just prior to coagulation, addition of chemicals at the pumps is feasible. The
pump selection should take into account possible effects on organic solids of shear in centri-
fugal units. Where problems are anticipated, lower speed units such as screw pumps
should be used. Baffled compartments or in-line static mixing devices are limited in their ef-
fectiveness as chemical mixing devices whether in water or wastewater treatment because:
1. Head losses of up to 3 ft are required.
2. G cannot be changed to meet requirements, but rather is a function of flow rate
through the units.
In mineral addition to biological wastewater treatment systems, coagulants may be added
directly to mixed biological reactors such as aeration tanks or rotating biological con-
6-3
-------
DRIVE MECHANISM
MOTOR
SUPPORT BEAMS
OVERFLOW
IMPIUER
FEED
FIGURE 6-1
IMPELLER MIXER
6-4
-------
tactors.
Based on typical power inputs per unit tank volume, mechanical and diffused aeration
equipment and rotating fixed-film biological contactors produce average shear intensities
generally in the range suitable for chemical mixing. Parker (13) indicated that an analysis of
datajfor 14 activated sludge plants revealed that G ranged'from 88 to 220 fps/ft with an aver-
age of 136 fps/ft. Localized maximum shear intensities vary widely depending on speed of
rotating equipment or on bubble size for diffused aeration. Camp (9) presented bases for
relating localized maximum shear intensities to bubble size in diffused aeration. For fine
bubble diffusion (1.5 mm bubbles) maximum intensity reaches 1500 fps/ft with higher val-
ues for coarse bubbles.
No similar development has been located for rotating mechanical aerating equipment, but it
appears that maximum localized shears range from little more than the basin mean value
for large, low-speed devices such as rotating biological contactors, to perhaps as high as 50
times the mean for high speed (1800 rpm) mechanical aerators. Questions have been raised
about detrimental effects of high speed aerators on settling of activated sludge.
When using polymers, manufacturer's recommendations should be sought on the mixing
conditions which optimize their effectiveness, and these should be supplemented by jar tests,
if possible. When coagulant aids are employed, provisions for multiple addition points
should be made at the rapid mixing basin and in the flocculator to optimize the perform-
ance of the coagulant.
6.3 Flocculation
The proper measure of flocculation effectiveness is the performance of subsequent solids
separation units in terms of both effluent quality and operating requirements, such as filter
backwash frequency. Effluent quality depends greatly on the reduction of residual primary
size particles during flocculation, while operating requirements relate more to the floe vol-
ume applied to separation units.
For water treatment using alum or iron coagulants and flow-through flocculation (as op-
posed to solids-contact units) traditional designs have been based on G* of up to 100 fps/ft
andiGt values of 0.3 to 1.5 x 10$ (10) and GCt values of 10-100 (3). The wide ranges of
these parameters may reflect genuine differences between waters (or wastewaters) but may
also reflect different design approaches. Hudson (7) has suggested use of Gt values of 2 x
105 which he claimed would produce high density floe with settling velocities equivalent to
those of larger lower density floe produced at low G values. Camp (1) has suggested use of
higher G values and resulting lower floe volumes to get equivalent primary particle agglo-
meration but with lower solids loadings on subsequent separation units.
Values in the ranges above are certainly ample for wastewater flocculation in flow through
units. Because of the larger coagulant doses commonly used in wastewater treatment (espe-
cially with phosphorus removal) detention times and Gt values can generally be lower. Gulp
et al (12) recommend a maximum of 15 minutes detention for wastewater coagulation. Gulp
and Gulp (8) recommend using paddle speeds of 0.5 to 1.0 fpm.
6-5
-------
Tapered flocculatiqn in which the flow is exposed to decreasing G values as it passes
through the unit, can provide a rapid build-up of small dense floe with subsequent agglome-
ration at lower 73 into larger but still dense particles. (9) (10) (11). Use of mul-
ti-compartment flocculators not only permits tapered flocculation, but also greatly reduces
the high short-circuiting associated with single-compartment units. A wide variety of physi-
cal layouts are possible to achieve series flow through multiple compartments (10).
Flocculation units should have multiple compartments and should be equipped with adjust-
able speed mechanical stirring devices to permit meeting changed conditions. In spite of
simplicity and low maintenance, non-mechanical, baffled basins are undesirable because of
inflexibility, high head losses and large space requirements.
Mechanical flocculators may consist of rotary, horizontal-shaft reel units as shown in Fig.
6-2, rotary vertical shaft turbine units as shown in Fig. 6-3 or other rotary or recipro-
cating equipment. Features of these various type units are discussed and compared else-
where (9) (10) (11).
Tapered flocculation may be obtained by varying reel or paddle size on horizontal common
shaft units or by varying speed on units with separate shafts and drives. A typical series of G
values for successive compartments would.be 100, 50 and 20 fps/ft. In most cases, equip-
ment should provide overall Gt values up to 2 x 105 at maximum drive speed. Speed vari-
ation over a range of 1:3 or 1:4 should be possible (10).
.G values are determined from the hp actually transmitted to the fluid (water hp). This
should be distinguished from the total input hp which includes losses in motors, drives,
bearings, etc. It should be noted that G is a mean value for the entire flocculator volume.
Practical limits are set to localized high values at the flocculator blades or paddles by speci-
fying peripheral speeds below about 2 fps.
. In applications other than coagulation with alum or iron salts, flocculation parameters may
be quite different. Lime precipitates are granular and benefit little from prolonged floccula-
tion or very low terminal G values. At Lake Tahoe a detention of 4.5 min. proved adequate.
Gulp and Gulp (8) recommend a minimum of 5 min. but as much as 10 min may be needed
to assure complete dissolution and reaction of CaO. As in water softening practice, G val-
ues of 100 or more are desirable.
Polymers which already have a long chain structure may provide a good floe at low Gt val-
ues. Often the turbulence and detention in the clarifier inlet distribution is adequate.
Settling and effluent clarity in the activated sludge process can frequently be improved by'
controlled flocculation between the aeration tank and clarifier. Parker, et al, (14) showed
that flocculation at G = 40 to 60 fps/ft and detention of 20 to 30 min. could reduce the SS
in aerator effluent (after settling) by some 45 to 55 percent. The benefits of flocculation de-
pend on the level of turbulence in the aerator, and on the sludge age which affects the natu-
ral flocculating characteristics of the sludge. In the above study sludge with a sludge age of
10 days was better destabilized and benefited more from flocculation than did sludge with a
6-6
-------
carrot «um
JL
n
_u u_
1! K
II II II
J
FIGURE 6-2
MECHANICAL FLOCCULATION BASIN
HORIZONTAL SHAFT-REEL TYPE
MOTORIZED SPEED REDUCER
GUIDE BEARING
WATER PRESSURE LUBRICATED
'FIGURE 6-3
MECHANICAL FLOCCULATOR
VERTICAL SHAFT—PADDLE TYPE
(Courtesy of Ecodyne Corp.)
6-7
-------
sludge age of 3 to 4 days or 12 days.
This behavior may be interpreted in light of the observation by Dean (15) that activated
sludge contains an excess of natural anionic polymers. As sludge age increases these pol-
ymers are reduced—first to levels where destabilization is better—but then to levels below
the optimum.
6.4 Solids-Contact Processes
Solids-contact processes combine chemical mixing, flocculation and clarification in a single
unit designed so that a large volume of previously-formed floe is retained in the system. The
floe volume may be as much as 100 times that in a "flow-through" system. This greatly in-
creases the rate of agglomeration from particle contacts (11), and may also speed up chem-
ical destabilization reactions.
Solids contact units are of two general types: slurry-recirculation and sludge-blanket. In the
former, the high floe volume concentration is maintained by recirculation from the clari-
fication to the flocculation zone, as illustrated in Fig. 6-4. In the latter, the floe solids are
maintained in a fluidized blanket through which the wastewater under treatment flows up-
ward after leaving the mechanically stirred-flocculating compartment, as depicted in Fig.
6-5. Some slurry-recirculation units can also.be operated with a sludge blanket.
Solids-contact units have become popular in water treatment and are being increasingly
considered in advanced wastewater treatment because of the following advantages:
1. Reduced size and lower cost result because flocculation proceeds rapidly at high
floe volume concentration.
2. Single-compartment flocculation is practical because high reaction rates and the
slurry effects overcome short circuiting.
3. Units are available as compact single packages, eliminating separate units.
4. Even distribution of inlet flow and the vertical flow pattern in the clarifier improve
clarifier performance (16).
Equipment typically consists of concentric circular compartments for mixing, flocculation
and settling. Velocity gradients (G) in the mixing and flocculation compartments are devel-
oped by turbine pumping within the unit and by velocity dissipation at baffles. For ideal
flexibility it is desirable to independently control intensity of mixing (G) and sludge scraper
drive speed in the different compartments. Ives (3) indicates that slurry-recirculation solids
contact reactors in the water treatment field operate with a velocity gradient in the range of
60 to 120 fps/ft. Hudson and Wolfner (11) indicate that in water treatment solid-contact re-
actors.with variable-speed turbine-type agitators apply velocity gradients of 300 fps/ft in
the mixing zone while reaction zone values may vary from 100 fps/ft near entrance to 20
fps/ ft at the settling zone boundary. Comparably proportioned units are being used in
6-8
-------
RAPID MIXING AND RECIRCULATION
SLOW MIXING AND FLOC FORMATION
CHEMICAL INTRODUCTION
TREATED WATER
EFFLUENT
ON
CLEAR WATER
SEPARATION
SLUDGE RECIRCULATION
CLARIFIED
WATER
RAW WATER
INFLUENT
SEDIMENTATION
SLUDGE REMOVAL
FIGURE 6-4
SOLIDS CONTACT CLARIFIER WITHOUT SLUDGE BLANKET FILTRATION
(Courtesy of Ecodyne Corp.)
-------
EFFLUENT COLLECTOR FLUME
,
!
AGITATC
' f^ |i
r i] •/ '-j
!| ' ' "
..'.., „ , il. - *+-.£.-^ ,
R
CHEMICAL fflrO INltTS
f . r
. INFIUFNT
>dfflS§3
jji^*1*
SKIMMING
SLOT
• »»' 00000 OO O O
o ooooooooo»«i
SLUDGE
BLOW OFF
LINE
SAMPLE CONNS.J
AGITATOR
ARM
SLUDGE
CONCENTRATOR
SWING SAMPLE
INDICATOR -
PRECIPITATOR DRAIN
FIGURE 6-5
SOLIDS CONTACT CLARlFlER WITH SLUDGE BLANKET FILTRATION
(Courtesy of the Permutit Co.)
-------
wastewater treatment, but with little explicit consideration of G values.
Experience with solids-contact units in wastewater treatment has up to now been limited to
slurry-recirculation units. Gulp and Gulp (8) have expressed concerns about the use of
sludge-blanket units: septicity and uncontrolled blanket upsets under varying-load condi-
tions. Slurry-recirculation units not requiring sludge blankets or with minimum blanket
depths are not very sensitive to such upsets. Units equipped with scrapers have operated
without septicity problems treating secondary effluent at Nassau County, N.Y. (17) and at
Ely, Min. (18).
Operation of slurry-recirculation solids-contact units is typically controlled by maintaining
steady levels of solids in the reaction zone. For lime treatment of wastewater at Ely and
Blue Plains a solids concentration of 10 to 12 percent by weight was found most effective
(for phosphorus removal) (18) (19). For tertiary alum treatment at Nassau Gounty 45 per-
cent floe volume concentration proved most satisfactory (17).
Design features of solids-contact clarifiers should include:
1. Rapid and complete mixing of chemicals, feedwater and slurry solids must be pro-
vided.This should be comparable to conventional flash mixing capability and
should provide for variable control of Gt values, usually by adjustment of recir-
culator speed.
2. Mechanical means for controlled circulation of the solids slurry must be provided
with at least a 3:1 range of speeds. The maximum peripheral speed of mixer blades
should not exceed 6 ft/sec. Rushton and Mahoney (20) offer means of estimating
pumping capacity of mixers.
3. Means should be provided for measuring and varying the slurry concentration in
the contacting zone up to 50 percent by volume.
4. Sludge discharge systems should allow for easy automation and variation of vol-
umes discharged. Mechanical scraper tip speed should be less than 1 fpm with
speed variation of 3:1.
5. Sludge-blanket levels must be kept a minimum of 5 feet below the water surface.
6. Effluent launders should be spaced so as to minimize the horizontal movement of
clarified water.
Most'Of the above requirements are based on those cited in Water Treatment Plant Design
(10). Further considerations include skimmers and weir overflow rates. Skimmers should be
provided on all units since even secondary effluents contain some floatable solids and
grease. Overflow rates and sludge scraper design should conform to the requirements of
other clarification units.
6-11
-------
6.5 References
1. Camp, T. R., Floe Volume Concentration, Jour. AWWA, 60, 656 (1968).
2. O'Melia, C. R., The Coagulation Process: Application oj Research to Practice, Re-
port submitted to ASCE (Oct. 1969).
3. Ives, K. J., Theory oj Operation of Sludge Blanket Clarijiers, Proc. ICE (Br.), 39, 243
(Feb. 1968).
4. Weber, W. J., Jr., Physicochemical Processes Jor Water Quality Control, John Wiley
& Sons, Inc., New York (1972). '
5. Harris, H. S., et al, Orthokinetic Flocculation in Water Purification, J. San Eng. Div.,
ASCE, 92, SA6, 95 (Dec. 1966).
6. Parker, D. S. et al, Floe Breakup in Turbulent Flocculation Processes, J. San. Eng.
Div. ASCE, 98, SA1, 79-99 (Feb. 1972).
7. Hudson, H. E., Jr., Physical Aspects of Flocculation, Jour. AWWA, 57, 885 (July,
1965).
8. Gulp, R. L. and Gulp, G. L., Advanced Waste-water Treatment, Van Nostrand Rein-
hold Company, New York (1971).
9. Camp, T. R., Flocculation and Flocculation Basins, Trans. ASCE, 120, (1955).
10. Water Treatment Plant Design, American Water Works Ass'n, Inc., New York
(1969).
11. Hudson, H. E., Jr., and Wolfner, J. P., Design oj Mixing and Flocculating Basins,
Jour. AWWA, 59,1257 (Oct. 1967).
12. Gulp, G. L., et al., Physical-Chemical Wastewater Treatment Plant Design.
U.S. EPA Technology Transfer Seminar Publication (August 1973).
13. Parker, D. S., Effect of Turbulence on Activated Sludge Effluent Clarity, Paper
Presented at 12th Annual Northern Reginal Conference, Calif. WPCA, (Oct. 1970).
14. Parker, et al, Physical Conditioning of Activated Sludge Floe, JWPCF, 43, 9, pg.
1817, (Sept. 1971).
15. Dean, R. B., Colloids Complicated Treatment Processes, Environmental Science &
Tech. pg. 820, (Sept. 1969).
6-12
-------
16. Aitken, I.M.E., Reflections on Sedimentation Theory and Practice—Part I, Eff, and
Water Treatment Jour. (Br.), 74, 226 (Apr. 1967).
17. Oliva, J. A. Department of Public Health, County of Nassau, Personal Commu-
nication. (March 1973).
18. Westrick, J. J., U.S. EPA, NERC, Cincinnati, Ohio, Personal communication (April,
1973).
19. Kreissl, J. F., U.S.E.P.A., National Environmental Research Center, Cincinnati,
,Ohio, Personal Communication, (April 1973).
20. Rushton, J. H. and Mahoney, L. H., Mixing Power & Pumpage Capacity, Annual
Meeting of AIME, New York (Feb. 15, 1954).
6-13
-------
CHAPTER 7
GRAVITY SEPARATION
7.1 Introduction
Gravity separation refers to the removal of SS whose specific gravity difference from that
of water causes them to settle or rise during passage through a tank or basin under
quiescent conditions. Separation by settling is termed sedimentation; separation by rising is
termed flotation. The size of particles determines the fluid drag retarding this separation.
For a given specific gravity, smaller particles having greater surface area encounter more
drag and hence are more difficult to separate.
The factors affecting separation efficiency are discussed in depth for sedimentation, and
separate sections cover each of its major applications. The section on flotation indicates
special considerations pertaining to this process and deals with its applications. Finally two
sections deal with devices which enhance the performance of sedimentation basins.
7.2 Configuration of Sedimentation Units
The tanks or basins in which sedimentation is carried out (also frequently termed clarifiers)
may be classified as horizontal flow or vertical-flow according to the predominant direction
of the flow path from inlet to outlet. It should be noted that, depending on placement of
inlet? and outlets, certain designs—particularly small radial flow tanks—may have a flow
path with significant components in both horizontal and vertical directions.
7.2.1 Vertical-Flow Units
Vertical-flow applications in the U.S. have generally been limited to settling compartments
in flocculation-clarifiers, solids-contact units and activated sludge systems of similar
configuration (Aero-Accelator, Rapid Block, etc.). In Europe, vertical-flow basins have
been used extensively. Kalbskopf has illustrated several European designs (1).
Vertical-flow units may be annular or rectangular, and are generally narrower at the
bottom than at the top. In annular designs, the flow is distributed at the bottom along the
circumference of the tank and rises to peripheral or radial effluent weirs or launders. Flow
in rectangular tanks is distributed at the bottom along the length of the tank and rises to
longitudinal or transverse effluent weirs or launders.
Annular units have been built with outside diameters to 150 ft, but the width from inner
wall to outer wall is much less. Figures 6-4 and 6-5 illustrate annular, vertical-flow units.
7.2.2 Horizontal-Flow Units
In the U.S. horizontal-flow units, both rectangular and circular, are most often used for
7-1
-------
sedimentation applications. Tank proportions, inlet and outlet arrangements and types of
sludge and scum collecting equipment are summarized and discussed in the ASCE/WPCF
Manual for Sewage Treatment Plant Design (2). Individual bays of rectangular tanks
should have a length to width ratio of at least four.
Flow through rectangular tanks enters at one end, passes a baffle arrangement, and
traverses the length of the tank to effluent weirs. In narrow tanks, longitudinal collectors
scrape sludge to single or multiple hoppers at one end (Figure 7-1). In tanks with multiple
wide bays, the longitudinal collectors scrape sludge to a cross collector which then moves
the sludge to a central hopper. Circular designs employ three inlet/ outlet configurations
with corresponding flow paths as shown in Figure 7-2. In configurations 7-2 (a) and (c),
sludge is removed by mechanical scraping to a central hopper or draw-off. In configuration
7-2 (b) a hydraulic suction sludge removal system is employed.
7.3 Basic Factors Affecting Settling Tank Design
7.3.1 Hydraulic Loading
The basic parameter to which settling tank performance is related is the surface hydraulic
loading (Q/A). This is the inflow (Q) divided by the surface area (A) of the basin, and is
commonly expressed in units of gpd/sq ft.
Hazen (3) showed that under the following assumptions performance is a function of
surface loading alone:
1. Quiescent or non-turbulent flow
2. Uniform distribution of velocity over all sections normal to general flow direction
3. Discrete non-interacting particles
. 5 A.
4. No resuspension of particles once they reach the floor of the basin
Under these conditions all particles whose settling velocity (Vs) exceeds Q/A are removed.
In addition, in horizontal flow tanks particles of lower settling velocities are partially
removed in the proportion Vs/(Q/A).
In actual basins conditions depart in many respects from those assumed in Hazen's original
analysis. The most significant of these departures are:
1. Currents induced by inlets, outlets, wind and density differences may cause short
circuiting or dead spaces within the tank.
2. Turbulence due to forward velocity or currents in the tank retards settling.
3. Flocculent solids may agglomerate into larger particles during passage through the
.basin.
7-2
-------
INFLUENT
DRIVE SPROCKET
-ADJUSTABLE WEIRS
.WATER LEVEL
. RECESS FOR
/ DRIVE CHAIN
FLOW
SKIMMING
CHAIN a FLIGHT
CROSS COLLECTOR
SLUDGE HOPPER
AVERAGE
WATER
DEPTH
J I J -—^t^
f T IT- :-^r4
D
EFFUJENT
2"x 6" FLIGHTS
PIVOTING FLIGHT-1
A. WITH CHAIN AND FLIGHT COLLECTOR
TRAVELING
BRIDGE
BRIDGE
TRAVEL
COLLECTING •• 'HMYCL. ^ SKIMMING
SCUM
TROUGH
WATER LEVEL
SLUDGE
DRAWOFf
SLUDGE
HOPPER
SKIMMING POSITION
SLUOOE COLLECTION POSITION
- SCREW CROSS
COLLECTOR
B. WITH TRAVELING BRIDGE COLLECTOR
FIGURE 7-1
RECTANGULAR SEDIMENTATION TANKS
(Courtesy of FMC'Corp!)
7-3
-------
EFFLUENT
SLUDGE
INFLUENT
(a)CIRCULAR CENTER-FEED CLARIFIER WITH
A SCRAPER SLUDGE REMOVAL SYSTEM
INFLUENT
EFFLUENT
SLUDGE
(b)CIRCULAR RIM-FEED, CENTER TAKE-OFF CLARIFIER WITH A
HYDRAULIC SUCTION SLUDGE REMOVAL SYSTEM
-i-
jrf
1±
^g
INFLUENT
EFFLUENT
I
SLUDGE
(clCIRCULAR RIM-FEED, RIM TAKE-OFF CLARIFIER
FIGURE 7-2
TYPICAL CLARIFIER CONFIGURATIONS
7-4
-------
4. Sludge may be scoured and resuspended at high forward velocities.
5. When influent solids concentrations are high, particles settle as a mass rather than
discretely.
The subsections below indicate how investigators, most notably Camp (4), have attempted
to account for these departures by relating performance to additional parameters. The
relationships are not generally adequate to permit prediction of performance from design
values of the parameters, but they do provide insights helpful in deciding a number of tank
features such as shape, depth, inlet type, etc. In addition, such relationships offer guidance
in translating settling test results into sizing for full scale tanks. Procedures for conducting
and interpreting such tests have been outlined by O'Connor and Eckenfelder (5) and others
(6) (7) (8).
To account for departures of full scale tanks from ideal or test conditions scale-up factors in
the following ranges have been suggested (5):
Sizing Parameter Scale-Up Factor
•"••"Area" 1.25 to 1.75
Volume 1.5 to 2.0
These scale-up factors are not intended to cover extreme variations in flows or solids
loadings, or to allow for operation at temperatures significantly different from those in the
tests. Neither do they include standby capacity as needed for units critical to overall plant
performance. Smith (9) has discussed the use of excess capacity factors to provide for
standby and to cover expected variations in loadings.
7.3.2 Short Circuiting
Short circuiting can greatly reduce the removal efficiency of a settling tank. Effects are
most critical for flocculent suspensions whose removal is affected by detention time (Sec.
7.3.4), but depending on the current pattern, removal of discrete particles may also be
affected. Short circuiting is accentuated by high inlet velocities, high outlet weir rates, close
placement of inlets and outlets, exposure of tank surface to strong winds, uneven heating of
tank contents by sunlight, and density differences between inflow and tank contents.
Density-induced short circuiting can be a significant factor in secondary settling tanks
handling activated sludge mixed liquor (10). Inlet and outlet conditions, tank geometry, and
density differences due to influent SS concentrations produce steady short circuiting,
whereas effects of other factors are generally intermittent and unpredictable.
The degree of short circuiting can be measured using tracer studies. Figure 7-3 shows results
of such studies on four types of settling tanks (11), where short circuiting was due primarily
to inlet and outlet conditions and tank geometry. Studies of this type have confirmed that
such short circuiting is minimized in narrow, rectangular, horizontal-flow tanks and is most
serious in circular horizontal flow tanks. Although upflow tanks show the least short
7-5
-------
r»
O
n
o
z
O
5
Radial flow in Circular
Tank.
Horizontal RowinWiifa
Rectangular Tank.
Horixpnfal Flow in Narrow
Roctonqulor Tank.
Vertical Flow in Upflow
Ctaiqn of Tank.
\oo'/t
PERCENTAGE OF NOMINAL DETENTION PERIOD
FIGURE 7-3
RESULTS OF SALT-INJECTION TESTS WITH v
DIFFERENT TYPES OF SEDIMENTATION TANKS
-------
circuiting, practical problems in obtaining uniform initial flow distribution have limited their
use to small diameter units.
The degree of short circuiting in circular units can vary considerably, however, depending
on the type of inlet used. Inlet conditions have been shown to be more critical than those at
the outlets (12). For activated sludge final settling tanks, peripheral feed and certain
special-design center feed inlets have been shown to cause less short circuiting than
conventional center feed inlets (10) (13) (14).
Even where the degree of short circuiting can be measured or predicted, techniques for
evaluating the effect on tank performance (1) (15) are questionable as to their utility and
accuracy. Hence, the best design approach is to avoid short circuiting as far as possible,
thus minimizing uncertainty as to its effects. The most important factors to consider in
controlling short circuiting are dissipation of inlet velocity, protection of tanks from wind
sweep and uneven heating, and reduction of density currents associated with high inlet SS
concentrations (13).
Such density current short curcuiting is a particular problem in settling tanks for activated
sludge. Fitch (10) has presented estimates of the velocities of such currents as a function of
SS concentration, and has compared two fundamental approaches to preventing short
circuiting from this source. These are dynamic stabilization as proposed by Camp (4) and
density stabilization. Dynamic stabilization requires shallow basins with high forward
velocities. (Froude numbers of 5 x 10 5 or greater). The resulting friction losses, in theory,
counteract stratification and instability of flow. Density stabilization essentially establishes
an upflow type pattern by introducing the dense feed at low velocities and close to the tank
bottom. Fitch showed that low inlet velocities are essential to successful density
stabilization, and proposed a novel center inlet design to achieve such velocities (see Section
7.4).
7.3.3 Turbulence
Turbulence levels in a settling basin are normally difficult to estimate. The only exception is
turbulence due to drag from net forward velocity. Camp (4) has presented a basis for
estimating turbulence from this source and for compensating for its effects by increasing
tank area. Required increases vary directly with forward velocity in the tank and with the
desired removal rate.
Good design practice is to minimize other sources of turbulence such as inlet, outlet, wind
and density currents. These sources produce unpredictable levels of turbulence and may
increase short circuiting. Even where the degree of turbulence during sedimentation can be
definitely measured the effect on removal of flocculent particles is not easily predicted,
because agglomeration induced by turbulence can alter particle sizes and localized settling
velocities.
7-7
-------
7.3.4 Particle Agglomeration
For the flocculent suspensions handled in wastewater treatment, particle contact and
agglomeration continues during sedimentation. Two mechanisms produce particle contacts:
velocity gradients within the settling tank, and differential settling rates; each of which
permit faster moving particles to overtake slower ones. Depending on the nature of the
influent suspension, either mechanism can significantly affect both the size and settling
velocity of floe and the fraction of fine, unsettleable particles remaining in suspension.
Regardless of surface loading on a settling tank, attachment of smaller, unsettleable
particles onto larger ones of separable size is essential in attaining high SS removal
efficiences. In any case, these larger particles must have the opportunity to agglomerate to
sizes which will be removed at the maximum surface loadings applied to the tank.
Otherwise massive failure of the separation process will occur with significant loss of SS in
the effluent.
Camp (4) asserted that the rate of particle contacts due to differential settling depends only
on the characteristics of the suspension. Fitch (16) (17) maintained that the rate also
increased with tank depth. In either case, the total number of contacts occurring due to
differential settling is a direct function of detention time, which at a given surface hydraulic
loading is, in turn, a function of settling tank depth: In contrast, the rate of particle contacts
due to velocity gradients increases with forward velocity and hence decreases with depth.
For the 10 to 15 ft tank depths normally used in wastewater treatment in the U.S.,
agglomeration depends mainly on differential settling. For wastewaters such as raw sewage,
which agglomerate slowly under differential settling, detention time can have a significant
effect on settling tank performance (See Section 7.5).
Camp (4) urged the use of much shallower settling tanks, theorizing that the higher velocity
gradients would accelerate particle agglomeration sufficiently to more than offset the
reduction in detention time. Fitch (16) disputed this noting that-in stirred settling tests
velocity gradients comparable to those proposed by Camp provided little flocculation. In
any case, common U.S. practice has remained to design fairly deep tanks with low forward
velocities (about 1 fpm at mean flow) and to depend on some other means than gradients
due to forward velocity to achieve desired flocculation. Kalbskopf (1) indicated that in
Europe it is common, to design shallower (3 to 10 ft depth) primary settling tanks with
higher forward velocities (2.5 fpm at mean flow and up to 7.5 fpm at maximum flow).
Studies for the Emscher Mouth treatment facility (18) showed only minor variation of
primary effluent SS with forward velocity. Performance related much more to surface
loading. For any given surface loading, however, the best performance was at a velocity in
the range between 1.6 and 2.5 fpm.
It is well recognized that increasing velocity gradients by stirring the inlet zone of a settling
tank can often improve performance (1) (4) (19). Essentially this combines mechanical
flocculation and settling in a single tank. Compartmentation is desirable to reduce short
circuiting. The major advantage of such combined units is that a suspension can be
flocculated at decreasing G values (see Section 6.1) down to very low levels and then
delivered to sedimentation without subjecting the suspension to the shearing effects of
7-8
-------
collection and redistribution. Recognizing this advantage, several equipment manufacturers
offer combined units designed on this basis. Where flocculation is to be used to upgrade
performance of existing settling tanks, the possibility of locating flocculation mechanisms
directly in the tanks should be considered. (See U.S. EPA, Process Design Manual for
Upgrading Existing Wastewater Treatment Plants).
7.3.5 Bottom Scour
Where high forward velocities are used, the possibility of scouring previously deposited
sludge should be analyzed. As a rule of thumb, forward velocities should be limited to from
9 to 15 times the settling velocity of critical size solids to avoid scour (20).
7.3.6 Hindered Settling and Compaction
When a concentrated suspension such as activated sludge mixed liquor settles under
quiescent conditions, a distinct interface develops almost immediately between the sludge
and the clarified liquid above it (21). As illustrated in Figure 7-4, this interface subsides for
a time at a constant rate. This rate is termed the initial settling velocity of the sludge.
Because the accompanying upward displacement of liquid reduces this settling velocity to
below that of discrete particles of the sludge, the process is termed hindered settling.
As the sludge mass continues to settle an interparticle structure develops in the more
concentrated lower layers and the subsidence rate slows further. Figure 7-5 illustrates this
compaction or thickening of sludge in a full scale tank. If high sludge concentrations are to
be obtained, thickening rather than solids separation may control the tank sizing. Sizing
secondary settling tanks for activated sludge to meet thickening requirements is discussed in
Section 7.6.
7.4 Clarifier Design Considerations
7.4.1 General
In selecting the particular tank shape, proportions, equipment, etc. the designer should:
1. Provide for even inlet flow distribution in a manner which minimizes inlet
velocities and short circuiting.
2. Minimize outlet currents and their effects by limiting weir loadings (see Sec. 7.5
and 7.6) and by proper weir placement.
3. Provide sufficient sludge storage depths to permit desired thickening of sludge.
4. Provide sufficient wall height to give a minimum of 18 inches of freeboard.
5. Reduce wind effects on open tanks by providing wind screens and by limiting fetch
of wind on tank surface with baffles, weirs or launders.
7-9
-------
CLEAR WATER ZONE
HINDERED SETTLING
A\ CONSTANT COMPOSITION
TRANSITION. ZONE
VARIABLE COMPOSITION
COMPRESSION ZONE
ULTIMATE CONCENTRATION
TIME
CYLINDER
FIGURE 7-4
SCHEMATIC REPRESENTATION OF SETTLING ZONES
o.
UJ
(9
UJ
-------
6. Consider economy of alternative layouts which can be expected to provide
equivalent performance.
7. Maintain equal flow to parallel units. This is most important and often forgotten.
Equal flow distribution between settling units is generally obtained by designing
equal resistances into parallel inlet flow ports or by flow splitting in symmetrical
weir chambers.
7.4.2 Inlet Design
Inlet design for rectangular tanks, where the distance from inlet to outlet is large, is less
critical than for circular tanks where there is generally little separation between inlet and
outlet.
In rectangular tanks flow is distributed over the width of the tank by provision of multiple
inlets. Size and spacing vary considerably from one design to another. Small openings are
avoided in wastewater applications because of the possibility of fouling. Maximum spacings
are generally less than 10 ft. Target baffles are commonly provided to help dissipate the
velocity of the inlet jets. Distribution to multiple inlets in a rectangular tank usually
involves a manifold conduit. A method, developed by Dobbins, for design of inlets and
manifold conduits, is presented elsewhere (22).
The common type of center feed for circular tanks depends on symmetrical baffling to
distribute flow equally in all radial directions. The high degree of short circuiting with such
inlets has led manufacturers to develop several special inlet designs for circular tanks—both
center and peripheral feed.
Figure 7-2b and c show peripheral feed units. In these units, inlet ports discharge outside a
deep peripheral baffle and flow passes under this baffle to enter the tank. In a peripheral
feed unit manufactured by Lakeside Equipment Corp., the inlet line to the tank discharges
tangentially into a tapered race located behind a similar skirt baffle. The manufacturer
claims that the tangential motion imparted to the tank contents reduces short circuiting. In
model studies, the latter type of peripheral unit showed significantly higher removals of iron
floe than a similarly loaded center feed unit (23). This was attributed to better conditions
for particle agglomeration in the peripheral feed model.
A center feed inlet manufactured by Dorr-Oliver, Inc: has two races with
tangentialiy-introduced flows rotating in opposite directions. Shear between these rotating
flows dissipates the energy of the inlet velocity before the inflow leaves the feedwell (10).
The modular Energy Dissipating (MED) Feedwell, manufactured by Envirotech Corp.,
forces all flow to pass through honeycombs of small tubes, mounted vertically around the
entire feedwell periphery. The manufacturer claims that the honeycomb creates a laminar
flow pattern with uniform radial velocities and that periodically reversing the modules on
their pivot mountings(changing flow direction through them) will clear the honeycomb of
any accumulated solids.
7-11
-------
Available data comparing performance of primary and secondary clarifiers using special
and conventional inlets are presented in Table 7-1.
7.4.3 Economy
The two major elements in settling tank cost are the structure and the sludge collection
mechanism. Installed cost for the mechanism is typically 30 to 40 percent of the structural
cost. Structural costs for multiple- rectangular and circular tanks (horizontal flow) are
comparable, provided common-wall construction is used for the rectangular units (1) and
liquid depths are not more than about 10 ft. At greater depths, circular units with tank walls
designed as hoops show increasing savings. Single circular units are less expensive than the
same size rectangular basin. Where tanks must be covered, costs may favor rectangular
units because of their shorter roof spans. European data (1) indicate structural costs for
vertical flow units may run 50 percent higher than horizontal units of the same volume, but
the vertical tanks have deep conical bottoms eliminating the need for costly sludge collector
mechanisms.
Rotary collectors for circular tanks generally cost 20 percent less than chain-and-flight
collectors for comparable rectangular units. In addition, maintenance requirements for the
rotary units are decidedly lower. Travelling-bridge collectors for rectangular tanks
apparently compete favorably with rotary circular collectors in cost and ease of
maintenance. They are common in Europe but until recently have not found widespread
application in the U.S. A recent comparison (27) for secondary tanks showed a floating
travelling bridge collector with siphon sludge drawoffs to be decidedly cheaper than either
chain-and-flight or circular mechanisms.
7.4.4 Skimming
In rectangular tanks with chain-and-flight collectors, skimmings are moved toward their
discharge point by return travel of the flights at the tank surface (Figure 7-1 A). In circular
tanks skimmings are moved by travel of a surface arm attached to the rotary collector
(Figure 7-2). A surface arm can be similarly used in rectangular tanks with travelling-bridge
collectors. Discharge of scum from the settling tank may be continuous or intermittent
depending on quantity produced. Skimmer mechanisms are of two types: dipping-weir and
sloping-beach. In the first, a slotted tilting pipe or other weir device is positioned during
skimming so that scum overflows from the tank together with considerable water. In
sloping-beach units scum is raked mechanically up a beach leaving most of the water
behind. The latter are simple to provide on circular tank mechanisms where they are almost
standard. For rectangular tanks .with chain-and-flight collectors, however, a separate
mechanism is required to move scum up the sloping beach. In this application, use of sloping
beach rather than weir type it is desirable to minimize the moisture content in the scum and
facilitate subsequent handling. Where scum is to be pumped away from the tanks the less
expensive weir-type skimmers are generally preferable.
7-12
-------
TABLE 7-1
PERFORMANCE OF SPECIAL SETTLING TANK INLETS
Loading
Effluent SS
Special Inlet Type and
Application
Peripheral Feed (Rex-
Nord) — Activated
Sludge Final Clarifier
Peripheral Feed
(Lakeside Equipment)
Primary Clarifier
Modular Cell Inlet
Feed well (Enviro-
tech)
Location
Ann Arbor,
Mich.
Sioux Falls,
S.D.
Ewing-Lawrence,
N.J.
Odgen, Utah
Test
Period
5/22/61
to
6/22/61
8/1 2/58 to
9/27/58
6/70 to
12/70
10/70 to
5/71
Special
Inlet
gpd/sq ft
1408
2015
1000±
520-715
850-950
950-1150
Conventional
Center Feed
gpd/sq ft
951
401
1000±
520-715
850-950
950-1150
Special
Inlet
mg/1
11
30
67
31
31
31
Conventional
Center Feed
mg/1
14
30
81
46
48
54
Refer-
ence
24
24
25
26
26
26
-------
7.5 Primary Sedimentation
In theory, sizing of primary tanks may be regarded as a question of economics. Successive
increments of tank area (providing lower loadings and longer detentions) typically yield
diminishing returns in performance. At some point it becomes more economical to accept
higher loads in subsequent units rather than provide more primary tank capacity. Some
designs have even omitted primary tanks entirely. Von der Emde (28) has indicated this may
be advantageous if one or more of the following conditions apply:
1. Sludge from the facility is to be pumped away for treatment elsewhere
2. Problems are expected with odors in primary tanks or poorly settling sludge in
secondary tanks
3. Aerobic digestion or extended aeration processes are to be used.
As a practical matter, performance-loading relationships adequate for use in cost
optimization studies can currently be obtained only by extensive testing of the actual
wastewater in existing full scale or pilot facilities, taking into account variations in flows
and characteristics. The generalized performance-loading curves for sedimentation units
available in the literature (2) (8) (19) (29) are unsatisfactory even as a basis for predicting
performance at particular design loadings much less for cost optimization studies. Such
curves are based on average daily plant flows from diverse sources, and ignore effects of
diurnal flow variations and of major in-plant flows such as waste secondary sludge, which
may be recycled to the primary tanks. The effect of su'ch unaccounted-for factors may be
seen in the wide scatter of removal-loading data plotted in the WPCF/ ASCE Sewage
Treatment Plant Design Manual (2).
In the absence of reliable performance-loading relations, primary tank designs may be
based on the typical parameters shown in Table 7-2.
TABLE 7-2
TYPICAL DESIGN PARAMETERS FOR PRIMARY CLARIFIERS
Hydraulic Loading
Type of Treatment Average Peak Depth
gpd/sqft ft
Primary Settling Followed
by Secondary Treatment 800-1,200 2,000-3,000 10-12
Primary Settling with Waste
Activated Sludge Return 600-800 1,200-1,500 12-15
Sizing should be calculated for both average and peak conditions (if flow equalization is not
used) and larger size used.
7-14
-------
These parameters are applicable to normal municipal wastewaters primarily of domestic
origin and should provide SS removals of 50 to 60 percent.
Weir loading limitations between 10,000 and 30,000 gpd/ ft (24-hr basis) have been sug-
gested for primary tanks (19) (29). At usua.l surface loadings, up to 1200 gpd/sq ft, round
tanks with single peripheral weir fall in this range for all but very large diameters (> 100 ft).
Thus normal practice is to provide only the single weir. In contrast, at surface loadings
as low as 600 gpd/sq ft rectangular tanks with single transverse weirs across the effluent
end exceed this range if the tank length is over 50 ft. Although rectangular tanks with weir
rates of more than 100,000 gpd/ft have shown SS removal in the normal range (30), rec-
tangular tanks are commonly equipped with multiple weir troughs to provide loadings of
30,000 gpd/ft or less. However, weir loadings are not as critical for primary tanks as they
are for secondary clarifiers.
Sludge solids can be estimated directly from the expected SS removal, making sure to
include waste activated sludge returned to the primary tank in the solids loading. The sludge
volume can be calculated based on expected concentration. If sludge is properly thickened
in the primary tank and pumping is carefully controlled to avoid pulling excess water, solids
concentrations of 2 to 7 percent may be obtained. On this basis typical primary sludge vol-
umes for domestic sewage would range from 0.2 to 0.5 percent of plant flow. The concen-
tration used in particular estimates should be based on actual plant experience or at least on
settling/thickening tests. Quantities of skimmings are quite variable. On a sustained basis
few plants average over 1 cu ft/mg (free water decanted) but scum handling facilities should
be capable of moving peak loads of perhaps six times this amount.
7.6 Secondary Sedimentation
7.6.1 Tank Sizing—General
The approach to sizing secondary clarifiers varies with the type of biological process they
serve.
7.6.1.1 Tank Sizing For Trickling Filter Effluent
Clarifiers following trickling filters are basically sized on hydraulic loading. Solids loading
limits are not involved in this sizing. Where further treatment follows sedimentation, cost
optimization may be considered in sizing the settling tanks, but the effort of developing ade-
quate performance—loading relations is seldom justified. Typical design parameters for
clarifiers following trickling filters are presented in Table 7-3. In applying the hydraulic
loading values from the table to design, sizing should be calculated for both peak and aver-
age conditions and the largest value determined should be used. At the indicated hydraulic
loadings, settled effluent quality is limited primarily by the performance of the biological
reactor not of the settling tanks.
7-15
-------
TABLE 7-3
TYPICAL DESIGN PARAMETERS FOR SECONDARY CLARIFIERS
Type of Treatment
Settling Following
Trickling Filtration
Settling Following
Air Activated
Sludge (Excluding
Extended Aeration)
Settling Following
Extended Aeration
Settling Following
Oxygen Activated
Sludge with
Primary Settling
Hydraulic Loading
Average Peak
' gpd/sq ft
400-600 1,000-1,200
400-800 1,000-1,200
200-400
800
400-800 , 1,000-1,200
Solids Loading*
Average Peak
Ib solids/day/sq ft
20-30
20-30
50
50
10-12
12-15
12-15
25-35
50
12-15
*Allowable solids loadings are generally governed by sludge thickening characteristics
associated with cold weather operations.
7.6.1.2 Tank Sizing For Activated Sludge Mixed Liquor
Activated sludge settling tanks have two distinct functions: solids separation and production
of a concentrated return flow to sustain biological treatment. Figure 7-6 illustrates how
important final tank underflow concentration is in maintaining the level of active solids
(and hence treatment) in the aerator mixed liquor.
As indicated in Section 7.3.6, the initial separation of activated sludge solids involves
hindered rather than discrete settling. For this type of settling, tanks must be sized so the
maximum surface hydraulic loading is less than the minimum initial settling velocity (ISV)
expected at maximum mixed liquor concentration and at minimum temperature. If the
hydraulic loading exceeds the ISV massive failure and overflow of solids will result.
To perform properly while producing a concentrated return flow, activated sludge settling
tanks must be designed to meet thickening as well as solids separation requirements. The
critical element in thickening is the rate at which solids are transported downward and
removed in the tank underflow. This is termed the solids transport or solids flux capacity,
generally expressed in the units of solids loading. Ib/sq ft/day. When the actual solids
loading applied to a tank exceeds its transport capacity, solids are being added faster than
7-16
-------
Q+R
Q=WASTEFLOW V C=MLSS
i
f 0-1.
'•*«t"AERATloN°**oo^
i%"*. *V JANK * o »*•! « V
R« RECYCLE
C= Cu
R) Ul Aft = D X r,i
Q-f R , V
r - SECONDARV
SETTLING
MAHK -
^ "5=
<
V
<
r* — ^
C=Cu
iQ
1 - WASTE
; = cu
MLSS =
Q + R
Cu
FIGURE 7-6
DEPENDENCE OF MLSS CONCENTRATION ON
SECONDARY SETTLING TANK
UNDERFLOW CONCENTRATION <31)
7-17
-------
they are being removed. If this condition persists the blanket of solids in the tank will build
up and eventually overflow with drastic effects on effluent quality. If significant solids are
lost from the system, biological treatment efficiency will be impaired. Tank depth may be
important in containing blanket buildup from diurnal peaks in solids loading.
Dick (31) has analyzed solids transport capacity assuming both solids and (vertical) under-
flow velocity uniformly distributed over the plan area of the tank. Although these condi-
tions are approximated only in moderate size circular tanks, this analysis provides major in-
sight into the thickening process and represents the only rational and straightforward ap-
proach currently available for estimating solids transport capacity. Under the conditions as-
sumed, solids transport capacity depends on only two factors: the thickening characteristics
of the solids (i.e., the relation between subsidence rate and concentration within the sludge
blanket) and the tank underflow rate. To get a concentrated underflow requires a low sludge
return rate which in turn means low solids transport capacity. High underflow rates have
been resorted to for handling poorly compacting sludges. This is only partly effective since
while increasing solids transport capacity, higher underflow rates also increase solids load-
ings due to the higher sludge recycle.
Methods for developing hydraulic and solids loading parameters from tests of settling and
thickening characteristics are discussed in Section 7.6.3. Typical design parameters for clar-
ifiers in activated sludge systems treating domestic waste are given in Table 7-3. In applying
hydraulic and solids loading values from this table, sizing should be calculated for both
peak and average conditions and the largest value determined should be used.
Settling tests provide worthwhile guidance in selecting design loadings. They should certain-
ly be included wherever pilot study of biological treatment is warranted by unusual waste
characteristics or treatment requirements. Testing is essential in any case where proposed
loadings go beyond the upper limits shown in Table 7-3.
Sizing activated sludge settling tanks according to proper hydraulic and solids loading para-
meters protects against massive failure, but does not by itself'guarantee high quality
effluent. After separation of the mass of activated sludge solids, significant quantities of small,
slowly settling particles may still be left in the clarified liquor. The amount and character of
such residual suspended solids logically relate to the loading and operating conditions in the
aeration tank, but few spe'cific studies have explored such relations. A study in Baltimore,
Md. (32) indicated that sludges with poorer thickening characteristics left lower residual
solids in the effluent. This was confirmed by studies covering a number of plants in Sweden
(33).
As noted in Chapter 6, the concentration of these residual solids can be reduced by floccula-
tion of the mixed liquor between aeration and settling, or by use of recirculation-type or
sludge-blanket-type solids contact reactors. Finally, although rates are lower, flocculatiori
in the clear water zone still appears to be a significant mechanism in removal of solids not
already entrapped in the sludge mass as it settles. This indicates that basin depth and deten-
tion are important in getting effluent SS down to low levels. Mixed liquor settling tests run
at several treatment. rUants,ia,Sweden (33) showed that residual turbidity above the sludge
Mail code 3404T
1200 Pennsylvania Avenue NW 7-18
Washington, DC 20460
202-566-0556
-------
interface dropped significantly over the first hour.
7.6.2 Development of Loading Parameters from Mixed Liquor Settling Tests
7.6.2.1 Surface Hydraulic Loadings
Initial settling velocity (ISV) at actual mixed liquor concentration may be determined in a
single test simply by plotting the height of the sludge-liquid interface vs. time and noting the
slope of the straight line portion of the plot. The critical minimum ISV value for a particu-
lar system may be estimated from results of a number of individual tests. The designer
should attempt to establish relations between ISV and biological process parameters such
as mixed liquor concentration and organic loading. The selected ISV value should then re-
flect conditions most unfavorable to settling including correction for minimum expected
temperature. Finally a capacity factor as discussed in Section 7.3.1 should be applied to
convert the critical ISV to a hydraulic loading.
The resulting maximum surface hydraulic loading should not be exceeded by any sustained
maximum flow (say 4-hr duration). Initial settling velocities for mixed liquor from air acti-
vated sludge systems have been reported to range from 3 ft/hr to over 20 ft/hr (4) (21) (34).
For good settling (non-bulking) air activated sludges from municipal wastewaters the fol-
lowing design relation between the ISV and the mixed concentration has been suggested
(34):
Vi = 22.5e -338Cl
Where:
Vi = settling velocity in ft/hr
Ci = concentration in Ib/lb
Bulking sludges will show ISV values well below those indicated by this line. Sludges with
superior settling qualities may show considerably higher values.
7.6.2.2 Sludge Volume Index
The sludge volume index (SVI) widely used to guide operating control of the activated
sludge process, provides an approximate indication of sludge compaction characteristics.
The index is calculated by dividing the initial mixed liquor SS concentration (percent) into
the settled volume (percent of initial volume) occupied by the solids after one half hour
of settling.
The reciprocal of the SVI is often taken as an approximate indication of the maximum re-
turn sludge concentration which can be obtained with a given mixed liquor (100/SVI = per-
cent solids). The index has been used as a guide to sizing return sludge pumping require-
ments to maintain different mixed liquor concentrations (2). Although the SVI does not
give a direct indication of solids transport capacity, it has been suggested that for index
values of less than 100,.underflow concentrations below 1 percent and mixed liquor concen-
trations below 3000 mg/1, hydraulic rather than loadings will govern clarifier sizing (2).
7-19
-------
7.6.2.3 Solids Loading
Based on the analysis discussed in Section 7.6.1, Dick (31) has proposed a method for deter-
mining limiting solids transport capacity as a function of underflow rate, given a curve or
equation defining the relation of settling velocity to concentration. Dick and Young (35)
have formulated the method into a series of equations, assuming that the settling veloc-.
ity-concentration curve could be represented in the form:
V = a.c-n
where
V is settling velocity
c is concentration
and
a and n are appropriate constants for the units used.
The most serious problem in applying the method is determining the settling veloc-
ity-concentration relation. Dick suggests developing the relation, from a series of ISV tests
on the same mixed liquor at different initial concentrations (obtained by settling, decanting
clear liquid and resuspending the solids). There is a serious question whether a curve devel-
oped from such tests really represents the behavior of the solids in the sludge blanket of a
clarifier. Nevertheless this approach is the best presently available for estimating solids
transport capacity fom settling tests. Others suggested (6) are open to even more serious
objections.
In translating solids transport capacity to an allowable solids loading some safety factor
may be needed to allow for possible critical conditions (temperature, poor thickening char-
acteristics, etc.) not reflected in the test work.
In a design application trial solutions at different return sludge rates may be justified to de-
termine the effect on tank sizing of the different solids loadings and capacities that result at
the various underflow rates. Sizing should be based on peak solids loadings associated with
sustained maximum flows unless specific testing has justified a reduction taking advantage
of storage of peak solids by increases in sludge blanket height. Such storage should be
avoided with nitrifying sludges (34).
7.6.2.4 Settling Test Procedures
Although it has been demonstrated (36) (37) that factors such as column diameter, sludge
depth, dissolved oxygen and application of stirring can significantly affect the results of
settling tests, standard values for such factors or standard allowances for their variations
have not been adopted. Dick (36) has detailed test procedures and indicated (38) preference
for use of-sludge depths of 3 ft. column diameters of 3.5 in. or more and slow stirring at tip
speeds of 10 in./ min.
7-20
-------
7.6.3 Flow Stabilization and Density Currents
Two approaches to preventing short circuiting from density currents, were described in Sec-
tion 7.3.2: dynamic stabilization and density stabilization. An exhaustive study of shallow
activated sludge settling tanks has been made in Sweden (33). Included were tracer studies
on tanks under actual operating conditions and parallel quiescent and stirred settling col-
umn tests on the mixed liquor. These tanks, although designed for dynamic stabilization,
showed serious short circuiting. Due to flocculation in the tanks, however, effluent quality
was better than predicted from the quiescent settling results and actual detention times.
In the U.S. where final settling tanks commonly have design depths 10 ft or more, flow
stabilization depends totally on density. Unfortunately studies of the type conducted in
Sweden have apparently not been run on tanks designed for density stabilization. In the
side-by-side performance tests (Section 7.4) comparing special inlet designs for circular
tanks with conventional center feedwells, density stabilization could have been important,
but no data were taken to show the degree of short circuiting. Neither were any parallel
settling column tests run. Tracer studies of these special inlets generally have been run on
clear water, so they fail to show any effects of density stabilization. Even without these ef-
fects, special inlets displayed less short circuiting (13) (14).
In an attempt to minimize undesirable density current effects, several designs have varied
the placement of sludge drawoffs and effluent weirs in relation to the inlet. Sawyer (39)
pointed out the "submerged waterfall effect" that occurs when the density current reaches
the tank floor in its initial downward sweep. In conventional rectangular or circular tanks
where the sludge drawoff is located below the inlet the impact of the "waterfall" can dilute
the collected sludge and resuspend a portion of it. Peripheral-feed circular tanks avoid this
problem as do those equipped with suction-type mechanisms which remove sludge from the
entire tank floor. Even in tanks with centerfeed and center sludge drawoff, use of deep feed-
wells discharging at low velocities can minimize the problem (10). Rectangular tanks have
been constructed with sludge drawoffs located away from the inlet. Excellent results have
been obtained at New York City with sludge drawoffs at mid-length of the tanks (30). This
arrangement uses the density current to speed sludge removal but prevents the density cur-
rent from entering the outlet zone of the tank. Rectangular tanks may also be equipped with
suction-type sludge removal on traveling-bridge mechanisms.
In one special rectangular tank arrangement, effluent weirs are distributed over the length
of the tank, with baffles in the upper part of the tank to impede counter currents induced by
density current below and force vertical flow to the weirs. In peripheral feed circular tanks,
tests have shown that units with peripheral drawoffs located just inside the inlet channel
produce better effluent than units with weirs located more toward the tank center (40).
Weir hydraulic loadings of 15,000 gpd/ft at average design flows are suggested in the Ten
State Standards (29), with allowances of up to 20,000 gpd/ft where weirs are located so that
density currents do not upturn below them. Loadings of up to 100,000 gpd/ft have been
used without apparent problems in designs such as those of New York City where weirs are
well separated from density current effects (30).
7-21
-------
7,6.4 Sludge and Skimmings Removal
Suction-type sludge removal should be considered wherever sludge detention in the tanks is
critical and doubt exists about conveyance time for other mechanisms. Desirable features
in suction-type mechanisms include independent flow controls for each suction drawoff and
visible gravity sludge discharges.
Federal guidelines (41) require skimming equipment on secondary settling tanks to remove
floating sludge and any oily materials not separated in previous treatment. Scum quantities
generally are small in relation to those from primary tanks (0.1 cu ft/mg). Where no
primary tanks are included in plant process, scum quantities from secondary tanks could be
conservatively estimated on the same basis as for primary tanks. Effluent weirs should be
laid out to permit skimming the maximum possible portion of the tank surface.
Maximum practical concentrations of underflow from secondary clairifiers in activated
sludge systems range from 0.5 Jo 2.0 percent solids, depending on settling and compaction
characteristics of the sludge. Actual concentrations depend on the return sludge pumping
rate. Sludge concentrations of 3 to 7 percent solids may be obtained from secondary clari-
fiers iff trickling filter systems.
7.7 Chemical Sedimentation
Sedimentation of chemically coagulated or precipitated wastewaters is similar to sedimen-
tation of wastewaters without chemicals. The design of tanks can proceed on essentially the
same basis, provided special consideration is given to the effects of chemical treatment on
settling characteristics, sludge quantities, resistance of the sludge to movement by collecting
and pumping equipment, and the special maintenance problems encountered with lime
coagulation. Few data have been reported concerning settling characteristics of chemically
precipitated floe in wastewater treatment. Some data are available on chemical precipitation
in water treatment using similar chemicals (42).
From the literature it is apparent that actual surface loadings vary considerably from one
application to another (2) (43) (44) (45) (46) (47) (48) (49). This wide variation emphasizes
the importance of testing and pilot work in designing chemical precipitation facilities. In the
absence of testing indicating higher figures to be satisfactory, the following typical surface
loading rates may be used for sizing tanks (47) (50):
Chemical Peak Surface Loading
gpd/sq ft
Alum 500-600
Iron 700-800
Lime 1400-1600
In general, these design rates may be used for primary, secondary or tertiary applications.
It should be noted, however, that they are based on limited data and may be revised when
more experience is available.
7-22
-------
Sludge quantities from chemical precipitation can be estimated from the SS removal and
the stoichiometry of chemical reactions involved. Volumes depend on sludge concentrations
which are highly variable (1 to 15 percent) and are best determined by actual test. Equip-
ment suppliers should be consulted about strength and power of collector equipment to
handle the dense sludges expected from lime precipitation. Extra smooth piping glass-lined
or PVC, should be used for lime sludges. Average sludge productions determined from raw
wastewater coagulation by lime, iron and alum are 6,500, 1740 and 1120 Ib/mg respectively
(47). Average sludge volumes for the same locations for lime and iron are 10,000 and 13,000
gal/mg, respectively (47). Brown (51) observed a sludge production of 1894 Ib solids/mg
(6275 gal/mg) using alum for precipitating trickling filter effluent."
7.8 Flotation
7.8.1 Applications
This section deals with flotation induced by introduction of fine gas bubbles into waste-
water. Since most SS in municipal wastewater have specific gravity values only slightly
above 1, adhesion of the gas bubbles to the solids particles readily makes them buoyant.
For flotation of solids in municipal wastewater, gas bubbles must be quite fine (.01 to 0.1
mm); otherwise, their own rise rate prevents significant adhesion to the solid particles.
Three methods of introducing gas bubbles have been shown to create bubbles sufficiently
fine for flotation of municipal wastewater SS. Vacuum flotation and dissolved-air flotation
(DAF) both create conditions in which the wastewater is supersaturated with air at some
pressure. Upon reduction of that pressure, air comes out of solution as finely-divided bub-
bles. Auto-flotation can occur in algae suspensions if they become sufficiently super-
saturated with dissolved oxygen from photosynthesis. Vacuum flotation and autoflota-
tion are not often used because the former is expensive and the latter can only operate under
limited conditions of warm weather and bright sunshine (52). Diffused or submerged tur-
bine aerators create bubbles much too coarse for flotation of municipal wastewater solids.
Pressure and vacuum flotation units have found only limited application in treatment of
municipal wastewater. It has been difficult to justify using these units in conventional appli-
cations such as primary SS removal or mixed liquor clarification because sedimentation is
ordinarily cheaper, simpler and often provides better results.
Advantages which might favor use of flotation in special applications include: 1) higher sur-
face loadings, hence smaller tanks sizes (important where space is critical); 2) ability to
handle peak seasonal loads or storm flows (in some designs flotation may be used inter-
mittently to increase capacity of settling tanks); 3) effectiveness in removing solids which
. are difficult to settle.
Dissolved-air flotation has been suggested for separation of grit and scum in a single treat-
ment unit (2) (53) (54). Performance data for such applications are lacking, however. Be-
cause it can produce a float of much thicker consistency than settled activated sludge, dis-
solved-air flotation has been tried for mixed liquor solids separation. Full scale studies at
7-23
-------
Manassas, Va. (55) indicated that this application was not economically competitive with
conventional systems.
7.8.2 Dissolved-Air Flotation
As shown in Figure 7-7, dissolved air flotation (DAF) units commonly employ rectangu-
lar tanks with separate chain-and-flight scum and sludge collectors. Circular units are also
commercially available. The widest application for these units has been as thickners for
waste activated sludge. Units used for SS separation are similar, but design parameter val-
ues vary according to the application. To avoid fouling of pressurizing and pres-
sure-regulating equipment and excessive shearing of influent solids, a stream of recycled ef-
fluent is usually pressurized. Upon pressure release, this stream is blended with the inflow
to be treated. Other methods include pressurizing all or part of the influent stream.
Design of DAF units involves selection of values for a number of parameters including per-
cent recycle flow, operating pressure, pressurization retention time, air flow, and surface
hydraulic loading, solids loading (area basis) and float detention period. Variables reflec-
ting influent characteristics include flow, solids loading, liquid temperature and type and
quality of influent solids. Investigators have attempted to relate flotation performance to
the air to solids ratio and a number of other variables with a limited amount of success.
Mulbarger and Huffman (56) noted that float concentrations depend more on float deten-
tion time than solids loading. They related capture in flotation to a parameter equal to the
air to solids ratio divided by the product of surface hydraulic loading and dynamic viscosity.
Values of specific parameters used in actual applications vary widely. Typical ranges cited
are as follows (2) (50) (56) (57) (58) (59) (60) (61):
Parameter: Range:
Pressure, psig 25 to 70
Air to Solids Ratio, Ib/lb 0.01 to 0.1
Float Detention, min 20 to 60
Max. 24-hr
Surface Hydraulic Loading, gpd/sq ft 500 to 4000
Recycle, percent 5 to 120
Available data from specific applications are summarized in Table 7-4.
In flotation equipment special attention must be given to the inlet, outlet and collector
mechanism configurations. The flotation tank must permit aggregate rise with a minimum
of interference in the form of turbulence or obstructions and provide for removal of floated
froth, settled sludge and treated effluent. Effluent ports must be sufficiently submerged to
prevent interference with the froth on the surface. The inlet conditions of the flotation tank
are critical to proper performance. Baffles, walls, and other obstructive energy-dissipating
devices tend to destroy aggregate bonding with resulting loss in flotation efficiency. Also,
7-24
-------
SLUDGE REMOVAL MECHANISM
EFFLUENT
^WRECIRCULATION PUMP
FEED
RETENTION TANK
AIR DISSOLUTION
INFLUENT
RECYCLED
FLOW —5
SLUDGE
DISCHARGE
REAERATION PUMP
FIGURE 7-7
SCHEMATIC OF A DISSOLVED-AIR
FLOTATION UNIT
(Courtesy of Komline-Sanderson)
7-25
-------
TABLE 7-4
10'-
o\'
DISSOLVED-A1R FLOTATION APPLICATIONS
Flotation
Plant
Aker, Sweden
Klagerup, Sweden
Salemstaden,
Sweden
Bara, Sweden
Kungsor, Sweden
Design ,
Flow
mgd
0.4
0.12
2.16
0.08
1.93
Type of
Municipal
Wastewater
Primary Effluent
Primary Effluent
Primary Effluent
Aerator Mixed
Liquor
Unsettled Trick-
Chemical
Treatment
Coagulant
Alum(a)
Alum(a>
AIum(a)
Alum
AJum(a)
Dose
mg/1
100
159
175
-
145
Surface
Hydraulic
Loading
gpd/sq ft
2360
1180
2540
2480
4480
Detention
Time
hrs.
0.31
0.37
0.26
0.24
0.20
SS Performance Data
Inf. Eff. Removal
mg/1 mg/1 percent
71(0
77(0
60(O
_ _ _
97(d)
Ref. Remarks
(62)
(62)
(62)
(62)
(62)
Flen, Sweden 2.54
Prince William 1.0
County, Va.
Bellair, Texas (b)
Stockton, Calif. (b)
ling Filter
Effluent
Unsettled Trick-
ling Filter
Effluent
Aerator Mixed
Liquor
Aerator Mixed
Liquor
Lagoon Effluent
Alum(a)
None
Cationic 8
Polymer 30
Alum 75-225
3360
360
0.35
3.4
0.32
30
to
100
2000 70
17
94 12
to to
152 20
87
(62)
(56) Limiting solids
loading 15 Ib/lb
SF
(63)
(52) Includes filtration
(a) 30-60 min. of flocculation provided before flotation, (b) Pilot Plants.
(c) BOD removal; no SS data given, (d) P removal; no SS data given.
-------
turbulence in the. region of the froth will result in losses of floated solids. Ettelt (58) report-
ed several different designs of inlet structures in his prototype units. His tangential flow in-
let appears to offer considerable promise where such designs are compatible with the entire
structure.
Further discussion of design features of dissolved-air flotation units may be found in the lit-
eVature (56) (58) (59) (60).
7.9 Shallow Settling Devices
The potential advantages of multiple tray shallow settling devices have long been recognized
(3) (4), but early prototypes of such equipment were unsuccessful due to practical problems
of flow distribution and sludge removal. In recent years, shallow settling devices of im-
proved design, such as tube settlers, have been applied to water and wastewater treatment.
Tube settlers consist of bundles of small plastic tubes with hydraulic radii ranging from one
inch upward and lengths of 2 ft or more, depending upon the particular application. Square
tube sections are most common but hexagonal and other shapes have been used by various
manufacturers.
Tubes are commonly inclined steeply (60°) to horizontal and fabricated in modules, as
shown in Figure 7-8. These modules have beam strength which permits their installation in
settling tanks, as shown in Figures 7-9 and 7-iO. Clarifier influent is introduced beneath the
tube modules. The flow passes upward through the modules with the solids moving counter-
currently by gravity (Figure 7-11) and falling from the tube bottoms into the sludge collec-
tion zone beneath. The clarified effluent is collected above the tube modules.
Free standing package units with tubes only slightly inclined (5°) have found some appli-
cation in small chemical clarification/ filtration systems for tertiary wastewater treatment.
Tube settlers promote sedimentation in three ways: 1) the multiple tubes stacked one above
another provide an effective settling area several times that of the projection in plan of the
modules; 2) the small hydraulic radius of the tubes maintains laminar flow and promotes
uniform flow distribution; 3) in steeply inclined tubes, the movement of sludge against the
direction of flow favors particle contact and agglomeration. This additional flocculation off-
sets the reduction in their horizontal projected area caused by inclining the tubes. For alum
floe suspensions a given length.tube has been shown to provide most effective removal at an
inclination of about 45 degrees, and performance even at 60 degrees was comparable to that
when horizontal (64).
Tube settlers have been promoted both for reducing required size of settling tanks and for
improving their performance, but manufacturers presently tend to emphasize improved per-
formance and recommend the same surface hydraulic loadings for tanks equipped with tube
settlers as for conventional tanks.
Comparative data on performance of tanks with and without tube settlers (either
side-by-side or before-and-after) are shown in Table 7-5. The data are quite limited and
7-27
-------
^1
NJ
OO
FIGURE 7-8
MODULE OF STEEPLY INCLINED TUBES
(Courtesy Neptune Microfloc, Inc.)
-------
FIGURE 7-9
TUBE SETTLERS IN EXISTING CLARIFIER
SUPPORT MODULE
TUBE SETTLER
MODULES
FIGURE 7-10 .
PLAN VIEW OF MODIFIED CLARIFIER
7-29
-------
DIRECTION OF
FLOW
TO SLUDGE
COLLECTION
FIGURE 7-11
TUBE SETTLERS - FLOW PATTERN
7-30
-------
TABLE 7-5
TUBE SETTLER INSTALLATIONS
Plant Location
Hopewell Township,
Pennsylvania
Trenton.
Michigan
Lebanon.
Ohio
Operational Data
Using Tube Settlers
Type
Activated
Sludge
Activated
Sludge
Activated
Sludge
Plant
Design
mgd
0.13
6.5
0.75
Flow
Actual
mgd
0.13
5.6
1.25
Existing Facility
Location
Secondary
Clarifier
Secondary
Clarifier
Secondary
Clarifier
Overflow Rate
gpm/sq ft
0.34
-
-
0.61
Effluent SS
mg/1
60-70
-
-
61
Tube Over-
flow Rate
gpm/sq ft
-
2
0.56
0.85
Tank Over-
flow Rate
gpm/sq ft
—
0.68
0.29
0.61
Effluent SS
mg/1
—
27
8
30
Reference
65
66
67
-------
rather inconclusive as to the benefits obtained from the tube settlers.
Tube settlers have found wider application in water treatment than in wastewater. For
wastewater, tube settlers may find their best applications in tertiary coagulation and settl-
ing. They also may be of help in upgrading performance of units with serious short circuit-
ing problems.
When installed, settling tubes usually cover one half to two thirds of the basin area. To pre-
vent fouling of tubes, the remaining area between the inlet and tube area is arranged to pro-
vide scum removal. The portion of the basin equipped with settlers should have collecting
weirs at 15 ft or closer spacing to induce an even vertical flow distribution and reduce short
circuiting. '
Tube settler installations require a support grid (usually designed to support one man) and
surface baffles to separate the tube settler and scum collection area. Minimum basin depths
should be 10 to 12 ft.
Long term studies have revealed that in wastewater treatment the upper surface of the settl-
ing tubes becomes coated with sludge (68). Long term fouling of the tubes with grease or
rags has not been a problem, but in order to maintain a high degree of solids removal, it is
generally necessary to install an air grid and periodically interrupt flow to introduce air to
remove the sludge build-up on the tubes.
A typical air wash cycle consists of draining the tank to the level of the tubes and then al-
lowing air to rise up through the tubes. The air is supplied either from a fixed grid or a
scour system attached to the rake arm. After the air wash, approximately 15 to 25 min is re-
quired for the effluent SS to return to normal, e.g., from 60 mg/ 1 to 10 mg/1 before the
unit can be returned to service (68). A short quiescent period of no flow may also be needed
between the drain down and the air wash(69). Generally, required cleaning frequency varies
from one week to several months (70). Where serious sludge carryover conditions are ex-
perienced, however, it has proven difficult to prevent fouling with even the highest cleaning
frequency (71).
7.10 Wedge-Wire Settler
Wedge-wire settlers are wire matrices installed in clarification basins similarly to tube set-
tlers. They are designed to improve the quality of settled effluent from activated sludge or
trickling filter treatment.
The equipment can be installed in any conventional clarifier configuration. The settling de-
vice consists of a matrix constructed from parallel wires (See Figure 7-12) suspended be-
neath and parallel with the surface of the water so that the wastewater must pass upward
through the mesh before reaching the effluent weir. Over 200 secondary clarifier in-
stallations in England are equipped with wedge-wire settlers (72).
Typically, the wire in the matrix is triangular in cross-section and arranged with apex point-
7-32
-------
TANK WALL
INFLUENT
FINAL EFFLUENT OUT
WEDOE WIRE PANELS
FIGURE 7-12 ;
SIMPLE WEDGE WIRE CLARIFIER
4x3 TIMBER FRAME
SECURED BY RA6BOLTS TO
WALLS.
WEDGE WIRE PANELS.
2x2 ANGLE
IRON.
METAL BAFFLE PLATE.
4 1/2" WALL BUILT
UP FROM TANK
FLOOR.
OUTLET WEIR
FIGURE 7-13
INSTALLATION OF WEDGE WIRE PANELS
7-33
-------
ed downward, to provide 0.125 to 0.250 mm openings at the top surface. The openings com-
prise about 15 percent of the total area. Construction is of either stainless steel or aluminum
and the wires must be rigidly fixed. The wire "rack" is supported within the tank about
6-inches below the water surface on steel angles or other similar structural grid-work (See
Figure 7-13).
Used in conjunction with effluent launders at spacings of 15 ft or less, wedge-wire settlers
distribute flow quite uniformly over the entire area of the basin, producing a nearly ideal
upflow clarification zone above the wire (65). Finer particles which settle in this zone even-
tually coalesce into a sludge blanket which aids in removal of near-colloidal particles. When
the blanket builds in thickness and nears the water surface (2 to 4 days) cleaning is neces-
sary. The basin level is lowered below the wire level and the wire is hosed down to clean off
the accumulated solids. Total clean up time is about one hour (72).
Wedge-wire settler applications have been limited to relatively low surface hydraulic load-
ings, i.e., 600 gpd/sq ft with peaks to 800 gpd/sq ft. Effluent quality has been roughly re-
lated to flow rate (73). Under stable conditions an effluent SS of 15 to 20 mg/1 could be ex-
pected at 600 gpd/sq ft and 10 mg/1 at 300 gpd/sq ft. Results of side by side tests of secon-
dary clarifiers treating trickling filter effluent (150 mg/ 1 SS) indicated that standard clari-
fiers produced effluents from 7 to 77 mg/ 1 with an average of 41 mg/ 1 while identical
wedge-wire settlers produced effluents of 1.6 to 18 mg/ 1 with an average of 8 mg/1 (74).
All units operated at a relatively low rate of 300 gpd/sq ft. This low hydraulic loading ap-
peared significant to the wedge-wire settler performance, but it had little effect on the ef-
fluent quality of the conventional settling unit.
In activated sludge clarification use of the wedge-wire settlers reduced effluent SS to 8 to 16
mg/ 1 compared with 30 to 40 mg/ 1 from the same conventional clarifier at a rate of 600
gpd/sq ft (74).
Pullen (75) cited 5 clarifiers of small size (18,000 to 170,000 gpd), which were equipped with
wedge-wire screens, experiencing a 50 percent improvement in effluent SS quality.
Wedge-wire settlers are limited to multiple tank installations so that during shut down for
washing flow can be diverted to other clarifiers. Drain and wash down flow can be recycled
to pretreatment units or to sludge handling systems.
7-34
-------
7.11 References
1. Kalbskopf, K. H., European Practices in Sedimentation, in Water Quality Improvement
by Physical and Chemical Processes, Univ. of Texas Press, Austin, Texas (1970).
2. Sewage Treatment Plant Design, WPCF Manual of Practice No. 8, ASCE Manual
No. 36(1959).
3. Hazen, A., On Sedimentation, Trans. Amer. Soc. Civ. Eng., 53, 45 (1904).
4. Camp, T. R., Sedimentation and Design oj Settling Tanks, Trans., Am. Soc. Civil
Engrs., Ill, 895(1946).
5. O'Connor, D. J. and Eckenfelder W. W., Evaluation of Laboratory Settling Data for
Process Design in Biological Treatment of Sewage and Industrial Wastes, Vol. 2, Rh'ein-
hold Publishing Co. (1958).
6. Eckenfelder, W. W. and Ford, D. L., Water Pollution Control, Jenkins Book Publishing
Company, Austin, Texas (1970).
7. Weber, W. J., Physicochemical Processes for Water Pollution Control, Wiley— Inter-
science (1972).
8. Metcalf, L., and Eddy, H. P., Sewerage and Sewage Disposal, McGraw-Hill Book
Company, New York (1930).
9. Smith, R., Cost of Conventional and Advanced Treatment of Wastewater, Jour.
WPCF, 40, 1546 (Sept. 1968).
10. Fitch, E. B. and Lutz, W. A., Feedwells for Density Stabilization, JWPCF 32, 147
(1960).
11. Aitken, I.M.E., Reflections on Sedimentation Theory and Practice—Part I, Eff. and
Water Treatment Jour. (Br.), 7, No. 4, pg. 226 (Apr. 1967).
12. Geinopolis, A. and Katz, W. J., United States Practice in Sedimentation of Sewage
and Waste Solids, in Water Quality Improvement by Physical and Chemical Pro-
cesses, Univ. of Texas Press, Austin, Texas (1970).
13. Katz, W. J. and Geinopolis, A., A Comparative Study of Hydraulic Characteristics of
Two Types of Circular Solids Separation Basins in Biological Treatment of Sewage
and Industrial Wastes, Vol. 2 Rheinhold Publishing Company (1958).
14. Dague, R. R. and Baumann, E. R., Hydraulics of Settling Tanks Determined by Mod-
els, Presented at 1961 Annual Meeting of Iowa Water Pollution Control Association
(Reprinted by Lakeside Equipment Corp.)
7-35
-------
15. Villemonte, F.R., Rohlich, G.A., et al, Hydraulic and Removal Efjiciencies in Sedimen-
tation^asins -Third International Conference on Water Pollution Research, Munich,
Section II, Paper 16(1966).
16. Fitch, E. B. Sedimentation Process Fundamentals, in Biological Treatment of Sewage
: and Industrial Wastes, Vol. 2, Rheinhold Publishing, Co. (1958).
17. Fitch, E. B., Significance oj Detention in Sedimentation, Sewage and Industrial
Wastes, 29, 1123 (Oct. 1957).
18. Knop, E., Design Studies for the Emscher Mouth Treatment Plant, JWPCF 38, 1194
(July 1966).
19. Fair, G. M., Geyer, J. C. and Okun, D. A.; Water and Waste-water Engineering, 2,
John Wiley & Sons, New York: (1968).
20..' Ingersoll, A. C., McK.ee, J. E. and Brooks, N. H., Fundamental Concepts of Rectangu-
lar Settling Tanks Trans ASCE 121, 1179 (1956).
21. Eckenfelder W. W. and Melbinger, N., Settling and Compaction Characteristics of
Biological Sludges, Sewage and Industrial Wastes 29, 1114 (Oct. 1957).
22. Naval Facilities Engineering Command, U.S. Navy, Pollution Control Systems, Ch. 10
iri Civil Engineering Design Manual, DM-5.
23. Cleasby, J. L., Baumann, E. R., and Schmid, L., Comparison of Peripheral Feed and
Center Feed Settling Tanks Using Models, Report by Iowa Engineering Experiment
Station, Ames, Iowa, (February, 1962).
24. Reed, R. V., Rexnord, Personal Communication^ (April 1973).
25. Hikes, Burd, Lakeside Equipment Corp., Personal Communication (May, 1973).
26.- Envirotech, Municipal Equipment Division, Eimco Modular Energy Dissipating Clari-
fier Feedwell, Technical Brochure: MED 121, 10/72.
27. Mercer, R. H., Rectangular vs. Circular Settling Tanks, The American City, 98, (Oct.
1973).
28. Von der Emde, W., To What Extent are Primary Tanks Required?, Water Research, 6,
395(1972).
29. Recommended Standards for Sewage Works, Great Lakes—Upper Mississippi River
Board of State Sanitary Engineers (1968).'
7-36
-------
30. Gould, R., Wards Island Plant Capacity Increased by Structural Changes, Sewage and
Industrial Wastes, 22, 997 (1950).
31. Dick, Richard I., Role of Activated Sludge Final Settling Tanks, J. SED, ASCE, 96,
423 (Apr. 1970).
32. Keefer, C. E., Relationship of the Sludge Density Index to the Activated Sludge Pro-
cess, Jour. WPCF, 35, 1166 (1963).
33. Fischerstrom, C.N.H., Isgard, E. and Larsen, I., Settling of Activated Sludge in Hori-
zontal Tanks, J. SED, ASCE, 93, SA3, 73 (June 1967).
34. Suggested Peaking Considerations for Activated Sludge, Sanitary Engineering Staff
Report, Iowa State University (197)).
35. Dick, R. I. and Young, K. W., Analysis of Thickening Performance of Final Settling
Tanks, Proceedings of 27th Purdue Industrial Waste Conference, (1972).
36. Dick, R. I. and Ewin, B. B., Evaluation of Activated Sludge Thickening Theories, J.
SED, ASCE, 93, SA4, 9 (Aug. 1967).
37. Veselind, P. A.. Discussion of Evaluation of Activated Sludge Thickening Theories, J.
SED ASCE 94, SA1, 185 (Feb. 1968).
38. Dick, R. I., Thickening in Water Quality Imrovement by Physical and Chemical Pro-
cesses, Univ. of Texas Press, Austin, Texas (1970).
39. Sawyer, C. N., Final Clarifiers and Clarifier Mechanisms, Biological Treatment of
Sewage and Industrial Wastes, Vol. l,Reinhold Publishing Co, New York (1956).
40. Fall, E. B., Jr., Redesigning Existing Facilities to Increase Hydraulic and Organic
Loading, Vol. 43, pg. 1695, Jour. WPCF (1971).
41. Federal Water Quality Administration, Federal Guidelines for Design, Operation and
Maintenance of Waste-water Treatment Facilities (September, 1970).
42. Geinopolos, A., Albrecht, A.E., and Katz, W. J., The Character of Suspended Solids
and Basin Hydraulics are Key Factors in the Clarification of Water and Waste-water,
Industrial Water Engrg., 3, 10, 19 (Oct. 1966).
43. Green, O., Eyer, F., and Pierce, D., Studies on Removal of Phosphates and Related
Removal of Suspended Matter and BOD at Grayling, Michigan, Distributed by Dow
Chemical Co.
44. Hennessey, J., Keilinski, R., Beeghly, J. H., and Pawlak, T. J., Phosphorous Removal
at Ponliac, Michigan, Presented at U.S. EPA, WQO Design Seminar, Cleveland, Ohio
(Apr. 1971).
7-37
-------
45. Oliva, J. A., Dept. of Public Health, County of Nassau, Personal Communication
(March, 1973).
46. Gulp, R. and Gulp, G.,.Advanced Wastewater Treatment, Van Nostrand Reinhold Co.
(1971).
47. Kreissl, J. F. and Westrick, J. J., Municipal Waste Treatment by Physical-Chemical
Methods, U.S. EPA, National Environmental Research Center, Cincinnati, Ohio.
48, Water Treatment Plant Design, American Water Works Association, Inc., New York
(1969).
49. U.S. EPA, Advanced Wastewater Treatment as Practiced at South Tahoe, Proj. No.
171010 ELQ08/71 (August 1971).
50. Metcalf and Eddy, Inc., Wastewater Engineering, Collection, Treatment, Disposal,
McGraw-Hill, New York (1972).
51. Brown, James C., Alum Treatment of High-Rate Trickling Filter Effluent, Chapel
Hill, 'North^Carolina, Supplemental Information for U.S. EPA Technology Transfer
Design Seminar on Upgrading Trickling Filters, Salt Lake City, Utah, Nov. 13-15
(1973).
52. Parker,Denny S.,'et al, Algae Removal Improves Pond Effluent, Water and Wastes
Engineering, 10, 1, pp. 26-29 (January, 1973).
53. Wahl,A.J., Larson, C.C., et al, 1963 Operator's Forum, Jour. WPCF, 36, 401 (April,
1964).
54. Katz, W.J.,Solids Separation Using Dissolved-Air Flotation, in Air Utilization in the
Treatment of Industrial Wastes, University of Wisconsin (1958).
55. Mulbarger, M. C., et al, Manassas Va., Adds Nutrient Removal to Waste Treatment,
Water and Wastes Engineering, 6, 4, pp. 46-48 (April 1969).
56. Mulbarger, Michael C., and Huffman, Donald D., Mixed Liquor Solids Separation By
Flotation, Jour. SED, ASCE, 96, SA4, pp. 861-871 (Aug. 1970).
57. Levy, R. L., White, R. L., Shea,T. G., Treatment of Combined and Raw Sewages with
the Dissolved Air Flotation Process, Water Research, Pergamon Press, Great Brittain,
Vol. 6, pp. 1487-1500(1972).
58. Ettelt, G. A., Activated Sludge Thickening by Dissolved Air Flotation, Proc. 19th Pur-
due Ind. Wastes Conf. (1964).
7-38
-------
59. Vrablik, E. R., Fundamental Principles of Dissolved Air Flotation of Industrial
Wastes, Proc. 14th Purdue Ind. Wastes Conf. (1959).
60. Masterson, E. M.,and Pratt, J.W., Application of Pressure Flotation Principles to Pro-
cess Equipment Design, in Biological Treatment of Sewage and Industrial Wastes, Vol
II, Reinhold Pub., Co., New York (1958).
61. Ort. J. E., Lubbock WRAPS It Up. Water and Wastes Engineering, Vol. 9, No. 9, pp.
63-66 (Sept. 1972).
62. Experience of Chemical Purfication, National Swedish Environment Protection
Board (1969).
63. Anderson, L., The Permutit Company, Personal Communication (January 1973).
64. Gulp, G., Hansen, S., Richardson, G., High-Rate Sedimentation in Water Treatment
Works, Jour. AWWA, 60, 681 (June, 1968).
65. Hansen, S. P., Gulp, G. L. and Stukenberg, J. R., Practical Application of Idealized
Sedimentation Theory in Wastewater Treatment, Journal Water Pollution Control
Federation, 41, No. 8, pp. 1421-1444 (1969).
66. Neptune Mocrofloc Incorporated, City of Trenton Sewage Treatment Plant, Case His-
tory No. 27(1971).
67. Oppelt, E. T., Evaluation of High Rate Settling of Activated Sludge, Interim U.S..
EPA Internal Report, Advanced Waste Treatment Laboratory, Cincinnati, Ohio
(1973).
68. Slechta, A. F., Conley, W. R. Recent Experiences in Plant-Scale Application of the
Settling Tube Concept Jour. WPCF, 43, 1724 (August 1971).
69. Neptune MicroFLOC, Inc.Application Criteria for Tube Settling In Activated Sludge
Plant Secondary Clarifiers, Technical Release No. 3.
70. Slechta, A., Neptune Micro-FLOC, Inc., Personal Communication (May, 1973).
Hennessey, T. L., City Engineer, Trenton, Michigan, Personal Communication (April,
71. 1973).
72.. Sparham, V. R., Improved Settling Tank Efficiency by Upward Flow Clarification,
Journal Water Pollution Control Federation, 42, 801 (May 1970, Part 1).
73. Sparham, V. R., Personal Communication.
7-39
-------
(74).Crockford, J. B., Sr., Sparham, V. R., Developments to Upgrade Settlement Tank
Performance, Screening, and Sludge De-watering Associated with Industrial Waste-
water Treatment, Proc. 27th Purdue Ind. Wastes Conference (May 3, 1972).
75. Pullen, K. G., Methods of Tertiary Treatment, Pebble and Wedge Wire Clarifiers, Lich-
field Rural District Council, Pollution Monitor, October/November (1972).
7-40
-------
CHAPTER 8
PHYSICAL STRAINING PROCESSES
8.1 General
Physical straining processes are defined for the purpose of this manual as those processes
which remove solids by virtue of physical restrictions on a media which has no appreciable
thickness in the direction of liquid flow.
Physical straining devices may be grouped according to the nature of their straining action.
(See Table 8-1).
8.2 Wedge Wire Screens
8.2.1 Inclined Screens
Inclined screens, are typified by the Hydrasieve, (Figure 8-1). made by C-E Bauer, Division
of Combustion Engineering Inc., or the Hydroscreen made by Hydrocyclonics Corporation.
These devices were originally developed in 1965 for the pulp and paper industry to dewater
and classify pulp slurries having solids contents of 6 percent or less (1). The units operate by
gravity and function as an inclined drainage board with a screen of wedge wire construction
having openings running transverse to the flow.
The first full scale municipal application of Hydrasieves was at the Ohio Suburban Waste-
water Treatment Plant at Huber Heights in 1967 treating raw wastewater(l).
8.2.1.1 Equipment Details
The screen consists of three sections with successively flatter slopes on the lower sections.
(Figure 8-2). The screen wires are triangular in cross section as shown in Figure 8-3, and
usually spaced 0.06 in. apart for raw wastewater screening applications. In the Bauer
unit, these wires bend in the plane of the screen, as illustrated in Figure 8-4. They are
straight and transverse to the flow in the Hydrocyclonics unit.
Above the screen and running across its width is a headbox; Figure 8-2 shows two possible
inlet designs. A light-weight hinged baffle at the top portion of the screen reduces flow
turbulence in the Bauer unit. To collect the solids coming off the end of the screen several
arrangements can be used, including a trough with a screw conveyor.
Inclined screening units are generally constructed entirely of stainless steel. Lighter units
with a fiber glass housing and frame costing about 25 percent less (1) may also be obtained.
Dimensions and capacities for hydrasieve units are given in Table 8-2.
8-1
-------
TABLE 8-1
PHYSICAL STRAINING PROCESSES
oo
Principal
Applications
Pretreatment &
Primary Treat-
ment
n tt
n a
Secondary and
Tertiary SS
Removal
it it
it it
Device
Inclined
wedge-wire
stainless
steel screens
Rotary
stainless steel
wedge wire
screens
Centrifugal
screens
Micro-
Screens
Diatomite
filters
Ultra-
Filters
Hydraulic
Capacity
High flow rates
4-16 gpm/in of
screen width
16-112gpm/sqft
40-100gpm/sqft
Medium flow
rates 3 to 10
gpm/sq ft
Medium flow rates
0.5-1.0 gpm/sq ft
Low flow rates
5 to 50 gpd/sq ft
Straining
Surface
Coarse
.01 to .06 in
(250-1500
microns)
Coarse
.01 to .06 in
(250-1500
microns)
Medium 105
Medium (")
1 5-60 microns
N/ACb)
Fine 99
(a) These values typify the range of solids filtered by the media. Removals are a function of media thickness and
not media opening sizes.
(b) Straining occurs through particulate mat of solids on screening surface.
-------
FIGURE 8-1
HYDRASIEVE SCREENING UNIT
Courtesy C-E Bauer
SLUDGE
FEED
HEAOBOX
DRAIN (OR
ALTERNATE FEED)
FIGURE 8-2
HYDRASIEVE SCHEMATIC
8-3
-------
FIGURE 8-3
SCREEN DETAIL
FIGURE 8-4
CURVED SCREEN BARS
Courtesy C-E. Bauer
8-4
-------
Width
~Ti
2
3.5
4.5
5.5
6.5
7
14
21
28
35
Depth
"IT"
3.5
4
5
5
5
9.5
9.5
9.5
9.5
9.5
Height
ft
5
5
7
7
7
7.3
7.3
7.3
7.3
7.3
TABLE 8-2
SPECIFICATIONS OF HYDRASIEVES
OVERALL DIMENSIONS
Capacity
mgd
350 0.2
550 0.4
650 0.9
800 1.2
1000 1.5
1800 2.9
3600 5.8
5400 8.7
7200 11.6
9000 14.5
8.2.1.2 Process Description and Design
Influent wastewater enters and overflows the headbox, on to the upper portion of the screen.
On the screen's upper slope most of the fluid is removed from the influent. The solids mass
on the following slope, because it is flatter, and additional drainage occurs. On the screen's
final slope the solids stop momentarily, simple drainage occurs, and the solids are displaced
from the screen by oncoming solids (2).
In test studies and actual installations hydrasieves have been operated satisfactorily at
loading capacities of 4 to 16 gpm per in. of screen width (1). This hydraulic capacity is a
function of the viscosity (which is a function of the temperature of the fluid), the solids load-
ing, and the spacing of the individual slots. Slot width is selected by actual tests using
sample screens. Once the slot opening has been chosen the screen's capacity per foot of
width can be determined from empirical relationships. Since work to date has not been suf-
ficiently extended to actual municipal wastewater conditions, pilot studies should be the
prime basis for design.
Little quantitative work has been done on the solids loading capacity of a hydrasieve but
generally speaking, for good performance, the influent should be dilute enough for smooth
flow over the weir. Unit sizes designed to accommodate more than 1 mgd are available;
however, for pilot studies a 6-in. wide by 22 in. long screen can be used provided flow
rates are limited to 5 to 10 gpm (1).
8.2.1.3 Operating Experience
At the 3 mgd Huber Heights plant in Ohio hydrasieves ahead of trickling filters have
effectively replaced primary clarifiers. Using 3 hydrasieves 72 in. wide and 54 in. long with
a slot opening of 0.06 in., an average suspended solids removal of 25 percent was
obtained while the units operated over a flow range of 1.5 to 4.5 mgd. Roughly 1 cubic yard
of solids was removed per million gallons of wastewater with an average solids content of 12
to,15 percent (3) (4).
8-5
-------
Although inclined screens cannot remove SS to the same extent as a sedimentation tank,
they have been favorably received by operators because they do an excellent job of
removing trashy materials which may foul subsequent treatment of sludge handling units.
Their ability to remove fine grit is limited by size openings. Separate grit removal
equipment, if needed, should be installed after the inclined screens. In a pilot study at South
Buffalo Creek Sewage Treatment Plant at Greensboro, North Carolina, hydrasieve
suspended solids removals ranged from 10 to 30 percent with an average removal of 20 per-
cent (5).
At the U.S. EPA Blue Plains pilot study in Washington, D.C. (6) hydrasieves were installed
in an effort to eliminate operational problems of debris collection on the mixers and plugg-
ing of recycle and waste discharge lines. Although the screens eliminated these problems,
suspended solids removals varied from only 7 to 11 percent. The low removals were attrib-
uted to the wastewater's age (24 to 48 hours) (7).
An installation list is included (Table 8-3). Some of these installations are temporary;
hydrasieves are being used for short term alleviation of excess solids and flows coming into
plants which are to be abandoned when new facilities are built.
Operating experience in these installations varies as to cleaning and maintenance
requirements. Generally, a daily washing of the screen surfaces, which takes about 5
minutes, is sufficient for good screen performance. Washing is normally done with steam or
hot water to remove grease which accumulates and blinds the screen preventing passage of
wastewater through the screen, and resulting in poor separation (5).
Daily steam cleaning proved necessary at Freehold, N.J. (8) but other installations such as
Huber Heights required only monthly steam cleaning (3). Grease build-up requiring steam
cleaning appears to be related to low air and wastewater temperatures, exposure of units
and high grease content in wastewater.
Incidental to the removal of suspended solids in this process is the aeration of the separated
water. At the Huber Heights plant raw wastewater impinging on a Bauer screen has been
found to be aerated up to a level of 2 or more mg/ 1 of dissolved oxygen (3). A noticeable
reduction in odors from the grit removed in the subsequent chamber has also been claimed
along with the elimination of scum in the digester (3).
8.2.2 Rotating Wedge Wire Screens
Hydrocyclonics, to overcome grease blinding problems of its own wedge wire screen,
developed a rotating wedge wire screen which backwashes itself (Figure 8-5). Wastewater
passes vertically downward from the outside to the inside of the drum by gravity. The
screened wastewater then passes out through the lower half of the drum to a collection
trough.
8-6
-------
TABLE 8-3
WEDGE WIRE SCREENS MUNICIPAL TREATMENT INSTALLATIONS
All units below are Bauer units unless otherwise indicated:
PLANT & LOCATION REMARKS
Ohio Suburban Water Co. 3-4 mgd
Huber Heights, Ohio
Rochelle Treatment Plant
Rochelle, Illinois
Prophetstown Treatment Plant
Prophetstown, Illinois
Corinna Treatment Plant
Corinna, Maine
Bucks County Sewage Authority
Mr. MacNamara Executive Director
(Hydrocyclonics Units)
Upper Gwynedd
Towamencin Municipal Authority
Lansdale, Pa.
Irvine Ranch Water District
Irvine, California
Hercules—AWT Div.
Freehold Township, N.J.
Mr. Ron Lee, Engineer
Montgomery County Commissioners 10 mgd
Moraine, Ohio
STP Rogersville, Tennessee
(Hydrocyclonics Units)
Blue Plains Pilot Study
Washington, D.C.
S. Buffalo Creek STP 0.03 mgd
Greensboro, N. Carolina
8-7
-------
FIGURE 8-5
ROTATING WEDGE WIRE SCREEN AT NORTH CHICAGO S.T.P.
(Courtesy of Hydrocyclonics Corp.)
-------
Solids are retained on the outside of the drum and are removed by a fixed scraper blade. A
screen spacing of 0.06 inches is recommended for service on raw wastewater. In comparison
with static screens, the manufacturer claims the rotating units require less maintenance,
lower operating head and smaller space and produce dryer solids (9). Table 8-4 shows com-
parative design data for rotating and stationary wedge wire screens. (10).
TABLE 8-4
Parameter
S.S. Removal, percent
Flow rate
Wire Spacing
DATA SHEET-WEDGE WIRE SCREENS
Inclined Rotary
5 to 25 5 to 25
4 to 16 gpm 15 to 112 gpm/sq ft
per inch of
screen width
.01 to .06 in. .01 to .06 in.
Percent solids by wt.
Volume of solids
produced
12 to 15
1 to 2 cu yd of
solids per million
gallons of waste-
water
16 to 25
8.3 Microscreening
8.3.1 General Description
As shown in Figure 8-6 in its usual configuration a microscreen unit consists of a motor
driven rotating drum mounted horizontally in a rectangular chamber. A fine screening
media covers the periphery of the drum. Feedwater enters the drum interior through the
open end and passes radially through the screen with accompanying deposition of solids on
the inner surface of the screen. At the top of the drum pressure jets of effluent water are
directed onto the screen to remove the mat of deposited solids. The dislodged solids
together with that portion of the backwash stream which penetrates the screen are captured
in a waste hopper as shown in Figure 8-7. Solids flushed from the unit are sent to sludge
handling systems or recycled to the head of the plant. Units may be equipped with
ultraviolet lights to control biological growth on the screen media. Effluent passes from the
chamber over control weirs oriented perpendicular to the drum axis.
8-9
-------
>
FIGURE 8-6
TYPICAL MICROSCREEN UNIT
(Courtesy of Cochrane Division, Crane Co.)
-------
BACKWASH HOPPER
DRUM LIFT
SPRAY SYSTEM
-
MESH SYSTEM
FIGURES-?
MICRO-MATIC®STRAIM R
(Courtesy of Zurn I ml.. Inc.)
-------
8.3.2 Functional Design
The functional design of a microscreen unit involves:
1. Characterization of suspended solids in feed as to concentration and degree of
flocculation, as these factors have been shown to affect microscreen capacity,
performance and back washing requirements (9)(11)
2. Selection of unit design parameters which will assure sufficient capacity to meet
maximum hydraulic loadings with critical solids characteristics, and provide the
required performance over the expected range of hydraulic loadings and solids
characteristics.
3. Provision of backwash and supplimental cleaning facilities to maintain the design
capacity.
Table 8-5 shows typical values for microscreen and backwash design parameters for solids
removal from secondary effluents. Similar values would apply to direct microscreening of
good quality effluent from fixed film biological reactors such as trickling filters or rotating
biological contactors, where the microscreens replace secondary settling tanks (9). This
application is not widely practiced, however.
Microscreening has been used for the removal of algae from uncoagulated lagoon effluents.
At Bristol, England, algae reductions of 1565 to 450 algae per ml and 989 to 168 algae per
ml were achieved on astrerionella, cyclolella and synedra (12). However many classes of
algae, e.g. chlorella, are too small to be removed, even on fine screens (23 microns) and ex-
cessive loadings (up to 2 x 106 algae per ml) make this application a limited one.
The parameters of mesh size, submergence, allowable headless and drum speed [rpm =
peripheral speed/ -^ (diameter)j are sufficient to determine the flow capacity of a
microscreen with given suspended solids characteristics (13).
8-12
-------
TABLE 8-5
MICROSCREEN DESIGN PARAMETERS
Hydraulic
Loading
Typical Value
20-25 microns
75 percent of height
66 percent of area
5-10 gpm/sq ft
of submerged drum
surface area
Remarks
Range 15-60 microns
Head-loss (Hi.)
through Screen
3-6 in
Maximum under extreme con-
dition: 12-18 in. Typical
designs provide for overflow
weirs to bypass part of flow
when head exceeds 6-8 in.
Peripheral
Drum Speed
Typical Diameter
of Drum
Backwash Flow
and Pressure
15 fpm at 3 in. (Ht)
125-150 fpm at
6 in. (Ht)
10ft
2 percent of throughput
at 50 psi
5 percent of throughput
at 15 psi
Speed varied to control
extreme maximum speed
150 fpm
Use of wider drums
increases backwash require-
ments.
Among these parameters peripheral speed, hydraulic loading and major variations in mesh
size also affect performance on a given feed flow. In addition, drum speed and diameter
affect the wastewater flows and pressures needed to effect proper cleaning of the screen.
8-13
-------
8.3.3 Hydraulic Capacity
The Jilterability index developed by Boucher (14) quantifies the effect of the feed solids
characteristics on the flow capacity of a particular fabric. Boucher assumed that at any
constant laminar flow rate the headless, AP in ft, through any given strainer fabric would
increase exponentially with the volume passed per unittarea (V. in cu, ft/sq ft ):
AP _ PIV
In the above relation the filterability index is the exponential rate constant I (in I/ ft).
From the filterability index concept Mixon (13) developed hydraulic capacity relations for
continuous operation of a rotating drum microscreen, which can be expressed as follows:
Where:
H = mean flow velocity through submerged screen area (fps)
Q = total flow through microscreen (cfs)
A = submerged screen area (sq ft)
P = pressure drop across screen (ft)
CF = fabric resistance coefficient (ft/fps or sec) (clean fabric headless at 1 fps
approach velocity)
I = filterability index (1 / ft)
0 = decimal fraction of screen area submerged
R = drum rotational speed (rpm)
The expression'AP/CF represents the initial flow velocity through the clean screen as it en-
ters submergence.CF is a particular characteristic of the screen fabric, varying inversely
with mesh opening size as follows:
Mesh Size Fabric Resistance CF
mu ft/fps
15 3.6
23 1.8
35 1.0
60 0.8
8-14
-------
Limits on A P reflect screen fabric mechanical strength and expected operating conditions
for the unit. A typical value is 0.5 ft at normally-expected maximum flow.
The relation of parameters in the expression (I<£/R) shows that the effect of a higher index
or faster buildup of headless on the screen may be offset by maintaining a higher drum
rotational speed.
Figure 8-8 is a graphical representation of the above relation which Mixon obtained by
plotting Q/A against A.P/CF for various values of the parameter I0/R.
The graph shows lines of constant value for the ratio
E= Q/A
which is the ratio of the mean velocity through the screen to the initial velocity when the
screen enters submergence. Recognizing the effect of drum speed on performance Mixon
suggested selecting 107 R to keep the ratio E below 0.5. Above this limit he assumes that
insufficient opportunity is given for a mat to form on the drum and solids removal efficien-
cy is likely to suffer.
The filterability index may be determined by Boucher's laboratory procedure (14), by field
testing with an apparatus available from Crane Co. (11) or by analysis of test data from
pilot microscreen units using relations such as those proposed by Mixon or Boucher. In
some cases where numerous values of T have been obtained, a relationship to influent SS
loading may be obtained (15). Solids loading limitations such as those noted by Lynam, et
al, (16) would have broader applicability if they were related to a maximum filterability
index under which a particular microscreen could maintain a given capacity.
8.3.4 Performance
Suitable relationships have not been developed for quantitative predictions of microscreen
performance from knowledge of influent characteristics and key design parameters. Where
performance must be predicted closely, pilot studies should be made. Where close
prediction is less critical, performance data from other locations with generally similar
conditions may serve as a guide.
Table 8-6 provides performance data for a number of microscreen installations for SS
removal from secondary effluents including the first such installation made at Luton
Sewage Works, England, in the early 1950's. Table 8-7 lists additional American
installations provided by two manufacturers.
8-15
-------
O.004
O.002
0.01
O.O2
O.04 0.06O.080.I
AP CF FT/SEC
0.4 O.6 O.8 I.O
TYPICAL DESIGN RANGE
5-10 GPM/SQ FT
FABRIC
MARK 0
MARK I
MESH- MU
23
35
FIGURE 8-8
MICROSCREEN CAPACITY CHART(13)
8-16
-------
8-6
MICROSCREEN INSTALLATIONS
Drum
Dia.
SrTApn
LOCATION
Luton
Bracknell
Hambledon R.D.C.
Elmbridge
Leighton-
Linslade U.D.C.
Fleet U.D.C.
Esher U.D.C.
Hatfield R.D.C.
The Borough of
Bury St. Edmonds
Franklin Township
STP Murraysville,
Pa.
Letchworth
Basingstoke
Euclid, Ohio
Euclid, Ohio
Euclid, Ohio
Lebanon, Ohio
Hanover Park, 111.
MSD North Side
STP Chicago, 111.
Country
England
England
England
England
England
England
England
England
U.S.A.
England
England
U.S.A.
U.S.A.
U.S.A.
U.S.A.
U.S.A.
U.S.A.
Influent Source
Trickling Filter
Trickling Filter
Trickling Filter
Trickling Filter
Trickling Filter
Trickling Filter
Trickling Filter
Trickling Filter
Trickling Filter
Final-Settling
Tanks
Act. Sludge Final
Clarifiers
Act. Sludge Final
Clarifiers
Act. Sludge with
Chem.Precip.in
Primary Clarif.
(Fed)
Act. Sludge with
Chem.Precip.in
Final Clarif.
(Fed)
UNOX Final
Clarifiers
Act. Sludge
Final Clarif.
Act. Sludge
Final Clarif.
Act. Sludge
Final Clarif.
Width
ft
7.5x5,
" ~ 1
t
10x10,
10x10,
7,5x5,
,
10x10,
10x10,
5x3,
10x10,
2.5x2,
2.5x2,
2.5x2,
5x1,
5x1,
10x10,
12.5x30
Mesh
mu
60
35
35
23
35
35
-
.
23
23
23
23
23
23
23
35
23
23
, 23
Hydraulic
Load on
Submerged
Mo of Area Plant Flow
Units Max. Avg. Max. Avg.
gpm/sq ft mgd mgd
9.0-3
2 6.3 2.2 6.3 2.2
10.8 2.5 2.5 0.6
10.8 2.5 2.5 0.6
6.8 3.9 2.0 1.2
2 6.0 2.0 3.6 1.2
3 - 10.8 (Design Flow) .
3 - 2.55 (Design Flow)
9.0 (Design Flow)
5 - - 5.3 (Max. Flow to
Date)
2 7.8 - 4.0
1 4.3 3.3 Pilot Study
S 1.5 1.0 3.2 2.2
1 2.5 1.25 Pilot Study
40 gpm 20 gpm
1 2.5 1.25 Pilot Study
40 gpm 20 gpm
1 2.5 1.25 Pilot Study
40 gpm 20 gpm
1 1
1
1 S.3 2.6 1.5 0.8
15 (Design Flow)
Average
Suspended
Solids
Influent
mg/1
14
20
16.2
14
14
29
15
19
14
28
37
17
13.1
54
38
65
27
17
6-28
10
Effluent
mg/1
8
11
6.9
8
7.7
11
6
9
8
7
6
6.6
3.9
8
10
21
7
2
4-11
3
Removed Reference
percent
45 19
45 19
57
45 19
45
60 19
60 9,19
60 9
43 9,19
75 9
83+10 9
62 9,19,20
70 9,19
85 17
74 17
68 17
73 15
83 15
55(ave.) 16
67 9
-------
TABLE 8-7
MUNICIPAL MICROSCREENER INSTALLATIONS
Location
Arthur Bloom Apts.
Lancaster, Pa.
Ecological Utilities
North Miami, Fla.
City of Murfreesboro
Murfreesboro, Tenn.
MSD Chicago
Lemont, 111.
City of Cookeville
Cookeville, Tenn. 7.2
City of Dayton
Dayton, Tenn. 5.4
Department of Public Works
Erie, Pa. 45.0
Village of Pepperpike
Pepperpike, Ohio 1.0
Borough of Bellefonte
Bellefonte, Pa. 3.0
Good Samaritan Hospital
Islip, New York 0.2
Plant
Flow
mgd~
0.1
2.7
4.0
Units
No.
Courtesy
2
1
2
Unit
Sizes
D x L, ft
Zurn Industries
4x2
10x10
10x10
Straining
Media*
microns
20
(polyester)
20
(polyester)
20
(polyester)
10x10
10x10
10x10
10x15
6x6
6x6
4x4
21
21
20
(polyester)
21
21
21
21
* All fabrics stainless steel unless otherwise indicated.
8-18
-------
TABLE 8-7 (continued)
MUNICIPAL MICROSCREENER INSTALLATIONS
Location
Borough of Carroltown
Carroltown, Pa.
Cincinnati Dept. of Sewers
Cincinnati, Ohio
Muncie Mall
Muncie, Indiana
Opalaka Sewage Plant
Chesterland, Ohio
I.B.M.
Essex Junction, Vt.
Hot Springs Village
Hot Springs, Arkansas
Union "76" Oil Co.
Clarion, Pa.
Deer Creek State Park, Ohio
Oakbourne Hospital,
Westchester, Pa.
Sugar Creek STP,
Greene County, Ohio
Little Miami STP,
Greene County, Ohio
Westminster, Md.
Hammond, Ind.
East Wheatfield Township, Pa.
Plant
Flow
mgd
1.0
0.5
0.2
0.3
0.2
0.3
(Courtesy
0.1
0.3
0.1
6.0
6.0
3.0
1.5
0.1
Units
No .
2
1
1
1
1
1
of Crane Co.)
1
1
1
2
2
1
1
1
Unit
Sizes
DxL, ft
4x4
4x4
4x4
4x4
4x4
4x4
5x1
5x3
5x1
10x10
10x10
10x10
7.5x5
5x1
Straining
Media*
microns
21
21
21
21
35
21
23
23
23
23
23
23
23
23
8-19
-------
TABLE 8-7 (Cont'd)
MUNICIPAL MICROSCREENER INSTALLATIONS
Location
Hartville, Ohio
Louisville, Ky.
Browntown, Minn.
Salisbury, N.C.
Allen County, Ohio
Fairmont, Minn.
Parkway, Md.
Penn State University, Pa.
Bolingbrook, Illinois
Park Forest, Illinois
Sugar Creek, Ohio
Union Oil Co.
Harrisburg, Pa.
Chicago Metro Sanitary District
Chicago, Illinois
Ursuline Academy, Ohio
Jackson Township, N.J.
FWPCA Research Project, Phila.
Akron STP, Ohio
Wm. Henry Apts., Dowington, Pa.
Plant
Flow
mgd
2.5
0.8
0.1
6.0
2.5
6.0
15.0
0.3
0.5
2.0
0.3
0.1
2.0
0.3
0.1
0.2
3.0
0.2
Units
No.
2
2
1
2
2
2
5
1
1
2
1
1
1
1
1
1
1
1
Unit
Sizes
D x'L, ft.
7.5x5
5x3
5x1
10x10
7.5x5
10x10
10x10
5x3
7.5x5
7.5x5
5x3
5x1
10x10
5x3
5x1
5x3
7.5x5
5x1
Straining
Media*
microns
23
60
23
23
23
23
23
23
60
23
23
23
23
35
35
23
60
35
8-20
-------
TABLE 8-7 (continued)
Location
Franklin Twp., Pa.
Hempfield Twp., Pa.
Chelsea Ridge Apts., N.Y.
Harpeth Valley, Tenn.
University School
Cleveland, Ohio
Willoughby Hills
Cleveland, Ohio
Lionville, Pa.
Margate STP, Fla.
Lauderhill STP, Fla.
Bel-Aire STP,
Miami, Florida
Homestead, Fla.
Upper Sandusky, Ohio
Petosky, Michigan
Ravenna, Ohio
Bolingbrook STP, 111.
Commonwealth Edison, 111.
Lucas County, Ohio
Pymatunihg State Park, Pa.
Plant
Flow
mgd
4.0
6.0
0.2
1.5
0.1
0.1
0.7
3.5
3.5
0.9
0.2
1.5
2.5
1.5
0.5
2.0
0.2
0.1
Units
No.
2
2
1
1
1
1
2
2
2
1
2
. 2
2
2
1
1
2
1
Unit
Sizes
D x L, ft
10x10
10x10
5x3
7.5x5
5x1
5x1
5x3
10x10
10x10
7.5x5
5x1
7.5x5
10x10
7.5x5
7.5x5
10x10
5x1
5x1
Straining
Media
microns
23
35
23
23
23
23
23
23
23
23
23
23
23
23
60
23
23
35
8-21
-------
Figure 8-9 shows average operating results from a number of British tertiary microscreen
installations with various hydraulic loadings.
Figure 8-10 presents the results of three extended British studies on microscreening of
trickling filter and activated sludge secondary effluents.
Some general conclusions can be made about the microscreen as a device for removing SS
from secondary effluents:
1. Under best operating conditions microscreen units can reduce solids to as low as
5mg/l.
2. Although the SS removal pattern is irregular, performance tends to be better at
lower hydraulic loadings (Figure 8-10a).
3. Increases in influent suspended solids are reflected in the effluent but with
noticeable damping of peaks (Figure 8-10).
4. Microscreens are applicable in place of clarifiers to polish effluent from low rate
trickling filters, if the solids are generally low in concentration and well flocculated
(Figure 8-10b).
Other investigations provide insights beyond those cited above:
Data from Lebanon, Ohio, (15) show better removal with smaller mesh sizes.
Commenting on how solids characteristics affect microscreen performance, one British
study (8) notes that a clear non colloidal secondary effluent containing a reasonable amount
of suspended solids would result in a better effluent than colloidal effluent containing less
suspended solids.
Chemical application can unfavorably alter solids characteristics: Lynam, et al. (16)
reported poor removals when applying an alum flocculated secondary effluent directly to a
microscreen. In contrast, at Euclid, Ohio (17) a microscreen removed 74 to 85 percent of
the_ solids from the settled effluent of an activated sludge pilot plant with mineral addition
for phosphorus removal. The better performance at Euclid could be attributed to a tougher
biological nature of the effluent solids.
Microscreen suppliers (9) (18) stress the importance of minimizing shearing action on
microscreen influent to avoid breaking up flocculated particles. This is advanced as a reason
for settling limits on drum peripheral speed and for avoiding pumping ahead of microscreen
units.
Lynam, et. al, (16) indicate that microscreening at lower drum speeds yields better quality
effluent. This is attributed to better straining action through the thicker mat of solids which
builds up at low speeds.
8-22
-------
• PLANT OPERATED AT MAXIMUM
FLOW RATES OF 9.2 TO 10.8 gpm/s.f.
O PLANT OPERATED AT MAXIMUM
FLOW RATES OF 6.0 TO 6.8 gpm/s.f.
20
CO
V)
u.
u.
Ill
10
10 20 30
INFLUENT SS Mg/l
40
FIGURE 8-9
MICROSCREEN REMOVAL AT VARIOUS FLOW RATES
(After Isaac and Hibbert (19))
8-23
-------
CONCENTRATION OF SUSPENDED SOLIDS
8OK A A ~~ Filter Effluent
1 ft A - — Humus Tank Effluent
1 Microstrainer Effluent
SUSPENDED
SOLIDS
(p.p.m.)
OLa
7 21 5 19 2 1630142811 25 8 22 6 2O 3 17 1 15291226
April May June July Aug Sept Oct Nov Dec Jan
1962 1963
A. HAftPCNOEN SEWAGE WORKS
(Courtesy of Crane Co.)
§ 30
20
10
400
300
u. B
I °7
SUSPENDED SOLIDS (influent)
FLOW
INDICATES TIMES WHEN MICF.OSTBAINER WASHWATER JETS CHANGED
'TREATED WITH HYPOCHLORITE
SUSPENDED SOLIDS (effluent)
•v
—i- t—'-->_!_•> ' I I- L i i-i J I LI t I I I __ ]_ . r ] j | I I i i i I I ] i I | j | | LI | 11 j 111 I (
21 5 19 2 16 30 14 28 II 25 8 22 6 20 3 17 I 15 29 12 26 9 23 9 23
APR. MAY JUNE JULY AUG. SEPT. OCT. NOV. DEC. JAN. FEB. MAR.
1966 1967
-400"6
300^
200 g
u.
B. LETCHWORTH WATER POLLUTION CONTROL WORKS
Data were obtained from unulysi.s ot continuous records, each point representing
weekly mean hourly readings. Broken lines indicate break in operation of plant
FIGURE 8-10
MICROSCREENING OF TRICKLING FILTER
PLANT EFFLUENT
8-24
-------
SUSPENDED 25
SOLIDS
(mg E) 20
MICROSTRAINER
EFFLUENT
Sept Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec
1968 1969
C. FLEET U- D-C.
(Courtesy of Crane Co.)
FIGURE 8-PO (Continued)
8-25
-------
As shown in Figure 8-8 under peripheral speed and hydraulic loading limits set by
manufacturers and regulatory authorities, the ratio E for most microscreen designs actually
falls below 0.1, a lower limit suggested by Mixon for most efficient utilization of equipment.
The wide range of suggested design values underlines the need for developing quantitative
relations between removal efficiency and key design parameters.
8.3.5 Microscreen Construction
The basic screen support structure is a drum shaped, suitably stiffened rigid frame
supported on bearings to allow rotation. Designs using water lubricated axial bearings or
greased bearings located on the upper inside surface of the rotating drum allow
submergence well above the central axis.
Both plastic (polyester) and stainless steel are used for the microscreen media itself. Greater
mechnical strength, especially at higher temperatures, is the prime advantage stated for
stainless steel (9) (18). Greater economy and chemical resistance are pointed to as
advantages of plastic (18).
Depending on manufacturer, screen fabric is supplied either in small sections (8 in. x 12 in.)
supported by and fastened directly to the drum frame or in larger (18 in. x 24 in.) panels
consisting of fabric integrally bonded to a grid like supporting mesh of stainless steel. These
panels are attached to the drum frame.
One manufacturer offers a microscreen unit with an accordion or pleated outer surface to
achieve up to 30 percent more filtering surface within the same general dimensions of regu-
lar designs. The unit is 12.5 feet by 30 feet and has a rated capacity of 15 mgd. This unit is
shown in Figure 8-11.
In the past cast bronze and cast carbon steel were used as drum and frame construction
materials. The present practice is to use fabricated carbon steel. Generally, smaller units are
factory assembled in steel tanks while large units are placed in concrete structures.
Table 8-8 illustrates the approximate size and power requirements for various microscreen
units.
8.3.6 Screen Fabric and Operating Headless
Microscreen fabrics normally are woven of stainless steel or plastic (polyester with
polypropylene supporting grid) with openings in the range of 15 to 60 microns. .. .
Plastic fabric is less subject to chemical attack by strong chlorine or acid cleaning solutions.
Stainless steel can better withstand temperatures encountered in steam cleaning..' '
Suggested operating headloss limits for microscreens are based on observation of the effect
of differential head on screen life. Standard design calls for a 3 in. headloss at average flows
8-26
-------
FIGURE 8-11
MICROSCREEN UNIT WITH PLEATED OUTER SURFACE
(Courtesy of Cochrane Div., Crane Co.)
U S EPA Headquarters Library
Mail code 3404T
1200 Pennsylvania Avenue NW
" Washifiaton, DC 20460
2o2-bbb-0556
8-27
-------
TABLE 8-8
TYPICAL MICROSCREEN POWER AND SPACE REQUIREMENTS
Drum. Sizes
Source
Code
A
A
A
A
Diam.
It
5.0
5.0
7.5
10.0
Length
ft
1.0
3.0
5.0
10.0
Floor Space
Width Length
ft ft
8 6
9 14
11 16
14 22
Motors
Drive
BHP
0.50
0.75
.2.00
5.00
Wash Pump
BHP
1.0
3.0
5.0
7.5
Approx. Ranges
of Capacity
mgd
0.07-0.15
0.2 - 0.4
0.5 - 1.0
1.5 -3.0
B
B
B
4.0
6.0
10.0
4.0
6.0
10.0
7 15
10 17
14 22
0.75
2.00
5.00
1.0
1.5
5.0
0.2 - 0.4
0.5- 1.0
1.5-3.0
Code A Courtesy Crane Co., Cochrane Div.
Code B Courtesy Zurn Industries
and 6 in. for normally-expected maximum flows (9) (18). For occasional peaks (less than 3
percent of time) headlosses up to 24 in. can be tolerated. Crane Co. indicates that stainless
steel screens operated under the above conditions would have a life of 10 years; if operated
continuously at a 24 in. headloss the same screen might only last 6 months. (9)
8-28
-------
8.3.7 Hydraulic Control
Hydraulic control of microscreening units is effected by varying drum speed in proportion
to the differential head across the screen. The controller is commonly set to give a
peripheral drum speed of 15 fpm at 3 in. differential and 125 to 150 fpm at 6 inches (9)
(18). In addition, backwash flow rate and pressure may be increased when the differential
reaches a given level (9) (18).
The operating drum submergence is related to the effluent water level and headless through
the fabric. The minimum drum submergence value for a given installation is the level of
liquid inside the drum when there is no flow over the effluent weir. The maximum drum
submergence is fixed by a bypass weir which permits flows in excess of unit capacity to be
bypassed; at maximum submergence the maximum drum differential should never exceed
15 inches.
Effluent and bypass weirs should be designed as follows:
1. Select drum submergence level (70 to 75 percent of drum diameter) for no flow
over the effluent weir.
2. Locate top of effluent weir at selected submergence level.
3. Determine maximum flow rate.
4. Size effluent weir to limit liquid depth in effluent chamber above the weir to 3 in.
at the maximum flow rate.
5. Position the bypass weir 9 to 11 in. above effluent weir. (3 in. head on effluent weir
maximum flow plus 6 to 8 in. differential on drum at maximum drum speed
and maximum flow).
6. Size bypass weir length to prevent the level above effluent weir flow exceeding 12
to 18 in. at peak maximum flow or overflowing the top of the backwash collec-
tion hopper.
8.3.8 Backwashing
Backwash jets are directed against the outside of the microscreen drum as it passes the
highest point in its rotation. About half the flow penetrates the fabric,dislodging the mat of
solids formed on the inside (15). A hopper inside the drum .receives the flushed-off solids.
The hopper is positioned to compensate for the trajectory that the solids follow at normal
drum peripheral velocities.
Microscreen effluent is usually used for backwashing. Straining is required to avoid
clogging of backwash nozzles. The in-line strainers used for this purpose will require
periodic cleaning; the frequency of cleaning will be determined by the quality of the
backwash water.
8-29
-------
The backwash system used by Zurn employs two header pipes; one operates continuously at
20 psi, while the other operates at 40 psi when the unit receives a high solids loading. The
Crane system also uses two sets of jets but both operate continuously at pressures from 15
to 55 psi. Under normal operating conditions these jets operate at 35 psi. Once a day they
are operated at 50 psi for l/2 hour to keep the jets free of slime build-up. Should this
procedure fail to keep the jets clean, the pressure is raised to 55 psi. At this pressure the
spring loaded jet mouth widens to allow for more effective cleaning.
Backwash pressure is also increased to compensate for heavy solids loadings which require
the higher pressure for thorough cleaning. Crane reports that no major problems have been
encountered with this jet design (9).
Prior to 1967 Crane designed backwash systems to operate only at 15 psi. A pilot study in
Letchworth England (20) showed the superiority of the higher pressure system. Results of
this study showed:
1. Operation at 50 psi, as opposed to 15 psi, increased the process flow capacity
30 percent.
2. Suspended solids concentration in the backwash increased from 260 mg/ 1 at 15
psi to 425 mg/ 1 at 50 psi.
3. Water consumption of the jets as a percent of process effluent decreased from
5 percent at 15 psi to 2 percent at 50 psi.
In general, backwash systems are operated at as low a pressure as possible consistent with
successful cleaning. High pressure operation incurs added system maintenance, particularly
jet replacement, and is used only as needed.
8.3.9 Supplemental Cleaning
Over a period of time screen fabrics may become clogged with algal and slime growths, oil,
and grease. To prevent clogging, cleaning methods in addition to backwashing are neces-
sary.
To reduce clogging from algae and slime growth, Crane Company recommends the use of
ultraviolet lamps placed in close proximity to the screening fabric and monthly removal of
units from service to permit screen cleaning with a mild chlorine solution. While most liter-
ature sources say ultraviolet lamps are of value, one authority (21) feels these lamps are
uneconomical because they require frequent replacement. Zurn Industries claims that, be-
cause their screening fabric is completely bonded to the supporting material, crevices where
algae become lodged are eliminated and backwashing alone is sufficient to remove algal and
associated slime growths (18).
8-30
-------
Where oil and grease are present, hot water and/or steam treatment can be used to remove
these materials from the microscreens. Plastic screens with grease problems are cleaned
monthly (9) with hot water at 120° F (18) to prevent damage to the screen material. Down
time for cleaning may be up to 8 hours.
8.3.10 Operation
In starting a microscreening unit care should be taken to limit differential water levels
across the fabric to normal design ranges of 2 to 3 inches. For example, while the drum is
being filled it should be kept rotating and the backwash water should be turned on as soon
as possible. This is done to limit the formation of excessive differential heads across the
screen which would stress the fabric during tank fill-up.
Leaving the drum standing in dirty water should be avoided because suspended matter on
the inside screen face which is above the water level may dry and prove difficult to remove.
For this reason introducing unscreened waters, such as plant overloads, into the
microscreen effluent compartment should also be avoided (18).
If the unit is to be left standing for any length of time the tank should be drained and the
fabric cleaned to prevent clogging from drying solids.
8.4 Other Screening Devices
Conventional mesh screens have not been used with success in municipal wastewater
treatment. Recently, however, a centrifugal screen, the Sweco Concentrator, has
demonstrated its effectiveness. In this unit, influent is directed against the inside
of a rotating cylindrical screen cage (See Fig. 8-12). It is claimed that the rotational
speed (centrifugal force of 3 to 6 gravities) increases hydraulic capacity and, together with
the impingement angle, permits separation of solids finer than the screen openings (150 to
165).
Separated solids and the rejected portion of the liquid flow are removed from the bottom
of the unit while effluent is taken off at the periphery. Screen blinding is cleared by timer-
actuated spray cleaning systems which direct water jets against both the inner and outer
screen surfaces.
At Contra Costa County, California, a 60 inch unit treats 0.9 mgd wastewater containing
about 200 mg/ 1 SS (22). The waste flow is split into two streams, a small volume (15 per-
cent) concentrated stream and a supernatant (85 percent of influent) stream. The concen-
trated stream is settled in the conventional primary basin (formerly overloaded but now ca-
pable of good solids reduction at the lower hydraulic loading). The supernatant is treated by
flotation (the Sweco concentrator does not specifically remove floatables but thoroughly
aerates the wastewater aiding subsequent flotation) and finally, flotation and settled con-
centrate effluents are mixed and chlorinated before disposal. Overall SS reductions for the
system including concentration, flotation, settling and chlorination are reported as 70 to 80
percent. Similar removals are claimed for settled secondary effluent from aeration pro-
cesses (23). Design flow rates are claimed to range from 40 to 100 gpm/sq ft. The con-
8-31
-------
Influent Distribution Pans
Removable Screen Panels
X
ui
i j
Effluent Discharge
Concentrate Discharge
Rotating Screen Cage
Concentrate Collector
Influent
Effluent Collector
• FIGURE 8-12
THE SWECO CONCENTRATOR
(Courtesy of SWECO, Inc.)
-------
centrator is of epoxy-lined steel and screen construction is of stainless steel or polyester in
plastic frames. During operation the outside of the screen is washed every 20 minutes with
cold clarified water and the inside with 5 to 7 gpm hot water.
8.5 Diatomaceous Earth Filters
Diatomaceous earth (DE) filters have been applied to the clarification of secondary
effluents at pilot scale. No full-scale installations have been characterized in the literature.
They produce a high quality effluent but appear unable to handle the solids loadings
normally expected in this application.
DE filtration utilizes a thin layer of precoat formed around a porous septum to strain out
the suspended solids in the feedwater which passes through the filter cake and septum. The
driving force can be imposed by vacuum from the product side or pressure from the feed
side. As filtration proceeds, headless through the cake increases due to solids deposition
until a maximum is reached. The cake and associated solids are then removed by flow
reversal and the process is repeated. In the cases where secondary effluents have been
treated by this process, a considerable amount of diatomaceous earth (body feed) has been
required for continuous feeding with the influent in order to prevent rapid buildup of
headlosses. Generally, the DE filtration process is capable of excellent removal of
suspended solids but not colloidal matter.
A wide variety of diatomaceous earth (diatomite) grades are available for use. As might be
expected, the coarser grades have greater permeability and solids-holding capacities than do
the finer grades which will generally produce a better effluent. Some grades of diatomite are
pretreated to change their characteristics for improved performance. A number of vessel
configurations are available, with open-basin vacuum and vertical pressure designs most
common. (See Figures 8-13, and 8-14.)
Design criteria for diatomite filters have been discussed by Bell (24). The filtration cycle can
be divided into two phases, run time and down time. Down time includes the periods when
the dirty cake is dislodged from the septum and removed from the filter and when the new
precoat is formed. Run time commences when the feed is introduced to the filter and ends
when a limiting headless is reached. The single most important factor in secondary effluent
filtration by DE filters is the amount of body feed required during the filtration or run time.'
The body feed rate is the largest operating cost factor and strongly affects the operating
economics of the process. Similarly, it is related to cycle time between backwashing which
determines the installed filtering area, hence the capital cost economics. A ratio of 5 to 6
mg/ 1 of body feed per JTU of influent turbidity was required at San Antonio and the
possible need for a higher ratio was suggested (25). Another pilot study used a variety of
ratios and filtration rates and used both pressure and vacuum systems for secondary
effluent filtration (26). Some results from this study are shown in Table 8-9. Both studies
indicate that a precoat of about 0.1 Ib/sq ft of filter area and greater than 6 mg/ 1 of body
feed per JTU of turbidity should.be used.
An English study with a reversible-flow DE unit also resulted in uneconomical operating
conditions due to excessive body feed, short filter runs and high backwash water
8-33
-------
X
CORROSION PROOF
FILTER ELEMENTS
CORROSION PROOF
TANK
FILTERED WATER
OUTLET
ANGLE IRON
FRAME
RAW WATER
INLET
BAFFLE
FILTERED WATER
MANIFOLD
DRAIN
FIGURE 8-13
VERTICAL LEAF VACUUM FILTER
(Courtesy of Johns-Manville)
-------
FIGURE 8-14
VERTICAL LEAF PRESSURE FILTER, VERTICAL TANK
(Courtesy of Johns-Manville)
8-35
-------
TABLE 8-9
DIATOMACEOUS EARTH FILTRATION OF SECONDARY EFFLUENT
Turbidity
Flow Rate Body Feed In Out Run Length Type
gpm/sq ft mg/1 JTtJ JTU hr
0.53 42 5.5 0.8 19.5 Vacuum
0.75 33 5.2 0.8 10.7 Vacuum
1.0 19 4.4 0.4 5.4 Vacuum
0.50 50 8.2 3.1 50.0 Pressure
0.81 42 8.3 3.9 28.4 Pressure
1.0 45 7.5 3.0 31.0 Pressure
requirements (27). Acceptable operation was possibly only with very low influent solids (3
to 13 mg/1). None of these studies considered the possibility of recovering DE filter aid,
which could reduce the estimated costs significantly (28).
8.6 Ultrafiltration
8.6.1 General
Ultrafiltration (UF) is the title given to a form of membrane separation which employs
relatively coarse membrane separation at relatively low pressures. The process should be
differentiated from reverse osmosis which is a similar process used for dissolved solids
separation using fine membranes and high pressures. Ultrafiltration, using a thin
semi-permeable polymeric membrane is reported most successful in separating suspended
solids as well as large-molecule colloidal solids (0.002 to 10.0p.) from wastewater (29).
Fluid transport and solids retention are achieved by regulating pore size openings. Thus, the
Ultrafiltration process is a physical screening through molecular-sized openings rather than
one controlled by molecular diffusion.
A system employing high-MLSS aeration followed by UF operated at Pikes Peak since
1970 has proven capable of removing virtually 100 percent of the suspended matter and 93
to 100 percent of the associated BOD, COD and TOC from aerated mixed liquors (30).
8.6.2 Application
Several installations (29) (30) (31) have proven the ability of the activated
sludge-ultrafiltration process to remove all SS and almost all bacteria and BOD. These
systems are typified by Figure 8-15. Results of such installations are given in Table 8-10.
Because UF installations have all produced zero SS effluents, other parameters are given to
illustrate process capability. More detailed data on the Pikes Peak installation is given in
8-36
-------
POTABLE WATER SUPPLY
POTABLE
USE
oo
i
t^j
-j
DISCHARGE
DISINFECTION-
NON-
POTABLE
USE
ULTRA-
FILTRATION
CELL
MEMBRANE
HOLDING
TANK
GRINDER
HIGH-SOLIDS
ACTIVATED SLUDGE
PRESSURIZED
REACTOR
(AERATOR)
en
H
CO
IT"
FIGURE 8-15
SCHEMATIC FLOW DIAGRAM OF THE PIKES PEAK TREATMENT & REUSE SYSTEM
-------
Table 8-11. Coliform counts in all instances were above zero but were attributed to outside
contamination rather than passage through the membranes (30) (31).
TABLE 8-10
RESULTS OF ULTRAFILTRATION INSTALLATIONS
Effluent Superficial
Capacity BOD Coliform Flux Pressure Flow Rate
gpd mg/1 No./100 m sq ft final psi. > fps
Fabric Fire
Hose, 3,600 2-15 1-100* 18 8 22-27 4-6
Sandy Hook (av. 5)
Pikes Peak 21,000 1 0-11** 30 6 50 5-6
16 9
* includes deliberate upset tests
**attributable to external contamination
The major drawback of-ultrafiltration is the high capital and operating costs. Phosphate
and color removal are both negligible, but they may not be necessary in many places. The
high cost may be offset by compactness where space is a critical factor, such as on a ship or
a mountain top. A 6800 gpd shipboard installation was designed to occupy a volume of 6 x
8x9ft(31).
8.6.3 Design
The most important design considerations are:
1. Membrane area
2. Membrane configuration
3. Membrane material
4. Membrane life
5. Driving force.
8-38
-------
TABLE 8-11
SUMMARY OF PIKES PEAK DATA
oo
I
OJ
Parameter
BOD (mg/1)
COD (mg/1)
TOC (mg/1)
TSS (mg/1)
FLUX (gpd/sq ft)
MLSS (mg/1)
Fecal Coliform
(per 100ml)
CSSTP '70 Summit '70
72 days 22 days
Inf. Eff. Inf. Eff.
382 1 285 1
678 20 547 32
192 7.5 136 6.6
323 0 129 0
30 to 6
3954
0 0
Summit '71
49 days
Inf. Eff.
362 1.3
738 52
197 9.8
172 0
16 to 9
4156
6
Weighted
Summit '72 Average
83 days 226 days Percent
Inf. Eff. Inf. Eff. Removal
426 6.2 384 3 99.2
839 536 737 40.5 94.6
185 8.1 95.6
263 0 249 0 100
—
11 9
-------
8.6.3.1 Area
Membrane area is a function of flux which is determined by membrane construction and the
fouling characteristics of the wastewater. A flux of 8 gpd/sq ft has proven satisfactory at
the Pikes Peak installation and can be used as a normal design figure in calculating
necessary area. Membrane flux tends to decrease with time due to surface fouling. It has
been found that physical elimination of foulants, mostly organic acids and polarized
materials, lessens their flux-reducing effect (29). By operating the process at liquid velocities
of 3 to 10 fps parallel to the membrane surface, scouring of contaminants can be
accomplished and a more stable flux achieved (29). At such velocities, with normal
membrane fluxes, single-pass design would require impractically large membrane area.
Therefore, the wastewater is recirculated as shown in Figure 8-15. Some blowdown of
concentrated waste results to prevent excessive solids buildup. The blowdown can be
intermittent, at a rate sufficient to keep the MLSS within acceptable ranges, usually 4,000
to 15,000 mg/ 1 (29) (32).
8.6.3.2 Membrane Configuration
This aspect of design concerns the amount of membrane surface area which can be
incorporated into a module. Because of low membrane fluxes it is imperative to design the
module to maximize membrane surface area. One configuration adopted solely for
ultrafiltration is the storage battery configuration, as shown in Figure 8-16. The membrane
is cast on both external faces of a hollow plate. A number of these plates are arranged in a
parallel array. The edges of these plates face the incoming stream of solids and act as a
coarse screen which can be backwashed by reversing the direction of the approaching flow.
Other designs include tubular support elements over which a membrane is wound helically
or in which the membrane is enclosed in a continuous spiral.
8.6.3.3 Material
The membrane itself is made up of two basic layers:
1. Surface—an extremely thin homogenous polymer of 0.1 to 10.00 microns
(typical, 5.0 microns).
2. Surface support—an open cell of 5 to 10 mil thickness
The membrane, in turn, is supported on a porous sheet (paper) for added mechanical
support.
The thin surface layer controls the transport and rejection properties of the membrane.
Numerous means and types are available and can be tailored to the particular application.
Typical membrane specification ranges are listed in Table 8-12.
8-40
-------
Outlet
oo
Recirculating
flow inlet
Ultrafiltrate
Ultrafiltration
cartridge
ili
Cover
FIGURE %-\6
"STORAGE BATTERY" MEMBRANE MODULES
(Courtesy of Dorr-Oliver).
-------
TABLE 8-12
TYPICAL MEMBRANE SPECIFICATIONS
Material Most organic polymers
Water Permeability 7-290 gpd per sq ft at 30 psi
Molecular weight 340-45,000
Retentivity 60-100 percent
Maximum Operating
Temperature 50-120°C
Water permeability is used to characterize the porosity of the membrane, but does not
represent the stabilized, long-term flux on a process fluid. In the waste treatment field,
fluxes of 7 to 10 gallons per square foot of membrane surface per day are typical (31).
Given the current state of membrane manufacturing technology, almost any set of clean
water performance characteristics, without consideration for fouling can be produced. A
few of the leading manufacturers of ultrafiltration membranes are Romicon Corp., Abcor,
Inc. and Dorr-Oliver, Inc. Catalogues offering a wide variety of membranes are available.
8.6.3.4 Membrane Life
Membrane life is a function of fouling and required flux rates. A membrane may be
considered acceptable for a life span of 6 months in continuous operation with an initial flux
of 18 and a final flux of 8 gpd/sq ft. A plant must be designed for the lower figure and
membrane replacement made when the design figure is reached.
8.6.3.5 Driving Force
The driving force for transport of water through the membrane is pressure. Operation is
achieved at pressure gradients of approximately 25 psi. Total system pressures do not
exceed 50 psi. Very recent work has shown that vacuum extraction of the product can be
used advantageously in certain applications (31).
8-42
-------
8.7 References
1. Ginaven, M.E., The Hydrasieve—A New Simplified Solids-Liquid Separator; Paper
Trade Journal (January 19, 1970).
2. Ginaven, M.E., A New Low Cost Device for Solids Recovery from Effluent; Cost En-
gineering; 16,4, pp 4-10 (Oct., 1971).
3. Wittenmyer, James D., A Look at the Future Now, Presented at the Ohio Water Pol-
lution Control Conference (June 20, 1961).
4. Wittenmyer J.D., Operating Experience, Ohio Water Pollution Conference (June 16,
1972).
5. Wong, Alan, Personal Communication, Hazen and Sawyer.
6. Bishop, D.F. Jr., U.S. EPA Internal Monthly Report, May, 1972, Contract No.
68-01-0162.
7. O'Farrell, Tom, Personal Communication, Sanitary Engineer, U.S. EPA Blue Plains
Pilot Study.
8. Lee, R., Personal Communication, Hercules, Inc., AWT Division.
9. Hydrocyclonics Corp., Rotostrainer Bulletin.
10. Diaper, E.W.J., Personal Communication, Crane Co. (April, 1973).
11. Diaper, E.W.J., Tertiary Treatment by Micro straining. Water and Sewage Works,
776, pg. 202 (June 1969).
12. Diaper, E.W.J., Oxidation Pond Effluent Improvement, 49th Texas Water and Sew-
age Works Associations's Short School, College Station, Texas (March 1967).
13. Mixon, P.O., Filterability Index and Microscreener Design, Jour. WPCF, 42, pg. 1944
(Nov. 1970).
14. Boucher, P.L., A New Measure of the Filterability of Fluids with Application to Wa-
ter Engineering, ICE Jour. (Brit.), 24, 4, pg. 41'5 (1947).
15. Bodien, D.G., and Stenburg, R.L., Microscreening Effectively Polishes Activated
Sludge Effluent, Water and Wastes Engineering, 3, pg. 74 (Sept. 1966).
16. Lynam, B., Ettelt, G., and McAloon, F., Tertiary Treatment at Metro Chicago by
Means of Rapid Sand Filtration and Microstrainers, Jour., WPCF, 41, pg. 247 (Feb.
1969).
17. Havens and Emerson Ltd. Consulting Engineers, Report on Wastewater Treatment Pi-
lot Plant for the City of Euclid, Ohio (April 30, 1971).
8-43
-------
18. Sapper, John, Personal Communication, Zurn Industries (April, 1973).
19. Isaac, C.G. and Hibberd, R.L. The Use of Microslrainers and Sand Filters for Ter-
tiary Treatment. Water Research, (Britain) 6, pg. 465, (1972).
20. Trusdale, G.A., and Birkbeck, A.E., Tertiary Treatment of Activated—Sludge Ef-
fluent, Metropolitan and Southern Branch International Filtration Exhibition and
Conference, The Institute of Water Pollution Control (Brit.), 67, pg. 483 (1968).
21. Microscreening Plant for Effluent Polishing, Effluent and Water Treatment Jour., Per-
mutit, (July 1971).
22. Saucer, Victor, Concentrator and Floatation Cell Increase Wastewater Plant Capacity
Fourfold, American City 88, 4, pg. 49, (April, 1973).
23. Falby, W.J., Personal Communication, SWECO Inc., Los Angeles, Calif. (May,
1973).
24. Bell, G.R., Design Criteria for Diatomite Filters, Jour. AWWA, 54, pg. 1241 (Oct.
1962).
25. Wells, W.N. and Davis, D.W., Filtration of Activated Sludge Plant Effluent. Public
Works, 98, 4, pg. 94 (Apr. 1967).
26. Summary Report—Advanced Waste Treatment Research Program (1964-1967),
U.S.D.I., FWPCA Publication WP-20-AWTR-19 (1968).
27. Guiver, K., and Huntington,R., A Scheme for Providing Industrial Water Supplies by
the Re-use of Sewage Effluent. Water Pollution Control (G.B.), 70, pg. 74 (1971).
28. Filter and Recovery Systems Bulletin FF 206. Celite Technical Data, Celite Div.,
Johns-Manville Corp. (1971).
29. Smith, C.V. Jr., Di Gregario, D. and Talcott, R.M., The Use of Ultrafiltration Mem-
branes for Activated Sludge Separation, Purdue Ind. Waste Conf. 5/7/69. Water Re-
novation of Municipal Effluents by Reverse Osmosis, Water Quality Office, U. S. EPA.
Project EPA 17040 EOR (Feb. 1972).
30. Bemberis, Ivars, Personal Communication, MST Marketing, Dorr-Oliver Inc.
(March, 1973).
31. Bemberis, Hubbard and Leonard, Membrane Sewage Treatment Systems. American
Soc. of Agric. Engrs., 1971 Winter Meeting.
32. Schwartz, Warren A., Record Memorandum, Office of Program Co-ordination, U.S.
EPA, (July 6, 1972).
8-44
-------
CHAPTER 9
GRANULAR MEDIA FILTRATION
9.1 Introduction
9.1.1 Applications and General Description
This process, long applied in treatment of municipal and industrial water supplies, is
becoming widely used for waste water treatment both in upgrading existing conventional
plants and in designs of new advanced treatment facilities. Next to gravity sedimentation it
is the most widely used process for separation of wastewater solids. The following specific
applications have been noted (1):
1. Removal of residual biological floe in settled effluents from secondary treatment
by trickling filters or activated sludge processes.
2. Removal of residual chemical-biological floe after alum, iron, or lime precipitation
of phosphates in secondary settling tanks of biological treatment processes.
3. Removal of solids remaining after the chemical coagulation of wastewaters in
tertiary or independent physical-chemical waste treatment.
In these applications filtration may serve both as an intermediate'process to prepare
wastewater for further treatment (such as carbon adsorption, cli-noptilolite ammonia
exchange columns, or reverse osmosis) and as a final polishing step following other
processes.
Granular media, filtration involves passage of water through a bed of granular material with
resulting deposition of solids. Eventually the pressure drop across the bed becomes
excessive or the ability of the bed to remove suspended solids is impaired. Cleaning is then
necessary to restore operating head and effluent quality to acceptable levels. Most filters
operate on a batch basis, the entire unit being removed from service for periodic cleaning.
The time in service between cleanings is termed the run length. The head loss at which
filtration is interrupted for cleaning is called the terminal head loss.
9.1.2 General Design Considerations
Filter design involves selection of the following filter characteristics (1):
1. Filter configuration
2. Media sizes and depths and materials
3. Filtration rate (gpm/sq ft)
9-1
-------
4. Terminal head loss (ft of water)
5. Method of flow control
6. Backwashing design features
The major goal in design is to achieve effluent quality objectives at low capital and
operating costs. The most important characteristic in determining capital costs is the
filtration rate, which fixes the filter size. Operating costs are affected primarily by
filtration rate, terminal head loss, media characteristics and backwash design. The first
three filter characteristics determine the cost of power for operating head and the
production of the filter per run. The backwash design determines the cost per cleaning of
operator attention, washwater pumping, air scouring (compressor operation) and treatment
of dirty washwater. The cost of cleaning per unit volume treated (cost per cleaning divided
by production per run) depends on all four factors.
Section 9.3 discusses the interrelation and effects on performance of process variables
describing the characteristics of filters and of influent wastewaters. Many investigators have
attempted to relate filter performance quantitatively to these variables (2) (3) (4) (5) (6).
Unfortunately, such relations are of little help in predicting performance except under
specific conditions already explored in pilot work. In part, this is due to the wide variations
in the filtering characteristics of wastewater. solids and to the dearth of reproducible
objective data from well-conceived studies of wastewater filtration. In part, however, it may
be inherent in the nature of the filtration process that any fully general quantitative
relations describing it would be too complex for practical use. Nevertheless existing
theoretical relations are useful in providing: 1) general insight into filter behavior, 2)
frameworks for analysis of data from pilot investigations and 3) bases for comparing cost
effects of alternatives in specific applications.
9.1.3 Basis for Design
Wherever possible, designs should be based on pilot filtration studies of the actual waste
(Section 9.9). Such studies are the only way to assure:
1. Meaningful cost comparisons between different filter designs capable of equivalent
performance (7), i.e., producing the same output quantity and quality over the
same time period.
2. Most economical selection of filter rate, terminal head loss and run length for a
given media application.
3. Definite effluent quality performance for a given media application.
Pilot studies are also useful for determining effects of pretreatment variations or for
characterizing Jilterability in terms of performance attainable with a specific filter design.
Where there is no opportunity for pilot studies, parameters for workable designs can still be
9-2
-------
determined from the discussions of wastewater and filter characteristics in the sections
below. The parameters will necessarily be conservative and will tend to give more costly
designs and less assurance of effluent quality than parameters based on testing. Facilities
designed without pilot testing are likely to be small ones, for which the design should
provide long filter runs and minimize required operating attention.
Another approach to obtaining economical facilities is to prepare a functional specification
which will permit competitive bidding between suppliers of alternative filter systems.
Functional requirements should include:
1. Guarantees of specified performance; both capacity and effluent quality.
2. Guarantees of proposed values of all factors which affect operating costs such as
head or power requirements and backwash volume to be recycled.
Bids should be evaluated based on total present worth including operating costs, which
should be calculated by a predetermined formula using factors in the guarantee.
This approach will work best when bidders are supplied test results characterizing the
filterability of the waste flow. In any case, bidders should be given full information on the
wastewater and treatment to be provided ahead of filtration plus the opportunity of testing
effluents from such treatment, where already in operation.
9.2 Process Alternatives
Filter units generally consist of a containing vessel, the granular media, structures to
support or retain the media, distribution and collection devices for influent, effluent and
washwater flows, supplemental cleaning devices, and necessary controls for flows, water
levels or pressures.
Some of the more significant alternatives in filter layout are discussed below.
9.2.1 Alternative Flow Directions
Most filter designs employ a static bed with vertical flow either downward or upward
through the bed. The downflow designs traditionally used in potable water treatment
(Figures 9-1 and 9-2 (a) (d) and (e)) are most common, but recently a number of
installations have been designed for upward flow (Figures 9-3 and 9-2 (b)). The Eurdpean
biflow design (Figure 9-2 (c)) employs both flow directions with the effluent withdrawn
from the interior of the bed. Upflow washing is used regardless of the operating flow
direction. Two special filter designs employ horizontal radial flow through an annular bed.
Media is cycled downward through the bed, withdrawn at the bottom, externally washed,
and returned to the top. (See Section 9. 10).
9-3
-------
Operating-.
table
Rate of flow and loss
of head gages
vo
Operating
floor
Pipe gallery
floor
Filter drain
Filter to waste
Perforated
laterals
Cast-iron/
manifold
-Filter
floor
Filter bed wash-
'water troughs
Influent to filters
Concrete filter .
tank
Pressure lines to
hydraulic valves from
operating tables
Effluent to
clear well
Drain
FIGURE 9-1
TYPICAL RAPID SAND FILTER
-------
EFFLUENT
INFLUENT
30-40'i-
UNDERDRAIN —
CHAMBER
OVERFLOW TROUGH
GRID TO
RETAIN
SAND
EFFLUENT
EFFLUENT
INFLUENT
STRAINER
UNDERDRAIN
CHAMBER
INFLUENT
(a), CONVENTIONAL FILTER (b), UPFLOW FILTER
SINGLE MEDIA FILTERS
(c), BI-FLOW FILTER
30-40"
ANTHRAFILT
(d), DUAL MEDIA FILTERS
—COARSE MEDIA—
— INTERMIX<
ZONE \
•—FINER MEDIA ^
FINEST MEDIA
CHAMBER
...fc». •• ..• -.
ANTHRAFILT
\
28-48"
GARNET SAND
(e),MIXED-MEDIA FILTERS
(TRIPLE MEDIA)
FIGURE 9-2
FILTER CONFIGURATIONS
9-5
-------
COVER OPTIONAL
(FOR CLOSED SYSTEM)
"GRID"
DEEP SAND LAYER
GRAVEL LAYERS
1
INLET RAW WATER
WASH WATER
FILTRATE OUTLET
SAND "ARCHES"
SPECIAL VENT
AIR FOR
SANDFLUSH CLEANING
FIGURE 9-3
CROSS SECTION OF UPFLOW FILTER
9-6
-------
9.2.2 Gravity vs. Pressure Filtration
Filters may be designed with closed vessels permitting influent pressures above atmospheric
(Figure 9-4) or with open vessels where only the hydrostatic pressure over the bed is
available to overcome filter headlosses (Figure 9-1). Pressure units are generally preferable
where high terminal headlosses are expected or where the additional head will permit flow
to pass through downstream units without repumping (1) (8). They are most commonly
used in small-to-medium-sized treatment plants where steel-shell package units are
economical (8).
9.2.3 Media Alternatives
Figure 9-2 shows schematically a number of different filter configurations using fixed bed
media. The beds shown are all graded during upflow washing so that the finer material of a
given specific gravity is on top. It should be noted that the conventional single media filter
used in potable water treatment (Figure 9-2(a)) is generally unsatisfactory for wastewater
treatment because the wastewater solids cause a high headloss buildup at the fine surface
layer.
In upflow designs, flow passes first through the coarser media which for a given head loss
buildup has greater capacity for retaining filtered solids. This is advantageous in
lengthening filter runs and increasing output. Dual and multi-media (Figure 9-2 (d) and (e))
obtain the same effect under downflow operation by placing coarser layers of lighter
material over finer denser material. An alternative downflow single-media configuration
not shown attempts to get the same advantage from use of beds of uniform-sized coarse
media with depths of 60 in. or more. The effects of significant media characteristics such as
size gradation, specific gravity and depth, on filter performance are discussed in detail in
Section 9.3.
In filters using external wash, the media is not vertically graded; particle size distribution
tends to be the same throughout the bed.
9.2.4 Batch vs. Continuous Operation
It is normal practice to design filters to operate on a batch basis with entire units taken out
of service for cleaning according to schedule or as required. Several special designs,
however, provide more or less continuous cleaning, either externally with media cycled
through the bed, or in-place with techniques such as traveling backwash or air pulsing of the
bed and air mixing of the liquid above it.
9.2.5 Normal Design
Most of this chapter will relate to fixed bed systems with intermittent upflow washing. Both
upflow and downflow designs will be included as normal design. Proprietary designs using
non-fixed beds or special washing systems are discussed in Section 9.10.
9-7
-------
SIG PRESSURE
VESSEL-^
A
1 COUPLING
AIR RELEASED
L/IC
1
V
1 \
MEDIA
,
al p Q o o o o
V
9 9 9 9' 9 9 99 909
B 9/
-4
I2"x 16" MANHOLE
ON VERTICAL
OF TANK
FLANGE INFLUENT
BACKWASH WASTE
2 FLANGE
SURFACE WASH
JO"FLANGE EFFLUENT
AND BACKWASH
\
X2" FILTER DRAIN
FILTER SUPPORTS
AT 1/4 POINTS
ELEVATION
DISTRIBUTOR-
s
8'-O" O.D.
I2"x 16" MANHOLE
SURFACEWASH
MIXED MEDIA
LATERALS
S ECTI 0 N
FIGURE 9-4
TYPICAL PRESSURE FILTER
(Courtesy of Neptune Microfloc, Inc.)
SUPPORT GRAVEL
CONCRETE
9-8
-------
9.3 Process Variables
9.3.1 Performance Relations
The measures of filter performance are output quality and quantity. The variables which
determine or limit performance fall into two major groups: influent characteristics and the
physical characteristics of the filter. The latter include media characteristics, filtration rate,
available and applied operating head and the design and operating parameters of the filter
cleaning system.
In determining the fundamental limits on quality performance.the characteristics of prime
importance are those of the influent solids: concentration, strength, size, and the
physical-chemical properties governing adhesion of particles to each other or to the media
surfaces. Commonly a number of filters with different physical characteristics can come
close to the limiting quality performance for a given influent. In contrast, quality
performance of given filters can vary widely for different solids characteristics.
In determining output quantity from filters, the influent solids characteristics—especially
floe strength and solids concentration—are again very inportant, but the physical
characteristics of the filter become significant too.
At run lengths of 24 nr or more, output depends almost totally on filter rate. As run lengths
become shorter, however, the effects of downtime and washwater recycle during cleaning
become increasingly important (See Section 9.4). The washwater recycle volume depends
on the backwash flow rates and the wash cycle duration needed for adequate cleaning (see
Section 9.7). Factors governing backwash system design include:
1. Size distribution, depth and specific gravity of media
2. Nature of solids removed, principally their adhesion to the media and their
tendency to compact in a dense layer at the media surface
3. Type of supplementary cleaning provided.
Run length may be limited either by available head or by deterioration of effluent quality as
the filter bed becomes filled with solids (breakthrough). Which factor governs depends on
the interaction of several variables including:
1. Influent solids characteristics (all those which affect quality performance)
2. Flow rate
3. Temperature and viscosity of the wastewater
4. Media characteristics
5. The amount of head available.
9-9
-------
Headless in a clean bed varies directly with filter rate and inversely with grain size. In
determining head loss buildup, the most significant media characteristic is the grain or
pore size at the influent surface of the bed (or in some cases within finer denser layers of
multi-media filters). In downflow filtration through a graded bed, influent solids particles
larger than about 7 percent of the minimum grain size (9) will be removed by straining
provided their strength is sufficient to withstand the shear at the surface. Shear varies with
filter rate and liquid viscosity.
In surface straining, head loss increases exponentially with time or solids accumulation (See
Chapter 8). Where significant solids loads are removed predominantly by surface straining,
head loss buildup will be rapid, filter runs short and backwash frequency high. In addition
the solids removed at the surface tend to be compressed into a dense mat which is difficult
to remove in backwashing.
Removal of solids within the bed rather than just at the surface is termed depth filtration.
Both surface and depth filtration are usually involved to some degree in any given
application.
In depth filtration head loss tends to build up linearly with time or with solids
accumulation. Compression of the solids removed is limited by the granular structure of the
bed. For downflow filtration within a single media, the farther solids penetrate into the bed,
the slower will be the rate of head loss buildup, but the sooner solids will breakthrough into
the effluent.
The factors which determine breakthrough for a single media are the media size and depth,
the flow rate and the resistance of deposited materials to shear within the bed. Hudson (10)
suggested characterizing the resistance of solids to breakthrough by an index, K, calculated
from the physical characteristics of the filter and the head loss at which breakthrough
occurs. The expression for the index is:
K = Vd3H/L
Where:
V = filtration rate—gpm/sq ft
H = head loss at breakthrough—ft
d = effective size of media—mm
i L = bed depth—ft
9.3.2 Influent Characteristics
The influent characteristics of prime importance in determining filter performance are
those of the solids to be removed. The only significant characteristic of the wastewater
liquid—as opposed to the solids—is viscosity which varies with temperature. Its effects on
development of filter head loss are generally small, however, in comparison to the effects of
solids accumulations or filter rates.
9-10
-------
The characteristics of wastewater solids which govern or limit filter performance are
determined by the treatment processes ahead of filtration (see Section 9.1). In direct
filtration of secondary biological effluent the residual solids applied to the filter are
predominantly biological floe grown in the treatment process. In filtration of effluent fol-
lowing tertiary coagulation for phosphate removal the residual solids are predominantly
chemical floes. In filtration of chemically precipitated raw wastewater or primary effluent,
the solids consist of inorganic chemical floe with varying quantities of precipitated organics.
Loading, media and performance data for filter applications of the above three types are
shown in Tables 9-1, 9-2 and 9-3. Most data are for full scale installations but a few large
pilot facilities are included. The data are those ordinarily recorded in tests of filter
installations. These data show only that the systems filtering physical-chemical floe tend to
use somewhat lower filter rates and somewhat finer media (with dual or multi-media
configurations almost standard) and that effluent results for any given type of influent
source may vary considerably.
Compiling data from a number of filter installations treating biological effluents, Kreissl
(33) found removals ranging from 50 to 90 percent with a mean of about 70 percent,
provided influent solids were below 35 mg/ 1. Included were data for a variety of loading
rates media configurations and types of prior biological treatment. Subsequent compilation
of similar data for effluents from chemical treatment systems showed mean solids removals
of only 60 percent (34), indicating that on the average chemical floe tends to be more
difficult to filter.
The only influent solids characteristic included in routine filter testing is the concentration,
perhaps because it is the only one that is easily measured. A few special studies have
attempted to take into account other characteristics such as floe strength, particle size
distribution (concentration vs. size) and properties governing adhesion of particles to each
other or to the media. Some other studies have tried to distinguish differences in filter
performance according to parameters of the treatment prior to filtration. Outlined below
are a few significant additional insights into wastewater filtration provided by these special
studies.
9.3.2.1 Floe Strength
Biological floes tend to be significantly stronger or more resistant to shear than chemical
floes, at least those from alum or iron coagulants (2). Consequently, in filtering biological
floes, surface straining is generally significant and runs are almost always terminated by
excessive head loss. Breakthrough is rarely observed. In one study head losses as high as 30
ft were applied without deterioration of effluent quality (35). This contrasts with alum and
iron (hydroxide) floes which have been shown to penetrate readily into filters and to
breakthrough at relatively low heads ranging from 3 to 6 ft (2) (10) (36). In an isolated
instance where breakthrough of biological floe was observed, the index K was found to be
13.7, far above the range of 0.3 to 3.6 cited for alum or iron floe in water treatment (37). In
contrast to floes from other common coagulants, calcium carbonate precipitates are
strongly removed at the filter surface where they may form a dense compressed layer hard
9-11
-------
TABLE 9-1
RESULTS OF STUDIES OF
FILTRATION OF EFFLUENT FROM SECONDARY BIOLOGICAL TREATMENT
LOCATION
Luton, G.B.
•Hanover Park,
Illinois
Walled Lake-Novi,
Michigan
Louisville, Ky.
(Hite Creek)
Coldwater,
Michigan
i i r L. wr
FILTER
Gravity
Downflow
Gravity
Downflow
Gravity
Downflow
Pressure
Upflow
Pressure
Downflow
Gravity
Downflow
Pressure
Horizontal
Pressure
Downflow
i n r LI u u 11 i
SOURCE
Activated
Sludge
Trickling
Filter
Activated
Sludge
Activated
Sludge
Activated
Sludge
Activated
Sludge
Activated
Sludge
Trickling
Filter
MEDIA
type
Sand
Sand
Coal
Sand
Garnet
Sand
Coal
Sand
Mixed
media
Coal
Sand
Garnet
Coal
Garnet
Garnet
1
1
1
0
0
1
0
0
0
0
1
SIZE
mm
.2-1.3
.8- .9
.4- .8
-2
.4-1.8
.8-1 .0
.25-2.0
.0 -1.2
.45-0.55
.2-0.3
.8
.4-0.6
.2
DEPTH
in
36
36
30-]
12 r
6J
60
2 4~1
12J
30
16. Si
9 1
4.5J
20~]
20 f»
9 J
11 i ur\n\j LI j. i*
LOADING
gpm/ft2
1.6-4.0
3.4
5.0
2.0
4.0
2.0
4.0
5.0
2
4
6
8
10
3-4
3.4
4.9
IN
mg/1
25-50
28-35
13
14
16
14
15
13
16
15
16
13
18
7
27
21
OUT
rag/1
3-6
9-10
8
4
4
7
6
6
7
5
6
6
8
3
3
8
REMOVAL
percent
72-91
67-74
40
57
67
SO
67
54
56
67
62
54
55
57
89
62
LENGTH
hr
12
_
-
106
27
150
17
7
90
15
22
31
12
-
2.5-8
REFERENCE
(11)
(12)
(13)
(13)
(14)
(15)
(16)
(15)
9-12
-------
TABLE 9-1 (CONTINUED)
RESULTS OF STUDIES OF
FILTRATION OF EFFLUENT FROM SECONDARY BIOLOGICAL TREATMENT
LOCATION '
Bedford Township,
Michigan
Ventura, Cali-
fornia
West Hertfordshire,
G.B.
Ann Arbor,
Michigan
State College, Pa,
Springfield, Ohio
1 I ft UJ-
FILTER
Pressure
Downf low
Gravity
Deep Bed
Downf low
Immedium
Up flow
Pressure
Downf low
Pressure
Downf low
Gravity
Downf low
INf LUtN 1
SOURCE
Activated
Sludge
Trickling
Filter
Activated
Sludge
Activated
Sludge
Activated
Sludge
Contact
Stabi-
lization
MEDIA SIZE
type mm
Multi-
media
Sand 1-2
Gravel
Sand 1-2
Mu 1 1 i -
media
Sand
Sand 0.45
DEPTH
in
.
-
261
60J
-
84
10
HI LIKMULil,
LOADING
gpm/ft2
.
6
2.2
4.0
5.0
6.0
6
3-12
5.3
IN
mg/1
15
18
44
37
55
37
42
6
14
OUT
mg/1
3
7
2
4
7
10
5
1
5
REMOVAL LENGTH
percent hr
80 IS
61 6-18
95
90
87
73
88
85 6
64
(IS)
(17)
(18)
(19)
(20)
(21)
9-13
-------
TABLE 9-1 (CONTINUED)
RESULTS OF STUDIES OF
FILTRATION OF EFFLUENTS FROM SECONDARY BIOLOGICAL TREATMENT
LOCATION
Letchworth, England
Upper Stour Main
Drainage, Freehold
IVorks ,' England
Harpenden, U.D.C.
Rodbourne Works,
Swindon, England
Derby, England
Tharaeside, England
Thameside, England
Thameside, England
Ashton-Under-Lyne,
England
i i r c ur
FILTER
Pressure
Upflow
Gravity
Downf low
Gravity
Downf low
Gravity
Downf low
Simater
Radial Flow
Iramedium
Pressure
Upflow
Permutit
Upflow
Simater
Radial Flow
Iramedium
Upflow
i t\ r L U C N 1
SOURCE MEDIA SIZE
type mm
Activated Sand 1-2
Sludge
Activated Sand 0.5-2.5
Sludge
Trickling Sand 1 . 1
Filter
Trickling Sand 1.5-3
Filter
Trickling Sand 1-2
Filter
Activated Sand 1-3
Sludge
Activated Sand 0.60-1.20
Sludge
Activated Sand 0.5-1
Sludge
Trickling Sand 1-2
Filter
— — — — — niUKHULlL
DEPTH LOADING
in gpra/ft2
60 5.3
1.2-2.4
1-3
1.6-3.2
4-6
63 3.3
3.3
5.0
5.0
57 3.3
3.3
5.0
5.0
3.3
3.3
5.0
5.0
60 4.5-5.0
IN
mg/1
17
12
20
21
22
9
46
8
37
9
32
11
28
11
51
11
24
30
OUT
mg/1
7
5
5
5
9
2
8
6
10
1
7
4
5
3
7
4
10
8
nun
REMOVAL LENGTH
percent hr
60
58
75
75
60
74
84
20
74
86
78
60
83
74
86
62
58
80
REFERENCE
(22)
(22)
(22)
(22)
(22)
(23)
(23)
(23)
(24)
9-14
-------
TABLE 9-2
RESULTS OF STUDIES OF
FILTRATION OF CHEMICALLY TREATED SECONDARY EFFLUENT
LOCATION
Piscataway, Md .
Ely, Minnesota
Jefferson Parish,
La.
Nassau County,
N.Y.
Lake Tahoe,
California
lYFt Uh
FILTER
Pressure
Downf low
Gravity
Downf low
Upflow
Gravity
Oownf low
Two -Stage
Pressure
INhLUhNl
SOURCE
A.S . +2-Stage
Lime Cla- •
rif icat ion
High Rate
+2-Stage
Lime Cla-
rifica-
tion
T.F. tin-
Line Alum
Inj ection
A.S .+Alum
Clarifica-
tion
A.S.+Lime
Clarifica-
tion-*- Ammonia
MEDIA
type
Coal
Sand
Coal
Sand
Sand
Coal
Sand
Coal
Sand
Garnet
SIZE DEPTH
mm in
1.0 1 2~L
0.5 6 J
24~1
1?J
-
0.9min. 36~1
0.35 nin. 12_J
18T
12 f
6 J
LOADING IN OUT REMOVAL LENGTH 1 REFERENCE
gpm/ft2 mg/1 mg/1 percent hr
3 12 8 33 50 (263
2.3 8 <2 >75 24 (27)
3 40 21 48 2.5-6.5 (28)
2.5-3.5 2-10 0-2 80-90 16-48 (29)
2.8-4.0 9-15 0-1 93 4-60 (30)
Stripping
and Recar-
bonation
9-15
-------
TABLE 9-3
RESULTS OF STUDIES OF
FILTRATION FOLLOWING CHEMICAL TREATMENT OF PRIMARY OR RAW WASTEWATER
LOCATION
Washington, D.C.
Lebanon, Ohio
Washington, D.C.
Washington, D.C.
lire ur
FILTER
Gravity
Downf low
Gravity
Downf low
Gravity
Downf low
Gravity
Downf low
inrLucn i
SOURCE
Two-Stage
Lime Cla-
rification
Single
Stage
Lime Cla-
rification
Two-Stage
Lime Cla-
rification
Single
Stage
MEDIA
type
Coal
Sand
Coal
Sand
Coal
Sand
Coal
Sand
SIZE
mm
0.9
0.45
.75
.46
1.2-1.4
0.6-0.7
1.2-1.4
0.6-0.7
— — — — n i u R/\ u L i t,
DEPTH LOADING IN
in gpm/ft2 mg/1
IB"! 1.7-6.3' 14
6J
IS"^ 2.0 30
6_
24~1 2.4-4.4 139
24~l 2.3-4.3 123
I'/
rvuii
OUT REMOVAL LENGTH REFERENCE
ing/1 percent hr
6 70 12-50 (32)
10 67 - (31)
33 74 ^.24 (25)
23 81 >24 (25)
Lime Cla-
rification
-------
to remove during washing (36). Comparative data are lacking on the strength of floes from
precipitation of phosphates in wastewater using alum, iron or lime. It is reasonable to
assume, however, that they are similar to aluminum or ferric hydroxide floe.
Polymer filter aids may be added to the filter influent to strengthen weak chemical floes
thereby permitting operation at higher rates without breakthrough. Doses of 0.1 mg/ 1 or
less are often adequate (8). Polymers added as coagulant aids in upstream settling or
flocculating units may similarly strengthen the residual floe applied to the filters. Ample
head loss must be available to meet losses due to the tougher floe, and doses must be kept as
low as possible to avoid excessive head loss.
9.3.2.2 Particle Size
Floe particle sizes in settled biological effluent tend to be bimodally distributed. Mean sizes
for the two modes in one study were 3 to 5 microns and 80 to 90 microns (2). About half of
the weight was in each mode. Theoretical work (38) suggests that particles in the lower size
range are much less effectively removed by filtration than those in the higher range. Hence
for the best quality performance from filtration, the proportion of smaller size particles
must be reduced to a minimum by proper flocculation.
9.3.2.3 Filterability
The filterability of residual solids from secondary settling varies with solids retention time
and with liquid contact time in the biological process. For biological systems with higher
solids retention times and longer liquid contact times, filtered effluents tend to have lower
suspended solids. Gulp and Gulp (8) indicated the expected performance of multi-media
filters for plain filtration in secondary effluents as shown in Table 9-4.
TABLE 9-4
EXPECTED EFFLUENT SUSPENDED SOLIDS FROM MULTI-MEDIA
FILTRATION OF SECONDARY EFFLUENT
Effluent Type Effluent SS
mg/1
High Rate Trickling Filter 10-20
2-Stage Trickling Filter 6-15
Contact Stabilization 6-15
Conventional Activated Sludge 3-10
Extended Aeration 1-5
It is significant that the solids in extended aeration effluents filter particularly well, in as
much as they often settle poorly, leaving high concentrations in the secondary effluent. This
behavior may be understood from the flocculation studies of Parker, et al, (39) who found
that sludges with high solids retention times lose their tendency to agglomerate into larger
easily settleable particles, but increase in strength so that fewer are broken up into particles
of a size not readily filtered.
9-17
-------
9.3.2.4 Headloss Buildup vs. Solids Capture
While effluent quality reflects the solids which pass through the filters, headless
development reflects the amount and location of solids which deposit in the bed. Both solids
loading (solids concentration times flow rate) and filter efficiency are important in
determining the buildup of headloss with increasing solids capture Various studies relating
headloss buildup to solids capture show widely different results. This would be expected in
view of the wide range of solids characteristics, media characteristics and filter rates, and
the very different headloss patterns that result from surface and depth filtration. Baumann
and Cleasby (1) cite specific solids capture values (average over the filter run) ranging from
0.035 to 0.35 Ib/sq ft/ft of headloss. The variation was mainly in activated sludge effluent.
The trickling filter data, from a single plant in Ames, Iowa, showed values close to 0.07 lb/
sq ft/ft of headloss for a wide range of media sizes and filter rates. British data for trickling
filter effluent, however, showed specific capture values averaging 0.35 lb/ sq ft/ ft of
headloss over a filter run with initial values as high as 0.6 Ib/sq ft/ft of headloss (40). For a
fine (0.5 mm) media with low solids loadings, Tchobanoglous and Eliassien (2) reported
values an order of magnitude lower than the smallest cited by Baumann and Cleasby. It is
reasonable to expect the highest values of specific capture where the filter and influent
solids characteristics permit depth filtration and extremely low values where they promote a
high degree of surface straining.
9.3.2.5 Properties of Solids Affecting Adhesion
Available measures of the properties which affect adhesion of solids particles to other solids
or to media grains are limited to Zeta potential or the related electrophoretic mobility. Very
few studies have included such measures or attempted to relate performance to them.
Tchobanoglous (6) reported that reduction of natural negative electrophoretic mobility of
wastewater solids using cationic polymers improved filter quality performance. Where
sufficient polymer was added to reverse the negative charge of the particles performance,
though excellent at first, deteriorated rapidly after the first hour. With charge reversal,
initial performance apparently was aided by electrostatic attraction between the negatively
charged filter media grains and the positively charged wastewater solids. After an hour,
however, the grains were coated with positively charged solids, and the resulting
electrostatic repulsion interfered with filtration of further solids applied.
9.3.3 Physical Characteristics of the Filter
Most wastewater filter designs employ media configurations and loadings which minimize
surface straining and promote depth filtration (Section 9.4). A few special designs with fine
media (Section 9.6) are intended to remove solids primarily by surface filtration or
straining. These designs include provisions for overcoming the adverse effects of rapid
headloss buildup. Where surface filtration predominates, the media characteristics have
little effect on quality performance or head loss. In addition, removal of solids is quite
independent of filter rate or influent solids concentration (1). Hence the effects of physical
characteristics of filters are discussed below only in relation to depth filtration not surface
straining.
9-18
-------
9.3.3.1 Media Characteristics
The most important media characteristic in determining performance is size. Studies using
uni-size media have clearly demonstrated that finer media have greater removal efficiency
(2) (6) (7) (35) (41). Various investigators have related percent removal to powers of
diameter ranging from -1 to -3 (3). In finer media headloss per unit of removal (Ib/cu in. of
media) is also higher (2).
In a media graded from fine to coarse in the direction of flow, the highest solids
concentration is applied to the layers with the greatest removal efficiency. As a result,
removal is concentrated in a small depth with accompanying high headlosses.
In media graded from coarse to fine in the direction of flow, substantial penetration occurs
but most of the solids are removed in the coarser media where less head loss buildup results.
The finer layers, protected from heavy solids loadings, are available for polishing and to
prevent breakthrough as the coarser layers become filled with solids.
Media depth is most significant in coarse uniform beds. Because of the uniformity, the
efficiency of removal (as a percent of the solids applied to each depth) is nearly constant for
all layers of the filter. Penetration is substantial and extra depth is relied upon for polishing
and to retard breakthrough.
Size and specific gravity of media together are significant in determining expansion during
backwash and the degree of intermixing in multi-media beds.
9.3.3.2 Filter Rates
The effect of filter rates on quality performance can vary widely depending on application.
In filtering biological floe at reasonably low influent solids concentration, the effect on
effluent quality of rates up to 10 gpm/sq ft is not very significant (24) (37) (42). In a study
of ultra high rate filtration (43), operation at up to 32 gpm/sq ft still provided 50 percent
removals compared to 75 percent at 8 gpm/sq ft. With weaker chemical floes or with high
influent concentrations of biological solids (usually indicating poorly functioning biological
treatment) filter effluent quality tends to degrade at filter rates above about 5 gpm/sq ft
(33). Sudden changes in filter rates may affect effluent quality more adversely than
sustained higher rates.
Higher filter rates tend to increase solids penetration. In cases where this significantly
reduces surface removal, head loss buildup per unit volume filtered may actually be less at
.higher rates. This was illustrated in studies at Iowa State University involving settled
trickling-filter and lime-softening effluents (35) (36). For the trickling filter effluent,
production per run (at a given terminal head loss) increased slightly with filter rate over the
range tested up to 6 gpm/sq ft. For the lime effluent, production per run increased with
filter rate up to 5 gpm/sq ft and then decreased (36). Existence of an optimum rate, as in
the latter study, has been suggested as typical of combined surface and depth filtration (1).
It has also been suggested that the advantages of using a coarse top media layer may be lost
if the filter rate is not high enough to force solids into the bed and limit surface straining (7).
9-19
-------
9.3.3.3 Cleaning System Variables
In addition to upflow washing, some form of auxiliary scouring of the media appears
essential to adequately clean wastewater filters. If cleaning is inadequate, two serious
problems will develop: filter bed cracking and mud ball formation. Cracks open in filter
beds because of compression of excessively thick coatings on the filter grains. The resulting
localized heavy penetration of solids may both lower effluent quality and contribute to mud
ball formation.
Mud balls are compressed masses of filtered solids large and dense enough to remain in the
bed during backwashing. If conditions favoring their formation persist, mud balls tend to
increase in size and to sink deeper in the bed. Their presence increases head loss and may
lead to loss of effluent quality.
Both air scrubbing and surface or internal water jets have been used for auxiliary scouring
of the media. Air injected below the media produces shear as the bubbles rise through the
bed. Water jets, positioned at the top of the expanded bed, produce high shear around the
surface media, which is the most heavily loaded with solids. In multi-media beds, jets may
be similarly provided at the expanded height of the media interface.
The main upflow wash and the auxiliary scouring systems should be controlled
independently to permit use together or separately. Washing procedures are discussed in
Section 9.7. The key parameters for design of the cleaning system are the upflow wash rate
capacity and the air scour rate or surface wash rate capacity. Typically, upflow wash rates
are about 20 gpm/sq ft. The maximum capacity is selected to provide the desired degree of
fluidization and expansion of the media under critical high temperatures (See Section 9.7).
Capacities for auxiliary scouring are generally established empirically. Air scour rates
typically range from 3 to 5 scfm/sq ft, and surface wash rates from 1 to 3 gpm/sq ft.
9.4 Selection of Filtration Rate and Terminal Headless
Given adequate information on performance, the filter rate and terminal headless for a
particular media design should be selected by making economic tradeoffs between filter
size, operating head requirements and run length, all within the limits dictated by effluent
quality requirements. This section outlines procedures for such tradeoffs and provides an
alternative basis for selection where specific performance information is lacking.
9.4.1 Information for Economic Tradeoffs
Adequate information for making economic tradeoffs can be obtained only from pilot
studies of the specific media application. (See Section 9.9). Pilot studies should indicate the
buildup of headloss with time for various filter rates and for average and peak influent
solids concentrations. Results may be indicated in a form similar to Figure 9-5. With this
information it is possible to estimate the filter run length, the net production and the capital
and operating costs of the filter for the given influent solids concentrations and for different
combinations of filter rate and terminal headloss (See Sections 9.1 and 9.3).
9-20
-------
50-
40-
co
CE
13
O
X
O
z
UJ
30-
20-
10 -
i
0
CO
CO
o
UJ
a:
UJ
l
2
PEAK SOLIDS LOADING
AVERAGE SOLIDS LOADING
l
4
l
6
i
8
i
10
FILTER RATE (gpm / sq ft )
FIGURE 9-5
RUN LENGTH VS. FILTER RATE FOR
VARIOUS TERMINAL HEAD LOSSES
9-21
-------
In determining net production, allowances must be made, for downtime during cleaning and
for recycle of washwater through the treatment plant. The downtime effects are calculated
from the cleaning frequency, cleaning cycle duration and the number of individual filters.
Washwater recycle effects are calculated from the cleaning frequency and the backwash
rate and duration. Washwater recycle has no effect on net production if filter influent is
used for washing. Net production may be expressed as volume (filter rate x run length) or as
an average rate (gpm/sq ft) over one filter cycle (run length plus cleaning time). The net
production rate is almost the same as the filter rate for runs of 24-hr or more. For run
lengths below 10 to 12-hr the differences become significant (1); below 6 to 8-hr the effect
on production may be critical.
Short term peak loadings due to down time or recycle during backwash need not be
considered directly in economic tradeoffs. After the design filtration rate and terminal
headloss are determined, however, the design should be checked to assure that it can
accommodate these peaks within the available headloss and effluent quality limits. If not,
peak effects should be reduced or eliminated by increasing the number of filter units or by
providing equalizing storage for the backwash and wastewater flows.
The design should also be checked for its ability to handle the sustained peak loads imposed
when a unit is taken out of service for repairs. If the resultant shorter run lengths do
not provide enough capacity, the design may be revised as follows: peak hydraulic
loadings should first be reduced by increasing the number of filters keeping the total area
the same; if this reduction is not sufficient, the area should be increased beyond that
determined in the original design.
Before cost tradeoffs are made, the following must be defined:
1. Maxium flows and solids loadings for various durations up to 24-hr. A tentative
decision is required on the use of equalization to limit maximum wastewater flows.
2. Run length limits. The lower limit should be 6 to 8 hr to maintain reasonable net
production. The upper limit should be 36 to 48 hr to avoid anaerobic
decomposition of solids in the filter (1).
3. Head loss limits. For gravity filters allowable head losses generally are below 10 ft.
Use of heads much above this commits the design to pressure filters. Use of
pressure filters would be favored where pressurized discharge to following
facilities is needed (1). Gravity filters would be favored where the extra head for
pressure filters would require intermediate pumping but head for the gravity units
is available without such pumping.
4. Backwash design and expected cost per cycle. Manpower costs should reflect
whether the operation is to be automated. Backwashing costs should include costs of
treating recycled backwash in units ahead of the filters, and the recycled flow
should be deducted in determining the net productio
Mail cotib 3404T
1200 Pennsylvania Avenue NW
9-22 Washington, DC 20460
202-566-0556
-------
5. Space limitations. These may force use of higher filter rates.
6. Number of filter units. This should be tentatively selected to facilitate cost
estimates, but may be varied with little effect on the tradeoff calculations provided
labor is not a major factor in the operating cost per backwash. For reliability and
economy, a minimum of four to six units should generally be provided, with at
least two in even the smallest installations. Above these minimums, the number of
filter units, depends on the actual size of individual units. The practical maximum
size of gravity filters is about 800 sq ft.
In addition to limits indicated above, pilot testing may reveal: 1) upper limits on headloss or
rate required to avoid solids breakthrough and effluent quality deterioration, 2) an
optimum filter rate for minimizing headloss buildup. No filter rates lower than the
optimum should be considered in the tradeoffs.
9.4.2 Tradeoff Procedures
The following procedures are suggested for determining the most cost effective filter sizing,
design terminal head loss and run length. Figure 9-6, relating net production to filter rate
and run length, has been prepared to facilitate the analysis. The figure should be modified
before application if backwashing conditions differ significantly from those assumed in its
development.
1. From filter test data for average and maximum design influent solids con-
centration, prepare a headloss development plot (see example plot Figure 9-5).
2. Assume initial trial values for terminal headloss and filter run length. (See Item 12).
3. For the assumed terminal headloss and run length determine the filter rate from the
headloss development plot for maximum solids concentration.
4. For this filter rate and the assumed run length determine the net production rate
from Figure 9-5.
5. Determine filter sizing based on this net production rate and the maximum design
flow for a duration equal to the filter cycle time (run length plus downtime for
cleaning).
6. Estimate capital costs for filters based on above sizing and the design terminal head-
loss.
7. Determine average net production by dividing average flow by filter area.
8. Construct a plot of net production vs. filter rate based on run lenths to reach the
trial value of terminal headloss at various filter rates with average solids concentra-
9-23
-------
10
8
I 6
UJ
cc
5
z
o
o
§ «
cc
Q.
UJ
2 3
INFINITE RUN LENGTH
71
20 HOUR RUN LENGTH^
10 HOUR RUN LENGTH
6 HOUR RUN LENGTH
6 HOUR RUN LENGTH
NET PRODUCTION RATE AT AVERAGE SOLIDS
LOADING AND IO FT. TERMINAL HEADLOSS
I
I
I
345
FILTER RATE
6 7
(GPM/SQ. FT.)
8
IO
FIGURE 9-6
NET PRODUCTION RATE VS. FILTER RATE
FOR VARIOUS RUN LENGTHS
9-24
-------
tions. (See Example Figure 9-6).
9. From the plot in 8. determine filter rate and run length to provide average
net production.
10. Calculate operating costs based on the assumed terminal headloss and the run
length for average flow and solids loading.
11. Convert operating costs to present worth and add to capital cost to determine total
present worth.
12. Repeat above analysis assuming different values for terminal head loss and filter
runs. The objective is to find assumptions which minimize present worth, within
technological constraints.
It is suggested that a conservative initial value of 8 ft be assumed for terminal head-
loss with a run length of at least 8 hr at maximum solids concentrations. Subsequent
trials would explore use of higher headloss values to permit either longer runs or
higher filter rates whichever appears more advantageous. Judgement must be ap-
plied to minimize amount of calculation required.
9.4.3 Selection Without Pilot Testing
Where it is impossible to test proposed filter media on the actual influent, guidance may be
obtained from results with the same media treating similar influents. In the absence of spe-
cifically applicable test results, filter rates and headloss allowances should be very con-
servatively selected, based on ample estimates of influent solids concentrations. To assure
adequate capacity it is suggested that, as a minimum, sufficient filter area be provided to
handle the 24-hr design flow at 4 gpm/sq ft or the 4-hr maximum design flow at 6 gpm/sq
ft, whichever is more stringent. For predominantly chemical floe, the surface media should
be no finer than 1 mm and allowance should be made for a terminal headloss of 10 ft. For
filtration of biological solids in secondary effluent the following procedures are suggested in
selecting terminal headloss and final filter sizing:
1. For the minimum filter area as determined above, estimate headloss buildup
based on expected solids removals and the following values of specific
capture:
Minimum Media Size
at Influent Surface Specific Capture
mm Ib of solids removed/sq ft/ft of
headloss increase
1.8 0.07
1.3 0.035
9-25
-------
2. Avoid use of any finer surface media. Surface media coarser than 1.8 mm may per-
mit higher specific captures but problems of adequate cleaning must be considered
(See Section 9.7).
3. For the minimum filter area calculate the required head for 24-hr run length at aver-
age solids loading and for 8-hr run length at maximum (8-hr) solids loadings.
4. Provide for terminal headloss on the more critical basis above or use more than
minimum filter size and recalculate solids loadings and headloss requirements..
Designs based on the criteria above should be as flexible as possible to permit use of higher
rates or lower heads if operating experience shows this is possible. Flexibility to increase
rates is most valuable where capacity is to be increased in future stages. Flexibility in
pumping and control systems will permit head to be reduced to what proves necessary in
actual operation.
9.5 Filtration Media
9.5.1 Materials
Media commonly used in water and wastewater filtration include silica sand (sp gr 2.65),
anthracite coal (sp gr 1.4 to 1.6) and in special multi-media designs garnet (sp gr 4.2) or
ilmenite (sp gr 4.5).
As they occur in nature these materials are not of uniform size but instead typically have a
grain size distribution such as that shown in Figure 9-7. Fair and Geyer (44) discuss size
measures for irregular particles, equivalent diameters, shape effects, etc. Natural grain size
distributions frequently are close to geometrically normal, i.e.plot as a straight line on log
probability paper. As shown in Figure 9-7, grain size distributions are often characterized
by two points, the 10 percent and 60 percent size (dio and deo). These are sizes such that
the weight of all smaller particles constitutes respectively 10 or 60 percent of the whole.
Media is frequently specified in terms of effective size (d 10) and the uniformity coefficient
(deo/d 10).
It is possible to change the characteristics of a given media material by removing certain
size fractions. Coarser fractions may be seived out while finer fractions may be removed by
"scalping" (removing surface layers) after hydraulically grading material during upflow
washing. Fair and Geyer (44) present a method of calculating the size fractions which must
be removed to convert from one size distribution to another.
The most important modification for most media is to remove any very fine particles—say
those less than 80 percent of the effective size. Such fine material never constitutes more
than a small fraction of the media volume but, if not removed, may cause headlosses far
greater than would be expected for the given effective size.
With sufficient effort in size separation, it is possible to produce almost uniform media.
9-26
-------
U.S. Standard Sieves
K)
1" 3/4" 3/8" 4 6 10
20 40 60 100 200
Grain Diameter in Millimeters
FIGURE 9-7
GRAIN SIZE CURVE
0.1
0.01
-------
Such media are frequently used in experimental investigations, but most designers have not
considered the extra cost justified in full scale installations. Important exceptions should be
noted, however: One equipment supplier, Dravo, emphasizes use of uniform coarse media in
deep beds; Baumann (7) recommends use of uniform anthracite and sand in dual media
filters pointing out that the extra cost is probably not more than 1 percent of the overall
cost of the filters.
9.5.2 Dual and Multi Media
Upflow washing stratifies a bed in accordance with the settling velocities of the media
particles as determined by their size, shape and specific gravity (See Section 9.7). In a dual
or tri-media bed, although each media component is still graded fine to coarse in a
downward direction, lighter coarser media can be maintained above finer denser media.
This makes it possible to approximate a coarse-to-fine gradation in down flow filtration
units. Another advantage of dual or multi-media over a single medium is that mud balls
formed in the filter remain above the coal-sand interface where they are subject to auxiliary
scrubbing action (1).
The maximum settling velocities of media particles also determine the minimum wash rate
required for adequate fluidization of the bed during backwash. Hence, for a given media
size at the top of the bed, lower wash rates can be used if each media component is more
uniform and the top portion of the filtration is of anthracite rather than a heavier material.
Baumann and Cleasby (1) recommend that dual or multi-media be sized so the coarsest
(dgo) sizes of each component have about the same minimum fluidization velocity.
9.5.3 Pore Size and Intermixing of Media Components
The hydraulic behavior and filtration performance of any given media are more properly
related to pore size than to grain size. For single media component, pore size is directly
proportional to grain size, and the porosity (percent of volume represented by pores) is a
constant depending only on media shape. Coal which tends to be angular has a porosity of
almost 0,5 whereas sand porosity is closer to 0.4. In water treatment applications, coal
media, because of its greater porosity, has been found to give poorer removals but lower
pressure losses than sand of the same grain size (45).
The pore size in multi-component filter media depends on the degree of intermixing of the
components. With no mixing, pore size distribution simply follows that of the components.
With intermixing, however, the finer layers of the denser material below are dispersed into
the voids of the coarser layers of lighter material above. No precise methods have as yet
been demonstrated for calculating actual pore size, or even the degree of intermixing, from
the characteristics of media components. Where such information is of interest it may be
obtained from test columns or from experience with specific combinations of components in
other installations. Limiting size ratios have been proposed to control intermixing and to
avoid the extreme where lower density coarse media is overtopped by very fine high density
media. Camp (46) has hypothesized that for dual media filters with an interface size ratio
(coarsest coal/finest sand) of 2.8 no intermixing should result, whereas for a ratio of 4.0
9-28
-------
intermixing would occur over a depth of about 5 inches. Baumann (7) indicates only limited
intermixing (6 inches) and no overtopping with highly uniform coal and sand having a size
ratio of 3.35.
Gulp and Conley (47) indicate that to avoid overtopping in dual media beds, the
effective size of the coal grains must be no more than about three times the effective
size of the sand. Whatever the exact limiting ratio, designs using coarser
anthracite to accept higher solids loadings at lower head losses must also use
correspondingly coarser sand. Where stringent effluent quality or weak
floe conditions require a finer media component than the sand, garnet can be used.
Coal/garnet size ratios as high as five will not result in overtopping.
The significance and desirability of intermixing in dual or multimedia beds is a subject of
debate. Camp (46) has reported deliberate selection of dual media sizes to minimize
intermixing, whereas some manufacturers actively promote intermixing as advantageous in
three and four component media, claiming that controlled intermixing approximates a
"theoretically ideal" coarse to fine gradation of voids in the direction of filtration.
In side by side tests at Washington D.C. (See Table 9-3) on chemically treated effluent,
mixed (tri) media filters did show slightly better effluent quality performance than dual
media filters. However there was no evidence to demonstrate that this was due to
intermixing rather; than just to the fine, high specific gravity garnet present at the
bottom of the filters.
9.5.4 Specific Media Designs
Table 9-5 lists typical characteristics for several specific media configurations which have
been used in normal design of wastewater filters. All provide initial filtration through coarse
media either by upflow filtration or by downflow filtration using dual, tri or deep, uniform,
media. Configurations using fine single media and hence requiring special cleaning
provisions, are not included. Typical application conditions (floe strength, solids load) are
shown for each design. The dual-media designs for the most part employ depths ranging
from 30 to 36 inches. To allow for level variations due to uneven backwashing, sand depths
are set at 12 to 15 inches even though only the top few inches significantly affect removals
Minimum depths for the anthracite are 15 "to 18 inches. Greater depths may be necessary
where solids loads are heavy.
In tri-media designs the overall depths and the minimum depths of anthracite and of the
combined finer media are in the same range as in dual media designs.
The single media configurations employ depths of 60 inches or more. In downflow filtration
this great depth is intended to improve efficiency, while in upflow units it has an additional
purpose of adding weight to restrain the bed from uplift due to differential pressures during
operation. Where uplift exceeds the submerged weight of the media it will either fluidize the
bed or lift it in a "piston" effect (small diameter filters).
9-29
-------
TABLE 9-5
TYPICAL MEDIA DESIGNS FOR FILTERS
Ul
o
Coal
Sand
Garnet
Media
Design
Single
Single
Dual
Dual
Tri
Tri
^
Dia.
mm
—
—
0.9
1.84
1.0-1.1
1.2-1.3
Unif.
Depth Coeff.
in
• — —
—
36 <1.6
15 <1.1
17 1.6-1.8
30
Dia.
mm
1-2
2-3
0.35
0.55
0.42-0.48
0.8-0.9
Depth
in
60
72
12
15
9
12
Unif.
Coeff.
1.2
1.11
<1.85
< 1.1
1.3-1.5
—
Dia.
mm
—
—
—
—
0.21-0.23
0.4-0.8
Depth
in
—
—
—
—
4
6
— — — — i ypicai
Unif. Application
Coeff. Conditions
A
A
B
A
1.5-1.8 B
C
Reference
13
48
29
7
25
13
NOTES: A = Heavy Loadings, Strong Floe.
B = Moderate Loadings, Weaker Floe.
C = Moderate Loadings, Strong Floe.
-------
Some of the theoretical advantage of upflow coarse-to-fine filtration is lost because min-
imum grain sizes must be coarse enough to avoid excessive uplift.
Additional resistance to uplift is provided in many upflow designs by placing a restraining
grid on top of the media. The spacing between bars of the grid must be large enough to pre-
vent upward bed movement during filtration. Although these two requirements appear con-
tradictory, arching of the grains takes place between the bars, allowing a reasonably large
spacing, in the range of 100 to 150 times the diameter of the smallest grain size in the beds.
9.5.5 Selection of Media
Pilot testing is indispensable to provide the information necessary for meaningful com-
parison of different media designs or to assure the effluent quality performance of any
media design selected. Without pilot testing, the designer should select a media which, on
the basis of experience with similar influents, may be expected to provide good solids re-
moval with low head loss buildup. In general, any such media would include an ample depth
of coarse media followed by fine media in the size ranges indicated for dual media con-
figurations in Table 9-5.
Pilot testing to guide media selection should define headless development vs time for each
media design, under all test conditions. Suggested ranges for test conditions are given in
Section 9.9. If one media design clearly gives lower head buildup at all times and under all
test conditions it may be selected directly provided its backwash requirements are not ex-
traordinary. If different media provide essentially the same headloss development over the
range of test conditions, selection may be based on other factors. Where different media ap-
pear significantly better under different conditions, selection should be based on cost com-
parison of the alternative designs each at its most favorable rate, terminal headloss and run
length, determined as indicated in Section 9.4. Significant differences in backwash flows
should be taken into account.
9.6 Filter Control Systems
Major filter functions requiring monitoring and/or control are:
1. Head Loss
2. Effluent quality
3. Initiation of backwash where automatic
4. Flow rate through the filters
5. Backwash sequence, rate and duration.
An important tool in performance control is the automatic turbidimeter which can contin-
uously monitor the filter feed and product. This allows the operator to anticipate difficulties
9-31
-------
from changes in feed quality, and rapidly remedy process failures. In addition, these devices
allow the operator to rapidly evaluate the effects of changes in process variables and pro-
vide a continuous record of plant performance. All turbidimeters operate on the pinciple of
measurement of scattered or transmitted light. A variety of commercial instruments are
available.
Filter installations should be equipped with appropriate loss of head and flow indicators. In-
dividual filters should have multiple taps for pressure readings if full scale experimental
testing is desired.
Provisions for automatic or remote initiation of backwash by timers or based on headloss
or turbidity monitoring may be justified to reduce the need for operating attention.
Three types of flow control systems are used for filters:
1. Effluent rate control
2. Influent flow splitting
3. Variable declining rate control.
Features of these systems and of automatic backwash systems are described below.
9.6.1 Effluent Rate Control
This system, common in traditional water treatment plant designs, maintains a set flow for
each filter by throttling the effluent (see Figure 9-1). The throttle valve may be controlled
directly by mechanical linkage to a venturi controller or indirectly by a set point controller
linked to a pneumatic or hydraulic valve operator. The direct acting system is unsuitable if
flows to individual filters must vary over the day. The indirect system is complex and both
may be troublesome in maintenance. The system is also wasteful of head since available
head not needed in a clean filter is lost in the controller. In addition, control valves may
produce high frequency surges in the filter bed with accompanying loss of efficiency (10).
9.6.2 Influent Flow Splitting
In this system flow is evenly divided among filters in a splitter box located at or above the
level of the top of the filter boxes (see Figure 9-8a). The boxes themselves are made deep so
that the water level in them can build up to provide the maximum operating head needed
when the filter bed is dirty. A weir on the filter outlet maintains a constant back pressure or
minimum water level to prevent accidental dewatering of the bed.
Advantages of influent flow splitting include:
1. Rate controllers with attendant maintenance and surging problems are eliminated.
9-32
-------
Flow Splitting Tank
Inlet
Valve
Open
O
o
Closed
Backwash
Outlet
Valve
s
Filter Cell
No. 1
Wash Trough
nnj-Ln.
-
Filter Cell
No. 2
Wash Trough
n_rLn_n.
t
-*—
•+
Open
o
-------
2. Flow variations are distributed to filters automatically.
3. Head loss may be read directly from water levels in filter boxes.
4. Only a single master flow meter is needed.
5. Changes in filter rate are gradual because of time required for head to build up in
filter boxes.
Disadvantages are:
1. The head not needed for filter operation is lost in the drop between the splitter and
the filters.
2. Capital cost of filter box construction is increased by the greater depth.
9.6.3 Declining Rate Filtration
This system requires multiple filters. All operate under the same head but at different flow
rates depending on the degree of clogging. Under constant head the output from a single fil-
ter declines as the run progresses. The filter selected to be backwashed is always the one
which has been on line the longest and is most clogged. Total output from all filters is con-
trolled by varying the head applied. Figure 9-8b shows a variable declining rate filter.
The head on the filters may be controlled by varying either the upstream or downstream
water level (10) (49). With downstream water level control, an equalizing chamber must be
provided to limit the rate of change of head and hence of flow, when filters are taken off line
or restored to service. It is common to apply maximum design loadings to the filters as a
group and to limit maximum rates on individual clean filters to from 20 percent to 40 per-
cent above these design loadings.
Advantages cited for declining rate filtration (10) (49) include better effluent quality,
absence of surges, and significantly lower total head requirements. Less head is needed
because:
1. There is no loss due to throttling or due to free fall after flow splitting.
2. As rate declines turbulent head losses (underdrains, valves, etc.) reduce rapidly (in
proportion to second power of flow) making head available to overcome resistance
of clogged filter (proportional to flow).
For proper operating control, flow rates should be measured individually for each declining
rate filter. Only single indicators are needed for inlet and outlet levels on head loss, since
these are the same for all filters.
The chief disadvantage of this method of flow control is the need for a large volume of
9-34
-------
water storage upstream of the filter.
9.6.4 Backwash Control
Programmed backwash systems are widely used in current designs. Such systems consist of
interlocked controllers and timers programmed to open and close valves, make or break
siphons, start and stop pumps and blowers and limit backwash flows to control the rate, du-
ration and sequence of activities during backwash. Even where backwash is manually in-
itiated, the rest of the control system may be entirely automatic.
Proprietary systems, with various features are available from different manufacturers. One
such system, designed to operate with only a single control valve is shown in Figure 9-9.
9.7 Filter Cleaning Systems
9.7.1 Upflow Washing
Accumulated solids are removed from filters by a rapid upflow of washwater. The waste
flow is then recycled to some prior treatment unit, usually primary settling. Washwater
sources may include filter influent, filter effluent or effluent from subsequent treatment
units. Storage of washwater supply may be needed if rates required exceed the flow avail-
able. Recycled spent washwater flows should be equalized by storage so they do not disrupt
prior treatment processes. Backwash rates for most effective cleaning vary with media size
and density.
Baumann and Cleasby (1) recommend providing upflow wash capacity adequate to fluidize
the 90 percent finer size of each media component at the warmest expected water tempera-
ture. Figures 9-10 and 9-11 may be used to determine minimum upflow velocities or wash
rates to fluidize coal, sand, and garnet media of various sizes. Rates should be variable to
compensate for changes in temperature, viscosity, and hence bed expansion. Maximum hy-
drodynamic shear and most efficient cleaning have been shown to occur when the porosity
of the expanded media is about 0.7 (1). To reach this porosity in the surface layers of a
non-uniform sand (effective size = 0.4, uniformity coefficient = 1.47) requires almost 50
percent expansion (7). Because of its higher unexpanded porosity coal requires only 20 to 25
percent expansion to reach a porosity of 0.7 in the surface layers. In practice the backwash
rates generally used in filter designs, range up to 20 gpm/sq ft, and do not provide more
than 15 to 30 percent expansion except for very fine media (50). This means that higher
backwash durations (5 to 10 or at the extreme 15 minutes) and somewhat higher washwater
consumption are required than at rates which would provide the most efficient washing.
Whatever wash rate and duration are expected, design of piping, valves, pumps and storage
tanks should provide extra capacity of at least 25 percent.
Methods are available for predicting expansion of sand beds accurately (51), but have not
yet been satisfactorily extended to other media components. Expansion of multi-component
media can be rapidly obtained, however, from backwash tests in pilot columns.
9-35
-------
FFFLUEN1
c
-.
-
AIR
BACKWASH WATER
INFLUENT
ANTHRACITE
SAND
ROTO SCOUR
UNDERDRAIN
FIGURE 9-9
AUTOMATIC GRAVITY FILTER, SINGLE COMPARTMENT
(Courtesv^of Ecodyne Corp.)
-------
0.12
0.10
0.08
0.06
0.04
0.02
0.0
0.
COAL p= 1.7
= 1.73
o
-o-
I
I
p= .
SILICA SAND p = 2.65 o
GARNET SAND p= 4.13 A
I I I __ I
60
50
40
30
20
10
.ST
25 0.5 0.75
1 1.5 2
MEAN SIEVE SIZE, mm
FIGURE 9-10
MINIMUM FLUIDIZATION VELOCITY, Vmf,
TO ACHIEVE 10 PERCENT BED EXPANSION
AT 25°C(1).
2.5
9-37
-------
0.04
0.03
0.02
o.oi
1.0
0.5
.1
'o
Q.
0)
u
(5
u
U
o
CO
10
TEMP,
20
°C
30
FIGURE 9-11
EFFECT OF TEMPERATURE ON Vmf FOR
SAND AND COAL AND ON ABSOLUTE VISCOSITY
OF WATER (1)
9-38
-------
Several manufacturers offer coarse media filter systems with wash rates too low to fully
fluidize the bed. Auxiliary air scouring is provided, but its long term effectiveness in main-
taining the media in good condition has not been proven. Studies at Iowa State University
(52) indicate that air scour and simultaneous upflow washing at full fluidization can free a
dirty filter bed of mud balls which accumulate during multiple cycles using air scour follow-
ed by washing. Further work will study effectiveness of air scour and simultaneous upflow
washing at less than full fluidization.
9.7.2 Underdrains
Underdrains should distribute washwater as uniformly as possible over the area of the filter.
Excessive variation in washwater rate results in uneven and ineffective cleaning. Moreover,
the accompanying excessive jet action can lead to lateral displacement of gravel and clog-
ging of the underdrains with filter media.
In general, underdrain systems developed for water filtration may also be used in waste-
water applications. One of the first systems employed in water filtration consists of several
layers of graded gravel surrounding manifold piping positioned on the filter floor. Orifices
in the manifold piping provide preliminary distribution of the washwater as shown in Figure
9-12a. The final distribution is accomplished as the water moves upward through the gravel.
Fair and Geyer (44) provide rules of thumb for sizing lateral header systems and a basis for
hydraulic analysis.
Several commercial systems are available which employ patented false bottom distributors
in place of the manifold. The gravel is then placed on top of the false bottom. Leopold
Block andWheeler Filter Bottom systems are illustrated in Figure 9-12b and c. A system
manufactured by Dravo (Figure 9-12e) distributes both air during airwash and washwater
during normal backwash.
Table 9-6 shows the standard gravel design used in water filtration together with a modified
design suggested by Baylis (45) for use with higher backwash rates.
9-39
-------
A. HEADER LATERALS
(COURTESY OFTHEAWWA)
5/52" DIA. DISPERSION ORIFICES
APPROX 49 PER SO FT.-
6/8* OIA CONTROL ORIFICES
APPKOX. tPER 80. FT.
COMPENSATIN6 LATERAL
(•ECONDARY126.B SO IN&
FEED LATERAL(PRIMARV)
90 8 80 IN.
B. LEOPOLD BLOCK SYSTEM
(Court«»y F. B- Leopold Co-)
tuenx I/*"-
at* 1/4"-
LAYER I A"-1/4"
WOOD STRIPPH4C TO
PRgVPfTMIOUT I
WTO FUMC
<•-, •„ .;>v*- •
-.-•' »-- ***-
- PLUME AREA '1.5 [FILTER AREA)
(MM) I UO FT) J
I«T RCCOMMOOCD
If* MAXIMUM
FIGURE 9-12
UNDERDRAINS
9-40
-------
C. WHEELER FILTER BOTTOM
(Courtesy of B I F )
!_*"-•' . " " '*'- ** ' - * / ' * t - ..A ' * _ iJ*^
D. SUBFILL-LESS STRAINERS
(Courtesy of Ecodyne Corp.)
FIGURE 9-12 (Continued)
9-41
-------
RAW
WATER
STABILIZING
LAYER
MEDIUM
SUPPORTING
LAYER
AIR
DISTRIBUTION
PIPES
CLEAR WATER
a
WASH WATER
CONDUITS
CLEAR
WATER
FILTERING
FILTER BED
FINE
SUPPORTING
LAYER
COARSE
SUPPORTING
LAYER
M-BLOCKS
COVER
PLATES
FILTRATE OUT <—
*- RAW WATER IN
E. DRAVO M-BLOCK SYSTEM
(COURTESY OF DRAVO CORR )
FIGURE 9-12 (continued)
9-42
-------
TABLE 9-6
FILTER GRAVEL DESIGN
Standard
2-1/2
3-1/2
3-1/2
2
4
6
1/12 to 1/8
1/8 to 1/4
1/4 to 1/2
l/2to3/4
3/4to 1-1/2
1-1/2 to 3-1/2
Baylis
Depth
in.
5
2
2
4
2
2
4
1 to 2
1 / 2 to 1
1/4 to 1/2
1/8 to 1/4
1/4 to 1/2
1/2 to 1
1 to 2
Baylis found that with the standard design the upper layers of gravel could fluidize under
high backwash rates (above 20 gpm/ sq ft), and proposed placing a final layer of heavy
gravel over the finer gravel to prevent this fluidization. A 3-inch layer of coarse (1 mm) gar-
net or ilmenite above the gravel has also been suggested to overcome this problem (8).
Some systems eliminate the gravel layers entirely by using nozzles or porous plates. Advan-
tages and disadvantages of these systems are discussed elsewhere (1) (8). One system, manu-
factured by Graver Water Conditioning Co., utilizes a cast concrete false bottom with plas-
tic or metal strainers on about 12-inch centers as shown in Figure 9-12d.
9.7.3 Washwater Troughs
Except for small units, filters are commonly equipped with washwater troughs spaced on
about 6-ft centers. Hydraulic design of these devices is discussed in Fair and Geyer (44).
Troughs are installed with the aim of aiding uniform distribution of washwater and avoid-
ing dead spots which may retard the removal of dislodged solids from the filter box.
9.7.4 Auxiliary Cleaning
The function of auxiliary cleaning by air scour or surface wash is to loosen accumulated de-
posits from the filter media. The slimes and organic particulates encountered in wastewater
filtration cannot be completely loosened by normal backwash flow.
Typical surface wash equipment consists of either fixed or rotating pipe distributors fitted
with nozzles which are placed about 1 to 2 inches above the top of the bed. While the sur-
face wash is on, the backwash expansion is set at a lower rate than after the surface wash is
terminated. Surface washwater is supplied at 50 to 100 psi at rates approximating 1 to 3 gal/
min/sq ft of bed.
9-43
-------
With the widespread use of media permitting deeper floe penetration, the ability of the
above type of surface wash to clean mid-and lower portions of the bed has been questioned.
Wash jets positioned at lower levels in the bed may help to alleviate this problem.
Air scour systems have been increasingly used in an attempt to reduce washwater require-
ments and to effect cleaning of the deeper portions of the bed. Some concern has been ex-
pressed concerning loss of finer lighter media particles when air washing is used (8). Where
this is a problem, air scour should be applied separately from the backwash, with liquid in
the filter box drawn down below the washwater overflow level so that no overflow occurs
during air wash. Allowance must be made for 6 to 9 inches of water level rise due to air lift
(7). Although some of the lighter media may remain on the surface of the water and sub-
sequently be lost, the rate of such loss should generally be negligible.
To prevent air scour from disrupting gravel placement, air is usually injected through a grid
above the under-drain gravel. It may go directly into the underdrains where no gravel is
used.
9.7.5 Backwash Sequence
The cleaning cycle time (total downtime during one cleaning operation) includes time for
valve openings and closings, time for drainage of inflow from the filter and time for the ac-
tual upflow washing and auxiliary cleaning. Unless influent in the filter box is to be wasted,
drainage time should be calculated at normal filter rates. Valve openings and closings to
start and stop backwash and air scour should be gradual to keep from upsetting the media
gradation and structure.
A typical sequence for backwash with auxiliary surface wash is:
1. Shut influent and permit water level to drain down to top of the washwater troughs
or other washwater control weir.
2. Apply surface wash for 1 to 3 minutes.
3. Apply upflow wash and surface wash together 5 to 10 minutes as needed to flush
out solids.
4. Shut off auxiliary wash and apply backwash alone for 1 to 2 minutes at rate
needed to classify the bed.
5. Return bed to service.
A typical sequence for cleaning using upflow wash and air scour is:
1. Stop influent and lower the water level to a few inches above bed.
2. Apply air alone at 2 to 5 cfm/sq ft for 3 to 10 minutes.
9-44
-------
3. Apply water backwash at 2 to 5 gpm/sq ft with air on until water Is within one foot
of wash water trough.
4. Shut air off.
5. Continue water backwash at normal rate for usual period of time.
6. Apply backwash for 1 to 2 minutes at a rate required to insure hydraulic classifica-
tion of the filter media.
7. Return bed to service.
9.8 Filter Structures and General Arrangement
A typical wastewater filter consists of a tank or filter box containing an underdrain system,
media and sufficient overall depth to contain media during backwash. In gravity units, the
overall depth must also provide for operating submergence and freeboard. Influent, ef-
fluent, washwater and waste connections are provided. In addition, all wastewater filters
should have provisions for auxiliary cleaning.
Underdrains are designed to properly distribute the washwater during cleaning (see Section
9.7). During normal operation,underdrains~collect filter effluent (downflow operation) or
distribute influent (upflow operation). Washwater troughs and filter inlets (downflow) or
effluent launders (upflow) are located in the submerged zone above the media.
The recommendations for general arrangement and special structural features of concrete
filters presented in Water Treatment Plant Design (53) are fully applicable to wastewater
treatment applications.
For gravity filters of concrete construction, filter boxes are usually arranged in rows along
one or two sides of a common pipe gallery, narrow side toward the gallery. This maximizes
common wall construction and minimizes piping runs. Gravity filters may be of concrete or
steel shell construction. Concrete units are generally square or rectangular and steel units
circular. Sizes of gravity concrete units are limited to about 1000 sq ft (8); steel units are
generally smaller.
For filters using influent flow splitting (see Section 9.6) multiple filter boxes have been con-
structed as compartments in a single round or square tank (concrete or steel) with common
influent and waste piping located above the center of the tank and common washwater and
effluent piping around the outside base.
Steel shell package pressure filters are cylindrical units with either horizontal or vertical
axes. To minimize piping runs horizontal units are usually placed in rows with common pip-
ing along the ends. Vertical units are arranged in either rows or clusters. Horizontal pres-
sure units are less restricted in size than the vertical pressure units and hence are normally
used for plant capacities above 1 to 1.5 mgd (8). Where pressure units are used it is essential
9-45
-------
that manholes be provided for interior access both above the bed and below the under-
drains. Pressure filters should also be provided with a means for hydraulic removal of all
the filter media, and with sight glasses for observation of the bed.
In municipal water filtration plant designs it is common to totally enclose the pipe gallery
and to locate controls in an enclosed superstructure above the gallery and overlooking the
filters. In northern climates the filters themselves are usually included under the super-
structure.
Wastewater effluent temperatures are generally somewhat higher than the local natural wa-
ters, so in a given locality there is less justification for housing wastewater filters than water
filters. Piping and valves need to be protected in climates where freezing occurs either by
housing or by insulation and heating. Controls need housing or weather protected enclo-
sures in any climate. Local controls for each filter should be placed in a location from which
the backwashing filter can be observed.
9.9 Pilot Studies
Specific pilot study objectives may include:
1. Comparing performance of different media designs in a given application.
2. Establishing relations between flow rate, headloss and run length for a particular
media and application.
3. Establishing limiting headlosses and rates to assure required effluent quality and
no deterioration due to breakthrough.
4. Characterizing wastewater variability in terms of performance variations with a
specific media.
5. Studying effects of variations in chemical coagulation, flocculation, biological pro-
cesses or other pretreatment.
In all cases the variation of influent and effluent quality and headloss buildup with time
should be noted. Studies aimed at media selection may be limited to variations of size and
depth for a particular configuration or may involve parallel operation of quite different
media configurations.
Media selection tests should cover the range of operating conditions which may occur in the
actual design: typically flow rates of at least 6 to 8 gpm/sq ft..headlosses to 30 ft (urilesslT
lower limit is imposed by constraints) and runs to at least 24-hr (if headloss or break-
through do not limit). Effects of influent quality variations are revealed by conducting
long-term tests on the same influent. It is important however to include influent quality con-
ditions representative of the entire range expected in operation.
9-46
-------
Studies to select filter rates and terminal headlosses for a particular media will generally in-
clude more rates and cover a somewhat wider range than a study involving media selection.
Also extra effort should be given to including critical influent conditions (e.g. solids load-
ings and floe strength).
Studies aimed at characterizing the filterability of a waste (see Section 9.1) would generally
be run at a low rate (2 to 4 gpm/sq ft) using a media expected to give high effluent quality
but without excessive surface filtration. Changes in quality and headloss with elevation in
the bed should be observed. The widest possible range of influent conditions should be in-
cluded.
Studies to test pretreatment generally would employ a specific media and a single filter rate,
but would attempt to vary influent conditions by adjustment of pretreatment.
In judging what are critical influent conditions, variations in operating records of existing
pretreatment facilities should be thoroughly studied. Where facilities do not exist, records
for similar treatment at other locations should be considered.
Where existing biological pretreatment is inconsistent, with frequent periods of upset, it
would be prudent to determine and eliminate the causes of these upsets rather than count on
filtration as a cure-all for resulting problems (33).
Standard methods for conducting filtration pilot studies have not been established. The fol-
lowing is a list of typical equipment and practices.
1. Multiple filter tubes of transparent material with a minimum diameter of 4 to 6 in.
are utilized.
2. The tubes are fitted for either gravity or pressure operation.
3. A false bottom underdrain is utilized with either a gravel covered plate or strainer
backwash system.
4. Flow to filter units is set by a combination of positive displacement metering
pumps, weirs, control valves, etc. Declining rate control or influent flow splitting
may be used where important to test for design.
5. Sample taps are provided above and below the media, as well as at other locations
within the bed. Tap locations are generally located near the top of each type of
media used. If the effects of media depth variations are to be studied, however,
taps should be located at 3 to 4 in. intervals down the column. As an alternative,
parallel multiple columns of different lengths may be used.
Additional details of pilot filters are given in various references on filtration studies (7) (54)
(55).
9-47
-------
Pilot studies as outlined above cannot adequately determine effects of cleaning system de-
sign parameters. Cleaning of the filter bed is difficult to simulate in pilot scale because of
the small surface area of the beds utilized. The small area makes it impossible to study sur-
face wash and air scour. Results of water backwash may not be representative because of
the wall effect.
Information on cleaning performance can be obtained only in long term studies using large
pilot installations with filtering areas of 4 sq ft or more. Because such studies are expensive,
it may be desirable to design cleaning systems based upon experience from other studies
(50) (51) (52).
9.10 Special Designs
9.10.1 Radial Flow Filters
Some recently developed filters employ horizontal radial flow through media contained in
the anular space between concentric vertical cylinders. The inner cylinder acts as distributor
and draw off points are distributed around the periphery of the outer cylinder which forms
the filter vessel.
In the Simater unit (Figure 9-13) developed in England and marketed in the U.S. by Dravo,
the sand media is continuously moved downward, drawn off, washed in a separate tank and
returned to the top of the bed. A filter developed by Hydromation Corporation (Figure
9-14) has a batch external wash to clean its synthetic resin media.
A Simater filter was tested on biologically treated wastewater at the Essex River Authority
in England. The unit was run in parallel with two upflow units (23). Media size for the Sim-
ater filter was 0.5 to 1.0 mm, comparable to one of the upflow filters but much finer than
the other. Rates were not stated for the radial flow unit but were apparently in the same
range as for the upflow units—4 to 6 gpm/sq ft. All the filters gave SS reductions of 60 to
80 percent and effluent SS concentrations below 10 mg/ 1.
The Simater showed marked tolerance for short slugs at high influent solids. Prolonged op-
eration at high concentration resulted in clogging of the outlet screens, but this could "be pre-
vented by using higher rates of media washing.
9.10.2 Travelling Backwash Filter
Hardinge Corporation furnishes a fine media (0.48 mm sand), multi-compartmented filter
in which each compartment (8-in width) can be backwashed without stopping filtration in
the remainder of the filter. (See Figure 9-15). Backwash rates are similar to those for other
filters (up to 15 gpm/sq ft), but the backwash time for individual compartments is as low as
45 seconds. Hence, each compartment can be backwashed every 1 to 2 hours without ex-
cessive washwater use.
Each compartment has its own underdrain section. Media is supported on 1-in. porous
9-48
-------
H EAD TANK (OPT) *
FEED
ISLUDGE]
FILTER INLET
|FILTRAfE
[FILTRATE I
AIRLIFT TUBE
FIGURE 9-13
SIMATKR ULTl R
(Courtesy ol DravoCorp.)
9-49
-------
FIGURE 9-14
HYDROMATION IN-DEPTH FILTER
9-50
-------
A. Influent line.
B. Influent ports.
C. Influent channel.
D. Compartmented filter bed.
E. Sectionalized under-drain.
F. Effluent and backwash ports.
G. Effluent channel.
H. Effluent discharge line.
I. Backwash valve.
J. Backwash pump assembly,
K. Washwater hood.
L Washwater pump assembly.
M. Washwater discharge pipe.
N. Washwater trough.
O. Washwater discharge.
P. Mechanism drive motor.
O. Backwash support retaining springs.
R. Pressure control springs.
S. Control instrumentation.
T. Traveling backwash mechanism.
FIGURE 9-15
HARDINGE AUTOMATIC BACKWASH FILTER
(Courtesy Koppers Co., Inc.)
-------
plates over the underdrain. Flow from the underdrain sections discharges through individual
ports to a common effluent channel.
The travelling backwash consists of a rolling bridge carrying two pumps and equipped with
a hood extending over the length of a single compartment. The backwash pump draws wa-
ter from the effluent chamber and discharges it into the underdrain section for the com-
partment where the bridge is stationed. The wash water pump withdraws backwash flow
from the hood positioned over the compartment and discharges it to waste. Initiation of a
backwash cycle is controlled either by timer or by headless sensors.
Lynam (56) reported 68 percent removal of SS in uncoagulated activated sludge effluent by
Hardinge filters at an SS loading of 0.5 lb/ sq ft/ day and 11.5 inches headless. At 4.4 inches
of headless the removal at 0.4 Ib/sq ft/day was 75 percent. At 11.5 in. headloss the max-
imum hydraulic loading was 6.0 gpm/sq ft compared with 2.5 gpm/sq ft at 4.4 in. In the
same study, coagulation with alum did not improve performance.
9.10.3 Filter with On-Line Surface Scouring
Hydro-Clear Corporation offers a fully automatic, shallow bed, fine-media sand filter for
tertiary wastewater treatment. The media consists of 10 inches of 0.45 mm sand with a uni-
formity coefficient of 1.5 supported on a wire mesh above the underdrain system. This filter
combines air mixing of the water above the bed with air surging upward through the bed to
prolong run length. Influent flow splitting controls the flow to parallel units.
Typically, a filter run consists of a preset number of filter cycles. Each cycle begins by filter-
ing secondary effluent until a preset headloss is developed. Air mixing is then started in the
liquid above the bed to resuspend the solids collected on the media surface. After additional
headloss buildup, air trapped in the vented underdrain system is forced upward through the
bed for a short period. Solids removed by the air are resuspended by the air mixing, and the
cycle begins again. After the predetermined number of cycles, the filter is backwashed, en-
ding the run.
Data from Clark County, Ohio, indicate an average filtrate SS concentration of 4.8 mg/ 1
using effluent from a 0.2 mgd contact stabilization plant at 1.2 gpm/sq ft (57).
9.11 Slow Sand Filters
Slow sand filters consist of a layer of sand supported on graded gravel with an underdrain
system but no backwash system. The depth of the sand layer ranges up to 42 inches, and the
effective size is 0.25 to 0.35 mm with a uniformity coefficient of 2 to 3 (53). Secondary
effluent is applied, generally at a rate of about 3 gph/sq ft (8), and the filter is used until the
wastewater rises to the top of the filter wall. The filter is then removed from service,
drained, permitted to dry and then cleaned by manually removing the filtered solids.
Truesdale and Birkbeck (58) report only 60 percent SS removal for slow sand filters and a
cleaning frequency of once or twice per month. Rapid clogging of slow sand filters has been
9-52
-------
observed (59). Slow sand filters require large land areas and therefore, are not normally em-
ployed for large installations. Sand that is lost during cleaning must eventually be replaced.
Filters of the same construction, operated intermittently, have been used as combined phys-
ical-biological treatment for secondary effluent polishing. Intermittent operation permits
aerobic digestion of solids reducing somewhat the required frequency of maintenance. Area
requirements are still quite large, however, and generally limit applications to small
plants. The fact that maintenance is only required on an intermittent basis makes this type
of filter a viable process for upgrading existing lagoons which cannot meet effluent stan-
dards. Further discussion of this application can be found in the U.S. EPA manual, Up-
grading Existing Wastewater Treatment Plants and elsewhere (60).
9-53
-------
9.12 References
1. Baumann, R.E., and Cleasby, J. L., Design oj Filters for Advanced Wastewater Treat-
ment, Engineering Research Institute, Iowa State University, Ames, Iowa (October
1973).
2. Tchobanoglous, G. and Eliassen R., Filtration oj Treated Sewage Effluent. JSED,
ASCE, 96, 243 (April 1970).
3. Ives, Kenneth J. and Sholji, Ihsan, Research on Variables Affecting Filtration, Jour.
SED, ASCE, 91, 1 (Aug. 1965).
4. Hsiung, Kou-ying and Cleasby, J. L., Prediction of Filter Performance. Jour. SED,
ASCE, 91, 1 (June, 1965).
5. Ives, Kenneth J., Filtration: The Significance of Theory. Journal of The Institution of
Water Engineers. 25, 13 (Feb. 1971).
6. Tchobanoglous, G., Filtration Techniques in Tertiary Treatment. Journal WPCF, 42,
4, 604 (April, 1970).
7. Baumann, E. R., Design of Filters for Advanced Wastewater Treatment. Presented at
the Technology Transfer Design Seminar sponsored by the U.S. EPA, at Syracuse,
N.Y. (May 31, June 1, 2, 1972).
8. Culp, R. L., and Gulp, G. L., Advanced Wastewater Treatment, Van Nos-
trand-Reinhold Co., New York (1971).
9. Baumann, E. R., and Oulman, C.S., Sand and Diatomite Filtration Practice. Water
Quality Improvement by Physical and Chemical Processes, University of Texas Press,
Austin, Texas (1970).
10. Hudson, H.E. Jr., Declining Rate Filtration, Journal AWWA, 51, 1455 (Nov., 1959).
11. Convery, J. J., Solids Removal Processes. Nutrient Removal and Advanced Waste
Treatment Symposium. Presented by Federal Water Pollution Control Adminis-
tration, Cincinnati, Ohio (April 29-30, 1969).
12. Naylor, A. E., Evans, S. C., and Dunscombe, K. M., Recent Developments on the Ra-
pid Sand Filters at Luton. Water Poll. Control Jour. (Brit.) 66, 309 (1967).
13. Zenz, D. R., Weingarden, M. J., and Bogusch, E. D., Hanover Park Experimental
Bay Project (March 8, 1972).
14. Zenz, D. R., Lue-Hing, C., and Obayashi, A., Preliminary Report on Hanover Park
9-54
-------
Bay Project, U.S. EPA Grant #WPRD 92-01-68 (R2) (November, 1972).
15. University of Michigan short course, January 25-26, 1973. Reported by Thomas Hoo-
gerhyde, Michigan Department of Health.
16. Private Communication with J. Wiley Finney, Jr., Treatment results Hite Creek Ter-
tiary Plant, Louisville, Kentucky (April, 1973).
17. Ventura, California, East Side STP Test Report, Technical Bulletin, Dravo Corp., Wa-
ter and waste Treatment Division, Pittsburgh, Pa.
18. Wood, R., Smith, W. S. and Murray, J. K., An Investigation Into Upward Flow
Filtration, Water Pollution Control (British) 67: 421-426. (1968).
19. Private communication with H. M. Mueller, Jr., Neptune Microfloc (April, 1973).
20. Ripley, P. G., and Lamb, G. L., Filtration of Effluent from a Biological-Chemical
System, Water and Sewage Works, 12, 67 (February, 1973).
21. Performance Data Contained in Hydroclear Corporation Catalogue, Avon Lake,
Ohio, as tested by the Clark County Utilities Department, Springfield, Ohio (May,
1969).
22. Isaac, P. C. G. and Hibberd, R. L., The Use of Microstrainers and Sand Filters for
Tertiary Treatment, Water Research, Pergamon Press, 6, p. 465-474.
23. Guiver, K. and Huntingdon, R., A Scheme for Providing Industrial Water Supplies by
the Re-Use of Sewage Effluent, Water Pollution Control Journal, 70, 1, p. 75, 1971.
24. Michaelson, A. P., Under the Solids Limit at Ashton-Under-Lyne, Water Pollution
Control (Brit.), p. 533 (1971).
25. U.S. EPA Blue Plains Pilot Plant, Washington, D.C. Contract No. 6801-0161,
Monthly Reports (1972).
26. U.S. EPA Internal Monthly Reports, Piscataway, Md. (March-September, 1973).
27. U.S. EPA Internal Monthly Reports. Ely, Minn. (April-December, 1973).
28. Study of Upflow Filter for Tertiary Treatment. U.S. EPA Project No. 17030 DMA
(August, 1972).
29. Oliva, J. A., Department of Public Works, Nassau County, New York, Personal
Communication (March, 1973).
30. U.S. EPA, Advanced Wastewater Treatment As Practiced At South Tahoe. Project
9-55
-------
17010 ELQ (WPRD 52-01-67) (August, 1971).
31. Villiers, R. V., Berg, E. L., Brunner, C. N. and Masse, A. N., Municipal Waste-water
Treatment By Physical and Chemical Methods, Water and Sewage Works, R-62
(1971).
32. Bishop, D. F., O'Farrell, T. P., and Stamberg, J. B., Physical Chemical Treatment of
Municipal Waste-water. Presented before the 43rd Annual Meeting, WPCF, Boston,
Mass. (Oct., 1970)
33. Kreissl, J. F., Granular Media Filtration of Waste-water: An Assessment, Presented at
Seminar Filtration of Water and Wastewater, Ann Arbor, Mich. (Jan., 1973).
34. Kreissl, J. F., U.S. EPA NERC, Cincinnati, Ohio, Personal Communication.
35. Baumann, E. R. and Huang, J. C. Granular Filters for Tertiary Wastewater Treat-
ment. Accepted for publication, Journal of the Water Pollution Control Federation.
(Iowa State University, ERI 72051, Preprint) (February, 1972).
36. Cleasby, J. L. and Baumann, E. R. Selection of Sand Filtration Rates. JAWWA 54,
579 (May, 1962).
37. Misaka, Yasunao, et. al., Filtration of Activated Sludge Secondary Effluent Through
Sand and Anthracite-Sand Beds. The University of Wisconsin Water Resources Center
(1969).
38. O'Melia, C. R. and Stumm, W., Theory of Water Filtration. Jour. AWWA, 59, 1393,
(Nov. 1967).
39. Parker, D. S. et. al., Floe Breakup in Turbulent Flocculation Processes, J. San. Eng.
Div. ASCE 98, SA1, 79-99 (Feb. 1972).
40. Oakley, H. R. and Cripps, T. British Practice in the Tertiary Treatment of Waste-
water, JWPCF, 41, 36 (Jan. 1969).
41. Diaper, E. W. J. and Ives, J. J., Filtration Through Size-Graded Media. Jour. SED,
ASCE, 91, 89 (June, 1965).
42. Tebbutt, T.H.Y., An Investigation Into Tertiary Treatment By Rapid Filtration, Wa-
ter Research (Brit.), Pergamon Press, 5, p. 81 (1971).
43. Nebolsine, R., Poushine, I. and Fan, C. Y., Ultra High Rate Filtration of Activated
Sludge Plant Effluent, U.S. EPA Environmental Protection Technology Series,
EPA-R2-73-222 (April, 1973).
44. Fair, G., and Geyer, J., Water Supply and Waste Water Disposal, Chapter 24, John
9-56
-------
Wiley & Sons, Inc., New York (1954).
45. Water Quality and Treatment, American Water Works Association, Inc.,
McGraw-Hill, Inc., New York (1971).
46. Camp, T. R., Discussion of Anthracite Sand Filters. (Walter R. Conley), Jour.
AWWA, 53, 1478 (Dec. 1961).
47. Gulp, G. L. and Conley, W. R., High Rate Sedimentation and Filtration, Water Qual-
ity Improvement by Physical and Chemical Processes. University of Texas Press, Aus-
tin, Texas (1970).
48. Dravo Corporation, Effluent Polishing With Deep-Bed Filtration, Technical Bulletin,
7/WWT19.
49. Cleasby, J. L., Filter Rate Control Without Rate Controllers. Jour. AWWA, 61, 4,
181-185 (April, 1969).
50. Cleasby, J. L., Filtration, Chapter 4 in Physicochemical Processes for Water Quality
Control, Weber, W. J., Jr., Editor, Wiley-Interscience, New York, (1972).
51. Amirtharajah, A. and Cleasby, J. L., Predicting Expansion of Filters During Back-
washing, Jour. AWWA, 64, 1, 52-59 (Jan. 1972).
52. Cleasby, J. L., et. al., Optimum Backwash of Granular Filters. Engineering Research
Institute, Iowa State University, September 1973. (Presented at WPCF Conference
Cleveland, Ohio (Oct. 1973).
53. Water Treatment Plant Design, American Water Works Association, Inc., New York
(1969).
54. Kreissl, J. F., and Robeck, G. G., Multi-Media Filtration: Principles and Pilot Ex-
periments, Bulletin No. 57, School of Engineering and Architecture, University of
Kansas, Lawrence, Kansas (1967).
55. Summary Report, Advanced Waste Treatment. WP-20-AWTR-19, U.S. Dept. of the
Interior, FWPCA (1968).
56. Lynam, B., Ettelt, G., and McAloon, T. J., Tertiary Treatment at Metro Chicago by
Means of Rapid Sand Filters and Microstrainers, Jour. WPCF, 41, 247 (Feb. 1969).
57. Rogers, E., Clark County Utilities Department, Springfield, Ohio, Personal Commu-
nication (May, 1972).
9-57
-------
-58. Truesdale, G. A. and Birkbeck, A. E., Tertiary Treatment Processes Jor Sewage
Works Effluents. Water Poll. Control Jour. (Brit.) 66, 371 (1967).
59. New England Interstate Water Pollution Control Commission, A Study of Small,
Complete Mixing, Extended Aeration Activated Sludge Plants in Massachusetts,
(1961).
60. Marshall, G. R., and Middlebrooks, E. J., Intermittent Sand Filtration To Upgrade
Existing Wastewater Treatment Facilities, Utah Water Research Laboratory, PRJEW
115-2 Utah State University, Logan, Utah (February 1974).
9-58
-------
CHAPTER 10
COST ESTIMATES .
10.1 Introduction
The cost curves included in this chapter are based on: 1) actual installations, 2) projections
from pilot studies and other literature, and 3) manufacturers' information. In general, larg-
er capacity units show economy in both capital and operating costs. Since the added econo-
my of scale for plants greater than 100 mgd size is small, unit figures for this flow (or the
total area required at this flow) can be applied to larger plants. Costs for plants smaller
than 1 mgd vary too widely to permit effective use of general curves.
Cost calculations were based on an EPA-STP Index of 175 (July 1972 for U.S. Average).
The procedure for adjustment of costs to another cost index is outlined in Section 10.5.
10.2 Curve Content
The curves shown include all equipment and controls necessary for a working unit process.
Construction is assumed to include excavation and backfill in good soil on a level site. In
general cost curves do not include:
1. Buildings
2. Land
3. Pumping between processes
4. Sludge disposal
5. Yard piping
6. Special site conditions requiring pile foundations, rock excavation, etc.
7. Chemical feed equipment (given as separate curve)
8. Automated control (except as noted)
9. Engineering, Legal and Fiscal Costs.
10.3 Operation and Maintenance Costs
These are presented as curves or as a percentage of total capital costs and include normal
repairs expected during operation of units, but not breakdowns resulting from mis-
application or overloading. Chemical costs are not included in operating costs of chemical
feed systems, and should be allowed for separately based on actual cost and dosage. O&M
costs include power for normal operation but do not include external power costs for pump-
ing between units.
EPA Regulations, Title 40, Chapter 1, Part 35, Appendix A (Federal Register, 38, No. 174,
September, 1973), specify the useful life of various structures and equipment items to be ap-
plied in cost-effectiveness analysis. The regulations also specify use of 7 percent annual in-
terest in cost comparisons. In general, structural items have lives of 30 to 50 years, process
equipment 15 to 30 years, auxiliary equipment 10 to 15 years and electrical equipment 8 to
10 years. More specific information than given in the regulations may, in some cases, be ob-
tained from product manufacturers.
10-1
-------
10.4, Freight
Costs include freight allowances for equipment based on typical distances in the mid-west
and eastern parts of the country. For western installations costs may be up to 2 percent
higher due to extra shipping costs.
10.5 How to Use Cost Curves
1. Select capacity of unit(s) based on plant design capacity plus desired standby
capacity in unit(s).
2. Enter curve with capacity of unit(s) and read total cost.
3. Correct unit cost to current local cost index. EPA publishes current treatment
plant construction cost indices for 20 cities in the U.S. The current cost index
should be corrected for time and geographic location. After selecting the proj-
ected cost index use the following equation to compute corrected costs:
Corrected Total Cost =
($1000)
[Current Cost Index 1 x [Total Cost!
175 J L ($1000) J
4. The costs arrived at in Step 3 should be modified for the following items wherever
applicable:
a. Special instrumentation for automation or computer control
b. Special site preparation such as pile foundation, rock excavation, housing and
landscaping
c. Architectural requirements
10.6 Curve Description
The curves are primarily intended for preliminary cost comparisons between processes at
any flow or equipment size in the range covered. For actual designs a detailed specific cost
analysis should be made. Further data on costs for suspended solids removal processes are
available (1) (2) (3) (4) (5) (6).
10.6.1 Flocculators—Flash Mixers
Figure 10-1 presents capital and O&M cost estimates for flocculators and flash mixers.
10-2
-------
100
o
O
O
Q
UJ
(O
O
I
10 ,_
. FLO CGULATORS
CAPITAL COST FLOCCULATORS
& M. FLASH MIXER
CAP.ITAL COST
I FLASH MIXERS
j- r _...-.-. . T T .. t rr rr T n i n I
O.I
_ cr
< o.
7 CO
O
(T
Ul
Q.
O
O-OOI
10
CAPACITY MGD
.FIGURE 10-1
FLQCCULATORS—FLASH MIXERS
'COSTS ADJUSTED TO EPA—STP INDEX 175
100
10-3
-------
Capital costs ($1000) and O&M costs (
-------
1000-
0eooJ
o
0
-W- 900
INSTALLED
o 9
p p
i-
co
o
o
_i 80~
H
H
20-
lO
t
'
f
}•
rf>
S
•*
**
iiiSljiUM
EJ I
S
" — ; 'l
,
- ? '
— — ^ pi!. -
^ ™ _ . i !.......
._ _ — •* — -•• • • - 1 • i
;»•' = I
^ *
^ ^
^, 3*
2 3
iilimmHiliH:^
LEGEND : CAPJT)
= =fci= 3=
\L CO
Q ALUM — 200 mg/l
OR IRON — 100 mg/l
i ^ LIME — 300 mg/ 1
• Q POLYMER— 1.0 mg/l
i, ~~jT~ T j
'"> I 1
" .. .. (.\ .
1 " ^~ - "2 ( * •* ' ' •" 1
i|EEiEEEi;:||;;j j —
O fi M 1 FDR A LI J
v/iVi i»i. \ i vyr\ <™« ^w
^
r/
/
r
1
>
i
M
II
/
^
L
*
___..,
^<...
>
M==*
^ — I — 1— _ B
ONC
456789 2
10
— s -. i — --,
-- •---:;'? j'!
_ , ! __, i . .
l'' '
>' I1'
' i'''t I
i ! ii_. ._
)R POLYMER)
345
IMplI-10
§ ::::: :!::£: :3
:: ::::: ::: :: X
.. ','.. ..... ,E
•ft
-•• |i ••}
_....,
>\ 1
l;PJ::l-">
r:::li;:: i: >
'•'-'•'• ! ;: ?
i;;;|;;;;!!!
f
6 7 8 9 ' w
100
UJ
o
UJ _l
I- <
ui
o
oc
UJ
o.
o
CAPACITY MGD
.FIGURE 10-2
CHHMICAL FEED SYSTEMS
COSTS ADJUSTED TO EPA—STP INDEX 175
10-5
-------
For different chemical dosages calculate cost based on an equivalent plant flow calculated
as follows:
equivalent plant flow.= , I design I
[ flowj
expected dosage
dosage used to
develop cost curve
Each system includes: 1) a minimum of two volumetric or gravimetric automatic propor-
tioning feeders sized to provide 50 percent excess feed capacity; 2) pumps to deliver chem-
ical feed solutions to the process; and 3) 30-days bulk storage. Prices cover installed equip-
ment suitably corrosion-protected for the intended chemical service. Not included in capital
costs are buildings (except for dry chemical storage space), land, sludge disposal and ex-
ternal piping.
O&M costs include manpower for operation and normal maintenance, and power costs for
pumping, but do not include chemical costs.
10.6.3 Sedimentation Basins
Figure 10-3 presents estimates of capital and O&M costs for sedimentation basins.
Capital costs ($1000) and O&M costs ($1000/yr) are given for installations requiring total
basin surface areas of 1000 to 100,000 sq ft. Costs are based upon installations using two or
more units.
Included in capital costs are inlet appurtenances and sludge-collecting mechanisms circular
or rectangular tanks (steel or concrete), skimmers, scrapers, supports and walkways, and
sludge draw-off, all completely installed. Curves are applicable for end-feed, center-feed or
peripheral-feed designs for primary or secondary treatment applications.
O&M costs include manpower for operation and normal maintenance but do not include
chemical or pumping costs.
Not included in capital costs are land, buildings, covers, chemical feed equipment, pump-
ing, sludge pumping and disposal and external yard piping. Instrumentation is limited to
automatic sludge blowdown valves and lines.
10.6.4 Solids Contact
Figure 10.4 presents estimates of capital and O&M costs for solids-contact units.
Capital costs ($1000) and O&M costs ($1000/yr) are given for installations requiring total
basin surface area of 1000 to 100,000 sq ft. A minimum of two operating units was used to
estimate costs.
10-6
-------
1000-
O
o
o
(0
O
o
0.
<
0
100-
10-
9J
^
__ ""_. -PA Dl "
i V^Ar 1
( i '
( * '
0 !
„*"
?
r A l PflCT i ii
1 ML. \j vO 1 v
::::::::::: = = ::::;;*
^ ^
^
, "
, »'
....... ^
... _ f ...
.. ., *_ 5 _ . _ _
.1" ^ , n
,.' S
^MM-|Wc
— 5
r
^
t
*
..T.. j
._..._ 4! . -^ —
-_ ,8
t '
^ *
£ '
^
a M.!EEE;:!!!!EEE!
ro
.
1 j<-- -too
....... ^ ^ .. Ill
-L C^
i
! ; "!E ::gr ~ " ~" "i "~ ^t
• ^ ^ ^J
. ::... ...( 2 :. f^ u?
::S!?::::: _" = "| — UJ
:::::::::: :::::: LU
Q.
<^*>
iiiiimiiiiiii-miriu 8
- •- J- £•• Q
^? "
^^
........... ------- Q jfr
_ QL
EEEEEE;EEEE-EEEEEE o
| 2 3456789) 2 34567891 1-w
1 10 100
^SURFACE AREA 1000 SQ. FT.
FIGURE 10-3
SEDIMENTATION
COSTS ADJUSTED TO EPA—STP INDEX 175
10-7
-------
10
• I»I»B =========== =====sz=:rz::i::::::=s===s=::: ::::::;::
1000
o
o
o
40-
STALLED
t-
V)
o
O
5678 9
ioo
SURFACE AREA, 1000 SO. FT
FIGURE 10-4
SOLIDS CONTACT
COSTS ADJUSTED TO EPA—STP INDEX '75
10-8
-------
Included in the capital costs are concrete or steel tanks and slabs, turbine recirculators,
sludge scrapers, skimmers, inlet and outlet distributors, supports and walkways, sludge
drawoff, internal baffles, piping and accessories, all fully installed. Prices are applicable for
upflow-type solids-contact units with integral flash mixing and flocculating provisions.
O&M costs include manpower for operation, turbine mixer power, and normal mainte-
nance but do not include chemical or pumping power costs.
Not included in capital costs are buildings, land, chemical-feed equipment, external pump-
ing, sludge disposal and external yard piping. No flow, turbidity, conductivity, or other as-
sociated instrumentation is included.
10.6.5 Flotation
Figure 10-5 presents estimates of capital and O&M costs for dissolved-air flotation pro-
cesses.
Capital costs ($1000) and O&M costs ($10007yr) are given for installations requiring total
basin surface areas of 300 to 50,000 sq ft.
Included in the capital cost are all tanks and internals, air-pressurizing equipment, rec-
ycie-pumping equipment, operating valves and piping, all fully installed. No special corro-
sion protection is included except for normally-painted items.
Assumed O&M costs were 3 percent of capital costs including manpower for operation and
normal maintenance, power for normal pumping and air pressurization, but not including
chemicals.
Not included in capital costs are buildings, land, chemical feed equipment, sludge disposal
and external yard piping. Instrumentation is limited to pressure-sensing controls for normal
operation of the units.
10.6.6 Settling Tubes and Wire Septums
Figure 10-6 presents estimates of capital costs for tube settlers and wire septums.
Capital costs are given as $1000 for total required screen areas of 10 to 2000 sq ft. The esti-
mated operating labor requirement is 0.5 man-hrs/day mgd.
O&M costs for tube settlers can be estimated at 2 man-hr per basin per week. O&M costs
for wire septums can be estimated at 1 man-hr per basin per day.
Included in the capital costs of tube settlers are plastic tubes with 60° inclination and 21 in.
deep plus steel supports and additional effluent collector weirs. Wire septum costs include
stainless steel wires, all fully installed.
10-9
-------
cuyuuu
10
O*
•
O"
T,
r>
° 6
-«*- 9
4
0
UJ ,
-1
?
CO
z •
H 1000
CO 9
Oa .
•
07.
6:
< 5
H 4
0.
<-m.
O
2
IOO!
• *
ft
g •
*
•
0
9
,,9
-,ii!
.'
a M
g^
4
P*'*
^
^ ^
i
0 '
/
t -- j
; i —
\ •
i '
--s^--j •*
' '
(-,<- ---
.
_ ^ E - - . ..... U
_... —
!lE!|!:l===E|ii
:»;;;;;;; ==5«!;i:=
! —
j
----- ^
I
... __!.,.-
j
j -
g i '
2 ' ^
J!!::i!;iiii^|C/J
/,
F
^
/
>l
,i
_ ,,!:..
4\
. '
^
., '_. !
r T
'^- I' '
TAL COS
j ::::: :±i 2OO
. ^B. . .
T j|
, w
j *
too
:::::::::: Ul
]:::::":" o
i^E|=g; z
J:|5i^ <
i^»?;;;;: u
^l""f * <
jiffiMffl « ^
— ^* >•
mm a«:
. O Q^
z
....:..:.. ^ o
10 o
iiiiijiniiii o
— z "~
".'.".'.' ".'".' O '^*'
;T EEEEEEE; H
)[[-[ J L J (E
UJ
::::::::::: Q.
::i::i:::::: Q
I.O
I 4 967091
0.3 1.0
4967 Vo'o
3 4 is.'o
SURFACE AREA, 1000 SQ. FT.
FIGURE 10-5
FLOTATION
COSTS ADJUSTED TO EPA—STP INDEX 175
10-10
-------
o
o
o
o
UJ
O)
CO
o
0.
<
O
10,000 m
9 •
» :
4 •
» ::
2 •
1000
9. ..
0
6 •
• •
4 ••
••
2 <•
100-
9 '•
, ,!
1 .
5 ::
* I!
». ..
2 •
10- .
J
w
1
it
\ ~
'
^
, *
>
4
jjj
SC
-
r
t
^
-^ -*
\
H
^
'
/
X
WIRE SEPTU
1
I1
~ , - . . ,t . . . -
. ! .4
,! .,l.
---?- — ,>'
A - *' 1 '
> ' t> I1
J ' I1
* ('
,.2- -_.<('..
x^ ..'
i'
J r
— r
r
SETTLING Tl
(CIRCULAR B
^
^^
j
MS l^ljj,^::;
I"t LU»ff1 PM Lkf J4^
_--.(t....T| ? 3
~a'" T4" " '^*! ' '"
/ i *
— -;z.!-|-?Z---^
^ . - r i —
-£-- — "jiff"
?*---& JJ-J
t — --- i
\
BE
AS
^
i
7
(i
c
II
/
/
5
4
/
/
S
/
(^
^
). S
, -. r .... -
i:^..s
. . _ .1 _
" • - j * • •---• —
- _ - . !
~t ^
--3"- !•'•-
3 1
2 «' /
>' i1
__.^ .(!.._
/ •'
-3 -tl -
-^- -
_,2 _
9
*
SETTLING TUBE
RECTANGULAR B
A
-»_-. .g!..
-- |!
_.j: _^
i:__. _2 _
i 4
t . *
? . i
:;:: ::j !:::::: *-q
(-_-.. , i . _
--^ i---
!--..,! 3
i
, '
- - -^- J~'
'S —
AS INS)
" '
. *
- I -
jj _ _
r
f
f
^
;'
•J 6789^
0.5 1.0
2 3 456789
2 3 45678
10
100
INSTALLED AREA, 1000 SQ.FT.
FIGURE 10-6
WIRE SEPTUMS AND SETTLING TUBES
COSTS AUJUS1 tD TO EPA—STP INDEX 175
10-11
-------
Prices for wire septums are applicable to circular or rectangular designs. Prices for tube set-
tlers are given separately for circular and rectangular tank designs. No special corrosion pro-
tection is included except for normally-painted items.
Not included in capital costs are buildings, tanks, cleaning devices (air grids), sludge dis-
posal and external yard piping.
10.6.7 Wedge-Wire Screens
Figure 10-7 presents estimates of capital costs for wedge-wire screens. O&M costs are cal-
culated as described below.
Capital costs are given as $1000 for total required screen areas of 10 to 2000 sq ft. The esti-
mated operating labor requirements is 0.5 man-hrs/ day/ mgd.
Included in the capital cost are screens and screen supports. All screens are stainless steel
construction, but supports, baffles and distributors are steel. Rotating screens include mo-
tor drives. Prices are applicable for rotating screens and stationary screens. No special cor-
rosion protection is included except for normally-painted items.
Not included in capital costs are land, buildings, pumping, sludge handling and external
yard piping.
10.6.8 Microscreens
Figure 10-8 presents estimates of capital and O&M costs for microscreen equipment.
Capital costs ($1000) and O&M costs ($10007yr) are given for installations requiring total
screen areas of 100 to 10,000 sq ft. One unit for 1 to 2 mgd flow, two units for 3 to 4 mgd,
and three units for flows of more than 5 mgd were used to estimate costs.
Included in the capital costs are tanks, drums, screens, backwash equipment, drive motors
and all accessories for automatic operation, all fully installed. Prices are applicable for con-
crete or steel tank construction. No special corrosion protection is included except for nor-
mally-painted items.
O&M costs include operation, normal maintenance, and power for rotation and for
spray-water pumping, but do not include chemicals or external pumping power costs.
Not included in capital costs are buildings, land, pumping (except spray system), sludge dis-
posal and external yard piping. Instrumentation is limited to automatic valves and time
cycle or pressure sensing backwash control suitably panel-mounted.
10-12
-------
IOOO-"
8
7
6
6
4
3
2
0
0
o
•«*- 100 -
9
8
0 7
W 6
-I 6
J 6
» 4
CO
z 8
CAPITAL COST 1
o
at O -a os to 1 10
4
3
2
1.0
l(
=j
<-
?
T
I
" ... ( ROT
— "~:
- r
T 2
.
t '
, '
^ '"/
j. .._..., _j__
j )
_^_ _. .....
^ . .... _ _
2 3
3
-\
ATING \--~,~~-'-'.~
'-.---.~::-lt\-: — - = -*:': =
~ ' ~ '[ ~ £ ~~
1 .-•-•/
g '
^
4 '
j n
<
?
^
•1 — HI niiriiiiii 1 1 1
_.:"::; •;::::::: ::^
^
r
i '
1 1
__ (i!..., ...
p j i __*
1 j ~
P
. *
A ...-1..
4 «'
a - - - -
/ --
\- --"
= -t= = ioTAT OKI
-w 1 " 1 1 w IN
4B6789 2 3
100
:::::::::: :_::::::g
__ ... f
_.,_... .. _ — a''-
-,.-.. -t
_ ... j ...... ~ j---..
(... - -a - --
^fiiiiiiiiiKiiriiin
. *
^ *
. t
A R V -H-il -\-\4 444 'M
v
— j
^
--:j
4 6 6789.
1000
TH
^ '
t
- t -
*--
1
2000
SCREEN AREA, SQ. FT.
FIGURE 10-7
WEDGE WIRE SCREENS: ROTATING AND STATIONARY
COSTS ADJUSTED TO EPA—STP INDEX 175
10-13
-------
f
3
2
1000-
O 9
O 6
O 7
•w-
5
s «
_J
_J 3
<
»-
(/) 2
z
h-
C/>
O 100
» a9
<
1- 5
Q- 4
<
°
2
10 -
O.I
£_
x
(*"
/
4
e.
_ . .X
T
T
. I
- ±
_. -- g*
t
"f. "~
^l-z-----
i
t *
^ C +
^ . .
2
_
.
::::::: = — :::!£:
+-
CAPITAL
('
j
S
... ..j* .
t
t '
: i'
)
j" 3
„.,!:.. ..
,i ' ..
3 4
COS
2
x
^
.i'
i
_2
2
,'
.,,.
T
/
Z ..
_ 1 _
2
5 6 7 8 <
L
/
^
y
/
<
/
,
^1
?'
"
! '
(j'
(?__
. - . -
-(' -
-v -
2
!_ _
._-\. ._ i, ..
-•^oa M
rmmr
— f —
T
t
) 2 3
1.0
EEEE;^:|JEEEEE
^
- --->-
x**
,,•'
/
— , i _
. '
,j
> (
?
f
__,_._._..!... ___
.1
<'
4 5678
:::^~
.
. . . ^ _
^ '
<*
/
2
^
<_
/
9
10
»'
v
i 100
--• 10
1 0
•
20
UJ
O
z
<
z
UJ
o:
<
UJ
cr
UJ
Q.
o
o
o
or
UJ
a.
o
SCREEN AREA, 1000 SQ. FT.
FIGURE 10-8
MICROSCREENS
COSTS ADJUSTED TO EPA—STP INDEX 175
10-14
-------
O
O
o
o
UJ
O
o
a.
<
o
4 1 1 Mil
2
10,000
9
5
4
2
. j '
( i
1000 -::::
9
Q
6 --• j
5
4
2 £--•
0 1
1 •
100
fffflr0 a
,!
•;,i'-
i'
J:: = = = = :&
... _(
,!
... « - ...
,i I
2 34
M;; =
. ,!. -
1 ' ....
ftpm
::::;i-
\
f.
" ~ *
0
L (
5 6 7 e
- 1 n
., R
X)S
2
^
T
^
^
^
?
.1 '! _,
>i
S ._ _
f
1:: j!!::::: =
jl-_, _
9 23
::::::: :::;
,*'-.
t*
,,i'
. - l" .
I " .
- '
^
.s
,> ^
F
i
, !
• •
2
«L_
-. 2
i •
^
?<
><
.
e.
Tl
^l
ILU
"" ~ ^ ?
-^ ---
4 S 6 7 S 9
300
UJ
o
z
z
UJ
I-
z
100 <
' *
o
<
cr
<
UJ
o
o
or
UJ
a.
o
1.0
1.0
10
SURFACE AREA, 1000 SQ. FT.
FIGURE 10-9
MEDIA FILTERS
COSTS ADJUSTED TO EPA—STP INDEX 175
10-15
-------
10.6.9 Media Filters
Figure 10-9 presents estimates of capital and O&M costs for filtration equipment.
Capital costs ($1000) and O&M costs ($1000/yr) are given for filter installations with total
required surface areas of 200 to 10,000 sq ft. A minimum of three operating units and filter
bed depths of 4 to 6 ft. with sand and/or coal media were used to estimate costs.
Included in the capital cost are filter tanks, internals, media, operating valves and piping
and automatic backwash controls, all fully installed. The curve shown is an average curve
for upflow or downflow, gravity or pressure (up to 60 psig) designs of either concrete or
steel construction. Pressure filters are usually less expensive than gravity units below 3 to 6
mgd, but are considerably more expensive at larger flows.
10.7 References
1. Smith, R., Cost of Conventional and Advanced Treatment of Wastewater, Jour.
WPCF Vol. 41, pg. 1546 (1968).
*
2. Smith, R., and McMichael, W. F., Cost and Performance Estimates for Tertiary
Wastewater Treating Processes, USDI, FWPCA Report No. TWRC-9 (June 1969).
3. Evans, D. R., and Wilson, J. C., Actual Capital and Operating Costs for Advanced
Waste Treatment, paper presented at WPCF 43rd Annual Conference, Boston, Mass.
(Oct. 1970).
4. Black & Veatch Engineers, Estimating Costs and Manpower Requirements for Con-
ventional Wastewater Treatment Facilities, prepared for U.S. EPA, Project No. 17090
DAN (Oct. 1971).
5. Lynam, B., Ettelt, G., and McAloon, T., Tertiary Treatment at Metro Chicago by
Means of Rapid Sand Filtration and Microstrainers, Jour. WPCF, Vol. 41 pg. 247
(Feb. 1969).
6. Inder Jit Kumar, Clesceri, N. L. Phosphorus Removal from Wastewaters: A Cost
Analysis, Water and Sewage Works, Vol. 120, pg. 82 (March, 1973)
7. Gulp, R. L., and Culp, G. L., Advanced Wastewater Treatment, Van Nostrand Rein-
hold Company, New York (1971).
8. U.S. EPA, Advanced Wastewater Treatment As Practical At South Tahoe, Proj. No.
171010 ELQ08/71 (August 1971).
9. Sewage Treatment Plant Design, ASCE Manual of Practice 36 (WPCF M.O.P. 8),
New York (1972).
U S EPA Headquarters Library
Mail code 3404T
1200 Pennsylvania Avenue NW
Washington, DC 20460 10-16
202-566-0556
-------
10. Diaper, E. J., Personal Communication, Crane.Company, Cochrane Division, King of
Prussia, Pa. (Nov. 1972).
11. Ginaven, M. E., Personal Communication, The Bauer Bros. Co., Springfield, Ohio
(April, 1973).
12. Carpenter, D. A., Personal Communication, Komline-Sanderson Engineering Corp.
Peapack, N.J. (April, 1973).
13. Johnson, R., Personal Communication, E1MCO, Envirotech Corp., Salt Lake City,
Utah (April, 1973).
14. Dvorin, R., Personal Communication, Graver Water Conditioning Company, Union,
N.J. (April, 1973).
15. Diaper, E.W.J. Tertiary Treatment by Micro-Straining, Water and Sewage Works,
Vol.-116, pg. 202 (June 1969).
16. Eilers Richard, G., Condensed one page cost estimates for wastewater treatment U.S.
EPA, NERC, Cincinnati, Ohio (Nov. 1970).
17. Graham, A., Personal Communication, Neptune Microfloc, Inc., Corvallis, Oregon
(April, 1973).
10-17
------- |