United States
Environmental Protection
Office of Research and
Washington DC 20460
September 1993
Combined Sewer Overflow

                                            September 1993

  Combined Sewer Overflow
   U.S. Environmental Protection Agency

    Office of Research and Development
Center for Environmental Research Information
            Cincinnati, Ohio
                                   Printed on Recycled Paper

This document has been reviewed  in accordance with  the U.S. Environmental Protection Agency's peer and
administrative review policies and approved for publication. Mention of trade names or commercial products does
not constitute endorsement or recommendation for use.

Chapter 1  Introduction	  1

           Purpose	  1
           Scope and Organization	  1

 Chapter 2  Introduction to CSO Control Technologies	  3

           Unique Aspects of CSOs	  3
           Major Technologies To Be Addressed	  5
           Other Control Methods	  6
           References	  11

Chapter 3  Process Selection	  13

           Performance Goals	  13
           Data Requirements for Design of CSO Controls	  22
           Technology Selection	  30
           References	  34

Chapter 4  CSO Control Technologies	  37

           In-System Controls/ln-Line Storage	  37
           Off-line Near-Surface Storage/Sedimentation	  43
           Deep Tunnel Storage	  55
           Coarse Screening	  65
           Swirl/Vortex Technologies	  69
           Disinfection	  81
           References	  89

Chapter 5  Costs for CSO Control  Technologies	  93

           Sources of Cost Data	  93
           Construction Costs for CSO Controls	  93
           O&M Costs for CSO Controls	  94
           References	  95

                                            List of Figures
Figure                                                                                          Page

3-1   Approximate relationship between performance of CSO storage and storm size used
      as design basis—percent capture	  15
3-2   Approximate relationship between performance of CSO storage and storm size used
      as design basis—number of overflows	  16
3-3   Relationship between percent capture and number of overflows for CSO storage control	  17
3-4   Illustration of use of monitoring data to characterize first flush	  21
3-5   Illustration of knee-of-the-curve analysis	  22
3-6   Probability distribution of storm event volumes	  26
3-7   Illustration of intensity-duration curve	  27
3-8   CSO settling velocity—typical test equipment and sample results	  28
3-9   Effect of particle settling velocity and hydraulic loading rate on TSS removal efficiency	  31
3-10  Cost optimization of storage-treatment	  34
4-1   Example of a vortex valve	  38
4-2   Example of a wirbeldrossel and wirbelvalve	  40
4-3   Example of an inflatable dam	  41
4-4   Example of a motor-operated gate regulator	  41
4-5   Example of an elastomeric tide gate	  42
4-6   Flow schematic for typical storage/sedimentation facility, Newport, Rhode Island	  47
4-7   Plan and profile for a typical  rectangular storage/sedimentation facility	  49
4-8   Examples of tangential  inlet configuration	  58
4-9   The E-15 dropshaft	  59
4-10  The D-4 dropshaft	  60
4-11  Cutting cycle for a typical tunnel boring machine	  62
4-12  Tunnel system dewatering pump station	  64
4-13  Catenary-type mechanically cleaned bar screen	  67
4-14  Example  EPA swirl concentrator	  70
4-15  Example  Fluidsep vortex separator	  71
4-16  Example  Storm King hydrodynamic separator	  72
4-17  Particle settling velocities for grit and organic material in still water	  73
4-18  Typical gradation for grit and organic material	  73
4-19  Settling velocity profiles of combined and sanitary wastewaters	  74
4-20  Head vs. discharge  per linear foot of weir length for a circular weir	  76
4-21  Common layouts for swirl/vortex installations	  77
4-22  Graphical representations of log Yt/Y0 versus log CT	  82
4-23  Relationship between GT and bacterial kill	  83
4-24  Schematic of typical liquid sodium hypochlorite system	  86
4-25  Diffuser types	  87
5-1   Construction costs for CSO controls	  94
5-2   O&M costs for CSO controls	  95

                                            List of Tables
Table                                                                                         Page

3-1   Reported Performance of CSO Treatment Technologies	 18
3-2   Summary of Peak Hourly Rainfall Characteristics	 20
3-3   Rainfall Characteristics for Six U.S. Locations	 27
4-1   Typical Overflow Rates for Primary Settling Tanks	 45
4-2   Typical Detention Times for Primary Settling Tanks	 45
4-3   Typical Rectangular Tank Dimensions for Storage/Sedimentation Facilities	 50
4-4   Examples of Floor Slopes for Storage/Sedimentation Facilities	 50
4-5   Typical Dimensions of Circular Storage and Storage/Sedimentation Tanks	 51
4-6   Advantages and Disadvantages of Deep Tunnel Excavation Methods	 61
4-7   Chicago TARP Mainstream Pumping Station Capacities	 65
4-8   Examples of Bar Screen Installations  at CSO Control Facilities	 69
4-9   Comparison of Design Details for Swirl/Vortex Devices	 79
4-10 Examples of Swirl/Vortex Chamber Construction	 79
4-11  Examples of Sodium Hypochlorite Disinfection Systems at CSO Control Facilities	 85
5-1   Cost Equations for CSO Control Technology	 94

This manual is the product of the efforts of many individuals. Gratitude goes to each person involved
in the preparation and review of the document.
Donald Walker, Jonathan Golden, and David Bingham of Metcalf & Eddy, Inc., Wakefield, MA, and
Eugene Driscoll of Hydroqual, Inc., Mahwah, NJ, were the principal authors of this manual.
Technical Contributors
The following individuals provided invaluable technical assistance during the development of this
          Richard Field, U.S. EPA, Office of Research and Development, Storm and Combined
          Sewer Program, Edison, NJ
          Michael Brown, Doctoral Candidate, Cranfield Institute of Technology, School of Water
          Sciences, Woldingham, Surrey, UK
          Joseph Mauro, U.S. EPA, Office of Water, Municipal Support Division, Washington, DC
Peer Reviewers
The following individuals peer reviewed this manual:
          Michael Domenica, Massachusetts Water Resources Authority,  Boston, MA
          Kenneth Pew, Northeast Ohio Regional Sewer District, Cleveland, OH
          Robert Pitt, University of Alabama - Birmingham
Editorial Reviewers and Document Production
Heidi Schultz, Eastern Research Group, Inc. (ERG), Lexington, MA, provided editorial review and
produced this manual.
Technical Direction and Coordination
Daniel Murray, U.S. EPA, Office of Research and Development, Center for Environmental Research
Information, Cincinnati,  OH,  coordinated the preparation of this manual  and  provided technical
direction throughout its development.

                                             Chapter 1
Overflows from combined sewers during storm events
result in the discharge to receiving waters of untreated
sanitary sewage, which also may contain pre-treated
industrial  wastewaters  and  untreated  stormwater.
Combined sewer overflows (CSOs) contain  pollutants
that  are  present  in the  domestic  and  industrial
wastewaters, as well  as those in the urban stormwater
runoff that enters the  combined sewer system.

In many cases,  these discharges have an adverse
effect on  receiving water quality and attainment of
designated uses.  In  addition, since CSOs  are  point
source discharges  of untreated sanitary sewage, they
require National Pollutant Discharge Elimination System
(NPDES) discharge permits. In recent years, there has
been an enhanced  regulatory focus on  CSOs and their
control, and communities with combined sewer systems
are being called upon to develop and implement programs
for control of CSOs.

Control  technology applicable  to  CSOs has  many
aspects  in  common  with  treatment of  domestic
wastewaters in publicly owned treatment works (POTWs).
However, a number of unique aspects of CSOs must be
considered both for overall control program design and
for effective design and operation of treatment units.


This manual provides information to assist in selecting
and designing control measures for reducing pollutant
discharges from CSOs. The manual will be useful for
municipal  public works staff, design  engineers, and
regulatory agency  staff tasked with the development
and review of facility  plans and long-term CSO control

Scope and Organization

The manual provides design  information for six CSO
control technologies  that were selected because they
represent  the  most  commonly  employed techniques
used in programs developed to date. Additional control
techniques are identified  and described, but detailed
design guidance for  these techniques is  beyond the
scope of this manual.
Chapter 2 discusses the characteristics of CSOs and
the special considerations these features impose on the
development of effective designs for control systems
and unit processes. It describes the technologies for
which design guidance is provided,  and the basis for
technology  selection. Finally, Chapter 2 identifies and
discusses other potential CSO control techniques that
are not addressed by the manual.
Chapter 3 presents process selection considerations, and
describes how the six control technologies for which
design details are provided relate to each other and to
potential control programs for an overall system. The
first section discusses a variety of alternative  design
goals that may be imposed by a regulatory agency or
otherwise adopted for design, and describes how to
convert  such general requirements to a specific design
for a control system or treatment  unit. Factors that
influence the performance of each  selected control
technology  are identified,  and  the  advantages and
disadvantages associated  with  each technology are

Combined sewer systems with their associated overflow
points are relatively complex. Hydraulic conditions are
highly variable  due  to the  intermittent and variable
characteristics of rainfall. The quality characteristics of
CSO  flows, and hence their treatability, also can vary
significantly from location to location and from storm to
storm at a  given location. As a result, effective design
programs  usually  require   application of  computer
models, appropriate analysis and interpretation of rain
gage records, and quality and treatability characterizations
of the CSOs. The second section  of Chapter Three
describes the data requirements for  addressing these
elements and for developing the information that forms
the basis for the design details presented in Chapter 4.

Chapter 4 provides design details, rules of thumb, and
examples  of   practice  for the   following   control

•  In-System Controls/ln-Line Storage
•  Off-line Near-Surface Storage/Sedimentation

•  Deep Tunnel Storage

• Coarse Screening                                   and/or operational  features from reported operating

• Swirl/Vortex Technologies                             units also are Provided'
  D's'nf  t'o                                           Chapter 5 summarizes currently available capita! and
                                                      operation  and  maintenance  (O&M) cost  information
For each of these technologies, Chapter 4 summarizes    relating to  the selected control technologies.  The
process  theory,   design   parameters,  operational    information provided is suitable for preliminary indications
requirements, and  important related aspects  (e.g.,    of the general order of costs associated with the control
solids  handling, odor control).  Examples of design    measures, and the  need for site-specific estimates for
                                                      developing accurate local costs.

                                            Chapter 2
                        Introduction to CSO Control Technologies
Unique Aspects of CSOs

The  design criteria  and performance capabilities for
most unit operations  at POTWs  generally are well
established, based on the existing wealth of knowledge
regarding  municipal  wastewater  flow  and   load
characteristics, and decades of operating experience at
thousands of treatment facilities. This  same municipal
sewage is an  important component of the overflows
from combined sewers during storms.  Although mixed
with  surface  stormwater runoff  in varying degrees
during CSO events, both  the quality and  treatability
characteristics of CSO discharges and municipal sewage
are relatively similar. Differences are  more in degree
than in kind. As a result, experience with the treatment
of municipal wastewater provides  an important and
useful basis for the design and  operation of CSO
treatment technologies.
A number of important differences must be recognized,
however, because they have a  significant influence on
the effort needed to develop effective designs for CSO
control units. The differences derive from several unique
aspects of CSOs; the influences they have on the
design of CSO control programs and individual control
units are discussed briefly below. Specific considerations
are addressed in greater detail  in other sections of the
manual, where appropriate.

Performance Goals Are Not Uniform

CSO control goals may be governed by receiving water
quality-based and/or technology-based  requirements,
depending on the policies of state and federal permitting
authorities. A  considerably  wider range of potential
performance requirements may be applied for a CSO
area  compared  to   the  standard  "secondary"  or
"advanced" treatment objectives that apply in the case
of most POTWs.  Chapter  3  provides examples  of
commonly encountered  CSO  performance  goals.  In
addition, while  the technology and design  basis for
providing  secondary treatment at a POTW is well
established, CSO control concepts and design details
are influenced  by the specific performance goal that is
applied. Additional data and system analysis is required,
in relation to POTW design, because the development
of controls that meet a selected performance goal must
consider the variability of CSO flows and the frequency
of extreme events.

System Characteristics Are Site Specific

The characteristics of combined sewer systems are site
specific, and  have  an  important  influence on  the
elements  of an overall control  program,  the control
technologies that  are most  appropriate, and design
features of individual control units.

Many CSO control programs include clever modifications
or adaptations of features of the existing combined
sewer  collection/conveyance  system.  For  example,
relatively minor piping changes or regulator modifications
can  significantly affect  design  of  end-of-pipe control
units. The design, maintenance,  and operation  of
regulators  can have an important effect on the CSO
flows delivered to a control technology.

Typically, a combination of management practices and
CSO control technologies  is required to meet the CSO
program goal for a given community. Simply applying a
control technology at a particular overflow point usually
is not appropriate; effective control programs require
consideration of the system as a whole.

An  important factor contributing to the complexity of
CSO control is  the existence  of multiple overflow
locations,  which may discharge to different receiving
water bodies. Multiple overflows may be consolidated
for treatment or they may be addressed separately. In
some cases, the most effective design approach may
completely eliminate a particular CSO that discharges
at a sensitive location, by means of piping system
modifications to reroute the affected flows.

Often a number of possible alternatives exist for a given
community. The design of cost-effective CSO control
programs  thus  may involve  modifications to  the
collection  system to relocate or consolidate overflows,
or  to  reduce overflows  by  expanding  hydraulic
conveyance  capacity  at  critical  locations.  Such
modifications will  influence the number and design
capacity of any end-of-pipe CSO treatment technologies
that may be required to meet performance goals. As a

result, site-specific features of a local combined sewer
system impose an additional level of design considerations
than the typical POTW design program.

Design Basis Selection Is Not Standardized

Rainfall, which produces flows in the combined sewer
system, is the factor that determines when and where
overflows will  occur,  and the rates,  volumes, and
durations of the episodes. Rainfall amounts vary from
year to year, storm event to storm event, and  hour to
hour during individual  events.

Effective CSO control  requires that an  appropriate
design condition be selected from among the wide array
of naturally occurring conditions. In some  cases,  a
predetermined  "design   storm"  may   have  been
designated. While this can be used to provide a design
basis for a particular  CSO treatment unit, it leaves a
number of design issues unanswered. First, there  is
some  ambiguity  in converting a designated  design
storm  to CSO control system  design, depending  on
whether  the control  technology design  is  keyed to
volume or to flow rate.  In  cases where only a return
period is used in  the definition (e.g.,  10-year storm,
1-year storm), a wide variation in the volumes and rates
will occur depending on the arbitrarily selected duration
associated with the recurrence interval.  Even in the
case of a more complete designation, as in the 1-year,
6-hour  design storm,  ambiguity  exists. While  this
defines a volume,  it does not represent a volume for a
complete storm event, so inferences on performance of
storage units will be  uncertain. A rate  of flow  can be
extracted from this design condition in a number of
ways,  none  of which provide a confident  basis  for
design of a technology unit based  on  flow rates and
peak flows.

The  principal  issue  to   address   in  selecting  an
appropriate design basis for a treatment technology can
be stated as follows. While an arbitrarily selected design
storm condition provides a convenient starting  point to
size an  individual control unit and  develop  design
details, it provides no  information on the overall level of
performance that will be provided. This approach tacitly
assumes that the  designated design storm is, in fact,
the appropriate basis for meeting control objectives, and
that control units  based on it will provide acceptable
levels  of  performance.  Consider   situations  where
control performance goals are stated in other terms (see
Chapter 3).  If a CSO control program is to be designed
to  capture  or treat  some percentage  of  the total
combined sewer flow, or to limit untreated overflows to
some specified number per year, then  the designer is
required to  develop,  by appropriate analyses  of local
rainfall   and  collection  system   characteristics,  a
site-specific determination of the  appropriate design
condition to apply. Rainfall characteristics and analysis
issues are discussed further in Chapter 3.

In summary, the design storm approach comes from
flood control and urban drainage planning, where peak
flow is of concern for flood damage risk control. Smaller
storms are not an  issue.  For pollution control, even
small  storms can  result in standards violations and
environmental   harm,  especially  due  to  bacterial
contamination. As a consequence, the effect of control
measures on the annual frequency and volume of all
CSO  events  is  more   important  than  conditions
associated with a particular peak flow.

CSO Flows Are Intermittent and Highly

While analysis and interpretation of rainfall records can
provide useful information for identifying an appropriate
design basis, drainage area and conveyance system
characteristics impose site-specific influences on  the
flow rates and volumes that will occur during any storm
event  at locations in the  system where control
technologies will be applied. Each  combined sewer
system is unique, and most will be of sufficient extent
and complexity that deriving the necessary information
on combined sewer flows and CSOs usually requires
the application of simulation models such as EPA's
Storm Water Management Model (SWMM).

Data  requirements  for the proper application of such
models can be substantial, and are discussed in further
detail  in Chapter  3.  Data  requirements  include
information on physical features of the sewershed and
combined sewer system,  such as areas,  pipe sizes,
slopes,   regulator  and pump  station  design,  and
operating parameters. In addition, accurate information
on the characteristics of a number of individual storms
and the flows they produce at various locations in the
system is required to calibrate the model. This, in turn,
requires  flow monitoring, usually at multiple  locations
and for an adequate number of storm events to permit
an acceptable mode! calibration to be made.

Thus, a considerable study effort often  is required
simply to identify the flow regime that will prevail at the
point  of  application  of a  control  technology.  The
characteristics of the variable flows at  one or more
control locations must be developed by applying the
calibrated model to a  sufficiently long  sequence of
rainfall   events  to  provide  representative  results.
Additional  analysis  then  is   required  to  derive
appropriate  design  parameters  for the  treatment
technology to be evaluated  or applied, based on the
pattern of variable flows projected to be the inflows to
the unit.

Facilities for the control of CSOs must be capable of
performing under  a  wide range  of flow and  load

conditions. CSO facilities have influent hydrographs that
often are sharply peaked along with periods of no flow,
so that the concept of an "average flow" often used in
POTW design  has little physical meaning  to CSO
facilities.   Peak  flow  rate  is  an  important  design
parameter  for a number of  technologies,  and  the
designer  must select an appropriate value  from the
range of  extreme values that will apply at the control
location.  The  range in  actual flow  values  must be
developed by the simulation model, however,  insight to
the inherent variability and  recurrence  frequency of
extreme values is provided by the rainfall characteristics
summarized in Table 3-2 of Chapter 3.

CSO Quality and Testability Are Site Specific

Pollutant concentration levels in CSOs are site specific
and influenced by the strength of dry weather flows, the
age and  condition of the collection  system,  and the
amount of infiltration/inflow relative to the sanitary flow.

The  fact that  CSO  quality  characteristics  are  site
specific limits the confidence with which data from other
sites can be applied. A number of the quality features
also influences treatability parameters.  For  example,
the settling characteristics of solids in CSOs to be
treated by sedimentation technology are influenced by
local  system features, so monitoring and treatability
testing is advisable to develop control unit design

Higher pollutant concentrations may be associated with
the initial peak flows, depending on factors such as the
size and  slope of the piping system, the time  interval
between  storms, the drainage area characteristics and
response, and the solids accumulation in the collection
system. The existence and/or magnitude of high initial
loadings  can influence the design of a control  unit, as
discussed in Chapter 3.

Performance Data on CSO Controls Is Limited

Due to the highly variable nature of CSO flows, the
relationships between pollutant  removals and design
parameters, such  as hydraulic loading  rates and
detention times,  can  be difficult  to  establish with
reliability. Thus,  while many studies on pilot-scale or
one-of-a-kind installations have been published in the
past, a comprehensive data base covering performance
and design criteria for the most commonly employed
CSO control measures remains incomplete.

Contributing to this limitation is the cost and difficulty of
implementing effective monitoring programs to develop
operating data on existing  CSO facilities.  Unlike at
dry-weather POTWs,  where operating data  generally
can  be  collected  at  the  operator's  convenience,
collecting useful  data at CSO  facilities  requires  a
concerted effort by personnel to be available on short
notice and at the odd hours during which storm events
often  occur. In addition, because of the substantial
variability in  applied   flows  and   pollutant  loads,
monitoring programs to characterize performance must
extend over a sufficiently long period of time to reliably
determine performance level.

Because of the variability in flow and quality at any site,
the differences between sites, and the technical difficulty
and  cost of  developing comprehensive  performance
monitoring data, CSO control technologies represent a
situation where theoretical prediction of flow and quality
(based on models calibrated against limited data sets)
may provide  a more  accurate basis  for determining
design parameters and performance characteristics.

Major Technologies To Be  Addressed

This section identifies six CSO control technologies for
which detailed design guidance is provided in Chapter
4. The  technologies were selected based on the  fact
that  they are  currently  in wide  use and have been
demonstrated to be effective in  reducing CSO flows
and/or pollutant loads.

Separately, these control technologies will almost never
be sufficient to satisfy the needs of a comprehensive
CSO control program. For example, coarse screening
is rarely considered the  only technology to be applied
at an  overflow  point;  its  more common  use is in
providing pretreatment for any or all of the other control
technologies. In cases where bacterial levels produced
by CSOs cause designated use impairment, disinfection
is a necessary component of a CSO control system that
may include  any of the other technologies.  In  fact,
disinfection of CSOs often requires some level of solids
reduction by one of the other technologies for maximum
effectiveness and reliability. System-wide  CSO  control
may well  require application of different basic  control
technologies  or  combinations  at different overflow

A  comprehensive   control  plan   addressing   the
characteristics  of the combined sewer  system  and
overflows, which identifies the  impact  of  CSOs on
receiving  water  uses  and  establishes  performance
goals for the CSO control program, will provide the basis
for selecting and locating appropriate technologies (or
technology combinations) in the system.

A comprehensive control program design also should
consider  other  control methods,   such  as  those
discussed below. Any methods that  apply  should be
incorporated because some will affect the design basis
of a control technology, and many will be necessary to
assure  continuing effective operation  of a selected
treatment technology.

A comprehensive CSO control  program is likely to
incorporate one or more of the following technologies,
for which design details are provided in Chapter 4.

In-System Controls/ln-Line Storage

This technology seeks to optimize the use of existing
storage capacity in the collection system,  and maximize
the conveyance  of  combined  flows  to the POTW.
In-system controls typically are less costly than other,
more capital-intensive technologies such as  off-line
storage/sedimentation, and are attractive because they
utilize the existing facilities most efficiently. However,
they are not normally sufficient to provide the complete
degree of control required. The application of in-system
controls, and the feasibility of in-system storage is very
site specific. A variety of common regulator types, along
with control strategies and operational issues pertaining
to in-system controls, are addressed below.

Near-Surface Off-Line Storage/Sedimentation

This technology reduces overflow quantity and frequency
by storing all or a portion of the CSO that occurs during
storm events. In designs  providing sedimentation, flow
in excess of the tank volume passes through the units,
receiving  some measure of solids separation.  For
smaller storms, the tanks may provide 100-percent
capture. Stored flows are returned to the  interceptor for
conveyance to the POTW once system  capacity is
available. In some cases, flows may be conveyed to a
CSO treatment facility.  This manual  reviews basic
sedimentation   theory  as  applied  to  CSO  control
facilities, and  presents design criteria, examples of
design  details,  and  control  strategies  for  storage/
sedimentation facilities.

Deep Tunnel Storage

This technology provides storage and conveyance of
storm flows in large tunnels constructed well below the
surface. Tunnels  can provide large storage volumes
with  relatively  minimal  disturbance  to the  ground
surface,  which can  be very  beneficial  in congested
urban  areas. The components  of tunnel systems as
they  relate to CSO control are  described below.
Geotechnical aspects of tunnel construction are noted,
but are not addressed in detail in this manual.

Coarse Screening

This technology provides coarse solids removal, as well
as a degree of floatables removal.  Coarse screening
typically   is  provided  upstream  of   other  control
technologies, such as storage facilities or vortex  units
that are applied as off-line treatment units (rather than
as in-line regulator/degritter units). Aspects of coarse
screening as they relate to CSO applications and design
are presented below.
Swirl/Vortex Technologies

These devices  provide  flow regulation and solids
separation by inducing a swirling motion within a vessel.
Solids are concentrated  and removed  through an
underdrain, while clarified effluent passes over a weir at
the top of the vessel. The swirl/vortex devices described
in this manual include the U.S. Environmental Protection
Agency (EPA) swirl concentrator and commercial vortex

Conceptually, the EPA swirl concentrator is designed to
act as an in-line regulator device. In addition to flow
routing  or  diversion, it  removes  heavy solids and
floatables from the overflow. The commercial vortex
separators are based on the same concept as the EPA
swirl  concentrator, but include a  number  of design
modifications intended to improve solids separation.
Two commercial designs, the Fluidsep vortex separator
and the Storm King hydrodynamic separator have been
applied as  off-line  treatment  units.  Each type  of
swirl/vortex unit has a different configuration of depth/
diameter ratio,  baffles, pipe  arrangements,  and other
details designed to maximize performance.  The basis
of design  for  each  type of swirl/vortex  unit, and
examples of design details and control strategies, are
reviewed in this manual.


This process inactivates or destroys microorganisms in
overflows,  most  commonly  through  contact  with
chlorine, although a variety of disinfection technologies
are available  without chlorine.  Some of  the  more
common technologies include gaseous chlorine, liquid
sodium  hypochlorite,  chlorine   dioxide,  ultraviolet
radiation, and ozone. For disinfection of CSOs, liquid
sodium hypochlorite  is the most common technology.
This manual  focuses on the design  of liquid sodium
hypochlorite disinfection systems.

Other Control Methods

A variety of practices and control techniques can be
utilized to  supplement the  application  of  a control
technology at a CSO discharge location. An overall
control program should consider all  possibilities and
utilize  any  that  apply  for  the  local   situation.
Implementing a locally appropriate combination of the
practices discussed below can enhance performance of
the control technology applied and, in some cases, may
reduce the design size and cost of the basic control unit,
while maintaining the targeted performance level.

Control "practices" fall into one of three categories:

• Practices that  restrict  the rate and/or volume of
  stormwater  runoff  that enters  the combined sewer

• Pollution  prevention  practices  that  reduce  the
  quantity of pollutants that enter the system.

• Operation and maintenance practices for the combined
  sewer system  that improve its ability to contain wet
  weather flows  and deliver them to the POTW.
Examples of such control practices are described below.
A number of treatment technologies available for control
of CSOs are not addressed in this manual. Examples
include  dissolved   air floatation,  fine  screens  and
microstrainers,   high  rate  filtration,  and   biological
treatment. These technologies have been tested on a
demonstration scale  and found  to  be capable of
effective  CSO control.  Detailed  discussion of these
technologies is  outside  the scope of this manual;
however, a brief  discussion of each is provided later in
this section.

Flow Control Practices
Both infiltration of ground water into a combined sewer
system, and direct inflow of surface stormwater runoff
to the system can significantly influence the magnitude
and frequency of CSOs, and the size and cost of control
technologies. Reducing the quantity of infiltration and
inflow  (l&l)  will make  additional  system  capacity
available to  contain wet weather flows and  reduce the
magnitude of the CSOs reaching the control technology
application points.
Two examples of flow control  practices are infiltration
and inflow control, described below:
•  Infiltration  control: Sources of  infiltration  include
  ground water entering the collection system through
  defective pipe joints, cracked or broken  pipes, and
   manholes as  well  as  footing  drains and springs.
   Infiltration flow rates tend to be relatively constant
  and result in lower volumes than inflow contributions.
   Infiltration problems usually are not isolated,  and
  often reflect a more general sewer system deterioration.
   Extensive sewer rehabilitation typically is  required to
  effectively remove infiltration. The rehabilitation effort
   often must include house laterals, which are normally
   a significant source. Infiltration control will generally
   have a much smaller impact on CSO reduction  than
  will control applied to inflow.
•  Inflow  control:  Combined   sewer  systems  were
   designed  to  drain stormwater effectively  and to
   convey  sanitary  sewage. For a large percentage of
   storm events, the surface  runoff flows  are much
   greater  than  the  sanitary  sewage  flows in the
   combined  system.  CSO  control  efforts can be
   assisted either by diverting some of the surface runoff
   inflows  from  the combined sewer system  to an
  alternate surface drainage system or to ground water
  via  infiltration devices, or by retarding  the  rate at
  which these flows are permitted to enter  the system.

  Inflow of surface runoff can  be retarded by using
  special gratings or restricted outlet  pipes to modify
  catch basin inlets to restrict the rate at which surface
  runoff is permitted to enter the conveyance system.
  Inlet flow restrictions  may be designed  to produce
  acceptable levels of temporary ponding on streets or
  parking lot surfaces, allowing all  runoff to eventually
  enter the system at the inflow point, but reducing the
  peak flow rates that  the combined sewer  system
  experiences.  Flow detention  to  delay the entry of
  runoff  into  the collection system  by storing  it
  temporarily and releasing it at a  controlled rate also
  can  be accomplished by  roof-top  storage  under
  appropriate site conditions.

Peak   flow rates  in  downstream segments  of  the
collection system and at overflow points also may be
regulated by the installation of flow-restricting devices
at suitable  locations  in  upstream  portions  of  the
combined sewer system.  A variety  of  commercial
flow-restricting devices are available, such as the vortex
valves described in Chapter 4.

Eliminating the direct  connection of roof drains to the
CSO collection system and  causing this runoff to reach
the system inlets by overland flow patterns (preferably
via unpaved or vegetated areas) also can retard inflows.
When  site conditions permit, some  surface runoff flows
may be prevented from  entering the combined system
by diverting them via overland flow to pervious areas or
to separate storm drains. When these outlets  are not
available, excess surface runoff flows  may be diverted
to more favorable locations in the combined system, a
technique  referred to as flow-slipping (WPCF, 1989).
Also, depending on site conditions, it may be possible
to intercept some of the surface runoff flows by using
infiltration devices, and divert them to  ground water. In
coastal communities, repair and adequate maintenance
of tide gates  may  prevent flow intrusion  caused by
diurnal tide cycles.
The   implementation  ("retrofit")  of   l&l   controls in
developed areas on a scale necessary to substantially
control CSOs is difficult and may  be impractical. Flow
control techniques,  however,  may  be  useful  and
practical in selected problem  areas  for  addressing
specific  segments of  an  overall  system.  Control
measures  that  address  l&l  to  avoid  increasing
stormwater flows  in  a combined  system are  most
effectively  implemented  in  areas   currently  being
developed, where their use can be required as a
condition for development or reconstruction.

Pollution Prevention Practices

Pollution prevention measures include source controls
and other actions within a drainage basin that reduce
the amount of stormwater-related pollution entering the
combined sewer system. Source controls usually do not
require   large  capital  expenditures,   but  they  are
generally labor intensive,  and ongoing operation and
maintenance costs can be  high.

Controlling the use of problem materials  such  as
de-icing  salt,  fertilizers,  and  pesticides can  reduce
pollutants  entering  the   system.  Product  bans  or
substitutions (e.g., for plastic fast-food packaging) can
prevent  particular  problem  pollutants  from  being
generated.  Pollution prevention  also can  include
increased frequency  of  solid waste  collection, and
programs that enhance the environmentally responsive
disposal of bulk refuse, home renovation debris, and
household hazardous wastes.

Other control measures in this category are designed to
minimize accumulation  of pollutants  on streets and
other  tributary   land  areas   and  in catchbasins.
Implementing  these  measures decreases  pollutant
loadings to combined sewers by preventing or reducing
their entry.
Examples of pollution prevention practices applicable to
comprehensive control programs for reducing pollutant
discharges from CSOs are discussed below.

Street Cleaning

Street litter can  be  a significant  source  of certain
pollutants (e.g., floatables) entering receiving waters via
CSOs. An extensive monitoring program conducted by
the City  of New York (NYC DEP, 1993) concluded that
the major portion (perhaps 95 percent or more) of the
floatables in  CSOs originate as street litter. Litter on
streets can be removed by mechanical or manual street
cleaning. Street  sweeping  often  is  considered  a
practical  best management practice  (BMP) for  CSO
pollution control.  The effectiveness of street sweeping
depends on the rainfall frequency, sweeping frequency,
and other factors  such  as  street density  and the
prevalence of curbside parking. A major impediment to
street  cleaning in  densely populated  cities  is cars
parked adjacent to the curb; therefore, enforced parking
regulations  are an essential  component  of a street
cleaning program.

Public Education Programs

Education methods  may  consist of developing public
announcements,  advertising,  stenciling street  drain
inlets, and distributing information with water/sewer
bills. An important aspect of a public education program
for CSO control is to encourage the proper disposal of
sanitary and personal hygiene items disposed of through
household toilets. The New York City floatables study
(NYC DEP, 1993) determined that although these items
accounted for only about 5 percent of floatable materials
discharged  by   CSOs,   they  include  the  more
objectionable items which cause the greatest public
concerns and result in beach closings.

Anti-litter campaigns can  reduce the amount of street
litter and floatables that originate from CSO. Since it is
unrealistic  to  anticipate  widespread enforcement of
anti-litter  ordinances,  the   effectiveness   of   such
programs depends on public education. Citizen action
or education  programs instituted to focus on specific
issues such as those identified above, also will  raise
public awareness of the problems associated  with
CSOs and the  justification for the broader control

Recycle Programs

Crankcase oil, paints, cleaning agents, chemicals and
other household wastes,  as well as leaves and grass
clippings sometimes are disposed of in catchbasins or
street inlets. Proper disposal of these materials should
be addressed in public education  programs, but an
essential  element  of  an  effective  program is  the
availability  of suitable disposal mechanisms. Recycle
programs  that  establish disposal  locations and/or
collection  schedules  must   be  organized by   the
municipal agency, though they may utilize commercial
establishments  to  implement components  of  such
programs (e.g., acceptance of waste oil by service stations).

Fertilizer and Pesticide  Control

Fertilizers and pesticides washed off the ground during
storms contribute to runoff pollutant levels.  Controlling
use of these chemicals on  municipal  lands can help
reduce the pollutant load. For example, in urban areas
an important source of fertilizer and pesticide runoff that
can readily be controlled is the park systems. Individual
homeowner use of these chemicals is not likely to be a
major source of pollution in  urban  areas served by
combined  sewers.  Where   homeowners  do   use
fertilizers and/or pesticides, a public education program
is required to address this issue.

Soil Erosion Control

Controlling  soil erosion  is  important because  soil
particles carry nutrients and metals as well as contribute
sediment. The principal areas of potential  concern in
urban areas served by combined sewers are public
parks and  construction sites. Problems in public  parks
may require  limitations on the type of  use,  regrading,
and/or revegetation of eroding areas.

Commercial/Industrial Runoff Control

Certain  commercial  and  industrial  sites  can  be
responsible for disproportionate  contributions of some
pollutants (e.g., grit, oils, grease, and toxic materials) to
the combined sewers. Sources  of  potential concern
include gasoline stations, railroad yards, freight loading
areas,  and  parking lots.  In specific cases  where
significant pollutant  loadings to the combined sewer
system  are  contributed by well-defined locations  of
limited area,  pretreatment of runoff  from these areas
may be  a practical and  effective  control  measure.
Pretreatment measures can be required as part of a
community's  sewer  use  regulations.  Examples  of
pretreatment measures include oil/water separators for
problem service stations, or use of modified catch-basin
designs to enhance retention of oil and grease or solids.

Procedures for detecting and locating illicit connections
to separate storm drains by testing for specific chemical
tracers (U.S. EPA, 1993) can be applied to CSOs to
identify commercial  or industrial sources contributing
substantial levels of problem pollutants.

Operation and Maintenance Practices

Operation and maintenance (O&M)  program activities
that focus on combined  sewer system components
(regulators, tide gates, pump stations, sewer lines, and
catchbasins), can significantly influence the  level of
control applied to CSOs.

Regulator/Tide Gate Maintenance

Because of  the  debris normally present in combined
sewage, especially at times of storm flows, regulators
continually accumulate materials that cause clogging
and blockages. This  is  a particular problem  with
static-passive regulators.

The majority  of unnecessary  overflows  in  passive
regulators are caused  by  trash  blocking the entrance
orifice to the interceptor. Other causes of unnecessary
diversions at regulators include weir plates or dams that
are improperly set, damaged, or broken off. Improperly
operating  tide gates allow receiving  water to enter the
combined sewer system, and reduce the storage and
flow capacity of downstream interceptors that otherwise
would  be  available during wet weather. Trash  in
combined sewer flows and/or trash  and timber in the
receiving water body can cause  a tide gate to remain
partially   open.   Corroded  or  warped  gates,  or
deteriorated gate gaskets are also common causes of
improper operation.

Frequent inspection of CSO regulators and tide gates
will assure that  as  much  of the wet weather flow as
system capacity permits is retained in the system. The
size and characteristics of the combined sewer system
will influence the schedule, but where practical,  an
inspection several times per month, and after every storm,
will ensure that necessary repairs and maintenance to
clear debris and obstructions are performed in a timely

Sensors  to detect the presence  of water in overflow
lines during dry weather periods  (resulting from faulty
regulator or tide gate operation) are used to signal the
need for  maintenance action.

Maintenance  programs  that prevent the  systematic
overflow of unnecessary quantities of wet weather flows
can influence  the required design size and  cost of
treatment technologies. Furthermore, effective O&M for
collection system elements is necessary to ensure that
CSO controls operate in accordance with conditions for
which the design is based.

Pump Station Maintenance

Proper operation and maintenance will ensure  that
pump stations transfer the design flows. Inadequate
pumping  capacity may  result from changes in the
upstream area that cause flows to exceed the original
design basis, or from mechanical defects that cause
substandard performance. Inadequate capacity results
in back-up  of dry weather flow in the sewer system,
possibly  contributing to  dry weather overflows  and
resulting  in reduced storage and hydraulic flow capacity
for wet  weather  flows.  This contributes to CSO at
overflow  points upstream of the pump station, and also
prevents  the   wastewater  treatment   facility  from
maximizing treatment of  wet weather flows.

Removal of Sewer Line Obstructions

Blockages in interceptor sewers can cause back-ups
that create excess CSO overflows upstream  of the
blockage. A common cause of sewer line obstruction is
the deposition of solids,  which deplete the in-system
storage capacity otherwise available for wet weather
flow and  reduce the flow-carrying capacity of the sewers
(U.S.  EPA, 1984a). Other system obstructions may result
from roots growing into pipes and from collapsed pipes.

Removal of flow obstructions may include maintenance
activities to remove and prevent accumulations of
debris in parts  of  the  system that  experience  flow
restrictions.  Where flow obstruction  is the  result of
sediment accumulations, sewer  flushing  may be an
effective  control  measure. A maintenance program of
periodically  flushing sewers during  dry weather to
convey   settled  materials to dry weather  treatment
facilities  minimizes  the buildup of such sediments. In
cases where  a section of  the conveyance system
routinely  accumulates   sediment  deposits  at  a
substantial  rate, design and installation of a permanent
flushing  station or an in-line grit chamber  may be a
cost-effective approach.

Severe situations may require the application of sewer
cleaning measures to physically clean a segment of the
conveyance system. Hydraulic, mechanical, or manual
devices may be required to remove solids or resuspend
solids into the waste flow and carry them out of the
collection system (U.S.  EPA, 1984b).

Knowledge of the area and collection system assists in
determining the parts of  the system where particular
problems can  occur, and guides development of an
effective inspection  program. For example, collapsed
pipes occur most frequently in  areas with old sewer
lines or with current or  recent  construction activity.
Excessive infiltration also  is most likely to be a problem
in these areas, as well as where water tables are high.
The potential  for  obstructions  caused by  solids
deposition is greatest where velocities are low during
dry weather and for small-  and average-sized storms.
These conditions exist where lines are oversized for the
flows they normally convey  and/or where gradients are
flat. As a result, solids deposition obstructions likely will
recur where they are present. Lines with a history for
acquiring sediment  deposits should be scheduled for
cleaning on a regular basis.

Catchbasin Cleaning
In many communities, catchbasin cleaning is performed
infrequently and is targeted  towards maintaining proper
drainage system  performance  rather than  pollution
control. Regular cleaning  of catchbasins (once or twice
per year) can remove accumulated sediment and debris
that ultimately could be discharged from CSOs, thus
providing  some  degree  of pollution  control.  This
technique applies  for  true catchbasins,  that is inlet
chambers that provide a sump for retention of sediment
and debris. The technique does not apply for simple
drain inlets.

Other Control Technologies
Physical  treatment  technologies  other than  those
identified earlier in  this chapter and  that have been
applied to CSOs  include dissolved air floatation,  fine
screens  and  microstrainers,  dual-media  high  rate
filtration, and  biological treatment.  Sewer separation
also is considered  a control technology here and is
included  in the discussion that follows. Although the
technologies have  not been widely applied in CSO
control  applications, they  have been  tested  on  a
demonstration scale and are effective in CSO control. A
brief synopsis of these technologies as they apply to
CSO control is presented below.

Sewer Separation
Separation is the conversion of a  combined sewer
system into separate stormwater and sanitary sewage
collection systems. This alternative, historically considered
the ultimate answer to CSO pollution control, has been
reconsidered  in recent years because of cost and the
major disruptions to traffic and other daily community
activities associated with separation. Separate stormwater
runoff  also  contains  pollutants  (sediments,  organic
matter, bacteria, metals,  oils, floatables,  etc.), which
continue to be discharged to the receiving waters.

Several potential benefits of  sewer separation might
warrant its  consideration in  specific cases.  These

• Eliminating CSOs and preventing untreated sanitary
  sewage from  entering  the  receiving  waters during
  wet  weather periods. Sanitary sewage is a more
  objectionable  source of some pollutants,  such as
  TSS, sanitary  floatables, and bacteria.

• Reduced volume of flow to be treated at the POTW,
  thus reducing O&M costs, by eliminating surface
  runoff inflows during wet weather periods.

• Reduced infiltration and excess flow to a  POTW  if
  new sanitary sewers are constructed to replace old
  combined  sewers.

• Reducing upstream flooding as well as overflows in
  cases  where  the   existing combined  sewers  are
  undersized  and back  up frequently during storm

• Being more effective and economical than treatment
  facilities for remote segments of  a combined sewer
  system serving relatively small areas.

Dissolved Air Floatation

Dissolved air  floatation  (DAF) removes solids by
introducing fine air bubbles to wastewater. Air bubbles
attach to solid particles suspended in the liquid, causing
the solids to float to  the surface where  they can be
skimmed off. This technology has been tested in CSO
applications (U.S. EPA, 1972a, 1975a, 1977, 1979a). A
major advantage of DAF is its relatively high overflow
rate and short detention time, which results in reduced
facility size compared to conventional  sedimentation.
Oil and grease also are more  readily removed by DAF.
Operating costs for DAFs are high due to a large energy
demand,  and skilled operators are required  for  its

Fine Screens and Microstrainers

These devices remove solids through capture on screen
media. They have been tested for use in CSO control
(U.S.  EPA,   1970,  1971,  1973,  1974a). The most
common fine screening devices are rotary drum and
rotary disk devices. In the rotary drum screen, media is
mounted on  a rotating drum. Flow enters the end of the
drum and passes out through the filter media. Drum
rotational  speed usually is  adjustable. Solids retained

on the inside of the drum are backwashed to a collection
trough. Filter media aperture size typically ranges from
15 to 600  microns. The  rotary disk screen media is
mounted on a circular frame placed perpendicular to the
flow. Flow passes through the bottom half of the rotating
disk, which is submerged. Solids retained on the disk
are directed to a discharge launder using spray water.

One form of static screens features wedge-shaped steel
bars, with the flat part of the  wedge facing the flow.
These "wedge-wire"  screens typically have openings
ranging from 0.01  to 0.05  in.  These screens require
daily maintenance to  prevent clogging (Metcalf & Eddy,
Inc., 1991).

Screens are subject  to blinding from grease and "first
flush" solids loads. A means for providing high-pressure
backwash,  and  collecting  and conveying backwash
solids typically is required. Effective cleaning of screens
after storm events  using high-pressure steam  or
cleaning agents is required to maintain performance.

Removal efficiencies may be increased by decreasing
media aperture size, but smaller apertures are more
likely to blind. Coarse screening and disinfection facilities
often are provided in conjunction with microstrainers.


Dual-media high-rate filtration  has been piloted for
treatment of CSO flows (U.S.  EPA, 1972b, 1979b). A
two-layer bed, consisting  of coarse anthracite particles
on top of less coarse sand, was used. After backwash,
the less dense anthracite remains on top of the sand.
Filtration rates of 16 gal/f^/min or more were utilized
resulting in substantially  smaller area requirements
compared  with   sedimentation.  Demonstration  test
systems included pretreatment by fine-mesh screens.
The use of chemical coagulants improved performance
considerably. Filtration is more appropriately applied
after  pretreatment  provided   by   fine  screening.
Operation may be automated, but tends to  be O&M

Biological Treatment

Biological treatment  processes  have been tested in
CSO control applications (U.S. EPA,  1974b, 1975b,
1981a,b). Although they have potential to provide a high
quality effluent, disadvantages of biological treatment of
CSOs include:

• The biomass used to break down the organic material
  and assimilate nutrients in  the combined  sewage
  must be  kept alive during dry weather, which can be
  difficult except at an  existing treatment  plant.

• The land requirements for these types of processes
  can preclude their consideration in  an urban area.
• Operation and maintenance can be  costly  and
  facilities require highly skilled operators.

Some biological treatment technologies are utilized in
CSO control as  elements of a wastewater treatment
plant.  Pump-back or  bleed-back  flows  from  CSO
storage  facilities  commonly   receive   secondary
treatment at the treatment plant, once wet weather flows
have subsided. In a treatment plant that has maximized
the wet weather  flows that are  accepted, flows are
sometimes split,  with only a portion  of the primary
treated flows  receiving secondary treatment,  to  avoid
process  upset.  The  split flows  are  blended  and
disinfected for discharge.


When an  NTIS number is cited in a  reference, that
document is available from:
  National Technical Information Service
  5285 Port Royal Road
  Springfield, VA  22161

Metcalf & Eddy,  Inc.  1991. Wastewater engineering:
  treatment, disposal, reuse.  3rd edition. New  York:
NYC  DEP.  1993.  New  York City  Department  of
  Environmental Protection. Citywide floatables study:
  characterization and  quantification  of  sources  of
  floatable materials.  Report  by  HydroQual,  Inc.

U.S. EPA. 1970. U.S. Environmental Protection Agency.
  Microstraining  and disinfection of combined  sewer
  overflows. Report no. 11023 EVO 06/70 (NTIS PB
  195  674). (U.S. EPA called Federal Water Quality
  Administration in 1970). June.

U.S. EPA. 1971. U.S. Environmental Protection Agency.
  Demonstration  of  rotary screening  for combined
  sewer overflows. Report no. 11023 FDD 07/71 (NTIS
  PB 206 814). Portland, OR. July.

U.S.  EPA.  1972a.  U.S.  Environmental  Protection
  Agency. Screening/floatation treatment of combined
  sewer overflows. Report no. 11020 FDC 01/72 (NTIS
  PB 215 695). January.

U.S.  EPA.  1972b.  U.S.  Environmental  Protection
  Agency.  High-rate filtration  of  combined  sewer
  overflows. Report  no. 110EYI 04/72  (NTIS PB 211
  144). Cleveland, OH. April.

U.S. EPA. 1973. U.S. Environmental Protection Agency.
  Microstraining and  disinfection of combined sewer
  overflows:  phase  II. EPA/R2-73/124 (NTIS PB 219
  879). January.

U.S.  EPA.  1974a.  U.S.  Environmental  Protection
  Agency. Microstraining and disinfection of combined
  sewer overflows:  phase III. EPA/670/2-74/049 (NTIS
  PB 235 771). August.

U.S.  EPA.  1974b.  U.S.  Environmental  Protection
  Agency. Combined sewer overflow treatment by the
  rotating biological  contactor  process.  EPA/670/
  2-74/P50 (NTIS PB 231 251). June.

U.S.  EPA.  1975a.  U.S.  Environmental  Protection
  Agency. Treatment of combined sewer overflow by
  dissolved air floatation. EPA/600/2-75/033 (NTIS PB
  248 186). September.

U.S.  EPA.  1975b.  U.S.  Environmental  Protection
  Agency. Biological treatment of combined sewer
  overflow at Kenosha, Wisconsin. EPA/670/2-75/019
  (NTIS PB 242 126).  April.

U.S. EPA. 1977. U.S. Environmental Protection Agency.
  Screening/floatation  treatment  of  combined sewer
  overflows: volume I—bench scale and pilot plant
  investigations.  EPA/600/2-77/069a  (NTIS  PB 272
  834). August.

U.S.  EPA.  1979a. U.S.   Environmental  Protection
  Agency. Screening/floatation treatment of combined
  sewer  overflows: volume  If—full-scale operation
  Racine, Wisconsin.  EPA/600/2-79/106a (NTIS  PB
  80-130693). August.
U.S.  EPA.  1979b.  U.S.  Environmental  Protection
  Agency. Dual  process high-rate  filtration  of raw
  sanitary sewage and  combined  sewer overflows.
  EPA/600/2-79/015 (NTIS PB 296 626/AS). April.

U.S.  EPA.  1981 a.  U.S.  Environmental  Protection
  Agency.  Evaluation of a  treatment  lagoon  for
  combined  sewer overflow. EPA/600/2-81/196 (NTIS
  PB 82-105214).

U.S.  EPA.  1981b.  U.S.  Environmental  Protection
  Agency. Joint dry/wet weather treatment of municipal
  wastewater at Clatskanie, Oregon. EPA/600/2-81/061
  (NTIS PB  81-187262).

U.S.  EPA.  1984a.  U.S.  Environmental  Protection
  Agency. Procedures for estimating dry-weather sewage
  in-line pollutant deposition—phase II. EPA/600/2-84/
  020 (NTIS PB  84-141480).

U.S.  EPA.  1984b.  U.S.  Environmental  Protection
  Agency. Dry weather  deposition  and  flushing for
  combined  sewer overflow pollution control. EPA/600/
  2-79/133 (NTIS PB 80-118524).    '

U.S. EPA. 1993. U.S. Environmental Protection Agency.
  Investigation of inappropriate pollutant entries into
  storm drainage systems: a  user's guide. EPA/600/
  R-92/238 (NTIS PB 93-131472). January.

WPCF.  1989. Water  Pollution  Control   Federation.
  Combined  sewer  overflow  pollution   abatement:
  manual of practice no.  FD-17. Alexandria, VA.

                                              Chapter 3
                                         Process Selection
Performance Goals

While the overall objective of CSO control programs is
to protect or improve water quality and designated uses
of water bodies that receive CSO discharges, selecting
a treatment process and determining size or capacity
involves a focus on a more specific performance target.

Individual technologies are more effective for controlling
some  pollutants  or  pollutant  classes  than  others.
Depending on the designated use of the receiving
water, the specific pollutant classes that directly affect
that use assume primary importance  in developing
control plans. Therefore, the first criterion in selecting
thr most appropriate technology to apply in  a specific
situation is its control effectiveness for  pollutants that
directly affect the protected use. Some overlap exists,
but the following  list of uses, CSO pollutants,  and
control technologies generally indicates  the basic type
of control technology to  consider.
Considerations In Technology Selection

Use         Pollutants
life/fin fish
Solids, BOD,
COD, metals,
toxic organtas
Aesthetics    Solids, floatables
fine screens

storage/ sedimentation,
fine screens

fine screens
                  Screens, swirl/vortex,
Where the objective  of CSO control is to mitigate a
specific water quality problem in an area immediately
influenced by  a CSO, the  foregoing  considerations
should have a major influence  on treatment process
selection.  However,  in  many  cases  CSO  control
requirements are addressed in a more general manner.

Cause and effect relationships between CSO and water
quality impacts often are difficult to establish. The
intermittent and highly variable nature of CSO and the
complexities of many receiving waters impose significant
demands on monitoring data requirements. In addition,
impacts  on  aquatic life  resulting  from  long-term
sediment pollutant buildup and/or an already degraded
receiving water may not be obvious, or able to be related
to CSO in a definitive way. In such cases, performance
goals for  CSO control may be  governed by state  or
other regulations, and expressed as one of a number of
alternative technology-based goals.

Examples  of   commonly  used  technology-based
performance goals are:
• Percent capture: That a specified percentage of flow
  be captured and/or treated.

• Overflow frequency:  Reducing  the  number  of
  untreated CSOs per year to a specified number.

• Treatment  level:  Specifying  the pollutant removal
  efficiency of the CSO controls; often specified as the
  equivalent of primary treatment.

• First flush:  Providing  capture and/or treatment  of
  some  portion of  a total overflow, determined  to
  contain a major fraction of the pollutant load.

• Knee-of-the-curve: Basing the size of a control unit
  on  cost effectiveness  (i.e.,  where  significant
  increases in cost produce marginal  improvements in

Selection of the general control strategy and specific
performance  goals  may  depend  on  local,  state,  or
federal regulations,  a  community's  long-term  CSO
control plan, or requirements of a  permitting authority.
In practice, a combination of technology and  water
quality considerations may be used  in developing a
CSO control plan. Water quality considerations may be
used  to  guide  the  general  approach,  identify the
pollutant types of major concern at a particular location
(and  hence  the appropriate  technologies), address
spatial issues such as consolidation and/or relocation of
outfalls, and similar issues. Specific performance goals,
such as those listed above,  then would be used  to
develop the design basis for individual treatment units
or systems.
Implementing  any strategy requires an estimate of the
CSO flows or volumes that must be used as a design

condition in order to meet the selected  performance
goal.  These  estimates require  information on  local
rainfall characteristics and an appropriate engineering
analysis   of   the  system  layout  and   hydraulic
characteristics. In most cases, an appropriate analysis
may require development of a monitoring  program and
application of computer models, in order to adequately
assess the operation of the collection system and/or the
water quality  impacts in the receiving  water body.
Computer  models  such  as  EPA's  Storm   Water
Management Model (SWMM) (U.S. EPA,  1988) are
commonly applied to simulate the hydraulic behavior of
a combined sewer system, and develop the information
necessary to design CSO controls.

All strategies require an estimate of the design flow to
use  in  developing the  design  details discussed in
Chapter 4. If a design storm  has been specified (for
example, the 1-year, 6-hour event), then an analysis
would be made to determine the CSO flows that would
be delivered to a planned  control unit by that rainfall
condition. However, if the performance goal specifies a
desired result (e.g., four  overflows per year), then the
system  analysis  to be  performed must,  in  effect,
develop both the appropriate "design storm" to apply,
and the CSO flows and volumes to be used for detailed

Data requirements  for system modeling  analysis and
process selection and design are presented later in this
chapter. General  design  considerations that relate to
each of the  possible  performance goals for  a CSO
control are presented below.

Design for Percent Capture

This goal is properly defined as capture for subsequent
treatment.  It is used to define performance goals for
storage technology. When applied to off-line storage for
CSO, this performance  goal is used to  define the
additional  storage volume  required  to  capture  a
specified percentage of  either the volume  of current
CSOs or  the volume of  wet weather  flows  in the
combined  sewer system,  usually considered  on a
system-wide annual average basis. This assumes that
all captured flow is returned to the collection  system
during dry weather for  treatment  at  a  POTW. The
required storage used for design considers the  storage
provided by both the collection system and additional
storage devices.

This  performance  goal could   also  be  applied to
treatment  units  or  combination   storage-treatment
systems  located  at  individual  outfall  points  (e.g.,
sedimentation,  swirl/vortex, screens) if  the specified
percentage is considered to be  the percentage of the
total combined sewage volume  treated by the unit at
flow rates consistent with design criteria.
The percent capture  by a  storage unit could  be
expressed in an individual case as the percentage of
current  discharges from  an overflow point.  In  the
general  case,   however,  it  is  usually  and  more
appropriately considered to be a percentage of the total
wet weather combined sewage flows. This  definition
imposes more uniform requirements for different CSO
systems, and allows the designer  flexibility to optimize
the use of  both in-line and off-line storage. Percent
capture is most appropriately considered as a  long-term
average (or  in a year of average rainfall).

An  indication  of  the  approximate  total  storage
requirements for different capture efficiencies can be
developed by extracting from the rainfall record the
percentage of total rainfall volume resulting from storm
events that are equal or smaller than a selected design
storm  volume.  Figure  3-1   shows  the  relationship
between design storm size used as a basis for storage
design  and  the  percentage  of the combined sewage
flows that will be captured, based on 42-year rainfall
records  at six locations. The performance pattern  is
similar for all locations. As the size of the design storm
increases, the effect of regional differences in  rainfall on
percent capture become progressively smaller. Since a
design  storm (for example, the 1-year, 6-hour storm)
can be expressed  as a volume, the  approximate
performance of a storage control  using it as a design
basis can be estimated from the rainfall record.
A critical assumption for  this estimate relates to the
ability to physically place  the storage at the locations
where it is required. The results apply directly for the
analysis of  a single  overflow point.  If  applied for a
system with multiple overflows, the tacit assumption is
that the required  storage capacity  is appropriately
distributed  among the  overflow  locations.  Existing
systems with portions that are undersized, or  otherwise
sensitive to "kicking-off" an overflow at  particular
locations during very small storms would generate more
overflow events than the screening analysis  suggests.
However, where these situations are associated  with
relatively small drainage areas, the number of overflow
occurrences may  be disproportionately high, but the
effect on the estimate of the overall percent captured
may be minor.

The tacit assumption is that the aggregate of all storm
events  equal to or  smaller than a selected storm  size
that accounts for a particular percentage  of total rain
volume also estimates the corresponding percentage of
the total wet weather flow volumes. For example,  if 75
percent of the total amount of rainfall is delivered by
storms  equal to or smaller than 1  inch, then about 74
percent  of the  total wet weather flow volume will be
provided by the runoff from storms of this size. Thus,
storage capacity sufficient to contain the runoff from a
1-inch storm would retain, in such a case,  about 75

                                DESIGN STORM SIZE  V»  % CSO CAPTURE-STORAGE
oo -
90 "
80 -
70 -
60 *
50 -
40 -
30 "
— „







! 1 i















— ,,







— i —
.._ ...4....
t :
t :
• s
I i
! 1
i i
| j
~ ..

— «
r~ i i i i i i i
                       0      0.5      1       1.5      2     2.5      3      a?      4

                                         DESIGN STORM SIZE  (Inch**)

Figure 3-1.  Approximate relationship between performance of CSO storage and storm size used as design basis—percent capture.
percent of the wet weather flows. The absolute amount
will vary with the physical size and other features of the
drainage area and collection system, and will  be site
specific. Details must be developed by  an appropriate
model or other analysis,  but the  required storage
capacity inferred from the area rainfall characteristics
would include all forms of storage that may be provided
by the  overall  system. This would include capacity
resulting from both in-system storage and off-line basins
at the overflow points (as well as any natural depression
storage  that  attenuates  the  rainfall   reaching  the
combined sewers).
Maximization of in-system  storage  accordingly would
reduce  off-line storage capacity required to produce a
desired capture percentage, because the required
storage volume also would be accounted for, in  part, by
the wet weather flows that the  conveyance  system
retains and delivers to the POTW as wet weather flows.
An off-line storage basin is required to provide capacity
only for the  excess over capacity  provided by other
features of the overall system.

The captured volume  usually  is sent to the POTW
where it is treated before  ultimate discharge. At this
point it is a wet weather-related component of the
treatment plant flow. The ability of the POTW to process
captured flows that are returned to the system,  and the
treatment for those flows, are important  elements of the
overall  control program. An  important influence on the
performance of  CSO storage units is the rate at which
they  can  be emptied following a storm. A very large
storage volume may be provided,  but if it cannot be
substantially emptied during the interval between most
storms, it has the practical effect of providing a much
smaller "effective" volume.
Two factors may restrict the emptying rate. One is the
hydraulic capacity of the interceptors to which captured
flows  are returned  for conveyance to the POTW.  If
infiltration into the  combined sewer system  is great
enough that flows in the line remain high for extended
periods following a storm, the ability to remove captured
CSOs may be significantly restricted. The other factor
relates to the capacity of the POTW and its ability to
accept increased flows on a relatively consistent basis.
If a large storage volume is provided, and a substantial
percentage of wet weather flow is captured and then
returned to the system over extended periods during dry
weather, the  result will be an increase in the  average
dry weather  flow to the  POTW.  This flow could be

The approximate  ratio between  annual volumes of
sanitary sewage and stormwater in a combined sewer
area is site specific. The sanitary flow from the area  is
determined by the wastewater generated (gallons per
capita), and  by the population density (persons per
acre). Stormwater  runoff  depends principally on the
amount of rainfall and how impervious the area is.  In
areas such as San Francisco with  annual precipitation
in the order  of  15  or  20 inches, the annual sanitary
volume generated is perhaps 3 to 5 times greater than
the stormwater  volume. In  areas with 60 inches  of

annual precipitation (e.g., the  Gulf Coast), the ratio is
about 1 and the annual volumes generated  are about
equal. A large percentage of CSO areas are in regions
with about 30 to 40 inches of annual precipitation, and
annual  sanitary volumes may be about  2  times  the
stormwater volume generated.  Storage controls that
provide a high  percentage  of CSO  capture  and
distribute the returned flow to the system over much of
the dry weather periods can significantly  increase the
"average" dry weather flow delivered to the  POTW for
treatment. There also will be additional sludge and grit
generated by CSO control. This is an element of CSO
control planning that should be properly considered.

Design To Reduce Overflow Frequency

Overflow frequency usually is defined as the number of
overflows per year as a long-term average. Overflow
frequency is higher in wet years and lower in dry years.
This performance goal may be  applied to  a specific
discharge, or treated as a value applied to the combined
sewer system as a whole.

A preliminary indication of the design storm size used
as  a basis for CSO storage design can be  developed
by analyzing the rainfall record to determine the number
of  storm  events that  have larger volumes than  the
design  storm.  CSO  storage  units  normally are  not
emptied until after the end of a storm because available
system capacity is utilized  by wet weather flows,  so
event volumes that exceed the design  storm produce
                             an  overflow.  Figure  3-2 illustrates the  approximate
                             relationship  between   the   number  of  overflows
                             (expressed as an  annual average), and the design
                             storm  size for which  storage  is provided, based  on
                             analysis of rainfall records. Results are plotted using a
                             log scale to assist in reading the low values.

                             A relationship exists  between percent  capture and
                             number of overflows for any site using storage as the
                             basic  approach to control. Figure 3-3 illustrates the
                             "equivalence" between performance goals that specify
                             percent capture  and number  of overflows.  The plot
                             combines  the results for all six locations shown
                             separately on Figures 3-1 and 3-2. Of particular interest
                             is the indication that very large storage volumes and
                             capture efficiencies are required to  reduce the number
                             of overflows to 4 or 6 per year.  Independent support for
                             the relationship indicated is provided by results from a
                             CSO study for the city of Windsor, Ontario (Mahood and
                             Zukovs, 1993). A model analysis performed using the
                             complete  simulation model,  STORM, was  used to
                             define  the relationship between  CSO  capture and
                             overflow frequency, and the results compare quite well
                             with the general relationship indicated  by  Figure 3-3.

                             Design to comply  with  this performance goal would
                             require continuous rainfall-flow modeling  to verify and
                             refine the above approximations. Parts of the collection
                             system  with   marginal  excess  capacity produce  a
                             disproportionate  number of overflows compared with
                             the system as a whole.
                                  DESIGN STORM SIZE v«  OVERFLOWS per YEAR
	1- -
                     1 0 - -: :

j..—j....j—.4 .—i >'-.?•'• 4—^-...*"..|....j....4-—J—-T™j—*f™4—-i—*••"•*•
^....(-...<|.  | .+....».«.<.„.+.„.).4....«....|.«.4....|..»4.«.4...^.«.«..«.>.
1 _.>....;....4..»i...4...4...H....«.k..i....;    •  •  '•  -  '-  • '• '-
  •••.'.:     •.  •  '

                                                                          Newark NJ
       • i  :  i  i


                                M  g   ,
                                w....»...^...^ ....
              444	j.4..i.i4444.
Portland ME  "1~

Portland OR  _L.

Loultvill* KY   i

Chicago IL   —r
                                                 Attana QA
                                          f_j_4	f-4-|—j	"f-
                                          i    '•
                                             "tl"1	r i1"tTrtt "IT  1~ 1 f
                                              i j  ;   i  ^ i  I  •.  I '  !  I  i    '  |  ! I
                                              :  :  :  :  :  i  : :  :  :  i  ;  !  i  !  :  :  :
                                        1      1.5     2      2.5      3

                                           DESIGN STORM SIZE   (Inch**)
 Figure 3-2.  Approximate relationship between performance of CSO storage and storm size used as design basis—number of

                         COMBINED  RELATIONSHIPS  FOR   6  CITIES

                       Newark NJ , Portalnd ME , Portland OR, Louisville KY. Chicago IL, Atlanta GA
            I   1
                                                                           « *
                   30       40       50       60       70       80       90

                         %  of Total Wet Weather How Volume Captured by Storage

Figure 3-3.  Relationship between percent capture and number of overflows for CSO storage control.
This performance goal could be applied to CSO control
systems that provide treatment to CSOs, but continue
to discharge them at the overflow point. In such a case,
it would be necessary to establish that an "overflow"
referred to  an untreated  discharge.  In  addition, for
treatment  units that operate as flow-through  devices,
the variable inflows produce short duration peaks, some
amount of which may be bypassed. Since these peaks,
at times, occur at more than one period during the same
storm  event,  agreement as  to  what constitutes an
"overflow evenf is important to develop an appropriate

Design To Provide a Specified Treatment

Treatment technologies such as screens, sedimentation,
and  swirl/vortex designs  are applied  at individual
discharge locations to  remove pollutants from all or a
portion  of the  combined  sewer  overflows,  without
modifying the overflow volume discharged at that point.
Combination surface storage/sedimentation units are
an exception because they return captured flow to the
POTW for additional treatment. Also, the underflow (foul
sewer  discharge) from swirl/vortex units is retained  in
the conveyance system.  However, for pure treatment
technologies, essentially all flows are discharged at the
                                      overflow point in question, but pollutants are reduced or
                                      removed from these flows in accordance with the design
                                      size  and treatment efficiency of the unit or process

                                      Any  treatment device  has an  inherent performance
                                      capability  dictated by  its  design principles and  the
                                      removal mechanisms it employs. The pollutant mass
                                      removal efficiency it can apply to the flows delivered to
                                      it, even under optimum operating conditions, will vary
                                      for different pollutants.  For  example, removals by a
                                      particular unit could be nearly 100 percent for floatables,
                                      essentially  zero  for soluble  nutrients, and anywhere
                                      between 5 and 50 percent for total suspended solids
                                      (TSS). Partitioning between dissolved/paniculate fractions,
                                      particle sizes and settling velocity distributions, and
                                      settleable solids characteristics are  all site specific and
                                      extremely important for design. Selection of the treatment
                                      technology to be adopted should consider the pollutants
                                      of concern for  the outfall location and the average
                                      removal the type of unit is  capable of providing.

                                      A receiving water impact analysis can determine a
                                      specific treatment level, which  is used  to define a
                                      treatment level performance goal. An alternate application
                                      of this performance goal  is to specify a  technology-
                                      based  treatment level.  A common  requirement  is that

CSO control provide the equivalent of primary treatment.
Because of the variability of CSO discharge flows, when
this performance goal is applied  a design storm flow
rate also is specified  and treatment units are designed
for flows up to those produced by the design storm.
Reduced performance is accepted for extreme events.

Estimates  of  performance  characteristics  for  many
types of treatment units can be based on experience
with  their performance as unit processes in  a POTW.
The relationship is  not absolute because of differences
between CSOs and sanitary sewage.  The major factor
for CSOs is the intermittent  and highly variable nature
of the influent flows and pollutant concentrations. Other
factors, such as the reported tendency for CSO solids
to have higher settling  velocities and  proportionally
greater quantities  of floatables,   also  may have  an
important bearing on  design details.

Design/performance relationships could also be based
on  units  that treat CSOs.  The  difficulty is  that
performance data  on CSO  treatment units is limited,
and   because  the  characteristics  of  CSOs  vary
substantially, the results for one site may not apply very
well  for other CSO  sites. Table  3-1  summarizes the
range in reported performance for CSO treatment units.
Also, effectively monitoring  intermittent,  short-duration
events, with variable and usually high  flow  rates and
heavy particles, is  difficult and the number of separate

Table 3-1.  Reported Performance of CSO Treatment Technologies
                                  Percent Reduction

Technology                    TSS          BOD
events monitored often is quite limited. Further, results
from individual studies often are reported in such a way
as to make  it difficult or questionable to generalize
sufficiently for transfer to other situations. Because of
these considerations,  the Table 3-1 results  are best
used as a general indication of approximate performance

Treatment units that employ sedimentation are designed
on the  basis of overflow rates  and detention times,
which in turn are derived from particle settling velocities.
A  study of the relationship  between  solids  removal
performance and overflow rates for POTWs  suggests
that  the performance expected  from a  given design
overflow rate may also be affected by  other factors
(WEF, 1992). Since CSO quality and treatability tend to
be highly site specific, the design of treatment units for
CSO control should consider local analysis,  including
bench- and pilot-scale tests where reliable determination
of performance is  required.  Determining the actual
settling  velocity distribution of the solids in the CSO will
provide a more accurate assessment of the appropriate
design overflow rate. Procedures to consider for local
sampling and testing  programs  are further discussed
later in this chapter.
Design of a CSO treatment-type control to meet a mass
removal performance goal must consider two  elements
in combination.  One  is  the efficiency  of  pollutant
Without chemicals
Chemical assisted



General characteristics
General characteristics

U.S. EPA, 1977a
U.S. EPA, 1977a
Cottage Farm site                45            erratic
Cottage Farm site—later data       46            28
Prison Point site                 34            14
Chippewa Falls                  18-70         22-24
Columbus, OH—Whittier St.        15-45         15-35

Swirl Concentrator/Regulator

Washington, DC                 33            83
Decatur, IL                     (0-87)         (0-72)
                              avg 37        avg 34
General data                    40-60         25-60

Microstrainers                   50-95         10-50
Drum Screens                  30-55         10-40
Disc Screens                    10-45          5-20
Static Screens                   5-25          0-20
 General characteristics
 Average data for 1988-1992
 Average data for 1988-1992
 General characteristics
 General characteristics
 Test data8
 Test data"*

 General characteristics'"
U.S. EPA, 1974
MWRA, 1993
MWRA, 1993
U.S. EPA, 1974
U.S. EPA, 1974

Washington, DC, 1992
BGMAandCMT, 1987

U.S. EPA, 1977a
                               U.S. EPA, 1977a
                               U.S. EPA, 1977a
                               U.S. EPA, 1977a
                               U.S. EPA, 1977a
a 18% attributed to flow diversion, 15% to treatment by unit. Average flow was 32% of design peak; observed peak was 70% of design peak.
b No information was provided on removal by diversion or by treatment. Removals stated are presumed to be due to combination of treatment
  and diversion.
0 49 observations during four storm events. Individual range 0-87% TSS and 0-72% BOD reduction.

removal applied to the flows that are delivered to a unit
for treatment,  as discussed above. The other is the
percentage of total flow delivered to and processed by
the unit. When the performance goal is defined as the
equivalent of primary treatment and a design storm size
is specified, both of the required design conditions are
defined.  However,  when  a  treatment  level  goal  is
defined only in  terms of an overall pollutant mass
removal percentage, then the design analysis also must
determine the equivalent of the design storm size to be
used as a basis for the treatment unit hydraulic design.

A trial and error assessment of alternative design sizes
using a SWMM analysis may be required to make this
determination. However, since treatment unit sizing is
significantly influenced by  peak flow rates, a summary
of the frequency pattern of hourly rainfall  intensities,
extracted by analysis of rain gage data, can be helpful
in  identifying the most appropriate range of conditions
to explore by the model analysis. Table 3-2 summarizes
rainfall intensity characteristics  for six locations in the
United States. For each of a range of hourly intensities
(inches/hour) the table lists the following information
derived from a 42-year record and converted to an
average annual basis:

• The upper segment indicates the average number of
  storm events per year that include one or more hours
  in  which the intensity is greater than the intensity
  shown at the end of each row.

• The middle segment breaks down this  total, and
  indicates  that  as the reference  intensity increases,
  the number of "peak" hours  in any event becomes
  increasingly restricted to only one or two individual

• The lower segment lists the  percentage of the total
  rainfall volume contained in all the hours of rainfall
  that have intensities equal  to  or  lower than the
  reference value.

Practical considerations usually restrict SWMM model
runs to limited time spans. Multi-year simulations rarely
are  feasible,  and for  complex  systems  practical
simulation  periods  may be several months. Reliable
estimates of the annual average number of untreated
overflows or the percentage of total volume treated may
not, therefore, be provided by the model simulation and
limit  its ability to suitably evaluate alternate design
sizes.  However, the model output can be  analyzed/
interpreted to  identify the  magnitude of the  flow rates
(and rainfall intensity) that will produce an overflow, or
a peak condition at which the unit will be bypassed. The
information on the overall pattern of rainfall characteristics
listed  in Table 3-2, can  be used to determine the
frequency at which the flagged condition (an overflow
or a bypass) would occur as a long-term average, and
the fraction of the  total volume associated with that
condition. In  cases where the initial design selection
either was inadequate or excessive  in terms  of the
applicable performance goal, the rainfall relationships
indicated in  the table, coupled with the information
derived  from the model  simulation, can  be used  to
provide  guidance on  appropriate adjustments  to the
initially selected design parameters.

Design To Capture First Flush

The concept indicated by the term "first flush," is that in
the early stages of a  storm runoff or combined sewer
overflow event, a relatively small percentage of the total
flow contains a disproportionately large  percentage  of
the total pollutant mass associated with the overall
storm event.

Significant first flush effects are most likely to be present
with  small catchments,  flat  slopes,  low impervious
fractions, relatively simple conveyance system networks,
and lines with low dry weather flow velocities that permit
solids to accumulate in-line. For larger drainage areas,
and complex  piping networks, an array of separate first
flush conditions, may  occur but reach the downstream
location at which control  is to be applied at staggered
intervals that substantially attenuate and blend the
small-scale effects. In  addition, the design and capacity
of  the   conveyance   and  treatment  system  may
effectively retain that portion  of many of the storms in
which  the first flush effect is present.  Regulator
configuration   also  may  influence  first flush  solids
loadings in CSO flows. If the regulator that diverts flow
to the CSO control facility is  a side weir or high-level
outlet, much  of the heavy bed-load of grit associated
with the first flush may not be carried over to the CSO

Monitoring  data can be used to characterize the first
flush. One method of estimating first flush volumes was
employed in the design of first-flush tanks in Decatur,
Illinois  (BGMA and CMT, 1987). A sampling program
established "baseline" average dry weather flow (DWF)
concentrations for biochemical oxygen demand (BOD),
TSS,  and volatile  suspended solids (VSS). The first
flush was defined to start when the CSO  facility influent
concentrations  rose  above  the  baseline  level,  and
continued until the concentrations returned to the baseline
level. The first flush volume was then estimated by plotting
flow on the same axis as the pollutant concentration. A
sample plot is illustrated in Figure 3-4. The shaded area
indicates the  volume of the first flush. The limits of the
shaded area under the flow curve (Q) correspond to the
duration  in  which  the  pollutant  concentration was
greater than the baseline concentration.

A reasonably large event sampling program is desirable
to characterize the potential  significance of first flush
effects at the point  in the system where  a control
measure will be located. Various intensities/durations or

Table 3-2.  Summary of Peak Hourly Rainfall Characteristics
bypass Atlanta,
(in/hr) GA
Average Number of
Average Number of
1 hr
Storms Per Year Having
1 or More Hours With a Greater Intensity
63.6 51 .2
49.1 35.5
27.3 18.9
17.0 11.8
10.9 8.1
6.9 6.0
4.9 3.9
3.1 2.9
1.6 1.9
0.9 1 .3
0.6 0.9
0.5 0.8
for by-
Storms Per Year With Indicated Number of Hours of Greater Intensity
2 hr >2 hr
14.6 28.2
12.1 16.7
7.8 5.5
4.5 1 .8
2.0 0.8
1 .4 0.2

of Total Rainfall Volume

1 hr

2 hr >2 hr
15.1 28.2
11.9 13.8
5.5 3.9
2.9 1.2
1.8 0.6
0.9 0.2
0.5 0.1

With Hours Having
1 hr 2 hr
13.7 13.3
17.9 8.4
12.0 4.5
8.7 2.2
6.5 1.2
4.9 0.9
3.1 0.6
2.4 0.4
1.6 0.3
1.1 0.2
0.8 0.1
0.7 0.1

1 hr
2 hr >2 hr
10.2 29.6
7.8 15.4
3.7 4.0
1.7 1.2
1 .0 0.4
0.4 0.1
0.1 0.1

1 hr 2 hr >2 hr
15.3 11.4 30.1
17.5 90 15.6
13.3 4.8 4.0
8.6 2.6 1.1
6.2 1.3 0.5
4.5 0.6 0.3
2.8 0.3 0.2
1.9 0.2 0.1
1 hr 2 hr >2 hr
19.7 12.9 24.1
15.3 5.9 6.8
4.6 0.7 0.5
Equal or Lower Intensity


        BOD, TSS, or VSS Concentration
               Sample Time - minutes

  291 mg/LtorTSS
  203 rng/L tor VSS

Figure 3-4.  Illustration of use of monitoring data to
          characterize first flush.
more events should be sampled,  because variations
from event to event are expected. In  addition, when
such monitoring may be associated with reduced-scale
pilot testing, an effort should be made to ensure that the
influent samples used to characterize  a first  flush
represent conditions at the location where the full-scale
controls will be installed.

Whether a consistent and significant first flush effect
exists  at  the point  in  the  system at which control
measures are to be applied  is highly site specific, and
only by appropriate monitoring can its presence and
magnitude be determined. However, where appropriate
monitoring can demonstrate and adequately characterize
a first flush, then more cost-effective design of CSO
controls may be possible. Where the first flush effect is
significant and occurs reasonably consistently, some
control units may have smaller design sizes for a given
level  of  mass removal performance  than otherwise
would be the case. The design capacity of storage units
would  have a direct relationship to the presence of a
first flush; however, the design of treatment-type units
cannot be related to first flush in any practical manner,
unless applied in relation to integral equalization storage
that may be incorporated in the design.

A number of existing CSO  facilities are designed to
capture the  more concentrated combined flows that
may occur during the initial stages of a storm event. The
key to designing such a facility is to define the limits of
the first flush. First flush effects vary substantially from
storm  to  storm,  and sampling of combined flows
resulting from a range of storm durations and intensities
is required to indicate the duration of the first flush. For
example,  at  the  Massachusetts  Water Resources
Authority  (MWRA)  Cottage  Farm facility,  sampling
indicated that the concentration of pollutants dropped
off after 2 hours following activation of the facility (U.S.
EPA,  1977b). At  Bannockburn,  Scotland,  sampling
indicated  that  BOD  concentrations  dropped  to  65
percent of peak in 30 minutes, and 50 percent of peak
in 60 minutes (Henderson et al., 1981).
Sizing for the first flush is complicated by the fact that,
depending on the system configurations, part of the first
flush may not even enter the CSO facility. Sampling at
the Bannockburn facility indicated that the initial flow
into the facility had the highest pollutant concentrations.
The "rising leg" of the first flush, therefore,  must have
been  carried down  the  interceptor.  At the  MWRA
Cottage Farm facility, the downstream interceptor was
sized for 2.1  times the  mean  dry weather flow.
Depending on the timing of the storm, the  interceptor
may have available  capacity to carry part of the first

Design Based on Knee-of-the-Curve
The  term  "knee-of-the-curve" reflects  the  fact that
as  control   units  become larger,  the incremental
improvements in performance become progressively
smaller in relation to incremental increases in design
size and cost. This pattern is generally true, but often is
intensified in  the case of CSO control because the size
of increasingly rare events (volumes and flows) grows
logarithmically, and these rare events have an important
influence at the high  end of the performance scale.
The  knee-of-the-curve  technique is  widely used  to
establish the design  size of a specific control measure
or an overall CSO control system, by identifying  the
point at which costs increase disproportionately for only
a marginal improvement in performance. For any site,
cost will be  directly  related to design size, and cost
estimates, developed for units with a range of design
capacities, can be developed  and plotted against  the
corresponding  performance  levels   for   a  formal
knee-of-the curve assessment. Usually some degree of
subjectivity involving best professional judgment is used
to select the "knee," and it may  be  influenced by a
variety of legitimate non-technical factors.

The  knee-of-the-curve concept may be applied to  the
assessment of alternative design  sizes for a particular
control, when a particular type of  control measure has
been  identified  as the preferred approach.  In a case
where  a  CSO  control  assessment  is  made  on a
system-wide  basis  and  includes  consideration   of
alternate technologies as well as design sizes,  a cost
performance  relationship  developed  to examine  the
knee  of  the curve  may result  in  a pattern with
discontinuities,  rather than the  smooth curves that
usually  apply  for  the  assessment of  a  selected

For a simple illustration, consider a case where storage
technology has been  selected for  CSO control. For
either surface storage or tunnel storage, the cost generally
will increase  relatively uniformly as performance level
increases.  However,  space  availability or other site
constraints may place an  upper limit on the maximum
size and hence performance level (percent capture) that
surface storage could provide, so that performance
levels greater than this would require the use of tunnel
storage. Examples of site  constraints other than space
availability that have been  found  to  influence the
placement of CSO control structures include:
• Ownership (including multiple ownership) of site.
• The presence of other utilities on site.
• Historic landmarks.
• Traffic concerns.
• Ground-water conditions.
• Zoning and other land use issues.
• Neighborhood resident  concerns with visual  impact
  of proposed structures and/or odor problems.
For situations where the maximum  size of one  control
technology is constrained, a display of the overall cost
curve showing the relationship over a broad range of
performance levels shows a  discontinuous  step at the
technology break point,  when unit storage costs for
tunnels are significantly greater than for surface storage
units.  Figure  3-5   illustrates  the  cost-performance
relationships for a knee-of-the-curve design assessment.

Data Requirements for Design of CSO
Several classes of  data  are important for designing
CSO control systems or  individual treatment units  so
that they achieve a specified performance goal.
• Information  on the combined  sewage  and CSO
  volumes and flow rates used to design control units
  normally is developed using a simulation model such
Size & Cost
                                    Surface Storage
                               Size Limit for Surface
                               'Storage Based on Available Space
                      Performance •
 Figure 3-5.  Illustration of knee-of-the-curve analysis.
  as SWMM.  Data requirements for this model  are
  discussed below in the section on modeling analysis.

« Although the model can accept all available hourly
  sequences of rainfall contained in a rain gage record,
  it is  rarely practical to utilize the entire record. An
  appropriate year, season, month, or storm event must
  be selected.  Analyzing a rainfall record for the study
  area can provide information to guide this selection.
  Rainfall data sources and  analysis techniques  are
  discussed below in the section on rainfall analysis.

e Most combined sewer  systems are unique. Quality
  characteristics,  including pollutant particle size and
  settling velocity distributions, tend to be site specific.
  Appropriate   CSO  monitoring  and   bench-scale
  treatability tests to establish treatability  parameters
  can  be important for the  design of effective CSO
  controls. This issue is discussed below in the section
  on treatability analysis.

Modeling Analysis

Design of CSO control systems and individual units
requires a determination of the combined sewage flow
rates and volumes produced by rainfall on the drainage
area.  Most  collection and  conveyance systems  are
sufficiently large  and/or  complex  that  computations
using a computer model are necessary. EPA's Storm
Water  Management Model   (SWMM)  is  a common
choice to simulate the behavior of a combined sewer
system and approximate the characteristics  of  the
associated  CSOs. SWMM   simulates  the complex
time-varying physical process of rainfall onto land of
varying characteristics,  the  conversion of rainfall to
runoff, and the  collection  and transport of mixed
stormwater  runoff and sanitary  sewage through  the
collection system.  It can  develop  hydrographs  for
selected  nodes within a combined sewer system. Both
hydraulic and  pollutant routing are performed, and the
model can be  used for both single storm and long-term
(continuous) simulation.

Single storm simulation provides a detailed assessment
of the sewer system and overflow characteristics during
individual storm events. This analysis mode can provide
detailed predictions of flow and pollutant concentrations
and can illustrate the manner in which control strategies
or design alternatives affect these flows and concentrations.
The continuous simulation mode performs a long-term
analysis  based on an annual or seasonal rainfall record,
and is used to develop CSO flow and load statistics for
an  existing system as well as to simulate the effects of
different  control  options. It  provides information to
assess the CSO problem and the cost effectiveness of
abatement options. Both types of simulation are key to
understanding the behavior  of a  combined sewer
system, designing controls, and projecting CSO impacts
on  the receiving wateis.

Detailed  guidance  on  applying  SWMM  or  other
simulation models is provided in separate documents or
by contacting EPA's Center for Exposure Assessment
Modeling (CEAM), Athens, GA. The data required to
perform an analysis  using the  SWMM model  are
discussed below.

Structure and Input Data

For  convenience and  computational  manageability,
SWMM is constructed in modules or "blocks." Included
are modules that manage  information  and  data,  and
modules that perform modeling and evaluation tasks.
The Runoff and Transport Block modules determine the
combined sewer system data requirements  and relate
to the design of CSO controls.
The  Runoff Block is designed  to simulate the quantity
and quality of runoff from a drainage area. It utilizes a
rainfall  input  file  to   produce  hydrographs   and
pollutographs at selected locations in the system. These
serve as inputs to the Transport Block, which routes
them through the combined sewer system.  Input data
for this block includes:

• Meteorological data: Consists of rainfall and snowfall
  amounts, wind speeds, and air temperatures. All data
  should be  local to  the drainage area since  storm
  variations throughout the drainage area can affect the
  patterns of runoff.

• Surface quantity data: Consists of characteristics that
  define the drainage area, which includes information
  on  subcatchment areas, land  slope, sewer  inlet
  locations, infiltration, evaporation, depression storage,
  width  of overland  flow,  and  the  land  surface
  roughness. The model uses these factors (which  may
  be developed through evaluating aerial photographs,
  topographic maps,   field  testing,  or literature) to
  determine the amount and rate of runoff flow.

• Surface  quality  data:  Includes  information  for
  calculating the amount and type of pollutants carried
  by the overland runoff flows. For a fully implemented
  Runoff Block, surface quality data relating  to erosion,
  street dirt  build-up  and  rain water washoff,  street
  cleaning, snowplowing, and catch basin cleaning are
  needed. This  information  can  be  provided  by
  literature, field  data,  and/or through calibration of
  observed pollutant loading data.

The  Runoff Block is  based on segmentation of the
drainage   area   into  an   appropriate  number  of
subcatchments.  The  number is determined by the
features of the drainage area and by  the  layout and
characteristics  of  the collection  system.  Manning's
equation  is used to calculate the maximum rate of
runoff from each subcatchment area.  The maximum
subcatchment runoff is reduced by an infiltration  and
evaporation factor to produce a volumetric loading rate
for use as input to the Transport Block.

The  Transport Block routes inputs from the array of
locations where flows enter the collection system, and
simulates the quantity and quality of flow at locations of
interest. Inputs, generated by the Runoff Block or by
other means, are entered into the network at designated
inlet manholes. The calculated flows and concentrations
define the characteristics  of specific overflows, or may
be applied as input for another SWMM  block, such as
Storage/Treatment,  or for a separate receiving water
quality  model.   Information  required  by  this  block

• Transport data: Describes the physical characteristics
  of  the sewer system, which is conceptualized as  a
  network of conduits connected by non-conduit nodes
  representing manholes or regulators. Conduits are
  described  by  shape or element type, dimensions,
  length, slope,  roughness, and number of adjacent
  lines  (barrels). Transport data are  acquired from
  as-built drawings  and field inspections of the sewer

• Quality data: Describes pertinent characteristics of
  pollutants   the  model  routes  through  the   sewer
  system.  The  required  data  for  each  pollutant
  simulated  includes  first-order  decay  rate, specific
  gravity,  and solids particle size distributions.  These
  parameters allow the Transport Block to determine
  pollutant concentrations  after  decay, scour, and
  deposition processes are accounted for in the model.
• Internal storage data: Input as an individual element
  of the drainage system. Geometry, depth-discharge
  relationships, and initial pollutant concentrations of a
  storage unit are all required data. A storage unit may
  be used to simulate processes such as sedimentation
  or chlorination.

• Infiltration data: Used to assess the amount of flow
  and   pollutants entering  the  sewer  system from
  leaking pipes  and joints.  Infiltration flow may come
  from ground water, rain water, residual moisture, and
  base dry  weather infiltration.  Monthly degree day
  data  is used to determine infiltration from residual
  moisture. These data are  most useful when obtained
  directly from the  study area. Historical data from  a
  near-by study area, estimates by local professionals,
  and  estimates based on  country-wide observations
  are  useful alternatives when  direct  data are not

Additional Transport Block  data  requirements  include
subarea identification parameters, observed flow data,
and flow estimating  data.  Subareas are subdivisions of
the  drainage area identified  according  to  zoning
classifications.  Subareas  may  not  correspond  to
subcatchments defined in the Runoff module. Land use,

subarea  acreage, water  usage, population  density,
dwelling unit density, housing values, average income,
and process flows are data  required by the Transport
Block to  estimate  average  daily sewage flow  and
pollutant concentrations. This average sewage flow also
is  corrected for hourly and  daily fluctuations in dry
weather flows and concentrations. The Transport Block
includes a dry-weather flow model that may be used to
estimate  both  sanitary  sewage flow  and  pollutant
concentrations from the drainage area.

Transport of wet- and dry-weather flow by the Transport
Block is based on representation of flow as a kinematic
wave.   Disturbances   are  propagated  only   in  the
downstream direction and backwater effects are limited
to only  a single conduit. Surcharging  is modeled by
storing excess  flow at the upstream non-conduit until
capacity is freed to accept the stored flow. Equations for
gradually  varied, unsteady flow are used by Transport
to calculate the velocity  and depth of flow along the
sewer system.  This procedure is not appropriate for
analysis  of   system  segments  that  consistently
experience significant surcharge conditions.  In  such
cases, the Extran (Extended Transport)  Block must be

Practical Application and Calibration

Calibration of the model is the only practical way to
assure  that the substantial array of input parameter
values  assigned  accurately represents  the  system
being modeled. Model calibration consists of adjusting
selected input parameters so that the computed output
matches observed  monitoring results at key points in
the system. On completion,  a calibrated model should
provide a good match between predicted and observed
results,  when applied to data sets that were not used in
the calibration  step. When this  criterion  is met,  the
model is considered validated or verified.

Model calibration  requires  synoptic rainfall  and  flow
measurements for a number of storm events. Rainfall
data  should  be  obtained  from one  or more gages
located in or close to the drainage area. Flow monitoring
should be conducted at representative locations in the
system. Monitoring overflows at the largest CSO points
in systems with a number of different CSO locations is
advisable. Even the best model predictions  are  only
approximations of reality, and emphasis on the largest
CSOs will help to secure the best accuracy for the most
significant discharges,  and  thus the  best overall
estimate  of CSO volumes and pollutant loads for the
system as a whole. In addition to the overflows, flow
monitoring stations should be located at one or  more
points  in the  collection  system  upstream  of  the
regulators. This is necessary  to calibrate the model
parameters defining the drainage  area features that
control the conversion of rainfall to flows in the collection

An adequate number of storm events (usually 5 to 10)
should be monitored and used in the calibration. This is
important because of possible malfunctions in different
flow sensors at different times, and also to compensate
for variations in  rainfall distribution over the drainage
area. The point rainfall measurement at the rain gage
provides only an approximation of the distribution over
the entire drainage area, so  variations  in individual
events should be expected.

The availability of a sufficient number of different storm
events  also  will facilitate and  improve  the  model
calibration. Many users prefer to base the calibration on
the model's ability to reproduce the overall results from
a number of different  events, rather than attempt to
match instantaneous flow measurements.

Flow measurements almost always will be made using
automatic sensing units, but it is important to provide
operator inspection  and supervision.  Sensors tend to
clog and generate erroneous readings,  which  can
seriously  impede model  calibration efforts  if  left

In addition to timely maintenance of  the flow meters,
operator supervision also can assure that the location
being monitored is not surcharged at the time  of flow
measurement. The existence of a surcharged condition
is essential  information for reliable application of the
simulation model.

For lines that surcharge, the Extran Block must be used,
rather than  the normal  Transport Block. This adds
complexity to the model analysis, not so much in the
computations, but in  the  model's sensitivity  to the
accuracy of the input parameters assigned.  Where
Extran is employed, a  better definition of ground truth
(pipe diameters, slopes, lengths, etc.) is required.

Applying the model is  more cumbersome for systems
with many regulators  and overflow points.  Modeling
large,  complex  systems  also   imposes  significant
demands on available  computer memory for input and
output files.  This is a serious consideration for model
analyses performed on  PC  platforms.  The practical
effect is to impose a limit on the length of rainfall record
that can be analyzed.

For large or  complex systems, practical maximum
record lengths are a year or less,  and in some cases
could be less than a  6-month simulation. When the
period analyzed is short, it is  important  that the most
appropriate  period of the rain record be selected for
analysis,  so that the  model  provides representative
projections  of   system   or  control   performance.
Independent analysis of rain data can assist in selecting
an appropriate portion  of the overall rainfall record.

Many  applications  of  the  SWMM  model  do  not
independently compute  the  surface runoff  quality  as
above, and then combine it  with the sanitary sewage
quality characteristics to define the quality of the CSOs.
A common approach  is to directly  assign  values  for
quality characteristics of the combined sewage and
CSOs, based on local  CSO monitoring data or the
quality of wet-weather inflows to the POTW. When  an
average concentration value is assigned for all storms
(or a concentration correlated with storm size when the
data  so indicates),  the  model  will   not  reproduce
individual  event quality, but   provides  acceptable
estimates of pollutant loads averaged over the analysis
period.  Many analysts believe that such projections
have   equal  or  better  reliability,  because  of  the
uncertainties associated with the use of the build-up/
wash-off model  routines and the difficulty  commonly
encountered in calibrating this model component.

Rainfall Analysis

Precipitation  is  the driving  force that mobilizes and
transports pollutants via CSOs to receiving  waters.
Evaluating the precipitation characteristics of an area
can assist in analyzing issues such as the estimation of
CSO  pollutant loads, the water quality impacts they
produce, and the assessment  of control  strategies.
When the basic analysis is performed using a simulation
model such as SWMM, separate analysis of the rainfall
record can  assist in  selecting the most  appropriate
rainfall periods or events to use in the model analysis.

Rainfall is highly variable, and excursions of individual
events above typical or  average values of parameters
such   as  intensity, duration,  and  volume,  can  be
considerable.  The greater the  magnitude of rainfall
volume or intensity, the more larely it occurs. Statistical
analysis of  rainfall records can define the frequency
distribution (or probability of occurrence) of magnitudes
of the various rainfall parameters.  Such probabilities
often   are  expressed in  terms of return  period  or
recurrence interval.

Hourly rainfall records  for rain gages in the United
States  are  available  from the  National  Weather
Service's  (NWS's)  National  Climatic Data  Center
(NCDC)  in Asheville, North Carolina.  Each record is
identified by a unique 6-digit number,  consisting of a
2-digit state code followed by a 4-digit gage number.
The information  in a gage record includes the location,
latitude and longitude, elevation of the measurement
site, and the depth of hourly rainfall recorded (in 0.01
inches). The date and hour are recorded for each depth
in the record.

The NWS  generates regional summary reports from
time to time, but for all NWS gages, the basic data are
readily available in  electronic  format from both the
NCDC and  commercial sources. The  electronic data
include  records from  both  previously and currently
operating  stations in the 50  states. These records
include  daily rainfall  data  for over  25,000 stations,
hourly data for over 5,500 stations, and 15-minute data
from  over  2,700 stations.  Approximately 8,000  rain
gages currently  operate across  the country.  The
increased use  of personal computers over the past
decade  has allowed  interested parties to  conveniently
obtain and  analyze  the  actual data  for a  particular
location, avoiding the  need  to  rely on  published

A long-term sequence of  rainfall data can be analyzed
in  a  number of different ways to develop relatively
concise characterizations, which  may then be used
for  engineering  purposes.  Common  methods   for
characterizing  rainfall  include  total  volumes,  event
statistics,  return period/volume curves, and intensity-
duration frequency curves.  Each method is described

Total Volumes

A  common basis for  describing  rainfall is the total
volume  of  rainfall  occurring  each  year. The NWS
publishes annual totals as well as deviations from the
average for each rain gage in its network.  Definitions of
wet- and dry-year rainfall  can be made by comparing a
particular year's rainfall to  the  long-term average.
Monthly totals and averages also can be computed in
the  same  way  to  examine  seasonal  differences.
Evaluating  annual  or seasonal  rainfall totals is  a
common basis for selecting a specific time period to use
in detailed simulation modeling.

Event Statistics

Information may be developed on the characteristics of
individual storm "events"  for a site. If the sequence of
hourly volumes is grouped into separate  events, then
each event  may be  characterized  by  its duration,
volume,  average intensity,  and  the  time  interval
between successive events. The event  data can be
analyzed   using  standard   statistical procedures to
determine the mean and standard deviation, as well as
probability distributions and recurrence  intervals.  A
computer  program,  "SYNOP," performs  a  statistical
analysis of the rainfall data in a NWS record (U.S. EPA,
1989).  It segregates the  hourly rainfall values  into
independent  storm   events,  and  determines  the
parameters of  each  event (volume, duration, average
intensity, and interval since the preceding storm).  The
array of individual event values is analyzed to determine
the  mean and standard deviation  or  coefficient of
variation for selected  stratifications  of the data  set,
producing results sorted by month and by year, as well
as  for  the entire period of record. Outputs  include
information on the frequency distribution and recurrence

period for the event parameter values. This report also
presents summaries of rainfall statistics developed for
rain gages in different areas of the country.

Return  Period/Volume Curves

The frequency of occurrence for a given magnitude of
a storm event parameter such as volume may be shown
as a plot of its probability distribution, as illustrated in
Figure 3-6. The plot indicates that approximately 10
percent  (90th percentile) of the  storm events in the
42-year  record used for the analysis, deposited about
1.5 inches or more of rain. If the statistical  analysis
indicates an average of 60 storm events per year, this
would mean that  six  storms per year had a rainfall
volume equal or greater than 1.5 inches. Furthermore,
six events per year averages one event per 2 months,
so the 1.5-inch rain event could be characterized as the
storm-event volume with a 2-month return period, or the
"2-month storm."

Note  that this 2-month storm definition is not related to
duration; the  252 storms  equal or  greater than  1.5
inches during the 42-year period (10 percent x 42 year
x 60  per year) would  have a range of durations. The
return  period-volume  relationship  is  different than
corresponding relationships that are  based on rainfall
intensity-duration  (e.g., 1-year, 6-hour storm) curves
provided by other design storm determinations. The
relationships from a storm  "evenf analysis are more
appropriate for the assessment of storage requirements
because events, by  definition, provide storm volumes
that are typically followed by dry periods that average
several days, during which a  CSO storage  unit is

The specific values for storm parameters vary regionally,
but the event summary and distribution plot (Figure 3-6)
illustrates the  nature  of the  storm event  statistics
produced by any rain  gage. Table 3-3,  which lists
summary event statistics for rain  gages  in different
regions of  the country, illustrates  both  the general
similarity as well as the regional differences  in rainfall

Intensity-Duration Frequency Curves
Curves of this nature are another way of characterizing
the  variable  rainfall  at a  site.  A typical  set  of
intensity-duration frequency curves  is  illustrated  by
Figure  3-7.  For this  record, a  60-minute duration
(1-hour)  rainfall intensity of about  1  inch/hour occurs
once every year. One-hour intensities that are higher
than this occur more rarely (e.g., 2 inches/hour every
10 years).  Note also that rainfall with an average
intensity of 1 inch/hour also may occur over longer
durations, at less frequent recurrence intervals.

A major use of these families of rainfall characterization
curves is to design hydraulic structures (e.g., storm
drains, culverts), where short-duration peak flows must
                                           Storm Event  Volumes
                                Louisville,  KY
                                    N = 2910
                                Median = 0.42 inch
                                  cov = 0.99
                                 Mean = 0.61 inch
                                   1     5  10  20 30   50   70 80  90 95

                                          Percent of Storms Equal or Less Than

 Figure 3-6.  Probability distribution of storm event volumes.
99.9  99.99

Table 3-3.  Rainfall Characteristics for Six U.S. Locations3
                    Annual Averages
Event Rainfall Statistics Mean Storm Event
                                                                                       Storm Event Volumes
                                                                                       for Three Recurrence
                                                                                          Intervals'1 (in)
Number Rain
City of Stormsb Volume (in)
Atlanta, GA
Louisville, KY
Portland, ME
Newark, NJ
Chicago, IL
Portland, OR
a Based on 42 years of records, from 1949 through 1990.
Storm events greater than 0.1 inches with a minimum of 6 dry hours to separate storm events.
c Delta is the average interval between the midpoint of storm events.
The tabulation of recurrence interval volumes indicates rainfall volumes for events that recur on
average 3-,
Return Period
3 mo 6 mo 1 yr
6-, and
12-month intervals.


$    ~
                                    3 YR. STORU

                                       10 YR. STORM
                \ YR. STORU

                       2 YR. STORU

                   1  IN/HR.-25.4 mm/hr.
                         10                  100  200

                    DURATION (MIN.)

Rgure 3-7.  Illustration of intensity-duration curve (Moffa, 1990).

be considered to avoid local flooding. They also may be
applied to the design of CSO treatment units where
peak flow is a relevant design parameter.

These curves are developed by analyzing an hourly
rainfall record in such a way as to compute a running
sum  of volumes  for consecutive hours equal to the
duration of interest. The set of volumes for that duration
are then rank ordered, and based on the length in years
of the record, the recurrence interval for any rank/value
is determined. This rainfall analysis procedure is used
to calculate the local value for a design storm such as
a 1-year, 6-hour design condition.

Treatability Tests

CSO treatment technologies such as sedimentation
basins and swirl/vortex units remove particulate pollutants
           by employing design parameters and unit configurations
           that  enhance  the  separation  of solids  by  settling.
           Therefore, the settling velocity  of the pollutants in the
           CSOs  to be treated  has an important influence  on
           performance and should be considered in developing
           design details.

           Bench-scale tests can be employed to determine the
           typical  range of CSO settling velocities. A  number of
           variations of the  procedure and  equipment are used for
           these tests. Figure 3-8 is a schematic illustration of several
           settling test column designs. Test procedures vary with the
           design of the test apparatus  and are described briefly
           below for the units illustrated. Results obtained from either
           procedure can be analyzed to determine the probability
           distribution of particle settling velocities in the CSO sample
           tested, and displayed as illustrated on the plot shown in
           the lower portion of Figure 3-8.

           For  the  8-in-diameter  by 5-ft-tall test column with
           multiple sample ports, the column is filled to the top with
           a sample of the water to be tested, and then stirred to
           distribute the settleable solids uniformly throughout the
           column. A sample at time zero is analyzed to establish
           initial conditions. Samples are then withdrawn from
           each port at selected time intervals. They are analyzed
           to determine the concentration  of TSS that remains at
           the sample location after the elapsed time interval.

           For the other test unit,  clean water is  used to  fill the
           column up to the level of the butterfly valve at the upper
           end. A 500-ml sample is then added at the top of the
           cylinder. After the sample has been properly stirred, the
           butterfly valve is opened and particles in the sample
           settle through the body  of clean water.  Water samples
           containing settled suspended solids are collected from
           the bottom cone at  various elapsed times ranging from
           1 to  60 minutes, and analyzed for TSS. For  a 60-cm
           settling distance,  the  test  results will  provide the

        (J) 65mm




                                                                         -Sample 500ml
                                                                         • Butterfly valve

                                                                       — Plexiglass cylinder
                                                                         - Clean water
                                                                       i —Sample collection point
                                TYPICAL SETTLING COLUMNS
                       Initial TSS = -165 mg/L
Rutland, Vermont CSO
                     99.9    99     95  90  80      50      20   10  5
                                     PERCENT EQUAL OR GREATER
Figure 3-8.  CSO settling velocity—typical test equipment and sample results.

distribution of particle settling velocities in the range of
1 to 60 cm/min (2 to 120 ft/hr).

TSS normally is selected as the pollutant analyzed, but
other pollutants of concern also could be selected to provide
information on the settleability of BOD, for example.

A number of different approaches have been used to
analyze, summarize, and interpret settling column test
results.  A  procedure  that  provides  useful  results
recognizes that particulates are  present in combined
sewer overflows in a range of particle sizes and settling
velocities,  and that it is appropriate to characterize
settling  properties in  terms of a  settling  velocity
distribution. Such  distributions define the  cumulative
percentages of solids having settling velocities greater
than a specified value. Test data can be analyzed to
develop this information by following these steps:

1. For each measured concentration (each sample port
   and sample time),  the  difference  between  the
   concentration  and  the  amount present initially is
   used to compute a percent removal.

2. Since each combination of depth to a sample point,
   and elapsed time between the start of the test and
   the time the sample is taken, reflects  a specific
   distance settled in a  specific interval of time, each
   sample corresponds to a settling velocity  (ft/hr).
   Each such  value corresponds  to a percent removed
   value computed from the measured concentration.
   The results can be interpreted as the percentage of
   TSS in the sample that has settling velocities equal
   to  or greater  than the value represented by the
   settling distance and sample time.
3. The results then are plotted on graph paper with a
   probability scale to indicate the frequency distribution
   of pollutant settling velocities in the sample. CSO
   settling velocity characteristics developed in  this
   manner are shown in Figure 3-8.

Test results can  be displayed on  probability plots to
illustrate  the   range  in  settling  velocities  of  the
suspended solids (or other pollutants) in  the CSOs.
Results from  different samples at a  particular site or
from  different  CSOs  also can  be summarized  and
compared by listing the median (50 percent) and/or the
90th percentile settling velocity.

For example, settling tests conducted on CSO samples
at Washington, DC, indicated that for 11 samples during
eight  events,  the  median  settling  velocity  ranged
between about 1 and 30 ft/hr. For these same samples,
10 percent of the particles had settling velocities in the
range 10 to 300 ft/hr (Washington,  DC,  1992).

Median CSO settling velocities at other locations were
measured  in the range of  5 to 50 ft/hr in one city and
estimated from particle settling velocity data to be in the
range 50 to 100 ft/hr in another (Dufresne-Henry, Inc.,
1992). Both of these results can be considered to fall in
the same general range as the Washington data, given
the inherent variability. Several tests on New York City
CSOs, in contrast, provided median settling velocities
in the range of 1 to 3 ft/hr (NYC DEP, 1993). However,
the CSO quality at this location is suspected  to be
significantly influenced by ground-water infiltration.

The variability in available data from different locations
emphasizes the importance of local testing to determine
design parameters for CSO controls that are based on
sedimentation. The significant variability exhibited by
different samples at  the same  site indicates the
importance of (a) testing a sufficient number of samples
(at least 6 to 12 in most cases) to define a representative
range of conditions for the site, and (b) testing samples
from representative parts of a storm.

The latter is particularly important if the overflow has a
significant first flush effect, because particle sizes and
settling  velocity  are significantly different in different
parts of the overflow event. Part of the wide variation in
median settling velocities  in the Washington, DC, data
set is attributed to this factor.

Care is required in collecting samples used in settling
column tests to ensure that results provide representative
settling  characteristics  of the flows  to be treated.
Samples should be collected from an appropriate point
in the collection/overflow system. In addition, sampling
equipment  and  procedures should  be checked  to
assure that the sample itself properly collects the solids
that are present. Sample intake locations and  intake
velocities must be established  to capture the heavier
solids present in most CSO flows.
Settling velocity  test results can guide design decisions
by  determining  the removal  efficiency that a CSO
control  unit  will provide for  particles with selected
settling velocities, when the unit processes CSO flows at
the design  flow rate (or other applied rates of interest).

For sedimentation basins,  a  relationship  between
steady  state removal  efficiency  and  the  hydraulic
loading rate and particle settling velocity is provided by
the following equation (Fair and Geyer, 1954).
   R = fraction of initial solids removed (R x 100 =
       percent removal)
   n = a parameter that provides a measure of the
       degree of turbulence or short-circuiting, which
       tends to reduce removal efficiency
   vs = settling velocity of particles
 Q/A = rate of applied flow divided by surface area of
       basin (an  "overflow velocity," often designated
       the overflow rate)

The upper plot in Figure 3-9 illustrates the relationship
between the  design  size  of  a sedimentation basin
(expressed as an overflow rate, gpm/ft2 or ft/hr) and
removal  efficiencies for  a range  of particle  settling
velocities, as  provided by the above  equation  (U.S.
EPA, 1986).

For  swirl/vortex  units,   the  design   capacity  and
performance do not relate directly to an overflow rate
based on surface area, as is the case for sedimentation
basins. An indication of the relationships of interest as
they apply to  a swirl unit based on the EPA design is
provided by the lower plot in  Figure 3-9. This plot is
adapted from  information developed from a series of
experiments in  Sweden  on a 1.17-m  diameter swirl
concentrator,  supplemented with  data from  two  full
actual operating concentrators (Lygren and  Damhaug,
Figure 3-9 summarizes the interrelationship between
removal efficiency,  applied  flow  rate,  and  particle
settling velocity.  Comparing these  relationships with
local test data that defines the  distribution  of settling
velocities in the CSO flows to be treated,  will permit
estimates  of  expected   performance  relative  to  a
particular size fraction at any  design flow selected for
evaluation. In  addition, overall performance estimates
for a selected  unit design size  (over the range of
projected  CSO  flows)   may   be  developed   by
incorporating  the  relationships in  a simulation model
(e.g., the Storage-Treatment Block of SWMM). In  this
case, the continuous distribution of settling velocities
can  be  incorporated by dividing  the  distribution  into
three to five size fractions and  assigning an  average
settling velocity  to each.  The  performance  analysis
would be applied independently to each of the fractions.
Design  decisions  could  then  be  based  either  on
projected results for a selected size fraction, or on the
combined results for all fractions.

Technology Selection

Factors That Influence Technology Selection

Selecting a specific control technology, or as  is often the
case, the combinations of technologies to be incorporated
in a comprehensive CSO control program, requires an
engineering  evaluation  that  considers a  variety of
elements. The following elements usually have a major
influence on technology selection:

• The  layout and  hydraulic characteristics  of  the
   combined sewer system and overflow points.

• The  number and  size of  overflow locations,  the
   distance  they  are  separated, and the  feasibility of
   consolidating multiple overflows  at a point at which a
   control technology  can  be applied.
• The  magnitude (volumes, peak flow rates) that will
  occur at individual overflow control  locations, which
  influences  the   size   requirements   for  specific

• Available space and other siting constraints in the
  vicinity of CSO locations.

• The  performance goal that is  applicable  for the
  system  as a whole and/or for specific  overflow

• The  ability of  a particular technology to  provide
  effective control of  specific pollutants  that are
  contributing to designated use impairment.

Such  a  comprehensive  evaluation,  addressing the
system as a whole and  identifying control needs  at
specific CSO locations, and site features that influence
the applicability of individual technologies, is essential
to develop an effective CSO control program. Surveys
are necessary to establish characteristics of the study
area that  are  pertinent  to technology  selection.  In
addition, the  hydraulic characteristics  of the combined
sewer/CSO  system  must  be  characterized   by  a
monitoring/modeling  effort.   Performance/applicability
factors that relate to individual control technologies also
must be evaluated.

The principal performance factors that  relate to the
applicability of the individual CSO control technologies
addressed in Chapter 4 are discussed below.

In-System Controls/ln-Line Storage

All CSO control programs should incorporate in-system
controls  appropriate  for the features  of  the local
combined sewer system. Both the  measures that are
applicable and the degree of control  they will provide
are variable and site-specific. However, the increase in
in-line storage and reduction  in overflows produced by
in-system controls generally will be cost effective, since
they represent modifications of existing facilities. Many
of the  applicable measures will require relatively minor
engineering and cost commitments.

To the extent that the existing collection system can be
adapted to increase the retention of wet weather flows,
it will reduce the required design capacity and cost of
other control technologies that must  be applied. This
technology, therefore, is essential and  should be a
component of all control plans.

Increases  in the utilization of the  available  hydraulic
capacity of the existing  combined sewer  system for
containing wet weather flows will  increase the risk of
street  and basement flooding. This situation must be
carefully considered in establishing the maximum in-line
storage that can be safely achieved, and in many cases
will  require  a  model analysis  to define  acceptable
conditions. Even with model projections as a  guide,

                                   SEDIMENTATION   BASIN
                                       adapted from Rg 7 [EPA 1986]
                100  I	-^
                 40  -;
                                         Hydraulic Loading Rate
                                            gal/day /sq ft
                                         for short-circuit factor n > 3
                                                                     NOTE : 1 cm/sec = 118 Whr
                                   50             100           '150
                                 Particle Settling Velocity (feet per hour)

                                    SWIRL   CONCENTRATOR
                                        adapted from Fig 2 [Lygren 1986]
                                                                     NOTE : 1 cm/sec = 118 ft/hr
                                    50            100            150
                                 Particle Settling Velocity  (feet per hour)
Figure 3-9.  Effect of particle settling velocity and hydraulic loading rate on TSS removal efficiency.

weirs should  be raised in incremental  stages with
suitable  follow-up  monitoring.  Increased collection
system  storage also will  increase the  potential for
accumulating   sediment  deposits.  Operation  and
maintenance efforts can increase for the effective
application of  in-system controls, because effective
performance relies on continuing reliable operation of
regulators, tide gates,  and  pump stations,  and on
preventing sediment deposits  from  accumulating to
problem levels.


Storage technology is  probably the  most favorable
control technology to consider as a general rule. For the
volumes captured at a CSO location, storage removes
all  pollutants  (pathogens,  floatable   and dissolved
pollutants, solids) from discharge at that location. When
captured flows are delivered to the POTW, treatment is
provided at that point and discharge occurs at the plant
outfall.  For discharges at the overflow  point produced
by   storms   in  excess  of  the storage  capacity,
sedimentation in the storage unit provides a degree of
treatment, and where disinfection may  be required, all
or part of the storage volume may be utilized to provide
contact time.
However, despite  the   theoretical attractiveness of
storage technology, a  number of  factors  limit  its
applicability. For relatively high levels of control, storage
volumes  are large, with corresponding  costs. Cost
effectiveness may be very poor if storage technology is
applied separately to  multiple CSO  locations,  or to
CSOs with relatively small discharges. Consolidation of
a number of existing CSOs for treatment by a single
storage unit may involve substantial modifications to the
collection system  and  significant disruption of the
service area during installation. Storage units are often
difficult to site because of the need for adequate space
and other constraints imposed by land use in the area.

When all CSO flows captured by storage are returned
to the interceptor for conveyance to and treatment by
an  existing  POTW, the  treatment  plant  loadings
increase, as do the solids residual and sludge handling
requirements. The  ability of the POTW to respond  to
changes  in  operation  required to accommodate the
implementation of  CSO storage technology  must be
adequately addressed by the control program studies.

These factors apply to both surface and tunnel storage
technologies. Selection factors specific to each of these
technologies are discussed below.

Off-Line Near-Surface Storage/Sedimentation

This technology balances control provided by storage
to  capture  CSOs and  return them to the POTW for
treatment and direct discharge. It may limit the furnished
storage to a capacity that is feasible based on local site
constraints and cost considerations. However, for CSO
volumes  that  are  not  captured, it provides effective
sedimentation  treatment by utilizing efficient settling
basin design features. This control technology provides
effective reductions in suspended solids and associated
pollutants, is very effective for control of floatables, and
provides  contact time for  disinfection when control of
bacterial  levels is required. CSO flows captured by
storage will result in an  increase in flows and solids that
must be processed by the POTW.

While less costly  than tunnel  storage up to  certain
storage requirements, surface storage often is difficult
to site.  Combination  surface  storage/sedimentation
designs may permit more efficient utilization of available
space and optimize the level of control that can be
applied at a particular location. The overall performance
level that can  be achieved depends on the proportion
of the CSOs  that are captured and  returned  to the
POTW versus treated by sedimentation and discharged
at the overflow location.
The  sedimentation unit should  produce reductions in
pollutants  such  as   suspended  solids  and   BOD
equivalent to  primary  treatment. This control  should
completely remove floatables and effectively disinfect
for bacteria.

Deep Tunnel  Storage

Deep tunnels  are the most costly storage capacity, but
they are  less constrained  by considerations of surface
space availability  and construction  impacts   to the
community. Large storage volumes are more  readily
effected than in the case of surface storage, and tunnel
designs  also  can  provide  conveyance  of captured
volumes to the eventual treatment/discharge location.

Geotechnical  conditions  influence the feasibility of
applying this technology at any site, and consolidation
and  other collection system modifications may require
significant effort to deliver CSOs to the tunnel. This will
be site specific and depend on the tunnel alignment and
its location with respect to the existing overflows. As
with any storage technology, the ability of the POTW to
process  the captured flows must be considered in the
technology selection process. This may have particular
importance in the  case of tunnel storage, because of
the ability to capture very  large  percentages of the wet
weather  flows.

In low-lying areas with little topographical  relief, using
tunnels  to  control CSO may  serve the additional
purpose  of reducing flooding problems.

Coarse Screening

The  control provided  by  this technology is limited to
removal  of large  objects, and its principal use  is to

protect downstream  control units. It is addressed  in
Chapter 4 because  it is commonly included with a
variety of control technologies as part of the overall
control system.

Other than the removal of debris and floatable materials,
coarse screens accomplish little in the way of removal
of other pollutants. As such, they would  not normally be
considered as a stand-alone CSO control. There may
be exceptions, as in a case of relatively small CSOs
discharging to  segments  of water  bodies  where
aesthetic  appearance and floatables  are  the only
significant water quality impact.

Proper maintenance  and frequent cleaning of these
screens is important  both to ensure effective operation
and  to avoid  restrictions to flow, which might induce
upstream flooding.

Swirl/Vortex Technologies

These controls are attractive primarily because of their
relatively low cost and space requirements in relation to
the degree of treatment they can provide. They are
applied at individual overflow points to provide some
level of treatment before  direct discharge.

They provide effective removal of grit and heavy solids,
and  are quite effective  for control of floatables. For
settleable solids and BOD, the removal capabilities
appear to range from nominal to moderately low, and
performance is influenced to a major degree by  the
site-specific paniculate  settling  characteristics of  the
CSOs to be treated.  Pollutant removal is provided by a
combination of  solids separation  from the flows that
discharge and of the underflow that is returned to  the
combined sewer system. The latter often accounts for
a significant fraction of the  overall pollutant removals.
As a result, appropriate bench testing to define settling
velocities usually is  necessary to determine the basic
applicability of  this  technology at the site and  the
appropriate design parameters.

The very short residence time resulting from the typical
high rate  design limits the ability to provide disinfection
contact time, and for locations where bacterial control
is required, additional storage/contact chambers  are
usually incorporated  in the system design.


In all cases where CSO control must include reductions
in  bacterial  levels   to  protect  a  designated use,
disinfection technology must be applied. It essentially
will  never be  considered independent of  any other
control technology, because bacterial contamination is
not the exclusive water  quality issue for CSO control
and   because  pretreatment  to   reduce  solids  that
otherwise shield the  bacteria normally is necessary for
effective disinfection.
Adequate contact time  must be  provided, usually in
tanks associated with the control technology applied at
the overflow location.  However,  in cases  where the
outfall lines are sufficiently long to provide necessary
contact, they may be used.  For  control technologies
involving  storage  or  sedimentation  treatment,  the
storage volumes usually are sufficient to provide the
required disinfection contact time.  For CSO control
technologies such as  screens or  swirl/vortex units,
which have very short  residence  times, separate
provisions  must  be made  to  provide contact  for

Optimizing Storage-Treatment CSO Controls

For CSO control systems that utilize a combination of
storage and treatment methods, a  cost-optimization
analysis should be performed to determine the most
cost-effective combination of design sizes. For a given
performance level, the design capacity, and hence cost
of a treatment unit such as a sedimentation device, can
be reduced by using a storage unit that absorbs the
higher flows during parts of the storm, and delivers flow
to the treatment unit  at both a lower, averaged rate and
at a  more uniform influent rate.

The  upper plot in Figure 3-10 illustrates the concept by
indicating the array of combinations of storage capacity
and design capacity for a treatment unit. Large amounts
of equalization storage permit the installation of small
treatment units. Conversely, when little or no storage is
provided to capture  the short  duration peaks of  the
variable CSO flows, a  very  large treatment capacity
must be provided to avoid either degraded performance
due  to excessive applied rates, or the need to bypass
the treatment unit. When this relationship is developed
for a particular set of site conditions  (using a SWMM
model analysis or other suitable  procedure), and  the
relationship between increasing  size and increasing
cost is applied for both  storage and treatment units, a
summary relationship such  as that illustrated  by  the
lower plot  in Figure  3-9 can  be  developed. This
indicates the relationship between the reduction in  the
costs for a treatment unit as smaller design capacities
become applicable and the corresponding use of larger
and  more  costly storage  capacities. Total  cost of  the
combination system, which is the sum of the individual
unit   costs,  passes  through   a  minimum,  which
represents the most  cost-effective design combination.

The concept applies directly to the design of a control
system to be installed at a particular  CSO overflow
point, and  relates most  closely  to the off-line near-
surface storage/sedimentation technology discussed in
Chapter  4.  It can be  applied to the  design  of
disinfection facilities by associating  chemical dosage
and/or  type with  the  "treatment"  element,  and  the

                            All combinations provide (he
                            same operating capacity
                            (or the treatment unit and
                            hence the same performance
                 Required Treatment Unit Size
          Physical Size of Storage Device and Treatment Unit

      Figure 3-10.  Cost optimization of storage-treatment.

      contact time required with the "storage" element of the
      optimization analysis.

      The optimization concept generally is not applicable in
      the case of swirl/vortex technologies, because providing
      storage upstream of the inlet to these devices usually
      is considered inappropriate.

      Theoretically, the analysis also could be extended to the
      case where the control approach  is based on storage
      at overflow points, but where treatment at the POTW
      follows pump-back or bleed-back of captured overflows.
      Direct application of a cost optimization analysis in this
      case is complicated by several factors. For example, in
      most cases the capacity of the treatment units is already
      established by the existing facility design and  both the
      existing dry  weather flows in the  system  and the
      possibility that captured overflows from a number of
      overflow points may be involved. In addition, federal or
      state   regulations  may  impose requirements  for
      performance of  a storage unit  (such  as number of
      overflows  permitted or minimum percent capture) that
      may conflict with the strict cost optimization of an overall
      system that relies on maximizing the available treatment
      capacity of the POTW.

      However, in cases where the collection  system has the
      hydraulic capacity to deliver wet weather flows at rates
greater than a POTW can process, and the design of
additional parallel primary treatment facilities for these
flows  is being  considered,  using  an  optimization
analysis would increase the cost effectiveness of the

When an NTIS  number is cited in a reference, that
document is available from:
  National Technical  Information Service
  5285 Port Royal Road
  Springfield, VA 22161
BGMA and CMT. 1987. Bainbridge,  Gee, Milanski &
  Associates, Inc., and Crawford, Murphy & Tilly,  Inc.
  Sanitary district of Decatur combined sewer overflow
  operational study.
Dufresne-Henry,  Inc. 1992. Report on  Rutland CSO:
  basis of design. June.
Fair, G.M., and  J.C. Geyer. 1954. Water supply  and
  waste water disposal. John Wiley and Sons.
Henderson, R.J.A., W. McBain, and F. Pettigrew. 1981.
  The performance of an off-sewer storm-sewage tank.
  Water Poll. Control.

Lygren,  Elvind,  and  T.  Damhaug.  1986.  The  swirl
  concentrator as an urban runoff treatment device. In:
  Torno, H.C., J. Marsalek, and M. Desbordes, eds.,
  NATO ASI series vol. G10, urban runoff pollution.
Mahood, T.W., and G. Zukovs. 1993. A comprehensive
  stream  study.  In:  James,  William,   ed.,  New
  techniques  for  modeling  the  management  of
  stormwater quality impacts. Boca Raton, FL: Lewis
Moffa, RE.  1990. Control  treatment of contaminated
  sewer overflows. New York, NY: Van Nostrand Reinhold
  Environmental Engineering Series.
MWRA. 1993. Massachusetts Water Resources Authority.
  Interim report  on master planning and CSO facilities
  planning. Report prepared by Metcalf & Eddy, Inc.
NYC  DEP.  1993.   New  York  City  Department of
  Environmental Protection. Inner harbor CSO facility
  planning project (in  progress). Hazen  and  Sawyer,
  and HydroQual, Inc.
U.S. EPA. 1974.  U.S. Environmental Protection Agency.
  Urban stormwater management and technology: an
  assessment. EPA/670/2-74/040.
U.S. EPA. 1977a. U.S. Environmental Protection Agency.
  Urban stormwater  management  and  technology:
  update  and  users'  guide.  EPA/600/  8-77/014.

U.S.  EPA.  1977b.  U.S.  Environmental  Protection
  Agency. Cottage Farm combined  sewer detention
  and chlorination station, Cambridge, Massachusetts.

U.S. EPA. 1986. U.S. Environmental Protection Agency.
  Methodology  for analysis  of  detention  basins for
  control of urban  runoff quality. EPA/440/5-87/001
  (NTIS  PB  87-116 562).  Prepared for Office  of
  Water, Nonpoint Source Division. September.

U.S. EPA. 1988. U.S. Environmental Protection Agency.
  Stormwater  management   model  users'  manual
  version 4. EPA/600/3-88/001. August.
U.S. EPA. 1989. U.S. Environmental Protection Agency.
  Analysis of storm event characteristics for selected
  rainfall gages throughout the United States. November.

Washington, DC, Water and Sewer Utility Administration.
  1992. CSO abatement program, segment 1: performance
  evaluation. April.

WEF.  1992. Water Environment Federation. Design of
  municipal wastewater treatment plants volumes I &
  II. WEF manual of practice no. 8 and ASCE manual
  and  report on engineering practice no.  76.  2nd
  edition. Alexandria, VA: WEF, and New York, NY: ASCE.

                                             Chapter 4
                                   CSO Control Technologies
In-System Controls/In-Line Storage


One of the more readily  implementable and cost-
effective approaches to achieving immediate reductions
in CSO volumes is to utilize the available storage and
conveyance capacity of existing collection systems and
the  available  treatment  capacity at  the  POTW.
Optimizing the  use of in-line storage and maximizing
flows to the POTW reduces  overflow volumes by
allowing a larger fraction of the total flow from a storm
event to be conveyed to the POTW for treatment. This
control approach is only feasible if sufficient capacity is
available in the collection system and at the treatment

A number of "in-system" technologies or strategies can
contribute to maximizing in-line storage, maximizing flows
to the POTW, and reducing overflow volumes, including:

• Collection system inspection and maintenance.

• Tidegate maintenance and repair.
• Reduction of surface inflow.

• Adjustment of regulator settings.

• Enlargement of undersized pipes to eliminate flow

• Removal of obstructions to flow, such as sediments.

• Polymer injection to reduce pipe friction.

• In-system flow diversions through existing system

• Adjustment   and/or  upgrade  of pumping  station

• Partial separation  of storm drain connections from
  combined sewers.

• Infiltration removal.

Prior to  implementing these practices, communities
should undertake a program of detailed data collection,
flow monitoring, and  modeling to fully characterize the
features and behavior of their combined sewer systems.
The results of these system characterizations can be
used to evaluate the performance of each regulator, and
to identify the potential for utilizing storage in conduits
upstream of the regulator, eliminating flow bottlenecks,
and improving pump station operations. The goal of all
these  measures  is  to  optimize  the  storage  and
conveyance capacity of the combined system.

Many types of in-system controls can be implemented
with relatively minimal engineering and cost,  although
detailed hydraulic  analyses  are  required. Evaluating
and  optimizing system components  may  serve to
reduce  the overall scope  of a community's  CSO
problem, thereby reducing the size and/or number of
more capital-intensive CSO control facilities required to
meet  overall  CSO control goals. In-system controls
generally  will  be  most  effective  where  upstream
drainage systems consist of large-diameter pipe laid on
shallow gradients.

Disadvantages to  increasing in-line  storage through
implementation of  in-system controls may include an
increased risk of basement or street flooding, increased
opportunity for sediment deposition, and higher costs
associated  with  increases  in maintenance  efforts to
ensure that regulators, tidegates, and other features are
functioning  properly and are in good repair.

When evaluating potential in-system controls, criteria
can be set for allowable changes  in predicted flow
velocities and peak hydraulic grade line elevation. Using
output from a detailed system model such as SWMM,
time-varying flow  velocities  in  conduits subject to
backwater  could  be  reviewed  both  for  existing
conditions  and  for  conditions  once the  proposed
in-system controls are implemented. The criteria could
be  to  ensure that minimum carrying  velocities  are
restored once the storm  ends and the backwater
subsides. Similarly, changes to the peak  hydraulic
grade line in conduits upstream of the proposed controls
could be limited to a set maximum above existing
conditions,  or to a minimum depth below grade. Setting
criteria for  predicted increases to the hydraulic grade
line requires judgment, and the criteria could vary within
a given drainage  system,  based on  knowledge of
conduit depth, previous reports of flooding, tidal impacts,
and other factors.

This   manual   focuses   primarily   on   regulator
modifications and  control, and  the use of real-time
control systems, as a means to optimize in-line storage
and  conveyance  of  flows  to  the  POTW.   Sewer
separation  and infiltration  removal  are  two  more
capital-intensive collection system  control  strategies.
These strategies are discussed only briefly in Chapter
2, but  have been well  described by others  (U.S. EPA,
1974a, 1975a;  Metcalf & Eddy, Inc.,  1981). Similarly,
actions to increase in-system storage which are related
to maintenance, such as tidegate inspections and sewer
flushing, and actions associated with best management
practices,  are  not addressed  in  this manual.  Key
references provided at the end of this chapter address
these items (WPCF, 1989; Metcalf & Eddy, Inc., 1991 a;
Field, 1990).


Regulators control the amount of flow  that enters  an
interceptor from an  upstream combined system, and
provide an overflow relief point (the CSO)  for  flows in
excess of the interceptor capacity.  Regulators fall into
two  broad categories: static  and  mechanical.  Static
regulators feature  no moving parts and, once  set, are
usually not  readily adjustable. Examples of static
regulators include side weirs, transverse weirs, restricted
outlets,  swirl  concentrators   (flow  regulators/solids
concentrators),  and vortex valves. Mechanical regulators
are  more  readily  adjustable  and  may   respond  to
variations  in local flow conditions, or be controlled
through  a remote  telemetry  system.  Examples  of
mechanical regulators include inflatable dams, tilting
plate regulators, reverse-tainter gates,  float-controlled
gates, and motor-operated or hydraulic gates.

Many of the older float-operated mechanical regulators
are   erratic   in  operation   and  require  constant
maintenance.  In  Saginaw, Michigan,  many  existing
float-operated  regulators  were  replaced   by vortex
valves,  due  to   the  unreliability  and   excessive
maintenance associated with the mechanical regulators
(U.S.  EPA, 1985b). In Boston,  Massachusetts, many
float-operated  regulators have been replaced over the
years  with static regulators.

Following are  descriptions of types of  regulators and
gates  which  have been  installed in more recent CSO
control projects, or have been  used to replace older,
less reliable types. The reader is referred to other
sources (WPCF, 1989; Metcalf & Eddy, Inc., 1991a; and
Urbonas   and  Stahre,  1993)  for  more  detailed
descriptions of other regulator types.

Vortex Valves

Vortex valves  are  static regulators  that allow  dry
weather flow to pass without restriction,  but control
higher flows  by a vortex throttling action (Figure 4-1).
Vortex valves have been used to divert flows to CSO
treatment facilities, control flow out of storage facilities,
and   replace   failed  mechanical  regulators.   The
advantages  of  vortex  valves over  standard  orifices
include (Urbonas and Stahre, 1993):

• The discharge opening on the vortex valve is larger
  than the opening on  a standard orifice sized for the
  same discharge rate, thereby  reducing the risk  of

• The discharge from the vortex valve is less sensitive
  to variations  in  upstream head  than  a standard

Vortex valves are sized based on design flow and head.
Typically, a  vortex valve manufacturer will  provide a
table listing the ranges of flow and head appropriate for
the various models and sizes of vortex valves available.
The design  flow corresponds to the dry weather flow
capacity of the  downstream conveyance capacity. The
                              Controlled Flow
                                 Low Flow
                                 Free Discharge
 Figure 4-1.  Example of a vortex valve (Metcalf & Eddy, Inc.,

peak head is usually dictated by the maximum upstream
hydraulic grade line allowable without  undue  risk of
flooding. A high  outlet relief should be  provided to
ensure that upstream flooding does not occur.

Vortex valves used as CSO regulators are commonly
mounted  in   reinforced  concrete  structures.  Such
structures should be  channelized with benching to
eliminate "dead"  flow areas. The  valves are typically
constructed of stainless steel and  have been reported
to require little maintenance  (H.I.L. Technology, Inc.,
1988, 1991).

The European wirbeldrossel  and  wirbelvalve (Figure
4-2) are similar to the vortex valve, with a tangential inlet
and vertical bottom  outlet. The flow chamber for the
wirbeldrossel is horizontal and cylindrical, while the flow
chamber for  the  wirbelvalve is slanted  with a conical
bottom. These devices feature  an air supply pipe to
prevent cavitation which might otherwise develop in the
rotational flow. As with the vortex valve, the wirbeldrossel
and wirbelvalve provide greater restriction of flow  and
less sensitivity to upstream  head  than  similarly sized
standard orifices. In  Europe, these devices have been
used primarily to control the release of flow from storage
facilities (Urbonas and Stahre, 1993).

Inflatable Dams

An inflatable  dam is  a  reinforced rubberized  fabric
device which, when fully inflated, forms a broad-crested
transverse weir. When deflated, the dam collapses to
take the form of  the conduit in which it is installed. A
dam can be inflated with air, water, or a combination of
both.  Water provides the best control of the weir crest,
but air control is usually associated with lower costs for
equipment, particularly if rapid inflation-deflation cycles
are required. Air inflation is also required where the dam
may  be exposed to  freezing temperatures (APWA,

Inflatable dams are typically positioned to restrict flow
in an  outfall  conduit or  combined sewer trunk.  The
dams, which normally  remain fully inflated, can act as
regulators  by directing flow  into  an interceptor,  and
preventing the diversion of flow to an outfall until the
depth of flow exceeds the crest of the dam (Figure 4-3).
Inflatable dams are controlled by local or remote flow or
level sensing devices,  which regulate the height of the
dam to optimize  in-line storage and prevent upstream
flooding. The dam height is controlled by controlling the
air or water pressure in the dam.

Since inflatable dams are typically constructed of rubber
or strong fabric, they are subject to puncture by sharp
objects. These devices generally require relatively little
maintenance, although the associated air supply should
receive periodic inspection to ensure reliable operation
(WPCF, 1989).
Motor- or Hydraulically Operated Sluice Gates
Similar  to inflatable  dams, motor-  or hydraulically
operated gates typically operate in response to local or
remote flow or level sensing devices. Normally closed
gates can be located  on  overflow  pipes  to  prevent
overflows except  under  conditions  when  upstream
flooding  is  imminent.  Normally  open gates  can be
positioned to throttle flows to the interceptor to prevent
interceptor surcharging. Controls  can  be configured to
fully open or close gates, or to modulate gate position.
The  level  of  control  and   general  reliability  of
motor-operated gates make them well suited for use in
conjunction with real-time control systems (Figure 4-4).

Elastomeric Tide Gates
While not actually regulators, tide  gates are intended to
prevent the  receiving  water from  flowing back through
the outfall  and regulator  and into  the  conveyance
system.  Inflow  from  leaking  tide  gates  takes up
hydraulic capacity in the downstream interceptors and
increases the hydraulic load on downstream treatment
facilities. Elastomeric  tide gates provide an alternative
to the more traditional flap-gate style tide gates, which
are prevalent in many CSO communities. Tide gates
have  historically  required  constant  inspection  and
maintenance to ensure that the flaps are seated correctly,
and that  no objects or debris  prevent the gate  from
closing. Warpage, corrosion, and a tendency to become
stuck in one position also characterize flap-gate  style
tide gates. Elastomeric tide gates  are designed to avoid
the maintenance  problems associated with  the flap
gates (Figure 4-5). In particular, the elastomeric gates
are designed to close tightly around objects that might
otherwise prevent a flap gate from closing (Field, 1982).

Regulator Controls

Static Regulators
Static regulators have no moving parts and thus offer
no opportunity  for additional  control once  the weir
elevations or orifice  dimensions are set. Adjustable
weirs may allow some degree of manipulation between
storm   events.  Modifications  to  static  regulators,
however, generally can be achieved at relatively low
cost.  For example, if  a collection  system  hydraulic
model  indicates  that  upstream storage  could  be
optimized by raising  an existing  masonry weir in the
overflow pipe at a regulator, the immediate benefits of
such  actions may justify the cost  of the  relatively
minor  construction work  required to adjust  the  weir.
Modification of restricted outlets  requires more extensive
demolition and  excavation  work, and would  have to
achieve a greater reduction in overflow volume to match
the cost effectiveness of raising weirs.

                                                                    Aeration Pipe
Figure 4-2.  Example of a Wirbeldrossel and wlrbelvalve (Urbonas and Stahre, 1993).

                        To Interceptor

Figure 4-3. Example of an inflatable dam (WPCF, 1989).
Local Control

Local regulator control is most appropriate where a
single regulator is associated with an outfall, or where
the behavior of a regulator would not influence or be
influenced by the behavior of another regulator in the
system. Many of the older style mechanical regulators,
such as reverse tainter gates, respond  to local water
level in the sewer through a float mechanism. These
types of regulators are unreliable in many installations,
require constant maintenance, and are slow to react to
rapidly  changing flow conditions (U.S.  EPA, 1985b).
More advanced local control systems feature electronic
flow or water level monitoring devices, which  control
motor-operated  gates  or  inflatable  dams.  These
systems  appear  to  be  more  reliable  than  the old
float-controlled mechanical gates.

System-Wide Real-Time Control

Real-time control (RTC) systems can provide integrated
control  of regulators, outfall gates,  and pump station
                            Trunk Water
                            Level Controls
                            Outfall Gate
                 Depth of
                 Water in Interceptor
                 Controls Regulator
                                                              May Prevent Gate
Figure 4-4.  Example of a motor-operated gate regulator (WPCF, 1989).

Figure 4-5.  Example of an elastomeric tide gate (WPCF, 1989).

operations based on anticipated flows from individual
rainfall events. Feedback control adjustments are based
on actual flow conditions within the system. Computer
models associated  with  the  RTC  system  allow
evaluation of  expected system response  to control
commands before the commands are executed. As with
any  plan  for  improving  in-line storage,  to take  the
greatest advantage of  RTC, a collection system must
have relatively  flat upstream  slopes  and  sufficient
upstream storage and downstream interceptor capacity.

An RTC system typically includes the following control
features (James and Stirrup, 1986):

• Sensors to  detect flow, water level,  rainfall, and/or
  pollutant concentration.

• Circuitry and software to transform the signals from
  the sensors into numerical quantities.

• Circuitry   and  software   to  drive   the  control
  mechanisms (usually gates).

• Rainfall and/or runoff forecasting software running in
  real time.

• A computer acting as both data logger and controller.
• For an integrated RTC system, telemetry equipment
  for the communication  of data among the various

A control strategy based  on RTC must identify  the
control system constraints and evaluate alternatives for
developing  the optimum control strategy within those
constraints. Examples of constraints which could define
the limits of an RTC system include (Trotta et al., 1977):

• Capacities of interceptor and trunk sewers, storage
  facilities,  and POTW.

• Rainfall runoff forecast models.

• Data acquisition system.

• Computer hardware and software.

• Control timestep.

• Human error and equipment malfunction.

Evaluation of the costs and impacts of addressing these
constraints may dictate to what extent an RTC system
will be appropriate for a given community. For example,
if little excess capacity is available  in the collection
system, the benefits  of RTC may not justify the costs.
Similarly,  a  community  must  have  sufficient  staff
resources to inspect  and maintain the remote sensing
equipment to minimize the risk of equipment malfunction.

The  configuration of an RTC system depends on a
community's overall  CSO  control goal. For example,
control of average annual CSO volume may require a
different control strategy than control of mean annual
CSO  frequency.  For this reason, evaluation  of  the
feasibility of an RTC system should be conducted as
part of a long-term CSO control plan.

One of the  older RTC systems in the United States has
been operated by  the Municipality of Metropolitan
Seattle since 1973.  Regulators in the Seattle system
feature vertical in-line sluice gates to throttle flow into
the  interceptors  and  store  flow in  the  upstream
combined sewers. Overflow weirs upstream of the gates
prevent upstream flooding during extreme events. The
original computer-augmented treatment and disposal
(CATAD) control  system monitored flows and water
surface levels in  the combined system, and provided
integrated, remote control of pump station and regulator
gate operation.

A 1990 study evaluated the most cost-effective means
for upgrading the Seattle  RTC  system and  providing
additional storage of CSO flows. The study found  that
improvements to the RTC  system would be more cost
effective for  providing  additional in-line storage  than
construction of off-line storage facilities. Improvements
to the  RTC system  included  development  of a
hydrologic  model of  the CSO system, instrumentation
upgrades,  database development, integration of the

hydrologic  model  and  database  into  a  simulation
program, and control strategy design.

The  hydrologic  model can estimate flows into the
system  for up to 6  hours  into the future, based  on
predicted  rainfall and  current  conditions  within the
system. Information on current conditions is continually
input through a supervisory control and data acquisition
(SCADA)  system. Initial simulation  runs  assume a
relatively simple, local  control strategy.  Following the
initial   simulation,  an  optimal  control   strategy   is
computed using an optimizing algorithm. A flow-routing
model then checks the optimized strategy for feasibility.
Infeasible  strategies  are   rejected,   while  optimal
strategies which  are modeled to be feasible are sent to
the SCADA system for execution (Vitasovic et al., 1990).

With  the  continuing improvements to computer and
instrumentation technologies, and the cost-effectiveness
of in-system control vs. new facilities, RTC systems are
gaining in popularity for both CSO and stormwater control.

Off-Line Near-Surface
Off-line,  near-surface storage/sedimentation  facilities
consist of tanks that store and/or treat combined sewer
flows   diverted  from combined  trunk sewers and
interceptors. These facilities provide storage up to the
volume of the tanks, as well as sedimentation treatment
for flows that pass through the facilities in excess of  the
tank volume. Coarse screening, floatables control, and
disinfection  are  commonly provided. The  phrase
"near-surface" means that these facilities are constructed
at depths that allow the  use of traditional open-cut
excavation techniques, as opposed to the  deep tunnel
facilities described in the section  of this document on
deep  tunnel storage.

Process Theory

The  sizing  of sedimentation structures  at  POTWs
historically has been based on discrete particle settling
theory. Type 1 discrete particle settling theory relates
the terminal settling velocity  of a particle to an overflow
rate and ratio of depth to detention time, which can then
be used as a basis for sizing a sedimentation basin.
Detailed discussions of discrete particle settling theory
are provided in other texts (Metcalf & Eddy, Inc.,1991 a;
WEF, 1992). For CSO storage/sedimentation facilities,
the basic discrete  particle  settling model must   be
expanded to account for the  non-steady-state conditions
that occur as a  CSO storage/sedimentation tank fills,
overflows, and is dewatered.
A number of mathematical models have been developed
to predict sedimentation behavior in settling tanks. One
such  model, developed for preliminary evaluation of
storage/sedimentation facilities, identifies  four distinct
operating  phases that may occur in a CSO  storage/
sedimentation tank (Lessard and Beck, 1991):

• Fill, when flow is entering the tank but the tank is not

• Quiescent  settling,  when  the  tank  is partly or
  completely filled, with no flow entering or leaving the

• Dynamic  settling,  when   the  tank  is  full   and

• Draw, when  the tank is  being dewatered to the
  sanitary system.

This model can be  used to  estimate the  change in
selected pollutant concentrations and sludge mass over
time, given an array of input parameters such as tank
size, particle settling velocity, and percent settleable
fraction. The model provides an approach for estimating
both the performance of a storage/sedimentation tank,
and the subsequent  hydraulic and pollutant load on
the POTW resulting from dewatering of the  storage/
sedimentation tank.  The following aspects  of  CSO
storage/sedimentation  tank  behavior,  which   were
considered  in   developing  the  model,  summarize
storage/sedimentation  tank  dynamics (Lessard  and
Beck, 1991).


During  the initial  stages of  tank  filling,  turbulence
created by the influent flow prevents the contents of the
tank from  settling. As the tank fills, there may  be some
fraction of the total tank volume above which the influent
turbulence no longer inhibits sedimentation. For larger
tanks, the fraction  of total  volume  required before
sedimentation occurs is relatively small; for small tanks,
turbulence may influence a greater percent of the total
tank volume.

Quiescent Settling

The percentage of solids that settle during the quiescent
settling phase depends upon the settling velocity profile
and the duration of the settling  period. The extent of
quiescent  settling  that occurs  during the  lag period
between the end of the storm flows to the facility and
the start of the tank dewatering process may impact the
solids  loading rates on the POTW as the tanks are

Dynamic  Settling

Under dynamic settling conditions, the behavior of the
tank is essentially the same as a typical primary settling
tank  at a POTW.  Key parameters for  estimating
removals  during this phase include an average particle
settling velocity, overflow rate, detention time, and the

fraction of a particular pollutant that could be considered

Two issues are considered when evaluating the draw
phase and the effect  of  dewatering a  CSO storage/
sedimentation facility on the POTW:
• Pollutant concentration
• Timing of dewatering
The pollutant  concentration in the dewatering  flow
depends on the sludge handling process at the CSO
facility.  Some  tanks  are aerated or  provided  with
mechanical  means for resuspending  solids prior  to
pumpback.  Other  facilities may  provide  high-rate
dewatering  of  the clarified volume,  with separate
pumping of the sludge. Some facilities are dewatered
by gravity, with or without separate disposal of settled
solids. In general,  it  is assumed that the  sludge is
resuspended to a given depth, which corresponds to a
given fraction of the total tank volume. The load on the
POTW can then be estimated from the two volumes, the
concentrations  within  the  two  volumes,  and the
dewatering rate(s).
Ideally,  the  dewatering phase  should  be initiated  as
soon as capacity is available at the POTW and in the
conveyance system between the CSO facility and the
POTW.  This ensures that the greatest tank capacity is
available for the next storm. Theoretically, a dewatering
system   could  be  activated  automatically  on  a
telemetered signal from the POTW, once flow at the
POTW  drops below  a specified value. In  practice,
however, many  CSO facilities  are  designed  with
manually activated dewatering  systems. One  of the
benefits of manual activation is that POTW operators
can evaluate the impacts of sustained peak storm flows
on  POTW performance,  before potentially prolonging
the  peak flows by dewatering  the CSO facility. The
solids loads from a storage/sedimentation facility may
be  particularly  difficult to  handle  for a system which
already has been stressed. In addition, some POTWs
that serve combined systems may implement special
operating procedures during wet weather to mitigate the
impact  of peak storm  flows, particularly on biological
treatment systems. Delaying the dewatering of a  CSO
facility may impact the level of control achieved with a
storm that follows immediately.  However, if immediate
dewatering  of a  CSO facility results in the upset or
washout of a secondary treatment system, the overall
impact on receiving waters may be greater.

Process Design
Sizing criteria  for a storage/sedimentation  tank  vary,
depending upon the intended goal of the CSO control
facility. As discussed in Chapter 3, typical CSO control
goals may include:

• Providing a specified minimum treatment level, such
  as the equivalent of primary treatment for flows up to
  a specified design condition.

• Providing  a level  of  treatment required to  meet
  receiving water quality standards for storm flows up
  to a given recurrence interval.

• Providing full capture of the first flush (highest mass
  loadings during the storm event),  then a  reduced
  level of treatment for subsequent flows.

• Reducing the  number of annual  overflow events
  and/or total annual  overflow volume to a specified

Selection of the control strategy may depend on local,
state, or federal  regulations; a community's long-term
CSO control plan;  or requirements  of  a permitting
authority. All strategies require an estimate of the flows
expected under the given design condition. Computer
models,  such  as  SWMM  which  is described  in
Chapter 3, can be used to simulate the behavior of a
combined sewer system. System behavior can be
simulated for  a  specific storm of previously defined
characteristics, and on a continuous basis, reflecting the
precipitation  history  of  a  selected  time   period
(commonly  one  year).  The  continuous  simulation
models the impact  of succeeding storms on system
performance,  which  may  impact  facility sizing and
dewatering rates. Sizing based on modeled  flows is,
therefore, an iterative process.
General  concepts related to each of the above CSO
control goals are discussed in Chapter  3. Additional
process  design  considerations that apply to storage/
sedimentation facilities are described below.

Sizing To Provide a Specified Minimum
Treatment Level

The minimum treatment level most commonly identified
for CSO control facilities is primary treatment, although
facilities  also have been sized to provide a specified
minimum detention time for disinfection. Primary settling
tanks at POTWs are sized based on overflow rates and
detention times,  which are in turn derived from particle
settling velocities. Tables 4-1 and 4-2 present examples
of design overflow rates and detention times for primary
settling tanks (WEF, 1992).  Since  combined flows
generally have a higher  fraction of heavier solids than
separate  sanitary  flows,  the  values  presented  in
Tables 4-1  and 4-2  may be conservative for CSO
applications. In  addition, a  study of  the relationship
between solids removal performance and overflow rates
suggests that the performance expected from a given
design overflow  rate may be affected by other  factors

Table 4-1. Typical Overflow Rates for Primary Settling Tanks
         (WEF, 1992)
   Overflow Rate

nVWVd  gpdrtt2
Metcalf & Eddy,    Primary treatment
Inc., 1991 a; U.S.   followed by
EPA, 1975b       secondary:
WEF, 1992
WEF, 1992
Mississippi River
Board, 1978
  Average flow
  Peak flow

All units in service:
                                   32-48    800-1,200
                                   80-120  2,000-3,000
                  Maximum day flow   49     1,200
                  Peak flow          81     2,000
One unit out of
  Peak flow
Larger area of:

  Average flow
  Peak hour flow
                                   163    4,000
Table 4-2.  Typical Detention Times for Primary Settling Tanks
          (WEF, 1992)
 Detention Time
Metcalf & Eddy,
Inc., 1991 a

WEF, 1992

Fair et al., 1968

U.S. Army, 1978
 Primary treatment
 followed by
 Primary treatment
 followed by
 activated sludge
 1.5 to 2.5 hours, 2
 hours typical

 1 to 2 hours based
 on peak flows

 Minimum 2 hours in
 3m (10 ft) side water

 1.5 hours
                 All other conditions   2.5 hours
 such as design details, loading variations, recycle flows,
 temperature,  and  proportions  of  inert-to-volatile  and
 soluble-to-insoluble material in the flow (WEF, 1992).
 Sizing criteria should consider both overflow rates and
 detention times. Determining the actual settling velocity
 distribution of the solids  in the CSO to be treated, as
 described in Chapter 3,  will provide a  more accurate
 assessment of the appropriate design overflow rate.

 Since  influent hydrographs for CSO control  facilities
 tend to exhibit more sharply defined peaks, which  are
 not typical of POTW influent hydrographs, and periods
 of no flow, the concept of the "average" flow to a CSO
facility does not have as much physical significance as
at POTWs. Thus, if overflow rates or detention times are
used  as a basis for  design of a CSO control facility,
considering  peak flow  conditions  may  be  more
appropriate than attempting to define an "average" flow.
If the shape of the influent hydrograph is known, then
the total volume of flow that has entered the facility at
the timestep when the peak occurs can be determined
and compared against the total tank volume. If the peak
flow occurs before the tank is full,  then the actual
maximum overflow rate would occur on the falling leg
of the influent hydrograph.

The actual maximum overflow rate, therefore, would be
less than the calculated overflow rate associated with
the actual peak  influent  flow.  This  condition would
provide  a degree of conservatism in  the sizing of
facilities based on peak overflow rates.

Sizing To Meet Water Quality Standards

Water quality impacts influence the level  of treatment
that permitting agencies require CSO control facilities to
achieve. Wasteload allocation studies (which include
water quality monitoring and modeling) can be used to
determine the  effluent  characteristics  and level of
treatment needed to achieve a given water quality goal.
Once the required level of treatment is established, the
design can then evaluate tank sizes and configurations
to meet the required level of control.

Sizing To Capture First Flush
A number of CSO facilities have  been  designed to
capture  the  more concentrated combined  flows  that
may occur during the initial stages of a storm event. The
key to designing such a facility is to define the limits of
the first flush,  which will in turn define the  volume of
capture required. Chapter 3 defines and discusses the
use of the first flush as a design criterion.

Sizing To Reduce the Number of Overflow Events

Use of this criterion  requires continuous simulation to
estimate  the  flow associated with  storms  of various
recurrence intervals. As discussed in Chapter 3, the
definition of  an "overflow event" also must be clearly
established.  An overflow event  may be defined as a
treated overflow, an untreated overflow, or an event in
which  flow  receives less  than a  defined  level of
treatment. The established definition  may  determine
whether  a  storage-only  or  flow-through  storage/
sedimentation  facility is appropriate for the application.
A related consideration is whether flows in excess of the
design storm are to bypass the facility or  flow through
the facility and  potentially receive a lower  level of
treatment than provided under design conditions.

Process Flow

The  arrangement of unit processes at a storage/
sedimentation facility varies depending on the goal of
the CSO control strategy (i.e., primary treatment vs. first
flush capture) and hydraulic considerations. A typical
arrangement includes a regulator, bar screens, settling
tank(s), and an outfall. Disinfection may be required and
can be provided  upstream and/or downstream of  the
settling tanks. Figure 4-6 presents a schematic for a
typical layout of a storage/sedimentation facility.


Typical regulator/influent arrangements include:

•  Overflow from  remote regulator conveyed to CSO
   facility by influent conduit. Depending on the type of
   regulator used, the bedload could tend to remain in
   the interceptor. The regulator should be designed to
   maximize  the  in-system storage  capacity  of  the
   upstream pipe  network.

•  Overflow from sanitary wetwell. A common arrangement
   is to combine a CSO facility with a standard sanitary
   sewage pumping  station.  As  flow exceeds  the
   capacity of the sewage pumps, flow  in  the  pump
   station  influent channel overtops a weir, and passes
   to the CSO facility influent channel. This arrangement
   typically occurs downstream of coarse  screening
The  regulator or influent piping arrangement  should
provide some means for a relief overflow to protect the
facility from flooding. In Newport, Rhode Island (Figure
4-6),  a high flow relief weir  is  located in a manhole
adjacent to the side weir regulator controlling flow into the
facility influent conduit. The facility includes  provisions
for disinfecting the high  level overflow. In addition, a
positive  means  for isolating the  facility  should  be
provided. The Newport facility and the MWRA Prison
Point facility both feature a modulating influent gate to
throttle flow into the facility when the water surface in
the facility reaches a high level.

Bar Screens

Most  storage/sedimentation   facilities provide  some
means for coarse screening  of the influent, such as
catenary-type mechanically cleaned screens. Isolation
gates for the barscreens should be provided, along with
a  bypass or a redundant channel. Coarse screening is
described in more detail later in this chapter.

Influent vs.  Effluent Pumping

Hydraulic  conditions often require  influent  or effluent
pumping to or from a CSO facility. Operationally, effluent
pumping is preferable, since the pumps run only during
a  storm  large  enough to overflow the  tanks.  The
disadvantage of effluent pumping is that the elevation
of the entire facility must be set by the hydraulic grade
line of the influent conduit. If the influent conduits are
well below grade, then relying on gravity flow into the
facility  could  require  significant   excavation   and
construction. Influent pumping allows the  elevation of
the facility to be  set  independently  from  the influent
conduit elevations. Ideally, the facility elevation is set
such that the facility effluent, and possibly the facility
dewatering piping, flow by  gravity. If influent pumping is
selected, wetwell  and pump sizing must consider the
range of flows expected.


If  required,  disinfection is typically provided by liquid
sodium  hypochlorite, dosed by metering pumps,  and
paced by influent flow. Some facilities include additional
discharge control based on chlorine residual. Dosing
typically  is  applied at  the  tank  influent channel to
maximize contact  time.  Some  facilities (Newport,
Washington  Street; MWRA  Prison  Point) include  a
second dosing location after sedimentation to augment
the original dose.  One drawback to dosing upstream of
the settling tanks is that the flow at that location contains
a higher concentration of solids than downstream of the
tanks. Thus, while  contact time is  gained by dosing
upstream of the  tanks, actual  "contact"  between the
bacteria and the  disinfectant may not be as effective.
Auxiliary uses  of disinfection  may include  periodic
disinfection  of  sludge  piping  and  the  spray wash
system. Dechlorination  also  may be required due to
concerns over  the potential toxic impacts of chlorine
residuals  on the receiving waters.  Disinfection  and
dechlorination are discussed in more detail later in this

Solids Handling-Screenings

A number  of  methods have been used to handle
screenings  from  mechanically cleaned bar  screens.
Some methods  require  separate  disposal  of  the
screenings,  while others return the screenings to the
interceptor. Screenings handling options are discussed
in more detail later in this  chapter.

Solids Handling-Sludge  and Grit
Due  to the  intermittent operation  of  CSO  facilities,
mechanical  sludge collection  equipment typically is not
provided. Rather,  tank floors  are sloped to a center or
side trough  and flushing systems help direct solids to
the trough. The trough is  sloped back to a dewatering
system drawoff for either pumping or gravity flow back
to the interceptor. Exceptions to this arrangement are
the  circular open  tanks  in  the Springfield, Illinois,

                  A   /-eoa Effluent
                      / Conduit (To Outfall)
                      >  (Narranansfitt Bav^
                                                                                           Operations Building
                                                                             10" FM
                                                                             and Infiltration
                                                                             to WPCP Via
                                                                             36" FM)
                             (Narragansett Bay)
                                                                                            Flow Direction
                                                  Effluent  L Settling Tanks  Gate
                                                  Box               Influent
           Effluent Lift Station
                                                                                          Influent Flowmeter
                                        48" Settled
                                        Effluent Line
                                                             Manhole Station
                                                             10 + 64
                                                                                                       18" Overflow Connection
                                            60" Influent Sewer

                  Dewatering System

                  Flushing Water

                  Chlorine Solution

                  Flow Direction

                  Dewatering and
                  Infiltration Flow


                  Sluice gate

                  Slide Gate
24" Sewer        /-w  /-Overflow
(To Long Wharf    f I Structure
Pump Station)	H /  No. 2
                                                                                                              24" Sewer        T
                                                                                                              (From Sewer   >^
                                                                                                              District No. 2)
   72" Storm
   (From City)-
   • 30" Overflow Bypass
   -Overflow Structure
    No. 1
Figure 4-6.  Flow schematic for typical storage/sedimentation facility, Newport, Rhode Island (Metcalf & Eddy, Inc., 1991b).

Sanitary District, Lincoln and Taylorville facilities, which
feature  mechanical sludge scrapers that are activated
automatically on tank level. The potential solids loads
on the POTW during pumpback operations should be
considered  in  evaluating sludge and grit  handling

Design Details
Many of the design details for CSO storage/sedimentation
facilities are influenced by variations in treatment goals,
local  hydraulic  conditions,  and  other  site-specific
considerations. As a result,  a variety of  methods  or
strategies have been developed to address such issues
as facility activation, flow distribution, tank flushing, and
other aspects of facility design. The following examples
of design details have been incorporated into currently
operating   CSO  storage  or   storage/sedimentation

Facility Influent  Gates
Influent gates  can  prevent  dry weather flow from
entering the facility, control the rate of wet weather flow
to the  facility,  and  protect  the facility from flooding
or  activating  during equipment  failure  or  regular
maintenance. Examples of influent gate operations are:
• MWRA Cottage Farm: Flow passes over a weir to an
  inlet structure. Three 72-inch sluice gates in the inlet
  structure open automatically when  the liquid in the
  inlet structure reaches a high level. One gate is left
  open several  inches to  prevent  splashover from
  triggering facility startup (U.S. EPA, 1977a).
•  Saginaw,  Michigan, Hancock Street: The original
  design  provided a sonic  level sensor in the CSO
  facility influent conduit,  set to open sluice gates to
  the facility on high level, then close  the gates on low
  level. Once on line, gate operation was erratic and
  the controls were changed  to require manual closing
  of the gates on low level. Non-rising stem sluice gates
  were used, but a  1990 study recommended  using
   rising stem gates to avoid submergence of the stem
  threads (U.S. EPA, 1980).
•  MWRA  Prison Point: A  hydraulic sluice  gate  is
  controlled  by water level in an inlet structure and in
   the stormwater wetwell. The facility gate will modulate
   to throttle flow into the facility on a high flow condition
   in the inlet chamber. The gate also is provided with
   a manual remote control (Maguire,  1981).
•  Newport,  Rhode  Island,  Washington  Street: The
   normally open hydraulic sluice gate throttles flow into
   the facility on high level in the effluent screw pump
   wetwell   by  closing  and   reopening   in  2-inch
   increments on rising or falling high level.  High  level
   float switches in the influent channel and the effluent
   wetwell automatically close the gate completely as a
  backup in  case the  gate-modulating  system  fails
  (Metealf & Eddy, Inc.,  1991b).

Flow Distribution

CSO  storage/sedimentation  facilities commonly are
designed  with  multiple  rectangular  tanks.  Flow  is
distributed  such  that the tanks  fill sequentially  to
minimize maintenance  and  cleanup following small
storms. At some point in the process of filling, flows are
equalized so that all tanks are completely full before
overflow  occurs. Examples of flow distribution include
the following:

• MWRA Prison Point: The inlet structure has one dry
  weather flow (DWF) channel and three wet weather
  channels. On high level in the DWF wetwell, the DWF
  gate throttles causing flow to  backup in  the  inlet
  structure. On high level in the inlet structure the first
  storm gate opens. A bubble tube  in the first tank
  signals the next gate to open  before the first tank
  water surface reaches the effluent weir. The third gate
  operates in a similar manner. Once all gates are open
  and all tanks are at the same level, the water surface
  uniformly  rises to  the effluent  weir elevation. This
  operating strategy allows the tanks to fill sequentially,
  while  preventing overflow  until all tanks  are full.
  Available  storage  capacity in  the inlet  structure
  provides a delay between throttling of the DWF gate
  and  opening of the  first  storm  gate,  preventing
  short-cycling of the storm system (Maguire, 1981).

• Newport,  Rhode Island,   Washington Street:  The
  elevation of the influent sluice gates for one tank is
  lower than the elevation of the gates for the other two
  tanks.  Flow in the influent channel will pass into tank
  1 first, then  the second two. All gates are located
  below the effluent weir elevation, so all tanks will be
  completely full before the facility overflows. There are
  three manual sluice gates for each tank. All gates are
  normally kept open,  although  operators can  allow
  tanks 2 and  3 to fill sequentially by monitoring tank
  levels  and manually operating the gates (Metealf &
  Eddy, Inc., 1991b).

• Saginaw, Michigan, Hancock Street: Influent pumps
  initially discharge to one bay. When this bay is 60
  percent full,  the sluice gate to the second bay is
  opened. Two additional bays follow the second bay.
  Once all bays are at the 60 percent fill level, they all
  will fill evenly,  and thus all overflow at once (U.S.
  EPA, 1980).

Tank Configuration and  Details-Rectangular Tanks

Rectangular tanks are the most common configuration
for  CSO  storage/sedimentation  facilities.  A  major
advantage of rectangular tanks  is that  common-wall
construction or installation of interior walls allows the

total volume to be compartmentalized. As noted above,
sequential filling of compartments can reduce cleanup
efforts during smaller storms. Figure 4-7 presents a plan
and profile for a typical rectangular storage/sedimentation
tank. A number of design considerations are related to
tank configuration and appurtenant details:

Surface Geometry

For  rectangular  primary  settling  tanks at  POTWs,
length-to-width ratios are typically in the range of 3:1 to
5:1.  In theory, tank geometry is intended to  minimize
flow patterns that  might resuspend settled  solids or
otherwise disrupt the settling process. The critical scour
velocity is calculated as follows (WEF, 1992):
                  H = {[8k(s-1)gdyf}
  VH = critical scour velocity
        k = constant for type of scoured particles
        s = specific gravity of scoured particles
        g = acceleration due to gravity
        d = diameter of scoured particles
        f = Darcy-Weisbach friction factor

     Typical values for k range from 0.04 for unigranular
     sand to 0.06 for sticky, interlocking material. The units
     in  the equation may be either U.S.  customary or SI
     (Metcalf & Eddy, Inc., 1991a).
     In practice, the width of primary settling tanks commonly
     is  constrained by the  available width  of mechanical
     sludge collection  equipment. Since such equipment is
     typically  not provided  at CSO storage/sedimentation
     facilities,  the constraints on tank width are not as rigid.
     In addition, the general reliability of length-to-width ratio
     as a design tool has been questioned (WEF, 1992). For
     tanks intended for storage only, geometry is of even less
     concern.  Table 4-3  presents  typical dimensions for
                               -Overflow Conduit
- Collection Trough
[-Influent Channel
nf 1 Sluice Gate and 1
/ Target Bafflen \

/- Influent
S Pressure
<- Sluice Gate

^Coarse Bar
                                                  Effluent Force Main to
                                                  Treatment Plant Headworks

                                                                          Hnfluent Chamber
                Conduit —
| ^Walkway With Railway
III 1 1 ^Al 1 1 1
II 1 1*1 1 II
.r- Contact Chamber Dividing Wall
\- Effluent
Contact Chamber
Trough -v T*

                                                                                  r- Influent
                                                                                 /  Pressure
                                                                                '   Sewer
                                                                                    - Coarse
                                                                                   - De watering
Figure 4-7.  Plan and profile for a typical rectangular storage/sedimentation facility (Metcalf & Eddy, Inc., 1991a).

Table 4-3. Typical Rectangular Tank Dimensions for
         Storage/Sedimentation Facilities
                                Dimension (ft)
Length     Width
Metcalf & Eddy, Inc.,

WEF, 19921

Great Lakes-Upper
Mississippi River Board,
 10-15, 12


7 minimum
U.S. EPA, 1975b1
Milwaukee Humboldt Ave. 420 75
(Medina etal, 1981)
Saginaw Hancock St. 206 52
(U.S. EPA, 1980)
Southgate/Wyandotte, 222 362
(Hubbell, 1990)
MWRA Prison Point 94 23.5
(Maguire, 1981)
Newport, Washington St. 130 25
(Metcalf & Eddy, Inc.,
1 Values for primary settling tanks
2 Large tank is compartmentalized
settling tanks from  published texts and examples of
actual dimensions at CSO storage/treatment facilities.


As noted above, to facilitate solids handling, floors are
typically sloped to a center or side trough, which is in
turn sloped  back to the dewatering  system drawoff.
Flushing  systems  generally  are  required to ensure
complete solids removal. A number of installations have
been  developed in Europe with multiple  grooves or
channels in  the tank floor, in an  attempt  to maintain
self-cleaning velocities across the entire floor area. This
type of design is more expensive to build, and operating
experience shows that some flushing is still required to
prevent the  build-up of solids (Stahre  and Urbonas,
1993). Table 4-4 presents examples of  floor slopes at
installations in the United States.


Most facilities include influent baffles to dissipate energy
in  the influent  flow and to  minimize short circuiting
through the tank. Commonly of timber construction, the
design of the influent baffles for CSO storage/treatment
tanks is similar to influent baffles for rectangular primary
settling tanks.  Details of influent baffle designs can be
found in the literature (WEF, 1992; Stahre and Urbonas,
1993). Effluent scum baffles control the discharge of
floatables, a desired goal of CSO abatement strategies.
The effluent baffles for CSO storage/treatment tanks
typically consist of steel plates or concrete beams that
extend  below  the  water  surface. Mechanical scum
collection troughs are less common.

Flushing Systems

Flushing systems are provided at storage/sedimentation
facilities to assist in solids handling and general cleanup
following facility  activation.  Due to the  intermittent
operation of CSO control facilities, it is important to keep
tanks clean between storms to  minimize odor buildup
and maintain tank  capacity.  As tanks are dewatered,
heavier solids that have settled out will not necessarily
be  resuspended unless some mechanical means such
as  aeration or  mixing   is   provided.  Without  fully
resuspending  settled  solids,  pressurized  flushing
systems can be used to direct the settled solids towards
the sludge drawoff  hoppers.  The two types of flushing
systems commonly found in CSO facilities in the United
States  are  header-mounted spray nozzles,  located
around the perimeter of the tank, and high-pressure,
manually  controlled monitor  nozzles,  mounted  on
Spray headers are  used to wash down walls, but may
not be effective in moving solids along the bottoms of
the tanks. The advantage of  spray headers is that little
operator attention  is required. Monitor nozzles  are
intended primarily for moving solids along tank floors
towards central gutters and sludge drawoffs. They also
are effective for cleaning walls,  but  are  more  labor
intensive than the spray headers.
Some facilities feature only one type of system (Oakport,
California),  while  others  (Newport,  Washington   St.;
Saginaw, Hancock  St.) feature both. In some facilities
where monitor nozzles are provided, hose gates and
flexible hose are provided to reach  corners or  other
                               Table 4-4.  Examples of Floor Slopes for
                                         Storage/Sedimentation Facilities

                                               Slope (ft/ft)    Location
Floor slope to
center trough
MWRA Prison Point
(Maguire, 1981)
Saginaw, Michigan, Hancock
(U.S. EPA, 1980)
                                                 0.0870     Newport, Rhode Island,
                                                            Washington Street
                                                            (Metcalf & Eddy, Inc.,

                                Center trough       0.0167     Saginaw, Michigan, Hancock
                                slope to drawoff                Street
                                                            (U.S. EPA, 1980)

                                                 0.0115     Newport, Rhode Island,
                                                            Washington Street
                                                            (Metcalf & Eddy, Inc., 1991 b)

spots out of the range of the monitor nozzles (MWRA
Cottage Farm).

Some European installations feature "tipping flushers,"
which consist of cylindrical containers mounted above
the width of the tank by an off-center swivel mechanism.
Each container has an asymmetrical discharge opening
running the length of the container. At the end of a storm
event,  after the storage/sedimentation tank has been
dewatered, a  flushing water  pump fills  the tipping
flusher. When the tipping flusher is nearly full, the center
of gravity shifts, causing  the  container to rotate and
discharge its contents. The force of the sudden rush of
water moves remaining solids along the floor of the
tank. Tipping flushers may not be  effective  in tanks
longer  than  approximately   160  feet  (Brombach,
A number  of facilities located  near rivers utilize river
water  as  the   source   of   flushing   water.  The
Bannockburn,  Scotland,  facility,   which   features  a
circular tank, experienced silting of the river intake and
clogging of  spray nozzles; the  spray  system was
abandoned. The Saginaw, Hancock Street facility and
the Bay City, Michigan, facility  both use river water for
flushing, while the Chapaton, Michigan, Retention Basin
uses lake water.  The Chapaton  facility  can  use  its
flushing system to resuspend solids for dewatering, but
this option generally is not used.

The MWRA Prison Point facility uses river water from
the Charles  River for plant water. The river water  is
supplied  by  gravity  through  an  auto-backflushing
strainer. Astormwater stripping  pump provides a backup
source of  flushing water. The MWRA Cottage Farm
facility uses an on-site well as  a flushing water source.
The Newport Washington Street facility uses city water
as a source for flushing  water. The city  water is fed
through an air gap into a break tank. On low level in the
break tank, a valve on the city  water supply opens. On
high level, the  city water valve closes and a "Flushing
Water Ready1  indicator light is activated at the main
control panel.

Tank Configuration and Details-Circular Tanks

Circular tanks typically feature  one of two  types of inlet
configurations: center column feed with influent baffles,
similar to circular primary settling tanks; and tangential
feed, more common in Europe, which creates a swirling
motion within the tank. Tangential feed tanks are in-line
devices intended primarily for storage of the initial storm
flows, with an overflow located upstream of the tank. A
vortex valve throttles the underflow from a center bottom
drawoff to the downstream interceptor (Brombach, 1989).
Flushing systems for center-feed circular tanks typically
feature header-mounted nozzles and/or manual flushing
with hoses fed from yard hydrants. Tangential feed
tanks are  intended to  be essentially  self-cleaning.
Typical dimensions for circular tanks are presented in
Table 4-5.
Table 4-5. Typical Dimensions of Circular Storage and
         Storage/Sedimentation Tanks
                          Dimension (ft)

(Metcalf & Eddy, Inc..       40-150
1991 a)


Bannockburn, Scotland        66
(Henderson et al., 1981)

Decatur, Illinois,             85
McKinley Ave.
(BGMAandCMT, 1977)

Lincoln, Illinois              130
(CMT, undated)

Springfield, Illinois,           180
Spring Creek
(CMT, undated)

Springfield, Illinois,           210
Sugar Creek
(CMT, undated)
           12 typical
* 1:3 slope is on sides of tank near bottom

Tank Access

The need for a cover over a tank depends on a number
of  factors, including the  elevation of the tanks with
respect to grade, proposed development on or adjacent
to the site, public  safety and security, the potential for
odors, and general aesthetics. A common practice is to
provide a walkway inside the tank, above the maximum
water surface, to provide access for maintenance and
cleaning. Explosion-proof lighting and electrical fixtures
are required, since the  tanks could be exposed to
potentially explosive sewage gases or  chemical spills
that could be conveyed through the collection system.
Walkways must be provided with railings and toeplates,
as  required  by the appropriate building and  safety
codes.  Covered   tanks  must   be   provided  with
emergency exits, in  accordance with applicable  codes.
In  some  facilities,   portions of  the  roof  slab are
removable, allowing a compact front-end loader to be
lowered into the tank to assist in  grit handling.

Tank Dewatering Systems

Standard practice is to return the contents of tanks to
the sanitary collection system by gravity, if hydraulically
feasible,  or by  pumping.  Trade-offs between  influent
pumping with gravity effluent flow,  and gravity influent
flow with effluent pumping should be evaluated during

design.  Multiple tanks should have the capability of
sequential  dewatering. One  common  method  for
providing sequential dewatering is to pipe the dewatering
drawoff  lines to a common header. Motor-operated or
hydraulic valves on the drawoffs are set to open and
close sequentially. The dewatering header could also be
piped to rapid dewatering  pumps  and solids stripping
pumps.  Centrifugal pumps  are appropriate for the rapid
dewatering  of  storage/sedimentation tanks,  while
plunger  pumps should be considered for solids handling.
Examples of dewatering systems at currently operating
CSO facilities include:

• MWRA Cottage  Farm:  Tanks  are  dewatered by
  gravity to the interceptor. Approximately 1 hour is
  required to drain the full  1.2-million gallon capacity of
  the facility (U.S. EPA, 1977a).

• Bannockbum, Scotland: Tank is dewatered by screw
  pump to the interceptor  (Henderson et al., 1981).

• Oakport,  California:  Tanks  drain  by  gravity  to
  interceptor, one tank at a time. Sequential dewatering
  minimizes the amount of time that sludge is exposed
  to the air, thereby minimizing odors (McCormick et
  al., 1990).

• Southgate/Wyandotte, Michigan: Gravity dewatering
  to interceptor is controlled by a flow meter and control
  valve (Hubbell, 1990).

• MWRA Prison Point: Manually controlled hydraulic
  gates drain the tanks to  a stormwater wetwell, which
  is dewatered by  a 500-gpm  centrifugal stripping
  pump. The stormwater  wetwell is sized to hold the
  entire contents of one tank (Maguire, 1981).

• Newport,  Rhode  Island, Washington  Street: Two
  800-gpm vertical dry-pit non-clog centrifugal pumps
  are provided for rapid dewatering, while two 85-gpm
  duplex  plunger  pumps are  provided  for solids
  stripping.  One of each type of pump is intended as
  a standby, although a "rapid dewatering" mode allows
  both  centrifugal  pumps to run, then  both duplex
  plunger pumps (Metcalf & Eddy, Inc., 1991b).

Ventilation  and Odor Control

Tanks,  screenings  areas,  wetwells,  and other areas
within storage/sedimentation facilities that are exposed
to  sewage  gases must be ventilated to prevent the
accumulation of potentially explosive  and/or corrosive
gases  and  to control  the buildup  of moisture  and
condensation. Evacuation  of sewage gases is required
for personnel safety and  equipment longevity, while
control  of moisture is  important since damp surfaces
can provide an environment for bacteria that oxidize
hydrogen sulfide to sulfuric acid, resulting in accelerated
corrosion. Ventilation systems for process areas are
sized based on the volume of the space to be ventilated
and a specified ventilation rate (typically in terms of air
changes per hour).  For areas  exposed to  sewage
gases, ventilation rates of 6-12 air changes per hour
are typical  (U.S.  EPA,  1985a;  Great Lakes-Upper-
Mississippi River Board, 1978). Two-speed fans allow a
lower rate of ventilation when spaces are not occupied
and a higher rate when  personnel  are present. This
arrangement lowers operating costs for facilities that are
not continuously staffed. Intake louvers are sized based
on  a  maximum  face velocity  and  commonly  are
interlocked with supply and/or exhaust fan  operation.
The   designer   should   consult  applicable  codes,
ordinances,  standards, and manufacturer's literature to
establish design criteria for a particular application.

Where storage/sedimentation facilities are  located in
the vicinity of residential, commercial, or recreational
areas, or where the exhaust from the facility ventilation
system may otherwise impact a sensitive receptor,  a
means for controlling odors in the exhaust gas should
be considered. Two of the more common processes for
removing odors from  exhaust gas are wet scrubbing
and  activated  carbon adsorbtion.  Activated carbon
systems generally are applicable for  air flows  up to
2,000 cfm, while wet scrubbers are more appropriate for
air flows greater than 2,000 cfm (U.S. EPA, 1985a).
Activated carbon may also be applied as a polishing
step, following wet scrubbing.
Discharge  permits may be required  for  ventilation
exhaust  from  CSO  facilities. The designer should
consult current federal, state, and local codes regarding
exhaust air discharge  permit requirements.

Wet Scrubbers
Wet scrubbers  function  by providing contact  between
the exhaust  air and a scrubbing solution, which removes
the odorous compounds through one or more  of the
following processes (U.S. EPA, 1985a):

• Condensation of odorous vapors

• Removal  of odorous participates

• Odor adsorbtion into the scrubbing solution

• Odor reaction with an oxidizing scrubbing solution

• Emulsification of odorous gases in a chemical reagent

Wet scrubbers  can be arranged  in a vertical,  counter-
current flow configuration, or in a horizontal, cross-flow
configuration. A typical wet  scrubber system  includes
the following features:
• Scrubbing tower, housing the following:
  - scrubbing  solution spray system
  - inert packing material, to promote liquid/solid contact
  - mist eliminator, to minimize loss of scrubbing fluid
     in the exhaust air

  - scrubbing solution wet well
  - drain

• Exhaust  fan with  ductwork,  including bypass of
  scrubbing tower

• Scrubbing solution recycle pumping system

• Chemical storage and feed equipment, including:
  - metering pumps
  - storage tanks
  - piping, valves, and appurtenances
  — instrumentation and controls (pH, oxidation-reduction
    potential [ORP] probes)

Potassium  permanganate and sodium hypochlorite are
two of the more common oxidizing agents used in wet
scrubbers.  Scrubbing solutions also  may include an
acid or base  to keep the pH of the solution in an
optimum  range  for  odor  compound  removal  or
neutralization. Selection of the appropriate scrubbing
solution  should   be  based  on   the   types  and
concentrations  of the odor-causing constituents to be
removed. Pilot testing should be undertaken to establish
design and performance criteria for a full-scale system.

Some scrubber units are provided with a constant flow
of make-up  water  and a continuously  overflowing
scrubbing  solution wetwell. The constant renewal of
scrubbing solution minimizes the build-up of solids in
the scrubbing  solution resulting  from  the  reaction
between hydrogen sulfide and sodium hypochlorite. The
overflow from  these units is commonly piped to a
sanitary drain.

Activated Carbon Adsorption Unit

Activated  carbon   adsorbtion  units remove  odors
through  the  adsorbtion of odor-causing  compounds
onto the surface of the activated carbon media. Atypical
activated carbon system consists of a vessel supporting
a bed of granular activated carbon, an exhaust fan, and

Sodium  hydroxide-impregnated  carbon commonly is
used  in  odor control applications, as it has a higher
capacity to  remove  hydrogen   sulfide  than  non-
impregnated   or  potassium  hydroxide-impregnated
carbon. A further advantage of the impregnated carbon
is that it can be regenerated in-place,  using 50-percent
sodium  hydroxide  solution.  Non-impregnated spent
carbon must be removed from the vessel and thermally
regenerated in a multiple hearth furnace. If impregnated
carbon  is  to be specified, the  design must  include
appropriate chemical  piping, valves, and drains to
permit in-situ sodium hydroxide regeneration.
Design  of an activated carbon  odor control  system
should include the following steps (U.S. EPA, 1985a):

• Characterization of the exhaust air to be treated.

« Identification of effluent criteria.

• Selection of the adsorbent.

• Completion of  pilot  studies to establish expected
  performance, estimated useful life  of  carbon,  and
  design criteria.

• Application of pilot data to full-scale design.

The designer is referred to manufacturer's literature and
published design guides (U.S. EPA, 1985a) for additional
information pertaining to odor control system design.

System Controls and Operation

Due to the uncertainty  in  predicting the nature of
precipitation and storm flows, CSO facilities typically are
designed for automatic activation. Facility activation is
triggered by sensing of flow or water surface elevation,
as a  regulator  activates and flow  is diverted to the
facility. Regulator activation can be passive (i.e., sideweir,
orifice) or mechanical (i.e., inflatable dam). Mechanical
regulator activation can be triggered by a flow signal or
water surface elevation, or through a real time control
system, as described in the first section of this chapter.

In general, the simpler  the activation system, the more
reliable the  operation.  For example, at the Newport,
Rhode Island, Washington Street facility  (Figure 4-6),
flow enters  the facility influent  conduit from a static
sideweir regulator.  Once flow  reaches  the  facility,
catenary bar screens and a hypochlorite dosing system
activate from a mercury  float switch  in the  influent
channel. Flow enters the tanks through normally open
manual sluice gates, with the elevations of the gates set
so that one tank fills first. No sophisticated controls are
required until the  effluent  weir overflows,  causing
activation of the  effluent screw  pumps.  Examples of
activation strategies for other facilities are described
above (see "Design Details").

More sophisticated  controls are required  for influent/
effluent pumping, facility dewatering, and hypochlorite
dosing for disinfection.

Pumping Systems (Influent/Effluent)

Influent and/or effluent  pumping commonly is controlled
by wetwell elevation, as in a standard sanitary pumping
station. Wetwell  design,  pump  design,  and  control
strategy must consider the peaked  hydrograph typical
of storm flows and must either attenuate or be able to
respond to relatively rapid changes  in flow rate. Some
type of variable speed  control is  used for pumps.

Screw pumps are well suited for CSO applications since
the pump discharge varies directly with influent flow.
The   drawback  to  screw  pumps  is   the   space
requirement, especially for  higher lifts. Influent pumps
must be capable of passing solids that fit through coarse

Overflow relief must be provided to protect the facility
from flooding if the influent or effluent pump capacity is
exceeded. Relief could be  in the form of an overflow
weir, which activates when a wetwell backs up, or it
could be a remote regulator, which works in conjunction
with a modulating gate. On high level, the gate would
throttle, causing an  upstream regulator to activate. It is
advisable to have simple backup, such as a  mercury
float  switch, which would  close the facility  gate
completely on high-high level in the facility, in case the
gate modulating controls fail.

Facility Dewatering

As discussed under Process Theory, initiation of a
facility  dewatering  sequence typically  is  a manual
operation.  The operators   must ensure  that  the
conveyance system and POTW  capacity have  been
restored following  the storm.  Programmable  logic
controllers allow for automatic control and switch-over
from rapid dewatering pumps to solids handling pumps
once  the dewatering  sequence  is  initiated.  Where
gravity dewatering is possible, some facilities feature a
flow control valve for more precise regulation of flow to
the interceptor.
Grit can be troublesome  for pumpback operations.
Some facilities  have a separate means to  handle grit
and heavier solids, such  as providing  access  for
removal via a vacuum truck. Other facilities provide a
separate  set of pumps  for solids  removal.  Some
facilities aerate and mix the tank contents to facilitate
solids removal.


A detailed discussion of disinfection control strategies is
provided later in this chapter.

Process Variations

The design of  CSO storage/sedimentation facilities is
not standardized. Some early facilities were intended to
be experimental prototypes  for  evaluating emerging
technologies.   Process variations that   have  been
implemented are described below.

Fine  Mesh Screening with Sedimentation

This process variation is intended to augment solids and
floatables  removals  through  storage/sedimentation
tanks. A potential drawback to this variation is that static
fine screens tend to easily clog.
• MWRA Cottage Farm: Horizontal fixed screens with
  0.2-inch openings are located  between  the scum
  baffle and effluent weir. The screens are hinged to
  allow them to swing open  by hydraulic pressure if
  clogged.  Flow passes under  the baffle, up through
  screens,  then  over the  weir.  The  screens  are
  manually flushed during tank cleanup following  a
  storm (U.S. EPA, 1977a).

• Atlanta,  Georgia,  Intrenchment  Creek:  50-mesh
  (300-u.m) static screens are installed upstream of the
  sedimentation tanks. The screens were  subject to
  blinding during the  first flush  and when there were
  heavy grease loads (West et al., 1990).

Aeration/Mixing of Tank

At some facilities  where the tank is not intended to
function as a flow-through treatment facility,  aeration
and/or mechanical  mixing prevents sludge from turning
septic  and generating  odors,  and facilitates  solids
handling.  If aeration or  mixing  is  considered,  the
potential benefits  should be evaluated  against  the
additional operation and maintenance costs associated
with the aeration or mixing equipment.

• Decatur,  Illinois,  Seventh  Ward and Lincoln Park:
  Blowers provide  aeration until the first flush tank can
  be  dewatered. Aerating the tank  keeps contents
  mixed, minimizes odor problems, and makes the tank
  easier to clean after a storm (BGMA and CMT, 1977).
• Milwaukee,   Wisconsin, Humboldt Avenue:  Seven
  mechanical mixers resuspend solids once the storm
  is over to facilitate dewatering. Mixing reduces the
  cleanup effort after a storm (Medina et al., 1981).

Flow Balance  Method

The in-receiving water flow balance method  involves
using floating pontoons and flexible curtains to create
an in-receiving  water storage facility. CSO flows fill the
facility by displacing the receiving water that  normally
occupies the storage facility.  The  CSO flows are then
pumped to the collection system following a storm. The
technology has been used for CSO control in Brooklyn,
New York,  where floating pontoons were permanently
installed in the receiving water near the CSO outlets.
The feasibility  of this technology depends in  part on
whether the storage facility would significantly  impact
the  aesthetic  value  of  the surrounding  area,  and
whether the structure would hinder navigation. Other
site-specific concerns include the  availability of volume
due to tidal variations in coastal water, and the need to
protect from damage due to high winds or wave action.
Standard marina technology was used  in the Fresh
Creek,  Brooklyn  demonstration  project to  prevent

Deep Tunnel Storage


Deep  tunnel  storage often  is  considered  as  an
alternative to near-surface  storage/treatment facilities
where  space   constraints,   potential   construction
impacts,  and other  issues challenge the feasibility of
near-surface  facilities. The major advantage of  deep
tunnel storage is that relatively large volumes can be
stored and conveyed with little disturbance to existing
surface features. In  congested urban  areas,  near-
surface CSO control facilities can be difficult to site, and
the only available open spaces for  such  facilities often
include recreational areas such as river-front parks or
ball fields. Deep tunnel construction may  avoid some of
the issues arising  from the use of open spaces and
minimize the  disruptions associated with the extensive
open-cut  excavation   associated  with   near-surface

Deep tunnel construction does not completely avoid the
issues related to siting of open excavations, however,
as  excavations at  work shafts, access shafts, vent
shafts,  and  drop  shafts  are required,  along  with
excavations for  near surface consolidation conduits.
Handling and disposal of excavate can also  present
challenges in coordinating truck traffic and identifying
disposal  locations.  In Rochester,  New York, tunnel
excavate was successfully used to fill low areas around
an  airport and the  POTW,  as well  as other  areas
identified  by  the owner.  Allowable truck routes  were
specified to minimize impacts on local neighborhoods.
The feasibility of deep tunneling must be established
through  geotechnical  investigations,  as  well  as
evaluation of other  concerns such  as  the potential for
encountering  hazardous  wastes,  the  impacts  on
adjacent structures, and construction logistics.

A typical  deep tunnel system includes  the following

• Regulators, to divert and  control storm flows to the
  tunnel system.

• Consolidation   conduits,   to  convey  flows  from
  regulators to the tunnel system.

• Coarse screening, to remove large debris and protect
  downstream pumps.
• Vertical dropshafts, to deliver flow to the tunnel and
  dissipate energy.
• Air  separation  chambers, to allow release  of air
  entrained in the dropshafts.

• Tunnel, sized to store and convey flows from a given
  design condition.
• Access  shafts,  for  maintenance  personnel  and

• Vent shafts, for balancing air pressure.

• Dewatering system, to pump volume stored  in the
  tunnel to the POTW once conveyance system  and
  treatment capacity is restored.

• Odor control systems at certain venting locations.

Design considerations for each of these components
are described below.


Regulators control  flow  into the  interceptor from  a
combined trunk sewer and serve as the diversion point
for routing excess flows to the tunnel system. Regulator
types are  reviewed earlier in this chapter.

In  the  Metropolitan Water Reclamation District of
Greater Chicago's Tunnel and Reservoir Project (TARP),
sluice gates in the regulator structures regulate flow into
the tunnel system and can be opened in advance of an
expected storm (Dalton and Goyal, 1989).

Consolidation Conduits

Routing a deep tunnel to provide direct interception of
every CSO outfall in a community's  combined  sewer
system is not feasible. Rather,  construction of near-
surface consolidation conduits from selected outfalls or
regulators to a more centrally located deep tunnel is a
cost-effective alternative. The feasibility  and cost of
constructing consolidation conduits may be one of the
factors that influences the  configuration,  depth,  and
route of a deep tunnel system. Conversely, the route of
the tunnel system  may  dictate the  requirements for
consolidation conduits. Aspects to consider in evaluating
the use of near-surface consolidation conduits include:

• Potential disruptions to  traffic,  utilities, access to
  businesses, and other impacts on the community and
  the environment during construction of consolidation

• Cost of consolidation conduits compared to cost of
  multiple drop shafts; minimizing  the number of drop
  shafts typically will be cost effective.

• Impacts of near-surface  soil conditions on consolidation
  conduit construction methods (wetlands, unsuitable
  soils, bedrock elevation, etc.).
• Impacts of subsurface geology on tunnel construction
  methods and tunnel routing.

Consolidation conduits are  sized  based  on  the peak
flows from the design condition for which CSO control
is to be provided. For example, suppose a community's
long-term CSO control strategy is to route all flows from
a particular design storm to the deep tunnel system. A

computer model of the community's combined sewer
system, such as SWMM, generates values of the peak
flow from the design storm at each node where the
proposed consolidation conduits intercept flow. These
peak  flows  are used  for preliminary  sizing  of the
consolidation conduits. The model then is expanded to
include the proposed conduits, and subsequent model
runs are used  to refine  the  peak flows,  particularly
where one  conduit intercepts the flow from more than
one node. Through this iterative process, the optimum
size of the consolidation conduits can be developed.

Consolidation conduits sized to convey  peak storm
flows  may  be of substantial  size, and may  present
opportunities for  in-line  storage.  Gates  on  the
downstream end of consolidation conduits  can  be
operated in conjunction with a real-time control system
to optimize the use of storage and conveyance capacity
in the  tunnel   system.  Construction  methods  for
near-surface consolidation   conduits  can  include
open-cut excavation and soft ground tunneling.

Relief  points  must  be  provided  for  flows   to the
consolidation conduits  in excess of the design storm
peak flow. A convenient location for relief is where the
flow  enters the  consolidation conduit,   as these
structures are commonly located at existing regulators
or on  existing  outfalls.  In  some  cases,  one relief
overflow can serve a number of hydraulically connected
CSOs,   thus   reducing   future   monitoring  and
maintenance costs by reducing the number of overflow

Coarse Screening and Grit Removal

At the  downstream end of consolidation conduits,
coarse screening equipment can remove large, bulky
solids such as branches and logs before the flow enters
the dropshaft. For a system  with  a large number of
dropshafts,  providing  coarse screening  at  each
dropshaft may not be feasible, both in terms of cost and
operation and  maintenance  requirements.  For this
reason, screening facilities may be located in the tunnel
itself.  Coarse  screening and  grit removal  can  be
accomplished by screens and sumps located just before
the dewatering  pump station, usually a low point in the
tunnel system.  Screenings can be  removed by a rake
mounted on a  bridge crane  located at the  ground
surface. The rake is lowered to the underground trash
rack to retrieve and raise screenings to the surface for
disposal. A clamshell bucket can remove grit  material
from a sump located ahead of the  dewatering pumps.
The Chicago TARP system incorporates trash and grit
removal facilities at the intake tunnels of the dewatering
pump station (Variakojis and Quintanilla, 1989). In the
Rochester, New York,  deep tunnel system, screening
facilities were provided at the downstream ends of each
of the two main branches of the tunnel system. Coarse
screening equipment is discussed in more detail later in
this chapter.

Vertical Dropshaft

The function of the vertical dropshaft is to deliver flow
from the near-surface  conveyance system to the deep
tunnel system, dissipating the energy in the flow to the
extent possible and providing a means to  remove air
entrained  in the flow as it  passes down the shaft.
Dropshaft design considerations  and dropshaft types
include (Westfall, 1990):

Dropshaft Design  Considerations

Dropshafts have three basic components:

• Inlet structure, to provide the transition  between
  horizontal and vertical flow.

• Vertical shaft barrel,  to convey flow to the lower
  elevation and dissipate energy  in the flow.

• Bottom chamber, to dissipate energy in the vertical
  flow and provide a  means to separate and release
  air entrained in the dropshaft.

Dropshaft design is influenced by one or more of the
following factors (Westfall,  1990;  St. Anthony Falls
Hydraulic Laboratory, 1971).

Variable Discharge

CSO outfall hydrographs tend to show distinct peaks. A
consolidation conduit connecting a number of outfalls or
regulators to a dropshaft may provide some attenuation
of peak flows,  but the  dropshaft still will be exposed to
a range of flows and  must be capable of functioning
within the expected range.

Impact on Dropshaft Floor

The impact of the flow on the dropshaft floor depends
on the depth of the drop. For even relatively shallow
drops, provisions  should be made  to  reduce the
magnitude of the impact. Alternatives for dissipating the
energy in the flow include inducing a hydraulic jump in
the  shaft,  creating a  vortex  action  in  the shaft,
increasing wall friction  in the shaft,  and providing a
plunge pool in the  shaft floor. The plunge pool can be
created by forming a sump in the floor or by providing
a weir downstream of the shaft  to back up flow. The
depth of a plunge pool is determined from the Dyas
formula (WPCF/ASCE, 1974):
 h = depth of drop (ft)
 dc = critical flow depth in shaft inlet (ft)

Entrained Air

The air that becomes entrained in the flow as the flow
passes down the dropshaft can be both beneficial and
detrimental to the dropshaft operation. The benefits of
entrained air include:

• Minimizing    the   potential  for   subatmospheric
  conditions in the shaft, which would cause cavitation.

• Providing a cushion to absorb the impact of the falling

Disadvantages to entrained air include:

• Increasing  the volume  of the  flow,  which may
  increase the required dropshaft size.

• Requiring  a  separate  means  for removing the
  entrained air, to avoid a build up of pressure.

Headless in the Dropshaft

If the hydraulic gradeline in the tunnel is close to the
hydraulic  gradeline in the near-surface conveyance
system, excessive headless in the dropshaft can cause
flows to back up in the near-surface system.

Surge Relief

Six out of a total  of 40  vertical  dropshafts  in the
Rochester, New York, tunnel system were oversized to
provide surge relief from transient hydraulic pressures
that could develop in the tunnels (Holzbach, 1990).

Types of Dropshafts

Different types  of  dropshafts have been  developed
based on hydraulic studies. Selection of dropshaft type
depends  on  the relative  importance of the design
considerations noted above. Four of the more common
dropshaft types  are  discussed below (Westfall, 1990).

Drop Manholes

Drop manholes are used in near-surface conveyance
systems to drop flow from a higher sewer into a lower
sewer. The manholes  minimize turbulence  that would
otherwise  promote  the release of sewage  gas and
erosion of the  manhole.  A separate access  manhole
commonly is  provided to  the lower sewer,  so that
maintenance personnel can avoid climbing down the
wet shaft.  Drop manholes are suitable for drops up to
70 feet. For drops of greater than 70 feet,  one of the
following dropshaft types should be considered.

Vortex Dropshafts
Flow is introduced  into vortex dropshafts tangentially,
causing flow to spiral down the shaft in a vortex flow
pattern. Centrifugal  forces keep the flow in contact with
the wall, dissipating  energy through friction and creating
an open inner core, through which entrained air may
escape.  The vortex flow pattern forces air bubbles
toward the center core, further reducing the amount of
entrained air. The central air core also tends to maintain
atmospheric pressure in the shaft, thereby  minimizing
the potential for cavitation. A relatively smooth transition
from horizontal to the vortex  vertical flow  minimizes
turbulence at the  inlet,  which  also  minimizes  air
entrainment. Kinetic  energy remaining in the  flow is
dissipated either through creating a hydraulic jump in
the shaft, or by a plunge pool at the bottom of the shaft.
Five types of tangential inlet configurations  have been
developed (Figure 4-8), including (Westfall, 1990):
• Circular: The dropshaft is concentric with the vortex
  inlet, which has a horizontal floor.
• Scroll: The sides of the vortex inlet, with a horizontal
  floor, curl  towards the dropshaft.
• Spiral: The approach channel issues  a vortex flow
  into the dropshaft by winding downwards.
• Tangential: The  approach channel  contracts at its
  junction with the dropshaft so as to project the flow
  tangentially around the walls.
• Siphonic: Generally used for outlets from a reservoir,
  a series of siphons are located  around the entrance
  so as to produce a vortex flow down the dropshaft.
Hydraulic studies have suggested that the  spiral  and
tangential inlets perform best, while the tangential inlet
is easier to construct (Westfall, 1990). The vortex shaft
results in minimal air entrainment and significant energy
dissipation,  but headloss is significant. A vortex type
might  not be  appropriate where the  difference in
hydraulic gradient  between  the  tunnel   and   the
near-surface conveyance system is minimal.

Morning Glory
These drop structures feature a circular crested weir for
inlet control and  commonly  are used for reservoir
outlets. Flow characteristics are determined from weir
control, orifice control, and differential head  control.

Direct Drop Air Entraining Type
The inlet structure for this dropshaft allows flow to enter
the dropshaft radially, entraining air to the  extent that
the shaft flows  full. The full-flowing  shaft promotes
energy   dissipation  through   wall friction,  and   the
entrained air provides a cushion for absorbing energy
at the  bottom of the shaft.  A large air  separation
chamber is  required at the bottom of the shaft, along
with a separate venting system  to evacuate the air
before it enters the tunnel system.
Two variations of the direct drop style were developed
for Chicago's TARP system. Both  types are capable of

                        (a) Circular
                    (b) Scroll
                              (d) Tangential
                                                      (c) Spiral
                                                                                           (e) Siphon
Figure 4-8.  Examples of tangential inlet configuration (Westfall, 1990).

passing  a  wide  range  of  flows  at  approximately
20 percent of the headloss of similarly sized vortex-type
dropshafts. The E-15 (Figure 4-9) features an air vent
located inside the drop shaft. Slots in the vertical wall
separating the dropshaft from the air vent allow air to
be re-entrained in the falling water. Recirculating the air
provides  a  more consistent  mix  of  air  and water,
resulting in a more uniform flow in the dropshaft. The
air  separation chamber at the  bottom of the shaft
features a sloped roof to direct the released air back
towards the vent. The E-15 shaft was used for shaft
diameters up to 9 feet, with a maximum flow of 600 cfs.

The  D-4 dropshaft  was developed  for dropshaft
diameters greater than 9 feet (Figure 4-10). The air vent
for the D-4 shaft is a separate shaft, located downstream
of the dropshaft. The air vent feeds into the dropshaft
above the crown of the  incoming sewer, so air  is
recycled to the top. The air separation chamber for the
D-4 has a flat roof. The D-4 is suitable for dropshafts up
to 20 feet in diameter, and flows up to 4,500 cfs.

The air separation chambers  and shaft structures for
both   types   of  direct-drop   dropshafts  are large
structures, and must be suitably anchored to withstand
the forces and vibrations generated by the falling flow.
The floors of the air chambers in the TARP system are
lined with a  metal coating to minimize erosion from grit
carried in the flow. Drop shafts initially may be oversized
             to provide access to the tunnel during construction. The
             finished diameter of the dropshaft is sized based on the
             following  equation (Westfall, 1990):
                                         = 0.2
             Q = design flow (cfs)
             g = acceleration of gravity (ft/sec2)
             D = finished shaft diameter (ft)

             Over 200 of the direct-drop  style dropshafts, some of
             which have  been operating for 20  years,  have been
             installed in Chicago's TARP system.


             Sizing and routing of deep storage tunnels, as with any
             major CSO control  system, is a  complex  process,
             requiring the review of a series of alternatives before
             arriving  at  the optimal  configuration. The  storage
             volume required to meet a given CSO control goal is
             developed through the  use of  a system model  such
             as SWMM,  or a  more simplified  routing  technique.
             Alternatives encompassing variations in tunnel diameter,
             length, route, depth,  and construction methods that
             meet the required storage and conveyance needs are
             then developed and evaluated. Topics to consider in
      Shaft I.D.-r4
              L>C    ELEVATION     30oto75Q
            E .    K        L
          Design Discharge
          Shaft I.D.
          W (Min)
          N (Min)
          N x P (Min)(ft2)
          S (Min)
          Head Loss
                  SECTION B-B
2' -9"
10' -0"
1' - 0"
6' -8"
5' -8"
2' - 10'
2' -6"
20' -3 1/2" 15' -01/2" 11' -2 1/2"
12' - 0"
27' - 0"
29' - 0"
56' -0"
Equal to Pipe Dia.
6' -8"
20' - 3"
21' -6"
41' -91
3' -6"
(But Not To Exceed I.D. of Shaft)
14' -10"
16' - 0"
30' -10"

32' - 6"

24' - 0"
 Figure 4-9.  The E-15 dropshaft (Westfall, 1990).

t n

/-T\ Vl
o r:

I, F
ea uover

P ~"

                                TOP VIEW
Inlet Conduit
                                     Air Shaft
                              SECTION A-A

Figure 4-10.  The D-4 dropshaft (Westfall, 1990).

                 Air Separation
                                          Design Discharge
                                          Head Loss
20' - 0"
48' - 0"
40' - 0"
25' - 0"
55' - 0"
90' - 0"
20' - 0"
28' - 0"
36' - 0"
30' - 0"
19' -0"
42' - 0"
67' - 0"
15'- 0"
                                                      SECTION B-B
29' - 0"
24' - 0"
33' - 0"
53' - 0"
17' - 0"
developing tunnel sizing  and  route  alternatives are
discussed below.

Subsurface Conditions

A  substantial  subsurface  exploration  program  is
necessary to evaluate the feasibility of deep tunneling
and   to  identify  the  most   appropriate  tunneling
techniques.   Depth   to  bedrock,   rock  strength,
discontinuities and weaknesses in the rock structure,
and ground-water conditions can impact the selection
of tunnel route and construction method. Softer rocks
allow a faster penetration rate, but also may require
additional temporary and permanent support structures.
Harder rocks, while  requiring  less support, generally
require   a   slower  rate  of  excavation.  Excessive
ground-water  may  cause  tunnel  flooding  or local
collapse, and may require  special ground  treatment
techniques such as dewatering, grouting, or ground
freezing. Discontinuities, fault zones, and other areas of
local  weakness or variation may preclude the  use of
certain tunneling techniques. The elevation of the top of
the bedrock is an important consideration in setting the
vertical alignment. In some cases, it may also be cost
                             effective to route the tunnel at a deeper elevation if
                             doing so would  avoid difficult areas or "mixed face"
                             conditions  (Thompson  and   Dobbels,  1991).  In
                             Rochester,  New York, the  surface topography would
                             have allowed the downstream leg of the tunnel system
                             serving the west side of the city to be set at a constant
                             grade to the POTW. While  allowing  gravity flow to the
                             POTW, the downstream end of this vertical alignment
                             would  have  risen above  the  top  of bedrock. The
                             designer preferred to  keep  the entire segment in rock
                             to ensure maximum support against transient pressure
                             waves when the tunnel was filling or flowing full. The
                             final vertical  alignment,  therefore, consisted of  an
                             inverted siphon, which allowed a tunnel boring machine
                             (TBM) to excavate at a constant but steeper grade
                             entirely in rock to a downstream control structure, at
                             which point the rising  leg of the siphon was excavated
                             up to the POTW (Holzbach, 1990).

                             Encountering hazardous  materials can have a severe
                             effect on project schedule and cost. Delays may occur
                             to identify the nature and extent of the wastes, develop
                             and implement disposal and/or treatment options, and
                             obtain permits. An environmental assessment along the

route of a  proposed  tunnel  and  at  proposed  shaft
locations allows the  designer to identify routes that
avoid areas of  known or potential contamination. A
limited review for preliminary siting might include review
of available agency files and data on known locations
of oil or hazardous materials spills, and review of past
and   present  land   use  along   proposed  routes
(Thompson  and  Dobbels, 1991). During design, a more
detailed study, including soil borings, monitoring wells,
and   sampling,   may  be  required.   The  potential
requirement for  subsurface easements from property
owners also should be investigated.

Excavation Method

Three  methods  for deep rock  tunnel excavation  are
TBMs,  road  header   machines,  and  drill-and-blast
methods. These methods are  described below;  the
advantages and disadvantages  of each technique  are
summarized in Table 4-6.


TBMs feature a  rotating, cutting  head of a size equal to
the  full  cross-sectional  area  of  the tunnel  to  be
excavated.  Horizontal hydraulic thrust cylinders force
the rotating  cutting head forward into the rock. The body
of the TBM  is supported by hydraulic legs, which push
against the  walls of the tunnel, anchoring the TBM
against the thrust applied to the cutting head. When the
hydraulic thrust  cylinders are fully extended, the main
support legs are retracted and  the TBM is  advanced
forward  to  start the next cutting  cycle.  Figure 4-11
shows the cutting cycle for a typical TBM system. The
cutting heads feature a means to remove the excavated
material from the cutting area, commonly by discharging
the  excavate to a  conveyor system. The  type and
                         configuration  of  cutters  on  the cutting  head vary,
                         depending on the anticipated geologic  conditions.  In
                         rock, roller or disc cutters are more reliable than drag
                         cutters, which  are  more appropriate  for soft ground
                         conditions (Whittaker and Frith, 1990).

                         Rock Header Machines

                         These machines feature a relatively  small-diameter
                         boom-mounted rotary cutting head, supported by a
                         tracked base. The cutting head is worked back and forth
                         across the rock face by manipulating the boom. The
                         rock debris is picked up at the base of the rock face by
                         a conveyor system  running through the base of the
                         machine.  The most  common boom machines are the
                         milling-type and ripping-type. The milling-type generally
                         is best suited for hard rock conditions and features a
                         conical cutting head that rotates in the same axis as the
                         boom. The ripping-type  cutting head  rotates  on a
                         horizontal axis perpendicular to the boom and  is best
                         suited for softer rock conditions  (Whittaker and Frith,

                         Drill-and-Blast Methods

                         The steps involved  in drill-and-blast methods include
                         (Whittaker and Frith, 1990):

                         • Drilling: The number of blast  holes required in the
                           rock face depends on the diameter of the tunnel and
                           the strength of the rock. In addition to the blast holes,
                           a  number  of  burncut  holes,   without  explosive
                           charges, are drilled into the rock  face to relieve the
                           explosive stress. Drilling rigs may be mounted on the
                           haulage tracks or  on separate  tracks of wider gauge
                           than the haulage tracks, or they may be rubber tire-
                           or tread-mounted.
Table 4-6.  Advantages and Disadvantages of Deep Tunnel Excavation Methods

Method                       Advantages
Rockheader machines
Drill and blast methods
Rapid excavation to final tunnel
diameter and grade
Disturbance of surrounding rock
Well suited for long reaches of constant
cross section

One machine can excavate different
tunnel diameters
Typically lower lead times for delivery
than TBMs
If tunneling conditions change, machine
can be easily withdrawn to allow use of
drill and blast methods

Can be used in most rock conditions
• Cutting face can become jammed where
 high rock stresses create "squeezing" condition
• Usually long lead times required to fabricate
 new machines (use of reconditioned
 machines can reduce lead time)
• Usually not economical if multiple tunnel
 diameters are required

• Rate of advancement is lower than for TBMs
• Rate of advance depends more on operator
 skill and  rock fracture patterns
• Cannot typically apply as much force to
 rock as TBMs
  Relatively slow rate of advance
  Higher potential for damage to surrounding
  rock during blasting

                               Main Legs
                                                                               Step 1.
                                                                               Start of boring cycle. Machine
                                                                               clamped, rear support legs
                                                                              Step 2.
                                                                              End of boring cycle. Machine
                                                                              clamped, head extended, rear
                                                                              support legs retracted.
                                                                              Step 3.
                                                                              Start of reset cycle. Machine
                                                                              undamped, rear
                                                                              support legs extended.
                                                                              Step 4.
                                                                              End of reset cycle. Machine
                                                                              undamped, head retracted.
                                                                              Machine now ready for
                                                                              clamping and beginning boring
Rgure 4-11.  Cutting cycle for a typical tunnel boring machine (Whlttaker and Frith, 1990).

• Blasting: The strength of the explosive charge also
  depends on the diameter of the tunnel, the  local
  geology, and the proximity of adjacent structures. The
  weight of explosive can be limited by a specified
  maximum allowable  blasting vibration.  Following
  detonation, smoke fumes must be exhausted before
  personnel return to the blasting areas. Reversible
  fans in conjunction with ventilation tubes minimize the
  exposure of workers to the blasting fumes.

• Debris Clearance: Once personnel can safely return
  to the blasting areas, the roof, face, and sides of the
  tunnel are visually inspected and the loose  rock is
  removed.  A variety of methods and machines are
  available  to  handle  excavation,  including  rocker
  shovels and scraper action loaders, which discharge
  into rail-mounted cars, conveyors, or slurry pipelines.

• Ground Support: Temporary supports protect personnel,
  control overbreak, and support loosened rock around
  the perimeter of the tunnel.  Even tunnels in hard,
  competent rock may require temporary supports for
  localized weak zones. Common types of temporary
  support include arched  steel ribs, rock bolting, and
  shotcrete. Permanent linings in drill-and-blast tunnels
  typically  are   installed   once  the   excavation  is
  complete, although in larger tunnels, excavate removal
  and  concrete  placement  can  be  more  readily

In general, if geologic conditions are suitable for using
a TBM,  then tunnel routing  and  sizing should be
directed toward longer reaches of constant diameter
tunnel. Straighter tunnels are usually preferable, since
TBMs  have  a limited turning radius.  If  subsurface
conditions  preclude  use  of  TBMs,  then excavation
volume can be optimized by using tunnel segments with
different  cross-sectional  areas. In any case,  work
generally proceeds upgradient to minimize the opportunity
for  ground-water infiltration to pond at the work  face.
The direction of excavation can influence the location of
work shafts and tunnel layout.

Consolidation Conduit Layout

As  noted  above, tunnel routing must be developed in
conjunction with  near surface consolidation  conduit
design. The  capacity  for storage  in consolidation
conduits and upstream collection systems may reduce
the size and/or location of tunnels required.

Potential Operating Strategies

Strategies  to  control  flow into and through a tunnel
should be considered  when developing the optimal size
and configuration of the tunnels. For example, certain
diversion structures in the  Rochester, New York, tunnel
system are capable of  diverting flow to two different
sections of the tunnel system. If one system of the
tunnel  fills  more  rapidly than  the other,  remotely
controlled gates redivert flow to the tunnel section with
greater remaining  capacity  (Kent,  1992).  Designing
control structures  with  automatic gates allows such
operating  strategies,   especially  if   provided   in
conjunction with a real-time control system.

Access, Vent, and Work Shafts

Work shafts and access shafts are required to move
personnel, equipment, and materials in and out of the
tunnel  during construction and  once the tunnels  are
operational.  The size of construction work shafts may
be dictated  by the size of the  excavation machinery
used. The land around a work shaft commonly is used
as a staging area  for equipment and supplies, and is
subject to  heavy vehicular traffic  associated  with
excavate  removal. One study  estimated  that 2 to
3 acres are  required at each construction access shaft
location (Wheatley, 1991).
Permanent access shafts for tunnel maintenance  and
inspection may be developed from construction work
shafts  or incorporated into other structures, such as
screening houses or control structures.
Vent shafts provide for the passive movement of air in
and  out  of  the tunnel  during  filling and dewatering,
responding  to differences  in  pressure between  the
tunnel  and the  above-grade atmosphere. Vent shafts
may be provided with odor control. In Rochester, New
York, intermediate vent shafts  are  not provided with
odor control, whereas carbon adsorbtion odor control is
provided at a screening  facility on the upstream end of
the inverted siphon  to  the  POTW, and also at the
downstream end of the  inverted siphon. In addition to
odor-causing  compounds,  emissions  can  include
volatile organic compounds (VOCs).

Vent and access shafts typically feature a concrete pad
at grade, with either gratings or an appropriate cover to
secure the opening.  The  total  area required for
permanent access or vent shafts is 1/4 acre or less.

Dewatering Pump Station

Deep tunnel storage and conveyance systems typically
feature  a  dewatering  pumping   station  at  the
downstream end of the tunnel system, although gravity
drainage systems do exist. The pumping station may be
dedicated to the tunnel  system  or integral to a POTW
sanitary influent pumping station. If coarse screening is
not  provided at upstream locations,  then screening
facilities  should be provided  at  the  pump station.
Figure 4-12 shows a tunnel system dewatering pump
Tunnels  are sloped to ensure a  sufficient  carrying
velocity and to facilitate dewatering. A tunnel slope of
0.1 percent has resulted in minimal problems with grit

            Screenings Handling Area
  Scrubber Discharge

        Fan Building

   Clamshell -
  Rock     Screenings
                                                          Bndge Crane

                                                          Pump Chamber
                                  {  Egress Shaft
Figure 4-12.  Tunnel system dewatering pump station (MWRA, 1990).
deposition in Rochester, New York. Though much of the
pumping station  equipment  may  be located  below
grade, an above-ground  superstructure is  required to
house screenings handling equipment,  odor control
equipment,  and/or  electrical  and  operations control
equipment. Dewatering rates are determined by POTW
primary and secondary treatment capacities.

Active tunnel air venting  may be provided at pumping
stations for use during dry periods, when  anaerobic
conditions in tunnel sediment can produce  odors. The
method of venting can be by drawing air into the tunnel
through drop shafts and access shafts, and venting at
the pumping station,  or  by pushing air through  the
tunnel and out the shafts. The selected direction of air
movement may depend on the desired location for odor
control facilities and the location of sensitive receptors.

In Rochester, New  York, combined sewage from  the
east side of the city feeds a tunnel system that drains
to a 275-mgd pumping station.  This  pumping station,
which also handles sanitary flows, features nine pumps,
with the largest having a  capacity of 38,200 gpm. The
station provides a total lift of approximately  145 feet,
and was designed to allow expansion to a  capacity of
400 mgd.
Flows from the west side of Rochester feed a second
tunnel  system, which drains  by gravity to the POTW.
This tunnel system was originally planned to pass under
the Genessee River, with a gradient which would have
required downstream pumping to the POTW. The local
topography,  however,  would  allow  a conduit to  be
suspended over the Genesee River, which in turn would
raise the tunnel gradient high enough to allow gravity
flow to the POTW. A second benefit of this design was
that the  crossing of the Genesee  River gorge was
designed with an  integral pedestrian walkway, providing
access between  two parks on  opposite sides of the
river. At the downstream end of the inverted siphon on
the west side tunnel, gates allowed the siphon to drain
to the  main pumping station. This operating strategy
makes the volume of the siphon available for storage at
the start  of  the  storm  event  (Kent,  1992; Holzbach,

The pumping station on  the Chicago TARP Mainstream
System features  pumps housed in  two independent
underground  chambers, in  order to provide  backup
capacity.  Each housing can accommodate four pumps.
The pumps provide service at two rated heads, 150 feet
and 330  feet; therefore, the required pumping  head
range from 60 feet to 330 feet is obtained by throttling
cone valves. Vertical, single-suction volute pumps with

vertical, constant-speed motors are used, with pump
casings encased in concrete to support the motors and
control vibration. Energy for the 85,000 hp of pumping
(57,000 hp of high head and 28,000 hp of low head) is
supplied through two independent substations that tie
into  independent transmission lines with switchover
capability  (Variakojis and Quintanilla,  1989). Table 4-7
summarizes the capacities of the pumps (Dalton and
Goyal, 1989). Once a storage reservoir  is completed,

Table 4-7.   Chicago TARP Mainstream Pumping Station

No. of
Pumps    Rating at Maximum Head     Head Range
North Pumphouse

   1       330 cfs (213 mgd) at 330 ft

   1       220 cfs (142 mgd) at 330 ft

   2       330 cfs (213 mgd) at 150 ft

South Pumphouse

   1       330 cfs (213 mgd) at 330 ft

   1       220 cfs (142 mgd) at 330 ft

   1       330 cfs (213 mgd) at 150ft

   1       160 cfs (103 mgd) at 150ft
240-330 ft

240-330 ft

60-150 ft (1 future)

240-330 ft

240-330 ft

60-150 ft (future)

60-150 ft
this pumping station will have the capability of operating
in three modes:

•  Pumping from the tunnels to the POTW

•  Pumping from the tunnels to the storage reservoir

•  Pumping from the storage reservoir to the POTW

The pumping rate of storm flows to  the POTW is
controlled to maintain  a maximum flow to the POTW of
1.5 times DWF. Average DWF to the POTW is 903 mgd,
which leaves 455 mgd of capacity available for treatment
of storm flows (Dalton and Goyal, 1989).

A supervisory control and monitoring system  (SCMS)
provides the following remote control functions at the
Mainstream Pumping  Station (Dalton and Goyal, 1989):

•  Pump control  and monitoring

•  Power circuit breaker control  and monitoring

•  Dewatering tunnel gate control and monitoring

•  Tunnel discharge valve control and monitoring

•  Closed  circuit television control

 Tunnel System  Operation

Deep  tunnel storage  systems   can  have   various
operating  strategies.  The simplest operating  strategy
allows the  system to fill with  no restrictions  at the
dropshafts until the tunnels, dropshafts, and consolidation
conduits  are  filled  and  an  overflow  occurs.  An
alternative  operating  strategy is to prioritize areas
served by the system.  Under this strategy, flows to a
dropshaft of low priority are throttled in order to allow
inflow from a higher priority dropshaft. A higher priority
is assigned to a dropshaft that serves an area with a
more sensitive receiving  water.  A  real time control
system can be used to operate the  throttling gates. If
the volume  of captured CSO does not exceed  the
storage capacity, the stored  flow is pumped to  the
treatment facility during the  dry  weather period. The
Chicago TARP system  pumps excess volume  to large
pit-type storage reservoirs when  the tunnel system is
filled to capacity.

Storage of combined flows allows solids deposition in
the tunnel system. In unlined tunnels, a concrete invert
increases  velocities  and  reduces  solids  deposition.
Solids can be resuspended  by  flushing the  tunnels
during dry weather. Sources of flushing water include
existing interceptor sewers or surface waters. Flushing
water can be removed using high head pumps  capable
of handling the high  solids  concentration expected
during flushing. The required volume of flushing water
is determined by the diameter, length, and grades of the
tunnel sections.

Coarse Screening

Process Description

Coarse screening equipment, consisting of vertical or
inclined steel bars spaced evenly across a channel, with
or without mechanical raking apparatus, traditionally is
located at the headworks of POTWs to remove from the
influent flow large objects that might  otherwise  damage
downstream  equipment  or  clog downstream pipes.
Depending on the  clear spacing between the bars,
coarse screens also can entrain rags and floatables.
Coarse screening equipment is installed at CSO control
facilities,  both  for  the  protection of  downstream
equipment and to provide floatables  removal.

The types of bar screens used at CSO control  facilities
include trash racks, manually cleaned screens,  and
mechanically cleaned screens. The major features of
each type are (Metcalf & Eddy, 1991 a; WEF, 1992):

• Trash racks
  - Typically 1.5- to 3-inch clear  spacing between bars
  - Intended to remove large objects such as timber
     planks and stumps
  - Often followed by bar screens  with smaller clear

•  Manually cleaned bar screens
  - 1- to 2-inch clear spacing between bars
  - Bars set 30 to 45 degrees from the vertical

  - Screenings are manually raked onto a perforated
    plate for drainage prior to disposal
  - Commonly used in bypass channels for mechanically
    cleaned bar screens
• Mechanically cleaned bar screens
  - 0.25- to 1-inch clear spacing between bars
  - Bars set 0 to 30 degrees from the vertical
  - Electrically driven rake mechanism either continuously
    or periodically removes material entrained on the
    bar screen
  — Common types of mechanically cleaned screens
    include chain-driven, with front or back cleaning;
    reciprocating rake; catenary; and continuous

Automatically activated, mechanically cleaned screens
are   recommended   for   CSO  facilities.   Of  the
mechanically cleaned bar  screens,  the catenary type
most  commonly  is selected for CSO control facilities
(Figure  4-13).  Catenary  screens  are  rugged  and
reliable. All sprockets, bearings, and shafts are located
above the screenings channel, reducing the potential for
damage   and   corrosion    and  facilitating   routine
maintenance. The cleaning rake is held against the bars
by the weight of its chains, allowing the rake to be pulled
over large  objects that are stuck in the bars  and  that
might  otherwise  jam  the  rake   mechanism.  The
chain-driven, front or back cleaning types  are better
suited for  separate  sanitary  flows,  and  may be
susceptible to jamming and increased maintenance if
exposed to the  range of debris present in combined
flows. Reciprocating rake screens have limited capacity
to handle  peak  screenings  loads.  Continuous self-
cleaning screens tend to  have a  higher capacity for
solids handling than the more traditional design,  and
have  potential for application at CSO control facilities.
More detailed descriptions of screening devices are  in
the literature (Metcalf & Eddy, Inc., 1991a; WEF, 1992;
U.S. EPA,  1977b).

Process Design

The following discussion of design considerations for
coarse   screening  equipment  includes   hydraulic
considerations, equipment details, solids handling, and
process  flow. Supplemental information is available
from  equipment  manufacturers   and the   literature
(Metcalf & Eddy, Inc., 1991 a; WEF, 1992).

Hydraulic Considerations

A straight upstream channel provides the most uniform
flow velocity distribution across a screen, which, in turn,
will result in a more even distribution of solids entrained
on the screen. The headloss across a clean bar screen
can be estimated from the following equation (Metcalf
& Eddy, Inc., 1991 a):
                      0.7^  2g  J
 hL = headloss, ft (m)
0.7 = an empirical discharge coefficient to account for
     turbulence and eddy losses
 V = velocity of flow through the openings of the bar
     rack, ft/s (m/s)
  v = approach velocity in  upstream channel, ft/s (m/s)
  g = acceleration due to gravity, ft/s2 (m/s2)

For coarse screens, the approach velocity should be at
least 1.25 ft/s to minimize deposition, while the velocity
through the bars should be less than 3 ft/s to prevent
entrained solids from  being  forced through the bars
(Metcalf & Eddy, Inc., 1991a). Instrumentation  provided
with mechanically cleaned bar screens is configured to
trigger cleaning cycles  so that headloss across the bar
screen is limited to 6 inches.

At POTWs, solids deposited in the screenings channel
during  periods  of  low  flow  sometimes  can  be
resuspended during peak flow periods. At CSO control
facilities, however, peak flows  generally  occur during
the early part of the storm,  after which the flow gradually
tails off.  There is typically no equivalent to the diurnal
peak flows which could repeatedly scour the bar screen
channel. In addition, towards the end of  a storm, as
flows subside, backwater from a storage/sedimentation
tank effluent weir can create quiescent conditions in the
bar screen channel. A means to flush  or otherwise
handle  solids  deposited  in  the screenings  channel,
therefore, should be provided.

A redundant or standby bar screen should be provided
so that peak flow to the facility can be maintained with
one unit out of service. Screenings channels with sets
of stop log grooves or slide  gates will allow each bar
screen   to  be  isolated  from  the  influent  flow for

 Equipment Details

 Data  have  been  published  that  relate estimated
 screenings quantities to bar screen spacing for separate
 sanitary sewer systems, but these data do not apply to
 combined   systems,   which   may   produce  higher
 screenings loading rates during  storm events (WEF,
 1992). The 0.5- to 1.0-inch  range  for bar spacing is
 common for  mechanically cleaned  screens at CSO
 control facilities.

 Suitable  guards,  railings,  and  gratings  should  be
 provided around the screening equipment to  ensure
 operator  safety.  Electrical   fittings  and   devices
 associated with the screening equipment must conform
 to the  exposure rating for  the  space  in which the
 equipment  is located  (Great Lakes-Upper Mississippi
 River Board, 1978). If  screens discharge  to screenings

Rgure 4-13.  Catenary-type mechanically cleaned bar screen (WEF, 1992).

carts or containers, hoisting and transport equipment is
needed to transfer the screenings from the containers
to a truck for disposal. A monorail-mounted hoist can
provide such capability.

Solids Handling

The quantities of screenings removed at CSO facilities
can be highly variable, depending on the configuration
of the  drainage system, the time of year, the interval
between  storms,   and  other factors.   Average CSO
screenings loads  range from  approximately  0.5  to
11 cf/MG, with peaking factors based  on hourly flow
ranging from 2:1 to greater than 20:1. The bulk density
of screenings from combined flows range from 40 to
70 Ib/cf (WEF, 1992).

Methods for  handling  screenings from mechanically
cleaned bar screens vary among existing CSO control
facilities. Examples of handling methods include:

•  Newport, Rhode Island, Washington Street: Screenings
   from this storage/sedimentation facility are discharged
   into  1-cubic yard capacity bins, which are manually
   wheeled to a monorail hoist. A lifting/dumping frame
   is attached to the bin to keep the bin in  an  upright
   position while being carried on the hoist. The frame
   features a  release arm for dumping the contents of
   the bin into a truck for disposal (Metcalf &  Eddy, Inc.,

•  Atlanta,  Georgia, Intrenchment Creek: Screenings
   from the headworks of this  physical/chemical CSO
   treatment facility are  discharged into  conveyors,
   which carry the screenings to collection bins. During
   operation material has spilled repeatedly on the floor
   and the conveyor belts have jammed. A study of this
   facility's  operations  recommended  that   future
   facilities use an  alternative  screenings handling
   system (West et al.,  1990).

•  MWRA Cottage Farm: This detention facility features
   trash racks upstream of bar racks, both mechanically
   cleaned, catenary type. The trash rack screenings
   are  deposited  into dumpsters,  while the bar rack
   screenings are discharged to a sluiceway and flushed
   back to the interceptor (U.S. EPA, 1977a).

•  MWRA Prison Point: Screenings from this detention
   facility are discharged to a conveyor belt and carried
   to a sorting table. From the sorting table, screenings
   are  fed manually into a  hammermill grinder, then
   discharged   back  to  the   interceptor.   Material
   unsuitable   for  grinding,  such   as   bricks  or
   heavy-gauge scrap metal, is removed at  the sorting
   table and disposed of separately (Maguire, 1981).
Process Flow

As at POTWs, bar screens at CSO control facilities are
located at the head of the facility, upstream of the tanks
and process equipment, but typically downstream of the
regulator that diverts  flow to off-line facilities. Some
facilities, such as the MWRA Cottage Farm and Atlanta
Intrenchment   Creek  facilities,  are equipped  with
larger-opening  trash  racks upstream  of  smaller-
opening, mechanically cleaned bar screens. As noted
above, at Cottage Farm both the trash  racks  and bar
screens  are  catenary-type  mechanically cleaned. A
performance  study at  this  facility noted that very little
material  was  collected  on the  trash  racks,  and
suggested  that manually  cleaned  trash racks would
have been  suitable (U.S. EPA, 1977a).
At the  Decatur,  Illinois, McKinley Avenue facility, a
manually cleaned bar screen with 2-inch openings was
provided as a bypass for times when the mechanically
cleaned catenary  screen was out of service. During an
operational study,  concerns were raised that the manual
screen might become  subject to clogging if forced into
operation since personnel  may not be available at the
facility (which normally is not staffed) to  manually clean
it. The study recommended that if manual  bypass screens
were provided, they should  be sized to capture only
objects  that are large  enough to damage downstream
equipment  (BGMAand CMT, 1977).

System Controls and Operation
Controls for mechanically cleaned bar screens include
some combination of the following:
• Manual start/stop
• Automatic start/stop on timer
• Automatic start/stop on differential head
At CSO facilities, activation of mechanically cleaned bar
screens is  triggered by remote sensing  of flow into the
facility, or water level  in the  screening channel. Timed
cycles,  if used, should include a high differential head
override. Near-continuous operation may be  required
during the  initial  phase of the storm when a greater
portion of leaves, litter, and other solids  may be carried
in the initial peak flows  to the CSO facility. Automatic
start-up of  the bar screens should include a time delay
to prevent premature initiation  of  operation due to
transient surges in flow.
Table 4-8 presents examples of bar screen installations
at CSO control facilities. The table indicates bar screen
type, bar spacing,  screenings handling method, and
control sequence.

Table 4-8.  Examples of Bar Screen Installations at CSO Control Facilities

Location               Facility Type
                  Bar Screen
              Bar         Screening Disposal
              Spacing (in)  Method
Newport, Rhode Island
Washington St. (Metcalf
& Eddy, Inc., 1991b)

Atlanta, Georgia
Intrenchment Creek
(Westetal., 1990)

Cottage Farm (U.S. EPA.
Storage/treatment    Catenary
Storage/treatment    Trash rack
MWRA                Storage/treatment
Prison Point (Maguire,

MWRA                Coarse screening/
Constitution Beach       disinfection
(Hayden-Wegman, 1989)
Somerville Marginal
Pretreatment Facility
(Tighe & Bond, Inc.,

Fox Point
(Hayden-Wegman, 1991)

Decatur, Illinois
McKinley Ave. (BGMA
and CMT, 1977)

Washington, DC
NEB Swirl Facility
(O'Brien & Gere, 1992)
Coarse screening/
Coarse screening/
Swirl concentrator
(trash rack)
















Discharge to 27-cf carts Auto start on high level float
                    switch; continuous operation

Discharge to conveyor

Discharge to conveyor

Discharge to dumpster  Auto start on high level

Discharge to sluiceway  Auto start on high level

Discharge to conveyor.  Control by timer, differential
to sorting table, then    head, or manual
to hammermill grinder
                                            Discharge to carts
                    Automatic start/stop controlled
                    by water level
Discharge to conveyor  Auto start on water level; stop
to storage container    after time delay following gate
Discharge to container  Auto start/stop on water level
Discharge to cart
Discharge to 80-cf
fixed bins; bins
emptied by vacuum
Auto start/stop on differential
Control by timer or differential
Swirl/Vortex Technologies

Swirl concentrators and vortex separators are compact
flow throttling and solids separation devices that provide
flow regulation and a rough level of solids and floatable
removal in combined flows. The technology  originally
was applied  in England in the 1960s, and since  has
evolved into a number of  configurations. Three of the
more   common  technologies  are  the  EPA  swirl
concentrator,  the  Fluidsep vortex separator, and  the
Storm  King hydrodynamic separator.  The differences
among the three separator types came about through
experimental  work aimed at creating flow conditions
within  the unit  that  would  optimize the  liquid/solid

Although  each  of the three types  of separators  is
configured differently, the operation of each unit and the
mechanism for solids  separation are similar. Flow
entering the unit is directed around the  perimeter of a
cylindrical shell, creating a swirling, vortex flow pattern.
The  swirling  action throttles the influent flow,  and
causes solids to be concentrated  at the bottom of the
                                    unit. The throttled underflow containing the concentrated
                                    solids passes out through a foul  sewer outlet in  the
                                    bottom of  the unit,  while  the clarified supernatant
                                    passes out through the top of the unit. The underflow is
                                    typically discharged to the downstream interceptor for
                                    treatment  at  the  POTW.  Various baffle arrangements
                                    capture floatables in the supernatant. The floatables are
                                    carried out in the underflow when the unit drains, once
                                    storm flows subside.

                                    The mechanism for solids separation  is created by the
                                    flow patterns  within the unit. Flow initially follows a path
                                    around the perimeter of the  unit.  After one revolution,
                                    the flow is deflected into an  inner swirl pattern, which
                                    has  a lower velocity  than  the  outer swirl.  Gravity
                                    separation  occurs as particles follow a  "long  path"
                                    through the outer and inner  swirl. The quiescent inner
                                    swirl, as well  as tangential breakaway of particles from
                                    the cyclonic flow field and drag forces along the walls,
                                    bottom, and  in the shear zone between the inner and
                                    outer  swirl,   all  contribute  to  solids  separation.
                                    Secondary currents direct particles across the floor of
                                    the unit towards the foul sewer outlet. Examples of each
                                    of the three  common types of swirl/vortex units  are
                                    presented in  Figures 4-14 through 4-16.

   Foul Sewer
A Inlet Ramp
B Flow Detector
C Scum Ring
D Overflow Weir and Weir Plate
E Spoilers
F Floatables Trap
G Foul Sewer Outlet
H Floor Gutters
I  Downshaft
J Secondary Overflow Weir
K Secondary Gutter
Figure 4-14. Example EPA swirl concentrator (U.S. EPA, 1982).
Performance of swirl/vortex devices depends primarily
upon hydraulic throughput and the settling characteristics
of the  solids  in the  combined  flow.  The  settling
characteristics are particularly important. The EPA swirl
concentrator is most effective at removing solids with
characteristics  similar to grit (0.2 mm diameter,  2.65
specific  gravity),  while  the particle  settling  velocity
profile of the flow to be treated is required to predict the
removal  efficiency of the  Fluidsep  and  Storm  King
devices.   Determining  the  particle  settling  velocity
distribution in samples of the actual flow to be treated
is strongly  recommended  in order  to  better predict
actual solids removal efficiencies.

Opinions differ as to the effect  of turbulence and free
vortex formation on performance. The  EPA swirl is
            designed with vertical baffles to "reduce the rotational
            energy of the liquid above the weir plate, and between
            the scum ring and the weir" (U.S. EPA, 1982). These
            baffles, along with a flow deflector at the inlet, disrupt
            the free vortex flow, with the intent of creating a  more
            "gentle" swirl.  Laboratory  studies  showed that the
            baffles improve  solids separation performance  (U.S.
            EPA, 1972). The Fluidsep, by comparison, features no
            vertical baffles or flow deflectors, with the specific intent
            of encouraging free vortex flow. The researchers who
            developed the Fluidsep unit contend that the free vortex
            creates less turbulence than the disrupted flow patterns
            in the EPA swirl. The differences  among  the  three
            common types of  swirl  devices are  outlined in the
            section called Design Details.

                      Effluent Launder
                                                                                  Flow Pattern
                                                                                     Solids -

Figure 4-1S.  Example Fluldsep vortex separator (Metcalf & Eddy, Inc., 1991 a).
Process Design
The design of swirl concentrators and vortex separators
is  based on  the scale-up  of  empirical  data from
experiments  on model systems.  Researchers used
models  to  develop  optimum  configurations  and
dimensions for a given set of conditions (flow, solids
settling velocity distribution).  Model  dimensions were
then scaled-up to match the conditions of the intended
application, using Froude's Law of Similarity.
The  methodology for sizing swirl concentrators varies
among the three types of devices. A detailed procedure
for  the  design of the EPA swirl  concentrator has
previously been published (U.S.  EPA, 1982), while the
sizing  computations  for the Fluidsep vortex separator
and the Storm King hydrodynamic separator are mostly
proprietary. A general description of the methodology for
sizing  these  units is provided below. The reader is
referred to the EPA swirl concentrator  design procedures
(U.S. EPA, 1982) and the respective manufacturers of
Fluidsep  and  Storm King for  more  specific sizing
EPA Swirl Concentrator

During the 1970s, EPA conducted a series of performance
studies on secondary-flow-motion wastewater control/
treatment devices, which were the  prototypes for the
standard EPA swirl concentrator. The purpose of these
studies was to develop the optimum configuration  of
swirl chamber elements and dimensions to achieve flow
regulation and maximum  solids separation from  a
typical CSO flow.
The solids settling velocity distribution in the typical flow
used  in the studies was determined by  sampling a
number of grit chambers. The  actual settling velocity
distribution used  in the  EPA design is presented in
Figures 4-17 and 4-18. Average  settling velocity profiles
from other sampling studies of  combined and sanitary
wastewater are  presented  in  Figure 4-19.  For  the
prototype  performance studies, synthetic  solids were
created to match the  "standard"  settling  velocity
distribution developed  from the grit chamber studies
(U.S. EPA, 1972,  1975c).

               Dip Plate and Spillway
                   Support Frame

                 Top Baffle
                   Baffle Ptate
                  Venturi Plate   -
Overflow Chamber
Overflow to Outfal
                               ._  Foul Outlet
                                    Pipe to Sanitary Sewer
                                                                                Tangential Intake Pipe

                                                                             - Intake Deflector Plate
                                                                    	1 — Sanitary Outlet Pipe Spigot
              Concrete Chamber Poured In Place or
              Precast Segmental Shaft Rings
              (eg. Charcon One Pass)
          — Concrete Base (Poured in Place)
                                    Inlet Manhole
            Combined Sewer
               Storm King
                                                                          Si' Stormwater
                                                                             Outlet Manhole
   Control Manhole \ \   Stormwater Sewer

                                                          Max. Permitted Discharge
              Reg-u-Flow    Dry Weather
              Vortex Valve   Flow Bypass
Figure 4-16.  Example Storm King hydrodynamlc separator (H.I.L. Technology, Inc. undated).

         400 300 140
                   U.S. Standard Sieve Numbers
                   70   40  30      10




   8  3-0

   1  2-0




   0.03 0.06  I 0.10
    0.04 0.08
0.40 IO.BOI
   0.60 1.0
                       Particle Diameter, mm
                                       1 in = 2.54 cm
Figure 4-17. Particle settling velocities for grit and organic
           material in still water (U.S. EPA, 1982).

As a result of the studies with the synthetic waste, solids
removal performance was correlated with flow and the
ratio of swirl chamber diameter to inlet diameter (Da/D,).
A series of curves were produced that related discharge
to the parameter D2/D^, for a range of values for inlet
diameter and percent settleable  solids removal. Given
the flow and the desired settleable solids removal, the
curve will yield the value of Dz/D-i corresponding to the
selected  inlet  diameter,  assuming  a  ratio  of  unit
diameter to weir height (Da/H) of 0.25. Additional curves
were developed for revising the D2/D^ ratio based on a
Dg/H ratio of other than 0.25, since in some cases the
available head may make H a controlling parameter. All
of these  design curves are presented in U.S. EPA,
1982.  The  remainder of the unit is dimensioned  in
proportion to the values of D2, D^ and H determined in
the initial  steps.

In practice, the EPA swirl concentrator is intended to be
primarily an in-line flow regulator, and is not intended to
remove the lighter solids that may be found in combined
flows. A suggested method for estimating the  actual
solids removal performance of a  swirl concentrator
involves comparison of the actual solids settling velocity
profile of the flow to be treated with the theoretical
settling velocity profile for grit developed by EPA. From
                                                        3  4
                                                                  a 10
                                                                         U.S. Standard Sieve Numbers

                                                                         16   20  30   40  50
                                                                                             70  100  140
                                                                            1     0.6

                                                                          Grain Size in mm
U.S. Sieve Size
mm in.
5.0 (0.020)
2.0 0.08)
0.84 (0.034)
0.42 (0.017)
0.30 (0.012)
0.20 (0.008)
% Finer by Weight
Grit Organics
100 100
100 53
63 31
31 17
18 14
0 10
                                                   Figure 4-18.  Typical gradation for grit and organic material
                                                              (U.S. EPA, 1982).
                                                   Figure 4-17, grit particles of 0.2  mm have a settling
                                                   velocity of approximately 2.5 cm/sec. A swirl concentrator
                                                   designed  in  accordance with EPA recommendations
                                                   (U.S. EPA, 1982) is intended to remove 90 percent of
                                                   grit-sized  particles.  Given an actual  settling  velocity
                                                   distribution  curve,  the  swirl might  be  expected  to
                                                   remove  90  percent  of the  particles  with  settling
                                                   velocities  equal to or greater than 2.5 cm/sec. On this
                                                   basis, an  overall estimated solids  removal rate can be
                                                   Two factors may  contribute  to  the appearance  of
                                                   lower-than-expected removal efficiencies for the EPA
                                                   swirl. Since the expected performance is  based on the
                                                   removal of particles in the size and weight range of grit,
                                                   much of the material the swirl is specifically intended to
                                                   remove would  be in the bed load carried  along the
                                                   bottom of the interceptor. If the swirl unit  is installed in
                                                   an off-line arrangement, where flow is diverted to the
                                                   swirl through a side weir or  high-outlet regulator, then
                                                   the bed  load would  most likely be carried down the

                      U.S. EPA Swirl Concentrator
                      Solids Basis of Design
                      Upper Curve = Inorganic Grit
                      Lower Curve = Organic Settleable Solids
                      S.G. Upper = 2.65
                      S.G. Lower = 1.2
                                                                                              Saginaw, Ml
                                 Boston, MA
                                 (heavy commercial)
    Philadelphia, PA
Boston, MA
                                                                     San Francisco, CA
                        10     20     30      40     50     60     70     80     90     100

                                                % of Particles with Settling
                                             Velocity Less Than Stated Value
Rgure 4-19.  Settling velocity profiles of combined and sanitary wastewaters (Pisano, 1990).

interceptor and would not be part of the influent to the
swirl.  Similarly, a swirl unit located downstream  of a
storage tank may not receive the heavier fraction  of
solids that it is capable of removing.

In addition, the sampling apparatus used to assess  swirl
performance may not take  a representative sample of
the influent  flow.  A single sample draw-off located  at
mid-depth in the influent pipe might not pick up heavy
solids passing to the unit in the bed load. The sampler
intake  velocity also must  be sufficient to  carry the
heavier grit  particles into the sampler. Since the bed
load solids may constitute  a significant fraction of the
total solids mass loading, failure to pick up the bed load
solids in the swirl influent, effluent, and underflow could
skew the overall performance results.

Fluidsep  Vortex  Separator

As with the EPA swirl, the standard configuration of the
Fluidsep vortex separator evolved through a series of
studies aimed at optimizing the solids removal efficiency
of the  unit. However, rather than basing the design on
a "typical" CSO solids settling characteristic, as EPA did,
a generalized design procedure was  developed.  This
procedure can be  used for any given solids settling
distribution,  stated performance level  for the final  swirl
design, and design discharge.  Therefore, the user is
required to perform studies of the existing CSO solids
settling characteristics before  the  Fluidsep  vortex
separator can  be designed.

The first step  in the design procedure for site-specific
applications is to determine the  actual solids settling
velocity profile  in the CSO to be treated. Froude's Law
of Similarity then is applied to the "real-world" solids
distribution to  obtain a model solids distribution curve.
Model tracer removal  curves then are applied to the
model solids distribution curve to develop a curve of
predicted  removal  efficiency  versus  flow  for given
vessel geometries (Pisano, 1990). A nominal D/H  ratio
of 2.5  initially is used for the design, although the final
ratio chosen depends on  considerations such as the
available  head,  site  area,  and  construction  costs.
Designs are available with  D/H ratios  varying from 0.5
to 3.0. This information then is  used by Fluidsep to
design the final swirl through a proprietary process.

Since  the design is  based on actual solids settling
characteristics  determined through  sampling   CSO
discharges,  removal  efficiency  predictions for  the
Fluidsep vortex separator should be more precise than
for  the EPA swirl. However, inflow solids stratification
still can pose problems in accurately characterizing the
influent, thereby making removal efficiency calculations
Storm King Hydrodynamic Separator

The design of the Storm King hydrodynamic separator
is based on the solids settling velocity profile of the CSO
to be treated, and the hydraulic loading or overflow rate.
Through studies with its unit, the manufacturer of the
Storm King has developed a series of optimal overflow
rates  for particular ranges of solids to be removed.
Thus, if the  influent flow rate and  the distribution of
solids are known, the manufacturer can provide  a  unit
of diameter determined from the optimum overflow rate
for the given size of material to be removed. As with the
Fluidsep vortex separator,  the specific  dimensional
details for the vessel, as well as the inlet, outlet, and
other features of the Storm King, are proprietary. The
Storm King separator differs from the Fluidsep primarily
in the location of the solids outlet and the introduction
of internal components designed to stabilize secondary

General Hydraulic Considerations

In addition to the specific sizing procedures discussed
above,  the configuration of a swirl  unit,  or even  the
feasibility of  installing a swirl unit,  may depend upon
system  hydraulics. One  of the main considerations is to
determine to  what  level the influent  sewer  can  be
surcharged  without causing  upstream  flooding. This
elevation, in conjunction with the elevation head on the
effluent weir,  will  set  the  elevation of the unit with
respect to the influent sewer. For example, Figure 4-20
indicates an  empirically derived curve  of head versus
discharge per linear foot of weir length for a circular weir
(U.S. EPA, 1972). For a maximum flow of 300 cfs  and
a circular weir length of  60 feet, the discharge would be
5 cfs/ft, corresponding  to a head of approximately 3
feet. If entrance losses  were  neglected, the weir crest
then would be set 3 feet below the maximum hydraulic
grade line allowed in the influent sewer.

Since overloading the swirl units decreases performance,
the units are provided with some manner of overflow
weir to relieve  peak flows. The  EPA swirl design
recommends that flows to  the unit  be  limited  to twice
the design flow (U.S. EPA, 1982), and both the Fluidsep
and Storm King swirls are designed for the peak flow.
The crest elevation of the side overflow relief weir can
then be determined from the head on the weir  at the
maximum flow using Figure 4-20.

If the available head is minimal, the geometry of the unit
may be modified, in certain  instances, to reduce the
required head. Pumping the foul sewer discharge also
has been employed at  installations where the system
hydraulics prevent  gravity flow of the   foul  sewer
discharge to the interceptor.


                       r 90
                     0  CFS
28           56            84           112
     Discharge per Linear Foot (30.5 cm)
Figure 4-20. Head vs. discharge per linear foot of weir length for a circular weir (U.S. EPA, 1982).
Process Flow
Four of the common layouts for swirl concentrator and
vortex separator  installations  are  illustrated in Figure
4-21.  Selection of the best arrangement for a particular
application may  depend on  the overall system-wide
CSO control strategy, treatment goals, cost, and/or site

Off-Line, Stand Alone

With  this arrangement  (Figure  4-21 a),  a regulator
diverts flows in excess of the interceptor capacity to the
swirl/vortex unit. Diversion weirs and vortex valves are
examples of  regulators that have been  used for this
purpose. The supernatant or effluent is discharged from
the separator to the receiving water while the concentrated
foul sewer underflow is returned to the interceptor. An
upstream relief overflow, or high level weir in the swirl
unit typically  is  provided  to  protect the  unit during
extreme storm flows.

Off-Line, With Storage Tank

This arrangement (Figure 4-21 b) is similar to the off-line
stand alone arrangement except that the effluent from
the swirl/vortex unit discharges to a storage tank rather
than directly to the receiving water. The  foul sewer
underflow similarly is returned to the interceptor. The
potential benefit of locating a swirl/vortex unit in series
upstream of a storage tank is that the unit can capture
much of the grit and heavy solids that otherwise would
settle out in  the  storage tank. In  addition to making
cleanup easier, removal of  a fraction of the solids load
may  reduce  the  required  hypochlorite  dosage  if
disinfection is provided.
                    In-Line, Stand Alone or With Storage Tank
                    With this arrangement (Figures 4-21c,d), the swirl/vortex
                    unit acts as a regulator, throttling flow to allow only the
                    underflow  to  continue  down the  interceptor.  The
                    overflow from  an in-line unit passes either to  the
                    receiving water or to a storage tank. The unit is sized
                    such that peak dry weather flows pass to the interceptor
                    without being throttled. This arrangement  may be the
                    most appropriate for the EPA swirl concentrator, which
                    is  intended to  be primarily a  flow-regulating  device.
                    Locating the swirl device in an in-line arrangement also
                    will provide the best opportunity for the unit to remove
                    the heavier fraction of solids in the flow. One drawback
                    to an  in-line  arrangement  is the wear  and  tear
                    associated  with  constant exposure to flow.  Should
                    maintenance or repairs be required, provisions must be
                    made to handle the dry weather flow until the repairs
                    are completed.

                    Other Unit Processes

                    Many swirl/vortex facilities are designed with upstream
                    mechanically cleaned bar racks. The bar racks protect
                    the swirl unit from being clogged or otherwise damaged
                    by large objects carried in the  flow, and may improve
                    floatables   removal.  For  installations   where   the
                    underflow is pumped, the bar racks also help protect the
                    downstream pumping equipment. Swirl units have been
                    installed in England without upstream bar racks. At the
                    Alma Road swirl unit in Bristol, England, no blockages
                    of the foul sewer outlet were reported during the first 3
                    years  of  operation  (Smisson,  1968).  In-line   swirl
                    concentrator installations that do not require pumping of
                    the concentrated underflow may not require upstream

                    OFF - LINE
                                    ON - LINE
            Emergency Overflow
               Swirl/Vortex Unit
                                                         Tank Overflow
Figure 4-21.  Common layouts for swirl/vortex Installations (Pisano, 1990).
bar racks, as flow regulators  are not provided  with
upstream coarse screening.

Performance Considerations
As  noted above,  the performance of swirl units  is
influenced primarily  by solids  settling characteristics
and flow rate, and the apparent performance also  may
be affected by sampling techniques. Performance  data
for swirl units is  relatively scarce, due in part to the
relatively  small number of installations in the  United
States, and also to the difficulty in obtaining data  from
existing units. Performance evaluations and monitoring
of CSO control facilities present a number of logistical
problems due  to the  intermittent nature of  facility
operation. Sampling and analysis crews must be  able
to mobilize on short notice,  at  odd  hours, and for
uncertain durations in order to obtain the types of  data
necessary to  evaluate   performance.   Even  with
automatic sampling devices, supervision is required to
ensure that the sampling runs smoothly. In addition, as
noted above, sampling systems must be designed and
installed so as to obtain a true representative sample of
the solids distribution  in  the influent flow. Samples that
do not include  an appropriate fraction of the bedload
may result in underestimation of the influent solids load,
               which may in turn result in underestimation of the solids
               removal performance.

               The definition of performance for swirl  units  is  not
               uniform. Researchers have distinguished between  the
               removals  attributable  to solids separation,  and  the
               removals  attributable to the function of the unit as a
               regulator. In a performance evaluation of the Washington,
               D.C., Northeast Boundary (NEB) swirl facility (O'Brien
               & Gere, Inc., 1992), the overall removal efficiency was
               determined from the ratio of influent to effluent mass
               loadings. The removal efficiency due to flow diversion
               was estimated by assuming that the foul sewer solids
               concentration equalled the influent solids concentration.
               The theoretical effluent load due to flow diversion then
               was obtained by  subtracting the assumed foul sewer
               load from the influent load. The theoretical removal rate
               due to flow diversion was then calculated:
                 removal =
(influent load) - (theoretical effluent load)
             influent load
               The removal efficiency due to solids separation then
               was estimated by the difference between the actual
               overall removal efficiency and the theoretical removal
               rate due to flow diversion.

Evaluation  of the  performance  of  a Storm  King
hydrodynamic separator in Walsall, England (Hedges,
1992), identified a treatment factor e, such that
          e =
(foul sewer load/influent load)
(foul sewer flow/influent flow)
If a device provided no treatment, but simply divided the
flow, then e =  1. If some degree of separation and
concentration occurred, then e would be greater than 1.

Design Details

The interior features of swirl units vary among the three
general types: EPA swirl concentrators, Fluidsep vortex
separators, and  Storm King hydrodynamic separators.
In addition, since the technology has been evolving,
variation exists in the design details among currently
operating  swirl devices of the same general type. For
the EPA swirl concentrator, the dimensions and orientation
of troughs, baffles, and other details are developed from
the basic sizing  dimensions  "D1" and  "D2"  (inlet
diameter  and vessel  diameter)  in  accordance with
design procedures (U.S.  EPA, 1982). For the Fluidsep
and Storm King devices, sizing criteria for weirs, baffles,
and other details are not available in  the literature.
Table  4-9 compares the key  features of the three
common  types of  separator   units (refer  also  to
Figures 4-14 through 4-16).

Chamber Construction

Concrete  is the  most common material of construction
for the swirl chambers,  although  stainless steel and
epoxy-coated  carbon  steel  also have been used.
Trowelled finishes  and/or epoxy wall  coatings have
been  suggested to minimize clinging of material to the
chamber wall (U.S. EPA, 1984). The performance study
of the Washington, D.C., NEB swirl facility recommended
providing  smoother  coatings for the floors and walls,
and rounding off the edges of the floor gutters to reduce
turbulence. This study also recommended increasing
the floor slope to minimize the deposit of solids on the
floor  (O'Brien   &  Gere, Inc.,  1992).  Examples  of
chamber construction are presented in Table 4-10.

Interior Baffle/Cone Construction

Interior baffles, flow deflectors, and similar components
typically are constructed  of carbon or  stainless steel.
Corrosion in the Lancaster, Pennsylvania, EPA swirl,
prompted a  recommendation  to  use  stainless steel
(U.S.  EPA,  1984).  A  performance   study  of the
Washington, D.C.,  NEB swirl  facility presented the
following recommendations regarding interior hardware
for EPA swirls (O'Brien & Gere, Inc., 1992):

• Extend the floatables weir so that the weir skirt is
   directly over  the  entrance flow deflector baffle (as
   indicated in U.S. EPA, 1982). This adjustment creates
  a  more  continuous  "inner  wall"  of  water,  and
  therefore, a more quiescent inner settling zone.

• Locate weir plate hangers on the  inside face of the
  weir to reduce turbulence.

• Provide a continuous, smooth overflow weir plate and
  skirt to reduce turbulence.

• For multiple swirl units, set weirs to provide an even
  flow split at the most  typical flow  rates, rather  than
  the peak design flow.

When  a Storm King  hydrodynamic  separator  was
installed  for  the  Walsall  Metropolitan  Borough  in
England, the interior base cone was constructed of light
gauge stainless steel and served as  a  permanent form
for concrete. The vessel top, including dip plate, top
baffle plate, outlet chamber and channel, and floatables
trap, was fabricated in sections, then bolted together at
the site.  A concrete platform ring cast into the vessel
body supported  the top assembly (H.I.L. Technology,
Inc., undated).

Piping/Flow Control

The inlet pipe to a swirl unit should be laid at as shallow
a slope as possible to minimize turbulence, yet maintain
self-cleaning velocities.  EPA recommends sizing the
foul  outlet for three-times  DWF,  and  notes that it is
difficult to size the underflow pipe to act as a throttle.
The use of a sluice gate, motor-operated gate, or vortex
valve on the foul sewer discharge is recommended to
control the underflow  (U.S. EPA,  1982). An underflow
pipe sized  to throttle  flow is of smaller diameter and
more prone to clogging. The level of control required for
the underflow is governed by local hydraulic conditions
and downstream interceptor capacity.

At the Lancaster,  Pennsylvania,  and West Roxbury,
Massachusetts,  EPA  swirls,  the  foul  outlets  are
controlled by vortex valves,  located away from the swirl
chamber. Sediments observed in the lower-velocity flow
upstream of the vortex valves prompted a suggestion
that the vortex valves be cast into the base of the swirl
unit (U.S. EPA, 1984).

At the Decatur, Illinois, McKinley Avenue Swirl Facility
(BGMA  and CMT,  1977), a  vortex valve  on the
interceptor acts as a regulator,  allowing dry  weather
flows to pass uninhibited,  but throttling flows  greater
than the interceptor design capacity.  The throttled flows
back up and pass  over a weir to the McKinley Avenue
facility. During startup, adjustments to the overflow weir
were required to develop the design  head on the vortex
valve. Once the  proper head  was  developed, the
discharge from  the vortex valve followed its design
curve fairly well.  During  an  11-month  performance
study, the vortex valve did not become plugged.

Table 4-9.  Comparison of Design Details for Swirl/Vortex Devices

                   EPA Swirl Concentrator               Fluldsep Vortex Separator
                                                                          Storm King Hydrodynamic Separator
Flat, or less than 2% slope;  curved
gutters to direct solids and DWF to
foul sewer outlet
Foul sewer outlet   Asymmetric; orifice offset from center
                   of rotation of swirl
Influent entrance    Located at floor of unit; features flow
                   deflector, to cause first revolution flow
                   to be deflected inwards, minimizing
                   turbulence and creating interior
                   rotating water mass

Treated effluent    Past vertical, cylindrical scum baffle,
exit                then over vertical, cylindrical weir onto
                   horizontal clear water plate, leading to
                   central drop shaft; shaft passes down
                   through center of unit, discharging
                   below bottom of vessel

Baffles and flow    Series of vertical baffles at surface,
directing features   perpendicular to rotational flow,
                   dampens  free vortex formation -
                   influent flow deflector, effluent weir
                   scum baffle

Roatables removal Vertical flow deflector extends across
                   top of unit, directing floating material
                   to a channel;  channel crosses weir
                   plate to vertical vortex cylinder at wall
                   of overflow downshaft; vortex draws
                   material beneath weir plate, where it
                   Is dispersed and  contained under weir
                   plate and inside weir skirt; when
                   storm flow subsides, ftoatables are
                   drawn down and exit through the foul
Underflow control
Difficult to size underflow pipe to act
as throttle; suggest use of gate or
vortex valve to control underflow
        Floor sloped 4-6%;
        smooth; no gutters
                                      Conical outlet in exact alignment
                                      with center of rotation
                                      Located at floor; no flow deflector
                                      Flow passes up through annular
                                      opening between vertical, cylindrical
                                      scum baffle, and cone shaped
                                      baffle; flow passes over roof of
                                      vessel to outlet
                                      No baffles to dampen vortex; Inner,
                                      cone-shaped baffle deflects
                                      downward lighter particles carried
                                      up by secondary currents;
                                      cylindrical scum baffle.

                                      Vertical, cylindrical floatables skirt
                                      creates annulus with cone baffle;
                                      floatables retained behind skirt, and
                                      removed out foul sewer as flows
        Vortex action at underflow draw off
        throttles underflow, or vortex valve
        can be used
Roor sloped; no radial gutters
                                            Annular trough along base of interior
                                            cone, halfway between wall and
                                            center of unit

                                            Located at mid-depth of unit; no flow
                                            Flow passes up through annulus
                                            between horizontal, circular baffle
                                            plate, and cylindrical, vertical dip
                                            plate; flow passes over top of circular
                                            baffle to outlet
                                            Horizontal, circular top baffle;
                                            cylindrical dip plate located at shear
                                            zone; interior cone along centroidal
                                            rotational axis
                                            Floatables captured behind dip plate;
                                            manually removed, or passed to
                                            WWTP when water level subsides
Use vortex valve to control underflow
Table 4-10.  Examples of Swirl/Vortex Chamber Construction

Location                Swirl Type          Diameter (ft)    Chamber Construction
Walsall Metropolitan

Clevedon, Avon
Lancaster, PA
West Roxbury, MA
Decatur, IL
(McKinley Ave.)
Washington, DC
Presque Isle, ME
Burlington, VT
Decatur, IL
(7th Ward)
EPA Swirl
EPA Swirl
EPA Swirl
EPA Swirl
EPA Swirl
Fluidsep Vortex
Fluidsep Vortex
Carbon steel with epoxy coating
Stainless steel plate inside concrete shell
Concrete (proposed)
U.S. EPA, 1984
U.S. EPA, 1984
BGMA and CMT, 1977
O'Brien & Gere, Inc., 1992
WPCF, 1989
Ganley and McCarthur,
Wilcoxon and Hunsinger,
      Storm King

      Storm King
17.3        Concrete w/stainless steel base cone

19.7        Concrete
            H.I.L. Technology, Inc.,

            H.I.L. Technology, Inc.,

From the regulator, flow to the facility passes to a screen
house. A 54-in inlet pipe from the mechanically cleaned
bar screens to the swirl unit was installed at a "minimal"
slope.  During  operation, solids  were deposited in the
bar rack channels and the inlet pipe. During  cleanup
following a storm, these  solids had to be manually
flushed down through the swirl unit. Increasing the slope
of the inlet pipe would minimize solids deposition in the
pipe, and facilitate cleanup of the screening chamber.

At the Washington, DC, NEB swirl  facility (O'Brien &
Gere, Inc., 1992), inflatable dams serve as regulators,
diverting dry and wet weather flows to the facility. In wet
weather, pressure in the dams is increased to divert
more flow to  the facility  and  to optimize  in-system
storage. The  dams  also  can  be deflated  to relieve
upstream flooding and/or excessive  flow to the facility.
From the  dams, flow passes  through mechanically
cleaned bar racks, then to the  three 57-foot diameter
EPA swirl  concentrators.

During operation of the NEB  swirls,  air periodically
bubbled up through the effluent downshaft. The release
of this entrained air could possibly reduce the  capacity
in the downshaft and cause premature surcharging of
the overflow weir. Vents consisting of four vertical pipes
were installed in the downshaft of one of the units, and
were effective in  eliminating this problem. The  foul
sewer underflow was pumped to the interceptor using
torque-flow pumps.

For the Storm  King hydrodynamic separator installed for
the  Walsall Metropolitan  Borough,  England,  the  foul
sewer underflow is controlled by a vortex valve (H.I.L.
Technology, Inc., undated).
The Fluidsep  vortex separator system  in  Saginaw,
Michigan,  captures  "first flush"  flows  by  utilizing
in-system storage and then filling two  retention tanks
having vortex throttles to  permit continuous drainage.
When the "first flush" storage capacity is filled, pumped
discharge  to  vortex  solids separators commences.
Pumped underflow from the separators is detained by
a third tank, while treated overflow enters a fourth tank.
The continuous drainage from the two retention tanks
permits optimal use of  the separators and treatment
tanks, particularly for long intermittent storms (Pisano,

Spray Wash Systems

Automatic spray wash  systems reduce the manual
cleanup effort following activation of a swirl unit. The
potential benefit of an automatic wash system is greater
for multiple-unit installations. EPA  suggests  installing
two  spray headers in each swirl unit: one under the
horizontal circular weir plate, and one along the wall of
the  chamber,  above the  maximum water  level.  The
hydraulic pressure on the underside of the circular weir
plate may cause material to stick (U.S. EPA, 1982).

Operating experience has yielded  mixed results with
respect  to  spray  wash  systems.  The Lancaster,
Pennsylvania, EPA swirl was designed with two spray
headers, in  accordance with  EPA recommendations
(U.S. EPA, 1982). During a performance study, material
hung up on the spray header, and a manual wash down
system using hoses was suggested (U.S. EPA, 1984).
The West Roxbury EPA swirl had no spray wash system
or problems  regarding  clinging  material (U.S. EPA,
1984). The Decatur McKinley Avenue EPA swirl does
not have a spray  header  system. Washdown  is
accomplished  manually using  hoses from  adjacent
yard hydrants. A performance study  noted that post-
operation cleanup was time consuming, and suggested
installing an automatic spray  system in  the  swirl unit
(BGMA and CMT,  1977).  The EPA swirl units at the
Washington, DC, NEB facility were provided with spray
headers  around the walls of the  chambers. The spray
systems  worked well flushing material from the walls,
but were ineffective at moving material across the floor
of the  units and into the gutters (O'Brien & Gere, Inc.,
1992). The Fluidsep vortex separators installed at the
Decatur  Lincoln  Park and 7th Ward  Facilities  do not
feature automatic  spray wash systems. As with the
McKinley Avenue facility, washdown  of these units is
accomplished using hoses from adjacent yard hydrants
(Drake and Hunsinger, 1990).

Structural Covering

A roof over a swirl unit is not required  for functional
purposes,  but  can  be  beneficial  for safety and
aesthetics purposes. If a roof is provided, a manhole
should be provided over the floatables vortex cylinder
to allow  rodding of the floatables trap and foul sewer
outlet  (U.S.  EPA,  1982). Walk-in  access  is  most
convenient and safe for removal of  large  floatables. An
inspection  walkway  should  be  provided around the
perimeter of the vessel to allow access to the weir and
scum plate.

A domed cover,  grating,  and open  tank were  each
considered in the design of the McKinley Avenue EPA
swirl.  The advantage of a  dome  or  grate  is that
personnel, especially intruders, are protected  from
falling in. The disadvantage of a dome (or other type of
solid roof) is that it creates a confined space, requiring
ventilation, lighting, and appropriate safety procedures
before entering the unit for maintenance. An open-grate
covering eliminates these concerns, but also makes
maintenance more cumbersome by restricting access.
In the end, the McKinley Avenue swirl was constructed
without a roof or grating. These types of considerations,
including safety, maintenance access,  and cost of

appurtenances such as lighting and ventilation, should
be evaluated on a site-specific basis.

System Controls and Operation

One of the advantages of swirl/vortex units is that they
have no  moving parts, and thus operation of the  unit
itself is governed solely by hydraulics. However, upstream
regulators, mechanically cleaned bar screens, upstream
or  downstream  pumping  systems,  and  disinfection
systems  that may  be associated  with the swirl/vortex
units  typically  are provided with some degree of
automatic control  to allow unattended  startup  and
operation. Float switches or bubble tubes detect storm
flows  through  rising water levels.  These  devices
typically activate the mechanically cleaned bar screens
and the disinfection system, if provided.

Some  dewatering  and washdown  operations  are
designed for automatic control, while others are designed
for manual operation. As an example, automatic spray
wash systems could  be activated on dropping water
levels  in the swirl  unit.  Refer  to  the  sections on
In-System Controls/ln-System Storage, Coarse Screening,
and Disinfection for additional  details on the control
strategies for these operations.

Process Variations

EPA recommends  procedures for  modifying the basic
swirl design to serve as degritter, primary separator (for
sewage), and for treatment of erosion runoff (U.S. EPA,
1982).  In Rochester,  New York,  a pilot  program
evaluated the use  of a degritter and primary separator
for treating combined sewage. Flow passed through the
degritter first, then the primary separator. Tests were run
without chemical addition, with polymer,  with polymer
and alum, and  with  polymer, alum, and phosphorus
spikes. Chemical addition did not enhance performance.
The authors believe that the mode of chemical addition
was  responsible   for  the  lack of  improvement.
Inadequate velocity gradients and/or contact time due
to  in-line mixing of chemicals may not have provided
efficient floe development (U.S. EPA, 1979a).

Other studies demonstrate the ability of swirl degritters
designed in accordance with EPA procedures to meet
the projected removal rates for grit-sized particles (U.S.
EPA  1974b, 1977c,  1981).  These  devices may be
effective as off-line devices at treating swirl concentrator
underflows or CSOs  with  particularly high fractions of
grit-sized particles. The major difference  between a
swirl  concentrator  and a  swirl  degritter is that the
degritter features a conical bottom hopper for collecting
the settled grit and  delivering  it to a  separate grit
washing system. The swirl degritter has no capability to
regulate flow, as the bottom outlet is solely intended for
the removal of the grit.

Disinfection  of overflows is  a common goal of CSO
control strategies. Since liquid sodium hypochlorite is
the most  commonly selected means for disinfecting
CSOs, this  section focuses  on the design  of  liquid
sodium hypochlorite systems. Alternative technologies
such as ultraviolet radiation, ozone, gaseous chlorine,
and chlorine dioxide are identified  in Chapter 2 and
addressed in the literature (White, 1992; WPCF,  1986;
U.S. EPA, 1986; Chlorine Institute), but are not presented
in detail in this manual. Dechlorination is discussed at
the end of this section.

Process Theory

Disinfection  effectiveness is measured in terms  of
reduction in bacterial concentration. The overall chlorine
dose required to achieve a given bacterial kill is the sum
of the chlorine demand and the chlorine residual. Since
a number of substances found in CSOs can exert a
chlorine demand in  addition to the demand created by
bacteria, the actual chlorine demand of the combined
wastewater to be treated is best assessed using actual
laboratory studies or pilot tests.

The expected change in bacterial concentration as a
function of chlorine residual and  contact time can be
estimated using the Collins model, which was originally
developed from pilot studies on primary effluent (White,

               y, = y0 (1 + 0.23 Ct)'3

  yt = bacterial concentration after time t (MPN/100 ml)
 y0 = original bacterial concentration (MPN/100 ml)
  C = chlorine residual concentration after time t (mg/l)
   t = contact time (min)

This  equation, however, does not accurately reflect
bacterial  kills at low  values  of Ct.  A modified model
proposed  the  following  relationships,   which  are
illustrated graphically  in  Figure 4-22 (White, 1992):

  y,/y0 = 1.0 for Ct b
    b = X intercept when yt/y0 = 1
    n = slope of regression  line
This model requires laboratory or pilot data to define the
relationship between y/y0 and Ct. The values of b and
n in the equation are obtained from a regression line
through the data points.

The  relationships  described by  the  Collins  model
require thorough mixing and  ideal plug flow during the
contact time. Since the disinfection capability of chlorine
species  depends on  physical  contact between  the
chlorine-containing  molecules  and the bacteria,  the

   Log 1=0
              -Lag Time-*1
                                     -n = Slope
                    Arithmetic Plot of Y7VO = —

Rgure 4-22.  Graphical representations of log Yt/Y0 versus
           log CT (White, 1992).
effectiveness of the dispersion of the chlorine solution
into the flow can have a significant impact on the overall
disinfection efficiency.
Adequate mixing is important to ensure dispersion of
the chlorine solution in the flow. The intensity of mixing
can be defined by the velocity gradient, G, as follows
(White, 1992):
                    G =
 G = mean velocity gradient (sec"1)
 P = power requirement (ft x Ib/sec)
 H = absolute fluid viscosity (Ib x sec/ft2)
 V = mixing chamber volume (ft3)

For mixing by diffuser, the power requirement, P, can be
expressed as:

               P = (Q x h)/(3,960 x erf)

  Q = flow through diffuser (gpm)
   h = headless in diffuser (ft)
  eff = efficiency (assume  = 1 .0)

An  alternative  method for computing  the velocity
gradient in open channels (U.S. EPA, 1973a) is:
 G = velocity gradient (sec"1)
  v = velocity (ft/sec)
  s = slope of channel  (ft/ft)
  u. = absolute viscosity (cp)
The parameter "GT, equal to the product of the velocity
gradient and the contact time while under the influence
of mixing, has been identified in pilot studies as a key
to disinfection efficiency at low contact times.  It has
been  suggested that providing turbulence during the
contact time improves the level of bacterial kill  (U.S.
EPA,  1973a, 1979b).

One of the challenges to disinfecting CSOs is the limited
contact time typically available at CSO control facilities.
The  limited  detention  time  can  be  particularly
problematic due to the  higher solids concentrations in
CSOs compared with secondary POTW effluent. The
solids can shield  the  bacteria from  exposure to the
disinfecting agent,  reducing the  efficiency  of the
disinfection  process.   Separate   plug-flow  contact
chambers generally are not provided at CSO facilities
due to space and/or cost constraints. Contact time for
disinfection commonly is limited to the detention time in
the tanks (for a storage/sedimentation facility) and/or in
the outfall pipe. A number of studies published in the
1970s investigated means for achieving high bacterial
kills at lower contact  times,  using  increased mixing
intensity,  increased disinfectant  dosage,  chemicals
having a  higher  oxidation rate  than  chlorine,  or  a
combination  thereof (Crane Co., 1970; Geisser and
Garver,  1977;  Tift et  al., 1977;  U.S.  EPA, 1973a,b,
1974c,  1975d,  1976,  1979a,b).  These methods are
generally referred  to as "high-rate disinfection."

There is  no clear  definition as to what constitutes a
"high-rate" disinfectant dosage or a "high-rate" value for
GT, other than those values for which acceptable
reduction in bacteria concentrations are achieved at
detention times less than the conventional values of 15
to 30 minutes.  For example, one study on high-rate
disinfection produced the relationships between fecal
coliform  reduction, GT, and chlorine dosage presented
in Figure 4-23 (U.S. EPA, 1979b). AGT of 10,000 with
a detention time of 1  minute would  be considered a
high-rate  application.   Under these  conditions,  as
indicated in  Figure 4-23,  99.99  percent (4-log) fecal
coliform  reductions are achieved with a chlorine dosage
of approximately 5 mg/l. If the velocity gradient G were
reduced, a corresponding increase in chlorine dosage
is required to achieve the same level of bacterial kill at
the  same  1-minute detention  time.  Another  study
reported  approximately 99.99 percent total coliform
reduction with a 10 mg/l chlorine dosage and a GT of
approximately  5,000 (2-minute detention), with  G =
40 sec"1  (U.S. EPA, 1973a).

When using bench-scale  or pilot studies to assess the
chlorine  demand of the flow to be treated, consideration
should  be  given  to  the  principles   of  high  rate
disinfection,  particularly with regard to  the anticipated
available contact  time, and the effects of mixing and,
turbulence of projected required dosages.

                T=1 Win

                SS = 200 mg/l

                FC = 500,000 counts/100ml

                CI2 Dose, as shown, mg/l

                pH = 6.8
                      CI2 Dose
                       NOTE: All Syracuse Data Included
Figure 4-23.  Relationship between GT and bacterial kill (U.S. EPA, 1979b).
Process Design

Sizing  of disinfection  equipment is based on  the
required dose rate and expected flow. In some states,
the permitting agency may specify a maximum required
dosing capacity, such as 25 mg/l at peak flow. In other
states, the permitting agency may specify the maximum
allowable  bacteria  concentration,   such  as   1,000
MPN/100  ml fecal coliform. Where a maximum coliform
bacteria discharge limit is provided, some knowledge of
influent flow  bacteria concentration and chlorine
demand is required to estimate the  required dosage.
Given an average  influent bacteria concentration, y0, a
value for the required effluent bacteria concentration, yt,
and a specified minimum contact time (15 min at peak
flow  is typical [Metcalf &  Eddy 1991 a]),  the required
residual can be estimated from the Collins model. For
example,  typical primary effluent might have an initial
bacteria concentration y0 = 38x106 MPN/100 ml (White,
1992).  If the discharge requirement is 1,000 MPN/100
ml, then:

          1,000 = 38x106 x (1 + 0.23 Ct)'3

  Ct= 142
For t = 15 min, C = 9.47 mg/l
Allowing  for an  immediate (3-5  minutes)  demand of
8 mg/l, and a die away  demand during contact time
(25+ minutes) of 1 mg/l, then the estimated dosage for
the above example would be about 19 mg/l.

As  noted above, final sizing of the disinfectant dosing
system should  be based on pilot studies the actual
combined flow to be treated. If a storage/sedimentation
tank is to provide contact time, actual required dosages
may be higher  than  predicted by models, due to the
concentration of solids in the influent combined flow.
Pilot studies also should  allow  evaluation  of other
high-rate disinfection  techniques, such as increasing
mixing intensity and  using chemicals with  a higher
oxidation rate than hypochlorite. Equations such as the
Collins model estimate the relationship between dosage
and contact time for a  given  reduction  in  bacterial
concentration,   and  therefore  provide  a  basis for
designing pilot tests for assessing  actual  dosages
Once the required maximum dosage is established, the
metering pump capacity can be computed based on the
estimated strength of the  hypochlorite solution. The
strength of hypochlorite solution is presented as a "trade
percent," which  is essentially a  percent  by volume,
computed as:
           trade percent = -
                                                                                 available CI2

The weight percent of available chlorine is  the trade
percent divided by the specific gravity of the solution:
  weight percent available chlorine =
                          trade percent
                         specific gravity
The specific gravity of a hypochlorite solution of a given
strength  is  not  constant; rather, it  depends on the
amount of excess sodium hydroxide the manufacturer
uses to promote stability. For example, to compute the
required metering pump capacity for a CSO facility with
a peak flow of 40 mgd, using 10 percent (trade) sodium
hypochlorite (specific gravity = 1.14) and a 15 mg/l dose
rate, let:
Qp = metering pump capacity (gph)
Cp = concentration of stock sodium hypochlorite
 Q( = peak effluent flow (mgd)
 Cf = required sodium hypochlorite dose rate
Qp x
                 = Q, x Cf x 41 ,667 gph/mgd
              Cp = % available CI2 by weight =
                   = 8.8%                     -
              Cp = 8.8% x (10,000 mg/l)/1%
                   = 88,000 mg/l
 88,000 mg/l x Qp = 15 mg/l x 40 mgd
                   x 4 1,667 gph/mgd
              Qp = 284 gph

Process Flow

At  CSO  storage/treatment  facilities,  liquid  sodium
hypochlorite usually is introduced at the upstream end
of the facility to maximize the contact time in the tanks.
Some facilities also provide a second dosing location
downstream of the  storage  tanks to ensure that a
desired residual is achieved. CSO control facilities may
feature  upstream relief overflows to protect the facility
from flooding  during  extreme events. These  remote
overflows  also  may  be   suitable  locations   for
hypochlorite addition. They  can  be served by piping
from chemical storage and feed equipment located at
the  main  CSO  facility,  and controlled  by  remote
telemetry. Table 4-11 presents examples of hypochlorite
systems currently in  use at  CSO  control  facilities. A
schematic of a typical hypochlorite system is presented
in Figure 4-24.

In  addition to effluent  disinfection,  other  uses  of
hypochlorite at CSO facilities may include:

• Odor control scrubbers

• Sludge pipe disinfection

• Spraywash system
These applications are discussed earlier in this chapter.
Design Details

Components of typical hypochlorite systems include:
• Storage tank

• Metering pumps

• Dilution water supply

• Piping and valves

• Diffuser

• Chlorine residual analyzer

These components are discussed below.  Additional
information is available in  the literature (White, 1992;
WPCF, 1986; U.S. EPA, 1986).

Storage Tank

Sodium hypochlorite is delivered to a CSO facility in
liquid form by truck, while smaller quantities can  be
delivered in 55-gallon drums.  Sizing considerations for
bulk storage tanks include projected usage of chemical,
including auxiliary uses such as odor control, delivered
strength, and  rate of decomposition of hypochlorite.
Hypochlorite is delivered with solution strengths ranging
from 10 to 15 percent. Higher strength solutions require
less storage volume, but  deteriorate more rapidly. A
10 percent solution is most economical (White, 1992).
Chemical  storage tanks,  whether above  or below
ground, must meet the applicable  federal,  state, and
local regulations and fire codes.

Metering  Pumps

Because of the relatively low metering pump capacities
required for hypochlorite dosing, positive-displacement
diaphragm pumps commonly are used for this purpose.
Pumps can be controlled by variable speed drives,  by
stroke length  positioners, or  both.  Pacing  of  the
metering pumps  in proportion  to  flow is  a common
control strategy. Dosing locations at the influent end of
the facility should be paced by the influent flow, while
dosing locations at the effluent end of the facility should
be paced  by the effluent flow (White, 1992). The need
for  additional  control  of  dosage  based on  effluent
residual  may  depend  on  permit  requirements and
potential water quality impacts of a  high residual.

Dilution Water

Since sodium hypochlorite solution  dose rates are low
enough to be expressed in terms of gallons per hour,
dilution water is provided as a carrier fluid to ensure a
reasonable flow velocity.  Hypochlorite dilution water
typically is controlled by a solenoid valve and rotameter.
The valve opens when the  hypochlorite pumps  are
activated. The rotameter is set to provide sufficient flow
to maintain a minimum 2 ft/sec velocity in the piping to

Table 4-11.  Examples of Sodium Hypochlorlte Disinfection Systems at CSO Control Facilities

Facility Location     Facility Type        Dosing Location(s)  Startup Control     Dose Rate Control     Other Details
Newport, Rl
Washington Street
Storage/treatment Influent channel;
tank effluent
collection box
Influent channel
mercury float switch
Pump speed paced by
either influent or
effluent flow; manual
stroke adjustment
Manual dose control
for remote upstream
Newport, Rl
Wellington Avenue

Saginaw, Ml
Hancock Street
Oakport, CA
Prison Point



Cottage Farm
Effluent forcemain
Upstream of tanks
Influent lift station
wetwell, near pump

Influent channel
upstream of tanks

Each of four storm
pump suctions,
downstream of
Storage/treatment   Influent channel
MWRA             Coarse screening/
Somerville Marginal   disinfection
Pretreatment Facility
MWRA             Coarse screening/
Constitution Beach    disinfection
Fox Point
Coarse screening/
                  Upstream of
                  influent sluice
                  gates in inlet
                  Flow channel
                  through facility
Flow channel
through facility;
dewatering pump
Lead stormwater
pump activation

Influent pump
Opening of tank
inlet gate
                                                        Storm pump
                  Influent pump
                  Capacitance probe
                  in drainage channel
                  upstream of facility
                  opens inlet gates
                  and starts
                  equipment on high

                  Start/stop on high/
                  low water level in
Start/stop controlled
by ultrasonic depth/
velocity flowmeter
in flow channel
Pump stroke controlled
by effluent flow

Pump speed paced on
facility flow
Pump speed controlled
by flow; pump stroke
controlled manually

Pump speed controlled
by flow; pump stroke
controlled by CRA

CRA located halfway
down influent channel
controls dose to
maintain 1 ppm

Initial high first flush
dose provided for first
10 minutes, then
control switched to
pacing on effluent flow
Pump speed controlled
by CRA output;
manual stroke

Speed controlled by
facility flow; stroke
controlled by CRA
Progressive cavity
pumps provided for
dosing caused
operational problems
                                        Rotodip feeder
                                        provided for
                                        hypochlorfte dosing
the  diffuser. A manual  bypass  should  be  provided
around the solenoid and rotameter.


A typical sodium hypochlorite piping system includes:

• Calibration standpipe,  for checking  pump capacity.

• Pulsation dampeners, to absorb  the impulses from
  the diaphragm metering pumps and  provide smoother

• Back pressure valves, to ensure that the check valves
  on the metering pump discharges seat properly.

Where chemical storage and feed equipment is located
at a higher elevation than the dosing  locations, siphon
conditions may develop that could prematurely drain the
contents of a storage  tank. Backpressure valves can
                                        prevent the siphon from developing, depending on the
                                        valve  setting and the  relative elevations  of the fluid
                                        surface in  the  storage tanks and the  valve. Some
                                        installations also feature motor-operated  valves on the
                                        drawoff from the tank to prevent the contents of the tank
                                        from siphoning or leaking between storms.

                                        PVC   or CPVC  piping  with  solvent-welded  joints
                                        commonly is used for  hypochlorite systems. Solvent-
                                        welded joints are less susceptible to leakage as compared
                                        with screw-type joints (White,  1992).


                                        The diffuser disperses the sodium hypochlorite solution
                                        into the flow. Proper mixing  is critical to disinfection
                                        efficiency, since chlorine species must come in contact
                                        with bacteria to exert the bactericidal  effects. A variety

                                                    Hypochlorite Metering
                                                    Pump No.'
                            2" CLS Line
                            to Effluent
Valve -v
1 1/2" SW,
Line J ^
«-. , . *
Overflow! ^
t >.

^^ Rotameter-j
J*C|I^ \
(0 - 20 GPM)

_. [—Calibration
+ Stand Pipe
J — I 	 . ? r Pulsation
.., •. \ Ed fDampener
Hypochlorite i EJ J=±
ige Tank / ^ j .
-^ t t T
1 Jt. ' '
n ^ r Check
Ball Valve -> \ Valve|
H]^ J(0- 1 0GPM)
prj, _ l—Qiobg valve
L Gate Valve
rL / T. T
T f f ^i

In-Line — '
Static Mixer
Hypochlorite Metering
Pump No. 2-7 ^
1 rZ^ ?
_,*I- JH-JSU

Odor Contrc
System Met
Pumps -v
Back Pressure \
Drain-7 Valve ^H J
1? 1 Tl I
^i. XJlJ

1 — Hvoochlorite M


ig 1"CL
to 10"
ri4.HS 2" CLS Line
to Influent
,_,H_.,r 2" CLS Line
* Diversion
3 Line
... ,1" niSI inn
Lixtoudol — '•- 1-1 ~>to^Qu
rina Line
                                                                 Pump No. 3

Figure 4-24.  Schematic of typical liquid sodium hypochlorite system (Metcalf & Eddy, Inc., 1991b).
of diffuser types and configurations are commonly used,
including in-channel and in-pipe  arrangements.  One
common arrangement is to  suspend  a section  of
perforated PVC pipe across a channel or pipe. The exit
velocity of  the chlorine solution from across-channel
perforated pipes should be  approximately 25-30 ft/sec
(White, 1992). Examples of diffuser types are presented
in Figure 4-25. The reader  is referred to the literature
(White, 1992) for additional  information on diffusers.

Chlorine Residual Analyzer (CRA)

The  CRA  is  useful  as  part of a  control loop for
hypochlorite feed and for residual monitoring purposes.
Some discharge permits specify  a minimum effluent
residual, where residual  has  been correlated with  a
given level  of disinfection.  Some discharge permits
specify a maximum residual, where the concern is the
toxic effects of chlorine on the receiving waters. For
chlorine dose control based on chlorine residual, the
residual sample point  should  be  located immediately
downstream of the chlorine application point, so that the
sample  is taken within 15-30 seconds of application.
The total loop time between application of chlorine and
response of the metering pump controls to the chlorine
residual signal should range from 2 to 5 minutes (White,
1992). If the CRA is located near the sample point, the
sample  piping  may be coiled  or looped to create the
minimum detention time. CRAs should not be located,
however,  in  areas exposed to  explosion  hazards.
Although  the CRA  could  be   provided  with  the
appropriately rated enclosure, this enclosure would
have to be  opened periodically to replace the stock
chemicals and  provide  routine service. For chlorine
residual monitoring and reporting purposes, the sample
point should be  located as far  downstream  of the
dosage point as practical.
The design of a CRA installation for a CSO facility must
consider both the  higher effluent solids present in  CSO
effluent as compared  with secondary  POTW effluent,
and the intermittent operation of the facility. The higher
solids  concentration  suggests that the sample  may
need to be filtered to minimize the potential for clogging
or other  interference  with  the   CRA operation.  In
addition, bare  electrode cell  CRAs  may be   more
appropriate for CSOs than permeable membrane  cells.
Although the bare electrode cells use more chemical
than the permeable membrane cells, the bare electrode
cells may be  less sensitive to higher solids concentrations.

During periods between storms, when  the CSO facility
is inactive, clean  water should be continually flushed
through the CRA to prevent the buildup of slime or algae
and to keep the  analyzer cell fresh. A flow of 0.2 to
0.5 gpm  should  be sufficient for this purpose. The

          4 in Chlorine
          Solution Piping
4 in Perforated PVC

                Continuous Weld
                         Approx. 1/3
                         Pipe Diameter
    Flexible Connection
    (Rubber or Teflon)
                                                   {fl—Blind Flange
                                                     6 in
                                                     Steel Sleeve
                      6 in
              Steel Insert Connection
  CI,Gas  Jb
     ~~  U  Injector

   Injector Water Supply
                                                                 Injector  If —

                                                                    Water Supply
                                       Chlorine Solution
               3 or 4 in
               Chlorine Solution Line
               PVC or Other Material
                                           Water Level
                                                                                                             Flexible Hose
                                                                                                             1.5 in
             Minimum Submergence, 1.5 ft
                                    Typical Diffuser
                                    Nozzle for
                                    1.5 in Hose
Figure 4-25.  Typical dlffusers used to Inject chlorine solution: (a) single Injector for small pipe, (b) dual Injector for small pipe,
             (c) across-the-pipe diffuser for pipes larger than 3 ft In diameter, (d) diffuser system for large conduits, (e) single
             across-the-channel diffuser, and (f) typical hanging-nozzle-type chlorine diffuser for open channels (Metcalf &
             Eddy, Inc., 1991 a).

0.5 gpm should  be sufficient for this  purpose.  The
chemical feed in the analyzer  can be  deactivated
between storms, but the  electronics  should remain
powered. When the facility activates during a storm
event, the control system should ignore the signal from
the CRA until  the chemical feed system in the CRA is
reactivated, and  the CRA starts reading an  actual

System Controls and Operation

A common  method for controlling the hypochlorite dose
rate is through pacing the metering pump speed or
stroke in  proportion  to  facility flow. Some facilities
provide compound loops, where pump speed is paced
by flow and stroke is controlled  by chlorine residual.
Aspects of dosage control based on  plant flow and
chlorine residual  were described earlier.  Since CSO
control facilities typically  are not staffed on a full-time
basis, some means  of automatic activation  usually is
required. Simpler control systems  are considered more
reliable for  facilities without a staff. Examples of typical
startup controls include mercury  float switches in the
influent channel,  and startup on  activation of influent

One of the disadvantages  of hypochlorite disinfection
systems is that  sodium  hypochlorite  solutions are
unstable, and the strength of a  hypochlorite solution
decreases  over  time as the hypochlorite ions break
down. The  actual rate of hypochlorite decay depends
on  a number of factors.  Decay is accelerated  by
exposure to heat,  light, and certain heavy metal cations,
particularly  iron.  Sodium   hydroxide   is  added  to
hypochlorite solutions  by  manufacturers  to improve
stability. The decomposition rate increases with increasing
solution strength.

Routine testing of the hypochlorite strength should be
part of standard facility maintenance procedures. The
benefits of monitoring hypochlorite strength are:

• Due  to  the  intermittent nature  of CSO facility
  operations, quantities of  hypochlorite may remain in
  the storage tank for extended periods. Since the dose
  rate depends on the solution strength, overestimation
  of the solution strength  may result in underdosing the

• Over a period of time, records on the decay rate and
  quantities used of various strengths  of hypochlorite
  will allow the  operator  to determine  an optimum
  strength and delivery schedule.

The strength  of  a  hypochlorite  solution  in terms of
weight percent can be  determined by making  two
consecutive 1 MOO dilutions  of a sample of hypochlorite,
and  then  analyzing for  chlorine concentration. The
concentration in mg/l of the diluted sample is equal to
the concentration in weight percent of the  original

Process Variations


A  major disadvantage of chlorine-based disinfection
systems is that the residual chlorine concentration may
have a toxic effect on the receiving waters, due either
to the free chlorine residual itself or the reaction of the
chlorine with organic compounds in the effluent. One
report on chlorine toxicity concluded  that continuous
effluent concentrations of greater than 0.01  mg/l may
exert  toxic  effects  on  resistant organisms,  while
concentrations greater than 0.002 mg/l may harm many
aquatic organisms. A variety of stable  chloroorganic
compounds  may be formed  as  byproducts of the
chlorination of wastewater, and certain chloroorganics
can exert toxic effects on  aquatic species (White, 1992;
Brungs, 1973; Jolley, 1975). With the relatively  short
contact times  available at many CSO  control facilities,
high residuals can be of particular concern and may
require consideration of dechlorination alternatives. Two
common means for dechlorinating treated effluent are
through application of gaseous sulfur dioxide and liquid
sodium bisulfite solution. A sulfur dioxide dosing system
requires apparatus  similar to  a gaseous chlorine
disinfection  system,  while sodium bisulfite requires
apparatus similar to a liquid sodium hypochlorite dosing
system. Since the  hypochlorite disinfection systems
have been more reliable than gaseous chlorine systems
for intermittent operation in CSO control applications, a
liquid sodium  bisulfate system also may be preferable
to a gaseous sulfur dioxide system for dechlorination of
treated CSOs.
Sizing  of  a  sodium  bisulfite dechlorination system  is
based on the following relationships (White, 1992):
• Sodium bisulfate is usually provided as a 38 percent
• One  gallon  of 38  percent  sodium  bisulfite  is
  equivalent to 2.17  pounds of sulfur dioxide.
• Stoichiometrically,  0.9  parts  of sulfur dioxide are
  required to remove 1.0 part of chlorine; actual ratios
  can be as high as 1.05  parts sulfur dioxide to 1.0 part
Chemical  storage and  feed equipment for  sodium
bisulfite is identical   to  the  equipment  required  for
hypochlorite systems. Sodium bisulfite storage tanks
should  be   located  away  from  and  be  clearly
distinguished from hypochlorite tanks,  since the mixing
of these two chemicals results in a violent temperature
reaction (White,  1992).


When an  NTIS number is cited in a reference, that
document is available from:
  National Technical Information Service
  5285 Port Royal Road
  Springfield, VA 22161

EPA documents with a number that begins with 600 or
625 can be ordered from EPA's Office of Research and
  Center for Environmental Research Information
  Document Distribution Section (G-72)             '
  26 West Martin Luther King Drive
  Cincinnati, OH 45268
  513-569-7562 (phone)
  513-569-7566 (fax)
APWA.  1970. American  Public Works  Association.
  Combined  sewer regulation  and  management: a
  manual of practice. U.S. Department of the Interior.

BGMA and CMT.  1977. Bainbridge,  Gee, Milanski &
  Associates, Inc.,  and Crawford, Murphy & Tilly, Inc.
  Sanitary district of Decatur combined sewer overflow
  operational study.
Brombach, H. 1989. Combined sewer overflows control
  in West Germany: history, practice, experience. Proc.
  Eng. Found. Conf. Curr. Pract. Des. Crit. Urban Qual.
  Control. New York,  NY:  American Society of Civil
Brungs, W.A.  1973.  Effects  of  residual  chlorine  on
  aquatic life. WPCF 45:2180.
Chlorine Institute. Miscellaneous publications.
CMT.  Undated. Crawford, Murphy & Tilly, Inc. Printed
  project descriptions.
Crane Co. 1970. Crane Company Cochrane Division.
  Microstraining and disinfection of  combined sewer
  overflows. 11023EVO06/70. King of Prussia, PA. June.

Dalton, F.E., and B.B. Goyal. 1989. Complex operation
  of TARP mainstream system simplified by computer
  controls. Public Works. May.

Drake, J., and C. Hunsinger. 1990. Combined sewer
  overflow  treatment  utilizing   revolutionary  swirl
  concentrator, sanitary district of Decatur.  Presented
  at the Central States WPCA, Inc., Conference.

Fair, G.M., J.C. Geyer, and D.A. Okun. 1968.  Water and
  wastewater engineering vol. II. New York, NY: John
  Wiley and Sons,  Inc.

Field,  R. 1982. Storm and combined sewers: part of the
  treatment   process.  Water   Engineering   and
Field, R. 1990. Combined sewer overflows: control and
  treatment. In: P.E. Moffa, ed. Control and treatment
  of combined sewer overflows. New York, NY: Van
  Nostrand Reinhold.

Ganley,  R.C.,  and  D.A.  McCarthur.  1991.  CSO
  abatement for Burlington, Vermont. WET 3:2.

Geisser,  D.F.,  and  S.R.  Garver.   1977.  High-rate
  disinfection: chlorine versus chlorine dioxide. ASCE
  J. of the Environ. Eng. Div. EE6. December.

Great Lakes-Upper Mississippi River Board of State
  Sanitary Engineers. 1978. Recommended  standards
  for sewage works (ten states' standards). Albany, NY:
  Health Education Services, Inc.

Hayden-Wegman, Inc.  1989. Massachusetts  Water
  Resources  Authority Fox  Point  combined  sewer
  overflow facility operation and maintenance manual.

Hayden-Wegman,   Inc.   1991.   Operations   and
  maintenance manual, Constitution Beach  combined
  sewer overflow facility.
Hedges, P. 1992. The performance of a hydrodynamic
  separator combined sewer overflow. Hydro Research
  and Development  Seminar on  Combined  Sewer
  Overflows Practice and Performance.
Henderson, R.J.A., W. McBain, and F. Pettigrew. 1981.
  The performance of an off-sewer storm-sewage tank.
  Water Poll. Control.

H.I.L. Technology, Inc. 1988,1991. Technical brochures
  on Reg-U-Flo vortex valves.

H.I.L.  Technology,  Inc.  1992. Elimination  of crude
  discharge at Ladye Bay, Clavedon, Avon. Case study
  no. CS11-7/92.

H.I.L.  Technology, Inc. Undated.  Storm King dynamic
  separator combined sewer overflow treatment. H.I.L.
  publication  no. CS/SK04.  Sneyd Lane,  Walsall
  Metropolitan Borough, England.

Holzbach, J.F. 1990.  Deep sewer tunnels operate by
  gravity.  Paper presented at  the  International
  Symposium  on Unique Underground Structures,
  Denver, CO.

Hubbell, J.W. 1990.  CSO Experience in Michigan—a
  consultant's   perspective.   Presented  at   Water
  Pollution  Control Federation Specialty Conference
  Series on Control of CSOs, Boston, MA.

James, W., and M. Stirrup. 1986. On-site microcomputer
  control  of a  combined  sewer overflow  diversion
  structure.  Water   Forum  '86  Combined   Sewer
  Overflow Control.

Jolley,   R.J.    1975.   Chlorine-containing    organic
  constituents in chlorinated effluents. JWPCF 47:601.

Kent, M. 1992. Tunnel project helps stem CSO pollution.
  Public Works. September.

Lessard, M., and M.B. Beck. 1991. Dynamic simulation
  of storm tanks. Water Res. 25:4.

Maguire,  C.E., Inc. 1981. Operation and  maintenance
  manual, Prison Point pumping station, Cambridge, MA.

McCormick,  E.H.,  D.A. Giffes,  J. Hegdal, and M.J.
  Wallis.   1990.  EBMUD's   Oakport  wet  weather
  treatment  plant design, construction  and operation.
  Presented at WPCF Specialty Conference Series on
  Control  of CSOs, Boston, MA.
Medina, M.A., W.C.  Huber, and J.P.  Heaney.  1981.
  Modeling  stormwater storage/treatment  transients:
  applications.  J.  Environ. Eng.  Div. Proc. ASCE.

Metcalf & Eddy, Inc. 1981. Wastewater  engineering:
  collection and pumping of wastewater. New York, NY:

Metcalf & Eddy, Inc.  1991 a. Wastewater engineering:
  treatment, disposal, reuse 3rd ed., New York, NY:

Metcalf & Eddy, Inc. 1991b. Operations and maintenance
  manual  for the Newport, Rhode Island, Washington
  Street combined sewer overflow facility.

MWRA.   1990.  Massachusetts  Water  Resources
  Authority.  Final  combined sewer overflow  facilities
  plan and final environmental impact report. Prepared
  by CH2M Hill.

O'Brien & Gere, Inc.  1992.   Draft  report on  CSO
  abatement program segment I performance evaluation.
  Washington, DC: Water and Utility Administration.

Pisano, W.C. 1989. Swirl concentrators revisited: the
  American  experience and new German technology.
  In:  Roesner, L, B. Urbonas, and M. Sonnen, eds.
  Urban Runoff Water Quality Control.  New York, NY:
  American Society of Civil Engineers.

Pisano, W.C.  1990.  Physical  modeling experimental
  results  for design  of CSO  and stormwater vortex
  solids separator treatment devices. Presented  at the
  WPCF Annual Conference, Washington, DC.

Smisson,   B.  1968.   Design,  construction   and
  performance of vortex overflows. ICE  Symposium on
  Stormwater Overflows.

St.   Anthony  Falls   Hydraulic   Laboratory.   1971.
  Hydraulics of long vertical conduits and associated
  cavitation.   Report  no.  11034  FLU 06/71.   U.S.
  Environmental Protection Agency.

Stahre, P., and B. Urbonas. 1993. Stormwater: best
  management practices  and  detention  for  water
  quality, drainage, and CSO management. Englewood
  Cliffs, NJ: Prentice Hall.

Thompson,   D.E.,   and   D.J.   Dobbels.    1991.
  Geotechnical/geoenvironmental planning issues for
  tunneling  projects in  urban  settings.  Proc. Rapid
  Excavation and Tunneling Conference.

Tift, E.G.,  P.E. Moffa, S.L. Richardson, and R.I. Field.
  1977.  Enhancement  of  high-rate  disinfection by
  sequential addition of chlorine and chlorine  dioxide.
  JWPCF. July.

Tighe & Bond, Inc. 1990. Operations and maintenance
  manual  for Somerville CSO facility. Massachusetts
  Water Resources Authority Sewerage Division.

Trotta, P.D.,  J.W. Labadie, and N.S. Grigg.  1977.
  Automated control strategies for urban stormwater. J.
  of the Hydraulics Div., ASCE 103(HY12).

Urbonas, B, and P.  Stahre. 1993.  Stormwater:  best
  management  practices  and  detention for  water
  quality, drainage, and CSO management. Englewood
  Cliffs, NJ: PTR Prentiss Hall.

U.S.  Army.  1978.  Domestic  wastewater treatment.
  Tech-Manual No. 5-814-3, St. Louis, MQ.
U.S. EPA. 1972. U.S. Environmental Protection Agency.
  The swirl concentrator as a combined sewer overflow
  regulator facility. EPA/R2-72/008. NTIS  PB-214687.
  Washington, DC.

U.S.  EPA.  1973a.  U.S.  Environmental Protection
  Agency. High-rate disinfection  of combined sewer
  overflow.  In:  Combined  sewer  overflow  papers.
  EPA/670/2-73/077. NTIS PB-231836.

U.S.  EPA.  1973b.  U.S.  Environmental Protection
  Agency. Microstraining and disinfection of combined
  sewer  overflows—phase  II.  EPA/R2-73/124.  NTIS
  PB-235771. King of Prussia, PA. January.

U.S.  EPA.  1974a.  U.S.  Environmental Protection
  Agency.   Urban  stormwater  management  and
  technology: an assessment. EPA/670/2-74/040. NTIS

U.S. EPA. 1974b. U.S. Environmental Protection Agency.
  The swirl  concentrator as a  grit  separator device.
  EPA/670/2-74/026. NTIS PB-234175.

U.S.  EPA.  1974c.  U.S.  Environmental Protection
  Agency. Microstraining and disinfection of combined
  sewer overflows—phase III. EPA/670/2-74/049. NTIS
  PB-235771.  King of Prussia,  PA. August.

U.S.  EPA.  1975a.  U.S.  Environmental  Protection
  Agency. Handbook for  sewer system evaluation and
  rehabilitation. EPA/625/6-91/030. Washington, DC.

U.S.  EPA.  1975b.  U.S.  Environmental  Protection
  Agency. Process design manual for suspended solids
  removal. EPA/625/1-75/003a. NTIS PB-259147.

U.S.  EPA.  1975c.  U.S.  Environmental  Protection
  Agency.  Physical  and  settling  characteristics  of
  particulates in stormwater and sanitary wastewater.
  EPA/670/2-75/011. NTIS  PB-242001.

U.S.  EPA.  1975d.  U.S.  Environmental  Protection
  Agency.  Bench-scale   high-rate   disinfection   of
  combined sewer overflows with chlorine and chlorine
  dioxide.   EPA/670/2-75/021.   NTIS   PB-242296.
  Syracuse, NY. April.

U.S. EPA. 1976. U.S. Environmental Protection Agency.
  Proceedings  of workshop  on microorganisms  in
  urban stormwater (March 24,1975). EPA/600/2-76/244.
  NTIS PB-263030. Edison, NJ. November.

U.S.  EPA.  1977a.  U.S.  Environmental  Protection
  Agency. Cottage Farm combined sewer detention
  and chlorination station, Cambridge, Massachusetts.
  EPA/600/2-77/046. NTIS PB-263292.

U.S.  ERA.  1977b.  U.S.  Environmental  Protection
  Agency.  Urban  stormwater  management  and
  technology: update  and users'  guide. EPA/600/8-
  77/014. NTIS  PB-275654.
U.S.  EPA.  1977c.  U.S.  Environmental  Protection
  Agency. Field prototype  demonstration of the swirl
  degritter. EPA/600/2-77/185. NTIS PB-272668.
U.S.  EPA.  1979a.  U.S.  Environmental  Protection
  Agency.  Combined  sewer overflow  abatement
  program—Rochester,  NY:   volume  II pilot  plant
  evaluations. EPA/600/2-79/031 b. NTIS PB80-159262.

U.S.  EPA.  1979b.  U.S.  Environmental  Protection
  Agency. Disinfection/treatment of combined sewer
  overflows--Syracuse,  NY.  EPA/600/2-79/134. NTIS
  PB80-113459. Syracuse, NY. August

U.S. EPA. 1980. U.S. Environmental Protection Agency.
  Urban  stormwater  management and technology:
  case  histories.  EPA/600/8-80/035.   NTIS  PB81-

U.S. EPA. 1981. U.S. Environmental Protection Agency.
  Field evaluation of a swirl degritter at Tamworth, New
  South  Wales, Australia.  EPA/600/2-81/063.   NTIS

U.S. EPA. 1982. U.S. Environmental Protection Agency.
  Swirl and helical bend  pollution control devices.
  EPA/600/8-82/013. NTIS  PB82-266172.

U.S. EPA. 1984. U.S. Environmental Protection Agency.
  Swirl and helical  bend regulator/concentrator  for
  storm and combined sewer overflow control. EPA/600/
  2-84/151. NTIS  PB85-102523.
U.S.  EPA.  1985a.  U.S.  Environmental  Protection
  Agency. Design manual:  odor and corrosion control
  in sanitary sewerage systems and treatment plants.
  EPA/625/1-85/018. NTIS  PB88-184031.

U.S.  EPA.  1985b.  U.S.  Environmental  Protection
  Agency. In-system storage control for reduction of
  combined   sewer  overflow—Saginaw,   Michigan,
  executive   summary.  EPA/905/2-85/001 a.   NTIS

U.S. EPA. 1986. U.S. Environmental Protection Agency.
  Municipal  wastewater  disinfection  design manual.
  EPA/625/1-86/021. Washington, DC.

Variakojis, J.,  and  A.   Quintanilla.  1989.  Facilities
  planning study  update supplement and summary.
  The  Metropolitan  Water Reclamation  District of
  Greater Chicago.

Vitasovic, Z., R. Swarner and E. Speer. 1990. Real time
  control system for CSO  reduction. Water Environ.
  and Tech. March.

WEF.  1992. Water Environment Federation. Design of
  municipal wastewater treatment plants volumes I &
  II. WEF manual of practice no. 8, ASCE manual and
  report  on  engineering practice no.  76.  2nd ed.
  Alexandria, VA: WEF, and New York,  NY: American
  Society of Civil Engineers.
West, M.C., T.L. Moy, and H.A. Elmendorf. 1990. Atlanta
  combined sewer overflows present and future. ES
  Engineering Science Technical Paper.

Westfall,  D.E.  1990.  Dropshafts  for conveyance of
  combined sewer overflows to tunnels. Presented at
  the  Water  Pollution Control Federation Specialty
  Conference Series on  Control of Combined Sewer
  Overflows, Boston, MA.

Wheatley, N.J.  1991. Massachusetts Water Resources
  Authority  CSO  control   program.  Proc.  Rapid
  Excavation  and Tunneling  Conference.  Society of
  Mining, Metallurgy, and Exploration.

White,  G.C.  1992.  Handbook  of chlorination  and
  alternative  disinfectants, 3rd ed. New  York, NY: Van
  Nostrand Reinhold.

Whittaker, B.N.,  and  R.C. Frith. 1990.  Tunneling:
  design,  stability  and  construction.  London:  The
  Institute of Mining and  Metallurgy.

Wilcoxon, N., and C.R.  Hunsinger, Jr.  1991. Vortex
  solids separator: a new CSO technology. WET 3:6.

WPCF.  1986.  Water  Pollution  Control  Federation.
  Wastewater  disinfection,  manual  of  practice no.
  FD-10. Washington, DC.

WPCF.  1989.  Water  Pollution  Control  Federation.    WPCF/ASCE. 1974. Joint Committee of Water Pollution
  Combined  sewer  overflow  pollution  abatement,      Control Federation and the American Society of Civil
  manual of practice no. FD-17. Alexandria, VA.           Engineers. Design and construction of sanitary and
                                                     storm sewers.  WPCF Manual of  Practice No.  9;
                                                     ASCE Manual and Report on Engineering Practice
                                                     No. 37. New York, NY: WPCF/ASCE.

                             Costs for CSO Control Technologies
This chapter summarizes information on costs associated
with construction and operation of treatment technologies
presented in this report. This  information can be used
to develop preliminary budgetary estimates and provides
a basis for comparing technologies. The cost information
also can be useful in characterizing economic sensitivity
in relation to various design alternatives for a proposed
facility. Readers should consult the source documents
cited for additional detail.

The cost relationships reflect a consolidation of cost
data, some of them estimates, compiled from a number
of comprehensive cost assessments  that  extract data
from facility plans and from a  number of individual site
studies not included in the broad-based data sets. As
noted, actual costs for CSO facilities varied  considerably
for facilities of similar type and design capacity. This is
attributable to a  variety  of site-specific  factors that
influence project costs.

The primary parameter influencing the estimated cost
for a CSO treatment facility is either design flow rate or
storage volume. For treatment-based control measures
that are designed on the  basis of flow rate, costs are
related to flow rate in million gallons per day (MGD). For
storage facilities, the  primary design parameter  is
storage volume, or the storm size selected as a design
basis, and the cost relationship  is based  on storage
volume in million gallons (MG).

Sources of Cost Data

Cost  estimating  procedures  for  CSO storage and
treatment facilities have been available since the late
1970s. Updates have been developed at regular intervals.
Recent  cost  data for CSO  storage and treatment
facilities have been developed (U.S.  EPA, 1992) and
other sources of cost information exist (Field, 1990; U.S.
EPA, 1976, 1977a,b).

The cost curves developed from the literature are based
on a compilation  of information extracted from actual
construction costs, estimates developed for CSO facility
plans, consultant-supplied  data for  specific projects,
and other published sources. The costs provided for
CSO control do not include land acquisition; engineering,
legal,  fiscal, or administrative services; contingencies;
or construction loan interest. The exception to this is
screening facilities, where these cost components could
not be isolated and extracted.

Other cost  estimating  references,  such as those
extrapolated from POTW costs (U.S. EPA, 1978,1980),
may be  used  to refine estimates developed from CSO
costing references, but should be used cautiously. Many
POTW costing procedures are based on average daily
flow, while CSO facilities usually are based on peak flow
or storm volume. Possible differences in the relationship
between peak and average flow and  its translation to
the design surface area should be considered.

Construction Costs for CSO Controls

Figure  5-1  summarizes  the  relationship between
construction cost  and design  capacity for a range of
CSO control technologies. These cost curves are based
on recent studies  that assembled information on CSO
control costs (U.S. EPA, 1992), and reflect the cost of
the  basic structure  and  ancillary  equipment  (e.g.,
grates, valves, conduits). Associated pumping facilities
are included for some, but not all of the facilities. Land
acquisition and professional service costs are excluded.
The  source report provides details on cost elements
included for individual sites. Local cost estimates for
these elements should be  developed  and added for a
more complete estimate, whenever appropriate.

Most of the cost curves display a line of best fit through
the plotted data, in addition to the plotted points showing
cost versus design size for individual  sites. The linear
regression line is described  by an  equation of  the
following form:

         COST = aQb   or COST = aVb

where Q is the design flow in MGD for treatment units,
and  V is the storage volume in MG, and the values a
and  b are unique for each line fit.

The construction cost relationships shown by Figure 5-1
are summarized by the equations in Table 5-1.

Inspection of the relative scatter of individual site values
about the regression  lines in  Figure 5-1 provides a

si 100°
S 100
i 10
* 10
1 0.1
g 1
I «•'
Figure 5-1
Table 5-1.
CSO Cont

ju.g.ui }trniM4aooj»i.|j
; : I"1; i \ A
' '*•*' e'o fogf
\"l 	 ipTO?
- A i-^**T < "1 !t
o-^W -S> '• i i 1
;<» i!,i!i'n t i li!
LQE: ' '.-' ' '"-.
;.; i i dru »»(*> :ii
It 1 1 1 t t M ' mW I { II
Ij^Sh?: ! ;S
ili ! ]i!l! — c —
~ t ill

Hi ' i \W.\ i ,"l!
,Jf -
• 1 l!:
Surface Units 1
Deep Tunnels 1

i 1 II
9.1 1 10 100 1000
Storage Volume MG
i •« i d & \ fcfc
l\\\" " i I "iTj 1 	 '" I"3T$I ii ^,^'1 ! f i-!
;, ,:,:;.:, ^,a^£rf^ : -"-':
Iil| | I | a I kit 11,1 !lj. i I , f. ,1
:|||| 1 ! lli!!! ! ! 1 11 If! 1 ! 1 l.ifE
- .. i ~ t ,.
1 1 i i i i i i i I i i i 1 1 1 1 1 1 1 1 1 ! 1 1 1 [i
1 10 100 1000
• Design Flow MQD
=5 '"I'M S'iS'HI ! i1Jk!ll-!
•;-., SEDIMENTATION t *,!*...
| EMR-1600 4

!-M ^J*J*^rf^*^*^]"
Y^ lJ«— * ' 	
:ii;i """\"l 'I'1!!1 ' ! 	
1 ! ' 3 1 f "j
* i £ i i £ 1 1 -
! | ill!1!
Sedimentation Basin 1
Wd tor Chem Feed Eqpt|
, |M , ,,,!,,, 	 | , , | , , ,|
1 10 100 1000
Design Row MGD
. Construction costs for CSO controls.
Cost Equations for CSO Control Technology
rol Technology Cost Equation*
Storage basins 3.637 V0'828
Deep tunnels 4.982 V0'795
Swirl concentrators 0.176 Q°'611
Screens 0.072 Q0'843
Sedimentation 0.211 Q0-668
Disinfection 0.121 Q0'464
5 10 S
C 01 t Lit

SCREENS!," , - : , , ,:

^^SBp'i"!1 .'!':=
i i i 1 1 1 > i i i i < 1 1 1 1 i i i 1 1 1 1 1
1 10 100 1000
Design Flow MGD
5 I0 ^J
3 i!:r

ISINFECTION;: " ! ' ' !> 1 : ] - ' }
FNB » tsaa
- ii,i~i- i ,-.,..! ~ ,"i ' =
Ll^r^^^] ' '":
! '!;'n" <-"';': ''>: -
1 10 100 1000
Design Flow MGD
Design Range ENR Index
0.15 to 30 MG 4,800
1.8 to 2,000 MG 4,800
3 to 300 MG 4,800
0.8 to 200 MG 4,800
1 to 500 MG 4,500
1 to 200 MG 4,500
* V = volume (MG); Q = flow rate (MGD).
sense of the degree of uncertainty in the above cost
relationships. Reliable  cost estimates for a particular
CSO  control facility require a  conventional material
take-off and application of normal costing procedures,
when facility design is sufficiently advanced to support
such an analysis.
O&M Costs for CSO Controls

Operation and maintenance costs are difficult to predict
because of the intermittent  use of CSO treatment
facilities. Actual O&M costs are indicated in the data in
Figure 5-2 to be a function of the number of overflows
experienced by the facility, as well as the design capacity.

The frequency at which a specific CSO facility receives
an overflow event to process depends strongly on the
capacity of the collection system it serves and varies
from one year to another, depending on the amount and
pattern of rainfall experienced.

The  significant effect of  the frequency of activation is
illustrated by the O&M cost curves for screening facilities,
shown in Figure 5-2. All the major components of O&M
costs, such as energy consumption, labor requirements,
residuals disposal, and equipment maintenance  are a
function of the facility use. This, in turn, is a function of
the rainfall quantity and the number of wet weather
events; therefore, O&M  costs are highly site specific.
The O&M curves that appear in  published reports and
literature should be used cautiously.


When  an  NTIS number is  cited in  a reference, that
document is available from:
  National Technical Information Service
  5285 Port Royal Road
  Springfield, VA22161

Field, R. 1990. Combined sewer overflows: control and
  treatment. In: Moffa, RE.,  ed.,  Control and treatment
  of  combined sewer overflows. New York,  NY: Van
  Nostrand Reinhold.
U.S.  EPA. 1976. U.S. Environmental Protection Agency.
  Cost  estimating manual—CSO  storage-treatment.
  EPA/600/2-76/286 (NTIS PB 266 359).

U.S.   EPA.  1977a.  U.S.  Environmental  Protection
  Agency.  Urban  stormwater   management   and
  technology—update and  user's guide.  EPA/600/8-
  77/014 (NTIS PB 275  654).

U.S.   EPA.  1977b.  U.S.  Environmental  Protection
  Agency. Nationwide evaluation of CSOs and urban
  stormwater discharges—volume III: cost assessment
  and impacts. EPA/600/2-77/064b (NTIS PB 266 055).

U.S.  EPA. 1978. U.S. Environmental Protection Agency.
  Innovative and alternative technology assessment
  manual. EPA/430/09-78/009.

U.S.  EPA. 1980. U.S. Environmental Protection Agency.
  Treatability   manual—volume  IV:  cost estimating.
  EPA/600/8-80/042D (NTIS PB80-223084).

U.S.  EPA. 1992. U.S. Environmental Protection Agency.
  CSO needs survey, and in a report prepared for ERA
  entitled: cost estimates for select combined sewer
  overflow  control  technologies.  Office  of Water,
  Municipal Support Division. Washington, DC.
    W   100
                 ENR = 4500;
                                        10 O/F events /yr
                                        30 O/F events/yr
                        10            100
                        Design Flow MGO
                        10            too
                        Design Flow MGO
                        Design Flow MQD
Figure 5-2.  O&M costs for CSO controls.
    •U.S. GOVERNMENT PRINTING OFFICE: 1995-650-006/22022