Astoria, OR

5 MGD

Springfield, OR
6.5 MGD

Salem, OR
62 MGD

DESIGN

Now we’ll get into some design aspects of slow sand filtration. Aerial photo of Astoria, OR
(a 5 MGD) was taken by Frank Wolf in 2010. Other photos are from GoogleMaps
downloaded in 2013.

1
 GENERAL SCHEMATIC (WHO, 1974)

This is a schematic from the 1974 World Health Organization (WHO) “Slow Sand Filtration”
design manual (Huisman & Wood, 1974. pg 18).

2
 GENERAL SCHEMATIC (WHO, 1974)

This is figure 18 showing more details on the filter controls. Figure 18 is from the 1974
World Health Organization (WHO) “Slow Sand Filtration” design manual (Huisman & Wood,
1974. pg 64).

3
 GENERAL SCHEMATIC (USEPA)

This is another schematic from the USEPA.

4
 MORE DETAILED SCHEMATIC

A more detailed schematic is shown here, beginning with the raw water inlet into the filter
box. The filter box is equipped with a drain and water for backfilling with filtered water.
Supernatant (or headwater) filters through the sand bed and support gravel, out of the
under drains through a flow meter and control valve and into a flow control structure. The
adjustable weir keeps the sand bed from de-watering when the filtration rate declines
towards the end of a filter run. The weir is adjustable to facilitate draining the filter bed
during cleaning. Filtered water then flows to the clearwell for disinfection. Piezometers are
shown where they can be used to measure headloss across the filter bed as well as the
tailwater level.

5
 “GOOSENECK” WEIR

This is an example of a “Gooseneck” style weir construction showing process control

monitoring points.

6
 GRAVITY FED SYSTEM

Gravity Fed system

7
 TELESCOPING VALVE

Telescoping Valve

8
 DESIGN - SS VERSUS RAPID RATE
Parameter Slow Sand Filters Rapid Rate Filters
Influent Flow Continuous Intermittent

Filter Box  - designed to overflow  - not intended to overflow

Filter Media - sand  - sand

- GAC (rare) - anthracite
- GAC (more common)
Underdrain  
“Backwash”  - slow filling from bottom  - high rate (and low rate)
mechanisms of filter – removal of flow designed to suspend &
entrained air not media wash the media
expansion.
Surface Agitator 
(breaks up media
during backwash)

This slide compares slow sand and rapid rate filtration, both utilize similar media (sand),
however, there are some important differences. For example, slow sand filters are design
to operate continuously, where rapid rate plants are meant for intermittent operation.

9
 DESIGN - SS VERSUS RAPID RATE
Parameter Slow Sand Filters Rapid Rate Filters
Filtration Rate 0.03 – 0.1 gpm/ft2 2-4 gpm/ft2
Water Above Sand ~4-6 ft ~ 5 ft
Sand Bed Depth ~ 24-48 inches ~ 24-30 inches
Sand Effective Size (d10) 0.15 – 0.35 mm 0.45 – 0.55 mm
Retention Time above Sand 15 hrs 9 min
Retention Time in Sand Bed 3.2 hrs 2 min
Cycle Length 1-6 mo 1-4 days
Removal mechanisms Chemical, physical, and Chemical and physical (depends
biological (no chemicals) on proper coagulation)
Turbidity Removal Variable even if optimized < 0.1 NTU when optimized
< 5 NTU by regulation Good indicator of filter
Not indicative of filter performance and pathogen
performance or pathogen removal.
removal. Coagulation/flocculation
Sub micron particles are not removes even sub-micron
readily removed. particles.
Giardia Removal >3.0 log >3.0 log
Raw Water Turbidity <10 NTU 100+ NTU

Filtration rates for rapid rate filters is roughly 40 times that of slow sand filters and the sand
effective size is roughly twice that of slow sand media. Retention time above the slow sand
bed is measured in hours rather than minutes and the filter run is weeks long rather than
days long for rapid rate plants. The removal mechanism for slow sand filters incorporates a
biological process without the addition of any coagulation chemicals. Coagulation is critical
for effective rapid rate filtration. Due to the coagulation, rapid rate plants are far less
sensitive to elevated raw water turbidity. In spite of all these differences, the removal
efficiencies for Giardia are the same at 3.0-log and both are capable of producing finished
water with very low turbidity.

10
 COMMON DESIGN PITFALLS

Common Design Pitfalls

1. Inappropriate source water quality => inappropriate application
2. Not conducting a pilot study
3. Improperly designed under drains
4. Poorly designed filter piping
5. Inadequate flow control and air binding
6. Insufficient head loss allowed
7. Insufficient sand bed depth
8. Inappropriate filtration rate and variability
9. Poorly specified sand and gravel media (effective size, uniformity, etc.)
10. Poor access to filter bed for cleaning and re-sanding
11. Insufficient sample ports
12. Failure to have the operator involved in design process
13. Failure to provide a good O&M manual with filter cleaning/ripening protocols

Even with the best design, there are a number of variables that can have a big impact on
performance. Raw water characteristics like turbidity, color, and colloidal content for
example. Other critical variables include sand size and uniformity, flow control and
management of air binding, headloss development, sand bed depth, filtration rate and flow
variability. Allowing sufficient time to mature once a filter has been newly sanded (usually
4 – 6 weeks) and allowing the filter to ripen once cleaned (24 – 48 hours) are very critical to
optimal performance.

11
 DESIGN RECOMMENDATIONS -
EXAMPLE

Source: Vigneswaran, S. and C. Visvanathan. 1995

http://www.nesc.wvu.edu/ndwc/pdf/OT/TB/TB14_slowsand.pdf

There are a number of design references, such as the one shown here.

12
 DESIGN MANUAL - 1991

“Manual of Design for Slow

Sand Filtration“ . David
Hendricks & American Water
Works Association, 1991.
ISBN 978-0898675511

Excellent resource that

covers design in great detail

The “Manual of Design for Slow Sand Filtration” covers design in great detail.

13
 3 OTHER DESIGN REFERENCES
The 3 other main design references include:
1. Recommended Standards for Water Works (a.k.a., “Ten
States Standards”, 2012);
2. “Slow Sand Filtration for Community Water Supply”,
International Research Center for Community Water Supply
and Sanitation (Visscher et al., 1987)
3. “Slow Sand Filtration”, World Health Organization (Huisman
& Wood), 1974;

http://www.who.int/water_sani
tation_health/publications/ssf/e
http://10statesstandards.com/ http://www.irc.nl/page/4530 n/index.html

There are, however, 3 main design references that have either stood the test of time, like
the manuals on slow sand filtration produced by the World Health Organization and the
International Research Center for Community Water Supply and Sanitation or that are
commonly referenced by State regulatory agencies, such as the Ten States Standards.

14
DESIGN CRITERIA – 3 OTHER REFERENCES
The design specifications for these 3 sources are summarized here
Comparison of Design Specifications
(Design Period, Operation, and Filtration Rate, # of Units, and Supernatant Depth)

Reference WHO Manual IRC Manual Ten States Standards

(Huisman & Wood, 1974) (Visscher et al., 1987) (2012)

Design Period 7-10 Years 10-15 years Not Specified

Mode of Continuous 24 hr/day Not Specified

Operation

Filtration Rate 0.04 – 0.08 gpm/ft2 0.04 – 0.08 gpm/ft2 0.03 – 0.1 gpm/ft2
(flow rate ÷ (0.1 – 0.2 m/hr) (0.1 – 0.2 m/hr)
filter area)

Filter Units 2 minimum 2 minimum 2 minimum

(a.k.a., cells)

Supernatant 39 – 59 in, 79 in max 27 – 39 in, 60 in max 36 – 72 in

Depth (100 – 150 cm, 200 cm (70 – 100 cm, 150 cm (91 – 183 cm)
max) max)

15
DESIGN CRITERIA – 3 REFERENCES
Comparison of Design Specifications
(Minimum Sand Bed Depth)

Reference WHO Manual IRC Manual Ten States Standards

(Huisman & Wood, 1974) (Visscher et al., 1987) (2012)

Minimum Filter 28 – 35 in (70 – 90 cm) 18 – 35 in 19 in

Bed Depth* (45 – 90 cm) (48 cm)

*The design should add these minimum sand bed depths to the amount of sand anticipated
to be removed during cleanings throughout the design life of the filter (estimates of sand
removal can be determined based on cleaning data obtained during pilot testing). Filters
designed for harrowing only need to account for minor losses, since sand is not removed due
to scraping when using this method of cleaning.

16
DESIGN CRITERIA – 3 REFERENCES
Comparison of Design Specifications
(Filter Sand Effective Size and Uniformity Coefficient)

Reference WHO Manual IRC Manual Ten States Standards

(Huisman & Wood, 1974) (Visscher et al., 1987) (2012)

Filter Sand 0.15 – 0.35 mm 0.15 – 0.30 mm 0.15 – 0.30 mm

Effective Size
(d10)

Uniformity 1.5 – 3 <3–5 < 2.5

Coefficient (U)

Other specifications include:

1. Percent of fines passing the #200 sieve < 0.3% by weight
(can impact post sanding turbidity levels and length of filter to waste time
needed to “wash” the fines out)
2. Acid solubility < 5% (can impact sand grain characteristics, effective size, and
uniformity coefficient if acid soluble)
3. Apparent specific gravity > 2.55

17
 TEN STATES STANDARDS

Ten States Standards

http://10statesstandards.com/index
.html

Member States and Provinces

Illinois New York
Indiana Ohio
Iowa Ontario
Michigan Pennsylvania
Minnesota Wisconsin
Missouri

Since the Ten States Standards were developed with the participation of many state
agencies, are widely recognized, and lay out fairly concise specifications for slow sand
filters, I will use this reference to discuss a little about each specification.

18
TEN STATES STANDARDS
APPLICATION TO BE BASED ON STUDIES

4.3.4 Slow Sand Filters

The use of these filters shall require prior engineering

studies to demonstrate the adequacy and suitability of this
method of filtration for the specific raw water supply.

19
 PILOT TESTING

Pilot testing
Recommend 1-year pilot study in
order to account for seasonal
variability of source water
variables (required in some
cases)

Pilot testing is to determine

feasibility (i.e., “will it work
when we need it to?”), not just
to discover O&M issues.
2010. Walla Walla, WA Pilot filters

A pilot study of at least a year should be conducted to determine the suitability of slow
sand filtration for the available source water and required system demands. Pilot testing
can also uncover unanticipated O&M issues.

20
 PILOT TESTING BENEFITS
Pilot testing
Pilot testing is relatively easy and can be used anytime to test
cleaning procedures, model ripening times, or evaluate new
media.

<< It can be done on a

small scale
(1991 - Alsea, OR)

or a very large scale >>

(2010 - Walla Walla, WA)

Pilot tests can be small scale like the one on the left with a 12-inch diameter column or
large scale like the 4 pilot columns on the right. The one on the left was used to pilot
media used in Alsea, Oregon. The pilot study used by Walla Walla Washington in 2010 was
used to evaluate media from 3 different sources in 3 of the columns. A 4th column was
used to evaluate the effects of a roughing filter.

21
 INFO GAINED BY PILOT TESTING
Pilot testing can yield valuable information such as:
1. The flow to be expected (will the proposed design be enough
to meet demands or will more sand bed area be needed?).
2. Cleaning frequency. As sand is removed during cleaning, the
frequency of cleaning can yield information about how many
years the sand will last before re-sanding is needed.
3. O&M requirements that may change seasonally
4. If algae growth will have an adverse impact
5. Cold temperature effects (may require longer filter-to-waste
times after ripening.
6. Ripening time (Use plots of turbidity and coliform)

Pilot testing yields information on what raw water characteristics may adversely impact
performance and operation such as algae, cold temperatures, etc.. Pilot test results can be
used to evaluate different sand characteristics and determine how much filter area is
needed to meet the anticipated demand. It can also be used to estimate operation and
maintenance costs associated with cleaning and re-sanding.

22
 STUDIES SHOULD REPLICATE FULL SCALE

So what are the features of a full scale plant that should be considered in designing pilot
filters? Most properly designed slow sand filters have the same basic design elements,
with some variations. There is a raw water inlet (#2 in the diagram) and usually some way
to remove surface scum (#1) and drain the headwater for cleaning (#3). In a mature filter,
there is the schmutzdecke filter skin, filter bed, and underdrain system with some sort of
effluent flow control, which is typically either a valve or moveable weir. The underdrain
system also functions to drain the filters for cleaning or resanding. During cleaning, the
filter bed should not be completely drained, but rather drained just enough to allow the
bed to be walked upon or to allow machinery to safely operate during cleaning. In this
example, this level is dictated by the design of the outlet chamber (#11) and overflow weir
(#8). In other cases, this level is controlled by valves. The water level is maintained in this
example by the overflow weir indicated by #8 in the diagram. Once cleaned, there is often
the ability to slowly refill the filter bed with filtered water from another filter as shown by
#5. This allows air that gets into the filter bed during cleaning to be purged. This
“backfilling” continues until the headwater is roughly 1-ft above the filter bed, which
protects the bed from scouring that can occur when top filling with raw water commences.
There should be the ability to filter to waste for at least 48 hours until ripened (#4).

23
 PILOT FILTER
SCHEMATIC
Intent is to replicate full-
scale design.

Pilot testing schematics can

be simple, but should
clearly show key features
that can simplify operation
and improve data
collection.

A schematic is an excellent way to ensure that key features of a full-scale plant are
incorporated into a pilot filter. Pilot testing schematics can be simple, but should clearly
show key features that can simplify operation and improve data collection. This is a
schematic of the pilot filter used at Humbug Mountain State Park in Curry County, Oregon.

24
 PILOT FILTER SCHEMATIC - GAC
This design is set
up to evaluate a
layer of granulated
activated carbon
(a.k.a. GAC sandwich)

This design incorporates a layer of granular activated carbon for organics removal. This is
sometimes called a “GAC sandwich”. These plans are for a pilot filter for the City of Salem.

25
 PILOT FILTER
SCHEMATIC

This is a plan view

of a pilot filter
(shown above) used
by the City of Walla
Walla in 2010-2012.

This is another schematic of a pilot filter used by the City of Walla Walla in 2010-2012.

26
 PILOT FILTER
SCHEMATIC
This is a section view
of the filter used by
the City of Walla
Walla, WA.

This is a section view of the filter used by the City of Walla Walla, WA.

27
 PILOT FILTER
SCHEMATIC

This is another schematic of a pilot filter.

28
 PILOT FILTER COLUMN
Pilot Filter Material
1. PVC, concrete, fiberglass, etc. (5 gallon buckets have been
used). Design some durability into it in order to retain filter
for future studies
Pilot Filter Size
1. 8-12 ft high (replicate full scale filter)
2. 12 – 36 inch diameter
3. Diameter dictated by room needed to fit under drains,
sample ports, etc. and accommodate cleaning. A joint
constructed just above the sand bed can facilitate cleaning in
small diameter filters. A lip below the sand surface can help
eliminate side-wall effects (short-circuiting) of smaller
diameter filters.

Pilot filters can be made out of various materials, but should be made fairly durable so that
they can be retained for future studies. Pilot filters should be designed in order to
accommodate the entire sand bed, support gravel and under drain system as well as the
headwater that would be expected from the full scale installation. Larger diameter filters
can help mitigate short-circuiting at the sidewalls, however, a lip built into the filter below
the sand bed can help compensate for smaller diameter filters.

29
 PILOT FILTER MEDIA
Pilot Filter Media
1. The media should be the same as that intended to be used in
the full scale installation.
2. Multiple, identical filters should be used to evaluate various
sources or specifications of sand.
3. Pilot filter media should be delivered and washed as would be
done at full scale in order to help estimate the time needed to
wash out fines and for the filter to fully mature.
4. The filter bed and support gravel layers should be installed to
the same depth anticipated to be used at full scale.

Filter media and support gravels should be supplied, washed, and installed in a similar
manner to that anticipated with the full scale installation. This helps to determine the filter
wash-out and maturation periods that more closely resembles full-scale conditions. The
pilot filter should also be covered or left uncovered like the anticipated full-scale design.

30
 COVERED PILOT FILTERS?
Pilot Filter – to cover or not to cover
1. Pilot filters should be covered if the intended full-scale design
includes a cover. If not, it may be advantageous to have the
ability to cover the pilot filter in order to observe differences
in filter performance.

2. Note that covered

filters may not
develop a discernable
schmutzdecke, rather
they may exhibit a
layer of darker sand at
the surface when
maturing.
Tillikum Retreat Center, OR – Covered “Blue Future” filters

The pilot filter should also be covered or left uncovered like the anticipated full-scale
design. It may be advantageous to incorporate provisions for covering the pilot filter in
order to study differences in filter performance or to evaluate filter covers at a later date.
This photograph of excavated media was taken in 2012 at the Camp Tillicum Retreat in
Yamhill County, Oregon for a newly installed full-scale filter that was being filtered-to-waste
until fully mature.

31
 PILOT FILTER SAMPLE PORTS
Note the
sample ports
at various
locations to
evaluate
removal
mechanisms
throughout
the filter bed.

Here is another example of a pilot filter. Note the sample ports at various column depths.
This is not typical, but it demonstrates how a pilot filter can be used to evaluate the
removal mechanisms at work throughout the filter bed.

32
 PILOT TEST MONITORING –
RAW WATER
Raw Water

Sample location Parameter Sample frequency Laboratory or field

analysis needed
Raw water Turbidity Daily Field
Temperature Daily Field
Apparent color Weekly Field
pH Weekly Field
Alkalinity Weekly Field
Coliform (total and E. coli) Weekly Laboratory
Dissolved oxygen Weekly Field
UV 254 absorbance, TOC Monthly Field or laboratory
and/or THM formation analysis
potential
Iron and Manganese Monthly Laboratory
Algae identification and Quarterly or with algae Laboratory or field
enumeration (toxins if blooms identification.
indicated) Laboratory or field test
strips for toxins.

33
 PILOT TEST MONITORING –
FILTER EFFLUENT
Filter Effluent

Sample location Parameter Sample frequency Laboratory or field

analysis needed
Filter effluent Turbidity Daily Field
Temperature Daily Field
Apparent color Weekly Field
pH Weekly Field
Alkalinity Weekly Field
Coliform (total and E. Weekly Laboratory
coli)
Dissolved oxygen Weekly Field
UV 254 Absorbance, Monthly Field or laboratory
TOC and/or THM analysis
formation potential
Iron and Manganese Monthly Laboratory
Algal toxins If indicated by raw Laboratory or field test
water testing strips for toxins.

34
 PILOT TEST MONITORING -
OTHER
Other

Sample location Parameter Sample frequency Laboratory or field

analysis needed
Other Filter head loss Daily Field
Flow rate Daily and with changes Field
Filter run length Record cumulative days Field
Cleaning frequency Record events and Field
unusual circumstances
Depth of Sand Initial amount and Field
amount remaining after
each cleaning

35
 PILOT TEST CONCLUSIONS
Key pilot test conclusions that can influence design include…
1. Flow
• Will it meet system demands?
• What sand characteristics are most appropriate?
• How much filter area do I need?
• Do I need to account for slower flows due to cold temps or
should they be covered?
2. Cleaning.
• What frequency?
• How much ripening time? Cold water effects?
• How long can I go without a filter?
• Will I need multiple smaller filters, rather than fewer large
filters due to cleaning and ripening requirements?
• How long will the filter last & how deep will the bed need to
be to make it last given the cleaning required?

36
PILOT TEST PLAN AND REPORT
Document the pilot test plan and
results for future reference

37
TEN STATES STANDARDS
RAW WATER QUALITY

4.3.4.1 Quality of raw water

Slow rate gravity filtration shall be limited to waters

having maximum turbidities of 10 units and maximum
color of 15 units; such turbidity must not be attributable
to colloidal clay. Microscopic examination of the raw
water must be made to determine the nature and extent
of algae growths and their potential adverse impact on
filter operations.

38
 RAW WATER QUALITY
Recommended Limits for Raw Water
(Source Water Characteristics)

Turbidity < 10 NTU

(colloidal clays are not desirable)
True Color < 5 platinum color units
Coliform Bacteria < 800 /100 ml (CFU or MPN)

Dissolved Oxygen (DO) > 6 mg/l (filtered water DO > 3 mg/l)

Total Organic Carbon (TOC) < 3.0 mg/l
(low TOC to prevent DBP issues)
Iron & Manganese Both < 1 mg/l Each

Algae < 200,000 cells/L (depends upon type)

Even with the best design, there are a number of variables that can have a big impact on
performance. Raw water characteristics like turbidity, color, and colloidal content for
example. Other critical variables include sand size and uniformity, flow control and
management of air binding, headloss development, sand bed depth, filtration rate and flow
variability. Allowing sufficient time to mature once a filter has been newly sanded (usually
4 – 6 weeks) and allowing the filter to ripen once cleaned (24 – 48 hours) are very critical to
optimal performance.

39
TEN STATES STANDARDS
NUMBER OF FILTERS

4.3.4.2 Number

At least two units shall be provided. Where only two

units are provided, each shall be capable of meeting the
plant design capacity (normally the projected maximum
daily demand) at the approved filtration rate. Where
more than two filter units are provided, the filters shall
be capable of meeting the plant design capacity at the
approved filtration rate with one filter removed from
service.

40
NUMBER OF FILTERS
How do you determine a reasonable filter area?

An individual filter should be small enough to allow it to be cleaned in 1 day.

Determine the filter size as follows:

Area of 1 filter = (cleaning rate in ft2/person/hr)

x (no. of people available for cleaning)
x (hours allotted to cleaning)

Example:
Cleaning rate: 1,000 ft2/5 persons/hr (Cullen and Letterman, 1985)
(1” of sand hand shoveled with hydraulic conveyance)

Number of people: 2 minimum (think safety)

Hours estimated for cleaning: 2.5 hrs (desired)
Area of 1 filter: 1,000 ft2/5 persons/hr x 2 people x 2.5 hrs
= 1,000 ft2 => 20 x 50-ft filter

41
NUMBER OF FILTERS
Is there such thing as too small?

The minimum size of a filter depends upon the:

1. Cleaning method and equipment access needs
2. System demands
3. If covers are needed Direct, In-line, DE, Slow Sand, or Cartridge/Bag
4. Construction costs Filtration Plants

Huisman and Wood (1974)

and Sharp et al (1994)
indicate a minimum area for
one filter of about 1,000 ft2
(100 m2). This is due to
construction costs being
lower per ft2 with larger filters
(economy of scale).

Small modular units are

common and can be very cost
effective

42
 NUMBER OF FILTERS
How do you determine the number of filters needed?

Equation for number of filters needed: N = 1 + (Q / (HLR * A))

Where:

1. HLR = hydraulic loading rate (gpm/ft2)

2. Q = flow needed to meet demands (gpm)
3. A = The sand bed surface area of one filter bed (ft2)
4. N = total number of filter beds needed (assumes 1 filter is taken off-line for
cleaning and storage can meet peak hour demands)

Example: How many filters are needed, given a peak day demand of 250 gallons
per capita per day and a community of 600 people. The peak design filtration rate
is 0.1 gpm/ft2. A minimum rate of 0.05 gpm/ft2 has been identified through pilot
testing for operation during cold conditions and to accommodate filters left in
service that may be near the end of their filter run. There is also a desire to limit the
size of each filter to 20’x50’ in order to facilitate the cleaning:
1 + (250 gpcpd x 600 people x 1 day/1440 minutes) = 3.08 = 3 filters
(0.05 gpm/ft2 x (50-ft x 20-ft))

The number of filters needed, can be determined using this equation. This equation
assumes that the filters are all to be of equal size, only one filter is taken off-line at a time
for cleaning, and that peak day demand is the design flow for the plant (i.e., peak hour and
fire flow demands can be met by available distribution system storage). Notice, the desire
to keep the filter area to a manageable size in order to facilitate cleaning operations and
minimize filter down-time. A low rate of 0.05 gpm/sqft is chosen for design as it represents
what may be needed should scraping have to occur during colder water temperatures or for
filters left in service that may be near the end of their filter run during the remainder of the
year (cleaning should generally be scheduled to avoid very cold weather).

43
NUMBER OF FILTERS
Are more filters better?
Equation for minimum number of filters needed: N = 1 + (Q / (HLR * A))

In the previous example, 3 filters were determined to be needed, any 2 of which

are capable of meeting 100% of the peak day demand (PDD) to allow for 1 filter
being taken out of service for cleaning. This means that each filter is able to meet
50% of the PDD.
3 filters x 50% of PDD = a plant capacity of 150% x PDD
(1 filter off-line leaves 2 filters to meet 100% of PDD)

If 4 smaller filters were constructed, each filter would only need to be capable of
meeting ~33% of the peak day demand to allow for 1 being taken out of service.
4 filters x 33% of PDD = a plant capacity of 132% x PDD
(1 filter off-line leaves 3 filters to meet 100% of PDD)

The capital cost involved with fewer large filters should be carefully weighed
against the benefits of having a higher number of smaller filters (smaller overall
plant capacity, more operational flexibility, shorter time cleaning each filter,
although more filters to construct and maintain)

44
 SYSTEM DEMANDS
System demands and operation to consider…
1. 20-year planning horizon
2. Average day demands (ADD) – Design Goal
3. Peak day demands (PDD) – Design Goal
4. Peak hour demands (use storage)
5. Available storage (3 days ADD
recommended)
6. Account for cleaning/ripening (min 2 filter Tank for City of Astoria, OR

beds) – ability to meet PDD with largest

filter off-line.
7. Keep filtration rates below 0.1 gpm/ft2
8. Avoid rapid flow changes (strive for weekly
or monthly changes)
9. Plan for constant flow through filter
(constant supply of nutrients for biological
health)

For planning purposes, system demands should be estimated for a minimum of 20 years
into the future. Filters should be designed with enough surface area to meet peak day
demands with the largest filter out of service without making drastic flow changes. Even
though installations may be small, a minimum of two filters should be installed to allow for
taking one filter off-line for several days during cleaning and ripening. The filter area should
be large enough to meet these demands, while maintaining a filtration rate between 0.1
and 0.16 gpm/ft2. Should storage not be enough to meet peak hour demands, then the
filter area should be expanded to meet these demands as well. Typically this is not needed
when storage is capable of meeting 3 or more days of average demand.

45
TEN STATES STANDARDS
STRUCTURAL DETAILS & HYDRAULICS

4.3.4.3 Structural details and hydraulics

Slow rate gravity filters shall be so designed as to provide:

a. a cover,
b. headroom to permit normal movement by operating
personnel for scraping and sand removal operations,
c. adequate access hatches and access ports for handling of
sand and for ventilation,
d. an overflow at the maximum filter water level, and
e. protection from freezing.

46
 FILTER BOX
Filter boxes and earthen cells
1. Water tight. Filter boxes should be watertight, not merely to
prevent loss of treatment water, but to exclude ingress of
groundwater, which might contaminate the treated effluent.
If possible, ensure the floor is above the highest water table.

2. Allows for cleaning and re-sanding efforts.

3. Insulated from freezing (below ground, covered, or fully

enclosed) .

4. Covered as needed to prevent algae blooms and exclude

falling leaf litter.

5. Freeboard of 4 – 12” (10 – 30 cm) above overflow level.

47
 FILTER WALLS - VERTICAL
Vertical Walls
Short circuiting of the filter-bed along vertical walls can be mitigated by
using a keyway (6x8 cm), rough sloped walls, or a batter.

This diagram shows three mechanisms to prevent short circuiting along the interface with
vertical side walls and the filter sand bed. The WHO manual indicates that the most
effective precaution is to give the walls a slight outward batter, so as to obtain the
advantages of sloping walls and to use grooved or roughened surfaces.

48
 FILTER WALLS - VERTICAL
City of Sumpter (Baker Co)

3 cells
360,000 gpd
completed
In the spring of 2003

This photograph shows how a circular basin can be divided to provide more cells, making
cleaning easier since only 1/3 of the basin needs to be cleaned at a time.

49
 FILTER WALLS - VERTICAL
Vertical Walls
Ramps allow to get equipment and new sand in and old sand out.

Ramps allow to get equipment and new sand in and old sand out of vertical walled filter
boxes. This photograph shows the two of the filters for Falls City, OR.

50
 FILTER WALLS - VERTICAL
Vertical Walls
Removable stop “logs” allow
access to the filters used by the
City of Banks, OR.

Top of stop logs also serve as

scum outlet/overflow.

Removable stop logs allow for cleaning filters used by the City of Banks, OR.

51
 FILTER WALLS –
SLOPED
Design showing sloped walls
with liner

These schematics show the construction of an in-ground filter with a liner and sealed pipe
penetrations.

52
COVERED FILTERS
If covered and/or housed in a filter building, make sure
ample room exists to enable cleaning.

2012. Camp Yamhill in Yamhill County Oregon.

53
 COVERED FILTERS
Protection from freezing
Even if the filters are enclosed, like these “Blue Future” filters,
they may not provide enough protection from freezing
temperatures. These filters were eventually enclosed in a
building.

2009. Jewell School District #8, OR. “Blue Future” covered filter (left) and raw water control tank (right)

This photo shows a covered “Blue Future” filter installed at the Jewell School District #8 in
2009. The green tank on the left is the filter, the taller green tank on the right is the raw
water control tank, and the small grey tank in the middle is the effluent control tank. These
filters are designed to be harrowed rather than scraped. Harrowing will be discussed a little
later.

54
 COVERED FILTERS
Wickiup Water District (Clatsop Co)

Two 80’x30’ cells

120 gpm (0.025 gpm/sf)
Framework allows
for shade cloth to be
used during the
summer

55
 COVERED FILTERS
Thames Water (London England)
Positive air pressure support a plastic film cover at Thames
Water.

56
 SLOW SAND DESIGN
UNCOVERED VS. COVERED
Note the difference in biomass development following scraping
(Campos et. al, 2002/2006) – this should be considered in sizing system

Note the difference in biomass development following scraping (Campos et. al, 2002/2006).
This information can be obtained during a pilot study.

57
 SLOW SAND DESIGN
UNCOVERED VS. COVERED FILTERS
Parameter Uncovered Covered
Temperature More exposed to lower Less susceptible to temperature
temperatures which can effects
adversely impact biological
activity and increase filter
ripening times.
Algae Algal growth/blooms in the Not as susceptible to localized
headwaters can increase algae blooms.
clogging
Biomass Development Filter has a higher biomass and Overall biomass levels are lower
develops a more noticeable and schmutzdecke formation
schmutzdecke. may appear non-existent or
present as an easily suspended,
inert, black carbonaceous deposit
of about 1 mm in thickness.
Biomass is significantly
correlated to bacteria counts.
Removal Efficiency Equivalent May be adversely impacted by
lack of schmutzdecke layer

This table shows some of the differences experienced in filters with and without a cover.
Although there are some important differences, deciding to cover the filters may be
dictated by site constraints (space, temperatures, etc.) and the amount of filter area
needed (is it practical?).

58
TEN STATES STANDARDS
FILTRATION RATE
4.3.4.4 Rates of filtration

The permissible rates of filtration shall be determined by the

quality of the raw water and shall be on the basis of
experimental data derived from the water to be treated. The
nominal rate may be 45 to 150 gallons per day per square foot of
sand area (1.8 - 6.1 m/day), with somewhat higher rates
acceptable when demonstrated to the satisfaction of the
approving authority.

45 – 150 gpd/ft2
(0.031 – 0.10 gpm/ft2)

59
 FILTRATION RATE
(HYDRAULIC LOADING RATE)
Equation for Determine HLR: HLR = Q / (A * (N-1))
Where:

1. HLR = hydraulic loading rate (gpm/ft2)

2. Q = flow needed to meet demands (gpm)
3. A = The sand bed surface area of one filter bed (ft2)
4. N = total number of filter beds needed > 2 (“N-1” is the total
number of filters with 1 filter taken out of service for cleaning)

Example: Given a peak day demand of 250 gallons per capita per
day and a community of 600 people served by two 50’x20’ filters:
250 gpcpd x 600 people
(1,000 ft2/filter x (2 filters – 1 filter) = 150 gpd/ft2 (0.1 gpm/ft2)

This equation is used to determine the filtration rate (or hydraulic loading rate) of slow sand
filters.

60
FILTRATION RATE
Maximum
< 0.1 gpm/ft2

Rate may need

to be < 0.05
gpm/ft2 when
water temp <5
°C
Minimum
> 0.02 gpm/ft2
to keep biota
viable

61
TEN STATES STANDARDS
UNDERDRAINS
4.3.4.5 underdrains

Each filter unit shall be equipped with a main drain and an adequate
number of lateral underdrains to collect the filtered water.
The underdrains shall be
placed as close to the floor as
possible and spaced so that
the maximum velocity of the
water flow in the underdrain
will not exceed 0.75 feet per
second.

The maximum spacing of

laterals shall not exceed 3 feet
if pipe laterals are used.

62
 UNDERDRAINS
underdrains are
typically made using
perforated pipe
laterals (PVC, NSF-61)
due to minimal head
loss.

For larger installations,

laterals are typically 4-
8” in diameter, while
main drains are 12-18”
in diameter.

underdrains are typically constructed of PVC (NSF-61), which has minimal head loss.

63
 UNDERDRAIN CONFIGURATIONS
Common configurations include laterals that connect to a main
drain system. Smaller filters will often have only 1 main drain
like the one shown on the right.

Common configurations include laterals that connect to a main drain system. Smaller filters
will often have only 1 main drain like the one shown on the right.

64
 UNDERDRAINS
City of Cannon Beach

12” PVC header

4” laterals
48” lateral spacing
8” hole spacing

Access Ramp
for Cleaning/
Resanding

“Hydraulic Control
Manhole”

Here is a plan view of the underdrains for the City of Cannon Beach, Oregon.

65
 UNDERDRAINS
City of Astoria (5 MGD)

Filter Cell #2
1993 filter rebuild
15” PVC manifold
14” PVC header
6” laterals
60” lateral spacing
6” hole spacing
¼” drain holes

This shows a detail of the lateral pipe perforations of the underdrain system for filter cell #2
for the City of Astoria, Oregon.

66
 UNDERDRAINS
Velocity in laterals and main
drain should not exceed
0.75 fps (0.23 m/sec.)

Velocity in the laterals and main should not exceed 0.75 fps (0.23 m/sec). This diagram
shows the configuration of the main drain pipe and laterals. Note the spacing of laterals at
3 – 5 feet apart (1-2 meters). Drain holes should be 5/64” – 5/32” in diameter (2-4 mm)
and spaced every 4 to 12 inches apart (0.1 – 0.3 meters).

67
 UNDERDRAINS – LATERAL SPACING
Streamlines forming a
streamtube

Closer
lateral
spacing
leads to a
more even
distribution
of headloss
and more
consistent
filtration
rates across
the filter

This diagram illustrates how the spacing of laterals can impact the flow of water through
the underdrains (indicated by streamlines) and the resulting increase in headloss and
decrease in filtration rate with larger lateral spacing.

68
 UNDERDRAINS
City of Astoria (5 MGD)

Filter Cell #2
1993 filter rebuild
Keep laterals spaced away
from filter walls to avoid
short-circuiting of raw
water down the filter wall

Keep laterals spaced away from filter walls to help prevent sidewall effects where unfiltered
water can slip past the filter media down the sidewall.

69
 UNDERDRAINS

Underdrain Design Parameters Recommended Specification

Maximum Velocity in Laterals1 0.75 fps (0.23 m/sec)

Maximum Velocity in Main Drain1 0.75 fps (0.23 m/sec)

Spacing of lateral drain pipes1 36 inches (91.4 cm)

Spacing of bottom lateral drain 4 – 12 inches (0.1 – 0.3 m)

holes2 (include air release holes @ ends on top of laterals)

Diameter of drain holes2 5/64” – 5/32” (2-4 mm)

(needs to be determined through hydraulic calculations)

Material Non-Corrosive and meeting NSF-61 (e.g., PVC)

1 Source: 2012 Edition of the Recommended Standards for Water Works (Ten States Standards). Visscher et., al. (see footnote 2)

recommended 1.64 fps (0.5 m/sec).

2Source: Visscher, J.T., R. Paramasivam, A. Raman, and H.A. Heijnen. 1987. Slow Sand Filtration for Community Water Supply, Planning,

Design, Construction, Operation and Maintenance. Technical Paper No. 24, The Hague, Netherlands: International Reference Center for
Community Water Supply and Sanitation.

Design recommendations from the IRC manual are included here. One addition is the
provisions for air release holes at the ends on top of the laterals to purge air pockets in the
laterals upon initial filling.

70
TEN STATES STANDARDS
FILTER MEDIA
4.3.4.6 Filter material

a. Filter sand shall be placed on graded gravel layers for a minimum depth of 30
inches.

b. The effective size shall be between 0.15 mm and 0.30 mm. Larger sizes may
be considered by the reviewing authority; a pilot study may be required.

c. The uniformity coefficient shall not exceed 2.5.

d. The sand shall be cleaned and washed free from foreign matter.

e. The sand shall be rebedded when scraping has reduced the bed depth to no
less than 19 inches. Where sand is to be reused in order to provide biological
seeding and shortening of the ripening process, rebedding shall utilize a “throw
over” technique whereby new sand is placed on the support gravel and existing
sand is replaced on top of the new sand.

71
 FILTER MEDIA
Plan the work and provide an adequate budget so recommended media
specifications and placement practices are able to be followed – it will pay off
in the long run!

Plan the work and provide an adequate budget for media – it will pay off in the long run!

72
 SAND BED DEPTH RECOMMENDATIONS
MINIMUM SAND DEPTH GUIDELINE
20 – 24 INCHES
45
39
40
35
35
Inches 30
30 28
25 24 24 24 24
20 20 20 19 20 20 20
20
16
15

10

Various sources all fall within 16 – 35 inches, however, most recognize 20 – 24 inches as a
minimum level the sand bed ought to be allowed to operate with.

73
 Freeboard Air above water
SAND BED DEPTH (4 – 12”)
Headwater Sedimentation
According to the WHO manual, (39-59”)
biochemical and adsorption removal
mechanisms are in effect immediately
below the schmutzdecke down to a
depth of around 24-inches.
Schmutzdecke Biological
(1-2 cm)
Therefore the total bed thickness
Filter Sand Biochemical
would need to be at least 24 inches in (12-16”)
order for these two mechanisms to be
fully effective. Adsorption (8”)

Additional sand is needed to Sand allowance

accommodate the amount of sand for cleanings
anticipated to be removed due to Sand Support Support Gravel
cleanings over the design life. (15-24”)

As covered in the discussion on removal mechanisms, a certain amount of sand ranging

from 20 – 24” is needed to ensure that the removal mechanisms remain available for the
entire life of the filter bed. In order to ensure this, an added sand allowance is needed to
account for successive cleanings over the life of the filter.

74
 SAND BED DEPTH - FORMULA
Formula for determining depth of sand:

Di = [Y (R * fscraping)] + Df

Re-arrange to find design life: Y = (Di – Df) / (R * fscraping)

Where:
Y = years of operation before sand bed needs rebuilding
Di = initial sand bed depth (inches)
Df = final sand bed depth before rebuilding (inches)
R = sand depth removal per scraping (inches/scraping)
fscraping = frequency of scraping (scrapings/year)

This is used to determine the additional sand allowance needed to account for successive
cleanings over the life of a filter.

75
 SAND BED DEPTH - EXAMPLE
Example: Di = [Y (R * fscraping)] + Df

Given:
1. Di = initial sand bed depth (inches)
2. Df = 24 inches
3. fscraping = 6 cleanings per year
4. R = Removal of 1.3 cm (1/2”) of sand per cleaning
5. Y = 7-year design life (before re-sanding is needed)

Di = 7 yrs * (0.5 in/scraping * 6 scrapings/year)] + 24 in

= 45 inches
Therefore, an additional 21 inches (53 cm) of sand is
needed to allow for scraping over 7 years.

This example shows how one would use the same formula to determine the additional
sand allowance needed to account for successive cleanings over the life of a filter.

76
 FILTER MEDIA - SILICA
Silica sand
Durable
Inexpensive
Readily available

• The most important feature is the pore space in the media.

• Removal mechanisms occur in the pores where suspended

solids are trapped, microorganisms grow, and air and water
flow.

• Using media with an appropriate effective size and uniformity

ensures an optimal pore space.
•

Media selection is critical to proper operation and should be primarily silica sand, due to it’s
durability and availability.

77
FILTER MEDIA – GRAIN SIZE
There are many different
“grades” of sand available

“Sand” ranges from 0.0625 mm (#230 sieve) – 2.0 mm (#10 sieve)

78
 FILTER MEDIA - SIEVE SIZES

Sieves larger than the

#4 sieve are designated
by the size of the
openings in the sieve.

Smaller sieves are numbered according to the

number of openings per inch.

Example: A #10 sieve is made with 0.0237” thick

wire. What is the opening size?
9 wires/in x 0.0237”/wire = 0.213”
1” – 0.213” = 0.787”
Opening width = 0.787”/10 openings = 0.0787”
0.0787” = 2.0 mm

Sieves are numbered in one of two ways. Sieves larger than the #4 sieve are designated by
the size of the openings in the sieve. Smaller sieves are numbered according to the number
of openings per inch. The thickness of wire must be accounted for in determining the
opening. For example, a #10 sieve has 10 openings per inch. The openings are only 0.0787
inches (2.0 mm) because the sieve is made using 0.0237” thick wire. Since there are 9
wires within an inch to make up 10 openings per inch, the wire accounts for the remaining
0.213” inches.

79
 FILTER MEDIA - EFFECTIVE SIZE (D10)

Effective Size (D10)

Range should be 0.2 mm to 0.35 mm

• The effective size gives a good

indication of the permeability characteristics of sand.

• D10 is the size of grain such that 10% by weight of the total
sample is smaller. (i.e., 10% of the sand, by weight, is finer
than a given grain size).

D10 = 0.2 mm 90% #70 Sieve (70 openings/inch = 0.212 mm opening)

(10% passes
through to the pan) 10% (by weight) of this sample
is smaller than 0.2 mm
10% Pan

One critical sand specification is the effective size. The effective size (or diameter) is
typically expressed as D10 and indicates the grain diameter in millimeters at which 10% of
the total grains of a given sample are smaller and 90% of the total grains are larger, based
on weight. The effective size for slow sand media should be between 0.2 mm to 0.35 mm.
Ten States Standards recommends a range of 0.15 – 0.35 mm, but the 0.15 mm
specification often only restricts production without any added benefit.

80
 FILTER – DETERMINING D10
Effective Size (D10)
Range should be 0.2 mm to 0.35 mm

A sieve analysis is done to determine D10 by:

1. Passing a known amount of media through a
series of progressively smaller sieve sizes; and

2. Weighing the amount of media retained on each

sieve.

The effective size is determined by doing a sieve analysis. So for example, you take a
measured quantity of sand and weigh it. Then you sift it through a series of sieves with
progressively smaller screens and weigh the portion of sand retained on each sieve.

81
 FILTER MEDIA - Percentage retained on any sieve:
DETERMINING D10 = 100% x (weight of soil retained / total soil weight)
Cumulative percentage retained on any sieve:
= Σ percentage retained
Percentage passing the sieve:
= 100% - Σ percentage retained

Sieve # Diameter Mass of soil retained Percent Cumulative Percent

(mm) on each sieve retained Retained Passing
(g) (%) (%) (%)
20 0.850 5 1.00% 0.00% 100%
30 0.600 27.5 5.50% 5.50% 95%
40 0.425 85 17.00% 22.50% 78%
50 0.300 125 25.00% 47.50% 53%
70 0.212 128 25.50% 73.00% 27%
100 0.150 77.5 15.50% 88.50% 12%
140 0.106 40 8.00% 96.50% 4%
200 0.075 10 2.00% 98.50% 2%
Pan N/A 2.5 0.50% 99.00% 1%
Total => 500 gram sample

Then you determine the % of the sample, by weight, that passes (is finer than) each
successive sieve. This example shows how data from a sieve analysis is tabulated.

82
 FILTER MEDIA – DETERMINING D10
Effective Size (D10)
Range should be 0.2 mm to 0.35 mm
• The results are plotted (% passing vs. sieve/grain
size (mm)).
• D10 is where a horizontal line drawn from the 10%
passing mark intersects the grain size.

The tabulated results are plotted on a graph of % passing versus sieve (i.e., grain size). D10
is where a horizontal line drawn from the 10% passing mark intersects the sieve size – that
sieve size intersected is the D10.

83
 FILTER MEDIA - UC
Uniformity Coefficient (UC)
Range should be 1.5 – 3.0

• D60 is the size of grain such that 60% by weight of the total
sample is smaller. (i.e., 60% of the sand, by weight, is finer
than a given grain size).

D60 = 0.3 mm 30% #50 Sieve (50 openings/inch = 0.3 mm opening)

(50% on #100 sieve
plus 10% in pan)
D10 = 0.15 mm 50% #100 Sieve (100 openings/inch = 0.15 mm opening)
(10% passes
60% (by weight) of this sample is smaller
through to the pan)
than 0.3 mm (50% + 10%)
10% Pan
D60 = 0.3 mm
10% is smaller than 0.15 mm = d10
UC = D60/D10 = 0.3 mm/0.15 mm = 2.0

The uniformity coefficient, denoted by “U” or “Cu”, is determined by dividing D60 by D10.
Similar to D10, D60 is the sieve size through which 60% of the sample by weight passes and
40% is retained. You will not usually see D60 referenced, however, it is used to determined
the uniformity coefficient, which is an important specification for slow sand filter sand.

84
 FILTER MEDIA - UC
Uniformity Coefficient (UC) = 1.5 – 3.0

If the grain sizes vary greatly, the

smaller ones will fill the spaces between
the larger particles, making it easier for
the filter to clog.

The uniformity coefficient is related to the distribution of grain sizes of soils. Uniformly
graded soils have soil grains that are mostly the same size. This keeps the pore spaces
between the grains open. Well graded sand has a broader size distribution (higher
uniformity coefficient), which results in the fines filling up the pore spaces of the media
resulting in less space for biological removal mechanisms to work and higher head loss.

85
 FILTER MEDIA – DETERMINING UC

Quiz:
What is D10, D60 and UC for this sieve analysis?
Sieve Diameter Mass of soil Percent retained Cumulative Percent
# (mm) retained on (%) Retained Passing
each sieve (%) (%)
(g)

20 0.850 0 0.00% 0.00% 100%

30 0.600 30 30% 30% 70%

50 0.300 10 10% 40% 60%

70 0.212 40 40% 80% 20%

100 0.150 10 10% 90% 10%

200 0.075 5 5% 95% 5%

Pan N/A 5 5% 100% 0%

Quiz – What is d10, d60 and UC for this sieve analysis?

86
 FILTER MEDIA – DETERMINING UC Cumulative percentage retained on any sieve:
D10 = 0.15 mm = Σ percentage retained
D60 = 0.3 mm Percentage passing the sieve:
UC = d60/d10 = 0.3/0.15 = 2.0 = 100% - Σ percentage retained
Sieve Diameter Mass of soil Percent retained Cumulative Percent
# (mm) retained on (%) Retained Passing
each sieve (%) (%)
(g)

20 0.850 0 0.00% 0.00% 100%

30 0.600 30 30% 30% 70%

50 0.300 10 10% 40% 60% d60

70 0.212 40 40% 80% 20%

100 0.150 10 10% 90% 10% d10

200 0.075 5 5% 95% 5%

Pan N/A 5 5% 100% 0%

Answer: D10 = 0.15 mm, D60 = 0.3 mm and UC = 0.3/0.15 = 2.0. This example illustrates how
the mass retained on each sieve relates to the determination of d10, d60 and UC.

87
 FILTER MEDIA – DETERMINING UC
D10, D60 and UC can also be determined
graphically

This slide shows how to identify D10, D60, and UC graphically.

88
 FILTER MEDIA http://www.slowsandfilter.org

Recommended effective diameter (D10) – 0.2 mm to 0.35 mm

Shown here are sand sizes from 0.15 – 0.35 mm. Photo from www.slowsandfilter.org.

89
 Sieve Analysis – good sand

This chart illustrates the results of a sieve analysis for sand primarily within the
recommended specifications.

90
 Sieve Analysis – not so good sand

This chart illustrates the results of a sieve analysis for sand with an effective diameter larger
than the recommended specification, although the uniformity coefficient is just within the
specifications.

91
 OTHER MEDIA CONSIDERATIONS
1. % of fines passing #200 sieve
< 0.3% by weight
2. Acid solubility
< 5%
3. Apparent Specific Gravity
> 2.55
4. Minimum depth
20-24 inches before re-sanding
5. Availability
• Local supply options (keep transport costs low)
• Redundant/backup supply (e.g. 2 or more quarries)
• Ability to meet specifications
• Consider ability to clean/stockpile scraped media
6. NSF-61 or equivalent (tested for contaminants)

% of fines passing the #200 sieve should not be more than 0.03% by weight. The more
fines, the longer it will take for turbidity in a newly sanded filter to clear. Acid solubility
should be less than 5%. Media with more acid soluble content, think of limestone, could
eventually end up with an undesirable effective diameter or uniformity coefficient in the
presence of acidic waters. Sand beds are typically 30 to 36 inches in depth and should not
be allowed to drop below 20-24 inches. Some references indicate lower levels may still
provide adequate filtration, but as discussed earlier, this may inhibit removal mechanisms
that occur deeper in the sand bed. Another thing to consider is where the sand is going to
come from. Finding one or more local sources keeps transportation costs low. Not all
quarries can provide sand meeting the desired specifications. Some systems have made
provisions for cleaning and stockpiling sand that has been removed during the scraping
process. This sand should then be analyzed for conformance with the desired
characteristics and can then be re-used in subsequent re-sanding efforts.

92
 Sources of Sand

1. CEMEX, Vancouver, WA & Boardman, OR

2. Kleen Industrial Services, Danville, CA
3. Knife River Corporation, Corvallis, OR & Stayton, OR
4. Naselle Rock and Asphalt, Naselle, WA
5. Fazio Brothers, Vancouver, WA

Others???

Some suppliers of slow sand filtration sand.

93
 OTHER MEDIA CONSIDERATIONS
Monitoring Media Depth

Incorporate a means of monitoring media depth

• Keyway (also mitigates sidewall effects)
• Staff gage

2x4
Keyway

Design should incorporate a way to measure sand bed depth. A keyway can serve to both
indicate the minimum sand depth (when the top of the keyway is reached), while
interrupting flow down the sidewall.

94
 OTHER MEDIA CONSIDERATIONS
Monitoring Media Depth

Incorporate a means of monitoring media depth

“Marks on Liner Side”

This construction plan shows “marks” at 18” sand depth on the liner of a sloped filter wall
(cell #3) for the City of Astoria (1993)

95
 RECOMMENDED MEDIA SPECS
Media specifications (silica sand) - summary
Filter Sand Specification Recommended Range

Effective Diameter (d10) 0.2 – 0.35 mm

Uniformity Coefficient (U) 1.5 – 3.0

% fines passing #200 sieve < 0.3% by Wt.

Acid Solubility < 5%

Apparent Specific Gravity > 2.55

Minimum Depth 20-24 inches

Delivery/Installation Sand washed prior to installation

NSF/ANSI Standard 61 Certified or equivalent

This table summarizes some of the main specifications for sand and bed depth.

96
SUPPORT GRAVEL
Support gravel prevents migration of sand down to
underdrains, while allowing passage of filtered
water.

Proper gradation is key to prevent migration

Rounded rock is used to promote drainage

Example shown* is for a rapid rate plant

(City of Grants Pass)

Top Layer 1 Filter sand

(Silica sand w/ D10 = 0.45 mm- 0.55 mm)
Layer 2 #50 garnet sand
Layer 3 #12 garnet gravel
Layer 4 3/8” x 3/16” gravel
Layer 5 3/4” x 3/8” gravel
Bottom layer 1-1/2” x ¾” gravel

*Anthracite is on top of filter sand, but is not shown.

97
TEN STATES STANDARDS
SUPPORT GRAVEL
4.3.4.7 Filter gravel

The supporting gravel should be similar to the size and depth

distribution provided for rapid rate gravity filters. See
4.2.1.6.f.2. (e.g. 4.3.1.6.e. – Support Media (for rapid rate
gravity filters))

98
TEN STATES STANDARDS
SUPPORT GRAVEL, CONT.

4.3.1.6.e. – Support Media (for rapid rate gravity filters)

4.3.1.6.e.1. Torpedo sand (often used to backfill utility pipes)

A three-inch layer of torpedo sand shall be used as a supporting
media for filter sand where supporting gravel is used, and shall
have:

a. effective size of 0.8 mm

to 2.0 mm (1/32” – 5/64”)

b. uniformity coefficient
not greater than 1.7.

99
TEN STATES STANDARDS
SUPPORT GRAVEL, CONT.
4.3.1.6.e.2. Gravel - Gravel, when used as the supporting media shall consist of
cleaned and washed, hard, durable, rounded silica particles and shall not include
flat or elongated particles. The coarsest gravel shall be 2.5 inches in size when the
gravel rests directly on a lateral system, and must extend above the top of the
perforated laterals. Not less than four layers of gravel shall be provided in
accordance with the following size and depth distribution:

Size Depth

3/32 to 3/16 inches 2 to 3 inches

3/16 to 1/2 inches 2 to 3 inches
1/2 to 3/4 inches 3 to 5 inches
3/4 to 1 ½ inches 3 to 5 inches
1 ½ to 2 ½ inches 5 to 8 inches

Reduction of gravel depths and other size gradations may be considered upon
justification to the reviewing authority for slow sand filtration or when proprietary
filter bottoms are specified.

100
 SAND & GRAVEL
ASTM E11 standard sizes for woven wire test sieve cloth
Mesh Mesh Mesh Mesh
No. Size No. Size No. Size No. Size
(mm) (mm) (mm) (mm)
1” 25.0 7 2.80 20 0.85 60 0.250
¾” 19.0 8 2.36 25 0.71 80 0.180
½” 12.5 10 2.00 30 0.60 100 0.150
3/8” 9.5 12 1.70 35 0.50 120 0.125
4 4.75 14 1.40 40 0.425 140 0.106
5 4.00 16 1.18 45 0.355 170 0.090
6 3.35 18 1.00 50 0.300 200 0.075

There are standard mesh sizes for grading sand and gravel. Some of these mesh sizes likely
to be used in slow sand filters identified under ASTM E11 are shown here for reference.

101
 SUPPORT SAND
Concrete Sand
(a.k.a. “Torpedo Sand”)
• Fine Aggregate grade 1 or 2 (FA-1 or 2)
• Angular to sub-angular
• 3/8” x #100 mesh (9.5 x 0.15 mm)
• Washed and screened
• Used in production of ready mixed concrete
• Commonly used for pipe bedding & Backfill
• Meets ASTM C33 standard
Torpedo Sand

Photos of gravel from Phil’s Topsoil, Inc.,

http://www.philstopsoil.com/stone_and_sand.html

102
 SUPPORT
GRAVEL
Pea Gravel (CA-16) Pea Gravel
• Round to sub-angular river rock
• 3/8” x #16 (9.5 x 1.18 mm)
(1.18 mm ~ 3/64”)
• Washed but may contain smaller sediment
• Commonly used for decorative
landscaping,
• retaining wall backfill, or where finer
aggregates are needed when drainage
through stone is
• desired.

#8 Gravel (a.k.a. #8 pea gravel)

3/8” x #8 (9.5 x 2.36 mm) (2.36 mm ~ 3/32”)
#8 Pea Gravel

Top photo of gravel from Phil’s Topsoil, Inc.,

http://www.philstopsoil.com/stone_and_sand.html
Bottom photo of gravel from The Gravel Guy – Don Brown. http://thegravelguy.com/gravel/

103
 FILTER MEDIA

#57 gravel
1” x #4 (25.0 x 4.75 mm) (4.75 mm ~ 3/16”)

#4 gravel
1-1/2” x 3/4” (37.5 x 19 mm)

#4 Gravel #57 Gravel

Photos of gravel from The Gravel Guy – Don Brown. http://thegravelguy.com/gravel/

104
 FILTER MEDIA

3/4” washed Gravel (CA11) 3/4” River Rock

1-1/2” Washed Gravel (CA5) 1-1/2” River Rock

Photos of gravel from Phil’s Topsoil, Inc.,

http://www.philstopsoil.com/stone_and_sand.html

105
SUPPORT MEDIA
City of Salem, OR

Filter sand D10 = 0.27 – 0.33 mm

4” Top Layer #4 x #8 sieve
4” Middle Layer 1/2” x #4 sieve (3/16”)
10” Bottom Layer 7/8” x 1/2”

106
SUPPORT MEDIA
3 layers of support gravel can be adequate, but 4 or more layers is
recommended due to product and placement uncertainties.
D10 D90 Depth
Considerations
Layer (mm) (mm) (inches)
Durability
Top Layer 3/64” 1/16” 6
Cost
(1.0 mm) (1.4 mm)
Availability
Middle Layer 5/32” 7/32” 6
(4.0 mm) (5.6 mm)
Bottom Layer 5/8” 29/32” 6
(16 mm) (23 mm)
Each successive layer should be graded so that its smaller (D10)
particle diameters are not more than four times smaller than
those of the layer immediately below.
The grains of the bottom layer should have an effective
diameter of at least twice the size of the drain holes or slots.
* The gravel support using three layers as specified will work if the orifices into the under
drain pipe are less than 8 mm in diameter. If the orifices are larger, more than three layers
of gravel may be needed.

107
SUPPORT MEDIA Discard
Example of a 4 layer support media for use with 1/16”
a filter sand with an effective size (D10) of 0.2 2nd Layer (1/8” x 1/16”)
mm and 1/8” drain holes. 1/8”
3rd Layer (1/4” x 1/8”)
Discard
1/4”
0.8 mm 4th Layer (1/2” x 1/4”)
Top Layer (0.8-1.2 mm)
1/2”
1.2 mm
Discard
Discard

Support media for use with a filter sand with an effective size (d10) of 0.2 – 0.35 mm and 1/8” underdrain holes (for
¼” underdrain holes a fifth layer of ¾” x ½” or 1” x ½” gravel is needed).
Depth
Layer Size Range Specification
(inches)
Largest Size/Smallest Size = 1.5
1/32” 1/21”
Top Layer 3 Smallest Size in Top Layer /D10 Filter
(0.8 mm) (1.2 mm)
Sand = 4.0
Largest Size/Smallest Size = 2.00
2nd Layer 1/16” 1/8”
3 Largest Size/Smallest Size of Top Layer =
1/8” x 1/16” (1.588 mm) (3.175 mm)
3.97

Largest Size/Smallest Size = 2.00

3rd Layer 1/8” 1/4”
3 Largest Size/Smallest Size of 2nd Layer =
1/4” x 1/8” (3.175 mm) (6.35 mm)
4.00
Largest Size/Smallest Size = 2.00
Embed and bury
4th Layer 1/4” 1/2” Largest Size/Smallest Size of 3rd Layer =
underdrain piping
1/2” x 1/4” (6.35 mm) (12.7 mm) 4.00
with 1” of cover
Smallest Size = 2x Drain Diameter

108
SUPPORT MEDIA
4-5 Layer Option
Filter sand D10 = 0.2 mm
3” Top Layer #20 sand
3” Second Layer 1/8” x 1/16“
3” Third Layer 1/4” x 1/8”
3” Fourth Layer ½” x ¼” (4 layers work with 1/8” drain holes)
Bottom Layer ¾” x ½” (5 layers are needed with ¼” drain holes)

AWWA WHO (Huisman

AWWA
B100-09 & Wood)
AWWA AWWA B100-09
d10 top d10 lower/d10
B100-09 B100-09 d90/d10
d10 gravel upper (<3 where
4-5 Layer Option d90 (mm) d10 (in) d90 (in) & WHO d90 same layer
(mm) between d90/d10 of same
d10 > lower/d10 (<2)
4 and 4.5 layer < 2 or <
2xDrain? upper (<4) (< 1.4,
times d10 4 if d90/d10 of
WHO)
sand same layer < 1.4)
Sand 0.2 N/A N/A N/A N/A N/A N/A
Top Gravel 0.8 1.2 1/32 1/21 No 4.00 4.00 N/A 1.50
2nd Gravel 1.5875 3.175 1/16 1/8 No N/A 1.98 3.97 2.00
3rd Gravel 3.175 6.35 1/8 1/4 No N/A 2.00 4.00 2.00
4th Gravel 6.35 12.7 1/4 1/2 Yes N/A 2.00 4.00 2.00
5th Gravel 12.7 19.05 1/2 3/4 Yes N/A 2.00 3.00 1.50
Drain 3.175 N/A 1/8 N/A N/A N/A N/A N/A

109
 SUPPORT MEDIA
Rapid Rate Filter (City of Grants Pass)

Top Layer 1 (#50 garnet sand w/ D10 = 0.25 mm)

Layer 2 #12 garnet gravel
Layer 3 3/8” x 3/16” gravel
Layer 4 3/4” x 3/8” gravel
Bottom layer 1-1/2” x ¾” gravel

AWWA WHO (Huisman

AWWA
B100-09 & Wood)
AWWA AWWA B100-09
d10 top d10 lower/d10
Grants Pass using B100-09 B100-09 d90/d10
d10 gravel upper (<3 where
Garnet Sand as d90 (mm) d10 (in) d90 (in) & WHO d90 same layer
(mm) between d90/d10 of same
Filter Medium d10 > lower/d10 (<2)
4 and 4.5 layer < 2 or <
2xDrain? upper (<4) (< 1.4,
times d10 4 if d90/d10 of
WHO)
sand same layer < 1.4)
Sand 0.25 N/A N/A N/A N/A N/A N/A
Top Gravel 1.18 1.7 2/43 1/15 No 4.72 4.72 N/A 1.44
2nd Gravel 4.7625 9.5 3/16 3/8 No N/A 4.04 8.05 1.99
3rd Gravel 9.5 19.05 3/8 3/4 No N/A 1.99 4.00 2.01
4th Gravel 19.05 38.1 3/4 1 1/2 Yes N/A 2.01 4.01 2.00
5th Gravel 0 0 No N/A 0.00 0.00 #DIV/0!
Drain 6.35 N/A 1/4 N/A N/A N/A N/A N/A

110
SUPPORT MEDIA
Opal Creek, OR (Blue Future Filters)

Filter sand D10 = 0.3 mm

2” Top Layer 3/8” x 3/16”
2” Middle Layer 1” x 7/8“ (or 1” x 1/2”)
8” Bottom Layer 1-1/2” x 3/4”
Drain orifice = ¼” slots

AWWA WHO (Huisman

AWWA
B100-09 & Wood)
AWWA AWWA B100-09
Opal Creek, d10 top d10 lower/d10
B100-09 B100-09 d90/d10
Oregon d10 gravel upper (<3 where
d90 (mm) d10 (in) d90 (in) & WHO d90 same layer
(Blue Future (mm) between d90/d10 of same
d10 > lower/d10 (<2)
Filters) 4 and 4.5 layer < 2 or <
2xDrain? upper (<4) (< 1.4,
times d10 4 if d90/d10 of
WHO)
sand same layer < 1.4)
Sand 0.35 N/A N/A N/A N/A N/A N/A
Top Gravel 3.175 9.525 1/8 3/8 No 9.071429 9.07 N/A 3.00
2nd Gravel 22.25 25.4 7/8 1 Yes N/A 7.01 8.00 1.14
3rd Gravel 19.05 38.1 3/4 1 1/2 Yes N/A 0.86 1.71 2.00
4th Gravel 0 0 No N/A 0.00 0.00 #DIV/0!
5th Gravel 0 0 No N/A #DIV/0! #DIV/0! #DIV/0!
Drain 6.35 N/A 1/4 N/A N/A N/A N/A N/A

111
SUPPORT MEDIA
City of Salem, OR

Filter sand D10 = 0.27 – 0.33 mm

4” Top Layer #4 x #8 sieve
4” Middle Layer 1/2” x #4 sieve (3/16”)
10” Bottom Layer 7/8” x 1/2”

AWWA WHO (Huisman

AWWA
B100-09 & Wood)
AWWA AWWA B100-09
d10 top d10 lower/d10
B100-09 B100-09 d90/d10
City of Salem, d10 gravel upper (<3 where
d90 (mm) d10 (in) d90 (in) & WHO d90 same layer
Oregon (mm) between d90/d10 of same
d10 > lower/d10 (<2)
4 and 4.5 layer < 2 or <
2xDrain? upper (<4) (< 1.4,
times d10 4 if d90/d10 of
WHO)
sand same layer < 1.4)
Sand 0.27 N/A N/A N/A N/A N/A N/A
Top Gravel 2.36 4.75 4/43 3/16 No 8.740741 8.74 N/A 2.01
2nd Gravel 4.75 12.7 3/16 1/2 No N/A 2.01 5.38 2.67
3rd Gravel 12.7 22.25 1/2 7/8 Yes N/A 2.67 4.68 1.75
4th Gravel 0 0 No N/A 0.00 0.00 #DIV/0!
5th Gravel 0 0 No N/A #DIV/0! #DIV/0! #DIV/0!
Drain 6 N/A 17/72 N/A N/A N/A N/A N/A

112
SUPPORT MEDIA
More layers of support gravel is often needed,
due to product cost and availability.

For material size and layer depth, follow the latest:

1. Guidelines in Appendix D of ANSI/AWWA

B100 Standard; or

2. Ten States Standards

for slow sand filter
construction

113
SUPPORT MEDIA
Support Media Installation

1. For material washing/handling/delivery/installation

recommendations, follow the ANSI/AWWA B100 Standard

3. Desired layer elevations should be marked on filter wall

and each layer added and screeded level and even with the
mark.

4. The elevation of the top surface of each layer shall be

checked using water that is introduced into the filter as a
guide with the media within + 0.5 inch of the desired level
and the areas above and below the desired level within
10% of each other.

5. Support gravel should washed prior to placement of filter

sand (see AWWA B100 Standard).

114
 FILTER FABRIC?
City of Astoria
Filter Cell #2
(1993 filter rebuild)

(Note: CH2MHill did not install fabric – only removed it as part

Filter fabric
of rebuild)
between sand
and gravel
often gets
clogged as in
this case. This
fabric was
discarded and
is not needed
with proper
gravel
gradations

A filter fabric layer originally installed for filter cell #2 for the City of Astoria, Oregon was
later removed in 1993 due to clogging.

115
TEN STATES STANDARDS
SUPERNATANT WATER (HEADWATER)
4.3.4.8 Depth of water
on filter beds

Design shall provide a

depth of at least 3 – 6
feet of water above the
sand. Influent water shall
not scour the sand
surface.

Wickiup Water District, OR

116
HEADWATER
• Purpose is to provide driving head
• Provides retention/settling
• Little benefit to exceeding a depth of 4-5 feet
• Shallow levels may increase algae due to sunlight penetration
• Important to include:
1. Side stream influent piping if harrowing
2. Overflow
3. Drain Headwater Sedimentation
4 – 5 ft
4. Backflush piping (48-60”)

Schmutzdecke Biological
(1-2 cm)

117
 INFLUENT ENERGY DISSIPATION
“Influent water shall not scour the sand surface”
Energy Dissipation (to avoid sand scouring)

Energy dissipation
helps to keep
sand from
scouring and
erosion of the
filter cell liner.

2011. City of Corbett, OR

This is filter influent piping for the City of Corbett, OR.

118
 INFLUENT ENERGY DISSIPATION
Jewell School District #8 (Clatsop Co)

3 cells (Blue Future)

18 gpm (0.1 gpm/sf)
Completed in 2010
Cleaned using wet harrowing

Even in smaller covered filters, influent energy should be minimized.

119
 INFLUENT ENERGY DISSIPATION
Energy Dissipation (“Splash Plate”) for Jewell SD #8

2012. Jewell School District #8, OR

This picture shows a splash plate under the inlet piping for an enclosed slow sand filter for
the Jewell School District #8 in Clatsop County, Oregon.

120
 INFLUENT ENERGY DISSIPATION
Sections showing influent baffles

Overflow

M
Filter Inlet

Min sand
level
Gravel
Drain

M
Filter Inlet

These designs allow influent water to enter a filter bed without the risk of scouring the
sand bed.

121
 INFLUENT ENERGY
DISSIPATION

City of Salem, Oregon

influent baffles

These designs allow influent water to enter a filter bed without the risk of scouring the
sand bed.

122
TEN STATES STANDARDS
MONITORING & INFLUENT CONTROLS
4.3.4.9 Control appurtenances

Each filter shall be equipped with:

a. Influent and effluent sampling taps;
b. An indicating loss of head gauge or other means to measure
head loss;
c. An indicating rate-of-flow meter. A modified rate controller that
limits the rate of filtration to a maximum rate may be used.
However, equipment that simply maintains a constant water
level on the filters is not acceptable, unless the rate of flow onto
the filter is properly controlled. A pump or flow meter in each
filter effluent line may be used as the limiting device for the rate
of filtration only after consultation with the reviewing authority.

123
 HEADLOSS MEASUREMENT
Piezometers
Measurement of head loss can be accomplished with simple
piezometers mounted outside the filter and tailwater
structures.

hL (schmutzdecke)
hL (schmutzdecke + sand/gravel) hL (sand bed/gravel) hL (total headloss)
hL (drain orifice) hL (underdrain)
Schmutzdecke
hL (drain pipe)
HGL
Filter Sand

<=Piezometer Tailwater
4 to 5-layer Support Gravel

Underdrain

Screened probes at the top and bottom of the filter sand can allow easy measurement of
head loss with simple piezometers mounted outside the filter. This diagram illustrates how
headloss is greatest in the schumtzdecke and top few centimeters of sand towards the end
of the filter run. This diagram also shows how design ensures an even distribution of flow
by having a much higher headloss through the drain pipe orifices compared to the
underdrain piping.

124
TEN STATES STANDARDS
EFFLUENT CONTROLS

4.3.4.9 Control appurtenances, continued

Each filter shall be equipped with:

d. Provisions for filtering to waste with appropriate measures
for cross connection control;
e. An orifice, Venturi, or other suitable means of discharge
measurement installed on each filter to control the rate of
filtration.
f. An effluent pipe designed to maintain the water level above
the top of the filter sand.

125
KEY FLOW CONTROL ELEMENTS
Key flow control elements:
1. Effluent weir or controls to prevent air entrainment
2. Ability to fill from the top with raw water or the bottom
with filtered water from another cell – a flow meter is
needed to control this flow to a rate of 0.3 – 0.6 ft of
filter bed per hour (0.0374 – 0.0748 gpm/ft2).
3. Continuous operation (constant supply of nutrients)
4. Gradual flow rate changes (ideally no more often than
weekly or monthly)
5. Flexibility to change sources or use various combinations
of filter beds

126
 FILTRATION
RATE
Filtration rate should be continuous
1. Good for dissolved oxygen
2. Good for nutrient supply
3. Good for biological mechanisms
4. Influent flow should not scour sand surface
5. 0.1 gpm/ft2 maximum filtration rate
6. 0.03 gpm/ft2 minimum filtration rate
7. Cold temperatures may need lower filtration rates
(e.g., 0.05 gpm/ft2 when water temp < 5°C)
8. Controls should be in place to prevent the tail water
(effluent side) from dropping below the sand bed
during operation (e.g., an effluent weir) – this helps
prevent vacuum conditions and air entrainment.

The filtration rate should be continuous with rate changes needed to accommodate system
demands made gradually over a period of several days or weeks. Operating this way keeps
a constant supply of nutrients and dissolved oxygen needed for healthy biological activity.
Filtration rates should not exceed 0.1 gpm/ft2 and should not drop below 0.03 gpm/ft2.

127
FLOW CONTROL –INLET VS OUTLET
Flow control can be practiced at the inlet or outlet. Inlet flow control can be either
operated as constant rate or declining rate modes.

1. Inlet flow Control – Constant Rate

Uses a throttling valve plus a flowmeter or V-notch weir prior to each filter. The
operator uses the flow control valve to set the desired filtration rate. As the
resistance of the filter bed increases, the water level rises. When the
headwater level approaches the overflow pipe the bed should be cleaned.
• Requires less operator involvement
• Ensures a more constant rate of filtration
• Allows operator to see headloss development as headwater rises
• Low headwater at the beginning of filter runs may make filters more
vulnerable to freezing in the winter if filters are not covered or insulated
2. Inlet Flow Control – Declining Rate
Uses a hydraulic control valve with a flowmeter and valve at the raw water line
prior to each filter that regulates flow while maintaining a constant water
surface elevation above the filter. Effluent flow decreases as the filter plugs.
• Headwater level is not indicative of headloss development (piezometers or
pressure gages are needed)
• Decline in effluent rate or approach to terminal headloss indicates cleaning

128
 INLET
FLOW
CONTROL
Inlet
(Influent)
Control

a. Valve for raw water inlet and regulation of filtration rate

b. Valve for draining supernatant water layer
c. Valve for backfilling the filter bed with clean water
d. Valve for draining the filter bed and outlet chamber
e. Filter to waste valve
f. Valve for delivery of treated water to the disinfection clear well
g. Inlet weir
h. Calibrated flow indicator

Influent control regimes use a weir box and/or valves to control the flow of water into the
filter. The outlet weir structure is still in use to ensure that the water level never drops
below the sand bed during normal operation.

129
 CONSTANT • Starts off with a lower headwater level
• Influent valve is set to the desired rate (e.g.,
RATE INLET 0.03-0.1 gpm/ft2)
CONTROL • Outlet valve is fully open
• The filter will need to be cleaned when the
headwater approaches the overflow
Overflow

Filter
inlet To waste
or return
Filter to inlet
Drain

Filter
effluent
What goes into the filter will go out of the filter (filter effluent or to waste/inlet)
Headloss builds up towards the end of the filter run as it plugs up causing headwater to rise
Cleaning unplugs filter allowing headwater to drop
Excess water produced is re-circulated (pumped) to influent or sent to waste (gravity)

In inlet controlled filters, the rate of filtration is set by the filter inlet valve. Once the
desired rate is set, no further adjustment of the valve is needed. At first the headwater
level will be relatively low, but will gradually rise as the filter plugs. Once the level has
reached the scum outlet or overflow, the filter has to be cleaned. Inlet control reduces the
amount of work and keeps a constant rate of delivery of water into the filter.

130
 DECLINING RATE INLET
CONTROL
Inlet float control valve

Some systems use influent float

control valves in order to control
the headwater level above the
sand. This shows the location of
the inlet float control valve on the
inside of a small package filter.
2012. Camp Yamhill in Yamhill County Oregon.

Some systems use influent float control valves in order to control the headwater level
above the sand. This shows the location of the inlet float control valve on the inside of a
small package filter. [This photo is from Camp Yamhill in Yamhill County, Oregon.]

131
FLOW CONTROL –INLET VS OUTLET
3. Outlet Flow Control (declining rate)
Uses a control valve and flowmeter on the outlet pipe from each filter. As the
filter plugs, the filtration rate will decrease, even if the headwater level is
increased. The level of water on top of the filter can be controlled by using float
switches to turn on and off raw water pumps or control inlet control valves.
Excess water can also be diverted out an overflow and directed back to the
source.
• Most common.
• Fairly simple control method although operator involvement is higher if no
automation is used.
• Higher rates may be implemented faster for emergency situations, since
you don’t have to wait for headwater to rise as with constant rate influent
control.
• Ability to maintain higher headwater level provides better protection from
freezing.
• Higher headwater level provides raw water storage should influent flows
be interrupted due to power failure or intake shutdown due to damage or
to avoid high turbidity events.
• Headwater level is not indicative of headloss development (piezometers or
pressure gages are needed)

132
 OUTLET
FLOW
CONTROL
Outlet
(effluent)
control is
most
common
A. Raw water inlet valve
B. Valve for draining supernatant water layer
C. Valve for backfilling the filter bed with clean water
D. Valve for draining the filter bed and outlet chamber
E. Valve for regulation of the filtration rate
F. Filter to waste valve
G. Valve for delivery of treated water to the disinfection clear well
H. Outlet weir
I. Calibrated flow indicator

Effluent control regimes use a weir box and/or valves to control the flow of water out of the
filter. The outlet weir structure ensures that the water level never drops below the sand
bed during normal operation. This prevents a vacuum from developing and air being
entrained in the sand bed should headloss due to plugging increase to high levels.

133
 DECLINING • Starts off with a higher headwater level
• Influent valve is adjusted to keep filter from
RATE overflowing
OUTLET • Outlet valve is partially closed at first and then is
CONTROL gradually opened as the yield drops due to filter
plugging
Overflow • The filter will need to be cleaned when the yield
cannot keep up with demands
Filter
inlet
Overflow
Filter
Drain

Filter
effluent

What goes into the filter will go out of the filter (Excess can be overflowed if needed)
Headloss builds up towards the end of the filter run as it plugs up causing headwater to rise
Cleaning unplugs filter allowing yield to recover
Influent water is balanced with effluent or excess can be overflowed from headwater

In outlet control, the effluent is restricted at the beginning of the filter fun to keep flows
down to 0.1 gpm/sf or less, while the headwater level is maintained by adjusting the filter
inlet. Daily or every couple of days the valve has to be opened a little further to
compensate for the increase in headloss, causing a slight variation in the rate of filtration.
Inlet and outlet flows have to adjusted periodically to balance flows into and out of the
filter throughout the filter run.

134
 Float control

VALVES valve. Camp

Yamhill in Yamhill
Co. Oregon. 2012.
Select valves for the
purpose they are intended
to serve and the pressures
they are needed to
withstand
Telescoping
Valve. City of
Cannon Beach
in Clatsop Co.
Oregon. 2013.

Butterfly
valves. City of
Banks in
Washington
Co. Oregon.
2011.

135
GATE VALVES – ISOLATION OR THROTTLING
GATE VALVES
Gate valves contain a solid gate that is lowered for closing and raised for opening. This gate may be
in the form of a square, rectangle, circle, oval, or ellipse. There is very little pressure loss through a
gate valve and because they operate slowly, they are unlikely to cause water hammer. In the fully
closed position, gate valves provide a positive seal under pressure. However, under very low
pressure, i.e. 5 psi, light seepage would not be considered abnormal with this kind of valve. Gate
valves should always be left fully open or fully closed. Throttling or fine controlling of gate valves,
which places the gate into the flow of the liquid, can cause serious erosion of the gate. Most
sedimentation basin inlet valves are gate valves. Gate valves are also commonly used as main raw
water intake valves at the heads of water treatment plants.

136
BALL VALVES - ISOLATION
BALL VALVES
Description - Ball valves are very similar to plug valves, except have a ball-shaped plug with a hole
bored through its center that can be rotated to throttle flow. Ball valves are relatively simple and
trouble free, have low pressure drops, and open and close quickly, although opening or closing a ball
valve too quickly can cause water hammer.

Isolation – Allow quick, quarter turn on-off operation, making them good for isolation. With the
development of Teflon seals, ball valves have grown in popularity.

Throttling – Generally have poor throttling characteristics. Ball valves have a ported ball that can be
rotated to throttle the flow of clear water, however, they should be operated either fully open or fully
closed with any liquid containing particles that could scratch the ball.

Common Uses – They can be used for high or low pressure applications. Most water treatment plant
storage tank, day tank, and chemical feed line valves are ball valves.

137
BUTTERFLY VALVES – ISOLATION OR THROTTLING

BUTTERFLY VALVES
Description - Butterfly valves, like ball valves, operate with an adjustable circular disc mounted on a
shaft in the center of the valve that can be opened or closed with just a 1/4 turn.

Isolation – Not normally rated as bubble tight.

Throttling – Can be used for throttling, but should not be used for throttling for extended periods of
time.

Common uses - They are often used for backwash, filter-to-waste, and filter effluent valves. They are
generally used for handling large flows of gases or liquids, including slurries. Butterfly valves are also
commonly used as large water line valves because they are less expensive than similarly sized ball
valves. They are also very compact relative to flanged gate and ball valves.

138
GLOBE VALVES – PRECISE THROTTLING
GLOBE VALVES
Description - Globe valves have a casing that historically has been shaped more globe-like than
today’s models. Globe valves have a plug that fits into a seat within the main cavity area of the
globe. Like a gate, globe valves close slowly to prevent fluid hammer.

Isolation – Not typically used for isolation

Throttling -You can throttle the flow and they will not leak under low pressure when they are shut
off, but have relatively high head loss.

Common Uses - Flow and pressure control valves as well as hose bibs generally use the globe
pattern. The disadvantage of this design is that the "Z" pattern restricts flow more than the gate,
ball, or butterfly valves.

139
PLUG VALVES – ISOLATION OR THROTTLING

PLUG VALVES
Description - Like the gate valve, a plug valve has an unobstructed flow, yet requires only a 90 degree
turn to open it. It also requires very little headroom. Stem corrosion is minimal because there are no
screw threads. Almost all plug valves now are furnished with an elastomer-coated plug and will seal
off drip-tight.

Isolation – Plug valves can seal well and have a tight shutoff, however, some plug valves are made
with a reduced port, which means the valve is smaller than the adjoining pipe’s cross-sectional area,
leading to higher pressure drop – look for full bore plug valves if you need them.

Throttling – Not typically used for throttling, but they have been used for throttling.

Common Uses - Plug valves are available in much larger sizes than ball valves and are highly suitable
for use in wastewater plants.

140
TELESCOPING VALVES

TELESCOPING VALVES
Telescoping valves use a gasketed slip pipe to allow the outlet to be raised or
lowered. They can be fitted without weirs or with V-notch weirs.

141
TELESCOPING VALVES

TELESCOPING VALVES

Image to the right shows:

1. Traditional hand wheel
2. Self locking bevel gear
3. Clear acrylic stem cover tube
4. Remote (electric) actuator
5. Valve position indicator

142
TELESCOPING VALVES

TELESCOPING VALVES

Image to the right shows:

1. Rack and pinion
2. Worm gear box
3. Acrylic rack cover tube
4. Depth indication markings
5. Slip tube lubrication system

In a work gear, the worm can

move the wheel, but the wheel
cannot move the worm. In that
way the weight of the slip pipe
will not turn the hand wheel
when released.

143
 EFFLUENT WEIR

Depth

Depth
Filter skin Filter skin

weir absent Hydrostatic pressure weir present Hydrostatic pressure

Effluent weir prevents air binding caused by a drop in hydrostatic pressures low enough to cause a
partial vacuum in the filter media below the schmutzdecke . This can seriously impact water quality
because algal activity causes the supernatant water to become supersaturated with oxygen, which
would be released as tiny bubbles in the partial vacuum below the schmutzdecke – the condition
known as “air binding”. If air binding occurs in a part of the filter, the remaining filter can become
overloaded. If air binding is more widespread, the hydrostatic head above the filter can greatly
increase, causing a rupture in the schmutzdecke and breakthrough of pathogens.

144
EFFLUENT WEIR
Effluent piping can be
configured to simulate weir
in order to prevent
unplanned bed dewatering
and air binding

BioSand Filter (Dr. David Manz)

145
 EFFLUENT WEIR
Telescoping valve
with a V-notch weir
serves the same
function as an
effluent weir.

Latanick Equipment, Inc.

A telescoping valve serves the same function as an effluent weir. Water can either flow
over the V-notch weir into or out of the slip pipe.

146
 EFFLUENT WEIR
Telescoping Valve – Cannon Beach

City of Cannon Beach,

Oregon

Here is an example of a telescoping valve installed for the City of Cannon Beach, Oregon.

147
 EFFLUENT WEIR
Telescoping Valve – Cannon Beach

To Clearwell

Cannon Beach,
Oregon

Filter Effluent Pipes

(2 filters)

These photos show the interior of the basin outlet control structure with the two filter
effluent pipes and the single telescoping valve outlet pipe. A visible site tube allows
operators to see and measure the position of the telescoping valve. For Cannon Beach,
filtered water flows into the telescoping valve slip pipe and out to the clearwell.

148
 EFFLUENT WEIR
Telescoping Valve – City of Sumpter

8” Stainless Steel
Telescoping Valve

To Clearwell

City of Sumpter,
Oregon => Cannon Beach,
Oregon

For the City of Sumpter, water flows from the filter underdrain up through the telescoping
valve slip pipe and into the control structure.

149
 EFFLUENT WEIR
Floating effluent weirs

These are examples of floating effluent weirs.

150
TEN STATES STANDARDS
FILTER RIPENING

4.3.4.10 Ripening

Slow sand filters shall be operated to waste after scraping or

rebedding during a ripening period until the filter effluent
turbidity falls to consistently below the regulated drinking water
standard established for the system.

151
RIPENING NEW FILTERS
Basic steps to ripening a new filter are as follows:

1. Backfill slowly to displace air pockets at a rate of 0.3 – 0.6 feet

of filter bed depth per hour (0.0374 – 0.0748 gpm/ft2) until the
inlet jets are covered.
2. Set the weir plate with the crest at the level of influent jets
3. Begin top filing through the inlet jets and begin filtering to
waste.
4. The water in the filter box will rise slowly due to the
Schmutzdecke buildup and when the level reaches twice the
distance between the sand bed and influent jets, lower the weir
plate slowly so that the crest is at the level of the sand bed
surface.
5. Continue filter-to-waste until the filter is ripened as indicated
by turbidity < 1 NTU and coliform < 10 CFU/100 ml.

152
FILTER TO WASTE

1. Allows for cleaning newly sanded beds.

2. Allows for ripening without public health risk.
3. Air-gap is recommended to prevent cross-
contamination.

153
 DESIGN FOR WET HARROWING

Facilities Needed:
1. Access for harrowing equipment
2. Harrowed water influent distribution system
• Cross-flow (raw water)
• Up-flow (filtered water)
3. Harrowed wastewater collection system
4. Holding lagoon for the harrowed wastewater
5. Filter-to-waste piping
6. Provisions to prevent equipment from contaminating
filter bed

Facilities needed if designing for harrowing are shown here. Note the additional piping and
controls to allow cross-flow of raw water, which is used to flush debris out of the filter bed
during harrowing through a waste collection system, as well as up-flow of filtered, but
unchlorinated water, to prevent debris from being driven deeper into the sand bed during
harrowing.

154
 DESIGN FOR WET HARROWING
Wet Harrowing
Wet harrowing is a common method
of cleaning small filters.

Basic process:
1. Lower water level to ~6” above
the top of the sand.
2. Use a rake or rake-like
Mechanism
3. agitate top 2”-3” of sand
while slowly backflushing with
filtered, but unchlorinated water
4. Wastewater is collected through
A harrowing valve and waste piping

Wet harrowing is a common method of cleaning small filters. This is often accomplished
with just a stiff-tined garden rake. With the water level lowered to about 6” above the
sand, the top 2-3 inches of sand is agitated. The material suspended by the raking action is
then decanted from the top of the filter through a harrowing valve and waste piping. A
slow backflush using filtered (but unchlorinated water) helps keep the suspended material
from being driven down into the filter.

155
 DESIGN FOR WET HARROWING
Harrowing
Inlet float control valve

Harrowing
Valve &
Waste Line

2012. Camp Yamhill in Yamhill County Oregon.

This shows the location of the inlet float control valve as well as the harrowing waste line
on the inside of the filter. [This photo is also from Camp Yamhill in Yamhill County, Oregon.]

156
SLOW SAND DESIGN
SCRAPED VS. HARROWING
Parameter Scraped Harrowed
(wet harrowed)
Biomass Development Biomass and schmutzdecke take Biomass and schmutzdecke
longer to develop due to the restore at a faster rate, however,
removal of biomass the sudden release of nutrients
can cause dissolved oxygen to
dip as microbial grazing
intensifies. Keeping influent
water flowing and filtering to
waste at a higher initial rate can
help to replenish depleted
oxygen levels.
Removal Efficiency Equivalent once filter is Equivalent once a filter is
properly ripened properly ripened – usually takes
less time to accomplish this.
Filter life Impacted by removal of top Little media loss leads to longer
~2 cm of plugged sand layer filter life. Media is more
susceptible to deep bed clogging
if not done properly.

157
 OTHER DESIGN CONSIDERATIONS
Additional treatment may be needed for challenging waters

Examples:
Roughing Filters (turbidity)
Calcite Contactors (pH)
Ozone (DBP precursors)
Granular activated carbon (TOC/Color)
Filter mat (schmutzdecke removal)
Michigan Environmental
Aeration (low dissolved oxygen) Education Curriculum

City of Astoria, OR
Middle Lake Source 8-28-13
Photo by Gary McLauchlin
Aphanizomenon flos-aquae

Challenging source waters may require additional treatment. For example, roughing filters
may be used prior to slow sand filtration to combat high turbidity, calcite contactors may be
used after filtration to increase pH with corrosive waters, and ozone prior to filtration may
be needed to address DBP precursors. These processes may all be piloted if needed.

158
 OTHER – ROUGHING FILTERS
Roughing Filters

Indicated with pilot

test results.

May be required in the

future due to changes
In source water
quality.

Pre-planning
ensures flexibility

Pilot testing may reveal that roughing filters are needed. Changes to water quality over
time also may dictate the need for roughing filters at a future date. Pre-planning ensures
this flexibility.

159
 ROUGHING FILTERS, CONT.
Roughing Filters

Reduces the algae

and sediment load to
the filters

As with slow sand filters,

biological maturity is key to
optimal performance

Effectiveness
90% removal of particles > 10 microns (medium silt and larger)
72% removal of 2-5 micron particles (Cryptosporidium size particles)

Roughing filters can be 90% effective in removing particles larger than 10 microns and 72%
effective at removing particles in the 2-5 micron size range. This diagram shows the basic
elements of a roughing filter, with the gradual gradation of larger and larger size media.

160
 ROUGHING FILTERS, CONT.
Roughing Filters

• Can be upflow, downflow, or

horizontal

• Filtration rates of 0.12 – 0.62

gpm/ft2

• Cleaned by flushing at high

hydraulic rates. Hydraulic
surges can be generated by
rapid openings and closings
of the inlet and outlet valves

• Horizontal flow roughing

filters are considered to have
greater silt storage capacity
and lower hydraulic cleaning
needs than upflow or
downflow roughing filters.

Roughing filters can be upflow, downflow, or horizontal in configuration with filtration rates
of 0.12 – 0.62 gpm/ft2. They are cleaned by flushing at high hydraulic rates, sometimes
generated by the rapid opening and closing of inlet and outlet valves.

161
 ROUGHING FILTERS, CONT.
Roughing Filters
 Gravel size range from
0.2 – 2 in.

3/8” Rock

60”
1” Rock

1.5” Rock

162
 OTHER – PH CONTROL
Calcite Contactors

• Used to increase the pH in

Corrosive waters

• Limestone can work well

Privately owned
slow sand filter & limestone
calcite contactor
Photos by Stephen Tanner

Calcite contactors like this limestone contactor can help increase the pH in corrosive
waters.

163
 OTHER - OZONE
Ozone

• Used prior to or after filtration

for organics removal
(DBP precursors)

• Oxidizes iron and manganese

• Reduces some algal toxins

• Removes color, taste, and odor

causing compounds

• Increased O&M due to shorter

filter runs

Not commonly used in conjunction with slow sand filters in the Northwest is the use of
ozone. Ozone is an effective means of addressing DBP pre-cursors as well as high iron and
manganese. The use of ozone may lead to shorter filter runs and, hence, more frequent
filter cleaning.

164
 OTHER – ACTIVATED CARBON
Granular Activated Carbon

• Removes color, taste, and odor

causing compounds.
• Needs to be replaced when
depleted (i.e., deactivated)

Comparison of Slow Sand with and without a layer of GAC

Parameter SSF w/out GAC SSF with GAC
Cleaning Frequency 30 days No change
Color Removal 20% 50%
TOC reduction 20% 35-40%

TTHM Formation Potential 130 µg/L 60 g/L

Reduction (24-hr contact time)

Also not commonly used in conjunction with slow sand filters in the Northwest is the use of
GAC.

165
 OTHER – FILTER MATS
Nonwoven Synthetic Filter Mats

• Nonwoven synthetic fabric helps to concentrate the macro-

particle removals on the fabric layers, thereby avoiding the
need to remove sand.
• Fabric increase filter runs due to lower head loss development.
• Filter cleaning involves removal and cleaning of fabric
• Typically for filters smaller than about 3oo ft2 due to logistics of
cleaning the mat. Limit thickness to 1-1.5 inches (2-3 cm).
• Properties of Nonwoven Synthetic Fabrics:

• Thickness of 0.36 – 20 mm
• Bulk density 0.02 – 0.4 g/ml
• Mean fiber diameter 27-48 µm
• Porosity 0.56-0.99
• Specific surface area 13,000 – 14,000 m2/m3

Although not common, nonwoven synthetic fabrics may be used to assist with cleaning
while minimizing sand removal. It does this by trapping macro particulate matter in the
mat rather than the sand, which means that you do not have to scrape the sand to restore
headloss.

166
 OTHER – AERATION
Aeration

May be needed if
DO < 6 mg/l

May be required in
the future due to
changes in source
water quality.

Pre-planning
ensures flexibility

Pilot testing may reveal that aeration is needed. Changes to water quality over time also
may dictate the need for aeration at a future date. Pre-planning ensures this flexibility.

167
 MONITORING POINTS
The recommended minimum location points
for recording and monitoring include:

Source water for:

•Turbidity
•Flow
•Temperature
•pH
•Grab sampling of coliform, TOC, or other water
quality parameters

Supernatant for:
•Level
•Headloss
•Grab sampling of coliform, TOC, or other water
quality parameters

It is important to account for monitoring requirements as part of the design process. This
ensures that sample taps and flow monitoring is adequate to support operations.

168
 MONITORING
Individual filter effluent for:
•Flow rate and quantity
•Turbidity
•Grab sampling of coliform, TOC, or other water quality parameters
Combined filter effluent for:
•Flow rate and quantity
•Turbidity
•Grab sampling of coliform, TOC, or other water quality parameters
Finished water (post disinfection and storage used for disinfection contact
time) for:
•Flow rate and quantity
•pH
•Temperature
•Chlorine residual
•Grab sampling of coliform, TOC, or other water quality parameters
Finished water storage for:
•Effluent flows
•Level

169
 MONITORING HEAD LOSS
Head loss Measurement

On smaller facilities, routine visual observation of

the supernatant depth and recording of the flow
rate may be sufficient to monitor filter head loss
development.

On larger facilities, screened probes at the top and

bottom of the filter sand can allow easy
measurement of head loss with simple piezometers
mounted outside the filter, or through the use of a
differential pressure transducer connected to the
facility’s SCADA system.

Tracking this data will allow the operator to predict

and plan filter cleanings.

170
 Frequency and tasks are adapted from WHO,
1996. Fact Sheets on Environmental Sanitation,
O&M MANUAL Fact Sheet 2.12: Slow Sand Filtration

Frequency Labor Slow Sand Filter Maintenance Task

(person hours)

Daily 1-3 Check raw water intake

Check/adjust filtration rate
Check water level in filter
Check water level in clear well
Sample & check water quality (raw/finished
NTU, raw temp)
Check pumps
Enter observations in logbook

Weekly 1-3 Check & grease any pumps & moving parts
Check/re-stock fuel
Sample & check water quality (coliform)
Enter observations in logbook

1 – 2 months 5 / 1,000 ft2 Scrape filter beds

50 / 1,000 ft2 /12 inches of sand for Wash scrapings & store retained sand
re-sanding Check & record sand bed depth
(Letterman & Cullen, 1985) Enter observations in logbook

Design should include development of an Operation and Maintenance Manual. Some of

the tasks that should be included are shown here. More about this will be discussed later
as we get into operations.

171
 KEY REFERENCES – 1974, 1987
1. “Slow Sand Filtration”, World Health Organization (Huisman
& Wood), 1974;
2. “Slow Sand Filtration for Community Water Supply”,
International Research Center for Community Water Supply
and Sanitation (Visscher et al., 1987)

http://www.who.int/water_sanitat http://www.irc.nl/page/4530
ion_health/publications/ssf/en/ind
ex.html

Again, the key references are shown here.

172
 KEY REFERENCES - 1991

“Manual of Design for Slow

Sand Filtration“ . David
Hendricks & American Water
Works Association, 1991.
ISBN 978-0898675511

The “Manual of Design for Slow Sand Filtration” covers design in great detail.

173
 KEY REFERENCES - 2012
“Water Treatment Plant Design, 5th Edition”.
Stephen J. Randtke, Ph.D., P.E.; Michael B.
Horsley, P.E. Co-published by the American
Water Works Association; Environmental and
Water Resources Institute of American Society of
Civil Engineers; McGraw-Hill Professional. 2012.
ISBN: 9780071745727

Recommended
Standards for Water
Works (a.k.a., “Ten
States Standards”,
2012);

http://10statesstandards.com/

Here is a more recent publication, which covers design as well.

174
QUESTIONS ABOUT DESIGN?

175