Water and Wastewater Engineering

Подождите немного. Документ загружается.

11-36 WATER AND WASTEWATER ENGINEERING

without good maintenance they may overshoot the set point and “hunt” for the correct position, or

possibly attempt to balance the flow in steps that result in surges to the filters that remain on line.

In the weir splitting system, water flows in through a common channel and is split equally to

all operating filters by the weir. This system is the sim

plest method for splitting the flow, and rate

changes are made gradually without the hunting of control valves.

In declining rate filtration no active control or apportionment is used. Each filter receives a

different flow rate depending on the accumulated headloss. The cleanest filters receive the great-

est flow, and the flow through each filter declines as solids a

ccumulate. The advantage of this

system is that it can be constructed without instrumentation or flow control. The disadvantages

are that the operators have no indication of the flow rate or headloss, there is no method to control

the filtration rate, and the rate at the beginning of filtration after cleaning may exceed the design

filtration rate resulting in turbidity breakthrough.

Headloss Accommodation

The hydraulic gradient through the filter at various tim es during a filter run is illustrated in

Figure 11-15 . Negative head (less than atmos pheric pressure) develops in a filter when the

summation of headlosses from the media surface downward exceeds the depth of water to that

point. It is not uncommon to have a meter or more of headlo

ss in the upper 15 cm of a dirty

sand filter bed (Cleasby, 1972). When shallow water operating depths are provided, negative

head develops a short distance into the media. The negative pressure may cause bubbles of

dissolved oxygen and nitrogen to come out of solution. If these bubbles are trapped in the bed

(a phenomenon

called air binding ), they red uce the effective filtering area. This increases the

filtration rate through the remaining filter area, which results in a more rapid turbidity break-

through than otherwise would occur. In addition, the headloss rises dramatically.

The depth of water required to prevent a negative pressu re in the filter can be estimated by

solving Bernou

lli’s equation for the headloss between the water surface on the filter and the cen-

terline of the effluent pipe. Using the notation in Figure 11-16 :

() ()vp

22ggg

 



(11-21)

0.5

Near end

of filter run

45°

Zone of

negative

pressure

Near middle

of filter run

Clean filter

at design rate

Static pressure

(no flow)

Water surface

Coal

Sand

Under drain

1.0

Pressure, m (water gauge)

Depth from water surface

1.5 2.00

0.5

FIGURE 11-15

Pressure development within filter bed during filtration.

GRANULAR FILTRATION 11-37

where v

, v

 velocities at points 1 and 2, m/s

, p

 pressures at points 1 and 2, kPa

, z

 elevation heads at points 1 and 2, m

g  acceleration due to gravity, m/s

  specific weight of water, kN/m

 headloss through filter media, and underdrain system, m

If relative pressure is used, p

 0. The velocity at the surface of the water in the filter box

may be assumed to be zero. The datum is selected so that z

 0. Equation 11-21 may then be

reduced to

2

 

()

(11-22)

The depth of sand, gravel, underdrain, and the water above the media is equal to z

. The velocity

in the effluent pipe is v

. The depth of water above the filter media is selected to maintain a posi-

tive value for p

/ .

It is natural for designers to make the filter box as shallow as possible. Deeper boxes are

more expensive than shallow ones. Early in the 20th century it was common to design for media

submergence of 1.2 to 1.5 m for rapid sand filters. A majority of these plants (up to 80 percent)

reported air binding (Huds

on, 1981).

The available headloss for gravity filters is generally designed to be 2.5 to 3 m. Kawamura

(2000) recommends that a minimum depth of water above the filter bed of 1.8 m be provided and

that a preferred depth is 2.4 m. He also notes that the turbidity of gravity filters often begins to

increase when the net headloss is over 1.8 m. Consequently, the filter m

ust be backwashed even

though the headloss available may be 2.4 m or more.

E xample 11-8 illustrates the headloss calculation.

Effluent

Influent

/ 2g

/␥

FIGURE 11-16

S chematic of granular filter during filtration. The Notation for Equation 11-22 is shown

at right.

11-38 WATER AND WASTEWATER ENGINEERING

Example 11-8. Determine the depth of the filter box for Ottawa Island’s sand filter. Use the

clean bed headloss from Example 11-2 , and the velocity headloss and the media and underdrain

depths from Example 11-7 . Assume the minimum depth of water above the filter bed is 2.4 m.

Solution:

a . Calculate z

using the assumed water depth and other depths from Example 11-7 .

water depth depth of sand gravel and,,uunderdrain

mmmmmz

24 06 03 0 336.. ...

b. The velocity headloss from Example 11-7 is

2981

0165 017

m/s

or m

()

c. Calculate p

/  at the beginning of a filter run using the clean bed headloss calculated in

E xample 11-2 .

36 017 076 267 27



  .. . ..mmm or

d. Estimate the maximum allowable headloss.

36 017 343



  .. .mmh mh

Thus, the maximum headloss that will cause a negative pressure in the filter is one that

is 3.43 m.

e. The filter box depth is estimated as

box

depth of water depth of media depth of underdrain

factor of safety freeboard

A ssuming a factor of safety of 0.6 m and a freeboard depth of 0.6 m

box

mmmmmm24 09 03 06 06 48......

Comments:

1 . The elevation of the maximum water level in the filter box governs the hydraulic profile

of the upstream processes in the plant. This, in turn, has a significant impact on the cost

of the plant. This cost, as well as the cost of the filter itself, favor the design of shallower

rather than deeper filter boxes.

2. The use of low profile drains and those that do not need gravel allow for the

design of a

shallower filter box.

3. Though more expensive, a d eeper filter box provides for future expansion to deep bed

monomedium.

4. The technique used here for establishing the depth of the filter box is conservative. Monk

(1984) provides a detailed method for optimizing the depth of the filter box.

GRANULAR FILTRATION 11-39

In addition to providing for a minimum depth of water, operational control must be provided to

accommodate the increase in headloss during operation. Three methods are used to accommodate

the increase in headloss during filtration: (1) maintaining a constant head in the filter effluent by the

use of a modulating control valve, (2) maintaining a constant head in the filter effluent and

allowing

the water level to rise, and (3) maintaining a constant headloss and allowing the filtration rate to

decline. These are discussed in detail by Castro et al. (2005), Monk (1987), and MWH (2005).

Some Important Appurtenances

Many features of the filter design are beyond the scope of this text, but a few are sufficiently

noteworthy to identify them here.

A s noted in F igure 11-2 the turbidity levels rise immediately after filter backwash and then

drop off as the filter begins to clog. When this rise is averaged over the filter run, the change

in turbidity is sm

all. However, outbreaks of Giardiasis and Cryptosporidiosis caused by proto-

zoan cysts that are very resistant to chlorine disinfection make even this “small” perturbation a

potential health hazard (Hibler et al., 1987; Kramer et al., 1996). The design and operating solu-

tion is to provide for a “filter-to-waste” period immediately after backwashing. A diagram s

how-

ing a method for the design to implement filter to waste is shown in Figure 11-17 .

Because of outbreaks of Giardiasis and Cryptosporidiosis, filtration practice now includes a

provision for monitoring the effluent turbidity from eac h filter. It is us ed to guide the length of

Lateral drain

Washwater rate

controller

Bottom of

washwater

trough

Washwater

storage tank

Water level

while filtering

Water level

while washing

From coagulation

clarification basin

Filter rate

controller

Filter to

waste valve

Washwater

drain

Main drain

Seal

Filtered water

storage tank

Sand

Gravel

FIGURE 11-17

Diagrammatic section of a rapid sand gravity filter. A, B, C, D, and E are valves that may be

hydraulically or pneumatically actuated. Valve D permits wasting filtered water. The seal in the

effluent pipe keeps the pipe full at all times so that the rate controller will function.

11-40 WATER AND WASTEWATER ENGINEERING

time for filter effluent to be diverted to waste and to detect the onset of breakthrough of turbidity

that signals the end of a filter run.

Many very small particles can exist in the water with turbidity less than 1 NTU. This implies

that asbestos fibers, bacteria, viruses, and cysts may be passing through even with very low tu

bidities. Particle counters are used to detect these small particles and are recommended in con-

junction with turbidimeters to address this issue (Cleasby and Logsdon, 1999).

Design Criteria

Along with Tables 11-1 , 11-3 , 11-4 , and 11-5 , Tables 11-6 through 11-13 provide a summary of

the design guidance presented in this section.

Source: Adapted from MWH, 2005.

TABLE 11-6

Pretreatment conditions for different operating modes

Operating mode Pretreatment conditions

Conventional filtration Coagulation with alum or ferric chloride and

polymer followed by flocculation and

sedimentation. Can treat turbidities up to

1,000 NTU.

Direct filtration Coagulation with alum or ferric chloride and

polymer followed by flocculation, but not

sedimentation. Limited to raw water turbidities less

than 15

NTU.

In-line filtration Coagulation with alum or ferric chloride and

polymer. Flocculation is incidental. Sedimentation

is not provided. Limited to turbidities less than 10

NTU.

TABLE 11-7

Recommended dimensions of ordinary gravity rapid filters

Parameter

Reported

range of values Comment

Area of filter 25–100 m

Very large plant maximum is 200 m

Width of cell  6 m For “Off-the-shelf” troughs

L:W ratio 2:1 to 4:1

Depth 4–8 m Provide space for underdrain

Depth of water  1.8 m

Gullet

Width 0.4–2 m

Depth Varies from top of media to

bottom of underdrain

Measured from the bottom of wash

trough

Sources: Castro et al., 2005; GLUMRB, 2003; Kawamura, 2000.

GRANULAR FILTRATION 11-41

Parameter

Reported

range

GLUMRB

recommendation Recommended

Sand only

Effective size 0.35– 0.7 mm 0.45–0.55 mm

Uniformity coefficient 1.3–1.8 1.65

Shape factor (f)

0.7– 0.95

Porosity 0.4– 0.47

Specific gravity 2.65

Depth of medium 0.6– 0.75 m 0.6 m and  .76 m

Filtration rate 5–12 m/h 7.5

m/h

Backwash rate 30–60 m/h 37 m/h

Backwash duration

10–20 min 15 min

Surface wash rate

Revolving arms 1.2–1.8 m/h 1.2m/h

Fixed arms 4.9–10 m/h 4.9 m/h

Air scour Staged or special baffle

Anthracite coal only

Effective size 0.70– 0.75 mm 0.45–0.55 mm

Uniformity coefficient 1.3–1.8 1.65

Shape factor (f)

0.46–0.73

Specific gravity 1.45–1.75

Porosit

y 0.53–0.60

Depth of medium 0.6–0.75 m

Filtration rate 5–12 m/h 7.5 m/h

Backwash rate 37– 45 m/h 24 m/h

Backwash duration

10–20 min 15 min

Surface wash rate

Revolving arms 1.2–1.8 m/h 1.2 m/h

Fixed arms 4.9–10 m/h 4.9 m/h

Underdrain

Pipe lateral Yes

BlockYes

Air scourNo

Air scour Staged or special baffle

TABLE 11-8

Design criteria for single-medium filters

A ctual off-line time will be  30 min because of the time required to drain the filter and gradually come up to the

full backwash rate. An additional 30–40 minutes off-line is required for “filter-to-waste” to clear the bed of wash

water and dislodged turbidity.

Sources: Castro et al., 2005; Cleasby and Log

sdon, 1999; GLUMRB, 2003; Kawamura, 2000; MWH, 2005;

Reynolds and Richardson, 1996.

11-42 WATER AND WASTEWATER ENGINEERING

TABLE 11-9

Design criteria for dual-media filters

Parameter

Reported

range

GLUMRB

recommendation Recommended

Anthracite coal on top

Effective size 0.9–1.4 mm 0.8–1.2 mm

Uniformity coefficient 1.3–1.8 1.85

Shape factor (f)

0.46–0.73

Porosity 0.53–0.60

Specific gravity 1.45–1.75

Depth of medium 0.4–0.6 m

Sand on bottom

Effective size 0.35–0.7 mm 0.45–0.55 mm

Uniformity coeffic

ient 1.3–1.8 1.65

Shape factor (f)

0.7–0.95

Porosity 0.4–0.47

Specific gravity 2.65

Depth of medium 0.3–0.75 m 0.6 m and 0.76 m

Filtration rate 7–20 m/h  15 m/h

Backwash rate 30–60 m/h 24 m/h

Backwash duration

10–20 min 15 min

Surface wash rate

Revolving arms 1.2–1.8 m/h 1.2 m/h

Fixed arms 4.9–10 m/h 4.9 m/h

Underdrain

Pipe lateral Yes

BlockYes

Air scour No—if total depth of medium

0.75 m

Air scour 0.6–1.5 m

/min · m

if total

medium depth  1 m

A ctual off-line time will be  30 min because of the time required to drain the filter and gradually come up to the full

backwash rate. An additional 30–40 minutes off-line is required for “filter-to-waste” to clear the bed of wash water and

dislodged turbidity. If air scour is provided, the time will be even longer beca

use of the necessity of sequencing the air scour

and wash water.

Sources: Castro et al., 2005; Cleasby and Logsdon, 1999; GLUMRB, 2003; Kawamura, 2000; MWH, 2005; Reynolds and

Richardson, 1996.

GRANULAR FILTRATION 11-43

Parameter

Reported

range Recommended

Anthracite coal on top

Effective size 0.9–1.4 mm

Uniformity coefficient 1.4–1.75

Shape factor (f)

0.46–0.73

Porosity 0.53–0.60

Specific gravity 1.45–1.75

Depth of medium 0.4–0.5 m

Sand in middle

Effective size 0.45–0.55 mm

Uniformity coefficient 1.4–1.65

Shape factor (f)

0.7–0.95

Porosity 0.4–0.47

Spec

ific gravity 2.65

Depth of medium 0.15–0.3 m

Garnet on bottom

Effective size 0.20–0.35 mm

Uniformity coefficient 1.2–2.0

Shape factor (f)

0.6

Specific gravity 3.6–4.2

Depth of medium 0.075–0.15 m

Filtration rate 10–25 m/h 15 m/h

Backwash rate 37–45 m/h

Backwash duration

10–20 min 15 min

Surface wash rate

Revolving arms 1.2–1.8 m/h

Fixed arms 4.9–10 m/h

Underdrain

Pipe lateral Yes

BlockYes

Air scour No—if total depth of medium

0.75 m

Air scour 0.6–1.5 m

/min · m

if total medium

depth  1 m

TABLE 11-10

Design criteria for tri-media filters

Actual off -line time will be  30 min because of the time required to drain the filter and gradually come

up to the full backwash rate. An additional 30–40 minutes off-line is required for “filter-to-waste” to

clear the bed of wash water and dislodged turbidity. If air scour is provided, the time will be even longer

becau

se of the necessity of sequencing the air scour and wash water.

Sources: Castro et al., 2005; Cleasby and Logsdon, 1999; Kawamura, 2000; MWH, 2005; Reynolds and

Richardson, 1996.

11-44 WATER AND WASTEWATER ENGINEERING

Parameter

Reported range

of values Recommended

Anthracite coal

Effective size 0.9–1.0 mm

Uniformity coefficient 1.4–1.7

Shape factor (f)

0.46–0.73

Specific gravity 1.45–1.75

Porosity 0.53–0.60

Depth of medium 0.9–1.8 m

Filtration rate 10–25 m/h 15 m/h

Backwash rate 37–45 m/h

Backwash duration

15 min

Surface wash rate

Revolving arms 1.2–1.8 m/h 1.2 m/h

Fixed arms 4.9–10 m/h 4.9 m/h

Underdrain

Pipe lateral No

BlockYes

Air scourYes

Air scour 0.6–1.5 m

/min · m

TABLE 11-11

Design criteria for deep-bed monomedium filters

A ctual off-line time will be  30 min because of the time required to drain the filter and

gradually come up to the full backwash rate. An additional 30–40 minutes off-line is required

for “filter-to-waste” to clear the bed of wash water and dislodged turbidity. If air scour is provided,

the time will be even longer be

cause of the necessity of sequencing the air scour and wash

water.

Sources: Castro et al. 2005; Cleasby and Logsdon, 1999; Kawamura, 2000; MWH, 2005; Reynolds

and Richardson, 1996.

Parameter Criterion

Length 6m

Interval Particle travel distance 1 m

Height to weir edge

Sand filter Expansion  depth of trough  0.15 to 0.3 m

Anthracite 1.1–1.2 m

Freeboard5 cm

Installation Weirs level

TABLE 11-12

Wash trough design criteria

Sources: Castro et al., 2005; Cleasby and Logsdon, 1999; GLUMRB, 2003.

GRANULAR FILTRATION 11-45

TABLE 11-13

Approximate flow velocities for filter channels and piping

Channel or pipe Velocity, m/s

Influent conduit carrying flocculated water 0.3–0.6

Effluent conduit carrying filtered water 0.9–1.8

Backwash water conduit carrying clean wash water 2.4–3.5

Backwash water conduit carrying used wash water 1.2–2.4

Filter-to-waste connections3.6–4.6

11-7 OPERATION AND MAINTENANCE

Under steady-state conditions, the operation and maintenance of the filter system is routine. Rapid

sand filters following coagulation/flocculation are generally operated with filter run lengths be-

tween 12 and 96 hours with typical runs of about 24 hours. Some plants operate with longer

cycles. Longer runs may make cleaning difficult beca

use of compaction of the particulate matter

(Cleasby and Logsdon, 1999; Castro et al., 2005). Baumann (1978) recommends that at peak

solids and flow, run times should be greater than 15 hours and less than 24 hou rs. In the East

Lansing water treatment plant, with softening floc, run times with a 2 NTU influent to the filter

are limited to about 60 hours. With a m ore normal 0.

5 NTU influent, filter runs are terminated

at 120 hours to allow a 24-hour “float” for operational expediency. There is a general relation-

ship between the influent particulate concentration, the filtration rate, and the filter run. It can be

expressed as follows (Reynolds and Richards, 1996):

()runtimeat runtimeat

in in

runtimeat runtimeat

()

where v

a 1

, v

a 2

 filtration rates 1 and 2, m/h

, C

in 2

 influent particulate concentration, NTU

With changing raw water quality, equipment failure, power outages, and maintenance activi-

ties, rigorous attention to the filter system and upstream process is required. Three indicators are

used in evaluating the performance of the filter: filtered water turbidity, length of the filter run,

and the ratio of the volume of bac

kwash water to the volume of filtered water.

After backwash, when steady-state is achieved, the turbidity should always be less than 0.1

NTU. Deviation from this level is an indication of need for ad justment of the coagulation/floc-

culation/settling system.

Shorter filter runs may imply one or more of several problems. Exam

ples include air binding,

accumulation of mud balls, and poorly settling floc.

Increases in the ratio of wash water to filtered water imply difficulty in cleaning the filter.

This can result from deterioration (plugging) of the surface wash system or maldistribution of the

backwash water.

These and many other i

ssues are addressed in detail in Kawamura (2000) and the American

Water Works Association’s Filter Maintenance and Operations Guidance Manual (2002).