Page 420 - Fundamentals of Water Treatment Unit Processes : Physical, Chemical, and Biological

Rapid Filtration 375

(a) (b)
FIGURE 12.41 Mudballs from filters: beginning and fully formed. (a) A microscopic agglomerate of alum floc and anthracite. (Courtesy of
Grant Williamson-Jones, City of Fort Collins Utilities, Font Collins CO.) (b) Fully formed mudball. (Reprinted from Mackay, D.J., Opflow,
14(11), 1, 1988b. With permission. Copyright ß 1988, American Water Works Association.)

12.5.3.1 Mudballs and Surface Cracks 12.5.3.4 Operating Protocol
Figure 12.41a shows a small agglomerate of floc and media, An operating protocol recommended by Amirtharajah (1982)
which is the beginning stage of a mudball formation. Figure for both effective cleaning and minimizing media loss is as
12.41b shows an example of a fully formed mudball. If not follows:
disrupted by cleaning, the small agglomerate of floc and
media, seen in Figure 12.41a will grow in size during con- 1. Lower water level in filter to media surface.
tinued filter cycles with the formation of mudballs, as in 2. Begin water backwash.
Figure 12.41b. Baylis (1937, p. 1020) described the problem. 3. Introduce air slowly after bed is flooded by 80–160
The data presented showed that much of the filter bed mm (3–6 in.) of water, using low water velocity and
trouble originated from an accumulation of compacted coagu- high air loading rate which also prolongs the time to
lated material not removed by the washing system, and to reach the weir crest of the wash-water trough.
coatings on the sand grains. Shrinkage or settlement of the 4. Terminate air when the water level is about 160 mm
beds while in service usually is caused by a soft coating (6 in.) from the weir crest.
around the sand grains which is not removed by backwashing 5. Increase the water backwash rate to cause about 0.20
the filters. This shrinkage allows cracks to open along the bed expansion.
sidewalls and occasionally other parts of the bed.
The most effective control is to clean the media adequately The procedure requires considerable operator attention.
through backwash with surface-wash or air-wash. Surface-wash A SCADA system can be programmed to perform these
is especially effective in breaking up compacted surface layer, functions.
thought to be a precursor to mudball formation. The use of
polymers may also predispose the media to ‘‘stickiness,’’ and
development of mudballs. If mudballs build up to a great extent 12.6 PILOT PLANTS
the media may require replacement.
For reference, Section 3.5 describes the general issues of pilot
12.5.3.2 Floc-to-Grain Bonding plant experimentation. Examples of permanent pilot plants
The bonding of an aluminum or iron floc to a grain of sand or used as an adjunct to operation have been at Fort Collins,
anthracite is due to four forces: (1) van der Waals, (2) elec- Colorado (1988) and Bellingham, Washington (1993),
trical double-layer, (3) Born repulsion, and (4) structural respectively. The Fort Collins pilot plant was set up with
(Raveendran and Amirtharajah, 1995). The hydraulic shear two conventional treatment trains, with 37 L=min (10 gpm)
and grain-to-grain shear and impact forces developed by flow each, which can be changed to in-line or direct modes by
backwash must overcome these bonding forces in order to valve adjustment with flow directed to any one or all of three
dislodge attached particles. filter columns, each with a different media design. The pilot
plant has been used to address design questions as changes are
12.5.3.3 Practice contemplated and for a continuing array of operational
In backwash, the media expands and the grains are main- questions. The Bellingham pilot plant has a flow of 37
tained apart due to the nature of the pressure field associated L=min (5 gpm) per train, with three in-line treatment trains.
with an array of particles. Therefore, grain-to-grain contacts do Figure 12.42 shows one of the 305 mm (12 in.) square filter
not occur and fluid shear is the only removal mechanism columns. The pilot plant was installed to address process
(Hewitt and Amirtharajah, 1984). Both are necessary, how- design questions, anticipating plant expansion as population
ever, for effective cleaning. Therefore, an adjunct to bed fluid- increases. One question was whether increased filtration vel-
ization, that is, either air-wash or surface-wash, is required. ocity, as a means of handling increasing future demands,