Page 493 - Fundamentals of Water Treatment Unit Processes : Physical, Chemical, and Biological

448 Fundamentals of Water Treatment Unit Processes: Physical, Chemical, and Biological

the longest filter run. The answer depends on raw-water 14.5.2.1 SR Ranch, Colorado
characteristics, which may vary seasonally. Other questions The SR Ranch in Colorado had been using a 38 L=min
may relate to the effect of increasing HLR on the economics (10 gpm) diatomite filtration system starting about 1970.
of operation and on effluent quality. The headloss versus time The raw-water source was the Boyd Hanson Feeder Canal,
curves are of interest for each HLR so that the effect of near Loveland, Colorado, which was snow-melt water. The
increasing HLR can be anticipated. The pilot study can also turbidity range was 2.3–6.7 NTU during the period of testing,
provide guidelines for pre-coat and body feed and whether to which was typical during most of the year. Excursions to
add alum or a polymer for short periods, for example, in 30 NTU may occur, however, during spring runoff. Cysts of
responding to a turbidity event such as after a rainfall. In G. lamblia have been found in the canal. The system was
addition to its utility in design, a pilot plant can serve as a operated about 8 h=day with storage used to meet the off-line
continuing tool for operation and may be retained on-site demand of about 30–50 persons. The equipment was a U.S.
usually for a nominal cost. Army Erdlator system that had been obtained as military
surplus. The unit had supplied field troops with drinking
14.5.1.1 Functional Relationships
water from surface supplies. Since the system supplied 25
The questions for a pilot plant study may be formulated in persons, it was classified as a public water system under the
terms of the operative variables, for example, as given in 1974 U.S. Safe Drinking Water Act and therefore was
Section 14.3.2.1. In such a case, the dependent variables are required to meet drinking-water standards. For diatomite fil-
some function of the independent variables, that is, tration, the turbidity standard was 1 NTU set by EPA and
adopted by the Colorado Health Department. At the time of

. Effluent concentrations of selected contaminants the investigation in 1984, the pressure tank was found to have
. Rate of headloss increase a few pinpoint holes due to corrosion and a high rate of
headloss increase. The maximum pressure differential across
2 3
. Raw water quality
the filter cake was DP(cake) 207 kPa (30 psi). The manage-
6 . Grade of diatomite 7
6 7 ment was interested in updating the system and in operating
6 . Alum=polymer concentration 7
¼ f 6 7 (14:14) parameters being defined for meeting drinking-water stand-
6 . HLR 7
6 7 ards. Since the SR Ranch had a limited budget, an exploratory
4 . W(precoat) 5
gratis study was undertaken by Mr. Ray McIndoe, then with
. C(bodyfeed)
Johns Manville in Denver and Mr. Alan Wirsig, a consultant
recently retired from Johns Manville.
Empirical relationships generated from a pilot plant based on
Equation 14.14 are necessarily limited by time and budget 14.5.2.1.1 Pilot Plant
(a well-known truism), and so the study must be pared down The pilot plant used to investigate the issues at SR Ranch was
2
2
from including all of these variables. In paring down, know- a Johns-Manville 0.091 m (1.0 ft ) stainless steel unit shown
ledge of research and practice can serve as a guide, for in Figure 14.15a. Figure 14.15b shows the same pressure
example, diatomite grades for water treatment generally have vessel=septum assembly with some of the adjunct instrumen-
a size range of 18 < d(50) < 24; smaller causes excessive head- tation required to perform the pilot plant study. Such instru-
loss and larger does not, as a rule, remove sufficient turbidity. ments include a raw-water screw-type metering pump, a raw-
2
Also 2.44 HLR 4.88 m=h (1.0 HLR 2 gpm=ft ), with water flow meter, a pre-coat tank and pump for circulation, a
little effect on water quality if higher values are used. Whether body-feed tank and metering pump, pressure gage for the
2
to use HLR 4.88 m=h (2 gpm=ft ) in sizing units involves pressure vessel, air vent for the pressure vessel, a constant
trade-offs between capital and operating costs (Box 14.4). head overflow for the effluent, injection port in the raw-water
line, and sampling ports in the raw-water line and the effluent
line. The setup was moved from a laboratory at CSU to the
14.5.2 CASES
field site adjacent to the Boyd Hansen Feeder canal, the water
Two diatomite filtration pilot plant cases may illustrate the source, requiring only a few hours for the move. The setup
diversity of situations. Every case has a different aspect. was essentially the same as required for a full-scale plant.
Effluent turbidities 0.5 NTU for C512 and Hyflo and 0.7
NTU for Celatom FW20, with DP=Dt 5.5 kPa=h (0.8 psi=h)
BOX 14.4 PILOT PLANT STUDIES
for run length 207 kPa=5.5 kPa=h ¼ 38 h, which is much
Pilot plant studies are common in the field of diatomite longer than the needed 8 h. Effluent turbidity was 3.2 NTU for
filtration. Skid-mounted units are available, as a rule, the coarser grade, C545, which of course was not satisfactory.
from manufacturers of equipment or diatomite produ- Adding alum as a coating to the diatomite reduced markedly
cers. In the case of cylindrical elements, however, a the effluent turbidities for each of the grades, as expected,
pilot plant must be fabricated. A pilot plant study can without appreciable DP=Dt penalty. While simplicity of oper-
either confirm that a contemplated design will perform ation is a main goal in any diatomite system, the alum per-
as intended or possibly save funds in construction. formance provided knowledge that there is a ‘‘way out’’
should the need for lower effluent turbidities be necessary.