SESSION II
TERTIARY
TREATMENT FOR
WASTEWATER REUSE
ADVANCED WASTEWATER TREATMENT
DESIGN SEMINAR
RIVERSIDE, CALIFORNIA
MARCH 24, 1972
U.S. ENVIRONMENTAL PROTECTION AGENCY
NATIONAL ENVIRONMENTAL RESEARCH CENTER
ADVANCED WASTE TREATMENT RESEARCH LABORATORY
CINCINNATI, OHIO
-------
Program for
Advanced Wastewater Treatment
Design Seminar
Riverside, California
March 24, 1972
Session Introduction
Mr. John Merrell, Director 30 min. Introduction
Categorial Programs Division
EPA - Region IX
San Francisco, California
Session I Control of Nitrogen in Wastewater Effluents
Speaker
Time
Topic
Mr. Edwin F. Barth 45 min.
Biological Treatment Research Prog.
Environmental Protection Agency
NERC-Cincinnati, Ohio
Nitrogen control - general
considerations including
design of two sludge systems
Mr. Dennis Parker
Brown & Caldwell, Inc.
Consulting Engineers
San Francisco, California
45 min.
Design of Central Contra
Costa plant for nitrogen
control
Coffee Break
15 min.
Mr. John M. Smith 60 min.
Municipal Treatment Research Prog.
Environmental Protection Agency
NERC-Cincinnati, Ohio
Upgrading existing facilities
for nitrogen control including
denitrification
Drf Robert B. Dean, Chief
Ultimate Disposal Program
Environmental Protection Agency
NERC-Cincinnati, Ohio
30 min.
Physical-chemical methods
of nitrogen removal
30 min.
Questions and discussion
-------
-2-
Session II - Tertiary Treatment for Wastewater Reuse
Speaker Time Topic
Mr. A. N. Masse, Chief 30 min. Tertiary clarification
Municipal Treatment Research Prog.
Environmental Protection Agency
NERC-Cincinnati, Ohio
Dr. Robert B. Dean, Chief 45 min. Lime recovery
Ultimate Disposal Program
Environmental Protection Agency
NERC-Cincinnati, Ohio
Coffee Break
Mr. George M. Wesner, Engineer
Orange County Water District
Orange County, California
45 min.
Staff Engineer - County Sanitation 60 min.
Districts of Los Angeles County
Clarification and
filtration design
Disinfection of municipal
wastewater
30 min.
Questions and discussion
-------
CONTENTS
COAGULATION OF WASTEWATER
PHOSPHORUS REMOVAL BY LIME TREATMENT OF SECONDARY EFFLUENT
LIME RECOVERY AND REUSE
FILTRATION
DISINFECTION
-------
COAGULATION OF WASTEWATER
-------
CHAPTER 4
COAGULATION OF WASTEWATER
4.1 General
Wastewaters contain a wide variety of organic and inorganic suspended solids. These solids
may have been present in the raw wastewater or may have been precipitated from solution
during previous stages of treatment. All must be removed if a high quality effluent is to be
produced.
A number of factors will influence the rates at which wastewater solids may be removed by
sedimentation. Particularly important is particle size. Small particles in the colloidal range
will not settle out in a practical detention time and must be agglomerated into larger
particles which will settle at reasonable rates.
Small particles have very large surface areas per unit weight of solid with associated forces
which tend to keep the particles separated or in a stable state. Effecting aggregation of the
solids is a matter of overcoming these stabilizing forces by application of selected chemical
coagulants. It is not the purpose of this manual to cover the subject of coagulation in detail.
Only such information which is useful in process control is provided. More extensive cover-
age can be found in a number of publications for those who desire further information (1),(2),
(3).
The principle stabilizing forces are electrostatic repulsion from electrical double layers sur-
rounding particles suspended in water and physical separation of particles by films of water
adsorbed on the particle surfaces. The electrical double layer is due to an imbalance of ions
near the particle surface and imparts an electrical charge to particles which is generally
negative for domestic wastewater particles (4). The particles with like charges then tend to
repel one another. The magnitude of the charge on the double layer must be reduced to
permit the particles to come together for agglomeration. Reduction of the charge is a
principle function of the coagulant.
Principal natural destabilizing or aggregating forces are Brownian movement and Van der
Waals forces. Brownian movement is a constant random motion of small particles due to
collisions of the particles with thermally-agitated water molecules, while Van der Waals
forces are atomic dipole interactions which exist between all atoms. Van der Waals forces,
which are always attractive, predominate at short distances from the particle surfaces. If
Brownian movement causes two or more particles to approach within a short distance from
each other, and if the repulsive electrical forces have been sufficiently reduced, then the
particles will be held together by the Van der Waal forces. Other particles can be added to
the particle pair until the resulting aggregate or floe reaches such a size that it rapidly settles
out of solution.
These natural destabilizing forces are generally not sufficiently rapid to permit effective
solids removal from wastewater,but must be augmented by addition of chemical coagulants
to destabilize and tie the particles together and by application of hydraulically-or mechanic-
ally-applied mixing to promote rapid collisions of the destabilized particles. A number of
mechanisms have been proposed to explain the action of coagulants in promoting particle
4-1
-------
aggregation. These include charge reduction, physical or chemical bridging of coagulant
molecular chains between particles and physical enmeshment of particles in a mass of
precipitated coagulant. All of these mechanisms probably play a greater or lesser role in
effective coagulation.
Coagulants used in wastewater treatment include those used in potable water treatment with
some additions. These are: alum, sodium aluminate, ferric chloride, ferric sulfate, ferrous
chloride, ferrous sulfate, lime and organic polyelectrolytes. Other materials such as soda ash
or clays may be used as sources of alkalinity or weighting agents, respectively, to aid
coagulation. When added to water, salts of aluminum or iron react with the water or
alkalinity present in the water to form insoluble hydrolysis products. It is these hydrolysis
products which are the effective coagulating agents. These materials which are positively
charged in the neutral pH range adsorb on the negatively charged wastewater particles
reducing repulsive forces between the particles. The coagulant may also react with other
constituents of the wastewater, particularly anions such as phosphate and sulfate, forming
hydrolysis products containing various mixtures of ions. These various hydrolysis products
differ in their effectiveness as coagulants. The chemistry of the reactions is extremely
complex.
For each combination of coagulant and wastewater there is an optimum dosage of coagulant
and an optimum pH range for coagulation. These are the two parameters which are generally
controlled in operation of advanced waste treatment plants for solids removal through
coagulation.
4.2 Coagulation Control
Because coagulation represents a group of complex reactions, laboratory experimentation is
essential to establish and maintain the optimum coagulant dosage and the effect of im-
portant variables on the quality of coagulation of the wastewater under investigation. With
hydrolyzing coagulants three procedures may be followed for this purpose: the jar test,
measurement of zeta potential, and measurement of phosphate content of the wastewater.
Lime is a special case. Proper control of lime addition may usually be maintained by
measuring the pH or automatically titrating alkalinity after lime addition.
4.Z1 Jar Test
The single, most widely used test to determine dosage and other parameters is the jar test.
The equipment for this test and the directions for its proper performance have been pub-
lished^),(6),(7),(8). The jar test attempts to simulate the full scale coagiilation-flocculation
process and has remained the most common control test in the laboratory since its introduc-
tion in 1918. Since the intent is to simulate an individual plant's conditions, it is not
surprising that there has been no standardization of the test. The jar test as variously
performed does, however, have some elements of conformity. In its essentials, the jar test
simply consists of a series of sample containers, usually six, the contents of which can be
stirred by individual mechanically-operated stirrers. Water to be treated is placed in the
containers and treatment chemicals are added while the contents are being stirred. The range
of conditions, for example, coagulant dosages and pH, are selected to bracket the antici-
pated optima. After a short, 1-5 minute, period of rapid stirring to ensure complete dis-
persion of coagulant, the stirring rate is decreased and flocculation is allowed to continue
for a variable period, 10 to 30 minutes or more, depending on the simulation. The stirring is
then stopped and the floe are allowed to settle for a selected time. The supernatant is then
4-2
-------
analyzed for a variety of parameters. With wastewater the usual analyses are for turbidity or
suspended solids, pH, residual phosphorus and residual coagulant.
If desired, a number of supernatant samples may be taken at intervals during the settling
period to permit construction of a set of settling curves which provide more information on
the settling characteristics of floe than a single sample taken after a fixed settling period. A
dynamic settling test may also be used in which the paddles are operated at 2 to 5 rpm
during the settling period. This type of operation more closely represents settling conditions
in a large horizontal basin with continuous flow.
Several six-position stirrers are available commercially for running jar tests; one from Phipps
and Bird, (Phipps and Bird, Inc., Richmond, Va.),another from Coffman Industries, (Coff-
man Industries, Inc., Kansas City.Ka.), are shown in Figure 4-1. Standard laboratory mixers
have also been used; however, it is difficult to obtain reproducible mixing conditions using
different pieces of equipment. Various types of containers, usually beakers or jars, are used
to hold the samples. Improved mixing may be obtained by adding stationary plates in the
containers as described by Camp and Conklin (8). The Coffman stirrer has an attachment
which makes it possible to add coagulant to all containers simultaneously, however, good
results can be obtained by rapidly adding coagulant from a large graduated pipette to each
jar in sequence.
A simple apparatus, shown in Figure 4-2, can be constructed from tubing, rubber stoppers
and small aquarium valves to permit rapid sampling of supernatant. The unit is placed next
to the sample jars at the beginning of the settling period with the curved stainless steel tubes
dipping into the jars. At desired intervals the vent valve is covered with a finger, permitting
vacuum to draw samples into the small sample bottles. The needle valves are adjusted so that
supernatant is drawn into all the bottles at the same rate. When sufficient sample is ob-
tained, the vent is uncovered and the bottles are replaced with empties. The maximum
sampling rate is about once per minute.
Figure 4-3 shows typical types of settling curves which may be obtained. Curve A indicates a
coagulation which produced a uniformly fine floe, so small that at the end of 1 to 2 minutes
settling, the supernatant had a turbidity equal to that of the starting water due, in part, to
the fine floe which resisted settling. Settling was slow and the final turbidity was excessive.
This coagulation would not be satisfactory in an advanced wastewater treatment plant.
Curve B represents the most common type of settling rate obtained. During the first 5
minutes, the settling rate was practically a straight line on a semilog plot. Settling was rapid
and clarification was satisfactory. The coagulation represented by curve C shows that a
mixture of large rapid settling floe and small, slow-settling particles was obtained. Settling
was rapid for the first two minutes, but with little further clarification after that. High
residual turbidity may also have resulted from incomplete coagulation. Curve D represents
the ultimate in coagulation. Practically all of the floe particles were so large and dense that
97% settled within three minutes. Sedimentation was essentially complete within that time
since only 0.5% additional floe settled in the next 27 minutes. Final clarity of the super-
natant was entirely satisfactory. This coagulation was obtained with a coagulant aid.
Measurement of turbidity provides the most rapid indication of the degree of solids removal
obtained. The recommended procedure for turbidity measurement by light scattering is
given in the 13th edition of Standard Methods for Examination of Water and Wastewater;
however, other methods varying from simple visual evaluation to measurement of light
4-3
-------
Figure 4-1. JAR TEST UNITS WITH MECHANICAL (TOP)
AND MAGNETIC (BOTTOM) STIRRERS
4-4
-------
TWO-HOLE
STOPPER
TO VACUUM SOURCE
SAMPLE
BOTTLE
SAMPLE
TUBE
VENT-
VENT VALVE
MAIN CONTROL VALVE
REGULATING VALVES
PLUG
Figure 4-2. SIX-POSITION SAMPLER
-------
SETTLING TIME-MIN
Figure 4-3. SETTLING CURVES FREQUENTLY OBTAINED
4-6
-------
transmitted on a laboratory spectrophotometer can be used for purposes of comparison.
Measurement of residual suspended solids is the only procedure which gives the actual
weight concentration of solids remaining, but the procedure is too slow for purposes of
process control. Residuals of phosphate and coagulant in supernatant are usually of interest
and may be measured either manually or with automated equipment.
A typical jar test might be run as follows: Wastewater samples are placed in containers and
rapid mix is started at 100 rpm. Selected dosages of coagulant are rapidly added to the
containers covering the expected range of the optimum dosage and a timer is started. If a
polymer is to be used as a coagulant aid, it is added to each jar after two minutes and rapid
mixing is continued for one additional minute. The paddles are then slowed to 30 rpm and
mixing continues for 10 minutes. The paddles are then stopped and the sampling apparatus
previously-described is placed in position. At settling times of 1, 3, 5, 10 and 20 minutes
samples of supernatant are drawn for turbidity measurement. After the final turbidity
sample is drawn, a larger volume of supernatant may be decanted for more complete
analysis. Results are plotted as in Figure 4-4 for judgment as to the desired coagulant
dosage. The jar test may be repeated using a smaller dosage range around the observed
optimum to more closely locate the best dosage.
100
50
U
z
2 20
<
2
>. 10
H
5
3
H
2
SETTLING TIME-MIN
Figure 4-4. JAR TEST RESULTS
"O 24 mg/1
Ferric Sulfate
~—26 mg/1
28 mg/1 ¦
O— 30 mg/1
32 mg/1
mg/1
4-7
-------
If additional alkalinity is required to hold the coagulation in the optimum pH range, this
should be added to the samples ahead of the coagulant. Once an approximate optimum
dosage has been determined, it may be desirable to repeat the jar test with the optimum
coagulant dosage but using varying quantities of added alkalinity to give different pH values.
Experience in coagulating a given wastewater provides the best guide as to the best methods
for controlling the process.
4.2.2 Zeta Potential
Measurement of particle charge is another procedure which may be useful for control of the
coagulation process(9),( 10),( 11). The total particle charge is distributed over two concentric
layers of water surrounding the particle; an inner layer of water and ions which is tightly
bound to the particle and moves with it through the solution and an outer layer which is a
part of the bulk water phase and moves independently of the particle. The surface charge is
not measurable experimentally; however, the zeta potential which is the residual charge at
the interface between the layer of bound water and the mobile water phase can be deter-
mined with a commercially-available instrument.*
In the zeta potential measurement procedure, a sample of treated water containing floe is
placed in a special plastic cell under a microscope as shown in Figure 4-5. Under the
influence of a voltage applied to electrodes at the ends of the cell, the charged particles will
migrate to the electrode having a polarity opposite that of the particle. The velocity of
migration will be proportional to the particle charge and to the applied voltage. The particle
velocity can be calculated by observing the time it takes a particle to travel a given distance
across an ocular micrometer. The zeta potential can then be obtained from a chart which
combines the particle velocity with instrumental parameters. Detailed operating instructions
are supplied with the instrument.
To control the coagulation by zeta potential, samples of water while being mixed are dosed
with different concentrations of coagulant. Zeta potentials are then measured and recorded
for floe in each sample. The dosage which produces the desired zeta potential value is
applied to the treatment plant. Zeta potentials of floe produced in the plant may also be
measured as a means of control. The precise zeta potential which signals optimum coagula-
tion must be determined for a given wastewater by actual correlation with jar tests or with
plant performance as in Figure 4-6. The control point is generally in the range of 0 to -10
millivolts. If good correlations can be obtained between some zeta potential values and
optimum plant performance, then it is possible to make rapid measurements of particle
charge to compensate for major variations in wastewater composition due to storm flows or
other causes. Short term variations such as those due to sudden industrial waste dumps are
still beyond control with any present techniques because of the time lag between recogni-
tion of a problem with coagulation and adoption of a satisfactory change of coagulation
conditions.
4.2.3 Phosphate Monitoring
A third means of coagulant dosage control where the coagulant is being used to precipitate
phosphate as well as to remove solids is to automatically analyze the incoming wastewater
for soluble orthophosphate. The coagulant is then paced to maintain a selected ratio of
*A product of Zeta-Meter, Inc., /V. Y., N. Y.
4-8
-------
Figure 4-5. ZETA POTENTIAL APPARATUS
4-9
-------
+ 10
:
200 300
ALUM DOSAGE, (mg/1)
400
500
Figure 4-6. COAGULATION OF RAW SEWAGE WITH ALUM
4-10
-------
coagulant to phosphate either automatically or by frequent manual adjustment. The Techni-
con Auto-Analyzer* which is commercially available has been adapted for this purpose.
Dow Chemical Co. has developed an automatic system which will add ferric chloride in
proportion to soluble orthophosphate with corrections for varying flow and concentration
of ferric chloride.
Coagulant dosages which produce good phosphate removal will also generally result in good
solids removal if polymers are used to aid in flocculating the fine precipitated matter. Under
normal conditions, the polymer feed rate should be varied with flow to maintain a constant
dosage; however, the polymer feed rate does not have to be changed with changes in dosage
of the inorganic coagulant.
*A product of Ttchnicon Corporation, Tarry town, N. Y.
4-11
-------
4.4 References
1. Black, A.P., "Basic Mechanisms of Coagulation", JAWWA, 52, 492 (1960).
2. O'Melia, C.R., "A Review of the Coagulation Process", Public Works, 100, 87, (May
1969).
3. "State-of-the-Art of Coagulation", A Committee Report to the AWWA Research Com-
mittee, JAWWA, 63, 99 (1971).
4. Faust, S.D., and Manger, M.C., "Electromobility Values of Particulate Matter in Do-
mestic Wastewater", Water & Sew. Wks., 1J_1, 73 (1964).
5. Cohen, J.M., "Improved Jar Test Procedure", JAWWA, 49, 1425 (1957).
6. Black, A.P., Buswell, A.M., Eidsness, F.A., and Black, A.L., "Review of the Jar Test",
JAWWA, 49, 1414(1957).
7. Black, A.P., and Harris, R.H., "New Dimensions for the Old Jar Test", Water & Wastes
Engrg.. 6^49 (Dec. 1969).
8. Camp, T.R., and Conklin, G.F., "Towards a Rational Jar Test for Coagulation",
JNEWWA, 84, 325 (1970).
9. Black, A.P. & Chen, C., "Electrophoretic Studies of Coagulation and Flocculation of
River Sediment Suspension with Aluminum Sulfate", JAWWA, 57, 354 (1965).
10. Riddick, T.M., "Role of Zeta Potential in Coagulation Involving Hydrous Oxides",
TAPP1, 47, 1 71A (1964).
11. Riddick, T.M., Control of Colloid Stability Through Zeta Potential, Vol. 1, Zeta-Meter,
Inc., 1720 First Avenue, New York.N.Y. 10028.
4-13
-------
PHOSPHORUS REMOVAL BY LIME TREATMENT OF SECONDARY EFFLUENT
-------
Chapter 8
PHOSPHORUS REMOVAL BY LIME TREATMENT
OF SECONDARY EFFLUENT
8.1 Description of Process
8.1.1 Theory
Lime treatment of wastewater is essentially the same process as the familiar lime
softening of drinking water supplies. The objectives, however, are quite different. While
softening may occur, the primary objective is to remove phosphorus by precipitation as
hydroxyapatite. This reaction was described in Chapter 3.
During phosphorus precipitation other important reactions occur. The reaction of lime
with alkalinity, that results in calcium removal when carrying out softening, not only
takes place when treating wastewater, but may have a very important effect on the
general efficiency of the process. This reaction can be considered to take place in the
two following ways:
Ca(HC03)2 + Ca(OH)2 + 2 CaCO? + 2 H20
NaHC03 + Ca(OH)2 ~ CaCO? + NaOH + HjO
The first equation is that for softening. Some wastewaters do hot contain enough
calcium, however, for that equation to be satisfied. Calcium carbonate precipitation
may still occur, but by the second reaction. The reactions to form CaCOj are
important for two reasons: the lime consumption determines to a considerable extent
the lime dose required for operating the process, and the resulting CaCO-j acts as a
weighting agent to aid in settling of sludge.
Another reaction that may be important is precipitation of Mg(OH)2 as follows:
Mg2+ + 2 OH" + Mg(OH)2
This reaction does not approach completion until the pH is raised to 11. Magnesium
hydroxide is a gelatinous precipitate which aids colloid removal, but which hinders
sludge thickening and dewatering. Where the pH must be raised to 11 or above to
meet a phosphorus removal requirement or some other treatment objective, Mg(OH)2
formation must be considered.
In the two-stage lime treatment process which will be described below, it is necessary
to include recarbonation after the first stage to reduce the pH and precipitate the
excess lime as CaCOj in the second stage. The following reaction occurs:
Ca2+ + C02 + 2 OH" ~ CaCO? + H20
Carbon dioxide may also be used to lower pH after lime treatment. The important
reaction in this case would be the conversion of Co^2" to HCO3".
8 - 1
-------
A final reaction that is of major concern, where lime is to be recovered by sludge
recalcination, is as follows:
CaC03 A*. CaO + C02
The CaO produced would then be slaked to form Ca(OH)2 before use.
8.1.2 Treatment Systems
Two lime treatment systems may be used with wastewater, single-stage and two-stage.
Figure 8-1 shows a single-stage system. In single-stage treatment lime is mixed with
feed water to raise the pH to a desired value. Although the pH will depend upon the
required phosphorus removal, it is likely to be substantially less than 11, and may be
less than 10. Precipitation of phosphate and other materials, as indicated by reactions
discussed above, takes place. Time is allowed in the appropriate equipment for
precipitate particles to flocculate to sufficient size for good settling. The clarified water
from the settler may be discharged directly or may be filtered to improve solids
removal. Adjustment of pH with CC>2 may be necessary before discharge and will
almost certainly be required before filtration to prevent post-precipitation of CaCOj
from the unstable water. The settled lime sludge may be disposed of as landfill or
may be recalcined for recovery of lime. In the latter case, the sludge is thickened,
dewatered by centrifuge or vacuum filter, and calcined. The calcined product is then
slaked and reused. To avoid buildup of inerts in the lime, some of the sludge or
recalcined lime must be wasted or the inerts must be separated from the sludge before
calcination.
Two-stage treatment is somewhat more complicated than single-stage treatment. In
typical two-stage treatment, shown in Figure 8-2, enough lime is added to the water in
the first stage to raise the pH above 11. Precipitation of hydroxyapatite, CaCOj, and
Mg(OH>2 occurs. Consideration of the solubility product for CaCO^ and the
equilibrium between CO^" and HCO3" shows that the minimum solubility for Ca^+
occurs at a pH of about 10. At a pH of II or above there is a considerable
concentration of Ca present in the water. In two-stage treatment CC>2 is added after
the first-stage settler to bring the pH down to about 10 where CaCO^ precipitation
results. The CaCO^ is settled out and the clarified water is either discharged or sent
to filtration. As in the case of single-stage treatment, pH reduction may be necessary
before discharge and probably would be required before filtration.
8.2 Typical Performance Data
A number of full-scale and pilot-scale tertiary lime treatment plants are in operation
and more full-scale plants are beginning operation. Only one plant, that at South Lake
Tahoe, California, has operated for a significant period with recalcination of sludge for
lime recovery. Most aspects of this 7.5 mgd plant have been discussed in detail by
Culp and Culp (1). Although the data on handling of wastewater sludges for
recalcination and the recalcination process itself are limited mainly to this plant,
8 - 2
-------
WASTE
WASHWATER
FILTER
AID
WASTEWATER
FEED
CARBON
DIOXIDE
RECYCLED LIME
SLUDGE
CARSON
DIOXIDE
MAKEUP
LIIC
SUPERNATANT
CENTRATE
SLAKER
RAPID
MIX
CALCINER
THICKENEI
SLAKED
CENTRIFUl
FLOCCULATOR
RECARBONATO
FILTER
SETTLER
WASTE SLUDGE
LIME DISPOSAL
FIGURE 8-1 SINGLE STAGE LIME TREATMENT SYSTEM
-------
WASTE
WASHWATER
FILTER
WASTEWATER FEED
CARBON
DIOXIDE
CARBON
DIOXIDE
SLUDGE
SLUDGE
WASHWATER
SLUDGE TO RECALCINATOR
SLAKER
RAPID
MIX
FLOCCULATOR
RECARBONATOR
RECARBONATOR
FILTER
SETTLER
SETTLER
OR DISPOSAL
FIGURE 8-2 TWO STAGE LIME TREATMENT SYSTEM
-------
performance data on other parts of a lime treatment system are available from more
than one location. Extensive data have been reported from pilot plants located in
Washington, D. C., (2) and Lebanon, Ohio (3).
The lime treatment system at Tahoe is a two-stage clarification system which is
followed by pressure filters of the multimedia type. These contain 3 ft. of a mixture
of coal, sand, and garnet. The water passes through two filters in series. The first stage
clarifier is operated at a pH of 11. To reach that pH requires about 300 mg/1 CaO. A
small amount of polymer is used to improve flocculation. The pH is reduced to 9.6 by
recarbonation before the second-stage settler. Before filtration, the pH is further
reduced to 7.5. A small amount of alum (1 to 20 mg/1) or a combination of alum
and polymer is used as filter aid. Filter run length has varied between 4 and 60 hours.
Phosphorus removal at this plant always has been good and has improved as operating
experience has increased. By returning plant waste streams containing precipitated
phosphorus to the first stage flocculator, it is now possible to obtain routinely an
effluent with less than 0.1 mg/1 P. Before the filters the phosphorus concentration is
about 0.4 mg/1.
In addition to phosphorus removal, there is significant removal of organic materials.
During a period of intensive study of the first stage of clarification (unpublished data)
77% removal of BOD and 61% removal of COD were obtained. Although there was
not a large removal of suspended solids, there was a significant change in the character
of the solids from organic to largely inorganic. Further removal of organic materials
and suspended solids occurs over the remainder of the system. Culp and Culp report
typical filter effluents with a BOD of 3 mg/1, COD of 25 mg/1, and turbidity of 0.3
JTU.
Sludges from the first stage settler and the settler following recarbonation are sent to a
gravity thickener. Solids concentration increases during thickening from about 1% to
from 8 to 20%. Thickened sludge is then centrifuged to form a cake with from 30%
to more than 40% solids. The cake is next calcined in a multiple hearth furnace and
the recovered lime slaked for reuse. Initially it was planned to operate the centrifuge
for high solids recovery in the cake. This necessitates discarding a significant amount
of the recovered lime to prevent buildup of precipitated phosphate and other inerts.
Since operation began, it has been found that much of the phosphate and some other
inert materials can be. separated from the CaC03 in the sludge by operating with a
rather high solids concentration in the centrate. Approximately 90% of the phosphorus
can be removed in the centrate, for example, when 25% of the solids entering the
centrifuge are allowed to remain in that stream. This classification procedure results in
a loss of about 15% of the recoverable lime. A greater degree of utilization of
recalcined lime compensates for the loss, however, and the load on the calcining
furnace is reduced. Over three years of operation, mostly without classification in the
centrifuge, the average concentration of CaO in the recalcined product was 66%.
Recalcined lime made up 72% of the total used at the plant.
8 - 5
-------
Considerable operating data have been obtained by O'Farrell and Bishop (2) on a
two-stage system operated at the EPAWQO-DC Pilot Plant in Washington, D. C. This
50,000 gpd plant treated secondary effluent from a pilot activated sludge system at
the municipal treatment plant. The first stage pH was maintained at about 1 1.7, using
a lime dose of about 400 mg/1 as CaO, and the pH after recarbonation was
maintained at about 10.3. Ferric iron was added at a rate of 5 mg/l in the second
stage to improve flocculation. Water from the second-stage settler was filtered without
further pH adjustment through gravity-flow dual-media filters consisting of anthracite
coal over sand. The average filter run length was more than 50 hours.
Phosphorus removal for the system was similar to that obtained at Tahoe; 0.09 mg/l
as P was the average concentration remaining. Phosphorus concentration before
filtration averaged 0.13 mg/l.
In addition there was significant organic and suspended solids removal. The average
BOD was reduced from 15 mg/l to 2.1 mg/l before and 1.5 mg/l after filtration.
Suspended solids were reduced from 33 mg/l to 17 mg/I before filtration and 3.8
mg/l after filtration.
Sludge from the second stage settler was returned to the first stage and all sludge
removed from the system was from the settler of the first stage. This sludge contained
about 5% solids and constituted about 1.5% of the feed volume. Although there was
sludge thickening and calcining equipment available, only preliminary study was made
of lime reuse. Limited results showed improved thickening of the sludge when
recalcined lime was recycled to the system.
A 75 gpm single-stage pilot system was operated at the Lebanon, Ohio Sewage
Treatment Plant. Feed water was activated sludge effluent. This system, consisting of a
flocculator-clarifier followed by dual-media filters of anthracite coal and sand, is
described by Berg, et al. (3). Sludge was gravity thickened and discharged to sand
drying beds. The system has been operated over a pH range from about 9 to 11 with
most data taken at a pH of 9.5. Before filtration the pH was reduced to about 8.8
with sulfuric acid to prevent precipitation on the filter media. Filter run length
averaged about 90 hours.
Effluent phosphorus concentrations for the system are shown in Figure 8-3. The effect
of pH is clearly indicated. Since the average phosphorus concentration of the
secondary effluent was about 10 mg/l as P, approximately 95% removal was obtained
at a pH as low as 9.5. Before filtration the average phosphorus concentration was 0.75
mg/l at a pH of 9.5. Removal improved only slightly at higher pH because of the
presence of precipitated phosphorus in the effluent.
Average suspended solids concentration of the secondary effluent was 43.5 mg/l. This
was reduced to 16.5 mg/l in the clarifier. Much of the suspended matter in the clarifier
effluent consisted of inorganic precipitates. After filtration the water was of high
clarity. During a six month period when the biological treatment plant operated well,
turbidity averaged 0.2 JTU. Even when the suspended solids load from the activated
8 - 6
-------
••
10.5
9.0
9.5
10.0
11.0
5
CLARIFIES pH
FIGURE 8-3 EFFECT OF pH
ON PHOSPHORUS CONCENTRATION
OF EFFLUENT FROM FILTERS
FOLLOWING LIME CLARIFIER
8 - 7
-------
sludge plant was high, the lime treatment system produced water with a daily average
turbidity not exceeding 2.5 JTU.
During a two month period in which extensive measurement of organic removal was
made, total organic carbon was reduced by 63% over the clarifier and 68% over the
clarifier and filters. Organic carbon measurements taken at other times indicated
removals of from 55 to 74%. Occasional BOD and COD samples indicated removals of
86% and 62%.
Sludge was usually removed from the settler at 1.5 gpm giving a sludge concentration
of 2.8% solids. This was thickened to about 10% before being pumped to drying beds.
The sludge dewatered quickly to about 50% solids. Equipment was not available for
recalcining the sludge.
The lime requirement is an important consideration in lime treatment. This can vary
over a wide range depending on operating pH and water composition. From reactions
given earlier it was seen that alkalinity has an important effect on the lime dose.
Buzzell and Sawyer (4) have shown, for example, that for several wastewaters the lime
dose required to reach a pH of 11 correlated approximately with alkalinity.
Examination of the earlier equations would show also that calcium hardness can affect
lime dose. One part by weight of CaO can react with from 0.89 to 1.79 parts of
bicarbonate alkalinity expressed as CaCO^, the lower value applying to very soft waters
and the higher value to very hard waters. In addition to the reaction of lime with
hardness, other competing reactions occur in lime treatment of wastewater. Also, there
may be incomplete reaction of the lime. All of these complications make calculation
of lime dose difficult. The result is that, at present, determination of lime dose is
largely empirical. Approximate values have already been given for the plants at Tahoe
and Washington, D. C. Some approximate values for the Lebanon work are shown in
Table 8-1. It appears from the data already available that the lime dose will usually be
in the range of 300 to 400 mg/1 as CaO for two-stage treatment, and from 150 to
200 mg/1 where single-stage treatment is satisfactory.
Table 8-1
LIME REQUIREMENTS
Feed Water
Alkalinity
Clarifier pH
Approximate Lime
Dose
(mg/1 as CaCO-j)
(mg/1 of CaO)
300
9.5
185
300
10.5
270
400
9.5
230
400
10.5
380
8-8
-------
8.3
Criteria for Selection of Process
Lime treatment of secondary effluent represents a significantly higher capital cost and
somewhat higher operating cost for phosphorus removal than mineral addition to a
conventional treatment plant. Lime treatment has, however, several advantages over
mineral addition. It can be considered more dependable since it adds additional
flocculation and sedimentation steps to the system. Upsets in the conventional plant,
which would reduce the efficiency of phosphorus removal by mineral addition, would
have less effect on tertiary lime treatment. Because tertiary lime treatment is separate
from the conventional treatment plant, it adds flexibility to operation of the system.
For very high degrees of phosphorus removal, lime treatment would be the method of
choice, not only because of the inherent greater dependability mentioned above, but
because of the ability to produce an effluent slightly lower in phosphorus content.
Lime treatment decreases the total dissolved solids content of the water by removal of
hardness and alkalinity while mineral addition adds to the total dissolved solids. In the
case of alum addition, for example, 5.3 parts by weight of sulfate are added for each
part of aluminum. Lime treatment has the capability of removing turbidity to very low
levels. Where there are plans for recreational reuse or certain industrial reuses of the
treated water, low turbidity along with low phosphorus content of the lime treatment
effluent is very desirable. Lime treatment offers the opportunity for recovery of the
treatment chemical. At the present time there are no acceptable methods for recovery
of aluminum or iron salts.
The choice of single-stage or two-stage lime treatment depends partly upon the degree
of phosphorus removal required, but more importantly, on the alkalinity of the water.
Unless a high treatment pH is used, waters with low alkalinity, in the range of 150
mg/1 as CaCO^ or less, form a poorly settleable floe because of the low fraction of
dense CaCO^. A pH above 10 is needed even to obtain measurable CaCO-j production.
A pH of 11 or above is then needed to precipitate Mg(OH)2 which aids in settling of
fine particles. Since neither discharge nor reuse of the high-pH water is likely to be
acceptable, addition of a second treatment stage after recarbonation to a pH of about
10 becomes generally necessary. It would be possible to recarbonate to a much lower
pH and avoid the second treatment stage, but this would result in a high calcium
effluent and would eliminate the chance to produce high CaCO^ sludge in situations
where lime recovery was contemplated. Although most of the phosphorus removal
occurs in the high pH first stage, in accordance with the pH dependence of
phosphorus solubility as shown earlier by Figure 8-3, some removal does also occur in
the second stage. Another important reason for including the second stage is to assure
better control of clarification. With low alkalinity waters there is sometimes difficulty
in settling the sludge in the first stage even at high pH. The second stage settler with
its high CaCC^ sludge, prevents solids carryover when the first stage settler does not
operate properly.
With high alkalinity waters, a well settling floe is formed at pH values as low as 9.5.
There is no need for two-stage treatment unless the required degree of phosphorus
removal necessitates a high pH. In addition to obtaining a small degree of phosphorus
8-9
-------
removal, the second stage would be used in these cases to lower calcium content in
the effluent and to obtain high CaC03 sludge for lime recovery. There is not yet
available operating experience at enough plants to state positively the alkalinity at
which single-stage treatment will perform satisfactorily, from the standpoint of floe
settleability. At present, experience indicates that at an alkalinity of 150 mg/1 as
CaCO^, single-stage treatment probably cannot be used. Between 150 mg/1 and 200
mg/1 settleability will probably depend on the amount of organic floe present. Above
200 mg/1, settleability is likely to be satisfactory.
8.4 Description of and Criteria for Choice of Equipment
8.4.1 Single-Stage System and First Stage
of a Two-Stage System
Diagrams for single and two-stage lime treatment systems are shown in Figures 8-1 and
8-2. The first part of each system is made up of clarification equipment essentially the
same as that found in water treatment plants. It includes a chemical feeding system
(see Chapter 10), a rapid mix tank, a flocculator, a settler, and a means for sludge
removal. Mixing, flocculation, and settling may be carried out in separate vessels or
tanks, or they may be combined in one integrated unit. The system consisting of
separate tanks has been used for many years in water treatment and was the system
adopted for the first stage of the Tahoe lime treatment plant. The integrated type of
equipment is sometimes referred to as an upflow clarifier or a sludge blanket clarifier.
The latter term can be misleading, however, since not all such units operate with a
sludge blanket. A diagram of a unit that does not operate with a sludge blanket is
shown in Figure 8-4. Culp and Culp (1) recommend the system of separate tanks
because that system allows separate control of each part of the process. They point
out that such a system has greater flexibility in points of addition for chemicals. They
also point out that in the integrated units with a sludge blanket there may be
difficulty in controlling the blanket height and the blanket may become anaerobic,
leading to poor phosphorus and solids removal. Pilot studies at other locations have
shown that excellent solids removal can be obtained with sludge blanket equipment,
but that blanket instability could be a problem. On the other hand, pilot upflow units
designed to operate without a sludge blanket have proved to be very effective at
Washington, D.C. (2). Results available at this time indicate that both the system of
separate tanks and the integrated units without sludge blankets should be considered
for design of new plants.
The Tahoe plant is an excellent example of a system with the first stage consisting of
separate tanks for each unit process. At the present time the plant must be relied
upon heavily as a guide to design of such systems. Description of the equipment and
some design parameters are given by Culp and Culp. The rapid mix tank which is
located in one corner of the flocculator has about a 30 second residence time at
design flow. (The plant usually runs with a daily average flow of about one third of
design flow.) Mixing is accomplished with a vertical shaft mixer. The flocculator is a
8 - 10
-------
WASTEWATER LIME
EFFLUENT
SLUDGE
FIGURE Q-n TYPICAL UPFLOW CLARIFIER
8-11
-------
square tank with a depth of 8 ft. At design flow the residence time is 4.5 minutes.
The flocculator is provided with air agitation, but operating experience has shown that
this agitation is not necessary. The settler is circular with a center inlet. The depth is
10 ft and the design overflow rate is 950 gpd/ft . There are two sludge pumps, one a
variable-speed centrifugal, and the other a positive displacement Moyno. Provision has
been made to return part of the sludge to the rapid mix tank. The concept of
returning sludge to the section of the clarification equipment where precipitation is
taking place is called solids contact. The objective is to hasten precipitation and to
obtain larger precipitate particles which will settle well.
The ease with which flocculation occurs at Tahoe would suggest that a flocculator for
this service presents no design problems. Experience is, however, limited. Pilot testing
would be very desirable to determine the flocculating characteristics of a particular
wastewater. Jar tests may be of some value. Unfortunately, the effect of solids
contact, which experience at Tahoe and elsewhere indicates is beneficial, is difficult to
duplicate in a jar test. If jar tests indicated good flocculating characteristics without
benefit of solids contact, however, flocculation should be as good or better with solids
contact. When pilot studies are carried out, provision should be made for solids
contact.
Comments concerning the effect of solids contact in jar tests also apply to settling
rates determined from these tests. Rapid settling without benefit of solids contact
would be a strong indication that settling with solids contact would also be rapid.
Clarification equipment of the integrated type for use in municipal water treatment is
manufactured by a large number of companies. Units of this type used for treatment
of wastewater have been essentially of the same design as used for municipal water
supplies. The basins are usually circular with the rapid mix and flocculating sections
located at the center of the settler. Agitation for mixing and flocculating as well as
power for sludge scrapers is provided at the center of the settler. Solids contact is
usually provided, and may be carried out with best control by external circulation of
sludge to the mixing section. In some designs, however, internal recirculation is used.
There is considerable variation in the geometry and complexity of the mixing and
flocculating sections depending upon the manufacturer. The unit shown in Figure 8-4
is a relatively simply type. Units from three different manufacturers were used in pilot
work at Lebanon, Ohio and Washington, D. C. These and others at additional locations
have performed effectively. However, in some cases with operation of a sludge blanket,
difficulties have been encountered. A sludge blanket can be eliminated by avoiding
designs in which the wall separating the center compartments from the concentric
settler reaches close to the bottom of the settler. The design in Figure 8-4 prevents
sludge blanket formation.
There has been little effort to modify this type of equipment from use in water
treatment to use with wastewater. Experience at Tahoe would suggest, for example,
that less flocculation time, and probably less agitation, is required than is generally
provided. A unit such as that shown by Figure 8-4 may have about 40 minutes
flocculation time at design capacity for water treatment, nearly nine times that at
8 - 12
-------
Tahoe. Whether modifications in such equipment would be worthwhile remains to be
seen.
If a designer chooses to use equipment that is on the market, he does not have
complete flexibility in specifying the design. Fortunately, equipment satisfactory for
water supply treatment has proved satisfactory for wastewater, except that a lower
settler overflow rate is necessary. The single-stage treatment system at Lebanon, Ohio
was operated at a constant overflow rate of 1,440 gpd/ft . This proved satisfactory
even with a sludge blanket. At Washington, D. C., tests were made with settler rates as
high as 1,950 gpd/ft^ although the average rate was 1,120 gpd/ft^. For design, a peak,
dry weather overflow rate of 1,200 to 1,400 gpd/ft^, with the average somewhat lower,
appears reasonable.
8.4.2 Recarbonation
In a two-stage system recarbonation follows the first stage settler. In a single-stage
system and following the second stage of a two-stage system, pH reduction by
recarbonation may be necessary to make the effluent suitable for filtering or for
discharge. An excellent discussion of all aspects of wastewater recarbonation is given
by Culp and Culp (1). The reader is referred to that publication, as well as one by
Haney and Hamann (5), especially for information on sources of carbon dicixide.
Sources include stack gas from sludge incinerators and lime recalciners, CC>2 generators,
and commercial liquid CC^-
The equipment used for contacting the CC>2 and water may be simply a tank with the
gas being bubbled through the water. This is the system used at the Tahoe plant
where the height of water over the CC>2 source is 8 ft. Pilot studies at Washington,
D.C., with a tank depth of 11 ft and a turbine mixer to reduce bubble size and
distribute bubbles, showed almost 100% absorption of CC^-
When considering residence time, the recarbonation tank following the first stage of a
two-stage system must be differentiated from the recarbonation tank used just to
reduce effluent pH. The residence time of the recarbonator between the first and
second stages is not particularly important. At Tahoe the recarbonation tank is only
large enough to give a 5 minute residence time at design flow. In the Washington,
D.C. studies, the residence time was 15 minutes. A problem that may arise when
recarbonating wastewater is foam formation. This is most pronounced when the flow
of bubbles is concentrated in parts of the recarbonation tank. The CC>2 distributing
system should cover as well as possible the whole lateral area of the tank bottom. This
would be especially true if the residence time were as low as 5 minutes.
In the recarbonation tank used just for effluent pH adjustment, sufficient residence
time must be allowed for completion of reactions taking place. Culp and Culp
recommend 15 minutes. In the Tahoe plant only 4 minutes is provided at design flow,
but the water flows to storage ponds with a much longer residence time before the
water is filtered. Just as in the case of recarbonation between stages of a two-stage
system, good bubble distribution is important for prevention of excessive foaming.
8 - 13
-------
Since one reason for using a two-stage system is to prevent a high calcium
concentration in the effluent, the pH to which the water should be recarbonated
between stages is that for maximum conversion of the calcium to CaCO^. Where lime
recovery is practiced, maximum formation of CaCO^ is also desirable because it results
in maximum CaO production. The optimum pH is about 10. At Tahoe, for example,
10.3 was selected because tests showed that pH to give maximum CaCO-j production.
The pH to be selected when the objective of recarbonation is just pH reduction, will
depend on a number of factors. An effluent standard may determine this pH. If
stabilizing the water to prevent post precipitation is the objective, the pH may be
determined by consideration of the Langelier Index. If, as in the case of Tahoe, a
treatment step such as activated carbon adsorption is to follow lime treatment, this
will determine pH. In the latter case a pH of about 7.5 has usually been selected.
The CC>2 dose requirements can be calculated from the chemical reactions taking place
and a knowledge of the concentrations of the various forms of alkalinity in the water.
For design of a new plant, alkalinity data must be obtained from samples of liquor
taken from jar tests run at the same pH values as planned for the plant. Results must
be considered very approximate, but they do help in sizing CC>2 feeding equipment.
Culp and Culp discuss the calculations in detail. It was found in pilot plant work at
Washington, D. C. that the CC>2 dose for recarbonation between the stages of a
two-stage treatment system can be calculated reasonably well by considering the
reduction in calcium concentration that occurs. This reduction is due to CaCO?
formation with the CO-j coming from the CC^- The calcium content before
recarbonation can be determined satisfactorily from jar tests run at the desired
operating pH. The calcium content after recarbonation can also be determined
approximately by jar test. Results from several pilot plants indicate the soluble calcium
content at the pH of minimum solubility should be about 40 mg/1 as Ca. This figure
is probably just as good for calculations as a figure obtained from a jar test. The CC>2
dose in mg/1 is then equal to:
/44\
(Ca reduction in mg/1) — .
\40/
A safety factor of about 20% should be added to the calculated dose to compensate
for inefficiency in absorption.
8.4.3 Second Stage Flocculation and Sedimentation
In two-stage treatment, recarbonation is followed by the second stage flocculation and
settling. Experience with equipment for the second stage of treatment with wastewater
is limited. Two very different systems have been tested. The simplest is the system at
Tahoe. In the Tahoe plant the equipment used is just a longitudinal settler with a 30
minute detention time and a 2,400 gpd/ft^ overflow rate at design flow. No
flocculation equipment is provided. Culp and Culp (1) refer to the settler as a reaction
and settling basin. It has been found at Tahoe that a significant part of the
recarbonation reaction actually occurs in the settler. This is shown by the pH decrease
8 - 14
-------
of about 0.7 unit that occurs between the recarbonation tank and the settler outlet.
The clarification that results at Tahoe is unusually good for the simplicity of the
equipment.
Jar test and pilot plant work at Washington, D. C. indicated that flocculation after
addition of a flocculating aid was required to obtain satisfactory clarification of that
water. The pilot tests were run in upflow equipment of the integrated type discussed
earlier. Without a flocculating aid, fine CaC03 precipitate escaped the settler, even
at a settler overflow rate as low as 1,100 gpd/ft^. Using 5 mg/1 of Fe^+, good
operation was observed at overflow rates as high as 1,950 gpd/ft^. Hydraulic
limitations prevented testing at higher rates. Apparently, the good contact between
solids and water during flocculation was beneficial in reducing CaCO^ supersaturation
in the effluent. Precipitation in the filters following the second stage did not occur,
and no pH reduction was required.
In this work the sludge from the second stage was returned to the mixing section of
the first stage to serve as a weighting agent. In waters of low alkalinity or expected
high loads of biological solids, this procedure should be considered.
It is difficult to recommend the amount of flocculation to provide and the overflow
rate to use for second-stage treatment. Jar tests would be of little value because of the
complication of first raising the pH with lime and then lowering with CC>2. Pilot
testing may be desirable. Such a pilot system would have to include the first stage
equipment and recarbonation equipment, although these could be of crude design. In
the absence of pilot tests to indicate otherwise, flocculation should be provided in the
second-stage system. If filtration is included, a peak dry weather overflow rate of up
to 2,000 gpd/ft^ in the settler may be used. If filtration is not provided, a somewhat
lower overflow rate should be chosen to assure reasonable effluent quality during periods
when an upstream part of the system is not operating properly.
8.4.4 Filtration
The last step in a complete lime treatment system is filtration. Although there are a
variety of filter types that could be used, essentially all the recent test work has been
done with downflow filters of multimedia type. These are essentially the same in
design as a rapid sand filter, except that the media are graded from coarse at the filter
surface to fine at the filter outlet. This is accomplished by using media of different
densities with the largest particles being composed of the least dense material. The
result is a filter bed which has far more capacity for removing suspended solids than
an ordinary sand filter, without impairing effluent clarity. Both gravity and pressure
filters have been used. Each has certain advantages. Generally, however, pressure filters
are more appropriate at smaller plants.
Two media combinations have been tested extensively. These are dual-media of coal
and sand, and tri-media of coal, sand, and garnet. Filters used at Lebanon, Ohio
following a single-stage system (3) and filters at Washington, D. C. following a
two-stage system (2) were of the dual-media type. Both contained 18 in. of anthracite
8 - 15
-------
coal over 6 in. of sand. Average particle sizes of the media at Lebanon were 0.75 mm
and 0.46 mm; at Washington, D. C., 0.9 mm and 0.45 mm. At 2 gpm/ft the average
run length at Lebanon was about 90 hours using 8 ft of water as the terminating
pressure loss. At Washington, D. C., filter run length averaged about 50 hours for an
average rate of 3.4 gpm/ft^ and a terminating pressure loss of 7 ft of water. High
clarity waters were obtained in each case. Backwash in both cases was carried out at
20 gpm/ft . Backwash time was 5 minutes at Lebanon and 10 minutes at Washington,
D. C. in the latter case, a surface wash was included.
These run length figures give a rough idea of the results that may be expected at
other locations. For maximum run length at a given product quality, the depths of
media and their particle sizes should be optimized. Although this was not done for the
above-mentioned filters, pressure loss distribution in the filters at Lebanon as reported
by Berg, et al. (3) indicates that good solids storage in the anthracite was being
obtained. Small pilot filters could be used for optimization of the media.
A reasonable design rate for gravity-flow dual-media filters appears to be about 3
gpm/ft . Higher rates could result in inconveniently short runs during periods of high
solids load. Rates much lower than 3 gpm/ft^ result in excessive filter costs, not
justified by the longer filter runs. Very long filter runs could result in backwash
problems from biological activity in the filters. Recommendation of a 3 gpm/ft^ rate is
based upon having good operation of the second-stage settler. Frequent settler upsets
must be avoided.
Tri-media filters are used at the Tahoe plant. These are pressure filters which at design
flow operate at 5 gpm/ft^. Each bed holds 3 ft of mixed coal, sand, and garnet as
supplied by Neptune Microfloc. Two filters are used in series, and are backwashed in
series, usually when the head loss reaches 16 ft of water. Run length has varied from
4 hours during heavy solids load to about 60 hours under good conditions. Backwash
is carried out at 15 gpm/ft^. The backwash water is reprocessed through the lime
treatment system. More details are given by Culp and Culp (1).
From the available data, it is difficult to give precise criteria for choosing between
dual-media or tri-media filters following lime treatment. The use of garnet allows for a
smaller particle size at the bottom of the filter than is possible with sand. In case of
floe weakness, the tri-media filter offers, therefore, more protection from turbidity
breakthrough. The cost, however, is slightly higher. Where the tri-media fill is not used,
filters can still be backwashed at the onset of turbidity breakthrough. Experience with
dual-media filters on lime treated water has not shown sudden breakthrough to be an
important problem. Where water of the highest clarity is required, tri-media filters may
be of most value.
8.4.5 Sludge Handling
The sludge from a lime treatment system may be handled in two general ways. It may
be thickened, dewatered, and disposed of or it may be thickened, dewatered, and
8 - 16
-------
recalcined to recover lime for reuse. For small plants, recalcination will not be economically
competitive with disposal and should not be considered unless there are restrictions
on disposal which make that alternative difficult. It can be assumed generally that
for plants over 10 mgd, recalcination of lime sludge will be practical. Recalcination
may also be practical at plants somewhat smaller than 10 mgd, depending on the cost
of purchased lime and other local conditions.
Data given earlier indicate that the volume of sludge from lime treatment will vary
from 1.5% to several percent of feed volume. Sludge concentration will probably be in
the range of 1% to 5%. The actual weight of sludge will vary with the chemical
composition and amount of suspended solids in the feed water. Values have been
observed in the range of 4 to 7 lb/1,000 gal. Sludge production from the first stage of
a two-stage system or from a single-stage system can be estimated approximately by
weighing the dried sludge from jar tests run at the planned operating pH. Additional
sludge produced in the second stage can be assumed to be the CaCOg formed from
calcium concentration reduction during recarbonation. Calcium content before
recarbonation can be obtained from the above-mentioned jar tests and after recarbonation
can be assumed to be 40 mg/1 as Ca.
Design information for the Tahoe sludge handling system is reported by Culp and
Culp. (1) The reader is referred to that publication for further information. The Tahoe
plant is the only one which includes recalcination of sludge from the lime treatment
of wastewater, and that has been operated for a long period of time. Paramenters from
that system can be used as a rough guide for design. The gravity thickener has a
bottom scraper mechanism and is 8 ft deep. Overflow from the thickener is returned
to either the primary clarifier or the first stage of lime treatment. The design solids
loading is 200 lb/day/ft^ and the design overflow rate is 1,000 gpd/ft^. Some
preliminary data from pilot work at Washington, D. C. suggest these loadings are high.
The Tahoe equipment was sized, however, to handle a volume of sludge equal to
about 9% of the plant design flow. The actual volume of sludge should usually be less
than one third of that volume. At Washington, D.C., recalcined lime was found to
produce a sludge which thickened significantly faster than sludge from virgin lime.
If lime is not to be recovered from the sludge, the underflow from the thickener can
be placed directly on drying beds for final dewatering, or it can be dewatered by
centrifuge or vacuum filter.
If sludge recalcination is planned, the thickened sludge would be dewatered, by
centrifuging or filtering and fed to the calcining furnace. The centrifuge may have
advantages over filters for this purpose, such as the ability to separate phosphate
sludge from CaCOj. Lime from the furnace is stored for reuse. It, along with any new
lime, must be slaked before use. Sludge can be pumped from the thickener to the
centrifuge. Cake from the centrifuge must be transported by conveyor. At Tahoe the
centrate is sent to the primary clarifiers. It has been found at Tahoe that, by
operating the centrifuge with less than maximum solids capture, much of the
precipitated phosphorus can be retained in the centrate. This results in a higher
g - 17
-------
quality lime. This mode of operation requires a second centrifuge to remove the
phosphorus rich solids from the centrate. These solids may have value for use in
fertilizer.
The Tahoe dewatering and recalcining system is described by Culp and Culp. The
centrifuge is a 24 by 60 in. concurrent flow type. The calciner is a 14 ft-3 in. diameter,
6 hearth furnace operated at a top temperature of 1,850° F. Other types of
furnaces have been used in water treatment plants and these should also be applicable
to the sludge from wastewater treatment. The reader is referred to equipment
manufacturers for further information about centrifuges and furnaces.
8.4.6 Control of Lime and Carbon Dioxide Feed
The feeding of lime to a lime softening system can be controlled in a number of
ways. For the wastewater application, the most appropriate appears to be control of
pH, with the pH value being selected for good suspended solids removal or to meet a
phosphorus removal requirement. Where flow equalization is employed, pH sensing
alone should be sufficient. Where there is substantial diurnal variation in flow, better
control can be maintained by flow proportional-pH control. Systems are available
for this type of control.
Carbon dioxide dose is also controlled by pH. As in the case of lime feed, flow
proportional-pH control would be preferred where there is diurnal variation in flow.
There is a tendency for pH electrodes to become coated with precipitates and lose
sensitivity. The precipitate can be dissolved with acid. In some instances, however, it
has been found satisfactory to use manual control because of difficulties with the pH
control system.
8.4.7 Scale Formation on Equipment
Because the water in a lime treatment system is supersaturated to some degree with
CaCOj and other precipitating substances, there is a tendency for scale to form on
equipment and pipe surfaces. The problem is particularly serious with the lime slurry
from the slaker. This can quickly plug the slurry line to the rapid mix tank. There
should be easy access to all parts of this line for cleaning. At one small plant a
flexible hose was used to feed lime slurry. By periodically flattening or flexing the
hose, the scale was removed. The mechanical mixer in the rapid mix tank will also
become scaled and must be cleaned.
Scale can also form in sludge lines and the effluent line from the first-stage clarifier.
It is recommended that open troughs be used wherever possible. Provision should be
made for cleaning lines when open conduits are not possible.
Possible scaling of filters has already been mentioned. Recarbonation before the filters
will minimize scale formation.
8 - 18
-------
8.5 Capital and Operating Costs
Because of the short history of lime treatment of wastewater, there is a scarcity of
capital and operating cost information. Costs are available for the Tahoe plant and are
reported by Culp and Culp (1). Capital cost for the 7.5 mgd lime treating facility was
$1,115,000 and cost of the filters was an additional $705,000. For 1969 the estimated
operating costs exclusive of equipment amortization were 7.3^/1,000 gal. for lime
treatment without filtration and 2.81/1,000 gal. for filtration. Amortization of
equipment, based on costs adjusted to the 1969 national average, interest of 5% for 25
years, and the assumption of the plant operating at full capacity, would add
2.7^/1,000 gal. for treatment without filtration and 1.8<£/1,000 gal. for filtration. For
the plant operating at full capacity the total cost of operation would then be
10.01/1,000 gal. without filtration and 14.61/1,000 gal. including filtration. The
fraction of the cost resulting from amortization is a substantial 31% even with the
assumption of full capacity. There is an obvious need to keep equipment size to a
minimum. Flow equalization deserves strong consideration when planning for lime
treatment of secondary effluent to minimize the initial cost of new equipment.
Additional costs for lime treatment based on information from Tahoe and other
sources have been reported by Smith and McMichael (6). Tables 8-2 and 8-3 show
Table 8-2
CAPITAL COST OF LIME TREATING FACILITIES
Treatment
L0
Cost ($)
Plant Size (mgd)
20
100
Single-Stage without Filtration
100,000
1,200,000
5,500,000
Two-Stage without Filtration
160,000
1,500,000
7,900,000
Dual-Media Filtration
110,000
Table 8-3
510,000
2,300,000
TOTAL COST FOR LIME TREATMENT OF WASTEWATER
Treatment
LP
Cost (if 1,000 gal.)
Plant Size (mgd)
10
100
Single-Stage without Filtration
13
7
4
Two-Stage without Filtration
16
9
6
Dual-Media Filtration
8
3
1.4
costs based largely on their results. Capital costs have been updated to December,
1970. Amortization is at 6% for 25 years. For 1 mgd plants, recalcination equipment is
8 - 19
-------
not included. All lime would be purchased. The applicability of recalcination was
discussed earlier. Component costs making up the total cost figures represent an
average for the whole country. Local conditions may cause significant deviation from
these values.
8 - 20
-------
8.6 References - Chapter 8
1. Culp, R. L.( and Culp, G. L., Advanced Wastewater Treatment, Van Nostrand
Reinhold Company, New York (1971).
2. O'Farrell, T. P., and Bishop, D. F., "Lime Precipitation in Raw, Primary, and
Secondary Wastewater", Presented at the 68th National Meeting of AIChE,
Houston, Texas (March 1971).
3. Berg, E. L., Brunner, C. A., and Williams, R. T., "Single-Stage Lime Clarification of
Secondary Effluent", Wat. and Wastes Eng., 7:3, p 42 (1970).
4. Buzzell, J. C., and Sawyer, C. N., "Removal of Algal Nutrients from Raw
Wastewater with Lime", JWPCF, 39:10, Part 2, p R16 (1967).
5. Haney, P. D., and Hamann, C. L., "Recarbonation and Liquid Carbon Dioxide",
JAWWA, 61:10, p 512 (1969).
6. Smith, R., and McMichael, W. F., "Cost and Performance Estimates for Tertiary
Wastewater Treating Processes", Robert A. Taft Water Research Center, Report No.
TWRC-9 (June, 1969).
8 - 21
-------
LIME RECOVERY AND REUSE
-------
LIME RECOVERY AND REUSE
by
Robert B. Dean
Chief, Ultimate Disposal Research Program
Advanced Waste Treatment Research Laboratory
National Environmental Research Center
Cincinnati, Ohio U5268
Riverside, California
March 2b, 1972
-------
LIME RECOVERY AND REUSE
Contentb
1. Lime Recovery Aspects of Advanced Waste Treatment as Practiced at
South Tahoe, through page 369.
2. Lime Equations.
3. Selected Graphs from Mulbarger et al., December 19&9, "Lime Clarification,
Recovery, Reuse, and Sludge Dewatering Characteristics," JWPCF 4l(l2),
2070-85.
h. Cost data prepared "by R. Smith, 19&9> Figures b and 5.
-------
LIME RECOVERY AND REUSE ASPECTS
of
ADVANCED WASTEWATER TREATMENT
AS PRACTICED AT SOUTH TAHOE
by
SOUTH TAHOE PUBLIC UTILITY DISTRICT
SOUTH LAKE TAHOE, CALIFORNIA
for the
WATER QUALITY OFFICE
ENVIRONMENTAL PROTECTION AGENCY
Project 17010 ELQ (WPRD 52-01-67)
August 1971
For salt by the Superintendent ol Document!, U.S. Oovenunent Printing Office
Washington, D.C. 20402 ¦ Price 13.21
-------
EPA Review Notice
This report has been reviewed by the Environ-
mental Protection Agency and approved for
publication. Approval does not signify that
the contents necessarily reflect the views and
policies of the Environmental Protection Agency,
nor does mention of trade names or commercial
products constitute endorsement or recommen-
dation for use.
tl
-------
SECTION III
INTRODUCTION
Today there is widespread public recognition of the gross in-
adequacies of conventional secondary sewage treatment processes in
protecting the environment under many circumstances and the need for
the practical application of new treatment methods which produce re-
claimed water of superior quality.
At Lake Tahoe this need became apparent about ten years ago
because of special local conditions which will be described in more
detail later in this report. As a result, research and pilot plant studies
were undertaken in 1961 to reveal possible new processes for waste-
water reclamation. It was found that a tertiary sequence of treatment
including conventional activated sludge followed by chemical treatment,
mixed-media filtration, and granular carbon adsorption produced remark-
able improvements in water quality. These Improvements included
virtually complete removal of suspended solids, BOD, bacteria, and
other substances only partially removed by secondary treatment. In
addition, good removals were obtained of COD, color, odor, viruses,
phosphates, MBAS, and other substances which are relatively unaffected
by secondary treatment.
In 1963, a 2.5 mgd tertiary plant was designed for the South
Tahoe Public Utility District Incorporating these processes, plus facil-
ities for thermal regeneration of granular activated carbon for flows up
to 10 mgd. This plant was built in 1964 and 1965 and was placed into
operation during the summer of 1965.
In April 1965, further laboratory, pilot plant, and full-scale
plant studies were initiated at Tahoe with WQO, EPA demonstration grant
funds. These studies included recovery and reuse of coagulant, nitro-
gen removal, and data collection in connection with full-scale carbon
regeneration and reuse.
In 1966, plans and specifications were prepared to expand the
capacity of the entire South Tahoe plant from 2.5 to 7.5 mgd. Facilities
were planned for the recovery and reuse of lime as a coagulant and for
the incineration of all sludge produced. These plant additions were
completed and placed into operation on March 31, 1968. Also an exper-
imental ammonia stripping tower with a nominal capacity of only 3.75 mgd
(one half that of the basic plant) was built and placed into operation in
November 1968.
5
-------
This 7.5 mgd Water Reclamation Plant of the South Tahoe Public
Utility District is the largest and most complete scale advanced waste-
water treatment plant in the world, and was the first to be placed into
operation. It has been designated as a National Demonstration Plant
by the WQO, EPA under the Clean Waters Restoration Act.
WQO, EPA Demonstration Grant WRPD 52-01-67 not only provides
for studies of recovery and reuse of lime as a coagulant and ammonia
stripping, but also for complete detailed studies and reporting of the
entire plant operation including costs for a three year period ending Feb-
ruary 1971.
This report includes: a discussion of the history and purpose of
the Tahoe project, a description of the process used, design data for the
plant and export system, an outline of the on-the-job training of plant
personnel, a description of sample collection techniques and test pro-
cedures employed, detailed descriptions of each liquid processing step
and each solids handling procedure, detailed results of treatment and
plant operation, complete data on actual construction and operating
costs, information on the prolonged storage of reclaimed water in Indian
Creek Reservoir, conclusions and recommendations, and other miscell-
aneous related information.
Only the lime recovery aspects are included in this excerpt.
6
-------
Requirements for Ultimate Disposal of Sludge. Of major Import-
ance in process selection are the circumstances related to the handling
and ultimate disposal of the sludge produced. In locations where there
are available remote, large areas of land, almost any kind of sludge,
wet or dry, stable or decomposing, can be, and is, disposed of by haul-
ing or pumping to these land disposal sites. In many places this crude
method for sludge disposal probably will not be tolerated indefinitely, but
might suffice for the time being. Ocean disposal of sludge has long been
an easy way to evade the knotty problems involved in proper sludge dis-
posal. However, the nuisances which have been created and the damages
wrought to beaches and coastal waters have aroused the public to the
point where this method is now in almost universal public disfavor.
All of the more acceptable methods for sludge disposal involve
dewatering of the sludge. The ease with which sludge may be dewatered
is a prime factor in unit process sielection. There are many alternate ways
to process the liquid component of wastewater to secure the desired re-
sult at about equal costs, but there are very few ways to satisfactorily
25
-------
and economically dewater sludge. In sludge from certain wastewaters,
dewatering of mixtures of organic-chemical sludges may be satisfactory,
but care must be taken to check this out before designing a full scale
plant. Favorable pilot plant test results are an important prerequisite.
Another approach which has been used successfully is to'keep
entirely separate all organic sludge from all chemical sludges. Then,
conventional equipment used for either of these types of sludge can be
installed. Pilot tests are still highly desirable, even with this approach.
Heat treatment or anaerobic digestion of organic-chemical sludges
may condition these mixtures to make possible use of a wide range of de-
watering equipment. Chemical conditioners may accomplish this same
end, but dosages must be determined to ascertain economic feasibility.
At South Tahoe, the choices in sludge disposal were: incineration,
or digestion and hauling the sludge out of the basin provided a disposal
site could be found. Because of the long haul involved in sludge export,
the descision to incinerate all waste organic and chemical sludges was
quite easy as it was favored both by esthetics and economics. Existing
sludge digestors for the original 2.5 mgd plant were utilized in the ex-
panded plant for emergency storage of sludge. They also can be used to
pretreat sludge for easier dewatering if necessary.
One of the greatest single factors favoring the tertiary sequence
of treatment (in which advanced treatment follows conventional biological
treatment) is the sludge handling problem. At Tahoe, mixtures of primary,
waste activated, and chemical sludges proved to be prohibitively expen-
sive to dewater, but organic sludges and chemical sludges which were
kept separate dewatered readily and at low cost. This is one of the most
important factors to be evaluated prior to final plant design, particularly
in the decision between physical-chemical treatment and the tertiary
sequence.
-------
SECTION VI
THE TAHOE PROCESS FOR WASTEWATER
RECLAMATION
Water Reclamation Plant. The South Tahoe water reclamation
plant Is the most advanced full-scale wastewater treatment plant In the
world, although other similar plants are now under construction In other
places.
Treatment of wastewater consists of two basLc parts, liquid
processing and solids handling. The first two steps of liquid processing
are the conventional ones of primary,or solids separation, and secondary,
or biological oxidation. In addition, the advanced treatment provides
chemical treatment and phosphate removal, nitrogen removal, mixed-
media filtration, activated carbon adsorption, and disinfection.
The spllds handling system provides for Incineration of biolog-
ical sludge, regeneration and reuse of granular activated carbon, and
recalclnlng and reuse of lime, all by means of multiple hearth furnaces.
The furnaces are equipped with scrubbers and after-burners to prevent
air pollution.
A detailed description of the treatment plant processes will be
given In the normal order which they occur In flow through the plant,
first for the liquid processing and then for the solids handling.
Figure 5 is a schematic flow and process diagram.
Liquid Processing. All wastewater Is pumped to the reclam-
ation plant. Plant Influent may be prechlorlnated for odor control. The
raw wastewater Is then passed through a barmlnutor which screens and
shreds the coarse solids, then through Parshall flumes which measure
the plant Inflow.
The water then passes to either or both of two primary settling
tanks where the liquids and solids are separated by sedimentation. In
addition to the raw wastewater, the primary tanks may also receive the
overflow from the lime mud thickener, and the centrate from the lime
centrifuge or the sludge centrifuge. If desired, lime or polymer may be
33
-------
SECONDARY
TREATMENT
(BIOLOGICAL
TREATMENT)
PRIMARY
CHEMICAL TREATMENT
AND
PHOSPHATE REMOVAL
ACTIVATED
MAJOR TYPES
OF TREATMENT
PROVIDED
TREATMENT
NITROGEN
FILTRATION
OlSlNFECTfON
CARBON
REMOVAL
SOLIDS
ADSORPTION
SEPARATION)
«0*G EMEROENCT
HOC DING POND
reclumeo
water to
INDIAN
CREC*
RESERVOIR
IMC
Su#GC
U<»*
FILTERS
MRSHALL FLUMES
FLOW MEASURCMEN
AND DIVISION
tMMMUTORS
CHLORINE
APPLICATION
RETURN TO
SECONOARY BALLAST
PONOAT PLANT
REC4RS0NATI0N
ISTANO 0TI
AMMONIA
FLOW DIVISION BOX
STRIP**©
FLOCCULATlON
LUTHER PASS •
BOOSTER PUMP
STATION
WASTE WATER
flow
through
plant
CHEMICAL
ClAAiriER
PKiMAJnr
CLARlFICR
SECONDARY
CLARlFiERS
FINAL
BALLAST
POND
RAPtq
MIX
SECONDARY
BALLAST
POND
Ef FLUENT
STATION
rowcR
PUMP
STATION rCARSOMAHO
ACTlWCO
SLU06C
TERTIARY
PLANT INFLUENT
FOftCt MAMS
PRIMARY
SLUOtt PUMPS
CARBON
SLURRY PUNP9—
LIME SLUOGC
PUMPS
OLD D4GESTCWS
EMEROENCT
PRIMARY SUAM
STORACf
SCCONOAAY
SLUOCt PUMPS
LIME
THICKEKR
SLUOGC FLOW
DIVISION §01
OCWATERiNO
TANKS
WASTE ACTIVATED
ILU0«C
LUC SLUOCC
solios
HANDLING
LIME.AND
CARBON
RECLAMfl "ION
CARBON
DC-FINM6
TANKS
Rl CALCINED U«t
TO RC-USE
FIGURE 5
SCHEMATIC FLOW AND PROCESS DIAGRAM
-------
added ahead of the primary tanks. One of the primary tanks Is a rectang-
ular basin, and the other Is a circular basin. Both are equipped with
mechanical collectors for continuous sludge removal.
The next step In the process Is secondary treatment or biological
oxidation. One-third of the plant capacity utilizes the conventional plug-
flow activated sludge process with diffused air. The other two-thirds of
the secondary treatment capacity Is provided by a complete mix system
using a combination of diffused air and mechanical mixing, or surface
aeration. Secondary settling takes place in either of two circular basins,
each of which has Its own sludge recirculation pump station. Chlorine
may be added to the secondary settling basin Influent for control of sludge
bulking as necessary.
Secondary settling Is the end of the processing In conventional
treatment plants, so that the additional treatment provided beyond this
point Is all In the nature of advanced waste treatment.
The secondary effluent together with waste filter backwash water
enters a rapid mix basin where violent mechanical stirring occurs, and
where lime Is added to pH=ll+, which corresponds to a dosage of about
400 mg/1 of lime. Next comes a period of slow mixing and flocculatton
by air agitation. Following this, a polymer Is added, usually In the
amount of about 0.1 mg/1, as the flocculated hlgh-pH water flows to the
chemical clarlfler. The chemical clarlfler, a circular basin, separates
the liquid from the rather large quantities of lime sludge.
Overflow water from the chemical clarlflers flows by gravity to
a sump where either of two pumps lifts It to the top of the nitrogen re-
moval tower, the first tower of Its kind used In a municipal wastewater
plant. The tower has a nominal capacity of 3.75 mgd (one-half plant
capacity), and Is the only part of the full-scale plant which Is still con-
sidered to be experimental In nature. Water pumped to the tower has
pH=ll+, Indicating that the ammonia Is virtually all present as dissolved
gas, rather than as ammonium Ion In solution. Large quantities of air
must be circulated through the tower for best efficiency, so that an open
packing must be used In the tower to minimize head losses and power
requirements for the circulating fan.
The water Is distributed uniformly from a horizontal tray across
the top of the packing, which Is made of treated hemlock slats spaced at
1.5 In. vertically and 2 In. horizontally. As the water strikes a slat,
droplets are formed. Surface film thickness In the droplets Is at a mini-
mum as they are forming, and this favors the escape of ammonia gas from
the droplet. By circulating a great amount of air through the tower, the
35
-------
air surrounding the droplet Is kept at a low ammonia concentration, pro-
moting a maximum transfer of ammonia from the water to the air. Once
droplets are fully formed, very little further transfer of ammonia takes
place, so the droplets are coalesced on top of the next slat below and
new droplets are formed as the water falls off the slat, allowing further
escape of the ammonia. This process is repeated about 240 times In one
pass through the 24-ft-high tower.
Air in the tower flows across the descending droplets. The air
enters through the side louvers and travels horizontally to a central
plenum, where it is discharged vertically upward through a fan that has a
maximum capacity of 700,000 cfm. Tower loading rates are about 2.9gpm
of water per square foot, and about 390 cfm of air per gallon of waste-
water. The efficiency of the tower in removing ammonia will vary from
30 to 98 percent, depending principally upon air and water temperatures
and to a lesser extent upon hydraulic loading and air supply.
From the catch basin beneath the nitrogen removal tower, the
tower effluent passes over a weir for flow measurement, and then into
another basin beneath the tower, which contains three sections. The
first section is the primary recarbonatlon basin. Here, carbon dioxide
gas, supplied by compressing stack gasses Ln the incinerator building,
is dissolved in the water to lower the pH from 11 to 9.6. The recarbon-
ated water is then held in a contact basin for 30 minutes or more to allow
complete formation and some settlement of calcium carbonate. This basin
is equipped with equipment for continuous sludge removal. The third
section of the basin is the secondary recarbonation chamber, where the
pH is reduced from 9. 6 to any desired level, usually 7.5, by further
addition of carbon dioxide.
The recarbonated water then flows through two ballast ponds in
series. The ballast ponds are used to store water for backwashing the
mixed-media filters, and to reduce peak flows to the filters and carbon
columns.
Water Is pumped from the ballast ponds to the filters and carbon
columns. Ordinarily about 5 mg/1 of alum are added to the filter influent
water. Three pairs of mixed-media filter beds, developed especially for
filtering waste water, handle the plant flow.
The mixed-media bed, with pores graded coarse to fine In the
direction of flow, Is composed of coarse coal, with a specific gravity of
1.4; medium-sized sand, with a specific gravity of 2.65; and fine garnet,
with a specific gravity of 4 .5. A properly graded bed made of these
materials i-nvides an almost ideal coarse-to-fine filter. Because of the
different specitic gravities of the three materials, the particles retain
their desired position in the depth of the bed during backwashing. The
36
-------
fine filter media is supported on a conventional bed of graded gravel, to
which has been added an important new feature, a 3-in. layer of 16-mesh
garnet. This extremely heavy material (specific 4.5) positively prevents
any movement of the gravel bed below, and also prevents any penetration
of the fine garnet above into the gravel supporting bed. The filters are
equipped with rotary surface washers, and rotation indicator lights. Good
surface wash is essential to proper operation of the coarse-to-fine filters
because of the large quantities of particulates removed from the water and
stored throughout the depth of the bed.
Each pair of beds comprises a filter unit, and they operate in
series both during filtration and backwash. A total depth of 6 ft of fine
media is provided in each pair of beds. With coarse-to-fine filters, the
length of filter run is almost directly proportional to the depth of fine
media, so that the length of run for the two beds in series is about double
that for a single bed. By backwashing the two beds in series, about half
as much backwash water is discharged to the decant tank and then return-
ed at a slow rate to the rapid-mix basin for reprocessing.
When a pair of filters is plugged or no longer produces a high
quality effluent, it is automatically taken off the line, backwashed,
filtered-to-waste, and restored to service. This is all done automatically
and monitored through a control panel located in the filter building.
The filtered water next flows under pressure to the eight carbon
columns, which operate in paraLlel. Each column is 12 ft in diameter by
24 ft high, and contains about 2 2 tons of 8 X 30 mesh granular activated
carbon. Flow in the columns is of the moving bed, countercurrent type -
that is, water flows from the bottom to top of the column, while movement
of the carbon is down, the fresh carbon being added at the top and the
spent carbon removed at the bottom. This system permits frequent removal
and corresponding addition of carbon for maximum operating efficiency.
The water is in contact with the carbon for a period of 15 to 2 5
minutes, during which the adsorption of organics by the carbon takes
place. The carbon column effluent is colorless, odorless, low in organics,
and sparkling clear.
The high quality of the water following complete treatment vastly
improves the efficiency of chlorination, the final step in the liquid pro-
cessing. As compared to chlorination of ordinary secondary effluent, the
chlorination of the reclaimed water at South Tahoe is many times more ef-
fective, as virtually all of the chlorine-demanding materials, except am-
monia, present in the secondary effluent have been removed. With rapid
violent mixing at the point of chlorine application, good disinfection can
37
-------
be accomplished in the presence of ammonia. One explanation offered is
that chlorine reacts more rapidly with bacteria and viruses than with am-
monia .
Solids Handling System. As mentioned previously, all solid
plant wastes are processed in multiple hearth furnaces. The biological
and waste chemical sludges are incinerated, the lime mud is recalcined
and reused in the process, and the spent granular carbon is regenerated
and reused.
The materials to be incinerated include primary and waste activ-
ated sludge, waste chemical sludge, screenings, skimmings, and the
centrales from the lime and sludge centrifuges.
A mixture of all of these materials is pumped to the sludge cent-
rifuge, which operates at about 1,600 rpm, where a polymer is added and
the sludge is partially dewatered to a solids content of about 19 percent in
the cake. The cake is conveyed by belt to a multiple hearth furnace; the
fuel is natural gas. The furnace is operated at about 1,600 °F and re-
duces the sludge to an insoluble sterile ash, which may be disposed of on
the plant grounds. The furnace stack gas is cooled to 110°F and scrubbed.
There is no odor, smoke or steam plume, and the discharges meet all air
pollution codes. When the sludge dewatering or incineration equipment is
out of service, the old sludge digesters may be used for sludge storage.
Handling of the spent lime mud is similar to that just described
for biological sludge. Lime sludge from the chemical clarifier is pumped
to a sludge thickener which thickens the sludge to about 8 percent solids.
The thickened sludge is then pumped to a centrifuge for further dewatering
at 1,600 rpm. The cake contains about 40 percent solids. The lime is re-
calcined at about 1,850°F to a calcium oxide content of about 50 to 80 per-
cent. Again the stack gas is cooled and scrubbed so that there is no air
pollution. Only about 75 percent of the lime sludge is recalcined and re-
used. The other 25 percent is wasted, mostly hydroxyapatite, in the cen-
trate to the primary clarifier and thence to the sludge furnace, or directly
to the sludge furnace. Alternately, the centrate from the lime centrifuge
can be dewatered in a second centrifuge and the cake incinerated in the
sludge furnace.
The recalcined lime is conveyed pneumatically to a storage bin,
and then reused in the process. Makeup lime is unloaded from trucks
pneumatically and stored in a separate bin. Separate gravimetric feeders
and slakers are provided for the recalcined and makeup lime. Many of the
furnace operations are controlled from a panel on the main floor of the in-
cineration building.
38
-------
The carbon regeneration equipment is located in the filter build-
ing. After the carbon becomes saturated with materials removed from the
wastewater, it loses its capacity to absorb certain organics, and must be
regenerated. Originally in 1965, a break-through of MBAS (or detergents)
was the indicator that regeneration was needed. With the advent of the
soft detergents, a high COD content in the carbon column effluent has
become the indicator. In regeneration a carbon column is pressurized
and the carbon slurry is drawn off the bottom of the column to one of the
dewaterlng bins. Here the free moisture drains off in about 10 minutes,
leaving carbon with a moisture content of about 40 percent, which is
suitable for introduction of carbon to the furnace. Spent carbon is fed to
the furnace at a controlled rate by a screw conveyor equipped with vari-
able speed drive. The carbon regeneration furnace is operated at about
1,700°F in a limited oxygen atmosphere with the addition of steam. The
rate of feed to the furnace and the hearth temperatures are controlled by
the apparent density of the regenerated carbon, which is held at 0.48 to
0.49. As a check on regeneration efficiency, iodine numbers ( a relative
measure of adsorptive capacity) are run on carbon samples in the laboratory.
The carbon is regenerated to full virgin activity with an attrition loss that
is about 8 percent per cycle. The regenerated carbon is cooled in a quench
tank, pumped by a diaphragm slurry pump to wash tanks, washed to remove
carbon fines, and then returned to the top of the carbon column.
39
-------
SECTION VII
PLANT DESIGN DATA
The principal design criteria for the South Tahoe Water Recla-
mation Plant are tabulated below.
Item
Amount
Plant design average flow
7.5
mgd
Peak flow rate (except as noted below)
15.0
mgd
Peak flow rate (filters and carbon columns)
8.2
mgd
Maximum hydraulic rate
20.0
mgd
Plant design BOD (summer)
325
mg/1
Plant de.:;ign 30D (winter)
250
mg/1
Plant suspended solids (summer)
200
mg/1
Plant suspended solids (winter)
150
mg/1
Water temperature (summer)
17°
C
Atmospheric pressure (elevation 6,300 ft)
11.6
psi
Primary clarifier No. 1
Surface area
2,350
sf
Flow
2.7
mgd
Overflow rate
1,150 gpd/sf
Primary clarifier No. 2
Surface area
7,850
sf
Flow
4.8
mgd
Overflow rate
610 gpd/sf
Aeration basins 1,2, and 3, plug flow
Flow
2.7
mgd
Volume
115,000
cf
Detention (without recycle)
7.5
hrs
BOD loading
50 lbs/1,000
cf
41
-------
Item
Amount
Pumps to tertiary plant
No. 1
1,900
gpm
No. 2
3,800
gpm
No. 3
4,200
gpm
Surface wash booster
500
gpm
Mixed media filters
Flow
Units, 3 sets of 2 series beds
Hydraulic loading
Backwash rate
Area each bed
Surface wash flow
Waste backwash water receiving tank
Capacity
8.2
mgd
5 gpm/sf
15 gpm/sf
380 sf
0.6 gpm/sf
80,000 gals
Carbon columns (8),upflow countercurrent
Flow
Carbon volume, each column
Carbon depth, effective
Contact time
Hydraulic loading
Chlorination equipment
Three feeders, each
8.2 mgd
1,810 cf
14 ft
17 min
6.5 gpm/sf
2,000 lbs/day
Carbon regeneration furnace, 6-hearth,
54-inch diameter, gas-fired
Capacity, dry carbon
Sludge dewatering equipment, concurrent
flow centrifuges, 24" x 60"
Organic sludge
Number
Capacity, each, dry solids
Lime sludge
Number
Capacity
6,000 lbs/day
4 50 lbs/hr
1,650 lbs/hr
44
-------
Item
Amount
Sludge incineration furnace, 6-hearth
14'-3" diameter, gas-fired
Capacity, dry solids 900 lbs/hr
Lime recalcining furnace, 6-hearth
14'-3" diameter, gas-fired
Capacity, dry CaO 10 tons/day
45
-------
SECTION XVII
LIME RECOVERY AND REUSE
General. At a flow of 7.5 mgd through the water reclamation plant
approximately 17 tons (dry CaO basis) per day of lime mud would have to
be dewatered and disposed of. Since about 93% by weight of this lime mud
is in the form of calcium carbonate , disposal costs would include not only
dewatering and disposing of about 34 tons of water and solids but also the
loss of recoverable calcium oxide. By recovering the lime through recal-
cination, the total blow-down of waste solids is reduced to about 1.5 tons
of dry solids. The cost of recalcined lime as shown later would be slight-
ly more than that of new lime at 7.5 mgd; however, at this flow the reuse
of lime reduces by a factor of 20 the amount of water and sludge to be dis-
posed of and, therefore, effects a substantial cost savings.
Physical System. Lime mud is pumped from the chemical clarifier
and recarbonation reaction basin to a 30-foot diameter gravity thickener,
with a design overflow rate of 1000 gal/ft^/day. Thickened lime mud is
pumped to a 24" x 60" solid bowl concurrent flow centrifuge. The cake
from the centrifuge is carried by a belt conveyor to a 14.3 foot diameter,
six hearth furnace in which calcium oxide and carbon dioxide are produc-
ed. The recalcined lime is conveyed out of the furnace by gravity through
a crusher to a thermal disc cooler where lime temperatures are lowered
from 700°F to 100°-150°F, and then into a rotary air lock. The recalcined
lime is pneumatically conveyed from the rotary air lock to a 35-ton capac-
ity recalcined lime storage bin for eventual reuse. Stack gases, rich in
carbon dioxide, are scrubbed in a multiple tray scrubber before being ex-
hausted to the atmosphere. A portion of the gases are recycled to the re-
carbonation system. See Figure 41.
Solids wasting must be performed continuously to maintain an accept-
able calcium oxide content in the recalcined lime. Wasting can be accom-
plished by feeding recovered lime to the primary clarifier, by diverting
part of thickener influent to the primary clarifier, by conveying lime mud
cake from the centrifuge directly to the organic sludge furnace or by us-
ing the centrifuge to classify the phosphate and other inerts into the cen-
trate and the calcium carbonate into the dewatered cake conveyed to the
139
-------
IAIN
ITACI
INOUCEO //A
OIAFT
ri>
CENTRATE
10 priiary
CLARIF ICR
INFLUENT
CHANNEL
NO. I
CENTRIFU8C
FOR LIKE
ILU06E
TO CO
NO. 3
CENTRIFUGE
FOR LINE OR
SEVABE SLUDGE
CLOSED
OPEN
A BYPASS
r STACK
LINE BUS
THICKENER
HOT AIR
RIVERStlLE
SCREV
CONVEYOR
FEE0
INLET
IEVERSI9LI
FECO CONVEYOR
I
IYPASS -
DEfATEIIO
LINE SLU08E
HORACE BIN
CHERICAL
CLAR IF IER
FURNACE
HEARTH
LOAO IN6
u
DRAIN
L HE
IICALCINIRI
FURNACE
THERRAL
DISC.
ARR DRIVE
SHAFT
COOL INS
AIR FAN
FROM LINE
DELIVERY
TRUCK
KDp
PNEURATIC I
UNLOADINO J
AND
CONVEYINO
EQUIPMENT
CONIUSTION
All ILOIER
LIRE
LIRE
FEEOER
FEEOER
LIRE
LIRE
SLAKER
SLAIER
H
RECALC 1 NEO
il
FICSH
LIRE TO
llll TO
SPLITTER
smm»
aoi
eoi
RECALCINEO
LINE RLOVER
F 1
G 11 R
E
41
SOL
1 D
S H
A N
D L I N G
L
1 M
E S
L U
D G E
141
-------
recalclner. A second centrifuge dewaters the centrate from the first
machine and its cake is conveyed to the organic sludge furnace.
The centrate from the lead centrifuge may be returned to the prim-
ary clarifier instead of directed to the second machine. The second cen-
trifuge's centrate is returned to the primary. The spillage from the lime
conveyor belt is collected in a tray and returned to either the primary
clarifier or to the lime mud thickener.
Operating Practice. Since plant startup in 1968, steps have been
taken to remove all active lime or waste lime mud streams from the prim-
ary clarifier. As described in the "Chemical Treatment" section, feeding
active lime to the primary clarifier led to organic sludge dewatering prob-
lems and to some extent adversely affected primary clarification. Waste
lime mud streams being returned to the primary clarifier also affected, but
to a lesser extent, the dewatering characteristics of the raw and waste
activated sludges.
In January, April and May 1970, and from July 1970 to date, the
lime centrifuge has been used to classify inert materials out of the feed
to the recalcining furnace. At the same time, the classified centrate from
the first centrifuge has been dewatered and clarified in a second centri-
fuge and dried in the organic incineration furnace. The degree of classi-
fication and resulting economies are discussed later in this section.
Lime Mud Thickening. Total solids tests were used to evaluate
the efficiency of the lime mud thickener. To determine the amount of cal-
cium lost over the thickener launder the following procedures were used.
Basin overflows and sludge withdrawals were measured to determine the in-
fluent flow. The amount of calcium in the thickener influent was determin-
ed by heating the influent total solids samples in the laboratory muffle
furnace at 1850°F for two hours and then determining the available CaO
content. By this method, the calcium hydroxyapatite, Ca50H(P04)3, was
not included, and the value reflected only the calcium in CaC03 and
Ca(OH)2. Part of the influent sample was filtered 0.45 u millipore filter
and calcium hardness was run on the filtrate to determine the Ca(OH)2-
The difference in the two provided the pounds of usable calcium available
for thickening.
The pounds of usable calcium in the thickener overflow were found
by determining the difference in calcium content between acidified and fil-
tered samples.
143
-------
The amount of usable calcium lost over the thickener launder was
evaluated over an eight hour composite sampling period. Using the pro-
cedures described above, 0.04% by weight of the usable calcium coming
into the thickener was lost over the weir.
The true ability of the thickener to concentrate lime mud solids
could not be evaluated due to low plant flows. Thickener influent percent
solids were about 1% by weight. At chemical clarifier flows of 2,8 mgd
and thickener underflow rates of 36 gpm, the lime mud was thickened to
4.9% solids; whereas at 3.7 mgd through the chemical clarifier and 15 gpm
thickener underflow rate, lime mud was thickened to 8.3% solids.
Lime Mud Classification and Dewaterinq. During the lime centri-
fuge acceptance tests, three bowl speeds were evaluated to determine the
optimum centrifugal bowl speed. The three bowl speeds evaluated were
1600, 1800, and 2200 rpm. Each of the three evaluations were conducted
at 140:1 gear reduction ratio. The function of the gear unit is to drive the
conveyor at a fixed speed relative to the bowl. For the three bowl speeds
above, the conveyor revolved at approximately 11.5, 13, and 15.5 rpm. For
all three bowl speeds evaluated, the cake solids ranged from 36-38%, and
recovery or capture remained fairly constant. As a result, the lowest bowl
speed of 1600 rpm was chosen for plant operations because of the reduced
wear and maintenance problems.
Chemical Addition. Since the lime centrifuge is used primarily
for classification purposes and not dewatering, polyelectrolytes were not
used to condition the thickened mud prior to centrifuging. In the next
section the excellent classification results are discussed. The addition
of a polyelectrolyte to the centrifuge feed would defeat these purposes.
Classification Evaluation. Two methods were used to evaluate the
centrifugal classification of the inert calcium hydroxyapatite out of the
lime solids recovery stream. The first was to make four, ten-minute samp-
ling runs with high and low feed rates and pool depths within the centri-
fuge to optimize the classification process. An approximate twenty min-
ute interval was allowed after making adjustments to the centrifuge to
insure steady state conditions for a particular set of variables. The cen-
trifuge bowl speed was kept constant at 1600 rpm. During each ten-minute
period a sample was composited uniformly on the feed, centrate, and cake
steams. The centrifuge pool depth, cake volumetric flow rate and wet
144
-------
density, and centrate flow rate were recorded. The composited samples
were analyzed for total solids and then calcined in a muffle furnace at
1750°F. The total solids, percent capture and centrate flow rate were used
to compute the feed rate. After calcination, the samples were analyzed
for CaO, and samples were dissolved in acid and analyzed for magnesium
and orthophosphates.
With the volumetric flow rate and phosphate and calcium results,
balances were made around the centifuge for usable calcium, and the inert
materials, magnesium and phosphate.
The second method was to composite a sample of centrifuge feed,
centrate, and cake every two hours for an eight hour period. The centri-
fuge centrate flow rate was recorded every two hours also. As in the pre-
vious evaluation, the centrifuge bowl speed was maintained at 1600 rpm.
The resulting composite samples were analyzed for total solids, calcined
at 1750°F for 2 hours, and then the usable calcium was determined by an-
alyzing for percent calcium oxide. The dilutions and analysis for phos-
phate used for the first evaluation were repeated. The percent capture for
solids was computed with the analysis information, and the appropriate
balances around the centrifuge for usable calcium and phosphate were deter-
mined .
Classification Evaluation Results. A brief explanation of the eff-
ects of changing the pool depth in a centrifuge is necessary before the
classification results are explained. If the pool depth is increased, the
cake should get progressively wetter. It is possible to increase the pool
depth to the extent that the cake leaves the centrifuge as a wet slurry.
However, if the pool depth is decreased to just below this point, the
maximum capture should be attained. Conversely, if the pool depth is
decreased, the centrate should have a higher solids content, lower per-
cent solids capture, but the cake should be drier. The centrifuge pool
depth setting is on a scale of 1-10, with a low number corresponding to a
low pool depth and a high number, a high pool depth. As Figure 43 indi-
cates, the centrifuge performed as was expected, the higher pool depth
produced the higher captures for similar flow rates. The flow rate to the
machine also has a significant effect on the solids capture, with a mark-
ed decrease in solids recovery at the higher flows. The highest flow rate,
23 gpm, had a feed solids content of 6%, whereas the other four data points
had feed solids contents of 8%. Also the highest flow rate data point was
derived from an eight-hour sampling period, and the other four points were
from ten minute sampling periods.
145
-------
FIGURE 43
PERCENT SOLIDS CAPTURED VS FLOW RATE TO CENTRIFUGAL
O - POOL SETTING 7
• - POOL SETTING 3
0
20
10
25
30
5
15
FLOW RATE TO CENTRIFUGAL. GPM
-------
The removal of inert materials by classification into the centrate
stream provides many benefits to the lime treatment and solids recovery
system. A significant savings in fuel usage is realized, which is des-
cribed later in the recalcining portion of this section. Also the size of
the waste stream is reduced, a smaller amount of usable lime is lost,
which results in lower makeup lime dosage and costs. The effect of sol-
ids capture or recovery in the centrifuge on the removal of inert materials
from the feed to the recalcination furnace by classification is shown in
Figures 44 through 47. The data in Figures 44 and 45 is derived from the
two previous sampling runs described and from three runs of the optimi-
zation of the recalcination furnace described later. As would be expect-
ed, higher captures of solids from the centrifugal feed will result in a
lower percent increase in the usable calcium or active lime, CaO, in the
cake. At lower captures the lower specific gravity inert materials can be
separated from the usable lime. This is very evident in Figure 44, at 90%
solids capture there is only a 5% improvement in the percent CaO of cake
over the feed to the centrifugal. At 70-75% solids capture, a 12-15% im-
provement in the cake percent CaO over the centrifugal feed can be ex-
pected. Correspondingly in Figure 45, at 95% capture practically all of
the usable calcium is conveyed to the recalcination furnace, but so are a
majority of the inert materials. At lower solids recoveries a portion of
the usable calcium is lost in the centrate, but the higher active lime or
CaO content of the cake from better classification of the inerts more than
offsets this loss.
Since one of the high priorities of the treatment at South Tahoe is
to remove phosphorus from the wastewater, the logical question is how
efficient is the phosphorus removal from the lime solids recovery stream.
In Figure 46, the effect is shown of solids capture and classification on
the amount of phosphate wasted in the centrate. The four data points a-
bove 80% capture are taken from the ten-minute sampling periods and the
other point is taken from the eight-hour period. At 90% capture or 10% of
the solids entering the centrifuge being removed in the centrate, a 20%
reduction in phosphate in the cake can be expected. A centrate contain-
ing 20% of the solids entering the centrifuge will have almost 40% of the
phosphate entering the centrifugal. According to the trend indicated in
Figure 46, 90% of the phosphates in the centrifuge feed can be wasted
to the centrate at 75% solids capture or recovery, which corresponds well
with Figure 44, indicating a near maximum increase in the cake CaO over
the centrifuge feed CaO around 75% capture of solids. The removal of
magnesium from the centrifuge feed to the centrate by classification is
shown in Figure 47 . The four data points shown in Figure 47, are from
the ten-minute sampling runs; magnesium was not analyzed in the eight-
hour run. The fourth point in Figure 47 was disregarded in drawing the
147
-------
FIGURE 44
EFFECT OF CENTRIFUGAL CAPTURE ON THE CHANGE IN
CALCIUM OXIDE CONTENT OF THE CENTRIFUGAL CAKE
+10%
+15%
CaO
+20%
RELATIVE INCREASE IN CAKE CaO CONTENT
TO FEED CaO CONTENT
-------
FIGURE 45
PERCENT SOLIDS CAPTURED
VS
PERCENT OF USABLE CALCIUM IN FEED CONVEYED TO FURNACE
90
80
70
75
85
95
100
PERCENT OF USABLE CALCIUM IN FEED CONVEYED TO FURNACE
-------
FIGURE 46
PERCENT SOLIDS CAPTURED
VS
PERCENT OF P04 IN FEED WASTED IN CENTRATE
20
O
40
80
60
120
100
PERCENT OF P04 IN FEED WASTED IN CENTRATE
-------
FIGURE 47
PERCENT SOLIDS CAPTURED
VS
PERCENT OF MAGNESIUM IN FEED WASTED TO CENTRATE
100
o
iu
cc
D
95
<
u
S
D 90
O
O)
t-
z
111
o
cc
Ui
Q.
85
80
10
15
20
25
30
35
40
PERCENT OF MG IN FEED WASTED TO CENTRATE
-------
line, since the analysis for this run resulted in a 15% error in the magnes-
ium balance around the centrifugal. However, the same trend as the
wasting of phosphates in Figure 46 is indicated. At 80% capture, approxi-
mately 35% of the magnesium entering the centrifuge is classified into
the centrate.
To further exemplify the classification abilities of the lime centri-
fuge the organic sludge and recalcination furnaces' scrubber waters were
analyzed for ortho phosphorus (mg/1 PO4-P) for three different periods of
operation. The first period was under normal classifying operations with
the lead lime centrifuge classifying and the swing centrifuge dewatering
the centrate and conveying the waste inert materials to the organic sludge
furnace. The second period was when the recalcination furnace was remov-
ed from service for a short period for maintenance, but the organic solids
furnace continued to operate. The third period was when the lead lime
centrifugal was not used for classifying, but dewatering only, and the re-
sulting cake was being recalcined. The average results of the three per-
iods are shown in Table 14 .
TABLE 14
SCRUBBER WATER
ORTHO PHOSPHORUS CONTENT, mg/1 PO4-P
Operation Mode
Sludge Furnace
Recalcination Furnace
Total
Soluble*
Total
Soluble*
Classification
17.95
.56
2.4
.32
Not Recalcining
1.54
0.08
0.19
0.07
Lime Centrifuge
3.59
—
12.02
—
for dewatering only
~Soluble defined as passing a .45 u filter.
The water reclamation plant effluent is used for the source of sup-
ply for the furnace scrubbers. The normal ortho phosphorus content of
the plant effluent is in the range of .06 to 0.1 mg/1. When the recalcina-
tion furnace was out of service, it's scrubber water phosphorus content
was barely affected. During the same period, when the sludge furnace
was incinerating only organics, it's phosphorus content was 1.5 mg/1,
all particulate. When the lead lime centrifuge is used for dewatering
152
-------
purposes only, the lime centrate is directed to the primary where the lime
solids that are in the centrate settle out and are dewatered and dried or in-
cinerated with the organic sludges. During this period the phosphorus con-
tent of the sludge furnace scrubber water increased to 3.5 mg/1, as a re-
sult of the lime centrate being directed to the primary. Since the lime
centrifuge is operated with the highest possible solids capture when it is
used as a dewatering device, the majority of the inert materials are convey-
ed to the recalcination furnace with the lime cake. This resulted in a much
higher phosphorus scrubber water content of 12 mg/1. Only total ortho
phosphorus was analyzed during this period. During the period of normal
operations when the lead lime centrifuge is used for classification the
sludge furnace, drying the concentrated waste lime inerts increased to
18 mg/l, mostly particulate. Since the majority of the inert materials had
been classified out of the lime cake, the recalcination furnace scrubber
water content decreased to 2.4 mg/1, mostly particulate.
Lime Mud Recalcininq. Since April 1968 the District has success-
fully recalcined lime mud from the lime chemical treatment process. Over
this period makeup lime has accounted for only 28 percent of the calcium
oxide used. Average monthly CaO values in the recalcined lime have rang-
ed between 51.0% and 74.7% with the average over the entire period being
66.0%. Table 15 shows the operating data for the lime mud recalciner. A
reduction of approximately 40% in fuel requirements was achieved when
centrifugal classification was used.
In an effort to measure the usable calcium losses in the recalcina-
tion furnace, the assumption was made that all the calcium lost as fly ash
would be captured in the wet scrubber. The increase in calcium and phos-
phate in the scrubber water as a result of fly ash consisting of usable cal-
cium and inert calcium hydroxyapatite was measured by acidifying the
scrubber influent and effluent samples to pH 2.0, filtering with a .45 u
filter, neutralizing the filtrate to pH 7 and analyzing for calcium hardness
and phosphate. The phosphate was measured to determine the amount of
calcium combined with the hydroxyapatite. The difference in terms of cal-
cium, of the calcium hardness and the calcium combined with the hydroxy-
apatite provided the amount of usable calcium loss from the furnace. By
measuring the amount of usable calcium in the centrifuge cake entering
the furnace, described earlier in the classification and dewatering portion
of this section, the percent calcium losses from the furnace can be deter-
mined. The amount of usable calcium loss from the furnace in the scrubber
water was evaluated over an eight hour composite period. Using the anal-
ysis procedures and assumptions described above, 3.7% by weight of the
usable calcium entering the furnace during the eight hour period was lost
from the furnace.
153
-------
TABLE 15
LIME RECALCINER OPERATING DATA
No Centrifugal Centrifugal
Classification Classification
Period Oct. 69 - May 71 July 70 - Nov.70
Chemical Clarifier Flow, MGD 2 .95 3 .63
Feed, lbs/hour^) 435 756
Feed, % Solids 33.3 42
Electricity KWHR/day(2) 508
Fuel Requirements ,BTU/lb(3) 5500 3270
(1) Pounds of dry solids per hour
(2) Includes energy and demand charges for furnace support motors
and recalcined lime conveying system. February 70 to Decem-
ber 1970
(3) Natural gas at 18 psia and .860 BTU/ft^
154
-------
Optimum Furnace Conditions. Eades and Sandberg In their dis-
cussion of lime reaction parameters point out that, "Although the art of
lime burning has been practiced since ancient times, it was not until the
18th century that a scientific explanation of calcination was advanced. As
industrial and chemical technology developed, lime became an increasingly
important component in numerous reactions and processes. For these appli-
cations, the lime was judged primarily on its chemical purity, and minimal
amounts of silica, alumina, iron, and other impurities were desired. In
general, reaction rates were not thought important."
"Reaction rates of commercial limes became a matter of considerable
interest with the introduction of the basic oxygen converter steel furnace.
Operation costs for such converters are quite high, and steel producers
quickly began investigation of methods to lower the time per heat of steel.
Metallurgists became interested in the relationship between lime reactiv-
ity and slagging time. As a result of these and other studies, it quickly
became apparent that in steel making and many other industrial applications
reactivity of a lime was a more significant criterion for judging quality and
suitability of a lime than chemical purity."
The authors further state, "that pore space and calcium oxide (CaO)
crystallite size are the prime factors controlling reactivity of any given
lime. These in turn can be linked to calcining conditions with low tempera-
ture burning producing a highly porous, highly reactive lime and high tem-
perature burning producing a shrunken, dense lime with low porosity and
low reactivity".
According to the AWWA Standard for Quicklime and Hydra ted Lime
(AWWA B202-65), high-reactive, soft-burned lime will show a temperature
rise of 40°C in 3 minutes or less and the reaction will be complete within
10 minutes when tested by the method given in the standard. A medium-
reactive, medium-burned lime will show a temperature rise of 40°C in 3 to
6 minutes and the reaction will be complete in 10-20 minutes. For low re-
active, hard-burned lime more than 6 minutes will be required for a 40°C
temperature rise and the complete reaction time will take longer than 20
minutes.
Since it is possible to produce quicklime with the same calcium oxide
content, but with very different slaking properties, tests were performed on
the lime recalcining furnace. The purposes of the tests were to determine
the optimum recalcining temperature, feed rate and rabble rate in terms of
lime reactivity and furnace fuel requirements.
155
-------
The Rapid or EDTA Method for calcium oxide as described in Section
XI and the AWWA Slaking Rate Test at 400 rpm were used to determine the
reactive properties of the recalcined lime. At least an hour was allowed
after changes in process variables to permit the furnace to reach equilibrium.
A sample was then composited at 15-minute intervals from the No. 6 hearth
for an additional hour. Composite feed samples were also collected at the
same intervals and recalcined in the laboratory muffle furnace at 1800°F for
1 hour to insure that the potential input calcium oxide remained constant.
To establish a qualitative base line, calcium oxide and slaking rate
tests were performed on virgin makeup lime. The data presented in Figure
48 shows that the makeup 16 x 50 mesh granular quicklime used at South
Lake Tahoe is highly reactive. A 40°C temperature rise is reached in less
than 30 seconds and the reaction completed in seven minutes. Identical
results were achieved in three separate tests of the sample.
The effect of temperature on the recalcined lime activity at a constant
feed and rabble rate was first investigated. Table 16 shows that recalcin-
ing temperatures between 1600°F and 1900°F had a major effect on recalcin-
ed lime activity, although all three temperatures produced lime which, un-
der the AWWA standard, was considered to be highly-reactive. Within
this temperature range there was no indication that the lime was being over-
bumed, since the total reaction time was well within the time requirements.
Recalcining lime at 1900°F as opposed to 1800°F produced a 5% increase in
available calcium oxide, but very little improvement in an already accept-
able slaking rate.
At 1600°F the flour-like recalcined lime showed pronounced tenden-
cies to agglomerate into soft, easily crushed particles of 1/4 inch to 3/4
inch diameter. Many of the particles contained centers of unburned organ-
ic sludge. Additional evidence of unburned organic sludge was observed
in the Dewar Flask after the slaking test.
For the second phase, the rabble rate was varied between 1.5 and
2.0 rpm while the temperature and feed rate were held constant. At both
1900°F and 1600°F the variation of rabble rate showed very little effect on
the recalcined lime activity. Once again temperature proved to be the maj-
or variable. Table 17 shows the results of varying the rabble rate.
Finally,to determine the effect of feed rate, the temperature and rab-
ble rate were held constant at 1900°F and 1.5 rpm, respectively. The feed
rate to the centrifugal was run at 870, 820, and 450 pounds of dry solids
per hour. Table 18 shows that the total slaking time doubled when the feed
rate was reduced from 870 lbs/hr to 450 lbs/hr with no significant change
1 56
-------
in
FIGURE 48
VIRGIN LIME SLAKING RATE TEST
AVVWA STANDARD B202-65
90
80
70
oc
3
H
<
IT
LU
Q.
2
o
z
60
50
40
30
I
" 24.5°C
CALCIUM OXIDE CONTENT: 94.2%
3-MINUTE TEMPERATURE RISE: 56.6°C
TOTAL TEMPERATURE RISE: 58-8°C
TOTAL ACTIVE SLAKING TIME: 7 min
20
4 5 6
SLAKING TIME. MIN.
10
-------
TABLE 16
Effect of Temperature On
Recalcined Lime Activity
At a Constant Feed and
Rabble Rate
No. 3 Hearth
No. 4 Hearth
No. 5 Hearth
Recalcininq Temperature °F
1640 1630 1450
1900 17 10 1620
1900 1780 1600
Percent CaO
86
81
76
Slaking Rate:
60 Sec. Temp. Rise, °C
Total Temp. Rise, °C
Total Reactive Time, min,
50 48
51.5 48
2 1
41
41
1
Feed Rate, lbs/hr^) 870 800 870
Rabble Rate, rpm 1.5 1.5 1.5
( 1 ) Dry solids to centrifuge
158
-------
TABLE 17
Effect of Rabble Rate
On Recalcined Lime
Activity At a Constant
Feed Rate
Rabble Rate, rpm
2 .0 rpm 1. 5 rpm
1900°F Avq Temp M
Percent CaO
86
86
Slaking Rate
60 Sec. Temp. Rise,
°C
51.5
50
Total Temp Rise, °C
52
51.5
Total Reaction Time,
min.
1.5
2
Feed Rate (2), lbs/hr
940
870
'F Avq Temp M
Percent CaO
70
76
Slaking Rate
60 Sec. Temp. Rise,
°C
39
41
Total Temp Rise, °C
39.5
41
Total Reaction Time,
min.
2
1
Feed Rate , lbs/hr
830
870
( 1 ) Average of No. 4 & No. 5 Hearth temperatures
( 2 ) Dry solids to centrifuge
159
-------
TABLE 18
EFFECT OF FEED RATE ON
RECALCINED LIME ACTIVITY
AT 1900°F AND 1.5 RPM
RABBLE RATE
FEED RATE (2) lbs/hr
870 820 450
Percent CaO
86
90
89
Slaking Rate
60 Second Temp. Rise, °(
Total Temp. Rise, °C
Total Reaction Time, min.
50
51.5
2
50.5
53.5
2.5
50.5
53.5
4
(1) Average of No. 4 and No. 5 hearth temperatures
(2) Dry solids to centrifuge
160
-------
in the calcium oxide content. All three feed rates produced highly-react-
ive recalcined lime under the AWWA standard. However, the lower feed
rate produced a less reactive lime.
The optimum furnace conditions in terms of recalcined lime activity
appear to be about 1900°F on the fourth and fifth hearths at 1.5 to 2.0 rpm
rabble rate for 800-900 lbs/hr of dry solids to the centrifuge. The centri-
fuge was being used as a classifying device during this test period. On
the basis of the 75% capture that was obtained, the actual furnace feed
rate for the optimum conditions was 600 to 700 lbs/hr.
A slightly less reactive lime was obtained at an average recalcining
temperature of 1750°F. At the same loading rates and taking into consider-
ation the higher natural gas consumption, increasing the calcium oxide
content from 81% at 1750°F to 86% at 1900°F saved approximately $2.00 per
ton of dry solids fed to the furnace.
The soft burned recalcined lime produced at South Lake Tahoe has,
as previously mentioned, a flour-like texture. Individual particles are not
easily seen without magnification. The large surface area to volume ratio
makes the recalcined lime very easy to slake. Tables 16, 17 and 18 all in-
dicate that the total slaking time was four minutes or less; whereas the
highly reactive 16 x 50 mesh granular makeup lime required seven minutes
for total slaking.
Conclusions. At the 7.5 mgd design flow about 34 tons per day of
dewatered lime would have to be disposed of if lime recovery and reuse
were not practiced. Through lime recalcination the total blow-down of
waste solids is reduced to approximately 1.5 tons of dry solids. The cost
of recalcined lime is slightly more than that of new lime, however, the
reuse has and will avoid the costs of disposal and purchase or production
of C02- Lime recovery and reused at South.Lake Tahoe has demonstrated
the following:
1. Lime recalcination provided not only 72 percent of the lime used
for the past three years, but also a usable source of carbon dioxide.
2. Industrial gravity type thickeners are an effective device for
thickening lime mud containing organic solids with very low weir overflow
losses of usable calcium.
3. A concurrent flow centrifuge can be used to separate or classi-
fy phosphate rich inerts and magnesium from reusable calcium carbonate.
161
-------
4. Three centrifuge bowl speeds, 2200, 1800, and 1600 rpm, were
used to determine optimum bowl speed. All three speeds produced approx-
imately the same percent cake solids and recovery or capture. Consequent-
ly the lowest bowl speed, 1600 rpm, was selected for plant operations to
reduce wear and maintenance costs.
5. The lime centrifuge performed as expected with high solids
captures at high pool depths and lower flows to the machine. Conversely,
at low pool depths and higher flows solids capture was less.
6. At 9 tons of solids to the furnace per day, the optimum recal-
cining conditions were 1900°F on the number 4 and 5 hearths with a 1.5-
2 .0 rpm rabble rate.
7. Of the three parameters, recalcining temperature, rabble rate
and feed rate, temperature had the most effect on recalcined lime activity.
The CaO content in the recalcined lime was increased 15% by raising the
temperatures from 1600°F to 1900°F.
162
-------
SECTION XXV
CAPITAL AND OPERATING COSTS FOR CONVENTIONAL AND ADVANCED
WASTE TREATMENT
Introduction. The increasing interest by the general public in the
quality of the environment has stimulated several questions concerning
water pollution abatement. One of the most significant questions is what
the cost of cleaning up our nation's lakes and rivers will be. In some in-
stances, the addition of secondary treatment will be sufficient to reduce
the problem; however, in other areas one or more of the various advanced
waste treatment processes will be required to eliminate a pollution prob-
lem. A knowledge of the costs for the various degrees of conventional
and advanced waste treatment is essential in planning the nation's water
pollution abatement needs.
The purpose of this section is to present in detail the actual costs
of conventional and advanced waste treatment and to briefly review the
benefits of the plant scale conventional-advanced waste treatment scheme
used continuously since 1968 at South Lake Tahoe.
This section will show that the total cost at 7.5 mgd was $166/mg
for conventional waste treatment and $217/mg for advanced waste treat-
ment. These costs are based on producing the extremely high quality re-
claimed water described in earlier sections with 100 percent reliability.
Lesser requirements should demonstrate lower costs. Thus at South Lake
Tahoe, the cost of advanced waste treatment is approximately 30 percent
greater than the cost of conventional treatment.
The conventional and advanced waste treatment phases, with the
exception of the ammonia stripping tower, are designed for 7.5 mgd. The
ammonia stripping tower design capacity is 3.7 5 mgd.
Conventional treatment includes primary clarification, and both
plug flow and completely mixed activated sludge secondary treatment.
The activated sludge process is operated with a high organic loading and
low mixed liquor suspended solids and sludge age, to prevent nitrification
323
-------
and to keep the majority of the nitrogen in the ammonium ion form. The
mixed liquor is chlorinated at 2 mg/1 before clarification if nitrifying or-
ganisms begin to proliferate. Raw and waste activated sludges are de-
watered in centrifugals and then incinerated.
Advanced waste treatment begins with phosphorus removal and clar-
ification of the secondary effluent, using lime. The spent lime mud is
thickened in a gravity thickener, dewatered by centrifuging, and then re-
calcined in a multiple hearth furnace for reuse. Phosphorus rich lime mud
is classified in the centrifugal and wasted to the organic sludge system.
The effluent from the lime clarifier flows through an ammonia strip-
ping tower to a two-stage recarbonation system. Scrubbed stack gases
from the lime recalcining and sludge incineration furnaces are used to neu-
tralize the high pH water. The recarbonated effluent then is pumped to
mixed media filters and carbon columns. Two ballast ponds, each one
million gallons in capacity, float on the system in order to provide flow
equalization and supplemental filter backwash water. Spent carbon is
withdrawn periodically from the carbon columns, thermally reactivated in
a separate multiple hearth furnace, and then returned to the carbon col-
umns. The carbon column effluent is dosed with 2 mg/1 of chlorine and
then lifted 1,500 feet and through 27 miles to Indian Creek Reservoir in
Alpine County, California.
Assumption for Capital Costs. The capital costs include all equip-
ment and construction costs. They do not include design costs. The
equipment and construction costs were taken from District records of act-
ual contracts awarded for the various treatment phases. These contracts
were completed at various periods between 1960 and 1968. The EPA
Sewage Treatment Plant Construction Cost Index, Base Year 1957-1959 =
100, was used to adjust the capital costs to 1969. It was assumed that
the San Francisco Region Indexes were equivalent to South Lake Tahoe.
All costs were adjusted to 1969 San Francisco Index and then to the Nat-
ional Average Index for 1969. The 1969 Index values used were 136.2 for
San Francisco and 127.1 for the Nation. The 1969 replacement costs per
million gallons were based on the national average at 7.5 mgd design ca-
pacity assuming capital amortization of all the costs at 5 percent for 25
years and no federal assistance. In fact, the capital costs to the District
were much lower since federal grants from the USPHS, EDA, and EPA fin-
anced approximately 46 percent of the total construction cost for conven-
tional and advanced waste treatment.
Capital Costs. The capital costs are shown in Table 55 and also
later in the operational cost tables. Included in the capital cost figures
for a specific treatment phase are items common to several processes.
324
-------
TABLE 55
SOUTH TAHOE PUBLIC UTILITY DISTRICT, CALIFORNIA
CAPITAL COSTS FOR CONVENTIONAL AND ADVANCED WASTE TREATMENT PLANT
7.5 MGD DESIGN CAPACITY
Actual Estimated Estimated
Contract National Average R•placement
Treatment Phase
Total Construction
Cost Per Pha» 1"
Replacement Construction
Cost for 1969
Costs Per Ml
far 1969
CONVENTIONAL treatment
Primary
Activated Sludge
Organic Sludge "1
Chlorination
$ 692,000
1,247,000
583.000
9,000
$ 753,000
1.300.000
545.000
11,000
$ 19.50
33.60
14.10
0.30
TOTAL, Conventional Treatment
S2.531,000
$2,609,000
$ 67.50
ADVANCED TREATMENT
Nutrient Removal
Phosphorus Removal
Lime Treatment
Lime Recalcining
401,000
552,000
378,000
516,000
9.70
13.60
SUBTOTAL, Phosphorus Removal
$ 953,000
$ 894,000
$ 23.20
Nitrogen Removal1,1
327,000
310,000
8.00
Recarbonation
162,000
152,000
4.00
SUBTOTAL, Nutrient Removal
$1,442,000
$1,356,000
$ 35.20
Filtration
705,000
687,000
17.80
Carbon Treatment
Carbon Adsorption 1,1
Carbon Regeneration
656.000
193,000
632,000
199,000
16.30
5.20
SUBTOTAL, Carbon Treatment
S 849.000
$ 831,000
$ 21.50
TOTAL, Advanced Treatment
$2,996,000
$2,874,000
$ 74.50
TOTAL WATER RECLAMATION
Conventional Treatment
Advanced Treatment
2,531,000
2,996,000
2,609,000
2,874,000
67.50
74.50
TOTAL, WATER RECLAMATION
$5,527,000
$5,483,000
$142.00
(2| Construction costi n tsfcen from Dbfricl itcndi of (ctual contncts (warded for i
wioui periods between I960 md 1968. That corti hm oot been idjutfed to • com
rarions phases. CAttracts for eooitnictioa w
moaycar.
trt completed
-------
These items include buildings, electrical systems, piping, and controls
not specifically identifiable in the construction contracts. Both area and
volume relationships are used to proportion this capital cost to specific
treatment phases.
Total capital costs, based on 1969 national average replacement
costs, were 67.50 $/mg for conventional treatment and 74.50 $/mg for
advanced treatment. The 20 acre site where the present conventional and
advanced waste treatment processes are located was acquired in conjunc-
tion with the construction of the original 2.5 mgd primary and secondary
plant in 1960. Since the records did not show the site acquisition as a
separate item, the cost of the 20 acres was included only in the capital
cost of conventional treatment.
Assumptions for Operating Costs. The operating costs included in
this section are based on the plant design capacity of 7.5 mgd from Feb-
ruary 1969 to December 1970. During this period, the actual average mon-
thly influent flows varied between 1.79 mgd and 3.15 mgd. As a result of
recycling water from scrubber flows, backwashing filters, and backflowing
carbon columns, the filtration and carbon adsorption flows during the same
period varied between 3.13 mgd and 5.22 mgd.
To compare operating costs on the common basis of the plant design
capacity, it was assumed the total cost for fuel, chemicals, make-up lime,
and make-up carbon would increase in proportion to the flow. However,
the same assumption could not be made for electricity, operating and main-
tenance labor costs, equipment repair, and instrument maintenance. Costs
per day for electricity were adjusted upward to reflect the equipment char-
acteristics at 7 .5 mgd. The cost per day for operating and maintenance
labor, repair materials, and instrument maintenance were assumed to be
the same as at present and design flows.
The assumption that the plant is staffed as though it were operating
at design capacity was made for the following reasons. During the grant,
three individuals were required per shift to operate the plant. One oper-
ator spent approximately five hours in the laboratory analyzing samples
for the District's FWQA research grant, and three hours performing plant
related duties. The second operator spent an hour per shift measuring
flows and keeping track of expendables for the grant, and seven hours in
plant operation. The third individual controlled the operations within the
incineration building, which did not vary with flow. This extensive data
collection would not be taking place at 7.5 mgd nor would it be the usual
practice at other advanced waste treatment plants. Maintenance labor
and repair materials would be affected by equipment age but not necess-
326
-------
arily by flow since most of the equipment is running today. Because it
was the basic purpose of this section to show costs at 7.5 mgd in 1969
and 1970, age was not considered to be a factor.
It was assumed, however, that emphasis placed on each treatment
phase by the first two operators would be different during carbon regener-
ation periods when this additional phase must be covered.
No District operating and capital costs associated with sewers,
pump stations, janitorial work in administrative areas, minor plant modi-
fications, effluent export, FWQA research grant, and general Utility Dis-
trict administration were included. It is felt that these costs were not
typical operating costs for actual waste treatment.
Operating Cost Data Collection and Analysis. Each month the com-
puter provided machine listings of the individual operating costs per treat-
ment phase for current and design plant flows. The computer also deter-
mined and printed the average to date costs for the current and design
flows.
The monthly labor rates for the operations and maintenance included
the actual monies paid by the District to or in behalf of the individual for
straighttime, overtime, standbytime, holidays, vacation, sick leave, pre-
mium pay, Social Security, retirement fund, medical insurance, unemploy-
ment tax, and Workmen's Compensation. The labor rates represented all
monies paid divided by the actual hours worked.
Two methods were used to allocate labor costs to the various treat-
ment phases. For maintenance labor, the actual hours spent within a
specific treatment phase were fed into the computer. Overhead time, such
as supervision, lunch, coffee breaks and unreported time, was prorated
among the various treatment phases on the basis of actual hours reported.
The maintenance group included six individuals.
Operating labor was divided among the various treatment phases by
fixed percentages. Three shifts were used, seven days per week. The
basic shift included three operators. A total of 15 operators were needed
to cover the three shifts, chief operator, and vacation and sick leave make-
up. The fixed percentages used to allocate the operational labor hours
are shown in Table 56 . The fixed percentages were based on personal in-
terviews with each of the operators, and on observations by the authors
and the District administrative staff. Again, these percentages were stor-
ed in the computer as variable constants.
327
-------
TABLE 56
PERCENT OPERATIONAL LABOR DIVISION
PER TREATMENT PHASE
Treatment Phase
PLANT EXCLUDING INCINERATION BUILDING - 2 Operator^Shift
Without Carbon
Regeneration
During Carbon
Regeneration
u
to
CD
Primary Treatment
Secondary Treatment
Lime Clarification
Ammonia Stripping
Recarbonation
Filtration
Carbon Adsorption
Carbon Regeneration
INCINERATION BUILDING - 1 Operator/Shift
Organic Sludge Dewatering
Organic Sludge Incineration
Lime Mud Dewatering
Lime Mud Recalcining
Lime Clarification (Slakers)
Rfecarbonation (CO2 Compressors)
20.80
25.00
.17
.10
.08
,25
6.25
6.25
4.17
4
12
4
2
17
50
17
09
13.53
17.70
.83
.42
.83
4.17
4.17
25.00
6,
4,
4,
12 ,
4,
2 ,
25
17
17
50
17
09
TOTAL
100.00
100.00
-------
Operator responsibilities include process monitoring and adjust-
ments, chemical mixing, routine equipment servicing, and inside cleanup.
Electrical costs were divided among the treatment phases on the
basis of ampere-hours per month. The actual running time for each elec-
trical motor was one of the inputs for each month. The amperes for each
motor were stored in the computer as variable constants.
Cubic feet of natural gas used by the sludge incinerator, lime re-
calciner and carbon furnace and the amounts of chemicals used each month
were each measured separately.
Instruments were maintained by an outside contract. The costs per
treatment phase were determined by the number of instruments repaired or
calibrated per treatment phase. The contract included labor, parts, and
instrument replacement.
Repair material costs included replacement costs, material, repair
equipment purchase, and rental costs. This category also reflected out-
side labor cost, as well as material costs for work done outside the plant,
such as rewinding electric motors.
Unit Operating Costs. Costs of all commodities at Lake Tahoe are
high in comparison to the average of commodity costs in the USA. This is
due in large measure to the location in a mountainous area and to the Ta-
hoe area's tourist-based economy. In order to permit comparison with
costs which might be anticipated in other areas, Table 57 has been prepar-
ed to show the present cost of various commodities at Lake Tahoe. These
unit costs were stored in the computer as variable constants.
Primary Treatment. Raw sewage entering the plant passes through
a barminutor, a parshall flume and then into the primary clarifier. A rec-
tangular 2.7 mgd clarifier equipped with water spray scum collection and
a 4.8 mgd( 100 foot diameter circular clarifier with mechanical scum coll-
ection can be used. The underflow is degritted with a cyclone degritter.
Table 58 shows the operating and capital costs for primary treatment.
The barminutor, primary clarifier, degritter, and sludge and scum withdraw-
al pumps were considered to be part of primary treatment.
Secondary Treatment. The primary effluent flows by gravity to the
activated sludge secondary treatment system. This system consists of
three plug flow aeration basins (0.9 mgd each), two completely mixed
aeration basins (2 A mgd each) all in parallel, and two circular secondary
clarifiers, 2.0 and 5.5 mgd, respectively. Activated sludge is wasted to
329
-------
TABLE 57
UNIT COSTS^1) 1969 AND 1970
Labor <2)
Operations
Maintenance
Electricity^)
Fuel ^
Make-up Lime (Quicklime)^)
Chemicals
Chlorine
Polymer - Sludge Dewatering
Polymer - Lime Coagulation
Polymer - Filtration
Alum - Filtration (Liquid)(®)
Activated Carbon Make-up^
$ 6.11 /hour
5.05 /hour
12.10/1,000 kwh
0.0543/therm
28.83/ton CaO
114.00/ton
2.53/lb
1.92/lb
1.92/lb
0.014/lb
0.305/lb
(1) All appropriate unit costs are f.o.b. South Lake Tahoe and include
a 5% California sales tax.
(2) Labor costs include all direct and indirect monies paid upon the
employees behalf. The rates are averages for 1969 and 1970.
(3) Includes energy and demand charges.
(4) Natural gas at about 860 BTU/cu ft at 6,200 feet elevation and
billed on the basis of interruptable service.
(5) Average calcium oxide content 93.6%.
(6) Liquid alum weight 11.08 lbs/gal. Dry alum equivalent 49%.
(7) Activated carbon at 30 lbs/cu ft.
330
-------
Overall Costs - Organic Sludge Dewaterinq, Incineration, and
Disposal. The operating and capital costs shown in Table62 represent
the total cost for organic sludge handling and disposal excluding dump-
ing charges.
Lime Coagulation. Chemical coagulation (with lime) of the secon-
dary effluent to pH 11 is accomplished with a rapid-mix flocculation basin,
followed by a 100 foot diameter conventional clarifier. To reach pH 11 for
the South Lake Tahoe wastewater, a lime dose of 300 mg/1 of calcium
oxide is required. A polymer, at a 0.1 to 0.3 mg/1 dose, is added just as
the water leaves the flocculation chamber to improve clarification.
The lime coagulation system typically removes 9 5 percent of the
phosphorus it receives. In the clarifier effluent, phosphorus concentra-
tions range between 0.2 and 0.7 mg/1 PO4-P and turbidity levels between
1 and 10 SJU.
Table 63 shows the specific operating and capital costs for this
treatment phase. The lime storage bins, slakers, floe basin, clarifier,
polymer feed system, and sludge draw-off pumps were considered to be
part of the lime coagulation system. As shown in Table 56, operating lab-
or from both the general plant and the incineration building were charged
to lime clarification. Maintenance labor and repair material costs for
1969 and 1970 represented primarily costs associated with the slakers,
lime buildup on the flash mixer, sludge removal from the lime mixing bas-
in, and cleaning of the lime slurry line from the slakers to the flash mix-
ing basin.
Lime Mud Dewaterlng. Lime mud is pumped from the chemical clar-
ifier and reaction basin to a gravity thickener for solids concentrations.
In turn, the thickened mud is pumped by either of two variable speed pump
units to a 24-inch x 60-inch solid bowl centrifugal for dewaterlng. Phos-
phorus is wasted from the system in the form of calcium hydroxyapatite by
operating the lime centrifugal so that 10-30 percent of the solids entering
the machine come out in the centrate. A second centrifugal clarifies the
centrate and the phosphorus-rich cake is conveyed to the organic sludge
furnace.
Operating and capital costs for the dewaterlng phase of lime recal-
cination are shown in Table 64 . The items of equipment included in this
phase were the thickener, lime sludge pumps to the centrifugal, the cen-
trifugal itself and lime cake conveyor to the furance. In 1969 and 1970,
most of the maintenance labor and repair costs were related to lime mud
line cleaning and centrifugal repair. The dewaterlng costs also included
the second centrifugal when it was being used to dewater lime centrate for
336
-------
TABLE 63
OPERATING AND CAPITAL COSTS
LIME COAGULATION AT 7 . 5 MGD
OPERATING COSTS $/DAY
Electricity 5.22
Make-up Lime 168.19
Polymer 22.18
Operating Labor 30.67
Maintenance Labor 9.59
Repair Materials 3.14
Instrument Maintenance 3.04
Total Operating Cost 242 .03
TOTAL COSTS PER MG $ / MG
Operating 32.27
Capital 9.70
Total 41.97
338
-------
TABLE 64
OPERATING COSTS
LIME MUD DEWATERING AT 7.5 MGD
$/DAY
8.92
21. 47
16. 04
8. 81
0. 00
55. 24
LIME MUD DEWATERING
Electricity
Operating Labor
Maintenance Labor
Repair Materials
Instrument Maintenance
Total Operating Cost
TOTAL OPERATING COST
Per MG Plant Influent $ 7.37/MG
Per ton CaO Recalcined 5.49/ton
CaO
339
-------
the lead centrifugal.
The costs for drying and disposing of the dewatered high phosphate
lime mud were included in the costs for organic sludge incineration and
ash disposal.
Lime Mud Recalcining. Dewatered lime mud is conveyed from the
centrifugal to a 14.3 foot diameter,six hearth furnace for recalcining at
1800-1900°F. The maximum capacity of the furnace is about 20 tons of
dry solids per day. The recalcined lime exits by gravity through a crusher
into a thermal disc cooler, and then is pneumatically conveyed to the re-
calcined lime storage bin for eventual reuse.
For the purpose of determining operating and capital costs, all
equipment and functions from the lime furnace to the lime storage bin were
considered to be part of the lime recalcining system. These costs are
shown in Table 65. The majority of the maintenance costs were associat-
ed with the conveying of recalcined lime from the furnace to the storage
bin.
Overall Costs - Lime Dewatering and Recalcining. The operating
and capital costs shown in Table 66 represent both dewatering and recal-
cining costs. If lime mud dewatering costs are considered common to any
lime clarification process, then the costs of lime recalcining are compet-
itive with buying new lime, particularly when disposal costs are included.
At South Lake Tahoe new lime was purchased at $28.83/ton CaO, whereas
lime recalcination cost $31.61/ton CaO.
Nitrogen Removal by Ammonia Stripping. Following lime clarifica-
tion, ammonia stripping is utilized at South Lake Tahoe to remove the nut-
rient nitrogen from the wastewater. The stripping process includes two
constant speed pumps, a cross flow cooling tower with a two-speed re-
versible 24-foot fan, a concrete collection basin below the tower, and a
flow measurement weir on the basin exit. The tower has an average de-
sign air-to-water ratio of 250 cu ft/gal, and a nominal capacity of 3.75
mgd.
Tower removal efficiencies have varied from 30 to 90 percent, de-
pending on air temperature and extent of calcium carbonate buildup on the
fill before cleaning. During the winter when the air temperature is lower
than 32°F (0°C.), the tower is bypassed to prevent ice buildup on the
fill. Tower influent NH3-N concentrations have ranged from 15 to 30 mg/1,
with effluent values from 3 to 15 mg/1.
340
-------
TABLE 65
OPERATING COSTS
LIME MUD RECALCINING AT 7.5 MGD
LIME MUD RECALCINING $/DAY
Electricity 6.55
Natural Gas 142 .83
Operating Labor 64.36
Maintenance Labor 12 .68
Repair Materials 8.31
Instrument Maintenance 3 .63
Total Operating Cost 238 .36
TOTAL OPERATING COST
Per MG Plant Influent $ 31.78 MG
Per Ton CaO Recalcined 23 .70/ton
CaO
341
-------
TABLE 66
OPERATING AND CAPITAL COST
LIME MUD DEWATERING AND RECALCINING
AT 7 .5 MGD
LIME MUD DEWATERING $/MG $/Ton CaQ
Operating Costs 7.37 5.49
Capital Costs 2.90 2.16
Total 10.27 7.65
LIME MUD RECALCINING
Operating Cost 31.78 23.70
Capital Cost 10.60 7.91
Total 42.38 31.61
TOTAL OPERATING AND CAPITAL COSTS 52.65 39 .26
342
-------
SECTION XXVII
REFERENCES
OPERATOR TRAINING.
1. Anon., "Elementary Mathematics and Basic Calculations", Water &
Sewage Works Magazine, Scranton Publ. Corp., Chicago, Illinois.
2. Anon., "Operator Short Course", Water & Wastes Engineering, New
York.
3. Culp, Russell L., "The Operator of Wastewater Treatment Plants",
Public Works Magazine, Ridgewood, N. J. (1970).
4. Texas Water & Sewage Works Association, Austin, Texas, Manual for
Sewage Plant Operators.
5. W.P.C.F., Washington, D.C., "Operation of Wastewater Treatment
Plants", WPCF Manual of Practice No. 11.
6. W.P.C.F., Washington, D.C., "Wastewater Treatment Plant Opera-
tor Training Course Two", WPCF Publication No. 14.
TEST PROCEDURES.
7. American Public Health Assoc. , New York City, Standard Methods for
the Examination of Water and Wastewater. 12th Edition.
INCINERATION OF WASTE SOLIDS.
8. Albertson, O. E. and Guidi, E. E., Jr., "Centrifugation of Waste
Sludges", Journal Water Pollution Control Federation, p. 607 (1969).
9. Alford, J. M. , "Sludge Disposal Experiences at North Little Rock,
Arkansas", Tournal Water Pollution Control Federation, p. 175 (1969).
10. Bird Machine Co., South Walpole, Mass., Operating Manual, (1966).
365
-------
11. Blattler, Paul X. , "Wet Air Oxidation at Levittown", Water & Sewage
Works, p. 32 (February, 1970).
12. Cardwell, E. C., "Dewatering by Mechanical Means", Proc. 10th
San. Eng. Conf.
13. Ettelt, G. A., and Kennedy, T. J., "Research and Operational Exper-
iment in Sludge Dewatering at Chicago", Tournal Water Pollution Con-
trol Federation, p. 248 (1966).
14. Genter, A. L., "Computing Coagulant Requirements in Sludge Condi-
tioning", Transactions, American Society of Civil Engineers, p. 641,
(1946).
15. Jones, W. H., "Sizing and Application of Dissolved Air Flotation
Thickeners", Water & Sewage Works, Ref. No. R—177—194 (Nov. 1968).
16. Jordan, V. J., and Scherer, C.H., "Gravity Thickening Techniques at
a Water Reclamation Plant", Tournal Water Pollution Control Feder-
ation, p. 180 (1970).
17. McCarty, P.L., "Sludge Concentration-Needs, Accomplishments, and
Future Goals", Journal Water Pollution Control Federation, p. 493(1966).
18. Nichols Engineering Co., Bulletin No. 238A, "Sludge Furnaces".
19. Schepman, B. A., and Cornell, C.F., "Fundamental Operating Vari-
ables in Sewage Sludge Filtration", Sewage and Industrial Wastes,
p. 1443 (1956).
20. Sharman, L. , "Polyelectrolyte Conditioning of Sludge", Water and
Wastes Engineering, p. 8 (1967).
21. Tenney, M. W. , and Cole, T.G., "The Use of Fly Ash In Condition-
ing Biological Sludges For Vacuum Filtration", Tournal Water Pollution
Control Federation, p. R281 (1968).
22. Tenney, Mark W., et al, "Chemical Conditioning of Biological Sludges
for Vacuum Filtration", Tournal Water Pollution Control Federation,
p. Rl, (1970).
366
-------
23. Tomas, C. M., "The Use of Filter Presses for the Dewatering of Sew-
age and Waste Treatment Sludges", Paper presented at the 42nd Annual
Conference of the WPCF, Dallas, Texas (October, 1969).
CHEMICAL TREATMENT AND WASTE SOLIDS.
24. Camp, T. R. , "Flocculatlon and Flocculatlon Basins", Transactions
American Society of Civil Engineering, p. 1 (1955).
25. Camp, T. R. , "Floe Volume Concentration", Journal American Water
Works Association, p. 656 (1968).
26. Chemical Feeder Guide, BIF Co., Providence, Rhode Island (1969).
27. Hudson, H. E., "Physical Aspects of Flocculation", Tournal American
Waterworks Association, p. 855 (1965).
28. LaMer, V. K. , and Smellie, R. H., Jr., "Flocculation, Subsidence,
and Filtration of Phosphate Slimes", I. General, Tournal Colloid
Science, p. 704 (1956).
29. Lea, W. L., Rohlick, G. A., and Katz, W. J., "Removal of Phosphates
from Treated Sewage", Sewage and Industrial Wastes, p. 261 (1954).
30. Mulbarger, M.C., etal, "Lime Clarification, Recovery, Reuse, and
Sludge Dewatering Characteristics", Tournal Water Pollution Control
Federation, p. 2070 (1969).
31. Morgan, J.J. and Englebrecht, R.S., "Effects of Phosphates on Co-
agulation and Sedimentation of Turbid Waters", Tournal American Water
Works Association, p. 303 (1960).
32. O'Melia, C. R. , "A Review of the Coagulation Process", Public Works
p. 87 (May 1969).
33. Rose, J.L., "Removal of Phosphorus By Alum", Presented at the FWPCA
Seminar on Phosphate Removal, Chicago, Illinois (June 26, 1968).
34. Sawyer, C.N., "Some New Aspects of Phosphates in Relation to Lake
Fertilization", Sewage and Industrial Wastes, p. 768 (1952).
35. Sebastian, Frank, and Sherwood, Robert, "Clean Water and Ultimate
Disposal", Water & Sewage Works, (August 1969).
367
-------
36. Schmid, L.A., and McKlnney, R.E., "Phosphate Removal by a Lime-
Biological Treatment Scheme", Tournal Water Pollution Control Feder-
ation, p. 1259 (1969).
37. Stumm, W. and Morgan, J.J., "Chemical Aspects of Coagulation",
Journal American Water Works Association, p. 971 (1962).
38. Walker, J.D., "High Energy Flocculation", Tournal American Water
Works Association, p. 1271 (1968).
39. Wuhrmann, K., "Objectives, Technology, and Results of Nitrogen and
Phosphorus Removal", Advances in Water Quality Improvement, I,
University of Texas Press, Austin, Texas (1968).
40. Wukasch, R. F., "The Dow Process for Phosphorus Removal", Present-
ed at FWPCA Seminar on Phosphate Removal, Chicago, Illinois (1968).
RECARBONATION.
41. Anon., "Submarine Burners Make CO2 For Softening Recarbonation",
Waterworks Engineering, p. 182 (1963)..
42. Compressed Gas Assoc., Pamphlet G-6, "Carbon Dioxide", 2nd Edi-
tion, New York (1962).
43. Compressed Gas Assoc., Pamphlet G-6 IT, "Tentative Standard for
Low Pressure Carbon Dioxide Systems At Consumer Sites", New York
(1966).
44. Culp, R.L., and Culp, G.L., Advanced Wastewater Treatment, Van
Nostrand Reinhold Co., 450 W. 33rd St., New York, N.Y. (1971).
45. Fair, M.F. and Geyer, J.C., Water Supply and Waste Disposal, John
Wiley & Sons, Inc., New York, N.Y. (1954).
46. Handbook of Compressed Gases, Compressed Gas Assoc. , Reinhold
Book Corp., New York (1962).
47. Haney, PaulD. and Hamann, Carl L. , "Recarbonation and Liquid Car-
bon Dioxide", Journal American Water Works Association, p. 512 (1969).
48. Hoover, Charles P., Water Supply and Treatment, Eighth Edition,
Bulletin 211, National Lime Association, Washington, D. C.
368
-------
49. Ross, R.D., Industrial Waste Disposal, Reinhold Book Corp., New
York, N.Y. (1968).
50. Ryznar, John W., "A New Index For Determining Amount of Calcium
Carbonate Scale Formed by a Water", journal of the American Water
Works Association, (April 1944).
51. Scott, L.H. , "Development of Submerged Combustion For Recarbona-
tion", lournal American Water Works Association, p. 93 (1940).
52. Walker Process Co., Bulletin No. 7-W-83, Carball CO? For Recar-
bonation, (May 1966).
LIME RECOVERY AND REUSE.
53. Aultman, W.W. , "Reclamation and Reuse of Lime in Water Softening",
Tournal American Water Works Association, p. 640 (1969).
54. Black, A.P., and Eidsness, F.A., "Carbonation of Water Softening
Plant Sludge", Journal American Water Works Association, p. 1343(1957).
55. Crow, W.B., "Techniques and Economics of Calcining Softening Slud-
ges-Calcination Techniques", Tournal American Water Works Assoc*
p. 322, (1960).
56. Eades, J.L. , and Sandberg, P.A., "Characterization of the Properties
of Commercial Lime by Surface Area Measurements and Scanning Elec-
tron Microscopy", The Reaction Parameters of Lime, ASTM STP 472,
American Society for Testing and Materials, pp 3-24 (1970).
57. Mulbarger, M.C., et al, "Lime Clarification, Recovery, Reuse and
Sludge Dewatering Characteristics", Journal Water Pollution Control
Federation, p. 2070 (1969).
58. Nelsen, F.G., "Recalcination of Water Softening Plant Sludge", Jour.
American Water Works Association, p. 1178 (1944).
NITROGEN REMOVAL.
59. Cillie, G.G., et al, "The Reclamation of Sewage Effluents for Domes-
tic Use", Third International Conference on Water Pollution Research,
Munich, Germany, Section II, Paper I, WPCF, Washington, D.C.(1966).
60. Culp, R.L., "Nitrogen Removal by Air Stripping", Presented at the 2nd
Annual Univ. of Cal. Sanitary Engineering Research Lab. Workshop,
Tahoe City, California, (June 26, 1970).
369
-------
- -
LIME CLARIFICATION, RECOVERY AND REUSE
Precipitation
++
5Ca + UOH" + 3HPO4"
Ca(0H)2 + Ca(HC03)2
-1-+
Mg + Ca(OH),
Ca50H(P01|)3^ + 3H20
2CaC0^ + 2HpO
Mg(OH) A/ + Ca
++
Recarbonation
4 +
Ca + 20H + C02 + H20
Mg(OH) + 2C02 ^
CaCO-,4^ + 2HpO
Mg++ + 2HCO-
Recalcination
CaC03 > CaO + C02
Mg(0H)o
MgO + HgO
Hydration
CaO-MgO + H20 > MgO + Ca(OH)2 + heat
TABLE 1
-------
FIGURE 1
EXPERIMENTAL SYSTEMS
("C" AND"D")
SLAKE
MIX
FLOC
SETTLE
co2_^
MIX
FLOC
SETTLE
^ -
("B" AND"D")
MAKE UP S C°2 mmm*
LIME
U
GRAVITY
THICKENER
REFERS TO
FLOW STREAM,
SEE TABLE 2
VACUUM
FILTER
WATER
WASTE
FURNACE
SYSTEM "A"
TOTAL TREATMENT
SYSTEM "B"
TOTAL TREATMENT WITH SLUDGE CARBONATION
SYSTEM "C"
SPLIT TREATMENT (0.5Q to 2"d STAGE)
SYSTEM "D"
SPLIT TREATMENT WITH SLUDGE CARBONATION
NOTE: LIME DOSE FOR SYSTEMS "C" AND "D" IS Yi OF
THE LIME DOSE USED FOR SYSTEMS "A" AND "B"
-------
FIGURE 2
LIME REQUIREMENT FOR pHall.O AS A FUNCTION
OF THE WASTEWATER ALKALINITY
o
(Q
o
-J
GO
E
1
500
o
t-H
rH
Al
X
Cl
400
or
O
Li_
Q
C>
1 ¦ 1
300
LiJ
cn
2
O
1—
<
200
cn
i—
2
LiJ
O
2
O
o
100
LiJ
_J
O- DATA FROM THIS INVESTIGATION
- DATA FROM OTHERS WITH REFERENCE
0 100 200 300 400 500
WASTEWATER ALKALINITY mg/L-CaC03
-------
FIGURE 4
SLUDGE THICKENING CHARACTERISTICS
vs.
MAGNESIUM HYDROXIDE/CALCIUM CARBONATE RATIO
cr 40
ii
o
O o:30
O LU
3 j~
_l u-
co <
Q
LU
20
h- O
i— en
UJ LU
CO Q_
MEAN ± STANDARD DEVIATION
0 0.10
Mg(0H)2/CaC03
0 0.10
Mg(0H)2/CaC03
-------
CVJ
<;
CO
Q
JO
I
LlI
I—
<
o:
cr
LlJ
FIGURE 5
LIME DOSE vs. FILTER YIELD
MEAN ± STANDARD DEVIATION
IMMERSION FRACTION = 0.25
VACUUM = 15" Hg
3 RECYCLES
- -NO RECYCLES
x
3 RECYCLES
NO RECYCLE
ii
100 200 300 400 500
LIME DOSE mg/L CaO TO ACHIEVE A pH>11.0
-------
FIGURE 6A
SYSTEM "A"
TOTAL TREATMENT
-------
FIGURE 6B
SYSTEM "B"
TOTAL TREATMENT WITH SLUDGE CARBONATION
1200
1200
1000
1000
800
800
400
400
200
200
0
LEB.
LEB.
ADJ. P04 ¦ P
LEB.
ADJ. ALK.
BAT.
ADJ. ALK.
-------
FIGURE 7
SLAKING CHARACTERISTICS OF RECOVERED
LIME AFTER THREE RECYCLES
-97% CaO (STANDARD)
.71% CaO (LEBANON "B")
56% CaO (LEBANON "C")
,22% CaO (BATAVIA "D")
I I
4 6
TIME (minutes)
8
10
-------
- 15 -
references
1. Slechta, A. P., and Culp, G. L., "Water Reclamation Studies at the
South Tahoe Public Utility District." Jour. Water Pollution Control
Federation. 22, 5, 787 (May 1967).
2. Machis, A., "What Third-Stage Sewage Treatment Means." American City,
110 (September 1967)*
3. Sawyer, C. N., "Some Aspects of Phosphates in Relation to Lake
Fertilization." Sev. and Ind. Wastes, 2b, 768 {1962).
b. Rudolfs, W., "Phosphates in Sewage and Sevage Treatment. I. Quantities
of Phosphates." Sev. Wks. Jour., 1£, U3 (19^7).
5* Owen, R., "Removal of Phosphorus from Sewage Plant Effluent with Lime."
Sew, and Ind. Wastes, 25, 5^ (1953).
6. Malhotra, S. K., Lee, G. F., and Rohllch, G. A., "Nutrient Removal
from Secondary Effluent by Alum Flocculation and Lime Precipitation."
Int. J. Air Water Poll., 8, U87 (1964).
7. Karanik, J. M., and Nemerov, N. L., "Removal of Algal Nutrients."
Water and Sewage Works, 112, 460 (1965).
8. Rand, M. C., and Nemerov, N. L., "Removal of Algal Nutrients from
Domestic Wastewater." Report No. 9, Department of Civil Engineering,
Syracuse University Research Institute (1965).
9. Buzzell, J. C., and Sawyer, C. N., "Removal of Algal Nutrients from
Raw Wastewater with Lime." Jour. Water Pollution Control Federation,
^2, Rl£ (October 1967).
10. Stander, G. J., "Final Report on Reclamation of Purified Sewage
Effluent for Augmentation of Domestic Water Supply of Windhoek."
Rational Institute for Water Research Council for Scientific and
Industrial Research, CSIR Report No. CWATU, Pretoria, South Africa
(1965).
U. Bishop, D. F., Marshall, L. S., O'Farrell, T. P., O'Connor, B.,
Dobbs, R. A., Griggs, S. H., and VlUlers, R. B., "Studies on
Activated Carbon Treatment." Jour. Water Pollution Control Federa-
tion. 22, 188 (February 1967).
12. Van Wazer, J. R., "Phosphorus and Its Compounds." Vol. I., Intersclence
Publishers, New York, N. Y. (1958).
-------
- 16 -
13. Stumm, W., "Discussion on: Rohlich, G. A., "Methods for the Removal of
Phosphorus and Nitrogen from Sewage Plant Effluents." Advances in Water
Pollution Research, Vol. 2, Pergamon Press, The MacMillan Company, New
York, N. Y. (196U).
14. Dickerson, B. W., and Farrell, P. J., "Laboratory and Pilot Plant Studies
on Phosphate Removal from Industrial Waste Water." Uoth Annual Con-
ference of the Water Pollution Control Federation, New York, N. Y.
(October 1967).
15. Russell, G. D., and Russell, G. S., "The Disposal of Sludge from a
Lime-Soda Softening Plant as Industrial Waste." 9th Annual Industrial
Waste Conference, Purdue University, Lafayette, Indiana (May 10-12, 195*0 •
16. Black, A. P., "Disposal of Softening Plant Wastes - Lime and Lime-Soda
Sludge Disposal." Jour. American Water Works Association, kl, 9, 819
(19^9).
IT. Black, A. P., and Eidsness, F. A., "Carbonation of Water Softening
Plant Sludge." Jour. American Water Works Association, k9, 13^3 (1957)•
18. Nelson, F. G., "Recalcination of Water Softening Sludge." Jour. American
Water Works Association, 36, 1178 (19WO.
19. "Investigation of Recovery and Disposal of Solids from the Water Treat-
ment Process." Minneapolis City Council Water Department, Sanitary
Engineering Report No. 127-5 (1959)•
20. Crow, W. B., "Techniques and Economics of Calcining Softening Sludges -
Calcination Techniques." Jour. American Water Works Association, 52,
1, 322 (I960).
21. "Lime - Handling, Storage, and Use in Water and Wastewater Treatment."
BIF, No. 1, 21-24, Providence, Rhode Island (1962).
22. ASTM Specification C25-58.
23. Jackson, M. L., "Soil Chemical Analysis." Prentice-Hall, Inc.,
Englewood Cliffs, N. J. (i960).
-------
TvVO CLARIFIER LIME CLARIFICATION PROCESS
WITHOUT CHEMICALS
Capital Cost, Operating & Maintenance Cost, Debt Service
vs.
Design Capacity
Cost Adjusted to March, 1969
na
10.0 10
0.10
20.0
10.0
1.0
t> 7 3 9 10
10.0
S 6 7 8 9 10
100.
0.10
Design Capacity, millions of gallons per day
C = Capital Oost, millions of dollars
A = Debt Service, cents per 1000 gallons (4J5 - 25 yr.)
O & M = Operating and Maintenance Cost, cents per 1000 gallons
T = Total Treatment Cost, cents per 1000 gallons
16
Figure 4
-------
LIME RECALCINATION PLUS MAKE UP LIME
FOR USE WITH LIME CLARIFICATION
Capital Cost, Operating & Maintenance Cost, Debt Service
vs.
Design Capacity
Cost Adjusted to March, 1969
10.0 10
20.0
10.0
1.0
4 5 6 7 M 10
0.10
1.0 10.0 100.
Design Capacity, millions of gallons per day
C = Capital Cost, millions of dollars
A = Debt Service, cents per 1000 gallons (4% - 25 yr.)
0 tc M = Operating and Maintenance Cost, cents per 1000 gallons
T = Total Treatment Cost, cents per 1000 gallons
20
Figure 5
-------
FILTRATION
-------
CHAPTER 9
DEEP BED FILTRATION
9.1 General
With the exception of gravity sedimentation, deep-bed filtration is the most widely used
unit process for liquid-solids separation. Until recently its use was generally confined to the
treatment of municipal and industrial water supplies. The primary reason fpr its recent
adoption in wastewater treatment has been the need to upgrade effluents from conventional
treatment plants. Installations may use direct filtration of activated sludge or trickling filter
effluents, without the addition of chemical agents. Also, deep-bed filters are employed in
systems for phosphorus removal from secondary effluents and in physical-chemical systems
for the treatment of raw wastewater. In these latter cases chemical coagulation, flocculation
and sedimentation precede the filters as in water treatment plants.
In essence, the unit process of deep-bed filtration encompasses exhaustion of the bed
followed by a regeneration. Water containing suspended solids is passed through a bed of
granular material resulting in deposition of the suspended solids in the bed. Eventually the
pressure drop across the bed becomes excessive or the ability of the bed to remove sus-
pended solids is impaired. Thereupon filtration is stopped and the bed is cleaned prior to
being placed back in service.
9.2 Filter Design
At the present time virtually all deep bed filters utilized for waste treatment are "rapid
sand" type, i.e., downflow, static bed, with batch or semicontinuous operation. In this
section, design information on this traditional type of filter will be reviewed. In a later
section of this chapter, descriptions and design information for some new concepts in deep
bed filtration will be given.
Figure 9-1 (3) is a cut-away view of a typical "rapid sand type" filter, illustrating most of its
components. In essence the filter is a box containing filter media, an underdrain system, a
backwash system, flow control systems and various conduits for bringing feedwater and
wash water to and conveying filtrate and used wash water away from the filter. Figure 9-2
illustrates a pressure filter. It can be seen that there is little difference between the pressure
and gravity flow filters except for the pressure housing. Because of size restrictions on
pressure filters they are equipped with a simpler wash water collection trough system than
gravity filters.
9.2.1 Design Parameters
Process design of a filter includes determination of:
a. type and size of filter media
b. depth of filter
c. rate, duration and timing of water backwash, air scour and surface wash
9-1
-------
Rate of flow and loss
Operating
table
Operating
floor
Pipe gallery
floor
Filter drain
Filter to waste
Wash line
Perforated
laterals
Filter bed wash -
water troughs
Influent to filters
Wash troughs
Filter sand
Graded gra
Concrete filter
tank
Pressure lines to
hydraulic valves frc
operating tables
Effluent to
clear well
Cast-iron,
manifold
Figure 9-1 TYPICAL RAPID SAND FILTER
PRESSURE
GAUGES
RAW WATER
INLET
\,
MANHOLE
INIET BAH IE
FILTERED
WATER
OUTLET
MANUAL X"
Mb IT i PORT
VALVE
BACKWASH LINE
AND RATE SET VALVE
RINSC LINE ANO
RATE SET VALVE
OWE
GRADE
FILTER
MEDIA
STPAfNf P
' ST AO
'DOuBU
DISH
UNDERDRAiN
,STRUCTURAL
LEGS
(Courtesy of Permutit Co.. Paramus, N.J.)
Figure 9-2 PRESSURE FILTER
9-2
-------
d. filtration rate
e. type of chemical pretreatment and dose requirement
f. expected duration of filter runs.
Values for each of these parameters must be selected to yield a design which will produce an
effluent of the desired quality at a minimum cost.
The difficulty in arriving at an optimum process design is twofold. Many, if not all, of these
parameters are interdependent. Also, the present level of filtration theory can only semi-
quantitatively represent the interdependence of the process design parameters.
Process design information should be generated from pilot plant data. At best, present
filtration theory can only reduce the degree of pilot plant study required. All too often
process design of a filtration installation is performed according to "rule of thumb" ex-
perience. Use of such a procedure usually results in an overdesign (uneconomical) or an
underdesign (system failure to meet quality or quantity objectives.)
9.2.2 Pilot Plant Studies
A rather large number of interdependent parameters must be considered in pilot plant
studies. In addition, the range of variation of each which may have to be investigated is
quite large. Thus, a field pilot investigation cannot generally be of short duration. Pilot
studies up to a year in length may be required, but it should be possible to decrease this
time period by utilizing experience as a guide in establishing limits of variation of the
parameters.
Pilot studies can be physically described as consisting of a series of filter columns using an
available supply of feedwater. Provisions must be made to measure effluent quality with
time, pressure drops over various sections of the bed and over the entire bed with time and
backwash rates required to achieve specified degrees of media expansion. The studies should
be conducted at several flow rates, with several media types and size ranges, to various
terminal headloss values and with a variety of chemical pretreatments where appropriate.
The data resulting from these studies should be utilized to eliminate from further considera-
tion those situations which obviously cannot meet effluent criteria or cannot be justified
economically. Based on these analyses, a narrowed-down set of parameters can be used in a
second set of investigations. This second study may be more comprehensive than the first
series, including such variables as expected diurnal variation and slight changes in chemical
dose. This process may continue through several more rounds before an "optimum design"
is achieved. This final design should be given a lengthy test (one month or more) at con-
ditions as close as possible to those expected in the field. Even after this lengthy study, it is
probable that only a good approximation of field conditions will be obtained.
Specific details for conducting filtration pilot studies have not been established. The follow-
ing is a list of equipment and practices usually used:
a. Multiple filter tubes of transparent material with a minimum diameter of 2 to 3 in.
are utilized.
9-3
-------
b. The tubes are fitted for either gravity or pressure operation.
c. A false bottom underdrain is utilized with either a porous plate or strainer backwash
system.
d. Flow control is established by the use of a positive displacement pump or an effluent
throttle valve connected to a float.
e. Pressure taps are provided above ard below the media, as well as at other locations
within the bed. Tap locations are generally located near the top of each type of media
used.
Additional details of pilot filters are given in various references on filtration studies (1), (2).
Two areas of process design cannot be as adequately explored in pilot studies as the others
listed above. These are the parameters associated with cleaning the bed and the effect of
chemical treatment of the feed. Cleaning of the filter bed is difficult to simulate in pilot
scale because of the small surface area of the beds utilized. The small area makes it im-
possible to study surface wash and air scour. Results of water backwash may not be repre-
sentative because of the wall effect. Chemical treatment of the filter feed without floccula-
tion and sedimentation may be simulated in the pilot studies. If flocculation and sedimenta-
tion are employed after chemical treatment however, the performance of model flocculators
and clarifiers may not closely simulate that of full-scale field units. Thus the feed to the
filters may not be representative. Both areas of difficulty can be overcome by using large
pilot installations, i.e., minimum filter size of about 2' x 2' to study backwash, and mini-
mum clarifier size of about 10' diameter to simulate full-scale sedimentation. However,
going to this scale is not usually economically justified. If studies are being conducted at an
existing plant, i.e., one that has full-scale clarifiers or filters, advantage should be taken of
the situation.
Although pilot studies give the most reliable information on which to design filters, it
should be remembered that filtration is a mixture of art and science. Excessive fine tuning
of the process design should not be attempted. The chosen process design should provide a
flexible system which will do the job with an adequate safety factor at a reasonable cost.
Fine tuning of the process should be left to the plant operator.
9.2.3 Filter Media
Selection of the size, type and depth of the filtration media is the single most important
decision in the design of a filtration system. Unfortunately, it is impossible to specify
optimum media characteristics from theoretical considerations. Pilot studies should be con-
ducted with several sizes and types of media prior to making a decision.
Sand has historically been the filtration medium most commonly used, but anthracite coal
and, to a lesser extent, garnet have been employed. These substances occur in nature and are
used in filters in a graded size range. Although use of a uniform size medium would have
certain advantages, it is not economical to excessively restrict the size range.
A typical grain size distribution curve for a naturally occurring filter medium is given in
Figure 9-3. This curve is often a straight line on log-probability paper. Historically, the two
points used to characterize a medium are the 10 percent size and the 60 percent size. These
9-4
-------
U.S. Standard Sieves
1" 3/4" 3/8
40
60
100 200
100
60
ll
u
£
>>
QQ
40
0.01
0.1
1.0
100
Grain Diameter in Millimeters
Figure 9-3 GRAIN SIZE CURVE
-------
are defined as the particle size of a distribution such that the weight of all smaller particles
constitutes the stated weight fraction of the whole. Usually the specification is given as the
effective size (10 percent size) and the uniformity coefficient (ratio of the 60 percent size to
the 10 percent size). In general, good correlation has been found between clean water
headloss and effective size (3). Fair and Geyer (3) present a method of calculating the size
fractions which must be sieved or washed out to convert from one size distribution to
another.
Until recently most filter designs called for a single filtration medium, either sand or coal.
Typical effective sizes range from 0.35 mm to 0.8 mm with uniformity coefficients between
1.3 and 1.7 (4). These size specification ranges can be utilized for preliminary selection of
sizes to be evaluated in pilot studies.
In coal and sand media of identical size, it has been found that the solids removal with coal is
somewhat inferior, but the rate of pressure drop build-up is lower (5). This finding is
explained by the greater angularity of coal with consequent greater porosity.
The relationship of medium size to filter performance can be generalized. Smaller media are
more effective in removing suspended solids at the expense of increased pressure drop or
headloss buildup rates, i.e., shorter filter runs. Larger media have reduced initial headlosses
and pressure drop increases during the filter run, but may yield a higher suspended solids
concentration in the effluent.
The purpose of pilot studies is to quantify the above experiences for the wastewater under
study. The quantitative relationships developed must then be used in conjunction with
economic and physical design factors to define the optimum design. The historic trend has
been toward coarser sizes of filter media in order to attain higher flow rates without
reducing the length of filter runs. In order to assure an adequate effluent quality with coarse
media and high filtration rates, chemical treatment of the feed may be required.
Another recent trend in filtration has been the adoption of the multi-media concept. Con-
ventional single medium filters have a fine to coarse gradation in the direction of flow which
results from hydraulic gradation during backwash. This type of gradation is not efficient as
virtually all of the removal and storage must take place in the upper few inches of the filter
with a consequent rapid increase in headloss. A coarse to fine filter gradation is much more
efficient as it provides for much greater utilization of bed depth, using the fine media only
to remove the finer fraction of the suspended solids.
One method of obtaining a coarse to fine filter gradation is the dual-media filtration con-
cept. This employs the use of a layer of coarse anthracite coal over a layer of fine sand. The
sizes of the anthracite and sand are chosen so that the coarser but lighter anthracite (specific
gravity 1.6) will remain above the heavier (specific gravity 2.65) but smaller sand during
backwash. It is desirable to have the coal as coarse as possible to prevent surface blinding
and the sand as fine as possible to promote high degrees of removals. However, the disparity
in sizes cannot be too great lest overtopping of the coal by the sand would result. In general,
sand sizes much finer than 40 mesh are not utilized because the coal size required to prevent
overtopping by sand during backwash would be too small to allow high filtration rates. To
ascertain the degree of mixing which will occur during backwashing and its effects on
subsequent filter performance, pilot column studies are best utilized.
9-6
-------
An extension of the dual media concept is the mixed-media filter. Ordinarily this is a
tri-media filter of coal over sand over garnet (specific gravity 4.2). There is evidence that
judicious selection of the size of each medium will allow a degree of intermixing of the
media, such that a reasonable approximation of continuous coarse to fine gradation is
obtained. Figure 9-4 illustrates the conventional single medium filter, a non-mixed dual
media filter and an ideal coarse-to-fine filter. Typical designs of multimedia filter beds are
given in Table 9-1.
Table 9-1
TYPICAL MULTI-MEDIA DESIGNS (6)
Garnet
Sand
Coal
Design
Size
Depth
Size
Depth
Size
Depth
No.
(Mesh)
(Inches)
(Mesh)
(Inches)
(Mesh)
(Inches)
1
-40x80
8
-20x40
12
-10x20
22
2
-20x40
3
-10x20
12
-10x16
15
3
-40x80
3
-20x40
9
-10x20
8
Conley and Hsiung (6) present additional information on the design of multi-media filters
for a variety of applications.
Several studies have compared the performance of single medium, dual media and multi-
media filters (7), (8), (9). These studies indicated that in general the latter two types of
filters outperformed single medium filters. Better effluents were obtained at higher flow
rates, with longer filter runs.
9.2.4 Filter Bed Depth
Although it is known that single medium filters make effective use of only the top few
inches of the bed, the historical practice of designing deep beds (24 to 36 in.) has not been
abandoned. Dual and multi-media filter depths of the same magnitude are justified because
the full filter depth is utilized. In fact, deeper filters of this type may find utility.
9.2.5 Flow Rates
Traditionally, the design flow rate of rqost single medium filters was 2 gal/min/ft . Recently
these have been raised to 4 gal/min/ft or greater when coarser media were employed along
with higher terminal pressure drops. The multi-media filters have been successfully operated
at rates up to 8 gal/min/ft on a continuous basis. Rates of this magnitude or higher can be
anticipated for design in the future. Of course, pilot studies are essential in determining the
design rates of a full-scale installation.
9-7
-------
Cross-Section Through
Single-Media Bed
Such as Conventional
Rapid Sand Filter
DOG
s
Cross-Section Through
Dual-Media Bed
Coarse Coal Above
Fine Sand
Cross-Section Through
Ideal Filter
Uniformly Graded From
Coarse to Fine
From Top to Bottom
i u
\cit
V
Oi
jQy.j
N .
•x. ^ >
r' > '"V"1
K «- » .-"s
Figure 9-4 MEDIA COMPARISONS
9-8
-------
9.2.6 Filter Size, Layout and Housing
Both gravity and pressure filter systems of standard sizes are provided by various filter
manufacturers. Specific size details can be obtained from manufacturers' catalogues. Gravity
filters are usually constructed on site of concrete. Common feedwater and backwash water
supply header pipes generally travel down a central pipe gallery (see Figure 9-1) between
rows of filter elements. The gravity filters are usually enclosed in a filter building with
storage of finished water in the basement area. All pumps, motors, controls, etc., are housed
in the filter building.
Pressure units are more individualized than gravity filters. They are fabricated of steel and
are cylindrical in shape. Structural design is according to standard pressure codes. Vessels up
to 10 feet in diameter and 60 feet in length are available. It is essential that a manhole be
incorporated in the design to allow for maintenance. The filter should be designed with a
means for hydraulic removal of all the filter media. Sight glasses for observation of the bed
should also be incorporated in the design.
9.2.7 Filter Cleaning Systems
Termination of a filter run takes place when either of these events occurs:
a. The effluent does not meet the quality criteria
b. The bed pressure drop is excessive
Either event indicates that the filter is excessively dirty and must be cleaned. The import-
ance of an effective cleaning system on the filter system performance has been reviewed in
several recent publications (10), (11). If effective cleaning is not obtained, short filter runs
and poor effluent quality will result. Indeed in many respects, improper cleaning sets up self
perpetuating operational difficulties such as mud balls, filter media slime coating and media
cracking (5).
Backwashing the filter bed by flow reversal is the major, and in many cases, the only
method used to clean the bed. The sources of backwash water may include feedwater, filter
effluent or some other effluent downstream of the filter. It is virtually impossible to predict
the specific conditions required to insure cleaning of the bed by this technique. Experience
indicates that for a single medium filter provision should be made to backwash the bed for a
10- to 15-minute period at a rate which will insure fluidization of all the media. The
optimum requirements can only be determined by the plant operator once the plant is in
service. The plant operator should be instructed that excessive expansion of the bed is just
as bad as insufficient expansion. Optimum expansion varies with the size of the medium,
type of floe and penetration of floe.
The most important aspect of the design of the backwash system is to insure uniform
distribution of the wash water over the entire cross section of the filter area. Although the
function of the underdrain system of the filter is both to collect the filtrate and distribute
wash water, the latter function controls the design.
A variety of underdrain systems are available for use. The traditional system employs several
layers of graded gravel under the filter bed with a lateral header system positioned on the
9-9
-------
filter floor. The lateral header system is equipped with orifices and provides the preliminary
distribution of the wash water. The final distribution is accomplished as the water moves
upward through the gravel. Culp & Culp (12) provide rules for the design of a lateral header
system.
Several commercial systems are available which employ patented false bottom distributors
in place of the lateral header system. The gravel is then placed on top of the false bottom.
Diagrams of Leopold Block and Wheeler Filter Bottom systems are illustrated in Figure 9-5.
Some systems have been devised with the concept of eliminating the gravel layers. One
system utilizes porous plates over a false bottom, as shown in Figure 9-5. Another system
also utilizes a false bottom with strainers on 12" centers. The strainers are nozzles fabricated
of plastic and metal with many small openings.
Table 9-2 presents the usual specifications for gravel, as well as a new design suggested by
Baylis (5).
Table 9-2
FILTER GRAVEL DESIGN
Standard
Baylis
Depth
Size
Depth
Size
2-1/2"
1/12"-l/8"
5"
1" - 2"
3-1/2"
l/8"-l/4"
2"
l/2"-l"
3-1/2"
l/4"-l/2"
2"
l/4"-l/2"
2"
l/2"-3/4"
4"
l/8"-l/4"
4"
3/4"-1-1/2"
2"
l/4"-l/2"
6"
l-l/2"-3-l/2"
2"
l/2"-l"
4"
1" - 2"
It was found that with the old design the upper layers of gravel could fluidize if excessive
backwash rates were used. The Baylis design, which places a final heavy layer of gravel over
the finer gravel, prevents this fluidization.
After moving upward through the expanded bed the wash water is conducted out of the
filter by wash water troughs. In a pressure filter the inlet baffle serves as the wash water
collector. In a gravity filter separate troughs are used because of the large area which must
be serviced. In order to prevent non-uniform flow of wash water, the maximum lip to lip
distance between troughs cannot exceed six feet. The troughs are usually rectangular chan-
nels. A convenient width is assumed and the depth computed by a momentum analysis of
the backwater curve in the trough. Figure 9-6 is a nomogram for this solution. For example,
9-10
-------
SECTION A-A
A Leopold-block underdrain, with glazed-tile blocks furnishing passages
to the water, instead of using laterial pipes.
3
:.trnri
t T
Xf-r.' I
£
i
¦¦.'iTy. .! >c/'><'/•' W X/r*
u
a
SECTION A-A
The Wheeler-filter bottom, consisting of solid, inverted, truncated pyra-
mids with water connections at the apex of each pyramid, and the
pockets filled with cement or glazed earthenware spheres.
"71"
<5
f
¦r_
T -H»
J
I
SECTION A-A
Porous-plate filter bottoms.
Figure 9-5 FILTER BOTTOM AND UNDERDRAIN SYSTEMS
-------
teiM
iSoe®
13000
l2ooo
| I OOA
|CrOOO
3 COO
0000
7 do*
(,04C
4o»o
Figure 9-6
NOMOGRAM FOR SELECTING WASH TROUGH DIMENSIONS
9-12
-------
3
an 18 inch wide trough carrying a flow of 3.33 ft /sec will have a maximum water depth of
13.2 inches. The side wall depth should be increased by 2 to 3 inches to allow for freeboard.
Maintaining a clean bed by backwash alone has been found to be difficult when wastewater,
even after biological and/or chemical treatment, is the filter feed. Even periodic chlorination
during backwashing has not always been successful in preventing slimes. Consequently,
either surface wash or air scour should be employed in waste treatment filters. Multi- and
dual-media filters allow for greater floe penetration of the bed than single medium filters.
The use of surface wash and/or air scour is mandatory under these circumstances. In
essence, the function of air scour and surface wash is to loosen the accumulated deposits in
the filter. The normal backwash then flushes the deposits away.
A surface wash apparatus is illustrated in Figure 9-7. Either fixed nozzles or rotating pipes
fitted with nozzles are placed about 1 to 2 inches above the top of the bed. While the
surface wash is on, the backwash expansion is set at a lower rate than after the surface wash
is terminated. Surface wash water is supplied at 50 to 100 psi at rates approximating 1 to 3
gal/min/ft of bed.
Air scour is accomplished by injecting air into the underdrain system prior to initiating
water backwash. The following procedure has been recommended (12):
a. Stop influent and lower the water level to a few inches above bed.
3 2
b. Apply air alone at 2-5 ft /min/ft for 3-10 min.
2
c. Apply water backwash at 2-5 gpm/ft with air on until water is within one foot of
wash water trough.
d. Shut air off.
e. Continue water backwash at normal rate for usual period of time.
f. Apply backwash for 1-2 min. at a rate required to insure hydraulic classification of
the filter media.
Air backwash has the disadvantage of increased possibility for coal media losses. Although
small losses repeatedly occur due to air bubble attachment during normal backwashing
cycles, the danger of massive losses exists if air is applied during trough overflow periods.
Control is especially difficult in pressure filters as it is hard to observe the bed during
backwash.
9.2.8 Filter Control Systems
Recent improvements and advances have been made in the systems which are used to
control the operation of filters. As a result, where once a filter required a manual operation,
it is now possible to have a completely automated filtration plant. Not only have these
advances removed much of the drudgery from filter operation, but they have given the plant
operator powerful techniques for improving filter performance.
9-13
-------
Figure 9-7
PALMER FILTER BED AGITATOR
9-14
-------
Perhaps the most important of the recent advances is the use of automatic turbidimeters to
continuously monitor the filter feed and product. This allows the operator to anticipate
difficulties from changes in feed quality, and rapidly remedy process failures. In addition,
these devices allow the operator to rapidly evaluate the effects of changes in process vari-
ables and provide a continuous record of plant performance. All turbidimeters operate on
the principle of measurement of scattered or transmitted light. A variety of commercial
instruments are available.
Automatic backwash, initiated either by terminal pressure drop across the bed or by excess
turbidity in the effluent, should be considered for all new plants. The cost of these systems
may be justified by a reduction in labor. Automatic backwash systems are available from a
variety of manufacturers. It is important that these systems be equipped with delay mech-
anisms so that the wash water reservoir capacity is not exceeded if the wash cycles of several
filters overlap.
The automatic backwash system referred to above is essentially an electronics package
which operates valves, pumps, etc., by remote control. Two other automatic backwash
systems based on a different concept are commercially available. An automatic gravity filter
is depicted in Figure 9-8. No data are available for waste treatment application of these
systems.
A traveling backwash filter employs an 11-inch depth of sand supported on porous plates.
The bed is divided into many sections 8 inches wide. A traveling backwash assembly moves
from section to section as required to clean each section. Thus most of the filter is in
operation with only a small section being washed at any time.
In the past, one type of flow control system was utilized in most filters. This system
provides for constant flow with a constant water depth over the filter. Under this condition,
constant flow is maintained by varying the headloss downstream of the filter so that the
total headloss in the filter and downstream is constant. Headloss in the downstream line is
automatically adjusted by a throttle valve connected to a preset counterweight. Usually a
venturi with a variable opening diaphragm serves as the rate controller. Frequently main-
tenance of these rate controllers is troublesome.
An alternate system which does not require a rate controller to achieve constant flow has
recently been described (13). A weir in the inlet channel to the clear well provides a
constant back pressure on the filter. Flow into the filter box is over an inlet weir with the
crest set quite high above the bed surface. As headloss develops in the filter, the depth of
flow above the filter increases to maintain the filter flow constant. A disadvantage of this
method is the capital cost associated with building the filter box walls several feet higher
than required where a rate controller is used.
Another control concept is declining-rate filtration. This method is applicable only to
medium or large scale plants which utilize multiple filters. With this method the flow rate
through the filter is allowed to decline as the filter clogs. The filters are staggered in degree
of clogging so that the totaJ production of the plant is constant. This procedure is claimed
to produce better effluent "quality and longer filter runs because flow slows as the filter
clogs. Thus, the rate of headloss increase decreases and the probability of floe breakthrough
is lessened.
9-15
-------
EFFLUENT
VO
o\
AIR
BACKWASH WATER
INFLUENT
ANTHRACITE
SAND
ROTO SCOUR
UNDERDRAIN
Figure 9-8 AUTOMATIC GRAVITY FILTER, SINGLE COMPARTMENT
(Courtesy of Graver)
-------
A simple method of flow control which can be used with pressure filters is to pump at a
constant rate to each filter with a positive displacement pump. With this system, the pump
rate sets the flow rate and the pump discharge pressure rises as the bed clogs.
9.2.9 Chemical Pretreatment
Chemical pretreatment ahead of the filter is designed to:
a. Coagulate suspended solids to make them more amenable to sedimentation and/or
filtration.
b. React with soluble components which must be removed to form insoluble precipitates
for removal by sedimentation and/or filtration.
c. Adjust the strength of the floe to control the degree of penetration of solids into the
filter.
In order to ascertain optimum chemical doses to accomplish any of these objectives, pilot
studies are required.
While coagulants aid in removal of colloids, the floe formed may be relatively weak. The
major use of polymeric materials is to strengthen floe so that it will not penetrate through
the filter. The floe should not be excessively strengthened or it will not penetrate beyond
the filter surface, producing excessive pressure drop. A good rule to follow is that floe
breakthrough should coincide with the achievement of terminal headloss.
Chemicals are usually applied prior to sedimentation to remove the bulk of the solids before
the filter. Recent practice has employed further chemical addition just prior to filtration to
enhance the filterability of the feedwater solids.
9.3 Summary of Results of Filtration Studies
Rapid sand type filters have been used in three kinds of waste treatment systems: direct
filtration of secondary effluent, filtration of secondary effluent after chemical treatment
and filtration of raw or primary wastewater after coagulation and sedimentation. Repre-
sentative results from field studies in each of these areas will be discussed below.
9.3.1 Filtration of Secondary Effluent
Early work on filtration of secondary effluent took place in Europe. Truesdale and Birkbeck
(14) reported on tests run between October, 1949, and May, 1950, at the Luton Sewage
Works. Beds of sand 2 feet deep, ranging in size from 0.9 mm to 1.7 mm, exhibited 72 to 91
percent removal of suspended solids and 52 to 70 percent removal of BOD. Flow rates
rangedfroml.33to3.31mp.gal/min/ft . Air-scour-aided backwash was used once per day to
clean the bed. Backwash flow rate was 11.6 Imp. gal/min/ft2.
Naylor, Evans and Dunscome (15) later reviewed 15 years of studies of tertiary treatment at
Luton. A 3-foot deep bed of-10 to +18 mesh sand consistently provided an effluent of 4 to
6mg/l suspendedso!idsatfiowratesof3.3Imp.gal/min/ft . It was found best to wash the beds
every 12 hours.
9-17
-------
In the U.S.,most direct filtration work has been with activated sludge feed. At the Hyperion
Plant in Los Angeles, sand of 0.95 mm effective size was used in a shallow bed (11 inches
deep) traveling backwash filter. This study lasted for six months during which time 46
percent suspended soHds removal and 57 percent BOD removal were obtained. Filtration
rate was 2 gal/min/ft . Difficulty was encountered in cleaning the filters and performance
gradually deteriorated during the study. Use of a finer sand (0.45 mm effective size) in an
attempt to yield a better effluent was a failure due to very rapid clogging of the filter (16).
Much greater success utilizing the traveling backwash filter for activated sludge effluent
treatment was obtained by Lynam (17) in Chicago. The effective size of sand used in this
study was 0.58 mm. Suspended solids removal of 702percent and BOD removal of 80
percent were obtained at flow rates of 2 to 6 gal/min/ft . Terminal headloss was quite low
(11 inches of water.) The range of flows studied exhibited no significant difference in terms
of suspended solids removal.
A study of filtration of activated sludge effluent was presented by Tchobanoglous and
Eliassen (18). They found that flow rate (2 to 10 gal/min/ft ) had little effect on perform-
ance, but effective size of sand in the range of 0.4 mm to 1.2 mm had a very significant
effect. It was also determined that virtually all suspended solids removal took place in the
upper six inches of the bed. It was concluded that the floe strength of activated sludge is
quite high on the basis of the low degree of penetration even with very coarse media and
high flow rates. Poor removal of suspended solids (10 to 40 percent) was obtained in this
study, with filter depths greater than 6 inches producing no additional benefits. It was
found that a bimodal distribution of particle sizes existed for the activated sludge effluent.
Apparently, only the larger solids could be removed by the filters.
Culp and Culp (12) reviewed the work on plain filtration of secondary effluent with both
single medium and multimedia filters. They concluded that, with either type of filter, better
results would be obtained as the degree of self flocculation of the sludge increased. Thus, a
high-rate activated sludge effluent which contains much colloidal material should filter
poorly, while an extended aeration effluent should filter well. Multi-media filters exhibit a
marked superiority for filtration of activated sludge effluent because of the high volume of
floe storage available in the upper bed and the polishing effect of the small media. They
indicated the expected performance of multi-media filters for plain filtration of secondary
effluents, as shown in Table 9-3.
Table 9-3
EXPECTED EFFLUENT SUSPENDED SOLIDS FROM MULTI-MEDIA
FILTRATION OF SECONDARY EFFLUENT
Effluent Type
Effluent S.S. mg/1
High Rate Trickling Filter
10-20
2-Stage Trickling Filter
6-15
Contact Stabilization
6-15
Conventional Activated Sludge
3-10
Extended Aeration
1 - 5
9-18
-------
9.3.2 Filtration of Chemically-Treated Secondary Effluent
Treatment of secondary effluent by chemical coagulation, sedimentation and filtration has
been conducted at a number of installations. The purposes of this treatment procedure have
often been twofold, suspended solids and phosphorus removal. Unfortunately, in many
cases, performance results have been reported for complete systems, not each individual
process. Thus, analysis of the filter performance is not possible. In most of these installa-
tions, the filter is viewed essentially as a polishing device to capture solids which escape the
sedimentation tank. As Culp and Culp (12) have indicated, this philosophy may be wrong,
as modern filters can absorb much greater solids loads than older designs.
In conjunction with the studies at Chicago (17), coagulation with alum followed by filtra-
tion was evaluated. It was found that the alum treatment had little effect on the effluent
quality. It is probable that this result is due to the weakness of chemical floe compared to
activated sludge floe.
An advanced waste treatment plant has been used to renovate step aeration activated sludge
effluent in Nassau County, New York. Alum at 200 mg/1 is used to coagulate the waste-
water prior to sedimentation and filtration. The filters are dual-media containing 30 inches
of 0.9 mm coal over 6 inches of 0.35 mm sand. With the addition of 0.5 mg/1 of an anionic
polymer, effluent turbidity is maintained below 0.4 JTU. Run length varies from 8 to 24
hours depending on the solids load from the clarifier.
At Lebanon, Ohio, treatment of the activated sludge effluent with lime has been investi-
gated (19). The lime dose averaged 300 mg/1. Dual-media filters, consisting of 18 inches of
0.75 mm coal over 6 inches of 0.45 mm sand, followed clarification. The filters were
operated at 2 gal/min/ft , and were backwashed when the leadloss reached 9 ft of water.
Influent to the filters ranged from 13 to 36 mg/1 of suspended solids (turbidity 4 to 10
JTU.) Filter effluent ranged from 0.07 to 0.14 JTU.
At Lake Tahoe,lime is used to coagulate activated sludge effluent prior to sedimentation and
filtration p2). Multi-media beds 3 feet deep are utilized at an average flow rate of 5
gal/min/ft . Filter runs have varied from 4 to 60 hours. Polymers as well as alum have been
used to strengthen the floe. Normally, runs are terminated at headlosses of 8 ft of water.
Effluent turbidities are typically reduced to 0.3 JTU with correspondingly low values for
other parameters.
At the Environmental Protection Agency pilot plant at Washington, D.C., mineral addition
to the final phase of a step aeration activated sludge is practiced for phosphorus removal.
After sedimentation the effluent is filtered through parallel filters at 2.4 gal/min/ft . One
filter is a dual-media, while the other is a multi-media. It has been found that the multi-
media filter removes 5 to 10 percent more suspended solids than the dual-media filter. Filter
runs have been in the range of 24 to 32 hours. Typically, the filters reduce secondary
effluent suspended solids from 33 mg/1 to 8 mg/1 (20).
9.3.3 Filtration Following Chemical Treatment of Primary or Raw Wastewater
Clarification of raw wastewater followed by carbon adsorption is just emerging as a viable
treatment technology. This system employs filtration as part of the solids separation system.
Although several pilot installations employing this concept are in operation, filtration has
not been closely studied; thus data are sparse.
9-19
-------
At the EPA pilot plant in Washington, D.C., two-stage lime treatment of raw sewage fol-
lowed by sedimentation, filtration and granular carbon adsorption is being studied (21).
Dual-media filters (18 inches of 0.9 mm coal over 6 inches of 0.45 mm sand) are employed.
Cleaning is initiated at a headloss of 9 feet of water. Cleaning is performed automatically
with a surface wash rate of 3 gal/min/ft and an upflow rate of 20 gal/min/ft . Run lengths
averaged 50 hours during cold weather, but in warm weather the growth of slimes reduced
run lengths to less than 1 2 hours. Prechlorination of the filter feed was employed to restore
the run lengths to 50 hours. Filter effluent averaged 4.5 mg/1 of suspended solids over a
6-month period, which represented a 70 percent efficiency for the filter. This plant operates
on a programmed diurnal flow variation which produces a flow rate variation on the filter
from 1.7 to 4.3 gal/min/ft .
A similar system using single-stage lime treatment was run for several months at the EPA
installation in Lebanon, Ohio (22). Flow to the filters was 2 gal/min/ft . Suspended solids
in the filter effluent averaged 10 mg/1, which represented a 67 percent removal efficiency.
9.4 New Filtration Systems
The rapid sand type filtration system previously discussed has been a downflow, batch,
static bed system. Consequently, it suffers from a variety of process deficiencies including:
a. The need to stop the process periodically to clean the filter medium.
b. The limited ability to economically handle suspensions containing high concentra-
tions of suspended solids.
During the last decade a number of new filtration systems have been developed which are
aimed at overcoming these shortcomings. Several of these will be discussed below.
9.4.1 Upflow Filtration
As indicated previously, a major difficulty with downflow, single-media filtration is that
after backwash, the bed is graded fine to coarse in the direction of flow. If filtration is
conducted upflow, this difficulty is circumvented. However, once the headloss produced by
the upflow exceeds the buoyant weight of the filtration medium, fluidization with con-
sequent loss of filtration efficiency will result. One solution to this problem is to place a
restraining grid on or near the top of the filter medium to prevent fluidization. A diagram of
an upflow filter is illustrated in Figure 9-9 (courtesy of DeLaval.)
The spacing between bars of the grid must be large enough to allow the bed to expand
during backwash, but must be small enough to prevent upward bed movement during
filtration. It would seem that these two requirements are directly contradictory; however,
arching of the grains takes place between the bars, allowing a reasonably large spacing. Space
between the bars is usually in the range of 100 to 150 diameters of the smallest grain size in
the beds. During cleaning, air is first introduced to agitate the bed. After the air has broken
the arches, backwashing with water is started. Table 9-4 gives a summary of typical design
parameters for the upflow filters.
9-20
-------
COVER OPTIONAL
(FOR CLOSED SYSTEM)
"GRID
DEEP SAND LAYER
GRAVEL LAYERS
WASH WATER
FILTRATE OUTLET
SAND "ARCHES"
SPECIAL VENT
INLET RAW WATER
AIR FOR
SANDFLUSH CLEANING
Figure 9-9 CROSS SECTION OF UPFLOW FILTER
9-21
-------
Table 9-4
DESIGN PARAMETERS FOR UPFLOW FILTERS
Bed Material:
Sand
Bed Construction:
60 in. 1-2 mm
10 in. 2-3 mm
4 in. 10-15 mm
Flow Rate:
2-3 gal/min/sq ft
Backwash Rate:
To achieve minimum 20% expansion
Terminal Headloss:
6 to 20 feet of water
Boby and Alpe (23) reported on the performance of an upflow filter treating secondary
effluent at Totor, England. Average suspended solids removal was 85 percent, with the filter
effluent below 5 mg/1. These results were equal to or better than those obtained with
downflow filtration with the same size bed. It is claimed that this type of filter can absorb
higher loads of solids than a conventional filter.
9.4.2 Moving Bed Filter
The basic concept of a moving bed filter is the mechanical movement of the most heavily
clogged portion of the medium out of the zone of filtration with virtually no interruption of
the filtration process. The potential of such a process is operation at higher flow rates and at
much higher solids loadings than conventional systems. Superior cleaning of the filter media
should also be possible.
Johns-Manville Corporation has developed a moving bed filtration system. A diagram
illustrating the essentials of this system is given in Figure 9-10. Wastewater (A) flows
through the inlet pipe where chemicals, if required, are added at (B). The wastewater enters
the head tank (C) and then passes through the sand bed (D). The filtered water leaves
through the exit screens. When excessive headloss develops, the bed is pushed toward the
head tank by pressurizing a chamber separated from the bed by a flexible diaphragm. A
mechanical cutter (F) sweeps down over the face of the bed cutting off the top layers. These
then fall into the hopper (G) of the head tank. The sludge and sand are removed from the
head tank with the aid of an ejector using feedwater. The solids are hydraulically conveyed
to the sand washer (H) where filtered water or air and filtered water are used to backwash
the sand. Clean sand moves by gravity back to the base of the filter. The spent washwater is
sent to a sedimentation tank for removal of the wastewater solids. The operation of the
system is automated.
Under an Environmental Protection Agency contract (24), this system was evaluated by the
manufacturer for the treatment of raw wastewater, primary effluent and settled and
unsettled trickling filter effluents at the Bernards Township Sewage Treatment Plant. Alum
was used to precipitate phosphorus, and an anionic polymer was employed to prevent
excessive Hoc penetration. The results, given in Table 9-5, show excellent treatment
performance in these situations.
9-22
-------
WASH
WATER +
RECLAIM
WASTE
INFLUENT
1
i
N>
U)
SAND _
^AIA* washer
^ FILTERED WATER
FOR SAND WASHING
v— SAND
DRIVE
SYSTEM
B)
CHEMICALS
' F> CUTTER\
EXIT
SCREENS
/^\
FILTERED WATER
FOR SYSTEM REUSE
©
SAND ~ SOLIDS
TANK
SAND
FILTERED
WATER TO
DISCHARGE
Figure 9-10
SCHEMATIC DRAWING OF THE JOHNS-MANVILLE MOVING BED FILTER
-------
The moving bed filter is being evaluated at full scale (2 MGD) at the Borough of Manville
Sewage Treatment Plant, Manville, N.J. At this site, unsettled trickling filter effluent is the
feed. This study is being partially funded by an EPA demonstration grant.
Table 9-5
JOHNS-MANVILLE MOVING BED FILTER EVALUATION AT
BERNARDS TOWNSHIP SEWERAGE AUTHORITY TREATMENT PLANT
Unsettled
Trickling
Final Effluent
Filter
Primary
Raw
Parameter
w/o C
llorination
Effluent
Effluent
Wastewater
(mg/1)
In
Out
%
In
Out
%
In
Out
%
In
Out
%
Total P
9.37
0.51
95
19.1
0.99
95
14.6
1.13
93
21.5
2.16
91
Filterable P
8.03
0.1 1
99
14.9
0.62
96
13.2
0.58
96
18.6
0.79
96
Ortho P
7.80
0.10
99
12.4
0.53
96
9.8
0.38
96
13.2
0.57
95
BODs
65
12
80
55
3.8
93
67
12
82
115
19
84
Suspended
Solids
50
15
70
86
7.1
91
77
11
87
156
27
83
Turbidity
(JTU)
33
7
79
39
3.4
91
53
3.7
93
123
16.7
87
Alum: 200 mg/1 (commercial grade)
Polyelectrolyte: 0.5 mg/1 anionic
The moving bed filter system is currently available in modules of 2-bed and 4-bed configura-
tions. These modules can be used singly or in combination to accommodate large flow
requirements as necessary.
a. The flow rate through any unit or combination of units is dependent on the quality
of the incoming liquid and the discharge requirements. Flow rates up to 7.0 gal/
min/ft of exposed filter area are possible.
b. Filter bed dimensions - 48-inch face diameter, 60-inch length from the face of the
unit to the center line of exit screen.
9-24
-------
c. Module dimensions -
2-Bed Unit 4-Bed Unit
Plan length 20' 20'
Plan width 7' 14'
Height 17' 17'
Weight, empty 22,000 lb 44,000 lb
operating 78,0001b 156,0001b
Plan floor space,
module 140 ft2 280 ft2
system 350 ft 720 ft
d. Power requirements -
Two-Bed System Four-Bed System
Function Conn. HP Op. HP Conn. HP Op. HP
Diaphragm or
Sand-bed
Movement
3.0
3.0
6.0
6.0
Cutter
3.0
0.75
3.0
1.5
Sand Cleaning
System
3.75
3.75
7.5
7.5
Screen Wash
7.5
0.5
7.5
1.0
Chemical Pump
and Mixer
1.0
0.5
1.0
0.75
Air Compressor
1.0
0.5
1.0
0.75
19.25
9.0
26.0
17.5
e. Filter media - Hard sharp-grained quartz sand of filter grade quality (effective size 0.6
to 0.8 mm; uniformity coefficient 1.5)
f. Sand drive rate -
Linear - 12 in./jiour - maximum
Volume-25 ft /hour - maximum
Pump pressure - 75 psi - average
150 psi - maximum
g. Head tank - 30 minute detention time
h. Sludge settling tank - 30 minute detention time
i. Controls -
Low level and high level off-on
Differential pressure to actuate bed movement and cutter
Effluent turbidity to adjust chemical feed
9-25
-------
9.4.3 Radial Flow Moving Bed Filter
Recently,Dravo Corporation has introduced a radial flow moving bed filter to the American
market. At present, most applications of this system have been for industrial waste treat-
ment. In addition to the moving bed concept,the mam feature of this system is the radial
flow concept. This geometric configuration provides more filter area per unit volume than
downflow or upflow systems. As the liquid flows radially from the central core, it slows
down, providing increased opportunity for solids removal.
9.4.4 Radial Flow-External Wash-Filter
The Hydromation Corporation has developed a new concept in filtration. A diagram of this
filter is illustrated in Figure 9-11. It is a batch-type radial flow filter with external media
wash. When a filter run is terminated, the media is pumped out of the filter and upward into
a scrubber. The flow velocity in the scrubber is 20 ft/sec, which produces a very high degree
of turbulence, assuring good cleaning of the media. The clean media circulates around the
backwash loop to the radial flow bed.
This filtration system utilizes a polymer resin as the filter medium. It is claimed to have
superior dirt holding capacity compared to natural media. Because of the superior dirt
holding character and special cleaning system, flow rates of 10 to 20 gal/min/ft have been
purported.
9-26
-------
Polymer Media
Inlet
Conductor
mMr M
Outlet Septum
Figure 9-11 HYDROMATION IN-DEPTH FILTER
9-27
-------
9.5 References
1. Kreissl, J.F., & Robeck, G.G., "Multi-Media Filtration: Principles and Pilot Experi-
ments". Bulletin No. 57, School of Engineering and Architecture, University of Kansas,
Lawrence, Kansas (1967).
2. Summary Report, Advanced Waste Treatment. WP-20-AWTR-19, U.S. Dept. of the
Interior, FWPCA (1968).
3. Fair, G., Geyer, J., "Water Supply and Waste Water Disposal". Chapter 24, John Wiley
& Sons, Inc., New York (1954).
4. Water Treatment Plant Design, American Water Works Association, Inc., New York
(1969).
5. Water Quality and Treatment, American Water Works Association, Inc., McGraw-Hill,
Inc., New York (1971).
6. Conley, W.R., and Hsiung, K., "Design and Application of Multimedia Filters". Jour.
AWWA, 61, 97 (Feb. 1969).
7. Laughlin, J.E., and Duvall, T.E., "Simultaneous Plant-Scale Tests of Mixed-Media and
Rapid Sand Filters". Jour. AWWA, 60, 1015 (Sept. 1968).
8. Westerhoff, G.P., "Experience with Higher Filtration Rates". Jour. AWWA, 63, 376
(June 1971).
9. Miller, D.G., "Rapid Filtration Following Coagulation Including the Use of Multi-Layer
Beds". Proc. The Society for Water Treatment and Examination, 16, 3, 197 (1967).
10. Hirsch, A.A., "Backwash Investigation of a Proposed Simple Uniformity Control".
Jour. AWWA, 60, 570 (May 1968).
11. Johnson, R.L., and Cleasby, J.L., "Effect of Backwash on Filter Effluent Quality".
Jour. San. Eng. Div., ASCE, 92, 215 (Feb. 1966).
12. Culp, R.L., and Culp, G.L., Advanced Wastewater Treatment, Van Nostrand-Reinhold
Co., New York (1971).
13. Baumann, E.R., and Oulman, C.S., "Sand and Diatomite Filtration Practice". Water
Quality Improvement by Physical and Chemical Processes, University of Texas Press,
Austin, Texas (1970).
14. Truesdale, G.A., and Birkbeck, A.E., "Tertiary Treatment Processes for Sewage Works
Effluents". Water Poll. Control Jour. (Brit.) 66, 371(1967).
15. Naylor, A.E., Evans, S.C., and Dunscombe, K.M., "Recent Developments on the Rapid
Sand Filters at Luton". Water Poll. Control Jour. (Brit.) 66, 309 (1967).
16. Laverty, F.B., Stone, R., and Meyerson, L.A., "Reclaiming Hyperion Effluent". Jour.
San. Eng. Div., ASCE, 87, 6, I (Nov. 1961).
9-29
-------
9.5 References (Cont)
17. Lynam, B., Ettelt, G., and McAloon, T.J., "Tertiary Treatment at Metro Chicago by
Means of Rapid Sand Filters and Microstrainers". Jour. WPCF, 41, 247 (Feb. 1969).
18. Tchobanoglous, G., and Eliassen, R., "The Filtration of Treated Sewage Effluent".
Proceedings of the 24th Purdue Industrial Waste Treatment Conference, 1323 (1969).
19. Berg, E.L., Brunner, C.A., and Williams, R.T., "Single Stage Lime Clarification of
Secondary Effluent". Water & Wastes Engineering, 7, 3, 43 (March 1970).
20. Hais, A.B., Stamberg, J.B-, and Bishop, D.F., "Alum Addition to Activated Sludge with
Tertiary Solids Removal". Presented before AIChE National Meeting, Houston, Texas
(March 1971).
21. Bishop, D.F., O'Farrell, T.P., and Stamberg, J.B., "Physical-Chemical Treatment of
Municipal Wastewater". Presented before the 43rd Annual Meeting, WPCF, Bston,
Mass. (Oct. 1970).
22. ViUiers, R.V., Berg, E.L., Brunner, C.A., and Masse, A.N., "Treatment of Municipal
Wastewater by Lime Clarification and Granular Carbon". Presented before ACS,
Toronto, Canada (May 1970).
23. Boby, W., and Alpe, G., "Practical Experiences Using Upward Flow Filtration". Proc.
Society for Water Treatment and Examination, 16, 3, 215 (1967).
24. Phosphorus Removal Using Chemical Coagulation and a Continuous Countercurrent
Filtration Process, Final Report (17010 EDO), U.S. Dept. of the Interior, FWQA (June
1970).
9-30
4 U.S. GOVERNMENT PRINTING OFFICE: 1672 —
759-398/148
-------
DISINFECTION
------- |