EPA-670/2-75-050g
June 1975
Environmental Protection Technology Series
-------
EPA-670/2-75-050g
June 1975
DIRECT FILTRATION OF LAKE SUPERIOR
WATER FOR ASBESTIFORM FIBER REMOVAL
Appendix I
Diatomite Filters for Asbestiform Fiber Removal from Water
By
E. Robert Baumann
for
Black § Veatch, Consulting Engineers
Kansas City, Missouri 64114
Program Element No. 1CB047
Contract No. DACW 37-74-C-0079
Interagency Agreement EPA-IAG-D4-0388
USEPA, Region V and Corps of Engineers, St. Paul
Project Officer
Gary S. Logsdon
Water Supply Research Laboratory
National Environmental Research Center
Cincinnati, Ohio 45268
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
REVIEW NOTICE
The National Environmental Research Center, Cincinnati, has reviewed
this report and approved its publication. Approval does not signify
that the contents necessarily reflect the views and policies of the
U.S. Environmental Protection Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
11
-------
FOREWORD
Man and his environment must be protected from the adverse effects
of pesticides, radiation, noise and other forms of pollution, and the
unwise management of solid waste. Efforts to protect the environment
require a focus that recognizes the interplay between the components of
our physical environment -- air, water, and land. The National
Environmental Research Centers provide this multidisciplinary focus
through programs engaged in
• studies on the effects of environmental contaminants
on man and the biosphere, and
• a search for ways to prevent contamination and to
recycle valuable resources.
This report and its appendices present the results of pilot plant
filtration research for the removal of asbestiform fibers from drinking
water. The several appendices present detailed information on water
quality, pilot plant equipment and operation, individual filter run data,
asbestiform fiber and amphibole mass concentrations in raw and filtered
water, and diatomite filter optimization. Appendix I contains the diatomite
filter optimization study.
A. W. Breidenbach, Ph.D.
Director
National Environmental
Research Center, Cincinnati
111
-------
ABSTRACT
Pilot plant research conducted in 1974 at Duluth, Minnesota, demonstrated
that asbestiform fiber counts in Lake Superior water could be effectively
reduced by municipal filtration plants. During the study, engineering
data were also obtained for making cost estimates for construction and
operation of both granular and diatomaceous earth (DE) filtration plants
ranging in size from 0.03to 30 mgd.
During one phase of the pilot plant investigation, the diatomite
filters were operated in a way that yielded data used for computer
optimization of the DE filtration process, the POPO (Program for
Optimization of Plant Operation) results are presented in Appendix I.
IV
-------
ER! - 75082
PREPRINT
Project 651
Engineering Research Institute
IOWA STATE UNIVERSITY
AMES
DIATOMITE FILTERS FOR ASBESTIFORM FIBER REMOVAL FROM WATER
E. Robert Baumann
Anson Marston Distinguished Professor of Engineering
Acting Director, Engineering Research Institute
April 1, 1975
This paper is a summary of the results from vacuum and pressure
diatomite filter pilot plants operated by Black and Veatch, Con-
sulting Engineers of Kansas City, Missouri, for the removal of
asbestiform fibers from Lake Superior water at Duluth, Minnesota.
The work was conducted under contract with the U.S. Environmental
Protection Agency in cooperation with the U.S. Army Corps of En-
gineers and the City of Duluth. The author served as a consul-
tant to the engineers in planning and interpreting the pilot-
plant results.
A preprint of a paper to be published in the Journal American Water
Wo^ks Association, reproduced here by permission of the Association,
6666 West Quincy Avenue, Denver, Colorado 80235.
-------
TABLE OF CONTENTS
Page
Introduction 7
Quality of Raw Water 9
Potential Treatments with Diatomite Filters 15
Pilot Plant Results 25
Determination of Filtration Resistance 30
Determination of Optimum Plant Designs 36
Final Cost Comparisons 61
Conclusions and Recommendations 68
Acknowledgments 75
References 76
-------
LIST OF TABLES
Table 1. City of Duluth Water Supply, Lakewood Pumping
Station, raw water turbidity data, 1952-1972.
Table 2, Percentage of time given raw water turbidity
is equalled or exceeded, City of Duluth,
Lakewood Pumping Station, 1952-1972.
Table 3. Filter aid ratings (approximate).
Table 4. Water and/or filter media treatments selected
for optimization runs.
Table 5. Suspended solids removal in selected series of
filter runs.
Table 6. Input data, POPO (Program for Optimization of
Plant Operation), asbestos-turbidity removal,
pressure diatomite filter plant.
Table 7. Input data, POPO (Program for Optimization of
Plant Operation), asbestos-turbidity removal,
Lake Superior water, vacuum diatomite filter
plant.
Table 8. Capital construction costs of pressure diatoma-
ceous earth or diatomite water filtration plants
with varying filter rates, plant design for
30 mgd.
Table 9. Capital construction costs of vacuum diatoma-
ceous earth or diatomite water filtration
plants with varying filter rates, plant design
for 30 mgd.
Table 10. POPO input data, filter aid characteristics
(approximate values).
Table 11. Data input to POPO, pressure diatomite filter,
Case 5, raw water turbidity of 1.9 FTU.
Table 12. Computer output for data input shown in Table 11
Table 13. POPO output, pressure filter, all cases.
Table 14. POPO output, vacuum filter, all cases.
Page
12
14
16
26
29
38
39
41
41
43
46
47
49
50
-------
Table 15. Comparison of pressure and vacuum dlatomite
filtration costs optimized to remove 1.0
FTU turbidity.
Table 16. Data input to POPO, pressure diatomite filter,
Case 5, turbidity of 1.0 FTU when plant is
designed for 1.9 FTU.
Table 17. Final optimization of Case 5 pressure diatomite
filters operated at full range of water turbid-
ity to be encountered.
Table 18. Total annual pressure diatomite filtration
costs, Case 5, when plants designs are based
on raw water turbidities of 1.0, 1.9, and
5.8 FTU.
Table 19. Effect of 20 percent increase in filter aid
cost on annual unit filtration cost of three
plant designs.
Table 20. Revised cost data, total capital construction
costs for pressure (De Laval) diatomite fil-
tration plants with various filtration rates,
plant capacity of 30 mgd.
Table 21. Unit filter and labor costs for De Laval pres-
sure filters and the estimated plant labor re-
quirements .
Table 22. POPO optimum design of pressure filters (1-in.
diameter septa), raw water turbidity of 2.5 FTU.
Table 23. Average cost per thousand gallons for pressure
filtration of 2.5 FTU turbidity water, Cases
1-5, 50-year plant life.
Table 24. Amount of water filtered at each turbidity
level as a function of daily plant produc-
tion.
Table 25. POPO optimized plant operation for different
mean levels of turbidity expected in Lake
Superior during a year.
page
50
55
56
58
60
64
65
66
67
67
69
-------
page
Table 26. Approximate total annual pressure diatomite 70
filtration costs, Case 5, when plant is de-
signed for producing 30 mgd using raw water
containing 2.5 FTU of turbidity and operated
at 30 mgd.
Table 27. Approximate total annual pressure diatomite 71
filtration costs, Case 5, when plant is de-
signed for producing 30 mgd using raw water
containing 2.5 FTU of turbidity and operated
at 20 mgd.
-------
LIST OF FIGURES
page
Fig. 1. City of Duluth, Minnesota, Lakewood Pumping 13
Station, Lake Superior raw water turbidity,
1952-1972.
Fig. 2. Surface charges on Chrysotile fibers. 19
Fig. 3. Zeta potential, £, and filter cake resistance, 22
§, as a function of coating level for Alum
coated on Celite 536 (6).
Fig. 4. Zeta potential, £, and filter cake resistance, 23
§, as a function of coating level for C-31
coated on Celite 545 (6).
Fig. 5. Turbidity and pressure drop as a function of 32
run length, Run 79, Case 5.
Fig. 6. Development of a Beta prediction equation for 33
Case 5.
Fig. 7. Beta prediction relationships for Cases 1-4. 35
Fig. 8. Sections through typical pressure and vacuum 51
diatomite filter units.
Fig. 9. Optimum design characteristics of pressure 53
diatomite filters (Table 12), Case 5.
Fig. 10. Body feed required for optimum plant operation, 61
Case 5 plant based on raw water turbidity of
1.9 FTU.
Fig. 11. Body feed levels required to provide a given 61
run length, Case 5 plant optimized for handling
1.9 FTU turbidity water.
-------
DIATOMITE FILTERS FOR ASBESTIFORM FIBER
REMOVAL FROM WATER
E. Robert Baumann,
Anson Marston Distinguished Professor of Engineering
Engineering Research Institute, Acting Director,
Iowa State University,
Ames, Iowa
Introduction
During the period from May through September, 1974, two diatomite
filter pilot plants (10 gpm) were operated at Duluth Minnesota, for
the removal of suspended solids, including asbestiform materials, from
Lake Superior water. A total of 86 runs was made using a pressure
diatomite filter (ERD-2), and a total of 122 runs was made using a
vacuum diatomite filter (BIF-2). All runs were made with filtration
rates between about 0.8 and 1.6 gpm/sq ft of filter area. The raw
water turbidity varied from 0.5 to about 1.2 FTU, and the water
temperature was between 33 °F and 48 °F.
Data of the following type were collected during each filter run:
1. Filtration rate (constant),
2. Head loss every half-hour,
3. Water temperature,
4. Precoat and body feed types and amounts and filter media
coating types and amounts used,
5. Raw and filtered water turbidity every half-hour.
Raw and filtered water samples were taken during about every fifth
run for detailed analysis for asbestiform fibers.
-------
During the series of filter runs, the variables which could be
controlled included:
1. Filtration rate,
2. Grade and amount of precoat filter aid used,
3. Grade and amount of body feed filter aid used,
4. Coating type and amount applied to the body feed and/or
precoat filter aid used,
5. Type and amount of polymer used as a coagulant during the
filter run.
The test series were conducted in three phases:
1. A preliminary test series conducted to train the operators,
develop test and data recording procedures, and gain plant
operating experience,
2. An "asbestiform fiber removal" efficiency series of runs to
determine which treatments gave promise of effectively re-
moving asbestiform fibers from suspension (This series was
designed to narrow down the treatment possibilities.)>
3. An "optimum-design" series of runs to determine filter cake
resistance measurements using treatment techniques found
most successful in the phase two runs. Cake resistance
measurements are necessary to optimize the design of dia-
tomite filters to provide least-cost filtration of the Lake
Superior water. These runs were also used to evaluate the
expected suspended solids (asbestiform fiber) removal to be
accomplished by diatomite filtration.
-------
This paper provides a brief summary of the results of the signi-
ficant runs on which the optimum design of a diatomite filtration plant
for removal of asbestiform fibers are based. The paper will consider
the following items:
1. The quality of the raw water,
2. The potential treatments that might be expected to produce
the desired quality of filtered water,
3. The results achieved,
4. The determination of filtration resistance,
5. The determination of optimum plant designs,
6. Conclusions and recommendations.
Quality of Raw Water
The design of a filter plant must be based both on the quantity of
the filtered water which must be produced and on the quality of the
raw water that must be filtered. In optimizing filter plant operating
costs, it is not necessary to design a diatomite filter plant to oper-
ate under the worst raw water conditions that will be encountered. If
it were, the plant might operate under optimum conditions during only
a few days per year. On all other days, it would be overdesigned. It
is more satisfactory to design a filtration plant for optimum operation
using a raw water quality that will be equalled or exceeded, say 90
or 95 per cent of the time. The economics of doing this will be ex-
plained in more detail later in this paper.
-------
The main purpose of filtering Lake Superior water, for example,
for use as a public water supply would be for the reduction of suspend-
ed solids, principally the suspended solids in the form of asbestos
fibers. Unfortunately, it is not customary to routinely measure sus-
pended solids levels in water supply sources, and interest in asbestos
fibers in Lake Superior waters has been generated only recently (1973) .
Thus, adequate data on levels of suspended solids and asbestos fibers
in Lake Superior water are not available. Such data are necessary for
predicting the quality of raw water which will have to be filtered.
The turbidity of water, however, is related to the level of suspended
solids present and is a reliable indicator of the presence of sus-
pended solids.
Until 1962 the U.S. Public Health Service drinking water standards
ik-
limited the turbidity of drinking water to a value less than 10 JTU .
In 1962 a new drinking water standard was issued in which the allow-
able drinking water turbidity was reduced to 5 JTU [1]. With the re-
cent passage of the "Safe Drinking Water Act" by the U.S. Congress in
December, 1974, it is expected that the Environmental Protection Agency
will again revise the drinking water turbidity standard by requiring a
drinking water turbidity less than 1.0 JTU. In 1968 the American Water
Works Association adopted a turbidity "goal" (which water works should
seek to achieve) of only 0.1 JTU [2]. Fortunately, the Duluth water
works has collected Lake Superior raw water turbidity data at the Lake-
wood Pumping Station for a number of years. These records can be used
*
Jackson Turbidity Units. In this study, a formazin turbidity standard
was used and thus the turbidity results are expressed in Formazin
Turbidity Units (FTU).
10
-------
to evaluate the quality of water which a filtration plant must be de-
signed to handle.
Table 1 summarizes the raw water turbidity of Lake Superior
water measured at the Lakewood Pumping Station from 1952-1972. Each
year, the turbidity of the Lake water is recorded during those hours
when the turbidity exceeds 1 JTU. The number of hours during which
the turbidity is in the range 1-5, 6-10, 11-15, 16-20, etc. are record-
ed by year. There were only 4 hours in 21 years that the turbidity
was as high as the 76-100 range, and a total of only 124 hours in 21
years that it exceeded a turbidity of 20 JTU. During the 21 years the
water turbidity was below 1.0 JTU 89.95 per cent of the time.
Figure 1 is a plot of Lake Superior water turbidity as a func-
tion of the per cent of time the turbidity was equal to or less than a
given value. Figure 1 shows the data on logarithmic probability paper
so that the data plots as a straight line. Table 2 lists the per cent
of time that the water has a turbidity less than the given value.
Ninety per cent of the time the raw water has a turbidity less than
1.0 JTU and 95 per cent of the time the turbidity is less than 1.9 JTU.
The working values at the bottom of Table 2 will be used in evaluating
diatomite filter plant design and operating costs. During the phase
three pilot plant tests, the lake water turbidity was between 0.5 and
1.4 FTU, values that are exceeded only about 8 per cent of the time.
11
-------
Table 1. City of Duluth water supply, Lakewood Pumping Station raw water turbidity data, 1952-1972
Year
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
Totals
Turbidity
Percent of
than Given
No. of
Total
Hours
l-»-Ki^ < t-v Ti
Per Cent
Time
Hours Recorded* < 1
6062
5871
5815
5913
6076
6050
5705
5964
5806
5907
6049
5556
5457
5668
5786
5766
5821
5949
6310
5814
5826
123,171 12
Tine Turbidity
Value
244
315
281
172
108
81
31
232
422
147
508
530
376
656
284
744
866
892
772
1315
3406
, 384 Mean
Less
96.0
94.6
95.2
97.1
98.2
98.7
99.4
96.1
92.7
97.5
94.1
90.5
93.1
88.4
95.1
87.1
85.1
85.0
87.1
77.4
41.5
89.95
< i
89.95
1-5
106
233
263
128
93
55
10
143
315
123
478
445
326
635
263
538
547
727
689
1034
3271
10,422
< 5
98.41
Number tours
6-10
90
46
12
41
15
10
21
59
68
18
26
59
38
21
21
136
143
111
67
167
108
1,277
< 10
99.45
11-15
26
36
3
0
0
0
0
13
15
0
4
25
7
0
0
22
84
29
7
96
23
390
< 15
99.76
Turbidity Within
16-20
12
0
3
3
0
10
0
10
24
0
0
0
1
0
0
14
52
8
0
15
0
152
< 20
99.88
21-30
10
0
0
0
0
6
0
7
0
8
0
1
4
0
0
21
20
17
0
3
0
97
< 30
99.96
Indicated JTU Ranges
31-40
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
14
0
1
0
4
21
99.97
41-50
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
6
0
2
0
0
10
-------
5040 30 20
10
2 1 0.5 0.20.1 0.05 0.01
1.0
50 60 70 80
95
98 99 99.899.9 99.99
PEJBCENT OF TIME TURBIDITY IS EQUAL
TO OR LESS THAN GIVEN VALUE
Fig. 1. City of Duluth, Minnesota, Lakewood Pumping Station,
Lake Superior raw water turbicRty, 1952-1972.
13
-------
Table 2. Percentage of time given raw water turbidity is
equalled or exceeded, City of Duluth, Lakewood Pumping Station,
1952-1972.
Raw Water
Turbidity, JTU
1.0
2.0
3.0
4.0
5.0
10.0
20.0
30.0
50.0
Working values
1
1.9
5.8
20.5
57.5
Per Cent of Time
Equal or Less than
89.95
95.3
97.2
98.2
98.7
99.6
99.9
99.99
99.985
from Fig. 1
90.0
95.0
99.0
99.9
99.99
14
-------
Potential Treatments with Diatomite Filters
Diatomite filter aids are produced by several manufacturers in
several grades. Table 3 is a table of filter aid grades currently
available on the market. The table lists the trade name of filter aid
grades from several manufacturers that are approximately "equivalent."
Filter aids are considered "equivalent" when they produce approximately
the same flow rate and filtered solution clarity under the same oper-
ating conditions when filtering a standard sugar solution. Grades
that produce the highest clarity will also produce the lowest flow
rate. In general, the high clarity filter aids are composed of very
small particles of filter aid (mean size in the range of 3-6 (j,m) and
high flow rate filter aids are composed of larger particles (mean size
in the range of 20-40 |j,m) . High clarity filter aids are cheaper, pro-
duce a better clarity of filtered water, and produce the highest rate
of pressure drop increase across the filter and, therefore, shorter
filter runs.
In the pilot plant work, therefore, the filter aids and/or filter
aid treatments are selected which will produce the desired clarity of
filtered water or the desired removal of asbestiform fibers with the
lowest rate of head loss increase. The rate of head loss increase is
measured by a filter cake resistance index, 3 [3,4]. In the pilot plant
studies all grades of filter aid were not available; filter aid pur-
chases were made periodically of grades that were found to be poten-
tially useful. The principal grades used in the study, in order of
15
-------
Table 3. Filter aid ratings.
Standard Ratios8
Flow Rate
100
125
200
300
400
700
800
950
1000
1800
2500
3000
4500
5500
Clarity
1000
1000
995
986
983
970
965
963
960
948
940
936
930
927
Eagle-Picher
Celatom FP-2
Celatora FW-2
Celatom FP-4
Celatom Ftf-6
Celatom FW-10
Celatom FW-12
Celatom FW-14
Celatom FW-18
Celatom FW-20
Celatom Fff-40
Celatom FW-50
Celatom FM-60
Celatom FW-70
Celatom FW-80
Johns -Manville
Filter Cel
Cellte 505
Standard Super-
Cel
Cellte 512
Byflo Super Cel
Cellte 501
Cellte 503
Celite 535
Cellte 545
Celite 550
Celite 560
Dicalite
215
Superaid, DF
Speed flow
Special Speed-
flow, 231
341
Speedplus , 689
CP-100
375
CP-5
Speedex, 757
4200, CP-8
4500
5000
___
aBased on bomb filter tests with 60 Brix raw sugar solution, 80 C.
16
-------
decreasing water clarity achieved, were (Table 3):
FW-6, Celite 512
FW-10
FW-12, Hyflo Super Cel
FW-20, Celite 503
FW-50, Celite 535 (Aquacel)
As the mean particle size of the filter aid gets smaller, the total
surface area of a given weight of material increases. In order to re-
move very small particles of suspended solids such as asbestiform
fibers, the particles must be transported through the fluid over to the
surface of the filter media, and then the particles must become attached
to the surface of the filter media. Both particle transport and attach-
ment are enhanced as filter media particle size is decreased. Diato-
mite filter aid normally carries a negative surface potential. Also,
most normally encountered suspended solids also carry negative surface
potentials. When such particles are filtered through negatively charged
filter media, the transport forces have more difficulty in moving the
particles to the media surface where attachment can take place. Thus
pretreatment of the water or pretreatment of the filter aid can be de-
signed and used to change the surface potential either of the suspended
solids or of the filter media. One must be changed to carry a surface
charge opposite to that of the other.
The removal of asbestiform fibers by such treatments presents an
anomaly. Two types of asbestiform fibers are found in Lake Superior -
amphibole particles and chrysotile particles. Amphibole particles have
fiber diameters (or widths) from 0.1 to about 2 to 3 pm. Chrysotile
17
-------
particles have a fiber diameter that is consistently less than 0.08
to 0.1 (j,m. Asbestiform fibers, by definition, are particles which have
a length/width ratio equal to or greater than three. Both fiber forms
are usually present in roughly equivalent numbers (see Table 4). Thus,
because of the different sizes of the two asbestiform particles, the
mass of amphibole particles is consistently and significantly greater
than the mass of chrysotile particles.
Unfortunately, current analytical techniques do not always produce
the same fiber count when identical samples are analyzed by two or
more laboratories• An order of magnitude difference in such results is
not unusual. If a series of samples is processed by the same labora-
tory, however, more consistent results can be achieved. Standard labo-
ratory procedures are still under development.
During this study, several methods were used to evaluate the qual-
ity of the raw and filtered water:
Amphibole fibers per liter (Transmission Electron Microscope Count)
Chrysotile fibers per liter ( " " " " )
Total fibers per liter ( " " " " )
Mass of suspended solids, mg/1 (X-Ray Diffraction)
Mass of amphibole fibers, mg/1 (X-Ray Diffraction)
Turbidity, FTU (Hach 2100A Turbidimeter)
Amphibole asbestos fibers are negatively charged (Fig. 2) [5],
whereas chrysotile asbestos fibers are positively charged. Thus water
or media treatments that are designed to enhance removal of one species
will interfere with the removal of the other species. In general,
transport forces are a minimum when particles are about 0.5-2.0 \j, in
18
-------
RELATIVE
SURFACE CHARGE
-------
size. Significantly smaller particles are transported effectively by
diffusion forces. Since the size of the amphibole particles is in the
size range that would be difficult to transport and since they are
present in larger numbers, water and filter media treatments used were
designed to enhance removal of these negatively charged particles.
In diatomite filtration, the prime filter media consists of a pre-
coat layer of filter aid with a thickness of about 1/8 in. (0.1-0.2 Ib
of filter aid/sq ft filter). The addition of body feed filter aid
throughout a filter run is designed primarily to maintain a desirable
permeability of the case which is formed (suspended solids removed by
the filter plus the body feed filter aid). It does, however, also pro-
vide an increasing depth of filter media as the run progresses and
thus serves to increase both the area of media surface available for
particle adsorption and the opportunities for transport of particles
to the media surface.
There are only a limited number of water and/or media treatments
that can be expected to improve solids removal on a diatomite filter:
Use of filter media without chemical treatment of the water or the
media.
• Use of a finer grade of media as precoat to improve retention, but
with a greater rate of head loss increase.
• Use of a greater depth of precoat, but at an increased cost for
the filter aid.
• Use of two filter media, a fine media for the precoat to improve
solids retention and a coarser media for the body feed to improve
the hydraulic characteristics of the case. The operation of the
20
-------
filter would be slightly more complicated by the use of two fil-
ter media.
Use of chemical pretreatments of the filter media [6].
•Application of an organic or inorganic coating to the filter
media used as precoat to improve particle retention.
•Application of an organic or inorganic coating to the filter media
used as body feed to improve particle retention.
•Application of such a coating to both precoat and body feed fil-
ter aid.
Use of chemical treatment of the raw water.
•Use of inorganic and/or organic coagulants to change the particle
characteristics to improve their retention in the diatomite filter
media.
In view of the normally expected effectiveness of diatomite filters
for the removal of suspended solids from low turbidity waters (0-30 FTU),
no provision was made for the coagulation-flocculation-sedimentation
normally provided with single, dual, or multi-media granular filters.
When used, coagulants were added immediately prior to the main filter
pump without provision for more than a few seconds of mixing time before
the water reached the filter media.
Since the amphibole particles had a negative surface (zeta) poten-
tial, it was expected that a positive coating applied to the filter
media would be most effective. Two such coatings had been used in pre-
vious studies - a cationic polymer, C-31, manufactured by the Dow Chem-
ical Company, and alum. Figures 3 and 4 show the effect of coating
level at a pH of 6.5 on the surface potential of Celite filter media
21
-------
>
E
ALUM AS AL(OH)3
WITH CELITE 535
0.001 0.002 0.003 0.004
COATING LEVEL, gm coating per gm of diatomite
0.005
Fig. 3. Zeta potential, C, and filter case resistance, £, as a
function of coating level for Alum coated on Celite
535 (6).
22
-------
80
PURIFLOC 601 WITH CELITE 545
0.001 0.002 0.003 0.004
COATING LEVEL, gm coating per gm of diatomite
0.005
Fig. 4. Zeta potential, C, and filter case resistance, £, as a
function of coating level for C-31 coated on Celite
545 (6).
23
-------
coated with alum and C-31, respectively. Both coatings effectively
provide a positive surface characteristic to the filter media at a
water pH of 6.5. At Duluth, the water in Lake Superior during the period
of the pilot plant operations had a normal pH from seven to eight.
Previous studies had demonstrated that C-31-coated media had a negative
surface potential at pH values above 7.0 and a positive surface poten-
tial at pH values below 7.0 [6]. Figure 2 indicates that chrysotile
asbestifonn particles carry a positive surface potential at pH values
between about two and thirteen. No similar data is available to eval-
uate the effect of pH on the surface potential of amphibole particles
or of alum-coated filter media. It was hypothesized, however, that
alum-coated filter medias would carry a positive surface potential,
and the amphibole particles would carry a negative surface potential in
the Lake Superior water pH range of seven to eight. Use of other media
coatings was based on an "Edisonian" approach (try it and see if it
would work) rather than on detailed knowledge of the coating charac-
teristics .
The treatments designed and used for the removal of amphibole
particles included:
Using different grades of filter aid.
Celite 535
Celite 503
Hyflo Super Gel
Celite 512
etc.
24
-------
Coating precoats and/or body feed with
C-31 (a Dow cationic polymer)
Alum
573-C
A-23 (a Dow anionic polymer)
Magnifloc 985H (a Cyanamid non-ionic polymer)
Coagulation of the water prior to filtration
Cat-Floe B (Calgon Corporation cationic polymer)
Of all the treatments used, the treatments listed in Table 4 of-
fered the most potential for use in a diatomite filtration plant at
Duluth based on the pilot plant studies.
Pilot Plant Results
As the pilot plant runs were made, routine analyses were made of
the run results. A number of problems were encountered which seriously
affected the ability to interpret adequately the test results. These
included the following:
1. The time required for analysis of samples collected for as-
bestos fiber determinations was so long (four to six weeks
frequently) that the periodic results available were of little
use in adjusting the test schedule.
2. The early test results indicated serious passage of filter
aid into the filter effluent due to the fact that the filter
septa were not fine enough to retain the finer precoat filter
aids. Subsequently, a two-layer precoat was used. The first
25
-------
Table 4. Water and/or filter media treatments tested for optimization
runs.
Case
Treatment
a,b
Precoat Alum-coated C-512
Body Feed C-512
Coagulant None
Precoat Alum-coated C-512
Body Feed Hyflo Super Cel
Coagulant None
Precoat Alum-coated Hyflo Super Cel
Body Feed Alum-coated Hyflo Super Cel
Coagulant None
Precoat Alum-coated Hyflo Super Cel
Body Feed Hyflo Super Cel
Coagulant Cat-Floe B (0.59 mg/1)
Precoat r Alum-coated C-512
Body Feed Aquacel (C-535)
Coagulant Cat-Floe B (0.33 mg/1)
All alum coatings at 1% by weight of Al
All filter aids produced by Johns-Manville Products Corp.
Cat-Floe B is a cationic (primary coagulant) polyelectrolyte
produced by the Calgon Corporation.
26
-------
layer consisted of coarse filter aid (FW-50 or C-535) used to
retain the finer precoats found necessary for good filtration.
The second layer consisted of the grade of filter aid con-
sidered necessary to achieve successful suspended solids re-
duction. Only the second precoat layer was ever coated. The
first precoat layer was considered to be a part of the septum.
3. In early runs, the major problem encountered concerned the
consistency of suspended solids removal. In many runs, the
turbidity level (used as a measure of suspended solids) re-
mained low for only a portion of the filter run and then in-
creased to levels approaching turbidity levels in the raw
water.
The phase two series of runs conducted to determine which treat-
ments gave promise of removing asbestiform fibers were made under
fairly standard conditions:
Filtration rate 1.0 + gpm/sq ft
Precoat weights 0.1 Ib/sq ft of coarse media (part of
septum)
0.1 Ib/sq ft of test media
Body feed rate 50-70 mg/1 (more or less only to pro-
vide run of at least 12 hours and no
more than 2 days)
In order to classify success or failure in achieving the desired
water quality, the primary objective was to vary precoat, body feed,
and filter aid coating treatments until the average filtered water
turbidity reached a consistent goal of 0.1 FTU or better. That goal
was reached when either the C-512 or Hyflo Super Gel (or other equiva-
lent filter aids) were used with an alum coating. Other coatings did
27
-------
not consistently produce such water quality. For example, the C-31
polymer coatings used in the basic Lake Superior water did not prove
consistently effective in turbidity reduction.
Once it was established that alum-coated filter aids used as pre-
coats could produce the desired water quality, additional studies were
made to evaluate the use of various types and amounts (mg/1) of body
feed for providing both better turbidity removal and better cake hy-
draulic characteristics during the filter run. In this series of runs,
the body feed rate was different in each of four or five runs in the
series. The variation in the body feed rate made it possible to opti-
mize the filter design for use with that "treatment." When the asbes-
tos fiber analyses were ultimately evaluated, the five cases described
in Table 4 were found to provide the suspended solids removal summarized
in Table 5. The average turbidity of the final effluent in each series
of runs was as follows:
Filtered Water Turbidity. FTU
Case 1 0.09 (Runs 59-62)
Case 2 0.10 (Runs 55, 57, 58)
Case 3 0.05 (Runs 67, 68, 69)
Case 4 0.05 (Runs 73, 74, 75)
Case 5 0.06 (Runs 79, 80, 81)
Analysis of the data in Table 5 indicated that amphibole asbestos
fiber removal generally was above 80 per cent, and when finished water
turbidity was 0.05 or 0.06 FTU fiber removal was over 95 per cent. In
most cases the amphibole fiber removal was significantly better than
chrysotile fiber removal. This may be due to several conditions:
28
-------
Table 5. Suspended solids removal in selected series of filter runs.
Treatment
Case
1
2
3
4
5
ORF
ORF
NWQL
NWQL
UMD
B&V
ORF
ORF
NWQL
NWOL
UMD
B&V
ORF
ORF
NWQL
NWQL
UMD
B&V
ORF
ORF
NWQL
NWQL
UMD
B&V
ORF
ORF
NWQL
NWQL
UMD
B&V
Suspended Solid
- Amphibole fibers/liter
- Chrysotile fibers/liter
- SS, mg/1
- Amphibole fibers, mg/1
- Total fibers/liter
- Turbidity, FTU
- Amphibole fibers/liter
- Chrysotile fibers/liter
- SS, mg/1
- Amphibole fibers, mg/1
- Total fibers/liter
- Turbidity, FTU
- Amphibole fibers/liter
- Chrysotile fibers/liter
- SS, mg/1
- Amphibole fibers, mg/1
- Total fibers/liter
- Turbidity, FTU
- Amphibole fibers/liter
- Chrysotile fibers/liter
- SS, mg/1
- Amphibole fibers, mg/1
- Total fibers/liter
- Turbidity, FTU
- Amphibole fibers/liter
- Chrysotile fibers/liter
- SS, mg/1
- Amphibole fibers, mg/1
- Total fibers/liter
- Turbidity, FTU
Concentration
in Raw Water
300,000
720,000
0.81
0.02
25,300,000
0.73
10,000
170,000
0.81
0.06
8,980,000
0.70
780,000
330,000
0.41
0.07
17,800,000
0.58
1,000,000
500,000
0.6
0.04
15,400,000
0.58
900,000
100,000
0.31
0.08
20,300,000
0.48
Concentration
in
Filtered Water
0
170,000
0.04
< 0.003
810,000
0.09
40,000
40,000
0.16
< 0.003
1,530,000
0.10
0
540,000
0.008
< 0.003
270,000
0.05
0
1,000,000
0.03
< 0.003
640,000
0.05
20,000
60,000
0.04
< 0.003
810,000
0.06
Per Cent
Removal
100
76
95
85
97
•88
?+
77
80
95
83
86
100
?+
98
96
99
91
100
?+
95
93
96
91
98
40
87
96
96
88
ORF = Ontario Research Foundation
NWQL = EPA National Water Quality Laboratory, Duluth
UMD = University of Minnesota, Duluth
B&V = Black and Veatch
29
-------
1. The precoat coating used was designed for the removal of
negatively charged particles similar to the amphibole par-
ticles;
2. The chrysotile fibers are very small and positively charged,
and their removal was undoubtedly retarded by the precoat
treatments used;
3. There must be an anomaly in the measurement techniques used
to count chrysotile fibers, since the results frequently
showed more (by a factor of two) fibers leaving in the
filter effluent than there were present in the raw water
(Cases 3 and 4).
To date no criteria have been established for evaluating what re-
moval of asbestos fibers is desired. Accordingly, the results of all
five cases were analyzed in detail to determine the optimum design
characteristics of the diatomite filtration plant needed ^n each case
to produce the lowest cost of filtered water.
Determination of Filtration Resistance
The design of a diatomite filtration plant requires specification
of the following items:
1. The design filtration rate, gpm/sq ft,
2. The terminal pressure drop to be built up across the filter
cake, feet of water,
3. The precoat type and weight,
4. The body feed type and amount to be used.
30
-------
The research group at Iowa State University has developed and
published mathematical models for and described techniques which can
be used in the optimum design and operation of diatomite filtration
plants [3,4,7].
In order to optimize filter design, it is first necessary to
evaluate the filter cake resistance that is generated when any given
ratio of suspended solid to body feed exists in a filter cake. In
order to accomplish this, it is necessary to have a record of:
1. The head loss generated across the filter cake as a function
of time during a filter run,
2. The average raw water turbidity during the filter run,
3. The average body feed rate, CD, in mg/1, during the filter run.
These data were collected routinely in each run (Fig. 5).
The data for head loss vs time from each run in all five cases
were fed into a computer, and the filter cake resistance of each
_2
filter cake (B, in ft ) was calculated. The BID (Beta Index Deter-
mination) program developed at Iowa State University was used for this
purpose. The results of each run provided two related parameters:
1. The cake resistance, represented by B,
2. The proportion of solids to body feed in the cake as repre-
sented by the ratio, T/CD (Turbidity, FTU/Body feed, mg/1)
If two or more (preferably many more) runs are made under similar
conditions using different T/CD ratios, it is possible to determine a
so-called Beta prediction equation which can be used to predict the
filter cake resistance that will result for any "reasonable" ratio of
T/CD in the filter cake. For example, Fig. 6 shows a plot of Beta as
31
-------
18
NJ
0
8 10 12 14 16 18 20 22 24
RUN LENGTH, hours
Fig. 5. Turbidity and pressure drop as a function of run length,
Run 79, Cose 5.
-------
io.or
X
OL
0.1
RUN 79
80
81
3 1.20xl06 0.38 x IO6 0.160xl06
RAW WATER TURBIDITY FTU
0.42 0.50 0.52
CD, mg/l
27.4 71.6 135.3
T/CD 0.01533 0.00698 0.00384
97.2
97.8
95.0
FILTERED WATER TURBIDITY FTU
0.05 0.06 0.07
p - PREDICTION EQUATION
p = 108.7254
PRECOAT
BODY FEED
COAGULANT
FW -50
C -512, ALUM
COATED
AQUACEL
CAT FLOC B
0.33 mg/l
i i i i i
0.001
0.01
(T/CD)
0.10
Fig. 6. Development of a Beta prediction equation for Case 5.
33
-------
a function of T/CD for Case 5. The figure includes the calculated
value of B, the raw water turbidity, the body feed, the run T/CD value,
and R, the correlation coefficient describing the statistical fit of
the head loss to time data for the run. An R value of 100 would indi-
cate perfect correlation. The results of the three runs shown in Fig. 6
were used to determine a Beta prediction equation.
The Beta prediction equation is frequently expressed in the form
Beta = 10a(T/CD) CDC §d where § is a measure of the cake resistance
of the clean filter aid. However, in this study the Beta prediction
equations are expressed in the form
Beta = 103(T/CD)b
The Beta prediction equation exponents determined in each case are as
follows:
a b
Case 1 7.207 0.735
Case 2 6.987 0.760
Case 3 9.252 1.393
Case 4 9'.053 1.628
Case 5 8.725 1.458
Figure 7 shows a plot of the Beta vs T/CD for Cases 1-4. Case 5
is not shown since it falls very close to Case 4. The results for
Cases 1 and 2 are typical of Beta prediction curves obtained in other
studies using filter aids without chemical treatment of the body feed
34
-------
lo.or
.001
o.oi
TURBITY, FTU
BODY FEED, mg/l
Fig. 7. Beta prediction relationships for Cases 1-4.
35
-------
filter aid. Use of C-512 for body feed provides a higher cake resis-
tance at the same T/CD ratio than does Hyflo Super Gel because it is a
finer filter aid. When the Hyflo Super Cel used as body feed is coated
with alum, the cake resistance increases, and increases at a faster
rate with increasing values of T/CD. It does, however, provide the
best effluent quality. Case 3, because of the higher cake resistance,
should also provide the highest filtration cost. In both Cases 4 and 5
the use of Cat-Floe B as a coagulant provides excellent water quality,
but the filter cake resistance is less than in Case 3. Both cases
should provide nearly the same filtration cost. The data for Case 5 is
probably the most reliable, since there was a better fit of the indivi-
dual Beta values to the curve. With the 5 Beta prediction equations,
it was possible to design the optimum diatomite filtration plants for
each case so that each plant would operate with maximum economy.
Determination of Optimum Plant Designs
Diatomite filters can be designed as pressure filters (head losses
across filter cakes to about 200 ft of water maximum) or as vacuum fil-
ters (head loss across filter cakes of 20-22 ft). Vacuum filters are
easier to operate and maintain, but the cake is under a pressure less
than atmospheric, and dissolved gases in the water can come out of
solution and collect in the cake causing an unnecessary resistance in-
crease. This was observed to happen in the pilot filter runs, parti-
cularly with water at about 33-40 °F. In general, pressure filters are
more economical than vacuum filters under higher turbidity loadings
36
-------
and/or where power costs are relatively low. The filter cake resistances
should be the same using vacuum or pressure filters provided gases are
not collected in the cake. Therefore, the Beta prediction equations
listed in the previous section were used to optimize the design of both
pressure and vacuum diatomite filtration plants.
A computer "Program for Optimization of Plant Operation" (POPO),
developed at Iowa State University, was used to find the optimum design
characteristics of the plants in each case. It was assumed initially
that the design flow of the plant was 30 mgd, and costs are based on
the production of 30 mgd during each day of operation.
The operation of POPO requires the input of 20 items of data.
Tables 6 and 7 list the data input into the computer for pressure and
vacuum diatomite plant design, respectively. In general, the data are
self explanatory, but some items deserve some additional comment (see
Table 6).
Item 4. This is the assumed rate of interest that cities on
Lake Superior might pay on the bonded indebtedness of
the first cost.
Item 6. The turbidity of the raw water varies up to about
57.6 FTU. A turbidity of 1.0 is exceeded only 10 per
cent of the time. The plant design is repeated using
other turbidity values to provide the least possible
yearly plant operating costs.
Items 7,
10, 12,
and 16 These will be explained later.
37
-------
Table 6. Input Data, POPO (Program for Optimization of Plant Operation),
asbestos-turbidity removal, Lake Superior water, Pressure
Diatomite Filter Plant.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
Design Flow
Salvage Value
Energy Conversion
Interest Rate
Plant Life •
Solids, C
s
XI Index •
Temperature
Precoat Weight
Precoat Density
Septum Diameter --•
Beta Prediction
Unit Flow Rate
Body Feed
Terminal Head
Diatomite Cost -•
First Cost
Area, sq ft
1,000
3,472
6,944
20,832
41,667
80,000
100,000
Power Cost
Labor Cost
($5/hr)-1.25 men
2.0 men
4.0 men
6.0 men
Area, sq ft
1,000
3,472
6,944
20,832
41,667
80,000
100,000
20. Backwash Cost
30 mgd
30% First Cost
70%
6.68%
20 Years
1.0 FTU (by turbidity)
ft/lb
33 °F
0.15 Ib/sq ft
Ib/cu ft
3.5 in
0.5/0.1/6.0 gpm/sq ft
10/5/150 mg/1
20/5/150 ft
$/ton
$/sq ft
$616
376
259
155
134
122
120
1.627 c/KWH
$/sq ft per month
$1.87
1.30
1.04
0.69
0.52
0.40
0.35
10,30 gal/sq ft, rain
38
-------
Table 7. Input data, POPO (Program for Optimization of Plant operation),
asbestos-turbidity removal, Lake Superior vater, vacuum
diatomite filter plant.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
Design Flow
Salvage Value
Energy Conversion
Interest Rate
Plant Life
Solids, C
s
XI Index
Temperature
Precoat Weight —
Precoat Density -•
Septum Diameter -•
Beta Prediction -•
Unit Flow Rate --•
Body Feed
Terminal Head
Diatomite Cost —
First Cost
Area, sq ft
3,472
6,944
20,832
41,667
Power Cost
Labor Cost Area, sq ft
($5/hr) 1.25
3,472
6,944
20,832
41,667
20. Backwash Cost
30 MGD
30% First Cost
70%
6.68%
20 Years
1.0 FTU (turbidity)
ft/lb
33 °F
0.15 Ib/sq ft
Ib/cu ft
Flat
0.5/0.1/6.0 gpm/sq ft
10/5/150 mg/1
20 ft
$/ton
$/so ft
$479
355
257
237
1.627 C/KWH
$/sq ft per month
1.17
0.936
0.621
0.468
10,30 gal/sq ft, min
39
-------
Item 8. This is the temperature of the water under the worst
plant operating conditions. When the water is warmer,
the operating costs will be lower.
Item 11. This is the diameter of the septa in pressure filters.
For vacuum filters, flat septa are specified.
Item 13. This tells the computer to compute filtration costs for
filtration rates from 0.5 to 6.0 gpm/sq ft in increments
of 0.1 gpm/sq ft.
Item 14. This tells the computer to compute filtration costs for
body feed rates from 10 to 150 mg/1 in increments of
5 mg/1.
Item 15. This tells the computer to compute filtration costs for
terminal head losses from 20 ft to 150 ft in increments
of 5 ft.
Item 17. This tells the computer the cost per square foot of the
filter area required as related to the total filter area
requirement. The computer interpolates between these
values when other areas are required. The cost of the
pressure filters was derived from Table 8 and the cost
of the vacuum filter was derived from Table 9.
Item 19. This tells the computer the monthly cost per square foot
of the filter area required as related to the total
filter area requirement. The computer interpolates
between these values when other areas are required. The
pressure filter labor requirement is shown at a wage
rate of $5/hr. It is assumed that each man indicated is
40
-------
Table 8. Capital construction costs of pressure diatomaceous earth or
diatomite water filtration plants with varying filter rates,
plant design for 30 mgd.b
Item
Building and
associated costs
Filter equipment
Body feed and precoat
slurry feeders and
storage tanks
Equipment set-up and
assembly
Totals
0.5
($xio6)
2.288
2.150
0.500
0.645
$5.583
Filter Flow
1.0
($X106)
1.430
1.075
0.400
0.324
$3.229
Rate, gpm/sq
3.0
($X106)
1.144
0.350
0.200
0.105
$1.800
ft
6.0
($X106)
0.858
0.189
0.200
0.058
$1.305
Cost data from Black and Veatch, Consulting Engineers.
Based on lowest cost filtration equipment currently commercially available.
Table 9. Capital construction costs of vacuum diatomaceous earth or
diatomite water filtration plants with varying filter rates,
plant design for 30 MGD.
Item
Building and
associated costs
Filter equipment
Body feed and precoat
slurry feeders and
storage tanks
Equipment set-up
and assembly
Totals
0.5
($X106)
2.288
5.437
0.500
1.631
$9.856
Filter Flow
1.0
($X106)
1.430
2.718
0.400
0.815
5.363
rate, gpm/sq
3.0
($xl06)
1.114
0.937
0.200
0.218
2.469
ft
6.0
($X106)
0.858
0.468
0.200
0.140
$1.666
Coat data from Black and Veatch, Consulting Engineers.
Based on costs of filtration equipment currently commercially available.
41
-------
required 24 hrs per day. More men are required with
larger filter areas. The vacuum filter labor costs are
assumed to be 90 per cent of those needed for pressure
filters.
Item 20. This tells the computer that 10 gal of water are required
for backwashing each square foot of filter and that for
30 min the filter will be out of service for backwashing.
In each computer run, the costs are computed as if the water al-
ways had the turbidity and temperature shown. In general, lower operating
costs will be obtained since the turbidity is less than 1 FTU 90 per
cent of the time, and the water temperature is more than 33 °F (and
easier to filter) for some of the year.
Items 7, 10, and 16 refer to the characteristics of the filter aid
grade used. In all POPO runs the characteristics of the body feed fil-
ter aid were used in these items. Table 10 lists the appropriate values
for the filter aids used in Cases 1-5. The 5 index (item 7) measures
the cake resistance of the pure diatomite without any suspended solids.
The filter aid density (item 10) is used to measure the thickness of
filter cake that builds up during filtration. The per ton cost of
filter aid (item 16) represents the December 1974 cost delivered to
the Lake Superior region.
The following computer runs were made to determine the optimum
plant design and operating costs at each given turbidity level:
42
-------
Table 10. POPO input data, filter aid characteristics (approximate values).
Filter Aid Cost, $/ton |, ft/lb Density, Ib/cu ft
Celite-512 132 5.5 x 109 21.0
Hyflo Supercel 134 4.74 x 109 20.7
Celite-503 146 3.05 x 109 19.9
Celite-545 (Aquacel) 152 1.77 x 109 19.7
FW-20 144 2.57 x 109 20.7
43
-------
a
Turbidity Level Assumed . FTU
Case 1
Case 2
Case 3
Case 4
Case 5
Pressure Filter Vacuum Filter
1.0 1.0
1.0 1.0
1.0, 1.9, 5.8, 20.5 1.0
1.0, 1.9, 5.8, 20.5, 57.5 1.0
1.0, 1.9, 5.8, 20.5, 57.5 1.0
a
Working values from Table 2.
In the solution of the POPO runs on the computer, two items of
computer output were usually requested.
1. A listing of the data input into the computer,
2. A list of the ten cheapest operating combinations of filter
design (filtration rate, terminal head loss, body feed rate),
the Beta value of cake resistance produced, the filter area
required, the optimum length of filter run, the thickness of
the filter cake at the end of the run, and the filter operating
costs. The operating costs ($/mg) include the following
separately indicated items: Total Cost
First Cost
Operating Costs
Labor and Maintenance
Power
Filter Aid
Total Cost, $/Month
44
-------
In order to evaluate the changes that would occur in plant design
and operating costs if the Beta resistance varied above and below the
predicted value, the same data as outlined above were output for Beta
resistance values of 50, 75, 100, 125, 150 and 175 per cent of the Beta
prediction equation computed values. An example of the data input and
the (100 per cent [3 value) computer printout is shown in Tables 11 and
12. Table 11 shows the computer input data required by the computer
to optimize the design of pressure diatomite filters for the removal of
1.9 FTU of turbidity at a water temperature of 33 °F using Case 5 pre-
coat and body feed treatments. Table 12 shows the 100 per cent {3 value
computer output data for this input. All ten least-cost combinations
of design data show about the same per 1000 gallon filtration cost of
5.89-5.90 cents. In all cases the economic filtration rate is 1.5-1.7
gpm/sq ft. In all cases the body feed rate is 6 or 8 mg/1. In general,
the optimum terminal head varies from 120 to 140 ft, with the best op-
timum (lowest monthly cost) showing a terminal head of 130 ft of water.
These values are, therefore, probable best operating conditions and are
entered into a summary table (Table 13) in the appropriate place, as
follows:
Case 5, turbidity 1.9 FTU
Optimum filtration rate 1.60 gpm/sq ft
Optimum terminal pressure 130 ft of water
Optimum body feed 8 mg/1
Predicted length of run 14.6 hr
Predicted terminal cake thickness 0.14 in
45
-------
Table 11. Data input to POPO, pressure diatomite filter, Case 5, raw
water turbidity of 1.9 FTU.
C JOB X5. ASBESTOS-TUpqi[
0
0
TRI AL 1
RUNS
79,80f81 I
0 PPES$URE D. E.
f\
FINAIJ TUPBIDIT1
1 DE^ IGN FLOW
? SAL,
3 ENE
4 INT
5 PL^
6 SOL
7 XI
8 TEN
9 PRE
LVAGE VALUft
PGY CONVERSION
EREST PATf
,NT LIFE
IDS (CS)
INDFX
IPERATUPE
COAT WFIG>|rr
1C PPFfCOAT DENSITY
1 1 SFPJ TUM DIAMETER
12 BE{A PREDICT;
13 UNJT FLOW RAT
14 BOC
1 «S TER
If D I 4
1 7 F IP
IV FEED (Cf
MINAL HEAt
iTOMITE CO<
ST COST ^
* i
18 POVfER COST
19 LA$OR COST
i
!
*
ON
T
»
T
2T BACKWASH COST
flFCf 1| N
>ITY REMOVAL; DULUTHJ MINNESOTA
kQUACFL * (
PLANT
' * 0 • ^ ^
v>
30
70
!•- - -- -
;ATFLOC B
MGD
PERCENT FIRST COST
PERCENT
...
6. 68 PERCENT
20 j YEARS i
1.9 '
1 .77F9
33
MG/L (BY TURBIDITYJ
FT/LB
DEGREES P
C'. 15 I LB/SF
19.7 LB/CF
3.5
8.725/1 .i
0 .4/0.1/1
6/2/30
100/10/2<
152
INCHES
5800/0/0
. 8
0
S/TON
APFA 1 S/SF
1000
3473
6944
20832
41667
8000C
1 OOQOC
1.62'
AREA
1000
3472
6944
2C832
41667
8COOC
1 00000
61 6
376
, 259
' 155
1 34
I 122
120
' CENTS/KWH
1 S/SF PER
1.87
1.30
1 ,C4
0.69
0.52
0.40
0.35
10, 30 GAL/SF, N
1
GSFM
MG/L ;
FT
.
'
i
1 !
MONTH !
r
,
f
I IN ;
46
-------
Table 12. Computer output for data input shown in Table 11.
FLO* TEt
HF>
GSFM FT
.60 ;
.7C :
• TO ;
.60 :
.70 ;
• no :
• ao ;
.sc i:
.70 11
.60 11
.60 i:
.50 i:
.60 1'
.5C 1J
.70 1<
•70 i:
.50 1'
.60 i;
.sc i:
.SC 1'
1 . 5C 1 <
1 . *0 1 <
1 . 50 1 !
1.40 i:
1 . 60 1 •
.40 % 1!
.sc i:
.6C It
• 30 i :
• 6C 1<
M CF B
o
PPM 1<
0 6
C 6
fcO 6
if) 6
0 8
C 6
C 8
C 6
0 8
0 4
0 8
0 8
o a
f a
0 8
C 8
>0 6
>0 8
0 6
to a
C 8
o a
SC 8
10 8
>0 8
(C 8
0 8
0 8
0 8
0 8
f T A Tin
4 -2
FT H«
f
5549 15
E549 15.
5549 13,
5549 16,
3648 15,
5549 13,
3648 14.
5549 17,
3648 14,
3648 16.
(
4864 14,
4864 16,
4864 15,
4864 15,
4864 13.
4864 13
7399 15,
4864 13.
7399 14,
4864 17,
f
6080 14,
6080 16,
6'J80 15,
6^flO 15,
60 BO 13,
6C80 17,
6C80 13,
6080 12,
608I"1 17,
tear 14,
E AREA 1
SO FT
IFTA INDICf
6 13545
0 12763
fl 123*7
9 135*3
9 12713
4 12108
2 12077
7 14384
6 12778
4 13517
IFTA INDIC!
6 13583
5 14421
7 13541
3 14467
<5 12804
9 12848
5 14457
5 13632
4 1450?
fl 14381
IETA INDICf
3 14509
3 15466
3 14466
2 15512
5 13631
5 15425
3 14558
6 13676
6 16619
4 13592
HICK *
*
IN * TO"
S = 75 P!
r. 3 * s<
0. 3 * 5(
'. 3 * 51
0. 4 * - 5(
C. 5 * 5(
C. 3 * 5(
C. 1 5 * b<
C.14 * 5(
C.15 * 5(
O.I 5 * 5<
S = 100 PJ
C.14* 5>
r. 15 * 5!
C . 1 5 * 5«
r. 14 * 5-
0.14 * 5<
C.14* 5<
0.13 * 5<
0.14 * 5<
0.13* 5'
C.15 * 5<
S = 125 Pf
0.14 6
C.14 6:
0.4 6
C . 4 6
f. 4 6
r. 5 e
C. 4 6
C . 4 6
0.14 6
r . 14 6
COSTS. *
AL 1ST
CCENT OF t
.3 22.2
.3 21.9
i.3 21.9
.3 ??.2
.3 21. 9
.3 21.7
.3 31. 6
..3 22.5
.3 21.9
.3 22.2
RCF.NT OF t
.9 22.2
i.9 22.6
.9 22.2
1.1 22.6
.0 21,9
'. 0 22. i
.0 22.6
'.0 22.3
'.-! 22.6
.0 22.5
RCENT OF [
.1 22.6
.1 23.0
.2 22.6
.2 23.0
.2 22.3
.2 23.0
.2 22.7
.2 22.3
.2 23.4
.3 22.3
PF.R MILLI'
LAy<
OPFR MAir
REOICTEO \
34.0 13.;
34.4 13.i
34.3 13.:
34.1 3.!
34.4 3.;
34.6 3.
34.7 3.C
33.8 3.<
34.4 3.;
34.2 13.!
REOICTED \
36.7 13.'
36.4 13.'
36.7 13.:
36.4 13.)
37.0 13.:
37.0 13.:
36.4 13. f
36.7 13.!
36.4 13.1
36.5 13. f
•REDICTEO >
38.5 13.*
38.2 14.1
38.6 13. f
38 . 1 14.2
38.9 13.'
38.3 14.]
38.5 13. r
38.9 13. t
37.8 14.'
39. C 13. «
IN G ALL i IN 1
PQWR t- I
ALOES
8.8 1
9.5 1
e.e ;
9.5 :
e.s ;
9.5 ;
1.8 ;
8.8 1
s.i i:
8.1 li
•
ALUES
9.6 i:
9.5 i:
10.3 :
8.8 :
10.3 :
9.6 '
10.3 ;
8.8 '
9.6 :
10.3 ;
ALUES
. 10.3
10.3 :
n.o :
9.6
11.0 '
11.0
9.6 1!
10.3 I1
9.6 i:
11.8 1
-- « TJ
* rr
IO » I,
.7 * si :
.ft * 51
.3 * 51
. 1 * 51 .
.4 * 51
'. 1 * 51
.9 * 51
.2 * 51
l.i* 51 :
».6 * 51
1.6 * S3"
1.1 * 53'
1.0 * 53
.7 53
J.5 53
1.2 53
1.3 531
.3 53i
1.0 531
.5 * 53!
.4 * 55
1.7 * 55"
.7 * 55-
..4 * 55"
.3 * 55!
1. 1 * SSi
i.l * 55<
i.O * 55*
.7 * 55«
1. 7 * 55(i
AL
ST
MO
09
15
?*
41
142
154
155
ift6
77
184
'30
44
'60
72
78
91
103
n
15
16
'52
61
74
86
,^9
128
31
134
49
172
-------
Required filter area 13,583 sq ft
Operating cost 5.89 c/1000 gal
Monthly operating cost $53,730 (30 mgd)
If cake resistance goes up or down from the predicted value, the
optimum design conditions will change slightly, producing the new con-
ditions shown in Table 12. In general, if unusual operating conditions
result the computer output will readily indicate what they are. With-
out such conditions the optimum design condition would be considered
acceptable; it has been transferred to the summary Table 13.
Tables 13 and 14 list the optimum design conditions for pressure
and vacuum filters, respectively, for various levels of influent water
turbidity when the plant is optimized for that turbidity (see Fig. 8) .
The data indicate that under similar conditions, the pressure filters
are significantly more economical than vacuum filters. For example,
when the plants are designed to remove 1.0 FTU of turbidity under op-
timum conditions, the comparison of plant operating costs shown in
Table 15 results. Any of the treatments (cases) might be used success-
fully, depending on the quality desired in the filtered water. The
costs follow the range predicted previously on the basis of cake re-
sistance. Since the vacuum filter is limited in terminal pressure
drop to 20-22 ft of water and cannot take advantage of the relatively
low power costs to reduce overall filtration costs, the costs of vacuum
filtration will increase faster than the costs of pressure filters de-
signed for higher water turbidity levels. On this basis, vacuum fil-
ters were eliminated from further consideration.
48
-------
Table 13. POPO output, pressure filter, all cases.
Turbidity, FTU
Case Raw Filtered
1 1.0 0.09
2 1.0 0.10
3 1.0 0.05
1.9
5.8
20.5
4 1.0 0.05
1.9
3.5
5.8
20.5
57.5
5 1.0 0.06
1.9
5.8
20.7
57.5
Optimum Operation Conditions
Q.gpm/
sq ft
3.05
3.20
1.30
1.10
0.80
0.50
1.80
1.50
1.30
1.10
0.90
0.60
1.80
1.60
1.10
0.80
0.50
HL3
50
40
155
195
200
200
110
130
150
180
200
200
110
130
180
200
200
CD
5
5
9
10
20
35
10
12
14
16
28
42
6
8
10
18
26
Cost,
0/1000 gal
3.97
3.83
6.47
7.29
9.47
13.64
5.52
6.10
6.75
7.37
9.65
12.60
5.32
5.89
7.16
9.51
12.62
Run Length,
hours
20-22
28.4
13.4
13.0
11.0
10.0
14.1
14.3
13.1
14.5
9.8
10.2
16.1
14.6
15.5
12.1
13.1
HL = terminal pressure drop across filter cake.
49
-------
Table 14. POPO output, vacuum filters, all cases.
Turbidity, FTU
Case
1
2
3
4
5
Raw
1.0
1.0
1.0
1.0
1.0
Filtered
0.09
0.10
0.05
0.05
0.06
Q,gpm/sq ft
2.65
3.15
0.75
1.20
1.20
Optimum Operating Conditions
HL
20
20
20
20
20
CD
5
5
30
25
15
Cost
C/1000 gal
5.04
4.63
12.44
8.64
8.39
Run Length,
hours
9.9
12.0
7.5
8.8
8.9
Table 15. Comparison of pressure and vaccum diatomite filtration costs
optimized to remove 1.0 FTU turbidity.
Case
1
2
3
4
5
Total Operating Costs,
Pressure Filters
3.97
3.83
6.47
5.52
5.32
C/1000 gal
Vacuum Filters
5.04
4.63
12.44
8.64
8.39
Vacuum
Filter Cost
Pressure
Filter Cost
1.27
1.21
1.92
1.57
1.58
50
-------
SEPTUM
DISCHARGE
TO WASTE
FILTER HOUSING
RECIRCUIATION
TO FILTER PUMP
PRECOAT POT
B—FROM FILTER PUMP
DISCHARGE
TO WASTE
SEPTUM
— INFLUENT
RECIRCUIATION
LINE
— EFFLUENT
Fig. 8. Sections through typical pressure and vacuum diatomite
filters (Table 15), Case 5.
51
-------
Table 13 lists data for three cases for which the filters are op-
timized for higher turbidity levels. In Case 3, when the water tur-
bidity is at ?0.5 FTU, the least cost of filtration proved to be 13.64 c/1000
gallons. (In all cases the terminal pressure drop in pressure filters
was limited to 200 ft of water.) With optimized plants Cases 4 and 5
can filter the same turbidity of water (20.5 FTU) for 9.65 and 9.51 c/1000
gallons, respectively. Figure 9 shows the optimum filter operating
characteristics as related to TTHW water turbidity for Case 5 pressure
filters. Since Case 5 filter operation produces essentially the sane
filtrate quality as did the Case 3 and Case 4 treatment, further study
was made to determine the approximate average operating cost of Case 5
filters over a period of a year where the filter might encounter raw
water turbidities as high as 57.5 FTU for a short period of tine.
In the plant design procedure the next question of concern is,
"What base turbidity should be used for plant optimization to minimize
the total annual operating costs of the plant?" A study of this ques-
tion was made for Case 5. Table 13 indicates the following design
conditions for raw water turbidities of 1.0, 1.9, and 5.8 FTD:
Filtration Rate, Terminal Head, Run Length,
Raw Water Turbidity. FTP gpn/sq ft ft of water hours
1.0
1.9
5.8
1.80
1.60
1.10
110
130
180
11.0
14.6
15.5
Let us assume, for example, that the plant design is based on a raw
water turbidity of 1.9 FTU and the filters are designed to operate at
1.60 gpm/sq ft to a terminal head of 130 ft of water. When the turbidity
52
-------
PLANT OPTIMIZED AT ALL TURBIDITY LEVELS
FILTRATION RATE
20 30 40 50 60
RAW WATER TURBIDITY, FTU
70
Fig, 9. Optimum design characteristics of pressure diatom!te
filters (Table 12), Case 5.
-------
is either more or less than 1.9 FTU, the only operating change that can
be made to provide minimum operating costs under the fixed flow rate-
terminal pressure drop conditions is to adjust the body feed to a new
level which provides minimum operating costs for the fixed filter plant
and the new turbidity conditions. POPO can be used for this purpose.
For example, Table 16 shows the new computer input data for Case 5
when the plant designed to handle 1.9 FTU turbidity water actually
handles a 1.0 FTU turbidity water. The computer printout data indicated
that a body feed of 5 mg/1 will provide a 22-hr run and a filtration
cost of 5.37 c/1000 gallons. When the plant was designed to filter
1.0 FTU turbidity water, the plant operating cost was only 5.32 C/1000
gallons (Table 13) .
The results of this series of POPO runs is shown in Table 17.
When the plant design is based on a 1.0 FTU trubidity, the optimum run
lengths become shorter and filtration costs become greater as turbidity
increases. It is doubtful whether plant operators could keep up with
backwashing duties during the 3.65 days a year (Table 2) that water
turbidity is greater than 5.8 FTU and the filter runs are less than
3-4 hours long, unless the filter backwash operation were automated.
When the plant design is based on a turbidity of 1.9 FTU or 5.8
FTU (instead of on 1 FTU), the cost of filtration with lower turbidi-
ties goes up and the cost of filtration with higher turbidities goes
down. In general, plants designed on the basis of a 1.9 or 5.8 tur-
bidity would be easier to operate, since both give longer runs with
higher turbidities. However, a more detailed analysis can be made to
determine the average annual filtration costs, assuming the normal
54
-------
Table 16. Data Input to POPO, pressure diatomite filter, Case 5, tur-
bidity of 1.0 FTU when plant is designed for 1.9 FTU.
Asbestos-Turbidity Removal: Lake Superior Water, Design Based
on Turbidity of 1.9 FTU, Final Turbidity: 0.06 FTU
1 Design Flow
2 Salvage Value
3 Energy Conversion
4 Interest Rate
5 Plant Life
6 Solids (CS)
7 XI Index
8 Temperature
9 Precoat Weight
10 Precoat Density
11 Septum Diameter
12 Beta Prediction
13 Unit Flow Rate
14 Body Feed (CD)
15 Terminal Head
16 Diatomite Cost
17 First Cost
18 Power Cost
19 Labor Cost
30 MGD
30% First Cost
70%
6.68%
20 years
1.0 FTU (by turbidity)
1.77E9 ft/lb
33 °F
0.15 Ib/sq ft
19.7 Ib/cu ft
3.5 in.
8.725/l.:45800/0/0
1.60/0/0 gpm/sq ft
5/1/20 mg/1
130/0/0 ft
152 $/ton
Area, sq ft
1,000
3,472
6,944
20,832
41,667
80,000
100,000
1.627 C/KWH
Area, sq ft
1,000
3,472
6,944
20,832
41,667
80,000
100,000
$/Sq ft
$616
376
259
155
134
122
120
$/sq ft per month
$1.87
1.30
1.04
0.60
0.52
0.40
0.35
20 Backwash Cost
10,30 gal/sq ft, min
55
-------
Table 17. Final optimization of Case 5 pressure diatomite filters oper-
ated at full range of water turbidity to be encountered.
Raw
Turbidity,
FTU
1.0
1.9
5.8
20.5
57.5
1.0
1.9
5.8
20.5
57.5
1
1.9
5.8
20.5
57.5
Plant Based
Q,
gpm/sq ft
1.8
1.8
1.8
1.8
1.8
PPlant Based
1.60
1.60
1.60
1.60
1.60
Plant Based
1.10
1.10
1.10
1.10
1.10
on 1.0 FTU
HL,
ft of water
110
110
110
110
110
on 1.9 FTU
130
130
130
130
130
on 5.8 FTU
180
180
180
180
180
Turbidity
CD,
mg/1
6
9-10
20
49
104
Turbidity
5
8
16
40
84
Turbidity
3
3
10
25
50
Optimization
Cost,
C/1000 gal
5.32
5.96
8.04
13.59
24.18
Optimization
5.37
5.89
7.59
12.08
20.56
Optimization
5.82
6.13
7.16
9.87
14.92
Run Length,
hours
16.1
10.4-10.9
5.2
2.5
1.5
22
14.6
6.8
3.2
1.9
50.5
33.8
15.5
6.9
3.7
56
-------
water quality variations. During a year the plant will filter a total
of 30 million gal/day x 365 days or 10,950,000 thousand gallons per
year. The gallons filtered when the turbidity is at each level would
be:
Turbidity, FTU
0-1
1-1.9
1.9-5.8
5.8-20.5
20.5-57.5
Per cent of Time
Turbidity Experienced
90.0
5.0
4.0
0.9
0.09
Fraction
0.90
0.05
0.04
0.009
0.0009
Thousand
Gallons
9,855,000
547,500
438,000
98,550
10,950
Total water filtered - 10,950,000
When these gallons of water filtered at each turbidity level are mul-
tiplied by the unit cost of filtering water at that turbidity level
(Table 17), the total annual unit costs of filtering water in plants
optimized for operation at turbidity levels of 1.0, 1.9, and 5.8 FTU
are shown in Table 18. The results of this table indicate that the
annual unit filtration cost is essentially the same for plant designs
based on 1.0 and 1.9 FTU turbidities. Filtration costs in a plant de-
signed on the basis of a raw water turbidity of 5.8 FTU would be only
6.24 per cent higher. This would indicate that the final filtration
plant design should be based on a turbidity between 1.9 and 5.8 FTU
and the annual unit filtration cost would be between 5.56 and 5.93 C/1000
gallons.
Once a filtration plant has been built, the first cost and the
power requirements of the plant are fixed. The major plant operating
57
-------
Table 18. Total annual pressure diatomite filtration costs, Case 5, when plants are designed on raw water
turbidities of 1.0, 1.9, and 5.8 FTU.
Turbidity,
FTU
0-1.0
1.0-1.9
1.9-5.8
5.8-20.5
20.5-57.5
Total
1,000 Gallons
Filtered Yearly
9,855,000
547,000
438,000
98,550
10,950
10,950,000
Yearly
1.0
Unit Cost,
$/1000 gal
0.0532
0.0596
0.0804
0.1359
0.2418
(Table W)
0.0555
Plant Design Based on Turbidity
FTU 1.9 FTU
Yearly Cost, Unit Cost, Yearly Cost,
$ $/1000 gal $
$524,286 0.0537 $529,214
32,631 0.0589 32,248
35,215 0.0759 33,244
13,393 0.1208 11,905
2,648 0.2056 2,251
(Table 17)
$608,173 $608,862
0.0556
5.8 FTU
Unit Cost, Yearly Cost,
$/1000 gal $
0.0582 $573,561
0.0613 33,531
0.0716 31,361
0.0987 9,777
0.1492 1,634
(Table 17)
$649,814
0.0593
00
-------
costs that can escalate with the passage of time due to inflation in-
volve power costs, labor costs, and the costs of filter aid. With the
higher turbidities encountered, from 60-80 per cent of the total plant
operating costs are involved in the cost of filter aid. A study was
made, therefore, to determine what effect a 20 per cent increase in
filter aid cost would have on the total cost of filtration. The results
are summarized in Table 19. The results indicate that such an increase
in filter aid costs, a probable increase in the years ahead, would make
the plant design based on a 1.9 FTU turbidity the most economical plant
to build. The probability of further filter aid cost increases would
tend to suggest that the plant be based on a raw water turbidity of
about 2.5-3.5 FTU.
If the plant design were based on a turbidity of 1.9 FTU, the
operator would control the turbidity/body feed ratio by measuring water
turbidity at say 4-hr intervals (more frequently under rapidly changing
conditions) and could determine the optimum body feed required from
Fig. 10. The filter run lengths expected at various turbidities are
shown in Table 17. When the turbidity exceeds about 20.5 FTU (about
10 hr/yr), the runs would be less than 3.2 hr long. In such cases it
might be necessary to forego operating economy in order to obtain an
operable filter run length. Figure 11 shows how much body feed would
have to be added to obtain filter run lengths of various time periods
at raw water turbidities of 5.8, 20.5 and 57.5 FTU. An alternative
to operation with such high water turbidity would be to:
1. Provide a raw water storage basin to store low turbidity
water that could be used during storms when water turbidity
might increase for short periods,
59
-------
•pain §f pat} Apeq jo
euret
OSO*>99$ lt8 0001/^90*9 »ie3 !"«««¥
ZEt'l ^££91*0
98S'01 2^01*0
9*6'2E IISLO'O
Z8V<7C 16290*0
<702'<7iS 8S6SO*0
028'1£9$ l»i 0001/
BlE'El <7lfEl*Q
028'SC l^iO'O
^E^'EE 29190*0
19E'9^S WSO'O
Zn'9£9^ l»i U001/
O90'£ IWLl'Q
«i'si SEsro
69E'8£ 9£80'0
8£^'^E 6S90*0
8£E'S^S$ 2ESfO*0
$ '5ieo ^Bi OQ01/£
Itnuuy 'lies STUft
W80TO
ZilZO'O
80010*0
1^900 '0
'*LL'S
^0281*0
82§CO*0
ZE910'0
^OT'0'0
'*T8'S
9£OfO
JE^O'O
0861*0
zaro
»..»uif»W
?B*1 AA/
£060*0
1810*0
^800 ' 0
*?SOO*0
3 SOD }Bnuuv
nsi'o
^620*0
9E10*0
18ITO
QiiO'O
09EO*0
S910'0
IOTO'0
18UTili0 '
\i /A
!l60-0
9UO*0
E190*0
28SO*0
8021*0
6i£0*0
68§0'0
81WO
6SEVO
^080*0
96£0'0
2£SO*0
aai^i P1V 3B31T.I
B40|g| 3i@D 31Ufl
£'«-S'OZ
£f02-8'£
8*S-6*1
6*1-1
1-0 niJ B'S
£ *£S-S*02
S*02-8*S
8*S-6*1
6*1-1
i-o nu 6*1
S'OZ-8'S
8*S-6*1
6*1-1
1-0 AiLI 1
sSu^H ^31PH^nl uo
^jfPTSani *»5»tt paifg ulftaa suvi£
uo
-------
PLANT FLOW RATE = 1.6 gpn/sq ft
HEAD = 130 ft of water
J_
_L
JL
8 12 16 20 24
TURBIDITY, FTU
28
32 36
Fig. 10. Body feed required for optimum plant operation. Case 5
plant baMd on raw water tuffaidity of 1.9 FTU.
28
24
C20
J
£16
O
12
Z
8
_LLL
J.
0 20 40 80 120 160 200 240 280 320
BODY FEED, mg/l
Fig. 11. Body feed levels required to provide a given run length.
Case 5 plant optimized for handling 1.9 FTU turbidity
water.
-------
2. Provide several days of filtered water storage capacity so
that the plant flow rate could be reduced during the rare
periods of peak raw water turbidity.
Final Cost Comparisons
In order to bring the costs developed to this point in this paper to
the same comparative basis used in developing costs of direct-filtra-
tion, granular media filters for the removal of asbestiform fibers
from Lake Superior water, several design assumptions had to be adjusted.
These included the following design bases:
•The pressure filter costs were to be based on use of filters in-
corporating 1-in diameter filter septa.
•The plant operating costs were to be based on the 30 mgd capacity
plant producing an average plant output of only 20 mgd.
•The plant was to be designed for a 50-yr life with an interest
rate of 5.625 per cent. (This represented an interest rate re-
duction from the federal government interest rate estimated to
be 6.68 per cent to a tax-free municipal bond interest rate of
5.625 per cent.)
•The plant costs were to include a sludge disposal cost of $0.01
per pound of sludge produced and provide for higher operating and
maintenance costs.
In accomplishing a final cost comparison on the adjusted bases, it
was decided that the final plant design and annual operating costs
should be based on use of a pressure, diatomite filtration plant op-
timized to filter raw Lake Superior water when the plant turbidity was
62
-------
2.5 FTU. Table 20 lists the revised cost data used to determine the
per square foot cost of the final plant design. Table 21 shows the
basic information used to determine the plant first cost and the plant
operation and maintenance costs for use in the POPO computer program.
These data were then input into the computer to accomplish the following:
•To determine for Cases 1-5 the optimum design characteristics of
the final diatomite pressure filtration plant to have a capacity
of 30 mgd.
•To use the plant size determined for the 30 mgd plant to determine
the costs of operating that plant under optimum operating condi-
tions to produce 20 mgd of filtered water.
Table 22 is a summary of the costs of the filters for each case. These
data are not comparable to any previous data cited in this paper. These
costs must be increased to provide for the cost of the chemicals used
to treat the filter media and the raw water and to cover the cost of
sludge disposal. Table 23 lists the total cost of filtering 2.5 FTU
Lake Superior water through pressure diatomite filters having a 50-yr
useful life.
A more useful estimate of the annual average cost of filtering
water was obtained for Case 5. Table 24 lists the amount of water
which would have to be filtered at each turbidity level during the
year as a function of yearly plant production. Rather than assume that
all of the water filtered would be at the level in the top of the range,
as was done in Tables 18 and 19, a mean turbidity for the range was
assumed. The Case 5 pressure filter characteristics optimized for
filtering water at a turbidity of 2.5 FTU (Table 22) were then used to
63
-------
Table 20. Revised cost data3, total capital construction costs for
pressure (De Laval) diatomite filtration plant with various
filtration rates, plant capacity of 30 mgd.
Filter Flow Rate
Item
Building and
Associated Costs
Filter Equipment
Body Feed and Precoat
Slurry Feeders and
Storage Tanks
Equipment Set-up
and Assembly
Piping and Wiring
Equipment
Present Worth of 50-year
Life Replacement Items
TOTAL
-0.5
($x!06)
2.288
5.824
0.800
1.747
0.874
0.255
$ 11.788
1.0
($x!06)
1.430
2.953
0.600
0.886
0.443
0.129
$ 6.441
, gpm/sq
3.0
($x!06)
1.114
0.984
0.400
0.295
0.148
0.043
$ 2.984
ft
6.0
($x!06)
0.858
0.492
0.300
0.148
0.074
0.022
$ 1.894
Cost data supplied by Black and Veatch, Consulting Engineers.
64
-------
Table 21. Unit filter and labor costs for De Laval pressure filters and
the estimated plant labor requirements.
Plant Filtration
Capacity, Rate,
MGD gpm/sq ft
30
30
30
30
Largest commercial
Plant
Operation
Body Feed
Laboratory
Maintenance
Filter Floor
0.5
1.0
3.0
6.0
filter has 585
Plant Labor
72
1 man
1 man
1 man
4 men
Filter
Area, Cost,
sq ft $/sq ft
41,667 282.90
20,833 309.17
3,944 429.72
3,472 545.51
sq ft.
Requirements
No. of Filters
36 12
1 man 1 man
1 man 1 man
1 man 1 man
2 men 1 man
No. of
Filters3
72
36
12
6
6
1 man
1 man
% man
% man
Total Manpower
per Shift
7 men
5 men
4 men
Cost per Year $ 241,500 $ 172,500 $ 138,000
Cost per Month
per sq ft $ 0.48 $ 0.69 $ 1.66
3 men
$ 103,500
$ 2.48
65
-------
Table 22. POPO optimum design of pressure filters (1 in. diameter septa),
raw water turbidity of 2.5 FTU.
Plant Raw HL, Length
Output, Water Q, ft of CD, Cost3 of run,
Case mgd turb, FTU gpm/sq ft water mg/1 c/1000 gal hrs.
1
2
3
4
5
30
20
30
20
30
20
30
20
30
20
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.50
1.66
2.70
1.80
1.15
0.77
1.60
1.06
1.60
1.06
60
60
50
50
195
195
135
135
135
135
4
4
4
4
12
9
13
10
8
6
5.13
6.27
4.94
6.06
8.32
9.71
7.05
8.29
6.85
8.11
23.5
56.0
29.6
71.4
13.0
25.6
14.3
27.2
15.6
30.6
p
Does not include chemical (except filter aid) or the sludge disposal
costs.
66
-------
Table 23. Average cost per thousand gallons for pressure filtration of
2.5 FTU turbidity water, Cases 1-5, 50-year plant life.
Plant
Production,
Case mgd
1 30
20
2 30
20
3 30
20
4 30
20
5 30
20
Optimized
Cost
C/1000 gal
5.13
6.27
4.94
6.06
8.32
9.71
7.05
8.29
6.85
8.11
Chemical and
Sludge
Disposal Cost,
c/1000 gal
0.111
0.083
0.090
0.069
0.558
0<450
0.480
0.415
0.346
0.288
Total
Filtration Cost,
C/1000 gal
5.24
6.35
5.03
6.13
8.88
10.16
7.53
8.71
7.20
8.40
Table 24. Amount of water filtered at each turbidity level as a func-
tion of daily plant production.
Raw Water Turbidity, FTU
Range
0-1
1-1.9
1.9-5.8
5.8-20.5
20.5-57.5
Assumed
Mean
0.75
1.50
3.85
13.15
39.0
Annual Plant
at Each
1000
20 mgd
6,570,000
364,667
292,000
65,700
7,300
Production
Range,
gal
30 mgd
9,855,000
547 , 000
438,000
98,550
10,950
Total
7,300,000 10,950,000
67
-------
evaluate the per unit cost of filtering water at other assumed mean
turbidity levels representative of the ranges of turbidity to be en-
countered throughout the year. The costs for producing water at rates
of 30 mgd and 20 mgd are listed in Table 25. Since Table 24 lists the
volumes of water which must be filtered at each turbidity range in a
year and Table 25 lists the unit cost of filtering water at each tur-
bidity range, it is possible to calculate the total annual average unit
cost for pressure filtration of 30 mgd and 20 mgd of Lake Superior water
when the plant was originally designed to filter water with a turbidity
of 2.5 FTU. Tables 26 and 27 show the calculations required to deter-
mine the average annual cost of filtering Lake Superior water. The
total cost for filtering 20 mgd and 30 mgd through the 30 mgd plant
were 7.81 and 6.40 c/1000 gal, respectively.
Conclusions & Recommendations
The pilot plant studies of diatomite filtration for the removal
of asbestiform fibers from Lake Superior water indicate that the follow-
ing conclusions may be drawn.
1. Several filter operating conditions are capable of consistent-
ly producing a filtered water turbidity in the range of 0.05-
0.07 FTU and also can accomplish 95-98 per cent removal of
asbestiform fibers.
2. The filter operating conditions considered most satisfactory
for providing best filtered water quality would involve either
the use of:
68
-------
Table 25. POPO optimized plant operation for different mean levels of
turbidity expected in Lake Superior during a year.
Turbidity, FTU
0.75
1.50
3.85
13.15
39.00
0.75
1.50
3.85
13.15
39.00
Average Plant
Body Feed, mg/1
4
6
11
26
56
Average Plant
4
4
8
19
41
Production - 30 mgd
Run Length, hr
37.7
22.7
11.8
5.4
2.9
Production - 20 mgd
89.7
41.9
22.8
10.2
5.2
a
Unit Filtration
Costb, C/1000 gal
6.06
6.43
7.34
9.89
15.03
7.53
7.79
8.47
10.38
14.19
filtration rate = 1.60 gpm/sq ft, Head = 135 ft of water.
Exclusive of chemical costs and sludge disposal costs.
CFiltration rate = 1.06-gpm/sq ft, Head = 135 ft of water.
-------
Table 26. Approximate total annual pressure dlatonite filtration costs, Case 5, when plant is designed for producing
30 mgd using raw water containing 2.5 FTU of turbidity and operated at 30 mgd.
Mean
Turbidity,
FTU
0.75
1.50
3.85
13.15
39.0
Precoat Usage
Run Precoat
Length, Precoats Precoat, Per Cent Used,
hrs Per Year Ib/sq ft of Year Ib/sq ft/yr
38.2 229.3 34.4 0.900 30.96
23.2 377.6 56.6 0.050 2.83
12.3 712.2 106.8 0.040 4.272
5.9 1484.7 222.7 0.009 2.004
3.4 2576.5 386.5 0.0009 0.347
Total 40.413
Body Feed Usage
Thousand
Body Feed, Gallon* Pounds
•g/1 Per Year Per Year
4 9,855,000 328,369
6 547,000 27,339
11 438,000 40,134
26 98,550 21,344
56 10,950 5,108
10,950,000 422,294
Cost
Unit
Filtration Cost Total Cost
C/1000 gal $
6.06 597,213
6.43 35,172
7.34 32,149
9.89 9,747
15.03 1,646
Cat-Floe B Cost - '
x 10,950,000 thousand gal x $0.35 -
Al« Co., - «,.*13
Soda Ash Cost - 40.413
U.W. ., ft
i* e/t *.. 25 Ib $0.0791
x 13,546 sq ft x 453.6 lb x ^—
Sludge Disposal Cost - (40.413 -r- x 13,546 sq ft + 422,294 Ib) x 7 -
sq It ID
Total Annual Cost
10,501
2,017
2,387
9,697
$ 700,529
Annual Unit Cost, c/1000 gal - 10> 950^ thousand gal ' *-«°
-------
Table 27. Approximate total annual pressure diatouite filtration costs, Case 5, whan plant Is designed for producing
30 ogd using raw water containing 2.5 FTU of turbidity but operated at 20 mgd.
Mean
Turbidity,
FTU
0.75
1.50
3.85
13.15
39.0
Precoat Usage
Run Precoat
Length, Precoats Precoat, Percent Used,
hrs Per Year Ib/sq ft of Year Ib/sq ft/yr
90.2 97.1 14.6 0.900 13.14
42.4 206.6 31.0 0.500 1.55
23.3 376.0 56.4 0.040 2.26
10.7 818.7 122.8 0.009 1.11
5.7 1,536.8 230.5 0.0009 0.21
Total 18.27
Body Feed Usage
Thousand
Body Feed, Gallons Pounds
mg/1 Per Year Per Year
4 6,570,000 218,912
4 364,667 12,133
8 292,000 19,466
19 65,700 10,402
41 7,300 2,494
7,300,000 263,407
Cost
Unit
Filtration Cost, Total
c/1000 gal Cost, $
7.53 494,721
7.79 28,408
8.47 24,732
10.38 6,820
14.19 1,036
Cat-Floe B Cost -
x 7,300,000 thousand gal x $0.35 -
Al- Cost - 18.27 x 13,546 .„ ft x
Soda Ash Cost - 18.27 x 13,546 sq ft x
Sludge Disposal - (18.27 ~-j- x 13,545 sq ft + 263,407 lb) x — - -
sq it ID
Total Annual Cost
Annual Unit Cost, c/1000 gal
7,001
912
1,079
5,109
$ 569,818
7.81 c/gal
-------
Alum-coated Hyflo Super Gel (or equivalent grade) as body
feed and precoat or
Alum-coated C-512 filter aid (or equivalent grade) as pre-
coat plus a continuous coagulant feed to the filter in-
fluent water of Cat-Floe B polymer.
3. The analysis of filter operating data indicates that vacuum
diatomite filters are significantly more expensive in producing
low per unit cost filtered water and would be very difficult
to operate under high turbidity conditions in the raw water.
4. Pilot plant data were used in least cost design of several
alternate plants. The results suggest that a low-cost plant
designed for a 20-yr life and on the basis of the turbidity
(1.9 FTU) equalled or exceeded only 5 per cent of the time
could produce 30 mgd of filtered water at a cost of 5.56 c/1000
gal.
5. Once such a plant is constructed, the major item contributing
to increased cost of plant operation will be increases in the
cost of power, labor, and filter aid. A 20 per cent increase
in filter aid cost will increase the annual average unit cost
of water filtration to 5.77 <:/1000 gal, an increase of 3.78
per cent. This increase could be reduced by reoptlmizing the
body feed using the POPO computer program each time a signi-
ficant labor, power, or filter aid cost increase was encountered.
Further computer analysis would probably indicate that the best
protection against filter aid price increases significantly
affecting unit filtration costs would occur if the filtration
72
-------
plant were designed on the basis of a water turbidity of about
2.5-3.5 FTU.
6. A pressure diatomite filter with a 50-yr life can be optimized
to filter water with a turbidity of 2.5 FTU and operated over
an average year to produce 20 mgd or 30 mgd of low turbidity
water at an average annual unit cost of 7.81 or 6.40 c/1000
gal, respectively. The costs include the cost of chemical
and sludge disposal.
7. All optimized diatomite filtration plants would experience
short filter runs and a difficult operating experience during
periods of high water turbidity. These difficulties during
such periods can be reduced by:
•Providing raw or filtered water storage adequate to carry
the plants over peak lake water turbidity periods. A raw
or filtered water storage period of 0.5-1.5 days would be
required in the example cited in this paper,
•Abandoning the concept of economic filter aid use during
high turbidity periods and increasing the rate of body
feed to secure longer than optimum length runs during
periods of short filter cycles obtained with optimum
body feed rates.
8. The annual per unit filter operating costs cited are conserva-
tive since they are based on a constant level of turbidity
during each time period and on a constant water temperature
of 33 °F. The water turbidity is a maximum of 1.0 FTU 90 per
cent of the time, not an average of either 1.0 or 0.75, as
73
-------
used in calculating filtration costs. With warmer summertime
water, the filtration costs would be lower since longer filter
runs would occur with the better water viscosity. The opti-
mized costs cited in this paper are probably accurate to
within a + 5 per cent range. Any diatomite filtration plant
designed, built, and operated without such optimization could
be expected to experience operating costs significantly higher
than the costs developed here.
9. For maximum operating economy the availability of a computer
and the computer programs used in this study would permit
plant personnel to adjust plant operation to changes that occur
naturally with time (power, chemical, labor costs, etc.),
including evaluations of conditions that were not encountered
in the pilot-plant operations.
10. The margin for error in plant design can be significant when
extrapolation of pilot-plant data is made significantly beyond
the conditions existing during the pilot plant period. It
would be comforting, prior to final plant design, to operate
a pilot plant with raw water turbidities in the 5-20 range,
with lower body feed rates than were used (down to 4 mg/1 when
turbidity is in the range of 0.5 FTU) and with a wider range
of alum coating levels on the precoat filter aid and Cat-Floe
coagulant dosages. Such additional pilot plant experience
could provide better plant operating guidance but probably
would not significantly alter the costs determined in this
paper.
74
-------
ACKNOWLEDGMENTS
This paper is a stannary of the results from vacuum and pressure
diatomite filter pilot plants operated by Black and Veatch, Consulting
Engineers of Kansas City, Missouri, for the removal of asbestiform
fibers from Lake Superior water at Duluth, Minnesota. The work was
conducted under contract with the U.S. Environmental Protection Agency
in cooperation with the U.S. Army Corps of Engineers and the City of
Duluth. The author served as a consultant to the engineers in planning
and interpreting the pilot plant results.
The author wishes to express his appreciation to Mr. Gary Logsdon,
project officer for the U.S. Environmental Protection Agency, for his
cooperation and encouragement during the study. The personnel of Black
and Veatch, Consulting Engineers were without exception most helpful in
all aspects of the work - George Bollier who conducted the pilot plant
work At Duluth, John R. Stukenberg, Kendall M. Jacob, John Schmidt,
J. H. Robinson, and Paul Haney, who participated in facilitating the
collection, interpretation, and dissemination of the study results in
many ways.
75
-------
REFERENCES
1. Public Health Service Drinking Water Standards. U.S. Department
of Health, Education, and Welfare, Public Health Service.
Washington, D.C. Public Health Service Publication No. 956 (1962).
2. Quality Goals for Potable Water, Statement of Policy. Jour. AWWA,
60:12:1317-1322 (Dec. 1968).
3. Dillingham, James H., Cleasby, J. L., & Baumann, E. R. Diatomite
Filtration Equations for Various Septa. Jour. San. Engr. Dlv.,
Proc.. ASCE, SAI:41-55 (Feb. 1967).
4. Dillingham, James H., Cleasby, J. L., & Baumann, E. R. Prediction
of Diatomite Filter Case Resistance. Jour. San. Engr. Div..
Proc.. ASCE. SAI:57-76 (Feb. 1967).
5. Lusmer, John H. Asbestos-Containing Filter Materials. Joint
Filtration Symposium at the 78th National Meeting, AIChE, Salt
Lake City, Utah. Paper no. Ih, (Aug. 19, 1974).
6. Baumann, E. R., & Oulman, C. S. Polyelectrolyte Coatings for
Filter Media. Filtration and Separation. 7:6:682-690 (Nov./Dec. 1970)
7. Dillingham, J. H., Cleasby, J. L., & Baumann, E. R. Optimum
Design and Operation of Diatomite Filtration Plants. Jour. AWWA.
58:6:657-672 (Jun. 1966).
76
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO. 2.
EPA-670/2-75-050g
TITLE AND SUBTITLE
DIRECT FILTRATION OF LAKE SUPERIOR
WATER FOR ASBESTIFORM FIBER REMOVAL
Appendix I
AUTHOR(S)
E. Robert Baumann
(Consultant to Black § Veatch)
PERFORMING ORGANIZATION NAME AND ADDRESS
Black § Veatch, Consulting Engineers
1500 Meadow Lake Parkway
Kansas City, Missouri 64114
2. SPONSORING AGENCY NAME AND ADDRESS
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
3. RECIPIENT'S ACCESSION-NO.
5. REPORT DATE
June 1975; Issuing Date
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
10. PROGRAM ELEMENT NO.
1CB047; ROAP 21AQB; Task 024
11. CONTRACTJOeWW NO.
DACW 37-74-C-0079
IAG #EPA-IAG-D4-0388
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
5. SUPPLEMENTARY NOTES
This work conducted through interagency agreement between EPA Region V and the Corps
of Engineers, St. Paul District. See also EPA-670/2-75-050a, b, c, d, e, and f.
6. ABSTRACT
Pilot plant research conducted in 1974 at Duluth, Minnesota, demonstrated that
asbestiform fibers counts in Lake Superior water could be effectively reduced by
municipal filtration plants. During the study engineering data were also obtained
for making cost estimates for construction and operation of both granular and
diatomaceous earth (DE) filtration plants ranging in size from 0.03 to 30 mgd.
During one phase of the pilot plant investigation, the diatomite filters were
operated in a way that yielded data used for computer optimization of the DE
filtration process. The POPO (Program for Optimization of Plant Operation)
results are presented in Appendix I.
7. KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Asbestos
Amphiboles
Serpentine
Water supply
Filtration
Water treatment
Pilot plants
3. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
b. IDENTIFIERS/OPEN ENDED TERMS
Mixed media filtration
Diatomaceous earth fil-
tration
Asbestiform
Chrysotile
Fiber removal
Duluth (Minnesota)
Lake Superior
19. SECURITY CLASS (This Report)
UNCLASSIFIED
20. SECURITY CLASS (This page)
UNCLASSIFIED
c. COS AT I Field/Group
13B
21. NO. OF PAGES
81
22. PRICE
PA Form 2220-1 (9-73)
TT
U. S. GOVERNMENT PRINTING OFFICE: 1975-657-59VS'tOO Region No. 5-11
------- |