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

-------