United States
Environmental Protection
Agency
f/unieipal Environmental Research
Laboratory
Cincinnati OH 45268
EPA 600 2 79 125
August 1979
Research and Development
&EPA
Seattle Tolt Water
Supply Mixed
Asbestiform
Removal Study
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, US. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. Special Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution -sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-125
August 1979
SEATTLE TOLT WATER SUPPLY
MIXED ASBESTIFORM REMOVAL STUDY
by
Gregory J. Kirmeyer
Water Quality Division
Seattle Water Department
Seattle, Washington 98144
Grant No. 804422
Project Officer
Gary S. Logsdon
Drinking Water Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio 45268
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental Research
Laboratory, U.S. Environmental Protection Agency, and approved for publica-
tion. 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
The Environmental Protection Agency was created because of increasing
public and government concern about the dangers of pollution to the health
and welfare of the american people. Noxious air, foul water, and spoiled
land are tragic testimony to the deterioration of our natural environment.
The complexity of that environment and the interplay between its components
require a concentrated and integrated attack on the problem.
Research and development is that necessary first step in problem
solution and it involves defining the problem, measuring its impact, and
searching for solutions. The Municipal Environmental Research Laboratory
develops new and improved technology and systems for the prevention,
treatment, and management of wastewater and solid and hazardous waste
pollutant discharges from municipal and community sources, for the preser-
vation and treatment of public drinking water supplies, and to minimize
the adverse economic, social, health, and aesthetic effects of pollution.
This publication is one of the products of that research; a most vital
communications link between the researcher and the user community.
This report presents the results obtained and conclusions drawn from
pilot plant filtration research on the removal of naturally occurring
asbestiform fibers from a protected mountain water source. Appendixes not
contained in this report are available separately and present detailed
information on water quality, pilot plant equipment and operation, indi-
vidual filter runs, ambient conditions and cost.
Francis T. Mayo, Director
Municipal Environmental Research
Laboratory
111
-------
ABSTRACT
In 1975, the U.S. Environmental Protection Agency (EPA) discovered
naturally occurring asbestos fibers in the Tolt River, a 100 million gallon
per day (MGD) source of water supply for the City of Seattle. Two types of
fibers were detected amphibole and chrysotile. These fibers are in the
submicron size range and are thought to have opposite surface charges.
Although earlier research had been conducted on amphibole fiber removal from
Lake Superior, little was known about methods of removal for chrysotile. To
reduce lead time should fiber removal become necessary, the Seattle Water
Department obtained research funding from EPA to develop methods of removal
for these contaminants.
The research study gathered strong field evidence that both types of
asbestos can be removed if certain operating conditions are met. Simply
meeting the 1.0 nephelometric turbidity unit (NTU) specified by the National
Interim Primary Drinking Water Regulations is not adequate to insure fiber
removal; the American Water Works Association goal of 0.10 NTU must be met.
The pilot plant research was conducted to learn what combinations and
variations of unit processes would effectively remove amphibole and chryso
tile. Those studies confirmed that a static mixer could be used to blend
chemicals with the raw waters and that a short period of flocculation was
needed to properly condition floc particles during the winter months when
waters were cold and turbidities were highest. Direct filtration of the
conditioned water through granular media filters was effective at rates as
high as 10 gallons per minute per square foot (gpm/ft 2 ). The most effective
pretreatments developed for asbestos removal include alum at 10 milligrams per
liter (mg/l) with lime for pH control, alum at 3 mg/l along with cationic polymer
at 2 mg/l, and cationic polymer alone at 3 mg/l. Using these treatments,
amphibole counts can consistently be reduced down to the detection limit of
0.01(106) fibers/liter. Chrysotile removal is much more difficult to achieve.
However, when finished water turbidity was <0.10 NTU, 50% of the time chryso-
tile counts were reduced from an average of 7.1(106) fibers/liter in the raw
water down to not statistically significant levels r 0.02(10 6 ) fibers/liter]
in the finished water.
Removal of asbestos fibers by filtration was found to be quite sensitive
to changes in the treatment process. Turbidity spikes in the filtered water
of less than 0.35 N ru have caused increases in asbestos counts from levels
which were not statistically significant, to over 12 million fibers/liter.
Thus, vigilant control over the chemical additions and filter breakthrough
is a necessity to insure consistent asbestos removal. To enable continuous
process control, relationships between finished water asbestos counts and
iv
-------
turbidity were developed. The turbidimeter can be calibrated easily by
the treatment plant operator and can be used to indicate if asbestos is
being successfully removed. Thus, the operator does not have to wait for
days or sometimes weeks to get asbestos results from an electron microscope
and can take corrective action immediately if needed. Also, by relying on
the turbidimeter, the utility can substantially reduce the number of asbestos
analyses, which now cost about $400 per sample.
The conceptual design of the asbestos removal treatment plant for the
Tolt includes static mixers, flocculators and granular media filters; no
sedimentation basins are needed. Project costs are estimated at $25 million
and operating costs are about $1.2 million per year. The plant would be
capable of removing amphibole down to the detection limit and chrysotile
down to levels which are not statistically significant.
This report was submitted in fulfillment of Grant No. 804422 by the
Seattle Water Department under the sponsorship of the U.S. Environmental
Protection Agency. This report covers a period from May, 1976 to
November, 1978 and work was completed March, 1979.
V
-------
CONTENTS
Page
Foreword. jjj
Abstract
List of Figures V i i i
List of Tables X
Acknowledgments Xii
Sections
I Introduction 1
II Conclusions 8
III Recommendations 13
IV Equipment Design, Installation and Operation 14
V Sampling and Analyses 30
VI Results and Discussion 33
VII Design Considerations 82
VIII Operational Considerations 90
IX Cost Analysis 91
X References 95
XI Glossary 97
XII AppendixA....
A-i Results of Asbestiform Analyses on Streams
Feeding the South Fork Tolt Reservoir
A-2 Complete Chemical Analysis of South Fork
Tolt River Water Supply
A3 Efficiency Data for Selected Filter Runs
A-4 Summary of Raw and Finished Water Asbestiform Counts
XIII Index of Unattached Appendixes ill
v ii
-------
FIGURES
Number Page
1 Major facilities of the Seattle Distribution System. . 2
2 Pilot plant flow schematic . 16
3 Flow diagram for filter columns 19
4 Flow diagram for backwash wastewater treatment system. 20
5 Raw water amphibole vs. turbidity 37
6 Raw water chrysotile vs. turbidity 39
7 pH vs. alum dosage for raw Tolt water 41
8 Finished water turbidity and aluminum residual vs. pH. 42
9 Finished water turbidity vs. alum dosage 44
10 Finished water turbidity vs. Catfioc dosage 48
11 Finished water turbidity vs. 573C dosage 49
12 Headloss vs. time rung #93 & #93iir . . 52
13 UFRV vs. filter loading rate 54
14 Net water produced per 24 hours vs. the filter loading rate. 55
15 Operating data for run #l74 57
16 Operating data for run #174cc 58
17 Operating data for run #174CMM 59
18 Amphibole vs. finished water turbidity . . . . . 61
19 Chrysotile count vs. finished water turbidity. . . . 62
20 Probability plot for chrysotile when finished water
turbidity is < 0.10 NTU 63
21 Probability plot for chrysotile when finished water
turbidity is >0.10 NTU 64
22 Finished water turbidity and chrysotile vs. time - run #21 66
23 Finished water turbidity and chrysotile vs. time - run #24 67
24 Finished water chrysotile counts and turbidity vs. time
run#l20 70
viii
-------
Nuniber Page
25 Finished water turbidity and chrysotile vs. timerun #174MM 72
26 Raw and finished water turbidity vs. time run #160MM . . . 76
27 Interface settling velocity vs. solids concentration . . . . 80
28 Batch settling curve for waste solids 81
29 Preliminary process diagram 84
ix
-------
TABLES
Number Page
1 Dimensions of unit processes 17
2 Motor description 21
3 Power factor information 22
4 Field electrical data on mixer motor 22
5 Mixing information on 2-inch Kenics static mixer 24
6 Mixing intensities for fiocculator . . . 26
7 Characteristics of media tested 27
8 Chemicals tested 29
9 Raw water quality characteristics. . . 33
10 Raw water asbestos counts 35
11 Estimated lime dosages [ mg/i Ca(OH) 2 J. 40
12 Alum dosage vs. operating parameters . 43
13 Preferred chemical treatments alum 45
14 Alum and lime dosage vs. cationic polymer dosage 45
15 Preferred chemical treatments alum plus cationic polymer 46
16 Catfioc Tl dosages and turbidity 46
17 573C dosages and turbidity 46
18 Comparison of one with three static mixers . . . . 47
19 Results from filter runs with arid without flocculation . . 50
20 Linear regression equations for the various filter medias. 53
21 Chrysotile results from various filter medias 56
22 Results from run #11 .. . . . . 65
23 Results from runs #21 and #24 68
24 Results from run #53 68
25 Results from run #62 69
26 Resultsfromrun#120 69
x
-------
Number Page
27 Results from run #174MM 71
28 Comparison of Millipore and Nuclepore analysis methods . . . 73
29 Organic removal data 75
30 Trihalomethane results on unfiltered and filtered Tolt water 77
31 Filtration system design criteria 83
32 Characteristics of static mixers 85
33 Determining the filter area 86
34 Resulting filter loading rates 86
35 Suggested guidelines for granular media filters 87
36 Backwash wastewater characteristics 88
37 Clarifier/Thickener area requirements based on batch flux
curves 88
38 Plant construction costs 92
39 Annual plant operating costs 93
xi
-------
ACKNOWLEDGMENTS
The execution of this study involved the team approach in which
representatives from health and regulatory agencies, consultants, analysts
and manufacturers met and/or conversed via conference calls periodically
to give advice and direction to the Citys project officer. The cooperation
and assistance of all participants is gratefully acknowledged.
A special thanks is due to Mr. Roy R. Jones, Physical Scientist,
Region X, EPA, for helping the City during the grant application phase,
to Dr. Edwin Boatman, Electron Microscopist, University of Washington, for
his timely and accurate asbestos analysis, and to Mr. Richard R. Harbert,
Mechanical Engineer, CH2M/Hill Consulting Engineers*, for his help in design
of the pilot plant and evaluation of plant data.
The cooperation and assistance provided by the following personnel is
also gratefully acknowledged.
Mr. William A. Mullen Water Supply Division
Mr. Steve Has elman Region X, USEPA
Seattle, Washington
Mr. A. R. Galiler Laboratory Branch
Mr. R. H. Riech Region X, USEPA
Seattle, Washington
Mr. Kenneth Merry Water Supply and Waste Section
Department of Social and Health Services
Olympia, Washington
Ms. Marilyn Mah Electron Microscopist
University of Washington
Seattle, Washington
Mr. Ric Tower Treatment Equipment Company
Lake Oswego, Oregon
Mr. William Owens Neptune Microfloc, Inc.
Dr. Andrew Hsiung Corvallis, Oregon
Mr. Leland E. Montgomery
*Mr. Harbert is now employed with RH2 Engineering, Kirkland, Washington.
xii
-------
Mr. Lee Miller HIAC Instruments Division
Mr. Gerald C. West Pacific Scientific
Montclair, California
Mr. Ken Olson Pacific Scientific
Seattle, Washington
Mr. J. N. McKee Turbitrol Company
Mr. E. R. Carter Atlanta, Georgia
Mr. Norm Ward CH2N/Hill Consulting Engineers
Mr. Collier Martin Bellevue, Washington
Mr. Carl Hamann CH2M/Hill Consulting Engineers
Corvallis, Oregon
Dr. Cliff G. Thompson Black, Crow and Eidsness, Inc.
Montgomery, Alabama
Mr. Dominic Zicardi American Cyanaxnid Company
Ren ton, Washington
Mr. John W. McCurdy Calgon Corporation
Portland, Oregon
Mr. Newell Dimen Koch Engineering Company, Inc.
Wichita, Kansas
Arrangements for performance of this research were made through the
Water Supply Division, Region X, U.S. Environmental Protection Agency,
, Washington.
xiii
-------
SECTION I
INTRODUCTION
BACKGROUND INFORMATION
Seattle Service Area
The Seattle Water Departments service area includes Seattle proper arid
also most of the metropolitan area. It extends from Edmonds on the north to
just south of Des Moines and easterly from Puget Sound to Lake Sainmamish at
the foot of the Cascade range. Approximately 992,000 people are served
directly or indirectly by the Seattle Water Department, which is about 83%
of the population of King County. Average total water consumption is 161 MGD
and summer peak day use has reached 350 MGD. Approximately twothirds of
this water is sold directly to the Seattle Water Department retail customers
and the remaining portion is delivered at wholesale rates to suburban water
districts and municipalities, who in turn sell it to their direct service
customers.
Water Sources
The primary sources of Seattle water are the Cedar and Tolt rivers on
the western slope of the Cascade Mountains in eastern King County. These two
rivers furnish high quality mountain water of unusual clarity. The Cedar
River system, which has been in operation since 1901, serves the southern
region of Seattle and parts of South King County; the Tolt, which was added
to the City system in 1964, generally serves the area north of the Ship Canal
and the water purveyors on the east side of Lake Washington. The two sources
feed the distribution system from opposite ends, a feature which affords
excellent hydraulic characteristics. Figure 1 illustrates the major facili
ties of the system.
Water Treatment
Currently, the treatment of both water supplies consists of intake
screening, chlorination and fluoridation, Prior to distribution, the water
is rechiorinated as it leaves the open intown reservoirs to assure the free
chlorine residual necessary to maintain adequate levels of disinfection.
Compared to most other major water utilities in this country, the present
treatment is minimal. Many are required to extensively treat their water
with coagulation and filtration techniques to render it safe to drink. The
1
-------
_______ nohomish_Co._______
- King Co. ______
Transrni
ssion Lines
S. W. D.
Service
rea
!atershed Boundary
Landsburg
Diversion
!atershed
Ki tti tas
Co.
ities f the Seattle Distri
bution System.
-------
Water Departments emphasis on watershed protection, coupled with the large
storage reservoir system to meet demands during periods of high turbidity,
has enabled Seattle to meet National Interim Primary Drinking Water Requlations
without major modifications to the existing facilities.
Existing Facilities on the Tolt
Water originates in a 13,390 acre area watershed, which is narrow and
steep and has been prone to slide activity. Ownership in the watershed is
about equally divided between Weyerhaeuser Company, the Forest Service and
the Water Department. Ten major streams feed the South Fork Tolt Reservoir,
a 56,000 acrefeet, manmade storage reservoir with a maximum pool elevation
at 1760 ft. The water enters a 33inch diameter, 4.8mile pipeline through
one of three intakes, which are located at different depths in the reservoir.
Upon discharge from the pipeline, water enters the pressure Regulating Basin,
a 88 acrefeet reservoir at an elevation of 760 ft. The purpose of the pipe-
line and the Regulating Basin is to dissipate 1000 ft of head so that the
water is at an acceptable pressure prior to chlorination. Water then enters
a 66inch diameter transmission line where it is chlorinated and fluoridated
and subsequently discharged into the Seattle distribution system. The firm
yield flow on the Tolt is 60 MGD with a design hydraulic capacity of 100 MGD.
It is critical that water enter the transmission line at an elevation of
760 ft; otherwise, the necessary peak flow of 100 MGD would be unattainable.
There is a proposal at this time by Seattle City Light Department to con-
struct a second pipeline parallel to the existing line between the South Fork
Tolt Reservoir and the Regulating Basin to enable generation of power rather
than dissipating the excess head. Presently, there is a small turbine which
supplies power to the existing treatment station located on the pipeline just
upstream from the Regulating Basin.
Internal Corrosion Problem
The Seattle Water Department has recognized the existence of a corrosion
problem for some time. In 1964, when the Tolt supply was brought into full
operation, customer complaints concerning rusty water and fixture staining
in that service area had doubled in frequency within a short period of time.
Consequently, the Tolt supply was considered much more aggressive than the
older Cedar source.
The corrosiveness of the water is a result of several related factors:
1. The dissolved oxygen concentration is very high; often at saturated
conditions due to the low water temperature, the turbulence in the
system and the lack of oxygenconsuming material in the water.
2. The water is very soft and has a low pH. The hardness is only 9.0
mg/l as CaCO3 and after chlorination with gaseous chlorine and
fluoridation with fluosilicic acid, the pH of the Tolt water is
in the range of 5.86.2.
3. There is insufficient natural calcium and bicarbonate alkalinity in
the water to form protective calcium carbonate scale on the pipe sur-
faces.
3
-------
4. There is a relatively high halogen/alkalinity ratio, 1.5-2.5, which
causes areas of low pH to develop and results in conditions favorable
to pitting type corrosion which is prevalent in much of the old
galvanized pipe.
A recently completed study 2 has recommended a comprehensive corrosion
control plan with two features a water treatment program and a material
selection program. The treatment program will raise the pH and alkalinity
of both supplies and possibly supplement silicates on the Tolt supply.
Treatment facilities should be on line in 1981. The material selection
program will help the customers choose corrosion resistant materials and
will initiate plumbing code changes to allow the expanded use of materials
such as plastic pipe. Since filtration techniques can increase the corro
sivity of a water supply, a treatment plant for turbidity or asbestos
removal must include provisions for addition of corrosion control chemicals.
Discovery of the Asbestos Fibers
In January of 1975, the U.S. Environmental Protection Ageny (EPA) was
asked to determine if a water system utilizing asbestos cement (A/c) pipe
could be located and sampled to supplement information for a loop study which
was being conducted at the Municipal Environmental Research Laboratory in
Cincinnati. A water district served by the Seattle Tolt supply was contacted
and they agreed to participate in the sampling program. Samples were gathered
from two locations within the water districts distribution system and at a
control point from the Tolt transmission pipeline prior to any exposure to
the A/C pipe. Analysis indicated that there were no statistical differences
between fiber concentrations found in samples collected from the water dis-
trict and those from the Tolt pipeline. Followup sampling confirmed the
presence of both amphibole and chrysotile fibers in the Tolt transmission
line and in the South Fork Tolt Reservoir. Samples from ten major streams
which feed the supply reservoir were gathered and analyzed to determine if
a point source existed, results of these analyses are contained in Appendix A.
All of the feeder streams contained asbestos fibers in varying concentrations.
Since the Tolt supply serves up to 100 MGD to a large population of the Pacific
Northwest, the Seattle Water Department applied for and was awarded a $150,000
EPA research grant to study methods of fiber removal to reduce lead time should
a treatment plant become necessary.
Related Asbestos Removal Study
In June, 1973, the EPA reported the discovery of a large amount of
asbestos in Lake Superior the source of drinking water for Duluth, Minnesota.
Subsequent electron microscope analysis revealed up to one billion amphibole
fibers/liter. A Federal Court ruling indicated that these fibers resulted
from a mining company which had been dumping talconite tailing waste directly
into the lake at its iron ore processing plant in Silver Bay, 50 miles north-
east of Duluth. The discovery of these fibers in the lake and in the Duluth
tap water prompted many environmental and epideiniological studies. Methods
to remove these fibers were developed in a joint study conducted by Black and
Veatch Consulting Engineers, the EPA and the Corps of Engineers. 3 Today, a
water treatment facility is in operation and is removing amphibole fibers from
the Duluth drinking water.
4
-------
Health Implications
Occupational studies indicate that a human health hazard may exist for
ingested asbestos since the death rates due to digestive system cancer are
elevated in asbestos workers. This finding may be related to the swallowing
of fibers that were inhaled and cleared from the respiratory system via the
respiratory clearance mechanism. 4 One investigator has found the incidence
of gastrointestinal (GI) cancer in asbestos workers to be 23 times the level
in the general population; other researchers have also found significant
increases in GI cancer. 4 The effect is more evident after 20 years or more
from first exposure. The doseresponse relationship between occupational
exposure to asbestos and the risk of GI cancer is poorly defined. Therefore,
it is difficult to evaluate the potential health hazard of low levels of
ingested asbestos.
It appears that the physical and fibrous properties of asbestos, rather
than its chemical composition or presence of trace elements, determines its
carcinogenity. 4 Studies involving autopsies of asbestos workers are generally
inconclusive but do suggest that asbestos may penetrate digestive system
tissue. A recent study in Minnesota indicates that asbestos fibers were
found in the urine of people drinking water containing fibers. Studies
involving the feeding of asbestos to animals are generally inconclusive 4
mainly because insufficient numbers of test and control animals were used.
However, there is again some evidence that asbestos may penetrate digestive
system tissue and migrate to other locations in the body. 4
There have been at least four epidemiological studies conducted to
determine if ingestion of asbestos fibers is a possible health hazard. Three
of the studies, which were conducted in Duluth, ,5 Quebec, Canada 6 and in
Connecticut, 7 indicated no correlation between fiber count and cancer inci-
dence. A fourth study 8 conducted in the San Francisco Bay Area is the first
study to suggest a link between asbestos ingestion and cancer incidence. EPA
indicates that results from the San Francisco study are not definitive enough
to justify extensive modifications in water supply treatment and distribution
and that additional studies should indicate more clearly whether there is a
need for changes in drinking water practices. 9
An epidemiological study is presently being conducted locally by the
Fred Hutchinson Cancer Research Center in the Puget Sound Area. It will
compare cancer rates of persons ingesting various quantities of asbestos
fibers from drinking water in Everett, Seattle, Tacoma and an area on the
Kitsap Peninsula where industrial exposure to asbestos was known to exist.
In summary, the potential hazards associated with asbestos ingestion
through drinking water are not well defined at this time. The National
Academy of Sciences indicates that no acute hazard exists but they do suggest
minimizing exposure through effective coagulation and filtration of water
containing asbestos. 1 -° EPA has set no standard for asbestos in drinking
water.
5
-------
STUDY OBJECTIVES
The formal objective of this research effort was to determine the most
feasible method for reducing asbestiforin count in the Tolt water. The scope
of work included:
1. Developing methods to improve chrysotile removal by the use of
threestage mixing in combination with three conditioning
chemicals in granular media filtration;
2. Optimizing asbestos removal using anionic, cationic and nonionic
polymers;
3. Conducting filtration experiments at 5-6 gpm/ft 2 with granular
media filters;
4. Determining the effect of mixing intensity on filtration and
comparing back mixing with inline mixing techniques;
5. Developing an operating tool which indicates quickly and
economically if asbestos removal is occurring;
6. Collecting design data and developing cost information; and
7. Confirming amphibole results gathered during the Duluth asbestos
removal study. 3
Once the grant was approved, a literature review of the properties of
asbestos fibers and methods for their removal was conducted. CH2M/Hill
Consulting Engineers was awarded a contract to help with the collection of
design data aid cost information and the University of Washington agreed
to perform the asbestos analysis. A study plan was developed, bench scale
filtration studies were conducted and pilot testing was then undertaken.
SCOPE OF THE STUDY
During the period from January 1977 through September 1978, a small
scale water filtration plant was operated at the Regulating Building on the
Tolt supply. The primary goal of the pilot filtration study was to investi-
gate various filtration systems to remove amphibole and chrysotile asbestos
from the raw Tolt water. Inline, direct and conventional filtralicn processes
were investigated. Various mixing intensities were tested with different
chemical coagulants and dosages. Several granular medias were compared at
filter loading rates up to 10 gpm/ft 2 .
The study was conducted using a team approach where the consultant,
the EPA, the Washington State Department of Social and Health Services, and
the University of Washington, met or conversed via conference calls several
times during the pilot study. This approach enabled the principal investi-
gator to modify the study as needed and insured that all necessary information
had been collected by the end of the pilot study.
6
-------
ELEMENTS OF THE STUDY
The study consisted of the following six primary elements that included
various secondary topics.
Water Quality
1. Seasonal Variation of Water Quality Parameters.
2. Relationship of Raw Water Turbidity with Asbestiform Counts.
3. Asbestos Size Distribution.
Methods of Analysis
1. Comparison of Asbestos Analysis Methods.
2. Particle Counting Techniques.
3. Development of Turbidimeter as Operational Tool.
Pretreatment for Filtration
1. Effectiveness of Alum, Ferric Salts and Polymers.
2. Mixing Requirements Including Flocculation.
Filtration Process
1. Comparison of Granular Media Filters.
2. Direct Filtration Techniques.
3. Filter Loading Rates Up to 10 gpm/ft 2 .
4. Rate of Headloss BuildUp.
5. Air Binding Problems.
Backwash Water
1. Sludge Production.
2. Sludge Settling Characteristics.
3. Sludge Treatment Train.
Engineering Analysis
1. Preferred Techniques for Asbestos Removal.
2. Development of Design and Operating Criteria.
3. Cost Estimates.
7
-------
SECTION II
CONCLUSIONS
RAW WATER
Asbestos
Two types of naturally occurring asbestos fibers contaminate the Seattle
Tolt water supply amphibole and chrysotile. Amphibole fiber counts averaged
1.6(106) fibers/liter and ranged from<0.04(10 6 ) up to 5.7(106) fibers/liter;
chrysotile fiber counts averaged 7.1(106) fibers/liter andranged from l.2a0 6 ).
fibers/liter up to 25.8(106) fibers/liter.
Turbidity
The turbidity of the raw water fluctuates from about 0.10 nephelometric
turbidity unit (NTU) up to about 5 NTU throughout the year depending upon the
season. Asbestos counts are positively correlated with turbidity.
FINISHED WATER
Turbidity Removal
Finished water turbidities can consistently be reduced down to <0.10 NTU
with granular media filtration if proper pretreatment is applied.
Chemical Dosages
The most effective pretreatment combinations are as follows:
Chemical Dosage
Chemical Dosage
Chemical Dosage
Aluminum Sulfate 710 mg/i
(Alum)
Alum 3-5 mg/l
Cationic 3 mg/i
Polymer
Lime [ Ca (OH) 211 14 mg/l
(pH range = 6.16.7)
Cationic Polymer 2 mg/l
Filter Aid 0.10.3
(Nonionic or mg/l
anionic poly
mer)
Filter Aid 0.10.3
(Nonionic or mg/i
anionic polymer
Filter Aid 0.020.25
(Nonionic or mg/i
anionic polymer)
mg/i = Milligrams per liter
pH control is very critical when using alum and pH must be maintained
between about 6.1 and 6.7 units for effective destabilization to occur.
8
-------
Preferred Treatment Train
The preferred treatment train includes static mixers, flocculators and
granular media filters; no sedimentation basins are needed. Direct fil-
tration techniques are capable of meeting both process and water quality
goals.
Static mixers functioned as well as mechanical mixers in dispersing
chemicals into the process stream and they use less energy and require
less maintenance.
A flocculator must be included in the process train to meet both water
quality and process goals at low water temperatures and turbidities which
exceed 1.5 NTh.
Filtration Considerations
All filter medias which were tested met both water quality and process
goals.
As the filter loading rate increased, the unit filter run volume (filter
efficiency) dropped and net water produced per 24 hours increased.
Filter loading rates up to 10 gallons per minutes per square foot (gpm/ft 3
were found to be effective at reducing finished water turbidity down to 0.10
NTU and at reducing asbestos fiber concentrations down to not detectable (ND)
or not statistically significant (NSS) levels.
When filter loading rates were between 5.5 gpm/ft 2 and 7.5 gpm/ft 2 ,
there was little difference noted in the water production efficiencies among
the various filter medias tested. At loading rates less than 5.5 gpm/ft 2 ,
the dual media with a coarse coal was most efficient and at rates greater
than 7.5 gpm/ft 2 , the mixed media demonstrated a higher efficiency.
The air binding problems, which were encountered with the cold, oxygen
saturated water during pilot testing, can be eliminated in the fullscale
plant by increasing the depth of water over the top of the filter surface
thereby maintaining a positive head throughout the filter media.
Asbestos Removals
Amphibole --
When amphiboles were detected in the raw water and when finished water
turbidity was < 0.10 NTU, 52 out of 57 asbestos results (91%) had levels of
amphibole which were at or less than the detection limit of 0.01(106) fibers!
liter. In all 5 cases where .amphiboles were greater than 0.01(106) fibers!
liter, 4 or fewer fibers were counted, indicating that results were NSS.
Turbidity spikes and poor destabilization were consistently associated
with high amphibole counts in the finished water.
9
-------
Chrysotile
Chrysotile fibers are more difficult to remove than amphibole; neverthe-
less, when finished water turbidities were < 0.10 NTU, 50% of the time,
chrysotile counts were < 0.02(106) fibers/liter. When turbidities were
>0.10 NTU, 50% of the time, chrysotile was < 0.27(106). This indicates
that there was ten times more asbestos present in the finished water when
turbidity was>0.10 NTU than when turbidity was < 0.10 NTU.
Turbidity spikes, filter breakthrough and poor destabilization were
consistently associated with very high finished water chrysotile counts.
Removal Techniques
Direct filtration techniques using standard water treatment chemicals
can consistently remove amphibole down to ND levels and chrysotile down to
NSS levels if certain operating conditions are met.
Analysis Procedures
Both Millipore and Nuclepore asbestos analysis methods yielded virtually
the same results on raw and finished water samples.
Trihalomethane Reduction
Levels of trihalomethanes in the filtered water were on the order of
6070% less than in the unfiltered water.
High Turbidity Removal
Direct filtration techniques can be used to reduce raw water turbidities
of about 20 NTU down to levels of 0.5 NTU.
BACKWASH CONSIDERATIONS
Volume
Backwash water to clean the filters will be drawn from a 5 million
gallon (MG) finished water storage reservoir and will comprise an estimated
3.1% of the plant production water at nominal flow rates.
Treatment
Treatment for the backwash wastewater must be provided to meet the
state pollution control requirements. Due to the high quality water of
the receiving streams in the area, complete recycle of the liquid waste
and land disposal of the solids will most likely be required.
DESIGN CONSIDERATIONS
Treatment Train
The preferred treatment train includes an energy dissipating system;
three static mixers; four flocculation chambers, each with three stages of
10
-------
mixing; eight granular media filters; a backwash wastewater treatment system;
and a finished water storage clearwell with provisions for addition of
corrosion control chemicals.
Filtration System
Either dual or mixed media filters will meet both process and water
quality goals.
The suggested nominal filter loading rate is 4.2 gpm/ft 2 at 60 million
gallons per day (MGD) and 6.9 gpm/ft 2 at 100 MGD.
Rates up to 10 gpm/ft 2 can produce water with turbidities of < 0.10 NTU
and can consistently remove asbestos fibers down to NSS levels; however,
filter production efficiency is significantly decreased at the higher filter
loading rates.
Deep filter boxes are needed to prevent air binding problems from reduc-
ing filter efficiency and creating problems during backwashing.
Water Quality
The preferred treatment train and filtration system will meet the National
Interim Primary Drinking Water Regulations, 1 will remove asbestos, turbidity,
color, trihalomethane precursors, aluminum,will fulfill state guidelines
regarding direct filtration plants and will meet state pollution control
requirements.
Monitoring
Each filter will be equipped with continuously recording turbidimeters
and the proposed plant should include a gravity flow pilot filter identical
to the fullscale plant along with a coagulant control center.
Construction Materials
Asbestosbearing products should be avoided when selecting construction
materials.
OPERATING PARAMETERS
Water Quality
To insure asbestos removal, the finished water turbidity should be main-
tained at or preferably below 0.10 NTU and backwash procedures should be
initiated at the first sign of turbidity breakthrough.
Head loss
The plant should be operated to a terminal headloss of 10 feet (ft).
11
-------
COST
The estimated project costs are $24,747,000 with annual operation and
maintenance costs of $1,212,000. This equates to an annual cost of $3,019,000.
12
-------
SECTION III
RECOMMENDAT IONS
This section of the report makes specific recommendations concerning
additional research needs and further evaluations in the area of asbestiform
removal and filtration capabilities.
Analyze the Seattle asbestiform count data to determine the size dis-
tribution of asbestos particles in the raw and filtered water.
Determine if high rate, coarse media, deep bed filtration is effective
at removing asbestos fibers.
Compare a continuous recording turbidimeter with a continuous recording
particle counter on a filter effluent to determine if there are any advantages
to using a particle counter as an operational tool over the simple turbidi
meter.
Compare the effluent from a declining rate versus a constant rate filter
to determine the advantages and disadvantages of each.
Determine what problems may arise from an overdose of a filter aid.
Determine why filtration of a surface water can cause it to be more
corrosive and what measures can be taken to reduce the resulting problems.
Develop a standard procedure for the asbestiform analysis.
Publish a state of the art manual on removal of asbestiform particles
from the water to enable water utilities to modify either plant facilities
or operations to remove the fibers.
13
-------
SECTION IV
EQUIPMENT DESIGN, INSTALLATION AND OPERATION
GENERAL
The pilot study was conducted on a 20 gallon per minute (gpm) Waterboy
(WB27) package plant, which is manufactured by Neptune Nicrofloc, Inc. The
unit was installed in the chlorine room at the Tolt Regulating Basin. Since
the unit was designed to operate as a conventional filtration plant, modi-
fications were made to the basic unit to enable several different treatment
trains to be tested. Because of a serious air binding problem which developed
in the filter media due to insufficient water over the top of the filter,
three additional filter columns were constructed and operated in parallel
with the Waterboy filter. Photographs of the pilot plant apparatus are
presented in Appendix B.
RAW WATER SOURCE
Water was supplied to the pilot unit by tapping the injector water supply
for the chlorinators and running a 2inch supply line to the pilot unit.
Injector water is pumped from the forebay of the Regulating Basin up to an
elevated storage tank and then it flows by gravity to the injectors. This
is the same water which is supplied directly to the customers of the Seattle
Water Department except that the water had not been fluoridated or chlorinated
at that point.
PERSONNEL
The Tolt Regulating Building is manned 24 hours per day, 7 days per week,
by a trained operator. The operators, who are employees of the Seattle Water
Department, were given all drawings and manuals for the pilot unit. They
became quite familiar with the operation of the unit and gathered all necessary
operating data such as headloss, turbidity, flow rate, chemical feed rates, etc.
The study was conducted in three phases. First, there was an initial
breakin period where all personnel became familiar with the operation of
the pilot apparatus. Next was a lengthy intensive testing period during
which numerous different treatment trains and dosages were evaluated to deter-
mine the feasible asbestos removal methods. Lastly, there was an optimal
design phase in which filter loading rates up to 10 gpm/ft 2 were tested and
14
-------
special studies were made involving excessive raw water turbidities and
sludge treatment techniques.
AMBIENT CONDITIONS
To determine what relationship existed, if any, between ambient condi-
tions in the watersheds and raw water quality, and to document conditions
during the study, several readings including the water level in the South
Fork Tolt Reservoir, the turbidity, the precipitation, valve operations,
water use, the wind direction and velocity, and the temperature were collected
These are contained in Appendix B.
PILOT PLANT
Waterboy
The unit processes on the package plant included a hydraulic rapid mix
chamber, a flocculation basin with picket flocculators, tube settlers and a
granular media filter. Media installation and initial startupwere supervised
by a representative of Neptune Microfloc, Inc. The plant is rated at 20 gprn
and can be operated either on a manual or automatic basis.
Modifications
Since the unit was to be used as a pilot plant, Neptune Microfloc, Inc.,
provided a variable speed flocculator so that various mixing intensities could
be investigated. Several other modifications were made on the unit by the
Seattle Water Department. Additional piping was installed to enable any unit
process to be bypassed or eliminated simply by operating a valve or removing
a pipe or fire hose. Figure 2 illustrates the possible flow schematics and
Table 1 lists pertinent dimensions.
Three 2inch static mixers were supplied by EPA and they were installed
with the unit. Mechanical mixers and 55gallon drums were also part of the
flexible treatment train. Water would flow under pressure into the first
drum and then could discharge by gravity into a second drum or into any other
unit process in the treatment train.
A flow controller, which would maintain a preset flow rate regardless of
incoming line pressure, was installed on the raw water pipe and an effluent
meter was placed on the finished water line. These devices provided assurance
that a constant flow rate entered the plant and was discharged from the filter.
Two turbidimeters were installed, one on the raw water and one on the
finished water. These meters fed information to a strip chart recorder to
insure that a permanent record of turbidity was kept. Bubble traps were later
installed on each turbidimeter to prevent air bubbles from giving false tur-
bidity readings.
Filtered water was discharged to an 8,000gallon storage tank with an
overflow near the top of the tank. Excess water discharged to a small stream
15
-------
MEC!-!AMICAL BACK MIXERS
TO LI
R A 14
WATER
CHEMICAL
A B C ,D,ADDITjON
E,F,G,H POINTS
STATIC MIXERS
H
o I
ISHED
AT ER
TO
STORAGE
TA N K
FIGURE 3
Figure 2. Pilot plant flow schematic.
-------
TART..E 1. DIMENSIONS OF UNIT PROCESSES
Finished
Water Storage
Tank
Depth
(cm)
Volume
Units as Indicatec
Unit Process
Length (cm)
Width or Diameter
(cm)
Detention Time*
I-
- 1
Hydraulic
Rapid Mixing
Chamber
62.9 cm
243/4 in
27.3 cm
103/4 in
165.1 cm
65 in
.028 m 3
3
10 ft
4.7 mm
Mechanical
Back Mixing
Chamber
N/A
58.4 cm
23 in
7.1 cm
28 in
0.19 m 3
6.7 ft 3
3.1 mm
Static Mixing
Pipe
45.7 cm
.
18 in
5.1 cm
2 in
N/A 933 cm 3
3
0.033 ft
0.92 sec
Flocculation
Chamber
Backwash
Settling
Chamber
62.9 cm
243/4 in
6.1 m
20 ft
57.1 cm
221/2 in
165.1 cm 0.59 m 3
3
65 in 20.9 ft
0.53 m 3.88 m 3
1.75 ft 140 ft 3
9.8 mm
N/A
1.2 m
4 ft
N/A
3.1 m
10 ft
Theoretica1 detention time
cm = centimeter
cm 3 = cubic centimeter
m = meter
m 3 = cubic meter
in inch
2.l3m
7ft
at 60.5 1/mm (16 gpm).
N/A = not applicable
gal = gallon
ft3 cubic foot
mm = minute
sec = second
16.1 m 3
549.5 ft 3
N/A
-------
adjacent to the Regulating Building and stored water was used for backwashing
the Waterboy filter.
Early in the pilot studies, air binding in the Waterboy filter became a
serious problem. Apparently, the pressure in the filter would drop below one
atmosphere and dissolved gases, which were present in the raw water, would
come out of solution and plug the voids in the filter media. This caused two
problens: (1) rapid buildup of headloss as the filter run progressed, and
(2) loss of anthracite media during filter backwash because of air bubble
release at the start of backwash.
Since headloss was rising so rapidly, it became evident that filter
design data could not be gathered from the Waterboy filter. To collect rep-
resentative design data, three 4Όinch diameter filter columns were installed
in the basement of the Regulating Building (see Figure 3). A small stream
of water was diverted from the chamber above the Waterboy media to the filter
columns to maintain a column of water about 10 ft high over the top of the
filter media. Maintaining a positive head throughout the column filter media
held the gases in solution and prevented the air binding problem. Having
three columns in parallel also allowed comparisons to be made among three
different types of filter media at one time.
The loss of media during backwash of the Waterboy filter due to entrapped
air was minimized by modifying backwashing procedures. The anthracite coal
was stirred mechanically to release as much air as possible before backwash
water was brought into the filter underdrain system. Electronic modifications
also gave the plant operator complete control over the amount of backwash
water allowed to the unit at any time. Water was brought into the filter
chamber very slowly to prevent anthracite coal from being carried out with
the backwash water.
Backwash Wastewater
Backwash wastewater from the Waterboy was discharged to a weir box and
then into a plywood settling basin (see Figure 4). Emergency overflows from
the settling basin were discharged to an earthen lagoon. Water was normally
held in the settling basin until the solids had settled and then the clarified
water was discharged to the North Fork Tolt River. Solids were periodically
removed from the basin and tests were performed to determine treatment tech-
niques for the sludge.
Mixing Intensities
Mixing intensities for each unit process were estimated using both field
data and empirical calculations. A discussion of this information follows.
Hydraulic Rapid Mixing Chamber --
This chamber is part of the WB27 pilot filtration unit. Water enters
the bottom of the chamber under pressure and flows upward spilling over a
rectangular weir into the flocculation chamber. Mixing intensity is estimated
using the following formula:
18
-------
FROM
FIGURE
I-
GRU .2_ FLOO
BASEMENT
DIVERSION
CHEMICALLY TREATED
UNFILTERED WATER
GRANULAR
COLUMNS
FINISHED WATER
FINISHED WATER
F-igure 3. Flow diagram for filter columns.
-------
BACKWASH WATER
FINISHED
FLOC WATERBOY
WATER TO RECEIVING
CHAMBER GRANULAR
STORAGE STREAM
MEDIA
TAN K
FILTER
BACKWASH WASTEWATER 4
EMERGENCY
LAGO .II.II...j
EARTHEN
EMERGENCY
OVERFLOW
PLYWOOD
WEIR BACKWASH
WASTEWATER SETTLED WASTEWATER
BOX SETTLING TO STORM DRAIN
CHAMBER
Figure 4. Flow diagram for t ackwash wastewater treatment system.
-------
/ 62.4 (H )
G=V T,
where G = Velocity Gradient in seconds 1 (sec
H = Headloss in ft
( lb sec
,Ct Viscosity in pounds seconds per square foot ft 2 I
T = Time in seconds (sec)
Assumptions/Data:
H = 5.4 ft
A at 8.5°C = 2.883(lO ) ( ibsec )
T = 282 sec from Table 1
/ 62.4(5.4 ft ) 1
= 203 sec
G = V 2.883(1O ) ; lb sec 1 282 sec
This also yields a GT value = 57246.
Mechanical Back Mixing Chambers -
Two 55gallon drums with mechanical mixers were used as back mixing
chambers. Four baffles were placed in each drum to insure that a vortex was
not formed. Information found on the mixer motor name plate or gathered from
the manufacturer is presented below.
TABLE
2.
MOTOR
DESCRIPTION
HP
1/3
Phase = 1
Revolutions
per
Minute
1725 mpere = 6.2
Volts
=
115
Hertz = 60
21
-------
TABLE 3. POWER FACTOR INFORMATION
Loading
J\rnperes
Power Factor
Efficiency
Full Load
6.2
0.63
54%
3/4 Load
5.9
0.56
59%
1/2 Load
5.6
0.47
42%
OLoad
5.0
-
-
The following data was gathered by varying the number of propellers on
the motor shaft and then measuring the voltage and current drawn.
TABLE 4. FIELD ELECTRICAL DATA ON MIXER MOTOR
No. of Propellers Miperes Drawn Voltage Comments Regarding
on Shaft (Amps) (Volts) Motor
7 11.0 108 Very warm.
6 9.5 110 Very warm.
5 8.0 111 Marginal.
4 7.0 112 Good.
3 6.2 113 Good.
2 5.5 114 Good.
1 5.0 115 Good.
0 5.0 116
The formula used to estimate G factors is as follows:
I
G=\/ V,q
where G = Velocity Gradient in sec 1
lbsec
Absolute Viscosity it
V = Volume of Basin (ft 3 )
Amps * Volts * PF
P = Water Horse Power = 746
PF = Power Factor
22
-------
Assumptions/Data:
,kat 8.5°C = 2.883(10 ) lb 2 sec
V = 6.7 ft 3 = Volume used in drum
3 propellers and 2 propellers.
For 3 propellers:
550(6.2 amps) (113 volts)0.63
G (6.7 ft )74 2.883(l0 ) lb sec 1 = 1298 sec 1
also yielding a dimensionless GT value = 245113.
For 2 propellers:
/ 550(5.5 amps) (115 volts)0.46 = 1049 sec 1
G V (6.7 ft )746 [ 2.883(l0 )l
also yielding a dimensionless GT value = 195114.
Static Mixing Pipes
Three Kenics static mixers were loaned to the Seattle Water Department by
the Municipal Environmental Research Laboratory, EPA, for use during the pilot
studies. EPA performed hydraulic testing on the mixers which allowed calcula-
tion of pressure drop and mixing intensity at various flow rates. The formula
used to calculate mixing intensity is:
/ 62.4 H
G V A * T
where G Velocity Gradient in sec 1
H = Headloss in ft
lbsec
= Absolute Viscosity in ft 2
T = Time in sec
Assumptions/Data:
,i at 8.5°C = 2.883(l0 ) lb ec
23
-------
Table 5 contains mixing information at various flow rates.
TABLE 5.
MIXING INFORMATION ON 2-INCH KENICS STATIC
MIXER
Flow Rate
(qpm)
Time
(sec)
Velocity
(ft/sec)*
Headloss
unit
pJig** ft H2O
Velocity
Gradient
(sec -)
Dimensionless
GT Value
8
1.8
0.83
4.8 0.21
502
904
12
1.2
1.25
11 0.49
940
1128
16
0.92
1.63
19 0.85
1414
1301
20
0.73
2.05
30 1.34
1993
1455
24
0.61
2.46
44 1.96
2637
1609
28
0.53
2.83
58 2.59
3252
1724
*ft/sec = feet per second.
= millimeters of mercury.
Flocculator
The flocculator is powered by a variable speed motor with an adjustable
rheostat. The flocculator paddles (area = 6.5 ft 2 ) are attached to a vertical
shaft and rotate in a clockwise direction. Two methods were employed to
estimate the energy input to the water by the paddles: (1) Substitution into
an empirical formula and (2) field measurements made with a torque meter.
The empirircal formula is presented as follows:
I - A
G= J
V AV
where Cd = Coefficient of Drag of Flocculator Paddles Moving Perpendicular to
Fluid
A = Area of Paddles in ft 2
Mass Fluid Density, slugs/ft 3
-2= Relative Velocity of Paddles in Fluid in ft/sec, usually about 0.7
to 0.8 of Paddle Tip Speed
G = Velocity Gradient in sec 1
V = Volume of Flocculation Chamber in ft 3
lbsec
A = Absolute Viscosity ft 2
24
-------
Assumptions/Data:
A = 6.5 ft 2
/2= 1.94 slugs/ft 3
2-= 0.7 Tip Velocity (Tip Velocity varied from 0 to 1.78 ft/sec)
, 4 (at 8.5°C = 2.883(l0 ) lb ec
V = 20.9 ft 3
Cd = 1.8
Information gathered on the flocculator is summarized in Table 6.
The second method of estimating energy input involved measuring torque
with the torque meter, calculating actual power input to the water, and then
tabulating G and GT factors. The formula used is as follows:
G = / 2 (RPM)Tq
V6o V
where RPM = Revolutions/mm
Tq = Torque in ft - lb
lb-sec
Absolute Viscosity in ft 2
V = Volume in ft 3
G = Velocity Gradient in sec 1
Assumptions/Data:
RPM = 019
Tq = 023 ft sec as measured on torque meter
,, at 8.5°C = 2.833(l0 ) lbsec
V = 20.9 ft 3
25
-------
Empirical
Calculations
-1
sec
Transducer Output (Tq )
With
Water
ft lb
98 57624
114 67032
140 82320
174 102310
294 119950
259 152290
276 162290
*Detention
time assumed to be 9.8 minutes at 16 gpm from Table 1.
% on Floccu].ator
Rheostat
Paddle
RPM
TABLE 6. MIXING INTENSITIES FOR FLOCCULATOR
Tip Velocity
ft/sec
G
GT*
Without
Water
ft lb
Torque Meter Date
M
Output
1
G, sec
40
0
1
0 00
2 10
r
0
0
--
0
0
0
9
0
10
13/4 0.16 5
2940
2 10
20
31/2
0.33 15
9 2 l0
411
1.5
30
17640
30
5-1/2 0.52 30
17640
2- 10
3
..-- - - -
GT*
8
0.75
52
50
10 0.94 73
60
70
12 1.13 96
14
1.31 20
80
90
15
1.41
18
100
134
30576 210 616
42924 210 819
56448 2 22
2 10 11 26
2 10 14 30
2 10 20 25
19
7
7.5
9.5
12.5
16
21.5
20 28 { 23
78792
103488
1.78
193 113484 3 10
-------
Review of Table 6 indicates that actual field measurements with the
torque meter resulted in hiGit and GT values that were consistently higher
than those from the empirical relationship; however, the differences are
remarkably small when one considers the many variables which were included
in the evaluation.
Filter Media Tested
Media Characteristics -
There were four different types of fIlter media evaluated during the
pilot tests and the characteristics of each as supplied by the manufacturer
are presented in the following tables.
TABLE 7. CHARACTERISTICS OF MEDIA TESTED ___
Table 7A
Neptune Microfloc Mixed Media
Type MM
TV
pe CMM
ffective SiZE
(mm)
Jniformity
oefficienl
rhicknes
(cm)
ffective Siz
(mm)
Iniformity
:oefficien
:hicknes
(cm)
Anthracite
Coal
1.0 1.1
1.7
45,7
1.0 1.1
1.7
53.3
Sand
0.42 0.55
1.8
22.9
0.42 0.52
1.4
17.8
Fine Garnet 0.18 0.32 2.2 7.6 0.18 0.32 2.2 5.1
Table 7B
Turbitrol
Type FC
;Effective SizE Jniformity hicknes
(mm) oefficient (cm)
Anthracite
Coal 0.92 1.28 50.8
Sand 0.40 1.30 25.4
Dual Media
Type CC
ffective Siz iniformity hicknes
(mm) oefficien (cm)
1.1 1.31 50.8
0.40 1.30 25.4
Notes: MM = Mixed media, sand MS-6.
CMM = Mixed media, sand MS-l8.
PC = Dual media with fine coal.
CC = Dual media with coarse coal.
mm = Millimeter.
27
-------
Other pertinent information on the medias may be found in Appendix B.
The media in the Waterboy (Type MM) was installed under the direction of the
Neptune Microfloc Service Representative. The media in the filter columns was
installed by the principal investigator. The placement was accomplished by
adding about 1 more media to the bottom layer than was actually needed,
backwashing the filter, skimming about onehalf of the excess, backwashing
again, and then skimming the remainder of the excess media.
Water Production Efficiencies -
Three methods were used to evaluate filter efficiencies; The Unit
Filter Run Volume (UFRV) -, Percent Efficiency and Net Water Produced per
24 hours.
The UFRV is the volume of water passing through a given filter surface
area during the course of a filter run. The units are gallons per square
foot per run. Trussel] suggests a minimum UFRV of 5000 gal/ft 2 /run because
filter production efficiency begins to drop off rapidly below that value.
The formula for UFRV is as follows:
UFRV = * T
2
where UFRV = Unit Filter Run Volume, gal/ft
run
LR = Average Filter Loading Rate, gpm/ft 2
T = Time, Length of Filter Run in minutes
The Percent Efficiency (%) is the net water produced per filter run
divided by the total water produced times 100. The formula is presented
below:
r(LR * T) - BW 1
% Efficiency = 100 [ LR * T
where LR = Average Filter Loading Rate, gpm/ft 2
T = Time, Length of Filter Run in minutes
BW = Amount of Backwash Water used, 200 gallons per ft 2 per backwash.
The Net Water Produced per 24 hours is the total volume of water produced
per unit of filter area in 24 hours less the amount of backwash water used to
clean the filter per 24 hours. This term also accounts for filter down time
during backwash. The formula is presented below:
NP = (LR * 1440) - NBW LBw+ (LR * T)]
28
-------
where NP = Net Water Produced, gallons/ft 2 /24 hours
LR = Average Filter Loading Rate, gpra/ft 2
NBW = Number of Backwashes per 24 hours
BW = Amount of Backwash Water Used, 200 gallons per ft 2 per backwash
T = Time, 15 minutes down time per backwash
Method of Operation --
The Waterboy filter was operated as a constant rate filter to a headloss
of 8 ft while filter columns were operated in either one of two modes constant
or declining rate. Filter runs on the columns were terminated ataheadlossof
10 ft and loading rates as high as 10 gpm/ft 2 were tested.
Chemical Additions
Several types and combinations of inorganic and organic chemical coagu
lants were investigated and the chemicals used are listed below.
TABLE 8. CHEMICALS TESTED
Type of Chemical
Name or Use Manufacturer
Alum Inorganic Coagulant
Ferric Chloride Inorganic Coagulant
Lime, Hydrated pH Control
1986N Nonionic Polymer American Cyanainid Co.
1849A Anionic Polymer American Cyanaxnid Co.
573C Cationic Polymer American Cyanamid Co.
N17 Nonionic Polymer Dow Chemical Co.
A23 Anionic Polymer Dow Chemical Co.
CA253 Anionic Polymer Calgon Corp.
CA233 Nonionic Polymer Calgon Corp.
CATFLOC T-1 Cationic Polymer Calgon Corp.
29
-------
SECTION V
SAMPLING AND ANALYSIS
SAMPLING PROCEDURES
Grab samples were gathered during the filter runs and analyses were
performed on these samples. pH and temperature measurements were performed
at the pilot plant location. Other analyses including aluminum, calcium,
conductivity, alkalinity, tannin and lignin, color, iron, dissolved oxygen,
suspended solids and particle counts were conducted in the Seattle Water
Department laboratory according to the 14th edition of Standard Methods 1 - 2
(where applicable). Asbestos analyses were performed by an electron micro
scopist at the University of Washington. Trihalomethane analyses were per-
formed by the Laboratory Branch, Region X, EPA.
TURBIDITY
During normal filtration runs, turbidity information was taken directly
from inline, Model 1720, Hach turbidimeters. When asbestos samples were
gathered, turbidity analysis was performed directly on a portion of each
individual water sample using the laboratory, Model 2lOOA, Hach turbidimeter.
Thus, correlations between turbidity and other parameters such as asbestos
and particle counts could be defined if they existed. The on-line field
turbidimeters (Model 1720) were calibrated against the laboratory turbidi-
meter (Model 2100A), which is calibrated using formazin standards.
Early in the pilot studies, air bubbles which were formed in the granular
media filter would become entrained in the finished water and give false
turbidity readings. This situation was eliminated by installing bubble traps
on the lines which lead to the recording turbidimeters.
PARTICLE COUNTER
A HIAC particle counter was loaned to the Seattle Water Department by
Pacific Scientific Company for use during the early portion of the pilot
studies. The counter was equipped with 12 separate channels and had a sensor
that could detect and count particles which were between 1 and 60 microns in
diameter. The instrument was calibrated using a stock solution of spheres
with a known particle distribution which were provided by the manufacturer.
All samples were degassed prior to being analyzed. Normally finished water
30
-------
samples were run directly on the instrument; however, raw water often had to
be diluted prior to analysis.
ASBESTOS FIBER ANALYSIS
General
The samples were gathered in quart cubitainers and were normally delivered
to the University of Washington within 24 hours after the time of collection;
thus, preservation techniques were not used. Concerning analysis, suspended material
in the water is collected on a Millipore filter, which is later dried in an
asbestosfree oven. A small disc is cut from the filter and it is then placed
upside down on a carbon-coated specimen grid. The filter is gently dissolved
in a condenser washer, leaving the asbestos and other suspended materialont1
grid. The grid is then placed in the electron microscope (EM) and magnified
for counting. Before a fiber can be counted, it must be identified as being
asbestos, preferably by type. There were two types of asbestos fibers found
in the Telt water amphibole and chrysotile. The identification of these
fibers is based on three factors: Morphology (size, shape and appearance),
crystal structure and elemental composition. 1 - 3
Identification
Since asbestos fibers generally appear to have sharp edges and often
ragged or broken ends, they can be readily distinguished from most biological
debris and many inorganic fibers. Identifying fibers by their crystal struc-
ture is done in the EM using selected area electron diffraction. An intense
beam of electrons penetrates a section of a selected fiber, and if a diffrac-
tion spot pattern appears on the screen, the substance is crystalline. The
arrangements of the spots depend on the atom layers in the crystal and are
therefore a sort of identifying fingerprint. Diffraction spots from the
amphibole fibers form rows of uniformly spaced dots while the three double-dot
diffraction pattern is an important way to identify chrysotile asbestos. The
elemental composition is determined by the Xray energy dispersive analysis
system (EDS) attached to the EM. The electron beam is focused on the fiber
of interest and Xrays are produced as a result of the interaction between
the electrons and the atoms on the surface of the fiber. A lithium drifted
silicon 1 i(Li)J detector converts the X-rays into voltage pulses proportional
to their Xray energies. After accumulating Xrays for a preset period of
time, the EDS unit displays the data graphically as a spectra of peaks. Since
atoms of different elements produce Xrays having different energies, the
positions of the peaks show which elements are present in the fiber. The
height of the peaks gives an indication as to the amount of each element
present. Since all types of fibers have a basic silicone structure but differ
in the amount of magnesium, iron, calcium and sodium which they contain, the
EDS can differentiate among thetypesof asbestos present. Identifying and
counting the fibers on a significant portion of the sample grid requires
between 1 and 4 hours. The number of fibers per liter present in the sample
is determined by multiplying the average number of fibers per grid opening
by a factor that contains the ratio of the area of the grid opening to the
filter area and the sample volume.
31
-------
Laboratory Analyst
The samples were analyzed by a team of microscopists at the University
of Washington (UW) headed by Dr. Edwin Boatman, School of Public Health. The
capability to analyze for waterborne asbestos was developed at the UW shortly
before the pilot studies began and since that time, the EPA has utilized that
capability in several asbestos surveys. Sample results were normally avail-
able within a couple weeks after collection and the cost of each analysis was
$250 per sample.
32
-------
SECTION VI
RESULTS AND DISCUSSION
RAW WATER
Quality
The Toit water supply is a high quality source of water originating from
rainfall and snowmeit runoff in the north Cascade mountains and possesses the
following water quality characteristics. (A complete chemical analysis sheet
may be found in Appendix A.)
TABLE 9. RAW WATER QUALITY CHARACTERISTICS
Parameter Value
pH 6.65 (units)
Alkalinity 5.0 (mg/i CaCO 3 )
Hardness 9.0 (mg/i CaCO 3 )
Conductivity 24 (micromhos)
Dissolved Oxygen 13 (mg/i, Saturated)
Temperature 2-10 0Centigrade (°C)]
Aluminum 0.21 (mg/i)
Coior 18 (units)
Tannin/Lignin 0.25 (mg/i)
Corrosivity Highiy Corrosive
Turbidity Range: 0.10-5 NTU; Average = 0.75 NTU
Bacteriological Counts Range: 165/100 milliliters (ml);
Average = 9/100 ml
Amphibole Range <0.045.7 (106) fibers/liter;
Average = 1.6(106) fibers/liter
Chrysotile Range l.2(106)_25.8 (106) fibers/liter;
Average = 7.1(106) fibers/liter
33
-------
The water is quite soft, with little buffering capacity and currently
meets the maximum contaminant level for turbidity set forth by the National
Interim Primary Drinking Water Regulations 1 - without filtration. Disinfection
is accomplished with gaseous chlorine.
Turbidity
Turbidity exceeds the desirable goal of 1.0 NTU on a seasonal basis. It
normally drops throughout the summer months to near 0.10 NTU and then begins
to rise slowly with the fall precipitation. Highest turbidities normally
occur during January, February, March, April and December. The turbidity has
not exceeded the 5 NTU maximum contaminant level since the National Interim
Primary Drinking Water Regulations 1 became effective in June of 1977.
The Tolt watershed is narrow and steep and has been prone to landslides
for many years. Roads have been built and intensive logging activity has been
carried out. This activity has increased the potential for high turbidity,
especially during heavy fall and winter precipitation. Further, fluctuations
in reservoir levels, especially the lower levels during the winter months,
cause large areas of bank to be exposed. These bare, muddy banks are extremely
vulnerable to erosion by heavy rains and wind and wave action. Fortunately,
during periods of high turbidity, the bacteriological quality of the water has
been excellent.
Asbestiform Counts
Table 10 lists the results of both raw water amphibole and chrysotile
counts which were gathered during the study. Appendix B contains micrographs
of both types of asbestos fibers found in Tolt water. These fibersare present
naturally in the streams which feed the South Fork Tolt Reservoir andhaveI
found in several water supplies throughout western Washington.
Amphibole fibers range in length from 0.3 to 7.5 microns and counts have
ranged from <0.04(106) fibers/liter up to 5.7(106) fibers/liter. Amphibole
counts fluctuate with the season of the year and appear to be related to raw
water turbidity as indicated by the statistical correlations developed below.
Statistical evaluation including all data points
Average Turbidity = 0.99 NTU
Standard Deviation = 0.82 NTU
Average Amphibole = 1.28(106) fibers/liter
Standard Deviation = 1.58(106) fibers/liter
Linear Regression: A = 1.0 + 0.25 (TURB)
where A = Amphibole in fibers/liter * 106
TUPB = Turbidity in NTU
34
-------
TABLE 10. RAW WATER ASBESTOS COUNTS
Date
of
Collection
Jan. 24, 77
Feb. 2, 177
Feb. 9, 177
Feb. 17, 77
Feb. 23, 77
Mar. 3, 77
Mar. 25, 77
Apr. 11, 77
Apr. 21, 77
May 6, 77
May 18, 77
June 1, 77
June 9, 77
June 29, 77
July 13, 77
Sept. 2, 77
Oct. 6, 77
Nov. 7, 77
Nov. 16, 77
Jan. 12, 78
Jan. 12, 78
Feb. 14, 78
June 8, 78
Sept. 4, 78
Turbidity
NTU
1.4
1.4
1.4
1.3
1.15
1.0
0.66
0.60
0.62
0.61
0 56
0.54
0.50
0.35
0.35
0.35
0.38
0.55
0.85
3.30
3.40
1.80
0.30
0.36
Amphibole
fibers/liter
5.70(106)
3.31(106)
3.06(106)
3.46(106)
4.33 (106)
1.76(106)
2. 18 (106)
2.4 (106)
0.94(106)
0.65(106)
0.90(106)
<0.29(106) (ND)
0.70 (106) (NSS)
<0.12(106) (ND)
<0.07(106) (ND)
<0.05(106) (ND)
<0.07(106) (ND)
<0.14(106) ( )
<0.14(106) (ND)
0.19(106) (NSS)
<0.10(106) ( )
Z0.07(10 6 ) (ND)
Z0.04(106) (ND)
0.07 (106) (NSS)
Chrysotile
fibers/liter
8.89(106)
5.12(106)
16. 39 (106)
13.0 (106)
13.29(106)
13.14(106)
25. 80 (106)
9.40(106)
4.25(106)
3. 82 (106)
2.80(106)
8.40(106)
3.60(106)
2.52(106)
2.81(106)
1.20(106)
3. 60 (106)
3.61(106)
4.62 (106)
5. 38 (106)
11.56(106)
3. 90 (106)
2.00(106)
1.84 (1O )
correlation Coefficient (R) = 0.13
Standard Error of Estimate (Se) 1.57(106) fibers/liter
Filter
Run
Number
3
4
5
6
11
12
21
24
29
33
44
51
53
62
70
89
93
108
ill
120
120
135
161
174
35
-------
Statistical evaluation excluding two outlying data points collected on
January 12, 1978
Average Turbidity = 0.77 NTh
Standard Deviation = 0.42 NTU
Average Amphibole = 1.38(106) fibers/liter
Standard Deviation = 1.62(106) fibers/liter
Linear Regression: A = 0.5 + 2.5 (TURB)
Correlation Coefficient (R) = 0.67
Standard Error of Estimate (Se) = 1.20(106) fibers/liter
As these regressions indicate, when all data points are included in the
statistical analysis, the correlation coefficient is very weak, 0.13. After
exclusion of two samples collected during unusually turbid water conditions
(3.3, 3.4 NTU), the correlation coefficient increases to 0.67, which gives a
good indication that there is a positive relationship between raw water tur-
bidity and amphibole counts. Figure 5 contains a graph of the data with the
line of best fit.
Concerning chrysotile, these fibers r nge in length from 0.1 up to 8
microns and counts have ranged from 1.2(10 ) up to 25.8(106) fibers/liter.
Statistics relating chrysotile counts to raw water turbidity are listed below.
Statistical evalution including all data points
Average Thrbidity = 0.99 NTU
Standard Deviation = 0.82 NTU
Average Chrysotile = 7.1(106) fibers/liter
Standard Deviation = 5.6(106) fibers/liter
Linear Regression: C = 5.5 + 1.6 (TURB)
where C = Chrysotile in fibers/liter * io 6
TURB = Turbidity in NTU
Correlation Coefficient (R) = 0.23
Standard Error of Estimate (Se) 5.45(106)
36
-------
5.0
LINEAR REGRESSION
0
0
-4
4.O
Li
-4
U,
Li
U-
Li
-J
0
a-
- .4
4
A=-O.5+2.5 (TURB)
R0.67
Se=1.2x10 6
1.0
POINTS NOT USED
1.5
2.0
TURBIDITY (NTU)
3.0
Figure 5. Raw water amphibole vs. turbidity.
-------
Statistical evalution excluding outlying data points collected on
January 12, 1978 and March 25, 1978
Average Turbidity = 0.78 NTU
Standard Deviation = 0.41 NTU
Average Chrysotile = 6.10(106) fibers/liter
Standard Deviation = 4.3(106) fibers/liter
Linear Regression: C = 1.36 + 6.1 (TURB)
Correlation Coefficient (R) = 0.66
Standard Error of Estimate (Se) = 3.2(106) fibers/liter
As with the amphibole, when outlying data points are excluded from the
analysis, there is a positive correlation (R = 0.66) between raw water chryso
tile counts and turbidity. Figure 6 contains a graph of the data with the
line of best fit.
FINISHED WATER
General
This section discusses goals for the quality of the finished water and
treatment trains needed to meet these goals. It deals with asbestos removal
techniques, develops relationships between finished water asbestiform counts
and turbidity and documents the removal of trihalomethane precursors while
using direct filtration treatment methods. Conditions surrounding each filter
run (i.e., flow rate, chemicals used, dosage, etc.) are presented in pendix C.
Headloss and turbidity information have been plotted for each filter run and
these figures may be found in Appendix D.
Water Quality Goals
Removal of all amphibole and chrysotile asbestos was the primary goal of
the research effort. To effect good fiber removal, a finished water turbidity
goal of < 0.10 NTU was established. Turbidities of 0.10 NTU are readily
attainable with filtration and such techniques had been effective at removing
amphibole fibers down to the detection limit at Duluth, Minnesota. 3 To pre-
clude problems with reflocculation in the distribution system, the lunerican
Water Works Association (AWWA) goal of <0.05 mg/i for aluminum in the finished
water was also established. It should be noted that EPA has set no standard
for asbestos levels in drinking water at this time and as such, complete fiber
removal was a selfimposed goal.
38
-------
Point Not Used
0
Points Not
Used
0
cD
-4
LU
I-
-J
U)
LU
U-
U
LU
-J
I
U)
>-
L)
..U
U
U
LINEAR REGRESSION
C=1.36+6.1 (TURB)
R0.66
Se=3 2x10
U
U
0
0
TURBIDITY (NTU)
0
Figure 6. Raw water chrysotile vs. turbidity.
-------
Chemical Treatments to Remove Turbidity
Alum Coagulation -
Soft, unbuffered, low turbidity waters are one of the most difficult to
destabilize and aggregate. 14 To determine what effects the addition of alum
would have on the pH of this water, titrations were performed and Figure 7 was
developed. pH is of concern because alum is only effective over a relatively
narrow pH range. It was evident that even very low dosages of alum signifi-
cantly depressed the pH and that a buffer of some type was needed to properly
pretreat the water before filtration. Lime is normally used for pH control
and Table 11 lists estimates of lime dosages required to maintain a pre-
selected pH at various alum dosages.
- TABLE 11. ESTIMATED LIME DOSAGES* 1& ii Ca(OH) j
Alum
mg/l
pH=6.7
pH=8.2
5
1.9
5.8
10
3.7
7.7
15
5.6
9.6
30
11.2
15.1
50
18.7
22.6
*Calculated values.
With an alkalinity of only 5 mg/l as CaCO3, the water is very sensitive
to both alum and lime addition and numerous filter runs did not meet water
quality goals due to a slight under or over feed of lime. Since the problem
persisted, tests were initiated to better define the optimum pH range for
destabilization. Figure 8 was developed and indicates that pH should be
maintained between 6.1 and 6.7 pH units. Field data indicated that between
1 and 4 mg/i of Ca (OH) 2 is required to maintain the pH in this range at a
dosage of 10 mg/i of alum. This compares favorably with the estimated dosage
of 3.7 mg/i of Ca(OH)2 which is listed in Table 11. The AWWA goals of 0.05
mg/l for aluminum and 0.10 NTTJ for turbidity in the finished water can con-
sistently be met in this pH range. Although these data were collected from
numerous filter runs conducted under various water quality conditions, they
indicate that pH is a very critical factor in the destabilization of Tolt
water.
Based on laboratory jar tests, an alum dosage of 10 mg/i was chosen as a
starting point for the pilot filter runs. This dosage, in conjunction with
granular media filtration, was very successful at removing turbidity from the
raw water under practically all conditions encountered during the testing
phase. Finished water turbidity would normally drop rapidly to < 0.10 NTU
within a half an hour and would remain at that level until breakthrough or
terminal headloss occurred. The addition of lime for pH control was critical
for consistent removal as was the use of a filter aid. (Approximately 100
filter runs were conducted at a dosage of 10 mg/i of alum.)
40
-------
7.
Figure 7. pH vs. alum dosage for raw Tolt water.
6.
=
0
w
I
tJ I
F-
w
I -
4
10
ALUM DOSAGE (mg/i)
40
41
-------
0.4
0.10
0.3 0.08
r
I
0. i6
0.2
0.04 -
I -
0.1
0.02
0
I I I I I I
5,7 5.9 6.1 6.3 6.5 6.7 6.9 7.1
pH
ALUMINUM
0
c TURBIDITY
0
0.
0
0
Figure 8. Finished water turbidity and aluminum residual vs. pH.
-------
To determine minimum dosages for effective turbidity removal, several
filter runs were conducted at dosages ranging from 3 to 10 mg/i. The infor-
mation from these runs is summarized in Table 12 and in Figure 9.
TABLE 12.
ALUM DOSAGE VS. OPERATING PARAMETERS
Raw Water
Alum
Finished Water Time to
Turbidity
Dose
Turbidity Steady State
Run #
NTU
mg/i
NTU or 0.10 NTU Comments
9
1.1
5.0
0.11 3 hours Slow breakin period.
10
1.15
3.0
0.9 c i hour No breakin period.
11
1.12
5.9
0.20 3 hours Slow breakin period.
22
0.62
5.6
0.1 2½ hour Slow breakin period.
23
0.63
6.0
0.1 2 hours Slow breakin period.
24
0.60
7.0
0.1 <1 hour Quickly reached steady state.
25
0.65
7.5
0.1 <1 hour Quickly reached steadyState.
26
0.60
7.0
0.1 <1 hour Quickly reached steadystate.
27
0.62
7.0
0.1 <1 hour Quickly reached steady state.
28
0.63
7.0
0.1 1½ hours Slow breakin period.
29
0.69
7.0
0.1 2 hours Slow breakin period.
50
0.56
8.0
0.1 <1 hour Quickly reached steady state.
104
0.46
8.5
0.1 <1 hour Quickly reached steady state.
105
0.48
8.5
0.1 ci hour Quickly reached steady state.
107
0.61
8.5
0.1 <1 hour Quickly reached steadystate.
108
0.70
8.5
0.1 <1 hour Quickly reached steadystate.
124
2.5
10.0
0.1 <1 hour Quickly reached steady state.
Dosages as low as 5 mg/i were successful at removing turbidity to ( 0.10
NTU; however, results at this dosage did not appear to be as consistent as at
the higher dosages. In addition, the lower dosages normally required a much
longer breakin period to reach steady state conditions.
Several dosages between 10 and 20 mg/l of alum were also tested and were
successful at removing turbidity down to <0.10 NTU. However, these higher
dosages did not appear to .offer any significant advantages over the 10 mg/l
dose.
Several nonionic and anionic polymers were used as filter aids. These
polymers included i986N, i849A, N17 and A23, and were fed directly onto the
top of the filter at dosages ranging from 0.020 to 0.25 mg/l. Generally, the
higher dosages were used to prevent filter breakthrough during periods when
raw water turbidity exceeded 1.0 NTU and at filter loading rates above 5gpri/ft 2 .
43
-------
1.0
0.8
I
I
4
U i
V)
-4
U-
0.2
0
a.
I U I U U U I U
o 1 2 3 4 5 6 7 8 9 10
ALUM DOSAGE (mg/i)
Figure 9. Finished water turbidity vs. alum dosage.
0
0
44
-------
Based on these pilot tests, the removal of turbidity down to levels
< 0.10 NTU is attainable using alum, lime and a filter aid in conjunction with
granular media filtration. The preferred dosage ranges and chemicals are
listed in Table 13.
TABLE 13. PREFERRED CHEMICAL TREATMENTS - ALUM
Chemical
Dosage
Alum
7-10 mg/i
Lime [ Ca (OH) 2 J to p11 between
6.1 and 6.7; alkalinity 4 mg/i.
1-4 mg/i
Nonionic or anionic filter
aid.
0.02.25
mg/i
Alum and Cationic Polymer --
One of the objectives set forth by the EPA was to investigate the effec-
tiveness of three conditioning chemicals in conjunction with granular media
filtration. To determine the optimum chemical dosages; various combinations
of alum, lime, cationic polymer and filter aids were investigated and are
listed in Table 14.
TABLE 14. ALUM AND LIME
DOSAGE VS. CATIONIC
POLYMER
DOSAGE
Alum Dosage
mg/i
CATFLOC
T-l
mg/i
Dosage
2
.2
.4
.8
1.6
3.2
4
.2
.4
.8
1.6
3.2
8 + Lime
.2
.4
.8
1.6
3.2
i6 + Lime
.2
.4
.8
1.6
3.2
Based on these screening tests, a combination of 35 mg/i of alum, 2 mg/i
of CATFLOC T-l and a filter aid was found to be quite effective at removing
turbidity down to < 0.10 NTU. The major advantage associated with this chemical
and dosage combination is that pH adjustment with lime is not necessary. Since
much smaller amounts of alum are used, both pH and alkaiinity are only slightly
affected.
Several nonionic and anionic filter aids including 1986N, CA233, CA253,
and A23 were tested during this portion of the pilot studies. These aids
prevented rapid breakthrough of the fioc particles from occurring. For exanple,
filter runs #64 and #81 indicated that although turbidity was removed to<0.iO
NTU with a dosage of 0.1 mg/i of CA233, breakthrough would occur several
hours before reaching terminal headloss. Increasing the dosage to 0.3 mg/i
(filter runs #65 and #82) prevented breakthrough from occurring until terminal
headioss was reached.
45
-------
The preferred dosages for the alum, cationic polymer and filter aid
combination are listed in Table 15.
TABLE 15. PREFERRED CHEMICAL TREATMENTS - ALUM PLUS CATIONIC POLYMER
Chemical Dosage
Alum 3-5 mg/i
Cationic Polymer 2 mg/i
Filter Aid (Nonionic or Anionic Polymer) 0.10.3 mg/i
Cationic Polymer Alone - -
Pilot tests were run to determine the effectiveness of using a cationic
polymer without alum. The dosages of cationic polymer, CATFLOC T-l and 573C,
were increased slowly at set intervals during separate filter runs with the
following results.
TABLE 16. CATFLOC T-l DOSAGES AND TURBIDITY
CATFLOC Tl Dosage (mg/i) Finished Water Turbidity (NTU)
0.12 0.35
0.20 0.34
0.40 0.34
0.60 0.32
0.80 0.33
1.2 0.25
2.0 0.15
3.0 0.1
TABLE 17. 573C DOSAGES AND TURBIDITY
573C Dosage (mg/i) Finished Water Turbidity (NTU)
0.12 0.19
0.20 0.18
0.40 0.18
0.60 0.19
0.80 0.14
1.6 0.095
2.4 0.10
3.2 0.08
46
-------
As indicated in these tables and in Figures 10 and 11, additions of
CATFLOC T1 and 573C were effective at removing turbidity down to < 0.10 NTU
if the dosage was increased to about 3 mg/i. The advantages associated with
this chemcail treatment are twofold. First, it does not affect the pH or
alkalinity of the water; thus, pH control with lime is not necessary. Sec-
ondly, the sludge generated from cationic polymers is normally denser and
easier to dewater than an alum sludge.
Ferric Chloride
A limited amount of testing was conducted using ferric chloride at
dosages ranging from 3 to about 20 mg/l. Results were not very encouraging.
A dosage of 69 mg/i with a filter aid could reduce turbidity to about 0.2 NTU
for a short period of time. Breakthrough would normally occur rapidly and
results were inconsistent.
Unit Processes
Static Mixers
Several filter runs were conducted to determine if static mixers would
effectively blend the treatment chemicals with the raw water and if so, how
many units would be needed to accomplish this task. Efficiency data from
filter runs #6A, B, C and D are listed in the following table and can be used
to compare results when different numbers of mixers were in use.
TABLE 18.
COMPARISON OF ONE WITH THREE STATIC MIXERS
Run
No.
No.
of
Static
Mixers
Filter Efficiencies
UFRV
(gal/ft 2 /run)
Net Water Produced
Efficiency
(%)
24 Hours
(gal/ft 2 /24 hours)
6A
3
1980
3785
89.9
6B
3
2160
3830
90.7
6C
1
2340
3868
91.5
6D
1
2430
3884
91.8
Review of the filter efficiencies resulting from the use of one or three
static mixers indicates that there was little or no discernable difference in
the results. A single static mixer appeared to mix the treatment chemicals
with the raw water as well as three mixers in series; thus, most runs ccnducted
during the study utilized only one mixer.
Back Mix Systems -
Several back mix systems were investigated including hydraulic and mechan-
ical mixers with and without the flocculator in the treatment train. There
appeared to be little difference among the various back mix systems tested.
Filter runs conducted with these systems would normally be terminated due to
turbidity breakthrough well before terminal headloss was reached.
47
-------
0.4
0.3i
I
>-
I
0.2
I
w
l J
U)
0 ii
0 1
CATFLOC T-1 DOSAGE (mgJl)
Figure 10. Finished water turbidity vs. Catfioc dosage.
I U
I
2
3
.
48
-------
.20
.15
- 10
LU
I
.
LU
C , )
-I
U..05 I
Ot__ I -I- U
1 2 3
CYANAMID 573C DOSAGE (mg/i)
Figure 11. Finished water turbidity vs. 573C dosage.
0
0
49
-------
Comparison of Static and Back Mix Systems -
Both static and back mix systems could be operated to produce an accept-
able quality finished water. One difference that was noted was the reason
for terminating a filter run. With the static mixers, there appeared to be a
quicker buildup of headloss and the run would be terminated because the
headloss reached 8 ft, rather than because of turbidity breakthrough. The
opposite was true with the back mix systems, headloss would build ata slower
rate and the run would normally be stopped because of turbidity breakthrough
not because it reached terminal headloss. The reason for this difference
in operational characteristics was not determined.
Flocculator
To determine if a flocculation basin should be included in the
treatment plant, the following set of filtration runs was conducted
without this unit process in the treatment train.
fullscale
with and
TABLE 19. RESULTS FROM FILTER RUNS WITH AND WITHOUT FLOCCULATION
Review of the data indicates that when the flocculator was deleted from
the treatment train, water production efficiencies were consistently unxthlc, er.
The runs conducted at 5-6 gpm/ft 2 were unable to remove turbidity down to<0.l0
NTU and even the runs conducted at 4 gpm/ft 2 experienced rapid breakthrough.
When the flocculator was placed back into the treatment train, the UFRV
and the filter efficiency rose immediately to much higher levels. Efficiencies
exceeded 95% and the turbidity was normally held to < 0.10 NTU until terminal
headloss occurred. The reason for the increased efficiency with the floccula
tor in place was not determined but the differences in operational capabilities
ilter
Flocculator
ilter Loading Rate
Number
UFRV
Filter
Run
No.
l27M
Yes/No
Yes
qpm/ft 2
of }burs
0.lON1U
10
Efficiency
gal/ft 2 /run %
4080 95.1
Comment
Maximum
7.0
Average
6.8
128M
Yes
8.0
6.5
16
6240
96.8
129M
No
6.1
6.1
0
0
0
rurbicIit
0.10 NTT
130M
No
6.0
6.0
0
0
0
Turbidity
>0.10 NT1
13114
No
4.0
4.0
8
1915
89.6
l32M
No
5.0
5.0
5
1500
86.7
133M
No
6.0
6.0
0
0
0
Turbidit
?040 NTt
l34M
Yes
4.0
4.0
25
5985
96.7
13514
Yes
4.0
4.0
23
5506
96.3
13614
Yes
5.0
5.0
16
4781
95.8
137M
Yes
6.0
6.0
22
7405
97.3
50
-------
were evident. These differences were most noticeable when the raw water
turbidity was >1.5 NTU and water temperatures were cold, 56°C. The cold
water conditions may have required more contact or reaction time before
filtering.
Filter Media
Loading Rates and Mode of Operation -
One of the initial objectives established by the EPA was to conduct
filtration experiments at loading rates of 56 gpm/ft 2 ; loading rates as
high as 10 gpm/ft 2 were investigated during the study using various filter
medias. The granular filters were operated in either a constant or declining
rate mode up to a terminal headloss of 10 ft.
Water Production Efficiencies
Methods used to evaluate filter efficiencies were described in detail
in SECTION IV, EQUIPMENT DESIGN, INSTALLATION AND OPERATION, and included
the Unit Filter Run Volume, Percent Efficiency and Net Water Produced per
24 hours. The minimum goal for the UFRV was set at 500011 gal/ft 2 /run for
the nominal flow rate of 60 MGD, which is equivalent to a filter efficiency
of 96%. Rather than setting a goal for Net Water Produced per 24 hours, this
parameter was used to establish the filter area required to produce the inaxi
muiii daily flow of 100 MGD.
As mentioned earlier in this report, air binding in the Waterboy filter
was a serious problem. It shortened filter runs to the point that very few
of the runs ever approached the minimum UFRV or percentage efficiency goals.
The headloss would progress at a normal rate until about 4 ft of headloss;
then, as the gases rapidly began coming out of solution, the rate of headloss
build-up would accelerate. Review of Figure 12 illustrates the two headloss
rates. To address the problem, three filter columns were designed and instailed
to maintain a positive head throughout the media for the duration of the filter
run. A comparison of the headloss buildup in the Waterboy filter and in a
filter column (Figure 12) illustrates the seriousness of the air binding prob-
lem. In addition to a reduced rate of headloss buildup, the columns demon-
strated much higher UFRVs and percentage efficiencies than the Waterboy filter.
Fortunately, the air binding problem can be eliminated in the fullscale plant
simply by insuring that an adequate depth of water is maintained over the
surface of the filter media. Other cold waters with high dissolved oxygen
and gas content occur in many areas of the Northwest and would probably cause
similar problems upon filtration. The problem is especially acute where direct
filtration techniques are used and extra freeboard in the filter boxes ought
to be considered in all such applications.
After the filter columns were installed, several typical filter runs were
selected and measures of water production efficiencies were calculated (see
Appendix A). Filter runs were terminated when turbidity was 0.l0 NTU or at
10 ft of headloss, thichever came first. Most of the runs analyzed had average
filter loading rates of 6 gpm/ft or above, and rates at the beginning of
declining rate runs were as high as 10 gpm/ft 2 . To simulate filter conditions
which would occur at peak summer flows of 100 MCD, data from four declining
rate runs (#153, #156, #157 and #161) were also analyzed in a slightly dif-
ferent manner. The rate at the beginning of these runs was 10 gpm/ft 2 and
51
-------
10
WATERBOY
I-
L U
L U
t O
t o
0
-J
0
C
LU
8
6
4
2
U ,
M
IKJTE DIFFERENT c
S LOP ES
4>
FILTER
COLUMN
2 4 6
8 10 12 14 16 18 20
TIME (HOURS)
Figure 12. Headloss vs. time - runs#93 & #93M.
-------
it was assumed that the filter run would terminate when the loading rate had
decreased to 6 gpm/ft 2 rather than allowing the run to progress to 10 ft of
headloss. This data provides a better picture of the filtersoperating under
stressed conditions. To evaluate data from the various medias under consi-
deration, sidebyside testing of the three filter columns was undertaken
and measures of water production efficiency for runs #139, #140, #141, #145,
#147, #153*, #156*, #157*, and #161* were plotted in Figures 13 and 14 and
linear regressions are presented in the following table.
TABLE 20. LINEAR REGRESSION EOUATIONS FOR THE VARIOUS FILTER MEDIAS
Filter Media
MM
FC
CC
UFRV
ga l/ft 2 /run
7(LR ) = 8442 488(LR)
= 6620 19
= 9308 610(LR)
Net Water
Produced per = 197 + l323(LR) = 151 + l322(LR)
24 Hours
gal/ft 2 /24 1Durs
= 397 + 1283(LR)
+LR = Average Filter Loading Rate in gpm/ft 2 .
Review of the Figures and the regression equations indicates that when
filter loading rates are between about 5.5 gpm/ft 2 and 7.5 gpm/ft 2 , there is
little difference in the UFRVt5 of the different medias. Assuming the data
can be extrapolated outside that range, the differences become more exaggerated.
At loading rates less than 5.5 gpxn/ft 2 , the dual media with coarse coal (CC)
would produce more water per filter run and is more efficient than either dual
media with fine coal (FC) or the mixed media (MM). At rates exceeding 7.5
gpm/ft 2 , the mixed media (MM) produced more water per filter run. This finding
may be related to a more even distribution of floc particles throughout the
mixed media which becomes evident at the higher loading rates. The linear
regressions for the net water produced per 24 hours were about the same for all
three medias.
By analyzing all the data contained in Appendix A3, the following linear
regressions are obtained.
UFRV (gpm/ft 2 /run) 8773 572(LR)
Net Water Produced per 24 Hours (gal/ft 2 /24 hours) = 665 + 1239(LR)
where LR = Average Filter Loading Rate (gpm/ft 2 )
*Effjcjencjes were calculated as if filter runs were terminated at 10 ft of
headloss and then at 6 gpm/ft 2 .
53
-------
(gpm/ft 2 )
a
4
U
0
4 5 6 7 8
FILTER LOADING RATE
Figure 13.
UFRV vs. filter loading rate.
4
a
Legend: MM
FC a
cc =
C
S.-
4-
I-
LI.-
U
0
0
.
I
a
4
U
6000
5000
4000
3000
U
U
a
54
-------
11
10,000
9,000
8,000
CsJ
C.J
-1-
4-
I-
C D
-
CD
C D
-7
c1. I,
LU
I
I
LU
6,000
5,000
4,000
4 5 6 7 8
FILTER LOADING RATE (gpm/ft 2 )
Figure 14. Net water produced per 24 hours vs. the filter loading rate.
55
Legend
MM =
FC =
cc =
-------
These equations combine data from all filter medias tested, are quite
similar to the ones found in Table 20 and will be used to estimate the required
filter surface area in SECTION VII, DESIGN CONSIDERATI 1 IS.
One filter run, #174, was conducted using a different combination and
depth of the various mixed media components. To reduce the rate of headloss
buildup, the depth of garnet and sand was reduced and the depth of anthracite
was increased. In addition, the uniformity coefficient of the sand was
slightly lower, 1.4 for filter media CMM, as opposed to 1.8 for filter media
MM. Results of that run indicated that there were virtually no differences
between average filter loading rates, the UFRVs and the net water produced
per 24 hours for the various medias. Figures 15, 16 and 17 illustrate the
operating parameters throughout the filter run for each media as well as water
quality data which is discussed in the following section.
In summary, there were advantages and disadvantages associated with both
dual and mixed media filters. Both types of filters met process goals and
efficiencies were ractica1ly the same 2 when the filter loading rate was between
5.5 and 7.5 gpm/ft. Below 5.5 gpm/ft , the coarse coal dual media (CC) was
more efficient and above 7.5 gpm/ft 2 , the mixed media (MM) demonstrated a
higher efficiency.
Water Quality Generated --
Since removal of asbestos fibers was a prime consideration of the study,
a comparison of the chrysotile counts in the finished water generated from
the various medias was made and results are shown below.
TABLE 21. CHRYSOTILE RESULTS FROM VARIOUS FILTER MEDIAS
Run No.
Hour
Into Run
Filter Column
MM
106
fibers/liter
FC
106
fibers/lite
CC
106
fibers/litei
CMM
106
ibers/liter
161
1
0.Ol(NSS)
0.0l(ND)
<0.0l(ND)
4
0.02(NSS)
0.04(NSS)
cO.Ol(ND)
10
-
<0.0l(ND)
0.02(NSS)
0.0l(NSS)
13
<0.0l(ND)
0.03(NSS)
cO.Ol(ND)
15
<0.01(ND)
0.0l(ND)
<0.Ol(ND)
174
0
0.07
0.Ol(NSS)
0.09
1
0.1
0.04(NSS)
0.02(NSS)
1
0.36
<0.03(ND)
0.94
2
0.03(NSS)
0.Ol(NSS)
-
<0.Ol(ND)
3
-------
0.5
# l7t lM
0.4
0.3
Ό0
0
r.
s-I
a)
-I -I
-r i
U )
s - i
a)
p
- r i
f r i
U C l )
-J
z
8
C )
H
f r i
0
C l)
C)
0.2
0.
10
9
8
7
6
-I i
1 1 4
5
U)
U)
0
4Q
3
2
1
Q = Turbidity
*=
Filter
ing Rate
Headloss
0 = Chrysotile
14 16 18
TIME (Hours)
22 24
Figure 15. Operating data for run #174MM.
-------
#174CC
w
0
H
4
0
4- I
- H
i- I
U)
14
0
r 1
124
U )
ui
OD
0
C -)
i-1
H
0
U )
C )
0.5
0.4
0.3
0.2
0.1
0.0
- I i
1 ti
C )
C l )
0
Q
0
Q Turbidity
Filter
Loading
4 6 8 10 12 14 16 18 20 22 24 0
TIME (Hours)
Figure 16. Operating data for run #174CC.
-------
= i1ter
loading
tl=
Chrysol
0 2 4 6 8 10 12 14 16 18 20 22 24
5
4
4J
cz 4
C l )
0
TIME (Hours)
#1 74CMM
9
w
0
a)
1-4
a)
C l )
E4
0 1
0
C)
H
E-
0
U)
:i1
C)
1.0
0.8
0.6
0.4
0.2
0.0
8
7
6
0 = Turbidity
3
2
1
Figure 17. Operating data for run #174CMM.
-------
When one considers the complexity of the asbestos analysis, the data
indicate that all filter medias are equally effective at removing asbestos
fibers during the normal course of a filter run. With the exception of
samples collected during a turbidity spike and at the very beginning of
the filter run #174, all samples had ND or NSS levels of chrysotile.
Finished water turbidities were also virtually the same for all of the medias
being tested.
Asbestos Removals
Two objectives established by the EPA at the outset of the study were
to: (1) confirm earlier amphibole removal results gathered during the Duluth
asbestos removal study, 3 and (2) develop methods for chrysotile removal.
Finished water asbestos results are contained in Appendix A.
Amphibole --
Review of the removal data indicates that amphiboi fibers can consist-
ently be removed down to the detection limit of 0.01(10 ) fibers/liter.
Removal efficiencies were between 98.3 and 100%. Where amphiboles were
detected in the raw water and when turbidity was < 0.10 NTU, 52 out of 57
results (91%) had levels of amphibole which were at or less than the detec-
tion limit. In all 5 of the cases where amphibole fibers were>0.Ol(l0 6 )
fibers/liter, 4 or fewer fibers were counted by the analyst, which means
that the results were not statistically significant. 3 out of the 5 cases
occurred when alum without lime addition was being employed as a treatment
method and the other 2 occurred at an alum dosage of about 7 mg/i rather
than the optimal 10 mg/l. These are possible explanations for the outlying
data points but by no means are they conclusive evidence. As indicated in
Figure 18, when the finished water turbidities exceed 0.10 NTU, amphibole
counts are much higher and there is a noticeable scatter to the data. These
data points represent times when either poor destabilization was occurring
or when turbidity breakthrough had already taken place. In conclusion,
amphibole fibers can consistently be removed down to the detection limit
if certain operating conditions are met. This confirms the earlier amphibole
research and operating data at Duluth, Minnesota. 3
Chrysotile
Review of the chry otile data listed in Appendix A indicates that although
excellent removals can be achieved, it is more difficult to remove than amphi-
bole and results are more variable. Figure 19 indicates that there is a very
tight pattern of data when finished water turbidity is < 0.10 NTU and consider-
able scatter above that point. As with the amphibole results, the scattered
chrysotile data above 0.10 NTU are normally associated with either poor
destabilization, turbidity spikes or breakthrough. Data from filter runs
#111, #120, #161 and #174 indicate that chrysotile can consistently be removed
down to not detectable or not statistically significant levels when good
filtration is occurring (i.e., < 0.10 NTU). Figure 20 can be used to estimate
what levels of chrysotile were present at any given time when finished water
turbidity was <0.10 NTU. For example, 50% of the time when turbidity was
<0.10 NTU, finished water chrysotile counts were < 0.02(106) fibers/liter.
Figure 21 estimates levels of chrysotile present when turbidity was 0.lO NTU
and indicates than 50% of the time, chrysotile counts were <0.27(106)
60
-------
1.0
0
0
Region of Most Successful Removal
0
00
0 00 0
NOTE: The number below a data point
represents the number of times
a given turbidity and fiber
count occurred.
I I I
I
0.1 0.2 0.3 0.4 0.5
FINISHED WATER TURBIDITY (NTU)
Figure 18. Amphibole vs. finished water turbidity.
0.1
0.01
o.
I - .
I-u
I
I-I
(I )
i u
-4
w
-4
=
0
0
0
ii
ii
23
I
61
-------
a
10.0
5.0 Region of Most Successful Removal
LiJ
F-
4
-J
0
U.Z
w
4
9 E l 0
a
V)
I. a
a 0 a
a
C-) U
L
_J
-4
I
C.1 a
>-
= DI 0
C) 3D
0.05
0
9
NOTE: The number below a data point
represents the number of times
a given turbidity and fiber
a a count occurred.
0.01 rlrjIrlrlI I
0.2 0.3 0.4 0.5
FINISHED WATER TURBIDITY (NTU)
Figure 19. Chrysotile count vs. finished water turbidity.
62
-------
0.35
0.30
0
C
.
LU
I .
L I)
LU
I -I
U-
E 0.15
UJ
-J
I
C
0.10
0.05
1 2 5 10 20 30 40 50 60 70 80 90 95
PROBABILITY OF CHRYSOTILE BEING LESS THAN OR EQUAL TO GIVEN VALUE (%)
Figure 20. Probability plot for chrysotile when finished water turbidity is < 0.10 NTU.
Note: Finished water turbidity < 0.10 NTU.
S
50% of the time, chrysotile is < 0.02(106
fibers/liter.
98
-------
Note: Finished water> .0.l0 NTU.
50% of the time, chrysotile is <0.27(106)
fibers/liter.
iu.uuuuiuuuiiuuuuuuuiuuuiuuuuuiuiuuuuiuuuuiuiuiuuuuuuuuuuuiiiiiuiiuiwdli
I I I I I I I
2 5 10 20 30 40 50 60 70 80 90 95
PROBABILITY OF CHRYSOTILE BEING LESS THAN OR EQUAL TO GIVEN VALUE %
Figure 21. Probability plot for chrysotile when turbidity is >0.10 NTU.
2.0
0
i .
0.9
U)
07
H 0.6
U) 0.5
z
0
U
1Z 0.3
H
0
U)
0.2
U
0 i
a
.
U
a
a
a
a
a
a
a
a
a
a
a
I
-------
fibers/liter. Thus, over 10 times more asbestos was present in the finished
water when turbidity was >0.10 NTU as compared to when turbidity was <0JD NTU.
In conclusion, chrysotile fibers can consistently be removed to levels which
are NSS if certain operating conditions are met. This accomplishes the most
important objective established by the EPA for this study.
Relationship to Turbidity --
One of the objectives of the study established by the principal investi-
gator was to develop an operating tool which would indicate quickly when
effective asbestos removal was occurring. To be an operating tool, it needed
to be simple to calibrate and operate and had to give immediate results. An
online turbidimeter would fit the criteria if it could be correlated with
asbestos removal 0 To provide the background for development of this rela-
tionship, several filter runs and asbestos counts are reviewed in detail.
Filter Run #11 . This filter run was conducted at 5.9 mg/l of alum and
never reached the finished water goal of 0.10 NTU. Chrysotile results are
presented in the following table.
TABLE
22.
RESULTS
FROM Rtfl
#11
Chrysotile Counts
Sample
Location
Hour
Into
Run
Turbidity,
NTU
fibers/liter
PAW
FINISHED
6
6
1.15
0.20
13.39(106)
1.64 (106)
Although the finished water was meeting the 1.0 NTU maximum contaminant
level established by the National Interim Primary Drinking Water Regulations, 1
these results indicate that asbestos removal efficiencies were poor, only 87.7%
and that chrysotile counts were extremely high in the finished water.
Filter Runs #2land#24 . These filter runs were conducted using 7 and
9.2 mg/l of alum, respectively, along with lime for pH control and a polymer
as a filter aid. Results are presented in the following table and in Figures
22 and 23.
These results indicate that when finished water turbidity was < 0.10 NW,
then chrysotile removal was excellent (99.399.8%). At Hour 7 when the fin-
ished water turbidity spiked at 0.34 NTU, the finished water chrysotile rose
from NSS levels up to 12.25(106) fibers/liter (removal efficiency = 52.5%).
The spikes probably occurred because of a temporary overfeed of lime causing
pH to increase to levels that were not conducive to proper floc formation.
Even though 0.34 NTU is well within the 1.0 NTU maximum contaminant level,
poor destabilization was occurring and this resulted in very high asbestos
counts in the filtered water. High levels of chrysotile also coincided with
the turbidity spike in run #24.
65
-------
RAW WATER TURBIDITY = 0.70 NTU
RAW WATER CHRYSOTILE COUNT = 25.8 (io )
1.2
1.0
0.8
0.6
0.4
0.2
%
I
I
w
I
w
V )
C-
. -) C)
i4 I
LiJ
-J
ifl U)
a.
LU
a
U
C)
L)
LU
-J
C)
tZ U)
U >
=
(-)
LU
I-
-T
C)
LU
=
U)
4
, 4
- LL.
CHRYSOTILE
0
TURBIDITY
0 2 4 6 8
10
12
TIME (HOURS)
0
18
Figure 22.
Finished water turbidity and chrysotile vs. time - run 21.
-------
12
1.2
RAW WATER TURBIDITY = 0.60 NTU
Q
RAW WATER CHRYSOTILE COUNT = 9.4 (106) FIBERS/LITER __
LU
I
I -j
L I)
>- LU
1- 8
U-.
I
I-
CHRYSOTILE
(-)
LIJ
LU
I-
-J
I-
TURBIDITY
( I )
C . Li
>-
O.4i
U-
0
I I I I - I I
0 2 4 6 8 14 18 20
TIME (HOURS)
Figure 23. Finished ter turbidity and hrysoti1e vs. time - run 24.
-------
TABLE 23. RESULTS FROM RUNS #21 and #24
Chrysotile Counts
Sample Location
Hour Into Run Turbidity, NTU
fibers/liter
RAW (Run #21)
7 0.66
25.8 (106)
FINISHED (Run #21)
2 0.065
0.16(106)
FINISHED (Run #21)
6 0.06
0.06(106)
FINISHED (Run #21)
7 0.34
12.25(106)
FINISHED (Run #21)
8 0.07
9.19(106)
FINISHED (Run #21)
12 0.059
0.09(10 )
RAW (Run #24)
7 0.60
9.4 (106)
FINISHED (Run #24)
6 0.085
0.34(106)
FINISHED (Run #24)
7 0.36
6.2 (106)
FINISHED (Run #24)
13 0.07
0.13(106)
Filter Run #53.
This filter run was conducted using
aid. Asbestos results are presented
3 mg/i of CATFLOC
below.
T1 without a filter
TABLE 24. RESULTS FROM RUN #53
Sample Hour Into
Turbidity Chrysotile
Amphibole
Location Run
NTU fibers/liter
fibers/liter
RAW 9
0.50 3.0 (106)
0.70 (106) (NSS)
FINISHED 3
0.13 0.27(106)
<0.02(106) (ND)
FINISHED 9
0.24 11.2 (106)
1.2 (NSS)
Turbidity in the Hour 3 sample was slightly >0.1 NTU and asbestos removals
were good for amphibole but only marginal for chrysotile. The Hour 9 sample
was collected during a turbidity spike and it contained higher levels of both
amphibole and chryosile than did the raw water. This indicates that the floc
particles, which contain concentrated levels of asbestos, were probably shear 1
from the media and entered the filtered water. Thus, operations tending to
break the floc should be avoided.
Filter Run #62 . This filter run was conducted using 3 mg/i of alum, 2
mg/l of CATFLOC Pi and 0.1 mg/l of a nonionic polymer as a filter aid. Fin-
ished water turbidities were just at the 0.10 NTU goal until breakthrough
began to occur late in the run. Asbestos results are presented in Table 25.
68
-------
TABLE 25. RESULTS FROM RUN #62
Sample
Location
Hour Into
Run
Turbidity
NTU
Chrysotile
fibers/liter
Amphibole
fibers/liter
RAW
16
0.35
2.52 (106)
- (0.1 (106) (ND)
FINISHED
3
0.1
0.34(106)
<0.02(106) (ND)
FINISHED
9
0.105
0.24(106)
<0.02(106) (ND)
FINISHED
16
0.37
2.28(106)
0.1 (106)
It is interesting to note that although amphibole fibers were not detected
in the raw water, one fiber was found in the finished water after breakthrough
occurred. The filtration process was evidently concentrating the fibers to
the point that they were detected in the finished water after breakthrough
had occurred. The sample that had detectable amphibole also indicated a very
poor removal efficiency for chrysotile, only 9.5%.
Filter Run *120 . This filter run provides an excellent example of how
finished water turbidity and asbestos counts track during filter breakthrough.
The results are presented in the following table and in Figure 24.
TABLE
26.
RESULTS FROM RUN
#120
Chrysotile Count
Sample Location
Hour
Into
Run
Turbidity,
NTU
fibers/liter
RAW
9
3.3
5.38(106)
FINISHED
1
0.14
0.14(106)
FINISHED
2
0.090
0.19(106)
FINISHED
6
0.08
0.07(106)
FINISHED
9
0.065
<0.01(106) ()
FINISHED
12
0.06
0.01 (106) (NSS)
FINISHED
13
0.07
<0.01(106) ( )
FINISHED
15
0.062
0.01(106) (NSS)
FINISHED
16
0.071
0.01 (106) (NSS)
FINISHED
17
0.071
0.02(106) (Nss)
FINISHED
18
0.115
<0.01 (106) (ND)
FINISHED
10
0.14
0.19(106)
FINISHED
20
0.28
0.14 (106) (NSS)
FINISHED
21
0.48
0.28(106)
FINISHED
22
0.57
0.57(106)
FINISHED
23
1.2
1.25(106)
69
-------
1.0
0
a
-4
0.8
w
4 ,
-J
In
0.6
LU
-J
I
a
U)
= 0.4
1.2
1.0
0.8
0.6
0.4
0.2
I
>-
I
-4
I
0.2
Figure 24. Finished water chrysotile counts and turbidity vs. time-run #120.
1.2
RAW WATER TURBIDITY = 3.3 NTU
CHRYSOTILE
TURBIDITY
2
4
6
8
10 12 14
16 18 20 22 24
70
-------
This filter run indicates that high chrysotile counts in the filtered
water coincide with filter breakthrough as indicated by a rising finished
water turbidity. This run also illustrates that direct filtration techniques
can consistently remove chrysotile down to ND or NSS levels even during
periods of high raw water turbidities (3.33.4 NTU).
Filter Run #174 . Filter run #174 was conducted using 10 mg/l of alum,
lime for pH control and a nonionic polymer as a filter aid. Initial filter
loading rates were as high as 10 gpm/ft 2 and the declining rate filtration
mode of operation was used. To test the sensitivity of the turbidimeter in
detecting changes in the operation of the filter, the alum feed pump was dis-
continued and finished water turbidity was monitored continuously during the
process. Within about 15 minutes after disconnecting the pump, the finished
water turbidity began to rise rapidly. By adding the estimated detention time
of the water in the flocculation chamber and in the free space above the filter
media, a time of about 13.5 minutes is obtained. Although both chambers are
complete mix systems as opposed to plug flow reactors, a comparison of these
two times, 15 vs. 13.5 minutes, indicates that the simple turbidimeter func-
tions exceptionally well as a troubleshooting t instrument which could have
alerted an operator that something was wrong somewhere in the system. Within
15 minutes after the alum pump was placed back into operation, the finished
water turbidity returned to < 0.10 NTU and stayed at that low level for the
duration of the filter run. Asbestos analyses conducted before, during and
after pump shutdown are presented in Table 27 and in Figure 25. These data
indicate that when the alum pump was off and finished water turbidity had
risen only to 0.20 NTU, there was a noticeable increase in chrysotile counts
from the very low levels previously occurring.
TP BLE
27.
RESULTS FROM RUN #174MM
Turbidity
Chrysotile
P lum Pump
Sample Location
Hour
Into
Run
NTU
fibers/liter
On/Off
RAW
1½
0.36
1.84(106)
FINISHED
0
0.089
0.07(106)
On
FINISHED
1
0.075
0.1 (106)
On
FINISHED
1½
0.20
0.36(106)
Of f
FINISHED
2
0.079
0.03(106) (NSS)
On
FINISHED
3
0.078
<0. 01(106) (ND)
On
FINISHED
4
0.075
0.01(106) (NSS)
On
FINISHED
5
0.079
0.Ol(l0 6 )(NSS)
On
FINISHED
6
0.072
0.01(106) (NSS)
On
FINISHED
7
0.075
0.02(106) (NSS)
On
FINISHED
FINISHED
14
18
0.078
0.075
<0.01(106) (ND)
<0.01 (106) ()
On
On
71
-------
CHRYSOTILE = 0.36 (io )
0.30
11
>- 0.4 0.2
0.3 CHRYSOTILE
:: . FL TURBIDITY o.i
I I I I I I I I I I I I I I I I I I I I I
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Alum Off Alum On
TIME (HOURS)
Figure 25. Finished water turbidity and chrysotile vs. time-run #174MM.
-------
In conclusion, the data from numerous filter runs indicate that the
continuous, online turbidimeter can be used as an operational tool to
indicate when excellent asbestos removals are occurring. The turbidimeter
does not necessarily relate to specific asbestos concentrations in the
finished water but nevertheless it can alert the treatment plant operator
when poor destabilization, a turbidity spike or breakthrough are occurring.
These conditions invariably cause both turbidity and asbestos counts to
increase in the finished water.
Comparison of Different Methods of Asbestos Analysis -
About two-thirds of the way through the research study, the EPA project
officer asked if the scope of work might be expanded to compare the results
obtained from two different analysis methods the Millipore - 5 and the Nude
pore 15 . There are several procedural differences in the analysis methods, the
major ones being that filters with different pore sizes are used and in the
Nuclepore method, the fibers are coated with carbon. The Millipore procedure
uses a filter that has a pore size of 0.45 microns whereas the Nuclepore
filter has a pore size of 0.1 microns. To make the comparison, one raw water
sample and ten finished water samples were gathered, split and analyzed by
each method. Results of analysis are presented in Table 28.
TABLE 28.
COMPARISON
OF MILLIPORE AND NUCLEP
ORE ANALYSIS METHODS
Run No.
Hour
Into
Run
Turbidity
Amphibole
Chrysotile
Millipore Nuclepore
i 6 106
Millipore Nuclepore
106 106
161CC
161CC
1
10
0.092
0.069
<0.0l(ND) O.03(ND)
c0.0l(ND) 0.03(ND)
c0.0l(ND) 0.03(NSS)
0.01(NSS) O.03(ND)
16lFC
1
0.095
<0.0l(ND) -cO.l4(ND)
<0.O1(ND) 0.14(ND)
161MM
161MM
1
10
0.135
0.079
cO.Ol(ND) <0.l4(ND)
< .0.Ol(ND) 0.03(ND)
0.Ol(NSS) 0.l4(ND)
-------
sample was collected during a turbidity spike and based on previous results
should have had detectable chrysotile. The Millipore method indicated non-
detectable levels whereas Nuclepore filtration indicated that 0.71(106)
fibers/liter were present. Samples collected during the same turbidity
spike from other filter columns indicated that chrysotile was present when
analyzed by Millipore methods. Overall, the results of analysis by both
methods were virtually the same.
Particle Counts
Early in the pilot testing phase, the particle counts on hourly samples
throughout filter runs were measured and results are contained in Appendix B.
This data indicates that particle counts tracked fairly well with turbidity;
however, there were some unexplained spikes in the counts. The particle
counter is undoubtedly much more sensitive to treatment changes than the
turbidimeter and the former could be used to optimize chemical dosages and
fine tune treatment plant operations. The particle counter is more difficult
to calibrate and maintain and requires a much higher level of skill to operate
than a turbidimeter; therefore, the particle counter was not judged suitable
as an online tool for indicating the acceptability of filter effluent.
High Turbidity Removal
Since direct filtration techniques were being employed and were quite
effective at removing turbidity and asbestos fibers from the raw water, it
became desirable to determine the capability of these same techniques during
abnormally high raw water turbidities (>5 NTU). Therefore, tests were con-
ducted to determine (1) if the process flow schematic, which included a static
mixer, a flocculator and granular media filters, would remove high levels of
turbidity; (2) what practical upper limit of raw water turbidity could be
removed; and (3) which of the test filter medias was most effective under
the stressed turbidity conditions.
As determined from previous pilot tests, a finished water turbidity
level of < 0.10 NTtJ was needed to effectively remove asbestos fibers. For
this series of tests, meeting the National Interim Primary Drinking Water
Regulations 1 for turbidity was the primary goal; therefore, a finished water
goal of <0.5NTU was established. Occurrences causing high turbidity, such as
landslides or flooding, are considered short term phenomena and asbestos
fiber removal would likely be of secondary importance during these brief
periods. It should be noted that the Tolt rarely, if ever, will exceed a
turbidity of 5 NW at the outlet from the Regulating Basin.
To achieve the high turbidities, soil from the banks of the South Fork
Tolt Reservoir was gathered and mixed with raw Tolt water. The slurry was
allowed to settle to remove heavy suspended material and the supernatent was
decanted into another container for feeding into the pilot plant. Equipment
limitations prohibited the investigator from injecting the slurry at a point
upstream of the static mixer and consequently it was fed directly into the
head end of the flocculation chamber. Turbidities between 5 and 34 NW were
tested and results were encouraging. (Data for runs #150, #151, #154, #159,
#160, #162 and #163 are listed in Appendix C.)
74
-------
As expected, the filter run efficiencies would decrease substantially
as raw water turbidities were increased; nevertheless, 10 mg/i of alum along
with lime and nonionic polymer could reduce finished water turbidity to
0.5 NTU when raw water turbidities were less than about 20 NTU. Run #160
yielded interesting results and these are illustrated in Figure 26. Raw
water turbidities were fairly constant at about 15 NTU and finished water
was normal1y 0.l0 NTU. At Hour 6, the raw water turbidity spiked at 34 NTU
and finished water turbidity rose immediately above the 0.5 NTU goal.
As the raw water turbidity was lowered to its original level, the finished
water turbidity concurrently dropped back to acceptable levels. Dual media
filters with coarse coal normally yielded more water per 24 hours than either
the dual media with fine coal or the mixed media filters. Results under the
high raw water turbidity conditions were favorable and indicate that direct
filtration techniques would produce an acceptable finished water for short
periods of time until the turbidity receded to lower levels.
Organic Removal
The Tolt River supply has been plagued with color problems since it was
brought on line in the mid1960s, although the color has been steadily
decreasing since that time. Some of the color probably results from naturally
occurring organic material originating in the watershed and from swampy areas
within the reservoir itself. Since colorcausing organic precursors in the
water, such as tannins and lignins, could form chloroform and other trihalo
methanes (TENs) upon chlorination, the efficiency of filtration at removing
these materials was documented and is listed in Table 29.
TABLE
29. ORGANIC
REMOVAL DATA
Sample
Color
Tannin/Lignin
Run No.
Location
Units
ing/l
%
Removal
111
RAW
FINISHED
-
0.17
0.08
53%
120
RAW
FINISHED
17
8
0.25
0.08
68%
130
RAW
FINISHED
17
5
0.25
0.07
72%
135
RAW
FINISHED
17
3
0.23
0.07
70%
174
RAW
FINISHED
RAW
16
12
17
0.08
0.035
0.08
56%
175
FINISHED
FINISHED
FINISHED
RAW
12
12
14
15
0.03
0.03
0.025
0.075
62%
62%
69%
175
FINISHED
FINISHED
FINISHED
12
12
12
0.03
0.03
0.03
60%
60%
60%
75
-------
o 1 2 3 4 5 6 7 8 9 10 11 12 13
TIME (HOURS)
2.0
1.9
High Turbidity Filter Run
Run 160 MM
1.8
1.7
40
35
30
25
20
15
10
5
1.6
=
I
I
I-
Lj I
I-
c c
cc
3
1.5
1.4
1.3
1 2
1.1
.0
RAW
I-
>-
I-
I
LU
I
cc
LU
=
( 1
I -I
U-
I
I
I
I
I
I
I
I
I
I J
0.7
0.6
0.5
0.4
0.3
FINISHED
0.2
0.1
14
15 16
Figure 26. Raw and finished water turbidity vs. time for run #160 MM.
-------
These data indicate that from onehalf to twothirds of the tannins and
lignins can consistently be removed using direct filtration techniques. Based
on these encouraging results, further testing was performed to document actual
THM concentrations before and after filtration and results are contained in
Table 30.
*Ail results in micrograms per liter unless otherwise noted.
**CHBr2C1 and CHBr3 were detected in none of the sample (<0.l, g/l).
tFree residual chlorine.
The unfiltered water used in these tests was raw Toit water gathered
from a tap in the Regulating Building. The filtered water was collected from
the Waterboy pilot plant and had been pretreated with 10 mg/i of alum, lime and
a nonionic polymer prior to filtration. The flow schematic included a single
static mixer, a flocculator and a mixed media granular filter. The filtration
rate was 4 gpm/ft 2 and the raw and finished water turbidities were 1.0 NTU and
0.14 NTU respectively. All samples were transported to the Seattle Water
Department laboratory in specially prepared glass containers, where they were
adjusted to a pH of about 8 and an alkalinity of 25 mg/I with lime, sodium
bicarbonate and sodium silicate. Water was then fluoridated to 1 mg/l and
TABLE 30. TRIHALOMETHANE RESULTS ON UNFILTERED AND FILTERED TOLT WATER*
I
)ay Parameter**
7°C
Total
THM
%Removal
25°C
Total
THN
%Reinova:
Unfiltered Filtered
Water Water
pH=8.25 pH=8.0
tJnfiltered Filtered
Water Water
pH=8.25 pH=8.0
CHC 1 3
0 CHBrC1 2
Total THM
Re5jduai ] t(mg/1)
27
1.4
28
1.8
9.5
1.0
Ii
2.4
35
61% 1.4
36
1.8
5.5
1.1
7
2.4
80%
CHC1 3
CHBrC1 2
Total THM
CHC1 3
50 15
2.8 2.4
53 17
1.5 2.0
65 21
75
68% 3.0
78
0.8
60
26
2.7
29
2.0
30
63%
CHBrC1
Total THM
Residual CJ 2 (mg/i)
2.1
67
1.2
1.8
23
1.5
68%
2.9
63
0.3
2.6
33
1.8
48%
CHC1 3
CHBrC1 2
Total THN
Residual Cl 2 (mg/i)
CHC13
10 CHBrC12
Total THM
ResidualCl 2 t(mg/l)
65
3.1
68
0.8
90
3.4
93
0.7
19
0.7
20
2.0
25
2.3
27
2.0
135
70% 4.2
139
0.05
140
71%
144
0.05
45
1.8
47
1.6
60
4.7
65
1.2
66%
55%
77
-------
dosed with about 2.5 mg/i of chlorine. After these treatments were applied,
the water was placed into vials that were capped with Teflon seals az d the
samples were taken to the Region X, EPA, laboratory, incubated at 7°C or 25°C
and analyzed by the EPA/Bellar purge and trap method 16 for THM.
The results indicate that the same techniques that are used to remove
asbestos down to ND or NSS levels, will also substantially reduce THM pre-
cursors in Tolt water. Removal efficiencies for THM are highly consistent
and are typically on the order of 6070%. The removal efficiencies for tannin
and lignin are very similar to the THM removals although one cannot be certain
that they are the actual precursors. The chlorine demand of filtered water
is significantly less than the demand exerted by unfiltered water.
WASTEWATER TREATMENT
Quantity
The volume of wastewater to be treated depends upon the quantity used to
backwash the granular media filters and has been conservatively estimated at
200 gal/ft 2 /backwash. The number of backwashes per unit of time will depend
on the efficiency of the filter at different filter loading rates. These
parameters are selected in SECTION VII, DESIGN CONSIDERATIONS, and they will
provide the basis for estimating the total volume of backwash wastewater
generated at various plant flows.
Quality
The solids that are removed from the filters during backwash result from
two sources the alum added during the treatment process and the suspended
material present in the raw water.
An estimate of the solids produced from the addition of alum can be made
based on the following reaction. 7
A1 2 (S04)3 . 14 H20 + 3 Ca(HCO3)2
2 Ai(OH)3 + 3 CaSO4 + 6 CO 2 + 14 H20
The molecular weights of A12(SO4)3 . 14 H20 vs. 2 Ai(OH)3 are in a ratio
of 3.8 to 1. Thus, if 10 mg/i of alum is used in the treatment process, it
will result in approximately 2.63 mg/i of sludge being generated. This uates
to 0.022 pounds of sludge per 1000 gallons (lb/i000 gal) of water treated.
The concentration of suspended solids in raw Tolt water fluctuates through.
out the year but is normally 1 to 2 mg/i and never exceeds 3 mg/i. Since
filtration techniques are removing even the micron size asbestos particles,
it is probably removing most other suspended matter as well. If the suspended
solids concentration is 2 mg/l, then the solids removed amount to approximately
78
-------
0.017 ib/l000 gal treated. The total sludge generated from both the chemical
additives and the solids in the raw water is approximately 0.039 lb/i000 gal
water treated. This estimate falls within the range of values (0.038-0.061
lb/l000 gal) found by Black and Veatch Consulting Engineers during their
evaluation of Lake Superior water. 3
Based on the 0.039 lb/i000 gal water treated estimate, a UFRV of 5000
gal/f t 2 /run, and a volume of 200 gai/ft 2 /backwash; the suspended solids con-
centration in the backwash water would be an estimated 117 mg/l. A range in
suspended solids concentrations from 106 mg/i to 227 mg/i was obtained by
Black and Veatch Consulting Engineers; 3 thus, the 117 mg/i concentration
appears to be a reasonable estimate. Suspended solids tests were performed
on the backwash water from the Waterboy filter and the filter columns and
the solids concentrations were only about onehalf of the 117 mg/i estimate.
Proportionately larger quantities of backwash water are normally required to
clean filter columns and relatively small filters when compared to large scale
filters and this may have diluted the solids concentrations in these samples.
This is one possible explanation for the discrepancy between the field values
and the theoretical estimate.
Settleabiiity
Tests were performed to determine the settling characteristics of sludge
that had been generated from backwashing the Waterboy filter and held in the
settling basin Although the sludge resulted from several different water
treatment techniques, holding and concentrating the solids was the only method
that would provide enough quantity of sludge for the tests. Using a 6-inch
diameter settling column to minimize wail effects and a mechanical stirring
device that rotated at 1 RPM to enable water to move more freely upward
through the solids, interface settling velocities (ISV t s) were documented
at various sludge concentrations (Figure 27). Based on these tests, Figure
28, a batch flux curve was prepared and it can be used to estimate the lim-
iting flux at various underf low concentrations. 4 Review of the flux curve
indicates that underfiow concentrations of 23% are feasible using gravity
thickening techniques. Details of these tests are presented in Appendix B.
79
-------
60
500
400
300
I d
T j
4- 1
(1)
200
100
5 10 15 20 25 30
SOLIDS CONCENTRATION (grams/liter)
Figure 27. Interface settling velocity vs. solids concentration.
80
-------
50
40
>1
30
c 1
.1J
4- 1
-Q
Ii.i
10
Figure 28.
S
.
5 10 15 25
SOLIDS CONCENTRATION (grams/liter)
Batch settling curve for waste solids.
30
81
-------
SECTION VII
DESIGN CONSIDERATIONS
PREFERRED TREATMENT TRAIN
The preferred treatment train includes static mixers, flocculators and
granular media filters. The plant will have a maximum hydraulic capacity of
100 MGD and a nominal flow rate of 60 MGD. Since the study demonstrated
that direct filtration techniques met both process and water quality goals,
no settling basins will need to be incorporated into the design. The proposed
plant will remove amphibole fibers to ND levels, chrysotile fibers to NSS
levels, turbidity to 0.10 NTU, tannins and lignins to <0.07 mg/l, 6070%
of the THM precursors, color to <15 units, and will incorporate corrosion
control apparatus and a coagulant control center. The design also includes
provisions for treatment of backwash wastewater which will comply with the
discharge requirements of the state pollution control agencies. Asbestos
free materials will be used in construction of the plant. The design criteria
is summarized in Table 31 and the unit processes are illustrated in Figure 29.
A brief description of the unit processes is presented below.
FILTRATION PLANT
Energy Dissipators
To reduce the pressure head off of the Tolt pipeline between the South
Fork Tolt Reservoir and the Regulating Basin, energy disspiators will be
constructed ahead of the static mixers to reduce pressure from the existing
head of 160900 ft down to between 2040 ft. It will be equipped with both
fine and coarse controls to enable accurate adjustment of head exiting the
unit. If Seattle City Light Department decides to develop power on the Tolt
supply, water would flow through their proposed turbines into a stilling basin
and would then discharge into the existing Water Department Regulating Basin.
In this case, the water would need to be pumped from the Regulating Basin to
the static mixers rather than flowing through the energy dissipators as is
assumed at this time.
Static Mixers
To provide initial mixing of chemicals into the process stream, three
static mixers are planned to provide flexibility throughout the flow range.
Characteristics of these units, which require no outside power source and
very little maintenance, are described in Table 32.
82
-------
GENERAL DESIGN CRITERIA
FILTRATION SYSTEM DESIGN CRITERIA CHEMICAL STORAGE AND DOSAGE
PLANT CAPACITY MOD
NOMINAL
DESIGN HYDRAULIC CAPACITY
PRIMARY TREATED WATER QUALITY GOALS
TURBIDITY UNITS
COLOR UNITS
TASTE
ODOR
FLUORIDE
PH UNITS
ASBESTIFORM COUNT 100 FIBERS/LITER
AMPHIBOLE
CHRYSOTILE
ALKALINITY mg/I
SILICATE mg/I
ALUMINUM mg/I
NUMBER OF BASINS
NOMINAL CAPACITY. MOD
EACH
COMBINED
DESIGN HYDRAULIC CAPACITY, MGD
EACH
COMBINED
FLOCCULATION BASINS
DETENTION TIME AT NOMINAL
DESIGN FLOW, MINUTES
WATER DEPTH
STAGES OF FLOCCULATION
PADDLE WHEEL DIAMETER FT
VELOCITY GRADIENT, 0. FT SEC/FT
1ST STAGE
2ND STAGE
3RD STAGE
BASIN INLET VELOCITY, FT/SEC
BASIN OUTLET (DIFFUSION WALL)
VELOCITY, FT/SEC
AVAILABLE SPEED VARIATION RATIO
NUMBER OF FILTERS (DUAL)
FILTER AREA EACH FT 2
FILTRATION RATE, GPM/FT 2
NOMINAL
DESIGN HYDRAULIC CAPACITY
0 FILTER CAPACITY, EACH, MGD
o NOMINAL
1.0 DESIGN HYDRAULIC CAPACITY
8.3 SYSTEM CAPACITY, MOD
NOMINAL )4.2 GPP.j/FT 2 )
PEAK (6.9GPM/FT
0.01 TOTAL STATIC HEAD ACROSS FILTER, FT
2025 NET AVAILABLE INDICATED HEAD, FT 111
9 AT4.2GPM/FT 2
0.06 AT 6.9 GPM/FT 2
AT 10.0 GPM/FT 2
DESIGN MAXIMUM BACKWASH RATE
GPM/FT 2
DESIGN BACKWASH WATER USAGE
AS PERCENT OF RAW WATER (MAX.)
BACKWASH SUPPLY
RESERVOIR. MG
CAPACITY EACH BACKWASH PUMP 121 GPM
4 5
15,000
NUMBER OF AIR WASH BLOWERS 121 2
CAPACITY AIR WASH BLOWERS 121
26 EACH SCFM 4315
100 (1) AS MEASURED FROM FILTER WATER SURFACE
TO CENTER LINE OF EFFLUENT PIPE AHEAD
OF FLOW TUBE.
(2) ONE PUMP REDUNDANT
CLEAR WATER RESERVOIR DESIGN CRITERIA
CHEMICAL FORM AND METHOD
OF DELIVERY
4.2 ALUM
10.0 POLYMER
CATIONIC
7.5 ANIONIC/NONIONIC
12.8
LIME: FILTRATION
PLUS CORROSION
CHLORINE: PRE
PLUS POST
FLUORIDE LIQUID FLUOSILICIC
ACID. TANK TRUCK
ANTICIPATED DOSAGE RANGE, mg/I
MIN. AVG. MAX.2-WEEKS PEAK
2 10 20 30
2 5 7
0.1 0.2 0.3
10.3 12 18
0.5 1.5 3 5
0.5 1.0 1.0 1.0
1 4 10 15
3 7.5 15 30
8
1250
03
I. . . )
RAPID MIX DESIGN CRITERIA
NUMBER. STATIC TYPE 2 OR 3
FLOCCULATION BASIN DESIGN CRITERIA
60
100
14
12.7
12.1
11.1
LIQUID-TANK TRUCK
LIQUID-TANK TRUCK OR DRUM 0.5
LIQUID-TANK TRUCK OR DRUM 0.01
BULK QUICKLIME (CoOl 1.3
PNEUMATIC
CONVEYOR TRUCK
LIQUID-CL 2 1TON
CYLINDER
SODIUM
SILICATE
12 SODIUM BICARBONATE
5 SOLIDS HANDLING
LIQUID-TANK
TRUCK
15
20
3
15
NUMBER
CAPACITY. MG
MAXIMUM DEPTH, FT
1.0
5
25
2100 SURGE/CLARIFIER/THICKER
2-2000 GPM RETURN PUMPS
2100 GPM UNDERFLOW PUMPS
230 THICKENER/CONDITIONER
275 GPM UNDERFLOW PUMPS
2SOLID BOWL CENTRIFUGES. 750PM
TABLE 31
FILTRATION SYSTEM-DESIGN CRITERIA
0.25
2:1
-------
TO
REGULATION
BASIN
U
z
0
I-.
U
_ >- Ow
D j
- DI
-
-J
0
0.
w
2
0
C.)
w
2
0
I
FILTRATION
6 MG RESERVOIR
RAW WATER
TRANSMISSION
PIPELINE
(EXISTING)
PLANT
SERVICE WATER
PUMPS
SUPER N ATA NT
RETURN PUMPS
SURGE /CLARIFIERS
FINISHED WATER,
TO CITY
TO LAND
THICKENERS
SLUDGE DISPOSAL
FIGURE 29.
PRELIMINARY PROCESS DIAGRAM.
-------
TABLE 32. CHARACTERISTICS OF STATIC MIXERS
No. of
Units
Flow
Capacity
Diameter
Length
Pressure
Drop
No. of
Elements
Approximate
Weight of
Unit (1-b)
2
25 MGD
3 ft
9
ft,6
in
6.3 ft
3
2500
1
50 MGD
4 ft
12
ft,6
in
8.4 ft
3
4000
Flocculators
When water temperatures were cold and turbidity was> 1.5 NTU, use of the
flocculators demonstrated certain advantages including better turbidity
removal and longer filter runs. In addition, the Washington State Department
of Social and Health Services suggests a minimum contact time of 15 minutes
before filtration in direct filtration treatment plants, 1 - 8 thus the floccula-
tion chambers will also satisfy these requirements. Four flocculation chambers
are proposed, each with a 15-minute detention time at the design hydraulic
capacity of 100 MGD. A tapered threestage system with variable speed floc
culator paddles is suggested.
Filtration Gallery
Filter Loading Rates --
Based on the filtration rate studies and the equations and criteria
listed below, the proposed filter loading rates are developed and presented
in Table 33.
Equations Used:
UFRV = 8773 572(LR)
NP = 665 + 1239(LR)
where UFRV = Unit Filter Run Volume (gal/f t 2 /run)
NP = Net Water Produced (gal/ft 2 /24 hou 9 )
LR = Average Filter Loading Rate (gpm/ft
Minimum Criteria:
UFRV = 5000 gal/ft 2 /run 1-1 at 60 MGD
UFRV = 4000 gal/ft 2 /run at 100 MGD
Filter must produce a net volume of 100 MG/24 hours.
85
-------
TABLE 33. DETERMINING THE FILTER AREA
Assuming a
Minimum UFRV
of
at a
Capacity
of
Average Filter
Loading
Rate
Net Water
Produced
Filter Area
Required
5000 gal/ft 2 /run
60 MGD
Nominal Rate
6.6 gpm/ft 2
8850
gal/ft 2 /24
hours
6780 ft 2
4000 gal/ft 2 /run
100 MGD
Design
Hydraulic
Capacity
8.3 gpm/ft 2
10950
gal/ft 2 /24
hours
9130 ft 2
Based on these calculations, the limiting factor is the flow rate at 2
design hydraulic capacity, not the nominal rate; thus, a minimum of 9130 ft
of filter area is required to produce 100 MGD. Assuming that there will be
10,000 ft 2 of filter surface area divided into 8 filter boxes, the following
filter loading rates are developed.
TABLE 34.
RESULTING FILTER
LOADING RATES.
8
Filters Operating
7
Filters Operating*
Flow
(Average Rate)
gpm/ft 2
gpm/ft 2
60 MGD
4.2
4.8
100 MGD
6.9
79**
*One filter down for backwashing.
**Filter will be designed o handle hydraulic loading rates up to 10 gpm/ft 2 .
Extra freeboard in the filter boxes must be provided for two reasons.
First, it will be needed to maintain a positive head throughout the filter media
up to 10 ft of headloss and secondly, to accommodate fluctuations in water
levels inherent in declining rate filters. Rate controls will be of the type
that either constant or declining rate filtration will be attainable. Contin-
uous recording turbidimeters will be installed on the effluent from each filter
as well as the clearwell effluent.
Filter Media
Measures of water production efficiency including the UFRV, net water
produced per 24 hours and % efficiency indicate little differences among the
various medias evaluated in the loading rate range between 5.5 gpm/ft 2 and
7.5 gpm/ft . Below 5.5 gpm/ft 2 , the coarse coal, dual media (CC) filter
produced more water and above 7.5 gpm/ft 2 , the mixed media filter (MM) exhib-
ited the best efficiency. Most of the year, the plant will operate within
the range of 5.57.5 gpm/ft 2 or less and as such any of the filter medias
tested would be acceptable from a process or water quality standpoint. To
continue sidebyside testing of dual and mixed media filter, to encourage
lower bids for the filter media, and to take advantage of the benefits of
86
-------
both dual and mixed media; it may be desirable to specify both types of filter
media in the fullscale plant. Listed below in Table 35 are selected guidelines
for media selection. (This information is general in nature and should not be
used as is for bidding purposes.)
T2 BLE 35. SUGGESTED GUIDELINES FOR GRANULAR MEDIA FILTERS
Mixed Media
Dual Media
Effective Sizc
(Tm ! )
Uniformity
oeffidenI
hicknes
(cm)
ffective Size
(mm)
Uniformit
oefffcient
rhickness
(cm)
! nthraci te
Coal
1.0 1.1
1.7
53
0.9 1.1
1.3
51
Sand
0.42 0.52
1.4
18
0.40
1.3
25
Garnet
0.18 0.32
2.2
5
Backwash System
Since polymer additions were necessary to meet both water quality and
process goals, it is especially important to have a good backwashing system
to insure removal of the soft, adherent floc particles from the filter media.
An air scour, water rinse type system is suggested. Provisions should be made
for addition of a polymer to the backwash water to precoat the filter to
preclude a turbidity spike as the clean filter is placed back into operation.
Backwash water will be drawn from the finished water storage reservoir at the
plant. Assuming that 200 gallons f water per ft 2 of filter will be used per
backwash for cleaning the filter, 1 the backwash volume requirements are
estimated to be 3.1% of the water produced at the nominal filtration rate
of 4.2 gpm/ft 2 (60 MGD) and 4.2% at the design hydraulic capacity of 6.9
gpm/ft 2 (100 MGD). The figure of 200 gal/ft 2 /backwash is considered to be
a conservative estimate since air scour backwash systems normally require
significantly less volume of water than simply hydraulic or fluidized washes - 1
The assumptions made in sizing the filter were based on the premise that the
volume of backwash water would not exceed 5% of the water produced (95% filter
efficiency) at the peak loading rate; thus, a volume up to about 5 MGD could
periodically be required for backwashing purposes.
FILTERED WATER STORAGE A1 D CORROSION TREATMENT
To provide backwashing water and to dampen pipeline flow fluctuations,
a 5 MG finished water storage reservoir is proposed. It would be baffled and
will provide contact time for chlorine disinfection. In addition, the Tolt
is known to be a very corrosive water at this time and since filtration would
likely increase the problem, provisions for adding corrosion control chemicals
to the storage basin would be included. The overflow elevation of the clear
well must be no less than elevation 760 ft to provide enough water at an
adequate head to the Tolt supply line.
87
-------
WAS TEWATER TREATMENT
Characteristics
Based on the filter loading rates selected previously and the associated
filter efficiencies, the wastewater characteristics are listed in the follow-
ing table.
TABLE
36. BACKWASH
WASTEWATER
CHARACTERISTICS
Water
Produced
MGD
Efficiency
%
Wastewater
Produced
MGD
lb Solids
1000 gal
Produced
lb Solids
Day
Estimated Solids
Concentration
(mg/l)
60
3.1
1.86
0.039
2340
151
100
4.2
4.2
0.039
3900
111
Thickening
Using the batch flux curve shown in Figure 28, thickener area estimates
for several underf low concentrations are presented in the following table.
TABLE 37.
CLARIFIER/THICKENER AREA
REQUIREMENTS BASED ON BATCH
FLUX CURVES
Solids
Minimum
Water
Produced
MGD
Wastewater
Volume
MGD
Underflow
Concentration
(gram/liter)
Limiting Production
Flux (m of Solids
(1b/ft 2 /day) Day J
Surface
Area Required
(ft 2 )
60
1.86
20
25
30
35
42 2340
33 2340
25 2340
18 2340
56
71
94
130
100
4.2
20
25
30
35
42 3900
33 3900
25 3900
18 3900
93
118
156
217
In addition to the thickening requirements as dictated by the batch
settling curves, there are also hydraulic requirements that must be met to
enable a clarifier/thickener to operate properly. These requirements are
normally based on an allowable surface overflow rate and have been developed
by designers based on field experience. Textbooks normally list values between
600 and 1800 gallons per day per square foot (gpi/ft 2 ). - 4 Using the lower
value of 600 gpd/ft 2 and the peak wastewater flow of 4.2 MOD, this yields a
minimum surface area of 7000 ft 2 . Comparing 7000 ft 2 with 217 ft 2 (Table 37);
the surface overflow rates, not the thickening requirements, govern the surface
area of clarifier/thickener. With two circular clarifiers in operation, each
88
-------
would need to be a minimum of 66 ft in diameter. To absorb periodic slugs
of wastewater from the backwashing system; to provide a smooth, even flow of
clarified wastewater back to the head of the plant; and to provide some
extra storage before the dewatering process; the suggested design includes
two 100 ft diameter surge/clarifiers in parallel followed by two 30 ft
diameter thickeners with provisions for chemical treatment. Assuming the
waste was thickened to a 2% solids concentration, this would amount to
about 60 tons of sludge per day. Consideration should also be given to
the method for recycling of the clarified wastewaters. Since directly
recycled waters can change the raw water characteristics, provisions for
discharging the clarified water to the Regulating Basin prior to recycle
should be considered.
Dewatering
Several methods of sludge dewatering are available and they normally
involve the use of centrifuges, press filters or vacuum filters. These
devices can produce a cake with a solids concentration of 1540%. If the
waste was dewatered to a 15% solids concentration, this would result in about
2920 tons of sludge per year that would require ultimate disposal. For
purposes of cost estimation, two centrifuges were included in the waste
treatment train to dewater the thickened sludge. More thorough studies
of this subject need to be undertaken before an actual dewatering method
can be chosen.
89
-------
SECTION VIII
OPERATIONAL CONSIDERATIONS
To insure that asbestos fibers will be removed down to ND or NSS levels,
optimizing the operation of the filtration plant is of utmost importance. To
aid in the operations and monitoring functions, the facility should include a
gravity flow pilot plant of not less than 20 gpm capacity which is identical
to the fullscale plant. A coagulant control center and a computer would
facilitate treatment plant operation as would a particle counter. Turbidi-
meters on the effluent from each filter and on the finished water storage
reservoir should be mandatory and the finished water turbidity should be
maintained at < 0.1 NTU. Backwash should be initiated at terminal headloss
of 10 or when the effluent turbidity curve on the recorders begins to rise
upward indicating that turbidity breakthrough is beginning to occur.
90
-------
SECTION IX
COST ANALYSES
Construction and operating cost estimates have been developed byCH2WHill
Consulting Engineers for a 60 MGD (100 MGD peak flow) water filtration plant
for the Tolt supply. The American Association of Cost Engineers divides con-
struction estimating into three categories. 19 To better understand the level
of confidence to be placed in the estimates presented herein, a brief des-
cription of those categories is listed below.
1. Order of Magnitude - Approximate, prepared without detailed
data, prepared from cost curves with various factors applied
to scale up or down, considered accurate within +50% or
30%.
2. Budget Estimate - Prepared from flow sheets, lay-outs and
equipment details, considered accurate within +30% or 15%.
3. Definitive Estimate - Prepared from explicit engineering
data such as near complete set of plans and specifications,
considered accurate within +15% or 5%.
CONSTRUCTION COSTS
The estimate prepared and presented here reflects an anticipated level
of confidence which would fall between categories 1 and 2 above, or accurate
within +40% or 20%. Table 38 outlines the project costs and Appendix B
contains more detailed documentation of cost information.
OPERATING COSTS
Table 39 shows a breakdown of operation and maintenance costs. It has
been assumed that the plant is manned continuously and that labor costs carry
a 50% overhead burden. No allowance has been nade for vehicles or resident
housing.
91
-------
TABLE 38. PLANT CONSTRUCTION COSTS
$ 6,152,000
145,000
358,000
638,000
3,091,000
1,290,000
230,000
731,000
97,000
562,000
68,000
815,000
184,000
1,679,000
582,000
75,000
$16,697,000
668,000
17,365,000
4,341,000
$21,706,000
1,408,000
310,000
54,000
75,000
1,193,830
TOTAL PROJECT COST $24,747,000**
*The general category includes all cleaning and grubbing of stumps, brush, etc.,
after the logging operation; all structural excavation for all facilities and
storage reservoir; all trench excavation for piping, etc; concrete for side-
walks, curbs and any other concrete not specific to a particular facilityall
asphaltic concrete for paving, parking spaces and driveways; miscellaneous
metals not specific to a particular facility, for example, handrails, etc.,
some allowance for equipment not specific to any facility for example, main-
tenance of grounds equipment; instrumentation controls, electrical, etc., not
specific to a particular facility; all yard piping, fittings, valves, etc.,
not specific to a particular facility; and drainage ditch for environmental
safety of nearby reservoir.
** osts are based on October 1978 dollars, Engineering News Record (ENR) index
for Seattle, Washington, 3194. Estimates do not include costs for land (812
acres), logging the proposed site, escalation during design and constructic3 ,
mileage allowances for workers or permit fees. Assumptions include bridges
and roads meet weight capacities needed to transport supplies and workers;
labor availability reasonable; scope of work as defthed; contractor avail-
ability reasonable; and good access on current roads to site.
General*
Landscaping
Headworks and Energy Dissipation
Flocculation Basins
Filtration Complex
Chemical Complex
Control Building
Sludge Disposal
Surge Clarifiers
Thickeners
Dewatering Complex
Pump Station
Reservoir, 5 MG
Backwash and Service Water
Electrical
Instrumentation and Control
Base
Data Logging Computer
Base Cost
Movein, Bond & Insurance (4%)
Subtotal
Contingency Allowance (25%)
TOTAL CONSTRUCTION COST
Final Engineering Design (6 % of Construction Cost)
Resident Inspection
Surveying
Soils Analysis
Sales Tax (5 % of Construction Cost)
92
-------
TABLE 39. ANNUAL PLANT OPERATING COSTS
1. Labor
Chief Operator & Chemist
Operators & Maintenance
2 @ $33,000
10 8 $27,000
= $ 66,000
= 270,000
$336,000
2. Chemicals
Alum
Lime
Polymer
Sodium bicarbonate
Sodium silicate
Chlorine
Fluoride (25% strength)
912 ton/yr 8 $100/ton
759 ton/yr 8 $57/ton
383,250 lb/yr @ $0.70/lb
686 ton/yr 8 $230/ton
365 ton/yr 8 $330/ton
137 ton/yr 8 $270/ton
461 ton/yr 8 $83/ton
= $ 91,200
= 53,600
= 268,300
= 157,800
= 120,400
= 37,000
= 38,750
767,000
3. Power (Average 72,400 kilowatthours/month)
4. Maintenance and Repairs
22,900
Assume 3%/yr of initial cost of major equipment
ANNUAL OPERATING COST
86,100
$1,212,000
TOTAL ANNUAL COSTS
The amortized project costs plus the annual operating costs are presented
below as the estimated total annual costs in October 1978 dollars for the Tolt
treatment plant.
20
CR (P L) (crf i% n) + Li
where CR = Annual cost of capital recovery
P = Present worth
L = Salvage value at n years
crf i% n = Capital recovery factor at i% interest rate after n years
93
-------
assuming i = 7%
n = 40 years (yr)
L = 40% of first cost
CR = [ 24,747,000 0.4(24,747,000) (0.07501) + [ 0.4(24,747,000):] (0.07)
= 1,113,763 + 692,916
= $1,806,679
TOTAL ANNUAL COST = CAPITAL RECOVERY + ANNUAL OPERATING COSTS
= $1,806,679 + $1,212,000 = $3,018,679
or approximately $3,019,000/yr
94
-------
SECTION X
REFERENCES
1. National Interim Primary Drinking Water Regulations, Federal Register,
Vol. 40, No. 248; Wednesday, December 24, 1975; pgs 5956659574; and
Background Used to Develop Regulations, EPA-570/9-76003, Office of
Water Supply, U.S. Environmental Protection Agency.
2. Seattle Water Department, City of Seattle, Final Environmental Impact
Statement for the Proposed Seattle Corrosion Control Plan, SEPA Entry
No. 1246, November 14, 1978.
3. Black and Veatch Consulting Engineers. Direct Filtration of Lake
Superior Water for Asbestiform Fiber Removal, EPA670/275-050a,
U.S. Environmental Protection Agency, Cincinnati, Ohio, June, 1975.
4. Hallenbech, W. H. and C. S. Hesse. A Review of the Health Effects
of Ingested Asbestos, Reviews on Environmental Health, Vol. II, No. 3,
1977.
5. Levy, Barry S., Eunice Sigurdson, Jack Mandel, Emaline Laudon and John
Pearson, Investigating Possible Effects of Asbestos in City Water:
Surveillance of Gastrointestinal Cancer Incidence in Duluth, Minnesota,
2 xnerican Journal of Epidemiology, Vol. 103, No. 4, 1976.
6. Wigle, D. T., Cancer Mortality in Relation to Asbestos in Municipal
Water Supplies, Archives of Environmental Health, Vol. 32, pgs 185190
1977.
7. Harrington, J. Malcolm and Gunther F. Craun, J. Winter Meigs, Philip J.
Landrigan, John T. Flannery and Richard S. Woodhull. An Investigation
of the Use of Asbestos Cement Pipe for Public Water Supply and the
Incidence of Gastrointestinal Cancer in Connecticut 1935-1973, P merican
Journal of Epidemiology, Vol. 107, No. 2, pgs 96103, 1978.
8. Kanarek, Marty Steven, Draft Interim Report entitled, Asbestos in
Drinking Water and Cancer Incidence, University of California, Berkeley,
1978.
9. EPA Position Paper on Draft Interim Report entitled, Asbestos in Drinking
Water and Cancer Incidence (see Reference #8), September 26, 1978.
95
-------
10. National Academy of Sciencies, Drinking Water and Health, Safe Drinking
Water Committee, 1977 (Library of Congress #77089284), prepared at the
request of and funded by the U.S. Environmental Protection Agency,
Contract No. 68013139.
11. Trussell, R. Rhodes, Application of Treatment Technology, prepared for
U.S.E.P.A. Environmental Research Center Technology Transfer Seminar
on Designing and Upgrading Drinking Water Systems, Portland, Oregon,
May 2526, 1977.
12. American Public Health Association, American Water Works Association
and Water Pollution Control Federation, Standard Methods for the
Examination of Water and Wastewater, Fourteenth Edition, 1975.
13. Millette, James R. Analyzing for Asbestos in Drinking Water. News
of Environmental Research in Cincinnati, Municipal Environmental Research
Laboratory, U.S.E.P.A., January 16, l976
14. Weber, Walter J., Jr. Physiochemical Processes for Water QualityControL
Wiley Interscience, A Division of John Wiley & Sons, Inc. New York,
1972, Library of Congress Catalog Card No. 7737026.
15. Anderson, C. H. and J. M. Long, Preliminary Interim Procedure for Fibrous
Asbestos, Transmission Electron Microscopy Method, EPAAnalyticalthanistry
Branch, Athens, Georgia, July, 1976.
16. Bellar, EPA/Bellar Purge and Trap Method for THM Analysis.
17. Sawyer, Clair N. and Perry L. McCarty, Chemistry for Sanitary Engineers,
Second Edition, McGrawHill Book Company, New York, 1967, Library of
Congress Catalog Card No. 6720179.
18. Kirner, John C., Regional Engineer, Washington State Department of Social
and Health Services, Memorandum, Subject: Proposed Criteria for the 1 ccept-
ance of Direct Filtration Water Treatment Plants, To: Water Supply and
Waste Section Operations Staff, February 16, 1977.
19. American Association of Cost Engineers, The Cost Engineers Notebook,
Section AA4.000, pg 10, January, 1978.
20. Grant, Eugene L. and W. Grant Ireson. Principles of Engineering Economy,
Fourth Edition, The Ronald Press Company, New York, 1964, Library of
Congress Catalog Card No. 6466236.
96
-------
SECTION XI
GLOSSARY
A/c Asbestos cement
alum - Aluminum sulfate.
amp Amperes.
A Area of paddles. 2
BW - Amount of backwash water used, 200 gal/ft /backwash.
°c Degree Centrigrade.
CC Dual media with coarse coal.
CMM Mixed media, sand MS-lB.
Cd Coefficient of drag.
cm Centimeter.
CR - Capital recovery.
crf Capital recovery factor.
EDS - Energy dispersive analysis system.
EN Electron microscope.
EPA U.S. Environmental Protection Agency.
FC Dual media with fine coal.
ft Foot or feet.
ft 2 Square foot.
ft/sec Foot per second.
G - Velocity gradient.
gal Gallon.
GI 2 Gastrointestinal.
gpd/ft Gallons per day per square foot.
gpm Gallons per minute.
gpm/ft 2 Gallons per minute per square foot.
GT Velocity gradient times the time.
H Headloss in feet.
i Interest rate.
in - Inch
ISV Interface settling velocity.
L Salvage value. 2
LR - Average filter loading rate, gpm/ft
- Pound.
4tX 0 }- Pounds per 1000 gallons.
m Meter.
m 2 Square meter.
m 3 Cubic meter.
MG Million gallon.
MGD Million gallons per day.
97
-------
mg/i Milligram per liter.
mm - Minute.
mm - Millimeter.
MM Mixed media, sand MS6.
mmHg Millimeter of mercury.
n Number of years.
N/A - Not applicable.
NBW Number of backwashes per 24 hours.
ND - Not detectable.
NP - Net water produced, gal/ft 2 /24 hours.
NSS Not statiscally significant.
NTU - Nephelometric turbidity units.
P Water horsepower.
PF Power factor.
pH Negative logarithm of the hydrogen ion concentration.
R Correlation coefficient.
RPM - Revolutions per minute.
Se Standard error of estimate.
sec Seconds.
sec 1 Seconds 1 .
T - Time.
THM - Trihalomethane.
ton/yr Ton per year.
Tq - Torque.
TURB - Turbidity.
A Absolute viscosity.
UFRV - Unit Filter Run Volume.
UW - University of Washington.
Relative velocity of paddles in fluid.
V Volume of basin.
yr Year.
/7 Mass fluid density, slugs/ft 3 .
Percent.
98
-------
SECTION XII
APPENDIX A
Ai RESULTS OF ASBESTIFORM P NALYSES ON STREAMS FEEDING THE SOUTH FORK
TOLT RESERVOIR.
A-2 COMPLETE CHEMICAL ANALYSIS OF SOUTH FORK TOLT RIVER WATER SUPPLY.
A-3 WATER PRODUCTION EFFICIENCIES FOR SELECTED FILTER RUNS.
A-4 SUMMARY OF RAW AND FILTERED WATER ASBESTIFORM COUNTS.
99
-------
APPENDIX A-i. RESULTS OF ASBESTIFORM ANALYSES ON STREAMS FEEDING THE SOUTH
FORK TOLT RESERVOIR.
Sample Point Descriotion
TW 1 CONSULTANT CREEK Consultant Creek; east bank, 100 upstream
from Spur 70 culvert; R 9E, T 26N, Sec 29,
swΌ.
Lat. 470, 4215; Long. 121°, 4115.
TW 2 RAINBOW CREEK Rainbow Creek, west bank, 5 upstream from
Spur 70 culvert; R 9E, T 26N, Sec 29, SEΌ.
Lat. 27°, 4214; Long. 121°, 4O18.
TW 3 HORSESHOE CREEK Horseshoe Creek; east bank, 15 upstream
from Spur 70 culvert; R 9E, T 26N, Sec 28,
SEΌ.
Lat. 470, 4216; Long. 121°, 3917
TW 4 SOUTH FORK TOLT South Fork Tolt River; north bank, 50
RIVER upstream from Spur 70 bridge; R 9E, T 26N,
Sec 25, swΌ.
Lat. 470 4225.
TW 5 PHELPS CREEK Phelps Creek; north bank, 75 upstream from
Spur 50 bridge; T 9E, T 26N, Sec 25, SWΌ.
Lat. 47°, 4218; Long. 1210, 3602.
TW 6 SKOOKUM CREEK Skookum Creek; east bank, 25 upstream from
Spur 50 bridge; R 9E, T 26N, Sec 26, SWΌ.
Lat. 470, 42O4; Long. 121°, 3724.
IW 7 SIWASH CREEK Siwash Creek; west bank, 10 upstream frcm
Spur 50 sulvert; R 9E, T 26N, Sec 34, NWΌ.
Lat. 47 , 42OO; Long. 1210, 3821.
isq 8 DOR YPHY CREEK Dorothy Creek, west bank, 10 upstream from
Spur 50 culvert; R 9E, T 26N, Sec 33, NWΌ.
Lat. 470, 415O; Long. 121°, 3947.
TW 9 CRYSTAL CREEK Crystal Creek; west bank, 75 south of Spur
50, above concrete dam; R 9E, T 26N, Sec 32,
swΌ.
Lat. 470, 4129; Long. 121°, 4108.
TW 10 SOUTH FORK TOLT
RESERVOIR
100
-------
APPENDIX A-i (CONTINUED)
1 ESULTS
OF ANALYSES*
Chrysotile
Possible
Sample
Lower
Date
Location
fibers/liter
Amphibole
fibers/liter
Vol.,
ml
Limit of
Detection
Comments
June 15, 1975 TW 1 5 x 1O 4 2.5 x
TW2 ND ND 5xl0 4
TW3 ND ND 2.5x10 4
TW4 ND ND 2.5x10 4
TW5 ND ND 2.5x 10 4
TW6 ND ND 2.5x10 4
2.5x10 4 2.5x10 4
4 4
TW8 2.SxlO 2.5x10
TW 9 ND ND 100 2.5 x l0
6 5 Preserved
July 29, 1975 TW 10 1.3 x io6 1.7 x 10 50 1 x 10 with HgCl .
6 6 Not
TW 10 1.3 X 10 1.9 X 10 50 1 x 10 Preserved.
Sept.28, 1975 TW 1 2.3 x 10 2.8 x 10 200 2.5 x 1O
TW 2 2 x lO 1.5 x 10 100 5 x 1O 4
P 3 ND ND 100 5 x 1O
TW 4 5 x 1O 7.5 x 10 200 2.5 x 1O 4
TW 5 1.5 x 10 7.5 x 1O 200 2.5 x 1O 4
TW 6 ND ND 100 5 x 10
TW 7 ND ND 200 2.5 x l0
TW 8 7.5 x 1O ND 200 2.5 x
TW 9 7.5 1.3 x 10 200 2.5 x 1O 4
TW 10 1.6 x io 6 x 1o 5 50 1 x
Cubetainer
Blank ND ND 1000 5 x 1O 3
*Analyses performed by Mr. Jack Murchio, University of California at Berkeley.
ND = Not Detected.
101
-------
APPENDIX A-2. COMPLETE CHEMICAL ANALYSIS OF SOUTH FORK TOLT RIVER W7 L R SUPPLY -
JANUARY 20, 1978*
Raw Treated
Alkalinity, Total (as CaCO 3 ) 4.5 1.4
Alkalinity, bicarbonate (as CaCO3) 45 1.4
Aluminum .21 .21
Barium <0.015 <0.03
Cadmium <0.002 <0.002
Calcium 2.9 3.3
Carbon Dioxide, free (calc.)t 3.0 -
Chloride 1.8 2.6
Chromium <0.006 <0.006
Color, standard units 18 18
Copper <0.008 <0.008
Fluoride <0.1 1.01
Hardness (as CaCO3) 8.0 9.0
Hardness, grains per gallon 0.47 0.53
Iron 0.18 0.20
Lead <0.015 <0.015
Lithium, g/l <0.15 <0.15
Magnesium 0.50 0.43
Manganese 0.013 0.005
Mercury, inorganic 1eachable,,j g/1 <0.1 <0.08
Mercury, total,,qg/l <0.05 <0.05
Nitrogen Nitrate (as N) 0.13 0.45
Nitrogen Ammonia (as N) <0.01 <0.01
Nitrogen Organic (as N) 0.065 0.04
Nickel <0.01 <0.01
Oxygen, Dissolved 13.0 12.4
Oxygen, % of saturation 101 105
pH 6.65
Phosphorus, Total Orthophosphate <0.002 0.002
Phosphorus, Filtrable Orthophosphate <0.002 <0.002
Phosphorus, Acid Hydrolysable 0.011 0.014
Potassium 0.27 0.22
Residue, Total 24 23.5
Residue, Nonfiltra.ble i
Silica, Reactive 3.8 4.2
Silver <0.001 <0.001
Sodium 1.15 1.02
Specific Conductance, micromhos 24 27½
Strontium 0.009 0.010
Sulfate 1.9 1.95
Temperature, °C 4 6
Turbidity, NTU 3.8 3.1
TanninLignin 0.25 0.20
Zinc <0.004 0.008
*Results in mg/i, except as noted 0
tCalculated.
102
-------
APPENDIX A-3. WATER PRODUCTION EFFICIENCIES FOR SELECTED FILTER RUNS.
Run #
Column
Maximum Average
Filter Filter
Loading Loading
Rate Rate
gpm/ft 2 gpm/ft 2
UFRV
gal
ft 2 run
Net Water
Produced
per
24 Hours
Filter
Efficiency
84
MM 6
6
3600
7944
94.4
85
MM 8
8
2880
10240
93.1
86
MM
6
6
4026
8017
95.0
87
MM
6
6
4464
8098
95.5
88
MM
6
6
4836
8121
95.9
90
MM
6
6
6480
8253
96.9
91
MM
8
8
2400
9984
91.7
95
MM
6
6
3960 8007
95.0
96
MM
8
8
2880 10240
93.1
98
MM
6
6
3600
7944
94.4
99
MM
8
8
2880
,
10240
93.1
101
MM
6
6
1
3960
8007
95.0
102
MM
8
8
2880
10240
93.1
104
MM
6
6
4320
8060
95.4
105
MM
8
8
3840
10560
94.8
109
110
MM
MM
6
6
6
6
2880
4298
7770
8057
93.1
95.3
j
I
112
MM
6
6
5040
8143
96.0
123
MM
6
5.8
5905
7930
96.6
124
MM
6
5.2
6500
7100
96.9
125
MM.
7
4.5
5152
6171
96.1
I
127
MM
7
6.8
4080
9072
95.1
128
MM
8
6.5
6240
8914
96.8
137
MM
6
5.6
7405
7768
97.3
138
MM
7
6.7
4020
8930
95.0
138
FC
7
6.5
4680
8760
95.7
139
MM
7
6.5
4303
8739
95.4
139
FC
7
6.5
3900
8640
94.9
139
CC
7
6.25
4875 8400
95.9
140
MM
8
7.1
5112 9470
96.1
(continued)
103
-------
APPENDIX A3 (CONTINUED).
Ufl #
Column
Maximum
Filter
Loading
Rate
gpm/ft 2
Average
UFRV
Filter
Loading
gal
Rate
gpm/ft 2 ft 2 run
Net Water
Filter
Produced
I Efficiency
per
24 Hours
140
140
FC
CC
8
8
7.4
7.3
4440
5690
9888
9872
95.5
96.5
141
141
141
MM
FC
CC
6
6
6
5.9
5.7
6.0
5568
4446
5760
7902
7608
8190
96.4
95.5
96.5
145
145
145
MM
FC
CC
6
6
6
5.4
4.5
4.7
5184
6750
7050
7355
6223
6508
96.1
97.0
97.2
147
147
147
MM
FC
CC
8
8
8
7.7
6.3
6.1
6060
6048
7686
10520
8630
8451
96.7
96.7
97.4
153
153
153
MM
FC
CC
10
10
10
7.3
6.8
6.0
5256
S712
5040
9893
9248
8143
96.2
96.5
96.0
156
156
156
MM
FC
CC
10
10
10
6.9
7.1
6.6
5796
5964
5544
9390
9672
8991
96.5
96.6
96.4
157
157
157
MM
FC
CC
10
10
10
6.3
5.2
4.0
6048
6240
6720
8630
6820
5537
96.7
96.8
97.0
161
161
161
MM
FC
CC
10
10
10
6.7
6.5
6.1
6432
6280
5856
9197
8916
7909
96.9
96.6
96.6
174
174
174
MM
CMM
CC
10
10
10
6.6
6.7
6.7
7128
7236
7638
9115
9257
9257
97.2
97.2
97.4
(continued)
104
-------
APPENDIX A-3 (CONTINUED)
Run *
Column
Maximum
Filter
Loading
Rate
gpm/ft 2
Average
Filter
Loading
Rate
gpm/ft 2
UFRV
al
ft -run
Net Water
Produced
per
24 Hours
Filter
Efficiency
%
153
MM
10
7.6
5016
10253
96.0
153
FC
10
7.3
4380
9769
95.4
153
CC
10
6.8
3672
8976
94.5
156
MM
10
7.5
4500
10050
95.5
156
FC
10
7.6
4104
10096
95.1
156
CC
10
7.5
3825
9925
94.8
157
157
MM
FC
10
10
7.4 3774
7.5 3600
9785
9862
94.7
94.4
157
CC
10
7.2
3024
9321
93.4
161
MM
10
7.8
5244
10540
96.2
161
FC
10
8.6
5418
11627
96.3
161
CC
10
8.7
5481
11768
96.3
174
MM
10
7.3
4270
9902
95.3
174
CMM
10
7.6
4560
10190
95.6
174
CC
10
7.5
3600
8925
94.4
NOTES:
* Runs
MM =
FC =
CC =
CMM =
terminated at 6
Finished Sample
Finished Sample
Finished Sample
Finished Sample
gpm/ft 2 instead of 10 headloss.
from Mixed Media Filter Column with MS-6 Sand.
from Dual Media Filter Column with Fine Coal
from Dual Media Filter Column with Coarse Coal
from Mixed Media Filter Column with MS-is Sand.
105
-------
APPENDIX A-4. SUMMARY OF
RAW AND FINISHED WATER ASBESTIFORM COUNTS.
Run #
Hour
Into
Th n
Turbidit
(NTU)
Raw (fibers/liter)
Finished (fibers/liter)
Removal
Amphibole
(106)
Chrysotile
(106)
Amphibole
(106)
Chrysotile
(106)
Amphibole
(%)
hrysotil
(%)
3R
3F
8
8
0.1
5.7
5.7
8.9
8.9
0.04(NSS)
0.09
99.4
99.0
4cR
4cF
5
5
1.4
0.1
3.31
331
5.12
5.12
O.05(NSS)
0.09
98.5
98.2
5cR
5cF
5
5
1.4
0.08
3.06
3.06
16.39
16.39
c0.01(ND)
0.15
100
99.1
6R
6F
4
4
1.3
0.07
3.46
3.46
13.0
13.0
0.05(NSS)
0.15
98.6
98.8
11R
11F
6
6
1.15
0.26
4.33
4.33
13.29
13.29
0.42
1.64
90.3
87.7
12dR
12dF
7
7
1.0
0.09
1.76
1.76
13.14
13.14
O.01(NSS)
0.13
99.4
99.0
21R
21F
21F
21F
21F
21F
7
2
6
7
8
12
0.66
0.065
0.06
0.34
0.07
0.059
2.18
2.18
2.18
2.18
2.18
2.18
25.8
25.8
25.8
25.8
25.8
25.8
0.0l(NSs)
0O1(NSS)
0.72(NSS)
<0.01(ND)
<0.O1(ND)
0.16
0.16(NSS)
12.25
0.19
0.09
99.5
99.5
67.0
100.0
100.0
99.4
99.4
52.5
99.3
99.7
24R
24F
24F
24F
7
6
7
13
0.60
0.085
0.36
0.062
2.4
2.4
2.4
2.4
9.4
9.4
9.4
9.4
O.04(NSS)
O.6(NSS)
O.04(NSS)
0.34
6.
0.13
98.3
75.0
98.3
96.4
72.3
98.6
29R
29F
29F
10
10
17
0.62
0.090
0.10
0.94
0.94
0.94
4.25
4.25
4.25
O.01(ND)
O.O1(ND)
0.07
0.22
100.0
100.0
98.4
94.8
(continued)
106
-------
APPENDIXA-4 (CONTINUED).
Run
Hour
Into
Run
bd
r 1 1
(NTU)
Raw (fibers/liter)
Finished (fibers/liter
Removal
Amphibole
(106)
Chrysotile
(106)
Amphibole
(106)
Chrysotile
(106)
Amphibole
(%)
hrysotilE
(%)
33R
33F
33F
6
6
9
0.61
0.062
0.053
0.65
0.65
0.65
3.82
3.82
3.82
<0.01(N0)
0.0l(ND)
0.06(NSS)
0.22
100
100
98.4
94.2
44R
44F
44F
11
11
13
0.56
0.042
0.09
0.90
0.90
0.90
2.8
2.8
2.8
0.0l(ND)
<0.01(ND)
0.26
<0.01(ND)
100.0
100.0
90.7
100.0
51R
51F
51F
9
3
3
0.54
0.07
0.07
0.29(ND)
ND
ND
8.4
8.4
8.4
<0.01(ND)
0.01(ND)
0.06(NSS)
0.20
99.3
97.6
53R
53F
53F
9
3
9
0.50
0.13
0.24
0.70(NSS)
0.70
0.70
3.6
3.6
3.6
-------
APPENDIX A-4 (CONTINUED).
Run #
Hour
Into
bdt
(NTU)
Raw (fibei/liter)
inished (fthers/liter
Removal
Amphibole
(106)
Chrysotile
(106)
Amphibole
(106)
Chrysotile
(106)
Amphibole
(%)
hrysotih
( )
93R
93MM
33MM
33MM
93MM
93F
18
3
12
14
17
18
0.38
0.098
0.10
0.065
0.074
0.08
<0.07 (ND)
ND
ND
ND
ND
ND
3.6
3.6
3.6
3.6
3.6
3.6
0.02(ND)
<0.0l(ND)
0.0l(ND)
<0.0l(ND)
<0.0l(ND)
0.09(NSS)
0.53
0.07
0.18
0.04(NsS)
97.5
85.3
98.9
98.9
1O8R
108MM
108MM
108MM
108MM
108F
4
2
10
15
16
18
0.55
0.12
0.073
0.081
0.068
0.082
c0.14(ND)
ND
ND
ND
ND
ND
3.61
3.61
3.61
3.61
3.61
3.61
<0.02(ND)
0.0l(ND)
<0.0l(ND)
0.0l(ND)
<0.01(ND)
0.34
0.03(NsS)
0.14
0.07(NSS)
0.13
90.6
99.2
96.1
98.1
96.4
111R
111MM
111MM
111MM
111MM
111MM
111MM
flF
12
7
12
16
17
18
19
19
0.85
0.084
0.19
0.082
0.09
0.10
0.22
0.072
c O.14(ND)
ND
ND
ND
ND
ND
ND
ND
4.62
4.62
4.62
4.62
4.62
4.62
4.62
4.62
<0.01(ND)
0.02(ND)
<0.01(ND)
0.01(ND)
0.02(ND)
<0.02(ND)
<0.01(ND)
0.09
0.07(Nss)
0.07(NsS)
0.03(Nss)
0.07(NSS)
0.15
0.04(NsS)
98.1
98.5
98.5
99.4
98.5
96.8
99.1
.20R
20R
.20MM
.20MM
120MM
.20MM
.20MM
.20MM
.20MM
.20MM
.20MM
.20MM
.20MM
20MM
9
15
1
2
6
9
12
13
15
16
17
18
19
20
3.3
3.4
0.14
0.096
0.08
0.065
0.06
0.07
0.062
0.071
0.071
0.115
0.14
0.28
0.19(NsS)
<0.10( 183)
0. 19(Nss)
0.19(NSS)
0.19(NSS)
0.19(NsS)
0.19(NSS)
O. 19(NS S)
O.19(NsS)
0.19(NsS)
0.19(NsS)
0.19(NsS)
0.19(NSS)
0.19(Nss)
5.38
10.1
5.38
5.38
5.38
5.38
5.38
5.38
5.38
5.38
5.38
5.38
5.38
5.38
0.02(ND)
O.02(ND)
-------
APPENDIX A-4 (CONTINUED).
Run #
Hour
Into
Run
Turbidt
(NTU)
Raw (fibers/liter)
inished (fthers/liter)
Removal
Amphibole
(106)
Chrysotile
(l06
Amphibole
(106)
Chrysotile
(106)
Amphibole
( )
Chrysotil
(%)
120MM
120MM
120MM
21
22
23
0.48
0.57
1.2
0.19(NSS)
0. 19(NSS)
0.19(Nss)
5.38
5.38
5.38
&0.04(ND)
< .0.07(ND)
<0.08(ND)
0.28
0.57
1.25
94.8
89.4
76.8
135R
135MM
135MM
135MM
135MM
135MM
135MM
135MM
135MM
135MM
135MM
135MM
9
0
1
2
3
4
11
13
21
22
23
24
1.8
0.28
0.14
0.10
0.096
0.090
0.074
0.11
0.065
0.089
0.072
0.065
<0.07(ND)
ND
ND
ND
MD
ND
ND
ND
ND
ND
ND
ND
3.9
3.9
3.9
3.9
3.9
3.9
3.9
3.9
3.9
3.9
3.9
3.9
<0.02(ND)
<0.02(ND)
0.01(ND)
<0.0l(ND)
<0.0l(ND)
<0.01(ND)
<0.0l(ND)
(0.01(ND)
<0.0l(ND)
<0.0l(ND)
-------
APPENDIX A-4 ( CONTINULD) .
Run *
FlOUt
Into
Run
Turbidit
NTU)
Raw (fibers/liter)
Finished (fbers/liter)
Removal
Amphibole
(106)
Chrysotile
(106)
Amphibole
(106)
Chrysotile
(106)
Amphibole
(%)
:hrysotile
(%)
74MM
2
0.079
<0.Ol(ND)
O.03(NSS)
98.4
74MM
3
0.078
<0.Ol(ND)
<0.Ol(ND)
100.0
74NM
4
0.075
-
<0.Ol(ND)
0.Ol(NSS)
99.4
74MM
5
0.079
<0.Ol(ND)
O.0l(NSS)
99.4
.74MM
6
0.072
0.0l(ND)
0.0l(NSS)
99.4
.74NM
7
0.075
(0.Ol(ND)
0.02(NSS)
98.9
.74MM
14
0.078
c0.Ol(ND)
<0.Ol(ND)
100.0
74MM
18
0.075
<0.Ol(ND)
<0.0l(ND)
100.0
.74CMM
0
0.095
<0.0l(ND)
0.09
95.1
.74CNN
1
0.070
c0.Ol(ND)
0.02(NSS)
98.9
74CMM
1.5
0.21
<0.02(ND)
0.94
48.9
74CMM
2
0.068
cO.Ol(ND)
<0.Ol(ND)
100.0
.74CNN
3
0.070
(0.Ol(ND)
0.03(NSS)
98.4
174CNN
4
0.071
(0.01(ND)
0.01(NSS)
99.4
L74CMM
5
0.062
<0.01(ND)
0.06(NSS)
96.7
74CNN
6
0.089
0.01(ND)
0.02(NSS)
98.9
74CNN
7
0.075
<0.0l(ND)
0.01(NSS)
99.4
.74CNN
8
0.070
(0.01(ND)
0.Ol(NSS)
99.4
.74CNN
14
0.065
0.0l(ND)
0.01(NSS)
99.4
.74CNN
18
0.079
0.01(ND)
0.05(NSS)
97.3
74CC
0
0.097
0.01(ND)
0.Ol(NSS)
99.4
.74CC
1
0.07
0.01(ND)
0.04(NSS)
97.8
.74CC
1.5
0.20
0.03(ND)
C0.03(ND)
100.0
.74CC
2
0.070
<0.Ol(ND)
0.0l(NSS)
99.4
L74CC
3
0.071
<0.0l(ND)
0.01(ND)
100.0
174CC
4
0.090
0.Ol(ND)
0.0l(ND)
100.0
174CC
5
0.071
C0.Ol(ND)
0.0l(ND)
100.0
L74CC
6
0.081
C0.0l(ND)
0.0l(ND)
100.0
L74CC
7
0.068
<0.Ol(ND)
0.06(NSS)
96.7
174-CC
8
0.070
0.0l(ND)
<0.0l(ND)
100.0
L74CC
14
0.070
0.0l(ND)
C0.Ol(ND)
100.0
174CC
19
0.070
<0.Ol(ND)
<0.Ol(ND)
100.0
NOTES:
Raw Sample
Finished Sample
Finished Sample
Finished Sample
Finished Sample
Finished Sample
None Detected
Not Statistically Significant (4 or fewer fibers actually counted)
R =
F =
MM =
CNN =
FC =
CC =
ND =
NSS =
from Waterboy
from Mixed Media Filter Column with MS6 Sand
from Mixed Media Filter Column with MS-18 Sand
from Dual Media Filter Column with Fine Coal
from Dual Media Filter Column with Coarse Coal
110
-------
SECTION XIII
INDEX OF UNATTACHED APPENDIXES
B-i PHOTOGRAPHS OF PILOT PLANT APPARATUS.
B-2 WEATHER AND OPERATING CONDITIONS AT THE SOUTH FORK TOLT RESERVOIR
AND TOLT REGULATING BASIN.
B-3 SUMMARY OF INFORMATION ON GRANULAR MEDIA FILTERS.
B-4 MICROGRAPHS OF ASBESTOS FIBERS.
B-5 PARTICLE COUNT DATA.
B6 FILTER RUN PILOT TESTING AT HIGH RAW WATER TURBIDITIES.
B-7 SLUDGE SETTLEABILITY TESTING.
B-8 DETAILED DOCUMENTATION OF COST INFORMATION.
C CONDITIONS SURROUNDING EACH FILTER RUN.
D PLOTS OF HEADLOSS AND TURBIDITY VS. TIME FOR EACH FILTER RUN.
ii]
-------
TECHNICAL REPORT DATA
(Please read Inuructions on the reverse before completing)
1. REPORT NO. 2.
EPA600/ 279125
3. RECIPIENTS ACCESSIOI*NO.
4. TITLE AND SUBTITLE
SEATTLE TOLT WATER SUPPLY MIXED ASBESTIFORM
REMOVAL STUDY
5. REPORT DATE
August 1979 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Gregory J. Kirmeyer
8. PERFORMING ORGANIZATION REPORT NO.
. . PERFORMING ORGANIZATION NAME AND ADDRESS
Water Quality Division
Seattle Water Department
1509 South Spokane Street
Seattle,_Washington__98144
10. PROGRAM ELEMENT NO.
1CC614 Sos 1 Task 06
11. CONTRACT/GRANT NO.
R804422
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research LaboratoryCin, ,,OH
Office of Research & Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final 5/76 - 11/78
14.SPONSORING AGENCY CODE
PA 6 0 14
15.SUPPLEMENTARYNOTES Appendices B and C, EPA600/279153, are supplementary to this
main report. Additional supplementary material is in Appendix D, EPA600/279126.
Project Officer: Gary Logsdon 513/6847345.
l 6 ABSTRACTF 0 r 1 1/2 years the Seattle Water Department conducted direct filtration
pilot plant studies at the Tolt Reservoir, obtaining data on techniques to remove
amphibole and chrysotile asbestos from drinking water. Research showed that filtered
water turbidity should be 0.1 ntu or lower in order to effectively remove fibers.
Flocculation was necessary but sedimentation was not. Amphibole fibers are more
readily removed than chrysotile, but both types could be reduced to below detectable
limits or to not statistically significant counts by treatment with alum, lime and a
filter aid (nonionic or anionic polymer); or alum, cationic polymer and a filter aid;
or cationic polymer and a filter aid. Asbestos fiber content of filtered water
increased sharply when filtered water turbidity rose above 0.1 ntu because of filtra-
tion rate changes, interruption of chemical feed, or turbidity breakthrough associated
with the end of the filter rue. Asbestos fibers in the concentrations encountered in
this study (raw up to 20 x 10 f/L, filtered down to 0.01 x 10 f IL) can not be
detected by a turbidimeter; however, the association of rising fiber counts and
turbidities in filtered water would enable a plant operator to estimate fiber removal
by observing turbidity if the filter was operated in the manner done in this work.
17. KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OFEN ENDED TERMS
C. COSATI Field/Group
Asbestos, Coagulation, Electron
microscopes, Filtration, Pilot plants,
Potable water, Turbidity, Water
treatment
Seattle, Washington
Tolt Reservoir
Fiber removal,
Chrysotile, Amphibole,
Direct filtration,
Flocculation
13 B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
126
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73) US GOVSRtIMSI(T rnNGOFF1cE 1979 -657-060/5451
------- |