Filter ripening sequence reduction by physical and chemical variation of backwashing
by Kelly Orville Cranston
A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in
Environmental Engineering
Montana State University
© Copyright by Kelly Orville Cranston (1987)
Abstract:
The period of initial effluent quality degradation from water filtration systems is known to reduce the
overall water quality produced by a filter plant. A limited amount of work has been conducted in the
past to describe this phenomenon and to develop methods to reduce it. The research undertaken for this
thesis was intended to further describe the mechanisms of the initial effluent degradation period and to
investigate alternative methods of reducing it.
The research project utilized a dual media, in-line pilot filtration plant with varied primary coagulants
and raw water sources. The effects of various coagulants injected into the backwash water and the
variation of several physical aspects of backwashing on the initial effluent degradation periods were
investigated. From the data gathered in this research a more comprehensive theory concerning the
mechanisms and timing of events occuring in the initial period of degradation has been developed.
The following generalizations concerning the results can be made: 1. The backwash coagulant yielding
the best results was generally the same as the primary coagulant system.
2. The optimum time of injection of this coagulant into the backwash water corresponded to the time
required to completely displace the backwash water into the filter unit.
3. The backwashing volume required to minimize the initial degradation period is that required to
displace the retained particles of filtration out of the filter unit.
4. Variation of the remnant volume above the media does not affect the magnitude of the initial period
of degradation, only the timing at which events occur. In systems utilizing backwash coagulants,
increasing this volume can enhance the effects of the backwash coagulants.
5. Incremental opening of a filter unit can substantially reduce the magnitude of the initial period of
degradation when compared to an instantaneously opened filter. FILTER RIPEETING SEQUENCE REDUCTION BY
PHYSICAL AND CHEMICAL VARIATIONS
OF BACKWASHING
by
Kelly Orville Cranston
A thesis submitted in partial fulfillment
of the requirements for the degree
of
Master of Science
in
Environmental Engineering
MONTANA STATE UNIVERSITY
Bozeman, Montana
January, 1987
■
main u b .
ii
APPROVAL
of a thesis submitted by
Kelly Orville Cranston
This thesis has been read by each member of the thesis
committee and has been found to be satisfactory regarding
content, Engl i s h usage, format, citations, bibliographic
style, and consistency, and is ready for submission to the
College of Graduate Studies.
Date
Chairperson, Graduate Committee
Approved for the Major Department
Approved for the College of Graduate Studies
iii
STATEMENT OF PERMISSION TO USE
In presenting this thesis in partial fulfillment of the
requirements
for
a master's
degree
at
Montana
State
University, I agree that the Library shall make it available
to borrowers under rules of the Library.
from this thesis are allowable
without
Brief quotations
special
permission,
provided that accurate acknowledgement of source is made.
Permission for extensive quotation from or reproduction
of this thesis may be granted by my major professor,
his
absence,
by
the
Director
of
Libraries
when,
or in
in
the
opinion of e i t h e r , the proposed use of the material is for
scholar Iy purposes.
Any copying or use of the material in
this thesis for financial gain shall not be allowed without
my written permission.
Signature
Date
TABLE OF CONTENTS
Page
APPROVAL........
ii
STATEMENT OF PERMISSION TO USE........................
TABLE OF CONTENTS...... .................... . . :...... .
LIST OF TABLES.........................
LIST OF FIGURES........................ ........ ........
ABSTRACT....................................
iii
iv
vii
viii
xi
CHAPTER
1.
INTRODUCTION......
I
2.
RESEARCH OBJECTIVES.............................
4
3.
FILTER RIPENING:
A LITERATURE REVIEW..........
6
Backwash...........
Filter Ripening Theory.......................
Filter Media Preconditioning During Backwash
6
8
13
4.
EXPERIMENTAL RESEARCH...................
Experimental Methods.........................
Experiments Conducted . . ......................
A.
Polymer as Primary Coagulant......
1.
Polymer in Backwash Wat e r ......
2.
Alum in Backwash Water..... .
3.
Variation of Backwash Volume...
4.
Variation of Injection Time....
5.
Polymer Overdosing Effects....
6.
Variation of Remnant Volume....
7. Operation o f .Pre-ripened Filter
B . Alum as Primary Coagulant..........
1. Polymer in Backwash Water.....
2. 'Alum in Backwash Water.........
3.
Alum/polymer in Backwash Water.
16
16
24
26
26
27
27
27
27
" 28
28
30
30
30
30
V
4.
5.
6.
7.
8.
9.
10.
11.
C.
D.
E.
5.
Variation of Injection Time....
Variation of Backwash Volume...
Variation, of Remnant Volume....
Backwash Effluent Turbidity....
Zeta Potential of Influent....
Operation of Pre-ripened Filter
Injection Above Media.........
Zeta Potential of Backwash
Effluent.....................
Alum/polymer as Primary Coagulant..
1. Polymer in Backwash Water. .....
2. Alum in Backwash Water.......
3.
Alum/polymer in Backwash Water.
Tracer Studies......................
1 . Variation of Remnant Volume....
2. Continuous Dye Injection at
Station C .......................
3.
Backwash Tracer of Pilot Plant.
4.
Tracers on Filters at Bozeman
Plant...........................
Pilot Studies at Bozeman Plant.....
1. Polymer in Backwash Water.....
2.
Alum in Backwash Water..... . . .
3. Alum/polymer in Backwash Water.
RESULTS AND DISCUSSION.................... .....
Development of a Revised Filter Ripening
Theory........................................
The Remnant Stage.............. '........
Influent Mixing and Particle
Stabilization Stage................ . . . .
Filter Media Conditioning Stage.......
Results And Evaluation of the Use of
Backwash Coagulants.... .....................
The Polymer Experiments' A-I and A - 2 . .. .
The Alum Experiments B- I , B- 2 , and B-3.
The Alum/polymer Experiments C - I , C-2,
and C - 3 ..... '___ '.............. .........
The Bozeman Water Treatment Plant Study
Experiments E-I , E-2 , and E - 3 ..... . . . .
Backwash Coagulant Summary..... .'......
Optimizing Backwash Coagulant Injection Time
Variation of Remnant Volume Above Med i a....
Variation of the Backwash Water Volume.....
6.
PRACTICAL APPLICATION SUMMARY................
Optimum Backwash Coagulants. .................
Optimum Injection Time of Backwash
Coagulants....................................
30
30
31
31
31
31
31
32
32
32
32
33
33
33
33
34
35
35
36
36
36
37
37
40
54
60
64
65
74
84
88
98
99
103
105
HO
HO
112
vi
O p t i m u m B a c k w a s h V o l u m e ......
Optimization of Remnant Volume..............
Incremental Filter Opening...........
7.
CONCLUSIONS........
112
113
113
115
REFERENCES CITED.......................................
119
SELECTED BIBLIOGRAPHY...................................
122
V
i
-
vii
/'
I
LIST OF TABLES
Table
Page
1.
Pilot Plant Filter Media Size Distribution...
16
2.
Bentonite Particle Size Distribution...........
18
I
N
sT
viii
LIST OF FIGURES
Figure
'
1.
Characteristics ofInitial Effluent
Quality.....
9
2.
Laboratory Pilot Filtration PlantSchematic........
17
3.
Particle Size Distribution for Min-u-sil 3 0 .....
19
4.
Actual Pilot Plant Effluent Turbidity Strip Chart
22
5.
Pilot Plant Schematic at Bozeman Water Treatment
Plant............................... ...............
25
Pilot Plant Filter Unit Measurements and
Detention times....................................
29
Proposed Characteristics of Initial Effluent
Degradation................... ;...................
38
Varia t i o n of -Remnant V o l u m e Above Filter
' Media.
Experiments A- 6, B- 6 and D - 1 ..........
42
Filter Ripening Sequence at Bozeman Water
Treatment Plant with Tracer Study................
45
Filter Ripening Sequence at Helena Water
Treatment Plant with Instantaneous and
Incremental Filter Opening.........................
49
6.
7.
8.
9.
10.
Page
11;■ Variation of Backwashing Volume.
Experiments
A-3 , B-5 , and B-7 ................................
12.
13.
14.
15.
51
Zeta Potential of Backwash Effluent Particles.
Experiment B - I l .................................
53
Zeta Potential of Influent Particles Compared
to Effluent Turbidity.
Experiment B - 8 ..........
56
Coagulant Injection Above and Below Filter Media.
Experiment B - I O ....... '.........................
59
Operation of Pre-ripened Filter with Polymer
as Primary Coagulant.
Experiment A -8............
61
ix
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
Operation of Pre-ripened Filter with Alum as
Primary Coagulant.
Experiment B - 9 .. ..............
Determination of Optimum Backwash Alum Dose in
' Polymer Primary System.
Experiment A-2 ..........■
62
66
Determination of Optimum Backwash Polymer Dose in
Polymer Primary System During Winter Season.
Experiment A-Ia.............................. -....
68
Determination of Optimum Backwash Polymer Dose in
Polymer Primary System During Summer Season.
Experiment A - I b ....................................
69
Determination of Optimum Backwash Polymer Dose in
Polymer Primary System During Autumn Season.
Experiment A-I c.... ...........
70
Backwash Coagulant Optimization Summary with
Polymer Primary Coagulant System...................
71
Polymer Overdosing of Backwash Water During
Winter Season. Experiment A - 5 ....................
73
Determination of Optimum Backwash Polymer Dose in
Alum Primary System During Summer Season.
Experiment B- 1 .........
75
Determination of Optimum Backwash Alum Dose in
Alum Primary System During Summer Season.
Experiment B - 2 .....................................
77
Variation of Remnant Volume Above Filter Media
with Alum Primary Coagulant. Experiment
B-.6.....
80
Determination of Optimum Backwash Alum/polymer
20:1 Dose in Alum Primary System During Summer
Season.
Experiment B - 3 ...........................
81
Backwash Coagulant Optimization Summary for Alum
Primary Coagulant System........ .................
83
Determination of Optimum Backwash Polymer Dose in
Alum/polymer 20:1 Primary System During Summer
Season.
Experiment C - I ...........................
86
Backwash Coagulant Optimization Summary for .
Alum/polymer 20:1 Primary Coagulant System........
87
x30.
Determination of Optimum Backwash Alum Dose in
Alum/polymer 20:1 Primary System During Summer
Season.
Experiment C - 2 .....................",.....
89
31. ' Determination of Optimum Backwash Alum/polymer
Dose in Alum/polymer 20:1 Primary System During
Summer Season.
Experiment C- 3............ ........
90
32.
33.
34.
35.
36.
37.
38.
39.
40.
Determination of Optimum Eackwash Polymer Dose in
Bozeman Water Treatment Plant Influent System
During Autumn Season.
Experiment E - I .... .......
92
Determination of Optimum Backwash Alum Dose in
Bozeman Water Treatment Plant Influent System
During Autumn Season.
Experiment E-2... .........
93
Backwash Coagulant Optimization Summary With
Bozeman Water Treatment Plant Influent System....
94
Determination of.Optimum Backwash Alum/polymer
20:1 Dose in Bozeman Water Treatment Plant
Influent System During Autumn.
Experiment E-3...
95
Optimization of Backwash Coagulant Injection Time
with Polymer as Primary and Backwash Coagulant
Systems.
Experiment A-4 .........................
100
Optimization of Backwash Coagulant Injection Time
with Alum as Primary and Backwash Coagulant
Systems.
Experiment B - 4 .........................
101
Summary of Optimization of Backwash Coagulant
Injection T i m e .....................................
102
Variation of the Volume of Backwash Water with
Polymer as Primary Coagulant.
Experiment A-3..
107
Variation of the Volume of Backwash Water with
Alum as Primary Coagulant.
Experiment B - 5 ......
108
xi
ABSTRACT
The period of initial effluent quality degradation from
water f i l t r a t i o n systems is known to reduce the o v e r a l l
water quality produced by a filter plant.
A limited amount
of work has been conducted in the past to describe this
p h e n o m e n o n and to d e v e l o p m e t h o d s to r e d u c e it.
The
research undertaken for this thesis was intended to further
describe the mechanisms of the initial effluent degradation
period and to i n v e s t i g a t e a l t e r n a t i v e methods of reducing
it .
The research project util i z e d a dual media, in-line
pilot filtration plant with varied primary coagulants and
raw water sources.
The effects of various coagulants
injected into the backwash water and the variation of
s e v e r a l physical aspects of backwashing on the initial
effluent degradation periods were investigated.
From the
data gathered in this research a more comprehensive theory
concerning the mechanisms and timing of events occuring in
the initial period of degradation has been developed.
The following generalizations concerning the results
can be made:
1.
The backwash coagulant yielding the best, results
was generally the same as the primary coagulant system.
2.
The o p t i m u m ti m e of i n j e c t i o n of th i s c o a g u l a n t
into the backwash water corresponded to the time required to
completely displace the backwash water into the filter unit.
3.
The tackwashing volume required to minimize the initial
degradation period is that required to displace the retained
particles of filtration out of the filter unit.
4.
Vari a t i o n of the remnant
volu m e above the media does
not a f f e c t the m a g n i t u d e of the i n i t i a l
p e r i o d of
degradation, only the timing at which events occur.
In
systems u t i l i z i n g backwash coagulants, increasing this
volume can enhance the effects of the backwash coagulants.
5.
Incremental opening of a filter unit can substantially
reduce the magnitude of the initial period of degradation
when compared to an instantaneously opened filter.
I
CHAPTER I
INTRODUCTION
Operators of municipal water f iltration plants have
long
recognized
turbidity
units.
a
brief
immediately
period
following
of
increased
backwashing
of
effluent
the
filter
Research has indicated that the period of initial
effluent
degradation
is a function
of the remnant water
remaining in the filter at the end of backwash [3] and/or a
function
of
the
influent
[7].
The subsequent
period of
effluent quality improvement, or "filter ripening" has been
related to the accumulation of influent particles within the
pores of the media resulting in increased capture of further
particles
that
an
[10,
12],
Additional research has also suggested
increased
transport
of
potentially
pathogenic
microorganisms through the filter unit [9] may be associated
with
the
initial
initial
period
subsequent
period
of
of
effluent
effluent
impro v e m e n t
is
to
degradation.
quality
be
The
degradation
termed
in
the
and
current
research as the "filter ripening sequence".
Several methods have been suggested and
researched
to
reduce or remove the period of poor quality effluent from
filter effluent.
The most important methods considered have
2
been
the addition
backwash
water
of p o l y e l e c t r o l y t e s
and
quality effluent.
"filtering
to
(polymers)
waste"
of
the
to the
poorest
The addition of polymers to the backwash
water has been suggested
as a means of "preconditioning"
the filter by adsorption of the polymer to the filter media
[5,
8,
15].
The
adsorbed
significant Iy reduce
polymers
have
been
shown
to
the effluent
turbidity in the initial
stages of filtr a t i o n [5, 8, 15].
The use of a "filter to
waste" period at the beginning of a filter run can also be
effective in restricting the initial period of poor quality
effluent.
However,
due to the significant length of the
filter ripening sequence, sometimes two hours or more [4],
the
"filter
consumption
reduce
to
period
of raw water.
or remove
developed
waste"
may
Therefore
research
the filter ripening
in a manner minimizing
induce
an
excess
conducted
to
sequence should
be
the consumption
of raw
water.
A literature review revealed that the use of backwash
coagulants other than polymers,
physical
or the variation of several
parameters of backwashing,
to reduce the magnitude
and duration of the filter ripening sequence, has not been
extensively
this
thesis
researched.
was
present knowledge.
The current
conducted
to
fill
experimental
some
of
the
work for
gaps
in
3
As a result of the experimental work conducted in the
current
research
project,
and from a review of previous
theories, a more encompassing hypothesis has been developed
on the mechanisms of the filter ripening sequence.
also
been
developed
concerning
the
variation
Data has
of physical
and chemical parameters of backwashing to determine their
importance in affecting this phenomenon.
som e
methods
have
bee n
determined
for
From this data,
the
optimized
reduction of the initial period of effluent degradation.
4
CHAPTER 2
RESEARCH OBJECTIVES
The o v e r a l l
further
affect
describe
the
the filter
media, in-line
this
objec t i v e
physical
ripening
filtration
phenomena will
of this research project is to
and
chemical
sequence
unit.
phenomena
which
of a deep-bed,
dual
A better
understanding
of
allow development of methods to reduce
the magnitude and duration of the filter ripening sequence.
The" individual research objectives are as follows:
1.
Develop
physical
a more
and
encompassing
chemical
hypothesis
phenomena which
to describe
define
the
the magnitude
and duration of the filter ripening sequence.
2.
Determine
the
most
effe c t i v e
backwash
coagulant
to
reduce the magnitude and duration of the filter ripening
sequence
used;
for each of the three primary c oagulant systems
polymer,
aluminum.
sulfate,
and
aluminum
s u l f a t e / p o l y m e r combination.
The optimum dosages of each
coagulant
for
will
be determined
Bozeman, Montana tap water.
The
each system using.the
seasonal variations of the
optimum dosages of backwash polymer will be described.
3.
Determine
the
optimum
time
for
the
injection
of
coagulant into the backwash water with respect to completion
5
of the backwashing phase.
4.
Determine the effect on the magnitude and duration of
the filter ripening sequence of changing the volume above
the filter
during
5.
media into which the remnant water is displaced
backwash.
Determine
the effect on the magnitude and duration of
the filter ripening sequence of changing the total volume of
backwash water
6.
used to backwash the filter at a given rate.
Link the results of the laboratory pilot plant study to
the Bozeman, Mon t a n a Water Treatment Plant in terms of the
reduction
ripening
of
the
sequence.
magnitude
and
duration
The pilot plant will
of
the
filter
be transported to
the BWTP so it may util i z e the same raw influent derived
direct Iy from the plant's flocculation units.
6
CHAPTER 3
FILTER RIPENING:
A LITERATURE REVIEW
Backwash
During the operation of a deep-bed filter unit, a point
is reached in which either the water head above.the filter
media has built Up to an excessive l e v e l , or an increase in
the effluent turbidity occurs.
At this time, the filter is
taken off line and backwashed.
The backwash is conducted by
r e v e r s i n g the flow of water through the filter in order to
remove the particulate material and chemical coagulant floes
held within the filter media.
been shown by Amirtharajah
This backwash process has
[1] to be
most efficient for
removal of attached particles from the media when the media
bed is f luidized from 30 to 50%.
when
the
drag
sufficient
to
force
of
suspend
position of expansion.
the
the
water
media
The f l u idization occurs
against
grains
in
the
media
is
a particular
The drag forces which fluidize the
bed have been e m p i r i c a l l y expla i n e d by Fair et al [6] as a
function of the. v e l o c i t y of the fluid through the bed and
the expanded porosity of the bed.
Le
L
I- f
(I)-
I
Where:. Ve = wash rate as superficial velocity
Vs = bed grain settling velocity
q
= 0.2 to 0.3, depending on the shape of the
media grains and the flow regime.
f
= porosity of the packed bed
Le = expanded bed depth
Studies
L
= packed bed depth
by
Amirtharajah
[1]
have
shown
that
during
backwash,
particle collisions are insignificant in terms of
removing
attached
particles
from the media
grains.
The
major mechanism of cleaning has been determined to be the
hydrodynamic shear that occurs around the particle while it
is in the backwash f lowstream.
This maximum shear occurs
within a size graded media at bed expansions of 30 to 50%.
In
general,
backwashing
alone
is
fairly
effective
in
removing part i c l e s and floes from within the pores of the
filter media.
the
use
of
Studies by Regan [13], however, indicate that
air
scour
provides better removal
grains.
Air
scour
in a d d i t i o n
to a w a t e r
backwash
of particles attached to the media
is g e n e r a l l y
not
used
in the United
States, though a surface wash is frequently used to break up
mud b a l l s and caked mud which occur on top of the filter
media during filtration.
8
Filter Ripening Theory
Work conducted by Amirtharajah and.Wetstein [3] showed
that the initial effluent quality from a filter used over
several filtration runs could be divided into three periods;
the lag period, the rising limb culminating in two turbidity
p e a k s , and a long receding limb (Figure I).
Amirtharajah
and Wetstein [3] proposed that the lag period was due to the
clear backwash water in the under drains up to the bottom of
the media,
during
the rising
collisions
of
limb was
the
due to p articles
settling
media
at
derived
the
end
of
backwash, and the receding limb was due to the dispersion of
the
media
derived
particles
fro m
the
filter
and
the
accumulation of particles in the media pores.
Recent work by Francois and Van Haute
altered
filter
Amirtharajah
ripening
and Wetstein’s [3]
sequence.
They
[7] has somewhat
description
concluded
of the
that the peak
turbidity is more related to the influent water (they say
95%) than to the remnant water as Amirtharajah and Wetstein
had proposed.
that
Francois and Van Haute go further to explain
the ripening
period
of the filter coincides with a
change in pore structure of the filter bed.
initial
turbidity
breakthrough
to
the
They relate the
breakdown
of
the
i n i t i a l l y placed weak hydroxide floes within the pores of
the media due to the rapid increase of v e l o c i t y gradients
EFFLUENT QUALITY (TURBIDITY)
CLEAN
BACKWASH
WATER
FUNCTION OF
BACKWASH REMNANTS
FUNCTION OF INFLUENT
RECEDING LIMB
RISING
LIMBs
WITHIN
MEDIA
I
I
ABOVE
MEDIA
BELOW
M EDIA
Figure
I.
Characteristics of Initial Effluent Quality.
10
as p articles begin to accumulate.
floes
are
scoured
amounts.
back
Francois
overdosing
the
filter run will
and
primary
into
Van
The loos e l y deposited
suspension
Haute
coagulant
also
at
in
considerable
discovered
the
that
beginning
of
a
decrease the ripening peak and shorten the
time for it to occur.
They attribute this to an increased
blocking rate of the pores and dead zones.
The primary peak
of degradation is thus assumed to be due to a lack of filter
efficiency
because
of
an
inadequate
pore
structure
and
passage of the initial weak floes through the filter media.
The
rate
of
pore
blocking
is
suggested
to
be
strongly
influenced by the chemical treatment of the raw water.
Studies
by Payatakes
et al
[12] and O'Melia and Ali
[11] have shown that the improving phase of filter ripening
is due to the a c c u m u l a t i o n of particles within the media
flow channels.
Payatakes [12] used visual data to show that
the main mechanism causing alteration of the geometry of the
flow channels within the filter media was
This
throat
clog g i n g
resulted
in
an
throat
increase
clogging.
of
local
capture efficiency explaining the improving phase of filter
ripening.
The study by O'Melia and Ali [11] using a polymer
coagulant system, mathematically related the improving phase
of f i l t r a t i o n to
the a c c u m u l a t i o n
of particles
and the
formation of dendrites and particle chains within the.media
pores.
The c o n s t a n t l y a c c u m u l a t i n g particles within the
11
media
are
thought
to
continually
improve
the
effluent
quality by the improved capture of influent particles by the
dendrites.
From
this
extension
improving
of
work
the
phase
O ’M e l i a
particle
and
Ali
capture
of filter ripening.
[11]
deve l o p e d
theory
to model
an
the
The equation not only
includes the c o l l e c t o r efficiency of the original filter
grain,
but it also displays the collector efficiency of the
filter grain and its associated particles
collected
during
filtration.
(2 )
Where:
np
=
Collection efficiency
of a
retained
particle.
nr
=
Single
collector
efficiency
of
a
particle and its retained particles.
Ap
=
Collector
efficiency
factor
of
retained particle.
Diameter of suspended particle.
dp
N
=
Number
of
particles
that act as
collectors
A
=
Collision efficiency factor.
dc
=
Diameter of collector
n
=
Single collector efficiency
12
A
further
equation
was
developed
based
on
a mass
balance about a differential volume element of the filter in
which
the
removal
retained
varies
particles
with
time.
act
This
as c o l l e c t o r s
equation
so that
is modified
by
simplifying assumptions and numerically integrated as a step
function.
coefficients
E x perimental
and for
data
was
calibration
used
to
estimate
the
of the .model. .. Though the
resulting equation is somewhat emperical,
it does correspond
well with data collected on experiments during their study.
- - ^ x n x A x
(1 -f) x
I + np x Ap x B x vo x dp^
(3 )
4;
Where:
ni
=
x Xno x A te
Particle
- ( 4 x ( 1 - f) x nr
^
concentration
-I
xOic))
I
in
the depth L
at time i .
Single collector removal
(nr i-1)
efficiency for
for the (i-1) time step.
Particle retention fraction on media.
B
te
=
Depth in media.
L
no
=
Initial particle concentration.
Bed porosity.
f
vo
Time
=
Fluid velocity.
13
A rational
Payatakes
approach
to interpreting-the
and 0'M e l i a would
be to assume
studies
that
by
in actual
plant operation, the methods of accumulation they described,
dendrites versus pore clogging,
will act synergisticalIy in
varying degrees of importance to provide increased effluent
quality during the improving stages of the filter ripening
sequence.
that
a
It wou l d be difficult to r e a l i s t i c a l l y assume
system
utili z i n g
alu m
and
a system
utilizing
a
polymer as primary coagulants would act exactly the same in
respect to formation of dendrites or pore clogging due to
the varying nature of the chemicals.
Filter Media Preconditioning During Backwash
Beginning with the study by Harris [8] in 1970,
researchers
have
primarily polymers,
looked
into
the
use
of
several
coagulants,
added into the backwash water
in order
to reduce the magnitude and duration of the filter ripening
sequence.
These
studies
have
assumed
that
portion of the beneficial effects of adding
at
least
a
a polymer into
the backwash water would be the adsorption of the polymer
onto the grains of the filter media.
The adsorbed polymer
would subsequently improve the capability of the filter to
remove
the
particles initially passing through the filter,
14
quickly
clogging
the pores,
thus reducing the transport
of
particles through the filter.
Studies by Yapijakis [15],
plants
for
the
100 mgd
Croton N.Y. City water
addition of a small
Newark
conducted using the pilot
N.J.
treatment
and
for
plants,
the
300 mgd
showed that the
amount of polymer (0.15 mg/1) to the
backwash water s u b s t a n t i a l l y reduced the duration of the
filter
ripening
sequence.
The
studies
by H a r r i s
[8]
indicated that similar results could be obtained using a
polymer dose of 0.10 mg/1.
Further study by Yapijakis
[11]
indicated that when polymer had been added to the backwash
water,
the settling characteristics of the backwashed floes
displaced
to
the
settling
basins
was
much
improved.
The study by Francois and Van Haute [7] was conducted on a
pilot plant flocculating domestic waste water with alum in
c onjuction with a non-ionic polymer.
During the last 15
minutes of the backwashing cycle, the polymer was added to
the backwash water.
These studies showed similar results to
the previous r e s e a r c h e r s ; both a decrease in the initial
degradation
peak
and
a decrease
filter ripening period.
in
the
duration
of
the
They attribute this effect to the
abil i t y of poly m e r s to form larger floes than hydroxides,
thus filling the pores of the filter media earlier reducing
the time required for ripening.
15
A study by Chen
[5]
concerning the preconditioning of
a filter during backwash,
again indicated an improvement in
the effluent quality in cases where polymer was added to the
backwash water.
Another focus of the study was to determine
if filter preconditioning during backwash would also reduce
the
concentration
filter.
of microorganisms
passing
through
the
Klebsiella was added into the influent water and a
determination
of the
quantities
passing
through
the filter
during normal filtr a t i o n and during preconditioning were
made.
This study did not show a very good c orrelation in
removal
turbidity
efficiency
between
in the runs where
the
filter
microorganisms
and
the
bed preconditioning was
used.
Logsdon's study [9] concerning the same microorganisms,
including Giardia1 did show an increase of microbes passing
through the filter in the initial stages,
filter ripening sequence progressed.
decreasing as the
It is possible that
when preconditioning is used to improve the initial effluent
from
the
filter
in
terms
of
turbidity,
that
a similar
improvement will occur with the removal of microbes.
16
CHAPTER 4
EXPERIMENTAL RESEARCH
Experimental Methods
The experiments for this study were conducted using the
in-line, dual media filtration pilot plant in the Montana
State University Environmental Engineering Laboratory.
pilot
plant
filter
(Figure
plexiglass unit with a
2)
consists
of
The
a 6" by
media of anthracite coal and sand.
Pilot Plant Filter Media Size Distribution [14].
Table I.
SIZE CHARACTERISTIC
(effective size)
SAND
COAL
0.46
0.86
DlO
(mm)
D60
(mm)
0.62
1.25 .
D90
(mm)
0.70
1.52
1.35
1.46
Uniformity coefficient
The
gpm/sq.
6"
plant
was
contin u o u s l y
ft. using a raw water
operated
at a rate of 4
"mixed within the pilot plant
unit, composed of tap water with bentonite for the polymer
studies and tap water with a silica clay (Min-u-sil 30) as
the
turbidity
source
for
the
alum studies.
The primary
reason for different turbidity sources was that the polymer
17
A.
B.
C.
D.
E.
F.
G.
H.
I.
J.
K.
L.
INFLUENT WATER (TAP)
A RT IFICIAL WAT ER PREPARATION
T URBIDITY FEED
PRIMARY C OA GULANT FEED
RAPID MIX UNIT
DUAL M EDIA FILTER UNIT
BACKWASH W ATER (TAP)
B ACKWASH C OA GULANT FEED
KOMAX STATIC MIXER
HACH SCATTER 4 TURBID IM E TE R
HEWLE T T- P AC KA R D STRIP CHART RECORDER
AIR SCOUR INJECTION
Figure
2.
Laboratory Pilot Filtration Plant Schematic.
13
could not e f f e c t i v e l y remove the M i n - u - s i l and the alu m
could not e f f e c t i v e l y remove the bentonite with the given
water.
The particle size distribution for the Min- u - s i l 30
was determined by Trusler
[17] using a standard hydrometer
(ASTM D 422) for the weight distribution,
image analyzer
3).
for the particle
and an Omnimet
count distribution
(Figure
The bentonite particle distribution was supplied by
Wyo-Ben Inc. and is.as follows:
Table
2.
Bentonite
Particle
Size
Distribution
Screen size
# 200 (0.074 mm)
80 %
# 325 (0.045 mm)
50 to 60 %
The blended
unit
to
% passing
a rapid
mix
raw water was pumped from the mixing
unit
r e s p e c t i v e coagulants;
where
it was
blen d e d
with
3.5 mg/1 Cat Floc TL polymer,
mg/1 alum,
or 16:0.8 mg/1 a l u m / polymer combination.
respective
primary
coagulant
dosages
were
optimized
the
19
The
by
varying the dosages, administered to the pil o t plant under
c o n t r o l l e d conditions.
From the rapid mix unit the water
passed into the filter unit.
The filtered water then passed
through the underdrains and the full
flow
vol u m e
passed
through a Hach Scatter 4 turbidimeter where turbidity was
19
PERCENTAGE
FINER
o % BY WEIGHT
• % BY NUMBER
PARTICLE
Figure
3.
SIZE (/m )
Particle Size Distribution for Min-u-sil 30.
(Trusler 14)
20
c o n t i n u o u s l y monitored and recorded on a Hewlett - P a c k a r d
strip chart.
The flowrate was routinely checked by timed
displacement.
Although the maximum recommended flow through the Hach
turbidimeter is about I liter/min.,
trial runs indicated
that under actual flow conditions of up to 4 liters/min. the
turbidity m easurements could be made with a good degree of
accuracy.
Following a cycle of filtration, which generally was 50
minutes,
the filter was backwashed.
The backwashes were
conducted as follows:
1.
The filter media was air-scoured for I minute in order
to remove
a buildup of coagulant bal l s on top of the
media.
2.
The b a c k w a s h
water
was
circulated
through
the
turbidimeter bypassing the filter until the backwash
water turbidity stabilized.
3.
The filter media was then backwashed for a period of 5
minutes at a p p r o x i m a t Iy 21 gpm/sq. ft. (30% expansion
[I]).
4.
At the end of backwash,
again
c irculated
the backwash influent water was
through
the
turbidimeter
and
the
turbidity monitored until it stabilized.
5.
At
the
same
monitored,
time
the pilot
the
backwash
water
was
being
plant was operated bypassing the
21
filter
in
order
to
stabilize
the
turbidity
coagulant dosages in the filter unit influent.
and
This
generally took about 3 minutes.
6.
Upon
stabilization
influent,
of
the
backwash
and
the
the valves were switched to bring the filter
unit on-line to the turbidimeter.
sequence
water
was
monitored
with
The filter ripening
the Hewlett-Packard
strip
chart.
7.
During experiments where
coagulants
were injected
into the backwash water, the coagulant was introduced
into
the backwash water
means
of a S i g m a
finger
at the base of the media by
pump.
Thi s
water
passed
through a Komax static mixing unit before it reached
the filter.
During the experiments,
infl u e n t
control.
temperature, flowrates, pH, and
turbidity were routi n e l y monitored
All
measurements
of
c oagulant
for quality
dosages
were
determined by timed displacement into a graduated cylinder.
Influent turbidities were determined from grab samples using
a Hach 2100A. nephelometer.
F o l l o w i n g monitoring of the filter ripening sequence
with the strip c h a r t , (see Figure 4 for actual chart data
for experiment B - 2) the v a l u e s from the strip chart were
plotted on standard graph paper, time versus turbidity.
The
o >
O 2
Ua2" - t=
—i
-O
en > I—
3>
•
Oi-IZ
IO —I
EFFLUENT TURBIDITY (NTU)
IQ 70
^
C=
—- Z
I-* Ja
I— * J » Z
« IN)□
-Pa O
3
IQ
m
CO
^
—•
"O
m
o
IN) J » —I
O
3
O
J a i-i
IN) <C
CH m
—•
CO
Ja
O
PN
-Pa J a
O Ja
IN)
3
CPi
LO
^
-I
Z
3>
CO
z
Ja
I-
C=
3=
cn 3= o
CO 3 a O
IN) CO
LQ
Figure
4.
Actual Pilot Plant Effluent Turbidity Strip Chart.
IN)
IN)
23
area under the curves being proportional
to the total mass
of part ic les passing through the filter during the given
time period.
The plotted curves were then planimetered to
a predetermined
time where
the turbidity had
stabilized to the final value.
approximately
Each curve was planimetered
in two separate cycles of three passes around the area.
these
two
cycles
varied
by
more
than
2%,
the
area
If
was
planimetered again.
For
ea c h
series
of
run s
utilizing
a
particular
coagulant, sufficient pilot plant runs were conducted and
recorded until similar filter ripening curves were obtained.
At this point varying dosages of coagulants were injected
into the backwash water as p r e v i o u s l y d e s c r i b e d , with low
dosages tested first followed by higher dosages
to prevent
the "memory effect" noted by earlier researchers [15].
The . p l o t t e d
curves
fr o m
the s e
pl an im ete red as p r e v i o u s l y described.
ru n s
we r e
th en
The value s of area
obtained from these curves were compared to the control run
area and were plotted on graph paper as a percent of the
control
area
plotted
points
optimum dosage
versus
backwash
del ineate
a
of co agu lan t
coagulant
curve, which
to be added
water under the given conditions.
dosage.
These
determines
the
to the backwash
The curve also gives an
indication of the improvement in effluent quality which can
24
be obtained
backwash
by adding
given
dosages
of coagulant
to the
water.
A pilot plant study was conducted at the Bozeman Water
Treatment Plant (BWTP) using the same filter unit as in the
pilot plant studies (see Figure 5).
At the BWTP the pilot
plant was operated using the water from the f l o e culators
leading
to
the
plant
filtration
siphoned from the flocculaters
into
units.
This
the pilot
water was
filter
unit.
The turbidity was monitored using the same Hach Scatter 4
t u rbidimeter and strip chart as used in the pilot studies.
The unit was backwashed with water from the B W T P , the same
water used at the plant for backwashing.
These procedures
were conducted in exactly the same manner as tests conducted
in the laboratory pilot plant for comparison of results.
Experiments Conducted
The following is a list of the experiments conducted to
determine the effects of varying operating parameters and
backwash c oagulants
filter
ripening
on the magnitude and duration of the
sequence,
the
and
filter
to delineate
boundaries
of
ripening
alphanumeric
labeling of these experiments
index the related figures in the text.
the physical
sequence.
The
is also used to
25
A.
B.
C.
D.
E.
F.
G.
Figure
BOZEMAN WATER TRE AT M EN T PLANT FLOCCULATOR
DUAL M EDIA PILOT PLANT FILTER UNIT
BACKWASH C OA GULANT FEED
KOMAX STATIC MIXER
HACH SCATTER 4 TUR BI D IM E TE R
HEWLIT T- P AC KA R D STRIP CHART RECORDER
SURFACE SCRUB WATER INJECTION
5.
Pilot Plant Schematic at Bozeman Water
Treatment Plant.
26
A.
Polymer as primary coagulant
Th i s
set
of
,
experiments
was
conducted
using.the
laboratory pilot plant with Cat Floc TL, a medium molecular
weight cationic p o l ye lec tro ly te,
at a dose of 3.5 mg/1 as
the primary coagulant for direct filtration.
I.
Pol ym er
in Backwash Water
This set of experiments was conducted by injecting
varying
dosages
of
polymer
(Cat
Floc
TL)
into
the
backwash water for the full 5 minute backwash duration.
The dosages used started at the
s m all es t
value
increased in dosage with subsequent experiments.
and
The
dosages of polymer injected into the backwash water
were
determined by measuring the time required to pump
a particlular volume from a graduated cylinder.
Due to the variati on of the water quality during
the y e a r , the tests were conducted over three seasons
as follows:
a.
Winter/spring.
were
b.
Twenty-nine pilot plant runs
conducted.
Summer.
Nineteen
pilot
plant
runs
were
conducted.
c.
Early
fall.
conducted.
Twelve pilot plant runs
were
27
2.
Alu m
in Backwash Water
This
set
of
experiments
was
conducted
in
the
summer season with the same methods as described above.
Eighteen pilot plant runs were conducted.
3.
Var ia ti on
Th i s
control
set
of
experiments
conditions
backwash water.
was
of Backwash Volu me
was
conducted
with no coagul ant s
added
under
to the
The volume used to backwash the filter
varied by using
increasingly
shorter
time periods
to backwash the filter at the standard 21 gpm/sq.
rate.
4.
ft.
Six pilot plant runs were conducted.
Var ia ti on
of Injection
Time
This set of experiments was conducted during the
winter
season by injecting the optimum dose of polymer
determined in Experiment A-Ia into the backwash water
at
varying
times
backwash phase.
5 minutes,
with
respect
to
the
end
of
The polymer was injected for the full
then the
last
4,3,2,1 and 0.5 minutes
backwashing.
Nine pilot plant runs were conducted.
5.
Ov erd osi ng
Pol ym er
the
of
Effects
During the winter it was noticed that with very
small doses of polymer added to the tap water, a fog of
overdosing
form ed,
which
turbidity of the water.
quickly,
increased
the
A series of experiments were
conducted involving displacement of known quantities of
28
polymer
mixed
with
the
backwash
water
into
the
turbidimeter and subsequent monitoring of the turbidity
of
this
water
in
the
turbidimeter, under
static
conditions over a given time period.
6.
V a ri ati on of Remnant V o l u m e
Th i s
series
setting
the
heights
above
of
backwash
the
shown on Figure 6.
experiments
discharge
filter
was
conducted
valve
media.
at
These
by
different
heights
are
Varying this height above the media
has the effect of varying the remnant volume left above
the media f o l l o w i n g backwash.
Six pilot plant runs
were
conducted.
7.
Operation of Pre-ripened Filter
This
operating
series
of
the filter
experiments
plant
was
conducted
as p r e v i o u s l y
described
until the filter had ripened to a stable effluent.
this
point
drained
the
out
water
standing
above
the
this top portion
without
visible
media.
was
The filter unit was then
filled with backwash water containing no
all
media
At
of the filter unit by means of a v a l v e
dir ect ly above the media.
backwashed
by
of the
filter
disturbing
pa rti cle s
were
the
above
coagulant and
the media was
ripened media until
removed
from
above
the
At this time the filter was brought back on
29
FILTER DATA
I .D. = 6"x6"
Area = 0.25 feet sq.
Flowrate = 4 qpm/ft. sq.
Velocity = 0.535 f t/min.
14.38 min
COAL T
Figure
6.
HACH TURBIDIMETER
Pilot Plant Filter Unit Measurements and
Detention times.
30 •
line and the results monitored.
were
B.
Two pilot plant runs
conducted.
Alum as Primary Coagulant
1.
Pol ym er
in Backwash Water
This series of experiments was conducted in the
sam e
m a n n e r as
the
polymer
primary
experiments previously described.
were
conducted.
2.
Alu m
coagulant
Six pilot plant runs
in Backwash water
This series of experiments was conducted in the
sam e
m a n n e r as
the
polymer
primary
experiments previously described.
Ten pilot plant runs
were
conducted.
3.
A l u m / po lymer in Backwash Water
This
experiment
coagulant
determined
the
optimum
a l u m / p o l y m e r (20:1) dose using Cat Floc TL, injected
into the backwash water
to reduce the magnitude and
duration of the filter ripening sequence.
Nine pilot
plant runs were conducted.
4.
Var ia ti on
of Injection
The optimum dose
Time
of al u m determined
in B - 2 was
used in the backwash water of these experiments.
pilot plant runs were conducted.
5.
Variation
of Backwash Volume
Four pilot plant runs were conducted.
Five
31
6.
Variation of Remnant Vol um e
Pilot plant runs with and without the optimum dose
of alum coagulant in the backwash water were conducted.
Twenty pilot plant runs were conducted.
7.
Backwash Effluent Turbidity
During
the
ba ckwash
of
the
filter
unit
samples were taken at 30 second time intervals from the
overflow gutter of the pilot plant.
monitored
for
turbidity
These samples were
using
the
Ha c h
2 10 0 A
nephelometer.
8.
Zeta Potential of Influent
- These
experiments
manner as described
in A - 7.
A series
for
were
conducted
the polymer
of tests were
without the optimum dose
in
the
same
primary coagulant
conducted with
of alum coagulant
in B-2 injected into the backwash water.
and
determined
Four pilot
plant runs were conducted.
9.
Operation of Pre-ripened Filter
This series
manner as A-8.
10.
of
tests
was conducted in the same
Two pilot plant runs were conducted.
Injection Above Media
This series of experiments was conducted in the
same
manner
backwash
as
previously
injection
described
experiments
except
for
that
the
other
instead
of
injecting the backwash coa gul ant bel ow the media,
it
32
was
injected
into
the
filter
unit abo ve
the media.
This experiment was designed to minimize the contact
time of the backwash coagula nt with the filter media.
Two pilot plant runs were conducted.
11.
Zeta Potential
Effluent
of Backwash Effluent
samples
were
obtained from the backwash
water effluent at 30 second intervals.
were
monitored
.previously
for
turbidity
described.
Two
These samples
and Zeta Potential
pilot
plant
runs
as
were
. conducted.
C.
Alum/polymer (Cat Floc TL) 20:1 as Primary Coagulant
1.
Pol ym er
This
in Backwash Water
experiment determined the optimum polymer
dose injected into the backwash water
to reduce the
magnitude and duration of the filter ripening sequence
under summer conditions.
Six pilot
plant runs were
conducted.
2.
Alu m in Backwash Water
This
injected
experiment
determined
int o
backwash
the
the
optimum alum dose
water
to
r e duc e
the
magnitude and duration of the filter ripening sequence
under summer co n dt i o n s .
conducted.
Eight pilot plant runs were
33
3.
Alum/polymer in Backwash Water
This
experiment
determined
the
optimum
alum/polymer (Cat Floc TL) 20:1 dose, injected in to the
backwash water
the
filter
to reduce the magnitude arid duration of
ripening
sequence
under
summer
conditions.
Seven pilot plant runs were conducted.
Tracer Studies
1.
Vari a t i o n
of Remnant
Vol u m e
At each depth interval above the media used in the
prev i o u s experiments where the remnant water
volume
above the media was varied, 'a dye trace was conducted.
A slug of Rhodamine WT was . injected into the inlet of
the filter
unit.
The effluent
dye concentrations were
sampled in the Hach scatter 4 turbidimeter after the
fluid had passed through the filter unit.
in al l
The flowrate
of the dye trace studies was maintained
gpm/ft.sq.'
The dye samples were kept in a water bath
approximately
filtered
at 4
the
same
to stabilize
temperature
as the water
the concentration
values
being
to be
determined with the flouromenter. Dye concentrations
were determined using a Turner Flourometer and recorded
in ppb.
2.
Four pilot plant runs were conducted.
Continuous Dye Iniection at Station
"C"
A dye trace was conducted while the filter unit was
34
operating with the influent at the 4.17' station above
the media.
During these tests,
Ehodamine Wt dye was
c o n t i nu ou sly injected into the rapid mix unit of the
filter
plan t.
conducted
Effluent
on samples
dye
we r e
taken from the Hach turbidimeter
at regular time intervals.
conducted measuring
concentrations
One pilot plant run was
the dye concentration as it was
injected into the filter unit, and after stabilization
of the eff lue nt dye concentration,
the dye injection
was stopped and a test was conducted to determine the
displacement time of the dye from the filter.
All dye
tracer studies were conducted at the same 4 gpm/ft sq.
flow rat e as used in the filtra tio n experiments.
Two
pilot plant runs were conducted.
3.
Backwash Tracer of Pilot Plant
A pilot
plant
backwash dye trace was conducted
with the influent valve at the 4.17' station above the
filter
media.
One
experiment
was conducted injecting
dye continuously into the backwash water with the same
pump used to inject the backwash coagulants, the dyed
backwash water then passed through the Komax static
mixer
and
on
through
the
filter
unit.
The
dye
concentrations were measured from the effluent hose at
reg ula r intervals. . The other part of the experiment
involved
stopping
the
injection
of
dye
into
the
35
backwash water after the effluent dye concentration had
stabilized,
then
measuring
the
effluent
dye
concentration as the dye was displaced from the filter
unit.
4.
One pilot plant run was conducted.
Tracers on Filters at Bozeman Plant
Two tracer studies were conducted at the Bozeman
Water Treatment Plant in conjunction with work being
conducted by Buc kl in [4] grant.
These tracer studies
were conducted by continuously injecting a solution of
flour ide
(sodium
flou ros ili cate)
into
water of one of the filters at the plant.
from
the
filter
was
monitored
the
influent
The effluent
for
flouride
conce ntr ati on by use of col orime try conducted with a
Bausch
and
conducted
Lomb
to determine
of a filter unit.
injected
effluent
the
spectrometer.
into
the
Another
backwash
study
was
displacement
time
The flouride tracer was continuously
one
of
the
filter
units
concentration had stabilized.
filter
was
conce ntr ati on
backwashed
and
until
the
At this point,
the
flouride
from the backwash gutter effluent was
monitored until the end of backwashing.
Turbidity was
also monitored.
E.
Pilot Studies at Bozeman Plant
This series of experiments were conducted using the
36
laboratory pilot plant which had been transferred to the
Bozeman Water Treatment Plant.
The influent for
the pilot
plant was derived from the BWTP flocculators which contained
an average polymer
Floc T .
I.
coagulant
concentration
of 3.5 mg/1
Cat
The influent turbidity averaged about 3 NT U.
Polymer in Backwash Water
This
experiment
determined
the optimum dose of
polymer injected into the backwash water to reduce the
magnitude and duration of the filter ripening sequence.
Six pilot plant runs were conducted.
, 2.
Alu m
in Backwash Water
This
experiment
determined
the optimum dose of
alu m injected into the backwash water
to reduce the
magnitude and duration of the filter ripening sequence.
Five pilot plant runs were conducted.
3.
Alu m/polymer in Backwash Water
This
experiment
alum/polymer
reduce
the
ripening
conducted.
determined
20:1 injected
magnitude
sequence.
and
Five
the optimum dose of
into the
backwash water
duration
pilot
of
plant
the
to
filter
runs
we r e
37
CHAPTER 5
RESULTS AND DISCUSSION
Development of a Revised Filter Ripening Theory
Analy si s
of
the
results
of
these
experiments
has
allowed improvement of the theories previously developed by
Amirtharajah and Wetstein [3] and has also incorporated the
ideas of O'Melia and Ali
[7].
[11] and Francois and Van Haute
The modified theory of filter ripening will now be
explained
and
documented
with
experimental
data
where
required.
Filter ripening can be predominantly divided into three
major stages:
1.
The remnant
2.
The influent
stage.
water
mixing
and
particle stabilization
stage.
3.. The filter media conditioning stage.
Figure
7. depicts
the
proposed
initial eff lue nt degradation.
characteristics
The figure is composed of
ideas developed from the current research,
B u ck li n
of
the
[4], from Ami rtharajah and Wetstein
research
of
[3] and from
REMNANT STAGE
INFLUENT MIXING AND
PARTICLE STABILIZATION
STAGE
FILTER M EDIA CONDITIONING
STAGE x
EFFLUENT TURBIDITY
ACCUMULATION OF
PARTICLES IN FILTER
MEDIA PORES
u>
CO
BELOW
MEDIA
WITHIN
MEDIA
Figure
7.
ABOVE
MEDIA
INFLUENT MIXING WITH
ABOVE MEDIA REMNANTS
INFLUENT
Proposed Characteristics of Initial Effluent
Degradation.
39
Francois and Van Haute [7].
The current research did not
show al l of the features of the filter ripening sequence
that
have
been
noted
by
the
oth e r
researchers.
The
variation of results is probably due to the varying nature
of
in div id ual
operation.
fi ltr ati on
A brief
plants
description
and
of
ripening sequence is as follows:
their
the
methods
proposed
of
filter,
The "remnant stage" has
been shown by dye trace studies in the current research to
be a function of the water remaining in the filter unit at
the end of backwash.
Amirt ha ra jah
From the research of Buck lin [4] and
and Wetstein
[3],
the
further divided into the "lag phase",
phase",
remnant
stage can be
the "media disturbance
and the "upper filter phase".
The lag phase is a
function of the turbidity of the backwash water which
is
generally
is
fair ly
low.
The
period
of
low
turbidity
f o l l o w e d by a peak of turbidity,
or the media disturbance
phase,
to
which
has
been
proposed
be a function of the
particles derived from the collisions of the settling media
at the end of backwash, and from the shearing of particles
from the media grains at the beginning of the filter run.
The
subsequent
filter phase,
period
of decreasing
turbidity,
is a function of the volume and
or upper
turbidity of
the remnant water remaining above the media at the end of
backwash.
The true effects from this period may be masked
by excessive turbidity derived during the media disturbance
40
phase.
The "influent
mixing
and particle
stabi lizat ion
stage", which constitutes the largest and longest peak of
turbidity,
is apparently due to the partial stabilization of
the influent floes as they disperse into the coagulant-free
backwash remnant water.
not
The partialIy stabilized floes will
attach to the media grains as well as correct Iy treated
floes,
unit.
thus they w i l l
be transported through the filter
The final period, or the "filter media conditioning
stage", occurs after the partialIy stabilized particles have
been displ ac ed from the filter unit, and pro pe rl y treated
particles begin to accumulate within the filter media pores.
The
final
eff lu en t
quality
of' the
filter
is
therefore
reached as the filter media is conditioned with the influent
particles.
Each portion of the filter ripening sequence
will be discussed individual Iy in more detail.
The Remnant Stage
From
the
pre vious
discussion, the remnant stage is
r
associated with the fluid remaining in the entire filter
unit at the end of backwash.
This fluid inc ludes that in
the under d r a i n s , the filter media, and that in the remnant
v o lu me above the filter media up to the backwash gutters.
In Amirt ha ra jah
being
filter
called
and Wetstein's
the
remnant
stage
theory
[3],
consisted
what
of
ripening sequence up to the imp ro vi ng
is
now
the entire
phase.
The
comparison can be made by analysis of Figure I and Figure 7.
41
The fluid in the underdrains was associated with the "lag
phase",
the fluid
"rising
limb"
and
remnant
vo lu m e
in the media was
the
above
first
the
peak
media
associated
with
the
and
the
with
the
of turbidity,
was
associated
second peak of turbidity due to the exponential Iy increasing
particle concentration in the fluid directly above the media
up to the overflow gutter.
Ex ha us t i v e
conducted
studies (over 190 pilot plant runs) were
in this
coagulants.
research project
utilizing
three
primary
The results of these studies, esp ec i a l l y the
polymer and alum studies,
varied significantly in terms of
the magnitude and duration of the filter ripening sequence.
The polyme r studies A-I
the
beginning
experiments
of
was
and A - 2 , show an initial peak at
filtration,
which
shown
due
to
be
through
to
the
small
scale
dislodging
of
particles from the pilot plant piping during the opening of
the fluid control valves.
comparison
to
dye
tracer
The second peak is associated,
studies
(Figure
8),
with
by
the
remnant fluid v o l u m e remaining within and abo ve the media
f o l l o w i n g backwash.
The effluen t water qua lit y begins to
impr ov e the instant the inf lu en t water enters the filter
media.
The alum and a l u m / p o l y m e r experiments,
hand,
show
a relatively
long
period
of
on the other
low
turbidity,
associated directly with the water in the filter unit up to
OF MAX IM U M (PPB)
0L
DYE CONCENTRATION
M
Figure
8.
Variation of Remnant Volume Above Filter
Media.
Experiments A-6, B-6 and D-I.
43
the interface with the influent.
This time period is the
same length as the entire filter ripening sequence in the
poly me r
studies.
compared
to
the
Again
dye
the. length of this p e r i o d , when
tracer
studies
on Figure
8,
can
be
relate d to the v o l u m e of remnant water up to the interface
of the influent water.
turbidity,
Following this initial period of low
the turbidity
rapidly
increases as the influent
water begins to enter the turbidimeter.
This stage will be
explained in the next section.
In both the alum and polymer studies, the.lengths of
the
periods
described
have
been
we l l
documented
with
a
series of pilot plant runs where the remnant volumes above
the filter media were varied,
and B-6.
ie. experiments in series A-6
From Figure 8, when the time periods of the filter
ripening curves are compared with the dye tracer study D — I
and with th eor eti cal detention times,
indeed
the
remnant
stages
are
it can be seen that
directly
related
to
the
remnant fluid in the filter unit up to the backwash overflow
gutter.
This data, compared to data derived from the Bozeman and
H e le na
water
treatment
W e t s t e i n ’s [3]
terms
work
[4]
and
Ami rtharajah
does not correspond we l l
of the presence and timing
degradation.
second
earl ier
plants
of the initial
peak
and
in
of
The data, does however correlate well with the
degradation
peak.
The
discrepancy
between
the
44
studies
indicates that a volume of turbidity involving the
first peak on the filter
present
in the
ripening
pilot plant
sequence curve
studies
is not
conducted here.
The
absence of this first peak may be due to the more efficient
backwashing of the filter media in the current study by the
use of air scour,
in comparison with the other studies which
did not use air scour.
From
the
studies
conducted
at
the
Bozeman
Water
Treatment Plant (D-4 fl uoride tracer, Figure 9), it can be
seen that the first peak of effluent degradation,
though of
very short duration, occurs at about the same time as the
th eor eti cal
detention
concu re nt ly
with the first presence of inf luent
through the filter.
time
time
for
the
The a r r i v a l
early stage in filtration,
detention
to
top
of
the
media
and
passing
of the inf luent at this
when compared to the theoretical
it to occur,
indicates
short-circuiting
in the filter unit.
The early arrival is also partially due
to
flowrate,
the
very
high
effluent
valves
are
ope n
completely) during the filter to waste cycle in the initial
stage of filter operation.
This initial peak also occurs at
the same time the fi.lter to waste valve switches off and the
filter
eff lue nt
passes
appear to be sev er al
to the
c l e a r w e l I.
Since
there
things occuring at the same time it
would be very difficult, without further study, to determine
the exact cause of the first degradation peak in the BWTP
Flouride tracer - Run
FLOURIDE TRACER % OF MAXIMUM
- - -o
Flouride tracer - Run
CD ZD CO
O
Q
i-O
Turbidity - Run
Cl O
<3- UJ
?.
Turbidity - Run I
u c OJ
4->-IO) +->4O Ol-I-
-I
- ?
= 4 . 5 n pm/ft. sq
= 3.4 qpm/ft. sq
CU +->
I—
Figure
'O +-1
9
Filter Ripening Sequence at Bozeman Water
Treatment Plant with Tracer Study.
46
filter
units.
however,
Some
limiting
assumptions
to narrow down the possibilities.
can
be
made,
Due to the short
duration of the initial p e a k , it can be assumed that the
a r ri va l of inf lue nt at about this time is not res ponsi ble
for the peak since the filter could not "ripen” fast enough
to produce the peak shown.
of the filter to
On the other hand,
waste valve
the switching
or the shear occuring within
the filter media due to the high flo wrate at the start of
the
filt er
run
initial peak.
may
be
r eas on abl e
exp lan at io ns
It is not very likely,
for
though possible,
the
that
the switching of the v a l v e s
could produce such a peak of
turbidity.
mo s t
likely
the
filter
part ic le s
backwash
This
pe a k
derived
and
the
from
beginning
of
the
is
associated
media
filter
at
the
r u n , as
with
end
of
other
researchers have suggested [3],
The laboratory pilot plant studies,
on the other hand,
were conducted using an air scour phase and were relatively
short in duration compared to the BWTP studies [4].
These
differences can explain the lack of the initial peak in the
pilot plant studies.
of
filtration
would
The air scour and short time duration
allow
a much
smaller
buildup
of
particles on the media grains during a cycle of filtration
and backwash.
Thus when the
filter
unit was
started
up
again fewer particles would be dislodged than in the Bozeman
47
plant
filters,
and
thus
this
initial
peak would
not
be
present.
From this discussion it can be seen that the remnant
stage can be further subdivided into three sub-stages which
are
reminiscent
Wetstein [3].
of
those . proposed
by
Amirtharajah
and
The "lag phase" associated with the water in
the underdrains,
"the media disturbance" phase associated
with the dislodging of particles within the media at the end
of backwash
and
beginning
of
filtration,
and the "upper
filter phase" associated with the remnant water remaining
above the filter media at the end of backwash.
During
the
"l a g
phase"
the
turbidity
will
approximately the same as that of the backwash water.
be
Small
peaks of turbidity can occur in this period of time due to
particles dislodged in the piping by the movement of valves,
as seen in the pilot plant studies.
The "media disturbance
phase".was not present in the pilot plant studies
but was
clearly seen in the data gathered at the Bozeman and Helena
plants [4].
seen
due
to
During this phase a turbidity increase may be
the
di slo dgi ng
of
particles
as
the
media
compacts and media particles collide with each other at the
end of backwash [3].
off
the
media
production.
Additi ona l particles may be sheared
grains
as
the
filter
is
put
bac k
into
The initial shear within the filter media pores
may cause the dislodging of particles from the media grains
48
which were in stable configurations of attachment before
b a c k w a s h , but which after the backwash were reoriented on
the media particles,
and were in unstable configurations.
Upon opening of the filter, these particle configurations
which are now in a new flo w r e g i m e , may try to reorient on
the filter media surface.too great,
into
the
If the initial hydraulic shear is
these particles will be swept from the media and
effluent.
If,
increased gradually,
of
the
other
h a n d , the
This
is
these particles will have more time to
reorient on the grain of attachment or attach
grain.
flow
phenomenon,
whi le
difficult
to another
to
prove
in
theory, is clearly shown in results from data collected at
the
H e le na
plant
[4].
At this
partic ula r
plant,
the
operators have the capability to open flow from the filters
at wh a t e v e r rate they desire.
A comparison of the filter
ripening sequence curves derived from a filter which had
been opened to full
opened
flow immediately,
inc remental Iy
depicted
in Figure
10.
over
The
a
thirty
with the same filter
minute
comparison
shows
difference in both magnitude and duration
period
are
a dramatic
of the curves.
Since the only difference in the filter operation is the
rate of opening,
opening will
it can safely be assumed that
apply a lower shear to the particles attached
on the media grains,
reorient
incremental
or attach
thus giving them more opportunity to
to another
media
grain.
Since
these
INCREMENTAL OPENING SCHEDULE
TIME (minutes)
FLOWRATE (mqd)
EFFLUENT TURBIDITY
O to 3
3 to 5
5 to 7
7 to 9
9 to 11
11 on
Filter //7 instantaneous opening
>£»
VO
Filter #7 incremental opening
TIME (MINUTES)
Figure
10.
Filter Ripening Sequence at Helena Water
Treatment Plant with Instantaneous and
Incremental Filter Opening.
50
particles are not transported out of the filter media,
will
accelerate
the filter,
the capture
of influent
particles
they
entering
thus reducing the magnitude and duration of the
entire filter ripening sequence.
The "upper filter
water
above
gutter. ' In
stage" is associated with the remnant
the filter
media
Amirtharajah
and
to the top of
Wetstein's
the
study
backwash,
[3]
they
related the second degradation peak to the concentration of
p articles
in the remnant water standing above the media.
From the data
derived
in experiments
A-3,
B-5,
and
B - 7,
shown in Figure 11, it can be seen that there is a very good
correlation
with
the
overall
magnitude
of
the
filter
ripening sequence when compared with the concentration of
particles
left
in
the
foil owing backwash.
remnant
volume
above
media-
In Figure 11, an.inflection point in
the B-5 and B-7 curves occur at about 2 minutes.
B - 7 curve,
the
From the
it can be seen that this is the point at which
the majority of the p articles in the.remnant volume above
the media have been
removed.
At backwashing times less
than 2 minutes, the magnitude of the B-5
curve increases
dramatically due to the increased concentration of particles
in the remnant volume- above the media.
seen in the A-3 curve.
do not
reflect
experiment, w i l l
what
be
This affect is not
However the A series of experiments
is
occurring
ignored.
This
at
the
BWTP
obser v a t i o n
thus
this
tends
to
OO
UJ
O
CO
EXPERIMFHT B-7
Backwash effluent turbidity
EXPERIMENT B-5
ALUM PRIMARY COAGULANT
Variation of backwashing
time at 20 liters/minute
EXPERIMENT A-3
POLYMER PRIMARY COAGULANT
Variation of backwashing
time at 20 liters/minute
r:
Z=
O
«vJ-OD
Cvi
BACKWASH E FFLUEMT T URBIDITY (N
O
Ul
O > O
LO C£ CO
ZD r-H
O
O
C
UI
or
O
_ J CXj
O «—I
or
h
ro
o
O
O
U O
O r-4
O
CXJ
---
1
Figure
2
11.
3
4
TIME (MINUTES)
5
6
Variation of Backwashing Volume.
A- 3 , B -5, and B-7.
Experiments
52
verify
the
concerning
results
an increase
of
Amiftharajah
and
of turbidity during
the filter ripening sequence.
Wetstein
this
[3]
portion
of
The increase is due to the
exponentialIy increasing concentration of particles from the
bottom to the top of the remnant
f o l l o w i n g backwash.
volu me above
In the alum studies
the media
it was seen
that this portion of the filter ripening sequence had a very
low turbidity,
this
stage.
only sl i g h t l y increasing up to the end of
The polymer
studies showed basically the same
thing, a slight increase in turbidity as the fluid above the
media was c o m p l e t Iy di splac ed through the unit.
From
analy sis of Figure 9 at the BWTP it can be seen that a great
deal
of short-circuiting occurs within the filter unit.
may thus be difficult,
in any large scale filter unit
It
to
c l e a r l y define each of these three phases' proposed in the
remnant
stage.
Another mechanism that occurs within the remnant stage
which
tends
to
increase
the
mob il ity
of
the
particles
remaining within the filter unit fol l o w i n g backwash,
is
the partial stabilization of these particles.by contact with
the backwash water
conducted
to
during backwashing.
determine
the
Zeta
Experiment B-Il was
potential
of
particles
removed from the filter unit during backwash with respect to
time.
Figure '12 shows the resu lts of the two experiments
conducted.
In both runs, as the majority of the particles
53
RUN 2
- -CK
backwash effluent
TURBIDITY (NTU)
ZETA p o t e n t i a l (MV)
RUN I
TIME (MINUTES)
Figure
12.
Zeta Potential of Backwash Effluent Particles.
Experiment B-Il.
54
in the filter unit had been displaced,
had
also
increased.
stabilization
extended
water.
of
these
contact
The
Thi s
time
higher
Zeta
indicates
particles
with
the Zeta potential
the
potential
had
that
occurred
coagulant-free
partial
by
their
backwash
of these particles means
that they will be less likely to be removed by attachment to
the filter
potential
magnitude
grains
[10].
of
the
than
This
p articles
with a lower negative
phenomenon will
remnant
tend to
increase
portion of the filter
the
ripening
sequence.
Influent Mixing and Particle Stabilization Stage
Following backwash and at the start of a new filtration
cycle,
as the infl u e n t water passes into the filter unit
above the filter media,
a certain portion of this water will
mix with the ”coagula n t - f r e e " backwash remnant water.
The
dispersion of the formerly destabilized particles and floes
into this water causes
become
some of the particles to possibly
partially stabilized,
or weakened,
thus reducing the
effectiveness of their filtration. The partial stabilization
of these particles results
in a major portion of the filter
ripening sequence magnitude as they pass through the filter
media and are displ a c e d from the filter unit.
this phenomenon,
In terms of
the polymer studies cannot be considered
55
v a l i d when compared
to the. alum
or to the Bozeman plant
studies since the ripening sequence ends at the end of the
remnant phase and the particle stabilization stage does not
occur.
The alum studies,
on the other hand, correlate well
with the results obtained at the Bozeman and Helena water
treatment plants.
The analysis of the alum study_(B series)
data will be thus used to explain the phenomenon of influent
water
mixing.
E xperiments
B-8
and
B-IO
quantify this phenomenon.
plant
run s
were
designed
to
In experiment B-8,
were, c o n d u c t e d ,
two
control
further
four pilot
runs
without,
coag u l a n t in the backwash w a t e r , and two with coagulant.
For each run the Zeta potential
and turbidity were taken
from grab samples obtained slightly above, the filter media
during the initial stages of filtration.
In Figure 13, both
the Zeta potential and influent turbidity values (corrected
to correspond
to effluent
turbidity at the turbidimeter by
means of c o r r e l a t i o n with the D-I dye trace) were, plotted
concurrently
ripening
during
the
sequence.
the
increase
with
effluent
From
control
runs
turbidity
Figure
the
13
of
it can
beginning
of
the filter
be
the
of the effluent corresponds to the
seen
that
turbidity
diluted
steadily increasing turbidity of the influent water.
but
At the
same time, the Zeta potential, which was the most negative
for
the
initial
influent
water
(indicating
partially
EFFLUENT TURBIDITY (NTU)
In
?n
INFLUENT TURBIDITY (NTU) ZETA POTENTIAL (MV
IO
20
- 6 - 8
-10
56
(137) control
(138) control
12
I
TIME (MINUTES)
Figure
13.
Zeta Potential of Influent Particles Compared
to Effluent Turbidity.
Experiment B-8.
57
stabilized
particles)
begins
to
decrease
to
the
stable
influent Zeta potential required for optimum filtration.
This means
that the particles
diffusing into the remnant
water become partially stabilized and thus will pass through
the filter unit much more readily creating the major peak in
turbidity of the filter ripening sequence.
Further tests conducted using coagulant in the backwash
water show the effect of a s tabilized particle dispersing
into the remnant backwash water containing alum.
clearly shows that,
particles will
water.
and
as expected,
the Zeta potential of these
decrease as they are mixed with this remnant
This result corresponds
Van
Figure 13
Haute
[7]
data
somewhat with the Francois
indicating
decreased
initial
degradation peaks when the influent water was overdosed with
coagulant.
The initial overdose would have the affect of
reducing the stabilization of the influent particles as they
mix
with
the
remnant
water
by
maintaining
a more
even
coagulant/particle concentration ratio as this dispersion
takes place.
In
the
coagulant
pilot
plant
runs
shown
in Figure
in the backwash water, the initial
13,
with
peak occuring
at about 2 minutes and dissipating by about 5 minutes,
is
due to the formation of aluminium
hydroxide
in
the
overdosing
backwash
remnant
These floes may
water
by
the
micro-fIocs
of
coagulant.
become attached to the filter media thus
58
may
enhance
the effects
of addition of coagulant
to the
backwash water in producing a reduction in the magnitude and
duration of the filter ripening sequence.
Experiment B-IO (Figure 14) was designed to determine
if the effect of coagulant introduction into the backwash
water
was
more
important
as
a media
conditioner,
or
if
destabilization of the stabilized particles in the influent
mixing and particle stabilization stage was more important.
In B - I O , the a l u m c oagulant was injected bel o w the filter
media
in
the
conducted.
sam e
manner
as
In a subsequent test,
the
standard
tests
were
the coagulant was added to
the backwash water only above the filter m e d i a .so that a
minimum of excess coagulant would contact the filter media.
In both tests,
the results were essentially the same as far
as reduction of the peak of the filter ripening curve was
concerned.
the
These results indicate that the prevention of
stabilization
of particles
during the backwashing of a
filter unit by the addition of coagulants to the backwash
water is probably more important, at least in alum systems,
than is conditioning of the filter media by a coating
coagulant.
tends
to
of
Injection of coagulant above the media also
suppress
the
formation
of
the
initial
peak
of
turbidity which is due to the overdosing of coagulant in the
remnant backwash water.
EFFLUENT TURBIDITY (NTU)
PILOT
INFLUENT
BACKWASH
PLANT TURBIDITY
COAGULANT
RUN
NTU
DOSE mq/1
2B9-1
16
0.0
2B9-2
13
0.0
2B10-1
17
18.4
2B10-2
17
17.6
2B10-1 inject below media
2B10-2 inject above media
12
Figure
14.
16
20
TIME (MINUTES)
24
28
Coagulant Injection Above and Below Filter
Media.
Experiment B-10.
60
'
Filter Media Conditioning Stage
0 ’M e l i a and Ali [25], Yapijakis [11], Payatakes [18],
and
by
Francois
and
Van
Haute
[12],
proposed
that
the
a c c u m u l a t i o n of p articles within the media results in the
improving phase of the filter ripening sequence, whether by
dendrite
formation,
accumulation.
conducted
pore clogging,
Observation
using
the
or discrete
of the numerous
pilot
plant
in
particle
filtration runs
this
study
coincide with the results of these researchers.
tend
to
Experiments
A-8 and B-9 (Figures 15 and 16) were designed to determine
the r e l a t i v e importance of the a c c u m u l a t i o n of particles
within
the
filter,
or "filter
media
conditioning"
with
respect to final effluent quality.
In the experiments,
the filter was operated in each
case until a stable effluent quality was produced.
ripening,
Upon
the water above the media was displaced from this
v o l u m e with backwash water by means of a backwashing port
placed direcly above the media.
back on line.
and
the
filter
quality.
In all. cases, passage of the backwash water
dispersed
media
The
The filter was then brought
did
influent
not
at the interface through the
signif i c a n t l y
post-ripened effluent
affect
the
effluent
quality remained about
the same as the i n i t i a l l y ripened effluent quality.
This
indicates that the conditioning of the filter media is the
most important characteristic during the filter ripening
/
POLYMER PRIMARY COAGULANT
PILOT PLANT RUN
DESCRIPTION
(174)
No disturbance of media
during displacement of
fluid above media.
Slight disturbance of
media during displacement
of fluid above media.
Top 2" severly disturbed
during displacement of
fluid above media.
Initial filter ripening
sequence
EFFLUENT TURBIDITY (NTU)
^r
o
CO
o
CU
o
CTl
H
C ontinuation o f filtration through post-ripened filter
6
8
10
TIME (MINUTES)
Figure
15.
Operation of Pre-ripened Filter with Polymer
as Primary Coagulant.
Experiment A-8.
EFFLUENT TURBIDITY
PILOT
PLANT
RUN
177) 2B9-1
178 2B9-2
INFLUENT
TURBIDITY
NTU
Continuation of 2B9-1 and 2B9-2
after backwashing out remnant
volume above media
Figure
16.
Operation of Pre-ripened Filter with Alum as
Primary Coagulant. Experiment B-9.
63
sequence and it is not just the dispersion of the partially
stabilized influent particles from the filter unit.
In the experiments which utilized c oagulants in the
backwash water,
A-I,.2, B-1,2,3, and C-1,2,3,
it appears that
an accelerated filter conditioning is taking place.
It has
been proposed by other researchers [8, 15] that in polymer
systems a portion of the polymer added to the backwash water
will
adsorb to the media particles.
will
create
This adsorbed polymer
bridges and projecting dendrites within the
pores of the filter and will aid in the removal of particles
from the backwash water and in the initial influent as it
passes
through
the
filter,
by
the
accele r a t i o n
process of particle accumulation within the media.
of
the
In alum
systems, the alum will probably not be as likely as polymer
to create the bridges between media grains and projecting
dendrites.
of
In these systems,
parti c l e s
will
be
the accelerated
due to the
accumulation
des t a b i l i z a t i o n of the
initial influent particles which wil l->more readily attach to
the media grains.
, ■
. In the same framework of analysis it can be assumed
that a filter plant using incremental filter opening w i l l
also produce the effect of accelerated filter ripening.
The
partially stabilized particles which are usually sheared off
the
media
grains
at
the
start
of
the
filter
run
under
immediate opening, w i l l be retained within the media in an
.64
incrementalIy opened
system,
thus providing an accelerated
accumulation of particles.
From this discussion, it can be seen that the improving
phase of f i l t r a t i o n is larg e l y due to the conditioning of
the filter media,
both by the accum u l a t i o n
within
and
the
media
by
the
adsorption
directly to the media grains,
of particles
of
coagulants
e s p e c i a l l y with the use of
polymers. '
..
Resnlts and Evaluation of the Use of Backwash
Coagulants
Four
primary
sets
of experiments,
coagu l a n t s
and
each utilizing
turbidity
sources,
different
were
used
to
e v a l u a t e thie use of c oagulants in the backwash water as a
means of reducing the magnitude and duration of the filter
ripening
sequence.
The
Aluminium sulfate (alum),
four
primary
coagulants
were:
Cat Floc TL cationic polymer,
a
combination of alum and Cat Floc TL at a 20:1 ratio, and Cat
Floc T (at the BWTP pilot study).
The
alum
studies
were
conducted
using
an
artificial water composed of tap water and a silica clay
(Min-u-sil 30), while the polymer study was conducted using
tap
water
with
bentonite
as
the
turbidity
source.
The
primary reason for the different turbidity sources was that
the polymer could not effectively remove the Min-u-sil, and
65
al u m could not e f f e c t i v e l y remove the bentonite with the
given water.
The "E" series of experiments conducted at the
BWTP util i z e d the same raw water and primary coagulant as
the plant,
as it was siphoned from the flocculation unit at
the plant. The use of these diverse systems of raw water, in
terms
of
particulate
matter
and
the
different
primary
coagulants,
gives an e x c e l l e n t insight to the workings of
the
ripening
filter
sequence.
These
studies
wi l l
be
discussed individually.
The Polymer Experiments A-I and A-2
In the A-2 study,
to
evaluate
its
ef f e c t i v e n e s s
primary coagulant
(Figure
17),
it
alum was added to the backwash water
system.
when
used
with
a polymer
From the A-2 series of graphs
is apparent
that
a dose
of alum was not
found which reduced the magnitude or duration of the filter
ripening sequence.
known,
possibly
The chemical reasons for this are not
due
to
the
different
mechanisms
destabilization occurring in the two systems.
of
The pH of the
water remained at about 7.5, and the temperature about 13
degrees C. throughout the test.
The A-I studies,
on the other hand,
using polymer in
the backwash water showed a significant reduction of the
magnitude of the filter ripening sequence.
The duration of
EFFLUENT TURBIDITY (NTU)
(79)
(77)
81)
85)
PILOT
PLANT
RUN
PCA6
PA6
PA8
PAll
INFLUENT
TURBIDITY
NTU
BACKWASH
COAGULANT
DOSE mq/1
TIME (MINUTES)
Figure
17.
Determination of Optimum Backwash Alum Dose in
Polymer Primary System.
Experiment A-2.
67
the sequence was not reduced since the length is a direct
function of the time required for the remnant water to be
enti r e l y
removed
influent
water.
from the
filter
Experiments
and its r eplacement
A-la,b,c
illustrate
by
this well
(Figures 18, 19, and 20).
The effective reduction of the filter ripening sequence
magnitude
for
each
polymer
dose
was
determined
by
planimetering the areas under the filter ripening curves out
to
a time
where
These areas,
the
effluent
turbidity
had
stabilized.
as a percent of the control run area with no
coag u l a n t in the backwash w a t e r , were plotted
dose used.
F.rom this graph (Figure 21),
each coagulant could be determined.
versus the
the optimum dose of
The optimum dosages of
backwash coagulants are as follows;
A-2, alum - none,
A■.
1
la, polymer under winter conditions - 0.35 mg/1, • A - 1 b ,
polymer
under
summer
conditions
- 1.8 mg/1,
A- Ic, polymer
under autumn conditions - none.
Since this
seasons,
set of experiments was conducted over three
the varying effects
apparent.
at about
8 C , the
results was about 0.35 mg/1.
of
the
water
optimum
conducted
to
dose
for
the
best
At doses higher than this the
quickly
increased
produced from an overdosing effect.
(A - 5) w e r e
became readily
During the winter/spring season with the water
temperature
turbidity
of the seasons
quantify
due
to
a
fog
A series of experiments
the
effects
of
this
(13)
(27)
(25)
(10)
(22)
(16)
TURBIDITY (NTU)
-PP 15
PILOT
PLANT
RUN
PC 8
PP 15
PP 13
PP 5
PP 10
PP 7
INFLUENT
TURBIDITY
NTU
22
-
19
20
10
18
BACKWASH
COAGULANT
DOSE mq/1
0
1.2
0.8
0.2
0.4
0.1
EFFLUENT
PP 10
6
8
10
TIME (MINUTES)
Figure
18.
Determination of Optimum Backwash Polymer Dose
in Polymer Primary System During Winter Season.
Experiment A-Ia.
PILOT
plant
run
EFFLUENT TURBIDITY (NTU)
PRAV2C7
PRAV2P5
PRAV2P6
PRAV2P7
PRAV2P8
PRAV2P9
PRAV2P10
PRAV2P11
PRAV2P12
PRAV2P13
PRAV2P14
6
8
INFLUENT
TURBIDITY
NTU
19
15
20
23
18
20
37
22
38
16
16
BACKWASH
COAGULANT
DOSE mq/1
0.0
0.17
0.17
0.34
0.085
0.6
0.9
1.3
1.6
2.0
2.5
10
TIME (MINUTES)
Figure
19.
Determination of Optimum Backwash Polymer Dose
in Polymer Primary System During Summer Season.
Experiment A-Ib.
INFLUENT
TURBIDITY
NTU
BACKWASH
COAGULANT
DOSE mq/1
EFFLUENT TURBIDITY (NTU)
PILOT
PLANT
RUN
TIME (MINUTES)
Figure
20.
Determination of Optimum Backwash Polymer Dose
in Polymer Primary System During Autumn Season.
Experiment A- Ic .
% OF CONTROL AREA (TOTAL CURVE TO 14 MINUTES)
POLYMER AS
BACKWASH COAGULANT
A-Ia = winter/sprinq
A-Ib = summer
A-Ic = early fall
ALUM AS BACKWASH
COAGULANT
summer
153% 9
198% 9
10.1 mg/1
Final effluent
%
of control
POLYMER DOSE IN BACKWASH WATER mq/1
10.0
20.0
ALUM DOSE IN BACKWASH WATER mn/l
Figure
21.
Backwash Coagulant Optimization Summary with
Polymer Primary Coagulant System.
72
overd o s i n g
on the
basis
given polymer dosage over
can
be
seen
that
of the turbidity
varying time.
the effects
increase for a
From Figure 22 it
of overdosing
under winter
conditions with this water can be very severe,
especially at
higher doses and at higher standing times.
As the water
conditions
changed and the temperature
rose to about 12 C 1 the effective dose for optimum reduction
of the magnitude of the filter ripening sequence changed.
From the A-Ib
curve
(Figure
21),
it can be seen that the
optimum dose occurs at about 1.8 mg/1, and the overdosing
effect is not present until a much higher dose is reached.
The o v e r a l l
e f f ectiveness of the polymer in the backwash
water for the two seasons is about the s a m e .
the temperature was falling,
obtained.
far different
In autumn as
results were
Regard less of what dose was used, the series of
tests on Figure 21 show that there is not a reduction of the
magnitude or duration of the filter ripening sequence.
The seasonal changes are probably due to a combination
of physio-chemical characteristics of the raw water, such as
t e m p e r a t u r e , hardness, color and d i s s o l v e d solids.
the
Bozeman
water
treatment
plant
different mountain stream sources,
gets
water
Since
from
it may be possible
two
that
variations in mixing of the two waters will effect the water
quality characteristics enough to vary the results of these
tests, especially with the seasonal variations included.
TURBIDITY (NTU) DUE TO POLYMER OVERDOSING
73
T urbidity @
90 minutes
Tur bi d it y @
■10 minutes
T ur bidity @
5 minutes
0.4
0.8
1.2
1.6
2.0
POLYMER DOSE (mg/1) IN BACKWASH WATER
Figure
22.
Polymer Overdosing of Backwash Water During
Winter Season.
Experiment A-5.
74
The Alum Experiments B - I , B-2, and B-3
These studies were conducted using polymer, alum, and a
20:1 combination of alum/polymer respectively as coagulants
in the backwash water when the primary coagulant system was
alum.
The pH of the water
remained
at about
7.5 and the
water temperature at about 12 to 13 degrees C. throughout
the
tests.
In the B - I set of experiments, polymer was added to the
backwash water of the alum primary coagulant system.
The
results of this experiment can be seen on Figure 23.
The
results show a significant reduction of the magnitude and of
the filter ripening sequence.
due
to
the
adsorbing
other
media
This reduction is most likely
conditioning
to the media grains
r e s e a r c h e r s ' [8,
15].
effect
of
the
polymer
that has been described by
This
phenomenenon
is
better
indicated by the "memory effect" noted by Yapijakis [15] and
in studies conducted in this research project.
refers to the
This effect
improved effluent in subsequent runs where no
polymer was injected,
following a filter run where polymer
had been injected in to the backwash water.
The most likely
e x p l a n a t i o n of this o b s e r v a t i o n is the adsorption of the
polymer
onto
the
media
grains.
A small
portion
of the
reduction may be due to the d e s t a b i l i z a t i o n of p a r t i a l l y
stabilized particles occurring in the influent mixing phase.
EFFLUENT TURBIDITY (NTU)
(88)
(122)
(123)
(124)
(125)
(126)
(127)
PILOT
PLANT
RUN
AC2
AP36
AP37
AP38
AP39
AP40
AP41
INFLUENT
TURBIDITY
NTU
BACKWASH
COAGULANT
DOSE mq/1
TIME (MINUTES)
Figure
23.
Determination of Optimum Backwash Polymer Dose
in Alum Primary System During Summer Season.
Experiment B-I.
76
Study
water.
B - 2 was
conducted
using
alum
in the
backwash
These results are much more significant in terms of
the filter ripening sequence reduction than are the polymer
studies as can be seen on Figure 24. This effect has never
been documented before.
an initial
peak
of
During the filter ripening sequence
turbidity
occurs
at
a p p roximately
I
minute, which from the data is apparently a function of the
dose of alum used.
This peak occurs in the remnant phase of
the ripening sequence and is most
likely
due to overdosing
of the clear backwash water with alum and the subsequent
formation of m i c r o - f l o c s composed of a l u m i n i u m hydroxide.
Towards the end of the remnant stage and into the influent
mixing
and
particle
d r a s t i c a l l y reduced,
dispersion
stage,
the
turbidity
is
compared to the control run without
alum in the backwash water.
Since alum is not as likely to
attach to the filter media grains as is polymer, it can be
assumed
that
most
of
this
effect
is
due
to
the
d e s t a b i l i z a t i o n of p articles which had become partial Iy
stabilized
in the influent mixing stage.
The reduction of
the filter ripening sequence magnitude is most likely due to
the properly
destabilized
to the media grains
particles,
floes
by
particles
accele r a t i n g
being quickly attached
the capture
of further
or by the strengthening of the initial
a
longer
contact
time
with
proposed by Francois and Van Haute [7].
the
influent
coagulant
as
The attachment of
EFFLUENT TURBIDITY (NTU)
I A
V V
\
PILOT
PLANT
RUN
(104) A C 18
(105) AA19
(106) AA20
(107) AA21
(111) AA25
(112) AA26
(113) AA27
\
INFLUENT
TURBIDITY
NTU
BACKWASH
COAGULANT
DOSE mg/1
\
4
Figure
8
24.
12
16
20
TIME (MINUTES)
24
28
32
36
Determination of Optimum Backwash Alum Dose in
Alum Primary System During Summer Season.
Experiment B-2.
78
the micro-flocs
produced by the initial
overdosing of alum
in the backwash water to the filter media may also provide
active
sites
1
within
the
media
.
a c c u m u l a t i o n of particles.
experiment
is
the
■
further
accele r a t i n g
the.
Another benefit seen by this
substantial
reduction
of
the
final
effluent turbidity which was not seen in the B-I experiments
using polymer in the backwash water.
Experiments
B - 6 , B - 8 , and
B-10,
were
developed
to
further quantify the effect of particle destabilization with
alum.
Experiment B-IO was planned
to minimize
the contact
time of the al u m injected into the backwash water with the
media by injecting the alum into the backwash water above
the filter media.
Injection of coagulant above the media
a l l o w e d the remnant water containing alum to pass through
the media one time versus twice for the standard experiments
injecting alu m b e l o w the media.
results
previously
described
As can be seen from the
in F i g u r e
essentially no difference in results.
14,
there
is
From this observation
it can be assumed that in the alum systems, the reduction of
the magnitude of the filter
ripening
sequence
is primarily
due to the destabilization of the influent particles which
were
normally
partially
stabilized
in the
influent
mixing
stage, and prob a b l y only m i n i m a l l y by conditioning of the
filter media by contact with the backwash coagulant.
79
Experiment
text,
more
B-8 (Figure
clearly
13) as described earlier in the
shows
the
destabilization
particles occurring in this stage of filtration.
demonstrates
the
reduction
influent particles,
the
cases
where
of the
Zeta
of
The figure
potential
of the
in the initial stages of filtration,
alum
was
used
in
the
the
backwash
in
water.
Experiment B - 6, which varied the remnant v o l u m e remaining
above the media following backwash,
hypothesis.
From
Figure
25
it
again demonstrates this
can
be
seen
that
in
the
experiments where alu m c oagulant was used in the backwash
water,
when the remnant
volume was increased,
the reduction
of the filter ripening sequence magnitude became greater and
greater.
This
probably
results
from the greater
detention
time of the remnant water containing alum with the influent
water
and
the
formation
of
stronger
floes
or
better
destabilized particles within this stage of filtration.
Study
B-3
alum/polymer
was
conducted
in a 20:1 ratio
using
injected
a
combination
into
water of the alum primary coagulant system.
the
of
backwash
It can be seen
that very similar results are obtained as with alum alone by
o bserving
coagulants
Figure
at
26.
this
Thus
ratio
the combination of these two
of
al u m
to
polymer,
does
not
improve the effectiveness of the reduction of the magnitude
of the filter ripening sequence over the use of alum alone.
The
expectation
was
that
the
polymer
would
act'
EFFLUENT TURBIDITY (NTU)
PILOT
INFLUENT
PLANT
TURBIDITY
RUN
NTU
(88) AC2
19
(91) ACS
17
(92) 2AC6
14
(93) 2AC7
19
(94) 3AC8
19
(95) 3AC9
18
(97) OACll
18
(98) O A C I2
17
(117) A A 3 1 .
14
(119) 2 A A 3 3 ? \ 17
(120) 3AA34X \ \ 20
(121) O A A 3 S y
BACKWASH
COAGULANT
DOSE mc|/l
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
18.0
18.0
18.0
18.0
\ \ \ 16
FLUID DEPTH
ABOVE MEDIA
feet
2.17
2.17
4.17
4.17
6.17
6.17
0.17
0.17
2.17
4.17
6.17
0.17
OO
O
12
16
TIME (MINUTES)
Figure
25.
Variation of Remnant Volume Above Filter Media
with Alum Primary Coagulant. Experiment B-6.
EFFLUENT TURBIDITY (NTU)
PILOT
PLANT
RUN
Y
/-
(129)
(130)
(131)
(132)
(133)
(134)
(135)
(136)
INFLUENT
TURBIDITY
NTU
BACKWASH
COAGULANT
DOSE mg/1
AC43
AAP44
AAP45
AAP46
AAP47
AAP48
AAP49
AAP50
TIME (MINUTES)
Figure
26.
Determination of Optimum Backwash Alum/polymer
20:1 Dose in Alum Primary System During Summer
Season.
Experiment B-3.
82
s y n e r g i s t i c a l l y with the alum,
creating polymer bridging
between p articles and from adsorbed polymer on the media
working with the alum to provide
over the sole use of alum.
experiment,
plant.
from
however
an additional
improvement
This was not the case in this
it may work in an actual
filtr a t i o n
As with the alum experiment B-2, a peak of turbidity
overdosing occurs throughout the remnant stage as this
material' is disp l a c e d out of the filter unit.
this
portion
of
the
ripening
sequence,
Following
a substantial
i m p r o v e m e n t in effluent quality is obtained f o l l o w e d by a
s i g n i f i c a n t l y .impr o v e d
control
final
effluent
quality
over
the
runs.
In an operating plant u t i l i z i n g alum or a l u m / p o l y m e r
injection into the
backwash water,
The o v e r a l l
effluent
quality could be s u b s t a n t i a l l y improved by the follo w i n g
measures;
filtering
of turbidity,
water
above
coagulant
effl u e n t
removal
to waste of the coagulant
injection
the
the
media,
last
quality
of the coagulant
I or
or
by
stopping
2 minutes
i m provement
would
into
induced
the backwash
injection
of. the
of backwashing.
be
peak
a result
of
The
the
or suppression of the backwash coagu l a n t induced
peak of turbidity at the beginning of the filter run.
Figure
27 is a graphical
representation of the backwash
c o a g u l a n t dose o ptimizations for each backwash coagulant.
The figure shows a comparison of the areas under the filter
% OF CONTROL AREA (TOTAL CURVE TO 32 MINUTES)
B-I
POLYMER AS
BACKWASH COAGULANT
B-2
ALUM AS
BACKWASH COAGULANT
B-3
ALUM/POLYMER 20:1
BACKWASH COAGULANT
B-3 witho u t initial peak
B-2 without initial peak
Final effluent
% of control
ALUM DOSE IN BACKWASH WATER mq/1
1.0
2.0
3.0
POLYMER DOSE IN BACKWASH WATER mg/1
Figure
27.
Backwash Coagulant Optimization Summary for
Alum Primary Coagulant System.
84
ripening
curves
for
each
coagulant
at
varying
dosages.
Experiments B-2 and B-3 were also plotted as if the initial
peak
of
turbidity
caused
w a t e r , was filtered
by
overdosing
to waste.
of
the
backwash
From a comparison
of the
curves it can be seen that the B-2 and B-3 experiments both
provided better results than the B-I polymer experiments.
The optimization curves for B-2 and B-3 with the initial
peak filtered to waste provide even better results, as from
the graph it can be seen that
these curves are a c t u a l l y
lower than the final quality effluent line.
The Alum/polymer Experiments C-I, C-2, and C-3
A
series
of
experiments
similar
in
polymer and alum studies described earlier,
nature
were
to
the
conducted
using.a combination of a l u m / p o l y m e r (Cat Floc TL 20:1) as
the
primary
coagulant.
This
system
util i z e d
the
same
turbidity source (Min-u-sil 30) and tap water as the alum
studies.
The pH of the water remained at about 7.5 and the
temperature
at about
12 to
13 degrees
C. throughout
the
tests.
The
exhibited
control
runs
for. this
series
of
experiments
a higher quality effluent and a filter
ripening
sequence of less magnitude and duration at a lower coagulant
dosage than the "B" series of experiments u tilizing
alone as the primary coagulant.
alum
85
In the C-I experiments (Figure 28) using polymer in
the backwash water,
the magnitude of the filter ripening
sequence was reduced about 25%.
showing
the
However, from analysis of
Figure
29
backwash
coagulant
curves,
it can be seen that the total reduction of the area
under the curve was less than 10%.
the
filter
considerably
ripening
sequence
less that the total
optimization
The total reduction of
in
experiment
C-I
is
reduction observed in the
alum primary coagulant experiments B—I using polymer in the
backwash water.
This may be due to the polymer
already
present in the primary.coagulant system play i n g the major
role in particle bridging within the media.
The control
runs of the combination system are almost exactly equivalent
to the alu m primary c oagulant
runs using polymer
in the
backwash water (Figure 23), in terms of the magnitude and
duration of the filter ripening sequence.
This is a further
indication of the polymer in the primary coagulant system
alre a d y
playing
a major
role
in- the . reduction
of
the
magnitude and duration of the filter ripening sequence.
Experiments C-2 and C-3 also exhibited the initial
turbidity peak seen in experiments B-2 and B-3, due to the
overdosing
coagulants
of the backwash
but
not w i t h
water
the
alum
and
addition.of
contrast to the B-2 and B-3 experiments,
alum/polymer
polymer.
In
the C-2 and C-3
PILOT
PLANT
RUN
(157)
(159)
(160)
(161)
(162)
INFLUENT
TURBIDITY
NTU
BACKWASH
COAGULANT
DOSE mq/I
APC15
APP17
APP18
APP19
APP20
TIME (MINUTES)
Figure
28.
Determination of Optimum Backwash Polymer Dose
in Alum/polymer 20:1 Primary System During
Summer Season.
Experiment C-I.
I
% OF CONTROL AREA (TOTAL CURVE TO 32
C-3 alum/polymer
C-I polymer
C-3 without initial
peak
Final effluent % of control
C-2 without
initial peak
1.5
2.0
2.5
3.0
POLYMER DOSE IN BACKWASH WATER mq/1
5
10
15
20
25
30
35
40
ALUM DOSE IN BACKWASH WATER mg/1 (ALUM/POLYMER 20:1)
Figure
29.
Backwash Coagulant Optimization Summary for
Alum/polymer 20:1 Primary Coagulant System.
88
experiments (Figures 30 and 31),
displayed lower initial
turbidity peaks than the ones using only alum.
The C-2 and C-3 experiments using alum and alum/polymer
respectively
in
results.
From
including
the
the
backwash
Figure
initial
29
water
in
the
turbidity
show
very
different
optimization
peak
the
curves
alum/polymer
combination undoubtedly provides a more optimum reduction of
the filter ripening sequence at a lower coagulant dosage
than does alu m alone. However, when the initial turbidity
peak is considered as filtered to waste,
approximately the same.
both curves are
This indicates that a small amount
of polymer in the backwash w a t e r , when a l u m is used as a
backwash
coagulant,
may
reduce
the
initial
peak
of
overdosing.
The
Bozeman
Water
Treatment
Plant
E x p e r i m e n ts
E-l,
E-2, and E-3
The
Bozeman
pilot
Water
experiments.
plant
filter
Treatment
unit was transferred to the
Plant
for.
thi s
series
of
This series was conducted in a similar fashion
as were the preceding experiments,
but instead of filtering
an a rtificial water, the influent for the pilot plant was
the same influent
as used in the Bozeman
plant
filters.
This series of experiments was intended as a c omparison of
EFFLUENT TURBIDITY (NTU)
PILOT
INFLUENT
TURBIDITY
PLANT
NTU
RUN
(150) APA8
(152) APAlO
(153) APAll
(154) APA12
(155) A P A I3
(156) APA14
(157) APC15
BACKWASH
COAGULANT
DOSE mg/1
TIME (MINUTES)
Figure
30.
Determination of Optimum Backwash Alum Dose in
Alum/polymer 20:1 Primary System During Summer
Season.
Experiment C-2.
EFFLUENT T URBIDITY (NTU)
(143)
(144)
(145)
(146)
(147)
(148)
(149)
PILOT
PLANT
RUN
APCl
APAP2
APC3
APAP4
APAP5
APAP6
APAP7
INFLUENT
TURBIDITY
NTU
BACKWASH
COAGULANT
DOSE mci/1
12
16
TIME (MINUTES)
Figure
31.
Determination of Optimum Backwash Alum/polymer
Dose in Alum/polymer 20:1 Primary System During
Summer Season.
Experiment C—3.
91 •
the results achieved with artificial waters in the lab to
results obtained with an actual water filtered by a plant.
During this series of experiments conducted in early
autumn,
and
the water influent turbidity fluctuated between 2.5
3.9 NTU,
the pH remained at about 6.7,
and the water
temperature fluctuated between 3 and 4 degrees C . . The plant
influent
was
flocculated
with
3.5
mg/l
of
Cat
Floc
T
polymer.
The
34,
overall
results
of
and 35) did not compare
the
with
tests (Figures 32, 33,
the
series
of polymer
primary coagulant experiments (A-1 and A-2, Figures 17,
19,
and
20)
experiments
conducted
displayed
in
the
a truncated
lab.
filter
The
lab
polymer
ripening
sequence
which ended at the culmin a t i o n of the remnant stage.
study
conducted
using
Bozeman plant water,
18,
The
on the other
h a n d , was similar in nature to the alum and a l u m / p o l y m e r
studies (B-1,2,3 and C-1,2,3) conducted in the lab as the
filter ripening curves displayed the entire filter ripening
sequence.
This difference in results is most likely due to
the differing influent turbidities at the plant and
laboratory
pilot
plant
studies.
The
pilot
in the
plant
was
operated in the lab with turbidities ranging from 10 to 30
NTU, the Bozeman water treatment plant influent was about 3
NTU.
In
resulted
the
pilot
plant
studies,
in the formation of very
the
higher
turbidity
large floes in the upper
EFFLUENT TURBIDITY (NTU)
PILOT
PLANT
RUN
INFLUENT
TURBIDITY
NTU
(185)
(186)
3.
3.
BACKWASH
COAGULANT
DOSE mq/1
D-IC DYE TRACE
IO
KJ
TIME (MINUTES)
Figure
32.
Determination of Optimum Backwash Polymer Dose
in Bozeman Water Treatment Plant Influent System
During Autumn Season. Experiment E-I.
INFLUENT
TURBIDITY
NTU
BACKWASH
COAGULANT
DOSE mq/1
EFFLUENT TURBIDITY (NTU)
PILOT
PLANT
RUN
TIME (MINUTES)
Figure
33.
Determination of Optimum Backwash Alum Dose in
Bozeman Water Treatment Plant Influent System
During Autumn Season.
Experiment E-2.
.---- —
0
E-2 ALUM AS
BACKWASH COAGULANT
E-3 ALUM/POLYMER 20:1
AS BACKWASH COAGULANT
=5 ^-
E-3 without
initial peak / ^ _ ^ / E-2
E-2 without
initial peak
%
OF CONTROL AREA (TOTAL CURVE TO 60 MIN
E-I POLYMER AS
BACKWASH COAGULANT
ALUM DOSE IN BACKWASH WATER mg/1
1.0
2.0
3.0
POLYMER DOSE IN BACKWASH WATER mo/1
Figure
34.
Backwash Coagulant Optimization Summary With
Bozeman Water Treatment Plant Influent System.
INFLUENT
TURBIDITY
NTU
BACKWASH
COAGULANT
DOSE mg/1
EFFLUENT TURBIDITY (NTU)
PILOT
PLANT
RUN
TIME (MINUTES)
Figure
35.
Determination of Optimum Backwash Alum/polymer
20:1 Dose in Bozeman Water Treatment Plant
Influent System During Autumn.
Experiment E-3.
96
portion of the filter unit which quickly
of the
filter
media.
media
as soon
plugged
as the influent
the pores
reached
the
The influent at the plant did not display these
large floes,
thus the accumulation of particles in the pores
would occur at a much slower rate.
The use of flocculators
at the plant in contrast to the in-line f iltration in the
laboratory study may also be another factor in the differing
results.
• The
E-I
series
of
experiments
was
polymer (Cat Floc T) in the backwash water.
conducted
using
From Figure 32
it can be seen that the beginning of the initial turbidity
peak coincides with
trace
study.
. This
the initial
indicates,
arrival
of the D-IC dye
as described before,
primary peak of turbidity is a function
that the
of the influent.
The E-I series of experiments failed to find a suitable dose
of polymer added to the backwash water which would reduce
the magnitude and duration of the filter ripening sequence.
These results compare with the A-Ic series of experiments
(Figure 20) conducted in the autumn season in the 'lab, even
though
the
two
curves
are
very
different.
It may
be
d ifficult to equate the results obtained in the lab using
polymer
in the backwash water with the results
the plant for all
of the curves.
obtained
at
three seasons due to the differing nature
It may be possible, however,
that since a
reduction in the filter ripening sequence was observed in
97
the winter and summer season tests in the lab,
that similar
results may be achieved at the plant during these seasons.
Experiment E-2'was conducted using alum in the backwash
water.
initial
By observation of Figure 33, it can be seen that, the
peak of turbidity
caused
by overdosing
is present.
It is also apparent that some reduction of the magnitude and
duration of the filter ripening sequence is achieved by the
use of alum in the backwash water.
Figure 34
shows the backwash coagulant optimization
curves for the E series of experiments.
From the E-2 curves
plot t e d in Figure 34, it can be seen that the optimum dose
of alum is about 22 mg/1.
to
such
a narrow
diff i c u l t
range
Since this reduction is confined
of
for the Bozeman
coagulant
plant
dosage,
to util i z e
it
alum
may
in
be
the
backwash water, as the changing conditions of the water most
likely would affect this dose almost daily.
Figure
35
shows
the
results
for
the
E-3
series
of
experiments util i z i n g a 20:1 ratio of a l u m / p o l y m e r in the
backwash water.
From o b s e r v a t i o n of Figures 32 and 33 it
can be seen that this coagulant reduced the magnitude and
duration
of
the
filter
ripening
sequence
by
a moderate
degree, but not as great as was achieved using alum alone.
The results,
however, were much more consistent in terms.of
reduction of the filter ripening sequence over a wider range
of coagulant
dosages,
making this system more feasible for
98
plant use than alu m alone.
was
again
pres e n t ,
but
The initial peak of turbidity
it was
not
as
large
as
in
the
In summarizing the backwash coagulant experiments,
the
experiments utilizing alum alone.
Backwash Coagulant Summary
evidence indicates that the optimum coagulant for use in the
backwash
water
filtration.
is
the
same
coagulant
used
for
primary
This phenomenon is probably due to the effects
of the backwash water remnants interfacing with the influent
water in the remnant volu m e above the filter media in the
initial stages of a filter run.
Since the primary coagulant
was (or should have been) optimized for the best quality
effluent,
the
chemical
conditions
that
exist
at
the
interface of the two fluids should be the same when the same
coagulant is used in the backwash water.
different
systems,
c o agulants
in
the
from
the
primary
while providing some reduction in the magnitude of
the filter ripening sequence,
the
backwash
Systems utilizing
optimum conditions
will
not necessarily provide
for the maintenance
of destabilized
particles and the formation of stronger floes in the initial
influent.
The
acceleration
of .the
a c cumulation
of
particles
within the media is a p p a r e n t l y by far the most important
aspect of the reduction of the filter ripening sequence.
99
Thus,
it can be proposed that the proper physico - c h e m i c a l
conditions
must
exist
in
the
initial
influent
optimum accumulation
of particles to occur.
of
filter
polymer
to
the
media
may
for
the
The adsorption
also
serve
as
an
accelerator of particle accumulation.
Optimizing the Backwash Coagulant Injection Time
A series of experiments were conducted with alum and
polymer as the primary coagulants (A-4 and B-4), in terms of
optimizing
the
time
backwash water.
as
determined
of
injection
of .coagu l a n t
into
the
The optimum doses of alum and/or polymer,
in
the
backwash
coagulant
optimization
experiments A-I and B-2 (Figures 21 and 27), were injected
into the backwash water of the filter u n i t .at varying times
measured from the end of the backwashing phase.
of these experiments,
The results
as explained earlier, were plotted on
graph paper as turbidity
versus time (Figures 36 and 37).
The areas under the curves were planimetered and compared to
the initial
control
were injected
minutes.
run in which the respective coagulants
into the backwash water
for the full
five
These v a l u e s were plot t e d as percent' of control
area versus time of injection from the end of the backwash.
The results of these curves are plotted and compared to the
dye
trace
D-3,
in
Figure
38.
As was
expected,
in
both
INFLUENT
POLYMER
TURBIDITY
INJECTION
NTU
TIME MIN.
none
full 5 min.
last 4 min.
last 3 min.
last 2 min.
last I min.
last 0.5 min
BACKWASH
COAGULANT
DOSE mg/1
OCT
EFFLUENT TURBIDITY (NTU)
PILOT
PLANT
RUN
TIME (MINUTES)
Figure
36.
Optimization of Backwash Coagulant Injection Time
with Polymer as Primary and Backwash Coagulant
Systems.
Experiment A-4.
ALUM
INJECTION
TIME
full 5 min
none
INFLUENT
TURBIDITY
NTU
BACKWASH
COAGULANT
DOSE mq/1
TOT
107)
(114)
(115)
(116)
(117)
PILOT
PLANT
RUN
AA21
AA28
AA29
A A 30
AA31
TIME (MINUTES)
Figure
37.
Optimization of Backwash Coagulant Injection Time
with Alum as Primary and Backwash Coagulant
Systems.
Experiment B-4.
% OF MAXIM U M DYE CONCENTRATION (PPB)
% OF CONTROL AREA
(TOTAL CURVE TO 32MIN. ALUM, 14 MIN. POLYMER)
1
Figure
2
38.
3
4
5
6
7
TIME (MINUTES) FROM END OF BACKWASH
Summary of Optimization of Backwash Coagulant
Injection Time.
102
EXPERIMENT D-3
Backwash dye trace
at 20 liters/minute
EXPERIMENT A-4
POLYMER PRIMARY COAGULANT
Variation of time of
injection of polymer 0.4 mg/1
into backwash water
EXPERIMENT B-4
ALUM PRIMARY COAGULANT
Variation of time of
injection of alum 18 mg/1
into backwash water
103
ca.es,
the o p t i m a injection time is essential!, the same as
the minimum time necessary for the water in the filter unit
to
be c o m p l e t e l y
displaced
b, the backwash water.
indicates that the coagulant must be evenly
dispersed
the backwash remnant water for the optimum results
achieved,
in
both
the
alum
and
polymer
This
studies.
into
to be
This
further illustrates the importance of providing the correct
P h y s i c o - c h e m i c a l e n v i r o n m e n t in the upper
portion of the
remnant water which first contacts the i n f l u e n t water,
in
order to achieve the optimum reduction of the magnitude and
duration of the filter ripening sequence.
Variation of Remnant Volume Above MeHi=
From the study B-6 (Figures 8 and 25) it is apparent
that the remnant
volume which is left in place above the
media following backwash plays a very important role in the
ac.te r i s 11 c s
of
the
filter
ripening
sequence.
The
primary characteristic affected by this volume is the timing
at which
events
occur
within
the
filter ripening
sequence.'
From the Prior discussion concerning the filter ripening
sequence theory,
it was seen that the length of the remnant
stage will theoretically be dependent on the length of time
it takes for the remnant water to he displaced from, and the
influent water.to be displaced into,
the filter unit.
.This
104
time w i l l be dependent on the physical detention time and
the dispersion characteristics of a particular filter unit.
Data
derived
polymer,
from
the
experiments
involving
alum
and
when combined with the dye trace study D-3 (Figure
8) shows conclusively that the length of the remant stage is
directly proportional to the detention time and thus to the
volume of water which remains above the media at the end of
the backwash.
In
the polymer experiments A-la,b,c,
2, (Figures 17, 18,
and A-
19, 20) this detention time predicted
the end of remnant stage, w h i l e in the alum studies B — 1 , 2,
.3, (Figures 23,
24,
26) it predicted the beginning of the
particle dispersion stage.
In both cases it does h o w e v e r ,
c l e a r l y predict the end of the remnant stage.
Varying the
remnant v o l u m e above the media does not appear to greatly
affect the magnitude of the filter ripening sequence.
the
alum
turbidity
studies,
but
a
however,
lengthed
a
slight
duration
in
decrease
the
in
total
In
peak
filter
ripening sequence occurs with increasing v o l u m e above the
media. The total number of particles passing through the
filter in each case wi l l therefore be about the same.
Another factor arises in the case where a coagulant is
used in the backwash water.
involving
The alum studies were conducted
injection of the optimum alum dose determined in
experiment B-2 into the backwash water.
It was found i n .the
experiment
that
series
B-6,
(Figure. 25)
the
effluent
105
actually improved as the remnant volume above the media was
increased.
This may be attributed to better mixing of the
influent water with the remnant water at their interface
with an increased detention time.
contains
the
opportunity
optimum
is
alum
developed
Since the remnant water
coagulant
for
dose,
a
destabilization
particles at the remnant water/influent interface.
better
of
the
This is
also an indication that a majority of the effects of adding
alum to the backwash water can be directly attributed to the
reduction of the particle dispersion phase,
and not with
preconditioning of the filter media with alum.
In
t e r m s of
a
plant
design
incorporating
injection of coagulants into the backwash water,
On the other hand,
of
increasing
the depth above the media to the backwash gutters
worthy endeavor.
the
may be a
a benefit can not be
seen for decreasing the vol u m e of remnant water above the
media in a conventional filtration system.
Variation of the Backwash Water Volume
In the work of Amirtharajah and Wetstein
[3] it was
proposed that the magnitude of the second peak of the filter
ripening
sequence
was
proportional
to
the
number
of
particles remaining above the media in the remnant water at
the end of backwash.
It thus
could
be assumed
that
the
106
magnitude of this curve would be reduced by a longer washing
period decreasing the particle concentration in this portion
of the filter unit.
On the other hand, reducing the time of
backwash would leave more particles in the remnant water and
thus would increase the magnitude of the filter ripening,
sequence.
A series of experiments
A - 3 , and B - 5, (Figures 39. and
40) were conducted to determine the effect of varying the
total vol u m e (length of time at 20 liters /min (21 gpm/sq.
ft.)) used to backwash the filter.
The results
of these
tests, in terms of the planimetered areas under the filter
ripening curve compared to a control run,
were compared
with the B-6/D-3 plot showing the turbidity of the effluent
backwash water versus time (Figure 11).
polymer
A-3
studies
(Figure
39)
A comparison of the
indicates
only
a small
increase in the magnitude of the filter ripening sequence
with a very large increase in remnant water turbidity.
The
alum study B - 5 , (Figure 40) on the other hand, shows a very,
close correlation between the turbidity of the remnant water
and the magnitude of the filter ripening sequence as seen on
Figure 11.
This c o rrelation c l o s e l y corresponds with the
results
Amirtharajah
particles
of
and
Wetstein
coagulated with alum.
[3]
who
used
iron
The difference in results
between the alum and polymer experiments may be due to the
polymers forming larger stronger floes very quickly, thus
INFLUENT
TURBIDITY
NTU
BACKWASHING
TIME
MINUTES
107
EFFLUENT TURBIDITY (NTU)
31)
32)
33)
34)
35)
PILOT
PLANT
RUN
BWTV2
B WT V 3
BWTV4
BWTV5
BWTV6
TIME (MINUTES)
Figure
39.
Variation of the Volume of Backwash Water with
Polymer as Primary Coagulant.
Experiment A-3.
PILOT
PLANT
RUN
BACKWASHING
TIME
MINUTES
A C 14
AC15
AC16
AC17
108
EFFLUENT TURBIDITY (NTU)
100)
101)
(102)
(103)
INFLUENT
TURBIDITY
NTU
TIME (MINUTES)
Figure
40.
Variation of the Volume of Backwash Water with
Alum as Primary Coagulant.
Experiment B-5.
109
clog g i n g the media pores much more quickly than the alum
destabilized particles.
as noted before,
other
experiments
sequences.
Thus
The series of polymer experiments,
did not closely follow the results of the
because
the
of the
polymer
truncated
experiments
filter
are
ripening
not
to
be
considered as important as the other experiments in terms of
results.
HO
CHAPTER 6
PRACTICAL APPLICATION SUMMARY
The data derived in this study is intended to be more
closely
related
theory.
It. is therfore only fitting that it be described, in
such a mariner.
to
practical
a p plication
There are b a s i c a l l y
five
than
to
categories
practical
application'which
can be used to optimize
reduction
of
and
the
ripening sequence.
magnitude
pure
duration
of
the
of
the
filter
They are as follows:
1.
Optimum backwash coagulants.
2.
Optimum injection time of backwash coagulants.
3.
Optimum time or volume used to backwash a filter unit.
4.
Optimum
remnant volume left above the
media
following
backwash.
5.
Incremental filter opening.
Each of these categories will be discussed individually.
Optxanni Backwash Coagplants
For each i n d i v i d u a l
system of primary coagulant,
and
particle
f i ltration plant,
a different
influent water characteristics
characteristics
wil l
affect
the
type
and
quantity of c oagulant to be used in the backwash water for
Ill
optimum results.
The influent water characteristics,
as
shown by the results of this s t u d y , can vary seasonally,
thus the optimum coagulant type and dosage may be affected.
Due to the v a r i a b i l i t y of each system,
it is recommended
that a pilot plant study be conducted on a seasonal basis to
determine the optimum coagulant type and dosage to be used.
In general it may be assumed that the primary coagulant
system used in the plant would give the optimum results when
used as a backwash coagulant.
This would give a starting
point for the pilot plant study.
In systems where alum or
a l u m / p o l y m e r are to be used in the backwash water,
it is
crucial that the first period of increased turbidity caused
by the overd o s i n g of the backwash water with coagulant be
filtered to waste.
If this water is added to the clearwel I
the increased coagulant dosage in this water may induce the
same overdosing affect in the entire clearwelI volume.
length
of
time
required
for
filtering
to
waste
can
The
be
determined by conducting a tracer survey on the filter unit
in question.
The detention time required for the remnant
backwash water to be displaced from the filter unit will be
the time required for remo v a l of the induced turbidity of
overdosing.
112
Optimum Injection Time of Backwash Coagplants
It
has
been
determined
that
the
optimum
time
injection of coagulant into the backwash water is also
of
the
same time required to displace the backwash water into the
entire
filter
volume.
This
time
can
be
determined
by
conducting a tracer survey on the filter unit and using the
80 to 90% displa c e m e n t
injection.
By injecting
..only during this period,
ripening
as the
sequence
and
time
coagulant
required
into
for optimum
the backwash water
the maximum reduction of the filter
the
lowest
chemical
usage
can be
simultaneously achieved.
Optimum Backwash Volume
It
has
been
determined
magnitude and duration
in
of the
this
filter
research
ripening
that
sequence can
be a function of turbidity remaining in the remnant
abo v e
the
magnitude
media
following, backwash.
of the filter
ripening
In
sequence
the
order
volume
for
the
to be minimized,
the turbidity of the remnant volume must also be minimized.
The
minimum
volume
of. water
required
reduction of the remnant turbidity is
for
the
maximum
related to the time
required to effectively displace the filtration remnants out
of the filter unit, with clean backwash water.
be determined by a tracer survey,
by m onitoring
the
effluent
This time can
or more easily determined
turbidity
from the
backwash.
113
Once the effluent turbidity has been reduced by about 95%,
no further reduction of the magnitude of the filter ripening
sequence will be achieved.
Optimization of Reanant Volume
From this
study it was determined
that reducing the
remnant v o l u m e above the media w i l l not have a very large
effect on reduction of the magnitude of the filter ripening
sequence.
The reduction of this volume will only reduce the
time required for the primary degradation peak to occur.
On
the other hand, if a plant is to be designed with the intent
of adding, coagulants to the backwash water, an increase of
the
volume
desirable.
of
the remnant water
The increased volume
above
the
media
may
be
can allow a longer mixing
time of the influent water with the remnant water, forming
stronger
floes
and
better
reducing the magnitude
d e s tabilized
and duration
of the
sequence possibly by a substantial
particles,
thus
filter ripening
degree.
It. should be
noted here that the data derived from these experiments was
for
a
direct
filtration
pilot
plant
and
may
not
be
applicable to a conventional filter plant.
Incremental Filter Opening ,
If a filter unit is opened at a low rate of flow and
very gradually increased to full flow over a pre-determined
114
period of time,
occur.
First,
at least
two beneficial mechanisms w i l l
the initial
influent
particles which are
poorly destabilized or weakly flocculated will have a better
chance of adhering to the media grains as they pass through
the
filter
due
to the lower hydrodynamic shear.
the parti c l e s which have remained attached
grains
at
the
opportunity
end
of
backwashing
to re-oreint
will
to the media
have
On the given media
Secondly,
a better
grain or an
adjacent grain while the hydrodynamic shear within the pores
of the filter media is relatively low.
part i c l e s
in
the
filter
media
The accumulation of
by these mechanisms
accelerate the ripening of the filter unit.
will
115
CHAPTER 7
CONCLUSIONS
The results of this research have further described the"
mechanisms of and have given methods for the control of the
initial degradation of effluent quality from a backwashed
deep-bed
filter.
The
c o nclusions
of
this
study
are
as
follows:
I.
The
initial
effluent
degradation,
and
subsequent
effluent quality improvement, termed the "filter ripening
sequence" in this research, has been further quantified in
terms of the following steps:
A.
The remnant stage.
This
stage
of
with
the
associated
remaining
relatively
remnant
in the underdrains,
low
water
turbidity
of
backwashing
within the media,
above the media to the backwash gutter.
is
and
The particles
within this stage have been partially stabilized by the
coagulant-free backwash water and thus will more easily
pass
through
the
filter
media.
A
pea k
of
hig h
turbidity may occur within this stage due to particles
sheared
off
of
the
media
at
the
beginning
of
the
116
f iltration
cycle,
or at
the end of backwash as the
media particles collide with each other.
B.
Influent
As
mixing
the
and
influent
coagulant-free
particle
water
remnant
water
''
stabilization
disperses
above
the
stage.
into
the
media,
the
parti c l e s in the influent water which had p r e v i o u s l y
been d e s t a b i l i z e d by primary coagulants wil l become
partially stabilized or the floes will become weakened,
a l l o w i n g these p articles to easily pass through the
filter media.
This mechanism results in the largest
peak of turbidity in the filter ripening sequence.
C.
Filter media conditioning stage.
This
stage
ripening".
is
also
referred
to
as
"filter
This stage of the filter ripening sequence,
is associated with the accumulation of particles within
the pores of the media resulting in a gradual reduction
in effluent turbidity until a stable effluent quality
is
2.
of
The
obtained.
use of coagulants in the backwash water as a means
reducing
ripening
magnitude
sequence
determined
effective
the
that
as
can
other
backwash
be
and
very
duration
effective.
coagulants
and
filter
has
polymers
in
It appears
the
It
besides
coagulants,
s i g n i f i c a n t l y more effective.
of
some
as
been
are
cases,
though
the
117
optimum coagulant type will
primary
coagulant
be the same as that used as the
for filtration.
The backwash
coagulants
appear to work by preventing stabilization of the initial
influent particles and forming stronger floes between these
particles as they disperse into the backwash remnant water.
Poly m e r s
effect,
tend
to work
not
only
by
this
destabi l i z a t i o n
but also by adsorption to the filter media providing
sites for influent particle attachment.
effect
of
these
mechanisms
allows
The synergistic
acceleration
of
the
a c c u m u l a t i o n of particles within the media resulting in a
reduced
magnitude
and
duration
of
the
filter
ripening
sequence.
3.
It was determined that the optimum time, for injection of
coagu l a n t into the backwash water, in terms of the maximum
reduction
of
the
ripening sequence,
magnitude
and
duration
of
the
filter
and the minimal coagulant usage, was the
same time required to c o m p l e t e l y displace the filter unit
volume with backwash water.
4.
It was determined that the optimum volume of water used
to backwash the filter, in terms of the maximum reduction of
the magnitude and duration
of the
filter ripening sequence
and the mini m a l water use, is the same as that required to
displace
unit.
the majority
of the p a r t i c l e s , . from the filter
• 118
5.
It
was
determined
remaining above
wi l l
not
have
that
varying
the
remnant
vol u m e
the filter media at the end of backwash
a significant
affect
on the magnitude
duration of the filter ripening sequence.
or
The variation of
this volu m e w i l l only shift the time at which the events
occur within the filter
determined
that
in
ripening
direct
sequence.
filtration
coagulants in the backwash water,
It was
systems
also
using
increasing the volume of
remnant water above the media can significant Iy enhance t.hs
effects of the backwash coagulants.
6.
Based
on data obtained at the Helena,
Montana Water
Treatment Plant [4], an incremental opening of a filter unit
versus
an instantaneous
opening may possibly
reduce the.
magnitude and duration of the filter ripening sequence by a
significant
degree.
REFERENCES CITED
120
REFERENCES CITED
1.
Amirtharajah,
A.,
"Optimum Backwashing of Sand Fil­
ters",
Journal of Environmental Engineering Division
of ASCE, October,.1978.
2.
Amirtharajah, A., "The. Interface Between Filtration and
Backwashing",
Water R e s e a r c h , V o 1 . 19, No. 5, 1985,
pp 581.
3.
Amirtharajah, A., Wetstein, D. P., "Initial Degradation
of Effluent Quality During Filtration." Journal of the
American Water Works A s s o c i a t i o n , September, 1980, pp
518.
4.
Bucklin, K.,
A summary of data collected from the
Bozeman and Helena Water Treatment plants for Master's
Thesis at Montana State University, 1986.
5.
Chen, C . T.,
"Filter Preconditioning to Reduce Initial
Degradation in Effluent Water Quality." Masters Thesis
at U n i v e r s i t y of Cincinnati, Department of Civil and
Environmental Engineering, 1986.
6.
Fair,
G. M.,
Geyer, J. C., Okun, D. A.,
Wastewater
Engineering; Water Purification and Wastewater Treat­
ment and Disposal
John W i l e y Publisher, New York,
1967.
7.
Francois,
R . J.,
Van Haute, A. A., "Backwashing and
Conditioning of Deep Bed Filter."
Water R e s e a r c h ,
Vo 1 . 19, No. 11, 1985, pp 1357.
8.
Harris, W. L., "High Rate Filter Efficiency"
Journal
of the American Water Works A s s o c i a t i o n , 62:8:515,
August, 1970.
9.
Logsdon,
G . S., Rice, E. W., "Evaluation of Sedimenta­
tion and Filtration for Microorganism removal."
To be
published by USEPA Drinking Water Research Division,
Cincinnati, Ohio, 1985.
121
10.
0'Melia,
G.
R.,
"Particles,
Pretreatment,
and
Performance in Water Filtration.", Journal of the
Environmental Engineering Division of ASCE, Vol. Ill,
No. 6,1985, pp 874.
11.
O ’Melia, C. R., All, W. "the Role of Retained Particles
in Deep Bed Filtration",
Progress in Water Technology,
Vol. 10, No. 5/6, 1978, pp 167.
12.
Payatakes,
A.
C.,
Park, H. Y., Petrie, J., "A Visual
Study of Particle Deposition and Reentrainment During
Depth Filtration of Hydrosols with a Polyelectrolyte/'
Chemical Engineering Science, 36:1319:1981.
13.
Regan,
M. M., "Optimization of Particle Detachment by
C o l l a p s e - P u l s i n g During Air Scour", M a s t e r ’s Thesis
Submitted to Montana State University, 1984.
14.
Trusler, S . L.,
"Turbulent Rapid Mixing in
Direct
Filtration", Master’s Thesis Submitted to Montana State
University, 1983.
15.
Yapijakis, C., "Direct Filtration:
Polymer in Backwash
Serves Dual Purpose."
Journal of the American Water
Works Association, August, 1982, pp 426.
122
SELECTED BIBLIOGRAPHY
SELECTED BIBLIOGRAPHY
Black,
A.. P ., Birkner,
F . -B.,
Morgan,
J . J .,
"Destabilization of Dilute Clay Suspensions With
Labeled Polymers." Journal of the American Water Works
Association, V o 1 . 57, No. 19, 1965,pp 1547;
Cleasby, J. L.,
"Backwashing
of Granular Filters"
Journal of the American Water Works A s s o c i a t i o n ,
69:2:115, February,•1977
Cleasby, J . L.,
Baumann,
E., "Backwash ofGranular
Filters Used in Wastewater Filtration."
USEPA report
No. EPA-600/2-77-016, Office of Resource Development,
USEPA Cincinnati, Ohio, April, 1977.
Edzwald, J . K.,
"ConventionalWater Treatment
and
Direct Filtration:
Treatment and Removal of Total
Organic Carbon and Trihalomethane Precursors." Organic
Carcinogens in Drinking Water:
Detection, Treatment,
and Risk A s s e s s m e n t ,' N . M . R a m , E. Calabrese, and R . F .
Christman, Editors, Wiley Publishing, New York, 1986.
Edzwald, J.
K., Becker,
W. C., Tambini, S . J.,
"Orga n i c s ,
Polymers,
and P e r f o r m a n c e in Direct
Filtration." Feb. 1986, To be published in Journal of
the Environmental Engineering Division of ASCE.
Ghosh,
M.
M.,
Jordan,
T. A.,
Porter,
R. L.,
"Physico-chemical Approach to Water and Wastewater
Filtration."
Journal of the Environmental Engineering
Division of A S C E , V o 1 . 101, EE I, 1975, pp 71.
Habibian, M . T., 0'Melia, C. R., "Particles, Polymers,
and P e r f o r m a n c e in F i l t r a t i o n . "
J o u r n a l of the
Environmental Engineering Division of ASCE, Vol.- 101,
1975, pp 567.
Hudson,
H. E.,
"Filter Washing Experiments at the
Chicago Experimental Filtration Plant." Journal of the
American Water Works Association, 27:11:1547, November,
1935.
124
10.
0 ’Melia,
C.
R .,
"Particles,
Pretreatment,
and
Performance in Water Filtration.", Journal of the
Environmental Engineering Division of ASCE, Vol. Ill,
No. 6,1985, pp 874.
11.
O'Melia, C. R., Ali, W. "The Role of Retained Particles
in Deep Bed Filtration",
Progress in Water Technology,
Vol. 10, No. 5/6, 1978, pp 167.
12.
Payatakes, A. C.,
Park, H. Y., Petrie, J., "A Visual
Study of Particle Deposition and Reentrainment During
Depth Filtration of Hydrosols with a PoIy electrolyte."
Chemical Engineering Science, 36:1319:1981.
13.
Regan,
M. M., "Optimization of Particle Detachment by
C o l l a p s e - P u l s i n g During Air Scour", Master's Thesis
Submitted to Montana State University, 1984.
14.
Trusler, S . L.,
"Turbulent Rapid Mixing in
Direct
Filtration", Master's Thesis Submitted to Montana State
University, 1983.
15.
Yapijakis, C., "Direct Filtration:
Polymer in Backwash
Serves Dual Purpose."
Journal of the American Water
Works Association, August, 1982, pp 426.
MONTANA STATE UNIVERSITY LIBRARIES
1762 1001 3844 3