Transcript Water Treatment

Water Treatment
Monroe L. Weber-Shirk
School of Civil and
Environmental Engineering
Where are we going?
 Unit processes* designed to
 remove _________________________
Particles and pathogens
 remove __________
___________
dissolved chemicals
 inactivate ____________
pathogens
 *Unit process: a process that is used in similar
ways in many different applications
 Unit Processes Designed to Remove Particulate
Matter
 Screening
 Sedimentation
 Coagulation/flocculation
 Filtration
Empirical design
Theories developed later
Conventional Surface Water
Treatment
Raw water
Filtration
Screening
Alum
Polymers
sludge
Coagulation
sludge
Cl2
Disinfection
Flocculation
Storage
Sedimentation
Distribution
sludge
Screening
 Removes large solids
 logs
 branches
 rags
 fish
 Simple process
 may incorporate a mechanized trash
removal system
 Protects pumps and pipes in WTP
Sedimentation
the oldest form of water treatment
uses gravity to separate particles from water
often follows coagulation and flocculation
Sedimentation: Effect of the
particle concentration
Dilute suspensions
Particles act independently
Concentrated suspensions
Particle-particle interactions are significant
Particles may collide and stick together
(form flocs)
Particle flocs may settle more quickly
At very high concentrations particleparticle forces may prevent further
consolidation
Sedimentation:
Particle Terminal Fall Velocity
 F  ma
Identify forces
Fd  Fb  W  0
Fb
ρ p  particle density
Fd
ρw  water density
p p g
W  _______
g  acceleration due to gravity
C D  drag coefficient
pr wg
Fb = "________
Fd  CD AP  w
Vt 2
2
 p  particle volume
projected
Ap  particle cross sectional area
Vt  particle terminal velocity
W
4 gd ( r p - r w )
Vt =
3 CD
rw
Drag Coefficient on a Sphere
4 gd ( r p - r w )
Vt =
3 CD
rw
18
Stokes Law
100
10
1
laminar
Reynolds Number
turbulent
10
00
00
0
10
00
00
00
10
00
00
10
10
00
0
24
Re
10
00
Cd 
1
0.1
0.
1
Drag Coefficient
1000
10
0
Vt 
d 2 g  p   w 
Re 
turbulent
boundary
Vt d 

Reynolds Number for a Floc
Find the diameter of the largest floc that
falls with laminar flow characteristics
Density of water 1000 kg/m3
Viscosity of water 0.001 kg/(m s)
1
Re 
Vt d 

fluid properties
4 gd   p   w 
Vt 
3 CD
w
45
What is the coefficient of drag? _____
For flocs CD=45/Re because flocs aren’t spheres



R

d
 4 g
 p  w  w


 3 CD






2/3
1.6
particle diameter (mm)
R
4 gd   p   w 
d w
3 CD
w
9
8
d
V
1.4
1.2
1
0.8
0.6
Turbulent
0.4
0.2
Laminar
0
1
1.05
1.1
1.15
1.2
7
6
5
4
3
2
1
0
1.25
terminal velocity (mm/s)
Laminar Flow Boundary
particle specific gravity
Flocs larger than ≈1 mm have turbulent flow characteristics
The flow field around most flocs (water treatment) is laminar
Settling zone
Inlet zone
 long rectangular
basins
 4-6 hour
retention time
 3-4 m deep
 max of 12 m
wide
 max of 48 m
long
Sludge out
Sludge zone
Outlet zone
Sedimentation Basin
Q
Vh 
A
flow rate
Inlet zone
Horizontal velocity
WH
Vertical velocity
Sludge out
Outlet zone
Sedimentation Basin:
Critical Path
Vh
Vc
Sludge zone
L
Vc = terminal velocity that just barely ______________
gets captured
What is Vc for this sedimentation tank?
Vc 
H

H
Sedimentation Basin:
Importance of Tank Surface Area


  residence time
Time in tank
  WHL  volume of tank
Q
Q
HQ
Q
Vc  


As


LW
H
W
A s  top surface area of tank
Vh
H
Vc
L
Want a _____
small Vc, ______
large As, _______
small H, _______
large .
Suppose water were flowing up through a sedimentation tank. What Q
would be the velocity of a particle that is just barely removed? Vc =
As
Lamella
 Sedimentation tanks are
commonly divided into
layers of shallow tanks
(lamella)
 The flow rate can be
increased while still
obtaining excellent
particle removal
Lamella decrease distance particle
has to fall in order to be removed
Outlet
zone
Inlet
zone
Design Criteria for
Sedimentation Tanks
Settling zone
Sludge zone
_______________________________
Minimal turbulence (inlet baffles)
_______________________________
Uniform velocity (small dimensions normal to velocity)
_______________________________
No scour of settled particles
_______________________________
Slow moving particle collection system
_______________________________
Q/As must be small (to capture small particles)
Sedimentation of Small
Particles?
How could we increase the sedimentation
rate of small particles?
Increase d (stick
particles together)
d g  p   w  Increase density difference
Increase g (centrifuge)
2
Vt 
18
(dissolved air flotation)
Decrease viscosity
(increase temperature)
Particle/particle interactions
 Electrostatic repulsion
 In most surface waters, colloidal surfaces are negatively
charged
stable suspension
 like charges repel __________________
 van der Waals force
 an attractive force
 decays more rapidly with distance than the electrostatic
force
 is a stronger force at very close distances
Electrostatic
Layer of
counter ions
+ +++++
++++
+ ++++ +
++
van der
Waals
Energy Barrier
Increase kinetic energy of
particles
increase temperature
stir
+ ++
+ +
+ ++
+
Decrease magnitude of energy
barrier
change the charge of the particles
introduce positively charged
particles
Coagulation
Coagulation is a physical-chemical process
whereby particles are destabilized
Several mechanisms
adsorption of cations onto negatively charged
particles
decrease the thickness of the layer of counter
ions
sweep coagulation
interparticle bridging
Coagulation Chemistry
The standard coagulant for water supply is
Alum [Al2(SO4)3*14.3H2O]
Typically 5 mg/L to 50 mg/L alum is used
The chemistry is complex with many
possible species formed such as AlOH+2,
Al(OH)2+, and Al7(OH)17+4
The primary reaction produces Al(OH)3
Al2(SO4)3 + 6H2O2Al(OH)3 + 6H+ + 3SO4-2
pH = -log[H+]
Coagulation Chemistry
Aluminum hydroxide [Al(OH)3] forms
amorphous, gelatinous flocs that are heavier
than water
The flocs look like snow in water
These flocs entrap particles as the flocs
settle (sweep coagulation)
Coagulant introduction with
rapid mixing
The coagulant must be mixed with the water
Retention times in the mixing zone are
typically between 1 and 10 seconds
Types of rapid mix units
pumps
hydraulic jumps
flow-through basins with many baffles
In-line blenders
Flocculation
Coagulation has destabilized the particles
by reducing the energy barrier
Now we want to get the particles to collide
We need relative motion between particles
Brownian motion is too slow (except for tiny
particles)
_________
Differential _____________
sedimentation rates
Turbulence shears the water
__________
Flocculation
Turbulence provided by
gentle stirring
Turbulence also keeps large
flocs from settling so they
can grow even larger!
High sedimentation rate of
large flocs results in many
collisions!
Retention time of 10 - 30
minutes
Flocculator Design (Prior to 1992):
The “shear is dominant” assumption
 Velocity gradient (G)
 P is power input to
 mechanical paddles
 Hydraulic (head loss)
Drag coefficient = 2 for flat
plate perpendicular to flow
Cd V A
P
2
P   ghl Q
 Recommended G and G values
 G – 20 to 100 /s
 G – 20,000 to 150,000
3
Cd V 2 A
Fd 
2
P
G
V

 Based on the (incorrect) assumption that the   Q
primary collision mechanism was fluid shear
These values were obtained empirically, so even though
the theory was wrong the values might be right!
Improved Model Development
 Transport mechanisms
 Diffusion
 Shear
 Differential Sedimentation
 Monodisperse vs Heterodisperse suspensions
 Rectilinear models ignored near field effects of
hydrodynamic and electrostatic repulsion and van
der Waals attraction
 Curvilinear models incorporated these near field
effects
Heterodisperse, Rectilinear
Flocculation
dnk
1
 rk 
  i, j  ni n j     i, k  ni nk

dt
2 i  j k
all i
Change in number
concentration of
size k particles
We double
counted the
formation of
these particles
Number
concentration of
size i particles
Collision frequency
[1/cm3]
between two
particles of sizes i
and j
[cm3/s]
Rectilinear Collision Frequency –
Transport Mechanisms
Brownian motion
2kT
 Br  i, j  
3
1 1 
    di  d j 
 di d j 
k is Boltzmann’s constant 1.38 x 10-16
g  cm 2
s2  K
3
1
 Sh  i, j    di  d j  G
6
Shear
G is the average velocity gradient 1/s
Differential Sedimentation
Assumes laminar flow
Vt 
d 2 g  p   w 
18
Add them all up
3
g
 DS  i, j  
 p  w  di  d j  di  d j

72
2
g
2
2 
 DS  i, j  
 p   w  di  d j
di  d j 


18
4
  i, j    Br  i, j    Sh  i, j    DS  i, j 
Differential Sedimentation
Rectilinear model
aj
Curvilinear model
Use trajectory
analysis to get
Critical path
ai
ai + aj
xc
Herterodisperse, Curvilinear
Flocculation
 Hydrodynamic interactions prevent collisions water between particles must move out of the way
 Van der Waals attractive force promotes collisions
- become significant at small separation distances
 Electrostatic repulsion prevents collisions –
diffuse layer of ions rich in those with charge
opposite to that of the surfaces is induced in the
fluid surrounding each particle
Dominant Collision Mechanisms
This model assumes
floc density is
independent of floc
size
100
10
dj (m)
Plot conditions
G = 10/s
T = 20°C
p = 1.1 g/cm3
1
Differential Sedimentation
Fluid Shear
Brownian Motion
0.1
0.1
1
10
di (m)
100
Curvilinear Simplified Conclusions
Shear is only important for particles within
a factor of 5 of the same size
Diffusion is important if the small particle is
less than 1 m and the large particle is less
than about 20 m
Differential sedimentation is important if
one of the particles is greater than 20 m
Application of Results
1
  i, j  ni n j

2 i  j k
 Increasing the concentration of large particles will
increase the collision rate
 Turbulence can be used to keep large particles in
suspension
 Need high fluid velocities at bottom of tank
 Could use grid or jet turbulence
 Recirculate large particles by providing upflow zone
 No need to try to optimize the fluid shear mechanism
since the differential sedimentation mechanism is
more efficient
Mechanical Flocculators
Mechanical flocculators are preferred in the
Global North
Speed of the mechanically operated paddles can
be varied (but do operators vary this?)
Disadvantages
Motors, speed controllers, gear boxes (to
reduce speed), and bearings may not be
maintainable
Require electricity
Hydraulic Flocculators
 Flocculation parameters are a function of flow and
thus cannot be adjusted independently
 Head loss is often significant
 Cleaning may be difficult, but appropriate designs
can accommodate cleaning
 Types
 Vertical flow
 Horizontal flow
 Tapered (to reduce shear as flocs grow larger?)
 Gravel bed flocculators (related to filters!)
Potential Research Project
I don’t know if anyone has designed a better
flocculator based on this new understanding
of the importance of keeping large particles
in suspension
But there is at least one existing design that
keeps particles in suspension
100
dj (m)
10
Sludge Blanket
Flocculator/Sedimentation Tank
1
Differential Sedimentation
Fluid Shear
Brownian Motion
0.1
Retention time 1-3 hours
Turbidity less than 900 NTU
Differential
sedimentation causes
large flocs to collide
with small Brownian
particles
0.1
1
10
100
di (m)
Critical velocity
Overflow channel
36 - 100 m/day
Water inlet
High velocities keep
particles moving up
Raw water with alum (Brownian
motion)
Water Outlet
Sludge blanket
Desludging valve
Flocculator Design
 Keep particles suspended
 requires tanks designed to resuspend particles that settle
 Keep shear levels low so that particles don’t break apart
 We need some data for appropriate shear levels
velocity gradient x viscosity
 Shear =_________________________
 Use the velocity gradient G that is recommended for
flocculator design and calculate the shear!
 The existing designs were based on the wrong theory, yet
they work.
 How can we reconcile this?
Basic Mechanism of Bed Load
Sediment Transport
 drag force exerted by fluid
flow on individual grains
 retarding force exerted by
the bed on grains at the
interface
 particle moves when
resultant passes through (or
above) point of support
V
h
force of drag will vary with time
Grains: usually we mean incoherent sands,
gravels, and silt, but also sometimes we
include cohesive soils (clays) that form larger
particles (aggregates)
Fd
Fg

point of support
Threshold of Movement
Force on particle due to gravity
4
Fg  g r 3
3
2
Force on particle due to shear stress Fshear   or
We expect movement when
 o  g
o
2
2d
tan 
3
 tan 
gd 3
 or 2
4
g r 3
3
 tan 
4

Fg  g r 3
3
important dimensionless parameter
Fshear   or 2
Shields Diagram (1936)
1
 Shields
Suspension
o

 gd
Saltation
0.1
0.056
Threshold of
movement
No movement
0.01
u*  gh
hf
L
1
10
Re d 
u*d 100

Laminar flow of bed
1000
Turbulent flow of bed
G and  and biggest particles
du
 0    G
dy
 Shields 
d
o
 gd
Assume floc density is 1100 kg/m3
How large were the flocs that are kept in
suspension given empirical design for G?
 Shields
G
 g  Shields
kg 
1

0.001
20
to
175



m  s 
s

d
kg 
m

100 3  9.8 2   0.05 
m 
s 

G

 gd
 = 0.001 kg/(m s)
 = 100 kg/m3
g = 9.8 m/s2
Shields = 0.05
d  0.41mm  3.6mm
Recommended G and G values:
Turbidity or Color Removal
Velocity
gradient (G)
(1/s)
Type
without
20-100
solids
recirculation
With solids 75-175
recirculation
G
20,000150,000
125,000200,000
0*
(Pa)

(s)
1000-1500
0.020-0.1
1100-1700
0.075-0.175
* Estimated from G assuming viscosity of 0.001 kg/(m s)
Suggested Design Process
residence time of about 30 minutes
Maximize vertical mixing to keep heavy
flocs in suspension
Keep shear levels less than 0.2 Pa* to avoid
floc breakup
Or perhaps 0.2 Pa is required to keep heavy
floc in suspension
*We need some research to see if this shear level is correct!
Vertical-flow Baffled Flocculator
u* 
0

 0   u*2
 0   fV
u*  V f
d
2
V
0.1
 D
f  Cp 
l 

0.05
0.04
0.03
0.02
0.015
friction factor
0
V
f
H
0.01
0.008
0.006
0.004

D
laminar
0.002
0.001
0.0008
0.0004
0.0002
0.0001
0.00005
0.01
1E+03
smooth
1E+04
1E+05
1E+06
Re
1E+07
1E+08
0.2 Pa
m
V
 0.1
kg 
s

1000
0.02

3 
m 

K180 bend= 2.5 - 4
hex
V2
K
2g
Scaling floc sed down to POU
 Need to reduce fluid velocities to avoid turning the
sedimentation tank into a CMFR (____________
Completely
___________________)
mixed flow reactor
 Batch may work best to ensure good
sedimentation
 Consider recycling flocs from previous batches
 Continuous flow could take advantage of big flocs
(upflow flocculation/sedimentation)
Coagulants
Inorganic
Aluminum Sulfate (alum)
Ferric chloride
Organic
Chitosan
Moringa oleifera
Dosage – 10 to 100 mg/L based on “jar”
tests
Jar Test
Mimics the rapid mix, coagulation,
flocculation, sedimentation treatment steps
in a beaker
Allows operator to test the effect of
different coagulant dosages or of different
coagulants
Suggests a batch technique for POU
turbidity removal
Upflow Flocculator/Sedimentation
Tank particle capture
What is the size of the smallest
floc that can be captured by this
tank with critical velocity of 100
m/day?
We need a measure of real water
treatment floc terminal velocities
Research…
Physical Characteristics of Floc:
The Floc Density Function
 Tambo, N. and Y. Watanabe (1979). "Physical
characteristics of flocs--I. The floc density
function and aluminum floc." Water Research
13(5): 409-419.
 Measured floc density based on sedimentation
velocity (Our real interest!)
 Flocs were prepared from kaolin clay and alum at
neutral pH
 Floc diameters were measured by projected area
Floc Density Function:
Dimensional Analysis!
 floc   w
 Floc density is a function of
w
__________
floc size
 Make the density dimensionless
d floc
 Make the floc size
d particle
dimensionless
 d floc 
  floc   w 
 Write the functional


  f 
w 
relationship

 d particle 
 After looking at the data
n
 d floc 





floc
w
conclude that a power law


  a 
w 
relationship is appropriate

 d particle 
d
Model Results
 d floc 
  floc   w 


  a 
w 

 d particle 
 For clay assume dparticle was 3.5 m (based on
Tambo and Watanabe)
 a is 10 and nd is -1.25 (obtained by fitting the
dimensionless model to their data)
 The coefficient of variation for predicted
dimensionless density is
 0.2 for dfloc/dparticle of 30 and
 0.7 for dfloc/dparticle of 1500
 The model is valid for __________flocs
in the size
clay/alum
range 0.1 mm to 3 mm
nd
Additional Model Limitation
 This model is simplistic and doesn’t include
 Density of clay
 Ratio of alum concentration to clay concentration
 Method of floc formation
 Data doesn’t justify a more sophisticated model
 Are big flocs formed from a few medium sized
flocs or directly from many clay particles?
 Flocs that are formed from smaller flocs may tend to be
less dense than flocs that are formed from accumulation
of (alum coated) clay particles
Model Results → Terminal Velocity
 d floc 
  floc   w 


  a 
w 

 d particle 
4 gd   p   w 
Vt 
3 CD
w
nd
45 Flocs aren’t spheres
Cd 
Re
Vt  d floc
4 gVt  w
3 45
 d floc 
a

d
 particle 
4 gt  w d floc  d floc 
Vt 
a


3 45
 d particle 
2
nd
nd
Re 
Cd 
Vt d floc  w

45
Vt d floc  w
Floc Density and Velocity
(Approximate)
  floc   w 



w


floc density
0.1
1000
floc density
Vt (m/day)
0.01
0.001
100
10
0.1
 floc ______
1030 kg/m3
0.4 mm
1
floc diameter (mm)
10
floc terminal velocity (m/day)
2
g

d
4 t w floc  d floc 
Vt 
a


3 45
 d particle 
nd
Flocculation/Sedimentation:
Deep vs. Shallow
 Compare the expected performance of shallow and deep
horizontal flow sedimentation tanks assuming they have
the same critical velocity (same Q and same surface area)
More opportunities to
______
collide with other
particles by _________
differential
____________
sedimentation or
Brownian motion
________________
deeper
Expect the _______
tank to perform better!
Flocculation/Sedimentation:
Batch vs. Upflow
Compare the expected performance of a batch
(bucket) and an upflow clarifier assuming they
have the same critical velocity
How could you improve the performance of
the batch flocculation/sedimentation tank?
Frictional Losses in Straight Pipes
0.1
 D
f  Cp 
l 

0.05
0.04
0.03
friction factor
0.02
0.015
0.01
0.008
0.006
0.004
laminar
0.002
0.001
0.0008
0.0004
0.0002
0.0001
0.00005
0.01
1E+03
smooth
1E+04
1E+05
1E+06
Re
1E+07
1E+08

D