Performance of commercial chevron mist eliminators


McNulty, K.J.; Monat, J.P.; Hansen, O.V.

Chemical Engineering Progress 83(5): 48-55

1987


Test methods have been developed for evaluating the performance of chevrons with respect to capacity, pressure drop, removal efficiency, and plugging tendency. The first three of these parameters are presented for 18 different commercial vertical-flow chevrons. For the chevrons tested and the range of test conditions studied, the capacities ranged over a factor of 2, the pressure drops ranged over a factor of 10, and the droplet penetration ratio ranged over a factor of 20 or more. All of the chevrons tested gave essentially complete removal of droplets larger than 70 Am in diameter, and some gave essentially complete removal of 10 Am droplets. These data can be used, along with other selection criteria, to compare various chevron designs for retrofit or new installations.

SPECIAL
REPORT
Performance
of
Commercial
Chevron
Mist
Eliminators
Test
methods
developed
for
the
evaluation
of
chevrons
can
predict
accurately
the
merits
of
various
commercial
chevron
designs.
Kenneth
J.
McNulty,
Jamie
P.
Monat,
Koch
Engineering
Co.,
Inc.,
Wilmington,
MA
01887
Ole
V.
Hansen,
Koch
Engineering
Co.,
Inc.,
Wichita,
KS
67208
U
nit
operations
that
involve
gas-liquid
contact
com-
prise
an
important
part
of
chemical
engineering.
With
such
unit
operations,
it
is
often
desirable
and
sometimes
essential
to
separate
the
gas
and
liquid
phases
following
contact.
For
the
case
where
the
gas
is
the
contin-
uous
phase
and
the
liquid
is
present
as
discrete
droplets
dispersed
throughout
the
gas,
phase
separation
can
be
made
by
one
of
a
variety
of
mist
eliminators.
K.
J.
McNulty,
technical
director
of
Koch
Engineering
R&D,
is
pri-
marily
responsible
for
technical
guidance
of
product
R&D
on
tower
packings,
trays,
mist
eliminators,
and
static
mixers.
Areas
of
partic-
ular
interest
include
gas-liquid
mass
transfer,
gas
absorption
with
chemical
reaction,
fluid
mechanics,
and
droplet
dynamics.
He
re-
ceived
his
degrees
from
Penn
State
(B.S.)
and
MIT
(Sc.D.)
in
chemical
engineering.
J.
P. Monat
was
manager
of
Koch
Engineering
R&D
when
this
work
was
done,
and
is
now
associate
director
of
R&D
at
Koch
Membrane
Systems,
Inc.
The
author
of
articles
on
flue
gas
desulfurization,
air
pollution
control,
energy
conservation,
and
laser-based
pollutant
monitoring,
his
current
interests
include
mist
elimination
from
gas
streams
and
membrane-based
liquid
filtration
systems.
Dr.
Monat
earned
his
B.S.E.
degree
in
aerospace
and
mechanical
sciences
at
Princeton
Univ.
and
his
M.S.
and
Ph.D.
degrees
in
environmental
en-
gineering
at
Stanford
Univ.
0.
V.
Hansen,
manager
of
the
Mist
Elimination
Group
of
Koch
Engi-
neering
Company,
Inc.,
has
been
responsible
for
that
area
since
join-
ing
Koch
in
1978.
Before
that
he
was
technical
marketing
director
of
Beltran
Associates,
manager
of
marketing
of
Monsanto
Enviro-Chem
Systems,
and
a
chemical
engineer
with
Fisher
Body
Division
and
3M
Co.
He
has
a
B.S.Ch.E.
degree
from
Wayne
Univ.,
a
B.S.
(equivalent)
in
Meteorology
from
Texas
A
&
M,
and
is
a
registered
Professional
Engineer
in
Michigan.
48
Different
types
of
mist
eliminators
are
effective
for
cap-
turing
different
ranges
of
droplet
size
(1).
In
this
article.
at-
tention
will
be
focused
on
chevron
or
baffle-type
mist
elim-
inators
in
which
the
gas
flow
is
in
the
upward
direction.
In
this
configuration,
mist
collected
by
inertial
impaction
on
the
baffles
must
drain
downward
in
the
form
of
large
drops
or
streams
countercurrent
to
the
rising
gas.
Of
the
three
common
types
of
mist
eliminators
(fiber
bed.
wire
mesh,
and
chevron),
the
chevron
type
is
limited
to
the
removal
of
the
largest
droplets
(>10
ktm).
Its
advantages
over
other
types
of
mist
eliminators
include:
low
pressure
drop,
high
capacity,
and
ability
to
handle
"dirty"
streams
that
would
quickly
plug
other
mist
eliminators.
In
a
pre-
vious
paper
(/),
test
methods
were
described
for
the
mea-
surement
of
chevron
performance,
including
the
use
of
a
la-
ser
droplet
sizing
interferometer
to
measure
droplet
re-
moval
efficiency
as
a
function
of
droplet
size.
These
test
methods
have
been
used
to
evaluate
the
performance
of
18
different
commercial
chevrons.
These
data
can
be
used
to
predict
chevron
performance
for
new
designs,
retrofits,
or
existing
installations
and
to
compare
the
relative
merits
of
various
commercial
chevron
designs.
Performance
parameters
The
important
performance
parameters
for
chevron
mist
eliminators
are
capacity,
pressure
drop,
droplet
removal
ef-
ficiency,
and
plugging
tendency.
These
parameters
are
all
interrelated
and
should
all
be
considered
together
when
Chemical
Engineering
Pmgress
Table
1.
Various
commercial
chevrons
tested.
Chevron
Number
Description
1.
4-pass
sinusoid
with
hooks,
1
1
/
4
in.
spacing,
plastic
2.
3-pass
modified
zig-zag,
1
in.
spacing,
45°
angle
to
flow,*
plastic
3.
4-pass
sinusoid
with
hooks,
1
3
/
a
in.
spacing,
plastic
4.
4-pass
zig-zag
with
hooks,
2
in.
spacing,
45°
angle
to
flow,
stainless
steel
(SS)
5.
4-pass
zig-zag,
no
hooks,
2
in.
spacing,
45°
angle
to
flow,
SS
6.
Corrugated
sheet
metal
packing,
1
/
2
in.
corrugation
height,
12
in.
thick,
45°
angle
to
flow*
7.
Corrugated
sheet
metal
packing,
1
in.
corrugation
ht.,
12
in.
thick,
30°
angle
to
flow
8.
Corrugated
sheet
metal
packing,
1
in.
corrugation
ht.,
12
in.
thick,
45°
angle
to
flow
9.
Corrugated
sheet
metal
packing,
'/2
in.
corrugation
ht.,
6
in.
thick,
45°
angle
to
flow
10.
3-pass
modified
zig-zag,
1.5
in.
spacing,
45°
angle
to
flow,
plastic
11.
3-pass
modified
zig-zag,
1
in.
spacing,
45°
angle
to
flow,
SS
12.
2-pass
modified
zig-zag,
0.75
in.
spacing,
SS
13.
3-pass
zig-zag
with
hooks,
2
in.
spacing,
45°
angle
to
flow,
SS
14.
3-pass
zig-zag,
no
hooks,
2
in.
spacing,
45°
angle
to
flow,
plastic
15.
2-pass
separated
zig-zag,
1
3
/
8
in.
spacing,
30°
angle
to
flow,
SS
16.
2-pass
modified
zig-zag,
0.75
in.
spacing,
plastic
17.
2-pass
wing-shaped
blade,
3
3
/
4
in.
spacing,
plastic
18.
3-pass
zig-zag,
no
hooks,
2
in.
spacing,
45°
angle
to
flow,
SS
"'Angle
to
flow"
signifies
the
angle
between the
principal
gas
flow
direction
and
the
chevron
blade.
For
corrugated
sheet
metal,
it
Is
the
angle
between
the
ridges
(or
grooves)
and
the
principal
flow
direction.
SI
Conversion:
mm
=
in.
x
25.4
comparing
the
performance
of
alternative
chevrons.
We
have
obtained
data
on
capacity,
pressure
drop,
and
removal
efficiency;
however,
tests
comparing
the
plugging
tendency
of
various
commercial
chevrons
have
not
been
routinely
conducted.
Therefore,
if
plugging
is
anticipated
to
be
a
problem,
it
will
be
necessary
to
obtain
plugging
data
on
the
chevrons
of
interest
or
to
make
some
reasonable
estimate
of
the
relative
plugging
tendency
of
alternative
chevrons
based
on
their
geometry.
Chevrons
with
narrow
blade
spacing
[e.g.,
less
than
about
1.5
in.
(38
mm)]
or
tortuous
flow
paths
would
be
expected
to
be
more
prone
to
plugging.
The
18
commercial
chevrons
tested,
Table
1,
are
de-
scribed
in
generic
terms
defining
the
number
of
"passes,"
blade
shape,
blade
spacing,
material
of
construction,
and
geometric
details.
Figure
1
illustrates
the
various
general
shapes
of
chevrons
listed
in
Table
1.
For
chevrons
with
rel-
atively
simple
geometry,
the
number
of
passes
is
the
num-
ber
of
straight
blade
sections
in
the
equivalent
zig-zag
chev-
ron.
All
of
the
shapes
shown
in
Figure
1
(except
for
the
cor-
rugated
sheet
metal)
may
be
considered
to
have
two
passes
since
they
can
all
be
approximated
by
a
simple
zig-zag
baf-
fle
having
two
straight
sections.
The
spacing
is
the
center-
to-center
distance
between
adjacent
blades
or
baffles
as
measured
in
the
horizontal
plane.
Capacity
The
capacity
(maximum
gas
velocity)
of
a
chevron
is
lim-
ited
by
the
phenomenon
of
reentrainment.
From
a
design
standpoint,
it
is
convenient
to
rate
both
capacity
and
pres-
sure
drop
performance
in
terms
of
F,
the
superficial
flow
factor,
defined
by:
F.
=
Nri
;•,,
(1)
where
U
5
is
the
superficial
gas
velocity
approaching
the
Table
2.
Hydraulic
performance
of
chevrons.
Chevron
Number
Inlet
Loading
=
0.3
gpm/ft
2
Inlet
Loading
gpm/ft`
=
2.5
R'
*
Euler
Number'
(F,),*
AP,
in.
11,0
R*
*
(F;),*
AP,
in.
H,0
1
3.93
1.10
1.21
3.60
1.10
1.43
11.7
2
3.29
0.43
1.08
3.09
0.43
1.23
5.89
3
2.97
0.40
1.0
2.77
0.40
1.21
9.63
4
3.10
1.10
1.09
3.00
1.10
1.14
32.6
5
3.00
0.57
1.0
2.80
0.57
1.0
17.8
6
4.10
1.40
1.0
3.60
1.40
1.28
13.1
7
4.30
1.50
1.18
3.91
1.50
1.51
5.35
8
3.70 0.53
1.05
3.35
0.53
1.11
7.12
9
3.20
0.70
1.07
2.80
0.70
1.29
8.51
10
3.39
0.27
1.05
3.15
0.27
1.41
4.47
11
4.00
0.20
1.17
3.70
0.20
1.47
2.89
12
4.40
0.51
1.11
3.90
0.51
1.26
3.75
13
3.20
0.94
1.05
3.05
0.94
1.08
28.4
14
3.00
0.57
1.04
2.50
0.39
1.07 17.7
15
3.60 0.23
1.0
3.30
0.23
1.0
5.41
16
3.80
1.00
1.0
3.50
1.00
1.09
8.56
17
2.30
0.40
1.18
1.80
0.30
1.33
20.5
18
3.20
0.51
1.0
2.85
0.51
1.05
15.3
*Critical
F,
above
which
reentrainment
occurs
in
ft/s
(Ibm/ft')
'•
R
is
the
ratio
of
wet
to
dry
pressure
drop
below
the
F,
at
which
loading
begins.
'Eu
=
2APg,/p
u
Ll:
where
AP
is
in
lbf/ft'
rather
than
In.
H,0.
SI
Conversion:
m'/m'
h
=
gpm/ft
2
x
2.445;
m/s
=
ft/s
(Ibm/ft
2
)
x
1.113;
N/m
1
=
in.
H
2
0
x
249.
May
1987
49
IL
500
-
400
-
300
-
200
-150
100
90
80
AP
70
PRESSURE
80
(N/m
2
)
DROP
-
50
15
19
9
10
2
I I
-
2
;
F
s
1((tisec)
Otarn/11
3
1
1/2
i
15
I
I
1
f
5
8
7
8
9
10
S
ZIGUG
I
ZIGZAG
NUSOID
PEAK
VALLEY
WING
SHAPED
ADJACENT
SHEETS
IN
CONTACT
AT
PEAKS
8
VALLEYS
CORRUGATED
SHEET
METAL
//
SEPARATED
ZIGZAG
Figure
1.
Shapes
of
commercial
chevrons.
chevron.
Using
F,
rather
than
velocity,
permits
the
same
ca-
pacity
and
pressure
drop
data
to
be
used
independent
of
gas
density.
[The
units
of
F,
used
are
ft/s
(Ibm/ft
.3
)
1 /2
.
Metric
equivalents
are
given
for
velocity
in
m/s
for
air
at
ambient
conditions.]
The
critical
F,
i.e.,
the
F„
at
which
reentrainment
just
be-
u
g
(m/s)
al
20°C
and
1.0
atm
2
1
i
5
2
3
4
5
8
7
6
9
10
1
1 1
1
1
15
18
6
AP
PRESSURE
DROP
(Inches
H2O)
-40
-
30
13
-20
17
145
18
gins
to
occur,
is
a
function
of
the
liquid
loading.
Values
of
the
critical
F,
at
liquid
loadings
of
0.3
and
2.5
gpm(gal/min)/
ft
2
(0.73
and
6.1
m
3
/m
2
h)
are
listed
in
Table
2
for
18
commercial
chevrons.
These
values
were
obtained
with
a
reasonably
well
developed,
symmetric
velocity
profile
at
the
inlet
of
the
chevron.
For
nonuniform
velocity
profiles,
the
chevron
must
be
overdesigned
to
prevent
local
reentrain-
ment
from
areas
of
the
chevron
where
the
gas
velocity
would
be
higher
than
the centerline
velocity
for
well-developed
flow.
Also
given
in
Table
2
is
the
critical
pressure
drop
for
each
chevron.
This
is
the
pressure
drop
observed
at
the
critical
F,
where
incipient
reentrainment
occurs.
For
most
chev-
rons,
the
critical
pressure
drop
is
independent
of
liquid
loading,
at
least
for
clean
conditions.
It
therefore
provides
a
very
simple
means
of
determining
whether
a
chevron
in
a
particular
field
installation
is
being
operated
above
the
reentrainment
point.
This
is
particularly
useful
since
it
does
not
require
a
knowledge
of
either
the
gas
flow
rate
or
the
liquid
loading.
The
capacity
of
these
chevrons
may
be
compared
to
the
capacity
of
other
mist
elimination
devices.
A
variety
of
mesh
pad
mist
eliminators
have
been
tested
using
the
same
equipment
and
procedures
as
for
the
chevron
test
data.
A
typical,
general-purpose
stainless
steel
wire
mesh
pad
hav-
ing
a
thickness
of
6
in.
(150
mm),
a
density
of
9
lbm/ft"
(144
kg/m"),
and
a
wire
diameter
of
0.011
in.
(0.28
mm)
gives
a
critical
F,
of
2.7
(3.0
m/s)
at
the
2.5
gpm/fe
(6.1
m"Irn
2
h)
loading.
This
is
only
about
16%
lower
than
the
mean
value
for
the
chevrons
of
Table
2
at
corresponding
conditions.
However,
tests
conducted
with
a
supposedly
equivalent
plas-
tic
pad
consisting
of
an
interlaced
structure
of
plastic
mon-
ofilaments
arranged
transverse
to
the
gas
flow
indicated
a
critical
F,
of
only
1.75
(1.9
m/s).
Chevron-type
mist
elimi-
nators
can
be
used
in
combination
with
knitted-wire
mesh-
type
mist
eliminators
to
improve
the
overall
collection
effi-
ciency
on
smaller
particles.
In
fouling
service,
the
chevron
will
also
reduce
the
required
maintenance
and/or
replace-
ment
frequency
compared
to
the
use
of
a
mesh
pad
alone.
Pressure
drop
The
dry
pressure
drops
of
the
chevrons
are
shown
as
a
function
of
F,
in
Figure
2.
At
a
given
flow
rate
(constant
F,).
the
pressure
drops
can
differ
by
an
order
of
magnitude.
Under
typical
conditions,
the
flow
through
a
chevron
is
quite
turbulent.
The
pressure
drop
arises
primarily
from
the
loss
of
kinetic
energy
as
the
gas
changes
direction
within
the
chevron.
The
equivalence
between
pressure
energy
and
kinetic
energy
can
be
derived
from
the
Bernoulli
equation
and
is
given
by:
2
15
10
-
06
-
05
04
XP
N
=
P
4
2
g,
(2)
03
1
I
1
1
This
equation
forms
the
basis
of
the
commonly
used
veloc-
ity
head
concept
(2)
for
estimating
the
pressure
drop
of
a
tortuous
flow
path.
Figure
2.
Dry
pressure
drop
for
vertical-flow
chevrons.
50
Chemical
Engineering
Progress
2
25
3
4
5
6
7
8
9
10
F
s
Itlisec
(Ibm/11
3
)
112
Figure
3.
Pressure
drop
vs.
F,
for
Chevron
No.
10.
Equations
1
and
2
indicate
that
the
pressure
drop
of
a
chevron
should
be
proportional
to
the
square
of
F,.
Pres-
sure
drop
curves
on
log-log
coordinates
should
therefore
have
a
slope
of
about
2;
observed
slopes
are
quite
close
to
this
value.
The
pressure
drop
performance
of
the
various
chevrons
can
be
quantitatively
represented
by
the
Euler
number:
AP
2g,
2g„AP
Eu
p
y
(3)
The
Euler
number
gives
the
number
of
velocity
heads
of
pressure
loss
for
the
chevron
based
on
the
superficial
gas
velocity.
The
lower
the
Euler
number,
the
more
energy-effi-
cient
the
chevron.
Euler
numbers
calculated
at
an
F,
of
2.5
ft/s
(Ibm/ft
3
)
1 /2
(2.8
m/s)
are
given
in
the
last
column
of
Table
2.
While
the
dry
pressure
drops
shown
in
Figure
2
provide
a
good
indication
of
the
relative
flow
resistance
of
the
chev-
rons,
the
actual
pressure
drop
during
operation
can
be
sig-
nificantly
higher
than
the
dry
pressure
drop,
particularly
if
the
chevron
is
operated
close
to
the
critical
P.
Pressure
drop
curves
(AP
vs.
F,)
have
been
generated
for
all
the
chevrons
in
Table
1
at
liquid
loadings
of
0.3
and
2.5
gpml
ft
2
(0.73
and
6.1
m
3
/m
2
h).
A
pressure
drop
model
de-
scribed
below
permits
a
reasonably
accurate
estimate
of
the
wet
pressure
drop
for
most
chevrons.
U
g
(m/s
al
20°C
and
1
atm)
2
3
4
5
6
7
8
9
10
200
150
100
90
60
70
60
AP
PRESSURE
DROP
so
(11/m
2
)
40
30
26
10
09
08
07
0
0
gprn/11
7
0.9
gprnit1
3
(0
73
61
3
(m
2
0)
A
z
5
gprn/11
3
(6
1
6
0
(6
2
01
AP
0
AP
025
PRESSURE
DROP
02
fin
H20)
075
010
009
0
08
007
00
04
06
05
0
06
005
0
04
/
2
5
gpm/11
2
3.0
gpm/6
2
10
1
.
3,62
1
6
10
73
m
3
/
6
1
2
61
I
I
I
I
10
r
Performance
data
presented
here
for
18
commercial
chevrons
can
be
used
to
predict
chevron
performance
for
new
designs,
retrofits,
or
existing
installations.
Figure
3
shows
a
typical
set
of
pressure
drop
curves
for
loadings
of
0,
0.3,
and
2.5
gpm/ft
2
(0,
0.73,
and
6.1
m"Im
2
h)
for
Chevron
No.
10.
Below
an
P
of
2.5,
the
wet
pressure
drop
curves
lie
parallel
to
the
dry
line
and,
at
the
lower
liq-
uid
loading,
may
coincide
with
it.
Above
an
F,
of
2.5,
the
drag
of
the
gas
inhibits
the
drainage
of
the
liquid,
and
ad-
ditional
liquid
is
held
up
in
the
chevron
resulting
in
a
higher
pressure
drop.
This
condition
may
be
called
"loading"
and
is
similar
to
loading
in
a
packed
tower
(3).
As
F,
is
further
increased,
the
critical
F,
is
reached
at
the
critical
AP
where
reentrainment
begins.
While
the
loading
point
varies
some-
what
from
chevron
to
chevron,
a
value
of
F,
at
loading
of
2.5
(2.8
m/s)
is
typical.
Below
the
loading
point,
there
is
a
constant
ratio
between
the
wet
and
dry
pressure
drops,
which
is
related
to
the
liq-
uid
holdup
in
the
chevron.
If
it
is
assumed
that
the
Euler
number
does
not
change
when
the
blades
get
wet,
the
effect
of
liquid
holdup
in
the
chevron
will
be
to
increase
the
ac-
tual
gas
velocity
through
the
chevron.
If
it
is
also
assumed
that
a
uniform
liquid
film
of
thickness
b
exists
on
two
adja-
cent
baffles
separated
by
a
distance
b.
the
ratio
of
actual
velocities
in
the
wet
and
dry
condition
will
be
approxi-
mately
(1_1
6
)„„
b
(11)„.
b
26
(4)
Here
it
is
assumed
that
the
length
of
the
cross-sectional
area
for
gas
flow
between
the
blades
is
much
greater
than
its
width
(i.e.,
than
the
blade
separation).
Equations
3
and
4
give
the
following
relationship
for
wet
pressure
drop
below
loading
AP„.
=
Eu
2g,
The
average
film
thickness
in
Eq.
5
is
undoubtedly
a
func-
tion
of
liquid
physical
properties
(4)
and
will
change
from
system
to
system.
However,
for
present
purposes,
the
brack-
(5)
May
1987
51
ea
REMOVAL
EFFICIENCY
(%)
AO
DROPLET
PENETRATION
RATIO
10
30
60
DROPLET
DIAMETER
loMI
Figure
4.
Droplet
removal
efficiency
at
5.3
ft/s
(1.6
m/s).
100
00
/13,18
eco
REMOVAL
EFFICIENCY
I%)
XI
DROPLET
PENETRATION
RATIO
tO
20
10
DROPLET
DIAMETER
(A41
Figure
5.
Droplet
removal
efficiency
at
10.3
ft/s
(3.1
m/s).
I
I
I
ea
REMOVAL
EFFICIENCY
I%I
DROPLET
PENETRATION
RATIO
20
30
40
00
DROPLET
DIAMETER
loM/
Figure
6.
Droplet
removal
efficiency
at
12.0
ft/s
(3.6
m/s).
AP,
2g,
Eu
(2.5)
2
R
(7)
In
arriving
at
an
optimum
design,
it
is
often
necessary
to
make
a
compromise
between
removal
efficiency
on
the
one
hand
and
pressure
drop
and
plugging
tendency
on
the
other.
eted
term
can
be
replaced
by
R,
the
pressure
drop
ratio
for
wet
to
dry
pressure
drop
below
the
loading
point.
Then
the
equation
for
wet
pressure
drop
below
the
loading
point
be-
comes:
Eu
FR
2g,.
(6)
For
F,
<
2.5
Values
of
R
measured
at
0.3
and
2.5
gpm/ft
2
(0.73
and
6.1
m
3
/m
2
/h)
are
listed
in
Table
2
for
the
chevrons
tested.
Above
the
loading
point,
pressure
drops
can
be
deter-
mined
by
interpolation
between
the
pressure
drop
at
load-
ing
and
the
pressure
drop
at
reentrainment.
Using
linear
in-
terpolation,
Eu
(2.5)
2
R
F
,
2.5
AP„
=
+
2g,
(
F
s
2.5
For
F,
>
2.5
This
equation
generally
gives
somewhat
conservative
re-
sults
(higher
pressure
drops
than
measured).
Other
methods
of
interpolation
could
be
used
for
greater
accuracy
but
are
more
complicated.
For
the
data
in
Figure
3
at
a
liquid
load-
ing
of
2.5
gpm/ft
2
(6.1
m
3
/m
2
h),
Eq.
7
predicts
pressure
drops
ranging
from
—2%
to
16%
higher
than
those
mea-
sured.
Removal
efficiency
Removal
efficiency
data
for
commercially
available
verti-
cal-flow
chevrons
are
presented
for
different
superficial
gas
velocities
in
Figures
4-6.
Also
shown
is
the
droplet
penetra-
tion
ratio
(DPR)
which
is
one
minus
the
fractional
removal.
The
DPR
is
the
ratio
of
the
flux
or
concentration
of
drop-
lets
at
the
outlet
to
that
at
the
inlet
for
a
particular
droplet
size.
All
the
curves
display
generally
similar
shapes.
Even
at
the
lowest
velocity,
all
chevrons
are
essentially
100%
effec-
tive
in
removing
dropets
that
are
larger
than
70
Am
in
di-
ameter.
At
the
highest
velocity,
several
chevrons
exhibited
essentially
100%
removal
of
10
Am
droplets.
This
is
re-
markable
since
it
is
generally
believed
that
only
knitted-wire
mesh
pads
can
be
used
effectively
for
droplets
below
10
Am.
The
droplet
penetration
ratios
can
vary
drastically.
For
example,
Figure
4
indicates
a
20-fold
range
in
DPR
for
15
ttm
droplets,
and
the
ranges
shown
in
Figures
5
and
6
are
even
greater.
In
general,
trends
predicted
by
theory
(1)
are
observed:
removal
efficiency
increases
with
gas
velocity,
droplet
size,
and
angle
of
inclination
and
decreases
with
52
Chemical
Engineering
Progress
Table
3.
Typical
mist-size
distributions.
Mist
1%
by
wt.
Smaller
Than
(pm)
10%
by
wt.
Smaller
Than
(pm)
50%
by
wt.
Smaller
Than
(pm)
90%
by
wt.
Smaller
Than
(pm)
99%
by
wt.
Smaller
Than
(i.im)
fir
s
o
Mist/Acid
Plants:
ping
Tower
Exhaust
p
r
imary
Absorbing
Tower
98%
Acid
Production
0.4
0.1
0.8
0.8
1.7
10.0
10.0
Oleum
Production
s
e
tondary
Absorbing
0.2
0.5
0.8
2.5
Tower
0.5
1.6
2.5
5.0
Ammonia
Scrubber
0.3
0.4
0.7
2.0
25
s
u
i
f
oric
Acid
Plants
0.3
26
(General)
ph
o
gphoric
Acid
Mist/Acid
0.5
5
Plant
up.plow
Cooling
Tower
200
300
400
500
600
paded
Cross-Flow
Tower
150
200
500
800
1,100
Venturi
Scrubber
40
100
175
300
500
Reverse-Jet
Scrubber
100
250
500
1,250
2,000
Evaporator
(25
in.
disengaging
space)
20
50
130
240
300
Sieve-Tray
Tower
(5.25
in.
disengaging
space)
250
600
1,100
1,800
2,500
Cooler-Condenser
0.1
5
10
20
35
2-Fluid
Nozzle
(Atomizing
1
15
35
90
120
Air
Pressure
less
than
80
psig)
SWle-Fluid
Nozzle
60
200
500
1,500
2,000
(P
=
100
psig)
Reference
D1005,
D.R.,
and
E.D.
Kennedy,
Chem.
Eng.
Prog.,
p.
70
(Sept.,
1978).
Perry,
R.H.,
"Chemical
Engineers'
Handbook,"
5th
Ed.,
McGraw-Hill,
pp.
18.60,
(1973).
Stern,
A.C.,
"Air
Pollution,"
3rd
Ed.,
Vols.
I
and
IV,
Academic
Press,
iv.
80,
295
(1976,
1977).
Atkinson,
D.S.F.,
and
W.
Strauss,
J.
A.P.C.A.,
28
(11),
p.
1114
(Nov.,
1978).
G.K.,
and
T.F.
Holmes,
U.S.
Patent
#4,374,813,
"Reverse
let
Scrubber
Apparatus
and
Method"
(Feb.
22,
1983).
Boll,
R.H.,
et
al.,
J.
A.P.C.A.,
24
(10),
p.
934
(Oct.,
1974).
SI
conversion:
mm
=
in.
x
25.4;
kPa
=
psi
x
6.89
blade
spacing.
Where
the
velocities
of
Figures
5
and
6
are
above
the
reentrainment
velocity
for
a
particular
chevron,
it
is
impossible
to
obtain
meaningful
removal
efficiency
data.
Which
commercial
chevron?
The
above
data
can
be
used
to
evaluate
and
select
com-
mercially
available
chevrons
for
retrofit
or
new
installa-
tions.
In
arriving
at
an
optimum
design,
it
is
often
neces-
sary
to
make
a
compromise
between
removal
efficiency
on
the
one
hand
and
pressure
drop
and
plugging
tendency
on
the
other.
To
do
so,
it
is
necessary
to
have
some
knowledge
of
the
droplet-size
distribution
entering
the
chevron
(or,
at
least,
to
know
which
chevron
has
given
acceptable
removal
in
the
past
so
that
the
relative
performance
of
alternatives
can
be
weighed).
Without
some
knowledge
of
the
inlet
drop-
let-size
distribution,
it
is
impossible
to
accurately
design
or
specify
a
chevron
to
meet
an
overall
liquid
removal
crite-
rion.
Unfortunately,
inlet
droplet-size
distributions
are
sel-
dom
known
accurately.
Table
3
gives
some
droplet-size
dis-
tributions
for
some
of
process
equipment
as
reported
in
the
literature.
These
data
should
be
considered
only
as
a
rough
indication
of
what
the
size
distribution
might
be.
It
is,
for
example,
apparent
from
this
table
that
chevrons
would
not
he
applicable
to
the
removal
of
acid-plant
mists.
Where
data
or
correlations
are
available,
it
is
sometimes
possible
to
calculate
a
droplet-size
distribution
for
the
par-
ticular
operating
conditions
and
geometric
configuration
of
the
process
equipment.
Nozzle
manufacturers
frequently
May
1987
53
60
50
d95
0
,
M)
40
CHEVRON
NO
15
CHEVRON
NO
10
CHEVRON
NO
14
30
20
10
0
characterize
the
droplet-size
distribution
produced
by
their
nozzles
at
various
conditions.
If
such
information
is
not
available,
the
correlation
of
Mugele
(5)
using
the
"upper
limit"
distribution
can
be
used
to
predict
droplet-size
distri-
butions
for
various
atomizing
devices.
Even
so,
the
chevron
may
not
receive
the
direct
output
of
the
spray
generating
device
and
so
will
"see"
a
different
distribution.
The
best
way
to
determine
inlet
droplet-size
distributions
is
to
measure
them
directly.
This
is
particularly
applicable
for
retrofits
where
the
operating
process
equipment
pre-
sumably
exists
and
measurements
can
be
made.
Although
various
techniques
are
available
(1)
for
measuring
droplets
in
the
size
range
for
chevrons
(>10
pcm),
most
are
expen-
sive
and
difficult
to
use
in
the
field.
One
simple
technique
that
has
been
successfully
used
in
our
tests
and
that
seems
to
give
good
agreement
with
measurements
made
with
the
laser
droplet
sizing
interferometer
is
the
impaction
tech-
nique
of
May
(6).
In
this
technique,
a
magnesium
ribbon
is
burned,
and
a
microscope
slide
is
coated
with
a
layer
of
magnesium
oxide
smoke.
When
this
slide
is
exposed
for
a
short
time
to
the
mist-laden
stream
the
droplets
form
im-
paction
craters
in
the
MgO
layer.
These
craters
are
then
viewed,
sized,
and
counted
with
a
light
microscope
to
get
the
droplet-size
distribution.
While
some
trial-and-error
is
involved
in
getting
the
proper
crater
density
(which
de-
pends
on
mist
loading
and
exposure
time),
this
technique
can
be
used
for
field
measurements.
Other
selection
criteria
for
a
particular
application
that
must
be
carefully
considered
along
with
pressure
drop,
ca-
pacity,
and
removal
efficiency
include:
Potential
for
plugging
or
fouling
of
the
chevron
Cleanabi
I
ity
Ruggedness
toward
system
upsets
or
frequent
manual
cleaning
Resistance
to
corrosion
or
chemical
attack
Uniformity
of
gas
velocity
profiles
Turndown
requirements
and
performance
Physical
properties
if
significantly
different
from
am-
bient
air-water
Inlet
droplet-size
distribution
and
liquid
loading
Experience
and
support
capabilities
of
vendor
Initial
cost
and
delivery
Ease
of
installation
and
removal
3
4
5
6
10
11
12
Ug
(FT/SEC)
Figure
7.
Droplet
removal
vs.
superficial
gas
velocity.
Design
examples
Design
Example
1:
Retrofit
for
Increased
Capacity
Prob-
lem.
A
wet
FGD
scrubber
is
operating
successfully
with
Chevron
No.
14
at
an
F,
of
2.25
ft/s
(lbm/ftl
1 /2
(2.50
m/s)
and
continuous
spray
washing
from
below
at
2.5
gpm/ft'
(6.1
m'/
m
2
h).
The
gas
flow
through
the
system
must
be
increased
by
25%
with
no
increase
in
pressure
drop
across
the
chev-
ron
and
with
approximately
the
same
removal
efficiency.
Which
chevron
should
be
used?
Solution.
Table
2
indicates
that
the
critical
F,
for
the
ex-
isting
chevron
is
2.5.
For
the
same
relative
approach
to
the
point
of
reentrainment,
the
new
chevron
should
have
a
crit-
ical
F,
which
is
higher
by
25%
or
3.12
ft/s
(1b/ft)''
(3.47
m/
s).
From
Table
2,
Chevrons
1,
6,
7,
8,
10,
11,
12,
15,
and
16
qualify
on
the
basis
of
being
able
to
handle
the
higher
gas
flow.
Since
FGD
is
a
highly
fouling
service,
chevrons
with
blade
spacings
less
than
1
1
/
4
in.
(32
mm)
or
with
inac-
cessible
internal
flow
passages
(corrugated
sheet
metal)
can
be
eliminated.
This
reduces
the
candidates
to
Chevrons
1,
10,
and
15.
Chevron
No.
1
can
be
eliminated
on
pressure
drop.
The
data
indicate
that
the
pressure
drop
for
Chevron
No.
1
at
the
new
flow
will
be
almost
75%
greater
than
that
of
Chev-
ron
No.
14
at
the
original
flow.
Furthermore,
Chevron
No.
1
would
likely
be
the
most
difficult
to
keep
clean
since
it
has
the
narrowest
blade
spacing
and
the
greatest
number
of
Ug
(m/s)
0
0.5
,o
1.5
2
0
25
3.0
35
54
Chemical
Engineering
Progress
All
of
the
chevrons
tested
gave
complete
removal
of
droplets
larger
t
han
70
,um
in
diameter,
and
some
gave
complete
removal
of
10
,um
dfoplets.
pa
yses.
Both
Chevrons
10
and
15
give
a
lower
pressure
drop
at
the
new
flow
than
Chevron
14
at
the
original
flow.
Re-
m
oval
efficiency
can
be
conveniently
compared
by
deter-
m
iffing
d,
from
Figures
4-6,
i.e.,
the
droplet
size
for
which
a
C
hevron
achieves
95%
removal.
Values
of
d„
are
shown
in
Figure
7
as
a
function
of
gas
ve
locity
for
Chevrons
10,
14,
and
15.
Assuming
a
density
of
0
.
065
lbm/ft
3
(1.043
kgle)
for
saturated
flue
gas,
the
veloc-
i
t
y
corresponding
to
the
original
flow
rate
is
8.8
ftls
(2.7
ml
s)
while
that
corresponding
to
the
new
flow
is
11.0
ftls
(3.4
mis)•
From
Figure
7
the
d
;
,
5
for
Chevron
No.
14
at
8.8
ftls
(
2,7
mis)
is
24.6
Am.
At
11.0
ftls
(3.4
mis),
Chevron
No.
10
g
ibes
a
d,
of
27.8
Am.
From
a
practical
standpoint,
this
is
c
iose
enough
to
be
considered
acceptable.
By
retrofitting
the
system
with
Chevron
No.
10,
the
capacity
can
be
in-
cr
eased
by
25%
and
the
pressure
drop
reduced
by
43%
while
maintaining
nearly
the
same
removal
efficiency.
Design
Example
2:
Retrofit
for
Reduced
Maintenance.
A
u
tility
was
experiencing
severe
maintenance
problems
with
Chevron
No.
1
in
their
FGD
scrubbers.
Although
collection
e
fficiency
was
adequate
to
prevent
any
carryover
of
mists
to
downstream
equipment,
rapid
pluggage
of
the
chevron
blades
with
solids
resulted
in
rapid
deterioration
of
the
plastic
chevron
blades
due
to
frequent
manual
cleaning.
Both
high
labor
costs
and
high
chevron
module
replace-
m
ent
costs
made
is
desirable
to
select
a
chevron
design
with
le
55
tendency
to
plug.
The
utility
selected
a
chevron
design
with
the
widest
blade
s
pacing
to
resist
pluggage
and
increase
the
ease
of
cleaning
did
washing.
They
selected
Chevron
No.
17,
fabricated
from
the
same
plastic,
as
a
replacement
for
Chevron
No.
1.
Fewer
maintenance
problems
were
experienced
and
the
chevron
was
easier
to
wash.
However,
an
extreme
buildup
of
solids
in
the
ducts
downstream
of
the
FGD
scrubber
was
noted.
This
carryover
was
due
to
the
reentrainment
of
collected
liquid
from
Chevron
No.
17
since
it
was
operating
well
be-
yond
its
maximum capacity
(F,
critical).
Solution.
Chevron
No.
10
was
selected
to
replace
Chev-
ron
No.
17
in
the
FGD
scrubber.
The
capacity
of
Chevron
No.
10
(F,
critical)
was
satisfactory
for
the
operating
condi-
tions,
which
eliminated
the
reentrainment
problems.
The
smooth
and
open
design
of
Chevron
No.
10
greatly
reduced
pluggage
and
improved
washing
over
the
original
Chevron
No.
1,
yet
it
provided
sufficient
collection
efficiency
to
vir-
tually
eliminate
carryover
of
mist
and
solids
downstream
from
the
chevron.
Conclusions
Test
methods
have
been
developed
for
evaluating
the
per-
formance
of
chevrons
with
respect
to
capacity,
pressure
drop,
removal
efficiency,
and
plugging
tendency.
The
first
three
of
these
parameters
are
presented
for
18
different
commercial
vertical-flow
chevrons.
For
the
chevrons
tested
and
the
range
of
test
conditions
studied,
the
capacities
ranged
over
a
factor
of
2,
the
pressure
drops
ranged
over
a
factor
of
10,
and
the
droplet
penetration
ratio
ranged
over
a
factor
of
20
or
more.
All
of
the
chevrons
tested
gave
es-
sentially
complete
removal
of
droplets
larger
than
70
Am
in
diameter,
and
some
gave
essentially
complete
removal
of
10
Am
droplets.
These
data
can
be
used,
along
with
other
se-
lection
criteria,
to
compare
various
chevron
designs
for
ret-
rofit
or
new
installations.
Acknowledgment
The
authors
gratefully
acknowledge
the
work
of
Brian
Murray
and
Ilya
Michelson
in
operating
the
pilot
plant
and
reducing
the
experimental
data.
Literature
cited
1.
Monat,
J.P.,
et
al.,
Chem.
Eng.
Prog..
82
(12),
p.:32
(1986).
2.
Lapple,
C.E.,
"Fluid
and
Particle
Mechanics,"
U.
of
Delaware,
New-
ark,
DE
(1956).
3.
Sherwood,
T.K.,
and
R.L.
Pigford,
"Absorption
and
Extraction,"
McGraw-Hill,
New
York,
pp.
236,
248
(1952).
4.
Portalski,
S.,
Chem.
Eng.
Sd.,
18,
p.
787
(1963).
5.
Mugele,
R.A.,
and
H.D.
Evans,
Ind.
and
Eng.
Chem..
43
(6),
p.
1317
(1951).
6.
May,
K.R.,
J.
Sci.
Inst.,
17,
p.
128
(1950).
Nomenclature
F,
=
superficial
F
factor
defined
by
Eq.
1,
ftls
(Ibm/ft')
=
gas
velocity,
ftls
(m/s)
AP
=
pressure
drop.
lbf/ft'
or
in.
11.0
(1)a)
=
gas
density,
IbmIft'
(kg/m'1
g,
=
dimensional
constant
=
32.17
Ihm•ftls
.
•lbf
Eu
=
Euler
number,
dimensionless
h
=
spacing
between
adjacent
chevron
blades
=
liquid
film
thickness
on
blades
=
ratio
of
wet
to
dry
pressure
drop
below
loading
4.1P,
=
critical
pressure
drop
above
which
reentrainment
occurs,
in.
H
D
critical
F,
above
which
reentrainment
occurs.
ills
(Ihmill')
c!„.,
=
droplet
size
for
which
95%
removal
is
obtained,
pim
May
1987
55