Design of complex wire-mesh mist eliminators


Brunazzi, E.; Paglianti, A.

AICHE Journal 46(6): 1131-1137

2000


Knitted wire-mesh mist eliminators have widespread application in many industrial plants. Despite their extensive use, the open literature regarding them is really limited. Some experimental data and mechanistic models have been published for common knitted wire-mesh mist eliminators formed from a single metal pad. This type of mist eliminator can be used in most distillation and absorption columns, but because of the poor removal efficiency, cannot be used in operations involving acid mist, fine fog resulting from liquid condensation from a saturated vapor, oil mist from compressed gases, and natural-gas dehydration applications. Moreover, other possible problems may arise when the separator is fed with high liquid and gas flow rates, because these conditions can induce flooding in the mist eliminator. In both of these cases, common wiremesh mist eliminators do not perform satisfactorily, and therefore complex wire-mesh mist eliminators have to be installed to improve separation efficiency or to increase allowable liquid loadings while avoiding flooding phenomena. This article presents a mechanistic model based on a set of new experimental data obtained by investigating performance of commercial complex eliminators.

Design
of
Complex
Wire-Mesh
Mist
Eliminators
Elisabetta
Brunazzi
and
Alessandro
Paglianti
Dept.
of
Chemical
Engineering,
Industrial
Chemistry
and
Materials
Science,
University
of
Pisa,
1-56126
Pisa,
Italy
Knitted
wire-mesh
mist
eliminators
have
widespread
application
in
many
industrial
plants.
Despite
their
extensive
use,
the
open
literature
regarding
them
is
really
limited.
Some
experimental
data
and
mechanistic
models
have
been
published
for
common
knitted
wire-mesh
mist
eliminators
formed
from
a
single
metal
pad.
This
type
of
mist
eliminator
can
be
used
in
most
distillation
and
absorption
columns,
but
because
of
the
poor
removal
efficiency,
cannot
be
used
in
operations
involving
acid
mist,
fine
fog
re-
sulting
from
liquid
condensation
from
a
saturated
vapor,
oil
mist
from
compressed
gases,
and
natural-gas
dehydration
applications.
Moreover,
other
possible
problems
may
arise
when
the
separator
is
fed
with
high
liquid
and
gas
flow
rates,
because
these
condi-
tions
can
induce
flooding
in
the
mist
eliminator.
In
both
of
these
cases,
common
wire-
mesh
mist
eliminators
do
not
perform
satisfactorily,
and
therefore
complex
wire-mesh
mist
eliminators
have
to
be
installed
to
improve
separation
efficiency
or
to
increase
allowable
liquid
loadings
while
avoiding
flooding
phenomena.
This
article
presents
a
mechanistic
model
based
on
a
set
of
new
experimental
data
obtained
by
investigating
performance
of
commercial
complex
eliminators.
Introduction
In
chemical
plants,
removal
of
entrained
liquid
from
gas
or
vapor
streams
may
be
required
not
only
to
recover
valuable
products
or
to
protect
downstream
equipment
from
corrosive
liquids,
but
may
also
be
necessary
to
improve
emission
con-
trols.
Selection
of
the
proper
collecting
equipment
depends
mainly
on
the
size
distribution
of
the
entrained
liquid
droplets.
The
present
work
regards
separation
of
entrained
liquid
from
the
gaseous
current
and
considers
droplet
size
from
around
1
micron
upwards,
particular
attention
being
focused
on
the
overlapping
size
region
between
the
coarsest
mist
par-
ticles
and
the
finest
spray
particles.
The
aim
of
this
article
is
to
study
two
of
the
main
problems
arising
in
gas—liquid
sepa-
ration
when
common
wire-mesh
eliminators
are
used:
how
to
obtain
a
high
separation
efficiency
even
with
a
large
quantity
of
liquid
entering
the
collector;
and
how
to
obtain
a
high
separation
efficiency
when
a
large
number
of
small
droplets
with
diameter
of
a
few
microns
are
present.
Wire-mesh
contactors
are
made
by
knitting
wires
to
form
a
layer
that
can
be
rolled
spirally
to
form
cylindrical
elements
(which
are
commonly
used
for
small-diameter
applications)
or
folded
into
several
layers
to
form
a
pad
of
the
desired
Correspondence
concerning
this
article
should
be
addressed
to
A.
Paglianti.
thickness.
In
the
present
article,
two
evolutions
of
the
origi-
nal
wire-mesh
contactors
have
been
analyzed
experimentally:
"the
multilayer
separators,"
formed
from
two
or
three
metal
pads
in
series,
and
"the
composite
separators,"
which
consist
of
knitted
mesh
eliminators
that
incorporate
a
multifilament
yarn
into
the
basic
wire-mesh
structure
(see
Figure
1).
The
metal
wire
used
in
the
knitted
multilayer
generally
has
a
di-
ameter
in
the
80-280-µm
range
and
the
typical
thickness
used
for
each
pad
is
between
20
and
100
mm.
The
composite
sepa-
rators
are
usually
made
from
the
same
type
of
basic
metal
wire
that
is
used
for
the
common
metal
wire
mesh.
Fiber
diameters
range
from
9
to
30
microns,
depending
on
the
fiber
material.
Common
materials
of
the
multifilament
yarns
are
polypropylene,
Dacron,
Teflon,
and
glass
fiber.
The
selection
of
the
fiber
material
depends
on
the
requirements
imposed
by
the
process
conditions
(such
as
corrosion
resistance,
tem-
perature,
liquid
loads).
For
instance,
polypropylene
and
Dacron
yarns,
which
are
cheaper
than
Teflon
yarns,
are
ap-
propriate
for
process
temperatures
up
to
50°C
and
70°C,
re-
spectively,
whereas
Teflon
yarns
can
be
used
up
to
180°C.
Composite
separators
made
from
glass-fiber
yarns
exhibit
an
extraordinarily
high specific
surface,
therefore
allowing
high
removal
efficiency,
but
can
only
treat
gas
streams
with
low
liquid
loads.
Some
articles
have
been
published
on
the
use
of
common
wire-mesh
mist
eliminators.
All
of
these
articles
essentially
AIChE
Journal
June
2000
Vol.
46,
No.
6
1131
4
4
4
(b)
(d)
Figure
1.
Composite
wire-mesh
mist
eliminator,
(a)
sin-
gle
layer,
(b)
two-layer
pad,
(c)
four-layer pad,
(d)
eight-layer
pad.
suggest
how
to
install
mist
eliminators
properly.
Attention
has
been
focused
on
identifying
the
maximum
gas
and
liquid
ve-
locities
to
avoid
flooding
in
working
conditions
and
to
evalu-
ate
separation
efficiency.
Semiempirical
equations
based
on
the
Souders-Brown
rela-
tion
are
commonly
used
(York,
1954).
However,
this
method
of
designing
wire-mesh
mist
eliminators
is
very
rough,
be-
cause
it
does
not
take
into
account
either
the
drop
size,
on
which
the
collection
efficiency
is
strongly
dependent,
or
the
liquid
load
that,
as
pointed
out
by
York
and
Poppele
(1963),
can
induce
flooding
of
the
pad.
Some
articles
in
the
last
few
years
have
shown
that
it
is
possible
to
predict separation performances
based
on
a
mechanistic
description
of
the
separation
phenomena.
Col-
lection
efficiency
of
impingement-type
separators
involves
three
different
separation
mechanisms:
inertial,
interception,
and
diffusion
capture
(Gerrard
et
al.,
1986;
Holmes
and
Chen,
1984;
Feord
et
al.,
1993).
Holmes
and
Chen
(1984)
showed
that
only
inertial
capture
plays
an
important
role
in
separa-
tion
efficiency
for
wire-mesh
separators.
This
implies
that
the
total
separation
efficiency
can
be
evaluated
only
by
taking
into
account
the
contribution
due
to inertial
capture
and
ne-
glecting
interception
and
diffusion
capture.
Some
relations
have
been
published
to
evaluate
the
inertial
capture
effi-
ciency
for
a
single
wire
target,
"1sT
(Langmuir
and
Blodgett,
1946;
Pich,
1966).
All
these
relations
agree
that
the
inertial
capture
efficiency
is
a
function
of
the
Stokes
number,
St,
de-
fined
as:
p
i
.
u
dd
St—
18.114
g
.
where
u
is
the
superficial
gas
velocity,
p
i
is
the
density
of
the
liquid
in
the
droplet,
Ag
is
the
gas
viscosity,
and
d
d
and
d„,
indicate
the
droplet
and
target
diameters,
respectively.
Based
on
this
analysis,
separation
efficiency
of
common
wire-mesh
mist
eliminators
can
be
evaluated
by
considering
the
separation
efficiency
of
the
single
target
and
by
taking
into
account
the
packing
geometry.
Brunazzi
and
Paglianti
(1998)
suggested
the
relation:
n
Tz
M
=
1
(
1
7
/sr
)
m
[
1
7
1sr
Tz
(2)
where
M
is
the
number
of
"reference"
cells
present
in
the
pad,
Tz
is
the
number
of
layers
necessary
to
fill
each
cell,
and
n
is
the
number
of
layers
that
form
the
separator.
M
is
com-
puted
as
a function
of
n,
and
of
the
number
of
layers,
Tz:
n
M
=
int
=
n
1.
The
number
of
layers
necessary
to
fill
each
cell
is
given
by
deg
It=
.
This
can
be
evaluated
if
the
geometric
characteristic,
d
eg
,
of
the
separator
is
known.
From
a
geometrical
analysis,
Brunazzi
and
Paglianti
(1998)
suggested
that
d
eg
-
4
ir
.€
d
w
a
e
(
5
)
where
z
is
the
distance
between
two
successive
layers,
e
is
the
packing
void
fraction,
and
a
e
its
specific
surface.
Finally,
both
the
mechanistic
model
suggested
by
Brunazzi
and
Paglianti
(1998)
and
the
semiempirical
equation
pro-
posed
by
Carpenter
and
Othmer
(1955)
can
be
used
to
com-
pute
the
collection
efficiency
of
common
wire-mesh
separa-
tors.
Both
of
them
correctly
predict
the
separation
efficiency
of
a
separator
if
the
pad
is
thicker
than
65
mm.
On
the
other
hand
some
problems
arise
with
thinner
pads
because
the
semiempirical
equations
available
in
the
literature
tend
to
underestimate
their
separation
efficiency,
while
thinner
pads
are
often
used
in
multilayer
separators.
In
fact,
some
new
tendencies
have
emerged
in
the
last
few
years
in
the
develop-
ment
of
wire-mesh-type
collectors.
Often
the
mist
eliminator
is
made
up
of
two
or
three
metal
mesh
pads
in
series.
Some-
times
the
first
pad,
made
of
fine
wires,
operates
as
a
coa-
lescer
and
is
followed
by
a
second
pad
that
works
as
the
ac-
tual
separator.
However,
when
high
efficiency
is
required
in
the
presence
of
a
high
entrained
liquid
load,
a
first
pad,
made
from
a
low-density
mesh,
reduces
the
liquid
load
arriving
at
the
second
pad.
This
second
pad
can
therefore
have
a
higher
density,
assuring
a
high
removal
efficiency
even
for
small
droplet
size.
Another
important
field
of
application
where
common
wire-mesh
mist
eliminators
can
give
some
problems
is
the
separation
of
droplets
with
dimensions
of
just
a
few
microns.
BErkholz
(1989)
showed
that
if
a
pad
made
from
220
micron
metal
wire
is
used,
high
gas
velocities
and
thus
pressure
drops
as
high
as
20
mbar
are
necessary
to
obtain
a
dp
50
of
1
mi-
cron.
d
50
is
the
diameter
of
the
particles
that
can
be
sepa-
(1)
(3)
(4)
1132
June
2000
Vol.
46,
No.
6
AIChE
Journal
Malvern
drY
Transmitter
unit
To
fan
Receiver
unit
Wire
mesh
separator
Containing
box
Compressed
ai
Water
Carrier
gas
Figure
2.
Experimental
test
apparatus.
rated
with
an
efficiency
greater
than
50%.
The
same
separa-
tion
efficiency
can
be
obtained
with
pressure
drops
as
low
as
2
mbar
when
wires
of
4
microns
are
used.
This
experimental
evidence
shows
that
when
liquid
load
is
low,
it
is
possible
to
obtain
high
separation
efficiency,
even
for
small
droplets,
while
maintaining
low
pressure
drops.
Because
the
applica-
tion
just
discussed
requires
thin
pads,
constructed
from
ei-
ther
metal
wires
or
composite
wires,
already
existing
semiem-
pirical
equations
and
models
cannot
be
used.
The
Experimental
Loop
Experimental
collection
efficiencies
were
determined
as
a
function
of
droplet
size
and
gas
velocity
in
atmospheric
work-
Table
1.
Geometric
Characteristics
of
Multilayer
Packing
Packing
Specific
Void
Wire
Dia.
Dens.
Area
Fraction
Pad
Thick.
Style
(mm)
(kg/m
3
)
(m
2
/m
3
)
(-)
(mm)
A
0.27
116
216
0.985
30
B
0.27
143
267
0.982
105
C
0.27
274
509
0.965
20
ing
conditions
in
an
experimental
loop
designed
and
built
at
the
Chemical
Engineering
Department
of
the
University
of
Pisa.
For
this
purpose
mist
eliminators,
furnished
by
Costacurta
S.p.A.
VICO,
were
tested.
Air
and
water
were
used
as
working
fluids.
The
experimental
rig
mainly
consists
of
a
spray-generation
circuit
and
a
carrier
air
circuit.
The
spray
was
generated
using
an
ultrasonic
nozzle
fed
by
a
volu-
metric
pump,
giving
liquid
flow
rates
ranging
from
0
to
2600
L/h,
and
by
a
compressor
supplying
air
at
6
atm
at
flow
rates
up
to
280
Nm
3
/h.
The
test
section,
shown
in
Figure
2,
consisted
of
a
3-m-long
metal
measuring
section
with
a
rectangular
cross
section,
120
mm
wide
and
190
mm
high.
The
separator
was
installed
hori-
zontally
with
respect
to
the
upflow
of
gas.
A
Malvern
Particle
Sizer
instrument,
based
on
measurements
of
the
diffraction
of
an
He-Ne
laser
beam
by
droplets
moving
through
the
mea-
suring
section,
was
used
to
accurately
measure
the
total
con-
centration
and
volumetric
droplet
distribution
(Brunazzi
and
Paglianti,
1998).
Acquisitions
were
carried
out
both
upstream
and
downstream
of
the
separator.
Each
datum
represents
the
average
of
six
different
acquisitions.
The
accuracy
of
the
measured
efficiency
is
quite
high
and
the
maximum
uncer-
tainty
of
the
measured
efficiency
is
below
5%.
This
article
analyzes
one
metallic
multilayer
separator
and
several
composite
industrial
wire-mesh
mist
eliminators.
The
metal
wires
used
in
all
the
tested
separators
are
made
in
AISI
316.
The
composite
separators
tested
in
the
present
work
are
50
mm
thick.
The
main
geometric
characteristics
of
each
packing
are
shown
in
Table
1
and
Table
2,
respectively.
Fiber
diameter,
fiber
material,
and
specific
surface
area
were
all
varied
in
this
work
in
order
to
investigate
the
influence
of
each
parameter
on
the
separation
performance.
Simulation
Model
of
Complex
Wire-Mesh
Behavior
The
model
that
will
be
presented
is
based
on
the
following
hypotheses
that
were
also
used
by
Brunazzi
and
Paglianti
Table
2.
Geometric
Characteristics
of
Composite
Packings
Style
Metal
Wire
Dia.
(mm)
Fiber
Dia.
µm)
Void
Fraction
Metal
Wt.
(kg/m
3
)
Fiber
Wt.
(kg/m
3
)
Fiber
Material
D
0.27
28
0.970
143.5
10.6
Polypropylene
E
0.27
28
0.958
143.5
21.8
Polypropylene
F
0.27
21
0.976
143.5
13.9
Teflon
G
0.27
22
0.971
143.5
15.3
Dacron
H
0.27
9
0.974
143.5
21.5
Glass
fiber
0.27
9
0.965
143.5
44.2
Glass
fiber
L
0.27
28
0.961
190
14
Polypropylene
M
0.27
28
0.944
190
28.8
Polypropylene
N
0.27
21
0.968
190
18.4
Teflon
0
0.27
22
0.961
190
20.2
Dacron
P
0.27
9
0.965
190
28.4
Glass
fiber
Q
0.27
9
0.954
190
58.3
Glass
fiber
AIChE
Journal
June
2000
Vol.
46,
No.
6
1133
4
w.
I
m
4
2
p
m
'
IF
(1998)
for
common
wire-mesh
mist
eliminators:
(a)
no
reen-
trainment,
(b)
no
buildup
of
liquid,
and
(c)
no
mixing
after
passage
through
each
layer.
The
first
two
hypotheses
are
also
found
in
the
model
suggested
by
Carpenter
and
Othmer
(1955),
whereas
the
last
hypothesis
represents
one
of
the
dif-
ferences
between
the
present
model
and
the
work
by
Carpen-
ter
and
Othmer
(1955).
The
model
presented
by
Brunazzi
and
Paglianti
(1998)
sug-
gested
that
the
separator
should
be
schematized
as
a
series
of
reference
cells.
Notwithstanding
its
simplicity,
the
model
allows
a
satisfactory
prediction
of
separation
efficiency
to
be
made.
However,
its
range
of
application
is
limited
to
common
pads
made
from
a
single
metal
wire
type
and
to
homoge-
neous
separators:
therefore,
it
cannot
be
applied
to
predict
separation
efficiency
of
multilayers
or
composite
mesh
sepa-
rators.
An
extension
of
the
model
to
predict
multilayer
efficiency
can
easily
be
performed.
These
new
mist
eliminators
can
be
computed
as
a
series
of
homogeneous
pads.
Therefore
the
total
efficiency
of
the
multilayer
pad
made
by
m
pads
can
be
computed
as
7
/:
=
1—
11
;11(
1
—7
/,),
(6)
where
The
free
volume
of
the
composite
pad,
e
c
,
can
be
evalu-
ated
as
7F 7F
(
1
=
71.
d,
2
.'i
m
+7
1
d;•
N
f
I
f
=
0
e.)
[1
+
N
f
(
2
i
d
).
1
41,
(10)
where
E
m
is
the
free
volume
of
the
metal
pad
supporting
the
nonmetallic
fibers.
In
a
similar
way,
it
is
possible
to
evaluate
the
specific
sur-
face
area
of
composite
pad
as
(
a
e
=
7F
d
m
I
m
+
N
f
IT
df
lf
=
a
m
1+
Ar,•
d
f
if
'
=
=
(11)
'
d
m
I
m
where
a
m
is
the
specific
surface
area
of
the
metal
pad
sup-
porting
the
nonmetallic
fibers.
Now,
it
is
necessary
to
evaluate
the
diameter
of
an
equiva-
lent
wire,
d
e
,
that
has
the
same
length
as
the
metal
wire
but
shows
the
specific
surface
of
the
composite
pad.
The
equiva-
lence
of
the
specific
surface
area
gives
n.—
fi
M.
m=1—
(1—
71„)
111
'
1—
"
(
7
)
7F
d
e
I
m
=
IF
d
m
•I
m
+
N
f
IF
df
lf
(12)
and
M
i
is
the
number
of
"reference"
cells
present
in
the
ith
pad,
n
,
is
the
number
of
layers
necessary
to
fill
the
cell
of
the
ith
pad,
n
i
is
the
number
of
layers
that
form
the
ith
pad,
and
is
the
efficiency
for
a
single
wire
target
of
the
ith
pad.
The
evaluation
of
the
characteristic
length
of
each
layer,
d
eg
.,
can
be
made
in
the
same
way
as
for
the
common
wire-mesh
mist
eliminator,
because
of
their
geometrical
analogy.
While
an
extension
of
the
model
is
quite
easy
for
multi-
layer
separators,
it
is
more
complex
to
define
the
geometrical
characteristics
of
composite
wire-mesh
mist
eliminators.
The
vendors
usually
give
the
weight
of
metal
wires
for
a
unit
volume
of
pad,
W
m
,
and
the
weight
of
the
fibers
for
a
unit
volume,
w
f
;
therefore,
the
free
volume
and
the
surface
area
have
to
be
evaluated.
The
length
of
metal
wire,
I„„
and
of
fiber,
/
f
,
for
a
unit
volume
of
the
pad
can
be
computed
by
geometrical
analysis
as
and
4
w
f
1
1
2
where
N
1
is
the
number
of
single
fibers
that
make
up
the
nonmetallic
monofilament,
d
m
and
d
f
are
the
metal
wire
di-
ameter
and
the
fiber
diameter,
respectively,
and
p
m
and
p
f
are
the
metal
and
fiber
densities.
which
can
be
rewritten
as
d
e
=
d
m
•(1+
N
f
i
d
1
4).
(13)
The
last
parameter
that
has
to
be
evaluated
in
order
to
use
the
mechanistic
model
suggested
by
Brunazzi
and
Paglianti
(1998)
represents
the
characteristic
length,
d
eg
,
of
the
refer-
ence
cell.
These
authors
defined
this
length
as
Cross
section
d
eg
4
Wetted
perimeter
(14)
where
the
wetted
perimeter
is
a
function
of
the
length
of
metal
wire,
I
m
,
of
the
packing
cross
section,
A,
and
of
the
distance
between
two
successive
layers,
z,
and
can
be
evalu-
ated
as
P
m
=
I
m
A
z.
(15)
For
a
composite
mesh,
the
wetted
perimeter
has
to
be
computed
taking
into
account
the
presence
of
nonmetallic
fibers,
and
therefore
it
can
be
evaluated
as
P
e
=
P
m
•(1+
N
I
f
f
df
d
m
I
m
Finally,
the
characteristic
length
of
a
composite
mesh
can
be
(8)
(9)
(16)
1134
June
2000
Vol.
46,
No.
6
AIChE
Journal
Sep
ara
t
ion
e
ffic
.
(17)
100
V
%
X
.
*
#A
80
7
,
•D
o
v
i4
_
v
x
60—
%
#
40
20
7
o
Style
D
x
Style
E
Style
F
Style
G
O
Style
H
0
Style
I
Style
L
0-
Style
M
Style
N
#
Style
0
*
Style
P
Style
Q
0
,
A
x
A
0
Sep
a
ra
t
io
n
e
ffi
c
i
computed
as
d
eg
4
Wetted
perimeter
d,
1,
P.•(1+N
f
•—
d
:
•—
in
i
4.e,
4.
71"
e
c
d
m
(
(i
m
f
-
I
a
m
z•
1+
N
•—
f•
d
1f)
.
Equation
7
makes
it
possible
to
compute
the
separation
efficiency
if
the
efficiency
of
a
single
target,
S
ST
,
is
known.
As
pointed
out
by
Lucas
(1983),
the
theoretical
analysis
by
Langmuir
and
Blodgett
(1946),
which
allows
the
evaluation
of
separation
efficiency
of
a
single
target,
can
induce
underesti-
mation
of
the
separation
efficiency
when
it
is
applied
to
an
array
of
targets
that
are
close
to
each
other.
This
effect
is
probably
due
to
a
mutual
influence
between
single
targets.
By
taking
this
experimental
observation
into
account,
Brunazzi
and
Paglianti
(1998)
showed
that
separation
effi-
ciency
of
a
common
wire-mesh
mist
eliminator
can
be
prop-
erly
evaluated
if
the
following
empirical
relation
is
intro-
duced
as
a
closure
equation:
if
St
<1,
then
71,
57
,
=
St,
(18)
whereas,
if
St
1,
then
n
sr
=1,
(19)
the
Stokes
number
being
defined
by
Eq.
1.
Therefore,
the
Stokes
number
can
be
easily
evaluated
for
common
wire-mesh
mist
eliminators
and
for
metal
multilayer
pads.
Using
a
simi-
lar
approach
to
that
suggested
for
a
metal
pad,
the
Stokes
number
for
a
composite
mesh
separator
has
been
defined
as
p
i
.
u
4
St=
18.
lig
.
d
e
'
(20)
where
the
target
diameter
of
the
composite
wire
has
been
assumed
to
be
equal
to
the
equivalent
wire
diameter
(see
Eq.
17).
This
simplifying
hypothesis
is
justified
by
analysis
of
Figure
3,
which
shows
a
plot
of
all
the
experimental
data
obtained
with
composite
separators.
It
can
be
noted
that
even
if
the
specific
surface
area
varies
within
the
range
of
1946-10,320
m
2
/m
3
,
the
experimental
data
obtained
at
Stokes
numbers
of
greater
than
1
display
separation
efficiencies
of
nearly
100%.
These
results
agree
with
the
experimental
trends
obtained
by
Brunazzi
and
Paglianti
(1998)
for
pads
made
from
a
single
metal
wire.
Analysis
of
Experimental
Results
When
a
gas/liquid
separator
has
to
be
chosen,
the
first
problem
is
to
decide
which
kind
of
separator
to
use.
If
a
sep-
arator
made
from
a
single
type
of
wire
is
used,
higher
liquid
removal
efficiency
can
be
obtained
by
increasing
the
density
of
the
pad,
but
this
can
lead
to
flooding
problems.
Therefore,
0
1
10
Stokes
number,
St
Figure
3.
Separation
efficiency
vs.
Stokes
number.
All
experimental
data
obtained
with
composite
pads.
if
a
common
pad
made
from
a
single
type
of
wire
is
used,
and
if
it
has
to
work
at
high
liquid
loads,
reentrainment
problems
can
only
be
avoided
by
working
at
low
gas
velocity
with
a
consequent
lowering
of
removal
efficiency.
For
instance,
to
avoid
flooding
of
a
pad,
such
as
a
York
wire
mesh
mist
eliminator
type
931
in
18-8
stainless
steel,
the
maximum
liquid
entrainment
load
is
4,564
kg/(m
2
.h)
with
an
air
velocity
of
3.7
m/s
(see
York
and
Poppele,
1963).
The
separation
performance
of
this
pad
can
be
evaluated
using
the
model
suggested
by
Brunazzi
and
Paglianti
(1998),
but
to
evaluate
the
nonseparated
liquid
rate,
the
liquid-drop
distri-
bution
has
to
be
known.
If
the
liquid
distribution
is
assumed
to
be
equal
to
the
distribution
suggested
by
Garner
et
al.
(1983)
for
steam—water
evaporators,
the
liquid-drop
concen-
tration
in
the
outlet
stream
is
evaluated
as
15.5
ppm.
If
a
higher
liquid
separation
efficiency
is
required,
a
more
effi-
cient
pad
has
to
be
used
but,
as
pointed
out
by
York
and
Poppele
(1963),
flooding
phenomena
will
occur.
Multilayer
separators
can
be
the
solution
in
this
case
(York,
1993).
These
collectors
allow
separation
efficiency
to
be
improved
at
high
liquid
loads,
while
maintaining
the
pressure
drop
low.
From
the
theoretical
point
of
view,
a
multilayer
separator
can
be
analyzed
as
a
series
of
single
pads.
Unfortunately,
no
experimental
data
on
separation
efficiency
have
been
pub-
lished
yet.
Figure
4
shows
a
comparison
between
present
ex-
perimental
data
and
the
computed
values
obtained
using
the
100
0
u=1
m/s
80
o
u=2
m/s
0
u=3
m/s
60
40
20
7
10
15
20
Droplet
diameter,
d
d
(microns)
Figure
4.
Separation
efficiency
of
a
multilayer
pad
vs.
drop
diameter.
Comparison
between
experimental
measurements
and
the
new
model.
Cross
section
4
A
e
e
z•(1+
N
f
df
if
)
d,,,
1.
AIChE
Journal
June
2000
Vol.
46,
No.
6
1135
100
80
Style
Sep
a
ra
t
io
n
e
ffic
iency,
ti
(
%)
60
Style
M
40
7
20
-
Style
N
Style
L
Style
L
o
Style
M
Style
N
o
Style
0
10
15
Droplet
diameter,
d
d
(microns)
0
0
20
10
Pressu
re
drop,
(m
ba
r
)
0.1
Sep
a
ra
t
io
n
e
ffic
o
Present
experimental
data
-
—Burkholz,
1989
10
d95
(microns)
Figure
5.
Pressure
drop
vs.
d
95
.
Comparison
between
present
experimental
measurements
and
BErkholz's
relation.
approach
suggested
in
the
present
work.
This
figure
shows
that
the
model
allows
the
separation
efficiency
to
be
evalu-
ated
with
an
acceptable
degree
of
accuracy.
Using
this
model,
it
is
thus
possible
to
design
or
to
opti-
mize
new
separators
that
allow
high
separation
efficiency
even
with
high
liquid
entrainment
loads.
For
instance,
as
pointed
out
before,
when
a
150-mm-thick
York
wire-mesh
mist
elimi-
nator
type
931
is
used
with
a
liquid
load
of
4564
kg/(h•m
2
)
and
a
gas
velocity
of
3.7
m/s,
15.5
ppm
of
liquid
are
still
present
in
the
gas
phase
downstream
of
the
separator.
But
if
a
multilayer
separator
is
used
in
a
configuration
consisting
of
85
mm
of
the
931
type
and
65
mm
of
the
421-type
mist
elimi-
nator,
for
instance,
the
liquid
load
on
the
second
pad
can
be
evaluated
as
0.36
kg/(h•m
2
).
Therefore
flooding
phenomena
are
avoided
and
a
higher
separation
efficiency
is
achievable.
In
this
case,
a
liquid
concentration
of
8.5
ppm
in
the
gas
phase
downstream
the
separator
has
been
estimated.
More
interesting
results
can
be
obtained
if
composite
pads
are
used.
As
pointed
out
by
BErkholz
(1989),
composite
pads
allow
high
separation
efficiencies
to
be
obtained
even
for
small
droplets.
Unfortunately,
few
experimental
data
have
been
published
on
the
separation
performances
of
these
types
of
separators.
BErkholz
(1989)
proposed
a
simple
and
reli-
able
empirical
way
to
estimate
the
diameter
of
the
particles
that
can
be
separated
with
an
efficiency
greater
than
95%,
labeled
d
95
.
He
suggested
plotting
d
95
against
pressure
drop
to
verify
whether
a
relation
linking
these
two
parameters
ex-
ists.
Figure
5
shows
a
comparison
between
BErkholz's
rela-
tion
and
experimental
data
from
the
work
presented
in
this
article.
It
can
be
noted
that
the
new
data
agree
with
the
Brukholz
empirical
relation
(1989).
From
an
empirical
point
of
view,
the
equation
suggested
by
BErkholz
(1989)
can
be
used
to
predict
d
95
,
but
unfortunately
no
model
has been
developed
to
predict
liquid
separation
efficiency
of
compos-
ite
pads.
Therefore,
in
order
to
predict
the
separation
effi-
ciency,
some
preexisting
models,
tested
for
single
metal
lay-
ers,
have
been
modified.
Most
of
the
published
equations
re-
fer
to
Carpenter
and
Othmer's
model
for
common
single
metal
wire
pads.
These
authors
suggested
the
following
em-
pirical
equation:
2
=
1
1
a
Ilsr
—)
3
e
7F
100
80
--
60
--
40
20
—I
0
o
Experimental
data
d
=
d
(1,
=d
0
5
10
15
20
Droplet
diameter,
d
d
(microns)
Figure
6.
Separation
efficiency
vs.
drop
diameter.
Comparison
between
experimental
measurements
with
the
Carpenter
and
Othmer
(1955)
model
using
4=
d„,
(con-
tinuous
line)
and
4=
d
f
-
(dashed
line).
Packing
style
H,
0.5
m/s
superficial
gas
velocity,
50-mm
pad
thickness.
where
a
e
is
the
specific
surface
area
of
the
separator,
z
the
distance
between
two
successive
layers,
n
the
number
of
lay-
ers,
and
the
efficiency
of
a
single
target.
Because
is
a
function
of
the
Stokes
number,
if
the
use
of
the
equation
is
to
be
extended
to
a
composite
pad,
it
is
necessary
to
define
the
diameter
d
e
.
There
are
two
possible
choices:
d
e
can
be
assumed
either
equal
to
the
metal
wire
diameter
or
equal
to
the
fiber
diameter.
Figure
6
shows
a
comparison
between
the
present
experimental
data
and
the
Carpenter
and
Othmer
model
using
the
diameter
of
the
metal
wire,
d
m
,
and
the
fiber,
d
f
,
when
defining
the
Stokes
number.
Analysis
of
the
figure
shows
that
the
modified
model
by
Carpenter
and
Othmer
(1955)
overestimates
the
separation
efficiency
when
the
fiber
diameter
is
used
in
the
Stokes
number,
whereas
it
largely
underpredicts
the
measured
efficiency
when
the
metal
wire
diameter
is
used
in
the
Stokes
number.
Therefore,
Carpenter
and
Othmer's
model
cannot
be
extended
to
composite
pads
without
introducing
large
errors
in
the
prediction
of
the
sepa-
ration
efficiency.
At
the
same
time,
it
was
found
that
Carpen-
ter
and
Othmer's
model
also
induces
large
errors,
even
when
the
equivalent
diameter
suggested
in
the
present
work
(Eq.
17)
is
used.
Figure
7
shows
a
comparison
between
the
experimental
data
obtained
in
this
article
and
the
new
model.
The
acquisi-
Figure
7.
Separation
efficiency
vs.
drop
diameter,
effect
of
specific
surface.
Comparison
between
experimental
measurements
with
the
new
model.
Superficial
gas
velocity
0.75
m/s,
50-mm
pad
(21)
thickness
(packing
style
M,
0;
packing
style
0, 0;
packing
style
L,
A;
packing
style
N,
*).
1136
June
2000
Vol.
46,
No.
6
AIChE
Journal
80
7
C
Sep
ara
t
io
n
e
ffic
60
7
40
20
7
4
6
8
10
Droplet
diameter,
d
d
(microns)
Figure
8.
Separation
efficiency
vs.
drop
diameter,
effect
of
superficial
gas
velocity.
Comparison
between
experimental
measurements
and
the
new
model.
Packing
style
D,
50-mm
pad
thickness.
obtained
in
this
article
shows
that
this
new
model
can
be
used
both
for
predicting
separation
efficiency
of
multilayer
pads
and
to
predict
the
separation
efficiency
of
composite
separators.
The
proposed
model
allows
the
measured
effi-
ciency
to
be
predicted
with
sufficient
accuracy
even
though
no
adjustable
parameter
has been
used.
This
result
is
signifi-
cant
because
no
mechanistic
model
has
yet
been
published
in
the
literature
to
predict
the
separation
efficiency
of
complex
wire-mesh
mist
eliminators.
This
could
be
an
important
im-
provement,
since
an
increasing
number
of
industrial
separa-
tors
are
composite
or
multilayers
and
only
empirical
relations
are
available
to
predict
their
separation
efficiency.
The
new
model
presented
in
this
article
allows
the
contribution
of
each
single
pad,
metallic
or
composite,
to
be
evaluated,
and
could
therefore
be
used
for
the
design
and
the
optimization
of
complex
separation
units.
100
0
0
u=
1.25
m/s
u=
0.75
m/s
i=
0.75
m/s
A
u=
1
m/s
u=
1.0
m/s
o
u=
1.25
m/s
Co
mp
u
te
d
e
ffic
iency,
11
(
%)
tions
refer
to
four
packings
with
different
geometrical
prop-
erties,
working
at
the
same
gas
load.
The
specific
surface
of
the
separators
was
varied
in
the
range
of
1972-5209
m
2
/m
3
.
It
can
be
noted
that
the
new
model
and
experimental
data
agree
for
all
the
pads
analyzed.
Figure
8
shows
a
comparison
between
computed
and measured
efficiency
of
a
pad
working
at
different
gas
loads.
Also
in
this
case,
the
model
presented
in
this
article
predicts
the
experimental
performances
of
the
separator
with
a
sufficient
degree
of
accuracy.
Finally,
a
parity
plot
is
shown
in
Figure
9,
which
compares
the
experimental
data
obtained
in
this
work
and
the
com-
puted
values
obtained
using
the
present
model.
Analysis
of
the
figure
shows
that
the
new
model
makes
it
possible
to
evaluate
separation
efficiency
with
sufficient
accuracy
for
all
the
separators
analyzed
in
this
work,
notwithstanding
the
fact
that
no
adjustable
parameters
have
been
introduced.
Conclusions
The
experimental
data
on
droplet
removal
efficiency
pre-
sented
in
this
article
were
obtained
using
a
laser-based
droplet
sizer,
the
Malvern
Particle
Sizer.
This
article
has
presented
a
new
model
for
predicting
removal
efficiency
of
complex
wire-mesh
mist
eliminators.
Analysis
of
the
experimental
data
100
80
60
7
40
7
20
0
20
40
60
80
100
Measured
efficiency,
0
06)
Figure
9.
Separation
efficiency.
Comparison
between
experimental
measurements
and
the
new
model
(composite
pads).
Acknowledgments
This
publication
is
based
on
work
supported
by
Costacurta
S.p.A.
VICO,
Via
Grazioli
30,
20161
Milan,
Italy.
The
authors
would
like
to
thank
Ing.
A.
Vitaletti
and
Ing.
N.
Gizzi
for
their
helpful
assistance,
and
Ing.
B.
Mondello
and
Ing.
A.
Lu-
ongo
for
some
useful
discussions.
Notation
d
eg
=
equivalent
diameter
of
the
mesh
n
n
=capture
efficiency
7r
=3.14159...
p
=wire
mesh
packing
w
=
wire
Literature
Cited
Brunazzi,
E.,
and
A.
Paglianti,
"Design
of
Wire
Mesh
Mist
Elimina-
tors,"
AIChE
J.,
44,
505
(1998).
BErkholz,
A.,
Droplet
Separation,
VCH,
Weinheim,
Germany
(1989).
Carpenter,
C.
L.,
and
D.
F.
Othmer,
"Entrainment
Removal
by
a
Wire-Mesh
Separator,"
AIChE
J.,
1,
549
(1955).
Feord,
D.,
E.
Wilcock,
and
G.
A.
Davies,
"A
Stochastic
Model
to
Describe
the
Operation
of
Knitted
Mesh
Mist
Eliminators,
Com-
putation
of
Separation
Efficiency,"
Trans.
Inst.
Chem.
Eng.,
71,
282
(1993).
Garner,
F.
H.,
S.
R.
M.
Ellis,
and
J.
A.
Lacey,
"Gas
in
Liquid
Dis-
persion,"
Chemical
Engineers'
Handbook,
5th
ed.,
Chap.
18,
R.
H.
Perry
and
C.
H.
Chilton,
eds.,
McGraw-Hill,
New
York,
p.
81
(1983).
Gerrard,
M.,
G.
Puc,
and
E.
Simpson,
"Optimize
the
Design
of
Wire-Mesh
Separators,"
Chem.
Eng.,
93(21),
91
(1986).
Holmes,
T.
L.,
and
G.
K.
Chen,
"Design
and
Selection
of
Spray/Mist
Elimination
Equipment,"
Chem.
Eng.,
91(21),
82
(1984).
Langmuir,
I.,
and
K.
B.
Blodgett,
U.S.
Army
Air
Forces
Tech.
Rept.,
5418
(1946).
Lucas,
R.
L.,
"Gas-Solid
Separations,"
Chemical
Engineers'
Hand-
book,
5th
ed.,
Chap.
20,
R.
H.
Perry
and
C.
H.
Chilton,
eds.,
Mc-
Graw-Hill,
New
York,
p.
81
(1983).
Pich,
J.,
Aerosol
Science,
Chap.
9,
C.
N.
Davies,
ed.,
Academic
Press,
New
York
(1966).
York,
0.
H.,
Glitsch
International
Companies
Bulletin,
5245-1/93
(1993).
York,
0.
H.,
"Performance
of
Wire-Mesh
Demisters,"
Chem.
Eng.
Prog.,
50,
421
(1954).
York,
0.
H.,
and
E.
W.
Poppele,
"Wire
Mesh
Mist
Eliminators,"
Chem.
Eng.
Prog.,
59,
45
(1963).
Manuscript
received
July
6,
1999,
and
revision
received
Jan.
10,
2000.
AIChE
Journal
June
2000
Vol.
46,
No.
6
1137