Office of Air and Waste Management
Air Pollution Training Institute
Control of Particulate Emissions
Manual:TrainingCourse413
APRIL 1975
-------
Control of Particulate Emissions
Conducted by
CONTROL PROGRAMS DEVELOPMENT DIVISION
Air Pollution Training Institute
Research Triangle Park, North Carolina 27711
The Control of Particulate Emissions manual has been prepared
specifically for the trainees attending the course and should not
be included in reading lists or periodicals as generally available.
-------
\v/5
&
'/
This is not an official policy and standards
document. The opinions, findings, and conclusions
are those of the authors and not necessarily those
of the United States Environmental Protection Agency.
Every attempt has been made to represent the
present state of the art as well as subject areas
still under evaluation. Any mention of products,
or organizations, does not constitute endorsement
by the United States Environmental Protection Agency.
-------
?
<
AIR POLLUTION TRAINING INSTITUTE
MANPOWER AND TECHNICAL INFORMATION BRANCH
CONTROL PROGRAMS DEVELOPMENT DIVISION
OFFICE OF AIR QUALITY PLANNING AND STANDARDS
77?e XUV Pollution Training Institute (1) conducts training for personnel working on
the development and improvement of state, and local governmental, and EPA air
pollution control programs, as well as for personnel in industry and academic insti-
tutions; (2) provides consultation and other training ass/stance to governmental
agencies, educational institutions, '^-iustrial organizations, and others engaged in
air pollution training activities; and (3) promotes the development and improve-
ment of air pollution training programs in educational institutions and state, regional,
and local governmental air pollution control agencies. Much of the program is now
conducted by an on-site contractor, Northrop Services, Inc.
One of the principal mechanisms utilized to meet the Institute's goals is the intensive
short term technical training course. A full-time professional staff is responsible for
the design, development, and presentation of these courses. In addition the services
of scientists, engineers, and specialists from other EPA programs, governmental
agencies, industries, and universities are used to augment and reinforce the Institute
staff in the development and presentation of technical material.
Individual course objectives and desired learning outcomes are delineated to meet
specific program needs through training. Subject matter areas covered include air
pollution source studies, atmospheric dispersion, and air quality management. These
courses are presented in the Institute's resident classrooms and laboratories and at
various field locations.
Robert G. Wilder
Program Manager
Northrop Services, Inc.
i J. Schueneman
Chief, Manpower & Technical
Information Branch
-------
Contents
CONTROL OF PARTICULATE EMISSIONS
Section _!_
Motivation for Control Measures
Section 2
Introduction: Collection Equipment
Introduction: The Collection of Particles
from a Gas Stream
Section 3
Conversion Factors
Section 4
Gas Properties Basic Concepts
Section 5
Particle Settling Dynamics
Terminal Velocities of Spherical Particles
Section 6
The Separation of Particles into Size-
Fractions
Section 7
Notes on the Analysis of Particle Size
Distributions
Section 8
The Effective Particle Size
Section 9
Representation of Particle-Size Data
Statistical Presentation of Data
Size-Efficiency Curves
Section 10
Settling Chambers
Section 11
Cyclones
Section 12
Miscellaneous Dry Inertial-Type Collectors
Section 13
Wet Collectors: Introduction
Section 14
Collection of Particles on Cylindrical and
Spherical Obstacles
Section 15
The Gravity Spray Tower
Section 16
Venturl Scrubbers
Section 17
Collectors with Self-Induced Sprays
Section 18
Impingement Type Scrubbing Tower
Section 19
Wet Centrifugal Collectors
Section 20
Wet Dynamic Predpitator
Section 21
Disintegrator Scrubbers
Section 22
Fabric Filtration
Section 23
Fabric Filtration - Basic Concepts
Section 24
Fabric Filtration Operations and Industrial
Applications
Section 25
Fabric Filtration - Mathematics of Bag-
Filter Operation
-------
Contents
Section 26
Effect of Changing Permeability, Varying
Flow Rate, and Non-Laminar Head Loss
Section 27
Electrostatic Precipitators Operation and
Industrial Applications
Section 28
High Temperature Gas Cleaning
Section 29
Sanitary Disposal of Collected Material
Section 30
Cost of Collection Equipment
Section 31
The Sylvan Chart
Section 32
Class Problems
-------
SLTTIGN 1
Motivation for Control Measures
-------
MOTIVATION FOR CONTROL MEASURES
I MOTIVATION FOR CONTROL
MEASURES IN INDUSTRY
A From the standpoint of the industrial plant,
air pollution control is motivated by:
1 The economic value of recovered
material, or the value of the cleaned
gas,
2 The desire on the part of management
to maintain satisfactory relations both
with the workmen and the neighbors in
the surrounding community, and
3 The legal requirements as covered by
ordinances and statutory regulations.
II THE ECONOMIC VALUE OF
RECOVERED MATERIAL, OR THE VALUE
OF THE CLEANED GAS
A Industrial management recognizes that
production results in two kinds of materials:
1 Saleable products, which are a source
of profit.
2 Waste products, which are a charge
against production.
a This charge against production is
increased by the cost of abating
airborne wastes.
B The Economic Value of Recovered Material
1 The cost of air cleaning equipment, in
some instances, is paid for from
salvaged material even though the pri-
mary reason for the installation is the
prevention of an air pollution problem
to the plant or neighborhood.
a Some examples are:
1) Flour dust in bakeries
2) Brass griding dust in metal
finishing
3) Ore dust from crushing and
milling
4) Line dust from kilns
5) Sugar dust from dryers and
coolers
In some cases, the collection equip-
ment is of primary importance, not
to air pollution prevention, but to the
economic operation of the manufacturing
process itself. However, satisfactory
performance of the collector is of
benefit both from a manufacturing and
air pollution standpoint.
a Some examples are:
1) The manufacture of carbon black
2) The sintering and roasting of
lead ores
3) The pulverizing of chemicals
In many cases, the installation of
collectors cannot be justified from an
economic standpoint.
a Any justification at all is then only
on the basis of establishing good
labor and community relations.
Ill LABOR AND COMMUNITY RELATIONS
A To a large extent, the degree to which
particulates must be removed from a
flue gas before emission to the atmosphere
is governed by labor and community
relations.
B The workmen and the community are
concerned with:
1 The deposit of coarse particles which
settle in the general area of the stack
and create problems of a localized
nature.
a Such settlement is mostly a nuisance
to the neighborhood property, but
PA. C. pm. 17a.9.60
1
-------
Motivation for Control Measures
may also cause damage by acceler-
ating or inducing oxidation or
chemical attack.
b Examples of coarse emissions:
1) The foundry cupola and lime
hydrator. Settled material on
roofs cause endless main-
tenance in material removal
and roof repair.
2) The aggregate dryer, the cement
kiln, the stoker-fired boiler,
the pulverized fuel boiler.
Coarse emissions settle on
porches, ledges, and laundry
over an extended area from
the point of emission.
'.'•) The exhaust of ferrous grinding,
heat treat furnances, plating
and caustic tanks, can dis-
charge materials that damage
the finish of automobiles.
Good housekeeping on the plant
grounds, both inside and outside
has an aesthetic effect on the work-
men and members of the community.
C The cumulative effect of all the above on
labor and public relations cannot be
overestimated.
1 When it is understood that plant manage-
ment is making a continuous effort to
remedy or improve air-polluting
emissions, and working conditions are
relatively clean, both the employee
and the community are inclined toward
the company.
IV LEGAL REQUIREMENTS
A Regulations Limiting the Opacity of the
Plume
The deposit of fine particles that
travel for long distances in the atmos-
phere contributing problems of an
area-wide nature.
a Such particles can be a nuisance,
a health hazard, reduce visibility,
and cause soiling and damage to
materials and vegetation.
b I'CxanipJe.s of fine particle emissions:
I) Molting operations which emit
fumos such as zinc oxide and
ferrous oxide.
2) (' irbonaceous matter from burn-
ing of coal, oil, gas, and
rubbish.
The plurr.e appearance
a A "dirty" plume has a psychological
effect on those viewing it.
4 The cleanliness of the plant area,
both inside and out.
1 Nearly every air pollution control
ordinance now existing has its origin
in smoke control regulations which
were designed to limit the density of
stack emissions by placing restrictions
on the opacity of the plume.
2 The smoke regulation concept has
carried over into present day air
pollution ordinances, in that some
restrictions on the opacity of industrial
stack gases are specified.
3 Examples of ordinances which recognize
the Ringelmann Chart as a visual gauge
of permissible visible emission are
shown in Table 1.
4 At present, there are no legal require-
ments that effluent gases be free from
all visible contaminants and the
possibility of such restrictions in the
foreseeable future is remote.
However, from the standpoint of public
relations, a stack discharge containing
sufficient visible contaminants to be
conspicuous should be avoided wherever
possible.
-------
Motivation for Control Measures
Table 1. DEFINITION OF ORDINANCE CLASSIFICATIONS
BASED ON SMOKE DENSITY*
CLASS I:
Ordinances which allow only Ringelmann No. 1
smoke:
A With no exception stated;
B Except for stated periods when cleaning or
building fires or other reason.
CLASS II:
CLASS III:
CLASS IV:
CLASS V:
CLASS VI:
Ordinances which allow short periods of No. 2
smoke:
A But may not exceed No. 3 for stated periods
for fire cleaning or building;
B And may exceed No. 3 for stated periods for
fire cleaning or building.
Ordinances which allow No. 2 smoke at all times:
A But may not exceed No. 2 at any time;
B But may not exceed No. 3 for fire cleaning or
building;
C And may exceed No. 3 for fire cleaning or
building.
Ordinances which allow short periods of No. 3
smoke:
A But may not exceed No. 3 at any time
B And may exceed No. 3 for fire cleaning and
building.
Ordinances which allow No. 3 smoke at all times:
A But may not exceed No. 3 at any time:
B And allow periods in excess of No. 3 without
specifying or in addition to fire cleaning or
building.
C And may exceed No. 3 for fire cleaning or building.
Ordinances which define smoke density using the
Umbrascope and not in terms of the Ringelmann
Chart
CLASS VII:
Ordinances which do not, or only loosely, define
the smoke density prohibited.
Air Pollution Abatement Manual, Manufacturing Chemists Assoc. 1953.
-------
Motivation for Control Measures
B Regulations limiting the emission of
particulate matter in terms of weight
units.
1 In some areas, attempts have been
made to reduce allowable stack
emissions to a quantitative basis. Thus,
ordinances have been enacted which in-
clude restrictions on either the con-
centration (weight per unit volume) or
mass rate (weight per unit time) of
emission of pollutant.
2 Legal requirements concerning parti-
culate emission vary over a wide range
of operating conditions, (see "Selected
Dust Emission Limitations of Typical
Communities in Various Population
Groups, " in Appendix)
3 Examples of dust loading prohibitions
a 75% collection entering collector,
minimum
b 85% collection entering collector,
minimum
e 90% collection entering collection
d 0.85 Ibs. per 1000 Ibs. gas or air
e 0.85 Ibs. per 1000 Ibs. of gas
adjusted to 12% CCX
iL
f 0.85 Ibs. per 1000 Ibs. of gas - 50%
cxeess air
g 0.85 Ibs. per ]000 Ibs. of gas
adjusted to 50% excess air. Maxi-
mum 0.5 Ib.s. per 1000 Ibs. of gas
shall be larger than 325 mesh.
h 0.55 Ibs. per 1000 Ibs. of gas
adjusted to 50% excess air. Maxi-
mum 0.2 Ibs. of dust larger than
325 mesh.
i 2 Ibs. per 1000 Ibs. of gases at
12% CO must collect 75%
tL
j 0.30 gr/cu.ft. at 500°F and 50%
exress air
k 0.30 gr/cu.ft. at 500°F and 50%
excess air not to exceed 0. 2 gr/cu.
ft. larger than 325 mesh.
1 0.425 gr/cu.ft. at 500°F and 50%
excess air.
m 0.75 gr/cu.ft. and 50% excess air
of which not more than 0.4 gr/cu.
ft. shall be larger than 325 mesh.
n 0. 75 gr/cu. ft. at stack temperature,
not more than 0. 2 gr/cu. ft. retained
on a 300 mesh U.S. Standard sieve.
Excess air not to exceed 50% at
full load.
o 0. 75 gr/cu. ft. at 500°F and 50%
excess air of which not more than
0. 2 gr/cu.ft. shall be larger than
325 mesh.
p 0.75 gr/cu.ft. at 500°F and 50%
excess air of which not more than
0. 2 gr/cu.ft. with gas at 850°F
shall be larger than 325 mesh.
q Process weight table (see ''Selected
Dust Emission Limitation of
Typical Communities in Various
Population Groups" in Appendix)
For compliance to laws, the engineer
must be able to determine the con-
centration of particles being emitted
from an industrial operation under a
variety of specifications including
temperature, pressure, carbon dioxide,
excess air, pounds of flue gas, certain
particle sizes, and efficiency of
collection.
Shortcomings of laws specifying per-
missible concentrations in terms of
grains per cubic foot.
a In industrial problems, dilution with
room or outside air is a possibility
which could produce "permissible"
concentrations for many manufactur-
ing processes without any collection
equipment.
-------
Motivation for Control Measures
b Examples
1) Electric furnances: If the furnace
is ventilated by roof monitors,
wall exhausters, and the like,
an exhaust volume of 100, 000
cfm may be required. If the
furnace has a local exhaust
hooding attached to the roof
ring, metal fumes and smokes
may be exhausted by 20, 000
cfm making collection of
pollutants more practical with
the lower gas volume. Yet
the removal of 80% of the solids
with a collector in the second
case would have emission
concentrations in gr/cu.ft.
the same as the case of the
larger ventilation where no
collection equipment is
employed.
2) Foundry cupola: A foundry cupola
with an open charging door may
emit 15, 000 cfm; one with a
closed top bell charging method,
3, 000 cfm. Pounds of solids
released during the melting
operation will be identical,
pounds of solids collected by
the air cleaning device the
same, as well as pounds of
solids discharged to the atmos-
phere. Yet the grains/cu. ft.
will be 5 times greater in the
top bell charge arrangement
due to the drastically reduced
gas volume involved.
V SELECTION OF COLLECTION
EQUIPMENT
A In the field of air pollution, it is necessary
to consider economical design that will
reduce a health hazard, a damage potential,
(*-> below the threshold of
or a nuisance,
complaint, or
law.
below that established by
It is not necessary to design control
equipment to remove all particles
with complete efficiency unless pro-
fitable recovery of as much valuable
material as possible is the goal.
Following such a concept allows the
air to function to its useful capacity
as a waste disposal medium.
B The present dynamic state of air pollution
abatement appears to lean in the direction
of more stringent control in the future.
This may mean the need of attaining
higher operating efficiencies of control
equipment in the future than is accepted
at present.
1 One safe recommendation in equip-
ment selection is (reference 4): Select
the collector that will allow the least
possible amount of contaminant to
escape and still be reasonable in first
cost and maintenance. However, for
some applications, even the question
of reasonable cost and reasonable
maintenance must be sacrificed to
meet established standards for air
pollution control or to prevent damage
to health or property.
REFERENCES
1 Rogers, S. M. "A Review and Appraisal
of Air Pollution Legislation in the
United States. " Presented at the
Golden Jubilee Meeting, APCA, St.
Louis, Mo. June 4, 1957.
-------
Motivation for Control Measures
2 Air Pollution Abatement Manual. Mano=_. 5 Grindle, A.J. "The Cupola Emission
facturing Chemists Association, Inc. Problem and Its Solution. " Presented
1952. to Semi-Annual Meeting, East Central
Section, APCA, Harrisburg. Sept. 25,
3 Omara, R., and Flodin, C. R. "Engineer- 1953.
ing Design Factors in Dust and Fume
Recovery Systems. " JAPCA 8, No. 1
May 1958.
6 Kane, J. M. "High Temperature Gas Clean-
4 Kane, J. M. "Operation and Effectiveness ing. " Paper presented to Air Pollution
of Dust Collection Equipment. " Heating and Smoke Prevention Association.
and Ventilation. Aug. 1952. Roanoke. May 7, 1951.
-------
2
SECTION 2
Introduction: Collection Equipment
Introduction: The Collection of Particles
from a Gas Stream
-------
INTRODUCTION: COLLECTION EQUIPMENT
I TWO BROAD CATEGORIES OF
COLLECTION EQUIPMENT
A Collection equipment may be divided into
two broad categories:
1 Dry collectors
2 Wet collectors
II DRY COLLECTORS
A The dry collectors fall into the following
classifications:
1 Settling chambers
2 Centrifugal separators
a Dynamic precipitators
b Cyclone
1) Simple cyclone (large diameter)
2) High efficiency (long cone)
3) Multicyclone
3 Inertial separators
a Baffle chamber
b Impingement type
c Louver type
4 Fabric collectors
5 Electrostatic precipitators
III WET COLLECTORS
A The wet collectors fall into the following
classifications
1 Gravity spray tower
2 Wet impingement scrubber
3 Self-induced spray deduster
(orifice type)
4 Disintegrator
5 Wet Dynamic precipitator
6 Venturi scrubber
IV FACTORS IN THE SELECTION AND
DESIGN OF COLLECTION EQUIPMENT
A Carrier Gas Properties
1 Temperature
2 Pressure
3 Humidity
4 Density
5 Viscosity
6 Dewpoint for condensable components
7 Electrical conductivity
8 Corrosiveness
9 Toxicity
10 Flammability
B Particulate Properties
1 Particle size and size distribution
2 Particle shape
3 Particle density (absolute and bulk)
4 Stickiness, build up tendencies, and
flowability
5 Hygroscopic properties
PA. C. pm. 19. 9. 59
-------
Introduction: Collection Equipment
6 Agglomeration tendencies and floe
stability (dispersibility)
7 Electrical conductivity
8 Corrosiveness
9 Flammability
]0 Toxicity
11 Abrasiveness
12 Flowability
C Conditioning
1 The actual deposition efficiency of a
collector may be modified by condition-
ing the carrier gas stream, the partic-
ulates, or the collecting surface.
2 Conditioning of the particle
a Condensation on the particle surface
b Klocculation of the particles
1) Natural
2) Mechanical
3) Electrical
4) Sonic
c Deposition of solids on liquid
droplets
d Treatment of the particle surface
e Electrical charging of the particle
f
g
'j Conditioning of the carrier gas stream
a Heating, cooling
b 11 um id if ieation
4 Conditioning of the collecting surface
a Viscous substances
b Irrigation
c Electrostatic
d Heating, cooling
e
f
D Manufacturing Process Factors
1 Volumetric gas rate collector must
handle
a (Retention time required in the
collector)
b (Velocity through the collector)
2 Particle concentration collector must
handle
3 Permissible pressure drop across
the collector
E Collector Operation Considerations
1 Maintenance
2 Continuity of operation. (Must it be
shut down and started up? Must it
take varying loads? etc.)
3 Safety and health protection
a Toxic hazard
b Explosion and flammability hazard
4 Type of labor required and availability
of such labor
5 Disposal of the collected material
a Waste disposal
b Product recovery
-------
Introduction: Collection Equipment
F Construction and Installation Factors
1 Floorspace requirements
2 Headroom requirements
3 Availability of utilities
a Water
b Steam
e Compressed air
d Electricity (A. C. , D. C. , high
voltage)
4 Auxiliary equipment
a Fans, blowers, compressors
b Pumps
c Motors and drives
d Shaking and rapping devices
e Conveyors, air lacks, rakes,
storage bins, etc.
f Cleanout ports, access doors,
explosion doors, etc.
g Electrical substation or transformer
h Timer, alarms, etc.
i Sludge tanks, treatment tanks,
agitators, etc.
j Valves, dampers, automatic valves,
regulators
5 Materials of construction
a Weather protection requirements
b Insulation or jacketing requirements
c Pressure requirements
d Temperature limitations
e Corrosion resistance
f Erosion resistance
REFERENCES
1 Lapple, C. A. "Dust and Mist Collection, "
Air Pollution Abatement Manual,
Manufacturing Chemists Association,
Inc. 1951.
2 Perry, J. H. Chemical Engineers'
Handbook, McGraw-Hill Book Co.
Inc. N. Y. 1950.
-------
INTRODUCTION: THE COLLECTION OF PARTICLES
FROM A GAS STREAM
I PHASES INVOLVED IN THE COLLECTION
OF PARTICLES FROM A GAS STREAM
A The collection of particles from a gas
stream involves three distinct phases:
I Deposition on a collecting surface
2 Retention on the collecting surface
3 Removal from the collecting surface
II DEPOSITION OF A PARTICLE ON A
COLLECTING SURFACE
A To enable deposition of a particle on a
collecting surface, there is need for:
1 A resultant force upon the particle in
the direction of the collecting surface,
2 A collecting surface upon which the
particle is deposited, and
:', Sufficient time for the particle to reach
the collecting surface before the particle
reaches the outlet of the collecting
device.
B There are six mechanisms by which a
resultant force may be created upon a
particle to cause it to migrate toward
a collecting surface, or cause it to be
directly intercepted.
1 Gravity settling
2 Flow line interception
3 Inertial deposition
4 Diffusional deposition
5 Electrostatic precipitation
6 Thermal precipitation
III RETENTION OF A PARTICLE ON A
COLLECTING SURFACE
A The fact that particles are "deposited"
on a surface is no assurance that they
are "collected".
1 To be "collected", they must remain
on the collecting surface until
intentionally removed.
B The problem of retaining a deposit on a
surface is basically one of having suffi-
ciently high surface forces to counteract
the dislodging tendencies of the fluid
shear of the carrier gas stream as the
gas passes over the deposit.
IV REMOVAL OF A PARTICLE FROM A
COLLECTING SURFACE
A For any collection equipment, some means
must be provided for removing the accu-
mulated deposit, either periodically or
continuously.
Removal of deposited material assumes
outstanding importance in some instances
1 Although deposit removal is usually
purely a problem of mechanical design,
it must be considered in terms of
overall collection efficiency.
I'A. C. pm. 18a. 0. GO
-------
3
SECTION 3
Conversion Factors
-------
CONVERSION FACTORS
Page
TEMPERATURE 2
PRESSURE 3
AREA 4
VOLUME 5
FLOW 6
WEIGHT 7
CONCENTRATION 8
LENGTH 9
EMISSION RATES 10
VELOCITY 11
PA.C.ge.33.12.73
-------
CONVERSION FACTORS - TEMPERATURE
1 Given Units
Degrees
Fahrenheit
Degrees
Centigrade
Degrees
Rankin
Degrees
Kelvin
Desired Units
°F
1. 8°C + 3'2
°R - 460
1. 8(°K-273) + 32
°C
. 5555 x
(°F - 32)
. 5555 x
(°R - 492)
°K - 273
°R
°F + 460
1.8°C + 492
1. 8(°K-273) + 492
°K
5555 x
(°F-32) + 273
°C + 273
. 5555 x
(°R-492) + 273
-------
CONVERSION FACTORS - PRESSURE
Sxs. Desired
^^s^ units
Given ^^N,.
units ^^s«s^
Smm
cm- s e c ^
dyn^s
o m^
^ m
ft-secZ
poundals
ft'2
gn>f
c m 2
*f
*f
m^
A tmosphe res "
gmm
2
c m-sec
1
14. 882
980. 665
478. 80
6. 8948
X 104
1. 0133
X 10B
dynes
cm2
1
14. 882
980. 665
478. 80
6. 8948
X 104
1. 0133
x 10s
*m
ft- sec 2
6. 7197
X 10"?
6. 7197
X ID-2
1
65. 898
32. 174
4. 6331
X 103
6. 8087
X 104
poundals
ft2
6. 7197
X 10"2
6. 7197
X ID'2
1
65. 898
32. 174
4. 6331
X 103
6. 8087
X 104
firnf
cm2
1. 0197
X 10~3
1. 0197
y. io~3
1. 5175
x 10"2
1. 5175
X 10~2
1
4. 8824
X 10"1
70. 307
1. 0332
X 103
»f
ft2
2. 0885
X 10"3
2. 0885
X ID'3
3. 1081
X 10"2
3. 1081
X 10"2
2. 0482
1
144. 00
2. 1 162
X 103
*f
m2
1. 4504
X 10"5
1. 4504
X 10'5
2. 1584
X 10"4
2. 1584
X 10"4
1. 4223
X 10"2
6. 9444
X 10"3
1
14. 696
"Atmospheres"
9. 8692
X 10" '
9. 8692
x io-?
1. 4687
X 10"5
1. 4687
X 10"5
9. 6784
X 10~4
4. 7254
X 10'4
6. 8046
X 10~2
1
To convert a value from a given unit to a desired unit, multiply the given value by the factor opposite the given units
and brneath the desired units.
-------
CONVERSION FACTORS - AREA
en
D
c
0)
J>
O
Square
Inch
Square
Foot
Square
Yard
Square
Mile
Acre
Square
Centimeter
Square
Decimeter
Square
Meter
Squa re
Kilometer
Desired Units
Square
Inch
1
144
1296
40. 144
x 108
62. 73
x 107
15. 5x10-2
15. 5
15. 5 x 102
15. 5 x 10b
Square
Feet
6. 9444
x 10"3
1
9
2. 788
x 107
4. 3560
x 104
10. 764
x ID'4
10. 764
x 10-2
10. 764
10. 764
x 1C6
Square
Yard
77. 1605
x ID"5
0.1111
1
3. 098
x 106
4840
1. 1960
x ID'4
1.1960
x 10-2
1.1960
1.1960
x 106
Square
Mile
2. 49
x 10-1°
3. 587
x 10"8
3. 228
x ID"7
1
15. 625
xlO'4
3. 8610
x 10'11
3. 8610
x 10-9
3.8610
x 10-7
3.8610
x 10-1
Acre
15. 94
x ID'6
2. 296
x lO'5
2. 066
xlO'4
640
1
2. 471
x 10'8
2.471
x 10-6
2.471
x ID"4
2. 471
x 102
Square
Centimeter
6. 452
929. 0341
83.61
x 102
2. 589998
xlQlO
4046. 873
x 104
1
1 x 102
1 x 104
1 x 1010
Square
Decimeter
6. 452
x 10-2
929. 0341
x 10-2
83.61
2. 589998
xlQ8
4046. 873
x 102
1 x lO'2
1
Lx 102
1 x 108
Square
Meter
6. 452
x 10-4
929. 0341
x lO'4
•83. 61
x lO'2
2. 589998
x 106
4046. 873
1 x 10-4
1 x 10-2
1
1 x 106
Square
Kilmeter
6.452
x 10-10
929. 0341
x ID'10
83.61
x 10-8
2. 589998
4046. 873
x 10'6
1 x 10-10
1 x 10-8
1 x 10-6
1
To convert a value from a given unit a desired unit, multiply the given value by the factor opposite the given units
and beneath the desired unit.
-------
CONVERSION FACTORS - VOLUME
^S^Desired
GiveriV^Units
Units ^v>^^
Cubic
Yard
Cubic
Foot
Cubic
Inch
Cubic
Meter
Cubic
Decimeter
Cubic
Centimeter
Liter
Cubic
Yard
1
3. 7037
x 10"
2. 143347
x io"5
1. 30794
1. 3079
x io"3
1. 3079
x io"6
1. 3080
x io"3
Cubic
Foot
27
1
5. 78704
x io"4
35. 314445
3. 5314
x io~2
3.5314
x io"°
3. 5316
x io~2
Cubic
Inch
4.6656
x io4
1728
1
6. 1023
x io4
•61.023
6. 1023
x io"2
61.025
Cubic
Meter
0. 764559
2. 8317
xio-2
1. 63872
x io"5
1
0.001
1 X l(f 6
1. 000027
x io"3
Cubic
Decimeter
764. 559
28. 317
1.63872
x 10"
1000
1
i x io"3
1.000027
Cubic
Centimeter
7.64559
X 10°
2. 8317
X 10
16. 3872
1 X IO6
1000
1
1000. 027
Liter
764. 54
28. 316
1. 63868
X io"2
999. 973
. 99997
9.99973
x 10"
1
To convert a value from a given unit to a desired unit, multiply the given value by the factor opposite the given units
and beneath the desired units.
-------
CONVERSION FACTORS - FLOW
\Desired
^^wUnits
Given ^^w^
Units ^^IN
M3
sec
j^.
mm
M3
hour
_ni
sec
jtL
min
n*
hour
L
sec
L
min
cm.3
sec
cm3
min
«L
SPL
1
0.0167
2.778
<10-5
28.317
x io"3
4.7195
xio"1
7.3658
X 10"6
,.000027
x io"3
,.6667
x io"5
1 < IO'6
1.6667
X io"8
-ill
min
60
1
16.667
x ID'3
1.699
28.317
X 10'3
4. 7195
X10-4
6.00016
XiO-2
1.000027
XKf3
6 X 10~5
-6
1 X 10
M3
hour
3600
60
1
101.94
1.699
28.317
x io'3
3.C
6.00016
-2
X 10
3.6 X 10~3
6'X IO"5
f:3
sec
35.3144
0. 5886
98.90
-4
x 10
i
16.667
X 10
2.778
-4
X 10
35. 316
X 10"
5.886
X 10~4
3.5314
x io"5
5.886
x io'7
ft3
min
21. 1887
X IO2
35.3144
0.5886
60
1
1G.667
X 10~3
2. 11896
35.316
X 10'3
2. 1189
x io"3
0.3531
x io"4
ft_3_
hour
12.7132
xio4
21. 189
X IO2
35.3144
3600
60
1
127. 138
2. 11B96
1.271
X IO"3
2. 11887
X io"3
L
sec
999.973
16.667
27. 777
x io'2
28. 316
47. 193
X 10~2
7.866
X 10"
1
1.6667
X 10~2
9.99973
X 10~4
5.9998
x io"2
L
mm
59.998
X 10
999.973
16.667
16.9896
xio2
28.316
0.4719
60
1
5.9998
_2
X 10
9.99973
X io"4
3
cm
sec
6
1 X 10
16.667
3
X 10
2.777
X 10"
2.8317
xio4
4.7195
X IO2
78.658
1000.027
16.667
1
60
3
cm
min
6 X IO7
6
1 X 10
1.666
4
X 10
1.699
X IO6
2.8317
4.7195
2
X 10
16.667
1000.027
16.667
X IO"3
1
To convert a value from a given unit to a desired unit, multiply the given value by the factor opposite the given units and beneath the desired unit.
-------
CONVERSION FACTORS-WEIGHT
Micro-
gram
Milli-
gram
gram
Kilog ram
grain
Ounce
(avdp)
Pound
(avdp)
Ton
(U.S. shoi
Tonne
(metric)
Desired Units
Micro-
gram
1
1 x 103
1 x 10b
1 x 109
64. 799
x 103
28. 349
x 106
453. 59
x 106
905. 185
-t) x 109
1 x 1012
Milli-
gram
1 x 1Q-J
1 x 10 3
1 x 10fa
64. 799 H
28. 349
x 103
453. 59
x 103
907. 185
x 10b
1 x 109
gram
1 x 10-b
1 x 10-3
1
1 x 103
64. 799
x lO-3
28. 349
453. 59
907. 185
x 103
1 x 10b
Kilo-
gram
1 x 10-9
1 x lO"6
1 x 10- 3
1
64. 799
x lO'6
28. 349
x 10'3
453. 59
x ID"3
907. 185
1 x 103
grain
15. 4124
x 10~6
15. 4324
x lO'3
15. 4324
15. 4324
x 103
1
437. 5
7000
14 x 10b
1. 543 xlO
Ounce
(avdp)
3. 5274
x 10'8
3. 5274
x 10-5
3. 5274
x 10-2
35. 274
22. 857
x ID'4
1
16
3. 2
x 104
3. 5274
x 104
Pound
(avdp)
2. 2046
x 10-9
2. 2046
x 10-b
2. 2046
x 10~3
2. 2046
1. 4286
x ID'4
62. 5
x 10-3
1
2000
2204. 62
Ton
(U. S. shorl
1.1023
x ID"12
1.1023
x 10-9
1. 1023
x lO-6
1. 1023
x ID"3
7. 143
x 10'8
3. 125
x 10-5
5 x 10~4
1
1. 10231
Tonne
») (metric)
1 x 10-12
1 x 10-9
1 x 10-b
1 x lO"3
64. 799
x 10-9
28. 349
x 10-6
453. 59
x 10-6
0. 907185
1
-------
CONVERSION FACTORS - CONCENTRATION
•M
•rH
C
C
ID
>
•H
-------
CONVERSION FACTORS - LENGTH
*S^L>pSln.d
,.. ^"X- Units
Given ^^
Units ^Vw
Infh
l-'fJOl
Yard
Mile
Mic rtjn
MilllmctiT
(.Vntimru.' r
M c [ e r
Kilomctc-r
In<.h
1
12
36
6. 330(1
x ln4
3. 937
X 1(T5
3 . '.) 3 7
x,o-2
3. 937
xio-1
39. 37
3. 037
,,o4
Foot
83. 33
X 10~3
I
3
5280
32. 808
X10-7
32. 808
-4
X 10
32.808
*,o-3
32. 808
X 10~'
32.808
x,o2
Yard
27.778
x 10- 3
3333
1
1760
10.94
xio-7
10.94
xio-4
10.94
X 10'3
10. 94
X lo"1
10.94
X 102
Mile
1.578
x io"5
1. 894
x lo'4
5. 682
-4
X 10
1
62. 137
xio-11
. 62. 137
X 10'8
62. 137
x io"7
62. 137
x io'5
62. 137
-2
X 10
Micron
2. 54
A 10
30.48
X 10
91.44
4
X 10
1. 6094
9
X 10
1
1 X 10
1 X 104
6
1 x 10
9
1X10
Millimeter
25. 4
304. 8
914. 4
1. 6094
« io6
-3
1 X 10
1
10
3
1 < 10
6
1 X 10
Centimeter
2.54
30.48
91.44
1. 6094
X )0
-4
1 X 10
0. 1
1
1 X 10
5
1 < 10
Meter
2. 54
x io~
30.48
x 10"
91.44
X 10"
1 . 6094
x,o3
1 X JO'6
•J
1 X 10
1 X 10~2
1
1 X 10
Kilometer
2. 54
X 10~5
30.48
x io'5
91.44
x io~5
.1. 60.94
IX ,0-9
1 X io"
1 X 10~5
1 X io"
1
To convert a value from a given unit to a desired unit, multiply the given vaJue by the factor opposite the given units and hr neat h the (Jcsj red units.
-------
CONVERSION FACTORS - EMISSION RATES
^^^ Desired
^^•s^ units
Given ^^^(^
units ^^x,.
grns/ sec
gms/min
kg/hr
kg/ day
Ibs/min
Ibs /lu-
tes /day
tons / hr
tons /day
tons/yr
gms/sec
1.0
1. 6667
X 10-2
2. 7778
X 10- '
1. 1574
x'lcr2
7. 5598
1. 2600
X ID" l
5. 2499
X ID"3
2.5199
X IO2
1. 0500
X 10
2. 8766
X 10-2
gms/ min
60.0
1. 0
16. 667
6. 9444
X ID'1
4. 5359
X 102
7. 5598
3. 1499
X ID'1
1.5120
X IO4
6. 2999
X 102
1.7260
kg/hr
3. 6
6. 0
X 10-2
1.0
4. 1667
X ID"2
2. 7215
X 10
4. 5359
x io-i
1. 8900
X ID"2
9.0718
X 102
3. 7799
X 10
1.0356
X 10"1
kg/ day
8. 640
X 10
1. 4400
2. 4000
X 10
1.0
6. 5317
X 102
1.0886
X 10
4.5359
X 10"1
2. 1772
X 104
9.0718
X 102
2.4854
Ibs/min
1. 3228
X ID"1
2. 2046
X 10-3
3.6744
X 10-2
1.5310
X 10-3
1.0
1. 6667
X 10-2
6. 9444
X lO"4
3. 3333
X 10
1. 3889
3.8052
X 10"3
Ibs/hr
7. 9367
1. 3228
x io-i
2. 2046
9. 1860
X lO-2.
60. 0
1.0
4. 1667
X 10-2
2. 0
X 103
8.3333
X 10
2. 2831
X 10'1
Ibs/day
1. 9048
X 102
3. 1747
5. 2911
X 10
2. 2046
1.44
X 103
24.0
1. 0
4. 8000
X 104
2.0
X 103
5.4795
tons/hr
3. 9683
X 1C-3
6. 6139
X 10"5
1. 1023
X 10-3
4. 5930
X 10"5
3. 000
X lO-2
5. 0000
X lO"4
2. 0833
X 10'5
1.0
4. 1667
X 10"2
1. 1416
X 10"4
tons /day
9.5240
X 10"2
1. 5873
X 10"3
2. 6456
x io-2
1. 1023
X 10-3
7. 2000
X ID'1
1. 2000
X ID"2
5. 0000
x io-4
24.0
1.0
2. 7397
X IO"3
tons/yr
3. 4763
X 10
5. 7938
X ID"1
9. 6563
4. 0235
X 1C' !
2. 6280
X IO2
4. 3800
1.8250
X lO'l
8. 7600
X 103
365.0
1.0
To convert a value from a given unit to a desired unit, multiply the given value by the factor opposite the given units and beneath
the desired units.
-------
CONVERSION FACTORS - VELOCITY
"x^^ Desired
^X*XS^ units
Given ^^X^
units ^"""^h
m/ sec
ft/ sec
ft/ rnin
km/hr
mi/hr
knots
mi/ day
ml sec
1. 0
3. 0480
/, 10-1
5. 0080
x 10~3
2. 7778
X ID' J
4. 4707
X 10' l
5. 1479
X 1Q-1
1. 8627
x 10- 2
ft/ sec
3. 2808
1. 0
1. 6667
X 10-2
9. 1 134
X ID"1
1, 4667
1. 6890
6. 1111
X. 10"2
ft/min
1. 9685
X 102
60
1. 0
5. 4681
X 10
88. 0
1. 0134
X 102
3. 6667
km/hr
3. 6
1. 0973
1. 8288
X ID'2
1. 0
1. 6093
1. 8533
6. 7056
X ID'2
mi/hr
2. 2369
6. 8182
X 10'1
1. 1364
X ID'2.
6. 2137
X ID'1
1. 0
1. 1516
4. 1667
x io-2
knots
1. 9425
5. 9209
X 10'1
9. 8681
X 10'3
5. 3959
X 10'1
8. 6839
X 10'1
1. 0
3. 6183
X ID"2
mi/day
5. 3687
X 10
1. 6364
X 10
2. 7273
X ID"1
1. 4913
X 10
24
2. 7637
X 10
1. 0
To convert a
opposite the
value from a given unit
;iven units and beneath
to a desired unit, multiply the given value by the factor
the desired units.
-------
SECTION 4
Gas Properties-Basic Concepts
-------
GAS PROPERTIES-BASIC CONCEPTS
I. EXPRESSION OF GAS-TEMPERATURE
A. The Fahrenheit and Celsius Scales
The range of units on the Fahrenheit
scale between freezing and boiling
is 180; on the Celsius or Centigrade
scale, the range is 100. Therefore,
each Celsius-degree is equal to 9/5
or 1.8 Fahrenheit-degree. The
following relationships convert one
scale to the other:
F - 1.8 C + 32 (I - 1)
(°F 32)/1.8 (I - 2)
C -
where
F - degrees Fahrenheit
C » degrees Celsius or
degrees Centigrade
B. Absolute Temperature
Experiments with perfect gases have
shown that, under constant pressure,
for each change in Fahrenheit-degree
below 32°F the volume of gas changes
1/491.67. Similarly, for each
Celsius-degree, the volume changes
1/273.16. Therefore, if this change
In volume per temperature-degree is
constant, the volume of gas would,
theoretically, become zero at
491.67 Fahrenheit-degrees below 32°F,
or at a reading of -459.67°F. On the
Celsius or Centigrade scale, this
condition occurs at 273.16 Celsius-
degrees below 0°C, or at a temperature
of -273.16°C.
P VAfM J.T
32° F
- 459.6° f
Absolute rero
Cftntigrad*
0°C
491.6
fahrenheit-
degrees
Afeiolut*
491.6° R
273
centigrade -
degrees
0°R
- 373*&
273°K
0°K
Absolute zero
Figure 1 Temperature-Scale relationship
Revised by R.T. Shigehara, Assistant Chief, Engineering Section,
Office of Manpower Development, Institute for Air Pollution Training,
National Air Pollution Control Administration
PA.SS.31b.2.70
-------
Gas Properties - Basic Concepts
Absolute temperatures determined by
using Fahrenheit units are expressed
as degrees Rankine (°R); those deter-
mined by using Celsius units are ex-
pressed as degrees Kelvin (°K). The
following relationships convert one
scale to the other:
3F + 459.67
°K = °C + 273.16
(I - 3)
(I - 4)
Relationship of the various temperature
systems are shown graphically in
Figure 1. The symbol "T" will be
used in thi-s outline to denote
absolute temperature and "t" will be
used to indicate Fahrenheit or Celsius
degrees.
TI. EXPRESSION OF GAS PRESSURE
A. Definition of Pressure
A body may be subjected to three
kinds of stress: shear, compression,
and tension. Fluids are unable to
withstand tensile stress; hence, they
are subject to shear and compression
only. Unit compressive stress in a
fluid is termed pressure and is ex-
pressed as force per unit area
(e.g. Ib./in^ or psi, gm,/cm^) .
Pressure is equal in all directions
at a point within a volume of fluid,
and acts perpendicular to a surface.
B. Barometric Pressure
Barometric pressure and atmospheric
pressure are synonymous. These
pressures are measured with a
barometer and are usually expressed
as inches, or millimeters, of mercury.
Standard barometric pressure is the
average atmospheric pressure at sea
level, 45° north latitude at 35°F.
It is equivalent to a pressure of
14.696 pounds-force per square inch
exerted at the base of a column of
mercury 29.921 inches high. Weather
and altitude are responsible for
barometric pressure variations.
C. Gage Pressure
Measurements of pressure by ordinary
gages are indications of the
difference in pressure above, or below,
that of the atmosphere surrounding the
gage. Gage pressure, then, is
ordinarily the pressure indicated by
the gage itself. If the pressure of
the system is greater than the
pressure prevailing in the atmosphere,
the gage pressure is expressed as a
positive value; if smaller, the gage
pressure is expressed as a negative.
The term, "vacuum," designates a
negative gage pressure.
The abbreviation, "g," is used to
specify a gage pressure. For
example, psig, means pounds-force
per square inch gage pressure.
D. Absolute Pressure
Because gage pressure (which may
be either positive or negative)
is the pressure relative to the
prevailing atmospheric pressure,
the gage pressure, added algebrai-
cally to the prevailing atmospheric
pressure (which is always positive),
provides a value that has a datum
of "absolute zero pressure." A
pressure calculated in this manner
is called absolute pressure". The
mathematical expression is:
where:
abs
abs
atm
g
P + P
atm g
(II - 1)
absolute pressure
= atmospheric pressure
= gage pressure
The abbreviation, "a," is sometimes
used to indicate that the pressure
is absolute. For example, psia,
means pounds per square inch
absolute pressure. The symbol "P"
by itself without the subscript
"abs" will also be used in this
outline to indicate absolute
pressure.
Equation II - 1 allows conversion
of one pressure system to the other.
Relationship of the pressure systems
are shown graphically in Figure
2.1 using two typical gage
pressures, (1) and (2). Gage
pressure (1) is above the datum
from which gage pressures are
measured, and, hence, is expressed
as a positive value; gage pressure
(2) is below the gage pressure
datum, and, therefore, is ex-
pressed as a negative value.
-------
Gas Properties Basic Concepts
<
i
pd)
i
i
0—
(2) '
atm
Gage
Pressure
Datum
" P g(2)
y
i
P(2)
Absolute
Pressure
Datum
Figure 2. 1 Gas-Pressure relationship
E. The Concept of Pressure-Head
Pressure-head is the height of a
column of fluid required to produce
a given pressure at its base.
Fluid of -<
density Pf
T
h
1
Figure 2.2 P-ti relationship
The relationship between pressure and
pressure-head is:
p . Pf(-8-) h (II - 2)
c
where: P - pressure, force/area
p - density of fluid, mass/volume
f
g » local acceleration due to
gravity, length/time^
g = dimensional constant
h ™ pressure-head in terms of
Pf, length
Pressure-head may be expressed in terms
of any fluid that is convenient; e.g.
Hg or H20.
F. Dalton's Law of Partial Pressure
When gases, or vapors (having no chemical
interaction) are present as a mixture
in a given space, the pressure exerted
by a component of the gas-mixture at a
given temperature is the same as it
would exert if it filled the whole
space alone. The pressure exerted by
one component of a gas-mixture is called
its partial pressure. The total pressure
of the gas-mixture is the sum of the
partial pressures.
III. THE LAW OF IDEAL GASES
A. The Laws of Boyle and Charles
1. Boyle's Law
Boyle's Law states that, when
the temperature (T) is held
constant, the volume (V) of a
given mass of a perfect gas of
a given composition varies in-
versely as the absolute
pressure, i.e.:
where: <* - proportional to
Charles ' Law
Charles' Law states that, when
the volume is held constant,
the absolute pressure of a
given mass of a perfect gas of
a given composition varies
directly as the absolute
temperature, l.e:
a T
at constant V
-------
Gas Properties - Basic Concepts
B. The Law for Ideal Gases
Both Boyle's and Charles' Law are
satisfied in the following equation:
where:
PV
mRT
(III - 1)
P
V
m
M
T
R
absolute pressure
volume of a gas
mass of a gas
molecular weight of a gas
absolute temperature
universal gas-constant
The units of R depend upon the
units of measurement used in the
equation. Some useful values are:
1.
1544 (ft)
2. 21.83
3. 554.6
(Ib -mole) ( R)
(in. Hg) (ft3)
~(lb -mole) (°R)
™ 1
(mm Hg) (ftj)
lib -mole) (°R)
m
In the above units of R:
3
V
m
ft~
Ib
m
M = Ib /Ib -mole
„ m m
T = °R
P = lbf/ft2 for (1)
in.Hg
mm Hg1
for (2)
for (3)
Any value of R can be obtained
by utilizing the fact, with approp-
riate conversion factors, that there
are 22.414 liters per gm -mole
1 "
or 359 ft per Ib -mole at 32°F
and 29.92 in. Hg.™ Problems to
illustrate this will be shown later
in this discussion.
IV. CALCULATION OF APPARENT MOLECULAR
WEIGHT OF GAS MIXTURES
Utilizing Dalton's law of partial
pressure and the ideal gas law, the
following equation can be derived for
calculating the apparent molecular
weight of a gas mixture:
VI.
M ,
mix "
where: ^mix.
xMx (IV-1)
apparent molecular
weight of a gas-mixture
proportion by volume
of a gas—component
molecular weight of a
gas component
In all other equations (except where
specifically noted), the symbol "M" will
be used to denote the molecular weight of
a pure gas or a gas-mixture.
V. GAS DENSITY
Gas density can be determined by re-
arranging equation III - 1 and letting
density p » r:
p - l_^ (V - 1)
R T
where: P » density
P - absolute pressure
M • molecular weight
T - absolute temperature
R =• universal gas constant
Another method of determining density is
by utilizing the fact that there are
22.414 liters per gm-mole or 359 ft3
per Ib -mole at 32°F and 29.92 in. Hg.
Ib
(V 2)
Ib
M
Ib -mole
m
ft"
359
ff
Ib -mole
m
492° R
'P in. Hg
T° R
29.92 in. Hg
In this equation, M- Ib /Ib -mole,
mm
T-R°, P-in. Hg, and p-lb /ft3.
m
VISCOSITY
A. Origin and Definition of Viscosity
Viscosity is the result of two
phenomena: (1) intermolecular
cohesive forces, and (2) momentum
transfer between flowing strata
caused by molecular agitation
perpendicular to the direction of
motion. Between adjacent strata of
-------
Gas Properties - Basic Concepts
• •• ••••••••••••••••••••••»• ••••••••••••^••••B • •••
^ f I t
Figure 3
a flowing fluid a shearing stress
occurs which is directly pro-
portional to the velocity gradient
(Figure 3)
dv
where: g = dimensional constant
7 unit shearing stress between
adjacent layers of fluid
dv
civ
velocity gradient
,1 proportionality constant
(v Lscosity)
The proportionality constant, \i, is
called the coefficient of viscosity,
or merely, viscosity. It should be
noted that the pressure does not
appear in equation VI - 1, in-
dicating that the shear (T) and the
viscosity (\>) are independent of
pressure. (Viscosity actually in-
creases very slightly with pressure
but this variation is negligible
in most engineering problems.)^
B. Kinematic Viscosity
Kinematic viscosity is defined
,!i-cording to the following re-
iati onship:
(VI -2)
v = kinematic viscosity
y = viscosity of the gas
p density of the gas
Note the absence of dimensions of force.
C. Liquid Viscosity Versus Gas Viscosity.
1. Liquid Viscosity
In a liquid, transfer of momentum
between strata having a relative
velocity is small compared to the
cohesive forces between the
molecules. Hence, shear stress T
is predominantly the result of
intermolecular cohesion. Because
forces of cohesion decrease rapidly
with an increase in temperature,
the shear stress decreases with an
increase in temperature. Equation
VI - 1 shows that shear stress is
directly proportional to the
viscosity. Therefore, liquid
viscosity decreases when the
temperature increases.
2. Gas Viscosity
In a gas, the molecules are too far
apart for intermolecular cohesion to
be effective: Thus, shear stress
is predominantly the result of an
exchange of momentum between flowing
strata caused by molecular activity.
Because molecular activity increases
with temperature increases, the shear
stress increases with a rise in the
temperature. Therefore, gas
viscosity is increased when the
temperature increases.
D. Determination of Viscosity of Gases
The viscosity of a gas for prevailing
conditions may be found accurately from
the following formula:
(VI - 3)
where: y = viscosity prevailing
y = viscosity at 0 C and
prevailing pressure
T = absolute prevailing
temperature (°K)
n = an empirical exponent
(n = 0.768 for air)
-------
Gas Properties Basic Concepts
Viscosity of air at 1 Atmosphere *
Temperature
Degree Degree
Centigrade Fahrenheit
-100 —i
— -100
0 —
100 .
200
300 —
400 •
500 •
600 •
700
800 '
900
1000-
—200
300
400
-500
— 600
-700
I—800
-900
-1000
-1100
-1200
-1300
-1400
-1500
-1600
-1700
-1800
Viscosity
Centipoises
— 0.1
- 0.09
- e.08
-0.07
: 0.06
i-0.05
:-0.04
-0.03
\>
-0.02
-0.01
- 0.009
-0.008
-0.007
r o.oo6
-0.005
(l)centipoise
(10)"2 gm
cm-sec
(lO) poise
2. 09 (10) "5
'b - sec
2
ft
2. 09 (10) "5
slug
ft - sec
6.72(lO)"4
Ib
m
ft - sec
*l'erry, J.H. Chemical Engineer's Handbook, McGraw-Hill Book Co., New York, 1950.
Figure 4
-------
Gas Properties - Basic Concepts
.036
.034
.032
.030
.028
.026
.024
.022
.020
Q.
u
- .018
*/)
8 -016
>
.014
.012
.010
.008
.006
.004
.002
-CO,
H20.
CH4
100
200
300
400
500
600
700
Temperature, "F
Figure 5 Viscosity at 1 Atmosphere
-------
Gas Properties Basic Concepts
Table 1 Ratios of Specific Heats of gases at 1 Atm. Pressure*
Compound Formula Temperature Ratio of
°C Specific
Heats
Acetaldehyde
Acetic acid
Acetylene
Air
Ammonia
Argon
Benzene
Bromine
Carbon dioxide
disulfide
monox Ldc
Ch lori ne
Ch lo reform
C2H40
C2H402
C2H2
NH3
A
C&H6
Br2
CO 2
CS2
CO
C12
CIIC13
30
136
15
-71
925
17
-78
-118
15
15
-180
0-100
90
20-350
15
-75
100
15
-180
15
100
1.14
1.15
1.26
1.31
1.36
1.403
1.408
1.415
1.310
1.668
1.76(?)
1.67
1.10
1.32
1.304
1 .37
1.21
1.404
1.41
1.355
1.15
Compound Formula
Cyanogen
Cyclohexane
Dichloro-
difluorme thane
Ethane
Ethyl alcohol
ether
Ethylene
Helium
Hexane (n-)
Hydrogen
bromide
chloride
(CM) 2
C6H12
CC12F2
C2H6
C2H60
C4H100
C2H4
He
C6H14
H2
HBr
HC1
Temperature Ratio of
°C Specific
Heats
15
80
25
100
15
-82
90
35
80
100
15
-91
-180
80
15
-76
-181
20
15
100
1.256
1.08
1.139
1.19
1.22
1.28
1.13
1.08
1.086
1.18
1.255
1.35
1.660
1.08
1.45
1.453
1.597
1.42
1.41
1.40
The viscosity of air and other gases at
various temperatures and at a pressure
of 1 atmosphere may be taken from
Figures A and 5.
VII. SPECIFIC HEAT
A. Definition of Specific Heat
The specific heat of a. gas is the
amount of heat required to change
the temperature of a unit-mass of
gas one temperature-degree. Units
of specific heat are, therefore,
(Btu/lb ) (°F) or calories/(gm )
(°C) depending upon the dimensional
system used.
Heat may be added while the volume
or pressure of the gas remains
constant. Hence, there may be two
values of specific heat: (1) the
specific heat at constant volume
(C ), and (2) the specific heat
atvconstant pressure (c ).
P
"Perry, -Mi. Chemical Engineer's Handbook, McGraw-Hill Book Co. New York. 1950.
-------
Gns Properties - Basic Concepts
Table 1 Ratios of Specific Heats of gases at 1 A Cm. Pressure
Compound Formula Temperature Ratio of
°C Specific
Heats
1
Hydrogen (continued)
1
cyanide
iodide
sul f i do
1 odi ne
I sobutane
Krypton
Mercury
Methane
Methyl acetate
;l 1 collO I
other
HCN
III
H2S
12
V'lO
Kr
Hg
cn4
C_,H602
CII40
C2II60
65
140
210
20-100
15
-45
-57
185
15
19
360
600
300
15
-80
-115
15
77
6-30
1.31
1.28
1.24
1.40
1 .32
1 .30
1.29
1 .30
1.11
1.68
1.67
1.113
1.16
1.31
1.34
1.41
1.14
1 . 203
1.11
Compound Formula Temperature Ratio of
°C Specific
Heats
Methylal
Neon
Nitric oxide
Nitrogen
Nitrous oxide
Oxygen
Pentane (n-)
Phosphorus
Potassium
Sodium
Sulfur dioxide
Xenon
C3H8°2
Ne
NO
N2
N20
°2
C5H12
P
K
Na
SO 2
Xe
13
40
19
15
-45
-80
15
-181
100
15
-30
-70
15
-76
-181
86
300
850
750-920
15
19
1.06
1.09
1.64
1 .400
1 .39
1.38
1 .404
1.47
1 .28
1.303
1.31
1.34
1.401
1.415
1.45
1.086
1.17
1.77
1.68
1.29
1 .66
Because the heat energy added at
constant pressure is used in raising
Lhf temperature and doing work
against the pressure as expansion
takes place, C is greater than
C P
v
Specific Heat Ratio
Specific heat ratio (R) is defined as:
C
k = —5L- (VII - 1)
Specific heat ratios for e,ases are
shown in Table 1.
C. Determination of Specific Heat for a
Gas-Mixture
The specific heat for a mixture of gases may
be calculated from:
P(mix)
v (mix)
BxCp(x) (VII 2)
Bx Cv(x) (VII 3)
-------
Gas Properties Basic Concepts
where:
"p (mix)
C
v (mix)
"p U)
(x)
specific heat at con-
stant pressure for gas-
mixture
specific heat at con-
stant volume for the
gas-mixture
proportion by volume of
a gas-component
specific heat at con-
stant pressure for a
gas-component
specific heat at con-
stant volume for a gas-
component
For ordinary temperature (for example, about
80° F as experienced at the metering device in
atmospheric or source sampling work) the
specific heats may be assumed to be constant.
VIII. REYNOLD'S NUMBER
A. Definition
A typical inertial force per unit
volume of fluid is p v ,
a typical viscous force per unit
volume of fluid is y v
The first expression divided by the
second provides the dimensionless
ratio known as Reynold's Number:
Re
inertial force
viscous force
(VIII - 1)
where: p
Re
density of the fluid (mass/
volume)
velocity of the fluid
dimensional constant
a linear dimension
viscosity of the fluid
Reynold's Number
The linear dimension, L, is a length
characteristic of the flow system.
It is equal to four times the mean
hydraulic radius, which is the cross-
sectional area divided by the wetted
perimeter. Thus for a circular pipe
L = diameter of the pipe; for a
particle settling in a fluid medium,
L = diameter of the particle; for a
rectangular duct, L - twice the
length times the width divided by the
sum; and for an anulus such as a
rotameter system, L - outer diameter
minus the inner diameter.
B. Laminar and Turbulent Flow
1 Laminar Flow
In laminar flow, the fluid is con-
strained to motion in layers
(or laminae) by the action of
viscosity. The layers of fluid
move in parallel paths that re-
main distinct from one another; any
agitation is of a molecular nature
only. Laminar flow occurs when
Reynolds' Number for circular
pipes is less than 2000 and less
than 0.1 for particles settling
in a fluid medium.
2 Turbulent Flow
In turbulent flow, the fluid is not
restricted to parallel paths but
moves forward in a haphazard
manner. Fully turbulent flow
occurs when Reynolds' Number is
greater than 2500 for circular
pipes and greater than 1000 for
settling particles.
IX. Summary of Useful Equations
A.
Temperature
°F
°R
°K
°R
where:
1.8 °C
+ 32
C
i
1 .
46
273
The larger the Reynold's Number, the
smaller is the effect of viscous
forces; the smaller the Reynold's
Number, the greater the effect of
the viscous forces.
a °v
o K
degrees Fahrenheit
degrees Centigrade
or Celsius
degrees Ranklne
degrees Kelvin
10
-------
(las Properties Hasic Concepts
B. Pressure
P
F + P
abs atm g
where:
1 std atm « 14.696 Ih /in
E 2316.22'. Ib /ft2
3 29.921 in. Hg
a 760 mm Hg
P = pressure
p - density
h - pressure head or height
g = gravitational acceleration
g *= dimensional constant
Subcripts
abs = absolute
atm atmosphere
g gage
f fluid
C. Ideal Gas Law
PV
m RT
M
154A (1V(ft)
(Ib -mole) ( k)
v m
21.83
(in. Hg )(ft )
Ub"
= 554 (ram Hg)(ft )
(Ib -mole)( R)
m
1 Ib -mole - 359 ft3 at 32°F and 29.92 in. Hg
1 urn -mole - 22.414 liters at 0°C and
m
760mm Hg
where P = absolute pressure
V = volume
m = mass
M = molecular weight
R = gas constant
T = absolute temperature
D. Apparent Molecular Weight
1MB
x x
molecular weight
B = proportion by volumi
Subscript
mix = gas mixture
x = component
.
mix
where: M
P
(359
where: p density
P = absolute pressure
R = gas constant
T = absolute temperature
F Viscosity, h
,-4
1 cp s 6.72 X 10 " Ib
m
ft sec.
r,. Revnold'--; '
Re
L v f,
whe re : N
Ke
L
Reynold's 'number
= linear dimension
D for circular pipe
= D for snhericjL pn
11
-------
Gas Properties - Basic Concepts
= D-D, for a rotameter
V = velocity
P, - fluid density
Mf = fluid viscosity
D = diameter
D = tube diameter
DC = float diameter
D = particle diameter
REFERENCES:
1. J.K. Uennard, Elementary Fluid
Mechanics, John Wiley and Sons,
Inc., New York (1947).
2. M.B. Lemon and M. Ference,
Analytical Experimental Physics,
The University of Chicago Press,
Chicago (1946).
3. J.H. Perry, Chemical Engineers
Handbook, McGraw-Hill Book Co.,
Inc., New York (1950).
REVIEW PROBLEMS - BASIC CONCEPT
Problem 1. In source sampling or in
evaluating control equipment it is
essential to know the volumetric flow rate
within a ductwork to enable the calculation
of pollutant emission rates or to properly
size the control equipment. The
volumetric flow is usually determined by
first measuring the velocity of the gas
stream. A pitot tube is probably the
most commonly used device for this purpose.
However, the pitot tube does not measure
velocity directly; it measures velocity
pressure or velocity head, usually in
inches of water. The velocity head is then
related to velocity by the following
equation.
Fluid of
density p.
T
h
1
I
P
P-h relationship
where: v
C
P
g
h
velocity
pitot tube coefficient
acceleration due to gravity
velocity head in terms of
the flowing fluid
a. The velocity head Is in terms of
the flowing fluid. Thus if air is flowing,
h Is the height of air that would be
supported by the velocity pressure or kinetic
energy of the flowing air.
Using C - 1.0 (empirically determined),
g = 32.2 ft/§ec2, density of the flowing
fluid - 0.075 lbm/ft3, and Ap (velocity head
) = 1.0 in. H.O, calculate v in
b. In the above problem la, the density
of the flowing fluid was given. This
normally must be determined by taking
measurements of P, T, and M and using the
equation p - PM. Assume that your stack
RT
pressure reads in in. Hg and T in °K.
Calculate an appropriate value of R (use
Ib
the dimension — — B^- — for M) .
-mole '
—
ID
Answer:
39.30 <%?
(lbm-mole) (°K)
in in, HO
ft/sec.
Answer: 66.8
c. In order to determine the molecular
weight of a gas mixture, its composition
must be known. An orsat .analyzer is
commonly used for this purpose, especially
for combustion effluents. If the pro-
portions by volume on a dry basis are:
12
-------
Gas Properties Basic Concepts
CO, = .10 M = 44
0, <= .08 M - 32
M = 28
Problem 2
Calculate the apparent molecular weight
(assume B of H^O = 0.10, M = 18).
Answer: 28.72 m
Ib -mole
m
d. If t - 77 C, P =31 in. Hg,
stack pressure P = -1 in. Hg, and M =
&
28.72 Ib calculate the density. Check
m
Ib -mole
m
your results using equation V-2.
Answer: .0626 Ib
m
Reynold's number, N
Re'
is a dimension-
less number. It is the ratio of the inertial
force to the viscous force in a flow system.
The importance of this number lies in the
area of predicting the behavior of particles
in a fluid medium or in determining flow
characteristics.
The small particles are of particular
interest in air pollution work. The rate
at which a particle settles due to gravity
or migrates due to electrostatic or
centrifugal forces is a function of N .
If N <0.1,J the particle is said to be
following Stoke's Law, i.e.
2
v = gD (p - p ) for gravitational settling.
P P P P
where: v settling velocity
P
g acceleration due to gravity
D particle diameter
e. Utilizing all the previous in-
formation, rewrite the relationship between
v and the pitot tube measurements in terms
of actual measurements taken.
particle density
fluid density
fluid viscosity
For the following case:
Pf = .075
Ih
ft"
Answer:
g
RT Ap
PM
Ib
ft
, P= 29.92 in. Hg
p = 128
P
M= 29 m/ m-mole; what is the largest
size particle that would settle in Stoke's
Law range? ( 1 ft = 30.48 * 10* n)
Answer: 27y
13
-------
5
SECTION 5
Particle Settling Dynamics
Terminal Velocities of Spherical
Particle
-------
PARTICLE SETTLING DYNAMICS
David R. Hemenwav
i NTRODUCTJ ON
Tho force of gravity, or attraction between
masses . was first described in analytical terms
by Sir Isaac Newton. This force has since been
recognised as being essentially constant for
attraction of a mass to earth - if it is less
than 1 mile above the earth's surface. The
distance from the center of the earth to the
equator is 3,063 miles. The gravitational ac-
celeration for this distance is 32.09 ft./sec."
The distance from Che center of the earth to the
North Pole is 3,050 miles with.,a gravitational
acceleration of 32.26 ft./sec.'. The accelera-
tion on .1 particle therefore is, for all practi-
cal purposes, constant regardless of its location
because of the very small variations in gravi-
tational attraction. Thus, the acceleration
value normally used is 32.2 ft./sec.-.
DESCRIPTION OK _FORCES
The weight of a particle is described by
the following equation:
Weight - V * P *
P P
u.oi
where :
is the
V is the volume of the particle,
p t p
density of the particle, g^ is a gravitational
constant, and g is the gravitational acceleration,
which Is considered to be constant as described
.ibove .
The second force that can be observed is
the effect of buoyancy. For example, when a
block of wood is placed in a pail of water it
floats becau.se it is supported by the buoyant
force of the water. If a piece of iron is
placed In the water it will sink to the bottom
of the pail because the buoyant force of the
water is not sufficient to cause it to float.
However, the piece of iron does not weigh as
much in the pail of water as it does in air
(see Figure 1).
IRONH
Figure 1.
The loss of weight (buoyant force) is equal
the weight of the volume of displaced vater.
For objects that are totally immersed in water
(or any other fluid) this loss of weight can
be calculated by the following equation:
where:
Buovant Force - V * Pf * g
P f T
p is the density of tha displaced fluid
or suspending medium.
Mr. Hemenwav is an Environmental Engineer with the
National Air Pollution Control Administration
Ai.pir. 100. 11.69
-------
Particle Settling Dynamics
When a particle or mass is placed in a
fluid, the above two forces (buoyancy and weight)
will produce a resultant force. If the buoyant
force is greater than the weight of the mass,
then the mass will begin to rise in the fluid
(see Figure 2). In the case of normal parti-
culate matter suspended in the atmosphere, the
weight of a particle usually exceeds the buoyant
force exerted on it, and the particle settles.
buoyancy
buoyancy
wei g ht
weighs
Figure 2.
The resultant force on the particle will accel-
erate it in the direction of the force, and in
so doing friction will be created between the
fluid and the particle. Thus, three forces are
exerted on a body placed in any fluid medium.
These forces are friction, buoyancy and weight
(see Figure 3).
DEFINITION OF TERMS
A = The projected cross-sectional area
of a particle
C = Drag Coefficient
D = Diameter of a spherical particle
P
p = Terming.! Settling Velocity
g = local acceleration due to gravity
g = dimensional constant
K = Cunningham Correction Factor for
discontinuous fluids
R = Reynold's Number
V = Volume of a particle
P = Particle Density
P = Fluid Density
Fluid Viscosity
All units are dependent on the system
of measurement used.
Figure 3.
DERIVATION OF SETTLING VELOCITY EQUATION
In the process of passing through a fluid,
a particle will cleave, or displace, the fluid
immediately in front of the particle, imparting
momentum to the fluid. In the process of pro-
ducing momentum, the force produced can be des-
cribed by Newton's Law of momentum, which states
that the force produced is equal to the pro-
duction of momentum per unit time.
The amount of momentum produced in the
fluid is, then, a function of the fluid density
(a measure of the fluid mass), the maximum vel-
ocity to which the fluid is accelerated by the
action of the moving particle, and the volume
of fluid that is equal to that of a cylinder
containing a cross-sectional area equal to that
of the particle and a length equal to the dis-
tance traveled by the particle in a unit time
(see Figure 4).
-------
Particle Settling Dynamics
volume
of fluid
displaced
Figure 4.
The force of momentum, commonly called
frictional force, is described by Newton's law
of momentum: proportional to the factors
affecting the production of momentum. Therefore,
the frictional force can be formulated as
follows:
where:
C' is the friction factor for correlating
U with f . However, the equation for friction
force is usually written in the following form:
C_ * P, * A * f 2
Friction Force D f p
where:
(5.0)
C will, by definition, be different than
C' by a factor of 2. Both C' and C are related
to the Reynold's Number because it is a measure
of the flow pattern surrounding the particle.
The friction force, as shown by equation
5.0, will increase as the square of the velocity
of the particle increases. As the particle ac-
celerates, it will reach the terminal settling
velocity where the buoyancy and friction forces
will equal the weight of the particle. The
velocity will remain constant at this point
because there will be no resultant force to
accelerate the particle. At the point of e-
qualization of forces, equations 1.0, 2.0,
and 5.0 may be grouped as follows:
Weight buoyant force + friction force
V * P * g
P P „
v * P *
p f
(6._0)
^2
Frictional Force
where:
P * A *
f * U
P
(3.0)
A is the cross-sectional area of the part-
icle perpendicular to the direction of movement,
f is the velocity of the particle, and U is the
P
maximum velocity to which the fluid is acceler-
ated in the process of being displaced by the
particle.
The velocity, U, to which the fluid is ac-
celerated will be directly related to the vel-
ocity of the particle, f , by a factor that
depends upon the flow pattern surrounding the
particle. Thus, the equation for friction force
can be written as:
Friction Force
* A * f'
(4.0)
By cancellation of the g terms equation 6.0
can be rewritten as follows: (6.1)
V *
P
(f )'
p
;- Vp*Pf
2 * V *
P
A *
y
A*,
(P -
_E
(6.2)
Equation 6.2 will be valid under the following
conditions:
1. When the fluid is continuous with res-
pect to the particle and the particles
are larger than 1.0 micron. Particles
smaller than 1.0 micron are able to
slip between the molecules of the fluid.
Because the friction force is caused
by bombardment of the particle by the
-------
Particle Settling Dynamics
molecules of the fluid, the friction
force becomes intermittent as the par-
ticle slips between the molecules. The
particle, therefore, increases in vel-
ocity as it slips between the molecules.
The actual average velocity will there-
fore be higher than predicted by the
equation, where the friction force is
assumed to be constant.
2. When the particle falls in a free state,
or when it is not hindered by other
particles in the surrounding fluid.
If the particle concentration is great
enough to cause additional friction
forces (from bombardment by other
particles and the resulting transfer
of momentum), the settling velocity
will decrease.
3. When the movement of the particle is
not slowed by hitting a boundary. The
fluid is then assumed to be infinite
in extent.
Equation 6.2 can be further simplified wher
particles under investigation are assumed
to be spherical. Therefore, the projected
area of the particle (A), and the volume
of the particle (v ) may be given by the
P
following equations:
(7.0)
(7.1)
where:
D is the diameter of the sphere.
Substituting equations 7.0 and 7.1 will
reduce equation 6.2 to:
and
2 * TT * D
(8.0)
/4 * D *
E_
1/2
(8.2)
* CD * Pf
As previously stated, the friction factor
or drag coefficient, C , in equation 8.2 depends
upon the flow pattern surrounding the particle.
This flow pattern can be characterized by the
Reynold's Number^( see figure 5 ). For example,
in air, a Reynold's Number of more than 1000
indicates the turbulent flow range while the
laminar flow range occurs at a Reynold's
Number of less than 1.0.
.laminar flow
region
turbulent _|
flow region
Reynolds number, Re
Relation between the drag coefficient (log10Cn)
and the Reynolds number (logioRe' Ior spheres.
Figure 5.
Many authors have attempted to write equations
to describe the variation of the drag coef-
ficient with the Reynold's Number. For example,
Langmuir and Blodgett wrote the following equa-
tion to describe the drag coefficient for
Reynold's Number ranging from 1 to 100
(see figure 6):
(9.0)
J24
R
1 + 0.197R °'63 + 0.0026R
e i
l.38\
- )
4 * D *
P
3 *
(8.1)
For a Reynold's Number ranging from 1000 to
200,000 the Drag Coefficient is relatively
constant; it varies only in the range from
0.38 to 0.50. This region is usually referred
to as the Newton Region and an average Drag
Coefficient of 0.44 is often used in calcul-
ations (see Figure 7).
-------
Particle Settling Dvns
10
Q 10 4
- 10 3
c
3 10
*S 10
u
01
0 1.0
Q
10 '
\
V
\
V
\
V
\
•• I -
I,^N TKV
^TO'j'W
Jss»vJ '"
?*>„
* •*" 4
*^**r.
-—
X
Thus, for a Reynold's Numbei
200,000, equation 8.2 become
f = 1.74 g DP (PP Pf>
Pf '
The flow region of greatest
cous flow region occurring
1/2
(10.0)
id4 10 10 10' i.o 10 io2 io3 10* ioJ io6
Reynolds number Re
Relation between the drag coefficient (log10CD)
and the Reynolds number (logioRe) for spheres.
Figure 6.
of less than 0.1. For this region, Stokes cal-
culated the Drag Coefficient from the following
formula:
(11.0)
where
io5
Q 10 "
U
3
^ 10
1 10'
**-.
0 10
U
I? i.o
Q
10 '
newtonian
region
\
\
\
s
\
\
\
\
\
\
^^
' 1 ", "I "
-:fi-
^»JIj
^LJ
^^i^wW^ftiy
; -• -. |-":- 1 A \^"\
,-,:..J,:l'-i...-jb.\ J
-4-3-2-1 2345
6
Reynolds number , Re
Relation between the drag coefficient
and the Reynolds number (logj0Re) for spheres.
Figure 7.
(ii.D
and
JJ is the viscosity of the fluid.
which, when substituted into equation 8.2,
reduces the settling equation to the familiar
"Stokes Law" equation:
18
(P _
P
(11.2)
For particles smaller than 1.0 micron, the fluid
will become discontinuous, resulting in higher
settling velocities than can be predicted by
Stokes' Law. To correct for this effect,
Cunningham deduced that the Drag Coefficient
should be written as:
K * R
m e
(12.0)
-------
Particle Settling Dynamics
wnere K is the Cunningham correction factor to
account for particle slippage.
Thus, Stokes' Law becomes
(12.1)
TABLE 1. K FOR AIR AT ATMOSPHERIC PRESSURE
m
18,
Table 1 shows the effect of temperatures and
particle diameters on the Cunningham correction
coefficient.
Particle
diameter
(microns)
0.1
0.25
0.5
1.0
2.5
5.0
10.0
Cunningham
factor ( Km
70°F
2.88
1.682
1.325
1.160
1.064
1.032
1.016
correction
)
212°F
3.61
1.952
1.446
1.217
1.087
1.043
1.022
500°F
5.14
2.528
1.711
1.338
1.133
1.067
1.033
BIBLIOGRAPHY
(1) Lapple, C.E. in J.H. Perry Edition;
Chemical Engineer's Handbook, 3rd Edition,
pg. 1021, McGraw-Hill, New York (1950).
(2) Langmuir, I. & Blodgett, K., American Air
Force Tech. Report #5418 (1946).
(3) Strauss, W., Industrial Gas Cleaning,
Pg. 122, Pergamon Press, 1966~.
(4) Torobin, L.B. and Guavin, W.H. Canad. J.
Chem. Engineering
Volume 37 Pg. 129, 167, 224 1959
Volume 38 - Pg. 142, 189 1960
Volume 39 - Pg. 113 1961
(5) Stern, A.C., Editor, Air Pollution, 2nd
Edition, Academic Press, 1968.
Volume I - Pg. 58
Volume II - Pg. 270
(6) Brown, George G., etal. Unit Operations,
Pg. 70, John Wiley & Sons, Inc. New York,
1950.
(7) Lapple, C.E., Fluid and Particle Mechanics,
Pg. 281, University of Delaware, Newark,
Delaware, March 1956.
(8) Orr, C., Jr. & Dallavalle, J.M., Fine
Particle Measurement, McMillan, 1959.
-------
Equivalent Standard
Terminal Velocities of
Spherical Particles
From: Lapple, C.E., "Fluid and 1Q2
Particle Mechanics", University
of Delaware, 1956
101
10°
-a
c
o
Terminal velocities of spherical g
particles of different density S 10"'
settling in air and water at a>
70 degrees fahrenheit under the ^
action of gravity. •"-
o
1 Iff2
DO
j— »
CO
1 lO'3
6
£
Notes :
1. Numbers on curves represent -5
true (not bulk or apparent)
specific gravity of particles
2. Stokes-Cunningham correction
Theoretical
Screen Mesh
§ § §
• 0
i44-
10 O C
(N CN O Is
^ -0 T C-
1
.
0 /
4^
::" r-.Zt"
•-^A
!T
/ /<
!//
' /
, /
/i
! ft
• i-/y!
i / /
1 />
if
'//
- \—f- — ,
' r/
i / /
t \//
< fc=
LI J j j
//N
^r
0 t ^
-' / ^
b' 0
33 )
f * I
77 '
r/ /
///
J^-i
f~ ' ,
I2^t^
/A'
r—f--t
//
TTfl
N — -
0 0
O if)
i !
} C
- a
8
J
[ /
/
'' ^f. >
!;'//
£---
' •<
;
w&\
f /
''v*
o /
^
/
^
•0 CD *o <
•o -* rt c
I 1
^
.' t
J{j - 1
f J j
/ .'
T~T
A £
<$*!
£/ ,
f / (
/ / 1
///
f/[/\
/J
7
v* r
i r~~t~
' ' f
f i / 1
> //
X
'
c
n o ^r
N CN •-
N
O
1
*£,
1 ^^X
' ' ^^
[P
T
.X
/;
>
' j^
, x1
'-• ^^
LK]/
j
CD « ~*
j 1 1
^
^^ * '
y
S*s"'
f
*
S*'
~^ — *^
. r
r t*
r *
(?
n CN
CO
1
d '
• '
, j
,
„,
-;
U-flj
1 '
( '
! 10'
10'
10'
r,-2
10
,-4
10-
particles settling in air.
3. Physical properties used:
10 100 1,000
Particle diameter,microns
10'
10,000
Fluid
Air
Water
Temp.
F
70
70
Viscosity
centipoise
0.0181
0.981
Density
Ib . mass ,.<_
cu. ft.
0.0749
62.3
7
PA.C. pm.40b.2.70
-------
SECTTON e
The Separation of Particles into
Size-Fractions
-------
THE SEPARATION OF PARTICLES INTO SIZE-FRACTIONS
I Accurate and rapid methods of analysis of
particles in a sample consist of:
A Dividing the sample into closely-sized
fractions, followed by
B The determination of weight, size, or
count in each of the fractions
II Processes commonly used for separation
of particles into size-fractions are:
A Screening (or sieving) for grading "coarse"
particles
H Procedures based on the motion of particles
in a fluid for grading of "fine" particles
1 Klutriation procedures
2 Sedimentation procedures
III SIEVING
A Procedure
1 In the process of sieving, the particles
are passed through a series of vibrating
screens, the openings of which get
progressively smaller.
B Screen Six.e
1 There are two series of screen -scale
.sieve.s commonly used in Hie llmled
State's at present. These- are the U. S.
and Tyler, the si/.e.s of which are shown
in Table 1.
C Lower Particle-Size Limit
1 Particles are graded down to a fineness
of 74-microns (Tyler or U. S. 200 mesh)
and may be extended to a lower limit of
43 or 44 microns (Tyler and U. S. 325
mesh).
D Need for Specifying the Type of Analysis
Used
1 Because of the wide variation possible
in making the actual test, and of
variations in particle shape, densities,
and properties of the material (such as
hygroscopicity and stickiness) it is
impossible to specify a single process
universally applicable to all particles.
2 Sieving to completeness is impossible.
Hence, it is necessary to define the
endpoint by one of the following methods:
a By specifying a standard time of
sieving
1) This may be good when there is
no great variation in fineness.
b By specifying that screening must
be continued until the weight of the
particles passing the sieve per
minute is less than a stated
percentage of the total weight of
sample taken.
1) This is complicated. It is
unsatisfactory for routine
work, but is good practice
for control equipment acceptance
tests.
c B> stating that sieving must be
continued until the weight of material
passing the sieve per minute is less
than a certain percentage of ihe weight
of the residue on the sieve considered.
1) This is very sound practice, but
extremely complicated.
3 Since there is no single screening
procedure adaptable to all particles,
and wide variations in methods used are
inevitable, the results of any sieve-
analysis should be accompanied by a
description of the procedure followed in
making the size-fractions.
PA.C.pm. 65. 9. 60
-------
Separation of Particles into Size-Fractions
Table 1. STANDARD U.S. AND TYLER SCREEN SCALES
U.S.
Sortn
Salt,
Meth
Number
400
325
270
230
200
170
140
120
J»X)
80
70
60
50
45
40
35
30
25
20
13
16
14
12
10
8
7
6
5
4
I ncket
0.0015
0.0017
0.0021
0.0024
0.0029
0.0035
0.0041
0.0049
0.0059
0.0070
0.0083
0.0098
0.0117
0.0138
0.0165
0.0197
0.0232
0.0280
0.0331
0.0394
0.0469
0.0555
0.0661
00787
0.0937
0.111
0.132
0.157
0.187
Nominal
Aperotnrt
Width
Mieront
37
14
63
62
74
88
105
125
149
177
210
250
297
350
420
600
590
710
840
1,000
1,190
1,410
1,680
2.000
2,380
2,830
3.360
4,000
4,760
Win
Diameter,
IndM
0.0010
0.0014
0,0016
0.0018
0.0021
0.0025
0.0029
0.0034
0.0040
0.0047
0.0055
0.0064
0.0074
0.0087
0.0098
0.0114
0.0130
0.0146
0.0166
0.0189
0.0213
0.0240
0.0272
0.0299
0.0331
0.036
0.040
0.044
0.060
Tyler
Serin
Seal*.
Uetk
Number
400
826
270
250
200
170
150
115
100
80
65
60
48
42
35
32
28
24
20
16
14
12
10
9
8
7
6
5
4
/neJUf
0.0015
0.0017
0.0021
0.0024
0.0029
0.0035
0.0041
0.0049
0.0068
0.0069
0.0082
0.0097
0.0116
0.0138
0.0164
0.0195
0.0282
0.0276
0.0828
0.0390
0.046
0.056
0.066
0.078
0.093
0.110
0.131
0.166
0.186
Nominal
Apcrature
Width
Miermt
37
43
63
61
74
88
104
124
147
176
208
246
295
351
417
495
589
701
883
991
1,168
1,897
1,661
1,981
2,862
2,794
3,827
8,962
4,699
Win
Oiamttrr.
Indut
0.0010
0.0014
0.0016
0.0016
0.0021
0.0024
0.0026
0.0038
0.0042
0.0066
0.0072
0.0070
0.0092
0.0100
0.0122
0.0118
0.0126
0.0141
0.0172
0.0286
0.026
0.028
0.086
0.033
0.082
0.0328
0.086
0.044
0.066
IV AIR ELUTRIATION
A Definition
1 Air elutriation is the process of
separating particles into size-fractions
by varying the velocity of an upward
current of air.
B Application
1 Size-separation by air elutriation is
desirable when the particles to be
analyzed are, in practice, subjected
to the grading action of air flow.
C Upper Particle-Size Limit
1 Air elutriation methods are employed
for the analysis of particles of sub-
sieve size (less than 80 or 40 microns).
D Principle of Particle-Size Separation
1 The upward velocity of air required
to just separate a given particle-size
is calculated from Stokes' Law. A
convenient form of the law is expressed
in the following:
-------
Separation of Particles into Size-Fractions
D2g
-jti
y
for N
<0.1
(i)
Where:
upward velocity of air (cm/sec)
size of particle to be removed
(microns)
local acceleration due to gravity
(cm/sec^)
density of the particle (gm/cm )
absolute viscosity of the air
(poises)
E The Roller Analyzer (Figure 1)
1 Description
Four vertical tubes, D-l, D-2, D-3,
and D-4, are mounted so that they may
be rotated about a central post thus
bringing the tubes successively into
position over U-tube, C.
U-tube C is constructed of heavy-walled
Pyrex tubing, 1-inch inside diameter.
Tubes D-l, D-2, D-3, and D-4 are
stainless steel and highly polished on
the inside. The inside diameters are
respectively 9.0, 4.5, 2.25, and 1.125
inches (these may vary according to
design). All tubes are grounded
electrically.
Gooseneck, E, is -^-inch stainless steel
tubing, polished on the inside. At the
end of E is a machined rubber stopper
on which is fitted an extraction thimble,
F. which collects the particles blown
over from the vertical tube to which it
is connected.
Tho elulrlalor conilili of the following
principal parti 01 shown In lh« drawing.
;yAi':'Alr Inlef
-J'l^'Cornftrettot
, p-4, •\Yf'f Cha^«j',||
• "...MI T u ' £•' Glasi Ooo»«ri«?l(ft
il U Tuba i '-. , .....,..'.' ittif^j,
;°" Chamber ' P) PpP?r.^^^'l^
ff Chambef . v O. Rotary fappiriftM
. • . ' ' i ' l./i-ti .. ' ." •' '> ..Vi '.I- "li'.tfllM
Roller-Analyzer
2 Operation
A weighed quantity of particles (25 grams)
is placed in U-tube, C. An air jet from
Figure 1
The Thermix Corporation, Greenwich, Conn.
-------
Separation of Particles into Size-Fractions
motal noz/.le, IB, impinges on the
hoi torn-center of U-tubc, C. Nozzles
of desired sizes are installed to provide
proper dispersing action for the size-
fraction being separated. The flow rate
is measured by a flowmeter attached to
air-hose, A.
The rubber-tipped hammer head, G-l,
actuated by a motor-driven cam, keeps
U-tube, C, constantly agitated. The
mass of particles in U-tube, C, moves
as a whole in a direction opposite to
that of the air in the jet. The mass of
particles also.rotates in a clockwise
direction so that fresh portions of the
mass are continuously exposed to the
action of the air jet.
Tapper, G-2, affects the rate of separa-
tion and also the uniformity of fractiona-
tion by shaking down into U-tube, C,
those particles that repose in the lower
cones of the vertical tubes.
The lowest size-fraction is separated
first; and, for each size-fraction a new
thimble is used. Before each run, the
air is conditioned to constant humidity
(the air is drawn through 1:1 sulfuric
acid solution) and blown through the
thimble until the thimble reaches
constant weight.
The analyzer must be calibrated to
determine the endpoint of each run.
V SEDIMENTATION IN LIQUIDS
A Definition
1 Sedimentation is the process of
separating particles into size-fractions
according to their velocity of fall in a
column of liquid at rest.
B Upper Particle-Size Limit
1 Sedimentation methods are employed for
the analysis of particles of sub-sieve
size (less than 80 or 40 microns).
C Principle of Particle-Size Separation
1 The velocity of fall of a given particle
in a liquid under known conditions is
calculated by using Stokes' Law. A
convenient form of the law is expressed
in the following:
-8
p(S)
for N <0.1 (2)
Where:
p(S)
D
vertical velocity of the
particle (cm/sec)
local acceleration due to
gravity (cm/sec^)
density of the particle
(gm/cm3)
density of the liquid (gm/cm3)
size of the particle (microns)
absolute viscosity of the
liquid (poise)
D A Sedimentation Method (Reference 5)
1 Prepare assemblies as in Figures 2, 3,
and 4.
2 Determine the viscosity and density of
the liquid, and the density of the
particles to be analyzed.
3 Calculate the time of fall over a vertical
distance of 6-cm for (0-10) and (10-20)
micron fractions (Figure 3); and the
time of fall over a vertical distance of
20-cm for fractions larger than 20-
microns (Figure 4).
4 For the (0-10) micron fraction, place
one-half gram of the sample in the
300 ml beaker shown in Figure 3. Fill
the beaker to the upper mark with the
liquid and stir well to obtain good
dispersion.
-------
Separation of Particles into Size-Fractions
1
Equipment for Determination of Particle Size by Liquid Sedimentation
Figure 2
Western Precipitation Corp. , Los Angeles, California
2 so rt I,
5uc-i-lon Y\ & rgcJuafet
I \Cylfi de. r
-J_J 1
\ -Post-Lion of 300ml.
Figure 3 Beaker
Ti
Figure 4
m*
-------
Separation of Particles into Size-Fractions
5 Stop stirring and immediately start the
timer.
6 At the end of the time calculated in
Step 3 for the (0-10) micron fraction,
raise the beaker so that the tube
connected to the filter flask is properly
placed at the lower mark on the beaker.
Rapidly draw off the liquid to the lower
mark. The particles drawn off are
smaller than 10 microns.
7 Filter the liquid drawn off through a
Gooch crucible.
8 Without emptying the beaker, fill it
once again with liquid to the upper
mark. Stir well.
9 Repeat steps 4, 5, 6, and 7 until the
quantity of minus 10-micron particles
remaining in the beaker is negligible
(15 to 25 decantations, or more, may
be necessary).
10 Dry the material collected on the filter,
and weigh. This weight represents the
(0-10) micron fraction.
11 Repeat steps 3 through 9 for the (10-20)
micron fractions using assembly shown
in Figure 3.
12 Repeat steps 3 through 9 for the (20-44)
micron fraction using assembly shown
in Figure 4.
1M Material remaining after the (20-44)
micron fraction is termed residue,
weighed, and reported as the plus 44
micron fraction.
14 AfU'r weighing the crucibles containing
the dust fractions, take portions from
each crucible and examine under a
microscope to observe the conformity
of the "actual" size-fractions to the
"calculated" size-fractions. Also
examine the fractions for agglomerates
and other irregularities.
VI NEED FOR SPECIFYING THE METHOD
OF PARTICLE -SIZE ANALYSIS
The size of a given particle determined by
one method of analysis will not necessarily
be duplicated for that same particle using
another method of analysis.
For example: In the sieving analysis, any
particle regardless of its shape, that will
pass a 325 mesh (Tyler) sieve is reported
as having a size of 44-microns because the
sieve apertures are squares averaging
44-microns on a side. But, if this same
particle were subjected to air elutriation
or sedimentation in a liquid, it does not
necessarily follow that the size of 44-microns
will be determined. Let's say that the 44-
micron particle as measured by the sieve
is a flat disc. This same particle, then, will
settle slowly in a sedimentation method of
analysis, and according to Stokes1 Law will
have a size much smaller than 44-microns.
Hence, in reporting particle-sizes, it is
necessary to specify the method of analysis.
VII A BASIC DEFECT OF PARTICLE-SIZING
TECHNIQUES (References 7, 8)
A The Dispersion of Particles in the Flue and
in the Particle-Sizing Apparatus
1 In a sample of particulate matter
extracted from a flue, the original
state of dispersion of particulates in
the carrier gas is permanently destroyed.
2 During the particle-sizing procedure,
the particles are redispersed. How-
ever, there is no certainty that the
original dispersion that prevailed in
the flue is duplicated in the particle-
sizing apparatus. In fact, a re-creation
of the original dispersion is very
unlikely, especially when the fine
particles are considered.
-------
Separation of Particles into Size-Fractions
B Examples of Particle Aggregation and
Dispersion
1 Figure 5 illustrates the problems of
redispersion of a dust sample.
A. Loose, Porous Aggregate
O
°
o
°
O
" O
o
o o o
O O o (-.
a °
B. Closely Packed Aggregate
C. Small particles adhering to a large one
modification of effective size when
prepared for sizing by liquid sedi-
mentation procedure.
c Figure 5C depicts a large particle to
which smaller ones are adhering.
During the dispersing action in
particle-sizing apparatus, the smaller
particles are detached.
d Figure 5D illustrates the air sizing
method in which particles originally
well dispersed are imperfectly
redispcrsed.
SELECTED REFERENCES
1 Roller, P.S. U.S. Bureau of Mines
Technical Paper No. 490. 1931.
2 Roller, P.S. Ind. Eng. Chem. (Anal. Ed.)
3; 212-16. 1931.
3 Roller, P.S. Measurement of Particle
Size With an Accurate Air Analyzer.
Proc. Am. Soc. TejEaingJWtls_. 35th
Part II,
Annual Meeting 1932.
Technical Papers, p.
3_2:
607.
O
D. Well dispersed particles
Figure 5
a Figure 5A represents a single
aggregate, somewhat loose and
porous as it might exist in a flue.
When deflocculated in preparation for
microscopic sizing or liquid sedimen-
tation, the density and effective size
arc quite different from the original.
I) Kiguro r>H shows a closely packed
aggregate which may undergo drastic
4 The Problem Characteristics Industrial
Dust. The Thermix Corporation.
Greenwich, Conn.
5 Particle jize Analysis. Bulletin No.
G402R. Western Precipitation Corp.
Los Angeles, Calif.
6 Traxler, R. N. and Baum, L. A. H.
Determination of Particle Size
Distribution in Mineral Powders by
Air Elutriation. Rock Products.
June, 1934.
7 Hemeon, W. C. L. , Raines, G.F., Jr.,
and Puntereri, S. D. Rating of Dust
Collectors According to Dust Settling
Velocities. Paper 60-54, APCA, 53rd
Annual Meeting. Cincinnati, Ohio.
May 26, 1960.
8 Hemeon, W. C. L. Dust Particle Inertia
and Various Consequences. ASHRAE
Journal Section, Heat Pip, and Air
Cond. 247-250. ~Feh>r~uary, 1957.
-------
7
SECTION 7
Notes on the Analyses of Particle
Size Distributions
-------
(Excerpts)
NOTES ON THE ANALYSES OF PARTICLE
SIZE DISTRIBUTIONS
Richard Dennis
GCA CORPORATION
GCA TECHNOLOGY DIVISION
Pollution Control Laboratory
Bedford, Massachusetts
May 1972
PA.C.pm.102.4.73
-------
SECTION 1
MATHEMATICAL REPRESENTAILON OF PARIICLE SIZE*
The simplest representation of a particle size distribution is a
size-frtquency curve which, shows the number of particles, N, present
for an\ specified diameter, D. Since most dusts are composed of an in-
finite range of particle sizes, it is first necessary to classify par-
ticles according to some consistent pattern- Then N may be defined as
che number of particles within a specified size group having finite
boundaries and typified by some average diameter, D. For most particle
distributions, a characteristically skewed curve results. Figure 1, It
is convenient to graph the fraction of the total number of particles, P
rather than the absolute number of particles, N, within a size range
>- n)
CD (X
Q-i •-
O •-
O
•fl
D
1
Par: ic If Diame-- r - D
Figure 1 Typical size frequency distribution
~Matnematical derivations of sizt parameters in Sections 1 and 2 arc
based largely on analytical methods described by W E Ran/ in Tech
Rpt No 1, University of Illinois Engineering Experiment Station,
April 30, 1950 USAEC SO-1000,
-------
to facilitate the derivation of descriptive size parameters. When no
mathematical relation between P and D has been stated, the general
equation for the curve in Figure 1 is
P = f(D)
The fraction of particles, dF, within the size range, dD, is ex-
pressed as
dF = f(D> dD
Bv introduction cf the proper constant (included in the term f(D)i
the value 1 0 can be assigned to the cumulative fraction of all parti-
cles within the distr ibut. too form D = 0 t o D = *• On this basis, the
toral area under the curve represents 100 percent of the particles,
i e., the fraction F = 1.0. Similarly, the area between any arbitrary
size limits D and D depicts the fraction of the total number of par
t ic les witr>in tnat range Thus
CO
f-^
i f(D> dD = 1
v-'
o
and
n
^ 2
/ ^
I f. (D i dD = fraction in range D to D7
X
The cumulative distribution, which indicates the fraction of the
total number of particles in the diameter range 0 to D, appears as
-------
D
r
' f(D) dD (i)
In all cases where the surface area and volume of uniformly
shaped particles (spherical or non-spherical) are proportional to D
3
and D , respectively, Equation (1) forms the basis for defining the
following diameters:
(1) D = number average particle diameter, the particle diameter
which multiplied by the total number of particles will give the sum
of all the particle diameters.
(2) D - surface average particle diameter, the diameter of the
s —
particle whose surface multiplied by the total number of particles
will give the total surface of all the particles.
(3) D - volume average particle diameter, the diameter of the
particle whose volume multiplied by the total number of particles
will give the total volume of all the particles.
3 2
(4) D = D /D = diameter of the particle with the same ratio
vs v s
of volume to surface as that exhibited by all particles in a given
s amp 1e.
In addition to mean diameters, other cumulative distribution
functions can be defined:
Cumulative surface area distribution function = the fraction of
the total surface of all the particles contributed by particles with
diameters from D = 0 to D = D.
-------
D
F(D2) = 1/D? / D2f(D) dD
Cumulative volume distribution function = the fraction of the
total volume of all the particles contributed by particles with diam-
eters from D = 0 to D = D.
D
3 3 P 3
F(D ) - 1/D / D f(D) dD (3)
^ U
o
Median Diameters
(1) D , or M = number or count median diameter, the particle
nmd g r
diameter where half the total number of particles have diameters greater
than and half have diameters less than D
nmd'
D
ouna
r
nmd
/ f(D) dD - / f(D) dD - | (4)
o D
(2) Dm or M = mass median diameter, the particle diameter
where one half the mass of the total number of particles is represented
by particle diameters less than or greater than D ,.
mmd
mmd oo
/ D3f(D) dD = / D3f(D) dD = ^ (5)
\J \J £
o D ,
mmd
Since the size distributions outlined in the previous section have
been outlined in general form, it is necessary to find some mathematical
relationship between P and D if solutions other than by graphical
-------
integration are desired. Several curve fitting methods have been pro-
(1 2)
posed including the commonly used Hatch-Choate equations ' which
have been derived in the next section. Other size distribution func-
tions have been described by Rosin-Rammler, Roller, Nukiyama-
Tanasawa, ' Gaudin-Schulhmann/ ^ Wynn-Dawes, Sichel, ' Kottler ,
and Dalla Valle. The criterion for selecting one relation in pref-
erence to another should be the goodness of fit to experimental measure
ments. All techniques are the same in principle, that is, the term
f(D) is usually defined by D (or function of D) and two arbitrary con-
stants, the first, some average diameter and the second, a measure of
homogeneity (range of sizes). If the Hatch-Choate equations are used,
average diameter may be represented by M and the homogeneity factor
o
byo-g
-------
ROLLER
The Roller distribution appears in the form:
w = a d exp(— T)
where w is the weight percent of the particles possessing diameters less
than d, and again, a and b are constants. By differentiation, the weight
and number-frequency distributions appear respectively as:
dw
= a
and:
d(d)-aV2 II
d4 d2
d n a / 1 . b \ / b
,,.. = - 1 T + —q~ ) exP \ ~ T
d(d) p\2 Z. 2. I \ d
d2 d2
where p = particle density.
A graph of the function ln(w/d^) versus -7, which should appear as a
straight line, permits estimation of the constants a and b.
NUKIYAMA-TANASAWA
The Nukiyama-Tanasawa distribution is often applied for the analysis
of atomized liquids where an extremely broad size range may be encountered
Three constants, a, b, and c, must be used in this system. In equation
form:
5a
-------
Here, n refers to the number of particles within the diameter range A d.
Proper selection of the constant c, which does not change significantly
for a given nozzle regardless of flow rate, allows for a graphical solu-
tion to Equation when In ( —^ ) ( --^-r ) is plotted against d . Again,
valid results obtain only when the plot is linear.
WYNN-DAUES
The Wynn-Dawes relationship has been applied to the analyses of mine
dusts in which mixtures of dust from several sources may be found. As a
result, bi-modal distributions frequently occur.
— = Q a exp(- a D) + (l-Q)p exp(- p D)
In the above equation,, n refers to the number of particles in the
size range represented by diameter D, and Q, a and P are constants.
Specifically, Q is the fraction of dust characterized by the constant Q
and 1-Q the fraction characterized by (3. If Q is equal to 1.0 (implying
that the dust arises from a single source and probably can be described
by a uni-modal system), the simplified equation results, viz:
d n /• rxi
- = a exp(- a D)
5b
-------
ROSIN-RAMMLER
The Rosin-Rammler equation may provide a convenient means of norma-
lizing size distribution measurements when the cut-off points for the
smallest and largest particles present are well defined, e.g., the frac-
tion of particles retained between two screen sizes. The weight percent
of particles, w, having a diameter greater than d is expressed as:
w = 100 exp (- ad )
where a and b are characteristic constants for the distribution. The
weight-frequency distribution obtained from the above equation is:
= - 100 a b d^-^expC- adb)
The constant a increases with decreasing particle size and the con-
stant b is somewhat analogous to standard deviation. By rearrangement of
Equation , viz:
In In (—) = In a + b In d
w
graphing of experimental measurements on log-log paper gives b as the
slope of the straight line of best fit. The value for a is then obtained
by calculation
5c
-------
TABLE 1
TABULATED SIZING DATA FOR
100 PARTICLE MEASUREMENTS
o>
Portion
Size
1
2
3
4
5
6
7
8
9
10
(2)
Diameter
Microns
0.44
0.62
0.88
1.25
1.76
2.50
3.53
5.00
7.07
10.0
(3)
Number of
Particles
19
29
18
12
7
8
2
4
0
1
(4)
Cumulative
Number and
Percent
19
48
66
78
85
93
95
99
99
100
(5)
Cumulative
Number
119
148
166
178
185
193
195
199
199
200
(6)
Cumulative
Percent
59.5
74.0
83.0
89.0
92.5
96.5
97.5
99.5
99.5
100
Curve 1, Figure 2 Curve 5, Figure 2
15
-------
10.0
CO
z:
o
(T
O
I
.80
)PE
— —
Ds 55-
1.57 -^
1.49 |||
1.62 ^ji
Dv
2.15
2.43
- 1 ^
.11 i . . i
—
^ir
• • •
He tU-.-.
US it
_ t
1 I
\ \ > '
— — — — — -
-—••-•••••• '
-H-rr i
ftj=t 4 i : ' r
---)
t j
t P
J >
^9
p==3 iu_^-_ .
^Fl
i rrhr-
i TTTF
p
1 'j ^
ir
'^\f M
fy I
2
1 L 2
tj-L
^
T TT-; :
'IT4
1 ! ' ,
•LliT^
~ ' "T" —
fflf
tJr\\
j j j
mm
jSr
Jf
71
'
T
J^
p— ^
-4—
fe^'-J
^^
^^5 •"::::
^
p^
^t±t
f — — r~
. — r->_
— > 1 I i
! 1 ' : i
*M
m-r n^
rp:: —
r—
1
=J
j
1
1 j
i
Jt^
— 1
^ 1 |.^
—
— -
,
^
—
LLi_l 1
rrr
TT
3»
M-p
;i i
i i i
: !
i ' '
. _
— — r
5V
T— t
~A
L
- 1
^
i
1
1= ---,
____
rr?: t i — ! — TT=
±~
^
4 —
T-1—
0.01
1.0
10
30 50 70
90
99
99.99
PER CENT LESS THAN OR EQUAL TO STATED SIZE
Figure 2. Particle Size Distributions from Table 1.
-------
TABLE 3
DATA TABULA!CON FOP PARTICLE SIZING BY TRUNCA1KD
MULTIPLE TRAVERSING SYSTEM*"
1 2
Tr -i verse
Number 0.25 .44
.44 .62
1 19 29
2
3
4
5
6
7
8
9
10
Tonal 19 29
Sum ~?r per
Traverse 19 29
Cum a lac i ve
Percent 19 48
Porton Size and Size Range in Microns
345678
.62 .88 1.25 1.76 2.50 3 53
.88 1 25 1.76 2.50 3-53 5.00
18 12 78 2 4
11 5 4 1
3 1
5 2
0
2
18 12 18 13 14 10
18 12 9 6.5 3-5 16
66 78 87 93 5 97.0 98.6
9
5.00
7.07
0
1
1
0
2
1
3
2
10
1.2
99.8
10
7.07
10.0
1
0
0
0
1
0
1
0
0
0
3
0.3
100. 0+
Total
Number
100
122
127
134
137
140
144
146
146
146
See Reference 8 - Section 1-
-------
TABLE U
NUMBER OF PARTICLES TO BE COUNTED TO
ACHIEVE ANY GIVEN ACCURACY
Weight % of Particles
in Any Size Range 2 5 10 15 20
(m)
Expected Accuracy No. of Particles to Be Counted
(s) (m)
2%
1%
0.52
0.2%
0.1%
(3)*
(6)
16
100
1+00
(8)
25
100
625
2,500
25
100
1+00
2,500
10,000
56
225
900
5,600
22,500
100
1+00
1,600
10,000
Uo.ooo
31
-------
01B127-430H
O»
200
RATIO Ds/Mg , Dv /Mg , Dsmd /Mg , MgV Mg
500 IOOC
Figure 6. Graphical estimation of characteristic particle diameters (D6, Dv, Dsrad and M ')
from count median diameter (Mg) and geometric standard deviation (
-------
REFERENCES
1. Hatch, T. and Choate, S.P., "Statistical Description of the Size
Properties of Non-Uniform Particulate Substances," J. Franklin
Inst 207. 369 (1929).
2. Drinker, P and Hatch, T., "Industrial Dusts", 2nd Ed., Ch. 10.
McGraw-Hill Book Co., New York (1954).
3. Rosin, P. and Rammler, E., Z. Verden Ing. 71. 1 (1927); J. Inst.
Fuel I, 29 (1933).
4. Roller, P.S., "Law of Size Distribution and Statistical Descrip-
tion of Particulate Materials," J. Franklin Inst. 223. 609 (1937).
5. Nukiyama, S. and Tanasawa, Y., Trans. Soc. Mech. Engrs. (Japan)
5, 63 (1939).
6. Gaudin, A.M. and Hukki, R.T., "Principles of Communition Size and
Surface Distribution," Min. Tech. Nov. T-P 1779 (1944).
7 Wynn, A.H.A. and Dawes, J.G., "The Size Classification of Airborne
Dusts in Mines," S.M.R.E. Res. Rpt. No. 28 (South Africa) (1951).
8. Sic>
-------
REFERENCES (cont.)
15. Smith, J.G. and Duncan, A.J., "Sampling Statistics and Applications,"
McGraw-Hill Book Co , Inc., New York (1945).
16 First, M.W and Silverman, L., "Air Sampling with Membrane Filters,"
Arch. Ind. Hyg. & Occ. Med. T_, 1 (1953).
17 Millipore Filter Corporation, Bed-ford, Massachusetts, "Bibliography,
A Reference Listing of Published Information Concerning Applications
of the Milltpore Filter," March, 1961.
18- Chamot, E.M. and Mason, C.W., "Handbook of Chemical Microscopy,"
Vol I, 3rd Ed., John Wiley & Sons, Inc., New York (1958)
19. Whitby, K.T., "Determination of Particle Size Distribution Apparatus
and Techniques for Flour Mill Dust," Univ. of Minn. Eng. Expt. Sta-
tion Bull No. 32 (January, 1950).
20. Dennis, R. , Johnson, G.A., First, M.W. and Silverman, L., Performance
of Commercial Dust Collectors - Report of Field Tests," U.S.A.E.G.
Report No NYO-1588, Harvard University (November 2, 1953).
21 Corn. M. , "Statistical Reliability of Particle Size Distributions
Determined by Microscopic Techniques," AIHAJ 26, 8 (1965).
22 Dixon, W J and Massey, F.J., Jr., "Introduction to Statistical
Analysis," 2nd Ed., p. 291, McGraw-Hill, New York (1957).
2J Fatrs, G.L., "Developments in the Technique of Particle Size Analysis
by Microscopical Examination," J. Roy. Micros. Soc. 71. 209 (1951)
36
-------
8
SECTION 8
The Effective Particle Size
-------
THE EFFECTIVE PARTICLE SIZE
All particles arr not spheres. For example,
liquid droplets and solid particles produced '
by condensation are usually spherical, but
solid particles generated by comminution are
non-spherical and irregular in shape (subse-
quent abrasion may, however, produce well-
rounded particles).
The theories applied to particle-size
measurement by air elutriation and sedimen-
tation in liquids are derived for spherical
particles. Therefore, the size of a non-
spherical particle calculated by use of these
equations expresses the diameter of a sphere
that behaves in a fluid as the particle in
question in the range of Stokes1 Law.
Only for spherical particles will the "sizes"
measured by the different methods of analysis
theoretically always be the same. The
following demonstrates the significance of
the "size" of non-spherical particles in
measuring devices and particle-collection
equipment.
I THE EQUIVALENT DIAMETERS OF A
NON-SPHERICAL PARTICLE
The equivalent diameters of a non-spherical
particle are defined as follows:
Symbol
6
A
Dd
Name
Volume
diameter
Surface
diameter
Drag
diameter
Definition
Diameter of a sphere
having the same volume
as the particle in question
Diameter of a sphere
having the same surface area
as the particle in question
Diameter of a sphere
having the same resistance
to motion as the particle in
question
Remarks: In the case of irregular particles,
the sizes may differ, not only according
to the method of measurement, but also
according to the orientation of the particle.
Of the basic diameters, only the volume
diameter (6) is independent of the
orientation of the particle. But, in
industrial operations as well as in methods
of particle-size measurement, the particles
can often be assumed to be randomly
oriented. Hence, the mean value of the
projected areas (A ) and the mean of the
resistances to motion (drag) can be taken
in such cases. Also, since the mean
projected area (Ap) of a particle is a
constant fraction ('/) of the surface area
(Cauchy's theorem), the surface area
rather than the projected area is used
as one of the basic properties.
It is important to note that there is no
experimental evidence (Ref. 1) that the
drag of an irregularly shaped particle
is proportional to the relative velocity
and viscosity of the fluid, even under
conditions where Stokes' Law is valid.
Only if this is true will the drag diameter
(D^) of a given particle have the same
numerical value in different fluids and at
different velocities. However, the
resistance of an ellipsoid is theoretically
proportional to the viscosity and velocity
(Lamb 1932) and this is likely to be the
case also for other non-spherical shapes.
II THE EFFECTIVE PARTICLE-SIZE
IN IMPINGEMENT, CENTRIFUGAL,
SEDIMENTATION, AND ELUTRIATION
MECHANISMS IN WHICH STOKES1 LAW
APPLIES
A The free-body diagram at terminal
velocity.
PA. C. pm. G3. 9, GO
-------
The Effective Particle Size
Buoyancy —g-
P g
Drag
Weight - p g -^
1 The driving force:
Driving Force p g
TT 6
6
3
P g
P g
2 The equilibrium equation:
The particle settles at a constant
velocity when the driving force equals
the drag.
Sir (j. f D ,
r p(s) d
(P
P(s)
III
(P P) g D
18 When NRe <°'1 (1)
Where:
The size (D )
P
It is the numerical value of (Dp) that is
calculated from equation (1) when using
particle-size measurement methods in
which Stokes' Law is valid. It is seen
that (Dp) consists of two equivalent
diameters: the volume diameter (6)
and the drag diameter (D^). Hence in
processes where the velocity of a
particle through a fluid depends on a
driving force that is proportional to the
volume of the particle, (Dp) as
determined by Stokes' Law is the
effective size.
The effective size (Dp) as determined
by air elutriation and liquid sedimenta-
tion procedures is the operative size
in filtration, impingement, and
centrifugal apparatus in which particles
are removed from a gas stream by
forcing their precipitation upon a surface.
THE EFFECTIVE PARTICLE SIZE IN
THERMAL PRECIPITATION
D
(2)
A thermal gradient provides a driving force
upon a particle suspended in a gas (the
particles are driven away from a hot body).
The theory of driving force (Ref. 1) due to a
thermal gradient indicates that the force
depends on the projected area of the particle
(or, since the particles are normally in
random orientation, the force depends on
-------
The Effective Particle Size
the surface; area). The velocity of a particle
moving in a thermal Hold would therefore be
dependent upon the size ( ^ ) However
inadequate the: theory of thermal force may
be, the important point to note is that the
effective particle size in thermal precipita-
tion is not necessarily equivalent to the size
determined by analyses employing Stokes'
Law (equation 1) when non-spherical particles
are involved.
V THl'J KKKECTIVE PARTICLE SIZE IN
DIKI'HJSIONAL PR I-XJLPITATION
Particles are driven toward regions of lower
particle-concentration as a result of their
Brownian movement (the particles are
bombarded by the molecules of fluid). The
driving force depends only on the concentra-
tion of particles; hence, the effective size
is simply the drag diameter (Dcj).
tV THE EFFECTIVE PARTICLE SIZE IN
ELECTROSTATIC FIELD
Electrically charged particles may be driven
through a fluid by applying an electrostatic
field. The driving force depends on the
charge of the particle. If the particle does
not have the maximum charge it is capable
of carrying, the velocity is dependent only
on the drag diameter (Djj). The maximum
charge that a particle can carry is propor-
tional to its surface area. Therefore, the
velocity of a fully charged particle in an
electrostatic field depends on the size [
It should be noted that the effective size in
electrostatic precipitation is not necessarily
Dp as det-ermined by Stokes' Law (equation 1).
REFERENCE
1 Hawksley, P.O. The Physics of Particle
Size Measurement: Part I. Fluid
Dynamics and the Stokes' Diameter.
The British Coal Utilization Research
Assoc. Monthly Bulletin.
NoT 4. April, 1951.
Vol XV.
-------
9
SECTION 9
Representation of Particle-Size Data
Statistical Presentation of Data
Size-Efficiency Curves
-------
REPRESENTATION OF PARTICLE-SIZE DATA
METHODS FOR EXPRESSING THE
RESULTS OF PARTICLE-SIZE
MEASUREMENTS
Distribution of particle-size in a mixture
of particulate material is commonly rep-
resented by plotting either, or both, of
the following curves:
1 A frequency distribution curve
2 A cumulative distribution curve
II THE FREQUENCY DISTRIBUTION CURVE
A Examples of Tabulation of Data
1 The tabulations in Tables 1 and 2 are
simply statements of the amounts of
material falling within each range of
size (size-fraction).
B Example of Frequency Distribution Curves
1 When gas-borne particles are produced
in industrial operations, there is a
tendency to form a "preferential" parti-
cle-size. As a result, particle-size
distributions tend to approximate a
probability relationship with a peak at
the preferential size. This fact is
demonstrated in Figure 1.
Table 1
22
Size
range
(v)
0 5
5 - 10
10 - 20
20 40
40 80
80 160
160 - 320
+ 320
% by wt.
in size
range
2. 3
2. 7
9. 1
13. 3
21. 1
34. 5
15. 6
1. 4
Table 2
Particle Si'/.e
Figure 1
Size
range
(M-)
0 - 2
2 4
4-6
6 8
8 10
10 - 12
12 14
14 - 16
16 18
18 - 20
20 22
22 24
24 26
26 28
28 30
30 32
32 34
34 36
36 - 38
38 - 40
+ 40
% by wt.
in size
range
3. 0
7. 5
8.5
8. 0
7. 0
6. 0
5. 2
4. 8
4. 0
3.8
3. 2
3. 0
2. 5
2. 3
2. 2
2. 0
2. 0
1. 5
1.5
1.0
21. 0
J>
/0.S
/
±Y
3 +-
4-$
t%,X
7/-°
/^O
/J,0
7 -t,J>
/ff,r
/?
/"^C1
PA. C. pm. 62a5. 61
-------
Representation of Particle-Size Data
Curve (1) shows a "normal" proba-
bility distribution in which the dis-
tribution is symmetrical about the
preferential size.
This form of curve is rarely en-
countered for dusts formed by com-
minution. However, it may be found
for particles, such as fumes formed
by vapor phase reaction and condensa-
tion, or for tar and acid mists.
Curve (2) resembles the normal
probability except that it is skewed
(off center).
This is the type of curve usually
obtained for comminuted dusts.
Tables 1 and 2 would plot skewed
similar to Curve (2).
Curve (3) demonstrates that some
materials may show more than one
preferential size.
C Construction of the Frequency Distribution
Curve
1 Type of paper
a Usually, frequency distribution
curves are plotted on regular co-
ordinate (linear) paper.
2 The ordinate and abscissa
a The percentage by weight (or the
frequency) is plotted as the ordinate.
b The average particle-size of each
size-range (size-fraction) is plotted
as the abscissa.
Enrollee: Plot a frequency distribu-
tion curve on Figure 2 using data
given in Table 2.
3 Significance of the size-range (size-
fraction)
a Need for a single system of selecting
size-ranges.
1) Since the curve is plotted using
average particle-size within each
size-range, for a given dust there
may be several positions of the
curve, each position depending
on the series of size-ranges used.
Thus, it is essential that a single
system of selecting the extent of
each size-fraction be adhered to
in order to make accurate inter-
pretations from the plotted results.
Systems for selecting size-ranges
(size-fractions)
1) Select equal arithmetic incre-
ments of size (See Table 2)
2) Choose size-ranges bounded by
sizes having the same ratio to
each other (See Table 1). (This
is equivalent to taking equal in-
crements of the logarithm of the
particle-sizes, a method used in
the log-probability type of fre-
quency distribution).
3) Divide the frequency (plotted as
the ordinate) by the micron-range
of the increment involved, pro-
viding a quotient of "frequency
per micron. " This method has
the advantage that no strict sysr
tern of selecting size-ranges need
be followed.
D Prediction of a Size-Distribution of a
Particulate Sample
It is evident from the above that a large
number of points are necessary to fix the
position of the frequency distribution curve.
Hence, if this were the only method of
representing the whole of a particulate
sample, particle-size analysis would be
an extremely laborious and time-consuming
procedure.
The following demonstrates methods of
predicting the distribution of particles in
a particulate sample by splitting the
sample into a few size-fractions.
-------
ID
13
i-t
(D
CD
o
0
o
Hi
Figure 2 Frequency distribution curve for data in Table 2
CO
H-
N
fD
tu
rt
CD
-------
Representation of Particle-Size Data
III THE SEMI-LOGARITHMIC CUMULATIVE
DISTRIBUTION CURVE
B Construction of a Semi-Logarithmic
Cumulative Distribution Curve
A Example of Tabulation of Data
Table 3
Size- % by wt. in
Range size-range
(1)
0 5
5 - 10
(2)
2. 3
2. 7
% by wt. smaller
than largest size
in range
(3)
2. 3
5.0
10 20
20 - 40
40 - 80
80 160
IfiO - 320
+ 320
9. 1
13. 3
21. 1
34. 5
14. 1
27. 4
48. 5
83. 0
15. 6 98. 6
1.4 ! 100.0
1 Type of paper
a Semi-logarithmic graph paper is
used.
2 The ordinate and the abscissa
a The particle-size is plotted on the
logarithmic abscissa
b The percent by weight (or frequency)
less than the size indicated on the
abscissa is plotted on the linear
ordinate.
c Example
The tabulation of column (3) of Table 3
is a statement of the percent of the
total weight of the sample attributable
to all particles less than an indicated
size.
in
c
jc 5
^ OJ ^
en £ U
en cti co
cu cj .
^-i -rH IH
T3 rt
Particle-size
(logarithmic scale)
Figure 3
-------
Representation of Particle-Size Data
1) A portion of such a plot will be a
straight line as between A and B
of Figure 3. However, this
straight line portion includes only
about 70% of the total sample; 15%
of the total sample lies beyond
point B, and the other 15% before
point A.
Hence, it is erroneous to assume
that a cumulative distribution
curve plotted on semi-log paper
is a straight line over the entire-
range of particulate sizes in the
sample.
Enrollee: Plot a semi-logarithmic
cumulative distribution curve on
Figure 4 using data given in Table
2.
3 Significance of the size-range
The selection of size-ranges is not
important
Prediction of Size Distribution of a Particu-
late Sample
As is evident from the above, the cumula-
tive distribution curve is not a straight
line for the entire range of particle-size
in a sample. Hence, a somewhat large
number of size-fractions must still be
used to determine the proper position of
the curve, especially the position of the
extremes.
IV THE LOG-PROBABILITY CUMULATIVE
DISTRIBUTION CURVE
A Example of Tabulation of Data
The data is tabulated as indicated by
Table 3.
B Construction of a Log-Probability Cumula-
tive Distribution Curve
1 Type of paper
a A special coordinate paper is used,
known as log-probability paper.
One scale is logarithmic, the other
is a special probability type.
2 The ordinate and the abscissa
a The particle-size is plotted on the
logarithmic ordinate
b The percent by weight (or frequency)
less than the size indicated on the
logarithmic ordinate is plotted on the
probability scale as the abscissa.
3 Example
a The entire plot becomes a straight
line if the frequency distribution
plot is skewed as in Curve (2) of
Figure 1.
Enrollee: Plot the data of Table 2
on Figure 5.
_0)
"d
% by weight less than si/.e indicated
(probability scale)
-------
IJ-L
100
90
80
70
60
50
40
30
20
10
0
iTTT
! I I
J--1.
1
1.5
10
15
20
30 40 50 60
Figure 4 Semi-log frequency distribution for data in Table 2
-------
Representation of Particle-Size Data
V THE ARITHMETIC PROBABILITY
CUMULATIVE DISTRIBUTION CURVE
A Example of Tabulation of Data.
The data is tabulated as indicated by
Table 3
Construction of an Arithmetic-Probability
Cumulative Distribution Curve
1 Type of paper
a A special coordinate paper is used,
known as arithmetic-probability
paper. One scale is linear, the
other is a special probability type.
2 The ordinate and abscissa
a The particle-size is plotted on the
linear ordinate.
b The percent by weight (or frequency)
less than the size indicated on the
linear scale is plotted on the proba-
bility scale as the abscissa.
3 Example
a The entire plot becomes a straight
line if the frequency distribution is
"normal" as in Curve (1) of Figure 1.
VI USEFULNESS OF PROBABILITY
CUMULATIVE DISTRIBUTION CURVES
A A Relatively Few Size-Fractions Need Be
Analyzed
1 If a few observations of cumulative
percent undersize give points which
fall on a straight line when plotted on
either arithmetic or logarithmic-
probability paper, one is reasonably
justified in taking the straight line for
the cumulative distribution curve.
2 For some approximations it may be
satisfactory to know the amount of
particulate less than only two specific
sizes in order to find two points for the
plotting of a straight line.
3 In some cases, it may be known that a
plotted distribution of a particular dust
always has the same slope. Hence, for
approximations, only one point may be
necessary to provide an indication of
distribution over the entire particle-
size range.
0)
CD 'ol
% by weight less than si?f
(probability scale)
-------
V
99.99 99 9 99 B
Figure D - Log-Probability Distribution for Data in Table 2
99 99 JS 90 80 70 60 50 40 30 20 10 5 21
0.5 O.J 0.1 t.G5 0.01
99.8 M.9" 99.99
-------
B Frequency distribution tabulation (as shown
in Table 1), and curves (as shown in Figure
1) may be plotted from data provided by the
probability cumulative distribution plot.
1 The frequency distribution curve can be
constructed by taking the change in
cumulative percentage for each small
increment of size.
For example: If 24% of the total
weight of sample is due to size less
than 1. 0 micron and .30% due to size
less than 3. 0 microns, then 6% of the
total weight of the sample is due to
sizes between 1. 0 and 3. 0 microns.
Since the frequency curve plots on
average size of an increment, the
plotting points become 6% and | 1 + 3
REFERENCES
1 Encyclopedia of Chemical Technology,
Volume 12. The Interscience Encyclo-
pedia, Inc New York. 1954.
2 Lapple, C. E. Fluid and Particle Mechanics.
U. of Delaware. Newark, Delaware.
1951.
3 The Problem Characteristics, Industrial
Dust. The Thermix Corporation.
Greenwich, Conn.
4 Particle Size Analysis. Bulletin No. G402R.
Western Precipitation Corporation.
Los Angeles, Cal.
5 Lapple, C. E. Representing Distribution
of Particle Size. Heating, Piping, Air
Conditioning 18: 108. February, 1946.
6 Hawksley, P. G. The Physics of Particle
Size Measurement- Part I. Fluid
Dynamics and Stokes' Diameter. The
British Coal Utilization Research Assoc.
Monthly Bulletin, Volume XV, No. 4.
April, 1951.
-------
STATISTICAL PRESENTATION OF DATA
Arithmetic Probability Cumulative Frequency Distribution
98 95 90 $p 70 40 SO 0 30 20 10 5 1
r
A
/
2 5 10 >
plotting position
/
X
/
/
/
/
/
/
f
/A
0304030607080 90 93 9
1
N + 1
If data are normally distributed, as N *• ™ median = mean
84.13
50
PA
.S. 17.3.67
-------
Statistical Presentation of Data
Log Probability Cumulative Frequency Distribution
98 95
9®
§0 TO
SO 40 30 20
10
9
G
7
6
5
4
3
2
8
7
6
5
4
3
2
1
A
y
yX^
1^"
>^
r '
I 5 TO 20 3
plotting position =
>
s
S
J>
/
/
>
X
>
^
x
>^
^A'
^
0 40 90 60 70 80 90 95 91
1
N + 1
If data are log-normally distributed, as N-
geometric mean
median
so
-------
SIZE-EFFICIENCY CURVES
I EFFICIENCY GUARANTEES ON
COLLECTION EQUIPMENT
A Efficiency guarantees, not outlet concentra-
tion guarantees, are made.
Particle collection equipment usually
operates upon a particulate-gas mixture
that is difficult to specify exactly, is
likely to be highly variable, and over which
there is little positive control. Therefore,
the manufacturer of the collector will
guarantee only efficiency, since efficiency
will vary less than other performance
indices with changes in the quantity or
characteristics of the feed.
On the other hand, the customer may be
primarily interested in the concentration
of particulate matter in the exit gases,
and hence, desires that this be the subject
of the guarantee. Loadings in exit gases
cannot be guaranteed unless the inlet
loadings are specified. But, specifying
inlet and outlet loadings is the same as
making an efficiency guarantee.
B Efficiencies are expressed in terms of:
1 Overall efficiency, or
2 Size-efficiency
II SI/.K-EFFICIENCY CURVES
A ATI or requirements for a valid test have
heen met, and particle-sizing methods
agreed upon, efficiency tests are run.
M Steps Leading to the Construction of a
Si/,e -Efficiency Curve
1 The overall efficiency by weight is
calculated.
2 Size-distribution curves for inlet and
outlet, inlet and catch, or outlet and
catch, are constructed on log-probability
paper.
3 From the size-distribution plots,
percent by weight in selected size-
fractions are determined, and the
efficiency of collection for each size-
fraction calculated.
4 The size-efficiency curve is drawn by
plotting the mid-point of each size-
fraction (as the absissa) against the
appropriate efficiency (as the ordinate)
on regular graph paper.
C Size-efficiency curves should specify:
1 Method of size-analysis
2 Carrier-gas (assumed air unless
specified)
3 Temperature and pressure of the
carrier-gas
4 Grain loading
5 Nature of the particulates
6 True density of the particulates
7 Flocculation characteristics of the
particulates
8 Type and dimensions of the collector
III GENERAL CLASSIFICATION OF
EFFICIENCY
50-80
80-95
95-99
99-99.9
above 99. 9
Classification
Low efficiency
Medium efficiency
High efficiency
Very high efficiency
Ultra-high efficiency
PA. C. pm. 06. 9. 60
-------
10
SECTION 10
Settling Chambers
-------
SETTLING CHAMBERS*
I DESCRIPTION
A gravity separator consists of a chamber
in which the velocity of the carrier gas is
made to drop suddenly so that the particles
in the gas stream will settle by gravity.
Reduction in velocity is accomplished by'
expanding the ducting into a chamber of
suitable dimensions to obtain the desired
low velocity.
II TYPES OF SETTLING CHAMBERS
A Simple Settling Chamber
The simple settling chamber consists
of a single compartment (See Figure 1).
dull collection hoppers
Figure 1. HORIZONTAL FLOW SETTLING CHAMBER
Howard Settling Chamber
In this separator, horizontal plates
parallel to the line of flow of the
carrier gas stream are used to obtain
f\ number of comnartments in which the
particles mav settle (See Figure 2).
Shelves between compartments may have
vortical dimensions as small as
one-inch.
Figure 2. HOWARD SETTLING CHAMBER
III DESIGN PARAMETERS
A Theory
Consider Figure 3 below
Figure 1. SIMPLE SETTLING CHAMBER
*Manual outline as revised by R. T. Shigehara,
January, 1.970
I'A. <:.,.",. 2 Jr. I-/'>
-------
Settling Chambers
If a particle enters the chamber at a
vertical distance hp above the lower
level of the chamber height Hc, it must
fall this distance hp before it tra-
verses the horizontal distance Lc if the
particle is to be collected by the cham-
ber.
Assuming that the particle velocity
Vx is the same as the velocity of the
gas stream, and that we have a uniform
velocity profile, the residence time 6
for all particles within the chamber is:
(EQ 1)
A particle of size D will settle some
distance hf) in time 9. If hg > Hc, these
particles wil] be collected with 100%
efficiency. If hg < Hc, whether a given
particle will be collected or not depends
upon Its position h_. For example, if
h0 = 0.5 H(., all particles of size Dp
wlio.se position hp ^_ 0.5H will be collected
and nil p.'irticles where hp> 0.5HC will not
be collected. Therefore, the fractional
t'f Tir-icncy Is 0.50. This Is true assuming
Ihnl all Llie particles of size Dp are
uniformly distributed over Ilc.
Willi this i.n mind, we can write:
(I'Q .;)
Theoretical Sije
Efficiency Equation: (EQ 4);
Settling Chamber
where E = fractional efficiency of
particles of size Dp
Vy = particle settling velocity
Vx = gas velocity through the
chamber
LC = chamber length
H = maximum distance particle must
settle in order to be collected.
Stoke's law offers a reasonable approxi-
mation of Vy for particles of concern
(see Table 1). A generally accepted
rule of thumb: Stoke's law applies when
Stoke's law says
;y(s)
g Dp2 (Pp - p)
18L1
(EQ 5)
where Vy(s) = settling velocity in Stoke's
law range
g = acceleration due to gravity
Dp = particle size
pp particle density
p = gas density
V = gas viscosity
where Ep is the fractional efficiency for
,\ eiven particle size Dp. Hc can now be
expanded to identify the maximum distance
that the particles must settle in order to
I'i? collected.
Tlivj distance hg that a given size particle
will settle is dependent upon the settling
voloci tv Vv
Thus:
Review the assumptions that concur with
Stoke's law in the manual outline titled
''Particle Settling Dynamics", (Section 4
of the Combustion Evaluation Manual).
The horizontal velocity Vx can be rewritten
in terms of the volumetric effluent flow
rate and chamber dimensions:
(KQ 3)
»c Bc
(EQ 6)
I I follin-.'.s Lliat (substituting liQl and EQ3
i ill o l'f'2 ) :
where H is the chamber height and B
is the chamber width. C
-------
Settling Chambers
Table 1. SETTLING VELOCITIES OF SPHERICAL PARTICLES OF
UNIT DENSITY IN AIR (D
Temperature: 20°C(68°F): Pressure 760mm Hg.
Particle diameter Experimental
microns
cm/sec
o-i
0-2
0-4
1-0
2
4
10
20
40
100
400
1000
8-7
2-3
6-8
3-5
1-19
5-0
3-06
1-2
4-8
24-6
157
382
x 10-
x 10-"
x 10-"
x 10-3
x 10-2
x 10-2
x 10-1
Calculated
from Stokes' law
cm/sec
8-71 x
2-27 x
6-85 x
3-49 x
1-19 x
5-00 x
3-06 x
1'2
5
25
483
3050
io-5
10-*
10-"
io-3
IO-2
io-2
IO-1
The maximum settling distance HC of
the
Howard settling chamber is the chamber
height Hc divided by (N + 1) number of
trays.
B Design Considerations
A Gas Velocity
The velocity of the gas stream should
be kept as low as possible because
the settling rate of dust decreases
when the turbulence of the gas is in-
creased.
But, although it is in the streamline
region of flow where the most effect-
ive settling is obtained, it is not
practical to reduce the gas stream
velocity to the streamline region
because of the large and impractical
cross-sectional area that would be
required.
For practical purposes, and so that
the velocity will not be so great as
to re-entrain the settled particles,
the generally accepted rule of thumb
is: that velocities below 10 ft/sec
are satisfactory for most materials.^)
Table 2 lists some pickup velocities
of various materials.
Table 2. PICK UP VELOCITIES OF VARIOUS MATERIALS
(3)
Material
Aluminium chips
Asbestos
Non-ferrous foundry dust
Lead oxide
Limestone
Starch
Steel shot
Wood chips
Wood sawdust
Density
g/cm3
2-72
2-20
3-02
8-26
2-78
1-27
6-85
1-18
-
Median
size
microns
335
261
117
14-7
71
64
96
1370
1400
Pick up
Velocity
ft/sec
14-2
17-0
18-8
25-0
21-0
5-8
15-2
13-0
22-3
-------
Settling Chambers
B Efficiency Equation
The actual magnitudes of gravitation-
al settling velocities Vy(s) as
determined by Stokes' law are not
used in settling chamber design be-
cause of inherent factors that are
impossible to predict by a theoreti-
cal mathematical expression.
For example, all particles do not
necessarily have undisturbed free-
fall; agglomeration during settling
may change the original particle size;
some particles may be re-entrained.
EQ A is only theoretical and does not
take into account eddy currents and
uneven distribution of air flowing
through the chamber.
Because of the above factors, EQ 4 in-
cluding EQ 5 and EQ 6 should be written
gpp 2(pp-p)
where: K =» empirical factor; a factor
of 0.5 frequently used when
no other information is
available
g = gravitational acceleration
p = Particle density
PP = gas density
y = gas viscosity
Q = volumetric flow rate
B = chamber width
L = chamber length
N - number of parallel chambers,
1 for a simple chamber and
N trays + 1 for Howard
settling chamber.
IV APPLICATION
Despite the fact that settling chambers
are simple in design and can be manu-
factured from almost any material, they
are infrequently used because of the
extremely large space requirements and
the relatively low efficiency. Where
settling chambers are used, they are
normally followed by a more efficient
collecting device.
Efforts have been made to improve the
efficiency of gravity settling chambers
by the use of baffles and various other
methods. The Howard settling chamber
is an example of these efforts. But
this type of settling chamber was never
widely used because of the difficulty in
removing the settled dust from the
horizontal trays.
A Metal Refining
The combination settling chamber and
cooling device has been widely used
in the metal refining industry to
partially collect large particulates
and to reduce the gas temperature to
the final collecting device. One of
the more common types, the "hair pin
cooler", is shown in Figure 4.(1)
Figure 4. Hair-pin cooler
B Arsenic Trioxide
Arsenic trioxide from smelting arsen-
ical copper ores has always been
collected in brick settling chambers
known as "kitchens".
C Foodstuffs
In the manufacture of various food-
stuff, simple settlement is the
first step in dust recovery, achieved
by spraying the condensed liquids
into large chambers.
The effluent air is then passed to
second stage cleaners (cyclone) and
the exhaust re-circulated to the
spray chambers.
D Boilers
Power and heating plants may employ
settling•chambers.
-------
Settling Chambers
REFERENCES:
1. W. Strauss, "Industrial Gas Cleaning," 4. Joglekar, G. D. and Subramanian, N. R.
Pergamon Press, Oxford 1966, pg. A Single Vane Cyclone Separator. J.
144-159 Sci. Industr. Res. 14A. 1955.
2. C; E. Lapple in J. H. Perry, Ed. Chemical 5. Drinker, P. and Hatch, T. Industrial
Engineer's Handbook, 3rd Edition, Dust. McGraw-Hill Book Co., N. Y.
p. 1021, McGraw-Hill, N. Y. (1950). 1954.
"1. J.. Balif, L. Greenburg, and A. C. Stern, 6. Lapple, C. E. Fluid and Particle
Am. Ind. Hyg. Association Quarterly Mechanics, U. of Delaware, Newark,
9, 85 (1948) as referenced in (1) Del. 1956.
-------
11
SECTION n
Cyclones
-------
CYCLONES
D. James Grove*
I INTRODUCTION
The cyclone collector is an inertial separa-
tor without moving parts where a confined
vortex is formed which produces sufficient
centrifugal force to drive the suspended part-
iculate to the collector wall. Although it is
simple, inexpensive, and can be constructed
from many materials, there are few applications
where collection efficiencies exceed 80-90%.
Cyclones with a body diameter less than__nine__
inches have been arj^££rJly deslgnated_as
''^ej}__g££j:cig£QY!!-_"cycIgnesTj'As~^Il"l)'e shown
•"Ta t eT, smaTTerbody diameTers create larger
separation forces, and consequently (up to
some practical limit) provide higher efficienc-
ies.
II MECHANISM OF PARTICLE COLLECTION
The basic components of a cyclone are shown
in Figure 1. These include a cyliner, a tang-
Cleaned-Gas Exit
Dust Laden
Gas Inlet
Cylinder
Cone
Dust Hopper
Collected Dust
Figure 1. BASIC CYCLONE COMPONENTS
ential gas inlet, a cone to deliver the
collected dust to a central disposal point,
a dust hopper, and an axial gas outlet, part
of which extends into the cylinder. The con-
fined vortex of a cyclone is illustrated in
Figure 2. The gas enters tangentially into
the annular space between the cyclone body
and the outlet tube, and spirals downward
in what is called the main vortex. Near the
bottom of the cone, the spiralling gases be-
gin to move upward in the vortex core. The
spiralling action of the gases causes the
particles to be driven to the walls by cent-
rifugal force, where they are moved towards
the dust discharge by the force of gravity
and the downward movement of the main vortex.
It should be noted that the spiral motion of
both vortices is in the same direction. The
tangential velocity (how fast the gases are
swirling) is lowest near the wall and at the
center of the cyclone. It reaches a maximum
at a point approximately 60-70% of the way
in from the wall to the center.
Eddy
Main Vortex
Vortex Core
Figure 2.
VORTEX AND EDDY FLOWS IN A
TYPICAL CYCLONE DUST COLLECTOR
DESIGN.
Not only are there variations in the tang-
ential velocity at different points in the
cyclone, there are also vertical eddies and
what is called inward drift. The inward drift
is a radial gas flow which moves toward the
center of the cyclone, opposing the movement
of particles. While vertical eddies can exist
In the cone, the most troublesome are those
present in the annular region near the gas
inlet. The eddies, which are caused by the
vortices and not by the gas outlet extension,
can carry particles directly from the gas
inlet to the gas outlet.
All of these variations combined together make
the problem of determining the separation forces,
and consequently the efficiency, much more
difficult.
*Chemical Engineer, Institute for Air Pollution Training
PA.C.pm.104.4.73
-------
III DETERMINATION OF CRITICAL PARTICLE SIZE
AND CUT SIZE
There are two sizes which are commonly used
to relate to the efficiency of a cyclone. The
equations given for both of these are empirical
relationships, and their derivation will not
be presented here. They should not be used for
original calculations, but rather for comparing
the efficiencies of similar cyclones operating
at different conditions.
, £h,e
size of the smallest _^^^
'removed completely "(removed with JLOO% efficiency)^
"from a 3TisT 'TaHen gas stfreamT
[D ] _
pjcr-
V9 p (D - d0)
27T N v± (p
- P)
= diameter of gas outlet, ft.
= number of revolutions the
gas stream makes (5-10 is
typical)
= inlet gas velocity, ft/sec
= density of the particulate,
lbs/ft3
= density of gas, lbs/ft
= viscosity of the gas, Ibs/
ft-sec
where: [D ] = critical particle size
p cr v
D = diameter of cyclone body, ft.
A size-efficiency curve is a plot of the re-
lationship between different sized particles
and the efficiencies with which they are
removed in a certain cyclone. An example of
such a curve is shown in Figure 3. If we
determine [D ] from one of these curves, it
would be the diameter which corresponds to
100
80
60
20
20 40 60 80 100
Particle Size - Microns
Figure 3. SIZE EFFICIENCY CURVE
120 140
-------
100% on the efficiency scale. For the curve
shown in Figure 3, the [D ]^ would be approx-
imately 90 microns.
Since it is rather difficult to determine
[Vcr accurately fr°"i the graph, the cut size
is often determined instead. The cut size is
defined as the size of the particle which is
removed with 50% efficiency.
[D ]
p cut =
p^T
* 2ir M -u
2ir N v (p - p)
1 p
(2)
where:
[D ] = cut size
p cut
w = width of the gas in-
let , ft.
Referring back to Figure 3, the [D ] would
be 12 microns. p cut
IV PRESSURE DROP DETERMINATION
The pressure drop across a cyclone collector
will generally range between one and seven
inches of water and it is usually determined
empirically. An equation does exist which can
be used for relating the pressure drops for a
cyclone operating at several different condi-
tions, or for geometrically similar cyclones.
0.0027 Q
Ap =
O)
k d
D / \ D
where: Ap = pressure drop, Inches of water
Q = volumetric flow rate at the Inlet,
cu. ft./sec.
L = height of the gas inlet, ft.
L = height of the cylinder, ft.
L = heicht of the cone, ft.
cone
k = dimensionless factor descriptive
of the cyclone inlet vanes
= 0.5 without vanes
= 1.0 for vanes that do not expand
the entering gas or touch the
gas outlet wall ("a" in Figure 4)
<= 2.0 for vanes that expand the
entering gas and touch the gas
outlet wall ("b" in Figure 4)
V CONFIGURATIONS AND RELATIVE DIMENSIONS-
Effects of Cyclone Performance
The dimensions of a cyclone are of primary
importance when considering efficiency and
pressure drop. A widely recommended cyclone
with an ordinary tangential inlet would have
the following proportions, based on the cyclone
body diameter D:
Table 1.
Cylinder length (L ) = 2D
Cone Length (L ) = 2D
cone
Outlet extension (into cyclone) length (L )
= 0.675 D °
Inlet height (L ) = 0.5 D
Inlet width (w ) = 0.25 D
Outlet diameter (d ) 0.5 D
o
Most design changes intended to increase
collection efficiency also increase pressure
drop. Conversely, recovering energy from the
outlet vortex is the only method of reducing
the pressure drop which does not also reduce
efficiency.
Increasing the inlet velocity will increase
the efficiency, although the relationship is
very complex. There is also an upper limit,
about 100 feet/sec, above which there is in-
creased turbulence which in turn causes re-
entrainment of the separated dust and reduced
efficiencies. The pressure drop will Increase
as a function of the inlet velocity squared,
or twice the velocity yields four times the
pressure drop.
The length of the cyclone body determines the
residence time during which the particles are
subject to the separating forces, and Increas-
ing this length will increase efficiency. Also
dust which has been entrained in the vortex
core will have more time to become reseparated.
Increasing the body diameter to outlet diameter
ratio will also increase efficiency, although
the optimum ratio is between 2 and 3.
As can be deduced from the definition of "high
efficiency" cyclones, decreasing the body
diameter will increase the efficiency. This is
due to increased separation forces caused by
the smaller vortex radius. In a very small
cyclone, however, the dimensional clearances
are so small that plugging occurs easily. Small
cyclones also experience bouncing of larger
-------
particles and local turbulence which reduce
efficiencies, and diameters of of 2 to 3 inches
seem to be a practicaT~mSffiu^"™"™°'*~-~-=~™~~~~™=
The cone portion of the cyclone is not neces-
sary to convert the downward vortex to an up-
ward one, although its presence does reduce
the length of cyclone needed to effect this
reversal. The main purpose of the cone is to
deliver the collected particles to a central
point for easy handling and disposal. The
cone cannot be too small in diameter at the
bottom, or the vortex core will contact the
walls and re-entrain the collected dust.
Dust discharge design is just as important in
reducing this re-entrainment, due to the high
turbulence and velocities present near the
discharge. The static pressure in the vortex
core may be slightly negative, and this will
tend to draw the collected dust up away from
the discharge. The best solution is some type
of mechanical device, such as a rotary valve,
a double-flap valve, a screw conveyer, or a
dip leg. To be successful, the mechanism must
achieve continuous, complete, and immediate
removal of the separated dust and prevent in-
flow of gas from the hopper.
As the inlet air enters the annular region at
the top of the cyclone, it is squeezed by the
existing gas to about half its inlet width.
This causes a significant pressure loss, which
can be reduced by adding vanes to the annular
area (see Figure 4). The presence of the vanes,
Figure 4. INLET VANES
however, reduces the efficiency, apparently
due to the prevention of vortex formation in
the annulus. Helical or involute inlets (Figure
5) are attempts to reduce interference between
the incoming gas and the vortex already present
in the annulus. Axial inlets are free from
most of these problems, but they introduce new
problems. The inlet vanes must be designed so
that they impart adequate rotation to the gas,
and yet resist erosion and plugging.
Figure 5. CYCLONE INLETS
-------
The eddy current in the annular region requires
that the gas outlet extend into the cyclone to
prevent excessive amounts of dust from passing
directly from the inlet to the outlet. Usually
this extension ends just below the bottom of
the inlet. Devices which permit the gases to
leave the gas outlet tube tangentially have
been successful in reducing the pressure loss
without sacrificing the efficiency.
Since the pressure drop in a cyclone is caused
by the vortex and not by wall friction, rough
walls actually reduce the pressure drop due to
the suppression of the vortex formation. They
also greatly reduce the collection efficiency,
due to increased turbulence and re-entrainment.
VI CARRIER GAS AND PARTICULATE CHARACTERISTICS
Effects on Cyclone Performance
Changesin particle size, density ,,_and con-_
cenfTratTqK do not .have a' significant effect
on the pressure drop across a cyclone. As far
as the efficiency is concerned, however, this
is not the case. Efficiency will increase with
increases in partTcTe density, mean particle
size, and concentration. The larger a particle
is, and the more it weighs, the better the
separation forces. The effect of increased
dust loading (concentration) Is not quite so
obvious, but it is in part caused by the small
particles getting swept to the walls by the
many large particles.
From equation 2 or 3, we can see directly the
effect of the carrier gas properties on the
efficiency, in terms of the cut size or critical
particle size. A lower value for [D ] or
[D ] means a higher efficiency, and this
p cut
would result from a lower viscosity or a lower
gas density. Temperature and pressure will
affect the density and viscosity, although
this is not generally significant. A change
in pressure would produce only a slight change
in density, and a change in temperature would
Increase one and decrease the other, with a
small net effect.
Although there may be a slightly higher effi-
ciency using several cyclones in series, the
additional pressure drop Is usually sufficient
to make series operation disadvantageous.
Occasionally a large diameter cyclone is used
as a precleaner for small multiple cyclones
to remove the larger particles that cause plug-
ging in the multiple units.
VIII FURTHER MATHEMATICAL CONSIDERATIONS
In order to obtain a prediction of overall
cyclone efficiency, a size efficiency curve
for a given cyclone under a given set of op-
erating conditions, and a particle size dis-
tribution are needed. The size efficiency curve
is characteristic of the cyclone, treating
the particular powder used in the test, when
running at the throughput of the test. Cor-
rections for differences in particle grading,
particle densities, gas viscosities, and gas
flow rates can be made using the equations
already presented, although care must be taken
in making large transpositions.
If the calculations are based on the cut size,
equation 2 will be the starting point:
[Vcut
2rr N v± (p - p)
(2)
In order to determine the effect of the var-
ious parameters on the cut size (or the ef-
ficiency) , a ratio is taken of the test cyc-
lone and conditions to the new cyclone and
conditions
9 y wj
p cut (test) T2TT N v (p - p)
test (A)
[Dp]cut (new) —
2'fr N v, (p - p)
i p new
VII TYPF.S AND ARRANGEMENTS OF CYCLONES
individual high efficiency, small diameter
cyclones have a small capacity, and they must
ho operated in parallel to handle typical gas
volunirs. They generally have a common gas in-
let, dust hopper, and gas outlet, and can be
arranged in banks of several hundred cyclones
each. A whole new set of design problems arises
witli this arrangement, but it is advantageous
in that we can get a higher efficiency com-
pared to one large cyclone, with about the
same pressure drop.
If, for example, the effect of changing vis-
cosity is to be determined, holding all the
other parameters (w , N, v , p , p) constant,
the ratio would be:
p cut (test)
p cut (new)
IV (test)
y (new)
(5)
-------
which can also be written:
[D ] , , = [D ]
p cut (new) p cut
(test.
(new)
(test)
(6)
By the same reasoning, the result of a change
in gas inlet velocity would be:
[D ]
P cut (new)
p cut (test)
(test)
[v. (new)
(7)
Repeating this procedure, the effect of
changing N, p , and p can also be determined.
Lapple (5) states that the above approach is
applicable only for geometrically similar
cyclones. Based on the dimensions given in
Table 1, Lapple presented a method for deriv-
ing the size efficiency curve from the cut
size alone. Once the cut size [D ] has
p cut
been determined, the efficiency for any size
particle D can be arrived at by calculating
D /[D ] , and then using the graph (Figure
6) to determine the efficiency. For example,
if the cut size is 20 microns and we want the
efficiency for a 10 micron particle, D /[D ]
3 P p cut
is 0.5, and the efficiency is about 22%. A
complete size efficiency curve can be drawn
by getting the efficiency at several different
particle sizes.
Collection Efficiency, %
3 o SD fe o c
/
/
/
/
^
f
'
./
s
Jf^
^l^~~
— '
.1 in
y
0.3 0.4 0.5
Particle Size Ratio, (D /D )
P PC
^Figure 6. CYCLONE EFFICIENCY VERSUS--
""\ PARTICLE SIZE RATIO
""-•--^ (LAPPLE, 1951).
According to the Air Pollution Engineering
Manual (6), experimental data has(compared
favorable with Lapple's correlation, except
for slightly lower efficiencies than those
calculated for D /[D ] ratios of 2-3.
p p cut
Apparently, Lapple's correlation nuy be suf-
ficiently accurate for an engineering estima-
tion of many cyclone applications.
Gallaer (3) determined that if the size ef-
ficiency curve of a cyclone is plotted on
semi-log paper, as in Figure 7, a straight
TT.T
99
95
90
80
70
60
50
40
20
/
/
/
-/-
/
/
/
&
/
V
/
/
/
/
°0 5 10 15 20 25 30 35 40 4!
Figure 7. SIZE EFFICIENCY CURVE
-------
line results. The equation that represents
this line is:
(d)
where:
overall efficiency
where:
J(d)
= fractional efficiency
of a particle with
diameter d
= particle diameter,
microns
= a constant for the
particular cyclone in
question
The particle size distribution of a typical
dust, when plotted on semi-log paper, also
produces a straight line (Figure 8). The
It should be noted that this formula is use-
ful even for those dusts whose particle size
distribution does not plot as a straight line
on semi-log paper, as long as the value of 6
approximates the slope of the line for those
values of d which significantly affect the
cyclone efficiency, i.e. the smaller particles.
In order to use equation 10, the particle
size distribution curve and size efficiency
curve must be drawn as straight line approxi-
mations, and convenient points are picked off
the straight lines to determine K and S. It
is then a simple matter to determine the over-
all efficiency.
An extension of this approach yields perhaps
an even easier method for the calculation of
the overall efficiency.
3 8 6 4 2
Percent Greater than d
Figure 8. PARTICLE SIZE DISTRIBUTION
equation of this line is:
Z -
(d)
where:
'(d)
the fraction, by weight,
of the total dust having
a larger particle size
than d
Since the cut size, like the value of ^, is
a function of the particular cyclone being
used, a: could be defined as:
(ID
]
p cut
6 = a constant for the part-
icular distribution in
question
Given these equations, it has been shown that
the overall efficiency can be represented by:
where:
a constant
The mean particle size, like the value of 6,
is a function of the particle size distribu-
tion being used, so 6 could be defined as:
+ 6
(12)
p mean
-------
where: y = a constant
Inserting equations 11 and 12 into equations
8, 9, and 10:
T-I -I V.^/ L" J / ^ /I o N
Inlet Gas
Velocity
Pressure Loss
20 to 70 ft/sec (usually
about 50 ft/sec)
0.5 - 2.0 inches of water
for simple cyclones, 2 to 7
inches for high efficiency
units
7
Z. , = e
p mean
(14)
Particle Size
]
p cut
(15)
With E.
3 ] [D ]
p cut p mean
°'5> d = [Dp]cut' and "ith Z(d)
J(d)
0.5, d = [D ] . Inserting these values
' p mean &
into equations 13 and 14, it is found that
x = y. Equation 15 then reduces simply to:
[D
p mean
(16)
[D
p cut
]
p mean
Therefore, if the mean particle size is known,
and the cut size Is known, the overall effi-
ciency can be very quickly calculated.
Not only can the efficiency by calculated
originally from this equation, but the effect
on the efficiency of changing either [D ]
or [D ]mean can be quickly calculated. Before
using this method, however, it is imperative
to determine if the distribution is truly
represented by an essentially straight line.
IX SUMMARY
Gas Flow
Table 2.
30 to 50,000 cfm (some to
100,000 cfm) smaller units
must be arranged in parallel
to accommodate large volumes
1 to 200 microns at vary-
ing efficiencies
High Efficiency 20 to 40 microns for simple
on Normal Indust- cyclones, 10 to 30 microns
rial Dusts with for high efficiency units
Mean Particle
Size of
Particle
composition
Particle
concentration
solid and liquid
down to .1 grains/ft.,
although usually above 10
gralns/ft3, with no real
upper limit
REFERENCES
1. Stairmand, C. J., The Design and Perfor-
mance of Cyclone Separators, Trans. Inst.
Chem. Engrs., Vol. 29, British, 1951.
2. Caplan, K..J., All About Cyclone Collectors,
Air Engineering, pages 28-38, September
1964.
3. Gallaer, C.A., and J.W. Schindeler,
Mechanical Dust Collectors, J_. A. P_. £. A..,
Vol. 13, pages 574-580, December 1963.
4. Kane, J.M., Operation, Application, and
Effectiveness of Dust Collection Equipment,
Heating and Ventilating, August 1952.
5. Lapple, C.E., Processes Use Many Collec-
tion Types, Chemical Engineering, Vol. 58,
pages 145-151, May 1951.
6. Air Pollution Engineering Manual, 999-AP-40
pages 91-99, 1967.
Gas Temperature to 750°F
-------
SECTION 12
Miscellaneous Dry Inertia!-Type
Collectors
-------
MISCELLANEOUS DRY INERTIAL-TYPE COLLECTORS
LOUVER - TYPE DUST SEPARATOR
(Figure 1)
LOUVER TYPE COLLECTOR
Figure 1
A Mechanism of Particle Removal
1 The louver-type dust separator contains
a series of blades set at an angle to
the air stream.
a A large portion of the air (90% of the
total) stream passes through the
louvers.
b A smaller portion (10% of the total)
of the air stream (blowdown) con-
tinues in its original direction with-
out passing through the louvers.
2 The air which passes through the louvers
is forced to turn sharply in a rapid re-
versal of air flow.
a Particles contained in the air stream
impinge upon the blades and rebound
into that portion of the stream (blow-
down) which did not pass through
the louvers.
1) Hence dust particles are con-
centrated in the relatively small
volume of blowdown. Usually a
secondary collection system is
needed to deal with the blowdown.
B Efficiency
1 Efficiency is a function of louver
spacing; closer spacing provides
higher efficiencies.
2 Particles as low as 10-20u. may be re-
tained in excellent designs.
3 When used as a pre-cleaner, particles
of 50-100|i are usually removed.
C Advantages and Disadvantages
1 Advantages
a Simplicity
b Low cost of construction
c Low pressure drop for degree of
removal obtained
d Temperature and pressure limita-
tions are imposed only by materials
of construction.
2 Disadvantages
a Plugging due to buildup of particles
on the blades
b Abrasion difficulties
c Inability to handle tacky materials
PA. C. pm. 67. 9. GO
-------
Miscellaneous Dry Inertial-Type Collectors
II SCROLL TYPE DUST SEPARATOR
(Figure 2)
Dust gas to
secondary
Collector
Partly-Cleaned Gas
Figure 2
A Figure 2 shows the scroll type collector.
B The scroll collector is simply a dust col-
lecting fan which separates particles very
much like the louver-type dust separator.
Secondary collectors are needed to sepa-
rate the particles in the "skimmed off" air
stream. The large-volume gas stream,
stripped of large particles, may then pro-
vide a lighter load for more efficient
cleaning equipment.
C A grade efficiency curve is shown in
Figure 3.
Ill REVERSE NOZZLE IMPINGEMENT
COLLECTORS (Figure 4)
A Figure 4 shows a typical commercial
no/zle impingement collector.
B High efficiency is attained on ducts
larger than 10-20(1.
C These collectors are designed for pressure
drop in the range of 0. 1 1.5 inches of
water.
o
£
a
u
t—1
fa
fc
W
u
H
J
o
CJ
100
80
60
40
20
20 40 60
PARTICLE SIZE, \i
Figure 3
80
100
D Chief advantage lies in their greater
adaptability to existing flues or ducts
than other types of collectors.
E They may be used at elevated temperatures.
(If the dust is tacky, circulating water
films may be used to keep the elements
clean).
IV BAFFLE CHAMBERS (Figure 5)
A Baffle chambers employ fixed baffle plates
which cause the air stream to change di-
rection, thereby projecting the particles
into a dead air space where they settle by
gravity.
B Efficiency
1 Particles greater than 50ji are removed
efficiency.
C Application
1 Baffle chambers are used as precleaners
for more efficient collectors to reduce
the load of large diameter particles on
these units.
-------
Miscellaneous Dry Inertial-Type Collectors
Otogrdmotlc Pion vtew Stowing Gos Mo«
-------
Miscellaneous Dry Inertial-Type Collectors
Figure 5
V DRY TYPE PYNAMIC PRECIPITATORS
A Principle of Operation (Figure 6)
1 Dry type dynamic precipitators are
motor-powered separators in which
the dust is precipitated by dynamic
force produced by the action of numer-
ous specially shaped fan blades.
a The precipitated dust is forced along
the blade surfaces and discharged
into a dust storage hopper.
B Advantages and Limitations
1 Units are compact and space require-
ments are small.
2 Pressure drop is only about l/z inch of
water. However, pressure drop varies
and is a function of mechanical efficiency.
3 It functions both as dust collector and
fan.
4 It cannot handle fibrous, sticky
materials.
DRY TYPE
DYNAMIC PRECIPITATOR
Figure 6
5 Volumes up to 20, 000 cfm may be
handled by some designs.
C Operating Conditions
Gas flow
Gas temperature
Draft loss . . . .
Draft loss sensitivity
to cfm change ....
High efficiency of
removal for ordinary
industrial dusts with
mass median size
greater than
Efficiency sensitivity
to cfm change ....
Particle composition ,
Humid air influence.
up to 20, 000 cfm
to 750°F
a function of mechanical
efficiency. Usually
about YI in. w. g.
a function of mechanical
efficiency.
10-20(1
negligible
cannot handle fibrous,
sticky materials
may cause condensa-
tion and plugging
-------
Miscellaneous Dry Inertial-Type Collectors
REFERENCES 2 Joglekar, G. D. and Subramanian. A
Single Vane Cyclone Separator.
1 Perry, J. II. Chemical Engineer's Hand- Division of Industrial Physics, Na-
book. McGraw-Hill Book Co. New tional Physical Laboratory of India,
York. 1950. New Delhi, India.
-------
SECTION 13
Wet Collectors: Introduction
-------
WET COLLECTORS: INTRODUCTION
I Wet collectors increase particle removal
efficiency by two mechanisms.
A Re-entrainment of the collected particles
is prevented by trapping them in a liquid
film or stream and then washing the liquid
(and trapped particles) away.
B Fine particles are "conditioned" so that
their effective size is increased, thus
enabling them to be collected more
efficiently.
addition of wetting agents does not
significantly increase removal
efficiency.
3 Effect of solubility of particles
Solubility of ihe particles in the
droplets is not a factor in effectiveness.
(An exception is the case of concentrated
mist droplets, such as sulfuric acid.
These droplets may grow in size by
absorption of moisture when passing
through a humid chamber).
II PARTICLE CONDITIONING
Particle conditioning in wet collectors
involves the process of increasing the
effective size of the fine particles so that
they may be more readily precipitated. The
effective size may be increased by:
Forcing precipitation of fine particles on
liquid droplets, or
Promoting condensation upon fine particles
(which act as nuclei) when the water vapor in
a gas passes through its dewpoint.
A Conditioning by Forcing Precipitation of
Particles on Liquid Droplets
1 An example:
An example is the attachment of a
5-micron dust particle to a liquid
droplet 50-microns in diameter thereby
increasing its apparent mass 1000 fold
for collection purposes.
2 Effect of wetting agents in resisting
redispersion
Collision of solid particles with liquid
droplets is inelastic and because of
Van der Waal's forces, the agglomerates
resist redispersion. Therefore, the
III
Conditioning by Promoting Condensation
upon the Particle Surface
If the liquid spray causes the gas to pass
through its dewpoint, condensation will
take place upon the surface of the particles
when the particles act as nuclei. Thus,
the effective size of the particles is
increased under such conditions. This
mechanism is important for initially hot
gases containing relatively small dust
concentrations (say less than 1-grain/cf).
OPERATING PROBLEMS OF WET
COLLECTORS
A Corrosion
All water scrubbers have the inherent
problem of corrosion.
a Even when no chemically corrosive
constituent may be contained in the
carrier gas stream, the carbon
dioxide present contributes to
corrosion.
b When corrosive agents are contained
in the gas stream (SO2, chlorides,
fluorides, nitric acid, etc.),
will occur on wet metallic surfaces.
PA.C.pm. 75. 9. 60
-------
Wet Collectors: Introduction
B Krosion
Wot collectors that remove insoluble,
abrasive materials have troubles due
to erosion especially if removal is
dependent upon impingement velocities
or centrifugal action.
C Wet-Dry
Scrubbers are faced with problems at
wet-dry junctions, particularly at the
entrance of an installation.
a When dust concentrations and gas
temperature are high, there may be a
zone where dust build-up can occur
(by reason of moist dust layers).
D Mist Elimination
1 In all scrubbers, entrainment eliminators
are important to prevent carry-over of
droplets.
2 Many scrubbers have mist eliminators
built into their design.
3 When not incorporated in the design,
mist elimination is accomplished by
means of additional separators.
E Slurry Handling
1 For all scrubbers, a method must
be provided for handling the liquid
effluent. Slurries may be treated by
means of:
a Settling tanks
b Filters
c Liquid cyclones
d Further chemical or recovery
methods
e Disposal to sumps, streams, rivers
f Others
2 All these effluent handling methods have
their own unique engineering problems.
-------
SECTION 14
Collection of Particles on Cylindrical
and Spherical Obstacles
-------
COLLECTION OF PARTICLES ON CYLINDRICAL
AND SPHERICAL OBSTACLES
Particulates transported by a carrier gas
through a depth of cylindrical (fibers) or
spherical(granules) obstacles tend to be
precipitated upon the surface of the obstacles.
Van der Waal's and electrical forces cause
the particulates to adhere to the surfaces of
the obstacles resulting in the removal of the
particulates from the gas stream.
I MECHANISMS OF PARTICULATE
REMOVAL
A Screening (or sieving) is not the principal
mechanism
It can be shown that the sizes of the gas
passages through the depth of obstacles
are very much larger than the sizes of the
particulates collected.
B The principal mechanisms by which par-
ticulates are brought into contact with
the obstacles include:
1 Interception
2 Gravitation
'i Impingement
4 Diffusion
5 Electrostatic
fi Thermal
II INTERCEPTION
Particulates being carried by a flow of gas
tend to follow the streamlines around an
obstacle. By chance, a particle on one of
the streamlines may make contact with the
obstacle if the streamline passes the obstacle
at a distance less than the radius of the
particle. This type of removal is called
direct interception, and depends solely on
the position that a particle has in the gas
stream.
Ill GRAVITATION
As a particle passes by an obstacle, it may
fall (under the influence of gravitational
force) from the streamline along which it is
being carried and settle upon the surface of
the obstacle.
IV ELECTROSTATIC
Since a force of attraction exists between
bodies possessing electrostatic charges of
opposite polarity, it is possible for a
charged particle to be removed from the
gas stream by an oppositely charged obstacle.
However, when only the particle or obstacle
is charged, a charge may be induced upon
the uncharged component resulting in a
polarization force that can also effect
particle removal.
The effect of the electrostatic mechanism
of particle removal from a gas stream may
be significant when" the charge on the particle
or obstacle is high, and when gas velocity
is low. The significance of particle size and
obstacle size varies, depending on whether
the electrical attraction originates from
Coulomb or polarization forces.
The mechanism involved in a bed of fibers
or granules depends principally on the charac-
teristics of the particulates and obstacles in
the bed, and on the gas velocity.
V IMPINGEMENT TARGET EFFICIENCY
A The meaning of impingement target
efficiency
When an obstacle is placed in the path of
a particulate-laden gas stream (Figure 1)
the streamlines will diverge and pass
around the obstacle. The particles, how
ever, tend to leave the streamlines (along
PA. C. pm. OOa. 5. 61
-------
Collection of Particles on Cylindrical and Spherical Obstacles
which they are being carried) at the be-
ginning of the curvature and may impinge
upon the obstacle.
If like particles, initially within a cross-
section of the carrier-gas stream having
a radius of D_^ (measured from the central
2~
streamline) strike a cylindrical obstacle
of diameter Do, then D1 is termed the
"impingement target diameter" of the
obstacle for the particular particles being
considered.
The ratiom' | is called the "impingement
target efficiency" and is symbolized r)j.
In other words, it is the ratio of the cross-
sectional area of the gas stream cleaned
of particles (all of which are alike) to the
projected area of the obstacle.
If it is assumed that all particles are alike
and equally dispersed throughout the gas
stream, the "impingement target efficiency"
is the ratio of the weight of particulate
collected by the obstacle to the weight of
particulate that would pass on if the ob-
stacle were not there. Therefore, "im-
pingement target efficiency" is the effi-
ciency of removal by weight of like
particles by one obstacle.
B The Mathematical Expression
Impingement target efficiency (r}j) is a
function of the dimensionless ratio,
D0g
where:
vp/ofP(s)
impingement target efficiency for
uniformly dispersed like particles
and for one obstacle.
D a diameter of the obstacle
VD/O - relative velocity of the particle
(in the approaching gas stream)
to the obstacle
Stokes1 settling velocity
C Impingement Target Efficiency Curves'
(Figure 2)
Figure 2 demonstrates the relationship
between impingement targe efficiency
(r)j) and the dimensionless ratio
vP/ofp(s)
Note that there are two curves; one for
spheres and one for cylinders. The im-
pingement target efficiency for spheres is
higher than that for cylinders because the
streamlines diverge more sharply around
spheres.
cross-section of air stream
cleaned of particles
L.
D'
r
Impingement on a spherical obstacle
Figure 1
-------
Vo TP(S)
Figure 2 (reference 2)
-------
Collection of Particles on Cylindrical and Spherical Obstacles
100 E
80 —
80 J4
50 =-
40 ^
* 3oLU
I »-
•s
•5 10 .L
? »i
-Q i^
e 5 ,,-1
* 1
ST]"
2^
ir~"
.01
^=Z=
+t
4^
fW
T^
T4^
w~n
f?^
^
•£::£
:::±±
1 1 1
.(
*
t-
X*
a*,
^
i
):
-H-
±i
*
-H"
•xT
j_!_
;=z
»JT
>
-ff
2
r1
/^_
i.
TTT
rH-
^S
*^H
=^=*<
^
v<
JTf-
. ' ''
'• 1 '
.(
F
•^
i
^
yf-
S
±:
-N
S
i
i
)3
H-f
•X
g|
' I
-f
/
•H-l
^
^
^
>t<
4+
if
TTTT
•4-rf-
VH-
^5
#
TtTT
^~y
P
H
=Sr
i
TTT
^E!
7
j —
•W-(-r
TT
* 1
1>T
•S
-So
^
= 4
^4
nrr?
~^~ r~
NT
i
^=
^
r
4- * —
4^
ri —
J. J ;_
Ei±
i J
•"•-^
^fe
^^
F3^
!±f
- ^ •*—
'^^
. ; S
0.1
; — r
t— r
|=M-
EJ,
i 77
"V-j-
^4
^
f-r
(-H-
-H-
-T
«iS
J-^l
^,_
t
*^
~^T
^
1
*j
^
1>
***r
pt-r
34,
^
T"
2)2
q
^
•p
J
-1-
^
^
_L
^
0.
Tf*-
^BT
•^
^R
^^
^•'"V
T-J^
./'
^s;
-------
Collection of Particles on Cylindrical and Spherical Obstacles
100
.01
.02 .03
0.1 0.2 0.3
Gas Velocity (Ft/Sec)
i.o 2.0 ;s.o
Target Efficiency at Various Air Velocities for Different Diameter Dust
Particles (Ref. 2)
-4
(p =2.0) (Fibre diameter 10^) (Viscosity =1.8(10) poises)
P
10.
= a constant depending on the temper-
ature, viscosity, and particle
diameter (cm2/sec)
K = 1.45(10)
-17
where:
T absolute tempc raturc (°K)
^ absolute viscosity of the gas (poise)
D = particle diameter (cm)
2 For air below 100°C
2.45
vp/o Dp Do
where:
(3)
nD diffusional target efficiency for like
particles and one obstacle when gas
stream is air below 100°C (%)
v I relative velocity of the particle (in
P the approaching gas stream) to the
obstacle (ft/sec)
-------
Collection of Particles on Cylindrical and Spherical Obstacles
D diameter of the particle (microns)
Do = obstacle diameter (microns)
B Remarks
The above equations show that if target
efficiency due to diffusion (np) is to be
high, then the relative velocity (vp/o)
must be low. This is opposite to the re-
quirements for high impingement target
efficiency (r^) which demands high relative
velocity. This leads to a velocity zone
for a given obstacle where the efficiency
of removal will be low for a given particle-
size; that is, where conditions are poor
for both impingement and diffusion.
This is evident by observing the low por-
tions of the curves in Figure 4.
VII SIZE-EFFICIENCY
A The equations for target'efficiencies pro-
vide information on the removal of like
particles by only one obstacle, or, let's
say, removal of like particles by one
"treatment" of the gas stream. Since in
the actual course of filtration through a
bed of fibers or granules the gas stream
meets a number of obstacles and is there-
fore "treated" a number of times before
it exits, an efficiency equation taking all
"treatments" into account is necessary.
H The general equation for size-efficiency
E 1(1- n)s°
where So is small |_as in spray devices
(S0 - 5) and old cloth (SQ * 2)j (4)
E = 1 - e'1150
where SQ is large [_as in packed fiber or
granular beds and new cloth filters
(S0 * 50f] (5)
where:
E = efficiency of removal of a given
particle-size (size-efficiency).
The particle size is identified in
e - natural logarithmic base = 2. 718
S0 number of "treatments" received
by the gas stream
_ Total projected area of all obstacles in the filter
0 Cross-section of filter normal to the gas flow
rj = target efficiency of the individual
obstacles
C Size-efficiency for a bed of spherical
granules
E = 1 - e
E = 1 - e'"So
(6)
(7)
where:
E = efficiency of removal of a given
particle-size by a bed of spherical
obstacles. The particle-size is
identified in 17.
e = natural logarithmic base = 2. 718
n - target efficiency of the individual
shperical granules in the bed
L = depth of the bed
« = volume of the spherical granules
per unit volume of bed
Dp = diameter of the spherical granules
S0 = number of "treatments" received
by the gas stream as it passes
through the bed
s _ Total projected area of all obstacles in the bed
0 Cross-section of the bed normal to the gas flow
-------
Collection of Particles on Cylindrical and Spherical Obstacles
D Size-efficiency for a bed of cylindrical
fibers
Impingement P PP Vpl
1 e
irDr
(8)
Diffusion
RT
where:
E
(9)
efficiency of removal of a given
particle-size by a bed of cylindrical
obstacles. The particle size is
identified in rj.
volume of fibers per unit volume
of bed
e - natural logarithmic base = 2. 718
r) target efficiency of the individual
fibers in the bed
L depth of bed
DO diameter of the fibers
So * number of "treatments" received
by the gas stream as it passes
through the bed
_ Total projected area of all obstacles in the bed
° ~ Cross-section of the filter normal to the gas flow
where:
= diameter of the particle removed
from the gas stream
Do diameter of the obstacle
P density (mass) of the particle
g = local acceleration due to gravity
\i viscosity of the gas
v
-------
15
SECTION 15
The Gravity Spray Tower
-------
THE GRAVITY SPRAY TOWER
I IMPINGEMENT OF GRAVITATIONAL
SPRAY DROPS
A In a gravitational spray unit, there are a
number of liquid spherical obstacles
(droplets) falling in an empty tower by
the action of gravity in the path of rising
particles.
B The relationship between_impingement
target efficiency (nj) and, ^oS j i
• Vp/o
shown in Figure 1.
It is seen that the maximum efficiency
for the smaller particle sizes (say
less than 5(i) occurs for droplet size
of about SOOp.: and that for larger
particle sizes, the efficiency varies
little over the range of droplet sizes
500 to lOOOp..
Thus in gravitational spray towers,
there is little point in using very
fine spray sizes even if such were
available.
v i , in gravitational spray towers,
is the difference in the free-falling
velocities (Stokes1) of the droolets and
the: particle.
In practice, since the free-falling
velocity of the particle is sjnall com-
pared to the droplet, I Vp/o j may be
taken as the free-falling velocity of the
droplet.
II EFFICIENCY
A Inspection of Figure 2 shows that the ef-
ficiency £f a gravitational spray tower is
very low for particles below 1-2 microns.
B Figure 3 illustrates a size-efficiency curve
for a large industrial spray tower handling
70, 000 cfm. The tower is 22 ft in diameter,
66 ft high. Pressure drop is less than
1" water.
Krorn Figure 1, it is evident that for high
ctil!i>< tioi: efficiency by impingement
>;<•:•'• mils', be a small obstacle (DQ) and
n high relative velocjty i v / _~] between
the obstacle and partiele.
I [n
-------
TI (spheres) = 0
at o8
V , f , ,
P/o p(s)
= 24
nT (cylinders) = 0
V
/ / s
p/o p(s)
Figure 1.
-------
100
c
-------
The Gravity Spray Tower
o
o
rc
o
rf
20 L
\6
50
33
Particle Size, Microns
Size-efficiency Curve for Spray Tower
Figure 3
-------
. The Gravity Spray Tower
Hence, the dirty water may be re-
circulated until it contains quite a high
concentration of trapped dust particles.
a Therefore, there is a saving of
water, and perhaps a simplification
of effluent treatment and ultimate
waste disposal.
VI PRESSURE DROP
A The pressure drop is very small (less
than 1 in. w. g.)
Draft loss-
Efficiency-
less than 1" w. g.
very low for below
l-2n
Particle concentrations - relatively high (over
5 gr/cu ft)
Particle composition solid, liquid. Some
problems with corrosion.
Water usage •
about 18 gal/1000 cu ft
VII PERFORMANCE DATA
Gas flow over 70, 000 cfm
Gas temperature often used as pre-
cooler. Gas tem-
perature over 2000°F
may be reduced to
275°F.
Gas velocity about 3-5 fps
Treatment time about 20-30 seconds
REFERENCES
1 Stairmand, C. J. Dust Collection by Im-
pingement and Diffusion. Paper read
at the Inaugural Meeting of the Midland
Branch of A. Inst. P. Birmingham,
England. October 14, 1950.
2 Stairmand, C. J. The Design and Per-
formance of Modern Gas-Cleaning Equip-
ment. Paper read before the A. Inst.
P. London. November, 1955.
-------
16
SECTION 16
Venturi Scrubbers
-------
VENTURI SCRUBBERS
I MECHANISM OF PARTICLE REMOVAL
A Since for high collection efficiency of fine
particles by impingement there must be a
small obstacle (DQ) and high velocity of
approach of the gas stream relative to the
obstacle (vD/o), attempt is made to ap-
proach this ideal by.
1 An arrangement in which very small
water droplets (upon which the particles
impinge) are formed by the gas flow so
that the droplets are initially at rest
at the time of impact with the particles.
Even during the period of acceleration
of the droplet, high relative velocities
will be maintained since the particles
move at the velocity of the gas stream.
a Such an arrangement is incorporated
in the Venturi scrubber. See
Figure 1.
II OPERATION (Figure 1)
A Collection of Particles upon the Droplets
1 In the Venturi scrubber, the particulate-
laden gas passes through a duct which
incorporates a Venturi scrubber.
2 At the throat, high gas velocities of the
order of 200-600 fps are attained.
3 Coarse water spray is injected into the
throat by way of radial jets in quantities
of 5 to 7 gpm per M cfm of gas.
4 The high gas velocities at the throat
immediately atomize the coarse water
spray to fine droplets (about 50 microns).
5 Since, at their genesis, these fine drop-
lets are initially at rest relative to the
particles in the gas stream, it is at this
moment that collection efficiency is at
its maximum. (vr>fo> is maximum.
a The atomized droplets, being fine,
rapidly accelerate to the velocity
of the carrier gas; but even during
this short period, relative velocities
will be high and effective collision
between droplet and particle will
take place.
It is during the period before the
droplets attain the same velocity
as the gas stream that any relative
velocity between the droplets and
particle is obtained. For example,
a 100-micron droplet introduced
into a gas stream moving at 100
fps would accelerate to 90% of the
gas velocity in 16 inches; a 20-
micron droplet would reach 90%
of the gas velocity in 2 inches.
B Removal of the Dirty Droplets
1 As the gas decelerates after passing
through the throat, agglomeration of
the particle-laden droplets takes
place.
2 The large agglomerates are readily
removed by a cyclonic separator.
Ill EFFICIENCY
A Effect of Pressure Drop on Efficiency
1 The higher the pressure drop, the
higher the removal efficiency of
particles. See Figures 2 and 3.
2 Pressure drops across the Venturi
of 25-30 inches of water gage may be
expected.
3 Pressure drop can be increased (and
hence efficiency can be increased)
simply by increasing the gas velocity
and/or the water injection rate. See
Figure 4.
PA. C. pm. 68a. 5. 61
-------
Venturi Scrubbers
CO«E BUSIER DISC
TIIGmill C«S l»l!I
• HER WTEI
OUTLET INLET
CONTABIIUTED '/
MS KLET
Cyclone scrubber
Orifice scrubber
Venturi scrubber
Figure 1
-------
Venturi Scrubbers
100
95 •
10
20
30
40
Venturi Pressure Drop (in. w. g. )
Curve A: Rotary iron powder kiln
B: Lime kiln, asphalt plant
C: Iron cupola
D: Phosphoric acid plant (acid mist)
E: Incinerator (sodium oxide fumes)
Figure 2
(5)
10 20 30 40 50
Venturi Pressure Drop (in.w. g. )
Curve A: Cupola gases
B: Blast furnace gases
Figure
— , ^3
3
c 20
•H
£ 15
o
t-i
7 10
0)
$ s
to
CO
CU
/
/
^
^~
/
f
^
S~
'
.
/
S
J
s
.---
S
s
^^
^
*~
^*
s*
£ 0 2 4 6 8 C
Water/Gas Ratio, Gal/1000 cu. ft.
Relation Between Pressure- loss
and Water Usage in Venturi
Scrubber
Figure 4
4 When gas cleaning requirements change,
the only adjustment necessary to the
Venturi scrubber, in most cases, is
in the flow of scrubbing liquid to in-
crease the pressure drop. Thus higher
cleaning efficiency is accomplished
without modification or addition.
B Effect of Particle Concentration on
Efficiency
1 If the number of water droplets is held
constant and the number of particles
(concentration) is increased, the number
of collisions would be expected to in-
crease. In other words, collection
efficiency should increase as loading
increases.
2 This increase, however, is due not
only to the increased chances of particle
collision with droplets, but also due to
collisions between the particles
themselves.
C Size-Efficiency
1 The Venturi scrubber approaches
100% for all particles larger than 1. 5
to 2 microns.
2 Figure 5 shows a size-efficiency curve
for a Venturi scrubber(l). Sizes above
2-microns were obtained on special
-------
•u
d
-------
Venturi Scrubbers
TABLE 1
TYPICAL PERFORMANCE DATA FOR VENTURI SCRUBBER
(5)
Source of Gas
IRON i. STEEL INDUSTRY
(jidv liim Cupola
U*VKI'" Slrd L'uiivei Itt
MtcM Upi'ii Hc.1t 111 1 urn-' r it U[n
SIcL'l Opin Ik'Jilli FJMUJCI
(Oxygen Lanced)
Blast Furnace (lion)
Electric Fuinace
Electric Furnace
Rotary Kiln— Iron Reduction
Crushing & Screening
CHEMICAL INDUSTRY
Acid— Humidified SO,
(a) Scrub with Water
(b) Scrub with 40% Acid
Acid Concentrator
Copperas Roasting Kiln
Chlorosulfonic Acid Plant
Dry Ice Plant
Wood Distillation Plant
TiCI, Plant, TiO, Dryer
Spray Dryers
Flash Dryer
Phosphoric Acid Plant
NON-FERROUS METALS INDUSTRY
Blast Furnace (Sec. Lead)
Reverberalory Lead Furnace
Ajax Furnace — Aluminum Alloy
Zinc Sintering
Reverberatory Brass Furnace
MINERAL PRODUCTS INDUSTRY
Lime Kiln
Lime Kiln
Asphalt Stone Dryer
Cement Kiln
PETROLEUM INDUSTRY
Catalytic Reformer
Acid Concentrator
TCC Catalyst Regenerator
FERTILIZER INDUSTRY
Fertilizer Dryer
Superphosphate Den & Mixer
PULP & PAPER INDUSTRY
Lime Kiln
Lime Kiln
Black Liquor Recovery Boiler
MISCELLANEOUS
Pickling Tanks
Boiler Flue Gas
Sodium Disposal Incinerator
A
Contaminants
Iron, Coke. Silica Dusl
lion Ou'U.1
Uon & /tin Ouilp
lion Oxide
lion Die t. Coke Dust
Ferro-Manganese Fume
Feiro-Silicon Dust
•ron, Carbon
Tacomte Iron Ore Dust
H,SO. Mist
H,SO, Mist
H.-SO, Mist
H^O, Mist
Amine Fog
Tar & Acetic Acid
TiOv-HCI Fumes
Detergents. Fume & Odor
Furfural Dust
H.PO, Mist
Lead Compounds
Lead & Tin Compounds
Aluminum Chloride
Zinc & Lead Oxide Dusts
Zinc Oxide Fume
Lime Dust
Soda Fume
Limestone 4 Rock Dust
Cement Oust
Catalyst Dust
H,SO, Mist
Oil Fumes
Ammonium Chloride Fumes
Fluorine Compounds
Lime Dust
Soda Fume
Salt Cake
HCI Fumes
Fly Ash
Sodium (hide Fumes
pproximat
Size Range
(Microns)
1 II!
b ?
OS 1
5-,'
.5-20
1-1
.1-1
.5-50
.5-100
—
—
—
—
.5-1
—
.1-1
—
.1-1
.1-.8
.1-.9
.1-1
.05-.5
1-50
.3-1
1-50
.5-55
.5-50
—
.05-1
—
.1-50
.1-2
—
—
.1-3
.3-.1
s Loading
(Grains/ cf)
Inlet Exit
8 10
5-1 5
1-6
3-24
10-12
1-5
3-10
5-25
303'
406'
136*
198'
756'
25'
1080'
1-5
—
1-1.5
192'
2-6
1-2
3-5
1-5
1-8
5-10
.2-5
5-15
1-2
.09
136'
756'
.1-.5
309'
5-10
2-5
4-6
25%
1-2
.5-1
05-15
05-08
03-06
.Ol-.O?
.008- Ob
.04-.03
Average
Removal
Efficiency (%)
yb
98 D
35
99
99
99
.1-.3 92
.1-.3
.005-.01
1.7'
2.8'
3.3'
2.0'
7.8'
2.0-
58.0'
.05-.!
—
.05-.08
3.8'
.05-.15
.12
.02-.05
.05-.!
.1-.5
.05-. 15
.OI-.05
.05-. 15
.05-.!
.005
3.3*
8.0'
.05
5V
.05-. 15
.01-.05
.4-.6
23'
.05-.08
.02
99
99.9
99.4
99.3
97.5
99
98.9
90+
95
95
95
95+
98+
99
91
95
98
95
99+
99
98+
97 +
95+
97.5
98+
85 +
98+
99+
99
90
90+
98
9S
* Milligrams per cubic ft
\«t<-: The efficiencies shown abare are average values for a particiitar
-------
Venturi Scrubbers
silica dust powders and those for the
smaller sizes on dispersed non-patho-
genic bacteria. This size-efficiency
curve suggests a very high efficiency
for a comparatively simple piece of
equipment.
D Overall Efficiency
Table I shows some efficiencies of collec-
tion experienced by various installations.
IV ENERGY USAGE
A Since pressure drops of 30 inches water
gage correspond to 120 kwh per million
cubic feet of gas cleaned, efforts have
been made to reduce the pressure drop.
However, if pressure drop is reduced,
there is a tendency to reduce the efficiency
also.
B Additional high energy usage results from
the method of injecting water into the
Venturi throat. See Figure 6.
THE TYPE 8-V VK1STTUBI
The Chemico Type S-F Venturi Scrubber is particularly
recommended for these hard-to-handle situations: re-
moval of "sticky" solids from gases; recycling of
heavy slurries where water supplies are limited; and
recovery of process materials in concentrated form.
In the S-F Venturi, scrubbing liquid is introduced
through troughs at the top of the unit. The liquid
flows downwardly in a continuous film along the slop-
ing walls to the deflecting lips, which direct it across
the throat of the Venturi to be atomized by the force
of the high velocity gas.
TUB TYPE P-A VENTUSU
The Chemico Type P-A Venturi Scrubber is most ef-
fective in the very difficult applications requiring
efficient removal of sub-micron dust, fume, and mist
particles.
Chemical Construction Corp.
Figure 6
-------
Venturi Scrubbers
V PERFORMANCE DATA
Gas flow
Gas velocity through throat'
Pressure loss
Gas temperature
Overall efficiency
High efficiency on dusts with mass median
size greater than •
Humid air influence on efficiency-
Water usage
-- 200 to over 145, 000 cfm
--200-600 fps
--up to 25-30 in. water gage
— "unlimited"
usually high (97 99-<-%i
-5-7 gpm o.f water per
M cfm gas
REKKRENCES
1 Stairmand, C. J. The Design and Per-
formance of Modern Gas-Cleaning
Equipn."nt. Paper read before
A.Inst. P. London. November, 1955.
2 Nicklen, G. T. Some Recent Developments
and Applications of Scrubbers in In-
dustrial Gas Cleaning. Proceedings
APCA, V'nd Annual Meeting APCA.
Los Angeles. June, 1959.
3 Jones, W. P. Development of the Venturi
Scrubber. Ind. Eng. Chem. Nov. 1949.
4 Basse, B. Gases Cleaned by the Use of
Scrubbers. Blast Furnace and Steel
Plant. Nov. 1956.
5 Chemico Gas Scrubbers for Industry.
Bulletin M-104, Chemical Construction
Corporation, 525 West 43rd Street,
N.Y. 36, N.Y.
6 Venturi Scrubbers for Industry, Bulletin
M-103A, Chemical Construction Cor-
poration, 525 West 43rd St. N. Y. 36, N. Y.
-------
17
SECTION 17
Collectors with Self-Induced Sprays
-------
COLLECTORS WITH SELF-INDUCED SPRAYS
I MECHANISM OF PARTICLE COLLECTION
A In this equipment, the particle collection
zone is a spray curtain which is induced
by the gas flow itself through a specially
designed orifice. (The spray curtain is
followed by a spray eliminator).
B The Collection of Particles
1 Normal gas velocity of about 50 fps
creates droplets about 320y.
2 Collection of particles is mostly by
Impingement on the droplets during
the free-falling period of the droplets
and also during the period of the accel-
eration of the droplets from rest (when
high relative velocities are available).
IL APPLICATION
A Since there is an absence of ledges,
moving parts, and restricted passages,
these units are especially adapted to
materials like:
1 Magnesium and explosive dusts
2 Sticky or linty materials like metallic
buffing exhausts
III PERFORMANCE DATA
Efficiency - See figure 2 (page 3)
Water usage - 10-40 gal/1000 cfm gas cleaned
(Much or all of this water may
be recirculated).
Sensitivity - not particularly sensitive
to cfm change (at least within
+ 25% of the design rate).
Concentration - high concentrations (40 grains/
ft3) (There are no fine clear-
ances to cause chokage)
Maintenance - The whole apparatus is well
irrigated and periodic hosing-
down of the interior is easily
done. There is an absence of
moving parts. There may be cor-
rosion difficulties.
PA.C.pm.70.9.60
-------
Collectors With Self-Induced Sprays
Figure 1
-------
CD
c
n
o
100
40
20
12
16
20
24
28
32
36
40
o
o
:r
CO
ID
3
Q.
C
Particle Size, Microns
Size-Efficiency Curve for Self-Induced Spray Collector
Figure 3
CO
ID
-------
Collectors with Self-Induced Sprays
REFERENCES
First, M., et. al., "Performance Character-
istics of Wet Collectors," NYO - 1587
Waste Disposal, Harvard University. 1953
Stairmand, C.J., "Mist Collection by Im-
pingement and Diffusion", paper read
at the Inaugural Meeting of Midland
Branch of A. Inst. P., Birmingham,
England. Oct. 14, 1950.
Stairmand, C.J. "The Design and Performance
of Modern Gas-Cleaning Equipment," paper
read before the A. Inst. P., London.
November, 1955.
Kane, J.M. "Operation, Application, and
Effectiveness of Dust Collection Equip-
ment," Heating and Ventilating. Aug.
1952.
Nicklen, G.T. "Some Recent Developments
and Applications of Scrubbers in In-
dustrial Gas Cleaning," Proceedings,
APCA, 52nd Annual Meeting, Los Angeles.
June, 1959.
Magill, P.L. Air Pollution Handbook, Mc-
Graw-Hill Book Co., Inc. 1956.
-------
SECTION 18
Impingement Type Scrubbing Tower
-------
IMPINGEMENT TYPE SCRUBBING TOWER
I TYPES OF SCRUBBING TOWERS
A There are two types of scrubbing towers
commonly used:
1 Those employing impingement target
plates
2 Those employing beds of spherical
obstacles
B Particle Concentration
1 An important feature of this design is
freedom from chokage in spite of the
small holes in the orifice plates. This
is due to:
a The very violent circulation induced
below the targets by the air jets,
and
II TOWER WITH TARGET PLATES (Figure 1)
A Construction and Operation
1 This type of scrubber is a tower con-
sisting of a vertical shell in which are
mounted a large number of equally
spaced, circular, perforated (orifice)
plates.
a At one side of each orifice plate, a
conduit, called a downspout, is pro-
vided to pass the liquid to the plate
below.
b At the opposite side of the orifice
plate, a similar conduit feeds liquid
from the plate above.
2 Over each hole (about 3/1G" diameter)
in the orifice plate, a target plate is
positioned.
a The motion of the gas past the edge
of the holes in the orifice plate re-
sults in the formation of spray drop-
lets (about lOOp.). These droplets
are initially at rest and provide an
effective relative velocity between
particle and droplet for good
impingement.
b The particle-laden gas passes through
the holes in the plate and the particles
impinge upon the atomized droplets
and on the target plates.
b A preliminary spray zone which
helps to keep the orifice plate free
from deposits.
2 Concentrations of 40 grains/ft^ can
readily be handled.
C Efficiency
1 An example of a size-efficiency curve
is shown in Figure 2.
D Pressure Drop
1 Each plate imposes a pressure drop of
3 in. w. g.
Ill TOWERS WITH BEDS OF SPHERES
A Construction and Operation
(An example of a scrubbing tower with
beds of spheres is shown in Figure 3)
1 Large particles are removed by im-
pingement on wet surfaces and contact
with water spray in an area below the
filter bed.
2 Particle-laden gas then passes upward
through a bed of spheres. In the inter-
stices of the bed, the particles are
subjected to increased velocities which
results in their efficient impingement
upon the surfaces of the spheres.
PA. C. pm. 73. 9. 60
-------
Impingement Type Scrubbing Tower
Target plate
Water
level
Gas flow
ARRANGEMENT OF "TARGET PLATES"
IN IMPINGEMENT SCRUBBER
Water droplets atomized
at the edges of orifices
Downspout to
lower stage
MECHANISM OF IMPINGEMENT SCRUBBER
IMPINGEMENT
BAFFLE STAGE
AGGLOMERATING
SLOT STAGE
Peabody Engineering Corp.
Figure 1
-------
OJ
O
i-
o>
Q.
O
c:
CD
QJ
C
O
U
O)
100
20
14 6 8 10 12
Particle size, microns
Size-efficiency curve for wet-impingement scrubber
Figure 2
14
CTQ
0>
3
rt>
3
H
en
o
a
-------
Impingement Type Scrubbing Tower
Fan
Transition
Place
Main Body
C Pressure Loss
1 Pressure loss is 4-6 in. w. g.
D Efficiency
1 Efficiency is high on two micron-sized
particles and above.
E Water Consumption
1 Fresh water:
cleaned
per 1000 cfm gas
Sludge
Ejector
Overflow
2 Recirculated water: 3 gpm per 1000
cfm gas cleaned (Scrubbing liquid can
have high solids content).
F Capacity
1 Units handle 500 to 40, 000 cfm.
Figure 3
National Dust Collector Corp.
3 The high gas velocity through the
interstices of the packed spheres also
results in pulling water upward with
sufficient force to disintegrate the
water streams into a turbulent mist in
the zone above the filter bed. Here,
ultra- fine particles are trapped by the
mist and constantly flushed downward.
4 Mist carried by the upward flowing
cleaned gas is removed by passage
through a bed packed with porcelain
saddles.
B Particle Concentration
1 Such units have self- cleaning action
and there is freedom from build-up of
solids and ease of cleaning.
2 Concentrations of about 40 grains
are readily handled.
_, Recirculation
Pump
Ejector REFERENCES
Pan
1 First, M. et al. Performance Character-
istics of Wet Collectors. NYO-1587
Waste Disposal, Harvard University.
1953.
2 Stairmand, C. J. Mist Collection by Im-
pingement and Diffusion. Paper read
at the Inaugural Meeting of Midland
Branch of A. Inst. P. Birmingham,
England. October 14, 1950.
3 Stairmand, C. J. The Design and Per-
formance of Modern Gas-Cleaning
Equipment. Paper read before the A.
Inst. P. London. November, 1955.
4 Kane, J. M. Operation, Application, and
Effectiveness of Dust Collection Equip-
ment. Heating and Ventilating. August,
1952.
5 Nicklen, G. T. Some Recent Development
and Applications of Scrubbers in In-
dustrial Gas Cleaning. Proceedings
APCA, 52nd Annual Meeting, Los
Angeles. June, 1959.
6 Magill, P. L. Air Pollution Handbook.
McGraw-Hill Book Co. , Inc. 1956.
-------
19
SECTION 19
Wet Centrifugal Collectors
-------
WET CENTRIFUGAL COLLECTORS
I TYPES OF WET CENTRIFUGAL
COLLECTORS
A Irrigated Types
1 These rely upon the throwing of parti-
Cjles against wetted collected surfaces,
such as wetted walls or impingement
plates by centrifugal action.
B Spray Chamber Types
1 These depend upon impaction of the
particles upon spray droplets and the
subsequent precipitation of the "dirty"
spray droplets upon the wall of the
unit by centrifugal action.
II IRRIGATED TYPES (Figure 1)
WET CENTRIFUGAL COLLECTORS
Clean air outlet G
Entrainment separator H
Water inlet
Impingement plates
Dirty air inlet I
Disintegrator
Inspection door
Wet cyclone for
collecting heavy
material
Water and sludge
drain
Figure 1
Ami-rir.-in Air Filter Co.
A The efficiency of a centrifugal collector-
may be increased by irrigating its walls,
if the attendant disadvantages of a wet
system can be tolerated.
B Water distribution may be from low
pressure nozzles or gravity flow.
C Performance Data
Water rates 3-5 gal/1000 cfm of
gas treated
Draft loss 2\ to 6"
Draft loss sensitivity
to cfm change as (cfm) 2
High efficiency on
particles of mass
median greater than. . l-5(i
Efficiency sensitivity
to cfm change ....
Humid air influence
on efficiency ....
Gas temperature . .
yes
none
"unlimited"
III CYCLONE SPRAY CHAMBERS (Figure 2)
A Operation
1 The dust-laden gas enters tangentially
at the bottom and spirals up through
a spray of high velocity fine water
droplets.
2 The dust particles are collected upon
the fine spray droplets which are then
hurled against the chamber wall by
centrifugal action.
3 An unsprayed section above the nozzles
is provided so that the liquid droplets
containing the collected particles will
have time to reach the walls of the
chamber before the gas stream exits.
pr,i.
-------
Wet Centrifugal Collectors
PLEASE-ANTHONY CYCLONIC SPRAY SCRUBBER
cleaned gas
core buster disc
spray manifold
tangential
gas inlet
swinging inlet
damper
gas inlet
water inlet
water outlet
Figure 2
Chemical Construction Corp.
B Efficiency
1 Efficiency of dust removal is given by:
3n rWH
E
1 -e
2D0Q
where:
K = efficiency of collection
r) individual droplet efficiency
r radius of the cyclone (the
length of the path of the droplet)
W = volume rate of liquid through
the nozzle
D-., - diameter of the droplets
Q
volume rate of carrier gas
II = height of tower (The drops
should not be made too small
since entrainment may occur,
requiring an increase in the
height of the tower)
2 Operating Conditions
Gas flow 500-more than 25, 000
cfm
Gas velocity into
cyclone up to 200 fps
Separation factor . . . 50 to 300
Efficiency 97+% on dust above l(j.
High efficiency on
particles of mass
median greater than . 0. 5 5p-
Efficiency sensitive
to cfm change .... yes
Draft loss 2-6"w. g.
Draft loss sensitivity
to cfm change .... as (cfm)2
Water usage 3-10 gal/1000 cu ft of
gas cleaned
Humid air influence
on efficiency none
Gas temperature . . .pre-cooling of high
temperature gases
necessary to prevent
rapid evaporation of
fine droplets.
Power requirements . 1 to 3 HP/1000 cfm of
gas
REFERENCES
I First, M. , etal. Performance Char-
acteristics of Wet Collectors. NYO-
1587 Waste Disposal. Harvard
University. 1953.
2 Stairmand, C. J. Mist Collection by
Impingement and Diffusion. Paper
read at the Inaugural Meeting of
Midland Branch of A. Inst. P. Birming-
ham, England. October 14, 1950.
3 Stairmand, C. J. The Design and Per-
formance of Modern Gas-Cleaning
Equipment. Paper read before the A.
Inst. P- London. November, 1955.
-------
Wet Centrifugal Collectors
Kane. .1. M. Operation, Application, and
Effectiveness of Dust Collection Equip-
ment. Heating and Ventilating.
August, 1952.
Nicklen, G. T. Some Recent Developments
and Applications of Scrubbers in
Industrial Gas Cleaning. Proceedings
APCA, 52nd Annual Meeting. Los
Angeles. June, 1959.
Mi gill, P. L. Air Pollution Handbook.
M:Graw-Hill Book Co. , Inc. 1956.
-------
20
SECTION 20
Wet Dynamic Precipitator
-------
WET DYNAMIC PRECIPITATOR
I OPERATION (Figure 1)
A Wet dynamic precipitators combine the
dynamic forces of a rotating fan to cause
the particles to impinge upon numerous
specially shaped blades.
B A film of water is maintained on the blades
by spray nozzles.
High efficiency on
particles with mass
median greater than.
Efficiency sensitivity
to cfm change ....
.no
Water usage.
0. 5 to 1 gpm/1000
cfm gas
di rt and water
discharged at
blade tips
clean air
outlet
dirty air
inlet
water spray
nozzle
WET-TYPE DYNAMIC PRECIPITATOR
water and
sludge outlet
Figure 1
REFERENCES
1 First, M, , et al. Performance Character-
istics of Wet Collectors. NYO-1587
Waste Disposal. Harvard University.
1953.
2 Stairmand, C. J. Mist Collection by
pingement and Diffusion. Paper read
at the Inaugural Meeting of Midland
Branch of A. Inst. P. , Birmingham,
England. October 14, 1950.
3 Stairmand, C. J. The Design and Per-
formance of Modern Gas-Cleaning
Equipment. Paper read before the A.
Inst. P. London. November, 1955.
II PERFORMANCE DATA
PITSSU rc- drop. . . .
a function of mechan-
ical efficiency
Usually less than ]-
in. w. g.
Pressure drop
sensitivity to cfm
change
. . . '. . . . a function of mechan-
ical efficiency.
Particle concentration. .less than 1 grain/ft .
(For heavy loading,
a pre-cleaner may
be used to lighten the
load on the unit).
4 Kane, J. M. Operation, Application, and
Effectiveness .of Dust Collection
Equipment. Heating and Ventilating.
August, 1952.
5 Nicklen, G. T. Some Recent Developments
and Applications of Scrubbers in In-
dustrial Gas Cleaning. Proceedings
APCA, 52nd Annual Meeting, Los
Angeles. June, 1959.
6 Magill, P. L. Air Pollution Handbook.
M;Graw-Hill Book Co. , Inc. 1956.
PA. C. pm. 71. 9. 60
-------
SECTION 21
Disintegrator Scrubbers
-------
DISINTEGRATOR SCRUBBERS
I MECHANISM OF PARTICLE COLLECTION
A Since for high collection efficiency there
must be a small obstacle (D ) and a high
relative velocity between the obstacle and
particle (v , ), attempt is made to ap-
proach this^laeal by:
1 Shooting water drops at the particles
so that a high relative velocity (v . )
will be obtained (even if such velocities
are maintained for short periods) and
arranging that this be done so that a
very large number of impacts will be
achieved.
B Such action is incorporated in the disinte-
grator scrubber (Figure 1).
II OPERATION
A A disintegrator scrubber consists of an
outer casing containing alternate rows of
stator and rotor bars, the relative velocity
between adjacent bars being of the order
of 200 - 300 fps.
B Water is injected axially and is effectively
atomized into fine droplets (say 25y) by
the rapidly rotating vanes.
C The dust-laden gas also enter axially and
passes through the dense spray zone where
the particles are subjected to intense
bombardment by the water droplets.
Water inlets
Stators
Dirty gas inlet
Rotors
Clean gas
Exit
Effluent
Figure 1
PA.C.pm.69.9.60
-------
Disintegrator Scrubbers
III. PERFORMANCE DATA
Efficiency.
Pressure drop.
Energy usage.
Water consumption.
highly efficient. See figure 2
for a size-efficiency curve.
.less than 1-in. w.g.
high power requirements. Total
power consumption may be 16—20
HP per 1000 cfm gas cleaned. This
power is largely expended in atomizing
and accelerating the water.
.usually preceded by conventional
collectors as cyclones and scrubbers
to insure that low concentrations of
the order of % to % grains per cu ft.
are presented to the unit. These pre-
cautions are necessary to avoid build-
up in the disintegrator, which, run-
ning at high speed with fine clearance,
is particularly susceptible to trouble
if operated under unsuitable conditions.
-------
Disintegrator Scrubbers
100
40
20
6789 10
Particle Size, Microns
11
12 13
14 15
Size-Efficiency Curve for Disintegrator Scrubber
Figure 2
REFERENCES
1 First, M., et al. Performance Character-
istics of Wet Collectors. NYO-1587
Waste Disposal, Harvard University. 1953.
2 Stalrmand, C.J. Mist Collection by Im-
pingement and Diffusion. Paper read
at the Inaugural Meeting of Midland
Branch of A. Inst. P. Birmingham,
England. October 14, 1950.
3 Stairmand, C.J. The Design and Performance
of Modern Gas-Cleaning Equipment. Paper
read before the A. Inst. P. London.
November, 1955.
Kane, J. M. Operation, Application, and
Effectiveness of Dust Collection Equip-
ment. Heating and Ventilating. August,
1952.
Nicklen, G.T. Some Recent Developments
and Applications of Scrubbers in In-
dustrial Gas Cleaning. Proceedings
APCA, 52nd Annual Meeting. Los Angeles,
June, 1959.
Magill, P.L. Air Pollution Handbook
McGraw-Hill Book Co., Inc. 1956.
-------
22
SECTION 22
Fabric Filtration
-------
FABRIC FILTRATION
G. W. Walsh*
I FABRIC FILTRATION EQUIPMENT
TYPES AND SIZES
A Basic Unit
In its simplest form, the industrial fabric
filter (baghouse) consists of a woven or
felted fabric through which dust laden
gases are forced. A combination of factors
result in the collection of particles on the
fabric fibers. When woven fabrics are
used, a dust cake eventually forms which,
in turn, acts predominantly as a sieving
mechanism. When felted fabrics are used,
this dust cake is minimal or non-existent.
Instead, the main filtering mechanisms
are a combination of inertial forces, elec-
trostatic forces, impingement etc. , as
related to individual particle collection on
single fibers. These are essentially the
same mechanisms experienced in "Air
Cleaning Type" filters.
As particulates are collected, pressure
drop across the filtering media increases.
Because of fan limitations, the filter must
be cleaned. This cleaning is accomplished
"in-place" since the filter area is usually
too large and time between cleanings too
short to allow for filter replacement or
cleaning external to the baghouse.
In order to meet the challenge of a variety
of operating conditions and applications, a
multitude of proprietary designs exist.
Essential differences are related to:
1 Fabric
2 Cleaning mechanism
3 I^quipment geometry
4 Mode of operation
Depending on the above factors, equipment
will follow one of three systems, as shown
in Figure 1.1.
Figure 1. la shows "bottom feed" units in
which the dust-laden gas is brought through
the baghouse hopper and then to the interior
of the filter tube. Obviously, a portion of
the dust is removed in the hopper and never
reaches the fabric.
Figure 1. Ib shows "top-feed" units, in
which the dust-laden gas enters the top
of the filter tubes.
Figure 1. Ic shows units wherein the gas
passes from the outside of the filters to
the interior, or clean-air side. With this
arrangement, the dust inlet can be located
in many positions. The fabric can be
formed in a tubular shape, or it may be
in an envelope form.
B Baghouse Operation
1 Intermittent operation
The fundamental principles of operation
are embodied in the discontinuous-type
or intermittent units. For such filters
the entire area collects dust for a pre-
set filtration time; at the end of this
time the unit is taken out of service and
the whole area is cleaned of collected
dust. Examples of such collectors are
shown in Figures 1.2, 1. 3, and 1. 4.
These collectors are primarily utilized
for the control of small volume opera-
tions such as grinding, polishing, etc. ,
and for aerosols of a very coarse nature.
They are also used extensively for pilot-
plant studies and research. Many of
these baghouses are of the so-called
"unit" type, in which the fan and filter
are contained in a single piece of
equipment.
2 Continuous operation
For most air pollution control installa-
tions and major dust control problems
it is desirable to utilize collectors which
*Chief, Air Pollution Training, Training Program, SEC
pa.c.pm.90.5.6fo
-------
IP
cr
\4
o
13
(a) Bottom Feed (b) Top Feed (c) Exterior Filtration
Figure 1. 1 POSSIBLE FILTERING SYSTEMS
-------
Fabric Filtration
MICMAMttM
aife 18
(American Wheelabrator and Equipment Co.)
Figure 1. 2
(Pangborn Corporation)
Figure 1. 3
(W. W. Sly Manufacturing Co.)
Figure 1. 4
-------
Fabric Filtration
allow for continuous operation. This
end is accomplished by arranging
several filter areas in a parallel flow
system, and cleaning one area at a time
according to some pre-set mode of
operation. Examples of these control
devices are shown in Figures 1.5, 1. 6,
and 1.7.
In a multicompartment baghouse the
basic filter area is a compartment or
section (see Figure 1.5). Each section
or compartment is essentially the Same
as a discontinuous unit. In a reverse
flow baghouse the basic filter is one
envelope or filter bag (Figure 1.6).
Whereas a small portion of one filter
tube is the basic filter area in a re-
verse jet baghouse (Figure 1. 7).
C Filter Cleaning
The heart of any fabric collector is the
technique employed to remove dust from
the fabric. There are two general types
of cleaning; the first involves flexing the
fabric and the second involves a reverse-
flow of clean air. A breakdown of these
types is as follows, according to com-
monly accepted terminology:
1 Fabric flexing
a Mechanical shaking and rapping
This type of cleaning generally in-
volves the use of a "rocker arm-
lever assembly" to produce a motion
to the top of the filter tubes. The
motion may be generally horizontal
(sometimes concave upwards, some-
times concave downwards), vertical,
or cover a 90° arc from bottom to
top of swing. Vertical motion is
sometimes accomplished by rapping.
b Sonic cleaning
This type of cleaning utilizes sympa-
thetic vibrations from sound waves
to dislodge dutJ from the cake.
Sound waves are generated at low
frequency by means of an air horn.
Figure 1. 8 shows a typical location
of sound producing horns on the
clean-air side of the filter tubes.
c Collapse cleaning
(1.1)
To clean filters by the "collapsing"
technique small reversals in pres-
sure are created, such that AP
from the dirty air side to the clean
air side is slightly negative. This
causes the filter tube to deflate, and
hopefully, the dust cake is discharged.
In some cases, the tube is slowly
collapsed and "popped" open. If
desired, the bags can be collapsed
several times per cleaning period.
Obviously, the baghouse must be
equipped with suitable valves and
ductwork to achieve AP reversal.
d Pressure-jet or pulse-jet cleaning
For this method a "bubble" of com-
pressed air is injected at the top of
the filter tube. A schematic of the
bubble as it travels down the tube is
shown in Figure 1. 9, in combination
with collapse cleaning.* ' ' Arrange-
ment of the system when filtering on
the bag exterior is shown in Figure
1. 10.
2 Reverse-air cleaning
a Reverse-jet
This mechanism employs a high
velocity (small volume) jet of com-
pressed air, blown back through the
fabric, to dislodge collected dust.
Figure 1. 7 shows the usual arrange-
ment of mechanical cleaning devices.
In the typical "reverse-jet" filter
unit, cleaning can be conducted con-
tinuously, so that pressure differen-
tial across the unit tends to remain
constant.
b Reverse-flow cleaning
Both filter tubes and envelope col-
lectors can be cleaned by a reverse
-------
Fabric Filtration
incoming gases
Filtering
Fil tering
Fi 1 tering
J*LJ
to fan
All compartments filtering, dampers open
incoming gases
incoming gases
Shaking
Fi 1 tering
Fi 1 tering
to fan
One compartment shaking, balance filtering
incoming gases
1
\
/N
Fi 1 tering
s/
_ *
h
Shaking
h
\
Fi 1 tering
>>
/
Fil
S
^
V
/ v
te
\)
t
h
r
>
''ing
/
/^
Fi 1 tering
\/
J*
Shaking
Y
to ffln to fan
One compartment shaking, balance filtering One compartment shaking, balance filtering
Figure 1.5 TYPICAL PARALLEL FLOW SYSTEM FOR A
CONVENTIONAL MULTICOMPARTMENT BAGHOUSE
-------
Fabric Filtration
INLfl
HOTAHY DISCHARK
VALViS
Figure 1.6 CONTINUOUS OPERATING ENVELOPE COLLECTOR
WITH REVERSE-FLOW CLEANING MANIFOLD
-------
Fabric Filtration
REVERSE JET
(Koppers Co. , Inc. )
Figure 1. 7
-------
Fabric Filtration
Figure 1. 8 AIR HORNS WHICH AID IN CLEANING BAGS
_TL
Figure 1. 9 PRESSURE JET CLEANING
-------
Fabric Filtration
TO
IXHAUSTEH
A. FILTER CYLINDERS
B WIRE RETAINERS
C. COLLARS
D. TUIE SHEtT
[. VENTURI NOZZIE
F. NOZZLE Of ORIHCI
O. SOLENOID VAIVI
H. TIMER
J. AIR MANIFOLD
K. COUECTOH HOUSING
L. INLET
M. HOPPER
N. AIRLOCK
O. EXHAUST OUTLET
P. MANOMETER
O. UPPER PLENUM
COMPRESSED
AIR
SUPPLT AT
IOO P.S.I.O.
I
INDUCED n.OW
MATERIAL DISCHARO4
Figure 1. 10
flow of clean air. This would be at
low or atmospheric pressures, and
would utilize a much larger volume
of air than the reverse-jet action.
An arrangement used for envelope
collectors is shown in Figure 1.6.
In this case, the cleaning manifold
traverses the baghouse length,
cleaning one row of envelopes at a
time. Filtration, of course, is on
the outside of the envelopes. An
arrangement used for tubular bags
is shown in Figure 1.11. Note the
rings which are used to maintain
filter shape.
D Baghouse Size
Bag filter units are relatively large in so
far as dust collection equipment is con-
cerned. Equipment size, therefore, is of
importance to the buyer as a matter of
economics and feasibility.
For given manufacturers, the nominal
filter velocity is a major factor in deter-
mining equipment size. This does not
mean, however, that different types of
units vary in size according to their rated
nominal velocities. Other factors, such
as height limitations, bag spacing, and
the length to diameter ratio for tubular
bags, are also important. Table 1. 1 lists
approximate size ranges for various cate-
gories of fabric collectors, at nominal
velocities of 1 fpm and at nominal veloci-
ties considered as normal for the particular
collector.
E Filter Fabric
1 Types of fabric
Filter fabrics can be divided into the
woven or felted classifications. If
felted fabrics are used, filter cleaning
is limited to the pressure-jet and
reverse-jet classifications. When
woven fabrics are employed any clean-
ing technique may be used. In practice,
however, bag collectors cleaned by
the reverse-jet technique operate at
relatively high filter ratios. To ensure
high efficiency and to maintain low
pressure differentials these collectors
usually employ felted fabrics.
Woven fabrics can be sub-divided into
the following classes:
a Continuous-filament type, in which
the filaments used to form the fabric
strands are continuous in structure.
Such a fabric is characterized by a
smooth surface and absence of fibers
or tendrils, and can be constructed
only of synthetic materials.
b Texturized strand, in which the
fabric strands are mechanically
degraded or broken at the surface
to produce a fuzzy thread. The
texturized strand is usually woven
in the fill direction.
c Staple strand, in which the fabric
strands are formed from short
-------
Fabric Filtration
INSPECTION DOOR
DUST LADEN
AIR INLET
AIR REVERSAL VftlVE
IN NORMAL FILTERING
POSITION
CLEAN AIR
TO FAN
THIS COMPARTMENT
FILTERING
PARTITION
AIR REVERSAL VALVE
IN BACK-WASH POSITION
BACK-WASH AIR
THIS COMPARTMENT
IS BEING BACK WASHED
WITH CLEAN AIR.
ACCUMULATED DUST
DROPS INTO HOPPER
FILTER TUBES
HOPPER
UNIFLOW-BACKWASH DUST COLLECTOR*
Figure 1.11
*The Ducon Company, 147 East Second Street, Mineola, New York.
10
-------
Table 1.1 APPROXIMATE SIZE RANGES FOR FABRIC COLLECTORS
Collector
Reverse-Jet
(3)
Pressure- Jet
Conventional
tubular bags
Mechanical
Reverse
flow
Envelope
(fpm)
1.0
1.0
1.0
1.0
1.0
Collector volume
per 1,000 cfm
(ft3)
1,250
670
210 - 370
590
210 - 340
Collector
Floor-Area
per 1,000 cfm
(ft2)
57 - 294
111
26 - 50
30 •* 42
21 - 59
Uf(2)
' (fpm)
10
10
3
2
2
Collector volume
per 1,000 cfm
(ft3)
125
67
70 - 123
295
105 - 170
Collector
Floor-Area
per 1,000 cfm
(ft2)
5.7 - 29 . 4
11.1
8.7 - 16.9
15 - 21
10 . 5 - 29 . 5
(1) Does not include dust hopper.
(2) Common values for filter velocity.
A
7&
(3) As manufactured by Pulverizing-Machinery Company, N. J.
-------
Fabric Filtration
filaments. This type of construction
is necessary for the natural fibers.
The fabric is characterized by its
fuzzy appearance and forms a most
efficient filter because of the fibers
and tendrils which mat the surface.
2 Fabric properties
a Insofar as the materials of construc-
tion are concerned, prime factors
of importance are temperature limi-
tations, and chemical stability.
Other factors which should be evalu-
ated include "air permeability, "
resistance to abrasion and shrinkage.
Table 1. 2 lists some generally
accepted properties for various
materials commercially available
and in use today. Since felted
fabrics are made from wool and
orlon, it can be seen that the range
of applications for the reverse-jot
units is limited to temperatures
below 270°F and to dusts low in
alkali content.
b The actual weave patterns used for
a fabric will influence the filtration
process. This will be discussed
in greater detail in later sections.
Some concept as to why the weave
pattern should influence filtration
can be obtained by microscopic
examination. Figures 1.12, 1.13,
and 1. 14 illustrate the fact that
significant changes occur as both
materials of construction and weave
patterns are altered. * ' Details
of the fabrics shown are listed in
Table 1.3. These differences be-
come especially significant when a
scale on the order-of-magnitude of
particle diameters is used as a
reference.
12
-------
Table 1. 2 PROPERTIES OF FILTER FABRICS**
Fabric Filtration
FABRIC
Cotton
Wool
Nylon 6, 6 '"
HT-1 <"
Dacron '"
Orion ">
Creslan American Cyanamid Reg. Trademark
•Temperatures recommended by Industrial Gas Cleaning Institute
131 Union Carbide Reg. Trademark »> W, W. Crlswell Tradenime
**W.W. Criswell Company, Division of Wheelabrator Corporation, 800 Industrial
Higheay, Riverton, New Jersey.
13
-------
FIBERGLASS N2| FIBERGLASS N23 ~ti FIBERGLASS IS|2| FIBERGLASS N23
REFLECTED LIGHT
TRANSMITTED LIGHT
20 X MAGNIFICATION
^
u
7
~>
^
o
Figure 1.12
-------
DACRON A DACRON B
DACRON A
DACRON B
REFLECTED LIGHT
TRANSMITTED LIGHT
20 X MAGNIFICATION
Figure 1. 13
-------
FIBERGLASS NS| DACRON B
FIBERGLASS N*l DACRON B
REFLECTED LIGHT
TRANSMITTED LIGHT
20 X MAGNIFICATION
Figure 1.14
-------
Table 1.3 CONSTRUCTION DETAILS OF TEST FABRICS
Filter Fabric
Air Permeability (cfm/ft.2 at 1/2" HO)
2
Weight (oz./yd. )
Yarn Count
Filament Diameter (in.)
Yarn Diameter (in.)
Filaments Per Strand
Strands Per Yarn
Twists Per Inch
Weave
Finish
Fiberglass
No. 1 No. 2
13.84 11.67
8.41 8.67
55 x 50 55 x 54
No. 3
7.86
9.06
55 x 58
0.00025
0.0097
408
2
3.8
3/1 Crowfoot
Silicone Oil
Dacron
A B
28.06 14.62
5.51 6.06
82 x 62 82 x 76
0.00113
—
50
1
3.5
3/1 Twill
Silicone Oil
cr
2
o
S-
o'
-------
SECTION 23
Fabric Filtration-Basic Concepts
-------
FABRIC FILTRATION - BASIC CONCEPTS
I TERMS AND DEFINITIONS
As a beginning, consider filtration in an
apparalus similar to that shown in Figure 1.
Such a test apparatus is admittedly far re-
moved from filtration in a commercial filter
unit. I: will, however, provide a sound basis
for the development of those concepts and
equations necessary to understand full-scale
filtration.
As shown, the device would consist of an
inlet duct leading to an expansion chamber
and the actual filter media. Proper gaskets
and filter media support would be provided.
The cleaned gas would be exhausted through
: ho fan to almosphcrc-. Pressure differential
.•ross the filler media would be measured
by means of a U-tube or inclined manometer
.•ad expressed in inches I^O. The volumetric
flow role (Q), the area of the filter (Af) and
ihe mass of dusl filtered (W) are known or
arc measurable quantities.
The term "filter pressure differential" (AP)
has several synonyms which should be noted.
These are:
1 Pressure drop
2 Pressure differential
3 Head loss
A second fundamental term is the superficial
face velocity (Uf) for the carrier gas stream.
By definition:
Uf = ~ (1)
Synonyms for superficial face velocity are:
1 Filter velocity
2 Filter ratio
Direction of gas flow
and dust feed
I
AP
T
^
filter
\
Clean gas
(suction
out
system)
U-tube manometer
AP represents the filter pressure differential (inches H 0)
1. SCHEMATIC OF BASIC TEST APPARATUS
-------
Fabric Filtration Basic Concepts
3 Air-to-cloth ratio
4 Nominal velocity
In developing the concepts of filtration it is
fundamental to make the following assumptions:
1 Flow through the filtering media and
the dust cake is laminar in nature.
2 The aerodynamic properties of the
filtering media and dust cake are con-
stant for the filtering period.
3 The weight of dust filtered per unit of
time is constant.
Under these conditions, several plots of
pressure differential as a function of filtering
time might appear as shown in Figure 2.
At this point, it is necessary to evolve those
basic factors whi'ch will allow the curves of
Figure 2 to be evaluated in terms of assump-
tions 1. 2 and 3.
First, it should be noted that the pressure
drop is not a fundamental characteristic of
the system. The situation can be compared
to the pressure differential across an orifice.
For a given orifice the pressure drop can be
used as an indicator of the flow rate; for two
dissimilar orifices, however, the same pres-
sure drop would indicate different flow rates.
For a fixed laminar flow element, a direct
proportionality between pressure differential
and face velocity will exist. That is:
AP
AP
r
IT
(2)
In Figure 2 for example, the ratio AP/Ur. for
curves A, B, and C and at time = 0 is equal
to 0.5 in. H2O/fpm. This ratio will hold true
"as long as assumptions 1 and 2 are valid;
that is, either turbulent flow is not reached
or the filter media is not changed by elonga-
tion, compaction or other means.
10
O)
+->
^ 4
Curve
A
B
C
Fabric
type
Tim*
A P
t H20)
0.75
1 .5
3.0
U
(fpm)
1 .5
3.0
6.0
(grs ft')
5
2.0
1.0
End of cycle
AP
(in.
3.75
7.5
7.5
Uf
(fpm}
1 5
3.0
3.0
(g« ft3)
5
2.0
2.0
_L
20 40 60 80 100
Filter pressure differential, in H20
Figure 2. POSSIBLE PRESSURE TIME CURVES
-------
Fabric Filtration Basic Concepts
AP
The ratio — is called "Filter Drag" and
f
given the symbol S. In the equation form:
The rate at which dust is filtered can be cal-
culated from the equation:
= C
t p A
Q
C • U,. .
p f(avg.)
(4)
In other words, assumption 3 is fulfilled if
the product (Cn • UJ is constant.
Pf avg.
It is important to recognize the distinction
between assumption 3 and the often quoted
assumptions of "Constant Flow Rate" and
"Constant Dust Concentration. " Curves B
and C of Figure 2 reflect the difference. For
curves B and C assumption 3 was maintained
so that the product (C_ • Uf) remained
^ P 1 avg.
6. 0 (grains/ft ) per minute. For Curve C,
however, both Cp and U^ were varied. The
result is the non-linear relationship. The
straight line relationship (Curves A and B,
Figure 2) is the result of a constant feed
rate and constant flow rate.
It is at once obvious that changing the dust
feed rate will alter the slope of the line,
although the basic characteristics of the dust
cake need not be altered. This then, leads
directly to plotting Filter Drag (S) as a func-
tion of Filtered Dust Mass (W) in order to
delineate the basic performance character-
istics of the system under study. Replotting
the data of Figure 2 in these terms results
in a single line, as shown in Figure 3. This
representation is known as a "Basic Per-
formance Curve.
II INTRINSIC DUST CAKE PROPERTIES
If assumptions 1, 2 and 3 are valid the basic
performance curve will always be a straight
line; the slope of this line will depend on
3.0
E
CL
. 2.0
S-
d)
1.0
I
I
I
I
I
100 200 300 400 500 2 600
Filtered Dust Mass, grains/ft
Figure 3. BASIC PERFORMANCE CURVE FOR DATA OF FIGURE 2
-------
Fabric Filtration - Basic Concepts
properties of the deposited dust cake. An
indication of some of the basic parameters
influencing the properties of the cake can be
obtained by examination of several equations
developed for laminar flow through packed
beds, as listed below:
Chilton-Colburn
(2)
AP
Uf
JL 11
G F
c a
D
(5)
pe
Fair-Hatch
(3)
— = —
U, G KF&H
f c
9 A
Hatch
(4)
AP
U7
(5)
(1)
AP
f
G C,
c k
1 (1 -ex)
H- T~
D
2 A
_£_
V
P
(8)
Historically, each of these equations have
rosulled from an examination of D'Arcy's^
basic premise that resistance to flow is pro-
portional to flow rate and depth of media.
In the terms so far developed,
AS =
= k X
AL
(9)
Because of assumption 2 it can be said that
the change in depth of the dust cake (AL) is
proportional to the change in weight of dust
collected (AW). Therefore,
AS
or K
1 (AW)
K
AW
AS~
(10)
(11)
which is the inverse-slope of the Basic Per-
formance Curve. The proportionality constant,
K, is given the term "Dust Permeability, "
and has the dimensions (grains ft )l
(in. H2O/fpm).
From equations (5) through (8), it is apparent
that K will be greatly influenced by particle
size and size distribution, particle shape,
particle surface characteristics, manner of
cake formation, and gas viscosity (viscosity
is discussed in detail in a latter section.)
Excepting viscosity, it is apparent that these
factors cannot be measured in a dust cake.
Therefore, dust permeability will be a prime
factor for possible correlation with other
operating conditions. The terms shown in
equations (5) through (8) will be useful in
interpreting these results. Generally
speaking, high values of dust permeability
imply a dust "easy" to filter, while a low
value of dust permeability implies design
problems in terms of pressure differential,
filtering time, and filter velocity.
Examples of the manner in which permeability
varies from dust to dust are Shown in
Table l.<6>
/n\
According to the authors, each dust listed
in Table 1 met the conditions of laminar flow
and the cake properties were constant through-
out the experimental runs. Several factors
should be noted, however, which prevent
extrapolation of the general conclusions to
other situations:
A The dusts were deposited at a constant
filter ratio of 10.0 fpm. This, in itself,
is a high velocity that could force the
deposition of a cake in an already compact
condition. In actual practice, velocity of
deposition is a variable quality, as will
be explained later.
B The filter itself consisted of an unidentified
piece of cloth "stretched tightly to prevent
sagging, which was found to affect the
resistance. " The manner of drawing the
cloth is not noted. It can be expected that
these parameters, (i. e. , the cloth itself,
the fact that the cloth could not stretch or
change shape, and the manner of cleaning)
might influence dust permeability, and any
relationship between filter drag and weight
of dust collected.
-------
Fabric Filtration Basic Concepts
Table 1. PERMEABILITY OF SEVERAL TEST DUSTS AS A
FUNCTION OF PARTICLE SIZE DISTRIBUTION*6*
Material
Granite
Foundry
Gypsum
Feldspar
Stone
Lamp black
Zinc oxide
Wood
Resin (cold)
Oats
Corn
Particle Size
Coarse
<20 Mesh
4,430
11,300
7,300
4,430
11,300
<140 Mesh
3, 180
4,430
11,300
<375 Mesh
1,850
1, 110
1, 110
4, 430
Medium(b)
< 90 ^
1, 110
1, 110
730
1,850
< 45 u
636
795
Fine
< 20u
363
370
257
<2M
148
446
278
(a) Flocculated material not dispersed; size actually larger
(b) Theoretical size based on density of silica; no correction made for actual particle density
Table 1 also illustrates a difficulty that would
arise in attempting to correlate dust cake pro-
perties with particle properties. Thus, dust
permeability is listed as 4, 430 for Corn Dust
passing a 375 mesh sieve (aperture equal to 39
microns). Note, however, that dust perme-
ability of Corn Dust elul riated to sizes less
lhan 4Ti microns is only 7!)5. The experimental
techniques, therefore, produced two entirely
dil'ferenl dusts. This discrepancy does no I
negate Hie need for basic studies, but does
indicate a need for caution in extrapolating
such results to operating conditions.
Ill FACTORS INFLUENCING DUST CAKE
PROPERTIES
By the year 1955, the concepts of Williams,
el. al. ''J' had been further investigated by
/7\ ° J
Snyder and Pringu' and utilized by Hemeon.
/Q\
v '
.
They were also published by Lapple^) in the
3rd Edition of the Chemical Engineer's Hand-
book. As a result of their experiments,
(7)
however, Snyder and Pringv ' questioned the
validity of constant dust cake properties.
Two sets of data published by the authors
are shown in Figures 4 and 5.
Figure 4 shows the influence of the fabric
itself on rale of filter drag increase. From
these data, the authors concluded that for the
same particulate properties, dust cake
permeability increases as the fiber surface
area (square feet per square feet of fabric)
increases. To use an extreme example
illustrating this relation, consider the same
dust collected "on the surface" of a membrane
filter and also collected by means of a woven
cotton fabric. In the first instance, the cake
formed would not be influenced by irregulari-
ties in the surface or fibers projecting beyond
the surface; it would, in fact, be a pure dust
cake, with a clear dividing line between dust
and filter. This would not be the case when
using the cotton fabric. In fact, we would
logically expect a change in properties as the
-------
Fabric Filtration - Basic Concepts
2.0
E
CL
O
CM
:r
en
(O
i-
o
1.0
A. 1-12 Fiberglas fabric, low fiber surface
area per square foot of cloth.
B. Napped B-27 Orion, medium fiber surface
area.
C. B-26 staple Orion fabric napped both sides,
high fiber surface area.
200
300
400
500
600
Filtered Dust Mass, grains/ft
Figure 4. EFFECT OF FABRIC ON BASIC PERFORMANCE CURVE
2. or
o
IN)
cn
(O
$
A. High twist unnapped Orion, low
fiber surface area per square
foot of cloth.
B. Fiberstock Orion, high fiber
surface area.
TOO
Filtered Dust Mass, grains/ft
Figure 5. EFFECT OF FABRIC ON BASIC PERFORMANCE CURVE
-------
Fabric Filtration Basic gorurepts
dusl deposit first bridges the fabric pores,
then proceeds to fill surface irregularities,
and eventually, if sufficient quantities are
deposited, extends completely beyond the
range of fabric influences. With a relatively
coarse dust the depth of material per unit
increase in drag would undoubtedly be greater
than with fine materials. Therefore, the
likelihood of a linear relationship between
drag and dust mass is greater with coarse
materials when compared to fine dust.
Figure 6 should help visualize the situation.
Figure 7 further supports the argument. In
this case, different dusts are collected by the
same fabric. The data do not negate the
possibility of linear relationship for the finer
materials at higher dusl loadings on the filter.
It should be noted that the test apparatus,
although similar to that used by Williams,
differs in several important aspects:
A In the experiments of Williams, the nature
of the filtered dust was altered by sieving
and elutriation columns. Snyder and Pring,
C
on the other hand, passed the air-dust
mixture "through a settling chamber
(duplicating the action of the inlet chamber
of the Dustube collector)". It is not clear
if Snyder and Pring measured particle
properties (i. e. , grain loading and size
distribution) before or after the settling
chamber. This, in turn, makes it difficult
to generalize on the results obtained.
The fabric is apparently unsupported in
the apparatus used by Snyder and Pring.
Stretching of the cloth as drag forces
increased would undoubtedly lead to
changes in dust cake structure.
At the end of each run, the cloth was
cleaned by "shaking the fabric for a timed
period in a prescribed manner. " This
undoubtedly achieved better cleaning than
is experienced in practice, and would
necessitate the build-up of the entire dust
cake with each run. It has been our ex-
perience that this "over-cleaning" almost
always results in non-linear filter-drag vs.
dust mass curves, especially in the
beginning of a run.
Drag
Depth of fabric plus cake
Figure 6. DRAG AS A FUNCTION OF MASS OF COLLECTED DUST
-------
2.0
1000
2000
Filtered Dust Mass, grains/ft
3000 4000 5000 6000
7000
100
200
400
500
600
700
Filtered Dust Mass, grains/ft
8000
Wheelabrator Steel Scale
Dust
Moderately Coarse
2% less than
10 microns
Ground Limestone Dust
Coarsest, 32.33
minus 325 mesh
0-5
- 0.4
- 03
- 0.2
O
csi
01
fD
•o
Figure 7. EFFECT OF DUST TYPE ON FILTER RESISTANCE OF COTTON SATEEN CLOTH
-------
Fabric Filtration Basic Concepts
IV PERFORMANCE ON COMMERCIAL-SIZE
FILTER TUBES
When the filtration process is extended from
a single area, as previously described, to
I he area of a commercial size filter tube,
(he overall process becomes less uniform
from area lo area. An evaluation as to the
extent of these non-uniformities formed the
basis for- research by the Public Health
Service which began in 1957. The basic test
unit for1 this research is shown in Figure
With this unit it was possible to control I he
volumetric flow rale at a pro-set level,
maintain constant dust feed rates, vary shaking
conditions relative to amplitude, frequency,
direction and duration, and vary tube diameter
and length. The unit could be programmed
for eonl inuous cycling, so that equilibrium
could be approximated. Point measurements
of dusl mass and filter velocity could also
he made, using the probes shown in Figures
As a result of studies on this unit, it lias
been shown that dust cake properties vary
over a filter area. Typical "profiles" illus-
trating such variations are shown in Figures
11 and 12.
(12)
Further, it. can be concluded
that a non-uniform residual profile will exist
at the start of every filtration period after
the first, and that, as a result, filtration
no longer occurs uniformly over the entire
filter surface. Rather, different filter
velocities, different amounts of dust collected,
and different cake structure's will be found at
various locations.
In relation to dusl permeability, it should be
noted that "when considering larger areas such
as that of an entire filter bag, no longer is
matrix structure solely of import, but
now, the macroscopic structural features
or 'topography1 of the cake must be con-
sidered. In other words, the shape of the
mass and resistance profiles will have direct
and significant bearing on the resultant effec-
tive permeability of the dust collected on the
f-
(*)
M L L T (j NI f I ( IS
Nl VI ir, A I Hi nwf.HS
LOWER HOU !-' i NG
UU'j f LNTRfllNME-Nl CHAMHER
U'.T Ft" EDER
iGU-SPELO FAN
QHPER
8 FiLUR TUUES
9 SHAKING MECHANISM
10 DUST CATCH BINS
"IN PLACE"
I I OUT.T CATCH tJINS DUR-
ING FILT RAT lON
12 HOPPER DOOR
li DUST CATCH HOPPER
©
Figure 8. SCHEMATIC DRAWING OF TWO
BAG TEST UNIT
-------
Fabric Filtration - Basic Concepts
Figure 9.
Pb21°-Bi21° MASS PROBE
Figure 10. FILTER VELOCITY PROBE
filter. This seeming anomaly exists because
of the physical configuration of the cake over
the whole filtering area and it demonstrates
that data from laboratory bench-scale deter-
minations of permeability cannot be equated,
for design purposes, to the effective perme-
ability of the same dust on full-sized tubes."
Obviously, these same conclusions can be
extended by considering possible variations
from filter to filter in a single compartment.
V SUMMARY
A Dust Collection Mechanisms
1 Initial stages of filtration
During the initial stages of a filtration
cycle the collection of dust particles is
accomplished by means of interception,
intertial deposition, diffusional impac-
tion, and electrostatic forces, as the
dust laden gas is passed through a fabric
of woven or felted structure.
2 Dominant collecting mechanism
After some short interval of operation
sufficient dust will be retained to form
a dust cake. Once the dust cake has
been formed the dominant collecting
mechanism is that of sieving. Because
the sieving media is composed of the
material being collected pore sizes in
the sieve will be on the order of particle
diameters. Therefore, efficiency of
collection is high, and the fabric filter
becomes a positive collection device.
B Properties of Dust Cakes
1 Laminar flow elements
The collected dust cakes, being com-
posed of small particles and being
subject to low superficial face velocities
are considered as laminar flow elements.
At any instant in time, the ratio of
pressure drop across such an element
to the superficial face velocity through
the element is related to properties of
the system not easily measured at this
time.
C Residual Filter Drag
Residual filter drag is that drag value
which exists after a dust fabric combina-
tion has been cleaned. As such it is
dependent on many interacting factors and
relationships, such as mode of cleaning,
type fabric, the particular dust involved,
velocities of filtration, and, time in
service. Experience has indicated that
this residual resistance is a major portion
of the total resistance across a fabric
filter unit, despite references to the
contrary in the literature. At the present
time there would appear to be no method
whereby this quantity can be estimated,
although for most installations it will
probably lie within the range of 0. 5 to
1. 5 in. H2O/fpm.
D Dust Permeability
It can be seen from Figure 13 that after
the initial stages of filtration a constant
10
-------
Fabric Filtration Basic Concepts
TYTY'Y \
W0.1 W0.4 W0.5 W0.fe TO.IO W
-------
Fabric Filtration Basic Concepts
relationship exists between changes in
filter drag and increments of dust filtered.
From equations (5) through (8) it can be
surmized that the only factor changing the
drag is the cake thickness, L. Since it is
impossible to measure L it becomes easier
to consider cake thickness as proportional
to the weight of dust deposited. On this
basis, it can be said that under particular
conditions, the cake properties indicated
in equations (5) through (8) can be expressed
by one proportionality constant, the dust
cake permeability, which is calculated
from the equation:
K =
AW
AS
(11)
As discussed, it is important to know the
conditions under which K has been deter-
mined. Of special significance is the
filter area involved.
E Terminal Filter Drag
The terminal filter drag is that resistance
of the dust fabric combination which exists
immediately prior to filter cleaning.
12
-------
SECTION 24
Fabric Filtration Operations and
Industrial Applications
-------
FABRIC FILTRATION OPERATIONS
AND
INDUSTRIAL APPLICATIONS
I. INTRODUCTION
A. The Baghouse
B. Filtration Processes
C. Operational Parameters
II. THE FILTERING MEDIA
A. Fiber Types
B. Yarn Construction
C. Weaving
D. Fabric Treatments
E. Bag Construction
III. BAGHOUSE DESIGN
A. Cleaning Processes
B. Baghouse Construction
C. Ma i ntenance
D. Dust Disposal
IV. INDUSTRIAL APPLICATIONS
A. Cement Kilns
B. Foundry Cupolas
C. Steel Furnaces
D. Nonferrous Metal Furnaces
E. Carbon Black Plants
F. Grain Handling Operations
G. Other Applications
PA.C.pm. 103.4.73
-------
I. INTRODUCTION
A. The Baghouse
Filtration is the removal of solid particles
from a fluid by passing the fluid through
a filtering medium on which the particles
are deposited. The industrial fabric filter
or baghouse consists of a woven or felted
fabric through which dust laden gas is
forced. As particles accumulate, resistance
to gas flow increases and pressure drop
across the filter increases. Accumulated
deposits are removed periodically by clean-
ing of the fabric in order to maintain
practical head loss. Provision for cleaning
the fabric filters in place is a distinguish-
ing design characteristic of baghouses.
Baghouses are used to collect particles in
the size range of submicron fumes to powders
greater than 200 microns. Fabric filtering
materials have been developed to handle gas
temperatures up to 550°F and can withstand
most chemical reactions. Recovery of
collected material is also a characteristic
of baghouse operation.
Superficial filtration velocities (total
air volume filtered/total cloth area)
commonly called the air-to-cloth ratios are
a function of the quantity of ventilation
or process gas, the dust concentration, and
the flow resistance properties of the parti-
culate deposit. The air-to-cloth ratio is
generally in the range of 1 to 15 feet per
minute and pressure drop is about 3 to 10
inches of water.
Commercial collectors are available in sizes
from a few square feet of cloth up to several
thousand square feet. Smaller units may be
fabricated and assembled by production line
processes while larger units are usually
designed to meet the requirements of specific
applications and are assembled at the
installation site.
Costs for baghouses vary in proportion to
size and with respect to kind and arrange-
ment of fabric and cleaning apparatus. A
range of 0.35 to 1.25 dollars per cubic foot
of gas per minute is typical of initial
collector costs although actual installed
costs run up to three times the base cost.
Where very high efficiency of collection of
small particles is required, baghouse oper-
ation costs are lower than other types of
collection equipment. If fabric filters
are properly designed, installed, operated,
and properly maintained, collection efficien-
cies in excess of 99.9% may be obtained.
B. Filtration Processes
Fabrics used for removing dust and fumes from
gas streams are usually woven with relatively
large open spaces up to 100 microns in size.
Since collection efficiencies for dust parti-
cles of 1 micron or less may be.greater than
90%, the filtering process obviously cannot
be simple sieving. Dust particles are cap-
tured and retained on the fibers by means of
interception, impingement, diffusion, gravi-
tational settling, electrostatic attraction,
and sieving.
Direct interception is possible because the
flow through fabric filters is usually lami-
nar. A dust particle experiences direct
interception by the fabric filter when it
comes in contact with a fiber as the stream-
line passes by the fiber. The particle adheres
to the fiber because of van der Waals forces.
This occurs in baghouses for particles less
than 1 micron in size as inertia effects
become dominant for particles larger than
that.
Larger particles do have appreciable inertia
and do not follow a streamline when the
streamline is deflected from a straight path.
The probability of a particle contacting the
surface of the obstruction depends on the
size of the obstruction and the size and
inertia of the particle. Smaller fibers are
more effective as streamlines pass closer to
smaller obstructions than to larger obstruc-
tions. Particles with greater inertia are
more likely to strike and impinge on a col-
lecting surface than a particle with less
inertia.
For very small particles with diameters less
than 0.1 or 0.2 microns, diffusion is the
prominent mechanism of collection. Particles
as small as these no longer follow stream-
lines because collisions with gas molecules
occur, resulting is random Brownian motion
that increases the, chance of contact between
the particles and the collection surface.
Not a great deal of information is available
that describes the role electrostatics play
in the collection of particles by baghouses.
Electrostatics may not only assist filtration
by providing an attractive force between
dust and fabric, but also may affect agglom-
eration, fabric cleanability, and collection
-------
efficiency. Charging of particles is likely
due to frictional effects.
Once a mat of dust or filter cake is accumu-
lated, further collection is accomplished by
sieving as well as the other mechanisms. The
fabric material serves as a supporting struc-
ture for the filter cake. After bag cleaning,
some residual dust remains to aid in further
filtering. A short time after cleaning the
dust cake acts as a seive and becomes the
dominant collecting mechanism. It is during
this period that the extremely high efficienc-.
ies are reached.
dust and filter combination is influenced by
particle size, size distribution, particle
shape, surface characteristics, the manner of
dust cake formation, and the gas viscosity.
In general, a high value of permeability
implies a dust that is easy to filter. A low
value means that a high pressure drop may be
expected along with problems with filtering
time and filter velocity.
Table 1 shows some values of permeability
for different dusts and dust sizes. Note that
from the largest to the smallest particle
permeability may vary by a factor of 1000 or
more.
C. Operational Parameters
Clean filter air resistance to flow is depen-
dent on fiber structure and the weave of the
cloth. A tight weave offers more resistance
to flow than a looser weave at the same rate.
As previously determined, the air flow
through a fabric is laminar and because of
this flow resistance or pressure drop will
vary directly with flow. Air flow through a
fabric filter is usually described as the
superficial face velocity or air—to-cloth
ratio. By definition:
(1)
where Q is the volume flow rate of dirty air
through the filter and A is the surface area
of the filtering media. Since the pressure
drop, AP, is proportional to the superficial
face velocity, U , a constant known as the
filter drag, S, may be defined:
AP
S = —
n
(2)
The filtering process of a fabric filter is
a function of the dust cake. Factors that
influence the properties of the dust cake are
worth discussing. A term commonly used with
regard to fabric filters is permeability.
Permeability is the openness of a material to
the transmission of a fluid. It is not deter-
mined theoretically for fabric filters, but
is measured using the relationship:
AW
K =
AS
(3)
where K is the permeability, AW is the change
in weight of dust collected during a change
in filter drag, AS. The permeability of a
A system of bags in a baghouse is similar to
a parallel electrical system. An analogy may
be drawn between total electrical resistance
and total flow resistance, S , for a baghouse.
Pressure drop may be determined across each
bag in a baghouse providing an equivalent or
total pressure drop of
111 1
JL = + + +... (4)
where n is the number of bags.
Figure 1 shows the basic performance curve of
the baghouse operation. Residual drag is that
drag value which exists after a dust - fabric
combination has been cleaned. It is dependent
on mode of cleaning, type of fabric, type of
dust, filtration velocities, and time in ser-
vice. Experience shows that residual drag is
a. major portion of the total resistance across
a. fabric filter. It is estimated to be about
0.5 to 1.5 inches of water per foot per min-
ute. Terminal filter drag is the resistance
of the dust - fabric combination immediately
prior to cleaning. The terminal resistance of
the fabric filter during the filtering opera-
tion may be estimated using:
(5)
where ST is the terminal resistance of the
filtering element. S* is determined by ex-
K
tending the linear portion of the Basic
Performance Curve to the zero intercept. C
represents the particulate concentration in
the dirty gas stream and t is the elaped time
of operation. The term U • C • t is the
t P
change in weight of dust collected per unit
-------
Material
Granite
Foundry
Gypsum
Feldspar
Stone
Lamp black
Zinc oxide
Wood
Resin (cold)
OatB
Corn
Particle SUc
Coarse
-------
of fabric area. With this relationship the
time between required cleanings may be de-
termined if the terminal bag resistance is
set at some maximum level. The face velocity
may be calculated and the dust concentration
is known or may be sampled. The dust permea-
bility is determined from pilot tests or pre-
vious experience.
For multi-compartment baghouses one other
team must be defined. This is the equivalent
terminal drag of the baghouse system, S^. As
the number of bags or the number of compart-
ments increases the equivalent terminal drag
effectively decreases for the same filtering
conditions. The equivalent terminal drag may
be estimated by:
et
(6)
where m is a constant that is dependent on
the number of bags or compartments. Table 2
shows the values of m as a function of the
number of compartments, n.
Table 2. VALUES FOR m AS
FUNCTION OF n
1
2
3
4
5
6
7
8
9
10
1.0
0.78
0.72
0.68
0.65
0.63
0.61
0.60
0.59
0.58
11
12
13
14
15
16
17
18
19
20
0.57
0.56
—
0.55
--
—
0.54
--
—
0.53
II. THE FILTERING MEDIA
A. Fiber Types
The type of fabric selected for use in a bag-
house is an important consideration in a pro-
perly designed installation. The material
must be compatable with the temperature of the
gas and the chemical constituents of the
effluent. There are several types of fibers
used for filters in baghouses; each one has
characteristics which give it advantages over
the others.
For many years cotton has been the standard
fiber for most common dusts. It is inexpen-
sive, readily available, effective, and rea-
sonably durable. One major limitation is
temperature, as cotton cannot be used for gases
with temperatures greater than about 180°F.
Cotton is not recommended for effluent gases
of high acid or alkali content. It is used
extensively in applications for abrasive blast-
ing, rock crushing, and conveying.
Wool felt is used for many metallurgical
operations such as lead smelters and for
reverse-jet baghouses. The temperature limit
for wool is about 220°F and wool can resist
the action of acid effluents reasonably well.
Wool has been mixed with asbestos for some
applications.
Nylon is relatively high in initial cost, but.
has excellent resistance to abrasion and flex-
ing, and a resistance to many chemicals. Nylon
has a slick surface that allows for easy
cleaning of the fabric surface. However, other
synthetic fibers with similar properties are
used more frequently than nylon because the
thermal properties of nylon allow only 220°F
for continuous operation.
Dynel is an acrylic fiber that has low moist-
ure absorption, good strength, resilience,
and resistance to many chemicals, mildew
and bacteria. Dynel will not support combus-
tion and is used in applications where chem-
ical resistance is important. The maximum
temperature gas recommended is about 175°F.
Orion and Dacron have similar properties in
that both resist chemical reactions and have
good heat resistance. Orion is produced only
in staple form while Dacron may be purchased
in filament-type yarn. A filament-type fabric
is more easily cleaned. Because of this and
because Dacron is less expensive, many early
Orion users have switched to Dacron. The
temperature limit for both is about 275°F.
Teflon has been used as a baghouse material
for high temperature gases. The tetrafluoro-
ethylene fiber can withstand continuous oper-
ation temperatures of about A50°F to 500°F
and is inert to most chemicals except fluorine
and chlorine. The flex and abrasion strength
of Teflon bags is only fair and teflon is
expensive.
Of all the materials available for filtration,
glass fabrics have the highest resistance to
high temperatures and most chemicals. Glass
-------
Table 3. PROPERTIES OF FILTER FABRICS**
FABRIC
Cotton
Wool
Nylon 6, 6 '"
Nomex
Docron '"
Orion '"
Creslan "'
Dyn.l '"
Polypropylene
Teflon '"
Hberglas
Flltron "'
Helling
Temperature
Decomposes at
302° F
Chars at 572° F
480° F
Chars at 700° F
482° F
482° F
Softens
475° F
Softens
325° F
Softens
333° F
Decomposes at
750° F
1470° F
505° F
Softens
fMlllWM
T.SSSS-
180° F
200° F
200° F
400° F
275° F
260° F
250° F
160° F
200° F
500° F; emits
toxic gas at
450° F
550° F
270° F
Acid
Resistance
Poor
Very Good
Fair
More resistant than
Nylon; inferior to Dae-
ron I Orion.
Good to most mineral
acids. Dissolves partial-
ly in concentrated
H.SO..
Good lo excellent in
mineral acids.
Good in mineral acids.
Little effect even In
high concentration.
Excellent
Inert except to fluorine.
Fair to Good
Good to excellent.
Mkili
Resistance
Very Good
Poor
Excellent
Not as resistant as
Nylon; superior to Dae
ron i Orion.
Good in weak alkali.
Fair in strong alkali
Fair to good in weak
alkalis.
Good in weak alkalis.
Little effect even in
high concentration.
Excellent
Inert except to chlorine,
tri-fluoride and molten
Alkaline metals.
Fair to Good
Good
Ffei
Abrasion
Very Good
Fair to Good
Excellent
Good
Very Good
Good
Good to Very Good
Fair to Good
Excellent
Fair
Fair
Good to Very Good
<'> Du Pont Rag Tradamark (!) American CyaMmid Reg Tradamark
•Taniptralurti recommended by Industrial Gaa Cleaning Institute
111 Union Carbide Beg Trademark
"> W. W. Criiwell Tredename
**W.W. Criawell Company, Division of Wheelabrator Corporation, 800 Industrial
Higheay, Riverton, New Jersey.
fibers, however, have a low resistance to
abrasion and crushing, so special precautions
must be taken during the cleaning cycle.
Vigorous shaking is avoided and filtering
velocity is usually less than for other
fabrics on the same dust. Fiberglass bags
have been applied to gases with temperatures
up to about 550°F.
Table 3 shows most of the types of fibers
used for filter bags in industrial applica-
tions and some of the characteristics of each.
B. Yarn Construction
The construction of the yarn for filter cloth
is as important as the material used to make
the yarn. The weave, count, and finish are
characteristics that may be controlled in
making the filter cloth. The two main types
of yarns used in weaving are filament yarns
and staple yarns. Only synthetic fibers may
be made into filament yarns as synthetic
fibers are manufactured by extruding material
through a spinneret to form long individual
filaments. These filaments may be twisted
together to form high tensile strength multi-
filament yarns. Filament yarns have a slicker
surface than do staple yarns. Staple yarns
of synthetic fibers are produced in a similar
manner except that filaments of staple yarns
are much shorter and finer. The surface of
the synthetic staple filaments is often
textured by using compressed air to rough up
the surface as the filament is extruded from
the spinneret.
Cotton staple fibers are cleaned and drawn
into parallel order by carding and are event-
ually twisted into yarns by a spinning process.
Some of the properties of the spun yarn depend
on the spinning and the amount of twist. A
highly twisted yarn tends to resist penetration
of particles into the interstices of the yarn.
-------
Yarn number is a measure of linear density.
Direct yarn number is the mass per unit length
of yarn, while the indirect yarn number or
yarn count is the length per unit mass of yarn.
Classification of yarns is different for cot-
tons and synthetics. Cotton yarns have been
numbered by determining the weight in pounds
of 840-yard lengths or hanks, but more fre-
quently by determining the number of 840-yard
hanks per pound. For example, if 840 yards
weigh one pound, the yarn count is Is; if 30
such hanks weigh one pound, the count is 30s.
A heavy yarn would be Is, a medium yarn would
be 30s, while a light yarn may be 160s.
Woolen system yarn is measured by the number
of 300-yard hanks per pound, while worsted
system yarn is measured by the number of
560-yard hanks per pound. Man-made fibers
are usually measured using the denier system.
The denier is equal to the weight in grams of
9,000 meters of yarn. The cotton yarn count
may be obtained by dividing 5,315 by the
denier number.
C. Weaving
The production of a flexible material for
fabric filtration involves weaving. Most
felts used in filtration are first woven and
then given further treatment. Woven fabrics
are formed by interlacing yarns at right
angles on a loom, after which the fabric may
be further treated. The most common patterns
of interlacing for fabrics used for gaseous
filtration are known as twill and sateen or
satin. Plain weave fabric is sometimes used.
The twill weave may be recognized by the diag-
onal pattern formed by the filling yarn inter-
lacing more than one warp yarn. The values
of twill weave include its strength and drap-
ability. The diagonally arranged interlacings
provide greater pliability and resilience than
the plain weave. Twill weaves are frequently
tightly woven and will not get dirty or
blinded as quickly as plain weave, though
twills are more difficult to clean when they
do get soiled. Twill weaves are used where
strong construction is essential.
Satin weave is similar to twill weave in
construction except that a satin weave requires
five to twelve harnesses in construction
while a twill weave requires no more than
four. Satin weave differs in appearance from
the twill because the diagonal of the satin
weave is not visible. Since more harnesses
are required for satin weave, a greater
amount of fine yarn may be compressed into
a given space of cloth. This compactness
gives the fabric more body as well as less
porosity.
Plain weave is the simplest type of con-
struction and is consequently most inexpen-
sive to produce. On the loom the plain
weave requires only two harnesses as each
filling yarn alternates over and under the
warp yarns. If the yarns are close together,
the plain weave has a high thread count and
the fabric is firm and will wear well.
Felts used in fabric filtration are also
woven in their early stages, but subsequent
steps change the character of the material
from a woven fabric. Using a woven base
fabric called a scrim increases the strength
and stability of the fabric over a matted
felt fabric. Woolen felt is produced by
mechanically working a woolen scrim in warm
water in the presence of certain lubricants'
and chemicals in order to shrink the material.
Napping the surface of a material produces
a felt-like fabric. Needle punching is a
method of combining two or more layers of
fabric into a felt-like fabric. Usually one
layer is a scrim for strength, while the
others may consist of fibers of almost any
description or combination. In this way con-
siderable control over material properties is
possible. Non-woven production methods
include resin bonding, wet bonding (paper-
like materials), spun bonding, heat bonding,
chemical bonding, spray bonding, and stitch
bonding. Nearly all fibers used in fabric
filtration can be used in non-woven fabrics
and non-woven fabrics can be produced more
rapidly than by weaving.
D. Fabric Treatments
Dimensional stability is an important factor
in filter fabrics. Cotton and wool fabrics
must be preshrunk and synthetics are given a
corresponding treatment called heat-setting.
This process contributes to a more even bal-
ence of warp and filling yarn tension, pro-
vides better surface smoothness, reduces yarn
slippage, controls porosity, and eliminates
shrinkage. Man-made fibers frequently con-
tain one or more of the following additives:
- plasticizers to reduce flow viscosity
and improve low temperature flexibility,
- solvents used in wet spinning and in
coating,
- organic peroxides used as poly-
merization initiators,
- antioxidants added to reduce
oxidative deterioration during
manufacture, processing, and storage.
-------
2.0
A. 1-12 Fiberglas fabric, low fiber surface
area per square foot of cloth.
B. Napped B-27 Orion, medium fiber surface
area.
C. B-26 staple Orion fabric napped both sides,
high fiber surface area.
Filtered Dust Mass, grains/ft'
Figure 1. EFFECT OF FABRIC ON BASIC PERFORMANCE CURVE
2. or
High twist unnapped Orion, low
fiber surface area per square
foot of cloth.
Flberstock Orion, high fiber
surface area.
Filtered Dust Mass, grains/ft'
Figure 3. EFFECT OF FABRIC ON BASIC PERFORMANCE CURVE
-------
Other agents are added to fabrics such as
flow control agents, colorants, flame retard-
ants, stabilizers, ultraviolet absorbers, and
antistatic agents. Yarns are treated as
well as fibers. Yarn treatments include
surface addition of lubricants, antistatic
agents, and other mechanical operations.
Fabric finishing includes processes to improve
appearance or serviceability of the fabric.
Cotton and wool fabrics are usually cleaned and
bleached and are sometimes waterproofed and
treated to protect against mildew and fire.
Synthetic fabrics are sometimes treated with
water repellants and antistatic agents. Glass
fabrics are usually treated with silicones,
graphite, and other proprietary finishes to
reduce fiber-fiber abrasion during filter
cleaning.
E. Bag Construction
The most common shape of filter elements used
is a simple, circular cross—section tube.
Commercial bags are usually 5 or 6 inches in
diameter and from 5 to 30 feet long. Bag
widths vary from the usual 5 or 6 inches to
20 inches, determined mainly by the width of
the material used to make the bags. There is
no standard length to diameter ratio, as,
from a theoretical point of view, the length
to diameter ratio has no effect on efficiency
of collection of the bag. The factor limit-
ing the length of filter bags is the cleaning
requirement. Excessively long bags are more
difficult to clean and create problems by
rubbing against each other. A practical limit
for the length to diameter ratio is about 20
to 1.
The multiple tube bag is a tube-type bag of
oval cross section with vertical stitching
that effectively divides the bag into cir-
cular tubes. The multiple tube bag has the
advantage of greater filtering area for a
given floor space and helps break its own
filter cake when the blower is turned off and
the bag returns to its oval shape. The bag
requires a special mounting and is more expen-
sive for the same filtering area than a tub-
ular bag.
Envelope type bags are nearly as common as
tubular-type bags. Filtering elements are
flat panels of cloth stretched over a frame.
The panels are usually in.pairs. Wear is
increased because of friction between the
filter cloth and the wire frame support. One
advantage associated with envelope bags is
that a greater filtering area may be installed
in a given size volume than for other designs.
Filter bags are available commercially over
a broad range of dimensions. The fabric sur-
face per compartment required may be deter-
mined from information about the allowable
variation in gas flow with respect to process
ventilation, the availability of sizes of
commercial compartments or houses, and the
expected frequency of maintenance. Many
combinations of filter length, diameter, and
spacing are available in order to obtain the
least expensive baghouse. Maximum filter
packing and compact filter housing do not
necessarily give lowest cost baghouse.
Closely packed filter elements tend to wear
against one another, and make inspection and
maintenance difficult. Taller units give
lower cost per floor space than compact ones.
III. BAGHOUSE DESIGN
A. Cleaning Processes
As previously mentioned accumulated dust tends
to increase the pressure loss through a fabric
filter until a desirable maximum value is
reached. The filter must then be cleaned.
Most of the development toward fabric filter
equipment has gone into improved methods of
removing the accumulated filter cake from the
fabric. As a result, a great variety of
cleaning mechanisms are available. Four
objectives of cleaning, bags are: 1) to remove
the desired amount of deposit from the fabric
quickly; 2) leave enough residual deposit to
improve collection efficiency at start-up for
woven fabrics; 3) avoid damaging cloth or
using too much power, either of which can be
a substantial part of operating costs; and
4) avoid excessive dispersal of removed dust
so that this dust would not have to be refil-
tered. There are two general types of clean-
ing; the first involves flexitig the fabric
and the second involves a reverse-flow of
clean air.
Mechanical shakers are the most common type
of baghouse cleaning equipment. Electric
motors are used to oscillate the tops of the
bags either vertically or horizontally. This
is usually done under slight negative pressure
inside the bag so that more effective cleaning
may be accomplished.
Sonic cleaning utilizes sympathetic vibrations of
low frequency sound waves to vibrate the bag
frame work. This method Is not successful
with difficult to remove filter cakes since
the total amount of energy imparted to the
bags is low.
-------
incoming gases
Filtering
Filtering
Fi Itering
Incoming gases
to fan to fan
All compartments filtering, dampers open One compartment shaking, balance filtering
incoming gases
incoming gases
}
/N
Filtering
\/
_ A
*
)•
Shaking
*
<
Fill
s
- <«-*
1
/ s
er
1>
+
'J
>
ing
^
/N
Filtering
^
t
Filtering
Shaking
y
to fan to fan
One compartment shaking, balance filtering One compartment shaking, balance filtering
Figure 4. TYPICAL PARALLEL FLOW SYSTEM FOR A
CONVENTIONAL MULTICOMPARTMENT BAGHOUSE
Bag collapse Is frequently used to improve
cleaning efficiency by permitting a small
flow of air to flow in the reverse direction
causing the bags to collapse. The process
may be repeated several times for each clean-
ing. The bag collapse method is often used
in conjunction with the other cleaning methods
to improve efficiency.
A variation of the bag collapse method is the
pulse-jet cleaning. For this method a "bub-
ble" of compressed air is injected at the top
of the bag when the bag is collapsed. As the
Pulse of air moves down the bag the filter
cake is flexed and the collected material
falls through the bag.
The reverse-jet mechanism is an example of
reverse-air cleaning. A high velocity jet
of compressed air is blown back through the
fabric to dislodge the collected dust. This
method is sometimes too efficient and does
not leave sufficient residual for cleaning
at start-up. In a typical reverse-jet filter
unit, cleaning may be conducted continuously
in order to maintain a constant pressure
differential across the unit.
Reverse air flow at low or atmospheric pres-
sures is also used to clean baghouse filters.
This method is often used for envelope-type
baghouses. A large volume of air is required
for this method which often requires a sepa-
rate air blower for cleaning. Supporting
rings are occasionally used inside filter
bags so that they may maintain their shape
during cleaning.
B. Baghouse Construction
The location of the blower on a baghouae deter-
mines the type of unit. If the blower is
located on the clean air side, the baghouse
is referred to as a pullthrough baghouse. In
-------
(
\
^
./
\^
VJ
(a) Bottom Feed (b) Top Feed (c) Exterior Filtration
Figure 5. POSSIBLE FILTERING SYSTEMS
tliis position the blower is protected from
the dust or fume being handled. However, a
relatively airtight housing is required of
the housing. A pushthrough baghouse has the
blower located on the dirty air side, and the
sides of the housing may be left open. This
configuration is often used for hot gases as
.1 greater degree of cooling may be obtained.
The dust loading that is handled by a push-
thi<>ii|;li b lower often causes substantial wear
.md frequent maintenance problems for the
1) lower.
The structure of the housing for a baghouse
must be able to withstand a pressure differ-
entia] of 8 inches of water or more. This
c.lls for heavy gage metal and bracing of
wails for the housing and the hoppers. Pull-
through baghouses are generally more of a
structural problem than pushthrough baghouses
as baghouse structures can withstand internal
pressure more easily than external pressure.
Bottom-feed baghouses are the most common
configuration used in industry. Dirty air
enters the bottom of the bags and filtered
through the bag from the inside out. The
clean air is on the outside of the bag. A
tup-feed bag is similar to the bottom-feed
except the dirty air enters the top of the
bag. Because of gravitational effects a top-
feed baghouse seems the more logical choice;
however, mechanical problems of securing bags
at both the top and bottom for top-feed bag-
houses makes the bottom-feed type easier to
construct and, thus, the more common. One
other possible configuration is the exterior
filtration type. Dirty air is brought through
the bag from the outside of the bag to the
inside. This type of arrangement requires an
inner-bag support structure to keep the inside
of the bag open.
Hopper size is dictated by the frequency of
pick-up and disposal of collected material.
If the hopper does not have adequate capacity,
dust already collected becomes reentrained
and increases the total dust load on the filter
cloth. This increases filter resistance and
the performance of the baghouse is affected.
Mechanical rappers or vibrators are sometimes
provided on hoppers to assist the collected
dust to flow freely from the discharge gate.
Materials that tend to stick or cake in the
hopper may be moved by rapping.
C. Maintenance
10
-------
The hopper of a baghouse should be emptied at
least once a day. Inspections of the equip-
ments should be conducted regularly at inter-
vals of a week, month, or quarterly, depending
upon the nature of the dust collected, the
quantity of dust, and the general severity of
service. Moving parts must be greased and
serviced. All bags should be examined once
a week to determine if any show wear. Ripped
bags should be replaced immediately.
l!ag failures generally occur between one and
two years after installation. After one year
frequent replacements are required meaning
that the baghouse equipment is down a great
deal of the possible operating time. Many
baghouse operators replace all the bags in
a baghouse periodically before serious pro-
blems develop. In this way bag replacement
may be scheduled to coincide with plant shut-
down periods.
Replacement of one or several bags in a large
baghouse is sometimes unavoidable. The resis-
tance of a new bag in a baghouse during start-
up is very low and, as a result, the filtering
velocity through the new bag is many times in
excess of the normal rate. This could result
in blinding of the new bag. Blinding is a
plugging of the fabric pores to such an extent
that the resistance becomes excessively high
permanently. One solution to the problem of
high filtering velocities is to precoat the
bags with dust to establish a dust cake
immediately after installation. Some author-
ities require all bags to be precoated after
each cleaning cycle.
Li. Uust Disposal
The most common means of disposing of dust
i.'i I Lee ted by ,1 bnghouse is to transfer it
I rum the hopper into a truck and take it to
,i dump. Much ol the dust dropped from a hop-
per to .1 truck wil.l escape to the air if not
h.indlL'il propri iy, defeating I lie purpose of
Lin- bnghouse. A sJeeve of canvas is frequently
i njt.-ii led on the outlet of the hopper to
eliminate this problem. After the dust is in
Llie truck it is usually wetted to keep it in
UK truck during the trip to the dump. This
l>nuedure is suitable for installations where
i lie dust is collected once a day. When more
frequent collections are made, automatic or
seminutomatic methods are recommended. Con-
veyors may be used to collect the dust from
'-'ever.ii hoppers and discharge the dust into
i covered tote box.
IV. iNltUSTKlAL APPLICATIONS
The collection of dust from rotary cement
kilns has been a difficult problem. Large
volumes of gas are handled with high concen-
trations of very fine particles, high gas
temperatures, and for wet-process kilns, the
presence of large amounts of water vapor.
Often conventional cyclones are used to col-
lect the large particles in front of the
final cleanup collector. Besides baghouses,
electrostatic precipitators are the only
devices available for cleaning the gases from
cement kilns. Fabric filter applications
have obtained efficiencies of 99.5 percent,
outlet loading below 0.02 grains per standard
cubic foot, and plume opacities less than
10 percent.
B. Foundry Cupolas
Exhaust temperatures from a grey iron cupola
range from 1000°F to 7500°F. Effluent load-
ings are about 1.0 grain per cubic foot with
much of the emission in the form of a fine
metal oxide fume less than 0.5 micron in diam-
eter. Gas cooling and high temperature fabric
filters are required. Using evaporative cool-
ing off-gas temperatures are reduced to about
450°F before filtration through fiberglass
bags. Typical filtering velocities are about
2.5 feet per minute and bags are about 11-j
inches in diameter and about 15^ feet long.
C. Steel Furnaces
The exhaust from the electric arc steel fur-
nace is characterized by a high percentage of
oxides of iron, highly variable gas tempera-
ture, variable dust loadings, and highly
variable gas volumes during different process
cycles. The use of dilution air to provide
for gas cooling and In-plant dust control
causes changes in the volume of gas to be
cleaned. Stack temperatures may reach 750°F
and higher with closed hooded units. Fiber-
glass bags are used with air-to-cloth ratios
of about 1.5. Orion bags are also used at
gas temperatures of around 200°F. Orion bags
have about a five-year life.
Iron oxide fumes from oxygen lanced open
hearth furnaces may be collected efficiently
by fiberglass bags. Temperatures range to
about 500°F and the filtering velocity is
about 2 ft/rain. Reverse air flex cleaning
along with sonic horns are used. Inlet load-
ings may be as high as 20 grains per cubic
foot and outlet loadings may be as low as
0.007 grains per cubic foot.
D. Nonferrous Metal Furnaces
Fiberglass haghouses have been applied to
11
-------
secondary lead smelters for fume collection
at temperatures higher than 400°F. The high
temperatures eliminate the deposition of
organic tars on the bags. Satisfactory
results have been obtained with filtering
velocities of about 1.2 feet per minute and
cleaning by shaking. Fiberglass bags have
also been applied to primary copper and zinc
smelters with some success.
E. Carbon Black Plants
Nearly all of the carbon black plants in the
United States are equipped with fiberglass
baghouses. Gas temperatures range from about
400°F to 500°F. Gentle cleaning techniques
are used such as bag collapse and sonic horns.
Filtering velocity is about 1.5 feet per min-
ute and the average life is about a year. The
high temperature of the gas is useful for
keeping the acid mist in the effluent above
the dewpoint.
F. Grain Handling Operations
Dust emissions come from several different
sources in grain handling operations: clean-
ing, rolling, grinding, blending, and the
loading of trucks, rail cars, and ships. Con-
veying and storing grains also cause dust
emissions. Cyclones are used to collect grain
dusts larger than 10 microns. Baghouses with
mechanically shaken woven cotton bags remove
99.9 percent of grain particles in the size
range of 1 to 5 microns with air-to-cloth
ratios of about 5 to 1. Felted bags with
reverse-jet cleaning may handle air-to-cloth
ratios up to 15 to 1.
G. Other Applications
One full scale bag filterhouse has been
installed to filter the entire exhaust efflu-
ent of a utility boiler. Fiberglass filter
bags are used with alkaline additives to
remove essentially all of the submicron parti-
culate. matter and a large percentage of the
sulfur trioxide. Hie boiler in this case is
oil-fired.
Baghouses have been applied in special cases
such as tobacco steaming processes, asbestos
recovery in brake lining manufacture, wood-
working facilities, grinding wheel manufacture,
flour mills, and other processes that may
include pulverizing, grinding, conveying, or
drying. Small particles and fumes from
industrial practices may be collected in fabric
filters.
REFERENCES
1. Billings, C.E., et al, Handbook of Fabric
Filter Technology. Volume 1, Fabric
Filter Systems Study, National Technical
Information Service, Springfield, Va.,
December, 1970, PB-200-648.
2. U.S. Department of Health, Education, and
Welfare, Control Techniques for Particu-
late Air Pollutants, National Air Pollu-
tion Control Administration Publication
No. AP-51, 1970.
3. U.S. Department of Health, Education, and
Welfare. Air Pollution Engineering Manual,
by J. A. Danielson, Public Health Service,
Washington, D.C.: Government Printing
Office, Pub. No. 999-AP-40.
4. Sommerland, R.E., ''Baghouse Filters as
Applied to Power Plant Effluents," Heat
and Fluid Dynamics Department, John
Blizard Research Laboratory, Foster
Wheeler Corporation, Carteret, N.J.,
May, 1966.
12
-------
Table 4. APPROXIMATE SIZE RANGES FOR FABRIC COLLECTORS
Collector
Reverse- Jet
Pressure-Jet^
Conventional
tubular bags
Mechanic*!
Reverse
flov
Envelope
(fpm)
1.0
1.0
1.0
1.0
1.0
Collector volume
per 1,000 cfm
(ft3)
1,250
670
210 - 370
590
210 - 340
Collector
Floor-Area
per 1,000 cfn
(ft2)
57 - 294
111
26-50
30-42
21-59
(fpn)
10
10
3
2
2
Collector volume
per 1,000 cfm
125
67
70 - 123
295
105 - 170
Collector
Floor-Area
per 1,000 cfo
(ft2)
5.7 - 29.4
11.1
8.7 - 16.9
15 - 21
10.5 - 29.5
(1) Does not Include dust hopper.
(2) Comon values for filter velocity.
C3) As manufactured by Fulverizing-Machinery Company, N. J.
-------
25
SECTION 25
Fabric Filtration-Mathematics
of Bag-Filter Operation
-------
IV. FABRIC FILTRATION - MATHEMATICS OF BAG-FILTER
OPERATION
As one moves from concepts of filtration related to a single
elemental area, to filtration on commercial size filter tubes,
to a parallel arrangement of tubes in a compartment, to the
conventional multi-section baghouse, filtration becomes
progressively more dynamic and non-uniform. Velocities of
filtration are seen to be extremely variable and certainly not
constant at any particular point for a filtration period. Meaningful
design data, therefore, can only be gathered from actual operating
installations or well designed pilot plants. This data should include
basic filtration parameters, so that proper adjustments can be made
to other operating conditions and other baghouse designs. It is the
purpose of this section to explore the basic mathematics of multi-
compartmental operation in order to delineate these basic parameters
and to evolve those equations necessary for baghouse design. The , , ,.
materials contained in this section are based on the works of Stephan,
Sargent,' ' and Walsh,''*- ^) as modified and extended by the author.
A. Basic Equations and Assumptions
As explained in previous sections, it is generally agreed that dust
fabric combinations behave as laminar flow elements at any
instant in time; that is:
= constant = S (4. 1 )
With certain limits, and depending on the nature of fabric and
dust, it would also appear valid to consider that changes in
filter drag (AS) are directly proportional to changes in the
weight of dust collected per unit fabric area (AW); that is:
AS = (constant) X ( W ) = ^ AW (4.2)
For an elemental filter area, then, the following equation can
be used to describe a Basic Performance Curve:
ST = SR + (4'3a)
U, C • t
or S = S + — - 2 - (4.3b)
PA.C.pm. 92.5.66
-------
Where S* is an extension of the linear portion of the Basic
Performance Curve to the zero intercept.
For convenience, Figure 4. 1 has been reproduced to summarize
these terms and definitions. It is important to note that in
calculating S, synonymous measurement of A P and U, must be
made unless it is known that Uf is constant throughout the
filtration period.
Throughout this text, filtration period is the time during which
an elemental area is on-stream; that is, the time required to
go from Sj^ to Sn-,. Unless otherwise noted, the term S-^ will
be considered identical to SR, for convenience and ease of
analysis.
B. Parallel Flow System, Discontinuous Operation
In any practical filter unit (i. e. , a single filter tube, a
collection of tubes in a single compartment or section, or a
multi-compartmented unit) an analogy can be drawn between
an electrical circuit and the filtration system. This analogy
has been illustrated and described by Stephan for filtration
through a single tube. The illustration used is shown as
Figure 4. 2.
For the electrical circuit, Ohm's Law applied, as follows:
1=^- (4.4)
e
Where:
I-I^Iu.IJ. x1 /„ e»
-75— - p- + T5 - + T5— +....+ p- (4.5)
Ke Rl R2 R3 Rn
By substituting like terms, it follows that:
(4.6)
S
e
Wherein Sg has the dimensions (in. H^O/cfm). In filtration,
however, it is advantageous to relate resistance (filter drag)
to one square foot of filter area, so that equation (4. 6) is
usually written as:
4.2
-------
E
a
O
CVl
X
cr
a
a:
LU
b
TOTAL CYCLE
REPEATED TO ATTAIN EQUILIBRIUM
A
INTERVAL OF
CAKE REPAIR
*
Projected
Residual Dra
B
DEPOSITION OF HOMOGENEOUS
DUST MASS
AW
DUST PERMEABILITY
AW
TERMINAL
DRAG
Figure 4. 1
FILTERED DUST MASS,W(GRAINS/FT.2 )
- SCHEMATIC REPRESENTATION OF BASIC
PERFORMANCE PARAMETERS,
-------
S = ** (4.7)
6
and the units of S£ become (in. P
As in the electrical circuit, the effective filter drag (Se)
is related to local filter drag (Sj, S2, S3> ..... Sn) as
follows:
1 1 (aj) I(a2) 1 (a3) 1 (aj
s; = ZJKJ + 3-P9 + -5^39- + ...-+ 3^9 <4-8a>
If each elemental area is the same, then equation (4. 8a) can
be simplified to:
_ = _ . . . .
Se S2 SZ b3 n
(4. 8b)
Af
Where n = the number of equal areas, or - .
ct
Consider, as an example, a single filter tube as shown in
Figure 4.2. After cleaning, the distribution of local filter
drags will not necessarily be uniform over the entire bag
area. Such a non-uniform situation has been shown previously;
a further example is shown as curve S , in Figure 4. 3. For
°» •!•
this situation the average velocity was 1. 5 fpm and A P was
o, l
0.30 in. H2O. Using equation 4. 1, then, the velocity through an
elemental area, at 50% altitude (i.e., half-way up the bag) was
approximately
AP 0. 30 . , , IA o \
(4. 9a)
while the velocity at 90% altitude was approximately
uf = nre- = 5-oofp™ (4.9b)
The above are only approximate values since the filter drag of
the fabric was not included in S ,. This drag, plus the drag
of the dust would determine actual velocity profiles.
4.4
-------
"5
"2
Figure 4.2 ELECTRICAL ANALOGY FOR FILTER FLOW SYSTEM
In reference to Figure 4.3, it can be seen that the greater
air flow in the low resistance areas brought correspondingly
greater amounts of dust to the filter surface per unit of time.
This, in turn, increased filter drag at a greater rate, leading
to a "flattening" of the profile with time. As a result, a plot
of Se vs W would be inherently non-linear, despite a constant
dust permeability. The extent of this non-linearity would
depend on the uniformity of conditions after cleaning. These
same concepts can be applied to each filter tube in a compartment,
using the area of one tube as a basis. Likewise, they can be
applied from compartment to compartment in a multi-sectioned
baghouse using the entire filter area of one compartment as a
basic area. The mathematics of such a such a system has
been developed by Sargent.
For the sake of discussion, consider again the analogy shown
in Figure 4. 2, and assume dust permeability is a constant.
The rate of change of filter drag can then be written as follows:
dS
(4. lOa)
dS-
K
(4. lOb)
4. 5
-------
0.03 Q04 0.060.080.1 0.2 0.3 0.4 O.6 0.8 1.0
DUST MASS RESISTANCE (IN. H20/fpm)
2.0 3.0 4.O
Figure 4. 3 DEVELOPMENT OF DUST MASS RESISTANCE
PROFILE THROUGH A FILTRATION PERIOD
From equation (4. 1),
AP =
= etc.
(4.11)
Combining equations (4. 10) and (4. 11),
or
1 K
dS
dS
2 K
P
dS
_ ~"
~ar
(4. 12)
dS
_ f.
~ar
= etc.
(4.13)
Integrating equation (4. 13) from time t = o to time t = t,
C2 2 C2 2 C2 2 .
S - s = S - s = S - s etc.
(4.14)
4.6
-------
Where s., s?, s^, refer to filter drag values at time t = o.
Now if the volumetric flow is Q, and the total area is A,,
then
Q = A (U +U + n + . . . . + U,) (4.15)
1 1 12 *3 n
Combining equations (4. 10) and (4. 15),
KA /dS dS dS dS \
Q = ^- ( , + . + -,— + . . + .. n ) (4. I6a)
p \ t J
or,
QC U,. ,-C: dS. dS9 dS7 dS
P f(avg. ) P _ 1,2,3
= K --+- +
Integrating equation (4. 16) from time t = o to t = t
Q- C (t) n - n n = n
n = 1 n = 1
In summary, equation (4. 17) says that the total of all drag changes
that occur during a filtration period (i.e., ^ S ifs ) can be
calculated on the basis of inlet conditions (i. e. ,
The total of all drag changes, however, is not the same as the
change in effective drag, A&e, since the sub-changes are
arranged in parallel. This is true if we consider a single filter
tube or several sections in a multi-section baghouse. From this
we can conclude that control over baghouse AP will depend on the
distribution of filter drag as well as the total of all drag changes.
4. 7
-------
As a check on equation (4. 17) assume that at time t = o all
values of s = SR. Then, at time t = t all filter drag values
will equal S-, since filtration would have been at a uniform rate.
= n(ST) -n(Sp) (4. 18a)
f i *.
or, rearranging terms,
s, . - r -Vvs.)^-'
(4. 18c)
which is identical to equation (4. 3a)
C. Continuous Operation
In Part A and B we considered basic equations and assumptions
for a single incremental area and for a parallel arrangement
of areas, respectively. For section B it was necessary to assume
some undefined values of filter drag at the beginning of filtration.
In this section the equations will be extended to the continuous-
operating, multi-section baghouse. The same assumption of
constant K will be maintained.
For the multi-compartment baghouse, two operating schemes are
possible, depending on the number of compartments, filtration
time, cleaning time, and individual preference. These are as
follows:
1. One compartment maintained on a stand-by, so that when a
dirty section is taken off-stream for cleaning, a cleaned
unit is placed on-stream. The total number of compartments
equal n + 1, with n compartments always in service.
2. In most applications, all compartments are on-stream for
some finite portion of the first part of a cycle; one com-
partment is then taken off-stream, cleaned, and put back
4.8
-------
in service, such that one less than the total number of
compartments are filtering at the end of a cycle. Actual
performance will be similar to (1), provided cleaning time
is a small fraction of filtration time*. The following
treatment is based on operation (1); it is easily adopted to
(Z) for many practical situations.
As a beginning, consider a baghouse with n + 1 = 5. Assume
that each compartment is cleaned to the same degree, and
that filter drag after cleaning is equal for all sections and
is denoted as SR . An electrical analogy for the 4 sections
on-stream is shown in Figure 4.4. Because of the cyclical
nature of the operation (i. e. , cleaning each section in
sequence and to the same degree), a pattern of filter drag
values will exist after several cycles of operation.
Table 4. 1 illustrates the nature of the pattern.
H
co
»— i
CO
W
W
E ~ AP
R2~S2
R3~S3
R4
\
f
Q =
U
Figure 4. 4 ELECTRICAL ANALOGY FOR FOUR COMPARTMENT
BAGHOUSE
* NOT El For a multicompartment baghouse, the following terms will
be used:
1. Filtration time (t) - time required for a compartment to go
from Sp to S™.
Z. Cleaning time - time a given compartment is off-stream for
cleaning.
3. Cycle time (tc) - time between cleanings of two consecutive
compartments.
4. 9
-------
Table 4.1 PATTERN OF FILTER DRAG VALUES IN
MULTI-COMPARTMENT BAGHOUSE AT EQUILIBRIUM. CONDITIONS
Cycle
1
20
7 1
22
23
24
25
Start
Finish
Compartment
1 I 2
Off- R
stream
Start SR
Finish . SA
Start SA
Finish S_
D
Start S^
JD
Finish , S_
Start S-.
1
Finish
Start
Finish
ST
Off-
stream
SA
SA
SB
SB
sc
sc
ST
Off-
stream
SR
SA
3 4
SA SB
SB
SB
sc
sc
ST
Off-
stream
SR
SA
SA
SB
sc
sc
ST
Off-
stream
SR
SA
SA
SB
SB
5
sc
ST
Off-
stream s
SR
SA
SA
SB
SB
sc
sc
SC ST
= Residual filter drag = Drag after cleaning.
- Terminal filter drag = Drag before cleaning.
, Sg, S_ = Intermediate drag values.
4. 10
-------
For cycle 21, and substituting in equation (4. 14),
= = =
Solving simultaneous equations:
s2 s2 - s2 s2
SA - SR - SA - SR
s2 s2 - s2 s2
SA - SR ~ SB " SA
s2 s2 - s2 s2
SA - SR - SC - SB
s2 s2 - s2 s2
SA - SR - ST " SC
or, 4S^ - 4SR = S2 - S2 (4. 20)
Rearranging,
S" ,4 r< " o (T1 /^oi\
= 4S. - 3S (4. 21)
JL 2~\. XV
For a baghouse with n compartments on-stream at all times
(total of n = 1 compartments with one always idle)
equation (4. 21) can be generalized to:
S2 = nSJ; (n-l)S^ (4.22a)
or;
2
ST = N|nSA (n-l)SR (4.22b)
Where S. - filter drag at end of cycle for most recently
cleaned compartment (i.e., compartment 2, cycle 20;
compartment 1, cycle 21; compartment 5, cycle 22).
4. 11
-------
2
Dividing by SR, and rearranging terms,
29
SIC /C \ _L /-r, 1 \
A (b „,/£>.£.) T (n-JL;
= _Z 5^ (4. 23a)
n
In summary, equation (4.23) allows for the calculation of filter
drag at the end of one cycle of operation for the compartment
just placed in service, provided values for S™, Sp, and n are
known. S. is not a terminal drag. It is an intermediate value
on the Basic Performance Curve for the compartment in question.
In like manner it can be shown that:
(4.23b)
(4.23c)
We can generalize equations (4.23a, b, c) by letting x be the
designation of any compartment in a baghouse, such that
x = 1 for most recently cleaned compartment
x = n for compartment at terminal conditions.
Sx = Drag of any compartment at end of cycle.
Accordingly,
Repetitive solutions of equation (4.24) will give the pattern
of filter drags for each compartment at the end of a cycle. *""
Substitution of these values in equation (4. 8b) will then give S .
4. 12
-------
The use of equation (4. 7) will give P at the end of the cycle.
Further manipulation of equation (4. 8b) and (4. 7) will give
P at the start of the cycle.
By way of example, consider a baghouse under the following
conditions of operation:
U,, . = 3.0 fpm with n compartments on- stream
S =4.0 in. H2O/fpm
S_ - 0. 5 in. H,O/fpm
n
then,
(S /S )* + (n-1) 6
— T R - = L =11-5 (4.25)
no v '
x = 1 (11. 5) - 0 = 11. 5 S1 (4.26a)
and -— =3.4
= 2 (11.5) - 1 = 22 Q (4. 26b)
O~
and * = 4. 7
= 3 (11.5) - 2 = 32.5 S (4. 26c)
and _±- = 5.7
- )= 4 (11.5) - 3 = 43 S (4.26d)
R/ and -3 = 6. 6
S
/•s
I -iN = 5 (11.5) - 4 = 53.5 S (4.26e)
R/ and = 7.3
4. 13
-------
i\ = 6 (11.5) - 5 = 64 S6 (4.26f)
R/ and =8
R
Using equation (4. 8b),
eT
<4' 27a>
and
S
R _ .294 + .213 + .175 + . 151 + .137 + . 125 1.095
--- ----- -
e
T
Se = = 2. 74 in. H0/fpm (4. 27c)
Consequently, from equation (4. 7),
P = (3.0 fpm)(2.74 in.H2O/fpm) (4.28)
= 8. 22 in.H2O
At the start of a cycle we have physically substituted a
clean compartment for the dirtiest section, so that in
S6 SR
equation (4. 27a), -5— has been deleted and -p— added.
SR SR
Thus, equation (4. 27a) becomes,
6S
tt + Trr + Trr + iT5 + Tr3 + T!+ T 4 (4-29a)
R
and .'. , Se = = 1'52 in> H°/fPm (4. 29b)
4. 14
-------
Also,
P = (3. 0)(1. 52) = 4. 56 in. HO (4. 30)
In summary, the six -compartment baghouse would cycle
between 4.56 in. H?O and 8.22 in. H?O, assuming constant
flow. Flow variations could be taken into account provided
the inlet dust mass flow (grains/min. ) was constant.
D_ Approximate Solutions for Equation (4. 24)
While repetitive solution of equation (4. 24) will provide an
exact solution to the problem of predicting maximum and
minimum Ap's of a multi- compartment baghouse, its use
is cumbersone and time consuming. Therefore, a less
involved technique is desirable, and is derived as follows:
In equation (4. 24), let
(S /S )2 + (n-1) - 1
Z= — - - £— -- (4.31)
Equation (4.24) then reduces to:
'S
X ' = xZ + 1 (4. 3 la)
or,
S
X
= JxZ + 1 (4. 31b)
From equation (4. 8b)
c n = n
ni>p ~L S /c
-*= = VSx (4.32)
Utilizing equations (4. 31b) and (4. 32), tabular computations
can be run for selected values of Z, as shown in Tables 4.2
through 4. 7. For any value of n, in these tables, column 3
4. 15
-------
represents S_/SR while column 6 represents Se /S_. When
ST/SR is plotted as a function of Se /SR (FigureT4. 5) a family
01 curves is produced. Each of these curves (excepting n = 1)
is slightly non-linear for the lower values of S^/S^ „ As can
be seen, the general slope of these curves tena toward some
limiting values, with the most significant changes in general
slope occurring as n increases from 1 to 6.
If desired, a family of curves similar to those shown in
Figure 4. 5 could be used in place of equation 4. 24, with some
slight loss in accuracy. To facilitate its use, a larger chart
could certainly be prepared. An alternate procedure is
possible in situations wherein rapid calculation is of prime
importance; the basis for this approach is to approximate the
relationship between S—/S.- and Se.p/SR for each value of n by
a straight line. By way ofexample, we have chosen to draw
this line through the point where each curve intersects ST/SR= 5. 0
The slope of each line can then be calculated (designated as m
below) and an approximate relation between SeT,/Sp and S^/Sp
developed, so that IK
SeT=SR+mAS (4.33a)
(C -Urt)
= SR + m -^ - (4- 33b)
Values for m are listed in Table 4. 8.
4.16
-------
Table 4. 2. SOLUTION TO EQUATIONS (4. 31b) and (4, 32)
for
Z = 0. 5
1
X
1
2
0
4
5
6
7
8
9
10
11
12
13
14
15
2
xZ + 1
1. 5
2. 0
2. 5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7. 0
7 5
8.0
8.5
3
JxZ + 1
or
VSR
1.225
1. 414
1. 581
1. 732
1. 871
2.000
2. 121
2.236
2. 345
2.449
2. 550
2.646
2. 739
2. 828
2. 916
4
STJ/S
R' x
0. 816
0. 707
0. 633
0. 577
0. 534
0. 500
0. 471
0. 447
0.426
0.408
0. 392
0. 378
0. 365
: 0.354
0. 343
5
2 SD/S
R7 x
or
" SR/SeT
0. 816
1. 523
2. 156
2.733
3.267
3. 767
4.238
4.685
5. Ill
5. 519
5.911
6.289
6. 654
7. 008
7. 351
6
x i
^ sD/se
R' e-j'
or
VSR
1.225
1. 313
1. 391
1.464
1.530
1. 593
1.652
1.737
1. 761
1.812
1.861
1. 908
1.954
1.998
2.040
4. 17
-------
Table 4. 3. SOLUTION TO EQUATIONS (4. 31b) and (4. 32)
for
Z = 1.0
1
X
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
2
xZ + 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
3
x(xZ + 1
or
Sx/SR
1.414
1.732
2.000
2.236
2.449
2.646
2. 828
3.000
3.162
3.317
3.464
3.605
3.742
3.873
4. 000
4. 123
4
VSx
.707
. 577
.500
.447
.408
.378
.354
. 333
.316
.301
.287
.277
.267
.258
.250
.242
5
SSR/Sx
or
HSR/SeT
.707
1.284
1.784
2.231
2.639
3.017
3.371
3.704
4.020
4.321
4.608
4.885
5. 152
5.410
5.660
5.902
6
X
nsR/seT
or
SeT/SR
1.414
1.558
1.682
1.793
1.895
1.989
2.076
2.160 !
2.239
2.314
2.387
2.456
2.523
2.588
2.650
2.711
4.18
-------
Table 4. 4. SOLUTION TO EQUATIONS (4. 31b) and (4. 32)
for
Z = 2
1
:
X
1
2
; 3
I
4
5
6
7
8
9
10
11
12
13
14
15
16
2
xZ + 1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
33
3
JxZ + 1
\l ./V*— I 1 X
or
Sx/SR
1.732
2. 236
2. 646
3.000
3. 317
3. 605
3.873
4. 123
4. 359
4. 582
4. 796
5. 000
5. 196
5.385
5. 568
5. 744
4
SR/Sx
. 577
.447
. 378
. 333
. 301
. 277
.258
. 242
.229
. 218
. 208
. 200
.192
. 186
. 180
. 174
5
S S /S
•" ~a/ v
Jtx X
or
. 577
1.024
1.402
1. 735
2. 036
2. 313
2. 571
2.813
3.042
3.260
3. 540
3.740
3. 932
4. 118
4.298
4. 472
1
6
i
X
or
seT/sR
1.732
1.953
2. 140
2.305
2.456
2.594
2.723
2.844
2.958
3.067
3. 107
3.208
3.306
3.400
3.490
3.578
4. 19
-------
Table 4. 5. SOLUTION TO EQUATIONS (4. 31b) and (4. 32)
for
Z = 4
1
X
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
2
xZ + 1
5
9
13
17
21
25
29
33
37
41
45
49
53
57
61
65
3
JxZ + 1
or
Sx/SR
2.236
3.000
3.605
4. 123
4. 582
5.000
5. 385
5.744
6. 083
6.403
6.708
7. 000
7.280
7.550
7.810
8.062
4
O-|-| / S
K. x
.447
. 333
.277
.242
. 218
.200
. 186
. 174
. 164
.156
. 149
. 143
. 137
. 132
. 128
. 124
5
SSR/SX
or
.447
. 780
1.057
1.299
1. 517
1.717
1. 903
2.077
2.241
2.397
2. 546
2.689
2.826
2. 958
3. 086
3.210
6
X
IS /Se
K T
or
SeT/SR
2.236
2.564
2.838
3.079
3.296
3.494
3.678
3.852
4.016
4. 172
4.320
4.463
4.600
4.733
4.861
4.984
4.20
-------
Table 4. 6. SOLUTION TO EQUATIONS (4, 31b) and (4. 32)
for
Z = 8
1
X
1
2
xZ + 1
9
2 17
3
4
25
33
5 ' 41
6 • 49
i
7 57
8
9
10
11
12
13
14
15
16
65
73
81
89
97
105
113
121
129
3
JxZ + 1
or
Sx/SR
3. 000
4. 123
5. 000
5.744
6.403
7. 000
7. 550
8. 062
8. 544
9. 000
9.434
9.849
10.247
10. 630
11. 000
11.358
4
Xx XI
. 333
.242
. 200
. 174
.156
. 143
. 132
. 124
. 117
. Ill
. 106
. 102
. 098
. 094
.091
.088
5
SS^/S
R' x
or
.333
.575
. 775
.949
1. 105
1.248
1.380
1. 504
1. 621
1. 732
1. 838
1.940
2. 038
2. 132
2. 223
2.311
6
x
or
3.000
3.478
3.871
4.215
4.525
4.808
5.072
5.319
5.552
5.774
5.985
6. 186
6.379
6.567
6.748
6.923
4.21
-------
Table 4. 7. SOLUTION TO EQUATIONS (4. 31b) and (4. 32)
for
Z = 16.0
1
X
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
2
xZ + 1
17
33
49
65
81
97
113
129
145
161
177
193
209
225
241
257
3
x|xZ +1
or
s /sD
x' R
4. 123
5. 745
7. 000
8. 062
9.000
9. 849
10.630
11.358
12. 042
12.689
13. 304
13.892
14.457
15.000
15. 521
16. 031
4
STJ/S
R' x
0.242
0. 174
0. 143
0. 124
0. Ill
0. 102
0. 094
0. 088
0. 083
0.079
0.075
0. 072
o. 069
0. 067
0. 064
0. 062
5
S VSx
or
n VseT
0.242
0.416
0. 559
0.683
0.794
0.896
0. 990
1.078
1. 161
1.240
1.315
1.387
1.456
1. 523
1. 587
1. 549
6
X
"V^T
or
SeT/SR
4. 132
4.808
5.367
5.856
6.297
6.696
7.071
7.421
7.752
8.064
8.365
8.652
8.928
9-192
9- 582
10. 329
4.22
-------
Table 4. 8. VALUES FOR m AS FUNCTION OF n
1
n
1
2
3
4
5
6
7
8
9
10
m
1.0
0.78
0.72
0.68
0. 65
0.63
0.61
0.60
0.59
0. 58
n
11
12
13
14
15
16
17
18
19
20
m
0. 57
0. 56
- -
0. 55
-
- -
0. 54
- -
- -
0. 53
Consider then, the problem previously stated, wherein
U,., . = 3.0 fpm with n compartments on-stream
f(avg. )
ST = 4. 0 in. H2O/fpm
SR = 0. 5 in. H2O/fpm
n =6
4.23
-------
Substituting in equation (4. 33a)
Se = 0.5 + (0.63)(4. 0-0. 5) (4. 34a)
= 0. 5 + 2. 2 = 2. 7 in. H2O fpm (4. 34b)
and, from equation (4. 1)
P = 2.7(3. 0) = 8. 1 in. H7O (4.35)
max £
as compared to 8.22 in. H,O in previous calculations.
In view of other factors, the difference does not seem significant.
It can be said, then, that the maximum filter drag for a bag-
house with n units on-stream at all times is, within practical
limits, a direct function of the terminal and residual drags of
a single compartment, and a factor which is dependent on the
number of compartments.
An approximate design equation which can be used for all units
is:
(4.33)
Adjustments of Equation to Other Operating Conditions
As noted in the beginning of Section IV-C, the continuous operation
of multi-compartmented baghouses can take place in essentially
two ways:
1. Operation with n compartments on-stream at all times,
and a stand-by section available to replace a dirty unit.
2. Operation with n compartments for the first part of a
cycle, and n-1 compartments for the final portion of a
cycle.
Pressure differential readings from "the clean-air side to the
dirty-air side" for a cycle of operation in each of the above
cases can be represented by the schematics shown in Figure 4. 5.
For Figure 4. 5a, Ap increases in a non-linear manner until
the end of a cycle. At this time, a clean compartment is
substituted for a dirty compartment, and AP immediately
returns to the minimum value.
4.24
-------
5. 0
1=1
= 2
4.0
3 .0
.oi—
i.o I-
n= 4
1
L ..1 L 1 L
1.0 2.0 3. 0 4. 0 5.0 6. 0 7. 0
ST/SR
Figure 4. 5 DESIGN CURVES FOR MULTI-COMPARTMENT BAGHOUSE
-------
Q
a)
CD
CO
CD
h
a,
/
A
/
-SeT'Uf
One cycle of operation
r\ Compartments on-stream.
Time
Figure 4. 5a OPERATION ACCORDING TO CASE (1)
SeT-Uf
Q
CD
co
CD
CD
h
0,
Compartments on-stream
H - 1
Compartments
on-stream
One cycle of operation
Time
Figure 4. 5b OPERATION ACCORDING TO CASE (2)
4.26
-------
For Figure 4. 5b, Ap again increases in a non-linear manner.
Near the end of the cycle, however, a dirty compartment is
taken off-stream for cleaning. At this time AP jumps to a
higher level since the filter area has been reduced (i. e. , average
velocity increased through compartments remaining on-stream).
Filtration continues with n-1 compartments until the cleaned
unit is put on-stream. At that time AP falls to a minimum.
For this mode of operation, Se will be approximately equal
to that of Figure 4. 5a, provided cleaning time is not a significant
portion of filtration time. For example, consider a four-
compartment baghouse with a cycle time of 30 minutes and a
cleaning time of 2 minutes. In this case each compartment
will collect dust for a period of (4)(30) - 2 or 118 minutes.
Thus, the compartment will be off-stream approximately 2%
of the time; this should not seriously alter the pattern of filter
drags from that previously developed.
In comparing the two modes of operation some slight confusion
is possible regarding the (weight of dust)/(square foot of fabric)
collected in a given compartment. For all cases this can be
calculated as follows:
W = C • U,, x-t (4. 36)
p f(avg.) v '
where
t = filtration time; i. e. , interval between the time a clean
compartment is put on-stream and the time taken off-
stream.
Characterization of Filter Units
To complete this discussion of "Mathematics for Fabric
Filter Operation" it is necessary to develop those equations
which will allow for a field determination of critical
characteristics. These characteristics are:
1. Average residual drag of a single compartment (SR)
2. Average terminal drag of a single compartment (S_,)
3. Residual velocity (Uf )„
4.27
-------
The importance of S-. and S — has been shown. The significance
of Uf will be demonstrated in latter sections. At this time,
however, Ufp can be defined as that velocity which exists in a
cleaned compartment immediately after it has been put in
service. As such, it may be several times greater than U^, .
and, therefore, may be important in determining filter °' '
efficiency, dust permeability, and filter blinding. These
critical operating characteristics may be determined from
pressure differential measurements taken across two points
common to the inlet and outlet sides of all compartments, if
the total flow through the system is known. Such measurements
will produce a curve of the type shown in Figure 4. 6. At time
t on this figure a freshly cleaned compartment has come
on- stream, and during the interval t -t, filtration proceeds with
a given number of compartments (n). When t, is reached one
compartment is at its terminal condition and is taken out of
service for cleaning. The pressure differential immediately
increases from ^P to Ap During this instant there is no
change in the drag 01 the remaining units as the flow through the
filters is laminar. Filtration then proceeds with n-1 compart-
ments until the cleaned filters are back in service. At this
time (tc) the pressure differential decreases from ^PA to Pi .
Again, there is no change in the drag of the units continually
filtering. These instantaneous changes in pressure differential
are indirect measurements of terminal and residual drag in a
particular compartment and can be used with the relations which
follow to characterize a particular mode of operation.
Thus, at time t^ and pressure differential ^F^' equation (4. 8b)
can be rewritten as follows:
(4. 37)
According to the mode of operation the compartment with
drag = S™ is now taken off -stream, so that at time = t, and
pressure differential
(4.38)
Combining equations (4. 37) and (4. 38)
4.28
-------
o
c^
«
CO
W
u
3
1—4
H
W
W
fa
H
tf
w
tf
P-
TIME
Figure 4.6. SCHEMATIC PRESSURE DIFFERENTIAL
CURVE FOR MULTI-COMPARTMENT BAGHOUSE
This then allows for calculation of S— from measurements of
Sp-r, and S« .
er e
At time t and pressure differential A P,,
(4.40)
Again, according to the mode of operation, a compartment
with filter drag So is now brought on-stream, so that at
time t and pressure differential
n
+
R
R
(4.41)
4.29
-------
Combining equations (4.40) and (4.41) produces an equation
which allows for the calculation of Sj^, as follows:
1 n n"1- (4.42)
4
R eR "e
As an example, consider a baghouse with n = 6 and the following
values of AP and Uf:
AP. = 5.32 in. H_O Uf =3.5 fpm
1
AP2 = 8.22 " Uf =3.0 "
AP3 = 8.60 " Uf =3.4 "
AP. = 8.70 " Uf =3.4 "
4 4
Effective filter drags can then be calculated for the baghouse.
Thus,
Se. = Se = ^iH = 1.52 in. H,O/fpm
•L R. J • D L*
\\ \\
Se = ^4° =2.53
e3 3.4
Se4=«0°=2.56 " -
Using equation (4. 39)
1 6 5
2T74~ " Z75T
= 0.22
and ST = 4. 5
Using equation (4. 42),
1 - 6 5
' TT5T "
and S_ =0.5
XV
4.30
-------
In summary, ST = 4. 5
SR = 0. 5
AS = 4. 0
For the sake of discussion, assume that cycle time is equal
to 30 minutes for the above problem. That is, filtration time
equals 6(30) or 180 minutes. If, in an attempt to lower AP,
cycle time is reduced to 20 minutes, the following will result:
1. Filtration time will be reduced to 6(20) or 120 minutes.
2. From equation (4. 3b),
„ ... 120 ., ,_
s = 4-° 2-67
3. From equation (4. 33b), and Table 4. 8
SeT = 0. 5 + 0. 63(2.67) = 1.68 in. H2Q/fpm
4. At U, =3.0 fpm, therefore,
A P? = l. 68(3. 0) = 5. 1 in. H?O compared to the original
8. 22 in. H2O.
In a similar fashion the equation could be used to predict various
values of A P as n is varied.
As mentioned in the beginning of section IV-F, a characteristic
of bag filters which undoubtedly influences filtration is the
filter velocity itself. For any two units operating with the
same Uf, ., entirely different velocities of filtration may
exist thrWjfn 'each compartment. Each individual pattern
will be a function of S™, S^ , and n.
1 iv
The velocity in a compartment at any time can be estimated,
provided the filter drag of the compartment (S ) and the effective
drag of the baghouse (S ) is known at that time. Thus, for any
compartment,
S = *j£- (4. 43)
-A. \J £
X
4. 31
-------
While for the entire baghouse,
AP
6 ^Vvg.)
therefore, since AP is the same,
(4. 44)
(4.45)
It would appear that a particularly critical value of D£
would be the highest value reached in a compartment, x
This would be the velocity in a clean compartment as it is
just put on-stream, and is designated Ur
R
Accordingly,
S
eR
U
f(avg. )
(4.46)
An approximate velocity pattern for a baghouse with n = 6 is
shown in Figure 4.7.
AVERAOC VELOCITY : 3.0 lorn
RESIDUAL COMPARTMENT niTEH 'J«AG ! CW IN. H O/tpm
TERMINAL COMPARTMENT FII.TEH IRAS z.ow. Mjo/ipm
COMPARTMENT
6
30 40 50
TIME (MINUTES)
Figure 4.7 VELOCITY PATTERN IN 6-COMPARTMENT
BAGHOUSE AS FUNCTION OF TIME
4. 32
-------
The fact that such velocity variations exist would appear to
emphasize the need to gather data on filtration through actual
field studies of multi-compartment units, or by studies of
multi-compartment pilot plants. The data obtained by techni-
ques outlined in Sections I and II must certainly be used with
caution. This is not to say, however, that a better understanding
of filtration cannot be achieved by such basic studies.
G. Summary of Useful Equations and Tables.
AP
f
1
ST
1
5~ ~
**•
n
Se
T
n
s —
eR
n- 1
e3
n - 1
64
(4.1)
(4.2)
1 '• ^- + . . • • + g^- (4. 8b)
12 n
(4.24)
= SD + mAS (4. 33a)
~ t\.
AW = C • U,, .-t (4. 36)
P f(avg.) '
(4.39)
(4.42)
S
Uf - --- U,, . (4.46)
fR -SJ- f(avg.)
4.33
-------
Table 4. 8. APPROXIMATE VALUES FOR m AS FUNCTION OF n
n
1
m
1.0
2 1 0.78
3
4
5
6
7
8
9
10
,
0.72
0.68
0.65
0.63
0. 61
0.60
0.59
0.58
n
11
12
13
14
15
16
17
18
19
20
m
0.57
0. 56
- -
0. 55
- -
- -
0. 54
- -
- -
0. 53
4.34
-------
26
SECTION 26
Effect of Changing Permeability,
Varying Flow Rate, and Non-Laminar
Head Loss
-------
EFFECT OF CHANGING PERMEABILITY,
VARYING FLOW RATE, AND NON-LAMINAR HEAD LOSS
Id*-Hi conditions for a baghouse include con-
stant inlet conditions, a constant dust per-
meability, and the absence of non-laminar
head loss elements such as valves, duct
work, i'tc. In this section we will explore
the pffc.-i t of certain of these parameters on
the theories so-fax- advanced, especially in
relation to predictions of baghouse AP.
A Graphical Solution to Design Equations
To account for a non-linear basic
performance curve, it will be necessary
tn develop a graphical solution to the
equations of Section IV.
In Section IV, it was shown that under
( onstant inlet conditions the total of all
drag changes will be the same from cycle
to cycle (Equation 4. 17). As discussed,
!.hr changes in drag from compartment to
compartment will follow a definite pattern
(j',jb]e '\. 1), with the greatest change
occurring in the compartment just put
on-stream. The equation for drag of any
compartment at the end of a cycle was
derived in Section IV (Equation 4. 24vi; this
drag also represents drag in the same
compartment at the beginning of the next
cycle. This concept is shown in graphical
form in Figure ,r>. 1.
The vertical scale of Figure 5. 1 represents
filter area expressed as percent of the
total. Thus, using the example of Section
iV-C, where n G, each filter area
represent.:--, 100 or Hi. 7% of the total.
The horizontal scale represents the filter
drag ratio S, /S|^ as defined by equation
(I.Li'O. if, Ihcn, V/L calculate SX/SR for
each compartment at the end of a cycle,
a profile of drag values within the bag-
house can be plotted. This is shown for
the example previously cited, as the
right-hand boundary of the cross-hatched
sections ni Figure 5. 1. A profile of drag
values at the start of a cycle can also be
represented. Thus, for the compartment
just cleaned, ST/SR= 1.0. In addition,
it is known that the drag of one compart-
ment at the end of a cycle equals the
drag of the adjacent compartment at the be-
ginning of a cycle. The result is the left-
hand boundary of the cross-hatched sections
on Figure 5. 1.
The cross-hatched areas represent the
change in drag for each compartment of
a multicompartment unit during one cycle
of operation. From cycle-to-cycle a
given compartment will proceed, in
sequence, through the incremental
changes shown.
Now, if r; is increased beyond 6, the
incremental changes per compartment
will become smaller and smaller. Ifr;
is allowed to increase indefinitely, the
cross-hatched areas will approximate a
smooth curve. This curve represents
the Drag Profile in a multicompartment
baghouse with an infinite number of filter
areas. For such an installation, the
pressure drop would be constant.
For convenience, equation (4. 24) has been
solved for n - 10 and selected values of
ST/SR. The results are summarized in
Table 5. 1. By plotting this data and
connecting the points with smooth curves,
it is possible to generate a family of
curves for r\ °° as shown in Figure 5. 2.
By reversing the "construction" process
values of SX/SR can easily be determined
giving the solution to equation 4. 24 by
graphical means.
B Effect of Varying Permeability
As developed, the drag profiles shown
in Figure 5. 2 represent conditions when
K is constant. If K is not constant, but
assuming it follows some regular pattern,
a drag profile will still exist, but its
shape will differ from those shown. It
-------
100
80
60 ...
g
o
M
0)
O.
a 40
Q)
M
M
I)
20
I I I I I I I I I 1
13
3
3.
3
TO
s
fD
P
cr
1.0 2.0 3.0 4.0 5.0 6J3 7.0
ao
Filter Drag Ratio, S /
V SR
Figure 5.1 Drag Profile with S / = 8.0 and n = 6
T/SR
-------
Table 5. 1 SUMMARY OF SOLUTION TO EQUATIONS (4. 24)
SELECTED VALUES OF S , AND r) = 10.
*/*
2
3
4
5
6
7
8
1
1. 140
1. 342
1. 581
1. 844
2. 121
2.408
2. 702
2
1. 265
1. 612
2.000
2.408
2. 828
3. 256
3. 688
3
1. 378
1. 844
2. 345
2. 864
3.391
3.924
4.461
4
1.483
2. 049
2. 646
3.256
3.873
4. 494
5. 119
5
1. 581
2. 236
2. 915
3. 606
4.301
5.000
5. 701
6
1. 673
2. 408
3. 162
3.924
4. 690
5.459
6.229
7
1. 761
2. 569
3. 391
4. 219
5.050
5. 882
6. 716
8
1. 844
2. 720
3. 606
4. 494
5. 385
6. 277
7. 169
9
1. 924
2. 864
3. 808
4. 754
5. 701
6. 648
7. 530
10
2.000
3. 000
9.000
5.000
6. 000
7. 000
8. 000
-------
Figure 5.2 Ideal Drag Profiles for r, = °° and Selected Values of S /
V i
d
OJ
a
m
ai
o.
100
80
60
2.0 3.0 4.0 5.0 6.0 7.0 8.0 = S.
3 40
20
n
0
3
q
3'
ffq
ft)
3
u
cr
1.0
3.0 4.0 5.0
7.0 8.0
-------
Effect of Changing Permeability
follows, then, that by assuming various
shapes for the drag profile, different
drag changes per compartment per cycle
can be graphically determined. With this
information, velocities in each compart-
ment at the beginning and end of a cycle
can be calculated and eventually the Basic
Performance Curve for a given compart-
ment cn.n be approximated.
By way of example, assume the following:
ST - 0. G in. H2O/fpm
SR 0. 2 in. H2O/fpm
Uf -
4 fpm with n compartments
on-stream
Kur-ther a.ssuine four possible variations
in the Drag- Profile ranging from the ideal
(Cast- A) to an extreme variation (Case D),
as illustrated in Figure 5. 3. By means of
previously derived equations Serp/So and
•VIJ/SR can be calculated; also, assuming
;i negligible but finite cleaning time,
ij
S(:./SR can also be determined.
Since Uf i.s constant, values of AP can
then lie calculated. These results are
summariy.ed in Table 5. 2, wherein the
subscripts are the same as used on
Figure (4. (i). Approximate AP curves
are shown in Figure T>. 4.
From value:, of Se,,,/,So and Sfp/Sp,.
and using equation (4.45), filter velocities
c an be calculated for each compartment
;it the beginning and end of a cycle. These
:irc shown in Figure r>. T>, with approximate
connecting curves. From these curves,
nverage velocities can be determined for
;ich compartment allowing for a calcula-
lion of AW. liy plotting Sx as a function
of AW the Basir: Performance Curves
of Figure .5. >i arc obtained.
It can be seen from the above illustrations
that as the Drag Profile is skewed to the
right:
1 Baghouse AP generally increases
and the magnitude of the jump from
AP2 to APg increases.
2 Residual velocity (UfR) increases,
resulting, perhaps, in increased
bleeding.
3 The Basic Performance Curve becomes
further removed from the "ideal"
straight line between SR and S'p.
If desired, any Basic Performance Curve
can be matched to a Drag Profile by
"trial and error". The nature of the
basic data and the accuracy required would
determine the desirability of such an
approach.
C Use of Equations (4. 38) and (4. 41) when
K is not Constant
Of some importance is the effect of a
non-linear Basic Performance Curve on
determination of S-p and Sp^ from a bag-
house AP curve. The original derivation
was based solely on changes in AP, as a
compartment was taken off-stream and
then returned to service. The jumps in
AP (Figure 4:6) were shown to be
functions of ST and S^.
To demonstrate the effect, equations
(4. 38) and (4.41) have been solved using
the data of Table 5. 2. The results of
these calculations are summarized in
TabJe 5. 3, indicating that the non-linear
Basic Performance Curve has little
effect on determinations of S-p and SR
from baghouse AP.
It should be noted, however, that the use
of such results to predict AP at other
operating conditions would treat K as a
constant. The predictions, therefore,
would be on the low side.
-------
Effect of Changing Permeability
c
01
CJ
n
-------
Effect of Changing Permeability
Table 5. - SUMMARY OF BAGHOUSF FIl.rER DRAG AND PRESSURE
DIFFERENTIALS FOR HYPOTHETICAL, CASES
Case A
Case R
Se2;
bR
2. 20
2.45
2.61
3e3/s
R
2. 06
2 . 36
o r -^
1 . 36
1. 48
1.5,
Po
1. 76
_^ __
1. 96
2.09
Po
2.06
2. 36
2. 53
i L'ase D
D Kffeet of Fan Characteristics on
Baghouse Performance
Situations wherein the rate of dust
entering a baghouse (grains/ min. ) and
the volumetric flow rate are variable,
will, of course, differ from eases per-
\ iou:,tv discussed. Treatment of the
data in these instances will depend on
tlii specific installation. A special case
whii h can be analysed in a general
1'ash'on, however, exists when the rate
•i( dust ma •;> entering a baghouse is
constant and \-olumetric flow variations
arc a result of tan characteristics. For
;his situation, the theories so far
developed will apply relative to filter
drae. 11 is emr intention, ho.vever, to
demonstrate how fan characteristics
•.oil influence baghouse AF and ^Tf-
Two extreme cases era be visualized in
looking at a baghouse as part of an overall
process:
1 The magnitude of energy losses other
than tliat associated with the baghouse
is large, so that changes in baghouse
drag do not materially affect fan
output.
2 The magnitude of energy losses other
than that associated with the baghouse
is negligible, so that fan performance
is basically related to changes in
filter drag.
It is the latter case that concerns us
hert, Consider, for example, that
a curve similar to Figure ^4. G) has
been predicted on the basis of constant
flow rate, and farther assume that a
-------
Effect of Changing Permeability
o
IN
3.0
2.0
.0
It
-O
Case A
iiii
3.0
2.0
.0
OQ
-O
Case B
i i i
Time
Time
•rH
T3
a
en
en
-------
Effect of Changing Permeability
Figure 5.5 Velocity patterns in hypothetical problem
EJ
o.
8.0
4.0
0
8.0
4.0
8.0
4.0
A
Case A
Case B
Case C
01234 5
Time
01 2345
Time
012345
Time
8.0
4.0
0
Case D
012345
Time
-------
Effect of Changing Permeability
0.8
0.4
Case A
12
16
20
e
o.
00
n
n
o
0.8
0.4
0.8
0.4
12
16
20
12
16
20
0.8
0.4
12
16
20
Dust Mass ( graines/ft )
Figure 5.6 Basic performance curves for hypothetical problem
10
-------
sffVci of Changing Permeability
TABLE 5. 3 VALUES OF ST AND SR
CALCULATED FROM FIGURE 5.4
USING EQUATIONS (4. 38) and (4. 41)
Original Values
Cast.- A
Case B
Case C
Case D
ST
0. 60
0. GO
0. 58
0. 56
0. 59
SR
0. 20
0. 20
0. 19
0. 20
0. 20
"backwardly inclined" fan has been
chosen such that it will deliver its
"rated capacity" at AP2-
From the fan performance curve,
(Figure 5. 7) it can be seen that as
the static pressure at the fan decreases,
fan output increases. Therefore,
when a clean compartment is put on-
stream and AP decreases, fan output
will increase. This, in turn, will
increase Uf so that baghouse AP will
come to equilibrium at some value
higher than predicted when considering
constant flow rate conditions.
Pressure to Fan Discharge is propor-
tional to the effective drag of the
baghouse.
By means of this proportion, it is
possible to estimate flow rates at
various values of Se- For example,
consider Case D of the previous
section, wherein:
0.404 in. H2O/fpm
0. 558 in. H2O/fpm
Further assume that Uf_ 4 fpm and
that the Minimum Design Flow Rate
isAf Uf2.
When a cleaned compartment is put
on-stream, then the ratio Fan Static
Pressure/Fan Discharge will equal:
0. 404
0. 558
0. 724
(5. 1)
Entering Figure 5. 8 from the right.
Fan Discharge is determined as 113. 5%
of Minimum Design Flow Rate.
Therefore:
u
(113. 5)
fi = •±-u^'» -Too-
- 4. 55 fpm
(5.2)
l''or convenience- in correcting for
inc reused flow rate, we have first
replottrd the Static Pressure curve
of Kigure 5. 7 in generic terms, such
lhat Kan Discharge is expressed as a
percent of Minimum Design Flow Rate
and Static: Pressure expressed as a
percent of Rated Static Pressure (See
Figure 5. 8). It is then possible to
construct a second curve with the ratio
of Static Pressure to Fan Discharge
plotted as a function of Fan Discharge
(also shown on Figure 5. 8). For the
system under consideration (i.e.,
essentially all head loss associated
with the baghouse) the ratio of Static
and:
also:
= 0. 404 (4. 55)
1. 84 in. H2O
( 5. 3)
0 404
UfR = -—y- (4.55) = 9.2 fpm (5.4)
Under conditions of constant velocity, it was
shown that:
and;
AP 1. 63 in. H2O
UfR = 8. 2 fpm
11
-------
Figure 5.7 Fan performance curve
10.0
10
ID
OQ
5'
crp
o
CN
33
to
*j
CO
8.0
6.0.
4.0
2.0.
Mechanical
Efficiency
Brake Horse Power
50
10 15 20 25 30
Fan Delivery ( 1000 cfm )
35
W
rH
0)
,
O
01
M
cu
I
en
Cl)
CO
M
O
X
a
H
P5
rc
tu
cr
-------
Figure 5.(
Static pressure curve
120
ICO
80
60_
40 .
20 _
Siaiic Pressure
Static Pressure: Discharge Ratio
1.0
0.8
0.4
0.2
0.
01
oc
20
40
60
80
100
120 140 160
Fan Discharge ( Percent DF Minimum Design Flow Rate )
-------
Effect of Changing Permeability
It can be concluded, then, that the use of
a "backwardly-inclined" fan has the effect
of increasing AP, and increasing Uf• ,
compared to constant flow-rate conditions.
It should be noted that the increase in Uf
occurs at a time when the system is sus-
ceptible to bleeding. This is especially
significant when "velocity profiles" along
the length of a filter bag are considered
and if variations from filter to filter are
considered.
E Influence of Non-Laminar Head Loss
In many fabric filter units, resistance to
flow is predominantly associated with the
dust-fabric combination per se. In other
units, significant energy loss can be
associated with the entire baghouse, but
it is still possible to determine that
portion of the overall AP associated with
the dust-fabric combination because of
common inlet and outlet chambers. In
still other units, significant resistance
to flow is caused by valves, duct-work,
elbows, etc. in series with each compart-
ment. In these cases, the volumetric
flow rate through a compartment will be
influenced by the added elements. The
equations and concepts previously derived
will not be applicable since the energy
loss associated with these components will
be approximately proportional to (Uf)2
instead of Uf.
By way of example, consider one section
of a baghouse, in which the resistance of
the dust-fabric combination is S and the
resistance of other components is R. The
pressure differential across each compon-
ent, then, is:
AP (dust-fabric) r Uf
(5.5)
AP (valves, elbows, etc.) = Uf2 • R (5.6)
Head loss across the entire compartment is:
AP total AP + A
2
Uf S + U/R
(5.7)
(5.8)
and the resistance of the compartment
becomes:
Y =
AP total
U.
= S + Uf R
(5.9)
and is seen to be a function of U*.
The addition of this resistance which is
a junction of Uf would, then, act as an
equalizer in the system. It would, for
instance, result in lower values of
which, in turn, would reduce the rate
of dust deposition in the most recently
cleaned compartment. The baghouse would
reach equilibrium, but at a higher AP and
showing less variation from AP^ to APj.
If desired, the non-laminar head losses
can be introduced in the derivations of
Section III. The resulting equations, how-
ever, would, require a computer for their
solution.
To illustrate the effect of non- laminar
head loss (either in series with individual
compartments, or in series with the entire
filter area), as the average filter velocity
(Uf avg. ) is increased or decreased,
consider the following illustration:
Assume a four compartment baghouse in
which S for each compartment is 1, 2, 4
and 6 in. H2°/fpm, and Uf = 3. 0 fpm.
Also assume R is the same for each
compartment and that R - 0. 5
in.H2O/(fpm)2.
For the entire baghouse, then:
R)
(5. 10)
Also,
jl
"§„
And,
1.917
2.09
(5.12)
From equation (5. 10),
Ye = 2.09 + (3. 0) (0. 5) = 3. 59 in. H2 O/fpm
(5.11)
14
-------
Effect of Changing Permeability
and,
AP = 3. 59(3. 0) 10.77 in. H2O (5.12)
If Uf is reduced to 1. 5 fpm, then
Ye 2. 09 + (1. 5) (0. 5) 2. 84 in. H0O/fpm
LJ
(5. 13)
and,
AP =2.84(1.5) - 4. 26in. H2O. (5.14)
Thus, in the example cited above, reduc-
ing the flow rate by the ratio 1/2 has
reduced AP by the ratio 1/2. 5.
F Discussion
In this Section we have attempted to show
the effect of several parameters normally
existing in any installation that create
"non-ideal" conditions. Each of these was
tested idependently; no attempt was made
to assess their combined impact on pre-
dicted operating pressure differentials or
other potential problems. Such an
analysis, of course, would be specific to
a given installation.
Perhaps the area of greatest need is
further data on the extent of non-laminar
head loss for the "typical" bag filter unit.
In present designs the head loss must be
at least 2 3 inlet velocity heads, with
perhaps an average of 4 5. Under such
conditions the "typical" baghouse will
automatically have a pressure drop of
say 3 inches H2O without the filter tubes
in place. It would seem that changes in
design could reduce this figure considerably.
15
-------
27
SECTION 27
Electrostatic PreclpHators Operation
and Industrial Applications
-------
ELECTROSTATIC PRECIPITATORS
OPERATION AND INDUSTRIAL APPLICATIONS
H. L. Engelbrecht
I Introduction
A. Particle Charging
B. Particle Transportation
II Influences on the Performance of an
Electrostatic Precipitator
A. Migration Velocity
B. Specific Collecting Surface
C. Particle Size and Dust Concentration
D. Gas Velocity
E. Electrical Wind
F. Dust Resistivity
III Special Design Considerations
A. Discharge Electrodes
B. Collecting Electrodes
C. Rapping Cycle Control
D. Electrical System Control
IV Operation of an Electrostatic Precipitator
A. Gas Conditioning
B. Gas Distribution
V Typical Industrial Precipitator Applications
A. Precipitators for Flue Gases from
Power Stations
B. Precipitators for the Iron and Steel
Industry
C. Precipitators for the Cement Industry
D. Precipitators for the Chemical and
Non-Ferrous Metallurgical Industry
E. Precipitators for Waste Gases from
Incinerators
VI Summary
-------
ELECTROSTATIC PRECIPITATORS
OPERATION AND INDUSTRIAL APPLICATIONS
Heinz L. Englebrecht*
I Introduction
In general, dust collectors can be class-
ified as dry process mechanical collectors
(cyclones, bag filters), wet process
mechanical collectors (scrubbers), or dry
or wet process precipitators (electrostatic
precipitators). Selection depends largely
on the specific dust collection problem,
the economical situation, and the suitabil-
ity of the equipment for the user.
Each one of these collectors has a prefer-
red range of application as far as the
required collecting efficiency is concern-
ed.
Figure 1. COLLECTING EFFICIENCIES
Wet-Process
vtttftM^ Precipitator
Dry-Process
precipitator
Wet-Process Mechanical Collector
"""""*
I Dry-Process Mechanical Collector
Collecting Efficiency
Since these ranges overlap at certain
efficiencies, the selection of the equip-
ment is limited by economical considera-
tions .
Total Cost
f////////Ah\\\\\\\\\\\\\VM Mechanical Collector
7////////////////////K88&
Investment
Cost
Electrostatic Precipitator
Operating Cost
Figure 2. COST FOR DUST COLLECTING
EQUIPMENT FOR EQUAL TOTAL
COSTS
Other factors to be considered are code
requirements and the availability of
electric power, water, compressed air, and
waste treatment facilities.
Electrostatic precipitators have advantages
that in principle they can be designed for
any desired degree of collecting efficiency
and they will deal effectively with dust
or moisture particles in the sub-micron
range. The dust is recoverable in either
dry or wet state and flow resistance and
power consumption are usually less than
for all other methods of dust collection.
Industrial electrostatic precipitators are
normally single stage (charging and collect-
ing of the particles in the same stage or
electrical field)precipitators. The number
of fields in series or the number of elect-
rical sections in parallel will vary with
the size of the precipitator.
Electrostatic precipitators can be class-
ified according to the mode of operation
as either dry process precipitators or
wet process precipitators; according to
the collecting surface as either pipe type
precipitators or plate type precipitators;
or by direction of the inlet gas flow as
either horizontal gas flow precipitators
or vertical gas flow precipitators.
The predominant group is the dry process,
plate type, horizontal gas flow precipi-
tator as shown in Figure 3.
meters
Consulting Engineer, Wheelabrator Corporation
Figure 3. DRY PROCESS, PLATE
TYPE HORIZONTAL GAS
FLOW PRECIPITATOR
-------
Electrostatic Precipitators Operation and Industrial Applications
Electrostatic precipitation requires a
discharge electrode (usually negatively
charged) of small cross sectional area
(such as a wire), and a collecting elect-
rode (usually positively charged and at
ground potential) of large surface area
(such as a plate or tube).
In operation, a unidirectional (DC) high
potential electrical field of the order
of 10 to 70,000 volts Is set up across'the
electrodes. The dirty gas stream is passed
between these electrodes. A typical gas
passage is shown in Figure A.
Normally a negative charge is applied to
the discharge electrode of an industrial
electrostatic precipitator.
As the ions move, they attach themselves
to the neutral dust particles causing them
to migrate toward the appropriate electrode.
Most of the dust particles pickup negative
ions. This is because the negative ions,
moving toward the positive collecting elect--
rode, have a greater distance to travel than
the positive ions which are formed near and
migrate toward the negative, or discharge
electrode. As the charged particles contact
the electrodes, they become neutral after
which it is relatively simple to remove them
by rapping, washing or gravity flow.
Figure 4. TYPICAL GAS PASSAGE
IN AN ELECTROSTATIC
PRECIPITATOR
At certain critical voltages, ionization
of the air molecules (dissociation into
negative and positive ions) takes place at,
or near, the surface of the discharge or
negative electrode and is evidenced by a
"corona" (no sparking). Figure 5 shows
negative and positive corona discharges.
Negative Positive
Figure 5. CORONA DISCHARGE
Figure 6. CHARGING OF PARTICLES
Therefore, there are four steps involved
in electrostatic precipitation: (1) elect-
rically charging the particles by means of
ionization, (2) transporting the charged
particles to a collecting surface, by means
of the force exerted upon them by the
electrical field, (3) neutralizing the
electrically charged particles precipitated
on the collecting surface, and (4) removing
the precipitated particles from the collect-
ing surface to a receptable external to the
precipitator.
A. Particle Charging
If there is a sufficient source of ions,
the number of charges which a dielectric
spherical particle can obtain is:
E r*
o
(1)
where n • saturation number of charges
8 on a particle
Eo - electrical field strength
-------
Electrostatic Precipitators Operation and Industrial Applications
r = particle radius
e = electronic charge = 4.8 x
10l° esu
K = dielectric constant (for
minerals, K = 1 to 10; for
water K = 80)
A maximum of three times as many charges
can be placed on a conducting particle
compared with an insulating particle.
However, since neither extreme is gen-
erally found, the number of charges on
closely sized particles of varying
dielectric constant is about the same.
For example, with a precipitation field
of about 4,500-volts per cm, a 1-micron
particle acquires a charge of the order
of 250 electronic charges, while a 10
-micron particle acquires about 25,000
electronic charges.
The electrostatic force on the 10-
micron particle is about 300 times
the gravitational force, assuming a
unit density spherical particle.
(See Figure 7.)
The existence of these relatively
large forces is what accounts for the
great effectiveness of electrostatic
precipitators to separate fine particles
from gases. Because these forces are
exerted directly on the particles
themselves, there is no need to exert
the necessary inertial separating
forces upon the particle by turning
the stream-lines and increasing the
velocity of the carrier gas as is done
in mechanical collectors.
B. Particle Transportation
The driving force Fs separating charged
particles from the gas is:
•rl
§
While the maximum number of charges is
time independent, the actual number
depends upon the rate at which a part-
icle is charged., A precipitator will
create about 10 to 10° ions/cm , in-
dicating that particles are fully
charged in 0.1 second or less.
Charges also accumulate on particles
due to thermal diffusion of ions in
the gas. The rate depends on ion den-
sity, ion mobility, and time, and is
significant only for particles less
than 1 micron in diameter.
F = n e E
s s p
(2)
10 XG
10 _
10 -
10
u 0. 1 1 10 100
Particle Diameter, microns
Figure 7. ELECTROSTATIC FORCliS ACTING ON
PARTIULKS UNDER TYPICAL PRECIPITATOR
CONDITIONS.
where nse is the total charge of the
particle and Ep represents the collect-
ion field intensity. This force F is
counteracted by a resistance force Fj,
which in the region of Stoke's law is:
Fd = 6 IT r y W (3)
where: r = radius of particle, y =
viscosity of the gas, and W = velocity
of the particle.
The particles are accelerated by the
electrical force until the resistive
force of the gas just equals the
driving force, i.e. Fg-Fj.
n eE = 6ir r y W
s p
(4)
The dust then travels at essentially
constant velocity to the collecting
surface. The velocity of the particle
then is:
W
n e E
s P
6 TT r \i
(5)
For particles larger than 1 micron
in diameter, the total charge is
n e - D E r
s o
where: D - 1 + 2 [ K - 1
K + 2
(6)
-------
Electrostatic Precipitators Operation and Industrial Applications
Therefore, the final migration velo-
city is
W = D Eo Ep
(7)
6 TT y
where: W = Migration velocity, cm/sec
E = Peak effective electrical
0 charging field, esu/cm
E = Average electrical field
p at collecting electrode,
esu/cm
r = Particle radius, cm
y = Gas viscosity, poise
II INFLUENCES ON THE PERFORMANCE OF AN
ELECTROSTATIC PRECIPITATOR
The collecting efficiency of an electro-
static precipitator follows an exponential
function: .
-fW
E = 1 - e ^ (8)
where: E = efficiency
e = base of natural logarithm
A = collecting surface
Q = gas volume
W = migration velocity
^
Setting jr = f (specific collecting surface)
we have: „
E = 1 - e-f W (9)
Therefore, the collecting efficiency is
a function of the specific collecting
surface f (or the collecting surface A
and gas volume Q), and migration velocity
W.
A. Migration Velocity
The migration velocity of a given particle
size is expressed by equation 7 as:
E E
D o P r
6 TT y (7)
W
With EQ being a function of the discharge
current i , the migration velocity also
becomes a function of the discharge cur-
rent.
With increased current the probability of
the dust being charged increases and thus
the collecting efficiency increases. After
the particles reach their maximum pos-
sible charge, a further increase in current
will not result in a consequent increase
in efficiency. Furthermore, W is related
to the intensity of the electrical field
Ep. The higher Ep, or applied voltage,
the more the particles will be attracted
towards the collecting surface. The limit-
ation is set by the breakdown voltage of
the gas, so that actually the precipitator
voltage should always be lower than the
breakdown voltage. A third controlling
factor is the dielectric constant of the
dust, represented as D. D-values for in-
sulating materials vary between 1.0 and
2.5, while for water the value is 2.9.
This is one of the reasons than an increase
in the dew point temperature normally
favors the electrostatic precipitation of
the suspended particles. The gas viscosity
y and the particle diameter r also influence
the value of the dielectric constant.
Values of W determined from commercial in-
stallations do not agree with theoretical
values obtained. The experimental values
are generally less than half the theoret-
ical values for W. The theoretical calcul-
ations do not consider the effects of re-
entrainment, which occurs when plates are
rapped or when there is uneven gas and
dust distribution.
For commercial design purposes, the value
of the migration velocity W must be deter-
mined experimentally or by using the data
of previous installations.
B. Specific Collecting Surface
The specific collecting surface is
a constant for a specific precipitator
operating with a certain constant gas
volume. The greater the specific col-
lecting surface, the higher the collect-
ing efficiency; f increases with in-
creasing collecting surface A and with
decreasing gas flow Q.
A number of variations to the basic
efficiency equation can be derived.
Consider a simple two plate precipi-
tator of length L, height H, spacing
between plates d, and gas velocity v.
The gas volume can be written as
Q - H d v (10)
Substituting
d v
-------
Electrostatic Precipitators Operation and Industrial Applications
Gives
,2 L W
d v
(11)
Figure 8.
SCHEMATIC SHOWING
DIMENSIONS.
If precipitatdr efficiency is to be
increased, this may be accomplished
by reducing spacing between plates
d; increasing precipitator length L;
reducing gas velocity v; or increas-
ing the drift velocity W.
Economic considerations generally
determine how these parameters are
optimized.
The gas velocity v can also be written
as L^
v = ~t
where t is the residence time of the
gas in the precipitator. This gives
From practical considerations, it is
desirable to make a precipitator as
small as possible consistent with
obtaining a satisfactory cleaning
performance. The physical dimensions
can be reduced only if the drift
velocity can be increased to more
closely approach theoretical values.
To accomplish this, one must increase
the average field strength, Ep, and
improve plate rapping, power, and gas
distribution.
C. Particle Size and Dust Concentration
A typical particle size analysis be-
fore and after an electrostatic pre-
cipitator shows a distinct difference,
with the dust after the precipitator
being much finer (see Figure 9).
Range of Particle Size
In Clean Gas
4-1
3
.
-rl
1-1
Gas
Particle Size, microns
Figure 9. PARTICLE SIZE DISTRIBUTION
IN DIRTY AND CLEAN GAS.
The particle size analysis and the
overall collecting efficiency enable
the grade efficiency curve of the
precipitator to be determined. From
this, the effective migration velocity
for a given particle size can be
calculated.
-(2 t W)
(12)
Obviously, the longer the gas is ex-
posed to the electrical forces,the
higher the efficiency. This is only
done, however, by decreasing the cap-
acity, since less gas will be handled
in a given precipitator.
E Average1
Fraction Size
Efficiency
W Average
Migration Velocity
JO
M
12
8
04-
Particle Size, microns
Figure 10. CURVES OF GRADE EFFICIENCY
AND MIGRATION VELOCITY.
-------
Electrostatic Precipitators Operation and Industrial Applications
This empirically determined relation
of the W-value to the particle dia-
meter is dependent on the dust content.
If the dust content is small, the
rated migration velocity has to be
reduced, since the probability of
the dust being charged, precipitated
and collected in the hopper is smaller.
On the other hand, when the dust
content is very high, the flow of
the current between the electrodes
is reduced, and so is the current
density, which also requires a low
migration velocity. This space charge
effect will increase both with the
concentration of the dust and also
with its degree of fineness.
D. Gas Velocity
Depending on the design of the col-
lecting surface, all dry-process
electrostatic precipitators have a
natural mechanical (inertial and
gravitational) collection efficiency,
which will increase with the gas
velocity to a maximum and then de-
crease due to higher re-entrainment
losses. Therefore, each collecting
plate has an optimum point for the
gas velocity, which is inherent with
its design and also with the pertinent
characteristics of the dust and the
carrier gas. The mechanical collecting
efficiency can be as high as 40 to
50 percent. The optimum gas velocity
can practically be used when the
specific application allows a high
migration velocity to be used.
E. Electrical Wind
It is well known that the electrical
wind is an attendant phenomenon with
corona discharge. The glow point on
the wire or on a sharp barb can be
regarded as a tiny nozzle in which
the gas molecules are accelerated to
a considerable velocity,owing to the
high field strength,so that a high-
pressure zone builds up in front of
it and a low-pressure zone behind it.
The hissing noise which accompanies
this discharge supports this concept.
The measured wind velocities attain-
ed considerable values. The curves
for the barbs in two directions and
in a single direction showed that
the latter arrangement caused a
considerably more powerful flow of
air in the direction to the plate.
-fe]
Figure 11. ELECTRICAL WIND VELOCITY
A. BARBED WIRE IN ONE
DIRECTION.
B. BARBED WIRE IN TWO
DIRECTIONS.
C. STAR-SHAPED WIRE.
Results obtained with dust-free gases
show that the wind caused by pointed
discharge electrodes had a velocity
near the collecting surface comparable
to the average gas velocities in in-
dustrial electrostatic precipitators.
Thus, an electrically charged particle,
which follows the force lines towards
the plate due to the eletrostatic
field, receives an additional impulse
in the same direction. The resultant
of the two directions of motion, per-
pendicular and parallel to the collect-
ing surface, is a preferential direction
of precipitation, which because of the
direction of the gas leads to non-sym-
metrical accumulations of dust on the
plate.
F. Dust Resistivity
Industrial precipitators normally
operate in the temperature range of
200 to 400°F, with less favorable
electrical conditions for precipitation
towards the higher end of this range.
With higher temperature, the break-
down voltage decreases, the viscosity
of the carrier gas increases, and in
the temperature range of 200 to 320°F
resistivity of most of the dusts in-
creases.
Specific dust resistivities range from
104 to 1014 ohm-cm. From 104 to lOll
no precipitation problems occur. Dust
resistivities higher than 10H ohm-cm
cause the accumulated dust on the
collecting surface to act as an in-
sulator.
-------
Electrostatic Precipitators Operation and Industrial Applications
Permanent Layer of Dust
tu
a
a
M-l
M
en
oo
•r-l
-fcJ
o
a
Dust Layers of
various densities
in
3
Q
Thickness
Figure 12. DUST DEPOSIT
The electrical charges of the dust
particles create a potential and thus
an intense electrical field in this
layer of dust, which at certain points,
for example at enclosed pores of gas,
forms an electrical ionization. A
corona discharge, similar to the one
at the discharge electrode.occurs.
This so-called secondary ionization
reduces the breakdown voltage con-
siderably, with a simultaneous increase
in current. The collection efficiency
Is also reduced considerably.
Practically all of the dusts have
certain semi-conductor characteristics,
i.e. an electrical resistance, varying
with temperature and moisture content
of the carrier gas. Normally in the
lower temperature range, the dust
resistivity is determined by the
.surface conductivity, which is a
function of the adsorbed or chemically
combined moisture.
10
.13
12
10
10" •
10
Dew Point
Gas Temperature
100
200 300
400
Figure 13.
EFFECT OF TEMPERATURE AND
MOISTURE ON RESISTIVITY
OF THE DUST
With increasing temperature the mois-
ture associated with the dust evapo-
rates, causing the dust resistivity
to increase. With a further increase
in temperature the conductivity of the
dust particles increases and this causes
the dust resistivity to drop again.
If the dry dust is cooled in a dry
atmosphere the dust resistivity will
increase over the previously experienc-
ed maximum (shown by the dashed lines
in Figure 13). This shows the influence
of moisture content on dust resisti-
vity.
The presence of 803 in the carrier gas
favors the electrostatic precipitation
process. For a number of dusts, the
resistivity is reduced below the critical
limit of 10-*- ohm-cm with very small
amounts of SO-j.
Collecting Efficiency
99% - Constant
180
160
140
120
100
0 2 4 6 8 10 12 M 16 re 20 22 24
S03, [PPM/VOL]
Figure 14. TEST OF COLLECTION
EFFICIENCY WITH
S03 INJECTION.
-------
Electrostatic Precipitators Operation and Industrial Applications
Tests at a pilot precipitator showed
that under constant conditions the
migration velocity doubled with an
303 injection of 22 ppm. This essential
improvement was due mainly to a re-
duction in dust resistivity caused by
the S03 injection. (See Figure 15).
No S0_ - Content
III SPECIAL DESIGN CONSIDERATIONS
An electrostatic precipitator, requires
the following basic components: (1) Dis-
charge system with high voltage energizing
set, (2) Collecting surface, (3) Rapping
system for discharge and collecting
electrodes and (4) Suitable casing with
means to receive the precipitated dust.
Roaster
Gas
SO Approximately 3mg/Nm
Gas Temperature °C
100
200 300 400 500
Figure 15.
EFFECT OF 803 ON
DUST RESISTIVITY
The presence of 803 is especially con-
ducive for the precipitation of dust
with oxide constituents, such as Si02,
MgO, and A1203, which react either
slowly or not at all with water to
form hydroxides
Another conditioning system for dust
particles is the injection of phos-
phorous pentoxide (P20^) or other
similar agents.
O
•H
O
01
G.
10
10
700
156 350
Gas Temperature
Figure 16.
DUST RESISTIVITY AND
BREAKDOWN VOLTAGE,
MEASURED IN THE WASTE
GAS OF A HOT-BLAST CUPOLA.
Figure 17. ELECTROSTATIC PRECIPITATOR
A. Discharge Electrodes
By selecting an electrode system
especially suited to match the re-
quirements of a specific application,
an improvement in collecting effici-
ency can be obtained.
With the composition of dust in
earlier days, discharge electrodes
with spikes pointed toward the col-
lecting electrodes offered no ad-
vantage. Spiked discharge electrodes
increased the collecting efficiency
in one installation, but in another
installation gave an increase in
the precipitator current with the
same or even decreased efficiency,
making its use doubtful.
Figure 18. DISCHARGE ELECTRODES
-------
Electrostatic Precipitators Operation and Industrial Applications
As a result of intensive research
work, however, it has been establish-
ed that with the barbs suitably
arranged and of the right length,
a corona discharge of high intensity
can be produced. This corona dis-
charge is capable of penetrating
the ionized dust space and to a
certain extent the deposited dust
layer itself. Thus, the two obstacles,
high electrical resistance and
space ionization, can be sufficient-
ly overcome.
\ \ FYt-clearing barbtd wire
t \ — Att*r-cl«3nng itar wwt
Figure 19. MAX ACR-OVER CURRENT
AS A FUNCTION OF PRECIPITATOR
INLET TEMPERATURE. NORMAL
GAS VOLUME.
For example in a two field precipi-
tator operating at nearly constant
conditions, i.e. constant gas volume
and raw gas dust content, both
fields drew a considerably higher
current after the first field was
equipped with spiked discharge
electrodes. The increased current
was very distinct in the temperature
range between 200 and 400°F. The
dust resistivity in this case had
a maximum of 5 x IflU ohm-cm at a
gas temperature of 230°F and was
well above lO^ ohm-cm between 170
and 380°F.
B. Collecting Electrodes
To achieve optimum efficiency the
collecting electrodes should meet
the following requirements:
1. An electrically smooth surface
giving a high breakdown voltage.
2. A good mechanical collecting
efficiency to prevent deposited
dust from reentrainment into the
gas flow during continuous op-
eration.
3. A high absorption capacity to
prevent reentrainment of dust
during or shortly after rapping.
4. Favorable vibration character-
istics to dislodge the accumu-
lated dust when the plate is
rapped.
Besides these basic requirements
for the operation of the precipi-
tator, collecting plates should
combine a high physical strength to
withstand the influences of the
environment, have enough rigidity
to stay aligned, and have a low
weight per area ratio for easy trans-
portation and handling.
Figure 20. COLLECTING PLATES
The development in the past ten
years included the step from the
box type electrode (Figure 20a) to
the folded electrode (Figure 20b)
and finally to the roll formed
electrode (Figure 20c).
Comparing the characteristics of
these electrodes, using the basic
requirements as criteria, it was
found that the breakdown voltage of
the roll formed electrode is sub-
stantially higher than for the box
type electrode. The folded electrode
with its long flat center part less
the highest breakdown voltage, but
it is much less effective in pre-
venting reentrainment and has less
adsorption capacity.
The box type electrode is only
slightly better in terms of reen-
trainment than the roll formed
electrodes, but this disadvantage
of the roll formed electrodes is
more than compensated for by the
better vibration characteristics.
-------
Electrostatic Precipitators Operation and Industrial Applications
Collecting plates should have good
dust retaining and absorption char-
acteristics. A picture representing
the gas flow in a single passage can
be obtained by using a water flow
model. (See Figure 21) The quiescent
zones formed by the bends of the
electrode outside of the main gas
flow serve to collect the major por-
tion of'the dust and shield it on
the way down into the hopper.
4,4m -
Figure 21. GAS FLOW USING A WATER FLOW MODEL.
The last requirement for a collect-
ing electrode is a favorable vibra-
tion characteristic. The impacts of
the rapping system, which consists
of a hammer arrangement on a rotat-
ing shaft at the lower end of the
collecting surface, have to be trans-
mitted over the total area. This
arrangement has the advantage in
that the dead weight of the plates
and the accumulated dust do not need
to be lifted by the rapper.
Figure 23. VERTICAL ACCELERATION
OF PLATE DURING RAPPING
Furthermore it leads to an accele-
ration of the plates vertically to
the surface, which is of utmost
importance for the dislodging of
the accoumulated dust. The amplitudes
of this acceleration, measured
vertically to the surface of the
collecting plates, should be at
least 100 g (one hundred times the
acceleration due to the gravity).
Otherwise substantial accumulations
of dust can occur.
Collecting efficiencies for differ-
ent types of collecting surfaces can
be compared by plotting the obtained
migration velocity versus the pro-
;duct of migration velocity and
specific collecting surface. For a
given value of wf, it can be noted
that the roll formed electrodes give
higher wf - values than other pocket
or flat plate electrodes.
10 50 80 90 9i 97 90 99
COLLECTING EFFICIENCY
Figure 22.
RAPPER SHAFT
WITH BEARING
Figure 24. COLLECTING EFFICIENCIES FOR
DIFFERENT TYPES OF COLLECTING
SURFACES.
10
-------
Electrostatic Precipitators Operation and Industrial Applications
C. Rapping Cycle Control
Collecting electrodes engineered to
give the maximum absorption capa-
city, (maximum dust collection in
pockets not subject to reentrainment)
account for approximately 60 per
cent of the dust. The remainder will
collect on the outside of the pockets
and can be reentrained under the
influence of the rapping of the plates.
From the moment the particle or ag-
glomerate falls away from the plate,
it is subject to the effect of grav-
itational force (settling velocity
Vf) and the velocity of the gas V.
The path followed by the dust part-
icle in the direction of the hopper
will thus correspond to the result-
ant R shown in Figure 25. If a line
is drawn parallel to R at the far
end of the precipitator outlet, one
can find which part of the dust
collected on the collecting plates
will drop into the hopper, and which
part will be carried out of the pre-
cipitator by the moving gas stream.
f, . AREA OF LOSS
H RESULTANT
V • GAS VELOCITY
V, - SETTLING VELOCITY
Figure 25. TWO-STAGE ELECTROSTATIC
PRECIPITATOR.
The area of loss can be expressed
as a loss factor to become part of
a revised W - value. It should be
pointed out that the area of loss
only affects that part of the dust
exposed to the gas stream and not
the part settling inside a hollow
plate or In the dead air space of
a profiled plate, in which case
the vector v is zero, since the
dust travels down a quiescent path
Into the hopper. The retaining cap-
acity of an electrode is thus a
factor of major importance.
Referring again to the loss area,
it is evident that the size of this
area depends on the slope of the
settling line R. At constant gas
velocity, as the settling velocity
increases, the size of the loss area
decreases and the collecting effi-
ciency is improved. The heavier
the agglomerates are, the higher
the probability that they will drop
into the hopper before being carried
out of the precipitator. With respect
to the efficiency of the precipitator
it would be favorable if the dust
could be dislodged in larger agglo-
merates. This then requires a
specific thickness of the dust layer
accumulated on the collecting plate
and also a rapping intensity adjusted
to this thickness.
The results of a rapping test using
different time intervals can be
shown by plotting the current of a
photo-cell located in the outlet of
the precipitator as percentage of
variation Against the rapping cycle.
Since the currents are inversely
proportional to the dust burden,
the lowest dust content and highest
collecting efficiency respectively
are obtained in the peak of this
curve.
4O IOO BO 2OO ZiQ \ 3OO
RAPPING |
Figure 26.
RAPPING CYCLE ["«*JT£jr
RAPPING CYCLE OF
ELECTROSTATIC PRECIPITATOR
FOR CLEANING WET BOTTOM
BOILER GASES.
D. Electrical System Control
Electrostatic precipitators operate
with high voltage direct current,
normally between 25 and 60 KV, and
in special cases up to 100KV. The
current usually is in the milliamp
range with exceptions up to one or
two amps.
-------
Electrostatic Precipitators Operation and Industrial Applications
Basically an energizing set for an
electrostatic precipitator has the
following components:
1. high voltage transformer
2. rectifier
3. regulating unit
4. control unit
High-voltage transformers do not
present any special problems as far
as their use for an electrostatic
precipitator is concerned.
Various types of rectifiers have
been used starting with rotary syn-
chronous mechanical rectifiers, which
were replaced by vacuum tube rect-
ifiers in the early 1950's. Almost
simultaneously suitable selenium
rectifiers were developed. Today
silicon rectifiers are used, which
normally are installed in the trans-
former tank, thus giving a compact
and reliable equipment.
Step-Up
Transformer
Rotary
Synchronous
Rectifier
Figure 27.
ROTARY SYNCHRONOUS
MECHANICAL RECTIFIERS.
When electrostatic precipitators
were first put into operation, it
became apparent that proper adjust-
ment of the electrical input was
required to obtain maximum perform-
ance. At first these adjustments
were made manually by personnel
assigned to the task.
In recent years automatic control
devices have been developed to main-
tain the desired operating point.
Since with increasing voltage the
corona discharge and the field in-
tensity increase, the precipitator
operation is highly a function of
the applied voltage. The voltage is
limited to the breakdown voltage
between the two electrodes and has
to be controlled to prevent exceed-
ing the breakdown level. This level
is not constant and varies with dust
concentration, dust composition, gas
composition, gas temperature, pressure,
mositure content, process variations,
etc.
There is no other practical way to
determine the breakdown level than
to reach this level In operation.
The control unit of the energizing
set therefore brings the operating
voltage to the breakdown level, then
reduces it to a lower level and
raises it again. The reduction of
the voltage has to be done as fast
as possible whereas the rise is at
a much slower rate. In this way the
voltage is controlled in a zigzig
line along the breakdown level.
o a)
-d oo
.*! n)
nj 4-1
<0 i-l
n o
M >
Precipator
Voltage
Time
Figure 28. AUTOMATIC VOLTAGE CONTROL
The rate at which the breakdown
level is reached greatly varies for
different .applications. The slope
of the rise of the voltage must
therefore be adjustable over a
broad range. In some installations
the breakdown level is reached more
than a hundred times in a minute,
whereas in others, it is only reached
once in several minutes. When a flash-
over occurs due to a voltage in
excess of the breakdown level, the
voltage in the precipitator drops
practically to zero and so does the
collecting efficiency. To reduce the
loss, the control has to extinguish
the flash as soon as possible and
bring the voltage in the precipita-
tor back up again.
In a single phase bridge current,
the arc-over extinguishes Itself
when the voltages pass through zero.
For this reason most of the pre-
cipitators are energized with single
phase circuits.
The regulating units for the trans-
12
-------
Electrostatic Preclpitators Operation and Industrial Applications
former rectifier began with a simple
resistor in series, followed by a
step-up transformer, an auto-trans-
former, a saturable core reactor
(transductor), and finally a thyr-
istor. Either auto-trans formers
or saturable core reactors combined
with an automatic control are con-
sidered standard equipment for an
electrostatic precipitator. The
only difference between the two is
that the auto transformer controls
the voltage, whereas the saturable
core reactor controls the current.
Although voltage control is generally
preferred, it is of minor importance
when operating close to the break-
down voltage. A current controlling
amplifier has the advantage of limit-
ing the current in case of a short
curcuit.
Step-Up
Transformer
AC
Supi
Saturable 8
Rector §
poooo
lv 1 I
0 0
1
' Rectifier
Precipitator
Control
Figure 29. PRECIPITATOR POWER CONTROL
BY USE OF SATURABLE REACTOR.
The block diagram shows a typical
precipitator control circuit.
Input power is applied through
suitable switching and protective
devices. These provide for compli-
ance with electrical code require-
ments and for safety and maintenance
considerations. Delivery of power
to the step-up transformer and
rectifier is modulated by a main
control element introduced in series
with the input to the transformer.
This element is usually a saturable
reactor, but electronic devices of
suitable rating such as thyristors
can also be used.
The action of the main control
element is determined by a signal
furnished by the amplifier section
of the equipment. The function of
this amplifier is to modify the
signals it receives by Increasing
their amplitude to make them suitable
for driving the main control element.
Direct manual control may be
accomplished by setting components
of the amplifier section or by
introduction of a manually-control-
led signal in place of the amplifier
section.
In an automatic system the signal
conditioning network compares the
Switching
13°wer and
Input Protective
Equipment
Direct
Manual
Control
Preset
Signals
— »-
— »•
Main
Control
Element
Step-Up Rectifier
Transformer (Tube or
j Semi- Con-
i ducted)
1
i I
! L
Amplifier
Signal
Condition-
ing
1
1
1
1
(Precipit
1
1
1
1
1
Feedback j
^ . J
Signals
ator
I
Figure 30. GENERALIZED BLOCK DIAGRAM - INDUSTRIAL PRECIPITATOR CONTROL.
13
-------
Electrostatic Precipitators Operation and Industrial Applications
signals from the actual operation
of the precipitator against preset
signals and the difference is fed
to the amplifier.
IV OPERATION OF AN ELECTROSTATIC PRECIPITATOR
The operation of an electrostatic precip-
itator, when properly designed and in-
stalled, will always follow some basic
mechanical and electrical laws. From the
available know-how it can be predicted
which conditions would favor the operation
of an electrostatic precipitator and how
these conditions can be achieved.
Once an electrostatic precipitator is
built, its design and physical dimensions
can only be changed with considerable
efforts and expenses. This leads to the
question of how the optimum operating con-
ditions can be established and maintained.
For a given installation this problem can
be split into four basic parts:
(1) Installing and operating a suitable
gas and dust conditioning system if re-
quired, (2) achieving and maintaining an
effective gas distribution ahead of the
electrostatic precipitator, (3) establish-
ing an optimum rapping cycle for the
collecting and discharge electrodes, and
(A) operating the precipitator with an
adequate electrical control system for
the energizing sets.
A. Gas Conditioning
Electrostatic precipitators for some
applications, such as the cement in-
dustry, steel mills, and some metallur-
gical furnaces, will only operate
satisfactorily If the gas temperature
at the precipitator inlet is reduced
to a specific level and the water dew
point raised to a required level.
L<5
20
GOOD'
100 no 120
POOR
130 140 150 160 °C
TEMPERATURE —
Figure 31. AREA OF GOOD OPERATION AS A
FUNCTION OF DEW POINT AND
TEMPERATURE.
This lowering of the gas temperature
and simultaneous raising of the dew
point can be effectively carried out
by installing an evaporation cooler
ahead of the dry process precipitator.
The evaporation cooler uses the princi-
ple of direct injection of a fine
spray water.
Different arrangements can be used,
but the most suitable design consists
of a standing cooling tower, using
the concurrent flow principle, with
the gas inlet and atomizers arranged
at the top of the tower.
1
J77J
Concurrent Flow
Countercurrent Flow
Figure 32. GAS CONDITIONING SYTEMS
The most suitable means of injecting
the water, in view of the large quan-
tities Involved, is by atomizers.
o
9-?
0)
O to
C 01
111 N
t-l i-l
t* to
P
O to
O 4J
O 01
• 0.
rH O
01 M
oi a
40.
30.
20-
10
\
I
1
1
I
\
V
N.^ Droplet Diameter »
01 U
> 01
•H *J
4J n]
<0 3
r-l
01 1W
40
30
20
10
100 200 300 400 500/«
100 200 300 400 500^
Figure 33.
SIZE DISTRIBUTION
OF WATER DROPLETS
14
-------
Electrostatic I'recipitators Operation and Industrial Applications
• -.o^ :MiS*0V?! * '
O~s» «W*
.<
> v- °
», o. «
arvST;,*' •
5go
« -v*'o.
The required treatment time for the
gas is given by the time it will take
for all the droplets to be evaporated.
The smaller the droplets, the faster
they will evaporate, and this will be
proportional to the square of the in-
itial droplet diameter.
A major aspect of the dimensional de-
sign of an evaporation cooler is the
determination of the overall height
or required volumetric capacity. The
volumetric capacity then corresponds
to a required treatment time since
each cooler of a given volume will
give a specific treatment time for a
constant gas volume, regardless of its
diameter.
^1000
t
^ 800
<
i 6°°
»--
5( 400
i 1(5') 1.1 K)1
„
a.6 a? as as
w 15 -
Figure 34. CALCULATION OF THE HEIGHT
OF AN EVAPORATION COOLER
In addition to the necessary uniform
gas and water droplet distribution
over the cross section of the cooling
tower, a control system has to be pro-
vided to ensure that a sufficient
amount of water is always sprayed in-
to the system and is always completely
evaporated. Inlet and outlet gas
temperatures at the evaporation cooler
as well as gas volume or heat content
of the gas can be used as parameters
for the control system.
B. Gas Distribution
Depending on gas and dust conditions
and the required collecting efficiency,
the gas velocities in an industrial
electrostatic precipitator are between
2.5 and 8.0 fps. A uniform gas dis-
tribution is of prime importance for
the precipitator operation, and it
should be achieved with a minimum
expenditure and pressure drop. This
is not always easy, since gas veloci-
ties in the duct ahead of the pre-
cipitator may be 30 to 100 fps in
order to prevent dust build-ups. Com-
paring the cross-section of the inlet
duct works and precipitator, ratios
up to 1:10 or 1:20 have to be coped with.
Suitable distribution models have to
be tested and the results related to
the full size unit considering the
laws of dynamic similarities.
The results of these tests can be
plotted on coordinates in the open
duct or precipitator area. These so-
called velocity profiles can now be
compared and the most suitable one
chosen for the final installation.
The results of these tests may be so
close together, that it is quite dif-
ficult to decide which arrangement to
use. For this purpose, a calculation
program can be used, which sets a
scale by establishing comparable local
collecting efficiencies for different
parts of the cross-sectional area of
the precipitator.
Measuring
Plane I
uannuajaaa.
luonnnnaa
iJiauntuuuL
tfr^-^
II
Measuring
Plane I
jUDHaaanm i
LantmaaaaJ
LDDUOJBJUULI
iJLHanjannnj
Figure 35. GAS DISTRIBUTION PROFILES
AFTER FLAP TYPE DEFLECTORS
15
-------
Electrostatic Precipitators Operation and Industrial Applications
V TYPICAL INDUSTRIAL PRECIPITATOR APPLICATIONS
A. Precipitators for Flue Gases From
Power Stations
One of the major applications for
electrostatic precipitators is de-
dusting of flue gases from coal fired
boilers in thermal power stations.
The tightening of regulations for the
abatement of air pollution is confront-
ing many power stations with the task
of installing high efficiency dust
collectors. The trend is towards using
low grade coal with a high ash content
and with the considerable increase in
the size of the boilers the problem
becomes even more significant. When
boilers were relatively small compared
to present standards, the dust emission
did not attain such high proportions.
Modern firing systems, in connection
with the variety of fuels used, con-
front the precipitator maufacturer
with new problems which require con-
tinuous extensive investigations.
The precipitator installation in Fig-
ure 36 is designed for a boiler with
a total capacity of 770 tons of steam
per hour. The gas volume from this dry
bottom, pressurized boiler is above
500,000 scftn. The collecting efficiency
of the two electrostatic precipitators
is 99.1 per cent with a guaranteed
residual of 0.044 gr/scf.
The two units in Figure 37 have a
turbine power of 250 MW each. Four
Babcock boilers with cyclone burners
and molten ash discharge generate
236,000 scfm into each one of the four
electrostatic precipitators. Tests
showed that the guaranteed collecting
efficiency was met with a clean gas
dust residual of 0.0326 gr/scf.
The introduction of boilers with wet
bottom ash discharge or with cyclone
burners necessitated dealing with very
small particles with up to 80 per cent
less than 10 microns. This dust has
also a number of other unpleasant
characteristics. It is difficult to
remove from the collecting surfaces
by rapping since it is very fine. Be-
cause of the high temperatures in the
fire box, the dust containes a high
percentage of vaporized (sublimated)
minerals which stick to the plates and
have a high electrical insulating
effect.
Figure 36. POWER PLANT
16
-------
Electrostatic Precipltators Operation and Industrial Applications
Figure 37. POWER PLANT
17
-------
Electrostatic Precipitators Operation and Industrial Applications
This very fine dust also induces a
space charge effect. Both the space
charge and the insulating layer of dust
on the collecting surfaces reduce the
current flow. Thus the desired collect-
ing efficiency can be attained only if
the size of the collecting surface is
correspondingly increased.
Due to the very steep slope in the peak
of the resistivity curve, changes of
20 to 40°F in gas temperature can affect
the operation of an electrostatic pre-
cipitator considerably. (See Figure 39).
5
o
f-
i/1
Co
LJ
CC
U~l
D
D
140 200 30O 4OO
GAS TEMPERATURE. °F
39. COM?AR[SON OF DUST
KI-.SISTIVITY IN CYCLONE
BOILER AND SF.AG TAG 1501LER.
Jn general, the size of an electrostatic
p rt'fi p 1 tator will largely depend on
1 ] ue gu.s volume, the required efficiency,
the firing system used, the composition
df thf coal (sulfur, moisture, volatile
matter, ash), flue gas temperature,
i-oke content of the dust, particle
size, electrical resistance of the dust,
;)nd the dust content of the flue gas.
Despite the possibility of future com-
plications, the requirements of the air
pollution codes can be met with a pro-
perly sized and operated electrostatic
precipitator. Flue gas precipitators
with efficiencies up to 99.9 per cent
and more are in operation.
li. 1'recipitntors for the Iron and Steel
industry
1 . Sintering Plants
Sintering has proved to be a highly
effective method of agglomerating
fine ores before they are charged
into the blast furnace. There are
two main dust producing sources in
a sintering plant: the sintering
belt and the sinter crushing and
conveying system.
Modern sinter belts have capacities
up to 10,000 tons per day and the
exhaust gas volume may be up to 1.2
million cfm. Dust contents range
from 0.22 to 1.3 gr/acf with 70
per cent of the dust having a part-
icle size of less than 20 microns.
Gas temperatures range from 210 to
320°F with dew points ranging from
90 and 120°F. The dew point influ-
ences the electrical resistivity
of the dust and thus the collecting
efficiency. In general low dew
points complicate the operation.
As an example of this application,
a precipitator for a sinter belt
with a capacity of 5,000 tons per
day is shown in Figure 40. The gas
volume is approximately 600,000 cfm
and the clean gas dust residual is
less than 0.01 gr/acf.
Precipitators for sinter crushing
and conveying systems have to deal
with dust inlet loadings up to 11
gr/acf. Electrical resistivity of
the dust increases sharply with the
air temperature and can reach
critical values. Residuals down to
0.01 gr/acf can be achieved.
2. Coke Oven Plants
The crude gas from coke oven plants
contains tar, which is removed by
condensing the gas and collecting
the droplets with an electrostatic
precipitator. Residual dust concen-
trations in the clean gas after the
precipitator can be as low as 0.0005
gr/scf.
3. Blast Furnace Plants
The application shown in Figure 42,
called the Venturion filter, is
unusual in the U.S.A. The installa-
tion is a combination of two diff-
erent dust collecting processes
giving an extremely high collecting
efficiency.
18
-------
Electrostatic Precipitators Operation and Industrial Applications
Figure 40. SINTERING PLANT
i BIBI^BBlMM
Figure 42. LONGITUDINAL SECTION
THROUGH PRESSURE TYPE
VENTURION PRECIPITATOR
19
-------
electrostatic Precipitators Operation and industrial AppHcatk
The raw gas enters the Venturi
washer and is greatly accelerated
at its narrowest cross section.
Water is injected simultaneously
at this point to obtain an intensive
mixing of gas and water. The dust
laden gas is washed to a large ex-
tent and cooled. After a 180° change
of direction the gas enters the
electrical fields of the electro-
static precipitator. The cleaned
gas is directed to the secondary
Venturi stages installed behind a
partition on both sides of the
electrical fields. The secondary
Venturis are primarily used to cool
the gas further. They also have a
slight scrubbing effect.
Before the cleaned gas leaves the
Venturi filter, its pressure is re-
duced to the desired pressure.
For example, in a pressurized
blast furnace installation in South
Africa designed for 225,000 scfm,
the clean gas residual is guaranteed
for less than 0.0044 gr/scf.
4. Steel Works
a. Open Hearth Furnaces
In this application the waste
gas from the open hearth fur-
nace is cooled by waste-heat
boilers and/or water sprays.
Dust loadings during an oper-
ating cycle vary widely, rang-
ing from as little as 0.1 g~f
cf to as high as several grains
per cubic foot during the
oxygen blow period. The mois-
ture content of open hearth
gas is about 7 to 8 per cent
by volume with a dew point in
the neighborhood of 100°F.
Because the extremely fine
dust (60 per cent less than 5
microns), care should be taken
to prevent re-entrainment.
The dust is readily precipitable,
but the moisture content must
be maintained at a certain
level.
b. Basic Oxygen Processes
It is general practice to burn
the CO gases with slight excess
air in waste-heat boilers. It
is immaterial whether a dry-
operated precipitator with an
evaporation cooler or a wet-
operated precipitator with a
Figure 43. BLAST FURNACE
20
-------
Electrostatic Precipitators Operation and Industrial Applications
saturator is installed. In
those instances where it ap-
pears to be of economic interest
to recover the dry dust (1.2
to 1.5 per cent of the pig
iron charge with an iron con-
tent of approximately 70 per
cent), and where the provision
of water is costly, it is log-
ical to use a dry operated
precipitator.
If, however, the plant should
already have a waste water
purification system, it is
advisable to install the smaller
wet operated electrostatic
precipitator.
Waste heat boilers entail con-
siderable capital outlay, and
the value of the steam generated
does not always warrant the
large investment.
C.
For this reason, alternative
solutions are sometimes favored:
radiant heat boilers and jack-
eted cooling stacks. Here too
it is immaterial whether the
subsequent electrostatic de-
dusting is done by the dry or
wpt process.
For example, a dry process
precipitator with an evaporation
cooler for the waste gases of
an LDAC converter will dedust
the gas below the visibility
limit of the so called brown
smoke. The capacity of the
converter is 80 tons and the
gas volume is between 88,000
and 95,000 scfm.
Precipitators for the Cement Industry
In the cement industry electrostatic
precipitators can be used to dedust
Figure 44. BASIC OXYCKN PROCESS
21
-------
Electrostatic Precipitators Operation and Industrial Applications
waste air or waste gases from raw mat-
erial drying, crushing and grinding
processes, rotary kilns, shaft kilns,
and cement mills.
Normally dry process plate-type pre-
cipitators are used, the casing made
either of steel or concrete. Corrosive
gases may require the use of aluminuir
alloy internal parts.
The d imcMisions of the precipitator de-
pend on the required collecting effi-
ciency, gas volume, and the process
used for the production of the clinker,
especially the gas temperature and
moisture content of the waste gases.
The operating temperatures of the
precipitator may range from 390 to
570°F. if the kilns are charged with
wet material and between 210 and 355°F
for the semi-wet process.
Figure 45. CEMENT PLANT
Figure5 46.
22
-------
Electrostatic Preclpitators Operation and Industrial Applications
If a dry process is used, the dust can
be collected either at temperatures up
to 660°F in relatively large precipi-
tators or at temperatures from 195 to
285°F in smaller precipitators by first
cooling and humidifying the gases. In
both cases the collected dust should
be completely dry.
Waste gases from dryers and grinding
mills at temperatures of about 210°F
can easily be dedusted with electro-
static precipitators.
0. Precipltator for the Chemcial and the
Non-Ferrous Metallurgical Industry
This application deals mostly with the
recovery of valuable metallic oxides
and salts from the waste gases from
metallurgical processes, e.g. shaft
furnaces, converters, stationary or
rotating reverberatory furnaces, re-
fining furnaces, electric furnaces,
roasters (multiple hearth, rotary and
fluidized bed furnaces), and sinter
machines for lead and zinc ores.
The gases from these applications are
normally cooled in waste heat boilers,
air coolers, evaporation coolers, or
Venturi scrubbers and then dedusted in
either pipe or plate type precipitators.
Normally a single stage cleaning system
is sufficient, but for gases with a
recoverable sulfur dioxide content
(e.g. roaster gases), multiple stage
cleaning systems are used.
Since most of these gases are highly
corrosive, precipitators operating
close to the acid dew points of the
gases have to be made out of corrosion
resistant materials, such as lead,
plastics or ceramics.
1. Electrostatic Precipitators for
Roaster Plants
Pyrite (FeS2) is fed into a turbu-
lent layer roaster. The exhaust
gas S02 containing is normally
cooled to 570 to 750°F in a waste
heat boiler and followed by a dry
process precipitator designed for
the elevated gas temperatures. If
the dust content is too high,
cyclones are used ahead of the pre-
cipitator.
Figure 47. COLLECTION OF SULFURIC ACID
MIST FROM ROASTER PLANTS.
23
-------
]ectrostatic Prccipltators Operation and Industrial Applications
Older incinerator installations, even
those using electrostatic precipitators,
received frequent complaints about burn-
ed or partially burned paper particles
which led to the addition of mechanical
collectors, such as arrestor plates after
the precipitator. Tests have, shown that
these particles are a prime subject for
reentrainment unless directly precipitat-
ed into a pocket or a dead air space of
the collecting surface. This required
the design of a special collecting plate
which eliminates this problem but must
be used with the proper gas velocity in
the precipitator. The gas velocity in
the electrical field has to be high
enough to cause an eddy current in the
pocket, but also low enough to prevent
re-entrainment from the flat parts of
the collecting surface.
The heat content of the gases exhausted
from the incinerator can be used in a
waste heat boiler to be converted into
steam either for heating or power genera--
ting purposes. The economics of such a
system will vary considerably and may
not be justified for smaller incinerators.
The flue gases leaving the incinerator
with temperatues up to 2200°F can be
cooled either directly or indirectly.
Indirect cooling using water or air re-
quires a substantial investment with
high maintenance costs.
1. Municipal Incinerators with Waste-
Heat Boilers
The refuse incinerator installation
includes eight electrostatic precip-
itators for the waste gases of four
large incinerators with waste-heat
boilers. Burning municipal refuse
with a lower heat content of 1500
kcal/kg resulted in a dust content
between 0.45 to 2.2 gr/scf. Gas
temperatures at the inlet of the
precipitator varied between 450 and
510"F with a water dew point between
115 and 145°F. When operating at de-
sign conditions, i.e. 440 tons of
refuse per day for each incinerator,
the gas volume of each precipitator
is approximately 141,500 cfm.
Tests showed clean gas dust residuals
between 0.0074 and 0.026 gr/scf with
collecting efficiencies between 98
and 99.5 per cent. These tests showed
that the operating conditions varied
considerably with variations in the
refuse analysis.
Figure 49. MUNICIPAL INCINERATOR
25
-------
Electrostatic. Precipitators Operation and Industrial Applications
Figure 50. INCINERATOR COAL-FIRED BOILER
2. Combination of Incinerator and Coal-
fired Boiler
A combined central heating and power
station is designed to burn either
refuse or bituminous coal, or a vari-
able mixture of both. Both firing
systems are located in separate cham-
bers and the flue gases pass a common
boiler and are subsequently dedusted
in an electrostatic preclpitator.
Test after commissioning showed com-
bined dust collecting efficiencies up
to 99.9 per cent.
3. Industrial Incinerator with Waste-
heat Boiler
The industrial incinerator installation,
which includes a waste heat boiler,
is mainly fired with paper, wood,
cardboard, and a substantial amount
of rubber, artificial leather and
rubbish.
Tested clean gas dust residuals aver-
aged 0.035 gr/scf or 0.0175 gr/acf.
The water dew point was approximately
104°F which is considerably lower
than for municipal refuse. The effi-
ciency of the electrostatic precipi-
tator is clearly related to the burn-
ing process. If for example carbon
black is formed, the efficiency will
decrease. It is important that enough
air is available to achieve complete
combustion.
Direct cooling can be done by simply
mixing false air with the gas but
unfortunately the gas volume increases
3.8 times when gases are cooled down
from 1800 to 570°F. All equipment
then has to be sized for this volume,
i.e. gas ducts, precipitator, fan and
stack.
Further it has to be considered that
In some air pollution codes the allow-
able stack discharge is specified for
either 12 per cent CO. or 50 per cent
excess air. Due to the dilution of
the air the C02 content in the gas
decreases, the actual allowable dust
residual decreases, too.
Direct cooling ean also be donfi by
injecting water into the gas stream
ahead of the precipitator. This
water injection will take place in
an evaporation cooler. The gas enters
the cooler with a temperature of 1800
to 2200°F and the water is sprayed
into it by high pressure atomizers.
26
-------
Electrostatic Precipitators Operation and Industrial Applications
The amount of heat required to eva-
porate the water will reduce the
gas temperature to approximately 570°F.
A cooling system using water sprays
to cool gases down from 1800°F to
570°F will increase the standard gas
volume by a factor of approximately
1.4 but, the actual gas volume will
decrease by a factor of approximately
0.6.
Since the installation costs for an
evaporation cooler are quite high,
sometimes in the range of the elect-
rostatic precipitator itself, part of
the incinerator can be used as a cool-
ing zone.
Figure 51.
INDUSTRIAL INCINERATOR WITH
WASTE HEAT BOILER
4. Industrial Incinerator without Heat
Recovery
The gas cooling system ahead of the
precipitator uses direct cooling by
water injection into an evaporation
cooler. Gas inlet temperatures are
between 1800 and 2200°F; the outlet
temperature is held constant at 570°F.
An automatic control system for the
water sprays is provided.
Figure 52. INDUSTRIE INCINERATOR
WITHOUT HEAT RECOVERY
The electrostatic precipitator has
an outlet gas dust residual of less
than 0.022 gr/scf.
Figure iJ. COOLING SYSTEM IN AN INCINERATOR
Due to the limited size of LUC in-
cinerator, cooling to 570°F cannot
be achieved within the incinerator,
but the gas can be cooled to approxi-
mately 900 to 1100°F by water sprays
installed at the end of the incinera-
tor and further cooled with air to
570°F. Thus the standard gas volume
entering the precipitator will be
approximately twice the gas volume
leaving the incinerator.
The moisture content of the gas at
the inlet of the precipitator is
approximately the same as at the outlet
of the incinerator.
VI Summary
The problem of air pollution is one which
grows with out modern civilization and,
generally, is a direct result of it. The
separation of suspended particles from
gases is one of the basic scientific and
technical problems of the industrial era.
27
-------
Electrostatic Precipitators Operation and Industrial Applications
Electrostatic precipitators, which can
be used for a wide range of applications,
can solve the problem of collecting even
the finest particles. There is no funda-
mental limit to the degree of cleaning
attainable, and, in practice, most pre-
cipitator installations operate in the
range of 90 to 99 per cent efficiency,
with some as high as 99.99 per cent.
The high collecting efficiency, the low
flow resistance, the ability to treat
huge gas quantities at high gas temperatures,
and the ability to cope sucessfully with
corrosive atmospheres and particles account
for the wide acceptance and diverse appli-
cations of the electrostatic precipitator
process.
28
-------
28
SECTION 28
High Temperature Gas Cleaning
-------
HIGH TEMPERATURE GAS CLEANING
I COOLING HOT GASES
A Fume collection problems ordinarily in-
volve handling hot gases (above 600°F.)
Thus, the problem of applying collection
equipment to fume-producing operations
is largely one of cooling the gases so as
to prevent heat damage to the collector.
B The following means of cooling are em-
ployed, either singly or in various
combinations:
1 Radiation cooling
2 Heat recovery equipment
3 Spray (evaporative) cooling
4 Admission of tempering air
5 Control of the manufacturing process
can at times be used to control gas
temperature
C Radiation Cooling
1 U-tube coolers
a Radiation cooling with conventional
U-tube condensers is effective for
high temperature gases when the
temperature gradient through the
tube walls is large. (Heat transfer
rates vary from 5BTU/hr/ft2/°F for
HR stainless steel in the range
AT = 1600 F to 0. 5 BTU/hr/ft2/°F
for carbon at lower values of AT.)
b Disadvantages
1) U-tube coolers require a relative-
ly large amount of ground space.
Considerable lengths of heavy
steel duct work from which heat
is radiated at a low rate are
necessary.
2) They lack flexibility with respect
to final temperature adjustment.
For example, inadequate cooling
may occur in hot weather and
overcooling in the winter time.
3) Usually a large capital investment
is required.
D Heat Recovery Equipment
1 Waste heat boilers
a In these devices the hot polluted
gases are passed over tubes through
which water is flowing. Hence, there
is a transfer of heat from the gases
to the water. The result is a cooling
of the gases and an economical use
of the heat they release.
b Although they require a large capital
investment, the financial return in
the form of steam or power usually
makes their use economically
feasible.
c Process operations which are inter-
mittent in nature cannot support
waste heat boilers.
2 Tubular heat exchangers
a These devices use cool outside air
which is forced around tubes carry-
ing the hot polluted gases.
b They are somewhat economical on
space requirements.
c They offer more flexibility in final
temperature adjustment than do U-
tubes or waste-heat boilers.
PA. C. pm. 26a. 1. 61
-------
High Temperature Gas Cleaning
3 Cleaning, .kefQre_.h.e_at. recovery
equipment
a Advantages
1) Fume deposition on the heat
recovery surfaces of the boiler
would be minimized. Thus there
would be improved heat transfer
and reduced maintenance problems.
2) Heat recovery equipment may be
eliminated if economical to do so.
b Disadvantages
1) Higher temperature resistant
materials would be necessary.
In many cases, the problem is
at present insurmountable.
2) The elevated temperature means
increased viscosity of gas and
consequently higher resistance.
3) Increased temperature also means
decreased gas density and con-
sequent reduction in blower
performance.
4) Large gas volumes are involved
at the elevated temperatures.
4 Cleaning after heat recovery equipment
a Advantages
1) Lower gas volumes are involved
since temperature is reduced.
2) Viscosity goes down, and density
goes up. Hence resistance is
reduced.
3) Heat recovery equipment may
act as an agglomerator of particles
which makes the particles easier
to collect by inertial or filtration
procedures.
4) Reduced particle concentration
may result because of collection
on the heat recovery equipment
surfaces.
5) Blower performance is improved.
6) Corrosion and temperature re-
sistant materials to tolerate the
prevailing conditions are readily
available.
b Disadvantages
1) Heat transfer is not good.
E Spray Cooling
1 Cooling involves the complete evapora-
tion of water spray by the hot polluted
gases.
2 With proper controls, initial gas
temperatures above 2000°F may be
reduced to 275°F by evaporative cooling.
At the same time, the dew point may
be maintained at less than 200°F.
3 The additional volume of gas, due to
the water vapor, is nominal.
4 Since a gas cleaning system installed
for air pollution abatement is often a
non-productive expenditure, the equip-
ment cost must be kept at a minimum.
For this reason, spray cooling of hot gases
is the most attractive of the available
methods of gas cooling since it is reason-
able in first cost, easily maintained, and
results in only & nominal increase in the
volume of gas to be cleaned.
F Admission of Tempering Air
1
Cooling gases already high in water
vapor content is frequently accomplished
by the admission of outdoor air through
bleed-in dampers.
2 Precise regulation of gas temperature
is readily accomplished.
3 Disadvantages
a One disadvantage lies in the high
resultant volume to be filtered, thus
requiring much larger collectors and
accessory equipment.
b In some cases where gas being
cleaned contains explosive gases,
the admission of oxygen may be ob-
jectionable for reasons of safety.
-------
High Temperature Gas Cleaning
G Comparison of Cooling Methods
] The following hypothetical problem
illustrates the difference in cooling
methods. (2)
a In all cases, it is assumed that
30, 000 cfm of air at 600°F contain-
ing 0. 01 Ib water vapor per Ib of
dry air is to be cooled to 275°F.
An 80°F ambient temperature is
assumed, tempering air is taken
at 80°F with 0. 01 Ib water vapor
per Ib of dry air. Water is avail-
able at 70°F. Extraneous heat
losses are not taken into account.
b Calculations
Wt. dry air (#/min)
\Vt, water vapor (#/min)
Wi. mixture (#/min)
j/'werfKje radiation loss (BTU/ft /hr/°F,)
KTU/miri to be removed
Kfg'd U-tube surface (fO
Length of 42" dia. pipe (ft)
1
Wt. 10 gauge U-tube, less hoppers (#)
Wt. tempering air (#/min)
**
Wt. cooling water (#/min)
Vol. tempering air at 275°F (cfm)
Vol. cooling water vapor at 275°F (cfm)
Voi . original mixture at 275 F (cfm)
Vol. to be cleaned at 275°F (cfm)
% clotii filter area required
Dt-w point
Cooling Method
U-tube
1096. 1
10.9
1107., 0
1.0
89, 443
15, 248
1,387
85,770
20, 629
20, 629
100
58°F
Temp. Air
1096. 1
10.9
1107.0
89,443
1,547.6
28,838
20,629
49,467
240
58°F
Water
1096.1
10.9
1107.0
89,443
90.30
2, 692
20, 629
23,322
113
125°F
-------
High Temperature Gas Cleaning
II
If spray cooling were used, from...
operating viewpoint, the filtration
temperature is a safe 155°F above
the dew point. With proper spray
cooler design, no difficulties would
be encountered.
HIGH TEMPERATURE GAS CLEANING
WITH CLOTH FILTERS*2)
a For example, when high tempera-
atures are encountered, DuPont's
Orion (R) acrylic fiber for temper-
ature not exceeding 275°F has
found application.
Where a high efficiency of removal
of solids from hot gas is necessary,
filtration through cloth may provide
a solution.
A Introduction
The enactment of stringent atmospheric
pollution codes has brought additional
interest in cloth bag filtration as a
means of reducing the mass rate of
emission of pollutants to permissible
levels.
The development of new and higher
temperature resistant fabrics has
increased the usefulness of cloth fabrics
in the high temperature gas cleaning
field.
a However, the success of such a
collector depends largely on the de-
gree of accuracy with which operat-
ing conditions can be determined,
or estimated, and the pains taken
in the design of the cloth filter and
accessory equipment in order to
provide continuous operation with a
minimum of maintenance.
B Examples of Hot Filtration
450° F
2000"F
Flue
Electrical
.. Precipitator
Furnace
Primary
Cooler
275' f
, Secondary
Cooler
. Cloth Filter
Figure 1
Carbon Black Filtration System
-------
High Temperature Gas Cleaning
Furnance
Rotary
Dryer
By-pass
Figure 2
Asbestos Rock Drying
Installation
Electric
Furriance
Tempering Air
(Dust Control Point)
I Figure 3
ilectric Casting Furnance
Installation
Fan
Canopy
lood
Water Jacket
Cooler
watef
4 water
^"~ ^X"
s
Tube Cooler
v
Air
°F
«
225'F
J
\AA/
s,
\
Figure 4
Reverberatory Brass Furnance
Installation
Furnance
Cloth Filter
-------
High Temperature Gas Cleaning
1950°F
n
Cupola
.jpolas
Water
Supply
t
450'F
Quencher
By-Pass
MVMVK/
Secondary Fan
Cooler
Cloth Filter Stack
Figure 5
Iron Foundry Cupola Installation
rt ft
Secondary
Cooling Air
ywv
) Cloth Filter
Screw
Conveyor
Tempering
Air
• Figure 6
Rock Wool Cupola Installation
Stack
Fan
-I JJ 1
' ' "
Spray Cooler
Furnances
Figure 7
Secondary Lead Smelting Installation
35O'F ,.PM$
v
Cloth Filter
-------
High Temperature Gas Cleaning
III HIGH TEMPERATURE GAS CLEANING
WITH ELECTROSTATIC PRECIPITATOR
A Introduction
the collection of fine particles which
may be less than 0. 001 microns and
not greater than 3 microns in many
cases.
In dealing with the control of stack
appearance, one is concerned with
The collection of fines in this range
may be efficiently done by electrostatic
precipitators.
Powdered
Coal
Raw Feed
Klin - 11' x 175'
Production = 1800 BBLS/day
K11n end temp. = 1800°F
Boiler exit temp. • 430°F
Cottrell = Concrete shell with
graded resistance electrodes
Cottrell efficiency = 86%
No dust returned
Clinker
Cooler
Waste
Heat
Boiler
Cottrell
Stack
Figure 8
Dry Process Cement Plant.
Kiln 11'x 360'
Production
Moisture in Gases to CMP = 10 to 12%
CMP inlet temp. 400° to 700°F.
Cottrell = Steel Shell - Metallic Electrodes
CMP Efficiency =95%
All Dust returned to system
Slag &
Lime rock
Cottrell
Powdered
.Coal
M/C
Clinker
Cooler
Fan
T
To
Raw
Mill
Rotary CMP
Dryer Unit
Stack
Figure 9
Dry Cement Process Using Slag
and Lime Rock
-------
High Temperature Gas Cleaning
Oyster
Shells
Fuel
Slurry
Storage
Rotary
Clinker
Cooler
fil
Kiln- 10' x 235'
Production - 1240 BBLS/day
Moisture in slurry 41%
Multiclone inlet temp. - 800 -900°F.
Cottrell Inlet temp. * 130°F.
CottreU » Redwood Shell & Pipes
CottreU .efficiency * 87.8%
800°-000°F.
M/C
Kiln
End
Housing
Cottrell
Preclpitator
Scrubber
130°F. Fan
Wood
Stack
Figure 10
Wet Cement Process Using Oyster Shells
DAM PIUS
• EGINEHAIOB
Figure 11
Flow Diagram for Open Hearth Gas
Cleaning Installation
-------
High Temperature Gas Cleaning
T)| cleaned gas
to boilers
Figure 12
Flow Diagram for Dry Cleaning Ferro-Manganese Blast Furnace Gas
AUTOMATIC STACK DAMPER
TRANSFORMER CABINET
HEAT EXCHANGER
CONDITIONER Two Point
SCHEMATIC DIAGRAM
of
TYPICAL INSTALLATION
on
HOT BLAST CUPOLA
\—/ MAIN BLOWER
Figure 13
-------
High Temperature Gas Cleaning
IV HIGH TEMPERATURE GAS CLEANING
WITH VENTURI SCRUBBERS
A Venturi scrubbers have been successfully
employed in high temperature gas cleaning.
Waste gases can be handled up to 1600°F
with high efficiency of particulate removal.
B Some applications are:
Gas Out~
44,000 SCFM (Dry Basis)
DAMPER
Gas at 125° - 165°F. Sat'd.
0.05 GR/SCF
S
Gas in--
44,000 SCFM
at 500° - 1600°F.
1 GR/SCF
P-A VtNTURI
_ _CYCJ.ONIC.
_^_ SEPARATOR
—»•
HIGH STATIC FAN
"! THICKENER.
RECIRC.
am
.at 40 PS]G
40~00~GPATM'A'K?i5p4WATER
TO
Figure 1A
Open Hearth Furnance
CONVERTER
lliMWLCflM.
•oMIJ'F...
RECKU UMK , •'- KCOVEtY,
Figure 1'J
Oxygen Steel Converter
10
-------
High Temperature Gas Cleaning
REFERENCES Making Processes. Air Repair 4,
No. 4, 189-196. February, 1955.
1 Pring, R. T. Bag Filtration of Aerosols.
Heating and Ventilating. December, 4 O'Mara, R. et al. Dust and Fume
1952. Problems in the Cement Industry. Air
Repair 4, No. 4. February, 1955.
2 Pring, R. T. Filtration of Hot Gases.
Air Repair. May, 1954. 5 Richardson, H. L. Scope of the Furnace
Fume Control Problem. Iron and
3 Silverman, L. Technical Aspects of High Steel Engineer. January, 1956.
Temperature Gas Cleaning for Steel
11
-------
29
SECTION 29
Sanitary Disposal of Collected Material
-------
SANITARY DISPOSAL OF
COLLECTED MATERIAL
In considering the choice of control equip-
ment, one must consider the safe and sanitary
disposal of the material collected by the con-
trol equipment. There are times when the
recovered pollutant is a valuable byproduct
and can be marketed profitably. But this is
not, always the case. Usually, the storage or
disposal of the material is a definite problem.
One that can add considerably to the cost of
the original control equipment.
In any waste disposal problem the overall
effect of the process on the environment must
be considered. If the solution of an air pol-
lution problem creates a problem in the
water pollution field, the problem is definitely
not solved. The reverse is also true. Take
for example the practice which exists in a
number of sewage treatment plants -- that of
disposing of the gases resulting from the
digestion of sewage sludge by burning of the
gases in a flare. Improperly designed flare
mechanisms have a tendency to blow out,
thereby allowing unburned hydrogen sulfide
to escape to the surrounding environment,
causing a definite odor nuisance. Suppose
the system is foolproof and all of the hydro-
gen sulfide is burned completely. The pro-
per combustion process is now an efficient
producer of sulfur dioxide, and as such
constitutes another potential air pollution
problem.
Years ago, one of the methods of treating
spent plating wostes, was the acidification
of a batch of spout wastes with sulfuric acid.
The cyanide fraction of the waste was con-
verted to hydrogen cyanide gas and vented to
thc> atmosphere. Being a very toxic gas, the
potential hazard to the atmospheric environ-
ment was very great. Disposal of plating
wastes to the waterways led to an equally
had problem leading to numerous fishkills.
Here, by correcting an air pollution problem,
a water pollution problem resulted.
There are several avenues of approach used in
the disposal of air pollutants removed from
emissions;
A Reuse of contaminant in original process.
B Conversion or use of waste product as a
saleable item; e.g., recovery of SO9 and
conversion into sulfuric acid; usage of
flyash as fine aggregate in building blocks.
C Storage of reclaimed product above
ground.
D Burial of contaminant in a sanitary landfill.
E Burial of contaminant at sea e. g. , encase-
ment of radioactive wastes in concrete
containers and dumping at sea.
F Discharge of materials to sewerage
system.
G Treatment of the waterborne wastes re-
sulting from control activities at the
plant prior to discharging to a sewerage
system.
It should be pointed out that although the con-
version of a waste product into a saleable item
could be considered as the most satisfactory
solution to the problem of disposal of the col-
lected waste, the demand for the product may
preclude any immediate sale, e.g., conversion
of sulfur dioxide into elemental sulfur. One
plant at present has a daily increase in its
storage area of ten tons of sulfur per day.
This company stored the sulfur above ground,
and was faced with the problem of curtailing the
emissions of sulfur dusts to the surrounding
environment due to wind errosion.
The disposal of collected dusts by sanitary
landfill is in certain cases the safest method
of disposal. A sketch of a landfill operation
is shown in Figure 1.
The landfill method in itself is relatively in-
expensive, assuming that land is available
for the operation, and assuming that the
collected material is such that can be compacted
and buried. In considering the economics of
PA.C.pm. 55a. 5.61
-------
Sanitary Disposal of Collected Material
Undisturbed
earth
Source: U. S. Government Printing Office.
"TB ENG 1 Sanitary Fill Method of Disposing
of Garbage and Refuse. " 1943.
Whether or not the effluent from a wet type
collecting device can be discharged directly
to a sewerage system or watercourse, de-
pends upon the effect of the waste on the
sewage treatment facilities or the watercourse
considered. There are numerous contami-
nants that can effectively interfere with
anaerobic and aerobic treatment processes,
even in low concentrations, e.g., heavy
metals (chromium, zinc, lead, copper etc.);
acids both organic and inorganic; and
alkalies. Certain compounds, can create a
severe taste and odor problem in the water
supply if discharged to a watershed, e.g.
aromatic compounds (phenols, naphthalene,
benzene, etc.) Pretreatment of the wastes
at the plant may be called for.
this method of disposal, the cost of trans-
porting the collected material from the
collector site to the disposal site must be
considered.
Conclusion: In any waste disposal problem,
the effect of the pollutant on the overall
environment, the air and land and the water,
must be considered.
-------
30
SECTION 30
Cost of Collection Equipment
-------
COST OF COLLECTION EQUIPMENT
I INTRODUCTION
There are two questions management asks
when they decide to act upon an air pollution
problem. These questions are: Whal is the
best way to solve this problem and what is
it going to cost? Cost is of vital importance,
for engineered control of a pollution problem
.seldom pays its way in recoverable product
or by-product. Rather, regardless of the
value of the community good will gained, the
decrease in potential damage toHhe plant
property, or achievement of better surround-
ings for employees, abatement costs increases
costs of production or service and, therefore,
decreases profit.
It is difficult to generalize about abatement
costs with meaningful values because most
pollution problems, even those in similar
industries, usually have attending circum-
stances that will vary the cost from one
specific application to another. Rather, what
follows, is a list of items one might consider
in attempt ing to estimate engineering costs in
air pollution abatement. Involved are capital
cosls, i. <:. , I he costs of estimating require-
ments, pur-chase:, site1 preparation and inslal-
Jalion and operating costs, i. e. , the costs of
ul.ililie.s and maintenance. Real property
required in pollution abatement consists of
(I) the abatement, device or .system itself, and
(2) those; accessories to the abatement device
necessary for its operation. Such accessories
may be divided into those functions associated
with: (J) movement of air, (2) movement of
liquid, (3) storage and disposal of separated
materials, (4) construction and support, and
(5) control instrumentation. These accessories
n ••:: very important and should be considered
i ••stimating both capital and operating costs.
II CAPITAL COSTS
A Kaclors that must be considered in estimat-
ing capital costs of a specific type of
equipment, include:
I Capacity of the abalemen! equipmem or
sy .<-: t e rn
2 Accessories to the abatement equipment
3 Installation costs
4 Special requirements or problems to
be solved
B Capacity
1 Inverse rule: Abatement equipment
prices per cfm handled will vary with
the magnitude of the cfm involved.
Usually the smaller the cfm the higher
cost per cfm.
2 Break point: In all designs there is a
point where the price per cfm levels
off, i.e., above this point, regardless
of the magnitude of the cfm the cost
per cfm remains the same. This "break
point", however, varies with the design.
C Accessories
Some sellers include all accessories, such
as blowers, motors, ductwork, etc. Other
sellers do not, and in this case, these
accessories must be obtained additionally
by the purchasers. To be considered are:
1 Air movement equipment
a fans and blowers
b electrical; motors, starters wire
conduit, switches, etc.
c hoods, duct works, gaskets, dampers,
etc.
2 Liquid movement equipment (in wet
abatement systems)
a pumps
b electrical: motors, starters, wire
conduit, switches, etc.
c piping and valves
d set tling tanks
PA.C.pm. 49. 10. 50
-------
Cost of Collection Equipment
3 Storage and disposal equipment
a dust storage hoppers
b sludge pits
c drag lines, track way, road way, etc.
4 Support construction
a structural steel work
b cement foundation, piers, etc.
c insulation (thermal)
d vibration and/or anti wear materials
e protective cover
5 Instrumentation; measurement and/or
control of:
a air and/or liquid flow
b temperature and/or pressure
c operation and capacity
d power
e opacity of flue gas (smoke meters,
etc.)
D Installation Cost
Installation costs vary considerably with
different types of equipment.
1 Method of shipment and labor required
at site. This varies according to how
the manufacture delivers the equipment
and/or accessories.
a completely assembled
b sub assemblies
c completely knocked down
2 Accessability: Is is necessary to:
a remove or relocate existing
equipment,
b provide ladders and servicing plat-
forms for maintenance?
3 Utilities: Do these have to be extended
or increased?
a electric power
b water
c sewerage: Will the liquid effluent
create a stream pollution problem?
d steam lines
E Special Requirements: Must the system
be designed for
1 Resistance to corrosion,
2 Resistance to abrasion or excessive
wear,
3 High or low process temperature,
4 Weather protection,
a wind
b flood
c temperature extremes
III OPERATING COSTS
A Utilities
1 Electric power
2 Water
3 Disposal
4 Steam
B Maintenance
1 Lubrication
2 Surface protection; cleaning and
painting
3 Replacement parts or structure
a equipment wear or failure: belt,
bags, filters, etc.
-------
Cost of Collection Equipment
b abrasion
c corrosion
IV COST COMPARISON
To provide a comparison of various abate-
ment systems Table I has been prepared.
Conditions under which device operates are
assumed similar (cfm, temperature, dust
loading, size analysis, etc.) unless other-
wise stated. The values indicated are not
intended to be used in a cost estimate for
some particular equipment but rather to pro-
vide an overall picture of comparison of
cost.
Table I -- Economic Comparison of Various Collection Systems (Reference 3)
3, 000 cfm, 68 F; 5 gr/cu. ft)
EQUIPMENT
Simple
cyclone
High eff.
cyclone
Irrigated
cyclone
Mull icyclone
Electrostatic
precjpitator
Irrigated
electrostatic
Conventional
fabric filte r
Reverse -jet
fabric filter
G ravitational
Spray tower
Wet impinge-
ment scrubber
Self- induced
Spray type
Venturi
•) -rubber
Lnsintegrator
Sc: rubber
>-,
o
c
V
o
£
w
65. 3
84.2
91.0
93, 8
94. 1
99. 0
99 9
99. 9
96 3
97. 9
93. 5
99. 7
98. 5
Capital Cost
($)
Total per
( 1) cfm
9, 240 14
17,640 .28
21,840 .36
19, 320 . 31
85,960 1.43
147, 840 2. 46
49,280 .81
47,600 78
51,240 84
28,840 48
24. 360 42
42,000 70
66, 640 1. 12
01
tc
3 C
S§-i
PH £ *
3. 7
4.9
3.9
4. 3
0.6
0. 6
4.0
5.0
1.4
6. 1
6. 1
22.0
--
s£
0 «»
0,
4, 732
6, 328
5, 634
5, 544
1, 736
3, 136
5. 264
11, 172
6, 650
8, 120
7, 896
29, 596
63, 560
r— t
0) ^
ta **--
QJ U
rt o
> 0
•* 0
T-l
--
4.0
--
...
2. 5
--
--
18. 0
3.0
0. 6
7. 0
5. 0
-&
o
o
. ^
«-£
a »
ES
--
--
1, 848
._
--
1, 232
--
--
9, 240
1, 540
308
3, 388
2, 380
Q)
g
? H
QJ ^"">
d •»
3
2
168
168
420
168
700
lr 120
8, 940
(2).
7. 560
fa)
840
840
560
840
560
i
v a
&8f.
13 w-C
s .s~
H15,4)
4, 900
6, 496
7, 952
5, 712
2, 436
5, 488
14, 168
18, 732
16, 730
10, 500
8, 764
33,824
66, 500
_ «
rt " t.
^j bo ^
-** ti ~?
g-rt^
°^5)
924
1, 764
2, 184
1, 932
8, 596
14, 784
4, 928
4, 700
5, 124
2, 884
2,436
4,200
6,664
Total
Cost
*/yr
5,824
8, 260
10, 136
7, 644
11,032
20, 272
19. 096
23,492
21, 854
13, 384
11, 200
38, 024
75, 104
C/1000cf
0.02
0. 029
0.034
0.027
0.038
0.070
0.066
0.082
0.075
0.047
0.038
0. 128
0.257
( 1) Includes accessories and erection
(2) Includes complete change of bags once a year
(3) Includes complete change of bags twice a year
(4) Assuming 8000 hr/yr operation
(5) 10% of capital cost
-------
Cost of Collection Equipment
REFERENCES PHS, U.S. Gov't Printing Office.
Washington, D. C. 1959.
1 Kane, J.M. Operation, Application, and
Effectiveness of Dust Collection Equip- 3 Stairmand, C.J. The Design and Fer-
ment. Heating and Ventilating. formance of Modern Gas-Cleaning
August, 1952. Equipment. Paper read before the
Institute of Chemical Engineers.
2 Lapple, C.E. Engineering Control of London. November, 1955.
Air Pollution. Proc. National Con-
ference on Air PoUution. USDHEW,
-------
31
SECTION 31
The Sylvan Chart
The class problems have been specifically
prepared for the use of students attending the
Control of Particulate Emissions course and
should not be taken out of context.
-------
SYLVAN CHART
The Sylvan Chart outlines many dust control
problems in terms of two important variables,
concentration and mean particle size. The
chart makes an ideal method of reporting
ranges of dust conditions encountered from
typical dust producing operations with the
number of operations reported limited only
by thr availability of the required field test
data.
The data provides a guide for collector per-
formance and furnishes an approximation of
collection efficiency and mass mean particle
size of effluent material. P-rediction of mass
m^an particle size of contaminant effluent is
based on the particle size distribution slope
for- a material being the same on the inlet
and discharge from collectors operating on
the impaction theory, i. e. , centrifugals,
most wet collectors, and fabric arresters.
For collectors in the electrostatic group,
deviation slopes differ and Sylvan Chart
cannot be used to approximate the mean
partii-le size of the effluent.
As dusts found in practice rarely consist of
particles of a given size, to define them one
must specify not only the size but also the
rohilivo amounts of each size. In order to
make a si/e and si'/,e distribution of a dusl,
it is necessary to make a particle size analy-
sis such as sieve, microscopic, elutriation,
;irnl sedimentation. For purposes of present-
ing, ' ompnring, analyzing, or extrapolating
pnrliele size aim lysis in dust control work,
logarithmic probability graph paper is norm-
ally used. On this paper, particle size is
plotted on a logarithmic scale against the
cumulative weight percent larger (or smaller)
than that size on an integrated probability
scale. This plot frequently gives straight
lines or lines of relative small curvature.
With this eurve, the distribution function can
be stated in terms of two parameters, the
median size and standard geometric deviation.
The latter is expressed as the 84. 13% size
divided by the 50% size. Using these parame-
ters in a mathematical function, the deviation
lines shown on the chart have been computed
for- industrial dusts.
Example: A suitable collector will be selected
for a lime kiln to illustrate the use of the
Sylvan Chart. Referring to the chart, the
concentration and mean particle size of the
material leaving the kiln can vary between
3 and 10 grains/cu. ft. with 5 to 10 microns
range of mass mean particle size. Assume
an inlet concentration of 7. 5 grains/cu. ft.
and inlet mean particle size of 9 microns.
Projection of this point vertically downward
to the collection efficiency portion of the
chart will indicate that a low resistance
cyclone will be less than 50% efficient; a
high efficiency centrifugal will be between
60 and 80% efficient and a wet collector,
fabric arrester and electrostatic preeipitator
will be 97% plus efficient. The latter three
collectors are often preceded by a precleaner
so a high efficiency centrifugal will be
selected. Using the average line of this group,
group, the efficiency will be 70%. Therefore,
the effluent from this collector will have a
concentration of 7. 5 (1. 00 .70) 2.25
grains/cu. ft. Draw a line through the initial
point witri a slope parallel to the lines
marked "industrial dusts. " Where deviation
is not known, the average of this group of
lines will normally be sufficiently accurate
to predict the mean particle size in the
collector effluent. Where this line intersects
the horizontal line marked 2. 25 grains/cu. ft. ,
a vertical line through the point will indicate
the effluent mean particle size of fi. 0 microns.
A projection of triis point of collector effluent
vertically downward shows that a second high
efficiency centrifugal will be less than 50%
efficient. A wet collector, fauric arrester
and electrostatic preeipitator will be not less
than 93% efficient. Selection of a good wet
collector will show an efficiency of 98%. The
effluent leaving this collector will have a
concentration of 2. 25 ( 1. 00 . 98) . 045
grains/cu. ft. Using the line initially drawn,
at the point where it intersects the line of
. 045 grains/cu. ft. will indicate a mean
particle size in the effluent of 1. 6 microns.
I'A. ('. pm. !). .r>. 57
-------
S) Ivan Chart
RANGE OF PARTICLE SIZES, CONCENTRATION, & COLLECTOR PERFORMANCE
COMPILED BY S. SYLVAN APRIL 1952: COPYRIGHT 1952 AMERICAN AIR FILTER CO. INC.
ACKNOWLEDGtMENTSOF PARTIAL SOURCES OF DATA REPORTED
FRANK W G AMF.RICAN AIR FILTER - SI2E AND CHARACTER ISTICS OF AIR BORNE SOLIDS 1931
I JRST AN'.i DRINKER ARCHIVES OF INDUSTRIAL HYGIENE AND OCCUPATIONAL MEDICINF APRIL 1952
-------
32
SECTION 32
Class Problems
-------
Problem 1. Volumetric Flow Rate
The volumetric flow rate from a process is a factor that affects
the efficiency of a collection device. A test yielded the following
results:
Flowing gas - dry air
Temperature (t) = 300°F
Static pressure (PB) 14 in. 1^0 vacuum
Atmospheric pressure (Patra) ~ 29.6 in. Hg
Average velocity head = 0.7 in. H,,0
Duct dimensions » 20 in. by 15 in.
Calculate the following:
a. Absolute temperature (T) in °R
Answer: 760
b. Absolute pressure (P , ) in in. Hg. Specific gravity of
mercury is 13.6.
Answer: 28.6
c. Has Density (p) in lbra/ft3. To solve the problem, use (1) the ideal gas
law and (2) the equivalence of 359 ft'/lbm - mole at 32°F and 14.7 psia
Molecular weight (M) of air - 29 lbm/lbm - mole and gas constant
(R) 21.83 (in. Hg * ft1) / (lbm mole * °R)
(1) [deal Gas Law
-------
(2) Equivalence
d. Volumetric flow rate (Q) in cfm. Density of 1^0 (PH 0^ " 62.4 Ib /ft
Answer: 8559
If the velocity doubles, what will be the velocity head?
e. Volumetric flow rate (Q) in cfm at 250°F and -18 in. HO
Answer: 8081
-------
Problem 2. Pressure Drop
The pressure drops for most cyclone installations are said to range from
2 to 7 in. w.g. How would you measure the pressure drop across a cyclone?
Is this static head loss or total head loss? What causes this loss? Is
static head loss approximately equal to total head loss? Explain:
-------
Problem 3. Reynold's Number
A particle 70 microns (\i) in diameter settles at a rate of 1 tps in air
at 300°F and 28.6 in. Hg absolute. Determine:
Gas Viscosity in ll^/ (ft * sec)
Answer: 1.55 * 10~5
b. Reynold's Number (NRe>. Refer to problem 1 for gas density.
1 ft » 30.48 * 10%
What is tho settling region of the particle?
-------
Optional Frob Jem. Reynold's Number
Determine the largest particle size that will settle in ajr at 70 ^
under Stokes's Law for dusts with specific gravities of 1, 2, and J
What Is the N equation?
b. Solve for Dp' In terms of p and fill in table below. Use Stoke's
extended region.
pair 0.018 cp at 70°F
ojU i..';?S lbra/(t1 ai 70°F and 1A. 7 psia
N
Re
0. 1
9-8-
1
37
2
29
3
25
-------
Problem 4. Molecular Weight of Gas Mixture and Review
An effuent gas has the following composition: a
Component
C02
CO
0?
N;.
H;,0
Vol. %
11
0.5 i.
6
76
6.5
M''
44
28
32
28
18
-*,«/-
f,40
(,4?-
M,
-------
c. Stoke's settling velocity in fps for D = 50u (s.g.=3).
1 ft.=30.48 * 10V P Answer;
What is the settling region for this particle?
Was Stoke's law assumption correct?
Answer: Yes
d. Velocity head in in. H20 if v- 80 fps
Answer: 0.94
n
-------
Problem 5.) Cumulative and Frequency Distribution Curves
Particle size analysis is an important consideration in evaluating
control equipment. The data can be represented by plotting a
frequency distribution curve and/or a cumulative distribution
curve or represented with the mean and standard deviation.
Particulate matter from a grinding operation is known to have
a log-normal distribution from previous analysis. Because of
this, only four size fractions were used for a particle size
analysis. For the data given below:
Table 1. Particle Size Analysis
,,-y
Particle Size
(microns)
0-10
10-20
20-40
+ 40
Total Weight in
Each Fraction
(%)
36.9
19.1
18.0
26.0
Cumulative
Z
Jo
f,;/
I a. Determine the geometric mean and geometric deviation. To
do this, construct a on
analysis (Use Figure 1).
graph paper for the particle size
Geometric Mean
Geometric Deviation •
b. Plot a. frequency distribution curve. To do this, arbitrarily
select size fractions desired. Then, using the
curve, determine the amount in each size
fraction (Use Table 2 and Figure 2).
& <2!L
-------
Figure 1 Cumulative distribution curve
c
o
i.
0.01 0.1 0.2 12 5 10 20 30 40 50 60 70 80 90 95 98 99 99.8 99.9 99.99
% by weight less than stated size
-------
Table 2. Frequency Distribution
Size
Fraction
M
0-1
1-2
2-3
3-4
4-6
6-8
3-10
10-12
12-14
14-16
16-18
18-20
20-30
30-40
40-50
>50
Lower
Size
0
2.5
24,2
31.0
36.9
41.8
46.0
50.0
53.0
56.0
67.0
74.0
78.8
% by Weight
Upper
Size
2.5
7.0
31.0
36.9
41.8
46.0
50.0
53.0
56.0
67.0
74.0
78.8
100.0
in Fraction
Difference
2.5
4.5
6.8
5.9
4.. 9
4.2
4.0
3.0
3.0
11.0
7.0
4.8
Per
Micron
2.5
4.5
3.4
2.95
2.45
2.1
2.0
1.5
1.5
1.1
0.7
0.5
Average
Size
P
0.5
1.5
7
9
11
13
15
17
19
25
35
45
/o
-------
6 iii
-rrt-rr
THi:
^
m
::~1
:n±
l- + -:f-
"ti -iii t-
nt.
,tiitH
ittn
"irlt
-tfftT
tir:
t Jii:
n;t
±nr
ft
.lU-nU-
Jl.
-I --•
MB
ffi
i
12
16 20 24
Particle diameter, microns
28
36
40
Figure 2 Frequency distribution curve
-------
Problem 6. Data Representation
If a log-normal or normal distribution curve exists, the mean
and standard deviation or geometric deviation are sufficient to
describe the data.
Given the following data for log-normal distribution, plot the
cumulative distribution curve on log-probability graph paper.
Source
Open Hearth
Fly Ash(b)
Cement Kiln*^
(d)
Gray-Iron Cupola
Fly-Ash, Cyclone
type Furnace
Mass Mean
Dia.
-------
100
80
60
50
40
30
20
10
8
-------
Problem 7. Frequency Distributions
a. Sketch the frequency distribution curve for the following:
50
Log-probability curve
c
o>
3
cr
01
t-
Log size
Frequency distribution curve
2.
50
Log-probability curve
u
c.
01
Log size
Frequency distribution curve
b. If you mix two different log-normal distributions with different
means, what general curve will result on log-probability paper?
Mixture of 1 and 2
50
so
Log-probability curves
-------
Problem 8: Settling Chambers
Two small heating plants, one using a traveling grate stoker
and the other a spreader stoker, desire to install a settling
chamber. The conditions of operations are listed in Table 1.
Table 1: Optional Data
travel ing
grate
spreader
Chamber Width (ft.)
Chamber Height (ft.)
Chamber Length (ft.)
Volumetric Flow Rate (scf/sec^
Flue Gas Temperature (°F)
Flue Gas Pressure (in. Hg gage)
Dust Concentration (gr/scf)
Mass Mean Diameter »(u)
Geometric Deviation a
Particle Specific Gravity^
10.8
2.46
15.0
70.6
446.0
0.0
0.23
72.0
1.95
2.65
10.8
2.46
15.0
70.6
446.0
0.0
1.22
57.0
4.06
2.65
1
Note: Standard conditions - 32°F and 29.92 in. Hg
Estimate the overall collection efficiency assuming actual
settling velocity • *5 theoretical Stoke' B settling velocity.
3 Particle size data from Anderson, D.H., et al.,
Pure Air for Pennsylvania, Penn. Dept. of
Health, November, 1961.
-------
(1) In order to determine the overall efficiency we need a size
efficiency curve. Assuming vy = 2 vy(s)> what is the size
efficiency equation?
Note that the chambers and operational conditions are identical
for both plants. Will the size efficiency curve be the same
for both? Yes No
(2) To obtain the size efficiency curve we need to solve the
equation in (1).
(a) What is the equation for v , ?
(b) Substitute (2a) into the size efficiency equation.
(c) Gather all constants into the factor K.
-------
(d) Do we know or can we calculate all the values in K?
If so, solve for K. Do not forget to correct Q
to operating conditions.
L •
B
g -
<"
u -
Q -
- P}
Since we wish to use p for D , place all conversion factors into
K. P
Answer: 1.136 * 10
-------
(e) By selecting arbitrary values for D . we can determine
E , or vice versa. What is the minimum particle size
in p
that can be collected at 100% efficiency.
Answer: 94
(f) Fill in the table below and check the size efficiency
curve in Figure 1. If it does not coincide, check your
caIculations.
8100
_P
100
92
-------
90
90
Size efficiency curve
for settling chamber
Particle size ( microns )
Figure 1
-------
»•- '. - .^r —-"-]"-T^^-'
o
300
200 t
o
s_
CJ
60 --—'-=± i ^ 1=
0.01 0.1 0.2 125 10 20 30 40 50 60 70 80 90 95 98 99
% less than indicated size
99.6 99.9 99.99
Figure 2 Cumulative Distribution curve
-------
(3) Besides the size efficiency curve we need to know the particle
size distribution of the inlet dust.
(a) Plot the size distribution curve for the inlet
dust in Figure 2.
traveling grate stoker
Y
V H c V V ™-
84.13 50 15.87 S
1.95X72p
140.3 y -36.9 y
spreader stoker
(b) Fill in the tables provided and calculate the overall
efficiency. Plot the outlet dust size distribution in
Figure 2.
b. Kscircate the amount and size distribution of material escaping
the .settling chamber. See(a3b).
-------
Table 1. Traveling Grate Stoker
Size
Fraction
p
0-20
20-30
30-40
40-50
50-60
60-70
70-80
80- 90
fr 44
Total
L "_"
InLet
%
2.7
6.9
9.4
10.5
10.5
9.5
7.0
9.5
34
100
gr/scf
0.0062
0.0159
0.0216
0.0242
0.0242
0.0218
0.0161
0.0218
0.0782
0.2300
'p
1.1
7.1
14.0
23.0
34.0
48.0
64.0
86.0
100
Hopper
Catch
gr/scf
0 . 0001
0.0011
0.0030
0,0056
0.0082
0.0105
0.0103
0.01.87
0.0782
0.135;
Outlet
gr/scf
0.0061
0.0148
0.0186
0.0186
0.0160
0.0113
0.0058
0.0031
0.0000
0.0943
%
6.5
15.7
19.7
19.7
17.1
12.0
6.2
3.3
0
100.1
Cum. %
6.5
22.2
41.9
61.6
78.6
90.6
96.8
100.1
Inlet -
inlet
CaLrli
Inlet'
.1357
.23
.59
-------
Table 2, Spreader Stoker
Size
Fraction
0-20
20-30
30-40
40-50
50-60
60-70
70-80
80-94
+ 94
TOTAL
Inlet
7.
23.0
9.5
4.5
3.5
4.5
36.0
100
gr/ecf
.2806
.1159
.0549
.0427
.0549
.4392
1.2200
E
P
%
1.1
7.1
48.0
64.0
86.0
100.0
Hopper
Catch
gr/acf
.0031
.0082
.0264
.0273
.0472
.4392
.6031
Outlet
gr/scf
.2775
.1077
.0285
.0154
.0077
0
.6169
%
45.0
17.5
4.6
2.5
1.2
0
%
45.0
62.5
96.3
98.8
100.0
-------
Problem 9. Settling Chamber
A settling chamber is 30 feet long, 6 feet high, and 12 feet wide.
The air flowrate at 500°F and 29.92 in. Hg is 4,320 acfm and the
entrained particles have a specific gravity of 2.5. What is the
diameter in microns of the smallest particle that can be removed
with 100%.efficiency if the effective settling velocity in the
chamber is h. the Stokes settling velocity?
-------
Problem 10. Cyclone
An efficiency teat on a cyclone produced the following data:
Efficiency - 94.7%
Specific gravity of dust - 2.75
Inlet dust concentration -1.0 grains/ft
Gas viscosity - 0.025 cp
Dust Analysis:8
Size Fraction
M
0-5
5-10
10-15
15-20
20-25
25-30
30-35
35-40
40-45
> 45
Hopper Catch
wt. %
0.5
1.4
1.9
2.1
2.1
2.0
2.0
2.0
2.0
84.0
Outlet
wt. TL
76.0
12.9
4.5
2.1
1.5
0.7
0.5
0.4
0.3
1.1
a C.A. Mau "Tho Elimination of Dust from Asphalt Plants"
Air Repair . 102-104 CNovember, 1953).
-------
005 001
OJDI
10 20 30 « 50 60 70 §0 M
% by weight less than indicated size
M 99
99.8 99.9 99.99
"•*- re 3 Pa[ le
-------
s. Draw the size-efficiency curve (Use Figure 2).
Size Fraction
u
0-5
5-10
10-15
15-20
20-25
25-30
30-35
35-40
40-45
> 45
TOTAL
Hopper
%
0.5
1.4
l,'l
3,>
1 '
tf >
2.0
2.0
2.0
2.0
84.0
100.0
gr/ft3
.0047
.0133
,01 SO
,d'^j1
0. / "1 1
.0189
.0189
.0189
.0189
.7955
.9469
Outlet
%
76.0
12.9
^
*>l
/^
0.7
0.5
0.4
0.3
1.1
100.0
gr/ft3
0.0403
0.0068
6.024-
6.VII
Ofirt
0003
0003
0002
0002
0006
0,OS3\
Inlet
gr/ft3
.0450
.0201
,,0304-
,02/3
'Ojo?
.0192
.0192
.0191
.0191
.7961
.9999
%
4.5
2.0
J,0
J,l
J-l
1.9
1.9
1.9
1.9
79.6
m
Eff .
%
10.4
66.2
#;;-
if ]
76,1
98.4
98.4
99.0
99.0
99.9
94.7
What is the cut size?
c ./ If a geometrically similar cyclone, one-half the diameter of the test cyclone,
/ is to be operated at three-quarters of the inlet velocity and a gas viscosity
of 0.020 cp, estimate the cut size. /
y
/ /
tf- f*rr/fr
X'
,
,•'_-
-------
d. Draw the size efficiency curve for the condition in (c). Use Figure 2.
Efficiency
20
40
50
60
70
90
Test Cyclone
D
P
Geometrically Similar Cyclone
D
P
-------
100
80
c
cu
o
40
20
TT
10 15 20 25
Particle size,/i
Figure 2 Size-efficiency curve
30
33
40
-------
Optional Problem: Cyclone
The following results were obtained from a cyclone servicing a hot-mix
paving plant.(1)
Vent
Dryer
3715 scfm
200 °F
743 Ib/hr dust
22,050 scfm
430 F
4,720 Ib/hr dust
Multiple cyclone
1525 Ib/hr dust
Table I. Collection Efficiency Data
Dust Particle
Size u
0-5
5-10
10-20
20-50
> 50
Size Analysis, wt . %
Inlet
6.2
9.4
13.8
22.9
47.7
Outlet
19.3
31.9
31.6
15.1
2.1
The data was plotted on a log-probability paper as shown in Figure 1.
Draw the size efficiency curve.
(I) Taken from Air Pollution Engineering Manual, PHS Publ. No. 999-AP-40, Cinn. (1967)
-------
100 99.99 99.9 99.g
Figure 1 Particle Size Distribution
99 98 95 «0 80 70 £0 50 40 30 20 10
2 I OJ 0.2 0.1 0.05 0.01
4;
o
0.01 0.05 0.1 0.2 0.5 1 2
10
20 SO 40 M SO 70
CO,
% Less Th*n
-------
Size Fraction
W
0 5
5 10
10 15
15 - 20
20 - 25
25 - 30
30 - 35
35 - 40
40 45
> 45
Total
Inlet
%
Ib/hr
Outlet
%
Ib/hr
Hopper
%
Ib/hr
Eff.
%
-------
H--
L.T_4-J—I—1.-I—i ' '—.--
Eff.
100
IT
-1-4-
80
60
•H-
h-
40
20
10
15
20
25
30
35
40
45 .
-------
Problem 11. Cyclone
A single cyclone handling 6,000 cfm has the following dimensions:
Diameter, D
Inlet width, W
Inlet height, 1^
Length of cone.l
6 cone
Height of cylinder, 1
Exit duct diameter, d
cyl
Carrier gas density, Ib /ft3
in
There is no entry vane
4 ft
1 ft
2 ft
8 ft
8 ft
2 ft
0.075
In an attempt to increase efficiency, a group of new cyclones is to be
designed with the same geometrical proportions and same pressure drop
as the single cyclone. Diameter of the small cyclone is to be 6 inches.
Determine the following:
a.. What will be the dimensions of the new cyclones?
Diameter, D
Inlet width, W
Inlet height, 1
Length of cone, 1
cone
Height of cylinder, 1 ^
Exit duct diameter, d
o
Dimensions, ft
Old
4
1
2
8
8
2
New
0.5
-------
b. How many of the small cyclones will be necessary to handle the original
flow rate at the same resistance?
.(1) What is the equation for the pressure drop across a cyclone?
(2) If the resistance must be the same for both cyclones, the ratio
of Ap to Ap must be equal to
old * new
(3) What then is the relationship between Ap and Ap ?
* old T new
(4) Solve for the number of new cyclones.
What Is the ratio of the inlet velocities for the old and new cyclones?
-------
Problem 12. Cyclone
A cyclone 8 inches in diameter carrying a. gas (essentially air)
at 340°F with an inlet velocity of 50 feet per second removes a
5 micron particle (sp.g. » 2.5) with 50% efficiency. If this
cyclone were operated at 170°F with an inlet velocity of 25 feet
per second, estimate the size of particle removed with 50%
efficiency. Assuming that the equation for efficiency as a function
of [Dp] cut and [Dp] mean (Eqn. 16 in the course manual) is valid
for this case, and given a mean particle size ([Dp] mean) of 30
microns, what are the values for the old and new efficiencies. „/•
- •/
/
37
-------
X
/<
-------
Problem 13. Electrostatic Precipitator
The following data represent results from the operation of a pilot-plant
electrostatic preclpitator on a 200,000 scfm power-plant effluent:
Dust Loading
Inlet = 3 grains/scf
Outlet => 0.50 grains/scf
Dust Analysis:
Size Fraction
(u)
0-10
10-20
20-44
+ 44
Inlet
(%)
30
20
12
38
Outlet
(%)
60
18
4
18
If the final precipitator is made twice as long as the pilot plant:
a. Calculate the percent removed in each size range and overall
percent efficiency for both units.
(l)Calculate the size and overall collection efficiencies for the
given size fraction for the pilot plant (Use Table 1).
Table 1. Pilot Plant Precipitator
Size
Fraction
P
0-10
10-20
20-44
> 44
TOTALS
Inlet
%
30
20
12
38
Iff
gr/scf
0.90
0,£0
0,3 (>
/,(!-
? /-i
<.*> ' '--•
Outlet
%
60
18
4
18
foo
gr/scf
0.30
0,010
0,MO
6f^o
0,£
Hopper
gr/scf
0.60
as/
0.34-
1,05
J.f.
Efficiency
%
66.7
8S>0
7/,^
-------
99.9
99.8
99.7
99.6
99.5
99.4
99.2
99
- 4-I--
n::
ft
-4
4
.J4r
-H+
l_.l
Ml!
•itH
J- ; : -i
t|T|
3E
f:
£
•STri
.r."
3^
t£
m
4P
14
=rM
J.
44-
¥i
T--E
W
mr
|-i4>4
rSEE
14
«
>
u
c
-------
(2) What is the efficiency equation for the precipitator
e = A * °'058 * FX * F2 * rD
Q
Noting the efficiency equation, how does doubling the precipitator
length affect the factor e ?
(3) Calculate the size and overall collection efficiencies for the given
size fraction for the final precipitator (Use Table 2). A calculation
aid is provided in Figure 1.
Table 2. Final Precipitator
Size
Fraction
V
0-10
10-20
20-44
-44
TOTALS
Conditions
Old
EP1
66.7
or O
-^C-1-/ LS
7A +
7<2>l
"i
1.10
/^
cZy'
C/,JJ
New
"2
2.20
J>,$
jf f:
.J. /
EP2
89.0
7jr^
Tlil
?7,*h-
Inlet
gr/scf
0.90
0,&d
QJ&
////-
J>,0
Hopper
gr/scf
.801
0,SB/
4,360
/,/?&
~3-&-$-
&i(^O
(4) Plot the size efficiency curve in Figure 2.
b. Estimate the efficiency for final precipitator based on overall
efficiency of pilot plant:
-------
-.-"f-
80
.-Hi
60
o
c
a*
•i—
o
20
10
15
20 25
Particle size,
30
35
40
45
Figure 2 Size efficiency curve
-------
Optional Problem. Electrostatic Precipltator
The following data represent results from the operations of a
pilot-plant electrostatic precipitator on a power-plant effluent:0
Dust Loading:
Inlet
Hopper
351 Ib /hr
210 lbm/hr
m
Particle Size Distribution
(microns)
0-10
10-20
20-44
+ 44
% In Fraction
Inlet Hopper
86
8
4.2
1.8
93.3
2.6
1.5
2.6
Calculate the following data, assuming the final precipitator is
made twice as long as the pilot plant:
a. Percent removed in each size range and overall - percent
efficiency for both units.
Table 1. Pilot Plant Precipitator
Size
Fraction
1.1
0-10
10-20
20-44
44
TOTALS
Inlet
%
Ib /hr
m
Hopper
%
Ib /hr
ra
Outlet
Ib /hr
m
Efficiency
%
a C.R. Flodin and H.H. Haaland, "Some Factors Affecting Fly-Ash Collector
Performance on Large Pulverized Fuel-fired Boilers," Air Repair 5,
27-32 {May, 1955).
-------
Table 2. Final Precipitator
Size
Fraction
u
0-10
10-20
20-44
> 44
TOTALS
Condition
Old
EP1
Bl
New
«2
V
Inlet
Ib /hr
m
Hopper
lbm/hr
%
Hopper
Actual
%
94.6
2.2
1.3
1.9
100
b. Estimate the efficiency for final precipitator based on overall
efficiency of pilot plant:
-------
Optional Problem. Electrostatic Preclpitator
If for the previous problem the particle size distribution were to
change can estimate of overall efficiency be made?
-------
Problem 14. Electrostatic Precipitator
An ele-etrostatlc precipitator operates at an overall efficiency of
95% when the gas temperature is 300°F. Estimate the new overall
efficiency when the gas temperature is 400°F, the precipitator
handles the same mass flowrate of gas, and all other operating
parameters remain constant.
-------
Problem 15. Electrostatic Precipitator
An industrial installation has two electrostatic precipitators each
designed to accomodate 72,000 cfm. Given:
Migration velocity: 0.4 feet/sec
Collecting surface: 14,400 sq. ft per precipitator
Calculate the collecting efficiency of each precipitator
(1) Write the efficiency equation. ~IZZZL-~- X
(2) Solve for E
Answer: 99.2
b. If one precipitator is shut down and the total gas volume is
treated in one precipitator, calculate the collecting efficiency
(migration velocity remains constant).
(1) At this point a graphical method will be illustrated. Note
that a plot of efficiency (E) vs. collecting surface per
unit gas volume (A/Q) on semi-log paper will yield a family
of straight lines, each slope depending on the migration
velocity (v ).
Mathematically
E = 1-e -A/Q «
P
or (1-E) = e - A/Q Vp
In (1-E) = - v A
P Q
Draw the line for vp =° 0.4 ft/sec in Figure 1.
-------
99.a
Eff.,
99.5
-Hr
-rr
Ft1,1
l-T
r
t! -
T:
99
98
tit
TT
Mr
95
TTtr
90
M i
ri
i i
Hi
80
50
I- t-
10 12 14
A/Q, sees/ft
16 18 20 22 24 26
-------
(2) Calculate A/Q and determine the efficiency from the graph.
Answer: 91.0
Increased production results in an increased gas volume of 84,000 cfm
per precipitator. New requirements call for a collecting efficiency
of 99.84% when both precipitators are on line, and a minimum efficiency
of 96.0% with all the gas going through one precipitator. Assuming
migration velocity to be constant at 0.4 ft/sec, for both conditions,
what size precipitator should be used to meet both requirements?
Answer: 22,400 ft2
-------
Optional Problem. Electrostatic Precipitator *
Given: A horizontal-flow, single stage electrical precipitator consisting of
two ducts formed by plates 8 ft wide by 12 ft high on 10 inch-centers,
handling 3,600 cfm with 2 grains fly ash (pulverized coal)/ft .
Estimate the drift velocity. Find the collection efficiency and the
outlet dust emissions in Ib/hr for
(a) Assuming uniform gas velocity.
Answer: 5.42 Ib/hr
(b) Assuming the velocity through one of the duct
is 50% greater than the average and 50% lower in
other
Answer: 9.27 Ib/hr
Table 1. Typical Drift Velocities
Application
Pulverized coal (fly ash)
Paper mills
Open-hearth furnace
Secondary blast furnace (80% foundry iron)
Gypsum
Hot phosphorous
Acid mist (H2SOA)
Acid mist (Ti02)
Flash roaster
Multiple-hearth roaster
Portland cement manufacturing (wet process)
Portland cement manufacturing (dry process)
Catalyst dust
Gray iron cupola (iron-coke ratio-10)
Drift
velocity
ft/sec
0.33 to 0
0.25
0.19
0.41
0.52 to 0
0.19 to 0
0. 19 to 0
0.25
0.26
0.33 to 0
0.19 to 0
0.25
0.10 to 0
(V
44
64
25
25
37
23
12
* Taken from Ai
r Pollution Engineering Manual, PHS Publ. No 999-AP-40, Cinn. (1967)
5"!
-------
Problem 16. Venturi Scrubber
A venturi scrubber has been operated in a phosphoric acid plant to
collect phosphoric acid mist. Operating data include the following
Throat Velocity, vt = 218 fps (t = 68°F)
Water Injection Rate, L = 1.44 ft3 / 1000 ft3 (t - 86°F)
Particulate Loading
Inlet = 8.50 gr PZ DS/ ft3
Outlet = 0.123 gr P 0 / ft3
Particle Size Data
Inlet: mean = 1.61, std. dev. =1.56
Outlet: mean = 0.85, std. dev. - 1.73
Mist Density, P = 116 lbm/ ft3
From the above data, the size-efficiency curve for the scrubber
can be shown to be as listed below:
Particle Dia. Efficiency
(H) %
0.2 31.5
0-4 64.0
0.6 88.3
0.8 96.0
1-0 98.2
a. Compare the theoretical size-efficiency curve (K = 1.52) with the
measured curve.
(1) Write down the applicable equation.
E - 1 - e - e
e KL ^
i(/ V D 2 p
P kp
18 Dw p
1.6 * IP1* + 28.5 L l'5
i, J.A., Contant, C.E., Ind. &Eng. Chem., v.50, n.8, 1958.
-------
Where: K • constant, determined experiment^?ly
L - liquid injection rate, ft3/1000 ft3
" - throat velocity, fps
D • water droplet diameter, y
" impaction efficiency, dimensionless
(2) Calculate the droplet size.
2
(3) Calculate \Jj as a function of D
Use D in y. P
-------
99.9
99.8
99.7
99.6
99.5
99.4
99.2
"
3
T1
-::--±.
-^
J~
l_
-r
o
c
o>
u
I T
1-1-
r -i-
T
-H-
1
. f;t+.
£E
Uib
m*
96
95
9-:
92
0
t:4-:
b(-
t
ftt
^-1
w~
f
-Hi
f'-
T1-
JO
H-t
--^-
ri ti
T-I---
til
-•-hn-
rr]U.
'TnT
iiiiiii
-l-i
4-r- -H-^-f
"trn;
r- -.
i
34
txponent, f.
-------
(4) Calculate B as a function of D in u .
P
(5) Using Figure 1 fill in the table below, and plot the data
in Figure 2.
DP (y)
0.2
0.4
0.6
0.8
1.0
e
E (%)
b. Compute a theoretical grade-efficiency curve for v = 554
(three times the original velocity)
(1) Calculate D at v - 654 fps.
w t
-------
LT
c
O)
o
O)
Q.
u
c
90
70
60
50
40
30
20
10
I !
--U-
T
0.5
1.0
1.5
Particle size ( microns )
Size efficiency curves for Venturi scrubber
Figure 2
-------
2. Calculate 1(1 as a function of D 2 where D is in microns.
P P
3. Calculate z as function of particle size D , where D is
in microns. P ^
4. Calculate the theoretical grade efficiency curve from the
relationship E - 1 - e - e.
D
P
(Microns)
0.2
0.4
0.6
0.8
1.0
z
E
%
-------
Problem 17. Venturi Scrubber
A venturi scrubber is to be installed on four processes.
All four applications will be using a venturi scrubber having
a throat cross section of 10" x 4.0* and a correlation
coefficient of 1.20. Using the following equations developed
by Ranz and Wong, and by Nukiyama and Tanawawa.
1 - e
ILL
7.38
v D
18
* .488 * 10~2
1 .5
Where: L - Liquid injection rate - gal./lOOO cf.
D • Average Particle Size - microns
D • Average Droplet Diameter - microns
w
v - Throat velocity of gas -ft/sec
v = Gas viscosity - centipoises
p = Particle density - lb /ft3
p m •
K «= Dimensionless correlation coefficient
Q - Flow Rate
a. Calculate the efficiency of impaction for each of the following:
Process
M.M.D. ,p
Q, cfm
L,gal/1000cf
P ,lb /ft3
p m
t, °F
1
2.00
50,000.00
8
120
68
2
1.00
50,000.00
8
120
68
3
1.00
70,710.00
8
120
68
4
1.00
50,000.00
16
120
68
-------
b. Explain why the efficiency decreases with decrease in average particle
si ze.
-------
What Is the effect of increasing velocity or liquid injection rate
on efficiency? Why?
loC
-------
Problem 18. Venturl Scrubber
In a venturi scrubber, the velocity at the throat is 328 ffeet per
second. The temperature of the carrier gas in 86°F. The density
Ibm
of the dust particles to be collected is 187 —- . The liquid
ftJ
injection rate is 2.0 ft3/1000 ft3 of air. What is the minimum
size particle in microns that can be removed with 98% efficiency?
K for the system is 1.52.
-------
Problem 19. Fabric Filter
Pilot-plant tests have been run with a bottom feed, single-section
(discontinuous operation) baghouse on the exhaust gases from an
electric-arc steel furnace. The fume concentration was measured
ajLjS^JLgr/jfJ^ • FUtEStifln—time (i.e., time between filter
cleaning) was maintained at 90 minuses. Volumetric flow thru the
baghouse was varied for a seTTe"s"™oF tests which produced the
'following pressure drop data.3
uf
fpm
2
3
4
5
AP (in. H20)
APr
1.5
1.9
2. it
3.3
APt
2.7
3.5
4.6
6.6
It is desired to install a multi-section (continuous operation)
baghouse consisting of six compartments. The time between
cleaning consecutive compartments (cycle time) is to be 30 minutes.
Calculate the maximum pressure drop that will be anticipated with all
compartments on stream, for each velocity level given in the
table.
W.W Campbell & R.W. Fullerton, " Development of an Electric-
"urnace Dust-Control System" JAPCA 12, 574-577, 590 (Dec., 1962).
-------
a. Write the applicable relationships.
S = S + miS
et r
Where: S •= Filter drag, in. H20/fpm
AP - Pressure drop, in.H20
Uf - Filter Velocity, fpm
j
C - Particulate concentration, gr/ft
6 * time, min,
K = Dust permeability
Subscripts:
t = terminal
r = residual
e = effective, when multiple compartments are used.
o
fvj
CD
-C
0
AP3 = Se3 Uf3
AP2 = se2 uf2
AP1 = Sel Ufl
"t. o
b c
6, minutes
-------
(el
b. From the experimental data done on a single section baghouse,
determine the dust permeability (K) for ^uch of the filter
velocities. Write the applicable equation below:
The figure below illustrates graphically what is
taking place for U, - 2 and 5. Study it, then complete the
table for Uf - 3 and 4:
200
W, gr/ft'
300
2.7
1.35
AP
1.5
r
0.75
AS
0,._60
150
1,
1,1$
6.6
1.32
3.3
0.66
rr
0^6-
3A1
;. Plot K vs Uf on Figure 1. Then explain
why K is not a constant.
-------
CL
O
J-
Ol
O)
200
100
-ff
T-|-rr
J±i
:.T:.:J-
345
. - Nominal face velocity ( fpm )
Figure 1 Permeability of dust cake as a function of face velocity
-------
d. Knowing K, write the equation for estimating Set for a six compartment
baghouse.
e. Calculate AS for the compartments in new baghouse for each Uj.
(tic
I'f
2
3
4
5
AS
S
r
0.63 AS
set
APt
f. "loL AP vs. U and compare with given data.
-------
Maximum pressure drop with all compartments on steam as a function of face velocity
o 6
-H~r
I I ! I
r
H-l-
T i
t±J
di
TT
------- |