EPA-450/3-81-005a
Control Techniques
for Particulate Emissions
from Stationary Sources -
Volume 1
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
September 1982
For sole by the Superintendent of Documents, 0.3. Government Printing Office, Woshtagton, D.C. 2M02
-------
This report has been reviewed by the Emission Standards and Engineering
Division of the Office of Air Quality Planning and Standards, EPA, and
approved for publication'; Mention of trade names or commercial products
is not intended to constitute endorsement or recommendation for use.
Copies of this report are for sale fay the Superintendent of Documents,
U.S. Government Printing Office, Washington, D.C. 20402, and the National
Technical Information Services, 5285 Port Royal Road, Springfield,
Virginia 22161.
-------
CONTENTS
VOLUME 1
Page
Contents of Volume 2 vii
List of Figures ix
List of Tables xix
Symbols ~ xxi
Conversion Factors xxiii
Glossary xxv
1. INTRODUCTION 1-1
2. BACKGROUND 2-1
2.1 Trends and Projections in Particulate Emissions 2-1
2.1.1 Air quality and particulate matter emission trends 2-1
2.1.2 Projections for future control programs and
emissions 2-2
2.2 Sources of Suspended Particulate Matter 2-3
2.2.1 Natural emission sources 2-4
2.2.2 Manmade sources of particulate 2-4
2.2.3 Transported particulate 2-9
2.3 Measurement of Particulate from Stationary Point Sources 2-11
2.3.1 Mass concentration measurement 2-11
2.3.2 Particle size analysis 2-16
2.3.3 Analysis of particulate samples 2-19
2.3.4 Chemical analysis and analysis of trace elements 2-20
3. ALTERNATIVE PARTICULATE CONTROL APPROACHES 3-1
3.1 Energy Source and Fuel Selection 3-1
3.2 Process Optimization 3-3
3.2.1 Modification of process feed materials 3-3
3.2.2 Elimination of process steps 3-4
3.2.3 Changes in process particle characteristics 3-5
3.3 Exhaust Gas Cleaning • 3-5
3.3.1 Applicable regulations 3-6
3.3.2 Source characteristics 3-6
' i i i
-------
3.3.3 Control device design limitations 3-7
3.3.4 Control device reliability 3-8
3.3.5 Control device costs and financial assistance 3-8
4. PARTICULATE CONTROL SYSTEMS 4.1-1
4.1 Introduction 4.1-1
4.1.1 Particle characteristics and behavior 4.1-1
4.1.2 Selection and application of particle
control devices 4.1-11
4.1.3 Control system design 4.1-11
4.2 Mechanical Collectors 4.2-1
4.2.1 Types of mechanical collectors 4.2-1
4.2.2 Operating principles of mechanical collectors 4.2-13
4.2.3 Design of mechanical collectors 4.2-25
4.2.4 Operation and maintenance of mechanical collectors 4.2-29
4.3 Electrostatic Precipitators 4.3-1
4.3.1 Types of electrostatic precipitators 4.3-1
4.3.2 Operating principles of electrostatic precipitators 4.3-8
4.3.3 Design of electrostatic precipitators 4.3-23
4.3.4 Operation and maintenance of electrostatic
precipitators 4.3-58
4.4 Fabric Filter 4.4-1
4.4.1 Types of fabric filters 4.4-1
4.4.2 Operating principles of fabric filters 4.4-9
4.4.3 Design of fabric filters 4.4-16
4.4.4 Operation and maintenance of fabric filters 4.4-36
4.5 Wet Scrubbers 4.5-1
4.5.1 Types of participate scrubbers 4.5-1
4.5.2 Operating principles of particulate scrubbers 4.5-11
4.5.3 Design of particulate scrubbers 4.5-35
4.5.4 Operation and maintenance of particulate scrubbers 4.5-44
4.6 Incinerators 4.6-1
4.6.1 Types of incinerators 4.6-1
4.6.2 Operating principles of incinerators 4.6-2
4.6.3 Design of incinerators 4.6-10
4.6.4 Operation and maintenance of incinerators 4.6-20
5. FUGITIVE EMISSION CONTROL 5-1
5.1 Sources of Fugitive Emissions 5-1
5.2 Control of Industrial Process Fugitive Emissions 5-2
iv
-------
5.2.1 Ventilation systems • 5-6
5.2.2 Optimization of equipment and operation 5-12
5.2.3 Wet suppression 5-13
5.3 Control of Fugitive Dust 5-16
5.3.1 Wet suppression 5-16
5.3.2 Stabilization 5-17
5.3.3 Specialized fugitive emission control techniques 5-19
6. ENERGY AND ENVIRONMENTAL CONSIDERATIONS 6-1
6.1 Energy Requirements 6-1
6.1.1 Fan energy requirements 6-2
6.1.2 Control device energy requirements 6-4
6.1.3 Hopper heaters and vibrators 6-13
6.1.4 Solids discharge and transport 6-13
6.1.5 Ultimate disposal 6-13
6.1.6 Other considerations 6-14
6.2 Secondary Pollutant Generation 6-14
6.2.1 Electrostatic precipitators 6-14
6.2.2 Incinerators 6-15
6.3 Liquid Waste Management 6-15
6.3.1 Regulatory requirements 6-15
6.3.2 Control techniques 6-16
6.4 Solid Waste Management 6-21
6.4.1 Regulatory requirements 6-22
6.4.2 Waste recycle 6-22
6.4.3 Waste disposal 6-24
6.5 Noise Management 6-25
6.6 Radiation Control 6-25
7. COSTS OF PARTICULATE CONTROL EQUIPMENT AND FUGITIVE EMISSION
CONTROL TECHNIQUES 7-1
7.1 Particulate Control Equipment Cost Analysis 7-2
7.1.1 Capital costs 7-2
7.1.2 Annualized costs 7-4
7.1.3 Other cost considerations 7-4
7.2 Methodology for Analyzing Cost of Particulate Control
Systems 7-5
-------
7,2.1 Capital costs 7-5
7.2.2 Annualized costs 7-8
7.3 Cost Curves for Various Particulate Control Systems 7-10
7.3.1 Equipment costs 7-11
7.3.2 Particulate control system costs 7-11
7.4 Cost of Fugitive Emission Control 7-28
8. EMERGING TECHNOLOGIES , 8-1
8.1 Advanced Scrubbing Techniques 8-1
8.1.1 Air Pollution Systems, Inc. electrostatic
precipitator 8-2
8.1.2 TRW charged droplet scrubber 8-5
8.1.3 University of Washington electrostatic droplet
scrubber 8-5
8.1.4 Steam hydro scrubber 8-11
8.1.5 Two-phase jet scrubber 8-11
8.1.6 Flux force/condensation scrubbing 8-15
8.2 Advanced Electrostatic Precipitation Techniques 8-15
8.2.1 Pulse energization 8-17
8.2.2 Two-stage ESP precharging 8-21
8.2.3 Flue gas conditioning for EPS's 8-23
8.2.4 Development of high temperature/high pressure
electrostatic precipitation 8-27
8.3 Advanced Filtration Techniques 8-28
8.3.1 Electrostatically augmented fabric filtration 8-29
8.3.2 Electrostatically augmented filtration through
fiber beds 8-29
8.3.3 Granular bed filtration 8-32
8.3.4 Barrier filtration 8-33
8.4 High-Gradient Magnetic Separation 8-34
8.5 Agglomeration Techniques 8-38
8.5.1 Sonic agglomeration 8-38
8.5.2 Magnetic agglomeration 8-43
-------
SUMMARY TABLE OF CONTENTS
VOLUME II
Section
9, SOURCES OF PARTICULATE EMISSIONS AND CONTROL TECHNIQUES , 9.1-1
9.1 Stationary Source Selection 9.1-1
9.2 Stationary Combustion Sources 9.2-1
9.3 Refuse Incinerators 9.3-1
9.4 Open Burning 9.4-1
9.5 Chemical Process Industry 9.5-1
9.6 Food and Agricultural Industry 9.6-1
9.7 Mineral Products 9.7-1
9.8 Metallurgical Industry 9.8-1
9.9 Petroleum Industry 9.9-1
9.10 Forest Products Industry 9.10-1
9.11 Lead-Acid Battery Manufacturing 9.11-1
9.12 Fugitive Dust Sources 9.12-1
vn
-------
-------
LIST OF FIGURES
Number Page
2-1 Estimated average contributions to nonurban TSP levels 2-5
2-2 EPA Method 5 particulate sampling apparatus 2-13
2-3 S-type pi tot tube and manometer 2-15
4.1-1 Aerosol distribution 4.1-3
4.1-2 Histogram of a lognormal size distribution 4.1-4
4.1-3 Cumulative lognormal size distribution 4.1-5
4.1-4 Bi-modal aerosol distribution 4.1-5
4.1-5 Impact!on of particles on a target in a moving gas
stream 4.1-6
4.1-6 Interception of a particle on a target in a moving
gas stream • 4.1-7
4.1-7 Diffusion of a particle to a target in a moving gas
stream 4.1-8
4.2-1 Howard multi-tray settling chamber 4.2-2
4.2-2a Simple momentum separator 4.2-3
4.2-2b Simple momentum separator . 4.2-4
4,2-2c Baffle-type momentum separator 4.2-4
4.2-3 Louvered shutter type collector 4.2-5
4.2-4 General types of cyclones 4.2-7
4.2-5 Typical simple cyclone 4.2-8
4.2-6 Flow pattern in a double vortex cyclone 4.2-9
4.2-7a Typical multi-cyclone collector 4.2-11
4.2-7b Individual tube from multi-cyclone collector 4.2-11
IX
-------
LIST OF FIGURES (continued)
Number Page
4.2-8a Fixed impeller straight-through cyclone 4.2-12
4.2-8b Bank of fixed impeller straight-through cyclones with
secondary cyclone dust collector 4.2-12
4.2-9 Typical size efficiency curve for settling chamber 4.2-15
4.2-10 Penetration of dust through a settling chamber serving
a sinter plant 4.2-16
4.2-11 Momentum separator 4.2-17
4.2-12 Penetration of fly ash through two momentum separators 4.2-17
4.2-13a Types of mechanically aided separators 4.2-18
4.2-13b Penetration curves for mechanically aided separators 4.2-19
4,2-14 Penetration curve predicted by Leith and Licht approach 4.2-21
4.2-15 Cyclone penetration as a function of particle size
ratio 4.2-22
4.2-16 Penetration curves for multicyclone tubes of different
diameter 4.2-23
4.2-17 Penetration curve for double vortex cyclone 4.2-24
4.2-18 Partial pluggage of multiple cyclone inlet vanes 4.2-32
4.3-1 Typical ESP with insulator compartments 4.3-2
4.3-2 Three types of wet ESP's 4.3-7
4.3-3 Basic processes involved in electrostatic precipitation 4.3-10
4.3-4 Typical temperature-resistivity relationship 4.3-15
4.3-5 Penetration as a function of particle size for an ESP
on a kraft pulp mill recovery boiler 4.3-16
4.3-6 Comparison of experimental penetration as a function
of particle diameter to the McDonald (1978) computer
model under normal SCA conditions 4.3-20
4.3-7 Penetration, pulverized-coal-fired boiler
(cold-side ESP) 4.3-21
-------
LIST OF FIGURES (continued)
Number Page
4.3-8 Computed versus actual penetration for cold-side
ESP on a western subbituminous-fired boiler2 4.3-22
4.3-9 Predicted precipitator penetration for bark/fossil
fuel-fired boilers 4.3-24
4.3-10a Precipitator penetration versus specific collection area
and precipitation rate w 4.3-26
4.3-lOb Precipitator penetration as a function of specific
collection area and modified precipitation rate
parameter w, 4.3-26
4.3-11 Distribution of ash-to-Btu ratio and log (resistivity)
for a single fuel field 4.3-27
4.3-12 Mechanical sectionalization of a precipitator 4.3-29
4.3-13 Typical fly ash precipitator voltage-current
characteristics, five fields in series, no ash
resistivity problem 4.3-32
4.3-14 ESP current wave form with and without si 1 iron controlled
rectifiers 4.3-34
4.3-15 Time periods are shown as control system reacts to a
spark impulse F after steady-state operation 4.3-35
4.3-16 Precipitator charging system and wire hanging system 4.3-38
4.3-17 Various combinations of electrical sectional ization in
an ESP 4.3-40
4.3-18 Vibrator and rapper assembly and precipitator high-
voltage frame 4.3-43
4.3-19 Typical fly ash type pneumatic vacuum system 4.3-48
4.3-20 Effect of two different methods of gas distribution
of flue characteristics in an ESP 4.3-50
4.3-21 Examples of two inlet plenum designs that generally
cause gas distribution problems 4.3-51
4.3-22 Expansion inlet plenums showing two methods of spreading
the gas patterns , 4.3-51
4.3-23 Internal view of one type of rectifier console showing
component parts 4.3-53
xi
-------
LIST OF FIGURES (continued)
Number
4.3-24 Flow diagram of sulfur burning flue gas conditioning
system 4.3-57
4.3-25 Comparison of mean penetration results 4.3-59
4.3-26 Precipitator log sheet 4.3-65
4.3-27 Typical operating curve to meet emission regulations
with partial malfunctions of ESP 4.3-76
4.4-1 Small shaker type baghouse 4.4-2
4.4-2 Reverse air baghouse 4.4-4
4.4-3 Continuous reverse air-cleaning system for flat
filter sleeves 4.4-5
4.4-4 Reverse air collector 4.4-6
4.4-5 Pulse jet baghouse 4.4-8
4.4-6 Initial mechanisms of fabric filtration 4.4-9
4.4-7 Baghouse performance, lead sinter machine 4.4-11
4.4-8 Baghouse performance, industrial boiler 4.4-11
4.4-9 Fabric filter penetration 4.4-12
4.4-10 Effect of air-to-cloth ratio on outlet concentration 4.4-13
4.4-11 Penetration correction term as a function of pressure
drop and
-------
LIST OF FIGURES (continued)
Number Page
4.4-18 Effect of gas temperature (continuous) on life of
glass fabric bags 4.4-26
4.4-19 Typical fabric weaves 4.4-29
4.4-20 Cross section of a thimble protecting bottom of bag 4.4-35
4.4-21 Dust penetration around snap-ring attachment 4.4-39
4.4-22 Gas jet adjacent to pin-hole 4.4-40
4.4-23 Abrasion damage at bag inlet 4.4-41
4.4-24 Bag failure location records 4.4-43
4.5-1 Spray tower scrubber 4.5-3
4.5-2 Vane type scrubber 4.5-5
4.5-3 Packed tower scrubber 4.5-6
4.5-4 Moving-bed scrubber 4.5-7
4.5-5 Tray scrubber 4.5-9
4.5-6 Venturi scrubber 4.5-11
4.5-7 Throat sections of variable throat venturi scrubbers 4.5-12
4.5-8 Theoretical single drop collection efficiency due to
diffusion and impaction 4.5-13
4.5-9 Theoretical penetration curves for various-sized
packed-bed scrubbers 4.5-17
4.5-10 Theoretical penetration curve for impingement plate
scrubber 4.5-19
4.5-11 Penetration curve for an impingement plate scrubber
on a rotary salt dryer 4.5-20
4.5-12 Theoretical penetration curve for a venturi scrubber
illustrating effect of throat velocity 4.5-23
4.5-13 Theoretical penetration curves for venturi scrubber
illustrating effect of liquid-to-gas ratios 4.5-24
4.5-14 Predicted venturi scrubber performance for f = 0.25 4.5-25
4.5-15 Comparison of Calvert's model results against measured
penetration data 4.5-26
xi i i
-------
LIST OF FIGURES (continued)
Number Page
4.5-16 Comparison of Yung, Calvert, and Barbarika model
against measured penetration data 4.5-28
4.5-17 Comparison of venturi scrubber outlet loadings to
static pressure drops for oil fired lime kilns 4.5-30
4.5-18 Correlation of coal-fired boiler scrubber outlet
dust loadings with theoretical power consumption 4.5-31
4.5-19 Comparison of predicted penetration as calculated in
Equation 4.5-19 and measured penetration 4.5-33
4.5-20 Liquid entrainment separators 4.5-39
4.6-1 Typical thermal incinerator- 4.6-3
4.6-2 Effect of air velocity and particle diameter on the
combustion rate of carbon 4.6-6
4.6-3 Fuel required to oxidize different concentrations
of combustible vapor 4,6-11
4.6-4 Tubular recuperator 4.6-13
4.6-5 Fixed-bed, pebble-stone, regenerative afterburner 4.6-13
4.6-6 Packed-bed flame arrestor 4.6-15
4.6-7 Corrugated metal flame arrestor with cone removed and
tube bank pulled partly off the body 4,6-15
4.6-8 Typical forced draft oil burner 4.6-17
4.6-9 Service temperature ranges for refractories 4.6-18
5-1 Hood design 5-8
5-2 Hood location 5-8
5-3 Air flow direction 5-8
5-4 Belt conveyor ventilation for fugitive emissions
control 5-9
5-5 Hopper and bin chute and conveyor-loading ventilation
for fugitive emissions control 5-10
5-6 Bag-filling fugitive emissions control 5-11
xiv
-------
LIST OF FIGURES (continued)
Number ' Page
5-7 Wet dust suppresion system applied to material handling
operation 5-15
6-1 Incremental energy requirements for fans 6-3
6-2 Energy required for transformer-rectifier set 6-7
6-3 Energy required for ESP insulator heaters and purge air
fans 6-9
6-4 Energy required for pumps 6-11
6-5 Energy required for stack gas reheat 6-12
6-6 Sedimentation tank or "clarifier" • 6-17
6-7 Vacuum filter 6-21
7-1 Cost of electrostatic precipitators; carbon steel
construction, thermally insulated, FOB factory 7-12
7-2 Cost of fabric filters, carbon steel construction, FOB
factory 7-13
7-3 Cost of mechanical collectors, carbon steel construction,
FOB factory 7-14
7-4 Cost of incinerators, FOB factory 7-15
7-5 Cost of venturi scrubbers, unlined throat, carbon steel
construction, FOB factory 7-16
7-6 Capital and annualized costs of fans and 30.5 m length
of duct 7-17
7-7 Capital and annualized costs of fan driver for various
head pressures 7-18
7-8 Capital and annualized costs of electrostatic precipi-
tators, carbon steel construction 7-19
7-9 Capital and annualized costs of fabric filters, carbon
steel construction 7-20
7-10 Capital and annualized costs of fabric filters, stainless
steel construction . 7-21
xv
-------
LIST OF FIGURES (continued)
Number Page
7-11 Capital and annualized costs of mechanical collectors,
carbon steel construction 7-22
7-12 Capital and annualized costs of incinerators 7-23
7-13 Capital and annualized costs of venturi scrubbers, carbon
steel construction 7-24
7-14 Capital and annualized costs of venturi scrubber,
stainless steel construction 7-25
8-1 APS electrostatic scrubber 8-3
8-2 Fraction efficiency performance of APS electrostatic
scrubber 8-4
8-3 TRW charged droplet scrubber 8-6
8-4 TRW charged droplet scrubber fractional efficiency
performance 8-7
8-5 University of Washington electrostatic droplet scrubber
schematic 8-9
8-6 University of Washington electrostatic droplet scrubber
fractional efficiency performance 8-10
8-7 Lone Star Steel steam-hydro air cleaning schematic 8-12
8-8 Lone Star Steel steam-hydro air cleaning fractional
efficiency performance 8-13
8-9 Aeronetics two-phase jet scrubber schematic 8-14
8-10 Aeronetics two-phase jet scrubber fractional efficiency
performance 8-16
8-11 Pulse energization voltage-current relationships for
various pulse frequencies 8-19
8-12 Comparison of DC and pulse energization voltage-current
relationships with same discharge electrode 8-19
8-13 Southern Research Institute precharger ESP assembly
drawing 8-22
8-14 Southern Research Institute precharger ESP fractional
efficiency performance 8-24
xvi
-------
LIST OF FIGURES (continued)
Number Page
8-15 Apitron electrostatic-filter cutaway view 8-30
8-16 Apitron electrostatic-filter fractional efficiency
performance 8-31
8-17 High gradient magnetic separator schematic
representation 8-36
8-18 High-gradient magnetic separator fractional efficiency
performance 8-37
xvi i
-------
-------
LIST OF TABLES
Number Page
2-1 Estimated Participate Emissions from Manmade Sources,
1977 2-7
2-2 Potential Industrial Sources of Fugitive Particulate
Emissions 2-10
3-1 Particulate Emission Reduction Potential of Various
Energy Sources 3-2
4-1 Particle Capture Mechanisms Normally Active in
Conventional Particulate Control Devices 4.1-12
4.2-1 Major Types of Mechanical Collectors 4.2-1
4.2-2 Effects of Operating Conditions on Cyclone Performance 4.2-23
4.3-1 Design Power Density 4.3-39
4.3-2 Reaction Mechanisms of Major Conditioning Agents 4.3-55
4.3-3 Example Effects of Changes in Normal Operation on ESP
Control Set Readings 4.3-75
4.4-1 Recommended Temperature Limits for Various Commercial
Fabrics 4.4-24
4.4-2 Chemical Resistance of Common Commercial Fabrics 4.4-27
4.5-1 Major Types of Wet Scrubbers 4.5-2
4.5-2 Typical Liquid-to-Gas Ratios for Wet Scrubbers 4.5-32
4.5-3 Typical Scrubber Pressure Drop 4.5-34
4.5-4 Properties of metals used as materials of construction
for wet scrubbers and auxiliary components 4.5-41
4.6-1 Auto-ignition Temperatures of Organic Compounds 4.6-5
4.6-2 ASTM Classification of Fire Clay Refractories 4.6-17
4.6-3 Commonly used Castable Fire Clay Refractories 4.6-17
xix
-------
LIST OF TABLES (continued)
Number Page
4.6-4 General Physical and Chemical Characteristics of
Classes of Refractory Brick 4.6-19
5-1 Industrial Process Fugitive Emission Sources and
Applicable Control Techniques 5-3
5-2 Fugitive Dust Emission Sources and Applicable Control
Techniques 5-4
6-1 Typical Static Pressure 6-2
7-1 Average Cost Factors for Estimating Capital Costs 7-6
7-2 Cost Adjustment Factors for Emission Control Systems 7-7
7-3 Example Factors for Annual!zed Costs 7-9
7-4 Typical Costs of Wet Suppression of Industrial Process
Fugitive Particulate Emissions .7-29
7-5 Cost Estimates for Wet Suppression of Fugitive Dust 7-31
7-6 Cost Estimates for Stabilization of Fugitive Dust 7-32
7-7 Cost Estimates for Sweeping and Flushing of Fugitive
Dust Sources 7-33
8-1 CDS Design Summary • 8-8
8-2 Comparison of High-Gradient Magnetic Separator and
Conventional Technology 8-38
8-3 Results of Industrial Tests with Sonic Agglomeration 8-40
xx
-------
SYMBOLS
Symbol Definition
a cross-sectional area
A collection plate area of an electrostatic precipitator
A wetted surface area
C Cunningham correction factor
c dust loading
d diameter
d , drop diameter
d, fiber diameter
d aerodynamic particle diameter
D gas diffusivity
D particle diffusivity
E charging-field strength
\-f
£ precipitation field strength
g acceleration of gravity
h height
H. liquid holdup
K inertia parameter
K t inertia parameter at throat velocity
pt
K2 resistance coefficient of dust cake
1 bed depth
n number of plates or stages
Npp Reynolds number
xxi
-------
Symbol Def1nj ti on
p total pressure
p static pressure
p corona power
P. penetration
q particle charge
Q, liquid throughput
Q gas throughput
*y
r, drop radius
r. radius of hole
n
T absolute temperature
t time
tR residence time
v gas velocity
y
v pickup velocity
of solid fraction in fiber bed
A static pressure change
q efficiency
(j gas viscosity
p. liquid viscosity
p gas density
y
p. liquid density
p particle density
pR bulk resistivity
Py in-situ resistivity
e porosity
uj migration velocity
xxi i
-------
CONVERSION FACTORS
1 meter (m) = 3.281 ft
1 meter (m) = 3.937 x 101. in.
1 meter2 (m2) = 1.076 x 101 ft2
1 meter3 (m3) = 1.308 yd3
1 meter3 (m3) = 3.532 x 101 ft3
1 meter/second (m/s) = 196.8 ft/min
1 meter/second (m/s) = 3.281 ft/s
1 meterVsecond (m3/s) = 2.119 x 103 ftVmin
1 meter3/second (m3/s) = 1.585 x 10s gal (U.S. liquid)/min
1 meterVsecond (m3/s) = 2.282 x 107 gal (U.S. liquid)/day
1 kilogram (kg) = 2.205 Ib
1 kilogram (kg) = 1.102 x 10"3 short tons (2000 Ib)
1 kilogram/meter3 (kg/m3) = 1.284 x IQ-2 lb/ft3
1 kilogram/meter3 (kg/m3) = 8.98 x 101 grains/ft3
1 joule (0) = 9.479 x 10'4 Btu (mean)
1 joule (J) = 2.778 x 10~7 kWh
1 watt (W) = 1.340 x 10-3 hp
1 pascal (Pa) = 1.45 x 10"* lbf/in.2 (psi)
1 pascal (Pa) = 4.019 x 10"3 in. H20
1 pascal second (Pa-s) = 0.672 Ib/ft2-s
1 kilopascal (kPa) = 4.019 in. H20
xxi 11
-------
-------
GLOSSARY
ABSORPTION.1 Transfer of molecules from the bulk of the gas to a liquid
surface followed by diffusion to the bulk of the liquid.
ADIABATIC SATURATION.1 A process by means of which an air or gas stream is
saturated with water vapor without adding or subtracting heat from the
system.
AERODYNAMIC DIAMETER. The diameter of a unit density sphere having the same
aerodynamic properties as an actual particle.
AEROSOL. A dispersion of solid or liquid particles of microscopic size.
AGGLOMERATION. The combination of smaller particles due to collisions.
AIR, DRY. Air containing no water vapor.
AIR-TO-CLOTH RATIO (A/C). The volumetric rate or capacity of a fabric
filter; the volume of air (gas) cubic meter per minute, per square
meter of filter medium (fabric).
ATOMIZATION. The reduction of liquid to a fine spray.
BACK CORONA. Localized electrical breakdown of a dust layer, producing
positive ions, which degrade or neutralize the intended charging
process.
BAROMETRIC SEAL.1 A column of liquid used to hydraulically seal a scrubber,
or any component thereof, from atmosphere or other part of the system.
BLAST GATE.2 A sliding plate installed in a supply or exhaust duct at right
angles to the duct for the purpose of regulating air flow.
BLINDING (BLINDED).2 The loading, or accumulation, of filter cake to the
point where capacity rate is diminished.
BURNER.1 A device for the introduction of fuel and air into a furnace at
the desired velocities, turbulence, and concentration to establish and
maintain proper ignition and combustion of the fuel.
CASCADE IMPACTOR. A particle-sizing device in which progressively increas-
ing inertia! forces are used to separate progressively smaller particle
sizes.
xxv
-------
CHEVRON MIST ELIMINATOR.1 Series of diagonal baffles installed in a gas
stream, designed to separate fine droplets of liquid from the gas by
means of inertia! impaction on the surfaces of the baffles.
COCURRENT. Flow of scrubbing liquid in the same direction as the gas
stream.
COLLECTION EFFICIENCY.1 The ratio of the weight of pollutant collected to
the total weight of pollutant entering the collector.
CONDENSATION.1 The physical process of converting a substance from the
gaseous phase to the liquid or solid phase via the removal of heat, the
application of pressure, or both,
CONTACT CHARGING. Charging of particles by contacting them and then re-
leasing them from a charged surface.
CORONA CURRENT. Measure of the current flow from the transformer to its
electrical section in an electrostatic precipitator (ESP).
COUNT.2 The number of warp yarns (ends) and filling yarns (picks) per inch.
Also called thread count.
CROSSFLOW. Flow of scrubbing liquid normal to the gas stream.
CROWFOOT SATIN.2 A 3/1 broken twill arranged 2 threads right, then 2
threads left. Also called 4 shaft satin, or broken crow weave.
CUNNINGHAM FACTOR. A correction factor to account for slippage of fine
particles moving through a discontinuous gaseous medium.
CURRENT DENSITY. Corona current level per unit area of collection surface
of an electrostatic precipitator (current per plate).
CYCLONE. A device in which the velocity of an inlet gas stream is trans-
formed into a confined vortex from which inertial forces tend to drive
particles to the wall.
DAMPER.2 An adjustable plate installed in a duct to regulate gas flow.
DEHUMIDIFY.* Reduction of water vapor content of a gas stream.
DEMISTER. A mechanical device used to remove entrained water droplets from
a scrubbed gas stream.
DENIER.2 The number, in grams, of a quantity of yarn, measuring 9000 meters
in length. Example: A 200-denier yarn measuring 9000 meters weighs
200 grams. A 200/80-yarn indicates a 200-denier yarn composed of 80
filaments. Usually used to describe continuous multifilament yarns of
silk, rayon, Orion, Dacron, Dynel, Nylon, and similar materials.
xxvi
-------
DENSITY.2 The ratio of the mass of a specimen of a substance to the volume
of the specimen. The mass of a unit volume of a substance.
DIELECTRIC STRENGTH. The maximum potential gradient that may exist In a
material without the occurrence of electrical breakdown.
DIFFUSION (AEROSOL). Random motion of particles caused by repeated colli-
sions of gas molecules.
DIFFUSION (LIQUID).1 The spontaneous intermingling of miscible fluids
placed in mutual contact, and accomplished without the aid of mechani-
cal mixing.
DIFFUSION CHARGING. Process of transferring electrical charge to particles
by random movement of electrons and ions; the effective charging mech-
anism for submicrometer-sized aerosols.
DIFFUSIOPHORESIS. Force acting on a particle, effecting movement due to a
vapor condensation gradient, resultant of differences in molecular
impacts on opposite sides of a particle.
DIMENSIONAL STABILITY.2 Capability of fabric to retain finished length and
width, under stress, in hot or moist atmosphere.
DRAFT.1 A gas flow resulting from the pressure difference between the
incinerator, or any component part, and the atmosphere, which moves the
products of combustion from the incinerator to the atmosphere. (1)
Natural draft: the negative pressure created by the difference in
density between the hot flue gases and the atmosphere. (2) Induced
draft: the negative pressure created by the vacuum action of a fan or
blower between the incinerator and the stack. (3) Forced draft: the
positive pressure created by the fan or blower, which supplies the
primary or secondary air.
DRAG FORCE. Resistance of a viscous medium due to relative motion of a
fluid and object.
DUST.2 Solid particles less than 100 micrometers created by the attrition
of larger particles.
DUST LOADING.2 The weight of solid particulate suspended in an airstream
(gas), usually expressed in terms of grains per cubic foot, grams per
cubic meter, or pounds per thousand pounds of gas.
DUST PERMEABILITY.2 The mass of dust (grains) per square foot of media
divided by the resistance (pressure drop) in inches water gauge per
unit of filtering velocity, feet per minute. Not to be compared with
cloth permeability.
ELECTROSTATIC FIELD. The position-dependent electrostatic force per unit
charge, made up of two components—one related to applied voltage and
electrode geometry, the other related to space change due to the pres-
ence of electrons, ions, and charged particles.
xxv ii
-------
ENTRAPMENT SEPARATOR (DEMISTER). That part of a gas scrubber designed to
remove entrained liquid droplets from a gas stream by centrifugal
action, by impingement on internal surfaces of the scrubber, or by a
bed of packing, mesh, or baffles at or near the scrubber gas outlet.
ENTRY LOSS.2 Loss in total pressure caused by air (gas) flowing into a duct
or hood.
EXCESS AIR.1 Air supplied for combustion in excess of that theoretically
required for complete combustion; usually expressed as percentage of
theoretical air (130% excess air).
FABRIC.2 A planar structure produced by interlacing yarns, fibers, or
filaments. (1) Knitted fabrics produced by interlooping strands of
yarns, etc. (2) Woven fabrics are produced by interlacing strands at
more or less right angles. (3) Bonded fabrics or a web of fibers held
together with a cementing medium which does not form a continuous sheet
of adhesive material. (4) Felted fabrics or structures built up by the
interlacing action of the fibers themselves without spinning, weaving,
or knitting.
FEEDSTOCK.1 Starting material used in a process. Can be raw material or an
intermediate product that will undergo additional processing.
FIELD CHARGING. Process of transferring electrical charge to particles
induced by high electric field strengths in the interelectrode space;
the effective charging mechanism for particles greater than 1 micro-
meter.
FIELD STRENGTH. A force field created by a large potential difference
between surfaces of different polarity; measured by the potential
difference divided by distance between surfaces.
FILAMENT.2 A continuous fiber.
FILL.2 Crosswise threads woven by loom.
FILL COUNT.2 Number of fill threads per inch of cloth.
FILTER MEDIUM. The substrate support for the filter cake; the fabric on
which the filter cake is built.
FILTER VELOCITY. The velocity at which the air (gas) passes through the
filter medium, or the velocity of approach to the-, medium. The filter
capacity rate.
FLY ASH. Finely divided particles of ash entrained in flue gases arising
from the combustion of fuel. The particles of ash may contain unburned
fuel and minerals.
xxviii
-------
FUGITIVE DUSTS. A type of participate emission made airborne by forces of
wind, man's activity, or both—such as construction sites, tilled land,
or windstorms.3
FUGITIVE EMISSIONS. Particles generated by industrial or other activities
which escape to the atmosphere not through primary exhaust systems but
through openings such as windows, vents and doors, ill-fitting oven
doors, or poorly-maintained equipment.3
FUGITIVE EMISSIONS. Industrial process such as emissions from a point,
area, or line source other than a stack, flue, or control system.
Emissions escape to the atmosphere from a defined industrial process
flow stream because of leakage, materials charging/ handling, inade-
quate operational control, lack of reasonably available control tech-
nology, transfer, or storage.
FUME.1 Fine solid particles predominately smaller than 1 micrometer in
diameter suspended in a gas. Usually formed from high-temperature
volatilization of metals or by chemical reaction.
GALVANIC SERIES.1 A list of metals arranged according to their relative
tendencies to corrode. When dissimilar metals are joined together in
an electrolytic solution, the one closest to the "active" end of the
galvanic series corrodes preferentially to the one closest to the
"passive" end.
GRAVITY, SPECIFIC.2 The ratio of the mass of a unit volume of a substance
to the mass of the same volume of a standard substance at a standard
temperature. Water is usually the standard substance. For gases, dry
air at the same temperature and pressure as the gas is often the stan-
dard substance.
GRID.1 A stationary support or retainer for a bed of packing in a packed
bed scrubber.
HEADER.1 A pipe used to supply and distribute liquid to downstream outlets.
HUMIDITY, ABSOLUTE.2 The weight of water vapor carried by a unit weight of
dry air or gas.
HUMIDITY, RELATIVE.2 The ratio of the absolute humidity in a gas to the
absolute humidity of a saturated gas at the same temperature,
HYDROPHILIC MATERIAL. Particulate matter that adsorbs moisture.
INERTIA. Momentum; tendency to remain in a fixed direction, proportional to
mass and velocity.
INTERCEPTION. A type of aerosol collection related to impaction, in which
an aerosol impacts the side of an obstacle because of reduced mobility
across streamlines.
xxix
-------
INTERELECTRODE SPACE. The space between the discharge electrode and the
collection plate; the active particle-charging region in an electro-
static precipitator.
INTERSTICES.2 The openings between the inter!acings of the warp and filling
yarns; the voids.
ION, GASEOUS. A gas molecule that loses or gains one or more electrons.
IONIZATION. Generation of free electrons that become attached to gas mole-
cules, forming ions.
ISOKINETIC SAMPLING. Matching the gas velocity at the sampling probe en-
trance to the gas velocity of the localized gas stream to collect a
representative particle size distribution.
LIQUOR.1 A solution of dissolved substance in a liquid (as opposed to a
slurry, in which the materials are insoluble),,
LOG-NORMAL DISTRIBUTION. A series of points that can be defined by a geo-
metric mean value and a geometric standard deviation.
MEAN FREE PATH. The average distance between successive collisions of gas
molecules; related to molecular size and number per unit volume.
MIGRATION VELOCITY. The average drift velocity of charged particles normal
to the direction of gas movement; also known as precipitation rate
parameter, a measure of the efficiency of collected particles to the
volume of gas treated and the area of the collection plate.
MOBILITY. A measure of response per unit force; the ease of motion relative
to the magnitude of the force-inducing motion.
MONOFILAMENT.2 A continuous fiber of sufficient size to serve as yarn in
normal textile operations.
MULLEN BURST. The pressure necessary to rupture a secured fabric specimen.
MULTIFILAMENT (MULTIFIL).2 A yarn bundle composed of a number of filaments.
NAPPING PROCESS.2 A process to raise fiber of filament ends (for better
coverage and more surface area), accomplished by passing the cloth over
a large revolving cage or drum of small power-driven rolls covered with
card clothing (similar to a wire brush).
NEEDLED FELT.2 A felt made by the placement of loose fiber in a systematic
alignment, with barbed needles moving up and down, pushing and pulling
the fibers to form an interlocking of adjacent fibers.
NONWOVEN FELT. A felt made by needling, matting of fibers, or compression
with a bonding agent for permanency.
xxx
-------
OPACITY. Measure of the fraction of light attenuated by suspended particu-
late.
PARTICLE. Small discrete mass of solid or liquid matter.
PARTICLE SIZE. An expression for the size of liquid or solid particle.
PARTICULATE MATTER. As related to control technology, any material except
uncombined water that exists as a solid or liquid in the atmosphere or
in a gas stream as measured by a standard (reference) method at speci-
fied conditions. The standard method of measurement and the specified
conditions should be implied in or included with the particulate matter
definition.
PARTICULATE MATTER, ARTIFACT. Particulate matter formed by one or more
chemical reactions within the sampling train.
PENETRATION. Fraction of suspended particulate that passes through a col-
lection device.
PERMEABILITY, FABRIC. The capability of air (gas) to pass through a fabric.
Measured on Frazier porosity meter or Gurley permeometer. Not to be
confused with dust permeability.
PENTHOUSE (ESP).1 Weatherproof gas-tight enclosure over the electrostatic
precipitator that contains the high-voltage insulators.
pH.l A measure of acidity-alkalinity of a solution; determined by calculat-
ing the negative logarithm of the hydrogen ion concentration.
PLAIN WEAVE.2 Each warp yarn passing alternately over each filling yarn.
The simplest weave, 1/1 construction; also called taffeta weave.
PLATE AREA. The effective area of both sides of the collecting surfaces in
an electrostatic precipitator.
POLYDISPERSITY. A particle size distribution consisting of different size
particles.
PRESSURE, STATIC. The pressure exerted in all directions by a fluid; mea-
sured in a direction normal to the direction of flow.
PRESSURE, TOTAL. The algebraic sum of the velocity pressure and the static
pressure.
PRESSURE, VELOCITY. The kinetic pressure in the direction of gas flow.
PRIMARY PARTICULATE MATTER. Particulate matter emitted directly into the
air from identifiable sources.
PRIMARY STANDARD. The national primary ambient air quality standard which
defines levels of air quality that are necessary to protect public
health.
xxx i
-------
PRIME COAT (PRIMER).1 A first coat of paint applied to inhibit corrosion or
to improve adherence of the next coat.
QUENCH.1 Cooling of hot gases by rapid evaporation of water.
RAPPER (ESP). Device for imparting acceleration of the collecting surface
to dislodge the deposited particles.
RAPPER INSULATOR. A device that electrically isolates a rapper from the
high-voltage system of an electrostatic precipitator, yet permits the
transmission of mechanical forces.
>
REFRACTORY.1 Ceramic material used for the lining of vessels, ducts, and
pipe for protection from heat, abrasion, or corrosion; also used for
insulation.
RESISTIVITY. The impedance offered to charge transfer across a dust layer;
defined by the ratio of electric field intensity to the current per
unit area passing through the dust layer.
REYNOLDS NUMBER, FLUID. A dimension!ess quantity in fluids to describe the
ratio of inertia! to viscous forces.
REYNOLDS NUMBER, PARTICLE. A dimensionless quantity in aerosol science to
describe the ratio of inertia! to viscous forces relative to the parti-
cle.
SATEEN.2 Cotton cloth made with a 4/1 satin weave, either as warp sateen or
filling sateen.
SATURATED GAS.1 A mixture of gas and vapor to which no additional vapor can
be added, at specified conditions. Partial pressure of vapor is equal
to vapor pressure of the liquid at the gas-vapor mixture temperature.
SATIN WEAVE.2 A fabric usually characterized by smoothness and luster.
Generally made warp face with a great many more ends than picks. The
surface consists almost entirely of warp (or filling) floats in con-
struction 4/1 to 7/1. The intersection points do not fall in regular
lines, but are shifted regularly or irregularly.
SECONDARY PARTICULATE MATTER. Particulate matter formed in the atmosphere
by physical and/or chemical gas-to-aerosol conversion mechanisms.
SECONDARY POLLUTANT. A pollutant not emitted into the air from a pollution
source, but formed in the air from the reactions of primary pollutants
(often photochemically).
SEEPAGE. The migration of particles through a freshly cleaned fabric.
SIZE DISTRIBUTION. Distribution of particles of different sizes within a
matrix of aerosols; numbers of particles of specified sizes or size
ranges, usually in micrometers.
xxxn
-------
SLURRY.1 A mixture of liquid and finely divided insoluble solid materials.
SMOKE. Small gasborne particles resulting from imcomplete combustion;
particles consist predominantly of carbon and other combustible mate-
rial; present in sufficient quantity to be observable independently of
other solids.
SNEAKAGE. Portion of a gas stream that bypasses the intended collection
area in an electrostatic precipitator.
SOOT. An agglomeration of carbon particles impregnated with "tar," formed
in the incomplete combustion of carbonaceous material.
SPECIFIC GRAVITY.1 The ratio between the density of a substance at a given
temperature and the density of water at 4°C.
SPRAY NOZZLE.1 A device used for the controlled introduction of scrubbing
liquid at predetermined rates, distribution patterns, pressures, and
droplet sizes.
SPUN FABRIC.2 Fabric woven from staple (spun) fiber; same as staple.
STAPLE FIBER.2 Manmade fibers cut to specific length (1% in., 2 ft, 2k in.,
etc.); natural fibers of a length characteristic of fiber, animal
fibers being the longest.
STOKES NUMBER. Descriptive of the particle collection potential of a speci-
fic system; the ratio of particle-stopping distance to the distance a
particle must travel to be captured.
STREAMLINE. The visualized path of a fluid in motion.
SUSPENDED PARTICULATE MATTER. Participate matter in the ambient atmosphere,
as determined by a specific reference method; material generally refer-
red to as total suspended particulate (TSP); consists of particles
within the size range of 100 to 0.1 micrometer in diameter.
TENSILE STRENGTH.2 The capability of yarn or fabric to resist breaking by
direct tension. Ultimate breaking strength.
TEMPERATURE, ABSOLUTE.2 Temperature expressed in degrees above absolute
zero.
TERMINAL SETTLING VELOCITY. The steady-state speed of a falling particle
after the equilibration of gravitation, drag, and buoyant forces has
occurred.
TRANSFORMER-RECTIFIER SETS. Electrical device used in electrostatic precip-
itators to rectify a.c. to d.c. and to transform low voltage to high
voltage.
xxxm
-------
THREAD COUNT. The number of ends and picks per inch of a woven cloth.
TURBULENT FLOW. A type of flow in which the fluid passes in a nearly ran-
dom, fluctuating motion.
TWILL WEAVE.2 Warp yarns floating over or under at least two consecutive
picks from lower to upper right, with the point of intersection moving
one yarn either outward and upward or downward on succeeding picks,
causing diagonal lines in the cloth.
VAPOR. The gaseous form of substances that are normally in the solid or
liquid state and whose states can be changed either by increasing the
pressure or by decreasing the temperature.
WARP.2 Lengthwise threads in loom or cloth.
WARP COUNT.2 Number of warp threads per inch of width.
WET/DRY LINE.1 The interface of hot, dry particulate-laden gas and cooling
or scrubbing liquid, at which an accumulation of solids can occur.
WOVEN FELT.2 Predominantly a woven woolen fabric heavily fulled or shrunk,
with the weave completely hidden by the entanglement of the woolen
fibers.
GLOSSARY
REFERENCES
1. Industrial Gas Cleaning Institute. Wet Scrubber Terminology. Publica-
tion WS-1, July 1975.
2. Industrial Gas Cleaning Institute. Fundamentals of Fabric Collectors
and Glossary of Terms. Publication F-2, August, 1972.
3. PEDCo Environmental, Inc. Technical Guidance for Control of Industrial
Process Fugitive Particulate Emission. EPA-450/3-77-010, March 1977.
xxxiv
-------
SECTION 1
INTRODUCTION
This document is a revision of "Control Techniques for Particulate Air
Pollutants,"1 which was published in January 1969. Changes and advances in
the technology of particulate control have made parts of the original docu-
ment obsolete. This second edition contains up-to-date information on the
emission reduction capabilities, costs, energy requirements, and environ-
mental impact of available control techniques, as required in Section 108(b)
of the Clean Air Act of 1977.
As in the first edition, the control techniques are based on informa-
tion from many technical fields. The methods and principles of operation of
many of the techniques have been known for years, but much experience has
been gained in their applications since 1969. The document also discusses
techniques that are still in various stages of research and development,
even though these new techniques are not yet available for general use.
Recent scientific data summarized in "Sulfur Oxides-Suspended Particu-
late Air Quality Criteria Document" have led to increased concern about
particles in the inhalable size range. This revision includes an expansion
of information on control effectiveness as a function of particle size.
Information on the conversion of gaseous pollutants to aerosols (secondary
particulate matter) has also been incorporated. The revised document re-
flects increased interest in fugitive particulate emission sources. Infor-
mation has been added on techniques to prevent and control these emissions.
The document is issued as two volumes. Volume 1 presents basic techni-
cal information on particulate emissions and control techniques; Volume 2
deals with control technology applied to major categories of pollu-
tant-emitting sources. The volumes are intended as general references for
technical personnel in regulatory agencies and in the private sector.
1-1
-------
Because the document is a general summary, it should not be used as the
basis for developing or enforcing control regulations.
Volume 1 has eight sections. Section 2 presents fundamentals of aero-
sol mechanics, trends in emissions and air quality, and sampling procedures.
The definition of particle size is of special importance because of the
differences among the definitions in the technical literature. Section 3
discusses the general ways in which particulate emissions can be minimized—
that is, by energy souYce and fuel selection, by process selection and
modification, and by exhaust gas cleaning.
Section 4 presents detailed information on exhaust gas cleaning tech-
niques. The operating principles, control effectiveness, and maintenance
requirements are summarized for each major category of techniques. Sec-
tion 5 discusses fugitive dust and industrial process fugitive emissions.
Most exhaust gas cleaning techniques concentrate particulate matter into a
liquid or solid waste stream. Accordingly, Section 6 presents information
on the environmental impact of these materials. This information is accom-
panied by an evaluation of energy requirements for particulate control tech-
niques.
Section 7 discusses the cost of particulate control. All the costs are
in first-quarter 1980 dollars, unless otherwise indicated. Section 8 pre-
sents the state of development of novel particulate control concepts.
Section 9 discusses the sources of specific particulate emissions and
the technology generally used to control emissions from novel sources.
Volume 2 consists entirely of Section 9, "Sources of Particulate Emissions
and Control Techniques." The page numbers in Volume 2, therefore, go from
page 1-2 (this page) to page 9.1-1, the first page of Section 9.
The data are given in metric units specified in the International
System of Units (SI). Conversion factors are listed in the front of the
report, as are important symbols.
1-2
-------
REFERENCES
1. U.S. Department of Health, Education, and Welfare. National Air Pollu-
tion Control Administration. Control Techniques for Particulate Air
Pollutants. AP-51, January 1969.
1-3
-------
-------
SECTION 2
BACKGROUND
This section provides background information relative to particulate
emissions and how they are measured. The generation of particulate matter
from stationary combustion sources, industrial processes, and fugitive emis-
sion sources is discussed, along with the secondary formation and transport
of particulate matter. Sampling and analytical methods of evaluating par-
ticulate matter are also described.
2.1 TRENDS AND PROJECTIONS IN PARTICULATE EMISSIONS
Particulate matter is generated by a variety of physical and chemical
mechanisms, and is emitted to the atmosphere from numerous sources, in-
cluding combustion, industrial process, fugitive emission, and natural
sources. Particulate matter is composed of finely dispersed liquids and
solids, including soot; dust; organic substances; inorganic substances, such
as sulfur compounds, metallic oxides, and salts; and other substances. The
chemical composition of particulate matter varies with source characteris-
tics, geographic area, and season of the year.
An estimated 12.4 teragrams (Tg) of particulate matter is emitted from
manmade sources in the United States each year.1'2 The major contributors
are stationary fuel combustion sources and industrial processes, which con-
tribute 39 and 42 percent, respectively. Transportation sources account for
9 percent of the total, and solid waste disposal and forest fires each
account for approximately 4 percent of the total.1'2
2.1.1 Air Quality and Particulate Matter Emission Trends
In 1971, the U.S. Environmental Protection Agency (EPA), promulgated
both primary (health-related) and secondary (we Ifare-related) National
Ambient Air Quality Standards (NAAQS) for particulate matter. Air assess-
ment in 1979 showed that 395 counties were classified as "nonattainment"
2-1
-------
(not having achieved the NAAQS) for particulate matter, the Nation's most
commonly monitored pollutant.
Ambient particulate levels at a monitoring site may be viewed as the
sum of particulate from traditional sources, nontraditional sources, natural
sources, and transported particulate. Each of these source types may at
times account for a significant portion of the measured particulate at a
particular monitoring site.3 In the 1970's, particulate emissions from
traditional sources were estimated to have decreased 46 percent nationally.
Dominant in this national trend were decreases in particulate emissions from
industrial processes and from fuel combustion.
2.1.2 Projections for Future Control Programs andEmissions
Traditional sources are still the dominant cause of nonattainment in
many urban areas with heavy industry; these sources may be contributing any-
where from 15 ug/m3 in residential areas to over 60 ug/m3 in industrial
neighborhoods.3 Nontraditional sources (e.g., reentrainment of road dust,
fugitive dust emissions from construction and demolition operations, and
dust from unpaved roads and parking lots) may contribute from 25 to 35 (jg/m3
to citywide particulate levels, and thus prevent some urban areas from at-
taining the standard.
Nonattainment of air quality standards is also caused by natural,
transported, and secondary particulates. Natural particulates are not
amenable to control, Nonurban sulfates and other secondary particulates,
transported primary particulates (up to 30 pg/m3 in densely populated metro-
politan areas), and urban secondary particulates (<10 ug/m3 formed from
gaseous emissions within the urban area) can be controlled by regional and
local planning and by control of gaseous pollutants.3
Annual variations in precipitation have affected the annual average
particulate levels by as much as 20 ug/m3. Although meteorology cannot be
controlled, the meteorological variations with time and location must be
considered in air quality analysis and planning.
New sources will be subject to New Source Performance Standards (NSPS)
as NSPS are developed for various industries. The Clean Air Act requires
that the EPA develop NSPS to prevent new air pollution problems by requiring
the installation of the best available technology considering cost, health
2-2
-------
and environmental impact, and energy requirements during initial construc-
tion. Designed to allow industrial growth without undermining air quality
management goals, NSPS are established at a national level so that control
requirements for new sources are uniform and consistent, regardless of
location.
The impact of NSPS on air quality will become more significant as new
plants are constructed and as older sources are either modified or replaced.
Before amendments to the Act were promulgated in 1977, a study was under-
taken to estimate the potential emissions impact of the NSPS program.
Particulate emissions from stationary sources in 1990 were predicted to be
in excess of 15 Tg/yr if new sources were required to meet only the State
standards. If six NSPS were set for particulate matter each year during the
1980's, the 1990 emissions were predicted to be considerably less than
10 Tg/yr.4
Attention is being given to particle size distribution to distinguish
the fraction of particulate matter that is likely to be inhaled and to the
chemical characterization of individual components. EPA has a network of
National Air Monitoring Sites (NAMS) in major urban areas to measure both
particle size and chemical composition. Monitoring at these sites provides
particulate data for enough urban areas with populations greater than 50,000
to permit the continued characterization of national trends in particulate.
Improvements in the characterization of particulate components help in the
assessment of shifts in the nature of particulate matter and in the selec-
tion of appropriate control strategies. A continuing concern is that a net
improvement in overall particulate levels could mask a shift in composition
toward the smaller particles.
2.2 SOURCES OF SUSPENDED PARTICULATE MATTER
Particulate matter is emitted into the atmosphere from a variety of
manmade and natural sources, and is sometimes formed in the atmosphere by
conversion of natural and anthropogenic gaseous constituents into particu-
late. This section discusses the sources of particulate matter and the
sources of gaseous precursors to particulate matter.
2-3
-------
2.2.1 Natural Em1s s i on Sources
Particulate emissions from natural sources are estimated to exceed
emissions from manmade sources on a worldwide basis.3 In areas remote from
urban population centers with concentrations of industrial and transporta-
tion facilities, natural particulate emissions are typically responsible for
more than half of the average measurable levels. Figure 2-1 indicates the
estimated average nonurban particulate levels in the continental United
States by region and the estimated fractions of these levels that are
attributable both to natural and to manmade sources, and to particulate
transported into the region.3
The most important of the natural particulate sources are soil and rock
debris, forest fires, volcanoes, and ocean salt spray. On land, wind-
entrained dust from soil and rock debris is the largest direct source of
particulate. Wind-entrained dust concentrations vary across the continent,
and they vary according to weather conditions. For example, in the Great
Plains region of the United States, wind erosion of soil is estimated to
produce more particulate than other sources in the region, and also can lead
to high dust concentrations over large areas. Although most duststorms
occur in the spring, they can be a problem in other seasons. Rainfall and
soil erosion of an area also influence the frequency and severity of dust-
storms.
The contributions of volcanoes and forest fires to particulate levels
can vary greatly. During most years, emissions from volcanoes do not con-
tribute a large proportion of the natural particulate emissions; but in
episodes such as the eruption of Mount St. Helens, Washington, starting in
May 1980, volcanic particulate is a major component of .regional particulate
levels. Although the contribution of forest fires cannot be accurately
determined, forest fires are important to urban air quality because such
fires are frequently adjacent to urban areas.3
Ocean salt spray, probably the largest particulate emission source, has
a limited effect that extends only a short distance inland.
2.2.2 Manmade Sources of Particulate
Manmade sources contribute less to overall particulate levels on a
nationwide basis than natural sources (Figure 2-1), but they are important
2-4
-------
40
35-
30
en
3.
§ 25
2
^
UJ
o
§ 20
cs
15
10
D CONTRIBUTION FROM
LOCAL MAN-MADE SOURCES
pig TSP TRANSPORTED
t^il FROM NEIGHBORING REGIONS
H NATURAL (BACKGROUND) TSP
WEST
MIDWEST
EAST
Figure 2-1. Estimated average contributions to nonurban TSP levels.
2-5
-------
with respect to air quality because they are usually concentrated near
population centers where they are generated. Such anthropogenic particulate
is the predominant component of particulate levels in many urban areas.3
Manmade particulate sources are sometimes grouped into four broad cate-
gories: stationary fuel combustion, industrial processes, solid wastes
disposal, and other significant sources. Table 2-1 summarizes the estimated
contributions from each of these categories.5
The concepts of fugitive emissions and fugitive dust must be addressed
in this discussion. Both of these terms refer to nonstack emissions of
particulates. Fugitive emissions result when particulate from industrial
operations finds its way to the atmosphere through building vents, windows,
doors, and leaks in hooding and ductwork or when particulate is emitted from
the open-air loading and transfer of materials. Fugitive dust, on the other
hand, is an emission that becomes airborne by the forces of the wind in
combination with man's activity. These emissions include windblown dust
from construction sites, paved and unpaved roads, tilled farmlands, and
raining operations. Fugitive dust, then, usually originates from nontradi-
tional sources, but can originate from natural causes such as duststorms.
2.2.2.1 Particulate Emissions From Stationary Combustion Sources. Par-
ticulate matter emitted from stationary combustion sources represents ap-
proximately 35 to 50 percent of the total particulate generated in the
United States by anthropogenic sources.2'3 Combustion source particulate
includes fly ash, soot, and sulfur oxide aerosols. Fly ash consists of
inorganic material from the fuel, which is not destroyed during combustion
and is subsequently entrained with the flue gas. Fly ash includes inorganic
oxides, salts, and trace metals.6 Soot consists o«f unburned carbon par-
ticles and polycyclic organic compounds formed under oxygen-deficient or
low-temperature combustion conditions. Sulfur oxide aerosols are emitted in
gaseous form from combustion sources that burn fuels containing sulfur. The
sulfur oxides often condense into aerosols, which can contribute to overall
particulate levels. Particulates formed from precursor gases after their
emission into the atmosphere, such as sulfur oxide aerosols, are termed
"secondary pollutants."
2-6
-------
TABLE 2-1. ESTIMATED PARTICIPATE EMISSIONS FROM MANMADE SOURCES, 1977s
_ Emission sources _ Tg/yr
Stationary fuel combustion
Utility boilers 3.4
Industrial boilers 1.2
Residential, commercial, institutional boilers 0 . 2
Subtotal 4.8
Industrial processes
Chemicals and petroleum refining 0.3
Metals refining 1,3
Mineral products 2.7
Miscellaneous industrial processes 1. 1
Subtotal 5.4
Solid waste disposal
Subtotal 0.4
Other significant sources
Transportation 1.1
Forest fires and agricultural burning 0.6
Miscellaneous 0.1
Subtotal -1.8
Total emissions from all sources 12.4
2-7
-------
Stationary combustion sources that emit particulate matter include
utility boilers, industrial boilers, residential space heating, and commer-
cial and industrial space heating. Utility boilers account for more than
half of the emissions in this category, followed by industrial boilers
(Table 2-1). The actual quantities of particulate matter emitted from a
stationary combustion source are dependent on the size of the source, the
efficiency of combustion, the efficiency of collection, and the amounts of
sulfur and ash in the fuel. Fuels commonly used in stationary combustion
sources include coal, oil, natural gas, and wood products. These fuels,
ranging from low-grade, high-sulfur coals to clean-burning natural gas, vary
considerably in their sulfur and ash content.
2.2.2.2 Emissions From Industrial Processes. Particulate emissions
from industrial processes are formed primarily through the mechanisms of
grinding, impaction, breakup of liquids, condensation, and chemical reac-
tion. These emissions are characterized by a wide range of particle sizes
and chemical compositions. The particles are chemically related to the
processes from which they are generated. Thus, smelting and metallurgical
operations, for instance, produce a large proportion of submicrometer-sized,
condensed metallic fumes. Carbon particles and organic condensables such as
tars are emitted from chemical operations related to the textile,
petroleum/petrochemical, and plastics industries.
Grinding procedures such as those used in rock crushing, flour milling,
and sanding operations produce predominantly large particles. High-
temperature processes and those using volatile compounds generally produce
aerosols of submicrometer size as a result of the condensation of vaporized
solids or liquids. Such processes include pyrolysis, vaporization of lubri-
cating or process oils, and metallurgical operations. Other sources that
produce substantial amounts of submicrometer particulate emissions are lime
kilns, pulp-mill recovery furnaces, and cement and asphalt plants.
Industrial-process fugitive emissions have the potential to contribute
significant quantities to particulate burdens, especially in the vicinity of
the source. With stack emission controls improving, the relative importance
of fugitive emissions is growing. Because these emissions generally enter
the atmosphere near ground level and at low velocities, their localized
2-8
-------
Impact on air quality can be large. Major industries with potential fugi-
tive emissions are listed in Table 2-2.2'7
2.2.2.3 Particulate Emissionsfrom Solid Wastes Disposal. Incinera-
tion and open burning were traditionally the most common methods of solid
waste disposal in urban areas. Under the pressure of more stringent pollu-
tion regulations, large municipal incinerators and industrial installations
have been upgraded and controlled. With the continuing trend toward land-
fills, recycling, and the use of combustible rubbish as a fuel substitute,
the decline of particulate emissions from solid waste disposal is expected
to continue.3
2.2.2.4 Other Significant Particulate Sources. The remaining sources
of particulates that could be considered significant may be grouped into two
categories. One consists of sources created directly by some manmade
action, such as emissions from vehicular exhausts, vehicular tire wear, and
construction and demolition activities. The other category consists of
re-entrainment sources, both natural and vehicle-related. Particulate
matter can accumulate on city streets and other paved areas from unpaved
roads and lots, truck spillage, sand and salt applied for snow control, and
sediments washed over roadways during heavy rains. These particulates can
become re-entrained by heavy winds and by vehicular traffic, and can produce
an area-wide source of temporarily suspended particulate.3
2.2.3 Transported Particu1 ate
As indicated in Figure 2-1, transported particulate represents a major
portion of the average nonurban particulate levels in the continental United
States. It is also evident that the importance of transported particulate
increases as air masses move with the prevailing wind patterns from the
Pacific Ocean across the continent. Transported particulate is both manmade
and natural, and consists of both primary and secondary particulates. Move-
ment of particulate from the source over a distance less than 100 km from
the monitoring site is considered to be short-range transport; movement over
more than 100 km is considered long-range transport.
The formation and transport of secondary particulate warrants special
attention because the highest concentration of secondary particulate is in
the size range of 0.01 to 1.0 urn.8'9 Particulate in this size range is
2-9
-------
TABLE 2-2, POTENTIAL INDUSTRIAL SOURCES OF FUGITIVE
PARTICULATE EMISSIONS2'7
Food/agricultural industry
Alfalfa dehydrating
Cotton ginning
Grain terminals
Grain processing
Metallurgical industry
Primary and secondary aluminum
Metallurgical coke
Primary copper
Ferroalloys
Iron
Steel
Primary and secondary lead
Gray iron and steel foundries
Primary and secondary zinc
Secondary brass/bronze
Mineral products industries
Asphalt concrete
Brick
Castable refractories
Cement
Ceramics/clays
Concrete batching
Coal cleaning
Glass
Gypsum
Lime
Phosphate rock
Stone quarrying
Potash production
Sand and gravel
Diatomaceous earth
Forest products industry
Lumber and furniture
2-10
-------
active in scattering light, and is also inhalable into the alveolar area of
the lungs. The main ingredients in formation of secondary particulate are
sunlight and gases such as sulfur oxides, ammonia, nitrogen oxides, ozone,
water vapor, hydrocarbons, and oxygen. The sulfur oxides and nitrogen
oxides are primary pollutants emitted in large quantities from combustion
sources and internal combustion engines, but can also be precursors of
secondary participate matter. Secondary participates arising from these
precursors include nitrates such as nitric acid, sulfates such as ammonium
sulfate and sulfuric acid, and organic particulates formed by the reactions
of volatile organic compounds in the presence of sunlight.
2.3 MEASUREMENT OF PARTICULATE FROM STATIONARY POINT SOURCES
Particulate matter emitted from point sources may be measured to deter-
mine compliance with applicable emission limitations, to evaluate control
equipment performance, or to establish emission factors. Many of the test
methods, however, are subject to biases that may influence the validity of
the results. The test procedures discussed here have been developed to
minimize or eliminate these biases in obtaining representative samples.
2.3.1 Mass Concentration Measurement
The most precise method of determining the mass concentration of par-
ticulate matter in a gas stream is to collect the entire volume of gas and
the particulate matter and to determine the mass concentration from this
sample. This procedure, however, is feasible only with a few sources (with
very low volumetric flow rates). Procedures for sampling small portions of
a gas stream to obtain a representative sample of the total gas stream have
been developed by various groups. Examples of these procedures are EPA
Reference Methods 5 and 17, American Society for Testing and Materials
(ASTM) Method D2928-71, and the American Society of Mechanical Engineers
(ASME) Power Test Code 27. The predominant test procedure for characteri-
zation of particulate matter is EPA Reference Method 5, "Determination of
Particulate Emissions from Stationary Sources," Appendix A, 40 CFR 60.
Quality assurance checks in Method 5 and use of the method with EPA Methods
1, 2, 3, and 4 help ensure the accuracy of mass concentration determina-
tions.
2-11
-------
Method 5 is based on extractive filtration. Gas is extracted isokinet-
ically; that is, the velocity of the gas entering the sampling nozzle is
equal to the gas velocity passing by the nozzle at that sampling point.
Extraction is done through a nozzle to an externally heated filter held at
120°±14°C. The particulate matter is captured in the sampling probe and on
the filter, and then the filtered gases are passed through a series of
impingers to remove moisture and other components before they pass through a
dry gas meter. The sampling apparatus is shown as Figure 2-2. Isokinetic
conditions must be maintained within ±10 percent of 100 percent for a valid
test. In a gas stream with both large and small particles, sampling rates
lower than 100 percent isokinetic can bias the sample toward larger par-
ticles, and can strongly bias the mass concentration calculations. The
reverse is true with sampling rates above 100 percent isokinetic, in which
the bias toward smaller particles would result in an apparent mass concen-
tration that is lower than the actual emission rates.
Establishing isokinetic sampling rates depends on the characteristics
of the individual sampling train and on determination of gas velocity, gas
volumetric flow rate (EPA Method 2), gas molecular weight (EPA Method 3),
and gas moisture content (EPA Method 4). The location and suitability of
the sampling site and the location of the sampling points to provide a
representative sample of the gas stream are performed according to the
procedures of EPA Method 1. Thus the use of EPA Method 5 depends on the
proper use of-other EPA test methods, each of which affects whether the mass
concentration data will be representative of the actual emissions from a
stationary source. A brief review of these methods and their uses with
Method 5 is necessary in evaluating test results.
The EPA Method 1 specifies criteria for selecting the sampling location
and the location and number of sampling points. Emphasis is on locating the
sampling locations away from flow disturbances such as fans, bends in ducts,
and duct expansion or contraction points. The duct is divided into equal
areas, and sampling points are at the centroid of each area. Generally, the
closer a sampling location is to upstream or downstream disturbances, the
more sampling points are needed to obtain a representative sample. The
existence of cyclonic flow is checked because angular velocity patterns can
lead to erroneous velocity determinations and to nonisokinetic sampling
conditions, which can result in biased mass concentrations.
2-12
-------
(S3
I
THERMOMETER
HEATED AREA \
\ lp FILTER HOLDER
THERMOMETER
TEMPERATURE
SENSOR
PROBE
REVERSE-TYPE
PITOT TUBE
'-SILICA GEL
VACUUM LINE
"kifcixti-r**"* L*~m, I/
PITOT MANOMETER THERMOMETERS "MPINGERS ^ICE BATH
O O BY-pAJS VALVE
ORIFICE
DRY GAS METER AIR-TIGHT
PUMP
Figure 2-2. EPA Method 5 participate sample apparatus.
-------
The EPA Method 2 is used to determine local velocity pressures for
establishing isokinetic sampling rate. Average gas velocity and volumetric
flow rate may be calculated from a traverse of all sampling points. Typi-
cally, this method uses a Type S pitot tube because it yields a higher
reading at a given velocity pressure than a standard pitot tube and because
it is resistant to plugging. A Type S pitot tube is shown in Figure 2-3.
The EPA Method 3 is used to determine gas molecular weight, a value
needed in determining gas velocity and volumetric flow rate. On combustion
sources, an Orsat analysis for oxygen, carbon dioxide, and carbon monoxide
is typically performed to assist in determining excess air and F-factor
calculations for heat input and mass concentrations. Values obtained can
aid in the determination of representative source-operating conditions and
in the calculation of mass emissions in units specified by various stan-
dards.
The EPA Method 4 is used in the determination of moisture in stack
gases. Although the moisture content is determined from the impinger catch
of Method 5, a value for moisture content must be assumed for isokinetic
flow calculations. For gas streams with low moisture content (e.g.,
3 to 10 percent), the errors in isokinetics caused by assuming the wrong
moisture content are relatively small. At high moisture content (greater
than 10 percent), however, a small error in the estimated moisture can lead
to sampling rates outside the acceptable range of 90 to 100 percent. For
example, if the estimated moisture was 50 percent and the actual was 45
percent, the calculated isokinetic rate would be 111 percent. Method 4 may
be used to aid in establishing a moisture content value for use in
Isokinetic sampling calculations. When high moisture content is encountered
during particulate sampling, care should be taken to properly heat the probe
and filter to avoid premature condensation.
For many sources, the amount of particulate matter captured on the
filter is a function of temperature. In-stack filtration methods allow fil-
tration to occur at approximately the same temperature as that of the stack
gas. Thus, the amount of particulate captured should vary among sources,
depending on the stack temperature and on the degree to which the particu-
late is temperature dependent. By use of an external filter of defined
filtration characteristics at 120°±14°C, the captured particulate on the
2-14
-------
1.90-2.54 CM
(0.75-1.0 IN.)
_L_d
T~=
TEMPERATURE SENSOR
TYPE S PITOT TUBE
LEAK-FREE
CONNECTIONS
Figure 2-3. Type S pilot tube and manometer.
2-15
-------
filter and in the heated probe is defined. This is the "front-half catch"
of the Method 5 sampling train, Condensible participate allowed to pass
into the impingers along with water vapor is the "back-half catch." In
certain jurisdictions, the back-half catch is included with the front-half
catch as total particulate.
Significant biases may be introduced into the sampling results if the
particulate is strongly temperature dependent and if the sample is not held
at the proper temperature. Even though the filter holder is placed in a
heated box within the 120°±14°C range, excessive heating or cooling in the
sampling probe can affect the results. Of most concern is excessive probe
heating, which can be caused either by a high probe-temperature setting or
by a stack gas temperature higher than 120°C. Although the heated box may
be at the proper temperature, the actual gas stream filtration temperature
may be much higher. An excessive probe or filter temperature is of more
concern in compliance tests, since the high temperature could tend to bias
the mass concentration low.
Where there is no temperature dependency, EPA Method 17 uses an in-
stack filter for particulate capture. As in Method 5, the sampling rate is
isokinetic; EPA Methods 1, 2, 3, and 4 are used with the particulate samp-
ling.
2.3.2 Particle Size Analysis
As part of the particulate emission characterization, a distribution of
particulate sizes may be useful for determining control equipment param-
eters. The cascade inertia! impactor is the device most commonly used for
particulate sizing. The sampling train consists of the probe, a precutter
such as a cyclone, and the cascade impactor.
The cascade inertia! impactor technique provides a distribution of
aerodynamic particle diameters. A cascade impactor usually has 5 to 10
stages of decreasing orifice diameters. The impactor is usually assembled
to give an alternating pattern of orifice plates and collection plates. As
the orifice size decreases, the gas velocity through each orifice increases.
Larger particles cannot overcome the inertia! force imparted to them through
the orifice and thus impact the collector plate. Smaller particles have
less inertia, and so the gas stream carries them to the next stage. The
2-16
-------
last stage is usually followed by a filter to capture the smallest particles
that have escaped impaction. Gravimetric methods are used in analysis of
each stage to determine particle size distribution, geometric mass median
diameter, and geometric standard deviation. The results of cascade
impactors are influenced by the deposition of particulate in the probe. For
example, one test indicated that at a velocity of 15 m/s, 33 percent of the
10-umA particles were collected in the probe.10
Cascade impactors are typically in situ (i.e., in-stack) devices used
with isokinetic sampling rates. When samples are obtained in situ at the
stack temperature, the particle size distribution should be representative
of the actual particle size distribution in the duct. Failure to sample
isokinetically results in the particle size distribution being biased and
unrepresentative. A bias toward larger particle sizes occurs with under-
isokinetic sampling (velocity entering nozzle is lower than the localized
gas velocity), and bias toward smaller sizes occurs with overisokinetic
sampling.
Cascade impactors are provided in stages with nominal values for aero-
dynamic cut-size diameters. Calibration procedures are usually provided by
the impactor manufacturer. Each impactor should be calibrated periodically
to determine the actual value of the cut-size diameter for each stage.
Cascade impactors are susceptible to several problems. First, in gas
streams with high particulate loadings, material may build up on the stages
quickly and thus shorten the available testing period. Second, particle
reentrainment and bounce can result in the particle size distribution being
prejudiced toward smaller particles. Finally, fracturing of the larger
particles at the impaction stage may lead to generation of fine particulate
and to a consequent bias toward small particle sizes.
Although in situ cascade impactors are probably more common, extractive
external cascade impactors are also in use. The particle size distributions
obtained with these models are representative of the temperature of the
impactor when there is temperature dependency of the particulate. Extrac-
tive cascade impactors may be used in observing temperature effects on
particle size distribution and particle growth. The results from in situ
and extractive cascade impactors should not be combined, because the samp-
ling conditions are often different. Sample losses to the walls of the
2-17
-------
sampling probe are a potential problem with extractive samplers. The ef-
fects of over- and underisokinetic sampling are similar with both types of
impactors.
Cyclones are used for in situ and extractive aerodynamic particle
sizing, but not to the extent of cascade impactors. The aerosol sample en-
ters the cyclone through a tangential inlet and follows a vortex flow pat-
tern. Particles that cannot follow the gas streamlines move outward toward
the cyclone wall and, depending on cyclone geometry, gas flow rate, and
particle size, may reach the cyclone walls and be collected. By use of a
series of cyclones of different geometric dimensions at a constant flow
rate, particles can be removed according to size from a gas stream. The
fractionating capability of cyclones is not predictable by theoretical means
to the degree of accuracy possible with impactors. The advantages of cy-
clones over impactors is that large samples can be acquired and particle
reentrainment is not so great.
Realtime particle sizing/counting has received minimal application to
characterization of source emissions chiefly because the techniques require
low mass concentrations. The instruments used in realtime analyses include
optical devices, diffusion batteries, condensation nuclei counters, and
electrical mobility analyzers.
Laboratory size distribution analysis of collected particulate samples
is often performed instead of in situ procedures. The results of these
methods must be interpreted with great caution, however, because the
original flue gas particle size distribution is almost impossible to recon-
struct under laboratory conditions. Particles or particle groups may be
altered from their gas-stream state by additional agglomeration or particle
breakup during sample collection. Size distribution results based on sedi-
mentation and elutriation, centrifuging, sieving, and electronic counting
are meaningful only when the effects of sample collection and redispersion
are negligible or clearly known.
Microscopic analysis is regarded as the fundamental technique for
counting and sizing particles. This procedure involves manual or computer-
ized microscopic examination of a prepared slide containing a representative
sample of the aerosol. Careful procedures must be followed in preparing the
slide so that the aerosol sample is not altered from its in-stack state.
2-18
-------
Although, microscopic examination of participate matter does not yield size
information in terms of aerodynamic diameters, it can produce useful data on
particle surface features, agglomeration, size, composition, and shape.
2.3.3 Analysis of Particulate Samples
2.3.3.1 General Considerations. Following the collection of a partic-
ulate sample at the sampling site, the sample must be analyzed to obtain a
quantitative or qualitative measurement. Generally, the exposed filter or
collection surface is returned to a laboratory for analysis. During this
transfer, care must be exercised to avoid loss of fibers or particulate
matter from the filter or the collection surface and to protect the sample
from damage or conditions that may affect the analytical results. Special
filter cartridges or filter holders are often used to safeguard the sample.
Also transferred with the sample is an information record containing the
site and sampler identification, the quality assurance data, and other
pertinent information.
2.3.3.2 Artifact Mass. Artifact particulate matter can be formed on
the surfaces of alkaline filter materials, such as glass fiber, by oxidation
of acid gases in the sample air. Formation of artifact particulate results
in an artificially high particulate measurement. This surface-limited ef-
fect usually occurs early in the sample period, and is a function of the
filter pH value and the presence of acid gases. Although the artifacts
usually account for only a small part of the collected particulate, the
error can be significant when sampling periods are short or when the total
amount of particulate matter collected is very small.
Reactions between various particulates collected on the filter are also
possible. Although such reactions may not significantly affect gravimetric
determinations, they may affect the chemical analysis of the sample.
2-3.3.3 Loss of Volatiles. Volatile particles collected on the filter
may be totally or partially lost during subsequent sampling, during trans-
port to a laboratory for analysis, or during storage before the postexposure
weighing. Filters are normally analyzed as soon after collection as prac-
tical, but some loss of volatiles is inevitable.
2-19
-------
2.3.3.4 Gravimetric Determination. Filters are conditioned and then
weighed before and after exposure to determine the particulate weight as the
net weight gain of the filter or the front-half of other collection sur-
faces. Mass concentration is determined directly by dividing the weight of
the collected participate by the total (standard) volume of air sampled. In
addition to potential errors described earlier, errors are also possible in
the weighing process.
2.3.4 Chemical Analysis and Analysis of Trace Elements
Most methods for trace element analysis of particulate material use
spectroscopic detection. The detectors respond to the presence of only an
element and provide no information about chemical compounds. Most methods
do not indicate the oxidation state of the element.
2.3.4.1 Atomic Absorption Spectrometry. Atomic absorption spectrom-
etry11 is widely used for quantitative elemental analysis of airborne parti-
cles. It usually involves an acid extraction and excitation of the solution
by a flame. Light with a wavelength characteristic of the one element of
interest traverses the flame. The amount of light absorbed is related to
the quantity of the element present.
Individual elements must be determined sequentially. Thus, although
any element can be determined for which a lamp is available to produce the
characteristic light, particulate samples are often large enough to allow
only half a dozen determinations. Moreover, some trace elements present in
the particles (including antimony and arsenic) may require the application
of special methods.
Atomic absorption is subject to significant interferences and can lead
to substantial errors. If recognized, these errors generally can be ac-
counted for or eliminated to produce good quantitative analyses at the ex-
pense of additional effort on each sample. Despite its drawbacks, atomic
absorption spectrometry is a useful method for elemental analysis of partic-
ulate.
2.3.4.2 Optical Emission Spectrometry. A variety of methods can be
used to excite rather loosely bound electrons in elements and to observe
2-20
-------
characteristic emissions as de-excitation occurs. The wavelength is charac-
teristic of the element, and the intensity is an indication of the quantity
of the element present. The most desirable technique is argon plasma exci-
tation of an acid extract of the particulate matter.12
Plasma spectrometry offers more advantages than atomic absorption
since sample preparation and analysis rates are essentially equal. The
spectrometry techniques, however, can simultaneously determine up to, per-
haps, 50 elements with detection limits about the same as those of flame
atomic absorption. Spectrometry is also more interference-free than atomic
absorption, although not totally so.
2.3.4.3 Spark Source Mass Spectrometry. Spark source mass spectrom-
etry13 can analyze particulates separated from the filter and oxidized, or
can be extracted with an acid. Spectral interferences are important, but
generally can be overcome by use of a spectrometer with high resolution.
The precision can be about 30 percent (relative standard deviation) in
careful analytical work. As with any multielement technique, its accuracy
may depend on the element and on the matrix. The advantage of this tech-
nique is that one can simultaneously estimate the quantity of every non-
volatile element in the periodic table and do so with roughly equal sensi-
tivity.
2.3.4.4 Neutron Activation Analysis. Neutron activation analysis13'14
consists of a variety of distinct methods, all of which produce unstable
nuclei that emit gamma radiation. The energy and intensity of the gamma
rays are indicators of the element and its quantity. Instrumental thermal
neutron activation analysis is most commonly used. In this approach, a
nuclear reactor is used to produce unstable nuclei. Neutron activation
analysis can simultaneously determine up to 25 elements in particulate
samples. Another advantage is that particles can be analyzed as received
directly on the filter surface.
2.3.4.5 X-Ray Fluorescence Spectrometry. X-Ray fluorescence spectro-
metry13'15 involves excitation of tightly bound electrons and observation of
the X-ray emission as de-excitation occurs. Excitation may be done by a
variety of techniques, but use of an X-ray generator is the most common.
2-21
-------
The technique may be either multielement (up to perhaps 30) energy disper-
sive detection or wavelength dispersive detection (up to perhaps 10 ele-
ments). Only elements with atomic numbers greater than that of magnesium
can be analyzed. Particles can be analyzed nondestructively, directly on a
filter; however, if samples are not thin and of uniform surface texture,
certain corrections must be made. Interferences are common and must be
considered, and adequate calibration can be a problem.
2.3.4.6 Electrochemical Methods. Electrochemical methods have been
used to a limited extent to determine a small number of elements in par-
ticles. These methods include potentiometry with ion-selective electrodes,
polarography, and anodic stripping voltammetry.13 Electrochemical methods
have few advantages for particle analysis aside from the low initial capital
costs of equipment relative to that needed for other techniques.
2.3.4.7 Chemical Methods. Many wet chemical procedures constitute the
classical methods used for trace element analysis of particulate. In gener-
al, a color-forming reagent is used, and the amount of an element is deter-
mined by the extent of color development. Probably the best known of these
procedures is based on the use of dithiocarbazone (dithizone)16 as the
colorimetric reagent for lead. Wet chemical procedures are labor-intensive
and slow, compared with spectral techniques, particularly since only one
element can be determined at a time. Interferences can also be a problem.
2.3.4.8 Analysis of Organics. Procedures for estimating the total
mass of benzene-extractable organic material in particulate matter have been
used occasionally. A portion of the front-half catch is placed in a Soxhlet
extractor and refluxed with benzene for several hours. Then the benzene is
volatilized, and the mass of the residue is measured. This procedure
presents problems because of special requirements for handling benzene.
Methods of identifying and determining individual organic species
abound. These methods use different sequences of solvent extractions that
separate groups of different organic species on the basis of solubility.
Solutions are often subjected to chromatographic separation with mass spec-
tral detection. For organic compounds that are volatile up to about 300°C,
gas chromatography-mass spectrometry (GC-MS) can be used.1T For organic
2-22
-------
species with lower volatility, liquid chromatography might be used. High-
performance liquid chromatography (HPLC)18 is typically used, but none of
these procedures permits a high rate of analysis.
For analysis of one species of longstanding interest, benzo-a-pyrene
(BaP), thin layer chromatography (TLC) with fluorescence detection is often
used, and HPLC procedures have been proposed. The TLC procedure requires a
cyclohexane extraction, spotting, and development of a TLC plate, with
fluorescence detection. This procedure is more interference-free than some
HPLC methods, and it has a higher production rate.19
2-23
-------
REFERENCES
1. U.S. Environmental Protection Agency. National Air Pollutant Emissions
Estimates, 1970-1978. EPA-450/4-80-002. Research Triangle Park, N.C.
January 1980.
2. Adamson, L. F., and R. B. Bruce. Suspended Particulate Mattel—A
Report to Congress. U.S. Environmental Protection Agency, Environ-
mental Criteria and Assessment Office. Research Triangle Park, N.C.
October 1978.
3. U.S. Environmental Protection Agency. National Assessment of the Urban
Particulate Problem, Vol. I. EPA-450/3-76-024. Research Triangle
Park, N.C. July 1976.
4. U.S. Environmental Protection Agency. Priorities and Procedures for
Development of Standards of Performance for New Stationary Sources of
Atmospheric Emissions. EPA-450/3-76-020. Research Triangle Park, N.C.
May 1976.
5. U.S. Environmental Protection Agency. National Air Quality Monitoring
and Emissions Trends Report, 1977. EPA-450/2-78-052. Research
Triangle Park, N.C. December 1978.
6. Faoro, R. B., and T. B. McMullen. National Trends in Trace Metals in
Ambient Air, 1965-1974. U.S. Environmental Protection Agency.
Research Triangle Park, N.C. EPA-450/1-77-003. February 1977.
7. U.S. Environmental Protection Agency. Technical Guidance for Control
of Industrial Process Fugitive Particulate Emissions. EPA-450/3-77-
010. Research Triangle Park, N.C. March 1977.
8. Whitby, K. T., R. B. Husar, and B. Y. H. Lia. The Aerosol Size Distri-
bution of Los Angeles Smog. J. Aero. Atmos. Chem. (ed. G. H. Hidy)^
Academic Press, New York, 1971.
9. Clark, W. E., and K. T. Whitby. Concentration and Size Distribution
Measurements of Atmospheric Aerosols and a Test of the Theory of Self-
Preserving Size Distributions. J. Atmos. Sci., Vol. 24, 1967.
10. Smith, W. B., et al. Sampling Data Handling Methods for Inhalable
Particulate Determination. Southern Research Institute. Birmingham,
AL. March 1981.
11. Ahearn, A. I. Trace Analysis by Mass 'Spectrometry. Academic Press,
New York, 1972. I
2-24
-------
12. Fassel, V. A. Quantitative Elemental Analyses by Plasma Emission Spec-
troscopy. Science, Vol. 202, 1978.
13. Morrison, G. H. Trace Analysis Physical Methods. Wiley Interscience,
New York, 1965.
14. Lambert, J. P. F., and F. W. Wilshire. Neutron Activation Analysis for
Simultaneous Determination of Trace Elements in Ambient Air Collected
on Glass-Fiber Filters. Anal. Chem., Vol. 51, 1979.
15. Goold, R. W., C. S. Barrett, J. B. Newkirk, and C. 0. Ruud. Advances
in X-ray Analysis. Kendall/Hunt, Dubuque, Iowa, 1976.
16. Snell, F. D. Photometric and Fluorometric Methods of Analysis, Metals.
Part I. John Wiley, New York, 1978.
17. McFadden, W. H. Techniques of Combined Gas Chromatography/Mass Spec-
trometry. John Wiley, New York, 1973.
18. Kirkland, J. J. Modern Practice of Liquid Chromatography. John Wiley,
New York, 1971.
19. Swanson, D., et al. A Rapid Analytical Procedure for the Analysis of
BaP in Environmental Samples. Trends Fluoresc., Vol. 1, 1978.
2-25
-------
-------
SECTION 3
ALTERNATIVE PARTICULATE CONTROL APPROACHES
Control of particulate emissions may be achieved either by prevention
of particle generation or by collection of particles entrained in effluent
gas streams. Prevention is preferable, both economically and environmental-
ly; however, opportunities for this control approach are limited. A more
commonly available strategy is process modification or optimization to
improve particle collectibility. The costs of particulate control systems
can be reduced, and their reliability can be improved by increasing the
particle size, reducing the particulate mass loading, or reducing the vari-
ability of process operation; these measures can be combined to improve
control. It is sometimes possible to modify particle characteristics to
maximize collectibility.
These control measures represent alternatives to, or supplements to,
installation of conventional particulate control systems. The following
sections deal with these broad issues. Fuel switching is the most common
and the most useful means to "prevent" particulate emissions. General
approaches to process optimization illustrate possible benefits to sub-
sequent control system performance. Finally, the control devices are brief-
ly addressed. More detailed information on particulate control systems is
in Section 4, and summaries of important processes are in Volume 2 of this
document.
3.1 ENERGY SOURCE AND FUEL SELECTION
Energy substitution can be an effective and useful technique for re-
ducing particulate emissions from stationary combustion sources. Substitu-
tion has special value in control of small and old sources, for which the
cost of effective control devices might be expensive relative to the worth
of the facility. Application of this approach is contingent on fuel avail-
ability and on the reduction in emissions to be achieved.
3-1
-------
The particulate reduction potential may be estimated initially by com-
parison of published emission factors for various fuels. Because of site-
specific combustion characteristics and highly variable fuel properties,
such emission factors can provide only an estimate of the emission reduction
potential. An emission factor analysis is presented in Table 3-1 for a
pulverized-coal-fired power boiler.1 The use of natural gas reduces emis-
sions by 50 percent relative to No. 6 fuel oil and by 80 percent relative to
coal firing with low-efficiency collection. Other clean fuels now avail-
able, such as distillate oil and refuse-derived fuels, could be included in
a similar analysis. Because of the limited supplies of naturally clean
fuels, this control option is severely restricted to use with only marginal
boilers or furnaces where gas cleaning is not economically feasible.
TABLE 3-1. PARTICULATE EMISSION REDUCTION POTENTIAL OF VARIOUS
ENERGY SOURCES
Energy source
Natural gas
No. 6 fuel oil
Bituminous coal
Assumed
particulate emission
control , %
0
0
98.0
Particulate emission,
ng/J (lb/106 Btu) of
delivered energy
21 (0.048)a
47 (0.108)b
140 (0.320)
aBased on an emission factor1 of 15 lb/106 ft3 for gas with a
heating value of 37,300 kJ/m3 and on an estimated conversion
efficiency of 31.4 percent.
Based on an emission factor1 of 5 lb/1000 gal for oil to
0.3 percent sulfur content with a heating value of 42,000
kJ/liter and on an estimated conversion efficiency of 30.8
percent.
Development of synfuels will ultimately provide a more plentiful supply
of clean fuels. Many synfuels, such as high-Btu gas and liquid solvent-re-
fined coal, are believed to have combustion characteristics and particulate
emission rates similar to those of natural gas. Many of these products
could possibly be used without add-on particulate control devices. Other
synfuels may have low ash content, but undergo moderate carbon losses that
demand some degree of gas cleaning. Little information is yet available on
particulate emission rates of synfuels.
3-2
-------
Substitution of electric-powered devices for onsite power generation
may provide relief from control requirements. This option again is most
appropriate for small, economically marginal units. In this option, the
particulate control requirement is simply transferred to the power genera-
tion facilities, most of which are equipped with high efficiency particulate
control systems.
3.2 PROCESS OPTIMIZATION
Process optimization involves modifications of process feed materials,
process unit functions, and process variables to eliminate or reduce
particulate emissions.
Optimization of the process may reduce the volume of exhaust gases or
alter the particle size distribution. A change in particle size distri-
bution may allow a broader selection of abatement equipment, and emission
standards may be achieved by application of equipment with lower energy
consumption.
Selection of a process and implementation of process modifications may
affect other process requirements, yields, or nonparticulate emissions.
Therefore, in evaluation of process changes as means of particulate control,
the total impact of the changes must be considered.
3.2.1 Modification of Process FeedMaterials
The physical properties of feed materials, such as particle size,
chemical composition, and moisture, may have significant effects on emis-
sions from industrial processes. The effects vary from process to process,
and the relationships must be developed on a process-by-process basis. In
general, where heated air is used for drying, fine particles in the process
feeds cause an increase in particulate emissions. Screening or cleaning of
raw materials can reduce the particulate emissions per ton of product.
Following are examples of process modifications that can reduce emissions
resulting from the properties of feed materials.
3.2.1.1 Phpsphate Rock Dryers. The screening or washing of phosphate
rock before drying can reduce the weight of fine materials in the feedstock.
These fines, referred to as phosphatic clay, have a substantial impact on
the emission rate from uncontrolled dryers. The particulate loadings in
3-3
-------
dryer exhaust gases at several facilities has varied from 1.14 x io3 to 8.00
x 10s kg/ms, depending on the condition of rock being processed. Removal of
fine materials from the feed can permit operation of the control device at
lower energy consumption and higher efficiency.
3.2.1.2 Secondary Lead and Copper Smelters. Operation of cupolas and
blast furnaces at lead and copper smelters is affected by the composition
and size of feed materials. The presence of fines in the charge reduces the
furnace thermal efficiency, and requires the addition of excess coke. This
results in higher combustion air pressure and flue gas volume. Fluctuation
of the charge composition and the inability of the furnace to respond rap-
idly can cause severe swings in furnace temperature and exhaust gas volume;
these changes can create positive pressures at charge doors, and generate
fugitive emissions during charging. Control of the size and composition of
the feed can reduce the capacity requirements for abatement equipment to
contain and control the fugitive emissions.
3.2.1.3 Textile Fabric Coating. Curing of chemical coatings and fin-
ishes on textile substrates normally leads to the release of condensible
hydrocarbon aerosols. The condensed organics can exhibit high opacity at
relatively low particulate loading. The emissions normally result from
chemicals incorporated into the fibers in previous processes. Removing
these components prior to treatment has been shown to be effective in reduc-
ing particulate emissions from these processes. Application of latex coat-
ings to textile fabrics that contain surfactants and plasticifiers has
resulted in high opacities, and has necessitated the use of afterburners to
control emissions. The need for afterburners can be reduced by modification
of the chemical composition of the surfactants used in the coating system.
In one instance, the surfactant was commercially available ammonium stea-
rate. The emission was composed of oleic and palmitic acids, which had been
introduced into the process as impurities in the ammonium stearate. When
the stearate was converted to a purer form and the usage rate reduced,
opacity of emissions dropped from 100 to 25 percent.
3.2.2 Elimination of Process Steps
The manufacture of products can require many individual process steps
involving simple functions. The transfer of materials from one process to
3-4
-------
another can result in fugitive particulate losses. The loss of product can
increase the cost of the finished marketed material, and also necessitates
the application of pollution abatement devices. Often, a careful analysis
of the number, order, and types of process steps can enable a company to
reduce the number of emission points and to reduce emissions by eliminating
repetitive and wasteful handling of materials.
The process changes many include installation of longer conveyors,
transfer of product by pneumatic instead of open conveyors, or combination
of process steps. In view of the energy required to transport materials and
due to the cost of ductwork and abatement equipment, the elimination of even
a single emission point can be cost effective.
3.2.3 Changes In Process Particle Characteristics
The particle size of the product being processed or handled can have a
direct effect on the particulate emission rate. The wetting or agglomera-
tion of materials can increase the effective particle size and the effi-
ciency of control equipment. An example of change in particle characteris-
tics that can reduce emissions is the transfer of wood fibers by using
cyclones and pneumatic systems in the fiberboard industry. The uncontrolled
emission of fibers from a cyclone handling 5000 kg/h of fiber can be as high
as 300 kg/h. Prior to transfer, partial polymerization of the heat-setting
resin that coats the fibers can reduce the emission to less than 5 kg/h.
3.3 EXHAUST GAS CLEANING
In a number of industrial processes, particulate emissions cannot be
controlled satisfactorily by fuel switching or process optimization. In
such processes, abatement of emissions to within the regulatory limits is
usually achieved by adding exhaust gas cleaning devices. Among the many
devices available for exhaust gas cleaning are cyclones, multicyclones, and
other mechanical devices; shaker-type, reverse-air, and pulse-jet fabric
filters; hot- and cold-side electrostatic precipitators; spray chamber,
venturi, and packed-bed scrubbers; and incinerators. Each of these devices
operates according to one or more of the basic physical or chemical princi-
ples discussed in Section 4.1, and each has distinctive advantages and
disadvantages, as discussed in Sections 4.2 through 4.6.
3-5
-------
Selection of a control device may involve a complex set of variables
including regulatory limitations; the nature of the emissions source and its
exhaust gases; the removal efficiency of each device; and long-term relia-
bility, ease of maintenance, and total costs of installing and operating the
device. Useful information is available in journals and other technical
literature; in U.S. EPA publications; in publications of control device
vendors and their representatives; through trade associations, professional
organizations, and their'.technical committees; and through paid consultants.
References 2 through 8 are examples of many publications which can provide
initial direction in the search for information about exhaust gas cleaning.
3.3.1 Applicable Regulations
Air pollution control regulations vary with jurisdiction, and are sub-
ject to periodic revisions. Some regulations are promulgated at the Federal
level: for example, New Source Performance Standards (NSPS),9 National
Emissions Standards for Hazardous Air Pollutants (NESHAPS),10 National
Ambient Air Quality Standards (NAAQS), and Prevention of Significant Deteri-
oration (PSD). Other air emission regulations are promulgated at the State
or local level: for example, those in State Implementation Plans (SIP}.11
Certain regulatory and enforcement functions related to air emissions are
retained at the Federal level, but many enforcement functions have been
delegated to the States. Some States, in turn, have delegated much of their
authority to municipalities, counties, or regional air quality agencies.
The first step in determining regulatory requirements for a particular
installation is to determine which local, State, and Federal agencies have
promulgated applicable regulations. These agencies should be contacted for
preliminary advice on the applicability and interpretation of current regu-
lations. If analyses show that fuel switching and process optimization are
not feasible, the technical and economic analyses of the various exhaust
gas cleaning devices must be begun.
3.3.2 Source Characteristics
The characteristics of the source must be defined as completely as pos-
sible to select the most appropriate control device. Single value estimates
of exhaust gas flow rates, temperatures, and particulate loadings usually
are not necessarily sufficient for the reliable design of a control device.
3-6
-------
characteristics of a source that are important to the design of a control
device also include the process operating schedule, variability of fuel
quality, the type of the raw materials, and the variability in process
operations. These characteristics determine the variability of such exhaust
gas parameters as temperature, moisture content, gas flow rate, particulate
concentration, particulate size distribution, and concentrations of chemical
constituents. A control device must be selected to provide the desired
efficiency and reliability over the anticipated range of exhaust gas condi-
tions. For example, the fabric in a fabric filter system must be chosen to
withstand both the expected high and low temperature excursions and the
typical or average temperature.
3.3.3 Control Device Design Limitations
Each type of device is limited in its capability to remove particulate
from exhaust gases. In general, the particulate removal efficiency of each
device is specific for each set of operating conditions. To provide the
desired level of abatement, a device must be properly matched for the
process conditions, the design must incorporate a sufficient factor of
safety to handle unexpected contingencies, and the design limitations must
not be exceeded because of increases in production rates after the device is
installed.
3.3.3.1 Mechanical Collectors. Mechanical collectors are efficient
for large particulate, but they cannot be expected to adequately remove fine
particulate. Nor can they be expected to perform well on processes in which
flow rates are extremely variable. A large increase in gas flow rates may
increase the efficiency of mechanical collectors, but may also reduce their
life by increasing abrasion. In contrast, a large reduction in gas flow
rates will significantly decrease the efficiency.
3.3.3.2 Electrostatic Precipitators. Electrostatic precipitators must
be sized to accommodate the expected gas flow rates, particulate concentra-
tions, and fly ash resistivities. The specific collection area (SCA) must
be great enough to handle the expected range of conditions. Exhaust gas
loadings and fly ash resistivities must not be altered to such an extent
that the SCA limitations of the precipitator are exceeded. In general, a
3-7
-------
properly designed electrostatic precipitator will collect nearly all partic-
ulate greater than 1 nn>A in diameter, but participate penetration increases
as particle size decreases.
3.3.3.3 Fabric Filters. The principle limitations of fabric filters
are the temperature limitations of available fabrics and the effects of
fabric failures on the penetration of particulate. A properly sized fabric
filter operating under dry conditions and within the temperature limitations
of the fabric can provide extremely efficient collection over a wide range
of particle sizes. Immediate replacement of broken filter bags is important
in maintaining a low particulate emission rate.
3.3.3.4 Wet Scrubbers. A variety of wet scrubbers are available, each
with certain performance limitations. Most wet scrubbers can operate effi-
ciently in collecting large particulate (>2 pmA); some are also efficient
for very small particulate (<0.2 jjmA). Therefore, within many wet scrubbers
there is a medium particle size "window" for which removal is less effi-
cient. Particulate collection generally decreases as energy input decreas-
es.
3.3.4 Control Device Reliability
The ultimate purpose of a normally efficient exhaust gas-cleaning
device can be compromised if it suffers from frequent malfunctions. Mal-
functions can reduce the performance of control equipment or cause periods
of uncontrolled emissions; on occasion, plant production must be halted
while the device is repaired. Malfunctions are caused by design deficien-
cies such as undersizing or omission of important ancilliary features or by
improper operation and maintenance. Many of the common malfunction modes as
well as design features and operating and maintenance procedures that can
reduce the frequency and severity of malfunctions are described in Sec-
tion 4.
3.3.5 Control Equipment Costs and FinancialAssistance
A financial management decision regarding the purchase and operation of
an exhaust gas-cleaning device should be aimed at selecting a device that
will provide efficient reliable service over the desired service life at the
lowest possible total annualized cost. An annualized cost determination
3-8
-------
must consider the amortization of control device investment, the direct and
indirect costs of operating and maintaining the device, any credits for re-
covered particulate, and the effects of the device, if any, on production
rates.12'13'14 Tax relief or special financing schemes that could favorably
affect the total annualized cost are sometimes available.13'15
3.3.5.1 Installed Costs of Control Equipment. The installed costs of
control equipment include the costs of engineering and designing the equip-
ment, the costs of materials of construction, cost of manufacturing, cost of
transportation, and costs of labor and equipment during installation. Other
costs associated with equipment installation may include the costs of pro-
duction loss during installation and the costs of an initial stack test to
verify performance of the device.
3.3.5.2 Direct and Indirect Operating Costs. Direct costs of operat-
ing a control device include the costs of utilities (e.g., electric power)
to run the equipment and the costs of labor for operating the equipment.
Other costs can include those for disposal of any collected particulate (if
the collected particulate has economic value, this can be a negative cost or
a net savings), the costs of periodic stack tests, and the costs of lost
production due to control device malfunctions. A particulate control device
may also have a positive or negative effect on production rates of some
processes. Maintenance costs for a control device include the costs of
replacement parts, of maintaining a spare parts inventory, and of labor
associated with routine preventive maintenance or emergency maintenance.
Indirect operating costs include overhead, parts replacement, insur-
ance, taxes, and amortization ("capital recovery") of the investment.
Overhead is traditionally expressed as a percentage of operating and main-
tenance labor, whereas the other indirect costs are normally computed as a
percentage of the control device investment.
3-9
-------
REFERENCES
1. U.S. Environmental Protection Agency. Compilation of Air Pollution
Emission Factors. AP-42, 1977.
2. U.S. Environmental Protection Agency. Office of Administration, Infor-
mation Division. Need Air Pollution Information? Promotional Litera-
ture.
3. Industrial Gas Cleaning Institute. IGCI - Who We Are and What We Do.
Promotional Literature. Stamford, Connecticut.
4. Catalog and Buyer's Guide - Pollution Equipment News. Rimbach Pub-
lishing, Inc., Waseca, Minnesota, Published annually.
5. Equipment Buyer's Guide Issue - Chemical Engineering. McGraw-Hill, New
York, N.Y. Published annually.
6. Air Pollution Control Association. Journal of the Air Pollution Con-
trol Association. Pittsburgh, Pennsylvania. Published monthly.
7. Pollution Engineering. Technical Publishing. Barrington, Illinois,
Published monthly.
8. Pollution Equipment News. Rimbach Publishing, Inc., Waseca, Minnesota.
Published bimonthly.
9. U.S. Environmental Protection Agency. Standards of Performance for New
Stationary Sources—A Compilation. EPA-340/1-77-015, October 1977.
10. U.S. Environmental Protection Agency. National Emissions Standards for
Hazardous Air Pollutants—A Compilation as of April 1, 1978. EPA-340/
1-78-008, 1978.
11. U.S. Environmental Protection Agency. Air Pollution Regulations in
State Implementation Plans. EPA-450/3-78-050 through EPA 450/3-78-104,
56 documents, updated annually.
12. U.S. Environmental Protection Agency. Choosing Optimum Management
Strategies—Pollution Control Systems. EPA-625/3-77-008, 1977.
13. U.S. Environmental Protection Agency. Choosing Optimum Financial
Strategy for Pollution Control Investments, EPA-625/3-76-005, 1976.
3-10
-------
14. Neveril, R. B. Capital and Operating Costs for Selected Air Pollution
Control Systems. EPA-450/5-80-002, 1980.
15. U.S. Environmental Protection Agency. Major Financial Assistance Pro-
grams Available for Industrial Pollution Control Expenditures. EPA-
340/1-77-023, 1977.
3-11
-------
-------
SECTION 4
PARTICULATE CONTROL SYSTEMS
Since "Control Techniques for Particulate Air Pollutants1" was origi-
nally published in 1969, there have been substantial advances in the capa-
bilities of particulate matter removal devices. This section presents
information on the many sophisticated and diverse control systems presently
in commercial use. Each of the five major categories of systems is dis-
cussed in terms of types available, basic operating principles, design
factors, and operation and maintenance considerations. Particle collection
capability with respect to particle size is emphasized.
4.1 INTRODUCTION
Introductory information is presented on particle behavior and charac-
teristics. Each particle control device takes advantage of one or more of
the particle aerodynamic properties to remove particulate matter from the
gas streams. Basic considerations in the application of particulate matter
control devices as part of a system are also discussed.
4.1.1 Particle Characteristics and Behavior
It is helpful to understand basic principles of particle behavior in
order to design a control device, to measure particulate matter concentra-
tions, or to evaluate control device performance. More extensive informa-
tion is available in aerosol physics texts such as references 2 through 10.
4.1.1.1 Particle Size and Shape. There are a wide variety of particle
shapes. Spherical particles are usually generated in high temperature
processes and in some cases the particle is partially or completely due to
the condensation of vapors as the gas stream cooled. Small particles can
link with other particles to yield a flocculant. Such flocculants tend to
be fragile and can break apart during sampling or during passage through
control devices. Fibrous particles result from the processing of certain
4.1-1
-------
natural biological and mineral materials. Asbestos is one type of fibrous
particle. Depending on the chemical composition of the parent material, it
is also possible to generate a flake-type aerosol. Mechanical attribution
type processes such as grinding, sawing, and polishing can yield irregular-
ity shaped particles. Such aerosols are generally larger than particles
formed by condensation of vapors and other common types of aerosols.
Liquid aerosols are comprised almost exclusively of spherical particles
or flocculants of spherical particles. In certain processes it is also
possible to generate solid particles covered with an outer layer of liquid
material. Solid aerosols can occur in any of the forms shown and in less
common forms such as cubes and rods. Factors governing the ultimate shape
of the particles include the chemical composition of the material and the
characteristics of the process.
Particle size is the most important characteristic affecting behavior
in gas streams, and it is a governing factor in the extent to which the
particle scatters visible light and therefore contributes to plume opacity.
The particle size range of interest to air pollution control studies if
generally from 0.01 to 100 micrometers. Under the SI system of metric
units, micrometer (pm) is the standard unit of particle size, and is 10 6
meters.
The many definitions of particle size are ultimately based on the size
measurement method. For example, microscopic methods measure the projected
area of particles. The projected area can be variously defined, depending
on the means used to convert the dimensions of irregularly shaped and
fibrous particles into a single "diameter" value. Likewise, sampling
methods, such as cascade impactors can be used to measure particle size
determined by observed behavior in a gas stream. The various measurement
methods yield many size definitions that are not necessarily consistent with
each other.
The most common definition for particulate control evaluation is the
aerodynamic particle diameter, which is defined as the equivalent unit
density sphere having the same aerodynamic characteristics as the actual
particle. It is the product of the Stokes diameter times the square root
of the product of particle density times the Cunningham slip correction
factor:
4.1-2
-------
dp = dps
,0.5
(Eq. 4.1-1)
Throughout this document, particles sizes are given in terms of aerodynamic
diameter (pinA) unless otherwise noted. The aerodynamic diameter is most
closely related to the behavior of particles in control devices, and it is
the "size" normally measured by stack sampling methods.
Each of the particle size measuring methods yields sets of data indi-
cating the quantities of particulate in a number of size categories. These
data can be compiled into a histogram, as shown in Figure 4.1-1 to graph-
ically illustrate the aerosol distribution. The terms used to characterize
the distributions of particles sizes are illustrated in Figures 4.1-2 and
4.1-3.
PARTICLE DIAMETER, urn
Figure 4.1-1. Aerosol distribution.
(Reprinted from: Silverman, L., Billings, C.E.,
and First, M. W. Particle Size Analysis in In-
dustrial Hygiene. Academic Press, 1971, p. 237.)
For example, the mean and the median are not equal for skewed distributions,
as shown in Figure 4.1-1. The median particle size, by definition, divides
the frequency distribution in half; 50 percent of the aerosol mass has
particles with a larger diameter, and 50 percent has particles with a smal-
ler diameter. A second measure of central tendency is the mean, which is
simply the sum of all observations divided by the number of size categories
used to construct the histogram; the mean is sensitive to the quantities of
material at the extremes of the distribution, so relatively few large parti-
cles could shift the observed mean to larger levels.
4.1-3
-------
For many industrial sources, the particle distribution approximates a
lognormal . distribution function. When the log of the particle diameter is
plotted against the frequency of occurrence, a normal bell-shaped curve
(shown in Figure 4.1-2) results, which is characterized by the geometric mean
diametei—the sum of the logs of the observations divided by the number of
size categories.
C_3
as
UJ
=>,
UJ
tse.
1 2 3 4 5 6 7 8 .9 12 14 18 22
PARTICLE DIAMETER, ym
Figure 4.1-2. Histogram of a lognormal size distribution.
(Reprinted by permission from Silverman, L., Billings, C. L.
and First, M. W. Particle Size Analysis in Industrial Hygiene.
Academic Press, 1971, p. 236.)
Both the geometric mean and the standard deviation can be determined
easily by plotting the distribution data as a log-probability plot as shown
in Figure 4.1-3. The geometric mean is the diameter equivalent to the 50
percent probability, and the standard deviation is the slope of the line.
The latter can be determined simply by dividing the geometric mean by the
particle size at the 15.78 percent probability (size d± in Figure 4.1-3) or
by dividing the particle size at the 84.13 percent probability (size d2 in
Figure 4.1-3) by the geometric mean size.
4.1-4
-------
O1
LU
20 -
Mill I
8 10 16
2 3 4567
. PARTICLE DIAMETER, urn
Figure 4.1-3. Cumulative lognormal size distribution.
(Reprinted by permission from Silverman, L. , Billings,
C. E. , and First, M. W. Particle Size Analysis in In-
dustrial Hygiene. Academic Press, 1971, p. 239.)
Aerosol distributions may exhibit more than one peak. The hypothetical
distribution shown in Figure 4.1-4 is called bimodal. This type of aerosol
distribution is more difficult to characterize. In some cases, it may be
possible to handle the distribution as two separate lognormal distribution
aerosols. More detailed discussions of aerosol distributions are presented
in references 4, 11, and 12.
o-
4 " 6 " 8 10 12 14"
PARTICLE DIAMETER, ym
16 18
Figure 4.1-4. Bi-modal aerosol distribution.
4.1-5
-------
4.1.1.2 Aerodynamic Properties. Each type of participate control
device uses one or more particle collection mechanisms. Those are the funda-
mental physical tools available to an equipment designer.
Impaction. Inertia! impaction is the mechanism most frequently used to
remove particulate matter. Particles have a much greater mass, and there-
fore much greater inertia when in motion, than the surrounding gas. Heavy
particles resist changes of gas flow and cross gas streamlines because
of their inertia. As a gas stream approaches an obstacle, the gas molecules
pass on either side of it, leaving the particle propelled toward the obsta-
cle by its inertia. If the particles are too small they flow with the gas
molecules and pass around the obstacle. Figure 4.1-5 illustrates these
phenomena relative to different sized particles. Low-energy impaction
conditions separate particles with aerodynamic diameters above 50 p.mA.
High-energy impaction conditions effect separation of particles with aero-
dynamic diameters above a few tenths of a micrometer. Particles with aero-
dynamic diameters below a few tenths of a micrometer can not be separated by
Inertia! impaction under normal conditions.
Water Droplet
Particle-
, Stream!ines
Figure 4.1-5. Impaction of particles on a target in a moving gas stream.
4.1-6
-------
Efficiency of impaction collectors is related to an impaction parameter:
a ratio of drag to viscous forces.
to relate impaction efficiency.
The Stokes number, K,, is commonly used
(Eq. 4.1-2)
where
I 18 M Dr
%*
Kj = Stokes number.
C = Cunningham slip factor, dimensionless.
d = particle diameter, umA.
p = particle density, g/cm3.
v = particle velocity, cm/s.
D = diameter of collector, cm.
M = gas viscosity, kg/m - s.
The impaction mechanisms become progressively more effective as the impaction
parameter increases. The strong particle size dependence of inertial impac-
tion is indicated by the fact that impaction is proportional to the square of
the particle size.
Particle separation from a gas stream can also occur by a mechanism
known as interception. The particle is intercepted by an obstacle if the
particle radius is as large as or larger than the streamline displacement.
Interception results in an increase over the amount of particle collection
that is predicted by impaction alone. Interception is similar to, and can be
considered a form of impaction. As illustrated in Figure 4.1-6, interception
occurs when particle size and gas streamline displacement values are compa-
rable. Particles in the size range above a few micrometers are susceptible
to interception.
»
-------
Diffusion. Particles in the same size range as molecules (10~3 urn) and
up to a few tenths of a micrometer experience random movement due to colli-
sions with gas molecules. The diffusion rate of a particle (D ) can be
determined by the Stokes-Einstein equation:
D = CKT
p 3?iudp
(Eq. 4.1-3)
where
D = diffusivity of particle, cmVs.
K = Boltzman constant, (g - cm2/s2 - °K).
T = absolute temperature, K.
The effectiveness of this particle collection mechanism is proportional to
the particle diffusivity which is inversely proportional to the particle
size, as indicated above. Thus high diffusion rates occur for very small
particles (0.001 to 0,1 urn), but diffusion is negligible for large particles.
Particle diffusion is illustrated in Figure 4.1-7.
Trajectory
Particle
Gas strearr.l i res
WaterDroplet
Figure 4.1-7. Diffusion of a particle to a target in a moving gas stream.
Settling. Movements of particles in the atmosphere are influenced by
two main forces: gravity and drag. When'particles are settling, gravita-
tional force is pulling downward and drag is pushing upward. After suffi-
cient time, the forces become equilibrated, and the particles reach their
4.1-8
-------
terminal settling velocity. To determine the settling velocity (V.) for
ideal situations, the Stokes terminal velocity equation can be used:
dO (PD " Pa^
V= P 9 (Eq. 4.1-4)
As with inertia! impaction, the effectiveness of settling is proportional to
the square of the particle diameter. This mechanism becomes important in air
pollution control systems only when the particle size is above 5 (jmA and when
the gas flow conditions are not highly turbulent.
El ectrostat 1c Attracti on. Particles are charged with unipolar ions, and
are subjected to a strong electrical field. Movement of particles to a
collection surface is primarily dependent on the balance between the elec-
trostatic force and the aerodynamic resistance. Random diffusion charging
contributes to the initial charging of the particles. Effectiveness of
electrostatic attraction is basically related to the square of the particle
aerodynamic diameter because larger particles can sustain a greater number of
charges.
4.1.1.3 Physical Phenomena. Condensation and agglomeration can alter a
particle size distribution and thereby have a significant impact on the
aerodynamic behavior of the particles. Light scattering is important be-
cause opacity of the stack effluent is sometimes used to evaluate the effec-
tiveness of the particulate control systems.
Condensation. Condensation of water vapor on suspended particles
occurs whenever a degree of supersaturation is available around the parti-
cles. In industrial gas cleaning, three methods are available to effect
particle growth by condensation: (1) mixing produces supersaturation by
combining two saturated gas streams of different temperatures; (2) steam
injection introduces steam into the gas stream, and (3) an adiabatic expan-
sion method, which is available in venturi scrubbers.13 A psychrometric
chart indicates the appropriate temperature levels and amounts of water
vapor available for use in particle growth by condensation.
The main benefit from condensation of water vapor on particles is the
increase in their mass and size, which makes them easier to collect in
inertial removal systems. Collection can be further enhanced by taking
4.1-9
-------
advantage of forces acting on the particles induced by a temperature gradi-
ent (thermophoresis) and vapor condensation (Stefan flow).14 Superimposition
of these effects can improve collection.
4.1.1.4 Agglomeration. When particles collide with each other during
diffusion or turbulent motion they may adhere to each other and become
combined or agglomerated particles. The rate of agglomeration depends
mainly on particle concentration and is virtually independent of particle
size, as shown by the following differential equation:
- — = kN2 (Eq. 4.1-5)
dt
where
k = rate constant.
N = number of particles per unit volume at time t.
Integration of Equation 4.1-5 from time zero to time t for an initial parti-
cle concentration (N ) is:
i - i = kt (Eq. 4.1-6)
o
Agglomeration rate constant values (k) can be estimated for homogeneous and
heterogeneous systems by the following:
homogeneous system: k = ~ (Eq. 4.1-7)
2KTC
heterogeneous system: k = —— (Eq. 4.1-8)
for reference purposes, the homogeneous rate constant for air at 20°C is 3 x
10"10 cms/s. Values for agglomeration rate constants are also influenced by
pressure and turbulence conditions. Increased pressure and greater turbu-
lence separately enhance the probability of particle collision, and thus
increase the agglomeration rates.7
4.1.1.5 Light Scattering. When white light strikes a suspended parti-
cle, a certain amount of light (depending on particle size, shape, composi-
tion, and surface configuration) is scattered irregularly in all directions.
Particles scatter light in a degree proportional to their sizes. Supermi-
crometer particles scatter light proportional to diameters to the sixth
power.8
4.1-10
-------
4.1.2 Selection and Application of Particulate Control Devices
Each application of a participate control system is unique to a degree.
No general selection method can guarantee an environmentally and economical-
ly acceptable installation. Instead, it is necessary to carefully match
process effluent characteristics with regulatory requirements of control
device performance capabilities and control system costs. This section
introduces some of the general issues involved in the selection and applica-
tion of parti cul ate control systems. More detailed information is presented
in Sections 4.2 through 4.6.
4.1.2.1 Performance Capabilities. Most particulate control devices
operate as combinations of the particle collection mechanisms discussed in
Section 4.1.1. One or more mechanisms may be operative in any given device;
the limitations of such mechanisms become the performance limits of the
device. Table 4.1-1 presents the mechanisms normally active in the major
types of particulate control devices. The effectiveness of the mechanisms
depends on the design of the unit (e.g., gas velocity, device geometry,
liquid utilization rate). Performance is typically characterized in terms of
collection efficiency:
inlet mass outlet mass
The penetration fraction (P+), defined in Equation 4.1-10, is a somewhat
simpler measure of control device performance; and it is related to collec-
tion efficiency as indicated in Equation 4.1-10.
Penetration = outlet mass loading = ^ff^^/wQ (Eq. 4.1.10)
inlet mass loading J H
The penetration term is easier to use when evaluating high efficiency con-
trol devices and series of particulate control devices, and when using the
computerized models developed for certain types of collectors. Due to the
size dependence of the particle collection mechanisms the collection effici-
ency (or penetration) is a size-specific value. A set of values for various
particle sizes is a penetration curve.
4.1.3 Control System Design
Proper selection of particulate control systems requires simultaneous
consideration of regulatory requirements, performance limits, effluent
4.1-11
-------
TABLE 4.-1-1. PARTICLE CAPTURE MECHANISMS NORMALLY ACTIVE IN CONVENTIONAL
PARTICIPATE CONTROL DEVICES
Control device
Principal particle
capture mechanism
Particle size
dependence
Settling chamber
Momentum separator
Large-diameter single
cyclone
Small-diameter multiple
cyclones
Fabric filters
Electrostatic precipitator
Wet scrubber
Incinerator
gravity settling
gravity settling
inertia! separation
inertia! separation
inertia! separation
impaction on dry surfaces
interception
diffusion to dry surfaces
electrostatic attraction
gravity settling
impaction on surfaces
impaction on liquid
droplets
diffusion to wetted surfaces
diffusion to liquid droplets
particle oxidation
V
V
V
V
V
p
1/dp
d 2 and
V
V
V
1/d.
Based on particle capture mechanism.
characteristics, and cost. The control device must have the capability to
maintain continuous compliance regardless of short-term fluctuations in the
effluent composition, flow rate, and particle size distribution.
Control Device Sizing. Most particulate control devices (for example,
baghouses and electrostatic precipitators can suffer performance degradation
4.1-12
-------
at high effluent gas stream velocities. A fundamental design problem is the
sizing of the control device to balance the need for a large unit (low gas
velocity) with the capital cost. The consequences of undersizing (high gas
velocities) can be frequent noncompliance periods. The sizing of a control
device should also accommodate anticipated operational changes; for example
increases in process throughput and/or effluent gas temperature can lead to
inadequate efficiency. A related problem is failure to properly size the
support systems, such as solids removal equipment, by inadequately assessing
the effluent conditions such as particulate mass loadings or bulk density.
Instrumentation. Control devices installed without proper instrumenta-
tion may be prone to frequent malfunctions and excessive emission periods.
Examples are a fabric filter on a high-temperature source without tempera-
ture monitors and a wet scrubber on a combustion source without pH monitors.
Proper instrumentation is necessary to provide an early warning of impending
problems and to assist in diagnosis of underlying factors.
Accessibility. No particulate control system is completely mainte-
nance-free. Control devices are normally subjected to multiple physical
insults including abrasion, corrosion, mechanical shocks, moisture, high
temperature, and high voltage. The designer and purchaser must decide what
additional capital cost is justifiable to minimize future maintenance costt
Power Input - The collection efficiency of a particulate control system
is generally associated with power input. As power input increases the
particulate emissions usually decrease. What emission level is affordable
and justifiable, given the energy demand? This question is complicated by
uncertainty over actual effluent gas stream characteristics (e.g., particle
size distribution) and the lack of site specific empirical models relating
power input to collection efficiency.
Corrosion - Catastrophic failure due to corrosion can result from
inaccurate assessment of gas stream conditions during steady-state or start-
up conditions. A contributing factor can be failure to consider the varia-
bility of effluent vapor concentrations caused by variability of process
operations. To the degree economically feasible, particulate control equip-
ment should be designed and fabricated to withstand worst-case conditions.
4.1-13
-------
Abrasion - Large particles suspended in a fast moving gas stream are
abrasive. Fabric filters are particularly vulnerable to abrasion near the
gas inlets. Precleaners can be installed, but usually increase both capital
cost and energy demand (increased fan energy).
Moisture and Freezing - Moisture and freezing can adversely affect wet
scrubbers and any control device using compressed air (pulse-jet fabric
filters and electrostatic precipitators with compressed air rappers). The
solution to moisture problems in dry collectors are the inclusion of dryers
on air compressors and drains on air reserve tanks. Wet scrubber lines
should have drainage capability, particularly when operation is noncontinu-
ous.
Ventilation Systems - The two basic parts of a ventilation system are
the hood or air intake for initial capture of the particulate matter and the
ductwork for transport of the particulate-laden gas stream to the control
device. Inadequate design of a ventilation system can compromise overall
performance.
The hood must be sized and oriented to capture the maximum quantity of
particles without requiring excessive gas volumes (a trade-off between
performance and energy consumption). It makes little sense to install a
high-efficiency control device if a major portion of the particles are not
captured initially. The hood should be as close as possible to the point of
generation without interfering with equipment movement; it should be ori-
ented to minimize cross-drafts and to take advantage of thermal drafts.
The ductwork leading from the hood (or pickup point) to the control
device must be sized to provide the needed transport velocity—generally
between 15 and 25 m/s, depending on particle size distribution.15 Layout of
ductwork should minimize energy losses indicated by static pressure drops,
and should minimize air inleakage. If the source is hot, insulation may
minimize temperature drops.
Solids Removal Equipment - The basic functions of the solids removal
equipment are to remove collected particulate from the device as rapidly as
it is collected and to deliver the material to an environmentally acceptable
disposal area. Solids removal is one of the most frequent problem areas
affecting particulate control equipment.
4.1-14
-------
Most particulate systems operated at elevated temperature must use
hopper heaters, insulation, or weather enclosures, or combinations of these
to keep the collected particulate hot to maintain free flow. Hopper temper-
ature control can also reduce corrosion resulting from condensation.
Delivery of the solids to a disposal site or to a temporary storage
site must be done without resuspension of the material. As with inadequate
hood capture, a seemingly small degree of resuspension can compromise any
gains achieved by a high-efficiency collector. This often happens when
solids are discharged with a significant free fall (directly below a dis-
charge valve or between two conveyors) or when a temporary storage pile is
unprotected from winds.
Fans - Movement of particulate-laden gas stream from the process,
through the control device, and out the stack is controlled by the fan. Fan
selection is critical to proper operation of the overall system. Although
the forward curved design has high efficiency, it is vulnerable to partic-
ulate buildup on the blades so it is very rarely used in parti cul ate control
systems. The backward-curved design also has relatively high energy effi-
ciency, and it is also susceptible to particulate buildup. This type of fan
is normally used only on the "clean side" of the particulate control device
to provide induced draft. The most rugged type is the radial blade design
fan which can withstand high dust loadings without excessive vibration and
therefore, can be used on either the clean or the dirty side.
There can be fan problems with fans initially selected properly. They
must provide adequate gas flow and static pressure despite nonideal ventila-
tion system design or fan inlet configuration. Air inleakage can lead to
lower-than-expected gas stream temperatures and hence to higher horsepower.
In summary, the design of a control device involves the balancing of
numerous factors. The design decisions must be made specifically for the
source. The ultimate success of the control system depends at least par-
tially on a realistic evaluation of the characteristics of the effluent gas
stream and the entrained particulate matter distribution.
4.1.3.1 System Operation and Maintenance. Continuous compliance with
air pollution control regulations depends largely on proper operation and
4.1-15
-------
maintenance such as preventive maintenance programs, recordkeeping pro-
cedures, and operator training programs. An important aspect of training
is full awareness of potential safety problems.
Prevention Maintenance - An integral part of a preventive maintenance
program is routine inspection of equipment. Depending on the potential for
malfunctions and the consequences of excess emissions (toxicity, quantity),
inspection could be done daily, weekly, or monthly. For most equipment, the
internal and external conditions of the equipment must be noted during
inspection.
An adequate spare parts inventory should be maintained to prevent
excessive downtime or excessive emissions while operating under nonoptimal
conditions. Determination of what is necessary depends partially on the
local availability of supplies and on the costs of parts relative to the
cost of equipment nonavailability.
Recordkeeping - The logical first step in recordkeeping is to make sure
that instruments are properly located and are functioning normally. Most
instruments, even supposedly simple devices, require calibration. There is
little sense in faithfully keeping records that are incorrect or misleading.
Only the data that must be evaluated to determine developing problems
or nonoptimal conditions need to be recorded. If large quantities of un-
necessary data are logged, the meaningful data may be lost.
In addition to the normal operating records, diagnostic records should
be recorded during forced outages or routine maintenance periods to indicate
the type and location of component failures (e.g., bag failure location, dis-
charge wire breakage type and location). Such diagnostic records can be as
simple as copies of repair work orders with comments by maintenance per-
sonnel .
Training - Particulate control systems are complex and expensive.
Operator training is helpful to ensure proper and safe operation. Training
should emphasize procedures for startup and shutdown to minimize damage to
the unit. Early signs of developing problems should be stressed.
The importance of safety training for operating and maintenance person-
nel cannot be overemphasized. A particulate control system may represent a
combination of a very large number of potential hazards, including oxygen
4.1-16
-------
deficiencies, toxic gases, high temperatures, high voltages, hot dust (hop-
pers), high noise levels, and moving machinery. As a minimum, training
should address confined space entry procedures, selection and use of protec-
tive equipment, and safe work practices.
4.1-17
-------
REFERENCES
1. U.S. Department of Health, Education, and Welfare. National Air Pollu-
tion Control Administration. Control Techniques for Particulate Air
Pollutants. AP-51, January 1963.
2. Fuchs, N. A. Mechanics of Aerosols. Pergamon Press, London, 1964.
3. Davies, C. N. Air Filtration. Academic Press, Inc., New York, 1973.
4. Silverman, L., C. E. Billings, and M. W. First. Particle Size Analysis
in Industrial Hygiene. Academic Press, Inc., New York, 1971.
5. Green, H. L., and W. R. Lane. Particulate Clouds, Dusts, Smokes, and
Mists (2d ed.). Van Nostrand, Princeton, New Jersey, 1964.
6. Davies, C. N. Aerosol Science. Academic Press, Inc., 1964.
7. Hesketh, H. E. Understanding and Controlling Air Pollution. Ann Arbor
Science Publishers, Inc., Ann Arbor, Michigan, 1974.
8. Hesketh, H. E. Fine Particles in Gaseous Media. Ann Arbor Science
Publishers, Inc., Ann Arbor, Michigan, 1977.
9. Drinker, P., and T. Hatch. Industrial Dusts (2d ed.). McGraw-Hill,
New York, 1959.
10. Hidy, G. M., and J. R. Brock. The Dynamics of Aerocolloidal Systems,
Vol. 1. Pergamon Press, London, 1970.
11. Yeshida, T., et al. Particle Growth by Condensation and Coagulation—
Basic Research of its Application to Dust Collection. In; Symposium
on the Transfer and Utilization of Particulate Control Technology,
Vol. I. EPA-600/7-77-004d, February 1979.
12. Calvert, S., J. Goldshmid, D. Leith, and N. Jhaveri. Feasibility of
Flux Force/Condensation Scrubbing for Fine Particle Collection. EPA-
650/2-73-036, October 1973.
4.1-18
-------
4.2 MECHANICAL COLLECTORS
Mechanical collectors comprise a broad class of participate control
devices that utilize the gravity settling, inertial, and dry impaction
mechanisms. Because the;ir performance capability is limited to relatively
large particles and regulatory requirements have become more stringent, use
of mechanical collectors has gradually declined and they are now used pri-
marily as precleaners. Mechanical collectors are reasonably tolerant of
high dust loadings, are not susceptible to frequent malfunction if properly
designed and operated, and are adequate control devices for some applica-
tions.
There is great diversity in the design and operating principles of the
various types of mechanical collectors. Most penetration performance data
for mechanical collectors were obtained from 1940 through 1970. Since 1970
attention has shifted to more sophisticated particulate control devices.
Consequently, only limited field data are available.
4.2,1 Types of Mechanical Collectors
The major classes of commercially available mechanical collectors are
listed in Table 4.2-1.
TABLE 4.2-1. MAJOR TYPES OF MECHANICAL COLLECTORS
Type
Particle capture mechanism
Sett!i ng chamber
Elutriator
Momentum separator
Mechanically aided collector
Inertial centrifugal collector
gravity settling
gravity settling
gravity settling, inertial
collection
inertial collection
inertial collection
4.2-1
-------
The general characteristics of these devices are described in the remainder
of this subsection. Later subsections address operating principles, design,
operation, and performance.
4.2.1.1 Sett!ing Chambers. Large particulate is removed by gravita-
tional settling in settling chambers to protect downstream equipment from
abrasion and excessive mass loadings.1
The two basic types are the simple expansion chamber and the multiple-
tray settling chamber (Figure 4.2-1). The latter is a set of horizontal
collection plates that reduce the distance a particle must fall to reach the
collecting surface.1'2 Thus the multiple-tray unit can collect somewhat
smaller particles, which settle more slowly.2
GAS OUTLET
GAS INLET
Figure 4.2-1. Howard multi-tray settling chamber.
Reprinted with permission of McGraw-Hill Book Company.
John H. Perry, Chemical Engineers Handbook, 3rd edition. 1950
The settling chambers should be designed for low velocities with a
minimum of turbulence so that the settling of particles is not disturbed.
Typical superficial velocities range from 0'.3 to 3 m/s.3'4 Gas stream
distribution across the chamber inlet is important.
4.2-2
-------
4.2.1.2 Elutriators. An elutriator consists of one or more vertical
tubes or towers in series into which a dust-laden gas stream passes upward
at a velocity defined by the gas flow rate and the tube cross-sectional
area.
Large particles with terminal settling velocities greater than the
upward gas velocity are separated and collected at the bottom of the cham-
ber. Smaller particles with lower settling velocity are carried out of the
collector. The particle size collected may be varied by changing the gas
velocity.
When size classification is desired for disposal or reintroduction into
a process, a series of collectors may be used with increasing cross-sec-
tional area. Typical uses of elutriators are in secondary metal operations,
food and agricultural processes, and petrochemical industries.
4.2.1.3 Momentum Separators. The momentum separator uses a combina-
tion of gravity and particle inertia (momentum) to settle particles onto
surfaces. The particles are separated from the moving gas stream by pro-
viding a sharp change in direction of gas flow so that momentum carries the
particles across the gas streamlines and into the hopper.
The simplest versions provide a 90- to 180-degree turn to separate
large particles5'6 (see Figure 4.2-2, a and b). Baffles can be added to
increase the number of turns and thereby provide a modest increase in
collection (see Figure 4.2-2c).
Figure 4.2-2. a. Simple momentum separator.
Reprinted from: Alden, J. L. Design of In-
dustrial Exhaust Systems, 3rd Ed., The Indus-
trial Press, New York, 1959.
4.2-3
-------
Figure 4,2-2. b. Simple momentum separator.
Reprinted from: R. F. Jennings, J. Iron Steel
Inst. Vol. 164, page 305, 1950.
Figure 4.2-2. c. Baffle-type momentum separator.
Reprinted from: L. Theodore and D. W, Buonicore. In-
dustrial Air Pollution Control Equipment for Partic-
ulates, CRC Press, Inc., page 66, 1976.,
The louver collector, a type of momentum collector shown in Figure
4.2-3, consists of a series of flat plates (blades) set at an angle to the
gas stream. A large portion of the gas stream is required to make a sharp
turn to pass through the plates. The momentum of the particles in the air
stream results in movement of the particles in a path parallel to the louver
surface and across the gas stream. Separation of the particles from the gas
stream leads to concentration of the larger particles in a small portion of
4.2-4
-------
the gas stream.7'8 Penetration is a function of louver spacing and gas
volume.
SHUTTERS
PARTLY CLEANED
GAS TO MAIN
CONTROL DEVICE
COLLECTED DUST CARRIED OFF BY
10 PERCENT BLEED STREAM
Figure 4.2-3. Louvered shutter type collector.
Reprinted from: Stairmand, C. J. Trans. Insti.
Chem Engro - Vol. 29, page 356, 1951.
4.2.1.4 Mechanically Aided Separators. The separation mechanism of
mechanically aided separators, like that of momentum collectors, is inertia.
Mechanical acceleration of the effluent gas stream increases the effective-
ness of the inertia separation so that these devices can collect smaller
particles than the momentum devices. The improved performance, however, is
gained at the expense of higher energy cost. Also, the devices are subject
to abrasion by the action of large-diameter particles at medium to high
velocities.
4.2-5
-------
The most common of the mechanically aided collectors is a modified
radial blade fan. The dust-laden air enters the collector perpendicular to
the blade rotation, and by momentum the particles cross the air stream and
concentrate at the side of the collector casing. The rapid acceleration of
the gas stream imparted by the mechanical rotation of the blades maintains
the concentrated particles in a narrow band which is then drawn off for
particle separation in a more efficient collector.
Many collectors use this design principle, and many variations of this
method are used to concentrate the particles into a smaller gas volume. In
addition to modifications, scroll collectors and skimmers are also used.
4.2.1.5 Cyclones. The cyclone collector is similar to the momentum
collector in that inertia is used to separate the particles from a turning
gas stream. In the cyclone the gas stream makes one or more circular turns,
followed by a 180-degree turn to the outlet duct. Combined effects of a
greater number of turns and higher gas velocity improve the particle collec-
tion capability above that of momentum collectors.
Cyclones can be classified into four basic categories according to the
methods used to remove the collected dust and to introduce the gas stream
into the unit.9 Figure 4.2-4 illustrates the four types of cyclone collec-
tors.
A vortex is created within the cylindical section of the cyclone by
either injecting the gas stream tangentially or by passing the gas stream
through a set of spin vanes. Because of inertia, the particles migrate
across the vortex gas streamlines and concentrate near the cyclone walls.
Near the bottom of the cyclone cylinder the gas stream makes a 180-degree
turn, and the particles are discharged either downward or tangentially into
hoppers below. The treated gas passes upward and out of the cyclone.
Simple cyclones. The simple cyclone consists of an inlet, cylindrical
section, conical section, gas outlet tube, and dust outlet tube. A typical
tangential-inlet, axial-outlet simple cyclone is shown in Figure 4.2-5.
Particle separation is a function of the gas throughput and the cyclone
cylindrical diameter. Particle inertia increases with increases in gas flow
4.2-6
-------
TOP VIEW
SIDE VIEW
TOP
VIEW
a. Tangential inlet, axial dust b. Tangential inlet, peripheral dust
outlet outlet
fTOP VIEW
SIDE VIEW
ii
TOP
VIEW
SIDE
VIEW
c. Axial inlet, axial dust outlet d. Axial inlet, peripheral dust dis-
charge
Figure 4,2-4. General types of cyclones.
Adapted from: Caplan, Air Pollution, A. Stern, Editor, Academic Press, 1968.
4.2-7
-------
GAS
IN-*
GAS
INLET
r
-------
Medium-efficiency single cyclones are usually less than 4 m in diameter
and operate at static pressure drops of 0,50 to 1.50 kPa.4 Overall collec-
tion efficiency is a function of the inlet particle size distribution.
Axial—inlet, axial-discharge cyclones. One common type of axial-inlet,
axial-discharge cyclone is called the double-vortex cyclone. The collector
consists of an air inlet at the bottom of a long cylinder with stationary
turning vanes. The gas stream is placed in a vortex flow pattern upward
through the cylinder. A secondary vortex is generated by either stationary
vanes or injection nozzles outside of the inner vortex moving in a downward
motion. The particles in the inner vortex move across the stream lines and
into the downward-moving outer vortex. The concentrated particles are
separated at the bottom of the unit as the outer vortex changes direction.
The flow pattern is illustrated in Figure 4.2-6.
Intel
Secondary air pressure
maintains hiflh
centrifugal action
i^»c Secondary air flow
2 ^^creates downward spiral
?of dust and protects
/ outer walls from abrasion
*,Dust is separated from
"^ gas by centrifugal force,
is thrown toward outer wall
and into downward spiral
, Falling dust is deposited
in hopper
Figure 4.2-6. Flow pattern in a double-vortex cyclone.
Courtesy of Aerodyne Development Corporation.
4.2-9
-------
The static pressure drop of the collector is between 3 to 6 kPa.
Reported separation efficiencies are >99 percent for particles >6 prnA and
>95 percent for particles >1 pmA.10
Multiple axial-inlet, axial-outlet cyclones. A multiple cyclone
consists of numerous small-diameter cyclones operating in parallel. The
high-efficiency advantage of small-diameter tubes is obtained without
sacrificing the ability to treat large effluent volumes. A typical unit is
shown in Figure 4.2-7a.
The individual cyclones, with diameters ranging from 15 to 60 cm,
operate at pressure drops from 0.5 to 1;5 kPa. The inlet to the collection
tubes is axial, and a common inlet and outlet manifold is used to direct the
gas flow to a number of parallel tubes. A single tube from a typical multi-
ple cyclone is illustrated in Figure 4.2-7b. The number of tubes per col-
lector may range from 9 to 200 and is limited only by space available and by
the ability to provide equal distribution of the gas stream to each tube.
Properly designed units can be constructed and operated with a collection
efficiency of 90 percent for particles in the 5 to 10 umA range.4
Variation in performance is achieved by the use of several multicyclone
banks in series or by the withdrawal of 10 to 20 percent of the gas.
Multiple axial-inlet, peripheral-discharge cyclones. The axial-inlet/
peripheral-discharge cyclone is a variation of the multicyclone collector in
that the gas is placed in a vortex motion by a fixed vane and the dust is
concentrated at the collector tube wall by inertia! force. The central core
of the gas stream is relatively free of particles and is allowed to exit the
collector tube axially. The outer portion of the gas stream (vortex) is
withdrawn by an induced-draft fan for removal of the concentrated particles
by a more efficient collector. The secondary collector system is operated
at a lower static pressure than the main air flow to reduce reentrainment of
the concentrated dust (Figure 4.2-8a).
A number of fixed impeller tubes are arranged in parallel and have a
common inlet and outlet manifold. The secondary gas stream is passed
through a settling chamber, then the concentrated particles are removed by a
small high-efficiency cyclone. The secondary gas volume typically con-
stitutes 10 to 20 percent of the primary gas stream. High collection
efficiency may be achieved by using the collector tube banks in series and
parallel arrangements to provide optimum flow volumes in each tube (Figure
4.2-8b).
4.2-10
-------
b. Individual tube
from multicyclone
collector.
a. Typical multicyclone collector.
Figure 4.2-7. Multicyclone collector.
Figure a reprinted from: Joy Manufacturing Co. and Figure b reprinted
from: Howden, James & Co. Ltd., 195 Scotland Street, Glasgow; C-5
4.2-11
-------
efficiency may be achieved by using the collector tube banks in series and
parallel arrangements to provide optimum flow volumes in each tube (Figure
4.2-8b).
Figure 4.2-8a. Fixed-impeller straight-through cyclone.
Annuior du»t tlols
i—j .
Figure 4.2-8b. Bank of fixed-impeller straight-through cyclones with
secondary cyclone dust collector.
Reprinted with permission of Davidson and Co. Ltd.. Belfast Publica-
cation, Ref. No. 387/61.
4.2-12
-------
4.2.2 Operating Principles of Mechanical Collectors
Fundamental operating principles of the various mechanical collector
designs are discussed in this subsection, with emphasis on theoretical
aspects of penetration and pressure drop. Information concerning these
specific collector types can be transferred with a reasonable degree of
confidence to other mechanical collectors of similar geometric configuration
with similar particle capture mechanisms.
4.2.2.1 Penetration.
Settling chambers. The following equations are a condensed summary of
the approach presented by Theodore and Buonicore.4 A similar approach is
described in Crawford.1 The performance limitations of settling chambers
are examined by determining the fraction of particles of a given size that
will be collected during the gas "treatment" time, tR, defined in Equation
4.2-1.
(Eq. 4.2-1)
where tR = residence time of gas stream in the settling chamber, s
B = chamber width, m
L = chamber length, m
H = chamber height, m
Q = gas flow rate, ms/s
The vertical distance, h, through which a particle of specified aerodynamic
diameter, d., will settle in this time period is simply tD times the
1 K
terminal settling velocity, V,, of that particle.
h = Vt X tR (Eq. 4.2-2)
where V. = terminal settling velocity of particle size i, m/s
The terminal settling velocity is calculated from Equation 4.2-3.
2
S P d.
L- (Eq. 4.2-3)
4.2-13
-------
where g = acceleration of gravity, 9.806 m/s2
p = density of particle, kg/in3
d. = aerodynamic diameter of particle,
u = gas viscosity, kg/m-s
Equation 4.2-3 is reasonably accurate for particles with Reynold
Numbers <1, which in this case is equivalent to aerodynamic particle diam-
eters, £80 pmA. Note that settling velocity is proportional to the aero-
dynamic particle diameter' squared and inversely proportional to the gas
viscosity.
The ratio of h/H represents the fraction of particles of a given size
that will be collected.
V. x tR VtBL
n/H = I 1 = -S— (Eq. 4.2-4)
(q x t^)/BL q
Penetration, P., is simply 1 minus the fractional efficiency.
V+BL
P1 = 1 - h/H = 1 - -=g- (Eq. 4.2-5)
The total penetration is the sum of the penetrations for the various
particle size increments. This simple approach is reasonably valid for
particles in the diameter range of 1 to 80 pmA, which is the range normally
of interest. To extend this type analysis to a wider particle size range,
see Reference 4. .
Equation 4.2-4 is based on the assumption of laminar flow throughout
the chamber. Turbulent conditions will lead to higher penetration (lower
efficiency) than predicted, especially for the smaller particle sizes.
As the gas velocity increases, the limiting velocity at which deposited
particles may be reentrained into the gas stream may be exceeded. A theore-
tical equation for pickup velocity, V , is:
[4 gd, (p.- p)l°'5
Vp = [ k^ CEq" 4"2'6)
4.2-14
-------
where V = pickup velocity, m/s
g = acceleration of gravity, m/s2
d. = particles diameter, i, m
p = particle density, kg/m3
p = gas density, kg/m3
Figure 4.2-9 shows a typical penetration curve for a settling chamber
and Figure 4.2-10 shows a penetration curve based on dust measurements at a
sinter plant.
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0 10 20 30 40 50 60 70 80 90 100
PARTICLE SIZE, ym
Figure 4.2-9. Typical size efficiency curve for settling chamber.
Adapted from data presented in Theodore and Buonicore, page 87.
4.2-15
-------
QUARTZ,
SPECIFIC
GRAVITY
2.6
IRON OXIDE;
SPECIFIC
GRAVITY
4.5
SIMPLE SETTLING CHAMBER
HEIGHT 3.05 m
WIDTH 3.05 m
VOLUME 950 m3
0 10 20 30 40 50 60 70 80 90 100
PARTICLE DIAMETER, urn
Figure 4.2-10.
Penetration of dust through a settling chamber
serving a sinter plant.
Adapted from data presented in Jennings, R. F.,
J. Iron Steel Inst. (London) Vol. 164,
page 305, 1950.
Momentum separators. Particles are removed from a gas stream by use of
the inertia of particles. A gas stream containing the particles is required
to make a sharp change in direction.
The separation in most devices of this type is a combination of momen-
tum mechanisms and particle settling mechanisms. The particles penetrate
through the gas stream and are settled by gravity in hoppers. Collection
may be by impingement on impaction surfaces or by use of secondary inertia
separation to remove the concentrated particles (Figure 4.2-11).
A penetration curve, adapted from Reference 7, for a commercially
available momentum separator is provided in Figure 4.2-12.
4.2-16
-------
CONVERGING
PASSAGE
INLET::
CIRCULATING
FLOW RE-ENTRY WALL
CLEANING WALL
OUTLET
CHAMBER
1- CLEANED
-DIRECT FLOW It FLUE GAS
-DUST SLOT
HOPPER
COLLECTED FLY ASH
FLY ASH OUTLET
Figure 4.2-11. Momentum separator.
Reprinted from: Jackson, R., British Coal Utilization Research Association
Bulletin, Vol. 24, p.'221, 1962.
0 10 20 30 40 50 60-70 80 90 100110120130140
FLY ASH PARTICLE DIAHETER, ymA
Figure 4.2-12. Penetration of fly ash through two momentum separators.
Adapted from data presented by Strauss, page 214.
Industrial Gas Cleaning., Pergamon Press, 1975.
4.2-17
-------
Mechanically aidedseparators. The theory of collection for mechani-
cally aided separators is similar to that used in momentum separators. The
velocity of the gas stream and the turn made by the gas stream are generated
mechanically in a fan or other device. The high radial tip velocity of the
moving mechanical surface and the momentum generated by the increased gas
velocity move the particles to the perimeter of the moving surface. The
particle concentration is increased in the outer portion of the fan and
separated from the main gas flow by a secondary gas flow. The normal
secondary flow is between 10 and 20 percent of the total system volume. Two
types of mechanically aided separators are shown in Figure 4.2-13a.
Aiternotive » i
gas outlet "JUT. [.__.
Ouilj 901 // .
to t«co«»arjf //' /
col IK lor
Cfeoned gas
K> HOC*
• Inlet
cutlet
Primary
dust
separate*
Dust outlet
TGos in!*'
SCROLL TYPE
COMPOUND SCROLL TYPE
Figure 4.2-13a. Types of mechanically aided separators.
Scroll-type collector drawing reprinted from: Stairmond,
C. J. and Kelsey, R. H., Chemistry and Industry, pp. 1324,
1955. Compound Scroll-type drawing reprinted by permission
of Buell Ltd., George St. Parade, Birmingham B3 1QQ
The theory of collection is not developed for collectors of this type,
and the penetration curves are developed empirically. Figure 4.2-13b shows
penetration curves for the scroll- and compound-scroll type collectors. The
secondary collector may be a conventional centrifugal collector such as a
cyclone.
4.2-18
-------
COMPOUND SCROLL
COLLECTOR
"0 20 40 60 80 100 120 140
AERODYNAMIC PARTICLE DIAMETER,
Figure 4.2-13b. Penetration curves for mechanically aided separators.
Reprinted from: Strauss, Industrial Gas Cleaning,
Pergamon Press, page 265, 1975.
Cyclonic separators. The performance of cyclonic separators is
dependent on particle size. Theoretical relationships have been based on
two different parameters, the critical particle diameter and the 50 percent
cut size diameter. The former is the smallest particle size collected with
100% collection efficiency (penetration of zero). Summaries of a number of
these equations are presented by Calvert et a!.11 and Strauss,7 Such
equations provide only a general indication of whether a cyclone will be
adequate for a specific application.
Theoretical models based on the 50 percent cut size have been developed
by Lapple,12 Leith and Licht,13 and Kalen and Zenz.14 Differences in the
equations result primarily from the assumptions used to explain particle
behavior with the cyclone. These and other approaches are described in
detail in Strauss,7 Crawford,1 Licht,15 and Theodore and Buonicore.4
The Leith and Licht model is based on the general concept of radial
mixing within the cyclone cylinder due to a combination of turbulent mixing
and particle bounce.
4.2-19
-------
The dependence of the particle size-specific penetration, P., on
cyclone design parameters and gas stream characteristics is presented in
Equation 4.2-7, which was developed by Licht.1S
-2
P. =e
~v~ 2n + 2 "
KQpp (n+1)
L D 18M
• 1 ~
, n+1
1
(Eq. 4.2-7}
Parameters used in Equation 4.2-7 are defined in Equations 4.2-8 to 4.2-13.
The factor, n, is a temperature-dependent parameter originally presented by
Alexander.16
n = 1 - (T/283)°'3[l - 0.67 (D)0'14]
8[(V + 0.5V.J/D3]
«. ri- .- ..........
2 [B/D]2
Vs = J [S - 0.5a][D2 - D
= f)2(n-S) + JD«[l + f
d = D - (-
$, = 2.3 De (D2/ab)
0'33
where
S = height of outlet tube extension, m
a - height of inlet duct, m
D = diameter of outlet tube, m
D = diameter of cyclone cylinder, m
h = height of cyclone cylinder, m
8, ~ "natural length," m
H = height of cyclone and cone, m
B = diameter of cyclone cone outlet, m
b = width of cyclone inlet, m
p = particle density, kg/m3
Q = gas flow rate, ms/s
(Eq. 4.2-8)
(Fa 4 2-91
^ q* J
(Eq. 4.2-10)
(Eq. 4.2-11)
(Eq. 4.2-12)
(Eq. 4.2-13)
4.2-20
-------
M = gas viscosity, kg/m-s
P.. = penetration of a particle having an aerodynamic diameter of i,
dimension!ess
The Leith and Licht equations give a fractional penetration curve of
the type shown in Figure 4.2-14. The cyclone dimensions and gas flow
characteristics used in the calculation of this curve are presented in
Reference 15.
Lapple's cyclone performance model involves the calculation of the
particle cut diameter, d50, using Equations 4.2-14 and 4.2-15
0.9
0.8
0.7
0.6
o
I—I
S 0.5
LU
°- 0.4
0.3
0.2
0.1
T I
I I
3456
PARTICLE DIAMETER,
8
Figure 4.2-14. Penetration curve predicted by Leith and Licht approach.
Adapted from: D. Leith and W. Licht, American Institute of
Chemical Engineers Symposia Series 68, 1972
4.2-21
-------
dso = 0.308 [(9
« - (V/Q).
)/2nNtv.(p -pG)]
°'5 (Eq. 4.2-14)
\j v* is I p \J
t TTU^ ^ 4'2~15)
where dso = particle diameter collected with 50% efficiency, umA
Up = gas viscosity, kg/m-s
B = width of gas inlet, m
v. = inlet gas velocity, m/s
p = particle density, kg/m3
pg = gas density, kg/m3
Nt = effective number of gas turns, dimension!ess
V = volume of cyclone, m3
Q = gas flow rate, m3/s
D = diameter of cyclone cylinder, m
Calculation of the volume of the cyclone and Equations 14 and 15 yield the
dgo- This is a useful parameter for comparing various cyclones. If a
fractional penetration curve is desired, the dso can be used in conjunction
with a generalized curve as shown in Figure 4.2-15 from Theodore and
Buonicore.4 A more detailed discussion of this approach is provided in the
latter reference. Information concerning theoretical performance models is
presented in Licht,15 Theodore and Buonicore,4 Kalen and Zenz,14 and
Theodore and DePaola.17 The general relationship between particle penetra-
tion and cyclone parameters is summarized in Table 4.2-2.
0.9
0.8
0.7
0.6
0.5
0.4
0.3
_"S^ I III!!
O.I
I I
I I I I 1 I L.
0.3 0.4 O.bQ.V
PARTICLE SIZE RATIO, (d|)/cipc)
Figure 4.2-15. Cyclone penetration as a function of particle size ratio.
Adapted from: Theodore and Buonicore, Industrial Air Pollution Control
Equipment for Particulates, CRC Press, 1976.
4.2-22
-------
TABLE 4.2-2, EFFECTS OF OPERATING CONDITIONS ON CYCLONE PERFORMANCE
Variable
Gas flow rate
Particle density
Gas viscosity
Dust loading
Relationship
Pl
P2~
Pl
P2
Pl
P2~
Pl
P2~
ftf
r(S - "<£
_Cpi ^_
@r
/c v 0.182
0.5
Reference
Licht,15 Theodore
Buonicore4
r Licht,15 Theodore
Buonicore4
Licht,15 Theodore
Buonicore4
Baxter18
and
and
and
The fractional efficiency of a cyclone system may be improved by reduc-
ing the cyclone diameter and using multiple cyclones to .handle the gas flow.
Penetration curves for several common multiple-cyclone tubes are shown in
Figure 4.2-16.
LARGE DIAMETER TUBES
MEDIUM DIAMETER TUBES
SMALL DIAMETER TUBES
0
10
20 30 40 50 60
AERODYNAMIC PARTICLE DIAHETER, pmA
Figure 4.2-16. Penetration curves for multicyclone tubes of different
diameter.
Adapted from: Theodore and Buonicore, Industrial Air Pollution Control
Equipment for Particulates, CRC Press, page 129, 1976.
100
4.2-23
-------
The overall penetration curve for a number of cyclones in parallel,
such as in a multiple-cyclone tube bank, may differ from the efficiency of
individual tubes because of interferences from the inlet and outlet, gas
distribution, dust stratification, reentrainment from the hopper, inleakage,
axial vane wear, plugging, and variation in pressure drop across individual
tubes.
Axial inlet - axial discharge (double vortex) cyclones. The funda-
mental mechanisms involved in the double vortex cyclones have been discussed
in detail by Schmidt;19 however, theoretical models are not available. A
general penetration curve prepared by Aerodyne Corporation is presented in
Figure 4.2-17.
I I II
I I I
4 1 J.
"0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
AERODYNAMIC PARTICLE DIAMETER,
Figure 4.2-17. Penetration curve for double vortex cyclone.
Reprinted with permission of Aerodyne Corporation, Cleveland, Ohio.
The reduced penetration claimed in the <2-|jmA size range may be partially
due to reduced particle reentrainment resulting from particle bounce.
Additional information is available in Klein.20
4.2.2.2 Static Pressure Drop. Typical cyclones have static pressure
losses ranging from 0.25 to 1.5 kPa. A number of factors contribute to
this, including the kinetic energy losses in the cyclone vortex, cyclone
4.2-24
-------
cylinder wall friction, and entry and exit duct functions. The kinetic
energy losses are normally considered to be the dominant factor.
One common means to calculate static pressure drop, AS. P., is based on
"inlet velocity heads." Licht15 summarizes the basic approach in
Equations 4.2-16 and 4.2-17. Equation 4.2-16 was originally developed by
Shepard and Lapple.21
AS. P. = 7.6 x 10"6 p_V? Nu (Eq. 4.2-16)
u i n
NH = 16 a b/D (Eq 4.2-17)
where AS. P. = static pressure loss, m of H20
PG = gas density, kg/m3
V. = gas inlet velocity, m/s
N,, = number of velocity heads
a = inlet duct width, m
b = inlet duct height, m
D = diameter of exit tube, m
This simple approach for single cyclones indicates that the static
pressure is proportional to the square of the inlet velocity and is directly
proportional to the inlet/outlet area ratio. Because of the dependence on
gas inlet velocity, it follows that static pressure drop is proportional to
the square of the gas flow rate. Other pressure drop equations share the
gas- flow- rate squared dependence, but differ with respect to the cyclone
dimensional factors taken into account. For more information on these
approaches, see Strauss,7 Theodore and Buonicore,4 and Byers.22
4.2.3 Design of Mechanical Collectors
Proper design of mechanical collectors is necessary to ensure unit
operation at optimum efficiency and to minimize malfunctions. This section
addresses both the design needed to achieve desired penetration levels and
the factors necessary for continuous compliance.
4.2.3.1 Settling Chambers. Proper design of settling chambers in-
cludes not only specifying the volume adequate for overall collection effi-
ciency, but also ensuring uniform gas distribution and minimum turbulence.
4.2-25
-------
A nonuniform gas distribution in the inlet can result in locally high
velocities and nonuniform dust concentrations through the chamber. The high
velocities lead to increased penetration of fine particulates. Gas distri-
bution may be improved by the use of gradual turns in the ductwork,
straightening vanes, and perforated plates.
The design must prevent air inleakage into the chamber and dust hopper.
Inleakage increases turbulence, causes dust reentrainment, and prevents dust
discharge from the hopper. Inleakage through the hopper can be prevented by
use of a rotary air lock, flapper valve, or other means of sealing the dust
discharge. Inleakage through the chamber shell can be prevented by use of
proper welding practices and proper materials of construction. When a
chamber is used to process a gas stream at high temperatures with high
moisture content, the inleakage of cold air causes local gas quenching and
condensation. The condensation can cause corrosion, dust buildup, and
plugging of the hopper. The use of thermal insulation can reduce radiant
heat loss and prevent operation below the dew point.
Materials of construction must provide extended life under conditions
of corrosion and abrasion. The composition and gauge of materials should be
specified on the basis of the expected composition of the gas stream.
Access doors and cleanout doors should be included to allow inspection and
cleanout of duct work, distribution vanes, and hoppers.
Since the large particles are separated at the front of the collectors,
it is expected that the maximum weight and volume of material will accumu-
late near the collector inlet. The design must incorporate special access
for frequent dust removal from this area to maintain operation. Normal
instrumentation for settling chambers consists only of an indicator of
differential static pressure. An increase in static pressure drop can
indicate plugging of turning vanes or distribution plates.
4.2.3.2 Momentum Separators. The design of momentum separators must
provide sufficient volume to allow settling of materials separated from the
high-velocity gas stream and materials of construction hard enough to
survive high abrasion.
As with all collectors, the design must include methods of sealing dust
discharge from hoppers to prevent inleakage. The methods may include use of
4.2-26
-------
rotary air locks, flapper valves, or other positive sealing devices.
Inleakage through the hopper or shell results in changes in the gas distri-
bution, interferes with dust discharge, and may cause condensation or corro-
sion.
Because of the high velocities used to separate the particles from the
gas stream and the impaction of these particles on surfaces that direct the
gas flow, the materials of construction must have high abrasion resistance.
Access must be provided for inspection and cleanout of the gas passages
and dust hoppers. The access must be large enough to allow the changing of
surfaces subject to high abrasion such as liners or blast plates.
Normal instrumentation consists only of differential static pressure
indicators. The plugging of gas passages or abrasion of baffle plates may
be indicated by a shift in normal static pressure drop. Such effects,
however, should be detected by visual inspection of the collector before a
significant change in static pressure is noted.
4.2.3.3 Mechanically Aided Separators. Because of the high rotational
speed of mechanically aided collectors, the major design considerations are
abrasion and vibration. Because of the abrasion associated with removal of
large particles, the design must include an impeller made of materials with
high abrasive resistance. The inlet of the collector is designed to provide
even wear of the impeller. If uneven wear occurs, the resulting imbalance
could cause bearing or impeller failure.
The effects of abrasion and material buildup on the impeller are mini-
mized by operating the units at low speeds (typically 400-800 rpm). Buildup
of sticky or wet materials on the impeller may be reduced by a water spray
in the inlet. The coating of water reduces adhesion of particles and acts
as a wet fan.
The periodic sheeting of materials from the blade tips results in
moderate vibration. The housing and structural support should be either
spring-mounted or rigidly fixed to a substantial foundation, depending on
ability to isolate the system from ductwork.
Since the mechanically aided collector acts as the prime air mover in
the system, the air volume is affected by all of the normal variables that
affect fans. The typical fan curve may not be applicable because of con-
tinuous changes in surface contour resulting from wear and material buildup.
4.2-27
-------
Because of the constant change in operating conditions, the penetration
curve is subject to change.
An increase in impeller weight caused by buildup can increase drive
belt slippage and can increase brake horsepower. The design should allow
for multiple drive belts or direct drive and high motor horsepower.
4.2.3.4 Cyclones.
Simple cyclones. The design of simple cyclones includes sizing of the
cyclone to provide an inlet velocity that will ensure high separation
efficiency without excessive turbulence. Typical values range between 10
and 25 m/s.4 Turbulence at high inlet velocities results in abrasion of the
cyclone wall and gas outlet duct. These areas can be reinforced by use of
materials of extra hardness or thickness. Turbulence can be reduced by use
of inlet configurations such as helical or involuted designs.
The design should use flush welds and should avoid components that
increase roughness, such as rivets. Internal disturbances reduce efficiency
by creating areas of turbulence and causing particles to bounce from the
wall to the inner vortex. The design should also include methods of sealing
the dust discharge and preventing gas inleakage, such as use of a rotary air
lock, flapper valves, or other devices.
Multiple cyclones. The design of a multiple-cyclone system entails
specification of the number of individual tubes needed to handle the gas
volume without exceeding the maximum gas flow per tube. If the inlet
velocity of the tube is excessive, the resulting turbulence increases pene-
tration. At excessive inlet velocities abrasion also increases. In
general, the maximum inlet velocity may be limited by the abrasiveness of
the particles to be collected and by the abrasion resistance of inlet vanes.
The arrangement of the tube bank generally should be in a square
matrix. The use of internal baffles in the inlet can allow a collector of
shallow depth to be used without abnormal gas distribution across the width.
Because of the possibility of static pressure differential across large tube
banks and the resulting hopper short-circuiting, baffles should be included
in the hopper or multiple hoppers should be provided.
Access should be provided both to clean and dirty collector plenums for
dust cleanout, inspection, and replacement of tubes. The access doors
4.2-28
-------
should be large enough to permit safe entry and removal of tubes. Gasket
materials used to seal tube assemblies, access doors, etc., should have long
life at elevated temperatures. A major cause of failure of well-designed
systems is gas penetration through weld gaps and gasket leaks from the dirty
side to the clean side.
The collection hopper should be equipped with devices that provide
positive sealing of the dust outlet, e.g., rotary valve, flapper valve, or
other devices to prevent air inleakage. Air leakage can cause dust dis-
charge, hopper bridging, and condensation.
The collector should be equipped with differential static pressure
monitoring equipment to determine static pressure drop.
4.2,4 Operation andMaintenance of Mechanical Collectors
Even well-designed equipment can fail because of improper operation and
inadequate maintenance. Despite the apparent simplicity of mechanical
collectors, regular maintenance is necessary.
4.2.4.1 Settling Chambers. The most common failure mode of settling
chambers is plugging of the chamber with collected dust. In simple collec-
tors the plugging can result from hopper bridging or rotary air lock
failure. In more complex collectors, such as Howard's settling chamber,
plugging of the individual gas passages can occur. Such failures can be
prevented or minimized by use of hopper level indicators or by continuous
monitoring of the dust discharge. Scheduled internal inspection can deter-
mine areas of inleakage and condensation, both of which may cause hopper
bridging.
4.2.4.2 Momentum Separators. The most common failure modes of
momentum separators are hopper plugging and baffle plate erosion. Plugging
of hoppers can be reduced by use of hopper level indicators. Erosion of
baffle plates and collector shell can be reduced by the use of extra thick-
ness in areas subject to abrasion. Periodic internal inspection of the
collector is recommended to identify and correct areas of high abrasion and
air inleakage.
4.2.4.3 Mechanically Aided Separators. The most common failure modes
of mechanically aided separators are abrasion and structural disintegration.
4.2-29
-------
The impact of an abrasive dust on impeller surfaces at high tip speeds
causes rapid erosion. The impeller can become unbalanced; and if the
unbalance is not corrected, structural failure can occur. The attachment of
sticky or tacky particulate to the impeller can also cause vibration.
The use of vibration detectors and periodic external inspection can
indicate impending failure from erosion or material buildup. Routine in-
ternal inspection of the impeller with removal of collected material can
extend collector life.
4.2.4.4 Simple Cyclones. Simple cyclones fail most often from abra-
sion and plugging of the dust outlet tube. Abrasion occurs in areas oppo-
site the gas inlet and in the lower areas of the cone. Wear can increase if
inlet velocities are high (>25 m/s) and particles can erode the gas outlet
tube. Internal roughness caused by poor welding or faulty fabrication can
cause local turbulence, which increases erosion and particle bounce into the
inner vortex. Plugging of the dust outlet tube can be reduced by use of a
large diameter outlet tube and by proper sizing of the rotary air lock.
Routine internal inspection of the inlet duct and internal surfaces for
areas of abnormal wear should be conducted to prevent shell failure. Liners
and replaceable wear plates are recommended where abrasive dust is
collected.
4.2.4.5 Multiple Cyclones. Factors which may contribute to reduced
performance include erosion, plugging, corrosion, and hopper recirculation.
Hopper recirculation in reverse flow type systems occurs when tubes at
the rear of the bank have a slightly lower static pressure drop. This can
occur whenever flow distribution is nonideal or when the outlet tubes are
shorter than those in the front23'24 (a common design). In these cases a
portion of the gas in the front tubes can pass out the bottom and then enter
the discharge of tubes in the back rows.
Hopper recirculation can lead to substantially reduced collection effi-
ciency.24 Some of the ways to minimize this problem include segregating the
hopper and equalizing the lengths of all discharge tubes.23
Ambient air inleakage into the hopper area can interfere with the dust
discharge in the tubes and lead to increased emissions. These leaks may
occur at hopper flanges, access hatches or through the solids discharge
4.2-30
-------
valves. Severe leakage on multicyclones serving combustion systems can be
identified by measurement of the flue gas oxygen content before and after
the collector.
Plugging of the gas inlet vanes leads to partial or total failure of
individual tubes to establish an adequate vortex. Partial plugging allows
dust penetration through the affected tube (Figure 4.2-17). Complete plug-
ging results in an increased flow of gas through the remaining tubes and an
increase in overall collector static pressure drop. Partial plugging of a
significant number of tubes can cause variation in pressure drop from tube
to tube, cause gas short circuiting through the dust hopper, and cause an
increase in dust reentrainment at the tube dust outlet.
Plugging of outlet tubes or the solids discharge is common.23 The
latter is generally due to poor hopper discharge practices, which allow
solids to accumulate to the bottom of the tubes.23 According to Barrow,25 a
single plugged tube can reduce collection efficiency by as much as 25 per-
cent.
Excessive gas velocities at the entrance of the multicyclone can lead
to erosion of the outlet tubes. Once a small hole is created, the large
differential static pressure between the inlet and outlet leads to very high
gas velocities through the hole. Rapid erosion then leads to substantially
reduced collection efficiency.
Corrosion in multicyclone collectors is usually minimized by avoiding
operation at or near the acid vapor dewpoint.23 In certain cases, insu-
lating the units and ductworks leading to it may be advisable.
Regular inspection of the multicyclone should be performed to minimize
the above problems. Static pressure taps should be installed on the inlet
and outlet ducts so that the static pressure drop can be measured on a
regular basis by either a portable gauge or a permanently mounted instru-
ment. Because of the tendency of these taps to plug, they should be cleared
before any measurements are attempted.
For units serving combustion sources, gas temperature and flue gas
oxygen content should be measured on a routine basis. This value should be
compared with typical values for the specific unit to identify deterioration
of the unit and/or operation below the dewpoint.
4.2-31
-------
JSi
00
JSi
Figure 4.2-18. Partial pluggage of multiple cyclone inlet vanes.
(Courtesy of PEDCo Environmental, Inc.)
-------
The rate of solids discharge should be checked frequently. Bridging of
solids in the hopper, failure of the solids discharge valve, or severe
internal problems may all be identified by this simple technique. All of
these situations demand immediate attention.
At least once a year, internal inspections should be conducted after
the unit has been purged by personnel qualified in proper confined-area-
entry procedures. The internal inspection should identify plugging,
erosion, and corrosion problems. In some cases it may be advisable to use
tracers or smoke to identify gasket and weld leaks.
4.2-33
-------
REFERENCES
*-,
1. Crawford, M. Air Pollution Control Theory, McGraw-Hill, Inc., New
York. 1976, p. 236.
2. Sargent, 6. Gas/Solid Separation. Chemical Engineering. 78(4):11-12.
February 15, 1971.
3. Munson, 0. S. Dry Mechanical Collectors. Chemical Engineering.
75(22):147-151. October 14, 1968.
4. Theodore, L., and A. J. Buonicore. Industrial Air Pollution Control
Equipment, CRC Press, Cleveland, Ohio, 1976, p. 65.
5. Alden, J. L. Design of Industrial Exhaust System, 3rd ed. The Indus-
trial Press, New York. 1959.
6. Jennings, R. F. J. Iron Steel Inst. 164:305. 1950.
7. Strauss, W. Industrial Gas Cleaning, 2nd ed, Pergamon Press. 1975,
p. 213.
8. Stairmund, C. J. Trans Instr. Chem Engrs. 29:356. 1951.
9. Caplan, K. Mechanical Collectors. In: Air Pollution, Vol. Ill, 2nd
ed. Stern, A., Editor. Academic Press. New York. 1968.
10. Aerodyne Development Corporation, Series "SV" Dust Collector, Bulletin
No. 1275-SV.
11. Calvert, S., J. Goldschmidt, D. Leith, and D. Mehta. Wet Scrubber
Handbook, Vol. I. U.S. EPA. Research Triangle Park, N.C. Publication
No. EPA-Rl-72-118a. 1972.
12. Air Pollution Engineering Manual. U.S. Environmental Protection Agency
Publication, AP-40. Research Triangle Park, N.C. 1973. p. 95.
13. Leith, D., and W. Licht. A.I.Ch.E. Symposium Series. 68:196.
November 26, 1979.
14. Kalen, B., and F. A. Zenz. A Theoretical-Empirical Approach to Salta-
tion Velocity in Cyclone Design. The Ducon Company. Bulletin No.
C-207. 1973.
4.2-34
-------
15. Licht, W. Air Pollution Control Engineering. Marcel Dekker, Inc.
1980.
16. Alexander, R. Mck. Proc. Austral. Inst. of Min. and Met (N.S.).
152:202. November 3, 1949.
17. Theodore, L., and V. DePaola. Predicting Cyclone Efficiency. JAPCA,
30:1132-33. October 1980.
18. Baxter, W. In: Air Pollution, Vol. II, 3rd ed. Stern, A. Editor.
Academic Press. New York. 1977.
19. Schmidt, K. R. Staub. 23:491-501. 1963.
20. Klein, H. Staub. 23:501-08. 1963.
21. Shepard, G. B. , and C. E. Lapple. Ind. Eng. Chem. 31:972. 1939.
22. Byers, R. L. Gravitational and Dry Centrifugal Collectors and Air
Pollution Control. In: Integrated Engineering Solution to Overall
Pollution Control: Air, Water, and Solid Waste Problems. AICHE. New
York. 1971.
23. Dey, A., J. Maloney, and J. D'Imperio. Inertia! Separators. In: Air
Pollution Engineering Manual. U.S. Environmental Protection Agency,
Research Triangle Park, N.C. Publication No. AP-40. 1973. pp. 91-99.
24. National Asphalt Paving Association. The Maintenance and Operation of
Exhaust Systems in the Hot Mix Batch Plant. Information Series 52,
NAPA. Riverdale, MD. 1975.
25. Barrow, A. J., Jr. Particulate and S02 Control Technology for the
Small and Medium Coal-Fired Boiler. Presented at the 1970 Industrial
Coal Conference. Purdue University. October 7-8, 1970.
4.2-35
-------
-------
4.3 ELECTROSTATIC PRECIPITATORS
Electrostatic precipitators (ESP's) are high efficiency particulate
collection devices applicable to a variety of source categories and gas
conditions. Particle collection is done by application of electrical energy
for particle charging and collection. Efficiencies of 99.9+ percent are
possible, depending upon application, ESP design, and gas and particle
characteristics.
4.3.1 Types of ElectrostaticPrecipitators
Electrostatic precipitators used for controlling particulate emissions
may be placed in two general categories: dry and wet. The major difference
is in the method by which the particulate is removed from the collection
electrodes. Each of these categories of precipitators may be further sub-
divided on the basis of electrode geometry and application.1 The dry ESP
with pi ate-type collection electrodes and pyramidal hoppers is the pre-
dominant type in industrial applications. Regardless of the type of precip-
itator and its geometry, the particle capture is accomplished by electro-
static attraction.
4.3.1.1 Dry Precipitators. Examples of the industrial sources that heavily
utilize dry ESP's for control of particulate emissions are utility boilers,
cement kilns, kraft pulp recovery boilers, and metallurgical furnaces. Each
industrial application requires different designs for the gas conditions and
particulate characteristics.
The basic functions of an ESP are (1) to impart a charge to the partic-
ulate, (2) to collect the charged particulate on a surface of opposite
polarity, (3) to remove the collected particulate from the collecting sur-
face in a manner that minimizes reentrainment of the particulate into the
gas stream, and (4) to discharge the collected particulate. Each dry ESP
with horizontal gas flow must have a shell to enclose the collection and
discharge electrode system, collection electrodes, discharge electrodes, a
high-voltage transformer-rectifier for application of electrical power to
the ESP, a system of rappers to remove particulate from the collection and
discharge electrodes, and a system to remove the collected particulate from
the precipitator proper. Figure 4.3-1 shows the basic features of a
4.3-1
-------
•US DUCT ASSY
INSULATOR
COMPARTMENT
VENTILATION SYSTEM
HIGH VOLTAGE
SYSTEM RAPPER
INSULATOR
COMPARTMENT
HAILING
HIGH VOLTAGE
SYSTEM UPPER-.
SUPPORT FRAME
24 1n.
MANHOLE
TRANSFORMER/RECTIFIER
REACTOR
PRIMARY LOAD
RAPPER
ROL PANEL
ELECTRICAL
EQUIPMENT
PLATFORM
COLLECTING
SURFACES
HIGH VOLTAGE
ELECTRODES
WITH WEIGHT
COLLECTING
SURFACE
RAPPERS
HOPPER
Figure 4,3-1.
Typical ESP with insulator compartments (courtesy
of Western Precipitation),2
4.3-2
-------
weighted-wire discharge electrode ESP. Another type of design uses a rigid
frame discharge electrodes mislead of weighted cones.
4.3.1.1.1 Precipitator housing or shell. Gases entering the precipi-
tator are usually ducted so that gas velocities are high enough to prevent
significant particulate fallout in the ductwork. These velocities (typical-
ly 20 to 35 m/s) are much too high for the ESP to capture the particles.
Thus, the precipitator shell not only encloses the gas treatment area but
also constitutes a large expansion volume in the ductwork to reduce gas
velocity to between 1 to 2 m/s. This reduced velocity is necessary to
allow sufficient residence time in the ESP for the particles to migrate to
the collection electrodes and to avoid reentrainment to the gas stream.
The precipitator shell is typically made of carbon steel that is insu-
lated to reduce corrosion of the shell. Doors are usually provided between
each field to allow internal inspection and maintenance when the precipi-
tator is off-line. Depending upon design and manufacturer, the precipitator
may be equipped with a penthouse to house high-voltage insulators.
4.3.1.1.2 Discharge electrodes. Dry plate ESP's are of two types:
weighted-wire and rigid-frame discharge electrode systems. The discharge
electrode provides the charge to the particulate in the treatment zone.
Weighted-wire designs (Figure 4.3-1) have dominated ESP service in the past.
Wires are suspended from a high-voltage frame at the top of the electrical
section, and shrouds are provided to reduce sparking between the wires and
the end of the collection electrode. The weighted-wire design typically has
a lower initial cost than rigid-frame designs, and closer spacings are
allowed between collection electrodes and discharge electrodes. Rigid-frame
designs, however, tend to provide lower maintenance requirements.
Both discharge electrode configurations have excellent collection capa-
bility. The low initial cost of a weighted-wire design is typically offset
by high maintenance costs caused partially by wire breakage. The reverse is
true for rigid-frame designs for which the high initial costs are usually
offset by low maintenance costs. Warping of the discharge electrode frame
because of wide thermal swings is generally not a problem with a properly
designed unit. Electrically, both designs are capable of delivering similar
power levels to the ESP for particulate collection. Generally, the rigid-
frame design operates at higher voltages and lower current densities than
4.3-3
-------
the weighted-wire designs in a given application because of the wider dis-
charge electrode to collection electrode spacing. These voltage-current
characteristics may be better suited to collection of high-resistivity
dusts.
4.3.1.1.3 Collection electrodes. The collection electrodes usually
consist of a plate with stiffeners to provide support. Some manufacturers
incorporate baffles (doubling as plate stiffeners) to provide regions of low
gas turbulence that enhance particulate capture. The charged particles that
migrate across the gas stream under the force of electrostatic attraction
build up a layer of dust on the collection electrodes (plates).
4.3.1.1.4 High voltage transformer-rectifier (T-R). Operation of the
ESP depends upon electrical power being supplied by high-voltage trans-
former-rectifiers (T-R). The function of the T-R set is to convert a low-
voltage AC power to high-voltage DC. In most industrial applications the
voltage applied is DC negative (negative corona) because higher power can be
applied to the ESP at lower sparking values than is possible with positive
corona units. Typically, a silicon-controlled-rectifier (SCR) circuit
controls precipitator current "phase," and the high-voltage power supply
includes an automatic control unit to control delivery of optimum voltage-
current performance. Most modern ESP's also include linear reactors to
modify the current waveforms and provide stability during sparking.
Most ESP's have a number of T-R sets; the number is determined by the
manufacturer's preference in conjunction with the desired collection effi-
ciency, degree of sectionalization, and degree of redundancy needed. The
voltage and current ratings of the T-R's must be matched to the application,
electrode geometry, and gas and particulate characteristics. Typically, in-
put voltage is 460 V, three-phase, 60 Hz AC with an output voltage between
45 and 70 kV DC. The maximum rated current output of the T-R sets is usual-
ly in the range of 250 to 1500 mi Hi amperes.
4.3.1.1.5 Insulators. Insulators are needed to prevent grounding of
the high-voltage power supply system with the precipitator shell. Insula-
tors are often made of a ceramic material selected for its high dielectric
strength and resistance to most components in the gases to which it may be
exposed. Insulators are used where the high-voltstge supply penetrates the
4.3-4
-------
precipitator shell and is connected with the discharge electrode system and
wherever the high-voltage system comes close to the precipitator shell or
plates.
4.3.1.1.6 Rappers. For dry ESP's, some version of rapping must be
used to remove particulate from collection surfaces. Rapper types include
electric vibrators and gravity impact hammers. Rappers are provided for
both the collection plates and discharge electrodes. Excessive dust buildup
in the precipitator will degrade the performance of the precipitator and is
usually evidenced by reduced power to the precipitator.
The number of rappers necessary for effective cleaning depends upon the
type of rapper, the ESP configuration, and other design considerations.
Each installation requires fine-tuning of both rapper intensity and rapping
frequencies to minimize reentrainment of dust. The rapping intensity
depends upon the particulate characteristics and the amount of collection
area and discharge electrode to be cleaned per rapper. Rapping frequency
generally decreases as the gas travels from inlet to the outlet of the ESP.
It should be noted that excessive rapping can degrade ESP performance as
much as insufficient or ineffective rapping.
4.3.1.1.7 Solids discharge. Pyramidal hoppers are generally used for
collecting partlculate. Discharge from hoppers may be accomplished by means
of screw conveyors, drag conveyors, or pneumatic conveying systems. In the
pulp and paper industry flat bottom, tile-lined precipitators that utilize
drag conveyors are common on recovery boilers. Solids discharge can repre-
sent a significant problem in the operation of an ESP in that excessive
buildup of material can cause an electrical shortage'or misalignment of ESP
internal components.
4.3.1.1.8 Gas distribution equipment. Effective utilization of elec-
trical energy supplied to the ESP depends upon well-balanced air flow across
the ESP. In new installations the requirements for extensive gas turning
vanes, because of sharp bends in ductwork, can be reduced. Retrofitted
ESP's, however, can present problems because of space limitations requiring
sharp bends in the ductwork immediately before and after the precipitator.
Gas turning vanes and some type of diffusion plate (e.g., perforated plates)
are typically utilized to balance the gas flow.
4.3-5
-------
4.3.1,2 Wet Precipitators. Wet preeipitators are used primarily in
the metallurgical industry, usually operating below 75°C. Until the late
1960's their use was restricted mostly to acid mist, coke oven off-gas,
blast furnace, and detarring applications. Their use in other areas is
rapidly increasing with the need for increased control efficiencies. The
new applications include sources with sticky and corrosive emissions.
Because of inherent temperature range limitations, wet ESP's are not used
for boiler installations.
The fundamental difference between a wet and a dry ESP is that a thin
film of liquid flows over the collection plates of a wet ESP to wash off the
collected particulates. In some cases, the liquid is also sprayed in the
gas flow passages to provide cooling, conditioning, or a scrubbing action.
When the liquid spray is used, it is precipitated with the particles, pro-
viding a secondary means of wetting the plates. Three different wet ESP
configurations are shown in Figure 4.3-2.
4.3.1.2.1 Plate-type (horizontal flow). The effluent gas stream is
usually preconditioned to reduce temperature and achieve saturation. As the
gas enters the inlet nozzle, its velocity decreases because of the diverging
cross section. At this point, additional sprays may be used to create good
mixing of water, dust, and gas as well as to ensure complete saturation
before the gas enters the electrostatic field. Baffles are often used to
achieve good velocity distribution across the inlet of the ESP.
Within the charging section, water is sprayed near the top of the
plates in the form of finely divided drops, which become electrically
charged and are attracted to the plates, coating them evenly. Simultan-
eously, solid particles are charged; they "migrate" and become attached to
the plates. Since the water film is moving downward by gravity on both the
collecting and discharge electrodes, the particles are captured in the water
film, which is disposed of from the bottom of the precipitator in the form
of slurry.
4.3.1.2.2 Concentric-plate. The concentric-plate ESP consists of an
integral tangential prescrubbing inlet chamber followed by a vertical wet-
ted-wall concentric-ring ESP chamber.3'4 Concentric cylindrical collection
electrodes are wetted by fluids dispensed at the top surface of the collec-
tion electrode system. The discharge electrode system is made of expanded
4.3-6
-------
KATE* SUW.T
SPRAT NOZZLES
CORONA MIRES
A. Plate type (horizontal flow)
Reprinted with permission of Academic
Press. Stern, Arthur C., editor, AIR
POLLUTION, 3rd edition, Vol. IV. 1977.
GAS aow IN
SAS ROW our
HI6H
VOLTAGE
LEADS
KATER PIPES
GAS FLOW
HOOD AND STACK
TRANSITION
PRECIPITATOR
SECTION
TYPICAL HETAL
ELECTRODE
DISCHARGE CAGE
CONTINUOUS FILK OF
LIQUOR FLOWS DOWN
POSITIVE COLLECTION
ELECTRODE SURFACES
(CTLINDER WALLS)
VENTURI INLET TO
PRECIPITATOR SECTION
BASE
C. Conventional pipe type
Courtesy of the McILvaine Co.,
THE ELECTROSTATIC PRECIPITATOR
MANUAL.
B. Concentric plate type
Reprinted with permission of Academic
Press. Stern, Arthur C., editor. AIR
POLLUTION, 3rd edition, Vol. IV. 1977.
Figure 4.3-2. Three types of wet ESP's.
4.3- 7
-------
metal with uniformly distributed corona points on the mesh background. This
system is intended to combine the high, nearly uniform, electric field
associated with a parallel plate system and the nearly uniform distribution
of corona current density associated with closely spaced corona points.
Higher gas flows can be handled by adding concentric electrode systems and
by increasing the length of each electrode,
4.3.1.2.3 Conventional pipe-type. This system consists of vertical
collecting pipes, each containing a discharge electrode (wire-type), which
is attached to the upper framework and held taut by a cast-iron weight at
the bottom. A lower steadying frame keeps the weights and thus the wires in
position.
The upper frame is suspended from the high-voltage insulators housed in
the insulator compartments, which are located on top of the precipitator
shell (casing). Heating and ventilating systems help to prevent accumula-
tion of moisture and dust in the insulator compartments.
The washing system usually consists of internal nozzles located at the
top of the plates.5 At specified intervals, the tubes are washed thorough-
ly. During the washing, the louver damper to the exhaust fan is closed to
prevent carryover of droplets.
4.3.1.3 Two-stage Precipitators. The two-stage ESP was originally
designed to purify air and is used in conjunction with air-conditioning
systems. Cleaning of incoming air at hospitals and at industrial and com-
mercial installations is a typical application. As an industrial partic-
ulate collector, the device is used for control of liquid particles dis-
charged from such sources as meat smoke-houses, asphalt paper saturators,
pipe coating machines, and high-speed grinding machines.
Two-stage ESP's are limited almost entirely to the collection of liquid
particles that will drain readily from collection plates. Two-stage precip-
itators cannot control solid or sticky materials, and become ineffective if
particle concentrations exceed 1.0 g/m3.
4.3.2 Operating Principles of Electrostatic Precipitators
4.3.2.1 Basic Processes. The three basic processes involved in elec-
trostatic precipitation are (1) the transfer of an electric charge to sus-
pended particles in the gas stream, (2) the establishment of an electric
4.3-8
-------
field for removing the particles to a suitable collecting electrode, and (3)
the removal of the particle layers from the precipitator (Figure 4.3-3}.
4.3.2.1.1 Corona generation. As the high-voltage DC current passes
through the discharge wire, it produces an electrical corona, which can be
defined as an ionization of gas molecules by electron collisions in regions
of high field strength near the discharge wire.6 The strength of the elec-
tric field varies inversely with the distance from the discharge wire.
Three sources of electrons are used to initiate the so called "ava-
lanche" of collisions (1) naturally occurring ionizing radiation, (2) photo-
ionization because of the corona glow, and (3) in high-temperature applica-
tions, thermal ionization at the electrode surface.6 These electrons and
positive ions as well, move under the influence of an electric field and
carry charges, but the current generated from the flow of these carriers is
too low to be of significance. Under the influence of sufficiently high
voltage, these free electrons are accelerated to a velocity high enough that
collision with a gas molecule will break an electron loose from the outer
shell of the molecule, creating a positive ion and another free electron.7
This phenomenon is repeated many times, thus the name "avalanche." The end
result is a large accumulation of positive ions and negative electrons in
the region of the corona.
The corona can be either positive or negative; but the negative corona
is used in most industrial precipitators since it has inherently superior
electrical characteristics that enhance collection efficiency under most
operating conditions.
4.3.2.1.2 Electric field. The electric field results from application
of high voltage to the ESP discharge electrodes, and the strength of this
electric field is a critical factor in determining ESP performance. Space
charge effects from charged particles and gas ions may interfere with gener-
ation of the corona and reduce the strength of the electric field. The
space charge effect is often seen in the inlet fields of art ESP where parti-
culate concentration is the highest. From a practical standpoint, the
strength or magnitude of the electric field is an indication of the effec-
tiveness of an ESP. The magnitude of the,electric field can be mathemat-
ically determined by use of derivations of Poisson's equations including
charge carriers and their mobilities.7
4.3-9
-------
M
O
FREE
ELECTRONS
REGION OF\
CORONA GLOW\
CORONA GENERATION
•e
ELECTRONS
ELECTRON
DUST
PARTICLE
GAS
MOLECULE
CHARGING
1
Figure 4.3-3. Basic processes involved in electrostatic precipitation.7
{Oglesby, Sabert Jr., "Electrostatic Precipitation" SRI Bulletin/Winter 1971.)
Courtesy of Southern Research Institute.
-------
4.3.2.1.3 Charging mechanisms. Particle charging and subsequent
collection take place in the region between the boundary of the corona glow
and the collection electrode, where gas particles are subject to the gener-
ation of negative ions from the corona process. Charging is generally done
by field and diffusion mechanisms. The dominant mechanism varies with
particle size.
In field charging, ions from the corona are driven onto the partj^les
by the electric field. As the ions continue to impinge on a dust particle,
the charge on it increases until the local field developed by the charge on
the particle causes such distortion of the electric field lines that they no
longer intercept the particle, and no further charging takes place. This is
the dominant mechanism for particles larger than about 0.5 umA.
The time required for a particle to reach its saturation charge varies
proportionally to the ion density in the region where charging takes place.
Under normal conditions with sustained high-current levels, charging times
are only a few milliseconds. Limitation of current because of high resis-
tivity or other factors can lengthen charging times significantly and cause
the particles to travel several meters through the precipitator before
saturation charge is approached.
The waveform of the secondary voltage can further affect the charging
times. The rectified unfiltered voltage has peaks occurring at regular
intervals, which match the frequency of the primary voltage. Thus, the
electric field varies with time, and the dust particles in the interelec-
trode region are subject to time-varying fields. The particle charging is
interrupted for that portion of the cycle during which the charge on the
particle exceeds that corresponding to the saturation charge for the elec-
tric field existing at the time. This further lengthens the charging times
and, in the case of high-resistivity dust, degrades precipitator perform-
ance.7
Diffusion charging is associated with ion attachment resulting from
random thermal motion; this is the dominant charging mechanism for particles
below about 0.2 umA. AS with field charging, diffusion charging is in-
fluenced by the magnitude of the electric field, since ion movement is
governed by electrical as well as diffusional forces. Neglecting electrical
forces, an explanation of diffusion charging is that the thermal motion of
4.3-11
-------
molecules causes them to diffuse through a gas and contact the particles.
The charging rate decreases as a particle acquires a charge and repels
additional gas ions, but charging continues to a certain extent because
there is no theoretical saturation or limiting charge other than the limit
imposed by the field emission of electrons. This is because the distribu-
tion of thermal energy ions will always overcome the repulsion of the dust
particle.7
The particle size'-range of about 0.2 to 0.5 pmA is a transitional
region in which both mechanisms of charging are present but neither is
dominant. Fractional efficiency test data for precipitators have shown
reduced collection efficiency in this transitional size range, where diffu-
sion and field charging overlap.
A more comprehensive theory8 has been derived that analyzes the diffu-
sion and field charging mechanisms simultaneously. The ion density distri-
bution near a particle is determined in terms of the local electric field,
and the rate at which ions reach the particle due to their thermal veloc-
ities is calculated statistically. A computer is used to calculate the
theoretical charging rate, since it cannot be expressed in closed algebraic
form. Results of this work indicate that the total charge accumulated by a
particle is strongly dependent on the electric field strength, the diameter
of the particle, the numerical density of the ions, and the residence time
of the particle in the charging region. Other variables cited that can have
a significant effect on particle charging rates are the gas temperature,
electrical mobility of the ions in the gas, and the dielectric constant of
the particulate material.
Sufficient rapping force must be applied to produce a rapid accelera-
tion perpendicular to the gas flow so that the dust shears off the plate.
Sproull9 indicates that rapping is optimum if the dust layer slides down the
plate vertically after each rap, making its way down the plate in the dis-
crete steps until it finally reaches the hopper.
With a tenacious dust that adheres stubbornly to the plate, vibrations
can be induced perpendicularly to the gas flow direction, in addition to the
necessary shear action, resulting in a scattering of the agglomerate and
subsequent reentrainment of relatively large fractions of the dust. In
general, the dust should be allowed to fall freely off the plate, as some-
times occurs with high-resistivity dust when rapping is done with "power
4.3-12
-------
off." The other extreme is with low-resistivity dust, whose reentrainment
can be caused by only a light rap.
Recent studies have investigated reentrainment caused by rapping in
terms of the percentage of material reentrained and its particle size dis-
tribution.10,11 One report describes the testing of six full-scale ESP
installations. Losses from rapping ranged from over 80 percent of the total
mass emissions from one hot-side unit to 30 percent of emissions from cold-
side units. The losses consist mostly of relatively large particles, pri-
marily those larger than 2.0 umA in diameter. Tests of a pilot-scale pre-
cipitator showed that rapping emissions decreased as time between raps was
increased.3
The intensity and frequency of rapping are usually greatest at the
inlet sections, decreasing as the gas moves through the ESP. The outlet
section is usually rapped only lightly, since the reentrained dust is not
recollected. The visible puffs that often appear as a result of rapping can
be used with a transmissometer to optimize the frequency and intensity of
rapping for each section of the ESP.
4.3.2.1.4 Importantprocess parameters influencing ESP performance.
The process parameters that most influence the design and operation of ESP's
are the particle properties such as resistivity and particle size distribu-
tion, and gas properties such as process temperature and flow. Once these
factors are determined, the designer can estimate the size of the ESP needed
to meet applicable emission regulations.
Corona current flows through the collected dust layer to reach the
collection electrode. With dry ESP's, high resistivity affects ESP effi-
ciency by limiting the current and voltage. If electrodes are clean, the
voltage can be increased until a sparking condition is reached. The maximum
voltage is determined principally by the gas composition and ESP dimensions.
When dust is deposited on the collection electrode, however, the voltage at
which sparking occurs is reduced because of the increased electric field at
the dust surface. As dust resistivities increase, the voltage at which
sparking occurs decreases. At values of resistivity above approximately
1012 ohm-cm, the voltage must be reduced so that sparks will not propagate
across the interelectrode space. At very low values of current and voltage,
4.3-13
-------
dust breakdown can occur. This can result in a back corona in which posi-
tive ions form and flow back toward the discharge electrode, neutralizing
the negative charge previously applied and thereby limiting ESP performance.
In addition to reducing the performance of an ESP, high-resistivity
dust can cling more tenaciously to collection electrodes than particles with
intermediate resistivity. A much greater rapping acceleration must then be
applied to the electrode to remove the dust layer. This can cause severe
reentrainment or damage to a precipitator that is not designed to withstand
such high acceleration.
Particle resistivity depends primarily on the chemical composition of
the ash, the ambient flue gas temperature, and the amounts of water vapor
and S03 in the flue gas.12 At low temperatures <250°C, current conduction
occurs principally along the surface layer of the dust and is related to the
absorption of water vapor and other conditioning agents in the flue gas.
For fly ash, resistivity is primarily related in an inverse manner to the
amount of SQ3 and moisture in the flue gas. Burning of low-sulfur coal
releases smaller amounts of S02, which is oxidized to S03. A higher-
resistivity fly ash results, except at temperatures below about 250°C, where
significant amounts of S03 are absorbed onto the fly ash particles.
At elevated temperatures >25Q°C, conduction takes place primarily
through the bulk of the material, and resistivity depends on the chemical
composition of the material. For fly ash at temperatures above 250°C resis-
tivity is generally below about 1010 ohm-cm. Resistivity has been shown to
decrease with increasing amounts of sodium, lithium, and iron.13
The range of operation of cold-side fly ash precipitators is 110° to
20Q°C> a range in which conduction takes place by a combination of the
surface and bulk mechanisms and resistivity of the ash is highest. Figure
4.3-4 illustrates this relationship.
4.3.2.1.5 ESP performance as a function of particle size. The per-
formance of an ESP varies considerably with changes in the particle size
distribution. Particles from 0.1 to 1.0 umA are the most difficult for an
ESP to collect. Usually, the greatest penetration through an ESP is by
particles 0.2 to 0.4 pmA (Figure 4.3-5). This penetration is probably
caused by a transition from field charging to diffusion charging.
4.3-14
-------
1000/TEMP., °K
E
u
I
JC
o
UJ
(X
10'
10
13
10
12
10
11
10
10
TO3
3.2 2.8 2.4 2.0 1.6 1.2
. i i i i ii i i i i
SURFACE \ |
RESISTIVITY \ VOLUME
RESISTIVITY
COMPOSITE
OF SURFACE
AND VOLUME,
ESISTIVITY^
70 150 250 400
100 200 300
600 800 1000
TEMP., °C
Figure 4.3-4. Typical temperature-resistivity relationship,
(reprinted with permission of Academic Press, Stern,
Arthur C., editor. AIR POLLUTION, 3rd edition,
Volume IV. 1977).7
4.3-15
-------
0.05
0.04
0.03
5
I—I
t=
0.02
0.01
0 1 2 3 .4 5 6 7 8 9 10
AERODYNAMIC PARTICLE DIAMETER, ymA
Figure 4.3-5. Penetration as a function of particle size for
an ESP on a kraft pulp mill recovery boiler.9
4.3-16
-------
Another problem posed by particles <1 |jmA in diameter is the reduction
in operation current associated with the electrical space charge of these
fine particles. Introducing large quantities of fine particles at the inlet
of the ESP can increase the electric field at the collecting plate, weaken
the field near the discharge electrode, and suppress corona generation.
This is known as corona quenching; it occurs when many incoming fine par-
ticulates acquire the same negative charge as the ions producing the charge,
resulting in an electrical repulsion that tends to reduce operating current.
Performance as a function of particle size has been measured at many
installations and has been the subject of computer modeling. Probably the
best known and most versatile model is the EPA-Southern Research Institute
Mathematical Model.14 This model can be used for sizing and troubleshooting
ESP's as well as for predicting penetration.
The effects of gas velocity distribution, mass entrainment, and gas
sneakage on penetration can also be modeled. The following descriptions of
the EPA-SoRI model is excerpted from Ensor et a!.15 Computations are based
on the Deutsch equation. The migration velocity for particles of a given
diameter is assumed to be constant over the collection area. The efficiency
is calculated for each particle size over an incremental length of the
precipitator. The incremental lengths must be sufficiently small to assure
a constant electric field and particle charge. The total efficiency is
integrated over each particle size over the total length of the ESP. The
calculation involves an iterative solution based on an initial estimate of
the total collection efficiency. Empirical corrections to the migration
velocities for each particle diameter are used to verify the model. The
space charge due to particulates is calculated in order to determine the
reduced free ion density used to obtain the particle charge. The electrode
spacing and applied voltage are used to calculate the electric field migra-
tion velocity. The nonideal behavior created by gas velocity distribution
is determined empirically. The .correction factor F is determined using this
relationship:
F = 1 + 0.77 01'786 + 0.0755 a An. (3-—) (Eq. 4.3-1)
where
a = the normalized standard deviation of the gas distribution
4.3-17
-------
rj = the ideal collection efficiency.
The nonideal penetration is then calculated by:
Pt = exp (- |) (Eq. 4.3-2)
where
P = penetration
k = constant predicted under ideal conditions by:
k = &r\ Or——) (Eq. 4.3-3)
Gas sneakage is another factor that can deteriorate ESP performance when gas
bypasses the electrified areas by flowing through the hopper or through the
high-voltage insulation space. A correction factor, B, analogous to the gas
flow quality factor, F, can be used as a divisor for the migration velocity
in the exponential argument of the Deutsch equation.
B = £H (1 - q) (Eq. 4.3-4)
NS AnCsn + (i - sn)(i - n) s]
where
N = number of baffled sections
S
S = fractional amount of gas sneakage per section.
The model assumes that perfect mixing follows each baffled section. This
model also considers the effects of rapping reentrainment on collection ef-
ficiency. It is assumed that the fraction of material that is reentrained
does not vary with particle size or position and that the reentrained mate-
rial is perfectly mixed in the gas stream after rapping. The effect on
penetration is determined by the empirical relationships:
Y1 = (0.155) x°'905 — cold-side ESP
Y2 = (0.618) x°*894 — hot-side ESP
Y is the rapping emissions and X is the mass removed by the last electrical
section; units are mg/DNm3.
Efficiencies with no plate rapping are used by the model to determine
the mass removed by the last electric particle field. A log-normal approxi-
4.3-18
-------
matlon was used for particle size of the rapping emissions based on data
from six different ESP field tests.
Figure 4.3-6 presents a comparison of data from the EPA-SoRI model with
test data from the George Neal No. 3 ESP, operating on a low-sulfur pulver-
ized-coal-fired boiler.15 Best agreement comes with a = 0.6, S =0.1, and
a rapping mass median diameter of 6 [jmA. The curves do not match, however,
at the low or high particle size range. The ideal values for velocity
distribution and sneakage are believed to be higher than normal for a modern
ESP and more the actual occurrence at the George Neal plant. Back corona
was not included in the model, and this is believed to be the cause of most
of the performance degradation. The results from this ESP were unusual in
that the curves for penetration versus particle diameter showed a double
peak in penetration, one at about 0.2 umA and the other at about 1 umA.
Also, a major limitation of the model was the use of correction factors for
nonideal conditions, which were unobtainable under normal plant operating
conditions. For example, the gas velocity distribution is often obtained
with velocity traverses with the boiler and ESP off and the draft fans
operating. Thus, the comparison of the field data with data from the theo-
retical model was qualitative rather than quantitative.15
Some ESP manufacturers also have developed models for use in sizing
precipitators. An example is a Research Cottrell (RC) model used to predict
penetration of particulate through an ESP as a function of particle size.
Although much simpler than the EPA-SoRI model, it includes many of the same
concepts in calculating penetration as a function of particle size.
Predicted penetration as a function of particle size by the RC model is
presented in Figure 4.3-7, for a cold-side ESP on a pulverized-coal-fired
boiler. The computer model shows the expected maximum penetration in the
0.2 to 0.4 MmA size range. Comparison of the computer model with test data
shows an excellent correlation, as indicated in Figure 4.3-8.
The RC model has been modified in more recent publications16 to include
a reentrainment factor, which can be used to estimate emissions due to
rapping reentrainment at sources such as bark/fossil-fuel-fired boilers.
Limitations of this rapping reentrainment factor are similar to the ones in
the EPA-SoRI model: (1) overall migration velocity remains constant, (2)
the fraction of material reentrained remains constant for different particle
4.3-19
-------
0.1 r
0.01
2=
CD
0.001
EXPERIMENTAL DA1A
(O Submicron Particle Mode)
Sn =0.1 a - 0.6 RAP HMD
Sn = 0.1 o = 0.6
Sn « 0.1 o=0.2
Sn « 0 o=0
6.0
RAP HMD = 2.0
RAP HMD « 6.0
OVERALL EFFICIENCY 97.45
SCA 748 ft2/(1000 ft3/min) 147 M2/(m3/sec)
0.1
1 10
DIAMETER (microns)
100
Figure 4.3-6. Comparison of experimental penetration as a
function of particle diameter to the McDonald (1978)
computer model under normal SCA conditions.16
4.3-20
-------
0.1
o
(—t
h-
o.oi
0.001
IJ,J
I I I I I I
III 1
x = 12, 0 = 3.8
(inlet)
0.01 0.020.03 0.050.08.10 .2 .3 .4 .5.6..8^0 2 3 4 5*6 8 10
PARTICLE AERODYNAMIC DIAMETER, y
Figure 4.3-7. Penetration, pulverized-coal-fired boiler
(cold-side ESP).2
4.3-21
-------
1.0
0.1
0.01
0.001.
I I I
ACTUAL FIELD DATA -
A- 98 3%
Q- 98.8%
0- 98.6%
0- 99.21
X- 99.5%
(Inlet: x=3,2; 0=2.72)
PREDICTED BY ESP MODEL
FOR n = 99%
i tii ii
i V ill ii
I I I I I
0.1 0.2 0.3 0.5
1.0
3 456 810
20 30 50 80 100
AERODYNAMIC PARTICLE DIAMETER, u
Figure 4.3-8. Computed versus actual penetration for
cold-side ESP on a western subbituminous-fired boiler.2
4.3-22
-------
sizes, and (3) the fraction of material reentrained remains constant for
each mechanical section.
Figure 4.3-9 shows predicted penetration as a function of particle size
for four different levels of particle reentrainment, based on the RC model.
Note that the maximum penetration is still in the 0.2 to 0.4 \jank size range.
Experimental field test data are needed to support the validity of the
reentrainment factors.
4.3.3 Design of Electrostatic Freeipitators
This section deals with ESP sizing techniques and design considerations
for the major ESP components.
4.3.3.1 Sizing Equations and Techniques. The first equation for
predicting particle collection probability was developed by Anderson in
1919. It was derived again in 1922 by Deutsch, who used a different method.
In various forms, this equation, n. = 1-e , has become the basis for
estimating precipitator efficiency on the basis of gas flow, precipitator
size, and precipitation rate. In this equation, n is the precipitator
collection efficiency, A is the total collecting electrode surface area, V
is the gas flow rate, and w is the migration velocity of the particles.
When determined empirically, the precipitation rate parameter, w, includes
effects of rapping losses, gas flow distribution, and particle size distri-
bution.
The Deutsch-Anderson model assumes that particulate concentration is
uniform across any section perpendicular to the gas flow of an ESP. This
assumption is made because of the turbulence of the gas, which takes the
particles near the collection surface and allows them to become electrically
charged. A serious limitation in use of the Deutsch-Anderson equation is
that it does not account for changes in the particle size distribution and
subsequently in the effective migration velocity, as precipitation proceeds.
This limitation affects the accuracy of sizing estimates for units operating
at very high efficiencies (approximately 98 percent and above) because of
the change in w with particle size.
In practice, factors such as particle reentrainment and gas leakage
cannot be accounted for theoretically. Also, some of the most important
4.3-23
-------
CO
ro
0.2
0.1
§ 0.05
g
S 0.02
LU
0-
0.01
0.005
0.002
0.001
0.005
I II I I I I I
rvj
»
O
I I 1(111
x = 15.0, a = 4.0
(INLET)
i i i i i 1 I I
I i l I I 1 1
\o
O O OOOOO«—
»—^
AERODYNAMIC PARTICLE DIAMETER, y
Figure 4.3-9. Predicted precipitator penetration for bark/fossil-fuel-fired boilers
(k = 0.6392 x 10+1* m"1*; n = 4 fields; r = percent reentrainment).
-------
physical and chemical properties of the particles and gases often are un-
known. Therefore, most designers use an effective precipitation rate param-
eter, w , that is based mainly on field experience rather than theory.17
Data from operating installations form a general basis for selection of w ,
and these data are modified to fit the situation being evaluated. Thus, w
becomes a semiempirical parameter that can be used in the Deutsch-Anderson
equation or its derivatives to estimate the collection area required for a
given efficiency and gas flow.
The most important parameters that determine w in practice are resis-
tivity, particle size distribution, gas velocity distribution through the
ESP, particle loss due to reentrainment, rapping and gas sneakage, ESP
electrical conditions, and requied efficiency.18
Matts and Ohnfeldt developed a semiempirical modification of the
Deutsch-Anderson equation that essentially removes the size dependency from
w.19 This equation is n, = 1-e wk^ . In most cases, k equals approxi-
mately 0.5. The modified migration velocity, w. , can be treated as being
independent of charging voltage and current levels and of particle size
distribution within an ESP, as precipitation proceeds in the direction of
gas flow. Other changes, however, such as in properties of the gas entering
the ESP, resistivity, and size distribution, produce a change in w, , just as
they change the conventional w. Other investigators have proposed other
forms of the basic penetration equation.20,21
Another design technique applied to existing installations or new
processes to aid in a full-scale design is the pilot-scale precipitator.
The main problem with use of a pilot-scale ESP is that the pilot unit almost
always performs better than a full-scale unit because of better gas flow
distribution, sectionalization, and electrode alignment.1,17 The result is
operation at higher current densities and voltages than in a full-scale
unit. Application of a scale-up factor, as in spark-limited operation of a
pilot-scale ESP, can cause uncertainties in sizing the full-scale ESP.
Therefore, pilot precipitator data should be supplemented as fully as possi-
ble by basic data on particle and gas properties, especially resistivity.12
The most important parameters that affect the size of an ESP are col-
lection area, gas velocity, aspect ratio', and structural considerations.
4.3-25
-------
Some variation of the Deutsch-Anderson equation is generally used to es-
timate the required collection plate area. Figure 4.3-10 presents the
relationships of specific collecting areas (SCA's) developed with the
Deutsch-Anderson w and Matts and Ohnfeldt w.,.
6 K
0.001
100 200 300 400 500 600
SPECIFIC COLLECTION AREA,
ft2/1000 acfm
Metric Conversion: ft2/1000 acfm x .055 = mVlOQQ m3/sec
Figure 4.3-10a. Precipitator penetration versus specific
collection area and precipitation rate w .1*
0.001
0.002
0.003
0.005
z 0.01
2 0.02
g 0.03
g 0.05
5 o.i
0.2
0.3
0.5
1.0
100 200 300 400 500 600
SPECIFIC COLLECTION AREA,
ft2/1000 acfm
Metric conversion: ft2/1000 acfm x .055 = m2/1000 mVsec
Figure 4.3-10b. Precipitator penetration as a function of specific
collection area and modified precipitation rate parameter w. .19
4.3-26
-------
The use of sizing equations is only part of the procedure for deter-
mining the final collection area. Each manufacturer has a method of comput-
ing required plate area, usually involving the use of computer models to
assist in the sizing procedure, determination of the amount of redundancy
requested by the user or believed necessary by the manufacturer, and some
amount of judgment. -A recent example of the current approaches of some
manufacturers involves the sizing of ESP's for highly variable fuels.18
Based on ash constituents, fuel sulfur, and the resultant resistivity, and
assuming log-normal relationships for log (resistivity) and ash-to-Btu
ratio, a contour ellipse is drawn from the bivariate normal distribution
(Figure 4.3-11) using the Matts-Ohnfeldt variation of the worst-case fuel.
Iso-SCA lines are then plotted on the contour ellipse and the SCA line
tangent to the contour ellipse defines the required collection area for the
most probable worst case.18
o
o
X= ASH-TO-BTU RATIO
Figure 4.3-11. Distribution of ash-to-Btu ratio
and log (resistivity) for a single fuel field.
Courtesy of the Air Pollution Control Association
Designers usually calculate a hypothetical average value for gas veloc-
ity from gas flow and cross section of the precipitator, ignoring the local-
ized variances within the precipitator. The primary importance of the
hypothetical gas velocity is to minimize potential losses through rapping
and reentrainment. Above some critical velocity, these losses tend to
increase rapidly because of the aerodynamic forces on the particles. This
critical velocity is a function of gas flow, plate configuration, precip-
itator size, and other factors, such as resistivity. Values for gas veloc-
ity in fly ash precipitators range from 0.9 to 1.2 m/s in high-resistivity,
4.3-27
-------
cold-side ESP applications, and in all low- resistivity applications, hot- or
cold-side. For most applications, the values range from 0.9 to 1.7 m/s.
Aspect ratio is defined as the ratio of the length to the height of gas
passage. Although space limitations often determine precipitator dimen-
sions, the aspect ratio should be high enough that reentrained dust carried
forward from inlet and middle sections can be collected. In practice,
aspect ratios range from 0.5 to 1.5. For efficiencies of 99 percent or
higher, the aspect ratio .should be at least 1.0 to 1.5 to minimize carryover
of collected dust.
One of the first structural parameters to be determined is the width of
the precipitator(s). This value is dependent on total number of ducts,
which is calculated as follows:
n = y_ (Eq. 4.3-5)
where
n = number of ducts
q _ ^Q-j-gi gas volumetric throughput, m3/s
V = gas (treatment) velocity, m/s
h = plate height, m
s = plate spacing, m.
Treatment velocity, V, is a function of resistivity of the fly ash. Values
of V should range from 1.0 to 1.2 m/s in high-resistivity, cold-side ESP ap-
plications, and in low-resistivity applications, hot-side or cold-side. For
most other applications the values should range from 1.0 to 1.5 m/s.
Plate spacing, s, is more or less fixed by the precipitator manufac-
turer and his experience with different types of fly ash, by velocity dis-
tribution across the precipitator, and by plate type. Plate spacing usually
ranges from 15 to 40 cm. Most precipitators in the United States have
spacing of 22.8 cm, but precipitator designers are now showing a great deal
of interest in larger spacings.
Plate height is selected from consideration of simultaneously maintain-
ing the required treatment velocity and also maintaining an adequate aspect
4.3-28
-------
ratio. Plate heights usually range from 7.2 to 14.4 m. The practical
limitation on plate height imposed by structural stability is obvious. Each
manufacturer limits the practical plate heights in accordance with his
overall design.
The width of the box is indicated by the total number of ducts. Cham-
ber (parallel) sectionalization is iiv^the direction across the gas flow,
whereas series sectionalization is in the direction of gas flow (Figure
4.3-12).
4TH SECTION
3RD SECTIOK
2ND SECTION
1ST SECTION
1
o
CO
«r
CD
(DIRECTION OF GAS FLOW
INDICATED BY ARROW)
Figure 4.3-12. Mechanical sectionlization of a precipitator.2
4.3.3.2 Casings. The casing should be of gas-tight weatherproof
construction. Major casing parts are the inlet and outlet transition ducts,
shell and hoppers, inspection doors, and insulator housing. The casing
should be fabricated of materials suitable for the application. The shell
should be reinforced to handle the following: maximum positive or negative
pressure static or dead loads of all components, including any equipment
located on the roof, superstructure weights, hoppers, or dust loads; loads
and movements imposed by connecting flues and dynamic loading caused by
vibrators or rappers; and environmental stresses such as those imposed by
wind, snow, and earthquake.22 In addition, design must provide for the
overall expansion of the casing caused by the high flue gas temperature.
The shell and insulator housing should form a grounded steel chamber com-
pletely enclosing all of the voltage equipment to ensure the safety of
personnel.
4.3.3.3 Dust Hoppers. Hoppers collect the precipitated dust and
deliver it to a common point for discharge. The most common type of hopper
4.3-29
-------
is pyramidal, converging to a round or square discharge. If the dust is to
be removed by screw conveyor, the hopper usually converges to an elongated
opening that runs the length of the conveyor. In applications where the
dust is very sticky and may build up on sloping surfaces, hoppers are not
recommended; the casing is extended to form a flat-bottomed box under the
ESP. The dust is removed by drag conveyors.
Plugging is a major problem with hoppers. Manufacturers have used
designs incorporating vibrators, heaters, poke holes, baffles, large dis-
charge flanges, and steep hopper wall angles (55 to 65°) to reduce these
problems, but this problem persists at certain installations.
The hopper should usually not be used for storage. The trend is toward
larger hoppers so operators can respond to hopper plugging before electrical
grounding occurs or before physical damage is done to the electrodes.
Some manufacturers offer a high-ash fail-safe system that automatically
deenergizes high-voltage equipment if high-ash levels are detected. Some
type of reliable ash level detection is recommended for most hopper designs.
If the preliminary design indicates potential problems with discharge of
ash, the discharge flange should be no less than 25 cm diameter. Transition
from a rectangular hopper to a round outlet should be accomplished without
ledges or projections. Heaters have been found to be especially beneficial
in the discharge throat and up to one-third the height of the hopper. A
low-temperature probe and alarm might also be considered. Hopper installa-
tion and enclosing the hopper areas are beneficial in reducing heat loss in
the hopper and discharge system.
Hopper aspect ratio (height to width) is an important consideration in
minimizing reentrainment caused by gas sneakage to the hoppers. Low aspect
ratio hoppers can be corrected by vertical baffling.
4.3.3.4 Power Supplies. A precipitator power supply consists of four
basic components: a transformer, a high-voltage rectifier, a control ele-
ment, and a control system sensor. The system is designed to provide volt-
age at the highest level possible without causing arc-over (sustained spark-
ing) between the discharge electrode and collection surface.
4.3.3.4.1 Transformer-recti fi ers. The unit converts low-voltage
alternating current to high-voltage unidirectional current suitable for
4.3-30
-------
energizing the precipitator. The transformer-rectifiers and radio-frequency
(RF) choke coils are submerged in a tank filled with a dielectric fluid.
The RF chokes are designed to prevent high-frequency transient voltage
spikes caused by the ESP from damaging the silicon diode rectifiers.23
The T-R sets should be matched to ESP load. The ESP will perform best
when all T-R sets operate at 70 to 100 percent of rated load without exces-
sive sparking or transient disturbances that reduce maximum, continuous-load
voltage and corona power inputs.23 Over a wide range of gas temperatures
and pressures in different applications, practical operating voltages range
from 15 to 80 kV at average corona current densities of 100 to 3200 mA2/1000
m2 (10 to 20 mA/1000 ft2) of collecting area. Over 1500 mA, T-R set inter-
nal impedances are low, which increases the difficulty of achieving stable
automatic control. The highest impedance possible that is commensurate with
the application and performance requirements should be used. This often
means more sectionalization with smaller T-R sets. The high internal imped-
ance of the smaller T-R sets facilitates spark quenching. Smaller electri-
cal sections localize the effects of electrode misalignment and permit
higher voltages in the remaining sections.
In general, current ratings should .increase from inlet to outlet fields
(3 to 5 times for many fly ash precipitators). Typical current voltage
characteristics of a five-field fly ash ESP without ash resistivity problems
are presented in Figure 4.3-13.24
4.3.3.4.2 Subcircuits. During normal operation, optimization of
applied power to the precipitator is accomplished by automatic power con-
trols, which vary the input voltage in response to a signal generated by the
sparkover rate. Although older ESP's used saturable reactors for power
control, modern ESP's use silicon-controlled rectifiers. Provisions are
also included to make the circuit current sensitive to overload and to allow
control in the event that spark level cannot be reached. Although the
circuits may vary among installations, many of the features described below
are common.
4.3.3.4.3 Silicon controlled rectifiers (SCR's). When the circuit
breaker and control circuit on/off switch are closed, power flows through
the current-limiting reactor, current transformer, and current signal trans-
former to the primary of the high-voltage transformer. The SCR's act as a
4.3-31
-------
1075
10
20 30 40.
PRECIPITATOR, M (AVER/kGE)
Figure 4.3-13. Typical fly ash precipitator voltage-current
characteristics, five fields in series,
no ash resistivity problem.24
4.3-32
-------
variable impedance and control the flow of power in the circuit. An SCR is
a three-junction semiconductor device that is normally an open circuit until
an appropriate gate signal is applied to the gate terminal, at which time it
rapidly switches to the conducting state. Its operation is equivalent to
that of a thyroton. The amount of current that flows is controlled by the
forward blocking ability of the SCR's. This blocking ability is controlled
by the firing pulse to the gate of the SCR. The current-limiting reactor
reshapes the current wave form and limits peak current due to sparking.
Current wave form with and without SCR's is illustrated in Figure 4.3-14.
The firing circuit module provides the proper phase-controlled signal
to fire the SCR. The timing of the signal is controlled by (1) the potenti-
ometer built in the module, (2) the signal received by the automatic con-
troller, and (3) the signal received by the spark stabilizer.
The automatic control circuit performs three functions: spark control,
current-limit control, and voltage-limit control.
4.3.3.4.4 Spark control. Spark control is based on storing electrical
pulses in a capacitor for each spark occurring in the precipitator. If the
voltage of the capacitor exceeds the preset reference, an error signal will
phase the mainline SCR's back to a point where the sparking will stop.
Usually this snap-action type of control will tend to overcorrect, resulting
in longer downtime than is desirable. At low sparking rates, about 50
sparks per minute, the overcorrection is more pronounced and causes reduced
voltage for a longer period, with subsequent loss of dust and low effi-
ciency.
Proportional control, another method of spark control, is also based on
storing of electrical pulses for each spark occurring in the precipitator.
The phaseback of the mainline SCR's, however, is proportional to the number
of sparks in the precipitator. The main advantage of proportional control
over spark control is that the precipitator determines its own optimum spark
rate, based on four factors: temperature of the gas, ash resistivity, dust
concentration, and internal condition of the precipitator. In summary, with
proportional spark rate control, the precipitator determines the optimum
operating parameters. With conventional spark control, the operator selects
the operating parameters, which may not be correct. Figure 4.3-15 shows
4.3-33
-------
VQLTAGE-NESATIVE CORONA
'PEAK
AVERA3E
0
f / 27rTIMI:
v /
W-«— AC VOLTAGE WAVE
CURRENT WITHOUT SCR
TT/2
IT
2ir TIME
APPROX.
CURRENT WITH SCR
A.A7
Of IT
I* »
2ir TIME
APPROX.
Figure 4.3-H. ESP current wave form with and without
silicon controlled rectifier$.25
Reprinted with permission of POWER Magazine, copyright McGraw-Hill,
Inc. 1975.
4.3-34
-------
-15
0
g -15
£
a.
I
+15
-IS
Prtct'pitster
current
torelope
Timt perioifs
•Sate output
I70mtt/ittc.
rtsponse
output
Longtirv
n«» cat-tM
output
Figure 4.3-15. Time periods are shown as control system reacts to a spark
impulse F after steady-state operation. Voltage-start ramp is rapid, then
switches to slow until cycle is completed."
Reprinted with permission of POWER Magazine,
copyright McGraw-Hill, Inc. 1975.
4.3-35
-------
voltage current characteristics as an automatic control system reacts to a
spark inpulse.
Some precipltators operate at the maximum voltage or current settings
on the power supply with no sparking. In collection of low-resistivity
dusts, where the electric field and the ash deposit are insufficient to
initiate sparking, the no-spark condition may arise. The fact that the
precipitator is not sparking does not necessarily indicate that the unit is
underpowered. The unit may have sufficient power to provide charging and
electric fields without sparking.
4.3.3.4.5 Voltage-limit control. The voltage-limit control feature of
the automatic control module limits the primary voltage of the high-voltage
transformer to its rating. A transformer across the primary supplies a vol-
tage signal that is compared to the setting of the voltage control, as in
the case of the current limit. The voltage control setting is adjusted for
the primary voltage rating of the high-voltage transformer. When the pri-
mary voltage exceeds this value, a signal is generated that retards the
firing pulse of the firing module and brings the primary voltage back to the
control setting.
4.3.3.4.6 Current-limit control. For current-limit control, a trans-
former in the primary circuit of the high-voltage transformer monitors the
primary current. The voltage from this transformer is compared with the
setting of the current control, which is adjusted to the rating of the
transformer-rectifier unit. If the primary current exceeds the unit's
rating, a signal is generated, as with spark control, which retards the
firing pulse of the firing circuit and this brings the current back to the
current-limit setting.
^ . ',
With all three control functions properly adjusted, the control unit
will energize the precipitator at its optimum or maximum level at all times.
This level will be determined by conditions within the precipitator and will
result in any one of the three automatic control functions operating at its
maximum, i.e., maximum voltage, maximum primary current, or maximum spark
rate. Once one of the three maximum conditions is reached, the automatic
control will prevent any increase in power to reach a second maximum. If
changes within the precipitator so require, the automatic control will
switch from one maximum limit to another.
4.3-36
-------
Other features include secondary overload circuits and an undervoltage
trip capability in the event that the voltage on the primary of the high-
voltage transformer falls below a predetermined level and remains below that
level for a period of time. A time-delay relay is also used to provide a
delay period in the annunciator circuit while the network of contacts is
changing position for circuit stabilization because of an undervoltage
condition.
Panels containing component modules, the SCR power circuit, DC overload
circuits, relays, control transformers, resistors, main contactor, current
transformer, and other components are mounted in the control cabinet and
should be completely accessible for servicing. Positive ventilation for the
control unit is provided by an intake fan located near floor level. Venti-
lating air is exhausted through an opening (grill-protected) in the upper
rear of the control unit.
The transformer enclosure is usually a square metal housing bolted to
the top of the transformer tank. The enclosure protects the transformer
bushings and electrical connections from weather and also ensures, via a key
interlock system, that none of the electric connections or bushings can be
handled until the associated control cabinet has been deenergized and
grounded.
The transformer pipe and guard are used to feed the high-voltage output
of the transformer-rectifier to the support bushings, which in turn are con-
nected to the upper high-tension support frame, from which the discharge
wires are suspended (Figure 4.3-16).
4.3.3.4.7 Electrical energization/sectionalization. The way in which
a precipitator is energized strongly affects its performance. Electrical
energization involves the number and size of the transformer-rectifier (T-R)
sets, the number of electrical sections, half-wave/full-wave (HW/FW) opera-
tion, and changes in the voltage-current characteristics as precipitation
proceeds in the direction of gas flow.
Selection of design power density is often conveniently based on re-
sistivity of the dust. Table 4.3-1 illustrates design values of average
power density as a function of resistivity for the fly ash applications.
4.3-37
-------
SUPPORT INSULATOR
HOUSING
ACCESS
DOOR
VIBRATION
ISOLATORS
HIGH VOLTAGE
BUS DUCT
BUS CONDUCTOR
HIGH VOLTAGE
SWITCH
TRANSFORMER-
RECTIFIER
PROTECTOR
TUBE
DISCHARGE ELECTRODE
SUPPORT FRAME
DISCHARGE ELECTRODE
TENSIONING WEIGHT
WEIGHT GUIDE FRAME
Figure 4.3-3,6. Predpltator charging system and
wire hanging system.26
Courtesy of The Mcllvaine Co., THE ELECTROSTATIC PRECIPITATOR MANUAL.
4.3-38
-------
TABLE 4.3-1. DESIGN POWER DENSITY2
Resistivity,
ohm- cm
104-7
107-8
1Q9_10
10n
1012
>1012
Power density,
w/m2 of
collecting plate
40
30
25
20
15
10
In a cold-side precipitator an average operating voltage may be between
25 and 45 kV for 23 cm spacing, whereas for a hot-side precipitator typical
values range from 20 to 35 kV for 23 cm spacing. Knowing power density and
operating voltage, one can estimate the current density. The density value
of collecting electrode is not constant for each point in the precipitator.
At the inlet section, where the dust loading is greatest, the voltage-
current characteristics differ significantly from those at the outlet, since
the probability of corona suppression is greater at the inlet and the per-
centage of fine particles is greater at the outlet.
Some powering arrangements are shown in Figure 4.3-17 for a variety of
field and cell (chamber) arrangements. The main advantage in splitting a
mechanical section by both chamber and section is to provide greater relia-
bility; this is achieved at an increase in cost.
Reliability of the precipitator relates not only to sectionalization of
a given collection area but also to the addition of collection area or elec-
trical sections. At the discretion of the designer and in accordance with
specifications of the user, the degree of reliability can be defined in
terms of a redundant capacity, which is a function of anticipated failure
and time between maintenance periods. In this context, redundancy may be
defined as that additional area in a precipitator that compensates for the
"normal" level of unavailable collecting area. The degree of additional
collection area will depend upon the application and the manufacturer's
4.3-39
-------
3AS FLOW
1
i
i
PA 11-11
r~"~-
F~"
i
PA. 1
i
i
t
i
i
i
t
PA 1
i-
~l
~®
>-@)
~©
2-12
t
i
1
i
t
~@
~®
-®
3-13
j_ _,
i b
t — . j
I 1
i i
1 r
j-.-r.-r.—
PA 1
PA _
"®
"®
~®
"®
GAS
Q
PA
_~}
i 1
i_ i
PA
i 1
1 IN
D
i 1
1 H
l_ J
PA
i — "i
i 1
i 1
i >-
i 1
1 1
i 1
i 1
4-14 PA
1 — TRAn
1 TRA,
CTCI
r ILL
rn i
FLOW
1
§>i i
i _i
21 -11
@H~~]
®H 1
22-12
^.i 1
®1__j
®c::
®rn
23-13
1
®
-------
experience. To provide a reliable yet cost-competitive design, the designer
must have detailed information on the composition of the particulate and the
physical and chemical parameters of the gas stream.
The highest efficiency of a precipitator is achieved when useful power
input is maximized. The number, size, and mode of operation (half-wave or
full-wave) of the T-R sets can be manipulated to provide the required cur-
rent density within each electrical section of the precipitator. Full-wave
energization involves the powering of two bus sections from a common power
supply, whereas with double-half wave, the two sections are energized
separately. The selection of half-wave or full-wave operation will depend
on the source-specific parameters in the optimization of ESP power input.
In spark-limited operation on a cold-side precipitator treating high-
resistivity ash, half-wave operation allows time during the off half cycle
to recover from the sparking condition (spark quenching). Complete decay of
the charging field and of collection efficiency during the off half cycle is
avoided because of the capacitive effect of high-resistivity fly ash, which
tends to maintain the field potential.
In operation of hot-side precipitators, fly ash resistivity is reduced
by the increase in operation temperature, and the capacitance effect of the
fly ash is reduced. Thus, the charging field decays faster in half-wave
than in full-wave operation.
In summary, the design considerations in sectionalization and energiza-
tion are based on maximizing the power input to the precipitator to achieve
the highest efficiency from a given collection area while minimizing loss of
performance as a result of various potential failures. Reliability of pre-
cipitator performance is a function of application and design experience.
Precipitator energization depends on the sectionalization configuration and
the current density to be supplied to each electrical section, as determined
by chemical and physical characteristics of the dust, dust loading, and
characteristics of the gas stream. The number, size, and mode of operation
of the T-R sets can be fitted to the sectionalized configuration after the
bus section design has been established.
4.3.3.5 Electrode Characteristics.
4.3.3.5.1 Discharge electrodes. Discharge electrodes may be cylin-
drical or square wire, barbed wire, or stamped of formed strips of metal of
4.3-41
-------
various configurations. The geometry of the electrode determines the cur-
rent-voltage characteristics; e.g., the smaller the wire or the more pointed
its surface, the greater the observed value of current for a given voltage.
Discharge electrodes may be suspended from an insulating superstructure
with weights at the bottom holding them tightly in place, or they may be
rigidly mounted on mats or frames. The weighted-wire type must be stabil-
ized against swinging in the gas stream. An example of rigid wire systems
is shown in Figure 4.3-18. the rigid type of discharge electrode system
requires a high degree of quality control during fabrication and erection,
making it more costly than the weighted-wire system. Replacement or repair
is expensive and time consuming. Larger casings are generally required
because of the greater spacing between plates. Two potential problems with
discharge electrodes are summarized below.
4.3.3.5,2 Electrical erosion. In situations where high-current sparks
or continuous sparking must be tolerated, the larger discharge electrodes
will provide more protection against erosion than will smaller sizes.
Shrouds should be included at both the top and bottom of weighted-wire
electrodes, and all interelectrode high-voltage and grounded surfaces should
have smooth surfaces to minimize spark-over. Transformer-rectifier sets
should be matched to precipitator load, and automatic spark controllers
should keep voltage close to the sparking theshold. Contact between the
electrode and the stabilizing frame should be solid to prevent sparking.
With rigid discharge electrodes, substantial reinforcement should be pro-
vided at the point where the electrode is attached to the supportframe, so
that a significant amount of metal must be lost before failure occurs. The
rigid type discharge electrode system has an advantage in that if a wire
does break, there is less chance of it falling against a plate and shorting
out a section of the ESP.
4.3.3.5.3 Mechanical fatigue. Mechanical connections in the discharge
electrode structure should be designed to minimize flexing and reduction in
cross-sectional area at junction points. Connections should be resistant to
vibration and stress, and electrodes should be allowed to rotate slightly at
their mounting points. Reducing the number of welds is also desirable.
Keeping the total unbraced length of electrode as short as possible will
4.3-42
-------
DISCHARGE
ELECTRODE
VIBRATOR
DISCHARGE ELECTRODE
VIBRATOR AND
INSULATOR ASSEMBLY
DISCHARGE
ELECTRODE
VIBRATOR
COLLECTING
ELECTRODE
RAPPER
COLLECTING ELECTRODE RAPPER
AND INSULATOR ASSEMBLY
Figure 4.3-18. Vibrator and rapper assembly and
precipitator high-voltage frame.
4.3-43
-------
help minimize mechanical fatigue. Adequate tension during the construction
process is vitally important for weighted-wire electrodes, and hard spring
wire should be used to prevent kinking.
Schneider et al.22 emphasized that electrode wire failure can be kept
to a minimum provided:
1) Reasonable care is taken during erection in alignment of casings
and surfaces.
2) The support, guide, and stabilizer system is well designed.
3) Reliable, properly adjusted voltage supplies are provided.
4) There is good operating maintenance of the dust handling system.
4.3.3.5.4 Collection electrodes. Collection plates are commercially
available in lengths ranging from 1 to 3 meters and heights from 3 to 15
meters. Generally, these panels are grouped with the precipitator to form
independently rapped collection modules. The total effective length of
these plates divided by their effective height is referred to as the aspect
ratio. Aspect ratios larger than 1.0 provide longer residence time for the
gas and reduce penetration, all other things being equal. Although a varie-
ty of plates are commercially available, their functional characteristics
are not substantially different. Collection plates should be straight and
parallel with the discharge electrodes when assembled. This alignment
depends on care during fabrication, shipping, storage in the field, and
erection.
The ruggedness of the plate support system is also important since in
many designs it must also transmit rapper energy to the plates. Each design
should be examined for its operating limits with various types of rappers.
The effects of vibration and impact loading at all welded points should be
considered. There should also be consideration for adjustment of plate
alignment if necessary after shakedown. Enough spacers should be provided
to maintain alignment and allow for temperature variations.22
Rapper anvils attached to either dust plate supports or rapper header
beams should be durable enough to withstand the stress of rapping and main-
tain alignment (no bending of flanges or other local deformations).
4,3-44
-------
4.3.3.6 Rapper Characteristics.
4,3.3.6.1 Rapper types. For the weighted-wire design rapping impulses
are provided by either single impulse or vibratory type rappers, which are
activated either electrically or pneumatically.
The electromagnetic or pneumatic impulse type rappers are usually
better for collection electrodes and difficult applications, as a vibrator
usually cannot generate sufficient operating energies without being
damaged.28 The magnetic-impulse, gravity-impact rapper is a solenoid elec-
tromagnet consisting of a steel plunger surrounded by a concentric coil,
both enclosed in a watertight steel case. The control unit contains all the
components (except the rapper) needed to distribute and control the power to
the rappers for optimum precipitation.
During normal operation, a dc pulse through the rapper coil supplies
the energy to move the steel plunger. The plunger is raised by the magnetic
field of the coil and then is allowed to fall back and strike a rapper bar,
which is connected to the collecting electrodes within the precipitator.
The shock transmitted to the electrodes dislodges the accumulated dust.
The electrical controls provide separate adjustments so that rappers
can be assembled into different groups and each group adjusted independently
to achieve optimum rapping frequency and intensity. The controls are ad-
justed manually to provide adequate release of dust from collecting plates
while preventing undesirable stack puffing.
In some applications, the magnetic-impulse, gravity-impact rapper is
also used to clean the precipitator discharge wires. In this case the
rapper energy is imparted to the electrode supporting frame in the same
manner, except that an insulator isolates the rapper from the high voltage
of the electrode supporting frame.
The vibrator is also an electromagnetic device—a coil that is ener-
gized by alternating current. Each time the coil is energized, the vibra-
tion is transmitted to the high-tension wire supporting frame and/or collec-
ting plates through a rod. The number of vibrators applied depends on the
number of high-tension frames and/or collecting plates in the system.
The control unit contains all components necessary for operation of the
vibrators, including adjustments for vibration intensity and the length of
the vibration period. Alternating current is supplied to the discharge wire
vibrators through a multiple cam-type timer to provide the sequencing and
time cycle for energization of the vibrators.
4.3-45
-------
The number of rappers, size of rappers, and rapping frequencies vary
with the manufacturer and the nature of the dust. Generally, one rapper
unit is required for 110 to 150 m2 of collecting area. Discharge electrode
rappers serve from 350 to 2000 m of wire per rapper. Intensity of rapping
generally ranges from about 35 to 70 J, and rapping intervals are adjustable
over a range of approximately 50 to 600 s.
For each installation, a certain intensity and time period of vibration
produces the best collecting efficiency. Insufficient intensity in the
discharge vibrators may cause heavy buildup of dust on the discharge wires,
which can lead to the following adverse operating conditions:
It reduces the spark-over distance between the electrodes, thereby
limiting the power input to the precipitator.
It tends to suppress the formation of negative corona and the produc-
tion of unipolar ions required for the precipitator process.
It alters the normal distribution of electrostatic forces in the treat-
ment zone. Unbalanced electrostatic fields can cause the discharge
wires and the high-tension frame to oscillate.
Rigid frame designs generally utilize mechanical hammer rappers. In
these installations, each frame is rapped by one hammer assembly mounted on
a shaft. A low-speed gear motor.is linked to the hammer shaft by a drive
insulator, fork, and linkage assembly. Rapping intensity is governed by the
hammer weight, and rapping frequency is governed by the speed of the shaft
rotation.
4.3.3.7 Solids Removal Equipment. Solids removal from ESP's may be
performed by pressure or vacuum system in large systems such as utility
applications, and by screw conveyor in many smaller industrial applications.
Dust can also be wet sluiced directly from the hoppers. Once conveyed from
the hoppers, the dust can be disposed of dry, or wet sluiced to a holding
pond.
4.3.3.7.1 Removal from the hopper. An air seal is required at each
hopper discharge. Air locks provide a positive seal, although tipping or
air operated slide gate check valves are also used. Heaters, vibrators,
and/or diffusers are frequently considered because bridging in the hoppers
occurs at least occasionally. In trough-type hoppers, a paddle-type con-
veyor has been shown to be the best means of transporting the dust to the
4.3-46
-------
air lock. Dust valves are often oversized to help facilitate removal of
dust from the hopper.
4.3.3.7.2 Pneumatic systems. Figure 4.3-19 shows a typical fly ash
type pneumatic vacuum system. The length of a vacuum system is limited by
the configuration of the discharge system and the altitude above sea level.
Pressure systems are applied when the limits for vacuum systems are exceed-
ed. When the number of hoppers exceeds about 20 and the length of the
system is too great for a vacuum system, combination vacuum pressure systems
may be used.
A vacuum is produced hydraulically or with mechanical vacuum pumps.
Positive displacement blowers are used with pressure systems. Vacuum sys-
tems use electric valves and slide gates, whereas pressure systems use air
locks and slide gates.
Materials of construction are extemely important in selection of a
solids removal system. The chemical composition of the dust and of the
conveying air, and temperatures at various points in the conveying system
should be determined.
When material characteristics are known (material density, particle
size, concentration, and the physical characteristics of the conveying air
or gas) the required conveying velocity can be determined. The design rate
is usually set at 20 percent above the theoretical maximum conveying capac-
ity to avoid plugging.29
Storage facilities for pneumatic dust handling generally are equipped
with cyclones, and often with a fabric filter. The dust is then conditioned
with water and/or a wetting agent and transported by truck or rail to a
disposal site or is mixed with water and pumped to a disposal pond.
4.3.3.8 Gas Distribution Equipment. Proper gas flow distribution is
critical for optimum precipitator performance. Areas of high velocity can
cause erosion and reentrainment of dust from collecting surfaces or can
allow gas to move virtually untreated through the precipitator. Improper
gas flow distribution in ducts leading to the precipitator result in dust
accumulation on surfaces and high pressure losses.
Devices such as turning vanes, diffusers, baffles, and perforated
plates are used to maintain and improve gas flow distribution. A diffuser
4. 3-47
-------
«H*4«» COLIICTOK
Co
I
4*
00
Figure 4.3-19. Typical fly ash type pneumatic vacuum system.29
Courtesy of Allen-Sherman-Hoff Company.
-------
consists of a woven screen or a thin plate with a regular* pattern of small
openings. The effect of a diffuser is to break large-scale turbulence into
many small-scale turbulent zones, which in turn decay rapidly and in a short
distance coalesce into a relatively low-intensity turbulent flow field.30
Two or three diffusers may be used in series to provide better flow than
could be achieved with only one diffusion plate.30 Gas distribution devices
may require rapping for cleaning.
Katz31 stresses the need for uniformity in designing inlet and outlet
nozzles of ESP plenums and their distribution devices. Examples of good and
bad flue and distribution device design for inlet and outlet flues are shown
in Figures 4.3-20 through 4.3-22, respectively.
In multiple-chamber ESPs, louver-type dampers are commonly used for gas
proportioning. At the inlet, however, guillotine shutoff dampers should not
be used for proportioning because they tend to destroy proper gas distribu-
tion to a chamber.22
Gas sneakage through hoppers can be caused by poor gas distribution.
Expansion-type plenums or top entry plenums cause gas vectors to be directed
towards the hopper, and if multiple perforated plates do not fit well in the
lower portion of the plenum or if the lower portion has been cut away be-
cause of dust buildup, gas is channeled into the hoppers.31 Short-circuit-
ing the ESP and/or reentrainment is the end result.
4.3.3.8.1 Gas flow models. Gas flow models are used to determine the
location and configuration of gas flow control devices. Although flow model
studies are not always effective in developing the desired distribution,
they are at least a qualitative indicator of the distribution.
Temperature and dust loading distributions are also important to effi-
cient ESP operations. It is generally assumed that the temperature of the
flue gas is uniform. This is not always true, however, and the effects of
gas temperature on ESP electrical characteristics should be considered.32
Dust loading distributions are not modeled at present for ESP's. It is
assumed that the dust is evenly distributed in the gas, and that as long as
the gas distribution is of a predefined quality, no dust deposition problems
will occur. However, problems such as poor flue design, poor flow patterns
at the inlet nozzle of the ESP plenum, and flow and wall obstructions can
cause dust deposition that gas flow models cannot anticipate.31
4.3-49
-------
A
(3) PERFORATED
DISTRIBUTOR PLATES
(1) PERFORATED
DISTRIBUTOR PLATE
SAS FLOW
RIGHT
:HIGH FLOW:
.LOW FLOW
WRONG
Figure 4.3-20. Effect of two different methods of aas distribution
on flue characteristics in an ESP.30
Courtesy of McGraw-Hi11.
4.3-50
-------
Cm Flow
ELEVATION
IUVATION
\
Figure 4.3-21 . Examples of two inlet plenum designs that
generally cause gas distribution problems.3>
Figure 4.3-22, Expansion inlet plenums showing two methods of spreading
the gas pattern,31
Courtesy of Precipitator Technology, Inc.
4.3-51
-------
4.3.3.9 Instrumentation. Instrumentation necessary for proper moni-
toring of ESP operation can be categorized by location; i.e., T-R sets,
rappers/vibrators, hoppers/dust removal systems, and external items.
4.3.3.9.1 T-R sets. Power input is the most important measure of the
ESP performance. Thus any new ESP should be equipped with the following:
Primary current meters
Primary voltage meters
Secondary current meters
Secondary voltage meters
Spark rate meter (optional)
These meters are considered essential for performance evaluation and
troubleshooting. Figure 4.3-23 shows a typical control cabinet and T-R set
instrumentation.
Data loggers (mainly for digital automatic control systems) are avail-
able to help speed troubleshooting and reduce operating labor. Oscillo-
scopes are also useful in evaluating power supply performance and identify-
ing the type of sparking (multiple burst versus single arc) but there is
little demand for such devices.
It is also possible to use feedback signals from transmissometer, full
hopper detectors, gas conditioning systems, rappers, and suitable process
fault indicators in conjunction with the automatic control unit to provide
optimum performance under all conditions.23 An example is automatic phase-
back of T-R sets when hoppers are overfilled, preventing the burning of
discharge wires.
4.3.3.9.2 Rappers/vibrators. Microprocessor type technology is avail-
able for a high degree of rapper control flexibility and ease of mainten-
ance. For example, in order to prevent control damage from ground faults,
new controls will test each circuit before energizing it. If a ground fault
occurs, the control will automatically bypass the grounded circuit and
indicate the problem on an LED display, thus permitting fast location and
solution of the problem.33
Instrumentation should be used in conjunction with a transmissometer to
help in troubleshooting ESP problems. Separate rapping instrumentation
should be provided for each field. Readings of frequency, intensity, and
4.3-52
-------
Figure 4.3-23. Internal view of one type of rectifier
console showing component parts.
Courtesy of Koppers Co., Inc.
4.3-53
-------
cycle time can be used with T-R set controls to properly set rapper frequen-
cy and intensity, in the case of the weighted-wire electrodes.
For rigid frame mechanical rappers, cycle time and rap frequency of
both internal and external types are easy to measure. Individual operation
of internal rappers is not easily instrumented, nor is intensity control
possible without a shutdown of the ESP.34
i& ••' -
4.3.3.9.3 Hoppers. Instrumentation should be provided for detecting
full hoppers and for operation of the dust valve and dust removal system.
Level detectors can utilize gamma radiation, sound capacitance, pres-
sure differential, or temperature.31 The alarms should be located so that
filling of hoppers does not occur but frequent alarms are avoided. A low-
temperature probe and alarm can be used in conjunction with the level detec-
tor. Control panel lights are used to indicate the operation of hopper
heaters and vibrators.34
Zero motion switches are used on rotary air lock valves to detect mal-
function, as well as on screw conveyors. Pressure switches and alarms are
normally used with pneumatic dust handling systems to detect operating prob-
lems.
4.3.3.9.4 Fans. Fans should be equipped with static pressure gages,
ammeters, and vibration indicators to assist in detecting abnormal operating
conditions.34
4.3.3.10 Gas and Particle Conditioning. In the United States, gas and
particle conditioning agents are used primarily to improve ESP collection
efficiency when they are operated on low-sulfur coal having high-resistivity
fly ash. Some older ESP's designed for higher-sulfur coal usually cannot
operate efficiently enough to meet emission regulations on low-sulfur coal
without flue gas conditioning. Other alternatives are increasing the size
of the ESP, cooling the gas below the design temperature of the ESP, or
switching to a hot-side ESP. Some hot-side ESP's appear to require chemical
conditioning in order to perform as designed.
4.3.3.10.1 Conditioning agents and mechanisms. The compounds other
than water that are now used or under study as conditioning agents are
sulfur trioxide, sulfuric acid, ammonia, ammonium sulfate, triethyl amine,
4.3-54
-------
compounds of sodium, and compounds of transition metals. Table 4.3-2 sum-
marizes these conditioning agents and their mechanisms of operation.
TABLE 4.3-2. REACTION MECHANISMS OF MAJOR CONDITIONING AGENTS
Conditioning agent
Mechanism(s) of action
Sulfur trioxide and sulfuric acid
Ammonia
Ammonium sulfate
Triethyl amine
Sodium compounds
Compounds of transition metals
Potassium sulfate and sodium chloride
Condensation or adsorption on fly ash
surfaces; may also increase cohesive-
ness of fly ash. Reduce resistivity.
Mechanism is not clear; various ones
proposed:
Resistivity modification
Increase in ash cohesiveness
Enhances space charge effect.
Little known about the actual mechan-
ism; claims made for the following:
Resistivity modification
Increase in ash cohesiveness
Enhances space charge effect.
Experimental data lacking to substan-
tiate which of the above is predomi-
nant.
Particle agglomeration claimed; no
supporting data.
Natural conditioner if added with
coal
Resistivity modifier injected into
gas stream.
Postulated that they catalyze oxida-
tion of S02 to S03; no definitive
tests with fly ash verify this
postulation.
In cement and lime kiln ESP's:
Resistivity modifiers in the
gas stream
NaCl: natural conditioner when
mixed with coal.
4.3-55
-------
Sulfur trio'xlde and sulfuric acid are the most widely used conditioning
agents in the United States. The primary mechanism is condensation or
adsorption on ash. Handling of both of these compounds is difficult because
they are highly corrosive and toxic liquids that must be vaporized before
injection to the flue gas. A typical S03 conditioning system is illustrated
in Figure 4.3-24.
Ammonia is less widely used in the U.S. mainly because of its incon-
sistent rate and lack of a clear indication of the mechanism by which it
acts.35 In some cases, ammonia shows a significant resistivity reduction,
while in other cases it does not.
Ammonium sulfate is believed to be a main component of many commercial
formulations presently used.35 It is thermally decomposed to ammonia, sul-
fur trioxide, and water after being injected into the flue gas at high tem-
perature about 600°C (355°F) and reforms ammonium sulfate as the tempera-
ture decreases. Although a number of mechanisms have been claimed to be
responsible for the action of ammonium sulfate, experimental test data are
not available to document which mechanism is predominant.35
Triethylamine is not being used in the U.S. but other proprietary form-
ulas used here claim to have the same operating mechanisms, namely, agglom-
eration. There is doubt as to whether agglomeration actually occurs since
no tests that measure a change in particle size distribution with triethyla-
mine have been conducted.
Sodium compounds have been used for both cold- and hot-side ESP's, and
Interest in these compounds stems from their role as a natural substance in
reducing resistivity in coal ash. Sodium can be added either with the coal
fed into the boiler or by conventional means as a solid powder or aqueous
solution in the gas stream. The 1-atter would probably result in a reduction
in the resistivity since it is co-precipitated with the ash rather than
being chemically incorporated into it. Full-scale tests of sodium (soda
ash) both in the lab and in the field have shown it is the conditioning
agent of choice for hot-side ESP's.36 A solution type injection appeared to
offer the most effective utilization of the chemical.
Tests with anhydrous sodium carbonate injected into the flue gas pre-
ceding a cold-side pilot ESP treating ash from a boiler firing low-sulfur
coal showed a sixfold reduction in ash resistivity, an improvement in
4.3-56
-------
LIQUID SULFUR
LIQUID SULFUR
STORAGE
METERING
PUMP
CONTROLLED TO
8Q00-825°F
CONDITIONED
FLUE GAS TO
PRECIPITATOR
U.S. Patent No. 3.993.429
Figure 4.3-24. Flow diagram of sulfur burning
flue gas conditioning system.
Courtesy of Robert L. Reveley (Vice President Wahlco, Inc.)
4.3-57
-------
achievable current' densities, enhancement of particulate collection effi-
ciencies, and an improvement in fractional efficiency collection capability
(see Figure 4.3-25).
Vanadium pentoxide and ferric sulfate have been added to flue gas on an
experimental basis as potential catalysts for the formation of S03, but
neither has given definitive results.
Potassium sulfate, a water-soluble alkalai, is another compound that
has been tested in full-scale applications as a conditioning agent for
preheating kilns in Brazil in addition to a conditioning tower.38 An aque-
ous solution of potassium (0.4% increase in water-soluble K) reduced
resistivity from 1013 to 1011 ohm-cm and improved ESP performance. Addi-
tional tests with potassium sulfate and sodium chloride at an ESP installa-
tion on a coal-fired lime kiln in South Africa yielded similar results.38
Sodium chloride is not recommended as a conditioning agent for the cement
industry because of the adverse effect of even small amounts of chloride on
kiln clinker quality. Addition of dry sodium chloride to the coal being
ground for firing in a Danish lime kiln led to considerable improvement in
ESP performance.38
In summary, flue gas conditioning is often successful in improving ESP
performance by reducing dust resistivity or through other mechanisms.
Conditioning agents should not be seen as a cure-all for ESP problems,
however; they cannot correct problems associated with an underpowered or
undersized ESP, poor gas distribution, misaligned plates and wires, or
inadequate rapping. Thus, any existing installation should be carefully
evaluated to determine that poor ESP performance is due only to high resis-
tivity and not the above-mentioned problems. Conditions for injection of a
chemical conditioning agent should also be carefully studied. Inadequate
mixing of the conditioner can result in performance below that expected.
4.3.4 Operation and Maintenance of Electrostatic Precipitators
4.3.4.1 Safety Considerations. An important aspect of ESP operation
and maintenance is the safety of personnel. Besides the problems and pre-
cautions necessary because of confined area entry, the dangers of high volt-
age must be considered. Even after the power supply has been shut down, a
residual charge may be retained by the ESP, which behaves much like a large
4.3-58
-------
5
4
o
I—I
S 1-0
o 0.8
|«Mt
1 0.6
§
0.4
0.2
FLYASH
EhFLYASH SODIUM
I
I
III
1111
I I
.00 .60 .40 .30 .20 .10 .05 .02 .01 .004 .001
PENETRATION
Figure 4.3-25. Comparison of mean penetration results.37
4.3-59
-------
capacitor. Prior to entry of personnel into the ESP all safety interlock
procedures should be followed so that the ESP, is properly grounded and
electrically discharged.
4.3.4.2 Modes of Failure in ESP's. Numerous mechanical and electrical
failures can occur in ESP's. Often these failures have synergistic effects
that cause other malfunctions. Generally the modes of failure are either
mechanical or electrical, although the symptoms of a malfunction may be
manifested as both.
The most common problems associated with ESP malfunctions are discharge
wire breakage, plugged hoppers causing excess buildup of material, and
failure of rappers or vibrators. Other problems are insulator failures,
inadequate electrical energization, and changes in the process operation
away from the specified design criteria. A brief discussion of the common
failures and their effects on emissions is presented below.
4.3,4.2.1 Discharge wires. One of the most common problems associated
with suspended wire electrode ESP's is wire breakage, which typically causes
an electrical short circuit between the high-tension discharge wire system
and the grounded collection plate. The electrical short trips the circuit
breaker and disables a section of the ESP, which remains disabled until the
broken discharge wire is removed from the unit.
Electrical erosion, the predominant cause of failures, occurs when re-
peated electrical sparkovers or arcs occur in a localized region. Heating
and vaporization of a minute quantity of metal occur with each spark.
Sparkover at random locations will cause no serious degradation of the
discharge electrode. Repeated sparkover at the same location, however, can
remove significant quantities of material, with subsequent reduction of
cross-sectional area and ultimate failure at that point.39
Localized sparking can be caused by misalignment of the discharge elec-
trode during construction or by variations in the electric field resulting
from "edge" effects of adjacent discharge and collection electrodes at the
top and bottom of the plates. Measures that will eliminate failure at these
points are adding shrouds and providing a round surface at the edge of the
collection electrode to reduce the tendency for sparking.39
..*V , ••
Electrical erosion can also be caused by "swinging" of electrodes,
which can occur when the mechanical resonance frequency of the discharge
4.3-60
-------
wire and weight system is harmonically related to the electrical frequency
of the power supply. The power supply adds energy to the swinging wire, and
sparking occurs with each close approach to the collection plate. This
action leads to erosion of the electrode and mechanical failure.39
Poor workmanship during construction can also cause electrical failure
of the discharge electrode. If pieces of the welding electrode remain at-
tached to the collection plate, localized deformation of the electric field
can lead to sparking and'-failure of the discharge electrode.
Mechanical fatigue occurs at points where wires are twisted together,
and where mechanical motion occurs continually at one location. This occurs
at the top of a discharge electrode where the wire is twisted around the
support collar. Methods of reducing mechanical fatigue include selection of
discharge electrode material that is resistant to cold work annealing after
attachment.
Chemical attack is caused by a corrosive material in the flue gas,
which can occur when high-sulfur coal is burned and flue gas exit tempera-
tures are low and near the acid dew point. Use of ambient air to purge
insulator compartments also can cause the temperature to drop below the acid
dew point in a localized region. Corrosion can be minimized by operation at
higher flue gas temperatures or by use of hot, dry air to purge insulator
compartments. Use of good insulation around the ESP shell to maintain high
temperatures also provides adequate protection within the usual range of
operating temperatures and fuel sulfur contents.39
Failure of some discharge wires can be expected in all ESP's, although
many rigid-frame units experience many fewer problems with discharge wire
breakage than do weighted-wire designs. If it occurs in a random manner,
this wire breakage will not significantly degrade ESP performance. After a
number of wires have broken and have been removed, they will have to be
replaced during some scheduled outage. It is important that wire breakage
does not occur excessively in any given gas passage between any two plates.
Excessive wire breakage will result in ineffective charging and collection
of particulate in the passage missing many wires.
4.3.4.2.2 Particulate removal system. Failure of the particulate
discharge and removal system will allow material to build up in the gas
treatment zone. This buildup of material can cause misalignment of both the
4.3-61
-------
collection plates and the discharge electrodes. Arcing between the wires
and plates will usually occur. In some applications, collection plates can
be warped, as well as discharge wire frames in rigid-frame designs. An
outage will be required for repair and realignment of the components since
the electrical section would be rendered useless. Buildup of materials also
can lead to the formation of clinker-like material between the discharge
electrodes and the collection electrodes. This clinker-like material is
formed by fusion of the dust when the high voltage of the ESP is passed
through it.
Proper design of particulate discharge and removal systems provides the
best means of preventing problems. Particulate characteristics such as bulk
density, flow characteristics, and agglomeration should be considered in the
initial design. Materials are generally more free flowing when hot, and
efforts to maintain the wall temperature of the hopper above 200°C will help
reduce hopper problems.61 Insulation of the entire hopper, and windbreaks
around the hopper area can add to the available heat in the hopper area.
Heaters should be installed from the bottom apex to at least 2 meters high
on hoppers on the inlet sections of the ESP and 1 meter high on the outlet
fields. Ratings of 6 to 10 W per m2 of hopper surface area should be satis-
factory.31 Reduction of inleakage through hopper access doors and cleanout
ports during and transfer from the hopper to the conveyor mechanism will
reduce the effects of cooling and condensation in the hopper as well as
reducing the reentrainment of collected dust.
Hoppers often contain baffle plates to prevent gas sneakage or by-
passing of the gas stream through the hoppers. Where the baffle plates are
close to hopper works, a "bridge" of material can be formed causing a build-
up of dust into the gas treatment zone. This problem can be avoided in the
design stage.
The use of vibrators on hoppers may cause as many problems as they pre-
vent. If the particulate tends to pack and agglomerate, vibrators may cause
hopper bridging rather than preventing it. The feasibility of using hopper
vibrators should be decided for each individual application.
4.3.4.2.3 Rappers. In dry removal systems, the dust must be removed
from the collection surface periodically. Effective rapping depends upon
4.3-62
-------
agglomeration of the material on the plate to minimize reentrainment of the
dust. The key to rapping is to avoid excessive rapping.31
4.3.4.2.4 Insulator/bushings. Suspension insulators support and
isolate the high-voltage parts of an ESP. As mentioned earlier, inadequate
pressurization of the top housing of the insulators can cause ash deposits
or moisture condensation on the bushings, which may cause electrical break-
down at the typical operating potential of 45 kV dc.
Corrective or preventive measures include inspection of fans that
ventilate the top housing, availability of a spare fan for emergencies, and
heating of insulators to prevent condensation.
4.3.4.2.5 Electrical energization. Electrical energization must be
adequate to charge the particles, maintain the electric field, and hold the
collected dust to the collection plates. Among several possible causes of
failure to achieve the required level of power input to the ESP, the follow-
ing are most common:39
0 High dust resistivity
. ° Excessive dust accumulation on the electrodes
0 Unusually fine particle size
0 Inadequate sectionalization
0 Improper rectifier and control operation
0 Misalignment of electrodes
0 Inadequate power supply range.
If a precipitator is operating at a spark-rate-limited condition but current
and voltage are low, the problem can commonly be traced to high-resistivity
dust, electrode misalignment, or uneven corona resulting from buildup on the
discharge electrode. The effects of high resistivity are discussed in more
detail in Section 4.3.2 in terms of conditions specific to utility industry,
where resistivity presents the greatest problem.
Failures in ESP controls can prevent the system from achieving the
level of power required for normal operation. Following are the most common
malfunctions in controls:
1. Power failure in the primary system
2. Transformer or rectifier failure in secondary system caused by:
a) Insulation breakdown in transformer
4.3-63
-------
b) Arcing in transformer between high-voltage switch contacts
c) Leaks or shorts in high-voltage structure
d) Contamination of the insulating field.
The most effective measure for correction of control failures is a good
maintenance program in which the controls are checked periodically for
proper operation. A daily log of instruments that register current, volt-
age, and spark rate can also indicate potential proolems.
4.3.4.3 Preventive Maintenance Schedule for ESP's.
4.3.4.3.1 Daily maintenance. An accurate log should be kept on all
aspects of precipitator operation including electrical data, changes in
rapper and vibrator operation, fuel quality, and process operations. Such a
log can provide information for diagnostic troubleshooting when any change
in performance occurs. For example, it is obvious that gross departures
from normal readings on the T-R meter and transmissometer indicate trouble.
It is not so widely recognized that small variations, often too slight to be
noticed without checking daily readings, can indicate impending trouble.39
Problems that usually affect precipitator performance gradually, rather
than suddenly, include (1) air inleakage at heaters or in ducts leading to
the precipitator, (2) dust buildup on precipitator internals, and (3) de-
terioration of electronic control components. Such problems are often indi-
cated by a slight but definite drift of daily meter readings away from base-
line values.1
An operator should usually not try to correct deviant meter readings by
adjusting control set points. An automatic control response range should
accommodate normal variations in load. When major changes occur, such as
would result from firing a coal substantially different from that for which
the precipitator was designed, the precipitator manufacturer should be
called in to retune the installation.39 If no such major changes have
occurred, then variant meter readings indicate problems that must be
detected and corrected. Figure 4.3-26 illustrates a log of electrical
readings that are checked several times at a coal-fired utility boiler
installation. These readings are used in troubleshooting ESP operations.
Probably 50 percent of all electrical set tripouts are caused by ash
buildup. Short of set tripout, buildup above the top of hoppers can cause
4.3-64
-------
PHECiriTATOR LOG SMiET
NO
y.
K
*M
fct
*.
8
NO
to
VI
w
I*J
1 PREC'PITATOR
(vpTn CASLE ) KV
AtfG SPARK RATE
P*»TH CABLE 2 *v
TRAMS PRi VOLTS V
TRJttvS PR> CURRENT
PPTR CABvE ! KV
A^G $P*RK R*TC
PPtR C*Bi-E 2 KV
TRftNS PRi VOLTS V
PPTR AVG CURRENT MA
TRANS PR' CURRENT
PPTR CABLE 1 KV
AVG SI^ARK RATE
PPTR CABLE 2 KV
TRANS PRi VOLTS V
PPTR AVG CURRENT MA
IRAN 5 PRi CURRENT
PPTfi CABLE" ' KV
AVO SPARK RATE
PPTR CABLE 2 XV
TRAM$ Pfi* VOI.TS V
PPTft *vs CURRENT MA
TRANS PR! CURRENT
£ PRCOPiTATQR
PPTR CA0tC 1 KV
*VO SPARK RATE
PPTR CABLE 2 *v
TRANS PHI VOLTS V
PPTR AVG CURRENT MA
TRANS PRi CURRENT
PPTR CABLE t KV
AVG 5»ARK RATE
PPTR CABLE 2KV
TRAN5 PRi VOLTS V
PPTR AVG CURRENT MA
TRAMS PRJ CURRENT
PPTR CABU t KV
AVG SPAR* RATE
PPTR CABLE 2 KV
fRANS PHl VQLlS V
PPTft AVG CURRENT MA
TRANS PRI CURRENT
PPTR CA&.C 1 KV
AVG SPARK RJtlf
PPTR CABLE 2 KV
TRAWL PRi VLLTS V
PPTR AVO CURRCNT MA
TRANS PR' CURRENT
12 M10
JAM
6AM
9AM
!2N0D*t
3PM
6PM
9PM
Figure 4.3-26. Precipitator log sheet.
4.3-65
-------
excessive sparking that erodes discharge electrodes.1 Further, the forces
created by growing ash pil'es can push internal components out of position,
causing misalignment that may drastically affect performance. Sometimes
operators attempt to preserve alignment by welding braces to hold collec-
ting-electrode plates in position. This practice may be inadvisable because
restraining the plates reduces the effectiveness of the rapping action that
keeps them clean.39
Although various indicators and alarms can be installed to warn of
hopper-ash buildup and of ash-conveyor stoppage, the operator can double
check by testing temperature at the throat of the hopper. If the tempera-
ture of one or more hoppers seems comparatively low, the hopper heaters may
not be functioning properly. Generally, however, low temperature indicates
that hot ash is not flowing through the hopper and that bridging, plugging,
or failure of an automatic dump valve has held ash in the hopper long enough
for it to cool. The ash subsequently will build up to the top of the hop-
per.39
If the temperatures of all hoppers seem low, the ash-conveyor system
should be checked; the system may have stopped or dust agglomeration may be
so great that the conveyor can no longer handle all of the fly ash.
Hopper plugging is sometimes caused by low flue gas temperature, which
permits moisture condensation. The temperature of the gas entering the ESP
may be too low, or ambient air may be leaking into the flue gas duct.
Hoppers are particularly prone to plugging during startup after an outage,
when they are cold and usually damp.
Daily checking of the control room ventilation system minimizes the
possibility of overheated control components, which can cause the control
set points to drift and can accelerate deterioration of sensitive solid-
state devices.
A daily check of all hopper and ESP access doors to detect gas irrleak-
age is recommended. Gas inleakage can cause excessive sparking, corrosion,
and particle reentrainment.
4.3.4.3.2 Weekly maintenance. Rapper solenoid-coil failures, fairly
common during the period when high voltage was used, are rare with modern
low-voltage equipment.39 Still, a weekly check of all units is advisable.
Rapper action should be observed visually, and vibrator operation confirmed
4.3-66
-------
by touch. If inadequate rapping force is suspected, an accelerometer
mounted on the plates could be used to verify that rapping acceleration is
adequate (often, up to 30 G is required). This is best done on a pretest
check.
Control sets must be checked internally for deposits of dirt that may
have penetrated the control cabinet filter. Accumulation of dirt can cause
false control signals and can damage such large components as contactors and
printed circuits.
Finally, filters in the lines supplying air to control cabinets and to
the precipitator top housing should be checked and cleaned if necessary to
prevent plugging.39
4.3.4.3.3 Monthly maintenance. Most new precipitators incorporate
pressurized top housings that enclose the bushings through which high-
voltage connections are made to the discharge electrodes within the precip-
itator box.
Pressurization ensures that if gas inleakage occurs where the bushings
penetrate the precipitator hot roof, gas will flow into the precipitator
rather than out from it. Leakage from the precipitator into the housing
could cause ash deposits or moisture condensation on the bushings, with risk
of electrical breakdown at the typical operating potential of 45 kV dc.
Monthly maintenance also should include inspection of bushings visually
and by touch for component vibration, checks of differential pressure to en-
sure good operation of the fan that pressurizes the housing, and manual
operation of the automatic standby fan to make sure it is service-ready.
4.3.4.3.4 Quarterly maintenance. Quarterly maintenance includes
inspection of electrical-distribution contact surfaces. These should be
cleaned and dressed and the pivots should be lubricated quarterly if not
more frequently, since faulty contacts could cause false signals,39 Fur-
ther, because transmissometer calibration is subject to drift, calibration
should be verified to prevent false indications of precipitator performance.
4.3.4.3.5 Semi-annual maintenance. Inspection, cleaning, and lubrica-
tion of hinges and test connections should be performed semi-annually. If
this task is neglected, extensive effort eventually will be required to free
test connections and access doors, often involving expensive downtime.
4.3-67
-------
Performance tests may be required at any time; they should not be delayed
while connections are made usable. An effective preventive measure is to
recess fittings below the insulation.39
Inspection of the exterior for corrosion, loose insulation, surface
damage, and loose joints can identify problems while repair is still possi-
ble. Special attention should be given to points at which gas can leak out
as fugitive emissions.39
4.3.4.3.6 Annual maintenance. Scheduled outages must be long enough
to allow thorough internal inspection of the precipitator. Following is a
summary of items to be checked during an annual inspection, abstracted
primarily from Reference 40.
1. Dust accumulation—The upper outside corners of a hopper usually
show the greatest accumulation. A spotlight can be used to check for dust
buildup, eliminating the need to enter the hopper.
2- Corrosion—Inaccessible parts of the ESP are often attacked by
corrosion. Access doors and frames, which are difficult to insulate, are
usually attacked first. Condensation can occur in penthouses that contain
support insulators; the penthouses are at lower temperatures than the gas,
and moisture is added also by purge air from the outside.
Corrosion can occur at several places in the ESP housing—the underside
of roof plates, the outside wall, the space between outside collecting sur-
face plates and sidewall, the back of external stiffening members that act
as heat sinks, and any area not continuously subject to gas flow such as
corners and the upper portion of the hopper connection to inlet and outlet
ducts. All gas connections should be checked for inleakage of oil, gas, or
air.
Corrosion in these areas can be minimized by keeping interior surfaces
hot and by effective thermal insulation of outside surfaces. Use of heaters
during routine shutdowns or operation at low loads also may help prevent
corrosion.
^' Rappers—Maintenance of the magnetic-impulse, gravity-impact
rapper has been discussed. Many of the rigid-wire ESP's, however, have
mechanical rappers. The drives for collecting and discharge electrode
rappers should be checked for high motor temperature, unusual noise, and
level and condition of the lubricant.
4.3-68
-------
Mechanical rappers should be checked for excessive wear, shifting of
point of impact, free movement of wire-frame rapper release, free movement
of hammers, and wear on hammer shaft bushings.
^' Hoppers—On both weighted-wire and rigid-wire precipitators, the
hopper discharge should be checked for such objects as broken pieces of rap-
pers, wires, and scale. Presence of foreign objects indicates a problem
that should be investigated further.
5. Gas distribution plates—Although perforated plates usually do not
become plugged, uneven distribution can sometimes cause plugging of a por-
tion of the plates. If a rapping system is not used, manual cleaning is
required.
6. Discharge electrodes—Frames in rigid-frame, discharge-electrode
precipitators should be centered between two rows of collecting surface
plates with a maximum deviation of +0.6 cm. Discharge wires must be
straight and securely connected to the discharge frame.
Weighted-wire precipitators should be checked for missing or dropped
weights. Removal of a broken wire that is not replaced should be recorded
on a permanent log sheet.2 In addition the location of broken wires,
location on the wire where the break occurred, and the cause of the break,
(erosion, corrosion, etc.) should be recorded. Discharge wires should be
cleaned manually as required.
7. Collecting electrodes—Collection plates should be inspected for
warping due to excessive heat or hopper plugging. Corrosion of lower por-
tions of the plates and portions of plates adjacent to door openings indi-
cates air inleakage through hoppers or around doors,2 Plates should be
cleaned manually as required.
8. Suspension insulators—When insulators become heavily coated with
moisture and dust, they may become conductive and crack under high-voltage
stress. Cracks can be spotted with a bright light during inspection.
Faulty insulators can cause excessive sparking and voltage loss and can fail
abruptly or even explode if allowed to deteriorate.
9- Housing—Thick dust deposits on interiors of housings indicate
high gas velocities resulting from excessive gas volumes, a condition that
should be corrected.
4.3-69
-------
4.3.4.3.7 Situational maintenance. Certain preventive maintenance and
safety checks are so important that they should be performed during any
outage of sufficient length, without waiting for scheduled downtime. Air
load readings should be compared with baseline values to detect possible
deterioration in performance. Readings taken immediately upon restoring the
precipitator to service can serve as a check on any changes resulting from
maintenance done during the outage. All maintenance performed on this
situational basis should be recorded for diagnostic purposes.
Critical internal alignments should be checked whenever an outage al-
lows; any misalignment warrants immediate corrective action. Interiors of
control cabinets and top housing should be checked during any outage of 24
hours or more and cleaned if necessary. Any outage of more than 72 hours
provides an opportunity to check grounding devices, alarms, interlocks, and
other safety equipment and to clean insulators and bushings.39
4.3.4.4 Optimization of Performance and Energy Consumption.
Preoperational checklist. Before placing the equipment in operation,
plant personnel should perform a thorough check and visually inspect the
system components in accordance with the manufacturer's recommendations.
Some of the major items that should be checked are summarized below:
Control Unit
Proper connections to control
Silicon Rectifier Unit
Rectifier-transformer insulating liquid level
Rectifier ground switch operation
Rectifier high-voltage connections
High-voltage bus transfer switch operation
High-Tension Connections
High-tension bus duct
Proper installation
Installation of vent ports
Equipment Grounding
Precipitator grounded
Transformer grounded
Rectifier controls grounded
4.3-70
-------
High-tension guard grounded
Conduits grounded
Rapper and vibrator ground jumpers in place
4,3.4.4.1 Air load tests. After the precipitator is inspected (i.e.,
preoperational check adjustment of the rectifier control and check of safety
features), the air load test is performed. Air load is defined as energi-
zation of the precipitator with minimum flow of air (stack draft) through
the precipitator. Before introduction of an air load or gas load (i.e.,
entrance of dust-laden gas into the precipitator), the following components
should be energized:
Collecting plate rappers
Perforated distribution plate rappers
High-tension discharge electrode vibrators
Bushing heaters - housing/compartments
Hopper heaters - vibrators - level indicators
Transformer rectifier
Rectifier control units
Ventilation and forced-draft fans
Ash conveying system
The purpose of the air load test is to establish reference readings for
future operations, to check operation of electrical equipment, and to detect
any improper wire clearances or grounds not detected during preparation in-
spection. Air load data are taken with the internal metal surfaces clean.
The data consist of current-voltage characteristics at intervals of roughly
10 percent of the T-R mi Hi amp rating, gas flow rate, gas temperatures, and
relative humidity.
For an air load test, the precipitator is energized on manual control.
The electrical characteristics of a precipitator are such that no sparking
should occur. If sparking does occur, an internal inspection must be made
to determine the cause. Usually, the cause is (1) close electrical clear-
ances and/or (2) the presence of foreign matter that has been left inside
the precipitator.
After the precipitator has been in operation for some time, it may be
necessary to shut it down to perform internal inspections. At such times,
4.3-71
-------
it would be of interest to take air load data for comparison with the
original readings.
4.3.4.4.2 Gas load tests. The operation of a precipitator on gas load
differs considerably from operation on air load with respect to voltage and
current relationships. The condition of high current and low voltage char-
acterizes the air load, whereas low current and high voltage characterize
the gas load. This effect governs the operation of the precipitator and the
final setting of the electrical equipment.
4.3.4.5 ESP Startup Procedures. The exact startup procedures for any
ESP are dependent upon the application and the manufacturers' recommended
procedures. The following general guidelines are usually applicable.
Prior to startup it should be confirmed that the hoppers are empty
before closing all hatches. Any lumps of material that may impede the flow
of dust should be removed. Several hours before process startup the hopper
heater should be turned on so that hoppers are warm.
Insulator surfaces should be heated by early startup of insulator
heaters or by the purge air system to prevent electrical tracking over the
surface of the insulators.
The point at which the ESP is energized will depend upon the individual
application. Some processes can be started up with the ESP fully energized,
whereas others have explosive gas compositions that must be considered.
Generally, the temperature at each field is higher than the moisture dew
point before that field is energized to minimize sparking.
It is usually recommended that the ESP be energized under manual con-
trol to reduce spark-over during startup and that the power supplied to each
section be increased gradually. Newer automatic controls that sense the
spark-over threshold can be set to spark-limit mode and placed on automatic
control. As the temperature and the process stabilize, the controller will
bring the power level to its maximum control level.
Rappers should be turned off through the process startup and probably
for a period of time after the ESP is energized to allow formation of an
adequate layer of dust on the plates. The amount of time between startup
and rapper initiation will depend upon the application and on previous
experience.
4.3-72
-------
4.3.4.6 Normal Operation. Although electrical portions of a precipi-
tator require very little maintenance, the items enumerated below should be
attended regularly if the equipment is to give optimum service. It is
considered good practice to assign one plant operator on each shift the task
of checking and recording data on electrical equipment at the start of the
shift.
The cycle of inspection and maintenance during normal operation
includes the following components:
T-R sets and associated equipment and controls
Transformer enclosure
Pipe and guard
Vibrators
Plate rappers
Top housing
Insulator compartments
Upper high tension frame
Discharge wires
Collecting plates
Lower precipitator steadying frame
Dust collection point (dry or wet bottom}
Hoppers and screw conveyors
Precipitator shell.
4.3.4.6.1 Using T-R set meters for troubleshooting. An operator can
utilize the meters to aid in diagnosing other problems with an ESP. General
examples of the effect of changing conditions in the gas stream and within
the ESP on control set meters are presented below:40
1. When the gas temperature increases, the voltage will increase and
the current will decrease. Arcing can develop. When the gas temperature
decreases, the voltage will decrease and the current will increase.
2. When the moisture content of the gases increases for any given
condition, the current and voltage will also tend to increase in value.
3. If reduced voltage occurs because of a spark-over, a rise in mois-
ture may allow for an increase in the precipitator voltage level.
4. An increase in the concentration of the particulate will tend to
elevate voltages and reduce current flow.
4.3-73
-------
5. A decrease in the particle size will tend to raise voltage while
suppressing current flow.
6. A higher gas velocity through the precipitator will tend to raise
voltages and depress currents.
7. Air inleakage may cause spark-over in localized areas, resulting
in reduced voltages.
8. A number of precipitator fields in series will show varying read-
ings, with voltage-current ratio decreasing in the direction of gas flow.
9. If a hopper fills with dust causing a short, the voltage will be
drastically reduced and the current will increase.
10. If a discharge electrode breaks, violent arcing can be observed,
with the meters swinging between zero and normal. ,
11. If a transformer-rectifier unit shorts, voltage will be zero at a
high current reading.
12. If a discharge system rapper fails, the discharge wires build up
with dust; the voltage increases to maintain the same current level.
13. If a plate rapper fails, the voltage decreases to maintain a cur-
rent level under sparking conditions.
Table 4.3-3 presents specific examples of the effect of changing condi-
tions on ESP control set readings. The operator can use T-R set readings
along with other systematic inspection procedures to optimize performance of
an ESP.
4.3.4.6.2 Use of ESP performance curves. Although ESP collection
efficiency is reduced by malfunctions such as breakage of discharge wires
and deterioration of power supply components, rectifiers, insulators, and
similar equipment, a unit can often be kept in compliance with particulate
emission regulations by reducing its load. Figure 4.3-27 (top graph)
illustrates collection efficiency of a four-field utility ESP with 24 bus
sections as a function of the gross boiler load, depending on the number of
bus sections out and whether they are in series or parallel. The bottom
graph shows the efficiency needed by the ESP to meet a state regulation of
0.38 lb/106 Btu as a function of the ash content of coal (assuming a heating
i1
value of 11,000 Btu/lb).
These types of graphs can be helpful to the operator. Knowing the ash
content of the coal he is firing and knowing which bus sections of his ESP
4.3-74
-------
TABLE 4.3-3. EXAMPLE EFFECTS OF CHANGES IN FORMAL OPERATION ON ESP CONTROL SET READINGS7
01
Condition
1.
2.
3,
4.
5.
6,
7.
8.
8.
10.
Normal full load
System load fall by 1/2
System load constant, but
increase in dust load
Gas temperature Increases
Gas temperature decreases
ESP hopper fills with dust
Discharge electrode breaks
Transformer rectifier shorts
Discharge system rapper falls
Collection plate rapper fails
Effect
Gas volume and dust concentra-
tion decrease, resistance
decreases
Resistance Increases
Resistance rises, sparking
increases because of increased
resistivity
Resistance decreases
Resistance decreases
Resistance may fall to 0 (may
vary between 0 and normal if
top part of electrode is left'
swinging Inside the ESP).
Violent instrument fluctuations.
Arcing can be heard outside the
ESP.
No current passes from T-R set
to the ESP
Dust builds up on discharge
electrodes. Resistance Increases
because corona discharge decreases
Additional voltage required to
keep current constant.
Sparking increases. Voltage
must be reduced to keep current
constant
Primary
voltage
V, a.c.
300
260
350
300-350
280
180
0-300
0
330
265
Primary
current
A, a.c.
50
51
40
50-60
52
85
0-50
100+
50
50
Secondary
current
mA, d.c.
_
i
200
230
175
20-250
210
300
0-200
0
200
200
-------
en
S9.0
98.0
a:
o
97.0
96.0
95.0
£ 98.0
*& 97.0
&• 96.0
95.0
10
CURVE A
95.3---^
i
30C
CURVE B
A4 B4 C4 04 E4 F4
A3 83 C3 03 E3 F3
A2 02 C2 02 E2 F2
A1 B1 Cl D1 El F1
COLLECTOR SECTION ARAAHGiHEHT
SECT10H5 OUT
0
1 OUT
2 OUT IN PARALLEL
3 OUT IK PARALLEL
4 OUT IN PARALLEL OR
2 OUT IN A SERIES
S OUT IN PARALLEL OR
2 OUT IN SERIES AND 1
6 OUT IN PARALLEL OR
2 OUT IN A SERIES AND
GAS FLOW
OUT IN PARALLEL
2 OUT IN PARALLEL
EXAMPLE: LOAD 290 HW
SECTIONS OUT A1, A2, C2, F4
(CURVC A) EFFICENClf AT 290 W WITH 2 OUT
IK A SERIES AND 2 OUT IN PARALLEL - 95.3
COAL-ASH 14*
MOISTURE 101
(CURVE D) EFFICIENCY REQUIRED TO MEET
STATE REGULATIONS - 96.5S
TO MEET STATE REGULATIONS REDUCE
LOAD TO 210 m
12
14
16 18
ASH IN COAL, t
20
22
24
Figure 4.3-27.
Typical operating curve to meet emission regulations
with partial malfunctions of ESP.2
-------
are inoperative, he can tell from the top graph how much the boiler load
must be reduced to keep emissions in compliance with regulations. Charts of
this type should be developed for each boiler-ESP combination.
4.3.4.6.3 ESP shutdown procedures. ESP shutdown procedures will
impact upon the success of maintaining deposits on the collection surface at
a workable and acceptable thickness.31 It is necessary to keep materials
from hardening on plates yet maintain acceptable stack conditions during
this period. The exact procedures must be determined on an individual
basis.
Methods to accomplish maximum cleaning include reducing of air flow
through the unit while reducing power to the ESP. Rappers are kept at
normal settings. Gradually the fields are deenergized as the fields are
cleaned. The last fields of the ESP are kept energized while the system
cools and captures most of the reentrained dust.31
Hopper evacuation should remain on at least 1 hour after all rappers
are shut off and 2 to 3 hours after the fans have been shut down to remove
all dust. The object is to remove the dust when it's easiest, i.e., when
the temperature is still above 200°C.31
After cold shutdown of the ESP, the unit may be entered according to
safety interlock procedures incorporated in that particular unit. All
safety recommendations and procedures should be followed.
4.3-77
-------
REFERENCES
1. Smith, Wallace B. et al. Procedures Manual for Electrostatic Precipi-
tator Evaluation. Southern Research Institute. Birmingham, Alabama.
EPA-600/7-77-059. June 1977.
2. Szabo, M.F. and R.W. Gerstle, Operation and Maintenance of Particulate
Control Devices on Coal-Fired Utility Boilers. PEDCo Environmental,
Inc. EPA-600/2-77-129. July 1977.
3. A.P. deSeversky, U.S. Patent 3,315,445 April 25, 1967.
4. Jaasund, S.A. and M.R. Mazer, "The Application of Wet Electrostatic
Precipitators for the Control of Emissions from Three Metallurgical
Processes," presented at symposium entitled Particulate Collection
Problems Using Electrostatic Precipitators in the Metallurgical
Industry, June 1-3, 1977.
5. Gooch, J.P. and A.H. Dean, "Wet Electrostatic Precipitator System
Study," SRI, Birmingham, Alabama, EPA-600/2-76-142, PB 257 128, May
1976.
6. Mcllvaine Electrostatic Precipitator Manual, Chapter VI, Section L,
August 1976, p. 5613.
7. Oglesby, Sabert, Jr. , and Grady B. Nichols. Electrostatic Precipita-
tion. In: Air Pollution, 3rd Edition, Vol. IV. Engineering Control
of Air Pollution. Academic Press. New York, San Francisco, London.
1977. pp. 189-256.
8. Portius, D.H., et al. Fine Particle Charging Experiments. Southern
Research Institute. EPA-600/2-77-173. August 1977.
9. Sproull, W.T. (title) APCA Volume 22, p. 181. ' 1972 (as presented in
Reference 8).
10. Spencer, Herbert W., III. Rapping Reentrainment in a Nearby Full-Scale
Pilot Precipitator. EPA-600/2-76-140. May, 1976.
11. Gooch, J.P. and G.H. Marchant, Jr. Electrostatic Precipitator Rapping
Reentrainment and Computer Model Studies. Southern Research Institute.
EPRI FP-792, Volume 3. Final Report August, 1978.
12. White, H.J. Electrostatic Precipitation of Fly Ash, Part II, Journal
of the Air Pollution Control Association. February 1977.
4.3-78
-------
13. Bickelharyst, R.E. Journal of the Air Pollution Control Association.
1974.
14. McDonald, Jack R. A Mathematical Model of Electrostatic Precipitation
(Revision I) Volume 1, Modeling and Programming. Southern Research
Institute EPA-600/7-78-111A, June 1978.
15. Ensor, et al. Evaluation of the George Neal No. 3 Electrostatic
Precipitator. Meteorology Research, Inc. and Stearns Roger, Inc.
prepared for Electric Power Research Institute. EPRI FP-1145. August
1979.
16. Szabo, M.F., and R.W. Gerstle. Operation and Maintenance of Partic-
ulate Control Devices in Kraft Pulp Mill and Crushed Stone Industries.
PEDCo Environmental, Inc. EPA-600/2-78-210, October, 1978.
17. White, H.J, Electrostatic Precipitation of Fly Ash, Part III. Journal
of the Air Pollution Control Association. March 1977.
18. Frisch, N.W., and D.W. Coy. Specifying Electrostatic Precipitators for
High Reliability. In: Symposium on Electrostatic Precipitators for
the Control of Fine Particles, pp. 131-157. EPA-650/12-7S-016, January
1975.
19. Matts, S. and P.O. Ohnfeldt. Efficient Gas Cleaning with S.F. Electro-
static Precipitators. Flakt, A.B., Svenska Flakt labricken, June 1973.
20. Feldman, P.L. Effects of Particle Size Distribution on the Performance
of Electrostatic Precipitators. Research Cottrell, Inc. Bound Brook,
New Jersey. Presented at the 68th Annual Meeting of the Air Pollution
Control Association. June 15-20, 1975. No. 75-02-3.
21. Cooperman, P. and G.D. Cooperman. Precipitator Efficiency for Log-
normal Distributions. In: Symposium on the Transfer and Utilization
of Particulate Control Technology. EPA-600/7-79-044a. February, 1979.
22. Schneider, G.E. et al. Selecting and Specifying Electrostatic Precip-
itators. Enviro Energy Corp. Chemical Engineering, May 26, 1975, pp.
94-108.
23. Hall, H.J. Design and Application of High Voltage Power Supplies in
Electrostatic Precipitation. In: Symposium on Electrostatic Precipi-
tators for the Control of Fine Particles. pp. 159-189. EPA-650/
2-75-016, January 1975.
24. Hall, H.J. Design, Application, Operation and Maintenance Techniques
for Problems in the Electrical Energization of Electrostatic Precipi-
tators. In: Proceedings - Operation and Maintenance of Electrostatic
Precipitators, Michigan Chapter - East Central Section Air Pollution
Control Association. April 1978. pp. 19-34.
25. Diulle, Walter, Precipitator Performance Hinges on Control. Envirotech
Corporation Power, January, 1975, p. 24.
4.3-79
-------
26. Mellvaine Electrostatic Precipitator Manual, Chapter VI, Section L,
August, 1979, p. 5613.
27. Air Pollution and Industry. Van Nostrand Rheirihold Company, New York,
1972. R.D. Ross, Editor.
28. Ito, R. and K. Takimoto. Wide Spacing E.P. is Available In Cleaning
Exhaust Gases from Industrial Sources. In: Symposium on the Transfer
and Utilization of Particulate Control Technology. Volume 1: Electro-
static Precipitators. EPA-600/7-79-044a. February, 1979.
29. Allen Sherman Hoff Company. A Primer on Ash Handling Systems. 1976.
30. White, H.J. Electrostatic Precipitation of Fly Ash, Part IV. Journal
of the Air Pollution Association. April 1977.
31. Katz, J. The Art of Electrostatic Precipitation, 1978.
32. Engelbrecht, H.L. Air Flow Model Studies for Electrostatic Precipi-
tators. In: Symposium on the Transfer and Utilization of Particulate
Control Technology: Volume I - Electrostatic Precipitators. Indus-
trial Environmental Research Laboratory. Office of Energy, Minerals,
and Industry. EPA-600/7-79-044a. February 1979.
33. Lynch, J.G., and D.S. Kelly. A Review of Rapper System Problems Asso-
ciated with Industrial Electrostatic Precipitators. Air Correction
Division, UOP, Inc. In: Proceedings Operation and Maintenance of
Electrostatic Precipitators. Michigan Chapter - East Central Section,
Air Pollution Control Association. April 1978. pp. 46-61.
34. PEDCo Environmental, Inc. Data from IGCI manufacturers used in:
Design Considerations for Particulate Control Equipment. Prepared
under contract for the U.S. Environmental Protection Agency. December
1979.
35. Dismukes, E.B. Flue Gas Conditioning in Coal Fired Power Plants in the
United States. In: Second US/USSR Symposium on Particulate Control
EPA-600/7-78-037 March, 1978.
36. Lederman, P.B. et al. Chemical Conditioning of Fly Ash for Hot Side
Precipitation. In: Symposium on the Transfer and Utilization of
Particulate Control Technology, Vol. I, Electrostatic Precipitators,
pp. 79-98. EPA-600/7-79-044a. February, 1979.
37. Schliesser, S.P. Sodium Conditioning Test with EPA Mobile ESP. In:
Symposium on the Transfer and Utilization of Particulate Control Tech-
nology, Vol. I, Electrostatic Precipitators, pp. 205-240.
EPA-600/7-79-044a. February, 1979.
i
38. Peterson, H.H. Conditioning of Dust with Water Soluble Alkali Com-
pounds. In: Symposium on the Transfer and Utilization of Particulate
Control Technology, Vol. I, Electrostatic Precipitators, pp. 99-112.
EPA-600/7-79-004a. February, 1979.
4.3-80
-------
39. Bibbo, P.P., and M.M. Peaces. Defining Preventive Maintenance Tasks
for Electrostatic Precipitators, Research Dottrel!, Inc. Power.
August 1975, pp. 56-58.
40. Engelbrecht, H.L. Plant Engineer's Guide to Electrostatic Precipitator
Inspection and Maintenance, Air Pollution Division of Wheelabrator
Frye, Inc., Plant Engineering. April 1976, pp. 193-196.
4.3-81
-------
-------
4.4 FABRIC FILTERS
This section addresses the basic operating principles, design criteria,
and operation and maintenance practices for major types of commercially
available fabric filters. The particle collection mechanisms of fabric
filtration include inertia! impaction, Brownian diffusion, interception,
gravitational settling and electrostatic attraction. Particles are collected
either on a dust cake supported on a fabric or on the fabric itself.
4.4.1 Types of Fabric Filters
Although the basic particle collection mechanisms utilized in various
fabric filters are relatively similar, the equipment geometry and mode of
fabric cleaning are exceptionally diverse. This diversity is partially due
to the broad applicability of these devices, which demands various perform-
ance capabilities and physical characteristics. Diversity of fabric filters
also results from the individual contributions of numerous equipment and
fabric vendors.
Although fabric filters can be classified a number of ways, the most
common way is by method of fabric cleaning. The three major categories of
fabric cleaning methods are mechanical shaking, reverse air cleaning, and
pulse jet cleaning.
4.4.1.1 Mechanical Shaking. A conventional shaker-type fabric filter
is shown in Figure 4.4-1. Particulate-laden gas enters below the tube sheet
and passes from the inside bag surface to the outside surface. Particles
are captured on a cake of dust that gradually builds up as filtration con-
tinues. At regular intervals a portion of this dust cake must be removed to
enhance gas flow through the filter. The dust cake is removed manually on
small systems and mechanically on larger systems. Mechanical shaking of the
filter fabric is normally accomplished by a rapid horizontal motion induced
by a mechanical shaker bar attached at the top of the bag. The shaking
creates a standing wave in the bag and causes flexing of the fabric. The
flexing causes the dust cake to crack, and portions are released from the
fabric surface. Some of the dust remains on the bag surface and in the
interstices of the fabric. The cleaning intensity is controlled by bag
tension and by the amplitude, frequency, and duration of shaking. The
4.4-1
-------
OVERMWJNTED
EXHAUSTER
INLET CHAMiER &
HOPPER BARHE
Figure 4.4-1. Small shaker-type baghouse (courtesy of
Carborundum Division of Flakt, Inc.)
4.4-2
-------
residual dust cake provides a minimal resistance to gas flow, causing a
static pressure drop that is higher than that of a new clean fabric. Woven
fabrics are used in shaker-type collectors. Because of the low cleaning
intensity, the gas flow is stopped before cleaning to eliminate particle
reentrainment and allow dust cake release. The cleaning may be done by bag,
row, section, or compartment.
Gas flow through shaker-type fabric filters is usually limited to a low
superficial velocity of less than 1 m/min. This means that the total gas
flow rate (at operating temperature and pressure) divided by the total cloth
area available should not exceed the stated "velocity." This parameter is
usually referred to as the air-to-cloth (A/C) ratio and is expressed in
m3/m2-min. High A/C values may lead to excessive particle penetration or
blinding, which reduces fabric life.
Mechanical shaker-type units differ with regard to the shaker assembly
design, bag length and arrangement, and type of fabric. All sizes of control
systems can use the shaker design.
4.4.1.2 Reverse Air Cleaning. Particles can be collected on a dust
cake on either the inside or outside of the bag. A small cylindrical unit
with external surface filtering is shown in Figure 4.4-2. In this design
the bags are arranged radially and are suspended from an upper cell plate.
The inner and outer row reverse-air manifolds rotate around the unit and
stop at each bag to induce reverse flow. In this manner the entire baghouse
need not be temporarily isolated to allow dust-cake removal. A reverse air
panel filter is shown in Figure 4.4-3. The reverse air cleaning manifold
traverses the rows of filter panels, cleaning all six layers simultaneously.
A somewhat larger reverse air filter is shown in Figure 4.4-4.
Regardless of design differences, reverse-air cleaning is accomplished
by reversal of the gas flow through the filter media. The change in direc-
tion causes the surface contour of the filter surface to change (relax) and
promotes dust-cake cracking. The flow of gas through the fabric assists in
removal of the cake. The reverse flow may be supplied by cleaned exhaust
gases or by a secondary fan supplying ambient air.
4.4-3
-------
INNER CLEAN
REVERSE AIR ROW REVERSE AIR
PRESSURE BLOWER DH1VE MOTOR AIR MANIFOLD OUTLET
OUTER ROW
REVERSE AIR
MANIFOLD
FABRIC
FILTER TUBES
HEAVY DUST
DROPOUT
MIDDLE ROW
REVERSE AIR
MANIFOLD
PRE-CLEANINQ
BAFFLE
Figure 4.4-2.
Reverse-air baghouse (courtesy of
Carter-Day Company).
4,4-4
-------
Inlet
Screen frame.
and
•filter bogs
• Walkway
Cleor sir side
Walkwa
Outl«t
•Casing
.Reverse olr
cleaning manifold
Grid wall
Manifold drive
and sheaves
Rotary discharge
valves
Figure 4.4-3. Continuous reverse air-cleaning system for flat
filter sleeves (courtesy of Flakt, Inc.).
4.4-5
-------
Figure 4.4-4. Reverse-air collector.
(Courtesy of MikroPul Corporation)
4.4-6
-------
In filters with internal cake collection, cleaning is accomplished
during off-line operation with compartments isolated. The filter bag may
require anti-collapse rings to prevent closure of the tube and dust bridg-
ing. Cake release may be increased by rapid reinflation of the bag, creat-
ing a snap in the surface, followed by a short period of reverse air flow.
Fabrics in reverse-air collectors may be woven or felt. The felts are
normally restricted to external surface collection.
Reverse-air filters are usually limited to A/C ratios of less than 1.0
m3/m2-min, but the ratio may be higher in certain applications. Most high-
temperature (>250°C) baghouses are of this type.
4.4.1.3 Pulse-Jet Clearn'ng. On pulse-jet fabric filters, particle
capture is achieved partially on a dust cake and partially within the fabric.
Filtering is done on exterior bag surfaces only. A small pulse-jet baghouse
is illustrated in Figure 4.4-5; this device is typical of many small instal-
lations. The bags, supported by inner retainers (sometimes called cages),
are suspended from an upper cell plate. Compressed air is supplied through
a manifold-solenoid assembly (not shown) into the blow pipes shown in an end
view. Venturis mounted in the bag entry area are intended to improve the
jet pump effect. The classifier shown at the gas inlet is intended to
prevent large particles from abrading lower portions of the bag.
A sudden blast of compressed air is injected into the top of the bag.
The blast of air creates a traveling wave in the fabric, which shatters the
cake and throws it from the surface of the fabric. The cleaning mechanism
is classified as fabric flexing and with some degree of reverse air flow.
Felted fabrics are normally used in pulse-jet-cleaned collectors, and the
cleaning intensity (energy) is high. The cleaning normally proceeds by
rows, all bags in the row being cleaned simultaneously. The compressed gas
pulse, delivered at 550 to 800 kPa results in local reversal of the gas
flow. The cleaning intensity is a function of compressed gas pressure.
Pulse-jet units can operate at substantially .higher A/C ratios than the
types previously discussed. Typical ratio ranges are 1.5 to 3.0 m3/m2-min.
The plenum pulse cleaning method is a variation of the pulse-jet clean-
ing mechanism; in this method an entire section of bags is pulsed with a
blast of compressed gas from the clean air plenum. The intensity of plenum
4.4-7
-------
TUBE SHEET
CLEAN AIR PLENUM
PLENUM ACCESS—
BLOW PIPE
INDUCED FLOW
BAG CUP & VENTURI
BAG RETAINER
Figure 4,4-5.
Pulse jet baghouse (courtesy of George A. Rolfes
Company).
4.4-8
-------
4*4*2
Principles of Fabric Filters
The two factors of basic importance in fabric filter operation are
particle capture and static pressure loss. Particle capture mechanisms on a
microscopic level are not fully understood. Macroscopic behavior, the net
result of all microscopic processes, indicates that fabric filter collection
is not highly size-dependent as would be expected in view of the collection
mechanisms. The static pressure loss results from forcing the gas stream
through the fabric and dust cake.
4.4.2.1 Particle Capture. Pore sizes (open areas) of the woven fabric
through which the contaminated gas stream passes (open areas) range from 10
to 100 urn, depending on fabric construction and fiber characteristics.1
Initially the particles easily penetrate this filter. As cleaning con-
tinues, some particles are retained upon filter elements (normally fibers)
because of the combined action of the classified collection mechanisms shown
in Figure 4.4-6.2 This figure is a simple modification of the mechanism
diagrams of Figures 4.1-5, 4.1-6, and 4.1-7. As the dust cake builds up,
additional "targets" are available to collect particles; accordingly,
penetration drops to very low levels.
DIRECT
INTERCEPTION-
N ^»
S
DIFFUSION^.
\ f- ELECTROSTATIC
wx\ ATTRACTION
— GRAVITATIONAL
SETTLING
INERTIAL
IMPACTION
Figure 4.4-6. Initial mechanisms of fabric filtration (courtesy of
CRC Press, Inc. Theodore, Louis and Anthony J. Buonicore.
Industrial Air Pollution Control Equipment for Particulates. 1976).
4.4-9
-------
Within the dust cake, inertia! impaction is the dominant collection
mechanism. The forward motion of the particles results in impaction on
fibers or on already deposited particles.2'3 Although increasing gas veloc-
ities favor impaction, they reduce the effectiveness of Brownian diffusion.
Increasing the fabric po.rosity also reduces diffusional deposition.4 Grav-
ity settling of particles as a method of collection is usually assumed to be
negligible.2 This effect should be considered at low velocities, how-
ever.4'5 Electrostatic forces may affect collection; however, the impact on
commercial-scale equipment is not fully understood.6 The magnitude of the
electrical charges may depend relative humidity.6 The source of the charges
could be triboelectric interaction between the particles and the fabric or
triboelectric interaction before the particles reach the fabric6 - 9 Donovan,
et al.6 have concluded that the latter is more important.
The net result of the particle collection mechanisms is potentially
high-efficiency removal and a penetration curve of the types presented by
Turner10 in Figures 4.4-7 and 4.4-8. These data indicate only a weak par-
ticle size dependence. The form of the curve is confirmed by data in Figure
4.4-9 from a shaker type fabric filter (silicon-graphite coated fiberglass
bags) serving a spreader stoker boiler.11 The penetration value scale does
not exceed 0.01 so that the shape of the curve can be illustrated.
Dennis and Klemm have proposed that leaks through unblocked pores of
100 to 200 urn in diameter in woven fabrics are partially responsible for the
lack of a strong particle size dependence evident in filters using woven
fabrics.12'13
A direct relationship between penetration and woven fabric pore concen-
tration was observed by Hall, Dennis, and Surprenant.14 Gas velocities
through the pores may be several orders of magnitude above the average face
velocity.14
The particulate matter emission rate through the pores should be a
function of the inlet gas mass loading and the air-to-cloth ratio. The
latter determines the actual velocity through the pores. Figure 4.4-10
illustrates the relationship between face velocity and outlet concentration
from a series of bench-scale tests. Hall, et al.14 conclude that the higher
emissions at the higher air-to-cloth ratios are attributable to particle
reentrainment through pinholes.
4.4-10
-------
100
&98
w
97
••3
196 -
94
EPA Control Systems Laboratory
Single-point Impactor Data
(shake cycle not included)
— Overall mass efficiency-;99.7%
I
1.0 2.0
Particle size, /»m
3.0
4.0
Figure 4.4-7. Baghouse performance, lead sinter machine.
100
EPA Control Systems Laboratory
Single-point Impactor Data
(not verified)
' Normal A/C ratio»3:l
1 High A/C ratio «6:1
Overall mass efficiency at normal A/C ratio=99.76%
Overall mass efficiency at high A/C ratio =99.51%
I I I
1.0
2.0
Particle size, /»m
3.0
4.0
Figure 4.4-8. Baghouse performance, industrial boiler.
4.4-11
-------
_
'e 5
o>
v>
3
o
en 4
en
3
O
i-
1
0
T—~r
Symbol
D.
Load Air-to-Cloth Ratio
Mw ft3/min/ft2
6 1.87
11 2.47
2.74
0.1 1.0-
Particle Diameter, Specific Gravity = 2.0 (g/cm )
Figure 4.4-9. Fabric filter penetration (adapted from
•, . • Reference 11). .
10
4.4-12
-------
20 40 60 80 100 I2O 140
FABRIC LOADING !W),
Figure 4.4-10. Effect of air-to-cloth ratio on outlet concentration.
Dennis and Klemm have presented a computerized model useful for pre-
dicting performance of shaker-type and reverse-air-type fabric filters.
This model is discussed in references 12, 13, and 15.
Penetration through felted bags used in pulse jet fabric filters is
believed due to (1) direct passage through the fabric and residual cake
(especially following pulse cleaning and (2) seepage of particles through
the fabric.16'" particulate emissions can increase substantially at high
air-to-cloth ratios.16'" Developmental work on models applicable to pulse
jet filters are discussed by Dennis, Wilder, and Harmon,18 and Leith and
Ellenbecker.17
Emissions from baghouses are not necessarily limited to particle pene-
tration through fabrics. Localized bypassing of filter elements can occur
through gaps in sheet welds, around poorly seated seals and gaskets, and
through bag tears.™'1* The outlet particle size distribution would not
4.4-13
-------
appear substantially different than that of the inlet although some very
large diameter particles may not be able to negotiate a pathway through the
narrow gaps.
Theodore and Reynolds have developed the following orifice equation for
calculating the increase in penetration due to pinholes, tears, and missing
bags.19
(Eq. 4.4-1)
where
p = increase in penetration.
Q (Eq. 4.4-2)
U'J3
Ld (t + 460)
where
4> = dimensional parameter.
Q = system gas volume, acfm.
L = number of broken bags.
d = diameter of orifice, in.
t = temperature, °F.
AP = fabric filter pressure drop, in. H20.
For convenience the increase in penetration is plotted against pressure
drop and <}> (Figure 4.4-11). The significance of bag failure can be seen
from the following example. If two bags, 6 in. in diameter are removed from
a 1415-ms/min baghouse operating at a static pressure drop of 1.15 kPa, the
collection efficiency is reduced by 4.7 percent from 99.5 percent to 94.8
percent. The outlet concentration increases from 4.56 x 10 kg/m3 to
-4
4.72 x 10 kg/m3. Note that the penetration does not depend directly on
the total number of bags in the collector or the percentage of the total
number failed. The penetration depends only on the absolute number of bags
failed. The pressure drop may be reduced as the bag failures increase, and
care must be taken to adjust the value if a significant number fail.
4.4.2.2 Pressure Drop. The static pressure drop across the fabric
filter is the sum of the static pressure drop across the cleaned fabric and
that across the accumulated cake. The latter is a function of time since
4.4-14
-------
1.0
O-
O£.
LU
O
IXJ
at:
oc
o
0.1
0.01
UJ
Q-
0.001
Numbers on curves are
value of *
Ldz (t + 460)
where Q 1s 1n acfm,
d in Inches, and
t 1n °F.
0.5
Figure 4.4-11.
1248
PRESSURE DROP (in. H20)
16
Penetration correction term as a function
of pressure drop and $.
4.4-15
-------
the last cleaning. The curve shown in Figure 4.4-12 represents the uniform
deposition of dust of a completely cleaned fabric. The slope of the line is
calledj<2» the specific resistance coefficient.
The shape of the curve in previously operated systems is not usually
like that shown in Figure 4.4-12 since it is difficult to completely clean
the fabric. Dennis and Klemm12'13 have proposed that partial cleaning
results in areas where the dust cake has partially "flaked off" and areas
where a substantial dust cake remains. This is illustrated in Figure 4.4-13.
Using a parallel .flow approach, they have modeled gas flow behavior through
these different areas. The model has adequately predicted static pressure
profiles in laboratory and commercial fabric filter systems.11'15
Generally, static pressure drop is proportional to the inlet dust
loading, air-to-cloth ratio, fabric structure and cleaning system cycle and
intensity. Most units are designed to operate at differential static pressures
of 0.5 to 2 kPa, however, some units operate at differential pressures up to
3 kPa.
The static pressure profiles for shaker and reverse air systems tend to
have a distinct sawtooth pattern if there are only a few compartments. On
larger systems, this pattern disappears and the static pressure drop across
the system is relatively constant.20
4.4.3 Design of Fabric Filters
A complete characterization of the effluent gas stream is important in
the design of a fabric filter system. This would include: gas flow rate,
minimum and maximum gas temperatures, acid dew point, moisture content,
presence of large particulate matter, presence of sticky particulate matter,
particulate mass loading, and presence of potentially explosive gases or
particulate. Given accurate data on these effluent characteristics,an
appropriate collector can be designed for the required degree of control.
Selection of the type of fabric, and dimensions of the bag normally must be
done in conjunction with the design of the cleaning system. The size of the
fabric filter depends on the air-to-cl.oth ratio necessary and the number of
compartments expected to be out-of-seryice for maintenance or cleaning at
any given time. The overall costs must be balanced against needs for good
accessibility and instrumentation, both of which favor improved maintenance
and» therefore, improved performance.
4.4-16
-------
DESCRIPTION
CO
<
K
G
OL
LJ
b
u.
u
Ifl
<
cc
Ul
MAXIMUM POSSIBLE CLEANING
HIGHLY EFFICIENT CLEANING
AVERAGE CLEANING RANGE-
MECHANICAL SHAKING
AVERAGE CLEANING RANGE-
COLLAPSE WITH REVERSE
FLOW
AVERAGE FABRIC LOADING
Figure 4.4-12. Filter drag profiles.
-------
Figure 4.4-13. Fly ash dislodgement from 10 ft x 4 in. woven glass
bag (inside illumination).
4.4-18
-------
4.4.3.1 Effluent Characteristics. The effluent characteristics should
be quantified to the extent possible before fabric filter design is com-
pleted. Variability of these conditions should be considered.
Gas Temperature. The gas stream temperature and its variability over
time determine the fiber and finish selected for the filter bag. The tempera-
ture of gases emitted from industrial processes may vary over several hundred
degrees in short periods of time. Low gas stream temperatures may go below
the gas moisture and acid dew points, and high temperatures may exceed the
maximum that the fabric will tolerate. The extremes of temperature must be
determined before fabric selection.
The temperature of the gas may be modified by using heaters (indirect
or direct fired) to increase the gas temperature above the dew point or by
using coolers to reduce the gas temperature to one compatible with the
fabric. Methods commonly used to increase or maintain gas temperature are
insulation of ductwork, use of direct-fired afterburners and heat exchangers,
and steam tracing. The use of direct-fired afterburners may serve two
purposes: to elevate the gas temperature above the dew point and to remove
organics that may blind the fabric surface.
Methods of cooling the gas include dilution, radiative cooling, and
evaporative cooling.2'21 .The use of dilution air to moderate gas tempera-
ture is the simplest approach, but it may increase the capital cost because
the volume of ambient air required may necessitate a substantially larger
filter. Figure 4.4-14 shows the increased capacity required when gas tem-
perature is reduced to 560°K by use of 300°K dilution air.
Radiative cooling does not require increased collector size, but it
does require an investment in greater duct length. Also, the static pres-
sure drop of the system may be increased by the increased duct length.
Figure 4,4-15 shows the reduction in gas temperature achieved as a function
of duct length with gases at an initial temperature of 1150°K. Radiation
cooling of gases at temperatures above 115Q°K requires exotic construction
materials and may be uneconomical.21 The cooling of gases at temperatures
below 800°K requires extensive surface area and is usually uneconomical
relative to the cost of the greater baghouse capacity required by. other
cooling methods.21
4.4-19
-------
1300
1100
s
UJ
I
i
900
700
500
300°K AIR USED
FOR DILUTION
0 100 200 300 400
INCREASE IN BAGHOUSE CAPACITY, X
Figure 4.4-14. Added capacity needed in baghouse when hot gases
are cooled by dilution with ambient air (reprinted by permission.
Vanderhoeck, P., Chemical Engineering, May 1, 1972).
1200
1000
800
600
500
400
T
ASSUMING * I.I « DIMETER DUCT KAHOUNG 1,400 «»
Of 1150'K BAS AT l.WO •/•in. CONTAINING H
MATER. 161 CO;. 310*K AMB1EKT TEMPERATURE.
30 60 90
LENOTH OF DUCT (FROM HOT GAS SOURCE), m
670°K
640° u
g
610
s-
580
550
520
20 g
40 g
_j
60 §
80 g
100 3
120
Figure 4.4-15. Radiation effectiveness in cooling hot gases (re-
printed by permission; Vanderhoeck, P. Chemical
Engineering, May 1, 1972).
Evaporative cooling is accomplished by injection of water into the gas
stream. The energy drawn from the gas stream to vaporize the water leads to
a reduction in gas temperature. The cooling is accomplished rapidly and in
a small space. The system must be designed to reduce gas temperature to a
point above the gas dew point, to prevent carryover of unvaporized water
droplets into the filter, and to prevent spray water impact on duct walls or
4.4-20
-------
liners.21 Figure 4.4-16 shows the increase in the necessary baghouse capacity
as a result of evaporative cooling (desired gas temperature of 560°K).
Presently, bags made of graphitized, siliconized fiber glass can withstand
this temperature. The water temperature has a relatively minor impact on
the quantity of water required.
1500
1300
1100
900
700
500
300°K WATER USED
10 ZO 30 40 50
INCREASE IN BAGHOUSE CAPACITY, %
Figure 4.4-16. Added capacity needed in baghouse when hot gases
are cooled by evaporating cooling (reprinted by permission:
Vanderhoeck, P. Chemical Engineering, May 1, 1972).
Particle characteristics. The particle size distribution of the dust
must be considered in design of the collector and in fabric selection.
Particle size distribution affects both dust cake porosity and abrasion of
the fabric. The presence of fine particles in the gas stream can create a
heavy dust cake and increase the static pressure drop through the cake.22
The small particles can also cause fabric bleeding.
The presence of large abrasive particles can reduce bag life and may
require the use of a precleaner or gas distribution devices in the collec-
tion system (see the inlet diffuser of Figure 4.4-5). Moreover, because the
resistance of the fabric to abrasion is greatly reduced when particles
strike tangentially, the presence of large particles may require modifica-
tion of gas inlet design. For certain sources such as spreader stoker
boilers, a mechanical collector ahead of the fabric filter may provide
protection from the large quantity of >10 umA particles. This would also
allow protection from glowing embers and thus reduce the risk of hopper
fires.23
4.4-21
-------
Particulate matter with a high carbon content may be generated during
periods of improper combustion. Fabric filter design should include ways to
minimize hopper fires.24'25 These could include continuous removal of
collected material, limited air entry into hoppers, and a fire-sensing
system.20'24
Sticky particulate can be difficult to remove from the fabric surface.
A survey done by Billings and Wilder indicated blinding of bags due at least
partially to sticky particulate was the most frequent problem reported.26
Improper combustion can lead to problems of this type and result in sub-
stantially increased static pressure drop.27
Gas Composition - Factors of importance regarding the gas composition
include moisture content and acid dew point. If a fabric filter is operated
at close to the acid dew point, there is a substantial risk of corrosion
especially in localized spots close to hatches, in dead air pockets, in
hoppers, or adjacent to heat sinks such as external supports.28'29 If the
operating temperature drops below the water dew point, either during start-up
or at normal operation, blinding of the bags can occur. The presence of
trace components, such as fluorine, can attack certain fabrics.
4.4.3.2 Fabric Selection. Fabric selection is usually based on the
prior experience in similar applications. Important factors to consider
are:
Dust penetration
Continuous and maximum operating temperatures
Chemical degradation
• Abrasion resistance
Cake release
Pressure drop
Cost
Cleaning method
Fabric construction.
The choice of fabric ultimately affects pressure drop, selection of cleaning
method, outlet concentration, and the life of the fabric under operating
conditions.
Temperature. Degradation of the polymer in synthetic and natural
fabrics is accelerated temperature. The rate of decay is also enhanced by
the actions of moisture, chemicals, and abrasive particles. Temperatures at
4.4-22
-------
which reasonable performance can be expected under normal conditions are
given in Table 4.4-1, which gives the continuous maximum operating tempera-
tures recommended by fabric manufacturers and summarized by Theodore and
Buonicore.2
The maximum short-term temperature represents the temperature at which
rapid deterioration will result in immediate failure. For synthetics, this
is the temperature at which polymer softening occurs and causes permanent
elongation. Figure 4.4-17 shows the reduction in strength of Nomex^fabric
with increasing temperature,30
Glass fabrics are sensitive to abrasion between fibers and are normally
I I? I
coated with either Teflorc-'or silicon-graphite. Polymer finishes can degrade
with increasing temperature. Figure 4.4-18 shows the effect of gas temperature
on the finish of glass fiber bags.31
Chemical degradation. Chemical degrading of the fabric is caused by
the breaking of polymer chains within the fiber structure. The degrading
may be from acid hydrolysis, alkali attack, or, in the case of glass fiber
fabrics, conversion of the structure to a noncrystalline form that has lower
strength.
As the chain length of a polymer is reduced by chemical attack, it
loses strength. The chemical attack may be accelerated by moisture or metal
catalysts in the dust impregnated in the fibers. The rate of attack in-
creases with temperature.
Chemical composition of the gas stream, moisture content, and temper-
ature must be considered in selection of the fabric. Table 4.4-2 indicates
the ratings of commercial fabrics with respect to chemical resistance. Note
that resistance is a relative term that does not imply total resistance to a
specific chemical. Also resistance may be greatly reduced by cyclic opera-
tion under different conditions and concentrations.
Abrasion resistance. Resistance to abrasion is a relative term that
indicates the ability of a fabric to provide extended service in collecting
abrasive dust. Resistance can be modified by fabric construction, fabric
finish, and shapes of the particles collected.
4.4-23
-------
Table 4.4-1. RECOMMENDED TEMPERATURE LIMITS FOR VARIOUS COMMERCIAL FABRICS2
Fabric
Cotton
Wool
Nyloifc/
Dynel®
Polypropylene
Orion©
Dacron©
Nomex©
Teflon©
Fiberglass
Stainless stee
(type 304)
Generic name
Natural fiber cellulose
Natural fiber protein
Nylon polyanride
Modacrylic
Polyolefin
Acrylic
Polyester
Nylon aromatic
Fluorocarbon
Glass
Type yarn
Staple
Staple
Filament spun
Filament spun
Filament spun
Spun
Filament spun
Filament spun
Filament spun
Filament spun
bulked
Maximum temperature range, °K
Long periods
of time
350
370
370
350
370
390
410
490
500
.530
1000
Short periods
of time
380
400
400
390
400
410
440
530
530
590
No data
Melting
temperature, °K
420 decomposes
575 chars
520
440 softens
440
520 softens
420
640 decomposes
670 decomposes
1070
1700
-------
60
**•"%
•M
"O
» 50
O
*f~
X)
(O
u-
M_ 40
O
*
c
I 30
M
rc
1 20
co
h-
1 10
LU
\
^
CONDITIONS
500 h exposure
500 ppm S02
4* 02
6% H20
3/1 air-to-cloth ratio
(m3/m2 - min)
2 cpm pulse
power house dust
NOMEX
R 14-oz FELT
POLYESTER-18-pz FELT
100
200
TEMPERATURE, °F
300
400
Figure 4.4-17. Effect of acid and temperature on strength of Nomex
and polyester fabrics (reprinted by permission:
E. I. du Pont de Nemours and Company).
4.4-25
-------
CJ
I
o
800
700
600
500
400
300
200
100
-COMMERCIAL SILICONE
FINISH
-COMMERCIAL SILICONE
PLUS GRAPHITE
FINISH
\
100 200 300
TEMPERATURE, C
400
35
30
25
20
15 o
10
Figure 4.4-18. Effect of gas temperature (continuous) on life of glass
fabric bags (reprinted by permission: Menardi and Company).
4.4-26
-------
Table 4.4-3. CHEMICAL RESISTANCE OF COMMON COMMERCIAL FABRICS2
Fabric
Cotton
Wool
Nylon
Dynel®
Polypropylene
Orion®
Dacron®
Nomex
Teflon
Fiberglass
Polyethylene
Stainless steel
(type 304)
Generic name
Natural fiber
cellulose
Natural fiber
protein
Nylon polyamide
Modacrylic
Polyolefin
Acrylic
Polyester
Nylon aromatic
Fluorocarbon
Glass
Polyolefin
Type yarn
Staple
Staple
Filament spun
Filament spun
Filament spun
Spun
Filament spun
Filament spun
Filament spun
Filament spun
bul ked
Filament spun
Acid
resistance
Poor
Very good
Fair
Good-very
good
Excellent
Good-
excellent
Good
Fair
Excellent
Fair- good
Very good-
excel! ent
Excellent
Fluoride
resistance
Poor
Poor- fair
Poor
Poor
Poor
Poor- fair
Poor-fair
Good
Poor- fair
Poor
Poor-fair
Alkali
resistance
Fair- good
Poor- fair
Very good-
excellent
Good-very
good
Excellent
Fair
Fair-good
Excellent
Excellent
Fair
Very good-
excellent
Excellent
Flex and
abrasion .
resistance
Fair- good
Fair
Very good-
excellent
Fair- good
Very good-
excellent
Fair
Very good
Very good-
excel 1 ent
Fair
Poor
Good
I
fVJ
-------
Pressure drop. Static pressure drop must be considered In selection of
the fabric. The residual pressure drop affects the cost of operation due to
an increase in fan horsepower. In applications where the fan capacity is
limited, the increase in pressure drop can reduce gas volume and reduce
capture and transport velocity in the ventilation system.
Cleaning method. The method of removing the dust cake is closely
related to fabric construction and fiber type. With woven fabrics that are
subject to abrasion or flex damage, gentle cleaning methods such as low-frequency
shaking or reverse air can be used. With felted fabrics, a more intense
cleaning method is required, such as high-pressure reverse air or pulse-jet
cleaning. An improper combination of fabric and cleaning method (e.g.,
intense shaking of glass bags) can cause premature failure of the fabric,
incomplete cleaning, or blinding of the fabric (complete plugging of pores)..
«
Fabric construction. Filter fabrics commonly used in operating facil-
ities are either woven or felted. Unlike woven fabric, felt is a genuine
filter medium and is more efficient in collection of particulate at compar-
able filtering velocities; it is, however, more expensive. Felted fabrics
are composed of randomly oriented fibers and are.relatively thick. Needling
the fibers meshes them and forms a strongly bonded fabric. The thickness of
felt provides for maximum particle impingement, but increases the static
pressure drop (reduces permeability). Felted fabrics are normally used in
pulse-type units and are operated at high A/C ratios.
Woven fabrics are characteristically used in shaker and reverse-air
filters, and are operated at relatively low A/C ratios. Woven fabric is
made up of filament or staple (spun) yarns in a variety of patterns, having
various spacings between the yarns, with a specific finish that is designed
to retain or shed filter cake, depending on the application. Seven of the
most common weave patterns are shown in Figure 4.4-19. Plain weave is
lowest in initial cost, and has the least porosity and greatest particle
retention; however, its potential for blinding is greatest. Twill weave has
medium retention and blinding characteristics and has reasonable permeability.
Twill weave also exhibits the best resistance to abrasion. Sateen weave has
the lowest particle retention, is easiest to clean of accumulated dust, and
has lowest potential for blinding.
4.4-28
-------
ft
tt
ft
It
(
ft
'j
/
2
i
^
2
2
5
2
y
Plain or Taffeta,
Weave 1/1
3/2 Regular Twill
"Crowfoot" Satin or
4 Shaft Satin
No. 4 Harness
2/1 Regular
Twill
4/1 Satin
(Sateen if Cotton)
No. 5 Harness
3/1 Regular
Twill
a.
QC
O
UJ
2/2 "Broken" Twill
or Chain Weave
Other Popular Weaves:
Drill = 2/1 L.H. Twill, or 3/1 Twill
Herringbone = a type of broken twill
Basket Weave = extension of plain weave
Gabardine = regular or steep twill with
higher warp than fill count
Figure 4.4-19. Typical fabric weaves (reprinted by permission:
Industrial Gas Cleaning Institute).
4.4-29
-------
The permeability of woven fabrics depends on the type of fiber, tightness of
twist, size of yarn, type of weave (geometric pattern), tightness of weave
(thread count), and type of fabric finish.
Woven fabrics may be provided with a number of finishes. Cotton fabrics
may be preshrunk to maintain dimensional stability, i.e., resistance to
stretching or shrinking in any direction, which could adversely affect other
fabric characteristics. Spun fiber fabrics may be napped on the surface
receiving the dust load. Napping is the process of pulling fibers out of
the yarn bundles to form a soft pile; this promotes the formation on the
fabric surface of a dust cake that does not penetrate the interstices of the
fabric. Synthetic fabrics may be heat-set to ensure dimensional stability
and provide a smooth surface with uniform permeability. Any fabric may be
silicone treated (also used in combination with graphite and Teflon-3 to
improve abrasion resistance, to facilitate cake release, and to reduce
moisture absorption.
4.4.3.3 Selection of Cleaning Technique. A number of factors are
considered in the selection of a cleaning technique for a fabric filter
system. Primarily, the physical and chemical properties of the flue gas and
particulate must be clearly defined. Specific case studies involving simi-
lar processes, together with laboratory studies, are often the most informa-
tive guideposts for design parameters. Critical interdependences in clean-
ing technique selection are particulate characteristics/cleaning method,
specific resistance/cleaning method, and cleaning method/service life.
Constraints imposed by intermittent or continuous operation and by availa-
bility of space must also be considered.
If the dust cake is released readily from the fabric, reverse air
cleaning may be adequate. Reverse air can be used in combination with
mechanical shaking. Felted fabrics generally are not cleaned by reverse air
because of their greater structural depth and, hence, greater dust retentiv-
ity. Bag tensioning and reduction of reverse air flow rates minimize the
degree of bag flexure and thus reduce the risk of accelerated bag wear.
These measures also prevent complete collapse of the bag, which makes cake
removal difficult. The rate of flexure is probably the controlling factor
with respect to fabric failure.
4.4-30
-------
Reverse-air and mechanical-shake units are capable of being cleaned
only while the unit or a single compartment is off line. In most combina-
tion reverse-air and mechanical-shake systems, bag collapse and/or flexure
caused by flow reversal are the major dust dislodging forces.
In shaker systems, the fabric cleaning action is defined by quantifying
shaking frequency, shaking amplitude, and duration of the shaking interval.
Tensioning of the bags is important in determining average amplitude and
acceleration of the bag. If the bag is too slack, the transmission of
cleaning energy over the entire length of the bag is incomplete, with the
result that cleaning is inefficient and nonuniform. This can lead to
abrasion damage and reduced bag life.
4.4.3.4 Sizing. The size of a fabric filter system is determined by
the gas volume to be filtered and the A/C ratio at which the filter can be
operated in view of fabric type, dust cake properties, and cleaning method.
The area of fabric surface is determined by multiplying the total gas flow
by the recommended A/C ratio.
Penetration is directly related to the effective air-to-cloth ratio in
the system, with substantially increased emission levels at high air-to-
cloth ratios.13'14'34 Accordingly, the lowest possible face velocity
consistent with economic constraints should be specified.
The minimum number of compartments in shaker-type and reverse-air units
is related to the maximum allowable increase in A/C ratio as one or more
compartments are removed from service. The following additional factors
should be considered for compartmentalization in fabric filter design:
1. Large compartments may necessitate oversized and uneconomical
cleaning equipment.
2. Large compartments contain more bags and, therefore, present a
higher probability that bag failure will occur in a single com-
partment. This eventuality could necessitate intermittent
replacement of bags.
3. Large compartments require large ducts and dampers; small com-
partments require more numerous, but smaller, ductwork branches
and dampers.
4. Large compartments require larger solids-collection hoppers.
4.4-31
-------
5. Larger compartments cool more slowly when brought off line for
maintenance.
Most of these considerations favor smaller, more numerous compartments.
The compartments of shaker and reverse air should be arranged so that
maintenance of all filter tubes is relatively simple. The number of rows of
bags along each walkway should be minimized so that the bags nearest the
compartment walls can be reached without disturbing too many bags near the
walkway. The appropriate "bag reach" is determined by the layout of tubes
and the diameter. Normally there are no more than 3 to 4 bags deep.35'36
4.4.3.5 Solids Removal Equipment. As solids are cleaned from the
filtering fabric, they fall to a collection hopper for ultimate removal.
The "fluid" properties of the collected solids are important in design and
operation of these systems, and they may be markedly different from the
properties of the material from which they originated. Fine dusts, for
example, tend to pack more readily than coarser materials; moreover, conden-
sation formed in the filter device may cause solid material to agglomerate.
Both of these factors can make solids disposal difficult.
Various design features can help prevent the clogging of solids collec-
tion hoppers. The. hopper should be designed with a steep valley angle;
angles of 55 to 70 degrees are recommended. Hoppers should also include
R
large discharge openings, smooth coatings (i.e., epoxy or Teflon ) on inside
surfaces, and minimal ledges or other obstructions on sidewalls. The top of
the hopper sidewall should drop vertically and begin the slope to the dis-
charge point at least one bag diameter below the bottom of the bags to allow
proper dust discharge. At least 0.3 m of clearance should be provided
between hopper walls and any internal partitions to allow easy discharge.
Heaters and insulation can be installed in hoppers to prevent condensa-
tion and caking of collected material. Jets of hot air can be used, to
fluidize material in the hopper and keep it free flowing.
Solids are generally removed from the hopper by means of a discharge
valve, which removes ash from the hopper while preserving the pressure
differential between the dust conveyance system and the fabric filter sys-
tem. ;
4.4-32
-------
The solids are transported to a collection point by means of screw con-
veyors, pneumatic (either vacuum or positive systems) systems, and wet
sluicing systems. Screw conveyors, which are common on small systems, work
well in a variety of applications but are sometimes cumbersome.
Pneumatic systems are not limited to straight-line runs as are screw
conveyors, and, therefore, are more flexible. Particularly abrasive solids
must be accounted for in the design of pneumatic systems by appropriate
materials of construction and in some applications by installation of
replaceable wear plates at turns. Sluice systems are normally found in
coal-fired boiler applications for transport of the boiler bottom ash.
4.4.3.6 Instrumentation. Reliable operation of a fabric filter is
favored by the use of the following instrumentation:
1. Thermocouples or other temperature-measuring instruments located
at the device inlet.
2. Inlet/outlet differential static pressure gages.
3. A single-pass transmissometer (opacity meter).
4. Compressed air pressure gage.
5. Fan motor ammeter.
In lieu of differential pressure gages, it is sometimes simpler to
install static pressure taps where appropriate and use a portable meter to
obtain readings. This approach reduces problems of meter moisture damage,
meter corrosion, and plugging of lines. Where permanent differential static
pressure gages are used, the static pressure lines should be as short as
possible and free of 90-degree elbows. Copper tubing has been found to be
less susceptible to deterioration than the polypropylene lines commonly
used.
Recording temperature meters are especially useful in identifying high
or low-temperature excursions, which rapidly destroy fabrics. As a less
expensive alternative, high-temperature indicators composed of colored fiber
or temperature-sensitive plugs may be used.
The single pass transmissometer may not provide an accurate measurement
of effluent opacity; however, it. is useful in identifying problems. A
significant leak is detected in a specific compartment by a drop in the
opacity when that compartment is off-line for maintenance.20
4.4-33
-------
The instrument readouts are best mounted on a master control panel as
close as possible to other process monitoring displays. The readings of
thermocouples, pressure differential gages, and transmissometers can all be
electronically recorded for permanent records.
4.4.3.7 Fire and Explosion Protection. Provision should be incor-
porated into the fabric filter design to protect personnel and equipment in
the event of an explosion. Such events can occur even in units supposedly
collecting inert particulate matter.24'25
One common means to minimize damage is to install explosion vents to
release generated gases at the onset of an explosion. The two basic types
of vents are "the diaphragms and the free-hanging door. Information con-
cerning the sizing and location of explosion vents is presented in references
37 and 38.
4.4.3.8 Other Factors. Several potential operating problems can be
minimized by means of proper design.
In the case of pulse jet filters, water and oil in the compressed air
lines can be deposited on the interior bag surfaces. Resultant blinding can
be minimized by using driers on the compressers and drains on the bottom of
the supply manifold. The blow tubes should be firmly secured at the far end
so that the shock does not shear the retaining pin and allow the blow tube
to wander.
Abrasion at the bottom of bags in reverse-air and shaker-type collec-
tors can be reduced by design of a good thimble arrangement. The thimbles
should be at least one bag diameter long to prevent abrasion caused by
particulate "turning the corner" at the cell plate and being thrown to the
outside by inertia! force.23'24 The thimbles act as flow straighteners and
protect the bottom of the bag from excessive abrasion. A properly designed
unit is shown in Figure 4.4-20. Note the rounded edge on the top of the
thimble to reduce cutting of the fabric even if tension is not optimum. For
units in which the base snaps into the tube sheet, a thimble can be added
which extends downward to provide the same type of abrasion protection as
that shown in the illustration.
A bypass may be advisable, especially, when process startup or upset
conditions could generate sticky particulate or result in gas temperatures
4.4-34
-------
Clamp
Bag
Thimble
Tube Sheet
Figure 4.4-20. Cross section of a thimble protecting bottom of
bag (reprinted .by permission: Mr. E. W. Stanly).
below the acid vapor or water dew points. These could also be used in con-
junction with a spark sensor to reduce risk of fire.
Inlet and outlet dampers should be provided in compartmented systems to
allow on-line maintenance. The dampers must be designed to provide positive
sealing so as to protect maintenance personnel from toxic gases.
The reverse-air fan provides cleaning of the bags by reversing the
system gas flow. The fan must be designed to deliver the necessary gas
volume at a pressure drop greater than the resistance of the filter.
Welds around tube sheets, thimbles and hoppers should be continuous.
Tack welding leaves gaps through which a relatively large quantity of gas
can pass untreated. The crevices created by task welds also provide sites
for corrosion.
Access doors should be large enough that maintenance personnel can con-
veniently enter while wearing safety equipment such as self-contained re-
breathers. These should be secured by several firm, yet easy-to-remove
4.4-35
-------
latches. The use of a large number of bolts discourages routine access,
which is necessary at most installations. A chain should be available
adjacent to the door to secure the door during periods when personnel are
inside. The chain provides additional security beyond that of the lock-out
system.
Additional information regarding fabric filter design is references 1,
2, 35, 40, 41, and 42. Models are available to predict the operations
characteristics of reverse-air and shaker-type fabric filters.12'13'15
4.4.4 Operation and Maintenance of Fabric Filters
The long-term satisfactory performance of fabric filters is at least
partially dependent on proper operating procedures and a preventive mainte-
nance program.
4.4.4.1 Startup/shutdown. A fabric filter is especially vulnerable
to corrosive vapors and sticky particulate during startup and shutdown. In
some cases, it may be advisable to bypass the collector until the effluent
gas stream temperature is above the acid dew point temperature.39
Prior to startup, a precoat of the material to be collected can be
placed on the bags by the injection of suitable material into the inle^
duct. This precoat is valuable for new fabrics in that it aids in the
conditioning of the fabric. Exposure of the new fabric without the precoat
could lead to deposition of fine particles within the fabric itself or the
collection of hard-to-remove sticky material on the fabric surface.20*23
Typically, a material similar to that to be collected is used as a precoat;
however, prudence should be applied because of potential problems. Use of
materials such as lime, for instance, has resulted in blinding problems
because of the hygroscopic nature of lime which lead to the formation of a
lime mud on the fabric surface.39
4,4.4.2 Fabric tension. The adjustment of bag tension is important
in ensuring adequate bag life and minimum particulate emissions.33*39'40'43
The bag should be tight enough to avoid excessive fiber-to-fiber and bag-to-
bag abrasion, but not so tight as to exceed the tensile strength of the bag
during cleaning. It may be necessary to check bag tension soon after startup.39
4.4.4.3 Cleaning system. Operation of the cleaning system should be
evaluated regularly. Cleaning intensity and frequency can have a direct and
4.4-36
-------
substantial impact on both bag life and emissions. For example, Dennis
and Wilder44 found in one installation that penetration from a shaker-type
unit was related to the shaker amplitude. Ladd, et a!.33 reported that
adjustment of shaker frequency contributed to a reduced rate of bag failure.
The gas flow rate and static pressure available in a reverse air fan can
also be an important operating variable.27
4,4.4.4 Solids removal. Accumulation of solids within the hopper can
lead to major operational problems. A partially filled hopper can lead to
particulate matter reentrainment and abrasion of the lower portions of the
bags.25 The material in the hopper can gradually cool and bridge over.
Ultimately, bridging could lead to a restriction of gas flow to the bags and
to a buildup of material into the bags. Collected material with a substantial
combustible content can be prone to fires.
On a frequent basis, operators should confirm positively that solids are
being discharged. For units with rotary valves, the use of a long (0.5 m),
brightly colored rod attached to the end of the shaft has proved useful in
determining from a distance that a rotary valve has stopped. Motion sensors
can be used on rotary valves and screw conveyors.
4.4.4.5 Fabric and Component Repair. When bag failure is the result
of localized abrasion or mechanical damage, small pinholes or tears are
usually sealed by adhesive and sewn with thread. The adhesive and thread
should be compatible with the original fabric in the properties of shrinkage,
temperature tolerance, and chemical resistance. Successful repair depends
on the strength and condition of the bag. Patching may not be successful
on bags that have been operated for a long period and have undergone chemical
and thermal degradation. Bag repair must be considered relative to the cost
of bag replacement. Repair becomes economically attractive when bags are
extremely large (>30-cm diameter) or when the fabric is expensive.
The blinding of bags because of process upset, operation below dew
point, or moisture inleakage can increase filter pressure drop. If the bags
are new and have not been subjected to chemical or thermal degradation, it
is possible to reduce the fabric resistance by laundering. Although it is
not available for all fabric types, laundering can sometimes make it possible
4.4-37
-------
to put off bag replacement until a later date. Care must be taken to
prevent bag shrinkage or chemical attack during cleaning. Even though the
expected bag life may even be reduced, the overall cost may be lower than
that incurred by total bag replacement or operation at higher static
pressure.
Reuse of bags and cages after a fire in a baghouse is usually not
possible and almost never advisable. High temperatures can warp the cages
>
to the extent that bag-to-bag abrasion results. Cages may be reused only if
they are not corroded or bent. Each cage must be carefully inspected before
Installation.
4.4.4.6 Maintenance Inspection. Regular inspection of the inside of
bags is necessary to confirm that the system is in compliance with
regulatory requirements and that there are no developing problems. Safety
procedures must be strictly followed. Diagnosis of prevailing operation is
done primarily by observation of "clean" side deposits resulting from pene-
tration of dust. If the penetration is local to specific bags or seals, the
pattern created by the dust on the tube sheet may indicate the point of
penetration. Figure 4.4-21 shows the characteristic pattern of a low
velocity dust penetration at the snap-ring attachment in a shaker baghouse.
The small depressions that look like craters are the penetration points. In
the early stages of penetration the pattern can also be highlighted with a
fluorescent dye and an ultraviolet light source. The use of fluorescent
dyes is not practical where total failure of a bag has occurred or the
problem has existed for an extended period.
Moderate clean-side deposits may be caused by a single small pinhole in
a bag. Again, the pattern of dust generated by penetrating gas flow can
indicate the location of holes. Figure 4.4-22 shows the dust pattern
indicated by the clean area on the tube sheet as a result of the gas
impingement. Impaction of the particles on bags on the opposite side of the
collector is indicated by the discoloration. The abrasion caused by this
high-velocity jet results in cascading bag failures in the collector.
Abrasion can occur if an adequately designed precleaner or baffle plate
is not used to remove large, abrasive particles. Figure 4.4-23 shows the
abrasive damage caused by large, sharp particles impinging on the bag surface
4.4-38
-------
Figure 4.4-21. Dust penetration around snap-rin attachment.
(Courtesy of PEDCo Environmental, Inc.)
4.4-39
-------
O>
•«J
O
o
o> •
i— O
o c
o c
•«J 0)
c o
(U J-
O -i-
••-3 C
"O UJ
(O
o
•»J O
d) O
•f-jLLl
o.
CO
CM
CVJ
d)
J_
4.4-40
-------
rs
o
•*•> •
r
fO
•M OJ
« E
O) O
TO S-
« -r-
E >
«S C
c o
o o
«i- Q
(/) UJ
(O O-
»Q <*-
-------
near the bag cuff. In this design the installation of thimble extensions,
blast plate, or precleaner would reduce the abrasive damage and extend bag
life. Most abrasion problems occur near the bottom of the bags directly
above the thimbles. Hopper overflow can also be detected at the bottom of
bags, especially in bags located in corners or along the outer walls of the
compartment.
4.4.4.7 C1earn*ng System Operation. The cleaning system (shaker,
•-:.<:•*
reverse air, pulse jet) should be checked for proper operation. When the
cleaning system malfunctions, an accompanying increase in filter static
pressure drop is usually noted. The motor and cam arrangement should be
inspected for wear and linkage failure. The dampers and fans of reverse-air
systems should be checked for correct operation. The pulsing of individual
solenoid/diaphragm systems on pulse-jet cleaning mechanisms should be
checked. The compressed air pressure should be checked, and proper operation
of compressed air dryers should be verified.
During operation, the stack opacity should be checked regularly. A
spike in visible emissions can often be related to the pulse cleaning of a
specific row of bags, thus aiding in the eventual identification of the bag
or bags causing the problem. On compartment type reverse-air or shaker
units, a reduction in opacity during cleaning of a specific compartment is
indicative of a problem in that compartment.
4.4.4.8 Preventi veMai ntenance. A preventive maintenance program
should be aimed at reducing bag failure. The program should include routine
servicing of mechanical equipment including gears, bearings, and pneumatic
cylinders and also should include a complete external and internal
inspection of the system at frequent regularly scheduled predetermined
intervals.
The operator should maintain records indicating system pressure drop,
temperature, date of bag replacement and location of the bags, and changes
in process operation. These data may then be used to diagnose failure
mechanisms and provide direction in preventing recurrence of failures.
An example of the types of records useful in diagnosing recurring
problems is shown in Figure 4.4-24. The top figure reflects a random type
4.4-42
-------
BAG FAILURE DATA SHEET
Date
System
Mod. No..
OOOOOOOOOO
oooooooooo
~)OO©OOO©OC
oooooooooo
oooooooooo
~)O(§)OOO(S)OOC
oooooooooo
OQO®OOOOOO
oooooooo@o
OOOOOQOOOO
im.tr r
lnsp.by:
oooooooooo
oooooooooo
oooooooooo
oooooooooo
oooooooooo
Figure 4.4-24.
Bag failure location records. (Courtesy of
Richard P. Bundy, Standard Havens, Inc.)
4.4-43
-------
pattern which could be due to a large variety of problems. Symptoms of a
baffle problem are shown in the lower figure. In addition to these figures,
a simple elevation sketch of each bag removed should be prepared showing the
location and type of damage. Another set of diagnostic records which has
proven useful is a record of the frequency of bag failures. This can be
used to identify a condition which has arisen recently or when a set of bags
is reaching the end of the useable life.
4.4-44
-------
REFERENCES
1. Billings, C. and J. Wilder, (GCA Corporation, Boston). Handbook of
Fabric Filter Technology, Vol. I. Prepared for U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina.
Publication No. APTD-0690.
2. Theodore, L. and A. J. Buonicore. Industrial Air Pollution Control
Equipment for Particulates. CRC Press. Cleveland. 1976.
3. Perkins, H. C. In: Air Pollution, A. Stern (ed.). New York,
McGraw-Hill. 1974.
4. Paretsky, L.L., Theodore R. Pfeffer, and A.J. Squires. J. Air
Pollution Control Association. 21:4. April 1971.
5. Anderson, D.M. and L. Silverman. Harvard Air Cleaning Laboratory.
AEC Report No. NYO-4615. 1958.
6, Donavan, R.P., J. H, Turner, and J. H. Abbott. Passive Electrostatic
Effects in Fabric Filtration. Second Symposium on the Transfer and
Utilization of Particulate Control Technology. In: Vol. 1. Control
of Emissions from Coal Fired Boilers. F. P. Venditti, J. A. Armstrong,
and M. Durham (ed.). U.S. Environmental Protection Agency Publication
No. EPA-600/9-80-039a. September .1980. pp. 476-493.
7. Donavan, R.P., R.L. Ogan, and J.H. Turner. The Influence of Electro-
statically Induced Cage Voltage Upon Bag Collection Efficiency
During the Pulse-jet Fabric Filtration of Room Temperature Flyash.
In: Proceedingsof the Third Symposium on Fabric Filters for
Particle Collection. U.S. Environmental Protection Agency, Research
Triangle Park, N.C. Publication No. EPA-600/7-78-087. June 1978.
8. Frederick, E. J. of the Air Pollution Control Association,
24: 1164-1168. December 1974.
9. Frederick, E. Chemical Engineering. 68: 107, June 1971.
10. Turner, J. Extending Fabric Filter Capabilities. J. of the Air
Pollution Control Association. 24: 1182-1187. December 1974.
11. Enson, D. S., R. G. Hooper, and R. W. Schech. Determination of the
Fractional Efficiency, Opacity Characteristics, Engineering and
Economic Aspects of a Fabric Filter Operating as a Utility Boiler.
Electric Power Research Institute. Report EPRI FP-297. November 1976.
12. Dennis, R. and H. Klemm. Modeling Coal Fly Ash Filtration with Glass
Fabrics. Third Symposium on Fabric Filters for Particulate Collection,
N. Surprenant. U.S. Environmental Protection Agency, Research Triangle
Park, N.C. Publication No. EPA-600/7-78-087. June 1978. pp. 13-40.
4.4-45
-------
13. Dennis, R. and H. A. Klemm. Fabric Filter Model Format Change, Vol.
I, Detailed Technical Report. GCA Report to U.S. Environmental Protec-
tion Agency, Contract No. 68-02-2607, Task No. 8. GCA Report GCA-TR-
78~51-G(1), January 1969.
14. Hall, R. R., R. Dennis, and N. F. Surprenant. Fibers, Fabrics, Face
Velocity and Filtration. A Specialty Conference on the User and Fabric
Filtration Equipment III. E. R. Fredrick, (ed.). APCA Specialty
Conference Proceedings. 1978. pp. 156-169.
15. Dennis, R. H. A. Klemm. Verification of Projected Filter System Design
and Operation, pp. 143-160. Symposium on the Transfer and Utilization
of Particulate Control Technology, Vol. 2, Fabric Filters and Current
Trends in Control Equipment. F. P. Venditti, J. A. Armstrong, and M.
Durham, (ed.). U.S. Environmental Protection Agency. Publication
No. EPA-600/7-79-044b. February 1979.
16. Leith, D., M. W. First, M. Ellenbecker, and D. D. Gibson. Performance
of a Pulse-jet Filter at High Filtration Velocities. In: Symposium on
the Transfer and Utilization of Particulate Control Technology:
Vol. 2, Fabric Filters and Current Trends in Control Equipment.
F. P. Venditti, J. A. Armstrong, and M. Durham, (ed.,). U.S. Environmental
Protection Agency. Publication No. EPA-600/7-79-044b. February 1979.
pp. 11-26.
17. Leith, D. and M. J. Ellenbecker. Theory for Penetration in a Pulse-
jet Cleaned Fabric Filter. Paper 80-30.1. Presented at the 73rd Annual
Meeting of the Air Pollution Control Association. Montreal, Quebec.
June 22-27, 1980.
18. Dennis, R., J. W. Wilder, and D. L. Harman. The Mechanics of Pulse-jet
Filtration, Paper 80-30.6. Presented at the 73rd Annual Meeting of the
Air Pollution Control Association. Montreal, Quebec. June 22-27, 1980.
19. Theodore, L. and J. Reynolds. Performance Equations Describing Effects
of Bay Farline on Baghouse Outlet Loadings. Presented at the Air Pollution
Control Association Annual Meeting. Cincinnati, Ohio. June 24-29, 1979.
20. Perkins, R. P. The Case for Fabric Filters on Boilers. Semiannual
Technical Conference on Air Pollution Equipment. Philadelphia, Pennsylvania.
April 23, 1976.
21. Vandenhoeck, P. Cooling Hot Gases Before Baghouse Filtration. Chemical
Engineering. Vol. 79, No. 9: 67-70. May 1, 1972.
22. Bergmann, L. New Fabrics and Their Potential Application. J. Air
Pollution Control Association. 24: 1187-92. December 1974.
4.4-46
-------
23. Perkins, R. P. State-of-the-art of Baghouses for Industrial Boilers.
Presented at the Industrial Fuel Conference. West Lafayette, Indiana.
October 5-6, 1972.
24. Rolschau, D. W. Matching a Baghouse to a Fossil Fuel Fired Boiler.
In: Symposium on the Transfer and Utilization of Particulate
Control Technology: Vol. 2, Fabric Filters and Current Trends in
Control Equipment. F. P. Venditti, J. A. Armstrong, and M. Durham,
(ed.). U.S. Environmental Protection Agency, Research Triangle Park,
N.C., Publication No. EPA-600/7-79-044b. February 1979. pp. 211-217.
25. Brandt, E. F. Operation and Maintenance of Fabric Filters on Coal-
Fired Boilers. In: Proceedings: Operation and Maintenance
Procedures for Gas Cleaning Equipment. E. R. Fredrick (ed.). APCA
Specialty Conference Proceedings. 1980. pp. 153-161.
26. Billings, C. E. and J. E. Wilder. Major Application of Fabric Filters
and Associated Problems. Paper No. 15. In: Proceedings of the
Symposium on Control of Fine-Particulate Emissions from Industrial
Sources. San Francisco, California. January 15-18, 1974. pp. 329-372.
27. Tracy, G. W. Operation and Western Precipitation Fly Ash Baghouses
Since Sunbury. In: Proceedings: Operation and Maintenance
Procedures for Gas Cleaning Equipment. E. R. Fredrick (ed.).
APCA Specialty Conference Proceedings. 1980. pp. 162-181.
28. Mappes, T. E. and R. D. Terns. An Investigation of Corrosion in Parti-
culate Control Equipment. U.S. Environmental Protection Agency,
Research Triangle Park, N.C. Publication No. EPA-340/1-81-002.
February 1981.
29. Prinz, R. T. Reducing Baghouse Maintenance by Design. Minerals
Processing. 11: 8-13. May 1970.
30. E. I. duPont de Nemoursand Company, Bulletin No. E-13544.
31. Menardi and Company. Bulletin. Segundo, California.
32. Pearson, G. L. Experience at Coors with Fabric Filters-Firing Pulver-
ized Western Coal. In: Second Symposium on the Transfer and
Utilization of Particulate Control Technology, Vol. 1, Control of
Emissions from Coal Fired Boilers, F. P. Venditti, J. H. Armstrong, and
M. Durham, (ed.). U.S. Environmental Protection Agency, Research
Triangle Park, N.C. Publication No. EPA-600/9-80-039a. September 1980.
pp. 359-370.
33. Ladd, K. L. Jr., R. Chambers, S. Kunka, and D. Harman. Objectives
and Status of Fabric Filter Performance Study. In: Second Sym-
posium on the Transfer and Utilization of Particulate Control Tech-
nology. Vol. 1, Control of Emission "rom Coal Fired Boilers, F. P.
Venditti, F. A. Armstrong, and M. Durham (ed.). U.S. Environmental
Protection Agency, Research Triangle Park, N.C. Publication
No. EPA-600/9- 0-039a. September 1980. pp. 317-341.
4.4-47
-------
34. Mycock, J. C. A Pilot Plant Study of Various Filter Media Applied
to a Pulverized Coal-fired Boiler In: Symposium on the Transfer
and Utilization of Particulate Control Technology. Vol. 2, Fabric Fil-
tration Current Trends in Control Equipment. F. P. Venditti, F. A.
Armstrong, and M. Durham, (ed.). U.S. Environmental Protection
Agency, Research Triangle Park, N.C. Publication No. EPA-600/7-79-044b.
February 1979.
35. Simon, H. T. Baghouses. In: Air Pollution Engineering Manual.
2nd Edition, J. A. Danielson (ed.). U.S. Environmental Protection
Agency, Research Triangle Park, N.C. Publication No. AP-40.
May 1973.
36. Wright, R. J. Customizing Baghouses for Clinker Coolers, Rock Products.
November 1972. pp. 94-129.
37. Miller, R. L. Explosion Pressure Relief. In: Proceedings: The
User and Fabric Filtration Equipment III. E. R. Fredrick (ed,).
APCA Specialty Conference Proceedings. 1978. pp. 98-112.
38. Reinauer, T. V. Guidelines in Application of Explosion Vents. In:
Proceedings, The User and Fabric Filtration Equipment IV. E. R. Fredrick
(ed.). APCA Specialty Conference Proceedings. 1978. pp. 113-119.
39. Perkins, R. P. and J. F. Imbalzano. Factors Affecting Bag Life
Performance in Coal-Fired Boilers. In: Proceedings, The User and
Fabric Filtration Equipment III. E. R. Fredrick (ed.). APCA
Specialty Conference Proceedings. 1978. pp. 120-144.
40. Szabo, M. F. and R. W. Gerstle. Operation and Maintenance of Partic-
ulate Control Devices on Coal-Fired Utility Boilers, U.S. Environmental
Protection Agency, Research Triangle Park, N.C. Publication No.
EPA 600/2-77-129. July 1977. pp. 3-85.
41. Strauss, W. Industrial Gas Cleaning. Pergamon Press. New York, 1975.
42. Smith, G. L. Engineering and Economic Considerations in Fabric Fil-
ters. J. Air Pollution Control Association. 24: 1154-56.
43. Bundy, R, P. Operations and Maintenance of Fabric Filters. In:
Proceedings, Operation and Maintenance Procedures for Gas Cleaning
Equipment. E. R. Fredrick, (ed.). APCA Specialty Conference
Proceedings. 1980. pp. 139-152.
44. Dennis, R. and J. E. Wilder. Factors in the Collection of Fine
Particulate Matter with Fabric Filters. Paper No. 17. In:
Proceedings, Symposium on Control of Fine-Particulate Emissions from
Industrial Sources. San Francisco, California. January 15-18, 1974.
pp. 385-424.
4.4-48
-------
4.5 WET SCRUBBERS
Wet scrubbers comprise a set of control devices with similar particle
collection mechanisms, primarily: inertia! impaction and Brownian diffusion.
Accordingly, these scrubber systems generally exhibit strong particle-size-
dependent performance. Among scrubber types substantial differences exist
with regard to their effectiveness, the greatest differences occurring in
the particle size range of 0.1 to 2 umA.
The various types of commercially available wet scrubbers are described
in Section 4.5.1. Fundamental operating principles are presented in
Section 4.5.2. Parameters of interest and performance limits are discussed.
Emphasis is on the capability for collection of particles smaller than 5 umA
in diameter.
Considerable progress made in the understanding of wet scrubber per-
formance since the first edition of this document was published is reflected
in practical design considerations in Section 4.5.3. Operation and mainte-
nance factors that enhance long-term performance are presented in
Section 4.5.4.
4.5.1 Types of Particulate Scrubbers
In this discussion major categories of scrubbers are grouped on the
basis of similar mechanisms. In Table 4.5-1, major categories of scrubbers
are listed in order of increasing performance capabilities and energy
requirements.
Scrubber liquids are used for particle collection in several distinct
ways. The most common method is to generate droplets, which are then
intimately mixed with the gas stream. Particles are also collected on water
layers or sheets surrounding of packing material by directing the particle-
laden gas stream through an intricate path around the individual packing
elements. A third method is to pass high-velocity gas through a vapor to
generate "jets" of liquid to collect particles. This is the least common of
the three liquid characteristics.
4.5.1.1 Preformed Spray Scrubbers. Preformed spray scrubbers require
the least energy of the various scrubbers, and they consequently allow the
4.5-1
-------
TABLE 4.5-1. MAJOR TYPES OF WET SCRUBBERS*
Category
Particle capture
mechanism
Liquid collection
mechanism
Types of scrubbers
Preformed-spray
Packed-bed
scrubbers
Tray-type
scrubbers
Mechanically
aided scrubbers
Venturi and
orifice
scrubbers
(gas atomized
scrubbers)
Inertia!
impaction
Inertia!
impaction
Inertial
impaction
Diffusional
impaction
Inertial
interception
Inertial
impaction
Diffusional
impaction
Droplets
Sheets, droplets
(moving bed
scrubbers)
Droplets, jets,
and sheets
Droplets and
sheets
Droplets
Spray towers
Cyclonic spray
towers
Vane-type cyclonic
towers
Multiple-tube
cyclones
Standard packed-bed
scrubbers
Fiber-bed scrubbers
Moving-bed scrubbers
Cross-flow scrubbers
Grid-packed scrub-
bers
Perforated-plate
Impingement-plate
scrubbers
Horizontal impinge-
ment-plate (baffle)
scrubbers
Wet fans
Disintegrator
scrubbers
Standard venturi
scrubbers
Van" able-throat
venturi scrubbers:
flooded disc, plumb
bob, movable blade,
radial flow, varia-
ble rod
Orifice scrubbers
List not intended to be all inclusive.
highest particulate penetration, especially of small-diameter particles.
Most preformed spray scrubbers are highly efficient only for particles
larger than 5 pmA in diameter.1-3
4.5-2
-------
4,5.1.1.1 Spray tower. A spray tower is the simplest type of scrubber,
consisting of a chamber containing an array of spray nozzles (Figure 4.5-1).
GAS OUTLET
DEMISTER
LIQUOR INLETS
GAS INLET
LIQUOR OUTLET
Figure 4.5-1. Spray tower scrubber
Particulate-laden gases pass vertically up through the tower while the
liquid droplets fall by gravity down through the gas flow. Particles
collide with the droplets, are collected into the liquor, and carried out of
the scrubber. Collection is limited by the terminal settling velocity of
the droplets. The spray tower scrubber has low particle removal capability,
but it is often useful for treating effluent gas streams having high mass
loadings of large diameter particulate matter.4
4.5-3
-------
4.5.1.1.2 Cyclonic spray tower. A cyclonic spray tower is similar to
the spray tower scrubber except that the gas stream is given a spiral motion.
In one typical configuration particulate-laden gases enter the scrubber
vessel tangentially at the bottom and pass upward in a spiral motion around a
centrally located array of spray nozzles. Droplet migration is crosscurrent
to the gas flow. Cyclonic spray towers generally operate with a static
pressure drop of 1 to 2 kPa.3 Penetration of particles less than 2 umA in
diameter is typically quite high.5
4.5.1.1.3 Vane type Cyclonic Spray Tower. A vane-type cyclonic scrubber
utilizes a system of vanes rather than a tangential inlet to impart cyclonic
motion (Figure 4.5-2).
4.5.1.1.4 Multiple-tube Cyclones. Another type of cyclonic scrubber
consists of multiple miniature tubes, each with a separate liquid supply.
In this design the gases flow in a downward pattern in contrast to the
upward flow of other cyclones.
4.5.1.2 Packed BedScrubbers. In the typical packed-bed scrubber
liquid introduced near the top trickles down through the packed bed. The
liquid flow spreads over the packing into a film with a large surface area
(Figure 4.5-3). The liquid can be introduced concurrent or crosscurrent
with the gas flow. Packing materials include raschig rings, pall rings,
berl saddles, tellerettes, intalox saddles, and materials such as crushed
rock.3 Packed beds are also constructed with metal grids, rods, or fiberous
pads. These scrubbers are often used for gas transfer or gas cooling, both
of which are facilitated by the large liquid surface area provided on the
packing.3
Plugging of a bed can occur if the gas to be treated is too heavily
laden with solid particles.3'4 A general rule for many applications is to
limit the use of packed beds to service in which particulate concentrations
are less than 0.45 g/m3. Moving-bed scrubbers that have less propensity for
plugging (Figure 4.5.4) are packed with low-density plastic spheres, which
are free to move within the packing retainers.
Packed-bed scrubbers are reported to have low penetration for particle
sizes down to 3 umA and can sometimes remove a significant fraction of
particulate in the range of 1 to 2 umA. The standard countercurrent
' 4.5-4 "
-------
CLEAN GAS
OUTLET
DIRTY GAS
INLET
CYCLONIC LIQUID
ENTRAINMENT
SEPARATOR
SPINNING VANES
LIQUOR INLET
LIQUOR OUTLET
Figure 4.5-2. Vane type scrubber (courtesy of the
Ducon Company, Inc.)*
4.5-5
-------
GAS OUTLET
DEMISTER
GAS INLET
LIQUOR INLET
LIQUOR OUTLET
Figure 4.5-3. Packed tower scrubber (courtesy of Air
Pollution Industries, Inc.)
-------
CLEAN AIR OUTLET
SCRUBBING
LIQUOR
RETAINING
GRIDS
MIST ELIMINATOR
MOBILE PACKING
SPHERES
HOT GAS INLET
Figure 4.5-4.
SLURRY DISCHARGE
Moving-bed scrubber (courtesy of UOP - Air
Correction Division).
4.5-7
-------
arrangement requires the greatest liquid flow and can best handle heavier
loadings. Crosscurrent packed-bed scrubbers require much less liquid flow,
usually operate at lower static pressure drops, and rarely suffer from
plugging. Concurrent packed-bed scrubbers are reportedly more efficient
than other packed-bed scrubbers for the smaller particulates, but they
typically operate at higher pressure drops.
4.5.1.3 TrayType Scrubbers. A tray-type scrubber typically consists
of a vertical tower with one or more perforated plates mounted inside trans-
versely to the shell. In such a scrubber the liquid flows from top to
bottom, and the gas flows from bottom to top. Gases in the scrubber mi*
with the liquid passing through the openings in the plates.
The perforated plates of a tray-type scrubber are often equipped with
impingement baffles or bubble caps over the perforations (Figure 4.5-5).
The gas passing upqard through a perforation is forced! to turn 180 degrees
into a layer of liquid. The gas bubbles through the liquid, and particulate
is collected in the liquid sheet. The impingement baffles are below the
liquid level on the perforated plates and are, therefore, continuously
washed clean of collected particles. Penetration through a typical impinge-
ment plate is low for particles larger than 1 umA,3 but penetration of sub-
micrometer particulate is higher than with some higher-energy scrubbers.
Pressure drop through a typical baffle plate is roughly 0.4 kPa per stage.3
Addition of plates increases the scrubber pressure drop, but does not pro-
portionally decrease the penetration of submicrometer particulate.6
One additional variation of the tray scrubber is the horizontal baffle-
type scrubber. In this type of scrubber the direction of gas flow is hori-
zontal and the baffle section is mounted vertically .in the scrubber. The
scrubbing liquid is introduced concurrently with the gas.
4.5.1.4 Mechanically Aided Scrubbers. The mechanically aided scrubbers
utilize a mechanical rotor or fan to shear the scrubbing liquid into dispersed
droplets. These scrubbers use a specially designed stator and rotor arrangement
to produce very finely divided liquid droplets that are effective in capture
of fine particulate. The low penetration of fine particulate, however, is
achieved at a high energy cost.1-3 Because both wet-fan and disintegrator-
4.5-8
-------
GAS OUTLET
DEMISTER
SAS
INLET
LIQUOR
INLET
LIQUOR
OUTLET
Figure 4.5-5. Tray scrubber (courtesy of the Koch
Engineering Company).
4.5-9
-------
type mechanically aided scrubbers are subject to particulate buildup or
erosion at the rotor blades, they are often preceded by precleaning devices
for removing coarse particulate.1'4 Mechanically aided scrubbers generally
do not perform well in air containing more than 1 g/m3 of particulate,1
4,5.1.5 Venturi and Orifice Scrubbers. Venturi and orifice scrubbers
are perhaps the most common particulate removal devices, in part because
they allow lower penetration of small particles than most other types of
scrubbers. These scrubbers accomplish superior particulate collection by
generating small liquid droplets in the turbulent zone in a manner that
creates a high initial relative velocity between the droplets and the parti-
culate. Inertia! impaction capture of particulate by the scrubbing liquid
is more efficient in these highly turbulent processes, but a price is paid
in energy consumption to achieve the low penetration.
,. .4.5.1.5.1 Ve.ntuH. scrubbers. The sjmple venturi scrubber, often
called a gas atomizing spray scrubber, consists of a series of sprays upstream
from a converging and diverging "throat" section (Figure 4.5-6). As the gas
approaches the venturi throat, the velocity and turbulence increase. The
high gas turbulence atomizes the liquid into small droplets and increases
interaction between the droplets and the particulate. Pressure drops in
venturi scrubbers can range from less than 1 to 40 kPa.
4.5.1.5.2 Variable-throat venturi scrubbers. Pressure drop and venturi
performance are partially dependent on gas velocity through the venturi.
Several variations of the standard venturi scrubber have been developed to
allow the venturi throat dimensions to be changed as the rate of gas flow
changes. Among these scrubbers are the plumb-bob venturi, the flooded-disc
venturi, the moveable-blade venturi, the radial-flow venturi, and the variable-
rod venturi. Several of these venturi throats are illustrated in Figure 4.5-7.
4.5.1.5.3 Orifice scrubber. In an orifice scrubber, sometimes referred
to as an entrainment or self-induced spray scrubber, the gas stream passes
over a pool of scrubbing liquid at high velocity just before entering an
orifice. The high velocity of the gas induces ("entrains") a spray of
scrubbing liquid droplets, which interact with the particulate in and
immediately after the orifice. Orifice scrubbers have moderate pressure
drops (0.8 to 4.0 kPa) and low penetration of particulate 2 to 3 ymA in
diameter and larger.
4.5-10
-------
4.5.2 Operating Principles of Particulate Scrubbers
The fundamental principles that govern particle penetration and total
static pressure drop are examined both in a general manner and with respect
to classes of wet scrubbers.
4.5.2.1 Penetration. Particulate matter collection in wet scrubbers
is highly size-dependent because of the fundamental characteristics of the
inertia! impaction and diffusion processes. Figure 4.5-8 illustrates the
theoretical single-droplet collection efficiency resulting from these two
phenomenon as calculated by Crawford7 for an example case. This particular
example illustrates a predicted minimum collection efficiency reached at
Q.l-umA particles; the size at which both mechanisms become ineffective.
This general curve applies to most scrubbers; however, the location and
magnitude of the efficiency minimum depends on the specific unit.
GAS INLET
LIQUOR
INLET
GAS OUTLET
LIQUOR OUTLET
Figure 4.5-6. Venturi scrubber.
4.5-11
-------
DIRTY GAS IN
>x^f SPRAYS
DIRTY GAS IN
MOVABLE
.VENTUR1
BLADES
TO LIQUID ENTRAINMENT
SEPARATOR
a. Movable-blade venturi
PLUMB BOB
ACTUATOR'
LIQUOR
INLET
MOVABLE
•PLUMB
BOB
TO LIQUID ENTRAINMENT
..• .SEPARATOR
b. Plumb-bob venturi
DIRTY GAS IN
DIRTY GAS IN
LIQUOR
INLET
V N
1
MUVAbLl
DRUM
TO LIQUID ENTRAINMENT
SEPARATOR
c. Radial-flow venturi
\
LIQUOR
INLET
-LIQUOR
SPRAY
MOVABLE
LOOPED
DISK
DI!
--*" TO
LIQUID
NTRAINMENT
SEPARATOR
T
d. Flooded-d1sc venturi
Figure 4.5-7. Throat sections of variable throat venturi scrubbers
(courtesy of Industrial Gas Cleaning Institute, Inc.).
4.5-12
-------
UJ
UJ
ex.
1.0
0.1
10-
10"
10-
10'
10-*
0.01
Z
0.1
1.0
10
PARTICLE DIAMETER
Figure 4,5-8. Theoretical single-drop collection efficiency
due to diffusion and impaction (reprinted by
permission: Crawford, M. Air Pollution Control
Theory, McGraw Hill Co., New York, 1976).
4.5-13
-------
Analysis of the particle collection capability of wet scrubbers can be
based on (1) the fundamental particle collection mechanisms and (2) the
empirical contact power approach. The latter method is based on the premise
that penetration is proportional to the power expended in the scrubber.8-13
This premise is logical because high energy consumption implies high
-fflj-
relative gas-water velocities, high water utilization, and fine droplet
formation, all of which favor impaction, the dominant collection mechanism.
Limitations of the contact power analysis can be attributed to the difficulty
of handling nonideal operating conditions such as pooir gas-liquid distribution
and particle shattering during high-energy scrubbing. Also, this type of
analysis is not amenable to situations in which particle collection mechanisms
other than inertia! impaction are important.
Penetration analyses based on the fundamental particle collection
mechanisms involve the identification of the dominant physical phenomenon
leading to particle capture. Following is a partial list of the collection
mechanisms. t
Collection medium Capture phenomenon
Droplets Inertial impaction
Interception
Brownian diffusion
Liquid sheets (layers) Inertial compaction
Interception
Brownian diffusion
Electrostatic attraction
Liquid sheets Inertial impaction
Interception
Diffusion
Bubbles Inertial impaction
Interception
Brownian diffusion
Electrostatic attraction
For each control device, penetration relationships are based on anticipated
particle collection mechanisms. The accuracy of the resulting equations
depends on the proper assignment of the mechanisms and on the accuracy of
the mechanism expressions. Penetration expressions presented in the Scrub-
ber Handbook3 for selected wet scrubbers are provided in the following para-
graphs.
4.5-14
-------
4.5.2.1.1 Preformed-spray scrubbers. The most effective particle
collection mechanism is inertia! impaction to liquid droplets. The
penetration calculations depend on droplet size and flow characteristics
(i.e., countercurrent, cocurrent, and crossflow). For countercurrent
conditions, the following equation is applicable:
(Equation 4.5-1)
where
P. = e -
"0.75 vt
. rd(vt -
i^r
V.
V
.V
droplet terminal settling velocity, cm/s.
impaction parameter, dimension! ess.
scrubber height, cm.
droplet radius, cm.
gas superficial velocity, cm/s.
v. =
r\, -
Z =
r . =
VG =
2± - liquid-to-gas ratio, dimensionless.
The parameter n. is calculated according to Equation 4,5-2.
where
Ki
(K.. + 0.7)
(Eq. 4.5-2)
Ki
impaction parameter for particles having aerodynamic diameters
of i.
The variables that control the penetration rates from a preformed-spray
scrubber include scrubbing zone height, superficial gas velocity, particle
aerodynamic diameter, liquid-to-gas ratio, and spray droplet size. These
variables are equally important in cocurrent and crosscurrent preformed spray
scrubbers, which are discussed further in Calvert et al.3 These scrubber
types share the characteristic penetration curve presented earlier for
impingement-plate scrubbers.
4.5.2.1.2 Packed-bed scrubbers. The packed-bed scrubber primarily
utilizes inertia! impaction of particles on water "sheets" on the packing
material. Turbulent diffusion may contribute some additional particle
capture in the <0.2-umA size range. Equation 4.5-3 is based strictly on
impaction.
4.5-15
-------
f—ifr-rl
L 2(j+r) (e-Hd) J
P1 « e -
(Eq. 4.5-3)
where
P. = penetration value for a specific particle size, dimensionless.
j = "channel" width factor, dimensionless.
e = bed porosity, dimensionless.
H . = liquid holdup in bed, dimensionless.
Z = height of packed section, m.
d = packing element diameter, m.
K. = inertia! impaction parameter, dimensionless. '.
The parameter K. is common to most penetration equations applicable to con-
trol devices based on impaction. The higher the value of K., the higher the
collection efficiency for particles with an aerodynamic diameter of i. The
impaction parameter is a strong function of particle size, as indicated in
Equation 4.5-4.
M
(Eq. 4.5-4)
where
v = gas velocity through the bed, calculated as the volumetric flow
S3 .
rate divided by the scrubber cross-sectional area, cm/s.
u ss gas viscosity at actual temperature, g/s-m.
d. = aerodynamic particle diameter i, urn (g/cm3) ' .
Combining Equations 4.5-3 and 4.5-4 provides a means of calculating penetra-
tion for specific particle sizes, which can then be added to determine total
penetration. Typical values for the parameter, j, are 0.160 to 0.190.3 The
smaller the packing material, the lower j should be. Bed porosity, e, also
depends on the packing size and normally varies from 0.60 to 0.95.3 Calvert
et al. list bed porosities for use in selecting types and sizes of packing.
Liquid holdup is normally assumed to be negligible.3
Figure 4.5-9 presents a typical penetration curve for a packed tower.
This theoretical curve is based on the above mathematical relationships and
use of the parameter values specified. It is apparent that there is a very
4.5-16
-------
1.20
1.00-
0.80
0.60-
0.40-
0.20-
300 eM/sec
0.165
0 1 23 4 5
AERODYNAMIC DIAMETER, ymA
Figure 4.5-9. Theoretical penetration curves for various-sized
packed-bed scrubbers.
strong interdependence between particle size and penetration. Actual pene-
tration values for a specific facility (at a specific time) might vary by
more than a factor of 2 from those calculated from Equation 4.5-3. Never-
theless, the penetration curve will resemble that of Figure 4.5-11. Gener-
ally, packed-bed scrubbers are relatively ineffective for sources with a
major fraction of the particulate emissions in the <2.0-ymA size range.
4.5-17
-------
Optimization of an existing unit could be done by changing gas velocity, bed
height, and packing type. The liquid-to-gas ratio (L/G) is not a control-
ling factor with the packed bed scrubber, as it is with other scrubber
tvoes.
4.5.2.1.3 Tray-type scrubbers,. An impingement-plate scrubber is typical
of penetration conditions created in a tray-type scrubber. The theoretical
penetration of this type of scrubber is illustrated in Figure 4.5-10. The
dominant particle-collection mechanism is inertial impaction to droplets
formed as the gas stream passes the impingement hole. The penetration rate
of particles of aerodynamic diameter i is calculated according to Equation
4.5-5.
P. = 1 -
Ki
K. +0.7
°'5 (Eq. 4.5-5)
where
P. - penetration through an impingement plate scrubber of a particle
with aerodynamic diameter of i, dimensionless.
K. - impaction parameter of a particle with aerodynamic diameter of i,
dimensionless.
The impaction parameter is calculated by use of Equation 4.5-4 with the
velocity term being the velocity of the gas at the vena contracta after it
passes through the hole. Calvert et al. suggest using a typical vena con-
tacta velocity of 1.43 times the gas velocity in the hole.3 In Equation
4.5-4 the "collector" diameter is taken as diameter of the impingement hole.
The theoretical penetration derived by use of Equation 4.5-5 is shown
in Figure 4.5-10. A comparable data set14 for a full-scale impingement
scrubber or a rotary salt dryer is shown in Figure 4.5-11.
Performance of the impingement plate scrubber is controlled by the same
basic factors as those of other wet scrubbers utilizing inertial impaction to
water droplets. Those factors are aerodynamic particle size, relative
velocities of particles and droplets, and liquid-to-gas ratio. The impinge-
ment plate scrubbers and the entire family of tray scrubbers operate the
inertial impaction mechanisms somewhat less effectively than the venturi
scrubbers; accordingly, they allow higher penetration in the range of 0 to 1
pmA.
4.B-18
-------
1.0
0.8
0,6
0.4
0.2
91,100 acfm @ 170°F
1,000 gpm LIQUOR FLOW
3/32-in. HOLES ON 3/16-iru CENTERS
13-ft. DIAMETER TRAYS
I
1234
AERODYNAMIC PARTICLE DIAMETER, ymA
Figure 4.5-10. Theoretical penetration curve for impingement
plate scrubber.
4.5-19
-------
1.0
0.8
0.6
0.4
0.2
TWO-STAGE IMPINGEMENT SCRUBBER
1-41 Nm3An1n
0.035 nr LIQUOR/min.
Ap = 3kPa
1234
AERODYNAMIC PARTICLE DIAMETER, ymA
Figure 4.5-11. Penetration curve for an impingement plate
scrubber on a rotary salt dryer.
4.5-20
-------
4.5.2.1.4 Venturl and Orifice Scrubbers The penetration equation
developed by Calvert et al. for venturi and orifice scrubbers is Equation 4.5-6,
based on inertia! impaction of particles onto water droplets.3
0.036
-0.7-k.f+ 1.4 In
1 V 0.7 / MCff +0.7,
0.49
1 1
where
P. = penetration value for particles with aerodynamic diameters of i.
p = droplet density, kg/m3.
d. - droplet diameter, m.
v = superficial gas velocity in venturi throat, m/s.
jj = gas viscosity, kg/m-s.
Q,/Q = liquid-to-gas ratio, dimensionless.
k- = impaction parameter, dimensionless.
f = nonuniformity correction factor, dimensionless.
Selection of a proper f factor is important in making an accurate
performance prediction. Calvert et al. suggest f values ranging from a low
of 0.10 to a high of 0.70 with typical values being 0.25 to 0.50.3 Low
values are applicable to scrubbers having high liquid-to-gas ratios and/or
having hydrophobic particles. Increasing f values leads to substantially
reduced penetration predictions. i The droplet diameter is normally calculated
using the equation of Nukiyama and Tanasawa presented below as Equation 5.4-7.
9
where
d . = Sauter mean droplet size, cm.
VQ = gas velocity in venturi throat, cm/s.
Q-./Q = liquid-to-gas ratio, dimensionless.
•3
4.5-21
-------
Boll et al.ls determined that Eq. 4.5-7 Is most accurate at a throat veloc-
ity of 4570 cm/s. At throat velocities and liquid-to-gas ratios typical of
commercial units the mean droplet size is estimated with an accuracy +50
percent. As an alternative, Boll suggests Equation 4.5-8.
to o
Iy1/g
drf = 283.000 + 793 1/g/ (Eq. 4.5-8)
vg
where
d . = sauter mean droplet size, urn.
Vg = gas velocity in throat, ft/s.
QL/QQ = liquid-to-gas ratio, gal/1000 ft3.
The impactor parameter takes into account the fundamental variables that
influence inertial impaction, namely, aerodynamic particle size and par-
ticle-droplet relative velocity. This is parameter calculated from Equation
4.5-9.
V(d.)2
(Eq. 4.5-9)
where
K. = impaction parameter, dimensionless.
—4 0 *>
d. = aerodynamic particle diameter i, cmxlO (g/cm3) .
u = gas viscosity at actual temperature, g/s-cm.
d . = sauter mean droplet size, cm.
V., = particle-droplet relative velocity, cm/s.
The theoretical penetration curves that can be generated from Equations
4.5-6 to 4.5-9 are shown in Figures 4.5-12 and 4.5-13. These figures illus-
trate the strong influence of throat velocity and liquid-to-gas ratio in
these equations. Results of this approach are conveniently summarized in
Figure 4.5-14 for a case with f = 0.25.3'16
A comparison of Cal vert's model (Equations 4.5-6, 4.5-7, and 4.5-9) and
actual scrubber performance data17 is shown in Figure 4.5-15. The empirical
f factor strongly affects the degree in which the model fits a specific
scrubber.
4.5-22
-------
8-1 » 1.5 liters/m3
THROAT VELOCITIES
1234
AERODYNAMIC PARTICLE DIAMETER, ymA
Figure 4.5-12. Theoretical penetration curve for a venturi scrubber
illustrating effect of throat velocity.
4.5-23
-------
VG - 10,000 m3/sec
f = 0.25
T = 25°C
123
AERODYNAMIC PARTICLE DIAMETER,
Figure 4.5-13. Theoretical penetration curves for venturi scrubber
illustrating effect of liquid-to-gas ratios.
4.5-24
-------
ce:
UJ
«=c
(—1
a
I—
o
o
Q
O
Oi
LU
-------
1.0
o
2 o.i
tj
UJ
0.
0.0
Calvert's model
f = 0.5
Experimental Data
0.1
Calvert's model
f « 0.25
1.0
PARTICLE DIAMETER, ymA
10
Figure 4.3-15.
Comparison of Calyert's Model results against
measured penetration data.
4.5-26
-------
18
Yung, Calvert, and Barbarika have presented a refined model, which
incorporates a number of changes from the Calvert model, including elimination
of the f parameter. A simplified equation is presented in Equation 4.5-10
for a scrubber in which all particle capture occurs in the throat. It should
be noted that the inverse tangent function should be expressed in radians.
K i 0.5-
(Eq. 4.5-10)
(Eq. 4.5-11)
xgJ LHgJ LUDOJ
C d..2 p_u^
,. 4.5-12)
CDO = 55/NRfi (for 100 < NRfi <500) (Eq. 4.5-13)
ude = 2 [1 - x2 + (X4 -X2)0'5] (Eq. 4.5-14)
x = 1 + 0.187 tDO pg (Eq. 4.5-14)
dd pl
where
1. = venturi throat length, cm.
Cno = drop drag coefficient at throat inlet, dimensionless.
p = gas density, g/cm3.
y
p, = drop density, g/cm3.
drf = drop diameter, cm.
NR = drop Reynolds number, dimensionless.
Mq = gas viscosity, kg/cm-s.
Up. = gas velocity in throat, cm/s.
Q, = liquid flow rate, cms/s.
Q = gas flow rate, cm3/s.
u*S = ratio of liquid drop velocity at throat exit to gas velocity at
exit, dimensionless.
C = Cunningham correction factor, dimensionless.
4.5-27
-------
Figure 4.5-16 is a comparison of the Yung et al. model against the same
scrubber performance data presented earlier in Figure 4.5-15. The revised
model is considered more accurate than the earlier Calvert model, however,
it is also much more complex.18 In another comparison with field units,
Calvert, Barbarika and Monahan17 have concluded that the revised model
adequately predicts penetration with the following qualifications.
1.
1.0
o
5 0.1
CC
tj
tu
a,
ui
fe
o.oi
Actual penetration of submicrometer particles is less than pre-
dicted.
2. Actual penetration of particles greater than 1 urn is greater than
predicted.
3. Accurate measurement of the liquid-to-gas ratio improves the
predicted penetration curve.
Experimental data
Predicted by
Equation 5-23
and by infinite
throat model
0.1
1.0
PARTICLE DIAMETER, yroA
10
Figure 4.5-16. Comparison of Yung, Calvert, and Barbarika
Model against measured penetration data.
4.5-28
-------
Optimization of venturi scrubber design based on modification of Cal-
vert's model3 has been discussed by Leith and Cooper.19 Other theoretical
approaches for calculating particle collection in venturi scrubbers have
been described by Crawford7 and Strauss.10
An empirical approach based on the "contact" power utilized has been
described by a number of investigations. The contact power approach is
based on the concept that penetration is directly related to the energy
input into the gas-liquid contact.8-13 For gas-atomized scrubbers, the
power consumption, PC is approximated by Equation 4. 5-16. 12
pQ = 0.158 AP ' (Eq. 4.5-16)
where
\
PG = power consumed, hp/1000 acfm.
AP = gas phase pressure drop, in. W.C.
When the liquid stream adds a significant fraction of the total energy,
equations presented in references 12 and 13 may be used. The total energy,
PT, is the sum of the gas- and liquid-phase power consumption.
PT = PG + PL (Eq. 4.5-17)
The scrubber collection efficiency, NT, is then expressed as shown in
Equation 4.5-18.
T
where
N = aP^ (Eq. 4.5-18)
NT = number of transfer units (^()), dimension-
less.
PT = total power consumption, hp/1000 acfm.
a,y = constants, dimensionless.
The constants are parameters dependent on the characteristics of the partic-
ulate matter. Using this approach, one assumes that no independent effects
can be attributed to throat velocity, liquid-to-gas ratio, scrubber design
and other parameters.
In certain cases good correlations can be achieved using this approach.
Figures 4.5-17 and 4.5-18 show the relationship between outlet loadings and
static pressure drop (an approach similar to Equation 4.5-18). Hesketh11
4.5-29
-------
0.20
0.10
0.09
~ 0.08
o
0.07
0.06
0.05
0.04
10
grains/DSCF = 12.44 (AP, inches w.c.)"1*61
o
o r Correlation = 0.756
20 30
PRESSURE DROP, 1ru w.c.
40
50
Figure 4.5-17. Comparison of venturi scrubber outlet loadings to
static pressure drops for oil-fired lime kilns.
4.5-30
-------
0.08
0.01
2 345
THEORETICAL POWER CONSUMPTION, hp/1,000 icfm
*
Figure 4.5-18. Correlation of coal-fired boiler scrubber outlet dust
loading with theoretical power consumption.12
4.5-31
-------
has developed an empirical equation (4.5-19) relating static pressure drop
and penetration. As shown in Figure 4.5-19 there is a good relationship
between the two.
-1 A.1
Pt = 3.47 AP •L*^ (Eq. 4.5-19)
A more complete description of the contact power approach is available in
references 8 through 13.
4.5.2.1.5 Other scrubber types. Many other scrubber types are
available, for most of which the penetration curves are similar to those
presented in this section. Additional information on wet scrubber operating
principles is given in References 1, 3, 5, 7, and 10,,
4.5.2.2 StaticPressure Drop. Factors affecting the static pressure drop
in a wet scrubber include scrubber geometry, gas velocity, and the liquid-
to-gas ratio. Typical liquid-to-gas ratios are listed in Table 4.5-2 and
typical static pressure drops are listed in Table 4.5-3. Calvert et al.3
have summarized equations useful for predicting pressure drop in various
types of scrubbers. A detailed summary of static pressure drop equations
applicable to venturi scrubbers is presented by Yung, Calvert, and
Barbari ka.1S ;
TABLE 4.5-2. TYPICAL LIQUID-TO-GAS RATIOS FOR WET SCRUBBERS
Liquid-to-gas ratio,
Scrubber type liters/m3
Venturi 0.70 - 1.00
Cyclonic spray tower 0.70-1.30
Spray tower 1.30 - 2.70
Moving bed 1,30 - 2.70
Impingement plate 0.40 - 0.70
Packed bed 0.10 - 0.50
4.5-32
-------
1.000
0.500
0.200
0.100
O.OBO
UJ
Q-
0.020
0.010
0.005
0.002
1
2 4 6 8 10 20
STATIC PRESSURE DROP, in, w.c.
60 80 100
Figure 4.5-19. Comparison of predicted penetration as calculated in
Equation 4.5-19 and measured penetration (reprinted by permission
from Hesketh, J. Air Poll. Control Assoc. Vol. 24, No. 10, 1974).
4.5-33
-------
TABLE 4.5-3, TYPICAL SCRUBBER PRESSURE DROP
Pressure drop,
Scrubber type kPa
Venturi 1.5 - 18.0
Centrifugal (cyclonic) spray 0.25 - 0.8
Spray tower 0.25 - 0.5
Impingement plate 0.25 - 2.0
Packed bed 0.25 - 2.0
Wet fan 1.0 - 2.0
Self-induced spray (orifice) 0.5 - 4.0
Irrigated filter (filter bed
scrubber) 0.05 - 0.8
Calvert et al.3 presented a simple approach for calculation of venturi
static pressure, as shown in Equation 4.5-20.
AP = 0.001 V*(Q,/QJ (Eq. 4.5-20)
t I y
where
AP = static pressure drop, cm. W. C.
V. - throat velocity, cm/s.
t !
Q,/Q = liquid-to-gas ratio, (cms/s)/(cm3/s).
Yung et al.18 have discussed a modified form of this equation as shown in
*
Equation 4.5-22. The parameter u. is the same as that described earlier
with respect to Equation 4.5-14.
AP = 0.001 ue Vt (Q/Qg) (Eq- 4.5-21)
Hesketh11 described an equation including throat area, Equation 4.5-22.
AP = (Vt2pGA0'133L°*78)/1270 (Eq. 4.5-22)
where
AP = static pressure drop, in. w. c.
V. = throat velocity, ft/s.
PQ = gas density, lb/ft3.
A = throat cross-sectional area, ft2.
L = liquid-to-gas ratio, gal/1000 acfm.
4.5-34
-------
4.5.3 Design of Particulate Scrubbers
4.5.3.1 Sizing of Wet Scrubbers. The overall dimensions of a scrubber
are established to provide the design gas velocity within the various sec-
tions of the vessel. Care must be taken to ensure an even gas velocity
throughout the mixing and demisting sections. Sharp turns should be avoided
in ductwork, and distribution baffles should be used where short-circuiting
would otherwise be likely to occur. In general, scrubbers operate at higher
average gas velocities than fabric filters or ESP's and therefore are usual-
ly more compact.1
4.5.3.2 Nozzle Selection and Liquid Distribution. The liquid distri-
bution system in a scrubber is intended to provide even distribution of
properly sized liquid droplets for contacting particulate in the mixing
zone. For this purpose, nozzles must be selected that will properly atomize
the liquid. The two general categories of nozzles use either hydraulic
pressure or compressed air to atomize the water. The various types of
hydraulic-pressure nozzles produce hollow-cone, solid-cone, or fan-shaped
sprays with various spray angles. Two-fluid nozzles use compressed air
instead of water pressure as the primary force for atomizing the water.
Use of two-fluid nozzles is especially attractive when an extremely small
droplet distribution is desired and when fluid viscosity is a problem.
Most nozzles produce a broad spectrum of droplet sizes rather than one
distinct size. It is often convenient, however, to express droplet size by
a single value such as the average diameter or the Sauter mean diameter (the
hypothetical droplet whose ratio of surface area to volume is equal to that
of the overall spray). In general, increasing water pressure (or air pres-
sure in two-fluid nozzles) will reduce the average or Sauter mean droplet
diameter. The droplet size population may be modified by additives such as
propan-1-ol but not necessarily by detergents.20 Hesketh21 has reported
that the addition of nonionic, low foaming surfactants reduced outlet dust
loadings 50 percent by improving particle wettability and/or atomization.
Because many scrubbers are operated with recirculating slurries, the
design of nozzles introducing the scrubbing liquid is of critical impor-
tance. With some scrubbers (packed towers', gas-atomized units) extensive
4.5-35
-------
distribution is needed at the point of liquid entry, but the scrubber ele-
ments provide liquid distribution. With other scrubbers the requirements
range from a hollow-cone spray to a full-cone spray. The most stringent
requirements are for high-pressure sprays that create the small water drop-
lets needed for high efficiency in scrubbers whose primary consumption of
energy is in the nozzles (preformed-spray scrubbers).
Nozzle plugging may be a major problem.22 The area formerly sprayed by
a plugged nozzle becomes subject to scaling and heat damage. The critical
factor in nozzle design is the minimum internal orifice. As the scrubber
size increases, the nozzle size normally increases. Under highly abrasive
conditions, ceramic nozzles may be needed. Special care must be taken when
installing and removing these nozzles to avoid breakage. Heat can also
cause breakage if a proper method of installation is not followed.23
4.5.3.3 Presaturatorsfor Hot Gases. The scrubbing of particulate
from hot gases presents special problems not associated with gases at am-
bient temperature. The heat in hot gases can evaporate substantial portions
of the scrubbing liquid droplets and adversely affect liquid/ particle
contact. In a hot gas stream, some droplets may evaporate before particulate
contact and others may evaporate after particulate contact, and thus cause
the particulate to be reentrained. Hot gases can also damage scrubber mate-
rials, especially fiberglass-reinforced plastics. Presaturation can be
economical when cooling of the gases permits the use of less expensive
materials of construction in the scrubber and when the lower volume of the
cooled gases allows the use of smaller scrubber vessels, fans, dampers, and
ducts.2*
High-temperature gases are usually cooled to near saturation by spray
quenching prior to entry into a scrubber. In most scrubber applications
approximately 1% to 2% times the theoretical evaporation demand is required
to quench the gases because of the kinetics of the cooling process.24 As in
the scrubbing process, nozzle design and arrangement are important in
quenching. As size of the quench water droplet decreases, the kinetics of
I
the cooling process increases and the evaporation demand becomes closer to
theoretical.
Quenching is frequently accomplished with scrubbing liquor rather than
clean water. In some applications, however, the use of scrubbing liquor for
4.5-36
-------
quenching can reduce scrubber performance. Most recirculating scrubbing
liquors contain very high levels of both suspended and dissolved solids. As
quench water evaporates, these solids can be reentrained into the gas stream
and must be collected again in the scrubber. Dissolved solids in evapo-
rating quench liquor can form fine particulate in the size range that may
escape collection in a scrubber.25 If fine particulate are regenerated in
this manner, the net effect is that contaminants are returned to the scrub-
ber inlet gas stream in a form more difficult to collect. It is usually
best, therefore, to use the cleanest water available for presaturator of a
scrubber. This can be accomplished by adding the makeup water directly to
the quench system rather than adding all makeup water to a common sump for
the quench and scrubber liquors. Using clean water alone for the quench process
is even better.
4.5.3.4 Staging of Scrubbers. When higher collection efficiencies are
required than can be obtained in a single scrubber, scrubbers are sometimes
staged to improve efficiency. When scrubbers are placed in series, end to
end, the penetration of particles of one size through the series is the
product of individual penetrations of particles of that size (see Equation
4.5-23). The corresponding pressure drop is the sum of individual pressure
drops per stage (see Equation 4.5-24). There is a diminishing overall
effectiveness in successive stagings, however, if the individual penetration
through a single stage by the finer particulate in a gas stream is consid-
erably higher than the penetration of the coarser particulate. Most of the
coarse particulate might be collected in the first stages while each succes-
sive stage collects only a small fraction of the remaining fine particulate
at a relatively high energy cost. Staging is most common with tray-type
scrubbers and packed bed scrubbers and is usually limited to three stages.
Staging can also be done with two types of scrubbers, such as a spray-tower
precleaner followed by a packed-bed scrubber.
P. = P. x P. x P. p. (Eq. 4.5-23)
\ nl ^2 3 ^
where
P. = total penetration of particles of aerodynamic diameter i
1t
P. = penetration of particles of aerodynamic diameter i from stage 1
h
4.5-37
-------
4Pt = ^ + &P2 + ••• + APn (Eq, 4.5-24)
4.5.3.5 Liquid Entrainment Separators. Wet scrubbing of particulate
is a two-step process, the second step being separation of the scrubbing
liquid droplets from the gas stream. This step is important in the ultimate
collection of particulate because poor liquid separation will cause reen-
trainment of the particulate.
There are four baste types of liquid entrainment separators26 or "de-
misters" (Figure 4.5-20). The mesh-pad and chevron types utilize inertia!
impaction of the liquid droplets to cause their agglomeration and removal.
The centrifugal and cyclonic types utilize centrifugal inertia to collect
the liquid droplets.
Plugging can be a persistent problem in mesh and chevron mist elimi-
nators in certain applications. Centrifugal-type mist eliminators are less
prone to plugging. Plugging can usually be minimized by continuous or
intermittent spraying of the mist eliminators.
4.5.3.6 Liquid Handling Facilities. Water usage and waste disposal
may become critical factors in the final selection of a wet scrubber. The
quantity of particulate collected, the size distribution of the particulate,
and the presence of dissolved contaminants in the scrubbing liquid have
great bearing on the amount of water and the type of liquid handling facil-
ities needed. Present water quality regulations require that most new
scrubber installations and many existing scrubber installations recirculate
scrubbing liquors to prevent the contamination of surface waters with ,the
collected air contaminants (Section 6). Recircufation, however, tends to
concentrate the dissolved scrubbing liquor contaminants.
Liquid handling facilities in recirculating scrubber systems usually
include a slurry pump; a. makeup water pump; a settling basin or pond; and
associated piping, valves, and spray nozzles. It is sometimes necessary
to construct multiple settling basins or to install clarifiers with drag
chains or rotary sludge collectors to settle and remove suspended solids
from scrubbing liquors. Additional procedures for liquid handling can
Include filtration; chemical treatment, for example to control pH level
and aid flocculation; and many other treatments common to industrial waste-
water treatment facilities (Section 6).
4.5-38
-------
GAS OUT
MESH TYPE MIST ELIMINATOR
(Courtesy of Koch Engineering
Company, Inc.)
GAS IN
TANGENTIAL INLET
CENTRIFUGAL MIST COLLECTOR6.
GAS OUT
CHEVRON MIST ELIMINATOR
GAS IN
CYCLONIC MIST
COLLECTOR0
Figure 4.5-20. Liquid entrainment separators (courtesy
of Industrial Gas Cleaning Institute, Inc.).
4.5-39
-------
Because scrubbing slurries are often corrosive and abrasive, all
liquor handling pumps, piping, nozzles, and valves must be constructed of
resistant materials or be lined with suitable protective materials. Since
slurries can also cause plugging, it is advantageous to install cleanout
traps and service hatches in many components. In some systems reliability
can be ensured only by installing duplicate pumps.
4.5.3.7 Materials of Construction. Materials of construction for
scrubber applications must be carefully selected to withstand corrosive and
abrasive agents in the gases or liquors, and to withstand any high tempera-
tures that may occur. If the process conditions are properly defined before
design begins, the experienced scrubber manufacturer can design a scrubber
that will withstand its service environment. Many scrubbers, however, fail
because inappropriate materials are selected after a superficial investiga-
tion of process conditions, or because insufficiently resistant materials
are substituted to reduce costs.27
Investigation of each scrubber application should include chemical
analysis of the raw materials, combustion products,, and scrubbing liquids.
The operating histories of any scrubber installations in similar applica-
tions should also be reviewed. Finally, review of the literature about
materials performance is recommended; and when materials performance data
are not available, in situ coupon tests may be required. After all relevant
information has been compiled, the designer prepares a list of materials
suitable for the expected service. Selection of materials of construction
from this list of candidates will be based in part on the relative costs.
Although the above mentioned procedures should be followed in selection
of materials, some general aspects of materials applications can be men-
tioned. Table 4.5-4 lists the major types of metals available for scrubbers
and ancilliary components, together with their major properties, general
corrosion behavior, and relative costs. Not listed are the nonmetallic
materials such as fiberglass-reinforced plastic, ceramics, protective coat-
ings, and wood, which are appropriate in many scrubber applications.
4.5.3.8 Instrumentation. Proper instrumentation is vital to the
monitoring of scrubber performance. Many installations require instrumenta-
tion with associated alarms and interlocks to protect valuable components
4.5-40
-------
TABLE 4.5-4. PROPERTIES OF METALS USED AS MATERIALS OF CONSTRUCTION
FOR WET SCRUBBERS AND AUXILIARY COMPONENTS28 30
Metal
Properties/uses
Corrosion resistance
Cast iron
High strength; low ductility;
brittleness; hardness; low
cost
Carbon steel
Martensitic
stainless
steel
(410, 416,
420, 440c)
Ferritic
stainless
steel
405
430
442,446
Good strength, ductility,
workability; low cost
and
Chromium alloy, hardenable by
heat treatment; typically used
for machine parts; costs 2 to
5 times more than carbon steel
Chromium alloy, not hardenable
by heat treatment; costs 2 to
4 times more than carbon steel
Modified for weldability
General-purpose, often used for
chimney liners
Used in high-temperature service
Ordinary cast irons exhibit
fair resistance to mildly
corrosive environments; high-
silicon cast irons exhibit
excellent resistance in a
variety of environments
(hydrofluoric acid is an
important exception); cast
irons are susceptible to
galvanic corrosion when
coupled to copper alloys or
stainless steels
Fair to poor in many environ-
ments; low pH and/or high
dissolved solids in moist or
immersion service leads to
corrosion; properly applied
protective coatings give
appropriate protection in
many applications; susceptible
to galvanic corrosion when
coupled to copper alloys or
stainless steels
Good
Good; better than martensitic
stainless steels; resists
stress corrosion; better
chloride resistance than
austenitic stainless steels
Good resistance to atmospheric
corrosion
(continued)
4.5-41
-------
TABLE 4.5-4 (continued)
Metal
Properties/uses
Corrosion resistance
Austenitic
stainless
steel
201, 202
301
302
304
304L
310
316
316L
Nickel alloy
Inconel
Monel3
Hastelloyfc
and
Chlorimet
Titanium
Chromium and nickel alloy; not
hardenable by heat; hardenable
by cold working; nonmagnetic
Types 201, 202, 301, 302, 303,
304, and 304L cost 3 to 5 times
more than carbon steel; types
310, 316, 316L, and 321 cost 4
to 10 times more than carbon
steel
Nitrogen added, used as a substi-
tute for 301 and 302
Good hardenability
General-purpose
General-purpose
Modified for weldability
Used in high-temperature service
Used in corrosive environments
Improved weldability
Good strength; costs over 10
times more than carbon steel
High strength; light weight (60%
that of steel); costs over 10
times more than carbon steel
Excellent; better than
martensitic or ferritic
stainless steel (except
for ha'lides)
Superior corrosion resistance;
good acid resistance; resistan'
to hot organic acids; good
pitting resistance
Excellent resistance in most
environments; not resisant
in strong oxidizing solutions
such a:; ammonium and HN03
Good resistance to stress
corrosion
Good resistance to hydrofluoric
acid
Excellent overall resistance
Exceptional resistance at
ambient temperatures
Excellent resistance at other
temperatures, except that
crevice corrosion is possible
in chloride solutions above
250°F
Registered trademark of Huntington Alloys, Inc.
Registered trademark of the Stalite Division of Cabot Corporation.
Registered trademark of the Duriron Company, Inc.
4.5-42
-------
from malfunctions such as loss of water pressure or a process temperature
runaway.
Every major scrubber system should include a meter to measure static
pressure drop across the scrubber and a meter to indicate water flow through
the scrubber. Static pressure drop can be measured with a differential
pressure gauge or manometer. Care must be taken in the design of the tubing
and fittings to prevent plugging and to allow easy cleaning, and tubing
materials should be selected to withstand the service expected. For
example, certain plastics can melt when exposed to high temperatures, some
plastics become excessively brittle at low temperatures, and polypropylene
tubing is degraded under continuous exposure to sunlight. Water flow rates
can be measured by in-line flow meters or doppler type indirect flow meters.
A less expensive and less accurate method of flow measurement is the use of
a pump pressure gauge calibrated to indicate flow rates. Open-channel type
flow-measuring devices such as the Parshall flume are sometimes useful,
although the preferred measuring point for liquor flow is between the pump
outlet and the scrubber spray nozzles.
For sources that generate hot gases, the gas temperatures must be
monitored if the scrubber contains materials that cannot withstand high
temperatures. A high-temperature alarm and/or an interlock system is
usually installed to shut down the process or to bypass the scrubber system.
Alarm systems can also be included to indicate low water levels in
orifice-type scrubbers. In systems that include presaturators, water flow
through the scrubber and the presaturator should be measured individually.
Where gas temperatures vary widely, it is sometimes necessary to install
temperature feedback instrumentation that controls the water flow rates to
the presaturator.
Scrubber instrumentation often includes liquor pH indicators, fan
ammeters, and fan vibration sensors. The pH meters are needed when pH of
the scrubbing liquor must be closely controlled. Maintaining clean, ac-
curately calibrated probes, although often difficult, is essential to the
success of pH control. Fan ammeters and tachometers can be used in conjunc-
tion with the manufacturer's fan performance curves to provide an estimate
of gas flow through the scrubber system, or these instruments can be used to
4.5-43
-------
provide a quick comparison of the system's performance with previous per-
formance. It is helpful in all scrubber systems to provide small ports in
the ducting before and after the fans, the scrubber vessels, and the pre-
saturators to allow pitot velocity traverses, gas temperature measurements,
and static pressure measurements.
4.5.4 Operation and Maintenance of Particulate. Scrubbers
4.5.4.1 Common Malfunctions. Wet scrubbers can provide continuous,
reliable service when they are operated properly and maintained regularly.
Poor operation and maintenance leads to component failure. Most scrubber
failures result from abrasion, corrosion, solids buildup, and wear of
rotating parts. Common failure modes for individual components are
discussed below.
4.5.4.1.1 Nozzle plugging. Nozzle plugging is one of the most common
malfunctions in scrubbers.22 Plugged nozzles reduce the liquid-to-gas ratio
or cause maldistribution of the liquid. Nozzle plugging results from improper
nozzle selection, excessive solids in scrubbing liquors, poor pump operation,
and poor sump design. Remedies for nozzle plugging include replacement with
nozzles of a different type, frequent cleaning, and reduction of liquor
solids content by increasing liquor blowdown and makeup water rates. Because
presaturator nozzles are especially prone to plugging, the quench water
should be limited to fresh water or very dilute liquors. Many quench nozzles
cannot tolerate greater than 2 percent solids in the liquid.23 Nozzle
plugging can be detected by observing the liquid spray pattern the nozzles
produce. If the nozzles are not accessible while the pumps are operating,
they should be checked during scrubber shutdowns for evidence of caking over
the nozzle openings. A decrease in water flow rate during scrubber operation
is an additional symptom of nozzle plugging.
4.5.4.1.2 Solids buildup. Solids buildup is another problem common to
wet scrubbers and one that is often difficult to control. The two types of
solids buildup are sedimentation and chemical scaling. Sedimentation occurs
when a layer of particles becomes attached to a surface or settles in areas
of low turbulence. Sedimentation can lead to plugging of pipes and ducts or
4.5-44
-------
buildup on internal parts. Chemical scaling results from a chemical reac-
tion of two or more species to form a precipitate on the surfaces of scrub-
ber components.
Solids buildup may occur in piping, sumps, scrubber packing, instrumen-
tation lines, or ductwork, and may lead to reduced scrubber efficiency and
major equipment failure. Most scrubbers using open pipes cannot reliably
tolerate liquor slurries of over 15 percent solids by weight. It is usually
best to maintain solids content at less than 6 to 8 percent.23 Techniques
to control scaling include increasing the liquid-to-gas ratio, controlling
pH, providing greater residence time in the holding tank, and adding other
chemical agents such as dispersants. Solids buildup can be detected by
inspection of accessible components and by inspection of the inner surfaces
of piping, tubing, and ductwork at removable fittings and hatches.
4.5.4.1.3 Corrosion. Corrosion problems arise frequently in wet
scrubbers, especially when the gases being cleaned contain acid-forming
compounds or soluble electrolytic compounds. The combustion of fossil
fuels, especially coal, coke, and residual fuel oil, yields oxides of
sulfur, which can produce sulfuric acid in scrubbing liquors. Metals-
refining processes, such as copper and lead smelting, can also produce
oxides of sulfur. Combustion of polyvinyl chloride plastics, commonly found
in incinerator feedstock, can produce hydrochloric acid in scrubbing liquors.
Rotary aggregate dryers and similar process equipment can produce chlorides
or fluorides, depending on the composition of the aggregate. The phosphate
fertilizer industry and the feldspar industry are especially troublesome
sources of fluorides. Acids and electrolytes in general are corrosive to
mild steels, chlorides are corrosive to many stainless steels, and fluorides
are harmful to nearly all stainless steels except certain specially formu-
lated (and expensive) high-nickel alloys.28 Recirculation of scrubbing
liquors greatly increases the concentrations of any corrosive agents they
contain.
Prevention of corrosion is best handled through proper choice of mate-
rials of construction and through pH control. When a pH control system is
to be the principal defense against corrosion, regular maintenance at frequent
intervals is necessary, especially at the pH electrodes. Another common
operating problem occurs when scrubber liquor blowdown rates are reduced to
4.5-45
-------
limit the emission of pollutants into surface waters. Reducing or
eliminating blowdown can so greatly increase the acid and electrolyte
concentrations in the liquor that otherwise acceptable materials of con-
struction become ineffective against corrosion.
4.5.4.1.4 Abrasion. Abrasion can occur where gases or scrubbing
liquors containing high concentrations of abrasiva particulate are in the
turbulent mode or are subjected to a sudden change in flow direction.
Typical wear areas in scrubbing systems include venturi throats, walls of
centrifugal mist collectors - near the inlet duct, and elbows in the
ductwork.23 Solutions to abrasion wear include the use of precleaning
devices and the use of large-radius turns in ductwork.
4.5.4.1.5 Wear of rotating equipment. Rotating equipment including
fans, pumps, and clarifiers must receive special attention in scrubber
service because of potential abrasion, plugging, and corrosion. Key wear
areas in these components include the bearings and any components rotating
in the fluid stream.31 :
Fan wear is a common problem. Forced-draft fans often suffer abrasion
because of exposure to particulate-laden gases. Wear problems in forced-
draft fans can be addressed by the use of special wear-resistant alloys, by
reduction of fan rotation speeds (by installing a larger fan), or by moving
of the fan to an induced-draft location on the clean air side of the scrubber
system. Induced-draft fans can undergo corrosion or solids buildup on the
blades if mist is carried over from the liquid entrainment separator.
Induced-draft fan problems can be addressed by use of corrosion-resistant
materials or by improving liquid entrainment separation.
Pump wear is also a common problem in scrubber systems. Pump housings,
impellers, and seals are subject to abrasion and corrosion by scrubber
slurries. Rubber linings and special-alloy pump materials are often used to
reduce abrasion and corrosion of the housings or impellers. Installation of
a water flush in the seals can help reduce wear of the seals.31
4.5.4.2 Preventive Maintenance. Preventive maintenance is an impor-
tant tool in assuring the continuous operation of scrubber systems. Preven-
tive maintenance programs for scrubbers should include periodic inspection
of equipment, replacement of worn parts, periodic cleaning of components
4.5-46
-------
prone to plugging, maintenance of an adequate spare parts inventory, and
recording of all maintenance performed on scrubber equipment.
All instrumentation such as differential pressure gauges, scrubbing
liquor flow meters, pump pressure gauges, and fan ammeters should be
observed at least once per work shift. All equipment should be inspected
regularly at regular Intervals, determined by the severity of service and
the likelihood of component failure. Failure-prone items include nozzles
and pumps handling slurries, forced-draft fans handling particulate-laden
gases, induced-draft fans downstream of inadequate liquid entrainment
separators, wear plates, pH probes, and bearings. These items should be
inspected as often as once per shift depending on the likelihood of failure.
Such components as ductwork and induced-draft fans handling clean, dry gases
should be inspected monthly.
All worn parts and malfunctioning equipment should be serviced as they
are discovered to prevent deterioration of system performance and to prevent
damage to equipment. An inventory of spare parts must be maintained in
stock for replacement of nozzles, bearings, pump seals, liners for pumps
with replaceable liners, pump impellers, wear plates for fans wheels with
wear plates, pH probes, and valve parts.30 Records should be made of all
maintenance performed and all parts replaced. This information is useful in
planning subsequent preventive maintenance schedules and in determining the
type and number of replacement parts needed.
4.5-47
-------
REFERENCES ,
1. Theodore, L. and A. J. Buonicore, Industrial Air Pollution Control
Equipment for Particulates. CRC Press, Cleveland. 1976.
2. Vincent, E. Wet Collection Devices. In: Air Pollution Engineering Manual
U.S. Environmental Protection Agency Publication No. AP-40, May 1973.
pp. 99-106.
3. Calvert, S., J. Goldshmid, D. Leith, and D. Mehta. Wet Scrubber System
Study, Volume I: Scrubber Handbook. U.S. Environmental Protection
Agency, Research Triangle Park, N.C. Publication No. EPA-R2-72-118a,
August 1972.
4. Sargent, G. D. Dust Collection Equipment, Chemical Engineering.
January 27, 1969. pp. 130-150.
5. Licht, W. Air Pollution Control Engineering. Marcel Dekker, Inc. New
York. 1980,
6. The Mcllvane Company. The Wet Scrubber Handbook. Northbrook,
Illinois. 1974.
i
7. Crawford, M. Air Pollution Control Theory. McGraw Hill Book Company.
New York. 1976.
8. Semrau, K. T., C. L. Witham, and W. W. Kerlin. Energy Utilization by
Wet Scrubbers. U.S. Environmental Protection Agency, Cincinnati, Ohio.
Publication No. EPA-600/2-77-234, November 1977.
9. Semrau, K. T. and C. L. Witham. Wet Scrubber Liquid Utilization.
U.S. Environmental Protection Agency, Cincinnati, Ohio. Publication
No. EPA-650/2-74-108, October 1974.
10. Strauss, W. Industrial Gas Cleaning, 2nd Edition. Pergamon Press,
Inc. New York, N.Y. 1975.
11. Hesketh, H. Fine Particle Collection Efficiency Related to Pressure
Drop, Scrubbant and Particle Properties, and Contact Mechanism. J. Air
Pollution Control Association. 24; 939-942. October 1974.
12. Ranade, M. B. and E. R. Kashdan. Design Guidelines for an Optimum
Scrubber System. Second Symposium on the Transfer and Utilization of
Particulate Control Technology. U.S. .Environmental Protection Agency,
Cincinnati, Ohio. Publication No. EPA-600/9-80-039a, September 1980,
pp. 538-560.
13. Engineering Science, Inc. Scrubber Emissions Correlation Final Report
to U.S. Environmental Protection Agency. Contract No. 68-01-4146, Task
Order 49. May 1979.
4.5-48
-------
14. Calvert, S., N. Jhaveri, and S. Yung. Fine Particle Scrubber
Performance Tests, U.S. Environmental Protection Agency,
EPA-650/2-74-093, October 1974.
15. Boll, R. H., L. R. Flais, P. W. Maurer, and W. L. Thompson. Mean
Drop Size in a Full Scale Venturi Scrubber via Transmissometer. J. of
Air Pollution Control Association. 24: 934-938. October 1974.
16. Calvert, S., N. C. Jhaveri, and S. Yung. Fine Particle Scrubber Per-
formance Tests. U.S. Environmental Protection Agency Publication
No. EPA-650/2-74-093, October 1974.
17. Calvert, S., H. F. Barbarika, and G. M. Monahan. Evaluation of Three
Industrial Particulate Scrubbers, U.S. Environmental Protection Agency,
Cincinnati, Ohio. Publication No. EPA-600/2-78-032, February 1978.
18. Yung, S., S. Calvert, and H. Barbarika. Venturi Scrubber Performance
Model, U.S. Environmental Protection Agency, Cincinnati, Ohio.
Publication No. EPA-600/2-77-172, August 1977.
19. Leith, D. and D. W. Cooper. Venturi Scrubber Optimization. Atmo-
spheric Environment. Volume 14, 1980. pp. 657-664.
20. Atkinson, 0. S. F. and W. Strauss. Droplet Size and Surface Tension
in Venturi Scrubbers. J. Air Pollution Control Association. 28: 1114-18.
November 1978. pp. 1114-1118.
21. Hesketh, H. Atomization and Cloud Behavior in Wet Scrubbers. Pre-
sented at the Symposium in Control of Fine Particulate Emissions from
Industrial Sources. San Francisco, California. January 15-18, 1974.
pp. 455-478.
22. National Asphalt Pavement Association. The Maintenance and Operation
of Exhaust Systems in the Hot Mix Batch Plant. Information Series 52,
Second Edition. February 1978.
23. Schifftner, K. C. Venturi Scrubber Operation and Maintenance. Pre-
sented at the U.S. EPA Environmental Research Information Center
Seminar on Operation and Maintenance of Air Pollution Equipment for
Particulate Control, Atlanta, Georgia. April 1979.
24. Industrial Gas Cleaning Institute, Inc. Scrubber System Major Auxil-
iaries. Publication WS-4. Stamford, Connecticut. 1975.
25. Kalika, P. W. How Water Recirculation and Steam Plumes Influence
Scrubber Design. Chemical Engineering. July 28, 1969. pp. 133-138.
26. Industrial Gas Cleaning Institute, Inc. Basic Types of Wet Scrubbers.
Publication No. WS-3. Stamford, Connecticut. June 1976.
4.5-49
-------
27. Mappes, T. and R. D. Terns. An Investigation of Corrosion in Particu-
late Control Equipment. U.S. Environmental Protection Agency, Research
Triangle Park, N.C. Publication No. 340/1-81-002. February 1981.
28. Nation Association of Corrosion Engineers. NACE Basic Corrosion
Course. Houston, Texas. May 1977.
29. Benzer, W. C. Steels. Chemical Engineering 77(22). October 12, 1970.
30. Fontana, M. G. and N. D. Greene. Corrosion Engineering. McGraw Hill.
New York, N.Y. 1967.
31. Czuchra, P. A. Operation and Maintenance of a Particulate Scrubber
System's Ancillary Components. Presented at the U.S. EPA Environmental
Research Information Center Seminar on Operation and Maintenance of Air
Pollution Equipment for Particulate Control. Atlanta, Georgia. April
1979.
4.5-50
-------
4.6 INCINERATORS
Incinerators are seldom, If ever, used solely to remove participate
matter because they tend to be more expensive and more energy intensive than
alternative control techniques. Applications are restricted to sources of
combustible matter with low gas flow rates and low particulate concentra-
tions. Principal among these are curing ovens, textile coating, charcoal
manufacturing, food processing, and certain chemical processes. The par-
ti cul ate- laden gas stream from these sources normally contains other pollu-
tants, such as volatile organic compounds (VOC), carbon monoxide, and odor-
ous compounds. Particulate control in the incinerator may in fact be only
ancillary to the control of malodors or VOC. Information concerning use of
incinerators for gaseous pollutant control is available in Reference 1.
An incinerator vaporizes and oxidizes particles. It is the only par-
ticulate control system that does not concentrate the particulate matter for
subsequent disposal.
4.6.1 Types of Incinerators
Three basic types of incinerators are used for particulate matter
removal: direct, thermal, and catalytic. Because the catalytic type is
prone to severe operating problems with particulate-laden gas streams, its
use is limited.
4.6.1.1 Direct and Thermal Combustion. These two basic designs are
similar and rely on simple combustion without the aid of a catalyst to
oxidize organics, essentially to water and carbon dioxide. The basic dif-
ference between the two types is that in direct combustion the gas stream
contains organic gases or vapors in sufficient concentration to sustain
combustion most or all of the time; in thermal combustion the gas streams
are lean, usually well below the lower explosive limit (LEL) for the par-
ticular organic gas or vapor. Thus thermal incinerators require appreciable
auxiliary fuel to achieve effective combustion, whereas direct afterburners
often require only a pilot flame to initiate combustion and to sustain
combustion during periods when the gas stream is lean. Many direct in-
cinerators are open flares. Where thermal incinerators are used, the gas
stream usually contains enough oxygen to burn the organic contaminants.
Gas streams vented to direct incinerators are often too rich in organics
4.6-1
-------
(concentrations are above the Upper Explosive Limit), and air must be intro-
duced to initiate combustion. A typical thermal incinerator is illustrated
in Figure 4.6-1.
4.6.1.2 Catalytic Combustion. This type of incinerator uses catalysts
to initiate and promote oxidation at temperatures well below those required
for thermal incinerators. The combustibles-laden gas stream is preheated
and passed through a catalyst bed to oxidize vapor phase organics, predomi-
nantly to carbon dioxide and water vapor.
The combustible contaminant concentration must be below the lower ex-
plosive limit. Commonly, catalysts are metals of the platinum family and
exist as a thin coating on an inert support material. Catalytic combustion
is not normally recommended for organic particulate removal because the
surface of the catalyst can become coated with particulate matter and there-
by inhibit the oxidation reaction. This type of incinerator is not dis-
cussed further.
4.6.2 Operating Principles of Incinerators
Combustion involves many complex, interrelated reaction mechanisms
between the fuel, fuel decomposition intermediates, and oxygen. Depending
on reaction conditions, results can include partially oxidized species such
as carbon monoxide, aldehydes, and organic acids, or simply carbon dioxide
and water vapor. The latter occurs only when the combustion processes
approach completion.
As with any combustion process, the basic variables for particulate
matter incineration are reaction temperature, reaction time, and reactant
mixing (turbulence). For solid particles, the reaction zone is confined to
the surface. At low temperatures the combustion rate is limited by the
chemical reaction rate, whereas at higher temperatures the chemical reaction
rate is so rapid that the rate of air supply to the surface controls the
combustion rate.2'3
Combustion of liquid droplets and volatile solids occurs away from the
surface of the particle, and combustion rate may be dependent on the rate of
heat transfer to the surface, which causes evaporation and thermal decompo-
sition of the solid. Combustion is influenced by the gas velocity, the rate
of mixing, and the supply of oxygen.4'5
4.6-2
-------
FLAME SENSOK-
STRAIGHTENING
VANES
REFRACTORY
INSOLATION
TURBULENT EXPANSION ZONE
•Till. MILL
UHITIZID IMHIHtnM
SAMPLE PORT
TEMPERATURE SENSOR
-•LOWIH
-INSULATION
Figure 4.6-1. Typical thermal incinerator.
4.6-3
-------
The temperature in the combustion zone surrounding the particulate
matter may exceed the temperature at the interior of the particle and in the
surrounding gas by several hundred degrees. Heat transfer is largely by
radiation from the incandescent surface of the particle, or from the incan-
descent carbon formed as an intermediate step in the combustion process.6
4.6.2.1 Reaction Temperature. The principal! requirement regarding
temperature is that the auto-ignition temperature of all species being
burned must be exceeded by approximately 100° to 2!00°C. This allows for a
margin of error to account for nonideal combustion conditions, heat losses,
and unknown particle composition. Operation at less than the auto-ignition
!
temperature means that combustion reactions are not initiated. Instead, the
particles are simply being heated, with possibly some volatilization.
Emissions at the stack may not exhibit any noticeable opacity; however,
downwind the vapors may recondense as secondary particulate matter. The
auto-ignition temperatures of selected 'organic compounds are presented in
Table 4.6-1.7
The reaction temperature also influences the rate of the combustion re-
actions. Most direct flame burners operate in the 650° to 820°C temperature
range to obtain maximum combustion within the limits of flame contact, mix-
ing, and residence time in the furnace.8
Figure 4.6-2 illustrates the effect of air velocity and particle dia-
meter on the combustion rate of carbon.9'10'11 The effects of particle
size, reaction temperature, combustion gas composition, and gas velocity on
the combustion rate of carbon, coal, and several other compounds have been
investigated.2"5'12'14
4.6.2.2 Reaction Time. Preheat (induction) and combustion times will
dictate the overall residence time of the particulate matter in the incin-
erator. The residence time requirement will determine both combustion
chamber dimensions and particle penetration.
The time required to heat the waste gas to peak furnace temperature is
|! ' • ,
dependent on the burner combustion intensity and inlet gas temperature.
Values of combustion intensity will vary from 400 kJ per cubic meter per
second for low-pressure gas jet mixers to 20,000 kJ per cubic meter per
second for premix mechanical burners. A typical value is 5600 kJ per cubic
meter per second for premix high-pressure gas jet multiple-port burners.
4.6-4
-------
TABLE 4.6-1. AUTO-IGNITION TEMPERATURES OF ORGANIC COMPOUNDS
Organic compounds
Auto-ignition
temperature, oK
Acetone
Ammonia
Benzene
Butadiene
Butyl alcohol
Carbon disulfide
Carbon monoxide
Chlorobenzene
Cresol
Cyclohexane
Dibutyl phthalate
Ethyl ether
Methyl ether
Ethane
Ethyl acetate
Ethyl alcohol
Ethyl benzene
Ethyl chloride
Ethylene dichloride
Ethylene glycol
Ethylene oxide
Furfural
Furfural alcohol
Glycerin
Hydrogen
Hydrogen cyanide
Hydrogen sulfide
Kerosene
Maleic anhydride
Methane
Methyl alcohol
Dichloromethane
Methyl ethyl ketone
Mineral spirits
Petroleum naphtha
Nitrobenzene
Oleic acid
Phenol
Phthalic anhydride
Propane
Propylene
Styrene
Sulfur
Toluene
Turpentine
Vi nyl acetate
Xylene
810
920
850
720
640
400
920
950
830
540
680
460
620
305
760
700
740
790
690
690
700
670
760
670
850
810
535
530
750
810
745
910
310
520
520
770
635
990
860
780
780
760
505
825
525
700
770
4.6-5
-------
-£ 1000
u
u
Jf
CN
O
U
»
CN
O
U
•z
O
100
<-> -r
< <
UJ ,5
01 Jl
< .T
U .
u
u. <
u. a:
UJ .
HI Oi
10
1
1000 1200 1400 1600 1800 2000
TEMPERATURE,'°K
2200
Figure 4.6-2. Effect of air velocity and particle diameter
on the combusiton rate of carbon.
(D = particle diameter).
4.6-6
-------
Combustion time required is dependent on particle size, oxygen content
of the furnace atmosphere, furnace temperature, particle composition, gas
velocity, and mixing of combustibles. Combustion times for several dif-
ferent materials have been determined and correlated on the basis of the
following equations:15
*d = CpR'Tmdpo) (96 * D pg} (Eq' 4'6"1}
tc = (pdpQ) (2Kspg) (Eq. 4.6-2)
where
t. = diffusion-controlled combustion time, sec
t = chemical reaction rate-controlled combustion time, sec
c
p = density of carbon residue or coke, g/cm3
R' = universal gas law constant, 82.06 atm cm3/mole/K
T = mean temperature of stagnant gas film, K
x = original diameter of particle, cm
= combustion mechanism = 1 for CO and 2 for C02 formation
D = diffusion coefficient of oxygen at temperature T , gm/cm2
p = partial pressure of oxygen in combustion air, atm
y
K = surface reation rate coefficient, g/cm2 sec atm.
K may be calculated by Equation 4.6-3 for soot and by Equation 4.6-4 for
o
coke and carbon residue.
Ks = (1.085 x 10*Ts ^)(e •"»•""" "'s) (Eq. 4.6-3)
K = 8710 e"35'700/RTs
Ks 871° e S (Eq. 4.6-4)
where
T = surface temperature.
Equation 4.6-1 holds at high temperature, zero gas velocity, and large
particle sizes. The equation can be corrected for the effects of gas velo-
city and turbulence by use of the dimensionless Nusselt conventional heat
transfer relationship for spherical particles:16
4.6-7
-------
NNu = 2 + 0.68 [Npr 1/3 x NRe %] (Eq. 4.6-5)
where
^Nu = Nusselt Number
Npr - Prandtl Number, a function of the physical properties of the
gas
NO = Particle Reynolds Number, a function of the physical proper-
ties of the gas, particle diameter, and gas velocity.
The Nusselt Number NN = h (d /k) = 2 at zero gas velocity, where h =
convectional heat transfer coefficient (cal/cm2 °C sec); d = particle dia-
meter (cm). The inverse function, h, of the stagnant gas film thickness,
d_/2, surrounding the particle and directly proportional to the thermal
conductivity of the furnace atmosphere, k (cal/cm2 °C cm sec).
The film thickness decreases with increasing velocity and decreasing
particle size to such an extent that the combustion rate for particles smal-
ler than 100 micrometers is limited only by chemical kinetics at normal in-
cinerator temperatures.
Equation 4.6-1 1s of limited value in design because particles larger
than 100 micrometers are easily collected by other gas cleaning devices and
would require excessive retention time and furnace volume.
Equation 4.6-2 holds for particle sizes smaller than 100 micrometers
and for temperatures at which the combustion rate is determined by chemical
kinetics.
Total combustion time for a carbon residue-forming particle then be-
comes:
tr = t. -H td • Kv + tc (Eq. 4.6-6)
where
t = total residence time, sec
t. = induction time, sec
K = volatile matter correction factor determined by the equation:
Ky = (1 + E/100)/(l + E/100 - V/100) (Eq. 4.6-7)
4.6-8
-------
where
E = percent excess air
V = percent volatile matter
The combustion time for hydrocarbon liquid droplets larger than 30
micrometers at zero gas velocity may be computed using the following equa-
tion:11
where
td = 29,800/Pg MwT"1'75 dpQ2 (Eq. 4.6-8)
M = molecular weight
W
T = furnace temperature, K.
The combustion time of particles smaller than 30 micrometers is depend-
ent on the combustion rate of the hydrocarbon vapors.
The time required to burn a 5 x 10 6 cm soot particle of 2 grams/cubic
centimeter density in a furnace atmosphere containing 0.20 atmosphere of
oxygen at 1260°C can be computed using the Equations 4.6-2 and 4.6-3. The
time required would be 0.51 second.
Total residence time in the furnace, including heat-up time from 100°C,
would be induction time + combustion time = total time, or:
tr = 0.208 + 0.510 = 0.718 second
In practice, minimum gas furnace retention time is about 0.30 second at
a temperature of 920°C. Particle retention time may be increased by design-
ing- the combustion chamber in the shape of a cyclone with a small tangential
inlet, and by introducing the gases at a high velocity.
4.6.2.3 Heat Transfer. The transfer of heat from burner flame to
gaseous and parti cul ate matter is an important factor in determining the
furnace size, operating temperatures, and fuel requirements of direct flame
contact incinerators. Heat transfer is best achieved by mixing when gases
are burned, and best achieved by radiant heat transfer when parti cul ate
matter is burned.16'17
For purposes of burning parti cul ate matter, radiant heat transfer and
furnace temperature uniformity may be increased by increasing the emissivity
of the burner flame. This can be accomplished by limiting the air supply to
4.6-9
-------
produce a sooty flame, by using high carbon-to-hydrogen ratio (C/H) fuels,
by adding soot, or by using fuel oil (by carburetion).
4.6.3 Design of Incinerator^
Factors that must be considered in incinerator design for particle-
containing gaseous waste include fuel requirements, burner selection, pro-
tection systems, heat recovery, refractory type, and instrumentation.
Design methodologies must take into account the complex interdependences of
operating parameters and the highly variable characteristics of many
sources. Semi-empirical approaches based on previous experience with analo-
gous sources are generally used.7
4.6.3.1 Fuel Requirements. Fuel requirements and burner capacity may
be determined by means of a heat balance, using the heat of combustion of
the fuel and the sensible heat needed to raise the temperature of the waste
gas and the products of combustion up to the desired combustion temperature.
The heating value of the contaminant must be deducted to determine net fuel
requirements.8'18 20
Insurance underwriters ususally limit the heating value of the waste
gas stream to combustible vapor concentrations of less than one-fourth of
the lower explosive limit. For organics this is equivalent to about 480 kJ
per standard cubic meter. The combustible particulate matter normally
contributes only a negligible heating value because of the low grain load-
ings.
Figure 4.6-3 presents the energy requirements for a thermal incinerator
at various influent gas stream coloric loads. Similar curves applicable to
a given facility could be generated by use of the procedures described in
Reference 2.
Energy requirements are inversely proportional to the influent gas
temperature—the higher the temperature, the less the sensible heat that
must be added to raise the influent gas stream to combustion temperature.
i
4.6.3.2 Heat Recovery. Energy requirements may be substantially
reduced by use of heat recovery equipment. The additional capital and
maintenance cost must be weighed against the energy savings.
4.6-10
-------
GAS FLOW, scfm
Figure 4,6-3. Fuel required to oxidike different concentrations
of combustible vapor (heat content of combustible
particulate assurance negligible).
4,6-11
-------
Heat recovery equipment used to recover heat from the flue gas may be
grouped into two classifications: recuperative and regenerative. Recupera-
tive heat exchangers, which recover heat on a continuous basis, include
crosscurrent-flow, countercurrent-flow, and cocurrent-flow heat exchangers.
For a given heat flow and temperature drop, heat exchanger surface require-
ments will be the least in the countercurrent-flow heat exchanger. The
crosscurrent-flow, U-shaped recuperator shown in Figure 4.6-4 is obviously
subject to fouling if combustion effectiveness decreases to the point that
large particles remain in the effluent. Cold-side deposits may also occur.
Cocurrent-flow heat exchangers are often used where a moderate level of
heat recovery is required. The cost of countercurrent-flow heat exchanger
construction may be greater than that for cocurrent-flow because operation
at lower temperatures (near the dewpoint) may require use of special alloys
or alloy steels.
Regenerative heat exchangers recover heat by intermittent heat exchange
through alternate heating and cooling of a solid. Heat flows alternately
into and out of the same exchanger as air and flue gas flows are periodi-
cally reversed. Regenerative heat exchangers are of fixed- and moving-bed
types. ]
A fixed-bed, pebble-stove, regenerative afterburner is shown in Figure
4,6-5. When gas is passed through the pair of pebble-type regenerators
connected back to back, the gas is heated on the upstream side and cooled on
the downstream side. When the upstream bed and gas temperature drop, gas
flow is reversed and the heat transfer process is repeated. Particulate
matter is effectively retained and incinerated. Heat recovery efficiencies
in excess of 95 percent can be achieved.21
A commonly used rotary regenerative heat exchanger consists of a parti-
tioned rotating cylinder containing a heat sink and heat transfer surfkce
',
area. The cylinder is partitioned along its axis by appropriate gas seals,
so that hot flue gas and cold waste gas may be passed through the heat
exchanger on opposite sides of the cylinder. Heat is absorbed from the hot
flue gas by the heat exchanger surface and transferred by the .continuous
rotation of the heat exchange surface to the cold waste gas side, where the
heat is absorbed by the incoming cold gases. Heat recovery efficiency
ranges from 85 to 95 percent.21
4.6-12
-------
NATURAL GAS
||P^*^*^i®
TO CHIMNEY
BLOWER
FROM KILN
Figure 4.6-5. Fixed-bed, pebble-stone, regenerative afterburner.
780 TO 840 K
WASTE GAS INLET
,330 TO 360 K
i 950 TO
980 "K
CLEAN
STACK
EXHAUST
n
AUXILIARY FUM£
BURNER INCINERATOR
INCINERATOR
WITH TUBULAR
RECUPERATOR
RECUPERATOR
Figure 4.6-4. Tubular recuperator.
4.6-13
-------
4.6,3.3 Hood and Ducts. Furnace inlet gases and vapors from paint and
varnish cooking kettles and other sources must be maintained at temperatures
above condensation to avoid exhaust duct fouling. Collection ductwork is
usually insulated and may be heated by means of an external duct that serves
to recover heat from the flue gas, which reduces fuel requirements.
Duct gas velocities are usually high, ranging from 1000 to 1700 m/min,
to prevent the settling of particulate matter, to effect a high heat re-
covery rate between the flue gas and furnace feed gas, and to minimize the
danger of flashback and fire hazards.7*22'23
Other safety devices for minimizing fire hazards may include diluting
vapors to below the lower explosive limits, using flame arresters, and in-
cluding a wet scrubber between the direct flame combustor and the vapor
source. Dilution of the vapors may be accomplished fay recirculation of a
portion of the flue gas, which would substantially reduce fuel requirements.
Flame arrestors may consist of a packed bed of pebbles, metal tower
packing, aluminum rings (Figure 4.6-6), or corrugated metal gridwork (Figure
4.6-7), in conjunction with a blast gate or other pressure-release device.
Flashback through the bed is prevented by bed gas velocities, by pressure
drop, and by cooling the flame to below combustion temperatures.24*25
Other types of flame arrestors include spray chambers, wet seals, and
dip legs. Wet flame arrestors have the disadvantage of cooling and humidi-
fying the exhaust gas with a consequent increase in fuel requirements. Wet
sprays are capable of partial removal of the solids.
Flame arrestors, regardless of type, should not be relied upon as the
primary defense against explosions.7 The combustibles content of the gas
stream should be continuously kept at less than 25 percent of the lower ex-
plosive limits, with consideration given to known fluctuations in process
operation.
4.6.3.4 Burners. The burner, the key operational component of an
incinerator system, is based on the type of service and capacity required.
Three major types of gas burners are available:7 raw gas burners,
premix gas burners, and forced-draft gas burners. The raw gas burner relies
on an induced draft of air to mix combustion air with the gas. This burner
consists of a cluster or ring of holes for the raw gas generation jets. The
premix burner is fed an already proportioned fuel/air mixture, which is
4.6-14
-------
vxxxxxxxxx,
vxxxxxxxxx
///ftttftf
PACKING
'/tt/ttfff
WASTE GAS
BLAST GATE
TO
COMBUSTOR
Figure 4.6-6. Packed-bed flame arrester.
TUBE BANK
SHELL
HANDLE
Figure 4.6-7. Corrugated metal flame arrester with cone
removed and tube bank pulled partly off the body.
4.6-15
-------
Ignited at the burner orifice or ring. Obviously, care must be taken to
avoid flashback. The forced-draft burner involves separate delivery of fuel
and combustion air to the burner. After mixing, the air and gas are ignited
by a pilot.
Oil burners are similar to the raw gas and forced-draft gas burners.
The principal difference is the addition of an oil atomizing system to
ensure intimate mixing of the fuel with air. A typical forced-draft oil
burner is shown in Figure 4.6-8. In this model, compressed air is used to
atomize the oil. The combustion air could be either contaminated effluent
or ambient air.
4.6.3.5 Construction Materials. Incinerator surfaces that will be
exposed to high temperatures and erosive or corrosive conditions must be
constructed of alloys capable of withstanding high temperatures or must be
lined with refractory materials.
4.6.3.6 Metals. Underwriters Laboratories, Inc., limits temperatures
at which alloy steels are used to approximately 100°G below the temperature
at which scale formation occurs. Martensitic and ferrjtic stainless steels
are recommended for use in areas that are exposed to wide ranges of tempera-
ture and to corrosive conditions.26 Temperature limitations for other
metals and alloys are determined by design stress and safety requirements.20
4.6.3.7 Refractories. Refractories used in direct-flame incinerators
"" V::: . •• „ ' .," .1" •..
increase radiant heat transfer, act as a support structure, and are resist-
ant to abrasion and corrosipn. They also must be capable of withstanding
thermal shock. Fire clay refractories are commonly used in incinerator
construction because of low cost, spall resistance, and long service life.
Fire clay refractory bricks are classified (Table 4.6-2) into maximum ser-
vice classes according to American Society for Testing and Materials (ASTM)
standards.27 Their softening points, as determined by pyrometric cone
equivalent (PCE), help determine their maximum service class.28 Other
requirements include limits on shrinkage, spelling loss, and deformation
under load. Commonly used castable fire clay refractories (Table 4.6-3) are
of two ASTM classes.29 ,
4.6-16
-------
COMBUSTION
AIR
NEBULIZER
NOZZLE
Figure 4.6-8. Typical forced draft oil burner.
TABLE 4.6-2. ASTM CLASSIFICATION OF FIRE CLAY
REFRACTORIES
Refractory type
Low heat duty
Intermediate heat duty
High heat duty
Super heat duty
PCE
19
29
21-32
33
Temperature,
1520
1640
1680-1700
1745
°c
TABLE 4.6-3. COMMONLY USED CASTABLE FIRE CLAY
REFRACTORIES
ASTM
No.
24
27
Temperature, °C
1310
1480
Density,
kg/m3
4.5-5.3
6.9-7.8
Special properties
Insulating, light-weight
General -purpose
Service temperature ranges of various refractories for corrosive condi-
tions are shown in Figure 4.6-9 and in Table 4.6-4. The literature contains
further information.27'28
4.6.3.8 Instrumentation. Minimum instrumentation for an incinerator
system consists of a temperature indicator that enables an operator to
determine if gas temperature is too low for effective oxidation of particu-
late matter or too high for the materials of construction.
4.6-17
-------
ov
I
5000
4500
4000
U- 3500
IU
5 3000
HI
-------
TABLE 4.6-4.
GENERAL PHYSICAL AND CHEMICAL CHARACTERISTICS
OF CLASSES OF REFRACTORY BRItK
Type of brick
Silica
High-duty fireclay
Super-duty fire-
clay
Acid-resistant
(type H),
Insulating brick
High-alumina
Extra-high alumina
Mullite
Chrome-fired
.
Magnesite-chrone
lx>nded.b
Magnesite-chrome
fired.
Magnesite-chrome
high-fired.
Magneslte-bondedb
Magncsite-fired
Zi rcon
Zirconia
(stablized).
•Silicon-carbide
Graphite
Typical
chemical
composition
SiO, 95%
SiO, 542
£.
Al20a 4035
Si Oj 52%
AljCh 422
Si 02 59%
Al20a 34S
Varies
AUO, 50-
85X
A1203 90-
99X
A1203 7 «
Chrome
ore 100X
MgO 50-
8015
CR2°3 5-
18X
F«20 3-
13*
Al,0, 6-
J in
SiO. 1.2-
5%
MgO 95%
ZrOg 57%
Si 02 33*
2r02 941
CaO 4%
SiC 80-
90%
C 97%
Approx.
bulk
density.
kg/m3
7.23
8.42
8.80
8.92
1.89-
4.71
10.69
11.62
9.62
9.62
12.25
11.44
11.31
11.31
11.31
11.18
12.57
15.40
10.06
6.60
Fusion
point,
oc
1950
1718
1743
1670
Varies
1760
1870
1650
2010
1810
Varies
Varies
2150
1950
2650
2300
3540
Chemical
nature
Acid
Acid
Ac-id
Acid
Slightly
acid
Neutral
Slightly
acid.
Neutral
Basic
Basic
Acid
Slightly
acid
Slightly
acid
Neutral
Deforma-
tion
under hot
loading
Excellent
Fair
Good
Poor
Poor
Good
Excellent
Excellent
Fair
Good
Excellent
Excellent
Good
Good
Excellent -
Excellent
Excellent
Excellent
Apparent
porosity
*
21
18
15
7
65-85
20
23
20
20
12
20
18
11
19
25
23
15
16 ''
Perme-
ability
High
Moderate
High
Low
High
Low
Low
Low-
Low
Very
low.
High
High
Low
Moderate
Very low
Low
Very low
Low
Hot
strength
Excellent
Fair
Fair
Poor
Poor
Good
Excellent
Good
Good
Good
Good
Excellent
Good
Good
Excellent
Excellent
Excellent
Excellent
Thermal
shock
resistance
Poor8
Fair
Good
Good
Excellent
Good
Good
Good
Poor
Excellent
Excellent
Excellent
Good
Good
Good
Excellent
Excellent
Excellent
Chemical resistance
to acid
Good
Good
Good
Insoluble in acids
except HF and
boiling phos-
phoric.
Poor
Good except for HF
and aqua regia.
Insoluble in most
acids.
Fair to good
Fair except to
strong acids.
Soluble in most
acids.
Very slight
Very slight
Slight reaction
with HF.
Insoluble
to alkali
Good at low temp-
eratures.
Good at low temp-
eratures.
Good at low temp-
eratures.
Very resistant in
moderate con-
centrations.
Poor
Very slight at-
tack with hot
solutions.
Slight reaction
Poor
Fair resistance
low tempera-
tures .
Good resistance
low tempera-
tures .
Very slight
Very slight
Attacked at
tempera-
tures.
Insoluble
aGood above 650°C.
Chemically bonded,
Dissociates above 1700°c,
-------
The addition of a temperature controller provides process control
capability not available with only a passive temperature indicator. The
controller is a feedback system that adjusts fuel input to follow changes in
gas temperature.7 In extreme situations, the fuel input is entirely shut off
to protect the incinerator. The controller thereby provides a degree of
protection and a means of continual adjustment to match influent character-
istics.
A flame-sensing device is useful for rapid shutdown of fuel supply in
case of flame failure. This device reduces the potential for accumulation
of explosive gases within the incinerator during the flame outage. It
should observe both the burner and the pilot flame.
4.6.4 Operation and Maintenance of Incinerators
Proper operation of an incinerator system requires control of contami-
nant quantity and characteristics and requires regular maintenance of the
burners. As with other particulate control devices, complete instrumentation
and an effective preventive maintenance program are necessary.
4.6.4.1 Burners. Burners are high-maintenance items because of the
high temperatures, refractory, and small orifices.7 When the contaminated
gas stream is used as the combustion air, fouling of the orifices and/or
deposits in the air delivery lines can occur. Impurities within the oil can
lead to similar problems. A second problem is improper sizing of the bur-
ner(s). This can lead to low gas temperatures resulting in incomplete
oxidation of particulate matter. Burners with poorly adjusted air-fuel
ratios can generate soot, which fouls downstream heat exchange surfaces.
Minimizing of burner problems is facilitated by providing a means of
visually checking the flame for proper luminosity, length, and stability.
Also, an adequate inventory of spare parts should be kept. If fouling
continues, a precleaner may be economical. Finally, such incinerator in-
strumentation as the flame sensor and the temperature controller should be
checked regularly.
4.6.4.2 Effluent Characteristics. Variability of effluent quantity
and heat content should be minimized by controlling the process operation.
Excess concentrations of combustible gases and vapors can lead to high
temperature excursions, which damage the incinerator refractory or shell.
4.6-20
-------
High gas flow rates lead to poor particle oxidation resulting from decreased
residence time and decreased reaction temperature. Undesirable reaction
products such as carbon monoxide, aldehydes, and organic acids also result
from poor combustion.
Contaminants containing sulfur or chlorine compounds may be oxidized to
highly corrosive species such as hydrochloric acid vapors and sulfuric acid
vapors. These could attack either the refractory or the shell of the in-
cinerator.27 Precleaning or special materials of construction are required
when these contaminants are present.
4.6-21
-------
REFERENCES
1. U.S. Environmental Protection Agency. Control Techniques for Volatile
Organic Emissions from Stationary Sources. Evaluation draft. February
1978.
2. Kobayashi, K. Combustion of a Fuel Droplet. In: Fifth Symposium on
Combustion, Pittsburgh. Reinhold Publishing Co., New York, Pub. No.
1955, 1954.
3. Spaulding, D. B. Heat and Mass Transfer in the Combustion of Liquid
Fuels. American Society of Mechanical Engineers. Published by the
Institution of Mechanical Engineers, Story's Gale, London, England,
September 11-13, 1951.
4. Hottel, H. C,, G. C. Willians, and H. C. Simpson. Combustion of Drop-
lets of Heavy Liquids Fuels. In: Fifth Symposium on Combustion,
Pittsburgh. Reinhold Publishing Co., New York, Pub. No. 1955, 1954.
5. Nlshiwaki, N. Kinetics of Liquid Combustion Processes. Evaporation
and Ignition Lag of Fuel Droplets. In: Fifth Symposium on Combustion,
Pittsburgh. Reinhold Publishing Co., New York, Pub. No. 1955, 1954.
6, Orning, A. A. Combustion of Pulverized Fuel—Mechanism and Rate of
Combustion of Low Density Fractions of Certain Bituminous Coals.
Trans. Am. Soc. Mech. Eng., Vol. 64, 1942.
7, Ross, R. Incinerators, An Operation and Maintenance of Air Pollu-
Control Equipment. Technomic Publishing, 1977.
8. Danielson, J. A. (ed.) Air Pollution Engineering Manual. U.S. Public
Health Service, National Center for Air Pollution Control, Cincinnati,
Ohio. PHS Pub-999-AP-40, 1967.
9. Yagi, S,, and D. Kunii. Combustion of Carbon Particles in Flames and
Fluidized Beds. In: Fifth Symposium on Combustion, Pittsburgh. Rein-
hold Publishing Company, New York, Pub. No. 1954.
10. Browning, J. A., T. L. Tyler, and W. G. Kran Effect of Particle Size
on Combustion of Uniform Suspensions. Ind. Eng. Chem., 49(1):142-147,
January 1957.
11. Godsave, G. A. E. Studies of the Combustion of Drops in a Fuel
Spray—The Burning of Single Drops of Fuel. In: Proceedings of the
4th Symposium on Combustion, Cambridge, Massachusetts. Williams and
Wilkins Co., Baltimore, Maryland. Pub. No. 1953, September 1952.
4.6-22
-------
12. Gregory, C. A., Jr., and H. F. Calcote. Combustion Studies of Droplet-
Vapor Systems. In: Proceedings of the 4th Symposium on Combustion,
Cambridge, Massachusetts. Williams and Wilkins Company, Baltimore,
Maryland. Pub. No. 1953, September 1952.
13. Smith, D. F., and A. Gudmundsen. Mechanism of Combustion of Individual
Particles of Solid Fuels. Ind. Eng. Chem., 23(3):277-285, March 1931.
14. Parker, A. S. , and H. C. Hottel. Combustion Rate of Carbon. Ind. Eng.
Chem., 28(11):1334-1341, November 1936.
15. Field, M. A., D. W. Gill, B. B. Morgan, and P. G. W. Hawksley. Combus-
tion of Pulverized Coal. British Coal Utilization Research Associa-
tion, Leatherhead, England, 1967.
16. Eckert, E. R. G., and R. M. Drake, Jr. Heat and Mass Transfer. 2nd
ed. McGraw-Hill, New York, 1959.
17. Topper, L. Radiant Heat Transfer From Flames in a Turbojet Combustor.
Ind. Eng. Chem., 46(12):2551-2558, December 1954.
18. Goodel, P. H. Industrial Ovens Designed for Air Pollution Control. J.
Air Pollution Control Assoc., Vol. 10, pp. 234-238, 1960. (Presented
at the 52d Annual Meeting of the Air Pollution Control Association, Los
Angeles, California, July 22-26, 1959.)
19. Vandaveer, F. E. , and C. G. Sedeler. Combustion of Gas. American Gas
Assoc., Inc., New York, 1965.
20. Ingels, R. M. The Afterburner Route to Pollution Control. Air Eng.,
No. 6, pp. 39-42, June 1964.
21. Perry, R. H. , C. H. Chilton, and S. D. Kirkpatrick Chemical Engineers'
Handbook. 4th edition. McGraw-Hill, 1963.
22. Sly Manufacturing Company. Industrial Air Pollution Control. Cleve-
land, Ohio. Bulletin 204, 1967.
23. Weil, S. A. Burning Velocities of Hydrocarbon Flames. Inst. Gas
Tech., Chicago, Illinois. Research Bulletin 30, 1961.
24. National Board of Fire Underwriters. Standards for Ovens and Furnaces.
NBFU No. 864, August 1963.
25. Radier, H. H. Flame Arresters. J. Inst. Petr., Vol. 25, pp. 377-381,
1939.
26. Underwriters Laboratories. Standards for Safety—Commerical—Indus-
trial Gas Heating Equipment--UL. 795. November 1952, Revised 1960.
27. Trinks, W. , and M. H. Mawhinney. Industrial Furnaces, 5th ed., Vol. I.
John Wiley and Sons, 1961.
4.6-23
-------
28. Burst, J. F., and J. A. Spieckerman A Guide to Selecting Modern
Refractories. Chera. Eng., 74(16):85-104, July 1967.
29. Griffiths, J. C., and E. V. Weber. Influence of Port Design and Gas
Composition on Flame Characteristics of Atmospheric Burners, American
Gas Association Laboratories, Cleveland, Ohio. Research Bulletin 77,
1958.
4.6-24
-------
SECTION 5
FUGITIVE EMISSION CONTROL
The failure of many areas to attain the National Ambient Air Quality
Standards for particulate matter has prompted reexamination of the partic-
ulate problem. Conventional stationary emission sources have been con-
trolled, but in many cases, fugitive emissions constitute a large percentage
of total emissions;1 thus attainment of the standards for particulate matter
necessitates the control of these fugitive emissions.
5.1 SOURCES OF FUGITIVE EMISSIONS
Particulate that becomes airborne is either industrial process fugitive
emissions or fugitive dust.
Industrial process fugitive emissions include particulates that are
emitted from industry-related operations and that escape to the atmosphere
through windows, doors, vents, and the like rather than through a primary
exhaust system such as a stack, a flue, or a control system. Fugitive
emissions may escape to the atmosphere from indoor manufacturing operations,
materials handling, transfer, and storage operations, and other industrial
processes. Sometimes they are emitted more directly to the atmosphere from
out-of-doors industrial processes such as coke ovens, quarry rock crushing,
and sandblasting. Fugitive emissions can result from poor maintenance of
process equipment or from the operation of processes without concern for the
environment; for example, they can result from leakage around coke oven
doors that cannot be properly sealed because of warpage.
Fugitive dust is generally either natural or man-associated particulate
that becomes airborne by the wind, man's activity, or both. Examples are
windblown particulate from unpaved dirt roads, tilled farmlands, and exposed
surface areas at construction sites and from duststorms.
5-1
-------
Table 5-1 lists sources of industrial process fugitive emissions and
the broad range of controls applicable to these emissions.
Table 5-2 lists sources of fugitive dust ("native soil that becomes
airborne"). Although some sources are in the area of processes (for exam-
ple, surface mines), they can be differentiated; overburden removal and haul
roads at surface mines are fugitive dust sources, whereas removal of a
product (e.g., coal) is a fugitive emission source.
5.2 CONTROL OF INDUSTRIAL PROCESS FUGITIVE EMISSIONS
As shown in Table 5-1, the following techniques are used to control
process fugitive emissions: ventilation systems, process optimization, and
wet suppression. Although some processes have been controlled by all three,
others are controllable by only one or two of these techniques.
Selection of the proper control technique requires consideration of
factors such as the industrial processing facility, the characteristics of
the exhaust stream, and the secondary multimedia impacts. Selection must be
site-specific.
Ease of control varies with the age and basic design of a facility. A
fugitive emission control system for a new plant can be integrally incor-
porated into the overall design of the plant, whereas a retrofit application
requires that the system be adapted to the configurations of the existing
plant. The retrofit must be built within fixed-space limitations without
interfering with operation of the process. In general, the more congested
the plant layout, the harder it is to retrofit most fugitive emission con-
trol systems.
The location of the facility and the fugitive emission source within
the facility can also affect the selection of the control technique. For
example, controls required for a storage pile of fine material near a public
road could differ from those for a storage pile well within the plant bound-
aries, i
Product quality also affects control technique selection. For example,
dry collection of fugitive emission may be needed so that it can be returned
to the product stream. Also, wet suppression may have limited use because
it has an adverse impact on product quality.
5-2
-------
TABLE 5-1. INDUSTRIAL PROCESS FUGITIVE EMISSION SOURCES
AND APPLICABLE CONTROL TECHNIQUES2
Source
Material handling
Transf erri ng/convey i ng
Loach" ng/unl oadi ng
Storage
Baggi ng/packagi ng
Material crushing and screening
Mineral mining
Drilling
Blasting
Extraction
Waste disposal (tailings)
Metallurgical operations
Furnace charging
Furnace tapping (product and
slag
Casting
Mold preparation
Casting shakeout
Slag disposal
Sintering
Coke oven charging
Coking (leaks)
Coke pushing
Coke quenching
Ventilation
and collection
X
X
X
X
X
X
X
X
X
X
X
X
X
Process
optimization
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Wet
suppres-
sion
X
X
X
X
X
X
X
X
X
5-3
-------
TABLE 5-2. FUGITIVE DUST EMISSION SOURCES AND APPLICABLE
CONTROL TECHNIQUES2'3
Source
Unpaved roads
Construction
activities
Dust from
paved roads
Off-road motor
vehicles
Overburden
removal /storage
Reclamation
efforts
Inactive
tailings piles
Disturbed soil
surfaces
Agricultural
ti 11 i ng
Wet sup-
pression
x .
X
X
Stabili-
zation
X
X
X
X
Speed
reduction
X
X
Surface
cleaning/
transpor-
tation
controls
X
X
Wind-
breaks
X
X
X
Good
oper-
ati ng
prac-
tices
X
X
X
X
5-4
-------
The major exhaust stream characteristics that collectively influence
selection of the control technique include particle size distribution;
temperature, moisture content, presence of corrosive gases; and physical and
chemical characteristics and associated toxicity of the particulate.
The size of particulates from many metallurgical fugitive emission
sources is predominantly below 5 jjmA, which in the case of add-on control
systems often dictates the need for a fabric filter.4 Exceptions are
sources in which the particulates have a relatively large diameter, which
sometimes can be sufficiently controlled by high efficiency cyclones.
Since most process fugitive emission exhaust streams are at either
in-plant or ambient temperatures, provisions such as heat-resistant fabric
filter material for excess temperatures are generally not required. In some
applications, however, hoods near furnaces must be water-cooled to withstand
initial high temperatures. Because most fugitive emissions have approxi-
mately the same moisture content as the ambient or in-plant air, little
insulation or reheat is generally needed for protection against condense-%
tion.
Physical and chemical characteristics of fugitive emissions also
influence selection of the control technique. The most critical physical
characteristics that relate to the type of control and to the material of
construction are abrasiveness (related to particle size and morphology),
hygroscopy, and true density; the most critical chemical characteristic is
corrosiveness.
Control of fugitive emissions can create secondary multimedia environ-
mental effects, which also must be controlled. These secondary effects
include solid waste disposal, water pollution, generation of additional
fugitive emissions, and noise. If not controlled, problems associated with
poor procedures for disposal of the fugitive emissions collected in a fabric
filter (when return to the processing system is not feasible) may exceed the
original problem. An example would be dumping collected materials into an
open truck, hauling them in an open truck to a landfill, and dumping them
into a landfill that is not adequately protected from wind erosion and
surface runoff. Thus adequate precautions must be taken in the selection of
a control technique, to avoid creating new fugitive sources.
5-5
-------
5.2.1 Ventilation. Systems
5.2.1.1 Localized Hoodj ngand Enclosure. Capture and ventilation
systems are used with appropriate particulate removal devices (e.g., a
fabric filter or wet scrubber) to control fugitive emissions. Sources
amenable to this type of control include processes such as materials hand-
ling (conveyors, elevators, feeders, loading and unloading, product bagging,
and storage silos); solids beneficiation (crushing, screening, and other
classifying operations); mining (drilling and crushing), and others (furn-
aces, dryers).2'4-8
In general, systems of capture near the process (localized hoods or
canopy hoods), as opposed to ventilation of an entire building, are desir-
able from an economic and occupational exposure standpoint. For example,
the use of canopy hoods to control fugitive emissions from electric arc
furnaces can be expected to require 40 to 50 percent of the flow rate re-
quired for building evacuation.4 Local hoods (nearer the emission source
than canopy hoods) may require even less air flow for capture.
The capture effectiveness of a ventilation system varies greatly and
depends on many parameters. The properties of the emission source, location
of the hood or enclosure, possibilities of external disturbances such as
wind and vibrations, and operator errors can all affect the capture effi-
ciency. In general, however, capture efficiencies of more than 90 percent
can be expected with proper design. For example, 90 percent efficiency is
attainable on the enclosure of basic oxygen furnaces.8
Detailed guidelines for the proper design of ventilation systems can be
obtained from several sources.9-11 The important factors in the proper
design of hooding, enclosures, ducting, and ventilation are:
0 Hood design and placement ,
0 Air velocity and flow direction required for proper capture of
particles
0 Volume of air required
0 Minimization of static pressure drop.
5-6
-------
Hoods should be designed to enclose the emission source as much as pos-
sible. Total enclosure, however, may be limited by the need for easy access
to the process. Hoods must have adequate flow rates and face velocities
(air into the hood) to capture the particles and to impose mininup pressure
drop on the system. Sometimes the hood can be located and shaped so that it
is aligned with the general flow direction of the contaminant.
Air flow past the emission source must be sufficient to capture the
contaminant, and must be directed to minimize exposure of workers to hazard-
ous fumes. Proper air flow rates depend on the specific properties of the
emission stream and the particulate matter. Duct systems must complement
the hoods in efficiency of operation. Fittings and other components should
be adequate to prevent excess pressure drop. When overall hood and duct
dimensions are selected, the expected static pressure drop in the system can
be calculated readily by using standard techniques.11
5.2.1.2 Building Evacuation. General ventilation of an entire build-
ing is receiving increased attention as a technique of fugitive emission
control because of the space entailed in capturing emissions with many local
hoods.2 Ductwork is installed on the building roof, and large fans draw the
particulate-laden gases from the building through a collection device. This
may require improvement of roof supports because of the added weight, and
the installation of opening/closing doors to minimize leaks. A fabric
filter is typically the control device selected for building evacuation.
Overall capture and collection efficiency for such a device is estimated to
be from 90 to more than 95 percent.5
Total building evacuation has its drawbacks, however. One is the large
air flow rates required. Another is that the evacuation systems sometimes
collect enough gas to sufficiently ventilate the workplace overall, but
still leave an unacceptable level of local pollutant concentrations because
of "deadspots" in the air flow pattern.5
Building evacuation systems have been applied to electric arc furnace
melt shops in the iron and steel industry and to a converter building in the
copper industry.5 These systems range in size from 235 to 285 m3/s and
larger.5
5-7
-------
BELT
ENCLOSING
7—-HOOD
HOPPER
\ HOPPER
"t
GOOD
BAD
Figure 5-1. Hood design.'
Figure 5-2, Hood location.
9
w
SLOT-
lit
PLATING
TANK
PLATING
TANK
GOOD BAD
Figure 5-3. Air flow direction.
5-8
-------
9 m MAX.
ro
f
60 era MIN.
/ /
/
7
V
RUBBER
SKIRT
COVER CONVEYOR BETWEEN TRANSFER POINTS.
EXHAUST AT TRANSFER POINTS AS REQUIRED,
EXHAUST MINIMUM ADDITIONAL 0.5 nT/s per in. OF
BELT WIDTH AT A MAXIMUM OF 9-m INTERVALS.
ENTRY LOSS = ENTRY LOSS FACTOR FOR TAPERED
HOOD X DUCT VELOCITY PRESSURE.
DUCT VELOCITY =1100 m/min MINIMUM.
Figure 5-4. Belt conveyor ventilation for fugitive emissions control.
10
5-9
-------
CONVEYOR
ENCLOSED
LOADING POINT
45? OR MORE
CLOSED CONVEYOR (
.OCATE REMOTE FROM
LOADING POINT
X
•CONVEYOR TO
HOPPER AND BIN
CHUTE
TO BIN
LOCATE REMOTE FROM
LOADING POINT
Figure 5-5. Hopper and bin chute and conveyor-loading ventilation for
fugitive emissions control.'"
5-10
-------
HOOD ATTACHED TO BIN
PRINCIPAL DUST SOURCE
SCALE SUPPORT
Figure 5-6. Bag-filling fugitive emissions control.
10
5-11
-------
5.2.2 Optimization of Equipment and Operation
Control by proper operation and maintenance practices primarily in-
volves the elimination of fugitive emissions resulting from process upsets,
leaks, and poor housekeeping. In addition, prompt cleanup of spills on the
ground or floor by vacuum systems will prevent spills from becoming air-
borne. A full-time cleanup crew may be required in some industries. Also
included in this category is the optimization of the capture efficiency of
the hooding systems of primary source control devices. Examples include:
Precautions to ensure that a cupola is not overloaded, to eliminate the
possibility of backpressure from the primary control system and "puf-
fing" fugitive emissions from the charging door opening
Maintenance of coke oven doors and seals to eliminate leaks during
coking
Prompt repair of electric arc furnace hooding damaged by overhead
charging crane
Conscientious periodic application of chemical suppressant to inactive
storage piles and tailings areas
Increase in the vent rate of a canopy hood system for an electric arc
furnace in a gray iron foundry
Prompt cleanup of spills from trippers of a clinker conveying system in
a portland cement plant
Increase in the vent rate of a primary control device to eliminate or
minimize leaks.
A change in the process or raw materials can be an effective technique
of reducing fugitive emissions. For example, using only clean scrap in
metal-melting furnaces or removing crankcase oil prior to automobile salvage
can reduce fugitive emissions. Changing the process (from cupola to elec-
tric arc furnace or from bucket elevator to pneumatic conveyor) is an effec-
tive way to minimize fugitive emissions at the source.
During the transfer of dusty materials from a conveyor or stacker to
another conveyor or stockpile, fine materials can be separated from large
materials by wind and/or the falling action of the materials. A simple
technique for reducing dusting is to shorten the fall distance by using
hinged-boom conveyors, rock ladders, telescoping chutes, lowering wells, or
5-12
-------
other devices.12 The hinged-boom conveyor, which can raise or lower the
.conveyor belt and thus reduce the fall distance at the transfer point, can
reduce emissions by an estimated 25 percent.4 Rock ladders allow the
material to fall short distances in a steplike fashion; the direction of
travel on successive steps is reversed to reduce the momentum that the
material receives from the previous fall and to reduce the resulting dust-
ing. This technique can reduce emissions by 50 percent.4 Telescoping
chutes carry the materials from the discharge point to the receiving point,
but the materials are not exposed. Estimated control efficiencies of 75
percent are possible. Lowering wells, or perforated pipes, allow materials
to flow out of the pipe above the pile surface; The dusting from the impact
of the falling materials is retained inside the pipe, and the material is
protected from wind action. This technique can reduce emissions by an
estimated 80 percent.4
Confinement by covering or enclosure basically involves the partial or
complete seclusion and/or shielding of the source of the fugitive dust or
the industrial, process fugitive emissions. The design strategy is to pre-
vent the fugitive particulate matter from becoming airborne by the wind or
by the mechanics of the process. These control techniques range from small
enclosures over conveyor transfer points (for protection from the wind and
from turbulence of the moving belt) to building structures (for complete
confinement of materials storage areas). Other examples include conveyor
system enclosures, weighted tarpaulin covers for inactive material storage
piles, partial windbreaks in the prevailing upwind direction from limestone
quarry surge pile areas, and partial open-ended shelters with shrouds for
railroad car loading and unloading.2 The total enclosure of a transfer
point can reduce fugitive emissions by an estimated 70 percent.4
5.2.3 Wet Suppression2'4'5
Fugitive emissions from materials handling and beneficiation can be
controlled by spraying liquid on the materials. Wet suppression techniques
include applications of water, chemicals, and foam. The point of applica-
tion is most commonly at the conveyor feed and discharge points, but some
are at conveyor transfer points and equipment intakes. Wet suppression with
water only is a relatively inexpensive technique; however, it has the inher-
ent disadvantage of being short-lived. Control with chemicals (added to
5-13
-------
water for improved wetting) or foam is longer lasting, but more expensive
than water alone. A wet suppression system is shown in Figure 5-1.
A wetting agent breaks down the surface tension of water, and allows it
to spread further, penetrate deeper, and to wet the small particles better
than untreated water. Mechanical agitation of the materials causes the
small particles to agglomerate. For effective control, the spray should be
applied at each point where the particles might be fractured, allowed to
free fall, or subjected to strong air currents.14
When applied to loading or unloading operations, wet suppression tech-
niques can reduce airborne dust to some extent. The loading process natur-
ally disturbs the materials, but water sprays with wetting agents cause
small dust particles to adhere to larger pieces so as not to become en-
trained. This technique is not suitable for many materials that cannot be
readily wetted.2
Foam is effective in dust suppression because small particles (1 to 50
pmA in diameter) break the surface of the bubbles in the foam when they come
in contact with it, and these particles become wetted. (Larger particles
only move the bubbles away.) The small wetted particles then must be
brought together or brought in contact with larger particles to achieve
agglomeration. If foam is injected into free-falling aggregate at a trans-
fer point, the mechanical motion causes the contact of particle with bubble
and subsequent contact of particle with particle.
Electrostatically charged fog is another type of wet suppression sys-
tem. It differs from conventional water sprays, in -that the droplets carry
a charge of static electricity. Because most fine particulates carry a
natural electrical charge, particle collection can be improved via elec-
trostatic attraction if the water spray droplets are charged to the opposite
polarity. The charged water droplets exert attractive forces on the op-
positely charged particles, and each droplet collects more particles as it
travels through the dust-laden gas.5'15
Spray systems on material-hand!ing operations are estimated to reduce
emissions by 70 to 95 percent.4 Spray systems achieve an estimated 80
percent reduction at rail car unloading stations at iron and steel plants.4
Spray systems can also reduce load-in and wind erosion emissions from
storage piles of various materials. Emissions from load-in can be reduced
5-14
-------
TO KILN
en
SECONDARY
CRUSHER/
SCREENS
PRIMARY
CRUSHER
INCOMING
WATER LINE
I
n
DUST CONTROL
AGENT
PROPORTIONER
Figure 5-7. Wet dust suppression system applied to material handling operation
Note: Wet suppression at fine mesh screens must be regulated properly to avoid
blinding of screens.
13
-------
by an estimated 70 to 90 percent,2'4*16 and wind erosion can be reduced by
an estimated 80 to 90 percent.2'4
5.3 CONTROL OF FUGITIVE DUST
Table 5-2 lists the major fugitive dust sources and the applicable
control techniques. These techniques are wet suppression; stabilization
(physical, chemical, and vegetative); and specialized methods such as speed
reduction, street cleaning, windbreaks, industrial transportation control,
and good operating practices.
5.3.1 Wet Suppression2 >4
Wet suppression of dust with water or with water plus a wetting agent
(surfactant) is an effective temporary control for fugitive dust from un-
paved roads, cattle feedlots, stockpiles, waste heaps, mining, and con-
struction activities. Water alone is not as effective because the high
surface tension of water prevents it from wetting, spreading, and pene-
trating. The addition of a wetting agent aids water penetration into the
material and helps to promote particle agglomeration.17 Wet suppression of
fugitive dust on exposed surfaces such as haul roads is usually accomplished
by spraying from tank trucks. Fixed pipeline spray systems have also been
used on main haul roads that are relatively permanent, such as those at
mines and large industrial facilities.
When water is used for wet suppression, repeated applications are
necessary. Evaporation and runoff cause its effectiveness to be temporary.
In some regions, water usage is limited as a suppressing agent because of
its scarcity.
The control efficiency of wet suppression techniques depends on local
climatic conditions, the properties of the fugitive source, and the length
of effectiveness provided by the control. Estimated control efficiencies of
80 percent have been reported for cattle feedlots.16 Extensive watering of
the soil may reduce emissions from construction operations as much as 70
percent.16 Wetting of access roads twice a day with 2.3 liters of water per
square meter will suppress fugitive dust from normal construction practices
an estimated 30 to 50 percent.
5-16
-------
5.3.2 Stabilization
Stabilization techniques isolate fugitive dust sources from external
disturbances such as wind and traffic. This can be done physically by
adding a layer of material on exposed surfaces; chemically by using mate-
rials that help to bind dust and particulates; and vegetatively by planting
of trees, shrubs, and grass over the surface.
5.3.2.1 Physical Stabilization. For inactive waste stockpiles and
steep slopes, stabilization can be accomplished by mixing a layer of rock,
soil, crushed or granulated slag, bark, wood chips, and straw with the top
layer.18 For dirt roads, paving is the most effective control. The most
widely used low-cost pavement is a bituminous asphaltic chip seal over a
granular base or a stabilized soil base. Maintenance requirements depend on
vehicle traffic and locale, but they generally include a second chip seal
after 1 year, followed by another in approximately 5 years. Gravel is
sometimes used on unpaved roads, but it is less effective than the chip
seal.19 In one study of road paving, an estimated 85 percent control effi-
ciency was cited.16 In a study comparing various methods of controlling
emissions from unpaved roads in Arizona, gravel paving of unimproved dirt
roads produced an estimated annual control efficiency of 50 percent.3
Road carpets are a recent development for controlling fugitive dust
from unpaved roads. A water-permeable polyester fabric is laid between the
roadbed (subsoil) and the coarse aggregate (e.g., gravel or crushed rock)
road ballast, which separates the fine soil particles in the roadbed from
the coarse aggregate. This fabric prevents fine materials from reaching the
road surface and thereby reduces fugitive dust. Any fine materials (<15
umA) in the road surface would be washed into the road ballast during rain-
fall. Fine dust in the ballast passes through the fabric into the subsoil
or to the edge of the road, but fines in the subsoil cannot be pumped up
into the ballast. It is the minimization of fines in the road surface
material that effects the reduction in fugitive dust.20
5.3.2.2 Chemical Stab11ization. Chemical stabilization requires the
use of materials that, upon drying, bind with surface particles to form a
protective crust. It acts in much the same way as physical controls (by
5-17
-------
isolating the surface from climatic factors), and it is often used in combi-
nation with vegetative stabilization. Chemical stabilization can be applied
to unpaved roads and airstrips, to waste or tailings piles, to disturbed
soil surfaces, or to reclaimed areas. Many types of chemical stabilizers
are available, and they can be applied with water or separately.
Approximately 400 hectares of the inactive Kennecott Copper tailings
area west of Salt Lake City have been successfully stabilized by aerial
chemical applications.21 Estimated control efficiencies of chemical stabil-
ization for a variety of sources are as follows.14
Source Efficiency, %
Unpaved roads 50
Construction—completed cuts and fills 80
Tailings piles 80
Cattle feedlots 40
The effectiveness of chemical stabilization on unpaved roads varies accord-
ing to the amount of traffic. Because heavy traffic tends to break up the
surface crust and to pulverize particles, it causes greater entrainment of
particles into the atmosphere.
The results of a study of various chemical stabilizers for dust control
on unpaved roads were encouraging.19 The stabilizers were applied to sec-
tions of an unpaved road with an average daily traffic of 140 vehicles and
with a surface soil silt content of 28 percent. Some of the chemicals were
applied to the surface by spray and others were mixed to a depth of 7.6 cm
and then compacted. After 5 months of surface stabilization, control effi-
ciencies of 83 to 95 percent were still being achieved. After 14 months and
several bladings, control efficiencies ranged from 9 to 54 percent, depend-
ing on the chemical used. When road stabilizers were worked into the road
surface, reductions in emissions of 80 to 95 percent were achieved after 5
months, and 12 to 84 percent after 14 months and several bladings. These
results show that working the stabilizer into the road surface causes it to
remain effective longer.3'19
5.3.2.3 Vegetative Stabilization. Vegetation is effective in the
stabilization of a variety of exposed surfaces. In many cases, modifica-
tions first must be made to the surface or the surrounding terrain.
5-18
-------
Sometimes, physical and chemical stabilization techniques are used with
vegetative stabilization. Vegetative stabilization is restricted to inac-
tive areas where the vegetation will not be mechanically disturbed once it
has been planted. Emission sources that can be controlled by vegetative
covers include mineral waste piles, road shoulders, reclaimed land, and
disturbed soil surfaces.3'16
In coal mining and preparation, both fine and coarse waste materials
(low-grade coal, ash, carbonaceous and pyrite shale, slate, clay, and sand-
stone) are produced. Because coal wastes are acidic, they must be treated
chemically and physically before vegetative stabilization can be implemented
effectively. Chemical treatment involves the addition of an alkaline
material such as limestone, fly ash, phosphate rock, or treated municipal
sewage sludge. Physical treatment involves covering the waste with soil so
that it will support vegetative growth. The use of acid-tolerant vegetative
species is recommended even when the soil has been neutralized.
Mineral mining and beneficiation produce wastes in the form of over-
burden, gangue, and tailings. Vegetative stabilization is normally no
problem with overburden and gangue, but it may be difficult to apply to
tailings because of a lack of nutrients, a concentration of saline or toxic
compounds, and extreme pH conditions. Tailings piles therefore must be
treated or covered with topsoil.
Control efficiency of vegetative stabilization varies considerably
according to the amount and type of cover. One report estimated a control
efficiency of 50 to 80 percent on tailings piles.16 In areas that are less
than hospitable to plant growth (e.g., the arid Southwest), reductions in
fugitive dust could be as little as 25 percent;3 whereas reductions could
approach 100 percent in areas that easily support dense vegetative covering.
5.3.3 Specialized Fugitive Emission Control Techniques
Some fugitive dust control techniques are relatively specific to cer-
tain processes, and thus are not as widely applicable as those just discus-
sed. Sometimes they are used to augment some of the techniques already
described.
5.3.3.1 Speed Reduction. Reducing the speed of vehicles traveling
over unpaved roads can lower the fugitive dust emissions because it reduces
5-19
-------
turbulence. Speeds of less than 48 km/h cause emissions to vary in propor-
tion to the square of the vehicle speed.22 Based on uncontrolled speeds of
64 km/h, reduced speeds would produce the following estimated reductions in
fugitive dust from unpaved roads.23
Vehicle
speed, km/h
64
48
32
24
Estimated
emission
reduction, %
0
25
65
80
Street Cleaning. Street cleaning ha
5.3.3.2
technique of reducing reentrainment of dust from paved roads. Essentially,
• '.f. . ;
three types of cleaners are now in use: broom sweepers, vacuum and regener-
ative-air sweepers, and flushers. Broom sweepers use a rotating gutter
broom to sweep material from the gutter into the main pickup broom, which
rotates to carry the material into the truck hoppers. The regenerative
air-sweeper uses an air blast to direct material into a collection hopper,
and flushers use jets of water to move material to the gutters.3
Results of field studies examining municipal street-cleaning techniques
in Kansas City, Cincinnati, and other cities were inconclusive; most of the
data showed that cleaning produced no effective reduction in emissions.3*24
Estimates of the effectiveness of cleaning paved roadways at industrial
i
facilities have estimated efficiencies of broom sweeping, vacuum sweeping,
and water flushing have been estimated at 70, 75, and 80 percent, respec-
tively.4 The broom and vacuum sweeping control efficiencies were based on
biweekly cleaning, and the water flushing was based on weekly cleaning.
5.3.3.3 Industrial Transportation Control. Another techniques for
controlling fugitive dust from paved or unpaved roads at industrial facil-
ities is to reduce the amount of vehicular traffic by providing perimeter
parking and by bussing employees to their work areas. This technique has
been used at western coal mines and at an ir*on and steel plant.
5-20
-------
The reduction in emissions that can be achieved by transportation
control is directly proportional to the reduction in vehicular traffic.
This type of control can be used with paving and street cleaning.
5.3.3.4 Windbreaks. Wind contributes significantly to fugitive dust
emissions. Reduction of surface windspeed by erecting physical barriers
(windbreaks) perpendicular to the wind direction is a logical technique for
reducing emissions. Windbreaks surrounding an agricultural field can reduce
soil erosion. The effectiveness of a windbreak extends downwind for a
distance of 10 times the barrier height.25 A barrier 7.6 m in height will
control erosion 76 m downwind of the barrier. Windbreakers can reduce
erosion by 12 percent on a typical 600-m long field; they are, however,
generally considered infeasible for the control of large fields.3
5.3.3.5 GoodOperating Practices. Good operating practices can mini-
mize fugitive dust from construction and from earthmoving activities. Such
practices include the following.2'3
Minimizing fall distances when dumping material from frontend loaders
Washing vehicle undercarriages prior to their leaving construction
sites (to eliminate mud carryout)
Limiting, by proper scheduling of activities, the exposure time of
cleared land
Little information is available on the reduction of fugitive dust that could
be achieved by good operating practices; however, reductions in emissions
should be proportional to reductions in dust-generating activities.
5-21
-------
REFERENCES . ,
1. Dunbar, D. R. Overview of the Fugitive Emission Problem—1979 SIP
Revisions. Presented at the 3rd Symposium on Fugitive Emissions:
Measurement and Control, San Francisco, California, October 23-25,
1978. EPA-600/7-79-182, August 1979.
2. Jutze, G. A., et al. Technical Guidance for Control of Industrial Pro-
cess Fugitive Particulate Emissions. EPA-450/3-77-010, March 1977.
3. Richard, G., and D. Safriet. Guideline for Development of Control
Strategies in Areas with Fugitive Dust Problems. EPA-450/2-77-029,
October 1977.
4. Bonn, R., T. Cuscino, Jr., and C. Cowherd, Jr. Fugitive Emissions From
Integrated Iron and Steel Plants. EPA-600/2-78-050, March 1978.
5. Daugherty, D. P., and D. W. Coy. Assessment of the Use of Fugitive
Emission Control Devices. EPA-600/7-79-045, February 1979.
6. Wallace, D., and C. Cowherd, Jr. Fugitive Emissions from Iron Found-
ries. EPA-600/7-79-195, August 1979.
7. Standards Support and Environmental Impact Statement Volume I: Pro-
posed Standards of Performance for Grain Elevator Industry. EPA-450/2
-77-OOla, January 1977.
8. Nicola, A. G. Fugitive Emission Control in the Steel Industry. Iron
and Steel Engineer 53(7):25, July 1976.
9. American Conference of Governmental Industrial Hygienists. Industrial
Ventilation, A Manual of Recommended Practice. 14th ed. Committee on
Industrial Ventilation, Lansing, Michigan, 1976.
10. Hagopian, J. H., and E. K. Bastress. Recommended Industrial Ventila-
tion Guidelines. Department of Health, Education and Welfare, Pub. No.
(NIOSH) 76-162, January 1976.
11. American Society of Heating, Refrigeration and Air Conditioning Engi-
neers, Inc. ASHRAE Handbook and Product Directory, 1976 Systems. New
York. 1976.
12. Weant, G. E., III. Characterization of Particulate Emissions for the
Stone-Processing Industry. Research Triangle Institute, Research
Triangle Park, North Carolina. May 1975.
5-22
-------
13. PEDCo Environmental, Inc. Reasonably Available Control Measures for
Fugitive Dust Sources. Prepared for Ohio Environmental Protection
Agency. March 1980.
14. Jutze, G. A., and K. Axetell. Investigation of Fugitive Dust. Vol. I.
Sources, Emissions and Control. EPA-450/3-74-036a, June 1974.
15. Hoenig, S. A. Fugitive and Fine Particle Control Using Electrostat-
ically Charged Fog. EPA-60Q/7-79-078, March 1979.
16. Siebel, R. J. Dust Control at a Transfer Point Using Foam and Water
Sprays. U.S. Department of the Interior, Bureau of Mines. Pub. No.
TPR 97, 1976.
17. Evans, R. J. Methods and Costs of Dust Control in Stone Crushing
Operations. U.S. Department of the Interior, Bureau of Mines. Pub.
No. 8669, 1975.
18. Dean, K. C., R. Havens, and M. W. Giants. Methods and Costs for Stabi-
lizing Fine -Sized Mineral Wastes. U.S. Department of the Interior,
Bureau of Mines. RI 7896, 1974.
19. Sultan, H. A. Soil Erosion and Dust Control of Arizona Highways.
Part IV. Final Report Field Testing Program. Arizona Department of
Transportation, November 1976.
20. Blackwood, T. R. Assessment of Road Carpet for Control of Fugitive
Emissions from Unpaved Roads. EPA-600/7-79-115, May 1979.
21. Jutze, G. A., K. Axetell, and R. S. Amick. Evaluation of Fugitive Dust
Emissions from Mining. EPA-600/9-76-001, 1976.
22. Cowherd, C., Jr., K. Axetell, C. M. Guenther, and G. A. Jutze. Devel-
opment of Emission Factors for Fugitive Dust Sources. EPA-450/3-74-
037, June 1974.
23. Compilation of Air Pollutant Emission Factors. AP-42, July 1979.
24. Axetell, K. Control of Reentrained Dust from Paved Streets. EPA-907/
9-77-007, July 1977.
25. How to Control Soil Blowing. U.S. Department of Agriculture. Farmers
Bulletin, No. 2169, July 1961.
5-23
-------
-------
SECTION 6
ENERGY AND ENVIRONMENTAL CONSIDERATIONS
Energy is consumed in the operation of participate control systems.
Determining the optimum energy usage rate for an installation involves
careful balancing of performance requirements, availability requirements,
and operating costs. Energy demands of conventional particulate control
systems, discussed in Section 6.1, emphasize the potentials for and the
limitations of energy conservation. Some control devices have the potential
for generating small quantities of secondary pollutants, not originally
present; environmental impacts of these pollutants are discussed in Section
6.2.
Particulate control devices concentrate the particulate matter en-
trained in a gas stream into a solid or liquid effluent stream for disposal
by alternative means. If disposal of the concentrated particulate is
handled improperly, the original air pollution problem can become a water
pollution or a solid waste problem. Environmental factors considered in
disposal of wastes collected in particulate control devices are addressed in
Sections 6.3 and 6.4. Other environmental considerations associated with
the operation of particulate control devices include noise management and
radiation control; these considerations are discussed in Sections 6.5 and
6.6.
6.1 ENERGY REQUIREMENTS
The goal of any evaluation of energy usage is to identify means of
minimizing the energy demand without sacrificing system performance or
predisposing the system to future maintenance problems. A complete and
realistic energy inventory is the basis for such an optimization program.
The energy requirements of a control System are simply the sum of the
energy requirements of each component. Accordingly, in the following
6-1
-------
sections, the energy demands for major components are evaluated as a
function of gas flow rate; the total system energy requirements are the sum
of the energy demands of each component at that flow rate. The results are
estimates which indicate the relative importance of various energy demands
within the systems. Optimization of energy usage is based on these
estimates.
6.1.1 Fan Energy Requirements
Each control device introduces a static pressure loss into the effluent
gas-handling system. Typical values are shown in Table 6-1. The hoods and
ducts, both before and after the control device, introduce additional static
pressure losses (1 to 10 kPa). The fan must be sized to deliver the desired
gas flow rate at the total static pressure drop associated with the control
system.
TABLE 6-1. TYPICAL STATIC PRESSURE
Control device type
Electrostatic precipitator
Mechanical collector
Fabric filter
Wet scrubber
Range of
static pressure loss, kPa
a
0.1- 0.5
0.2- 1.0
0.5- 2.5
0.5-25.0
aTo convert kPa to inches of water, multiply by 4.019.
To calculate the portion of total fan energy that is chargeable to the
control device (as opposed to the ventilation system for the process), the
fan curves must be analyzed. A comparison is made between power require-
ments at the actual operating point and power requirements for a similar fan
system without the control device static pressure loss. Subtraction yields
the incremental energy requirements due to the particulate control device.
Approximate incremental fan energy requirements due to the total pressure
losses of the ventilation system and the control device are shown in Figure
6-1;1,2 these general curves do not apply to any specific facility or manu-
facturer. These incremental energy requirements must be adjusted for the
gas temperature change during passage through the system. Most control
6-2
-------
10'
10'
S-
JC
<£>
QZ
UJ
S105
10/
"1IIII
I TT
I 1 I I- I L
_ kWh = 1.06 a 10
Where
= pressure drop, kPa
Q = gas flow rate, m3/min
h = time, 7000 hours
a = specific gravity, 0;7 (150°C)
n = efficiency, .65
10*
ID3 104
GAS FLOW RATE, «n3/min
Figure 6-1. Incremental energy requirements for fans.
6-3
-------
devices cool the gas stream 5° to 25°C. At lower temperatures, power
requirements are greater because the density of the gas is increased.4
One means of reducing incremental fan energy requirements is to
minimize static pressure losses. With electrostatic precipitators, this may
be accomplished by a smoother gas entry and exit the box. Better gas
distribution yields energy savings at the same time that performance is
improved. With fabric filters energy reduction may result from more
frequent cleaning, which allows a lower average static pressure drop across
the fabric and dust cake. For all other types of particulate control
devices, a decrease in static pressure drop is normally associated with a
reduction in performance.
The ventilation system gas flow may be improved. All sharp squared-off
turns in ducting should be replaced by smooth gradual turns having lower
flow resistance.1 Likewise, transitions in duct size should be as gradual
as possible. Unnecessary turns should be eliminated. Flow of the ,gas
stream into and out of the fan is particularly important.1'2'3
Reductions in gas flow requirements due to movement of the hood closer
to the particulate generation source could have a substantial impact in that
static pressure loss through ductwork is proportional to the square of the
gas flow rate.1 Insulation and sealing of ducts (to reduce cold air
inleakage) may be effective on large systems transporting gas at elevated
temperatures.
For cyclic processes, the fan may be shut down or dampered down when-
ventilation is not needed. There is a risk, however, that the gas tempera-
tures will repeatedly pass below the acid and water dewpoints, and lead to
corrosion. Care must be exercised when a particulate control system is
temporarily shut down.
6.1.2 Control Device Energy Requirements
Generalized energy demand requirements are presented for conventional
particulate control devices. These do not include requirements for hopper
heaters and vibrators or for solids transport equipment, both of which are
discussed separately. These estimates represent approximate values only.
6.1.2.1 Fabric Filter Subsystems. The fabric cleaning apparatus is
the dominant energy consumer within the control device. The total energy
6-4
-------
requirements are calculated as the power consumption rate times the time
that the cleaning apparatus is energized. Card, Inc.5 estimates that
reverse-air fans and shaker motors use 0.5 hp/1000 ft2. At an average air-
to-cloth ratio of 0.5 mVmin per m2 and an operating time of 2 minutes
per hour, the yearly (7000 hour) energy usage (kwh) is 1.87 x Q (ms/min). This
is an order of magnitude below that for the fan requirements. Regardless of the
type of cleaning, energy demand is considerably less than fan energy re-
quirements, based on a typical fabric filter static pressure drop of
1.5 kPa.
Energy can be reduced by reducing the frequency and intensity of clean-
ing; but in most cases, this will result in higher average static pressure
losses, and thus in higher fan energy demand. In view of the relative mag-
nitudes of energy requirements for cleaning apparatus and fans, less clean-
ing may be counterproductive. The exception is conversion to a fabric with
better cake-release properties, which requires less cleaning without in-
creasing the fan energy demand.
6.1.2.2 Electrostatic Precipitator Subsystems. Four precipitator
components use electrical energy:
0 Transformer-rectifier sets (T-R) - These convert alternating
current at line voltages to high-voltage direct current, and they
supply the discharge electrodes that enable particle charging and
collection.
0 High-voltage insulators - These isolate discharge wire rappers
from the high-voltage frame. Heaters are normally advisable to
prevent surface condensation, which allows "leakage" of current;
insulator heaters can be operated continuously or intermittently.
0 Rappers - These remove accumulated solids from collection plates,
discharge wires, and gas distribution plates. The rapper system
is operated continuously with the activation frequency for an
individual rapper normally greater on inlet fields than on outlet
fields. Rappers can be powered either electrically or by com-
pressed air.
0 Penthouse blowers - These purge the upper housing of the precipi-
tator of vapors that could condense on insulators and cause cur-
rent "leakage." They are normally operated continuously.
By far the most important energy use is by the T-R sets, which directly
control particle charging and collection. Both the peak voltage and the
total power input rates influence the level of penetration.6 There have
6-5
-------
been trends to increase both over the last 10 years; for example, design
power input rates have been increased from a range of 3 to 10 w/(mVmin)
to about 20 to 30 w/(ms/min).6>7'8 The T-R set energy demand as a function
of design input rates is shown in Figure 6-2.
Comparison of Figures 6-2 and 6-1 clearly indicates the importance of
T-R set energy consumption in an electrostatic precipitator system, where
the operative mechanism is electrostatic attraction rather than inertia!
impaction. Impaction systems are normally operated by imposing a gas flow
resistance, which is reflected in the fans high-energy demand. No such
resistance is necessary in electrostatic precipitator systems.
In actual operation, the power input to a precipitator system is not
constant over time. Substantial variability may be introduced by the input
mass loadings and the particle characteristics.6*7'9 High inlet mass con-
centrations lead to current suppression, particularly in the inlet fields.7
Likewise an increase in particle resistivity may reduce power input through-
out the unit. Power input, as calculated in Equation 6-1, may vary 30 to 50
percent daily, depending on fuel characteristics and process operating
conditions. ;
ET-R = * (Pc X Pv * *
T R i = 1 V ci vi
where Ey_R = total energy input to power supplies,
n = number of transformer-rectifier (T-R) sets,
PC. = primary current of T-R set i,
Pv. = primary voltage of T-R set i,
t. = time T-R set i operational.
of = power factor
Normally, the power input to the electrodes, E-r_R, cannot be reduced.without
an adverse impact on penetration. New power supplies are designed to maxi-
mize power input.6'7
The power consumption of the rapper system is the sum of the power
input to each rapper times the fraction of time that each is activated.
Because the rappers normally operate only a small portion of the time, even
on inlet fields, the energy consumption is more than an order of magnitude
6-6
-------
10'
I I I I I I I I
LU
CO
BASIS: 7000 h/yr
10
i i i i i
i i
GAS FLOW RATE, ms/min
Figure 6-2. Energy required for transformer-rectifier set.
6-7
-------
less than that for the T-R sets. Actual energy demand depends on the number
of rappers installed, the frequency of activation, and the intensity of
operation.
Reduced rapper energy use may offer several secondary benefits. Lower
intensity and lower frequency rapping of outlet fields can reduce reentrain-
ment losses in certain installations. Reduced rapping intensity may also
lower the probability of electrode misalignment, and thereby reduce future
maintenance costs. There is a limit, however, on the extent to which rapper
energy savings are feasible. That limit is indicated by simultaneous in-
creases in primary voltages and decreases in secondary currents to the
various energized sections. When these changes occur, excessive solids
accumulate on collection plates, and precipitator performance begins to de-
teriorate. The solids buildup can dampen rapper shocks transmitted down the
electrodes, and thus further aggravate the solids problem.6
High-voltage insulator heaters are operated continuously while the pre-
cipitator is operational. Total energy consumption, indicated by curve A of
Figure 6-3, depends on the number of heaters and on the rated power input.
Despite the continuous operation, the total requirements are relatively
minor because of the small number of heaters. This component is a rela-
tively poor candidate for energy conservation because inadequate heating; can
lead to vapor condensation on insulator surfaces and to voltage reductions
and penetration increases.
Energy demand for purge air fans is shown by curve B in Figure 6-3. As
with insulator heaters, the requirements depend on the number of blowers
used and on their sizes. These fans are typically operated continuously.
Potential energy savings are very limited, and are gained at the risk of
increased condensation on high-voltage insulators.
6.1.2.3 Wet Scrubber Subsystems. Components using energy in a wet
scrubber system are identified below. In most cases, the instrumentation
power demand is small relative to the three listed components.
0 Pumps - for liquor recirculation, makeup water supply, and purge.
0 Agitators - for mixing chemicals for neutralization of scrubbing
liquor.
6-8
-------
1IIIIII)
10'
>>
gio
10*
J I 111 I I I
1 1 J 1111
10*
KT 10-
GAS FLOW RATE, mVmin
Figure 6-3. Energy required for ESP insulator heaters and
purge air fans.
6-9
-------
0 Reheat - by steam injection or by indirect oil-fired burners for
increasing stack temperatures and thus improving the meteorolog-
ical dispersion characteristics of the effluent.
The energy required to operate pumps is a function of the liquid flow
rate and the head.10 Centrifugal pumps are common in wet scrubber systems.
General power requirements for pumps are shown in Figure 6-4; the liquid-to-
gas (L/G) ratios apply only to the recirculation pump. The purge and makeup
flow rates are generally only 1 to 5 percent of total recirculation flow,
and the energy demands are low relative to the recirculation pumps. Energy
-„ i
demand, of pumps can be reduced either by reducing recirculation flow or by
i
modifying spray nozzles, but both changes can lead to reduced performance.
Energy demand for stack gas reheat depends primarily on the gas flow
rate and on the degree of heating desired. Typically, the scrubbed gas
stream is at the adiabatic saturation temperature (normally 50°C) and the
necessary reheat is 25° to 50°C. The energy requirements are shown in
Figure 6-5,
Stack gas reheating prevents objectionable ground-level concentrations
of unremoved pollutants. Means of minimizing reheat costs include increas-
ing stack height, increasing stack gas exit velocities, and eliminating
pollutants not treatable in the wet scrubber.11
6.1.2.4 Mechanical Collector Subsystems. Mechanical collector subsys-
tems are passive. None of the components within the device use energy.
Solids removal equipment is addressed in a later section.
6.1.2.5 Incinerator Subsystems. The incinerator is the only particu-
:::' ~"~' " " v """""""""" ™ ™"niiiinni1 nni"11™1 |
late control device directly using fossil fuels. The quantity used depends
on gas flow rate, combustion temperature, inlet gas temperature, gas compo-
sition (H20, C02), and heating value of gas stream. Means of calculating
energy requirements are discussed in Section 4.6.
Reduced fuel requirements leading to reduced combustion chamber temper-
atures generally leads to higher particle penetration, especially for parti-
cles that are difficult to volatilize. A heat exchanger could prevent
reduced temperatures, but poor incinerator performance could cause fouling
of the hot side. '.
6-10
-------
10'
es
oc
UJ
10" —
10
Q(H)(S.G.Hhl
flow rate, L/min
head of fluid, m,30
specific gravity, 1.0
efficiency, 0.65
hours of operation, 7000
GAS FLOW RATE, mVmin
Figure 6-4. Energy required for pumps.
6-11
-------
INLET TEMPERATURE 125°F
15% RADIATION LOSS
TO2 103
GAS FLOW RATE, m3/min
Figure 6-5, Energy required for stack gas reheat.
6-12
-------
6.1.3 Hopper Heaters and Vibrators
All dry participate control devices—Including mechanical collectors,
fabric filters, and electrostatic preclpitators--occasionally require hopper
heaters and vibrators to maintain free-flowing discharges of solids. Nor-
mally, heaters are used only on larger devices when high-temperature gas
streams are being treated. The energy requirements for heaters and vibra-
tors depend on facility size but are independent of control device type.
Heaters and vibrators are normally operated in cycles. Hopper heaters
are thermostatically controlled to maintain temperatures above both the acid
and the water dewpoints. Vibrators are activated by timers from 1 to 20
times an hour; the energized time period can vary from several seconds to a
minute. Energy requirements for heaters and vibrators are a relatively minor
part (<10%) of total system energy demand.
Some energy savings may be realized by derating the heaters and reduc-
ing the vibrator on-time; such savings, however, are gained at the risk of
maintenance problems, and such changes should be carefully planned. Failure
to properly discharge solids can lead to misalignment of precipitator elec-
trodes, to fabric deterioration, or to plugging of mechanical collector
tubes.
6.1.4 Solids Discharge and Transport
Dry particulate control devices normally incorporate some type of
solids discharge valve at the bottom of the hopper and sometimes at screw
conveyor transfer points. Solids can be transported by pneumatic systems,
pressurized systems, screw conveyors, or drag conveyors. The screw conveyor
is normally used with small systems.
For solids transport and discharge equipment, the energy requirements
are normally directly proportional to the mass of material. With transport
equipment, the distance moved must also be considered.
6.1.5 Ultimate Disposal
Energy is needed to transport solids from a temporary storage site
(e.g., a pile or covered pit) to the ultimate disposal site (e.g., land-
fill). Transport requires energy, but the demand is a relatively small
6-13
-------
fraction of the overall system energy demand because the solids are in a
concentrated, manageable state.
6.1.6 Other Considerations
Recent increases in energy costs have motivated some operators to con-
sider recycling certain treated gas streams to occupied working areas. This
may have a substantial benefit on costs of space heating, but this should be
adopted cautiously, with full consideration of potential occupational health
impacts of control system malfunctions.
6.2 SECONDARY POLLUTANT GENERATION
Participate control devices have the potential for generating limited
quantities of gaseous and particulate air pollutants. Minimizing the gener-
ation of these pollutants requires an understanding of their physical and
chemical formation mechanisms. Unfortunately, these mechanisms are not
fully understood. :
6.2.1 Electrostatic Precipitators
Ozone can be formed in negative-corona electrostatic precipitators.6
Concentrations of 5 to 20 ppm have been reported.6 Ozone is toxic at con-
centrations normally encountered in precipitators. Accordingly, strict
confined entry procedures should be followed before maintenance is begun.
Ozone generation is only partially understood. Most mechanisms include
ionization of molecular oxygen during a spark incident or ionization of
molecular oxygen due to absorption of high-energy ultraviolet light emitted
by the corona. The former mechanism should be controllable by the reduction
of spark rate, which is now possible because of advanced power supplies.
The latter mechanism appears more difficult to minimize without affecting
precipitator performance. The relative importance of these mechanisms is
not known.
There is limited evidence that some sulfur dioxide is oxidized to
sulfur trioxide in a negative-corona precipitator.12 Such substances should
either form sulfuric acid aerosols or condense on available particle
surfaces. In either case, the generated sulfate would be indistinguishable
from that generated in the combustion process. The mode and extent of
sulfate formation are not well understood.
6-14
-------
6.2.2 Incinerators
Any combustion process can form some nitrogen oxides. Formation
mechanism is thought to be reasonably represented by the following reac-
tions, which collectively are referred to as the Zeldovich mechanism.*
1. N2 + 0 -> NO + N.
2. M + 02 •* M + 20 (M = any third body molecule).
3. N + 02 •* NO + 0.
Reaction 1 has high energy; accordingly, the mechanism is active only at gas
temperatures exceeding 1400°C. There is also a strong impact by the oxygen
content caused by reaction 6-2. Basically, these reactions are important
only within the flame; after the gases leave the combustion zone, the reac-
tions cease, and nitric oxide remains.
Considerable research has been devoted to techniques for suppression of
nitrogen oxides formation in coal-, oil- and gas-fired boilers, but very
little of this work is directly applicable to burners of the scale and type
used for particulate incineration. Available means of reducing nitrogen
oxides generation include reduced flame temperatures and reduced excess air.
6.3 LIQUID WASTE MANAGEMENT
The major sources of water pollutants are effluents from scrubbers and
from sluicing systems for removing particulate from hoppers. An indirect
source is the leachate produced when rain and surface runoff percolate
through collected particulate matter that has been disposed of improperly.
6.3.1 Regulatory Requirements
Regulatory requirements that apply to effluents streams from particu-
late control devices are similar to those that apply to the process being
controlled. Effluents from scrubbers, wet electrostatic precipitators, and
controls devices using wet sluicing systems are regulated along with other
plant sources under the Federal Water Pollution Control Act (Clean Water
Act) as "direct dischargers". Direct dischargers are point sources that
* The symbol M refers to any third body molecule.
6-15
-------
must conform to numerical limits on various pollutants under Federal
effluent guidelines developed on an industry-by-industry basis and based on
the Best Practical Control Technology (BPCT). Effluent guidelines based on
BPCT and BACT apply to existing sources at the points of discharge from the
plant treatment facility. New effluent sources are subject to demonstrated
BACT, processes, operating methods, or other alternatives including (where
practicable) a standard permitting no discharge of pollutants.
A rigorous discussion of water pollution regulation is beyond the scope
of this report. Along with the basic Federal effluent guidelines, other
subsections of The Clean Water Act specify control requirements for
i
"priority pollutants" and chemical industry pollutants, which might be
subject to rules and regulations under the Toxic Substances Control Act
(TSCA) enacted by Congress in 1976.
Most States have EPA-approved National Pollutant Discharge Elimination
Systems (NPDES) permit programs. Such programs enable State to issue
permits to sources that comply with the requirements of the Clean Water Act,
provide for public participation in the permit issuing process, and give
EPA, the Corps of Engineers, and other States the opportunity to object to
the issuance of a permit. Particulate controls that generate effluent
streams should be included on the NPDES permit.
In general, particulate control systems to which effluent regulations
apply are subject to the same regulations as the" source being controlled.
i
For example, a scrubber controlling a chemical process subject to TSCA might
collect toxic air pollutants; the presence of these toxic substances in the
scrubber effluent could bring the scrubber under the TSCA guidelines.
6.3.2 Control Techniques
The appropriate treatment for scrubber wastewaters can be selected on]y
after the wastewaters have been completely characterized. Various constit-
uent pollutants of scrubber liquors, such as suspended solids, dissolved
solids, toxic metals, biodegradable organics, and acids or caustics require
different types of treatment to meet regulatory discharge requirements.
6.3.2.1 Primary Treatment. Suspended solids, found in nearly all
scrubber wastewaters, are removed by sedimentation (commonly referred to as
6-16
-------
primary treatment). Sedimentation is accomplished by allowing the waste-
water to flow slowly through a large basin or pond so that suspended parti-
cles collect by gravity. Clarifier basins (Figure 6-6) use automatic
mechanical devices to continuously remove accumulated sludge, whereas set-
tling pond sludge is removed in a batch fashion. Increasing the size of a
sedimentation basin or pond increases the wastewater detention time, and
thus improves sedimentation efficiency. Sedimentation of very small parti-
cles can be improved by adding flocculants to the wastewater prior to sedi-
mentation to cause the fine particles to agglomerate into larger, more
easily separated particles.13*14
13
Figure 6-6. Sedimentation tank or "clarifier."
6.3.2.2 Secondary Treatment. Some scrubber wastewaters may contain
biodegradable organic compounds which, if untreated, will exert a biological
oxygen demand (BOD) on receiving waters. Such organic compounds can be
found in scrubber wastewaters at pulp and paper plants, wood products
plants, food processing plants, and other industries. Such BOD-causing
wastes must receive biological treatment, commonly referred to as secondary
treatment, prior to discharge. Failure to treat BOD-causing wastes can
6-17
-------
cause the growth of aerobic bacteria in receiving waters, the accompanying
depletion of dissolved oxygen, the death of naturally occurring organisms in
the water, and the growth of anaerobic odor-causing bacteria.
Secondary treatment is accomplished by allowing aerobic microorganisms
to biologically degrade the BOD-causing organic compounds before their re-
lease into surface waters. This biological activity requires a constant
influx of oxygen into the wastewater and sufficient detention time. Second-
ary treatment is usually accomplished by oxidation ponds, trickling filters,
or activated sludge units. Additional sedimentation must follow trickling
filters and activated sludge units to remove suspended bacteria from the
wastewater.1S'1S
6.3.2.3 Tertiary Treatment. Although primary and secondary treatments
may remove most suspended solids and BOD from wastewaters, but advanced
wastewater treatment, sometimes referred to as tertiary treatment, are
needed for pollutants such as nitrogen, phosphorus, inorganic acids, non-
biodegradable organics, and heavy metals.
In many particulate scrubbers, especially those at phosphate rock mines
and at fertilizer plants, the liquors contain phosjahorus and nitrogen and
excessive discharge of phosphorus and nitrogen disrupts the ecological
balance in lakes and streams by stimulating profuse growth of algae. Phos-
phorus is removed by adding coagulants such as alum, lime, and ferric chlo-
ride to convert the phosphorus to an insoluable form.13 These additions
cause the phosphorus to precipitate, coagulate, and settle from the waste-
water. Coagulants for phosphorus removal can be added before primary treat-
ment, before secondary treatment (if is required), or as a separate tertiary
treatment.
Three major processes are used to remove nitrogen from wastewaters.
Biological nitrification-denitrification is the aerobic biological conver-
sion of nitrogenous matter into nitrates (nitrification), followed by ana-
i
erobic biological conversion of nitrates into nitrogen gas for release to
the atmosphere (denitrification); both steps resemble the activated sludge
process in secondary treatment, but different microorganism colonies develop
in the vessels, and denitrification occurs without oxygen. A second method
is ammonia stripping by raising the pH of the wastewater and passing the
6-18
-------
water through a stripping tower, from which gaseous ammonia is released to
the atmosphere; scrubbing liquor nitrogen must be in the form of ammonium
ions for stripping to be effective. The third method is selective ion ex-
change, which resembles a home water softener; ammonium ions in solution are
exchanged for sodium or calcium ions displaced from an insoluble exchange
material.13
Scrubbing liquors from cast iron cupolas, certain chemical processes,
incinerators, steel mills, and many other industries can contain soluble
organics such as phenols and benzene, or colloidal oils, which are resistant
to biological breakdown during secondary treatment. These impurities, often
referred to as "refractory organics," can be responsible for unwanted colors
or tastes in water, and many are suspected carcinogens. Refractory organics
are generally removed from wastewater by carbon adsorption: wastewater is
passed through a bed of granular activated carbon, where the organics are
adsorbed onto the carbon surfaces until the carbon becomes saturated and has
to be regenerated or replaced.13
In scrubber wastewaters, toxic heavy metals such as chromium from metal
plating operations, or lead, mercury, and copper require special treatment.
Soluble metals can be removed by one of two methods. One is high-pH lime
coagulation, which is especially attractive if phosphorus must also be
removed; the other method is selective ion exchange, which is especially
attractive if nitrogen must also be removed.13'15
Some wastewaters contain significant colloidal material even after co-
agulation-sedimentation in primary or secondary treatment. If this colloi-
dal material is not suitable for discharge because of high turbidity or its
chemical nature, filtration can be accomplished by passing the wastewater
through a granular bed of sand or other small particles. As the filter
becomes plugged, it can be cleaned by briefly reversing the flow ("backwash-
ing") at a high flow rate. Backwash wastewaters, usually less than 5 per-
cent of total flow, must be recycled to the wastewater treatment plant.13
6.3.2.4 Other Treatment Considerations. Several other points must be
considered when selecting a treatment or wastewater from a scrubber. Should
scrubber wastewaters be combined with other plant wastewater streams or
treated separately. If the wastewater contains toxic materials or has ex-
treme pH values, the routing of scrubber effluents directly into plant (or
6-19
-------
municipal) treatment facilities can "poison" biological treatment processes.
Scrubber liquors which contain neither biodegradeable organics nor toxic
materials and which require no secondary or tertiary treatment, can be
treated most economically by primary treatment separate from plant treatment
facilities. Other liquors can be routed directly to combined treatment
facilities, or they can be treated for pH control before being routed to
i
combined facilities. If scrubbing liquors are changed in a batch fashion
instead of by continuous blowdown, flow equalization facilities may be re-
quired to prevent overload of plant treatment facilities.
6.3.2.5 Sludge Handling. Purifying scrubber wastewaters can lead to
another problem—sludge handling. Sludges withdrawn from treatment proces-
ses are still largely water (often more than 90%). Thus, sludge treatment
must separate solids from the large amounts of water, return the separated
water to the wastewater plant for reprocessing, and dispose of the separated
solids in an environmentally appropriate manner according to applicable
regulations (Section 6.4).
Several processes are available for the dewatering of sludge. One or
more of these processes may be required to properly dewater a particular
sludge. A common first step in dewatering is sludge conditioning, where
coagulants such as ferric chloride, lime, or organic polymers are added to
more easily separate sludge solids from water. After conditioning, sludges
are often thickened by gravity settling in vessels similar to wastewater
clarifiers to reduce sludge volume by a factor of 2 or more. Biological
sludges (i.e., from secondary treatment clarifiers) often require sludge
"stabilization" to breakdown organic solids so that they are more stable
(less odorous and less putrescible). Stabilization can be accomplished in
anaerobic and aerobic biological sludge digesters.13
Many thickened or stabilized sludges receive final dewatering by vacuum
filtration (Figure 6-7). A vacuum filter consists of a cylindrical drum
covered with a filtering material or fabric partially submerged in a vat of
conditioned sludge. A vacuum is applied to the inside of the drum to ex-
tract the water, leaving the solids or "filter cake" on the filter medium.
A blade scrapes the filter cake from the filter medium as the drum
rotates.13 Some sludges are filtered more readily if the filter medium is
6-20
-------
precoated with various dusts, usually applied daily. Another method of
final dewatering is drying on sandbeds. After dewatering by vacuum filtra-
tion or sand drying, sludges are either incinerated or landfilled, depending
on their properties.
Figure 6-7. Vacuum filter.13
6.4 SOLID WASTE MANAGEMENT
Substantial quantities of solids and sludges are collected in particu-
late control systems. Ideally, these materials are recycled to partially
offset control costs and to avoid disposal costs. Unfortunately, the physi-
cal and chemical characteristics of these materials frequently render them
noncompetitive in the very limited markets presently available. This sec-
tion describes the general chemical and physical properties that determine
recycle potential and disposal requirements. Information specific to par-
ticular industries and processes is available in the Background Information
6-21
-------
Documents for the Standards of Performance for New Sources and in the
Industrial Process Profiles for Environmental Use.
6.4.1 Regulatory Requirements
The major Federal regulations affecting solid waste disposal were
established under the Resource Conservation and Recovery Act (RCRA). Under
this Act, the generation, treatment, and disposal of solid (hazardous)
wastes are strictly regulated. In most situations, a permit is required by
a generator (e.g., a facility that generates scrubber sludge), regardless of
whether the generator disposes of the hazardous material on or off site. In
addition, the generator, transporter, and the disposal facility must com-
plete a manifest form every time hazardous waste is shipped, so that the
',. ' ,1,1 ' ,"i,,i!i"! ,i' • ,
waste can be tracked from "cradle to grave" and to eliminate illegal dis-
posal of hazardous wastes.
When waste from a wet scrubber or a wet electrostatic precipitator is
disposed of in an underground injection well, a permit must be obtained,
according to the Safe Drinking Water Act of 1974. The Act established a
national program to prevent underground injections that endanger drinking
water sources.
6.4.2 Waste Recycle
. f
Properties of accumulated materials determine the extent to which reuse
is economically attractive. A partial list of important physical and chemi-
cal properties is provided below:
Physical properties Chemical properties
Loss on ignition pH
Carbon content Soluble fraction
Particle size distribution Trace element composition
Moisture content
Pozzolanic activity
"Loss on ignition" and "carbon content" measure similar properties, in that
carbonaceous materials normally constitute most of the combustible fraction
of collected solids. Loss on ignition of fly ash from a pulverized-coal -
fired boiler normally ranges from 0.2 to 17 percent by weight. The normally
acceptable range is 6 to 12 percent by weight if fly ash is to be used as a
filler in portland concrete and asphalt concrete; this strict range limits
6-22
-------
the use of fly ash as a filler. In 1972, only 11 percent by weight of the
fly ash from pulverized coal was used as a filler. Since the carbon content
of fly ash from a stoker-fired boiler is typically 25 percent or more, this
material is unsuitable for use in concrete.
Particle size distribution is important. Excessive fines measured by
the minus-325 mesh fraction generally inhibit use. For use as filler in
cement, the fine fraction should not exceed 12 percent by weight; however,
the reported values for ash from pulverized-coal-fired boilers are 7 to 60
percent. The fines of any solids collected in particulate control devices
depend on the effectiveness of agglomeration in the hopper and in solids
transport equipment. (Agglomeration is caused by condensation of moisture
and inorganic vapors on particle surfaces during cooling.) Excessive fines
renders any use unattractive because of the fugitive dust created during
materials handling.
Moisture content influences recycle potential in a variety of ways.
Low moisture content (<3.5% by weight) increases potential dusting problems,
as discussed above, and increases the explosive potential of dusts. High
moisture content (>20-40% by weight) leads to materials-handling problems
because the solids begin to agglomerate and to cake, and may have an adverse
impact on the process fuel requirements. For these reasons, the solids
collected in wet scrubbers are rarely recycled.
The chemical composition of the dry particulate catch can affect use.
For example, the alkali content of portland cement kiln dust must be below
the limits stated in product specifications. Often it is possible to use
solids from the inlet fields of the precipitator, but not from subsequent
fields, which remove the majority of entrained alkali particles.
The potential for groundwater contamination from waste disposal sites
depends partially on the quantity of water-soluble compounds that could
leach out. Fly ash from pulverized-coal-fired boilers contains approxi-
mately 1 to 5 percent by weight water-soluble compounds. The major dis-
solved compounds include sulfates, chlorides, and calcium ions; arsenic,
mercury, and cadmium compounds are moderately soluble. The residue from
municipal incinerator collectors can have water-soluble fractions as high as
15 percent by weight.
6-23
-------
6.4.3 Waste Disposal
Waste characteristics and regulatory requirements must be fully con-
sidered in the selection of a disposal technique. Possible disposal tech-
niques include (1) placement in lined or unlined ponds, (2) placement in a
landfill, either as-received or after fixation treatment, and (3) deep well
injection.
Unlined ponds are satisfactory when leachate from the waste liquors or
slurries can be controlled either by special pond construction (e.e., under-
drainage systems) or by soil permeability properties. Various techniques
are available for analysis of pond leakage include soil resistivity moni-
toring networks and underdrainage system inspection sumps.
Among the liner materials available to improve leachate security in
ponds are various synthetic materials and clays. Flexible liners have an
estimated life of 20 to 25 years; nonflexible liners are more permanent.
Liner thicknesses range from 0.025 to 0.075 cm for synthetic materials such
as polyethylene and polyvinylchloride, from 30 to 40 cm for clays, and
approximately 15 cm for asphalt and concrete. Cost of the lining must be
weighed against the security offered by thicker liners.
Dry or dewatered material can be disposed of in a landfill. The sta-
bility of the fill material and the groundwater contamination potential
should be considered. For example, water runoff should be channeled around
these sites to minimize leaking of soluble compounds. Soil characteristics
and groundwater levels should be determined to avoid improper landfill loca-
tion. ' j
Treatment of the waste may be necessary to reduce potential landfill
problems. Scrubber sludges can be treated chemically to "fix" the material
into a physically stable, leach-resistant matrix. Wastes can also be fil-
tered to reduce moisture content.
Deep well injection generally is not economically attractive for dis-
posal of the quanitities and types of wastes discharged from particulate
control devices. >
6-24
-------
6.5 NOISE MANAGEMENT
Noise from particulate controls is generated by fans, ESP rapping, wet
scrubber pumps, and solids transportation systems. The noise levels from
most of these sources are usually negligible compared with noise from the
other plant sources. Objectionable noise levels attributable to particulate
control systems are most frequently generated by fans. If there is a need
to decrease the noise level from fans, the rotational speed should be
reduced and the additional capacity shifted to a parallel fan. Sound insu-
lation can also be used for fans as well for other components, such as ESP
rappers.
6.6 RADIATION CONTROL
Radiation sources associated with particulate controls are limited to
nuclear level indicators for hoppers. Radiation from such indicators is
usually minor and not sufficient to warrant the requirement that plant per-
sonnel wear nuclear badge detectors or dosimeters. These potential radia-
tion exposure areas, however, must be marked in accordance with Federal
regulations. Periodic checks could be made for radiation leakage with a
Geiger-Mueller detector.
6-25
-------
REFERENCES
1. American Conference of Governmental Industrial Hygienists. Industrial
Ventilation Manual,, 16th edition. Lansing, Michigan, 1980.
2. Jorgensen, R. , editor. Fan Engineering 7th edition, Buffalo Forge
Company, Seventh Edition, 1970.
3. Pollak, R. Selecting Fans and Blowers. Chemical Engineering, 80(3),
January 23, 1973.
4. Heath, C. Energy Analyses. National Asphalt Paving Association.
Publication 52, October 1978.
5. Neveril, R, B. Capital and Operating Costs of Selected Air Pollution
Control Systems. EPA-450/5-80-002. December 1978.
6. Katz, J. The Art of Electrostatic Precipitation. S & S Printing
Company, Pittsburgh, Pennsylvania, 1979.
7. Szabo, M., and R. W. Gerstle. Operation and Maintenance of Participate
Control Devices in Kraft Pulp Mill and Crushed Stone Industries.
EPA-600/2-78-210, 1978. ;
8. Szabo, M., and Y. Shah. Inspection Manual for Evaluation of Electro-
static Precipitator Performance. EPA-340/1-78-007, February 1979.
9. White, H. Electrostatic Precipitation of Fly Ash. Air Pollution Con-
trol Association, Pittsburgh, Pennsylvania, July 1977.
10. Karassik, I., and W. Krutzsch. Centrifugal and Axial Pumps. In:
Standard Handbook for Mechanical Engineers, 7th edition. McGraw-Hill
Book Company, New York, N.Y., 1967.
11. PEDCo Environmental Specialists, Inc. Flue Gas Desulfurization Process
Cost Assessment. Draft report for U.S. Environmental Protection Agency
under contract no. 68-01-3150.
12. Brown, W. R. and E. E. Stone. Sulfur Dioxide Conversion under Corona
Discharge Catalysis. U.S. Public Health Services contract no. PH-86-
67-2, March 1965.
13. Environmental Pollution Control Alternatives: Municipal Wastewater.
U.S. EPA Technology Transfer. EPA-625/5-76-012, 1976.
6-26
-------
14. Metcalf and Eddy, Inc. Wastewater Engineering. McGraw-Hill Book
Company, New York, N.Y., 1972.
15. Weber, Jr., W. J. Physiochemical Processes for Water Quality Control.
Wiley-Interscience, New York, N.Y., 1972.
6-27
-------
-------
SECTION 7 ,
COSTS OF PARTICIPATE CONTROL EQUIPMENT AND
FUGITIVE EMISSION CONTROL TECHNIQUES
The selection of a particulate control technique depends upon many
factors, such as degree of required emission reduction, gas stream charac-
teristics, and cost. This section deals with the costs of purchasing,
installing, and operating various particulate control devices and tech-
niques. The particulate control equipment addressed includes electrostatic
precipitators, fabric filters, mechanical collectors, incinerators, and
scrubbers. The fugitive emissions control techniques evaluated include wet
suppression and stabilization.
The total cost of a parti cul ate control system is influenced by many
factors. The cost of the control and auxiliary devices can be highly vari-
able in view of the many options that may be applicable. Even when the type
of equipment is selected, different materials of construction and penetra-
tion levels affect costs to a large extent. Auxiliary instrumentation
useful to improve reliability and to reduce maintenance can also have a
great effect on cost:
Retrofit applications can result in installation costs that greatly
exceed the costs of a new installation. Space restrictions, difficult
tie-ins, and outdated process equipment can cause added expense.
In addition to the collection costs of particulate, a cost estimate of
residue disposition must be made. Collected solids or sludges may consti-
tute a significant expense for disposal, or, when the particulate consists
of recovered product, a valuable credit. Facilities for treatment of wet
scrubber sludges may be necessary.
Labor expense affects both capital and annualized costs in the form of
installation, operation, and maintenance. This expense also carries with it
sizable overhead costs.
7-1
-------
Given compliance with other limiting factors, cost-effectiveness often
provides a primary criterion for choosing among various pollution control
alternatives. Cost-effectiveness is a measure of the total cost of a speci-
fied reduction in emissions. Computation of cost-effectiveness must take
into account all annualized costs including direct operating costs and
capital charges. ;
7.1 PARTICULATE CONTROL EQUIPMENT COST ANALYSIS
7.1.1 Capital Costs
Capital costs of a particulate control system include the cost of the
purchased equipment; i.e., the control device and its accessories, and all
installation costs. Installation costs are divided into direct costs,such
as those for foundations, piping, and painting, and indirect costs such as
those for engineering and supervision. Equipment costs generally form the
basis for estimating total capital costs for a particulate control system.
There are several methods for estimating the total capital costs of
particulate removal systems. The accuracy of any method is directly related
to the amount and detail of the information available. Simple cost-esti-
mating methods dependent solely on the type of unit and its capacity are the
least accurate; methods requiring preliminary engineering drawings and
specifications and detailed energy balances are generally the most accurate.
A discussion of several estimating techniques follows.
An order-of-magnitude estimate of total capital costs is based on the
average cost for equipment of a particular type and capacity. For air
pollution control facilities, these parameters are usually determined by gas
volume to be treated and the desired pollutant removal efficiency. The
range of average capital costs can be very wide; thus this simple method of
\
estimating costs is the least accurate.
Another method of estimating total capital costs is that in which
factors are applied to the cost of the major pieces of equipment as a means
of estimating the remainder of the costs. Experience gained from previous
projects (historical data) provides the multiplication factors used in this
method. There are several variations of the factor method, each requiring
7-2
-------
different input and producing different degrees of accuracy. The Lang,
Chi 1 ton, and Guthrie procedures are three variations that were developed for
use in estimating the cost of chemical plant construction. The method of
cost estimation used to obtain the cost curves presented in Section 7.3 is a
variation of the Lang method that EPA has developed for application to air
pollution control systems. This system is fully explained in Reference 1.
When total capital cost data are available for a system of similar
design but of different'capacity from that required for a particular appli-
cation, a scaled estimate can be utilized. Scaled estimates are usually
derived by use of the following equation:
E2 = E1 (^)n (Eq. 7-1)
where
Eg = cost of desired control device
E, = cost of scaled control device
r« = capacity of desired control device
r-, = capacity of scaled control device
n = exponent relation
This equation specifies that a log-log plot of capacity versus cost
should be a straight line with slope n. With respect to equipment costs, n
has been shown to average 0.6 and is referred to as the six-tenths factor.
Although the six-tenths factor is most accurate when applied to a single
piece of equipment, it can also be used for estimating complete system
costs. Its use should be limited, however, to cases where no other costs
are readily available for the desired control device.
The most accurate methods of capital costs estimation require complete
drawings and specifications, material and energy balances, site surveys, and
other engineering effort. The extent of the data obtained will determine
the accuracy provided by the estimate. Accuracy of ±5 percent is possible
if sufficient information is available. In contrast, however, cost spreads
7-3
-------
of 20 percent are sometimes encountered in formal bids as a result of market
conditions, interpretation of plans and specifications, subjective assess-
ment of installation difficulties, and error.
i
7.1.2 Annualized Costs
,*mmmmmmm~—,, . . n.i.i^—.n.-,. , ,..— |
Annualized costs of a particulate control system include direct costs
such as operating labor and materials, maintenance, replacement parts,
utilities, and the costs of particulate disposal; also included are indirect
costs such as overhead, insurance, taxes, and capital recovery, and credits
derived from recovery of particulate product.
Direct annualized cost estimates are obtained by applying unit costs of
utilities, labor, and materials to the estimated requirements for these
items. Indirect costs are derived generally by combining percentages of the
capital costs with a percentage of labor charges for operation and mainte-
nance and capital recovery costs. Capital recovery costs depend on interest
rates, the useful life of the equipment, and the equipment's salvage value,
if any.
7.1.3 Other Cost Considerations
Certain aspects of cost analysis are always affected by some degree of
uncertainty. When new technology is included, an adequate data base may not
be developed for proper evaluation. Differences in the expected service
lives of alternative systems may require appropriate adjustments in cost
comparison. Changes in labor rates, material costs, or fuel costs also may
affect the accuracy of cost analysis.
Inflation is always an important consideration,, since inflation rates
are subject to change, reflecting other economic factors. Cost indexes are
available to aid in adjusting past costs into current dollars.
The cost of retrofit applications can be difficult to assess by use of
typical cost analyses. Additional engineering input is important to account
properly for potential additional expenses for such items as site prepara-
tion, overtime labor, utility system modifications, space restrictions, and
lost production. By the methods described in Reference 1, retrofit cost
curves have been developed and are compared with those for grass-roots
installations in Section 7.3.
7-4
-------
7.2 METHODOLOGY FOR ANALYZING COST OF PARTICULATE CONTROL SYSTEMS
The method of cost analysis used in the analyses presented here is
identical to that used in Reference 1, which should be referred to for a
detailed explanation of cost analysis of air pollution control systems. A
brief explanation of the method is included in this section.
7.2.1 Capital Costs
Purchased equipment costs provide the basis for estimating the remain-
ing capital costs; i.e., direct and indirect installation costs, for partic-
ulate control systems. The purchased equipment costs include the price of
the control device, auxiliary equipment, instruments and controls, taxes,
and freight. Although the cost of a device and its auxiliaries may be
fairly standard for a particular size and type, the costs of instrumentation
and freight can vary considerably, depending upon the type and location of
application.
Installation costs are derived by applying the applicable cost factors
shown in Table 7-1. Many of the individual items in the installation cate-
gories, direct and indirect, are subject to site-specific adjustment. In
the direct cost category, erection and handling, site preparation, and
facilities and buildings are subject to adjustment. Purchased equipment
cost does not, however, directly determine the work needed in preparing the
site of erecting buildings. These cost items are more dependent upon the
nature of the facility and whether the application is new or retrofit. The
costs of foundations and supports, electrical work, piping, insulation, and
paintings are all generally proportional to the purchased equipment cost,
and adjustments are not deemed necessary.
Many indirect costs can vary considerably. Engineering and super-
vision, construction and field expenses, and construction fee depend to some
degree on site-specific conditions and therefore require appropriate adjust-
ment.
Table 7-2 indicates that total direct and indirect installation cost
factors can range widely. The importance of knowledge of the specific
application is clear; use of the appropriate cost adjustments can be the
most important aspect of this cost analysis.
7-5
-------
Table 7-1. AVERAGE3 COST-FACTORS FOR ESTIMATING CAPITAL COSTS
Cost factors
Direct costs
1.
Purchased equipment costs
a) Control device
b) Auxiliary equipment
c) Instruments and controls
d) Taxes
e) Freight
Subtotal
2.
Installation direct costs
a) Foundations and supports
b) Erection and handling
c) Electrical
d) Piping
e) Insulation
f) Painting
g) Site preparation
h) Facilities and buildings
Subtotal
Indirect costs
3.
Installation indirect costs
a) Engineering and supervisior
b) Construction and field
expense
c) Construction fee
d) Startup
e) Performance test
f) Model study
g) Contingencies
Total6
Electrostatic
Precipitator
0.82
0.10
0.03
0.05
1.00
0.04
0.50
0.08
0.01
0.02
0.02
b
b
1.67
0.20
0.20
0.10
0.01
0.01
0.02
0.03
2.24
Wet
Scrubber
0.82
0.10
0.03
0.05
1.00
0.06
0.40
0.01
0.05
0.03
0.01
b
b
1.56
0.10
0.10
0.10
0.01
0.01
0
0.03
1.91
Fabric
filter
0.82
0.10
0.03
0.05
1.00
0.04
0.50
0.08
0.01
0.07
0.02
b
b
1.72
0.10
0.20
0.10
0.01
0.01
0
0.03
2.17
Incinerator
0.82
0.10
0.03
0.05
1.00
0.08
0.14
0.04
0.02
0.01
0.01
b
b
1.30
0.10
0.05
0.10
0,02
0.01
0
0.03
1.61
These average factors may require adjustments for individual estimates.
As required.
The relative costs for items 2(g) and 2(h) must be added to these average
totals.
7-6
-------
Table 7-2. COST ADJUSTMENT FACTORS FOR EMISSION CONTROL SYSTEMS1
Adjustment factor
Cost
adjustment
factor
Instrumentation
1. Simple, continuous manually operated
2. Intermittent operation, modulating flow with emissions
monitoring instrumentation
3. Hazardous operation with explosive gases and safety backups
Freight
1. Major metropolitan areas in continental U.S.
2. Remote areas in continental U.S.
3. Alaska, Hawaii, and foreign
Handling and erection
1. Assembly included in delivered cost with supports, base,
skids included. Small to moderate size equipment
2. Equipment supplied in modules, compact area site with ducts
and piping less than 70 meters. Moderate-size system
3. Large system, scattered equipment with long runs. Equip-
ment requires fabrication at site with extensive welding
and erection
4. Retrofit of existing system; includes removal of existing
equipment and renovation of site. Moderate to large
system
Site preparation
1. Within battery limits of existing plant; includes minimum
effort to clear, grub, and level
2. Outside battery limits; extensive leveling and removal of
existing structures; includes land survey and study
3. Requires extensive excavation and land ballast and
leveling. May require dewatering and pilings
Facilities and buildings
1. Outdoor units, utilities at site
2. Outdoor units with some weather enclosures. Requires
utilities brought to site, access roads, fencing,
and minimum lighting
3. Requires building with heating and cooling, sanitation
facilities, with shops and office. May include rail-
road sidings, truck depot with parking area
Engineering and supervision
1. Small-capacity standard equipment, duplication of typical
system, turnkey quote
2. Custom equipment, automated controls
3. New process or prototype equipment, large system
Construction and field expenses
1. Small-capacity systems
2. Medium-capacity systems
3. Large-capacity systems
Construction fee
1. Turnkey project, erection and installation included in
equipment cost
2. Single contractor for total installation
3. Multiple contractors with A&E firm's supervision
Contingency
1. Firm process
2. Prototype or experimental process subject to change
3. Guarantee of efficiencies and operating specifications
0.5 to 1.0
1.0 to 1.5
0.2 to 1.0
1.5
2
0.2 to 0.5
I
1.0 to 1.5
0.5
1 to 2
3
0.5
1
1.5
0.5
1
2
3.3 to 5
5 to 10
7-7
-------
7.2.2 Annual 1zed Costs
Annualized costs are categorized in Table 7-3, which also gives example
cost factors. Direct operating costs are such that rates for labor, mate-
rial, and utilities can be applied to estimates of requirements for these
items. The rates can be obtained as average figures from sources such as
the Bureau of Labor Statistics and the Federal Energy Regulatory Commission,
or rates for specific areas can be obtained from actual consumers.
The requirements for operating labor and supervision depend on system
variables, such as degree of automation, and operational variables, such as
continuity of operation and number of shifts. Maintenance requirements
depend upon the nature of the gas stream; e.g., corrosiveriess or abrasive-
ness, construction materials, and system size and type. Labor and mainte-
nance costs can be estimated from information presented in Reference 1.
Utility requirements are derived by use of the following formulas:
Fans
kwh = 8.3 Q (Ap)(SG)(h)
10 n
where:
kWh = energy usage in kilowatt-hours
Q = actual volumetric flow rate, ms/s
Ap = pressure loss, Pascals
n, = efficiency, usually 60 to 70 percent
h = hours of operation
S6 = specific gravity as compared to air
Pumps
kwh = 9.80 Q(H)(SG)(h)
where:
7-8
-------
TABLE 7-3. EXAMPLE FACTORS FOR ANNUALIZED COSTS1
Direct operating costs
Cost factor
Operating labor
Operator
Supervisor
Operating materials
Maintenance
Labor
Material
Replacement parts
Utilities
Electricity
Fuel oil
Natural gas
Plant water
Water treatment and cooling water
Steam
Compressed air
Waste disposal
$7.87/man-hour
15% of operator
As required
$8.66/man-hour
100% of maintenance labor
As required
10.012/MJ
$125/m3
$0.071/m3
$0.67/m3
$0.27/m3
$11.7/Mg
$0.0007/m3
$6-12/Mg
Indirect operating costs
Overhead
Property tax
Insurance
Administration
Capital recovery cost
80% of operating labor and
maintenance labor
1% of capital costs
1% of capital costs
2% of capital costs
0.16275 (as an example of
10% and an equipment life
of 10 years)
Credits
Recovered product
As required
All costs are in December 1977 dollars.
7-9
-------
kWh = energy usage in kilowatt-hours
h = hours of operation
Q = flow rate, m3/s ;
H = head of fluid, m
SG = specific gravity
H = efficiency, usually 60 to 70 percent
Waste disposal costs do not always apply; when products are recovered,
such material constitutes a credit. ;
Indirect operating costs are based on both direct operating costs and
capital costs. Overhead costs are a straight percentage of wages and
salaries; they cover expenses for such items as fringe benefits and cafe-
terias.
Capital recovery costs are derived by use of the following equation:
Capital recovery cost = capital costs x ^* 2J— (gq. 7-4)
(1 •*- i)n- 1
where:
i = annual interest rate
n = capital recovery period, years
Specific information regarding overhead cost factors and equipment
lives can be obtained from Reference 1.
7.3 COST CURVES FOR VARIOUS PARTICULATE CONTROL SYSTEMS ;
This section presents cost data for several particulate control sys-
tems. Three cost curves are given for each type of control technique,
representing purchased equipment costs, total capital costs, and annualized
costs. All costs are estimated from information contained in Reference 1
1
and updated to January 1980 dollars. The accuracy of any curve relative to
a specific application depends upon the similarities between the assumptions
used in the example and the conditions under which the system will actually
7-10
-------
be used. The curves should not be relied upon, however, to provide better
than ±50 percent accuracy. For more precise estimates, it is recommended
that the reader apply the cost analysis method described earlier and
explained in detail in Reference 1.
7.3.1 Equipment Costs
Figure 7-1 through 7-5 show the estimated purchased costs, F.O.B.
factory, for five state-of-the-art particulate control device categories:
(1) electrostatic precipitators, (2) fabric filters, (3) mechanical collec-
tors (cyclones), (4) incinerators, and (5) venturi scrubbers. The curves
represent flange-to-flange costs and generally include internal electricals
and controls. Instrumentation is not included because it is usually pro-
vided as an optional feature. The cost curves are presented in terms of
dollars versus exhaust gas volume. This relationship is based on a number
of simplifying assumptions, which allow one to obtain quick, conceptual or
study estimates with a minimum of effort. It must be borne in mind that
these simplifications can lead to anomalous results at the extremes of the
ranges, since the curves are presented in the form y = ax and are based on
regression analyses.
7-3.2 Particulate Control System Costs
On the basis of information in Reference 1 and the equipment cost
curves in Figures 7-1 through 7-5, a number of cost curves have been devel-
oped that prouide conceptual or study estimates of the capital and annual-
ized costs of complete air pollution control systems. These curves provide
costs for grass-roots installations. A retrofitted installation generally
costs 10 to 30 percent more than a grass-roots installation and, depending
on specific difficulties at a given site, the costs can be calculated on the
basis of the latter percentage.
Annualized costs are based on 8700 h/y operation time. Since the
annualized costs vary with operating time, the annualized costs for opera-
tions of less than 8700 h/y will be lower than those shown in Figures 7-6
through 7-14. For example, the annualized costs for 2000 h/y operation as a
percent of the costs for 8700 h/y operation are approximately as follows:
7-11
-------
10,000
8000
6000
4000
to
CC
s
o
CO
en
I
oo
o
o
o
2000
1000
800
600
400
200
100
10
T I
SCA = nr/dOOO nTYmln)
n = COLLECTION EFFICIENCY
FOR DUST
HAVING
HIGH
RESISTIVITY
FOR DUST
HAVING
.MODERATE-
TO-LOW-
RESISTIVITY
I
J I
20
30 40 50
100
200 300 400
EXHAUST GAS RATE, m3/s
Figure 7-1. Cost of electrostatic precipitators; carbon steel
construction, thermally insulated, FOB factory.
(Instruments and controls and taxes not included.)
7-12
-------
6000
4000
2000
O
Q
O
CO
a:
<
CO
o
o .
o
1000
800
600
400
200
100
80
60
40
20
i i r
i i
TYPE OF CLEANING MECHANISM
I
AIR-TO-CLOTH RATIO (m/s)
CURVE 1 0.46 TO 1.0
CURVE 2 0.61 TO 1.0
CURVE 3 2.12 TO 1.0
BAG'MATERIAL
CURVE 1 NYLON
CURVE 2 NYLON
CURVE 3 NOMEX
J_
10
20
30 40 50
100
200 300 400
EXHAUST GAS RATE, nfVs
Figure 7-2. Cost of fabric filters, carbon steel construction, FOB factory,
(Instruments and controls and taxes not included.)
7-13
-------
100
80
60
i T
ac.
o
00
a\
DC.
Z
«c
CO
O
o
40
20
10
8
6
345 10
EXHAUST GAS RATE, m3/s
20
I L
30 40
Figure 7-3. Cost of mechanical collectors, carbon steel construction,
FOB factory. (Instruments and controls and taxes not
included.)
7-14
-------
1000
800
600
400
oo
o
o
o
00
cy>
>-
OS
<
CO
o
t/J
o
o
200
100
80
60
40
20
10
i—r
T T
CURVE 1 CATALYTIC UNIT, 351 HEAT RECOVERY
CURVE 2 CATALYTIC UNIT, NO HEAT RECOVERY
CURVE 3 THERMAL UNIT, 35% HEAT RECOVERY
CURVE 4 THERMAL UNIT, NO HEAT RECOVERY
l
_L
I
_L
2 34
EXHAUST GAS RATE,
10
20
30 40
m3/s
Figure 7-4.
Cost of incinerators, FOB factory. (Instruments
and controls and taxes not included.)
7-15
-------
1000
800
600
400
200
§
gg 100
1 80
o
"o 60
o
C_5
40
20
10
PRESSURE DROP, kPa
CURVE 1 15
CURVE 2 10
CURVE 3 5
10
Figure 7-5.
20 30 40 50
100
200 300 400
EXHAUST GAS RATE, mj/s
Cost of venturi scrubbers, unline'd throat, carbon steel
construction, FOB factory.
and taxes not included.)
(Instruments and controls
7-16
-------
1000
CAPITAL COST
ANNUALIZED COST
COST OF DUCT INCLUDES
ONE ELBOW
Figure 7-6.
EXHAUST GAS RATE, m /s
Capital and annualized costs of fans and 30.5 m length of duct.
7-17
-------
4000
2000
o
03
-------
8000
6000
4000
3000
2000
1000
o 800
o
i 600
>•
< 400
300
£ 200
o
o
100
80
60
40
20
NOTES A & C
NOTES B & C
NOTES A & D
INOTES B & D
n = COLLECTION EFFICIENCY
NOTE A = DUST HAVING HIGH RESISTIVITY
NOTE B = DUST HAVING MODERATE-TO-LOW RESISTIVITY
NOTE C = CAPITAL COSTS
NOTE D = ANNUALIZED COSTS
SCA = m2/(1000 mVmin)
_L
I
10
20
30 40 50
100
200 300 400
EXHAUST GAS RATE, itT/s
Figure 7-8. Capital and annualized costs of electrostatic precipitators,
carbon steel construction.
7-19
-------
6000
4000
2000
g
o
CO
en
Q£
-------
8000
6000 [~
4000
3000
2000
a:
a:
«c
1000
800
600
400
-------
1000
800
600
400
g
3 200
o
o
s
en
§ 100
1 80
o
"o 60
o 40
T T
20
10
I
I I I
CAPITAL COST
ANNUALIZED COST
I
1.0
2.0 3.0 4.0 5.0
10.0
20.0 30.0 40.0
EXHAUST GAS RATE, nT/s
Figure 7-11. Capital and annuallzed costs of mechanical collectors,
carbon steel construction.
7-22
-------
10,000
8000
6000
4000
DC
«c
§
en
a:
CO
o
o
o
2000
1000
800
600
400
200
100
80
60
40
I 1 T I
I I
ss/s
CAPITAL COST
ANNUALIZED COST
CURVE 1
CURVE 2
CURVE 3
CURVE 4
CATALYTIC UNIT, 351 HEAT RECOVERY"
CATALYTIC UNIT, NO HEAT RECOVERY
THERMAL UNIT, 35% HEAT RECOVERY '
THERMAL UNIT, NO HEAT RECOVERY
I
I
l
1.0
Figure 7-12.
2.0 3.0 4.0 5.0
10.0 20.0 30.0 40.0
3,
EXHAUST GAS RATE, nT/s
Capital and annualized costs of incinerators.
7-23
-------
6000
4000
2000
I I I
I T
1000
c/>
g 800
§ 600
o
CO
on
-------
8000
6000
4000
3000
2000
3 1000
^ 800
o
00
~ 600
1 400
300
CO
CD
200
100
80
60
40
30
1 T
Tl T
— CAPITAL COST
™-ANNUAL I ZED COST
PRESSURE DROP, kPa
CURVE 1 15
CURVE 2 10
CURVE 3 5
_L
10
Figure 7-14.
20 30 40 50
, 100
EXHAUST GAS RATE, m3/s
200 300 400
Capital and annualized costs of venturi scrubber,
stainless steel construction.
7-25
-------
Venturi scrubber 30 - 40 percent
Fabric filter 50 - 60 percent
Electrostatic precipitator 60 - 70 percent
Incinerator 25 - 35 percent
The annualized cost includes a disposal cost of $10 per ton, based on
disposal of a nontoxic substance. It also includes a capital charge based
on an assumed equipment life of 15 years and an opportunity cost of 15
percent.
Each of the control system cost curves includes the costs of auxiliary
equipment normally associated with such a system. In some instances one may
wish to know what the system would cost either with or without the ductwork,
fan, and fan drive. The capital and annual ized costs of the components are
shown in Figures 7-6 and 7-7.
7,3.2.1 Electrostatic Precipitator. Figure 7-8 presents cost curves
for systems utilizing an electrostatic precipitator housed in an insulated,
carbon steel shell. The assumption is made that the uncontrolled gas stream
is normally vented to a stack. Thus, the necessary fan and ductwork are
considered part of the process. Costs are presented for three levels of
control efficiency based on medium- and high-reactivity dust. For a given
collection efficiency, high-resistivity dust requires a greater SCA
(specific collection area) and the cost of the ESP is thus increased. For
purposes of estimating equipment costs, plate area was calculated according
to the Deutsch equation with particle drift velocities of 0.036 m/s for
high-resistivity dusts and 0.086 m/s for low-resistivity dusts. Dusts such
as fly ash from low-sulfur coal combustion and cement kiln dust have high
resistivity.
7.3.2.2 Fabric Filters. Fabric filters are commonly used across a
broad range of exhaust gas volumes. Low-temperature and low-volume exhaust
streams from conveyor transfer points are normally vented to a fabric
filter. On the other hand, high-temperature and high-volume exhausts from
electric arc furnaces are also often vented to a fabric filter. Figures 7-9
7-26
-------
and 7-10 present cost curves for a variety of fabric filter applications.
Costs are presented for filters utilizing each type of bag-cleaning mecha-
nism. The cost curves assume that the fan and drive are process equipment.
The control costs include tie-in ductwork, a dust handling conveyor, and a
dust storage bin. The costs of thermal insulation and heaters (necessary to
prevent condensation in some applications) are not reflected in the cost
curves. Separate curves are presented for stainless steel construction.
7.3.2.3 Mechanical Collectors. Capital and annualized cost curves for
mechanical collector systems are shown in Figure 7-11. System costs include
hooding to capture the exhaust at the emission point, ducting, a fan and
drive, and a dust storage bin. The system cost is based on carbon steel
construction. Collection efficiency for this type of system generally
ranges from 80 to 90 percent, depending on the particle size distribution
and inlet grain loading.
7.3.2.4 Incinerators. Incinerators are of two basic types, thermal3
and catalytic. Although thermal incinerators are less costly from a capital
cost standpoint, the fuel savings associated with catalytic units make them
attractive for compatible exhaust streams. Both types of units may recover
heat and thereby reduce the fuel requirements. The additional cost of the
heat exchangers must be compared with the fuel savings on a case-by-case
basis. Additionally, the use of catalytic incinerators for control of
particulate matter is limited to substances that will not blind or poison
the catalytic mesh. Figure 7-12 presents cost curves for both types of
units, based on an exhaust stream at 25 percent of the lower explosive limit
(LEL). The costs of units having a heat exchanger are based on a 35 percent
heat recovery rate. Exhaust streams that are amenable to incineration are
normally exhausted to the atmosphere. Thus for purposes of the cost curves
presented herein, the fan and drive are considered process equipment. The
cost curves include the cost of ductwork to tie the incinerator into the
process vent system.
7.3.2.5 Venturi Scrubbers. Venturi scrubber use ranges from control
of small process fugitive exhaust streams to control of high-volume point
Note: Direct-fired incinerators considered as thermal incinerators for
purposes of this analysis.
7-27
-------
sources such as basic oxygen furnaces. Figures 7-13 and 7-14 present cost
curves for a variety of pressure drops. The costs include a clarifier ,and
circulating pump for the scrubber liquor, a fan and drive, and ductwork
sufficient to tie the scrubber into the process exhaust stream.
7.4 COST OF FUGITIVE EMISSION CONTROL
There is no single method of estimating costs of control of fugitive
particulate emissions. Because of the great variety of sources and control
methods, cost estimation must be specific to the method of control.
Many industrial process fugitive emissions (IPFE) are controlled with
the same types of equipment used for controlling process emissions. The
three main approaches to control of IPFE are ventilation, wet suppression,
and optimization of operations. Ventilation makes use of hooding, ductwork,
enclosures, and control devices such as baghouses or ESP's, Use of similar
types of equipment, with the possible addition of some auxiliaries, permits
use of capital cost estimating techniques previously described and the cost
curves presented in Section 7.3. The annualized costs for most ventilation
control techniques may also be estimated from the curves in this section.
The costs of controlling IPFE with wet suppression techniques are not
amenable to the estimating procedures described in Section 7,2 because these
control techniques, applied to materials-handling operations such as con-
veyors and unloading stations, do not make use of the same types of equip-
ment used in controlling process emissions. Their cost depends on the
amount of material handled and the efficiency required.
Table 7-4 presents reported cost data for spray and charged fog systems
applied to rail car unloading and conveyor transfer stations. The foam-type
spray system at a single transfer point would have a total capital cost of
about $18,000; implementing such a system handling 2.0 Gg of material per
year at an integrated iron and steel plant has been estimated to cost
$240,000.2 Initial cost of the charged fog-type system at a single transfer
point would be about $15,000.3
Because most of the other techniques used in controlling fugitive
particulate emissions do not require large capital expenditures, annual
direct operating costs are presented in place of annualized costs. Indirect
annual costs are not considered.
7-28
-------
TABLE 7-4. TYPICAL COSTS OF WET SUPPRESSION
OF INDUSTRIAL PROCESS FUGITIVE PARTICULATE EMISSIONS
Rail car unloading station
(foam spray)
Rail car unloading station
(charged fog)
Conveyor transfer point
(foam spray)
Conveyor transfer point
(charged fog)
Estimated
control
efficiency,
%
80
NR
70-95
NR
Initial cost,
January 1980
dollars
37,000
128,000C
18,000
15,000d
Unit operating
cost, January
1980 dollars
NR
NR
0.022 to
0.055/Mg
material
treated
NR
NR = Not reported.
a Reference 2.
Reference 3.
Based on use of 16 large devices at $8,000 each.
Based on use of 3 small devices at $5,000 each.
7-29
-------
The cost of controlling fugitive dust emissions is generally dependent
I
upon the method of application, suppressant or stabilizer used, and desired
control efficiency. Wet suppression and stabilization are the two most
common approaches to dust control.
Table 7-5 presents capital and operating costs for wet suppression of
I, „ . . »•
fugitive dust from an unpaved road, a storage pile, and a disturbed or
unvegetated outdoor (exposed) area. The costs for the unpaved road are
based upon the use of a 11.35-m3-capacity, nonpressurized spray truck oper-
ating twice daily. The control costs for the storage pile are based upon
the use of a stationary, elevated water spray system including sprayers,
piping, pumping, wind instruments, and installation costs. Operating costs
are not presented. For the exposed area, costs are based on the following
assumptions: the equipment includes piping and sprinklers, the unit is
moved by hand, and the application rate is 0.473 m3 of water per minute with
an effective spray radius of 33.5 meters.2
Table 7-6 presents cost data for application of stabilization tech-
niques to var.ious fugitive dust sources. The three sources are again an
unpaved road, a storage pile, and an exposed area. Methods of stabilization
are oiling, chemical and vegetative stabilization, paving, and use of
aggregate or chips.2
Operating costs for oiling and chemical stabilization of the unpaved
road surface are based on use of a 11.35-m3-capacity, nonpressurized spray
truck for application. The capital costs for chemical stabilization of a
storage pile are assumed to be the same as for wet suppression.2 The costs
of vegetative stabilization of exposed areas are highly variable as a result
of the wide range of climate and the physical and chemical properties of the
soil. Where topsoil is required, the costs may be higher than those shown
in Table 7-6.
Other techniques available for control of fugitive dust include street
sweeping, vacuuming, and flushing. Table 7-7 presents costs for these
control techniques. The lower capital cost figure for sweeping is based on
the use of a trailer-type sweeper; the higher capital cost figure is based
on use of a self-propelled unit with a spray bar. The initial cost figure
for flushing applies to a 1.135-m3 street flusher, excluding truck chassis.
Water requirements are significant.5
7-30
-------
TABLE 7-5. COST ESTIMATES FOR WET SUPPRESSION OF FUGITIVE DUST2
Source method
Unpaved road- regular
watering
Storage pile-regular
watering
Exposed area-watering
> Estimated
control
efficiency, %
50
80
50
Initial
capital cost,
January 1980
dollars
13,000/truck
14,000/system
2,000/hectare
Annual
operating cost,
January 1980
dollars
25,000/truck
NR
10-27 hectare
NR = Not reported.
7-31
-------
TABLE 7-6. COST ESTIMATES FOR STABILIZATION OF FUGITIVE DUST
Source method
Unpaved road-oiling3
Unpaved road-chemicals
(lignin or coherex)
Unpaved road- asphal tic
paving
Unpaved road-oil (double
chip surface)
Exposed areas-oiling3
Exposed areas-chemicals
Exposed areas-paving,
asphalt
Storage piles-surface
crusting chemicals
Exposed areas- vegetation6
Estimated
control
jfficiency, %
75
90-95
90
80
80
70
95
99
25-100
Initial
capital cost,
January 1980
dollars
1860/km
3700-9300/km
21,100-37,900/km
6700/km
250/hectare
2000/hectare
29,500/hectare
14,000/system
370- 1480/hec tare
Annual
operating cost,
January 1980
dollars
22,360/kmb .
3700-9300/km
3100-5600/km
1240-3100/kmd
NR '
75-150/hectare
NR
0.05-0.10/m2
NRf
NR = Not reported.
. Reference 2.
Based on monthly application.
. Based on resurfacing every five years and 15% opportunity costs.
Based on reapplication every 2-3 years and 15% opportunity costs.
_ Reference 4.
Dependent on type of vegetation planted, condition of soil, and climate.
7-32
-------
TABLE 7-7. COST ESTIMATES FOR SWEEPING AND FLUSHING OF
FUGITIVE DUST SOURCES2
Source method
Paved road- sweeping
Paved road-vacuuming
Paved road-flushing
Estimated
control
efficiency, %
70
75
80
Initial '
capital cost,
January 1980
dollars
5,000-15,0007
truck
28,000/truck
14,000/truck
Annual
operating cost,
January 1980
dollars
21,000/truck
26,000/truck
21,000/truck
Cost per kilometer depends on nature of process and the site.
7-33
-------
REFERENCES
1. Neveril, R. B. Capital and Operating Costs of Selected Air Pollution
Control Systems. EPA-450/5-80-002, December 1978.
2. Bohn, R., T. Cuscino, Jr., and C. Cowherd, Jr. Fugitive Emissions from
Integrated Iron and Steel Plants. EPA-600/2-78-050. March 1978.
3. Daugherty, D. P. and D. W. Coy. Assessment of the Use of Fugitive
Emission Control Devices. EPA-600/7-79-045. February 1979.
4. Richard, G., and D. Safriet. Guideline for Development of Control
Strategies in Areas with Fugitive Dust Problems. EPA-450/2-77-Q29.
October 1977.
5. PEDCo Environmental, Inc. Technical Guidance for Control of Industrial
Process Fugitive Particulate Emissions. EPA-4SO/3-77-Q1Q. March 1977.
7-34
-------
SECTION 8
EMERGING TECHNOLOGIES
With the growing awareness of the environmental and health impacts of
participate emissions, there has been a need for more effective control
technology. The emergence of new energy production methods also dictates
that participate control devices operate under more stringent conditions of
temperature, pressure, and flue gas properties. For these reasons there is
an increasing demand for advanced particulate removal systems to supplement
or replace the conventional control techniques.
The advanced particulate control systems have been developed only
recently. Most emerging techniques are essentially hybrids of technological
elements associated with first-generation systems. This section examines
some of the novel concepts being developed for control of particulate emis-
sions. Discussion of each concept is followed by a proposed overall design,
together with data that indicate system performance.
8.1 ADVANCED SCRUBBING TECHNIQUES
The prevailing limitation of conventional scrubbing technology is the
high energy usage required to capture submicron particulate and the asso-
ciated high operating costs. The limitations are principally due to the
inefficient use of available energy and the laws governing particle behav-
ior. In use of conventional scrubbing to capture the entrained particles
(which generally comprise less than 1 percent of the stream's mass), energy
is applied to the complete mass of the stream (gas molecules and particles).
Furthermore, exponentially increasing energy consumption is required in
conventional scrubbing to capture progressively smaller submicrometer-size
material. The combination of these two factors can render conventional
particulate scrubbing noncompetitive with other conventional removal tech-
niques.
8-1
-------
Many industrial processes emit gaseous and particulate-laden streams
containing corrosive, sticky materials. For these processes conventional
nonscrubbing techniques (i.e., baghouses, electrostatic precipitators, and
mechanical collectors) are not applicable for control purposes. Conse-
quently, environmental concerns with increasing energy costs combine to
encourage development of new scrubbing techniques.
The following subsections describe the technical and performance char-
acteristics of several emerging scrubbing techniques. Scrubbing systems
that are electrostatically and hydrodynamically enhanced are shown to be
more energy-efficient and effective in capture of submicrometer-size mate-
rial than their conventional counterparts.
8,1.1 AirPollution Systems, Inc. Electrostatic Scrubber (ES)
The ES system is basically an electrostatic charger followed by a con-
ventional venturi scrubber (Figure 8-1).1 The system incorporates a
patented high-intensity ionizer, which electrostatically charges the par-
ticles in the gas stream before they enter a conventional low-energy scrub-
ber. The ionizer, designed with a unique electrode configuration, produces
a high electric field strength (10 to 15 kV/cm) with high ion densities.
s
These levels of field strength and ion density, higher than those of the
conventional ESP, effectively charge both submicrometer- and supermicro-
meter-sized particles. The attractive force between the charged particles
and droplet is additive to the inertia! and other forces acting in the
venturi scrubber. These electrostatic forces account for the enhanced
collection of the complete size range of available particles in the gas
stream. Figure 8-2 illustrates the higher removal efficiency of fine parti-
culate with the electrostatic scrubber relative to the efficiency pf_a_
venturi scrubber.*
A field pilot test program with the UC electrostatic scrubber was con-
ducted on a urea prilling tower. The emission stream contained 85 percent
submicron material at a concentration of 91.5 mg/m3. At a pressure drop of
19.0 cm W.C. (7.5 in. W.C.), performance measurements by plant personnel
indicated an average overall efficiency of 93.5 percent and zero opacity.
8-2
-------
INSULATOR COMPARTMENT
VENT AND HEAT SYSTEM
INSULATOR
COMPARTMENT
TRANSFORMER
RECTIFIER
CATHODE.SUPPORT
HIGH-INTENSITY
IONIZER CATHODE
ANODE
WATER INLET
VARIABLE-THROAT -
VENTURI SCRUBBER
FLOODED ELBOW
DRAIN
Figure 8-1. APS electrostatic scrubber.1.
8-3
-------
1.0
0.5
0.2
0.1
0.05
0.02
0.01
0.005
0.002
0.001
MEASURED RESULTS
•EXTRAPOLATION
LOW~ENERGY
VENTURI SCRUBBER
ELECTROSTATIC
SCRUBBER
20
oc
0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0
AERODYNAMIC PARTICLE DIAMETER, ymA
— 90
99.5
J99.9
5.0 10
Figure 8-2. Fraction efficiency performance of
APS electrostatic scrubber.1
8-4
-------
Estimates of electrical operating power based on laboratory and pilot test
results, indicate 0.24 kW per ms/s to convert a low-energy venturi scrubber
into the equivalent of a high-energy scrubber with a UC ionizer module.1
8.1.2 TRW Charged Droplet Scrubber (CDS)
The CDS system passes water through small-diameter tubes, and electro-
statically atomizes and charges the water droplets.2 The droplets range in
diameter from 60 to 250. jjmA and have a high surface charge density. The
charged droplets are immediately exposed to and interact with slow-moving
(~2 m/s) dust-laden gas streams (Figure 8-3). The electric field in the
droplet-particle mixing region accelerates the charged droplets to high
velocities (-^30 m/s). These conditions of relatively slow-moving particles
and fast-moving charged droplets account for high collision rates due to
enhanced inertial and electrostatic collection mechanisms.
Successful results of laboratory and pilot field tests with the CDS
system were followed by installation of a 51,000 mVh demonstration unit on
a coke oven battery.3 Emissions consisted of fluctuating concentrations
(114 to 755 mg/in3) of submicron sticky hydrocarbons and micron-sized high-
conductivity carbon black. Overall removal efficiencies ranged from 91.0
percent to 94.3 percent, and fractional efficiencies ranged from 80 percent
to 99 percent (Figure 8-4) for various inlet loading and CDS operating
conditions. The CDS design summary is shown in Table 8-1.
3
Low total energy and water consumption, 1.41 to 2.0 W per m /s and 0.11
to 0.13 liter/m3, respectively, were demonstrated over most of the test
conditions. Capital and annualized costs are not available in the litera-
ture, but several CDS systems have been purchased and installed on indus-
trial and municipal incinerators, and in the iron and steel and the pulp and
paper industries in Japan.
8.1.3 University of Washington Electrostatic Droplet Scrubber (UWEDS)
The UWEDS system involves the use of electrostatically charged water
droplets to capture suspended particles electrostatically charged to the
opposite polarity of the droplets. The particles are negatively charged in
the corona section and flow into a scrubber chamber, into which positively
charged water droplets are sprayed. The scrubbed gas stream with entrained
8-5
-------
HIGH VOLTAGE
ISOLATION TUBING
COUICTOR PLMI ,
HfD WATil INLET
DC POWfH SUPPLY
WAHS/DUST
SLURRY
CARKY-OPf
SCRUBBING WATER
SLURRY DISCHARGE
TO SETUING POND
Figure 8-3, TRW charged droplet scrubber.'
8-6
-------
0,002
0.001
MEASURED RESULTS
EXTRAPOLATION
0.01 0.02 0.05 0.1 0.2 0,5 1.0 2.0
AERODYNAMIC PARTICLE DIAMETER, ymA
5.0 10
Figure 8-4. TRW charged droplet scrubber fractional
eff1ci ency performance.3
8-7
-------
TABLE 8-1. CDS DESIGN SUMMARY3
0 Three high-voltage scrubbing stages with 0.127-m collector plate spacing.
i
0 Flow cross-sectional area, 7.36 m2.
0 High-voltage electrode, type 316 stainless steel tubing, 19 mm diameter
flattened to 12.7 mm.
0 High-voltage electrodes contained 67 spray tubes each on 44.5-mm centers.
0 Spray tubes, titanium with 1.27-mm O.D. by 0.15-mm wall, protruding
25.4 mm from the electrode.
0 Collector plates 3.05 m long, 1.83 m high, 2.0 mm thick, mild steel.
0 Wall wash system covering each collecting surface.
droplets then flows into a mist eliminator consisting of a positively
charged corona section that removes the entrained droplets.4 The UWEDS
system is shown schematically in Figure 8-5.s
Pilot plant studies were conducted with a UWEDS mobile unit on an elec-
tric arc steel furnace and a coal-fired power plant. Overall collection
efficiencies ranged from 79.7 to 99.6 percent on steel furnace emissions and
from 99.6 to 99.99 percent on power plant emissions under various source iand^
control device operating conditions. Fractional efficiency ranged from 90
to 99.99 percent for fly ash particle sizes of 0.3 to 10 urn (Figure 8-6).6
Further results from the power plant tests indicate that the UWEDS system
operates with a specific collection area from 9.8 to 13.3 m2 per m3/s, and
with water consumption from 2.0 to 2.1 liters/m3. Total power consumption
is estimated at 0.8 kW per m3/s.5 No data have been published on annualized
costs for a full-scale system.
8.1.4 Steam Hydro Scrubber (SHS)
The Lone Star SHS system uses high pressure steam to move the gas
through the system, and to enhance particle collection with flux/force
condensation (F/C). (F/C effects are discussed in Section 8.1.6 .) The
fast-moving steam entrains gas, particles, and water droplets within the
mixing tube, where particle/droplet collision occurs. The mixing tube
8-8 '
-------
CHARGED
SPRAY
CHARGED
SPRAY
MIST
ELIMINATOR
SAS
OUTLET
WATER
OUTLET
WATER
OUTLET
AEROSOL
GENERATOR
BLOWER
AEROSOL
CHAMBER
CORONA
CHARGER
Figure 8-5. University of Washington electrostatic droplet
scrubber schematic.5
8-9
-------
MEASURED RESULTS
EXTRAPOLATION
0.0002
0.0001
0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0
AERODYNAMIC PARTICLE DIAMETER, umA
5.0 10
Figure 8-6. University of Washington electrostatic-droplet
scrubber fractional efficiency performance.6
8-10
-------
promotes particle/droplet collision through inertial and F/C collection
mechanisms. A shock-wave pattern is created in the mixing tube, further
enhancing particle/droplet collision by the induced turbulence. Upon exit-
ing the mixing tube, the stream is accelerated to achieve more complete
separation of the entrained materials in the cyclones. Centrifugal mecha-
nisms in the low-pressure-drop cyclones account for the impingement and
separation of the aerosols entrained from the gas stream. The SHS system is
depicted in Figure 8-7.7
A performance evaluation was conducted on sa commercial SHS system of
6.1 m3/s capacity controlling an open-hearth furnace.7 Overall collection
efficiency was measured from 99.78 to 99.95 percent under different source
and control device conditions. Fractional efficiencies ranged from 70 to
99.99 percent for particle sizes from 0.02 to 10 urn under different condi-
tions (see Figure 8-8). Estimates of total power consumption indicate 395
kW per m3/s if waste heat is not available, and M3 kW per m3/s if waste heat
is available for use as steam,5
8-1.5 Two-Phase Jet Scrubber (TPJS)
The Aeronetics TPJS system uses a nozzle designed to produce a two-
phase mixture of vapor and liquid droplets when fed pressurized, heated
liquid. The droplets are initially accelerated by the pressurized delivery
and become further accelerated by expansion due to evaporation of the hot
(200°C) water. Calculations by the developers indicate that the droplets
attain supersonic velocities (~300 m/s). The intense atomization and rela-
tively high droplet velocity produce a high probability of collision with
entrained particles. With typical droplet diameters (<100 urn) at the indi-
cated supersonic velocity, the impaction mechanisms would be effective for
particles as small as 0.2 umA diameter.5 An additional benefit of the TPJS
system is an induced draft, which eliminates or minimizes fan power require-
ments. Figure 8-9 is a schematic showing two options of the TPJS system.8
Performance evaluation of a commercial TPJS system was conducted at a
7.5-MW submerged arc ferro-alloy furnace.8 Average overall collection
efficiency was 95.9 percent under typical furnace and scrubber operating
conditions. Fractional efficiencies covered a wide range: 30 percent re-
moval of particles in the 0.03 to 0.10 urn size range and ~99 percent for
8-11
-------
OUTLET SAMPLING
LOCATIONS
INJECTION WATER
STEAM NOZZLE
INLET
ATOMIZER WATER
INLET SAMPLING
LOCATIONS
FLUE GAS FROM WASTE
HEAT BOILER. FED BY
OPEN HEARTH FURNACE
ATOMIZER SLURRY
PARTICLE
ACCELERATOR
CYCLONES
CYCLONE
Figure 8-7. Lone Star Steel steam-hydro air cleaning schematic.7
8-12
-------
0.0002
0.0001
0.01 0.02 0.05 0,1 0.2 0.5 1.0 2.0
AERODYNAMIC PARTICLE DIAMETER, pmA
99.95
99.99
5.0 10
Figure 8-8. Lone Star Steel steam-hydro air cleaning
fractional efficiency performance.^
8-13
-------
HOT .
LIQUID »
IN
GAS AND
LIQUID OUT
OPTION 1
TWO-PHASE
XJET NOZZLE
HEAT EXCHANGER
HOT GAS
MIXING SECTION
SEPARATOR
MAKE-UP
WATER
PUMP
STACK
WASTEWATER
TREATMENT
OPTION 2
Figure 8-9. Aeronetics two-phase jet scrubber schematic.8
(2 options shown)
8-14
-------
those in the 0.5 to 10 umA range (see Figure 8-1Q).8 Estimates of total
power consumption are 41 kW per m3/s if waste heat is not available and 3.8
kW per ms/s if waste heat is available for use as steam.5
8.1.6 Flux Force/Condensation Scrubbing
Flux force/condensation (F/C) effects are those that accompany the
condensation of water vapor from a gas stream; they are generally caused by
contacting hot, humid gas with colder liquid and/or by injecting steam into
saturated gas. Flux force/condensation also capitalizes on the growth of
particle mass and size due to condensation of water vapor on suspended
particles. Particle growth facilitates the collection of particles by
inertia! impaction. In practical terms, F/C scrubbing takes advantage of
forces acting on the particles induced by a temperature gradient (thermo-
phoresis), a vapor condensation gradient (diffusiophoresis), and vapor
condensation (Stefan flow). This advanced scrubbing method is adaptable to
various scrubber configurations; the ability to enhance particle collection
has been demonstrated with venturi, sieve plate, packed bed, and mobile bed
designs.9'10 Moreover, as particle size decreases, the advantages of F/C
scrubbing over conventional systems becomes greater because F/C collection
efficiency is virtually independent of particle size.
A demonstration F/C scrubbing system was built consisted of a sprayr
type quencher, a sieve plate column, and a spray-type cooling tower with an
induced draft fan.11 The demonstration was performed on a secondary metal-
recovery furnace emitting particles with a mean aerodynamic diameter of 0.75
.umA. The 12,000-m3/h demonstration system collected 90 to 95 percent of the
submicron stream at a pressure drop of 68 cm W.C. Under these source condi-
tions, a conventional high-energy scrubber would require pressure drops of
approximately 250 cm W.C. for 90 percent collection efficiency and 535 cm
W.C. for 95 percent.
8.2 ADVANCED ELECTROSTATIC PRECIPITATION TECHNIQUES
Conventional electrostatic precipitators often cannot effectively treat
particulate materials that have high electrical resistivity (i.e., greater
than 5 x 1Q10 ohm-cm). The difficulty arises from the presence of highly
resistive material on the collection surfaces and from the current density
8-15
-------
MEASURED RESULTS
1.0
0.5
0.2
0.10
2 0.05
i
LU
£ 0.02
0.01
0.005
0.002
0.001
20
50
90
92 2
fe
95
99 5
o.
99.2
99.5
99.9
0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0
AERODYNAMIC PARTICLE DIAMETER, ymA
5.0 10
Figure 8-10. Aeronetics two-phase jet scrubber
fractional efficiency performance,.8
8-16
-------
levels typical of conventional precipitators. In the collected dust layer
the electric field strength, determined as the product of the resistivity
and current density, may exceed the dielectrical strength of the material
and cause electrical breakdown. This breakdown generally leads to condi-
tions that cause the production of positive ions from the collected mate-
rial. These ions have the reverse polarity of the ions discharged from the
corona wires, and consequently they neutralize and degrade the intended
charging process. This 'series of events stemming from the electrical break-
down is generally referred to as back corona or reverse ioni.zation.
The difficulty of precipitating high-resistivity materials limits the
performance of conventional precipitators. The following subsections dis-
cuss technologies that are being advanced as a result of and as a remedy for
this difficulty.
8.2.1 Pulse Energization
The collection of high-resistivity dusts by conventional electrostatic
precipitators can be substantially improved by pulse energization. The
theoretical basis of the pulse concept was established over 30 years ago; in
recent years, investigators in the United States, Europe, Australia, and
Japan have developed this concept and brought it to the marketplace.
Pulses of appropriate duration and frequency superimposed on the DC
voltage provide higher peak voltages, reduce sparkover, increase field and
diffusional particle charging, improve current distribution, and permit
independent control of secondary voltage and current.12 Resistivity-limited
dusts (e.g., those from combustion of some low-sulfur western coals) are
more easily precipitated with pulse energization than by conventional means
because higher and more uniform ion densities and field strengths prevail
when the electrical limit is reached.
Pulsed energization systems superimpose a high-voltage impulse of very
short duration and steep wave front on an underlying, relatively constant
potential. This steady base voltage is maintained at a reduced level to
sustain the migration of ions and particles toward the collecting plates.
For treatment of high-resistivity dusts, the base voltage may be set below
the normal corona initiation voltage. The high-voltage impulses momentarily
raise the actual potential well above the sparking or back-corona limit that
would be experienced with conventional energization.
8-17
-------
Pulse frequency is set in the range of 10 to 400 pulses per second
(pps). Higher secondary levels of peak voltage and current are achievable
with progressively higher pulse frequency. Figure 8-11 shows current-vol-
tage curves obtained with a special pulse-discharge electrode, illustrating
the achievement of higher corona points at increasing levels of pulse fre-
quency. Figure 8-12 shows current-voltage curves of conventional DC systems
and pulse energization systems under the same conditions. The conventional
current-voltage curve is steep, indicative of back corona, whereas the
corresponding curve with pulse energization reflects higher operating vol-
tage without back corona.
Limited but promising results on the performance and economics of pulse
energization are now available in the literature.13 1G Laboratory and pilot
studies have steered this emerging technology into successful, short-term
full-scale demonstrations. Full-scale applications on a 35 MW pulverized-
coal-fired utility boiler12 and a 290 ton/day rotary lime kiln14 have demon-
strated the performance and economic advantages of pulse energization in
collecting high resistivity dust (^5 x 1011 ohm-cm).
These two full-scale demonstrations were accomplished by retrofitting
conventional ESP's with pulse energization electrical systems. Comparative
tests on both emission sources were conducted with conventional DC power
supply and a pulse generator. Results of collection efficiency measurements
show a 1.3 to 1.5 factor of improvement by the pulse systems over the con-
ventional systems.12'14 The improvement factor was appropriately determined
by taking the ratio of the modified migration velocity values calculated
from the performance data for both systems, by use of the following modified
Deutsch-Anderson equation:
Pt = exp -(WR A/V)m (Eq. 8-1)
where
P. = penetration
A = collection area, m2
V = volumetric flow rate, m3/s
w. = modified migration velocity, cm/s
K , . . . . i
m = exponent, using a common value for m = 0.5.
8-18
-------
0.01
g§ 0.001
Z3
O
0.0001
PULSE
DISCHARGE
ELECTRODE
400 pps
200 pps
100 pps
50 pps
20 pps
10 pps
20 30 40 50
APPLIED VOLTAGE, Up kV
Figure 8-11. Pulse energization voltage-current relationships for
various pulse frequencies.12
LU
OC
0£.
O
DC SUPPLY
ULSE SUPPLY
0.01
0 20 40 60 80
APPLIED VOLTAGE, UDC+Up kV
Figure 8-12. Comparison of DC and pulse energization voltage-current
relationships with same discharge electrode.12
8-19
-------
Additional indicators of the enhanced performance associated with pulse
systems are a significant reduction in stack opacity and spark rate, and a
large increase in peak operating voltage, both relative to performance of
the conventional systems.
With respect to potential applicability of pulsed energization systems,
the following should be kept in mind: 1) the reported improvement factors
represent a substantial reduction in precipitator size, but the pulsed
systems require more sophisticated and expensive electrical hardware; 2) the
pulse energization system will consume more electrical energy unless the
design incorporates an energy-conservation system; 3) the reported results
are limited and may not be representative of all applications; and 4)
further development and long-term demonstrations are needed for this emerg-
ing control technology.
8.2.1.1 Japanese Pulse Charging Systems. Further development of the
pulse energization concept has occurred in Japan. A "Bias-Controlled Pulse
Charging System" with a three-electrode configuration has been successfully
demonstrated on large-scale ESP's.16 This three-electrode configuration and
bias-controlled pulse charging extend the applicable limit of precipitators
to collection of dusts with resistivity up to 1015 ohm-cm. This novel
system also provides advantages in system stability (reducing process varia-
tions and upsets) and flexibility over a wide range of conditions. The
three-electrode system has also been applied to a 240 ton/h boiler plant
emitting low-resistivity, finely dispursed carbon particles. The relia-
bility and versatility of the three-electrode bias-control led system have
been demonstrated since 1976 on a space charge limiting exhaust at about 97
percent efficiency. Another novel system was installed in early 1978 to
treat an 11,000 m3/min exhaust from an iron-ore sintering furnace. With
dust resistivities ranging from 1011 to 1013 ohm-cm, a quasi-pulse ESP
system successfully collects 99 percent of the particulate emissions.16
8.2.1.2 Other Japanese Developments. Another Japanese development in
ESP technology is the use of wide-electrode spacing design (50 to 60 cm).
This design requires less plate area and allows operation at voltages ap-
proaching 200 kV. Full-scale wide-spacing ESP units have been used to
effectively limit sinter plant emissions to less than 0.05 g/m3 at resis-
tivity levels in the range of 1011 to 1013 ohm-cm.17
8-20
-------
Roof-mounted precipitators have been successful in the economical and
space-effective control of blast furnace emissions from steel plants. The
design includes vertical flow to allow natural convection for gas draft,
collecting electrodes made from conducting plastic plates, and intermittent
water irrigation.18 Collection of materials with resistivities in the range
of 1011 to 1012 ohm-cm is maintained at 95 percent efficiencies, and oper-
ating costs are 14 percent of those of a baghouse treating the same
effluent.
8.2.2 Two-Stage ESP Precharging
One approach to the problem of precipitating high-resistivity materials
is to separate particle charging from particle collection in a staged man-
ner. A two-stage system is being developed by Southern Research Institute
(SoRI) through sponsorship by the U.S. EPA.19 The SoRI system incorporates
a precharger with a novel electrode configuration and energization technique
in the first stage (Figure 8-13), and a downstream collector with a novel
corona discharge geometry in the second stage.
The first stage is similar to a conventional wire-to-plate design, but
has an additional (screen) electrode to control back corona. The additional
electrode consists of an open screen plate located close (1.9 cm) and paral-
lel to the collection plate. Operation of the first stage is similar to
that of a conventional ESP section, with the additional electrode being
energized separately. The screen electrode is energized with the same
polarity, but at a reduced voltage relative to the conventional discharge
electrode. The downstream collector is also similar to a typical ESP,
except that an open-mesh wire screen is used as a corona discharge elec-
trode.19
In operation of the first stage, particle charging is effected in the
usual manner, positive ions are produced, and then they are captured on the
screen electrode before they drift into the active charging region. The
screen electrode does not effect particle charging, but simply collects the
positive ions being produced from the collected highly resistive material.
The biased negative potential of the screen electrode assures capture of the
positive ions, and with the openings in the screen electrodes allows passage
of the negatively charged ions and particles to the plate.
8-21
-------
PRECHARGER
HOUSING
CERAMIC HIGH-
VOLTAGE INSULATORS
RECHARGER
HOUSING (REF)
CORONA DISCHARGE ELECTRODE
SUPPORT AND BUS BAR
SCREEN ELECTRODE
SUPPORT AND BUS BAR
HOPPER
PASSIVE ELECTRODES
SUPPORTS
SCREEN ELECTRODES
Figure 8-13. Southern Research Institute precharger ESP
assembly drawing.19
8-22
-------
Operation of the second stage is relatively conventional. The open-
mesh wire screen electrode was selected on the basis of laboratory experi-
ments to produce low current density with high electric field strength
values. Since particle charging is essentially completed in the first
stage, current levels are reduced in the second stage to minimize back-
corona formation and reentrainment of collected material.
Laboratory studies and pilot field tests with this design have produced
results that correlate with theoretical predictions. A field test with a
0.47 m3/s pilot-scale device was conducted on a utility boiler burning
low-sulfur coal and producing fly ash with a resistivity level measured at
2.0 x 1011 ohm-cm at 135°C.19 Performance measurements on the pilot two-
stage system showed an averaged collection efficiency of 97.7 percent, with
specific collection area equal to 50.4 m2 per ms/s. In performance tests
with the precharger off, emissions were seven times greater than with the
precharger on. Parallel testing with a mobile, pilot system of conventional
ESP design operating under the same conditions indicated a collection effi-
ciency of 91.9 percent. The penetrating emissions from the pilot conven-
tional system were 3.5 times those from the two-stage system. Fractional
efficiency data indicate 70 to 98 percent collection of particles in the
size range of 0.02 to 10 jjmA (Figure 8-14).
The cost of fabricating such a two-stage system is estimated to provide
a savings of approximately 40 percent over costs of a conventional ESP for
control of high-resistivity material.19
8.2.3 Flue Gas Conditioning for ESP's
Flue gas conditioning involves the use of additive materials to control
particle resistivity for the purpose of improving precipitator perform-
ance.20 Conditioning agents may also enhance performance by altering non-
electrical characteristics of the flue stream (e.g., particle agglomeration
and/or adhesion). Conditioning can be considered as a control of particle
resistivity, and includes three general means for control: particulate
composition, flue gas composition, and temperature. The combination of
these three factors accounts primarily for the resistivity levels associated
with electrostatic precipitation. Secondary factors to be considered in
suspected resistivity-related precipitator problems are thickness of the
8-23
-------
MEASURED RESULTS
1.0
0.5
0.2
0.1
o 0.05
ti
| 0.02
0.01
0.005
0.002
0.001
0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0
AERODYNAMIC PARTICLE DIAMETER, pmA
20
50
90
92
95 3
UJ
Ou
99
99.2
99.5
99.9
a:
£
5.0 10
Figure 8-14. Sourthern Research Institute precharger ESP
fractional efficiency performance.'*
8-24
-------
collected dust layer, the influence of electric field strength, and dust
aging characteristics.
This section describes the more popular means of flue gas conditioning
as indicated in published case studies. Gas conditioning is not a new
concept, but is considered here as an emerging technology for two principal
reasons: (1) gas conditioning is a cost-effective and attractive option for
a growing number of new and retrofit ESP installations; and (2) the appli-
cability and effectiveness of a conditioning agent or other means of condi-
tioning can only be verified on an experimental, site-specific basis, and
not by predictive or referencing techniques.
8.2.3.1 S03 Conditioning. The ultimate use of gas-phase S03 in
streams from utility boilers or from other industrial processes fired with
low-sulfur coal is becoming a popular, nonproprietary form of gas condi-
tioning. Several commercial means of producing gaseous S03 include (1) use
and combustion of elemental sulfur with appropriate processing, (2) use of
S02 with catalysis to form S03, and (3) purchase and direct use of S03.
Capital investment, operating cost, system reliability, and safety con-
siderations are the main determinants in selecting an S03 conditioning
system. Injection of trace quantities of S03 (3 to 30 ppm) in the gas
stream is cost-attractive and effective in reducing resistivity by 1 to 2
orders of magnitude for ESP temperature conditions less than 200°C.
Results of an S03 conditioning study21 of an ESP-controlled utility
boiler burning 0.6 percent sulfur coal showed an improvement of overall
collection efficiency from 91.3 to 98.8 percent. Upon injection of 25 ppm
of S03 particulate emissions were reduced by a factor of 8.3, measured
resistivity dropped from 6 x 1012 to 4 x 1010 ohm-cm at 143°C, and emission
of H2S04 vapor (<1 ppm) was the same as without conditioning. This success-
ful S03 injection reduced particulate emissions in an amount equivalent to
expanding the existing ESP capability by a factor of about 3.
Another study22 on utility boiler burning low-sulfur coal also demon-
strates the effectiveness of S03 conditioning on precipitator performance.
In this installation S02 is catalytically converted to S03 and injected
downstream of the air preheater at a rate corresponding to 32 ppm of S03 in
the flue gas stream. Stack measurements with the controlled condensation
8-25
-------
method indicated S03 concentrations of 10.9 and 8.1 ppm at the ESP inlet and
outlet, respectively. Mass loading measurements indicated that S03 condi-
tioning reduced particulate emissions by a factor of 2.4, corresponding to
an increase in collection efficiency from 79.2 to 95.4 percent relative to
baseline conditions without S03 injection.
"n! -J
Other case studies of S03 conditioning illustrate the effectiveness 'and
limitations of this method. Many cost estimates for new utility installa-
tions include S03 conditioning as an alternative control method. At least
two U.S. vendors of ESP's have indicated that the combination of a cold-side
precipitator with S03 conditioning is the most cost-effective method for
control of boilers burning low-sulfur western coal.23 25 Cost analyses show
savings of 20 to 80 percent in annualized costs for gas conditioning with
conventional (cold) "precipitators, relative to the costs of nonconditioned
cold precipitators, hot precipitators, and baghouses,26
8.2.3.2 Gas Conditioning With Water-Soluble Alkali Compounds. Certain
water-soluble alkali salts have been shown to be effective conditioning
agents in certain industrial processes. Laboratory experiments and full-
scale demonstrations show a sensitive relationship between resistivity and
the content of water-soluble alkali compounds in the collected material.
The following paragraphs deal with case studies in which potassium sulfate,
sodium chloride, and sodium carbonate are used as conditioning agents.
Note, however, that the descriptions of these materials as conditioning
agents do not imply that they are universally applicable conditioning
agents, effective on any given resistivity-limiting process stream. Rather,
the discussion of these agents, which are natural, common, and/or inexpen-
sive, is meant to indicate the sensitive relationship between resistivity
and trace quantities of readily available materials, and the resultant
impact on precipitator performance.
Potassium sulfate (K2S04) was demonstrated as an effective, condition-
ing agent on an 800-ton/day cement kiln process in Brazil.27 The K2S04 was
mixed in water as a 5 percent solution, then injected, atomized, and evapo-
rated in the gas stream ahead of the precipitator. The K2S04 solution was
injected at a rate of 500 liters/h, a negligible (0.02 percent K20) amount
compared with the normal content of this constituent (0.4 percent K20) in
8-26
-------
the raw meal. Performance measurements showed an increase in overall col-
lection efficiency from 75.0 to 86.7 percent because of conditioning, corre-
sponding to a twofold reduction in particulate emissions. Resistivity
measurements indicate a reduction from 1013 to 1011 ohm-cm at 300°C, attri-
butable to K2S04 conditioning.
Another full-scale demonstration with K2S04 was conducted successfully
on a coal-fired lime kiln in South Africa.27 A separate and parallel demon-
stration was made with sodium chloride.27 Each conditioning additive was
put into solution in the cooling water injected into the kiln. The results
of adding these materials were an increase in water-soluble K20 from 0.02 to
0.25 percent, and an increase in water-soluble Na20 from 0.04 to 0.27 per-
cent, in the precipitated dusts. Each conditioning agent reduced particu-
late emissions by a factor of 4 and resistivity by 2 orders of magnitude.
Note that sodium chloride was an effective agent for this process, but may
not be applicable in other industries because of incompatibility with the
process, product, or materials of construction.
Sodium carbonate has been used as a conditioning agent for hot-side
ESP's on at least two boilers fired with low-sulfur coal.28 At one instal-
lation, a 15 percent solution of Na2C03 is used as the conditioning medium;
at another, the conditioning agent is commercial-grade soda ash, which is
pulverized and fed pneumatically into the process stream as a dry material.
Both sodium carbonate systems are used to condition process streams in the
temperature range of 370° to 400°C, which is the range of the ESP treatment.
Particulate emissions have reportedly been reduced by a factor of 20 by
conditioning with Na2C03 solution and by a factor of 9 with solid Na2C03.
8.2.4 Development of High Temperature/High Pressure (HTHP) Electrostatic
Precipitation
Concern over the feasibility of high temperature/high pressure (HTHP)
electrostatic precipitation arose as technological developments in advanced
energy-producing processes indicated the need for HTHP cleanup systems. A
recent laboratory study of the characteristics of electrostatic precipita-
tion at high temperatures and pressures has yielded substantial and encour-
aging conclusions.29
8-27
-------
Bench-scale experiments with a concentric wire-pipe ESP design were
conducted at temperatures up to 1100°C (2000°F) and pressures up to 3550 kPa
(515 psia) with both negative and positive polarity energization.29 The
experimental program included the use of three gas mixtures: dry air, a
simulated flue gas, and a substitute (noncombustible) fuel gas. The experi-
mental results indicate higher particle-collecting efficiencies under condi-
tions of high temperature and pressure than those usually achieved with
conventional precipitator design and conditions. Higher operating voltages
are obtainable and are stable over the broad temperature/pressure range of
practical interest. These higher voltages provide increased electric field
strengths and promote higher probability of particle collection.
Experimental results demonstrate that potentially higher breakdown
l| ' .."J.." ."''.I
voltages are achievable with the combination of high temperature with appro-
priately high pressure. Concerns about achieving stable corona are resolved
with the understanding of a "critical pressure" concept. The critical
pressure is the lowest level of elevated pressure at which corona initiation
and sparkover .voltage levels coincide. Increasing the temperature rais.es
the critical pressure for achieving stable corona and consequently broadens
the operating pressure range. The general rule is that high pressure should
accompany high temperature. Failure to0 operate at or above the critical
pressure level will result in sparkover or breakdown without the necessary
corona formation.29
The critical pressure concept is applicable to both positive and nega-
tive polarity systems. With positive discharge systems, this critical
phenomenon is distinct and reproducible. With negative discharge systems
the pressure limit is not so well-defined, but the systems generally respond
to application of the critical pressure concept. At high gas densities,
higher operating voltages are achievable with negative than with positive
29
discharge.
i
8.3 ADVANCED FILTRATION TECHNIQUES
Filtration technology is being advanced through incorporation of elec-
trostatic designs, new filtration media, and novel filtration concepts.
Research has shown that natural or induced electrostatic phenomena can play
8-28
-------
a critical role in the operating and collection performance of conventional
filtration systems. Current studies are aimed at clarifying the effects of
electrostatic charge and electric field on the collectibility and cleana-
bility of conventional filter media. The advent of new filter materials is
promoting the use of filtration technology to control process emissions to
which conventional filtration has been inapplicable. High-temperature
filtration, currently not economical above 300°C, is likely to become com-
mercially available for .elevated temperature control applications. Future
requirements for control of high temperature/high pressure process streams
are being evaluated with new filtration concepts and requisite new filtra-
tion media. Filters made from ceramic materials are currently being evalu-
ated and developed for such applications. The following subsections de-
scribe these emerging filtration technologies.
8.3.1 Electrostatically Augmented Fabric Filtration
In electrostatically augmented fabric filtration, particles are pre-
charged before crossing a fabric filter, which may or may not be electro-
statically charged. Figure 8-15 shows a schematic of the pilot unit built
by American Precisions Industries, Inc., called the Apitron. Air enters the
precipitator section from below, then passes upward through the tubes of a
set of parallel wire-pipe precipitators, in which the particles are charged
and most are precipitated. The air then continues into and through the
bags, where final filtration takes place. Bag cleaning is initiated by
pulse jet flow. Studies on silica dust and various other particulate emis-
sions30'31 indicate that the Apitron yields high collection efficiencies
(Figure 8-16) and that air/cloth ratios can be much higher than those in an
uncharged filter of similar size. For example, a fabric filter utilizing a
pulse jet cleaning mechanism might require an air/cloth ratio of 5 ms/m2-min
whereas an electrostatically augmented filter operates at an air/cloth ratio
of 14 m3/m2-min.36
8.3,2 Electrostatically Augmented Filtration Through Fiber Beds
The concept of electrostatically augmented fiber bed filtration was
studied by Battelle Northwest on aerosols (NH4C1, Na20, and MgO) having mass
mean diameters less than 1 pm.32 The freshly generated particles were first
8-29
-------
COMPRESSED AIR
MANIFOLD
CLEAN GAS
OUTLETS
BAGS
DIRTY GAS
INLET
JET PULSE NOZZLE
INSULATOR
TUBE SURFACE
COOLING WATER
MANIFOLD
CORONA WIRE
HOPPER
DUST DISCHARGE
WATER OUTLET
Figure 8-15. Apitron electrostatic-filter cutaway view.
35
8-30
-------
MEASURED RESULTS
0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0
AERODYNAMIC PARTICLE DIAMETER, umA
5.0 10
Figure 8-16. Apitron electrostatic filter fractional
efficiency performance.
3-31
-------
drawn through a corona charging section and then through the fibrous beds,
D
which were made of stainless steel, polypropylene, or Teflon and had a void
fraction of 0.96. At velocities up to 1 m/s through a bed 30 cm thick, the
collection efficiencies were greater than 95 percent and pressure drops were
less than 1 cm H20,
8.3.3 Granular Bed Filtration
The general term "granular bed filtration" describes any filtration
system including a bed of discrete granules or particles as the filtration
medium. To prevent the particulate matter from plugging the interstices
between granules and causing excessive pressure drop, the device must incor-
porate some means for periodic or continuous removal of particles from the
collecting surfaces. Fixed-bed filters are cleaned periodically, generally
by a backwash of air to blow the dust out of the bed. Moving-bed filters
are cleaned continuously by replenishing the dirty bed with new granules and
separating the dust and granules by vibration.
The primary mechanisms for particulate collection in a bed of granular
solids are inertia! impaction, flow line interception, diffusional collec-
tion, and gravity settling. Interception is the mechanism for particulate
collection by gas convection, but collection by this mechanism is negligible
with a clean bed. As particles are deposited in the interstices to form a
cake, interception becomes increasingly important as the bed porosity and
flow channels are reduced. When too much particulate matter is deposited,
the pressure drop becomes excessive and the bed must be cleaned.
A series of studies was performed on the pressure drops and particulate
collection efficiencies of l.l-um~diameter latex spheres at a superficial
gas velocity of 50 cm/s.33 The collection bed is made up of iron shot.
With a bed depth of 3.2 cm at a gas velocity of 50 cm/s, the collection
efficiencies are 22 and 53 percent for shot granules of 620 and 490 umA
diameter, respectively. Efficiency can be increased to nearly 90 percent as
the bed depth increases, although with a significant increase in pressure
drop. Performance data on removal of fine particles with granular bed
filters are lacking for industrial applications, especially at high temper-
atures and pressures. Some preliminary studies on a pilot-scale fluidized
bed combustion unit are inconclusive because of plugging of the filter and
8-32
-------
34
the retaining screens that hold the granular beds. Combustion Power
Company is just beginning a series of "cold" filtration experiments on
35
moving granular bed filters, and results are pending.
8.3.4 Barrier Filtration
Barrier filtration with fiber beds, woven fabrics, and porous materials
can be used efficiently to remove particulates from gas streams. At high
temperatures and pressures, however, special materials, usually ceramics,
must be used to construct the bed matrix. Ceramic fibers can be woven into
a fabric, packed into a mat, or made into sheets and then used to form
filter "bags" comparable to those made of conventional fabrics. Their
particulate removal characteristics are comparable to those of conventional
barrier filters except that the ceramic material can withstand higher temp-
eratures. Barrier filtration operates by three mechanisms of particulate
oc
removal: direct interception, diffusion, and inertia! impaction. Parti-
cle collection efficiency increases as a dust cake builds up on the filter.
Collection efficiency can be improved simply by making a filter bed thicker
37
(thus increasing the pressure drop) or by reducing fiber diameter. Reduc-
ing fiber diameter is a desirable approach because it reduces filter weight
and bed thickness.
Tests were performed with several ceramic media samples to determine
penetration levels of dioctylphthalate (OOP) particles at air velocities of
5.5 and 16.5 cm/s under ambient conditions. The results indicated highest
efficiency for paper media, intermediate for felts, and low for woven mate-
rials. Moreover, the efficiencies of many of the samples were higher for
filtration of OOP smoke than were those of standard industrial-grade fil-
ters. Collection efficiencies of the woven ceramic materials were poor,
probably because of their more open structure. Paper media in general had
poor mechanical strength and did not survive pulse cleaning.
Initial screening and experiments indicate that "blanket" ceramic fiber
materials (felts) consisting of small-diameter fibers (3 |jmA) are the most
promising because of their combination of good filtration performance and
$C
relatively high strength. A filter was made of a layer of Saffil alumina
blanket insulation material approximately 1 cm thick contained between
stainless steel (304) screens. This configuration was tested with flyash
8-33
-------
from a pilot scale fluidized bed combustor. Tests were made over a period
of 200 hours. High collection efficiencies (greater than 99 percent) were
maintained with an air/cloth ratio of 9 m3/m2-min.S7 Further studies are
needed to develop more efficient cleaning techniques, to maximize air/cloth
ratios, and to further demonstrate the effectiveness and durability of such
devices.
Porous ceramic filters are also under study.38 The most promising con-
figuration is a ceramic cross-flow monolith (ThermaComb) produced by the 3M
Co. This material is composed of alternate layers of corrugation separated
by thin filtering barriers. Limestone test dust with a mass median diameter
of 1.4 mm was used to test the filter, and dust loadings were maintained at
levels from 2 to 7 g/m3. At .linear velocities of 0.41 m/min, collection
efficiencies were greater than 99 percent over a range of temperatures from
ambient to 970°K. Thus, porous ceramic filters are viable as barrier fil-
ters, and their performance should be studied relative to that of ceramic
fiber filters.
8.4 HIGH-GRADIENT MAGNETIC SEPARATION
t —
Magnetic separation has long been recognized as a method of removing
magnetic materials from mixtures, as in the separation of ferrous minerals
from ores. Within the last decade, the development of high-gradient mag-
netic separation (HGMS) techniques has enabled the efficient separation of
submicron particles of weakly paramagnetic materials from liquid streams at
high process rates.39 Generalized theory indicates that this technique can
be extended to the removal of particles from a gas stream. The fundamental
concept of HGMS is the interaction of the small paramagnetic particles with
a ferromagnetic wire in a magnetic field of uniform background.40 The
ferromagnetic wire induces regions of highly nonuniform field intensity,
exerting a net force on the particles and causing them to migrate to the
surface of the wire, where they are retained. This process is analagous to
enhanced filtration under a magnetic field except that the wire matrix is
much more open. ,
In its most simple form, the HGMS system consists of a canister packed
with fibers of a ferromagnetic material (steel wool), subjected to a strong
8-34
-------
external magnetic field (Figure 8-17). The resultant strong magnetic forces
near the edges of the fibers provide very efficient collection of fine para-
magnetic particles. The particles are removed from the gaseous stream as
they pass through the canister. The fiber matrix eventually becomes fully
loaded and must be cleaned. The overall particle collection efficiency is
theoretically a function of the applied magnetic field, filter mesh param-
eters (fiber diameter and magnetization, packing density, and length of mesh
in the direction of flow), particle parameters (diameter and magnetic sus-
ceptibility), and fluid parameters (viscosity and superficial velocity).41
Most of the applications of HGMS on a commercial scale have been in the
kaolin clay industry, where it is used in the removal of weakly magnetic
color bodies less than 2 umA in diameter. Studies also have been made on
fluid particle systems such as industrial waste process water from steel
mills and electroplating operations, on nuclear reactor coolants, and on
oils and hydraulic fluids. Potential has been demonstrated for application
of HGMS to the desulfurization of coal.
Application of HGMS to particulate removal from gas streams has been
studied only recently. A study was made on the particulate control of emis-
sions from basic oxygen furnaces and electric arc furnaces used for steel-
making.42 The collection efficiency in the experimental-scale study is sum-
marized in Figure 8-18, which indicates that HGMS is effective in removing
submicrometer particles of relatively high magnetic susceptibility from
high-velocity gas streams. Several industrial processes in the iron and
steel industry and the ferroalloy industry produce particulate emissions of
sufficiently high magnetic susceptibility to make HGMS a potentially attrac-
tive control technique. The same investigators made a rough economic analy-
sis of HGMS relative to the ESP and wet scrubber, as shown in Table 8-2.
Although capital costs of HGMS are relatively high, it is a competitive
process overall for highly efficient (greater than 99.9%) particulate matter
removal.
More research is needed to define optimum conditions for operation of
HGMS, along with a better understanding of the fundamental collection mecha-
nisms. Both experimental and theoretical studies should be done for a
variety of applications. Magnetic design, matrix size, and configuration
should be specified for each application, as well as methods of cleaning the
8-35
-------
FIELD
P*RTiClE—USOEN
FLUID IN
CLEAN FLUID
OUT
Figure 8-17. High gradient magnetic separator
schematic representation.117
8-36
-------
MEASURED RESULTS
EXTRAPOLATION
1.0
0.5
0.2
0.1
2 0.05
1
i
£ 0.02
0.01
0.005
0.002
O.OOlL
I I I
Of
LU
o.
20
*•—"•.
H;
50 S
90 ~
z
o
i—i
h-
92 |
o
LU
95 ^
99
99.2
99.5
DC
«c
Q_
99,9
0.01 0.02 0.05 0.1 0.2 0.5 1.0 2.0
AERODYNAMIC PARTICLE DIAMETER, ymA
5.0 10
Figure 8-18. High-gradient magnetic separator
fractional efficiency performance.^
8-37
-------
TABLE 8-2. COMPARISON OF HIGH-GRADIENT MAGNETIC SEPARATOR
AND CONVENTIONAL TECHNOLOGY
Device
Collection efficiency, %
Flange-to-flange cost, $ m3/s
$ per cfm
Power requirement, kW per m3/s
hp per 1000 cfm
ESP
99.9
4042
1.91
3.2
2.0
Scrubber
99.9
2762
1.30
24.2
15.3
HGMS
99.9
6653
3.14
2.2
1.4
All estimates are referred to clean gas at 343°C, -1.5 kPa, 11 percent water
by volume. Source, Ref. 39. ,
matrix when it is loaded with particles. It is reemphasized that for HGMS
applications the particles must exhibit some paramagnetic behavior. As a
consequence, this procedure is not widely applicable to many industrial gas
streams. It is best suited for gas streams containing paramagnetic parti-
cles or particles that can be made paramagnetic by seeding with a magnetic
material. ;
8.5 AGGLOMERATION TECHNIQUES
Particles in the submicron range are especially difficult to control by
conventional methods. Techniques for increasing the size of submicron par-
ticles by agglomeration appear attractive because the larger particles could
then be removed more effectively by conventional methods. Four such agglom-
eration techniques (thermal, turbulent, magnetic, and sonic) have been pro-
posed. Of these, the sonic appears the most promising and the most advanced
in experimental and industrial demonstrations. Magnetic agglomeration tech-
niques offer potential to enhance collection, although application is limi-
ted to process emissions with ferromagnetic particles. Thermal and turbu-
lent agglomeration techniques have been evaluated,43*44 but are not discus-
sed because of their limited potential for future development.
8.5.1 Sonic Agglomeration
Sonic agglomeration techniques have been researched for many years, the
observation that particles behave differently under the influence "of sound
8-38
-------
waves being made in 1931. The mechanisms of agglomeration are complex and
not well understood; as many as nine possible mechanisms have been postu-
lated.
The acoustic field affects the hydrody nami c forces and vibrational
movements of the particles within the flow field and thus accelerates colli-
sions between particles. Reference 45 presents a thorough discussion of the
pathways involved. Some of the theoretical and experimental analyses allow
the following conclusions regarding sonic particle agglomeration:46
1. The optimum frequency for agglomerating fine particles (2 (jmA and
less) is approximately 10 kHz. Particles larger than 2 umA serve
as collecting centers, and the agglomeration rate is directly
proportional to the number of such collection centers.
2. Highly polydispersed aerosols are easier to agglomerate than mono-
dispersed aerosols (i.e., if particles are all of nearly the same
size, acoustic agglomeration is not effective).
3. Acoustic agglomeration rates vary in proportion to the square root
of the acoustic intensity.
4. Physical properties of aerosol particles have comparatively little
effect on acoustic agglomeration.
5. Residence times for significant agglomeration are about 5 to 10 s.
6. Sound intensities of about 160 dB are required for effective
industrial applications.
7. Water sprays can enhance sonic agglomeration.
Table 8-3 shows some results of industrial tests with sonic agglomera-
tion (used in conjunction with other mechanical collection devices).47
The requirement of 160 dB sound intensities for sonic agglomeration can
be translated into a power requirement of about 1.6 kW per ms/s, which is
equivalent to about 750 kW for a modern pulverized coal -firing utility
station (470 m3/s volume flowrate.) Also, the intense sound levels create
noise that must be muffled. Recent studies indicate possibilities of lower-
ing the sound intensity requirements.48 Analysis also suggests that
resonating chambers producing standing waves could reduce the energy re-
quirements.51 Thus, acoustic agglomeration is considered to be the most
promising of the agglomeration methods for treating industrial gases.
8-39
-------
Table 8-3. RESULTS OF INDUSTRIAL TESTS WITH SONIC AGGLOMERATION
Aerosol
Zinc oxide
sublimate
fron
roasting
zinc ore
Zinc oxide
sublimate
from
copper
smelting
Zinc oxide
sublimate
from brass
melting
Coke gas
-------
Table 8-3 (continued)
Aerosol
Carbide furnace
smoke
Carbide furance
smoke
Gas furnace
black
Aerosol and gas stream properties
Particle
radius,
umA
0.2-15,
predomi-
nant 0.5
The sane
0.03-0.07
i
Gas furnace
black
Aggregated
gas black
Atomized
carbon black
Hard coal
black
0.03-0.07
0.5-15
0.1-0.2
0.5-1.0
Concentration
by weight,
fl/»3
0.25-2.8
0.25-2.8
1.2-12.6
1.2-2.1
0.5-2.5
26
0.5.2.4
Temperature
°C
20
120
40
40
82
80-90
Volume
treated,
mVh
5000
500
1700-2000
1700-2000
600
45
90-100
Agglomeration chamber
Type and
dimensions,
m
Experimental ,
reverse flow
The same, with
water addition
(5 g/ffl3)
Experimental ,
direct flow
1.1 dia, x 6.6
The same, with
water addition
Experimental ,
reverse flow
with water
0.5 dia.
Experimental,
rising steam
-0.29 dia. x
1.9
Experimental ,
reverse flow,
0,2 dia. x
2.5
Type of
s i ren
Static
Static
Dynamic,
radial
The same
Dynamic,
axial
Static
with
pump-off
Dynamic,
axial
Frequency,
kc/s
7-10
10.5
4
2-4
3
4.6
3.6
Intensity
W/cm2
0.5-1.0
0.5-1.0
0.1
1.0
0,10-0.14
Length of
sonic
treatment
s
4-6
4-6
4.5
1.2
10
7
3-4
Collector system
Type
«u It icy-
clones {in
parallel
The same
Two cyclones
1.3 in dia.
(in series)
The same
)ne or four
cyclones
{in
parallel)
two cy-
clones
and a
glass
cloth fil-
ter (in
series)
Cyclone
0.15 dia.
Amount
removed
without
sound, X
11
40
8-32
68-72
(30)a
68-74
(81)"
Amount
removed
wi thout
sound, X
94
86
83-90
99
95
99-98 (97)*
87 (97)b
CD
I
(continued)
-------
Table 8-3 (continued)
Aerosol
Sulfuric acid
fog
Natural
sulfuric
acid fog
Dilute
sulfuric
acid fog
Aerosol and gas stream properties
Particle
radius,
u»A
0.5-5.0
0.25-2.5
2.5-50,
predomi-
nant 7.5
Concentration
by we ignt,
g/m3
5-40
1
0.5-1.2
Teaperature
°C
180
50
20
Volume
treated.
nrVh
1700
40,000
1800
Agglomeration chamber
Type and
dimensions,
n
The same, 0.6
dia. x 6
Industrial ,
composite
flow, 2.4 dia.
x 10.5 (2 sets)
Experimental ,
reverse flow,
0,64 dia. x 11
Type of
siren
The same
Dynamic,
radial
Dynamic,
aerial
F regency,
te/s
2.15
2.25
1-2
Intensity
H/cm2
0.1
0.1
0.1
Length of
sonic
treatment,
s
3
4
7
Collector system
Type
Multicy-
clones (in
parallel)
Two cy-
clones (in
parallel)
Four cy-
clones (in
parallel)
Amount
removed
without
sound, 9!
84
69-72
Amount
removed
without
sound, %
99.6-99.9
90
78-82
ro
Result without cloth filter, in parentheses,
Result from water addition, in parentheses.
-------
The emergence of new energy-producing technologies may also necessitate
the use of novel particulate separation methods such as sonic agglomeration.
Sonic agglomeration in conjunction with a mechanical separator (cyclone) may
be effective under these conditions, and studies are being performed.49
8.5.2 Magnetlc Agglomeration
Magnetic fields can alter the motion of particles suspended in a gas
stream, depending on the magnetic permeability of the particles.50 By
adjustment of the strength of the magnetic fields to accommodate the nature
of the particles and the flow field, particle agglomeration may be enhanced.
Magnetic agglomeration will work best with ferromagnetic particles, although
theoretically it should also be effective with charged particles or par-
ticles having permanent or induced dipole moments.
Applications to industrial gas cleaning are limited bec'ause long resi-
dence times and strong magnetic fields are needed even with ferromagnetic
materials, and most particulate emissions are not ferromagnetic. Direct
capture of submicrometer particles under high-gradient magnetic fiels may be
more promising; this concept is discussed in further detail in a preceding
section.
8-43
-------
REFERENCES
1. Kearns, M. T., et al. Union Carbide's High Intensity Ionizer Applied
to Enhance a VentuH Scrubber System. In: Symposium on the Transfer
and Utilization of Particulate Control Technology, Vol. 3, Denver,
Colorado, pp. 73-84. EPA-600/7-79-044c. February 1979.
i
2. Lear, C. W. Charged Droplet Scrubber for Fine Particle Control:
Laboratory Study. EPA-600/2-76-249a. September 1976.
3. Krieve, W. F., and J. M. Bell. Charged Droplet Scrubber for Fine
Particle Control: Pilot Demonstration. EPA-600/2-76-249b. September
1976.
i
4. Pilot, M. J. , and D. F. Meyer. University of Washington Electrostatic
Spray Scrubber Evaluation. EPA-600/2-76-100. April 1976.
5. Cooper, D. W., et al. Evaluation of Eight Novel Fine Particle Collec-
tion Devices. EPA-600/2-76-035. 1976.
6. Pilot, M. J. , and G. A. Raemhild. University of Washington Electro-
static Scrubber Tests at a Coal Fired Power Plant. EPA-600/7-78-177b.
December 1978.
7. McCain, J. D., and W. B. Smith. Lone Star Steel Steam-Hydro Air Clean-
ing System Evaluation. EPA-650/2-74-028. April 1974.
8. McCain, J. D. Evaluation of Aeronetics Two-Phase Jet Scrubber. EPA-
650/2-74-129. December 1974.
9. Calvert, S., et al. Feasibility of Flux Force/Condensation Scrubbing
for Fine Particulate Collection. EPA-650/2-73-036. October 1973.
10. Calvert, S., and S. Gandhi. Fine Particle Collection by a Flux Force/
Condensation Scrubber; Pilot Demonstration. EPA-600/2-77-238.
December 1977.
11. Yung, S. , C. R. Chmielewski, and S. Calvert. Mobile Bed Flux Force/
Condensation Scrubbers. EPA-600/7-79-071. February 1979.
12. Lausen, P., H. Henrikson, and H. H. Peterson. Energy Conserving Pulse
Energization of Precipitators. In: Second Symposium on the Transfer
and Utilization of Particulate Control Technology. (EPA publication
pending.) 1980.
13. Belco Pollution Control Corporation. Technical Description of the
Belco Pulsed Power Supply. CPA 22-69-143. March 18, 1970.
8-44
-------
14. Feldman, P.C., and H.I. Milde. Pulsed Energization for Enhanced Elec-
trostatic Precipitation in High-Resistivity Applications. In: Sympo-
sium on the Transfer and Utilization of Particulate Control Technology.
pp. 253-274. EPA-600/7-79-044a.
15. Penney, G.W., and P.C. Gelfand. The Trielectrode Electrostatic Precip-
itator for Collecting High-Resistivity Dusts. APCA Journal, Vol. 28,
No. 1. pp. 53-55. January 1978.
16. Masuda, S., et al. Bias-Controlled Pulse Charging System for Electro-
static Precipitator. pp. 241-251. EPA-600/7-79-044a.
17. Ito, R., and K. Takimota. Wide Spacing E. P. is Available in Cleaning
Exhaust Gases from Industrial Sources. pp. 297-305. EPA-6QO/7-79-
044a.
18. Ito, S., et al. Roof-Mounted Electrostatic Precipitator. pp. 485-495.
EPA-600/7-79-044a.
19. Pontuis, D.H., D.V. Buhs, and I.E. Sparks. Field Evaluation of a
Two-Stage ESP for High-Resistivity Dusts. (To be published in Staub
journal.) 1980.
20. White, H.J. Industrial Electrostatic Precipitation. Addison-Wesley
Publishing Company, Reading, Massachusetts. 1963.
21. Dismukes, E.B., and J. P, Gooch. Fly Ash Conditioning with Sulfur
Trioxide. EPA-600/2-77-242. December 1977.
22. Patterson, R. , et al. Flue Gas Conditioning Effects on Electrostatic
Precipitators. In: Symposium on the Transfer and Utilization of
Particulate Control Technology, Vol. I. Denver, Colorado, pp. 169-
177. PB 295, 226. EPA~600/7-79-044a. February 1979.
23. Bubem'ck, D.V. Economic Comparison of Selected Scenarios for Electro-
static Precipitators and Fabric Filters. In: Section 14 Power Genera-
tion .1. Emission Control. Presented at the 70th Annual Meeting of
APCA. Toronto, Ontario, Canada. June 1977.
24. Atkins, R.S., and D.V, Bubenick. Keeping Fly Ash Out of the Stack.
Environmental Science and Technology, p. 657. June 1978.
25. Harrison, M.E. Economic Evaluation of Precipitator and Baghouse for
Typical Power Plant Burning Low Sulfur Coal. Presented at the American
Power Conference. Chicago, Illinois. April 24-27, 1978.
26. Breisch, E.W. Method and Cost Analysis of Alternative Collectors for
Low Sulfur Coal Fly Ash. pp. 121-129. EPA-600/7-79-044a.
27. Petersen, H.H. Conditioning of Dust with Water-Soluble Alkali Com-
pounds, pp. 99-111. EPA-600/7-79-044a.
28. Lederman, P.B., et al. Chemical Conditioning of Fly Ash for Hot-Side
Precipitation, pp. 79-98. EPA-600/7-79-044a.
8-45
-------
29. Bush, J.R., P.C. Feldman, and M. Robinson. Development of a High
Temperature/High Pressure Electrostatic Precipitator. EPA-6QO/7-77-
132. November 1977.
30. Helfrich, D.J., and T. Arimai. Electrostatic Filtration and the Api-
tron/Design and Field Performance. In: NovelI Concepts, Methods and
Advaned Technology in Participate Gas Separation. (T. Arimai, ed.).
31. Harmon, D.L, Electrostatically Augmented Particulate Collection
Devices. In: Proceedings of the First Workshop of Particulate Con-
trol. Kernforschungsanlage Julich Gmbh. March 1978.
32. Reid, D.L., and L.M. Browne. Electrostatic Capture of Fine Particles
in Fiber Beds. EPA-600/12-76-132. May 1976.
33. Yung, S.C., et al. Granular Bed Filter for Particulate Collection at
High Temperature and Pressure. Air Pollution Technology, Inc. Pre-
sented at EPA/DOE Symposium on High Temperature, High Pressure Particu-
late Control. Washington, D.C. September 1977.
34. Hoke, R.C., and M.W. Gregory. Evaluation of a Granular Bed Filter for
Particulate Control in Fluidized Bed Combustion. Exxon Research and
Engineering Company. Presented at EPA/DOE Symposium on High Tempera-
ture, High Pressure Particulate Control. Washington, D.C. September
1977. ;
35. Wade, G.L. Performance and Modeling of Moving Granular Bed Filters.
Combustion Power Company, Inc. Presented at EPA/DOE Symposium on High
Temperature, High Pressure Particulate Control. Washington, D.C.
September 1977.
36. Calvert, S., and R. Parker. Effects of Temperature and Pressure on
Particle Collection Mechanisms: Theoretical Review. EPA-600/7-77-002.
January 1977.
37. Shack!eton, M.A. High Temperature, High Pressure Particulate Control
with Ceramic Bag Filters. EPA-600/7-78-194. October 1978.
38. Drehmel, D.C., and D.F. Cilikerti. High Temperature Fine Particle
Control Using Ceramic Filters. Presented at EPA/DOE Symposium on High
Temperature, High Pressure Particulate Control. Washington, D.C.
September 1977.
39. Gooding, C.H., T.W. Sigmon, and L.K. Monteith. Application of High
Gradient Magnetic Separation to Fine Particle Control. EPA-600/2-77-
230. 1977.
40. Cummings, D.L., et al. Capture of Small Paramagnetic Particles by
Magnetic Forces from Low Speed Fluid Flows. AICHE Journal. 22:(569).
1976. ~~
41. Luborsky, R.E., and B.J. Drummond. High Gradient Magnetic Separation:
Theory vs. Experiment. IEEE Transactions or Magnetics. 2:(1696).
1975.
8-46
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-450/3-81-005a
3, RECIPIENT'S ACCESSION NO,
4. TITLE AND SUBTITLE
Control Techniques for Participate Emissions
Stationary Sources - Volume 1
From
5. REPORT DATE
September 1932
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3173 (Task No. 12)
PEDCo Environmental.Inc.
12. SPONSORING AGENCY NAME AND ADDRESS
DAA for Air Quality Planning and Standards
Office of Air, Noise, and Radiation
U.S. Environmental Protection Agency
Research Triangle Park. North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONS
:OR!NG AGENCY CODE
EPA 200/04
15. SUPPLEMENTARY NOTES
This document is issued per the requirements of Section 108 of the Clean Air
Act Amendments of 1977.
16. ABSTRACT
Control Techniques for Particulate Emissionsfrom Stationary Sources
Volumes 1 and 2 present recent developments of control techniques which have
become available since preparation of an earlier document entitled Control
Techniques for ParticulateAir Pollutants (AP-51).
Volume 1 of this document presents available data on characterization;
sampling methods and analytical techniques for particulate emissions; oarticle
behavior and characteristics; types of participate control systems, their
operating principles, design, operation, and maintenance; costs and environmental
consideration of particulate control techniques; and emerging technologies for
particul ate removal systems. A major portion of Volume 1 presents information
to quantify particulate removal efficiencies by particulate size for the
differing types of particulate removal systems.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lOENTtFIERS/OPiN ENDED TERMS C. COSATI Field/Group.
Particulates
Control Techniques
Costs
Emission Control
Operation and Maintenance
Design
13B
-B. DISTRIBUTION STA~£MEN1
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
460
JO. SECURITY CLASS (Thispage)
Unclassified
22. PRICED
EPA Fe'm 2220-1 (Rev. 4—77) PREVIOUS ECITIOM is OBSOLETE
*U.S. CGYEIU0SENT PRICING OFFICE: 1982—539-061/3003
-------
-------
-------
United States Office of Air. Noise, and Radiation
Environmental Protection - ., Office of Air Quality Planning and. Standards
Agency Research Triangle Park NC 27711
Official Business Publication No. EPA-450 3-81 -005a •' On*,,** *™<
Penalty for Private Use Pels Kid
* • rees raiu
Environmental
Protection
Agency
EPA 335
H vow *ddi«ss ttineomct, pttiMCiungt on tht atova Ubal.
(Mr o'i, »«xJ return !o 0v« »bov !«»r*«
M you do not tJevit tocontinu* r*c«inog inn t»cnnic»l r«pcwl
»r,»i CHECK HERE 2! . l**r o'f l»6«l and i*!v
------- |