United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park, NC 27711
EPA-452/R-97-001
October 1998
Air_
STATIONARY SOURCE CONTROL
TECHNIQUES DOCUMENT FOR
FINE PARTICUL ATE MATTER
-------
Stationary Source Control Techniques Document
for Fine Particulate Matter
EPA CONTRACT NO. 68-D-98-026
WORK ASSIGNMENT NO. 0-08
Prepared for:
Mr. Kenneth Woodard
Integrated Policy and Strategies Group (MD-15)
Air Quality Strategies and Standards Division
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
October 1998
Submitted by:
EC/R Incorporated
Timberlyne Center
1129 Weaver Dairy Road
Chapel Hill, North Carolina 27514
U.S. Environmental Protection Agency
77 West Jackson Boulevard, 12th FtoOf
Chicago, IL 60604-3590
-------
Disclaimer
This report has been reviewed by the Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, and has been approved for publication. Mention of
trade names or commercial products is not intended to constitute endorsement or
recommendation for use.
Copies
Copies of this document are available through the Library Services Office (MD-35),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711; or from the
National Technical Information Service, 5285 Port Royal Road, Springfield, VA 22161 (for a
fee). This document can also be found on the Internet at the U.S. Environmental Protection
Agency website (http:\\www.epa.gov/ttn/oarpg).
11
-------
CONTENTS
TABLES ix
FIGURES xi
1 INTRODUCTION 1-1
1.1 PURPOSE OF THIS DOCUMENT 1-1
1.2 OTHER RESOURCES 1-1
1.3 ORGANIZATION 1-2
1.4 REFERENCES 1-3
2 BACKGROUND 2-1
2.1 TRENDS IN AMBIENT PARTICULATE MATTER
CONCENTRATIONS AND PARTICULATE MATTER EMISSIONS ... 2-1
2.2 PROJECTIONS FOR FUTURE CONTROL PROGRAMS AND
EMISSIONS 2-1
2.3 SOURCES OF PM10 AND PM25 EMISSIONS 2-3
2.3.1 Point Sources 2-3
2.3.2 Area Sources 2-6
2.4 REFERENCES 2-6
3 MEASUREMENT 3-1
3.1 LIST OF EPA PM MASS MEASUREMENT TEST METHODS 3-1
3.2 EPA STATIONARY (POINT) SOURCE PM MASS MEASUREMENT
TEST METHODS 3-3
3.2.1 EPA Test Method 5 for Total PM Mass 3-3
3.2.2 EPA Test Method 5 Variations: 5A - 5H 3-6
3.2.3 EPA Test Methods for PMi0 from Stationary Sources 3-7
3.2.3.1 Method 201: Determination of PM10 Emissions--
Exhaust Gas Recycle Procedure 3-8
3.2.3.2 Methods 201A: Determination of PM10
Emissions—Constant Sampling Rate Procedure .... 3-8
3.2.4 EPA Test Method 17: Determination of PM Emissions from
Stationary Sources— In-Stack Filtration Method 3-8
3.2.5 Method 202 for Condensible PM Measurement 3-9
3.2.6 EPA Test Method 9: Visual Determination of the
Opacity of Emissions from Stationary Sources, and
Alternate Method 1 for the Use of Remote Lidar 3-9
3.2.7 Performance Specifications for Continuous Emissions
Monitoring Systems (CEM) Used to Monitor Opacity .... 3-10
3.3 OTHER STATIONARY (POINT) SOURCE PM MASS
MEASUREMENT TEST METHODS 3-10
iii
-------
CONTENTS (continued)
3.4 FUGITIVE PM MEASUREMENT METHODS 3-11
3.5 PARTICLE SIZE ANALYSIS 3-14
3.5.1 Cascade Impactors 3-14
3.5.2 Sampling Cyclones 3-15
3.5.3 Real-Time Size Distribution Measurement 3-15
3.5.4 Size Distribution of Bulk Samples 3-16
3.6 SPECIATION 3-17
3.6.1 EPA Test Method 29 for Metals and PM 3-17
3.6.2 EPA Office of Solid Waste Test Method 0010 (SW-846)41 3-18
3.6.3 Spectrometry 3-19
3.6.3.1 Atomic Absorption Spectrometry 3-19
3.6.3.2 Optical Emission Spectrometry 3-19
3.6.3.3 Mass Spectrometry 3-19
3.6.3.4 Neutron Activation Analysis 3-20
3.6.3.5 X-Ray Fluorescence Spectrometry 3-20
3.6.4 Electrochemical 3-20
3.6.5 Chemical 3-21
3.7 REFERENCES FOR SECTION 3 3-21
4 FUEL SUBSTITUTION AND SOURCE REDUCTION APPROACHES FOR
PARTICIPATE MATTER 4-1
4.1 FUEL SUBSTITUTION 4-1
4.1.1 Applicability 4-1
4.1.2 Emission Reductions with Fuel Switching 4-2
4.1.3 Costs 4-5
4.1.4 Other Impacts 4-6
4.2 PROCESS MODIFICATION/OPTIMIZATION 4-7
4.3 REFERENCES FOR SECTION 4 4-9
5 EXHAUST GAS CLEANING SYSTEMS FOR STATIONARY SOURCES .... 5-1
5.1 PRETREATMENT 5.1-1
5.1.1 Precollection Devices 5.1-1
5.1.1.1 Settling Chambers 5.1-1
5.1.1.2 Elutriators 5.1-4
5.1.1.3 Momentum Separators 5.1-4
5.1.1.4 Mechanically-Aided Separators 5.1-4
5.1.1.5 Cyclones 5.1-8
5.1.2 Collection Efficiency of Precollectors 5.1-12
5.1.2.1 Gravity Settling 5.1-12
5.1.2.2 Momentum Separators 5.1-12
5.1.2.3 Mechanically-Aided Separators 5.1-12
iv
-------
CONTENTS (continued)
5.1.2.4 Cyclones 5.1-16
5.1.3 Applicability 5.1-20
5.1.4 Costs of Precollectors 5.1-22
5.1.4.1 Capital Costs of Cyclones 5.1-22
5.1.4.2 Annual Costs of Cyclones 5.1-24
5.1.5 Energy and Other Secondary Environmental Impacts of
Precollectors 5.1-27
5.1.6 Flue Gas Conditioning 5.1-27
5.1.6.1 Sulfur Trioxide Conditioning 5.1-28
5.1.6.2 Ammonia Conditioning 5.1-29
5.1.6.3 Ammonium Compound Conditioning 5.1-30
5.1.6.4 Organic Amine Conditioning 5.1-31
5.1.6.5 Alkali Conditioning 5.1-31
5.1.7 Costs of Flue Gas Conditioning 5.1-32
5.1.8 Energy and Other Secondary Environmental Impacts of Flue Gas
Conditioning 5.1-32
5.1.9 References for Section 5.1 5.1-34
5.2 ELECTROSTATIC PRECIPITATORS 5.2-1
5.2.1 Particle Collection 5.2-1
5.2.1.1 Electric Field 5.2-1
5.2.1.2 Corona Generation 5.2-3
5.2.1.3 Particle Charging 5.2-4
5.2.1.4 Particle Collection 5.2-4
5.2.2 Penetration Mechanisms 5.2-5
5.2.2.1 Back Corona 5.2-5
5.2.2.2 Dust Reentrainment 5.2-6
5.2.2.3 Dust Sneakage 5.2-6
5.2.3 Types of Electrostatic Precipitators 5.2-6
5.2.3.1 Dry ESPs 5.2-6
5.2.3.2 Wet ESPs 5.2-7
5.2.3.3 Wire-Plate ESPs 5.2-7
5.2.3.4 Wire-Pipe ESPs 5.2-9
5.2.3.5 Other ESP Designs 5.2-9
5.2.4 Collection Efficiency 5.2-14
5.2.5 Applicability 5.2-17
5.2.6 Costs of Electrostatic Precipitators 5.2-20
5.2.6.1 Capital Costs 5.2-20
5.2.6.2 Annual Costs 5.2-24
5.2.7 Energy and Other Secondary Environmental Impacts 5.2-27
5.2.8 References for Section 5.2 5.2-28
-------
CONTENTS (continued)
5.3 FABRIC FILTERS 5.3-1
5.3.1 Particle Collection and Penetration Mechanisms 5.3-1
5.3.2 Types of Fabric Filters 5.3.5
5.3.2.1 Shaker-Cleaned Fabric Filters 5.3-5
5.3.2.2 Reverse-Air Cleaned Fabric Filter 5.3-7
5.3.2.3 Pulse-Jet Cleaned Fabric Filter 5.3-10
5.3.2.4 Other Fabric Filter Designs 5.3-14
5.3.3 Fabric Characteristics 5.3-15
5.3.4 Collection Efficiency 5.3-17
5.3.5 Applicability 5.3-19
5.3.6 Costs of Fabric Filters 5.3-22
5.3.6.1 Capital Costs 5.3-22
5.3.6.2 Annual Costs 5.3-26
5.3.7 Energy and Other Secondary Environmental Impacts 5.3-30
5.3.8 References for Section 5.3 5.3-31
5.4 WET SCRUBBERS 5.4-1
5.4.1 Particle Collection and Penetration Mechanisms 5.4-1
5.4.2 Types of Wet Scrubbers 5.4-2
5.4.2.1 Spray Chambers 5.4-3
5.4.2.2 Packed-Bed Scrubbers 5.4-3
5.4.2.3 Impingement Plate Scrubbers 5.4-8
5.4.2.4 Mechanically-aided Scrubbers 5.4-8
5.4.2.5 Venturi Scrubbers 5.4-11
5.4.2.6 Orifice Scrubbers 5.4-11
5.4.2.7 Condensation Scrubbers 5.4-14
5.4.2.8 Charged Scrubbers 5.4-14
5.4.2.9 Fiber-Bed Scrubbers 5.4-18
5.4.3 Collection Efficiency 5.4-18
5.4.4 Applicability 5.4-22
5.4.5 Costs of PM Wet Scrubbers 5.4-25
5.4.5.1 Capital Costs 5.4-25
5.4.5.2 Annual Costs 5.4-28
5.4.6 Energy and Other Secondary Environmental Impacts 5.4-32
5.4.7 References for Section 5.4 5.4-35
5.5 INCINERATORS 5.5-1
5.5.1 Incinerator Control Mechanisms 5.5-1
5.5.2 Types of Incinerators 5.5-3
5.5.2.1 Thermal Incinerators 5.5-3
VI
-------
CONTENTS (continued)
5.5.2.1.1 Discrete Burner Thermal Incinerator 5.5-4
5.5.2.1.2 Distributed Burner Thermal Incinerator ... 5.5-4
5.5.2.2 Catalytic Incinerators 5.5-4
5.5.2.3 Heat Recovery Equipment 5.5-7
5.5.3 Control Efficiency 5.5-9
5.5.3.1 Control Efficiency for Volatile Organic
Compounds 5.5-9
5.5.3.2 Control Efficiency for Particulate Matter 5.5-9
5.5.4 Applicability 5.5-9
5.5.5 Costs of Incinerators 5.5-11
5.5.5.1 Capital Costs 5.5-12
5.5.5.2 Annual Costs 5.5-15
5.5.6 Energy and Other Secondary Environmental Impacts 5.5-19
5.5.7 References for Section 5.5 5.5-19
6 INDUSTRIAL FUGITIVE EMISSION CONTROLS 6-1
6.1 ENCLOSURES AND VENTILATION 6-1
6.1.1 Local Ventilation Systems 6-2
6.1.2 Building Enclosure/Evacuation 6-5
6.2 OPTIMIZATION OF EQUIPMENT AND OPERATION 6-5
6.2.1 Source Extent Reduction and Improvement 6-5
6.2.2 Process Optimization/Modification 6-6
6.2.3 Leak Prevention and Detection and Other Good O&M Practices . . 6-8
6.3 COSTS OF HOODS 6-8
6.4 FUGITIVE DUST CONTROL 6-9
6.5 REFERENCES FOR SECTION 6 6-10
7 EMERGING TECHNOLOGIES 7-1
7.1 EMERGING FABRIC FILTER TECHNOLOGIES 7-1
7.1.1 Ceramics: Ceramic Filter Elements and Ceramic Fiber
Enhancement 7-1
7.1.2 Fine 1.1 dtex Fibers 7-1
7.1.3 Electrostatically-Stimulated Fabric Filtration (ESFF) 7-4
7.2 EMERGING ESP TECHNOLOGIES 1A
7.2.1 Sonic Horn Rappers 7-4
7.2.3 Cold-Pipe ESP Precharger 7-4
7.2.2 Alternating Charging and Short ESP Collector Sections (SUPER ESP)7-5
7.2.4 Advanced Computer-Based ESP Control Systems 7-5
7.3 EMERGING COMBINATION DEVICES 7-5
7.4 EMERGING SCRUBBER TECHNOLOGIES 7-6
7.4.1 Annular Orifice Venturi Scrubber 7-6
vii
-------
CONTENTS (continued)
7.4.2 Waterweb Mesh Scrubber 7-6
7.5 EMERGING MECHANICAL COLLECTOR TECHNOLOGIES 7-7
7.6 EMERGING FUGITIVE DUST CONTROL TECHNOLOGIES 7-7
7.6.1 High-Voltage PM Ionizer 7-7
7.6.2 Dry Fog 7-8
7.7 EMERGING SIMULTANEOUS POLLUTION CONTROL
TECHNOLOGIES 7-8
7.7.1 SNRB (SOx-NOx-Rox Box) Catalytic Fabric Filter 7-8
7.7.2 Catalyst-Coated Fabric Filters 7-8
7.8 REFERENCES FOR SECTION 7 7-9
APPENDIX A: LIST OF RESOURCE DOCUMENTS FOR PM AND PM
PRECURSOR CONTROL A-l
APPENDIX B: VATAVUK AIR POLLUTION CONTROL COST INDEXES ... B-l
Vlll
-------
TABLES
Table 2-1 Summary of Trends in PM10 Emissions from 1987 through
1993 (Reference 2) 2-2
Table 2-2 Detailed Breakdown of PM-2.5 and PM-10 Emissions by
Source Category (Reference 4) 2-4
Table 3-1 EPA Test Methods for PM 3-2
Table 3-2 EPA Test Methods 1 through 4: General Stack Sampling Procedures .... 3-4
Table 4-1 Potential PM10 Emission Reductions with Fuel Switching
(References 1, 4, and 5) 4-3
Table 4-2 Potential PM2 5 Emission Reductions with Fuel Switching
(References 1, 4, and 5) 4-3
Table 4-3 Average Prices of Coal, Oil, and Natural Gas
(References 6, 7, and 8) 4-5
Table 4-4 Potential SOX Reductions with Fuel Switching 4-6
Table 5.1-1 Characteristics of Common Cyclones 5.1-11
Table 5.1-2 Annual Cost Parameters for Cyclones (Reference 9) 5.1-25
Table 5.1-3 Annual Cost Factors for Cyclones (Reference 20) 5.1-26
Table 5.1-4 Costs of Flue Gas Conditioning 5.1-33
Table 5.2-1 PM-10 and PM-2.5 Cumulative Collection Efficiencies for
ESPs at Coal Combustors, Primary Copper Operations, and
Iron and Steel Production Operations (Reference 11) 5.2-17
Table 5.2-2 Typical Industrial Applications of Electrostatic Precipitators
(References 2 and 12) 5.2-19
Table 5.2-3 Capital Cost Factors for Electrostatic Precipitators (Reference 10) .... 5.2-22
Table 5.2-4 Annual Cost Parameters for Electrostatic Precipitators
(Reference 14) 5.2-25
Table 5.2-5 Annual Cost Factors for Electrostatic Precipitators (Reference 14) .... 5.2-26
Table 5.3-1 Recommended Gas-to-Cloth Ratios (acfm/ft2) for Common Industrial
Applications of Fabric Filters (References 4 and 13) 5.3-6
Table 5.3-2 Temperature Ranges, and Physical and Chemical Resistances
of Common Industrial Fabrics (Reference 2) 5.3-16
Table 5.3-3 PM-10 and PM-2.5 Cumulative Collection Efficiencies for
Fabric Filters at Coal Combustors, Ferroalloy Electric
Arc Furnaces, and Iron and Steel Production Operations (Reference 15) 5.3-19
Table 5.3-4 Typical Cleaning Methods and Fabrics for Industrial Applications
of Fabric Filters (Reference 2) 5.3-20
Table 5.3-5 Capital Cost Factors for Fabric Filters (Reference 16) 5.3-23
Table 5.3-6 Annual Cost Parameters for Fabric Filters (Reference 17) 5.3-29
Table 5.3-7 Annual Cost Factors for Fabric Filters (Reference 17) 5.3-30
Table 5.4-1 PM-10 and PM-2.5 Cumulative Collection Efficiencies for
Wet Scrubbers at Coal, Oil, Wood, and Bark Combustors; and
Coke Production Units (Reference 6) 5.4-21
ix
-------
TABLES (continued)
Table 5.4-2 Current Industrial Applications of Wet Scrubbers
(References 1, 2, and 8) 5.4-22
Table 5.4-3 PMio/PM2.5 Control Potential for Various Scrubber Designs 5.4-24
Table 5.4-4 Capital Cost Factors for a Typical Scrubber (Reference 9) 5.4-26
Table 5.4-5 Recommended Gas Velocities, Liquid/Gas Ratios, and
Pressure Drops for Paniculate Wet Scrubbers (Reference 10) 5.4-27
Table 5.4-6 Annual Cost Parameters for Particulate Scrubbers
(Reference 11) 5.4-33
Table 5.4-7 Annual Cost Factors for Particulate Scrubbers (Reference 11) 5.4-34
Table 5.5-1 PM Control Efficiencies for Thermal Incinerators in
Phthalic Anhydride Manufacturing Processes (Reference 10) 5.5-10
Table 5.5-2 Operational Requirements for Satisfactory Incinerator
Performance for Various Industrial Applications and
Control Levels (Reference 3) 5.5-11
Table 5.5-3 Capital Cost Factors for Thermal Incinerators (Reference 11) 5.5-13
Table 5.5-4 Incinerator Annual Cost Parameters (Reference 11) 5.5-16
Table 5.5-5 Annual Cost Factors for Incinerators (Reference 12) 5.5-17
Table 6.1 Estimated Control Efficiencies for Drop Height
Reduction Techniques (Reference 7) 6-6
Table 6-2 Parameters for Hood Cost Equation (Reference 8) 6-9
Table 7-1 Summary of Emerging PM Control Technologies 7-2
-------
FIGURES
Figure 3-1 EPA Test Method 5 Sampling Train 3-5
Figure 5.1-1 Expansion Settling Chamber (Adapted from Reference 2) 5.1-2
Figure 5.1-2 Multiple-Tray Settling Chamber (adapted from Reference 2) 5.1-3
Figure 5.1-3 Elutriators in Series (Reference 3) 5.1-5
Figure 5.1-4 Momentum Separators (References 2 and 3) 5.1-6
Figure 5.1-5 Mechanically-Aided Separator (Reference 3) 5.1-7
Figure 5.1-6 Illustration of the Double Vortex Within a Cyclone (Reference 1) 5.1-9
Figure 5.1-7 Four Basic Cyclone Types (adapted from Reference 2) 5.1-10
Figure 5.1-8 Standard Dimensions of a Cyclone (Reference 6) 5.1-11
Figure 5.1-9 Typical Multiple Cyclone (Reference 3) 5.1-13
Figure 5.1-10 Typical Fractional Collection Efficiency Curve for a
Settling Chamber (Reference 2) 5.1-14
Figure 5.1-11 Impact of Particle Density on Settling Chamber
Fractional Collection Efficiency (Reference 3) 5.^14
Figure 5.1-12 Typical Fractional Collection Efficiency Curve
for a Momentum Separator (Reference 2) 5.1-15
Figure 5.1-13 Typical Fractional Collection Efficiency Curve for a
Mechanically-Aided Separator (Reference 2) 5.1-15
Figure 5.1-14 Typical Cyclone Efficiency Curve in
Log-log (A) and Linear (B) Scales (References 6 and 15) 5.1-17
Figure 5.1-15 Dimensions of the Cyclone Inlet and Outlet Ducts
for an Optimized Cyclone Design, According to the
lozia and Leith Cyclone Efficiency Theory (Reference 17) 5.1-19
Figure 5.1-16 Cumulative Collection Efficiency Data for
Multiple Cyclones at a Residual Oil-Fired Boiler (Reference 5) 5.1-20
Figure 5.1-17 Cumulative Collection Efficiency Data for
Multiple Cyclones at Coal and Wood Bark Boilers,
With and Without Fly Ash Reinjection (Reference 5) 5.1-21
Figure 5.1-18 Total Capital Investment vs. Inlet Duct Area
for 0.2 ft2 < Duct Area < 2.64 ft2 (Reference 19) 5.1-23
Figure 5.1-19 Total Capital Investment vs. Inlet Duct Area
for Duct Area > 2.64 ft2 (Reference 19) 5.1-24
Figure 5.2-1 Cutaway view of Wire-Pipe Electrostatic Precipitator (Reference 2). ... 5.2-2
Figure 5.2-2 Wire-Plate Electrostatic Precipitator (Reference 2) 5.2-8
Figure 5.2-3 Wire-Pipe Electrostatic Precipitator (Reference 2) 5.2-10
Figure 5.2-4 Square, Hexagonal, and Circular Pipe Arrangements for Wire-Pipe
Precipitators (adapted from Reference 4) 5.2-11
Figure 5.2-5 Rigid Frame Electrode (Reference 2) 5.2-12
Figure 5.2-6 Various Discharge Electrodes and Collection Plate Designs
(Reference 2) 5.2-13
Figure 5.2-7 Concentric Plate Precipitator (Reference 3) 5.2-15
xi
-------
FIGURES (continued)
Figure 5.2-8 Cumulative Collection Efficiency Data for Electrostatic
Precipitators at Coal-Fired Boilers, Primary Copper Producers,
and Iron and Steel Production Operations (Reference 11) 5.2-18
Figure 5.2-9 Effect of Collection Efficiency on ESP TCI Costs
(Reference 14) 5.2-23
Figure 5.2-10 Effect of the Use of Corrosion Resistant Materials on
ESP TCI Costs (Reference 14) 5.2-23
Figure 5.2-11 TCI Costs for ESPs With and Without Various Standard
Design Features (Reference 14) 5.2-24
Figure 5.3-1 Collection Mechanisms of Fabric Filtration (Reference 3) 5.3-2
Figure 5.3-2 Fractional Efficiency of Fabric Filters vs. Particle Size
(Reference 2) 5.3-4
Figure 5.3-3 Cutaway View of a Typical Shaker Fabric Filter (Reference 1) 5.3-8
Figure 5.3-4 Typical Shaker Mechanism (Reference 2) 5.3-9
Figure 5.3-5 Typical Design of One Compartment of a Reverse-air Cleaning
Fabric Filter (Reference 2) 5.3-11
Figure 5.3-6 Reverse-Air Fabric filter with Traveling Mechanism and
External Cake Collection (Reference 3) 5.3-12
Figure 5.3-7 Schematic of a Pulse-Jet Fabric filter with Enlarged View of Pulse Inlet
Area (Reference 1) 5.3-13
Figure 5.3-8 Cumulative Collection Efficiency for Fabric Filters at
Coal Combustors, Ferroalloy Electric Arc Furnaces, and
Iron and Steel Production Operations (Reference 15) 5.3-18
Figure 5.3-9 Effect of Cleaning Mechanism on Fabric Filter Capital Costs
(Reference 17) 5.3-24
Figure 5.3-10 Effect of Gas-to-Cloth Ratios on Fabric Filter Capital Costs
(Reference 17) 5.3-25
Figure 5.3-11 Effect of the Use of Insulation and Stainless Steel on
Fabric Filter Capital Costs (Reference 17) 5.3-26
Figure 5.3-12 Effect of Fabric Type on Capital Costs - Reverse-Air
Fabric Filter, G/C = 2.5 (Reference 17) 5.3-27
Figure 5.3-13 Effect of Fabric Type on Capital Costs - Pulse-Jet
Fabric Filter, G/C = 5 (Reference 17) 5.3-27
Figure 5.3-14 Annual Fabric Filter Operating Costs (Reference 17) 5.3-28
Figure 5.4-1 Schematic Diagram of a Spray Tower Scrubber (Reference 2) 5.4-4
Figure 5.4-2 Schematic Diagram of a Cyclonic Spray Chamber Scrubber
(Reference 1) 5.4-5
Figure 5.4-3 Typical Packing Materials for Packed Bed Scrubbers (Reference 2). ... 5.4-6
Figure 5.4-4 Schematic Diagram of a Packed Tower Scrubber (Reference 2) 5.4-7
Figure 5.4-5 Common Plate Designs for Impingement Plate Scrubbers
(adapted from Reference 2) 5.4-9
xii
-------
FIGURES (continued)
Figure 5.4-6 Schematic Diagram of a Plate Tower Scrubber
(adapted from Reference 2) 5.4-10
Figure 5.4-7 Diagram of a Mechanically-aided Scrubber (Reference 1) 5.4-12
Figure 5.4-8 Schematic Diagram of a Venturi Scrubber with Cyclonic
Separation (Reference 1) 5.4-13
Figure 5.4-9 Diagram of an Orifice Scrubber (Reference 1) 5.4-15
Figure 5.4-10 Diagram of a Sludge Ejector in an Orifice Scrubber
(Reference 2) 5.4-16
Figure 5.4-11 Schematic Diagram of a Condensation "Growth" Scrubber
(adapted from Reference 4) 5.4-17
Figure 5.4-12 Schematic Diagram of Charged Wet Scrubber (adapted
from Reference 2) 5.4-19
Figure 5.4-13 Cumulative Collection Efficiency Data for PM Wet Scrubbers
at Coal, Oil, Wood, and Bark Combustion Sources, and Coke
Production Operations (Reference 6) 5.4-20
Figure 5.4-14 Venturi Scrubber Capital Costs, Inlet Gas Flowrate
< 19,000 ACFM (Reference 11) 5.4-29
Figure 5.4-15 Venturi Scrubber Capital Costs, Inlet Gas Flowrate
> 19,000 ACFM (Reference 11) 5.4-29
Figure 5.4-16 Impingement Plate Scrubber Capital Costs, Inlet Gas Flowrate
<77,000 ACFM (Reference 11) 5.4-30
Figure 5.4-17 Impingement Plate Scrubber Capital Costs, Total Gas Flowrate
>77,000 ACFM (Reference 11) 5.4-30
Figure 5.4-18 Vertical Packed-bed Scrubber Capital Costs (Reference 9) 5.4-31
Figure 5.4-19 Horizontal Packed-bed Scrubber Capital Costs (Reference 9) 5.4-31
Figure 5.5-1 Calculated Theoretical Residence Tunes for Various-sized
Coke PM in an Incinerator, at Various Temperatures 5.5-2
Figure 5.5-2 Schematic Diagram of a Discrete Burner Thermal Incinerator
(Reference 4) 5.5-5
Figure 5.5-3 Schematic Diagram of a Distributed Burner Thermal Incinerator
(Reference 6) 5.5-6
Figure 5.5-4 Schematic Diagram of a Catalytic Incinerator (Reference 4) 5.5-8
Figure 5.5-5 Total Capital Investment vs. Flow Rate for a Thermal
Incinerator with 0, 35, and 50 Percent Recuperative
Heat Recovery (Reference 12) 5.5-14
Figure 5.5-6 Total Capital Investment vs. Flow Rate for a Thermal
Incinerator with 85 and 95 Percent Regenerative Heat
Recovery (Reference 12) 5.5-14
Figure 5.5-7 Annual Cost Curves for Incinerators with Recuperative (REC)
and Regenerative (REG) Heat Recovery (HR) (Reference 11) 5.5-18
xin
-------
FIGURES (continued)
Figure 6-1 Schematic of a Slag-tapping Hood at a Blast Furnace
(Reference 5) 6-3
Figure 6-2 Schematic of a Local Ventilation System at a "Skip Hoist"
Loading Station (Reference 5) 6-4
xiv
-------
1. INTRODUCTION
The U.S. Environmental Protection Agency (EPA) has recently analyzed information
on the health effects of elevated concentrations of respirable particulate matter (PM) in ambient
air. This analysis lead to revisions of the national ambient air quality standards (NAAQS) for
PM. The EPA has also added a new "indicator" to measure respirable PM concentrations.
The previous indicator was PM10, which is defined as particle matter having a nominal
aerodynamic diameter of 10 micrometer (^m) or less. The additional indicator is based on
smaller particles, PM25, defined as PM less than or equal to 2.5 micrometer in aerodynamic
diameter.1
1.1 PURPOSE OF THIS DOCUMENT
The purpose of this document is to support the development of implementation
strategies for attaining revised ambient standards for PM, based on PM2 5 and PM10. This
document is a revision of the EPA's 1982 guidance on Control Techniques for Particulate
Emissions from Stationary Sources.2 The focus of this document is on the control of PM10 and
PM2 5 emissions from industrial sources. This document does not address nonindustrial
sources, such as residential wood combustion and windblown dust, which are covered by
separate guidance documents.
Although they account for a smaller fraction of national PMj0 emissions than
nonindustrial sources (see Section 2), industrial sources can have significant ambient impacts.
These can be especially important in urbanized areas which are typically centers of both
population and industrial activity. In addition, PM emissions from industrial sources tend to
be concentrated in the smaller size ranges, increasing their importance in the implementation of
a potential standard for PM25.
1.2 OTHER RESOURCES
The EPA has recently developed control techniques documents for a number of
nonindustrial sources of PM emissions. In addition, reports have been prepared assessing the
overall levels of control that could be achieved both in direct emissions of PM, and in
emissions of gaseous pollutants that can react to produce secondary PM. Secondary PM is
produced mainly from sulfur oxides (SOX), nitrogen oxides (NOX), ammonia (NH3), and
volatile organic compounds (VOC). These precursor gases react with one another and with
oxygen and water in the atmosphere to form condensible compounds. Appendix A gives a
summary of control techniques documents and other EPA documents available to support the
development of control strategies for primary PM10 and PM25 emissions, and emissions of
precursor gases for secondary PM.
1-1
-------
1.3 ORGANIZATION
This document is organized hi seven sections, including this introduction. Section 2
gives background information on trends hi PM air quality and emissions, projected future
impacts of control programs, and major current sources of PMi0 and PM2 5 emissions.
Section 3 discusses the methods available to measure PM emissions. These techniques
are needed to estimate the level of emissions from the source before and after control, as well
as determine the control efficiency of the PM control devices/techniques. This section
discusses established as well as innovative procedures that have been developed to measure the
mass and/or size of PM, especially for PM10 and PM25. Techniques for identifying and
measuring the chemical species of the PM are discussed as well.
Section 4 presents approaches for reducing PM emissions through the use of fuel
substitution and source reduction techniques, i.e. process modifications or optimization.
Section 5 is the heart of this document, and contains detailed descriptions of the
prunary devices used to control PM at stationary sources: electrostatic precipitators (ESPs),
fabric filters, wet scrubbers, and incinerators. For each of these control devices, the various
designs of the devices are discussed along with the principles of operation. The range of
control efficiencies for each device is then discussed and the source categories to which the
devices are applicable are presented. The capital and annual costs for each device are also
included in Section 5 along with the energy and other secondary environmental impacts of the
technologies. Section 5 begins with a discussion of pretreatment techniques, that is similar in
format to the primary device discussion. The pretreatment devices are used to reduce the PM
loading on the primary PM collection devices, in order to reduce the size and, potentially, the
costs of the primary control device, and to possibly increase the overall PM collection
efficiency.
Section 6 discusses industrial fugitive emission controls that include enclosures,
ventilation techniques, and optimization of equipment and operations. Where available, the
reported control efficiencies of the control measures are presented.
Section 7 discusses the emerging PM control technologies that are being investigated by
the EPA and industry to increase the control efficiency of PM control and/or to target fine
particles. Many of these technologies have been implemented in pilot- or full-scale operation.
1-2
-------
1.4 REFERENCES FOR SECTION 1
1. Review of the National Ambient Air Quality Standards for Paniculate Matter: Policy
Assessment of Scientific and Technical Information - External Review Draft. U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina. 1996.
2. Control Techniques for Paniculate Emissions from Stationary Sources-Volume 1.
(EPA-450/3-81-005a). U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina. September 1982.
1-3
-------
2. BACKGROUND
National ambient air quality standards (NAAQS) for paniculate matter (PM) were first
established in 1971. These standards applied to total suspended PM (TSP) as measured by a
high volume sampler. The sampler design favored the collection of particles with aerodynamic
diameters up to 50 fim. In 1987, the EPA changed the indicator for PM from TSP to PM,0.
The NAAQS levels for PM10 were set at a 24-hour average of 150 /^g/m3 (with no more than
one expected exceedance per year), and an annual average of 50 ywg/m3 (expected arithmetic
mean).1 In 1997, the EPA revised the form of the 24-hour (daily) PM10 NAAQS and
established PM2 5 as a new fine PM indicator.
Prior to the revision, the 24-hour PM10 NAAQS was met when the expected number of
days per calendar year with a 24-hour average concentration above 150 /xg/m3 was less than or
equal to one (averaged over 3 calendar years). The revised 24-hour PM10 standard is met when
the 99th percentile of the distribution of 24-hour concentrations at each monitor in an area for a
period of one year (averaged over 3 calendar years) does not exceed 150 /ig/m3. The annual
PM10 standard was not impacted by the 1997 revisions.
The new PM2 5 NAAQS are set at an annual mean concentration of less than or equal to
15 jug/m3 and a 24-hour (daily) concentration less than or equal to 65 /*g/m3. The annual
standard is met when the three year average of the annual arithmetic mean of the 24-hour
concentrations from single or multiple community-oriented monitors does not exceed 15
/ig/m3. The daily standard is met when the 98th percentile of the distribution of the 24-hour
concentrations for a period of one year (averaged over 3 calendar years) does not exceed
65 /ig/m3 at each monitor within an area.
2.1 TRENDS IN AMBIENT PARTICIPATE MATTER CONCENTRATIONS AND
PARTICIPATE MATTER EMISSIONS
The most recent EPA report on trends in ambient PM concentrations covers PM10 for
the years 1988 through 1996.2 Complete data on ambient PM10 concentrations are available
from 900 monitoring sites with urban, suburban and rural locations. The annual arithmetic
mean PM10 concentration for all sites (national average) during 1988 was 32 j*g/m3. By 1996,
the annual arithmetic mean concentration had decreased to 24 pig/m3, a 25 percent
improvement over 1988 levels. The trend of PM10 concentrations at urban and suburban sites
was essentially the same with the annual mean decreasing from about 34 /ig/m3 hi 1988 to
about 26 jug/m3 hi 1996. The annual arithmetic mean at rural sites in 1988 was 25 jiig/m3. The
mean decreased 20 percent to 20 /ig/m3 in 1996.
An independent analysis of PM10 trends, conducted by Darlington, et.al., found the
same improvements hi concentrations.3 Data from monitoring sites that reported at least one
reading each year from 1988 through 1995 (585 sites) to the Atmospheric Information
Retrieval System (AIRS) were used in this analysis. Nationwide, the analysis indicated a 24
2-1
-------
percent reduction hi PM10 concentrations from 34 jug/m3 hi 1988 to 26 ^ig/m3 in 1995. About
160 of the monitoring sites used by Darlington were located hi counties designated as
nonattainment for PM10 and 425 were located hi attainment counties. The annual mean
concentration for all the nonattainment counties in 1988 was 41 ^ig/m3. This was reduced by
26 percent to 31 /ig/m3 hi 1995. The mean hi the attainment counties decreased 20 percent,
from 30 /zg/m3 hi 1988 to 24 /ig/m3 hi 1995.
Table 2-1 shows trends hi PM10 emissions for major emission sources from 1987 to
1996. The table shows a good deal of fluctuation hi emissions, mainly due to changes hi
natural wind erosion. It must also be noted that fugitive dust emissions estimates in the table
are subject to a high degree of uncertainty (e.g. paved and unpaved roads, construction,
agricultural operations, and wind erosion). These fugitive dust emissions are overestimated
when compared to ambient measurements of the mineral-related components of PM25.
Ambient concentration data are not available to assess historical trends hi PM2 5 ambient
concentrations or emissions. However, visibility can be viewed as a surrogate measure of
trends hi fine particles hi the range of 2.5 micrometers and under. Particles hi this size range
contribute greatly to the scattering and absorption of light (known as light extinction). There
are two large contiguous haze areas hi the continental U.S. One encompasses the eastern U.S.
and the other includes the western Pacific states. There has been a marked decrease hi haze
over the 20 year period from 1970 to 1990 in the western Pacific states. The mid-continent
section of the eastern haze area has remained relatively constant over this period, whereas haze
levels have increased over a large portion of the eastern U.S.2
2.2 PROJECTIONS FOR FUTURE CONTROL PROGRAMS AND EMISSIONS
The EPA's Office of Policy, Planning, and Evaluation (OPPE) has projected emission
levels for PM10 and PM2 5 based on implementation of control programs required under the
Clean Air Act Amendments (CAAA) of 1990. Control programs for PM under Title I of the
CAAA are projected to have only a small impact on overall future emissions~a reduction of
about 3 percent for PM10 and less than 0.1 percent for PM2
2
•2.5-
Because of a lack of available data, OPPE's projections did not take into account the
impact of regulations for hazardous air pollutants (HAPs) under Title III of the CAAA.
However, a number of the Title HI HAPs are metals that are emitted hi fine PM, both in bulk
and trace quantities. Standards implemented for these paniculate HAPs will have some
impacts on PM10 and PM25 emissions.
Substantial reductions in SO2 emissions are projected by OPPE as a result of acid rain
control programs implemented under Title IV of the CAAA. In addition, reductions hi NOX
emissions are projected as a result of Title IV, and reductions hi both NOX and VOC are
projected as a result of ozone control programs under Title I. All of these pollutants are
2-2
-------
Table 2-1. Summary of Trends in PM-10 Emissions from 1987 through 1996 (Reference 2)
to
Estimated Emissions (million tons)
Fuel combustion - utility and industrial
Fuel combustion - residential wood and other
Metals processing
Other industrial
Open burning and other waste disposal
Motor vehicles and off-highway engines
Wildfires and managed burning
*Agriculture
""Natural wind erosion
*Paved and unpaved roads
•"Construction, mining, and quarrying
Total
1987
0.5 '
0.8
0.2
0.8
0.3
0.9
1.0
7.3
1.6
16.6
12.5
42.5
1988
0.5
0.9
0.2
0.8
0.3
1.0
1.7
7.5
18.1
18.3
12.0
61.3
1989
0.5
0.9
0.2
0.8
0.3
1.0
0.9
7.3
12.1
17.6
11.7
53.2
1990
0.6
0.6
0.2
0.8
0.3
0.9
1.2
5.1
2.1
13.5
4.6
29.9
1991
0.5
0.7
0.3
0.7
0.3
0.9
0.9
5.1
2.1
13.6
4.4
29.6
1992
0.5
0.7
0.3
0.7
0.3
1.0
0.8
4.9
2.2
13.3
4.8
29.5
1993
0.5
0.6
0.2
0.7
0.3
1.0
0.8
4.5
0.5
13.9
5.1
28.0
1994
0.
0.
0.
0.
0.
1.
1.
4.
2.
13.
5.
30.
5
6
2
7
3
0
0
7
2
9
8
9
1995
0.6
0.6
0.2
0.7
0.3
0.9
0.8
4.7
1.1
12.8
4.2
26.9
1996
0.6
0.6
0.2
0.7
0.3
0.9
0.8
4.7
5.3
12.7
4.5
31.3
* Emissions from agricultural operations, wind erosion, paved and unpaved roads, and construction are far too large when reconciled with levels
of the mineral-related components of PMj 5 measured in the ambient air.
-------
precursors of secondary PM. Therefore, emission reductions for these pollutants are expected
to produce reductions in the formation of secondary PM.
2.3 SOURCES OF PM10 AND PM2 5 EMISSIONS
Emission sources can be broadly classified as point sources and area sources. Point
sources are large emission sources that are treated on a point-by-point basis in emissions
inventories. These are typically industrial facilities, utilities, or large commercial or
institutional emission sources. Area sources are defined as emission sources that are too
numerous or dispersed to be treated individually in an emissions inventory. This category also
includes highway vehicles and nonroad engines and equipment.
The emissions discussed hi the next two sections are based on the 1990 National
Inventory. The 1990 National Inventory attributed approximately 90 percent of PM]0 and 70
percent of PM25 emissions to fugitive dust from agriculture, paved roads, unpaved roads, and
construction activities. While these are certainly major sources of PM emissions, the
confidence in these estimates is low. These estimates are believed to be high, and the
inventory is being reviewed and revised to improve these estimates. For this reason, the
following two sections discuss important sources of PM emissions hi general terms without
estimates of impact on emissions.
2.3.1 Point Sources
Particulate matter emissions from utility, industrial and commercial/institutional
combustion sources are small hi comparison with emissions from area combustion sources.
This is due both to superior combustion conditions, which result hi higher combustion
efficiencies, and also to add-on PM controls for coal combustion and some oil combustion
sources. Utility, industrial, and commercial/institutional combustion were the most significant
point sources of PM10 and PM2.5 hi 1990. Other significant industrial sources included metal
processing, mineral products processing, and wood products processing.2
2.3.2 Area Sources
Fugitive emissions from agriculture, paved roads, unpaved roads, and construction
activities represent a major portion of PM10 and PM2.5 emissions. However, as stated above,
these emissions appear to be overestimated when reconciled to ambient measurements of the
mineral-related components of PM2 5 In addition to these fugitive dust emission sources, area
source combustion categories including residential wood burning, wildfires, and prescribed
burning of forest residues were important sources. Highway vehicles, nonroad engines and
equipment, and open burning of wastes also made significant contributions.2
2-4
-------
2.4 CHARACTERIZATION OF PM2.5 CONCENTRATIONS
Ambient samples of PM2 5 from eight research studies are summarized hi Figures 1 and
2.4>5 The PM2 5 samples were chemically analyzed to determine the amounts of ammonium
sulfate, ammonium nitrate, carbon, and soil present. Ammonium sulfate and ammonium
nitrate are secondary particles formed in the atmosphere from the reaction of ammonia with
sulfur dioxide (SO2) and nitrogen oxides (NOX), respectively. Carbon and soil are primary
particles. These are generally emitted directly into the atmosphere, or generated by processes
such as wind erosion, construction, or traffic on paved or unpaved roads. The results of these
analyses for eastern states are shown hi Figure 1. Figure 2 summarizes the results for western
states.
Figure 1 indicates that hi the eastern states, PM25 was dominated by ammonium sulfate
particles which accounted for 40 to 60 percent of the total mass. Ammonium nitrate particles
contributed another 5 to 15 percent. Carbon particles, from sources such as incomplete
combustion, accounted for 30 to 40 percent of the PM2 5 mass. The fraction of soil hi the
eastern samples ranged from 5 to 10 percent.5
Figure 2 shows that only 5 to 15 percent of the PM2 5 was ammonium sulfate, and
ammonium nitrate accounted for 1 to 35 percent of the total mass. The percentage of carbon
from incomplete combustion ranged from 35 to 65 percent of the western samples. Soil
content in the western samples contributed 5 to 15 percent of the total mass of PM2 5.5
The following sections of this document address techniques for reducing primary
paniculate emissions from stationary combustion sources and industrial processes. As
illustrated above, secondary particles (ammonium sulfate and ammonium nitrate) comprise a
large percentage of the PM25 samples hi both the Eastern and Western United States. This is
indicative of the need to address emissions of sulfur dioxide, nitrogen oxides, and ammonia
when considering means of reducing PM2 5 concentrations in the future.
2-5
-------
Figure 1. PM25 Composition in the Eastern United States
Nonurban Locations
Boundary Waters
(5.1 //g/m3)
New England Average
(9.5 //g/m3)
Appalachian & Mid-Atlantic
(11.35//g/m3)
H Carbonaceous D Nitrate BSoil FJSulfate H Other
Nonurban Location
Mid-South
(12.1jt/g/m3)
Urban Locations
Rochester
(14.9/yg/m3)
Washington, D.C.
(19.2//g/m3)
Note: PM2 5 mass concentrations are determined on at least one year of monitoring at each location using a variety of non-Federal reference methods. They should not
be used to determine compliance with the PM2 5 NAAQS.
-------
Figure 2. PM2.5 Composition in the Western United States
Urban Locations
Spokane
(11.0//g/m3)
San Joaquin Valley
(Avg. - 30 //g/m3)
South Coast
(Basin Avg.-28//g/m3)
BCarbonaceous DNitrate HSoil DSulfate HOther
Western Phoenix
(13.5//g/m3)
Sonoran Desert
(4.3 //g/m3)
Nonurban Locations
Badlands
(4.5//g/m3)
Central Rockies
(3.1 //g/m3)
Sierra Nevada
(4.5//g/m3)
Note: PM2 5 mass concentrations are determined on at least one year of monitoring at each location using a variety of non-Federal reference methods. They should not
be used to determine compliance with the PM25 NAAQS.
-------
2.5 REFERENCES FOR SECTION 2
1. Review of the National Ambient Ah" Quality Standards for Participate Matter: Policy
Assessment of Scientific and Technical Information - External Review Draft. U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina. 1996.
2. National Air Quality and Emissions Trends Report: Report Number
EPA-454/R-97-013. U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Research Triangle Park, North Carolina. January 1998.
3. Darlington, T.L., Kahlbaum, D.F., Heuss, J.M., and Wolff, G.T. Analysis ofPM10
Trends in the United States from 1988 through 1995. Journal of the Air & Waste
Management Association. October 1997. 1070-1078.
4. Pace, T.G. "PM25 in the Ambient Air". Proceedings on the AWMA Specialty
Conference PM-, 5 - A Fine Particle Standard. Long Beach, California. January, 1997.
5. Pace, T.G. and Kuykendal, W.B. "Planning Tools for PM2.5 Emission Factors and
Inventories". Proceedings of the Air & Waste Management Association's 91st Annual
Meeting and Exhibition. San Diego, California. June 1998.
2-8
-------
3 MEASUREMENT
The determination of the control efficiency of PM control devices requires the use of
methods to determine the control device inlet and outlet PM emissions. This section discusses
established as well as innovative procedures that have been developed to measure the mass
and/or size of PM, especially for PM10 and PM2 5. Techniques for identifying and measuring
the chemical species of the PM are discussed as well.
The most precise method of determining the mass concentration of PM is to collect the
entire volume of gas (and PM) and to determine the mass concentration from this sample. This
procedure, however, is feasible only with a few sources where there are very low flow rates.
Procedures have been developed to sample small portions of the gas stream to obtain a
representative sample so that estimates of PM mass emissions can be made. These procedures
are called "extractive" methods, since a portion of the gas stream is removed from the source
and sampled elsewhere. Other more innovative procedures are being used to determine PM
mass concentrations in situ. Also, as part of a PM emission characterization of a source or
control device, the size distribution of the PM may be needed. This is especially true for PM2 5
emission determinations, since procedures to determine PM2 5 mass emissions directly are still
under development (see Section 3.5, below).
In the measurement of PM during extractive methods, it is important that the gas be
sampled isokinetically so that a representative sample of PM enters the sampling device. The
term "isokinetic" refers to the situation where the gas streamlines of the source gas are
preserved within the sampling probe so that the concentration and size distribution of the PM
hi the sample probe is the same as that hi the source effluent duct. The parameter that must be
controlled to establish isokinetics is the gas velocity within the sample probe, which must be
equal to the actual gas velocity at the sample point hi the source exhaust duct. Since the
sample probe will have a smaller diameter than the source exhaust duct and possibly a lower
temperature, the actual gas flow rate used to extract gas through the sampling probe must be
controlled to establish an isokinetic sampling velocity.
Anisokinetics, or the lack of isokinetics, can lead to either over or under sampling of
particles of a certain size. Sampling velocities less than isokinetic will lead to an
overestimation of larger-sized particles and a higher than actual PM mass concentration;
conversely, sampling velocities higher than isokinetic will lead to an overestimation of smaller
particles with a lower than actual PM mass concentration.
3.1 List of EPA PM Mass Measurement Test Methods
Table 3-1 lists the EPA test methods applicable to the measurement of PM mass
emissions. These methods are discussed further hi the next section. To obtain a detailed
3-1
-------
Table 3-1. EPA Test Methods for PM
EPA Method
Federal Register Reference
Description of Method
U)
Method 5
Method 5A
Method 5B
Method 5C
Method 5D
Method 5E
Method 5F
Method 5G
Method 5H
Method 201
Method 201A
Method 17
Method 202
Method 9
Performance Specification 1
Method 29
36 FR 24877 12/23/71
47 FR 34137 08/06/82
51FR42839 11/26/86
tentative
49 FR 43847 10/31/84
50 FR 07701 02/25/85
51FR42839 11/26/86
53 FR 05860 02/26/88
53 FR 05860 02/26/88
55 FR 14246 04/17/90
55 FR 14246 04/17/90
43 FR 07568 02/23/78
56 FR 65433 12/17/91
39 FR 39872 11/12/74
36 FR 24877 12/23/71
59 FR 48259 09/20/94
PM from stationary sources
PM from asphalt processing and asphalt roofing
Nonsulfuric acid PM
PM from small ducts
PM from (positive pressure) fabric filters
PM from wool fiberglass plants
Nonsulfate PM
PM from wood heaters - dilution tunnel
PM from wood heaters - stack
PM/PM-10 - exhaust gas recycle (EGR) procedure
PM/PM-10 - constant sampling rate (CSR) pocedure
In-stack filtration method for PM
Condensible particulate emissions from stationary sources
Visual determination of stack opacity; remote Lidar
CEMS for opacity at stationary sources
Metal emissions (and PM)
-------
description of these methods, the EPA's Technology Transfer Network, an electronic bulletin
board, can be viewed at http://www.epa.gov/ttn.
3.2 EPA Stationary (Point) Source PM Mass Measurement Test Methods
The following sections describe the EPA Test Methods for the sampling and analysis of
PM mass that include test methods for the measurement of total PM, PM10, condensible PM,
and opacity.
EPA Test Method 5, that measures total PM from stationary sources, is the
predominant test procedure used to measure PM mass emissions. The sampling train and
isokinetic sampling procedures described in Method 5 are also the basis for many other EPA
test methods. The Method 5 sampling train and procedures also has been modified and
adapted into test methods that are designed to measure other gas constituents, such as semi-
volatile compounds, in exhaust gases where PM is likely to also exist. In some cases, this is
because PM mass measurements are desired in addition to the target compounds; in other
cases, the PM is collected so as to remove the potential for interference with the measurement
of the target compounds.
Method 5 and the other stationary source measurement methods described below rely
on the use of EPA Test Methods 1 through 4. These methods describe the appropriate
techniques to be used to sample the exhaust gas from stationary sources, and also the
techniques used to obtain data on the physical and chemical characteristics of the exhaust gas
which are needed to calculate PM emissions. These auxiliary test methods and their variations
are listed in Table 3-2.
3.2.1 EPA Test Method 5 for Total PM Mass
This method is applicable for the determination of PM mass emissions from stationary
sources. Paniculate matter (PM) is withdrawn isokinetically from the source and collected on
a glass fiber filter maintained at a temperature in the range of 120 + 14 °C or another
temperature as specified in a regulation or approved for special purposes by the EPA for the
specific application. The PM mass, which includes any material that condenses at or above the
filtration temperature, is determined gravimetrically after removal of uncombined water.
A schematic of the sampling train used in this method is shown hi Figure 3-1.
Complete construction details are given in "APTD-0581: Construction Details of Isokinetic
Source-Sampling Equipment;"1 commercial models of this tram are also available. Changes
from APTD-0581 and allowable modifications of the train shown in Figure 3-1 can be obtained
3-3
-------
Table 3-2. EPA Test Methods 1 through 4: General Stack Sampling Procedures
EPA Test
Method
Description of Method
Method 1 Sample and velocity traverses for stationary sources
Method 1A Sample and velocity traverses for stationary sources with small stacks or ducts
Method 2 Determination of stack gas velocity and volumetric flow rate (Type S pitot tube)
Method 2A Direct measurement of gas volume through pipes and small ducts
Method 2B Determination of exhaust gas flow rate from gasoline vapor incinerators
Method 2C Determination of stack gas velocity and volumetric flow rate in small stacks or ducts
(standard pitot tube)
Method 2D Measurement of gas volumetric flow rates in small pipes and ducts
Method 2E Determination of landfill gas; gas production flow rate
Method 3 Gas analysis for carbon dioxide, oxygen, excess air, and dry molecular weight
Method 3 A Determination of oxygen and carbon dioxide concentrations in emissions from
stationary sources (instrumental analyzer procedure)
Method 3B Gas analysis for the determination of emission rate correction factor or excess air
Method 3C Determination of carbon dioxide, methane, nitrogen, and oxygen from stationary
sources
Method 4 Determination of moisture content in stack gases
-------
B
C
a
E
Figure 3-1. EPA Test Method 5 Sampling Train.
3-5
-------
from the EPA's Emission Measurement Technical Information Center.2 The operating and
maintenance procedures for the sampling train are described in APTD-0576: "Maintenance,
Calibration, and Operation of Isokinetic Source Sampling Equipment. "2 Correct usage of the
sampling train is important in obtaining valid results with this method.
3.2.2 EPA Test Method 5 Variations: 5A - 5H
The following methods are considered variations of Method 5 that target a specific
industry or type of PM emissions. The specifics of each method are summarized below and
include the differences between the method and Method 5, and any other noteworthy details.
Otherwise, the methods are largely identical to Method 5.
• Method 5A: Determination of PM Emissions from the Asphalt Processing
and Asphalt Roofing Industry. This method is similar to Method 5 except that
in this method the PM catch is maintained at a slightly lower temperature hi
Method 5A, 42°C vs. 120°C hi Method 5, and a precollector cyclone is used.
• Method 5B: Determination of Nonsulfuric Acid PM from Stationary
Sources. This method is similar to Method 5 except that the sample train is
maintained at a higher temperature hi Method 5B, 160°C vs. 120°C in Method
5, and the collected sample is heated hi the oven for 6 hours to volatilize any
sulfuric acid that may have collected. The nonsulfuric acid PM is then
determined by the method.
• Method 5C: Determination of PM in Small Ducts. A test method to address
PM measurement in small ducts is tentatively planned; no information about the
method is currently available.
• Method 5D: Determination of PM Emissions from Positive Pressure Fabric
Filters. Method 5D is similar to Method 5, except that it provides alternatives
to Method 1 hi terms of determining the measurement site, and location and
number of sampling (traverse) points. Since the velocities of the exhaust gases
from positive pressure fabric filters are often too low to measure accurately with
the type S pitot specified hi Method 2, alternative velocity determinations are
presented in Method 5D. Because of the allowable changes to site selection and
velocity determination hi Method 5D, alternative calculations for PM
concentration and gas flow are presented with the method.
Emission Monitoring and Analysis Division, Office of Air Quality Planning and Standards, U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina. 27711
3-6
-------
• Method 5E: Determination of PM Emissions from the Wool Fiberglass
Insulation Manufacturing Industry. This method is similar to Method 5
except that it measures both filterable and condensed PM enabling the
determination of total PM. A sodium hydroxide impinger solution is used to
collect the condensed PM.
• Method 5F: Determination of Nonsulfate PM Emissions from Stationary
Sources. This method is similar to Method 5 except that the sample train is
maintained at a higher temperature in Method 5F, 160°C vs. 120°C in
Method 5, and the collected sample is extracted with water to analyze for sulfate
content.
• Method 5G: Determination of PM Emissions from Wood Heaters from a
Dilution Tunnel Sampling Location. This method differs substantially from
Method 5 in that there are different sampling trains specified for Method 5G and
that the PM is withdrawn from a single point from a total collection hood and
sampling tunnel that combines the wood heater exhaust with ambient dilution
air. The PM is collected on two glass fiber filters in series, as opposed to only
one used in Method 5. The fiber filters are also maintained at a much lower
temperature in Method 5G, 32°C vs. 120°C in Method 5.
• Method 5H: Determination of PM Emissions from Wood Heaters from a Stack
Location. This method is more similar than Method 5G to Method 5, since the
filter is maintained at 120°C. Although, a dual filter sampling train from a
single point is used, as in Method 5G, the two filters are separated by the
impingers.
3.2.3 EPA Test Methods for PM10 from Stationary Sources
The following are two methods to measure PM10 emissions from stationary sources.
Both methods are in-stack procedures; one method uses exhaust gas recyclhlg and the other
constant sampling. Since condensible emissions not collected by these methods are also PM10
that contribute to ambient PMi0 levels, the EPA suggests that source PM10 measurements
include both in-stack PM10 methods, such as method 201 or 201 A, and condensible emissions
measurements to establish source contributions to ambient levels of PMi0, such as for emission
inventory purposes. Condensible emissions may be measured by an impinger analysis hi
combination with Method 201 and 201A, or by Method 202. Method 202 is discussed below
hi Section 3.2.5
3-7
-------
3.2.3.1 Method 201: Determination of PM10 Emissions-Exhaust Gas Recycle
Procedure
Method 201 applies to the in-stack measurement of PM10 emissions. In Method 201, a
gas sample is isokinetically extracted from the source. An in-stack cyclone is used to separate
PM greater than PM10, and an in-stack glass fiber filter is used to collect the PM^. To
maintain isokinetic flow rate conditions at the tip of the probe and a constant flow rate through
the cyclone, a clean dried portion of the sample gas at stack temperature is recycled into the
nozzle. The paniculate mass is then determined gravimetrically after removal of uncombined
water. Further information on this method can be found in the EPA document Application
Guide for Source PM10 Measurement with Exhaust Gas Recycle Sampling System.3
3.2.3.2 Methods 201A: Determination of PM10 Emissions—Constant Sampling Rate
Procedure
Method 201A is a variation of Method 201, and may be used for the same purposes as
Method 201. In Method 201 A, a gas sample is extracted at a constant flow rate through an in-
stack sizing device, which separates PM greater than PM10, attached to a PM sampling tram.
The sizing device can be either a cyclone that meets the specifications in the method or a
cascade impactor that has been calibrated using a specified procedure. Variations from
isokinetic sampling conditions are maintained hi the sampling train within well-defined limits.
With the exception of the PM10 sizing device and in-stack filter, this train is the same as an
EPA Method 17 train. The paniculate mass collected with the sampling train is then
determined gravimetrically after removal of uncombined water. Further information on this
method can be found hi the EPA document Application Guide for Source PMW Measurement
with Constant Sampling Rate.*
3.2.4 EPA Test Method 17: Determination of PM Emissions from Stationary Sources-
In-Stack Filtration Method
This method describes an in-stack gas sampling method that can be used in situations
where PM concentrations are not influenced by stack temperatures, over the normal range of
temperatures associated with the source category. Therefore, Method 17 eliminates the use of
the heated glass sampling probe and heated filter holder required in the "out-of-stack" Method
5, that is cumbersome and requires careful operation by usually trained operators. Method 17
can only be used to fulfill EPA requirements when specified by an EPA standard, and only
used within the stack temperature range also specified by the EPA. Method 17 is especially
not applicable to gas streams containing liquid droplets or which are saturated with water
vapor. Also, Method 17 should not be used if the projected cross-sectional area of the
probe/filter holder assembly covers more than 5 percent of the stack cross-sectional area.
3-8
-------
3.2.5 Method 202 for Condensible PM Measurement
This method applies to the determination of condensible paniculate matter (CPM)
emissions from stationary sources. It is intended to represent condensible matter as material
that condenses after passing through a filter. In Method 202, condensible PM is collected in
the impinger portion of a Method 17 type sampling train. The impinger contents are
immediately purged after the run with nitrogen gas to remove dissolved sulfur dioxide gases
from the impinger contents. The impinger solution is then extracted with methylene chloride.
The organic and aqueous fractions are then taken to dryness and the residues weighed. The
total of both fractions represents the condensible PM.
There is the potential for low collection efficiency at oil-fired boilers with this method.
To improve the collection efficiency at these sources, an additional filter should be placed
between the second and third impinger. In sources that use ammonia (NH,) injection as a
control technique for hydrogen chloride (HC1), NH3 can interfere with the determination of
condensible PM by Method 202 by reacting with HC1 in the gas stream to form ammonium
chloride, which is then measured as condensible PM. The method describes measures that can
be taken to correct for this interference.
The filter catch of this method can be analyzed according to the appropriate method to
speciate the PM. Method 202 also may be used in conjunction with the methods designed to
measure PM10 (Method 201 or 201 A) if the probes are glass-lined. If Method 202 is used in
conjunction with Method 201 or 201A, the impinger train configuration and analysis specified
in Method 202 should be used in conjunction with a sample train operation and front end
recovery and analysis conducted according to Method 201 or 201A. Method 202 may also be
modified to measure material that condenses at other temperatures by specifying the filter and
probe temperature. A heated Method 5 out-of-stack filter may be used instead of the in-stack
filter to determine condensible emissions at wet sources.
The following documents discuss the measurement of condensible PM and the
development of this method in more detail: Measurement of Condensible Vapor Contribution
to PM10 Emissions,5 A Review of Current Methods for Measuring Paniculate Matter Including
Condensibles from Stationary Sources,6 and Method Development and Evaluation of Draft
Protocol for Measurement of Condensible Paniculate Emissions.1
3.2.6 EPA Test Method 9: Visual Determination of the Opacity of Emissions from
Stationary Sources, and Alternate Method 1 for the Use of Remote Lidar
This method involves the determination of plume opacity by qualified observers that are
trained and certified according to procedures described in the method. Method 9 describes the
procedures that are to be used by these observers to determine plume opacity in the field. The
method also includes performance criteria that are applicable to the method variables which,
unless controlled, may exert significant influence upon plume appearance to the observer.
3-9
-------
Alternate Method 1 to Method 9 provides for the remote determination of opacity using
a Lidar (laser radar light detection and ranging) system that employs a ruby (red) laser. The
Lidar uses its own light source and, therefore, can be used hi the day or night. The
Alternative Method includes design, calibration, and performance evaluation procedures for
the Lidar system. The method is applicable to a stationary or mobile Lidar system.
3.2.7 Performance Specifications for Continuous Emissions Monitoring Systems (CEM)
Used to Monitor Opacity
The EPA has published performance specifications (PS 1) for the use of CEMS to
monitor opacity at new stationary sources (40 CFR 60, Appendix B). This performance
specification applies to opacity monitors installed after March 30, 1996. The CEMS monitors
PM using the principle of transmissometry of light. Light that has specific spectral
characteristics is projected from a lamp in the CEMS through the stack gas. The intensity of
the light after passing through the gas is attenuated (absorbed and scattered) by the PM and
then measured by a sensor. The percentage of the projected light attenuated is defined to be
the opacity. Transmittance is defined to be the opposite of opacity, i.e., opaque stack
emissions that attenuate all of the light will have a transmittance of zero and an opacity of 100
percent, and transparent stack emissions have a transmittance of 100 percent and opacity of
zero.
This performance specification establishes specific design criteria for an opacity
monitor/transmissometer CEMS. The performance specification also specifies installation,
calibration, and evaluating criteria to ensure proper performance of the CEMS.
3.3 Other Stationary (Point) Source PM Mass Measurement Test Methods
Other test methods for the measurement of PM mass include ASTM Method
D3685/D3685M8 and ASME Power Test Code 27.9 A piezoelectric quart crystal has been
used to measure "quasi-continuous" PM mass emissions.10 The device directly measures the
electrical frequency shift of the crystal due to the accumulating PM mass.
A triboelectric instrument has been developed for use as a continuous PM mass
emissions monitor by Auburn International, called the Triboflow CEM 2604. The triboelectric
effect is the transfer of electrical charge that takes place when two objects rub or abrade each
other. In the Triboflow CEM, the tribolelectrical charge that is transferred from PM in a duct
to a stainless steel probe is monitored as a current flow. Note that the triboelectric effect is
different from static electricity, which is the storage as opposed to the transfer of charge. The
Triboflow CEM is suited for monitoring the PM mass emissions of a relatively low level PM
source, such as at the exit of a high efficiency PM control device.11
3-10
-------
The use of CEMs for PM mass measurement is currently being investigated by EPA's
Office of Research and Development. An EPA "Application Guide" for the mass measurement
of PM2 5 is also under development, but has not been validated yet.12
3.4 Fugitive PM Measurement Methods
The following are brief descriptions of eight methods that are available to measure
PM10 from fugitive dust sources. These methods are described hi detail in the EPA document:
A Review of Methods for Measuring Fugitive PM-10 Emission Rates.13 Some of these methods
have been available for a long time while others have been developed recently. Many are not
frequently used and/or they were developed for other purposes (e.g., soil science). Guidance
on selecting the most appropriate method for a given source type has been developed14 and can
be found as an appendix in the above EPA review document (Reference 13). Another
document that discusses many of the methods described below is Techniques and Equipment
for Measuring Inhalable Paniculate Fugitive Emissions.15
Other documents that discuss the individual methods are referenced below with the
description of each method. Because many of these methods were developed before PM10 was
a concern, some of the methods manuals referenced here describe only the measurement of
total suspended PM. Therefore, although these earlier documents contain a substantial amount
of valuable information, the equipment described for use in the methods are likely to be
outdated. More current documents are referenced where available.
• Quasi-stack Method. This method consists of enclosing or hooding the fugitive
dust source, on either a permanent or temporary basis, with the use of a fan and
then sampling the exhaust isokinetically using EPA Test Methods 201 or 201A.
This method is considered to be potentially the most accurate because the entire
plume is captured and measured close to the source. Care must be taken,
however, not to artificially generate emissions from the source with the
sampling equipment. Additional information about this method can be found in
Technical Manual for the Measurement of Fugitive Emissions: Quasi-Stack
Sampling Method.16 More current equipment is described in Technical Manual:
Hood System Capture of Process Fugitive Emission^1 and Evaluation of an Air
Curtain Hooding System for a Primary Copper Convener, Volume 7.18
• Roof Monitor Method.19 This method may be the best means of measuring
fugitive PM when a number of processes are located within a building. The PM
is measured from all openings in the building and the total fugitive emission rate
is the sum of the emission rates from all the openings. This method is best used
when the building itself is construed as the "source." The method involves the
measurement of the PM10 concentration (with EPA Test Method 201 or 201 A) in
the duct exhaust, which is then multiplied tunes the air exit velocity and the
opening cross-sectional area to produce the emission rate. Since the dust
3-11
-------
concentration may vary across the duct opening, a number of sites along the
cross section of each duct should be sampled, as in stack testing (EPA Test
Method 1). If sampling is performed in an actual roof monitor vent, it is
recommended that sampling be done according to EPA Method 5D.
In the event that isokinetic sampling is not feasible because of the variation in
air velocity hi the openings, ambient PM10 measurement techniques may be
used. Ambient samplers that have met EPA criteria are described in the "List of
Designated Reference and Equivalent Methods" issued by the EPA's
Atmospheric Research and Exposure Assessment Laboratory.6 Reference 13
contains the list as of February 8, 1993.
Upwind-downwind Method.20 In this method, ambient PM10 concentrations are
measured upwind and downwind of a dust source. The difference between the
two concentrations is considered to be the PMi0 concentration due to the fugitive
emission source. Using wind speed, direction, and other meteorological data
obtained during the PM10 sampling period, the emission rate is determined using
dispersion models. The EPA "Industrial Source Complex" model is being
revised to develop an improved deposition term to make it more accurate for use
with PM.
While the upwind-downwind method is considered the most versatile of the
fugitive measurement methods, it has also been considered the least accurate,
since only a small portion of a greatly diluted plume is sampled. Recent studies,
however, have found that fugitive emission rates estimated using the upwind-
downwind method are within a factor of two 80 percent of the time as with the
quasi-stack method.21
Exposure Profile Method.22 This method consists of using a number of
ambient samplers (typically 4 or 5) at several heights along a vertical tower (4 to
10 meters hi height) equipped with nozzles and flow rate adjustments to sample
the fugitive PM plume isokinetically. The tower is also equipped to measure
wind speed and direction. The towers are placed downwind of the source, with
ambient samplers (1 to 4) also placed upwind of the source to determine the
background PM concentration. Ambient data obtained from these samplers are
used to determine the total mass flux of dust emitted from the source; this is
done by integration of the dust exposure values obtained from the various
sampling points.23
b U.S. Environmental Protection Agency, Office of Research and Development, Atmospheric Research and
Exposure Assessment Laboratory, Methods Research and Development Division (MD-77), Research Trian
Park, North Carolina. 27711. (919)541-2622 or 4599.
3-12
-------
The Exposure Profile Method is largely constrained to situations where
sampling close to the source is possible, except where extensions to the towers
are used.24 However, even with these extensions, the exposure profiling method
may not be practical for sampling large area sources. Losses of PM10 may occur
if the source is close to the ground, but since this dust would not become
airborne and contribute to emissions, the measurement of these "relevant
emissions" is possibly just as desirable as the actual emissions from the source.13
The exposure profiling method is considered more accurate than the upwind-
downwind method and is considered to be comparable to the roof monitor
method13 and is considered well suited to roads.13
Portable Wind Tunnel Method. This method is applicable to wind-generated
fugitive emissions only. It was developed in the 1970's to study the effects of
wind-blown sand on vegetation and to quantify the sources of wind erosion. It
has been used since to quantify wind-generated emissions from exposed soil and
coal storage piles.25-26'27 The portable wind tunnel consists of a vacuum cleaner-
shaped device, the mouth of which is placed directly on the surface to be tested
using an airtight seal. A fan draws air through the mouth of the tunnel, through
a long tube into a raised duct section where PM sampling can occur. Sampling
can be performed using EPA Method 201 or 201 A. The fugitive dust emission
rate is equal to the particle concentration times the tunnel flow rate. Over flat
ground, the tunnel centerline wind speed can be related to wind speed at 10
meters altitude.
Scale-Model Wind Tunnel Method. This method involves creating a wind
tunnel that resembles the source or terrain to be sampled with the use of, in
many cases, a to-scale recreation of the source within the wind tunnel.
Parameters such as turbulence, velocity profile, wind shear, and other physical
quantities, such as air moisture and terrain roughness, are usually duplicated
within the wind tunnel. The advantage of using a scale-model wind tunnel is
that the individual parameters affecting dust emissions can then be controlled.
The disadvantage is that the relationship between the tests and actual field
measurements is "uncertain."
Tracer Method. This method uses either a gas or particles as a tracer for dust
from the source to be measured. Common tracers are sulfur hexafluoride
(SF6)28 and fluorescent or phosphorescent materials or coatings. The assumption
is that the tracer plume will strongly resemble the dust plume if the tracer is
released hi the same place and time as the dust. Downwind measurements of the
tracer and dust concentrations are used to quantify the (upwind) dust emission
rate by a direct proportion using the (upwind) tracer release rate (i.e. emission
rate). A study was done to determine the accuracy of the gas tracer method.
3-13
-------
This study determined that using a correction factor of 1.03 to calculate the
source emission rate increases the accuracy of the method.29
• Balloon Method.30'25 This method is a variation of the exposure profiling
method, discussed above. In the Balloon Method, a balloon is used to suspend
ambient samplers at varying heights instead of the sampling tower used hi the
exposure profiling method. This method is especially suited for sampling large
area sources and/or sources which may not be closely approached. The problem
with the Balloon method is that often the sampling is not done isokinetically,
since once the balloon is aloft, nozzles cannot be changed. If variable flow rate
sampling is not possible, the fixed flow rate in the samplers may also contribute
to anisokinetics. However, isokinetic sampling is less critical to accurate
measurement of PM10 than for total suspended PM.31
3.5 Particle Size Analysis
The size distribution of a paniculate dust stream is sometimes desired to determine the
emissions from a source or collection efficiency of a PM control device. Various measurement
approaches are available to determining the size distribution of a paniculate stream that include
cascade impactors, sampling cyclones, centrifugal separators, and more advanced techniques
that utilize lasers. Note that EPA Test Method 201 and 201A can be used directly to
determine PM10 emissions and collection efficiency; however, for PM2 5 and other particle
sizes, one of the methods discussed below is needed to determine emissions or collection
efficiency.32
3.5.1 Cascade Impactors
Cascade impactors are a widely used method to size particles that have been
commercially available for source testing since the early 1970's,33 and have a relatively well-
developed theoretical basis.32'34 Impactors collect particles by inertial impaction and utilize a
series of plates (discs) or stages with various-sized holes 0ets) that alter the velocity of the gas
passing onto the next stage. Particles of a specific size or larger will impact each plate, while
smaller particles will pass through to the next plate. The plates are coated with a sticky
material (substrate) that causes the PM impacting the plates to be irreversibly collected. The
selection of cascade impactor substrates is an important part of impactor use.32'35
Cascade impactors generally can determine particle sizes between 0.3 to 16 //m,32 with
low pressure impactors commercially available (Pollution Control Systems Corp.)c that
measure particles between 0.02 and 10 /urn.36'37 The major limitation of cascade impactors is
that only a small amount of PM (usually less than 10 mg) can be collected on each stage;32
Pollution Control Systems Corporation, Seattle, Washington.
3-14
-------
therefore, the gas sampling volume/time must be adjusted to accommodate for this upper limit.
Cascade impactors may also be subject to biases towards small particles because of particle
bounce and reentrainment, and because of fracturing of larger particles during impaction.
The proper operation of cascade impactors for source testing is described in the EPA
document, Procedures for Cascade Impactor Calibration and Operation in Process Streams?*
3.5.2 Sampling Cyclones
Cyclone samplers operate in the same manner as cyclones used for PM collection, in
that the gas with PM is forced to spin so that some of the PM hits the cyclone walls and is
collected. PM above a certain size specific to the cyclone will be largely collected and
particles below a certain size will mostly pass through the cyclone uncollected. Individual
sampling cyclones are only able to determine the mass of particles either above and below a
specific size. However, if various-sized sampling cyclones are used in series, PM sizes over a
(relatively broad) range can be determined. The limitations of the sampling cyclones are that
they are inadequate for sampling gases with low PM concentrations.32
In the late 1970's, the EPA developed the Source Assessment Sampling System (SASS)
as a broad source assessment tool to screen for PM, organics, and inorganics hi one sampling
train.39'40 The SASS train included a series of three cyclones and a back-up filter that separated
PM into the following size categories: > 10 yum, 3 to 10 //m, 1 to 3 /urn, and < 1 /urn.
Because the SASS train was constructed out of stainless steel to better withstand field
conditions, the recovery rates for the chemical analyses were not as high as sampling trains
made out of glass, such as the Office of Solid Waste's method, SW846-Method 0010 (Modified
Method 5) described in Section 3.6.2.41 The OSW method, however, does not fractionate PM
by size.
A system of five cyclones nested in series has also been developed42'43 and is
commercially available from vendors, including Sierra Instruments, Inc.d The five cyclones
have progressively smaller cut points, with reported d^'s of 5.4, 2.1, 1.4, 0.65, and 0.32 /urn.
3.5.3 Real-Time Size Distribution Measurement
An Aerodynamic Particle Sizer (APS) has been developed that can determines the "real-
time" size of particles by measuring their velocity as the particles accelerate through a plate
orifice.44 The particle velocity measurements are made with a laser Doppler velocimeter45>46
The instrument is commercially available from TSI Inc.6
d Sierra Instruments, Inc., Carmel Valley, California.
e TSI Inc., St. Paul, Minnesota.
3-15
-------
The in situ measurement of PM using lasers have been reported to be successful for
small particles, less than 1 ,uni.47 The laser is used to heat the PM and, based on the PM
cooling rates, size differentiation is possible. This method has the advantage of a very short
measurement time period (microseconds), so that the effect of rapidly changing processes on
the particle size distribution can be determined.
Acoustical techniques are also being used to size PM. This technique relies on the
detection and measurement of elastic waves arising from the impact of the particles on a
surface, such as a plate,48 or other particles.49 The acoustic signal can be measured
inexpensively by a high fidelity piezoelectric transducer48 or with the use of a simple
microphone.49 The latter method is better suited to regularly-shaped PM, while the former is
not recommended for high number concentration particle streams (> 10? per m3).
An optical method has been developed from an EPA funded study by Insitec
Measurement Systems/ The method, known as "Transform method for Extinction-Scattering
with Spatial resolution" (TESS), is a patented technique based on laser light scattering.50 TESS
measures total particle concentration. The measurement is independent of particle
composition, velocity, and size distribution. The particle concentration is measured as a ratio
of scattered to transmitted light.
A diffusion battery that can be used in a source situation has been developed.51
Diffusion batteries utilize arrays of screens, tubes, or plates, which present a large surface area
for the deposition of particles by Brownian diffusion. Paniculate matter of different sizes can
be classified with diffusion batteries because of the difference in diffusion coefficients for
particles of different sizes.52'53
Other real-time particle sizing/counting methods are available. These devices include
optical devices, condensation nuclei counters, and electrical mobility analyzers.54-55-56-57-58-59
Some of these devices may not be appropriate for source level PM concentrations and/or can
not properly account for particle refractive index or shape.
3.5.4 Size Distribution of Bulk Samples
Particulate matter can be collected from the source and later analyzed for the particle
size distribution in the laboratory using various available techniques. These techniques should
be used with caution, however, because the original flue gas particle distribution may be
altered by agglomeration, particle breakup, chemical reactions, or loss of volatiles that occur
during sample collection and storage. Also, artifact mass may be formed from filter materials,
such as glass fiber, that oxidize in contact with acid gases in the sample air. Therefore, the
f Insitec Measurement Systems, San Ramon, California.
3-16
-------
size distribution results obtained with these methods are meaningful only if the effects of
sample collection and storage are negligible or clearly known.
Particle size of PM collected on filters can be determined with scanning electron
microscopes (SEM)60. The particle volume size distribution is estimated from the size
distribution determined by the SEM. SEM can be used down to 0.01 yum.61
Another device used for particle sizing electrically stimulates the particles that are
resuspended in a conductive fluid. The amplitude of the resulting electrical pulse generated by
each particle is measured as the particles pass individually through an orifice or aperture.
Because the electrical pulse is proportional to the particle volume, particle diameter can be
estimated by assuming a spherically-shaped particle. Instruments such as a Coulter® Counter62
and Elzone®11 Electrozone63 are commercially available to provide this type of analysis.64'65
These devices can provide analysis of particle with sizes >20 ju.m and as low as 0.35,wm.62
A Bahco Micro-Particle Classifier is the particle sizing device recommended by the
Industrial Gas Cleaning Institute1 to determine particle size and collection efficiency of
mechanical collectors, for particles of 1 to over 20 /urn in diameter. Procedures for calibration
of the Bahco are found hi the American Society of Mechanical Engineers Power Test Code 289
that describes procedures to be used for determining the size of fine PM found in dust and
smoke. The Bahco is one of a class of instruments called centrifugal classifiers, that hi effect
determine the terminal velocity distribution of the PM, which can be related to particle size
using Stokes' Law. In reality, these instruments determine the diameter of equivalent solid
spherical particles and, therefore, may not provide useful information if the PM is significantly
nonspherical.
An across-duct laminar flow device has been developed that collects PM for later size
analysis by a Malvern laser diffraction size analyzer.66 This instrument also measures the
electric charge of the particles so that it can be used to optimize ESP operation.
3.6 Speciation
Extractive PM sampling procedures, such as EPA Test Method 5, allow for the
chemical speciation of the collected PM. Most methods for speciation of PM use spectroscopic
detection, although other analytical procedures can also be used. The following sections
discuss the various techniques that can be used to speciate PM. Descriptions of two EPA test
g Coulter Electronics, Hialeah, Florida.
h Particle Data Laboratories, Ltd., Elmhurst, Illinois.
1 Industrial Gas Cleaning Institute, Inc., Stamford, Connecticut.
3-17
-------
methods that detail procedures to collect as well as analyze PM for individual species are also
included below.
3.6.1 EPA Test Method 29 for Metals and PM
Method 29 is applicable to the determination of antimony (Sb), arsenic (As), barium
(Ba), beryllium (Be), cadmium (Cd), chromium (Cr), cobalt (Co), copper (Cu), lead (Pb),
manganese (Mn), mercury (Hg), nickel (Ni), phosphorus (P), selenium (Se), silver (Ag),
thallium (Tl), and zinc (Zn) emissions from stationary sources. This method may be used to
determine total PM emissions, in addition to the metals emissions, if the prescribed procedures
and precautions are followed.
In Method 29, a stack sample is withdrawn isokinetically from the source; PM
emissions are collected in the probe and on a heated filter, and gaseous emissions are then
collected in an aqueous acidic solution of hydrogen peroxide (analyzed for all metals including
Hg) and an aqueous acidic solution of potassium permanganate (analyzed only for Hg). The
recovered samples are digested, and appropriate fractions are analyzed for Hg by cold vapor
atomic absorption spectroscopy (CVAAS) and for Sb, As, Ba, Be, Cd, Cr, Co, Cu, Pb, Mn,
Ni, P, Se, Ag, Tl, and Zn by inductively coupled argon plasma emission spectroscopy (ICAP)
or atomic absorption spectroscopy (AAS). Iron (Fe) can be a spectral interference during the
analysis of As, Cr, and Cd by ICAP. Aluminum (Al) can be a spectral interference during the
analysis of As and Pb by ICAP.
Graphite furnace atomic absorption spectroscopy (GFAAS) is used for analysis of Sb,
As, Cd, Co, Pb, Se, and Tl if these elements require greater analytical sensitivity than can be
obtained by ICAP. If desired, AAS may be used for analysis of all listed metals if the
resulting in-stack method detection limits meet the goal of the testing program. Similarly,
inductively coupled plasma-mass spectroscopy (ICP-MS) may be used for analysis of Sb, As,
Ba, Be, Cd, Cr, Co, Cu, Pb, Mn, Ni, As, Tl and Zn.
Method 29 is discussed hi more detail hi the EPA document Multiple Metals Stack
Emissions Measurement Methodology.61 Method 29 is virtually identical to two other EPA
methods:
EPA's Office of Solid Waste method SW-848-0012,68 "Methodology for the
Determination of Metals Emissions from Hazardous Waste Incineration and
Similar Combustion Sources;" and
• EPA's Office of Solid Waste method: "Methodology for the Determination of
Metals Emissions from Hazardous Waste Incineration and Similar Combustion
Sources," that was developed for boilers and industrial furnaces (BIF)69.
3.6.2 EPA Office of Solid Waste Test Method 0010 (SW-846)41
3-18
-------
This method is used to determine the amount of semivolatile organic constituents
hi exhaust gas, as well as to determine total PM, as per EPA Method 5; hence, the common
name of "Modified Method 5." OSW Test Method 0010 can be used to detect polychlorinated
biphenyls, chlorinated dibenzodioxins and dibenzofurans, poly cyclic organic matter, and other
semivolatile organic compounds. The method is used by OSW to determine the destruction
and removal efficiency of the principle organic hazardous constituents from waste incinerators.
Test Method 0010 is an EPA Test Method 5 sampling train modified to include a high
efficiency filter and a packed bed of XAD-2 or foam resin. The filter collects the organic-
laden PM and the packed bed adsorbs the semivolatile organic species in the flue gas. The
method includes a description of a variety of comprehensive chemical analyses to determine the
identity and concentration of the organic material. Analysis of the filter only yields PM
speciation information.70
3.6.3 Spectrometry
Spectrometry is a common technique used to determine the species present in PM.
However, since the spectrometric detectors respond to the presence of only the element, they
provide no information about chemical compounds and, in most cases, do not indicate the
oxidation state of the element.
3.6.3.1 Atomic Absorption Spectrometry
Atomic absorption Spectrometry (AAS)71 usually involves some type of (acid) extraction
of the analyte followed by excitation of the solution in a flame. Light with a wavelength
characteristic of the element of interest traverses the flame. The amount of light absorbed is
then related to the quantity of the element present. In AAS, individual elements must be
determined sequentially. Thus, although any element can be determined for which a lamp is
available to produce the characteristic light, most PM samples are large enough to allow only
half a dozen determinations. Some elements, if present hi the PM, (such as antimony and
arsenic), may require the application of special methods. Atomic absorption is subject to
interferences which can lead to substantial errors. If recognized, however, these errors
generally can be accounted for or eliminated to produce good quantitative analyses.
3.6.3.2 Optical Emission Spectrometry
Optical emission Spectrometry (OES) involves the excitation of the loosely bound
electrons hi elements to observe their characteristic emissions as de-excitation occurs. The
wavelength of the resultant light is characteristic of the element, and the intensity is an
indication of the quantity of the element present.
The most desirable OES technique is argon plasma excitation.72 Plasma Spectrometry
offers more advantages than AAS, with similar sample preparation, analysis rates, and
3-19
-------
detection limits. The OES techniques can simultaneously determine anywhere up to 50
elements. Optical emission spectrometry is, in general, more interference-free than atomic
absorption.
3.6.3.3 Mass Spectrometry
Mass spectrometry (MS) is a currently expanding field of analysis, whose growth has
been fueled by the need for increased accuracy and detection of highly toxic pollutants that are
present in small concentrations hi the environment. Spectral interferences with MS are
important, but generally can be overcome by use of a spectrometer with high resolution. As
with any multi-element technique, the accuracy of MS depends on the elements to be analyzed
and on the sample matrix. The advantage of MS is that every element in the periodic table can
be simultaneously detected, with roughly equal sensitivity in the parts per million (ppm) and
sub-ppm ranges. The MS techniques are also able to distinguish between isotopes, which is
sometimes desirable hi the determination of isotope ratios.
Two specific types of MS techniques applicable to PM samples are spark source mass
spectrometry73 and laser ion source MS, sometimes called laser microprobe mass analysis
(LAMMA).74 Both techniques use an energy source to vaporize and partly ionize small
amounts of the PM which, under specially controlled conditions, are then accessible for MS
analysis. LAMMA has the advantage hi that it can also determine particle size and shape with
the same resolution as a light microscope (approximately 0.13 ^m).
3.6.3.4 Neutron Activation Analysis
Neutron activation analysis (NAA) consists of a variety of distinct methods, all of
which produce unstable nuclei that emit gamma radiation.75'76 The energy and intensity of the
gamma rays are indicators of the element and its quantity. Instrumental thermal NAA is the
most commonly used method for PM. In this approach, a nuclear reactor is used to produce
unstable nuclei. Neutron activation analysis can simultaneously determine up to 25 elements in
on PM sample. Another advantage is that particles can be analyzed directly on the collecting
filter surface.
3.6.3.5 X-Ray Fluorescence Spectrometry
X-Ray fluorescence spectrometry involves excitation of tightly bound electrons and
observation of the X-ray emission as de-excitation occurs.75'77 Excitation may be done by a
variety of techniques, but use of an x-ray generator is the most common. The technique may
be either multielement (up to perhaps 30) energy dispersive detection or wavelength dispersive
detection (up to perhaps 10 elements). Only elements with atomic numbers greater than that of
magnesium can be analyzed. Particles can be analyzed nondestructively, directly on a filter.
Interferences are common and calibration can be a problem.
3-20
-------
3.6.4 Electrochemical
Electrochemical methods have been used to a limited extent to determine a small
number of elements in PM samples. Some of these methods are: potentiometry with ion-
selective electrodes, polarography, and anodic stripping voltametry.75 Electrochemical
methods have few advantages for PM analysis aside from the low initial capital costs of
equipment relative to other techniques.
3.6.5 Chemical
Many wet chemical procedures constitute the classical methods used for trace element
analysis of paniculate. In general, a color-forming reagent is used, and the amount of an
element is determined by the extent of color development. Probably the best known of these
procedures is based on the use of dithiocarbazone (dithizone)78 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 tune. Interferences can
also be a problem.
Methods for estimating the total mass of benzene-extractable organic material hi PM are
available. In this technique, a portion of the front-half catch from EPA Test Method 5 is
placed in a Soxhlet extractor and refluxed with benzene for several hours. The benzene is then
volatilized and the mass of residue is measured.
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 spectral detection. For organic compounds that are volatile up to about 300°C, gas
chromatography-mass spectrometry (GC-MS) can be used.79 For organic species with lower
volatility, liquid chromatography might be used. High-performance liquid chromatography
(HPLC)80 is typically used, but none of these procedures permits a high rate of analysis.
For analysis of one organic species of longstanding interest, benzo-a-pyrene (BaP), thin
layer chromatography (TLC) with fluorescence detection has been used. This procedure
requires a cyclohexane extraction, spotting, and development of a TLC plate, with
fluorescence detection. This TLC procedure is more interference-free than some HPLC
methods and has a higher yield rate.81
3-21
-------
3.7 References for Section 3
1. Martin, R.M. Construction Details of Isokinetic Source-Sampling Equipment
(APTD-0581). U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina. April 1971.
2. Rom, J.J. Maintenance, Calibration, and Operation of Isokinetic Source Sampling
Equipment (APTD-0576). U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina. March 1972.
3. Application Guide for Source PM10 Measurement with Exhaust Gas Recycle Sampling
System (EPA/600/3-88-058). U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina. 1988.
4. Application Guide for Source PMj0 Measurement with Constant Sampling Rate
(EPA/600/3-88-057). U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina. 1988.
5. Measurement of Condensible Vapor Contribution to PM10 Emissions
(EPA-600/D-89/103; NTIS PB89-224521). U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina. June 1985.
6. A Review of Current Methods for Measuring Particulate Matter Including Condensibles
from Stationary Sources (EPA-600/3-89/020; NTIS PB89-169973). U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina. February
1989.
7. Method Development and Evaluation of Draft Protocol for Measurement of
Condensible Particulate Emissions (EPA-450/4-90-012; NTIS PB90-24805).
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina. May
1990.
8. 1995 Annual Book of Standards, Section 11: Water and Environmental Technology,
Volume 11.03: Atmospheric Analyses. American Society of Testing and Materials,
Philadelphia, Pennsylvania. 1995
9. Determining the Properties of Fine Particulate Matter: Power Test Codes. American
Society of Mechanical Engineers, New York, New York. 1996.
10. Fissan, H. and D.F. Schulze-Froehlich. "A New Piezoelectric Quartz Crystal for
Particle Mass Determination" in Proceedings: Advances in Particle Sampling and
Measurement (EPA-600/9-89-004; NTIS PB89-166615), Daytona Beach, Florida,
October 1981. February 1989.
3-22
-------
11. Tacelli, R.W., R.E. Newton, and J.M. Andrews. "Triboelectric Instrument for Use as
a Continuous Particulate Emission Rate Monitor," in Continuous Emission Monitoring:
A Technology for the 90s, Proceedings of an International Specialty Conference. Air &
Waste Management Association, Pittsburgh, Pennsylvania. 1993.
12. Personal Communication. Ward, Tom, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, with Jones, Donna Lee, EC/R Inc., Durham,
North Carolina. August 27, 1996.
13. A Review of Methods for Measuring Fugitive PM-10 Emission Rates
(EPA/454-R-93-037; NTIS PB94-204203). Office of Air Quality Planning and
Standards, U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina. November 1993.
14. Protocol for the Measurement of Inhalable Particulate Fugitive Emissions from
Stationary Sources. Draft report prepared under EPA Contract 68-04-3115 by TRC
Environmental Consultants, Inc., Wethersfield, Connecticut, for Mr. D. Bruce Harris,
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina.
March 1980.
15. Techniques and Equipment for Measuring Inhalable Particulate Fugitive Emissions
(EPA-600/9-82-005d; NTIS PB-83-149617). U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina. 1982.
16. Technical Manual for the Measurement of Fugitive Emissions: Quasi-stack Sampling
Method (EPA-600/2-76-089c). U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina. 1976.
17. Technical Manual: Hood System Capture of Process Fugitive Emissions
(EPA-600/7-86-016; NTIS PB86-19044). U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina. 1986.
18. Evaluation of an Air Curtain Hooding System for a Primary Copper Converter, Volume
I (EPA-600/2-84-042a: NTIS PB84-160514). U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina. 1984.
19. Technical Manual for the Measurement of Fugitive Emissions: Roof Monitor Sampling
Method for Industrial Fugitive Emissions (EPA-600/2-76-089b). U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina. 1976.
20. Technical Manual for the Measurement of Fugitive Emissions: Upwind/Downwind
Sampling Method for Industrial Emissions (EPA-600/2-76-089a). U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina. 1976.
3-23
-------
21. Gengxin, H., X. Liguo, and H. Yanfeng. "A Study of Diffusion Models Applied to
Dust Emissions from Industrial Complexes." Environmental Monitoring and
Assessment. 22 (89-105). 1992.
22. Development of Emission Factors for Fugitive Dust Sources (EPA-450/3-74-037; NTIS
PB-238262). U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina. 1974.
23. Garman, G and G.E. Muleski. Example Test Plan for Point or Nonuniform Line
Sources. Prepared under EPA Contract 68-DO-0123, Work Assignment 11-44.
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina
1993.
24. Clayton, P., S.C. Wallin, B.J. Davis, and A.C. Simmonds. Methods for Determining
Paniculate Fugitive Emissions from Stationary Sources (NTIS PB-85-181717).
National Technical Information Service, Washington, DC. 1984.
25. Improved Emission factors for Fugitive Dust from Western Surface Coal Mining
Sources (EPA-600/7-84-048; NTIS PB84-170802). U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina. 1984.
26. Cowherd, C. A New Approach to Estimating Wind-Generated Emissions from Coal
Storage Piles. Air Pollution Control Association Specialty Conference Proceedings
(SP-51): Fugitive Dust Issues in the Coal Use Cycle. Air & Waste Management
Association, Pittsburgh, Pennsylvania. 1983.
27. Iron and Steel Plant Open Source Fugitive Emission Control Evaluation
(EPA-600/2-83-110; NTIS PB84-110568). U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina. 1983.
28. On the Use of SF6-Tracer Releases for the Determination of Fugitive Emissions (EPA-
600/9-82-005d; NTIS PB83-149617). U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina. 1982.
29. Frankel, R.G. Fugitive Dust Emissions: Accuracy of Gas Tracer Determinations.
Masters Report. University of North Carolina at Chapel Hill, Chapel Hill, North
Carolina. 1995.
30. Balloon Sampling to Characterize Particle Emissions from Fugitive Sources
(EPA-600/9-82-005d; NTIS PB83-149617). U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina. 1982.
31. Davies, C.N. The Entry of Aerosols hi Sampling Heads and Tubes. British Journal of
Applied Physics. 2:291. 1968.
3-24
-------
32. Shendrikar, A.D. and D.S. Ensor. "Sampling and Measurement of Trace Element
Emissions from Paniculate Control Devices," in Toxic Metals in the Atmosphere,
J.O. Nriagu and C.I. Davidson, Eds. John Wiley and Sons, New York, New York.
1986.
33. Pilat, M.J. D. S. Ensor, and J.C. Bosch. Source Test Cascade Impactor. Atmospheric
Environment. 4:671-679. 1970.
34. Marple, V.A. and B.Y.H. Liu. Characteristics of Laminar Jet Impactors.
Environmental Science and Technology. 8:648. 1974.
35. Felix, L.G., G.I. Clinard, G.E. Lacey, and J.D. McCain. Inertial Cascade Impactor
Substrate Media for Flue Gas Sampling (EPA-600/7-77-060). U.S. Environmental
Protection Agency, Washington, DC. 1977.
36. Pilat, M.J. G.A. Raemhild, E.B. Powell, G.M. Fiorette, and D.F. Meyer.
Development of Cascade Impactor System for Sampling 0.02 to 20-micron Diameter
Particles (FP-844, Volume 1). University of Seattle, Seattle Washington. 1978.
37. Nelson, P.A., D.S. Mummey, and W.D. Snowden. "Ultra-Fine Cascade Impactor
Particle Size Data Relationships to Opacity: Case Histories" in Proceedings: Advances
in Particle Sampling and Measurement (EPA-600/9-89-004; NTIS PB89-166615),
Daytona Beach, Florida, October 1981. February 1989.
38. Procedures for Cascade Impactor Calibration and Operation hi Process Streams
(EPA-600/2-77-004). U.S. Environmental Protection Agency, Washington, DC.
1977.
39. Source Assessment Sampling System: Design and Development (EPA-600/7-78-018).
U.S. Environmental Protection Agency, Washington, DC. February 1978.
40. Merrill, R.G., J. Lewtas, and R.E. Hall. "Source Assessment Sampling System
(SASS) Versus Dilution Tunnel Sampling," hi Proceedings: Advances in Particle
Sampling and Measurement (EPA-600/9-89-004; NTIS PB89-166615), Daytona Beach,
Florida, October 1981. February 1989.
41. Test Methods for Evaluating Solid Waste, Third Edition. Report No. SW-848. U.S.
Environmental Protection Agency, Office of Solid Waste and Emergency Response,
Washington, DC. 1986.
42. Smith. W.B. and R.R. Wilson. Development and Laboratory Evaluation of a Five-
Stage Cyclone System (EPA-600/7-78-008). U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina. 1978.
3-25
-------
43. Smith, W.B. and R.R. Wilson. A Five-Stage Cyclone System for In Situ Sampling.
Environmental Science and Technology. 13(11). November 1979.
44. Agarwal, J.K. and R.J. Remiarz. "Development of an Aerodynamic Particle Sizer" in
Proceedings: Advances in Particle Sampling and Measurement (EPA-600/9-89-004;
NTIS PB89-166615), Daytona Beach, Florida, October 1981. February 1989.
45. J.C. Wilson and B.Y.H. Liu. Aerodynamic Particle Size Measurement by Laser-
Doppler Velocimetry. J. Aerosol Science. 11:139-150. 1980.
46. Agarwal, J.K. and L.M. Fingerson. Real-Time Aerodynamic Particle Size
Measurement with a Laser Velocimeter. TSI Quarterly. V(l). 1979.
47. Roth, P. and V. Filippov. In Situ Ultrafine Particle Sizing by a Combination of Pulsed
Laser Heatup and Particle Thermal Emission. J. Aerosol Science. 27(1):95-104.
1996.
48. Buttle, D.J., S.R. Martin, and C.B. Scruby. Particle Sizing by Quantitative Acoustic
Emission. Research in Nondestructive Evaluation. 3(1): 1-26. 1981.
49. M.F. Leach, G.A. Rubin, and J.C. Williams. Particle Size Distribution
Characterization from Acoustic Emissions. Powder Technology. 19(2): 157.
March/April 1978.
50. Correspondence and product literature from Mr. Hugh North, Malvern/Insitec, San
Ramon, California, to Mr. Kenneth Woodard, U.S. EPA. January 14, 1998.
51. Knapp, K.T. "Elemental Composition of Sized Profiles Emitted from Stationary
Sources," hi Proceedings of the International Symposium on Recent Advances in
Paniculate Science and Technology held on December 8-10, 1982, in Madras, India.
Department of Chemical Engineering, Indian Institute of Technology, Madras, India.
1986.
52. Chen, B.T., Y.S. Cheng, and H.C. Yeh. Tests of the Size resolution and Size
Accuracy of the Lovelace Parallel-Flow Diffusion Battery. Am. Ind. Hyg. Assoc. J.
52(2):75. February 1, 1991.
53. Barr, E.B., Y.S. Cheng, and H.C. Yeh. Size Characterization of Carbonaceous
Particles Using a Lovelace Multijet Cascade Impactor/Parallel-Flow Diffusion Battery
Serial Sampling Tram. Aerosol Science and Technology. 10(1):205. 1989.
54. Hinds, W.C. Aerosol Technology: Properties, Behavior, and Measurement of
Airborne Particles. John Wiley and Sons, New York, New York. 1982.
3-26
-------
55. "Particle Sizing-Past and Present" (Particle Sizing Editorial Review). Filtration and
Separation. July/August 1993.
56. McCain, J.D., W.G. Kistler, J.W. Ragland, R.L. Merritt. "Aerosol Size and
Chemistry at a Coal-Fired Power Plant" in Proceedings: Advances in Particle
Sampling and Measurement (EPA-600/9-89-004; NTIS PB89-166615), Daytona Beach,
Florida, October 1981. February 1989.
57. Gushing, K.M., R.R. Wilson, W.E. Farthing, and D.B. Harris. "A Comparison of
Several Particle-Sizing Techniques" hi Proceedings: Advances in Particle Sampling
and Measurement (EPA-600/9-89-004; NTIS PB89-166615), Daytona Beach, Florida,
October 1981. February 1989.
58. Steele, W.J., W.B. Smith, W.E. Farthing, and R.C. Carr. "A Sensitive Paniculate
Monitor for Measuring the Approximate Opacity for Fabric Filter Systems" hi
Proceedings: Advances hi Particle Sampling and Measurement (EPA-600/9-89-004;
NTIS PB89-166615), Daytona Beach, Florida, October 1981. February 1989.
59. Sadakata. M., M. Motegi, Y. Mashio, T. Ishada, A. Harano, and H.J. Kim. Growth
of Primary Soot Particles hi an Atmospheric Premixed Methane Oxygen Flame.
Chemical Engineering Research and Design, Part A: Transactions of the Institute of
Chemical Engineers. 73(42): 142-146. March 1995.
60. Huggins, F.E. D.A. Kosmack, G.P. Huffman, and R.J. Lee. "Coal Mineralogies by
SEM Automatic Image Analyses," in Scanning Electron MicroscopyII. SEM Inc.,
AMF O'Hare, Chicago, Illinois. 1980.
61. Davidson, D.L. "The Analysis of particles by Electron Microscopy and
Spectroscopy," hi Panicle Characterization in Technology, Vol. I: Applications and
Microanalysis, J.K. Beddow, Ed. CRC Press, Inc., Boca Raton, Florida. 1984.
62. van de Plaats, G., H. Harps, and L. Willeams. Size Determination of Conductive
Particles with a Coulter® Counter. Proceedings of the Particle Size Analysis
Conference, Loughborough Tech. University, Loughborough, England. September
1981.
63. Berg, R.H. Electronic Size Analysis of Subsieve Particles by Flowing Through a Small
Liquid Resistance (STP No. 234). American Society for Testing and Materials,
Philadelphia, Pennsylvania. 1958.
64. Karuhn, R.F. and R.H. Berg. "Practical Aspects of Electrozone® Size Analysis," in
Particle Characterization in Technology, Vol. I: Applications and Microanalysis,
J.K. Beddow, Ed. CRC Press, Inc., Boca Raton, Florida. 1984.
3-27
-------
65. Kinsman, S. "Particle Size Instrumentation-Coulter® Counter," in Panicle
Characterization in Technology, Vol. I: Applications and Microanalysis, J.K. Beddow,
Ed. CRC Press, Inc., Boca Raton, Florida. 1984.
66. Corbin, R.G., J.K. Horrocks, D. Towell, and C. Wainwright. Measurement of
Particle Size and Charge from an Electrostatic Filter. Filtration and Separation.
24(4):248-251. July-August 1988.
67. Multiple Metals Stack Emissions Measurement Methodology (EPA/600-A-93/089;
NTIS PB93-185734). U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina. 1993.
68. "Methodology for the Determination of Metals Emissions in Exhaust Gases from
Hazardous Waste Incineration and Similar Combustion Sources," in Test Methods for
Evaluating Solid Waste: Physical/Chemical Methods (SW-846, Third Edition).
U.S. Environmental Protection Agency, Office of Solid Waste and Emergency
Response, Washington, D.C. September 1988.
69. "Methodology for the Determination of Metals Emissions from Hazardous Waste
Incineration and Similar Combustion Sources," in Methods Manual for Compliance
with the BIF Regulations Burning Hazardous Waste in Boilers and Industrial Furnaces
(EPA/530-SW-91-010). U.S. Environmental Protection Agency, Office of Solid Waste
and Emergency Response, Washington, DC. December 1990.
70. Screening Methods for the Development of Air Toxics Emission Factors
(EPA-450/4-91-021). U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina. September 1991.
71. Ahearn, A.I. Trace Analysis by Mass Spectrometry. Academic Press, New York,
New York. 1972.
72. Fassel, V. A. Quantitative Elemental Analyses by Plasma Emission Spectroscopy.
Science, Vol. 202, 1978.
73. Morrison, G. H. Trace Analysis Physical Methods. Wiley Interscience, New York,
1965.
74. Kaufman, R. and P. Wieser. "Laser Microprobe Mass Analysis in Particle Analysis,"
in Panicle Characterization in Technology, Vol. I: Applications and Microanalysis,
J.K. Beddow, Ed. CRC Press, Inc., Boca Raton, Florida. 1984.
75. Priorities and Procedures for Development of Standards of Performance for New
Stationary Sources of Atmospheric Emissions (EPA-450/3-76-020).
U.S. Environmental Protection Agency, Research Triangle Park, NC. May 1976.
3-28
-------
76. 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. 51. 1979.
77. Goold, R.W., C.S. Barrett, J.B. Newkirk, and C.O. Ruud. Advances in X-ray
Analysis. Kendall/Hunt, Dubuque, Iowa. 1976.
78. Snell, F.D. Photometric and Fluorometric Methods of Analysis: Metals, Part I. John
Wiley, New York, New York. 1978.
79. McFadden, W.H. Techniques of Combined Gas Chromatography/Mass Spectrometry.
John Wiley, New York, New York. 1978.
80. Kirkland, J.J. Modern Practice of Liquid Chromatography. John Wiley, New York,
New York. 1971.
81. Swanson, D., et al. A Rapid Analytical Procedure for the Analysis ofBaP in
Environmental Samples. Trends Fluoresc. 1. 1978.
3-29
-------
4. FUEL SUBSTITUTION AND SOURCE REDUCTION APPROACHES FOR
PARTICIPATE MATTER
This section discusses the use of fuel substitution and process optimization to achieve
reductions in PM emissions. In many cases, these practices can be easier and less expensive
than upgrading existing control technology or investing in new add-on controls.
Fuel substitution, or fuel switching, is typically used as a means of reducing emissions
from combustion sources, such as electric utilities and industrial boilers. It involves replacing
the current fuel with a fuel which emits less of a given pollutant when burned. Common
examples of this would be replacing coal with oil or natural gas at an electric utility plant.
Source reduction techniques generally consist of modifying or optimizing a given process to
improve its operation, since many PM emissions are the result of processes which are not
performing to their potential. These emissions can be reduced or eliminated by altering the
process.
4.1 Fuel Substitution
Fuel substitution can be an effective means of reducing emissions for many types of
processes which use fuel combustion to provide heat for the process or to produce electricity.
Fuel combustion is responsible for significant emissions of PM10 and PM2.5, as well as SOX and
NOX. Control devices, such as fabric filters and electrostatic precipitators are often the first
option for PM control for fuel combustion sources. However, add-on PM controls can require
a very large capital investment.
The type of fuel and process have a great impact on the PM emissions from
combustion. Coal, oil, and natural gas are the most common fuels used. Of these fuels, coal
combustion generally results hi the highest PM emissions. The four major types of coal are
bituminous, subbituminous, anthracite, and lignite;1 their characteristics and emissions are
very different. Oil is broadly classified as residual or distillate. Residual oils contain more
sulfur and ash which contribute to higher emissions. Fuel oils are also described by numbers.
Numbers 1 and 2 fuel oils are distillate, Nos. 5 and 6 are residual, and No. 4 fuel oil can be
distillate or a mixture of residual and distillate.1 Natural gas is a relatively clean-burning fuel
and typically results hi much less PM than oil or coal.1
4.1.1 Applicability
There are several considerations to determine if fuel switching is the best option for
reducing emissions from a given combustion source. For many older boilers, the expense
associated with new add-on PM controls or modifications to existing controls is not practical.
Fuel switching is an especially attractive option for these boilers because the capital investment
is usually small when compared to that of control devices.
4-1
-------
For fuel substitution to be practical, there must be a suitable replacement fuel available
at an acceptable cost. Prospective fuels must be evaluated using the criteria of performance,
availability, and cost.2 The first requirement is that the replacement fuel provides a significant
reduction in emissions versus the original fuel. The effect that the replacement fuel has on
emissions of other pollutants should be considered as well. For example, switching to low
sulfur coal to reduce SOX emissions may increase PM emissions.2 In many cases fuel
substitution will reduce more than one type of pollutant. For instance, substituting natural gas
for coal will reduce PM emissions and virtually eliminate SOX.
While most industrialized areas have access to a variety of fuels, some fuels may not be
practical in certain locations because of cost. Natural gas and fuel oil are generally supplied
by pipeline. Locations which are not near existing pipelines may find it expensive to arrange
for a natural gas or fuel oil supply. Smaller industrial or commercial units can rely on
delivery by truck. Since coal is typically supplied by railroad and the characteristics of coal
from different areas of the country vary widely, some types of coal may not be applicable as a
replacement fuel for a given location because they must be shipped from an unreasonable
distance.
In most cases, the process will have to be modified to accommodate switching to a
different type of fuel. For certain types of coal fired boilers, such as stokers, it may be
impractical to retrofit them to burn a liquid or gaseous fuel. Fuel switching will often require
retrofitting the current control device in addition to the process. Fuel substitution therefore,
would not be applicable to sources with excessive retrofit costs.
In addition to the requirements for source modification, fuel prices can be a
determining cost factor for fuel switching. Since coal, oil, and natural gas all have different
prices based on their heating values, fuel switching may also increase operating costs. The
costs of fuel substitution will be discussed further in section 4.1.3.
4.1.2 Emission Reductions with Fuel Switching
If fuel substitution is applicable to a given combustion process, it can result in
significant reductions in PM emissions. In general, PM and SO2 emissions are highest for coal
and lowest for natural gas. Tables 4.1 and 4.2 show potential PM10 and PM2.5 emission
reductions, respectively, with fuel switching. The tables provide matrices showing the
approximate emissions reductions for switching from bituminous coal to subbituminous coal,
from coal to oil, and from oil to gas. Distillate oil was not provided as a replacement fuel for
utility sources because it is not typically burned in utility boilers.
The emission reductions were calculated based on emission factors and fuel
composition.1 Emission factors are dependent on the type of fuel and the type of combustion
process which is employed. The potential reductions achieved by switching from bituminous
and subbituminous coal in Tables 4.1 and 4.2 were based on emissions from dry bottom
4-2
-------
Table 4.1. Potential PM10 Emission Reductions with Fuel Switching (References 1, 4, and 5)
u>
Estimated PM)0 Reductions with Replacement Fuel (percent)
Industrial Utility
Original Fuel Subbituminous
Bituminous Coal 21.4
Subbituminous Coal
Residual Oil3
Residual Oil includes No. 4, 5,
Distillate Oil is No. 2 fuel oil.
Table 4. 2. Potential
Residual Oil • Natural Gas Distillate Oil Subbituminous Residual Oil
62.9 98.2 99.0 21.4 69.5
52.8 97.7 98.8 -- 61.2
95.1 97.4
and 6 fuel oil.
PM25 Emission Reductions with Fuel Switching (References 1,4, and
Natural Gas
99.3
99.2
97.9
5)
Estimated PM2 5 Reductions with Replacement Fuel (percent)
Industrial Utility
Original Fuel Subbituminous
Bituminous Coal 21.4
Subbituminous Coal
Residual Oil'
Residual Oil Natural Gas Distillate Oil Subbituminous Residual Oil
7.4 93.1 99.1 21.4 14.8
91.2 98.8
92.5 99.0
Natural Gas
97.5
96.8
97.0
a Residual Oil includes No. 4, 5, and 6 fuel oil.
b Distillate Oil is No. 2 fuel oil.
-------
boilers, since dry bottom boilers are responsible for the highest percentage of PM emissions
from coal combustion.3 In addition, dry bottom boilers have lower PM10 and PM2 5 emission
factors than wet bottom boilers and all types of stokers.1 Emission factors for utility and
industrial coal combustion in the same type of boilers were assumed to be the same.1
Variation in PM emissions from oil combustion is due to differences in utility and industrial
units. Utility units tend to operate more efficiently than industrial units and, therefore, have
lower PM emissions. This is also the case with natural gas combustion.1
In terms of fuel composition, the ash content of the fuel is a major factor in
determining PM emissions. In general, the higher ash content a given fuel, the more PM will
be emitted when burned.1 For these calculations, an average value of 8.6 weight percent ash
(6.62 lb/106 Btu) in bituminous coal was used. The average ash content of subbituminous coal
was assumed to be 5.2 percent (4 lb/106 Btu).4 Paniculate emissions from oil combustion are
dependent on ash and sulfur content.1 Increasing sulfur content will increase PM emissions
from oil combustion because the sulfur inhibits complete combustion.1 Also, a small
percentage (1 % to 3%) of the sulfur hi oil is emitted as sulfate paniculate.1 Residual oil was
estimated to have ash content of 0.03 weight percent (0.016 lb/106 Btu) and sulfur content of
2.5 percent (1.3 lb/106 Btu).4 Distillate oil was estimated to have ash content of less than 0.01
percent (<0.005 lb/106 Btu) and sulfur content of 0.22 percent (0.115 lb/106 Btu).4 A sample
calculation of the potential PM10 emission reduction associated with switching from bituminous
coal to distillate oil follows.
The PMj0 emission factors for bituminous coal and distillate oil combustion in dry
bottom boilers are 2.3(A) Ib/ton and 1 lb/103 gallons, respectively.1 In the coal emission
factor, (A) refers to the ash content of the fuel. Because these factors are based on tons of
coal and gallons of oil, they must be converted into factors based on the heating value of the
fuel in order to be useful. This is done by dividing the emission factor by the heating value:
Bituminous coal: (2.3 lb/ton)(8.6)/(26,000,000 Btu/ton) = 0.761 lb/106 Btu
Distillate oil: (1 lb/103 gal)/(138,000,000 Btu/103 gal) = 0.007 lb/106 Btu
The potential reduction in PM10 emissions when switching from bituminous coal to
distillate oil is calculated by subtracting the emission factor for oil (EFoil) from the emission
factor for coal (EF^) and then dividing by the coal emission factor:
Potential reduction = [(EF^J - (EFoil)]/(EFcoal)
= [(0.761 lb/106 Btu) - (0.007 lb/106 Btu)]/ (0.761 lb/106 Btu)
= 0.99 or 99 percent
Tables 4.1 and 4.2 indicate that the maximum reductions in PMi0 and PM25 emissions
can be obtained by switching to from coal or residual oil to natural gas or distillate oil. The
reductions presented hi these tables were based on the average values discussed above; actual
reductions will vary with specific fuel composition.
4-4
-------
4.1.3 Costs
The costs associated with fuel substitution are related to retrofitting the current unit and
purchasing the replacement fuel. Retrofitting the combustion process to burn another fuel can
be a major undertaking, with the necessary modifications unique to each site. Generally,
switching from one kind of coal or grade of oil to another is less costly than switching from
coal to oil or natural gas. In some cases, the cost of modifying the combustion process to
utilize the new fuel makes the fuel substitution unpractical. Another possible retrofit cost is
related to the existing control devices, which also may require modifications to accommodate
the type of emissions associated with the new fuel.
A cost differential may also exist between fuels. If the replacement fuel is much more
expensive than the fuel which is currently in use, operating costs may noticeably increase.
Table 4.3 provides average prices for coal, oil, and natural gas, in terms of common units and
heating value.5-6-7 These prices will vary depending on the actual location hi the U.S.
Table 4.3. Average Prices of Coal, Oil, and
Natural Gasa (References 5, 6, and 7)
Average Price
Fuel
Subbituminous coalb
Utility
Industrial
Bituminous coalb
Utility
Industrial
Residual oil0
Distillate oil0
Natural gas0
Common Price
$27.01/ton
$32.37/ton
$27.01/ton
$32.37/ton
$390/103 gal
S616/103 gal
$1680/106 scf
$/MMBtu
1.35
1.62
1.04
1.25
2.60
4.47
1.68
a More current prices are available from the Monthly Energy
Review, published by the U.S. Department of Energy,
Washington, DC.
b 1995 average prices.
0 Average prices as of September, 1996.
4-5
-------
4.1.4 Other Impacts
In addition to reducing PM emissions, fuel substitution can also reduce emissions of
other pollutants, such as SOX and NOX. Potential SOX reductions with fuel switching are
provided in Table 4.4. Natural gas is especially effective for SOX control, eliminating nearly
100 percent of SOX. Coal burning power plants have been switching to Western coal as a
means of reducing SOX emissions,2 since western coals have lower sulfur contents than many
otherwise comparable Eastern coals. Unfortunately, low sulfur ash is more difficult to collect
hi ESPs, so that switching to Western coal will usually require flue gas conditioning or a
control device modification to maintain PM collection efficiency.2
Substituting natural gas for coal has been shown to be effective at reducing NOX
emissions. In 1992, Public Service Electric and Gas (PSE&G) demonstrated seasonal control
of NOX emissions by operating two utility boilers with natural gas instead of coal for the
3-month ozone season (June, July, and August).8
Table 4.4. Potential SOX Reductions with Fuel Switching
Estimated SOX
Subbituminous
Original Fuel Coal
Bituminous Coal 72.9
Subbituminous Coal
No. 6 Fuel Oild
No. 4 Fuel Oilb
No. 2 Fuel Oilc
Reductions
Lignite
Coal2
80.2
26.9
—
—
—
with Replacement Fuel (percent)
No. 4
Fuel Oil"
47.4
—
46.2
—
—
No.2
Fuel Oilc
91.2
69.5
91.5
84.3
—
Natural
Gas
99.9
99.9
99.9
99.9
99.7
a Lignite coal with high sodium ash content and sulfur content of 0.4 percent by weight.
b Distillate/residual mixture with average sulfur content of 1.35 percent by weight.
c Distillate oil with average sulfur content of 0.22 percent by weight, typically not used
hi utility boilers.
d Residual oil with average sulfur content of 2.5 percent by weight.
4-6
-------
4.2 Process Modification/Optimization
Process modification and/or optimization can be an effective means of reducing PM
emissions. Some general examples of process optimization include reducing the frequency of
mass transfer operations, improving operational efficiency, and the proper use of dust
collection devices at the point of generation.
Manufacturing can require many individual process steps involving simple functions.
Material transfer steps can cause fugitive PM emissions and costly loss of product. A careful
analysis of all process steps may reveal some unnecessary or repetitive steps which can be
eliminated, resulting in fewer fugitive PM emissions.9
Particle characteristics can also have a significant impact on PM emission rates.
Particle size has a direct effect in that larger particles settle more quickly and are more easily
collected hi control devices. Therefore, wetting and agglomeration techniques in general
increase particle size and the efficiency of control equipment.9 The performance of some
control devices, such as ESPs, is also influenced by the chemical composition of the particles.
Flue gas conditioning (see Section 5.1) is a means of altering the composition of particles and
improving the conditions for electrostatic precipitation.
are:
Some specific process modification/optimization techniques to reduce PM emissions
Changing from a cupola to an electric arc furnace.9
Changing from an (open) bucket elevator to more efficient (closed) pneumatic
conveyor.9
Screening out undersized coke (< 1 inch) to reduce blast furnace fugitive
emissions in primary metal smelting.10
Improving blast furnace combustion efficiency during primary lead smelting by
improving the furnace water cooling system.10
Eliminating fugitive PM from transporting, pouring, and stirring molten metal
by the use of continuous kettle dressing rather than manual hi primary metal
smelting (as is currently done hi only foreign facilities).10
Improving raw material quality, e.g. improve the quality of coke and suiter
concentrate used hi primary metal production.10
Cooling metal pots to reduce fume generation during kettle dressing hi primary
metal production.10
4-7
-------
Pumping (primary) metal directly to dross kettles using an electromagnetic
pump.10
Agglomerating blast furnace flue dust in an agglomerating furnace to reduce the
load on the fabric filter to improve its performance. This process completely
eliminates handling of the dust and the associated fugitive emissions, and
eliminates fugitive emissions from flue dust storage piles.10
Using permanent mold castings in gray iron foundries instead of green sand.
This is reported to reduce PM emissions by 99 percent.10
Pre-treating glass manufacturing raw materials to reduce the amount of fine
particles. Pretreatments include: presintering, briquetting, pelletizing, or liquid
alkali treatment.10
Replacing grease and oil lubricants (e.g. in glass manufacturing) with silicone
emulsions and water-soluble oils that eliminate the smoke generated from flash
vaporization of hydrocarbons from greases and oils that come into contact with
process materials.10
Tuning industrial boilers to achieve more efficient combustion to reduce PM
that occurs as a result of incomplete combustion.
ABB Industry Oy of Finland's Burning Image analyZER (BIZER) that allows
combustion control for kraft pulp mill recovery boilers. In this process,
infrared cameras view the smelt pile and provide digital image processing to
present burning information. This technique can be used for automatic burning
control.11
Hitachi, Ltd., of Japan's Oilless, Dry Centrifugal "Leak-Free" Compressors
eliminate fugitive leaks and save energy. The PM reduction is achieved through
energy efficiency. This device is currently being used at petroleum refineries.11
Poland has developed a coal pyrolysis technique that produces a better fuel.
Crushed dried coal is decomposed into gas and char hi a circulating fluidized
bed reactor. The gas is burned in a turbine and the char is mixed with coal and
pressed into briquettes of smokeless fuel called ECOCOAL. ECOCOAL has
1.2 to 1.7 times the thermal efficiency of coal with PM emissions up to
50 percent lower.11
Lurgi Metallurgies (FRG) QSL Process for Secondary Lead Smelting. Use of
a completely closed reactor designed to treat all grades results in > 90 percent
control of PM.11
4-8
-------
Fluidized-bed heat treatment technology for primary metal manufacturing
developed by Quality Heat Treatment Pty, Ltd., of Australia. A gas-phase heat
treatment process uses a fluidized bed of alumina particles and is completely
enclosed, enabling collection of fugitive PM emissions.11
Dow Chemical Ferroalloy Process. Pure oxygen is used instead of ah" hi a
closed furnace that produces no dust.
4.3 References for Section 4
1. Compilation of Air Pollutant Emission Factors (AP-42). volume I (Fifth Edition).
U.S. Environmental Protection Agency, Research Triangle Park, NC. January 1995.
2. Rembold, D. and R. Rupinkas. CAA Compliance: Fuel Switching Can Be Complex,
Yet Rewarding. Electrical World. April 1994.
3. Source Category Emission Reductions with Paniculate Matter and Precursor Control
Techniques. Prepared for K. Woodard, U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina (AQSSD/IPSG), under Work Assignment 11-16
(EPA Contract No. 68-03-0034), "Evaluation of Fine Paniculate Matter Control."
September 30, 1996.
4. Perry, R.H. and D.W. Green. Perry's Chemical Engineers' Handbook (6th Edition).
McGraw-Hill Publishing Company, Inc., New York, New York. 1984.
5. Quarterly Coal Report, October-December 1995 (DOE/EIA-0121(95/4Q)). Energy
Information Administration, Department of Energy, Washington, DC. May 1996.
6. Electric Power Monthly, April 1996 (DOE/EIA-0226(96/04)). Department of Energy,
Washington, DC. April 1996.
7. Industry Statistics. Oil & Gas Journal. September 16, 1996.
8. Control Costs for VOC and NOx Measures for Non-Traditional Sources for Ozone
NAAQS Review. Final Report for Work Assignment No. 06; EPA Contract
No. 68-D3-0034. U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina. September 30, 1994.
9. Control Techniques for Paniculate Emissions from Stationary Sources - Volume 1
(EPA-450/3-18-005a, NTIS PB83-127498). U.S. Environmental Protection Agency,
Office of Ah- Quality Planning and Standards. Research Triangle Park, North
Carolina. September 1982.
4-9
-------
10. Estimating and Controlling Fugitive Lead Emissions from Industrial Sources
(EPA-452/R-96-006). U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina. May 1996.
11. Assessment of International Air Pollution Prevention and Control Technology for Title
IX of the Clean Air Act Amendments. Draft Report to Congress. Office of Research
and Development, U.S. Environmental Protection Agency, Washington, DC. July
1996.
4-10
-------
5. EXHAUST GAS CLEANING SYSTEMS FOR STATIONARY SOURCES
This section discusses the exhaust gas cleaning systems for stationary sources that can
be used in industries with particulate-bearing exhaust streams. The devices discussed are
electrostatic precipitators (ESPs), fabric filters, wet scrubbers and incinerators (used for
streams with especially high VOC contents as well as PM). The section begins with a
discussion of pretreatment devices that can be used to reduce the PM loading onto the primary
control device and flue gas conditioning which can enhance particle collection.
Each section includes a description of the device, the collection mechanisms, and
discussion of different systems designs. Also included is a discussion of the applicability of
the device to the various processes to which the controls can be applied. The effectiveness of
the device, hi terms of the range of efficiencies for various types of systems and applications,
is then discussed, with a special focus on PMj0 and PM2 5. General curves relating particle
size and efficiency are included along with a discussion of the parameters influencing
efficiency and their quantitative impacts. Two types of efficiency, fractional and cumulative,
are discussed in this section. Fractional efficiency refers to the efficiency of a control device
for a particular size of particle only, such as 10 /-on hi aerodynamic diameter. Cumulative
efficiency is the efficiency of a control device for a particular particle size and all the particles
smaller that size particle, such as PMi0, which includes all particles with aerodynamic
diameters of 10 /mi or smaller.
Costs of the devices are discussed, and include cost tables and curves derived using
standard EPA protocol. A discussion of the parameters affecting cost and the relationship
between costs and these parameters is included hi each cost section. In many cases, the costs
have been updated to fourth quarter 1996 using the Vatavuk Air Pollution Control Cost Index
(VAPPCI). This index is provided hi Appendix B. The index is also published monthly hi
Chemical Engineering. Finally, the energy and other secondary environmental impacts, such
as water pollution and waste generation, are discussed along with potential mitigation
measures.
5-1
-------
5.1 PRETREATMENT
The performance of paniculate control devices can often be improved through
pretreatment of the gas stream. For PM control devices, pretreatment consists of two
categories: precollection and flue gas conditioning. Precollection devices remove large
particles from the gas stream, reducing the loading on the primary control device. Gas
conditioning techniques alter the characteristics of the particles and/or the gas stream to allow
the primary control device to function more effectively. Both types of pretreatment can lead to
increased collection efficiency and operating life, while reducing operating costs. The
performance of precollection devices is discussed in Section 5.1.2.
5.1.1 Precollection Devices
The vast majority of precollection devices are mechanical collectors. Mechanical
collectors are a class of devices that rely on gravity and inertia for particle collection. They
are used extensively in industry because of several advantages they possess. Mechanical
collectors have low capital costs, the ability to operate in harsh environments, and low
maintenance requirements because they lack moving parts.1 There are also disadvantages
associated with mechanical collectors, such as the relatively low collection efficiencies for
small particles. While this does prevent their use as primary collection devices in many
applications, it is not a major concern when mechanical collectors are used for precollection.
Some mechanical collectors can achieve high collection efficiencies, but only with the high
operating costs associated with large pressure drops.1 The five major types of mechanical
collectors are settling chambers, elutriators, momentum separators, mechanically aided
collectors, and centrifugal separators (cyclones); these devices are discussed separately below.
5.1.1.1 Settling Chambers
The simplest mechanical collectors are settling chambers, which rely on gravitational
settling as a collection mechanism. Settling chambers prevent excessive abrasion and dust
loading in primary collection devices by removing large particles from the gas stream.2
Despite low collection efficiencies, settling chambers are still used extensively. They are
particularly useful for industries that also need to cool the gas stream prior to treatment hi a
fabric filter. The mineral products and metals processing industries have several applications
for settling chambers. There are two primary types of settling chambers: the expansion
chamber and the multiple-tray chamber. In the expansion chamber, the velocity of the gas
stream is significantly reduced as the gas expands into a large chamber. The reduction in
velocity allows larger particles to settle out of the gas stream.3 Figure 5.1-1 shows a
schematic diagram of an expansion chamber, which consists of a simple chamber with
collection hoppers.2
A multiple-tray settling chamber, shown in Figure 5.1-2, is an expansion chamber with
a number of thin trays closely spaced within the chamber, which causes the gas to flow
5.1-1
-------
Dust-laden
gas
Hoppers
Cleaned gas
Figure 5.1-1. Expansion Settling Chamber (Adapted from Reference 2).
5.1-2
-------
Dust-laden
gas
Cleaned gas
Figure 5.1-2. Multiple-Tray Settling Chamber (adapted from Reference 2).
5.1-3
-------
horizontally between them.2 While the gas velocity is increased slightly in a multiple-tray
chamber, the collection efficiency generally improves because the particles have a much
shorter distance to fall before they are collected. An expansion chamber must be very large to
collect any small particles, but multiple-tray chambers have lower volume requirements for the
collection of small particles (sd5 /an).3
5.1.1.2 Elutriators
Like settling chambers, elutriators also rely on gravitational settling to collect particles.
An elutriator is made up of one or more vertical tubes or towers in series, where the gas
stream passes upward through the tubes. Larger particles whose terminal settling velocity is
greater than the upward gas velocity are collected at the bottom of the tube, while smaller
particles are carried out of the top of the tube. Size classification of the collected particles can
be achieved by using a series of tubes with increasing diameters, as shown in Figure 5.1-3.2-3
5.1.1.3 Momentum Separators
Momentum separators utilize both gravity and inertia to separate particles from the gas
stream. Separation is accomplished by forcing the gas flow to sharply change direction within
a gravity settling chamber through the use of strategically placed baffles. Typically, the gas
first flows downward and then is forced by the baffles to suddenly flow upwards. Inertial
momentum and gravity act hi the downward direction on the particles, which causes larger
particles to cross the flow lines of the gas and collect hi the bottom of the chamber.2-3
There are several common arrangements of baffles hi momentum separators, as
illustrated in Figure 5.1-4.2-3 Momentum separators are capable of collecting particles as small
as 10 fj.m at low efficiency (10-20 percent). These devices require less space than gravity
settlers, but have higher pressure drops.3
5.1.1.4 Mechanically-Aided Separators
Mechanically-aided separators rely on inertia as a separation mechanism. The gas
stream is accelerated mechanically, which increases the effectiveness of the inertia separation.
As a result, mechanically-aided separators can collect smaller particles than momentum
separators. Unfortunately, they also have higher operating costs as a result of higher pressure
drops. A common type of mechanically-aided collector is the modified radial blade fan,
shown in Figure 5.1-5.3 In this device, the gas stream enters at the center of the fan,
perpendicular to the blade rotation. The blades propel the particles across the gas flow lines,
where they are concentrated on the inside wall of the casing. From there, the particles are
diverted into a collection hopper while the gas continues out of the separator. Mechanically-
aided separators are subject to abrasive wear from large particles and clogging from particles
which cake or accumulate on the blades. Consequently, these devices have higher
maintenance requirements than other separators.2>3>4
5.1-4
-------
Dust-laden
gas
ft
V
^n
t
1
T
^
Cleaned gas
Large
particles
Medium
panicles
Small
particles
Figure 5.1-3. Elutriators in Series (Reference 3).
5.1-5
-------
Dust-Laden
Gas
Collected Dust
Dust-Laden
Gas
T T
Collected Dust
Dust-Laden
Gas
Collected Dust
Figure 5.1-4. Momentum Separators (References 2 and 3).
5.1-6
-------
Figure 5.1-5. Mechanically-Aided Separator (Reference 3).
5.1-7
-------
5.1.1.5 Cyclones
Cyclones use inertia to remove particles from a spinning gas stream. Within a cyclone,
the gas stream is forced to spin within an usually conical-shaped chamber. Cyclones operate
by creating a double vortex inside the cyclone body. The incoming gas is forced into circular
motion either by tangential inlet or by turning vanes in the axial inlet. The gas spirals down
the cyclone near the inner surface of the cyclone tube. At the bottom of the cyclone, the gas
turns and spirals up through the center of the tube and out of the top of the cyclone. Figure
5.1-6 illustrates the double vortex operation hi a cyclone.1
Particles in the gas stream are forced toward the cyclone walls by the centrifugal force
of the spinning gas, but are opposed by the fluid drag force of the gas traveling through and
out of the cyclone. For particles that are large, inertial momentum overcomes the fluid drag
force so that the particles reach the cyclone walls and are collected; while for smaller particles,
the fluid drag force overwhelms the inertial momentum and causes these particles to leave the
cyclone with the exiting gas. Gravity also causes the larger particles that reach the cyclone
walls to travel down into a bottom hopper. While they rely on the same separation mechanism
as momentum separators, cyclones are more effective because they have a more complex gas
flow pattern.2-3
Cyclone collectors are generally classified into four types, based on how the gas stream
is introduced and how the collected dust is discharged:
• Tangential inlet, axial discharge
• Axial inlet, axial discharge
• Tangential inlet, peripheral discharge
• Axial inlet, peripheral discharge
The first two types are the most commonly used cyclones. Schematic diagrams of the four
types of cyclones are provided hi Figure 5.1-7.2-3
Cyclone collectors can be designed for many applications, and they are typically
categorized as high efficiency, conventional, or high throughput. High efficiency cyclones are
likely to have the highest pressure drops of the three cyclone types; high throughput cyclones
can treat large volumes of gas with a low pressure drop.1-5 Each of these three cyclone types
have the same basic design. Different levels of collection efficiency and operation are
achieved by varying the standard cyclone dimensions, identified hi Figure 5.1-8,6 according to
the values7-8-9 shown hi Table 5.1-I.1-6
5.1-8
-------
Cleaned gas out
Tangential
inlet duct
Dusty
gas in
t
' x— Vortex-finder tube
•Gas flow path
Dust out
Figure 5.1-6. Illustration of the Double Vortex Within a Cyclone (Reference 1).
5.1-9
-------
Dust-laden gas
Tangential inlet.
Peripheral discharge
Dust
Dust-laden gas
Clean gas
Axial inlet.
Axial discharge
Dust
Dust-laden gas
Clean gas
Axial inlet.
Peripheral discharge
Dust
, dean gas
^
«-T
Tangential inlet.
Axial discharge
Dust
Figure 5.1-7. Four Basic Cyclone Types (adapted from Reference 2).
5.1-10
-------
0.
H
r—i
T:
5 :
| :
i I . . .
Figure 5.1-8. Standard Dimensions of a Cyclone (Reference 6).
Table 5.1-1. Characteristics of Common Cyclones
Cyclone Type
Cyclone Dimension
Body Diameter, DID
Inlet Height, alD
Inlet Width, bID
Gas Exit Diameter, DJD
Vortex Finder Length, SID
Body Length, hID
Cone Length, LJD
Dust Outlet Diameter, BID
High Efficiency
(I) (II)
1.0
0.5
0.2
0.5
0.5
1.5
2.5
0.375
1.0
0.44
0.21
0.4
0.5
1.4
2.5
0.4
Conventional
(HI) (IV)
1.0
0.5
0.25
0.5
0.625
2.0
2.0
0.25
1.0
0.5
0.25
0.5
0.6
1.75
2.0
0.4
High Throughput
(V) (VI)
1.0
0.75
0.375
0.75
0.875
1.5
2.5
0.375
1.0
0.8
0.35
0.75
0.85
1.7
2.0
0.4
Note: The various cyclone designs correspond to the following literature references:
cyclone I and V from Reference 7, cyclones II, IV, and VI from Reference 8,
and Cyclone III from Reference 9.
5.1-11
-------
A multiple cyclone, shown in Figure 5.1-9, is a type of high efficiency cyclone which
consists of many small diameter cyclones operating hi parallel. This arrangement allows for
the treatment of large flow rates at higher efficiencies than for single cyclones.2
The greatest limitation hi the use of cyclones is the energy needed to force the gas
through the narrow cyclone body. The pressure drop within the cyclone generally increases
with increasing gas flow rate and decreasing cyclone diameter. Cyclone pressure drop can be
estimated from a number of equations that are based on both theory and experimental data.10
5.1.2 Collection Efficiency of Precollectors
Mechanical precollectors have a wide range of collection efficiencies. Collectors which
rely only on gravity settling, such as settling chambers and elutriators, typically have the
lowest collection efficiencies. Cyclones are the most effective mechanical collectors, with
multicyclones achieving the highest collection efficiencies.
5.1.2.1 Gravity Settling
Gravity settling chambers are most effective for large and/or dense particles.
Collection efficency for PM10 is very low, typically less than 10 percent. The efficiency of
settling chambers increases with the residence tune of gas hi the chamber. Because of this,
gravity settling chambers are often operated at the lowest possible gas velocities.
Unfortunately, as the gas velocity decreases, the size of the chamber must increase. In reality,
the gas velocity must be low enough to prevent dust from becoming reentrained, but not so
low that the chamber becomes unreasonably large.3 Figure 5.1-10 presents a typical fractional
collection efficiency curve for settling chambers.2 The impact of particle density is illustrated
hi Figure 5.1-11.3 The more dense particles of iron oxide, with a specific gravity of 4.5, are
collected more efficiently than the quartz dust which has a specific gravity of 2.6.2-3
5.1.2.2 Momentum Separators
Because these devices utilize inertia hi addition to gravity, momentum separators
achieve collection efficiencies approaching 20 percent for PM10. Collection efficiency for
momentum separators will increase as the gas velocity increases. The pressure drop and
corresponding operating costs will also increase with gas velocity, so the optimum velocity
must be chosen to balance the efficiency and operating costs.2-3 Figure 5.1-12 presents
efficiency data for momentum separators collecting fly ash.
5.1.2.3 Mechanically-Aided Separators
Figure 5.1-13 provides efficiency curves for two types of mechanically-aided
separators.2 Mechanically-aided separators are capable of collection efficiencies approaching
30 percent for PM10.? Mechanically-aided separators generally produce more centrifugal
5.1-12
-------
Figure 5.1-9. Typical Multiple Cyclone (Reference 3).
5.1-13
-------
Fractional Collection Efficiency (%)
IS
n
o
N)
T)
g"
9
5
s
w
3
o
S'
I
9
S
S1
i-»
oo
CD
a
i'
O
rf
Fractional Collection Efficiency (%)
ooooooooooo
o
HT
00
R'
re
I
o
o
\
\
-------
I
Fractional Collection Efficiency (%)
Fractional Collection Efficiency (%)
Ui
-•
CX O
P
00 >—
o
I
I
n
n>
I
o
I
a
to
h3
s
g;
CO
fi'
n>
^
3
-------
force than cyclones, but they also have shorter residence times and more reentrainment as a
result of turbulence. A major advantage of these separators is their compact size.4
5.1.2.4 Cyclones
There are many factors which affect the collection efficiency of cyclones. Cyclone
efficiency has generally been shown to increase with the following parameters: (1) particle
size and/or density, (2) inlet duct velocity, (3) cyclone body length, (4) number of gas
revolutions in the cyclone, (5) ratio of cyclone body diameter to gas exit diameter, (6) dust
loading, and (7) smoothness of the cyclone inner wall.11
The cyclone efficiency will decrease with increases in the following parameters: (1)
gas viscosity, (2) cyclone body diameter, (3) gas exit diameter, (4) gas inlet duct area, and (5)
gas density. Another common cause of cyclone ineffectiveness is leakage of air into the dust
outlet. Specifically, this will decrease the efficiency for fine particles.11
Several approaches for estimating cyclone efficiency have been developed. Most
cyclone theories utilize a particle size term, called a "particle cut size," that defines the
particle size for a specific collection efficiency. Particles greater than the cut size will be
collected with greater than the specified efficiency, and smaller particles will be collected less
efficiently. Usually, the particle cut size corresponds to 50 percent collection efficiency and is
called the "d50." Another important cyclone sizing parameter is the "critical particle size."
Particles of this size and larger are captured with 100 percent efficiency. Two general types of
cyclone fractional efficiency curves are shown in Figure 5.1-14; the first curve "A" is
hyperbolic, the second curve "B" is sigmoid shaped. Most cyclone efficiency theories will
produce a curve similar to one of the two shown in this figure.6-15
Lapple12 developed a relatively simple model to predict cyclone efficiency that was
derived from particle motion theory and requires an assumption about the number of turns the
gas makes within the cyclone. Leith and Licht13 developed an efficiency theory that was based
on an approximate solution to theoretical particle motion equations using the assumption of
turbulence within the cyclone. Both these theories produce a cyclone efficiency curve of type
"A" in Figure 5.1-14. More recently, lozia and Leith14'15 developed a cyclone efficiency
theory based on theoretical particle motion that uses empirically developed coefficients. This
theory produces a cyclone efficiency curve of type "B" hi Figure 5.1-14. The lozia and Leith
theory was shown to predict cyclone efficiency better than the theories of Lapple, and Leith
and Licht, using cyclone laboratory test data available in the literature.15 The Lapple theory
and the lozia and Leith theory are discussed hi more detail below.
According to the Lapple theory,12 d50 is calculated as follows:
450 = [9 M b 1(2 ic AT Vs (p. - pg))] * (Eq. 5.1-1)
5.1-16
-------
i
-4
T1
t
B
i
h—*
Ji.
1/3 O
ON g-
O
R-
V3
n
n
B'
S
D.
r
5'
9
v_^
00
O
e.
W
r
g'
Fractional Collection Efficiency (%)
K~ts)l»>-UVl0t~-JOO*OO
ooooooooooo
Fractional Collection Efficiency (%)
5 8
B'
K
8 w
s
R
5-
•s
B:
-------
where dso is the diameter of particle collected with 50 percent efficiency (ft), \t. is the gas
viscosity (Ib/sec-ft), b is the cyclone inlet duct width (ft), N is the number of gas revolutions in
cyclone (estimated to be between 0.3 and 10, with a mean value of approximately 516), Vj is
the inlet duct gas velocity (ft/sec), pp is the particle density (lb/ft3), and pg is the gas density
(lb/ft3). The limitation on this equation is that N, the number of gas revolutions within the
cyclone, is unknown and estimates for this value do not take into account individual cyclone
design or other operating conditions. Also, the Lapple theory does not allow for calculation of
collection efficiency for other particle sizes.
The efficiency theory developed by lozia and Leith14'15 to predict cyclone fractional
collection efficiency utilizes an equation, called a "logistic equation," that approximates a
sigmoid-shaped efficiency curve:
CE = 1/(1 + (dso/d)&) (Eq. 5.1-2)
where CE is the control efficiency (expressed as a fraction for a particle of diameter d,
J50 is the diameter of the particle collected with 50 percent efficiency, and beta (6) is a
coefficient. lozia and Leith developed an equation to predict 6 from cyclone dimensions
using laboratory test data from a 25 cm diameter cyclone:15
In B = 0.62 - 0.87 In (d50) + 5.21 In (ab/D2) + 1.05 (In (ab/D2)2 (Eq. 5.1-3)
where dso, as above, is expressed hi centimeters; a is the cyclone inlet duct height, b is the
cyclone inlet duct width, and D is the cyclone diameter. An equation to predict 4,0 was also
developed:
dso = {(9 M 0/Oc z PP VJU2)} (Eq. 5.1-4)
where p is gas viscosity, Q is gas flow, z is approximately equal to the cyclone height minus
the height of the extension of the exit duct into the cyclone, pp is the particle density, and
Vfnax2 is the maximum tangential gas velocity within the cyclone and is calculated as below:
Vt^2 = 6. 1 V5 {(ab/D2)061 (Dc/D)-0-74 (HID)**} (Eq. 5 . 1-5)
where V{ is the gas inlet duct velocity; a, b, and D are as above; De is cyclone outlet duct
diameter; and H is the cyclone overall height (h + Lc).
lozia and Leith used their cyclone efficiency theory to optimize cyclone design.17 Using
a computerized cyclone optimization program, they developed curves to predict the cyclone
dimensions of a cyclone with the highest efficiency possible for a given situation.18 Figure
5.1-15 shows the inlet and outlet duct dimensions needed to achieve this optimized cyclone
design.17 The predictions of lozia and Leith, however, have not yet been tested hi full-scale
industrial applications.
5.1-18
-------
0.8 -
0.6
I"
u
0.2
1.5
2.5
3.5
Figure 5.1-15.
Aerodynamic d50 (micrometers)
Dimensions of the Cyclone Inlet and Outlet Ducts for an Optimized
Cyclone Design, According to the lozia and Leith Cyclone Efficiency
Theory (Reference 17).
For single cyclones, conventional cyclones can remove 10 /^.m particles with 85 to 90
percent efficiency, 5 //m particles with 75 - 85 percent efficiency, and 2.5 //m particles with
60 to 75 percent efficiency.6 High efficiency single cyclones can remove 5 ^m particles at up
to 90 percent efficiency, with higher efficiencies achievable for larger particles.1 High
throughput cyclones are only guaranteed to remove particles greater than 20 /um, although
collection of smaller particles does occur to some extent.19
Multicyclones are reported to achieve from 80 percent efficiency up to 95 percent
efficiency for particles 5 //m.1-5-19 In some cases, multiple cyclones have been used as primary
collection devices.19 Multiple cyclones are often used as precollectors at industrial combustion
operations. Figure 5.1-16 shows the collection efficiency for multiple cyclones at a boiler that
is burning oil. At many large industrial combustion units, PM emissions include significant
amounts of carbon which was not fully combusted. To improve the efficiency of these units,
collected fly ash from the multiple cyclones (or other precollection devices) is reinjected into
the combustion unit. This operation, known as fly ash reinjection, increases the paniculate
loading considerably, and leads to lower collection efficiencies for small particles.5 Figure
5.1-17 illustrates this effect, showing efficiency curves for multiple cyclones at coal and wood
boilers with and without fly ash injection.
5.1-19
-------
e
ta
"o
O
100
QO
1
80
70
60
50
40
30
20
10
0
1
h •
•—. .
**
^^
t-,
i-H
h -•
j
10 10
Panicle Size (urn)
Figure 5.1-16.
Cumulative Collection Efficiency Data for Multiple Cyclones at a
Residual Oil-Fired Boiler (Reference 5).
5.1.3 Applicability
Mechanical precollectors have very few limitations in their application as precollectors,
although they are not generally used where there is no coarse PM. Mechanical precollectors
can be used to treat small and large flow rates and remove a wide range of particle sizes.
Mechanical collectors are simple in design and inexpensive to purchase and operate. In
addition, their use will reduce the particulate loading on primary collection devices which will
extend operating life.2
The vast majority of dusts are suitable for collection in mechanical precollectors. One
exception may be sticky dusts, which can clog cyclones. In addition, mechanical precollectors
would not be effective for gas streams where the bulk of the PM is small (< 3 //m).
Mechanical collectors can be constructed out of various materials and are capable of operating
under any conditions which the construction materials allow. Typically, any industry which
uses large and relatively expensive control devices, such as a fabric filters or electrostatic
precipitators (ESPs), will also employ mechanical precollectors. Multiple cyclones are the
most common device for industrial applications, especially for boilers and other combustion
units that generate smaller particles. Some industries also use mechanical collectors for
product or catalyst recovery, since these collectors are nondestructive and allow reuse or sale
of the recovered material.2
5.1-20
-------
n
o
S'
5s n
3- o
K-
I
o
o l-t
n
o
tu
*
Cumulative Collection Efficiency (%)
§ §
Cumulative Collection Efficiency (%)
i-* N> UJ ^ Ui o ^j OO vD
oooooooooo
t
o'
II
S-
n
§•
GT)
I
s
\
\
C1 O
fa
V
-------
5.1.4 Costs of Precollectors
The costs of installing and operating a mechanical pre-collector include both capital and
annual costs. Capital costs are all of the initial costs related to collector equipment and
installation. Annual costs are the direct yearly costs of operating the device, plus indirect
costs such as overhead; capital recovery; and taxes, insurance, and administrative charges.
The following sections discuss capital and annual costs for mechanical collectors, referenced to
the third quarter of 1995 unless otherwise noted. Since cyclones are the most common and
generally most effective mechanical precollectors for industrial applications, this section will
focus only on the costs of these devices.
5.1.4.1 Capital Costs of Cyclones
The total capital investment (TCI) for cyclones includes all of the initial capital costs,
both direct and indirect. Direct capital costs are the purchased equipment costs (PEC), and the
costs of installation (supports, etc.). Indirect costs are related to the installation and include
engineering, construction, contractors, start-up, testing, and contingencies. The PEC is
calculated based on the cyclone specifications. The direct and indirect installation costs are
calculated as factors of the PEC. For cyclones, installation costs are generally low, with the
combination of direct and indirect costs assumed to be about 25 percent of the PEC. Hence,
the TCI for cyclones is typically calculated as 1.25 times the PEC.20
The most important parameter for sizing cyclones is the inlet duct area (A), which can
be calculated from the following equations:19
A = Q/V{ (Eq. 5.1-6)
A = (£(pp - pgV/i)1'33 4267 (Eq. 5.1-7)
where A is the cyclone inlet duct area (ft2), Q is the cyclone gas flow rate (ACFM), V-t is the
cyclone inlet duct velocity (ft/min), pp is the particle density (lb/ft3), pg is the gas density
(lb/ft3), fji is the gas viscosity (Ib/ft-sec), and dc is the critical particle size Gum).
By selecting an inlet duct gas velocity (V;) for Equation 5.1-2, the inlet duct area can be
determined and Equation 5.1-3 can be solved for the critical particle size (dc). The critical
particle size is defined for this equation as the smallest particle that the cyclone can collect
with 100 percent efficiency. Similarly, the inlet duct gas velocity can be calculated for a given
critical particle size.19
Cyclone costs are based on the inlet duct area, and include the cyclone, fan, motor,
supports, hopper (or drum), and rotary air lock. Figure 5.1-18 presents a cost curve for
cyclones with inlet duct areas between 0.2 and 2.64 ft2. The step in this curve at 0.35 ft2 is a
result of the fact that cyclones with inlet duct areas less than 0.35 ft2 do not require air locks,
5.1-22
-------
o
X
J3
13
30
25
20
15
§• 10
U
Inlet Area (ft2)
Figure 5.1-18.
Total Capital Investment vs. Inlet Duct Area for
0.2 ft2 < Duct Area < 2.64 ft2 (Reference 19).
which are dampers that prevent the gas in the cyclone from entering the dust hopper during
dust removal.19 For cyclones with required total inlet duct areas greater than 2.64 ft2, the total
inlet duct area must be divided equally between 2 or more cyclones, each with inlet duct areas
less than 2.64 ft2.
Figure 5.1-19 provides a cost curve for cyclones with total inlet duct areas greater than
2.64 ft2.19 The steps in this curve indicate the flow rates at which an additional cyclone
becomes necessary. The step function approximates a straight line. The cost curves in this
document are for carbon steel cyclones, other materials may increase costs.20
Capital costs obtained from this document can be escalated to more current values
through the use of the Vatavuk Air Pollution Control Cost Indexes (VAPCCI), which are
published monthly and updated quarterly in Chemical Engineering magazine. The VAPCCI
updates the PEC and, since capital costs are based only on the PEC, capital costs can be easily
adjusted using the VAPCCI. To escalate capital costs from one year (Costold) to another more
recent year (Cost,^), a simple proportion can be used, as follows:21
(Eq. 5.1-8)
Cost™ =
The VAPCCI for mechanical collectors from fourth quarter 1996 was 103.3.
5.1-23
-------
6 8
Total Inlet Area (ft2)
10
12
14
Figure 5.1-19.
Total Capital Investment vs. Inlet Duct Area for Duct Area > 2.64 ft2
(Reference 19).
5.1.4.2
Annual Costs of Cyclones
The total annual costs for a cyclone consist of both direct and indirect costs. Direct
annual costs are those associated with the operation and maintenance of the cyclone. These
include maintenance labor, maintenance materials, electricity, and dust disposal. Typical
nonhazardous dust disposal costs are $20-$30/ton, excluding transportation costs. Hazardous
dusts can cost ten times as much to dispose of.22 Cyclones are assumed to have no need for
operator and supervisor labor.20
The indirect annual costs for cyclones include taxes, insurance, administrative costs,
overhead, and capital recovery. All of these costs but overhead are dependent on the TCI.
Table 5.1-2 provides the parameters which impact annual costs and estimates of typical values.
Table 5.1-3 provides the annual cost factors for cyclones.20 Annual costs are very site-specific
and, therefore, difficult to generalize.
5.1-24
-------
Table 5.1-2. Annual Cost Parameters for Cyclones (Reference 9).
Parameter
Description
Typical Values
Direct Cost Parameters
Operating factor (OF)
Maintenance labor rate (MR)
Maintenance shift (MS) factor
Maintenance materials factor (MF)
Electricity rate (ER)
Dust disposal cost (DC)
Indirect Cost Parameters
Overhead factor (OV)
Annual interest rate (I)
Operating life (n)
Capital recovery factor (CRF)
Taxes (TAX)
Insurance (INS)
Administrative costs (AC)
Hours of cyclone operation per year
Maintenance labor pay rate
Fraction of maintenance shift on cyclone
Fraction of maintenance labor cost
Cost of electricity
Cost of dust disposal
Fraction of total labor and (MM) costs
Opportunity cost of the capital
Expected operating life of cyclone
Function of (n) and (I)
Fraction of the TCId
Fraction of the TCId
Fraction of the TCId
8,640 hr/yr
$14.00/hrb
0.25"
1.0b
$0.07/kW-hra
$20-$30/tona
0.60"
7 percent6
20 yearsb
0.0944C
0.01"
0.01"
0.02b
Estimated for 1996 from currently available information.
Estimates from "CO$T-AIR" Control Cost Spreadsheets (Reference 20).
Capital Recovery Factor is calculated from the following formula: CRF = {1(1 + I) f H- {(1 + I)"- 1},
where / = interest rate (fraction) and n = operating life (years).
The total capital investment (TCI) can be escalated to current values by using the Vatavuk Air Pollution Control Cost Indexes
(VAPCCI), described in Section 5.1.4.1.
-------
Table 5.1-3. Annual Cost Factors for Cyclones (Reference 20).
Cost Item
Formula
Factor
ON
Direct Costs
Maintenance labor (ML)
Maintenance materials (MM)
Electricity (E)
Dust disposal (D)
Total Direct Cost (DC)
Indirect Costs
Overhead
Capital Recovery
Taxes
Insurance
Administrative Costs
Total Indirect Cost (1C)
Total Annual Cost (DC + 1C)
(OF) X (MR) X (MS)
(MF)X(ML)
Fan Power X (ER)
(DC) x Tons per year
(OV)x(ML+MM)
(CRF)X(TCI)
(TAX) x (TCI)
(INS) x (TCI)
(AC) x (TCI)
A
A
E
2A + E + D
1.2 A
0.0944 TCI
0.01 TCI
0.01 TCI
0.02 TCI
1.2 A + 0.1344 TCI
3.2 A + 0.1344 TCI + E + D
Note: These values are also described in Table 5.1-2.
-------
5.1.5 Energy and Other Secondary Environmental Impacts of Precollectors
The secondary environmental impacts of cyclone operation are related to energy
consumption and solid waste generation. The energy demands for cyclones consist of
electricity requirements for fan operation. The fan power needed for a specific cyclone is
dependent on the pressure drop and can be estimated with the folio whig equations:19
AP = 2.36 XlO-7(Vj2) (Eq. 5.1-9)
Fan Power (kW-hr/yr) = 1.81 x 104(0(AP)(0 (Eq. 5.1-10)
where AP is the cyclone pressure drop (in. water), V{ is the inlet duct gas velocity (ft/min), Q
is the gas flow rate (ACFM), and t is the operating time (hr/yr).
Dust which is collected in mechanical collectors may be sold or recycled if it has
intrinsic value or can be recycled by use hi other materials, such as concrete. Otherwise, the
fly ash must be disposed. Most inert, nonhazardous dusts can be landfilled. Dusts which are
hazardous or reactive will typically require treatment or disposal in a secure landfill.
5.1.6 Flue Gas Conditioning
Gas conditioning is used to modify the characteristics of the gas stream and particles to
enhance particle removal hi the primary collection device. Flue gas conditioning at coal fired
power plants is the most widespread application of this practice. Usually, flue gas
conditioning involves the use of chemicals that are added to the gas stream to improve the fly
ash properties and electrical conditions hi electrostatic precipitators. See Section 5.2 for a
detailed discussion of ESP operation. Fabric filter and scrubber performance is far less
dependent on the chemical composition of the gas and particles, so these devices typically do
not employ chemical conditioning for particle removal. Any gas conditioning for fabric filters
or scrubbers usually consists of controlling the temperature and moisture of the gas stream.
Therefore, this section will only discuss flue gas conditioning for ESPs.
Flue gas conditioning is most often used to retrofit ESP's which are not operating up to
the design efficiency. This often occurs as a result of switching to low-sulfur coal,1-11 which
produces a high resistivity fly ash that is difficult to collect hi an ESP. Collection efficiency of
an ESP is dependent on the electric field strength and ion density; the adhesive and cohesive
properties of the fly ash; and the particle size and size distribution. Flue gas conditioning can
influence all of these parameters.23 Conditioning agents can improve ESP collection efficiency
with one or more of the following mechanisms:23
• Adsorb on surface of fly ash and reduce surface resistivity
• Adsorb on fly ash and change adhesion/cohesion properties
• Increase ultrafine particle concentration for space charge improvement
5.1-27
-------
• Increase sparkover voltage of flue gas (reduce back corona)
• Increase mean particle size
• Decrease acid dew point in flue gas
Common conditioning agents include sulfur trioxide (SO3), ammonia, ammonium compounds,
organic amines, and dry alkalis.11
The effect of flue gas conditioning on collection efficiency is difficult to quantify
because it differs greatly between applications. In some cases, a small dose of conditioning
agent will provide a significant improvement in collection efficiency while subsequently larger
doses show little additional improvement. Other users have reported steady increases in
collection efficiency with increasing doses.23 In Flue Gas Conditioning (Reference 23), a
variety of SO3 conditioning users reported increases in collection efficiency ranging from 1.7
percent up to 18 percent with SO3 doses ranging from 5 ppmv to 64 ppmv.23
5.1.6.1 Sulfur Trioxide Conditioning
The most commonly used flue gas conditioning agent for power plants in the U.S. is
SO3. Sulfur trioxide is injected into the gas stream after the air preheater, which is a heat
exchanger that uses the hot flue gases to preheat the combustion intake air. Sulfuric acid is
actually the active agent hi conditioning. Virtually all of the SO3 is hydrated to sulfuric acid at
the temperature and humidity of the flue gas stream. Sulfuric acid has a strong affinity for
water and readily dissociates to two hydrogen ions and one sulfate ion hi solution. This
property makes the solution highly conductive. Sulfuric acid also has a low vapor pressure; as
a result, sulfuric acid will not easily vaporize, even from a concentrated solution to a dilute
vapor phase. This combination of low volatility and high conductivity make sulfuric acid very
effective at lowering the resistivity of fly ash.11-23
Sulfuric acid reduces the resistivity of particles in ESP's by establishing a layer of
conductive solution on the particle surface through adsorption and/or condensation of sulfuric
acid and water. A layer of acid solution also forms on particles which have already been
collected on the collection plate.11-23
Sulfur trioxide used for gas conditioning is commonly generated by one of the
folio whig four methods: 1) vaporization of sulfuric acid solution, 2) vaporization of liquid
SO3, 3) vaporization of liquid sulfur dioxide (SO2) and subsequent catalytic oxidation to SO3,
and 4) combustion of liquid sulfur hi ah* to produce SO2 with subsequent oxidation to SO3.
Methods 3 and 4 are the most reliable, with method 3 the most economical.11'23
A typical dosage of SO3 ranges from 5 to 30 ppmv, though the dose may be as high as
70 ppmv. A dosage of 20 ppmv can decrease fly ash resistivity by two orders of
magnitude.11-23
5.1-28
-------
The flue gas temperature and PM composition of the gas stream determine how
effective SO3/sulfuric acid will be in reducing resistivity. At high temperatures (>200°C),
sulfuric acid is less effective for two reasons. First, temperatures above the acid dew point
prevent sulfuric acid from adsorbing or condensing in appreciable quantities on the particles.
Second, volume conduction through the bulk of the particles becomes dominant over surface
conduction at high temperatures. A conductive layer of sulfuric acid will only improve
surface conduction. If the PM includes a large amount of alkaline compounds, the PM will
react with the acid to form a nonconductive layer of salts. A higher dosage of SO3 will
provide excess acid to adsorb or condense on top of the salt layer.11-23
Users of SO3 conditioning have reported problems with corrosion, catalyst
deactivation, and over conditioning. At temperatures below the acid dew point, sulfuric acid
will condense in the pipes leading to the injection nozzles. This can lead to corrosion and
clogging of the lines. Installations that rely on catalytic oxidation of sulfur dioxide to produce
SO3 have found it necessary to replace the catalyst approximately once a year. Conditioning
with excessive doses of sulfuric acid can also over-condition the ash. In such cases, the
resistivity becomes too low and particles are easily reentrained. In addition, the acid can form
a solution which binds particles to the plates, making them difficult to remove with normal
rapping.11-23
5.1.6.2 Ammonia Conditioning
Ammonia conditioning has been shown to be an effective gas conditioning agent for
ESPs operating with low sulfur coals. In Australia, ammonia is the most popular ESP
conditioning agent, because Australian coal produces ash with very high resistivity. In the
U.S., ammonia conditioning has become important with the increased use of low sulfur coals,
that are being used to offset the generation of acid rain precursors, like SO2.U'23
Ammonia is also being used to optimize the operation of fabric filters.24 Both of these
applications are discussed below.
Ammonia Conditioning in ESPs. In an ESP, ammonia can be injected either as a vapor
or in solution before or after the air preheater. While the conditioning mechanisms of
ammonia have not been fully explained, two distinct effects have been reported. The first
effect is the enhancement of the space charge in the ESP. The ammonia reacts with sulfuric
acid vapor in the flue gas and forms a fume of fine ammonium salt particles. The fine
particles increase the resistivity of the gas phase hi the interelectrode space. The increased gas
resistivity increases the voltage drop and the field strength between the electrodes, allowing
for operation at higher voltages. The second effect is an increase in the cohesiveness of the
particles. Condensation of sulfuric acid hi the presence of ammonium salts contributes to the
adsorption of acid and salts to water on the surface of particles. The layer of surface deposits
is very sticky and increases the cohesive force between particles on the collectionrplate,
reducing reentrainment.11-23
5.1-29
-------
The effects of ammonia on resistivity are not entirely clear. Evidence has shown that
fly ash conditioned with ammonia may have resistivity that is less than, the same as, or greater
than unconditioned ash. Ammonia conditioning is very sensitive to temperature, and is more
effective at low temperatures (< 110°C). Ammonia conditioning has generally been shown to
increase collection efficiency; however, the increase hi ESP efficiency does not necessarily
correlate with a decrease in fly ash resistivity. Users of ammonia have reported problems with
plugging of nozzles and build-up of deposits on the discharge electrodes.11-23
Ammonia Conditioning in Fabric Filters.24 Ammonia additions to flue gas containing
SO3 produces ammonium sulfate compounds. These byproducts have been found to increase
ash cohesivity, thereby reducing PM emissions during the cleaning cycle of fabric filters. The
presence of ammonium sulfate compounds also produces a higher porosity dust cake that
results in a lower pressure drop hi the fabric filter. In addition, ammonia conditioning reduces
the corrosion and bag failure that results with SO3 condensation.
5.1.6.3 Ammonium Compound Conditioning
Ammonium compounds provide a more convenient method of conditioning with
ammonia. Sulfamic acid (NH3SO2OH), ammonium sulfate [(NH4)2SO4], and ammonium
bisulfate (NH4HSO4) are the most common ammonium compounds for conditioning. Several
proprietary conditioning agents also contain ammonium compounds. These compounds can
dissociate into ammonia and sulfuric acid hi the gas stream, and may provide the effects of
both ammonia and sulfuric acid (from S03) conditioning. Ammonium compounds can be
introduced upstream or downstream of the air preheater.11'23
Ammonium compounds can enhance ESP performance by decreasing particle
resistivity, increasing space charge, and/or increasing particle cohesion. Particle resistivity
can be reduced by ammonium sulfate and ammonium bisulfate. Both compounds behave like
sulfuric acid, forming a layer of conductive solution on the surface of the collected fly ash.
The sulfuric acid and ammonia formed when ammonium compounds decompose can react to
form fine salt particles which increase the space charge. These particles increase the gas
resistivity, which increases the voltage drop between the electrodes and allows for ESP
operation at higher applied voltage. The cohesion of the particles is also increased, which
increases collection efficiency, through condensation of sulfuric acid and through the
adsorption of compounds that form a viscous, sticky layer on the particles.11-23
Ammonium sulfate, ammonium bisulfate, and sulfamic acid are effective hi reducing
ash resistivity because they can form a layer of conductive solution on the particles or
decompose into SO3. Ammonium compounds can reduce resistivity by approximately a factor
of 5. While ammonium compounds are effective hi reducing resistivity, they are not as
effective as SO3. Problems of plugging in the ah" preheater have been reported when the
agents are injected prior to the preheater.11-23
5.1-30
-------
Apollo LPA-445, a proprietary conditioning agent containing ammonium compounds,
effectively increases ash cohesiveness. Apollo LPA-445 has also demonstrated that it
increases the number of fine particles when used as a conditioning agent. Space charge
increases with an increase in fine particles.11-23
5.1.6.4 Organic Amine Conditioning
Research on organic amines as conditioning agents has been conducted in laboratories
and pilot-scale ESP's. Of all the amines, triethylamine [N(C2Hs)3] has been studied most
extensively. Triethylamine is an organic nitrogen compound which behaves similar to
ammonia, but is a stronger base than ammonia. The conditioning mechanism for triethylamine
is not completely understood, but it appears to also reduce the fly ash resistivity by forming a
conductive layer on the surface of the fly ash.11-23
A dose of 60 ppm triethylamine has been demonstrated to reduce ash resistivity from
3 x 109 ohm-m to 5 x 107 ohm-m in a pilot scale boiler. A laboratory study has shown
resistivity reductions of 1 to 2 orders of magnitude for a 25 ppm dose at temperatures of
102°C to 150°C. Triethylamine is more effective with lower temperatures, less basic ash, and
higher doses of the agent.11-23
5.1.6.5 Alkali Conditioning
Many coals from the western U.S. produce fly ash with high resistivity when
combusted. It has been determined that the fly ash has low alkali content. Fly ash resistivity
has been found to be inversely proportional to the concentration of lithium and sodium in the
ash. Hence, alkali conditioning is used to reduce resistivity by increasing alkali concentration
hi the ash. Sodium compounds are the most widely used conditioning agents. Sodium sulfate
and sodium carbonate are commonly used. Sodium chloride has shown to be effective hi
laboratory tests, but it is not used industrially because of its potential for corrosion.11-23
Sodium compounds can reduce resistivity in two ways. The compound can be injected
prior to the ESP and collected with the fly ash. In this case, the sodium compound mixes with
the ash and serves as a conductive medium. It also reduces the resistivity of the layer of dust
which collects on the plates. The second means of reducing fly ash resistivity is unique to
alkali conditioning. Sodium compounds can be injected into the boiler and combusted with the
coal. In this method, the sodium is bound to the fly ash and reduces resistivity as the presence
of natural sodium would.11-23
Compounds of sodium are effective at reducing the resistivity of fly ash if the sodium
and ash are well mixed. A 1.0 to 1.5 percent concentration of sodium carbonate has been
demonstrated to reduce resistivity from 2.1 x 1010 to 3.7 x 109 hi a pilot scale ESP. Ash
resistivity reductions of two orders of magnitude have been reported from field tests.
5.1-31
-------
Problems with boiler fouling may occur when sodium compounds are introduced to the coal
prior to combustion.11-23
5.1.7 Costs of Flue Gas Conditioning
Costs of flue gas conditioning vary with the size of the power plant and the type of
conditioning installed. There are also several site specific factors which will influence flue gas
conditioning costs, such as the sulfur content of the fuel, fly ash resistivity, and the initial
collection efficiency of the ESP prior to flue gas conditioning. Table 5.1^ provides estimates
of capital and annual costs for flue gas conditioning at electric utility plants.23-25 The cost data
in Table 5.1-4 has been escalated to February 1996 where possible through the use of
Chemical Engineering magazine's "Plant Cost Index."26 No cost data were available for
ammonium compounds and organic amines; this is probably due to the limited industrial
experience with these conditioning agents.11-23
5.1.8 Energy and Other Secondary Environmental Impacts of Flue Gas Conditioning
Flue gas conditioning can lead to air emissions from compounds formed by the
conditioning agents.11-23 No energy impacts, however, are associated with flue gas
conditioning.
Sulfur trioxide conditioning can lead to emissions of sulfuric acid mist and particulate
sulfate compounds caused by the condensation of acid and formation of sulfate on fly ash
particles. The ESP performance, ash composition, and gas temperature effect the emission
rates of these compounds.11-23
Ammonium sulfate conditioning can result in increased sulfate emissions. Most
sulfate particle emissions are smaller than 1 fan. Ammonium phosphates can decompose to
form ammonia and either phosphorous pentoxide or condensed phosphates.11-23
Although ammonia conditioning in the presence of SO3 forms ammonium bisulfate
aerosol, the ammonium bisulfate should be collected by the ESP. A secondary benefit of
ammonia conditioning is that large amounts of ammonia will react with nitrogen oxides to
form elemental nitrogen.11-23
5.1-32
-------
Table 5.1-4. Costs of Flue Gas Conditioning
Conditioning
Agent
Capital Cost
($/kW)a-c
Annual Cost
(mills/kWh)a-b
Comments
Sulfur trioxide
6.24
Ammonia
0.25
Alkali
2.47 - 4.78
0.191 Capital costs developed from 1982 data
(Reference 23). Annual costs based on
current costs for sulfur ($300/Mg) from
Reference 25.
c
0.024 Capital costs data from Reference 23.
Annual costs based on 1996 costs for
ammonia ($325/Mg) from Reference 25,
and costs for steam ($17.22/Mg) escalated
from 1982 data from Reference 23.
0.03 Capital costs from Reference 23.
Annual costs based on 1979 data from
Reference 23; available information not
detailed enough to allow for escalation to
current prices.
Kilowatts of power plant capacity.
1 mill = $0.001.
Escalated to February 1996 using the Chemical Engineering Plant Cost Index (Reference 26).
-------
5.1.9 References for Section 5.1
1. Cooper, C.D. and F.C. Alley. Air Pollution Control: A Design Approach. Second
edition. Waveland Press, Prospect Heights, Illinois. 1994.
2. Control Techniques for Participate Emissions from Stationary Sources - Volume 1
(EPA-450/3-81-005a; NTIS PB83-127498). U.S. Environmental Protection Agency,
Office of Air Quality Planning Aid Standards, Research Triangle Park, North Carolina.
September 1982.
3. Theodore, L. and AJ. Buonicore. Air Pollution Control Equipment, Volume I:
Particulates. CRC Press, Boca Raton, Florida. 1988.
4. Perry, R.H. and D.W. Green. Perry's Chemical Engineers'Handbook. Sixth edition.
McGraw-Hill Publishing Company, New York, New York. 1984.
5. Compilation of Air Pollution Emission Factors (AP-42). Volume I (5th Edition). U.S.
Environmental Protection Agency, Research Triangle Park, North Carolina. January
1995.
6. Air Pollution Engineering Manual. A.J. Buonicore, and W.T. Davis, Eds. Air &
Waste Management Association, Pittsburgh, Pennsylvania; and Van Nostrand
Reinhold, New York, New York. 1992.
7. Stairmand, C.J. "The Design and Performance of Cyclone Separators." Trans. Ind.
Chem. Eng. 29. 1951.
8. Swift, P. "Dust Control hi Industry." Steam Heating Engineering. 38. 1969.
9. .Lapple, C.E. "Processes Use Many Collector Types." Chem. Eng. 58:144-151.
1951.
10. Leith, D. and D. Mehta. Atmospheric Environment. 7:527. 1973.
11. The Electrostatic Precipitator Manual (Revised). The Mcllvaine Company.
Northbrook, Illinois. January 1992.
12. Lapple, L.E. Industrial Hygiene Quarterly. 11:40. 1950.
13. Leith, D. and W. Licht. AIChE Symposium Series. 68:196. 1972.
14. lozia, D.L. and D. Leith. "Effect of Cyclone Dimensions on Gas Flow Pattern."
Aerosol Science and Technology. 10(3):491-500. 1989.
5.1-34
-------
15. lozia, D.L. and D. Leith. "The Logistic Equation and Cyclone Fractional Efficiency."
Aerosol Science and Technology. 13(1). 1990.
16. Friedlander, S.K, L. Silveraian, P. Drinker, and M.W. First. Handbook on Air
Cleaning. USAEC (AECD-3361, NYO-1572), Washington, DC. 1952.
17. lozia, D.L. and D. Leith. "Optimizing Cyclone Design and Performance." Filtration
and Separation. 24(4):272-274. 1989.
18. Leith, D. and D.L. Jones. Chapter 15: "Cyclones," in Handbook of Powder Science
and Technology. M.E. Fayed and L. Otten, Eds. Chapman and Hall, New York, New
York. 1997.
19. Vatavuk, W.M. Estimating Costs of Air Pollution Control. Lewis Publishers,
Chelsea, Michigan. 1990.
20. Vatavuk, W.M. "CO$T-AIR" Control Cost Spreadsheets. Innovative Strategies and
Economics Group, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency. Research Triangle Park, North Carolina. February 1996.
21. Vatavuk, W.M. "Escalate Equipment Costs." Chemical Engineering. December
1995.
22. OAQPS Control Cost Manual (Fourth Edition, EPA 450/3-90-006) U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards,
Research Triangle Park, North Carolina. January 1990.
23. Flue Gas Conditioning (EPA-600/7-85-005, NTIS PB85-173912). U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina. February 1985.
24. Oglesby, S., Jr. Future Directions of Paniculate Control Technology: A Perspective.
J. Air Waste Management Assoc. 40(8): 1184-1185. August 1990.
25. "Chemical Prices." Chemical Marketing Reporter. 250:7. August 12, 1996.
26. "Plant Cost Index." Chemical Engineering. July 1996.
5.1-35
-------
5.2 ELECTROSTATIC PRECIPITATORS
This section discusses the basic operating principles, typical designs, industrial
applications, and costs of electrostatic precipitators (ESPs). Collection of particles by
electrostatic precipitation involves the ionization of the stream passing though the ESP, the
charging, migration, and collection of particles on oppositely charged surfaces, and the
removal of particles from the collection surfaces. In dry ESPs the paniculate is removed by
rappers which vibrate the collection surface. Wet ESPs use water to rinse the particles off.
Electrostatic precipitators have several advantages when compared with other control
devices. They are very efficient collectors, even for small particles. Because the collection
forces act only on the particles, ESPs can treat large volumes of gas with low pressure drops.
They can collect dry materials, fumes, or mists. Electrostatic precipitators can also operate
over a wide range of temperatures and generally have low operating costs. Possible
disadvantages of ESPs include high capital costs, large space requirements, inflexibility with
regard to operating conditions, and difficulty in controlling particles with high resistivity.1
Disadvantages of ESPs can be controlled with proper design.
5.2.1 Particle Collection
Particle collection during electrostatic precipitation is the end result of several steps.
These steps include the establishment of an electric field, corona generation, gas stream
ionization, paniculate charging, and migration to the collection electrode. One typical ESP
arrangement is shown in Figure 5.2-1.2 In this illustration, the discharge electrode is a
weighted wire and the collection electrode is a pipe. A wire-pipe ESP would contain many
such wires and pipes.
5.2.1.1 Electric Field
The electric field plays an important role hi the precipitation process hi that it provides
the basis for generation of corona required for charging and the necessary conditions for
establishing a force to separate paniculate from the gas streams.2 An electric field is formed
from application of high voltage to the ESP discharge electrodes; the strength of this electric
field is a critical factor hi ESP performance.3
The electric field develops hi the interelectrode space of an ESP and serves a three-fold
purpose. First, the high electric field in the vicinity of the discharge electrode causes the
generation of the charging ions hi an electrical corona; second, the field provides the driving
force that moves these ions to impact with and attach then- charge to the particles; and thirdly,
it provides the force that drives the charged paniculate to the collection electrode for removal
from the effluent gas stream.2
5.2-1
-------
Insulator
Precipitator
Shell
Power
Supply
i
Discharge Electrode
Pipe Collection Electrode
Dirty
Gas
Figure 5.2-1. Cutaway view of Wire-Pipe Electrostatic Precipitator (Reference 2).
5.2-2
-------
The electric field in an ESP is the result of three contributing factors: the electrostatic
component resulting from the application of a voltage in a dual electrode system, the
component resulting from the space charge from the ions and free electrons, and the
component resulting from the charged particulate. Each of these factors may assume a
dominant role in the determination of the field in a given set of circumstances. For example,
the electric field in the vicinity of the first few feet of the inlet section of an ESP collecting
particulate from a heavily particulate-laden gas stream may be dominated by the particle space
charge; while the field in the outlet section of a highly efficient ESP is usually dominated by
the ionic space charge.2
The strength or magnitude of the electric field is an indication of the effectiveness of an
ESP.3 Two factors are critical to the attainable magnitude of the electric field in an ESP.
First, the mechanical alignment of the unit is important. If a misalignment occurs in a
localized region that results in a close approach between the corona and collection electrodes,
the sparking voltage for that entire electrical section will be limited. The second is the
resistivity of the collected particulate, which can limit the operating current density and
applied voltage that results in a reduced electric field.2
5.2.1.2 Corona Generation
The corona is the electrically active region of a gas stream, formed by the electric
field, where electrons are stripped from neutral gas molecules leaving positive ions. The
positive ions are driven in one direction and the free electrons in another. The necessary
conditions for corona formation include the presence of an electric field with a magnitude
sufficient to accelerate a free electron to an energy required to ionize a neutral gas molecule on
impact, and a source of electrons to act as initiating electrons for the process.2
Details of electric field generation were discussed above. In terms of electron sources,
there is always a supply of free electrons available from the ionization of gas molecules by
either cosmic rays, natural radioactivity, photoionization, or the thermal energy of the gas.2
The corona is generated by a mechanism which is commonly referred to as electron avalanche.
This mechanism occurs when the magnitude of the applied electric field is great enough to
accelerate the free electrons. When free electrons attain sufficient velocity, they collide with
and ionize neutral gas molecules. Ionization occurs when the force of the collision removes
an electron from the gas molecule, resulting hi a positively charged gas molecule and another
free electron. These newly-freed electrons are also accelerated and cause additional
ionization.2.
The corona can be either positive or negative; but the negative corona is used hi most
industrial ESPs since it has inherently superior electrical characteristics that enhance collection
efficiency under most operating conditions.3
5.2-3
-------
5.2.1.3 Particle Charging
Particle charging hi an ESP (and subsequent collection) takes place in the region
between the boundary of the corona glow and the collection electrode, where gas particles are
subject to the generation of negative ions from the (negative) corona process.3
Upon entering the ESP, the uncharged dust particles suspended hi the effluent gas
stream are exposed to a region of space filled with ions and, in the case of negative corona,
perhaps some free electrons. As these electrical charges approach the electrically neutral dust
particles, an induced dipole is established in the paniculate matter by the separation of charge
within the particles.2 As a dipole, the particle itself remains neutral while positive and
negative charges within the particle concentrate within separate areas. The positive charges
within the particle are drawn to the area of the particle closest to the approaching negative ion.
As a negative ion contacts the paniculate matter, the induced positive charges will retain some
electrical charge from the ion. This results hi a net negative charge on the previously neutral
paniculate. The presence of an electrical charge is required in order for the electric field to
exert a force on the particle and remove the particulate from the gas stream.2
Charging is generally done by both field and diffusion mechanisms. The dominant
mechanism varies with particle size. In field charging, ions from the corona are driven onto
the particles by the electric field. As the ions continue to impinge on the dust particles, the
charge on it increases until the local field developed by the charge on the particle causes a
distortion of the electric field lines so that they no longer intercept the particle and no further
charguig takes place. This is the dominant mechanism for particles larger than about 0.5
Diffusion charging is associated with ion attachment resulting from random thermal
motion; this is the dominant charguig mechanism for particles below about 0.2 jim. As with
field charguig, diffusion charguig is influenced by the magnitude of the electric field, since ion
movement is governed by electrical as well as diffusional forces. Neglecting electrical forces,
diffusion charguig results when the thermal motion of molecules causes them to diffuse
through the gas and contact the particles. The charguig rate decreases as the particle acquires
charge and repels additional gas ions, but charguig continues to a certain extent.3
The particle size range of approximately 0.2 to 0.5 jum is a transitional region hi which
both charguig mechanisms are present but neither dominates. Fractional efficiency test data
for ESPs have shown reduced collection efficiency hi this transitional size range, where
diffusion and field charguig overlap.3
5.2.1.4 Particle Collection
The final step hi particle collection hi an ESP involves the movement of the charged
particles towards an oppositely-charged electrode that holds the particles hi place until the
5.2-4
-------
electrode is cleaned. Typically, the collection electrodes are parallel flat plates or pipes that
are cylindrical, square, or hexagonal.2
The movement of particles toward the collection electrode is driven by the electric
field. The motion of larger particles (greater than 10 to 20 //m) will more or less follow a
trajectory determined by the average gas velocity and average particle electrical velocity.2 The
trajectory for smaller particles ( < 10 ^m) will be less direct, since the inertial effects of the
turbulent gas flow predominate over the electrical velocity induced by the relatively smaller
electric charge. The overall movement of smaller particles, however, will be towards the
collection electrode. The cumulative collection efficiency of an ESP is generally dependent
upon the fractional collection efficiency of these smaller particles, especially between 0.2 to
2.0 Mm in size.2
5.2.2 Penetration Mechanisms
There are several conditions which can reduce the effectiveness of ESPs and lead to
penetration of paniculate. These conditions include back corona, dust reentrainment, erosion,
saltation, and gas sneakage.
5.2.2.1 Back Corona
Back corona or reverse ionization describes the conditions where an electrical
breakdown occurs in an ESP. Normally hi an ESP, a corona is formed at the discharge
electrode, creating electrons and negative ions which are driven toward the (positive)
collection electrode by the electric field. This situation is reversed if the corona is formed at
the (positive) collection electrode. A corona at this electrode generates positive ions that are
projected into the interelectrode space and driven toward the discharge electrode.2
As the positive ions flow into the interelectrode space hi an ESP, they encounter
negatively charged paniculate and negative ions. The electric field from the charged
paniculate exceeds that of an ion at most distances. Therefore, the majority of the positive
ions flow toward the negatively-charged dust particles, neutralizing their charge. This
neutralization of charge causes a proportionate reduction hi the electrical force acting to collect
these particles.2
A second mechanism by which back corona may be disruptive to ESP collection is due
to a neutralization of a portion of the space charge that contributes to the electric field adjacent
to the collection electrode. The space charge component of the electric field near the
collection zone may be as much as 50 percent of the total field. Neutralization of the space
charge reduces the total collection force by the same fraction.2
5.2-5
-------
5.2.2.2 Dust Reentrainment
Dust reentrainment associated with dry ESP collection may occur after the dust layer is
rapped clear of the plates. The first opportunity for rapping reentrainment occurs when the
dust layer begins to fall and break up while falling. Dust particles are swept back into the
circulating gas stream. The second opportunity occurs as the dust falls into the hopper,
impacts the collected dust, and puffs up to form a dust cloud. Portions of this dust cloud are
picked up by the circulating gas stream. Some of the dust may be recollected.2
Direct erosion of the collected dust from the collection electrode can occur when gas
velocities exceed 10 feet per second (fps). Most ESPs have gas velocities less than 8 fps,
while newer installations have velocities less than 4 fps. Saltation is theorized to be a minor
form of reentrauiment which occurs as particles are collected. As a particle is captured and
strikes the collection electrode, it may loosen other particles which are resuspended hi the gas
stream. Other causes of reentrainment hi an ESP are electric sparking, air leakage through the
hopper, and electrical reentrainment associated with low resistivity particles.2
5.2.2.3 Dust Sneakage
The construction of an ESP is such that nonelectrified regions exist in the top of the
ESP where the electrical distribution, plate support, and rapper systems are located.
Similarly, portions of the collection hopper and the bottom of the electrode system contain
nonelectrified regions. Particle-laden gas streams flowing through these regions will not be
subjected to the collection forces and tend to pass through the ESP uncollected. The amount
of gas sneakage and bypassing through nonelectrified regions will place an upper limit on the
collection efficiency of an ESP.2
5.2.3 Types of Electrostatic Precipitators
Electrostatic precipitators are generally divided into two broad groups, dry ESPs and
wet ESPs. The distinction is based on what method is used to remove paniculate from the
collection electrodes. In both cases, particulate collection occurs hi the same manner. In
addition to wet and dry options, there are variations of internal ESP designs available. The
two most common designs are wire-plate and wire-pipe collectors. Electrostatic precipitators
are often designed with several compartments, to facilitate cleaning and maintenance.
5.2.3.1 Dry ESPs
Dry ESPs remove dust from the collection electrodes by vibrating the electrodes
through the use of rappers. Common types of rappers are gravity impact hammers and electric
vibrators. For a given ESP, the rapping intensity and frequency must be adjusted to optimize
performance. Sonic energy is also used to assist dust removal hi some dry ESPs. The main
components of dry ESPs are an outside shell to house the unit, high voltage discharge
5.2-6
-------
electrodes, grounded collection electrodes, a high voltage source, a rapping system, and
hoppers. Dry ESPs can be designed to operate hi many different stream conditions,
temperatures, and pressures. However, once an ESP is designed and installed, changes hi
operating conditions are likely to degrade performance.li2-3
5.2.3.2 Wet ESPs
The basic components of a wet ESP are the same as those of a dry ESP with the
exception that a wet ESP requires a water spray system rather than a system of rappers.
Because the dust is removed from a wet ESP hi the form of a slurry, hoppers are typically
replaced with a drainage system. Wet ESPs have several advantages over dry ESPs. They
can adsorb gases, cause some pollutants to condense, are easily integrated with scrubbers, and
eliminate reentrainment of captured particles. Wet ESPs are not limited by the resistivity of
particles since the humidity in a wet ESP lowers the resistivity of normally high resistivity
particles.2'4
Previously, the use of wet ESPs was restricted to a few specialized applications. As
higher efficiencies have currently become more desirable, wet ESP applications have been
increasing. Wet ESPs are limited to operating at stream temperatures under approximately
170°F. In a wet ESP, collected paniculate is washed from the collection electrodes with water
or another suitable liquid. Some ESP applications require that liquid is sprayed continuously
into the gas stream; hi other cases, the liquid may be sprayed intermittently. Since the liquid
spray saturates the gas stream hi a wet ESP, it also provides gas cooling and conditioning.
The liquid droplets in the gas stream are collected along with particles and provide another
means of rinsing the collection electrodes. Some ESP designs establish a thin film of liquid
which continuously rinses the collection electrodes.2-3
5.2.3.3 Wire-Plate ESPs
Wire-plate ESPs are by far the most common design of an ESP. In a wire-plate ESP, a
series of wires are suspended from a frame at the top of the unit. The wires are usually
weighted at the bottom to keep them straight. In some designs, a frame is also provided at the
bottom of the wires to maintain then" spacing. The wires, arranged in rows, act as discharge
electrodes and are centered between large parallel plates, which act as collection electrodes.
The flow areas between the plates of wire-plate ESPs are called ducts. Duct heights are
typically 20 to 45 feet.2 A typical wire-plate ESP is shown in Figure 5.2-2.2
Wire-plate ESPs can be designed for wet or dry cleaning. Most large wire-plate ESPs,
which are constructed on-site, are dry. Wet wire-plate ESPs are more common among smaller
units that are pre-assembled and packaged for delivery to the site.4 In a wet wire-plate ESP,
the wash system is located above the electrodes.2
5.2-7
-------
-------
5.2.3.4 Wire-Pipe ESPs
In a wire-pipe ESP, a wire that functions as the discharge electrode runs through the
axis of a long pipe, which serves as the collection electrode. The weighted wires are
suspended from a frame in the upper part of the ESP. The pipes can be cylindrical, square, or
hexagonal. An example of a wire-pipe design is provided in Figure 5.2-3. Previously, only
cylindrical pipes were used; square and hexagonal pipes have currently grown in popularity.
The space between cylindrical tubes creates a great deal of wasted collection area. Square and
hexagonal pipes can be packed closer together, so that the inside wall of one tube is the outside
wall of another.4 This situation is illustrated in Figure 5.2-4.
Wire-pipe collectors are very effective for low gas flow rates and for collecting mists.
They can use dry or wet cleaning methods, but the vast majority are cleaned by a liquid wash.
As with wire-plate collectors, the cleaning mechanism in a wire-pipe ESP is located above the
electrodes. These pipes are generally 6 to 12 inches in diameter and 6 to 15 feet in length.2
5.2.3.5 Other ESP Designs
Rigid-Frame Plate. This ESP design is very similar to the wire-plate ESP, with the
exception that the discharge electrode is a rigid frame, rather than a series of weighted wires,
that is placed between plates. The frame supports wire discharge electrodes. This type of
ESP operates in the same manner as the wire-plate and can be wet or dry. In general, the
rigid frame design is more durable than weighted wires, but has higher initial (capital)
expense.2-3 Rigid frames have become the preferred design in some industries, such as pulp
and paper.5 Figure 5.2-5 provides an example of a rigid frame-plate ESP.
Wide-Plate Spacing.6 The flow areas between the plates of a conventional wire-plate
ESP usually vary from 8 to 12 inches in width. A recent enhancement hi these units has been
wide-plate spacings of up to 20 niches. Wide spacing gives a higher collecting field strength
due to the resultant increase hi space charge, a more uniform current density, and higher
migration velocities. More variation in the discharge electrode geometry is also possible with
wide-plate spacing. Because of the increased efficiency associated with this technique, less
plate area is needed, thereby reducing the overall size and cost of the ESP.7
Electrode Variations.1 -2 In addition to the rigid frames, there are several other
variations of electrodes that are not as common. In some cases, completely rigid discharge
electrodes are preferred over weighted wires or rigid frames with wires.1 Other discharge
electrode designs are square wires, barbed wires, serrated strips of metal, and strips of metal
with needles at regular intervals. The barbs, serration, and needles on the discharge electrodes
help to establish a uniform electric field. In some cases, flat plates are used both as discharge
and collection electrodes. Collection electrodes are often modified with baffles to improve gas
flow and particle collection. Some ESPs use wire mesh rather than flat plates as collection
electrodes. Examples of discharge electrodes and collection plates are shown in Figure 5.2-6.
5.2-9
-------
High Voltoa.* Initilator
Comportment
To CUcn Co* Main
T«n*ioft
Support From*
EUctred*
Waighl
Figure 5.2-3. Wire-Pipe Electrostatic Precipitator (Reference 2).
5.2-10
-------
Discharge
\
Discharge Electrodes
Square Pipes
Hexagonal Pipes
Discharge Electrodes
Wasted Area
Circular Pipes
Figure 5.2-4. Square, Hexagonal, and Circular Pipe Arrangements for Wire-Pipe
Precipitators (adapted from Reference 4).
5.2-11
-------
iriMt support Support
Figure 5.2-5. Rigid Frame Electrode (Reference 2).
5.2-12
-------
Discharge Electrodes
ii
Collection Plate Electrodes
Baffle Plates
VWWWNA
VWNAAAAA
Zig Zag Plates
Tulip Plates
Plates
Figure 5.2-6. Various Discharge Electrodes and Collection Plate Designs (Reference 2).
5.2-13
-------
Concentric Plate.3 In this design, the ESP consists of vertical cylinders that are
arranged concentrically and act as collection electrodes. The walls of the cylinders are
continually rinsed by a thin film of liquid which is supplied by a system above the electrodes.
The discharge electrodes are made of wire mesh located between the cylinders. This type of
ESP is only operated as a wet ESP. The gas stream is wetted hi a scrubber before it reaches
the ESP. The concentric plate ESP is illustrated hi Figure 5.2-7.
Pulsed Energization.2 Some ESPs have experienced success with pulsed energization.
Conventional ESPs rely on a constant base voltage applied to the discharge electrode to
generate the corona and electric field. In pulse energization, high voltage pulses of short
duration (of a few microseconds) are applied to the discharge electrodes. A typical pulse
energization system will operate with pulse voltages on the order of 100 kilovolts (kV) rather
than the 50 kV used with conventional energization. The pulses produce a more uniformly
current distribution on the collection electrode.8 Pulses can be used alone or in addition to a
base voltage and have been shown to increase the collection efficiency of ESPs with poor
energization. Pulse energization has been used successfully hi the electric utility industry. The
Ion Physics Corp. has performed tests of this procedure at Madison Gas and Electric,
Madison, Wisconsin.9 This technique is, however, still evolving to permit a more rational
approach to pulse energization and, perhaps, to reduce the cost.6
Two-Stage ESP.2-3 All of the ESP designs mentioned previously have been single-stage
ESPs. In a single stage ESP, particle charging and collection take place simultaneously hi the
same physical location. Two-stage ESPs are different hi that particle charging takes place in a
separate section which precedes collection. Two-stage ESPs are best suited for low dust
loadings and fine particles. It is often used for cleaning air hi buildings.
5.2.4 Collection Efficiency
Electrostatic precipitators are capable of collecting greater than 99 percent of all sizes
of paniculate.1 Collection efficiency is effected by several factors including dust resistivity,
gas temperature, chemical composition (of the dust and gas), and particle size distribution.
The resistivity of a dust is a measure of its resistance to electrical conduction and it has
a great effect on the performance of dry ESPs. The efficiency of an ESP is limited by the
strength of the electric field it can generate, which hi turn is dependent upon the voltage
applied to the discharge electrodes. The maximum voltage that can be applied is determined
by the sparking voltage. At this voltage, a path between the discharge and collection
electrodes is ionized and sparkhig occurs. Highly resistive dusts increase sparking, which
forces the ESP to operate at a lower voltage. The effectiveness of an ESP decreases as a result
of the reduced operating voltage.2
5.2-14
-------
fitt FlflH
MHO MO STACK
TMMSITIM
ftKIMTATOi
KTAi
cucrnoc
OISCMMtt CMS.
COKTIHUOUS nU» OF
L10UOI riOHS OOW
eoutcrioi
rucT«oi
(CTLIMC*
IH.CT TO
UCTIOH
wsc
Figure 5.2-7. Concentric Plate Electrostatic Precipitator (Reference 3).
5.2-15
-------
High resistivity dusts also hold their electrical charge for a relatively long period of
time. This characteristic makes it difficult to remove the dust from the collection electrodes.
In order to loosen the dust, rapping intensity must be increased. High intensity rapping can
damage the ESP and cause severe reentrainment, leading to reduced collection efficiency.
Low dust resistivities can also have a negative impact on ESP performance. Low resistivity
dust quickly loses its charge once collected. When the collection electrodes are cleaned, even
with light rapping, serious reentrainment can occur.2
Temperature and the chemical composition of the dust and gas stream are factors which
can influence dust resistivity. Current is conducted through dust by two means, volume
conduction and surface conduction. Volume conduction takes place through the material itself,
and is dependent on the chemical composition of the dust. Surface conduction occurs through
gases or liquids adsorbed by the particles, and is dependent on the chemical composition of the
gas stream. Volume resistivity increases with increasing temperatures and is the dominant
resistant force at temperatures above approximately 350°F. Surface resistivity decreases as
temperature increases and predominates at temperatures below about 250°F. Between 250 and
350°F, volume and surface resistivity exert a combined effect, with total resistivity highest in
this temperature range.2-3
For coal fly ash, surface resistance is greatly influenced by the sulfur content of the
coal. Low sulfur coals have high resistivity, because there is decreased adsorption of
conductive gases (such as SO3) by the fly ash. The collection efficiency for high-resistance
dusts can be improved with chemical flue gas conditioning that involves the addition of small
amounts of chemicals into the gas stream (discussed in Section 5.1, Pretreatment). Typical
chemicals include sulfur dioxide (SO2), ammonia (NH3), and sodium carbonate. These
chemicals provide conductive gases which can substantially reduce the surface resistivity of the
fly ash.7-10 Resistivity can also be reduced by the injection of steam or water into the gas
stream.2
In general, dry ESPs operate most efficiently with dust resistivities between 5 x 103
and 2 x 1010 ohm-cm.2 Electrostatic precipitator design and operation is difficult for dust
resistivities above 1011 ohm-cm.2 Dust resistivity is generally not a factor for wet ESPs.1-2
The particle size distribution impacts on the overall performance of an ESP. In general, the
most difficult particles to collect are those with aerodynamic diameters between 0.1 and 1.0
fji,m. Particles between 0.2 and 0.4 /xm usually show the most penetration. This is most likely
a result of the transition region between field and diffusion charging. Figure 5.2-8 provides
cumulative collection efficiency curves for ESPs operating in the utility, copper, and iron and
steel industries. The curves were derived from emission factors.11 Table 5.2-1 presents the
cumulative collection efficiencies for PMj0 and PM25.
5.2-16
-------
Table 5.2-1. PM10 and PM25 Cumulative Collection Efficiencies
for ESPs at Coal Combustors, Primary Copper Operations,
and Iron and Steel Production Operations (from Reference 11)
Collection Efficiency (percent)
Application PM10 PM25
Coal-Fired Boilers
Dry bottom (bituminous) 97.7 96.0
Spreader stoker (bituminous) 99.4 97.7
Spreader stoker (anthracite) 98.4 98.5
Primary Copper Production
Multiple hearth roaster 99.0 99.1
Reverberatory smelter 97.1 97.4
Iron and Steel Production
Open hearth furnace 99.2 99.2
Sinter oven . 94.0 90.0
5.2.5 Applicability
Approximately 80 percent of all ESPs in the U.S. are used in the electric utility
industry. Many ESPs are also used in pulp and paper (7 percent), cement and other minerals
(3 percent), iron and steel (3 percent), and nonferrous metals industries (1 percent).1
Table 5.2-2 lists common applications of ESPs.12
The dust characteristics can be a limiting factor hi the applicability of dry ESPs to
various industrial operations. Sticky or moist particles and mists can be easily collected, but
often prove difficult to remove from the collection electrodes of dry ESPs. Dusts with very
high resistivities are also not well suited for collection hi dry ESPs. Dry ESPs are susceptible
to explosion in applications where flammable or explosive dusts are found.2
5.2-17
-------
Sl-Z'S
Cumulative Collection Efficiency (%)
s $ 5 $ s i
Cumulative Collection Efficiency (%)
I •..
r
•n
D
8
5
O V O •
a S si
II Ii
Cumulative Collection Efficiency (%)
W O O C30 >
II II II H II
(T)
i a
OQ
CD
O! tU
Bs
2. era
§ S
I. s.
S o
o>
2
o 3
.
o 2. o
£ o a
«r
or>
8'
1
ra
•B
§
a
00
f t
II I!
•S3 •? i-
p
a
-------
Wet ESPs can collect sticky particles and mists, as well as highly resistive or explosive
dusts. Wet ESPs are generally not limited by dust characteristics, but are limited by gas
temperatures. Typically, the operating temperatures of wet ESPs cannot exceed 170°F. When
collecting a valuable dust which can be sold or recycled into the process, wet ESPs also may
not be desirable, since the dust is collected as a wet slurry that would likely need additional
treatment.2-4
Electrostatic precipitators are usually not suited for use on processes which are highly
variable, since frequent changes in operating conditions are likely to degrade ESP
performance. Electrostatic precipitators are also difficult to install on sites which have limited
space because ESPs must be relatively large to obtain the low gas velocities necessary for
efficient particle collection.1
Table 5.2-2. Typical Industrial Applications
of Electrostatic Precipitators (from References 2 and 12)
Application
Source Category
Code
Type of ESP"
Utility Boilers
(Coal, Oil)
Industrial Boilers
(Coal, Oil, Wood, Liq. Waste)
Commercial/Institutional
Boilers
(Coal, Oil, Wood)
Chemical Manufacture
Non-Ferrous Metals Processing
(Primary and Secondary)
Copper
Lead
Zinc
Aluminum
1-01-002...004
1-02-001...005
1-02-009,-013
1-03-001...005
1-03-009
3-01-001...999
3-03-005
3-04-002
3-03-010
3-04-004
3-03-030
3-04-008
3-03-000...002
3-04-001
DESP, Wire-Plate
DESP, Wire-Plate
DESP, Wire-Plate
Site specific
DESP, WESP, Plate-Plate, Wire-Plate,
Wire-Pipe, Rigid Frame-Plate
DESP, WESP, Plate-Plate, Wire-Plate,
Wire-Pipe, Rigid Frame-Plate
DESP, WESP, Plate-Plate, Wire-Plate,
Wire-Pipe, Rigid Frame-Plate
DESP, WESP, Wire-Plate, Wire-Pipe
Rigid Frame-Plate
5.2-19
-------
Table 5.2-2, continued
Application
Source Category
Code
Type of ESP"
Other
Ferrous Metals Processing
Coke Production
Ferroalloy Production
Iron and Steel Production
Gray Iron Foundries
Steel Foundries
Petroleum Refineries and
Related Industries
Mineral Products
Cement Manufacturing
Stone Quarrying and Processing
Other
Wood, Pulp, and Paper
Incineration
(Municipal Waste)
3-03-011...014
3-04-005...006
3-04-010...022
3-03-003...004
3-03-006...007
3-03-008...009
3-04-003
3-04-007, -009
3-06-001...999
3-05-006...007
3-05-020
3-05-003...999
3-07-001
5-01-001
DESP, WESP, Wire-Plate, Wire-Pipe
WESP, Wire-Pipe
DESP, Wire-Plate
DESP, WESP, Wire-Plate, Wire-Pipe
DESP, Wire-Plate
DESP, WESP, Wire-Plate, Wire-Pipe
DESP, Wire-Plate
DESP, Wire-Plate
Site specific
DESP, WESP, Wire-Plate, Needle-Plate
DESP, Wire-Plate, Rigid Frame-Plate
DESP, Wire-Plate, Rigid Frame-Plate
• DESP = Dry ESP, WESP = Wet ESP.
5.2.6 Costs of Electrostatic Precipitators
The costs of installing and operating an ESP include both capital and annual costs.
Capital costs are all of the initial equipment-related costs of the ESP. Annual costs are the
direct costs of operating and maintaining the ESP for one year, plus such indirect costs as
overhead; capital recovery; and taxes, insurance, and administrative charges. Please refer to
Chapter 6 of the OAQPS Control Cost Manual for cost equations.13
5.2.6.1
Capital Costs
The total capital investment (TCI) for ESPs includes all of the initial capital costs, both
direct and indirect. Direct capital costs are the purchased equipment costs (PEC), and the
5.2-20
-------
costs of installation (foundations, electrical, piping, etc.). Indirect costs are related to the
installation and include engineering, construction, contractors, start-up, testing, and
contingencies. The direct and indirect installation costs are calculated as factors of the PEC.13
Table 5.2-3 presents the TCI cost factors for ESPs. There are several aspects of ESPs which
impact the PEC. These factors include inlet gas flow rate, collection efficiency, dust and gas
characteristics, and various standard design features. The PEC is estimated based on the ESP
specifications and is typically correlated with the collecting area in two ways, the Deutsch-
Anderson equation or the sectional method.13 Please refer to Chapter 6 of the OAQPS Cost
Manual (Reference 13) for ESP cost estimation equations.
Inlet Flow Rate. The inlet flow rate has the greatest effect on TCI because it
determines the overall size of the ESP. As the gas flow rate increases so does the ESP size
and, in turn, the costs. Typical gas flow rates for ESPs are 10,000 to 1,000,000 actual cubic
feet per minute (ACFM).2 Electrostatic precipitator costs increase approximately linearly with
gas flow rate, with the slope of the cost curves dependent on the other factors discussed below.
Collection Efficiency. Electrostatic precipitators are designed to achieve a specific
collection efficiency. The TCI costs of ESPs increase as greater efficiencies are achieved. To
attain higher collection efficiencies, ESPs must be larger to provide greater collection areas.
In addition, extremely high efficiencies may require special control instrumentation and
internal modifications to improve gas flow and rapping efficiency. Figure 5.2-9 shows the
effect of collection efficiency on TCI costs for an ESP.14
Dust Characteristics. Particle size distribution, adhesiveness, and resistivity are dust
characteristics that affect ESP costs. The size distribution of the dust influences the overall
ESP collection efficiency. For example, particles in the range of 0.1 to 1.0 /*m are the most
difficult for an ESP to collect. If many of the particles are in this range, it will be more
difficult to achieve a given collection efficiency and a larger, more expensive ESP will be
required. If the dust is very sticky, dry ESPs will need to be made of more durable (and
costly) materials to withstand the intense rapping needed to remove the dust from the
collection electrodes. For this reason, a wet ESP is often preferred for very sticky dusts,
which drives costs higher. Dust resistivity influences costs, since highly resistive particles
will require the added operating expense of flue gas conditioning or the use of wet ESPs.13
Gas Stream Characteristics. Important gas stream characteristics are temperature,
moisture, and chemical composition. Gas stream temperature affects particle resistivity and,
consequently, ESP efficiency and costs. Very moist streams and mists generally require the
use of wet ESPs. The chemical composition of the gas stream may restrict the construction
materials appropriate for the ESP. Most ESPs are constructed of carbon steel; however when
the stream is highly corrosive, more costly corrosion resistant materials such as stainless steel,
carpenter, monel, nickel, and titanium are needed.13 Figure 5.2-10 shows the impact of the
use of corrosion resistant materials on ESP TCI costs.14
5.2-21
-------
Table 5.2-3. Capital Cost Factors for Electrostatic Precipitators (from Reference 10)
Cost Item Factor
Direct Costs
Purchased equipment costs
ESP + auxiliary equipment As estimated (A)
Instrumentation 0.10 A
Sales taxes 0.03 A
Freight 0.05 A
Total Purchased Equipment Cost, (PEC) B = 1.18 A
Direct installation costs
Foundations and supports - 0.04 B
Handling and erection 0.50 B
Electrical 0.08 B
Piping 0.01 B
Insulation for ductwork 0.02 B
Painting 0.02 B
Total direct installation cost 0.67 B
Site Preparation and Buildings As required (Site)
Total Direct Cost, DC 1.67 B + Site
Indirect Costs (installation)
Engineering 0.20 B
Construction and field expense 0.20 B
Contractor fees 0.10 B
Start-up 0.01 B
Performance test 0.01 B
Model study 0.02 B
Contingencies 0.03 B
Total Indirect Cost (1C) 0.57 B
Total Capital Investment = DC + 1C 2.24 B + Site
Design Features. There are several design features that are considered standard for
most ESPs and which can add up to 50 percent of the PEC. These options include inlet and
outlet nozzles, diffuser plates, hopper auxiliaries (heaters, level detectors, etc.), weather
enclosures, stair access, structural supports, and insulation.13 Figure 5.2-11 shows ESP costs
with and without these standard design features.14 Wet ESPs and rigid-frame designs typically
have higher initial (capital) expenses than dry and wire-plate ESPs.
5.2-22
-------
a
i
a.
a
U
30
25
20
15
10
0.0
.99.9%
Efficiency
99.5%
Efficiency
-99.0%
Efficiency
95.0%
Efficiency
Note:
Costs are referenced to
fourth quarter 1996.
02 0.4 0.6
Inlet Flowrate (acfmxlO6)
0.8
1.0
Figure 5.2-9. Effect of Design Collection Efficiency on ESP TCI Costs (Reference 14).
o
»—H
X
2
'£,
3
100
90
80
70
60
50
40
30
20
10
0.0
0.2 0.4 0.6
Inlet Flowrate (acfinx 106)
0.8
.Titanium
Nickel
1.0
Monel
Carpenter
Stainless Steel
Carbon Steel
Note:
Costs are referenced
to fourth quarter
1996, and are for 99.
percent efficiency.
Figure 5.2-10.
Effect of the Use of Corrosion Resistant Materials on ESP TCI Costs
(Reference 14)
5.2-23
-------
25
20
§
3
'a.
I
15
10
5
0.0
0.2 0.4 0.6
Inlet Flowrate (acfm x 106)
0.8
.With
Standard
Options
Without
Standard
Options
Note:
Costs are referenced to
fourth quarter 1996,
and are for 99.5
percent efficiency.
1.0
Figure 5.2-11.
TCI Costs for ESPs With and Without Various Standard Design Features
(Reference 14).
5.2.6.2
Annual Costs
The total annual cost of an ESP consists of both direct and indirect costs. Direct
annual costs are those associated with the operation and maintenance of the ESP. These
include labor (operating, supervisory, coordinating, and maintenance), maintenance materials,
operating materials, electricity, dust disposal, wastewater treatment (wet ESPs), compressed
air (for rappers), conditioning agents, and heating or cooling costs.13 Some operating costs are
not applicable to all ESPs. For ESPs collecting dusts which have no value, dust disposal can
be expensive. Gas conditioning agents are used for ESPs that need to collect highly resistive
dusts. Some ESP installations also require heating or cooling of the gas stream for effective
operation. The cost of the heating fuel can be significant; cooling water costs generally are
not.
13
Indirect annual costs include taxes, insurance, administrative costs, overhead, and
capital recovery. All of these costs except overhead are dependent on the TCI. Table 5.2-4
lists the annual cost parameters that impact ESP costs, with typical values provided for each
parameter. Table 5.2-5 provides the annual cost factors for ESPs. It is difficult to generalize
these costs for all ESPs, since annual costs are very site-specific.13
5.2-24
-------
Table 5.2-4. Annual Cost Parameters for Electrostatic Precipitators (Reference 14).
Parameter
Description
Typical Values
"
Direct Cost Parameters
Operating factor (OF)
Operator labor rate (OR)
Operator shift factor (OS)
Supervisor labor factor (SF)
Coordinator labor factor (CF)
Maintenance labor (ML)
Maintenance materials factor (MF)
Electricity rate (ER)
Chemicals (C)
Compressed air (CA)
Wastewater treatment (W)
Waste disposal (D)
Indirect Cost Parameters
Overhead factor (OV)
Annual interest rate (I)
Operating life (n)
Capital recovery factor (CRF)
Taxes (TAX)
Insurance (INS)
Administrative costs (AC)
Hours of scrubber operation per year
Operator labor pay rate
Fraction of operator shift on scrubber
Fraction of operator labor cost
Fraction of operator labor cost
Dependent on plate collector area
Fraction of Purchased Equipment Cost
Cost of electricity
Cost of chemical conditioning agents
Cost of compressed air for rappers
Cost of treating wet ESP effluent
Cost of disposing of dust/sludge
Fraction of total labor and (MM) costs
Opportunity cost of the capital
Expected operating life of scrubber
Function of (n) and (I)
Fraction of the TCId
Fraction of the TCId
Fraction of the TCId
8,640 hr/yr
$12.50/hra
0.25b
0.15b
0.33"
Site specific
0.01b
$0.07/kW-hra
Site specific (sect. 5.1)
$0.18/1000 scf
$1.55-$2.55/1000gala
$20-30/tona
0.60"
7 percent11
20 years'1
0.0944C
0.01"
0.01b
0.02b
Estimated for 1996 from currently available information.
Estimates from "CO$T-AIR" Control Cost Spreadsheets (Reference 14).
Capital Recovery Factor is calculated from the following formula: CRF = {/(I + f)} + {(1 +1) "- 1},
where / = interest rate (fraction) and n = operating life (years).
The total capital investment (TCI) can be escalated to current values by using the Vatavuk Air Pollution Control Cost Indicies
(VAPCCI), described in Appendix B.
-------
Table 5.2-5. Annual Cost Factors for Electrostatic Precipitators (Reference 14).
Cost Item
Direct Costs
Labor
Operator (OL)
Supervisor (SL)
Coordinator (CL)
Maintenance (ML)
Maintenance materials (MM)
Electricity (E)
Chemicals (C)
Compressed air (CA)
Wastewater treatment (W)
Waste disposal (D)
Formula"
(OF) x (OR) x (OS)
(SF)x(OL)
(CF)x(OL)
Site specific
(MF)X(PEC)
Power11 x (ER)
Site specific
(CA)
(W)
(D)
Factor
A
0.15 A
0.33 A
ML
0.01 PEC
E
C
CA
W
D
Total Direct Cost (DC)
Indirect Costs
Overhead
Capital Recovery
Taxes
Insurance
Administrative Costs
Total Indirect Cost (1C)
Total Annual Cost (DC + 1C)
1.48 A + ML + 0.01 PEC + E + C + CA + W + D
(OV)X(OL+SL+CL
+ML+MM)
(CRF)x(TCI)
(TAX) X (TCI)
(INS) x (TCI)
(AC) x (TCI)
0.89 A + 0.6 ML
+ 0.006 PEC
0.1424 TCI
0.01 TCI
0.01 TCI
0.02 TCI
0.89 A +0.6 ML + 0.006 PEC + 0.1824 TCI
2.37 A + 1.6 ML + 0.016 PEC + 0.1824 TCI + E +
C + CA + W + D
a Includes values also described in Table 5.2-4.
b Equal to total power requirements, e.g. fan, pump, etc.
5.2-26
-------
5.2.7 Energy and Other Secondary Environmental Impacts
The environmental impacts of ESP operation include those associated with energy
demand, solid waste generation hi the form of the collected dust, and water pollution for wet
ESPs. The energy requirements for operation of an ESP consist mainly of electricity demand
for fan operation, and electric field generation, and cleaning. Fan power is dependent on the
pressure drop across the ESP, the flow rate, and the operating tune. Assuming a fan-motor
efficiency of 65 percent and a ratio of the gas specific gravity to that of air equal to 1.0, the
fan power requirement can be estimated from the following equation:13
Fan Power (kW-hr/yr) = 1.81 x lO"4 (V)(AP)(t) (Eq. 5.2-1)
where Vis gas flow rate (ACFM), AP is pressure drop (niches H2O), t is annual operating
tune (hr/yr), and 1.81 x 10"4 is a unit conversion factor.
The operating power requirements for the electrodes and the energy for the rapper
systems can be estimated from the folio whig relationship:13
Operating Power (kW-hr/yr) = 1.94 x 10'3 (A)(f) (Eq. 5.2-2)
where A is ESP plate area (ft2), t is annual operating time (in hr/yr), and 1.94 x 10~3 is a unit
conversion factor.
Wet ESPs have the additional energy requirement of pumping the rinse liquid into the
ESP. Pump power requirements can be calculated as follows:13
Pump Power (kW-hr/yr) = (0.746(Q,XZ)(Sg)(t)) I (3,960 rj) (Eq. 5.2-3)
where Qt is the liquid flow rate (gal/min), Z is the fluid head (ft), Sg is the specific gravity of
the liquid, t is the annual operating time (hr/yr), TJ is the pump-motor efficiency, and 0.746
and 3,960 are unit conversion factors.
Solid waste is generated from ESP operation in the form of the collected dust.
Although the dust is usually inert and nontoxic, dust disposal is a major factor of ESP
operation. With some ESP operations, the dust can be reused in the process or on the facility
or sold. Otherwise, the dust must be shipped off site. Water pollution is a concern for wet
ESPs. Some installations may require water treatment facilities and other modifications to
handle the slurry discharge from wet ESPs.2'13
5.2-27
-------
5.2.8 References for Section 5.2
1. Cooper, C.D and F.C. Alley. Air Pollution Control: A Design Approach. 2nd ed.
Waveland Press, Prospect Heights, Illinois. 1994.
2. The Electrostatic Precipitator Manual (Revised). The Mcllvaine Company,
Northbrook, Illinois. March 1996.
3. Control Techniques for Paniculate Emissions from Stationary Sources - Volume 1
(EPA-450/3-81-005a, NTIS PB83-127498). U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards. Research Triangle Park, North
Carolina. September 1982.
4. Steinsvaag, R. Overview of Electrostatic Precipitators. Plant Engineering. July 10,
1995.
5. Mastropietro, T. and P. Dhargalkar. Electrostatic Precipitator Designs Evolve to meet
Tighter Regulations. Pulp & Paper. September 1991.
6. Oglesby, S., Jr. Future Directions of Particulate Control Technology: A Perspective.
J. Air Waste Management Assoc. 40(8): 1184-1185. August 1990.
7. Scholtens, M.J. Air Pollution Control: A Comprehensive Look. Pollution
Engineering. May 1991.
8. C. Sedman, N. Plaks, W. Marchant, and G. Nichols. Advances in Fine Particle
Control Technology. Presented at the Ukraine Ministry of Energy and Electrification
Conference on Power Plant Air Pollution Control Technology, in Kiev, The Ukraine,
September 9 - 10, 1996.
9. ESP Newsletter. The Mcllvaine Company, Northbrook, Illinois. May 1996.
10. Rikhter, L.A. et. al. Improving the Efficiency of Removal of High-resistance Ash in
Electrostatic Precipitators by Chemical Conditioning of Flue Gases. Thermal
Engineering. 38:3. March 1991.
11. Compilation of Air Pollutant Emission Factors (AP-42). Volume I (Fifth Edition).
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina.
January 1995.
12. Source Category Emission Reductions with Particulate Matter and Precursor Control
Techniques. Prepared for K. Woodard, U. S. Environmental Protection Agency,
Research Triangle Park, North Carolina (AQSSD/IPSG), under Work Assignment 11-16
(EPA Contract No. 68-03-0034), "Evaluation of Fine Particulate Matter Control."
5.2-28
-------
September 30, 1996.
13. OAQPS Control Cost Manual (Fourth Edition, EPA 450/3-90-006) U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards.
Research Triangle Park, North Carolina. January 1990.
14. Vatavuk, W.M. "CO$T-AIR" Control Cost Spreadsheets. Provided by the Innovative
Strategies and Economics Group of the Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina.
February 1996.
5.2-29
-------
5.3 FABRIC FILTERS
This section addresses the basic operating principles of fabric filters, the cleaning
methods, fabric selection, costs, and applicability to various industries. Fabric filters are a
popular means of separating particles from a gas stream because of their relatively high
efficiency and applicability to many situations. Fabric filters can be made of either woven or
felted fabrics and may be in the form of sheets, cartridges, or bags, with a number of the
individual fabric filter units housed together hi a group. Bags are by far the most common
type of fabric filter, hence the use of the term "baghouses" to describe fabric filters hi general.
The major particle collection mechanisms of fabric filters are inertial impaction,
diffusion from Brownian motion, and interception. During fabric filtration, dusty gas is drawn
through the fabric by forced-draft fans. The fabric is responsible for some filtration, but more
significantly it acts as support for the dust layer that accumulates. The layer of dust, also
known as a filter cake, is a highly efficient filter, even for submicrometer particles. Woven
fabrics rely on the filtration abilities of the dust cake much more than felted fabrics.
Fabric filters possess some key advantages over other types of particle collection
devices. Along with the very high collection efficiencies, they also have the flexibility to treat
many types of dusts and a wide range of volumetric gas flows. Fabric filters can be operated
with low pressure drops. Fabric filters also have some potential disadvantages. In general,
they are limited to filtering dry streams. Also, high temperatures and certain chemicals can
damage some fabrics. Fabric filters also have the potential for fire or explosion, and can
require a large area for installation.1 Proper design can rnhiimize or eliminate these
disadvantages.
5.3.1 Particle Collection and Penetration Mechanisms
Particle capture during fabric filtration is mainly due to some combination of inertial
impaction, diffusion, and direct interception. Collection may also occur due to gravitational
sedimentation and electrostatic attraction, but usually to a lesser extent.2 Figure 5.3-1
illustrates these five particle collection mechanisms.3
Inertial impaction occurs as a result of a change hi velocity between a fluid, such as
air, and a particle suspended hi the fluid. As the fluid approaches an obstacle it will accelerate
and change direction to pass around the object. Depending on the mass of the particle, it may
not be able to adapt to the fluid acceleration and a difference hi velocity will develop between
the particle and fluid stream. Inertia will maintain the forward motion of the particle towards
the object, but the fluid will attempt to drag the particle around the obstacle. The resultant
particle motion is a combination of these forces of fluid drag and inertia. This results hi
impaction for the particles where inertia dominates, and by-pass for those particles
overwhelmed by fluid drag.2
5.3-1
-------
DIRECT
INTERCEPTION
DIFFUSION
INCRTIAL
IMPACTION
ELECTROSTATIC
ATTRACTION
— GRAVITATIONAl
SETTLING
Figure 5.3-1. Collection Mechanisms of Fabric Filtration (from Reference 3).
5.3-2
-------
Collection by diffusion occurs as a result of both fluid motion and the Brownian
(random) motion of particles. Diffusional collection effects are most significant for particles
less than 1 micrometer (/zrn) in diameter.2 Another collection mechanism, direct interception,
occurs when a particle comes within one particle radius of an obstacle. The path that the
particle takes can be a result of inertia, diffusion, or fluid motion.2
Gravitational sedimentation, i.e. the falling of individual or agglomerated particles, is a
minor collection mechanism for fabric filter operations.2 Electrostatic charge can play an
important role in particle collection and agglomeration hi some situations. In order to
maximize the electrostatic effect, the characteristics of the particles must be understood before
the fabric is selected. See Section 6, Emerging Technologies, for more discussion of
electrostatic effects in fabric filtration.
Because of the physics of each collection mechanism, the particle size will determine
the predominance of one collection mechanism over another. Generally, as particle size
decreases, the predominance of the diffusion collection mechanism increases, assuming other
parameters remain constant. As particle size increases, the impaction collection mechanism
will most likely increase. The combination of these two major particle collection effects
contributes to a minimum efficiency at a given particle size, as illustrated in Figure 5.3-2, a
plot of fractional fabric filter collection efficiency versus aerodynamic particle size.2
The fabric itself is also a factor hi particle collection and penetration. In the initial
stages of filtration where the fabric is usually bare, the fabric is responsible for some
filtration. More significantly, however, it acts as support for the dust layer that accumulates
over the course of operation of the fabric filter. The dust or filter cake is a highly efficient
filter, even for submicrometer particles. In terms of fabric type, woven fabrics rely on the
filtration abilities of the dust cake much more than felted fabrics.
The structure of the fabric, particularly for woven fabrics, is also very important to
particle collection. Large pores and a high free-space area within the fabric contribute to low
particle removal. Particle capture in woven fabrics is enhanced by small fibers (known as
fibrils) which project into the pores. Dust can deposit on the fibrils and bridge across the
pores, which allows a filter cake to build up and increases collection efficiency. Fabrics can
have similar pore sizes and very different collection characteristics because of the number of
fibrils they possess. The electrostatic properties of fibers are also critical. Different fibers
have different electrostatic and surface characteristics. The intensity of the electrostatic charge
of the fabric has a distinct effect on particle collection efficiency and is a function of the fabric
properties and surface roughness. The resistivity of the fabric influences charge dissipation
once particles have been captured. The rate of charge dissipation affects how the dust releases
from the fabric and how easily the fabric can be cleaned.
The gas-to-cloth (G/C) ratio is an important design consideration and has a major effect
on particle collection mechanisms. This is a ratio of the volumetric flow rate of gas per unit
5.3-3
-------
N)
I
I
W
I
o
I
•
sr
Fractional Efficiency (%)
o
*—
Gl
GO
B'
I
P
b
rt
in
B'
58
.1
4
£
-------
of filtering area, and is usually expressed in the units of cubic feet per minute of gas per
square foot of fabric [(ft3/min)/ft2]. Since these units can be reduced to feet per minute
(ft/min), the G/C ratio is also referred to as the face velocity.4 In general, as the face velocity
increases, the efficiency of impaction collection increases and diffusional collection efficiency
decreases.2 Higher face velocities allow for smaller fabric filters, all other things being
constant. However, as the face velocity increases, there is increased pressure drop, increased
particle penetration, blinding of fabric, more frequent cleaning, and reduced bag life. Table
5.3-1 shows recommended G/C ratios for various industrial dusts.4-16
The majority of the dust that penetrates the filter is a result of dust that is dislodged
during cleaning and penetrates to the clean side, and dust that is loosened during cleaning and
cannot resist dislodging when the flow resumes.2 The majority of submicrometer particles
penetrate the fabric by passing directly through the pores or by seepage.2 Seepage occurs
when particles migrate through the filter cake and the fabric by continuous capture and
reentrainment. Seepage is more common with smooth particles and with a lack of significant
electrostatic forces.2
5.3.2 Types of Fabric Filters
There are a wide variety of materials which can be woven or felted into effective
fabrics, and there are many different sizes and arrangements of bags that can be utilized.
Although the presence of a filter cake increases collection efficiency as the cake becomes
thicker, it also restricts the flow of gas. This increases the pressure drop and energy
requirements. To operate a fabric filter continuously, the dust must be cleaned from the filters
and removed from the fabric filter on a regular basis. Fabric filters are frequently classified
by their cleaning method. The three major types of fabric filter cleaning mechanisms are
mechanical shaker, reverse-air, and pulse-jet. These types are discussed below along with a
brief discussion of other less common types of cleaning methods and fabric filter
configurations.
5.3.2.1 Shaker-Cleaned Fabric Filters
Shaking has been a popular cleaning method for many years because of its simplicity as
well as its effectiveness. Shaker-cleaned fabric filters utilizing specially chosen woven fabrics
are more effective than other types of fabric filters for many applications.2 For small units,
shaking can be accomplished manually. Large fabric filters require mechanical shaking. For
both cases, the operation is basically the same. In general, dusty gas enters an inlet pipe to the
shaker-cleaned fabric filter and very large particles are removed from the stream when they
strike the baffle plate hi the inlet duct and fall into the hopper. The particulate-laden gas is
drawn from beneath a cell plate in the floor and into the filter bags. The gas proceeds from
the inside of the bags to the outside and through the outlet pipe. The particles are collected on
the inside surface of the bags and a filter cake accumulates.
5.3-5
-------
Table 5.3-1. Recommended Gas-to-Cloth Ratios (acfm/ft2)
for Common Industrial Applications of Fabric Filters
(References 4 and 13)
Dust
Alumina
Asbestos
Bauxite
Carbon Black
Coal
Cocoa
Clay
Cement
Cosmetics
Enamel Frit
Feeds, Grain
Feldspar
Fertilizer
Flour
Fly Ash
Graphite
Gypsum
Iron Ore
Iron Oxide
Iron Sulfate
Lead Oxide
Leather Dust
Lime
Limestone
Mica
Paint Pigment
Paper
Plastics
Quartz
Rock Dust
Sand
Sawdust
Silica
Slate
Soap, Detergents
Spices
Starch
Sugar
Talc
Tobacco
Zinc Oxide
Shaker or Reverse-Air
Woven Fabric
2.5
3.0
2.5
1.5
2.5
2.8
2.5
2.0
1.5
. 2.5
3.5
2.2
3.0
3.0
2.5
2.0
2.0
3.0
2.5
2.0
2.0
3.5
2.5
2.7
2.7
2.5
3.5
2.5
2.8
3.0
2.5
3.5
2.5
3.5
2.0
2.7
3.0
2.0
2.5
3.5
2.0
Pulse-Jet
Felted Fabric
8
10
8
5
8
12
9
8
10
9
14
9
8
12
5
5
10
11
7
6
6
12
10
8
9
7
10
7
9
9
10
12
7
12
5
10
8
7
10
13
5
5.3-6
-------
A typical mechanical shaker-cleaned fabric filter unit is shown in Figure 5.3-3.1 In
mechanical shaking units, the tops of the bags are attached to a shaker bar. When the bags
are cleaned, the bar is moved briskly, usually in a horizontal direction. This movement flexes
the fabric, causing the dust cake to crack and .fall away from the fabric and into the hopper. A
typical shaker mechanism is shown hi Figure S.3-4.2 Some amount of filter cake will remain
on the inside of the filter bag; as discussed above, this is desirable and also necessary to
maintain a consistently high collection efficiency. The amount of dust that is removed during
cleaning can be controlled by regulating the frequency, amplitude, and duration of the shaking
cycles. In some designs, reverse-air flow is used to enhance dust removal.
The flow of gas through the bags must be stopped during the cleaning cycle to allow
the filter cake to release from the fabric and to prevent dust from working through the bag
during the shaking. In order to accomplish this, shaker-cleaned fabric filters are often
designed with several separate compartments. Each compartment can then be isolated from
the gas flow and cleaned while the other compartments continue to filter the stream.
Shaker-cleaned fabric filters are very flexible hi design, allowing for different types of
fabrics, bag arrangements, and fabric filter sizes. This enables shaker-cleaned fabric filters to
have many applications, with only some limitations. Shaker-cleaning fabric filters need a dust
that releases fairly easily from the fabric, or the fabric will be damaged from over shaking and
bag failure will result. Glass fabrics hi particular are susceptible to degradation from
shaking.2 Most other filter fabrics are less brittle than glass and have longer service lives in
shaker-cleaned applications. The shaker mechanism itself also must be well designed and
maintained or it will quickly wear and lose effectiveness. As the shaker mechanism loses
effectiveness, the operator will often increase shaking intensity hi order to clean the bags
satisfactorily. Continuing this practice can eventually destroy the shaking mechanism.2
5.3.2.2 Reverse-Air Cleaned Fabric Filter
Reverse-air cleaning is another popular fabric filter cleaning method that has been used
extensively and unproved over the years.5 It is a gentler but sometimes less effective cleaning
mechanism than mechanical shaking.1 Most reverse-air fabric filters operate hi a manner
similar to shaker-cleaned fabric filters. The bags are open on the bottom, closed on top and
the gas flows from the inside to the outside of the bags with dust being captured on the inside.
However, some reverse-air designs collect dust on the outside of the bags. In either design,
reverse-air cleaning is performed by forcing clean air through the filters in the opposite
direction of the dusty gas flow. The change hi direction of the gas flow causes the bag to flex
and crack the filter cake. In internal cake collection, the bags are allowed to collapse to some
extent during reverse-air cleaning. The bags are usually prevented from collapsing entirely by
some kind of support, such as rings that are sewn into the bags. The support enables the dust
cake to fall off the bags and into the hopper. Cake release is also aided by the reverse flow of
the gas. Because felted fabrics retain dust more than woven fabrics and thus, are more
difficult to clean, felts are usually not used hi reverse-air systems.2
5.3-7
-------
SHAKER
MECHANISM
OUTUT
PIPE
INIET
PIPE
HOPPER
Figure 5.3-3. Cutaway View of a Typical Shaker Fabric Filter (from Reference 1).
5.3-8
-------
SHAKING LEVER
BAGS
• ' ECCENTRIC
r-.::;:^
TjTO^ •- ^ DUST OUT
5.3-4. Typical Shaker Mechanism (from Reference 2).
5.3-9
-------
There are several methods of reversing the flow through the filters. As with
mechanical shaker-cleaned fabric filters, the most common approach is to have separate
compartments within the fabric filter so that each compartment can be isolated and cleaned
separately while the other compartments continue to treat the dusty gas. A typical design of
one compartment of a reverse-air cleaning fabric filter is shown in Figure 5.3-5.2 One method
of providing the reverse flow air is by the use of a secondary fan or cleaned gas from the other
compartments. A second method is with a traveling air mechanism. An example of such a
mechanism is shown hi Figure 5.S-6.3 In this design, the dust is collected on the outside of the
bags. The air manifolds rotate around the fabric filter and provide reverse air to each bag,
allowing most of the bags to operate while a few of the bags are being cleaned.
Reverse-air cleaning alone is used only in cases where the dust releases easily from the
fabric. In many instances, reverse-air is used in conjunction with shaking or pulsing. A
relatively recent development has been the use of sonic horns to aid cleaning (see
Section 5.3.2.4). During cleaning, sonic blasts from horns mounted in the fabric filter assist
in the removal of dust from the bags. This is an important enhancement to fabric filtration.6-7
Sonic assistance is a very popular method for fabric filters at coal-burning utilities.1
5.3.2.3 Pulse-Jet Cleaned Fabric FUter
Pulse-jet cleaning of fabric filters is relatively new compared to other types of fabric
filters, since they have only been used for the past 30 years. This cleaning mechanism has
consistently gamed hi popularity because it can treat high dust loadings, operate at constant
pressure drop, and occupy less space than other types of fabric filters.8 Pulse-jet cleaned
fabric filters can only operate as external cake collection devices. A schematic of a pulse-jet
cleaned fabric filter is shown in Figure 5.3-7.1 The bags are closed at the bottom, open at the
top, and supported by internal retainers, called cages. Particulate-laden gas flows into the bag,
with diffusers often used to prevent oversized particles from damaging the bags. The gas
flows from the outside to the inside of the bags, and then out the gas exhaust. The particles
are collected on the outside of the bags and drop into a hopper below the fabric filter.
During pulse-jet cleaning a short (0.03 to 0.1 second) burst of high pressure (90 -
100 psig) air is injected into the bags. The pulse is blown through a venturi nozzle at the top
of the bags and establishes a shock wave that continues on to the bottom of the bag. The wave
flexes the fabric, pushing it away from the cage, and then snaps it back dislodging the dust
cake. The cleaning cycle is regulated by a remote tuner connected to a solenoid valve. The
burst of air is controlled by the solenoid valve and is released into blow pipes that have
nozzles located above the bags. The bags are usually cleaned row by row.
5.3-10
-------
,, REVERSE-
-------
MHCM CLCAM
•IOW NIVC.4M ,*!?_,
OWVE MOTOM AM MANIFOLD OUTUT
OUTM MOW
•ivtnst AM
MAWK>LO
INLET
MIAWOUST
DROPOUT
MIOOLC HOW
MCVCNtt AIM
MAMIFOLQ
Figure 5.3-6. Reverse-Air Fabric filter with Travelling Mechanism and External Cake
Collection (from Reference 3).
5.3-12
-------
Compressed air
manifold 100 psig
Remote cyclic—-I I
timer '—'
Solenoid
valve
Filter bag
Tube sheet
Venturi
Dirty air
\—J in
Figure 5.3-7. Schematic of a Pulse-Jet Fabric Filter with Enlarged View of Pulse Inlet Area
(Reference 1).
5.3-13
-------
There are several unique attributes of pulse-jet cleaning. Because the cleaning pulse is
very brief, the flow of dusty gas does not have to be stopped during cleaning. The other bags
continue to filter, taking on extra duty because of the bags being cleaned.9 In general, there is
no change in fabric filter pressure drop or performance as a result of pulse-jet cleaning. This
enables the pulse-jet fabric filters to operate on a continuous basis with solenoid valves as the
only significant moving parts.2 Pulse-jet cleaning is also more intense and occurs with greater
frequency than the other fabric filter cleaning methods. This intense cleaning dislodges nearly
all of the dust cake each time the bag is pulsed. As a result, pulse-jet filters do not rely on a
dust cake to provide filtration. Felted fabrics are used in pulse-jet fabric filters because they
do not require a dust cake to achieve high collection efficiencies. It has been found that woven
fabrics used with pulse-jet fabric filters leak a great deal of dust after they are cleaned.
Since bags cleaned by pulse-jet do not need to be isolated for cleaning, pulse-jet
cleaning fabric filters do not need extra compartments to maintain adequate filtration during
cleaning. Also, because of the intense and frequent nature of the cleaning, they can treat
higher gas flow rates with higher dust loadings. Consequently, fabric filters cleaned by pulse
jet can be smaller than other types of fabric filters in the treatment of the same amount of gas
and dust, making higher gas-to-cloth ratios achievable.
A disadvantage of pulse-jet units that use very high gas velocities is that the dust from
the cleaned bags can be drawn immediately to the other bags.2 If this occurs, little of the dust
falls into the hopper and the dust layers on the bags becomes too thick. To prevent this, pulse-
jet fabric filters can be designed with separate compartments that can be isolated for cleaning.
5.3.2.4 Other Fabric Filter Designs
The less common fabric filter designs of reverse-jet, vibrational, and sonic cleaning,
and cartridge filters are briefly described below.
Reverse-Jet Cleaning:2 Reverse-jet fabric filters have internal cake collection and
employ felted fabrics. Each bag is cleaned by a jet ring which travels up and down the outside
of the bag on a carriage. The rings blow a small jet of moderately pressurized air through the
felt, dislodging the dust on the inside of the bags. Reverse-jet designs are generally used when
high-efficiency collection is required for fine particles at low dust loadings, such as toxic or
valuable dusts. This fabric filter cleaning method mechanism provides high efficiency at high
G/C ratios, but its industrial application seems to be declining.
Vibration Cleaning:2 Vibration cleaning is similar to mechanical shaker cleaning.
However, in vibrational cleaning, the tops of the bags are attached to one plate, rather than a
series of shaker bars as with mechanical shaker cleaning. To clean the bags, the plate is
oscillated in a horizontal direction at a high frequency. This creates a ripple hi the bags which
dislodges the filter cake. Vibration cleaning is most effective for medium- to large-sized
5.3-14
-------
particles with weak adhesive properties, therefore this cleaning method is limited to
applications where fine particle collection is not needed.
Sonic Cleaning:6-10 Sonic cleaning is generally used to assist another cleaning method,
such as reverse air cleaning. Sonic horns are installed inside the fabric filter compartments,
where the bags are periodically blasted with sonic energy. The frequency and amplitude of the
soundwaves can be adjusted to maximize the effect for a given dust. The soundwave shock
causes a boundary layer to form in the filter cake; this allows more of the cake to be dislodged
during cleaning and, hence, improves cleaning efficiency. Over half of the reverse gas fabric
filters also use sonic horns, either continuously or intermittently.11
Cartridge Dust Collectors:2-12 Cartridge collectors are pleated fabrics that are contained
in completed closed containers, or cartridges. These collectors offer high efficiency filtration
combined with a significant size reduction in the fabric filter unit. A cartridge filter occupies
much less space than filter bags with the same amount of filtration media. In addition,
cartridge collectors can operate at higher G/C ratios than fabric filters. Cartridges can be
pulse cleaned, and some types can be washed and reused. Cartridge replacement is also much
simpler than filter bag replacement. However, this type of fabric filter has been limited to low
flow rate and low temperature applications. New filter materials and collector designs are
increasing the applications of cartridge filters.
5.3.3 Fabric Characteristics
Fabric selection is a very important feature of fabric filter operation. There are many
fibers that can be used effectively used as filters, with different properties that determine their
appropriate applications. In general, fibers can be made into woven or felted fabrics. The
cleaning method affects the fiber choice, since some fibers wear quickly and lose then-
effectiveness as a result of frequent flexing or shaking. The fabric type must also fit the
cleaning method, and the stream and particle characteristics. Woven fabrics are preferred for
shaker and reverse-air fabric filters. Felted fabrics are recommended for pulse-jet and reverse-
jet fabric filters. The use of felt is generally limited to external surface dust collection styles.2
The major gas stream characteristics to consider when selecting fabrics are temperature
and chemical composition. Most fabrics are degraded by high temperatures.2 Among the
variety of available fabrics, there is a wide range of maximum operating temperatures that can
be matched to the range of temperatures in the different applications. Some fabrics are also
easily degraded by acids, whereas others are highly resistant to acids. Alkalis, oxidizers, and
solvents are other types of chemicals that can damage filter materials.2 New fibers, such as
Ryton®, Gore-Tex®, and Chem-Pro®, are continually in development for high temperature and
other demanding applications.13 Ceramic fabrics, Nextel® for example, have recently been
developed and can function at temperatures up to 1000°F.13'14 Table 5.3-2 lists the maximum
operating temperatures, and physical and chemical resistances of various industrial-used
fabrics.2
5.3-15
-------
Table 5.3-2. Temperature Ranges, and Physical and Chemical Resistances of
Common Industrial Fabrics (from Reference 2 and 14)
Fabric Type
Cotton
Dacron
Orion
Nylon
Dynel
Polypropylene •
Creslan
Vycron
Nomex®
Teflon*
Nextel™
Wool
Glass
Maximum
Operating
Temperature
<°F)
180
275
275
225
160
200
275
300
400
450
1000
215
550
Physical Resistance
Dry
Heat
G
G
G
G
F
G
G
G
E
E
E
F
E
Moist
Heat
G
F
G
G
F
F
G
F
E
E
E
F
E
Abrasion
F
G
G
E
F
E
G
G
E
F-P
NA
G
P
Shaking
G
E
G
E
F-P
E
G
E
E
G
NA
F
P
Flexing
G
E
E
E
G
G
E
E
E
G
NA
G
F
Mineral
Acids
P
G
G
P
G
E
G
G
F-P
E
E
F
E
Chemical Resistance
Organic
Acids
G
G
G
F
G
E
G
G
E
E
E
F
E
Alkalies
F
F
F
G
G
E
F
G
G
E
E
P
G
Oxidizers
F
G
G
F
G
G
G
G
G
E
E
P
E
Solvents
E
E
E
E
G
G
E
E
E
E
E
F
E
Note: E = Excellent, G = Good, F = Fair, P = Poor, NA = Not available.
-------
The important particle characteristics to consider in fabric selection are size, abrasion
potential, and release potential. The average sizes of the particles can be a factor in the
selection of the type of weave or felt that is chosen for a particular application. With very
abrasive dusts, care must be taken to insure that the fabric will not wear out too quickly.
Moist or sticky dusts require a fabric that will easily release the dust cake, or that is coated
with some type of lubricant layer.2
Several different finishes and textures have been developed for fiberglass fabrics to
increase their use in filtration. There are also many coatings and chemical treatments available
to provide lubrication and other properties to fibers to improve then* performance.
5.3.4 Collection Efficiency
Well-designed and maintained fabric filters that are operated correctly should collect
greater than 99 percent of particles ranging hi size from submicrometer to hundreds of
micrometers.1 There are several factors which can affect the collection efficiency of fabric
filters. These factors include gas filtration velocity, particle characteristics, fabric
characteristics, and cleaning mechanism. In general, collection efficiency increases with
increasing filtration velocity and particle size. Other particle characteristics, as well as the
type of cleaning method, are key variables in fabric filter design. An improperly designed
fabric filter will not function as well as possible and will oftentimes impact efficiency.
For a given combination of filter design and dust, the effluent particle concentration
from a fabric filter is nearly constant whereas the overall efficiency of a fabric filter is more
likely to vary with paniculate loading.2 For this reason, fabric filters can be considered
constant outlet devices rather than constant efficiency devices. Constant effluent concentration
is achieved because at any given time part of the fabric filter is being cleaned. Unlike
cyclones, scrubbers, and electrostatic precipitators, fabric filters never really achieve a steady
state of particle collection.2 As a result of the cleaning mechanisms used in fabric filters, the
collection efficiency at a given time is always changing. Each cleaning cycle removes at least
some of the filter cake and loosens particles which remain on the filter. When filtration
resumes, the filtering capability has been reduced because of the lost filter cake and loose
particles are pushed through the filter by the flow of gas. This reduces the collection
efficiency. As particles are captured the efficiency increases until the next cleaning cycle.
Average collection efficiencies for fabric filters are usually determined from tests that cover a
number of cleaning cycles at a constant inlet loading.2
Earlier, Figure 5.3-2 showed typical fractional collection efficiency curves versus
particle size for fabric filters. Figure 5.3-8 shows cumulative collection efficiency curves for
fabric filters in operation hi the utility, ferroalloys, and the iron and steel industry industries,
respectively, that were calculated from reported test data.15 The collection efficiency data for
PM10 and PM25 are provided in Table 5.3-3.
5.3-17
-------
COAL-FIRED BOILERS
IRON AND STEEL PRODUCTION
I
I—»
00
u
$
— • — Spreader, Bituminous,
Data Quality C
— -•—•Spreader, Anthracite,
Data Quality. D
•••*•• Dry Bottom, Bituminous,
Data Quality C,E
1 10
Particle Size (jim)
100
FERROALLOY ELECTRIC ARC FURNACES
100
99
98
97
96
95
0
t
L.A *- - - « __
'""••. ""
*l
1" • •
'••*•*
— • — SOV.FeMn,
Data Quality B
--»--50y.FeSi,
Data Quality: B
• - -A • • Si Metal,
Data Quality. B
1 1 10 100
Particle Size (pm)
100
98
%
92
88
0
i
/ 1
A
,/"*"""
r~"
— • — Iron and Steel,
Desulfunzation,
Data Quality D,E
• • -A - - Gray Iron,
Cupolas,
Data Quality C,E
1 1 10 100
Particle Size (urn)
Note: Data quality refers to the data quality ratings
assigned to the emission factors from which the
efficiencies were calculated (from Reference 8).
A = excellent
B = above average
C = average
D = below average
E = poor
Figure 5.3-8. Cumulative Collection Efficiency Data for Fabric Filters at Coal Combustors, Ferroalloy Electric Arc
Furnaces, and Iron and Steel Production Operations (from Reference 15).
-------
Table 5.3-3. PM10 and PM25 Cumulative Collection Efficiencies for
Fabric Filters at Coal Combustors, Ferroalloy Electric Arc Furnaces, and
Iron and Steel Production Operations (from Reference 15).
Collection Efficiency (percent)
Application PMi0 PM:
•2.5
Coal-fired Boilers
Dry bottom (bituminous) 99.2 98.3
Spreader stoker (bituminous) 99.9 99.3
Spreader stoker (anthracite) 99.4 98.4
Ferroalloy Electric Arc Furnaces
Iron silicate 97.0 97.6
Iron manganese 98.3 98.7
Silica 96.3 96.9
Iron and Steel Production
Desulfurization 96.7 96.8
Gray iron cupolas 93.9 93.4
5.3.5 Applicability
Fabric filters can perform very effectively in many different applications. The variety
of designs and fabrics allows for adaptability to most situations. For most applications, there
are several combinations of cleaning method and filter fabric that are appropriate. Table 5.3-4
lists common applications of fabric filters and their recommended cleaning methods and
fabrics. Table 5.3-1 above provided the recommended G/C ratios for various applications of
fabric filters. There are also empirical methods for determining G/C ratios for a given
application, which are described hi Chapter 5 of the OAQPS Cost Manual.
Although fabric filters can be used hi many different conditions, there are some factors
which limit then- applications. The characteristics of the dust are one factor. Some particles
are too adhesive for fabric filters. While such particles are easily collected, they are too
difficult to remove from the bags. Particles from oil combustion are an example of a very
sticky dust, most of which is thought to be heavy hydrocarbons. For this reason, fabric filters
are not recommended for boilers which fire oil exclusively;4 however, fabric filters are often
used with boilers which fire oil as a secondary fuel.
5.3-19
-------
The potential for explosion is also a concern for certain fabric filters applications.
Some fabrics are flammable, and some dusts and stream components may form explosive
mixtures. If a fabric filter is chosen to control explosive mixtures, care must be taken when
designing and operating the fabric filters to eliminate conditions which could ignite the dust,
the stream, and the bags. In addition, the fabric filters should be designed to prevent operator
injuries in the event of an explosion.
Temperature and humidity are also limiting factors in the use of fabric filters.
Currently, there are few fabric filters hi applications where temperatures exceed 500°F for
long periods of time. However, new fibers which can operate at temperatures in the 900 to
1000°F range are commercially available and hi use at some installations. An example of such
a fabric is the ceramic fabric Nextel®. This fabric is very effective, but is also very expensive
and is priced much higher than Teflon®, the most expensive of the commonly used filter
fabrics.13 The high cost of new filter fabrics may discourage the use of fabric filters hi very
high temperature applications. Humidity can also be a problem when considering fabric
filters. Moist particles can be difficult to clean from the bags and can bridge over and clog the
hopper.1 Streams with high humidity can also require baghouses with msulation to maintain
temperatures well above the dew point to prevent condensation.
Table 5.3-4. Typical Cleaning Methods and Fabrics
for Industrial Applications of Fabric Filters (from Reference 2)
Application
Utility Boilers
(Coal)
Source
Category
Code
1-01-002... 003
Typical
Cleaning
Methods3
MS, PJ,
RA with SA
Typical Fabrics"
Fiberglass, Teflon, Teflon treated Glass
Industrial Boilers
(Coal, Wood)
Commercial/Institutional
Boilers
(Coal, Wood)
Non-Ferrous Metals
Processing (Primary and
Secondary)
Copper
Lead
Zinc
1-02-001...003
1-02-009
1-03-001...003
1-03-009
3-03-005
3-04-002
3-03-010
3-04-004
3-03-030
3-04-008
PJ, RA Fiberglass, Teflon, Teflon treated Glass
PJ, RA Fiberglass, Teflon, Teflon treated Glass
MS, RA Fiberglass, Dacron, Polypropylene, Nomex,
Teflon
MS, RA Polypropylene, Nomex, Teflon, Dacron,
Orion
MS, RA Polypropylene, Nomex, Teflon, Dacron
5.3-20
-------
Table 5.3-4. (continued)
Application
Source
Category
Code
Typical
Cleaning
Methods1
Typical Fabrics'1
Aluminum
Other
Ferrous Metals
Processing
Coke
Ferroalloy
3-03-000...002
3-04-001
3-03-011...014
3-04-005...006
3-04-010...022
3-03-003...004
3-03-006...007
MS, RA Nomex, Dacron, Teflon, Polypropylene
MS, RA Nomex, Dacron, Teflon, Polypropylene
Iron and Steel 3-03-008.. .009
Gray Iron
Foundries
Steel Foundries
3-04-003
3-04-007
3-04-009
MS, RA Dacron Combination, Teflon
RA, MS Fiberglass (coated with graphite, silicone,
Teflon), Dacron, Nomex
RA, MS, PJ Fiberglass with lubricant (silicone, graphite,
Teflon)
MS, RA Fiberglass (with silicone), Nomex, Orion
RA, MS, PJ Fiberglass, Nomex, Dacron, Teflon
Mineral Products
Cement
Coal Cleaning
Stone Quarrying
and Processing
Other
Asphalt Manufacture
Grain and Feed Milling
3-05-006...007
3-05-010
3-05-020
3-05-003...999
(except above)
3-05-001...002
MS, PJ, RA Dacron, Cotton, Wool, Nomex, Fiberglass
MS, PJ, RA Cotton, Orion, Dacron
MS, PJ, RA Fiberglass, Nomex, Teflon, Orion, Nylon,
Dacron, Dynel, Cotton
PJ, RA, MS Nomex, Fiberglass
PJ, RA Dacron
MS = Mechanical Shaking, RA = Reverse Air, PJ = Pulse Jet, SA = Sonic Assistance
Fabrics are not specified as woven or felted. Felted fabrics are generally recommended for
pulse-jet cleaning, woven fabrics are usually utilized for reverse-air and mechanical shaker
cleaning.
5.3-21
-------
5.3.6 Costs of Fabric Filters
The costs of installing and operating a fabric filter includes capital and annual costs.
Capital costs include all of the initial equipment-related costs of the fabric filter. Annual costs
are the direct costs of operating and maintaining the fabric filter for one year, plus such
indirect costs as overhead; capital recovery; and taxes, insurance, and administrative charges.
The following sections discuss capital and annual costs for various fabric filter designs. Capital
costs have been referenced to the third quarter of 1995. The major design consideration,
regarding cost, is the G/C ratio (G/C). The G/C is dependent on several factors and must be
optimized to balance the capital costs, hi terms of the fabric filters size, and the annual
operating costs, in particular the pressure drop.1
5.3.6.1 Capital Costs
The total capital investment (TCI) for fabric filters includes all of the initial capital
costs of the fabric filter, both direct and indirect. Direct capital costs are the purchased
equipment costs (PEC) and the costs of physically installing the equipment (foundations and
supports, electrical wiring, piping, etc.). Indirect capital costs are also related to installation
and include engineering, contractor fees, start-up, testing, and contingencies. The PEC is
dependent upon the fabric filter design specifications; direct and indirect installation costs are
generally calculated as a factors of the PEC.16 Commonly used factors for estimating fabric
filter capital costs are provided in Table 5.3-5.
There are several design factors which influence fabric filter PEC, and in turn the TCI
of fabric filters. Important factors include the inlet gas flow rate, the cleaning mechanism, the
type of dust, the dust loading, particle characteristics, gas stream characteristics, and fabric
type. Please refer to Chapter 5 of the OAQPS Control Cost Manual for cost equations.16
Gas Flow Rate: The inlet flow rate has the greatest impact on the costs of a fabric
filter, since it affects the necessary fabric filter size. For any one fabric filter cleaning type, as
the gas flow rate increases so does the fabric filter size and, consequently^ the costs. Fabric
filters typically treat flow rates from 10,000 to over 1,000,000 acfm.2 Although fabric filter
costs increase approximately linearly with gas flow rate, the slope of the cost curve depends on
the other design features that are discussed below.
Cleaning Mechanism: The fabric filter cleaning mechanism is the next most important
design feature hi terms of costs. Figure 5.3-9 presents cost versus gas flow rate curves for
different fabric filters cleaning mechanisms with Nomex® fabric treating similar streams.17
Cost curves for two different types of pulse-jet fabric filters, common-housing and modular,
are provided hi Figure 5.3-9, along with reverse ah* and mechanical shaker cleaning cost
information. Pulse-jet common-housing fabric filters are fabric filter that not taken off-line for
cleaning. Modular pulse-jet fabric filters are constructed with bags hi separate compartments
which can be taken off-line for cleaning.16 Figure 5.3-9 illustrates that reverse ah- and
5.3-22
-------
Table 5.3-5. Capital Cost Factors for Fabric Filters (from Reference 16)
Cost Item Factor
Direct Costs
Purchased equipment costs
Fabric filter + bags + auxiliary equipment As estimated (A)
Instrumentation 0.10 A
Sales taxes 0.03 A
Freight 0.05 A
Total purchased equipment cost (PEC) B = 1.18 A
Direct installation costs
Foundations and supports 0.04 B
Handling and erection 0.50 B
Electrical 0.08 B
Piping 0.01 B
Insulation for ductwork 0.07 B
Painting 0.02 B
Total direct installation cost 0.72 B
Site Preparation and Buildings As required (Site)
Total Direct Cost (DC) 1.72 B + Site
Indirect Costs (installation)
Engineering 0.10B
Construction and field expense 0.20 B
Contractor fees 0.10 B
Start-up 0.01 B
Performance test 0.01 B
Contingencies 0.03 B
Total Indirect Cost (1C) 0.45 B
Total Capital Investment = DC + 1C 2.17B + Site
mechanical shaker cleaning fabric filters have higher TCI costs than pulse-jet (with reverse air
units the highest), largely due to the much lower G/C ratios, which raises capital costs. In
terms of pulse-jet cleaning fabric filters, modular pulse-jet units are slightly more expensive
than common-housing units in terms of TCI.
Although different cleaning mechanisms can operate over different ranges of G/C
ratios, pulse-jet fabric filters generally operate at higher G/C ratios compared to shaker and
reverse-air models. For this reason, pulse-jet fabric filters are usually smaller (with lower TCI
costs) than other fabric filter designs that treat the same flow rate. However, the cleaning
mechanism is not chosen simply because of the resultant fabric filter size. Some cleaning
5.3-23
-------
O
1
H
18.0
16.0
14.0
12.0
10.0
8.0
6.0
4.0
2.0
0.0
-- Reverse- Air,
G/C = 2
____ Mechanical Shaker,
G/C = 2
- Pulse- Jet (modular),
OX? = 4.5
....... Pulse- Jet (common),
Note:
Costs are referenced to
fourth quarter 1996, and
include Nomex filter bags.
0.0 0.2 0.4 0.6
Inlet Flowrate (acfinx 106)
0.8
1.0
Figure 5.3-9. Effect of Cleaning Mechanism on Fabric Filter TCI (Reference 17).
mechanisms may not be recommended for certain dust types (See Table 5.3-4 above). In
addition, the choice of cleaning mechanism also affects the choice (and resultant costs) for
fabric and auxiliary equipment. When choosing between cleaning mechanisms, the PEC is
calculated for all applicable designs to determine the least expensive option.
Gas-to-Cloth Ratio: Dust type is most responsible for determining the correct G/C
ratio for a particular fabric filter. Each combination of dust and fabric filter cleaning method
has a recommended G/C ratio that in most cases has been arrived at through actual fabric filter
operations. For a given flow rate, a higher G/C ratios will result hi a smaller fabric filter and
lower TCI costs. Figure 5.3-10 shows typical cost curves for mechanical shaker and pulse-jet
fabric filters operating at G/C ratios appropriate for each cleaning type.17 Although
mechanical shaker cost curves are generally higher than pulse-jet (modular) cost curves, a
fabric filter with mechanical shaker cleaning and a high G/C ratio (curve C) has similar TCI
costs to a pulse-jet cleaned fabric filter with a low G/C ratio (curve D).
Dust Loading: The dust loading is a measure of the amount of dust per volume of gas
being treated that is generally expressed as the weight of dust per unit volume of gas
(e.g. grams per cubic foot (g/ft3)). While the type of dust generally determines the best G/C
ratio, the dust loading may cause adjustments to the recommended ratio. For high dust
loadings, the G/C ratio should be decreased so that more fabric is available to handle the high
dust levels.1 With low dust loadings, the G/C ratio can be increased, which hi turn will
reduce the fabric filter size.
5.3-24
-------
o
X
J5
U
I
18.0
16.0
14.0
12.0
10.0
8.0
6.0
4.0
2.0
0.0
0.0 0.2 0.4 0.6
Inlet Flowrate (acfinx 10s)
0.8
1.0
Mechanical Shaker
A. GC=1.5
E.G/C = 2.5
C. G/C = 3.5
Pulse-Jet (modular):
D. G/C = 5
E. GC = 9
F. G/C =13
Note:
Costs are referenced to
fourth quarter 1996, and
include Nomex filter bags.
Figure 5.3-10.
Effect of Gas-to-Cloth Ratio on Fabric Filter TCI (Reference 17).
Particle Characteristics: Particle size and adhesiveness are particle characteristics that
will influence fabric filter design and costs. The G/C ratio should be decreased for small
particles and increased for large particles.1 The adhesive properties of the dust will affect the
fabric and cleaning mechanism selection. Higher intensity cleaning mechanisms, like pulse-
jet, work best with sticky particles, as well as fabrics with coatings such as teflon or other
lubricants.
Gas Stream Characteristics: The two primary stream characteristics that influence
fabric filter design and capital costs are the temperature and chemical properties of the gas
stream. Both characteristics can have a major impact on the fabric selection, since the
available fabrics have widely varying resistances to heat and chemical degradation (see
Table 5.3-2 above). In addition, gas stream properties can affect the construction of the fabric
filter. High temperature streams require insulation of the fabric filters. Streams with highly
corrosive components will need a fabric filter constructed of corrosion-resistant stainless steel.
Insulation and corrosion-resistant materials can be very expensive additions to the cost of a
fabric filter, as shown hi Figure 5.3-11; the use of stainless steel, however, has a greater cost
impact on fabric filters than insulation.17
Fabric Type: Fabric type is usually selected to a great extent by the type of fabric
filter cleaning method, dust type, and the characteristics of the particles and the gas stream.
While these factors may limit the choices, there are usually at least two fabrics that can
perform satisfactorily in a given situation. There is a wide range of prices among the typical
fabrics, but it is not recommended that fabrics be chosen based on cost alone. Some higher
5.3-25
-------
°
14.0
12.0
10.0
e
1 8.0
•S 6.0
1
u 4.0
~S
& 2.0
0.0
0.0
Figure 5.3-11.
Insulation and
Stainless Steel
Stainless Steel Only
Insulation Only
.Basic Pulse-Jet
Fabric Filter
(modular G/C = 5)
Note:
Costs are referenced to
fourth quarter 1996 and
include glass filter bags.
0.2 0.4 0.6
Inlet Flowrate (acfm x 106)
0.8
1.0
Effect of Insulation and Stainless Steel on Fabric Filter TCI
(Reference 17).
priced fabrics have longer operating lives, resulting in lower maintenance and replacement and
costs.4 Figures 5.3-12 and 5.3-: 13 show capital cost curves for reverse-air and pulse-jet fabric
filters, respectively, with typical fabric types.17
5.3.6.2 Annual Costs
The total annual cost of a fabric filter consists of both direct and indirect costs. Direct
annual costs are those associated with the operation and maintenance of the fabric filter.
These include labor (operating, supervisory, and maintenance), operating materials,
replacement parts, electricity, compressed air (for pulse-jet), and dust disposal.
Disposal costs for collected dusts that have no reuse value can be high, comprising sometimes
over 50 percent of the annual costs. Indirect annual costs include taxes, insurance,
administrative costs, overhead, and capital recovery costs. All indirect annual costs except
overhead are dependent on the TCI. In most cases, annual costs are difficult to generalize
because they depend on many factors which can vary widely, even among similar fabric
filters. Table 5.3-6 lists the annual cost parameters that impact fabric filter costs, with typical
values provided for each parameter. Table 5.3-7 provides the annual cost factors for fabric
filters. It is difficult to generalize these costs for all fabric filters, since annual costs are very
site-specific.16
Electricity costs, however, are a significant portion of the annual costs for most fabric
filters. The fabric filters fans consume the majority of the electrical power; the cleaning
equipment also requires power. Fan power consumption is directly related to the pressure
5.3-26
-------
o
X
c
>*i^
•^^^
A*"*"*^^
^^
x*4^
Nomex
Polyestpr
Note
Costs are referenced 1
fourth quarter 1996.
0.0
0.2 0.4 0.6
Inlet Flowrate (ac'fm \\06)
0.8
1.0
Figure 5.3-12.
Effect of Fabric Type on TCI - Reverse-Air Fabric Filter, G/C=2.5
(Reference 17).
&
3>
12.0
10.0
8.0
6.0
4.0
3 2.0
0.0
0.0
0.2 0.4 0.6
Inlet Flowrate (acfinx 106)
0.8
. Teflon
. Nomex
Fiberglass
Polypropylene
Note:
Costs are referenced
to fourth quarter
1996.
1.0
Figure 5.3-13.
Effect of Fabric Type on TCI - Pulse-Jet Fabric Filter, G/C=5
(Reference 17).
5.3-27
-------
drop across the fabric filter, which hi turn is directly dependent upon the G/C ratio. As the
G/C ratio increases, so does the pressure drop and resultant electricity costs. As mentioned
above, increasing the G/C ratio will decrease the fabric filters size and capital costs. Fabric
filters are generally designed to operate at a specific pressure drop. The G/C ratio should be
selected to minimize the annual costs while maintaining the design pressure drop. Power
requirements for fans can be calculated by the following relationship:
Fan Power (kW-hr/yr) = 1.81 x
(Eq. 5.3-1)
where Vis the gas flow rate (ACFM), AP is the pressure drop (in. H2O), t is the operating
hours per year, and 1.81 x 10^ is a unit conversion factor. Once the fan power is determined,
it can be multiplied by the cost of electricity (hi $/kW-hr) to determine the electrical costs.16
Figure 5.3-14 shows annual operating cost curves for four different fabric filter types.17
The curves represent annual costs for the same fabric filter designs used in Figure 5.3-9. For
each fabric filter type, identical values for the design parameter listed hi Table 5.3-6 were used
to prepare the curves hi both figures. Although the same trend hi costs is observed for annual
costs as for TCI costs, where reverse-air fabric filter costs are highest and pulse-jet lowest, the
advantage of pulse-jet fabric filters hi terms of annual costs is not as distinct as with capital
costs.
(D
o
o
O
60
n.
O
I
7.0
6.0
5.0
4.0
3.0
2.0
1.0
0.0
0.0
0.4 0.6
Inlet Flowrate (acfinx 10s)
0.8
- Reverse- Air,
&C = 2
- Mechanical Shaker,
G€ = 2
____ Pulse- Jet (modular),
Pulse-Jet (common),
GC = 4.5
Note:
Costs are referenced to
fourth quarter 1996, and
1-0 include Nomex filter bag
replacement.
Figure 5.3-14.
Annual Fabric Filter Operating Costs (Reference 17).
5.3-28
-------
Table 5.3-6. Annual Cost Parameters for Fabric Filters (Reference 17)
Parameter
Description
Typical Values
Direct Cost Parameters
Operating factor (OF)
Operator labor rate (OR)
Operator shift factor (OS)
Maintenance labor rate (MR)
Maintenance shift (MS) factor
Electricity rate (ER)
Compressed air (CA)
Dust disposal (DD)
Bag capital rec. factor (BCRF)
Indirect Cost Parameters
Overhead factor (OV)
Annual interest rate (I)
Operating life (n)
Capital recovery factor (CRF)
Bag life (b)
Taxes (TAX)
Insurance (INS)
Administrative costs (AC)
Hours of fabric filter operation per year
Operator labor pay rate
Fraction of operator shift on fabric filter
Maintenance labor pay rate
Fraction of maintenance shift on f.f.
Cost of electricity
Cost of compressed air
Cost of disposing of dust
Function of (b) and (i)
Fraction of total labor costs
Opportunity cost of the capital
Expected operating life of fabric filter
Function of (n) and (i)
Expected operating life of filter bags
Fraction of the TCId
Fraction of the TCId
Fraction of the TCId
8,640 hr/yr
$12.50/hra
0.25"
l.lxOR"
0.125"
$0.07/kW-hra
$0.1871000 scf
$20-30/tona
0.5531°
0.60b
7 percent1"
20 yearsb
0.0944C
2 years"
0.01"
0.01b
0.02b
Estimated for 1996 from currently available information.
Estimates from "CO$T-AIR" Control Cost Spreadsheets (Reference 17).
Capital Recovery Factor is calculated from the following formula:
CRF = {i(l + 0«} - {(1 + On -1},
where / = interest rate (fraction) and n (or b) = operating life (years).
The total capital investment (TCI) can be escalated to current values by using the Vatavuk Air Pollution Control Cost Indicies
(VAPCCI), described in Appendix B.
-------
Table 5.3-7. Annual Cost Factors for Fabric Filters (Reference 17).
Cost Item
Formula*
Factor
Direct Costs
Labor
Operator (OL)
Maintenance (ML)
Electricity (E)
Compressed air (C)
Dust disposal (D)
Bag Capital Recovery
Total Direct Cost (DC)
Indirect Costs
Overhead
Capital Recovery
Taxes
Insurance
Administrative Costs
Total Indirect Cost (1C)
Total Annual Cost (DC + 1C)
(OF) x (OR) x (OS)
(OF) X (MR) X (MS)
Power" x (ER)
(CA) x scf per year
(DD) x tons per year
(BCRF)X(BAG)C
(OV)x(OL+ML)
(CRF)X(TCI)
(TAX) x (TCI)
(INS) x (TCI)
(AC) X (TCI)
A
0.55 A
E
C
D
0.5531 BAG
1.55 A + E + C + D+ 0.5531 BAG
0.93 A
0.0944 TCI
0.01 TCI
0.01 TCI
0.02 TCI
0.93 A + 0.1344 TCI
2.48 A + 0.1344 TCI + 0.5531 BAG + E + C + D
Includes values also described in Table 5.3-6.
Equal to total power requirements, e.g. fan, shaker, etc.
BAG = the capital cost of the filter bags only.
5.3.7 Energy and Other Secondary Environmental Impacts
The vast majority of energy demands for fabric filters are for fan operation. Other
minor energy requirements are for cleaning mechanism operation and air compression. The
fan power requirements can be calculated from the above mentioned formula. Energy
requirements for cleaning mechanisms are very site specific.16
The major secondary environmental impact of fabric filters is the generation of solid
waste. Fabric filters collect large amounts of particulate matter, which must be disposed of hi
many cases. The characteristics of the waste are ultimately dependent on the specific
installation. In most applications fabric filters collect dust which is nontoxic and suitable for
landfilling, but some dusts are valuable and can be recycled or sold. In some applications,
fabric filters may collect dusts which are toxic or hazardous. Such dusts will require special
handling and treatment prior to disposal.2
5.3-30
-------
5.3.8 References for Section 5.3
1. Cooper, C.D and F.C. Alley. Air Pollution Control: A Design Approach. 2nd ed.
Waveland Press, Prospect Heights, Illinois. 1994.
2. The Fabric Filter Manual (Revised). The Mcllvaine Company, Northbrook, Illinois.
March 1996.
3. Control Techniques for Particulate Emissions from Stationary Sources - Volume 1
(EPA-450/3-81-005a, NTIS PB83-127498). U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards. Research Triangle Park, North
Carolina. September 1982.
4. McKenna, J.D. and J.H. Turner. Fabric Filter-Baghouses I: Theory, Design, and
Selection (A Reference Text). ETS Inc., Roanoke, Virginia. 1993.
5. Jensen, R.M. Give Reverse-air Fabric Filters a Closer Look. Power, 139:2.
February 1995.
6. Schlotens, M.J. Air Pollution Control: A Comprehensive Look. Pollution
Engineering. May 1991.
7. Pontius, D.H. Characterization of Sonic Devices Used for Cleaning Fabric Filters. J.
Air Pollution Control Association. 35:1301. December 1985.
8. Belba, V.H., W.T. Grubb, and R. Chang. The Potential of Pulse-Jet Baghouses for
Utility Boilers. Part 1: A Worldwide Survey of Users. Journal of the Air and Waste
Management Association. 42:2. February 1992.
9. Carr, R.C. Pulse-Jet Fabric Filters Vie for Utility Service. Power. December 1988.
10. Carr, R.C. and W.B. Smith. Fabric Filter Technology for Utility Coal-Fired Power
Plants, Part V: Development and Evaluation of Bag Cleaning Methods in Utility
Baghouses. J. Air Poll. Control Assoc. 34(5):584. May 1984.
11. Oglesby, S., Jr. Future Directions of Particulate Control Technology: A Perspective.
J. Air Waste Management Assoc. 40(8): 1184-1185. August 1990.
12. Grafe, T. and K. Gregg. Baghouse and Cartridge Dust Collectors: A Comparison.
American Ceramic Society Bulletin. 72:9. September 1993.
13. Parkinson, G. A Hot and Duty Future For Baghouses. Chemical Engineering. April
1989.
5.3-31
-------
14. Croom, M.L. New Developments in Filter Dust Collection. Chemical Engineering.
February 1996.
15. Compilation of Air Pollutant Emission Factors (AP-42). Volume I (Fifth Edition).
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina.
January 1995.
16. OAQPS Control Cost Manual (Fourth Edition, EPA 450/3-90-006) U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards.
Research Triangle Park, North Carolina. January 1990.
17. Vatavuk, W.M. "CO$T-AIR" Control Cost Spreadsheets. Provided by the Innovative
Strategies and Economics Group of the Office of Air Quality Planning and Standards,
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina.
February 1996.
5.3-32
-------
5.4 WET SCRUBBERS
Wet scrubbers are PM control devices that rely on direct and irreversible contact of a
liquid (droplets, foam, or bubbles) with the PM. The liquid with the collected PM is then
easily collected. Scrubbers can be very specialized and designed in many different
configurations. Wet scrubbers are generally classified by the method that is used to induce
contact between the liquid and the PM, e.g. spray, packed-bed, plate. Scrubbers are also often
described as low-, medium-, or high-energy, where energy is often expressed as the pressure
drop across the scrubber. This section addresses the basic operating principles, designs,
collection efficiency, applicability, and costs of wet scrubbers.
Wet scrubbers have important advantages when compared to other PM collection
devices. They can collect flammable and explosive dusts safely, absorb gaseous pollutants,
and collect mists. Scrubbers can also cool hot gas streams. There are also some
disadvantages associated with wet scrubbers. For example, scrubbers have the potential for
corrosion and freezing. Additionally, the use of wet scrubbers can lead to water and solid
waste pollution problems.1 These disadvantages can be minimized or avoided with good
scrubber design.
5.4.1 Particle Collection and Penetration Mechanisms
The dominant means of PM capture in most industrial wet scrubbers is inertial
impaction of the PM onto liquid droplets. Brownian diffusion also leads to particle collection,
but its effects are only significant for particles approximately 0.1 micrometer (/mi) in diameter
or less.2 Direct interception is another scrubber collection mechanism. Less important
scrubber collection mechanisms utilize gravitation, electrostatics, and condensation.2
Inertial impaction in wet scrubbers occurs as a result of a change in velocity between
PM suspended in a gas, and the gas itself. As the gas approaches an obstacle, such as a liquid
droplet, the gas changes direction and flows around the droplet. The particles in the gas will
also accelerate and attempt to change direction to pass around the droplet. Inertial forces will
attempt to maintain the forward motion of the particle towards the object, but the fluid force
will attempt to drag the particle around the droplet with the gas. The resultant particle motion
is a combination of these forces of fluid drag and inertia. This results hi impaction for the
particles where inertia dominates, and by-pass for those particles overwhelmed by fluid drag.2
Large particles, particles i.e. greater than 10 ^m are more easily collected by inertial
impaction because these particles have more inertial momentum to resist changes in the flow of
the gas and, therefore, impact the droplet. Small particles (i.e. particles < 1 /^m) are more
difficult to collect by inertial impaction because they remain in the flow lines of the gas due to
the predominance of the fluid drag force.
Collection by diffusion occurs as a result of both fluid motion and the Brownian
(random) motion of particles. This particle motion in the scrubber chamber results hi direct
5.4-1
-------
particle-liquid contact. Since this contact is irreversible, collection of the PM by the liquid
occurs. Diffusional collection effects are most significant for particles less than 0.1 /um hi
diameter.2 Direct interception occurs when the path of a particle comes within one radius of
the collection medium, which in a scrubber is a liquid droplet. The path can be the result of
inertia, diffusion, or fluid motion.2
Gravitational collection as a result of falling droplets colliding with particles is closely
related to impaction and interception, and is a minor mechanism hi some scrubbers.2
Gravitational settling of particles is usually not a factor because of high gas velocities and
short residence tunes.3 Generally, electrostatic attraction is not an important mechanism
except hi cases where the particles, liquid, or both, are being deliberately charged, or where
the scrubber follows an electrostatic precipitator.3 Some scrubbers are designed to enhance
particle capture through condensation. In such cases, the dust-laden stream is supersaturated
with liquid (usually water). The particles then act as condensation nuclei, growing hi size as
more liquid condenses around them and becoming easier to collect by inertial impaction.2'4
The collection mechanisms of wet scrubbers are highly dependent on particle size.
Inertial impaction is the major collection mechanism for particles greater than approximately
0.1 /mi hi diameter. The effectiveness of inertial impaction increases with increasing particle
size. Diffusion is generally effective only for particles less than 0.1 /mi in diameter, with
collection efficiency increasing with decreasing particle size. The combination of these two
major scrubber collection mechanisms contributes to a minimum collection efficiency for PM
approximately 0.1 /mi hi diameter.5 The exact minunum efficiency for a specific scrubber will
depend on the type of scrubber, operating conditions, and the particle size distribution hi the
gas stream. Scrubber collection efficiency is discussed hi more detail hi Section 5.4.3.
5.4.2 Types of Wet Scrubbers
There are a great variety of wet scrubbers that are either commercially available or can
be custom designed. While all wet scrubbers are similar to some extent, there are several
distinct methods of using the scrubbing liquid to achieve particle collection. Wet scrubbers are
usually classified according to the method that is used to contact the gas and the liquid.
The most common scrubber design is the introduction of liquid droplets into a spray
chamber, where the liquid is mixed with the gas stream to promote contact with the PM. In a
packed-bed scrubber, layers of liquid are used to coat various shapes of packing material that
become impaction surfaces for the particle-laden gas. Scrubber collection can also be achieved
by forcing the gas at high velocities though a liquid to form jet streams. Liquids are also used
to supersaturate the gas stream, leading to particle scrubbing by condensation.
5.4-2
-------
5.4.2.1 Spray Chambers
Spray chambers are very simple, low-energy wet scrubbers. In these scrubbers, the
particulate-laden gas stream is introduced into a chamber where it comes into contact with
liquid droplets generated by spray nozzles. These scrubbers are also known as pre-formed
spray scrubbers, since the liquid is formed into droplets prior to contact with the gas stream.
The size of the droplets generated by the spray nozzles is controlled to maximize liquid-
particle contact and, consequently, scrubber collection efficiency.
The common types of spray chambers are spray towers and cyclonic chambers. Spray
towers are cylindrical or rectangular chambers that can be installed vertically or horizontally.
In vertical spray towers, the gas stream flows up through the chamber and encounters several
sets of spray nozzles producing liquid droplets. A de-mister at the top of the spray tower
removes liquid droplets and wetted PM from the exiting gas stream. Scrubbing liquid and
wetted PM also dram from the bottom of the tower in the form of a slurry. Horizontal spray
chambers operate in the same manner, except for the fact that the gas flows horizontally
through the device. A typical spray tower is shown in Figure 5.4-1.1>2-5
A cyclonic spray chamber is similar to a spray tower with one major difference. The
gas stream is introduced to produce cyclonic motion inside the chamber. This motion
contributes to higher gas velocities, more effective particle and droplet separation, and higher
collection efficiency.1 Tangential inlet or turning vanes are common means of inducing
cyclonic motion.5 Figure 5.4-2 provides an example of a cyclonic spray chamber.
5.4.2.2 Packed-Bed Scrubbers
Packed-bed scrubbers consist of a chamber containing layers of variously-shaped
packing material, such as raschig rings, spiral rings, and berl saddles, that provide a large
surface area for liquid-particle contact. These and other types of packings are illustrated in
Figure 5.4-S.2-5 The packing is held in place by wire mesh retainers and supported by a plate
near the bottom of the scrubber. Scrubbing liquid is evenly introduced above the packing and
flows down through the bed. The liquid coats the packing and establishes a thin film. In
vertical designs, the gas stream flows up the chamber (countercurrent to the liquid). Some
packed beds are designed horizontally for gas flow across the packing (crosscurrent).
In packed-bed scrubbers, the gas stream is forced to follow a circuitous path through
the packing,, on which much of the PM impacts. The liquid on the packing collects the PM
and flows down the chamber towards the dram at the bottom of the tower. A mist eliminator
(also called a "de-mister") is typically positioned above/after the packing and scrubbing liquid
supply. Any scrubbing liquid and wetted PM entrained hi the exiting gas stream will be
removed by the mist eliminator and returned to drain through the packed bed. A typical
packed-bed scrubber is illustrated in Figure 5.4-4.2-5
5.4-3
-------
DEMISTER
GAS IN
GAS OUT
\/
A
A A
A A
\/ \S \/
« DEMISTER WASH
*« DEMISTER WASH
*+
S*
INLET SLURRY
EFFLUENT SLURRY
EFFLUENT SLURRY
Figure 5.4-1. Schematic Diagram of a Spray Tower Scrubber (Reference 2).
5.4-4
-------
Dirty Gas :$&*j
Clean Gas
Clean Liquid
Dirty Liquid
Figure 5.4-2. Schematic Diagram of a Cyclonic Spray Chamber Scrubber (Reference 1).
5.4-5
-------
RASCHIC RING
USSIN6 RING
CROSS-PARTITION RING SINOU SPIRAL RINO
DOUftLE SPIRAL RING TRIPLE SPIRAL RINO
URL SADDLE
INTALOX SADDLI
Figure 5.4-3. Typical Packing Materials for Packed Bed Scrubbers (Reference 2).
5.4-6
-------
Discharge
Liquid Inlet
Liquid
Spray Distributor
Gas
Inlet
Overflow
- ,r
Suction
j Drain
Figure 5.4-4. Schematic Diagram of a Packed Tower Scrubber (Reference 2).
5.4-7
-------
In a packed-bed scrubber, high PM concentrations can clog the bed, hence, the
limitation of these devices to streams with relatively low dust loadings.5 Plugging is a serious
problem for packed-bed scrubbers because the packing is more difficult to access and clean
than other scrubber designs.2 Mobile-bed scrubbers are available that are packed with low-
density plastic spheres that are free to move within the packed bed.5 These scrubbers are less
susceptible to plugging because of the increased movement of the packing material. In general,
packed-bed scrubbers are more suitable for gas scrubbing than paniculate scrubbing because of
the high maintenance requirements for control of PM.1-2
5.4.2.3 Impingement Plate Scrubbers
An impingement plate scrubber is a vertical chamber with plates mounted horizontally
inside a hollow shell. Impingement plate scrubbers operate as countercurrent PM collection
devices. The scrubbing liquid flows down the tower while the gas stream flows upward.
Contact between the liquid and the particle-laden gas occurs on the plates. The plates are
equipped with openings that allow the gas to pass through. Some plates are perforated or
slotted, while more complex plates have valve-like openings. Figure 5.4-5 shows common
plate designs used in impingement plate scrubbers.2-5
The simplest impingement plate is the sieve plate, which has round perforations. In
this type of scrubber, the scrubbing liquid flows over the plates and the gas flows up through
the holes. The gas velocity prevents the liquid from flowing down through the perforations.
Gas-liquid-particle contact is achieved within the froth generated by the gas passing through
the liquid layer. Complex plates, such as bubble cap or baffle plates, introduce an additional
means of collecting PM. The bubble caps and baffles placed above the plate perforations force
the gas to turn before escaping the layer of liquid. While the gas turns to avoid the obstacles,
most PM cannot and is collected by impaction on the caps or baffles. Bubble caps and the like
also prevent liquid from flowing down the perforations if the gas flow is reduced.
In all types of impingement plate scrubbers, the scrubbing liquid flows across each
plate and down the inside of the tower onto the plate below. After the bottom plate, the liquid
and collected PM flow out of the bottom of the tower. A typical impingement plate scrubber
is shown in Figure 5.4-6.2-5 Impingement plate scrubbers are usually designed to provide
operator access to each tray, making them relatively easy to clean and maintain.2
Consequently, impingement plate scrubbers are more suitable for PM collection than packed-
bed scrubbers. Particles greater than 1 jim in diameter can be collected effectively by
impingement plate scrubbers, but many particles < 1 //m will penetrate these devices.5
5.4.2.4 Mechanically-aided Scrubbers
Mechanically-aided scrubbers (MAS) employ a motor driven fan or impeller to enhance
gas-liquid contact. Generally in MAS, the scrubbing liquid is sprayed onto the fan or impeller
blades. Fans and impellers are capable of producing very fine liquid droplets with high
5.4-8
-------
Sieve
Bubble Cap
Baffle
Top View O O O O O
Cross-sectional
View
Baffles
Gas
Gas
Figure 5.4-5. Common Plate Designs for Impingement Plate Scrubbers (adapted from Reference 2).
-------
Clean Gas
Scrubbing
Liquid
Dust Laden
Gas
Perforated
Plates
Scrubbing Liquid
Figure 5.4-6. Schematic Diagram of a Plate Tower Scrubber (adapted from Reference 2).
5.4-10
-------
velocities. These droplets are effective in contacting fine PM. Once PM has impacted on the
droplets, it is normally removed by cyclonic motion. Mechanically aided scrubbers are
capable of high collection efficiencies, but only with a commensurate high energy
consumption. An example of a mechanically aided scrubber is provided in Figure 5.4-7.1-2-5
Because many moving parts are exposed to the gas and scrubbing liquid in a MAS,
these scrubbers have high maintenance requirements. Mechanical parts are susceptible to
corrosion, PM buildup, and wear. Consequently, mechanical scrubbers have limited
applications for PM control.2-5
5.4.2.5 Venturi Scrubbers
A venturi, or gas-atomized spray, scrubber accelerates the gas stream to atomize the
scrubbing liquid and to improve gas-liquid contact. In a venturi scrubber, a "throat" section is
built into the duct that forces the gas stream to accelerate as the duct narrows and then
expands. As the gas enters the venturi throat, both gas velocity and turbulence increase. The
scrubbing liquid is sprayed into the gas stream before the gas encounters the venturi throat.
The scrubbing liquid is then atomized into small droplets by the turbulence in the throat and
droplet-particle interaction is increased. After the throat section in a venturi scrubber, the
wetted PM and excess liquid droplets are separated from the gas stream by cyclonic motion
and/or a mist eliminator. Venturi scrubbers have the advantage of being simple hi design,
easy to install, and with low-maintenance requirements.1 An example of a venturi scrubber is
provided in Figure 5.4-8.
The performance of a venturi scrubber is dependent to some extent on the velocity of
the gas through the throat. Several venturi scrubbers have been designed to allow velocity
control by varying the width of the venturi throat.2-5 Because of the high interaction between
the PM and droplets, venturi scrubbers are capable of high collection efficiencies for small
PM. Unfortunately, increasing the venturi scrubber efficiency requires increasing the pressure
drop which, in turn, increases the energy consumption.1
5.4.2.6 Orifice Scrubbers
Orifice scrubbers, also known as entrainment or self-induced spray scrubbers, force the
particle-laden gas stream to pass over the surface of a pool of scrubbing liquid as it enters an
orifice. With the high gas velocities typical of this type of scrubber, the liquid from the pool
becomes entrained in the gas stream as droplets. As the gas velocity and turbulence increases
with the passing of the gas through the narrow orifice, the interaction between the PM and
liquid droplets also increases. Paniculate matter and droplets are then removed from the gas
stream by impingement on a series of baffles that the gas encounters after the orifice. The
collected liquid and PM drain from the baffles back into the liquid pool below the orifice.2-5
Orifice scrubbers can effectively collect particles larger than 2 ^m in diameter.1-5 Some orifice
5.4-11
-------
Flag WPe
motor
mount
Mam
shaft
Flexible
coupling
Air inlet
Pump
housing
Drag
conveyor
Air outlet
Water distributor
Figure 5.4-7. Diagram
of a Mechanically-aided Scrubber (Reference 1).
5.4-12
-------
Clean gas out
o
Mist eliminator
Dirty gas in
Liquid in -•--;
Throat
Separator
Elbow
crossover
Liquid to settling
and rectrculation
Figure 5.4-8. Schematic Diagram of a Venturi Scrubber with Cyclonic Separation
(Reference 1).
5.4-13
-------
scrubbers are designed with adjustable orifices to control the velocity of the gas stream. A
typical orifice scrubber is shown in Figure 5.4-9.
Orifice scrubbers usually have low liquid demands, since they use the same scrubbing
liquid for extended periods of time.1 Because orifice scrubbers are relatively simple in design
and usually have few moving parts, the major maintenance concern is the removal of the
sludge which collects at the bottom of the scrubber. Orifice scrubbers rarely drain continually
from the bottom because a static pool of scrubbing liquid is needed at all times. Therefore,
the sludge is usually removed with a sludge ejector that operates like a conveyor belt. As the
sludge settles to the bottom of the scrubber, it lands on the ejector and is conveyed up and out
of the scrubber. Figure 5.4-10 shows a typical sludge ejector.2
5.4.2.7 Condensation Scrubbers
Condensation scrubbing is a relatively recent development hi wet scrubber technology.
Most conventional scrubbers rely on the mechanisms of impaction and diffusion to achieve
contact between the PM and liquid droplets. In a condensation scrubber, the PM act as
condensation nuclei for the formation of droplets. Generally, condensation scrubbing depends
on first establishing saturation conditions hi the gas stream. Once saturation is achieved,
steam is injected into the gas stream. The steam creates a condition of supersaturation and
leads to condensation of water on the fine PM in the gas stream. The large condensed droplets
can be removed by several conventional devices. Typically, a high efficiency mist eliminator
is also used.2'4
A high-efficiency condensation "growth" PM scrubber has been developed that is
suitable for both new and retrofit installations, and is designed specifically to capture fine PM
that escapes primary PM control devices. This type of scrubber utilizes a multistage process,
including pretreatment and growth chambers, that provide an environment that encourages the
fine PM to coagulate and form larger particles. A schematic diagram of this scrubber is
provided in Figure 5.4-11.4
5.4.2.8 Charged Scrubbers
Charged, or electrically-augmented, wet scrubbers utilize electrostatic effects to
improve collection efficiencies for fine PM with wet scrubbing. Since conventional wet
scrubbers rely on the inertial impaction between PM and liquid droplets for PM collection,
they are generally ineffective for particles with diameters less than 1 /um. Pre-charging of the
PM in the gas stream can significantly increase scrubber collection efficiency for these
submicrometer particles. When both the-particles and droplets are charged, collection
efficiencies for submicrometer particles are highest, approaching that of an ESP.2
5.4-14
-------
Clean gas
Foam
Orifice
Figure 5.4-9. Diagram of an Orifice Scrubber (Reference 1).
5.4-15
-------
Conveyor
Sludge Ejector
Figure 5.4-10. Diagram of a Sludge Ejector in an Orifice Scrubber (Reference 2).
5.4-16
-------
Water
Injection
Steam
. Injection
Dust
Laden
Gas
Stream
Particle
Coagulation
Chamber
Gas Stream
Conditioner
^•;y%/&|?pi"-:
•ih -•.!- .iffi&i* '
*«?•• S4.fv:-itr'|>;;i''
Clean Gas
Particle
Growth
Chamber
Water
Conventional
Particle
Removal
Figure 5.4-11.
Schematic Diagram of a Condensation "Growth" Scrubber (adapted from Reference 4).
-------
There are several types of charged wet scrubbers. Particulate matter can be charged
negatively or positively, with the droplets given the opposite charge. The droplets may also
be bipolar (a mixture of positive and negative). In this case, the PM can be either bipolar or
unipolar. Figure 5.4-12 is a schematic of a charged wet scrubber.2
5.4.2.9 Fiber-Bed Scrubbers
In a fiber-bed scrubbers, the moisture-laden gas stream passes through mats of packing
fibers, such as spun glass, fiberglass, and steel. The fiber mats are often also spray wetted
with the scrubbing liquid. Depending on the scrubber requirements, there may be several
fiber mats and an impingement device for PM removal included in the design. The final fiber
mat is typically dry for the removal of any droplets that are still entrained hi the stream.
Fiber-bed scrubbers are best suited for the collection of soluble PM, i.e. PM that dissolves hi
the scrubber liquid, since large amounts of insoluble PM will clog the fiber mats with tune.
For this reason, fiber-bed scrubbers are more often used as mist eliminators, i.e., for the
collection of liquids, rather than for PM control.2
5.4.3 Collection Efficiency
Collection efficiencies for wet scrubbers are highly variable. Most conventional
scrubbers can achieve high collection efficiencies for particles greater than 1.0 /mi hi diameter,
however they are generally ineffective collection devices for submicrometer (< 1 //m)
particles. Some unconventional scrubbers, such as condensation and charged, are capable of
high collection efficiencies, even for submicrometer particles. Collection efficiencies for
conventional scrubbers depend on operating factors such as particle size distribution, inlet dust
loading, and energy input. Figure 5.4-13 provides scrubber efficiency curves for coal and oil
combustion, wood combustion, and coke production. Table 5.4-1 presents the PM-10 and
PM-2.5 collection efficiencies.6
Conventional scrubbers rely almost exclusively on inertial impaction for PM collection.
As discussed above, scrubber efficiency that relies on inertial impaction collection mechanisms
will increase as particle size increases. Therefore, collection efficiency for small particles
(< 1 ^m) are expected to be low for these scrubbers. The efficiency of scrubbers that rely on
inertial impaction can be unproved, however, by increasing the relative velocity between the
PM and the liquid droplets. Increasing velocity will result hi more momentum for all PM,
enabling smaller particles to be collected by impaction. This can be accomplished hi most
scrubbers by increasing the gas stream velocity. Unfortunately, increasing the gas velocity
will also increase the pressure drop, energy demand, and operating costs for the scrubber.1>2>s
5.4-18
-------
U)
VO
Dust
Laden
Gas
Stream
Scrubbing
Liquid
Particle
Charging
Droplet Production
and Charging
Droplet -
Particle
Interaction
Clean Gas
Liquid
Droplet
Removal
Figure 5.4-12. Schematic Diagram of Charged Wet Scrubber (adapted from Reference 2).
-------
COAL AND OIL COMBUSTION
WOOD AND BARK COMBUSTION
100
r 90
70
60
50
O.I
1 10
Panicle Size (urn)
—•— Residual Oil,
DataQality C,D
- •* • • Dry Bottom Boiler,
Bituminous Coal,
Data Quality C.D
100
95
90
85
-Wood and Bark
Combustion,
Data Quality E
0.1
1 10
Panicle Size dim)
100
f'
s
COKE PRODUCTION
(00
95
go
75
0.1
1 10
Panicle Size (|im)
100
Coke Pushing,
Mobile Scrubber,
Data Quality: D
• Coal Preheating,
Ventun
Scrubber, Data
Quality. D
Note: Data quality refers to the data quality ratings assigned to
the emission factors from which the efficiencies were
calculated (Reference 6), as follows:
A = excellent
B = above average
C = average
D = below average
E = poor.
Figure 5.4-13.
Cumulative Collection Efficiency Data for PM Wet Scrubbers at Coal, Oil, Wood, and Bark
Combustion Sources, and Coke Production Operations (Reference 6)
-------
Table 5.4-1. PM-10 and PM-2.5 Cumulative Collection Efficiencies
for Wet Scrubbers at Coal, Oil, Wood, and Bark Combustors;
and Coke Production Units (Reference 6).
Collection Efficiency (percent)
Application PM-10 PM-2.5
Combustion Sources
Bituminous coal (dry bottom) 81.7 50.0
Residual oil 91.5 88.8
Wood and bark 93.3 92.1
Bark only 85.1 83.8
Coke Production
Coal preheating (venturi scrubber) 92.9 89.0
Coke pushing (mobile-bed scrubber) 95.2 89.0
Another factor which contributes to low scrubber efficiency for small particles is short
residence times. Typically, a particle is hi the contact zone of a scrubber for only a few
seconds. This is sufficient time to collect large particles that are affected by impaction
mechanisms. However, since submicrometer particles are most effectively collected by
diffusion mechanisms that depend on the random motion of the particles, sufficient tune hi the
contact zone is needed for this mechanism to be effective. Consequently, increasing the gas
residence time should also increase the particle/liquid contact tune and the collection efficiency
for small particles.2
An important relationship between inlet dust concentration (loading) and collection
efficiency for fine PM in scrubbers has been recently found.7 Collection efficiency for
scrubbers has been found to be directly proportional to the inlet dust concentration. That is,
efficiency will increase with increasing dust loading. This suggests that scrubber removal
efficiency is not constant for a given scrubber design unless it is referenced to a specific inlet
dust loading. In contrast, it has been shown that scrubber outlet dust concentration is a
constant, independent of inlet concentration.7
5.4-21
-------
5.4.4 Applicability
Wet scrubbers have numerous industrial applications and few limitations. They are
capable of collecting basically any type of dust, including flammable, explosive, moist, or
sticky dusts. In addition, they can collect suspended liquids (i.e. mists) or gases alone or with
PM simultaneously.1 However, while scrubbers have many potential applications, there are
some characteristics that limit their use. The most significant consideration is the relatively
low collection efficiency for fine PM, especially those less than 1.0 ju,m in diameter.
Therefore, conventional scrubbers may not be suitable for processes which emit many
submicrometer particles. As discussed above venturi, condensation, and charged scrubbers are
capable of collecting submicrometer particles at higher efficiencies than other scrubbers and,
therefore, can be used effectively in applications where there are a large percentage of fine PM
in the gas stream.2
Gas stream composition may also be a limiting factor in scrubber application for a
specific industry, since wet scrubbers are very susceptible to corrosion.1 The use of wet
scrubbers also may not be desirable when collecting valuable dust which can be recycled or
sold. Since scrubbers discharge collected dust hi the form of a wet slurry, reclaiming clean
dry dust from this slurry is often inconvenient and expensive.1 Because of design constraints,
paniculate scrubbers are generally not used in very large installations, such as utilities where
gas flowrates exceed 250,000 ACFM, since multiple scrubbers are needed once flowrates
exceed 60,000-75,000 ACFM.
Table 5.4-2 lists current applications of wet scrubbers.li2'8 It should be noted that the
level of PM control supplied by each of the scrubber types listed hi Table 5.4-2 will vary
according to the level of control currently required by each industry and/or facility. The
driving force for PM control in many industries and/or facilities is the Federal, State, and
local air pollution regulations. As more stringent PM regulations are put into place, a shift
toward the use of higher efficiency scrubbers is likely to occur. Table 5.4-3 rates the various
scrubber types according to then* potential for controlling fine particles.
Table 5.4-2. Current Industrial Applications
of Wet Scrubbers (References 1,2, and 8)
Application
Utility Boilers
(Coal, Oil)
Industrial Boilers
(Coal, Oil, Wood, Liquid Waste)
Commercial/Institutional Boilers
(Coal, Oil, Wood)
Source Category Code
1-01-002... 004
1-02-001. ..005,
-009, -Oil, -013
1-03-001... 005
1-03-009
Typical Scrubber Type
Venturi
Venturi, impingement plate (baffle)
Venturi
5.4-22
-------
Table 5.4-2, continued
Application
Source Category Code
Typical Scrubber Type
Chemical Manufacture
Non-Ferrous Metals Processing
(Primary and Secondary)
Copper
Lead
Aluminum
Other
Ferrous Metals Processing
Coke Production
Ferroalloy Production
Iron and Steel Production
Gray Iron Foundries
Steel Foundries
Asphalt Manufacture
Mineral Products
Coal Cleaning
Other
Wood, Pulp, and Paper
Food and Agriculture
Incineration
3-01-001...999
3-03-005
3-04-002
3-03-010
3-04-004
3-03-000...002
3-04-001
3-03-011...014
3-04-005...006
3-04-010...022
3-03-003...004
3-03-006...007
3-03-008...009
3-04-003
3-04-007, -009
3-05-001...002
3-05-010
3-05-003...999
3-07-001
3-02-001...999
5-01-001,
5-02-001, -005
5-03-001, -005
Packed-bed, venturi, fiber-bed
Spray chamber
Venturi, (cyclonic) spray chamber,
fiber-bed, charged
Spray chamber, packed-bed, venturi, charged
(Cyclonic) spray chamber
Charged, venturi, packed-bed (mobile)
Packed-bed, fiber-bed
Venturi
Venturi, impingement plate (baffle)
Venturi
Venturi
Venturi, fiber-bed
Venturi
Venturi, (cyclonic) spray chamber
Impingement, fiber-bed, packed-bed
Venturi, packed-bed, condensation
5.4-23
-------
Table 5.4-3. PM10/PM25 Control Potential
for Various Scrubber Designs
Scrubber Type
PM10/PM2.5
Control
Potential
Comments
Spray Chamber
Packed-Bed
Impingement Plate
Mechanically-aided
Venturi
Orifice
Condensation
Charged
Fiber-Bed
Fair Cyclonic are better than
conventional spray
Poor Useful for low dust loadings only
Good Not as good for PM < 1 /urn
Good High energy consumption to
achieve PM10/PM2 5 control
Good High energy consumption to
achieve PM10/PM2 5 control
Good Not as good for PM < 2 fj.m
Good Excellent control possible with
condensation "growth" scrubbers
Excellent Electric power costs add to overall
scrubber costs
Fair Useful for soluble PM only
5.4-24
-------
5.4.5 Costs of PM Wet Scrubbers
The costs of installing and operating a scrubber include both capital and annual costs.
Capital costs are all of the initial costs related to scrubber equipment and installation. Annual
costs are the direct yearly costs of operating the scrubber, plus indkect costs such as overhead,
capital recovery, taxes, insurance, and administrative charges. The following sections discuss
capital and annual costs for scrubbers, referenced to the third quarter of 1995 unless otherwise
noted.
5.4.5.1 Capital Costs
The total capital investment (TCI) for scrubbers includes all of the initial capital costs,
both dkect and indirect. Direct capital costs are the purchased equipment costs (PEC), and the
costs of installation (foundations, electrical, piping, etc.). Indirect costs are related to the
installation and include engineering, construction, contractors, start-up, testing, and
contingencies. The PEC is calculated based on the scrubber specifications. The direct and
indirect installation costs are calculated as factors of the PEC. Table 5.4-4 provides the TCI
factors for a typical scrubber.9-10
Wet scrubber costs are dependent upon the type of scrubber selected, the required size
of the scrubber, and the materials of construction. Scrubber sizing incorporates several design
parameters, including gas velocity, liquid-to-gas ratio, and pressure drop. Gas velocity is the
primary sizing factor. Increasing the gas velocity will decrease the required size and cost of a
scrubber. However, pressure drop will increase with increasing gas velocity. This will also
result in increased electricity consumption and, therefore, higher operating costs. Determining
the optimum gas velocity involves balancing the capital and annual costs. In most cases,
scrubbers are designed to operate within recommended ranges of gas velocity, liquid-to-gas
ratio, and pressure drop. These ranges are provided in Table 5.4-5.11
Another important scrubber parameter that affects costs is the temperature of the gas
stream at saturation once it has been cooled by the scrubber liquid. This temperature affects
the volumetric flowrate of the outlet gas and, consequently, the size of the scrubber. In
addition, the saturation temperature impacts the scrubbing liquid makeup and the wastewater
flowrate. The saturation temperature is a complex function of essentially three variables: the
temperature of the inlet gas stream, the absolute humidity of the inlet gas stream, and the
absolute humidity at saturation. Typically, the saturation temperature is determined
graphically from a psychometric chart once these three variables are known. For this
document, the sizing and costing of wet scrubbers were aided by the use of the CO$T-AIR
Control Cost Spreadsheets,12 that employ an iterative procedure for estimating the saturation
temperature.
5.4-25
-------
Table 5.4-4. Capital Cost Factors for a Typical Scrubber (Reference 10).
Cost Item Factor
Direct Costs
Purchased equipment costs
Scrubber + auxiliary equipment As estimated (A)
Instrumentation 0.10 A
Sales taxes 0.03 A
Freight 0.05 A
Total Purchased Equipment Cost (PEC) B = 1.18 A
Direct installation costs
Foundations and supports 0.06 B
Handling and erection 0.40 B
Electrical 0.01 B
Piping 0.05 B
Insulation for ductwork 0.03 B
Painting 0.01 B
Total direct installation cost 0.56 B
Site Preparation and Buildings (Site) As required
Total Direct Cost (DC) 1.56 B + Site
Indirect Costs (installation)
Engineering 0.10B
Construction and field expense 0.10 B
Contractor fees 0.10 B
Start-up 0.01 B
Performance test 0.01 B
Model study Model
Contingencies Q.Q3 B
Total Indirect Cost (1C) 0.35 B
Total Capital Investment = DC + 1C 1.91 B + Site + Model
5.4-26
-------
Table 5.4-5. Recommended Gas Velocities, Liquid/Gas Ratios,
and Pressure Drops for Particulate Wet Scrubbers
(Reference 9).
Liquid/Gas Ratio Pressure Drop
Scrubber Type Velocity (ft/sec) (gal/1000 ACFM) (inches H2O)
Venturi
Impingement plate
Spray chamber
Cyclonic spray chamber
90-400"
<14
1Q
105-140"
4-100
2-10
—
7
<100
2-3c
2-4
4-6
Packed tower
Vertical 2-6
Horizontal 4-8
Venturi throat velocity varies with pressure drop, volumetric flowrate, gas density, and
liquid/gas ratio as follows: v, = throat velocity (ft/sec) = C(AP/rg)°5, AP = pressure drop
(niches H2O), rg = gas density (lb/ft3), L/G = liquid/gas ratio (gal/1000 ACFM), C =
l,060exp(-0.0279L/G).
Varies with pressure drop and gas density.
Pressure drop per plate.
Once a scrubber has been properly designed and sized, the costs can generally be
expressed as a function of the inlet or total gas flowrate.9 Cost curves are shown below for the
following types of scrubbers: venturi, impingement plate, and packed tower.
All the estimates for scrubber capital costs have been escalated to third quarter 1995
dollars. However, the capital costs presented in this section can be escalated further to reflect
more current values through the use of the Vatavuk Air Pollution Cost Control Indexes
(VAPCCI), which are updated quarterly, available on the OAQPS Technology Transfer
Network (TTN), and published monthly in Chemical Engineering magazine. The VAPCCI
updates the PEC and, since capital costs are based only on the PEC, capital costs can be easily
adjusted using the VAPCCI. To escalate capital costs from one year (Costold) to another more
recent year (Cost^), a simple proportion can be used, as follows:13
Cost™ = Costold(VAPCCInew/VAPCCIold)
The VAPCCI for wet scrubbers for third quarter 1995 was 114.7.
5.4-27
-------
Venturi Scrubbers: Venturi scrubber costs are based on data for two ranges of gas
flowrates. Cost curves for scrubbers treating less than 19,000 ACFM are provided in
Figure 5.4-14. Cost curves for venturi scrubbers capable of handling greater than
19,000 ACFM but less than 59,000 ACFM are shown in Figure 5.4-15. For total flowrates
greater than 59,000 ACFM, the gas stream should be divided evenly and treated by two or
more identical scrubbers (with inlet flowrates of < 59,000 ACFM) operating in parallel.
The most common construction material for venturi scrubbers is carbon steel. Special
applications may require other materials, such as rubber-lined steel, epoxy-coated steel, fiber-
reinforced plastic (FRP), that will increase the cost of the unit.9 Separate cost curves for
carbon steel and other specialized materials are included in Figures 5.4-14 and 5.4-15.12
Impingement Plate Scrubbers: Impingement plate scrubber costs are dependent on the
number of plates and the total gas flowrate. The costs for impingement scrubbers are based on
data that corresponds to a total gas flowrate between 900 and 77,000 ACFM or above. For
total gas flowrates above 77,000 ACFM, multiple scrubbers are required. Figure 5.4-16
presents cost curves for impingement plate scrubbers with total gas flowrates between 900 and
77,000 ACFM. Cost curves for scrubbers with total flowrates above 77,000 ACFM are
shown in Figure 5.4-1712 and require the use of 2, 3, or 4 identical scrubber units. All the
cost correlations shown here are for sieve plate scrubbers with three plates. Impingement
plate scrubbers are usually constructed with carbon steel. Some applications may require more
expensive materials, such as coated carbon steel, FRP, or polyvinyl chloride (PVC).9
Packed-bed Scrubbers: The costs for packed-bed scrubbers depend on the inlet gas
velocity/column diameter, orientation of the column (vertical vs. horizontal), height of packing
material, and the presence of any auxiliary equipment. Figures 5.4-18 and 5.4-19 present
costs curves for two types of packed-bed scrubbers. Figure 5.4-18 presents a cost
curve for a small vertical column packed-bed scrubber. The costs for this unit vary with the
column diameter, which can range from 1 to 2.5 feet. Gas flowrates range from 200 to 1200
ACFM.9 For Figure 5.4-18, the scrubber is assumed to be constructed of FRP with 6 feet of
polypropylene packing. Costs also include the costs for a spray nozzle, liquid distributor, and
mist eliminator. Figure 5.4-19 provides a cost curve for a large packed-bed scrubber with
horizontal gas flow from 800 to 80,000 ACFM. Costs for this unit are based on the use of
PVC or FRP construction materials and a design that includes a spray section, a 1-foot packed
bed, and a mist eliminator.9 Capital and annual costs are also available from Chapter 9 of the
OAQPS Control Cost Manual (Reference 14).
5.4.5.2 Annual Costs
The total annual cost of a wet scrubber consists of both direct and indirect costs.
Direct annual costs are those associated with the operation and maintenance of the scrubber.
5.4-28
-------
U
o
160
140
120
100
80
60
40
20
. Rubber-Lined Carbon Steel or
Fiber-Reinforced Plastic
, Epoxy Coated Carbon Steel
Carbon Steel
Note-
Costs are referenced to
fourth quarter 1996.
2 4 6 8 10 12 14 16 18 20
Inlet Flowrate (acfin x 103)
Figure 5.4-14.
Venturi Scrubber Capital Costs, Inlet Flowrate < 19,000 ACFM
(Reference 11)
400
300
I 200
e
100
10
Figure 5.4-15.
20 30 40 50
Inlet Flowrate (acfln x 103)
60
70
80
. Rubber-Lined Carbon Steel or
Fiber-Reinforced Plastic
, Epoxy Coated Carbon Steel
Carbon Steel
Note:
Costs are referenced to
fourth quarter 1996.
Venturi Scrubber Capital Costs, Inlet Gas Flowrate > 19,000 ACFM,
<59,000 ACFM (Reference 11).
5.4-29
-------
s
'§•
I
0.35
0.30
025
020
0.15
0.10
0.05
0.00
20,000 40,000 60,000
Inlet Flowrate (acfrn)
80,000
Fiber-Reinforced Plastic
or Polyvmyl Chloride (PVC)
., _ _ » Coated Carbon Steel
100,000
. Carbon Steel
Note:
Costs are referenced to fourth
quarter 1996.
Figure 5.4-16.
Impingement Scrubber Capital Costs, Inlet Gas Flowrate < 77,000
ACFM (Reference 11).
2.0
2 if
nvestmen
0.0
0.0
Figure 5.4-17.
2 units
required
3 units
required
4 units
required
0.1 02 03
Total Flowrate (acfinx 106)
Fiber-Reinforced Plastic or
Polyvinyl Chloride (PVC)
• _ _. Coated Carbon Steel
. Carbon Steel
Note:
Costs are referenced to
fourth quarter 1996.
0.4
Impingement Scrubber Capital Costs, Total Gas Flowrate > 77,000
ACFM (Reference 11).
5.4-30
-------
i
X
^
1
tn
1
~s
'B.
^
u
H
20.0
18.0
16.0
14.0
12.0
10.0
8.0
6.0
4.0
2.0
0.0
^^
^^
^^
^^
•"""^
050
Figure 5.4-18.
Note:
Costs are referenced to
fourth quarter 1996.
1.00
150 2.00
Column Diameter (ft)
250
3.00
Vertical Packed-bed Scrubber Capital Costs (Reference 9).
100
so
g 60
VI
U
>
J3
3 40
8-
u
1 20
Note:
Costs are referenced to
fourth quarter 1996.
0 10 20 30 40 50 60 70 80 90 100
Inlet Flowrate (acfmxlO3)
Figure 5.4-19.
Horizontal Packed-bed Scrubber Capital Costs (Reference 9).
5.4-31
-------
These include labor (operating, supervisory, coordinating, and maintenance), maintenance
materials, operating materials, electricity, sludge disposal, wastewater treatment, and
conditioning agents.12 Heating and cooling may be required in some climates to prevent
freezing or excessive vaporation loss of the scrubbing liquid.2
Indirect annual costs include taxes, insurance, administrative costs, overhead, and
capital recovery. All of these costs except overhead are dependent on the TCI. Table 5.4-6
lists the parameters that impact wet scrubber annual costs with typical values provided for each
parameter. Table 5.4.7 provides the annual cost factors for scrubbers. Annual costs for
scrubbers are difficult to generalize because these costs are very site-specific.
5.4.6 Energy and Other Secondary Environmental Impacts
The secondary environmental impacts of wet scrubber operation are related to energy
consumption, solid waste generation, and water pollution. The energy demands for wet
scrubbers generally consist of the electricity requirements for fan operation, pump operation,
and wastewater treatment. Charged scrubbers have additional energy demands for charging
the water droplets and/or PM. Energy demands for wastewater treatment and charged
scrubbers are very site specific and, therefore, are not estimated here.2
The fan power needed for a scrubber can be estimated by the following equation:14
Fan Power (kW-hr/yr) = 1.81 x 10^(V)(AP)(r) (Eq. 5.4-1)
where Vis the gas flowrate (ACFM), AP is the pressure drop (in. H2O), t is the operating
hours per year, and 1.81 x 10^ is a unit conversion factor. Electricity costs for fan operation
can be determined by multiplying the cost of electricity (in $/kW-hr) by the fan power. Pump
power requirements for wet scrubbers can be determined as follows:14
Pump Power (kW-hr/yr) = (0.746(a)(Z)(S?)(r)) / (3,960 T\) (Eq. 5.4-2)
where Ql is the liquid flowrate (gal/min), Z is the fluid head (ft), Sg is the specific gravity of
the liquid, t is the annual operating time (hr/yr), r\ is the pump-motor efficiency, and 0.746
and 3,960 are unit conversion factors.
Wet scrubbers generate waste hi the form of a slurry. This creates a need for both
wastewater treatment and solid waste disposal operations. Initially, the slurry should be
treated to remove and clean the water. This water can then be reused or discharged. Once the
water is removed, the remaining waste will be hi the form of a solid or sludge. If the solid
waste is inert and nontoxic, it can generally be landfilled. Hazardous wastes will have more
stringent procedures for disposal. In some cases, the solid waste may have value and can be
sold or recycled.2
5.4-32
-------
Table 5.4-6. Annual Cost Parameters for Particulate Scrubbers (Reference 12).
Parameter
Description
Typical Values
Direct Cost Parameters
Operating factor (OF)
Operator labor rate (OR)
Operator shift factor (OS)
Supervisor labor factor (SF)
Maintenance labor rate (MR)
Maintenance shift (MS) factor
Maintenance materials factor (MF)
Electricity rate (ER)
Chemical cost (CC)
Chemical rate (CR)
Wastewater treatment (WT)
Throughput (T)
Waste fraction (WF)
Indirect Cost Parameters
Overhead factor (OV)
Annual interest rate (I)
Operating life (n)
Capital recovery factor (CRF)
Taxes (TAX)
Insurance (INS)
Administrative costs (AC)
Hours of scrubber operation per year
Operator labor pay rate
Fraction of operator shift on scrubber
Fraction of operator labor cost
Maintenance labor pay rate
Fraction of maintenance shift on scrubber
Fraction of maintenance labor cost
Cost of electricity
Cost of chemical conditioning agents
Rate of chemical use
Cost of treating scrubber effluent
Rate of liquid throughput
Fraction of throughput that is wast
Fraction of total labor and (MM) costs
Opportunity cost of the capital
Expected operating life of scrubber
Function of (n) and (I)
Fraction of the TCId
Fraction of the TCId
Fraction of the TCId
8,640 hr/yr
$12.50/hra
0.25"
0.15"
l.lxOR"
0.25"
1.0b
$0.07/kW-hra
$/lb (Site specific)
Ib/hr (Site specific)
$/gal (Site specific)
gal/hr (Site specific)
Site specific
0.60"
7 percent11
10 years6
0.1424°
0.01"
0.01"
0.02b
Estimated for 1996 from currently available information.
Estimates from "CO$T-AIR" Control Cost Spreadsheets (Reference 12).
Capital Recovery Factor is calculated from the following formula: CRF = {1(1 + I)} 4- {(1 + /)"- 1},
where / = interest rate (fraction) and n = operating life (years).
The total capital investment (TCI) can be escalated to current values by using the Vatavuk Air Pollution Control Cost Indexes
(VAPCCI), described in Section 5.4.5.
-------
Table 5.4-7. Annual Cost Factors for Particulate Scrubbers (Reference 11).
Cost Item
Formula"
Factor
Direct Costs
Labor
Operator (OL)
Supervisor (SL)
Maintenance (ML)
Maintenance materials (MM)
Electricity (E)
Chemicals (C)
Wastewater treatment (W)
Total Direct Cost (DC)
Indirect Costs
Overhead
Capital Recovery
Taxes
Insurance
Administrative Costs
Total Indirect Cost (1C)
Total Annual Cost (DC + 1C)
(OF) x (OR) x (OS)
(SF)x(OL)
(OF) X (MR) X (MS)
(MF)X(ML)
Power" X (ER)
(OF)X(CR)X(CC)
(OF)x(T)x(WF)x(WT)
(OV)X(OL+SL+ML+MM)
(CRF)X(TCI)
(TAX) X (TCI)
(INS) x (TCI)
(AC) X (TCI)
A
0.15 A
1.1 A
1.1 A
E
C
W
3.35 A + E + 'C + W + D
2.01 A
0.1424 TCI
0.01 TCI
0.01 TCI
0.02 TCI
2.01 A + 0.1824 TCI
5.36 A + 0.1824 TCI + E + C + W + D
Includes values also described in Table 5.4-6.
Equal to total power requirements, e.g. fan, pump, etc.
-------
5.4.7 References for Section 5.4
1. Cooper, C.D and F.C. Alley. Air Pollution Control: A Design Approach. 2nd ed.
Waveland Press, Prospect Heights, Illinois. 1994.
2. The Scrubber Manual (Revised). The Mcllvaine Company, Northbrook, Illinois.
January 1995.
3. Perry, R.H. and D.W. Green. Perry's Chemical Engineers' Handbook (6th Edition).
McGraw-Hill Publishing Company, Inc. New York, New York. 1984.
4. Sun, J., B.Y.H Liu, P.H. McMurry, and S. Greenwood. A Method to Increase
Control Efficiencies of Wet Scrubbers for Submicron Particles and Paniculate Metals.
J. Ak & Waste Management Association. 44:2. February 1994.
5. Control Techniques for Paniculate Emissions from Stationary Sources - Volume 1
(EPA-450/3-81-005a, NTIS PB83-127498). U.S. Environmental Protection Agency,
Office of Ah- Quality Planning and Standards. Research Triangle Park, North
Carolina. September 1982.
6. Compilation of Air Pollutant Emission Factors (AP-42). Volume I (Fifth Edition).
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina.
January 1995.
7. Lerner, B.J. "Paniculate Wet Scrubbing: The Efficiency Scam" in the Proceedings of
the A&WMA Specialty Conference on "Paniculate Matter; Health and Regulatory
Issues (VIP-49)" held on April 4-6, 1995, hi Pittsburgh, Pennsylvania. A&WMA,
Pittsburgh, Pennsylvania. 1995.
8. Source Category Emission Reductions with Paniculate Matter and Precursor Control
Techniques. Prepared for K. Woodard, U. S. Environmental Protection Agency,
Research Triangle Park, North Carolina (AQSSD/IPSG), under Work Assignment
11-16 (EPA Contract No. 68-03-0034), "Evaluation of Fine Paniculate Matter Control."
September 30, 1996.
9. Vatavuk, W.M. Estimating Costs of Air Pollution Control. Lewis Publishers,
Chelsea, Michigan. 1990.
10. Vatavuk, W.M. and Neveril, R.B., Factors for Estimating Capital and Operating
Costs, Chemical Engineering, November 3, 1980, pp. 157-162.
11. Schifftner, K.C. and H.E. Hesketh. Wet Scrubbers: A Practical Handbook. Lewis
Publishers, Chelsea, Michigan. 1986.
5.4-35
-------
12. Vatavuk, W.M. "CO$T-AIR" Control Cost Spreadsheets. Innovative Strategies and
Economics Group, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina. February 1996.
13. Vatavuk, W.M. Escalate Equipment Costs. Chemical Engineering. December 1995.
pp. 88-95.
14. OAQPS Control Cost Manual (Fourth Edition, EPA 450/3-90-006). U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards,
Research Triangle Park, North Carolina. January 1990.
5.4-36
-------
5.5 INCINERATORS
This section presents the basic operating principles, typical designs, industrial
application, and costs of incinerators used as control devices. An incinerator is the only PM
control device that does not concentrate the PM for subsequent disposal. An incinerator
utilizes the principles of combustion to control pollutants. Incinerators used as add-on control
devices are however, seldom used to remove only paniculate matter (PM); PM control is
usually desirable as a secondary treatment of a gas stream with a high volatile organic
compounds (VOC) content.1 The type of PM that is usually controlled by an incinerator is
commonly composed of soot (particles formed as a result of incomplete combustion of
hydrocarbons (HCs)), coke, or carbon residue. There are two basic types of incinerators used
as add-on control devices: thermal and catalytic. For purposes of PM control, the use of a
catalytic incinerator is limited because catalysts are subject to blinding from the PM.2
There are several advantages to using incinerators for waste air streams that contain
VOC and PM. These advantages are: simplicity of operation; capability of steam generation
or heat recovery in other forms; and capability for virtually complete destruction of organic
contaminants. Disadvantages include: relatively high operating costs (particularly associated
with fuel requirements); potential for flashback and subsequent explosion hazard; and
incomplete combustion possibly creating potentially worse pollution problems.3 High gas
velocities are usually required for incinerators used for PM control to prevent settling of PM.5
This may increase the incinerator size necessary to achieve the minunum required gas
residence time.
5.5.1 Incinerator Control Mechanisms
Incinerator control is based on the principle that at a sufficiently high temperature and
adequate residence time, any HC can be oxidized to carbon dioxide (CO2) and water. In an
incinerator, PM containing HCs is first vaporized to a gas and then oxidized.1
To achieve complete combustion, i.e. convert all the HC to CO2 and water, sufficient
space, time, turbulence and temperature high enough to ignite the constituents must be
provided by the incinerator. The "three T's" of combustion: time, temperature, and
turbulence, govern the speed and completeness of the combustion reaction. For complete
combustion, oxygen must come into close contact with the combustible molecule at sufficient
temperature and for a sufficient length of time for the reaction to be complete.2
The combustion time required for PM control is dependent on particle size and
composition, oxygen content of the furnace, atmosphere, furnace temperature, gas velocity,
and extent of mixing of the combustibles. For PM less than 100 yum hi diameter, the
combustion rate is controlled by chemical kinetics; for PM greater than 100 ,um, diffusion
controls the combustion rate.1 In collection devices (ESP's, fabric filters, scrubbers) diffusion
controls the collection rate of particles less than 1 /zm in diameter.
5.5-1
-------
For particles smaller than 100 /urn the time required for complete combustion can be
calculated using the following equation:1
tc = (p dp)/<2 K, pg)
where for coke and carbon residue,
and for soot,
= 8,710 exp(-35,700/RTs)
K, = (1.085 x 104 Ty") (exp(-39,300/RTs)
Eq. 5.5-1
Eq. 5.5-2
Eq. 5.5-3
where t,. is the combustion time for a chemical kinetics controlled reaction (sec), p is the
density of particle (g/cm3), dp is the diameter of particle (cm), K,. is the surface reaction rate
coefficient (g/cm2-sec-atm), pg is the partial pressure of oxygen in combustion air (arm), R is
the universal gas law constant (82.06 atm-cm3/mole-°K), Ts is the surface temperature of the
particle (assumed to be the incinerator temperature) (°K).
o
o
0>
s
1
.£>
I
u
9.00E-07
8.00E-07
O.OOE+00
8 10 12
Particle Size
14
18
20
Figure 5.5-1. Calculated Theoretical Residence Times for Various-sized Coke PM in an
Incinerator at Various Temperatures
5.5-2
-------
With the proper residence time, complete combustion should result hi > 99 percent
control of particles containing HCs. Figure 5.5-1 shows the theoretical residence tune needed
for > 99 percent control of various sized coke PM hi an incinerator operated from 1200-
2000°F calculated using the above equations.1
Although residence time and incinerator temperature are the primary incinerator
parameters affecting incinerator performance, other important parameters are the heat content
and water content of the gas stream, and the amount of excess combustion air (i.e. amount
above the stoichiometric amount needed for combustion). Combustion of gas streams with
heat contents less than 50 Btu per standard cubic foot of air (SCF) usually will require
supplemental fuel to maintain the desired combustion temperature. Supplemental fuel may
also be needed for flame stability, regardless of the heat content of the gas.4
For incinerators operated above 1400 °F, the oxidation reaction rates become much
faster than the gas diffusion mixing rate. As a result, the combustion reaction may be
hindered because sufficient oxygen molecules are not in proximity to the HCs. To ensure that
this does not occur, mixing must be enhanced via vanes or other physical methods.5
5.5.2 Types of Incinerators
As discussed above, there are two basic types of incinerators, thermal and catalytic.
Both types of incinerator may use heat exchangers to recover some of the heat energy from the
incinerator. Therefore, this section discusses both types of incinerators as well as heat
exchangers.
5.5.2.1 Thermal Incinerators
A typical thermal incinerator is a refractory-lined chamber containing a burner (or set
of burners) at one end. Thermal incinerators typically use natural gas to supplement the
caloric content of the waste gas stream. In a thermal incinerator, the combustible waste gases
pass over or around a burner flame into a residence chamber where oxidation of the waste
gases is then completed. The most recent guidelines for incinerators to promote more
complete destruction of VOC are:5
• A chamber temperature high enough to enable the oxidation reaction to proceed
rapidly to completion (1200-2000 °F or greater);
• Flow velocities of 20-40 feet per second, to promote turbulent mixing between
the hot combustion products from the burner, combustion air, and waste stream
components; and
• Sufficient residence time (approximately 0.75 seconds or more) at the chosen
temperature for the oxidation reaction to reach completion.
5.5-3
-------
The following sections discuss the two types of thermal incinerators: discrete burner
and distributed burner. Both types may also use heat recovery equipment. This equipment is
discussed in Section 5.5.2.3 below.
5.5.2.1.1 Discrete Burner Thermal Incinerator. In a discrete dual burner
incinerator, shown in Figure 5.5-2, the waste gas stream and combustion air feed into a
premixing chamber fitted with a (auxiliary) discrete fuel burner. In this chamber, both gases
are thoroughly mixed and pre-heated by the auxiliary burner. The mixture of hot reacting
gases then passes into the main combustion chamber where another (primary) burner is
located. This chamber is sized to allow the mixture enough time at the elevated temperature
for the oxidation reaction to reach completion. Energy can the be recovered from the hot flue
gases in a heat recovery section.6
5.5.2.1.2 Distributed Burner Thermal Incinerator. Thermal incinerators (that
use natural gas as the supplemental fuel) may also use a grid-type, or distributed, gas burner.
This gas burner configuration is shown in Figure 5.5-3. In a distributed thermal incinerator,
small gas flame jets on a grid surface ignite the vapors in the gas as it passes through the grid.
The grid acts as a baffle to promote mixing before the gases enters the second part of the
incinerator chamber. Because there are many small flames distributed on the entire cross-
section of the combustion chamber and the vapors are well-mixed, this arrangement enables
the gas vapors to burn at a lower chamber temperature and allows for the use of less fuel than
the discrete burner configuration, described above.4 In the discrete burner, vapors and
particles are more likely to survive the 'single large flame initially, so the chamber must be
maintained at a higher temperature to ensure complete combustion.
5.5.2.2 Catalytic Incinerators
A catalytic incinerator is not usually recommended as a control device for PM since the
PM, unless removed prior to incineration, will often coat the catalyst so that the catalyst active
sites are prevented from aiding in the oxidation of pollutants in the gas stream. This effect of
PM on the catalyst is called blinding.2 Despite this drawback, catalytic incinerators are
sometimes used for PM control in the chemical manufacturing and textile industries, and for
combustion sources such as 1C engines, boilers, and dryers.7 Therefore, a brief description of
this type of incinerator is included here.
Catalytic incinerators are very similar to thermal oxidation, with the primary difference
that the gas, after passing through the flame area, passes through a catalyst bed.5 The catalyst
has the effect of increasing the oxidation reaction rate, enabling conversion at lower reaction
temperatures than in thermal incinerator units. Catalysts, therefore, also reduce the
incinerator volume/size.5 Catalysts typically used for VOC incineration include platinum and
palladium. Other formulations include metal oxides, which are used for gas streams
containing chlorinated compounds.4
5.5-4
-------
Stack
Auxiliary
Burner
(Discrete)
Optional Heat
Recovery
V
Combustion
Chamber
Figure 5.5-2. Schematic Diagram of a Discrete Burner Thermal Incinerator (Reference 4).
5.5-5
-------
Stack
gas
inlet
Incinerator Chamber
\
1
Auxiliary Fuel
(Natural Gas)
Fan
Optional Heat
Recovery
Figure 5.5-3. Schematic Diagram of a Distributed Burner Thermal Incinerator (Reference 6).
5.5-6
-------
A schematic of a catalytic incinerator is presented hi Figure 5.S-4.4 in a catalytic
incinerator, the gas stream is introduced into a mixing chamber where it is also heated. The
waste gas usually passes through a recuperative heat exchanger (discussed below), where it is
preheated by post-combustion gas.11 The heated gas then passes through the catalyst bed.
Oxygen and VOCs migrate to the catalyst surface by gas diffusion and are adsorbed onto the
catalyst active sites on the surface of the catalyst where oxidation then occurs. The oxidation
reaction products are then desorbed from the active sites by the gas and transferred by
diffusion back into the gas stream.8
As discussed above, PM can rapidly blind the pores of the catalysts and deactivate the
catalyst over time. Because essentially all the active surface of the catalyst is contained hi
relatively small pores, the PM need not be large to blind the catalyst. No general guidelines
exist as to the PM concentration and size that can be tolerated by catalysts because the pore
size and volume of catalysts vary greatly.9 This information is likely to be available from the
catalyst manufacturers.
The advantages of catalytic combustion reactors over thermal incinerators, therefore,
include:5
• Lower fuel requirements,
• Lower operating temperatures,
• Little or no insulation requirements,
• Reduced fire hazards, and
• Reduced flashback problems.
.5
The disadvantages include:
• Higher capital costs,
• Catalyst blinding causes operational problems and/or higher maintenance
requirements (annual costs),
• PM may need to be precollected, and
• Spent catalyst that cannot be regenerated may need to be disposed.
5.5.2.3 Heat Recovery Equipment
Since the flue gas that is still hot after exiting the incinerator, heat may be recovered
with the proper auxiliary incinerator equipment. Heat recovery equipment for an incinerator
can be either recuperative or regenerative. Recuperative heat exchangers, that recover heat on
a continuous basis, include crosscurrent-, countercurrent-, and cocurrent-flow heat exchangers.
For a given heat flow and temperature drop, recuperative heat exchanger surface requirements
will be the lowest hi a countercurrent flow configuration.
5.5-7
-------
Auxiliary
Burners
Auxiliary
Burners
To Atmosphere
Stack
Fan
Catalyst Bed
Waste Heat
Recovery (optional)
Mang Chamber
Figure 5.5-4. Schematic Diagram of a Catalytic Incinerator (Reference 4).
5.5-8
-------
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 periodically reversed. The heat sink and heat transfer
area for regenerative heat exchangers can be either a fixed bed, a moving bed or a rotary
cylinder.1
5.5.3 Control Efficiency
5.5.3.1 Control Efficiency for Volatile Organic Compounds
Theoretically, all organic material, including VOC, are combustible with combustion
efficiency limited only by cost. On the basis of studies of thermal incinerator efficiency, it has
been concluded that at least 98 percent VOC destruction (or a 20 part per million by volume
(ppmv) VOC exit concentration) is achievable by all well-designed incinerators. An estimate
of 98 percent efficiency is predicted for thermal incinerators operating at 1,400°F or higher,
with at least 0.75 seconds residence tune.5 If a thermal incinerator is properly designed and
operated to produce the optimum conditions in the combustion chamber, it should be capable
of higher than 99 percent destruction efficiencies for nonhalogenated VOC, when the VOC
concentration in the gas stream is above approximately 2,000 ppmv.6
5.5.3.2 Control Efficiency for Particulate Matter
Controlled emissions and/or efficiency test data for PM hi incinerators are not
generally available hi the literature. Emission factors for PM in phthalic anhydride processes
with incinerators were available, however.10 The PM control efficiencies for these processes
were calculated from the reported emission factors and are shown hi Table 5.5-1. The PM
control efficiencies ranged from 79 to 96 percent control for total PM.
In EPA's 1990 National Inventory,7 incinerators were used as control devices for PM
to achieve from 25 to 99.9 percent control of PM10 at point source facilities. The VOC control
reported for these devices ranged from 0 to 99.9 percent. These ranges of control efficiencies
are large because they include facilities that do not have VOC emissions and control only PM
(these facilities would report 0 percent efficiency for VOC control), as well as facilities which
have low PM emissions and are primarily concerned with controlling VOC.
5.5.4 Applicability
Although incinerators can be used to any organic material, then- application is limited
to a range of gas vapor concentration. To prevent explosions, the vapor concentration must be
substantially below the gas lower flammable level (lower explosive limit [LEL]). As a rule, a
factor of 4 is employed to give a margin of for safety.2 Therefore, incinerators are not likely
to be used for processes with very high VOC content. The presence of halogens also requires
additional equipment such as scrubbers for acid gas removal.4
5.5-9
-------
Table 5.5-1 PM Control Efficiencies for Thermal Incinerators in
Phthalic Anhydride Manufacturing Processes (Reference 10)
PM Emission Factor
(Ib PM/ton product)
Process Unit
O-xylene Processing
Oxidation
Pretreatment
Distillation
Naphthalene Processing
Oxidation
Pretreatment
Distillation
Uncontrolled
138
13
89
56
5
38
Controlled
7
0.7
4
11
1
8
Calculated
Control
Efficiency
(percent)
95
95
96
80
80
79
Thermal incinerators can be designed to handle minor fluctuations hi flow rate.
However, processes with the potential for excessive fluctuations in flow rate (i.e., process
upsets) may not be suitable for incinerator use, since control efficiency could decrease outside
the acceptable range.4 Flares may be an appropriate control for processes with excessive
fluctuation potential. Table 5.5-2 presents the operating conditions required for satisfactory
incinerator performance hi various industrial applications.3 Note that the residence time and
incinerator temperature required for PM control is much higher than for non-PM sources.
An examination of the EPA's 1990 National Inventory,7 presented in showed that the
primary source categories hi which incinerators were used for PM control were:
Petroleum and Coal Production
Chemical and Allied Product Manufacturing
Primary Metal Industries
Electronic and Other Electric Equipment.
These source categories were identified from the reported data hi the 1990 National
Inventory,7 and correspond to facilities that reported PM10 control efficiencies for incinerators
likely to have been used as primary control devices.
5.5-10
-------
Table 5.5-2 Operational Requirements for Satisfactory Incinerator
Performance for Various Industrial Applications
and Control Levels (Reference 3)
Application
HC Control
HC + CO
Odor
Low control
Medium control
High control
Smokes/Plumes
White smoke (liquid mist)
HC and CO
Black smoke (soot and other
combustible PM)
Control
Level
(percent)
>90
>90
50-90
90-99
>99
>99
>90
>99
Residence
Time
(sec)
0.3-0.5
0.3-0.5
0.3-0.5
0.3-0.5
0.3-0.5
0.3-0.5
0.3-0.5
0.7-1.0
Temperature
(°F)
1 100-1250*
1250-1500
1000-1200
1100-1300
1200-1500
800-1000"
1250-1500
1400-2000
a Temperatures of 1400 to 1500°F may be required if there is a significant amount of
any of the following: methane, cellosolve, and substituted aromatics (e.g., toluene
and xylenes).
b Operation for plume abatement only is not recommended, since this merely converts
a visible hydrocarbon emission into an invisible one and frequently creates a new
odor problem because of partial oxidation in the incinerator.
5.5.5 Costs of Incinerators
The costs of installing and operating an incinerator include both capital and annual
costs. Capital costs are all of the initial equipment-related costs of the incinerator. Annual
costs are the direct costs of operating and maintaining the incinerator for one year, plus such
indirect costs as overhead; capital recovery; and taxes, insurance, and administrative charges.
The folio whig sections discuss capital and annual costs for incinerators, referenced to the
fourth quarter of 1996, unless otherwise noted.
5.5-11
-------
Incinerators designed for PM control are likely to have higher costs than incinerators
designed for VOC control, because of the higher temperatures and longer gas residence tunes
are needed for PM destruction (see Table 5.5-2). Incinerators designed for PM control are
also likely to need more supplemental fuel to maintain the higher temperatures and larger
combustion chambers to achieve the longer residence times. Since the incinerator cost data
presented below were probably derived for incinerators designed for VOC control only, the
actual costs for incinerators designed for PM control are likely to be higher.
The use of a catalytic incinerator for PM control is limited because catalysts are subject
to poisoning/blinding from PM;2 consequently, only thermal incinerator costs are discussed in
this section. For information on the costs of catalytic incinerators, consult Estimating Costs of
Air Pollution Control11 and EPA's "CO$T-AIR" Control Cost Spreadsheets.12
5.5.5.1 Capital Costs
The total capital investment (TCI) for incinerators includes all of the initial capital
costs, both direct and indirect. Direct capital costs are the purchased equipment costs (PEC),
and the costs of installation (foundations, electrical, piping, etc.). Indirect costs are related to
the installation and include engineering, construction, contractors, start-up, testing, and
contingencies. The PEC is calculated based on the incinerator specifications. The direct and
indirect installation costs are calculated as factors of the PEC.11 The equipment cost presented
hi Table 5.5-3 are the TCI cost factors for custom incinerators (as opposed to packaged units).
The flue gas flow rate and auxiliary fuel requirement are the most important sizing
parameters for a thermal incinerator. The former determines the equipment size and cost,
while the latter comprises most of annual operating and maintenance costs. These parameters
are interdependent, based on material and energy balances taken around the incinerator.9
Figure 5.5-5 shows total capital investment vs. flow rate (size) for a thermal incinerator
with recuperative heat recovery equipment.12 Three levels of heat recovery are shown hi
Figure 5.5-5: 0 percent, 35 percent, and 50 percent. For the purposes of the figure, the
thermal incinerator was assumed to operate at a combustion temperature of 1600°F and the
waste gas was assumed to have a heat content of 4 Btu/SCF. The curves illustrate two
phenomena: 1) the direct proportionality of capital cost to flow rate (size), and 2) the
proportionality of capital cost to heat recovery efficiency. That is, capital costs increase with
both increasing flow rate (size) and increasing heat recovery efficiency.
Figure 5.5-6 shows total capital investment vs. flow rate (size) for thermal incinerators
with 85 percent and 95 percent regenerative heat recovery systems.12 As in the previous
figure, the thermal incinerator was assumed to operate at a combustion temperature of 1700°F
and the waste gas was assumed to have a heat content of 4 Btu/SCF. Also, as hi the previous
figure, capital costs for incinerators with regenerative heat recovery systems increase with
increasing flow rate (size) and decrease with increasing heat recovery efficiency.
5.5-12
-------
Table 5.5-3 Capital Cost Factors for Thermal Incinerators (from Reference 11)
Cost Item Factor
Direct Costs
Purchased equipment costs
Incinerator + auxiliary equipment As estimated (A)
Instrumentation 0.10 A
Sales taxes 0.03 A
Freight 0.05 A
Total Purchased Equipment Cost (PEC) B = 1.18 A
Direct installation costs
Foundations and supports 0.08 B
Handling and erection 0.14 B
Electrical 0.04 B
Piping 0.02 B
Insulation for ductwork 0.01 B
Painting Q.Q1 B
Total direct installation cost 0.30 B
Site Preparation and Buildings As required (Site)
Total Direct Cost, DC 1.30 B + Site
Indirect Costs (installation)
Engineer ing 0.10B
Construction and field expense 0.05 B
Contractor fees 0.10 B
Start-up 0.02 B
Performance test 0.01 B
Contingencies Q.Q3 B
Total Indirect Cost (1C) 0.31 B
Total Capital Investment = DC + 1C 1.61 B + Site
5.5-13
-------
700,000
600,000
§
en
U
C
&
U
Heat Recovery
—g— 35% Heat Recovery
t 0% Heat Recovery
S 200,000
£
100,000
5,000 10,000 15,000 20,000 25,000 30,000 35,000 40,000 45,000 50,000
Outlet Flowrate (scfm)
Figure 5.5-5. Total Capital Investment vs. Flow Rate for a Thermal Incinerator with 0, 35,
and 50 Percent Recuperative Heat Recovery (Reference 12).
3.5
3.0
2.5
2.0
i
"5 0.5
0.0
95% Heat Recovery
85% Heat Recovery
0 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 100,000
Inlet Flowrate (acfm)
Figure 5.5-6. Total Capital Investment vs. Flow Rate for a Regenerative Thermal Oxidizer
with 85 and 95 Percent Heat Recovery (Reference 12).
5.5-14
-------
A comparison between the capital cost of incinerators with recuperative vs.
regenerative heat recovery systems shows that for the same size incinerator, the capital
investment of a regenerative heat recovery system is over twice the capital investment required
for an incinerator with a recuperative heat recovery system.
5.5.5.2 Annual Costs
The total annual cost of an incinerator consists of both direct and indirect costs. Direct
annual costs are those associated with the operation and maintenance of the incinerator. These
include labor (operating, supervisory, coordinating, and maintenance); maintenance materials;
operating materials; electricity; and supplemental fuel, if applicable.
Indirect annual costs include taxes, insurance, administrative costs, overhead, and
capital recovery. All of these costs except overhead are dependent on the TCI. Table 5.5-4
lists the annual cost parameters that impact incinerator costs, with typical values provided for
each parameter. Table 5.5-5 provides the annual cost factors for incinerators. It is difficult to
generalize these costs for all incinerators, since annual costs are very site-specific.11
The supplemental fuel and electricity requirements for an incinerator are likely to have
a large impact on incinerator annual costs. The requirements for each can be estimated from
incinerator design values. The auxiliary heat requirement to be supplied by the fuel, usually
natural gas, can be calculated using the incinerator design equations described below.
An incinerator is designed to handle a total volumetric gas flow rate (Qf) equal to the
waste gas inlet flow rate (QO, which is known, and auxiliary fuel gas flow rate (QJ:
Qf = Qi + Qa (Eq. 5.5-4)
and where the requirements for auxiliary fuel gas are determined with the following equation:
Qa = (x/y)(Qi) (Eq. 5.5-5)
for x = (LlCpfCTf-T^-CCpiCTi-T^-h, (Eq. 5.5-6)
y = h,-l.lCpf(Tf-Tr) (Eq. 5.5-7)
where Qf is the flue gas flow rate (SCFM), Qj is the inlet waste gas flow rate (SCFM),Qa is the
auxiliary fuel gas (heat) requirement (SCFM), Cpf is the mean heat capacity of gas leaving the
combustion chamber (Btu/SCF-°F), Cpi is the mean heat capacity of gas entering the
combustion chamber (Btu/SCF-°F), Tf is the combustion chamber temperature (°F), T; is the
waste gas inlet temperatures (°F), Tr is the reference temperature, equal to the inlet fuel
temperature (typically 70°F), 1^ is the waste gas heat content (Btu/SCF), and ha is the fuel
heating value (Btu/SCF).
5.5-15
-------
Table 5.5-4. Incinerator Annual Cost Parameters (from Reference 11)
Parameter
Description
Typical Values
Direct Cost Parameters
Operating factor (OF)
Operator labor rate (OR)
Maintenance Labor Rate (ML)
Operator shift factor (OS)
Maintenance shift factor (MS)
Electricity rate (ER)
Fuel (F)
Indirect Costs
Annual Interest Rate (I)
Operating Life (n)
Capital Recovery Factor (CRF)
Taxes (TAX)
Insurance (INS)
Administrative Costs (AC)
Yearly incinerator (INC) operation hours
Cost of operator labor
Cost of maintenance labor
Fraction of operator's shift spent on INC
Fraction of maintenance shift spent on INC
Cost of electricity
Cost of fuel (natural gas)
Opportunity cost of the capital
Expected operating life of INC
Function of (n) and (I)
Fraction of TCId
Fraction of TCId
Fraction of TCId
8,000
$12.50/hra
l.l(OR)a
0.5
0.5
$0.07/kW-hra
$2.30/103SCFa
7 percent0
10 yearsc?
0.0944"
0.01°
0.01C
0.02C
Estimated for 1996 from currently available information.
Estimates from "CO$T-AIR" Control Cost Spreadsheets (Reference 12).
Capital recovery factor is calculated from the following formula:
CRF= {1(1+I)} 4- {(1+7) -I},
where / = interest rate (fraction) and n = operating life (years).
The total capital investment (TCI) can be escalated to current values by using the Vatavuk Air
Pollution Control Cost Indicies (VAPCCI), described in Section 5.4.5.
-------
Table 5.5-5. Annual Cost Factors for Incinerators (Reference 12).
Cost Item
Formula
Factor
Direct Costs
Labor
Operator (OL)
Supervisor (SL)
Maintenance (ML)
Maintenance materials (MM)
Electricity (E)
Fuel (F)
Total Direct Cost (DC)
Indirect Costs
Overhead
Capital Recovery
Taxes
Insurance
Administrative Costs
Total Indirect Cost (1C)
Total Annual Cost (DC + 1C)
(OF) X (OR) X (OS)
(SF)X(OL)
(OF) X (MR) x (MS)
(MF)X(ML)
Power6 x (ER)
Fuelc X (FR)
(OV)X(OL+SL+ML+MM)
(CRF)X(TCI)
(TAX) X (TCI)
(INS) X (TCI)
(AC) x (TCI)
A
0.15 A
1.1 A
1.1 A
E
3.35 A + E + F
2.01 A
0.1424 TCI
0.01 TCI
0.01 TCI
0.02 TCI
2.01 A + 0.1824 TCI
5.36 A + 0.1824 TCI + E + F
a Includes values also described in Table 5.5-5.
b Equal to the total power requirements, i.e. electricity and fan.
c Equal to the auxiliary fuel requirements.
-------
Electricity to run the incinerator exhaust fan is calculated with the following equation:
Fan Power (kW) = (1.575x10^) AP Q / n (Eq. 5.5-8)
where AP is the system pressure drop (inches of water), Q is the waste gas volumetric flow
rate through system (ACFM), and n is the efficiency of fan and motor (generally 0.50-0.70).
Figure 5.5-7 shows annual operating cost curves for an example thermal incinerator
with recuperative heat recovery systems at three levels of heat recovery efficiency: 0, 35, and
50 percent, and 85 percent regenerative heat recovery. For these curves, the example
incinerator was assumed to operate 8,000 hours per year, at a combustion temperature of
1700°F, with a waste gas heat content of 4 Btu/SCF. Figure 5.5-7 shows that annual
operating costs for incinerators with recuperative heat recovery decrease with increasing heat
recovery system efficiency, and increase with increasing inlet flow rates (size).
Figure 5.5-7 also shows that annual costs for the example incinerator with regenerative
heat recovery increase with inlet flow rate. Regenerative thermal incinerators achieve higher
heat recovery (*85 percent vs. ^50 percent) at lower annual costs than recuperative systems.
However, the higher capital costs of regenerative systems (see Figure 5.5-6) compared
trecuperative systems (see Figure 5.5-5), present a trade-off in the choice incinerator type.
o
X
&
ao
5
S.
o
73
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0% Heat Recovery (HR)
35% Recuperative HR
50% Recuperative HR
85% Regenerative HR
10,000
20,000 30,000
Inlet Flowrate (acfm)
40,000
50,000
Figure 5.5-7. Annual Costs for Incinerators with Recuperative and Regenerative Heat
Recovery (Reference 12).
5.5-18
-------
5.5.6 Energy and Other Secondary Environmental Impacts
No liquid, solid or hazardous wastes are generated from the use of thermal
incinerators. As discussed above, the energy impacts of incinerator operation include that
associated with the energy required to run the fan and are proportional to the gas flow rate and
the system pressure drop.
Nitrogen oxides are also generated as air pollution during incineration. Because of the
lower operating temperatures of catalytic incinerators, less NOX is generated with this type of
incinerator. Based on the combustion of natural gas only, thermal incinerators have the
potential to generate 100 pounds (Ib) of NOX per 106 SCF of natural gas combusted, and
catalytic incinerators have the potential to generate 50 Ib of NOX per 106 SCF of natural gas.13
5.5.7 References for Section 5.5
1. Control Techniques for Particulate Emissions from Stationary Sources - Volume 1
(EPA-450/3-81-005a, NTIS PB83-127498). U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards. Research Triangle Park, North
Carolina. September 1982.
2. Theodore, L., and A.J. Buonicore. Air Pollution Control Equipment. Volume II:
Gases. CRC Press, Inc., Boca Raton, Florida. 1988.
3. Perry's Chemical Engineers' Handbook. Sixth Edition. R.H. Perry and D.W. Green,
Eds. McGraw-Hill, Inc., New York, New York. 1984.
4. Hazardous Air Pollutant Emissions from Process Units in the Synthetic Organic
Chemical Manufacturing Industry-Background Information for Proposed Standards.
Volume IB: Control Technologies. U.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards, Research Triangle Park, North Carolina.
November 1992.
5. Buonicore, A.J. "Incineration" in the Air Pollution Engineering Manual.
A.J. Buonicore, and W.T. Davis, Eds. Air & Waste Management Association,
Pittsburgh, Pennsylvania; and Van Nostrand Reinhold, New York, New York. 1992.
6. Reed, R.J. North American Combustion Handbook. North American Manufacturing
Company, Cleveland, Ohio. 1978.
7. 1990 National Inventory. (Available at earthl.epa.gov/pub/gopher/Emis.Inventory).
U.S. Environmental Protection Agency, Research Triangle Park, NC. January 1996.
5.5-19
-------
8. Control Techniques for Volatile Organic Emissions from Stationary Sources (EPA-
450/2-78-002). U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina. May 1978.
9. OAQPS Control Cost Manual (Fourth Edition, EPA 450/3-90-006). U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards,
Research Triangle Park, North Carolina. January 1990.
10. Compilation of Air Pollutant Emission Factors (AP-42). Volume I (Fifth Edition).
U.S. Environmental Protection Agency, Research Triangle Park, North Carolina.
January 1995.
11. Vatavuk, W.M. Estimating Costs of Air Pollution Control. Lewis Publishers,
Chelsea, Michigan. 1990.
12. Vatavuk, W.M. "CO$T-AIR" Control Cost Spreadsheets. Innovative Strategies and
Economics Group, Office of Air Quality Planning and Standards, U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina. February 1996.
13. Organic Chemical Manufacturing, Volume 4: Combustion Control Devices
(EPA-450/3-80-026). U.S. Environmental Protection Agency, Research Triangle Park,
North Carolina. December 1980.
5.5-20
-------
6. INDUSTRIAL FUGITIVE EMISSION CONTROLS
This section describes measures used to control fugitive PM emissions from industrial
sources. Fugitive PM emission sources can be divided into two broad categories-process
fugitive emission sources and fugitive dust emission sources. Process fugitive emissions
sources include emissions from mechanical and metallurgical operations that receive and/or
generate dusty material. Fugitive dust emission sources relate to the transfer, storage, and
handling of dusty materials and include those sources from which particles are entrained by the
forces of nature acting on exposed dusty surfaces or from vehicle motion on dusty roads.
The most widely used methods of controlling process fugitives are local ventilation and
building enclosure/evacuation. Both types of systems have their advantages and drawbacks,
but local ventilation is generally more cost effective. Process optimization, good operation
and maintenance (O&M), and other industry-specific practices can also be quite effective in
reducing process fugitive emissions. However, both the selection of the system and the
ultimate performance of the system are related to industry and facility-specific design and
operating characteristics.
For most industrial plants, paved and unpaved roads are the primary sources of fugitive
dust emissions. Fugitive dust emissions from handling operations for storage pile materials
are usually less significant in comparison to road sources, unless the moisture content of the
storage pile materials is extremely low. Emissions due to wind erosion of storage piles are
likewise less significant unless wind speeds are unusually high.1 Low wind speeds can result
in significant emissions if storage pile materials are fines (e.g. cement kiln dust or materials
collected by fabric filters or ESPs)
The control of road dust from both paved and unpaved roads, therefore, can achieve a
significant reduction hi fugitive dust emissions. Paving of unpaved roads; eliminating,
reducing, or managing truck transportation; and street cleaning are the most effective
techniques to reduce fugitive dust emissions from roads.
More information about fugitive dust emission sources and controls can be found in the
EPA publications Fugitive Dust Background Document and Technical Information Document
for Best Available Control Measures,2 and Compilation of Air Pollutant Emission Factors (AP-
42), Volume I: Stationary Point and Area Sources?
6.1 ENCLOSURES AND VENTILATION
Partial or full enclosures, windbreaks, hoods and other ventilation systems, and
complete building evacuation are widely used methods to capture and control fugitive PM
emissions. These methods are usually used with traditional stack PM control devices (e.g.
fabric filters or scrubbers) to collect the captured PM. Processes amenable to this type of
control include materials handling devices such as conveyors, elevators, feeders, loading and
6-1
-------
unloading operations, and bagging; solids benefication, such as crushing, screening, and other
classifying operations; mining, i.e. drilling and crushing; and furnaces, ovens, and dryers.7
6.1.1 Local Ventilation Systems
Local ventilation systems can consist of a "secondary" hood at a localized source of
PM emissions or large canopy-type hood suspended over the entire source. An example of a
secondary local hood is a mobile hood that is used to collect emissions from pots or other
containers that are set aside for cooling. Ventilation systems are usually uniquely designed to
conform with the facility configuration and need for process access; these factors, however,
can limit their performance as well as their design. Ventilation hooding and its ductwork may
be difficult to retrofit in some facilities due to space limitations. In addition, local ventilation
systems may limit personnel and equipment access. For these reasons, a local ventilation
system may not be a feasible method of process fugitive emissions control for some
operations. Design information about local ventilation systems in general and for specific
applications can be found in the most recent edition of the American Conference of
Governmental Industrial Hygienists (ACGIH) publication: Industrial Ventilation: A Manual of
Recommended Practice.
Most ventilation systems are designed to meet several objectives.1 First, the hood must
enclose the source to the degree possible without excessively interfering with the access
needed for normal operations. Second, the hood should be configured in such a way that
natural buoyancy or mechanical forces direct the plume into, rather than away from, the hood.
Finally, the system must be designed with sufficient exhaust ventilation to maintain
recommended face velocities at all hood faces. Typically, these velocities are in the range of
75 to 150 meters per minute. Additionally, for buoyant plumes that generate a natural draft,
the ventilation rate must exceed the plume generation rate, or "spillage" from the hood will
occur.1
Metal operations, both primary and secondary, generate a large quantity of fugitive PM
emissions. One of the major sources of metallurgical process fugitive emissions that can be
controlled by local ventilation are stationary-type furnaces such as blast furnaces,
reverberatory furnaces, and cupolas. Hoods may be designed to collect gas perpendicular to
the buoyant gas flow, in which high face velocities are required; other designs may be such
that the buoyant gas plume is directed into the hood. Figure 6-1 shows a local ventilation
system at a blast furnace slag tapping area.1-4 A very different local ventilation design would
be needed for nonstationary electric arc and rotary furnaces that rotate during operation. A
key feature of these systems is that charging and tapping occur hi the same general area.
Hence, hooding must be designed in such a way that it does not interfere with either operation.
Material handling operations can also be equipped with local ventilation to control
fugitive PM emissions. Figure 6-2 shows a local ventilation system at "skip hoist" loading
station that is part of a metallurgical operation.
6-2
-------
SLAG TAPPING HOOD
FRONT SURFACE
OF HOOO RAISED
USING CABLE AND
PULLEY SYSTEM
METAL OUC"!
SLAG
CONTAINER
Figure 6-1. Schematic of a Slag-tapping Hood at a Blast Furnace (from Reference 4).
6-3
-------
AIR INTAKE
NEW RAW
MATERIALS
CONFINEMENT
SKIP HOIST
ENCLOSURE
EXHAUST DUCTS
TO TOP OF BLAST
FURNACE
AIR INTAKE
SKIP HOIST
BUCKET
STEEL PLATE
BARRIER
FLUE OUST AND
AGGLOMERATED
FLUE DUST
SLAG
LIMESTONE CHIPS
Figure 6-2. Schematic of a Local Ventilation System at a "Skip Hoist" Loading Station
6-4
-------
An air curtain capture system is a specially designed local ventilation system that can
capture fugitive emissions from a process without interfering in normal operations, such as the
use of an overhead crane. With an air curtain, air is blown across the space above the PM-
generating operation using a plenum or row of nozzles designed to form an air sheet which
causes as little turbulence as possible. The curtain air, entrained air (from above and below
the curtain) and PM, including fine PM fumes, are captured by the exhaust system. Capture
of fugitive PM is greater than 90 percent. This type of system has been used successfully in
primary copper production.1
6.1.2 Building Enclosure/Evacuation
Enclosing and ventilating an entire building may be the only feasible control method
when the process operation is characterized by a number of small fugitive emissions sources.
A typical building evacuation system might consist of opposing wall-mounted ventilators that
force air across process equipment and out through an overhead plenum to a fabric filter.5 In
order to limit worker exposure to emissions and expel the heat generated by process
operations, large airflow rates are required. Thus, operational costs for this type of system
can be prohibitive. In addition, the need to keep the building enclosed during operation of the
ventilation system may be too restrictive on process operations, such as the movement of
forklifts and other equipment into and out of the building.1
6.2 OPTIMIZATION OF EQUIPMENT AND OPERATION
Optimization of equipment and operation includes: 1) limiting the amount of dust
available for emissions; 2) improving the arrangement of materials that generate dust;
3) optimizing the process so that less dusty material is used, generated, or made vulnerable to
air contact; 4) preventing or minimizing leaks; and 5) other good O&M procedures that reduce
PM emissions. In some industries, optimizing equipment and operation to reduce PM
emissions can also reduce operating costs if valuable products and/or raw materials can be
recovered and used.
6.2.1 Source Extent Reduction and Improvement
Source extent reduction measures are largely a function of good work practices and
include measures designed to reduce the volume and/or area of PM-generating materials
disturbed or reduce the frequency of disturbances and spills.1-7 These goals can generally be
achieved through good work practices and without a large investment in a control program.1
Examples of source extent reductions/improvements include:1
• Drop height reduction through the use of hinged-boom conveyors, rock ladders,
lower wells, etc.7 Table 6-2 lists estimated control efficiencies for
improvements through drop height reduction techniques;6
6-5
-------
Table 6-1. Estimated Control Efficiencies for Drop Height
Reduction Techniques (from Reference 6)
Control Efficiency
Technique (percent)
Lowering well or perforated pipe 80
Telescoping chute 75
Rock ladder 50
Hinged-boom conveyor 25
• Use of less dusty raw materials;
• Choke-feed or telescopic chutes to confine the material being transferred;7
• Increasing moisture retention in dusty areas;1
• Washing down or scraping conveyor belts regularly;1
• Performing PM-generating activities only as needed, e.g. in secondary lead
production, breaking of batteries only as needed to keep pace with the furnace;1
• Monitoring of feed materials to identify high PM-generating conditions;1
• Use of clean scrap in metal-melting furnaces;7 and
• Removing crankcase oil prior to automobile salvage.7
6.2.2 Process Optimization/Modification
Process optimization and/or modification can be an effective preventive measure for
process fugitive emission control. Also included in this category is the optimization of the
primary PM control devices and their capture systems. Some general techniques are:
• Mass transfer frequency reduction,
• Improved operational efficiency, and
• Use and proper operation of point-of-generation dust collection devices.
Some process-specific optimization techniques are:
6-6
-------
Some process-specific optimization techniques are:
• Designing a sulfuric acid plant at a primary lead smelter with sufficient capacity
to preclude the creation of back pressure and excess venting of the sinter
machine.1
• Changing from a cupola to an electric arc furnace.7
• Changing from an (open) bucket elevator to more efficient (closed) pneumatic
conveyor.7
• Screening out undersized coke (< 1 inch) to reduce blast furnace fugitive
emissions in primary lead smelting.1
• Improving blast furnace combustion efficiency during primary lead smelting by
improving the furnace water cooling system.1
• Injecting molten sodium in primary lead smelting kettle dressing to form liquid
matte rather than dross.1
• Eliminating fugitive PM from transporting, pouring, and stirring molten lead by
the use of continuous kettle dressing rather than manual in primary lead
smelting (as is currently done in only foreign facilities).1
• Improving raw material quality, e.g. improve the quality of coke and sinter
concentrate used in primary lead production.:
• Cooling lead pots to reduce fume generation during kettle dressing in primary
lead production.1
• Pumping (primary) lead directly to dross kettles using an electromagnetic
pump.1
• Agglomerating blast furnace flue dust in an agglomerating furnace to reduce the
load on the baghouse to improve its performance. This process completely
eliminates handling of the dust and the associated fugitive emissions, and
eliminates fugitive emissions from flue dust storage piles.1
• Using permanent mold castings in gray iron foundries instead of green sand.
This is reported to reduce PM emissions by 99 percent.1
6-7
-------
• Pre-treating glass manufacturing raw materials to reduce the amount of fine
particles. Pretreatments include: presintering, briquetting, pelletizing, or liquid
alkali treatment.
• Replacing grease and oil lubricants (e.g. in glass manufacturing) with silicone
emulsions and water-soluble oils that eliminate the smoke generated from flash
vaporization of hydrocarbons from greases and oils that come into contact with
process materials.1
6.2.3 Leak Prevention and Detection and Other Good O&M Practices
Good O&M practices can help to reduce fugitive PM emissions significantly. A key
aspect of a good O&M program to reduce PM emissions is a formalized leak prevention and
detection program. Examples of items that may be included in a program are: 1) adequate
design and prompt repairs of exhaust hood leaks; 2) maintenance of door and window seals;1
and 3) repair and/or prevention of warpage of oven doors to maintain proper seal.7
Good housekeeping practices and/or prompt response to process upsets, accidents, and
spills are also key elements in the control of fugitive dust. This prevents the build-up of dusty
material that can be resuspended into localized drafts. Good housekeeping practices include
the following procedures:1
• Washing down of building interiors regularly,
• Wetting floors during high dust periods,
• Use of oil-based sweeping compounds,
• Wet-wiping drums after the packaging of products,
A central vacuum system may be cost-effective for especially dusty operations.1 A full-time
clean-up crew may be required for some facilities to regularly implement the above
procedures.1-7
The proper operation of equipment is a good industrial practice to prevent fugitive dust
emissions. One example that may be applicable to a number of industries, especially hi
metallurgy, is the operation of furnaces so that they are not overloaded to eliminate the
possibility of back pressure from the primary PM control system as well as "puffing" during
opening of the charging door.7 Employee incentive programs to limit fugitive dust emissions
also have been used successfully in some industrie's.1
6.3 COSTS OF HOODS
Chapter 10 of the OAQPS Cost Manual provides information on estimating costs for
circular canopy, rectangular canopy, push-pull, slide-draft, and back-draft (slotted) hoods.8
Hood costs are estimated by using parameters from Table 6.3 in the following equation:
6-8
-------
Ch = aAfb
(6.1)
where Ch = hood cost ($)
Af = hood inlet (face) area (ft2)
a,b = equation regression parameters
Table 6.2. Parameters for Hood Cost Equation (Reference 8)
Type of Hood
Canopy, circular
Canopy, rectangular
Push-pull
Side-draft
Back-draft, slotted0
Back-draft, slotted*1
Back-draft, slotted
Back-draft, slotted
Back-draft, slotted
Material3
FRP
FRP
FRP
FRP
PVC
PVC
PP
FRP
Galvanized Steel
Equation Material
a
123
294
595
476
303
789
645
928
688
a FRP = fiberglass reinforced plastic, PVC = polyvinyl
For slotted hoods, equation range indicates the range in
than the total face area.
c Hoods with two rows of slots and no dampers.
Hoods with four rows of slots and manual slot dampers
b
0.575
0.505
0.318
0.332
1.43
0.503
0.714
0.516
0.687
Equation Range
(Af, ft2) b
2-200
2-200
2-200
2-200
0.6-2.0
1. 1-2.1
1. 1-2.1
1. 1-2.1
0.5- 1.3
chloride, PP = polypropylene
the area of the slot openings, which is much less
6.4 FUGITIVE DUST CONTROL
For information on the control of fugitive dust, please refer to Fugitive Dust
Background Document and Technical Information Document for Best Available Control
Measures (EPA-450/2-92-004) and Control of Open Fugitive Dust Sources
(EPA-450/3-88-008).
6-9
-------
6.5 REFERENCES FOR SECTION 6
1. Estimating and Controlling Fugitive Lead Emissions from Industrial Sources
(EPA-452/R-96-006). U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina. May 1996.
2. Fugitive Dust Background Document and Technical information Document for Best
Available Control Measures (EPA-450/2-92-004). U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina. September 1992.
3. Compilation of Air Pollutant Emission Factors (AP-42). Volume I (Fifth Edition).
U.S. Environmental Protection Agency, Research Triangle Park, NC. January 1995.
4. Coleman, R., Jr. and R. Vandervort. Demonstration of Fugitive Emission Controls at
a Secondary Lead Smelter hi Lead-Zinc-Tin 1980, J.M. Cigan, T.S. Mackey, and
T.J. O'Keefe (eds.). Proceedings of TMS-AJJV1E World Symposium on Metallurgy
and Environmental Control hi Las Vegas, Nevada, February 24-28, 1980. 1981.
5. Smith, R.D., O.A. Kiehn, D.R. Wilburn, and R.C. Bowyer. Lead Reduction hi
Ambient Air: Technical Feasibility and Cost Analysis of Domestic Primary Lead
Smelters and Refineries. Bureau of Mines, U.S. Department of the Interior,
Washington, D.C. 1987.
6. Bonn, R., T. Cuscino, Jr., and C. Cowherd, Jr. Fugitive Emissions from Integrated
Iron and Steel Plants (EPA-600/2-78-050). U.S. Environmental Protection Agency,
Washington, DC. March 1978.
7. Control Techniques for Paniculate Emissions from Stationary Sources - Volume 1
(EPA-450/3-81-005a, NTIS PB83-127498). U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards. Research Triangle Park, North
Carolina. September 1982.
8. OAQPS Control Cost Manual (Fourth Edition, EPA 450/3-90-006). U.S.
Environmental Protection Agency, Office of Ah" Quality Planning and Standards,
Research Triangle Park, North Carolina. January 1990.
6-10
-------
7. EMERGING TECHNOLOGIES
This section discusses the technologies that are currently being developed to enhance
the operation and/or collection efficiency of particulate control devices. In many cases, the
increase in collection efficiency is achieved through increase in collection of fine particles.
The sections below present emerging technologies for fabric filters, ESP's, scrubbers, and
mechanical control devices. Control techniques that allow for simultaneous control of PM
along with other pollutants are discussed here, despite the fact the PM collection efficiencies of
the devices may not be substantially higher than that in traditional devices; these devices are
included here because of the large savings in efficiency that are offered by the methods.
Table 7-1 summarizes the technologies presented in this section.
7.1 EMERGING FABRIC FILTER TECHNOLOGIES
Emerging technologies for fabric filters include 1) ceramic filter elements and ceramic
fiber enhancement, 2) fine 1.1 dtex fibers, and 3) electrostatically stimulated fabric filtration
(ESFF). These technologies are discussed below.
7.1.1 Ceramics: Ceramic Filter Elements and Ceramic Fiber Enhancement
Ceramic filters have become available (Altair, Ltd., UK; Didier, GmbH, FRG) that
can be used for high temperature PM filtration applications. Ceramic material is formed into
stiff cylindrical filter elements, called "candles." The tubes are generally 1 to 1.5 meters in
length, with outside diameters of 60 mm and a wall thickness of 10 to 20 mm. One end of the
tube is closed, and the other is open. The open ends of the tubes are mounted either vertically
or horizontally on a tubesheet, as with fabric bags. Tubes are generally cleaned by pulse jets.1
Ceramic fibers have been successfully processed into synthetic yarns and woven into
fabric filter bag material by 3M Inc. These ceramic-enhanced bags, called Nextel®, are
capable of high efficiency filtration (> 99 percent) of gas streams at temperatures up to
1400°F. High temperature operation saves the expense of gas cooling, reduces maintenance
due to condensation of corrosive gases, allows for energy recovery, and allows for PM
removal from the hot gas before other catalyst processes are performed. The Electric Power
Research Institute (EPRI) is currently testing the performance of these bags.2
7.1.2 Fine 1.1 dtex Fibers
New fine fibers, of 1.1 dtex (textile density - grains fiber per 10,000 meters of fiber )
have been developed by Dupont, GmbH (FRG), for high-efficiency fabric filtration. The fine
fibers are available in Nomex® and Teflon® materials. The fine fibers are half the weight of
standard Nomex fibers. The advantage of the new fibers is that while two (1.1 dtex) fine
fibers will weigh the same as one standard (2.2 dtex) fiber, the fine fibers have 40 percent
7-1
-------
Table 7-1. Summary of Emerging PM Control Technologies
Device
Technology
Benefit
Fabric Filter
1) Ceramics
ceramic filter elements and ceramic
fiber bag enhancement
2) Fine 1.1 dtex Fibers
3) Electrostatically-Stimulated Fabric
Filtration (ESFF)
1) High temperature capabilities
2) Higher fine PM control efficiency
3) Higher PM control efficiency and lower pressure
drop
ESP
1) Sonic Horn Rappers
2) Cold-pipe Precharger ESP
3) SUPER ESP
4) Advanced Computer-Based Control
System
1) Better plate cleaning efficiency, lower capital
costs, lower energy requirements, less
maintenance, and less downtime
2) Higher PM control efficiency, especially for high
resistivity PM
3) Higher PM control efficiency, especially for high
resistivity PM; smaller ESP; and eliminates need
for flue gas conditioning
4) Higher PM control efficiency
(continued)
-------
Table 7-1. (continued)
Device
Technology
Benefit
Combination Devices
COHPAC Hybrid ESP/FF
Higher PM control efficiency, especially for PM10;
smaller size, and less downtime than an ESP
Scrubber
1) Annular Orifice Venturi Scrubber
2) Waterweb Mesh
1) Higher PM control efficiency, especially for
PM25
2) Higher PM control efficiency, less clogging
Mechanical Collector
Fugitive Dust
Core Separator
Higher PM control efficiency
1) High-Voltage PM Ionizer
2) Dry Fog
1) Higher PM control efficiency (^100 percent
control) for PM 2:0.005 /^m, as well as
2:99 percent SO2 and NOX, and ^95 percent
VOCs
2) 10 percent of the water requirements for
conventional spray techniques
Simultaneous Control
1) SOx-NOx-Rox Box (SNRB) Catalytic 1) Controls SOx and NOx as well as PM (Rox)
Fabric Filter
2) Catalyst-Coated Fabric Filters
2) Controls NOx and PM
-------
more filter surface area. This enables the construction of denser and less porous filter media.
Potential improvements in fabric filter operation include higher efficiency at the same weight
of media, or lighter media with the same or better efficiency. These fine fibers are suitable for
operation in pulse jet fabric filters.3
7.1.3 Electrostatically-Stimulated Fabric Filtration (ESFF)
Electrostatically-stimulated fabric filters (ESFF) have been developed that
reduce fabric filter pressure drop and significantly reduce particle penetration. Electrically-
charged particles have been found to form highly porous dust layers in fabric filters.4 One
type of ESFF involves the placement of discharge electrodes (wires) axially inside reverse air
filter bags with conductive fibers woven into the bags. This generates an electric field
between the wire and the surface of the bag. A second type of ESFF uses external placement
of discharge wires within an array of pulse-jet bags.17 Another variation of ESFF includes the
placement of a pulse-jet ESFF module within an existing ESP, discussed hi section 7.3.5
7.2 EMERGING ESP TECHNOLOGIES
Emerging technologies for ESP's include 1) sonic horn rappers, 2) cold pipe
precharger ESP, 3) alternating charging and short ESP collector sections (SUPER ESP), and
4) advanced computer-based ESP control systems. These technologies are discussed below.
7.2.1 Sonic Horn Rappers
Sonic horn rappers for ESP cleaning have been developed by Atlantic Electric, Ltd.
(UK) for application in the electric utility industry to improve cleaning of ESP plates. The
sonic horns, used with magnetic impulse gravity impact (MIGI) rappers, were found to be
superior to tumbling hammers hi terms of lower capital costs, lower energy requirements, less
maintenance, and less downtime.6
7.2.2 Cold-Pipe ESP Precharger
A cold-pipe ESP precharger has been developed by Denver Research Institute. This
device circulates cool water through the pipe of a wire-pipe ESP precharger section. This
configuration reduces the resistivity of the dust layer collected on the outside surface of the
cold pipe and achieves a very high level of charge on the entrained dust particles in a very
short flow distance. Retrofit of existing ESPs with cold-pipe prechargers showed that the
effect of the addition of a cold pipe section exceeded the negative effects of back corona, and
significantly unproved the collection efficiency with high resistivity dusts.7-8
One potential application of this technology entails placing a cold-pipe precharger in-
line downstream from a conventional ESP that has at least one mam electrical section. The
cold-pipe precharger can have collection pipes substantially shorter than hi the main ESP, and
7-4
-------
can be operated with a current density 75 percent or lower than pipes hi the main ESP
section.9 Another application involves placing a cold-pipe precharger hi front of each
collection section of a wire-plate ESP. This second application is referred to as a "multistage"
ESP.17 A new concept of multi-stage ESP uses smaller sections preceded by cold-pipe
prechargers, discussed hi the next section.
7.2.3 Alternating Charging and Short ESP Collector Sections (SUPER ESP)
By separating the charging and collecting sections and optimizing each separately, a
new concept called SUPER ESP has evolved. Instead of long, wire-plate ESP sections, the
SUPER ESP uses a wire-pipe (cold-pipe) precharger previously described followed by an
abbreviated wire-plate collector section.10 Research has shown the collection efficiency of a
SUPER ESP section of 3-4 wires to be equivalent to a conventional ESP section of 7-8 wires
for lower resistivity dust, and a 12-14 wire section for high resistivity dust. Therefore a new
ESP using this concept may be only one-half to one-fourth the length of a conventional ESP;
or an existing ESP can be upgraded by replacing one section with two or more SUPER ESP
sections hi the same space.11
7.2.4 Advanced Computer-Based ESP Control Systems
The operation of an ESP can be optimized with the use of a computerized control
systems. The computer can be programmed to monitor and control the ESP parameters that
influence efficiency, and do so hi a way that exceeds the capability of manual techniques. For
example, the computer can identify the existence of back corona and then change the ESP
current or voltage settings so as to avoid or minimize the influence of the back corona.
Another example of advanced computer-based ESP control is a method used by The Mitsubishi
Company in Japan, where computer-controlled intermittent energization12 is accomplished
within an ESP, as needed to avoid back corona.17
7.3 EMERGING COMBINATION DEVICES
A combination fabric filter/ESP hybrid has been developed by EPRI and is called the
Compact Hybrid Paniculate Collector (COHPAC). This device involves using pulse jet fabric
filtration to capture PM that escape an ESP. The device has two designs called COHPAC I
and n. COHPAC I involves placing a (pulse jet) fabric filter downstream from an ESP.
COHPAC H utilizes a fabric filter in place of the last field(s) of an ESP.13-14-15-16 COHPAC
has been successfully tested hi pilot facilities at utility boilers.17 In a similar configuration to
COHPAC II, an EPA-developed innovation is to replace the last section of an ESP with pulse-
jet ESFF and maintain a high-voltage charging and collection field with the existing ESP
power supply. This results hi an order of magnitude lower emissions and significantly lower
pressure loss than for an uncharged fabric filter.5
7-5
-------
The COHPAC merges the advantages of an ESP and fabric filter. Advantages due to
the fabric filter component of the COHPAC are:17
• Less sensitive to changes in fuel composition, than an ESP, since a fabric filter
at steady state has constant outlet emissions.
• Better collection efficiency of PM10, as typical of fabric filtration.
• On-line maintenance can be utilized for the fabric filter section of the
COHPAC, resulting in less downtime.
Advantages due to an upstream ESP collector are:17
• The PM loading to the fabric filter is low enough to allow the fabric filter to
operate at a very high gas-to-cloth ratio (10-18 ft/min), without excessive
pressure drop or penetration.
• The collection efficiency of fabric filtration is enhanced because the PM has a
residual charge from the ESP.
7.4 EMERGING SCRUBBER TECHNOLOGIES
The emerging technologies for scrubbers include: 1) annular orifice venturi scrubber
and 2) waterweb mesh scrubber. These technologies are described below.
7.4.1 Annular Orifice Venturi Scrubber
The annular orifice venturi scrubber has been developed (Leisegang QTV-Process,
N.A.) that is an alkaline scrubbing process for the removal of PM and gaseous pollutants.
Flue gases are first cooled by injection of a lime liquor and then scrubbed in an annular orifice
venturi. The unit generates no wastewater and uses the heat from the flue gases to concentrate
residue into a paste with 50 percent solids. This device has been successfully used in MSW
incineration where, with a pressure drop of 16 in. water, 99 percent control efficiency of PM
* 2.5 fan. was achieved. At 24 in. water pressure drop, >99 percent control efficiency was
achieved for PM k 1.0 ^m and 97.5 percent for 0.5 /un PM.18
7.4.2 Waterweb Mesh Scrubber
A waterweb mesh scrubber has been developed by Mystaire Air Pollution Control
Systems, of Misonix Inc. of Farmingdale, New York, that is suitable for many types of PM,
gases, vapors, odors, and mists. Waterweb mesh is an extremely effective scrubber packing
that consists of sections of layered PVC-coated fiberglass that is compressed and bonded to
create thousands of microventuri passages where gas is forcibly dispersed through scrubbing
7-6
-------
liquid. The mesh is nonclogging and thoroughly mixes the gas and scrubbing liquid. This
scrubber is reported to be able to handle up to 50,000 CFM.19
7.5 EMERGING MECHANICAL COLLECTOR TECHNOLOGIES
An emerging technology for mechanical collectors is the "Core Separator" that has
been developed by LSR Technologies, Inc. for use with PM from coal combustion. The core
separator consists of a centrifugal separator and a conventional cyclone. The centrifugal
separator is a cylindrical chamber with a tangential inlet at the bottom and two outlets, which
generates a circular motion and centrifugal force. In the Core Separator, dust-laden gas enters
the centrifugal separator, where solid PM is forced to the wall outwards from the center of the
centrifugal separator leaving clean gas. This clean gas in the center of the separator is
exhausted to the atmosphere and the remaining dust-laden portion of the gas is sent to a
cyclone for additional PM removal. The cleaned gas from the cyclone is then recirculated
back to the centrifugal separator. The Core Separator system costs about three times more
than an equivalent-sized cyclone, but removes 80 percent of what usually penetrates the
conventional cyclone, to give 95 to 98 percent overall PM control.20-21
7.6 EMERGING FUGITIVE DUST CONTROL TECHNOLOGIES
Emerging technologies that may be applicable for PM control from fugitive PM
sources are 1) high-voltage PM ionizer and 2) dry fog. These technologies are discussed
below.
7.6.1 High-Voltage PM Ionizer
A technique that utilizes high-voltage electrical pulses to control air pollutants has been
developed by two companies. One device, called the Ion Blast, has been developed by Ion
Blast, Inc. of Vantaa, Finland. In the Ion Blast, gas is blown into a chamber where a 150
kilovolt (kV) current charges the PM in the air. The PM is then attracted towards a collection
surface where they become attached. The device is reported to have low energy requirements
and no moving, wearing, or replaceable parts. The Ion Blast is reported to be able to remove
nearly 100 percent of the PM, including biological matter such as viruses, as small as 0.005
fj.m. Only small quantities of gas, up to 5,000 SCFM, can be treated.22
A similar device, called the Pulsatech, has been developed by Pulsatron Technology,
Ltd. of Los Angeles, California. This technology was acquired from a Russian government
agency, which had implemented the device at four Russian industrial plants. The Pulsatech is
reported to achieve 99 percent destruction of SO2 and NOX and > 95 percent destruction of
VOCs, while ionizing the PM so that it collects on the chamber walls. The device uses a
22 kV charge and has a capacity of less than 1,500 SCFM.23
7-7
-------
7.6.2 Dry Fog
A device to control fugitive coal dust emissions from transfer points has been
developed and patented by Sonic Development Corp., Parsippany, New Jersey. In the "Dry
Fog" device, a fog of micrometer-size water droplets is generated by nozzles. The water
moistens the coal and triggers agglomeration of the dust particles to sizes large enough to
settle. Once the dust drops reach the ground, the water evaporates leaving dry coal. A unit
installed near the power plant of a grain processor uses water at a rate of 1 gal/mm, about
10 percent of that used hi conventional wetting techniques.24
7.7 EMERGING SIMULTANEOUS POLLUTION CONTROL TECHNOLOGIES
Emerging technologies that target the simultaneous control of PM and other pollutants
include: 1) SOx-NOx-Rox Box (SNRB) catalytic fabric filter, and 2) catalyst-coated fabric
filters. These technologies are discussed below.
7.7.1 SNRB (SOx-NOx-Rox Box) Catalytic Fabric Filter
A SOx-NOx-Rox Box (SNRB) has been developed by Babcock & Wilcox (Alliance,
Ohio) for use hi coal combustion, that controls PM ("Rox") as well as SOX and NOX. The
device consists of a pulse-jet fabric filter equipped with ceramic fiber filter bags for high
temperature application, alkali (sodium or calcium based) sorbent injection used for SO2
removal, and ammonia injection and selective catalytic reduction (SCR) used for catalytic
reduction and control of NOX. The catalyst is located inside the filter bags (see Section 7.7.2
below). The alkali sorbent and ammonia injection can be performed at the high temperatures
of the fabric filter. Consolidating the removal of three pollutants into one device saves on
capital and operating costs, and the need for flue gas cooling is eliminated. Efficiencies
reported are 99.9 percent for PM, 85 percent for SOX, and 90 to 95 percent for NOX.25-26
7.7.2 Catalyst-Coated Fabric Filters
Catalyst-coated fabric filters have been developed by the Energy & Environmental
Research Center of Grand Forks, North Dakota, and Owens Corning Fiberglass Corporation
use with combustion sources. Using a sol-gel process, high temperature fabric filters are
coated with a vanadium/titanium (V/Ti) catalyst. The catalyst-coated filters can then be used
hi utility or industrial hot-side fabric filter and are capable of simultaneous PM and NOX
control.27
7-8
-------
7.8 REFERENCES FOR SECTION 7
1. Butcher, C. Hot News in Ceramic Filters. Chemical Engineer. October 10, 1991.
2. Fischer, E.M. "Performance of Ceramic Fiber and Composite Filters at High
Temperatures," hi Clean Air Technology News, Institute of Clean Air Companies,
Washington, DC. Summer 1995.
3. Dilger, F. Lowering Emissions - More Effective Air Pollution Control with Fine Fibre
Filter Media. Filtration & Separation. March/April 1994.
4. Plaks, N. and B.E. Daniel. "Advances in Electrostatically Stimulated Fabric
Filtration," in Proceedings of the Seventh Symposium on the Transfer and Utilization of
Paniculate Control Technology (EPA-600/9-89-046b). U.S. Environmental Protection
Agency, Washington, DC. May 1989.
5. Plaks, N. and C.B. Sedman. Enhancement of Electrostatic Precipitation with
Electrostatically Augmented Fabric Filtration. U.S. Patent No. 5,217,511. June 8,
1993.
6. ESP Newsletter. The Mcllvaine Company, Northbrook, Illinois. March 1996.
7. Yamamoto, T., P. A. Lawless, and N. Plaks. Evaluation of the Cold Pipe Precharger.
IEEE Transactions on Industry Applications. 26(4): 639-645. July/August 1990.
8. Rinard, et al. "Development of a Charging Device for High-resistivity Dust Using
Heated and Cooled Electrodes," hi Proceedings of the Third Symposium on the
Transfer and Utilization of Paniculate Control Technology, Vol. II
(EPA-600/9-2-005b). U.S. Environmental Protection Agency, Washington, DC. 1982
9. Mosley, R.B., L.E. Sparks, and N. Plaks. Electroprecipitator with Suppression of
Rapping Reentrainment. U.S. Patent No. 4,822,381. April 18, 1989.
10. N. Plaks and L.E. Sparks. Electroprecipitator with Alternating Charging and Short
Collector Sections (SUPER ESP). U.S. Patent No. 5,059,219. October 22, 1991.
11. Plaks, N. "The SUPER ESP - Ultimate Electrostatic Precipitation," hi Proceedings:
1991 Symposium on the Transfer and Utilization of Paniculate Control Technology.
Electric Power Research Institute, Palo Alto, California. 1992.
12. Oglesby, S., Jr. Future Directions of Paniculate Control Technology: A Perspective.
J. Air Waste Management Assoc. 40(8): 1184-1185. August 1990.
13. Fabric Filter Newsletter. The Mcllvaine Company, Northbrook, Illinois. April 1996.
7-9
-------
14. Miller, R.L. "Combining ESPs and Fabric Filters for Paniculate Collection," in Clean
Air Technology News, Institute ofClean Air Companies, Washington, DC. Winter
1994.
15. Lamarre, L. COHPing with Particulates. EPRI Journal. July/August 1993.
16. Chang, R. Compact Hybrid Paniculate Collector (COHPAC). U.S. Patent
No. 5,158,580. October 27, 1992.
17. C. Sedman, N. Plaks, W. Merchant, and G. Nichols. Advances in Fine Particle
Control Technology. Presented at the Ukraine Ministry of Energy and Electrification
Conference on Power Plant Air Pollution Control Technology, in Kiev, The Ukraine,
September 9 -10, 1996.
18. The Scrubber Manual. The Mcllvaine Company, Northbrook, Illinois.
February 1987.
19. Personal communication. C. Thomas, Misonix Inc., Farmingdale, New York, with T.
Stobert, EC/R Inc., Durham, North Carolina. Mystaire® Scrubbing Systems. August
6, 1996.
20. ESP Newsletter. The Mcllvaine Company, Northbrook, Illinois. May 1996.
21. Wysk, S.R. and L.A. Smolensky. Novel Paniculate Control Device for Industrial Gas
Cleaning. Filtration & Separation. January/February 1993.
22. ESP Newsletter. The Mcllvaine Company, Northbrook, Illinois. April 1996.
23. High-Power Pulses Blast Pollutants. Chemical Engineer ing. 103(9): 21-22.
September 1996.
24. Reason, J. Dust Suppression System Doesn't Wet Coal. Power. March 1989.
25. Kudlac, G.A., et. al. SNRB Catalytic Baghouse Laboratory Pilot Testing.
Environmental Progress. 11:1. February 1992.
26. Gennrich T. Filter Bags Help Meet Paniculate Control Standards. Power
Engineering. August 1993.
27. Ness, S.R., et al. SCR Catalyst-Coated Fabric Filters for Simultaneous NOX and High
Temperature Paniculate Control. Environmental Progress. 14:1. February 1995.
7-10
-------
APPENDIX A
LIST OF RESOURCE DOCUMENTS FOR PM AND PM PRECURSOR CONTROL
The following is a list of documents that can be use as resources to identify control
techniques for paniculate matter (PM10 and PM2 5), as well as the PM precursors: sulfur
oxides, ammonia, nitrogen oxides, and volatile organic compounds.
PARTICIPATE MATTER
Control Techniques for Paniculate Emissions from Stationary Sources - Volume 1 (EPA-
450/3-8l-005a, PB83-127498). U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina. 1982. (change to 1996 update, when completed)
Guidance Document for Residential Wood Combustion Emission Control Measures
(EPA-450/2-89-015). U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina. September 1989.
Technical Information Document for Residential Wood Combustion Best Available Control
Measures (EPA-450/2-92-002). U.S. Environmental Protection Agency, Research Triangle
Park, North Carolina. September 1992.
Control of Open Fugitive Dust Sources (EPA-450/3-88-008). U.S. Environmental Protection
Agency, Research Triangle Park, North Carolina. September 1988.
Fugitive Dust Background Document and Technical Information Document for Best Available
Control Measures (EPA-450/2-92-004). U.S. Environmental Protection Agency, Research
Triangle Park, North Carolina. September 1992.
Agricultural Activities Influencing Fine Particulate Matter Emissions. Draft Final Report
prepared for Mr. K. Woodard, U.S. Environmental Protection Agency, Air Quality Strategies
and Standards Division, Research Triangle Park, North Carolina, under EPA Contract No. 68-
D3-0031, Work Assignment H-19. March 25, 1996.
SULFUR OXIDES
Control Techniques for Sulfur Oxide Emissions from Stationary Sources, Second Edition
(EPA-450/3-81-004). U.S. Environmental Protection Agency, Research Triangle Park, North
Carolina. 1981.
A-l
-------
AMMONIA
ApSimon, H.M., D. Cowell, and S. Couling. Assessing the Potential for Abatement of
Ammonia Emissions from Agriculture in Europe: the MARACCAS Model (Draft).
International Conference on Atmospheric Ammonia Emissions, Deposition and Environmental
Impacts, NETCEN, Culham, Oxford, England. October 2-4, 1995.
NITROGEN OXIDES
Meeting the 15-Percent Rate-of-Progress Requirement Under the Clean Air Act: A Menu of
Options. STAPPA/ALAPCO, Washington, D.C. September 1993.
Controlling Nitrogen Oxides Under Clean Air Act. STAPPA/ALAPCO and ESI International,
Washington, D.C. July 1994.
VOLATILE ORGANIC COMPOUNDS
Control of Volatile Organic Emissions from Existing Stationary Sources - Volume I: Control
Methods for Surface Coating Operations. EPA-450/2-76-028 (NTIS). U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park,
North Carolina. November 1976.
Control of Volatile Organic Emissions from Existing Stationary Sources - Volume II: Surface
Coating of Cans, Coils, Paper, Fabrics, Automobiles, and Light-Duty Trucks. EPA-450/2-77-
008. U.S. Environmental Protection Agency, Office of Air Quality Planning and standards,
Research Triangle Park, North Carolina. May 1977.
Control of Volatile Organic Emissions from Existing Stationary Sources - Volume ffl: Surface
Coating of Metal Furniture. EPA-450-2-77-032. U.S. Environmental Protection Agency,
Office of Ak Quality Planning and Standards, Research Triangle Park, North Carolina.
December 1977.
Control of Volatile Organic Emissions from Existing Stationary Sources - Volume TV:
Surface Coating for Insulation of Magnet Wire. EPA-450/2-77-033. U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park,
North Carolina. December 1977.
Control of Volatile Organic Emissions from Existing Stationary Sources - Volume V: Surface
Coating of Large Appliances. EPA-450/2-77-034. U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
December 1977.
A-2
-------
Control of Volatile Organic Emissions from Bulk Gasoline Plants. EPA-450/2-77-035. U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, Research
Triangle Park, North Carolina. December 1977.
Control of Volatile Organic Emissions from Storage of Petroleum Liquids in Fixed-Roof
Tanks. EPA-450/2-77-036. U.S. Environmental Protection Agency, Office of Air and
Quality Planning and Standards, Research Triangle Park, North Carolina. December 1977.
Control of Refinery Vacuum Producing Systems, Wastewater Separators, and Process Unit
Turnarounds. EPA-450/2-77-025. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Research Triangle Park, North Carolina. October 1977.
Control of Volatile Organic Compounds from Use of Cutback Asphalt. EPA^50/2-77-037.
U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards,
Research Triangle Park, North Carolina. December 1977.
Control of Hydrocarbons from Tank Truck Gasoline Loading Terminals. EPA-450/2-77-026.
U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards,
Research Triangle Park, North Carolina. October 1977.
Design Criteria for Stage I Vapor Control Systems - Gasoline Service Stations, (no document
number issued). U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Research Triangle Park, North Carolina. November 1975.
Control of Volatile Organic Emissions from Solvent Metal Cleaning. EPA-450/2-77-022.
U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards,
Research Triangle Park, North Carolina. November 1977.
Summary of Group I Control Technique Guideline Documents for Control of Volatile Organic
Emissions from Existing Stationary Sources. EPA-450/3-78-120. U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park,
North Carolina. December 1978.
Control of Volatile Organic Emissions from Existing Stationary Sources - Volume VI:
Surface Coating of Miscellaneous Metal Pails and Products. EPA-450/2-78-015.
U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards,
Research Triangle Park, North Carolina. June 1978.
Control of Volatile Organic Emissions from Existing Stationary Sources - Volume VJJ:
Factory Surface Coating of Flat Wood Paneling. EPA-450/2-78-032. U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park,
North Carolina. June 1978.
A-3
-------
Control of Volatile Organic Emissions from Existing Stationary Sources - Volume VIQ:
Graphic Arts - Rotogravure and Flexography. EPA-450/2-78-033. U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park,
North Carolina. December 1978.
Control of Volatile Organic Compound Leaks from Petroleum Refinery Equipment. EPA-
450/2-78-036. U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Research Triangle Park, North Carolina. June 1978.
Control of Volatile Organic Emissions from Petroleum Liquid Storage hi External Floating
Roof Tanks. EPA-450/2-78-047. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Research Triangle Park, North Carolina. December 1978.
Control of Volatile Organic Compound Leaks from Gasoline Tank Trucks and Vapor
Collection Systems. EPA-450/2-78-051. U.S. Environmental Protection Agency, Office of
Air Quality Planning and Standards, Research Triangle Park, North Carolina. December
1978.
Control of Volatile Organic Emissions from Manufacture of Synthesized Pharmaceutical
Products. EPA-450/2-78-029. U.S. Environmental Protection Agency, Office of Air Quality
Planning and Standards, Research Triangle Park, North Carolina. December 1978.
Control of Volatile Organic Emissions from Manufacture of Pneumatic Rubber Tires. EPA-
450/2-78-030. U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Research Triangle Park, North Carolina. December 1978.
Summary of Group n Control Technique Guideline Documents for Control of Volatile
Organic Emissions from Existing Stationary Sources. EPA-450/2-80-001. U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, Research
Triangle Park, North Carolina. December 1979.
Control of Volatile Organic Compound Emissions from Large Petroleum Dry Cleaners.
EPA-450/3-82-009. U.S. Environmental Protection Agency, Office of Ah- Quality Planning
and Standards, Research Triangle Park, North Carolina. September 1982.
Control of Volatile Organic Compound Emissions from Manufacture of High-Density
Polyethylene, Polypropylene, and Polystyrene Resins. EPA-450/3-83-008. U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, Research
Triangle Park, North Carolina. November 1983.
Control of Volatile Organic Compound Equipment Leaks from Natural Gas/Gasoline
Processing Plants. EPA-450/2-83-007. U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards, Research Triangle Park, North Carolina. December 1983.
A-4
-------
Control of VOC Fugitive Emissions from Synthetic Organic Chemical, Polymer, and Resin
Manufacturing Equipment. EPA-450/3-83-006. U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina. March
1984.
Control of Volatile Organic Compound Emissions from Air Oxidation Processes in Synthetic
Organic Chemical Manufacturing Industry, EPA-450/3-84-015. U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Research Triangle Park,
North Carolina. December 1984.
Control of Volatile Organic Compound Emissions from Reactor Processes and Distillation
Operations Processes in the Synthetic Organic Chemical Manufacturing Industry, draft. U.S.
Environmental Protection Agency, Office of Air Quality Planning and Standards, Research
Triangle Park, North Carolina. November 15, 1993.
Control of Volatile Organic Compound Emissions from the Application of Agricultural
Pesticides. EPA-453/R-92-011. U.S. Environmental Protection Agency, Office of Air Quality
Planning and standards, Research Triangle Park, North Carolina. 1993.
Reduction of Volatile Organic Compound Emissions from Automobile Refinishing. EPA-
450/3-88-009. U.S. Environmental Protection Agency, Office of Air Quality Planning and
Standards, Research Triangle Park, North Carolina. October 1998.
Halogenated Solvent Cleaners. EPA-450/3-89-030. U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
August 1989.
Organic Waste Process Vents. EPA-450/3-91-007. U.S. Environmental Protection Agency,
Office of Air Quality Planning and Standards, Research Triangle Park, North Carolina.
December 1990.
Reduction of Volatile Organic Compound Emissions from Application of Traffic Markings.
EPA-450/3-88-007. U.S. Environmental Protection Agency, Office of Air Quality Planning
and Standards, Research Triangle Park, North Carolina. August 1988.
VOC/HAP Emissions from Marine Vessel Loading Operations: Technical Support Document
for Proposed Standards. EPA-450/3-93-001a. U.S. Environmental Protection Agency, Office
of Air Quality Planning and Standards, Research Triangle Park, North Carolina. May 1992.
A-5
-------
APPENDIX B
VATAVUK AIR POLLUTION CONTROL COST INDEXES
The Vatavuk Air Pollution Control Cost Indexes (VAPPCI) are updated quarterly and
published monthly in Chemical Engineering magazine. For a detailed explanation of the
development and use of the VAPPCI, see Chemical Engineering, December 1995, pp 88-95.
Vatavuk Air Pollution Control Cost Indexes
(1st Quarter 1994=100.0)"
Control Device
Carbon adsorbers
Catalytic incinerators
Electrostatic precipitators
Fabric filters0
Flares
Gas absorbers
Mechanical collectors0
Refrigeration systems
Regenerative thermal oxidizers
Thermal incinerators
Wet scrubbers
1995
(Avg.)
110.7
107.1
108.2
102.7
107.5
105.6
103.0
103.0
104.4
105.9
112.5
IstQ.
1996
109.2
107.7
107.0
104.0
104.5
108.6
103.3
104.2
105.8
108.0
111.7
2ndQ.
1996
107.5
107.0
107.6
104.2
104.9
108.2
103.3
104.2
106.0
108.0
110.1
3rdQ.
1996
105.2
107.1
108.9
104.8
105.1
107.1
103.3
104.4
106.7
108.3
109.3
4th Q.
1996b
103.9
105.8
108.3
105.0
105.7
106.9
103.3
104.8
106.5
108.2
109.0
IstQ.
1997"
104.2
105.6
108.8
105.3
105.5
108.7
103.5
105.1
107.3
109.1
108.3
Index values have been rounded to the nearest tenth.
All fourth quarter 1996 and first quarter 1997 indexes are preliminary.
For fabric filters and mechanical collectors, each quarterly value shown is the average of the Producer
Price Indexes (PPIs) for the three months in question, divided by the average of the PPIs for January,
February, and March 1994 (i.e., first quarter 1994)
B-l
-------
TECHNICAL REPORT DATA
(Please read Instructions on reverse before completing)
1. REPORT NO.
EPA-452/R-97-001
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Stationary Source Control Techniques Document for Fine
Particulate Matter
5. REPORT DATE
October 1998
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
Office of Air Quality Planning and Standards
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
11. CONTRACT/GRANT NO.
68-D-98-026
12. SPONSORING AGENCY NAME AND ADDRESS
Director
Office of Air Quality Planning and Standards
Office of Air and Radiation
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/200/04
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Stationary Source Control Techniques Document for Fine Paniculate Matter presents recent
developments in the control of particulate matter which have become available since preparation of
an earlier document entitled Control Techniques for Particulate Emissions from Stationary Sources
- Volume 1 (1982).
This document focuses on fine particulate matter (PM10 and PM2.5 - particles with an
aerodynamic diameter less than or equal to an nominal 10 microns and 2.5 microns, respectively).
Information presented hi this document includes background on particulate matter emissions;
measurement methods for particulate matter; types of particulate control devices, their operating
principles, design, operation, and control efficiencies; costs and environmental effects of particulate
matter control systems; and emerging technologies for particulate matter control.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
c. COSAT1 Field/Group
Air Pollution
Pollution Control
Fine Particulate Matter
Costs
Air Pollution Control
Fine Particulate Matter
18. DISTRIBUTION STATEMENT
Release Unlimited
;19. SECURITY CLASS (Report)
Unclassified
21. NO. OF PAGES
242
20. SECURITY CLASS (Page)
Unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION IS OBSOLETE
------- |