Control Techniques
  for Particulate Emissions
 from Stationary  Sources -
             Volume  1
       Emission Standards and Engineering Division
           Office of Air, Noise, and Radiation
        Office of Air Quality Planning and Standards
       Research Triangle Park, North Carolina 27711

                September 1982
For sole by the Superintendent of Documents, 0.3. Government Printing Office, Woshtagton, D.C. 2M02

This report has been reviewed by the Emission Standards and Engineering
Division of the Office of Air Quality Planning and Standards, EPA, and
approved for publication';  Mention of trade names or commercial products
is not intended to constitute endorsement or recommendation for use.
Copies of this report are for sale fay the Superintendent of Documents,
U.S. Government Printing Office, Washington, D.C. 20402, and the National
Technical Information Services, 5285 Port Royal Road, Springfield,
Virginia 22161.


                                  VOLUME 1

Contents of Volume 2                                                  vii
List of Figures                                                       ix
List of Tables                                                        xix
Symbols                                       ~                        xxi
Conversion Factors                                                    xxiii
Glossary                                                              xxv

1.   INTRODUCTION                                                     1-1

2.   BACKGROUND                                                       2-1

     2.1  Trends and Projections in Particulate Emissions             2-1

          2.1.1  Air quality and particulate matter emission trends   2-1
          2.1.2  Projections for future control programs and
                   emissions                                          2-2

     2.2  Sources of Suspended Particulate Matter                     2-3

          2.2.1  Natural emission sources                             2-4
          2.2.2  Manmade sources of particulate                       2-4
          2.2.3  Transported particulate                              2-9

     2.3  Measurement of Particulate from Stationary Point Sources    2-11

          2.3.1  Mass concentration measurement                       2-11
          2.3.2  Particle size analysis                               2-16
          2.3.3  Analysis of particulate samples                      2-19
          2.3.4  Chemical analysis and analysis of trace elements     2-20


     3.1  Energy Source and Fuel Selection                            3-1

     3.2  Process Optimization                                        3-3

          3.2.1  Modification of process feed materials               3-3
          3.2.2  Elimination of process steps                         3-4
          3.2.3  Changes in process particle characteristics          3-5

     3.3  Exhaust Gas Cleaning              •                          3-5

          3.3.1  Applicable regulations                               3-6
          3.3.2  Source characteristics                               3-6

                                    ' i i i

          3.3.3  Control device design limitations                    3-7
          3.3.4  Control device reliability                           3-8
          3.3.5  Control device costs and financial  assistance        3-8

4.   PARTICULATE CONTROL SYSTEMS                                      4.1-1

     4.1  Introduction                                                4.1-1
          4.1.1  Particle characteristics and behavior                4.1-1
          4.1.2  Selection and application of particle
                 control devices                                      4.1-11
          4.1.3  Control system design                                4.1-11

     4.2  Mechanical Collectors                                       4.2-1

          4.2.1  Types of mechanical collectors                       4.2-1
          4.2.2  Operating principles of mechanical  collectors        4.2-13
          4.2.3  Design of mechanical collectors                      4.2-25
          4.2.4  Operation and maintenance of mechanical  collectors   4.2-29

     4.3  Electrostatic Precipitators                                 4.3-1

          4.3.1  Types of electrostatic precipitators                 4.3-1
          4.3.2  Operating principles of electrostatic precipitators  4.3-8
          4.3.3  Design of electrostatic precipitators                4.3-23
          4.3.4  Operation and maintenance of electrostatic
                    precipitators                                     4.3-58

     4.4  Fabric Filter                                               4.4-1

          4.4.1  Types of fabric filters                              4.4-1
          4.4.2  Operating principles of fabric filters               4.4-9
          4.4.3  Design of fabric filters                             4.4-16
          4.4.4  Operation and maintenance of fabric filters          4.4-36

     4.5  Wet Scrubbers                                               4.5-1

          4.5.1  Types of participate scrubbers                       4.5-1
          4.5.2  Operating principles of particulate scrubbers        4.5-11
          4.5.3  Design of particulate scrubbers                      4.5-35
          4.5.4  Operation and maintenance of particulate scrubbers   4.5-44

     4.6  Incinerators                                                4.6-1

          4.6.1  Types of incinerators                                4.6-1
          4.6.2  Operating principles of incinerators                 4.6-2
          4.6.3  Design of incinerators                               4.6-10
          4.6.4  Operation and maintenance of incinerators            4.6-20

5.   FUGITIVE EMISSION CONTROL                                        5-1

     5.1  Sources of Fugitive Emissions                               5-1

     5.2  Control of Industrial Process Fugitive Emissions            5-2


          5.2.1  Ventilation systems           •                       5-6
          5.2.2  Optimization of equipment and operation              5-12
          5.2.3  Wet suppression                                      5-13

     5.3  Control of Fugitive Dust                                    5-16

          5.3.1  Wet suppression                                      5-16
          5.3.2  Stabilization                                        5-17
          5.3.3  Specialized fugitive emission control  techniques     5-19

6.    ENERGY AND ENVIRONMENTAL CONSIDERATIONS                          6-1

     6.1  Energy Requirements                                         6-1

          6.1.1  Fan energy requirements                              6-2
          6.1.2  Control device energy requirements                   6-4
          6.1.3  Hopper heaters and vibrators                         6-13
          6.1.4  Solids discharge and transport                       6-13
          6.1.5  Ultimate disposal                                    6-13
          6.1.6  Other considerations                                 6-14

     6.2  Secondary Pollutant Generation                              6-14

          6.2.1  Electrostatic precipitators                          6-14
          6.2.2  Incinerators                                         6-15

     6.3  Liquid Waste Management                                     6-15

          6.3.1  Regulatory requirements                              6-15
          6.3.2  Control techniques                                   6-16

     6.4  Solid Waste Management                                      6-21

          6.4.1  Regulatory requirements                              6-22
          6.4.2  Waste recycle                                        6-22
          6.4.3  Waste disposal                                       6-24

     6.5  Noise Management                                            6-25

     6.6  Radiation Control                                           6-25

     CONTROL TECHNIQUES                                               7-1

     7.1  Particulate Control Equipment Cost Analysis                 7-2

          7.1.1  Capital costs                                        7-2
          7.1.2  Annualized costs                                     7-4
          7.1.3  Other cost considerations                            7-4

     7.2  Methodology for Analyzing Cost of Particulate Control
            Systems                                                   7-5

          7,2.1  Capital costs                                        7-5
          7.2.2  Annualized costs                                     7-8

     7.3  Cost Curves for Various Particulate Control  Systems         7-10

          7.3.1  Equipment costs                                      7-11
          7.3.2  Particulate control system costs                     7-11

     7.4  Cost of Fugitive Emission Control                           7-28

8.   EMERGING TECHNOLOGIES                                           , 8-1

     8.1  Advanced Scrubbing Techniques                               8-1

          8.1.1  Air Pollution Systems, Inc.  electrostatic
                    precipitator                                      8-2
          8.1.2  TRW charged droplet scrubber                         8-5
          8.1.3  University of Washington electrostatic droplet
                   scrubber                                           8-5
          8.1.4  Steam hydro scrubber                                 8-11
          8.1.5  Two-phase jet scrubber                               8-11
          8.1.6  Flux force/condensation scrubbing                    8-15

     8.2  Advanced Electrostatic Precipitation Techniques             8-15

          8.2.1  Pulse energization                                   8-17
          8.2.2  Two-stage ESP precharging                            8-21
          8.2.3  Flue gas conditioning for EPS's                      8-23
          8.2.4  Development of high temperature/high  pressure
                   electrostatic precipitation                        8-27

     8.3  Advanced Filtration Techniques                              8-28

          8.3.1  Electrostatically augmented fabric filtration        8-29
          8.3.2  Electrostatically augmented filtration through
                   fiber beds                                         8-29
          8.3.3  Granular bed filtration                              8-32
          8.3.4  Barrier filtration                                   8-33

     8.4  High-Gradient Magnetic Separation                           8-34

     8.5  Agglomeration Techniques                                    8-38

          8.5.1  Sonic agglomeration                                  8-38
          8.5.2  Magnetic agglomeration                               8-43

                            SUMMARY TABLE OF CONTENTS

                                  VOLUME II



     9.1  Stationary Source Selection                                 9.1-1
     9.2  Stationary Combustion Sources                               9.2-1
     9.3  Refuse Incinerators                                         9.3-1
     9.4  Open Burning                                                9.4-1
     9.5  Chemical Process Industry                                   9.5-1
     9.6  Food and Agricultural Industry                              9.6-1
     9.7  Mineral Products                                            9.7-1
     9.8  Metallurgical Industry                                      9.8-1
     9.9  Petroleum Industry                                          9.9-1
     9.10 Forest Products Industry                                    9.10-1
     9.11 Lead-Acid Battery Manufacturing                             9.11-1
     9.12 Fugitive Dust Sources                                       9.12-1


                               LIST OF FIGURES
Number                                                                Page
2-1       Estimated average contributions to nonurban TSP levels      2-5
2-2       EPA Method 5 particulate sampling apparatus                 2-13
2-3       S-type pi tot tube and manometer                             2-15
4.1-1     Aerosol distribution                                        4.1-3
4.1-2     Histogram of a lognormal size distribution                  4.1-4
4.1-3     Cumulative lognormal size distribution                      4.1-5
4.1-4     Bi-modal aerosol distribution                               4.1-5
4.1-5     Impact!on of particles on a target in a moving gas
            stream                                                    4.1-6
4.1-6     Interception of a particle on a target in a moving
            gas stream                                •                4.1-7
4.1-7     Diffusion of a particle to a target in a moving gas
            stream                                                    4.1-8
4.2-1     Howard multi-tray settling chamber                          4.2-2
4.2-2a    Simple momentum separator                                   4.2-3
4.2-2b    Simple momentum separator     .                              4.2-4
4,2-2c    Baffle-type momentum separator                              4.2-4
4.2-3     Louvered shutter type collector                             4.2-5
4.2-4     General types of cyclones                                   4.2-7
4.2-5     Typical simple cyclone                                      4.2-8
4.2-6     Flow pattern in a double vortex cyclone                     4.2-9
4.2-7a    Typical multi-cyclone collector                             4.2-11
4.2-7b    Individual tube from multi-cyclone collector                4.2-11

LIST OF FIGURES (continued)
Number                                                                Page
4.2-8a    Fixed impeller straight-through cyclone                     4.2-12
4.2-8b    Bank of fixed impeller straight-through cyclones with
            secondary cyclone dust collector                          4.2-12
4.2-9     Typical size efficiency curve for settling chamber          4.2-15
4.2-10    Penetration of dust through a settling chamber serving
            a sinter plant                                            4.2-16
4.2-11    Momentum separator                                          4.2-17
4.2-12    Penetration of fly ash through two momentum separators      4.2-17
4.2-13a   Types of mechanically aided separators                      4.2-18
4.2-13b   Penetration curves for mechanically aided separators        4.2-19
4,2-14    Penetration curve predicted by Leith and Licht approach     4.2-21
4.2-15    Cyclone penetration as a function of particle size
            ratio                                                     4.2-22
4.2-16    Penetration curves for multicyclone tubes of different
            diameter                                                  4.2-23
4.2-17    Penetration curve for double vortex cyclone                 4.2-24
4.2-18    Partial pluggage of multiple cyclone inlet vanes            4.2-32
4.3-1     Typical ESP with insulator compartments                     4.3-2
4.3-2     Three types of wet ESP's                                    4.3-7
4.3-3     Basic processes involved in electrostatic precipitation     4.3-10
4.3-4     Typical temperature-resistivity relationship                4.3-15
4.3-5     Penetration as a function of particle size for an ESP
            on a kraft pulp mill recovery boiler                      4.3-16
4.3-6     Comparison of experimental penetration as a function
            of particle diameter to the McDonald (1978) computer
            model under normal SCA conditions                         4.3-20
4.3-7     Penetration, pulverized-coal-fired boiler
            (cold-side ESP)                                           4.3-21

LIST OF FIGURES (continued)

Number                                                                Page

4.3-8     Computed versus actual penetration for cold-side
            ESP on a western subbituminous-fired boiler2              4.3-22

4.3-9     Predicted precipitator penetration for bark/fossil
            fuel-fired boilers                                        4.3-24

4.3-10a   Precipitator penetration versus specific collection area
            and precipitation rate w                                  4.3-26

4.3-lOb   Precipitator penetration as a function of specific
            collection area and modified precipitation rate
            parameter w,                                               4.3-26

4.3-11    Distribution of ash-to-Btu ratio and log (resistivity)
            for a single fuel field                                   4.3-27

4.3-12    Mechanical sectionalization of a precipitator               4.3-29

4.3-13    Typical fly ash precipitator voltage-current
            characteristics, five fields in series, no ash
            resistivity problem                                       4.3-32

4.3-14    ESP current wave form with and without si 1 iron controlled
            rectifiers                                                4.3-34

4.3-15    Time periods are shown as control system reacts to a
            spark impulse F after steady-state operation              4.3-35

4.3-16    Precipitator charging system and wire hanging system        4.3-38

4.3-17    Various combinations of electrical sectional ization in
            an ESP                                                    4.3-40

4.3-18    Vibrator and rapper assembly and precipitator high-
            voltage frame                                             4.3-43

4.3-19    Typical fly ash type pneumatic vacuum system                4.3-48

4.3-20    Effect of two different methods of gas distribution
            of flue characteristics in an ESP                         4.3-50

4.3-21    Examples of two inlet plenum designs that generally
            cause gas distribution problems                           4.3-51

4.3-22    Expansion inlet plenums showing two methods of spreading
            the gas patterns                ,                          4.3-51

4.3-23    Internal view of one type of rectifier console showing
            component parts                                           4.3-53


LIST OF FIGURES (continued)
4.3-24    Flow diagram of sulfur burning flue gas conditioning
            system                                                    4.3-57
4.3-25    Comparison of mean penetration results                      4.3-59
4.3-26    Precipitator log sheet                                      4.3-65
4.3-27    Typical operating curve to meet emission regulations
            with partial malfunctions of ESP                          4.3-76
4.4-1     Small shaker type baghouse                                  4.4-2
4.4-2     Reverse air baghouse                                        4.4-4
4.4-3     Continuous reverse air-cleaning system for flat
            filter sleeves                                            4.4-5
4.4-4     Reverse air collector                                       4.4-6
4.4-5     Pulse jet baghouse                                          4.4-8
4.4-6     Initial mechanisms of fabric filtration                     4.4-9
4.4-7     Baghouse performance, lead sinter machine                   4.4-11
4.4-8     Baghouse performance, industrial boiler                     4.4-11
4.4-9     Fabric filter penetration                                   4.4-12
4.4-10    Effect of air-to-cloth ratio on outlet concentration        4.4-13
4.4-11    Penetration correction term as a function of pressure
            drop and 
LIST OF FIGURES (continued)
Number                                                                Page
4.4-18    Effect of gas temperature (continuous) on life of
            glass fabric bags                                         4.4-26
4.4-19    Typical fabric weaves                                       4.4-29
4.4-20    Cross section of a thimble protecting bottom of bag         4.4-35
4.4-21    Dust penetration around snap-ring attachment                4.4-39
4.4-22    Gas jet adjacent to pin-hole                                4.4-40
4.4-23    Abrasion damage at bag inlet                                4.4-41
4.4-24    Bag failure location records                                4.4-43
4.5-1     Spray tower scrubber                                        4.5-3
4.5-2     Vane type scrubber                                          4.5-5
4.5-3     Packed tower scrubber                                       4.5-6
4.5-4     Moving-bed scrubber                                         4.5-7
4.5-5     Tray scrubber                                               4.5-9
4.5-6     Venturi scrubber                                            4.5-11
4.5-7     Throat sections of variable throat venturi scrubbers        4.5-12
4.5-8     Theoretical single drop collection efficiency due to
            diffusion and impaction                                   4.5-13
4.5-9     Theoretical penetration curves for various-sized
            packed-bed scrubbers                                      4.5-17
4.5-10    Theoretical penetration curve for impingement plate
            scrubber                                                  4.5-19
4.5-11    Penetration curve for an impingement plate scrubber
            on a rotary salt dryer                                    4.5-20
4.5-12    Theoretical penetration curve for a venturi scrubber
            illustrating effect of throat velocity                    4.5-23
4.5-13    Theoretical penetration curves for venturi scrubber
            illustrating effect of liquid-to-gas ratios               4.5-24
4.5-14    Predicted venturi scrubber performance for f = 0.25         4.5-25
4.5-15    Comparison of Calvert's model results against measured
            penetration data                                          4.5-26
                                   xi i i

LIST OF FIGURES (continued)
Number                                                                Page
4.5-16    Comparison of Yung, Calvert, and Barbarika model
            against measured penetration data                         4.5-28
4.5-17    Comparison of venturi scrubber outlet loadings to
            static pressure drops for oil fired lime kilns             4.5-30
4.5-18    Correlation of coal-fired boiler scrubber outlet
            dust loadings with theoretical power consumption          4.5-31
4.5-19    Comparison of predicted penetration as calculated in
            Equation 4.5-19 and measured penetration                  4.5-33
4.5-20    Liquid entrainment separators                               4.5-39
4.6-1     Typical thermal incinerator-                                4.6-3
4.6-2     Effect of air velocity and particle diameter on the
            combustion rate of carbon                                 4.6-6
4.6-3     Fuel required to oxidize different concentrations
            of combustible vapor                                      4,6-11
4.6-4     Tubular recuperator                                         4.6-13
4.6-5     Fixed-bed, pebble-stone, regenerative afterburner           4.6-13
4.6-6     Packed-bed flame arrestor                                   4.6-15
4.6-7     Corrugated metal flame arrestor with cone removed and
            tube bank pulled partly off the body                      4,6-15
4.6-8     Typical forced draft oil burner                             4.6-17
4.6-9     Service temperature ranges for refractories                 4.6-18
5-1       Hood design                                                 5-8
5-2       Hood location                                               5-8
5-3       Air flow direction                                          5-8
5-4       Belt conveyor ventilation for fugitive emissions
            control                                                   5-9
5-5       Hopper and bin chute and conveyor-loading ventilation
            for fugitive emissions control                            5-10
5-6       Bag-filling fugitive emissions control                      5-11

LIST OF FIGURES (continued)
Number    '                                                            Page
5-7       Wet dust suppresion system applied to material handling
            operation                                                 5-15
6-1       Incremental energy requirements for fans                    6-3
6-2       Energy required for transformer-rectifier set               6-7
6-3       Energy required for ESP insulator heaters and purge air
            fans                                                      6-9
6-4       Energy required for pumps                                   6-11
6-5       Energy required for stack gas reheat                        6-12
6-6       Sedimentation tank or "clarifier"      •                     6-17
6-7       Vacuum filter                                               6-21
7-1       Cost of electrostatic precipitators; carbon steel
            construction, thermally insulated, FOB factory            7-12
7-2       Cost of fabric filters, carbon steel construction, FOB
            factory                                                   7-13
7-3       Cost of mechanical collectors, carbon steel construction,
            FOB factory                                               7-14
7-4       Cost of incinerators, FOB factory                           7-15
7-5       Cost of venturi scrubbers, unlined throat, carbon steel
            construction, FOB factory                                 7-16
7-6       Capital and annualized costs of fans and 30.5 m length
            of duct                                                   7-17
7-7       Capital and annualized costs of fan driver for various
            head pressures                                            7-18
7-8       Capital and annualized costs of electrostatic precipi-
            tators,  carbon  steel construction                         7-19
7-9       Capital and annualized costs of fabric  filters, carbon
            steel construction                                        7-20
7-10      Capital and annualized costs of fabric  filters, stainless
            steel construction              .                          7-21

LIST OF FIGURES (continued)

Number                                                                Page

7-11      Capital and annualized costs of mechanical collectors,
            carbon steel construction                                 7-22

7-12      Capital and annualized costs of incinerators                7-23

7-13      Capital and annualized costs of venturi scrubbers, carbon
            steel construction                                        7-24

7-14      Capital and annualized costs of venturi scrubber,
            stainless steel construction                              7-25

8-1       APS electrostatic scrubber                                  8-3

8-2       Fraction efficiency performance of APS electrostatic
            scrubber                                                  8-4

8-3       TRW charged droplet scrubber                                8-6

8-4       TRW charged droplet scrubber fractional efficiency
            performance                                               8-7

8-5       University of Washington electrostatic droplet scrubber
            schematic                                                 8-9

8-6       University of Washington electrostatic droplet scrubber
            fractional efficiency performance                         8-10

8-7       Lone Star Steel steam-hydro air cleaning schematic          8-12

8-8       Lone Star Steel steam-hydro air cleaning fractional
            efficiency performance                                    8-13

8-9       Aeronetics two-phase jet scrubber schematic                 8-14

8-10      Aeronetics two-phase jet scrubber fractional efficiency
            performance                                               8-16

8-11      Pulse energization voltage-current relationships for
            various pulse frequencies                                 8-19

8-12      Comparison of DC and pulse energization voltage-current
            relationships with same discharge electrode               8-19

8-13      Southern Research Institute precharger ESP assembly
            drawing                                                   8-22

8-14      Southern Research Institute precharger ESP fractional
            efficiency performance                                    8-24

LIST OF FIGURES (continued)

Number                                                                Page

8-15      Apitron electrostatic-filter cutaway view                   8-30

8-16      Apitron electrostatic-filter fractional efficiency
            performance                                               8-31

8-17      High gradient magnetic separator schematic
            representation                                            8-36

8-18      High-gradient magnetic separator fractional efficiency
            performance                                               8-37
                                   xvi i


                               LIST OF TABLES
Number                                                                Page
2-1       Estimated Participate Emissions from Manmade Sources,
            1977                                                      2-7
2-2       Potential Industrial Sources of Fugitive Particulate
            Emissions                                                 2-10
3-1       Particulate Emission Reduction Potential of Various
            Energy Sources                                            3-2
4-1       Particle Capture Mechanisms Normally Active in
            Conventional Particulate Control Devices                  4.1-12
4.2-1     Major Types of Mechanical Collectors                        4.2-1
4.2-2     Effects of Operating Conditions on Cyclone Performance      4.2-23
4.3-1     Design Power Density                                        4.3-39
4.3-2     Reaction Mechanisms of Major Conditioning Agents            4.3-55
4.3-3     Example Effects of Changes in Normal Operation on ESP
            Control Set Readings                                      4.3-75
4.4-1     Recommended Temperature Limits for Various Commercial
            Fabrics                                                   4.4-24
4.4-2     Chemical Resistance of Common Commercial Fabrics            4.4-27
4.5-1     Major Types of Wet Scrubbers                                4.5-2
4.5-2     Typical Liquid-to-Gas Ratios for Wet Scrubbers              4.5-32
4.5-3     Typical Scrubber Pressure Drop                              4.5-34
4.5-4     Properties of metals used as materials of construction
            for wet scrubbers and auxiliary components                4.5-41
4.6-1     Auto-ignition Temperatures of Organic Compounds             4.6-5
4.6-2     ASTM Classification of Fire Clay Refractories               4.6-17
4.6-3     Commonly used Castable Fire Clay Refractories               4.6-17

LIST OF TABLES (continued)
Number                                                                Page
4.6-4     General Physical and Chemical Characteristics of
            Classes of Refractory Brick                               4.6-19
5-1       Industrial Process Fugitive Emission Sources and
            Applicable Control Techniques                             5-3
5-2       Fugitive Dust Emission Sources and Applicable Control
            Techniques                                                5-4
6-1       Typical Static Pressure                                     6-2
7-1       Average Cost Factors for Estimating Capital Costs           7-6
7-2       Cost Adjustment Factors for Emission Control Systems        7-7
7-3       Example Factors for Annual!zed Costs                        7-9
7-4       Typical Costs of Wet Suppression of Industrial Process
            Fugitive Particulate Emissions                           .7-29
7-5       Cost Estimates for Wet Suppression of Fugitive Dust         7-31
7-6       Cost Estimates for Stabilization of Fugitive Dust           7-32
7-7       Cost Estimates for Sweeping and Flushing of Fugitive
            Dust Sources                                              7-33
8-1       CDS Design Summary                                        •  8-8
8-2       Comparison of High-Gradient Magnetic Separator and
            Conventional Technology                                   8-38
8-3       Results of Industrial Tests with Sonic Agglomeration        8-40


Symbol         Definition
a              cross-sectional area
A              collection plate area of an electrostatic precipitator
A              wetted surface area
C              Cunningham correction factor
c              dust loading
d              diameter
d ,             drop diameter
d,             fiber diameter
d              aerodynamic particle diameter
D              gas diffusivity
D              particle diffusivity
E              charging-field strength
£              precipitation field strength
g              acceleration of gravity
h              height
H.             liquid holdup
K              inertia parameter
K t            inertia parameter at throat velocity
K2             resistance coefficient of dust cake
1              bed depth
n              number of plates or stages
Npp            Reynolds number

Symbol         Def1nj ti on
p              total pressure
p              static pressure
p              corona power
P.             penetration
q              particle charge
Q,             liquid throughput
Q              gas throughput
r,             drop radius
r.             radius of hole
T              absolute temperature
t              time
tR             residence time
v              gas velocity
v              pickup velocity
of              solid fraction in fiber bed
A              static pressure change
q              efficiency
(j              gas viscosity
p.             liquid viscosity
p              gas density
p.             liquid density
p              particle density
pR             bulk resistivity
Py             in-situ resistivity
e              porosity
uj              migration velocity
                                   xxi i

                             CONVERSION FACTORS

1 meter (m) = 3.281 ft
1 meter (m) = 3.937 x 101. in.
1 meter2 (m2) = 1.076 x 101 ft2
1 meter3 (m3) = 1.308 yd3
1 meter3 (m3) = 3.532 x 101 ft3
1 meter/second (m/s) = 196.8 ft/min
1 meter/second (m/s) = 3.281 ft/s
1 meterVsecond (m3/s) = 2.119 x 103 ftVmin
1 meter3/second (m3/s) = 1.585 x 10s gal (U.S. liquid)/min
1 meterVsecond (m3/s) = 2.282 x 107 gal (U.S. liquid)/day
1 kilogram (kg) = 2.205 Ib
1 kilogram (kg) = 1.102 x 10"3 short tons (2000 Ib)
1 kilogram/meter3 (kg/m3) = 1.284 x IQ-2 lb/ft3
1 kilogram/meter3 (kg/m3) = 8.98 x 101 grains/ft3
1 joule (0) = 9.479 x 10'4 Btu (mean)
1 joule (J) = 2.778 x 10~7 kWh
1 watt (W) = 1.340 x 10-3 hp
1 pascal (Pa) = 1.45 x 10"* lbf/in.2 (psi)
1 pascal (Pa) = 4.019 x 10"3 in. H20
1 pascal second (Pa-s) = 0.672 Ib/ft2-s
1 kilopascal (kPa) = 4.019 in. H20
                                   xxi 11


ABSORPTION.1  Transfer  of molecules  from  the bulk  of the gas to  a liquid
     surface followed by diffusion to the bulk of the liquid.

ADIABATIC SATURATION.1  A  process  by means of which an air or gas stream is
     saturated with water  vapor  without adding or subtracting heat from the

AERODYNAMIC DIAMETER.   The diameter of a unit density sphere having the same
     aerodynamic properties as an actual particle.

AEROSOL.  A dispersion of  solid or  liquid  particles of  microscopic  size.

AGGLOMERATION.   The  combination of  smaller particles  due to  collisions.

AIR, DRY.  Air containing no water vapor.

AIR-TO-CLOTH  RATIO (A/C).   The volumetric  rate  or capacity  of  a fabric
     filter;  the  volume  of air (gas)  cubic meter  per minute,  per square
     meter of filter medium (fabric).

ATOMIZATION.  The  reduction of liquid to a fine spray.

BACK  CORONA.   Localized  electrical  breakdown  of  a dust  layer, producing
     positive  ions,  which  degrade  or  neutralize  the   intended  charging

BAROMETRIC SEAL.1  A column of liquid used to hydraulically seal a scrubber,
     or  any  component thereof,  from atmosphere or other part of the system.

BLAST GATE.2  A  sliding plate installed in a supply or exhaust duct at right
     angles to the duct for the  purpose of regulating air flow.

BLINDING  (BLINDED).2   The  loading,  or  accumulation, of  filter  cake to the
     point where capacity rate is diminished.

BURNER.1  A device for the  introduction of fuel and air  into a furnace at
     the  desired velocities,  turbulence, and concentration to establish and
     maintain proper  ignition and combustion of the fuel.

CASCADE  IMPACTOR.   A  particle-sizing device in which progressively  increas-
     ing  inertia!  forces are used to separate progressively smaller particle

CHEVRON MIST  ELIMINATOR.1  Series  of diagonal baffles  installed in  a  gas
     stream,  designed  to separate  fine  droplets  of liquid from  the  gas by
     means of inertia! impaction on the surfaces of the baffles.

COCURRENT.   Flow  of scrubbing  liquid  in  the same  direction  as the  gas

COLLECTION EFFICIENCY.1   The  ratio  of the weight of pollutant  collected to
     the total weight of pollutant entering the collector.

CONDENSATION.1   The  physical  process of  converting  a  substance from  the
     gaseous phase to the liquid or solid phase via the removal  of heat,  the
     application of pressure, or both,

CONTACT CHARGING.   Charging of  particles  by  contacting  them and then  re-
     leasing them from a charged surface.

CORONA  CURRENT.   Measure of  the current flow from the transformer  to  its
     electrical section in an electrostatic precipitator (ESP).

COUNT.2  The number of warp yarns (ends) and filling yarns (picks) per inch.
     Also called thread count.

CROSSFLOW.  Flow of scrubbing liquid normal  to the gas stream.

CROWFOOT  SATIN.2  A 3/1  broken twill  arranged 2 threads  right,  then  2
     threads left.  Also called 4 shaft satin, or broken crow weave.

CUNNINGHAM  FACTOR.   A  correction  factor  to  account  for  slippage of fine
     particles moving through a discontinuous gaseous medium.

CURRENT DENSITY.   Corona  current level per unit area  of  collection  surface
     of an electrostatic precipitator (current per plate).

CYCLONE.  A  device in which  the velocity  of an inlet gas stream is  trans-
     formed into  a confined vortex  from which inertial forces tend to drive
     particles to the wall.

DAMPER.2  An  adjustable  plate  installed in  a duct  to  regulate  gas  flow.

DEHUMIDIFY.*  Reduction of water vapor content of a gas stream.

DEMISTER.   A  mechanical  device used to remove entrained water droplets from
     a scrubbed gas stream.

DENIER.2  The number, in grams, of a quantity of yarn,  measuring 9000  meters
     in length.   Example:  A  200-denier yarn measuring 9000 meters  weighs
     200  grams.   A 200/80-yarn  indicates  a 200-denier yarn  composed  of 80
     filaments.  Usually  used  to describe  continuous multifilament yarns of
     silk, rayon, Orion, Dacron, Dynel, Nylon, and similar materials.

DENSITY.2  The ratio  of  the mass of a specimen of a substance to the volume
     of the specimen.   The mass of a unit volume of a substance.

DIELECTRIC STRENGTH.   The maximum  potential  gradient that  may  exist  In  a
     material without the occurrence of electrical breakdown.

DIFFUSION  (AEROSOL).   Random motion of particles caused  by  repeated colli-
     sions of gas molecules.

DIFFUSION  (LIQUID).1   The  spontaneous  intermingling  of  miscible  fluids
     placed  in mutual  contact,  and accomplished without the aid of mechani-
     cal mixing.

DIFFUSION  CHARGING.   Process  of transferring electrical  charge to particles
     by random movement  of electrons and ions; the effective charging mech-
     anism for submicrometer-sized aerosols.

DIFFUSIOPHORESIS.  Force  acting on a particle, effecting movement  due  to a
     vapor  condensation  gradient,  resultant  of  differences in  molecular
     impacts on opposite sides of a particle.

DIMENSIONAL  STABILITY.2   Capability of fabric to retain finished length and
     width, under stress, in hot or moist atmosphere.

DRAFT.1   A  gas  flow  resulting from  the pressure  difference  between  the
     incinerator, or any component part, and the atmosphere, which moves the
     products  of combustion  from the  incinerator  to the  atmosphere.   (1)
     Natural  draft:    the  negative  pressure  created by  the  difference in
     density  between  the hot  flue gases and  the atmosphere.   (2) Induced
     draft:   the  negative pressure created by the vacuum action of a fan or
     blower  between the  incinerator and the stack.   (3)  Forced  draft:   the
     positive  pressure created by  the fan or  blower,  which  supplies  the
     primary or secondary air.

DRAG  FORCE.   Resistance  of a  viscous  medium  due to relative motion  of a
     fluid and object.

DUST.2  Solid  particles  less than 100  micrometers created  by the attrition
     of larger particles.

DUST  LOADING.2   The  weight  of solid particulate  suspended  in an airstream
     (gas),  usually expressed in terms of grains per cubic foot, grams per
     cubic meter, or pounds per thousand pounds of gas.

DUST  PERMEABILITY.2    The  mass of dust (grains) per square  foot  of media
     divided  by  the  resistance (pressure drop)  in  inches water  gauge per
     unit  of filtering velocity,  feet per minute.   Not to be compared with
     cloth permeability.

ELECTROSTATIC  FIELD.   The  position-dependent  electrostatic  force  per  unit
     charge,  made  up  of two components—one related  to applied  voltage and
     electrode geometry,  the other related to space change due to the pres-
     ence  of electrons, ions, and  charged particles.

                                    xxv ii

ENTRAPMENT SEPARATOR (DEMISTER).   That  part of a gas  scrubber  designed to
     remove  entrained  liquid  droplets  from  a gas  stream by  centrifugal
     action, by  impingement on  internal  surfaces  of the scrubber, or  by a
     bed of  packing,  mesh, or baffles  at or near the  scrubber  gas  outlet.

ENTRY LOSS.2  Loss in total pressure caused by air (gas) flowing into  a duct
     or hood.

EXCESS AIR.1   Air supplied for  combustion in  excess of  that  theoretically
     required for  complete combustion;  usually expressed  as  percentage of
     theoretical  air (130% excess air).

FABRIC.2   A planar   structure  produced  by  interlacing  yarns,  fibers,  or
     filaments.    (1)  Knitted  fabrics produced by interlooping strands of
     yarns, etc.    (2) Woven fabrics are produced  by  interlacing strands at
     more or less  right angles.   (3) Bonded fabrics or a web of fibers held
     together with a cementing medium which does not form a continuous sheet
     of adhesive  material.  (4) Felted fabrics or structures built  up  by the
     interlacing action  of the  fibers themselves without spinning, weaving,
     or knitting.

FEEDSTOCK.1  Starting material  used in a process.  Can be raw material or an
     intermediate product that will undergo additional processing.

FIELD  CHARGING.    Process  of transferring electrical  charge  to  particles
     induced by  high  electric  field strengths  in  the interelectrode  space;
     the effective  charging mechanism  for particles greater  than 1  micro-

FIELD  STRENGTH.   A  force  field  created  by a  large  potential difference
     between  surfaces  of  different  polarity;  measured  by  the  potential
     difference divided by distance between surfaces.

FILAMENT.2  A continuous fiber.

FILL.2  Crosswise threads woven by loom.

FILL COUNT.2  Number of fill threads per inch of cloth.

FILTER MEDIUM.   The  substrate support  for the filter  cake;  the  fabric on
     which the filter cake is built.

FILTER VELOCITY.    The velocity  at which  the  air (gas)  passes  through the
     filter medium, or  the velocity of approach to the-, medium.  The  filter
     capacity rate.

FLY ASH.   Finely divided  particles of ash entrained in  flue  gases  arising
     from the combustion of fuel.   The particles of ash may contain unburned
     fuel and minerals.

FUGITIVE DUSTS.  A type  of participate emission made airborne  by forces of
     wind,  man's activity, or both—such as construction sites,  tilled land,
     or windstorms.3

FUGITIVE EMISSIONS.   Particles  generated by industrial  or  other activities
     which  escape to  the atmosphere not through primary exhaust systems but
     through  openings such  as  windows, vents  and doors,  ill-fitting  oven
     doors, or poorly-maintained equipment.3

FUGITIVE EMISSIONS.   Industrial  process such  as  emissions  from  a  point,
     area,   or line  source  other  than  a  stack,  flue,  or control  system.
     Emissions escape to the atmosphere from a  defined industrial  process
     flow  stream  because of  leakage,  materials charging/  handling,  inade-
     quate   operational  control, lack of reasonably  available control  tech-
     nology, transfer, or storage.

FUME.1   Fine  solid  particles  predominately  smaller than  1 micrometer in
     diameter  suspended  in  a  gas.   Usually  formed from  high-temperature
     volatilization of metals or by chemical reaction.

GALVANIC SERIES.1  A list  of metals arranged  according to their relative
     tendencies to corrode.   When dissimilar metals are joined together in
     an  electrolytic  solution,  the  one closest to the  "active"  end  of the
     galvanic  series  corrodes  preferentially  to  the   one  closest to  the
     "passive" end.
GRAVITY, SPECIFIC.2  The  ratio of the mass of  a unit volume of a substance
     to  the  mass of the  same  volume of a standard substance  at a standard
     temperature.  Water  is  usually the standard substance.  For gases, dry
     air at  the  same temperature and pressure as the gas is often the stan-
     dard substance.

GRID.1   A  stationary support  or  retainer  for a bed of  packing  in a packed
     bed scrubber.

HEADER.1  A pipe used to supply and distribute liquid to downstream outlets.

HUMIDITY, ABSOLUTE.2  The  weight of water vapor carried by a unit weight of
     dry air or gas.

HUMIDITY, RELATIVE.2  The ratio  of the absolute  humidity in a gas to the
     absolute humidity of a saturated gas at the same temperature,

HYDROPHILIC MATERIAL.  Particulate matter that adsorbs moisture.

INERTIA.  Momentum; tendency to remain in a fixed direction, proportional to
     mass and velocity.

INTERCEPTION.  A  type of aerosol collection  related  to  impaction, in which
     an  aerosol  impacts  the side of an obstacle because of reduced mobility
     across streamlines.

INTERELECTRODE  SPACE.   The  space between  the  discharge electrode  and  the
     collection plate;  the active  particle-charging region  in  an electro-
     static precipitator.

INTERSTICES.2  The openings between the inter!acings of the warp and filling
     yarns; the voids.

ION, GASEOUS.   A gas  molecule that  loses  or gains one  or more electrons.

IONIZATION.  Generation  of free electrons that become attached to gas mole-
     cules, forming ions.

ISOKINETIC SAMPLING.   Matching the  gas  velocity at the  sampling probe  en-
     trance to  the gas  velocity  of  the  localized gas  stream  to collect a
     representative particle size distribution.

LIQUOR.1   A  solution of  dissolved substance in  a  liquid  (as  opposed to a
     slurry, in which the materials are insoluble),,

LOG-NORMAL DISTRIBUTION.   A  series of points that  can  be defined by a geo-
     metric mean value and a geometric standard deviation.

MEAN FREE  PATH.   The  average distance between successive collisions of  gas
     molecules; related to molecular size and number per unit volume.

MIGRATION VELOCITY.   The  average  drift velocity of charged particles normal
     to  the  direction  of gas  movement; also  known as  precipitation rate
     parameter, a measure of  the efficiency of collected  particles to  the
     volume of gas treated and the area of the collection plate.

MOBILITY.  A measure of response per unit force;  the ease of motion relative
     to the magnitude of the force-inducing motion.

MONOFILAMENT.2  A continuous fiber  of sufficient size to  serve  as  yarn in
     normal textile operations.

MULLEN BURST.   The  pressure  necessary to rupture a secured fabric specimen.

MULTIFILAMENT (MULTIFIL).2  A yarn bundle composed of a number of filaments.

NAPPING  PROCESS.2   A process  to  raise fiber of filament  ends  (for better
     coverage and more surface area), accomplished by passing the cloth over
     a large revolving cage or drum of small power-driven rolls covered with
     card clothing (similar to a wire brush).

NEEDLED FELT.2  A felt  made by the placement of loose fiber in a systematic
     alignment, with  barbed  needles  moving  up and down, pushing and pulling
     the fibers to form an interlocking of adjacent fibers.

NONWOVEN FELT.  A felt  made by needling, matting  of fibers,  or compression
     with a bonding agent for permanency.

OPACITY.  Measure of  the fraction of light attenuated by suspended particu-

PARTICLE.  Small discrete mass of solid or liquid matter.

PARTICLE  SIZE.   An  expression for  the size  of  liquid or  solid particle.

PARTICULATE MATTER.   As  related to control technology,  any material  except
     uncombined water that  exists as a solid or liquid in the atmosphere or
     in  a gas  stream as  measured by a standard (reference) method at speci-
     fied conditions.  The  standard method of measurement and the specified
     conditions should be implied in or included with the particulate matter

PARTICULATE MATTER,  ARTIFACT.   Particulate  matter  formed  by one  or more
     chemical  reactions within the sampling train.

PENETRATION.   Fraction of suspended particulate that passes  through  a col-
     lection device.

PERMEABILITY,  FABRIC.  The capability of air (gas) to pass through a fabric.
     Measured  on  Frazier porosity  meter or Gurley  permeometer.   Not to be
     confused with dust permeability.

PENTHOUSE (ESP).1   Weatherproof  gas-tight  enclosure over the electrostatic
     precipitator that contains the high-voltage insulators.

pH.l  A  measure of acidity-alkalinity of a solution; determined by calculat-
     ing the negative logarithm of the hydrogen ion concentration.

PLAIN  WEAVE.2   Each warp  yarn passing alternately  over each filling yarn.
     The simplest weave, 1/1 construction; also called taffeta weave.

PLATE  AREA.  The  effective area of both sides of the collecting surfaces in
     an  electrostatic precipitator.

POLYDISPERSITY.   A  particle size distribution consisting of different size

PRESSURE, STATIC.  The  pressure exerted in all directions by a fluid; mea-
     sured  in a direction normal to the direction of flow.

PRESSURE, TOTAL.  The algebraic sum of the velocity pressure and the static

PRESSURE,  VELOCITY.   The  kinetic  pressure in the  direction of  gas flow.

PRIMARY  PARTICULATE  MATTER.   Particulate  matter emitted  directly into the
     air from  identifiable sources.

PRIMARY  STANDARD.   The national primary ambient  air quality standard which
     defines  levels   of  air  quality that  are necessary  to  protect public

                                    xxx i

PRIME COAT (PRIMER).1  A first coat of paint applied to inhibit corrosion or
     to improve adherence of the next coat.

QUENCH.1  Cooling of hot gases by rapid evaporation of water.

RAPPER  (ESP).   Device  for imparting acceleration of  the  collecting surface
     to dislodge the deposited particles.

RAPPER  INSULATOR.   A device  that  electrically  isolates  a rapper  from  the
     high-voltage system  of an electrostatic precipitator, yet  permits  the
     transmission of mechanical forces.

REFRACTORY.1   Ceramic  material used  for the lining  of vessels,  ducts,  and
     pipe for  protection from  heat,  abrasion, or corrosion;  also  used  for

RESISTIVITY.  The impedance  offered to charge transfer across a dust layer;
     defined by the ratio  of electric  field  intensity  to the  current  per
     unit area passing through the dust layer.

REYNOLDS NUMBER, FLUID.   A  dimension!ess quantity in fluids to describe  the
     ratio of inertia!  to viscous forces.

REYNOLDS NUMBER, PARTICLE.   A dimensionless quantity  in  aerosol  science to
     describe the ratio of inertia! to viscous forces relative to the parti-

SATEEN.2  Cotton cloth made with a 4/1 satin weave, either as warp sateen or
     filling sateen.

SATURATED GAS.1  A mixture of gas and vapor to which no additional vapor  can
     be added,  at specified conditions.   Partial pressure of vapor is equal
     to vapor  pressure  of the liquid at the  gas-vapor mixture temperature.

SATIN  WEAVE.2   A  fabric usually  characterized  by   smoothness  and luster.
     Generally  made warp  face with a great many  more ends than picks.   The
     surface consists  almost entirely  of  warp (or filling)  floats in  con-
     struction  4/1  to  7/1.   The intersection points  do not fall  in regular
     lines,  but are shifted regularly or irregularly.

SECONDARY PARTICULATE  MATTER.  Particulate matter formed  in  the atmosphere
     by  physical  and/or  chemical  gas-to-aerosol  conversion  mechanisms.

SECONDARY POLLUTANT.  A pollutant  not emitted into the air from a pollution
     source, but formed in  the air from the reactions of primary pollutants
     (often photochemically).

SEEPAGE.  The  migration  of particles  through  a  freshly cleaned fabric.

SIZE DISTRIBUTION.   Distribution of  particles  of different  sizes  within a
     matrix of aerosols;  numbers  of particles of  specified sizes  or  size
     ranges, usually in micrometers.

SLURRY.1  A mixture  of  liquid and finely divided insoluble solid materials.

SMOKE.    Small  gasborne  particles  resulting  from  imcomplete  combustion;
     particles consist  predominantly of carbon and  other  combustible mate-
     rial; present in sufficient  quantity to be observable independently of
     other solids.

SNEAKAGE.  Portion  of a  gas  stream that  bypasses the  intended collection
     area in an electrostatic precipitator.

SOOT.   An  agglomeration  of carbon particles impregnated with  "tar,"  formed
     in the incomplete combustion of carbonaceous material.

SPECIFIC GRAVITY.1   The  ratio between  the density of a substance at a given
     temperature and the density of water at 4°C.

SPRAY NOZZLE.1  A  device used for the  controlled  introduction of scrubbing
     liquid  at predetermined  rates,  distribution patterns, pressures,  and
     droplet sizes.

SPUN FABRIC.2  Fabric woven from staple (spun) fiber; same as staple.

STAPLE  FIBER.2  Manmade fibers cut to specific length (1% in.,  2 ft, 2k in.,
     etc.);  natural   fibers  of  a  length characteristic  of  fiber,  animal
     fibers being the longest.

STOKES  NUMBER.  Descriptive of the particle collection potential of a speci-
     fic system; the ratio of particle-stopping distance  to the distance a
     particle must travel to be captured.

STREAMLINE.  The visualized path of a fluid in motion.

SUSPENDED PARTICULATE MATTER.  Participate matter in the ambient atmosphere,
     as determined by a specific reference method; material generally refer-
     red  to   as  total  suspended  particulate  (TSP); consists  of particles
     within the size range of 100 to 0.1 micrometer in diameter.

TENSILE  STRENGTH.2   The  capability of yarn or  fabric  to resist breaking by
     direct tension.  Ultimate breaking strength.

TEMPERATURE,  ABSOLUTE.2    Temperature  expressed  in  degrees above absolute

TERMINAL  SETTLING  VELOCITY.  The  steady-state speed  of a falling particle
     after the equilibration of  gravitation,  drag, and buoyant forces has

TRANSFORMER-RECTIFIER SETS.  Electrical device used in electrostatic precip-
     itators  to  rectify  a.c.  to  d.c.  and to transform  low voltage to high

THREAD  COUNT.   The  number  of  ends and  picks per  inch  of a  woven cloth.

TURBULENT FLOW.   A  type of flow in which the fluid passes in a nearly ran-
     dom, fluctuating motion.

TWILL WEAVE.2   Warp yarns  floating over or  under  at  least two consecutive
     picks from lower  to upper right, with the point of intersection moving
     one yarn  either outward  and  upward  or  downward on  succeeding picks,
     causing diagonal lines in the cloth.

VAPOR.   The  gaseous form  of substances  that are  normally  in  the solid or
     liquid state and  whose states can be  changed  either  by increasing the
     pressure or by decreasing the temperature.

WARP.2  Lengthwise threads in loom or cloth.

WARP COUNT.2  Number of warp threads per  inch of width.

WET/DRY  LINE.1  The  interface  of hot, dry particulate-laden gas and cooling
     or  scrubbing liquid,  at which  an  accumulation  of solids  can occur.

WOVEN FELT.2   Predominantly  a  woven woolen fabric heavily fulled or shrunk,
     with the  weave  completely hidden  by the  entanglement of  the woolen

1.   Industrial Gas Cleaning Institute.  Wet Scrubber Terminology.  Publica-
     tion WS-1, July 1975.

2.   Industrial Gas  Cleaning Institute.   Fundamentals  of Fabric Collectors
     and Glossary of Terms.  Publication F-2, August, 1972.

3.   PEDCo Environmental, Inc.   Technical Guidance for Control of Industrial
     Process  Fugitive  Particulate Emission.   EPA-450/3-77-010,  March 1977.

                                  SECTION 1

     This document is  a  revision of "Control Techniques for Particulate Air
Pollutants,"1 which was  published  in January 1969.   Changes and advances in
the technology of  particulate  control  have made parts of the original docu-
ment obsolete.   This  second edition contains up-to-date  information  on the
emission  reduction capabilities, costs,  energy requirements,  and  environ-
mental  impact of available control  techniques, as required in Section 108(b)
of the Clean Air Act of 1977.
     As in the  first  edition,  the control  techniques  are based on informa-
tion from many technical  fields.  The methods and principles of operation of
many of  the techniques  have been  known for years, but  much  experience has
been gained  in  their  applications  since 1969.  The document  also discusses
techniques that  are  still  in  various  stages  of research  and development,
even though  these  new  techniques  are not  yet  available for  general  use.
     Recent  scientific data summarized in "Sulfur Oxides-Suspended Particu-
late Air Quality Criteria Document"  have  led  to  increased  concern  about
particles in the inhalable  size range.  This revision includes an expansion
of  information   on  control  effectiveness  as a  function of  particle size.
Information  on  the conversion  of gaseous  pollutants  to aerosols (secondary
particulate  matter)  has  also  been  incorporated.  The revised document re-
flects increased interest  in  fugitive particulate emission sources.  Infor-
mation has been  added on techniques to prevent and control these emissions.
     The document is issued as two volumes.  Volume 1 presents basic techni-
cal  information  on particulate  emissions and  control  techniques;  Volume 2
deals  with  control   technology  applied  to  major  categories  of  pollu-
tant-emitting sources.   The volumes are  intended as  general  references for
technical personnel in regulatory agencies and in the private sector.

Because the  document is  a general  summary,  it should  not  be used  as the
basis for developing or enforcing control regulations.
     Volume 1  has  eight sections.   Section 2 presents fundamentals of aero-
sol mechanics, trends in emissions and air quality, and sampling procedures.
The  definition of  particle size  is of special  importance  because  of the
differences among  the  definitions  in the technical  literature.   Section 3
discusses the general ways in which particulate emissions can be minimized—
that  is,  by  energy  souYce  and  fuel  selection,  by process  selection and
modification, and by exhaust gas cleaning.
     Section 4 presents detailed information on exhaust  gas  cleaning tech-
niques.   The  operating principles,  control  effectiveness, and maintenance
requirements  are  summarized  for each major  category of  techniques.   Sec-
tion 5 discusses  fugitive dust  and industrial process  fugitive  emissions.
Most  exhaust  gas  cleaning techniques concentrate  particulate  matter  into a
liquid or solid waste  stream.   Accordingly,  Section 6  presents information
on the environmental  impact of these materials.  This information is accom-
panied by an evaluation of energy requirements for particulate control tech-
     Section 7 discusses the cost of particulate control.  All the costs are
in first-quarter  1980 dollars,  unless otherwise  indicated.   Section  8 pre-
sents the state of development of novel particulate control concepts.
     Section 9 discusses  the sources of specific  particulate emissions and
the  technology generally  used  to  control  emissions  from   novel  sources.
Volume 2  consists  entirely of  Section 9,  "Sources of Particulate Emissions
and  Control Techniques."   The  page numbers in  Volume 2,  therefore,  go from
page 1-2 (this page) to page 9.1-1, the first page of Section 9.
     The  data are  given  in  metric  units specified  in  the  International
System of Units (SI).   Conversion factors are listed in the  front  of the
report, as are important symbols.


1.   U.S.  Department of Health, Education, and Welfare.  National Air Pollu-
     tion  Control  Administration.    Control  Techniques for  Particulate Air
     Pollutants.  AP-51, January 1969.


                                  SECTION 2

     This  section  provides background  information relative  to  particulate
emissions and how  they  are measured.   The generation  of  particulate matter
from stationary combustion sources, industrial processes,  and fugitive emis-
sion sources  is  discussed,  along with the secondary formation and transport
of  particulate matter.   Sampling and analytical methods  of  evaluating par-
ticulate matter are also described.

     Particulate matter  is  generated  by a variety  of  physical  and chemical
mechanisms,  and  is  emitted  to  the  atmosphere  from  numerous  sources,  in-
cluding  combustion,  industrial  process,  fugitive emission,  and  natural
sources.   Particulate matter  is composed of finely  dispersed  liquids  and
solids, including soot; dust; organic substances; inorganic substances, such
as  sulfur  compounds,  metallic oxides, and salts; and other substances.  The
chemical composition  of particulate  matter  varies  with source characteris-
tics, geographic area, and season of the year.
     An estimated  12.4  teragrams (Tg) of particulate matter is emitted from
manmade sources  in the United States each year.1'2 The  major contributors
are  stationary fuel  combustion sources and  industrial  processes, which con-
tribute 39 and 42 percent, respectively.  Transportation sources account for
9  percent  of the  total,  and  solid  waste  disposal  and forest  fires each
account for approximately 4 percent of the total.1'2
2.1.1  Air Quality and Particulate Matter Emission Trends
     In  1971, the  U.S.  Environmental Protection  Agency  (EPA), promulgated
both  primary  (health-related)  and  secondary  (we Ifare-related)  National
Ambient  Air  Quality Standards (NAAQS) for  particulate matter.   Air assess-
ment  in 1979 showed  that 395  counties  were classified  as  "nonattainment"


(not  having  achieved the  NAAQS)  for particulate matter,  the  Nation's most
commonly monitored pollutant.
     Ambient particulate  levels at  a monitoring site may be  viewed as the
sum of particulate from traditional sources, nontraditional sources, natural
sources,  and  transported particulate.   Each of these  source  types  may at
times  account  for a  significant  portion  of the measured particulate at a
particular  monitoring  site.3   In  the  1970's,  particulate  emissions  from
traditional sources were  estimated to have decreased 46 percent nationally.
Dominant in this national trend were decreases in particulate emissions from
industrial processes and from fuel combustion.
2.1.2  Projections for Future Control Programs andEmissions
     Traditional  sources  are still  the  dominant cause  of nonattainment in
many urban areas with heavy industry; these sources may be contributing any-
where  from 15  ug/m3  in residential  areas to over  60 ug/m3  in  industrial
neighborhoods.3   Nontraditional  sources  (e.g.,  reentrainment  of  road dust,
fugitive  dust   emissions  from  construction  and demolition  operations,  and
dust from unpaved roads and parking lots) may contribute from 25 to 35 (jg/m3
to citywide particulate levels, and thus  prevent some  urban areas from at-
taining the standard.
     Nonattainment  of  air  quality  standards  is  also  caused by  natural,
transported,  and  secondary  particulates.   Natural  particulates  are  not
amenable  to  control,    Nonurban  sulfates  and other  secondary  particulates,
transported primary particulates (up to 30 pg/m3 in densely populated metro-
politan  areas), and  urban  secondary particulates  (<10 ug/m3 formed from
gaseous emissions within  the urban area)  can be controlled  by regional and
local planning  and by control of gaseous pollutants.3
     Annual  variations  in  precipitation  have  affected the annual  average
particulate levels by  as  much as 20  ug/m3.   Although meteorology cannot be
controlled,  the meteorological  variations with time and location  must be
considered in air quality analysis and planning.
     New sources will  be  subject to New Source Performance Standards (NSPS)
as NSPS  are developed  for various industries.   The  Clean Air Act requires
that the EPA develop NSPS to prevent new air pollution problems by requiring
the  installation  of  the best available technology  considering cost, health

and environmental  impact,  and energy requirements  during  initial  construc-
tion.   Designed  to  allow industrial growth without  undermining  air quality
management goals, NSPS  are  established  at a national  level  so that control
requirements  for new  sources  are  uniform  and  consistent,   regardless  of
     The impact  of  NSPS on  air quality will become more significant as new
plants are constructed and as older sources are either modified or replaced.
Before amendments  to the Act  were promulgated in 1977, a  study was under-
taken  to  estimate   the potential  emissions  impact of  the  NSPS  program.
Particulate emissions from  stationary  sources  in 1990 were  predicted to be
in  excess  of  15 Tg/yr  if new  sources  were required to meet  only  the State
standards.   If six NSPS were set for particulate matter each year during the
1980's,  the 1990  emissions were  predicted to  be  considerably  less  than
10 Tg/yr.4
     Attention  is being given to  particle size  distribution  to  distinguish
the fraction of particulate matter that  is  likely  to  be inhaled and to the
chemical characterization  of  individual  components.   EPA  has a network of
National Air  Monitoring Sites  (NAMS)  in major urban  areas  to measure both
particle size  and  chemical  composition.   Monitoring at these  sites provides
particulate data for enough urban  areas with populations greater than 50,000
to  permit  the  continued characterization of national trends in particulate.
Improvements  in the characterization of  particulate components  help in the
assessment  of  shifts in the nature  of particulate  matter and in the selec-
tion  of  appropriate control strategies.   A continuing concern is that a net
improvement in overall  particulate levels could mask a shift  in composition
toward the  smaller  particles.

      Particulate  matter is  emitted into the  atmosphere from a  variety of
manmade  and natural  sources,  and is sometimes  formed  in  the atmosphere by
conversion  of natural  and  anthropogenic  gaseous constituents into particu-
late.   This  section discusses  the sources of  particulate  matter and the
sources  of  gaseous  precursors  to particulate matter.

2.2.1  Natural Em1s s i on Sources
     Particulate  emissions from  natural  sources  are  estimated to  exceed
emissions from manmade  sources  on a worldwide basis.3  In areas remote from
urban population  centers with concentrations of  industrial  and transporta-
tion facilities, natural particulate emissions are typically responsible for
more than half  of the average measurable  levels.   Figure  2-1  indicates the
estimated  average  nonurban  particulate  levels  in  the continental  United
States  by  region  and  the estimated  fractions  of these  levels  that  are
attributable  both  to  natural and  to manmade  sources, and to particulate
transported into the region.3
     The most important of the natural particulate sources are  soil  and rock
debris,   forest  fires,  volcanoes, and  ocean salt  spray.   On  land,  wind-
entrained dust  from soil  and  rock  debris is the  largest  direct source of
particulate.  Wind-entrained  dust concentrations  vary across the continent,
and they  vary according  to  weather conditions.   For example,  in the Great
Plains  region of the  United  States, wind erosion  of  soil  is estimated to
produce more particulate than other sources in the region, and also  can lead
to  high dust  concentrations over  large  areas.   Although  most duststorms
occur in the  spring,  they can be a  problem in  other seasons.   Rainfall and
soil erosion  of  an  area also influence  the frequency and severity  of dust-
     The contributions  of volcanoes  and forest fires  to  particulate levels
can vary greatly.   During most years, emissions from  volcanoes do  not con-
tribute  a   large  proportion   of  the  natural  particulate  emissions;  but in
episodes such as  the eruption of Mount  St.  Helens,  Washington, starting in
May 1980, volcanic  particulate  is a major component of .regional particulate
levels.   Although  the  contribution  of  forest fires   cannot  be accurately
determined, forest  fires  are  important  to  urban  air  quality  because  such
fires are frequently adjacent to urban areas.3
     Ocean salt spray, probably the largest particulate emission source, has
a limited effect that extends only a short distance inland.
2.2.2  Manmade Sources of  Particulate
     Manmade  sources contribute  less to  overall  particulate  levels  on  a
nationwide  basis  than natural sources (Figure 2-1), but  they  are  important


     § 25



     § 20




Figure 2-1.  Estimated average contributions  to  nonurban TSP levels.

with  respect to  air  quality  because  they are  usually concentrated  near
population centers where they are generated.  Such anthropogenic particulate
is the  predominant component  of particulate  levels  in many  urban  areas.3
Manmade  particulate sources  are sometimes  grouped  into  four  broad cate-
gories:    stationary  fuel  combustion,  industrial  processes,  solid  wastes
disposal, and other significant sources.  Table 2-1 summarizes the estimated
contributions from each of these categories.5
     The concepts  of  fugitive  emissions and fugitive dust must be addressed
in this discussion.   Both of  these terms  refer  to nonstack emissions  of
particulates.   Fugitive emissions  result  when particulate  from industrial
operations finds  its  way to the atmosphere through building vents, windows,
doors, and leaks in hooding and ductwork or when particulate is emitted from
the open-air loading and transfer of materials.  Fugitive dust, on the other
hand, is an emission  that becomes  airborne  by the  forces  of  the  wind  in
combination  with man's  activity.    These  emissions  include windblown  dust
from  construction sites,  paved and unpaved  roads,  tilled  farmlands,  and
raining  operations.   Fugitive dust,  then,  usually  originates from nontradi-
tional  sources,  but can  originate  from natural causes  such as  duststorms. Particulate Emissions From Stationary Combustion Sources.  Par-
ticulate matter  emitted from  stationary  combustion  sources  represents ap-
proximately  35  to 50  percent  of the total   particulate generated in the
United  States  by anthropogenic  sources.2'3    Combustion source  particulate
includes  fly ash,  soot,  and  sulfur oxide  aerosols.   Fly ash  consists  of
inorganic material  from the  fuel,  which is  not destroyed  during combustion
and is subsequently entrained with the flue gas.  Fly ash includes inorganic
oxides,   salts,  and trace  metals.6   Soot  consists  o«f unburned  carbon  par-
ticles  and  polycyclic  organic  compounds  formed under  oxygen-deficient  or
low-temperature combustion conditions.  Sulfur oxide aerosols are emitted in
gaseous form from combustion sources that burn fuels containing sulfur.   The
sulfur oxides often  condense into aerosols, which can contribute to overall
particulate  levels.   Particulates  formed  from precursor gases  after their
emission  into  the  atmosphere,  such as sulfur oxide aerosols,  are termed
"secondary pollutants."

_ Emission sources _ Tg/yr

Stationary fuel combustion

Utility boilers                                                  3.4
Industrial boilers                                               1.2
Residential, commercial, institutional boilers                   0 . 2

     Subtotal                                                    4.8

Industrial processes

Chemicals and petroleum refining                                 0.3
Metals refining                                                  1,3
Mineral products                                                 2.7
Miscellaneous industrial processes                               1. 1

     Subtotal                                                    5.4
Solid waste disposal
     Subtotal                                                    0.4
Other significant sources

Transportation                                                   1.1
Forest fires and agricultural burning                            0.6
Miscellaneous                                                    0.1

     Subtotal                                                    -1.8

     Total emissions  from all sources                            12.4

     Stationary  combustion  sources  that  emit  particulate matter  include
utility boilers,  industrial  boilers,  residential space heating, and commer-
cial and  industrial  space  heating.   Utility boilers account  for  more than
half of the  emissions  in  this  category,  followed  by industrial  boilers
(Table 2-1).   The actual  quantities  of particulate  matter emitted  from  a
stationary  combustion  source are  dependent  on the size of  the source,  the
efficiency  of  combustion,  the efficiency of collection, and the  amounts of
sulfur  and  ash in the fuel.   Fuels commonly  used  in stationary combustion
sources include  coal,  oil,  natural  gas, and  wood products.   These  fuels,
ranging from low-grade, high-sulfur coals to clean-burning natural  gas, vary
considerably in their sulfur and ash content.   Emissions From Industrial Processes.   Particulate  emissions
from  industrial   processes  are formed  primarily through the  mechanisms of
grinding,  impaction,  breakup of  liquids,  condensation, and  chemical reac-
tion.  These emissions are characterized by a wide  range  of particle sizes
and  chemical   compositions.   The  particles  are  chemically related  to  the
processes from which  they are generated.  Thus,  smelting  and  metallurgical
operations,  for instance, produce a large proportion of submicrometer-sized,
condensed metallic fumes.  Carbon particles and organic condensables such as
tars   are   emitted  from  chemical  operations   related   to  the   textile,
petroleum/petrochemical, and plastics industries.
     Grinding procedures such as those used in rock crushing, flour milling,
and  sanding   operations   produce  predominantly  large  particles.   High-
temperature processes  and those  using  volatile  compounds  generally produce
aerosols of submicrometer  size as a result of the condensation of vaporized
solids or liquids.  Such processes include pyrolysis, vaporization of lubri-
cating  or  process oils,  and metallurgical  operations.  Other sources that
produce substantial amounts  of submicrometer particulate emissions are lime
kilns,  pulp-mill   recovery furnaces, and cement and asphalt plants.
     Industrial-process fugitive  emissions  have  the potential  to contribute
significant quantities to particulate burdens, especially in the vicinity of
the source.   With stack emission controls improving, the relative importance
of  fugitive emissions  is growing.  Because  these emissions  generally enter
the  atmosphere near  ground  level  and  at  low velocities,  their  localized

Impact on air  quality  can be large.  Major  industries  with potential  fugi-
tive emissions are listed in Table 2-2.2'7   Particulate Emissionsfrom Solid Wastes Disposal.    Incinera-
tion  and  open burning were  traditionally the most common  methods  of  solid
waste disposal in  urban  areas.   Under the pressure of more stringent pollu-
tion  regulations,  large  municipal  incinerators and industrial installations
have  been upgraded and controlled.   With the  continuing  trend toward  land-
fills, recycling,  and  the use of combustible  rubbish  as  a fuel  substitute,
the decline  of particulate emissions from solid waste  disposal  is  expected
to continue.3   Other Significant  Particulate  Sources.   The remaining sources
of particulates that could be considered significant may be grouped into two
categories.    One  consists  of   sources  created  directly  by some  manmade
action, such  as  emissions from vehicular exhausts, vehicular tire wear, and
construction  and  demolition activities.   The other  category  consists  of
re-entrainment  sources,  both  natural  and  vehicle-related.   Particulate
matter can  accumulate on  city streets  and  other paved  areas from unpaved
roads and lots,  truck spillage,  sand and salt applied for snow control, and
sediments washed  over roadways during heavy  rains.   These particulates can
become re-entrained by heavy winds and by vehicular traffic, and can produce
an area-wide source of temporarily suspended particulate.3
2.2.3 Transported Particu1 ate
      As indicated  in Figure 2-1, transported particulate represents a major
portion of the average nonurban particulate  levels in the continental United
States.   It  is also evident that the  importance  of transported particulate
increases  as  air  masses  move  with the  prevailing  wind  patterns  from the
Pacific Ocean  across the continent.  Transported particulate  is both manmade
and natural, and consists  of both primary and  secondary particulates.  Move-
ment  of  particulate from  the  source over a distance  less  than  100 km from
the monitoring site is considered to be  short-range transport; movement over
more  than 100  km is considered long-range transport.
      The  formation and  transport of secondary particulate warrants special
attention because the highest concentration  of  secondary particulate  is in
the  size  range  of 0.01  to  1.0  urn.8'9   Particulate  in  this  size  range is


   Food/agricultural industry

        Alfalfa dehydrating
        Cotton ginning
        Grain terminals
        Grain processing

   Metallurgical industry

        Primary and secondary aluminum
        Metallurgical coke
        Primary copper
        Primary and secondary lead
        Gray iron and steel foundries
        Primary and secondary zinc
        Secondary brass/bronze

   Mineral products industries

        Asphalt concrete
        Castable refractories
        Concrete batching
        Coal cleaning
        Phosphate rock
        Stone quarrying
        Potash production
        Sand and gravel
        Diatomaceous earth

   Forest products industry

        Lumber and furniture

active in scattering  light,  and is also inhalable into the alveolar area of
the lungs.  The  main  ingredients in formation of  secondary  particulate are
sunlight and gases  such  as sulfur oxides, ammonia,  nitrogen oxides,  ozone,
water  vapor,   hydrocarbons,  and  oxygen.   The  sulfur  oxides and  nitrogen
oxides are  primary pollutants  emitted  in large  quantities  from combustion
sources  and internal  combustion  engines,  but can  also  be precursors  of
secondary participate matter.   Secondary participates  arising  from  these
precursors  include  nitrates  such as nitric acid,  sulfates  such  as  ammonium
sulfate  and sulfuric  acid, and organic particulates formed by the reactions
of volatile organic compounds in the presence of sunlight.

     Particulate matter emitted from point sources may be measured to deter-
mine  compliance  with applicable  emission limitations,  to  evaluate control
equipment performance, or  to establish emission factors.  Many  of  the test
methods,  however,  are subject to biases  that may  influence  the validity of
the  results.   The  test procedures  discussed here  have been  developed to
minimize  or eliminate  these  biases in  obtaining  representative  samples.
2.3.1  Mass Concentration  Measurement
     The  most  precise method of determining the  mass  concentration of par-
ticulate  matter  in a gas  stream  is  to  collect the entire volume of gas and
the  particulate  matter and  to determine  the  mass  concentration  from this
sample.   This  procedure,  however, is feasible only with a few sources  (with
very  low volumetric flow  rates).  Procedures for  sampling small portions of
a  gas  stream  to  obtain a  representative  sample of the total  gas stream have
been  developed  by various  groups.   Examples of these procedures  are EPA
Reference  Methods 5  and  17,   American  Society   for Testing  and Materials
(ASTM)  Method D2928-71,  and the  American Society  of  Mechanical Engineers
(ASME)  Power  Test Code 27.  The  predominant  test procedure for characteri-
zation  of particulate matter  is  EPA  Reference Method 5,  "Determination of
Particulate  Emissions   from  Stationary  Sources,"  Appendix A, 40 CFR 60.
Quality  assurance checks   in Method 5 and use of the method  with EPA Methods
1,  2, 3, and  4  help ensure the  accuracy of mass concentration determina-

     Method 5 is based on extractive filtration.   Gas is extracted isokinet-
ically; that  is,  the  velocity  of the  gas entering the  sampling nozzle  is
equal  to  the gas  velocity passing  by  the nozzle  at  that  sampling  point.
Extraction is done  through a nozzle to an  externally  heated filter held  at
120°±14°C.  The particulate  matter  is  captured in the  sampling probe and  on
the  filter,  and  then the filtered gases  are passed  through a  series  of
impingers to remove moisture and other components before they pass through a
dry gas meter.  The sampling apparatus is  shown  as  Figure 2-2.   Isokinetic
conditions must be  maintained within ±10 percent of 100 percent for a valid
test.  In a  gas  stream with both large  and small particles, sampling rates
lower  than  100  percent  isokinetic  can bias  the sample  toward  larger par-
ticles, and  can  strongly bias  the mass  concentration  calculations.  The
reverse is true with sampling  rates above  100 percent  isokinetic, in which
the bias  toward smaller  particles would result in  an  apparent mass concen-
tration that is  lower than the  actual emission rates.
     Establishing isokinetic sampling  rates depends on  the characteristics
of the individual  sampling train and on determination of gas velocity, gas
volumetric  flow  rate  (EPA Method 2),  gas  molecular weight (EPA Method 3),
and  gas  moisture  content  (EPA  Method 4).   The location  and suitability  of
the  sampling site  and  the  location of the sampling  points to  provide  a
representative sample of  the   gas  stream  are  performed  according to  the
procedures of EPA Method 1.   Thus  the  use of  EPA Method 5 depends on the
proper use of-other EPA test methods, each of which affects whether the mass
concentration data  will  be  representative of the  actual emissions from  a
stationary  source.   A  brief review of these methods  and their  uses with
Method 5 is necessary in evaluating test results.
     The EPA Method 1 specifies criteria for selecting the sampling location
and the location and number of sampling points.   Emphasis is on locating the
sampling locations away from flow disturbances such as  fans, bends in ducts,
and  duct  expansion or contraction  points.   The duct is  divided into  equal
areas, and sampling points are  at the centroid of each  area.  Generally, the
closer a  sampling  location  is  to upstream or downstream disturbances, the
more  sampling  points  are needed  to obtain  a representative  sample.  The
existence of cyclonic flow is checked  because angular velocity patterns can
lead  to   erroneous  velocity determinations and  to nonisokinetic sampling
conditions, which can result in biased mass concentrations.


                    HEATED AREA   \
                               \    lp   FILTER HOLDER

                                                                          '-SILICA GEL
                                                                 VACUUM LINE
                                                      "kifcixti-r**"* L*~m, I/

                                        O O  BY-pAJS VALVE

                                    DRY GAS METER   AIR-TIGHT
                    Figure 2-2.  EPA Method 5 participate sample apparatus.

     The  EPA Method 2 is  used  to  determine  local  velocity  pressures  for
establishing  isokinetic  sampling rate.   Average gas velocity and volumetric
flow rate may be calculated from a traverse  of all  sampling points.   Typi-
cally,  this   method  uses a  Type S pitot tube because  it yields  a  higher
reading at a given velocity pressure than a standard pitot tube and because
it  is  resistant to plugging.  A  Type  S pitot tube is shown  in Figure 2-3.
     The  EPA Method 3 is used  to determine  gas  molecular weight, a  value
needed  in determining gas  velocity and volumetric flow rate.   On combustion
sources, an  Orsat  analysis  for oxygen, carbon  dioxide,  and carbon monoxide
is  typically performed  to  assist  in  determining  excess air  and F-factor
calculations  for heat input  and mass  concentrations.   Values obtained  can
aid  in  the  determination of  representative  source-operating  conditions  and
in  the  calculation of mass  emissions  in  units specified  by various  stan-
     The  EPA Method 4 is  used  in  the determination  of moisture  in  stack
gases.  Although the  moisture content is determined from the impinger catch
of  Method 5,  a  value  for  moisture content  must be  assumed  for isokinetic
flow  calculations.   For  gas  streams  with  low moisture  content  (e.g.,
3 to 10 percent),  the errors  in isokinetics caused  by  assuming  the  wrong
moisture content are relatively  small.   At  high moisture content (greater
than 10 percent),  however,  a small error in the estimated moisture can lead
to  sampling  rates   outside the  acceptable range of 90 to  100 percent.   For
example,  if  the estimated  moisture was  50  percent and the  actual was  45
percent, the  calculated  isokinetic  rate would be 111 percent.  Method 4 may
be  used  to  aid  in establishing  a  moisture  content   value  for use  in
Isokinetic sampling calculations.  When high moisture content is encountered
during particulate sampling, care should be taken to properly heat the probe
and filter to avoid premature condensation.
     For  many  sources,  the  amount  of particulate  matter captured on  the
filter is a function of temperature.  In-stack filtration methods allow fil-
tration to occur at approximately the same temperature as that of the stack
gas.  Thus,  the amount  of  particulate captured  should  vary  among sources,
depending on  the stack temperature and on the  degree  to which the particu-
late  is temperature  dependent.   By  use  of  an external filter  of defined
filtration characteristics  at  120°±14°C, the  captured  particulate on  the


1.90-2.54 CM
(0.75-1.0 IN.)
              TYPE S PITOT TUBE
           Figure 2-3. Type S  pilot tube and manometer.

filter and  in  the  heated probe is defined.   This  is the "front-half catch"
of  the Method 5 sampling train,   Condensible  participate allowed  to  pass
into  the  impingers  along with water  vapor  is  the "back-half catch."   In
certain jurisdictions,  the  back-half catch is included  with  the  front-half
catch as total particulate.
     Significant biases may  be  introduced into the  sampling  results if the
particulate is strongly temperature  dependent and if the sample is not held
at  the proper temperature.    Even  though the  filter holder  is placed  in a
heated box  within  the 120°±14°C range, excessive heating  or  cooling in the
sampling probe can affect the results.  Of most concern is excessive probe
heating, which can be caused either by a  high  probe-temperature  setting or
by  a  stack  gas temperature  higher than 120°C.   Although the  heated box may
be  at  the proper temperature, the actual gas  stream filtration temperature
may  be much higher.   An  excessive  probe  or  filter temperature  is  of more
concern in  compliance tests,  since the high  temperature could  tend to bias
the mass concentration low.
     Where  there  is  no temperature dependency,  EPA Method 17  uses  an in-
stack  filter  for particulate capture.   As in  Method 5,  the sampling rate is
isokinetic;  EPA  Methods 1, 2, 3, and 4 are used with the  particulate samp-
2.3.2  Particle Size Analysis
     As part of the particulate emission characterization, a distribution of
particulate sizes  may  be useful   for  determining control  equipment param-
eters.  The cascade  inertia!  impactor is the device most  commonly used for
particulate sizing.   The  sampling  train consists of the probe,  a precutter
such as a cyclone,  and the cascade impactor.
     The  cascade  inertia!  impactor  technique  provides  a distribution  of
aerodynamic particle diameters.   A cascade  impactor usually  has  5  to  10
stages  of decreasing orifice diameters.  The impactor is  usually assembled
to  give an  alternating pattern of orifice plates and collection plates.  As
the orifice size decreases,  the gas velocity through each orifice increases.
Larger particles cannot overcome the inertia!  force imparted to them through
the  orifice and thus  impact the  collector plate.   Smaller  particles  have
less  inertia,  and  so  the gas  stream  carries them to the  next stage.   The

last stage is usually followed by a filter to capture the smallest particles
that have  escaped impaction.   Gravimetric  methods are  used  in  analysis  of
each stage  to determine  particle  size distribution,  geometric  mass median
diameter,  and  geometric   standard   deviation.    The  results  of  cascade
impactors are influenced by the deposition of particulate in the probe.   For
example, one  test indicated that at a velocity of 15 m/s, 33 percent of the
10-umA particles were collected in the probe.10
     Cascade  impactors  are typically in situ  (i.e.,  in-stack)  devices  used
with isokinetic  sampling  rates.   When samples are obtained  in  situ at the
stack  temperature,  the particle size  distribution  should be representative
of  the actual particle  size distribution  in  the duct.   Failure to sample
isokinetically  results  in  the  particle size  distribution  being biased and
unrepresentative.   A  bias  toward  larger  particle  sizes occurs  with under-
isokinetic  sampling  (velocity  entering nozzle is  lower than the localized
gas  velocity),  and  bias  toward  smaller  sizes  occurs  with overisokinetic
     Cascade  impactors  are provided  in stages with nominal values for aero-
dynamic  cut-size  diameters.   Calibration  procedures are usually provided by
the  impactor  manufacturer.   Each impactor should be calibrated periodically
to  determine the actual  value of  the cut-size  diameter  for  each stage.
     Cascade  impactors  are susceptible to  several  problems.   First, in gas
streams  with  high particulate loadings, material may build up on the stages
quickly  and  thus shorten  the  available  testing period.   Second,  particle
reentrainment and bounce can result in the particle size distribution being
prejudiced  toward  smaller particles.   Finally,  fracturing  of  the larger
particles  at  the  impaction stage may  lead to generation of fine particulate
and  to a consequent bias toward small particle sizes.
     Although in  situ cascade impactors are probably more common, extractive
external cascade  impactors  are  also  in use.  The particle size distributions
obtained with  these  models  are representative  of the  temperature of the
impactor when there is  temperature  dependency of the particulate.   Extrac-
tive cascade impactors  may  be used  in  observing  temperature  effects  on
particle size distribution  and particle  growth.  The  results  from in situ
and  extractive  cascade impactors should not be  combined, because the samp-
ling conditions  are  often  different.   Sample losses  to the walls of the


sampling probe  are a  potential  problem with extractive  samplers.   The ef-
fects of over-  and underisokinetic sampling are  similar  with  both types of
     Cyclones  are  used  for  in  situ and  extractive aerodynamic  particle
sizing, but  not  to the extent of cascade impactors.  The aerosol sample en-
ters the cyclone through a tangential inlet and  follows  a vortex flow pat-
tern.  Particles  that  cannot  follow the gas streamlines move outward toward
the  cyclone  wall  and,  depending  on  cyclone geometry,  gas flow  rate,  and
particle size,  may reach the cyclone walls and be collected.    By  use  of a
series  of  cyclones  of  different  geometric dimensions  at a  constant  flow
rate, particles  can be  removed  according  to  size from  a  gas stream.   The
fractionating capability of cyclones is not predictable by theoretical means
to  the  degree of  accuracy possible with impactors.  The advantages  of cy-
clones  over  impactors  is  that  large  samples  can  be  acquired and particle
reentrainment is not so great.
     Realtime particle sizing/counting has  received  minimal application to
characterization  of source  emissions  chiefly because the techniques require
low mass concentrations.   The instruments used in realtime analyses include
optical  devices,  diffusion  batteries,  condensation  nuclei   counters,  and
electrical  mobility analyzers.
     Laboratory  size distribution  analysis  of collected particulate samples
is  often  performed  instead of  in situ procedures.   The results  of these
methods  must  be  interpreted  with  great  caution,  however,  because  the
original flue gas  particle size distribution is almost impossible to recon-
struct  under laboratory  conditions.   Particles  or particle  groups  may be
altered from  their gas-stream state by additional agglomeration or particle
breakup during sample  collection.   Size distribution results based on sedi-
mentation  and elutriation, centrifuging,  sieving, and  electronic counting
are meaningful only when the  effects of sample  collection and redispersion
are negligible or clearly known.
     Microscopic  analysis  is regarded  as  the  fundamental   technique  for
counting and  sizing particles.   This  procedure involves manual or computer-
ized microscopic examination of a prepared slide containing a representative
sample of the aerosol.   Careful  procedures must be followed in  preparing the
slide so  that the  aerosol  sample  is not altered  from  its  in-stack state.


Although, microscopic  examination  of participate matter does  not  yield size
information in terms of aerodynamic diameters, it can produce useful  data on
particle  surface  features,  agglomeration,  size,  composition,   and  shape.
2.3.3  Analysis of Particulate Samples  General Considerations.  Following the collection of a partic-
ulate sample  at  the sampling site, the sample  must be analyzed to obtain a
quantitative  or  qualitative measurement.   Generally,  the  exposed filter or
collection  surface  is returned  to a laboratory  for  analysis.   During this
transfer,  care must  be exercised  to avoid  loss  of  fibers  or  particulate
matter from the  filter or the collection surface  and to  protect the sample
from damage or conditions  that may affect the  analytical  results.   Special
filter cartridges  or  filter holders are often used to safeguard the sample.
Also transferred  with the  sample is an  information  record  containing the
site  and  sampler  identification,  the quality assurance  data,  and  other
pertinent information.   Artifact Mass.   Artifact particulate matter  can  be formed on
the surfaces  of alkaline filter materials, such as glass fiber, by oxidation
of acid  gases in the  sample air.  Formation of artifact particulate results
in an  artificially high particulate  measurement.   This surface-limited ef-
fect usually  occurs  early  in the  sample  period, and  is  a function of the
filter  pH  value  and  the  presence of  acid  gases.   Although  the artifacts
usually  account  for  only  a  small part  of  the  collected  particulate, the
error can  be  significant when sampling periods  are short or when the total
amount of particulate  matter collected is very small.
     Reactions between various particulates collected  on the filter are also
possible.   Although such reactions may not significantly affect  gravimetric
determinations, they  may affect the chemical analysis  of the sample.
     2-3.3.3   Loss  of Volatiles.  Volatile particles  collected on the  filter
may  be  totally or partially lost  during  subsequent sampling, during  trans-
port to  a  laboratory  for analysis, or during storage  before the postexposure
weighing.   Filters  are normally analyzed as  soon after collection as prac-
tical, but  some  loss  of volatiles  is  inevitable.

-------  Gravimetric Determination.   Filters are  conditioned  and then
weighed before and after exposure to determine the particulate weight as the
net weight  gain of  the filter  or  the front-half of  other  collection sur-
faces.  Mass  concentration  is determined directly by dividing the weight of
the collected participate by the total (standard) volume of air sampled.  In
addition to potential  errors described earlier, errors are also possible in
the weighing process.
2.3.4  Chemical Analysis and Analysis of Trace Elements
     Most  methods  for  trace element  analysis  of particulate  material use
spectroscopic detection.  The detectors  respond to the presence  of only an
element and provide  no information about  chemical  compounds.   Most methods
do not indicate the oxidation state of the element.   Atomic Absorption Spectrometry.   Atomic  absorption spectrom-
etry11 is widely used for quantitative elemental analysis of airborne parti-
cles.   It usually involves an acid extraction and excitation of the solution
by a  flame.   Light  with  a  wavelength characteristic of  the  one  element of
interest traverses  the flame.   The amount of  light absorbed  is  related to
the quantity of the element present.
     Individual  elements  must  be  determined sequentially.   Thus,  although
any element can be  determined for which  a lamp is available to produce the
characteristic  light,  particulate  samples  are  often  large  enough  to  allow
only  half  a dozen determinations.   Moreover, some trace elements present in
the particles (including  antimony and arsenic) may  require  the application
of special methods.
     Atomic absorption  is  subject to significant interferences and can lead
to substantial  errors.  If  recognized,  these  errors  generally  can  be ac-
counted for or  eliminated to produce  good  quantitative analyses  at the ex-
pense  of  additional  effort  on  each  sample.  Despite  its drawbacks, atomic
absorption spectrometry is a  useful method for elemental analysis of partic-
ulate.   Optical  Emission Spectrometry.   A  variety  of methods  can be
used  to  excite rather  loosely  bound  electrons  in elements  and  to observe

characteristic emissions as de-excitation occurs.  The wavelength is charac-
teristic of the  element,  and the intensity is an indication of the quantity
of the  element present.   The most desirable technique is argon plasma exci-
tation of an acid extract of the particulate matter.12
     Plasma  spectrometry  offers more  advantages  than atomic  absorption
since  sample  preparation  and  analysis  rates  are  essentially  equal.   The
spectrometry  techniques,  however, can simultaneously  determine  up to,  per-
haps,  50  elements with  detection limits  about the same as  those of flame
atomic  absorption.   Spectrometry is  also more  interference-free than atomic
absorption, although not totally so.  Spark  Source  Mass Spectrometry.   Spark  source  mass  spectrom-
etry13  can analyze particulates separated from  the  filter  and oxidized, or
can  be extracted with  an acid.  Spectral  interferences  are  important,  but
generally  can be overcome  by  use of  a spectrometer  with  high resolution.
     The precision can  be about 30 percent (relative  standard deviation) in
careful analytical work.   As with any multielement  technique, its accuracy
may  depend on the element  and  on the matrix.   The  advantage of this tech-
nique  is  that one can  simultaneously estimate the quantity  of every non-
volatile element  in  the periodic table and do  so with roughly equal sensi-
tivity.  Neutron Activation Analysis.  Neutron activation analysis13'14
consists  of  a variety  of distinct  methods,  all of  which  produce unstable
nuclei  that  emit gamma  radiation.   The  energy and intensity of  the gamma
rays are  indicators  of  the element  and  its  quantity.   Instrumental thermal
neutron  activation   analysis  is  most  commonly  used.   In  this  approach,  a
nuclear reactor  is  used  to produce  unstable  nuclei.   Neutron activation
analysis  can  simultaneously  determine  up  to  25  elements  in  particulate
samples.   Another advantage  is that particles  can  be analyzed as received
directly on the filter surface.   X-Ray  Fluorescence Spectrometry.   X-Ray fluorescence spectro-
metry13'15 involves  excitation  of tightly  bound electrons and  observation of
the  X-ray emission  as  de-excitation occurs.   Excitation  may be  done by a
variety of techniques,   but use of  an  X-ray generator  is  the most common.

The technique may  be either multielement (up  to  perhaps  30) energy disper-
sive  detection  or wavelength  dispersive detection  (up  to perhaps  10 ele-
ments).  Only  elements with  atomic numbers greater than  that of magnesium
can be  analyzed.   Particles can be analyzed nondestructively, directly on a
filter;  however,  if samples  are not  thin  and of  uniform surface texture,
certain  corrections must  be  made.    Interferences  are  common and  must be
considered, and adequate calibration can be a problem.   Electrochemical  Methods.  Electrochemical  methods  have been
used  to a limited  extent to  determine  a small  number of elements  in par-
ticles.  These  methods include potentiometry with ion-selective electrodes,
polarography, and  anodic stripping  voltammetry.13  Electrochemical  methods
have few advantages for particle analysis aside from the low initial  capital
costs of equipment relative to that needed for other techniques.  Chemical Methods.  Many wet chemical procedures constitute the
classical methods used for trace element analysis of particulate.  In gener-
al, a  color-forming reagent is used,  and the amount of an element is deter-
mined by the  extent of color development.  Probably the best known of these
procedures  is   based  on  the  use  of  dithiocarbazone  (dithizone)16  as  the
colorimetric reagent  for  lead.   Wet chemical procedures are labor-intensive
and  slow,  compared  with  spectral  techniques,  particularly since only  one
element can be  determined at a time.  Interferences can  also be a problem.   Analysis of Organics.   Procedures   for  estimating  the  total
mass of benzene-extractable organic material in particulate matter have been
used occasionally.  A  portion of the front-half catch is placed in a Soxhlet
extractor and  refluxed with benzene for several  hours.  Then the benzene is
volatilized,  and  the  mass of  the  residue  is  measured.   This  procedure
presents  problems  because  of  special  requirements  for  handling benzene.
     Methods  of  identifying  and  determining  individual organic  species
abound.  These  methods use different  sequences of solvent extractions that
separate groups of  different  organic species  on the basis  of solubility.
Solutions are  often subjected to chromatographic separation with mass spec-
tral  detection.   For organic compounds that are  volatile  up to about 300°C,
gas  chromatography-mass  spectrometry  (GC-MS)  can be used.1T   For  organic

species with  lower  volatility,  liquid chromatography might be  used.   High-
performance liquid  chromatography (HPLC)18  is  typically used, but  none  of
these procedures permits a high rate of analysis.
     For analysis  of one  species of  longstanding  interest,  benzo-a-pyrene
(BaP), thin layer chromatography  (TLC) with fluorescence detection is often
used, and  HPLC  procedures  have been proposed.   The TLC procedure  requires a
cyclohexane  extraction,  spotting,  and  development  of  a  TLC  plate,  with
fluorescence detection.  This  procedure  is more interference-free than some
HPLC methods, and it has a higher production rate.19


 1.  U.S. Environmental Protection Agency.  National Air Pollutant Emissions
     Estimates, 1970-1978.   EPA-450/4-80-002.   Research Triangle Park, N.C.
     January 1980.

 2.  Adamson,  L.   F.,   and  R.  B.  Bruce.   Suspended  Particulate Mattel—A
     Report  to Congress.   U.S. Environmental  Protection  Agency,  Environ-
     mental  Criteria  and Assessment  Office.   Research  Triangle Park, N.C.
     October 1978.

 3.  U.S. Environmental Protection Agency.  National Assessment  of the Urban
     Particulate  Problem,  Vol.  I.    EPA-450/3-76-024.   Research  Triangle
     Park, N.C.  July 1976.

 4.  U.S.  Environmental Protection Agency.   Priorities and  Procedures for
     Development  of  Standards of Performance for  New Stationary Sources of
     Atmospheric Emissions.   EPA-450/3-76-020.  Research Triangle Park, N.C.
     May 1976.

 5.  U.S. Environmental  Protection  Agency.   National Air Quality Monitoring
     and   Emissions   Trends   Report,   1977.    EPA-450/2-78-052.   Research
     Triangle Park, N.C.  December 1978.

 6.  Faoro,  R.  B.,  and T. B.  McMullen.   National  Trends in Trace Metals in
     Ambient   Air,   1965-1974.    U.S.   Environmental   Protection  Agency.
     Research Triangle  Park, N.C.  EPA-450/1-77-003.  February 1977.

 7.  U.S.  Environmental Protection Agency.   Technical  Guidance for Control
     of  Industrial  Process  Fugitive  Particulate  Emissions.   EPA-450/3-77-
     010.  Research Triangle Park, N.C.  March 1977.

 8.  Whitby, K. T., R.  B. Husar, and B. Y. H. Lia.  The Aerosol  Size Distri-
     bution  of Los Angeles  Smog.   J.  Aero.  Atmos. Chem.  (ed.  G. H.  Hidy)^
     Academic Press, New York,  1971.

 9.  Clark,  W.  E.,  and K.  T.  Whitby.   Concentration and Size  Distribution
     Measurements of  Atmospheric Aerosols and a Test of the Theory of Self-
     Preserving Size Distributions.  J. Atmos. Sci., Vol. 24, 1967.

10.  Smith,  W.  B.,  et  al.    Sampling  Data  Handling Methods  for Inhalable
     Particulate  Determination.   Southern Research Institute.    Birmingham,
     AL.  March 1981.

11.  Ahearn,  A.  I.   Trace  Analysis by  Mass  'Spectrometry.   Academic Press,
     New York, 1972.                                                     I


12.  Fassel, V. A.  Quantitative Elemental Analyses by Plasma  Emission  Spec-
     troscopy.  Science, Vol. 202, 1978.

13.  Morrison, G.  H.   Trace Analysis Physical Methods.  Wiley Interscience,
     New York, 1965.

14.  Lambert, J. P. F., and F. W. Wilshire.  Neutron Activation Analysis  for
     Simultaneous  Determination  of Trace  Elements  in  Ambient  Air  Collected
     on Glass-Fiber Filters.  Anal. Chem., Vol. 51, 1979.

15.  Goold,  R.  W.,  C. S.  Barrett, J.  B.  Newkirk, and C. 0. Ruud.  Advances
     in X-ray Analysis.  Kendall/Hunt, Dubuque, Iowa, 1976.

16.  Snell,  F. D.  Photometric and Fluorometric Methods  of Analysis,  Metals.
     Part I.  John Wiley, New York, 1978.

17.  McFadden,  W.  H.   Techniques  of  Combined Gas Chromatography/Mass  Spec-
     trometry.  John  Wiley, New York, 1973.

18.  Kirkland, J. J.   Modern Practice of  Liquid Chromatography.   John Wiley,
     New York, 1971.

19.  Swanson,  D.,  et  al.   A Rapid Analytical  Procedure for the  Analysis of
     BaP in  Environmental Samples.  Trends Fluoresc., Vol. 1,  1978.


                                 SECTION 3

     Control of  particulate emissions may be achieved  either  by prevention
of particle generation  or  by collection of particles entrained  in effluent
gas streams.  Prevention is preferable, both economically and environmental-
ly; however,  opportunities for  this  control  approach  are  limited.   A more
commonly  available  strategy  is  process  modification  or optimization  to
improve particle  collectibility.   The costs of  particulate control  systems
can be reduced,  and their  reliability can be  improved by  increasing  the
particle size, reducing the particulate mass loading, or reducing the vari-
ability of process  operation;  these measures  can  be combined  to  improve
control.   It  is  sometimes possible  to  modify particle  characteristics  to
maximize collectibility.
     These  control  measures represent  alternatives  to,  or supplements  to,
installation  of  conventional  particulate control  systems.   The following
sections deal  with  these  broad issues.   Fuel  switching  is the  most common
and  the  most useful  means  to "prevent"  particulate emissions.   General
approaches  to process  optimization  illustrate  possible  benefits  to  sub-
sequent control system performance.   Finally, the control devices are brief-
ly  addressed.  More detailed  information on particulate  control  systems is
in  Section  4,  and summaries of  important processes  are in Volume 2 of this

     Energy substitution  can  be  an  effective and useful  technique  for re-
ducing  particulate  emissions  from stationary combustion sources.  Substitu-
tion  has  special value in  control  of small  and old  sources,  for which the
cost of effective  control  devices might be expensive relative to the worth
of  the  facility.   Application of this approach is contingent on fuel avail-
ability and on the reduction in emissions to be achieved.

     The particulate reduction  potential  may be estimated initially by com-
parison of published emission  factors for various fuels.   Because  of site-
specific  combustion  characteristics  and highly  variable  fuel  properties,
such emission factors can provide only an estimate of the emission reduction
potential.   An  emission  factor analysis  is presented  in Table 3-1  for  a
pulverized-coal-fired power  boiler.1   The use of natural  gas  reduces emis-
sions by 50 percent relative to No. 6 fuel oil and by 80 percent relative to
coal firing  with low-efficiency collection.   Other clean  fuels  now avail-
able, such as distillate oil and refuse-derived fuels, could be included in
a  similar analysis.   Because   of  the limited  supplies of  naturally clean
fuels,  this  control  option is  severely restricted to use with only marginal
boilers or furnaces where gas cleaning is not economically feasible.
                               ENERGY SOURCES
Energy source
Natural gas
No. 6 fuel oil
Bituminous coal
particulate emission
control , %
Particulate emission,
ng/J (lb/106 Btu) of
delivered energy
21 (0.048)a
47 (0.108)b
140 (0.320)
        aBased on an emission factor1 of 15 lb/106 ft3 for gas with a
         heating value of 37,300 kJ/m3 and on an estimated conversion
         efficiency of 31.4 percent.
         Based on an emission factor1 of 5 lb/1000 gal for oil to
         0.3 percent sulfur content with a heating value of 42,000
         kJ/liter and on an estimated conversion efficiency of 30.8

     Development of synfuels will ultimately provide a more plentiful supply
of clean fuels.   Many synfuels, such as high-Btu gas and liquid solvent-re-
fined coal, are  believed to have combustion characteristics and particulate
emission rates  similar  to those  of natural gas.   Many of  these products
could possibly be used  without add-on particulate  control  devices.   Other
synfuels may have  low ash content, but undergo  moderate carbon losses that
demand some degree  of gas cleaning.  Little information is yet available on
particulate emission rates of synfuels.

     Substitution  of  electric-powered devices  for onsite power  generation
may provide  relief from  control  requirements.   This  option again  is  most
appropriate  for  small,  economically  marginal units.   In this option,  the
particulate  control requirement  is simply transferred to the  power genera-
tion facilities,  most of which are equipped with high efficiency particulate
control systems.

     Process optimization  involves  modifications  of process feed materials,
process  unit  functions,  and  process  variables  to  eliminate  or  reduce
particulate emissions.
     Optimization  of  the process may reduce  the  volume  of  exhaust gases or
alter  the  particle size  distribution.   A change  in particle  size distri-
bution  may allow  a broader selection of  abatement  equipment,  and emission
standards  may be  achieved  by  application  of  equipment with  lower energy
     Selection of  a process and implementation of process modifications may
affect  other  process  requirements,  yields,  or  nonparticulate  emissions.
Therefore, in evaluation of process changes as means of particulate control,
the total  impact of the changes must be considered.
3.2.1  Modification of Process FeedMaterials
     The  physical  properties  of  feed  materials,  such  as particle  size,
chemical  composition, and  moisture,  may  have  significant  effects on emis-
sions  from industrial processes.   The effects vary from process to process,
and the  relationships must be developed on  a process-by-process  basis.   In
general, where heated air is used for drying, fine particles in the process
feeds  cause  an  increase in particulate emissions.  Screening or cleaning of
raw  materials can reduce  the  particulate  emissions  per  ton of product.
Following  are examples  of process modifications  that can  reduce emissions
resulting  from the properties of feed materials.  Phpsphate  Rock Dryers.   The  screening or washing of phosphate
rock before  drying can reduce the weight of fine materials in the feedstock.
These  fines, referred  to as phosphatic clay,  have  a  substantial  impact on
the  emission rate from  uncontrolled  dryers.  The  particulate  loadings in


dryer exhaust gases at several facilities has varied from 1.14 x io3 to 8.00
x 10s kg/ms, depending on the condition of rock being processed.  Removal of
fine materials  from  the feed can permit operation  of the control device at
lower energy consumption and higher efficiency.  Secondary Lead and Copper Smelters.   Operation  of cupolas and
blast furnaces  at lead and copper smelters is  affected  by  the composition
and size of feed materials.  The presence of fines in the charge reduces the
furnace thermal  efficiency,  and requires the addition of excess coke.  This
results in  higher  combustion air pressure and flue gas volume.  Fluctuation
of the  charge  composition and the inability of  the furnace  to respond rap-
idly can cause  severe swings in furnace temperature and exhaust gas volume;
these changes  can create  positive pressures at  charge  doors, and generate
fugitive emissions during charging.   Control of the size and composition of
the  feed  can reduce  the  capacity requirements  for abatement  equipment to
contain and control the fugitive emissions.  Textile Fabric Coating.  Curing  of chemical coatings and fin-
ishes on  textile substrates  normally leads to  the release  of condensible
hydrocarbon  aerosols.   The condensed  organics can exhibit  high opacity at
relatively  low  particulate  loading.  The  emissions  normally  result  from
chemicals  incorporated into  the  fibers  in  previous processes.   Removing
these components prior to treatment has been shown to be effective in reduc-
ing particulate  emissions from these processes.   Application of latex coat-
ings  to  textile  fabrics that  contain  surfactants  and plasticifiers  has
resulted in  high opacities,  and has necessitated the use of afterburners to
control  emissions.  The need for afterburners can be reduced by modification
of the  chemical  composition of the surfactants  used  in  the  coating system.
In one  instance, the  surfactant was  commercially  available  ammonium stea-
rate.  The emission was composed of oleic and palmitic acids, which had been
introduced  into  the process  as impurities in the  ammonium  stearate.   When
the  stearate was  converted to a purer form and  the usage  rate reduced,
opacity of emissions dropped from 100 to 25 percent.
3.2.2  Elimination of Process Steps
     The manufacture  of products can  require  many  individual  process steps
involving simple  functions.   The transfer of materials from  one process to


another can result  in  fugitive particulate losses.   The loss of product can
increase the cost of the finished marketed material,  and  also necessitates
the application  of  pollution abatement devices.  Often, a careful  analysis
of the  number,  order,  and  types of process  steps can enable  a  company to
reduce the number of emission points and to reduce emissions by eliminating
repetitive and wasteful handling of materials.
     The  process changes  many  include installation  of   longer  conveyors,
transfer of product by pneumatic instead of open conveyors,  or combination
of process steps.  In view of the energy required to transport materials and
due to the cost of ductwork and abatement equipment, the elimination of even
a single emission point can be cost effective.
3.2.3  Changes In Process Particle Characteristics
     The particle size  of  the product being processed or handled can have a
direct effect on  the particulate emission rate.  The  wetting or agglomera-
tion  of  materials  can  increase  the  effective  particle size  and the effi-
ciency of control equipment.   An example of change in particle characteris-
tics  that can  reduce  emissions  is  the  transfer  of  wood fibers  by using
cyclones and pneumatic systems in the fiberboard industry.   The uncontrolled
emission of fibers from a cyclone handling 5000 kg/h of fiber can be as high
as 300 kg/h.  Prior to transfer, partial polymerization of the heat-setting
resin that  coats the  fibers can reduce  the  emission to  less  than 5 kg/h.

     In  a number of industrial  processes, particulate emissions cannot be
controlled  satisfactorily  by fuel   switching  or process  optimization.   In
such  processes,  abatement of emissions to within  the regulatory limits is
usually  achieved by adding  exhaust gas cleaning  devices.  Among  the many
devices  available for  exhaust gas cleaning are cyclones,  multicyclones, and
other  mechanical devices;  shaker-type, reverse-air,  and  pulse-jet fabric
filters;  hot-  and   cold-side electrostatic  precipitators;  spray  chamber,
venturi,  and  packed-bed scrubbers;  and incinerators.  Each of these devices
operates  according  to  one  or more of the basic physical or chemical princi-
ples  discussed   in  Section  4.1,  and  each has  distinctive  advantages  and
disadvantages, as discussed in Sections 4.2 through 4.6.


     Selection  of  a control  device  may involve a complex  set of variables
including regulatory limitations; the nature of the emissions source and its
exhaust gases;  the  removal  efficiency of each  device;  and  long-term relia-
bility, ease of maintenance, and total costs of installing and operating the
device.   Useful information  is  available  in  journals and  other technical
literature;  in U.S.  EPA publications;  in  publications  of  control  device
vendors and  their  representatives;  through trade associations, professional
organizations, and their'.technical committees; and through paid consultants.
References 2  through  8 are examples of many publications which can provide
initial direction  in  the search for information about exhaust gas cleaning.
3.3.1  Applicable Regulations
     Air pollution  control  regulations  vary with jurisdiction, and are sub-
ject to periodic revisions.  Some regulations are promulgated at the Federal
level:   for  example,  New  Source Performance  Standards  (NSPS),9  National
Emissions  Standards  for  Hazardous  Air   Pollutants   (NESHAPS),10  National
Ambient Air Quality Standards (NAAQS), and Prevention of Significant Deteri-
oration (PSD).  Other air emission regulations are promulgated at the State
or local  level:   for  example, those  in State  Implementation Plans (SIP}.11
Certain regulatory  and  enforcement  functions related  to air emissions are
retained  at  the  Federal  level,  but many  enforcement functions  have been
delegated to the States.  Some States, in turn, have delegated much of their
authority  to  municipalities,  counties, or regional   air quality agencies.
     The first step in determining regulatory requirements for a particular
installation  is to  determine which local, State,  and  Federal agencies have
promulgated  applicable  regulations.   These agencies should be contacted for
preliminary  advice  on the applicability and interpretation of current regu-
lations.   If  analyses show that fuel switching and process optimization are
not  feasible,  the  technical  and economic  analyses  of the  various  exhaust
gas cleaning devices must be begun.
3.3.2  Source  Characteristics
     The characteristics of the source must be defined as completely as pos-
sible to select the most appropriate control device.  Single value estimates
of exhaust gas flow  rates, temperatures,  and  particulate  loadings usually
are not necessarily sufficient for the reliable design of a control device.


characteristics of  a source that  are  important to the design  of  a control
device  also  include  the process  operating  schedule,  variability  of  fuel
quality,  the  type  of the  raw  materials,  and  the variability in process
operations.  These characteristics determine the variability of such exhaust
gas parameters as temperature,  moisture content, gas flow rate, particulate
concentration, particulate size distribution, and concentrations of chemical
constituents.   A  control  device  must  be  selected to provide  the desired
efficiency and reliability  over the anticipated range of exhaust gas condi-
tions.  For example,  the fabric in a fabric filter system must be chosen to
withstand  both the  expected high and  low temperature excursions  and the
typical or average temperature.
3.3.3  Control Device Design Limitations
     Each  type of device is limited in its capability to remove particulate
from exhaust  gases.   In  general, the particulate removal efficiency of each
device  is specific  for  each set  of  operating conditions.   To provide the
desired  level of abatement,  a  device must  be properly  matched  for the
process  conditions,  the design must  incorporate a  sufficient  factor  of
safety  to handle  unexpected contingencies, and  the design limitations must
not be exceeded because of increases in production rates after the device is
installed.   Mechanical Collectors.    Mechanical  collectors  are  efficient
for large  particulate, but they cannot  be expected to adequately remove fine
particulate.  Nor can they be expected  to perform well on processes in which
flow  rates are extremely variable.  A  large  increase  in gas flow rates may
increase  the  efficiency  of mechanical   collectors, but may also reduce their
life  by increasing abrasion.   In contrast,  a large  reduction  in gas flow
rates will significantly decrease  the efficiency.  Electrostatic  Precipitators.  Electrostatic precipitators must
be  sized  to accommodate the expected gas flow rates, particulate concentra-
tions,  and fly ash  resistivities.  The specific  collection  area (SCA) must
be  great   enough to  handle  the  expected range  of  conditions.   Exhaust gas
loadings  and  fly ash resistivities must  not be  altered  to such  an extent
that  the   SCA  limitations  of the  precipitator  are  exceeded.   In general,  a

properly designed electrostatic precipitator will collect nearly all partic-
ulate greater  than  1 nn>A in diameter, but participate penetration increases
as particle size decreases.   Fabric Filters.   The principle  limitations  of fabric filters
are  the temperature limitations  of  available  fabrics  and the  effects  of
fabric  failures  on  the penetration of particulate.  A properly sized fabric
filter operating under dry conditions and within the temperature limitations
of the  fabric  can provide extremely efficient  collection  over a wide range
of particle sizes.   Immediate replacement of broken filter bags is important
in maintaining a low particulate emission rate.  Wet Scrubbers.  A variety of wet scrubbers are available, each
with certain performance  limitations.   Most wet scrubbers can operate effi-
ciently  in  collecting large  particulate (>2 pmA); some  are also efficient
for very small particulate (<0.2 jjmA).  Therefore, within many wet scrubbers
there is a medium  particle size  "window"  for which  removal  is  less effi-
cient.    Particulate  collection  generally decreases as energy input decreas-
3.3.4  Control Device Reliability
     The  ultimate  purpose of  a  normally  efficient exhaust  gas-cleaning
device  can  be compromised  if it  suffers from  frequent  malfunctions.   Mal-
functions can  reduce the  performance of control  equipment  or  cause periods
of  uncontrolled  emissions;  on occasion,  plant  production must  be  halted
while the device is repaired.   Malfunctions are  caused  by design deficien-
cies such as  undersizing  or omission of important ancilliary features or by
improper operation and maintenance.  Many of the common malfunction modes as
well as design  features  and  operating and maintenance  procedures that can
reduce  the  frequency  and  severity  of malfunctions  are described  in Sec-
tion 4.
3.3.5  Control Equipment Costs and FinancialAssistance
     A financial  management decision regarding the purchase and operation of
an exhaust  gas-cleaning device  should be aimed at selecting  a device that
will provide efficient reliable service over the desired service life at the
lowest  possible  total  annualized cost.   An annualized  cost  determination


must consider the  amortization  of control device investment, the direct and
indirect costs of  operating  and maintaining the device, any credits for re-
covered particulate,  and the effects  of the device, if  any,  on production
rates.12'13'14  Tax relief or special financing schemes  that could favorably
affect the total  annualized cost are sometimes available.13'15  Installed Costs of Control Equipment.  The  installed  costs of
control equipment  include  the  costs of engineering and designing the equip-
ment, the costs of materials of construction, cost of manufacturing, cost of
transportation, and costs of labor and equipment during installation.  Other
costs associated with  equipment installation may include the  costs of pro-
duction loss during  installation  and the costs  of  an  initial  stack test to
verify performance of the device.  Direct and Indirect Operating Costs.   Direct  costs of operat-
ing  a  control  device include the costs  of  utilities (e.g., electric power)
to  run  the equipment  and  the  costs of  labor  for  operating the equipment.
Other costs can include those for disposal of any collected particulate  (if
the collected particulate has economic value, this can be a negative cost or
a  net  savings),  the  costs of  periodic  stack tests, and the  costs of lost
production due to control device malfunctions.  A particulate control device
may  also  have a  positive or  negative effect  on  production rates  of some
processes.   Maintenance costs  for  a  control  device  include the  costs of
replacement  parts,  of  maintaining  a  spare  parts  inventory,  and  of labor
associated  with  routine  preventive maintenance or  emergency maintenance.
     Indirect  operating costs  include overhead, parts  replacement, insur-
ance,  taxes,  and  amortization  ("capital   recovery")   of   the  investment.
Overhead  is  traditionally  expressed as a percentage of operating and main-
tenance labor, whereas  the other indirect costs  are normally  computed as a
percentage of the control device investment.


 1.   U.S.  Environmental  Protection  Agency.    Compilation  of  Air Pollution
     Emission Factors.  AP-42, 1977.

 2.   U.S. Environmental Protection Agency.  Office of Administration, Infor-
     mation Division.  Need  Air Pollution Information?  Promotional Litera-

 3.   Industrial Gas  Cleaning Institute.   IGCI - Who We  Are and What We Do.
     Promotional Literature.   Stamford, Connecticut.

 4.   Catalog  and  Buyer's  Guide -  Pollution  Equipment  News.   Rimbach Pub-
     lishing, Inc., Waseca, Minnesota, Published annually.

 5.   Equipment Buyer's Guide Issue - Chemical Engineering.  McGraw-Hill, New
     York, N.Y.  Published annually.

 6.   Air  Pollution  Control  Association.   Journal of the Air Pollution Con-
     trol Association.  Pittsburgh, Pennsylvania.  Published monthly.

 7.   Pollution  Engineering.   Technical  Publishing.   Barrington,  Illinois,
     Published monthly.

 8.   Pollution Equipment News.  Rimbach Publishing, Inc., Waseca, Minnesota.
     Published bimonthly.

 9.   U.S. Environmental Protection Agency.  Standards of Performance for New
     Stationary  Sources—A  Compilation.   EPA-340/1-77-015,  October  1977.

10.   U.S. Environmental Protection Agency.  National Emissions Standards for
     Hazardous Air  Pollutants—A  Compilation as of April 1, 1978.  EPA-340/
     1-78-008, 1978.

11.   U.S.  Environmental  Protection  Agency.    Air  Pollution  Regulations in
     State Implementation Plans.  EPA-450/3-78-050 through EPA 450/3-78-104,
     56 documents, updated annually.

12.   U.S.  Environmental   Protection  Agency.   Choosing  Optimum  Management
     Strategies—Pollution Control Systems.  EPA-625/3-77-008, 1977.

13.   U.S.  Environmental   Protection  Agency.   Choosing  Optimum  Financial
     Strategy  for Pollution  Control Investments,   EPA-625/3-76-005,  1976.

14.  Neveril, R. B.   Capital  and Operating Costs for Selected Air Pollution
     Control Systems.  EPA-450/5-80-002, 1980.

15.  U.S. Environmental  Protection  Agency.   Major Financial Assistance Pro-
     grams  Available for  Industrial  Pollution Control  Expenditures.   EPA-
     340/1-77-023, 1977.


                                  SECTION 4
                         PARTICULATE CONTROL SYSTEMS

     Since  "Control Techniques for Particulate Air Pollutants1"  was  origi-
nally published  in 1969,  there  have been substantial advances  in  the capa-
bilities  of  particulate  matter  removal   devices.   This  section  presents
information on the many sophisticated and  diverse  control  systems  presently
in  commercial  use.   Each  of the  five  major categories  of systems  is  dis-
cussed  in  terms   of  types  available,   basic  operating  principles,  design
factors,  and  operation and  maintenance  considerations.   Particle collection
capability with respect to particle size is emphasized.

     Introductory  information  is presented on  particle behavior and charac-
teristics.  Each  particle control  device  takes advantage of one or more of
the  particle  aerodynamic  properties to remove particulate matter  from the
gas  streams.   Basic considerations in the  application of particulate matter
control devices as part of a system are also discussed.
4.1.1  Particle Characteristics and Behavior
      It  is helpful  to  understand basic principles  of particle behavior in
order  to  design a control device,  to measure  particulate  matter concentra-
tions,  or to evaluate  control  device performance.   More extensive informa-
tion is available  in  aerosol physics texts such  as references  2 through 10.  Particle Size  and  Shape.   There are a wide variety of particle
shapes.   Spherical   particles  are  usually  generated  in  high  temperature
processes  and  in  some  cases the particle is partially or  completely due to
the  condensation  of  vapors  as the  gas  stream cooled.   Small  particles can
link with other particles to yield a flocculant.   Such  flocculants tend to
be  fragile and  can  break apart during sampling or during  passage through
control  devices.    Fibrous particles result  from the processing of certain


natural  biological  and mineral  materials.   Asbestos is one  type  of fibrous
particle.  Depending  on  the chemical  composition of the parent  material,  it
is  also  possible to  generate a  flake-type  aerosol.   Mechanical  attribution
type processes  such  as grinding, sawing, and polishing can yield  irregular-
ity  shaped particles.   Such  aerosols  are  generally  larger  than particles
formed by condensation of vapors and other common types of aerosols.
     Liquid aerosols  are  comprised almost exclusively of spherical particles
or  flocculants  of  spherical  particles.   In certain  processes  it  is  also
possible  to  generate  solid  particles  covered with an outer  layer of  liquid
material.  Solid  aerosols can  occur  in  any of the forms shown and in less
common forms  such as  cubes and rods.   Factors  governing  the ultimate shape
of  the  particles include  the chemical composition  of the material and the
characteristics of the process.
     Particle size  is the  most important characteristic  affecting  behavior
in  gas   streams,  and  it  is  a governing  factor in the  extent to which the
particle  scatters  visible light and therefore  contributes to plume  opacity.
The  particle  size  range  of interest  to air  pollution  control  studies  if
generally  from  0.01  to  100  micrometers.   Under the  SI  system  of  metric
units, micrometer  (pm) is  the standard  unit  of particle size, and is 10 6
     The many definitions  of particle  size  are  ultimately based on  the size
measurement method.   For example, microscopic  methods measure the projected
area  of  particles.   The  projected area  can be variously  defined, depending
on  the  means   used  to  convert  the   dimensions of  irregularly  shaped and
fibrous  particles  into  a  single  "diameter"  value.    Likewise,  sampling
methods,   such   as  cascade  impactors  can  be used  to  measure particle size
determined by  observed behavior  in  a  gas  stream.   The various  measurement
methods yield many  size  definitions  that are not necessarily consistent with
each other.
     The  most   common  definition for particulate  control  evaluation  is the
aerodynamic  particle  diameter,  which  is  defined  as  the   equivalent  unit
density  sphere  having the  same  aerodynamic characteristics as  the  actual
particle.  It  is the  product of  the  Stokes diameter times  the square root
of  the  product  of  particle density  times  the  Cunningham   slip  correction

                   dp = dps
(Eq.  4.1-1)
Throughout this document,  particles  sizes are given  in  terms  of aerodynamic
diameter  (pinA)  unless  otherwise  noted.   The  aerodynamic diameter  is  most
closely  related  to the behavior  of  particles in control  devices,  and it is
the "size" normally measured by stack sampling methods.
     Each of  the particle  size  measuring methods yields  sets  of data indi-
cating the quantities  of  particulate in a  number  of size categories.  These
data can be  compiled  into a  histogram,  as shown in  Figure 4.1-1 to graph-
ically  illustrate  the  aerosol  distribution.  The  terms  used  to characterize
the  distributions  of particles  sizes are  illustrated in Figures  4.1-2  and
                             PARTICLE DIAMETER, urn
                    Figure 4.1-1.  Aerosol distribution.
              (Reprinted from:  Silverman, L., Billings, C.E.,
              and First, M. W.  Particle Size Analysis in In-
              dustrial Hygiene.  Academic Press, 1971, p. 237.)
For  example,  the mean and the median are not equal for skewed distributions,
as shown  in Figure 4.1-1.  The  median  particle size, by definition, divides
the  frequency  distribution  in  half;  50  percent  of the  aerosol  mass  has
particles with  a larger diameter, and  50  percent has particles with a smal-
ler  diameter.   A  second  measure of  central  tendency is  the mean, which is
simply  the  sum  of all observations divided  by  the number of size categories
used to construct the histogram; the mean  is  sensitive  to the quantities of
material  at the extremes of  the distribution,  so relatively few large parti-
cles could  shift  the  observed mean to larger levels.

     For many  industrial  sources,  the particle  distribution  approximates a
lognormal . distribution  function.   When  the log  of  the  particle diameter is
plotted  against  the frequency  of  occurrence,   a  normal bell-shaped  curve
(shown in Figure 4.1-2) results, which is characterized by the geometric mean
diametei—the  sum  of the logs  of the observations  divided  by  the number of
size categories.
                   1    2   3    4  5  6 7 8 .9 12 14 18 22
                            PARTICLE DIAMETER, ym
         Figure 4.1-2.  Histogram of a lognormal size distribution.
        (Reprinted by permission from Silverman, L., Billings, C.  L.
       and First, M. W.  Particle Size Analysis  in  Industrial Hygiene.
                       Academic Press, 1971, p.  236.)

     Both  the geometric mean  and the  standard deviation  can  be  determined
easily by  plotting  the distribution data  as a log-probability plot  as  shown
in  Figure  4.1-3.   The  geometric mean  is the  diameter equivalent to the  50
percent  probability,  and  the  standard deviation  is the  slope  of the  line.
The  latter can  be  determined simply  by dividing  the  geometric mean by the
particle size at the 15.78 percent  probability (size d±  in  Figure 4.1-3)  or
by  dividing   the  particle  size at the 84.13 percent probability (size  d2  in
Figure 4.1-3) by  the geometric mean  size.

                  20 -
Mill  I
8  10   16
                             2    3   4567
                             . PARTICLE DIAMETER, urn
           Figure 4.1-3.  Cumulative lognormal size distribution.
           (Reprinted by permission from Silverman, L. , Billings,
           C. E. , and First, M. W. Particle Size Analysis in In-
             dustrial Hygiene.  Academic Press, 1971, p. 239.)

     Aerosol distributions  may  exhibit more than one peak.  The hypothetical
distribution shown  in  Figure 4.1-4 is called  bimodal.   This type of aerosol
distribution is  more difficult  to characterize.   In  some  cases,  it may be

possible to  handle the  distribution  as two  separate lognormal distribution
aerosols.   More  detailed discussions of aerosol  distributions are presented
in references 4, 11, and 12.
                              4 "   6  "   8    10    12    14"
                               PARTICLE DIAMETER, ym
           16    18
                Figure 4.1-4.   Bi-modal aerosol distribution.

-------   Aerodynamic Properties.    Each  type  of  participate   control
device uses one or more particle collection mechanisms.   Those are the funda-
mental physical tools available to an equipment designer.
     Impaction.  Inertia! impaction  is  the mechanism most frequently used  to
remove particulate  matter.   Particles  have  a much greater mass,  and there-
fore much  greater inertia when  in motion, than the  surrounding  gas.   Heavy
particles  resist changes  of  gas  flow and  cross  gas  streamlines  because
of their inertia.   As  a gas stream approaches an obstacle, the gas molecules
pass on  either side  of it,  leaving  the  particle  propelled toward the obsta-
cle by  its inertia.   If the particles  are too small they flow with  the gas
molecules  and  pass  around  the  obstacle.   Figure  4.1-5  illustrates  these
phenomena  relative  to  different  sized  particles.    Low-energy  impaction
conditions  separate  particles  with  aerodynamic diameters  above  50 p.mA.
High-energy impaction  conditions effect separation  of particles  with  aero-
dynamic diameters above  a few tenths of a micrometer.   Particles with aero-
dynamic diameters below  a few tenths of a micrometer can not be separated  by
Inertia! impaction under normal conditions.
                                                    Water Droplet
                                                     ,	Stream!ines
  Figure 4.1-5.  Impaction of particles on a target in a moving gas stream.

     Efficiency of impaction collectors is related to an impaction parameter:
a ratio of  drag to viscous forces.
to relate  impaction efficiency.
                                     The  Stokes  number,  K,, is commonly used
                                                                 (Eq.  4.1-2)
                               I    18 M Dr

    Kj = Stokes number.
     C = Cunningham slip factor, dimensionless.
    d  = particle diameter, umA.
    p  = particle density, g/cm3.
     v = particle velocity, cm/s.
    D  = diameter of collector, cm.
     M = gas viscosity, kg/m - s.
The impaction mechanisms become progressively more effective as the impaction
parameter increases.   The  strong  particle size dependence of inertial  impac-
tion is indicated by the fact that impaction is proportional to the square of
the particle size.
     Particle  separation  from  a  gas  stream  can also  occur by a  mechanism
known as  interception.  The  particle  is intercepted  by  an obstacle  if  the
particle radius  is  as  large  as or larger than  the  streamline displacement.
Interception results  in an  increase  over the amount  of  particle  collection
that is predicted by impaction alone.   Interception is  similar  to,  and  can be
considered a form of impaction.  As  illustrated in Figure  4.1-6,  interception
occurs when particle  size  and gas streamline displacement  values  are  compa-
rable.  Particles in  the  size range above a few micrometers are susceptible
to interception.
     Diffusion.   Particles  in the same size range as molecules (10~3 urn) and
 up  to  a few tenths of  a  micrometer experience random movement due to colli-
 sions  with  gas  molecules.    The diffusion  rate of  a  particle (D )  can  be
 determined by the Stokes-Einstein equation:
                           D  =   CKT
                            p   3?iudp
(Eq.  4.1-3)
    D  = diffusivity of particle, cmVs.
     K = Boltzman constant, (g - cm2/s2 - °K).
     T = absolute temperature, K.
The  effectiveness  of this  particle collection mechanism  is  proportional  to
the  particle diffusivity  which  is inversely  proportional  to the  particle
size,  as  indicated above.   Thus high  diffusion  rates occur  for  very  small
particles (0.001 to 0,1 urn), but diffusion is negligible for large particles.
Particle diffusion is illustrated in Figure 4.1-7.
                                                            Gas  strearr.l i res
 Figure 4.1-7.  Diffusion of  a particle to a target  in  a moving  gas  stream.

     Settling.  Movements of  particles in  the  atmosphere are influenced by
two main  forces:   gravity and drag.   When'particles are  settling, gravita-
tional force  is pulling downward  and  drag is pushing upward.  After suffi-
cient time,  the forces become equilibrated,  and the  particles  reach their

terminal  settling  velocity.   To determine  the  settling  velocity  (V.)  for
ideal situations, the Stokes terminal velocity equation can be used:

                              dO (PD " Pa^
                         V=  P        9                        (Eq. 4.1-4)
As with  inertia!  impaction,  the effectiveness of settling is proportional to
the square of the particle diameter.  This mechanism becomes important in air
pollution control systems only when the particle size is above 5 (jmA and when
the gas flow conditions are not highly turbulent.
     El ectrostat 1c Attracti on.  Particles are charged with unipolar ions, and
are  subjected to  a  strong  electrical  field.   Movement  of particles  to  a
collection  surface  is primarily  dependent  on the  balance  between the elec-
trostatic  force  and  the  aerodynamic  resistance.   Random  diffusion charging
contributes  to  the  initial  charging  of the  particles.    Effectiveness  of
electrostatic attraction  is  basically related to the square of the particle
aerodynamic diameter  because larger particles can sustain a greater number of
charges.  Physical Phenomena.  Condensation and agglomeration can alter a
particle  size distribution  and  thereby  have  a significant  impact  on  the
aerodynamic  behavior of  the particles.  Light  scattering  is  important be-
cause  opacity  of  the stack effluent  is sometimes used to evaluate the effec-
tiveness of the particulate control systems.
     Condensation.    Condensation  of  water  vapor  on  suspended  particles
occurs  whenever  a  degree  of supersaturation is available  around the parti-
cles.   In  industrial  gas cleaning,  three  methods are available  to effect
particle  growth  by  condensation:   (1)  mixing  produces  supersaturation  by
combining  two saturated  gas  streams of different temperatures;  (2) steam
injection  introduces  steam into the gas  stream, and  (3) an adiabatic expan-
sion  method,  which   is  available in  venturi  scrubbers.13   A psychrometric
chart  indicates  the  appropriate temperature  levels  and  amounts  of water
vapor  available for use in particle growth by condensation.
     The  main benefit from  condensation of water  vapor on particles is the
increase  in  their  mass  and size,  which makes  them  easier to  collect in
inertial  removal  systems.   Collection  can  be further enhanced  by taking


advantage of  forces  acting on the particles  induced  by a temperature gradi-
ent (thermophoresis) and vapor condensation (Stefan flow).14  Superimposition
of these effects can improve collection.   Agglomeration.   When particles collide  with  each other during
diffusion  or  turbulent motion  they  may adhere  to   each  other  and become
combined  or  agglomerated  particles.   The  rate  of   agglomeration  depends
mainly  on particle  concentration and  is virtually independent  of particle
size, as shown by the following differential equation:

                            - — = kN2                           (Eq.  4.1-5)
     k = rate constant.
     N = number of particles per unit volume at time t.
Integration of  Equation 4.1-5 from time  zero to time t for an initial parti-
cle concentration (N )  is:
                            i - i  = kt                          (Eq.  4.1-6)
Agglomeration rate constant  values (k) can be  estimated  for homogeneous and
heterogeneous systems by the following:
          homogeneous system:  k = ~	                          (Eq.  4.1-7)
        heterogeneous system:  k = ——                          (Eq.  4.1-8)
for reference  purposes, the  homogeneous  rate constant for air at 20°C is 3 x
10"10 cms/s.   Values  for agglomeration rate constants are also influenced by
pressure  and  turbulence conditions.   Increased pressure and  greater turbu-
lence  separately enhance  the probability  of  particle collision, and  thus
increase the agglomeration rates.7  Light Scattering.  When  white light strikes a suspended parti-
cle, a  certain  amount of light (depending  on  particle size, shape, composi-
tion, and surface  configuration)  is scattered irregularly in all directions.
Particles scatter  light in  a degree proportional  to  their sizes.  Supermi-
crometer  particles  scatter   light proportional  to diameters  to  the  sixth


4.1.2  Selection and Application of Particulate Control Devices
     Each application of  a participate control system is unique to a degree.
No general  selection  method can guarantee an environmentally and economical-
ly  acceptable  installation.   Instead,  it  is  necessary to  carefully  match
process  effluent  characteristics  with  regulatory  requirements  of  control
device  performance  capabilities   and  control  system  costs.    This  section
introduces  some  of the  general  issues involved in the selection and applica-
tion of  parti cul ate  control systems.   More detailed information is presented
in Sections 4.2 through 4.6.   Performance Capabilities.    Most particulate  control  devices
operate  as  combinations of the particle  collection  mechanisms discussed in
Section  4.1.1.   One  or  more mechanisms may be operative in any given device;
the  limitations of  such  mechanisms  become  the  performance  limits of  the
device.   Table  4.1-1 presents  the mechanisms  normally  active  in  the  major
types  of particulate control  devices.   The  effectiveness  of the mechanisms
depends  on the  design  of the  unit   (e.g.,  gas  velocity,  device geometry,
liquid utilization rate).   Performance is typically characterized in terms of
collection  efficiency:
                       inlet mass   outlet mass
The  penetration fraction  (P+),  defined  in  Equation  4.1-10, is  a somewhat
simpler measure  of control  device performance; and  it  is related to collec-
tion efficiency as indicated in Equation 4.1-10.
          Penetration = outlet mass loading = ^ff^^/wQ  (Eq. 4.1.10)
                        inlet mass loading               J        H
The  penetration term  is  easier to use when  evaluating high efficiency con-
trol  devices and  series  of particulate control  devices,  and when using the
computerized  models  developed  for  certain types of collectors.   Due to the
size  dependence  of the particle collection mechanisms  the collection effici-
ency  (or  penetration)  is a size-specific value.  A set of values for various
particle  sizes is  a penetration curve.
4.1.3  Control System Design
     Proper  selection of  particulate control  systems  requires  simultaneous
consideration   of  regulatory  requirements,   performance  limits,   effluent

                         PARTICIPATE CONTROL DEVICES
  Control device
  Principal particle
  capture mechanism
Particle size
Settling chamber
Momentum separator
Large-diameter single
Small-diameter multiple
Fabric filters
Electrostatic precipitator
Wet scrubber
gravity settling
gravity settling
inertia! separation
inertia! separation

inertia! separation
impaction on dry surfaces
diffusion to dry surfaces
electrostatic attraction

gravity settling
impaction on surfaces
impaction on liquid
diffusion to wetted surfaces
diffusion to liquid droplets
particle oxidation
     d 2 and
 Based on particle capture mechanism.

characteristics,  and  cost.   The  control  device must have  the  capability to
maintain  continuous  compliance regardless of  short-term  fluctuations  in the
effluent composition, flow rate, and particle size distribution.
     Control Device Sizing.   Most particulate control  devices  (for example,
baghouses and  electrostatic  precipitators can suffer performance degradation

at high effluent  gas  stream velocities.  A fundamental design problem is the
sizing of  the  control  device to balance  the  need for a  large  unit  (low gas
velocity) with the  capital  cost.   The consequences of  undersizing  (high gas
velocities) can be  frequent noncompliance periods.   The  sizing  of  a control
device should  also accommodate anticipated operational  changes;  for example
increases in process throughput and/or effluent gas temperature  can lead to
inadequate efficiency.   A  related  problem is  failure  to properly  size the
support systems,  such  as solids removal equipment,  by inadequately assessing
the effluent  conditions such  as  particulate mass  loadings  or  bulk density.
     Instrumentation.   Control  devices  installed without proper instrumenta-
tion may  be  prone  to  frequent malfunctions  and excessive  emission periods.
Examples are  a fabric  filter on a  high-temperature  source  without tempera-
ture monitors  and a wet scrubber on a combustion source without pH monitors.
Proper instrumentation  is  necessary to provide an early warning of impending
problems and to assist in diagnosis of underlying factors.
     Accessibility.   No  particulate  control   system  is completely mainte-
nance-free.   Control   devices  are  normally  subjected  to multiple  physical
insults  including  abrasion,   corrosion,  mechanical  shocks, moisture,  high
temperature,  and  high  voltage.   The designer and purchaser  must decide what
additional capital  cost is justifiable to  minimize  future  maintenance costt
     Power Input  -  The collection  efficiency of a particulate control system
is  generally   associated  with  power input.   As power  input  increases the
particulate emissions  usually  decrease.   What emission  level  is affordable
and justifiable,  given the energy demand?   This question  is complicated by
uncertainty over  actual effluent gas  stream  characteristics (e.g.,  particle
size distribution)  and the lack of site  specific  empirical models relating
power input to collection efficiency.
     Corrosion  -   Catastrophic  failure  due  to  corrosion  can  result  from
inaccurate assessment  of gas  stream conditions during steady-state or start-
up conditions.  A contributing factor can  be  failure  to consider the varia-
bility  of effluent vapor  concentrations  caused by  variability of process
operations.  To the degree economically feasible, particulate control equip-
ment should  be designed  and  fabricated  to withstand worst-case conditions.

     Abrasion -  Large particles  suspended  in a  fast moving  gas  stream are
abrasive.  Fabric  filters are  particularly vulnerable to  abrasion  near the
gas inlets.  Precleaners  can be installed,  but usually increase both capital
cost and energy demand (increased fan energy).
     Moisture and Freezing -  Moisture and  freezing can adversely  affect wet
scrubbers  and  any  control  device  using  compressed  air  (pulse-jet  fabric
filters  and  electrostatic precipitators  with compressed air  rappers).   The
solution to moisture problems  in dry collectors are  the  inclusion of dryers
on  air  compressors  and  drains  on  air  reserve  tanks.   Wet  scrubber  lines
should  have  drainage capability, particularly when operation  is noncontinu-
     Ventilation Systems  -  The two  basic parts of a  ventilation  system are
the hood or  air intake for initial capture of the particulate matter and the
ductwork for  transport of  the particulate-laden gas  stream to the control
device.   Inadequate design  of a  ventilation system  can compromise overall
     The hood must be sized and oriented to  capture  the  maximum quantity of
particles  without   requiring  excessive  gas  volumes  (a  trade-off  between
performance  and energy  consumption).   It  makes little  sense to install  a
high-efficiency control  device if  a major  portion of the  particles are not
captured initially.   The  hood  should be as close as possible to the point of
generation without  interfering  with equipment movement;  it should  be  ori-
ented  to minimize  cross-drafts  and to  take advantage  of thermal  drafts.
     The ductwork   leading  from the  hood  (or pickup  point) to the control
device  must  be  sized  to provide  the   needed  transport  velocity—generally
between  15 and  25  m/s, depending on particle size distribution.15  Layout of
ductwork should minimize energy  losses indicated by  static pressure drops,
and  should minimize air inleakage.  If the  source  is hot,  insulation may
minimize temperature drops.
     Solids Removal  Equipment  - The  basic   functions  of  the  solids removal
equipment are to remove collected particulate from the device as  rapidly as
it  is  collected and to deliver the material to an environmentally acceptable
disposal  area.   Solids  removal  is  one of  the  most  frequent  problem  areas
affecting particulate control equipment.


     Most  particulate  systems  operated  at  elevated  temperature must  use
hopper heaters,  insulation,  or  weather enclosures, or  combinations  of these
to keep the  collected particulate hot to maintain free flow.   Hopper temper-
ature control can also reduce corrosion resulting from condensation.
     Delivery  of the  solids  to a  disposal  site or  to a  temporary  storage
site must  be  done  without resuspension of the  material.   As  with inadequate
hood capture,  a seemingly  small  degree of  resuspension can  compromise  any
gains  achieved  by  a high-efficiency  collector.   This  often  happens  when
solids are discharged with  a significant free fall  (directly  below a dis-
charge valve  or between  two  conveyors) or when a temporary  storage  pile is
unprotected from winds.
     Fans  -  Movement  of particulate-laden  gas  stream  from  the  process,
through the  control  device,  and out the stack  is controlled by the fan.  Fan
selection  is  critical to  proper  operation of  the overall  system.  Although
the  forward  curved  design  has  high efficiency, it is  vulnerable to partic-
ulate buildup  on the blades so it is very rarely used in parti cul ate control
systems.    The  backward-curved design  also has  relatively  high energy effi-
ciency, and  it is  also susceptible to particulate buildup.   This type of fan
is normally  used only on the "clean  side"  of the particulate control device
to provide induced  draft.   The most rugged type is  the  radial  blade design
fan  which  can withstand  high dust  loadings  without  excessive vibration and
therefore, can be used on either the clean or the dirty side.
     There can be  fan problems  with  fans  initially  selected  properly.  They
must provide  adequate gas flow and static pressure despite nonideal ventila-
tion  system  design  or fan inlet  configuration.   Air  inleakage  can lead to
lower-than-expected  gas  stream temperatures  and  hence  to higher horsepower.
     In  summary, the  design  of a  control  device involves the balancing of
numerous  factors.   The design  decisions must  be  made specifically  for the
source.  The ultimate  success  of the  control  system  depends  at least par-
tially on  a  realistic evaluation of  the characteristics  of the effluent gas
stream and the entrained particulate matter distribution.   System Operation  and Maintenance.  Continuous  compliance with
air  pollution control  regulations depends  largely  on proper  operation and

maintenance  such  as  preventive  maintenance  programs,  recordkeeping  pro-
cedures,  and operator  training programs.  An  important aspect  of  training
is full awareness of potential safety problems.
     Prevention Maintenance -  An  integral  part  of a  preventive maintenance
program is  routine  inspection of equipment.  Depending  on  the  potential  for
malfunctions and  the consequences of excess emissions  (toxicity, quantity),
inspection  could be  done daily, weekly, or monthly.  For most equipment,  the
internal  and  external   conditions of  the equipment  must  be   noted  during
     An  adequate  spare  parts   inventory  should  be  maintained to  prevent
excessive  downtime  or  excessive  emissions while  operating  under nonoptimal
conditions.  Determination of  what  is  necessary  depends  partially  on  the
local  availability of  supplies  and on  the costs  of  parts relative  to  the
cost of equipment nonavailability.
     Recordkeeping - The logical first step in recordkeeping is to make sure
that  instruments  are  properly located  and are  functioning  normally.   Most
instruments, even  supposedly simple devices, require calibration.   There is
little sense in  faithfully keeping records that are incorrect or misleading.
     Only  the  data that must be  evaluated to determine developing  problems
or nonoptimal  conditions  need  to be recorded.   If large  quantities  of  un-
necessary data are logged, the meaningful data may be lost.
     In addition  to  the normal  operating  records,  diagnostic  records  should
be recorded  during  forced outages or routine maintenance periods to indicate
the type and location of component failures (e.g., bag failure location, dis-
charge wire  breakage  type and location).   Such diagnostic  records  can be as
simple  as  copies  of repair  work orders  with comments by  maintenance per-
sonnel .
     Training  -  Particulate  control  systems  are complex  and  expensive.
Operator training  is  helpful  to ensure proper and safe operation.   Training
should emphasize  procedures for  startup and shutdown  to minimize  damage to
the unit.   Early signs of developing problems should be stressed.
     The importance of  safety training for operating and maintenance person-
nel cannot  be  overemphasized.   A particulate control system may represent a
combination  of  a very  large  number  of  potential   hazards,  including  oxygen


deficiencies, toxic  gases,  high temperatures, high voltages, hot  dust  (hop-
pers),  high  noise  levels,  and  moving machinery.   As  a minimum,  training
should address confined  space  entry procedures,  selection and use  of protec-
tive equipment,  and safe work practices.


 1.  U.S. Department  of Health, Education, and Welfare.  National Air Pollu-
     tion  Control  Administration.   Control  Techniques  for  Particulate Air
     Pollutants.  AP-51, January 1963.

 2.  Fuchs,  N.  A.   Mechanics  of  Aerosols.   Pergamon Press,  London,  1964.

 3.  Davies,  C.  N.    Air  Filtration.   Academic Press,  Inc.,  New York,  1973.

 4.  Silverman, L., C.  E.  Billings, and M. W. First.  Particle Size Analysis
     in Industrial Hygiene.  Academic Press,  Inc., New York,  1971.

 5.  Green,  H.  L.,  and W.  R.  Lane.  Particulate  Clouds, Dusts,  Smokes, and
     Mists (2d ed.).   Van Nostrand, Princeton, New Jersey, 1964.

 6.  Davies, C. N.   Aerosol Science.  Academic Press, Inc., 1964.

 7.  Hesketh, H. E.   Understanding and Controlling Air Pollution.  Ann Arbor
     Science Publishers, Inc., Ann Arbor, Michigan, 1974.

 8.  Hesketh,  H.  E.   Fine  Particles  in  Gaseous  Media.   Ann  Arbor Science
     Publishers, Inc., Ann Arbor, Michigan, 1977.

 9.  Drinker,  P.,  and  T.  Hatch.    Industrial  Dusts (2d  ed.).   McGraw-Hill,
     New York, 1959.

10.  Hidy, G.  M.,  and  J.  R.  Brock.  The  Dynamics  of  Aerocolloidal Systems,
     Vol. 1.   Pergamon Press, London, 1970.

11.  Yeshida, T.,  et al.  Particle Growth  by Condensation and Coagulation—
     Basic Research  of its  Application to Dust  Collection.   In;  Symposium
     on  the  Transfer  and  Utilization  of  Particulate  Control  Technology,
     Vol. I.   EPA-600/7-77-004d, February 1979.

12.  Calvert, S.,  J.  Goldshmid, D.  Leith, and  N. Jhaveri.   Feasibility of
     Flux Force/Condensation  Scrubbing for Fine Particle  Collection.   EPA-
     650/2-73-036,  October 1973.

     Mechanical  collectors  comprise  a broad  class of  participate control
devices  that  utilize  the  gravity  settling,  inertial,  and  dry  impaction
mechanisms.  Because  the;ir performance capability  is  limited to relatively
large particles  and  regulatory requirements have become more stringent, use
of mechanical  collectors  has  gradually declined and they  are now used pri-
marily  as precleaners.   Mechanical  collectors  are reasonably  tolerant of
high dust  loadings,  are  not susceptible to frequent malfunction if properly
designed  and operated, and are adequate control devices  for some applica-
     There is  great  diversity in the design and operating principles of the
various  types  of mechanical collectors.  Most penetration performance data
for mechanical collectors  were obtained from 1940 through 1970.  Since 1970
attention  has  shifted  to more sophisticated particulate  control  devices.
Consequently, only limited field data are available.
4.2,1  Types of Mechanical Collectors
     The  major classes of commercially available mechanical  collectors are
listed in Table 4.2-1.
Particle capture mechanism
     Sett!i ng chamber
     Momentum separator
     Mechanically aided collector
     Inertial centrifugal collector
gravity settling
gravity settling
gravity settling, inertial
inertial collection
inertial collection

The general  characteristics  of these devices are described in the remainder
of this subsection.  Later subsections address operating principles, design,
operation, and performance.  Sett!ing  Chambers.   Large particulate  is  removed by gravita-
tional  settling  in settling  chambers to  protect  downstream equipment from
abrasion and excessive mass loadings.1
     The two basic types are the simple expansion chamber and the multiple-
tray  settling  chamber  (Figure  4.2-1).   The  latter  is a  set of horizontal
collection plates that reduce the distance a particle must fall to reach the
collecting  surface.1'2   Thus  the multiple-tray  unit can  collect somewhat
smaller particles, which settle more slowly.2
                                                                    GAS INLET
             Figure 4.2-1.  Howard multi-tray settling chamber.
          Reprinted with permission of McGraw-Hill Book Company.
      John H. Perry, Chemical Engineers Handbook, 3rd edition.  1950
     The  settling chambers  should be  designed  for  low velocities  with  a
minimum of  turbulence so  that the settling of  particles  is  not disturbed.
Typical superficial  velocities  range  from  0'.3  to 3  m/s.3'4   Gas  stream
distribution across the chamber inlet is important.

-------  Elutriators.   An  elutriator consists of one  or  more vertical
tubes or towers  in  series into which a  dust-laden gas stream passes upward
at  a velocity defined  by the  gas  flow  rate  and the tube cross-sectional
     Large  particles with  terminal  settling  velocities  greater than  the
upward gas  velocity  are separated and collected  at the  bottom of the cham-
ber.  Smaller particles with lower settling velocity are carried out of the
collector.   The  particle size  collected may be  varied  by  changing  the gas
     When size classification is desired for disposal  or reintroduction into
a process,  a series  of collectors may  be  used  with increasing cross-sec-
tional area.  Typical uses of elutriators are in  secondary metal operations,
food and agricultural processes, and petrochemical industries.  Momentum  Separators.   The  momentum separator uses a combina-
tion  of  gravity and  particle inertia  (momentum) to   settle particles onto
surfaces.   The  particles are  separated from the  moving  gas  stream  by pro-
viding a sharp  change in direction of gas flow so that momentum carries the
particles across the gas streamlines and  into the  hopper.
     The  simplest versions  provide  a  90-  to  180-degree  turn  to separate
large  particles5'6  (see  Figure  4.2-2,  a  and  b).  Baffles  can  be added to
increase  the  number of turns  and thereby  provide   a  modest  increase  in
collection  (see  Figure  4.2-2c).
               Figure 4.2-2.  a.  Simple momentum separator.
               Reprinted from:  Alden, J. L.  Design of In-
               dustrial Exhaust Systems, 3rd Ed., The Indus-
               trial Press, New York, 1959.

                Figure 4,2-2.  b.  Simple momentum separator.
                Reprinted from:  R. F. Jennings, J. Iron Steel
                Inst. Vol. 164, page 305, 1950.
             Figure 4.2-2.  c.  Baffle-type momentum separator.
           Reprinted from:  L. Theodore and D. W, Buonicore.  In-
             dustrial Air Pollution Control Equipment for Partic-
             ulates, CRC Press, Inc., page 66, 1976.,
     The  louver collector,  a type  of momentum  collector shown  in  Figure
4.2-3, consists of  a series of flat plates  (blades)  set at an angle to the
gas stream.  A  large portion of the gas  stream is required to make a sharp
turn to pass through the plates.   The momentum of the particles in the air
stream results in movement of the particles in a path parallel to the louver
surface and across the gas stream.   Separation of the particles from the gas
stream leads to concentration of the larger particles in a small portion of

the  gas  stream.7'8   Penetration is  a  function of  louver spacing  and  gas
                                               PARTLY CLEANED
                                               GAS TO MAIN
                                               CONTROL DEVICE
                  10 PERCENT BLEED STREAM
              Figure 4.2-3.  Louvered shutter type collector.
              Reprinted from:  Stairmand, C. J.  Trans. Insti.
              Chem Engro - Vol. 29, page 356, 1951.   Mechanically Aided Separators.   The  separation  mechanism of
mechanically aided separators, like that of momentum collectors, is inertia.
Mechanical  acceleration  of the effluent gas stream increases the effective-
ness  of the inertia  separation  so that these  devices  can collect smaller
particles  than  the momentum devices.   The improved performance, however, is
gained  at  the  expense of  higher energy cost.  Also, the devices are subject
to  abrasion by the  action of large-diameter particles  at medium  to  high


     The  most  common  of the  mechanically aided  collectors  is a  modified
radial blade fan.   The dust-laden air enters the collector perpendicular to
the blade  rotation,  and by  momentum the particles  cross  the  air stream and
concentrate at the  side of  the collector casing.  The rapid acceleration of
the gas  stream  imparted by  the mechanical rotation  of  the blades  maintains
the concentrated particles  in  a narrow  band  which  is  then  drawn  off for
particle separation in a more efficient collector.
     Many  collectors use  this  design principle, and many variations of this
method are used  to  concentrate the particles into a smaller gas volume.  In
addition  to  modifications,  scroll  collectors  and  skimmers are also  used.  Cyclones.   The  cyclone  collector is  similar to  the  momentum
collector  in that  inertia is  used to  separate  the  particles  from a turning
gas stream.  In the cyclone  the gas stream makes one or more circular turns,
followed  by  a 180-degree turn to the  outlet  duct.  Combined  effects  of a
greater number of turns and  higher gas velocity improve the particle collec-
tion capability above that of momentum collectors.
     Cyclones can be classified into four basic categories according to the
methods  used  to remove  the collected dust and to  introduce  the gas stream
into the  unit.9   Figure 4.2-4 illustrates the four types of cyclone collec-
     A vortex  is created within the  cylindical  section of  the cyclone by
either injecting the  gas stream tangentially or by  passing  the gas stream
through  a set  of  spin  vanes.   Because  of  inertia, the  particles migrate
across the vortex  gas  streamlines  and concentrate  near  the  cyclone walls.
Near  the bottom of  the cyclone cylinder the gas stream  makes  a 180-degree
turn,  and  the  particles are discharged either downward or tangentially into
hoppers below.   The treated gas passes upward and out of the cyclone.
     Simple cyclones.  The  simple  cyclone consists of an inlet, cylindrical
section, conical section, gas  outlet tube, and dust outlet tube.   A typical
tangential-inlet,  axial-outlet  simple  cyclone  is   shown  in  Figure  4.2-5.
     Particle separation is  a function of the gas throughput and the cyclone
cylindrical diameter.  Particle inertia increases with increases in gas flow

                        TOP VIEW
                     SIDE  VIEW
a.  Tangential inlet, axial dust     b.  Tangential inlet, peripheral dust
    outlet                               outlet
                 fTOP VIEW
                 SIDE VIEW
c.  Axial inlet, axial dust outlet     d.  Axial inlet, peripheral dust dis-

                  Figure 4,2-4.  General types of cyclones.

Adapted from:  Caplan, Air Pollution, A. Stern, Editor, Academic Press, 1968.



     Medium-efficiency single cyclones are  usually less than 4 m in diameter
and  operate at  static pressure drops of  0,50 to 1.50 kPa.4   Overall collec-
tion  efficiency is  a function  of the  inlet  particle size  distribution.
     Axial—inlet, axial-discharge cyclones.   One common type of axial-inlet,
axial-discharge cyclone  is called the double-vortex cyclone.   The collector
consists  of  an air  inlet at  the bottom  of a  long  cylinder with stationary
turning  vanes.   The  gas stream is  placed  in a  vortex flow  pattern upward
through  the cylinder.   A secondary vortex  is generated by  either stationary
vanes  or injection  nozzles outside of the  inner vortex moving in a downward
motion.   The particles in the inner vortex move across the stream lines  and
into  the   downward-moving  outer  vortex.   The  concentrated  particles  are
separated at the  bottom  of  the  unit  as  the  outer vortex changes direction.
The  flow pattern is  illustrated  in Figure 4.2-6.
                                                   Secondary air pressure
                                                   maintains hiflh
                                                   centrifugal action
                                               i^»c  Secondary air flow
                                               2 ^^creates downward spiral
                                                  ?of dust and protects
                                                 / outer walls from abrasion

                                                  *,Dust is separated from
                                                 "^ gas by centrifugal force,
                                                   is thrown toward outer wall
                                                   and into downward spiral
                                                  , Falling dust is deposited
                                                   in hopper
            Figure  4.2-6.   Flow pattern in a double-vortex cyclone.
                Courtesy of Aerodyne  Development Corporation.

     The  static pressure  drop  of  the  collector is  between 3  to  6  kPa.
Reported  separation  efficiencies are  >99 percent for particles  >6  prnA and
>95 percent for particles >1 pmA.10
     Multiple axial-inlet, axial-outlet cyclones.    A    multiple   cyclone
consists  of numerous  small-diameter  cyclones  operating in  parallel.   The
high-efficiency  advantage  of  small-diameter  tubes  is  obtained  without
sacrificing the ability  to treat large effluent volumes.  A typical  unit is
shown in Figure 4.2-7a.
     The  individual  cyclones,  with diameters  ranging  from  15  to   60  cm,
operate at  pressure  drops from 0.5 to 1;5 kPa.   The inlet to the collection
tubes is axial, and a common inlet and outlet manifold is used to direct the
gas flow to a number of parallel tubes.  A single tube from a typical multi-
ple cyclone is  illustrated in Figure 4.2-7b.   The  number of tubes per col-
lector may range from 9 to 200 and is limited only by space available and by
the ability to provide  equal  distribution of the gas stream to  each tube.
Properly  designed  units  can  be constructed and  operated with  a collection
efficiency of 90 percent for particles in the 5 to 10 umA range.4
     Variation in performance is achieved by the use of several  multicyclone
banks in series or by the withdrawal of 10 to 20 percent of the gas.
     Multiple axial-inlet, peripheral-discharge cyclones.   The  axial-inlet/
peripheral-discharge cyclone is a variation of the multicyclone collector in
that the  gas  is placed in a  vortex  motion by a  fixed vane  and  the  dust is
concentrated at the collector tube wall by inertia!  force.  The central  core
of the gas stream is relatively free of particles and is allowed to exit the
collector tube  axially.   The  outer portion of the gas  stream  (vortex) is
withdrawn by an  induced-draft fan for removal of the concentrated particles
by a more efficient  collector.   The secondary  collector system  is operated
at a lower static pressure than the main air flow to reduce reentrainment of
the concentrated dust (Figure 4.2-8a).
     A  number  of fixed  impeller tubes are arranged in  parallel  and have a
common  inlet  and  outlet manifold.   The  secondary gas  stream  is  passed
through a settling chamber, then the concentrated particles are removed by a
small  high-efficiency cyclone.   The  secondary  gas  volume  typically  con-
stitutes  10  to 20  percent  of the primary gas  stream.  High  collection
efficiency may  be  achieved by using the  collector  tube  banks in series and
parallel  arrangements  to provide optimum flow  volumes in each  tube  (Figure

b. Individual tube
from multicyclone
                                   a. Typical multicyclone collector.
                   Figure 4.2-7.   Multicyclone collector.

   Figure a reprinted from:   Joy Manufacturing Co.  and Figure b reprinted
     from:  Howden, James &  Co.  Ltd.,  195 Scotland  Street, Glasgow;  C-5

efficiency may  be achieved by  using  the collector tube  banks  in series and

parallel arrangements  to provide optimum  flow volumes in each  tube (Figure
          Figure 4.2-8a.  Fixed-impeller straight-through cyclone.
            Annuior du»t tlols
                                              i—j     .
   Figure 4.2-8b.   Bank of fixed-impeller straight-through cyclones with
                     secondary cyclone dust collector.

   Reprinted with permission  of  Davidson and Co.  Ltd..  Belfast Publica-
   cation, Ref. No.  387/61.

4.2.2  Operating Principles of Mechanical Collectors
     Fundamental  operating principles of  the various  mechanical  collector
designs  are  discussed  in this  subsection,  with  emphasis on  theoretical
aspects  of penetration  and  pressure drop.   Information  concerning  these
specific  collector types  can  be  transferred with  a  reasonable  degree  of
confidence to other mechanical collectors of similar geometric configuration
with similar particle capture mechanisms.  Penetration.
     Settling chambers.   The  following  equations  are a condensed summary of
the  approach  presented  by Theodore  and Buonicore.4  A similar  approach is
described  in  Crawford.1   The  performance  limitations  of settling chambers
are  examined  by determining the fraction of  particles  of a given size that
will be  collected during the gas "treatment" time,  tR, defined  in Equation

                                                                 (Eq. 4.2-1)
where     tR = residence  time of  gas  stream  in the  settling chamber,  s
           B = chamber width, m
           L = chamber length, m
           H = chamber height, m
           Q = gas flow rate, ms/s
The vertical  distance,  h, through which a particle of specified aerodynamic
diameter,  d.,  will   settle   in  this  time  period  is  simply  tD  times  the
            1                                                   K
terminal settling velocity, V,, of that particle.
                                  h = Vt X tR                    (Eq. 4.2-2)

where     V. = terminal settling velocity of particle size i, m/s
The terminal settling velocity is calculated from Equation 4.2-3.
                                   S P  d.
                                         L-                      (Eq. 4.2-3)

where   g = acceleration of gravity, 9.806 m/s2
       p  = density of particle, kg/in3
       d. = aerodynamic diameter of particle,
        u = gas viscosity, kg/m-s
     Equation  4.2-3  is  reasonably  accurate  for particles  with  Reynold
Numbers <1, which  in  this case is  equivalent to aerodynamic particle diam-
eters, £80  pmA.   Note  that settling  velocity is proportional  to the aero-
dynamic  particle diameter' squared and  inversely  proportional   to  the gas
     The ratio  of  h/H represents the fraction  of  particles  of a given size
that will be collected.
                                 V. x tR     VtBL
                         n/H =    I    1  = -S—                 (Eq. 4.2-4)
                               (q x t^)/BL    q

     Penetration, P., is simply 1 minus the fractional efficiency.
                         P1 = 1 - h/H = 1 - -=g-                  (Eq. 4.2-5)
     The total  penetration is  the sum of the  penetrations  for the various
particle  size  increments.   This  simple  approach  is reasonably valid for
particles in  the diameter range of 1 to 80 pmA, which is the range normally
of interest.  To extend this type  analysis  to  a wider particle  size range,
see Reference 4.  .
     Equation 4.2-4  is based  on the assumption of laminar  flow throughout
the  chamber.   Turbulent conditions will  lead  to  higher penetration (lower
efficiency)  than  predicted,   especially  for the  smaller  particle sizes.
     As the gas velocity increases, the limiting velocity at which deposited
particles may be reentrained into the gas stream may be exceeded.  A theore-
tical equation for pickup velocity, V , is:

                              [4 gd, (p.- p)l°'5
                         Vp = [	k^	                      CEq" 4"2'6)

where  V  = pickup velocity, m/s
        g = acceleration of gravity, m/s2
       d. = particles diameter, i, m
       p  = particle density, kg/m3
        p = gas density, kg/m3
     Figure 4.2-9  shows  a  typical penetration curve for  a  settling chamber
and Figure 4.2-10  shows  a  penetration curve based on dust measurements at a
sinter plant.



                     0    10   20   30  40  50   60  70   80  90  100
                                   PARTICLE SIZE, ym

     Figure 4.2-9.  Typical size efficiency curve for settling chamber.
       Adapted from data presented in Theodore and Buonicore, page 87.

                                IRON  OXIDE;
                       SIMPLE  SETTLING CHAMBER
                            HEIGHT   3.05 m
                            WIDTH    3.05 m
                            VOLUME   950 m3
                   0    10   20  30   40   50   60   70   80   90  100
                               PARTICLE DIAMETER,  urn
       Figure 4.2-10.
Penetration of dust through a settling chamber
   serving a sinter plant.
              Adapted from data presented in Jennings, R.  F.,
                   J. Iron Steel Inst.  (London) Vol.  164,
                              page 305, 1950.
     Momentum separators.   Particles are removed from a gas stream by use of
the inertia of particles.   A gas stream containing the particles is required
to make a sharp change in direction.
     The separation in  most devices of this type is a combination of momen-
tum mechanisms  and particle  settling  mechanisms.   The  particles penetrate
through the  gas stream and are settled by gravity  in hoppers.   Collection
may be by  impingement on  impaction surfaces or  by  use of secondary inertia
separation to remove the concentrated particles (Figure 4.2-11).
     A  penetration curve,  adapted from  Reference  7,  for a  commercially
available momentum separator is provided in Figure 4.2-12.


                 1-  CLEANED

                 -DUST SLOT
                                             COLLECTED FLY ASH
                                        FLY ASH OUTLET

                   Figure 4.2-11.   Momentum separator.

Reprinted from:  Jackson, R., British Coal  Utilization Research Association
                 Bulletin, Vol. 24, p.'221, 1962.
            0  10  20  30 40  50  60-70  80 90  100110120130140
                       FLY ASH  PARTICLE  DIAHETER, ymA

 Figure 4.2-12.  Penetration of  fly ash through two momentum separators.

           Adapted from data presented by Strauss, page 214.
             Industrial Gas Cleaning., Pergamon  Press, 1975.


     Mechanically  aidedseparators.   The  theory of  collection for  mechani-
cally aided  separators  is  similar to that used in momentum separators.   The
velocity of the gas  stream and the turn made by the gas stream  are  generated
mechanically  in  a fan or  other device.  The high radial tip velocity of the
moving  mechanical  surface and the  momentum generated  by  the  increased gas
velocity  move the  particles  to  the perimeter  of the  moving  surface.   The
particle  concentration  is  increased  in  the outer  portion  of the  fan  and
separated  from the main  gas  flow  by a  secondary  gas  flow.   The  normal
secondary flow is  between  10 and 20 percent of the total system volume.   Two
types of mechanically  aided separators are shown in Figure 4.2-13a.
                                 Aiternotive  » i
                                 gas outlet "JUT. [.__.
     Ouilj 901 // .
    to t«co«»arjf //' /
    col IK lor
Cfeoned gas
                      • Inlet
                                         Dust outlet
                                                                      TGos in!*'
                                                 COMPOUND SCROLL TYPE
            Figure  4.2-13a.   Types of mechanically aided  separators.
          Scroll-type collector drawing reprinted from:   Stairmond,
           C.  J.  and Kelsey, R. H., Chemistry and Industry,  pp.  1324,
           1955.   Compound Scroll-type drawing reprinted  by  permission
           of  Buell Ltd., George St. Parade, Birmingham B3 1QQ
     The  theory of collection  is not developed for collectors  of this type,
and the  penetration curves  are developed empirically.   Figure  4.2-13b shows
penetration curves  for the scroll- and compound-scroll type  collectors.   The
secondary  collector may  be a  conventional centrifugal  collector such  as  a

                           COMPOUND SCROLL
                    "0    20    40    60    80    100   120   140
                          AERODYNAMIC PARTICLE DIAMETER,
   Figure 4.2-13b.   Penetration curves for mechanically aided separators.
              Reprinted from:   Strauss, Industrial  Gas Cleaning,
                       Pergamon Press, page 265, 1975.

     Cyclonic separators.    The   performance  of   cyclonic   separators   is
dependent on  particle size.   Theoretical  relationships have  been  based  on
two different parameters,  the  critical particle diameter and the 50 percent
cut size  diameter.   The  former is the smallest particle size collected with
100% collection efficiency  (penetration  of zero).   Summaries of a number of
these  equations  are presented  by  Calvert  et a!.11  and  Strauss,7   Such
equations provide  only a  general  indication  of whether a  cyclone will  be
adequate for a specific application.
     Theoretical models based on the 50 percent cut size have been developed
by  Lapple,12  Leith  and  Licht,13  and Kalen and Zenz.14  Differences in the
equations  result  primarily  from  the  assumptions  used to explain  particle
behavior  with the  cyclone.    These  and other approaches  are  described  in
detail in Strauss,7  Crawford,1 Licht,15 and Theodore and Buonicore.4
     The  Leith  and  Licht  model  is  based  on  the  general  concept of radial
mixing within the  cyclone  cylinder due to a combination of turbulent mixing
and particle bounce.

     The  dependence  of  the  particle  size-specific penetration,  P.,  on
cyclone  design parameters  and gas stream  characteristics is  presented  in
Equation 4.2-7, which was developed by Licht.1S
P. =e
~v~ 2n + 2 "
KQpp (n+1)
L D 18M

• 1 ~
, n+1
                                                                 (Eq. 4.2-7}
Parameters used  in  Equation 4.2-7 are defined in Equations 4.2-8 to 4.2-13.
The factor,  n,  is  a temperature-dependent parameter originally presented by
                   n = 1 - (T/283)°'3[l - 0.67 (D)0'14]

                             8[(V + 0.5V.J/D3]
                                    «.      ri- .- ..........
                                    2 [B/D]2

                       Vs = J [S - 0.5a][D2 - D

              = f)2(n-S) + JD«[l + f

                           d = D - (-
                          $, = 2.3 De (D2/ab)
       S = height of outlet tube extension, m
       a - height of inlet duct, m
      D  = diameter of outlet tube, m
       D = diameter of cyclone cylinder, m
       h = height of cyclone cylinder, m
       8, ~ "natural length," m
       H = height of cyclone and cone, m
       B = diameter of cyclone cone outlet, m
       b = width of cyclone inlet, m
      p  = particle density, kg/m3
       Q = gas flow rate, ms/s
                                                                 (Eq.  4.2-8)

                                                                 (Fa  4 2-91
                                                                 ^ q*       J

                                                                 (Eq.  4.2-10)

                                                                 (Eq.  4.2-11)

                                                                 (Eq.  4.2-12)

                                                                 (Eq.  4.2-13)

       M = gas viscosity,  kg/m-s
      P.. =  penetration of a  particle  having an aerodynamic diameter of  i,

     The Leith  and Licht equations  give a fractional penetration curve  of
the  type shown  in  Figure  4.2-14.  The cyclone  dimensions and  gas flow
characteristics  used  in  the  calculation  of this  curve are  presented   in
Reference 15.
     Lapple's   cyclone  performance  model involves  the  calculation  of the
particle cut diameter,  d50,  using Equations  4.2-14  and 4.2-15



    S 0.5

    °- 0.4



                                        T     I
                               I     I
   Figure 4.2-14.   Penetration curve predicted by Leith and Licht approach.
         Adapted from:  D. Leith and W.  Licht,  American  Institute of
                        Chemical Engineers Symposia Series  68,  1972

dso = 0.308 [(9
 «  - (V/Q).
                                       )/2nNtv.(p  -pG)]
                                                         °'5      (Eq. 4.2-14)
                                      \j  v*     is I   p  \J
                      t      TTU^                                  ^ 4'2~15)
where   dso = particle diameter collected with 50% efficiency, umA
         Up = gas viscosity,  kg/m-s
         B  = width of gas  inlet,  m
         v. = inlet gas velocity,  m/s
         p  = particle density, kg/m3
         pg = gas density,  kg/m3
         Nt = effective number of  gas turns,  dimension!ess
          V = volume of cyclone, m3
          Q = gas flow rate,  m3/s
          D = diameter of cyclone  cylinder,  m
Calculation of  the  volume  of the  cyclone and  Equations  14 and 15 yield the
dgo-   This is  a  useful  parameter   for  comparing  various cyclones.   If a
fractional penetration  curve is desired,  the dso can be used in conjunction
with  a  generalized  curve   as  shown  in Figure 4.2-15   from Theodore  and
Buonicore.4  A  more detailed discussion of  this  approach is provided in the
latter  reference.   Information concerning theoretical  performance models is
presented  in   Licht,15  Theodore  and   Buonicore,4  Kalen  and  Zenz,14  and
Theodore and DePaola.17   The general  relationship between particle penetra-
tion and cyclone parameters is summarized in Table 4.2-2.
                      _"S^ I III!!
                                     I  I
                         I  I  I  I 1 I L.
                      0.3 0.4 O.bQ.V
                             PARTICLE SIZE RATIO, (d|)/cipc)
Figure 4.2-15.  Cyclone penetration  as  a  function of particle size ratio.
  Adapted from:  Theodore and Buonicore,  Industrial  Air  Pollution Control
                   Equipment for Particulates, CRC Press,  1976.

Gas flow rate
Particle density
Gas viscosity
Dust loading
r(S - "<£
_Cpi ^_
/c v 0.182


Licht,15 Theodore
r Licht,15 Theodore
Licht,15 Theodore

     The fractional efficiency of a cyclone system may be improved by reduc-
ing the cyclone diameter and using multiple cyclones to .handle the gas flow.
Penetration curves  for several  common  multiple-cyclone tubes  are  shown in
Figure 4.2-16.
                                        	 LARGE  DIAMETER  TUBES
                                        	MEDIUM DIAMETER TUBES
                                        	 SMALL  DIAMETER  TUBES
                  20     30    40     50    60
                      AERODYNAMIC PARTICLE DIAHETER,  pmA
Figure 4.2-16.   Penetration curves for multicyclone tubes of different
   Adapted from:   Theodore and Buonicore, Industrial Air Pollution Control
           Equipment for Particulates, CRC Press, page 129, 1976.

     The overall  penetration curve  for a  number  of cyclones  in  parallel,
such as  in  a multiple-cyclone tube bank, may  differ from the efficiency of
individual   tubes  because of  interferences  from the inlet and  outlet,  gas
distribution, dust stratification, reentrainment from the hopper, inleakage,
axial vane  wear,  plugging,  and variation in pressure drop across individual
     Axial  inlet - axial discharge (double vortex) cyclones.    The   funda-
mental  mechanisms involved in the double vortex cyclones have been discussed
in  detail  by Schmidt;19  however,  theoretical models are  not available.   A
general  penetration  curve prepared by Aerodyne Corporation  is  presented in
Figure 4.2-17.
                            I     I    II
I     I    I
                                4	1	J.
             "0  0.5  1.0  1.5 2.0 2.5  3.0  3.5 4.0  4.5  5.0 5.5  6.0
                        AERODYNAMIC PARTICLE DIAMETER,
        Figure 4.2-17.  Penetration curve for double vortex cyclone.
    Reprinted with permission of Aerodyne Corporation, Cleveland, Ohio.

The reduced  penetration claimed  in the <2-|jmA size  range  may be partially
due  to  reduced  particle  reentrainment  resulting  from  particle  bounce.
Additional information is available in Klein.20   Static  Pressure Drop.   Typical cyclones  have static pressure
losses  ranging from  0.25  to 1.5  kPa.   A  number  of factors  contribute  to
this,  including  the  kinetic  energy losses  in  the cyclone  vortex,  cyclone

cylinder wall  friction,  and entry  and exit  duct functions.   The  kinetic
energy losses are normally considered to be the dominant factor.
     One common means  to  calculate static pressure drop, AS. P.,  is based on
"inlet  velocity   heads."   Licht15   summarizes   the   basic  approach  in
Equations 4.2-16  and 4.2-17.  Equation 4.2-16  was originally developed by
Shepard and Lapple.21

                       AS. P. = 7.6 x 10"6 p_V?  Nu              (Eq.  4.2-16)
                                           u i   n
                          NH = 16 a b/D                         (Eq  4.2-17)

where AS. P. = static pressure loss, m of H20
         PG = gas density, kg/m3
         V. = gas inlet velocity, m/s
         N,, = number of velocity heads
          a = inlet duct width, m
          b = inlet duct height, m
         D  = diameter of exit tube, m
     This  simple approach  for  single  cyclones  indicates  that  the static
pressure is proportional to the square of the inlet velocity and is directly
proportional to  the inlet/outlet area ratio.   Because  of the dependence on
gas  inlet  velocity, it follows that static pressure drop  is proportional to
the  square of the  gas flow rate.  Other pressure  drop equations share the
gas- flow- rate  squared  dependence,  but differ  with respect  to  the cyclone
dimensional  factors  taken  into account.   For  more  information  on  these
approaches, see  Strauss,7 Theodore and Buonicore,4 and Byers.22
4.2.3   Design of Mechanical Collectors
     Proper  design  of mechanical  collectors  is  necessary  to  ensure  unit
operation  at  optimum efficiency and to minimize  malfunctions.  This section
addresses  both  the design needed to achieve  desired penetration levels and
the  factors necessary  for continuous compliance.   Settling Chambers.   Proper design  of settling  chambers in-
cludes  not only specifying the  volume adequate for  overall  collection effi-
ciency, but also ensuring  uniform  gas  distribution and minimum turbulence.


     A nonuniform gas  distribution  in the inlet can  result  in  locally high
velocities and nonuniform dust concentrations through the chamber.   The high
velocities lead to  increased  penetration of fine particulates.   Gas distri-
bution  may  be  improved  by  the  use  of gradual  turns  in the  ductwork,
straightening vanes, and perforated plates.
     The design must prevent air inleakage into the chamber and  dust hopper.
Inleakage increases turbulence, causes dust reentrainment, and prevents dust
discharge from the hopper.   Inleakage through the hopper can be  prevented by
use of a  rotary air lock,  flapper valve, or other means of sealing the dust
discharge.   Inleakage  through  the  chamber shell can be  prevented  by use of
proper  welding practices   and  proper  materials  of  construction.   When  a
chamber  is  used  to process  a  gas stream  at  high temperatures with  high
moisture content, the  inleakage of cold air causes  local  gas quenching and
condensation.   The  condensation can  cause  corrosion,  dust buildup,  and
plugging  of  the hopper.  The  use of thermal insulation  can reduce radiant
heat loss and prevent operation below the dew point.
     Materials  of  construction must provide extended life under conditions
of corrosion and abrasion.   The composition and gauge of materials should be
specified  on  the  basis  of the  expected  composition  of  the  gas  stream.
Access doors  and cleanout  doors should be included  to  allow inspection and
cleanout of duct work,  distribution vanes, and hoppers.
     Since the large particles are separated at the front of the collectors,
it is expected  that the maximum weight  and  volume  of material  will accumu-
late near  the collector inlet.   The design  must  incorporate special access
for  frequent dust  removal  from  this  area  to maintain  operation.   Normal
instrumentation  for settling   chambers  consists  only  of  an indicator  of
differential  static pressure.   An increase  in  static  pressure  drop  can
indicate plugging of turning vanes or distribution plates.  Momentum Separators.   The design of  momentum  separators must
provide sufficient  volume to  allow settling of materials separated from the
high-velocity  gas   stream   and  materials  of  construction  hard  enough  to
survive high abrasion.
     As with all collectors, the design must include methods of  sealing dust
discharge from hoppers to prevent inleakage.   The methods may include use of

rotary  air  locks,  flapper  valves,  or other  positive  sealing  devices.
Inleakage through  the  hopper  or shell  results in changes in the gas distri-
bution, interferes with dust discharge, and  may cause condensation or corro-
     Because of the  high  velocities used to separate the particles from the
gas stream  and  the impaction  of these  particles on surfaces that direct the
gas flow, the materials  of construction must have high abrasion resistance.
     Access must be provided for inspection  and cleanout of the gas passages
and dust  hoppers.   The access must be  large enough to allow the changing of
surfaces subject to high abrasion such  as liners or blast plates.
     Normal   instrumentation consists  only  of differential  static pressure
indicators.   The  plugging  of  gas passages or  abrasion  of  baffle plates may
be  indicated by  a  shift  in  normal   static  pressure drop.   Such effects,
however, should be detected by visual  inspection  of  the collector before a
significant change in static pressure is noted.  Mechanically Aided Separators.  Because of the high rotational
speed  of  mechanically  aided collectors,  the major design considerations are
abrasion and vibration.   Because of the abrasion associated with removal of
large  particles,  the design must include an impeller made of materials with
high abrasive resistance.   The inlet of the collector is designed to provide
even wear of the  impeller.  If  uneven wear occurs,  the resulting imbalance
could  cause bearing or impeller failure.
     The effects  of abrasion  and material buildup on the impeller are mini-
mized  by operating the units at low speeds (typically 400-800 rpm).  Buildup
of  sticky or wet  materials on  the  impeller may be reduced by a water spray
in  the inlet.   The coating of water reduces  adhesion of particles and acts
as  a wet fan.
     The  periodic  sheeting of  materials  from  the   blade  tips  results  in
moderate  vibration.   The  housing and structural  support  should  be either
spring-mounted  or rigidly  fixed  to a substantial  foundation, depending on
ability to  isolate  the system from ductwork.
     Since  the  mechanically aided collector acts  as  the prime air mover in
the system,  the air volume is  affected  by  all of the normal variables that
affect fans.   The  typical  fan curve may not  be  applicable because of con-
tinuous changes in  surface  contour resulting from wear and material buildup.


Because  of the  constant change  in  operating conditions,  the  penetration
curve is subject to change.
     An  increase  in impeller  weight caused by  buildup can  increase  drive
belt  slippage  and can  increase brake horsepower.  The  design  should  allow
for multiple drive belts or direct drive and high motor horsepower.  Cyclones.
     Simple cyclones.  The  design  of simple cyclones includes sizing of the
cyclone  to  provide an  inlet  velocity  that  will  ensure  high  separation
efficiency without  excessive turbulence.   Typical  values range  between 10
and 25 m/s.4  Turbulence at high inlet velocities results in abrasion of the
cyclone wall and  gas outlet duct.   These areas  can  be reinforced by use of
materials of extra  hardness or thickness.  Turbulence can be reduced by use
of inlet configurations such as helical or involuted designs.
     The  design  should  use flush  welds and  should avoid  components  that
increase roughness, such as rivets.  Internal disturbances reduce efficiency
by creating  areas of  turbulence  and  causing  particles to  bounce  from the
wall  to the inner vortex.  The design should also include methods of sealing
the dust discharge and preventing gas inleakage, such as use of a rotary air
lock, flapper valves, or other devices.
     Multiple cyclones.   The design  of  a  multiple-cyclone  system entails
specification of  the number of individual   tubes  needed to  handle the gas
volume  without  exceeding  the  maximum  gas  flow  per tube.   If the  inlet
velocity of  the  tube is excessive, the resulting turbulence increases  pene-
tration.   At  excessive  inlet  velocities  abrasion  also  increases.    In
general, the  maximum inlet velocity may be limited  by  the  abrasiveness of
the particles to be collected and by the abrasion resistance of inlet vanes.
     The  arrangement  of the  tube  bank generally  should be  in  a  square
matrix.  The  use of  internal baffles  in  the inlet can allow a  collector of
shallow depth to be used without abnormal gas distribution across the width.
Because of the possibility of static pressure differential across large tube
banks and  the  resulting hopper short-circuiting, baffles should be included
in the hopper or multiple hoppers should be provided.
     Access should be provided both to clean and dirty collector plenums for
dust  cleanout,   inspection, and  replacement  of tubes.   The access  doors


should be  large enough to  permit  safe entry and removal  of  tubes.   Gasket
materials used to seal tube assemblies, access doors, etc., should have long
life  at  elevated temperatures.  A major cause of  failure of well-designed
systems is gas penetration through weld gaps and gasket leaks from the dirty
side to the clean side.
     The  collection  hopper  should be  equipped with  devices  that  provide
positive sealing of  the  dust outlet,  e.g.,  rotary  valve,  flapper valve, or
other  devices  to prevent  air inleakage.   Air  leakage can  cause dust dis-
charge, hopper bridging,  and condensation.
     The  collector  should  be  equipped with  differential static  pressure
monitoring equipment to determine static pressure drop.
4.2,4  Operation andMaintenance of Mechanical Collectors
     Even well-designed equipment can fail because of improper operation and
inadequate  maintenance.    Despite  the  apparent  simplicity  of  mechanical
collectors, regular maintenance is necessary.   Settling Chambers.   The most  common  failure mode of settling
chambers  is plugging of  the chamber with collected dust.  In simple collec-
tors  the  plugging  can  result  from  hopper  bridging or rotary air lock
failure.   In   more  complex collectors,  such as Howard's  settling chamber,
plugging  of  the individual  gas passages  can  occur.  Such  failures  can be
prevented  or  minimized by  use of hopper  level indicators or by continuous
monitoring of  the  dust discharge.   Scheduled internal inspection can deter-
mine  areas of  inleakage  and condensation,  both  of which may cause hopper
bridging.    Momentum Separators.    The  most  common   failure  modes  of
momentum  separators  are  hopper plugging and baffle plate  erosion.  Plugging
of  hoppers can  be  reduced by  use of  hopper  level  indicators.   Erosion of
baffle plates  and  collector shell  can be reduced by the use of extra thick-
ness  in  areas subject  to  abrasion.    Periodic internal  inspection  of the
collector  is  recommended  to identify and correct areas of high abrasion and
air inleakage.   Mechanically  Aided Separators.   The  most common failure modes
of mechanically aided  separators are abrasion and structural disintegration.

The  impact  of  an abrasive  dust  on  impeller surfaces  at high tip  speeds
causes  rapid  erosion.   The  impeller  can  become  unbalanced;  and  if  the
unbalance is not corrected, structural  failure can occur.   The attachment of
sticky or tacky particulate to the impeller can also cause vibration.
     The  use of  vibration detectors and  periodic  external  inspection  can
indicate  impending  failure from  erosion  or material  buildup.   Routine  in-
ternal  inspection of the  impeller with removal  of collected material  can
extend collector life.   Simple  Cyclones.   Simple  cyclones fail  most  often from  abra-
sion and  plugging of the dust outlet tube.   Abrasion  occurs  in areas  oppo-
site the gas inlet and in the lower areas of the cone.   Wear can increase if
inlet velocities  are  high  (>25 m/s) and particles  can erode  the gas  outlet
tube.   Internal roughness  caused by poor welding or faulty fabrication  can
cause local  turbulence, which increases  erosion and particle bounce into  the
inner vortex.   Plugging  of the dust outlet  tube  can be reduced by use of a
large  diameter outlet tube  and by proper sizing  of  the rotary  air  lock.
     Routine internal inspection of the  inlet duct and  internal surfaces  for
areas of abnormal  wear should be conducted to prevent shell failure.   Liners
and  replaceable  wear   plates  are  recommended  where  abrasive   dust   is
collected.   Multiple Cyclones.   Factors which  may  contribute to  reduced
performance  include  erosion,  plugging,  corrosion, and  hopper recirculation.
     Hopper  recirculation  in  reverse  flow type systems occurs when tubes at
the  rear  of the bank have a  slightly lower  static  pressure drop.   This  can
occur  whenever flow  distribution  is  nonideal  or when  the  outlet  tubes  are
shorter than those  in the front23'24  (a common design).   In  these cases a
portion of the gas in the front tubes can pass out the  bottom and then enter
the discharge of tubes in the back rows.
     Hopper recirculation can lead to substantially reduced collection effi-
ciency.24  Some of the ways to minimize  this problem include segregating  the
hopper and equalizing the lengths of all discharge tubes.23
     Ambient air  inleakage into the hopper area can interfere with the dust
discharge in the tubes  and lead  to  increased emissions.  These  leaks  may
occur  at hopper  flanges,   access  hatches or  through   the  solids  discharge

valves.   Severe  leakage  on multicyclones serving combustion systems  can  be
identified by  measurement of the  flue  gas oxygen content before  and after
the collector.
     Plugging of  the gas  inlet vanes  leads  to partial or total  failure  of
individual  tubes  to  establish an adequate vortex.   Partial  plugging  allows
dust penetration  through  the  affected  tube (Figure 4.2-17).   Complete plug-
ging results in  an  increased  flow of gas through the remaining tubes  and  an
increase in  overall  collector static pressure drop.   Partial  plugging  of a
significant  number of  tubes can  cause variation  in  pressure  drop from tube
to  tube,  cause gas  short  circuiting through the dust hopper,  and cause  an
increase in dust reentrainment at the tube dust outlet.
     Plugging  of outlet  tubes  or  the  solids  discharge  is common.23   The
latter  is  generally  due  to  poor hopper  discharge practices, which allow
solids to accumulate to the bottom of the tubes.23  According to Barrow,25 a
single plugged tube  can  reduce  collection efficiency  by  as  much  as 25 per-
     Excessive gas velocities at  the entrance of  the  multicyclone can lead
to  erosion  of the  outlet tubes.   Once  a small hole  is  created,  the large
differential static pressure between the inlet and outlet leads to very high
gas  velocities through the hole.   Rapid erosion then leads to substantially
reduced collection efficiency.
     Corrosion in multicyclone collectors is  usually  minimized by avoiding
operation  at or  near  the acid vapor  dewpoint.23   In  certain cases, insu-
lating the units  and ductworks leading to it may be advisable.
     Regular inspection  of the  multicyclone should be performed to minimize
the  above  problems.   Static pressure taps should  be installed on the inlet
and  outlet  ducts so  that the static  pressure drop  can be  measured  on a
regular  basis  by either  a portable gauge or  a permanently  mounted  instru-
ment.  Because of the tendency of these  taps to plug, they should be  cleared
before any measurements are attempted.
     For  units  serving  combustion  sources,  gas  temperature  and  flue gas
oxygen content should be measured on a  routine basis.  This value should be
compared with typical values  for the specific unit to  identify deterioration
of  the unit  and/or operation  below the dewpoint.

                          Figure  4.2-18.   Partial  pluggage of multiple cyclone inlet vanes.
                                           (Courtesy of PEDCo Environmental, Inc.)

     The rate of solids discharge should be checked frequently.   Bridging of
solids  in  the  hopper,  failure  of the  solids  discharge valve, or  severe
internal problems  may all be  identified by this simple  technique.   All  of
these situations demand immediate attention.
     At least  once a year,  internal  inspections should  be  conducted after
the  unit  has been  purged by personnel  qualified in proper confined-area-
entry  procedures.   The   internal   inspection  should  identify  plugging,
erosion, and corrosion  problems.   In  some cases  it  may  be advisable to use
tracers or smoke to identify gasket and weld leaks.


 1.  Crawford,  M.   Air  Pollution  Control  Theory,  McGraw-Hill,  Inc.,  New
     York.  1976, p. 236.

 2.  Sargent, 6.  Gas/Solid Separation.  Chemical Engineering.   78(4):11-12.
     February 15, 1971.

 3.  Munson,  0.  S.   Dry  Mechanical  Collectors.   Chemical   Engineering.
     75(22):147-151.  October 14, 1968.

 4.  Theodore,  L.,  and  A.  J.  Buonicore.   Industrial  Air Pollution Control
     Equipment, CRC Press, Cleveland, Ohio, 1976, p. 65.

 5.  Alden, J.  L.   Design of Industrial Exhaust System, 3rd ed.  The  Indus-
     trial Press, New York.  1959.

 6.  Jennings, R. F.  J.  Iron Steel Inst.  164:305.  1950.

 7.  Strauss,  W.   Industrial  Gas Cleaning, 2nd  ed,   Pergamon Press.  1975,
     p. 213.

 8.  Stairmund, C. J.   Trans Instr. Chem Engrs.  29:356.  1951.

 9.  Caplan,  K.   Mechanical  Collectors.  In:  Air  Pollution,  Vol.  Ill, 2nd
     ed.  Stern, A., Editor.  Academic  Press.  New York.  1968.

10.  Aerodyne  Development  Corporation,  Series "SV" Dust Collector, Bulletin
     No. 1275-SV.

11.  Calvert,  S.,  J.  Goldschmidt,  D.  Leith, and D.  Mehta.   Wet Scrubber
     Handbook, Vol. I.   U.S. EPA.  Research Triangle Park, N.C.   Publication
     No. EPA-Rl-72-118a.   1972.

12.  Air Pollution Engineering Manual.  U.S.  Environmental Protection  Agency
     Publication, AP-40.   Research Triangle Park, N.C.  1973.  p. 95.

13.  Leith,   D.,   and W.   Licht.   A.I.Ch.E.  Symposium  Series.   68:196.
     November 26, 1979.

14.  Kalen, B.,  and F.  A. Zenz.  A Theoretical-Empirical Approach to  Salta-
     tion  Velocity in  Cyclone  Design.  The Ducon  Company.   Bulletin No.
     C-207.  1973.

15.  Licht,  W.   Air  Pollution  Control  Engineering.   Marcel  Dekker,  Inc.

16.  Alexander,  R.    Mck.   Proc.  Austral.  Inst.  of  Min.  and  Met  (N.S.).
     152:202.  November 3, 1949.

17.  Theodore,  L.,  and V. DePaola.   Predicting Cyclone Efficiency.   JAPCA,
     30:1132-33.  October 1980.

18.  Baxter,  W.   In:   Air  Pollution,  Vol.  II, 3rd  ed.   Stern, A.  Editor.
     Academic Press.  New York.  1977.

19.  Schmidt, K. R.  Staub.  23:491-501.  1963.

20.  Klein, H.  Staub.  23:501-08.   1963.

21.  Shepard,  G.  B. ,  and C.  E.  Lapple.   Ind. Eng.  Chem.   31:972.   1939.

22.  Byers,  R.  L.   Gravitational  and Dry  Centrifugal  Collectors and  Air
     Pollution  Control.    In:   Integrated  Engineering  Solution to  Overall
     Pollution  Control:   Air,  Water, and Solid Waste  Problems.   AICHE.   New
     York.  1971.

23.  Dey,  A.,  J.  Maloney, and J. D'Imperio.   Inertia!  Separators.   In:   Air
     Pollution  Engineering  Manual.   U.S.  Environmental  Protection  Agency,
     Research Triangle Park, N.C.   Publication No.  AP-40.   1973.   pp.  91-99.

24.  National  Asphalt  Paving Association.  The Maintenance and  Operation of
     Exhaust  Systems  in  the Hot Mix Batch Plant.   Information Series  52,
     NAPA.  Riverdale, MD.   1975.

25.  Barrow,  A. J.,  Jr.   Particulate and  S02 Control  Technology for  the
     Small  and Medium Coal-Fired Boiler.   Presented at the  1970 Industrial
     Coal  Conference.  Purdue  University.  October 7-8, 1970.


     Electrostatic  precipitators  (ESP's)  are  high efficiency  particulate
collection  devices applicable  to a  variety of  source categories  and gas
conditions.  Particle collection is done by application of electrical energy
for  particle charging  and collection.  Efficiencies  of 99.9+  percent are
possible,  depending upon  application,  ESP  design, and gas   and particle
4.3.1  Types of ElectrostaticPrecipitators
     Electrostatic precipitators  used  for  controlling particulate emissions
may be placed in two general categories:  dry and wet.   The major difference
is  in  the  method  by which  the particulate is removed from the collection
electrodes.  Each  of these categories of precipitators  may  be further sub-
divided  on the basis  of electrode geometry and  application.1  The dry ESP
with  pi ate-type collection  electrodes  and  pyramidal  hoppers  is  the pre-
dominant type in industrial applications.  Regardless of the type of precip-
itator and its geometry,  the  particle  capture is  accomplished by electro-
static attraction.  Dry Precipitators.  Examples of the industrial sources that heavily
utilize  dry ESP's  for control  of particulate emissions are utility boilers,
cement kilns, kraft pulp recovery boilers, and metallurgical furnaces.  Each
industrial  application requires different designs for the gas  conditions and
particulate characteristics.
     The basic  functions of an  ESP are (1) to impart a charge  to the partic-
ulate,  (2)  to  collect  the charged  particulate  on  a  surface  of opposite
polarity,  (3)  to  remove the collected particulate  from the collecting sur-
face  in  a  manner  that minimizes reentrainment of  the  particulate into the
gas  stream, and  (4)  to discharge the collected  particulate.   Each dry ESP
with  horizontal gas  flow  must have  a shell to  enclose the collection and
discharge  electrode system,  collection  electrodes, discharge electrodes, a
high-voltage  transformer-rectifier for  application of  electrical  power to
the  ESP,  a system of rappers  to  remove  particulate from the  collection and
discharge  electrodes,  and  a system to remove the collected particulate from
the   precipitator proper.    Figure 4.3-1  shows  the   basic  features  of  a

              •US DUCT ASSY
                               VENTILATION SYSTEM


 24 1n.
                                              PRIMARY LOAD
                                                      ROL PANEL

                                                                      HIGH VOLTAGE
                                                                      WITH WEIGHT
      Figure  4,3-1.
Typical  ESP with  insulator compartments (courtesy
     of Western Precipitation),2

weighted-wire discharge electrode  ESP.   Another type of design uses a rigid
frame discharge electrodes mislead of weighted cones.  Precipitator housing or shell.  Gases  entering the precipi-
tator are  usually  ducted  so that gas velocities  are high enough to prevent
significant particulate fallout in the ductwork.  These velocities (typical-
ly  20  to  35  m/s)  are much too  high  for the ESP to capture the particles.
Thus, the  precipitator shell  not  only encloses the gas  treatment area but
also  constitutes  a  large expansion  volume in the  ductwork  to  reduce gas
velocity  to between  1 to  2  m/s.   This  reduced  velocity  is  necessary to
allow sufficient residence  time  in the ESP  for the  particles to migrate to
the  collection  electrodes  and  to  avoid  reentrainment  to  the  gas stream.
     The precipitator  shell  is typically made of carbon steel that is insu-
lated to  reduce  corrosion of the shell.  Doors are  usually provided between
each  field to allow  internal  inspection  and maintenance  when the precipi-
tator is off-line.   Depending upon design and manufacturer, the precipitator
may be equipped with a penthouse to house high-voltage insulators.   Discharge electrodes.   Dry  plate ESP's  are  of  two types:
weighted-wire  and  rigid-frame discharge  electrode  systems.   The discharge
electrode  provides the  charge  to  the particulate  in the  treatment zone.
Weighted-wire designs  (Figure 4.3-1) have dominated  ESP service in the past.
Wires are  suspended  from  a high-voltage frame  at  the top of the electrical
section, and  shrouds  are  provided to reduce  sparking  between the wires and
the end of the collection electrode.  The weighted-wire design typically has
a  lower  initial  cost than  rigid-frame  designs,  and  closer  spacings are
allowed between collection electrodes and discharge  electrodes.  Rigid-frame
designs, however, tend to provide lower maintenance  requirements.
     Both  discharge electrode configurations have excellent collection capa-
bility.  The  low  initial  cost of a weighted-wire design is typically offset
by  high maintenance costs caused partially by wire breakage.  The reverse is
true  for rigid-frame  designs  for  which the high  initial  costs are usually
offset by  low maintenance costs.  Warping  of the  discharge electrode frame
because  of wide thermal  swings  is generally not a  problem with a properly
designed unit.  Electrically, both designs are capable of delivering similar
power levels  to  the ESP for particulate  collection.  Generally,  the rigid-
frame design  operates at  higher voltages  and  lower current densities than

the weighted-wire designs  in  a given application because  of  the wider dis-
charge  electrode to  collection electrode  spacing.    These  voltage-current
characteristics  may  be better  suited  to  collection of  high-resistivity
dusts.   Collection electrodes.   The  collection electrodes  usually
consist of  a  plate  with stiffeners to provide  support.   Some manufacturers
incorporate baffles (doubling as plate stiffeners) to provide regions of low
gas turbulence that enhance particulate capture.  The charged particles that
migrate across  the  gas  stream under the force of  electrostatic attraction
build up a layer of dust on the collection electrodes (plates).   High  voltage  transformer-rectifier (T-R).  Operation of the
ESP  depends  upon  electrical   power being  supplied  by  high-voltage trans-
former-rectifiers (T-R).   The function  of the  T-R set is  to convert a low-
voltage AC  power to  high-voltage  DC.  In  most industrial  applications the
voltage applied is DC negative (negative corona) because higher power can be
applied to  the  ESP  at lower  sparking values  than is possible with positive
corona  units.   Typically,  a  silicon-controlled-rectifier  (SCR)  circuit
controls  precipitator current  "phase,"   and  the high-voltage  power supply
includes an automatic control unit to control  delivery of optimum voltage-
current performance.   Most  modern  ESP's  also  include linear  reactors  to
modify the current waveforms and provide stability during sparking.
     Most ESP's  have  a number of T-R sets;  the number is determined by the
manufacturer's  preference  in conjunction with  the  desired collection effi-
ciency, degree  of sectionalization,  and  degree  of  redundancy needed.   The
voltage and current ratings of the T-R's must be matched to the application,
electrode geometry, and gas and particulate characteristics.   Typically, in-
put voltage is  460  V, three-phase, 60 Hz  AC with an output voltage between
45 and 70 kV DC.  The maximum rated current output of the T-R sets is usual-
ly in the range of 250 to 1500 mi Hi amperes.   Insulators.   Insulators  are  needed  to prevent grounding of
the  high-voltage power supply system with  the  precipitator  shell.  Insula-
tors  are  often  made of a  ceramic  material  selected  for its high dielectric
strength and  resistance  to most components in  the  gases  to which it may be
exposed.  Insulators  are used where the  high-voltstge  supply penetrates the


precipitator shell and  is  connected with the discharge electrode system and
wherever the  high-voltage  system  comes  close to the  precipitator shell  or
plates.   Rappers.   For dry  ESP's,  some version  of rapping  must  be
used to  remove particulate from collection  surfaces.   Rapper types include
electric  vibrators   and  gravity  impact  hammers.   Rappers are  provided for
both the collection plates and discharge electrodes.  Excessive dust buildup
in the precipitator  will  degrade the performance of the precipitator and is
usually evidenced by reduced power to the precipitator.
     The number of rappers necessary for effective cleaning depends upon the
type  of  rapper,  the ESP  configuration,  and other  design  considerations.
Each installation requires  fine-tuning of both rapper intensity and rapping
frequencies  to   minimize  reentrainment  of   dust.   The  rapping intensity
depends  upon  the particulate  characteristics and  the amount of collection
area and discharge  electrode  to be cleaned  per  rapper.   Rapping frequency
generally decreases  as  the gas travels from inlet to the outlet of the ESP.
It  should  be  noted  that excessive  rapping  can degrade  ESP  performance as
much as insufficient or ineffective rapping.   Solids  discharge.   Pyramidal  hoppers are generally used for
collecting partlculate.   Discharge from hoppers may be accomplished by means
of  screw conveyors,  drag conveyors, or pneumatic conveying systems.  In the
pulp and paper industry flat  bottom,  tile-lined  precipitators that utilize
drag conveyors are  common  on  recovery boilers.  Solids discharge can repre-
sent  a significant  problem  in  the  operation of an ESP  in  that excessive
buildup  of  material  can cause an electrical shortage'or misalignment of ESP
internal components.  Gas  distribution equipment.   Effective utilization of elec-
trical energy  supplied  to the  ESP depends upon well-balanced air flow across
the  ESP.   In  new  installations  the requirements for extensive gas turning
vanes,  because  of  sharp  bends  in  ductwork,  can  be  reduced.  Retrofitted
ESP's, however,  can  present problems because of space limitations requiring
sharp  bends  in the ductwork immediately before and after the  precipitator.
Gas turning vanes and some type of diffusion plate  (e.g., perforated plates)
are typically  utilized  to balance the gas flow.


     4.3.1,2   Wet Precipitators.   Wet  preeipitators  are used  primarily in
the  metallurgical  industry, usually  operating below  75°C.   Until  the late
1960's  their  use  was restricted  mostly to  acid  mist, coke  oven  off-gas,
blast  furnace, and  detarring  applications.   Their  use  in other  areas is
rapidly  increasing with the  need for  increased  control  efficiencies.   The
new  applications  include  sources  with  sticky   and  corrosive  emissions.
Because  of inherent temperature  range limitations,  wet ESP's  are  not used
for boiler installations.
     The fundamental difference between a wet and  a  dry ESP is that a thin
film of liquid flows over the collection plates of a wet ESP to wash off the
collected  particulates.   In some  cases,  the liquid  is  also  sprayed  in the
gas  flow passages  to provide cooling,  conditioning,  or a  scrubbing action.
When the liquid  spray  is used, it  is  precipitated with the particles, pro-
viding  a  secondary means  of wetting  the  plates.   Three  different wet ESP
configurations are shown in Figure 4.3-2.   Plate-type (horizontal flow).   The  effluent gas stream is
usually preconditioned to reduce temperature and achieve saturation.  As the
gas enters the inlet nozzle, its velocity decreases because of the diverging
cross section.  At this  point, additional sprays may be used to create good
mixing  of  water,  dust,  and gas  as well  as  to ensure  complete  saturation
before  the gas enters  the electrostatic field.  Baffles are  often used to
achieve good velocity distribution across the inlet of the ESP.
     Within  the  charging  section, water  is  sprayed  near  the top  of the
plates  in  the  form of  finely  divided drops,  which  become  electrically
charged  and  are attracted  to  the  plates,  coating them evenly.   Simultan-
eously, solid  particles  are charged;  they "migrate"  and become attached to
the  plates.  Since the  water film is moving downward by gravity on both the
collecting and discharge electrodes, the particles are captured in the water
film, which  is disposed of from the bottom  of the precipitator in the form
of slurry.   Concentric-plate.   The concentric-plate ESP  consists of an
integral tangential  prescrubbing  inlet chamber followed by  a vertical  wet-
ted-wall concentric-ring ESP  chamber.3'4  Concentric  cylindrical collection
electrodes are wetted by fluids dispensed at the top surface of the collec-
tion electrode system.   The discharge electrode system  is  made of  expanded

                                 KATE* SUW.T
A.   Plate type (horizontal flow)

Reprinted with permission of Academic
Press.   Stern, Arthur C., editor,  AIR
POLLUTION, 3rd edition, Vol.  IV.   1977.
                 GAS aow IN
                       SAS ROW our
                              KATER PIPES
                                                     GAS FLOW


C.   Conventional pipe type

Courtesy of  the McILvaine Co.,
                                            B.   Concentric plate type

                                            Reprinted with permission of Academic
                                            Press.   Stern, Arthur C., editor.  AIR
                                            POLLUTION,  3rd edition, Vol.  IV.  1977.
                   Figure 4.3-2.  Three types  of wet  ESP's.

                                        4.3- 7

metal with uniformly distributed corona points on the mesh background.  This
system  is intended  to  combine the  high,  nearly  uniform,  electric  field
associated with  a parallel  plate system and the nearly uniform distribution
of  corona current  density  associated with  closely spaced corona points.
Higher gas flows can be handled by  adding  concentric electrode systems and
by increasing the length of each electrode,   Conventional  pipe-type.   This  system consists  of vertical
collecting pipes,  each containing a  discharge  electrode  (wire-type), which
is  attached  to  the  upper framework  and held taut by a cast-iron  weight at
the bottom.  A lower steadying frame keeps the weights and thus the wires in
     The upper frame is suspended from the high-voltage insulators housed in
the  insulator  compartments, which are located on  top  of the precipitator
shell  (casing).   Heating and ventilating systems help to  prevent accumula-
tion of moisture and dust in the insulator compartments.
     The washing system usually consists of internal nozzles located at the
top of the plates.5   At specified intervals, the tubes are washed thorough-
ly.  During  the  washing,  the louver damper to  the  exhaust fan is closed to
prevent carryover of droplets.   Two-stage Precipitators.   The  two-stage  ESP  was  originally
designed  to  purify  air and  is used  in conjunction with  air-conditioning
systems.   Cleaning of  incoming  air at hospitals and  at  industrial and com-
mercial  installations  is a  typical  application.   As an  industrial partic-
ulate collector, the device  is used  for  control  of liquid particles dis-
charged  from such sources  as meat  smoke-houses, asphalt  paper saturators,
pipe coating machines, and high-speed grinding machines.
     Two-stage ESP's are limited almost entirely to  the collection of liquid
particles that will drain readily from collection plates.   Two-stage precip-
itators cannot control  solid or sticky materials,  and become ineffective if
particle concentrations exceed 1.0 g/m3.
4.3.2  Operating Principles  of Electrostatic Precipitators  Basic Processes.  The  three  basic processes  involved in elec-
trostatic precipitation are  (1) the  transfer of an  electric charge to sus-
pended particles in  the gas  stream,  (2)  the establishment of an electric


field for removing the particles to a suitable collecting electrode, and (3)
the removal  of the  particle  layers from  the precipitator  (Figure 4.3-3}.   Corona  generation.   As the  high-voltage DC  current  passes
through the  discharge wire,  it produces an electrical  corona,  which  can be
defined as an  ionization  of gas molecules by electron collisions in regions
of high field  strength  near the discharge wire.6  The strength of the elec-
tric  field  varies  inversely with  the distance from  the  discharge  wire.
     Three sources  of electrons  are used to  initiate the  so  called "ava-
lanche" of collisions (1) naturally occurring ionizing radiation, (2) photo-
ionization because  of the corona glow, and (3) in high-temperature applica-
tions, thermal  ionization at  the electrode  surface.6   These electrons and
positive  ions  as well, move  under  the  influence of an  electric field and
carry charges,  but  the  current generated from the flow of these carriers is
too  low  to  be  of significance.  Under  the influence  of sufficiently high
voltage, these free electrons are accelerated to a velocity high enough that
collision  with  a gas molecule  will  break an electron  loose from the outer
shell of  the molecule,  creating a positive  ion  and  another free electron.7
This phenomenon  is  repeated many times, thus the name "avalanche."  The end
result  is a large  accumulation of positive ions and  negative  electrons in
the region of the corona.
     The  corona  can be  either positive or negative;  but the negative corona
is  used  in  most  industrial  precipitators since  it  has inherently superior
electrical  characteristics that  enhance  collection  efficiency  under most
operating conditions.  Electric  field.  The electric field results from application
of  high  voltage  to  the ESP discharge electrodes, and  the strength of this
electric  field  is a critical factor in determining  ESP performance.   Space
charge effects from charged particles and gas ions may interfere with gener-
ation  of the  corona and  reduce  the strength  of the  electric  field.   The
space charge effect  is often seen in the inlet fields of art ESP where parti-
culate  concentration  is   the  highest.   From a  practical  standpoint,  the
strength  or  magnitude of the electric field  is  an  indication of the effec-
tiveness  of  an  ESP.  The magnitude of the,electric  field can be  mathemat-
ically  determined  by  use of  derivations  of  Poisson's  equations  including
charge carriers and their mobilities.7

                          CORONA GENERATION
                  Figure 4.3-3.   Basic processes involved in electrostatic precipitation.7
              {Oglesby,  Sabert Jr.,  "Electrostatic Precipitation" SRI Bulletin/Winter  1971.)
        Courtesy of Southern  Research  Institute.

-------   Charging mechanisms.   Particle   charging   and  subsequent
collection take place  in  the region between the boundary of the corona glow
and the collection  electrode,  where gas particles are subject to the gener-
ation of  negative  ions from the corona process.  Charging is generally done
by  field  and diffusion  mechanisms.   The  dominant  mechanism  varies  with
particle size.
     In field charging,  ions from the corona are  driven onto the partj^les
by the electric  field.   As the ions continue to impinge on a dust particle,
the charge on it increases until the local  field developed by the charge on
the particle causes such distortion of the electric field lines that they no
longer intercept the particle, and no further charging takes place.  This is
the dominant mechanism for particles larger than about 0.5 umA.
     The  time required for a particle to reach its saturation charge varies
proportionally to  the  ion density in the region where charging takes place.
Under normal  conditions with sustained high-current  levels,  charging times
are only  a few  milliseconds.  Limitation of current  because  of high resis-
tivity or other  factors can lengthen charging times significantly and cause
the  particles  to   travel  several  meters through  the  precipitator before
saturation charge is approached.
     The  waveform  of  the  secondary voltage can  further  affect the charging
times.   The   rectified  unfiltered  voltage  has  peaks occurring  at regular
intervals,  which match  the  frequency  of the  primary  voltage.   Thus,  the
electric  field  varies with  time, and the dust  particles  in  the  interelec-
trode region  are  subject to time-varying fields.   The particle charging is
interrupted  for  that  portion  of the cycle  during which the charge on the
particle  exceeds that  corresponding to the saturation charge for the elec-
tric  field existing at the time.  This further lengthens the charging times
and,  in  the  case  of high-resistivity  dust,  degrades precipitator perform-
     Diffusion  charging  is  associated with  ion  attachment  resulting from
random thermal motion; this  is the dominant charging mechanism for particles
below  about  0.2  umA.   AS  with  field  charging,  diffusion charging  is in-
fluenced  by  the  magnitude  of the  electric  field,  since ion  movement is
governed  by electrical as well as diffusional forces.  Neglecting electrical
forces, an explanation of diffusion charging is  that the thermal motion of


molecules causes  them to  diffuse through a gas and  contact  the particles.
The  charging rate  decreases  as  a  particle acquires  a charge  and  repels
additional  gas  ions,  but  charging  continues to  a certain  extent  because
there is  no  theoretical  saturation or limiting charge  other  than the limit
imposed by the  field emission of electrons.  This  is  because the distribu-
tion of thermal  energy ions will always  overcome  the  repulsion of the dust
     The  particle size'-range of about  0.2 to 0.5  pmA is  a  transitional
region  in which  both mechanisms of  charging  are  present but  neither  is
dominant.    Fractional efficiency test  data for  precipitators  have  shown
reduced collection efficiency in this transitional size range,  where diffu-
sion and field charging overlap.
     A more  comprehensive  theory8 has been derived that analyzes the diffu-
sion and  field  charging  mechanisms simultaneously.  The ion density distri-
bution near  a particle is determined  in  terms of  the local electric field,
and  the  rate at which ions  reach the particle due to  their  thermal  veloc-
ities  is  calculated  statistically.    A  computer  is  used to calculate  the
theoretical  charging  rate,  since it cannot be expressed in closed algebraic
form.  Results  of this work  indicate that the total charge accumulated by a
particle  is  strongly dependent on the electric field strength,  the diameter
of the particle,  the numerical density of  the ions,  and the residence time
of the particle in the charging region.  Other variables cited that can have
a  significant  effect on  particle charging  rates  are the  gas  temperature,
electrical mobility  of the ions  in the gas,  and  the dielectric constant of
the particulate material.
     Sufficient rapping  force must be applied to  produce  a rapid accelera-
tion perpendicular  to the gas flow  so that the dust shears  off the  plate.
Sproull9 indicates that rapping is optimum if the dust layer slides down the
plate vertically  after each rap, making  its way down  the plate in the dis-
crete steps until it finally  reaches the hopper.
     With a  tenacious  dust that  adheres stubbornly to the plate, vibrations
can be induced perpendicularly to the gas flow direction, in addition to the
necessary shear action,  resulting  in a  scattering  of the agglomerate  and
subsequent  reentrainment of  relatively  large  fractions  of  the  dust.   In
general, the dust should be  allowed to fall  freely off the plate, as some-
times occurs with high-resistivity  dust  when rapping  is done  with  "power

off."  The  other  extreme is with low-resistivity  dust,  whose reentrainment
can be caused by only a light rap.
     Recent  studies  have  investigated reentrainment  caused by  rapping  in
terms of the  percentage  of material reentrained and  its particle size dis-
tribution.10,11   One report  describes the  testing  of  six  full-scale  ESP
installations.  Losses from rapping ranged from over 80 percent of the total
mass emissions  from  one  hot-side unit to 30 percent of emissions from cold-
side units.   The  losses  consist mostly of relatively  large  particles, pri-
marily those  larger  than 2.0 umA in diameter.   Tests  of a pilot-scale pre-
cipitator showed  that  rapping emissions decreased as  time between  raps  was
     The  intensity  and  frequency  of  rapping are usually greatest  at  the
inlet sections,  decreasing as  the  gas moves  through the ESP.   The outlet
section  is  usually  rapped  only lightly,  since  the reentrained  dust is  not
recollected.  The visible puffs that often appear as a result of rapping can
be  used  with a  transmissometer to optimize the frequency and intensity of
rapping for each section of the ESP.  Importantprocess parameters influencing ESP performance.
The process parameters that most influence the design and operation of ESP's
are the  particle  properties such as resistivity and particle size distribu-
tion, and  gas properties such as process  temperature  and flow.   Once these
factors are determined, the designer can estimate the size of the ESP needed
to meet applicable emission regulations.
     Corona  current flows  through  the collected  dust  layer to  reach  the
collection  electrode.    With dry ESP's, high  resistivity affects ESP effi-
ciency by  limiting  the  current and voltage.  If  electrodes  are clean,  the
voltage can be increased until a sparking condition is reached.  The maximum
voltage is determined principally by the gas composition and ESP dimensions.
When dust  is  deposited on the collection electrode, however, the voltage at
which sparking  occurs  is reduced because of the increased electric field at
the  dust  surface.   As  dust resistivities  increase,  the voltage  at which
sparking  occurs  decreases.   At  values of  resistivity  above approximately
1012 ohm-cm,  the  voltage must be reduced  so that  sparks will not propagate
across the  interelectrode space.  At very low values of  current and voltage,

dust breakdown  can  occur.   This can result  in  a back corona in which posi-
tive ions  form and  flow back  toward  the  discharge electrode,  neutralizing
the negative charge previously  applied and thereby limiting ESP performance.
     In  addition  to  reducing  the  performance  of an  ESP,  high-resistivity
dust can cling more tenaciously to collection electrodes than particles with
intermediate resistivity.   A much  greater rapping acceleration must then be
applied  to  the electrode to remove the dust layer.   This  can  cause severe
reentrainment or  damage  to  a precipitator that is not designed to withstand
such high acceleration.
     Particle resistivity  depends  primarily on  the  chemical  composition of
the ash,  the ambient  flue  gas temperature,  and the amounts  of water vapor
and S03  in  the  flue gas.12   At low temperatures <250°C, current conduction
occurs principally along the  surface layer of the dust and is related to the
absorption  of  water vapor  and other  conditioning  agents in the flue gas.
For fly  ash, resistivity is primarily related in an  inverse manner to the
amount of SQ3  and  moisture  in the flue  gas.   Burning of  low-sulfur coal
releases  smaller  amounts   of S02,  which  is  oxidized to  S03.  A  higher-
resistivity fly ash results,  except at temperatures below about 250°C, where
significant amounts of S03 are  absorbed onto the fly ash particles.
     At  elevated  temperatures  >25Q°C,  conduction  takes  place  primarily
through  the  bulk of  the material, and resistivity  depends  on  the  chemical
composition of the material.  For fly ash at temperatures above 250°C resis-
tivity is generally  below about 1010 ohm-cm.  Resistivity has been shown to
decrease with increasing amounts of sodium, lithium, and iron.13
     The  range  of operation  of cold-side fly ash precipitators  is  110° to
20Q°C>  a range  in which  conduction  takes  place by  a combination  of the
surface and  bulk  mechanisms and resistivity of  the  ash is  highest.   Figure
4.3-4 illustrates this relationship.   ESP performance as a function of particle size.   The  per-
formance of  an  ESP  varies  considerably with  changes   in the particle size
distribution.  Particles from 0.1  to 1.0 umA are the  most  difficult for an
ESP to  collect.   Usually,  the  greatest  penetration  through an ESP  is  by
particles 0.2  to  0.4 pmA  (Figure  4.3-5).   This  penetration   is  probably
caused  by a transition from field charging to diffusion charging.

                       1000/TEMP., °K
             3.2    2.8   2.4   2.0   1.6   1.2
             	.   i   i   i   i   ii   i   i  i   i
 SURFACE        \   |
         OF SURFACE
         AND VOLUME,
70    150   250    400
  100    200   300
                                    600 800  1000
                       TEMP., °C
 Figure 4.3-4.   Typical temperature-resistivity relationship,
     (reprinted with permission of Academic Press,   Stern,
        Arthur C.,  editor.   AIR POLLUTION, 3rd edition,
                      Volume IV.  1977).7

      0      1      2      3     .4     5      6     7     8      9     10

                      AERODYNAMIC PARTICLE DIAMETER, ymA
      Figure 4.3-5.  Penetration  as  a  function  of particle size for

              an ESP on a  kraft pulp mill  recovery boiler.9

     Another problem posed  by  particles <1 |jmA in diameter is the reduction
in operation  current associated  with  the electrical space  charge  of these
fine particles.   Introducing large quantities of fine particles at the inlet
of the ESP  can  increase the electric field  at the collecting plate, weaken
the  field near  the discharge  electrode,  and suppress  corona  generation.
This  is  known as corona  quenching;  it  occurs when many  incoming fine par-
ticulates acquire the same negative charge as the ions producing the charge,
resulting in an electrical repulsion that tends to reduce operating current.
     Performance as  a  function  of particle size has been  measured at many
installations and has  been  the subject of  computer  modeling.   Probably the
best  known  and  most versatile model is the  EPA-Southern  Research Institute
Mathematical Model.14  This model can be used for sizing and troubleshooting
ESP's as well as for predicting penetration.
     The  effects of  gas  velocity  distribution,  mass entrainment,  and gas
sneakage  on  penetration can also be modeled.  The following descriptions of
the  EPA-SoRI  model  is  excerpted from Ensor et a!.15  Computations are based
on the  Deutsch  equation.   The migration velocity for particles  of a given
diameter  is assumed to be constant over the collection area.  The efficiency
is  calculated for  each  particle size  over  an  incremental  length  of the
precipitator.  The  incremental  lengths  must be sufficiently small to assure
a  constant  electric  field  and  particle  charge.   The total  efficiency is
integrated  over  each particle  size over the total length  of  the ESP.  The
calculation  involves  an iterative solution based on  an  initial  estimate of
the  total  collection  efficiency.   Empirical  corrections  to  the migration
velocities  for  each particle  diameter  are  used  to  verify  the  model.  The
space charge due to  particulates  is calculated  in  order  to determine the
reduced  free  ion density used to obtain the particle charge.  The electrode
spacing  and  applied voltage are  used to calculate the electric field migra-
tion  velocity.   The nonideal behavior  created by gas velocity distribution
is determined empirically.  The .correction factor F is determined using this

          F = 1  + 0.77 01'786 + 0.0755 a An.  (3-—)               (Eq. 4.3-1)
     a =  the  normalized standard  deviation of  the gas distribution


     rj = the ideal collection efficiency.
The nonideal penetration is then calculated by:
          Pt = exp (- |)                                         (Eq. 4.3-2)

     P = penetration
     k = constant predicted under ideal conditions by:
     k = &r\ Or——)                                              (Eq. 4.3-3)
Gas sneakage is another factor that can deteriorate ESP performance when gas
bypasses the electrified  areas by flowing through the hopper or through the
high-voltage insulation space.  A correction factor, B, analogous to the gas
flow quality factor,  F,  can be used as a divisor for the migration velocity
in the exponential argument of the Deutsch equation.

     B =	£H (1 - q)	                             (Eq. 4.3-4)
         NS AnCsn + (i - sn)(i - n)  s]
     N  = number of baffled sections
     S  = fractional amount of gas sneakage per section.
The model  assumes that  perfect mixing follows each  baffled section.   This
model  also  considers  the  effects of rapping reentrainment on collection ef-
ficiency.   It  is  assumed  that the fraction of  material  that is reentrained
does not vary  with particle size or position and that the reentrained mate-
rial  is perfectly mixed  in the  gas  stream after  rapping.   The  effect on
penetration is determined by the empirical relationships:

          Y1 = (0.155) x°'905 — cold-side ESP
          Y2 = (0.618) x°*894 — hot-side ESP
Y is the rapping  emissions and X is the mass removed by the last electrical
section; units are mg/DNm3.
     Efficiencies with no  plate rapping are used by  the model  to determine
the mass removed by the last electric particle field.   A log-normal approxi-

matlon was  used for  particle  size of  the rapping emissions  based  on data
from six different ESP field tests.
     Figure 4.3-6 presents a comparison of data from the EPA-SoRI model with
test data from  the  George Neal No. 3 ESP, operating on a low-sulfur pulver-
ized-coal-fired boiler.15  Best  agreement comes with a = 0.6, S  =0.1, and
a rapping mass  median diameter of 6 [jmA.  The curves do not match, however,
at  the  low  or high  particle  size range.  The  ideal  values  for velocity
distribution and sneakage are believed to be higher than normal for a modern
ESP  and  more the actual  occurrence at the George Neal  plant.   Back corona
was  not  included  in the model, and this is believed to be the cause of most
of  the performance  degradation.   The results from this  ESP were unusual in
that  the curves for  penetration versus  particle  diameter  showed  a double
peak  in  penetration, one  at about 0.2 umA and  the  other  at about 1 umA.
Also,  a  major  limitation of the model was the use of correction factors for
nonideal  conditions,  which were  unobtainable  under  normal  plant operating
conditions.   For  example, the  gas velocity distribution  is often obtained
with  velocity   traverses  with  the boiler  and ESP off  and the  draft fans
operating.   Thus, the comparison of the field data with data from the theo-
retical model was qualitative rather than quantitative.15
     Some ESP  manufacturers also  have developed  models  for  use  in sizing
precipitators.   An example is a Research Cottrell (RC) model used to predict
penetration  of  particulate through  an ESP as  a  function  of particle size.
Although  much  simpler than the EPA-SoRI model, it includes many of the same
concepts  in  calculating penetration as a function of particle size.
     Predicted  penetration as a function of particle size by the RC model is
presented  in Figure  4.3-7,  for a  cold-side ESP  on  a pulverized-coal-fired
boiler.   The computer model  shows the  expected  maximum penetration in the
0.2  to 0.4  MmA size  range.  Comparison of the computer model with test data
shows  an  excellent correlation, as  indicated in Figure 4.3-8.
     The  RC  model has been modified in more recent publications16 to include
a  reentrainment factor,  which  can be  used to  estimate  emissions  due to
rapping  reentrainment at  sources  such as  bark/fossil-fuel-fired boilers.
Limitations  of  this rapping reentrainment factor are similar to the ones in
the  EPA-SoRI model:   (1)  overall  migration velocity remains constant, (2)
the  fraction of material reentrained remains constant for different particle


    0.1   r
                                              EXPERIMENTAL  DA1A
                                              (O  Submicron Particle Mode)
                                              Sn =0.1   a - 0.6  RAP HMD
                                              Sn = 0.1   o = 0.6
                                              Sn « 0.1   o=0.2
                                              Sn « 0    o=0
                             RAP HMD = 2.0
                             RAP HMD « 6.0
                                        OVERALL EFFICIENCY 97.45
                                        SCA 748 ft2/(1000 ft3/min) 147 M2/(m3/sec)
1                       10
 DIAMETER (microns)
            Figure 4.3-6.  Comparison  of experimental  penetration as a
               function of particle  diameter  to the McDonald  (1978)
                   computer model under normal  SCA conditions.16

                                          I   I  I  I   I  I
                                                                    III  1
                                                 x  = 12,  0 =  3.8

     0.01  0.020.03    .2   .3  .4 .5.6..8^0     2   3  4 5*6  8 10
                         PARTICLE AERODYNAMIC DIAMETER,  y

            Figure 4.3-7.  Penetration, pulverized-coal-fired  boiler

                                (cold-side ESP).2

                                    I    I   I
                                                       ACTUAL FIELD DATA  -

                                                           A- 98 3%
                                                           Q- 98.8%
                                                           0- 98.6%
                                                           0- 99.21
                                                           X- 99.5%
                                                (Inlet:  x=3,2;  0=2.72)
           FOR n = 99%
                 i  tii  ii
       i    V   ill  ii
                                                               I   I  I  I   I
     0.1   0.2 0.3   0.5
3  456  810
20  30   50   80 100
                        AERODYNAMIC PARTICLE DIAMETER, u

             Figure 4.3-8.   Computed versus actual penetration  for
            cold-side ESP on a western subbituminous-fired boiler.2

sizes, and  (3) the  fraction  of material  reentrained remains  constant  for
each mechanical section.
     Figure 4.3-9 shows predicted penetration as a function of particle size
for four different  levels  of  particle reentrainment, based on the RC model.
Note that the maximum penetration is still in the 0.2 to 0.4 \jank size range.
Experimental  field  test  data are  needed  to  support  the validity  of  the
reentrainment factors.
4.3.3  Design of Electrostatic Freeipitators
     This section deals with ESP sizing techniques and design considerations
for the major ESP components.   Sizing Equations and Techniques.   The   first  equation  for
predicting  particle collection  probability was developed  by Anderson  in
1919.  It was derived again in 1922 by Deutsch, who used a different method.
In  various  forms, this equation,  n. = 1-e       , has  become  the basis  for
estimating  precipitator  efficiency on  the basis of  gas  flow, precipitator
size,  and   precipitation  rate.   In this  equation,  n  is the precipitator
collection  efficiency, A  is the total collecting electrode  surface area,  V
is  the gas flow  rate,  and w is the  migration velocity  of  the particles.
When  determined  empirically,  the precipitation  rate  parameter, w, includes
effects of  rapping  losses,  gas flow distribution, and particle size distri-
     The  Deutsch-Anderson model assumes  that particulate  concentration  is
uniform  across any  section perpendicular to the gas flow of  an ESP.  This
assumption  is made  because of the  turbulence of the  gas,  which takes  the
particles near the collection surface and allows them to become electrically
charged.   A serious limitation  in use of  the Deutsch-Anderson equation is
that  it  does  not account for changes  in  the particle size distribution and
subsequently  in the effective migration velocity, as precipitation proceeds.
This  limitation affects the accuracy of sizing estimates for units operating
at  very  high  efficiencies  (approximately 98  percent and above) because of
the change  in w with particle size.
      In  practice, factors  such as  particle reentrainment and gas  leakage
cannot  be  accounted for  theoretically.   Also,  some  of  the  most important




§ 0.05

S 0.02




                              I     II   I   I   I  I  I
                                                                                   I     I    1(111

                                                                                     x = 15.0,  a = 4.0
                              i      i    i   i   i   1  I  I
                                                                                   I     i    l   I   I   1  1
                              O    O   OOOOO«—

                                           AERODYNAMIC  PARTICLE  DIAMETER, y

     Figure 4.3-9.  Predicted precipitator penetration  for bark/fossil-fuel-fired boilers

                (k = 0.6392 x 10+1* m"1*; n = 4 fields;  r =  percent reentrainment).

physical and  chemical  properties  of the particles and gases  often  are un-
known.  Therefore, most designers use an effective precipitation rate param-
eter, w ,  that is  based  mainly on  field experience  rather  than theory.17
Data  from  operating installations  form a general basis for selection of w ,
and these  data  are modified to fit the situation being evaluated.  Thus, w
becomes a  semiempirical parameter  that can be used  in the Deutsch-Anderson
equation or its  derivatives to estimate the collection area  required for a
given efficiency and gas flow.
     The most important parameters that determine w  in practice are resis-
tivity, particle  size distribution,  gas velocity distribution  through the
ESP,  particle loss  due to reentrainment,  rapping  and  gas   sneakage,  ESP
electrical  conditions, and  requied efficiency.18
     Matts  and  Ohnfeldt   developed  a  semiempirical  modification  of  the
Deutsch-Anderson equation  that essentially removes the size dependency from
w.19  This equation is n,  = 1-e wk^      .   In  most cases, k equals approxi-
mately  0.5.   The  modified  migration velocity, w. , can be  treated as being
independent  of charging  voltage  and current levels  and of  particle  size
distribution  within an ESP,  as precipitation proceeds in  the direction of
gas flow.  Other changes,  however, such as in properties of the gas entering
the ESP, resistivity, and size distribution, produce a change in w, , just as
they  change   the  conventional w.  Other investigators have  proposed other
forms of the basic penetration equation.20,21
     Another  design  technique  applied to  existing  installations  or new
processes  to  aid  in  a full-scale design  is  the  pilot-scale precipitator.
The main problem with use of a pilot-scale ESP is  that the pilot unit almost
always  performs  better than  a full-scale  unit  because  of better gas flow
distribution, sectionalization,  and  electrode  alignment.1,17   The result is
operation  at higher  current  densities and voltages  than in  a full-scale
unit.   Application of  a scale-up factor, as in spark-limited operation of a
pilot-scale  ESP,  can  cause   uncertainties  in  sizing  the full-scale  ESP.
Therefore, pilot precipitator  data should be supplemented as fully as possi-
ble  by  basic  data on particle and gas properties, especially resistivity.12
      The most important parameters that affect  the  size  of an ESP are col-
lection  area, gas velocity,  aspect ratio', and structural  considerations.

Some variation  of the  Deutsch-Anderson  equation is  generally  used to  es-

timate  the  required  collection  plate  area.   Figure  4.3-10  presents  the

relationships  of  specific  collecting  areas   (SCA's)  developed  with  the

Deutsch-Anderson w  and Matts and Ohnfeldt w.,.
                  6                         K
                       100  200 300  400  500 600
                      SPECIFIC COLLECTION AREA,
                            ft2/1000 acfm

        Metric Conversion:  ft2/1000 acfm x .055 = mVlOQQ m3/sec
         Figure  4.3-10a.   Precipitator penetration versus specific
                collection area and precipitation rate w  .1*

            z 0.01
            2 0.02
            g 0.03
            g 0.05

            5 o.i
                       100  200 300 400  500 600
                     SPECIFIC COLLECTION AREA,
                           ft2/1000 acfm

        Metric conversion:  ft2/1000 acfm x .055 = m2/1000 mVsec

    Figure 4.3-10b.   Precipitator  penetration as  a  function of specific
      collection  area and  modified precipitation  rate parameter w. .19

     The use  of sizing equations  is  only part of the  procedure for deter-
mining the final collection area.  Each manufacturer has a method of comput-
ing  required  plate area,  usually  involving the  use of  computer models to
assist  in  the sizing  procedure,  determination of the  amount of redundancy
requested by  the user  or believed necessary by  the  manufacturer,  and some
amount  of  judgment.   -A  recent  example  of  the  current approaches  of some
manufacturers  involves the  sizing of  ESP's  for highly  variable  fuels.18
Based on ash  constituents,  fuel  sulfur,  and the  resultant resistivity, and
assuming  log-normal  relationships  for   log  (resistivity)   and  ash-to-Btu
ratio,  a  contour ellipse  is  drawn from  the bivariate  normal  distribution
(Figure 4.3-11)  using the Matts-Ohnfeldt variation  of  the worst-case fuel.
Iso-SCA  lines are  then  plotted on  the  contour ellipse  and the  SCA line
tangent to the  contour ellipse defines the required collection area for the
most probable worst case.18
                                X= ASH-TO-BTU RATIO

              Figure 4.3-11.  Distribution of ash-to-Btu ratio
               and log (resistivity) for a single fuel field.
              Courtesy of the Air Pollution Control Association
     Designers usually calculate a hypothetical average value for gas veloc-
ity from gas flow and cross section of the precipitator, ignoring the local-
ized  variances  within  the precipitator.   The primary  importance  of  the
hypothetical gas velocity  is  to minimize potential  losses through rapping
and  reentrainment.    Above  some critical  velocity,  these  losses  tend  to
increase rapidly because of the aerodynamic forces on  the particles.   This
critical velocity is a  function of gas  flow,  plate  configuration, precip-
itator  size, and other factors, such as resistivity.   Values for gas veloc-
ity in  fly ash precipitators range from 0.9 to 1.2 m/s in high-resistivity,

cold-side ESP applications, and in all low- resistivity applications, hot- or
cold-side.   For  most applications,  the  values  range  from 0.9  to 1.7 m/s.
     Aspect ratio is defined as the ratio of the length to the height of gas
passage.   Although  space  limitations often  determine  precipitator dimen-
sions,  the  aspect  ratio should be high enough that reentrained dust carried
forward  from inlet  and middle  sections  can be  collected.   In  practice,
aspect  ratios range from  0.5 to  1.5.   For  efficiencies  of 99  percent or
higher, the aspect ratio .should be at least 1.0 to 1.5 to minimize carryover
of collected dust.
     One of the first structural parameters to be determined is the width of
the  precipitator(s).    This  value  is  dependent on  total  number  of ducts,
which is calculated as  follows:
          n = y_                                                (Eq. 4.3-5)
     n = number of ducts
     q _ ^Q-j-gi gas volumetric throughput, m3/s
     V = gas (treatment) velocity, m/s
     h = plate height, m
     s = plate spacing, m.

Treatment velocity,  V,  is a function of resistivity of the fly ash.  Values
of V should range from 1.0 to 1.2 m/s in high-resistivity, cold-side ESP ap-
plications, and in low-resistivity applications, hot-side or cold-side.  For
most other applications the values should range from 1.0 to 1.5 m/s.
     Plate  spacing,  s, is  more or less fixed  by  the precipitator manufac-
turer and  his  experience  with different types  of  fly ash, by velocity dis-
tribution across the precipitator, and by plate type.  Plate spacing usually
ranges  from 15  to  40 cm.   Most precipitators in  the United  States have
spacing of  22.8  cm,  but precipitator designers are now showing a great deal
of interest in larger spacings.
     Plate height is selected from consideration of simultaneously maintain-
ing the required  treatment velocity and also maintaining an adequate aspect

ratio.    Plate  heights  usually range  from  7.2  to  14.4 m.  The  practical
limitation on plate height imposed by structural  stability is obvious.   Each
manufacturer  limits  the  practical  plate  heights  in  accordance  with  his
overall design.
     The width of  the box is indicated by the total number of ducts.   Cham-
ber  (parallel)  sectionalization is  iiv^the  direction  across the  gas  flow,
whereas  series  sectionalization is  in the  direction  of gas flow (Figure
                                  4TH SECTION
                                  3RD SECTIOK

                                  2ND SECTION
                                  1ST SECTION
        Figure 4.3-12.  Mechanical sectionlization of a precipitator.2   Casings.   The  casing  should  be  of  gas-tight weatherproof
 construction.  Major  casing parts are  the  inlet and outlet transition ducts,
 shell  and  hoppers,   inspection  doors, and  insulator  housing.   The casing
 should  be fabricated of materials suitable for  the application.   The shell
 should  be reinforced to handle  the  following:  maximum positive or  negative
 pressure  static or  dead loads  of  all components,  including any equipment
 located on the roof,  superstructure weights,  hoppers,  or dust loads; loads
 and movements  imposed by  connecting  flues  and dynamic  loading  caused by
 vibrators or rappers;  and environmental  stresses  such as those imposed by
 wind,   snow,  and earthquake.22   In addition,  design must  provide  for the
 overall expansion of the casing caused  by the  high flue gas temperature.
 The shell  and  insulator housing should  form  a  grounded steel chamber  com-
 pletely  enclosing  all of  the  voltage equipment  to ensure the  safety of
 personnel.   Dust  Hoppers.    Hoppers collect  the  precipitated  dust  and
 deliver it to a common  point for discharge.   The most  common type of hopper

is pyramidal,  converging  to a round or square discharge.  If the dust is to
be removed  by  screw conveyor, the hopper  usually  converges  to an elongated
opening that  runs the  length of the  conveyor.    In  applications  where the
dust  is  very sticky and  may  build  up on  sloping  surfaces,  hoppers  are not
recommended; the  casing  is  extended to form a  flat-bottomed  box  under the
ESP.   The dust is removed by drag conveyors.
     Plugging  is  a  major  problem  with  hoppers.   Manufacturers  have  used
designs incorporating  vibrators, heaters,  poke  holes,  baffles,  large  dis-
charge flanges,  and steep  hopper wall angles  (55 to 65°)  to reduce these
problems,  but this problem persists at certain installations.
     The hopper should usually not be used for storage.  The trend is toward
larger hoppers so operators can respond to hopper plugging before electrical
grounding occurs or before physical  damage is done to the electrodes.
     Some manufacturers offer a high-ash fail-safe system that automatically
deenergizes high-voltage  equipment   if  high-ash  levels  are  detected.   Some
type of reliable ash level detection is recommended for most hopper designs.
If the preliminary  design indicates potential  problems with  discharge of
ash,  the discharge flange should be no less than 25 cm diameter.  Transition
from a rectangular  hopper to a round outlet  should  be accomplished without
ledges or projections.  Heaters  have been found to be especially beneficial
in the  discharge throat  and up  to  one-third the height of the hopper.   A
low-temperature probe  and alarm  might also be considered.   Hopper installa-
tion and enclosing  the hopper areas are beneficial in reducing heat loss in
the hopper and discharge system.
     Hopper aspect  ratio  (height to width) is an important consideration in
minimizing  reentrainment  caused  by  gas sneakage to the hoppers.  Low aspect
ratio hoppers can be corrected by vertical baffling.   Power Supplies.  A  precipitator power  supply consists of four
basic  components:   a transformer, a high-voltage  rectifier,  a control  ele-
ment, and a control  system sensor.   The system is designed to provide volt-
age at the highest level possible without causing arc-over (sustained spark-
ing)  between the discharge electrode and collection surface.   Transformer-recti fi ers.    The   unit  converts  low-voltage
alternating  current  to  high-voltage  unidirectional  current  suitable  for

energizing the precipitator.  The transformer-rectifiers and radio-frequency
(RF) choke  coils are  submerged in a  tank filled with  a  dielectric fluid.
The  RF  chokes   are  designed  to prevent  high-frequency  transient  voltage
spikes  caused by  the  ESP  from  damaging  the  silicon  diode  rectifiers.23
     The T-R  sets  should be matched to ESP load.  The ESP will perform best
when all T-R  sets  operate at 70 to 100 percent of rated load without exces-
sive sparking or transient disturbances that reduce maximum, continuous-load
voltage  and  corona power  inputs.23  Over  a wide  range  of gas temperatures
and pressures in different applications,  practical operating voltages range
from 15 to 80 kV at average corona current densities of 100 to 3200 mA2/1000
m2  (10  to  20  mA/1000 ft2) of collecting area.   Over 1500 mA, T-R set inter-
nal impedances are  low, which increases the difficulty  of achieving stable
automatic control.   The highest  impedance possible that is commensurate with
the  application  and  performance requirements  should  be  used.   This often
means more sectionalization with smaller T-R sets.  The high internal imped-
ance of  the  smaller T-R sets facilitates spark quenching.   Smaller electri-
cal  sections  localize  the effects  of electrode  misalignment  and permit
higher voltages  in the  remaining sections.
     In general, current ratings should .increase from inlet to outlet fields
(3  to   5  times   for  many fly  ash  precipitators).   Typical  current  voltage
characteristics  of a five-field  fly ash ESP without ash resistivity problems
are presented in Figure 4.3-13.24   Subcircuits.    During  normal  operation,  optimization  of
applied  power to the  precipitator is  accomplished by  automatic power con-
trols, which vary the input voltage in  response to a signal generated by the
sparkover  rate.   Although  older  ESP's used  saturable reactors  for power
control,  modern  ESP's  use  silicon-controlled rectifiers.   Provisions  are
also included to make the circuit current sensitive to overload and to allow
control  in the  event  that spark  level  cannot  be reached.   Although  the
circuits may  vary  among installations, many of the features described below
are common.   Silicon controlled rectifiers (SCR's).   When  the  circuit
breaker  and  control circuit  on/off switch are  closed,  power flows  through
the current-limiting reactor, current transformer, and current signal trans-
former  to  the primary of the  high-voltage  transformer.   The SCR's act as a

     20           30          40.
   Figure 4.3-13.   Typical  fly ash precipitator voltage-current
              characteristics, five fields in series,
                   no ash resistivity problem.24

variable impedance and  control  the flow of power in the circuit.   An SCR is
a three-junction semiconductor device that is normally an open circuit until
an appropriate gate signal is applied to the gate terminal, at which time it
rapidly switches  to  the  conducting  state.  Its operation  is  equivalent to
that of a  thyroton.   The amount of current  that flows is controlled by the
forward blocking  ability  of the SCR's.   This blocking ability is controlled
by  the  firing pulse to  the gate of the SCR.   The  current-limiting reactor
reshapes  the current  wave form  and  limits peak  current due  to  sparking.
Current wave form with  and without SCR's is  illustrated  in  Figure 4.3-14.
     The  firing  circuit module provides the proper phase-controlled signal
to fire the SCR.   The timing of the signal is controlled by (1) the potenti-
ometer  built in  the module,  (2)  the signal received  by  the  automatic con-
troller, and (3) the signal received by the spark stabilizer.
     The automatic control circuit performs three functions:  spark control,
current-limit control, and voltage-limit control.  Spark control.  Spark control is based on storing electrical
pulses  in  a  capacitor for each spark occurring in the precipitator.  If the
voltage of the  capacitor exceeds the preset reference, an error signal will
phase  the mainline  SCR's  back to  a point where  the sparking will  stop.
Usually this snap-action type of control will tend to overcorrect,  resulting
in  longer downtime  than  is  desirable.   At  low  sparking rates,  about 50
sparks  per minute, the overcorrection is more pronounced and causes reduced
voltage  for  a  longer  period,  with  subsequent loss  of  dust and  low effi-
     Proportional control, another method of spark control, is also based on
storing of electrical  pulses for each spark occurring in the precipitator.
The phaseback of the mainline  SCR's, however,  is proportional to the number
of  sparks  in the precipitator.   The main  advantage of proportional control
over spark control is that the  precipitator determines its own optimum spark
rate, based  on  four  factors:   temperature of the gas, ash resistivity, dust
concentration, and internal condition of the precipitator.  In summary, with
proportional  spark  rate  control,  the  precipitator  determines  the optimum
operating parameters.  With conventional spark control, the operator selects
the  operating parameters,  which  may not  be  correct.  Figure  4.3-15 shows

     f    /  27rTIMI:
     v   /

               CURRENT WITHOUT SCR
2ir  TIME

               CURRENT WITH SCR
                    Of  IT
                    I*   »
            2ir   TIME
      Figure 4.3-H.  ESP current wave form with and without
                silicon controlled rectifier$.25
Reprinted with permission of POWER Magazine, copyright McGraw-Hill,

Inc.  1975.


            g -15

                                                     Timt perioifs
                              •Sate output
                                                    n«»     cat-tM
Figure  4.3-15.  Time periods  are shown  as control  system reacts to a  spark
impulse F after steady-state  operation.   Voltage-start ramp  is rapid,  then
                 switches to slow until  cycle is  completed."

                Reprinted with permission of POWER  Magazine,
                      copyright McGraw-Hill, Inc.  1975.

voltage current characteristics  as  an automatic control system  reacts  to a
spark inpulse.
     Some precipltators  operate  at the maximum voltage  or current settings
on  the power  supply with  no sparking.   In collection of low-resistivity
dusts,  where the  electric field  and the ash  deposit are  insufficient  to
initiate  sparking,  the  no-spark condition  may arise.  The fact  that the
precipitator  is not  sparking  does not necessarily indicate that the unit is
underpowered.  The  unit may  have sufficient power to  provide  charging and
electric fields without sparking.  Voltage-limit control.  The voltage-limit control feature of
the automatic control module  limits the primary voltage of the high-voltage
transformer to its rating.  A transformer across the primary supplies a vol-
tage  signal  that is compared to the setting of the voltage  control,  as  in
the case  of  the  current limit.   The voltage control setting is adjusted for
the primary  voltage  rating of the high-voltage transformer.  When the pri-
mary  voltage exceeds  this value,  a signal   is  generated  that  retards the
firing pulse of the firing module and brings the primary voltage back to the
control setting.  Current-limit control.   For  current-limit  control, a trans-
former  in the primary circuit of the  high-voltage  transformer  monitors the
primary current.   The voltage  from this transformer  is  compared  with the
setting  of   the  current  control,  which  is  adjusted  to  the rating  of the
transformer-rectifier  unit.    If the primary  current  exceeds the  unit's
rating,  a signal  is generated,  as  with spark  control,  which  retards the
firing  pulse  of  the  firing circuit  and  this  brings the current back to the
current-limit setting.
                                              ^ .        ',
     With all  three control  functions properly adjusted, the  control  unit
will energize the precipitator at its optimum or maximum level at all times.
This level will be determined by conditions within the precipitator and will
result  in any one of the three automatic control functions operating at its
maximum,  i.e.,  maximum voltage,  maximum primary current, or maximum spark
rate.   Once  one  of  the three maximum conditions is  reached,  the automatic
control will  prevent any  increase  in power to reach  a second  maximum.  If
changes  within  the   precipitator so  require,  the  automatic  control  will
switch  from one maximum limit to another.

     Other features include  secondary  overload circuits and an undervoltage
trip capability  in the event that  the  voltage on the primary  of  the  high-
voltage transformer falls below a predetermined level  and remains below that
level  for  a  period of time.  A  time-delay  relay is also used  to  provide  a
delay  period  in  the  annunciator circuit while  the network  of contacts  is
changing  position  for circuit  stabilization  because  of  an  undervoltage
     Panels containing component modules, the SCR power circuit, DC overload
circuits,  relays,  control  transformers, resistors, main contactor,  current
transformer,  and  other components  are  mounted  in  the control  cabinet  and
should be completely accessible for servicing.   Positive ventilation for the
control unit is  provided  by an intake fan located near floor level.   Venti-
lating air is  exhausted  through an opening (grill-protected)  in  the  upper
rear of the control unit.
     The transformer  enclosure  is  usually a square metal housing  bolted  to
the  top  of the  transformer tank.   The  enclosure protects  the transformer
bushings and electrical connections from weather and also ensures,  via a key
interlock  system,  that none of the electric connections  or bushings can be
handled  until  the associated  control  cabinet  has   been  deenergized  and
     The transformer pipe and guard are used to feed the high-voltage output
of the transformer-rectifier to the support bushings,  which in turn are con-
nected to  the upper  high-tension  support  frame,  from which  the  discharge
wires are  suspended (Figure 4.3-16).  Electrical energization/sectionalization.   The  way in  which
a  precipitator is  energized  strongly affects  its  performance.  Electrical
energization involves the number and size of the transformer-rectifier (T-R)
sets,  the  number of electrical sections, half-wave/full-wave (HW/FW) opera-
tion,  and  changes  in  the voltage-current  characteristics  as precipitation
proceeds in the direction of gas flow.
     Selection of design power  density is often conveniently  based on re-
sistivity  of the  dust.   Table 4.3-1  illustrates  design values of average
power  density  as  a function of  resistivity  for the fly ash applications.

                                                  BUS CONDUCTOR
                                          DISCHARGE  ELECTRODE
                                             SUPPORT FRAME
                                DISCHARGE ELECTRODE
                                   WEIGHT GUIDE FRAME
             Figure 4.3-3,6.  Predpltator charging system and
                           wire hanging system.26

                      TABLE 4.3-1.   DESIGN POWER DENSITY2
ohm- cm
Power density,
w/m2 of
collecting plate
     In a cold-side precipitator an average operating voltage may be between
25 and 45  kV  for 23 cm spacing, whereas for a hot-side precipitator typical
values range  from  20  to 35 kV for 23 cm spacing.   Knowing power density and
operating voltage, one  can estimate the current density.   The density value
of collecting electrode  is not constant for each point in the precipitator.
At  the inlet  section,  where  the  dust loading  is greatest,  the  voltage-
current characteristics differ significantly from those at the outlet, since
the probability  of corona  suppression is greater at  the  inlet and the per-
centage of fine particles is greater at the outlet.
     Some powering arrangements  are shown in Figure 4.3-17 for a variety of
field  and  cell   (chamber)  arrangements.   The main  advantage  in splitting a
mechanical section by  both chamber and section is to provide greater relia-
bility; this is achieved at an increase in cost.
     Reliability of the precipitator relates not only to sectionalization of
a given collection area but also to the addition of collection area or elec-
trical sections.   At  the discretion of the designer  and  in accordance with
specifications  of the  user,  the  degree  of  reliability  can be  defined in
terms  of  a redundant  capacity,  which is a function  of anticipated failure
and time  between  maintenance  periods.   In this context,  redundancy may be
defined as that  additional area in a precipitator  that compensates for the
"normal"  level   of unavailable collecting  area.   The degree  of additional
collection  area  will  depend  upon the  application and  the manufacturer's


PA 11-11
PA. 1
PA 1
j_ _,
i b
t — . 	 j
I 	 1
i 	 i
1 	 r
PA 1
PA _
i 	 1
i_ 	 i
i 	 1
1 	 IN
i 	 1
1 H
l_ J
i — "i
i 	 1
i 	 1
i >-
i 	 1
1 	 1
i 	 1
i 	 1
4-14 PA
1 — TRAn
1 	 TRA,
rn i
§>i i
i _i
21 -11
®H 1
^.i 	 1
experience.   To provide a reliable yet cost-competitive design, the designer
must have detailed information on the composition of the particulate and the
physical and chemical parameters of the gas stream.
     The highest efficiency  of a precipitator is achieved when useful  power
input is maximized.   The number, size, and mode  of operation (half-wave or
full-wave) of  the  T-R sets can  be manipulated  to provide the required cur-
rent density within  each electrical  section of the precipitator.  Full-wave
energization involves  the  powering of two bus  sections  from  a common power
supply,  whereas with double-half wave,  the  two  sections   are  energized
separately.   The  selection of half-wave or full-wave  operation  will  depend
on  the  source-specific  parameters  in the optimization  of  ESP power input.
     In  spark-limited operation on a  cold-side precipitator  treating high-
resistivity ash, half-wave operation allows time during  the  off half cycle
to recover from the sparking condition (spark quenching).  Complete decay of
the charging field and of collection efficiency during the off half cycle is
avoided  because of the capacitive effect of high-resistivity fly ash, which
tends to maintain the field potential.
     In  operation  of hot-side precipitators,  fly ash resistivity is reduced
by  the  increase  in operation temperature, and the capacitance effect of the
fly  ash is reduced.   Thus,  the  charging  field  decays  faster in half-wave
than in  full-wave operation.
     In  summary, the design considerations in sectionalization and energiza-
tion are based on maximizing the  power input to the precipitator to achieve
the highest efficiency from a given collection area while minimizing loss of
performance as a  result of various potential failures.   Reliability of pre-
cipitator performance is  a  function of  application and design experience.
Precipitator energization  depends on the sectionalization configuration and
the current density to be  supplied to each electrical section, as determined
by  chemical  and  physical  characteristics  of  the  dust, dust loading, and
characteristics of the gas stream.  The number,  size, and mode of operation
of  the  T-R  sets can  be  fitted to the sectionalized configuration after the
bus section design has been established.   Electrode  Characteristics.  Discharge electrodes.  Discharge electrodes  may  be cylin-
drical  or square  wire,  barbed wire, or stamped of  formed strips of metal of

various  configurations.   The geometry of the  electrode  determines  the cur-
rent-voltage characteristics; e.g., the smaller the wire or the more pointed
its surface,  the  greater the observed value of current for a given voltage.
     Discharge electrodes may be suspended from an insulating superstructure
with weights  at the  bottom holding  them  tightly in place, or  they may be
rigidly mounted on  mats or frames.  The weighted-wire  type must be stabil-
ized against  swinging in the gas  stream.  An  example of rigid wire systems
is  shown in  Figure  4.3-18.   the  rigid  type  of  discharge electrode system
requires a  high degree of quality control during fabrication  and erection,
making it more  costly than the weighted-wire system.  Replacement or repair
is  expensive  and  time  consuming.   Larger  casings   are  generally  required
because  of  the  greater spacing between plates.  Two potential  problems with
discharge electrodes are summarized below.,2  Electrical erosion.  In situations where high-current sparks
or  continuous sparking  must  be tolerated,  the  larger discharge electrodes
will  provide more   protection  against  erosion  than  will smaller  sizes.
Shrouds  should  be  included  at both  the  top  and bottom  of  weighted-wire
electrodes, and all  interelectrode high-voltage and grounded surfaces should
have  smooth  surfaces  to  minimize spark-over.   Transformer-rectifier sets
should  be   matched  to  precipitator  load,  and automatic  spark  controllers
should  keep voltage  close  to the sparking  theshold.  Contact  between  the
electrode and the stabilizing  frame should be  solid to  prevent sparking.
With rigid  discharge  electrodes,  substantial  reinforcement should  be pro-
vided at the  point  where the electrode  is attached  to the supportframe, so
that a  significant  amount of metal must be lost before failure occurs.  The
rigid type  discharge  electrode  system has  an advantage in that if a wire
does break, there is less chance of it falling against a plate and shorting
out a section of the ESP.  Mechanical fatigue.  Mechanical connections in the discharge
electrode structure  should be designed to minimize flexing and reduction in
cross-sectional  area at junction points.   Connections should be resistant to
vibration and stress, and electrodes should be allowed to rotate slightly at
their mounting  points.   Reducing the  number of welds is  also desirable.
Keeping  the total  unbraced  length of electrode  as   short  as  possible will

              Figure 4.3-18.   Vibrator and rapper assembly and
                        precipitator high-voltage frame.

help minimize mechanical  fatigue.   Adequate tension during the construction
process  is  vitally important for weighted-wire electrodes,  and  hard spring
wire should be used to prevent kinking.
     Schneider et  al.22  emphasized  that electrode wire failure  can  be kept
to a minimum provided:
     1)   Reasonable care  is  taken  during erection in alignment of casings
          and surfaces.
     2)   The support, guide, and stabilizer system is well designed.
     3)   Reliable, properly adjusted voltage supplies are provided.
     4)   There  is good  operating maintenance of  the  dust handling  system.   Collection electrodes.   Collection plates  are  commercially
available in  lengths  ranging  from  1  to  3  meters and heights from  3  to 15
meters.  Generally,  these  panels are grouped with the precipitator  to form
independently  rapped  collection modules.   The  total  effective  length  of
these plates divided by  their effective height is referred to as the aspect
ratio.   Aspect ratios  larger than 1.0 provide longer residence time for the
gas and reduce penetration, all other things being equal.   Although a varie-
ty  of  plates are  commercially available,  their  functional characteristics
are  not  substantially different.  Collection plates should be  straight and
parallel  with the  discharge  electrodes  when  assembled.  This  alignment
depends  on  care  during  fabrication,  shipping,   storage  in the  field,  and
     The ruggedness  of the plate support system  is  also  important since in
many designs it must also transmit rapper energy to the plates.   Each design
should be examined for its operating  limits with  various types  of rappers.
The  effects  of vibration and impact loading at  all  welded points should be
considered.    There  should  also  be  consideration  for adjustment of  plate
alignment if  necessary after shakedown.  Enough  spacers  should  be provided
to maintain alignment and allow for temperature variations.22
     Rapper anvils attached to either dust plate  supports or  rapper header
beams should be  durable  enough to withstand the stress of rapping and main-
tain alignment (no bending of flanges or other local deformations).

-------  Rapper Characteristics.
     4,  Rapper types.   For the weighted-wire design rapping impulses
are provided by  either  single impulse or vibratory  type  rappers,  which are
activated either electrically or pneumatically.
     The  electromagnetic  or  pneumatic  impulse  type  rappers  are  usually
better  for  collection  electrodes and difficult applications, as  a vibrator
usually  cannot   generate  sufficient  operating  energies   without  being
damaged.28  The  magnetic-impulse,  gravity-impact  rapper is a solenoid elec-
tromagnet consisting  of a  steel  plunger surrounded  by a concentric  coil,
both enclosed in a watertight steel case.   The control unit contains all the
components (except the rapper) needed to distribute and control  the power to
the rappers for optimum precipitation.
     During normal  operation, a  dc  pulse through the  rapper coil supplies
the energy to move the steel plunger.  The plunger is raised by the magnetic
field of the  coil  and then is allowed to fall back and strike a rapper bar,
which  is  connected  to  the collecting  electrodes within  the precipitator.
The  shock  transmitted  to  the  electrodes  dislodges the  accumulated  dust.
     The electrical  controls provide  separate  adjustments  so  that rappers
can be assembled into different groups and each group adjusted independently
to  achieve  optimum rapping  frequency and intensity.   The controls  are ad-
justed  manually  to  provide adequate  release of dust from collecting plates
while preventing undesirable stack puffing.
     In  some  applications, the  magnetic-impulse, gravity-impact  rapper is
also  used  to  clean the precipitator  discharge  wires.   In this  case the
rapper  energy is  imparted to  the electrode supporting  frame  in  the same
manner, except that  an  insulator isolates the rapper from the  high voltage
of the electrode supporting frame.
     The vibrator  is also  an electromagnetic device—a  coil that is ener-
gized by alternating  current.   Each  time the coil  is energized,  the vibra-
tion is transmitted to the high-tension wire supporting frame and/or collec-
ting plates through  a  rod.  The  number of  vibrators applied depends on the
number of high-tension frames and/or collecting plates  in the system.
     The control unit contains all components necessary for operation of the
vibrators,  including  adjustments for vibration intensity  and the length of
the vibration period.  Alternating current is supplied  to the discharge wire
vibrators through  a multiple  cam-type timer to  provide  the sequencing and
time cycle for energization of the vibrators.

     The  number  of rappers,  size of rappers, and  rapping frequencies vary
with the  manufacturer and  the nature  of  the dust.   Generally,  one rapper
unit is required  for  110 to 150 m2 of collecting area.  Discharge electrode
rappers serve  from 350 to 2000 m  of  wire  per rapper.  Intensity of rapping
generally ranges from about 35 to 70 J, and rapping intervals are adjustable
over a range of approximately 50 to 600 s.
     For each installation, a certain intensity and time period of vibration
produces  the  best collecting  efficiency.    Insufficient  intensity  in  the
discharge vibrators may  cause heavy buildup of dust on the discharge wires,
which can lead to the following adverse operating conditions:
     It  reduces  the  spark-over  distance  between  the electrodes,  thereby
     limiting the power input to the precipitator.
     It tends  to  suppress the formation of negative  corona and the produc-
     tion of unipolar ions required for the precipitator process.
     It alters the normal distribution of electrostatic forces in the treat-
     ment  zone.    Unbalanced  electrostatic fields  can cause  the  discharge
     wires and the high-tension frame to oscillate.
     Rigid frame  designs generally  utilize  mechanical hammer rappers.   In
these installations,  each frame is rapped by one hammer assembly mounted on
a  shaft.   A  low-speed gear motor.is  linked  to the hammer  shaft  by a drive
insulator, fork, and linkage assembly.  Rapping intensity is governed by the
hammer weight, and rapping frequency is governed by  the  speed of the shaft
rotation.   Solids  Removal  Equipment.   Solids  removal  from ESP's  may be
performed  by pressure  or vacuum  system in  large  systems  such  as utility
applications, and by screw conveyor in many smaller industrial applications.
Dust can  also  be  wet sluiced directly from the hoppers.  Once conveyed from
the  hoppers,  the  dust can be  disposed  of  dry, or wet  sluiced to  a holding
pond.  Removal  from  the  hopper.   An air seal  is  required at each
hopper discharge.   Air  locks  provide a positive seal, although  tipping or
air  operated slide gate  check valves  are  also  used.  Heaters,  vibrators,
and/or diffusers  are  frequently considered because bridging in the hoppers
occurs at least  occasionally.   In trough-type  hoppers,  a paddle-type con-
veyor has been shown to  be the best means of transporting  the dust to the

air  lock.   Dust valves  are  often oversized  to help  facilitate  removal  of
dust from the hopper.   Pneumatic systems.   Figure 4.3-19  shows  a typical  fly ash
type pneumatic  vacuum  system.   The length of a  vacuum system is limited by
the configuration of the discharge system and the altitude above sea level.
Pressure systems are applied when the limits for vacuum systems are exceed-
ed.  When  the  number  of  hoppers exceeds  about 20 and  the  length  of the
system is too great for a vacuum system, combination vacuum pressure systems
may be used.
     A  vacuum  is produced  hydraulically  or  with mechanical  vacuum pumps.
Positive displacement  blowers are used with  pressure  systems.   Vacuum sys-
tems use electric valves and slide gates, whereas  pressure systems use air
locks and slide gates.
     Materials  of  construction  are  extemely important  in  selection  of a
solids  removal   system.   The chemical  composition  of  the  dust and  of the
conveying air,  and  temperatures  at various  points  in the  conveying system
should be determined.
     When material   characteristics  are  known  (material  density,  particle
size,  concentration, and the physical characteristics  of the conveying air
or gas)  the required conveying velocity can be determined.  The design rate
is usually  set  at  20 percent above the theoretical  maximum conveying capac-
ity to avoid plugging.29
     Storage  facilities  for pneumatic dust  handling generally are equipped
with cyclones,  and often with a fabric filter.  The dust  is then conditioned
with water  and/or a wetting agent  and  transported by truck or  rail  to a
disposal site or is mixed with water and pumped to a disposal pond.  Gas Distribution  Equipment.   Proper gas  flow distribution is
critical for optimum precipitator performance.   Areas  of high velocity can
cause  erosion  and  reentrainment  of dust  from collecting surfaces  or can
allow  gas  to move  virtually untreated  through the precipitator.   Improper
gas  flow distribution in  ducts leading to the  precipitator  result in dust
accumulation on  surfaces and high pressure losses.
     Devices  such  as  turning  vanes,  diffusers,  baffles,  and  perforated
plates  are  used to  maintain and  improve  gas  flow distribution.  A diffuser
                                   4. 3-47

                                                                   «H*4«» COLIICTOK
                               Figure 4.3-19.   Typical fly ash  type pneumatic vacuum  system.29

                                           Courtesy of Allen-Sherman-Hoff Company.

consists of a woven  screen or a thin  plate  with a regular* pattern of small
openings.   The effect  of  a diffuser is to break large-scale turbulence into
many small-scale turbulent zones, which in turn decay rapidly and in a short
distance coalesce  into a  relatively  low-intensity turbulent  flow field.30
Two or  three  diffusers may  be  used  in series to provide  better  flow than
could be achieved with only one diffusion plate.30  Gas distribution devices
may require rapping for cleaning.
     Katz31 stresses the  need for uniformity in designing  inlet and outlet
nozzles of ESP plenums and their distribution devices.  Examples of good and
bad flue and distribution device design for inlet and outlet flues are shown
in Figures 4.3-20 through 4.3-22, respectively.
     In multiple-chamber ESPs, louver-type dampers are commonly used for gas
proportioning.  At the inlet, however, guillotine shutoff dampers should not
be used for proportioning because they tend to destroy proper gas distribu-
tion to a chamber.22
     Gas sneakage  through hoppers  can be caused  by  poor gas distribution.
Expansion-type plenums or top entry plenums cause gas vectors to be directed
towards the hopper, and if multiple perforated plates do not fit well in the
lower portion of the plenum  or  if  the lower portion has been  cut away be-
cause of dust buildup, gas is channeled into the hoppers.31  Short-circuit-
ing the ESP and/or reentrainment is the end result.   Gas flow  models.  Gas  flow models are used to determine the
location and configuration of gas flow control devices.  Although flow model
studies  are   not always effective  in  developing  the  desired distribution,
they are at least a qualitative  indicator of the distribution.
     Temperature and  dust loading distributions are also important to effi-
cient ESP  operations.   It is generally assumed  that  the temperature of the
flue gas  is  uniform.   This is not  always  true,  however, and the effects of
gas  temperature  on ESP  electrical  characteristics  should  be considered.32
     Dust  loading distributions  are not modeled at present  for ESP's.  It is
assumed that  the dust is evenly distributed in the gas, and that as long as
the gas distribution is of a  predefined quality, no dust deposition problems
will occur.   However,  problems  such as poor flue design, poor flow patterns
at  the  inlet nozzle of  the  ESP plenum, and flow  and wall  obstructions can
cause dust deposition  that gas flow models cannot anticipate.31


           (3) PERFORATED
           (1) PERFORATED
                     SAS FLOW
                                     .LOW  FLOW
    Figure 4.3-20.  Effect of two different methods  of aas  distribution
                    on flue characteristics in  an  ESP.30

Courtesy of McGraw-Hi11.

                                          Cm Flow
       Figure 4.3-21 .   Examples of two inlet plenum designs  that
              generally cause gas distribution problems.3>
Figure 4.3-22,  Expansion  inlet plenums showing two methods of  spreading
                             the gas pattern,31
Courtesy of Precipitator Technology, Inc.

-------  Instrumentation.   Instrumentation necessary  for  proper moni-
toring  of ESP  operation can  be  categorized by  location;  i.e., T-R sets,
rappers/vibrators,  hoppers/dust removal systems, and external  items.  T-R sets.  Power  input is the most important measure of the
ESP performance.   Thus any  new ESP should be  equipped with  the following:
                         Primary current meters
                         Primary voltage meters
                         Secondary current meters
                         Secondary voltage meters
                         Spark rate meter (optional)
     These meters  are considered  essential  for performance  evaluation and
troubleshooting.   Figure 4.3-23  shows  a typical control  cabinet and T-R set
     Data loggers  (mainly for  digital  automatic control  systems) are avail-
able  to  help speed  troubleshooting and  reduce operating  labor.   Oscillo-
scopes  are also  useful  in evaluating power supply performance and identify-
ing the type of sparking  (multiple burst  versus  single arc)  but  there is
little demand for such devices.
     It is also  possible to  use feedback signals from transmissometer, full
hopper  detectors,  gas conditioning  systems,  rappers, and  suitable  process
fault indicators in  conjunction with the automatic control unit to  provide
optimum performance  under all  conditions.23  An example is automatic phase-
back  of  T-R  sets  when  hoppers are overfilled,  preventing the  burning of
discharge wires.  Rappers/vibrators.  Microprocessor type technology is avail-
able  for  a high degree  of  rapper control flexibility and  ease of  mainten-
ance.    For example,  in order to prevent  control damage  from  ground faults,
new controls will test each circuit before energizing it.   If a ground fault
occurs,  the  control  will  automatically bypass  the  grounded  circuit  and
indicate  the  problem on an  LED display, thus  permitting  fast  location and
solution of the problem.33
     Instrumentation should be used in conjunction with a transmissometer to
help  in  troubleshooting ESP  problems.   Separate rapping  instrumentation
should  be provided for  each field.   Readings  of  frequency,  intensity,  and


Figure 4.3-23.   Internal  view of one type of rectifier
           console showing component parts.

             Courtesy of Koppers Co., Inc.


cycle time can be used with T-R set controls to properly set rapper frequen-
cy and intensity, in the case of the weighted-wire electrodes.
     For  rigid  frame mechanical  rappers,  cycle  time  and rap  frequency of
both internal and  external  types are easy to measure.   Individual operation
of  internal  rappers is  not easily  instrumented,  nor  is  intensity control
possible without a shutdown of the ESP.34
                                                    i&  ••'                 -   Hoppers.   Instrumentation should be provided  for detecting
full hoppers  and for operation  of the dust valve and dust  removal system.
     Level detectors  can utilize gamma  radiation, sound  capacitance,  pres-
sure differential,  or temperature.31  The alarms should  be  located so that
filling of hoppers  does  not occur but frequent  alarms  are avoided.  A low-
temperature probe and alarm can be used in conjunction with the level detec-
tor.   Control  panel  lights are  used to  indicate  the operation  of hopper
heaters and vibrators.34
     Zero motion switches  are  used on rotary air lock valves to detect mal-
function, as well  as  on screw conveyors.  Pressure  switches and alarms are
normally used with pneumatic dust handling systems to detect operating prob-
lems.   Fans.   Fans should be equipped  with static  pressure gages,
ammeters, and vibration indicators to assist in detecting abnormal operating
conditions.34  Gas and Particle Conditioning.  In the United States, gas and
particle  conditioning  agents are  used primarily to improve ESP collection
efficiency when they are operated on low-sulfur coal  having high-resistivity
fly  ash.   Some  older ESP's designed for  higher-sulfur coal  usually cannot
operate efficiently enough to  meet  emission regulations  on  low-sulfur coal
without flue  gas conditioning.   Other alternatives are increasing the size
of  the  ESP,  cooling  the gas below  the  design  temperature  of  the ESP, or
switching to a hot-side ESP.  Some hot-side ESP's appear to require chemical
conditioning in order to perform as designed.   Conditioning  agents and mechanisms.   The  compounds  other
than water  that  are now  used  or under  study  as  conditioning  agents are
sulfur  trioxide,  sulfuric  acid,  ammonia,  ammonium  sulfate,  triethyl amine,

compounds of sodium,  and compounds of transition metals.   Table 4.3-2 sum-

marizes these conditioning agents and their mechanisms of operation.

        Conditioning agent
       Mechanism(s) of action
Sulfur trioxide and sulfuric acid
Ammonium sulfate
Triethyl amine

Sodium compounds

Compounds of transition metals

Potassium sulfate and sodium chloride
Condensation or adsorption on fly ash
surfaces; may also increase cohesive-
ness of fly ash.  Reduce resistivity.

Mechanism is not clear; various ones

     Resistivity modification
     Increase in ash cohesiveness
     Enhances space charge effect.

Little known about the actual mechan-
ism; claims made for the following:

     Resistivity modification
     Increase in ash cohesiveness
     Enhances space charge effect.

Experimental data lacking to substan-
tiate which of the above is predomi-

Particle agglomeration claimed; no
supporting data.

Natural conditioner if added with
Resistivity modifier injected into
gas stream.

Postulated that they catalyze oxida-
tion of S02 to S03; no definitive
tests with fly ash verify this

In cement and lime kiln ESP's:

     Resistivity modifiers in the
     gas stream
     NaCl:  natural conditioner when
     mixed with coal.

     Sulfur trio'xlde and sulfuric acid are the most widely used conditioning
agents  in the  United  States.   The  primary mechanism  is  condensation  or
adsorption on ash.  Handling of both of these compounds is difficult because
they are  highly  corrosive  and toxic liquids that must  be vaporized before
injection to the  flue gas.  A typical S03 conditioning system is illustrated
in Figure 4.3-24.
     Ammonia  is  less widely  used  in the U.S. mainly  because  of  its incon-
sistent rate  and lack  of a  clear  indication of the  mechanism by which it
acts.35   In  some cases, ammonia shows a  significant resistivity  reduction,
while in other cases it does not.
     Ammonium sulfate  is  believed  to be a main component of many commercial
formulations presently  used.35   It is thermally decomposed to ammonia, sul-
fur trioxide, and water after being injected into the flue gas at high tem-
perature  about  600°C  (355°F)  and reforms ammonium sulfate as the tempera-
ture decreases.   Although a  number of mechanisms  have been  claimed to be
responsible for  the action of ammonium sulfate, experimental  test data are
not available to  document which mechanism is predominant.35
     Triethylamine  is not being used in the U.S. but other proprietary form-
ulas used here  claim to have the same operating mechanisms, namely, agglom-
eration.  There  is  doubt  as to whether agglomeration  actually occurs since
no tests that measure a change in particle size distribution with triethyla-
mine have been conducted.
     Sodium compounds  have  been  used for both cold- and hot-side ESP's, and
Interest  in these compounds  stems  from their role as a natural substance in
reducing  resistivity  in coal  ash.   Sodium can be added either with the coal
fed into  the boiler or by  conventional  means as a  solid  powder  or aqueous
solution in the gas stream.  The 1-atter would probably result in a reduction
in  the  resistivity since it  is co-precipitated with the ash rather than
being chemically incorporated  into it.   Full-scale  tests of  sodium (soda
ash) both in the lab  and  in  the  field  have shown it  is the conditioning
agent of choice for hot-side ESP's.36  A solution type injection appeared to
offer the most effective utilization of the chemical.
     Tests with  anhydrous  sodium carbonate  injected into  the  flue gas pre-
ceding  a  cold-side  pilot  ESP treating ash from a boiler  firing  low-sulfur
coal  showed a   sixfold reduction   in ash  resistivity, an  improvement  in


                             LIQUID SULFUR
                                                   CONTROLLED TO
                                                                      FLUE GAS TO
                                                  U.S. Patent No. 3.993.429
                Figure 4.3-24.   Flow diagram of  sulfur burning
                         flue  gas conditioning system.
  Courtesy of Robert L. Reveley  (Vice President  Wahlco, Inc.)

achievable  current' densities, enhancement  of particulate collection  effi-
ciencies, and an  improvement in  fractional  efficiency collection capability
(see Figure 4.3-25).
     Vanadium pentoxide and ferric sulfate have been added to flue gas on an
experimental  basis  as  potential  catalysts  for the  formation of  S03,  but
neither has given definitive results.
     Potassium sulfate,  a water-soluble  alkalai,  is another  compound that
has  been tested  in full-scale  applications  as  a  conditioning agent  for
preheating  kilns  in Brazil  in addition to a conditioning tower.38  An aque-
ous  solution  of  potassium  (0.4%  increase  in  water-soluble  K)  reduced
resistivity from  1013  to  1011  ohm-cm and improved  ESP  performance.   Addi-
tional tests with potassium sulfate and sodium chloride at an ESP installa-
tion  on  a  coal-fired  lime kiln  in South Africa yielded similar results.38
Sodium chloride  is  not  recommended as a conditioning agent  for the cement
industry because  of the adverse  effect of even small amounts of chloride on
kiln  clinker  quality.   Addition  of dry  sodium  chloride to  the coal  being
ground for  firing in  a Danish lime  kiln  led to considerable improvement in
ESP performance.38
     In summary,  flue  gas conditioning is often successful in improving ESP
performance  by  reducing  dust  resistivity  or through other  mechanisms.
Conditioning  agents should  not   be seen as a  cure-all for  ESP  problems,
however;  they cannot  correct problems associated  with an  underpowered or
undersized  ESP,   poor   gas  distribution, misaligned plates  and wires,  or
inadequate  rapping.   Thus,  any   existing installation  should  be  carefully
evaluated to determine  that poor ESP performance is due only to high resis-
tivity and  not the  above-mentioned problems.  Conditions for injection of a
chemical  conditioning  agent should  also  be  carefully  studied.   Inadequate
mixing of  the conditioner  can  result  in performance below  that  expected.
     4.3.4  Operation and Maintenance of Electrostatic Precipitators  Safety Considerations.   An important  aspect of ESP operation
and maintenance is  the  safety of personnel.   Besides  the  problems  and pre-
cautions necessary because of confined area entry,  the dangers of high volt-
age must  be considered.   Even after the  power supply has been shut down, a
residual charge may be  retained  by the ESP, which behaves much like a large

S   1-0
o   0.8
1   0.6
            .00   .60   .40 .30 .20 .10  .05   .02  .01  .004  .001
        Figure 4.3-25.  Comparison of mean penetration  results.37

capacitor.   Prior  to entry  of personnel into the  ESP  all  safety interlock
procedures  should  be  followed so  that the  ESP,  is properly  grounded and
electrically discharged.  Modes of Failure in ESP's.  Numerous mechanical and electrical
failures  can  occur  in ESP's.  Often these failures have synergistic effects
that cause  other malfunctions.   Generally the modes of  failure are either
mechanical  or electrical,  although the symptoms  of  a malfunction  may be
manifested as both.
     The most common problems associated with ESP malfunctions are discharge
wire  breakage,  plugged  hoppers  causing  excess  buildup  of material,  and
failure  of   rappers  or  vibrators.   Other  problems are  insulator failures,
inadequate  electrical  energization,  and  changes  in the  process operation
away from the specified design criteria.  A  brief  discussion of the common
failures and their effects on emissions  is presented below.
     4.3,4.2.1  Discharge wires.  One of the most common problems associated
with suspended wire electrode ESP's is wire breakage, which typically causes
an  electrical  short circuit between the high-tension  discharge wire system
and the  grounded collection plate.   The electrical  short  trips the circuit
breaker  and  disables  a  section of the ESP, which remains disabled until the
broken discharge wire is removed from the unit.
     Electrical  erosion,  the predominant cause of failures, occurs when re-
peated electrical sparkovers or arcs occur in a  localized region.   Heating
and  vaporization of a  minute quantity of  metal  occur  with  each  spark.
Sparkover at  random  locations will  cause  no serious  degradation  of the
discharge electrode.  Repeated  sparkover at the same location, however, can
remove  significant  quantities  of  material,  with  subsequent  reduction of
cross-sectional area and ultimate failure at that point.39
     Localized sparking can be caused by misalignment of the discharge elec-
trode during  construction or by variations in the  electric field resulting
from "edge"  effects  of  adjacent discharge and collection  electrodes  at the
top and  bottom of the plates.  Measures  that will eliminate failure at these
points are  adding  shrouds and providing a  round  surface  at the edge of the
collection electrode to reduce the tendency for sparking.39
            ..*V          ,                               ••
     Electrical  erosion  can  also  be  caused  by "swinging"  of electrodes,
which can occur when the mechanical  resonance  frequency  of  the discharge

wire and weight  system is harmonically related  to  the electrical frequency
of the power supply.  The power supply adds energy to the swinging wire, and
sparking  occurs  with  each  close  approach to  the collection  plate.   This
action leads to erosion of the electrode and mechanical failure.39
     Poor workmanship  during construction can also cause electrical failure
of the discharge  electrode.   If pieces of  the  welding electrode remain at-
tached to the  collection plate, localized deformation of the electric field
can lead to sparking and'-failure of the discharge electrode.
     Mechanical  fatigue  occurs at points where  wires  are twisted together,
and where mechanical motion occurs continually at one  location.  This occurs
at the  top of  a discharge  electrode  where the wire  is  twisted around the
support collar.  Methods of reducing mechanical  fatigue include selection of
discharge electrode  material  that is resistant  to cold work annealing after
     Chemical  attack  is  caused by  a corrosive material  in the  flue gas,
which can occur  when high-sulfur coal is  burned and flue gas exit tempera-
tures are  low and  near the  acid  dew point.   Use of  ambient  air to purge
insulator compartments also can cause the temperature  to drop below the acid
dew point in a localized region.  Corrosion can  be minimized by operation at
higher flue  gas temperatures  or  by  use  of hot, dry air  to purge insulator
compartments.  Use  of  good insulation around the ESP  shell to maintain high
temperatures  also  provides  adequate protection within the usual  range of
operating temperatures and fuel sulfur contents.39
     Failure of  some discharge wires can be expected  in all ESP's, although
many  rigid-frame units  experience  many fewer  problems with discharge wire
breakage  than  do weighted-wire  designs.   If it occurs  in a random manner,
this wire breakage  will  not  significantly degrade ESP performance.  After a
number of wires  have  broken  and  have been  removed,  they will  have  to be
replaced  during  some scheduled outage.  It is  important  that wire breakage
does not  occur excessively in any given gas passage between any  two plates.
Excessive wire breakage will  result  in  ineffective charging and collection
of particulate in the  passage missing many wires.   Particulate  removal system.    Failure  of  the   particulate
discharge  and removal  system will  allow material to build up  in  the gas
treatment zone.  This  buildup of material can cause misalignment  of both the


collection plates  and the  discharge  electrodes.   Arcing between  the wires
and plates will  usually  occur.   In some applications, collection plates can
be  warped,  as  well  as  discharge  wire frames  in rigid-frame  designs.   An
outage will be  required  for repair and realignment  of  the  components since
the electrical section would be rendered useless.   Buildup of materials also
can lead  to  the  formation  of clinker-like  material between  the  discharge
electrodes and  the  collection electrodes.   This clinker-like  material  is
formed by  fusion of  the dust when  the high  voltage of the  ESP  is passed
through it.
     Proper design of particulate discharge and removal  systems provides the
best means of preventing problems.   Particulate characteristics such as bulk
density,  flow characteristics, and agglomeration should be considered in the
initial design.   Materials  are generally  more free flowing when  hot,  and
efforts to maintain the wall temperature of the hopper above 200°C will help
reduce hopper problems.61   Insulation of the  entire hopper,  and windbreaks
around the  hopper area  can add to  the available heat  in  the hopper area.
Heaters should  be  installed from  the bottom  apex to at least 2 meters high
on  hoppers on the inlet sections of  the ESP and  1 meter high on the outlet
fields.  Ratings of 6 to 10 W per m2 of hopper surface area should be satis-
factory.31  Reduction  of inleakage  through hopper access doors and cleanout
ports  during  and transfer  from  the  hopper  to the  conveyor mechanism will
reduce the  effects of cooling and  condensation  in  the  hopper as  well  as
reducing the reentrainment of collected dust.
     Hoppers  often  contain  baffle plates  to prevent  gas  sneakage  or by-
passing of the  gas  stream through the hoppers.  Where the baffle plates are
close to hopper works, a "bridge"  of material can be formed causing a build-
up  of  dust  into the gas treatment zone.  This problem can be avoided in the
design stage.
     The use of vibrators on hoppers may cause as many problems as they pre-
vent.   If the particulate tends to pack and agglomerate, vibrators may cause
hopper bridging  rather than preventing it.  The feasibility of using hopper
vibrators should be decided for each individual application.  Rappers.  In  dry  removal  systems, the  dust must be removed
from the  collection surface  periodically.   Effective  rapping depends upon

agglomeration of the  material  on the plate to minimize reentrainment of the
dust.  The key to rapping is to avoid excessive rapping.31   Insulator/bushings.   Suspension   insulators   support  and
isolate the  high-voltage  parts of an ESP.  As mentioned earlier, inadequate
pressurization of the top housing of the  insulators  can  cause ash deposits
or moisture  condensation  on the bushings, which may cause electrical break-
down at the typical  operating potential of 45 kV dc.
     Corrective  or  preventive  measures   include  inspection  of fans  that
ventilate the top housing,  availability of a spare fan for emergencies, and
heating of insulators to prevent condensation.   Electrical  energization.   Electrical  energization  must  be
adequate to  charge  the particles, maintain the electric field, and hold the
collected dust to the collection plates.   Among several  possible  causes of
failure to achieve the required level of power input to the ESP, the follow-
ing are most common:39
     0    High dust resistivity
    . °    Excessive dust accumulation on the electrodes
     0    Unusually fine particle size
     0    Inadequate  sectionalization
     0    Improper rectifier and control operation
     0    Misalignment of electrodes
     0    Inadequate  power  supply range.
If a precipitator is  operating at a spark-rate-limited condition but current
and  voltage  are  low,  the problem can commonly be traced to high-resistivity
dust, electrode misalignment,  or uneven corona resulting from buildup on the
discharge electrode.   The effects of high resistivity are discussed in more
detail in Section 4.3.2 in  terms of conditions specific to utility industry,
where resistivity presents  the greatest problem.
     Failures  in ESP controls  can prevent  the  system from  achieving the
level of power required for normal operation.  Following are the most common
malfunctions in controls:
     1.   Power failure in  the primary system
     2.   Transformer or  rectifier  failure  in secondary  system caused by:
          a)   Insulation breakdown in transformer


          b)   Arcing  in transformer  between high-voltage  switch  contacts
          c)   Leaks or shorts in high-voltage structure
          d)   Contamination of the insulating field.
The  most effective  measure for  correction of  control  failures is  a good
maintenance  program  in which  the  controls  are  checked periodically  for
proper operation.   A daily  log of instruments that  register current, volt-
age, and spark rate can also indicate potential proolems.  Preventive Maintenance Schedule for ESP's.   Daily  maintenance.   An accurate  log should  be  kept on all
aspects  of  precipitator operation  including electrical  data,   changes  in
rapper and vibrator operation, fuel quality, and process operations.  Such a
log  can  provide  information for  diagnostic troubleshooting  when any change
in  performance  occurs.  For  example,  it  is  obvious that  gross departures
from normal  readings on the T-R meter and transmissometer indicate trouble.
It is not so widely recognized that small variations, often too slight to be
noticed  without  checking daily readings,  can  indicate  impending trouble.39
     Problems that usually affect precipitator performance gradually, rather
than suddenly, include (1)  air inleakage at  heaters or in ducts leading to
the  precipitator,  (2)  dust  buildup  on precipitator internals,  and (3) de-
terioration of electronic control components.   Such problems are often indi-
cated by a slight but  definite drift of daily meter  readings away from base-
line values.1
     An operator should usually not try to correct deviant meter readings by
adjusting  control  set  points.   An automatic  control  response range should
accommodate  normal  variations in  load.   When major changes  occur, such as
would result  from  firing a coal substantially different from that for which
the  precipitator  was   designed,  the  precipitator  manufacturer should  be
called  in to  retune  the  installation.39   If  no  such major  changes have
occurred,  then  variant  meter  readings  indicate  problems  that  must  be
detected  and  corrected.    Figure 4.3-26  illustrates  a  log  of electrical
readings  that are  checked  several  times  at a  coal-fired  utility  boiler
installation.  These readings are  used  in troubleshooting  ESP operations.
     Probably 50 percent of  all  electrical set tripouts  are caused by ash
buildup.   Short  of  set tripout,  buildup above the  top  of hoppers can cause


             PHECiriTATOR LOG SMiET










(vpTn CASLE ) KV
P*»TH CABLE 2 *v

PPtR C*Bi-E 2 KV
12 M10








Figure 4.3-26.  Precipitator log  sheet.

excessive  sparking  that erodes discharge electrodes.1   Further,  the forces
created by  growing  ash pil'es can push  internal  components  out of position,
causing misalignment  that  may drastically  affect  performance.   Sometimes
operators  attempt  to  preserve  alignment by welding braces  to hold collec-
ting-electrode plates  in position.  This practice may be inadvisable because
restraining the plates reduces the effectiveness of the rapping action that
keeps them clean.39
     Although  various  indicators  and  alarms  can be  installed to  warn  of
hopper-ash  buildup  and  of  ash-conveyor stoppage,  the operator  can double
check by testing temperature at the throat of  the hopper.   If the tempera-
ture of one or more hoppers seems comparatively low, the hopper heaters may
not be functioning  properly.   Generally, however, low temperature indicates
that hot ash  is not flowing through the hopper and that bridging, plugging,
or failure of an automatic dump valve has held ash in the hopper long enough
for it to  cool.   The  ash subsequently  will build up to the top of the hop-
     If the temperatures of  all  hoppers seem  low,  the ash-conveyor system
should be  checked;  the system may have stopped or dust agglomeration may be
so great that the conveyor can no longer handle all of the fly ash.
     Hopper plugging is  sometimes caused by low flue gas temperature, which
permits moisture condensation.  The temperature of the gas entering the ESP
may  be  too  low,  or   ambient  air may  be  leaking into the flue  gas duct.
Hoppers are particularly prone to plugging during startup  after  an outage,
when they are cold and usually damp.
     Daily  checking of  the control  room  ventilation   system  minimizes  the
possibility of overheated control  components,  which  can cause the control
set points  to drift and can  accelerate  deterioration of  sensitive solid-
state devices.
     A daily  check  of  all hopper and ESP access doors to detect gas irrleak-
age is recommended.  Gas inleakage can cause excessive sparking, corrosion,
and particle reentrainment.  Weekly maintenance.   Rapper solenoid-coil  failures, fairly
common during the  period when high voltage was used,  are  rare with modern
low-voltage equipment.39  Still,  a weekly check of  all  units  is  advisable.
Rapper action  should be  observed  visually, and vibrator operation confirmed


by  touch.    If inadequate  rapping  force  is  suspected,  an  accelerometer
mounted on the  plates  could be used to  verify  that rapping acceleration is
adequate (often,  up  to 30 G  is  required).   This is best done  on  a pretest
     Control  sets must  be checked internally for deposits  of  dirt that may
have penetrated the  control  cabinet filter.  Accumulation of dirt can cause
false control signals and can damage such large components as contactors and
printed circuits.
     Finally, filters  in  the  lines supplying air to control cabinets and to
the precipitator  top  housing  should be checked  and  cleaned if necessary to
prevent plugging.39   Monthly maintenance.   Most  new precipitators  incorporate
pressurized  top  housings  that  enclose  the bushings  through which  high-
voltage connections  are made  to the discharge electrodes within the precip-
itator                                                                  box.
Pressurization  ensures  that   if  gas  inleakage occurs  where   the  bushings
penetrate  the  precipitator hot  roof,  gas  will  flow into  the precipitator
rather  than  out  from  it.  Leakage from  the precipitator  into the housing
could cause ash deposits  or moisture condensation on the bushings, with risk
of  electrical  breakdown  at  the  typical  operating  potential  of 45  kV dc.
     Monthly maintenance  also should include  inspection of bushings visually
and by touch for component vibration, checks  of differential pressure to en-
sure  good  operation  of  the  fan that pressurizes  the housing,  and manual
operation  of the automatic  standby fan  to make sure  it is service-ready.    Quarterly maintenance.     Quarterly  maintenance   includes
inspection  of electrical-distribution  contact  surfaces.   These  should be
cleaned and  dressed and  the  pivots should be  lubricated  quarterly if not
more  frequently,  since faulty  contacts  could  cause false  signals,39  Fur-
ther,  because  transmissometer calibration  is subject  to  drift, calibration
should be verified to prevent false indications of precipitator performance.  Semi-annual maintenance.   Inspection, cleaning, and lubrica-
tion  of hinges  and test  connections should be  performed  semi-annually.  If
this task is neglected, extensive effort eventually will be required to free
test  connections   and  access  doors,  often  involving   expensive downtime.


Performance  tests  may be  required  at any time; they  should  not be delayed
while  connections  are made  usable.   An effective preventive  measure  is to
recess fittings below the insulation.39
     Inspection  of the  exterior for corrosion, loose  insulation,  surface
damage, and  loose  joints can identify problems while repair is still possi-
ble.   Special  attention  should be given to points at which gas can leak out
as fugitive emissions.39   Annual maintenance.   Scheduled  outages  must  be long enough
to allow  thorough  internal  inspection of the  precipitator.   Following is a
summary  of  items  to be checked during an  annual  inspection,  abstracted
primarily from Reference 40.
     1.   Dust accumulation—The  upper  outside  corners  of  a  hopper usually
show the  greatest  accumulation.   A spotlight  can be used to  check for dust
buildup, eliminating the need to enter the hopper.
     2-   Corrosion—Inaccessible parts of  the  ESP  are often  attacked by
corrosion.   Access doors and  frames, which are difficult  to insulate, are
usually  attacked first.   Condensation can occur in penthouses that contain
support insulators;  the  penthouses  are at lower temperatures  than the gas,
and moisture is added also by purge air from the outside.
     Corrosion can occur at several  places in the ESP housing—the underside
of roof  plates,  the outside wall, the space between outside collecting sur-
face plates  and  sidewall,  the back of  external  stiffening  members that act
as  heat  sinks,  and  any  area  not continuously  subject  to  gas  flow such as
corners and  the  upper portion of the hopper  connection to  inlet and outlet
ducts.  All  gas  connections should be checked for inleakage of oil, gas, or
     Corrosion in  these  areas  can be minimized by keeping interior surfaces
hot and by effective thermal insulation of outside surfaces.  Use of heaters
during routine shutdowns or  operation  at low loads  also  may help prevent
     ^'   Rappers—Maintenance   of    the   magnetic-impulse,   gravity-impact
rapper  has  been  discussed.   Many  of  the  rigid-wire  ESP's,  however,  have
mechanical  rappers.  The   drives for  collecting  and  discharge  electrode
rappers  should be  checked  for high  motor  temperature,  unusual  noise, and
level and condition of the lubricant.


     Mechanical rappers  should be  checked  for excessive wear,  shifting of
point of impact,  free  movement of wire-frame  rapper  release,  free  movement
of hammers, and wear on hammer shaft bushings.
     ^'    Hoppers—On  both  weighted-wire and  rigid-wire  precipitators,  the
hopper discharge should be checked for such objects as broken pieces of rap-
pers, wires,  and scale.   Presence  of foreign objects indicates a problem
that should be investigated further.
     5.    Gas distribution plates—Although perforated plates usually do not
become plugged, uneven distribution can  sometimes cause plugging of a por-
tion of  the plates.   If a rapping system is not used,  manual  cleaning is
     6.    Discharge electrodes—Frames  in  rigid-frame,  discharge-electrode
precipitators  should  be  centered  between  two  rows  of  collecting surface
plates  with  a maximum deviation  of +0.6  cm.   Discharge  wires   must be
straight and securely  connected to the discharge frame.
     Weighted-wire  precipitators  should  be  checked  for  missing or dropped
weights.   Removal  of a broken wire that  is  not replaced should be  recorded
on  a permanent  log  sheet.2    In  addition the  location of  broken wires,
location on  the  wire where the break  occurred,  and  the cause of the break,
(erosion,  corrosion,  etc.) should  be recorded.   Discharge  wires  should be
cleaned manually as required.
     7.     Collecting electrodes—Collection  plates  should be  inspected for
warping  due  to excessive heat or hopper  plugging.   Corrosion of lower por-
tions of the plates and portions of  plates  adjacent to door openings indi-
cates  air   inleakage  through  hoppers  or  around  doors,2  Plates should be
cleaned manually as required.
     8.    Suspension insulators—When  insulators  become heavily coated with
moisture and dust,  they may become conductive and crack under  high-voltage
stress.   Cracks  can  be  spotted  with  a  bright  light  during inspection.
Faulty insulators can  cause excessive  sparking and voltage loss  and  can  fail
abruptly or  even explode if allowed to deteriorate.
     9-    Housing—Thick  dust deposits  on  interiors  of  housings   indicate
high gas velocities resulting from excessive  gas  volumes,  a condition  that
should be  corrected.

-------  Situational maintenance.   Certain preventive maintenance and
safety  checks  are  so important  that  they should  be performed during  any
outage  of  sufficient length,  without  waiting for  scheduled  downtime.   Air
load  readings  should be  compared with baseline  values  to detect  possible
deterioration in performance.  Readings taken immediately upon restoring the
precipitator to service  can  serve as a check  on  any changes resulting from
maintenance  done   during  the  outage.   All  maintenance  performed  on  this
situational basis  should be recorded for diagnostic purposes.
     Critical  internal  alignments should  be checked whenever  an outage al-
lows; any  misalignment warrants  immediate corrective  action.   Interiors of
control  cabinets and  top housing should be checked  during  any outage of 24
hours or  more and cleaned if  necessary.   Any outage of more  than  72 hours
provides an  opportunity  to check grounding devices, alarms, interlocks, and
other safety equipment and to clean insulators and bushings.39  Optimization of Performance and Energy Consumption.
     Preoperational checklist.    Before  placing the  equipment  in  operation,
plant personnel should perform  a thorough check  and visually inspect the
system  components  in  accordance with  the manufacturer's  recommendations.
Some of the major  items that should be checked are summarized below:
     Control Unit
     Proper connections to control
     Silicon Rectifier Unit
     Rectifier-transformer insulating liquid level
     Rectifier ground switch operation
     Rectifier high-voltage connections
     High-voltage  bus transfer switch operation
     High-Tension  Connections
     High-tension  bus duct
     Proper installation
     Installation  of vent ports
     Equipment Grounding
     Precipitator grounded
     Transformer grounded
     Rectifier controls grounded

     High-tension guard grounded
     Conduits grounded
     Rapper and vibrator ground jumpers in place
     4,  Air load tests.  After  the  precipitator is inspected (i.e.,
preoperational check adjustment of the rectifier control and check of safety
features), the air  load  test is performed.  Air  load  is  defined as energi-
zation of  the precipitator  with  minimum flow of air  (stack  draft) through
the precipitator.   Before introduction  of  an air  load or gas  load (i.e.,
entrance of  dust-laden gas  into the precipitator),  the following components
should be energized:
     Collecting plate rappers
     Perforated distribution plate rappers
     High-tension discharge electrode vibrators
     Bushing heaters - housing/compartments
     Hopper heaters - vibrators - level indicators
     Transformer rectifier
     Rectifier control units
     Ventilation and forced-draft fans
     Ash conveying system
     The purpose of the air load test is to establish reference readings for
future operations, to check operation of electrical  equipment, and to detect
any improper  wire  clearances or grounds not detected during preparation in-
spection.  Air load data are taken with the  internal  metal  surfaces clean.
The data  consist  of current-voltage characteristics at intervals of roughly
10 percent of the T-R mi Hi amp rating, gas flow rate,  gas temperatures, and
relative humidity.
     For an  air  load test,  the precipitator is energized on manual control.
The electrical characteristics  of a precipitator are  such  that  no sparking
should occur.  If  sparking  does occur, an  internal  inspection must be made
to determine  the  cause.   Usually, the  cause  is  (1)  close electrical clear-
ances  and/or (2) the presence  of foreign matter that  has  been  left inside
the precipitator.
     After the precipitator  has been in operation for some  time, it may be
necessary to  shut  it down to perform  internal  inspections.   At  such times,

it  would be  of  interest to  take air  load  data  for comparison  with  the
original readings.  Gas load tests.  The operation of a precipitator on gas load
differs  considerably  from operation on air load with respect to voltage and
current  relationships.  The  condition of high current and low voltage char-
acterizes the  air  load,  whereas low current  and  high voltage characterize
the gas  load.  This effect governs the operation of the precipitator and the
final setting of the electrical equipment.  ESP Startup Procedures.  The  exact startup procedures for any
ESP  are dependent  upon the  application  and  the manufacturers'  recommended
procedures.   The following general guidelines are usually applicable.
     Prior  to  startup  it should  be  confirmed  that  the hoppers  are  empty
before  closing all  hatches.   Any lumps of material that may impede the flow
of  dust should  be removed.   Several hours before process startup the hopper
heater  should be turned on so that hoppers are warm.
     Insulator  surfaces  should  be  heated  by  early  startup of  insulator
heaters  or  by  the purge air system to prevent electrical  tracking over the
surface of the insulators.
     The point at which the ESP is energized will depend upon the individual
application.  Some processes can be started up with the ESP fully energized,
whereas  others  have  explosive  gas  compositions  that must  be  considered.
Generally,  the  temperature at  each field  is  higher than  the moisture  dew
point before that field is energized to minimize sparking.
     It  is  usually  recommended that the  ESP be  energized  under manual con-
trol to reduce spark-over during startup and that the power supplied to each
section  be  increased  gradually.   Newer  automatic  controls that  sense  the
spark-over  threshold  can  be  set to spark-limit mode and placed on automatic
control.  As the  temperature  and the process stabilize, the controller will
bring the power level to its maximum control level.
     Rappers should be  turned  off through the process  startup and probably
for  a  period of  time after the  ESP  is  energized to  allow  formation  of an
adequate  layer  of  dust on the plates.   The amount of  time  between startup
and  rapper  initiation  will  depend  upon the  application  and on  previous

-------  Normal Operation.  Although  electrical  portions of a precipi-
tator require very  little  maintenance,  the items enumerated below should be
attended  regularly  if  the equipment  is  to  give  optimum service.   It is
considered good practice to assign one plant operator on each shift the task
of checking and  recording  data on electrical equipment  at the start of the
     The  cycle  of  inspection  and  maintenance  during  normal  operation
includes the following components:
     T-R sets and associated equipment and controls
     Transformer enclosure
     Pipe and guard
     Plate rappers
     Top housing
     Insulator compartments
     Upper high tension frame
     Discharge wires
     Collecting plates
     Lower precipitator steadying frame
     Dust collection point (dry or wet bottom}
     Hoppers and screw conveyors
     Precipitator shell.   Using T-R set meters for troubleshooting.   An  operator can
utilize the meters to aid  in diagnosing other problems with an ESP.  General
examples  of  the  effect of changing  conditions  in the gas stream and within
the ESP on control set meters are presented below:40
     1.   When the  gas  temperature increases, the voltage will  increase and
the  current will decrease.   Arcing can  develop.   When  the gas temperature
decreases, the voltage will decrease and the current will  increase.
     2.   When  the  moisture  content of the  gases increases  for  any given
condition,  the  current  and  voltage will also  tend to  increase  in value.
     3.   If reduced voltage occurs  because of a  spark-over,  a rise  in mois-
ture may  allow for an increase in the precipitator voltage  level.
     4.   An  increase  in the concentration of  the particulate will tend to
elevate voltages and reduce current  flow.


     5.   A decrease  in  the particle size will  tend  to raise voltage while
suppressing current flow.
     6.   A higher  gas  velocity through the precipitator will tend to raise
voltages and depress currents.
     7.   Air  inleakage  may cause spark-over  in localized areas, resulting
in reduced voltages.
     8.   A number  of precipitator fields in series will show varying read-
ings,  with  voltage-current ratio  decreasing in the  direction  of gas flow.
     9.   If a  hopper fills with dust causing  a short, the voltage will be
drastically reduced and the current will increase.
    10.   If a discharge electrode breaks, violent arcing can  be observed,
with the meters swinging between zero and normal.                       ,
     11.  If a  transformer-rectifier  unit shorts, voltage will  be zero at a
high current reading.
     12.  If a  discharge  system rapper fails,  the discharge  wires build up
with dust; the voltage increases to maintain the same current level.
     13.  If a  plate  rapper fails, the voltage decreases to maintain a cur-
rent level under sparking conditions.
     Table 4.3-3 presents specific examples of the effect of changing condi-
tions  on  ESP control  set  readings.   The operator can  use T-R  set readings
along with other systematic inspection procedures to optimize performance of
an ESP.   Use  of  ESP performance  curves.    Although  ESP  collection
efficiency  is  reduced by  malfunctions  such as  breakage  of discharge wires
and  deterioration  of power supply components,  rectifiers,  insulators,  and
similar equipment,  a  unit can often be  kept  in compliance with particulate
emission  regulations  by  reducing its  load.   Figure  4.3-27  (top  graph)
illustrates collection  efficiency of  a  four-field utility ESP with  24 bus
sections as a  function of the gross boiler load, depending on the number of
bus  sections  out and  whether they are  in series or parallel.   The  bottom
graph  shows the efficiency needed by the  ESP  to meet a state regulation of
0.38 lb/106 Btu as a function of the ash content of coal (assuming a heating
value of 11,000 Btu/lb).
     These types of graphs can be helpful to the operator.  Knowing the ash
content of  the  coal he is  firing  and  knowing which  bus sections of his ESP










Normal full load
System load fall by 1/2

System load constant, but
increase in dust load
Gas temperature Increases

Gas temperature decreases
ESP hopper fills with dust
Discharge electrode breaks

Transformer rectifier shorts

Discharge system rapper falls

Collection plate rapper fails


Gas volume and dust concentra-
tion decrease, resistance
Resistance Increases

Resistance rises, sparking
increases because of increased
Resistance decreases
Resistance decreases
Resistance may fall to 0 (may
vary between 0 and normal if
top part of electrode is left'
swinging Inside the ESP).
Violent instrument fluctuations.
Arcing can be heard outside the
No current passes from T-R set
to the ESP
Dust builds up on discharge
electrodes. Resistance Increases
because corona discharge decreases
Additional voltage required to
keep current constant.
Sparking increases. Voltage
must be reduced to keep current
V, a.c.








A, a.c.








mA, d.c.







             £ 98.0
           *& 97.0
           &• 96.0
                                  CURVE A
                                                            CURVE B
                                                                                          A4 B4  C4  04 E4 F4
                                                                                          A3 83  C3  03 E3 F3
                                                                                          A2 02  C2  02 E2 F2
                                                                                          A1 B1  Cl  D1 El F1
                                                      COLLECTOR SECTION ARAAHGiHEHT

                                                       SECT10H5 OUT

                                                      1 OUT

                                                       2 OUT  IN PARALLEL

                                                      3 OUT  IK PARALLEL

                                                      4 OUT  IN PARALLEL OR
                                                      2 OUT  IN A SERIES
                                                       S OUT  IN PARALLEL OR
                                                       2 OUT  IN SERIES AND 1
                                                       6 OUT  IN PARALLEL OR
                                                       2 OUT  IN A SERIES AND
                                                                                                                   GAS FLOW
                                                                    OUT IN PARALLEL

                                                                    2 OUT IN PARALLEL
                                                       EXAMPLE:  LOAD 290 HW
                                                                 SECTIONS OUT A1, A2, C2,  F4
                                                          (CURVC A)  EFFICENClf AT 290 W WITH 2 OUT
                                                          IK A SERIES AND 2 OUT IN PARALLEL - 95.3

                                                          COAL-ASH 14*
                                                               MOISTURE  101

                                                          (CURVE D)  EFFICIENCY REQUIRED TO MEET
                                                          STATE REGULATIONS - 96.5S

                                                       TO MEET STATE REGULATIONS REDUCE
                                                              LOAD TO 210 m
  16      18
                      Figure 4.3-27.
           Typical  operating curve  to meet emission regulations
            with partial  malfunctions of ESP.2

are  inoperative,  he can  tell  from the  top graph how much  the  boiler load
must be reduced to keep emissions in compliance with regulations.  Charts of
this type should be developed for each boiler-ESP combination.   ESP shutdown procedures.   ESP   shutdown   procedures  will
impact upon the success of maintaining deposits on the collection surface at
a workable  and acceptable  thickness.31  It is  necessary to  keep  materials
from  hardening on  plates yet maintain  acceptable stack  conditions  during
this  period.   The  exact procedures  must  be determined  on an individual
     Methods  to  accomplish  maximum cleaning  include  reducing  of  air flow
through  the  unit  while  reducing  power  to the  ESP.   Rappers  are  kept at
normal  settings.   Gradually the  fields  are  deenergized as the  fields are
cleaned.  The last fields  of  the  ESP  are kept  energized  while the system
cools and captures most of the reentrained dust.31
     Hopper evacuation  should remain  on at least 1  hour  after  all rappers
are  shut  off  and  2 to 3  hours  after  the fans have been shut down to remove
all  dust.   The object is to remove the dust when  it's  easiest, i.e., when
the temperature is still above 200°C.31
     After  cold  shutdown of the  ESP,  the unit may be  entered according to
safety  interlock  procedures  incorporated  in  that  particular  unit.   All
safety recommendations and procedures should be followed.

 1.   Smith, Wallace B.  et al.   Procedures Manual for Electrostatic Precipi-
     tator Evaluation.   Southern Research  Institute.   Birmingham, Alabama.
     EPA-600/7-77-059.  June 1977.

 2.   Szabo, M.F. and  R.W.  Gerstle,  Operation and Maintenance of Particulate
     Control   Devices  on  Coal-Fired  Utility Boilers.   PEDCo Environmental,
     Inc.  EPA-600/2-77-129.  July 1977.

 3.   A.P. deSeversky,  U.S. Patent 3,315,445 April 25, 1967.

 4.   Jaasund,  S.A.  and  M.R.  Mazer,  "The  Application of  Wet Electrostatic
     Precipitators  for the  Control   of Emissions from  Three Metallurgical
     Processes,"  presented  at  symposium   entitled  Particulate  Collection
     Problems   Using   Electrostatic  Precipitators   in   the  Metallurgical
     Industry, June 1-3, 1977.

 5.   Gooch,  J.P.   and A.H.  Dean,  "Wet  Electrostatic  Precipitator  System
     Study,"   SRI,  Birmingham,  Alabama, EPA-600/2-76-142, PB 257  128,  May

 6.   Mcllvaine  Electrostatic  Precipitator  Manual,   Chapter  VI,  Section L,
     August 1976, p. 5613.

 7.   Oglesby, Sabert,  Jr. , and Grady B. Nichols.   Electrostatic Precipita-
     tion.   In:  Air  Pollution, 3rd  Edition, Vol.  IV.   Engineering Control
     of  Air  Pollution.  Academic Press.   New  York,  San Francisco, London.
     1977.  pp.  189-256.

 8.   Portius, D.H., et al.   Fine Particle  Charging Experiments.  Southern
     Research Institute.  EPA-600/2-77-173.  August 1977.

 9.   Sproull, W.T.  (title) APCA Volume 22, p.  181.  ' 1972 (as presented in
     Reference 8).

10.   Spencer, Herbert W., III.   Rapping Reentrainment in a Nearby Full-Scale
     Pilot Precipitator.  EPA-600/2-76-140.  May, 1976.

11.   Gooch, J.P. and  G.H.  Marchant,  Jr.  Electrostatic Precipitator Rapping
     Reentrainment and Computer Model Studies.   Southern Research Institute.
     EPRI FP-792, Volume 3.  Final Report August, 1978.

12.   White,  H.J.   Electrostatic Precipitation  of Fly  Ash, Part II, Journal
     of the Air Pollution Control Association.   February 1977.

13.   Bickelharyst,  R.E.   Journal  of the  Air  Pollution Control Association.

14.   McDonald, Jack  R.   A Mathematical Model of Electrostatic Precipitation
     (Revision  I) Volume  1, Modeling and Programming.   Southern Research
     Institute EPA-600/7-78-111A, June 1978.

15.   Ensor,  et  al.    Evaluation  of  the  George Neal  No.  3  Electrostatic
     Precipitator.   Meteorology  Research,  Inc.  and  Stearns Roger,  Inc.
     prepared for  Electric  Power  Research Institute.  EPRI FP-1145.  August

16.   Szabo,  M.F.,  and R.W.  Gerstle.   Operation  and Maintenance  of Partic-
     ulate Control  Devices  in Kraft Pulp Mill and Crushed Stone Industries.
     PEDCo Environmental, Inc.  EPA-600/2-78-210, October, 1978.

17.   White, H.J,  Electrostatic Precipitation of Fly Ash,  Part III.  Journal
     of the Air Pollution Control Association.  March  1977.

18.   Frisch, N.W., and D.W. Coy.  Specifying Electrostatic Precipitators for
     High  Reliability.   In:  Symposium  on Electrostatic  Precipitators for
     the Control of Fine Particles, pp. 131-157.  EPA-650/12-7S-016, January

19.   Matts, S. and P.O. Ohnfeldt.  Efficient Gas Cleaning with S.F. Electro-
     static Precipitators.  Flakt, A.B., Svenska Flakt labricken, June  1973.

20.   Feldman, P.L.  Effects of Particle Size Distribution  on the Performance
     of Electrostatic  Precipitators.   Research Cottrell,  Inc.  Bound Brook,
     New Jersey.   Presented at  the 68th Annual Meeting of the Air  Pollution
     Control Association.  June 15-20, 1975.  No. 75-02-3.

21.   Cooperman,  P.   and  G.D. Cooperman.   Precipitator  Efficiency  for Log-
     normal  Distributions.   In:  Symposium on  the  Transfer and Utilization
     of Particulate Control Technology.   EPA-600/7-79-044a.  February,  1979.

22.   Schneider, G.E.  et  al.   Selecting and Specifying Electrostatic Precip-
     itators.   Enviro  Energy Corp.   Chemical Engineering, May 26,  1975, pp.

23.   Hall,  H.J.   Design  and  Application of  High  Voltage Power Supplies  in
     Electrostatic  Precipitation.   In:   Symposium on  Electrostatic Precipi-
     tators  for  the  Control  of  Fine  Particles.   pp.  159-189.   EPA-650/
     2-75-016, January 1975.

24.   Hall,  H.J.   Design, Application, Operation and  Maintenance Techniques
     for  Problems in the Electrical  Energization of  Electrostatic Precipi-
     tators.   In:   Proceedings  -  Operation and Maintenance of Electrostatic
     Precipitators,  Michigan Chapter  -  East Central  Section Air  Pollution
     Control Association.  April 1978.  pp. 19-34.

25.   Diulle, Walter, Precipitator Performance Hinges on Control.  Envirotech
     Corporation  Power, January, 1975, p. 24.


26.  Mellvaine  Electrostatic Precipitator  Manual,  Chapter VI,  Section L,
     August, 1979, p. 5613.

27.  Air Pollution and  Industry.   Van Nostrand Rheirihold Company, New York,
     1972.   R.D. Ross, Editor.

28.  Ito,  R.  and K.  Takimoto.   Wide Spacing  E.P.  is  Available In Cleaning
     Exhaust Gases from  Industrial  Sources.   In:  Symposium on the Transfer
     and Utilization of Particulate Control Technology.  Volume 1:  Electro-
     static Precipitators.  EPA-600/7-79-044a.  February, 1979.

29.  Allen Sherman Hoff  Company.   A Primer on Ash Handling Systems.   1976.

30.  White, H.J.  Electrostatic Precipitation of  Fly Ash,  Part IV.  Journal
     of the Air Pollution Association.  April  1977.

31.  Katz,  J.   The Art of Electrostatic Precipitation, 1978.

32.  Engelbrecht, H.L.   Air  Flow  Model  Studies  for Electrostatic Precipi-
     tators.  In:  Symposium  on the Transfer  and  Utilization of Particulate
     Control  Technology:   Volume  I  - Electrostatic Precipitators.   Indus-
     trial   Environmental  Research Laboratory.   Office  of Energy, Minerals,
     and Industry.  EPA-600/7-79-044a.  February 1979.

33.  Lynch, J.G., and  D.S.  Kelly.   A Review of Rapper System Problems Asso-
     ciated with  Industrial  Electrostatic  Precipitators.   Air  Correction
     Division,  UOP,   Inc.   In:    Proceedings   Operation  and Maintenance of
     Electrostatic Precipitators.  Michigan Chapter - East Central Section,
     Air Pollution Control Association.  April 1978.  pp. 46-61.

34.  PEDCo  Environmental,  Inc.   Data  from  IGCI  manufacturers  used  in:
     Design  Considerations  for  Particulate   Control  Equipment.   Prepared
     under contract  for  the U.S.  Environmental Protection Agency.  December

35.  Dismukes, E.B.   Flue Gas Conditioning in  Coal Fired Power Plants in the
     United States.   In:    Second  US/USSR Symposium on  Particulate Control
     EPA-600/7-78-037 March, 1978.

36.  Lederman,  P.B.  et al.   Chemical  Conditioning  of Fly  Ash for Hot Side
     Precipitation.    In:    Symposium  on  the  Transfer  and Utilization of
     Particulate  Control  Technology,  Vol.  I, Electrostatic Precipitators,
     pp. 79-98.   EPA-600/7-79-044a.  February, 1979.

37.  Schliesser,  S.P.   Sodium  Conditioning  Test  with  EPA  Mobile ESP.   In:
     Symposium on the  Transfer  and Utilization of Particulate Control Tech-
     nology,    Vol.     I,    Electrostatic     Precipitators,    pp. 205-240.
     EPA-600/7-79-044a.   February, 1979.
38.  Peterson,  H.H.   Conditioning  of Dust with Water  Soluble  Alkali  Com-
     pounds.  In:  Symposium  on the Transfer  and  Utilization of Particulate
     Control  Technology,  Vol.  I,  Electrostatic  Precipitators,  pp. 99-112.
     EPA-600/7-79-004a.   February, 1979.


39.   Bibbo, P.P.,  and M.M.  Peaces.   Defining  Preventive  Maintenance Tasks
     for  Electrostatic   Precipitators,   Research  Dottrel!,  Inc.   Power.
     August 1975, pp. 56-58.

40.   Engelbrecht, H.L.  Plant Engineer's Guide to Electrostatic Precipitator
     Inspection  and  Maintenance,  Air  Pollution  Division  of  Wheelabrator
     Frye, Inc., Plant Engineering.  April 1976, pp. 193-196.


     This section addresses the basic operating principles, design criteria,
and  operation  and  maintenance practices  for  major types  of commercially
available  fabric  filters.   The  particle  collection  mechanisms  of  fabric
filtration  include  inertia!   impaction,  Brownian  diffusion,  interception,
gravitational settling and electrostatic attraction.  Particles are collected
either on a dust cake supported on a fabric or on the fabric itself.
4.4.1  Types of Fabric Filters
     Although the  basic particle collection  mechanisms  utilized in various
fabric filters  are relatively similar,  the  equipment geometry  and mode of
fabric cleaning  are  exceptionally diverse.   This diversity is partially due
to the broad  applicability of these devices, which demands various perform-
ance capabilities and physical characteristics.  Diversity of fabric filters
also  results  from  the individual  contributions  of numerous  equipment and
fabric vendors.
     Although fabric  filters  can  be classified a  number  of  ways, the most
common way  is by method of fabric  cleaning.   The three major categories of
fabric cleaning  methods are  mechanical  shaking,  reverse  air cleaning, and
pulse jet cleaning.  Mechanical Shaking.   A conventional  shaker-type fabric filter
is shown in Figure 4.4-1.  Particulate-laden  gas enters below the tube sheet
and  passes  from the  inside bag surface to  the outside  surface.  Particles
are  captured  on  a cake of dust that gradually builds up as filtration con-
tinues.  At regular intervals  a portion of this dust cake must be removed to
enhance  gas  flow through the  filter.  The  dust cake is removed manually on
small systems and mechanically on larger systems.  Mechanical shaking of the
filter fabric is normally accomplished by a  rapid horizontal motion induced
by  a mechanical  shaker  bar attached  at  the  top  of the  bag.   The shaking
creates  a  standing wave  in  the  bag and causes  flexing  of the fabric.  The
flexing  causes  the dust  cake to  crack, and  portions  are  released from the
fabric surface.   Some of  the dust  remains on  the  bag surface  and  in the
interstices  of   the  fabric.   The cleaning  intensity  is  controlled  by bag
tension  and  by  the  amplitude,  frequency,  and  duration  of  shaking.   The

                     INLET CHAMiER &
                     HOPPER BARHE
Figure  4.4-1.  Small  shaker-type baghouse (courtesy of
            Carborundum Division of Flakt, Inc.)

residual  dust  cake  provides a  minimal  resistance to  gas flow,  causing  a
static pressure drop  that  is higher than that of a new clean fabric.   Woven
fabrics are  used  in  shaker-type collectors.  Because  of the  low cleaning
intensity, the  gas  flow is  stopped before  cleaning  to  eliminate particle
reentrainment and allow dust cake release.  The cleaning may be done by bag,
row, section, or compartment.
     Gas flow through shaker-type fabric filters is usually limited to a low
superficial velocity  of less  than 1 m/min.  This  means  that the total gas
flow rate (at operating temperature and pressure) divided by the total cloth
area available should not  exceed the stated  "velocity."   This  parameter is
usually  referred  to  as the air-to-cloth (A/C) ratio  and is  expressed in
m3/m2-min.  High  A/C values may lead to excessive particle  penetration or
blinding, which reduces fabric life.
     Mechanical shaker-type  units  differ with regard to the shaker assembly
design, bag length and arrangement, and type of fabric.   All sizes of control
systems can use the shaker design.   Reverse Air  Cleaning.   Particles can be  collected  on  a dust
cake on either  the  inside  or outside of the bag.   A small cylindrical unit
with external  surface filtering  is shown in Figure 4.4-2.   In this  design
the bags  are  arranged radially and are  suspended  from  an upper cell  plate.
The inner and outer  row  reverse-air manifolds  rotate  around  the unit and
stop at each bag to induce reverse flow.  In this manner the entire baghouse
need not  be  temporarily isolated to allow dust-cake removal.   A reverse air
panel   filter  is shown  in  Figure 4.4-3.  The  reverse air cleaning manifold
traverses the rows of filter panels, cleaning all six layers simultaneously.
A somewhat larger reverse air filter is shown in Figure 4.4-4.
     Regardless of  design  differences,  reverse-air cleaning is accomplished
by  reversal  of the  gas flow through the filter media.   The change in direc-
tion causes  the  surface contour of the filter surface to change (relax) and
promotes  dust-cake  cracking.   The  flow of gas through the fabric assists in
removal  of the cake.   The reverse flow may  be  supplied  by cleaned exhaust
gases or  by a secondary fan  supplying ambient air.

                                 INNER         CLEAN
     REVERSE AIR                ROW REVERSE       AIR
                                MIDDLE ROW
                                REVERSE AIR
 Figure 4.4-2.
Reverse-air  baghouse  (courtesy of
Carter-Day Company).

 Screen frame.
 •filter bogs
      • Walkway
        Cleor sir side
                                                            .Reverse olr
                                                         cleaning manifold

                                                            Grid wall
                                                            Manifold drive
                                                           and sheaves
                                   Rotary discharge
Figure  4.4-3.    Continuous reverse air-cleaning system  for  flat
                   filter sleeves (courtesy  of Flakt,  Inc.).

Figure 4.4-4.  Reverse-air collector.
               (Courtesy of MikroPul  Corporation)

     In  filters with  internal  cake  collection,  cleaning is  accomplished
during off-line  operation with  compartments  isolated.  The  filter  bag may
require anti-collapse rings  to prevent closure of the tube  and dust bridg-
ing.  Cake release  may  be increased by rapid reinflation of the bag, creat-
ing a snap  in the surface,  followed by  a short period of reverse air flow.
Fabrics  in  reverse-air  collectors may  be woven  or  felt.   The  felts are
normally restricted to external surface collection.
     Reverse-air filters  are usually  limited to A/C ratios of less than 1.0
m3/m2-min, but  the  ratio may be higher in certain applications.  Most high-
temperature (>250°C) baghouses are of this type.   Pulse-Jet Clearn'ng.   On pulse-jet  fabric  filters,  particle
capture is achieved partially on a dust cake and partially within the fabric.
Filtering is done on exterior bag surfaces only.  A small pulse-jet baghouse
is illustrated in Figure 4.4-5; this device is typical of many small  instal-
lations.   The bags, supported by  inner  retainers  (sometimes  called  cages),
are suspended  from  an upper cell plate.   Compressed air is supplied through
a manifold-solenoid assembly (not shown) into the blow pipes shown in an end
view.   Venturis  mounted  in  the  bag entry area are  intended  to improve the
jet pump effect.   The  classifier shown  at  the  gas  inlet is  intended to
prevent large particles from abrading lower portions of the bag.
     A sudden  blast of compressed air is  injected into the top of the bag.
The blast of  air creates a traveling wave in the fabric, which shatters the
cake and  throws  it  from the surface of  the  fabric.   The cleaning mechanism
is  classified as fabric  flexing and  with some degree of  reverse air flow.
Felted fabrics are  normally used in  pulse-jet-cleaned collectors,  and the
cleaning  intensity   (energy)  is  high.   The  cleaning  normally  proceeds by
rows, all bags in the row being cleaned simultaneously.  The compressed gas
pulse, delivered at 550  to 800  kPa  results in  local reversal  of  the gas
flow.   The  cleaning intensity  is a  function  of compressed  gas pressure.
Pulse-jet  units can operate  at substantially .higher A/C ratios than the
types previously discussed.   Typical  ratio ranges are 1.5 to 3.0 m3/m2-min.
     The plenum  pulse cleaning method is a variation of the pulse-jet clean-
ing mechanism;   in  this  method an  entire section of  bags  is  pulsed with a
blast of  compressed gas from the clean air plenum.  The intensity of plenum

                       TUBE SHEET



Figure 4,4-5.
Pulse  jet baghouse (courtesy of George A.  Rolfes

                 Principles of  Fabric  Filters
     The  two factors  of  basic importance  in fabric  filter operation are
particle capture and static pressure loss.  Particle capture  mechanisms  on a
microscopic  level  are not fully understood.   Macroscopic behavior, the net
result of all microscopic processes, indicates that fabric filter collection
is not  highly  size-dependent as would be expected in view of the collection
mechanisms.  The  static pressure  loss results  from  forcing the gas  stream
through the fabric and dust cake.  Particle Capture.  Pore sizes (open areas)  of the woven  fabric
through which the  contaminated gas stream passes (open areas) range from 10
to  100 urn,  depending  on  fabric  construction  and  fiber characteristics.1
Initially  the  particles  easily  penetrate this  filter.   As cleaning  con-
tinues,  some particles are retained upon filter elements (normally fibers)
because of the combined action  of the classified collection mechanisms shown
in  Figure 4.4-6.2  This  figure is  a  simple  modification  of the  mechanism
diagrams  of  Figures  4.1-5,  4.1-6,  and 4.1-7.  As the  dust cake builds up,
additional  "targets"  are  available  to collect  particles;  accordingly,
penetration drops to  very  low  levels.
                                                   N ^»
                                                       \  f- ELECTROSTATIC
                                                    wx\    ATTRACTION
                                                             — GRAVITATIONAL
      Figure 4.4-6.  Initial mechanisms of fabric filtration  (courtesy  of
          CRC Press, Inc.  Theodore, Louis and Anthony J. Buonicore.
     Industrial Air Pollution Control Equipment for Particulates.  1976).

     Within  the dust  cake,  inertia!  impaction  is the  dominant collection
mechanism.   The forward  motion  of  the particles  results in  impaction  on
fibers or on already deposited particles.2'3  Although increasing gas veloc-
ities favor  impaction,  they  reduce the effectiveness of Brownian diffusion.
Increasing the  fabric  po.rosity also reduces diffusional deposition.4  Grav-
ity settling of particles as a method of collection is usually assumed to be
negligible.2   This  effect  should be  considered at  low velocities,  how-
ever.4'5  Electrostatic forces may affect collection; however, the impact on
commercial-scale equipment is  not fully understood.6  The magnitude of the
electrical charges may depend relative humidity.6  The source of the charges
could be  triboelectric interaction between the particles and  the fabric or
triboelectric interaction before the particles reach the fabric6 - 9  Donovan,
et al.6 have concluded that the latter is more important.
     The  net result  of the particle  collection mechanisms  is potentially
high-efficiency removal  and a  penetration curve of the types  presented by
Turner10  in  Figures  4.4-7 and 4.4-8.  These data indicate  only a weak par-
ticle size dependence.  The form of the curve is confirmed by data in Figure
4.4-9 from  a shaker type  fabric  filter  (silicon-graphite coated fiberglass
bags) serving a spreader stoker boiler.11  The penetration value scale does
not exceed 0.01 so that the shape of the curve can be illustrated.
     Dennis  and Klemm have  proposed that leaks  through unblocked pores of
100 to 200 urn in diameter in woven fabrics are partially responsible for the
lack  of  a strong  particle size  dependence  evident in  filters using woven
     A direct relationship between penetration and woven fabric pore concen-
tration  was  observed  by  Hall,  Dennis, and Surprenant.14   Gas velocities
through the  pores  may be several orders of magnitude above the average face
     The  particulate matter  emission rate  through the pores should  be  a
function  of the  inlet gas  mass loading  and  the air-to-cloth  ratio.   The
latter  determines  the  actual   velocity  through  the pores.    Figure 4.4-10
illustrates  the relationship  between face  velocity and outlet concentration
from a series of bench-scale tests.  Hall,  et al.14 conclude that the higher
emissions at the  higher air-to-cloth ratios  are attributable  to particle
reentrainment through pinholes.

        196 -

EPA Control Systems Laboratory
  Single-point Impactor Data
  (shake cycle not included)
           —  Overall mass efficiency-;99.7%
                       1.0          2.0
                              Particle size, /»m
Figure 4.4-7.   Baghouse  performance,  lead sinter  machine.
                     EPA Control Systems Laboratory
                       Single-point Impactor Data
                             (not verified)

               ' Normal A/C ratio»3:l
               1 High A/C ratio  «6:1

                Overall mass efficiency at normal A/C ratio=99.76%
                Overall mass efficiency at high A/C ratio  =99.51%

               	I	I	I
                Particle size, /»m
 Figure  4.4-8.    Baghouse  performance,  industrial  boiler.


'e 5
 en 4
Load       Air-to-Cloth Ratio

 Mw          ft3/min/ft2
 6              1.87
                                          11              2.47
      0.1                      1.0-

                      Particle  Diameter,  Specific  Gravity  =  2.0  (g/cm  )

              Figure 4.4-9.  Fabric filter penetration  (adapted from
             •, . •             Reference 11).                     .

                             20   40    60   80   100   I2O   140
                               FABRIC LOADING !W),
   Figure 4.4-10.   Effect  of air-to-cloth  ratio  on outlet concentration.
     Dennis  and Klemm  have presented  a  computerized model useful for  pre-
dicting  performance  of  shaker-type  and reverse-air-type  fabric filters.
This model is discussed in  references  12,  13, and 15.
     Penetration  through  felted  bags  used  in  pulse jet  fabric filters  is
believed  due to  (1) direct passage  through  the fabric  and residual  cake
(especially  following pulse  cleaning  and (2)  seepage of particles  through
the  fabric.16'"   particulate emissions  can increase substantially  at  high
air-to-cloth ratios.16'"   Developmental  work on  models applicable to pulse
jet  filters  are discussed  by Dennis,  Wilder,  and  Harmon,18  and Leith and
     Emissions from baghouses are not necessarily limited to particle pene-
tration through  fabrics.   Localized bypassing  of filter elements can occur
through gaps  in sheet  welds, around  poorly seated  seals  and gaskets,  and
through bag  tears.™'1*  The outlet  particle  size  distribution  would  not

appear  substantially  different than  that of  the  inlet although  some very
large diameter  particles  may not be able to negotiate a pathway through the
narrow gaps.
     Theodore and Reynolds have developed the following orifice equation for
calculating the  increase  in penetration due to pinholes, tears, and missing
                                                                 (Eq. 4.4-1)
    p  = increase in penetration.
                      Q                                          (Eq. 4.4-2)
         Ld  (t + 460)
     4> = dimensional parameter.
     Q = system gas volume, acfm.
     L = number of broken bags.
     d = diameter of orifice, in.
     t = temperature, °F.
    AP = fabric filter pressure drop, in. H20.
     For convenience the increase in penetration is plotted against pressure
drop and  <}> (Figure  4.4-11).   The  significance  of bag  failure  can be seen
from the following example.  If two bags, 6 in. in diameter are removed from
a 1415-ms/min  baghouse  operating at a static pressure drop of 1.15 kPa, the
collection efficiency  is reduced  by 4.7 percent from 99.5  percent to 94.8
percent.   The  outlet  concentration  increases  from  4.56 x 10    kg/m3  to
4.72 x 10   kg/m3.   Note that the  penetration does not depend  directly on
the total  number of  bags  in  the  collector or the percentage of the total
number failed.  The  penetration  depends only on the absolute number of bags
failed.  The pressure drop may be reduced as the bag failures increase, and
care must be taken to adjust the value if a significant number fail.   Pressure  Drop.   The  static pressure  drop across  the fabric
filter is  the  sum of the static pressure drop across the cleaned fabric and
that across  the accumulated  cake.   The latter is a function  of time since

              Numbers on curves are
              value of *
                   Ldz (t + 460)

              where Q 1s 1n acfm,
              d in Inches, and
              t 1n °F.

 Figure 4.4-11.

    PRESSURE  DROP  (in.  H20)
Penetration correction  term as  a function
 of pressure  drop and $.

the last cleaning.   The curve shown in Figure 4.4-12 represents the uniform
deposition of dust of a completely cleaned fabric.  The slope of the line is
calledj<2» the specific resistance coefficient.
     The shape  of the  curve in previously operated systems  is  not usually
like that  shown  in Figure 4.4-12 since  it  is difficult to completely clean
the  fabric.    Dennis and Klemm12'13  have  proposed  that partial  cleaning
results in areas where the  dust cake has partially "flaked  off"  and areas
where a substantial dust cake remains.  This is illustrated in Figure 4.4-13.
Using a parallel .flow approach, they have modeled gas flow behavior through
these different  areas.   The model has adequately predicted  static pressure
profiles in laboratory and commercial fabric filter systems.11'15
     Generally,  static pressure drop  is  proportional  to  the inlet  dust
loading, air-to-cloth ratio,  fabric  structure and cleaning system cycle and
intensity.   Most units are designed to operate at differential static pressures
of 0.5 to 2 kPa, however,  some units operate at differential  pressures up to
3 kPa.
     The static pressure profiles for shaker and reverse air systems tend to
have a distinct  sawtooth  pattern if there  are only a few compartments.  On
larger systems,  this  pattern disappears  and the static pressure drop across
the system is relatively constant.20
4.4.3  Design of Fabric Filters
     A complete  characterization of  the  effluent gas stream is important in
the design of a fabric filter system.   This would  include:  gas  flow rate,
minimum and  maximum  gas  temperatures,  acid  dew point,  moisture content,
presence of large particulate matter, presence of sticky particulate matter,
particulate  mass loading,  and presence of  potentially explosive  gases  or
particulate.    Given  accurate  data  on  these effluent  characteristics,an
appropriate  collector  can be  designed for the required  degree  of control.
Selection  of the type of fabric, and dimensions of the bag normally must be
done in conjunction with the design of the cleaning system.  The size of the
fabric filter depends  on  the air-to-cl.oth ratio necessary and the number of
compartments expected  to  be  out-of-seryice for  maintenance  or  cleaning at
any given  time.   The overall costs must be  balanced against needs for good
accessibility and  instrumentation, both  of which favor improved maintenance
and» therefore, improved performance.









                    AVERAGE  FABRIC  LOADING
                 Figure 4.4-12.  Filter drag profiles.

Figure 4.4-13.  Fly ash dislodgement from 10 ft x 4 in. woven glass
                    bag (inside illumination).

-------  Effluent Characteristics.  The effluent characteristics should
be quantified  to the  extent possible before  fabric filter  design  is com-
pleted.  Variability of these conditions should be considered.
     Gas Temperature.  The  gas stream temperature and  its variability over
time determine the fiber and finish selected for the filter bag.  The tempera-
ture of gases emitted from industrial processes may vary over several hundred
degrees in  short periods  of time.  Low gas stream temperatures may go below
the gas moisture and acid dew points, and  high temperatures may exceed the
maximum that the fabric  will tolerate.  The extremes of temperature must be
determined before fabric selection.
     The temperature of  the gas may be modified  by  using heaters (indirect
or direct fired) to  increase the gas  temperature  above the dew point or by
using  coolers  to  reduce the  gas  temperature to  one compatible with the
fabric.  Methods  commonly  used to increase or  maintain gas temperature are
insulation of ductwork, use of direct-fired afterburners and heat exchangers,
and  steam   tracing.   The use  of  direct-fired afterburners  may serve two
purposes:    to  elevate  the  gas temperature above the dew point and to remove
organics that may blind the fabric surface.
     Methods of   cooling  the gas  include  dilution,  radiative  cooling, and
evaporative cooling.2'21  .The  use of dilution  air to  moderate gas  tempera-
ture is the simplest approach, but it may increase the capital cost because
the  volume  of ambient air  required may necessitate  a substantially larger
filter.  Figure  4.4-14 shows the increased capacity  required when  gas tem-
perature is reduced to 560°K by use of 300°K dilution air.
     Radiative cooling does  not require increased  collector  size,  but it
does require an  investment  in greater duct length.   Also, the static pres-
sure  drop   of  the system  may  be increased by the increased  duct  length.
Figure 4,4-15  shows  the  reduction in gas temperature achieved as a function
of  duct  length  with  gases  at an  initial temperature  of 1150°K.   Radiation
cooling of  gases at temperatures  above 115Q°K requires exotic construction
materials and  may be uneconomical.21  The cooling of  gases at temperatures
below  800°K requires  extensive  surface  area  and is  usually uneconomical
relative  to the cost of  the greater baghouse capacity  required  by. other
cooling methods.21

               300°K AIR USED
               FOR DILUTION
                            0    100   200   300    400

                              INCREASE IN BAGHOUSE CAPACITY, X
       Figure  4.4-14.  Added capacity  needed in baghouse when hot gases
       are cooled by dilution with ambient air (reprinted by permission.
       Vanderhoeck, P.,  Chemical Engineering, May  1,  1972).





       Of 1150'K BAS AT l.WO •/•in. CONTAINING H
                            30          60          90
                            LENOTH OF DUCT (FROM HOT GAS SOURCE), m

640° u


20 g

40 g
60 §

80 g

100 3
       Figure 4.4-15.   Radiation  effectiveness in cooling  hot gases  (re-
               printed by permission; Vanderhoeck, P.  Chemical
                           Engineering, May  1,  1972).

     Evaporative  cooling  is  accomplished  by  injection  of water into  the gas
stream.   The energy drawn from the gas stream to vaporize the water leads to
a reduction  in  gas  temperature.   The cooling is accomplished rapidly and in
a small  space.   The  system must be  designed  to reduce gas temperature to a
point  above  the  gas  dew point, to  prevent  carryover  of  unvaporized water

droplets into the  filter, and to prevent spray water  impact on duct walls or

liners.21  Figure 4.4-16 shows the increase in the necessary  baghouse  capacity
as  a  result  of  evaporative  cooling  (desired  gas  temperature  of 560°K).
Presently, bags  made of graphitized,  siliconized fiber glass can  withstand
this temperature.   The water  temperature has a relatively minor  impact  on
the quantity of water required.
                              300°K WATER USED
                           10    ZO    30    40     50
                           INCREASE IN BAGHOUSE CAPACITY,  %
      Figure 4.4-16.  Added capacity  needed  in  baghouse  when hot gases
         are cooled by evaporating cooling (reprinted  by permission:
            Vanderhoeck, P.  Chemical  Engineering,  May 1,  1972).
     Particle characteristics.   The  particle  size distribution of the  dust
must  be considered  in  design  of  the  collector  and in fabric  selection.
Particle size  distribution affects both  dust  cake porosity and  abrasion of
the fabric.  The presence of  fine particles in the gas stream can create a
heavy  dust cake  and  increase the static pressure drop through  the  cake.22
The small particles can  also cause fabric bleeding.
     The presence of  large  abrasive  particles  can reduce bag  life  and may
require the  use of a precleaner or  gas distribution  devices  in the collec-
tion system (see  the inlet diffuser of Figure  4.4-5).  Moreover, because the
resistance  of  the  fabric  to abrasion  is  greatly reduced when  particles
strike  tangentially,  the  presence of large  particles may require modifica-
tion  of gas  inlet  design.    For certain  sources  such as  spreader stoker
boilers,  a mechanical   collector ahead  of  the  fabric  filter may  provide
protection  from  the  large quantity  of >10  umA particles.  This  would  also
allow  protection  from  glowing  embers and  thus reduce  the risk  of hopper

     Particulate matter with  a high carbon content may  be  generated during
periods of improper combustion.  Fabric filter design should include ways to
minimize  hopper  fires.24'25   These  could  include  continuous removal  of
collected  material,  limited   air  entry  into  hoppers,  and a  fire-sensing
     Sticky particulate can be difficult to remove from the fabric surface.
A survey done by Billings and Wilder indicated blinding of bags due at least
partially  to  sticky particulate was  the most  frequent  problem reported.26
Improper combustion can lead  to  problems of  this type and result in sub-
stantially increased static pressure drop.27
     Gas Composition -  Factors of importance regarding  the gas composition
include moisture content and acid dew point.   If a fabric filter is operated
at  close  to the  acid  dew point,  there is a substantial risk  of  corrosion
especially  in  localized  spots close  to hatches,  in  dead  air  pockets,  in
hoppers, or adjacent  to heat sinks such  as  external  supports.28'29  If the
operating temperature drops below the water dew point, either during start-up
or  at  normal  operation,  blinding  of the  bags  can occur.  The presence of
trace components, such as fluorine, can attack certain fabrics.  Fabric Selection.    Fabric selection is usually  based on the
prior  experience in  similar  applications.   Important factors  to  consider
          Dust penetration
          Continuous and maximum operating temperatures
          Chemical degradation
         • Abrasion resistance
          Cake release
          Pressure drop
          Cleaning method
          Fabric construction.
The choice of fabric ultimately affects pressure drop, selection of cleaning
method,  outlet  concentration, and  the  life of  the fabric under  operating
     Temperature.    Degradation of  the  polymer   in  synthetic  and natural
fabrics is  accelerated  temperature.   The rate of  decay  is  also enhanced by
the actions of moisture, chemicals, and abrasive particles.   Temperatures at

which  reasonable performance  can be expected  under normal  conditions  are
given  in Table  4.4-1,  which gives the continuous maximum operating tempera-
tures  recommended by  fabric manufacturers  and  summarized by  Theodore  and
     The maximum  short-term temperature represents the temperature at which
rapid  deterioration will  result in immediate failure.  For synthetics, this
is  the temperature at  which polymer softening  occurs  and causes permanent
elongation.  Figure 4.4-17  shows the reduction in strength of Nomex^fabric
with increasing temperature,30
     Glass fabrics are sensitive to abrasion between fibers and are normally
                        I I? I
coated with either Teflorc-'or silicon-graphite.  Polymer finishes can degrade
with increasing temperature.  Figure 4.4-18 shows the effect of gas temperature
on the finish of glass fiber bags.31
     Chemical degradation.   Chemical  degrading of  the fabric  is  caused by
the  breaking  of polymer  chains within the  fiber  structure.   The degrading
may  be  from  acid hydrolysis, alkali attack,  or,  in the case of glass fiber
fabrics, conversion of the structure to a noncrystalline form that has lower
     As  the  chain  length  of a polymer is  reduced by chemical  attack, it
loses strength.  The chemical attack may be accelerated by moisture or metal
catalysts  in  the dust  impregnated in  the  fibers.  The  rate  of attack in-
creases with temperature.
     Chemical  composition  of the gas  stream,  moisture content, and temper-
ature must be  considered in selection of the  fabric.  Table 4.4-2 indicates
the  ratings of commercial fabrics with respect to chemical resistance.   Note
that resistance  is a relative term that does not imply total resistance  to a
specific chemical.   Also resistance may be greatly reduced by cyclic opera-
tion under different conditions and concentrations.
     Abrasion resistance.   Resistance  to abrasion  is a  relative  term  that
indicates  the  ability of a  fabric to provide  extended service  in collecting
abrasive  dust.   Resistance  can be modified  by fabric construction, fabric
finish, and shapes of the particles collected.

Stainless stee
(type 304)
Generic name
Natural fiber cellulose
Natural fiber protein
Nylon polyanride
Nylon aromatic

Type yarn
Filament spun
Filament spun
Filament spun
Filament spun
Filament spun
Filament spun
Filament spun

Maximum temperature range, °K
Long periods
of time
Short periods
of time
No data
temperature, °K
420 decomposes
575 chars
440 softens
520 softens
640 decomposes
670 decomposes


» 50
M_ 40


I 30


1 20


1 10

       500 h exposure
       500 ppm S02
       4* 02
       6% H20
       3/1 air-to-cloth ratio
        (m3/m2 - min)
       2 cpm pulse
       power house dust
                        R 14-oz FELT

                                  POLYESTER-18-pz FELT
                             TEMPERATURE,  °F
Figure 4.4-17.  Effect of acid and temperature on  strength  of Nomex
           and polyester fabrics  (reprinted by permission:
                E. I. du Pont de  Nemours and  Company).








                             -COMMERCIAL SILICONE

                             -COMMERCIAL SILICONE
                                PLUS GRAPHITE
    100     200     300
          TEMPERATURE, C



                                                  15 o
Figure 4.4-18.  Effect of gas temperature (continuous) on life of glass
     fabric bags (reprinted by permission:  Menardi and Company).

                          Table 4.4-3.  CHEMICAL RESISTANCE OF COMMON  COMMERCIAL FABRICS2
Stainless steel
(type 304)
Generic name
Natural fiber
Natural fiber
Nylon polyamide
Nylon aromatic

Type yarn
Filament spun
Filament spun
Filament spun
Filament spun
Filament spun
Filament spun
Filament spun
bul ked
Filament spun

Very good
Fair- good
Very good-
excel! ent
Poor- fair
Poor- fair
Poor- fair

Fair- good
Poor- fair
Very good-
Very good-
Flex and
abrasion .
Fair- good
Very good-
Fair- good
Very good-
Very good
Very good-
excel 1 ent


     Pressure drop.  Static pressure drop must be considered In selection of
the fabric.  The residual pressure drop affects the cost of operation due to
an  increase  in fan  horsepower.   In applications where the  fan capacity is
limited,  the increase  in pressure  drop can  reduce  gas  volume  and reduce
capture and transport velocity in the ventilation system.
     Cleaning method.   The  method  of  removing  the  dust  cake  is  closely
related to fabric  construction and fiber type.  With woven fabrics that are
subject to abrasion or flex damage, gentle cleaning methods such as low-frequency
shaking or  reverse air  can  be  used.   With felted fabrics,  a more intense
cleaning method is required,  such as high-pressure reverse air or pulse-jet
cleaning.   An improper  combination of fabric  and  cleaning  method (e.g.,
intense shaking  of glass  bags)  can cause premature failure  of the fabric,
incomplete cleaning, or blinding of the fabric (complete plugging of pores)..
     Fabric construction.  Filter  fabrics  commonly used in operating facil-
ities  are  either  woven  or felted.   Unlike woven fabric, felt is a genuine
filter medium and  is more efficient in collection of particulate at compar-
able filtering  velocities; it is, however, more  expensive.   Felted fabrics
are composed of randomly oriented fibers and are.relatively thick.  Needling
the fibers meshes them and forms a strongly bonded fabric.   The thickness of
felt provides for maximum particle impingement,  but increases  the static
pressure drop  (reduces  permeability).   Felted fabrics are  normally used in
pulse-type units and are operated at high A/C ratios.
     Woven fabrics are  characteristically used  in  shaker  and reverse-air
filters, and are   operated at relatively  low  A/C ratios.    Woven fabric is
made up of filament or staple (spun) yarns in a variety of patterns, having
various spacings between  the yarns,  with a specific finish that is designed
to retain or shed  filter cake, depending  on the application.   Seven of the
most common  weave  patterns   are  shown  in Figure 4.4-19.   Plain  weave is
lowest  in  initial  cost,  and has  the  least porosity  and  greatest particle
retention; however, its potential for blinding is greatest.   Twill weave has
medium retention and blinding characteristics and has reasonable permeability.
Twill weave also exhibits the best resistance to abrasion.   Sateen weave has
the lowest particle  retention,  is easiest to clean of accumulated dust, and
has lowest potential for blinding.









Plain or Taffeta,
  Weave 1/1
3/2 Regular Twill
 "Crowfoot" Satin or
    4 Shaft Satin
    No. 4 Harness
                              2/1 Regular
                                 4/1 Satin
                              (Sateen if Cotton)
                                No. 5 Harness
                                                           3/1 Regular
2/2 "Broken" Twill
   or Chain Weave
                                   Other Popular Weaves:
                                   Drill = 2/1 L.H. Twill, or 3/1 Twill
                                   Herringbone = a type of broken twill
                                   Basket Weave = extension of plain weave
                                   Gabardine = regular or steep twill with
                                             higher warp than fill count
  Figure  4.4-19.   Typical  fabric  weaves  (reprinted by permission:
                   Industrial  Gas Cleaning  Institute).

The permeability of woven fabrics depends on the type of fiber, tightness of
twist, size  of  yarn,  type of weave  (geometric  pattern),  tightness of weave
(thread count), and type of fabric finish.
     Woven fabrics may be provided with a number of finishes.   Cotton fabrics
may  be preshrunk  to maintain  dimensional  stability,  i.e.,  resistance to
stretching or shrinking in any direction, which could adversely affect other
fabric characteristics.   Spun fiber  fabrics may be  napped on  the  surface
receiving the  dust load.  Napping  is the process of pulling  fibers  out of
the yarn  bundles to  form a  soft  pile;  this promotes the  formation  on the
fabric surface of a dust cake that does not penetrate the interstices of the
fabric.   Synthetic  fabrics  may be heat-set  to  ensure dimensional  stability
and provide  a  smooth  surface with uniform permeability.  Any  fabric may be
silicone  treated (also used  in  combination with  graphite  and  Teflon-3 to
improve  abrasion  resistance, to  facilitate  cake  release,  and  to  reduce
moisture absorption.   Selection of Cleaning Technique.   A  number  of  factors  are
considered  in  the  selection of  a  cleaning technique  for  a  fabric  filter
system.  Primarily, the physical  and chemical properties of the flue gas and
particulate must be  clearly  defined.   Specific case studies involving simi-
lar processes,  together with  laboratory studies, are often the most informa-
tive guideposts for design parameters.  Critical interdependences in clean-
ing  technique   selection  are particulate characteristics/cleaning  method,
specific  resistance/cleaning  method,  and  cleaning  method/service  life.
Constraints  imposed  by  intermittent or continuous operation and by availa-
bility of space must also be considered.
     If  the dust  cake is  released  readily from  the fabric,  reverse air
cleaning  may be  adequate.    Reverse  air  can  be  used in  combination  with
mechanical shaking.  Felted fabrics generally are not cleaned by reverse air
because of their greater structural  depth and,  hence, greater dust retentiv-
ity.   Bag tensioning and reduction  of reverse air flow  rates  minimize the
degree of bag  flexure  and  thus  reduce  the risk  of accelerated  bag wear.
These measures  also prevent  complete collapse of the  bag,  which makes cake
removal difficult.  The  rate of  flexure  is  probably  the  controlling factor
with respect to fabric failure.

     Reverse-air  and  mechanical-shake  units are  capable of  being cleaned
only while the  unit or a single compartment is  off line.  In most combina-
tion reverse-air  and mechanical-shake systems,  bag  collapse  and/or flexure
caused by flow reversal are the major dust dislodging forces.
     In shaker systems, the fabric cleaning action is defined by quantifying
shaking frequency,  shaking  amplitude,  and duration of the shaking interval.
Tensioning of  the  bags  is  important  in  determining average  amplitude  and
acceleration  of the  bag.   If  the  bag is  too  slack,  the transmission of
cleaning  energy over the entire  length of the  bag  is  incomplete,  with  the
result  that  cleaning  is   inefficient  and  nonuniform.   This  can  lead to
abrasion damage and reduced bag life.  Sizing.   The  size of a fabric  filter  system  is determined by
the gas volume  to be filtered  and  the  A/C ratio at which the filter can be
operated  in  view  of fabric type, dust cake properties, and cleaning method.
The area  of  fabric surface is  determined  by multiplying  the  total gas flow
by the recommended A/C ratio.
     Penetration is directly  related to the effective air-to-cloth ratio in
the  system,  with  substantially increased  emission  levels  at  high air-to-
cloth  ratios.13'14'34    Accordingly,  the  lowest  possible  face  velocity
consistent with economic constraints should be specified.
     The minimum number of compartments in shaker-type and reverse-air units
is  related  to  the  maximum allowable  increase  in A/C ratio  as  one or more
compartments  are  removed  from  service.   The following  additional  factors
should  be considered  for  compartmentalization  in  fabric  filter  design:
     1.   Large  compartments  may  necessitate  oversized  and  uneconomical
          cleaning equipment.
     2.   Large  compartments contain  more  bags and, therefore,  present a
          higher  probability that bag  failure will  occur in  a single com-
          partment.    This  eventuality   could   necessitate   intermittent
          replacement of bags.
     3.   Large  compartments require  large  ducts and dampers;  small  com-
          partments  require more  numerous,  but smaller,  ductwork branches
          and dampers.
     4.   Large   compartments   require   larger  solids-collection  hoppers.

     5.   Larger  compartments cool  more slowly  when  brought off  line for
     Most of these considerations favor smaller, more numerous compartments.
     The compartments  of  shaker  and reverse air  should  be arranged so that
maintenance of all filter tubes is relatively simple.  The number of rows of
bags along  each walkway  should  be  minimized  so that the bags  nearest the
compartment walls  can be  reached without disturbing too many bags near the
walkway.  The  appropriate  "bag reach" is determined by  the layout of tubes
and the  diameter.   Normally there are no more than 3 to 4 bags deep.35'36   Solids Removal Equipment.    As  solids  are  cleaned  from the
filtering fabric,  they fall  to  a  collection  hopper for  ultimate removal.
The "fluid" properties of the collected solids  are  important in design and
operation of   these  systems,  and  they may  be  markedly different  from the
properties  of the  material  from  which they  originated.   Fine  dusts, for
example, tend to pack more readily than coarser materials; moreover, conden-
sation formed  in  the filter device may cause solid material to agglomerate.
Both of these factors can make solids disposal difficult.
     Various design features can help prevent the clogging of solids collec-
tion hoppers.   The. hopper  should  be designed  with  a  steep  valley  angle;
angles  of 55  to  70 degrees  are  recommended.    Hoppers  should also include
large discharge openings,  smooth coatings (i.e., epoxy or Teflon ) on inside
surfaces, and minimal ledges or other obstructions on sidewalls.   The top of
the hopper  sidewall  should drop  vertically  and  begin  the  slope to the dis-
charge point at least one bag diameter below the bottom of the bags to allow
proper  dust discharge.   At  least  0.3  m  of  clearance should  be provided
between  hopper walls  and  any internal partitions to  allow easy discharge.
     Heaters and insulation can be installed in hoppers to prevent condensa-
tion and caking  of  collected material.   Jets  of  hot  air can be used, to
fluidize material in the hopper and keep it free flowing.
     Solids  are  generally removed  from  the hopper by means  of  a discharge
valve,   which   removes  ash  from  the  hopper while  preserving the  pressure
differential between  the  dust conveyance system and the fabric  filter sys-
tem.                                                                     ;

     The solids are transported to a collection point by means of screw con-
veyors,  pneumatic  (either vacuum  or  positive systems)  systems,  and  wet
sluicing systems.   Screw  conveyors,  which are common on small systems, work
well in a variety of applications but are sometimes cumbersome.
     Pneumatic systems  are not  limited to straight-line  runs  as  are screw
conveyors, and,  therefore,  are more flexible.  Particularly abrasive solids
must  be accounted  for in  the design  of pneumatic systems  by appropriate
materials  of  construction and   in  some  applications  by  installation  of
replaceable  wear plates  at turns.   Sluice  systems  are  normally  found  in
coal-fired  boiler  applications   for transport  of the  boiler  bottom ash.   Instrumentation.    Reliable  operation of  a fabric  filter  is
favored by the use of the following  instrumentation:
     1.   Thermocouples or other temperature-measuring  instruments located
          at the device inlet.
     2.   Inlet/outlet differential  static pressure gages.
     3.   A single-pass transmissometer (opacity meter).
     4.   Compressed air pressure gage.
     5.   Fan motor ammeter.
     In  lieu  of differential  pressure gages,  it is sometimes  simpler  to
install  static pressure  taps  where  appropriate and  use  a portable meter to
obtain  readings.   This approach  reduces  problems  of meter moisture damage,
meter corrosion, and plugging of  lines.  Where permanent differential  static
pressure  gages are used,  the  static  pressure lines  should  be as short as
possible  and  free of 90-degree elbows.   Copper tubing has been found to be
less  susceptible  to  deterioration  than  the polypropylene  lines  commonly
     Recording temperature meters are especially  useful in identifying high
or  low-temperature excursions,  which  rapidly  destroy fabrics.  As  a less
expensive alternative, high-temperature indicators composed of  colored fiber
or  temperature-sensitive plugs may be  used.
     The  single  pass transmissometer may  not  provide an accurate measurement
of  effluent  opacity;  however,  it.  is  useful  in  identifying  problems.   A
significant  leak  is  detected  in  a  specific compartment  by a  drop  in the
opacity when  that  compartment  is  off-line  for maintenance.20

     The instrument  readouts  are best mounted on a  master control  panel as
close  as  possible to  other process  monitoring  displays.   The  readings of
thermocouples, pressure differential  gages,  and transmissometers can all be
electronically recorded for permanent records.   Fire and Explosion Protection.    Provision   should  be  incor-
porated into  the  fabric filter design to protect personnel and equipment in
the event  of  an  explosion.   Such events can  occur  even in units supposedly
collecting inert particulate matter.24'25
     One common  means to minimize  damage  is to install explosion  vents to
release generated  gases at  the onset of an  explosion.   The two basic types
of  vents  are "the  diaphragms  and the free-hanging   door.   Information con-
cerning the sizing and location of explosion vents is presented in references
37 and 38.   Other  Factors.   Several potential  operating problems  can be
minimized by means of proper design.
     In the case  of  pulse jet filters, water and  oil  in the compressed air
lines can be deposited on the interior bag surfaces.   Resultant blinding can
be minimized  by  using driers  on the compressers and drains on the bottom of
the supply manifold.   The blow tubes should be firmly secured at the far end
so that the  shock does not shear the retaining  pin  and allow the blow tube
to wander.
     Abrasion at  the bottom of bags  in reverse-air  and shaker-type collec-
tors can be  reduced  by design of a good  thimble arrangement.  The thimbles
should  be  at  least  one  bag  diameter  long  to  prevent abrasion caused by
particulate "turning the  corner" at the cell plate  and being thrown to the
outside by inertia!  force.23'24   The thimbles act as flow straighteners and
protect the bottom of the bag from excessive abrasion.   A properly designed
unit is  shown in  Figure  4.4-20.   Note the  rounded  edge on  the  top  of the
thimble to reduce cutting of the fabric even if tension is not optimum.  For
units  in which the  base snaps into the  tube sheet,  a  thimble  can  be added
which  extends  downward to  provide  the  same type of abrasion protection as
that shown in the illustration.
     A bypass  may be  advisable, especially,  when  process  startup  or upset
conditions could  generate sticky particulate or result in gas temperatures

                                             Tube  Sheet
       Figure 4.4-20.   Cross section of a thimble protecting bottom of
              bag (reprinted .by permission:   Mr.  E.  W.  Stanly).
below the acid  vapor  or water dew points.   These could also be used in con-
junction with a spark sensor to reduce risk of fire.
     Inlet and outlet dampers should be provided in compartmented systems to
allow on-line maintenance.   The dampers must be designed to provide positive
sealing so as to protect maintenance personnel from toxic gases.
     The  reverse-air  fan  provides cleaning  of the bags by reversing  the
system  gas  flow.  The  fan must  be  designed to  deliver the necessary  gas
volume at a pressure drop greater than the resistance of the filter.
     Welds around  tube sheets,  thimbles  and hoppers  should  be  continuous.
Tack welding  leaves gaps  through which a relatively  large quantity  of  gas
can pass  untreated.   The  crevices created by task welds also provide sites
for corrosion.
     Access doors should be large enough that maintenance personnel can con-
veniently enter while wearing  safety equipment such  as self-contained  re-
breathers.   These  should  be  secured by several  firm,  yet  easy-to-remove

latches.   The  use of  a large  number of bolts  discourages  routine access,
which  is  necessary  at  most installations.   A chain should  be available
adjacent  to  the door  to secure the door during  periods  when personnel  are
inside.   The chain  provides additional security beyond that of the lock-out
     Additional  information  regarding fabric filter design is references 1,
2,  35, 40,  41, and 42.   Models are  available to predict  the  operations
characteristics  of  reverse-air  and  shaker-type  fabric  filters.12'13'15
4.4.4  Operation and Maintenance of Fabric Filters
     The  long-term  satisfactory  performance  of  fabric filters  is  at least
partially  dependent  on proper operating procedures and a preventive mainte-
nance program.   Startup/shutdown.  A  fabric filter  is  especially vulnerable
to corrosive vapors  and sticky particulate during startup and shutdown.   In
some cases,  it may  be advisable  to bypass  the collector until the effluent
gas stream temperature is above the acid dew point temperature.39
     Prior  to  startup,  a  precoat of  the material to  be collected  can  be
placed  on the  bags  by  the  injection of suitable  material   into the inle^
duct.   This  precoat  is valuable for new  fabrics  in  that  it aids  in  the
conditioning of  the  fabric.   Exposure of the new fabric without the precoat
could  lead to  deposition of fine particles within  the  fabric itself or the
collection  of   hard-to-remove  sticky  material  on the  fabric surface.20*23
Typically, a material  similar to that to be collected is used as a precoat;
however,  prudence should be applied because of potential problems.    Use of
materials  such  as  lime,  for  instance,  has  resulted in  blinding  problems
because of the hygroscopic nature of  lime which  lead  to the formation of a
lime mud on the  fabric surface.39
     4,4.4.2   Fabric tension.  The  adjustment of bag  tension is important
in ensuring  adequate  bag life and minimum particulate emissions.33*39'40'43
The bag should be tight enough to avoid excessive fiber-to-fiber and bag-to-
bag abrasion,  but not so tight as to exceed the tensile strength of the bag
during cleaning.  It may be necessary to check bag tension soon after startup.39   Cleaning system.    Operation of  the cleaning system should be
evaluated regularly.  Cleaning intensity and frequency can have a direct and

substantial impact on both bag life and emissions.   For example, Dennis
and Wilder44 found in one installation that penetration from a shaker-type
unit was related  to  the shaker amplitude.   Ladd, et a!.33 reported that
adjustment of shaker frequency contributed to a reduced rate of bag failure.
The gas flow rate and static pressure available in a reverse air fan can
also be an important operating variable.27
     4,4.4.4   Solids removal.  Accumulation of solids within the hopper can
lead to major operational problems.  A partially filled hopper can lead to
particulate matter reentrainment and abrasion of the lower portions of the
bags.25  The material  in the hopper can gradually cool and bridge over.
Ultimately, bridging could lead to a restriction of gas flow to the bags and
to a buildup of material into the bags.  Collected material with a substantial
combustible content can be prone to fires.
     On a frequent basis, operators should confirm positively that solids are
being discharged.  For units with rotary valves, the use of a long (0.5 m),
brightly colored rod attached to the end of the shaft has proved useful in
determining from a distance that a rotary valve has stopped.  Motion sensors
can be used on rotary valves and screw conveyors.   Fabric and Component Repair.  When bag failure is the result
of  localized abrasion  or mechanical  damage, small pinholes or tears are
usually sealed by adhesive and sewn with thread.  The adhesive and thread
should be compatible with the original fabric in the properties of shrinkage,
temperature tolerance, and chemical resistance.  Successful repair depends
on the strength and condition of the bag.  Patching may not be successful
on bags that have been operated for a  long period and have undergone chemical
and thermal degradation.  Bag repair must be considered relative to the cost
of bag replacement.  Repair becomes economically attractive when bags are
extremely  large  (>30-cm  diameter)  or  when the  fabric  is  expensive.
     The blinding  of bags because of  process upset, operation below dew
point, or moisture inleakage can increase filter pressure drop.  If the bags
are new and have not been subjected to chemical or thermal degradation, it
is possible to reduce the fabric resistance by  laundering.  Although it is
not available for all fabric types, laundering  can sometimes make it possible

to  put off  bag  replacement  until  a later  date.   Care  must be  taken  to
prevent bag  shrinkage  or chemical attack during  cleaning.   Even  though the
expected bag life may even be  reduced,  the overall cost may  be  lower than
that  incurred  by total  bag  replacement  or  operation at higher  static
     Reuse  of bags  and  cages  after a  fire  in  a  baghouse is usually not
possible and  almost  never advisable.  High temperatures can warp the cages
to the extent that bag-to-bag abrasion results.  Cages may be reused only if
they are not corroded or bent.   Each cage must be carefully  inspected before
Installation.   Maintenance Inspection.  Regular  inspection  of  the inside of
bags  is  necessary  to  confirm  that  the  system  is   in   compliance  with
regulatory  requirements  and that there are no  developing  problems.   Safety
procedures must  be  strictly followed.  Diagnosis of prevailing operation is
done primarily  by observation  of "clean" side deposits resulting from pene-
tration of dust.  If the penetration is local  to specific bags  or seals, the
pattern  created  by  the  dust  on  the tube sheet  may indicate  the  point of
penetration.   Figure  4.4-21  shows  the  characteristic  pattern   of  a  low
velocity dust penetration at  the snap-ring attachment in a shaker baghouse.
The small depressions that look like craters are the penetration points.  In
the early stages of  penetration the pattern  can  also  be highlighted with a
fluorescent  dye  and an  ultraviolet  light source.   The use of  fluorescent
dyes  is  not  practical  where  total  failure  of a  bag  has   occurred  or the
problem has existed for an extended period.
     Moderate clean-side deposits may be caused by a single small  pinhole in
a  bag.   Again,   the  pattern of  dust generated by  penetrating gas  flow can
indicate  the location  of  holes.  Figure  4.4-22  shows the  dust  pattern
indicated by the clean  area  on  the tube  sheet  as  a  result  of the gas
impingement.  Impaction of the particles on bags on the opposite side of the
collector is  indicated  by  the  discoloration.  The  abrasion caused  by this
high-velocity jet results in cascading bag failures in the collector.
     Abrasion can occur if an adequately designed precleaner or baffle plate
is  not  used to  remove large,  abrasive particles.   Figure  4.4-23 shows the
abrasive damage caused by large, sharp particles impinging on the bag surface

Figure 4.4-21.   Dust penetration around  snap-rin  attachment.
                (Courtesy of PEDCo Environmental, Inc.)

                                       o>   •
                                      i—  O
                                       o  c
                                      o  c
                                      •«J  0)
                                      c  o
                                      (U  J-
                                      O -i-
                                      ••-3  C
                                      "O UJ
                                      •»J O
                                      d) O


                                   •*•>   •
                                   •M  OJ
                                    «  E

                                    O)  O
                                    TO S-
                                    « -r-

                                    E  >
                                    «S  C
                                    c  o
                                    o o
                                   «i- Q
                                    (/) UJ
                                    (O O-

                                   »Q <*-
near the bag  cuff.   In this design the  installation  of thimble extensions,
blast plate,  or  precleaner would reduce the abrasive damage  and extend bag
life.  Most  abrasion problems  occur  near the  bottom of  the  bags directly
above the thimbles.   Hopper overflow  can also be detected at the bottom of
bags, especially  in  bags  located in corners or along the outer walls of the
compartment.   C1earn*ng System Operation.   The  cleaning   system  (shaker,
reverse air,  pulse jet)  should be checked for proper  operation.   When the
cleaning  system malfunctions,  an  accompanying  increase  in   filter  static
pressure drop is usually  noted.   The motor  and cam arrangement should be
inspected for wear and linkage failure.   The dampers and fans  of reverse-air
systems should be  checked for correct operation.   The pulsing of individual
solenoid/diaphragm  systems  on  pulse-jet  cleaning  mechanisms  should  be
checked.   The compressed air pressure  should be checked, and proper operation
of compressed air dryers should be verified.
     During  operation,  the  stack opacity should  be  checked  regularly.   A
spike in visible  emissions can often  be related to the pulse cleaning of a
specific row  of  bags,  thus aiding in  the eventual  identification of the bag
or  bags  causing  the problem.   On compartment  type  reverse-air  or  shaker
units, a reduction  in  opacity during  cleaning of a specific  compartment is
indicative of a problem in that compartment.   Preventi veMai ntenance.   A  preventive  maintenance  program
should be aimed at reducing bag failure.   The program should include routine
servicing of  mechanical  equipment including gears, bearings,  and pneumatic
cylinders  and  also   should  include  a  complete  external   and  internal
inspection  of  the system at  frequent  regularly  scheduled  predetermined
     The operator  should maintain records indicating system  pressure drop,
temperature,   date  of  bag replacement  and location  of the  bags,  and changes
in  process  operation.   These  data may then  be  used  to  diagnose failure
mechanisms  and  provide   direction  in  preventing  recurrence of failures.
     An  example  of  the  types  of  records  useful  in  diagnosing recurring
problems is shown  in  Figure 4.4-24.  The top  figure  reflects a random type



       Mod. No..
            im.tr r
Figure 4.4-24.
Bag failure location records. (Courtesy of

Richard P. Bundy, Standard Havens, Inc.)

pattern which could be due to a large variety of problems.   Symptoms  of a
baffle problem are shown in the lower figure.  In addition  to these figures,
a simple elevation sketch of each bag removed should be prepared showing the
location and type of damage.  Another set of diagnostic records which has
proven useful is a record of the frequency of bag failures.   This can be
used to identify a condition which has arisen recently or when a set  of bags
is reaching the end of the useable life.


1.    Billings, C.  and J. Wilder, (GCA Corporation, Boston).  Handbook of
     Fabric Filter Technology, Vol. I.  Prepared for U.S. Environmental
     Protection Agency, Research Triangle Park, North Carolina.
     Publication No. APTD-0690.

2.    Theodore, L.  and A. J. Buonicore.  Industrial Air Pollution Control
     Equipment for Particulates.  CRC Press.  Cleveland.  1976.

3.    Perkins, H. C.  In:  Air Pollution, A. Stern (ed.).  New York,
     McGraw-Hill.   1974.

4.    Paretsky, L.L., Theodore R. Pfeffer, and A.J. Squires.  J. Air
     Pollution Control Association.  21:4.  April 1971.

5.    Anderson, D.M. and L. Silverman.  Harvard Air Cleaning Laboratory.
     AEC Report No. NYO-4615.   1958.

6,    Donavan, R.P., J. H, Turner, and J. H. Abbott.  Passive Electrostatic
     Effects in Fabric Filtration.  Second Symposium on  the Transfer and
     Utilization of Particulate Control Technology.  In:  Vol. 1.  Control
     of Emissions from Coal Fired Boilers.  F. P. Venditti, J. A. Armstrong,
     and M. Durham  (ed.).  U.S. Environmental Protection Agency Publication
     No. EPA-600/9-80-039a.  September .1980.  pp. 476-493.

7.    Donavan, R.P., R.L. Ogan,  and J.H. Turner.  The Influence of Electro-
     statically Induced Cage Voltage Upon Bag Collection Efficiency
     During the Pulse-jet Fabric Filtration of Room Temperature Flyash.
     In:  Proceedingsof the Third Symposium on Fabric Filters for
     Particle Collection.  U.S. Environmental Protection Agency, Research
     Triangle Park, N.C.  Publication No. EPA-600/7-78-087.  June 1978.

8.    Frederick, E.  J. of the Air  Pollution Control Association,
     24: 1164-1168.  December  1974.

9.    Frederick, E.  Chemical Engineering.  68: 107,  June  1971.

10.  Turner, J.  Extending Fabric  Filter Capabilities.   J. of  the Air
     Pollution Control Association.   24: 1182-1187.  December  1974.

11.  Enson, D. S.,  R. G. Hooper, and  R. W. Schech.  Determination of the
     Fractional Efficiency, Opacity Characteristics, Engineering and
     Economic Aspects of a Fabric  Filter Operating as a  Utility Boiler.
     Electric Power Research Institute.  Report EPRI FP-297.   November  1976.

12.  Dennis,  R. and H.  Klemm.   Modeling Coal Fly Ash Filtration with Glass
     Fabrics.  Third Symposium  on  Fabric Filters for Particulate Collection,
     N. Surprenant.  U.S.  Environmental Protection Agency, Research Triangle
     Park, N.C.  Publication No. EPA-600/7-78-087.  June 1978.  pp.  13-40.

 13. Dennis, R. and H. A. Klemm.  Fabric Filter Model  Format  Change,  Vol.
     I, Detailed Technical Report.  GCA Report to  U.S.  Environmental  Protec-
     tion Agency, Contract No. 68-02-2607, Task No. 8.  GCA Report  GCA-TR-
     78~51-G(1), January 1969.

14.  Hall, R. R.,  R. Dennis, and N. F. Surprenant.   Fibers,  Fabrics,  Face
     Velocity and Filtration.  A Specialty Conference on  the  User and Fabric
     Filtration Equipment III.  E.  R. Fredrick, (ed.).  APCA  Specialty
     Conference Proceedings.  1978. pp. 156-169.

15.  Dennis, R. H. A. Klemm.  Verification of Projected Filter  System Design
     and Operation,  pp. 143-160.   Symposium on the Transfer  and Utilization
     of Particulate Control Technology, Vol. 2, Fabric Filters  and  Current
     Trends in Control Equipment.   F. P. Venditti, J.  A.  Armstrong, and M.
     Durham, (ed.).  U.S. Environmental Protection Agency.  Publication
     No. EPA-600/7-79-044b.  February 1979.

16.  Leith, D., M. W. First, M. Ellenbecker, and D. D.  Gibson.  Performance
     of a Pulse-jet Filter at High  Filtration Velocities.  In:   Symposium on
     the Transfer and Utilization of Particulate Control  Technology:
     Vol. 2, Fabric Filters and Current Trends in Control  Equipment.
     F. P. Venditti, J. A. Armstrong, and M. Durham,  (ed.,).   U.S. Environmental
     Protection Agency.  Publication No. EPA-600/7-79-044b.   February 1979.
     pp. 11-26.

17.  Leith, D. and M. J. Ellenbecker.  Theory for Penetration in a  Pulse-
     jet Cleaned Fabric Filter.  Paper 80-30.1.  Presented at the 73rd Annual
     Meeting of the Air Pollution Control Association.  Montreal, Quebec.
     June 22-27, 1980.

18.  Dennis, R., J. W. Wilder, and  D. L. Harman.  The  Mechanics of  Pulse-jet
     Filtration, Paper 80-30.6.  Presented at the 73rd Annual Meeting of the
     Air Pollution Control Association.  Montreal, Quebec.  June 22-27, 1980.

19.  Theodore, L.  and J. Reynolds.  Performance Equations  Describing  Effects
     of Bay Farline on Baghouse Outlet Loadings.  Presented at  the  Air Pollution
     Control Association Annual Meeting.  Cincinnati,  Ohio.   June 24-29, 1979.

20.  Perkins, R. P.  The Case for Fabric Filters on Boilers.  Semiannual
     Technical Conference on Air Pollution Equipment.   Philadelphia,  Pennsylvania.
     April 23, 1976.

21.  Vandenhoeck, P.  Cooling Hot Gases Before Baghouse Filtration.   Chemical
     Engineering.   Vol. 79, No. 9:  67-70.  May 1, 1972.

22.  Bergmann, L.   New Fabrics and  Their Potential Application.  J. Air
     Pollution Control Association.  24: 1187-92.  December 1974.

23.   Perkins, R. P.  State-of-the-art of Baghouses for  Industrial Boilers.
     Presented at the Industrial Fuel Conference.  West Lafayette,  Indiana.
     October 5-6, 1972.

24.   Rolschau, D. W.  Matching a Baghouse to a Fossil Fuel  Fired Boiler.
     In:  Symposium on the Transfer and Utilization of  Particulate
     Control Technology:  Vol. 2, Fabric Filters and Current Trends in
     Control Equipment.  F. P. Venditti, J. A. Armstrong, and M. Durham,
     (ed.).  U.S. Environmental Protection Agency, Research Triangle Park,
     N.C.,  Publication No. EPA-600/7-79-044b.  February 1979. pp.  211-217.

25.   Brandt, E. F.  Operation and Maintenance of Fabric Filters on  Coal-
     Fired Boilers.  In:  Proceedings:  Operation and Maintenance
     Procedures for Gas Cleaning Equipment.  E. R. Fredrick (ed.).   APCA
     Specialty Conference Proceedings.  1980. pp. 153-161.

26.   Billings, C. E. and J. E. Wilder.  Major Application of Fabric Filters
     and Associated Problems.  Paper No. 15.  In:  Proceedings of the
     Symposium on Control of Fine-Particulate Emissions from Industrial
     Sources.  San Francisco, California.  January 15-18, 1974. pp.  329-372.

27.   Tracy, G. W.  Operation and Western Precipitation  Fly  Ash Baghouses
     Since Sunbury.  In:  Proceedings:  Operation and Maintenance
     Procedures for Gas Cleaning Equipment.  E. R. Fredrick (ed.).
     APCA Specialty Conference Proceedings.  1980. pp.  162-181.

28.   Mappes, T. E. and R. D. Terns.  An Investigation of Corrosion in Parti-
     culate Control Equipment.  U.S. Environmental Protection Agency,
     Research Triangle Park, N.C. Publication No. EPA-340/1-81-002.
     February 1981.

29.   Prinz, R. T.  Reducing Baghouse Maintenance by Design.  Minerals
     Processing.  11: 8-13.  May 1970.

30.   E.  I. duPont de Nemoursand Company, Bulletin No. E-13544.

31.   Menardi and Company.  Bulletin.  Segundo, California.

32.   Pearson, G. L.  Experience at Coors with Fabric Filters-Firing Pulver-
     ized Western Coal.  In:  Second Symposium on the Transfer and
     Utilization of Particulate Control Technology, Vol. 1, Control of
     Emissions from Coal Fired Boilers, F. P. Venditti, J.  H. Armstrong,  and
     M.  Durham, (ed.).  U.S. Environmental Protection Agency, Research
     Triangle Park, N.C.  Publication No. EPA-600/9-80-039a.  September 1980.
     pp. 359-370.

33.   Ladd, K. L. Jr., R. Chambers, S. Kunka, and D. Harman.  Objectives
     and Status of Fabric Filter Performance Study.  In:  Second Sym-
     posium on the Transfer and Utilization of Particulate  Control  Tech-
     nology.  Vol. 1, Control of Emission "rom Coal Fired Boilers,  F. P.
     Venditti, F. A. Armstrong, and M. Durham (ed.).  U.S.  Environmental
     Protection Agency, Research Triangle Park, N.C.  Publication
     No. EPA-600/9- 0-039a.  September 1980. pp. 317-341.

34.  Mycock, J. C.  A Pilot Plant Study of Various  Filter Media  Applied
     to a Pulverized Coal-fired Boiler In:   Symposium  on the Transfer
     and Utilization of Particulate Control  Technology.  Vol.  2,  Fabric Fil-
     tration Current Trends in Control Equipment.   F.  P. Venditti,  F.  A.
     Armstrong, and M. Durham, (ed.).  U.S.  Environmental Protection
     Agency, Research Triangle Park, N.C.  Publication No.  EPA-600/7-79-044b.
     February 1979.

35.  Simon, H. T.  Baghouses.  In:  Air Pollution Engineering  Manual.
     2nd Edition, J. A. Danielson (ed.).  U.S.  Environmental Protection
     Agency, Research Triangle Park, N.C.  Publication No.  AP-40.
     May 1973.

36.  Wright, R. J.  Customizing Baghouses for Clinker  Coolers,   Rock  Products.
     November 1972. pp. 94-129.

37.  Miller, R. L.  Explosion Pressure Relief.   In:  Proceedings:   The
     User and Fabric Filtration Equipment III.   E.  R.  Fredrick (ed,).
     APCA Specialty Conference Proceedings.  1978.  pp.  98-112.

38.  Reinauer, T. V.  Guidelines in Application of  Explosion Vents.   In:
     Proceedings, The User and Fabric Filtration Equipment  IV.   E.  R.  Fredrick
     (ed.).  APCA Specialty Conference Proceedings.  1978.  pp.  113-119.

39.  Perkins, R. P. and J. F. Imbalzano.  Factors Affecting Bag  Life
     Performance in Coal-Fired Boilers.  In:  Proceedings,  The User and
     Fabric Filtration Equipment III.  E. R. Fredrick   (ed.).  APCA
     Specialty Conference Proceedings.  1978. pp. 120-144.

40.  Szabo, M. F. and R. W. Gerstle.  Operation and Maintenance  of  Partic-
     ulate Control Devices on Coal-Fired Utility Boilers, U.S. Environmental
     Protection Agency, Research Triangle Park,  N.C.   Publication No.
     EPA 600/2-77-129.  July 1977. pp. 3-85.

41.  Strauss, W.  Industrial Gas Cleaning.   Pergamon Press.  New York,  1975.

42.  Smith, G. L.  Engineering and Economic  Considerations  in  Fabric  Fil-
     ters.  J. Air Pollution Control Association.   24:  1154-56.

43.  Bundy, R, P.  Operations and Maintenance of Fabric Filters.  In:
     Proceedings, Operation and Maintenance  Procedures for  Gas Cleaning
     Equipment.  E. R. Fredrick, (ed.).  APCA Specialty Conference
     Proceedings.  1980. pp. 139-152.

44.  Dennis, R. and J. E. Wilder.  Factors in the Collection of  Fine
     Particulate Matter with Fabric Filters.  Paper No. 17.  In:
     Proceedings, Symposium on Control of Fine-Particulate  Emissions  from
     Industrial Sources.  San Francisco, California.   January  15-18,  1974.
     pp. 385-424.

     Wet scrubbers comprise  a  set of control devices with  similar particle
collection mechanisms, primarily:  inertia! impaction and Brownian diffusion.
Accordingly, these scrubber  systems  generally exhibit strong particle-size-
dependent performance.   Among  scrubber types substantial  differences exist
with  regard to their  effectiveness,  the greatest  differences  occurring in
the particle size range of 0.1 to 2 umA.
     The various types of commercially available wet scrubbers are described
in  Section  4.5.1.   Fundamental   operating  principles  are  presented  in
Section 4.5.2.   Parameters of interest and performance limits are discussed.
Emphasis is on the capability for collection of particles smaller than 5 umA
in diameter.
     Considerable  progress  made  in  the understanding of wet scrubber per-
formance since the first edition of this document was published is reflected
in practical design  considerations  in Section 4.5.3.  Operation and mainte-
nance   factors   that  enhance   long-term  performance  are   presented  in
Section 4.5.4.
4.5.1  Types of Particulate Scrubbers
     In  this  discussion major  categories of scrubbers  are grouped  on the
basis of similar  mechanisms.   In Table 4.5-1, major categories of scrubbers
are  listed  in order of  increasing  performance  capabilities  and  energy
     Scrubber  liquids are used for particle  collection  in  several distinct
ways.   The  most   common  method  is  to  generate droplets,  which are  then
intimately mixed with the gas stream.  Particles are also collected on water
layers  or  sheets  surrounding of packing material by directing the particle-
laden gas  stream  through  an intricate path  around the  individual  packing
elements.   A third  method  is  to  pass high-velocity gas through  a vapor to
generate "jets" of liquid to collect particles.   This is the least common of
the three liquid characteristics.   Preformed Spray Scrubbers.  Preformed spray  scrubbers require
the  least  energy  of the various  scrubbers,  and they consequently allow the

                 TABLE 4.5-1.  MAJOR TYPES OF WET SCRUBBERS*
Particle capture
 Liquid collection
Types of scrubbers
aided scrubbers
Venturi and
(gas atomized
Sheets, droplets
 (moving bed
 Droplets, jets,
   and sheets
Droplets and
Spray towers
Cyclonic spray
Vane-type cyclonic

Standard packed-bed
Fiber-bed scrubbers
Moving-bed scrubbers
Cross-flow scrubbers
Grid-packed scrub-

Horizontal impinge-
 ment-plate (baffle)

Wet fans

Standard venturi
Van" able-throat
 venturi scrubbers:
 flooded disc, plumb
 bob, movable blade,
 radial flow, varia-
 ble rod
Orifice scrubbers
 List not intended to be all inclusive.

highest  particulate  penetration,  especially  of small-diameter  particles.

Most  preformed  spray  scrubbers  are  highly efficient  only for  particles

larger than 5 pmA in diameter.1-3

     4, Spray tower.  A spray tower is the simplest type of scrubber,
consisting of a chamber containing an array of spray nozzles (Figure 4.5-1).
                                           GAS OUTLET


                                                LIQUOR INLETS
                                               GAS INLET
                                            LIQUOR OUTLET
                     Figure 4.5-1.  Spray tower scrubber

Particulate-laden  gases  pass  vertically  up  through the  tower  while  the
liquid  droplets  fall   by gravity  down  through the gas  flow.   Particles
collide with the droplets, are collected into the liquor, and carried out of
the  scrubber.   Collection is  limited by the  terminal  settling velocity of
the droplets.  The spray  tower scrubber has low particle removal capability,
but  it is often  useful for treating effluent gas  streams  having high mass
loadings of large diameter particulate matter.4

------- Cyclonic spray tower.  A  cyclonic spray tower  is  similar to
the spray tower scrubber except that the gas stream is given a spiral motion.
In  one  typical  configuration  particulate-laden gases  enter the  scrubber
vessel tangentially at the bottom and pass upward in a spiral motion around a
centrally located array of spray nozzles.  Droplet migration is crosscurrent
to  the  gas  flow.   Cyclonic  spray towers generally  operate with  a static
pressure drop  of 1 to 2 kPa.3   Penetration  of  particles  less than 2 umA in
diameter is typically quite high.5  Vane type Cyclonic Spray Tower.   A vane-type cyclonic scrubber
utilizes a system of vanes rather than a tangential inlet to impart cyclonic
motion (Figure 4.5-2).  Multiple-tube Cyclones.   Another type  of cyclonic scrubber
consists of  multiple miniature  tubes,  each with a  separate liquid supply.
In  this  design  the  gases  flow in  a downward  pattern  in contrast  to  the
upward flow of other cyclones.   Packed BedScrubbers.   In  the   typical  packed-bed  scrubber
liquid introduced  near the  top trickles down  through the  packed bed.   The
liquid flow  spreads  over  the packing into a film with a large surface area
(Figure  4.5-3).   The  liquid can  be  introduced concurrent  or crosscurrent
with  the gas  flow.   Packing materials  include raschig  rings, pall  rings,
berl  saddles,  tellerettes,  intalox  saddles, and materials  such  as crushed
rock.3  Packed beds are also constructed with metal grids, rods,  or fiberous
pads.  These scrubbers  are  often used for gas transfer or gas cooling, both
of  which are  facilitated by  the large liquid  surface area  provided on  the
     Plugging  of  a bed can occur  if the gas  to be  treated is  too heavily
laden with solid particles.3'4   A general rule  for  many  applications is to
limit the  use  of packed beds to service in which particulate concentrations
are less than 0.45 g/m3.  Moving-bed scrubbers that have  less propensity for
plugging (Figure 4.5.4) are packed with  low-density  plastic spheres, which
are free to move within the packing retainers.
     Packed-bed scrubbers are  reported to have  low penetration for particle
sizes  down to  3  umA and  can  sometimes  remove a significant  fraction of
particulate in the range of 1 to 2 umA.  The standard countercurrent

                                 '  4.5-4 "

                          CLEAN GAS
                                                      CYCLONIC LIQUID
                                                      SPINNING VANES
                                                      LIQUOR INLET
                                                      LIQUOR OUTLET
      Figure 4.5-2.   Vane type scrubber (courtesy of the
                    Ducon Company, Inc.)*

                                                             LIQUOR INLET
                                                             LIQUOR OUTLET
             Figure 4.5-3.  Packed tower scrubber (courtesy of Air
                          Pollution Industries, Inc.)

                     CLEAN AIR OUTLET
                                            MIST  ELIMINATOR
                                                 MOBILE PACKING
                                          HOT  GAS  INLET
   Figure 4.5-4.

Moving-bed scrubber (courtesy of UOP - Air
  Correction Division).

arrangement  requires  the greatest  liquid flow and can  best handle heavier
loadings.  Crosscurrent  packed-bed  scrubbers require  much less liquid flow,
usually  operate at  lower  static  pressure  drops,  and  rarely suffer  from
plugging.   Concurrent packed-bed  scrubbers  are  reportedly  more  efficient
than  other  packed-bed  scrubbers  for  the  smaller particulates,   but  they
typically operate at higher pressure drops.  TrayType  Scrubbers.  A  tray-type scrubber typically consists
of a vertical tower with one or more perforated plates mounted inside trans-
versely  to the  shell.   In such a  scrubber the liquid flows from  top  to
bottom,  and  the gas  flows  from bottom to top.  Gases  in  the scrubber mi*
with the liquid passing through the openings in the plates.
     The perforated plates  of a tray-type scrubber are  often equipped with
impingement  baffles  or  bubble caps over the perforations  (Figure 4.5-5).
The gas  passing  upqard through a perforation  is  forced!  to  turn 180 degrees
into a layer of liquid.  The gas bubbles through the liquid, and particulate
is collected  in the  liquid sheet.   The  impingement  baffles  are  below the
liquid  level  on the  perforated plates  and  are,  therefore,  continuously
washed clean of collected particles.  Penetration through a typical impinge-
ment plate is  low  for particles larger than 1 umA,3 but penetration of sub-
micrometer particulate  is  higher  than with some  higher-energy  scrubbers.
Pressure drop through  a  typical baffle plate is roughly 0.4 kPa per stage.3
Addition of plates increases  the scrubber pressure drop, but does not pro-
portionally decrease the penetration of submicrometer particulate.6
     One additional variation of the tray scrubber is  the horizontal baffle-
type scrubber.   In this  type of scrubber the direction of gas flow is hori-
zontal and the baffle section is mounted vertically .in the scrubber.   The
scrubbing liquid is introduced concurrently with the gas.  Mechanically Aided Scrubbers.  The mechanically aided scrubbers
utilize a mechanical  rotor or fan to shear the scrubbing liquid into dispersed
droplets.  These scrubbers use a specially designed stator and rotor arrangement
to produce very finely divided liquid droplets that are effective in capture
of fine  particulate.   The low penetration of  fine  particulate,  however,  is
achieved at  a  high  energy cost.1-3  Because both wet-fan and disintegrator-


               Figure 4.5-5.  Tray scrubber (courtesy of the Koch
                             Engineering Company).

type  mechanically aided  scrubbers  are  subject  to particulate  buildup  or
erosion at the  rotor blades, they are often preceded by precleaning devices
for removing  coarse  particulate.1'4   Mechanically aided scrubbers generally
do not perform well in air containing more than 1 g/m3  of particulate,1
     4,5.1.5  Venturi and Orifice Scrubbers.  Venturi  and orifice scrubbers
are perhaps  the most  common particulate  removal  devices, in  part because
they  allow  lower penetration  of small particles  than most  other types  of
scrubbers.   These scrubbers accomplish  superior particulate  collection  by
generating  small  liquid  droplets in  the  turbulent zone  in  a  manner  that
creates a high initial relative velocity between the droplets  and the parti-
culate.   Inertia!  impaction capture  of particulate by  the scrubbing liquid
is more efficient  in these highly turbulent processes,  but a price is paid
in energy consumption to achieve the low penetration.
   ,.  .	Ve.ntuH. scrubbers.   The  sjmple  venturi  scrubber,  often
called a gas atomizing spray scrubber, consists of a series of sprays upstream
from a converging and diverging "throat" section (Figure 4.5-6).  As the gas
approaches the  venturi throat,  the  velocity and  turbulence  increase.   The
high  gas  turbulence  atomizes  the liquid  into  small droplets and increases
interaction  between  the  droplets and the particulate.    Pressure  drops  in
venturi scrubbers can range from less than 1 to 40 kPa.  Variable-throat venturi scrubbers.   Pressure drop and venturi
performance  are  partially  dependent on gas  velocity  through  the venturi.
Several variations of  the standard venturi scrubber have  been  developed  to
allow  the  venturi throat dimensions to be changed as  the  rate  of gas  flow
changes.  Among these  scrubbers  are  the plumb-bob venturi, the flooded-disc
venturi, the moveable-blade venturi,  the radial-flow venturi,  and the variable-
rod venturi.   Several of these venturi throats are illustrated in Figure 4.5-7.  Orifice scrubber.  In an orifice scrubber, sometimes referred
to as  an  entrainment or self-induced spray scrubber, the gas stream passes
over  a pool  of  scrubbing liquid  at high  velocity just  before entering  an
orifice.  The high  velocity  of the  gas  induces  ("entrains") a  spray  of
scrubbing  liquid  droplets,  which   interact  with  the   particulate in  and
immediately  after the  orifice.   Orifice  scrubbers have  moderate pressure
drops  (0.8  to 4.0 kPa)  and low penetration  of particulate  2  to  3 ymA in
diameter and larger.

4.5.2  Operating Principles of Particulate Scrubbers
     The  fundamental  principles that  govern  particle  penetration and total
static pressure  drop  are examined both in a general manner and with respect
to classes of wet scrubbers.   Penetration.   Particulate  matter  collection  in wet scrubbers
is  highly size-dependent because of the  fundamental  characteristics of the
inertia!  impaction  and  diffusion  processes.   Figure  4.5-8 illustrates the
theoretical  single-droplet collection  efficiency resulting  from these two
phenomenon as  calculated by Crawford7 for an example case.  This particular
example  illustrates  a  predicted  minimum collection  efficiency  reached at
Q.l-umA  particles;  the  size at which both  mechanisms  become ineffective.
This  general   curve applies  to most  scrubbers; however,  the location and
magnitude of the efficiency minimum depends on the  specific unit.
                                                      GAS OUTLET
                                                  LIQUOR OUTLET
                        Figure  4.5-6.   Venturi scrubber.


    >x^f SPRAYS
                                                              DIRTY  GAS IN
a.  Movable-blade venturi
                                                   PLUMB BOB
                                                  TO LIQUID ENTRAINMENT
                                                    ..•  .SEPARATOR

                                               b.   Plumb-bob venturi
            DIRTY GAS IN
                                                     DIRTY GAS IN

     c.  Radial-flow venturi

                                                                             --*" TO
                                           d.   Flooded-d1sc  venturi
      Figure 4.5-7.  Throat  sections  of variable  throat  venturi  scrubbers
             (courtesy  of Industrial Gas Cleaning  Institute, Inc.).


                        PARTICLE DIAMETER
Figure 4,5-8.  Theoretical  single-drop collection efficiency
        due to diffusion  and impaction (reprinted by
      permission:   Crawford, M.  Air Pollution  Control
          Theory,  McGraw  Hill Co., New York,  1976).

     Analysis of  the  particle collection capability of wet scrubbers can be
based  on (1)  the  fundamental  particle  collection  mechanisms and  (2)  the
empirical contact power approach.  The latter method is based on the premise
that penetration  is proportional  to the power expended in the scrubber.8-13
     This premise is logical  because high  energy consumption implies high
relative  gas-water velocities,  high  water utilization,  and fine  droplet
formation, all  of which favor impaction, the dominant collection mechanism.
Limitations of the contact power analysis can be attributed to the difficulty
of handling nonideal operating conditions such as pooir gas-liquid distribution
and particle  shattering during  high-energy scrubbing.  Also, this  type of
analysis is not amenable to situations in which particle collection mechanisms
other than inertia! impaction are important.
     Penetration  analyses  based  on  the  fundamental  particle  collection
mechanisms involve  the identification  of the dominant  physical  phenomenon
leading to particle capture.   Following is a partial list of the collection
mechanisms.                                                               t
          Collection medium             Capture phenomenon
          Droplets                           Inertial impaction
                                             Brownian diffusion
          Liquid sheets (layers)             Inertial compaction
                                             Brownian diffusion
                                             Electrostatic attraction
          Liquid sheets                      Inertial impaction
          Bubbles                            Inertial impaction
                                             Brownian diffusion
                                             Electrostatic attraction
For each  control  device,  penetration relationships are based on anticipated
particle  collection mechanisms.   The accuracy  of the  resulting  equations
depends on  the proper  assignment  of the mechanisms  and on  the  accuracy of
the mechanism expressions.   Penetration  expressions  presented in the Scrub-
ber Handbook3 for selected wet scrubbers are provided in the following para-


-------   Preformed-spray scrubbers.   The  most  effective  particle
collection  mechanism   is   inertia!   impaction  to  liquid  droplets.   The
penetration  calculations  depend  on  droplet  size and  flow characteristics
(i.e.,  countercurrent,  cocurrent,   and  crossflow).    For  countercurrent
conditions, the following equation is applicable:
                                                  (Equation 4.5-1)
P. = e -
"0.75 vt
. rd(vt -

droplet terminal settling velocity, cm/s.
impaction parameter, dimension! ess.
scrubber height, cm.
droplet radius, cm.
gas superficial velocity, cm/s.
     v.  =
     r\,  -
     Z   =
     r .  =
     VG  =
     2±  -  liquid-to-gas ratio, dimensionless.
The parameter n. is calculated according to Equation 4,5-2.
               (K.. + 0.7)
                                                       (Eq. 4.5-2)
 impaction parameter for particles  having aerodynamic diameters
of i.
The  variables that  control  the penetration  rates  from  a preformed-spray
scrubber  include scrubbing zone  height,  superficial  gas velocity, particle
aerodynamic  diameter,  liquid-to-gas  ratio,  and spray  droplet size.   These
variables are equally important  in cocurrent and crosscurrent preformed spray
scrubbers,  which are discussed  further  in Calvert et al.3  These scrubber
types  share  the  characteristic penetration  curve  presented  earlier  for
impingement-plate scrubbers.   Packed-bed scrubbers.   The  packed-bed scrubber  primarily
utilizes  inertia!  impaction  of  particles on water "sheets"  on the packing
material.    Turbulent  diffusion  may  contribute  some  additional  particle
capture  in   the  <0.2-umA  size range.   Equation 4.5-3  is  based strictly on

               L 2(j+r) (e-Hd) J
     P1 « e -
                                                      (Eq.  4.5-3)

     P. = penetration value  for a  specific particle  size,  dimensionless.
      j = "channel" width  factor, dimensionless.
      e = bed porosity,  dimensionless.
     H . = liquid holdup  in bed,  dimensionless.
      Z = height of packed section, m.
     d  = packing element  diameter, m.
     K. = inertia!  impaction  parameter, dimensionless.                  '.

The parameter K. is common to most penetration equations applicable to con-
trol devices based on impaction.  The higher the value of K.,  the higher the
collection efficiency for particles with an aerodynamic diameter of i.   The
impaction parameter is a  strong function of  particle  size, as  indicated in
Equation 4.5-4.
                                                      (Eq.  4.5-4)
     v  =  gas  velocity through the bed,  calculated as  the  volumetric  flow
      S3                     .
          rate divided by  the  scrubber cross-sectional  area,  cm/s.
     u  ss gas viscosity at actual temperature, g/s-m.
     d. = aerodynamic particle diameter i, urn (g/cm3) '  .
Combining Equations 4.5-3  and  4.5-4 provides a means of calculating penetra-
tion for specific particle sizes, which can then be added to determine total
penetration.   Typical values for the parameter, j, are 0.160 to 0.190.3   The
smaller the packing  material,  the lower j should be.   Bed porosity, e,  also
depends on the packing size and normally varies from 0.60 to 0.95.3  Calvert
et al.  list bed  porosities  for use in selecting types and sizes of packing.
Liquid holdup is  normally  assumed to be negligible.3
     Figure 4.5-9 presents  a  typical penetration curve  for  a  packed tower.
This theoretical  curve  is  based on the above mathematical relationships and
use of  the parameter values specified.  It is apparent that there is a very

300 eM/sec
    0            1           23           4           5
                            AERODYNAMIC DIAMETER, ymA

     Figure 4.5-9.  Theoretical penetration curves for various-sized
                   packed-bed scrubbers.

  strong  interdependence between particle size and penetration.  Actual pene-
  tration values  for  a  specific facility (at a specific time) might vary by
  more  than a  factor  of 2  from those  calculated from Equation 4.5-3.  Never-
  theless, the penetration curve will resemble that of Figure 4.5-11.  Gener-
  ally, packed-bed  scrubbers are relatively ineffective for sources with a
  major fraction  of the particulate emissions in  the <2.0-ymA size range.

Optimization of an existing unit could be done by changing gas velocity, bed
height, and packing  type.   The liquid-to-gas ratio  (L/G)  is  not a control-
ling  factor with  the  packed  bed  scrubber,  as  it  is with  other scrubber
tvoes.  Tray-type scrubbers,.  An impingement-plate scrubber is typical
of penetration conditions  created in a tray-type scrubber.  The theoretical
penetration of this  type of scrubber is  illustrated in Figure 4.5-10.   The
dominant  particle-collection  mechanism  is inertial  impaction  to droplets
formed as  the  gas  stream passes the impingement hole.  The penetration rate
of particles of  aerodynamic diameter i is calculated according  to Equation
    P. = 1 -
              K. +0.7
°'5                                       (Eq.  4.5-5)
     P. - penetration  through  an impingement  plate  scrubber of  a particle
          with aerodynamic diameter of i, dimensionless.
     K. - impaction parameter  of a  particle with aerodynamic diameter of i,
The  impaction parameter  is calculated  by  use of  Equation 4.5-4 with  the
velocity term being  the velocity of the gas  at the  vena contracta after it
passes through the  hole.   Calvert et al. suggest using a typical vena con-
tacta  velocity  of 1.43  times  the  gas  velocity in the  hole.3   In Equation
4.5-4 the "collector" diameter is taken as diameter of the impingement hole.
     The theoretical penetration derived by use of Equation 4.5-5 is shown
in  Figure  4.5-10.   A  comparable data  set14  for a  full-scale  impingement
scrubber or a rotary salt dryer  is shown in Figure 4.5-11.
     Performance of the impingement plate scrubber is controlled by the same
basic factors as those of other wet scrubbers utilizing inertial  impaction to
water  droplets.   Those  factors are  aerodynamic particle size,  relative
velocities of particles and droplets, and liquid-to-gas ratio.   The impinge-
ment plate  scrubbers and  the  entire family  of tray  scrubbers  operate  the
inertial  impaction  mechanisms  somewhat less  effectively than the  venturi
scrubbers; accordingly, they allow higher penetration in the range of 0 to 1


                            91,100 acfm @ 170°F
                            1,000 gpm LIQUOR FLOW
                            3/32-in. HOLES ON 3/16-iru CENTERS
                            13-ft. DIAMETER TRAYS

  Figure 4.5-10.  Theoretical penetration curve for impingement
                         plate scrubber.


1-41 Nm3An1n
0.035 nr  LIQUOR/min.
Ap = 3kPa

   Figure 4.5-11.   Penetration curve for an impingement plate
                scrubber on a rotary salt dryer.

-------   Venturl and Orifice Scrubbers    The  penetration  equation
developed by Calvert et al. for venturi and orifice scrubbers is Equation 4.5-6,
based on inertia! impaction of particles onto water droplets.3
                                  -0.7-k.f+ 1.4 In
                                        1         V    0.7    / MCff +0.7,
1 1
     P. = penetration value  for particles with  aerodynamic  diameters  of i.
     p  = droplet density, kg/m3.
     d. - droplet diameter, m.
     v  = superficial gas velocity in venturi throat, m/s.
      jj = gas viscosity, kg/m-s.
  Q,/Q  = liquid-to-gas ratio, dimensionless.
     k- = impaction parameter, dimensionless.
      f = nonuniformity correction factor, dimensionless.
     Selection  of a  proper  f factor  is important  in making  an accurate
performance prediction.   Calvert et al. suggest f values ranging from a low
of  0.10  to a  high of  0.70  with typical  values being  0.25 to  0.50.3  Low
values are  applicable to scrubbers having high  liquid-to-gas ratios and/or
having hydrophobic particles.   Increasing  f values  leads  to substantially
reduced penetration predictions. i  The droplet diameter is normally calculated
using the equation of Nukiyama and Tanasawa presented below as Equation 5.4-7.
    d . = Sauter mean droplet size, cm.
    VQ = gas velocity in venturi throat, cm/s.
 Q-./Q  = liquid-to-gas ratio, dimensionless.

Boll et  al.ls  determined that Eq. 4.5-7 Is most accurate at a throat veloc-
ity of 4570  cm/s.   At throat velocities and liquid-to-gas ratios typical of
commercial units  the mean  droplet size  is  estimated with  an accuracy +50
percent.   As an alternative, Boll suggests Equation 4.5-8.
to  o
   drf = 283.000 +  793    1/g/                                    (Eq. 4.5-8)

    d . = sauter mean droplet size, urn.
    Vg = gas velocity in throat, ft/s.
 QL/QQ = liquid-to-gas ratio, gal/1000 ft3.
The  impactor parameter  takes into  account the  fundamental  variables that
influence  inertial impaction,  namely,  aerodynamic  particle size  and par-
ticle-droplet relative velocity.  This is parameter calculated from Equation
                                                                 (Eq. 4.5-9)
    K. = impaction parameter, dimensionless.
                                               —4        0 *>
    d. = aerodynamic particle diameter i, cmxlO   (g/cm3)   .
    u  = gas viscosity at actual temperature, g/s-cm.
    d . = sauter mean droplet size, cm.
    V., = particle-droplet relative velocity, cm/s.
The  theoretical  penetration  curves  that can  be generated  from Equations
4.5-6 to 4.5-9 are shown in Figures 4.5-12 and 4.5-13.  These figures illus-
trate the  strong influence  of throat velocity and  liquid-to-gas  ratio in
these equations.   Results of  this approach are  conveniently summarized in
Figure 4.5-14 for a case with f = 0.25.3'16
     A comparison of Cal vert's model  (Equations 4.5-6, 4.5-7, and 4.5-9) and
actual scrubber performance data17 is shown in Figure 4.5-15.  The empirical
f  factor strongly  affects  the degree  in which  the model fits  a specific

                                           8-1  » 1.5 liters/m3
                       THROAT VELOCITIES

Figure 4.5-12.   Theoretical penetration curve for a venturi scrubber
               illustrating effect of throat velocity.

                                    VG - 10,000 m3/sec
                                         f =  0.25
                                         T =  25°C
Figure 4.5-13.   Theoretical penetration curves for venturi  scrubber
           illustrating effect of liquid-to-gas ratios.


2  o.i
        Calvert's model
        f = 0.5
          	 Experimental  Data
                                             Calvert's model
                                             f « 0.25

                            PARTICLE DIAMETER, ymA
          Figure 4.3-15.
Comparison of Calyert's Model results against
measured penetration data.

     Yung, Calvert, and Barbarika   have presented a refined model, which
incorporates a number of changes from the Calvert model, including elimination
of the f parameter.  A simplified equation is presented in Equation 4.5-10
for a scrubber in which all particle capture occurs in the throat.  It should
be noted that the inverse tangent function should be expressed in radians.
                                                                       K  i 0.5-
                                                                (Eq. 4.5-10)

                                                                (Eq. 4.5-11)
                 xgJ LHgJ LUDOJ
           C d..2 p_u^
                                                                   ,. 4.5-12)
     CDO = 55/NRfi (for 100 < NRfi <500)                          (Eq. 4.5-13)

     ude = 2 [1 - x2 + (X4 -X2)0'5]                             (Eq. 4.5-14)

       x = 1 + 0.187    tDO  pg	                             (Eq. 4.5-14)
                         dd  pl
     1. = venturi throat length, cm.
   Cno = drop drag coefficient at throat inlet, dimensionless.
     p  = gas density, g/cm3.
     p, = drop density, g/cm3.
     drf = drop diameter, cm.
   NR  = drop Reynolds number, dimensionless.
     Mq = gas viscosity, kg/cm-s.
   Up. = gas velocity in throat, cm/s.
     Q, = liquid flow rate, cms/s.
     Q  = gas flow rate, cm3/s.
   u*S  = ratio  of liquid  drop  velocity at  throat exit  to gas velocity  at
         exit, dimensionless.
   C  = Cunningham correction factor,  dimensionless.


Figure 4.5-16  is a  comparison of the  Yung et  al.  model against  the same

scrubber performance  data presented earlier in  Figure  4.5-15.   The revised

model is considered  more accurate than the earlier  Calvert  model,  however,

it  is also much  more complex.18   In  another comparison  with field units,

Calvert,  Barbarika  and  Monahan17  have  concluded  that  the  revised  model

adequately predicts penetration with the following qualifications.
   5  0.1
Actual penetration  of submicrometer  particles  is less  than pre-
     2.   Actual penetration  of particles greater than 1 urn is greater than

     3.   Accurate  measurement  of  the  liquid-to-gas  ratio  improves  the
          predicted penetration curve.
               	Experimental data
               	Predicted by
                      Equation 5-23
                      and by infinite
                      throat model

                    PARTICLE DIAMETER, yroA
         Figure 4.5-16.  Comparison of Yung, Calvert, and Barbarika
                  Model against measured penetration data.

     Optimization of  venturi  scrubber design based on  modification  of Cal-
vert's model3  has been  discussed by Leith and  Cooper.19   Other theoretical
approaches  for calculating  particle collection  in venturi  scrubbers have
been described by Crawford7 and Strauss.10
     An  empirical  approach based  on the "contact" power  utilized  has been
described  by  a  number of  investigations.   The  contact power  approach  is
based  on the  concept that  penetration  is  directly  related to  the energy
input  into  the  gas-liquid  contact.8-13  For  gas-atomized  scrubbers,  the
power consumption, PC  is approximated by Equation 4. 5-16. 12
     pQ = 0.158 AP    '                                          (Eq.  4.5-16)
     PG = power consumed, hp/1000 acfm.
     AP = gas phase pressure drop, in. W.C.
When  the liquid  stream adds  a significant fraction  of the  total  energy,
equations presented in  references 12 and 13 may be used.  The total  energy,
PT, is the sum of the gas- and liquid-phase power consumption.
     PT = PG + PL                                               (Eq.  4.5-17)
The  scrubber  collection   efficiency,  NT,  is  then expressed  as shown  in
Equation 4.5-18.
     N  = aP^                                                   (Eq. 4.5-18)
     NT =  number of  transfer units  (^()), dimension-
     PT = total power consumption, hp/1000 acfm.
    a,y = constants, dimensionless.
The constants are parameters dependent on the characteristics of the partic-
ulate matter.   Using  this approach,  one assumes that no independent effects
can  be  attributed to throat velocity,  liquid-to-gas  ratio,  scrubber design
and other parameters.
     In certain cases good correlations can be achieved using this approach.
Figures 4.5-17  and  4.5-18 show the relationship between outlet loadings and
static  pressure drop (an approach similar  to  Equation 4.5-18).   Hesketh11


~  0.08
                            grains/DSCF = 12.44  (AP,  inches w.c.)"1*61
                                o         r  Correlation = 0.756
     20               30

        Figure 4.5-17.  Comparison of venturi scrubber outlet loadings to
                 static pressure drops for oil-fired lime kilns.

                           2            345

             THEORETICAL POWER CONSUMPTION, hp/1,000 icfm
Figure 4.5-18.   Correlation of coal-fired  boiler  scrubber outlet dust
            loading with theoretical  power consumption.12

has  developed  an empirical equation (4.5-19)  relating  static pressure drop
and  penetration.   As shown  in Figure 4.5-19  there is  a good relationship
between the two.
                 -1 A.1
     Pt = 3.47 AP •L*^                                          (Eq. 4.5-19)

A more  complete description  of the contact power  approach  is available in
references 8 through 13.   Other  scrubber types.    Many  other   scrubber  types  are
available, for most of  which the  penetration curves  are similar to those
presented in this section.  Additional  information on wet scrubber operating
principles is given in References 1, 3, 5, 7, and 10,,   StaticPressure Drop.   Factors affecting  the  static pressure drop
in a wet  scrubber  include scrubber geometry,  gas  velocity,  and the liquid-
to-gas  ratio.   Typical   liquid-to-gas  ratios are  listed  in  Table 4.5-2 and
typical static pressure  drops  are listed in  Table 4.5-3.   Calvert et al.3
have  summarized equations  useful  for predicting  pressure drop  in various
types of  scrubbers.   A  detailed  summary of static pressure drop equations
applicable  to   venturi   scrubbers  is  presented  by  Yung,  Calvert,  and
Barbari ka.1S                                                             ;
                                        Liquid-to-gas ratio,
          Scrubber type	liters/m3	
          Venturi                            0.70 - 1.00
          Cyclonic spray tower               0.70-1.30
          Spray tower                        1.30 - 2.70
          Moving bed                         1,30 - 2.70
          Impingement plate                  0.40 - 0.70
          Packed bed                         0.10 - 0.50

2        4    6   8  10       20
60  80 100
    Figure 4.5-19.   Comparison of predicted penetration  as  calculated  in
     Equation 4.5-19 and measured penetration  (reprinted by permission
     from Hesketh,  J.  Air Poll.  Control  Assoc.  Vol.  24,  No. 10,  1974).

                                           Pressure drop,
          Scrubber type     	kPa	
          Venturi                            1.5  - 18.0
          Centrifugal (cyclonic) spray       0.25 -  0.8
          Spray tower                        0.25 -  0.5
          Impingement plate                  0.25 -  2.0
          Packed bed                         0.25 -  2.0
          Wet fan                            1.0  -  2.0
          Self-induced spray (orifice)       0.5  -  4.0
          Irrigated filter (filter bed
           scrubber)                         0.05 -  0.8
     Calvert et  al.3  presented a simple approach for calculation of venturi
static pressure, as shown in Equation 4.5-20.

     AP = 0.001 V*(Q,/QJ                                       (Eq. 4.5-20)
                 t  I  y
     AP = static pressure drop, cm. W. C.
     V. - throat velocity, cm/s.
      t                                                                 !
  Q,/Q  = liquid-to-gas ratio, (cms/s)/(cm3/s).
Yung et  al.18 have discussed  a modified form of this  equation as  shown in
Equation 4.5-22.   The parameter  u.  is the same as  that described earlier
with respect to Equation 4.5-14.
     AP = 0.001 ue Vt (Q/Qg)                                  (Eq- 4.5-21)
Hesketh11  described  an  equation  including throat  area,  Equation 4.5-22.

     AP = (Vt2pGA0'133L°*78)/1270                               (Eq. 4.5-22)
     AP = static pressure drop, in. w. c.
     V. = throat velocity, ft/s.
     PQ = gas density, lb/ft3.
      A = throat cross-sectional area, ft2.
      L = liquid-to-gas ratio, gal/1000 acfm.

4.5.3  Design of Particulate Scrubbers  Sizing of Wet Scrubbers.  The overall dimensions of a scrubber
are established to  provide  the design gas velocity within  the various sec-
tions of  the vessel.   Care must be  taken to  ensure  an even  gas velocity
throughout the mixing and demisting sections.   Sharp turns should be avoided
in ductwork,  and  distribution baffles should be used where short-circuiting
would otherwise be likely to occur.   In general, scrubbers operate at higher
average gas velocities than fabric filters or ESP's and therefore are usual-
ly more compact.1  Nozzle  Selection and  Liquid  Distribution.   The  liquid distri-
bution  system  in  a  scrubber  is intended to  provide even distribution of
properly  sized  liquid  droplets  for  contacting  particulate  in  the mixing
zone.  For this purpose, nozzles must be selected that will properly atomize
the  liquid.   The  two  general  categories of  nozzles use  either hydraulic
pressure  or  compressed  air  to atomize  the  water.   The various  types of
hydraulic-pressure  nozzles  produce  hollow-cone,  solid-cone,  or fan-shaped
sprays  with  various  spray angles.   Two-fluid  nozzles   use  compressed air
instead  of water  pressure as  the  primary force  for atomizing  the water.
Use  of  two-fluid nozzles  is  especially attractive when  an extremely small
droplet  distribution is  desired and when  fluid viscosity  is  a problem.
     Most  nozzles produce  a broad spectrum of droplet sizes rather than one
distinct  size.   It  is often convenient, however, to express droplet size by
a single value  such as the average diameter or the Sauter mean diameter (the
hypothetical  droplet  whose  ratio of surface area to volume is equal to that
of the  overall  spray).   In general, increasing water pressure (or air pres-
sure  in two-fluid  nozzles) will reduce the average  or  Sauter mean droplet
diameter.  The  droplet  size population may be modified by additives such as
propan-1-ol  but  not  necessarily by  detergents.20  Hesketh21  has reported
that  the  addition of nonionic,  low foaming surfactants  reduced outlet dust
loadings  50 percent  by  improving particle wettability  and/or atomization.
     Because  many scrubbers  are  operated with  recirculating  slurries, the
design  of nozzles  introducing the  scrubbing  liquid  is  of  critical impor-
tance.   With some  scrubbers  (packed  towers',  gas-atomized  units) extensive

distribution is  needed  at the point of  liquid  entry,  but the scrubber ele-
ments provide  liquid distribution.  With other  scrubbers  the requirements
range from  a hollow-cone  spray to a  full-cone  spray.   The most stringent
requirements are  for high-pressure sprays that create the small water drop-
lets  needed  for  high  efficiency in scrubbers  whose primary consumption of
energy is in the  nozzles (preformed-spray scrubbers).
     Nozzle plugging may be a major problem.22  The area formerly sprayed by
a plugged nozzle becomes subject to scaling  and  heat  damage.   The critical
factor in  nozzle design  is  the minimum  internal orifice.   As the scrubber
size  increases,  the nozzle size normally  increases.   Under highly abrasive
conditions, ceramic  nozzles  may be needed.  Special care must be taken when
installing  and   removing  these  nozzles  to avoid breakage.   Heat  can  also
cause breakage if a proper method of installation is not followed.23   Presaturatorsfor Hot Gases.   The  scrubbing of  particulate
from  hot  gases  presents  special problems not  associated with gases  at am-
bient temperature.  The heat in hot gases can evaporate substantial portions
of  the   scrubbing  liquid  droplets and  adversely  affect  liquid/  particle
contact.   In a hot gas stream, some droplets may evaporate before particulate
contact and others  may  evaporate after particulate contact,  and thus cause
the particulate to be reentrained.   Hot gases can also damage scrubber mate-
rials,  especially  fiberglass-reinforced  plastics.    Presaturation  can  be
economical when  cooling  of the gases  permits  the  use of  less expensive
materials of construction  in the scrubber and  when the  lower volume of the
cooled gases allows  the use of smaller scrubber vessels, fans, dampers, and
     High-temperature gases  are usually  cooled to near  saturation  by spray
quenching prior to entry  into a  scrubber.   In  most  scrubber applications
approximately 1%  to  2%  times the theoretical  evaporation demand is required
to quench the gases because of the kinetics of the cooling process.24  As in
the  scrubbing  process,   nozzle  design  and  arrangement are   important  in
quenching.  As  size  of  the quench water  droplet  decreases,  the kinetics of
the cooling process  increases  and the evaporation demand becomes closer to
     Quenching  is frequently accomplished with scrubbing liquor rather than
clean water.   In some applications, however, the use of scrubbing liquor for


quenching  can  reduce  scrubber performance.   Most recirculating  scrubbing
liquors contain very high levels of both suspended and dissolved solids.   As
quench water evaporates, these solids can be reentrained into the gas stream
and must  be collected  again  in the  scrubber.   Dissolved solids  in evapo-
rating quench  liquor can form  fine  particulate in the size  range  that  may
escape collection  in  a  scrubber.25  If fine particulate  are  regenerated in
this manner, the  net effect is that contaminants are returned to the scrub-
ber inlet  gas  stream  in a form  more difficult to collect.   It is usually
best,  therefore,  to  use the cleanest water available  for presaturator of a
scrubber.   This can  be  accomplished  by adding  the makeup water directly to
the quench  system rather than adding all makeup water to a common sump for
the quench and scrubber liquors.  Using clean water alone for the quench process
is even better.  Staging of Scrubbers.  When higher collection efficiencies are
required than  can be obtained in a single scrubber,  scrubbers are sometimes
staged to  improve efficiency.   When  scrubbers  are placed in  series, end to
end,  the  penetration of particles of  one  size through  the series  is  the
product of  individual  penetrations of particles of that  size (see Equation
4.5-23).   The  corresponding pressure drop is the sum of individual pressure
drops  per  stage  (see  Equation  4.5-24).   There is  a diminishing  overall
effectiveness in  successive stagings, however,  if the individual penetration
through a  single  stage  by the finer  particulate  in  a gas stream is consid-
erably higher  than the  penetration of the coarser particulate.  Most of the
coarse particulate might be collected in the first stages while each succes-
sive stage  collects  only a small fraction of the remaining fine particulate
at  a  relatively  high  energy  cost.   Staging  is most  common with tray-type
scrubbers and  packed bed scrubbers and  is  usually limited to three stages.
Staging can  also  be done with two types of scrubbers, such as a spray-tower
precleaner followed by a packed-bed scrubber.
     P.  = P.  x  P.  x P.   	 p.                              (Eq. 4.5-23)
      \     nl     ^2     3       ^
   P.  = total penetration of particles of aerodynamic diameter i
   P.  = penetration of particles of aerodynamic diameter i from stage 1

  4Pt  = ^  + &P2  + ••• + APn                                (Eq, 4.5-24)   Liquid  Entrainment  Separators.   Wet scrubbing  of particulate
is  a  two-step process,  the second  step  being separation  of the scrubbing
liquid droplets from the gas stream.  This step is important in the ultimate
collection  of particulate because  poor liquid separation  will cause reen-
trainment of the particulate.
     There  are four baste types of  liquid entrainment  separators26 or "de-
misters"  (Figure  4.5-20).   The mesh-pad and  chevron  types  utilize inertia!
impaction of  the  liquid  droplets  to cause their  agglomeration and removal.
The centrifugal  and cyclonic  types utilize  centrifugal  inertia to collect
the liquid  droplets.
     Plugging  can  be  a  persistent problem in mesh and  chevron mist elimi-
nators in certain  applications.   Centrifugal-type mist eliminators are less
prone  to plugging.   Plugging  can  usually be minimized  by  continuous  or
intermittent spraying of the mist eliminators.   Liquid Handling Facilities.   Water  usage and  waste disposal
may become  critical  factors in the  final  selection of  a wet scrubber.  The
quantity of particulate collected, the size distribution of the particulate,
and the  presence  of  dissolved  contaminants  in  the  scrubbing  liquid have
great bearing  on  the amount of water and the type of liquid handling facil-
ities  needed.   Present  water quality  regulations  require  that  most  new
scrubber installations and  many existing scrubber installations recirculate
scrubbing liquors  to  prevent  the contamination of surface  waters with  ,the
collected air contaminants  (Section  6).   Recircufation, however,  tends to
concentrate the dissolved scrubbing liquor contaminants.
     Liquid  handling facilities  in recirculating scrubber  systems usually
include  a  slurry pump; a. makeup  water pump;  a settling basin or pond;  and
associated  piping,  valves,  and spray  nozzles.   It is  sometimes necessary
to  construct  multiple settling basins or to install  clarifiers  with drag
chains or  rotary sludge  collectors to settle  and remove  suspended solids
from  scrubbing  liquors.   Additional  procedures  for  liquid  handling  can
Include  filtration; chemical  treatment,   for example  to  control  pH level
and aid  flocculation;  and many other treatments common to industrial waste-
water treatment facilities (Section 6).


                                                 GAS OUT
(Courtesy of Koch Engineering
Company, Inc.)
                                                            GAS IN
                                                       TANGENTIAL INLET
                                         CENTRIFUGAL MIST COLLECTOR6.
                                                               GAS OUT
                                                                GAS IN
                                                       CYCLONIC MIST
      Figure 4.5-20.   Liquid entrainment separators (courtesy
           of Industrial Gas Cleaning Institute, Inc.).

     Because  scrubbing  slurries  are  often  corrosive  and  abrasive,  all
liquor handling  pumps, piping,  nozzles,  and valves must  be  constructed of
resistant materials  or be lined with suitable  protective  materials.   Since
slurries can  also cause  plugging,  it  is  advantageous to  install  cleanout
traps and service hatches in many components.  In  some  systems reliability
can be ensured only by installing duplicate pumps.   Materials of Construction.    Materials  of  construction  for
scrubber applications  must  be carefully selected to withstand corrosive and
abrasive agents  in  the gases or liquors,  and to withstand any high tempera-
tures that may occur.  If the process conditions are properly defined before
design begins, the  experienced scrubber manufacturer  can  design  a scrubber
that will withstand  its  service environment.  Many scrubbers, however, fail
because inappropriate  materials  are  selected after a superficial  investiga-
tion of  process   conditions,  or because insufficiently  resistant materials
are substituted to reduce costs.27
     Investigation  of  each  scrubber  application  should   include  chemical
analysis of the  raw materials, combustion products,, and  scrubbing liquids.
The operating  histories of  any scrubber  installations  in similar applica-
tions  should  also  be reviewed.   Finally,  review  of  the  literature  about
materials performance  is recommended;  and when  materials  performance data
are not available, in situ coupon tests may be required.   After all relevant
information has  been compiled,  the  designer prepares a list  of materials
suitable for the expected service.   Selection  of materials of construction
from this  list of candidates will be  based in part on the relative costs.
     Although the above mentioned procedures should be  followed in selection
of materials,  some general  aspects  of materials applications can  be men-
tioned.  Table 4.5-4 lists the major types of metals available for scrubbers
and ancilliary  components,  together with  their major properties,  general
corrosion behavior,  and  relative costs.   Not  listed are the nonmetallic
materials such as fiberglass-reinforced plastic,  ceramics, protective coat-
ings,  and wood, which are appropriate in many scrubber  applications.   Instrumentation.   Proper  instrumentation  is   vital  to  the
monitoring of scrubber performance.   Many installations require instrumenta-
tion with  associated alarms  and interlocks  to protect  valuable  components

           Corrosion resistance
Cast iron
High strength; low ductility;
 brittleness; hardness; low
Carbon steel
(410, 416,
420, 440c)


Good strength, ductility,
 workability; low cost
Chromium alloy, hardenable by
 heat treatment; typically used
 for machine parts; costs 2 to
 5 times more than carbon steel
 Chromium alloy, not hardenable
  by heat treatment; costs 2 to
  4 times more than carbon steel
Modified for weldability
General-purpose, often used for
 chimney liners
Used in high-temperature service
Ordinary cast irons exhibit
 fair resistance to mildly
 corrosive environments; high-
 silicon cast irons exhibit
 excellent resistance in a
 variety of environments
 (hydrofluoric acid is an
 important exception); cast
 irons are susceptible to
 galvanic corrosion when
 coupled to copper alloys or
 stainless steels

Fair to poor in many environ-
 ments; low pH and/or high
 dissolved solids in moist or
 immersion service leads to
 corrosion; properly applied
 protective coatings give
 appropriate protection in
 many applications; susceptible
 to galvanic corrosion when
 coupled to copper alloys or
 stainless steels

        Good; better than martensitic
         stainless steels; resists
         stress corrosion; better
         chloride resistance than
         austenitic stainless steels
         Good resistance to atmospheric

TABLE 4.5-4 (continued)
  Corrosion resistance
   201, 202
Nickel alloy



Chromium and nickel alloy; not
 hardenable by heat; hardenable
 by cold working; nonmagnetic
Types 201, 202, 301, 302, 303,
 304, and 304L cost 3 to 5 times
 more than carbon steel; types
 310, 316, 316L, and 321 cost 4
 to 10 times more than carbon

Nitrogen added, used as a substi-
 tute for 301 and 302

Good hardenability
Modified for weldability
Used in high-temperature service
Used in corrosive environments
Improved weldability
Good strength; costs over 10
 times more than carbon steel
High strength; light weight (60%
 that of steel); costs over 10
 times more than carbon steel
Excellent; better than
 martensitic or ferritic
 stainless steel (except
 for ha'lides)
Superior corrosion resistance;
 good acid resistance; resistan'
 to hot organic acids; good
 pitting resistance

Excellent resistance in most
 environments; not resisant
 in strong oxidizing solutions
 such a:; ammonium and HN03

Good resistance to stress

Good resistance to hydrofluoric
                                   Excellent overall resistance
Exceptional resistance at
 ambient temperatures
Excellent resistance at other
 temperatures, except that
 crevice corrosion is possible
 in chloride solutions above
 Registered trademark of Huntington Alloys, Inc.

Registered trademark of the Stalite Division of Cabot Corporation.

Registered trademark of the Duriron Company, Inc.

from malfunctions  such as  loss  of water pressure or a  process  temperature
     Every major  scrubber system  should include a meter to  measure static
pressure drop across the scrubber and a meter to indicate water flow through
the  scrubber.   Static  pressure  drop  can be  measured  with  a differential
pressure gauge or manometer.  Care must be taken in the  design of the tubing
and  fittings  to  prevent plugging  and to allow  easy cleaning,  and tubing
materials  should  be  selected  to  withstand  the  service  expected.   For
example,  certain plastics  can  melt when exposed to high temperatures,  some
plastics become  excessively brittle at  low  temperatures,  and polypropylene
tubing is degraded  under continuous exposure to sunlight.   Water flow rates
can be measured by in-line flow meters or doppler type indirect flow meters.
A less expensive  and less accurate method of flow measurement is the use of
a pump pressure  gauge calibrated to indicate flow rates.  Open-channel  type
flow-measuring  devices such  as  the  Parshall  flume  are  sometimes  useful,
although the preferred measuring point for  liquor flow  is  between the pump
outlet and the scrubber spray nozzles.
     For  sources  that generate  hot  gases,   the  gas temperatures  must  be
monitored  if the  scrubber contains  materials  that  cannot  withstand  high
temperatures.   A  high-temperature  alarm and/or an interlock  system  is
usually installed to shut down the process or to bypass  the scrubber system.
Alarm  systems  can  also  be  included  to  indicate  low  water  levels  in
orifice-type scrubbers.   In systems that  include  presaturators,  water flow
through  the  scrubber and the presaturator should  be  measured individually.
Where gas  temperatures vary widely,  it  is  sometimes necessary  to  install
temperature  feedback instrumentation that controls the  water flow rates to
the presaturator.
     Scrubber  instrumentation  often  includes  liquor   pH  indicators,  fan
ammeters, and  fan  vibration sensors.   The pH meters are needed  when pH of
the  scrubbing   liquor must be closely controlled.   Maintaining  clean,  ac-
curately  calibrated probes, although  often  difficult,   is  essential to the
success of pH control.  Fan ammeters and tachometers can be used  in conjunc-
tion with the  manufacturer's  fan performance  curves  to  provide  an estimate
of gas flow through  the scrubber system, or these instruments can be used to

provide a  quick comparison  of the system's performance with  previous per-
formance.   It  is  helpful  in all scrubber systems  to  provide small  ports in
the ducting  before and after  the  fans,  the scrubber vessels,  and  the pre-
saturators to  allow pitot velocity traverses,  gas temperature measurements,
and static pressure measurements.
4.5.4  Operation and Maintenance of Particulate. Scrubbers   Common  Malfunctions.   Wet  scrubbers can  provide continuous,
reliable service  when they are operated properly  and maintained regularly.
Poor operation and maintenance  leads  to component failure.   Most  scrubber
failures  result  from abrasion,  corrosion,  solids  buildup,   and  wear  of
rotating  parts.   Common  failure  modes  for  individual  components  are
discussed below.  Nozzle plugging.   Nozzle  plugging  is  one of the most common
malfunctions in scrubbers.22  Plugged nozzles reduce the liquid-to-gas ratio
or cause maldistribution of the liquid.   Nozzle plugging results from improper
nozzle selection, excessive solids in scrubbing liquors, poor pump operation,
and poor sump design.   Remedies for nozzle plugging include replacement with
nozzles of a  different  type,  frequent cleaning,  and reduction of  liquor
solids content by increasing liquor blowdown and makeup water rates.  Because
presaturator  nozzles  are  especially prone  to  plugging,  the  quench water
should be limited to fresh water or very dilute liquors.  Many quench nozzles
cannot  tolerate greater  than  2  percent solids  in  the  liquid.23   Nozzle
plugging can be detected  by observing the  liquid  spray pattern the nozzles
produce.  If the  nozzles  are not accessible while the  pumps are operating,
they should be checked during scrubber shutdowns for evidence of caking over
the nozzle openings.   A decrease in water flow rate during scrubber operation
is an additional symptom of nozzle plugging.  Solids buildup.  Solids buildup is another problem common to
wet scrubbers  and  one that is often difficult to control.   The two types of
solids buildup are sedimentation and chemical scaling.  Sedimentation occurs
when a  layer  of particles becomes attached to a surface or settles in areas
of low turbulence.  Sedimentation can lead to plugging of pipes and ducts or

buildup on internal  parts.   Chemical  scaling results from a  chemical  reac-
tion of two  or  more species to form a precipitate on the surfaces of scrub-
ber components.
     Solids buildup may occur in piping, sumps, scrubber packing, instrumen-
tation lines, or  ductwork,  and may lead to  reduced  scrubber  efficiency and
major  equipment  failure.   Most  scrubbers  using  open pipes cannot reliably
tolerate liquor slurries of over 15 percent solids by weight.   It is usually
best to maintain  solids content at less than  6 to 8 percent.23  Techniques
to  control  scaling include increasing  the  liquid-to-gas  ratio, controlling
pH, providing greater  residence time in the holding  tank,  and adding other
chemical  agents  such  as dispersants.   Solids buildup  can  be  detected by
inspection of accessible  components  and by inspection of the  inner surfaces
of piping, tubing, and ductwork at removable fittings and hatches.   Corrosion.    Corrosion  problems   arise   frequently  in  wet
scrubbers,  especially  when  the  gases  being  cleaned contain  acid-forming
compounds  or soluble   electrolytic  compounds.   The  combustion  of  fossil
fuels,  especially  coal,  coke,  and  residual   fuel  oil,  yields  oxides  of
sulfur,  which  can  produce  sulfuric acid  in  scrubbing  liquors.   Metals-
refining  processes,  such  as  copper  and  lead smelting,  can  also  produce
oxides of sulfur.  Combustion of polyvinyl  chloride plastics,  commonly found
in incinerator feedstock, can produce hydrochloric acid in scrubbing liquors.
Rotary aggregate  dryers and similar process equipment can produce chlorides
or  fluorides, depending on  the composition of the aggregate.   The phosphate
fertilizer  industry and  the  feldspar  industry  are  especially troublesome
sources of fluorides.   Acids  and electrolytes in general are corrosive to
mild steels, chlorides are corrosive to many stainless steels, and fluorides
are harmful  to  nearly  all stainless steels  except  certain specially formu-
lated  (and  expensive)  high-nickel  alloys.28   Recirculation  of scrubbing
liquors greatly  increases the  concentrations  of  any corrosive agents they
     Prevention of corrosion  is best handled through proper choice of mate-
rials  of  construction  and through pH control.  When  a  pH control system is
to be the principal defense against corrosion, regular maintenance at frequent
intervals  is necessary,  especially at  the  pH electrodes.   Another common
operating  problem  occurs  when scrubber liquor blowdown rates are reduced to

limit  the   emission   of  pollutants  into  surface  waters.    Reducing  or
eliminating  blowdown  can  so  greatly  increase  the  acid  and  electrolyte
concentrations  in the  liquor  that  otherwise  acceptable materials of  con-
struction become ineffective against corrosion.   Abrasion.   Abrasion can  occur  where  gases or  scrubbing
liquors  containing  high  concentrations  of abrasiva particulate are  in the
turbulent  mode  or  are  subjected to  a  sudden change  in  flow  direction.
Typical  wear areas  in  scrubbing systems  include venturi throats,  walls of
centrifugal  mist  collectors - near  the  inlet  duct,  and   elbows  in  the
ductwork.23   Solutions  to  abrasion wear  include  the  use of  precleaning
devices and the use of large-radius turns in ductwork.   Wear of rotating equipment.    Rotating  equipment  including
fans,  pumps,  and clarifiers  must  receive  special  attention in  scrubber
service  because  of  potential  abrasion,  plugging, and  corrosion.   Key  wear
areas in these  components include the bearings and  any components rotating
in the fluid stream.31                                                   :
     Fan wear  is  a  common problem.  Forced-draft fans often suffer abrasion
because  of  exposure to  particulate-laden  gases.   Wear  problems  in forced-
draft fans can  be addressed by the use of special wear-resistant alloys, by
reduction of fan  rotation speeds (by installing a larger fan), or by moving
of the fan to an induced-draft location on the clean air side of the scrubber
system.   Induced-draft  fans  can  undergo corrosion or  solids  buildup  on the
blades  if  mist  is  carried  over from the  liquid entrainment  separator.
Induced-draft  fan problems can  be addressed  by  use of corrosion-resistant
materials or by improving liquid entrainment separation.
     Pump wear is also a common problem in scrubber systems.  Pump housings,
impellers,  and seals  are  subject  to  abrasion and  corrosion by  scrubber
slurries.  Rubber linings and special-alloy pump materials are often used to
reduce abrasion and corrosion of the  housings or impellers.   Installation of
a water flush in the seals can help reduce wear of the seals.31   Preventive Maintenance.   Preventive maintenance  is  an  impor-
tant tool in assuring the continuous  operation of scrubber systems.  Preven-
tive maintenance  programs for scrubbers should include periodic  inspection
of  equipment,  replacement  of  worn  parts,  periodic cleaning  of  components

prone to  plugging,  maintenance  of  an adequate  spare parts  inventory,  and
recording of all maintenance performed on scrubber equipment.
     All  instrumentation  such  as differential  pressure  gauges,  scrubbing
liquor  flow  meters,  pump  pressure  gauges,  and  fan  ammeters  should  be
observed  at  least once per work shift.   All equipment  should  be inspected
regularly at  regular Intervals,  determined  by the severity  of service  and
the  likelihood  of component  failure.   Failure-prone items  include nozzles
and  pumps handling  slurries,  forced-draft fans  handling  particulate-laden
gases,  induced-draft  fans  downstream  of  inadequate  liquid  entrainment
separators, wear  plates,  pH  probes,  and  bearings.   These items should be
inspected as often as once per shift depending on the likelihood of failure.
Such components as ductwork and induced-draft fans handling clean, dry gases
should be inspected monthly.
     All  worn parts  and malfunctioning equipment should be serviced as they
are discovered to prevent deterioration of system performance and to prevent
damage  to equipment.   An  inventory  of  spare parts  must  be maintained in
stock for replacement of  nozzles,  bearings, pump  seals,  liners for pumps
with  replaceable  liners,  pump  impellers,  wear plates for fans wheels with
wear  plates,  pH probes,  and  valve  parts.30   Records should  be made  of  all
maintenance performed and all parts replaced.  This information is useful in
planning  subsequent  preventive  maintenance schedules and  in determining the
type and  number of replacement parts needed.

                                 REFERENCES                                 ,

 1.  Theodore,  L.  and  A.  J.  Buonicore,   Industrial Air  Pollution Control
     Equipment for Particulates.  CRC Press, Cleveland.  1976.

 2.  Vincent, E.  Wet Collection Devices.  In:  Air  Pollution Engineering Manual
     U.S.  Environmental  Protection Agency Publication  No.  AP-40, May 1973.
     pp. 99-106.

 3.  Calvert, S., J.  Goldshmid, D. Leith, and D. Mehta.  Wet Scrubber System
     Study,  Volume I:   Scrubber  Handbook.   U.S.  Environmental Protection
     Agency,  Research Triangle  Park,  N.C.   Publication No.  EPA-R2-72-118a,
     August 1972.

 4.  Sargent,  G.  D.    Dust  Collection  Equipment,  Chemical   Engineering.
     January 27, 1969.  pp. 130-150.

 5.  Licht, W.  Air Pollution Control Engineering.   Marcel Dekker, Inc.  New
     York. 1980,

 6.  The   Mcllvane  Company.   The  Wet  Scrubber  Handbook.    Northbrook,
     Illinois.  1974.

 7.  Crawford, M.  Air  Pollution Control Theory.  McGraw Hill Book Company.
     New York.  1976.

 8.  Semrau, K. T., C.  L.  Witham, and  W.  W. Kerlin.   Energy Utilization by
     Wet Scrubbers. U.S.  Environmental  Protection Agency, Cincinnati, Ohio.
     Publication No.  EPA-600/2-77-234, November 1977.

 9.  Semrau,  K.  T.  and C.  L.  Witham.   Wet  Scrubber  Liquid   Utilization.
     U.S.  Environmental  Protection  Agency,  Cincinnati,  Ohio.    Publication
     No. EPA-650/2-74-108, October 1974.

10.  Strauss,  W.   Industrial  Gas  Cleaning, 2nd  Edition.   Pergamon Press,
     Inc.  New York,  N.Y.  1975.

11.  Hesketh,  H.   Fine Particle  Collection  Efficiency Related   to Pressure
     Drop, Scrubbant and Particle Properties, and  Contact Mechanism.  J. Air
     Pollution Control Association.  24; 939-942.  October 1974.

12.  Ranade,  M.  B.  and E.  R.  Kashdan.   Design Guidelines  for an Optimum
     Scrubber  System.   Second Symposium on  the  Transfer and Utilization of
     Particulate Control Technology.   U.S. .Environmental Protection Agency,
     Cincinnati, Ohio.   Publication No.  EPA-600/9-80-039a,  September 1980,
     pp. 538-560.

13.  Engineering Science,  Inc.   Scrubber Emissions  Correlation  Final Report
     to U.S. Environmental Protection Agency.  Contract No. 68-01-4146, Task
     Order 49.  May 1979.


14.  Calvert, S., N. Jhaveri, and S. Yung.  Fine Particle Scrubber
     Performance Tests, U.S. Environmental Protection Agency,
     EPA-650/2-74-093, October 1974.

15.  Boll,  R.  H.,  L.  R.  Flais, P.  W.  Maurer,  and W.  L.  Thompson.  Mean
     Drop Size  in a Full Scale Venturi Scrubber via Transmissometer.  J. of
     Air Pollution Control Association.  24: 934-938.  October 1974.

16.  Calvert, S., N.  C. Jhaveri, and  S.  Yung.   Fine Particle Scrubber Per-
     formance  Tests.   U.S.  Environmental  Protection  Agency   Publication
     No. EPA-650/2-74-093, October 1974.

17.  Calvert, S.,  H.  F. Barbarika, and G.  M.  Monahan.   Evaluation  of Three
     Industrial Particulate Scrubbers, U.S. Environmental Protection Agency,
     Cincinnati,  Ohio.   Publication  No.  EPA-600/2-78-032,  February 1978.

18.  Yung,  S.,  S. Calvert, and  H.  Barbarika.   Venturi  Scrubber  Performance
     Model,   U.S.  Environmental   Protection  Agency,   Cincinnati,  Ohio.
     Publication  No. EPA-600/2-77-172, August 1977.

19.  Leith,  D.  and  D.  W.  Cooper.   Venturi  Scrubber  Optimization.  Atmo-
     spheric Environment. Volume 14, 1980.  pp. 657-664.

20.  Atkinson,  0.  S.   F.  and W.  Strauss.   Droplet Size and Surface  Tension
     in Venturi Scrubbers.  J. Air Pollution Control Association.  28: 1114-18.
     November 1978.  pp. 1114-1118.

21.  Hesketh,  H.   Atomization  and Cloud  Behavior in  Wet  Scrubbers.  Pre-
     sented  at  the Symposium in Control  of Fine  Particulate Emissions from
     Industrial  Sources.   San Francisco,  California.   January 15-18, 1974.
     pp. 455-478.

22.  National Asphalt Pavement  Association.  The Maintenance and Operation
     of Exhaust  Systems in the Hot Mix Batch Plant.  Information Series  52,
     Second  Edition.   February 1978.

23.  Schifftner,  K.  C.  Venturi  Scrubber Operation  and Maintenance.  Pre-
     sented  at   the   U.S.  EPA  Environmental   Research  Information Center
     Seminar  on  Operation  and Maintenance  of  Air  Pollution  Equipment  for
     Particulate  Control, Atlanta, Georgia.  April 1979.

24.  Industrial  Gas Cleaning Institute,  Inc.   Scrubber System Major Auxil-
     iaries.  Publication WS-4.  Stamford,  Connecticut.   1975.

25.  Kalika,  P.  W.    How Water  Recirculation and  Steam  Plumes Influence
     Scrubber  Design.   Chemical Engineering.   July  28, 1969.   pp.  133-138.

26.  Industrial Gas  Cleaning Institute, Inc.   Basic Types  of Wet Scrubbers.
     Publication  No. WS-3.  Stamford,  Connecticut.  June  1976.

27.  Mappes, T.  and  R.  D. Terns.  An  Investigation of Corrosion in Particu-
     late Control Equipment.  U.S. Environmental Protection Agency,  Research
     Triangle  Park,   N.C.   Publication No.  340/1-81-002.   February  1981.

28.  Nation  Association  of  Corrosion  Engineers.   NACE  Basic  Corrosion
     Course.  Houston, Texas.  May 1977.

29.  Benzer, W. C.  Steels.  Chemical Engineering  77(22).  October 12, 1970.

30.  Fontana, M.  G.  and N. D. Greene.  Corrosion  Engineering.  McGraw Hill.
     New York, N.Y. 1967.

31.  Czuchra,  P.  A.   Operation and  Maintenance of  a  Particulate Scrubber
     System's Ancillary Components.   Presented at  the U.S. EPA Environmental
     Research Information Center Seminar on Operation and Maintenance of Air
     Pollution Equipment  for Particulate Control.   Atlanta, Georgia.  April

     Incinerators  are  seldom,  If ever,  used  solely to  remove  participate
matter because they  tend to be more  expensive and more energy intensive than
alternative  control  techniques.   Applications  are restricted  to  sources  of
combustible  matter with low  gas  flow  rates  and  low  particulate  concentra-
tions.   Principal  among these  are curing  ovens,  textile coating,  charcoal
manufacturing, food  processing,  and certain  chemical  processes.   The  par-
ti cul ate- laden gas stream  from these sources normally contains other pollu-
tants, such  as volatile organic compounds  (VOC),  carbon  monoxide,  and odor-
ous  compounds.  Particulate control  in  the  incinerator  may in fact  be  only
ancillary to  the  control of malodors or  VOC.   Information concerning use of
incinerators  for  gaseous  pollutant  control  is  available  in Reference  1.
     An  incinerator  vaporizes and oxidizes particles.   It is  the  only par-
ticulate control  system that does not concentrate the particulate matter for
subsequent disposal.
4.6.1  Types of Incinerators
     Three  basic  types of   incinerators  are  used  for  particulate  matter
removal:   direct, thermal,  and  catalytic.   Because  the catalytic  type  is
prone  to severe  operating  problems  with particulate-laden gas streams,  its
use  is limited.   Direct and Thermal  Combustion.   These  two  basic designs  are
similar  and  rely  on  simple  combustion  without  the  aid of   a catalyst  to
oxidize  organics,  essentially to  water and carbon dioxide.   The  basic dif-
ference  between  the two types is that in direct combustion  the  gas stream
contains  organic  gases or  vapors  in  sufficient  concentration to  sustain
combustion  most  or  all  of the time;  in thermal  combustion the  gas  streams
are  lean,  usually well  below the lower  explosive limit  (LEL) for  the par-
ticular  organic  gas  or vapor.  Thus  thermal incinerators require appreciable
auxiliary  fuel  to achieve effective combustion, whereas  direct afterburners
often  require only  a  pilot flame  to  initiate  combustion  and to  sustain
combustion  during periods  when  the  gas  stream  is  lean.  Many direct  in-
cinerators  are open flares.   Where  thermal  incinerators  are  used,  the gas
stream  usually  contains enough  oxygen  to burn  the organic  contaminants.
Gas  streams  vented  to  direct  incinerators are often  too rich  in  organics


(concentrations are above  the  Upper Explosive Limit),  and air must be intro-
duced to initiate  combustion.   A typical thermal incinerator  is  illustrated
in Figure 4.6-1.  Catalytic Combustion.  This type  of incinerator uses catalysts
to initiate and promote  oxidation at temperatures well  below those required
for  thermal  incinerators.   The combustibles-laden  gas  stream is  preheated
and passed through a  catalyst bed to oxidize  vapor  phase organics, predomi-
nantly to carbon dioxide and water vapor.
     The combustible  contaminant  concentration  must be below the  lower ex-
plosive  limit.   Commonly, catalysts  are metals  of  the platinum  family and
exist as a  thin coating on an  inert  support material.   Catalytic combustion
is  not  normally  recommended  for  organic  particulate  removal  because  the
surface of the  catalyst  can become coated with particulate matter and there-
by  inhibit  the oxidation reaction.  This  type  of  incinerator  is  not  dis-
cussed further.
4.6.2  Operating Principles of  Incinerators
     Combustion  involves  many  complex,  interrelated  reaction  mechanisms
between  the  fuel,   fuel  decomposition intermediates,  and oxygen.   Depending
on reaction conditions,  results can include  partially  oxidized  species  such
as  carbon  monoxide,  aldehydes,  and organic acids,  or  simply carbon dioxide
and  water  vapor.   The   latter occurs  only when  the  combustion  processes
approach completion.
     As  with   any  combustion  process,   the  basic variables  for  particulate
matter  incineration  are  reaction  temperature,  reaction time, and reactant
mixing  (turbulence).   For solid particles,  the  reaction zone is  confined to
the  surface.    At  low  temperatures the combustion rate  is  limited  by the
chemical reaction  rate,  whereas at higher temperatures  the chemical reaction
rate  is so rapid  that the  rate  of  air supply  to  the  surface  controls the
combustion rate.2'3
     Combustion of liquid droplets and  volatile  solids  occurs away from the
surface  of  the particle, and combustion rate may be dependent on the rate of
heat  transfer  to the surface,  which  causes  evaporation and thermal decompo-
sition  of  the   solid.   Combustion  is influenced by the gas velocity, the rate
of mixing, and  the supply of  oxygen.4'5


                                           FLAME SENSOK-



                         •Till. MILL
                                    UHITIZID IMHIHtnM
                                 SAMPLE PORT

                           TEMPERATURE SENSOR

Figure 4.6-1.   Typical  thermal  incinerator.

     The  temperature  in  the  combustion  zone  surrounding the  particulate
matter may  exceed  the temperature at the interior of the particle and in the
surrounding  gas by  several hundred degrees.   Heat  transfer  is  largely  by
radiation from  the incandescent  surface of the particle,  or from the incan-
descent  carbon  formed as  an intermediate step  in the  combustion process.6   Reaction Temperature.   The  principal!  requirement  regarding
temperature  is  that  the   auto-ignition  temperature  of  all  species  being
burned must be exceeded by approximately  100° to 2!00°C.  This  allows  for a
margin of error to account for nonideal combustion  conditions,  heat losses,
and unknown  particle composition.   Operation at less  than  the auto-ignition
temperature  means  that combustion reactions  are not initiated.  Instead, the
particles  are  simply  being  heated,  with  possibly  some  volatilization.
Emissions  at the  stack may not  exhibit  any  noticeable  opacity;  however,
downwind  the vapors  may  recondense as  secondary particulate  matter.   The
auto-ignition  temperatures  of  selected 'organic  compounds  are  presented  in
Table 4.6-1.7
     The reaction  temperature  also influences the rate of the combustion re-
actions.   Most  direct flame burners operate  in the 650° to 820°C temperature
range to obtain maximum combustion within the  limits  of flame contact, mix-
ing, and residence time in  the furnace.8
     Figure  4.6-2  illustrates  the  effect  of air  velocity  and particle dia-
meter  on the  combustion  rate  of  carbon.9'10'11   The  effects  of  particle
size, reaction  temperature, combustion  gas composition, and gas  velocity  on
the combustion  rate  of  carbon,  coal,  and several other compounds have been
investigated.2"5'12'14   Reaction Time.   Preheat (induction) and  combustion times will
dictate  the overall  residence time of  the particulate matter  in the incin-
erator.   The  residence  time  requirement will   determine  both  combustion
chamber dimensions and particle penetration.
     The time  required to  heat the waste  gas  to  peak furnace temperature  is
                                                    |!            '   •     ,
dependent  on  the  burner  combustion intensity  and  inlet  gas  temperature.
Values of  combustion  intensity  will  vary from  400 kJ  per cubic meter per
second for   low-pressure  gas jet  mixers  to  20,000 kJ  per cubic meter per
second for  premix  mechanical burners.   A typical  value  is  5600 kJ per cubic
meter  per   second  for premix  high-pressure   gas  jet  multiple-port  burners.

           Organic compounds
temperature, oK
         Butyl alcohol
         Carbon disulfide
         Carbon monoxide
         Dibutyl phthalate
         Ethyl ether
         Methyl ether
         Ethyl acetate
         Ethyl alcohol
         Ethyl benzene
         Ethyl chloride
         Ethylene dichloride
         Ethylene glycol
         Ethylene oxide
         Furfural alcohol
         Hydrogen cyanide
         Hydrogen sulfide
         Maleic anhydride
         Methyl alcohol
         Methyl ethyl ketone
         Mineral spirits
         Petroleum naphtha
         Oleic acid
         Phthalic anhydride
         Vi nyl acetate

        -£  1000


        <->  -r
        <  <
        UJ  ,5
        01  Jl

        <  .T
        U  .
        u. <
        u. a:
        UJ  .
        HI Oi
 1000   1200   1400   1600   1800    2000

Figure  4.6-2.   Effect of air velocity and particle diameter
              on  the combusiton rate  of carbon.
                   (D  = particle diameter).

     Combustion time required  is  dependent on particle size,  oxygen  content
of  the  furnace  atmosphere,  furnace  temperature,  particle composition,  gas
velocity,  and mixing  of combustibles.   Combustion  times  for  several  dif-
ferent  materials  have  been  determined and  correlated on  the basis of  the
following equations:15

          *d = CpR'Tmdpo) (96 * D pg}                             (Eq' 4'6"1}
          tc = (pdpQ) (2Kspg)                                     (Eq. 4.6-2)
        t. = diffusion-controlled combustion time, sec
        t  = chemical reaction rate-controlled combustion time, sec
        p  = density of carbon residue or coke, g/cm3
        R' = universal gas law constant, 82.06 atm cm3/mole/K
        T  = mean temperature of stagnant gas film, K
        x  = original diameter of particle, cm
          = combustion mechanism = 1 for CO and 2 for C02 formation
        D  = diffusion coefficient of oxygen at temperature T  , gm/cm2
        p  = partial pressure of oxygen in combustion air, atm
        K  = surface reation rate coefficient, g/cm2 sec atm.
K   may  be calculated  by Equation 4.6-3  for soot and by  Equation  4.6-4 for
coke and carbon residue.
          Ks = (1.085 x 10*Ts ^)(e •"»•""" "'s)                    (Eq. 4.6-3)
          K  = 8710 e"35'700/RTs
          Ks   871° e          S                                  (Eq. 4.6-4)
          T  = surface temperature.
     Equation 4.6-1  holds at high temperature, zero  gas  velocity,  and large
particle  sizes.   The equation can be  corrected for the effects of gas velo-
city  and turbulence by  use of  the  dimensionless  Nusselt  conventional  heat
transfer  relationship for spherical particles:16

          NNu = 2 + 0.68 [Npr 1/3 x NRe %]                       (Eq. 4.6-5)
        ^Nu = Nusselt Number
        Npr - Prandtl Number, a function of the physical properties of the

        NO  = Particle Reynolds Number, a function of the physical proper-
              ties of the gas, particle diameter, and gas velocity.

     The Nusselt  Number  NN  = h (d /k) =  2 at zero gas velocity,  where h =
convectional heat  transfer coefficient (cal/cm2 °C sec); d   = particle dia-
meter  (cm).   The  inverse  function,  h, of the stagnant gas  film thickness,
d_/2,  surrounding  the particle  and  directly proportional  to the  thermal
conductivity of the furnace atmosphere, k (cal/cm2 °C cm sec).
     The  film thickness  decreases with  increasing velocity  and  decreasing
particle size to  such an extent that the combustion rate for particles smal-
ler than 100  micrometers is limited only  by  chemical  kinetics at normal in-
cinerator temperatures.
     Equation 4.6-1 1s of  limited value in design because  particles larger
than 100 micrometers  are easily collected by  other gas cleaning devices and
would require excessive retention time and furnace volume.
     Equation 4.6-2 holds  for  particle sizes  smaller than  100  micrometers
and for temperatures  at  which the combustion  rate  is  determined by chemical
     Total  combustion time  for  a carbon  residue-forming  particle  then be-
          tr = t. -H td •  Kv + tc                                 (Eq. 4.6-6)
          t  = total residence time, sec
          t. = induction time, sec
          K  = volatile matter correction factor determined by the equation:
          Ky = (1 + E/100)/(l + E/100 - V/100)                   (Eq. 4.6-7)

          E  = percent excess air
          V  = percent volatile matter
     The  combustion  time for  hydrocarbon  liquid  droplets  larger than  30
micrometers  at  zero  gas  velocity  may be computed using the  following equa-
          td =  29,800/Pg  MwT"1'75  dpQ2                       (Eq. 4.6-8)

          M  = molecular weight
          T  = furnace temperature, K.
     The combustion  time  of particles smaller than 30 micrometers is depend-
ent on the combustion rate of the hydrocarbon vapors.
     The time  required  to burn a 5 x 10 6 cm soot particle of 2 grams/cubic
centimeter  density  in  a  furnace  atmosphere  containing 0.20  atmosphere  of
oxygen  at  1260°C can be  computed  using the Equations 4.6-2  and 4.6-3.   The
time required would be 0.51 second.
     Total residence  time in the furnace, including heat-up time from 100°C,
would be induction time + combustion time = total time, or:
         tr = 0.208 + 0.510 = 0.718 second
     In practice,  minimum gas furnace retention time is about 0.30 second at
a temperature  of 920°C.   Particle retention time may be increased by design-
ing- the combustion chamber in the shape of a cyclone with a small tangential
inlet, and by introducing the gases at a high velocity.   Heat Transfer.   The transfer  of  heat from  burner  flame  to
gaseous  and parti cul ate  matter  is  an  important  factor in  determining the
furnace size,  operating temperatures, and fuel  requirements  of direct flame
contact incinerators.   Heat  transfer is best achieved  by  mixing when gases
are  burned,  and best  achieved by  radiant heat  transfer  when parti cul ate
matter is burned.16'17
     For  purposes of burning parti cul ate matter,  radiant  heat transfer and
furnace temperature  uniformity may be increased by increasing the emissivity
of the  burner  flame.  This can be accomplished by limiting the air supply to


produce a  sooty flame,  by using high carbon-to-hydrogen ratio  (C/H)  fuels,
by adding soot, or by using fuel oil (by carburetion).
4.6.3  Design of Incinerator^
     Factors  that  must  be considered  in  incinerator design for  particle-
containing  gaseous  waste  include  fuel  requirements,  burner  selection,  pro-
tection  systems,  heat   recovery,   refractory  type,  and   instrumentation.
Design methodologies must take into account the complex interdependences of
operating   parameters  and the  highly  variable  characteristics   of  many
sources. Semi-empirical  approaches  based on previous  experience  with  analo-
gous sources are generally used.7   Fuel Requirements.   Fuel  requirements  and burner  capacity may
be  determined by means  of a  heat  balance, using the heat of combustion of
the fuel and  the sensible heat needed  to  raise the temperature  of  the waste
gas and the products of combustion up to the desired combustion  temperature.
The heating value  of the contaminant must  be  deducted to  determine net fuel
requirements.8'18 20
     Insurance  underwriters  ususally limit the heating value  of  the  waste
gas stream  to combustible  vapor concentrations of  less than one-fourth of
the lower explosive  limit.   For organics this  is  equivalent  to  about 480 kJ
per  standard  cubic  meter.    The  combustible  particulate  matter  normally
contributes only a  negligible  heating  value because of the  low  grain load-
     Figure 4.6-3 presents the energy requirements for a thermal  incinerator
at  various  influent  gas  stream coloric loads.  Similar  curves  applicable to
a  given  facility could  be generated by  use of the  procedures  described in
Reference 2.
     Energy  requirements  are   inversely  proportional  to  the  influent  gas
temperature—the  higher   the  temperature,   the  less  the sensible  heat  that
must be  added to raise  the  influent gas stream to  combustion temperature.
                                                                        i   Heat  Recovery.    Energy   requirements  may   be  substantially
reduced  by  use of  heat recovery  equipment.   The  additional   capital  and
maintenance cost must be weighed against the energy savings.

                         GAS FLOW,  scfm

Figure 4,6-3.   Fuel required to oxidike different concentrations
        of combustible vapor (heat content of combustible
                particulate assurance negligible).

     Heat  recovery equipment used to  recover  heat from the flue  gas  may be
grouped into  two  classifications:   recuperative and regenerative.   Recupera-
tive  heat exchangers,  which recover  heat  on a  continuous  basis,  include
crosscurrent-flow,  countercurrent-flow,  and cocurrent-flow  heat exchangers.
For a  given  heat  flow and temperature drop,  heat exchanger surface require-
ments  will be  the  least  in the  countercurrent-flow  heat exchanger.   The
crosscurrent-flow,  U-shaped  recuperator shown  in Figure 4.6-4  is obviously
subject to fouling  if  combustion  effectiveness decreases to  the  point that
large particles remain  in  the effluent.   Cold-side  deposits may also occur.
     Cocurrent-flow  heat exchangers  are  often used where a moderate level of
heat  recovery is   required.   The cost of countercurrent-flow  heat exchanger
construction  may  be greater  than  that for  cocurrent-flow  because operation
at lower  temperatures  (near  the dewpoint) may  require  use  of  special  alloys
or alloy steels.
     Regenerative  heat  exchangers  recover  heat by intermittent heat exchange
through alternate  heating and  cooling of a solid.   Heat  flows alternately
into  and  out of  the same exchanger as  air  and flue gas flows  are periodi-
cally  reversed.   Regenerative heat  exchangers are of  fixed-  and  moving-bed
types.                                                                    ]
     A fixed-bed,  pebble-stove,  regenerative afterburner is shown in  Figure
4,6-5.  When  gas  is passed  through  the  pair  of  pebble-type  regenerators
connected  back  to  back,  the  gas is heated on the upstream side and cooled on
the downstream  side.  When  the upstream bed  and gas  temperature drop,  gas
flow  is   reversed  and  the  heat transfer process is repeated.   Particulate
matter is effectively retained  and incinerated.   Heat  recovery  efficiencies
in excess of 95 percent can be achieved.21
     A commonly used rotary  regenerative heat exchanger consists of a parti-
tioned rotating cylinder containing a heat sink and  heat  transfer surfkce
area.   The cylinder  is  partitioned  along its axis by appropriate  gas  seals,
so  that   hot  flue  gas  and  cold waste  gas  may be  passed  through  the  heat
exchanger  on  opposite  sides  of  the cylinder.   Heat  is  absorbed from the hot
flue  gas   by  the  heat  exchanger surface and  transferred by  the .continuous
rotation  of the heat exchange surface to the  cold waste gas side, where the
heat  is   absorbed  by  the  incoming cold  gases.   Heat recovery  efficiency
ranges from 85 to 95 percent.21


                             NATURAL GAS
          TO CHIMNEY
                                                         FROM KILN
 Figure  4.6-5.  Fixed-bed, pebble-stone,  regenerative afterburner.
                         780 TO 840 K
                                      WASTE GAS INLET
                                      ,330 TO 360 K
i 950 TO

980 "K
                                                          WITH TUBULAR
              Figure 4.6-4.   Tubular recuperator.

     4.6,3.3  Hood and  Ducts.   Furnace inlet gases and vapors from paint and
varnish cooking  kettles  and  other sources must be maintained at temperatures
above  condensation  to  avoid  exhaust duct  fouling.   Collection ductwork  is
usually insulated and may be heated by means of an external  duct that serves
to recover heat from the flue gas, which reduces fuel requirements.
     Duct gas velocities  are  usually high, ranging  from  1000 to 1700 m/min,
to prevent  the settling  of  particulate  matter,  to effect  a high heat  re-
covery  rate  between  the flue gas and  furnace feed gas, and  to  minimize  the
danger of flashback and fire hazards.7*22'23
     Other safety  devices for  minimizing fire hazards may  include diluting
vapors  to  below the lower explosive limits, using flame arresters,  and  in-
cluding  a  wet  scrubber  between  the  direct flame  combustor and  the  vapor
source.  Dilution  of the  vapors  may be  accomplished fay recirculation  of a
portion of the  flue  gas, which would substantially reduce fuel  requirements.
     Flame  arrestors  may  consist of a  packed bed  of pebbles, metal  tower
packing, aluminum rings  (Figure 4.6-6),  or corrugated metal  gridwork (Figure
4.6-7),  in  conjunction  with  a blast gate or other  pressure-release  device.
Flashback through  the bed is prevented  by  bed gas velocities,  by pressure
drop, and by cooling the flame to below combustion temperatures.24*25
     Other types  of flame arrestors include spray chambers, wet  seals,  and
dip  legs.  Wet  flame  arrestors have the  disadvantage  of  cooling and humidi-
fying the exhaust  gas with a consequent  increase  in fuel requirements.   Wet
sprays are capable of partial removal of the solids.
     Flame arrestors,  regardless of type,  should not be relied  upon  as  the
primary  defense against  explosions.7   The combustibles  content of the  gas
stream  should be  continuously kept at less  than  25 percent of the lower ex-
plosive  limits, with consideration  given to  known fluctuations  in  process
operation.   Burners.    The  burner,  the  key  operational   component of  an
incinerator  system,  is  based  on  the type of service  and capacity required.
     Three  major types  of gas  burners  are available:7   raw  gas burners,
premix  gas burners,  and forced-draft gas burners.   The raw gas burner relies
on an  induced  draft  of air to  mix  combustion air with the  gas.   This burner
consists of  a  cluster  or ring of holes for the raw gas generation jets.   The
premix  burner   is  fed  an already  proportioned fuel/air  mixture, which  is


                                         WASTE GAS
                                    BLAST GATE

        Figure 4.6-6.  Packed-bed flame arrester.
                                     TUBE BANK

Figure 4.6-7.   Corrugated metal flame  arrester with cone
     removed  and tube bank pulled partly off the body.

Ignited at  the burner  orifice or  ring.   Obviously, care  must be taken  to
avoid flashback.  The  forced-draft burner involves separate delivery of fuel
and combustion  air  to  the burner.   After mixing, the air and gas are ignited
by a pilot.
     Oil  burners  are similar  to the  raw gas and  forced-draft gas  burners.
The  principal  difference  is  the   addition  of  an  oil  atomizing system  to
ensure  intimate mixing  of the  fuel  with air.   A typical  forced-draft  oil
burner  is shown in Figure 4.6-8.   In  this model, compressed air  is  used  to
atomize the  oil.   The  combustion  air could be  either  contaminated  effluent
or ambient air.   Construction Materials.    Incinerator  surfaces   that  will   be
exposed to  high  temperatures and  erosive or  corrosive conditions must  be
constructed of alloys  capable of  withstanding high  temperatures  or  must  be
lined with refractory materials.   Metals.   Underwriters Laboratories,  Inc.,  limits temperatures
at which  alloy steels  are used to  approximately 100°G  below the temperature
at which  scale formation occurs.   Martensitic and  ferrjtic  stainless  steels
are recommended for  use in areas that are exposed to wide ranges of tempera-
ture  and  to  corrosive  conditions.26   Temperature  limitations  for  other
metals and alloys are determined by design stress and safety requirements.20   Refractories.   Refractories used  in  direct-flame incinerators
               "" V:::                   .  •• „  ' .,"              .1" •..
increase  radiant heat  transfer,  act as a  support structure, and are resist-
ant to  abrasion and corrosipn.   They also  must be  capable of withstanding
thermal  shock.  Fire clay refractories  are  commonly  used in  incinerator
construction  because  of low  cost,  spall resistance, and  long  service life.
Fire  clay refractory bricks  are classified (Table 4.6-2)  into maximum ser-
vice  classes  according  to American Society for  Testing and Materials (ASTM)
standards.27   Their  softening  points,   as  determined by  pyrometric  cone
equivalent  (PCE),  help  determine  their maximum  service  class.28   Other
requirements  include limits  on  shrinkage,  spelling  loss,  and  deformation
under load.    Commonly  used castable fire clay refractories (Table 4.6-3)  are
of two ASTM classes.29                                                      ,

               Figure 4.6-8.   Typical  forced draft oil  burner.
Refractory type
Low heat duty
Intermediate heat duty
High heat duty
Super heat duty

Temperature, °C
Special properties
Insulating, light-weight
General -purpose
     Service temperature ranges  of  various refractories  for corrosive condi-
tions are shown  in  Figure  4.6-9 and in Table 4.6-4.   The literature contains
further information.27'28   Instrumentation.   Minimum  instrumentation for an  incinerator
system  consists   of a  temperature  indicator that  enables  an  operator  to
determine if gas  temperature  is too low  for  effective oxidation  of particu-
late matter or too high for the materials of construction.

               U- 3500
               5 3000

                     TABLE  4.6-4.
Type of brick

High-duty fireclay

Super-duty fire-

(type H),

Insulating brick


Extra-high alumina




Zi rcon

SiO, 95%

SiO, 542
Al20a 4035
Si Oj 52%

AljCh 422
Si 02 59%
Al20a 34S


AUO, 50-
A1203 90-

A1203 7 «

ore 100X

MgO 50-

CR2°3 5-
F«20 3-
Al,0, 6-
J in
SiO. 1.2-

MgO 95%
ZrOg 57%
Si 02 33*
2r02 941
CaO 4%
SiC 80-
C 97%






























under hot













Excellent -















16 ''












Very low
Very low





























Chemical resistance
to acid



Insoluble in acids
except HF and
boiling phos-

Good except for HF
and aqua regia.

Insoluble in most
Fair to good

Fair except to
strong acids.

Soluble in most

Very slight
Very slight
Slight reaction
with HF.
to alkali
Good at low temp-
Good at low temp-

Good at low temp-

Very resistant in
moderate con-


Very slight at-
tack with hot
Slight reaction


Fair resistance
low tempera-
tures .

Good resistance
low tempera-
tures .
Very slight
Very slight
Attacked at
aGood above 650°C.
 Chemically bonded,
Dissociates above 1700°c,

      The   addition   of a  temperature  controller  provides  process  control
 capability not  available  with  only  a passive  temperature  indicator.   The
 controller is a feedback system that  adjusts fuel  input to follow changes in
 gas  temperature.7  In extreme situations, the fuel  input is entirely shut off
 to  protect  the incinerator.   The controller  thereby provides a  degree of
 protection and a means of continual  adjustment to match influent character-
      A  flame-sensing device  is  useful for rapid  shutdown  of fuel  supply in
 case  of flame  failure.   This device  reduces the potential for accumulation
 of  explosive  gases  within   the  incinerator during  the  flame outage.   It
 should  observe  both  the burner and the pilot flame.
 4.6.4  Operation and Maintenance of Incinerators
      Proper  operation of  an incinerator system  requires control  of contami-
 nant  quantity  and  characteristics and  requires  regular maintenance  of the
 burners. As  with other particulate control devices, complete instrumentation
 and an  effective preventive maintenance program are necessary.   Burners.   Burners  are  high-maintenance items  because  of the
 high  temperatures,  refractory, and  small  orifices.7   When the contaminated
 gas  stream is  used  as the  combustion air, fouling  of the  orifices  and/or
 deposits in  the air delivery lines can occur.   Impurities  within the oil can
 lead  to similar problems.   A  second  problem is improper sizing of  the  bur-
 ner(s).   This  can   lead  to  low  gas  temperatures  resulting  in  incomplete
 oxidation  of  particulate  matter.   Burners  with  poorly   adjusted  air-fuel
 ratios  can  generate soot,  which  fouls  downstream heat exchange  surfaces.
      Minimizing  of  burner problems  is facilitated by providing a  means of
 visually checking  the flame  for proper  luminosity,   length,  and  stability.
 Also,  an  adequate   inventory  of  spare  parts   should  be  kept.   If  fouling
 continues,   a precleaner may  be economical.  Finally, such  incinerator in-
 strumentation  as  the flame  sensor and the temperature controller  should be
 checked regularly.   Effluent Characteristics.   Variability   of   effluent  quantity
and heat content should be  minimized by controlling  the process  operation.
Excess  concentrations of  combustible gases and  vapors  can  lead  to  high
temperature  excursions,  which  damage the  incinerator refractory or  shell.


High gas  flow  rates  lead to poor particle oxidation resulting from decreased
residence  time  and  decreased  reaction  temperature.   Undesirable  reaction
products  such  as carbon  monoxide,  aldehydes, and organic acids  also  result
from poor combustion.
     Contaminants containing  sulfur  or chlorine compounds may be oxidized to
highly corrosive  species  such as hydrochloric acid  vapors and sulfuric acid
vapors.   These  could attack  either  the  refractory  or the shell  of the in-
cinerator.27   Precleaning or special materials of construction  are required
when these contaminants are present.


 1.  U.S.  Environmental  Protection Agency.  Control  Techniques for Volatile
     Organic Emissions from Stationary Sources.  Evaluation draft.  February

 2.  Kobayashi,  K.   Combustion of  a Fuel Droplet.   In:   Fifth Symposium on
     Combustion,  Pittsburgh.   Reinhold  Publishing Co.,  New York,  Pub.  No.
     1955, 1954.

 3.  Spaulding,  D.  B.   Heat  and  Mass Transfer in the  Combustion of Liquid
     Fuels.   American  Society  of  Mechanical   Engineers.   Published  by  the
     Institution  of Mechanical   Engineers,  Story's  Gale,  London,  England,
     September 11-13, 1951.

 4.  Hottel, H.  C,,  G.  C.  Willians, and  H.  C.  Simpson.   Combustion of Drop-
     lets  of  Heavy Liquids  Fuels.   In:   Fifth  Symposium on  Combustion,
     Pittsburgh.   Reinhold Publishing Co.,  New York,  Pub.  No.  1955, 1954.

 5.  Nlshiwaki,  N.   Kinetics  of  Liquid  Combustion  Processes.   Evaporation
     and  Ignition  Lag  of  Fuel Droplets.   In:  Fifth Symposium on Combustion,
     Pittsburgh.   Reinhold Publishing Co.,  New York,  Pub.  No.  1955, 1954.

 6,  Orning,  A.  A.   Combustion  of Pulverized  Fuel—Mechanism and  Rate of
     Combustion  of  Low  Density  Fractions  of  Certain  Bituminous  Coals.
     Trans. Am. Soc. Mech.  Eng., Vol. 64, 1942.

 7,  Ross,   R.    Incinerators,   An  Operation and Maintenance of Air Pollu-
     Control Equipment.  Technomic Publishing,  1977.

 8.  Danielson, J.  A.  (ed.)  Air Pollution  Engineering  Manual.   U.S.  Public
     Health  Service, National Center for Air  Pollution  Control,  Cincinnati,
     Ohio.  PHS Pub-999-AP-40, 1967.

 9.  Yagi,  S,,  and D. Kunii.   Combustion of Carbon  Particles  in Flames  and
     Fluidized Beds.   In:   Fifth  Symposium on Combustion, Pittsburgh.  Rein-
     hold Publishing Company, New York, Pub. No. 1954.

10.  Browning, J.  A.,  T.  L. Tyler,  and  W.  G.   Kran   Effect  of  Particle Size
     on  Combustion  of  Uniform Suspensions.  Ind.  Eng.  Chem.,  49(1):142-147,
     January 1957.

11.  Godsave,  G.  A.   E.    Studies  of  the  Combustion  of  Drops  in  a  Fuel
     Spray—The  Burning  of Single  Drops  of Fuel.   In:    Proceedings  of  the
     4th  Symposium on  Combustion,  Cambridge,  Massachusetts.  Williams  and
     Wilkins Co., Baltimore, Maryland.   Pub. No. 1953, September 1952.


12.  Gregory, C. A.,  Jr.,  and H.  F. Calcote.  Combustion Studies of Droplet-
     Vapor  Systems.   In:   Proceedings  of  the  4th Symposium  on Combustion,
     Cambridge,  Massachusetts.   Williams  and  Wilkins  Company,  Baltimore,
     Maryland.  Pub.  No. 1953, September 1952.

13.  Smith, D.  F., and A.  Gudmundsen.  Mechanism of Combustion of Individual
     Particles  of Solid Fuels.  Ind. Eng.  Chem.,  23(3):277-285, March 1931.

14.  Parker,  A.  S. ,  and H. C. Hottel.  Combustion Rate of Carbon.  Ind. Eng.
     Chem., 28(11):1334-1341,  November 1936.

15.  Field, M.  A., D.  W. Gill, B. B. Morgan, and P. G. W. Hawksley.  Combus-
     tion  of Pulverized Coal.   British  Coal Utilization  Research Associa-
     tion, Leatherhead,  England, 1967.

16.  Eckert,  E.  R.  G.,  and  R. M.  Drake, Jr.  Heat and  Mass Transfer.  2nd
     ed.  McGraw-Hill, New York, 1959.

17.  Topper,  L.   Radiant Heat Transfer From  Flames  in a Turbojet Combustor.
     Ind. Eng.  Chem., 46(12):2551-2558, December 1954.

18.  Goodel,  P.  H.   Industrial Ovens Designed for Air Pollution Control.  J.
     Air  Pollution  Control  Assoc.,  Vol.  10, pp.  234-238,  1960.  (Presented
     at  the  52d Annual Meeting of the Air Pollution Control Association, Los
     Angeles, California, July 22-26, 1959.)

19.  Vandaveer,  F. E. ,  and C. G.  Sedeler.   Combustion of Gas.  American Gas
     Assoc.,  Inc., New York,  1965.

20.  Ingels,  R.  M.   The Afterburner Route  to  Pollution  Control.   Air Eng.,
     No. 6, pp. 39-42, June  1964.

21.  Perry,  R.  H. ,  C.  H. Chilton, and S.  D.  Kirkpatrick  Chemical Engineers'
     Handbook.  4th edition.   McGraw-Hill,  1963.

22.  Sly  Manufacturing Company.   Industrial  Air  Pollution  Control.    Cleve-
     land, Ohio.  Bulletin 204, 1967.

23.  Weil,  S.  A.    Burning  Velocities   of  Hydrocarbon  Flames.   Inst.  Gas
     Tech., Chicago, Illinois.  Research  Bulletin  30,  1961.

24.  National Board  of Fire  Underwriters.  Standards  for Ovens  and Furnaces.
     NBFU No. 864, August 1963.

25.  Radier,  H.  H.   Flame Arresters.  J.  Inst.  Petr., Vol. 25, pp. 377-381,

26.  Underwriters  Laboratories.    Standards  for  Safety—Commerical—Indus-
     trial  Gas  Heating Equipment--UL.  795.  November  1952,  Revised 1960.

27.  Trinks,  W. ,  and M. H. Mawhinney.  Industrial Furnaces, 5th ed., Vol. I.
     John Wiley and Sons,  1961.


28.  Burst,  J.   F.,  and  J.  A.  Spieckerman    A Guide  to Selecting  Modern
     Refractories.  Chera. Eng., 74(16):85-104, July 1967.

29.  Griffiths,  J.  C.,  and  E.  V. Weber.   Influence of  Port  Design and Gas
     Composition  on  Flame Characteristics of  Atmospheric Burners,   American
     Gas Association  Laboratories, Cleveland,  Ohio.   Research  Bulletin 77,

                                 SECTION 5
                         FUGITIVE EMISSION CONTROL

     The failure  of many  areas  to attain the National  Ambient  Air Quality
Standards  for  particulate  matter has prompted reexamination  of  the partic-
ulate  problem.   Conventional  stationary  emission  sources  have been  con-
trolled, but in many cases, fugitive emissions constitute a large percentage
of total emissions;1 thus attainment of the standards for particulate matter
necessitates the control of these fugitive emissions.

     Particulate that becomes airborne is either industrial process fugitive
emissions or fugitive dust.
     Industrial  process fugitive  emissions  include  particulates  that are
emitted  from  industry-related operations and that escape  to  the atmosphere
through  windows,  doors, vents,  and the like rather than  through a primary
exhaust  system such  as a stack,  a flue, or  a control  system.   Fugitive
emissions may escape to the atmosphere from indoor manufacturing operations,
materials  handling,  transfer, and  storage  operations,  and other industrial
processes.   Sometimes they are emitted more directly to the atmosphere from
out-of-doors industrial  processes  such as coke ovens, quarry rock crushing,
and  sandblasting.   Fugitive  emissions  can result from poor  maintenance of
process equipment or from the operation of processes without concern for the
environment;  for example,  they can result  from  leakage around  coke  oven
doors that cannot be properly sealed because of warpage.
     Fugitive dust is generally either natural or man-associated particulate
that becomes  airborne  by  the wind, man's activity, or both.   Examples are
windblown particulate from unpaved dirt roads, tilled farmlands, and exposed
surface areas at construction sites and from duststorms.

     Table  5-1  lists sources  of industrial process  fugitive  emissions and
the broad range of controls applicable to these emissions.
     Table  5-2  lists sources  of fugitive  dust  ("native soil  that  becomes
airborne").  Although some  sources  are in  the area  of processes (for exam-
ple, surface mines), they can be differentiated;  overburden removal and haul
roads  at surface  mines are  fugitive  dust sources,  whereas  removal  of  a
product (e.g., coal) is a fugitive emission source.

     As  shown  in Table  5-1,  the following techniques are  used to  control
process fugitive emissions:   ventilation systems,  process optimization, and
wet suppression.  Although some processes have been controlled by all three,
others are controllable by only one or two of these techniques.
     Selection  of  the  proper  control  technique requires  consideration  of
factors  such  as the industrial processing  facility,  the  characteristics  of
the exhaust stream, and the secondary multimedia impacts.   Selection must be
     Ease of  control  varies with the age and basic design of a facility.   A
fugitive  emission  control  system for a new plant can  be  integrally incor-
porated into the overall design of the plant, whereas a retrofit application
requires  that  the  system  be  adapted to the configurations  of the existing
plant.  The  retrofit must  be built  within fixed-space limitations  without
interfering with operation  of the process.  In  general, the more congested
the plant  layout,  the harder it is  to  retrofit  most fugitive emission con-
trol systems.
     The  location  of the  facility  and the  fugitive  emission  source within
the facility  can also  affect the selection of  the  control  technique.   For
example, controls required for a storage pile of fine material  near a public
road could differ from those for a storage pile well  within the plant bound-
aries,                                                                   i
     Product quality also affects control technique selection.   For example,
dry collection of fugitive emission may be needed so that it can be returned
to the  product  stream.   Also, wet suppression may have limited use because
it has an adverse impact on product quality.

Material handling
Transf erri ng/convey i ng
Loach" ng/unl oadi ng
Baggi ng/packagi ng
Material crushing and screening
Mineral mining
Waste disposal (tailings)
Metallurgical operations
Furnace charging
Furnace tapping (product and
Mold preparation
Casting shakeout
Slag disposal
Coke oven charging
Coking (leaks)
Coke pushing
Coke quenching
and collection


















                  CONTROL TECHNIQUES2'3
Unpaved roads
Dust from
paved roads
Off-road motor
removal /storage
tailings piles
Disturbed soil
ti 11 i ng
Wet sup-
x .









ati ng





     The  major  exhaust stream  characteristics that  collectively  influence
selection  of the  control  technique  include  particle size  distribution;
temperature, moisture content, presence of corrosive gases; and physical and
chemical characteristics and associated toxicity of the particulate.
     The  size of  particulates  from many  metallurgical   fugitive  emission
sources  is  predominantly  below  5  jjmA,  which in the  case  of  add-on  control
systems  often dictates the  need  for  a fabric  filter.4  Exceptions  are
sources  in  which the  particulates have  a relatively large  diameter,  which
sometimes can be sufficiently controlled by high efficiency cyclones.
     Since  most  process  fugitive emission  exhaust  streams  are  at either
in-plant  or ambient temperatures, provisions  such  as heat-resistant fabric
filter material  for excess temperatures are generally not required.  In some
applications, however, hoods near furnaces must be water-cooled to withstand
initial  high temperatures.   Because  most fugitive emissions  have approxi-
mately  the  same  moisture content as the ambient  or  in-plant  air, little
insulation  or reheat  is  generally needed for protection  against condense-%
     Physical  and  chemical  characteristics  of  fugitive  emissions  also
influence  selection  of the  control  technique.  The  most  critical physical
characteristics  that  relate to  the  type of control  and to  the  material  of
construction  are abrasiveness  (related  to  particle size  and morphology),
hygroscopy,  and  true  density; the most  critical  chemical  characteristic  is
     Control  of  fugitive  emissions can create secondary multimedia environ-
mental  effects,  which also  must be  controlled.   These  secondary  effects
include  solid waste   disposal,  water pollution,  generation  of additional
fugitive  emissions, and noise.   If not controlled, problems associated with
poor procedures for disposal of the fugitive emissions  collected in a fabric
filter  (when  return to the processing system is not feasible) may exceed the
original  problem.   An example would be  dumping collected  materials  into  an
open  truck, hauling them in  an open truck to a  landfill,  and dumping them
into  a  landfill  that  is  not  adequately protected  from wind  erosion and
surface  runoff.   Thus  adequate precautions must be taken in the  selection of
a control technique, to avoid creating new fugitive sources.

5.2.1  Ventilation. Systems    Localized  Hoodj ngand  Enclosure.    Capture  and  ventilation
systems  are  used with  appropriate  particulate  removal  devices  (e.g.,  a
fabric  filter  or wet  scrubber)  to  control  fugitive  emissions.   Sources
amenable to  this  type of control  include  processes  such as materials hand-
ling (conveyors, elevators, feeders, loading and unloading, product bagging,
and  storage  silos);  solids  beneficiation (crushing,  screening,  and other
classifying  operations);  mining (drilling and  crushing),  and others (furn-
aces, dryers).2'4-8
     In  general,  systems of  capture  near the  process  (localized  hoods  or
canopy hoods),  as opposed to ventilation  of an entire building, are desir-
able from  an economic and occupational exposure  standpoint.   For example,
the  use  of  canopy  hoods  to  control  fugitive  emissions from  electric  arc
furnaces can be expected to  require  40  to 50  percent of the  flow rate  re-
quired for  building  evacuation.4   Local  hoods  (nearer  the emission source
than canopy hoods) may require even less air flow for capture.
     The capture  effectiveness  of a  ventilation  system varies greatly  and
depends on many parameters.  The properties of the emission source, location
of the  hood  or  enclosure,  possibilities  of  external  disturbances  such  as
wind and vibrations, and  operator errors  can  all affect  the capture effi-
ciency.   In  general,  however, capture efficiencies of more than 90 percent
can be expected with proper design.  For  example, 90  percent efficiency is
attainable on the enclosure of basic oxygen furnaces.8
     Detailed guidelines for the proper design of ventilation systems can be
obtained  from  several  sources.9-11  The  important  factors in  the proper
design of hooding, enclosures, ducting,  and ventilation are:
     0    Hood design and placement                                      ,
     0    Air  velocity  and  flow direction required  for proper  capture  of
     0    Volume of air required
     0    Minimization of static pressure drop.

     Hoods should be designed to enclose the emission source as much as pos-
sible.   Total enclosure, however, may be limited by the need for easy access
to the  process.   Hoods  must have adequate  flow rates  and face velocities
(air into the  hood)  to capture the particles and to impose mininup pressure
drop on the system.   Sometimes the hood can be located and shaped so that it
is aligned with the general flow direction of the contaminant.
     Air  flow  past the  emission source must  be sufficient to capture the
contaminant, and must be directed to minimize exposure of workers to hazard-
ous fumes.   Proper  air flow rates depend on  the specific properties of the
emission  stream  and the  particulate matter.  Duct  systems must complement
the hoods in efficiency of operation.  Fittings and other components should
be adequate  to prevent  excess pressure drop.   When overall  hood and duct
dimensions are selected, the expected static pressure drop in the system can
be calculated readily by using standard techniques.11  Building  Evacuation.   General  ventilation of an entire build-
ing  is  receiving increased  attention as a  technique of  fugitive emission
control because of the space entailed in capturing emissions with many local
hoods.2  Ductwork is installed on the building roof, and large fans draw the
particulate-laden gases from the building through a collection device.  This
may  require  improvement of  roof supports because of  the  added weight, and
the  installation  of  opening/closing doors  to  minimize  leaks.   A  fabric
filter  is typically  the control  device  selected for  building evacuation.
Overall capture  and  collection efficiency for such a device is estimated to
be from 90 to more than 95 percent.5
     Total building evacuation has its drawbacks, however.  One is the large
air flow  rates  required.   Another is that the evacuation systems sometimes
collect  enough gas  to  sufficiently ventilate  the workplace  overall, but
still  leave  an  unacceptable level of local pollutant concentrations because
of "deadspots" in the air flow pattern.5
     Building  evacuation systems have been  applied  to electric arc furnace
melt shops in the iron  and steel industry and to a converter building in the
copper  industry.5  These  systems  range in  size from  235 to  285  m3/s and

   \  HOPPER
                   Figure 5-1.  Hood design.'
                Figure 5-2,  Hood location.
               GOOD                       BAD
                Figure 5-3.   Air flow direction.

                                  9 m MAX.
60 era MIN.
/    /
                 EXHAUST MINIMUM ADDITIONAL 0.5 nT/s per in. OF
                   BELT WIDTH AT A MAXIMUM OF 9-m INTERVALS.
                 DUCT VELOCITY =1100 m/min MINIMUM.
   Figure 5-4.  Belt conveyor ventilation for fugitive emissions  control.

                                              LOADING POINT
                    45? OR MORE

                                                    .OCATE REMOTE FROM
                                                   LOADING POINT
                      •CONVEYOR TO
                      HOPPER AND  BIN
            TO BIN
                                    LOCATE REMOTE FROM
                                    LOADING POINT
 Figure 5-5.  Hopper and bin chute and conveyor-loading ventilation for
                      fugitive emissions control.'"

                                     HOOD ATTACHED TO BIN
                                        PRINCIPAL  DUST SOURCE
                                        SCALE SUPPORT
Figure 5-6.  Bag-filling fugitive emissions control.

5.2.2  Optimization of Equipment and Operation
     Control  by proper  operation  and  maintenance practices  primarily in-
volves the elimination of fugitive emissions resulting from  process upsets,
leaks, and  poor housekeeping.   In addition, prompt cleanup of spills on the
ground or  floor by  vacuum systems will  prevent  spills  from becoming air-
borne.  A full-time  cleanup crew may be  required  in  some industries.  Also
included in  this  category is the optimization  of  the capture efficiency of
the hooding  systems  of primary source control devices.  Examples include:

     Precautions to ensure that a cupola is not overloaded, to eliminate the
     possibility of  backpressure from the primary control  system and "puf-
     fing" fugitive emissions from the charging door opening
     Maintenance  of  coke  oven  doors  and  seals  to  eliminate  leaks  during
     Prompt  repair  of  electric arc  furnace  hooding  damaged by  overhead
     charging crane
     Conscientious periodic  application  of chemical  suppressant to inactive
     storage piles and tailings areas
     Increase in the  vent rate of a canopy hood  system for an electric arc
     furnace in a gray iron foundry
     Prompt cleanup of spills from trippers of a clinker conveying system in
     a portland cement plant
     Increase in the  vent rate of a primary  control  device to eliminate or
     minimize leaks.
     A change in  the  process or raw materials can be an effective technique
of  reducing fugitive  emissions.   For  example, using  only clean  scrap  in
metal-melting furnaces or removing crankcase oil prior to automobile salvage
can reduce  fugitive  emissions.   Changing the process  (from  cupola  to elec-
tric arc furnace or from bucket elevator to pneumatic conveyor) is an effec-
tive way to minimize fugitive emissions at the source.
     During  the transfer of  dusty  materials from a conveyor  or  stacker  to
another conveyor  or stockpile,  fine  materials can be  separated  from large
materials by wind  and/or the  falling  action  of  the materials.   A  simple
technique for  reducing  dusting is to  shorten the  fall  distance  by using
hinged-boom conveyors, rock  ladders,  telescoping  chutes,  lowering wells,  or


 other devices.12   The hinged-boom  conveyor,  which can  raise or  lower  the
.conveyor belt and  thus  reduce the fall distance  at  the  transfer point,  can
 reduce  emissions  by an  estimated  25 percent.4  Rock  ladders  allow  the
 material to  fall short distances  in  a  steplike fashion;  the  direction of
 travel  on  successive  steps  is  reversed  to  reduce  the momentum  that  the
 material receives  from  the  previous fall  and to  reduce  the resulting dust-
 ing.    This  technique  can  reduce  emissions  by  50  percent.4   Telescoping
 chutes carry the materials  from the discharge point to the receiving point,
 but  the  materials  are  not  exposed.   Estimated  control  efficiencies  of 75
 percent are  possible.   Lowering wells, or perforated pipes, allow materials
 to flow out  of  the pipe above the pile surface; The dusting from the impact
 of the  falling   materials is  retained inside  the pipe,  and the material is
 protected  from  wind  action.    This  technique  can  reduce  emissions  by an
 estimated 80 percent.4
      Confinement by  covering  or enclosure basically involves the partial or
 complete seclusion  and/or  shielding  of  the source of the  fugitive dust or
 the  industrial,  process  fugitive emissions.   The  design  strategy is to pre-
 vent the fugitive  particulate matter from becoming airborne  by the wind or
 by the  mechanics of  the process.  These control techniques range from small
 enclosures over  conveyor  transfer  points  (for protection from  the wind  and
 from  turbulence  of the moving belt)  to  building structures  (for complete
 confinement  of  materials  storage  areas).   Other examples  include conveyor
 system  enclosures,  weighted tarpaulin covers for inactive  material storage
 piles, partial  windbreaks  in  the prevailing upwind direction from limestone
 quarry  surge pile  areas, and partial open-ended shelters  with shrouds  for
 railroad  car loading  and  unloading.2   The total  enclosure of a transfer
 point can reduce fugitive emissions by an estimated 70 percent.4
 5.2.3  Wet Suppression2'4'5
      Fugitive emissions from  materials   handling and beneficiation  can be
 controlled by spraying  liquid on the materials.  Wet suppression techniques
 include applications  of water, chemicals, and foam.   The point of applica-
 tion  is  most commonly  at the conveyor feed and  discharge  points, but some
 are at conveyor transfer points and equipment intakes.  Wet suppression with
 water only is a  relatively  inexpensive technique; however,  it has the inher-
 ent  disadvantage of  being  short-lived.    Control  with  chemicals  (added to

water  for  improved wetting)  or foam is longer  lasting,  but  more expensive
than water alone.  A wet suppression system is shown in Figure 5-1.
     A wetting agent breaks down the surface tension of water, and allows it
to spread  further,  penetrate deeper, and to wet the small  particles better
than  untreated water.   Mechanical   agitation  of  the  materials  causes  the
small  particles  to  agglomerate.   For effective control, the spray should be
applied  at each point  where the particles  might be  fractured,  allowed to
free fall, or subjected to strong air currents.14
     When  applied to  loading or unloading operations,  wet suppression tech-
niques can  reduce  airborne dust to some extent.   The loading process natur-
ally  disturbs  the  materials,  but  water sprays with  wetting agents  cause
small  dust particles  to  adhere to  larger  pieces so  as  not to  become  en-
trained.    This technique  is  not suitable for many materials  that cannot be
readily wetted.2
     Foam  is effective  in dust suppression because small  particles (1  to 50
pmA in diameter) break the surface of the bubbles  in the foam when they come
in contact with it,  and  these particles become wetted.   (Larger particles
only  move  the bubbles  away.)  The  small   wetted particles  then must  be
brought  together or  brought  in  contact with  larger  particles  to  achieve
agglomeration.   If  foam is injected into free-falling aggregate at a trans-
fer point,  the  mechanical  motion causes the contact of particle with bubble
and subsequent contact of particle with particle.
     Electrostatically  charged fog  is another type  of  wet  suppression sys-
tem.    It differs  from conventional  water sprays,  in -that the droplets carry
a  charge  of  static  electricity.   Because  most fine  particulates  carry  a
natural  electrical  charge,  particle collection  can  be  improved  via  elec-
trostatic attraction if the water spray droplets are charged to the opposite
polarity.   The  charged water  droplets  exert attractive  forces  on  the  op-
positely charged particles,  and each droplet collects  more particles  as it
travels through the dust-laden gas.5'15
     Spray  systems  on material-hand!ing operations  are estimated to reduce
emissions  by 70  to  95 percent.4   Spray  systems  achieve  an estimated 80
percent reduction at  rail car unloading stations  at  iron  and  steel plants.4
     Spray  systems  can  also  reduce  load-in and  wind erosion  emissions from
storage piles  of  various  materials.  Emissions  from load-in  can  be  reduced


                                                                                                           TO KILN
                                                       WATER LINE
                      Figure 5-7.   Wet dust suppression  system applied  to material handling operation
                      Note:  Wet suppression at fine mesh  screens must  be regulated properly to avoid
                                                    blinding  of screens.

by an estimated  70 to 90 percent,2'4*16 and  wind  erosion can be reduced by
an estimated 80 to 90 percent.2'4

     Table  5-2 lists  the major  fugitive dust  sources and  the  applicable
control   techniques.   These  techniques are  wet suppression;  stabilization
(physical, chemical,  and  vegetative);  and specialized methods such as speed
reduction,  street  cleaning,  windbreaks,  industrial  transportation control,
and good operating practices.
5.3.1  Wet Suppression2 >4
     Wet  suppression  of  dust with water or with water plus a wetting agent
(surfactant)  is  an effective  temporary control for  fugitive  dust from un-
paved roads,   cattle  feedlots, stockpiles,  waste  heaps,  mining,  and  con-
struction  activities.   Water  alone is  not  as  effective because  the  high
surface  tension  of  water prevents  it from  wetting, spreading,  and pene-
trating.   The addition of a wetting agent aids water penetration into the
material  and  helps to promote particle agglomeration.17  Wet suppression of
fugitive dust on exposed surfaces such as haul roads is usually accomplished
by spraying from  tank trucks.   Fixed pipeline  spray  systems  have also been
used  on main  haul  roads that  are  relatively  permanent,  such as  those  at
mines and large industrial facilities.
     When  water   is   used  for  wet  suppression,  repeated  applications  are
necessary.  Evaporation  and  runoff  cause its effectiveness to be temporary.
In some regions,  water  usage  is  limited as  a  suppressing  agent  because  of
its scarcity.
     The  control  efficiency  of wet suppression techniques  depends on local
climatic  conditions,  the  properties  of the fugitive  source,  and  the length
of effectiveness provided by the control.   Estimated control efficiencies of
80 percent  have  been  reported for cattle feedlots.16  Extensive watering of
the  soil  may  reduce  emissions from construction  operations  as  much as  70
percent.16  Wetting of access roads twice a day with 2.3 liters of water per
square  meter  will  suppress  fugitive dust from normal construction practices
an estimated 30 to 50 percent.

5.3.2  Stabilization
     Stabilization  techniques  isolate  fugitive  dust sources  from  external
disturbances  such as  wind and  traffic.   This  can  be done  physically  by
adding a  layer of  material  on exposed  surfaces; chemically  by using mate-
rials that  help  to  bind dust and particulates; and vegetatively by planting
of trees, shrubs, and grass over the surface.   Physical Stabilization.    For inactive  waste  stockpiles  and
steep slopes,  stabilization can  be accomplished by mixing  a  layer of rock,
soil, crushed  or  granulated  slag,  bark, wood  chips,  and straw with the top
layer.18  For  dirt  roads, paving  is the most effective  control.   The most
widely used low-cost pavement is  a bituminous  asphaltic chip  seal  over a
granular base or a stabilized soil  base.  Maintenance requirements depend on
vehicle traffic  and locale, but they generally include  a  second chip seal
after  1  year,  followed by  another  in approximately  5 years.  Gravel  is
sometimes used on  unpaved  roads,  but  it  is  less  effective  than  the chip
seal.19  In  one  study  of road paving, an estimated 85 percent control effi-
ciency was  cited.16   In  a study  comparing various methods  of controlling
emissions from unpaved roads  in Arizona,   gravel paving  of unimproved dirt
roads produced an estimated annual  control  efficiency of 50 percent.3
     Road carpets are  a  recent development for controlling  fugitive dust
from unpaved roads.  A water-permeable polyester fabric is laid between the
roadbed  (subsoil) and  the coarse  aggregate (e.g., gravel  or crushed rock)
road ballast,  which separates the fine soil particles  in  the roadbed from
the coarse aggregate.  This fabric prevents fine materials from reaching the
road surface and thereby  reduces  fugitive  dust.   Any fine  materials (<15
umA) in  the road surface would be washed into the road ballast during rain-
fall.  Fine  dust  in the ballast passes  through  the fabric into the subsoil
or  to  the  edge  of  the road,  but  fines in the subsoil  cannot  be pumped up
into  the ballast.   It is  the  minimization of fines  in the  road surface
material that effects the reduction in fugitive dust.20   Chemical  Stab11ization.   Chemical  stabilization  requires  the
use  of  materials that, upon drying,  bind  with surface  particles to  form a
protective  crust.   It  acts  in much the same way  as  physical  controls  (by

isolating the surface from climatic factors), and it is often used in combi-
nation with vegetative stabilization.  Chemical stabilization can be applied
to  unpaved  roads and  airstrips,  to waste  or tailings  piles,  to disturbed
soil  surfaces,  or to  reclaimed areas.  Many  types  of chemical  stabilizers
are available, and they can be applied with water or separately.
     Approximately 400  hectares of  the  inactive Kennecott  Copper tailings
area  west of  Salt Lake  City have  been  successfully stabilized  by aerial
chemical applications.21  Estimated control efficiencies of chemical stabil-
ization for a variety of sources are as follows.14
                      Source                        Efficiency, %
       Unpaved roads                                     50
       Construction—completed cuts and fills            80
       Tailings piles                                    80
       Cattle feedlots                                   40
The effectiveness of  chemical stabilization on unpaved roads varies accord-
ing to  the  amount of traffic.  Because heavy  traffic tends  to break up the
surface crust and  to  pulverize particles,  it  causes  greater entrainment of
particles into the atmosphere.
     The results of a study of various chemical stabilizers for dust control
on  unpaved  roads  were encouraging.19  The  stabilizers were  applied to sec-
tions of  an  unpaved  road with an  average  daily  traffic of 140 vehicles and
with a  surface  soil  silt content of 28 percent.   Some of the chemicals were
applied to  the  surface by spray and others were  mixed to a depth of 7.6 cm
and then  compacted.   After 5 months of surface stabilization, control effi-
ciencies of 83 to 95 percent were still being achieved.  After 14 months and
several bladings, control  efficiencies ranged from 9 to 54 percent, depend-
ing on  the  chemical  used.   When road  stabilizers were worked into the road
surface,  reductions in  emissions  of 80 to  95  percent were achieved after 5
months, and  12 to 84 percent after 14 months and  several  bladings.  These
results show  that working the stabilizer into the road surface causes it to
remain effective longer.3'19   Vegetative Stabilization.    Vegetation  is effective  in  the
stabilization of  a variety  of exposed surfaces.   In  many  cases, modifica-
tions first must be made to the surface or the surrounding terrain.


Sometimes,  physical  and  chemical  stabilization  techniques  are used  with
vegetative  stabilization.   Vegetative stabilization  is  restricted  to inac-
tive areas  where  the  vegetation will not be  mechanically disturbed once it
has been  planted.  Emission  sources that can  be controlled  by vegetative
covers  include mineral  waste  piles,  road  shoulders,  reclaimed  land,  and
disturbed soil surfaces.3'16
     In coal  mining and  preparation,  both fine  and  coarse waste materials
(low-grade  coal,  ash,  carbonaceous and pyrite shale, slate, clay, and sand-
stone) are  produced.   Because coal wastes are  acidic,  they must be treated
chemically and physically before vegetative stabilization can be implemented
effectively.   Chemical  treatment   involves  the  addition  of   an  alkaline
material  such as  limestone,  fly ash, phosphate  rock,  or treated municipal
sewage sludge.   Physical  treatment involves covering the waste with soil so
that it will  support vegetative growth.  The use of acid-tolerant vegetative
species is recommended even when the soil has been neutralized.
     Mineral  mining  and beneficiation  produce  wastes in  the  form of over-
burden,  gangue,  and   tailings.    Vegetative  stabilization  is  normally no
problem with overburden  and  gangue, but  it may  be  difficult  to  apply to
tailings  because  of a lack of nutrients, a concentration of saline or toxic
compounds,  and extreme  pH conditions.   Tailings  piles  therefore  must be
treated or covered with topsoil.
     Control  efficiency  of  vegetative  stabilization  varies  considerably
according to  the  amount and type  of cover.   One report estimated a control
efficiency  of 50  to 80 percent on tailings piles.16  In areas that are  less
than  hospitable to plant  growth  (e.g., the  arid Southwest),  reductions in
fugitive  dust could be  as little as 25  percent;3  whereas reductions could
approach  100  percent in areas that easily support dense vegetative covering.
5.3.3  Specialized Fugitive Emission Control Techniques
     Some fugitive dust control techniques are  relatively specific to  cer-
tain processes,  and thus are not as widely applicable as those just discus-
sed.   Sometimes  they  are  used  to  augment  some  of  the  techniques already
described.   Speed Reduction.    Reducing  the  speed  of  vehicles traveling
over unpaved  roads can lower the  fugitive dust emissions because it reduces


turbulence.  Speeds of  less  than 48 km/h cause emissions to vary in propor-
tion to the  square of the vehicle speed.22  Based on uncontrolled speeds of
64 km/h,  reduced  speeds would produce the following estimated reductions in
fugitive dust from unpaved roads.23
speed, km/h
reduction, %
Street Cleaning. Street cleaning ha
technique of  reducing  reentrainment of dust from paved roads.  Essentially,
                                                    • '.f. .  ;
three types of cleaners are now in use:  broom sweepers, vacuum and regener-
ative-air  sweepers,  and  flushers.   Broom  sweepers  use  a  rotating  gutter
broom to  sweep material  from  the gutter into the  main pickup  broom,  which
rotates  to  carry  the material  into  the  truck hoppers.    The  regenerative
air-sweeper uses an  air  blast to direct material  into  a  collection hopper,
and flushers use jets of water to move material to the gutters.3
     Results of field studies examining municipal street-cleaning techniques
in Kansas City,  Cincinnati,  and other cities were inconclusive; most of the
data showed that cleaning produced no effective reduction in emissions.3*24
Estimates  of  the  effectiveness  of cleaning  paved roadways at industrial
facilities  have  estimated efficiencies of  broom sweeping,  vacuum sweeping,
and water  flushing  have  been  estimated at 70,  75,  and 80  percent, respec-
tively.4  The broom  and  vacuum sweeping control efficiencies were  based on
biweekly  cleaning,   and  the water  flushing was based on  weekly cleaning.   Industrial Transportation Control.   Another  techniques  for
controlling fugitive dust from paved or unpaved roads  at industrial  facil-
ities is  to reduce  the  amount of vehicular traffic by providing perimeter
parking and  by bussing  employees to their work areas.   This technique has
been used at western coal mines and at an ir*on and steel plant.

     The  reduction  in  emissions  that  can  be  achieved by  transportation
control  is  directly  proportional  to  the reduction  in  vehicular  traffic.
This type of control can be used with paving and street cleaning.  Windbreaks.   Wind contributes significantly  to  fugitive dust
emissions.   Reduction of  surface  windspeed  by erecting physical  barriers
(windbreaks) perpendicular  to  the  wind direction is a logical  technique for
reducing emissions.  Windbreaks surrounding an agricultural  field can reduce
soil  erosion.   The  effectiveness  of  a  windbreak extends  downwind  for  a
distance of  10  times the barrier height.25  A  barrier  7.6  m in height will
control  erosion  76 m downwind of  the barrier.   Windbreakers can  reduce
erosion  by  12  percent  on  a  typical  600-m long  field;  they  are,  however,
generally considered infeasible for the control  of large fields.3  GoodOperating Practices.   Good operating  practices  can mini-
mize fugitive dust from  construction and from earthmoving activities.  Such
practices include the following.2'3
     Minimizing fall  distances when dumping material  from  frontend loaders
     Washing  vehicle  undercarriages   prior  to  their leaving  construction
     sites (to eliminate mud carryout)
     Limiting,  by proper  scheduling  of activities,  the exposure  time  of
     cleared land
Little information is available on the reduction of fugitive dust that could
be  achieved  by good  operating practices; however,  reductions  in  emissions
should be proportional to reductions in dust-generating activities.

                                 REFERENCES                  .            ,

 1.  Dunbar,  D.  R.   Overview  of  the  Fugitive Emission  Problem—1979 SIP
     Revisions.   Presented  at  the  3rd  Symposium  on Fugitive  Emissions:
     Measurement  and  Control,  San  Francisco,  California, October  23-25,
     1978.  EPA-600/7-79-182, August 1979.

 2.  Jutze, G. A., et al.  Technical Guidance for Control  of Industrial Pro-
     cess  Fugitive Particulate  Emissions.   EPA-450/3-77-010, March 1977.

 3.  Richard,  G.,  and  D.  Safriet.   Guideline  for Development  of  Control
     Strategies in Areas with  Fugitive Dust Problems.   EPA-450/2-77-029,
     October 1977.

 4.  Bonn, R., T.  Cuscino, Jr., and C.  Cowherd,  Jr.   Fugitive Emissions From
     Integrated Iron  and  Steel Plants.   EPA-600/2-78-050, March  1978.

 5.  Daugherty, D.  P., and  D.  W.  Coy.   Assessment of the  Use of Fugitive
     Emission Control Devices.  EPA-600/7-79-045, February 1979.

 6.  Wallace, D.,  and  C.  Cowherd,  Jr.   Fugitive Emissions  from Iron Found-
     ries.  EPA-600/7-79-195, August 1979.

 7.  Standards Support  and Environmental  Impact Statement  Volume  I:  Pro-
     posed Standards of  Performance  for Grain Elevator Industry.   EPA-450/2
     -77-OOla, January 1977.

 8.  Nicola, A. G.   Fugitive Emission  Control   in the  Steel  Industry.  Iron
     and Steel Engineer 53(7):25, July 1976.

 9.  American Conference of  Governmental  Industrial Hygienists.   Industrial
     Ventilation,  A Manual of Recommended Practice.  14th ed.   Committee on
     Industrial Ventilation,  Lansing, Michigan,  1976.

10.  Hagopian, J.  H.,  and E.  K. Bastress.   Recommended  Industrial  Ventila-
     tion Guidelines.   Department of Health, Education and Welfare, Pub. No.
     (NIOSH) 76-162, January 1976.

11.  American Society  of Heating,  Refrigeration and Air  Conditioning Engi-
     neers, Inc.   ASHRAE  Handbook  and  Product Directory,  1976 Systems.  New
     York.  1976.

12.  Weant, G.  E., III.   Characterization of Particulate  Emissions  for the
     Stone-Processing  Industry.   Research  Triangle   Institute,  Research
     Triangle Park, North Carolina.  May 1975.


13.   PEDCo  Environmental,  Inc.   Reasonably  Available Control  Measures for
     Fugitive  Dust  Sources.   Prepared  for  Ohio  Environmental  Protection
     Agency.  March 1980.

14.   Jutze, G.  A., and K.  Axetell.   Investigation of Fugitive Dust.  Vol. I.
     Sources, Emissions and Control.  EPA-450/3-74-036a, June 1974.

15.   Hoenig, S.  A.   Fugitive and  Fine  Particle  Control  Using Electrostat-
     ically Charged Fog.  EPA-60Q/7-79-078, March 1979.

16.   Siebel, R.  J.   Dust Control  at  a Transfer  Point  Using  Foam and Water
     Sprays.   U.S.  Department of  the Interior,  Bureau of  Mines.   Pub. No.
     TPR 97, 1976.

17.   Evans,  R.  J.   Methods  and  Costs  of Dust  Control  in  Stone Crushing
     Operations.  U.S.  Department   of  the  Interior,  Bureau of Mines.  Pub.
     No. 8669,  1975.

18.   Dean, K. C., R. Havens, and M. W. Giants.  Methods and Costs for Stabi-
     lizing  Fine -Sized Mineral  Wastes.    U.S.  Department  of  the Interior,
     Bureau of Mines.  RI  7896,  1974.

19.   Sultan, H.  A.   Soil   Erosion and  Dust Control  of Arizona Highways.
     Part IV.   Final  Report  Field  Testing Program.   Arizona  Department of
     Transportation, November 1976.

20.   Blackwood,  T.  R.   Assessment  of Road  Carpet  for Control  of Fugitive
     Emissions from Unpaved Roads.   EPA-600/7-79-115, May 1979.

21.   Jutze, G.  A., K. Axetell, and  R.  S.  Amick.    Evaluation of Fugitive Dust
     Emissions from Mining.   EPA-600/9-76-001, 1976.

22.   Cowherd,  C., Jr.,  K.  Axetell, C. M. Guenther, and G. A. Jutze.  Devel-
     opment  of Emission Factors for  Fugitive  Dust Sources.   EPA-450/3-74-
     037, June 1974.

23.   Compilation  of  Air  Pollutant  Emission Factors.   AP-42,  July  1979.

24.   Axetell,  K.   Control of Reentrained Dust from Paved Streets.  EPA-907/
     9-77-007, July 1977.

25.   How to  Control  Soil  Blowing.   U.S.  Department of Agriculture.   Farmers
     Bulletin, No. 2169, July 1961.


                                  SECTION 6


     Energy  is  consumed  in  the operation  of participate  control  systems.
Determining  the  optimum energy  usage  rate  for  an installation  involves
careful  balancing of  performance requirements,  availability  requirements,
and  operating costs.   Energy demands  of conventional  particulate  control
systems,  discussed in  Section 6.1,  emphasize the  potentials for  and  the
limitations of energy conservation.  Some control  devices have the potential
for  generating small  quantities  of secondary pollutants, not  originally
present; environmental  impacts  of these pollutants are discussed in Section
     Particulate  control devices  concentrate  the  particulate matter  en-
trained  in a gas  stream into a solid or liquid effluent stream for disposal
by  alternative means.    If  disposal  of  the  concentrated particulate  is
handled  improperly,  the original  air pollution problem  can become  a water
pollution  or a solid waste problem.   Environmental factors  considered  in
disposal of wastes collected in particulate control devices are addressed in
Sections  6.3 and   6.4.   Other environmental  considerations associated with
the  operation of   particulate  control  devices  include  noise  management and
radiation  control; these considerations  are  discussed  in  Sections  6.5 and

     The  goal of  any evaluation  of  energy usage  is  to identify means  of
minimizing  the energy  demand without  sacrificing  system performance  or
predisposing  the   system to  future  maintenance  problems.   A  complete  and
realistic  energy  inventory  is  the basis for  such an  optimization program.
     The  energy requirements of a control System  are  simply the sum of the
energy   requirements  of  each component.   Accordingly,   in the  following


sections,  the  energy  demands  for  major  components  are  evaluated  as  a
function of  gas  flow rate; the total system energy requirements are the sum
of the  energy demands of each component at that flow rate.  The results are
estimates which  indicate the relative importance  of  various energy demands
within  the  systems.   Optimization  of  energy  usage  is  based  on  these
6.1.1  Fan Energy Requirements
     Each control device introduces a static pressure loss into the effluent
gas-handling  system.  Typical  values are shown in Table 6-1.  The hoods and
ducts, both before and after the control device, introduce additional static
pressure losses (1 to 10 kPa).  The fan must be sized to deliver the desired
gas flow  rate at the total static pressure drop associated with the control

                     TABLE 6-1.  TYPICAL STATIC PRESSURE
             Control device type
          Electrostatic precipitator
          Mechanical collector
          Fabric filter
          Wet scrubber
         Range of
static pressure loss, kPa
        0.1- 0.5
        0.2- 1.0
        0.5- 2.5
          aTo convert kPa to inches of water, multiply by 4.019.

     To calculate  the  portion  of total fan energy that is chargeable to the
control device  (as opposed to the ventilation  system  for the process), the
fan  curves  must be  analyzed.   A comparison is  made  between power require-
ments at the actual operating point and power requirements for a similar fan
system without  the control  device static pressure loss.  Subtraction yields
the  incremental  energy requirements due to  the particulate control device.
Approximate  incremental  fan energy  requirements due  to  the total pressure
losses of the  ventilation system and the control device are shown  in Figure
6-1;1,2 these  general  curves do not apply to any specific facility or manu-
facturer.    These  incremental energy  requirements must be  adjusted for the
gas  temperature change  during  passage  through  the   system.   Most control




                                 I  TT
                                                       I  1  I  I- I L
      _  kWh = 1.06 a 10

                = pressure drop,  kPa
              Q = gas flow rate,  m3/min
              h = time, 7000  hours
              a = specific gravity, 0;7 (150°C)
              n = efficiency,  .65
                 ID3                    104

                  GAS FLOW RATE, «n3/min

Figure 6-1.  Incremental energy  requirements for fans.

devices  cool   the gas  stream 5°  to  25°C.   At  lower temperatures,  power
requirements  are greater  because the  density  of the  gas  is  increased.4
     One  means  of  reducing  incremental  fan  energy requirements  is  to
minimize static pressure losses.   With electrostatic precipitators, this may
be  accomplished by  a  smoother  gas  entry and  exit  the  box.   Better gas
distribution  yields  energy  savings   at  the same  time that  performance is
improved.   With  fabric  filters  energy  reduction  may  result  from  more
frequent cleaning, which  allows  a lower average static pressure drop across
the  fabric and dust  cake.    For  all  other  types  of particulate  control
devices, a decrease  in  static pressure drop is normally associated  with a
reduction in performance.
     The ventilation system gas flow may be improved.  All  sharp squared-off
turns  in ducting should  be  replaced  by  smooth gradual turns having lower
flow resistance.1  Likewise,  transitions  in duct  size  should be  as gradual
as  possible.    Unnecessary  turns  should  be eliminated.   Flow  of the  ,gas
stream into and out of the fan is particularly important.1'2'3
     Reductions  in gas  flow  requirements  due to movement of the hood closer
to the particulate generation source could have a  substantial impact in that
static pressure  loss  through  ductwork is proportional  to the square of the
gas  flow  rate.1  Insulation  and  sealing  of  ducts  (to  reduce cold  air
inleakage)  may be effective  on  large systems  transporting  gas  at elevated
     For cyclic  processes, the fan may be  shut down or dampered down when-
ventilation is  not needed.   There is a risk, however, that the gas tempera-
tures will  repeatedly  pass  below the acid  and  water dewpoints,  and lead to
corrosion.   Care must  be exercised  when  a particulate control  system is
temporarily shut down.
6.1.2  Control Device Energy Requirements
     Generalized  energy demand requirements are presented  for conventional
particulate control  devices.   These  do not include  requirements  for hopper
heaters  and vibrators  or for solids transport  equipment, both of which are
discussed  separately.   These  estimates  represent approximate values  only.   Fabric Filter  Subsystems.   The  fabric cleaning apparatus is
the dominant  energy  consumer  within  the control  device.   The total  energy


requirements are  calculated as the  power consumption  rate times the  time
that  the  cleaning  apparatus  is  energized.    Card,   Inc.5 estimates  that
reverse-air fans and  shaker motors use 0.5 hp/1000 ft2.  At an average air-
to-cloth  ratio  of  0.5 mVmin  per m2  and an  operating  time  of  2  minutes
per hour, the yearly (7000 hour) energy usage (kwh) is 1.87 x Q (ms/min).   This
is an order of magnitude below that for the fan requirements.  Regardless of the
type  of  cleaning,  energy  demand  is considerably  less than fan  energy re-
quirements,  based  on a  typical  fabric  filter  static  pressure  drop  of
1.5 kPa.
     Energy can be reduced by reducing the frequency and intensity of clean-
ing; but  in  most  cases,  this will result  in  higher average static pressure
losses, and  thus  in higher fan energy demand.  In view of the relative mag-
nitudes of energy  requirements  for cleaning apparatus and fans, less clean-
ing may  be  counterproductive.   The exception is conversion to a fabric with
better cake-release properties,  which  requires  less  cleaning without in-
creasing the fan energy demand.    Electrostatic Precipitator Subsystems.    Four   precipitator
components use electrical energy:
     0    Transformer-rectifier  sets  (T-R)  -  These  convert  alternating
          current at  line  voltages to high-voltage direct current, and they
          supply the  discharge  electrodes  that enable particle charging and
     0    High-voltage  insulators -  These isolate  discharge  wire  rappers
          from the  high-voltage frame.   Heaters are  normally advisable to
          prevent  surface  condensation, which  allows "leakage" of current;
          insulator heaters  can  be operated continuously or intermittently.
     0    Rappers -  These  remove  accumulated solids from collection plates,
          discharge  wires,  and gas distribution plates.   The  rapper system
          is  operated continuously  with  the  activation  frequency  for  an
          individual  rapper  normally  greater  on inlet  fields than on outlet
          fields.    Rappers  can  be powered either electrically or  by com-
          pressed air.
     0    Penthouse blowers  -  These purge the  upper housing of the precipi-
          tator of  vapors  that could condense  on  insulators and  cause cur-
          rent "leakage."  They are normally operated continuously.
     By far the most  important energy use  is by the T-R sets, which directly
control  particle  charging  and  collection.  Both  the peak  voltage  and the
total  power input rates  influence the level  of  penetration.6  There have

been  trends  to increase  both over  the  last 10 years;  for example,  design
power  input  rates have  been increased  from  a range of 3  to  10 w/(mVmin)
to about  20  to 30 w/(ms/min).6>7'8  The T-R set energy demand as a function
of design input rates is shown in Figure 6-2.
     Comparison of Figures  6-2  and 6-1  clearly indicates  the  importance of
T-R  set  energy consumption  in  an electrostatic precipitator  system,  where
the  operative mechanism  is  electrostatic  attraction rather  than  inertia!
impaction.   Impaction systems are normally operated by  imposing  a  gas flow
resistance,  which is  reflected  in  the fans  high-energy  demand.  No such
resistance is necessary in electrostatic precipitator systems.
     In actual  operation, the  power input to a precipitator  system  is not
constant  over  time.   Substantial  variability may be introduced by the input
mass  loadings  and the particle characteristics.6*7'9  High  inlet mass con-
centrations  lead  to  current suppression, particularly in the inlet fields.7
Likewise an increase in particle resistivity may reduce power input through-
out the unit.  Power input, as calculated in Equation 6-1,  may vary 30 to 50
percent  daily,  depending  on fuel  characteristics  and process  operating
conditions.                                                              ;

                    ET-R = *     (Pc  X Pv  * *
                     T R   i = 1 V ci    vi
where     Ey_R = total energy input to power supplies,
             n = number of transformer-rectifier (T-R) sets,
           PC. = primary current of T-R set i,
           Pv. = primary voltage of T-R set i,
            t. = time T-R set i  operational.
             of = power factor
Normally, the power input to the electrodes, E-r_R,  cannot be reduced.without
an adverse impact on penetration.   New power supplies are  designed to maxi-
mize power input.6'7
     The  power consumption  of  the rapper  system  is the  sum  of the  power
input  to  each rapper  times  the  fraction  of time  that  each  is  activated.
Because the  rappers  normally operate only a small  portion  of the time, even
on inlet  fields,  the energy consumption is more than  an order of magnitude

              I    I   I  I  I I  I I
                                               BASIS:   7000  h/yr
                                         i   i  i  i  i
                                                                       i i
                                GAS FLOW RATE, ms/min
           Figure 6-2.   Energy required for transformer-rectifier set.

less than that for the T-R sets.  Actual energy demand depends on the number
of  rappers  installed,  the  frequency  of  activation,  and the  intensity  of
     Reduced rapper  energy use may offer several secondary benefits.  Lower
intensity and lower frequency rapping of outlet fields can reduce reentrain-
ment  losses  in certain  installations.   Reduced  rapping  intensity  may also
lower  the probability of electrode misalignment, and  thereby reduce future
maintenance costs.  There is a limit, however, on the extent to which rapper
energy  savings  are  feasible.   That limit is  indicated by simultaneous in-
creases  in  primary  voltages  and  decreases  in   secondary  currents to  the
various  energized sections.    When these changes  occur,  excessive solids
accumulate on  collection  plates,  and precipitator performance begins to de-
teriorate.  The solids buildup can dampen rapper shocks transmitted down the
electrodes,  and thus further aggravate the solids problem.6
     High-voltage insulator heaters are operated continuously while the pre-
cipitator is operational.  Total energy consumption, indicated by curve A of
Figure  6-3, depends  on the number of heaters  and on the rated power input.
Despite  the continuous  operation,  the total  requirements  are  relatively
minor  because  of the  small  number  of  heaters.   This component  is a rela-
tively poor candidate for energy conservation because inadequate heating; can
lead to  vapor  condensation on insulator surfaces and  to  voltage reductions
and penetration increases.
     Energy demand for purge air fans is shown by curve B in Figure 6-3.  As
with  insulator heaters,  the  requirements  depend on  the  number  of blowers
used  and on their  sizes.   These fans  are typically operated continuously.
Potential energy  savings  are  very limited,  and  are gained  at  the risk of
increased condensation on high-voltage insulators.   Wet Scrubber Subsystems.   Components  using  energy  in  a wet
scrubber  system  are  identified  below.   In most  cases,  the instrumentation
power demand is small relative to the three listed components.
     0    Pumps -  for liquor recirculation, makeup water supply, and purge.
     0    Agitators  -  for mixing chemicals for  neutralization of scrubbing

J	I  111 I I I
1   1 J  1111
               KT                     10-

                  GAS FLOW RATE,  mVmin
          Figure  6-3.  Energy required for ESP insulator heaters  and
                               purge air fans.

     0    Reheat -  by  steam injection or by  indirect oil-fired burners for
          increasing stack temperatures and  thus  improving the meteorolog-
          ical dispersion characteristics of the effluent.
     The energy  required to operate pumps  is  a function of the liquid flow
rate and the  head.10  Centrifugal pumps are common in wet scrubber systems.
General power requirements for pumps are shown  in Figure 6-4; the liquid-to-
gas (L/G) ratios apply only to the recirculation pump.  The purge and makeup
flow rates  are  generally  only  1 to 5 percent  of  total  recirculation flow,
and the energy  demands are low relative to the recirculation pumps.  Energy
                                                 -„                        i
demand, of pumps  can be reduced either by  reducing recirculation flow or by
modifying spray  nozzles,  but both changes  can  lead  to reduced performance.
     Energy demand  for  stack  gas reheat depends primarily  on  the  gas flow
rate and  on  the degree  of heating  desired.   Typically, the  scrubbed gas
stream is  at the  adiabatic saturation temperature  (normally  50°C) and the
necessary  reheat is  25°  to  50°C.   The energy  requirements  are  shown in
Figure 6-5,
     Stack gas  reheating prevents objectionable ground-level concentrations
of unremoved  pollutants.   Means of minimizing  reheat costs include increas-
ing  stack  height,  increasing  stack  gas  exit velocities,  and eliminating
pollutants not treatable in the wet scrubber.11  Mechanical Collector Subsystems.  Mechanical collector subsys-
tems are  passive.   None  of  the components  within  the device use energy.
Solids removal equipment is addressed in a  later section.  Incinerator  Subsystems.  The  incinerator is the  only particu-
              :::' ~"~'         " " v """"""""""	 	™	™"niiiinni1 nni"11™1                                      |
late control  device directly using fossil  fuels.  The quantity used depends
on gas flow  rate,  combustion temperature,  inlet gas temperature, gas compo-
sition (H20,  C02),  and  heating value of gas  stream.   Means of calculating
energy requirements are discussed in Section 4.6.
     Reduced fuel requirements leading to reduced combustion chamber temper-
atures generally leads to higher particle penetration, especially for parti-
cles that  are  difficult  to  volatilize.   A heat  exchanger could prevent
reduced temperatures,  but poor incinerator performance  could  cause fouling
of the hot side.                                                          '.

  10"  —
                                           flow rate, L/min

                                           head of fluid, m,30

                                           specific gravity,  1.0
                                           efficiency,  0.65

                                           hours of operation, 7000
                                GAS FLOW RATE,  mVmin
                     Figure 6-4.   Energy required for pumps.

                                INLET TEMPERATURE 125°F
                                15% RADIATION LOSS
             TO2                     103
               GAS  FLOW RATE, m3/min
Figure 6-5,  Energy required for stack gas  reheat.

6.1.3  Hopper Heaters and Vibrators
     All dry  participate control  devices—Including mechanical  collectors,
fabric filters, and electrostatic preclpitators--occasionally require hopper
heaters and  vibrators  to maintain free-flowing  discharges  of solids.   Nor-
mally,  heaters  are used  only on  larger  devices when  high-temperature gas
streams are being  treated.   The energy requirements  for  heaters  and vibra-
tors  depend  on facility  size but are  independent of  control  device type.
     Heaters and vibrators  are normally operated in cycles.  Hopper heaters
are thermostatically controlled to maintain temperatures above both the acid
and  the water dewpoints.   Vibrators are  activated  by timers from  1 to 20
times an hour;  the energized time period can vary from several seconds to a
minute.  Energy requirements for heaters and vibrators are a relatively minor
part (<10%) of total system energy demand.
     Some energy savings  may be realized by derating the heaters and reduc-
ing  the vibrator  on-time; such savings, however,  are gained at the risk of
maintenance problems, and such changes should be carefully planned.  Failure
to properly  discharge  solids can  lead to misalignment of precipitator elec-
trodes,  to fabric  deterioration, or to  plugging  of  mechanical  collector
6.1.4  Solids Discharge and Transport
     Dry  particulate  control  devices  normally incorporate  some  type of
solids  discharge  valve at  the bottom of the hopper and  sometimes at screw
conveyor transfer  points.  Solids can be  transported by pneumatic systems,
pressurized systems, screw conveyors, or drag conveyors.  The screw conveyor
is normally used with small systems.
      For solids transport and  discharge equipment,  the energy requirements
are  normally  directly  proportional to the mass  of material.  With transport
equipment, the distance moved must also be considered.
6.1.5   Ultimate Disposal
      Energy  is needed  to  transport solids  from a  temporary  storage  site
(e.g.,  a  pile  or  covered pit)  to the ultimate disposal site (e.g.,  land-
fill).   Transport  requires  energy,  but   the  demand is  a  relatively  small

fraction  of  the overall  system energy demand  because the solids  are  in a
concentrated, manageable state.
6.1.6  Other Considerations
     Recent  increases  in  energy costs have motivated some operators to con-
sider recycling certain treated gas streams to occupied working areas.   This
may have a substantial benefit on costs of space heating, but this should be
adopted cautiously, with full consideration of potential occupational health
impacts of control system malfunctions.

     Participate control  devices have the potential  for generating limited
quantities of gaseous and particulate air pollutants.  Minimizing the gener-
ation of  these pollutants  requires  an understanding  of  their  physical  and
chemical  formation  mechanisms.   Unfortunately,  these  mechanisms   are  not
fully understood.                                                         :
6.2.1  Electrostatic Precipitators
     Ozone can be  formed  in  negative-corona electrostatic  precipitators.6
Concentrations of  5 to 20 ppm have been  reported.6   Ozone is toxic at con-
centrations  normally  encountered in precipitators.   Accordingly,  strict
confined  entry procedures should be  followed before  maintenance  is begun.
     Ozone generation is only partially understood.   Most mechanisms include
ionization of molecular  oxygen  during  a spark  incident or ionization of
molecular oxygen due  to absorption  of high-energy ultraviolet light emitted
by the corona.  The former mechanism should be controllable by the reduction
of spark  rate, which  is  now  possible  because  of advanced  power supplies.
The latter mechanism appears  more  difficult to  minimize without affecting
precipitator  performance.   The relative  importance  of these mechanisms is
not known.
     There is limited  evidence  that  some  sulfur  dioxide  is  oxidized  to
sulfur trioxide in a negative-corona precipitator.12   Such substances should
either  form  sulfuric  acid  aerosols  or condense  on  available  particle
surfaces.   In  either  case,  the generated sulfate would be indistinguishable
from  that generated  in the  combustion  process.   The  mode and extent of
sulfate formation are not well understood.

6.2.2  Incinerators
     Any  combustion  process  can  form  some  nitrogen  oxides.   Formation
mechanism is  thought to  be reasonably  represented  by the  following reac-
tions, which  collectively are  referred to  as  the Zeldovich  mechanism.*
          1.   N2 + 0 -> NO + N.
          2.   M  + 02 •* M  + 20 (M = any third body molecule).
          3.   N  + 02 •* NO + 0.

Reaction 1 has high energy; accordingly, the mechanism is active only at gas
temperatures exceeding  1400°C.   There is also a strong impact by the oxygen
content  caused  by reaction  6-2.   Basically, these  reactions are important
only within the  flame;  after the gases  leave the combustion zone, the reac-
tions cease, and nitric oxide remains.
     Considerable research has been devoted to techniques for suppression of
nitrogen  oxides  formation  in  coal-,  oil-  and  gas-fired boilers,  but very
little of this  work is directly applicable to burners of the scale and type
used  for particulate  incineration.   Available means of  reducing nitrogen
oxides generation include reduced flame  temperatures and reduced excess air.

     The major  sources  of water pollutants are effluents from scrubbers and
from  sluicing systems  for  removing  particulate  from hoppers.  An indirect
source  is  the  leachate  produced when  rain and  surface  runoff percolate
through collected particulate matter  that has been disposed  of improperly.
6.3.1  Regulatory Requirements
     Regulatory  requirements that apply to effluents streams from particu-
late  control  devices are similar to those that apply to  the process being
controlled.   Effluents  from scrubbers, wet electrostatic precipitators, and
controls  devices  using  wet sluicing  systems  are  regulated along with other
plant  sources under  the  Federal Water  Pollution  Control   Act (Clean Water
Act)  as  "direct  dischargers".  Direct  dischargers are point sources that
* The  symbol M  refers to  any third  body molecule.

must  conform  to  numerical  limits  on  various  pollutants  under  Federal
effluent guidelines developed  on an industry-by-industry basis and based on
the Best Practical  Control  Technology (BPCT).  Effluent guidelines based on
BPCT and BACT  apply to existing sources at the points of discharge from the
plant treatment  facility.   New effluent sources are subject to demonstrated
BACT, processes,  operating methods, or other alternatives  including (where
practicable) a standard permitting no discharge of pollutants.
     A rigorous discussion of water pollution regulation is beyond the scope
of  this  report.   Along  with  the  basic  Federal effluent  guidelines,  other
subsections  of  The  Clean  Water  Act  specify  control  requirements  for
"priority  pollutants"  and chemical  industry  pollutants,  which might  be
subject  to rules  and  regulations  under the Toxic Substances  Control  Act
(TSCA) enacted by Congress in 1976.
     Most  States  have EPA-approved National  Pollutant Discharge Elimination
Systems  (NPDES)   permit programs.   Such  programs  enable  State to  issue
permits to sources that comply with the requirements of the Clean Water Act,
provide  for  public participation  in  the permit  issuing process, and give
EPA, the Corps of Engineers,  and other States  the  opportunity to object to
the  issuance  of  a  permit.   Particulate controls  that generate  effluent
streams should be included on the NPDES permit.
     In  general,  particulate  control  systems  to which  effluent regulations
apply are  subject to  the  same regulations as  the" source  being controlled.
For example, a scrubber controlling a chemical process subject to TSCA might
collect toxic air  pollutants;  the presence of these toxic substances in the
scrubber effluent could  bring  the  scrubber  under  the TSCA  guidelines.
6.3.2  Control  Techniques
     The appropriate treatment for scrubber wastewaters can be selected on]y
after the  wastewaters  have been completely characterized.   Various constit-
uent  pollutants  of  scrubber  liquors, such  as suspended  solids,  dissolved
solids, toxic metals,  biodegradable  organics, and acids or caustics  require
different  types  of treatment  to meet regulatory discharge  requirements.   Primary Treatment.   Suspended   solids,  found  in  nearly  all
scrubber wastewaters,  are  removed by  sedimentation (commonly referred to as

primary treatment).   Sedimentation is  accomplished by allowing  the waste-
water to flow  slowly through a large basin or pond so that suspended parti-
cles  collect  by gravity.   Clarifier  basins  (Figure 6-6)  use  automatic
mechanical  devices  to continuously remove accumulated  sludge,  whereas set-
tling pond sludge  is removed in a  batch  fashion.   Increasing the size of a
sedimentation  basin or  pond increases the  wastewater detention  time,  and
thus improves  sedimentation  efficiency.   Sedimentation of very small parti-
cles can be  improved by adding flocculants to the wastewater prior to sedi-
mentation  to cause  the  fine  particles  to  agglomerate  into  larger,  more
easily separated particles.13*14
             Figure 6-6.  Sedimentation tank or "clarifier."   Secondary Treatment.   Some  scrubber wastewaters  may contain
biodegradable organic compounds which, if untreated, will exert a biological
oxygen  demand  (BOD)  on  receiving  waters.  Such  organic compounds  can be
found  in  scrubber wastewaters  at  pulp  and  paper  plants,  wood  products
plants,  food  processing plants,  and other  industries.  Such  BOD-causing
wastes must  receive  biological  treatment,  commonly referred to as secondary
treatment,  prior  to  discharge.   Failure  to  treat BOD-causing  wastes can

cause the  growth  of aerobic bacteria in  receiving  waters,  the accompanying
depletion of dissolved oxygen, the death of naturally occurring organisms in
the water, and the growth of anaerobic odor-causing bacteria.
     Secondary treatment  is  accomplished by allowing aerobic microorganisms
to  biologically  degrade  the BOD-causing organic  compounds  before  their re-
lease  into surface  waters.   This biological  activity requires a constant
influx of oxygen into the wastewater and sufficient detention time.  Second-
ary treatment is usually accomplished by oxidation ponds, trickling filters,
or  activated  sludge units.  Additional  sedimentation  must  follow  trickling
filters  and  activated sludge  units  to  remove  suspended bacteria  from the
wastewater.1S'1S  Tertiary Treatment.  Although primary and secondary treatments
may  remove most  suspended  solids  and  BOD  from wastewaters,  but advanced
wastewater  treatment,  sometimes referred  to  as  tertiary  treatment,  are
needed  for pollutants such  as nitrogen, phosphorus,  inorganic acids, non-
biodegradable organics, and heavy metals.
     In many particulate scrubbers, especially those at phosphate rock mines
and  at  fertilizer plants,  the liquors contain  phosjahorus  and nitrogen and
excessive  discharge  of  phosphorus  and  nitrogen  disrupts  the  ecological
balance  in lakes  and streams by stimulating profuse growth of algae.  Phos-
phorus  is  removed  by adding coagulants such as alum, lime,  and ferric chlo-
ride to convert the  phosphorus  to an  insoluable  form.13   These  additions
cause the  phosphorus  to precipitate,  coagulate,  and settle from the waste-
water.   Coagulants for phosphorus removal can be added before primary treat-
ment, before secondary treatment (if is required), or as a separate tertiary
     Three major processes  are  used to  remove  nitrogen from wastewaters.
Biological nitrification-denitrification  is the  aerobic biological conver-
sion of nitrogenous matter into nitrates (nitrification),  followed by ana-
erobic  biological  conversion  of  nitrates into  nitrogen gas  for release to
the  atmosphere  (denitrification);  both steps resemble  the  activated sludge
process in secondary treatment, but different microorganism colonies develop
in  the  vessels,  and denitrification occurs without oxygen.   A second method
is  ammonia stripping by  raising the pH  of the  wastewater and passing the

water through  a  stripping tower,  from which gaseous  ammonia  is released to
the atmosphere;  scrubbing liquor  nitrogen must be in the  form of ammonium
ions for  stripping  to be effective.  The  third  method  is selective ion ex-
change, which resembles a home water softener;  ammonium ions in solution are
exchanged  for  sodium or  calcium  ions displaced from an  insoluble exchange
     Scrubbing liquors  from cast iron cupolas,  certain  chemical  processes,
incinerators,  steel   mills,  and many other  industries  can contain soluble
organics  such as phenols and benzene, or colloidal  oils,  which are resistant
to biological breakdown during secondary treatment.  These impurities, often
referred  to as "refractory organics," can be responsible for unwanted colors
or tastes  in water, and many are suspected carcinogens.   Refractory organics
are generally  removed from wastewater by  carbon adsorption:   wastewater is
passed  through a bed of granular activated carbon,  where  the organics are
adsorbed  onto the carbon surfaces until  the carbon becomes saturated and has
to be regenerated or  replaced.13
     In scrubber wastewaters, toxic heavy metals such as chromium from metal
plating operations,  or  lead, mercury, and copper require special treatment.
Soluble metals can be removed  by  one of two methods.  One is high-pH lime
coagulation,  which  is   especially  attractive  if  phosphorus  must also  be
removed;  the other method  is selective  ion  exchange,  which  is  especially
attractive if nitrogen must also be removed.13'15
     Some  wastewaters contain significant colloidal  material  even after co-
agulation-sedimentation  in primary  or secondary treatment.   If this colloi-
dal material  is  not  suitable for discharge because of high turbidity or its
chemical  nature,  filtration  can  be accomplished  by  passing  the wastewater
through  a granular  bed  of  sand  or other  small  particles.   As  the  filter
becomes plugged, it can be cleaned by briefly reversing the flow ("backwash-
ing") at  a high  flow rate.   Backwash wastewaters,  usually  less than 5 per-
cent of total  flow, must be recycled to the wastewater treatment plant.13   Other Treatment Considerations.   Several other  points must be
considered when selecting a treatment or wastewater from a scrubber.  Should
scrubber  wastewaters be  combined  with  other  plant  wastewater  streams  or
treated separately.   If  the wastewater contains toxic materials  or has ex-
treme pH  values,  the routing of scrubber  effluents directly  into plant (or


municipal) treatment facilities can "poison" biological treatment processes.
Scrubber  liquors which  contain  neither  biodegradeable organics  nor toxic
materials  and which  require  no  secondary  or tertiary  treatment,  can  be
treated most economically by primary treatment separate from plant treatment
facilities.   Other liquors  can  be  routed  directly to  combined treatment
facilities,  or they can  be treated for  pH control  before  being routed to
combined  facilities.   If scrubbing  liquors are changed  in  a batch fashion
instead of by continuous blowdown,  flow equalization  facilities  may be re-
quired to prevent overload of plant treatment facilities.   Sludge Handling.   Purifying  scrubber  wastewaters  can lead to
another problem—sludge  handling.   Sludges withdrawn from treatment proces-
ses are still  largely  water (often more than  90%).   Thus, sludge treatment
must separate  solids from the large amounts of water,  return the separated
water to the wastewater plant for reprocessing, and dispose of the separated
solids  in an environmentally  appropriate manner  according  to  applicable
regulations (Section 6.4).
     Several   processes are  available  for the dewatering  of  sludge.   One or
more of  these processes  may be  required  to properly  dewater a particular
sludge.   A  common first  step in dewatering  is sludge conditioning, where
coagulants such  as  ferric chloride,  lime,  or  organic  polymers are added to
more easily  separate sludge solids  from water.  After conditioning, sludges
are often thickened by  gravity settling  in  vessels similar  to wastewater
clarifiers to reduce sludge  volume  by a  factor of  2  or more.   Biological
sludges  (i.e.,  from  secondary treatment  clarifiers) often  require sludge
"stabilization"  to  breakdown organic  solids  so that  they are  more stable
(less  odorous and less putrescible).   Stabilization  can  be  accomplished in
anaerobic and aerobic biological sludge digesters.13
     Many thickened or stabilized sludges receive final dewatering by vacuum
filtration (Figure 6-7).   A  vacuum  filter consists  of a cylindrical  drum
covered with  a filtering material or fabric partially submerged in a vat of
conditioned  sludge.   A vacuum  is applied  to  the inside of  the  drum to ex-
tract  the water, leaving the solids or "filter  cake"  on  the filter medium.
A  blade  scrapes  the  filter  cake  from  the  filter  medium as  the  drum
rotates.13  Some sludges  are filtered more readily  if  the filter medium is

precoated  with  various  dusts,  usually  applied  daily.   Another  method  of
final dewatering  is  drying on sandbeds.   After dewatering by vacuum filtra-
tion or sand drying, sludges are either incinerated or landfilled, depending
on their properties.
                        Figure 6-7.  Vacuum filter.13
     Substantial quantities  of  solids  and sludges are collected in particu-
late  control  systems.  Ideally,  these materials  are recycled to partially
offset control costs and to avoid disposal costs.  Unfortunately, the physi-
cal  and  chemical  characteristics of these  materials  frequently render them
noncompetitive  in  the very limited markets  presently available.   This sec-
tion  describes  the general chemical and  physical  properties that determine
recycle  potential  and disposal  requirements.   Information specific to par-
ticular  industries  and processes is available  in the Background Information

Documents  for  the Standards  of  Performance  for  New Sources  and  in  the
Industrial Process Profiles for Environmental Use.
6.4.1  Regulatory Requirements
     The  major  Federal   regulations  affecting  solid  waste disposal  were
established under  the  Resource Conservation and Recovery Act (RCRA).  Under
this  Act,  the  generation,  treatment,  and disposal  of  solid  (hazardous)
wastes are  strictly  regulated.   In most situations, a permit is required by
a generator (e.g., a facility that generates scrubber sludge), regardless of
whether the generator disposes of the hazardous material on or off site.   In
addition,  the  generator, transporter,  and the disposal  facility  must com-
plete  a  manifest  form  every time  hazardous  waste  is  shipped,  so that the
                                                      ',. ' ,1,1 ' ,"i,,i!i"! ,i' • ,
waste  can be  tracked  from  "cradle to grave" and  to eliminate  illegal dis-
posal of  hazardous wastes.
     When waste  from  a wet scrubber or  a  wet electrostatic precipitator is
disposed  of in  an  underground injection  well,  a permit  must  be  obtained,
according  to  the  Safe  Drinking Water  Act of 1974.   The  Act established a
national  program to prevent  underground  injections  that  endanger drinking
water sources.
6.4.2  Waste Recycle
                                                 .     f
     Properties  of accumulated materials determine the extent to which reuse
is economically  attractive.   A partial list of important physical and chemi-
cal properties is provided below:
     Physical properties                     Chemical properties
       Loss on ignition                        pH
       Carbon content                          Soluble fraction
       Particle  size distribution              Trace element composition
       Moisture  content
       Pozzolanic activity
"Loss  on  ignition"  and "carbon content" measure similar properties, in that
carbonaceous materials  normally constitute most of the combustible fraction
of  collected  solids.    Loss  on ignition of fly ash  from a pulverized-coal -
fired boiler normally ranges from 0.2 to 17 percent by weight.  The normally
acceptable  range is  6 to 12 percent by weight if fly ash is to be used as a
filler in portland concrete and asphalt concrete;  this  strict  range limits

the use of  fly ash as a filler.   In  1972,  only 11 percent by weight of the
fly ash from pulverized coal was used as a filler.   Since the carbon content
of fly ash  from a stoker-fired boiler is typically 25 percent or more, this
material  is unsuitable for use in concrete.
     Particle  size  distribution is important.   Excessive  fines  measured by
the minus-325  mesh fraction  generally inhibit  use.   For use as  filler in
cement, the  fine  fraction  should not exceed  12  percent by weight; however,
the reported values  for  ash from pulverized-coal-fired  boilers  are 7 to 60
percent.   The  fines  of any solids collected  in  particulate  control devices
depend on  the   effectiveness  of agglomeration in  the hopper  and  in solids
transport equipment.   (Agglomeration is caused  by  condensation  of moisture
and inorganic  vapors  on  particle surfaces during cooling.)  Excessive fines
renders any use  unattractive  because of the fugitive  dust  created during
materials handling.
     Moisture  content influences  recycle potential  in  a variety  of ways.
Low moisture content (<3.5% by weight) increases potential dusting problems,
as discussed above, and  increases the explosive potential  of dusts.  High
moisture content  (>20-40% by  weight) leads  to  materials-handling problems
because the solids begin to agglomerate and to cake, and may have an adverse
impact on   the  process  fuel  requirements.    For these  reasons,  the solids
collected in wet  scrubbers are rarely recycled.
     The chemical  composition  of the dry particulate  catch  can  affect use.
For example, the  alkali  content of portland  cement kiln dust must be below
the limits  stated  in  product specifications.  Often  it is  possible to use
solids from the  inlet fields of the precipitator,  but not  from subsequent
fields, which remove the majority of entrained alkali particles.
     The potential  for groundwater contamination from waste disposal sites
depends  partially  on  the  quantity  of  water-soluble compounds  that could
leach  out.   Fly  ash  from pulverized-coal-fired boilers  contains approxi-
mately 1  to 5  percent by  weight water-soluble compounds.   The major dis-
solved compounds   include  sulfates,   chlorides,  and calcium  ions; arsenic,
mercury,   and  cadmium  compounds are  moderately  soluble.  The  residue from
municipal incinerator  collectors can have water-soluble  fractions as high as
15 percent  by weight.

6.4.3  Waste Disposal
     Waste  characteristics  and regulatory  requirements  must be  fully con-
sidered in  the  selection of a disposal  technique.   Possible  disposal  tech-
niques include  (1)  placement in lined or unlined  ponds,  (2)  placement in a
landfill, either  as-received or after fixation treatment, and (3) deep well
     Unlined ponds  are  satisfactory when leachate from the waste liquors or
slurries can be controlled either by special pond construction (e.e., under-
drainage  systems)  or by  soil  permeability  properties.   Various  techniques
are  available  for  analysis  of pond leakage  include  soil  resistivity  moni-
toring networks and underdrainage system inspection sumps.
     Among  the  liner  materials available  to improve leachate security in
ponds  are various  synthetic  materials and  clays.  Flexible liners  have an
estimated life  of  20  to 25 years;  nonflexible liners  are more  permanent.
Liner  thicknesses range  from 0.025 to 0.075 cm for synthetic materials such
as  polyethylene and  polyvinylchloride,  from 30  to 40  cm for  clays,  and
approximately 15  cm for  asphalt  and concrete.   Cost of  the  lining  must be
weighed against the security offered by thicker liners.
     Dry  or  dewatered  material can be disposed of  in a  landfill.  The sta-
bility  of the  fill material  and  the groundwater  contamination potential
should be considered.   For  example, water runoff should be channeled around
these  sites  to  minimize leaking of soluble compounds.  Soil characteristics
and groundwater levels should be determined to avoid improper landfill  loca-
tion.                        '                                            j
     Treatment  of  the waste  may  be necessary  to  reduce potential  landfill
problems.  Scrubber sludges  can be treated chemically to "fix"  the material
into a physically stable, leach-resistant  matrix.   Wastes  can  also  be fil-
tered to reduce moisture content.
     Deep well  injection generally is not  economically  attractive  for dis-
posal  of  the quanitities and types of  wastes  discharged  from particulate
control devices.                                                          >

     Noise from particulate  controls  is generated by fans, ESP rapping, wet
scrubber pumps,  and solids  transportation systems.  The  noise  levels from
most of  these sources  are usually negligible compared with  noise from the
other plant sources.  Objectionable noise levels attributable to particulate
control systems are most  frequently generated by  fans.   If there is a need
to  decrease  the  noise  level  from  fans,  the rotational  speed  should  be
reduced and the additional capacity shifted to a parallel fan.  Sound insu-
lation can  also be  used for  fans  as  well  for other components, such as ESP

     Radiation sources  associated with particulate  controls  are limited to
nuclear  level  indicators  for hoppers.   Radiation from such  indicators  is
usually minor  and not  sufficient to warrant the requirement that plant per-
sonnel wear nuclear badge detectors or dosimeters.   These potential radia-
tion  exposure areas,  however,  must  be  marked  in  accordance with Federal
regulations.   Periodic  checks  could  be made  for  radiation  leakage  with a
Geiger-Mueller detector.

 1.  American Conference  of  Governmental  Industrial Hygienists.  Industrial
     Ventilation Manual,, 16th edition.  Lansing, Michigan, 1980.

 2.  Jorgensen,  R. ,   editor.   Fan  Engineering 7th  edition,  Buffalo Forge
     Company, Seventh Edition, 1970.

 3.  Pollak,  R.   Selecting Fans and  Blowers.   Chemical  Engineering, 80(3),
     January 23, 1973.

 4.  Heath,  C.   Energy  Analyses.    National   Asphalt Paving Association.
     Publication 52,  October 1978.

 5.  Neveril, R, B.   Capital  and Operating Costs  of Selected Air Pollution
     Control Systems.  EPA-450/5-80-002.  December 1978.

 6.  Katz,  J.   The  Art  of  Electrostatic  Precipitation.    S  &  S  Printing
     Company, Pittsburgh, Pennsylvania, 1979.

 7.  Szabo, M., and R. W. Gerstle.   Operation and Maintenance  of Participate
     Control  Devices  in  Kraft  Pulp  Mill  and  Crushed  Stone Industries.
     EPA-600/2-78-210, 1978.                                             ;

 8.  Szabo, M., and  Y.  Shah.   Inspection Manual  for Evaluation of Electro-
     static  Precipitator  Performance.   EPA-340/1-78-007,  February  1979.

 9.  White, H.  Electrostatic  Precipitation of Fly Ash.  Air  Pollution Con-
     trol Association, Pittsburgh, Pennsylvania, July  1977.

10.  Karassik,  I.,   and  W.  Krutzsch.   Centrifugal and  Axial  Pumps.   In:
     Standard  Handbook  for Mechanical  Engineers,  7th edition.  McGraw-Hill
     Book Company, New York,  N.Y., 1967.

11.  PEDCo Environmental Specialists, Inc.  Flue Gas Desulfurization  Process
     Cost Assessment.  Draft report for U.S. Environmental Protection Agency
     under contract no.  68-01-3150.

12.  Brown, W.  R. and E. E.  Stone.   Sulfur Dioxide Conversion under Corona
     Discharge  Catalysis.   U.S.  Public Health  Services  contract no. PH-86-
     67-2, March 1965.

13.  Environmental Pollution Control  Alternatives:   Municipal Wastewater.
     U.S. EPA Technology Transfer.   EPA-625/5-76-012,  1976.

14.   Metcalf  and  Eddy,  Inc.    Wastewater  Engineering.   McGraw-Hill  Book
     Company, New York, N.Y., 1972.

15.   Weber, Jr., W.  J.   Physiochemical Processes for Water Quality Control.
     Wiley-Interscience, New York, N.Y., 1972.


                                  SECTION 7  ,

     The  selection  of  a  particulate control  technique  depends  upon  many
factors, such  as degree  of required emission reduction,  gas  stream  charac-
teristics,   and  cost.   This  section  deals with  the  costs  of  purchasing,
installing,  and  operating various  particulate  control  devices  and  tech-
niques.  The  particulate control equipment  addressed  includes  electrostatic
precipitators,  fabric  filters,  mechanical  collectors,  incinerators,  and
scrubbers.   The  fugitive  emissions control techniques  evaluated  include wet
suppression and stabilization.
     The total  cost of  a parti cul ate control  system  is  influenced  by many
factors.  The  cost  of  the control and auxiliary devices can be highly vari-
able in view  of  the many options that may be applicable.  Even when  the type
of  equipment  is selected,  different materials of construction  and  penetra-
tion  levels  affect costs  to  a  large   extent.   Auxiliary  instrumentation
useful  to   improve  reliability  and  to  reduce maintenance  can  also  have  a
great effect on cost:
     Retrofit  applications can  result  in  installation  costs  that  greatly
exceed  the  costs   of  a  new  installation.    Space  restrictions, difficult
tie-ins, and outdated process equipment can cause added expense.
     In addition to the  collection costs of particulate,  a  cost estimate of
residue disposition must be made.   Collected solids or  sludges  may  consti-
tute  a significant expense  for  disposal, or, when  the particulate  consists
of  recovered  product,  a  valuable credit.   Facilities  for treatment  of wet
scrubber sludges may be necessary.
     Labor  expense  affects  both  capital  and annualized costs  in the form of
installation,  operation,  and maintenance.   This expense also carries  with it
sizable overhead costs.

     Given  compliance  with other limiting  factors,  cost-effectiveness  often
provides  a primary  criterion  for choosing  among various pollution  control
alternatives.  Cost-effectiveness is  a measure of the total  cost of a speci-
fied  reduction in  emissions.   Computation  of cost-effectiveness  must  take
into  account  all   annualized  costs  including  direct  operating  costs  and
capital charges.                                                         ;

7.1.1  Capital Costs
     Capital  costs  of a  particulate  control system include the  cost of the
purchased  equipment;  i.e., the  control  device and  its  accessories,  and all
installation  costs.   Installation costs  are divided  into direct costs,such
as those  for  foundations, piping, and painting,  and  indirect costs  such as
those  for engineering and  supervision.   Equipment  costs generally form the
basis  for estimating total  capital  costs for a  particulate control  system.
     There  are several  methods  for  estimating  the  total  capital costs  of
particulate removal  systems.   The accuracy of any method is  directly related
to the amount and  detail of  the information available.  Simple cost-esti-
mating methods dependent  solely on the type of unit and its  capacity are the
least  accurate;   methods  requiring  preliminary  engineering  drawings  and
specifications and  detailed  energy  balances are generally the  most accurate.
A discussion of several estimating techniques follows.
     An order-of-magnitude  estimate  of  total  capital costs is  based on the
average  cost  for  equipment of  a  particular type  and  capacity.    For  air
pollution  control facilities,  these  parameters are usually determined by gas
volume  to  be  treated and  the  desired pollutant  removal  efficiency.   The
range  of  average capital  costs can be  very wide; thus this  simple method of
estimating costs is the least accurate.
     Another  method  of   estimating  total  capital  costs  is  that in  which
factors are applied to the cost  of the  major pieces of equipment as a means
of estimating the   remainder of the  costs.  Experience  gained from previous
projects  (historical data)  provides  the multiplication  factors  used  in  this
method.   There are  several  variations of the factor  method,  each requiring

different  input  and  producing  different degrees  of  accuracy.   The  Lang,
Chi 1 ton, and Guthrie  procedures  are three variations that were developed for
use  in  estimating the  cost of  chemical  plant construction.  The method of
cost estimation used  to obtain the cost curves presented in Section  7.3 is a
variation of  the  Lang method  that  EPA has developed for  application  to air
pollution control  systems.   This  system is  fully explained  in  Reference 1.
     When total  capital  cost data are  available  for  a system of  similar
design but of  different'capacity from that required  for a particular appli-
cation,  a  scaled  estimate can  be utilized.    Scaled estimates are  usually
derived by use of the following equation:

     E2 = E1 (^)n                                               (Eq. 7-1)

     Eg = cost of desired control device
     E, = cost of scaled control device
     r« = capacity of desired control device
     r-, = capacity of scaled control device
      n = exponent relation
     This  equation specifies  that  a  log-log plot  of  capacity versus cost
should be a  straight  line with  slope  n.   With respect to equipment costs, n
has  been shown to average  0.6 and  is  referred to as the six-tenths factor.
Although  the  six-tenths  factor is  most accurate when  applied to  a  single
piece  of equipment,  it  can   also  be  used  for  estimating  complete  system
costs.   Its  use  should  be limited, however,  to cases where no other costs
are readily available for the desired control  device.
     The most  accurate  methods of  capital costs  estimation require  complete
drawings and  specifications,  material  and energy balances, site surveys, and
other  engineering effort.  The  extent of the data  obtained will  determine
the  accuracy provided  by the estimate.   Accuracy of ±5  percent is  possible
if  sufficient  information is  available.  In contrast,  however,  cost spreads

of 20 percent  are sometimes encountered in formal  bids as a result of market
conditions,  interpretation  of plans  and specifications,  subjective  assess-
ment of installation difficulties, and error.
7.1.2  Annualized Costs
       ,*mmmmmmm~—,, . . n.i.i^—.n.-,. , ,..—	                                                  |
     Annualized  costs  of a  particulate control system  include  direct costs
such  as  operating  labor  and materials,  maintenance,  replacement  parts,
utilities, and the  costs of particulate disposal;  also included are indirect
costs such as  overhead,  insurance, taxes, and  capital  recovery,  and credits
derived from recovery of particulate product.
     Direct annualized cost  estimates are obtained by applying unit costs  of
utilities,  labor,  and  materials  to  the  estimated  requirements  for  these
items.  Indirect  costs are  derived generally by combining percentages of the
capital costs  with a percentage  of labor charges for  operation  and mainte-
nance and  capital  recovery  costs.  Capital recovery costs depend on interest
rates, the useful  life of the equipment,  and  the  equipment's salvage value,
if any.
7.1.3  Other Cost Considerations
     Certain aspects of  cost analysis are always  affected by some degree  of
uncertainty.   When  new technology is included, an  adequate data base may not
be  developed for  proper evaluation.   Differences  in the  expected  service
lives  of  alternative  systems may require  appropriate adjustments  in  cost
comparison.  Changes in  labor rates, material  costs,  or fuel costs also may
affect the accuracy of cost analysis.
     Inflation is  always an  important consideration,,  since inflation rates
are subject  to change,  reflecting other economic  factors.   Cost indexes are
available to aid in adjusting past costs into current dollars.
     The cost of retrofit applications can be  difficult to assess by use  of
typical cost analyses.   Additional engineering input is important to account
properly  for  potential   additional  expenses  for such  items  as  site prepara-
tion, overtime labor,  utility system modifications,  space restrictions, and
lost  production.   By  the methods described  in Reference 1,  retrofit  cost
curves  have  been  developed  and  are  compared with  those  for  grass-roots
installations in Section 7.3.

     The  method of  cost analysis  used  in  the  analyses  presented here  is
identical to  that used  in  Reference 1,  which should  be  referred to  for  a
detailed  explanation  of cost  analysis  of air pollution control  systems.   A
brief explanation of the method is included in this section.
7.2.1  Capital Costs
     Purchased equipment costs  provide  the basis for estimating  the  remain-
ing capital costs;  i.e.,  direct and indirect installation costs,  for  partic-
ulate control  systems.   The  purchased  equipment costs include the price  of
the  control  device,  auxiliary equipment,  instruments  and controls,  taxes,
and  freight.   Although  the  cost  of a device  and  its  auxiliaries  may  be
fairly standard  for  a particular size and type,  the costs of  instrumentation
and  freight  can vary considerably,  depending  upon the type and  location  of
     Installation costs are  derived by  applying the  applicable cost  factors
shown in  Table  7-1.   Many of the  individual  items  in the installation cate-
gories,   direct  and  indirect,  are subject  to  site-specific adjustment.   In
the  direct  cost category,  erection  and  handling,  site preparation,  and
facilities and   buildings  are  subject  to  adjustment.   Purchased  equipment
cost does not,  however,  directly determine the work needed  in preparing the
site of  erecting buildings.   These cost  items  are more  dependent upon the
nature of the  facility  and whether the application  is  new or retrofit.  The
costs of  foundations  and supports, electrical work,  piping,  insulation, and
paintings  are all generally proportional to  the purchased equipment  cost,
and adjustments are not deemed necessary.
     Many  indirect  costs   can  vary  considerably.   Engineering  and  super-
vision,  construction  and  field expenses,  and construction fee depend  to some
degree on site-specific  conditions and  therefore require appropriate  adjust-
     Table 7-2   indicates that  total direct  and indirect installation cost
factors  can   range   widely.   The  importance  of  knowledge of the  specific
application  is   clear;  use  of  the appropriate cost  adjustments  can be the
most important aspect of this cost analysis.

Cost factors
Direct costs

Purchased equipment costs
a) Control device
b) Auxiliary equipment
c) Instruments and controls
d) Taxes
e) Freight

Installation direct costs
a) Foundations and supports
b) Erection and handling
c) Electrical
d) Piping
e) Insulation
f) Painting
g) Site preparation
h) Facilities and buildings
Indirect costs

Installation indirect costs
a) Engineering and supervisior
b) Construction and field
c) Construction fee
d) Startup
e) Performance test
f) Model study
g) Contingencies

















   These average factors may require adjustments for individual estimates.
   As  required.
   The relative  costs for items 2(g) and 2(h) must be added to these average

                        Adjustment factor
      1.  Simple, continuous manually operated
      2.  Intermittent operation,  modulating  flow with emissions
            monitoring instrumentation
      3.  Hazardous operation with explosive  gases and safety backups

      1.  Major metropolitan areas in continental U.S.
      2.  Remote areas in continental U.S.
      3.  Alaska, Hawaii, and foreign

    Handling and erection
      1.  Assembly included in delivered cost with supports, base,
            skids included.  Small to moderate  size  equipment
      2.  Equipment supplied in modules, compact area site with ducts
            and piping less than 70 meters.   Moderate-size system
      3.  Large system, scattered equipment with  long runs.  Equip-
            ment requires fabrication at site with extensive welding
            and erection
      4.  Retrofit of existing system; includes removal  of existing
            equipment and renovation of site.   Moderate  to large

    Site preparation
      1.  Within battery limits of existing plant; includes minimum
            effort to clear, grub, and level
      2.  Outside battery limits; extensive  leveling and removal of
            existing structures; includes land  survey and study
      3.  Requires extensive excavation and  land  ballast and
            leveling.  May require dewatering and pilings

    Facilities and buildings

      1.  Outdoor units, utilities at site
      2.  Outdoor units with some weather enclosures.   Requires
            utilities brought to site, access roads, fencing,
            and minimum lighting
      3.  Requires building with heating and  cooling,  sanitation
            facilities, with shops and office.  May  include  rail-
            road sidings, truck depot with parking  area

    Engineering and supervision
      1.  Small-capacity standard equipment,  duplication of  typical
            system, turnkey quote
      2.  Custom equipment, automated controls
      3.  New process or prototype equipment, large  system

    Construction and field expenses
      1.  Small-capacity systems
      2.  Medium-capacity systems
      3.  Large-capacity systems

    Construction fee
      1.  Turnkey  project,  erection  and  installation included in
            equipment  cost
      2.  Single contractor for total  installation
      3.  Multiple contractors with  A&E  firm's supervision

       1.  Firm process
      2.  Prototype or experimental  process subject to change
       3.  Guarantee of efficiencies  and operating specifications
0.5 to 1.0
1.0 to 1.5
0.2 to 1.0
0.2 to 0.5


1.0 to 1.5

 1 to 2

 3.3 to  5
  5 to 10

7.2.2  Annual 1zed Costs
     Annualized costs  are  categorized in Table 7-3,  which also gives example
cost factors.   Direct operating costs  are  such that rates for  labor,  mate-
rial, and  utilities can be applied  to  estimates of requirements  for  these
items.   The  rates can  be  obtained as  average figures from sources  such  as
the Bureau of  Labor Statistics and the Federal Energy Regulatory Commission,
or rates for specific areas can be obtained from actual  consumers.
     The requirements  for  operating  labor  and supervision depend  on system
variables,  such as  degree  of automation, and  operational  variables,  such  as
continuity  of  operation   and  number  of shifts.   Maintenance  requirements
depend  upon  the  nature of the gas stream;  e.g.,  corrosiveriess  or  abrasive-
ness, construction  materials,  and  system size and type.  Labor and mainte-
nance costs  can  be estimated from  information  presented  in  Reference  1.
     Utility  requirements   are derived  by  use  of  the following  formulas:

          kwh = 8.3 Q (Ap)(SG)(h)
                   10  n
          kWh = energy usage in kilowatt-hours
            Q = actual volumetric flow rate, ms/s
           Ap = pressure loss,  Pascals
            n, = efficiency, usually 60 to 70 percent
            h = hours of operation
           S6 = specific gravity as compared to air
          kwh = 9.80 Q(H)(SG)(h)

       Direct operating costs
       Cost factor
Operating labor

Operating materials


Replacement parts

  Fuel oil
  Natural gas
  Plant water
  Water treatment and cooling water
  Compressed air

Waste disposal
15% of operator

As required
100% of maintenance labor

As required

     Indirect operating costs

Property tax



Capital recovery cost
80% of operating labor and
maintenance labor

1% of capital costs

1% of capital costs

2% of capital costs

0.16275 (as an example of
10% and an equipment life
of 10 years)
Recovered product
As required
  All costs are in December 1977 dollars.

          kWh = energy usage in kilowatt-hours
            h = hours of operation
            Q = flow rate, m3/s                                           ;
            H = head of fluid, m
           SG = specific gravity
            H = efficiency, usually 60 to 70 percent
     Waste disposal costs  do not always apply;  when  products are recovered,
such material constitutes a credit.                                       ;
     Indirect operating  costs are  based on both direct  operating  costs and
capital  costs.    Overhead  costs  are  a  straight  percentage  of wages  and
salaries; they  cover expenses  for such  items  as fringe  benefits  and cafe-
     Capital  recovery  costs are  derived by  use of the  following  equation:

     Capital recovery cost = capital costs x ^*	2J—          (gq. 7-4)
                                             (1 •*- i)n- 1
          i  = annual interest rate
          n = capital recovery period, years
     Specific  information  regarding  overhead   cost  factors  and  equipment
lives can be obtained from Reference 1.

     This section presents  cost  data for  several  particulate  control  sys-
tems.   Three cost  curves  are  given  for  each  type  of  control  technique,
representing  purchased  equipment  costs,  total capital costs,  and annualized
costs.   All  costs are  estimated  from information  contained  in  Reference  1
and updated to  January  1980 dollars.  The  accuracy of any curve relative to
a specific application  depends  upon the  similarities  between the assumptions
used in  the  example and the conditions  under which the  system will actually

be used.   The curves should not  be  relied upon, however, to provide  better
than ±50  percent accuracy.  For  more precise  estimates, it is  recommended
that  the  reader  apply  the  cost  analysis  method  described  earlier  and
explained in detail in Reference 1.
7.3.1  Equipment Costs
     Figure  7-1  through   7-5  show  the  estimated purchased  costs,  F.O.B.
factory,  for five  state-of-the-art particulate  control device  categories:
(1) electrostatic  precipitators,  (2) fabric filters, (3) mechanical  collec-
tors  (cyclones),  (4) incinerators,  and (5) venturi  scrubbers.   The  curves
represent  flange-to-flange costs  and generally  include  internal  electricals
and controls.   Instrumentation is  not  included because  it is  usually pro-
vided  as  an  optional feature.   The cost  curves are presented  in terms  of
dollars versus  exhaust  gas  volume.  This  relationship  is based  on  a number
of simplifying  assumptions, which  allow one to obtain  quick, conceptual  or
study  estimates  with a minimum of  effort.   It must be borne  in mind that
these  simplifications  can  lead to anomalous  results  at  the extremes  of the
ranges, since the  curves  are presented  in  the  form y = ax  and  are  based on
regression analyses.
7-3.2  Particulate Control System Costs
     On  the  basis of  information  in  Reference  1 and  the equipment  cost
curves  in  Figures  7-1 through  7-5,  a number  of cost curves have been devel-
oped  that prouide conceptual  or  study  estimates of the  capital  and annual-
ized costs  of complete  air pollution control systems.   These  curves provide
costs  for grass-roots installations.   A retrofitted installation  generally
costs  10  to  30  percent more than a grass-roots installation and,  depending
on specific  difficulties  at a  given site, the costs can be calculated on the
basis  of the  latter percentage.
     Annualized  costs  are  based on  8700  h/y operation  time.   Since  the
annualized  costs  vary with  operating  time, the annualized  costs for opera-
tions  of  less than  8700   h/y will  be  lower than those  shown in  Figures 7-6
through 7-14.   For example, the annualized costs for 2000 h/y operation as a
percent  of the costs  for 8700 h/y operation  are  approximately  as  follows:







                                      T  I
                                    SCA = nr/dOOO nTYmln)
                                      n = COLLECTION EFFICIENCY
                                                                  FOR  DUST
                                FOR DUST
                                      J	I
30  40  50
                                                              200   300  400
                                   EXHAUST GAS RATE, m3/s
             Figure 7-1.  Cost of electrostatic precipitators; carbon steel
                    construction, thermally insulated, FOB factory.
                   (Instruments and controls and taxes not included.)


 o .






                           i     i    r
                                    i    i
                                             TYPE OF CLEANING MECHANISM

                                             AIR-TO-CLOTH RATIO (m/s)
                                                  CURVE 1  0.46 TO 1.0
                                                  CURVE 2  0.61 TO 1.0
                                                  CURVE 3  2.12 TO 1.0

                                                  CURVE 1  NYLON
                                                  CURVE 2  NYLON
                                                  CURVE 3  NOMEX
30  40  50
200   300  400
                                         EXHAUST  GAS  RATE,  nfVs
           Figure  7-2.   Cost of  fabric  filters,  carbon  steel construction,  FOB  factory,
                        (Instruments  and  controls  and taxes not  included.)



                                     i   T



                                     345         10

                                      EXHAUST GAS RATE, m3/s
 I    L

30   40
          Figure  7-3.   Cost  of mechanical  collectors, carbon steel construction,

                        FOB factory.   (Instruments and controls and taxes not








                                           T	T



                                2      34

                                  EXHAUST GAS RATE,
30  40
                 Figure  7-4.
                        Cost of incinerators, FOB factory.   (Instruments

                        and controls and taxes not included.)



 gg  100

 1   80

"o   60
   CURVE 1  15
   CURVE 2  10
   CURVE 3   5
           Figure 7-5.
                           20   30   40 50
                       200    300   400
                               EXHAUST GAS RATE, mj/s
                    Cost of venturi  scrubbers,  unline'd throat, carbon steel
                        construction, FOB factory.
                        and taxes not included.)
                                                (Instruments and controls

                                                     CAPITAL COST
                                                     ANNUALIZED COST
                                                    COST OF DUCT INCLUDES
                                                    ONE ELBOW
   Figure 7-6.
                 EXHAUST GAS RATE, m /s
Capital and annualized costs of fans and 30.5 m length of duct.







o  800

i  600
<  400
£  200




                                            NOTES A & C
                                            NOTES B & C
                                            NOTES A & D
                                                                  INOTES  B  &  D
   SCA = m2/(1000 mVmin)
30  40  50
200   300  400
                                    EXHAUST GAS RATE, itT/s
        Figure 7-8.  Capital and annualized costs of electrostatic precipitators,
                     carbon steel construction.



   6000 [~









 3   200
 §   100
 1    80
"o    60
o    40
                                                                        T    T
I    I    I
                                                               CAPITAL COST
                                                               ANNUALIZED COST
                              2.0    3.0  4.0  5.0
                                                                 20.0  30.0  40.0
                                      EXHAUST GAS RATE, nT/s
           Figure 7-11.  Capital and annuallzed costs of mechanical collectors,
                         carbon steel construction.









                         I      1    T   I
                         I     I
                                    CAPITAL COST
                                    ANNUALIZED COST
                              CURVE  1
                              CURVE  2
                              CURVE  3
                              CURVE  4
               Figure 7-12.
                        2.0   3.0  4.0 5.0
         10.0      20.0  30.0 40.0
                               EXHAUST GAS RATE,  nT/s
                       Capital and annualized costs  of incinerators.



          I      I    I
                                     I    T
 g  800

 §  600




3  1000

^  800
~  600

1  400





                                         1    T
                             Tl    T
                                                 — CAPITAL  COST
                                                 ™-ANNUAL I ZED COST
                                                 PRESSURE  DROP,  kPa

                                                     CURVE 1  15
                                                     CURVE 2  10
                                                     CURVE 3  5

               Figure 7-14.
                              20    30  40  50
                 ,  100
200   300  400
                            Capital and annualized costs of venturi scrubber,
                            stainless steel construction.

          Venturi scrubber              30 - 40 percent
          Fabric filter                 50 - 60 percent
          Electrostatic precipitator    60 - 70 percent
          Incinerator                   25 - 35 percent
     The annualized  cost includes a  disposal  cost of $10 per ton,  based  on
disposal of  a nontoxic  substance.   It also includes a capital  charge  based
on  an  assumed  equipment life  of  15 years  and  an  opportunity cost of  15
     Each of  the  control  system cost  curves  includes  the costs  of auxiliary
equipment normally associated with such a system.   In some instances one may
wish to  know  what the system would cost either with or without the ductwork,
fan, and fan  drive.   The capital and  annual ized  costs  of the components are
shown in Figures 7-6 and 7-7.
     7,3.2.1   Electrostatic Precipitator.   Figure 7-8 presents cost  curves
for systems utilizing  an electrostatic precipitator housed  in an  insulated,
carbon steel  shell.   The assumption is made that the uncontrolled  gas stream
is  normally  vented to  a stack.  Thus,  the necessary  fan and  ductwork are
considered part of  the process.  Costs  are presented for  three   levels  of
control  efficiency  based on  medium-  and high-reactivity dust.  For  a  given
collection  efficiency,   high-resistivity  dust   requires  a   greater   SCA
(specific collection  area)  and the cost  of the ESP is thus  increased.   For
purposes of estimating  equipment costs, plate area was calculated according
to  the Deutsch  equation with  particle  drift velocities of 0.036 m/s for
high-resistivity dusts  and  0.086 m/s  for low-resistivity  dusts.   Dusts  such
as  fly  ash  from  low-sulfur coal combustion  and  cement  kiln dust have  high
resistivity.   Fabric Filters.   Fabric filters are  commonly  used across  a
broad range of  exhaust gas  volumes.   Low-temperature and low-volume  exhaust
streams  from  conveyor  transfer  points  are  normally vented   to a  fabric
filter.  On  the other  hand,  high-temperature and high-volume exhausts  from
electric arc  furnaces  are also  often vented to a fabric filter.  Figures 7-9

and 7-10  present cost  curves  for a  variety of fabric  filter applications.
Costs are  presented for  filters  utilizing each type of bag-cleaning  mecha-
nism.   The cost  curves  assume  that the  fan  and drive are process equipment.
The control  costs include tie-in  ductwork,  a dust handling conveyor,  and  a
dust storage bin.   The  costs of thermal  insulation and heaters (necessary to
prevent  condensation  in  some   applications)  are not  reflected in the  cost
curves.    Separate curves  are   presented for  stainless  steel  construction.  Mechanical Collectors.   Capital  and  annualized cost curves for
mechanical collector systems are  shown in Figure 7-11.   System costs  include
hooding  to  capture the  exhaust at the  emission point,  ducting,  a fan and
drive, and  a dust  storage bin.   The  system cost  is based on carbon steel
construction.   Collection  efficiency  for  this type   of  system  generally
ranges from  80 to  90  percent, depending on the particle  size distribution
and inlet grain loading.   Incinerators.   Incinerators are  of  two basic  types,  thermal3
and catalytic.   Although  thermal  incinerators are  less costly from a capital
cost standpoint,  the  fuel  savings associated with  catalytic  units make them
attractive for compatible  exhaust streams.  Both types  of  units may recover
heat and thereby reduce  the  fuel  requirements.  The additional  cost  of the
heat  exchangers   must  be compared with  the  fuel  savings on  a case-by-case
basis.   Additionally,   the  use of  catalytic  incinerators  for  control  of
particulate  matter  is  limited  to substances  that  will   not blind or  poison
the  catalytic mesh.   Figure  7-12 presents  cost   curves  for   both types of
units, based  on  an exhaust stream at 25 percent of the  lower explosive limit
(LEL).   The  costs of units having a heat exchanger are based on a 35 percent
heat  recovery  rate.  Exhaust  streams that are  amenable to  incineration are
normally  exhausted  to  the atmosphere.   Thus  for purposes  of the cost curves
presented  herein, the  fan  and drive are  considered  process  equipment.   The
cost  curves  include the  cost   of  ductwork to  tie  the  incinerator  into the
process  vent system.   Venturi  Scrubbers.   Venturi scrubber use  ranges  from  control
of  small process fugitive  exhaust streams  to control  of  high-volume point
   Note:    Direct-fired  incinerators  considered  as thermal  incinerators  for
  purposes of this analysis.

sources  such  as basic oxygen  furnaces.   Figures 7-13 and 7-14  present  cost
curves  for  a variety of pressure  drops.   The costs include a clarifier ,and
circulating  pump for  the  scrubber  liquor,  a fan  and  drive,  and  ductwork
sufficient to tie the scrubber into the process exhaust stream.

     There  is  no single  method of estimating costs of control  of  fugitive
particulate emissions.  Because  of the great  variety of sources and control
methods, cost estimation must be specific to the method of control.
     Many industrial  process fugitive  emissions (IPFE) are controlled  with
the  same types  of equipment  used for  controlling process emissions.   The
three main  approaches to  control  of IPFE are ventilation, wet  suppression,
and optimization of operations.   Ventilation makes use of  hooding, ductwork,
enclosures,  and  control devices  such as baghouses  or  ESP's,   Use of similar
types of equipment, with  the possible addition  of  some  auxiliaries, permits
use of  capital  cost  estimating techniques previously described  and  the  cost
curves presented in Section 7.3.   The annualized costs  for  most ventilation
control   techniques  may also  be estimated  from  the curves in this  section.
     The costs  of  controlling  IPFE  with wet  suppression techniques  are not
amenable to the  estimating procedures described in Section 7,2 because these
control   techniques,  applied  to materials-handling  operations such  as  con-
veyors  and  unloading  stations, do not make  use  of the same types of equip-
ment  used  in   controlling process  emissions.  Their  cost depends on  the
amount of material  handled and the efficiency required.
     Table 7-4 presents reported cost data for spray and charged fog systems
applied to  rail  car  unloading and conveyor transfer stations.   The foam-type
spray system at a  single  transfer point would have a total capital  cost  of
about $18,000;  implementing such  a  system  handling 2.0 Gg of  material  per
year  at  an integrated  iron  and steel  plant  has  been  estimated  to  cost
$240,000.2  Initial cost  of the charged fog-type system at a single  transfer
point would be about $15,000.3
     Because  most  of the other  techniques  used  in  controlling  fugitive
particulate  emissions  do  not  require  large  capital  expenditures,  annual
direct operating costs  are presented in place of annualized costs.  Indirect
annual costs are not considered.



Rail car unloading station
(foam spray)
Rail car unloading station
(charged fog)
Conveyor transfer point
(foam spray)
Conveyor transfer point
(charged fog)
Initial cost,
January 1980
Unit operating
cost, January
1980 dollars
0.022 to
NR = Not reported.

a Reference 2.

  Reference 3.

  Based on use of 16 large devices at $8,000 each.

  Based on use of 3 small devices at $5,000 each.

     The cost  of  controlling fugitive dust emissions  is  generally  dependent
upon the method  of application, suppressant or stabilizer  used,  and desired
control  efficiency.    Wet suppression  and stabilization  are  the  two  most
common approaches to dust control.
     Table  7-5 presents  capital and operating costs  for wet  suppression  of
                                                       I, „   .     .     »•	
fugitive  dust  from  an  unpaved road,  a storage  pile,   and  a disturbed  or
unvegetated  outdoor  (exposed)  area.   The  costs  for the  unpaved   road  are
based  upon  the use of a  11.35-m3-capacity, nonpressurized  spray  truck oper-
ating  twice daily.  The  control  costs  for  the storage pile  are based upon
the  use of a  stationary,  elevated water  spray  system  including  sprayers,
piping, pumping,  wind  instruments,  and  installation  costs.   Operating costs
are  not  presented.  For  the  exposed area,  costs  are  based on  the  following
assumptions:   the equipment includes  piping  and  sprinklers, the  unit  is
moved  by hand,  and the application  rate is 0.473  m3 of water  per  minute with
an effective spray radius of 33.5 meters.2
     Table  7-6 presents  cost  data  for  application  of  stabilization  tech-
niques  to   var.ious fugitive  dust sources.   The three  sources are  again  an
unpaved road,  a  storage pile, and  an exposed area.   Methods of stabilization
are  oiling,  chemical   and  vegetative  stabilization, paving,  and  use  of
aggregate or chips.2
     Operating  costs for oiling  and chemical  stabilization  of the  unpaved
road surface  are  based  on  use of a  11.35-m3-capacity,  nonpressurized spray
truck  for  application.   The  capital costs for chemical  stabilization  of a
storage pile are  assumed to be the  same as  for wet suppression.2   The costs
of vegetative  stabilization  of exposed areas are  highly variable  as a result
of the wide range of climate and the physical and chemical  properties of the
soil.  Where  topsoil is  required, the  costs  may be  higher than those shown
in Table 7-6.
     Other  techniques  available for  control  of fugitive  dust  include street
sweeping,  vacuuming,  and flushing.    Table  7-7  presents  costs  for  these
control techniques.  The lower capital  cost figure for  sweeping  is based on
the  use  of a  trailer-type sweeper;  the higher capital cost  figure  is based
on use of  a self-propelled  unit with  a spray bar.   The  initial cost figure
for  flushing applies to a 1.135-m3  street flusher, excluding truck chassis.
Water  requirements are significant.5


Source method
Unpaved road- regular
Storage pile-regular
Exposed area-watering
> Estimated
efficiency, %
capital cost,
January 1980
operating cost,
January 1980
10-27 hectare
NR = Not reported.

Source method
Unpaved road-oiling3
Unpaved road-chemicals
(lignin or coherex)
Unpaved road- asphal tic
Unpaved road-oil (double
chip surface)
Exposed areas-oiling3
Exposed areas-chemicals
Exposed areas-paving,
Storage piles-surface
crusting chemicals
Exposed areas- vegetation6
jfficiency, %
capital cost,
January 1980
370- 1480/hec tare
operating cost,
January 1980
22,360/kmb .
NR '
NR = Not reported.
.  Reference 2.
  Based on monthly application.
. Based on resurfacing every five years and 15% opportunity costs.
  Based on reapplication every 2-3 years and 15% opportunity costs.
_ Reference 4.
  Dependent on type of vegetation planted, condition of soil, and climate.

                          FUGITIVE DUST SOURCES2
Source method
Paved road- sweeping
Paved road-vacuuming
Paved road-flushing
efficiency, %
Initial '
capital cost,
January 1980
operating cost,
January 1980
Cost per kilometer depends on nature of process and the site.


1.   Neveril,  R.  B.  Capital  and Operating Costs  of  Selected Air Pollution
     Control Systems.  EPA-450/5-80-002, December 1978.

2.   Bohn, R., T.  Cuscino,  Jr., and C. Cowherd, Jr.  Fugitive Emissions from
     Integrated Iron and Steel Plants.  EPA-600/2-78-050.  March 1978.

3.   Daugherty,  D.   P.  and  D.  W.  Coy.   Assessment of  the Use  of Fugitive
     Emission Control Devices.  EPA-600/7-79-045.   February 1979.

4.   Richard,  G.,  and  D.  Safriet.   Guideline for  Development  of  Control
     Strategies  in  Areas  with Fugitive  Dust  Problems.   EPA-450/2-77-Q29.
     October 1977.

5.   PEDCo Environmental, Inc.   Technical  Guidance for Control of Industrial
     Process  Fugitive Particulate  Emissions.   EPA-4SO/3-77-Q1Q.  March 1977.

                                   SECTION 8
                             EMERGING TECHNOLOGIES

     With the  growing awareness  of the environmental and health  impacts  of
participate  emissions,  there  has  been  a  need for  more effective  control
technology.   The  emergence of  new energy  production methods also  dictates
that participate control  devices  operate under more  stringent conditions  of
temperature, pressure, and  flue gas properties.  For these  reasons  there  is
an increasing  demand  for  advanced particulate  removal systems to  supplement
or replace the conventional control techniques.
     The  advanced  particulate  control   systems  have  been  developed  only
recently.  Most emerging  techniques are essentially hybrids  of technological
elements  associated with  first-generation systems.   This  section  examines
some of  the  novel  concepts being developed for control  of particulate emis-
sions.    Discussion  of  each concept is followed by a proposed overall  design,
together with data that indicate system performance.

     The  prevailing limitation of  conventional scrubbing technology is the
high energy  usage  required  to capture  submicron  particulate and  the asso-
ciated  high  operating  costs.   The limitations are  principally due  to the
inefficient  use  of available  energy and the  laws  governing particle behav-
ior.   In use  of  conventional  scrubbing  to capture  the  entrained particles
(which generally comprise less than 1 percent of  the stream's mass), energy
is applied to  the  complete mass of the stream (gas molecules and particles).
Furthermore,  exponentially  increasing  energy consumption   is  required  in
conventional  scrubbing to  capture progressively  smaller submicrometer-size
material.   The  combination of  these  two factors  can   render  conventional
particulate  scrubbing  noncompetitive with  other  conventional removal  tech-

     Many  industrial  processes  emit gaseous  and particulate-laden  streams
containing  corrosive,  sticky  materials.    For  these processes  conventional
nonscrubbing  techniques  (i.e.,  baghouses,  electrostatic precipitators,  and
mechanical  collectors)  are not applicable  for  control  purposes.   Conse-
quently,  environmental  concerns with  increasing  energy costs  combine  to
encourage development of new scrubbing techniques.
     The  following  subsections describe the technical and performance char-
acteristics  of  several  emerging  scrubbing  techniques.   Scrubbing  systems
that  are electrostatically  and hydrodynamically enhanced  are  shown  to  be
more  energy-efficient  and effective  in capture  of  submicrometer-size mate-
rial than their conventional counterparts.
8,1.1  AirPollution Systems, Inc.  Electrostatic Scrubber (ES)
     The  ES  system  is  basically an electrostatic  charger followed  by a con-
ventional  venturi  scrubber   (Figure  8-1).1   The   system   incorporates  a
patented  high-intensity  ionizer,  which  electrostatically  charges the par-
ticles in the gas stream before they enter a conventional  low-energy scrub-
ber.  The ionizer,  designed with a unique  electrode  configuration,  produces
a high  electric  field  strength (10  to 15 kV/cm) with  high  ion  densities.
These levels of  field  strength and  ion  density, higher than those  of  the
conventional  ESP,  effectively  charge  both  submicrometer-  and  supermicro-
meter-sized  particles.   The attractive  force between the charged  particles
and  droplet  is  additive  to  the  inertia!  and  other forces  acting  in  the
venturi   scrubber.   These  electrostatic  forces  account  for  the  enhanced
collection  of the  complete size  range  of available particles  in  the  gas
stream.    Figure  8-2 illustrates the higher removal efficiency of fine parti-
culate  with  the electrostatic scrubber  relative to  the  efficiency pf_a_
venturi  scrubber.*
     A field pilot  test  program with the UC  electrostatic  scrubber was con-
ducted on a  urea prilling tower.   The emission  stream contained 85  percent
submicron material  at  a  concentration of 91.5 mg/m3.  At  a  pressure drop of
19.0  cm W.C.  (7.5 in.  W.C.),  performance measurements  by plant  personnel
indicated an average overall  efficiency of  93.5 percent  and zero  opacity.

                                  INSULATOR COMPARTMENT
                                  VENT AND HEAT SYSTEM



            Figure  8-1.   APS electrostatic scrubber.1.









                                          MEASURED RESULTS

    0.01  0.02    0.05   0.1   0.2      0.5   1.0    2.0
                            — 90
                       5.0    10
            Figure 8-2.  Fraction efficiency performance of
                     APS electrostatic scrubber.1

Estimates of  electrical  operating  power  based on laboratory and  pilot  test
results, indicate 0.24  kW  per ms/s to convert  a  low-energy venturi  scrubber
into the  equivalent of a  high-energy  scrubber with  a UC ionizer  module.1
8.1.2  TRW Charged Droplet Scrubber (CDS)
     The CDS  system  passes water through small-diameter tubes,  and  electro-
statically atomizes  and  charges  the water droplets.2  The  droplets  range in
diameter from 60 to  250. jjmA and  have a  high surface charge  density.   The
charged  droplets are  immediately  exposed to  and interact with  slow-moving
(~2  m/s)  dust-laden  gas  streams  (Figure  8-3).   The  electric  field  in  the
droplet-particle  mixing  region  accelerates  the  charged  droplets  to  high
velocities (-^30  m/s).   These conditions of relatively slow-moving particles
and  fast-moving  charged  droplets   account  for high  collision  rates  due  to
enhanced inertial and electrostatic collection mechanisms.
     Successful  results  of  laboratory and  pilot field tests  with the  CDS
system were followed by  installation of a 51,000 mVh demonstration unit on
a  coke oven  battery.3  Emissions  consisted  of  fluctuating  concentrations
(114 to  755   mg/in3)  of submicron sticky hydrocarbons  and micron-sized high-
conductivity  carbon black.   Overall  removal  efficiencies  ranged from  91.0
percent to 94.3  percent,  and fractional efficiencies  ranged  from 80 percent
to  99  percent  (Figure  8-4) for  various inlet  loading and CDS  operating
conditions.   The CDS design summary is shown in Table 8-1.
     Low total energy  and  water consumption,  1.41 to 2.0 W per m /s and 0.11
to  0.13 liter/m3,  respectively,  were demonstrated  over  most  of  the  test
conditions.    Capital  and  annualized  costs are not available in  the litera-
ture,  but  several CDS  systems   have  been  purchased and installed on  indus-
trial  and municipal  incinerators,  and in the iron and steel and the pulp and
paper  industries in Japan.
8.1.3  University of Washington  Electrostatic Droplet Scrubber (UWEDS)
     The  UWEDS  system  involves  the  use  of  electrostatically  charged water
droplets  to   capture  suspended  particles  electrostatically  charged  to  the
opposite polarity  of the droplets.  The particles are negatively charged in
the  corona  section  and  flow into  a scrubber  chamber, into which positively
charged water droplets are sprayed.  The  scrubbed gas stream with entrained

                      HIGH VOLTAGE
                     ISOLATION TUBING

                                                                     SCRUBBING WATER
                                                                     SLURRY DISCHARGE
                                                                     TO SETUING POND
                 Figure  8-3,   TRW charged droplet  scrubber.'

                                          MEASURED RESULTS
   0.01  0.02     0.05   0.1   0.2      0,5    1.0    2.0

5.0   10
         Figure 8-4.  TRW charged droplet scrubber fractional
                        eff1ci ency performance.3

                        TABLE 8-1.  CDS DESIGN SUMMARY3
0    Three high-voltage scrubbing stages with 0.127-m collector plate spacing.
0    Flow cross-sectional area, 7.36 m2.
0    High-voltage electrode, type 316 stainless steel tubing, 19 mm diameter
     flattened to 12.7 mm.
0    High-voltage electrodes contained 67 spray tubes each on 44.5-mm centers.
0    Spray tubes, titanium with 1.27-mm O.D. by 0.15-mm wall, protruding
     25.4 mm from the electrode.
0    Collector plates 3.05 m long, 1.83 m high, 2.0 mm thick, mild steel.
0    Wall wash system covering each collecting surface.

droplets  then  flows  into  a  mist  eliminator  consisting  of  a  positively
charged  corona  section  that  removes  the  entrained  droplets.4   The  UWEDS
system is shown schematically in Figure 8-5.s
     Pilot plant  studies  were  conducted with a UWEDS mobile unit on an elec-
tric  arc steel  furnace  and  a coal-fired  power plant.   Overall  collection
efficiencies ranged  from  79.7  to 99.6 percent on steel furnace emissions and
from 99.6 to  99.99 percent on power plant emissions under various source iand^
control  device  operating conditions.   Fractional  efficiency ranged  from 90
to 99.99 percent  for fly ash particle  sizes  of 0.3 to 10  urn  (Figure 8-6).6
Further  results  from the power plant  tests  indicate that  the  UWEDS system
operates with  a specific collection area from  9.8 to 13.3 m2 per  m3/s, and
with water  consumption from  2.0 to 2.1 liters/m3.   Total  power consumption
is estimated at  0.8 kW per m3/s.5  No data have been published on annualized
costs for a full-scale system.
8.1.4  Steam Hydro Scrubber (SHS)
     The  Lone  Star  SHS  system uses  high  pressure  steam  to  move  the gas
through  the  system,  and to  enhance  particle  collection with  flux/force
condensation  (F/C).   (F/C  effects  are discussed  in  Section  8.1.6  .)   The
fast-moving  steam entrains  gas,  particles,  and  water droplets within the
mixing tube, where particle/droplet collision occurs.  The mixing tube

                                    8-8                                   '



        Figure 8-5.   University of Washington  electrostatic  droplet
                            scrubber schematic.5

                                            MEASURED RESULTS
     0.01  0.02     0.05   0.1   0.2      0.5   1.0   2.0

5.0   10
       Figure 8-6.  University of Washington electrostatic-droplet
               scrubber fractional efficiency performance.6

promotes  particle/droplet  collision  through   inertial  and  F/C  collection
mechanisms.   A shock-wave  pattern  is  created in  the  mixing tube,  further
enhancing  particle/droplet  collision by the induced turbulence.   Upon  exit-
ing  the mixing  tube, the  stream  is  accelerated  to  achieve more  complete
separation of the entrained  materials in the  cyclones.   Centrifugal  mecha-
nisms  in  the low-pressure-drop  cyclones account  for  the  impingement  and
separation of the  aerosols  entrained from the gas stream.   The SHS system is
depicted in Figure 8-7.7
     A  performance evaluation was  conducted on sa  commercial  SHS  system of
6.1  m3/s  capacity controlling  an open-hearth  furnace.7   Overall  collection
efficiency was measured from  99.78 to 99.95 percent under  different source
and  control   device  conditions.   Fractional  efficiencies  ranged from  70  to
99.99  percent for particle sizes  from 0.02  to 10 urn under  different condi-
tions  (see Figure 8-8).   Estimates of total power  consumption  indicate 395
kW per  m3/s  if waste heat is not available,  and M3 kW per m3/s if waste heat
is available  for use as steam,5
8-1.5  Two-Phase Jet Scrubber (TPJS)
     The  Aeronetics TPJS  system uses  a  nozzle  designed  to produce a two-
phase  mixture of  vapor and  liquid  droplets   when  fed pressurized,  heated
liquid.  The  droplets  are  initially accelerated  by  the pressurized delivery
and  become further  accelerated  by expansion  due to evaporation  of the hot
(200°C)  water.  Calculations  by the  developers  indicate  that  the  droplets
attain  supersonic  velocities  (~300 m/s).   The  intense  atomization and  rela-
tively  high  droplet  velocity  produce a  high  probability of  collision with
entrained  particles.   With  typical droplet diameters (<100  urn)  at the indi-
cated  supersonic velocity, the  impaction mechanisms would  be effective for
particles  as  small as 0.2 umA diameter.5  An  additional  benefit of the TPJS
system  is  an  induced draft, which eliminates or minimizes fan power require-
ments.   Figure 8-9  is a schematic showing two options of the TPJS system.8
     Performance  evaluation of  a commercial TPJS system was  conducted at a
7.5-MW   submerged  arc  ferro-alloy  furnace.8   Average  overall  collection
efficiency was  95.9 percent  under typical  furnace and  scrubber operating
conditions.   Fractional  efficiencies  covered  a  wide  range: 30  percent re-
moval  of particles  in the 0.03  to 0.10  urn size range and ~99 percent for

                      OUTLET SAMPLING
                  INJECTION WATER
         STEAM  NOZZLE
                                    ATOMIZER SLURRY
     Figure 8-7.  Lone Star Steel steam-hydro air cleaning schematic.7

     0.01   0.02     0.05    0,1    0.2      0.5    1.0    2.0

5.0    10
        Figure 8-8.  Lone Star Steel steam-hydro air cleaning
                 fractional efficiency performance.^

HOT     .
                      GAS AND
                      LIQUID OUT
                                  OPTION  1
                 XJET NOZZLE
               HEAT  EXCHANGER
           HOT GAS
                                  OPTION  2
           Figure  8-9.   Aeronetics  two-phase  jet  scrubber  schematic.8
                              (2  options  shown)

those  in the  0.5  to 10  umA range  (see  Figure 8-1Q).8   Estimates of total
power  consumption  are 41 kW per  m3/s  if  waste heat is not available and 3.8
kW per ms/s if waste  heat is available for use as steam.5
8.1.6  Flux Force/Condensation Scrubbing
     Flux  force/condensation  (F/C)  effects  are  those  that  accompany  the
condensation  of  water vapor from a  gas  stream;  they are generally caused by
contacting hot,  humid gas with colder  liquid  and/or by injecting steam into
saturated  gas.  Flux force/condensation  also  capitalizes  on  the  growth  of
particle  mass  and size  due to  condensation  of water  vapor  on  suspended
particles.    Particle growth  facilitates  the  collection  of  particles  by
inertia!  impaction.  In  practical  terms,  F/C scrubbing takes advantage  of
forces  acting on  the particles  induced  by a  temperature  gradient (thermo-
phoresis),  a vapor  condensation   gradient  (diffusiophoresis),   and  vapor
condensation  (Stefan  flow).   This advanced  scrubbing  method is adaptable to
various  scrubber configurations;  the ability  to  enhance particle collection
has  been demonstrated with venturi, sieve  plate,  packed bed,  and mobile bed
designs.9'10   Moreover,  as  particle size  decreases,  the advantages  of F/C
scrubbing  over conventional  systems becomes  greater  because  F/C collection
efficiency is virtually  independent  of particle size.
     A demonstration F/C  scrubbing system  was  built consisted  of a sprayr
type quencher, a sieve plate column,  and a spray-type cooling tower with an
induced  draft fan.11  The demonstration  was performed on a  secondary metal-
recovery furnace emitting particles with a  mean aerodynamic  diameter of 0.75
.umA.   The  12,000-m3/h demonstration system  collected 90 to  95  percent of the
submicron  stream at  a pressure drop of 68 cm W.C.  Under these source condi-
tions,  a conventional  high-energy  scrubber would  require pressure drops of
approximately 250  cm W.C.  for 90  percent  collection  efficiency and 535 cm
W.C. for 95 percent.

     Conventional  electrostatic precipitators  often cannot  effectively treat
particulate  materials that  have  high  electrical  resistivity  (i.e., greater
than 5 x 1Q10 ohm-cm).   The difficulty  arises  from the  presence  of highly
resistive  material on  the collection surfaces and  from the current density


                                              MEASURED RESULTS




2 0.05

£ 0.02




        92  2
        99  5


       0.01   0.02     0.05  0.1    0.2     0.5    1.0   2.0

                       AERODYNAMIC PARTICLE DIAMETER, ymA
5.0   10
                Figure 8-10.   Aeronetics two-phase jet scrubber
                      fractional  efficiency performance,.8

levels typical  of conventional  precipitators.   In the collected  dust  layer
the electric  field strength,  determined  as the  product  of  the  resistivity
and current  density,  may  exceed the dielectrical  strength of the  material
and cause  electrical  breakdown.   This  breakdown  generally leads to  condi-
tions that  cause the production of positive  ions from the  collected  mate-
rial.   These  ions  have  the reverse polarity of the ions  discharged  from the
corona  wires,  and consequently  they neutralize  and  degrade  the  intended
charging process.  This 'series of events  stemming from the electrical  break-
down is generally referred to as back corona or reverse ioni.zation.
     The difficulty  of  precipitating high-resistivity materials  limits  the
performance of  conventional  precipitators.  The  following subsections  dis-
cuss technologies  that  are being advanced as a result of  and  as a remedy for
this difficulty.
8.2.1  Pulse Energization
     The collection  of  high-resistivity dusts by  conventional  electrostatic
precipitators  can  be  substantially  improved  by  pulse  energization.   The
theoretical basis  of  the  pulse concept  was established over 30 years ago; in
recent  years,  investigators  in  the  United States,  Europe,  Australia,  and
Japan have developed this concept and brought it to the marketplace.
     Pulses  of  appropriate  duration and  frequency  superimposed on the DC
voltage  provide higher  peak  voltages,  reduce sparkover,  increase field and
diffusional  particle  charging,  improve   current   distribution,  and  permit
independent control of  secondary voltage and current.12  Resistivity-limited
dusts  (e.g.,  those  from  combustion  of some  low-sulfur  western  coals)  are
more easily precipitated  with pulse energization  than  by  conventional  means
because  higher and more  uniform ion densities  and  field  strengths prevail
when the electrical limit  is reached.
     Pulsed  energization  systems superimpose a high-voltage  impulse of very
short  duration and steep  wave  front on  an underlying,  relatively  constant
potential.   This  steady  base voltage  is  maintained  at  a reduced  level  to
sustain  the  migration  of  ions  and  particles  toward  the  collecting plates.
For  treatment of  high-resistivity  dusts,  the  base voltage may be set below
the normal  corona initiation voltage.  The high-voltage impulses momentarily
raise the  actual potential well above the sparking or back-corona limit that
would be experienced with  conventional energization.

     Pulse  frequency is  set in  the range  of  10 to  400  pulses per  second
(pps).   Higher  secondary levels  of peak voltage and  current  are  achievable
with  progressively  higher pulse  frequency.   Figure 8-11  shows  current-vol-
tage  curves  obtained with a special pulse-discharge electrode,  illustrating
the  achievement  of  higher corona points at increasing levels of pulse  fre-
quency.  Figure  8-12 shows  current-voltage curves of conventional  DC systems
and pulse  energization  systems  under the same  conditions.   The  conventional
current-voltage  curve  is steep,   indicative  of  back corona,  whereas  the
corresponding curve  with pulse  energization reflects higher  operating  vol-
tage without back corona.
     Limited but promising  results on the performance  and  economics of pulse
energization are now available  in the literature.13 1G  Laboratory and pilot
studies  have steered  this  emerging  technology into  successful,  short-term
full-scale  demonstrations.   Full-scale applications on a  35 MW   pulverized-
coal-fired utility boiler12  and a 290 ton/day rotary lime  kiln14 have demon-
strated  the performance  and economic advantages  of  pulse energization  in
collecting high resistivity dust (^5 x 1011 ohm-cm).
     These  two  full-scale demonstrations  were  accomplished by  retrofitting
conventional ESP's  with pulse energization  electrical systems.   Comparative
tests  on  both  emission  sources  were conducted  with  conventional   DC  power
supply and  a pulse  generator.   Results of collection efficiency  measurements
show a 1.3 to  1.5  factor of improvement by the pulse systems over the  con-
ventional  systems.12'14   The  improvement factor was appropriately determined
by  taking  the  ratio of the modified  migration velocity  values  calculated
from the performance data for both systems, by use of the  following modified
Deutsch-Anderson equation:
                    Pt = exp -(WR A/V)m                            (Eq. 8-1)

          P. = penetration
           A = collection area,  m2
           V = volumetric flow rate, m3/s
          w. = modified migration velocity, cm/s
           K                        ,         .         . .  .                 i
           m = exponent, using a common value for m = 0.5.

             g§  0.001
                                                    400 pps

                                                    200 pps
                                                    100 pps
                                                     50 pps

                                                     20 pps

                                                     10 pps
                            20    30    40    50
                       APPLIED VOLTAGE,  Up kV
 Figure 8-11.  Pulse energization voltage-current relationships for
                     various pulse frequencies.12

                                           DC SUPPLY
                                                ULSE SUPPLY
                      0     20    40    60    80
                      APPLIED VOLTAGE, UDC+Up kV

Figure 8-12.  Comparison of DC and pulse energization voltage-current
            relationships with same discharge electrode.12

     Additional indicators of  the enhanced performance associated with pulse
systems  are  a significant reduction  in  stack opacity and spark  rate,  and  a
large  increase in peak  operating voltage, both  relative to  performance  of
the conventional systems.
     With respect  to  potential  applicability  of pulsed energization systems,
the following should be  kept  in  mind:   1) the reported  improvement factors
represent  a  substantial  reduction   in  precipitator  size,   but  the  pulsed
systems  require more  sophisticated and expensive electrical  hardware;  2)  the
pulse  energization system  will   consume  more  electrical  energy unless  the
design  incorporates  an energy-conservation system;  3) the  reported results
are  limited  and  may  not be  representative  of   all   applications;  and  4)
further  development and  long-term demonstrations are needed  for  this  emerg-
ing control technology.   Japanese Pulse  Charging Systems.   Further  development  of  the
pulse  energization concept  has occurred in Japan.  A  "Bias-Controlled Pulse
Charging  System"  with a three-electrode configuration  has been  successfully
demonstrated  on large-scale  ESP's.16   This three-electrode configuration  and
bias-controlled pulse charging extend the applicable  limit  of precipitators
to  collection  of  dusts  with  resistivity  up  to  1015 ohm-cm.   This  novel
system also provides  advantages  in system stability (reducing process  varia-
tions  and upsets)  and  flexibility  over  a wide  range of  conditions.   The
three-electrode system has  also  been applied  to a  240  ton/h boiler plant
emitting  low-resistivity,  finely dispursed  carbon  particles.   The  relia-
bility  and versatility  of  the three-electrode bias-control led  system  have
been demonstrated  since  1976  on a space charge limiting  exhaust at about 97
percent  efficiency.   Another  novel   system was installed  in early 1978  to
treat  an 11,000  m3/min  exhaust  from an  iron-ore sintering  furnace.  With
dust  resistivities  ranging  from 1011  to 1013  ohm-cm,   a  quasi-pulse  ESP
system  successfully  collects  99 percent  of  the particulate  emissions.16   Other Japanese Developments.  Another Japanese  development  in
ESP technology is the  use  of wide-electrode  spacing design  (50 to 60 cm).
This design  requires less  plate  area and  allows operation  at  voltages  ap-
proaching  200  kV.    Full-scale   wide-spacing  ESP  units  have been used  to
effectively limit  sinter plant emissions  to  less  than 0.05  g/m3  at  resis-
tivity levels  in the  range of 1011 to 1013 ohm-cm.17

     Roof-mounted precipitators  have been  successful  in the  economical  and
space-effective control  of blast  furnace  emissions from steel  plants.   The
design  includes  vertical  flow to  allow natural  convection  for gas  draft,
collecting electrodes  made from conducting plastic plates, and  intermittent
water irrigation.18  Collection  of materials  with resistivities  in  the range
of  1011  to 1012 ohm-cm  is  maintained  at 95 percent efficiencies,  and oper-
ating  costs  are  14  percent  of  those  of  a   baghouse  treating  the  same
8.2.2  Two-Stage ESP Precharging
     One approach to the problem of precipitating high-resistivity  materials
is  to  separate particle  charging  from particle collection in a  staged man-
ner.  A  two-stage system  is  being developed by  Southern  Research  Institute
(SoRI) through sponsorship  by  the U.S. EPA.19   The SoRI  system  incorporates
a precharger with  a  novel electrode configuration and  energization  technique
in  the  first stage  (Figure 8-13),  and  a downstream collector with  a novel
corona discharge geometry in the second stage.
     The first stage is  similar to a  conventional  wire-to-plate design, but
has an additional  (screen) electrode to control back  corona.   The additional
electrode  consists of  an open  screen plate located close (1.9 cm) and paral-
lel to  the collection plate.   Operation  of  the  first  stage is similar  to
that  of  a conventional   ESP  section,  with the  additional   electrode being
energized  separately.    The screen  electrode  is  energized   with  the  same
polarity,  but  at a  reduced voltage relative  to  the  conventional  discharge
electrode.   The  downstream collector is  also  similar  to  a typical  ESP,
except that  an open-mesh  wire screen  is  used  as  a  corona  discharge elec-
     In  operation  of  the  first  stage, particle charging is  effected  in the
usual manner,  positive ions are produced,  and  then they are  captured on the
screen  electrode before they  drift into  the  active  charging  region.   The
screen electrode  does  not effect particle  charging, but  simply  collects the
positive  ions  being produced  from the collected  highly  resistive  material.
The biased negative  potential  of the screen electrode assures capture of the
positive  ions,  and  with  the openings in the screen electrodes allows passage
of  the negatively charged  ions and particles to the plate.

                                                                                  CERAMIC HIGH-
                                                                                  VOLTAGE INSULATORS
                                                                                   HOUSING (REF)

                                                                                  CORONA DISCHARGE ELECTRODE
                                                                                  SUPPORT AND BUS BAR
                                                                                  SCREEN ELECTRODE
                                                                                  SUPPORT AND BUS BAR
                                                                                  PASSIVE ELECTRODES
                                                                                  SCREEN ELECTRODES
             Figure  8-13.   Southern Research  Institute  precharger ESP
                                        assembly drawing.19

     Operation  of  the second  stage is  relatively  conventional.   The  open-
mesh wire  screen electrode was  selected on the basis of  laboratory  experi-
ments  to  produce   low  current  density with  high  electric  field  strength
values.   Since  particle  charging  is  essentially  completed  in  the  first
stage,   current  levels  are  reduced  in  the second  stage  to minimize  back-
corona formation and reentrainment of collected material.
     Laboratory  studies and pilot field tests  with  this  design  have  produced
results  that  correlate with  theoretical  predictions.   A  field test with  a
0.47 m3/s  pilot-scale  device was  conducted  on  a  utility  boiler  burning
low-sulfur  coal  and producing fly  ash with a resistivity  level measured at
2.0 x 1011  ohm-cm  at 135°C.19   Performance measurements  on  the pilot  two-
stage  system  showed  an  averaged  collection efficiency of  97.7  percent,  with
specific collection  area  equal  to  50.4 m2  per ms/s.   In  performance  tests
with the precharger off, emissions  were  seven  times greater  than with the
precharger  on.   Parallel  testing with a mobile,  pilot system of conventional
ESP design  operating under  the same conditions  indicated  a collection  effi-
ciency  of  91.9  percent.  The penetrating  emissions  from the  pilot  conven-
tional   system were  3.5 times those  from  the  two-stage  system.   Fractional
efficiency  data indicate 70  to  98  percent collection  of particles in the
size range  of 0.02 to 10 jjmA (Figure 8-14).
     The cost  of fabricating  such a two-stage  system is  estimated  to provide
a  savings  of approximately 40 percent over costs of a  conventional  ESP for
control  of  high-resistivity material.19
8.2.3  Flue Gas  Conditioning for ESP's
     Flue  gas  conditioning  involves the use of additive  materials  to control
particle  resistivity  for the   purpose  of  improving precipitator  perform-
ance.20  Conditioning  agents  may  also enhance  performance by  altering  non-
electrical  characteristics  of the flue  stream  (e.g.,  particle  agglomeration
and/or  adhesion).   Conditioning can  be considered as a control of  particle
resistivity,  and  includes  three  general  means  for control:    particulate
composition,  flue  gas  composition,   and  temperature.   The  combination  of
these  three factors  accounts  primarily for the resistivity levels  associated
with  electrostatic  precipitation.   Secondary  factors  to  be  considered  in
suspected   resistivity-related precipitator problems  are  thickness  of the

                                           MEASURED RESULTS




o 0.05

| 0.02



    0.01  0.02     0.05   0.1   0.2     0.5    1.0    2.0



                                                                      95   3



5.0    10
        Figure 8-14.   Sourthern Research Institute precharger ESP
                   fractional  efficiency performance.'*

collected  dust  layer,  the influence  of electric  field  strength, and  dust
aging characteristics.
     This section describes the  more popular means of  flue  gas conditioning
as  indicated in  published case studies.   Gas conditioning  is  not  a  new
concept, but  is  considered here  as an emerging  technology  for two principal
reasons:  (1) gas conditioning is a cost-effective and attractive option for
a growing  number of  new and  retrofit  ESP installations;  and  (2)  the  appli-
cability and  effectiveness of a  conditioning agent or  other means of  condi-
tioning  can  only be  verified on an  experimental, site-specific  basis,  and
not by predictive or referencing techniques.   S03 Conditioning.    The  ultimate  use  of  gas-phase  S03  in
streams  from  utility boilers  or from other industrial processes  fired  with
low-sulfur  coal  is  becoming   a  popular, nonproprietary  form of  gas  condi-
tioning.  Several commercial  means  of producing gaseous  S03  include  (1) use
and  combustion  of elemental  sulfur with appropriate processing,  (2)  use of
S02  with catalysis  to form  S03,  and (3)  purchase and  direct use of  S03.
Capital  investment,   operating  cost,  system  reliability,   and safety  con-
siderations  are  the main determinants  in  selecting an  S03  conditioning
system.   Injection  of  trace   quantities  of  S03  (3  to  30  ppm) in the  gas
stream  is  cost-attractive and effective in  reducing  resistivity by  1  to 2
orders of magnitude for ESP temperature conditions less than 200°C.
     Results  of an  S03 conditioning  study21  of  an  ESP-controlled  utility
boiler  burning  0.6  percent  sulfur coal  showed  an  improvement  of  overall
collection  efficiency from 91.3  to 98.8 percent.   Upon  injection of  25 ppm
of  S03  particulate  emissions were reduced  by a factor  of 8.3,  measured
resistivity dropped  from  6 x  1012 to 4 x 1010 ohm-cm  at  143°C,  and emission
of H2S04 vapor  (<1  ppm) was the  same as  without conditioning.  This success-
ful  S03 injection reduced particulate emissions in an amount equivalent to
expanding the existing ESP capability by a factor of about 3.
     Another  study22  on utility  boiler  burning low-sulfur  coal  also  demon-
strates  the effectiveness of S03 conditioning  on precipitator performance.
In  this  installation S02  is catalytically  converted to  S03  and  injected
downstream  of the  air preheater at a  rate  corresponding  to 32 ppm of S03 in
the  flue gas stream.   Stack  measurements  with the controlled condensation

method indicated S03  concentrations  of 10.9 and 8.1 ppm at the ESP inlet and
outlet,  respectively.   Mass loading  measurements  indicated that  S03  condi-
tioning  reduced  particulate emissions  by a factor of  2.4,  corresponding to
an  increase  in collection  efficiency from 79.2 to 95.4 percent  relative to
baseline conditions without S03 injection.
                                                      "n! -J
     Other case  studies  of S03 conditioning illustrate the effectiveness 'and
limitations  of this method.   Many cost estimates for  new utility installa-
tions  include  S03  conditioning  as an alternative control  method.   At least
two U.S. vendors of ESP's have indicated that the combination  of a cold-side
precipitator with  S03  conditioning  is  the  most  cost-effective  method  for
control  of boilers  burning low-sulfur western coal.23 25  Cost analyses show
savings  of  20  to  80  percent  in  annualized costs for  gas conditioning with
conventional  (cold) "precipitators, relative  to the  costs of  nonconditioned
cold precipitators, hot precipitators, and baghouses,26   Gas  Conditioning With Water-Soluble Alkali Compounds.   Certain
water-soluble  alkali   salts have  been  shown  to  be  effective  conditioning
agents  in  certain  industrial  processes.   Laboratory experiments and full-
scale  demonstrations  show  a sensitive relationship  between resistivity and
the  content  of  water-soluble  alkali  compounds  in  the collected material.
The following  paragraphs  deal  with case  studies in which  potassium sulfate,
sodium  chloride,  and sodium  carbonate  are  used as  conditioning  agents.
Note,  however,  that  the  descriptions  of these  materials as  conditioning
agents  do  not  imply  that they are  universally  applicable  conditioning
agents,  effective  on  any given resistivity-limiting process stream.   Rather,
the  discussion of  these  agents,  which are natural,  common, and/or  inexpen-
sive,  is meant  to  indicate the  sensitive relationship between  resistivity
and  trace  quantities of  readily  available  materials,  and  the  resultant
impact on precipitator performance.
     Potassium sulfate  (K2S04) was demonstrated as an  effective,  condition-
ing agent on an  800-ton/day cement kiln  process  in  Brazil.27   The K2S04 was
mixed  in water as  a 5 percent solution,  then injected, atomized,  and evapo-
rated  in the gas stream  ahead of the precipitator.   The  K2S04  solution was
injected at  a  rate of 500  liters/h,  a negligible (0.02 percent  K20)  amount
compared with  the  normal  content of  this constituent (0.4 percent  K20) in

the raw  meal.   Performance  measurements  showed an increase in  overall  col-
lection efficiency from  75.0 to 86.7 percent because  of conditioning,  corre-
sponding  to  a  twofold   reduction   in  particulate  emissions.    Resistivity
measurements indicate a  reduction  from 1013 to 1011  ohm-cm  at  300°C,  attri-
butable to K2S04 conditioning.
     Another full-scale  demonstration with K2S04 was  conducted  successfully
on a coal-fired  lime  kiln in South  Africa.27  A separate and parallel  demon-
stration  was  made with  sodium chloride.27   Each  conditioning   additive  was
put into  solution  in  the cooling water injected  into the kiln.   The results
of adding these  materials were an  increase in water-soluble  K20  from 0.02 to
0.25 percent,  and an increase  in water-soluble Na20  from 0.04  to  0.27  per-
cent,   in  the  precipitated  dusts.   Each conditioning  agent  reduced particu-
late emissions  by a  factor of 4 and resistivity by  2  orders of magnitude.
Note that sodium chloride was  an effective agent for this  process,  but may
not be applicable in  other industries because of incompatibility  with the
process, product, or materials of construction.
     Sodium carbonate has  been used as  a  conditioning agent  for  hot-side
ESP's  on  at least two boilers  fired  with  low-sulfur  coal.28 At one instal-
lation, a 15 percent  solution of Na2C03  is  used  as the conditioning medium;
at another, the conditioning  agent  is commercial-grade soda ash,  which is
pulverized  and  fed  pneumatically into the process  stream as a  dry material.
Both sodium carbonate systems  are  used to condition  process streams  in the
temperature range of  370° to 400°C, which is the range of the ESP treatment.
Particulate emissions have  reportedly been  reduced  by a  factor of 20 by
conditioning with  Na2C03 solution  and by  a  factor of  9 with solid Na2C03.
8.2.4  Development of High Temperature/High Pressure (HTHP) Electrostatic
     Concern over  the feasibility  of high  temperature/high pressure (HTHP)
electrostatic precipitation arose  as technological developments in advanced
energy-producing  processes  indicated the  need for HTHP  cleanup systems.  A
recent  laboratory  study  of  the characteristics of electrostatic precipita-
tion at  high temperatures and pressures has  yielded  substantial and encour-
aging  conclusions.29

     Bench-scale  experiments with  a  concentric  wire-pipe  ESP design  were
conducted at  temperatures  up to 1100°C (2000°F) and pressures up to 3550 kPa
(515  psia)  with  both  negative  and  positive polarity  energization.29   The
experimental  program included  the use  of  three gas  mixtures:   dry  air,  a
simulated flue  gas,  and a substitute (noncombustible) fuel gas.  The experi-
mental results  indicate higher particle-collecting efficiencies under condi-
tions  of high  temperature  and  pressure  than those  usually  achieved  with
conventional  precipitator  design and  conditions.   Higher  operating voltages
are  obtainable  and  are stable over the broad  temperature/pressure range of
practical interest.   These higher voltages provide  increased electric field
strengths and promote higher probability of particle collection.
     Experimental  results  demonstrate  that  potentially  higher  breakdown
                         l|                           '     .."J.." ."''.I
voltages are  achievable with the combination of high temperature with appro-
priately high pressure.  Concerns about achieving stable corona are resolved
with  the  understanding of  a  "critical  pressure"  concept.   The  critical
pressure is the lowest  level of elevated pressure at which corona initiation
and  sparkover .voltage  levels coincide.   Increasing the  temperature  rais.es
the  critical  pressure for  achieving stable corona  and consequently broadens
the  operating pressure  range.   The general  rule is that high pressure should
accompany  high  temperature.   Failure  to0 operate at  or above  the critical
pressure level  will   result  in sparkover or  breakdown  without the necessary
corona formation.29
     The critical  pressure concept is applicable to  both  positive and nega-
tive  polarity  systems.   With  positive discharge  systems,  this  critical
phenomenon  is  distinct and reproducible.   With negative discharge  systems
the  pressure  limit is not so well-defined,  but the systems generally respond
to  application  of the   critical  pressure concept.   At high  gas  densities,
higher  operating  voltages are  achievable with  negative than  with positive
     Filtration  technology is  being advanced through  incorporation of elec-
trostatic  designs,  new filtration  media,  and  novel  filtration  concepts.
Research has  shown that natural  or induced  electrostatic  phenomena can play

a critical role  in  the operating and collection  performance  of conventional
filtration systems.   Current  studies  are aimed at clarifying  the  effects of
electrostatic  charge  and electric  field on  the  collectibility and  cleana-
bility of conventional  filter media.   The advent of  new filter materials is
promoting the  use of  filtration technology to control  process  emissions to
which  conventional   filtration   has   been  inapplicable.    High-temperature
filtration,  currently  not economical  above 300°C,  is likely  to become  com-
mercially available  for .elevated  temperature control applications.   Future
requirements for control of  high temperature/high pressure  process  streams
are  being  evaluated with  new filtration concepts and requisite new  filtra-
tion media.  Filters made from ceramic materials are currently being evalu-
ated  and developed  for  such applications.   The following  subsections  de-
scribe these emerging filtration technologies.
8.3.1  Electrostatically Augmented Fabric Filtration
     In  electrostatically augmented  fabric  filtration,  particles are  pre-
charged  before crossing  a  fabric filter,  which  may or may  not be electro-
statically charged.   Figure  8-15  shows  a schematic  of  the pilot  unit built
by American  Precisions  Industries,  Inc., called the Apitron.   Air  enters the
precipitator section  from below,  then  passes upward through  the  tubes  of a
set  of parallel  wire-pipe precipitators, in  which  the particles are  charged
and  most are  precipitated.    The air  then  continues  into and through  the
bags,  where  final  filtration  takes  place.   Bag cleaning  is  initiated by
pulse jet  flow.   Studies on  silica dust  and various  other particulate emis-
sions30'31  indicate that the  Apitron  yields  high  collection  efficiencies
(Figure  8-16)  and that air/cloth ratios  can be  much  higher than those in an
uncharged filter of similar size.  For  example,  a  fabric filter utilizing a
pulse jet  cleaning  mechanism might require an air/cloth ratio of 5 ms/m2-min
whereas  an electrostatically augmented filter operates at an air/cloth ratio
of 14 m3/m2-min.36
8.3,2  Electrostatically  Augmented Filtration Through Fiber Beds
     The  concept of  electrostatically  augmented fiber  bed  filtration  was
studied  by Battelle Northwest on aerosols (NH4C1, Na20,  and MgO) having mass
mean  diameters less than 1 pm.32  The freshly generated particles  were first

                                                                             JET PULSE NOZZLE
                                                                               TUBE SURFACE
                                                                                COOLING WATER
                                                                                CORONA WIRE
                                                       DUST DISCHARGE
                                  WATER OUTLET
             Figure 8-15.   Apitron electrostatic-filter cutaway  view.

                                       MEASURED RESULTS
0.01  0.02     0.05  0.1    0.2     0.5    1.0    2.0

5.0   10
       Figure 8-16.  Apitron electrostatic filter fractional
                      efficiency performance.

drawn  through  a corona  charging  section and then through the  fibrous  beds,
which were  made  of stainless steel, polypropylene, or Teflon  and had a void
fraction of 0.96.   At velocities up to  1  m/s  through a bed  30 cm thick,  the
collection  efficiencies  were greater than 95 percent and pressure drops were
less than 1 cm H20,
8.3.3  Granular Bed Filtration
     The  general   term  "granular bed  filtration"  describes any  filtration
system  including  a bed  of  discrete granules or particles as  the filtration
medium.   To prevent  the particulate  matter  from plugging the  interstices
between granules and  causing excessive pressure drop, the device must incor-
porate  some means for periodic  or  continuous removal of particles  from  the
collecting  surfaces.   Fixed-bed filters are cleaned  periodically,  generally
by  a backwash of  air to blow the  dust out of the bed.  Moving-bed filters
are  cleaned continuously by replenishing the dirty bed with  new granules  and
separating the dust and granules by vibration.
     The primary mechanisms for  particulate collection  in a bed of granular
solids  are  inertia!  impaction,  flow line  interception,  diffusional  collec-
tion,  and  gravity settling.   Interception is the mechanism for particulate
collection  by  gas  convection,  but collection by this mechanism is negligible
with a clean bed.   As particles  are deposited in the interstices to  form  a
cake,  interception becomes  increasingly important as  the  bed porosity  and
flow channels  are reduced.   When too  much particulate  matter  is deposited,
the pressure drop  becomes excessive and the bed must be cleaned.
     A  series  of  studies was performed on the  pressure drops and particulate
collection  efficiencies  of  l.l-um~diameter latex  spheres  at  a  superficial
gas  velocity of  50  cm/s.33   The collection  bed  is made up  of iron  shot.
With a bed depth  of  3.2 cm at a  gas  velocity of  50 cm/s,  the collection
efficiencies  are   22  and 53 percent for  shot granules  of  620  and  490  umA
diameter, respectively.  Efficiency can be increased to nearly 90 percent as
the  bed depth increases, although  with a significant increase  in  pressure
drop.   Performance data on removal  of  fine  particles with  granular  bed
filters are  lacking for  industrial  applications,  especially at high  temper-
atures  and  pressures.  Some  preliminary studies  on  a  pilot-scale  fluidized
bed  combustion  unit  are inconclusive because  of  plugging of the filter  and

the  retaining  screens  that  hold  the  granular  beds.     Combustion  Power
Company  is  just  beginning a  series of  "cold"  filtration  experiments  on
moving granular bed filters, and results are pending.
8.3.4  Barrier Filtration
     Barrier filtration with fiber beds, woven fabrics, and porous materials
can  be  used efficiently  to remove particulates from gas  streams.   At high
temperatures  and pressures, however,  special  materials,  usually ceramics,
must be  used  to construct the bed matrix.  Ceramic fibers can be woven into
a  fabric,  packed  into  a  mat,  or made into sheets and  then used  to form
filter  "bags"  comparable  to  those  made of  conventional fabrics.   Their
particulate removal  characteristics  are comparable to those of conventional
barrier  filters  except  that the ceramic material can withstand higher temp-
eratures.   Barrier  filtration operates by three  mechanisms  of particulate
removal:  direct  interception,  diffusion, and inertia! impaction.    Parti-
cle  collection  efficiency increases as a dust cake builds up on the filter.
Collection efficiency can be  improved simply by making a filter bed thicker
(thus increasing the pressure drop) or by reducing fiber diameter.    Reduc-
ing  fiber diameter  is  a  desirable approach because it reduces filter weight
and  bed thickness.
     Tests  were performed  with  several  ceramic  media  samples to determine
penetration levels  of  dioctylphthalate (OOP) particles at air velocities of
5.5  and  16.5  cm/s under  ambient conditions.   The results indicated highest
efficiency  for  paper media, intermediate for felts, and low for woven mate-
rials.   Moreover,  the  efficiencies  of many of  the  samples were higher for
filtration  of OOP  smoke  than were  those  of standard  industrial-grade fil-
ters.   Collection  efficiencies  of the woven  ceramic materials  were poor,
probably because  of their more open  structure.   Paper media  in general had
poor mechanical strength  and did not  survive pulse cleaning.
     Initial screening  and  experiments indicate  that "blanket" ceramic fiber
materials  (felts)  consisting of small-diameter  fibers  (3 |jmA) are the most
promising  because of their combination of  good filtration  performance and
relatively  high strength.    A filter was made  of a layer of  Saffil alumina
blanket  insulation  material   approximately 1  cm  thick  contained between
stainless  steel (304)  screens.   This configuration was tested with  flyash

from  a  pilot scale fluidized bed  combustor.   Tests were made over  a  period
of  200  hours.   High  collection efficiencies (greater than 99 percent)  were
maintained with  an air/cloth  ratio of 9  m3/m2-min.S7  Further  studies  are
needed to  develop  more efficient cleaning techniques, to maximize  air/cloth
ratios, and  to  further demonstrate the effectiveness  and durability of such
     Porous  ceramic filters  are also under study.38  The most promising con-
figuration is a  ceramic cross-flow monolith (ThermaComb) produced  by  the 3M
Co.  This  material  is composed of  alternate layers  of corrugation  separated
by  thin filtering  barriers.   Limestone test dust with a mass  median diameter
of  1.4 mm  was  used to test  the filter,  and dust loadings were maintained at
levels from  2  to  7 g/m3.   At  .linear velocities  of 0.41 m/min, collection
efficiencies were  greater than  99 percent over  a  range  of  temperatures from
ambient to 970°K.   Thus,  porous  ceramic  filters are viable  as  barrier fil-
ters, and  their performance  should be studied  relative to  that of ceramic
fiber filters.

                                                                          t —	
     Magnetic separation  has long  been  recognized  as  a method  of  removing
magnetic materials  from  mixtures,  as in the separation of ferrous  minerals
from  ores.   Within the  last decade,  the  development of high-gradient mag-
netic separation  (HGMS)  techniques  has  enabled the  efficient separation of
submicron  particles of weakly paramagnetic materials from liquid streams at
high process rates.39   Generalized theory indicates that this technique  can
be  extended  to  the removal of  particles from  a  gas stream.   The fundamental
concept of HGMS  is the interaction of the  small paramagnetic particles with
a  ferromagnetic wire  in a  magnetic field  of  uniform background.40   The
ferromagnetic wire induces  regions  of  highly  nonuniform  field  intensity,
exerting a net  force  on  the particles  and causing  them  to migrate  to  the
surface of the  wire,  where they are  retained.   This process  is  analagous to
enhanced filtration  under a  magnetic field except  that the wire matrix is
much more  open.                                                            ,
     In its  most simple form, the  HGMS system consists  of  a  canister packed
with fibers  of  a ferromagnetic material  (steel  wool),  subjected to a strong

external magnetic field  (Figure  8-17).   The resultant strong magnetic forces
near the edges  of  the fibers provide very efficient collection of fine para-
magnetic particles.   The particles  are  removed from  the gaseous  stream  as
they pass  through  the canister.   The fiber matrix  eventually  becomes fully
loaded  and  must be  cleaned.   The overall particle  collection  efficiency  is
theoretically a function of  the applied magnetic field, filter  mesh param-
eters (fiber diameter and  magnetization, packing density, and length of mesh
in  the  direction of  flow),  particle parameters (diameter and  magnetic sus-
ceptibility), and  fluid  parameters  (viscosity and  superficial  velocity).41
     Most of the applications of HGMS on a commercial  scale  have been in the
kaolin  clay  industry, where  it  is  used in the  removal  of  weakly magnetic
color bodies  less  than  2  umA in  diameter.   Studies also have  been  made  on
fluid particle  systems  such  as  industrial waste  process  water  from steel
mills and  electroplating  operations, on  nuclear  reactor  coolants,  and  on
oils and hydraulic  fluids.   Potential has been  demonstrated  for application
of HGMS to the desulfurization of coal.
     Application of  HGMS to  particulate removal  from  gas  streams  has been
studied only recently.   A  study  was made on the particulate  control of emis-
sions from  basic oxygen furnaces  and electric arc  furnaces  used for steel-
making.42  The  collection  efficiency in the experimental-scale study is sum-
marized  in  Figure  8-18, which indicates that HGMS  is  effective  in removing
submicrometer  particles  of  relatively   high  magnetic  susceptibility  from
high-velocity  gas  streams.   Several  industrial  processes  in the  iron  and
steel industry  and  the ferroalloy industry produce  particulate  emissions  of
sufficiently high magnetic  susceptibility to make HGMS a potentially attrac-
tive control technique.  The same investigators made a rough economic analy-
sis  of  HGMS relative to the  ESP  and wet  scrubber, as  shown  in Table 8-2.
Although  capital  costs  of  HGMS are  relatively  high,  it  is  a  competitive
process overall  for  highly efficient (greater than 99.9%) particulate matter
     More  research  is needed to define optimum conditions  for  operation  of
HGMS, along  with a  better understanding of the fundamental  collection mecha-
nisms.   Both  experimental   and  theoretical  studies  should  be  done  for  a
variety  of  applications.   Magnetic  design,  matrix size, and  configuration
should  be  specified  for  each application, as well  as methods of cleaning the


              Figure 8-17.  High gradient magnetic  separator
                          schematic representation.117


                                             MEASURED RESULTS




2 0.05


£ 0.02




I      I        I

50  S

90  ~

92  |



95  ^



       0.01  0.02     0.05  0.1    0.2     0.5    1.0   2.0

                       AERODYNAMIC PARTICLE DIAMETER, ymA
                                   5.0    10
                Figure 8-18.   High-gradient magnetic separator
                      fractional efficiency performance.^

                         AND CONVENTIONAL TECHNOLOGY
Collection efficiency, %
Flange-to-flange cost, $ m3/s
$ per cfm
Power requirement, kW per m3/s
hp per 1000 cfm
 All estimates are referred to clean gas at 343°C, -1.5 kPa, 11 percent water
 by volume.  Source, Ref. 39.                           ,
matrix when  it is loaded with  particles.   It is reemphasized that  for HGMS
applications  the  particles  must  exhibit some  paramagnetic behavior.  As  a
consequence, this procedure  is  not widely applicable  to  many industrial  gas
streams.    It  is best  suited for gas streams containing  paramagnetic parti-
cles or  particles that  can  be  made paramagnetic by seeding  with  a  magnetic
material.                                                                 ;

     Particles in the  submicron range are especially difficult to control by
conventional methods.  Techniques  for increasing the  size  of submicron par-
ticles by  agglomeration  appear  attractive because the larger particles could
then be  removed more effectively by conventional methods.  Four such agglom-
eration  techniques  (thermal,  turbulent,  magnetic, and sonic)  have been pro-
posed.    Of  these, the  sonic appears the most promising and the most  advanced
in experimental and  industrial  demonstrations.   Magnetic agglomeration tech-
niques offer  potential  to  enhance collection, although application  is limi-
ted to  process emissions with  ferromagnetic particles.   Thermal  and turbu-
lent agglomeration techniques have been evaluated,43*44  but  are  not discus-
sed because of their limited potential for future development.
8.5.1  Sonic Agglomeration
     Sonic agglomeration techniques  have been researched for many years,  the
observation that  particles  behave  differently  under  the  influence "of sound

waves being  made in 1931.   The mechanisms of agglomeration  are  complex and
not well  understood;  as  many as  nine  possible mechanisms have  been postu-
     The  acoustic  field  affects  the  hydrody nami c  forces   and  vibrational
movements of the  particles within the flow field and thus accelerates colli-
sions between particles.   Reference  45 presents a thorough discussion of the
pathways involved.  Some of the theoretical and experimental  analyses allow
the following conclusions regarding sonic particle agglomeration:46
     1.   The optimum  frequency  for agglomerating fine particles (2 (jmA and
          less)  is  approximately 10 kHz.   Particles larger than 2 umA serve
          as  collecting  centers,  and  the agglomeration  rate  is  directly
          proportional to the number of such collection centers.
     2.   Highly polydispersed aerosols are easier to agglomerate than mono-
          dispersed aerosols  (i.e.,  if particles are all of nearly the same
          size, acoustic agglomeration is not effective).
     3.   Acoustic agglomeration rates vary in proportion to the square root
          of the acoustic intensity.
     4.   Physical properties of aerosol particles have comparatively little
          effect on acoustic agglomeration.
     5.   Residence times for significant agglomeration are about 5 to 10 s.
     6.   Sound  intensities  of  about 160  dB  are  required  for  effective
          industrial applications.
     7.   Water sprays can enhance sonic agglomeration.

     Table 8-3  shows  some results of  industrial  tests  with sonic agglomera-
tion (used in conjunction with other mechanical  collection devices).47
     The  requirement  of 160 dB sound intensities for sonic agglomeration can
be  translated into a  power requirement  of  about 1.6 kW per  ms/s,  which is
equivalent  to  about  750  kW  for  a modern  pulverized coal -firing  utility
station  (470 m3/s  volume  flowrate.)  Also,  the  intense sound levels create
noise  that must be muffled.  Recent studies indicate possibilities of lower-
ing  the   sound  intensity  requirements.48   Analysis  also  suggests  that
resonating  chambers producing  standing  waves  could  reduce  the  energy  re-
quirements.51   Thus,   acoustic  agglomeration  is considered  to be  the  most
promising of the agglomeration methods for treating industrial gases.

                         Table 8-3.   RESULTS  OF  INDUSTRIAL TESTS  WITH SONIC  AGGLOMERATION
Zinc oxide
zinc ore
Zinc oxide
Zinc oxide
from brass
Coke gas

           Table 8-3 (continued)
Carbide furnace

Carbide furance

Gas furnace

Aerosol and gas stream properties
nant 0.5
The sane


Gas furnace
gas black

carbon black

Hard coal





by weight,



















Agglomeration chamber
Type and
Experimental ,
reverse flow

The same, with
water addition
(5 g/ffl3)
Experimental ,
direct flow
1.1 dia, x 6.6

The same, with
water addition
Experimental ,
reverse flow
with water
0.5 dia.
rising steam
-0.29 dia. x

Experimental ,
reverse flow,
0,2 dia. x
Type of
s i ren



The same

















Length of







Collector system
«u It icy-
clones {in
The same

Two cyclones
1.3 in dia.
(in series)

The same

)ne or four
two cy-
and a
cloth fil-
ter (in
0.15 dia.

sound, X






wi thout
sound, X





99-98 (97)*

87 (97)b


          Table  8-3 (continued)

Sulfuric acid

acid fog

acid fog
Aerosol and gas stream properties



nant 7.5

by we ignt,









Agglomeration chamber

Type and
The same, 0.6
dia. x 6

Industrial ,
flow, 2.4 dia.
x 10.5 (2 sets)
Experimental ,
reverse flow,
0,64 dia. x 11

Type of
The same



F regency,






Length of



Collector system

clones (in
Two cy-
clones (in

Four cy-
clones (in
sound, 9!


sound, %



         Result without cloth filter, in parentheses,
         Result from water  addition, in parentheses.

     The emergence of  new energy-producing technologies may also necessitate
the use of  novel  particulate separation methods such as sonic agglomeration.
Sonic agglomeration  in  conjunction with a mechanical separator (cyclone) may
be  effective  under  these  conditions,  and  studies  are being  performed.49
8.5.2  Magnetlc Agglomeration
     Magnetic  fields  can alter  the motion of  particles suspended  in  a gas
stream,  depending on   the  magnetic  permeability  of the  particles.50   By
adjustment of  the  strength of the magnetic fields  to accommodate the nature
of the particles  and the flow field,  particle agglomeration may be enhanced.
Magnetic agglomeration  will  work best with ferromagnetic particles, although
theoretically  it  should  also be  effective with  charged particles  or par-
ticles having permanent or induced dipole moments.
     Applications to industrial  gas cleaning are limited bec'ause  long  resi-
dence  times  and strong magnetic fields  are  needed  even with ferromagnetic
materials,  and most  particulate  emissions  are  not ferromagnetic.   Direct
capture of  submicrometer particles under high-gradient magnetic fiels may be
more promising;  this concept  is discussed in further detail  in a preceding

 1.  Kearns, M.  T., et  al.   Union Carbide's High  Intensity Ionizer Applied
     to Enhance  a VentuH  Scrubber System.  In:   Symposium on the Transfer
     and  Utilization  of  Particulate  Control  Technology,  Vol.   3,  Denver,
     Colorado,   pp.  73-84.  EPA-600/7-79-044c.  February 1979.
 2.  Lear,  C.   W.   Charged  Droplet  Scrubber  for  Fine  Particle  Control:
     Laboratory Study.   EPA-600/2-76-249a.  September 1976.

 3.  Krieve, W.   F.,  and  J.  M.  Bell.   Charged Droplet  Scrubber  for  Fine
     Particle  Control:   Pilot Demonstration.   EPA-600/2-76-249b.   September
 4.  Pilot, M.  J. ,  and D. F. Meyer.   University of Washington Electrostatic
     Spray Scrubber Evaluation.  EPA-600/2-76-100.  April 1976.

 5.  Cooper, D. W.,  et al.  Evaluation of  Eight Novel  Fine Particle Collec-
     tion Devices.  EPA-600/2-76-035.  1976.

 6.  Pilot, M.  J. ,  and  G. A. Raemhild.   University of  Washington Electro-
     static Scrubber Tests at a Coal  Fired Power Plant.   EPA-600/7-78-177b.
     December 1978.

 7.  McCain, J. D.,  and W. B. Smith.  Lone Star Steel Steam-Hydro Air Clean-
     ing System Evaluation.  EPA-650/2-74-028.  April 1974.

 8.  McCain, J.  D.   Evaluation  of Aeronetics Two-Phase  Jet Scrubber.   EPA-
     650/2-74-129.  December 1974.

 9.  Calvert,  S.,  et al.   Feasibility of  Flux  Force/Condensation Scrubbing
     for  Fine  Particulate  Collection.    EPA-650/2-73-036.    October  1973.

10.  Calvert, S., and  S.  Gandhi.  Fine Particle  Collection  by a Flux Force/
     Condensation   Scrubber;     Pilot   Demonstration.     EPA-600/2-77-238.
     December 1977.

11.  Yung,  S. ,  C. R.  Chmielewski, and S.  Calvert.  Mobile Bed  Flux Force/
     Condensation Scrubbers.  EPA-600/7-79-071.  February 1979.

12.  Lausen, P.,  H.  Henrikson,  and H.  H.  Peterson.   Energy Conserving Pulse
     Energization of  Precipitators.    In:   Second Symposium  on the Transfer
     and  Utilization of  Particulate  Control  Technology.   (EPA  publication
     pending.)   1980.

13.  Belco  Pollution  Control Corporation.  Technical  Description  of  the
     Belco Pulsed Power Supply.  CPA 22-69-143.  March 18, 1970.


14.   Feldman, P.C., and  H.I.  Milde.   Pulsed Energization for Enhanced Elec-
     trostatic Precipitation  in  High-Resistivity Applications.   In:  Sympo-
     sium on the Transfer and Utilization of Particulate Control Technology.
     pp.  253-274.  EPA-600/7-79-044a.

15.   Penney, G.W., and P.C. Gelfand.   The Trielectrode Electrostatic Precip-
     itator  for  Collecting High-Resistivity  Dusts.   APCA Journal, Vol. 28,
     No.  1.  pp.  53-55.  January 1978.

16.   Masuda, S., et al.   Bias-Controlled Pulse Charging System for Electro-
     static Precipitator.  pp. 241-251.  EPA-600/7-79-044a.

17.   Ito,  R.,  and  K.  Takimota.   Wide Spacing E. P. is Available in Cleaning
     Exhaust  Gases from  Industrial   Sources.   pp.  297-305.   EPA-6QO/7-79-

18.   Ito, S., et al.   Roof-Mounted Electrostatic Precipitator.  pp. 485-495.

19.   Pontuis,  D.H.,  D.V.  Buhs,  and I.E.  Sparks.    Field  Evaluation  of  a
     Two-Stage ESP for High-Resistivity  Dusts.   (To  be  published in Staub
     journal.)  1980.

20.   White,  H.J.   Industrial Electrostatic Precipitation.   Addison-Wesley
     Publishing Company, Reading, Massachusetts.  1963.

21.   Dismukes,  E.B.,   and  J.  P,  Gooch.   Fly Ash Conditioning  with Sulfur
     Trioxide.  EPA-600/2-77-242.  December 1977.

22.   Patterson,  R. , et al.  Flue Gas  Conditioning  Effects on Electrostatic
     Precipitators.   In:   Symposium  on  the  Transfer  and  Utilization of
     Particulate Control  Technology, Vol.  I.   Denver, Colorado,   pp.   169-
     177.  PB 295, 226.  EPA~600/7-79-044a.  February 1979.

23.   Bubem'ck, D.V.   Economic Comparison of Selected Scenarios for Electro-
     static Precipitators and Fabric Filters.  In:  Section 14 Power Genera-
     tion .1.   Emission Control.   Presented at  the  70th  Annual  Meeting of
     APCA.  Toronto, Ontario, Canada.  June 1977.

24.   Atkins,  R.S., and D.V,  Bubenick.   Keeping  Fly Ash  Out  of  the Stack.
     Environmental Science and Technology,  p. 657.  June 1978.

25.   Harrison, M.E.   Economic Evaluation  of Precipitator  and  Baghouse for
     Typical Power Plant Burning Low Sulfur Coal.  Presented at the American
     Power Conference.  Chicago, Illinois.  April 24-27, 1978.

26.   Breisch,  E.W.  Method and Cost Analysis  of Alternative Collectors for
     Low Sulfur Coal Fly Ash.  pp.  121-129.  EPA-600/7-79-044a.

27.   Petersen,  H.H.    Conditioning  of  Dust with Water-Soluble  Alkali   Com-
     pounds,  pp.  99-111.  EPA-600/7-79-044a.

28.   Lederman, P.B.,  et  al.   Chemical Conditioning  of  Fly Ash for Hot-Side
     Precipitation,  pp. 79-98.  EPA-600/7-79-044a.


29.  Bush,  J.R.,   P.C.  Feldman,  and M.  Robinson.    Development  of  a High
     Temperature/High  Pressure  Electrostatic  Precipitator.   EPA-6QO/7-77-
     132.  November 1977.

30.  Helfrich, D.J.,  and T. Arimai.  Electrostatic  Filtration and the Api-
     tron/Design  and  Field  Performance.   In:  NovelI  Concepts,  Methods and
     Advaned  Technology in  Participate  Gas  Separation.   (T.  Arimai, ed.).

31.  Harmon,   D.L,    Electrostatically  Augmented   Particulate  Collection
     Devices.  In:   Proceedings  of  the First  Workshop of Particulate Con-
     trol.  Kernforschungsanlage  Julich Gmbh.  March  1978.

32.  Reid,  D.L.,  and  L.M.  Browne.   Electrostatic  Capture of Fine Particles
     in Fiber Beds.  EPA-600/12-76-132.  May  1976.

33.  Yung, S.C.,  et al.   Granular Bed Filter for Particulate Collection at
     High  Temperature  and  Pressure.   Air Pollution  Technology,  Inc.  Pre-
     sented at EPA/DOE Symposium  on  High Temperature, High Pressure Particu-
     late Control.  Washington, D.C.  September 1977.

34.  Hoke,  R.C.,  and  M.W.  Gregory.  Evaluation of a  Granular Bed Filter for
     Particulate  Control  in Fluidized  Bed Combustion.   Exxon Research and
     Engineering  Company.   Presented at EPA/DOE  Symposium on High Tempera-
     ture,  High   Pressure  Particulate  Control.  Washington,  D.C.  September
     1977.                                                              ;

35.  Wade,  G.L.   Performance and Modeling of  Moving Granular Bed Filters.
     Combustion Power  Company,  Inc.   Presented at EPA/DOE Symposium  on High
     Temperature,  High  Pressure  Particulate  Control.   Washington,  D.C.
     September 1977.

36.  Calvert,  S., and R.  Parker.   Effects  of Temperature and  Pressure  on
     Particle Collection Mechanisms:  Theoretical Review.  EPA-600/7-77-002.
     January 1977.

37.  Shack!eton,   M.A.   High Temperature,  High Pressure Particulate  Control
     with Ceramic Bag Filters.   EPA-600/7-78-194.  October 1978.

38.  Drehmel,  D.C.,  and  D.F.   Cilikerti.   High  Temperature  Fine  Particle
     Control  Using Ceramic  Filters.   Presented at EPA/DOE Symposium  on High
     Temperature,  High  Pressure  Particulate  Control.   Washington,  D.C.
     September 1977.

39.  Gooding,  C.H.,  T.W.  Sigmon, and  L.K.  Monteith.  Application  of High
     Gradient  Magnetic  Separation to Fine Particle Control.   EPA-600/2-77-
     230.  1977.

40.  Cummings, D.L.,  et  al.   Capture  of  Small  Paramagnetic  Particles  by
     Magnetic  Forces from  Low  Speed Fluid Flows.  AICHE Journal.   22:(569).
     1976.                                                         ~~

41.  Luborsky, R.E., and B.J.  Drummond.   High Gradient Magnetic Separation:
     Theory  vs.   Experiment.    IEEE  Transactions  or  Magnetics.   2:(1696).


                                    TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
                                                             3, RECIPIENT'S ACCESSION NO,
   Control  Techniques for Participate Emissions
   Stationary Sources - Volume  1
  September 1932
                                                             8. PERFORMING ORGANIZATION REPORT NO.

   U.S.  Environmental Protection  Agency
   Office of Air Quality Planning and  Standards
   Research Triangle Park, North  Carolina 27711
                                                             10. PROGRAM ELEMENT NO.
             11. CONTRACT/GRANT NO.

               68-02-3173  (Task No.  12)
               PEDCo Environmental.Inc.
   DAA  for Air Quality Planning  and Standards
   Office  of Air, Noise,  and  Radiation
   U.S.  Environmental Protection Agency
   Research Triangle Park. North Carolina 27711
             14. SPONS
                    :OR!NG AGENCY CODE
               EPA 200/04
    This  document is issued  per the requirements of  Section 108 of the Clean Air
    Act Amendments of 1977.
          Control Techniques  for Particulate Emissionsfrom Stationary Sources
     Volumes 1  and 2 present  recent developments of  control  techniques which  have
     become available since preparation of an earlier document entitled Control
     Techniques for ParticulateAir Pollutants  (AP-51).

          Volume 1 of this document presents available  data on characterization;
     sampling methods and analytical techniques for  particulate emissions;  oarticle
     behavior and characteristics;  types of participate control systems, their
     operating principles, design,  operation, and maintenance; costs and environmental
     consideration of particulate control techniques; and  emerging technologies for
     particul ate removal systems.  A major portion of Volume 1 presents information
     to quantify particulate  removal efficiencies by particulate size for the
     differing types of  particulate removal systems.
                                 KEY WORDS AND DOCUMENT ANALYSIS
                                               b.lOENTtFIERS/OPiN ENDED TERMS  C.  COSATI Field/Group.
     Control Techniques
     Emission Control
     Operation and Maintenance

19. SECURITY CLASS (This Report)

               21. NO. OF PAGES

                                                JO. SECURITY CLASS (Thispage)
                           22. PRICED
 EPA Fe'm 2220-1 (Rev. 4—77)   PREVIOUS ECITIOM is OBSOLETE
                                                     *U.S. CGYEIU0SENT PRICING OFFICE: 1982—539-061/3003



 United States                             Office of Air. Noise, and Radiation
 Environmental Protection           -   .,    Office of Air Quality Planning and. Standards
 Agency                                   Research Triangle Park NC 27711
Official Business                           Publication No. EPA-450 3-81 -005a    •'                                                                                     On*,,** *™<
Penalty for Private Use                                                                                                                                               Pels Kid
 *                                                                             •                                                                                     rees raiu
                                                                                                                                                                   EPA 335
                                                                      H vow *ddi«ss ttineomct, pttiMCiungt on tht atova Ubal.
                                                                      (Mr o'i, »«xJ return !o 0v« »bov !«»r*«
                                                                      M you do not tJevit tocontinu* r*c«inog inn t»cnnic»l r«pcwl
                                                                      »r,»i CHECK HERE 2! . l**r o'f l»6«l and i*!v