Office of Air and Waste Management
   Air Pollution Training Institute
     Control of Particulate Emissions
      Manual:TrainingCourse413
               APRIL  1975

-------
  Control of Particulate Emissions
                     Conducted by
            CONTROL PROGRAMS DEVELOPMENT DIVISION
               Air Pollution Training Institute
         Research Triangle  Park, North Carolina  27711

The Control of Particulate Emissions manual has been prepared
specifically for the trainees attending the course and should not
be included in reading lists or periodicals as generally available.

-------
                     \v/5
  &
'/
This is not an official policy and  standards
document.  The opinions,  findings,  and  conclusions
are those of the authors and  not  necessarily  those
of the United States Environmental  Protection Agency.
Every attempt has been made to represent  the
present state of the art as well  as subject areas
still under evaluation.  Any  mention of products,
or organizations, does not  constitute endorsement
by the United States Environmental  Protection Agency.

-------

?
    <
                     AIR POLLUTION TRAINING INSTITUTE
             MANPOWER AND TECHNICAL INFORMATION BRANCH
                 CONTROL PROGRAMS DEVELOPMENT DIVISION
             OFFICE OF AIR QUALITY PLANNING AND STANDARDS
77?e XUV Pollution  Training  Institute  (1) conducts training for personnel working on
the  development  and improvement of state, and local  governmental,  and  EPA air
pollution control  programs,  as well as  for personnel  in  industry and academic insti-
tutions;  (2) provides  consultation  and  other  training  ass/stance  to  governmental
agencies,  educational institutions,  '^-iustrial organizations,  and  others  engaged  in
air pollution training  activities; and (3)  promotes  the development  and improve-
ment of air pollution training programs in  educational institutions and state, regional,
and local governmental air  pollution control agencies. Much of the program is now
conducted by an on-site contractor, Northrop Services, Inc.

One of the principal mechanisms utilized to  meet the Institute's goals is the intensive
short term  technical training  course.   A full-time professional staff  is responsible for
the  design,  development, and presentation of these courses.  In addition the  services
of  scientists,  engineers,  and  specialists  from   other EPA  programs,  governmental
agencies,  industries, and  universities are used to augment and reinforce the  Institute
staff in the development and presentation of technical material.
Individual  course objectives  and desired learning  outcomes  are delineated  to meet
specific program  needs through  training.   Subject matter areas covered include air
pollution  source studies, atmospheric  dispersion, and air quality management.   These
courses are presented  in  the Institute's  resident classrooms  and laboratories  and at
various field locations.
Robert G. Wilder
Program Manager
Northrop Services, Inc.
                                                        i J. Schueneman
                                                    Chief, Manpower & Technical
                                                       Information Branch

-------
                                       Contents
                                CONTROL OF PARTICULATE  EMISSIONS
 Section _!_
 Motivation for Control Measures
 Section  2
 Introduction:  Collection Equipment
 Introduction:  The Collection of Particles
   from a Gas  Stream
 Section  3
 Conversion Factors
 Section  4
 Gas Properties   Basic Concepts
 Section  5
 Particle Settling Dynamics
 Terminal Velocities of Spherical Particles
 Section  6
 The Separation of Particles into Size-
   Fractions
 Section  7
 Notes on the Analysis of Particle Size
   Distributions
 Section  8
 The Effective Particle Size
 Section  9
 Representation of Particle-Size  Data
 Statistical  Presentation of Data
 Size-Efficiency  Curves
 Section  10
 Settling Chambers
 Section 11
Cyclones
Section 12
Miscellaneous Dry Inertial-Type Collectors
Section 13
Wet Collectors:   Introduction
Section 14
Collection of Particles  on  Cylindrical and
  Spherical  Obstacles
Section 15
The Gravity Spray Tower
Section 16
Venturl Scrubbers
Section 17
Collectors with  Self-Induced Sprays
Section 18
Impingement  Type Scrubbing  Tower
Section 19
Wet Centrifugal  Collectors
Section 20
Wet Dynamic  Predpitator
Section 21
Disintegrator Scrubbers
Section 22
Fabric Filtration
Section 23
Fabric Filtration -  Basic Concepts
Section 24
Fabric Filtration Operations and Industrial
  Applications
Section 25
Fabric Filtration -  Mathematics of Bag-
  Filter Operation

-------
   Contents
Section 26

Effect of Changing Permeability, Varying
  Flow Rate, and Non-Laminar Head Loss

Section 27

Electrostatic Precipitators Operation and
  Industrial Applications

Section 28

High Temperature Gas Cleaning

Section 29

Sanitary Disposal of Collected Material

Section 30

Cost of Collection Equipment

Section 31

The Sylvan Chart

Section 32

Class  Problems

-------
SLTTIGN 1
Motivation for Control Measures

-------
                  MOTIVATION FOR CONTROL  MEASURES
 I   MOTIVATION FOR CONTROL
 MEASURES IN INDUSTRY

 A  From the standpoint of the industrial plant,
    air pollution control is motivated by:

    1  The economic value of recovered
      material,  or the value of the cleaned
      gas,

    2  The desire on the part of management
      to maintain satisfactory relations both
      with the workmen and the neighbors  in
      the surrounding community, and

    3  The legal requirements as  covered by
      ordinances and  statutory regulations.
II   THE ECONOMIC VALUE OF
 RECOVERED MATERIAL, OR THE VALUE
 OF THE CLEANED GAS

 A  Industrial management recognizes that
    production results in two kinds of materials:

    1  Saleable products, which are a source
      of profit.

    2  Waste products, which are a charge
      against production.

      a This charge  against production is
         increased by the cost of abating
         airborne wastes.
 B  The Economic Value of Recovered Material

    1  The cost of air cleaning equipment, in
      some instances, is paid for from
      salvaged material even though the pri-
      mary reason for the installation is the
      prevention of an air pollution problem
      to the plant or neighborhood.

      a  Some examples are:

         1) Flour dust in bakeries
         2) Brass griding dust in metal
               finishing
          3)  Ore dust from crushing and
                milling
          4)  Line dust from kilns
          5)  Sugar dust  from dryers and
                coolers
       In some cases, the collection equip-
       ment is of primary importance,  not
       to air pollution prevention,  but to the
       economic operation of the manufacturing
       process itself.  However,  satisfactory
       performance of the collector is of
       benefit both from a manufacturing and
       air pollution standpoint.

       a  Some  examples are:

          1) The manufacture of carbon black
          2) The sintering and roasting of
                lead ores
          3) The pulverizing of chemicals
       In many cases, the installation of
       collectors cannot be justified from an
       economic standpoint.

       a  Any justification at all is then only
          on the basis of establishing good
          labor  and community relations.
Ill  LABOR AND COMMUNITY RELATIONS

 A To a large extent, the degree to which
    particulates must be removed from a
    flue gas before emission to the  atmosphere
    is governed by labor  and  community
    relations.

 B The workmen and the community are
    concerned with:

    1  The deposit of coarse particles which
       settle in the general area of  the stack
       and create problems  of a localized
       nature.

       a  Such settlement is mostly a nuisance
          to the neighborhood property, but
PA. C. pm. 17a.9.60
                                           1

-------
Motivation for Control Measures
         may also cause damage by acceler-
         ating  or inducing oxidation or
         chemical attack.

      b  Examples of coarse emissions:

         1) The foundry cupola and  lime
              hydrator.  Settled material on
              roofs cause endless main-
              tenance in material removal
              and roof repair.
         2) The aggregate dryer,  the cement
              kiln,  the stoker-fired boiler,
              the pulverized fuel boiler.
              Coarse emissions settle on
              porches,  ledges, and laundry
              over an extended area from
              the point of emission.
         '.'•) The exhaust of ferrous grinding,
              heat treat furnances, plating
              and caustic tanks, can dis-
              charge materials that damage
              the finish of automobiles.
          Good housekeeping on the plant
          grounds, both inside and outside
          has an aesthetic effect on the work-
          men and members of the community.
  C The cumulative effect of all the above on
    labor and public relations cannot be
    overestimated.

    1  When it is understood that plant manage-
       ment is making a continuous effort to
       remedy or improve air-polluting
       emissions,  and working conditions are
       relatively clean, both the employee
       and the community are inclined toward
       the company.
IV  LEGAL REQUIREMENTS

 A Regulations Limiting the Opacity of the
    Plume
      The deposit of fine particles that
      travel for long distances in the atmos-
      phere contributing problems of an
      area-wide nature.

      a Such particles can be  a nuisance,
        a health hazard,  reduce visibility,
        and  cause soiling and  damage to
        materials and vegetation.

      b I'CxanipJe.s of fine particle emissions:

         I) Molting operations which emit
              fumos such as zinc oxide and
              ferrous oxide.
        2) (' irbonaceous matter from burn-
              ing of coal,  oil, gas,  and
              rubbish.
     The plurr.e appearance

     a  A "dirty" plume has a psychological
        effect on those viewing it.
  4  The cleanliness of the plant area,
     both inside and out.
     1  Nearly every air pollution control
       ordinance now existing has its origin
       in smoke control regulations which
       were designed to limit the density of
       stack emissions by placing restrictions
       on the  opacity of the plume.

    2  The  smoke regulation concept has
       carried over into present day air
       pollution ordinances,  in that  some
       restrictions on the opacity of industrial
       stack gases are specified.

    3  Examples of ordinances which recognize
       the Ringelmann Chart as a visual gauge
       of permissible visible emission are
       shown  in Table 1.

    4  At present,  there are no legal require-
       ments  that effluent gases be free from
       all visible contaminants and the
       possibility of such restrictions in the
       foreseeable future is remote.

       However, from the standpoint of public
       relations, a stack discharge  containing
       sufficient visible contaminants to be
       conspicuous should be avoided wherever
       possible.

-------
                                                 Motivation for Control Measures
       Table 1.  DEFINITION OF ORDINANCE CLASSIFICATIONS
                    BASED ON SMOKE DENSITY*
CLASS I:
Ordinances which allow only Ringelmann No.  1
smoke:
A  With no exception stated;
B  Except for stated periods when cleaning or
   building fires or other reason.
CLASS II:
CLASS III:
CLASS IV:
CLASS V:
CLASS VI:
Ordinances which allow short periods of No. 2
smoke:
A  But may not exceed  No.  3 for stated  periods
   for fire cleaning or  building;
B  And may exceed No. 3 for stated periods for
   fire cleaning or building.

Ordinances which allow No. 2 smoke at all times:

A  But may not exceed  No.  2 at any time;
B  But may not exceed  No.  3 for fire cleaning or
   building;
C  And may exceed No. 3 for fire cleaning or
   building.

Ordinances which allow short periods of No. 3
smoke:
A  But may not exceed  No.  3  at any time

B  And may exceed No. 3 for fire cleaning and
   building.

Ordinances which allow No. 3 smoke at all times:
A  But may not exceed  No.  3 at any time:
B  And allow periods in excess of No. 3 without
   specifying or in addition to fire cleaning or
   building.
C  And may exceed No. 3 for fire cleaning or building.


Ordinances which define smoke density  using the
Umbrascope and not in terms of the Ringelmann
Chart
CLASS VII:
Ordinances which do not, or only loosely, define
the smoke density prohibited.
 Air Pollution Abatement Manual,  Manufacturing Chemists Assoc.   1953.

-------
Motivation for Control Measures
B  Regulations  limiting the emission of
   particulate matter in terms of weight
   units.

   1  In some areas, attempts have been
      made to reduce allowable stack
      emissions to a quantitative basis.  Thus,
      ordinances have  been enacted which in-
      clude restrictions on either the con-
      centration (weight per unit volume) or
      mass rate (weight per unit time) of
      emission of pollutant.

   2  Legal requirements concerning parti-
      culate emission vary over a wide range
      of operating conditions,   (see "Selected
      Dust Emission Limitations of Typical
      Communities in Various  Population
      Groups, " in Appendix)

   3  Examples of dust loading prohibitions

      a 75% collection entering collector,
        minimum

      b 85% collection entering collector,
        minimum

      e 90% collection entering collection

      d 0.85 Ibs. per 1000 Ibs. gas or air

      e 0.85 Ibs. per 1000 Ibs. of gas
        adjusted to  12% CCX
                          iL

      f  0.85 Ibs. per  1000 Ibs. of gas - 50%
        cxeess air

      g 0.85 Ibs. per ]000 Ibs. of gas
        adjusted to  50% excess air.  Maxi-
        mum 0.5 Ib.s.  per 1000 Ibs. of gas
        shall be  larger than 325 mesh.

      h 0.55 Ibs. per  1000 Ibs. of gas
        adjusted to 50% excess air.  Maxi-
        mum 0.2 Ibs.  of dust larger than
        325 mesh.

      i  2  Ibs. per 1000 Ibs.  of gases  at
        12% CO   must collect 75%
               tL

     j  0.30 gr/cu.ft.  at 500°F and 50%
        exress air
k  0.30 gr/cu.ft. at 500°F and 50%
   excess air not to exceed 0. 2 gr/cu.
   ft.  larger than 325 mesh.

1  0.425  gr/cu.ft. at 500°F and 50%
   excess air.

m 0.75 gr/cu.ft. and 50% excess air
   of which not more than 0.4 gr/cu.
   ft.  shall be larger than 325 mesh.

n  0. 75 gr/cu. ft. at stack temperature,
   not more than 0. 2 gr/cu. ft.  retained
   on a 300 mesh U.S. Standard sieve.
   Excess air not to exceed 50% at
   full load.

o  0. 75 gr/cu. ft. at 500°F and 50%
   excess air of  which not more than
   0. 2 gr/cu.ft.  shall be larger than
   325 mesh.

p  0.75 gr/cu.ft. at 500°F and 50%
   excess air of  which not more than
   0. 2 gr/cu.ft.  with gas at 850°F
   shall be larger than 325 mesh.

q  Process weight table (see ''Selected
   Dust Emission Limitation of
   Typical Communities in Various
   Population Groups" in Appendix)
For compliance to laws,  the engineer
must be able to determine the con-
centration of particles being emitted
from an industrial operation under a
variety of specifications including
temperature,  pressure, carbon dioxide,
excess air, pounds of flue gas, certain
particle sizes, and efficiency of
collection.

Shortcomings of laws specifying per-
missible concentrations in terms  of
grains per cubic foot.

a  In industrial problems, dilution with
   room or outside air  is a possibility
   which could produce "permissible"
   concentrations  for many manufactur-
   ing processes without any collection
   equipment.

-------
                                                                Motivation  for Control Measures
       b  Examples

          1)  Electric furnances:  If the furnace
                is ventilated by roof monitors,
                wall exhausters,  and the like,
                an  exhaust volume of 100, 000
                cfm may be required.  If the
                furnace has a local exhaust
                hooding attached to the roof
                ring, metal fumes and smokes
                may be exhausted by 20, 000
                cfm making collection of
                pollutants more practical with
                the lower gas volume.   Yet
                the removal of 80% of the solids
                with a collector in the second
                case would have  emission
                concentrations in gr/cu.ft.
                the same as the case of the
                larger ventilation where no
                collection equipment is
                employed.
          2)  Foundry cupola:  A  foundry cupola
                with an open charging door may
                emit 15, 000 cfm; one with  a
                closed top bell charging method,
                3, 000 cfm.  Pounds of solids
                released  during the melting
                operation will be identical,
                pounds  of solids  collected by
                the air  cleaning device the
                same,  as well as pounds of
                solids discharged to the atmos-
                phere.  Yet the grains/cu. ft.
                will be  5 times greater in the
                top bell charge arrangement
               due to the drastically reduced
                gas volume involved.

V  SELECTION OF COLLECTION
EQUIPMENT

A  In the  field of air pollution, it is necessary
   to consider economical design that  will
   reduce a health hazard, a damage potential,
                 (*-> below the threshold of
or a nuisance,
complaint, or
law.
                    below that established by
     It is not necessary to design control
     equipment to remove all particles
     with complete efficiency unless pro-
     fitable recovery of as much valuable
     material as possible is the goal.

      Following such a concept  allows  the
      air to function to its useful capacity
      as a waste disposal medium.
B  The present dynamic state of air pollution
   abatement appears to lean in the direction
   of more stringent control in the future.
   This may mean the need of  attaining
   higher operating  efficiencies of control
   equipment in the  future than is accepted
   at present.

   1 One safe  recommendation in equip-
     ment selection is (reference 4):  Select
     the collector that will allow the least
     possible amount of contaminant to
     escape and still be reasonable  in first
     cost and maintenance.  However,  for
     some applications, even the question
     of reasonable  cost and reasonable
     maintenance must be sacrificed to
     meet established standards for air
     pollution  control or to prevent  damage
     to health  or property.

REFERENCES

1  Rogers, S. M.  "A  Review and Appraisal
     of Air Pollution Legislation in the
     United States.  "  Presented at the
     Golden Jubilee Meeting,  APCA, St.
     Louis,  Mo.  June 4,   1957.

-------
Motivation for Control Measures
 2  Air Pollution Abatement Manual.  Mano=_.      5  Grindle, A.J.  "The Cupola Emission
      facturing Chemists Association,  Inc.              Problem and Its Solution. " Presented
      1952.                                            to Semi-Annual Meeting,  East Central
                                                       Section, APCA, Harrisburg.  Sept.  25,
 3  Omara,  R., and Flodin, C. R.  "Engineer-           1953.
      ing Design Factors in Dust and Fume
      Recovery Systems. "  JAPCA 8, No.  1
      May 1958.
                                                  6  Kane,  J. M.  "High Temperature Gas Clean-
 4  Kane,  J. M.  "Operation and Effectiveness            ing. " Paper presented to Air Pollution
      of Dust Collection Equipment. "  Heating           and Smoke Prevention Association.
      and Ventilation.  Aug.  1952.                       Roanoke.  May 7, 1951.

-------
                                                2
SECTION 2
Introduction:   Collection Equipment
Introduction:   The Collection of Particles
  from a Gas Stream

-------
                  INTRODUCTION:  COLLECTION EQUIPMENT
 I  TWO BROAD CATEGORIES OF
    COLLECTION EQUIPMENT

 A  Collection equipment may be divided into
    two broad categories:

    1  Dry collectors

    2  Wet collectors


II  DRY COLLECTORS

 A  The  dry collectors fall into the following
    classifications:

    1  Settling chambers

    2  Centrifugal separators

       a  Dynamic precipitators

       b  Cyclone

          1)  Simple cyclone (large diameter)
          2)  High efficiency (long cone)
          3)  Multicyclone

    3  Inertial separators

       a  Baffle chamber

       b  Impingement type

       c  Louver type

    4  Fabric collectors

    5  Electrostatic precipitators


III  WET COLLECTORS

 A  The wet collectors fall into the following
    classifications

    1  Gravity spray tower

    2  Wet impingement scrubber
    3  Self-induced spray deduster
      (orifice type)

    4  Disintegrator

    5  Wet Dynamic precipitator

    6  Venturi scrubber


IV  FACTORS IN THE SELECTION AND
    DESIGN OF COLLECTION EQUIPMENT

 A  Carrier Gas Properties

    1  Temperature

    2  Pressure

    3  Humidity

    4  Density

    5  Viscosity

    6  Dewpoint for condensable components

    7  Electrical conductivity

    8  Corrosiveness

    9  Toxicity

   10  Flammability


 B  Particulate Properties

    1  Particle size and size distribution

    2  Particle shape

    3  Particle density (absolute and bulk)

    4  Stickiness, build up tendencies, and
      flowability

    5  Hygroscopic properties
  PA. C. pm. 19. 9. 59

-------
Introduction:  Collection Equipment
   6  Agglomeration tendencies and floe
      stability (dispersibility)

   7  Electrical conductivity

   8  Corrosiveness

   9  Flammability

  ]0  Toxicity

  11  Abrasiveness

  12  Flowability


C  Conditioning

   1  The actual deposition efficiency of a
      collector may be modified by condition-
      ing the carrier gas stream, the partic-
      ulates,  or the collecting surface.

   2  Conditioning of the particle

      a  Condensation on the particle surface

      b  Klocculation of the  particles

         1) Natural
         2) Mechanical
         3) Electrical
         4) Sonic

      c  Deposition of solids on liquid
         droplets

      d  Treatment of the particle surface

      e  Electrical charging of the particle

      f

      g

   'j   Conditioning of the carrier gas stream

      a   Heating, cooling

      b   11 um id if ieation
   4  Conditioning of the collecting surface

      a  Viscous substances

      b  Irrigation

      c  Electrostatic

      d  Heating,  cooling

      e

      f


D  Manufacturing Process Factors

   1  Volumetric gas rate collector must
      handle

      a  (Retention time required in the
         collector)

      b  (Velocity through the collector)

   2  Particle concentration collector must
      handle

   3  Permissible  pressure drop across
      the collector


E  Collector Operation Considerations

   1  Maintenance

   2  Continuity of operation.  (Must it be
      shut down and started up? Must it
      take varying  loads? etc.)

   3  Safety and health protection

      a  Toxic hazard

      b  Explosion and flammability hazard

   4  Type of labor required and availability
      of  such  labor

   5  Disposal of the collected  material

      a  Waste disposal

      b  Product recovery

-------
                                                         Introduction:  Collection Equipment
F  Construction and Installation Factors

   1  Floorspace requirements

   2  Headroom requirements

   3  Availability of utilities

      a  Water

      b  Steam

      e  Compressed air
      d  Electricity  (A. C. , D. C. ,  high
         voltage)

   4  Auxiliary equipment

      a  Fans, blowers,  compressors

      b  Pumps

      c  Motors and drives

      d  Shaking and  rapping devices

      e  Conveyors,  air lacks, rakes,
         storage bins, etc.

      f  Cleanout ports,  access doors,
         explosion doors,  etc.

      g  Electrical substation  or transformer

      h  Timer,  alarms, etc.
      i   Sludge tanks, treatment tanks,
         agitators, etc.

      j   Valves,  dampers, automatic valves,
         regulators

   5  Materials of construction

      a  Weather  protection requirements

      b  Insulation or jacketing requirements

      c  Pressure requirements

      d  Temperature limitations

      e  Corrosion resistance

      f  Erosion  resistance
REFERENCES

1  Lapple, C. A.  "Dust and Mist Collection, "
      Air Pollution Abatement Manual,
      Manufacturing Chemists Association,
      Inc.  1951.

2  Perry, J. H.  Chemical Engineers'
      Handbook,  McGraw-Hill Book Co.
      Inc. N. Y.   1950.

-------
            INTRODUCTION:  THE  COLLECTION OF PARTICLES
                                    FROM A GAS  STREAM
 I   PHASES INVOLVED IN THE  COLLECTION
    OF PARTICLES FROM A GAS STREAM
 A  The collection of particles from a gas
    stream involves three distinct phases:

    I   Deposition on a collecting surface

    2   Retention on the collecting surface

    3   Removal from the collecting surface
II   DEPOSITION OF A PARTICLE ON A
    COLLECTING SURFACE
 A  To enable deposition of a particle on a
    collecting surface,  there is need for:

    1  A  resultant force upon the particle in
      the direction of the  collecting surface,

    2  A  collecting surface upon which the
      particle is deposited,  and

    :',  Sufficient time for the particle to reach
      the collecting surface before the particle
      reaches the outlet of the  collecting
      device.
 B There are six mechanisms by which a
   resultant force  may be created upon a
   particle to cause it to migrate toward
   a collecting surface, or cause it to be
   directly intercepted.

   1  Gravity settling

   2  Flow line interception

   3  Inertial deposition
   4  Diffusional deposition

   5  Electrostatic precipitation

   6  Thermal precipitation


III  RETENTION OF A PARTICLE ON A
    COLLECTING SURFACE
 A The fact that particles are "deposited"
    on  a surface is no assurance that they
    are "collected".

    1   To be "collected", they must remain
        on the collecting surface until
        intentionally removed.
 B The problem of retaining a deposit on a
    surface is basically one of having suffi-
    ciently high surface forces to counteract
    the dislodging tendencies of the fluid
    shear  of the carrier gas stream as the
    gas passes over the deposit.
 IV REMOVAL OF A PARTICLE FROM A
    COLLECTING SURFACE
 A For any collection equipment,  some means
    must be provided  for removing the accu-
    mulated deposit, either periodically or
    continuously.
    Removal of deposited material assumes
    outstanding importance in some instances

    1  Although deposit removal is usually
       purely a problem of mechanical design,
       it must be considered in terms of
       overall collection efficiency.
 I'A. C. pm. 18a. 0. GO

-------
                                                3
 SECTION 3
Conversion Factors

-------
         CONVERSION FACTORS



                                           Page
 TEMPERATURE 	     2

 PRESSURE 	     3

 AREA  	     4

 VOLUME	     5

 FLOW  	     6

 WEIGHT	     7

 CONCENTRATION	     8

 LENGTH  	     9

 EMISSION RATES	    10

 VELOCITY 	    11
PA.C.ge.33.12.73

-------
CONVERSION FACTORS - TEMPERATURE

1 Given Units
Degrees
Fahrenheit
Degrees
Centigrade
Degrees
Rankin
Degrees
Kelvin
Desired Units
°F

1. 8°C + 3'2
°R - 460
1. 8(°K-273) + 32
°C
. 5555 x
(°F - 32)

. 5555 x
(°R - 492)
°K - 273
°R
°F + 460
1.8°C + 492

1. 8(°K-273) + 492
°K
5555 x
(°F-32) + 273
°C + 273
. 5555 x
(°R-492) + 273


-------
                             CONVERSION  FACTORS  -  PRESSURE
Sxs. Desired
^^s^ units
Given ^^N,.
units ^^s«s^
Smm
cm- s e c ^
dyn^s
o m^
^ m
ft-secZ
poundals
ft'2
gn>f
c m 2
*f
*f
m^
A tmosphe res "
gmm
2
c m-sec


1
14. 882


980. 665
478. 80
6. 8948
X 104
1. 0133
X 10B
dynes
cm2


1
14. 882


980. 665
478. 80
6. 8948
X 104
1. 0133
x 10s
*m
ft- sec 2
6. 7197
X 10"?
6. 7197
X ID-2
1


65. 898
32. 174
4. 6331
X 103
6. 8087
X 104
poundals
ft2
6. 7197
X 10"2
6. 7197
X ID'2
1


65. 898
32. 174
4. 6331
X 103
6. 8087
X 104
firnf
cm2
1. 0197
X 10~3
1. 0197
y. io~3
1. 5175
x 10"2
1. 5175
X 10~2
1
4. 8824
X 10"1
70. 307
1. 0332
X 103
»f
ft2
2. 0885
X 10"3
2. 0885
X ID'3
3. 1081
X 10"2
3. 1081
X 10"2
2. 0482
1
144. 00
2. 1 162
X 103
*f
m2
1. 4504
X 10"5
1. 4504
X 10'5
2. 1584
X 10"4
2. 1584
X 10"4
1. 4223
X 10"2
6. 9444
X 10"3
1
14. 696
"Atmospheres"

9. 8692
X 10" '
9. 8692
x io-?
1. 4687
X 10"5
1. 4687
X 10"5
9. 6784
X 10~4
4. 7254
X 10'4
6. 8046
X 10~2
1
To convert a value from a given unit to a desired unit,  multiply the given value by the factor opposite the given units
and brneath the desired units.

-------
                                            CONVERSION FACTORS - AREA

en
D
c
0)
J>
O
Square
Inch
Square
Foot
Square
Yard
Square
Mile
Acre
Square
Centimeter
Square
Decimeter
Square
Meter
Squa re
Kilometer
Desired Units
Square
Inch
1
144
1296
40. 144
x 108
62. 73
x 107
15. 5x10-2
15. 5
15. 5 x 102
15. 5 x 10b
Square
Feet
6. 9444
x 10"3
1
9
2. 788
x 107
4. 3560
x 104
10. 764
x ID'4
10. 764
x 10-2
10. 764
10. 764
x 1C6
Square
Yard
77. 1605
x ID"5
0.1111
1
3. 098
x 106
4840
1. 1960
x ID'4
1.1960
x 10-2
1.1960
1.1960
x 106
Square
Mile
2. 49
x 10-1°
3. 587
x 10"8
3. 228
x ID"7
1
15. 625
xlO'4
3. 8610
x 10'11
3. 8610
x 10-9
3.8610
x 10-7
3.8610
x 10-1
Acre
15. 94
x ID'6
2. 296
x lO'5
2. 066
xlO'4
640
1
2. 471
x 10'8
2.471
x 10-6
2.471
x ID"4
2. 471
x 102
Square
Centimeter
6. 452
929. 0341
83.61
x 102
2. 589998
xlQlO
4046. 873
x 104
1
1 x 102
1 x 104
1 x 1010
Square
Decimeter
6. 452
x 10-2
929. 0341
x 10-2
83.61
2. 589998
xlQ8
4046. 873
x 102
1 x lO'2
1
Lx 102
1 x 108
Square
Meter
6. 452
x 10-4
929. 0341
x lO'4
•83. 61
x lO'2
2. 589998
x 106
4046. 873
1 x 10-4
1 x 10-2
1
1 x 106
Square
Kilmeter
6.452
x 10-10
929. 0341
x ID'10
83.61
x 10-8
2. 589998
4046. 873
x 10'6
1 x 10-10
1 x 10-8
1 x 10-6
1
To convert a value from a given unit a desired unit, multiply the given value by the factor opposite the given units
and beneath the desired unit.

-------
                                 CONVERSION FACTORS - VOLUME
^S^Desired
GiveriV^Units
Units ^v>^^
Cubic
Yard
Cubic
Foot
Cubic
Inch
Cubic
Meter
Cubic
Decimeter
Cubic
Centimeter
Liter
Cubic
Yard
1
3. 7037
x 10"
2. 143347
x io"5
1. 30794
1. 3079
x io"3
1. 3079
x io"6
1. 3080
x io"3
Cubic
Foot
27
1
5. 78704
x io"4
35. 314445
3. 5314
x io~2
3.5314
x io"°
3. 5316
x io~2
Cubic
Inch
4.6656
x io4
1728
1
6. 1023
x io4
•61.023
6. 1023
x io"2
61.025
Cubic
Meter
0. 764559
2. 8317
xio-2
1. 63872
x io"5
1
0.001
1 X l(f 6
1. 000027
x io"3
Cubic
Decimeter
764. 559
28. 317
1.63872
x 10"
1000
1
i x io"3
1.000027
Cubic
Centimeter
7.64559
X 10°
2. 8317
X 10
16. 3872
1 X IO6
1000
1
1000. 027
Liter
764. 54
28. 316
1. 63868
X io"2
999. 973
. 99997
9.99973
x 10"
1
To convert a value from a given unit to a desired unit, multiply the given value by the factor opposite the given units
and beneath the desired units.

-------
CONVERSION FACTORS - FLOW
\Desired
^^wUnits
Given ^^w^
Units ^^IN
M3
sec
j^.
mm
M3
hour
_ni
sec
jtL
min
n*
hour
L
sec
L
min
cm.3
sec
cm3
min
«L
SPL
1
0.0167
2.778
<10-5
28.317
x io"3
4.7195
xio"1
7.3658
X 10"6
,.000027
x io"3
,.6667
x io"5
1 < IO'6
1.6667
X io"8
-ill
min
60
1
16.667
x ID'3
1.699
28.317
X 10'3
4. 7195
X10-4
6.00016
XiO-2
1.000027
XKf3
6 X 10~5
-6
1 X 10
M3
hour
3600
60
1
101.94
1.699
28.317
x io'3
3.C
6.00016
-2
X 10
3.6 X 10~3
6'X IO"5
f:3
sec
35.3144
0. 5886
98.90
-4
x 10
i
16.667
X 10
2.778
-4
X 10
35. 316
X 10"
5.886
X 10~4
3.5314
x io"5
5.886
x io'7
ft3
min
21. 1887
X IO2
35.3144
0.5886
60
1
1G.667
X 10~3
2. 11896
35.316
X 10'3
2. 1189
x io"3
0.3531
x io"4
ft_3_
hour
12.7132
xio4
21. 189
X IO2
35.3144
3600
60
1
127. 138
2. 11B96
1.271
X IO"3
2. 11887
X io"3
L
sec
999.973
16.667
27. 777
x io'2
28. 316
47. 193
X 10~2
7.866
X 10"
1
1.6667
X 10~2
9.99973
X 10~4
5.9998
x io"2
L
mm
59.998
X 10
999.973
16.667
16.9896
xio2
28.316
0.4719
60
1
5.9998
_2
X 10
9.99973
X io"4
3
cm
sec
6
1 X 10
16.667
3
X 10
2.777
X 10"
2.8317
xio4
4.7195
X IO2
78.658
1000.027
16.667
1
60
3
cm
min
6 X IO7
6
1 X 10
1.666
4
X 10
1.699
X IO6
2.8317
4.7195
2
X 10
16.667
1000.027
16.667
X IO"3
1
To convert a value from a given unit to a desired unit, multiply the given value by the factor opposite the given units and beneath the desired unit.

-------
CONVERSION FACTORS-WEIGHT

Micro-
gram
Milli-
gram
gram
Kilog ram
grain
Ounce
(avdp)
Pound
(avdp)
Ton
(U.S. shoi
Tonne
(metric)
Desired Units
Micro-
gram
1
1 x 103
1 x 10b
1 x 109
64. 799
x 103
28. 349
x 106
453. 59
x 106
905. 185
-t) x 109
1 x 1012
Milli-
gram
1 x 1Q-J

1 x 10 3
1 x 10fa
64. 799 H
28. 349
x 103
453. 59
x 103
907. 185
x 10b
1 x 109
gram
1 x 10-b
1 x 10-3
1
1 x 103
64. 799
x lO-3
28. 349
453. 59
907. 185
x 103
1 x 10b
Kilo-
gram
1 x 10-9
1 x lO"6
1 x 10- 3
1
64. 799
x lO'6
28. 349
x 10'3
453. 59
x ID"3
907. 185
1 x 103
grain
15. 4124
x 10~6
15. 4324
x lO'3
15. 4324
15. 4324
x 103
1
437. 5
7000
14 x 10b
1. 543 xlO
Ounce
(avdp)
3. 5274
x 10'8
3. 5274
x 10-5
3. 5274
x 10-2
35. 274
22. 857
x ID'4
1
16
3. 2
x 104
3. 5274
x 104
Pound
(avdp)
2. 2046
x 10-9
2. 2046
x 10-b
2. 2046
x 10~3
2. 2046
1. 4286
x ID'4
62. 5
x 10-3
1
2000
2204. 62
Ton
(U. S. shorl
1.1023
x ID"12
1.1023
x 10-9
1. 1023
x lO-6
1. 1023
x ID"3
7. 143
x 10'8
3. 125
x 10-5
5 x 10~4
1
1. 10231
Tonne
») (metric)
1 x 10-12
1 x 10-9
1 x 10-b
1 x lO"3
64. 799
x 10-9
28. 349
x 10-6
453. 59
x 10-6
0. 907185
1

-------
CONVERSION FACTORS - CONCENTRATION


•M
•rH
C
C
ID
>
•H

-------
                                      CONVERSION  FACTORS - LENGTH
*S^L>pSln.d
,.. ^"X- Units
Given ^^
Units ^Vw
Infh
l-'fJOl
Yard
Mile
Mic rtjn
MilllmctiT
(.Vntimru.' r
M c [ e r
Kilomctc-r
In<.h
1
12
36
6. 330(1
x ln4
3. 937
X 1(T5
3 . '.) 3 7
x,o-2
3. 937
xio-1
39. 37
3. 037
,,o4
Foot
83. 33
X 10~3
I
3
5280
32. 808
X10-7
32. 808
-4
X 10
32.808
*,o-3
32. 808
X 10~'
32.808
x,o2
Yard
27.778
x 10- 3
3333
1
1760
10.94
xio-7
10.94
xio-4
10.94
X 10'3
10. 94
X lo"1
10.94
X 102
Mile
1.578
x io"5
1. 894
x lo'4
5. 682
-4
X 10
1
62. 137
xio-11
. 62. 137
X 10'8
62. 137
x io"7
62. 137
x io'5
62. 137
-2
X 10
Micron
2. 54
A 10
30.48
X 10
91.44
4
X 10
1. 6094
9
X 10
1
1 X 10
1 X 104
6
1 x 10
9
1X10
Millimeter
25. 4
304. 8
914. 4
1. 6094
« io6
-3
1 X 10
1
10
3
1 < 10
6
1 X 10
Centimeter
2.54
30.48
91.44
1. 6094
X )0
-4
1 X 10
0. 1
1
1 X 10
5
1 < 10
Meter
2. 54
x io~
30.48
x 10"
91.44
X 10"
1 . 6094
x,o3
1 X JO'6
•J
1 X 10
1 X 10~2
1
1 X 10
Kilometer
2. 54
X 10~5
30.48
x io'5
91.44
x io~5
.1. 60.94
IX ,0-9
1 X io"
1 X 10~5
1 X io"
1
To convert a value from a given unit to a desired unit, multiply the given vaJue by the factor opposite the given units and hr neat h the (Jcsj red units.

-------
                           CONVERSION FACTORS  -  EMISSION RATES
^^^ Desired
^^•s^ units
Given ^^^(^
units ^^x,.
grns/ sec
gms/min
kg/hr
kg/ day
Ibs/min
Ibs /lu-
tes /day
tons / hr
tons /day
tons/yr
gms/sec
1.0
1. 6667
X 10-2
2. 7778
X 10- '
1. 1574
x'lcr2
7. 5598
1. 2600
X ID" l
5. 2499
X ID"3
2.5199
X IO2
1. 0500
X 10
2. 8766
X 10-2
gms/ min
60.0
1. 0
16. 667
6. 9444
X ID'1
4. 5359
X 102
7. 5598
3. 1499
X ID'1
1.5120
X IO4
6. 2999
X 102
1.7260
kg/hr
3. 6
6. 0
X 10-2
1.0
4. 1667
X ID"2
2. 7215
X 10
4. 5359
x io-i
1. 8900
X ID"2
9.0718
X 102
3. 7799
X 10
1.0356
X 10"1
kg/ day
8. 640
X 10
1. 4400
2. 4000
X 10
1.0
6. 5317
X 102
1.0886
X 10
4.5359
X 10"1
2. 1772
X 104
9.0718
X 102
2.4854
Ibs/min
1. 3228
X ID"1
2. 2046
X 10-3
3.6744
X 10-2
1.5310
X 10-3
1.0
1. 6667
X 10-2
6. 9444
X lO"4
3. 3333
X 10
1. 3889
3.8052
X 10"3
Ibs/hr
7. 9367
1. 3228
x io-i
2. 2046
9. 1860
X lO-2.
60. 0
1.0
4. 1667
X 10-2
2. 0
X 103
8.3333
X 10
2. 2831
X 10'1
Ibs/day
1. 9048
X 102
3. 1747
5. 2911
X 10
2. 2046
1.44
X 103
24.0
1. 0
4. 8000
X 104
2.0
X 103
5.4795
tons/hr
3. 9683
X 1C-3
6. 6139
X 10"5
1. 1023
X 10-3
4. 5930
X 10"5
3. 000
X lO-2
5. 0000
X lO"4
2. 0833
X 10'5
1.0
4. 1667
X 10"2
1. 1416
X 10"4
tons /day
9.5240
X 10"2
1. 5873
X 10"3
2. 6456
x io-2
1. 1023
X 10-3
7. 2000
X ID'1
1. 2000
X ID"2
5. 0000
x io-4
24.0
1.0
2. 7397
X IO"3
tons/yr
3. 4763
X 10
5. 7938
X ID"1
9. 6563
4. 0235
X 1C' !
2. 6280
X IO2
4. 3800
1.8250
X lO'l
8. 7600
X 103
365.0
1.0
To convert a value from a given unit to a desired unit,  multiply the given value by the factor opposite the given units and beneath
the desired units.

-------
                     CONVERSION FACTORS -  VELOCITY
"x^^ Desired
^X*XS^ units
Given ^^X^
units ^"""^h
m/ sec
ft/ sec
ft/ rnin
km/hr
mi/hr
knots
mi/ day
ml sec
1. 0
3. 0480
/, 10-1
5. 0080
x 10~3
2. 7778
X ID' J
4. 4707
X 10' l
5. 1479
X 1Q-1
1. 8627
x 10- 2
ft/ sec
3. 2808
1. 0
1. 6667
X 10-2
9. 1 134
X ID"1
1, 4667
1. 6890
6. 1111
X. 10"2
ft/min
1. 9685
X 102
60
1. 0
5. 4681
X 10
88. 0
1. 0134
X 102
3. 6667
km/hr
3. 6
1. 0973
1. 8288
X ID'2
1. 0
1. 6093
1. 8533
6. 7056
X ID'2
mi/hr
2. 2369
6. 8182
X 10'1
1. 1364
X ID'2.
6. 2137
X ID'1
1. 0
1. 1516
4. 1667
x io-2
knots
1. 9425
5. 9209
X 10'1
9. 8681
X 10'3
5. 3959
X 10'1
8. 6839
X 10'1
1. 0
3. 6183
X ID"2
mi/day
5. 3687
X 10
1. 6364
X 10
2. 7273
X ID"1
1. 4913
X 10
24
2. 7637
X 10
1. 0
To convert a
opposite the
value from a given unit
;iven units and beneath
 to a desired unit, multiply the given value by the factor
the  desired units.

-------
SECTION 4
Gas Properties-Basic Concepts

-------
                                   GAS PROPERTIES-BASIC CONCEPTS
    I.   EXPRESSION OF GAS-TEMPERATURE

        A.   The Fahrenheit and Celsius Scales

            The range of units on the Fahrenheit
            scale between freezing and boiling
            is 180; on the Celsius or Centigrade
            scale, the range is 100.   Therefore,
            each Celsius-degree is equal to 9/5
            or 1.8 Fahrenheit-degree.  The
            following relationships convert one
            scale to the other:
F  -  1.8  C  +  32      (I - 1)
      (°F     32)/1.8    (I - 2)
             C  -

            where
                     F  -  degrees Fahrenheit
                     C  »  degrees Celsius or
                           degrees Centigrade
                        B.  Absolute Temperature

                            Experiments with perfect gases have
                            shown that, under constant pressure,
                            for each change in Fahrenheit-degree
                            below 32°F the volume of gas changes
                            1/491.67.  Similarly, for each
                            Celsius-degree, the volume changes
                            1/273.16.  Therefore, if this change
                            In volume per temperature-degree is
                            constant, the volume of gas would,
                            theoretically,  become  zero at
                            491.67 Fahrenheit-degrees below 32°F,
                            or at a reading of -459.67°F.  On the
                            Celsius or Centigrade scale,  this
                            condition occurs at 273.16 Celsius-
                            degrees below 0°C, or at a temperature
                            of -273.16°C.
                 P VAfM J.T
                   32° F
               - 459.6° f
     Absolute rero
         Cftntigrad*
               0°C
491.6
fahrenheit-
degrees
                                                                     Afeiolut*
                                                         491.6° R
                                                           273
                                                           centigrade -
                                                           degrees
                                         0°R
                                                                              -  373*&
273°K
                                                                     0°K
                                                         Absolute zero
                               Figure 1    Temperature-Scale relationship
Revised by R.T. Shigehara, Assistant Chief, Engineering Section,
Office of Manpower Development,  Institute  for Air Pollution Training,
National Air Pollution Control Administration
PA.SS.31b.2.70

-------
 Gas Properties - Basic Concepts
       Absolute temperatures determined by
       using Fahrenheit units are expressed
       as degrees Rankine (°R);  those deter-
       mined by using Celsius units  are ex-
       pressed as degrees Kelvin (°K).   The
       following relationships  convert  one
       scale to the other:
              3F  +  459.67
       °K  =  °C  +  273.16
(I - 3)

(I - 4)
       Relationship of the various  temperature
       systems are shown graphically in
       Figure 1.  The symbol "T"  will be
       used in thi-s outline to denote
       absolute temperature and "t" will  be
       used to indicate Fahrenheit  or Celsius
       degrees.

TI.  EXPRESSION OF  GAS PRESSURE

    A.   Definition of Pressure

        A body may be subjected to  three
        kinds  of stress:  shear, compression,
        and tension.   Fluids  are  unable to
        withstand  tensile stress; hence,  they
        are subject to shear  and  compression
        only.  Unit compressive stress in a
        fluid  is termed pressure  and  is ex-
        pressed  as  force  per  unit area
        (e.g.  Ib./in^  or  psi,  gm,/cm^) .

        Pressure is equal  in  all  directions
        at  a point  within  a volume  of  fluid,
        and acts perpendicular to a surface.

   B.  Barometric  Pressure

       Barometric  pressure and atmospheric
       pressure are synonymous.  These
       pressures are measured with a
       barometer and are usually expressed
       as  inches,   or millimeters, of mercury.
       Standard barometric pressure is the
       average atmospheric pressure at sea
       level,  45°  north latitude at 35°F.
       It is equivalent to a pressure of
       14.696  pounds-force per square inch
       exerted at  the base of a column of
       mercury 29.921 inches high.  Weather
       and altitude are responsible for
       barometric  pressure variations.

   C.   Gage Pressure

       Measurements of pressure by  ordinary
       gages are indications  of  the
       difference  in pressure above,  or below,
       that of the atmosphere surrounding the
       gage.   Gage  pressure,  then,  is
    ordinarily the pressure indicated by
    the gage itself.  If the pressure of
    the system is greater than the
    pressure prevailing in the atmosphere,
    the gage pressure is expressed as a
    positive value; if smaller, the gage
    pressure is expressed as a negative.
    The term, "vacuum," designates a
    negative gage pressure.

    The abbreviation, "g," is used to
    specify a gage pressure.  For
    example, psig, means pounds-force
    per square inch gage pressure.

D.  Absolute Pressure

    Because gage pressure (which may
    be either positive or negative)
    is the pressure relative to the
    prevailing atmospheric pressure,
    the gage pressure, added algebrai-
    cally to the prevailing atmospheric
    pressure (which is always positive),
    provides a value that has a datum
    of "absolute zero pressure."  A
    pressure calculated in this manner
    is called absolute pressure".  The
    mathematical expression is:
                  where:
                            abs
                            abs
                            atm
                            g
             P     +  P
              atm      g
(II - 1)
             absolute pressure

          =  atmospheric pressure

          =  gage pressure
                           The abbreviation, "a," is sometimes
                           used to indicate that the pressure
                           is absolute.  For example, psia,
                           means pounds per square inch
                           absolute pressure.  The symbol "P"
                           by itself without the subscript
                           "abs" will also be used in this
                           outline to indicate absolute
                           pressure.

                           Equation II - 1 allows conversion
                           of one pressure system to the other.
                           Relationship of the pressure systems
                           are shown graphically in Figure
                            2.1  using two  typical gage
                           pressures, (1)  and (2).  Gage
                           pressure (1) is above the datum
                           from which gage pressures are
                           measured, and,  hence, is expressed
                           as a positive value;  gage pressure
                           (2) is below the gage pressure
                           datum, and, therefore, is ex-
                           pressed as a negative value.

-------
                                                              Gas Properties    Basic Concepts

<



i


pd)

i
i
0—
(2) '
atm

Gage
Pressure
Datum
" P g(2)
y
i
P(2)
Absolute
Pressure
Datum
Figure 2. 1 Gas-Pressure relationship
E.  The Concept of Pressure-Head
    Pressure-head is the height of a
    column of fluid required to produce
    a given pressure at its base.
         Fluid of -<
         density Pf
T
  h
1
     Figure  2.2   P-ti  relationship
    The relationship between  pressure  and
    pressure-head  is:

         p  .  Pf(-8-)  h        (II  -  2)
                    c
where:   P  -  pressure, force/area

         p  -  density  of  fluid, mass/volume
          f
                                                              g  »  local acceleration due to
                                                                    gravity,  length/time^

                                                              g  =  dimensional constant

                                                              h  ™  pressure-head in terms of
                                                                    Pf,  length

                                                         Pressure-head may be expressed in terms
                                                         of any fluid that is convenient; e.g.
                                                         Hg or H20.

                                                     F.   Dalton's Law of Partial Pressure

                                                         When gases,  or vapors (having no chemical
                                                         interaction)  are present as a mixture
                                                         in a given  space, the pressure exerted
                                                         by a component  of the gas-mixture at a
                                                         given temperature is the same as it
                                                         would exert  if it filled the whole
                                                         space alone.   The pressure exerted by
                                                         one component of a gas-mixture is called
                                                         its partial  pressure.  The total pressure
                                                         of the gas-mixture is the sum of the
                                                         partial pressures.

                                                    III.   THE LAW OF  IDEAL GASES

                                                          A.  The Laws of Boyle and Charles

                                                              1.  Boyle's Law

                                                                  Boyle's Law states that, when
                                                                  the temperature (T)  is held
                                                                  constant,  the volume (V) of a
                                                                  given mass  of a perfect gas of
                                                                  a  given composition varies  in-
                                                                  versely as  the absolute
                                                                  pressure,  i.e.:
                                                    where:       <* -   proportional to
                                       Charles '  Law

                                       Charles'  Law states  that,  when
                                       the volume is held constant,
                                       the absolute pressure of a
                                       given mass of a perfect  gas  of
                                       a given  composition  varies
                                       directly  as the absolute
                                       temperature,  l.e:
                                                                  a T
                                                                          at constant V

-------
Gas Properties - Basic Concepts
      B.  The Law for Ideal Gases

          Both Boyle's and Charles' Law are
          satisfied in the following equation:
 where:
              PV
                     mRT
                (III - 1)
          P

          V

          m

          M

          T

          R
absolute pressure

volume of a gas
mass of a gas
molecular weight of a gas

absolute temperature
universal gas-constant
          The units of R depend  upon the
          units of measurement used in the
          equation.  Some useful values are:
          1.
              1544  (ft)
          2.  21.83
          3.  554.6
   (Ib -mole)  ( R)
      (in. Hg)  (ft3)
     ~(lb -mole)  (°R)
         ™      1
     (mm Hg) (ftj)

     lib -mole) (°R)
        m
          In the above units of R:

                    3
              V
              m
  ft~
  Ib
                    m
              M =  Ib  /Ib -mole
                  „ m  m
              T = °R

              P = lbf/ft2 for (1)
                  in.Hg

                  mm Hg1
          for (2)
          for (3)
               Any value of R can be obtained
           by utilizing the fact, with approp-
           riate conversion factors, that there
           are 22.414 liters per gm -mole
                    1              "
           or 359 ft  per Ib -mole at 32°F
           and 29.92 in. Hg.™ Problems to
           illustrate this will be shown later
           in this discussion.
    IV.  CALCULATION OF APPARENT MOLECULAR
         WEIGHT OF GAS MIXTURES

             Utilizing Dalton's law of partial
         pressure and the ideal gas law,  the
         following equation can be derived for
         calculating the apparent molecular
         weight of a gas mixture:
                                     VI.
                                              M ,
                                               mix  "

                                          where:  ^mix.
                     xMx      (IV-1)

                      apparent molecular
                      weight  of a  gas-mixture

                      proportion by volume
                      of a gas—component

                      molecular weight  of a
                      gas component
         In all other equations  (except where
    specifically noted), the symbol  "M" will
    be used to denote the molecular  weight  of
    a pure gas or a gas-mixture.

V.  GAS DENSITY

    Gas density can be determined by re-
    arranging equation III - 1 and letting
    density p » r:

         p  -  l_^            (V - 1)
               R T

    where:   P  »  density

            P  -  absolute pressure

            M  •  molecular weight

            T  -  absolute temperature

            R  =•  universal gas constant

    Another method of determining density is
    by utilizing the fact that there are
    22.414   liters per gm-mole or 359 ft3
                                         per  Ib  -mole  at  32°F and 29.92 in.  Hg.
              Ib
                                (V 2)
   Ib
          M
Ib -mole
  m
                                                        ft"
                                                              359
                                                                    ff
                                                  Ib -mole
                                                    m
                                                                              492° R
'P in.  Hg
                                                                T° R
                                   29.92  in.  Hg
     In this equation, M- Ib /Ib -mole,
                            mm
     T-R°,  P-in. Hg, and p-lb /ft3.
                             m

     VISCOSITY

     A.  Origin and Definition of Viscosity

         Viscosity is the result of two
         phenomena:  (1) intermolecular
         cohesive forces, and (2) momentum
         transfer between flowing strata
         caused by molecular agitation
         perpendicular to the direction of
         motion.  Between adjacent strata of

-------
                                                                 Gas  Properties  -  Basic  Concepts
      • •• ••••••••••••••••••••••»• ••••••••••••^••••B • •••


                               ^ f     I t
                   Figure 3
        a flowing fluid a shearing stress
        occurs which is directly pro-
        portional to the velocity gradient
         (Figure  3)
                         dv
where:  g   =  dimensional constant

        7      unit shearing stress between
               adjacent layers of fluid
        dv
        civ
velocity gradient
        ,1      proportionality  constant
               (v Lscosity)


        The  proportionality  constant,  \i,  is
        called the  coefficient  of  viscosity,
        or merely,  viscosity.   It  should  be
        noted that  the  pressure does not
        appear in equation VI - 1,  in-
        dicating  that the shear (T) and  the
        viscosity (\>) are independent  of
        pressure.   (Viscosity actually in-
        creases very slightly with  pressure
        but  this  variation is negligible
        in most engineering  problems.)^

    B.  Kinematic Viscosity

        Kinematic viscosity  is  defined
        ,!i-cording to the following  re-
        iati onship:
                            (VI -2)
            v  =  kinematic viscosity

            y  =  viscosity of the gas

            p     density of the gas
Note the absence of dimensions of force.
C.  Liquid Viscosity Versus Gas Viscosity.
    1.   Liquid Viscosity

        In a liquid, transfer of momentum
        between strata having a relative
        velocity is small compared to the
        cohesive forces between the
        molecules.  Hence, shear stress T
        is predominantly the result of
        intermolecular cohesion.   Because
        forces of cohesion decrease rapidly
        with an increase in temperature,
        the shear stress decreases with an
        increase in temperature.   Equation
        VI - 1 shows that shear stress is
        directly proportional to the
        viscosity.  Therefore, liquid
        viscosity decreases when the
        temperature increases.
    2.  Gas Viscosity

        In a gas, the molecules are too far
        apart for intermolecular cohesion to
        be effective:  Thus, shear stress
        is predominantly the result of an
        exchange of momentum between flowing
        strata caused by molecular activity.
        Because molecular  activity increases
        with temperature increases, the shear
        stress increases with a rise in the
        temperature.  Therefore, gas
        viscosity is increased when the
        temperature increases.

 D.  Determination of Viscosity of  Gases

    The viscosity of a  gas for prevailing
     conditions may  be  found accurately  from
     the following  formula:
                                                                  (VI - 3)
                                     where:  y   =   viscosity prevailing

                                             y   =   viscosity at 0 C and
                                                     prevailing pressure

                                             T   =   absolute prevailing
                                                     temperature (°K)

                                             n   =   an empirical exponent
                                                     (n = 0.768 for air)

-------
Gas Properties   Basic Concepts
                                  Viscosity of air  at  1 Atmosphere  *
                   Temperature
            Degree          Degree
            Centigrade     Fahrenheit
                   -100 —i
                         — -100
                      0 —
100 .



200


300 —

400 •

500 •

600 •

700

800 '
900
1000-
                          —200
                            300
                            400
                           -500

                          — 600
                           -700
                          I—800
                           -900
                           -1000
                           -1100
                           -1200
                           -1300
                           -1400
                           -1500
                           -1600
                           -1700
                           -1800
                                                     Viscosity
                                                     Centipoises

                                                           — 0.1
                                                           - 0.09
                                                           - e.08
                                                           -0.07
                                                           : 0.06

                                                           i-0.05

                                                           :-0.04
                                                                               -0.03
                                                                               \>
                                                                               -0.02
-0.01
- 0.009
-0.008
-0.007
r o.oo6

-0.005
(l)centipoise
(10)"2 gm
cm-sec
(lO) poise
2. 09 (10) "5
'b - sec
2
ft
2. 09 (10) "5
slug
ft - sec
6.72(lO)"4
Ib
m
ft - sec
  *l'erry,  J.H.   Chemical Engineer's  Handbook,  McGraw-Hill Book Co.,  New York,  1950.
                                                 Figure  4

-------
                                                               Gas Properties - Basic Concepts
   .036



   .034



   .032



   .030




   .028



   .026




   .024



   .022



   .020

 Q.
 u
  - .018
 */)


 8  -016
>
    .014



    .012



    .010



    .008




    .006



    .004




    .002
                       -CO,
                                  H20.
                                      CH4
                   100
200
300
                                                      400
                                   500
                                   600
700
                                  Temperature,  "F
                            Figure 5     Viscosity at 1 Atmosphere

-------
 Gas  Properties    Basic  Concepts
Table 1 Ratios of Specific Heats of gases at 1 Atm. Pressure*
Compound Formula Temperature Ratio of
°C Specific
Heats
Acetaldehyde
Acetic acid
Acetylene


Air







Ammonia

Argon





Benzene
Bromine

Carbon dioxide



disulfide

monox Ldc


Ch lori ne
Ch lo reform
C2H40
C2H402
C2H2










NH3

A





C&H6
Br2

CO 2



CS2

CO


C12
CIIC13
30
136
15
-71

925

17

-78

-118

15

15

-180

0-100

90
20-350

15

-75

100

15

-180
15
100
1.14
1.15
1.26
1.31

1.36

1.403

1.408

1.415

1.310

1.668

1.76(?)

1.67

1.10
1.32

1.304

1 .37

1.21

1.404

1.41
1.355
1.15
Compound Formula
Cyanogen
Cyclohexane
Dichloro-
difluorme thane

Ethane





Ethyl alcohol

ether



Ethylene




Helium

Hexane (n-)

Hydrogen





bromide

chloride

(CM) 2
C6H12
CC12F2

C2H6





C2H60

C4H100



C2H4




He

C6H14

H2





HBr

HC1

Temperature Ratio of
°C Specific
Heats
15
80
25

100

15

-82

90

35

80

100

15

-91
-180

80

15

-76

-181

20

15
100
1.256
1.08
1.139

1.19

1.22

1.28

1.13

1.08

1.086

1.18

1.255

1.35
1.660

1.08

1.45

1.453

1.597

1.42

1.41
1.40
    The viscosity of air and other gases at
    various temperatures and at a pressure
    of 1 atmosphere may be taken from
    Figures A and 5.

VII.   SPECIFIC HEAT

      A.  Definition  of  Specific Heat

         The specific heat of  a. gas is  the
         amount  of heat required to change
         the temperature of  a  unit-mass  of
         gas one temperature-degree.  Units
of specific heat are, therefore,
(Btu/lb ) (°F) or calories/(gm )
(°C) depending upon the dimensional
system used.
Heat may be added while the volume
or pressure of the gas remains
constant.  Hence, there may be two
values of specific heat: (1) the
specific heat at constant volume
(C ), and (2) the specific heat
atvconstant pressure (c ).
                       P
"Perry, -Mi. Chemical Engineer's Handbook, McGraw-Hill Book Co. New York. 1950.

-------
                                                        Gns Properties  - Basic Concepts
Table 1 Ratios of Specific Heats of gases at 1 A Cm. Pressure
Compound Formula Temperature Ratio of
°C Specific
Heats
1
Hydrogen (continued)
1

cyanide


iodide
sul f i do


1 odi ne
I sobutane
Krypton
Mercury
Methane








Methyl acetate
;l 1 collO I
other

HCN


III
H2S


12
V'lO
Kr
Hg
cn4








C_,H602
CII40
C2II60



65
140
210
20-100
15
-45
-57
185
15
19
360
600
300

15

-80

-115

15
77
6-30



1.31
1.28
1.24
1.40
1 .32
1 .30
1.29
1 .30
1.11
1.68
1.67
1.113
1.16

1.31

1.34

1.41

1.14
1 . 203
1.11
Compound Formula Temperature Ratio of
°C Specific
Heats

Methylal


Neon
Nitric oxide


Nitrogen

Nitrous oxide



Oxygen



Pentane (n-)

Phosphorus

Potassium
Sodium
Sulfur dioxide
Xenon

C3H8°2


Ne
NO


N2

N20



°2



C5H12

P

K
Na
SO 2
Xe

13
40

19
15
-45
-80
15
-181
100
15
-30
-70
15
-76
-181

86

300

850
750-920
15
19

1.06
1.09

1.64
1 .400
1 .39
1.38
1 .404
1.47
1 .28
1.303
1.31
1.34
1.401
1.415
1.45

1.086

1.17

1.77
1.68
1.29
1 .66
Because the heat energy added at
constant pressure is used in raising
Lhf temperature and doing work
against the pressure as expansion
takes place, C  is greater than
C             P
 v
Specific Heat Ratio

Specific heat ratio (R)  is defined  as:
          C
    k  = —5L-       (VII - 1)
         Specific heat ratios for e,ases are
         shown in Table 1.

C.  Determination of Specific Heat for a
      Gas-Mixture

The specific heat for a mixture of gases may
be calculated from:
      P(mix)

      v (mix)
                   BxCp(x)     (VII    2)
                                                              Bx Cv(x)    (VII   3)

-------
Gas Properties   Basic Concepts
where:
        "p (mix)
        C
         v (mix)
         "p U)
            (x)
     specific heat at con-
     stant pressure for gas-
     mixture

     specific heat at con-
     stant volume for the
     gas-mixture
     proportion by volume of
     a gas-component

     specific heat at con-
     stant pressure for a
     gas-component

     specific heat at con-
     stant volume for a gas-
     component
 For ordinary  temperature  (for example, about
 80° F  as experienced at the metering device in
 atmospheric or source sampling work) the
 specific heats may be assumed to be constant.

 VIII.  REYNOLD'S NUMBER

   A.  Definition

       A typical inertial force per unit
       volume of fluid is p v  ,
       a typical viscous force per unit
       volume of fluid is y v
       The first expression divided by the
       second provides the dimensionless
       ratio known as Reynold's Number:
    Re
                    inertial  force
                    viscous force
                       (VIII  -  1)
where:  p
         Re
density of the fluid (mass/
volume)
velocity of the fluid

dimensional constant

a linear dimension

viscosity of the fluid

Reynold's Number
        The  linear  dimension,  L,  is a length
        characteristic  of  the  flow system.
        It is  equal to  four  times  the mean
        hydraulic radius,  which  is the cross-
        sectional area  divided by  the wetted
        perimeter.   Thus for a circular pipe
        L =  diameter of the  pipe;  for a
        particle settling  in a fluid  medium,
        L =  diameter of the  particle;  for a
        rectangular  duct,  L  -  twice the
        length times  the width divided by the
        sum; and for an anulus such as a
        rotameter system,  L  -  outer diameter
        minus  the inner diameter.

   B.   Laminar and  Turbulent  Flow

        1  Laminar Flow

          In  laminar flow,  the fluid  is con-
          strained  to motion  in layers
          (or laminae) by the action  of
          viscosity.  The layers  of  fluid
          move in parallel  paths  that re-
          main distinct from  one  another; any
          agitation  is of a molecular nature
          only.  Laminar  flow occurs  when
          Reynolds' Number  for circular
          pipes is  less than  2000  and less
          than 0.1  for particles  settling
          in a fluid medium.

        2  Turbulent Flow

          In turbulent flow,  the  fluid is not
          restricted to parallel  paths but
          moves forward in  a  haphazard
          manner.   Fully turbulent flow
          occurs when Reynolds'  Number is
          greater than 2500 for circular
          pipes and greater than  1000  for
          settling particles.

IX.   Summary of Useful Equations
                                                            A.
              Temperature

                  °F

                  °R

                  °K

                  °R

              where:
                                                                            1.8  °C
                                                                                    +  32
                                                                             C

                                                                            i
                                                                            1 .
                                                                                   46

                                                                                 273
        The larger the Reynold's  Number,  the
        smaller is the effect  of  viscous
        forces;  the smaller the Reynold's
        Number,  the greater the effect  of
        the viscous forces.
a °v
o  K

  degrees Fahrenheit

  degrees Centigrade
  or Celsius
  degrees Ranklne

  degrees Kelvin
 10

-------
                                                             (las  Properties    Hasic  Concepts
 B.   Pressure


         P
                F    + P
          abs    atm    g
where:
        1 std atm  «  14.696 Ih /in



                   E  2316.22'. Ib /ft2



                   3  29.921 in. Hg



                   a  760 mm Hg


        P  = pressure



        p  - density



        h  - pressure head or height



        g  = gravitational acceleration



        g  *= dimensional constant



        Subcripts



        abs = absolute



        atm   atmosphere



          g   gage



          f   fluid
C.   Ideal Gas Law



       PV

              m  RT

              M



              154A (1V(ft)
                   (Ib -mole) (  k)
                   v  m
             21.83
                    (in.  Hg )(ft )
                     Ub"
          =   554      (ram Hg)(ft )
                    (Ib -mole)( R)
                       m


1 Ib -mole  -  359 ft3 at 32°F and  29.92  in. Hg
 1 urn  -mole   -   22.414  liters at 0°C and
    m
                760mm Hg
                                                 where P = absolute  pressure



                                                       V = volume



                                                       m = mass



                                                       M = molecular weight



                                                       R = gas constant



                                                       T = absolute  temperature




                                                 D.  Apparent Molecular Weight



                                                              1MB
                                                                 x x


                                                                     molecular weight



                                                             B   =   proportion by  volumi



                                                     Subscript



                                                           mix   =   gas mixture



                                                             x   =   component
                                                           .
                                                          mix


                                                        where:  M
                                                        P

                                                                (359


                                                        where:   p      density


                                                                 P  =   absolute  pressure



                                                                 R  =   gas  constant



                                                                 T  =   absolute  temperature
                                                     F    Viscosity,  h
                                                                              ,-4
                                                                 1  cp  s  6.72  X 10  "  Ib
                                                                                       m

                                                                                    ft  sec.
                                                  r,.  Revnold'--;  '
                                                             Re
                                                                    L  v  f,
                                                   whe re :   N
                                                            Ke

                                                            L
                                                                      Reynold's  'number


                                                                  =   linear  dimension


                                                                      D  for circular  pipe


                                                                  =   D   for  snhericjL pn
                                                                                           11

-------
Gas Properties - Basic Concepts
               =   D-D,  for a  rotameter

           V   =   velocity

           P,   -   fluid  density

           Mf   =   fluid  viscosity

           D   =   diameter

           D   =   tube diameter

           DC   =   float  diameter

           D   =   particle diameter
                                                    REFERENCES:
                                                         1.  J.K. Uennard, Elementary  Fluid
                                                               Mechanics, John Wiley and  Sons,
                                                               Inc., New York (1947).

                                                         2.  M.B. Lemon and M. Ference,
                                                               Analytical Experimental Physics,
                                                               The University of Chicago  Press,
                                                               Chicago (1946).

                                                         3.  J.H. Perry, Chemical Engineers
                                                               Handbook, McGraw-Hill Book Co.,
                                                               Inc., New York (1950).
  REVIEW PROBLEMS - BASIC CONCEPT

  Problem 1.  In source sampling or in
  evaluating control equipment it is
  essential to know the volumetric flow rate
  within a ductwork to enable the calculation
  of pollutant emission rates or to properly
  size the control equipment.  The
  volumetric flow is usually determined by
  first measuring the velocity of the gas
  stream.  A pitot tube is probably the
  most commonly used device for this purpose.
  However, the pitot tube does not measure
  velocity directly; it measures velocity
  pressure or velocity head, usually in
  inches of water.  The velocity head is then
  related to velocity by the following
  equation.

Fluid of
density p.


T

h
1
I
P
P-h relationship
    where:  v
            C
             P
             g
             h
                  velocity

                  pitot tube coefficient

                  acceleration due to gravity

                  velocity head in terms of
                  the flowing fluid
     a.  The velocity head Is in terms of
the flowing fluid.  Thus if air is flowing,
h Is the height of air that would be
supported by the velocity pressure or kinetic
energy of the flowing air.

     Using C  - 1.0 (empirically determined),
g = 32.2 ft/§ec2,  density of the flowing
fluid - 0.075 lbm/ft3,  and  Ap (velocity head
           )  =  1.0 in.  H.O, calculate v in
                                                         b.  In the above problem la, the density
                                                    of the flowing fluid was given.  This
                                                    normally must be determined by taking
                                                    measurements of P, T, and M and using the
                                                    equation p - PM.  Assume that your stack
                                                                 RT
                                                    pressure reads in in. Hg and T in °K.
                                                    Calculate an appropriate value of R (use
                                                                    Ib
                                                    the dimension — — B^- —  for M) .
                                                                     -mole       '
                                                                  —
                                                                  ID
                                                                 Answer:
                                                                          39.30 <%?
                                                                               (lbm-mole) (°K)
in in,  HO
ft/sec.
                     Answer:   66.8
                                                           c.   In order to determine the molecular
                                                      weight of a gas mixture, its composition
                                                      must be known.   An orsat .analyzer is
                                                      commonly used for this purpose, especially
                                                      for combustion effluents.   If the pro-
                                                      portions by volume on a dry basis are:
 12

-------
                                                                  Gas  Properties    Basic Concepts
       CO,   =  .10         M  =  44

       0,    <=  .08         M  -  32

                           M  =  28
Problem 2
Calculate  the  apparent  molecular weight
(assume B   of  H^O   =  0.10,  M = 18).
                       Answer:  28.72    m
                                      Ib -mole
                                        m
     d.   If t - 77 C, P    =31 in. Hg,
stack pressure P  = -1 in. Hg, and M =
                &
28.72 Ib       calculate  the  density.   Check
        m
     Ib -mole
       m
your results using  equation  V-2.
                          Answer:  .0626 Ib
                                          m
     Reynold's number, N
                        Re'
is a dimension-
less number.   It is the ratio of the inertial
force to the viscous  force in a flow system.
The importance of this number lies in the
area of predicting the behavior of particles
in a fluid medium or in determining flow
characteristics.

     The small particles are of particular
interest in air pollution work.  The rate
at which a particle settles due to gravity
or migrates due to electrostatic or
centrifugal forces is a function of N  .

If N  <0.1,J  the particle is said to be
following Stoke's Law, i.e.
      2
v = gD  (p - p )  for gravitational settling.
 P    P   P   P
where: v     settling velocity
        P
       g     acceleration due to gravity
       D     particle diameter
     e.  Utilizing  all  the  previous  in-
formation,  rewrite  the  relationship  between
v and the pitot  tube  measurements  in terms
of actual measurements  taken.
             particle density

             fluid density

             fluid viscosity
                                                     For the following case:
                         Pf = .075
                                    Ih
                                                                                         ft"
      Answer:
                          g
                                  RT Ap
                              PM
                                                              Ib
                                                              ft
              ,  P= 29.92 in. Hg
                                                     p = 128
                                                      P
                                                     M= 29   m/   m-mole;  what is the largest
                                                     size particle that would settle in Stoke's
                                                     Law range?  ( 1 ft = 30.48 * 10* n)
                                                                                   Answer:  27y
                                                                                             13

-------
                                                5
SECTION 5
Particle Settling Dynamics
Terminal Velocities of Spherical
  Particle

-------
                           PARTICLE  SETTLING  DYNAMICS

                                        David R. Hemenwav
 i NTRODUCTJ ON

      Tho force of gravity,  or attraction between
 masses .  was first described in analytical terms
 by Sir Isaac Newton.   This  force has since been
 recognised  as being essentially constant for
 attraction  of a mass  to earth - if it is less
 than 1 mile above the earth's surface.   The
 distance from the center of the earth to the
 equator  is  3,063 miles.  The gravitational ac-
 celeration  for this distance is 32.09 ft./sec."
 The distance from Che center of the earth to  the
 North Pole  is 3,050 miles with.,a gravitational
 acceleration of 32.26 ft./sec.'.  The accelera-
 tion on  .1  particle therefore is, for all practi-
 cal purposes, constant regardless of its location
 because  of  the very small variations in  gravi-
 tational attraction.   Thus, the acceleration
 value normally used is 32.2 ft./sec.-.
 DESCRIPTION  OK _FORCES

      The  weight  of  a particle  is  described by
 the  following equation:
            Weight  - V   * P  *
                     P     P
    u.oi
where :
is the
     V  is the volume of the particle,
      p                      t          p
density of the particle, g^ is a gravitational
constant, and g is the gravitational acceleration,
which Is considered to be constant as described
.ibove .
     The second force that can be observed is
the effect of buoyancy.  For example, when a
block of wood is  placed in a pail of water it
floats becau.se it  is supported by the buoyant
force of the water.  If a piece of iron is
placed In the water it will sink to the bottom
            of  the pail because the buoyant force of the
            water is not sufficient to cause it to float.
            However, the piece of iron does not weigh as
            much in the pail of water as it does in air
            (see Figure 1).
                    IRONH
                                                      Figure 1.
            The  loss  of weight  (buoyant force) is equal
            the  weight of  the volume of displaced vater.
            For  objects that are totally immersed in water
            (or  any other  fluid) this loss of weight can
            be calculated  by the following equation:
           where:
                     Buovant Force - V  * Pf * g
                                      P    f   T
                p   is  the density of tha displaced fluid
            or  suspending medium.
 Mr.  Hemenwav  is  an  Environmental  Engineer with  the
 National  Air  Pollution  Control Administration
   Ai.pir. 100. 11.69

-------
 Particle  Settling Dynamics
     When a particle or mass  is  placed  in a
fluid, the above two forces  (buoyancy and weight)
will produce a resultant force.   If  the buoyant
force is greater than the weight of  the mass,
then the mass will begin to rise in  the fluid
(see Figure 2).   In the case  of  normal  parti-
culate matter suspended in the atmosphere, the
weight of a particle usually  exceeds the buoyant
force exerted on it, and the  particle settles.
       buoyancy
buoyancy
       wei g ht
weighs
 Figure 2.
 The resultant  force on  the particle will accel-
 erate  it  in  the  direction of the force, and in
 so  doing  friction will  be created between the
 fluid  and the  particle.  Thus, three forces are
 exerted on a body placed in any fluid medium.
 These  forces are friction, buoyancy and weight
 (see Figure  3).
                                    DEFINITION OF TERMS
                          A  = The projected cross-sectional area
                               of a particle
                          C  = Drag Coefficient
                                                       D  = Diameter of a spherical particle
                                                        P
                                                        p = Terming.! Settling Velocity
                                                       g  = local acceleration due to gravity
                                                       g  = dimensional constant
                                                       K  = Cunningham Correction Factor for
                                                            discontinuous fluids
                                                       R  = Reynold's Number
                          V  = Volume of a particle


                          P  = Particle Density


                          P  = Fluid Density
                               Fluid  Viscosity
                                                          All  units  are  dependent on the system
                                                          of measurement used.
 Figure  3.
                       DERIVATION OF SETTLING VELOCITY  EQUATION

                            In the process of passing through a fluid,
                       a particle will cleave, or displace, the fluid
                       immediately in front of the particle, imparting
                       momentum to the fluid.  In the process of pro-
                       ducing momentum, the force produced can be des-
                       cribed by Newton's Law of momentum, which states
                       that the force produced is equal to the pro-
                       duction of momentum per unit time.

                            The amount of momentum produced in the
                       fluid is, then, a function of the fluid density
                       (a measure of the fluid mass), the maximum vel-
                       ocity to which the fluid is accelerated by the
                       action of the moving particle, and the volume
                       of fluid that is equal to that of a cylinder
                       containing a cross-sectional area equal to that
                       of the particle and a length equal to the dis-
                       tance traveled by the particle in a unit time
                       (see Figure 4).

-------
                                                                       Particle Settling Dynamics
     volume
     of fluid
     displaced
Figure 4.

     The force of momentum, commonly called
frictional force, is described by Newton's law
of momentum:  proportional to the factors
affecting the production of momentum.  Therefore,
the frictional force can be formulated as
follows:
                                                    where:
                                                         C' is the friction factor for correlating

                                                    U with f  .  However, the equation for friction
                                                    force is  usually written in the following form:
                                                                      C_  * P, * A *  f  2
                                                     Friction  Force     D     f        p
                        where:
                                                                (5.0)
                             C  will, by definition, be different than
                        C' by a factor of 2.  Both C' and C  are related
                        to the Reynold's Number because it is a measure
                        of the flow pattern surrounding the particle.

                             The friction force, as shown by equation
                        5.0, will increase as the square of the velocity
                        of the particle increases.  As the particle ac-
                        celerates, it will reach the terminal settling
                        velocity where the buoyancy and friction forces
                        will equal the weight of the particle.   The
                        velocity will remain constant at this point
                        because there will be no resultant force to
                        accelerate the particle.  At the point of e-
                        qualization of forces, equations 1.0, 2.0,
                        and 5.0 may be grouped as follows:
                        Weight     buoyant force  +  friction force
                       V  * P  * g
                        P    P   „
                           v   * P   *
                           p    f
                                                                    (6._0)
                                                                     ^2
Frictional Force
where:
                   P  * A *
f  * U
 P
(3.0)
     A is the  cross-sectional  area  of  the  part-
icle perpendicular  to  the  direction of movement,
f  is the velocity  of  the  particle,  and U  is  the
 P
maximum velocity  to which  the  fluid is acceler-
ated in the process of being displaced by  the
particle.

     The velocity,  U,  to which the  fluid is ac-
celerated will be directly related  to  the  vel-
ocity of the particle, f  ,  by  a factor that
depends upon the  flow  pattern  surrounding  the
particle.  Thus,  the equation  for friction force
can be written as:
Friction Force
                          * A  *  f'
                                        (4.0)
                                                     By cancellation of the g  terms  equation 6.0
                                                     can be rewritten as follows:                 (6.1)
                                                    V  *
                                                     P
                           (f )'
                            p
                                  ;- Vp*Pf
2 * V  *
     P
                                                        A  *
                                                                                                y
                                              A*,
                                      (P  -
                                       _E	
(6.2)
                        Equation 6.2 will be valid under the following
                        conditions:

                             1.  When the fluid is continuous with res-
                                 pect to the particle and the particles
                                 are larger than 1.0 micron. Particles
                                 smaller than 1.0 micron are able to
                                 slip between the molecules of the fluid.
                                 Because the friction force is caused
                                 by bombardment of the particle by the

-------
Particle Settling Dynamics
       molecules of the fluid,  the friction
       force becomes intermittent as the par-
       ticle slips between the  molecules.   The
       particle, therefore,  increases in vel-
       ocity as it slips between the molecules.
       The actual average velocity will there-
       fore be higher than predicted by the
       equation, where the friction force is
       assumed to be constant.

   2.  When the particle falls  in a free state,
       or when it is not hindered by other
       particles in the surrounding fluid.
       If the particle concentration is great
       enough to cause additional friction
       forces (from bombardment by other
       particles and the resulting transfer
       of momentum), the settling velocity
       will decrease.

   3.  When the movement of the particle is
       not slowed by hitting a  boundary.  The
       fluid is then assumed to be infinite
       in extent.

 Equation  6.2  can  be further simplified wher
 particles  under investigation  are  assumed
 to be  spherical.   Therefore, the projected
 area of  the particle  (A), and  the  volume
 of the particle (v  ) may be given  by the
                  P
 following equations:
                            (7.0)
                           (7.1)
where:

     D  is the diameter of the sphere.

     Substituting equations 7.0 and 7.1 will
reduce equation 6.2 to:
                                                    and
          2 * TT * D
                                          (8.0)
                /4 * D  *
                   	E_
                                           1/2
                                          (8.2)
                       * CD * Pf
              As  previously  stated,  the friction factor
         or  drag  coefficient,  C  ,  in equation 8.2 depends
         upon  the flow  pattern surrounding the particle.
         This flow pattern can be  characterized by the
         Reynold's Number^(  see  figure 5 ).   For example,
         in  air,  a Reynold's Number  of more  than 1000
         indicates the  turbulent flow range  while the
         laminar  flow range  occurs at a Reynold's
         Number of  less  than 1.0.
               .laminar flow
               region
                                   turbulent _|
                                   flow region
                     Reynolds number,  Re

         Relation between  the  drag  coefficient (log10Cn)
         and  the Reynolds  number  (logioRe'  Ior spheres.

         Figure 5.

         Many authors have attempted  to write  equations
         to describe the variation  of the drag coef-
         ficient with the Reynold's Number.  For  example,
         Langmuir and Blodgett  wrote the following equa-
         tion to describe the drag  coefficient for
         Reynold's Number  ranging from 1 to  100
         (see figure 6):
                                                      (9.0)
        J24
        R
                                                                  1 + 0.197R °'63 + 0.0026R
                                                                            e              i
                                                 l.38\
                                                -    )
          4 * D  *
               P
               3 *
(8.1)
For a Reynold's Number ranging from 1000  to
200,000 the Drag Coefficient is relatively
constant; it varies only in the range from
0.38 to 0.50.  This region is usually referred
to as the Newton Region and an average Drag
Coefficient of 0.44 is often used in calcul-
ations (see Figure 7).

-------
                                                                        Particle Settling Dvns
10
Q 10 4
- 10 3
c
3 10
*S 10
u
01
0 1.0
Q
10 '

\







V
\







V
\








V
\



•• 	 I 	 -
I,^N TKV
^TO'j'W
Jss»vJ '"

?*>„
* •*" 4







*^**r.







-—







	







X
Thus, for a Reynold's Numbei
200,000, equation 8.2 become

f = 1.74 g DP (PP Pf>
Pf '


The flow region of greatest
cous flow region occurring
                                                                                 1/2
                                                                                         (10.0)
    id4  10   10   10'   i.o  10   io2  io3   10*  ioJ  io6

              Reynolds number   Re

Relation between the drag coefficient  (log10CD)
and the Reynolds number  (logioRe) for  spheres.

Figure 6.
of less than 0.1.   For this region,  Stokes cal-
culated the Drag Coefficient from the following
formula:
                     (11.0)
                                                     where
io5
Q 10 "
U
3
^ 10
1 10'
**-.
0 10
U
I? i.o
Q
10 '
newtonian
region
\





\
\





s
\





\
\





\
\





\






^^
' 1 ", "I "
-:fi-
^»JIj
^LJ
^^i^wW^ftiy
; -• -. |-":- 1 A \^"\
,-,:..J,:l'-i...-jb.\ J
-4-3-2-1 2345
6
              Reynolds number ,   Re

Relation between the drag coefficient
and the Reynolds number  (logj0Re) for spheres.

Figure 7.
                                                                              (ii.D
                                                     and
                                                         JJ   is  the  viscosity  of  the  fluid.

                                                     which,  when substituted  into equation 8.2,
                                                     reduces the settling  equation to the familiar
                                                     "Stokes Law"  equation:
            18
                                                                    (P _
                                                                      P
                                (11.2)
                                                     For particles smaller than 1.0 micron, the fluid
                                                     will become discontinuous, resulting in higher
                                                     settling velocities than can be predicted by
                                                     Stokes' Law.  To correct for this effect,
                                                     Cunningham deduced that the Drag Coefficient
                                                     should be written as:
                                                            K  * R
                                                             m    e
                                                                               (12.0)

-------
 Particle Settling Dynamics
 wnere K  is the Cunningham correction factor to
 account for particle slippage.

 Thus, Stokes' Law becomes
                                   (12.1)
 TABLE 1.  K  FOR AIR AT ATMOSPHERIC PRESSURE
            m
              18,
  Table  1  shows  the  effect of  temperatures and
  particle diameters  on  the Cunningham correction
  coefficient.
Particle
diameter
(microns)
0.1
0.25
0.5
1.0
2.5
5.0
10.0
Cunningham
factor ( Km
70°F
2.88
1.682
1.325
1.160
1.064
1.032
1.016
correction
)
212°F
3.61
1.952
1.446
1.217
1.087
1.043
1.022

500°F
5.14
2.528
1.711
1.338
1.133
1.067
1.033
                                            BIBLIOGRAPHY
(1)  Lapple,  C.E.  in J.H.  Perry Edition;
     Chemical Engineer's Handbook,  3rd Edition,
     pg.  1021,  McGraw-Hill,  New York (1950).

(2)  Langmuir,  I.  &  Blodgett,  K., American Air
     Force Tech.  Report  #5418  (1946).

(3)  Strauss,  W.,  Industrial Gas  Cleaning,
     Pg.  122,  Pergamon Press,  1966~.

(4)  Torobin,  L.B. and Guavin,  W.H.  Canad. J.
     Chem.  Engineering

          Volume  37    Pg.  129,  167,  224      1959
          Volume  38  - Pg.  142,  189            1960
          Volume  39  - Pg.  113                 1961
(5)   Stern,  A.C.,  Editor, Air Pollution, 2nd
     Edition, Academic Press, 1968.

          Volume  I - Pg. 58
          Volume II - Pg. 270

(6)   Brown,  George G., etal. Unit Operations,
     Pg.  70, John Wiley & Sons,  Inc. New York,
     1950.

(7)   Lapple, C.E., Fluid and Particle Mechanics,
     Pg.  281, University of Delaware, Newark,
     Delaware, March 1956.

(8)  Orr, C., Jr. & Dallavalle, J.M.,  Fine
     Particle Measurement,  McMillan,  1959.

-------
                                                              Equivalent Standard
Terminal Velocities of
Spherical Particles
From: Lapple, C.E., "Fluid and 1Q2
Particle Mechanics", University
of Delaware, 1956
101
10°
-a
c
o
Terminal velocities of spherical g
particles of different density S 10"'
settling in air and water at a>
70 degrees fahrenheit under the ^
action of gravity. •"-
o
1 Iff2
DO
j— »
CO
1 lO'3
6
£
Notes :
1. Numbers on curves represent -5
true (not bulk or apparent)
specific gravity of particles


2. Stokes-Cunningham correction
Theoretical
Screen Mesh
§ § §
• 0 
i44-
 10 O C
(N CN O Is
^ -0 T C-
1

















.
0 /
4^
::" r-.Zt"


•-^A
!T
/ /<
!//

' /
, /
/i
! ft
• i-/y!
i / /
1 />
if
'//
- \—f- — ,
' r/
i / /
t \//
< fc=




LI J j j
//N

^r
0 t ^
-' / ^
b' 0
33 )
f * I

77 '
r/ /
///
J^-i
f~ ' ,
I2^t^
/A'
r—f--t
//
TTfl







N — -
0 0
O if)
i !
} C
- a
8





J

[ /

/

'' ^f. >
!;'//
£---


' •<
;

w&\
f /

''v*
o /
^
/
^
•0 CD *o <
•o -* rt c
I 1


^
.' t
J{j - 1

f J j
/ .'
T~T
A £
<$*!
£/ ,
f / (
/ / 1
///
f/[/\
/J

7
v* r
i r~~t~
' ' f
f i / 1
> //
X

'











c
















n o ^r
N CN •-
N
O
1

*£,
1 ^^X
' ' ^^
[P
T



.X



/;
>
' j^
, x1
'-• ^^
LK]/

j




















CD « ~*
j 1 1
^
^^ * '
y
S*s"'


f
*
S*'

~^ — *^
. r
r t*
r *
(?























n CN
CO
1

d '



• '
, j
,
„,
-;
U-flj
1 '
( '
























                                                                                       !  10'
                                                                                         10'
                                                                                         10'
                                                                                          r,-2
                                                                                         10
                                                                                          ,-4
                                                                                         10-
particles settling in air.
3.  Physical properties  used:
 10        100       1,000
Particle diameter,microns
   10'
10,000
Fluid
Air
Water
Temp.
F
70
70
Viscosity
centipoise
0.0181
0.981
Density
Ib . mass ,.<_
cu. ft.
0.0749
62.3
                                                          7
                                                                        PA.C. pm.40b.2.70

-------
SECTTON e
The Separation of Particles into
  Size-Fractions

-------
                   THE SEPARATION OF PARTICLES INTO SIZE-FRACTIONS
 I  Accurate and rapid methods of analysis of
 particles in a sample consist of:

 A Dividing the sample into closely-sized
    fractions,  followed by

 B The determination of weight, size,  or
    count in each of the fractions
 II  Processes commonly used for separation
 of particles into size-fractions are:

 A Screening (or sieving) for grading "coarse"
    particles

 H Procedures based on the motion of particles
    in a fluid for grading of "fine" particles

    1  Klutriation procedures

    2  Sedimentation procedures
III  SIEVING
 A Procedure

    1  In the process of sieving,  the particles
       are passed through a series of vibrating
       screens,  the openings of which get
       progressively  smaller.
  B Screen Six.e

    1  There are two series of screen -scale
       .sieve.s commonly used  in Hie llmled
       State's at present.  These- are the U. S.
       and Tyler, the si/.e.s of which are shown
       in Table 1.
 C Lower Particle-Size Limit

    1  Particles are graded down to a fineness
       of 74-microns (Tyler or U. S.  200 mesh)
       and may be extended to a lower limit of
       43 or 44 microns (Tyler and U. S. 325
       mesh).
D  Need for Specifying the Type of Analysis
   Used

   1  Because of the wide variation  possible
      in making the actual test, and of
      variations in particle shape, densities,
      and properties of the material (such as
      hygroscopicity and stickiness) it is
      impossible to specify a single process
      universally applicable  to all particles.

   2  Sieving to completeness is impossible.
      Hence, it is necessary to define the
      endpoint by one of the following  methods:

      a  By specifying a standard time of
         sieving

         1)  This may be good when there is
              no great variation in  fineness.

      b  By specifying that screening  must
         be continued until the weight  of the
         particles passing the sieve per
         minute is less than a stated
         percentage of the total weight of
         sample taken.

         1)  This is complicated.  It is
              unsatisfactory for routine
              work, but is good practice
              for control equipment acceptance
              tests.

      c  B> stating that sieving must be
         continued until the weight of material
         passing the sieve per minute  is less
         than a certain percentage of ihe weight
         of the  residue on the sieve  considered.

         1)  This is very sound practice,  but
              extremely  complicated.

   3  Since there is no single screening
      procedure adaptable to all particles,
      and wide variations in methods used are
      inevitable,  the results of any  sieve-
      analysis should be accompanied by a
      description of the procedure followed in
      making the size-fractions.
 PA.C.pm. 65. 9. 60

-------
 Separation of Particles into Size-Fractions
                   Table 1.  STANDARD U.S. AND TYLER SCREEN SCALES
U.S.
Sortn
Salt,
Meth
Number
400
325
270
230
200
170
140
120
J»X)
80
70
60
50
45
40
35
30
25
20
13
16
14
12
10
8
7
6
5
4




I ncket
0.0015
0.0017
0.0021
0.0024
0.0029
0.0035
0.0041
0.0049
0.0059
0.0070
0.0083
0.0098
0.0117
0.0138
0.0165
0.0197
0.0232
0.0280
0.0331
0.0394
0.0469
0.0555
0.0661
00787
0.0937
0.111
0.132
0.157
0.187

Nominal
Aperotnrt
Width
Mieront
37
14
63
62
74
88
105
125
149
177
210
250
297
350
420
600
590
710
840
1,000
1,190
1,410
1,680
2.000
2,380
2,830
3.360
4,000
4,760


Win
Diameter,
IndM
0.0010
0.0014
0,0016
0.0018
0.0021
0.0025
0.0029
0.0034
0.0040
0.0047
0.0055
0.0064
0.0074
0.0087
0.0098
0.0114
0.0130
0.0146
0.0166
0.0189
0.0213
0.0240
0.0272
0.0299
0.0331
0.036
0.040
0.044
0.060
Tyler
Serin
Seal*.
Uetk
Number
400
826
270
250
200
170
150
115
100
80
65
60
48
42
35
32
28
24
20
16
14
12
10
9
8
7
6
5
4




/neJUf
0.0015
0.0017
0.0021
0.0024
0.0029
0.0035
0.0041
0.0049
0.0068
0.0069
0.0082
0.0097
0.0116
0.0138
0.0164
0.0195
0.0282
0.0276
0.0828
0.0390
0.046
0.056
0.066
0.078
0.093
0.110
0.131
0.166
0.186

Nominal
Apcrature
Width
Miermt
37
43
63
61
74
88
104
124
147
176
208
246
295
351
417
495
589
701
883
991
1,168
1,897
1,661
1,981
2,862
2,794
3,827
8,962
4,699


Win
Oiamttrr.
Indut
0.0010
0.0014
0.0016
0.0016
0.0021
0.0024
0.0026
0.0038
0.0042
0.0066
0.0072
0.0070
0.0092
0.0100
0.0122
0.0118
0.0126
0.0141
0.0172
0.0286
0.026
0.028
0.086
0.033
0.082
0.0328
0.086
0.044
0.066
IV  AIR ELUTRIATION

 A  Definition

    1   Air elutriation is the process of
       separating particles into size-fractions
       by varying the velocity of an upward
       current of air.

 B  Application
    1   Size-separation by air elutriation is
       desirable when the particles to be
       analyzed are, in practice, subjected
       to the grading action of air flow.
C  Upper Particle-Size Limit

   1  Air elutriation methods are employed
      for the analysis of particles of sub-
      sieve size (less than 80 or 40 microns).
D  Principle of Particle-Size Separation

   1  The upward velocity of air required
      to just separate a given particle-size
      is calculated from Stokes' Law.  A
      convenient form of the law is expressed
      in the following:

-------
                                                      Separation of Particles into Size-Fractions
     D2g
     -jti
y
for N
                              <0.1
(i)
      Where:
               upward velocity of air (cm/sec)

               size of particle to be removed
               (microns)

               local acceleration due to  gravity
               (cm/sec^)

               density of the particle (gm/cm )

               absolute viscosity of the air
               (poises)
E  The Roller Analyzer (Figure 1)

   1  Description

      Four vertical tubes, D-l, D-2,  D-3,
      and D-4,  are mounted so that they may
      be rotated about a central post thus
      bringing the  tubes successively into
      position over U-tube, C.

      U-tube  C is constructed of heavy-walled
      Pyrex tubing,  1-inch inside  diameter.

      Tubes D-l, D-2, D-3,  and D-4 are
      stainless steel and highly polished on
      the inside.  The inside  diameters are
      respectively 9.0, 4.5,  2.25,  and 1.125
      inches  (these may vary according to
      design).   All tubes are  grounded
      electrically.

      Gooseneck, E, is -^-inch  stainless steel
      tubing,  polished on the  inside.  At the
      end of E is a machined  rubber stopper
      on which is fitted an  extraction thimble,
      F.  which  collects the particles blown
      over  from the vertical tube to which  it
      is connected.
                                        Tho elulrlalor conilili of the following
                                        principal parti 01 shown In lh« drawing.
                                     ;yAi':'Alr Inlef
                                     -J'l^'Cornftrettot
                                  , p-4, •\Yf'f Cha^«j',||
                       • "...MI T u     ' £•' Glasi Ooo»«ri«?l(ft
                       il U  Tuba     i  '-.   , .....,..'.' ittif^j,
                    ;°" Chamber  '     P) PpP?r.^^^'l^
                       ff Chambef   . v O. Rotary fappiriftM
                          .  • .  ' '  i ' l./i-ti ..  ' ." •' '> ..Vi '.I- "li'.tfllM
                                             Roller-Analyzer
   2  Operation

     A weighed quantity of particles (25 grams)
     is placed in U-tube,  C.   An air jet from
                                                Figure 1

                               The Thermix Corporation, Greenwich,  Conn.

-------
Separation of Particles into Size-Fractions
      motal noz/.le, IB, impinges on the
      hoi torn-center of U-tubc, C.  Nozzles
      of desired  sizes are installed to provide
      proper dispersing action for the size-
      fraction being separated.  The flow rate
      is measured by a flowmeter  attached to
      air-hose, A.

      The rubber-tipped hammer head,  G-l,
      actuated by a motor-driven cam,  keeps
      U-tube, C,  constantly agitated.  The
      mass of particles in U-tube, C, moves
      as a whole in a direction opposite to
      that of the  air in the jet.  The mass of
      particles also.rotates  in a clockwise
      direction so  that fresh portions of the
      mass are continuously exposed  to the
      action of the air jet.

      Tapper,  G-2, affects the rate of separa-
      tion and also the uniformity of fractiona-
      tion by shaking down into U-tube,  C,
      those particles that repose in the  lower
      cones  of the vertical tubes.

      The lowest size-fraction is separated
      first; and,  for each size-fraction  a new
      thimble is  used.  Before each run, the
      air is conditioned to constant humidity
      (the air is  drawn through 1:1 sulfuric
      acid solution) and blown through the
      thimble until the thimble reaches
      constant weight.

      The analyzer must be  calibrated to
      determine the endpoint of each run.
V  SEDIMENTATION IN LIQUIDS
A  Definition

   1  Sedimentation is the process of
      separating particles into size-fractions
      according to their velocity of fall in a
      column of liquid at rest.
B  Upper Particle-Size Limit

   1   Sedimentation methods are employed for
      the analysis of particles of sub-sieve
      size (less than 80 or 40 microns).
C  Principle of Particle-Size Separation

   1  The velocity of fall of a given particle
      in a liquid under known conditions is
      calculated by using Stokes' Law.  A
      convenient form of the law is expressed
      in the following:
                       -8
 p(S)
                           for N   <0.1  (2)
      Where:
         p(S)
         D
vertical velocity of the
particle (cm/sec)

local acceleration due to
gravity (cm/sec^)

density of the particle
(gm/cm3)

density of the liquid (gm/cm3)

size of the particle (microns)

absolute viscosity of the
liquid (poise)
D  A Sedimentation Method (Reference 5)

   1  Prepare assemblies as in Figures 2,  3,
      and 4.

   2  Determine the viscosity and density of
      the liquid, and the density of the
      particles to be analyzed.

   3  Calculate the time of fall over a vertical
      distance of 6-cm for (0-10)  and (10-20)
      micron fractions (Figure 3); and the
      time of fall over a vertical  distance of
      20-cm for fractions larger  than 20-
      microns (Figure 4).

   4  For the (0-10) micron fraction, place
      one-half gram of the sample in the
      300 ml beaker shown in Figure 3.  Fill
      the beaker to the upper mark with the
      liquid and stir well to obtain good
      dispersion.

-------
                                    Separation of Particles into Size-Fractions
                                                                      1
Equipment for Determination of Particle Size by Liquid Sedimentation

                          Figure 2

   Western Precipitation Corp. , Los Angeles, California
                                         2 so rt I,
                         5uc-i-lon     Y\ & rgcJuafet
                                     I  \Cylfi de. r
       -J_J	1	

        \	-Post-Lion   of 300ml.
Figure 3    Beaker


                  Ti
                                            Figure 4
                     m*

-------
Separation of Particles into Size-Fractions
   5  Stop stirring and immediately start the
      timer.

   6  At the end of the time calculated in
      Step 3 for the (0-10) micron fraction,
      raise the beaker so that the tube
      connected to  the filter flask is properly
      placed at the lower mark on the beaker.
      Rapidly  draw off the liquid to the lower
      mark.  The particles drawn off are
      smaller than 10 microns.

   7  Filter the liquid drawn off through a
      Gooch crucible.

   8  Without  emptying the beaker, fill it
      once again with liquid to the upper
      mark.  Stir well.

   9  Repeat steps 4, 5,  6, and 7 until the
      quantity of minus  10-micron particles
      remaining in the beaker is negligible
      (15 to 25 decantations, or more, may
      be necessary).

  10  Dry the  material collected on the filter,
      and weigh.  This weight represents the
      (0-10) micron fraction.

  11  Repeat steps 3  through 9 for the (10-20)
      micron fractions using assembly shown
      in Figure 3.

  12  Repeat steps  3  through 9 for the (20-44)
      micron fraction using assembly shown
      in Figure 4.

  1M  Material remaining after the (20-44)
      micron fraction is termed  residue,
      weighed, and reported as  the plus  44
      micron fraction.

  14  AfU'r weighing  the crucibles containing
      the dust  fractions, take  portions from
      each crucible and examine under a
      microscope to observe the conformity
      of the "actual" size-fractions to the
      "calculated" size-fractions.  Also
      examine the fractions for agglomerates
      and other irregularities.
 VI  NEED FOR SPECIFYING THE METHOD
     OF PARTICLE -SIZE ANALYSIS

  The size of a given particle determined by
  one method of analysis will not necessarily
  be duplicated for that same particle using
  another method of analysis.

  For example:  In the sieving analysis, any
  particle regardless of its shape,  that will
  pass a 325 mesh (Tyler) sieve is reported
  as having a size of 44-microns because the
  sieve apertures are squares averaging
  44-microns on a side. But, if this same
  particle were subjected to  air elutriation
  or sedimentation in a liquid, it does not
  necessarily follow that the  size of 44-microns
  will be determined.  Let's  say that the 44-
  micron particle as measured by the sieve
  is a flat disc.  This same particle, then,  will
  settle slowly in a sedimentation method of
  analysis,  and according to  Stokes1 Law will
  have a size much smaller than 44-microns.

  Hence, in reporting particle-sizes,  it is
  necessary to specify the method of analysis.
VII  A BASIC DEFECT OF PARTICLE-SIZING
     TECHNIQUES  (References 7,  8)
  A  The Dispersion of Particles in the Flue and
     in the Particle-Sizing Apparatus

     1  In a sample of particulate matter
        extracted from a flue, the original
        state of dispersion of particulates in
        the carrier gas is permanently destroyed.

     2  During the  particle-sizing procedure,
        the particles are redispersed.  How-
        ever,  there is no certainty that the
        original dispersion that prevailed in
        the flue is  duplicated in the particle-
        sizing apparatus.  In fact, a re-creation
        of the original dispersion is very
        unlikely, especially when the  fine
        particles are considered.

-------
                                                   Separation of Particles into Size-Fractions
B  Examples of Particle Aggregation and
   Dispersion

   1  Figure 5 illustrates the problems of
      redispersion of a dust sample.
A.  Loose, Porous Aggregate
 O
°
o
                                     °
                                         O
                                         " O
                               o
                                     o  o  o
                                   O O o (-.
                                     a  °
B.   Closely Packed Aggregate
C.  Small particles adhering to a large one
         modification of effective  size when
         prepared for  sizing by liquid sedi-
         mentation procedure.

      c  Figure 5C depicts a large particle to
         which smaller ones are adhering.
         During the dispersing action in
         particle-sizing apparatus,  the smaller
         particles are  detached.

      d  Figure 5D illustrates the air sizing
         method in which particles originally
         well dispersed are imperfectly
         redispcrsed.

SELECTED REFERENCES

1  Roller,  P.S.  U.S.  Bureau of Mines
                        Technical Paper No. 490.  1931.


                   2 Roller, P.S. Ind. Eng. Chem. (Anal. Ed.)
                        3; 212-16.  1931.

                   3 Roller, P.S. Measurement of Particle
                        Size With an Accurate Air Analyzer.
                        Proc.  Am.  Soc.  TejEaingJWtls_.  35th
                                                   Part II,
                                                        Annual Meeting 1932.
                                                        Technical Papers,  p.
                            3_2:
                            607.
      O
D.   Well dispersed particles
                  Figure 5


      a   Figure 5A represents a single
         aggregate,  somewhat loose and
         porous as it might exist in a flue.
         When deflocculated in preparation for
         microscopic sizing or liquid sedimen-
         tation,  the density and effective size
         arc quite different from the original.

      I)   Kiguro r>H shows a  closely packed
         aggregate which may  undergo drastic
                                                  4  The Problem Characteristics Industrial
                                                        Dust.  The Thermix Corporation.
                                                        Greenwich,  Conn.
                   5  Particle jize Analysis.   Bulletin No.
                        G402R.  Western Precipitation Corp.
                        Los Angeles, Calif.

                   6  Traxler,  R. N.  and Baum,  L. A. H.
                        Determination of Particle Size
                        Distribution in Mineral Powders by
                        Air Elutriation.  Rock Products.
                        June,  1934.

                   7  Hemeon,  W. C. L. ,  Raines,  G.F., Jr.,
                        and Puntereri,  S. D.   Rating  of Dust
                        Collectors According to Dust Settling
                        Velocities.  Paper 60-54, APCA, 53rd
                        Annual Meeting.  Cincinnati, Ohio.
                        May 26, 1960.

                   8  Hemeon,  W. C. L.   Dust  Particle Inertia
                        and Various Consequences.   ASHRAE
                        Journal Section, Heat Pip,  and Air
                        Cond.   247-250.  ~Feh>r~uary,  1957.

-------
                                               7
SECTION 7
Notes on the Analyses of Particle
  Size Distributions

-------
                                        (Excerpts)

                                NOTES  ON THE ANALYSES  OF PARTICLE
                                    SIZE DISTRIBUTIONS
                                        Richard  Dennis
                                       GCA CORPORATION
                                   GCA  TECHNOLOGY  DIVISION
                                 Pollution Control Laboratory
                                     Bedford, Massachusetts
                                         May  1972
PA.C.pm.102.4.73

-------
                                SECTION 1



              MATHEMATICAL REPRESENTAILON OF  PARIICLE SIZE*



      The  simplest  representation  of  a particle  size  distribution  is  a


size-frtquency  curve  which, shows  the number  of  particles,  N,  present


for an\ specified  diameter,  D.  Since most dusts  are composed  of  an  in-


finite  range  of  particle  sizes, it  is first  necessary to classify  par-


ticles  according to some  consistent  pattern-  Then N may be defined  as


che number  of particles within  a  specified size group having  finite


boundaries  and  typified by some average  diameter, D.   For  most  particle


distributions,  a characteristically  skewed curve  results.  Figure  1,  It


is convenient to graph the fraction  of the total  number of particles, P


rather  than the  absolute  number of particles, N,  within a  size  range
    >-   n)
    CD   (X
    Q-i •-
      O •-
        O
        •fl
                                 D
                                  1
                         Par: ic If Diame-- r - D
             Figure 1   Typical size frequency distribution
~Matnematical derivations of sizt parameters in Sections 1 and 2 arc
 based  largely on analytical methods described by W E  Ran/ in Tech
 Rpt   No  1,  University of Illinois Engineering Experiment Station,
 April  30,  1950   USAEC SO-1000,

-------
 to facilitate the derivation of descriptive size parameters.   When no


 mathematical relation between P and D has  been stated,  the  general


 equation for the curve in  Figure 1  is



                               P = f(D)



      The fraction of  particles,  dF,  within the  size  range,  dD,  is  ex-


 pressed as



                             dF  = f(D> dD



      Bv introduction  cf  the  proper  constant  (included in the term  f(D)i

 the value  1  0 can be  assigned to  the  cumulative  fraction of all parti-

 cles  within  the  distr ibut. too  form D  = 0 t o D = *•   On this basis,  the

 toral  area under  the  curve represents 100  percent of the particles,

 i  e.,  the  fraction F  = 1.0.  Similarly, the area between any arbitrary

 size  limits  D and D  depicts the fraction of the total number  of par
 t ic les witr>in tnat range   Thus

                            CO
                            f-^

                           i   f(D> dD = 1
                          v-'
                           o


and

                n
                ^ 2
                / ^
               I   f. (D i  dD = fraction in range D  to D7

              X


     The cumulative distribution,  which  indicates the fraction of the

total number of  particles in the diameter  range 0 to D, appears as

-------
                                D
                                r
                                '  f(D) dD                         (i)
      In all cases where the surface area and volume  of uniformly
 shaped particles (spherical or non-spherical) are  proportional  to D
      3
 and  D ,  respectively,  Equation (1)  forms the basis for defining the
 following diameters:
      (1)   D  = number  average  particle  diameter, the particle diameter
 which multiplied by the total  number  of particles will  give the sum
 of all the particle diameters.
      (2)   D  - surface  average  particle diameter, the  diameter of the
            s                 —	
 particle  whose surface  multiplied by  the  total number  of particles
 will  give the  total surface  of  all  the  particles.
      (3)   D  - volume average particle  diameter, the diameter of the
 particle  whose volume multiplied by the  total number of particles
 will  give the  total volume of all the particles.
                  3  2
      (4)   D   = D /D  = diameter of the particle with the same ratio
            vs     v  s
 of volume  to surface as that exhibited by all particles in a given
 s amp 1e.

      In addition  to mean diameters, other cumulative  distribution
 functions can be  defined:
     Cumulative surface area distribution function  = the fraction  of
the total surface of all the particles contributed  by particles  with
diameters from D  = 0 to D = D.

-------
                                      D


                       F(D2) = 1/D?  /  D2f(D) dD
      Cumulative volume distribution function = the fraction of the



 total volume of all the particles contributed by particles with diam-



 eters from D = 0 to D = D.


                                      D

                          3       3  P  3
                       F(D ) - 1/D   /   D f(D) dD                  (3)
                                  ^ U

                                     o



 Median Diameters





      (1)  D  ,  or  M  = number or  count median diameter,  the particle
            nmd      g	      r


 diameter where  half the total number of particles  have diameters  greater



 than and half have  diameters  less  than D
                                         nmd'
                    D
                    ouna
                                    r
                                    nmd
/     f(D) dD -  /    f(D) dD - |               (4)


o                D
      (2)  Dm   or M  = mass median diameter, the particle diameter



where one half the mass of the total number of particles is represented



by particle diameters less than or greater than D   ,.
                                                 mmd



                  mmd               oo


                /     D3f(D) dD =  /    D3f(D) dD = ^             (5)
               \J                  \J                 £

                o                  D  ,
                                    mmd




     Since the size distributions outlined in the previous section have



been outlined in general form,  it is necessary to find some mathematical



relationship between  P  and D if solutions other than by graphical

-------
integration  are desired.   Several  curve  fitting methods have been pro-


                                                         (1 2)
posed  including the commonly used  Hatch-Choate equations   '   which



have been derived  in  the  next  section.   Other size distribution func-



tions have been described by Rosin-Rammler,    Roller,    Nukiyama-



Tanasawa,  ' Gaudin-Schulhmann/ ^ Wynn-Dawes,    Sichel,  ' Kottler ,



and Dalla Valle.      The criterion for  selecting one relation in pref-



erence to another  should  be the goodness of fit to experimental measure



ments.  All  techniques are the same in principle, that is, the term



f(D) is usually defined by D (or function of D) and two arbitrary con-



stants, the  first, some average diameter and the second,  a measure  of



homogeneity  (range of sizes).  If the Hatch-Choate equations  are used,



average diameter may be represented by M  and the homogeneity factor
                                        o


byo-g

-------
 ROLLER




      The Roller distribution appears in the form:






                           w = a d   exp(— T)






 where w is  the weight  percent of the particles  possessing diameters less



 than d,  and again,  a and  b are constants.   By differentiation,  the weight



 and number-frequency distributions appear  respectively as:
                      dw
                          = a
 and:
                     d(d)-aV2   II

                                  d4    d2
                     d n    a /  1    .  b	\    /  b

                     ,,..  = - 1 	T + —q~ ) exP \ ~ T
                    d(d)    p\2   Z.     2.  I    \  d

                                  d2    d2
where p = particle density.




     A graph of the function ln(w/d^) versus -7, which should appear as a




straight line, permits estimation of the constants a and b.




NUKIYAMA-TANASAWA
     The Nukiyama-Tanasawa distribution is often applied for the analysis



of atomized liquids where an extremely broad size range may be encountered



Three constants, a, b, and c, must be used in this system.  In equation



form:
                                  5a

-------
Here, n refers to the number of particles within the diameter range A d.




Proper selection of the constant c, which does not change significantly




for a given nozzle regardless of flow rate, allows for a graphical solu-




tion to Equation    when  In  ( —^ ) ( --^-r  )  is plotted against d .  Again,




valid results obtain only when the  plot is linear.






WYNN-DAUES






     The Wynn-Dawes relationship has been applied to the analyses of mine




dusts in which mixtures of dust from several sources may be found.  As a




result, bi-modal distributions frequently occur.
              — = Q a  exp(- a D) +  (l-Q)p exp(- p D)








     In the above equation,,  n refers to the number of particles in the




size range represented by diameter D, and Q, a and P are constants.




Specifically, Q is the fraction of dust characterized by the constant Q




and 1-Q the fraction characterized by (3.  If Q is equal to 1.0 (implying




that the dust arises from a single source and probably can be described




by a uni-modal system),  the simplified equation results, viz:
                          d n         /•    rxi
                            -  = a exp(- a D)
                                  5b

-------
 ROSIN-RAMMLER




      The Rosin-Rammler  equation may  provide  a  convenient means  of  norma-



 lizing size distribution measurements when the cut-off  points for  the



 smallest and largest  particles present are well  defined, e.g.,  the frac-



 tion of particles  retained between two screen  sizes.  The weight percent



 of particles, w, having a diameter greater than  d is expressed  as:





                          w = 100 exp (- ad  )






 where a and b are  characteristic constants for the distribution.   The



 weight-frequency distribution obtained from  the  above equation  is:
                         =  - 100 a b d^-^expC- adb)
     The constant a increases with decreasing particle size and the con-



 stant b is somewhat analogous to standard deviation.  By rearrangement of



 Equation     , viz:






                     In In (—)  =  In a + b In d
                             w





 graphing of experimental measurements on log-log paper gives b as the



 slope of the straight line of best fit.   The value for a is then obtained



by calculation
                                 5c

-------
          TABLE 1
TABULATED SIZING DATA FOR
100 PARTICLE MEASUREMENTS
o>
Portion
Size
1
2
3
4
5
6
7
8
9
10
(2)
Diameter
Microns
0.44
0.62
0.88
1.25
1.76
2.50
3.53
5.00
7.07
10.0
(3)
Number of
Particles
19
29
18
12
7
8
2
4
0
1
(4)
Cumulative
Number and
Percent
19
48
66
78
85
93
95
99
99
100
(5)
Cumulative
Number
119
148
166
178
185
193
195
199
199
200
(6)
Cumulative
Percent
59.5
74.0
83.0
89.0
92.5
96.5
97.5
99.5
99.5
100
            Curve  1,  Figure  2   Curve 5,  Figure  2
            15

-------
     10.0
CO
z:
o
(T
O
 I
.80
)PE
— —
Ds 55-
1.57 -^
1.49 |||
1.62 ^ji

Dv
2.15
2.43
- 1 ^
.11 i . . i

— 	
^ir

• • •



He tU-.-.
US it


_ t
1 I

\ \ > '

— — — — — -
-—••-•••••• '
-H-rr i
ftj=t 4 i : ' r
---) 	




t j

t P
J >
^9
p==3 iu_^-_ .
^Fl




i rrhr-
i TTTF
p
1 'j ^
ir
'^\f M
fy I

2
1 L 2
tj-L
^
T TT-; :
'IT4
1 ! ' ,

•LliT^
~ ' "T" —
fflf
tJr\\
j j j
mm
jSr
Jf
71
'

T
J^
p— ^
-4—



fe^'-J
^^
^^5 •"::::
	
^









p^

^t±t
f — — r~









	 . — r->_
— > 1 I i

! 1 ' : i
*M
m-r n^
rp:: —
	 r—
1



=J






j
1
1 j
i
Jt^
— 1
^ 1 |.^ 	

—
— -


,






^
—


LLi_l 	 1 	
rrr
TT
3»
M-p





;i i
i i i
: !
i ' '
. _
— — r
5V
T— t
~A
	 L
- 1
^




i

1















1= ---, 	
____
rr?: t i — ! — TT=



±~
^
4 —
T-1—
       0.01
                1.0
10
30   50  70
90
99
99.99
                 PER CENT  LESS THAN OR EQUAL TO  STATED  SIZE
                       Figure 2.  Particle Size Distributions from Table 1.

-------
                                            TABLE 3

                      DATA TABULA!CON FOP  PARTICLE SIZING BY TRUNCA1KD
                                  MULTIPLE  TRAVERSING SYSTEM*"

1 2
Tr -i verse
Number 0.25 .44
.44 .62
1 19 29
2
3
4
5
6
7
8
9
10
Tonal 19 29
Sum ~?r per
Traverse 19 29
Cum a lac i ve
Percent 19 48
Porton Size and Size Range in Microns
345678

.62 .88 1.25 1.76 2.50 3 53
.88 1 25 1.76 2.50 3-53 5.00
18 12 78 2 4
11 5 4 1
3 1
5 2
0
2




18 12 18 13 14 10

18 12 9 6.5 3-5 16

66 78 87 93 5 97.0 98.6

9

5.00
7.07
0
1
1
0
2
1
3
2


10

1.2

99.8

10

7.07
10.0
1
0
0
0
1
0
1
0
0
0
3

0.3

100. 0+


Total
Number

100
122
127
134
137
140
144
146
146
146






See Reference 8 - Section 1-

-------
                             TABLE U

             NUMBER OF PARTICLES TO BE COUNTED TO

                  ACHIEVE ANY GIVEN ACCURACY
Weight % of Particles
  in Any Size Range    2         5        10      15        20
         (m)
 Expected Accuracy        No.  of Particles to Be Counted
         (s)                             (m)
2%
1%
0.52
0.2%
0.1%
(3)*
(6)
16
100
1+00
(8)
25
100
625
2,500
25
100
1+00
2,500
10,000
56
225
900
5,600
22,500
100
1+00
1,600
10,000
Uo.ooo
                                31

-------
                                                                                              01B127-430H
O»
                                                                                         200
                                         RATIO   Ds/Mg , Dv /Mg , Dsmd /Mg , MgV Mg
500     IOOC
                    Figure 6.  Graphical  estimation of characteristic particle diameters  (D6, Dv,  Dsrad and M ')

                              from count median diameter (Mg)  and geometric  standard deviation (
-------
                               REFERENCES
  1.   Hatch,  T.  and Choate,  S.P.,  "Statistical Description of the Size
      Properties of Non-Uniform Particulate Substances," J. Franklin
      Inst  207. 369 (1929).

  2.   Drinker,  P  and  Hatch,  T.,  "Industrial Dusts",  2nd Ed.,  Ch.  10.
      McGraw-Hill  Book Co.,  New York (1954).

  3.   Rosin,  P.  and Rammler,  E., Z.  Verden Ing.  71.  1 (1927);  J. Inst.
      Fuel I, 29 (1933).

  4.   Roller, P.S.,  "Law  of  Size Distribution and  Statistical  Descrip-
      tion of Particulate Materials,"   J.  Franklin Inst.  223.  609  (1937).

  5.   Nukiyama,  S.  and Tanasawa, Y., Trans.  Soc. Mech.  Engrs.  (Japan)
      5,  63  (1939).

  6.   Gaudin, A.M.  and Hukki,  R.T.,  "Principles  of Communition Size  and
      Surface Distribution," Min.  Tech.  Nov.  T-P 1779 (1944).

  7    Wynn, A.H.A.  and  Dawes,  J.G.,  "The Size  Classification of  Airborne
      Dusts in Mines,"  S.M.R.E. Res. Rpt.  No.  28 (South Africa)  (1951).

  8.   Sic>
-------
                           REFERENCES (cont.)
 15.   Smith,  J.G.  and Duncan,  A.J.,  "Sampling Statistics and Applications,"
      McGraw-Hill  Book Co ,  Inc.,  New York (1945).

 16    First,  M.W  and Silverman,  L.,  "Air Sampling  with Membrane Filters,"
      Arch.  Ind. Hyg.  & Occ. Med.  T_,  1 (1953).

 17    Millipore  Filter Corporation,  Bed-ford,  Massachusetts,  "Bibliography,
      A  Reference  Listing of Published Information  Concerning Applications
      of  the  Milltpore Filter," March, 1961.

 18-   Chamot,  E.M.  and Mason,  C.W.,  "Handbook of Chemical Microscopy,"
      Vol   I,  3rd  Ed., John Wiley  & Sons, Inc.,  New York (1958)

 19.   Whitby,  K.T., "Determination  of  Particle  Size Distribution   Apparatus
      and Techniques  for Flour Mill  Dust," Univ.  of Minn.  Eng. Expt.  Sta-
      tion  Bull  No.  32 (January,  1950).

 20.   Dennis,  R. ,  Johnson, G.A., First,  M.W.  and Silverman,  L.,  Performance
      of Commercial Dust Collectors  -  Report  of  Field  Tests," U.S.A.E.G.
      Report  No  NYO-1588, Harvard  University (November 2,  1953).

 21    Corn. M. ,  "Statistical Reliability of Particle Size  Distributions
      Determined by Microscopic Techniques,"  AIHAJ  26,  8 (1965).

22    Dixon, W J   and  Massey,  F.J., Jr.,  "Introduction  to  Statistical
     Analysis," 2nd  Ed.,  p. 291, McGraw-Hill, New  York (1957).

2J    Fatrs, G.L.,  "Developments in the  Technique of Particle Size  Analysis
      by Microscopical  Examination," J.  Roy.  Micros. Soc.  71.  209  (1951)
                                       36

-------
                                               8
SECTION 8
The Effective Particle Size

-------
                               THE EFFECTIVE PARTICLE SIZE
 All particles arr not spheres.  For example,
 liquid droplets and solid particles produced '
 by condensation are usually spherical, but
 solid particles generated by comminution are
 non-spherical and irregular in shape (subse-
 quent abrasion may,  however,  produce well-
 rounded particles).

 The  theories applied to  particle-size
 measurement by air elutriation and sedimen-
 tation in liquids are derived for spherical
 particles.  Therefore,  the  size of a non-
 spherical particle calculated by use of these
 equations  expresses the diameter of a sphere
 that  behaves in a fluid as the particle in
 question in the range of Stokes1 Law.

 Only for spherical particles will  the  "sizes"
 measured by the  different methods of analysis
 theoretically always be the  same. The
 following demonstrates the  significance of
 the "size" of non-spherical particles  in
 measuring devices  and particle-collection
 equipment.

 I  THE EQUIVALENT DIAMETERS OF A
   NON-SPHERICAL  PARTICLE

 The equivalent diameters of a non-spherical
 particle are defined as follows:
Symbol
6
A
Dd
Name
Volume
diameter
Surface
diameter
Drag
diameter
Definition
Diameter of a sphere
having the same volume
as the particle in question
Diameter of a sphere
having the same surface area
as the particle in question
Diameter of a sphere
having the same resistance
to motion as the particle in
question
 Remarks: In the case  of irregular  particles,
 the  sizes may  differ,  not only according
 to the method of measurement,  but also
 according to the orientation of the  particle.
 Of the basic diameters,  only  the volume
 diameter (6) is  independent  of the
 orientation of the particle.  But,  in
 industrial operations as well as in methods
 of particle-size measurement,  the  particles
 can  often be assumed  to be randomly
 oriented.  Hence,  the  mean  value of  the
 projected areas (A ) and the mean of the
 resistances  to  motion  (drag)  can  be taken
 in such  cases.   Also,  since  the mean
 projected area  (Ap)  of a particle  is a
 constant  fraction ('/) of the  surface area
 (Cauchy's theorem),  the  surface area
 rather than  the  projected area is used
 as one of the basic  properties.
It  is  important to  note  that there is no
experimental  evidence (Ref. 1) that  the
drag  of an  irregularly shaped particle
is proportional to  the relative velocity
and viscosity of the  fluid,  even  under
conditions  where Stokes' Law  is valid.
Only  if this is  true  will the  drag diameter
(D^)  of a given particle  have  the same
numerical  value in different  fluids and  at
different velocities.  However, the
resistance  of an ellipsoid is theoretically
proportional to  the viscosity and velocity
(Lamb 1932) and this is  likely to be the
case  also for other  non-spherical shapes.
                                                  II  THE EFFECTIVE PARTICLE-SIZE
                                                      IN IMPINGEMENT,  CENTRIFUGAL,
                                                      SEDIMENTATION, AND ELUTRIATION
                                                      MECHANISMS IN WHICH STOKES1 LAW
                                                      APPLIES

                                                   A  The free-body diagram at terminal
                                                      velocity.
PA. C. pm. G3. 9, GO

-------
The Effective Particle Size
Buoyancy   —g-
                                  P g
                                                         Drag
                                           Weight  - p g   -^
   1  The driving force:
 Driving Force   p g
                      TT 6
                                 6

                                 3
                         P g
                       P  g
   2  The equilibrium equation:

      The particle settles at a constant
      velocity when the driving force equals
      the drag.
                            Sir (j. f     D ,
                               r p(s)  d
                (P
          P(s)
                                                III
        (P     P) g D

             18	When NRe  <°'1   (1)
      Where:
      The size (D )
                 P
      It is the numerical value of (Dp)  that is
      calculated from equation (1) when using
      particle-size measurement methods  in
      which Stokes' Law is valid.  It is seen
      that (Dp) consists of two equivalent
      diameters:  the volume diameter (6)
      and the drag diameter (D^).  Hence  in
      processes where the velocity of a
      particle through a fluid depends on a
      driving force that is proportional to the
      volume of the particle,  (Dp) as
      determined by Stokes' Law is the
      effective size.

      The effective size  (Dp) as determined
      by air elutriation and liquid sedimenta-
      tion procedures is the operative  size
      in filtration, impingement,  and
      centrifugal apparatus in which particles
      are removed from a gas stream  by
      forcing their precipitation upon a surface.
                                        THE EFFECTIVE PARTICLE SIZE IN
                                        THERMAL PRECIPITATION
              D
                            (2)
A thermal gradient provides a driving force
upon a particle suspended in a gas (the
particles are driven away from a hot body).
The theory of driving force (Ref.  1) due to a
thermal gradient indicates that the force
depends on the projected area of  the particle
(or,  since the particles  are normally in
random orientation, the force depends  on

-------
                                                                  The Effective Particle Size
 the surface; area).  The velocity of a particle
 moving in a thermal Hold would therefore be
 dependent upon the  size (  ^  )   However
 inadequate the: theory of thermal force may
 be,  the important point to note is that the
 effective particle size in thermal precipita-
 tion is not necessarily  equivalent to the size
 determined by analyses employing Stokes'
 Law (equation 1) when non-spherical particles
 are involved.
V  THl'J KKKECTIVE PARTICLE SIZE IN
   DIKI'HJSIONAL  PR I-XJLPITATION

Particles are driven toward regions of lower
particle-concentration as a result of their
Brownian movement (the particles  are
bombarded by the  molecules of fluid).  The
driving force depends only on the concentra-
tion of particles; hence,  the effective  size
is  simply the drag diameter (Dcj).
tV THE EFFECTIVE PARTICLE SIZE IN
   ELECTROSTATIC FIELD

 Electrically charged particles may be driven
 through a fluid by applying an electrostatic
 field.   The driving force depends on the
 charge of the particle.  If the particle does
 not have the maximum charge it is capable
 of carrying, the velocity is dependent only
 on the drag diameter (Djj).  The maximum
 charge that a particle can carry is propor-
 tional to its surface  area.  Therefore, the
 velocity of  a fully charged particle in  an
 electrostatic field depends on the size [
It should be noted that the effective size in
electrostatic precipitation is not necessarily
Dp as det-ermined by Stokes' Law (equation 1).
REFERENCE

1  Hawksley,  P.O.  The Physics of Particle
      Size  Measurement:  Part I.  Fluid
      Dynamics and the Stokes'  Diameter.
      The British Coal Utilization Research
      Assoc. Monthly Bulletin.
      NoT 4. April, 1951.
Vol XV.

-------
                                               9
SECTION 9
Representation of Particle-Size Data
Statistical Presentation of Data
Size-Efficiency Curves

-------
                      REPRESENTATION OF PARTICLE-SIZE DATA
    METHODS FOR EXPRESSING THE
    RESULTS OF PARTICLE-SIZE
    MEASUREMENTS

    Distribution of particle-size in a mixture
    of particulate material is commonly rep-
    resented by plotting either, or both, of
    the  following curves:

    1  A frequency distribution curve

    2  A cumulative distribution curve
II  THE FREQUENCY DISTRIBUTION CURVE

 A  Examples of Tabulation of Data

    1  The tabulations in Tables  1 and 2 are
       simply statements of the amounts of
       material falling within each range of
       size  (size-fraction).

 B  Example of Frequency  Distribution Curves

    1  When gas-borne particles  are produced
       in industrial operations, there is a
       tendency to form a "preferential" parti-
       cle-size.   As a result, particle-size
       distributions tend to approximate a
       probability relationship  with a peak at
       the preferential size.  This fact is
       demonstrated in Figure  1.
Table 1
22
Size
range
(v)
0 5
5 - 10
10 - 20
20 40
40 80
80 160
160 - 320
+ 320
% by wt.
in size
range
2. 3
2. 7
9. 1
13. 3
21. 1
34. 5
15. 6
1. 4
  Table 2
              Particle Si'/.e

                 Figure  1
Size
range
(M-)
0 - 2
2 4
4-6
6 8
8 10
10 - 12
12 14
14 - 16
16 18
18 - 20
20 22
22 24
24 26
26 28
28 30
30 32
32 34
34 36
36 - 38
38 - 40
+ 40
% by wt.
in size
range
3. 0
7. 5
8.5
8. 0
7. 0
6. 0
5. 2
4. 8
4. 0
3.8
3. 2
3. 0
2. 5
2. 3
2. 2
2. 0
2. 0
1. 5
1.5
1.0
21. 0

J>
/0.S
/
t%,X
7/-°
/^O
/J,0
7 -t,J>
/ff,r
/?
/"^C1
 PA. C. pm. 62a5. 61

-------
Representation of Particle-Size Data
         Curve (1) shows a "normal" proba-
         bility distribution in which the dis-
         tribution is symmetrical about the
         preferential size.

         This form of curve is rarely en-
         countered for dusts formed by com-
         minution.  However, it may be found
         for particles,  such as  fumes formed
         by vapor phase reaction and condensa-
         tion, or for  tar and acid  mists.

         Curve (2) resembles the  normal
         probability except that it is skewed
         (off center).

         This is the type of curve usually
         obtained for comminuted dusts.
         Tables 1 and 2  would plot skewed
         similar to Curve (2).

         Curve (3) demonstrates that some
         materials  may  show more  than one
         preferential size.
C Construction of the Frequency Distribution
   Curve

   1  Type of paper

      a  Usually,  frequency distribution
         curves are plotted on regular co-
         ordinate (linear) paper.

   2  The ordinate and abscissa

      a  The percentage by weight  (or the
         frequency)  is plotted as the ordinate.

      b  The average particle-size of each
         size-range  (size-fraction) is plotted
         as the abscissa.

         Enrollee:  Plot  a frequency distribu-
         tion curve on Figure 2 using data
         given in Table 2.

   3  Significance of the size-range (size-
      fraction)

      a  Need for a  single  system of selecting
         size-ranges.
         1)  Since the curve is plotted using
            average particle-size within each
            size-range,  for a given dust there
            may be several positions  of the
            curve,  each position depending
            on the series of size-ranges used.
            Thus,  it is essential that  a single
            system of selecting the extent of
            each size-fraction be adhered to
            in order to make  accurate inter-
            pretations from the plotted results.

         Systems for selecting size-ranges
         (size-fractions)

         1)  Select equal arithmetic incre-
            ments of size (See Table 2)

         2)  Choose size-ranges bounded by
            sizes having the same ratio to
            each other (See Table  1).  (This
            is equivalent to taking equal in-
            crements of the logarithm of the
            particle-sizes,  a  method  used  in
            the log-probability type of fre-
            quency distribution).

         3)  Divide the frequency (plotted as
            the ordinate) by the micron-range
            of the increment involved, pro-
            viding a quotient of "frequency
            per micron.  " This method has
            the advantage that no strict sysr
            tern of selecting size-ranges need
            be followed.
D  Prediction of a Size-Distribution of a
   Particulate Sample

   It is evident from the above that a large
   number of points are necessary to fix the
   position of the frequency distribution curve.
   Hence, if this were the only method of
   representing the whole of a particulate
   sample, particle-size  analysis would be
   an extremely laborious and time-consuming
   procedure.

   The following demonstrates methods of
   predicting the distribution of particles in
   a particulate sample by splitting the
   sample into a few size-fractions.

-------
                                                                                           ID
                                                                                           13
                                                                                           i-t
                                                                                           (D
                                                                                           CD
                                                                                           o
                                                                                           0

                                                                                           o
                                                                                           Hi
Figure  2     Frequency  distribution curve for  data in Table  2
                                                                                           CO
                                                                                           H-
                                                                                           N
                                                                                           fD
                                                                                           tu
                                                                                           rt
                                                                                           CD

-------
  Representation of Particle-Size Data
III  THE SEMI-LOGARITHMIC CUMULATIVE
    DISTRIBUTION CURVE
B  Construction of a Semi-Logarithmic
   Cumulative Distribution Curve
 A Example of Tabulation of Data

                    Table 3
Size- % by wt. in
Range size-range
(1)
0 5
5 - 10
(2)
2. 3
2. 7
% by wt. smaller
than largest size
in range
(3)
2. 3
5.0
10 20
20 - 40
40 - 80
80 160
IfiO - 320
+ 320
9. 1
13. 3
21. 1
34. 5
14. 1
27. 4
48. 5
83. 0
15. 6 98. 6
1.4 ! 100.0
   1  Type of paper

      a  Semi-logarithmic graph paper is
         used.
                                                       2  The ordinate and the abscissa

                                                          a  The particle-size is plotted on the
                                                             logarithmic abscissa

                                                          b  The percent by weight (or  frequency)
                                                             less than the size indicated on the
                                                             abscissa is plotted on the linear
                                                             ordinate.

                                                          c  Example
       The tabulation of column (3) of Table  3
       is a statement of the percent of the
       total weight of the sample attributable
       to all particles less than an indicated
       size.
                     in
                     c
                     jc   5
                     ^ OJ ^
                     en £ U
                     en cti co
                     cu cj .
                     ^-i -rH IH
                       T3 rt
                                         Particle-size
                                       (logarithmic scale)
                                           Figure 3

-------
                                                       Representation of Particle-Size Data
      1)  A portion of such a plot will be a
         straight line as between A and B
         of Figure 3.  However,  this
         straight line portion  includes only
         about 70% of the total sample; 15%
         of the total sample lies beyond
         point B,  and the other 15% before
         point A.

         Hence, it is erroneous to assume
         that  a cumulative distribution
         curve plotted on semi-log paper
         is a  straight line over the entire-
         range of particulate  sizes in the
         sample.

         Enrollee: Plot a semi-logarithmic
         cumulative distribution curve on
         Figure 4 using data given in Table
         2.

3  Significance of the size-range

   The selection of size-ranges is not
   important

Prediction of Size Distribution of a Particu-
late  Sample

As is evident from the above,  the cumula-
tive  distribution curve  is not  a straight
line  for the entire range of particle-size
in a  sample.  Hence, a somewhat large
number of size-fractions must still be
used to determine the proper  position of
the  curve, especially the position of the
extremes.
IV  THE LOG-PROBABILITY CUMULATIVE
    DISTRIBUTION CURVE

  A Example of Tabulation of Data

    The data is tabulated as  indicated by
    Table  3.

  B Construction of a Log-Probability Cumula-
    tive Distribution Curve

    1  Type of paper

       a  A special coordinate paper is  used,
          known as log-probability paper.
          One scale is logarithmic, the other
          is a special probability type.

    2  The ordinate and the abscissa

       a  The particle-size  is plotted on the
          logarithmic ordinate

       b  The percent by  weight (or frequency)
          less than the  size  indicated on the
          logarithmic ordinate is plotted on the
          probability  scale as the abscissa.

    3  Example

       a  The entire  plot  becomes  a  straight
          line if the frequency distribution
          plot is skewed as in Curve (2) of
          Figure 1.

          Enrollee:  Plot  the data of  Table 2
          on Figure 5.
                 _0)
                 "d
                              % by weight less than si/.e indicated

                                       (probability scale)

-------
                         IJ-L
100






 90






 80






 70






 60






 50






 40






 30






 20






 10






  0
                iTTT
                                                        ! I I
J--1.

1
             1.5
                                       10
15
20
30      40    50   60
                        Figure  4     Semi-log frequency distribution  for  data in Table 2

-------
                                                          Representation of Particle-Size Data
V   THE ARITHMETIC   PROBABILITY
    CUMULATIVE DISTRIBUTION CURVE

 A  Example of Tabulation of Data.

    The data is tabulated as indicated by
    Table 3
    Construction of an Arithmetic-Probability
    Cumulative Distribution Curve

    1   Type of paper

       a  A special coordinate paper is used,
          known as arithmetic-probability
          paper.  One scale is linear, the
          other is a special probability type.

    2   The ordinate and abscissa

       a  The particle-size is plotted on  the
          linear ordinate.

       b  The percent by weight (or frequency)
          less  than the size indicated on the
          linear scale is plotted on the proba-
          bility scale as the abscissa.

    3   Example

       a  The entire plot becomes a straight
          line if the frequency distribution is
          "normal" as in Curve (1) of Figure 1.
VI  USEFULNESS OF PROBABILITY
    CUMULATIVE DISTRIBUTION CURVES

 A A Relatively Few Size-Fractions Need Be
    Analyzed

    1  If a few observations of cumulative
       percent undersize give points which
       fall on a straight line when plotted on
       either arithmetic   or logarithmic-
       probability paper,  one  is reasonably
       justified in taking the straight line for
       the cumulative distribution curve.

    2  For  some approximations it may be
       satisfactory to know the amount of
       particulate less than only two specific
       sizes in order to find two points for the
       plotting of a straight line.

    3  In some cases,  it may  be known that a
       plotted distribution of a particular dust
       always has the same slope.   Hence, for
       approximations,  only one point may be
       necessary to provide an indication of
       distribution over the entire particle-
       size range.
                     0)
                 CD  'ol
                               % by weight  less than si?f

                                       (probability scale)

-------
                                                                                                                                                            V
99.99       99 9  99 B
                          Figure D  -  Log-Probability Distribution for Data in Table 2
                        99    99     JS     90       80    70   60    50   40    30    20      10     5       21
                                                                                                             0.5    O.J  0.1  t.G5    0.01
                                                                                                                   99.8 M.9"       99.99

-------
B  Frequency distribution tabulation (as shown
   in Table 1), and curves (as shown in Figure
   1) may be plotted from data provided by the
   probability cumulative distribution plot.

   1  The frequency distribution curve  can be
      constructed by taking the change in
      cumulative percentage for each small
      increment of size.
      For example:  If 24% of the total
      weight of sample is due to size less
      than 1. 0 micron and .30%  due  to size
      less than 3. 0 microns, then 6% of the
      total weight of the sample is due to
      sizes between 1. 0 and 3. 0 microns.
      Since the frequency curve plots on
      average size of an increment, the  	
      plotting points become 6% and | 1 + 3
REFERENCES

1   Encyclopedia of Chemical Technology,
      Volume 12.   The Interscience  Encyclo-
      pedia,  Inc   New York.  1954.
2  Lapple, C. E.   Fluid and Particle Mechanics.
      U.  of Delaware.  Newark,  Delaware.
      1951.

3  The Problem Characteristics, Industrial
      Dust.   The Thermix Corporation.
      Greenwich,  Conn.

4  Particle Size Analysis.  Bulletin No. G402R.
      Western Precipitation Corporation.
      Los Angeles, Cal.

5  Lapple, C. E.   Representing Distribution
      of Particle Size.  Heating, Piping, Air
      Conditioning 18: 108.  February,  1946.

6  Hawksley, P. G.  The  Physics of Particle
      Size Measurement-   Part I. Fluid
      Dynamics and Stokes' Diameter.  The
      British Coal Utilization Research Assoc.
      Monthly Bulletin, Volume  XV, No. 4.
      April,  1951.

-------
                 STATISTICAL PRESENTATION OF DATA
Arithmetic Probability Cumulative Frequency Distribution
98 95 90 $p 70 40 SO 0 30 20 10 5 1















r
A






/

2 5 10 >
plotting position





/







X






/







/






/







/






/






/
f






/A






0304030607080 90 93 9
1
N + 1

                If data are normally distributed, as N 	*• ™  median = mean
                                 84.13
                                                50
PA
  .S. 17.3.67

-------
Statistical Presentation of Data
                         Log Probability Cumulative Frequency Distribution
98     95
                       9®
§0    TO
SO   40   30    20
10
9
G
7
6
5
4
3
2
8
7
6
5
4
3
2
1


































A















y
yX^
1^"














>^
r '



I 5 TO 20 3
plotting position =











>
s














S
J>















/

















/
















>
X















>
^














x
>^















^A'
^













0 40 90 60 70 80 90 95 91
1
N + 1

                 If data are log-normally distributed, as N-
                 geometric mean
                                                                 median
                                        so

-------
                             SIZE-EFFICIENCY  CURVES
 I   EFFICIENCY GUARANTEES ON
    COLLECTION EQUIPMENT
 A  Efficiency guarantees,  not outlet concentra-
    tion guarantees, are made.

    Particle collection equipment usually
    operates upon a particulate-gas mixture
    that is difficult to  specify exactly,  is
    likely to be highly variable,  and over which
    there is little positive control.  Therefore,
    the manufacturer of the collector will
    guarantee only  efficiency,  since efficiency
    will vary less than other performance
    indices with changes in the quantity or
    characteristics of the feed.

    On the other  hand,  the customer may be
    primarily interested in the concentration
    of particulate matter in the exit gases,
    and hence,  desires that this  be  the subject
    of the guarantee.  Loadings in exit gases
    cannot be guaranteed unless  the inlet
    loadings are  specified.  But,  specifying
    inlet and outlet loadings is the same as
    making an efficiency guarantee.
 B  Efficiencies are expressed in terms of:

    1  Overall efficiency,  or

    2  Size-efficiency


II   SI/.K-EFFICIENCY CURVES
 A  ATI or requirements for a valid test have
    heen met, and particle-sizing methods
    agreed upon, efficiency tests are run.

 M  Steps Leading to the Construction of a
    Si/,e -Efficiency Curve

    1  The overall efficiency by weight is
      calculated.
    2  Size-distribution curves for inlet and
       outlet, inlet and catch,  or outlet and
       catch, are constructed  on log-probability
       paper.

    3  From the size-distribution plots,
       percent by weight in selected  size-
       fractions are determined, and the
       efficiency of collection  for each size-
       fraction calculated.

    4  The size-efficiency curve is drawn by
       plotting the mid-point of each size-
       fraction (as the absissa) against the
       appropriate efficiency (as the ordinate)
       on regular graph paper.

 C Size-efficiency curves should specify:

    1  Method of size-analysis
    2  Carrier-gas (assumed air unless
       specified)

    3  Temperature and pressure of the
       carrier-gas

    4  Grain loading
    5  Nature of the particulates
    6  True  density of the particulates
    7  Flocculation characteristics of the
       particulates

    8  Type  and dimensions of the collector

III  GENERAL CLASSIFICATION OF
    EFFICIENCY
       50-80

       80-95

       95-99

       99-99.9

       above 99. 9
Classification

Low efficiency

Medium efficiency

High efficiency

Very high efficiency
Ultra-high efficiency
 PA. C. pm. 06. 9. 60

-------
                                          10
SECTION 10
Settling Chambers

-------
                                         SETTLING CHAMBERS*
         I  DESCRIPTION

            A gravity separator consists of a chamber
            in which the velocity of the carrier gas is
            made to drop suddenly so that the particles
            in the gas stream will settle by gravity.
            Reduction in velocity is accomplished by'
            expanding the ducting into a chamber of
            suitable dimensions to obtain the desired
            low velocity.
        II  TYPES OF SETTLING CHAMBERS

            A  Simple Settling Chamber

               The simple settling chamber consists
               of a single compartment (See Figure 1).
                           dull collection hoppers
         Figure 1.   HORIZONTAL FLOW SETTLING CHAMBER
              Howard Settling  Chamber

              In this separator, horizontal  plates
              parallel  to  the  line  of  flow  of  the
              carrier gas  stream are used to obtain
              f\ number  of  comnartments  in which  the
              particles mav settle  (See Figure 2).
              Shelves between  compartments  may have
              vortical   dimensions as small  as
              one-inch.
      Figure  2.   HOWARD SETTLING CHAMBER



III  DESIGN PARAMETERS

     A  Theory

        Consider  Figure 3 below
      Figure 1.  SIMPLE SETTLING CHAMBER
*Manual outline as revised by R. T.  Shigehara,
 January, 1.970

 I'A. <:.,.",. 2 Jr. I-/'>

-------
Settling Chambers
     If a particle enters the chamber at a
     vertical distance hp above the lower
     level of the chamber height Hc, it must
     fall this distance hp before it tra-
     verses the horizontal distance Lc if the
     particle is to be collected by the cham-
     ber.

     Assuming that the particle velocity
     Vx is the same as the velocity of the
     gas stream, and that we have a uniform
     velocity profile, the residence time 6
     for all particles within the chamber is:
                                    (EQ 1)
     A  particle  of  size  D   will  settle some
     distance  hf)  in time 9.   If  hg  >  Hc,  these
     particles wil]  be collected with 100%
     efficiency.  If hg  <  Hc,  whether a given
     particle  will  be collected  or  not depends
     upon  Its  position h_.   For  example,  if
     h0 =  0.5  H(., all particles  of  size Dp
     wlio.se position  hp ^_ 0.5H  will  be collected
     and nil p.'irticles where hp> 0.5HC will not
     be collected.   Therefore, the  fractional
     t'f Tir-icncy Is 0.50.  This Is true assuming
     Ihnl all  Llie particles  of size Dp are
     uniformly distributed over  Ilc.
     Willi  this i.n mind,  we can write:
                                    (I'Q .;)
              Theoretical Sije
             Efficiency Equation:   (EQ  4);
                Settling Chamber
 where   E  = fractional efficiency of
             particles of size Dp
        Vy  = particle settling velocity
        Vx  = gas  velocity through the
             chamber
        LC  = chamber length
        H  = maximum distance particle must
             settle in order to be collected.

 Stoke's law offers a reasonable approxi-
 mation  of  Vy for particles of concern
 (see Table 1).   A generally accepted
 rule of thumb: Stoke's law applies when
 Stoke's  law  says
     ;y(s)
g Dp2 (Pp - p)

      18L1
(EQ 5)
where  Vy(s) =  settling  velocity in Stoke's
                law  range
           g =  acceleration due to gravity
          Dp =  particle  size
          pp    particle  density
           p =  gas  density
           V =  gas  viscosity
    where Ep is the fractional efficiency  for
    ,\ eiven particle size Dp.  Hc can now  be
    expanded to identify the maximum distance
    that the particles must settle  in order  to
    I'i? collected.

    Tlivj distance hg that a given size particle
    will settle is dependent upon the settling
    voloci tv Vv
                  Thus:
Review the assumptions that  concur  with
Stoke's law in the manual  outline titled
''Particle Settling Dynamics",  (Section 4
of the Combustion Evaluation Manual).

The horizontal velocity Vx can be rewritten
in terms of the volumetric effluent flow
rate and chamber dimensions:
                                     (KQ  3)
                                                                           »c Bc
                                 (EQ 6)
    I I  follin-.'.s Lliat  (substituting  liQl  and  EQ3
    i ill o l'f'2 ) :
where H   is  the  chamber height and B
is the chamber width.                C

-------
                                                                               Settling Chambers
                     Table 1.  SETTLING VELOCITIES OF SPHERICAL PARTICLES OF
                                     UNIT DENSITY IN AIR (D

                         Temperature: 20°C(68°F): Pressure 760mm Hg.
Particle diameter Experimental
microns
cm/sec
o-i
0-2
0-4
1-0
2
4
10
20
40
100
400
1000
8-7
2-3
6-8
3-5
1-19
5-0
3-06
1-2
4-8
24-6
157
382
x 10-
x 10-"
x 10-"
x 10-3
x 10-2
x 10-2
x 10-1





Calculated
from Stokes' law
cm/sec
8-71 x
2-27 x
6-85 x
3-49 x
1-19 x
5-00 x
3-06 x
1'2
5
25
483
3050
io-5
10-*
10-"
io-3
IO-2
io-2
IO-1





    The maximum settling distance HC of
the
    Howard settling chamber is the chamber
    height Hc divided by  (N + 1) number of
    trays.

B   Design Considerations

    A  Gas Velocity

       The  velocity of the gas  stream should
       be kept as low as possible because
       the  settling rate of dust decreases
       when the turbulence of  the gas is in-
       creased.
       But,  although it is in the streamline
       region of flow where the most effect-
                   ive settling is obtained, it is not
                   practical to reduce the gas stream
                   velocity to the streamline region
                   because of the large and impractical
                   cross-sectional area that would be
                   required.

                   For practical purposes, and so that
                   the velocity will not be so great as
                   to re-entrain the settled particles,
                   the generally accepted rule of thumb
                   is:  that velocities below 10 ft/sec
                   are satisfactory for most materials.^)

                   Table 2 lists some pickup velocities
                   of various materials.
                     Table 2.   PICK UP VELOCITIES OF VARIOUS  MATERIALS
                                                                      (3)
Material
Aluminium chips
Asbestos
Non-ferrous foundry dust
Lead oxide
Limestone
Starch
Steel shot
Wood chips
Wood sawdust
Density
g/cm3
2-72
2-20
3-02
8-26
2-78
1-27
6-85
1-18
-
Median
size
microns
335
261
117
14-7
71
64
96
1370
1400
Pick up
Velocity
ft/sec
14-2
17-0
18-8
25-0
21-0
5-8
15-2
13-0
22-3

-------
Settling Chambers
     B   Efficiency  Equation

        The  actual  magnitudes  of  gravitation-
        al settling velocities Vy(s)  as
        determined  by Stokes'  law are not
        used in settling chamber  design  be-
        cause of inherent factors that are
        impossible  to predict  by  a theoreti-
        cal  mathematical expression.
        For  example, all particles do not
        necessarily have undisturbed  free-
        fall; agglomeration during settling
        may  change  the original particle size;
        some particles may be  re-entrained.
        EQ A is only theoretical  and  does not
        take into account eddy currents  and
        uneven distribution of air flowing
        through the chamber.


     Because of the above factors, EQ 4  in-
     cluding EQ 5 and EQ 6 should be  written
               gpp 2(pp-p)
   where:   K  =»  empirical  factor;  a  factor
                 of  0.5  frequently  used when
                 no  other information is
                 available
           g  =  gravitational  acceleration
           p  =  Particle density
           PP  =  gas density
           y  =  gas viscosity
           Q  =  volumetric flow rate
           B  =  chamber width
           L  =  chamber length
           N  -  number  of  parallel chambers,
                 1 for a simple chamber and
                 N trays +  1  for Howard
                 settling chamber.

IV  APPLICATION

    Despite the fact that settling  chambers
    are simple in design and  can be manu-
    factured from almost any  material, they
    are infrequently used because of  the
    extremely  large  space requirements and
    the relatively low efficiency.  Where
    settling chambers are used, they  are
    normally followed by a  more efficient
    collecting device.

    Efforts have been made  to improve the
    efficiency of gravity settling  chambers
    by  the  use of baffles and various other
                                                        methods.   The  Howard  settling  chamber
                                                        is  an  example  of  these  efforts.   But
                                                        this type  of settling chamber  was never
                                                        widely used because of  the  difficulty in
                                                        removing the settled  dust from the
                                                        horizontal trays.

                                                        A   Metal Refining

                                                            The combination settling chamber  and
                                                            cooling device has been  widely used
                                                            in  the  metal refining industry to
                                                            partially collect  large  particulates
                                                            and to  reduce  the  gas temperature to
                                                            the final collecting device.   One of
                                                            the more common types, the  "hair  pin
                                                            cooler", is shown  in Figure 4.(1)
                                                              Figure 4.  Hair-pin cooler
                                                         B  Arsenic Trioxide

                                                            Arsenic trioxide from smelting arsen-
                                                            ical copper ores has always been
                                                            collected in brick settling chambers
                                                            known as "kitchens".

                                                         C  Foodstuffs

                                                            In the manufacture of various food-
                                                            stuff, simple settlement is the
                                                            first step in dust recovery, achieved
                                                            by spraying the condensed liquids
                                                            into large chambers.

                                                            The effluent air is then passed to
                                                            second stage cleaners (cyclone) and
                                                            the exhaust re-circulated to the
                                                            spray chambers.

                                                         D  Boilers

                                                            Power and heating plants may employ
                                                            settling•chambers.

-------
                                                                                Settling  Chambers
REFERENCES:
 1.   W.  Strauss,  "Industrial Gas Cleaning,"           4.  Joglekar, G. D. and Subramanian, N. R.
       Pergamon Press, Oxford 1966, pg.                    A Single Vane Cyclone Separator. J.
     144-159                                               Sci. Industr. Res. 14A.  1955.

 2.   C;  E.  Lapple in J. H. Perry, Ed. Chemical        5.  Drinker, P. and Hatch, T. Industrial
       Engineer's Handbook, 3rd Edition,                   Dust. McGraw-Hill Book Co., N. Y.
       p.  1021, McGraw-Hill, N. Y. (1950).                 1954.

 "1.   J.. Balif, L. Greenburg, and A. C. Stern,         6.  Lapple, C. E.  Fluid and Particle
       Am.  Ind. Hyg. Association Quarterly                 Mechanics, U. of Delaware, Newark,
       9,  85 (1948) as referenced  in  (1)                   Del. 1956.

-------
                                             11
SECTION n




Cyclones

-------
                                             CYCLONES

                                           D.  James  Grove*
 I  INTRODUCTION

 The cyclone collector is an inertial separa-
 tor without moving parts where a confined
 vortex is formed which produces sufficient
 centrifugal force to drive the suspended part-
 iculate to the collector wall. Although it is
 simple, inexpensive, and can be constructed
 from many materials, there are few applications
 where collection efficiencies exceed 80-90%.
 Cyclones with a body diameter less than__nine__
 inches have been arj^££rJly deslgnated_as
 ''^ej}__g££j:cig£QY!!-_"cycIgnesTj'As~^Il"l)'e shown
•"Ta t eT, smaTTerbody diameTers create larger
 separation forces, and consequently (up to
 some practical limit) provide higher efficienc-
 ies.

 II  MECHANISM OF PARTICLE COLLECTION

 The basic components of a cyclone are shown
 in Figure 1. These include a cyliner, a tang-

               Cleaned-Gas Exit
    Dust Laden
    Gas Inlet
                                  Cylinder
                                  Cone
                                Dust Hopper
                 Collected Dust

      Figure  1.   BASIC  CYCLONE COMPONENTS

 ential  gas  inlet,  a cone to deliver the
 collected  dust to  a central disposal point,
 a  dust  hopper, and an axial gas outlet, part
 of which extends into the cylinder. The con-
 fined vortex of a  cyclone is illustrated in
 Figure  2.  The gas  enters tangentially into
 the  annular  space  between the cyclone body
and the outlet tube, and spirals downward
in what is called the main vortex. Near the
bottom of the cone,  the spiralling gases be-
gin to move upward in the vortex core.  The
spiralling action of the gases causes the
particles to be driven to the walls by  cent-
rifugal force, where they are moved towards
the dust discharge by the force of gravity
and the downward movement of the main vortex.
It should be noted that the spiral motion of
both vortices is in the same direction. The
tangential velocity (how fast the gases are
swirling) is lowest near the wall and at the
center of the cyclone. It reaches a maximum
at a point approximately 60-70% of the  way
in from the wall to the center.
                            Eddy
                           Main Vortex
                           Vortex Core
                                                       Figure 2.
             VORTEX  AND  EDDY  FLOWS  IN A
             TYPICAL CYCLONE  DUST COLLECTOR
             DESIGN.
 Not  only  are  there variations  in  the  tang-
 ential  velocity  at different points in  the
 cyclone,  there are also  vertical  eddies  and
 what is called inward  drift. The  inward  drift
 is a radial gas  flow which moves  toward  the
 center  of the cyclone, opposing the movement
 of particles. While vertical eddies can  exist
 In the  cone,  the most  troublesome are those
 present in the annular region  near the gas
 inlet.  The eddies, which  are caused by  the
 vortices  and not by the  gas outlet extension,
 can  carry particles directly from the gas
 inlet to  the gas outlet.

 All  of  these variations  combined  together make
 the  problem of determining the separation forces,
 and  consequently the efficiency,  much more
 difficult.
 *Chemical Engineer, Institute for Air Pollution Training

  PA.C.pm.104.4.73

-------
 III  DETERMINATION OF CRITICAL PARTICLE SIZE
      AND CUT SIZE

 There are two sizes which are commonly used
 to relate to the efficiency of a cyclone. The
 equations given for both of these are empirical
 relationships, and their derivation will not
 be presented here. They should not be used for
 original calculations, but rather for comparing
 the efficiencies of similar cyclones operating
 at different conditions.
                                         , £h,e
 size of the smallest               _^^^
'removed completely "(removed with JLOO% efficiency)^
"from a 3TisT 'TaHen gas stfreamT
 [D ]   _
   pjcr-
V9 p (D - d0)

27T N v± (p
                       - P)
                                                           = diameter of gas outlet, ft.

                                                           = number of revolutions the
                                                             gas stream makes (5-10 is
                                                             typical)

                                                           = inlet gas velocity, ft/sec

                                                           = density of the particulate,
                                                             lbs/ft3

                                                           = density of gas, lbs/ft

                                                           = viscosity of the gas, Ibs/
                                                             ft-sec
 where:   [D ]     = critical particle size
           p cr             v
          D      = diameter of cyclone body,  ft.
A size-efficiency curve is a plot of  the  re-
lationship between different sized particles
and the efficiencies with which they  are
removed in a certain cyclone. An example  of
such a curve is shown in Figure 3. If we
determine [D ]   from one of these curves, it

would be the diameter which corresponds to
             100
             80
             60
             20
                        20        40        60        80        100

                                     Particle Size - Microns

                                  Figure 3.  SIZE EFFICIENCY  CURVE
                                                            120       140

-------
100% on the efficiency scale. For  the curve
shown in Figure 3,  the [D  ]^ would be approx-
imately 90 microns.

Since it is rather  difficult to determine
[Vcr accurately fr°"i the graph,  the cut size
is often determined instead. The cut size is
defined as the size of the particle which is
removed with 50% efficiency.
      [D  ]
       p  cut =
               p^T
              * 2ir M -u
                 2ir  N v  (p  -  p)
                       1    p
                                       (2)
     where:
              [D  ]      =  cut  size
               p  cut
              w        =  width  of  the  gas  in-
                         let  ,  ft.

Referring back to  Figure 3,  the  [D  ]    would
be 12 microns.                    p cut

IV  PRESSURE DROP  DETERMINATION

The pressure drop  across a cyclone  collector
will generally range between one  and  seven
inches of water and it is usually determined
empirically. An equation does  exist which can
be used for relating the pressure drops for a
cyclone operating  at several different condi-
tions, or for geometrically  similar cyclones.
           0.0027  Q
Ap =
                                          O)
         k d
                       D  /    \  D

where:  Ap = pressure drop, Inches  of water

        Q  = volumetric flow  rate at the  Inlet,
             cu. ft./sec.

        L  = height of the gas inlet, ft.

        L    = height of the  cylinder,  ft.

        L     = heicht of the cone, ft.
         cone

        k  = dimensionless factor descriptive
             of the cyclone inlet vanes

           = 0.5 without vanes

           =  1.0 for vanes that do not expand
              the entering gas or touch the
              gas outlet wall ("a"  in Figure 4)

           <= 2.0 for vanes that expand  the
             entering gas and touch the gas
             outlet wall ("b" in Figure 4)
                                                     V  CONFIGURATIONS AND RELATIVE DIMENSIONS-
                                                        Effects of Cyclone Performance


                                                     The dimensions of  a  cyclone  are of primary
                                                     importance when  considering  efficiency  and
                                                     pressure  drop. A widely  recommended  cyclone
                                                     with  an ordinary tangential  inlet would have
                                                     the following proportions, based on  the cyclone
                                                     body  diameter D:

                                                                     Table  1.
Cylinder length (L   ) = 2D

Cone Length (L    ) = 2D
              cone

Outlet extension (into cyclone)  length (L )
= 0.675 D                                °

Inlet height (L )  = 0.5 D

Inlet width (w ) = 0.25 D

Outlet diameter (d )    0.5 D
                  o
Most design changes intended to  increase
collection efficiency also increase pressure
drop. Conversely,  recovering energy from the
outlet vortex is the  only method of reducing
the pressure drop  which does not also reduce
efficiency.

Increasing the inlet  velocity will increase
the efficiency, although the relationship is
very complex.  There is also an upper limit,
about 100 feet/sec, above which there is in-
creased turbulence which in turn causes re-
entrainment of the separated dust and reduced
efficiencies.  The pressure drop will Increase
as a function of the  inlet velocity squared,
or twice the velocity yields four times the
pressure drop.

The length of the  cyclone body determines the
residence time during which the  particles are
subject to the separating forces, and Increas-
ing this length will  increase efficiency. Also
dust which has been entrained in the vortex
core will have more time to become reseparated.
Increasing the body diameter to  outlet diameter
ratio will also increase efficiency, although
the optimum ratio  is  between 2 and 3.

As can be deduced  from the definition of "high
efficiency" cyclones, decreasing the body
diameter will  increase the efficiency. This  is
due to increased separation forces caused by
the smaller vortex radius.  In a  very small
cyclone, however,  the dimensional clearances
are so small that  plugging occurs easily. Small
cyclones also  experience bouncing of larger

-------
particles and local turbulence which reduce
efficiencies, and diameters of of 2 to 3 inches
seem to be a practicaT~mSffiu^"™"™°'*~-~-=~™~~~~™=

The cone portion of the cyclone is not neces-
sary to convert the downward vortex to an up-
ward one, although its presence does reduce
the length of cyclone needed to effect this
reversal. The main purpose of the cone is to
deliver the  collected particles to a central
point  for easy handling and disposal. The
cone cannot  be too small in diameter at the
bottom, or the vortex core will contact the
walls  and re-entrain the collected dust.

Dust discharge design is just as important in
reducing this re-entrainment, due to the high
turbulence and velocities present near the
discharge. The static pressure in the vortex
core may be  slightly negative, and this will
tend to draw the collected dust up away from
the discharge. The best solution is some type
of mechanical device, such as a rotary valve,
a double-flap valve, a screw conveyer, or a
dip leg. To  be successful, the mechanism must
achieve continuous, complete, and immediate
removal of the separated dust and prevent in-
flow of gas  from the hopper.

As the inlet air enters the annular region at
the top of the cyclone, it is squeezed by the
existing gas to about half its inlet width.
This causes  a significant pressure loss, which
can be reduced by adding vanes to  the  annular
area (see Figure 4). The presence  of the  vanes,
            Figure 4.  INLET VANES

however, reduces the efficiency, apparently
due to the prevention of vortex formation in
the annulus. Helical or involute inlets  (Figure
5) are attempts to reduce interference between
the incoming gas and the vortex already  present
in the annulus. Axial inlets are free from
most of these problems, but they introduce new
problems. The inlet vanes must be designed so
that they impart adequate rotation to the gas,
and yet resist erosion and plugging.
                                     Figure  5.   CYCLONE INLETS

-------
The eddy current in the annular region requires
that the gas outlet extend into the cyclone to
prevent excessive amounts of dust from passing
directly from the inlet to the outlet. Usually
this extension ends just below the bottom of
the inlet. Devices which permit the gases to
leave the gas outlet tube tangentially have
been successful in reducing the pressure loss
without sacrificing the efficiency.

Since the pressure drop in a cyclone is caused
by the vortex and not by wall friction, rough
walls actually reduce the pressure drop due to
the suppression of the vortex formation. They
also greatly reduce the collection efficiency,
due to increased turbulence and re-entrainment.

 VI  CARRIER GAS AND PARTICULATE  CHARACTERISTICS
     Effects on Cyclone Performance
 Changesin particle  size,  density ,,_and  con-_
 cenfTratTqK do not .have a' significant effect
 on the pressure  drop  across  a  cyclone.  As  far
 as the efficiency is  concerned, however, this
 is not the case. Efficiency  will increase  with
 increases in partTcTe density, mean particle
 size, and concentration. The larger a particle
 is, and the more it weighs,  the better  the
 separation forces. The effect  of increased
 dust loading (concentration) Is not quite  so
 obvious, but it  is in part caused by the small
 particles getting swept  to the walls by the
 many large particles.

 From equation 2  or 3, we can see directly  the
 effect of the carrier gas  properties on the
 efficiency, in terms of  the  cut size or critical
 particle size. A lower value for  [D ]   or

 [D ]    means a higher efficiency, and  this
   p cut
 would result from a lower  viscosity or  a lower
 gas density. Temperature and pressure will
 affect the density and viscosity, although
 this is not generally significant. A change
 in pressure would produce  only a slight change
 in density, and a change in  temperature would
 Increase one and decrease  the  other, with  a
 small net effect.
Although there may be a slightly higher effi-
ciency using several cyclones in series, the
additional pressure drop Is usually sufficient
to make series operation disadvantageous.
Occasionally a large diameter cyclone is used
as a precleaner for small multiple cyclones
to remove the larger particles that cause plug-
ging in the multiple units.

VIII  FURTHER MATHEMATICAL CONSIDERATIONS

In order to obtain a prediction of overall
cyclone efficiency, a size efficiency curve
for a given cyclone under a given set of op-
erating conditions, and a particle size dis-
tribution are needed. The size efficiency curve
is characteristic of the cyclone, treating
the particular powder used in the test, when
running at the throughput of the test. Cor-
rections for differences in particle grading,
particle densities, gas viscosities, and gas
flow rates can be made using the equations
already presented, although care must be taken
in making large transpositions.

If the calculations are based on the cut size,
equation 2 will be the starting point:
  [Vcut
              2rr N v± (p  - p)
                 (2)
In order to determine the effect of the var-
ious parameters on the cut size (or the ef-
ficiency) ,  a ratio is taken of the test cyc-
lone and conditions to the new cyclone and
conditions
                     9 y wj
    p cut (test) T2TT N v  (p  - p)
                                    test  (A)
  [Dp]cut (new)   	—
                  2'fr N v,  (p  - p)
                        i    p       new
 VII  TYPF.S AND ARRANGEMENTS OF CYCLONES

 individual high efficiency, small diameter
 cyclones have a small capacity, and  they must
 ho operated in parallel to handle typical gas
 volunirs. They generally have a common gas in-
 let, dust hopper, and gas outlet, and can be
 arranged in banks of several hundred cyclones
 each. A whole new set of design problems arises
 witli this arrangement, but it is advantageous
 in that we can get a higher efficiency com-
 pared to one large cyclone, with about the
 same pressure drop.
If, for example, the effect of changing vis-
cosity is to be determined, holding all the
other parameters (w ,  N, v ,  p ,  p) constant,
the ratio would be:
    p cut (test)
    p cut (new)
IV  (test)

 y  (new)
                                          (5)

-------
which can also be written:

  [D ]    ,   , = [D ]
    p cut (new)     p cut
                          (test.
(new)

(test)
                                          (6)
By the same reasoning, the result of a change
in gas inlet velocity would be:
  [D ]
    P cut (new)
                   p cut  (test)
                                   (test)
                               [v. (new)
                                           (7)
Repeating this procedure, the effect of
changing N, p , and p can also be determined.
Lapple (5) states that the above approach is
applicable only for geometrically similar
cyclones. Based on the dimensions given in
Table 1, Lapple presented a method for deriv-
ing the size efficiency curve from the cut
size alone. Once the cut size [D ]    has
                                p cut
been determined, the efficiency for any size
particle D  can be arrived at by calculating
D /[D ]   , and then using the graph (Figure
6) to determine the efficiency.  For example,
if the cut size is 20 microns and we want the
efficiency for a 10 micron particle, D /[D ]
         3                            P   p cut
is 0.5, and the efficiency is about 22%. A
complete size efficiency curve can be drawn
by getting the efficiency at several different
particle sizes.
Collection Efficiency, %
3 o SD fe o c









/








/








/








/









^








f









'







./
s







Jf^








^l^~~








— '









.1 in









y







































                                                         0.3 0.4 0.5
                                                                Particle Size Ratio, (D /D  )
                                                                                       P  PC

                                                        ^Figure 6.   CYCLONE EFFICIENCY VERSUS--
                                                         ""\       PARTICLE SIZE RATIO
                                                             ""-•--^  (LAPPLE, 1951).
                                                    According  to  the Air Pollution Engineering
                                                    Manual  (6), experimental data has(compared
                                                    favorable  with Lapple's correlation, except
                                                    for  slightly  lower efficiencies than those
                                                    calculated for D /[D ]     ratios of 2-3.
                                                                     p   p cut
                                                    Apparently, Lapple's correlation nuy be suf-
                                                    ficiently  accurate for  an engineering estima-
                                                    tion of many  cyclone applications.

                                                    Gallaer  (3) determined  that if the  size ef-
                                                    ficiency curve of a cyclone is plotted on
                                                    semi-log paper, as in Figure 7, a straight
TT.T
99
95
90
80
70
60
50
40
20









/
/
/
-/- 	






/
/
/








&

/








V

/









/











/












/






































°0 5 10 15 20 25 30 35 40 4!
Figure 7. SIZE EFFICIENCY CURVE

-------
line results. The equation that represents
this line is:
     (d)
                                    where:
                  overall efficiency
       where:
                J(d)
= fractional efficiency
  of a particle with
  diameter d

= particle diameter,
  microns

= a constant for the
  particular cyclone in
  question
The particle size distribution of a typical
dust, when plotted on semi-log paper, also
produces a straight line (Figure 8).  The
It should be noted that this formula is use-
ful even for those dusts whose particle size
distribution does not plot as a straight line
on semi-log paper, as long as the value of 6
approximates the slope of the line for those
values of d which significantly affect the
cyclone efficiency,  i.e. the smaller particles.

In order to use equation 10, the particle
size distribution curve and size efficiency
curve must be drawn as straight line approxi-
mations, and convenient points are picked off
the straight lines to determine K and S.  It
is then a simple matter to determine the over-
all efficiency.

An extension of this approach yields perhaps
an even easier method for the calculation of
the overall efficiency.
3 8 6 4 2
                                        Percent  Greater  than d
                              Figure 8.  PARTICLE SIZE DISTRIBUTION
equation of  this  line  is:

    Z         -
      (d)
     where:
              '(d)
the fraction, by weight,
of the total dust having
a larger particle size
than d
Since the cut size, like the value of ^, is
a function of the particular cyclone being
used, a: could be defined as:


                                       (ID
                                                                   ]
                                                                  p cut
                6  =  a constant  for  the part-
                     icular distribution in
                     question

Given these equations, it has been shown that
the overall efficiency can be represented by:
                                           where:
                                                       a constant
                               The mean particle size, like the value of 6,
                               is a function of the particle size distribu-
                               tion being used, so 6 could be defined as:
             + 6
                                                                                            (12)
                                                                   p mean

-------
             where:   y = a constant

Inserting equations  11 and 12 into equations
8, 9, and 10:

   T-I      -I       V.^/ L" J   / ^           /I o N
                                                    Inlet Gas
                                                    Velocity

                                                    Pressure Loss
                                                          20 to 70 ft/sec (usually
                                                          about 50 ft/sec)

                                                          0.5 - 2.0 inches of water
                                                          for simple cyclones, 2 to 7
                                                          inches for high efficiency
                                                          units
   7
   Z.  , = e
                p mean
                                      (14)
                                                    Particle Size
                   ]
                 p cut
                                      (15)
With  E.
        3 ]                  [D ]
         p cut                p mean
°'5> d = [Dp]cut'  and "ith Z(d)
      J(d)
  0.5, d =  [D  ]     . Inserting these values
      '       p mean          &
into  equations 13 and 14, it is found that
x = y. Equation 15  then reduces simply to:
            [D
             p mean
                                       (16)
         [D
          p cut
           ]
          p mean
Therefore, if the mean particle size is known,
and the cut size Is known, the overall effi-
ciency can be very quickly calculated.

Not only can the efficiency by calculated
originally from this equation, but the effect
on the efficiency of changing either [D ]

or [D ]mean can be quickly calculated. Before
using this method, however, it is imperative
to determine if the distribution is truly
represented by an essentially straight line.
IX  SUMMARY
Gas Flow
      Table  2.

      30  to  50,000  cfm  (some  to
      100,000  cfm)  smaller units
      must be  arranged  in parallel
      to  accommodate large volumes
                                                          1 to 200 microns at vary-
                                                          ing efficiencies
                                                    High Efficiency   20 to 40 microns  for  simple
                                                    on Normal Indust- cyclones, 10  to  30 microns
                                                    rial Dusts with   for high efficiency units
                                                    Mean Particle
                                                    Size of
                                                    Particle
                                                    composition

                                                    Particle
                                                    concentration
                                                          solid and liquid
                                                           down to .1 grains/ft.,
                                                           although usually above 10
                                                           gralns/ft3, with no real
                                                           upper limit
                                        REFERENCES
                                                     1.   Stairmand,  C.  J.,  The  Design and Perfor-
                                                         mance of  Cyclone  Separators, Trans.  Inst.
                                                         Chem. Engrs.,  Vol.  29,  British,  1951.

                                                     2.   Caplan, K..J.,  All  About Cyclone  Collectors,
                                                         Air Engineering, pages  28-38, September
                                                         1964.

                                                     3.   Gallaer,  C.A.,  and J.W. Schindeler,
                                                         Mechanical  Dust Collectors,  J_. A.  P_. £. A..,
                                                         Vol. 13,  pages  574-580, December 1963.

                                                     4.   Kane, J.M., Operation,  Application,  and
                                                         Effectiveness  of  Dust  Collection Equipment,
                                                         Heating and Ventilating,  August  1952.

                                                     5.   Lapple, C.E.,  Processes Use  Many Collec-
                                                         tion Types, Chemical Engineering,  Vol.  58,
                                                         pages 145-151,  May 1951.

                                                     6.   Air Pollution  Engineering Manual,  999-AP-40
                                                         pages 91-99, 1967.
Gas Temperature   to 750°F

-------
SECTION 12
Miscellaneous Dry Inertia!-Type
  Collectors

-------
                   MISCELLANEOUS DRY INERTIAL-TYPE COLLECTORS
   LOUVER - TYPE DUST SEPARATOR
   (Figure 1)
          LOUVER TYPE COLLECTOR

                Figure 1

A  Mechanism of Particle Removal

   1  The louver-type dust separator contains
      a series of blades set at an angle  to
      the air stream.

      a  A large portion of the air (90% of the
         total) stream passes through the
         louvers.

      b  A smaller portion (10% of the total)
         of the air stream (blowdown) con-
         tinues in its original direction  with-
         out passing through the louvers.

   2  The air which passes through the  louvers
      is forced to turn sharply in a rapid re-
      versal  of air flow.
      a  Particles contained in the air stream
         impinge upon the blades and rebound
         into that portion of the  stream (blow-
         down) which did not pass through
         the louvers.

         1) Hence dust particles are con-
           centrated in  the relatively small
           volume of blowdown.  Usually a
           secondary collection system  is
           needed to deal  with the blowdown.


B  Efficiency

   1  Efficiency is a function of louver
      spacing; closer spacing provides
      higher efficiencies.

   2  Particles as low as 10-20u. may be  re-
      tained in excellent designs.

   3  When used as  a pre-cleaner, particles
      of  50-100|i are usually removed.


C  Advantages and Disadvantages

   1  Advantages

      a  Simplicity

      b  Low cost of construction

      c  Low pressure drop  for degree of
         removal obtained

      d  Temperature and  pressure limita-
         tions are imposed only by materials
         of construction.


   2  Disadvantages

      a  Plugging due to buildup of particles
         on the blades

      b  Abrasion difficulties

      c  Inability to handle tacky materials
PA. C. pm. 67. 9. GO

-------
  Miscellaneous Dry Inertial-Type Collectors
 II  SCROLL TYPE DUST SEPARATOR
    (Figure  2)
  Dust gas to
  secondary
  Collector
                         Partly-Cleaned Gas
              Figure 2
  A  Figure 2 shows the scroll type collector.

  B  The scroll collector is simply a dust col-
     lecting fan which separates particles very
     much like the louver-type dust separator.
     Secondary collectors are needed to sepa-
     rate the particles in the "skimmed off" air
     stream. The large-volume gas stream,
     stripped of large particles, may then pro-
     vide a lighter load for more efficient
     cleaning equipment.

  C  A  grade efficiency curve is shown in
     Figure  3.
Ill  REVERSE NOZZLE IMPINGEMENT
    COLLECTORS (Figure 4)

 A Figure 4 shows a typical commercial
    no/zle impingement collector.

 B High efficiency is attained on ducts
    larger than 10-20(1.

 C These collectors are designed for pressure
    drop in the range of 0. 1  1.5 inches of
    water.
o
£
a
u
t—1
fa
fc
W
                                                u
                                                H
                                                J
                                                o
                                                CJ
    100
                                                     80
                                                     60
                                                     40
     20
             20      40      60

                  PARTICLE SIZE, \i

                     Figure 3
                                                                                      80
                                                                                              100
   D  Chief advantage lies in their greater
      adaptability to existing flues or ducts
      than other types of collectors.

   E  They may be used at elevated temperatures.
      (If the dust is tacky, circulating water
      films may be used to keep the elements
      clean).

 IV   BAFFLE CHAMBERS (Figure 5)
   A  Baffle chambers employ fixed baffle plates
      which cause the air stream to change di-
      rection, thereby projecting the particles
      into a dead air space where they settle by
      gravity.
   B  Efficiency
      1   Particles greater than 50ji are removed
         efficiency.
   C  Application
      1  Baffle chambers are used as precleaners
        for more efficient collectors to reduce
        the load of large diameter particles on
        these units.

-------
                           Miscellaneous Dry Inertial-Type Collectors
Otogrdmotlc Pion vtew Stowing Gos Mo«
-------
Miscellaneous Dry Inertial-Type Collectors
                 Figure 5



V   DRY TYPE PYNAMIC PRECIPITATORS

 A  Principle of Operation (Figure 6)

    1  Dry type dynamic precipitators are
      motor-powered separators in which
      the dust is precipitated by dynamic
      force produced by the action of numer-
      ous specially shaped fan blades.

      a  The precipitated dust is forced along
         the blade surfaces and  discharged
         into a dust storage hopper.


B  Advantages and Limitations

   1  Units  are compact and space require-
      ments are small.

   2  Pressure drop is only about l/z inch of
      water.  However, pressure  drop varies
      and  is a function of mechanical efficiency.

   3  It functions both as dust collector and
      fan.

   4  It cannot handle fibrous, sticky
      materials.
                                                                 DRY TYPE
                                                         DYNAMIC PRECIPITATOR
                  Figure 6
    5  Volumes up to 20, 000 cfm may be
       handled by some designs.
  C Operating Conditions
Gas flow	
Gas temperature
Draft loss . .  .  .
Draft loss sensitivity
to cfm change ....
High efficiency of
removal for ordinary
industrial dusts with
mass median size
greater than	
Efficiency sensitivity
to cfm change ....
Particle composition ,
Humid air influence.
up to 20, 000 cfm
to 750°F
a function of mechanical
efficiency.  Usually
about YI in. w. g.

a function of mechanical
efficiency.
10-20(1

negligible
cannot handle fibrous,
sticky materials
may cause condensa-
tion and plugging

-------
                                                  Miscellaneous Dry Inertial-Type Collectors
REFERENCES                                    2  Joglekar, G. D.  and Subramanian.  A
                                                        Single Vane Cyclone Separator.
1   Perry, J. II.  Chemical Engineer's Hand-              Division of Industrial Physics, Na-
      book.  McGraw-Hill Book Co.  New                tional Physical Laboratory of India,
      York.   1950.                                       New Delhi, India.

-------
SECTION 13





Wet Collectors:  Introduction

-------
                      WET  COLLECTORS:  INTRODUCTION
I  Wet collectors increase particle removal
efficiency by two mechanisms.
 A Re-entrainment of the collected particles
   is prevented by trapping them in a liquid
   film or stream and then washing the liquid
   (and trapped particles) away.

 B Fine particles are "conditioned" so that
   their effective size is increased, thus
   enabling them to be collected more
   efficiently.
      addition of wetting agents does not
      significantly increase removal
      efficiency.

   3  Effect of solubility of particles

      Solubility of ihe particles in the
      droplets  is not a factor in effectiveness.
      (An exception is the case of concentrated
      mist droplets, such as sulfuric acid.
      These droplets may grow in size by
      absorption of moisture when passing
      through a humid chamber).
II   PARTICLE CONDITIONING

 Particle conditioning in wet collectors
 involves the process of increasing the
 effective size of the fine particles so that
 they may be more readily precipitated.  The
 effective size may be increased by:

 Forcing precipitation of fine particles on
 liquid droplets,  or

 Promoting condensation upon fine particles
 (which act as nuclei) when the water vapor in
 a  gas  passes through its dewpoint.
 A  Conditioning by Forcing Precipitation of
    Particles on Liquid Droplets

    1  An example:

      An example is the attachment of a
      5-micron dust particle to a liquid
      droplet 50-microns in diameter  thereby
      increasing its apparent mass 1000 fold
      for collection purposes.

    2  Effect of wetting agents in resisting
      redispersion

      Collision  of solid particles with  liquid
      droplets is inelastic and because of
      Van der Waal's forces, the agglomerates
      resist redispersion.  Therefore, the
                                                 III
   Conditioning by Promoting Condensation
   upon the Particle Surface

   If the liquid spray causes the gas to pass
   through its dewpoint,  condensation will
   take place upon the surface of the particles
   when the particles act as nuclei.  Thus,
   the  effective size of the particles is
   increased under such conditions.  This
   mechanism is important for initially hot
   gases containing relatively small dust
   concentrations (say less than 1-grain/cf).
   OPERATING PROBLEMS OF WET
   COLLECTORS
A  Corrosion
      All water scrubbers have the inherent
      problem of corrosion.

      a Even when no chemically corrosive
        constituent may be contained in the
        carrier gas stream,  the carbon
        dioxide present contributes to
        corrosion.

      b When corrosive agents are contained
        in the gas stream (SO2, chlorides,
        fluorides, nitric acid,  etc.),
        will occur on wet metallic surfaces.
PA.C.pm. 75. 9. 60

-------
Wet Collectors:  Introduction
B  Krosion
      Wot collectors that remove insoluble,
      abrasive materials have troubles due
      to erosion especially if removal is
      dependent upon impingement velocities
      or centrifugal action.
C  Wet-Dry
      Scrubbers are faced with problems at
      wet-dry junctions, particularly at the
      entrance of an installation.

      a  When dust concentrations and gas
         temperature are high, there may be a
         zone where dust build-up can occur
         (by reason of moist dust layers).
D  Mist Elimination

   1   In all scrubbers, entrainment eliminators
      are important to prevent carry-over of
      droplets.

   2   Many scrubbers have mist eliminators
     built into their design.
   3  When not incorporated in the design,
      mist elimination is accomplished by
      means  of additional separators.
E  Slurry Handling

   1  For all scrubbers,  a method must
      be provided for handling the liquid
      effluent.  Slurries may be treated by
      means of:

      a   Settling tanks

      b   Filters

      c   Liquid cyclones

      d   Further chemical or recovery
         methods

      e   Disposal to sumps,  streams, rivers

      f  Others
  2  All these effluent handling methods have
     their own unique engineering problems.

-------
SECTION 14
Collection of Particles on Cylindrical
  and Spherical Obstacles

-------
               COLLECTION OF PARTICLES  ON  CYLINDRICAL
                          AND  SPHERICAL OBSTACLES
 Particulates transported by a carrier gas
 through a depth of cylindrical (fibers) or
 spherical(granules) obstacles tend to be
 precipitated upon the surface of the obstacles.
 Van der Waal's and electrical forces cause
 the particulates to adhere to the surfaces of
 the obstacles resulting in the removal of the
 particulates from the gas stream.
 I   MECHANISMS OF PARTICULATE
    REMOVAL

 A  Screening (or sieving) is not the principal
    mechanism

    It can be shown that the sizes of the gas
    passages through the depth of obstacles
    are very much larger than the sizes of the
    particulates collected.

 B  The principal mechanisms by which par-
    ticulates are brought into contact with
    the obstacles include:

    1   Interception

    2   Gravitation

    'i   Impingement

    4   Diffusion

    5   Electrostatic

    fi   Thermal
II   INTERCEPTION

 Particulates being carried by a flow of gas
 tend to follow the streamlines around an
 obstacle.  By chance, a particle on one of
 the streamlines may make contact with the
 obstacle if the streamline passes the obstacle
 at a distance less than the radius of the
 particle.   This type of removal is called
 direct interception, and depends solely on
 the position that a particle has in the gas
 stream.
Ill  GRAVITATION

 As a particle passes by an obstacle, it may
 fall (under the  influence of gravitational
 force) from the streamline along which it is
 being carried and settle upon the surface of
 the obstacle.
IV  ELECTROSTATIC

  Since a force of attraction exists between
  bodies possessing electrostatic charges of
  opposite polarity, it is possible for a
  charged particle to be removed from the
  gas stream by an oppositely charged obstacle.
  However, when only the particle or obstacle
  is charged, a charge may be induced upon
  the uncharged component resulting in a
  polarization force that can also effect
  particle removal.

  The effect of the electrostatic  mechanism
  of particle removal from a gas stream may
  be significant when" the charge on the particle
  or obstacle is high,  and when gas velocity
  is low.   The significance of  particle size and
  obstacle size varies,  depending on whether
  the electrical attraction originates from
  Coulomb or polarization forces.

  The mechanism involved in a bed of fibers
  or granules depends principally on the charac-
  teristics of the particulates  and obstacles in
  the bed, and on the gas velocity.
 V  IMPINGEMENT TARGET EFFICIENCY

 A The meaning of impingement target
    efficiency

    When an obstacle is placed in the path of
    a particulate-laden gas stream (Figure  1)
    the streamlines will diverge and pass
    around the obstacle.  The particles,  how
    ever, tend to leave the streamlines (along
PA. C. pm. OOa. 5. 61

-------
Collection of Particles on Cylindrical and Spherical Obstacles
   which they are being carried) at the be-
   ginning of the curvature and may impinge
   upon the obstacle.

   If like  particles, initially within a cross-
   section of the carrier-gas stream having
   a radius of D_^ (measured from the central
              2~
   streamline) strike a cylindrical obstacle
   of diameter Do, then D1 is termed the
   "impingement target diameter" of the
   obstacle for the particular particles being
   considered.
   The ratiom' | is called the "impingement
   target efficiency" and is symbolized r)j.
   In other words,  it is the ratio of the cross-
   sectional area of the gas stream cleaned
   of particles (all of which are alike)  to the
   projected area of the obstacle.

   If it is assumed that all particles are alike
   and equally dispersed throughout the gas
   stream, the "impingement target efficiency"
   is the ratio of the weight of particulate
   collected by the obstacle to the weight of
   particulate that would pass on if the ob-
   stacle were not there. Therefore,  "im-
   pingement target efficiency" is the  effi-
   ciency of removal by weight of like
   particles by one obstacle.

B  The Mathematical Expression

   Impingement target efficiency (r}j) is a
   function of the dimensionless ratio,
                  D0g
   where:
                vp/ofP(s)
           impingement target efficiency for
           uniformly dispersed like particles
           and for one obstacle.
   D    a  diameter of the obstacle

   VD/O -  relative velocity of the particle
           (in the approaching gas stream)
           to the obstacle

           Stokes1 settling velocity
C  Impingement Target Efficiency Curves'
   (Figure 2)

   Figure 2 demonstrates the relationship
   between impingement targe efficiency
   (r)j) and the dimensionless ratio
                vP/ofp(s)
   Note that there are two curves; one for
   spheres and one for cylinders.  The im-
   pingement target efficiency for spheres is
   higher than that for cylinders because  the
   streamlines diverge more sharply around
   spheres.
                   cross-section of air stream
                   cleaned of particles
                   L.
                   D'
                   r
                         Impingement on a spherical obstacle
                                       Figure 1

-------
        Vo   TP(S)
Figure  2    (reference 2)

-------
  Collection of Particles on Cylindrical and Spherical Obstacles
100 E
80 —
80 J4
50 =-
40 ^
* 3oLU
I »-
•s
•5 10 .L
? »i
-Q i^
e 5 ,,-1
* 1
ST]"
2^
ir~"
.01
^=Z=
	 +t
4^
fW
T^
T4^
w~n
f?^
^
•£::£
:::±±
1 1 1
.(
*
t-
X*
a*,
^
i
):
-H-
±i
*
-H"
•xT
j_!_
;=z
»JT
>
-ff
2
r1
/^_
i.
TTT
rH-
^S
*^H
=^=*<
^
v<
JTf-
. ' ''
'• 1 '
.(
F
•^
i
^
yf-
S
±:
-N
S
i
i
)3
H-f
•X
g|
' I
-f
/
•H-l
^
^
^
>t<
4+
if
TTTT
•4-rf-
VH-
^5
#
TtTT
^~y
P
H
=Sr
i
TTT
^E!
7
j —
•W-(-r
TT
* 1
1>T
•S
-So
^
= 4
^4
nrr?
~^~ r~
NT
i
^=
^
r 	 	
4- * —
4^
ri —
J. J 	 ;_
Ei±
i J 	
•"•-^
^fe
^^
F3^
!±f
- ^ •*—
'^^
. ; S
0.1
; — r
t— r
|=M-
EJ,
i 	 77
"V-j-
^4
^
f-r
(-H-
-H-
-T
«iS
J-^l
^,_
t
*^
~^T
^
1
*j
^
1>
***r
pt-r
34,
^
T"
2)2
q
^
•p
J
-1-
^
^
_L
^
0.
Tf*-
^BT
•^
^R
^^
^•'"V
T-J^
./'
^s;

-------
                             Collection of Particles on Cylindrical and Spherical Obstacles
100
  .01
.02  .03
0.1     0.2  0.3

 Gas Velocity (Ft/Sec)
i.o      2.0 ;s.o
         Target Efficiency at Various Air Velocities for Different Diameter Dust
                                Particles  (Ref. 2)
                                                               -4
          (p  =2.0)  (Fibre diameter 10^)  (Viscosity  =1.8(10)    poises)
            P
10.
     =  a constant depending on the temper-
        ature,  viscosity, and particle
        diameter (cm2/sec)
     K =  1.45(10)
                   -17
   where:

   T      absolute tempc raturc (°K)

   ^      absolute viscosity of the gas (poise)

   D    = particle diameter (cm)
                                                    2  For air below 100°C
                                                                   2.45
                                                     vp/o Dp Do
                                                     where:
                                                                                       (3)
                                          nD     diffusional target efficiency for like
                                                 particles and one obstacle when gas
                                                 stream is air below 100°C (%)

                                          v  I     relative velocity of the particle (in
                                           P     the approaching gas stream) to the
                                                 obstacle  (ft/sec)

-------
  Collection of Particles on Cylindrical and Spherical Obstacles
     D      diameter of the particle (microns)

     Do   = obstacle diameter (microns)
  B  Remarks

     The above equations show that if target
     efficiency due to diffusion (np) is to be
     high,  then the relative velocity (vp/o)
     must be low.  This is opposite to the re-
     quirements  for high impingement target
     efficiency (r^) which demands high  relative
     velocity.  This leads to a velocity zone
     for a given  obstacle where the efficiency
     of removal will be low for a  given particle-
     size; that is, where conditions are poor
     for both impingement and diffusion.

     This is evident by observing the  low por-
     tions of the  curves in Figure 4.
VII  SIZE-EFFICIENCY

  A  The equations for target'efficiencies pro-
     vide information  on the removal of like
     particles by only one obstacle, or, let's
     say, removal of like particles by one
     "treatment" of the gas stream.  Since in
     the actual course of filtration through a
     bed of fibers or granules  the gas stream
     meets a number of obstacles and is there-
     fore "treated" a number of times before
     it exits, an efficiency equation taking all
     "treatments" into account is necessary.
  H The general equation for size-efficiency

            E    1(1- n)s°

     where So is small |_as in spray devices
     (S0 - 5) and old cloth (SQ *  2)j     (4)

            E =  1  -  e'1150
     where SQ is large [_as in packed fiber or
     granular beds and new cloth filters
     (S0  * 50f]                          (5)
   where:

      E  = efficiency of removal of a given
           particle-size (size-efficiency).
           The particle size is identified in

      e  - natural logarithmic base = 2. 718

     S0    number of "treatments" received
           by the gas stream
  _ Total projected area of all obstacles in the filter
0  Cross-section of filter normal to the gas flow

      rj  =  target efficiency of the individual
           obstacles
C  Size-efficiency for a bed of spherical
   granules
                                                          E = 1 - e
      E = 1 - e'"So
(6)


(7)
   where:
      E  =  efficiency of removal of a given
           particle-size by a bed of spherical
           obstacles.   The particle-size is
           identified in 17.

      e  =  natural logarithmic base = 2. 718

      n  -  target efficiency of the individual
           shperical granules in the bed

      L  =  depth of the bed

      «  =  volume of the spherical granules
           per unit volume of bed

     Dp  =  diameter of the spherical granules

      S0  =  number of "treatments" received
           by the gas stream as it passes
           through the  bed
                                                  s  _ Total projected area of all obstacles in the bed
                                                   0 Cross-section of the bed normal to the gas flow

-------
                                    Collection of Particles on Cylindrical and Spherical Obstacles
   D  Size-efficiency for a bed of cylindrical
      fibers
                                          Impingement	  P PP Vpl
              1   e
                          irDr
                           (8)
                                                  Diffusion
                                                                                RT
      where:

         E
                                         (9)
efficiency of removal of a given
particle-size by a bed of cylindrical
obstacles.   The particle size is
identified in rj.

volume of fibers per unit volume
of bed
         e   -  natural logarithmic base  = 2. 718

         r)     target efficiency of the individual
              fibers in the bed

         L    depth of bed

        DO    diameter of the fibers

        So  *  number of  "treatments" received
              by the gas  stream as it passes
              through the bed

    _ Total projected area of all obstacles in the bed
   ° ~ Cross-section of the filter normal to the gas flow
                                                        where:
                                              = diameter of the particle removed
                                               from the gas stream
                                                        Do    diameter of the obstacle
                                                  P     density (mass) of the particle

                                                  g   = local acceleration due to gravity

                                                  \i     viscosity of the gas

                                                  v 
-------
                                            15
SECTION 15





The Gravity Spray Tower

-------
                               THE GRAVITY SPRAY TOWER
I  IMPINGEMENT OF GRAVITATIONAL
   SPRAY DROPS

A  In a gravitational spray unit, there are a
   number of liquid spherical obstacles
   (droplets) falling in an empty tower by
   the action of gravity in the path of rising
   particles.

B  The relationship between_impingement
   target efficiency (nj) and,    ^oS    j  i
                           • Vp/o
   shown in Figure 1.
                                                       It is seen that the maximum efficiency
                                                       for  the smaller particle sizes (say
                                                       less than 5(i) occurs for droplet size
                                                       of about  SOOp.: and that for larger
                                                       particle  sizes, the efficiency varies
                                                       little over the range of droplet sizes
                                                       500 to lOOOp..

                                                       Thus in gravitational spray towers,
                                                       there is  little point in using very
                                                       fine  spray sizes even if such were
                                                       available.
        v i    ,  in gravitational spray towers,
      is the difference in the free-falling
      velocities (Stokes1) of the droolets and
      the: particle.

      In practice, since  the free-falling
      velocity of the particle is sjnall com-
      pared to the droplet, I Vp/o j  may be
      taken as the free-falling velocity of the
      droplet.
                                                 II  EFFICIENCY

                                                 A Inspection of Figure  2 shows that the ef-
                                                    ficiency £f a gravitational spray tower is
                                                    very low for particles below 1-2 microns.

                                                 B Figure 3 illustrates a size-efficiency curve
                                                    for a large industrial spray tower handling
                                                    70, 000 cfm. The tower is 22 ft in diameter,
                                                    66 ft high.  Pressure drop is less than
                                                    1" water.
Krorn Figure  1, it is evident that for high
ctil!i>< tioi: efficiency by impingement
 >;<•:•'• mils', be a small obstacle (DQ)  and
n high relative velocjty  i v / _~]  between

the obstacle and partiele.

I   [n 
-------
                             TI  (spheres) = 0



                             at   o8
                                V  , f  , ,
                                 P/o p(s)
                                          = 24
                             nT (cylinders) = 0
           V
              /    /  s
            p/o p(s)
Figure 1.

-------
    100
  c
  
-------
                                                                          The Gravity Spray Tower
o
o
rc
o
rf
     20 L
\6
                                               50
33
                                      Particle Size, Microns

                              Size-efficiency Curve for Spray Tower
                                            Figure 3

-------
                                                                     . The Gravity Spray Tower
        Hence, the dirty water may be re-
        circulated until it contains quite a high
        concentration of trapped dust particles.

        a  Therefore, there is a saving of
           water, and perhaps a simplification
           of effluent treatment and ultimate
           waste disposal.
 VI  PRESSURE DROP

  A The pressure drop is very small (less
     than 1 in. w. g.)
Draft loss-
Efficiency-
less than 1" w. g.
very low for below
l-2n
Particle concentrations - relatively high (over
                        5 gr/cu ft)
Particle composition	solid, liquid.  Some
                        problems with corrosion.
                                                  Water usage •
                        about 18 gal/1000 cu ft
VII  PERFORMANCE DATA

  Gas flow	over 70, 000 cfm
  Gas temperature	often used as pre-
                          cooler. Gas tem-
                          perature over 2000°F
                          may be reduced to
                          275°F.

  Gas velocity	about 3-5 fps
  Treatment time	about 20-30 seconds
REFERENCES

1  Stairmand, C. J.  Dust Collection by Im-
      pingement and Diffusion.  Paper read
      at the Inaugural Meeting of the Midland
      Branch of A.  Inst. P.  Birmingham,
      England. October 14,  1950.

2  Stairmand, C. J.  The Design and Per-
      formance of  Modern Gas-Cleaning Equip-
      ment.  Paper read before the A. Inst.
      P.  London.   November, 1955.

-------
                                            16
SECTION 16
Venturi Scrubbers

-------
                                VENTURI SCRUBBERS
 I   MECHANISM OF PARTICLE REMOVAL

 A  Since for high collection efficiency of fine
    particles by impingement there must be a
    small obstacle (DQ) and high velocity of
    approach of the gas stream relative to the
    obstacle (vD/o), attempt is  made to ap-
    proach this  ideal by.

    1  An arrangement in which very small
       water droplets (upon which the particles
       impinge) are formed by the gas flow so
       that the droplets are initially at rest
       at the time of impact with the particles.
       Even during the period of acceleration
       of the droplet,  high relative  velocities
       will be maintained since the  particles
       move at the velocity of the gas stream.

       a Such an arrangement  is incorporated
         in the Venturi scrubber.  See
         Figure 1.
II   OPERATION (Figure 1)

 A  Collection of Particles upon the Droplets

    1   In the Venturi scrubber, the particulate-
       laden gas passes through a duct which
       incorporates a Venturi scrubber.
    2   At the throat,  high gas velocities of the
       order of 200-600 fps are attained.

    3   Coarse water  spray is injected into the
       throat by way  of radial jets in quantities
       of 5 to 7 gpm per M cfm of gas.

    4   The high gas velocities at the throat
       immediately atomize the coarse water
       spray to fine droplets (about 50 microns).

    5   Since, at their genesis,  these fine drop-
       lets are initially at rest relative to the
       particles  in the gas stream,  it is at this
       moment that collection efficiency is at
       its maximum.  (vr>fo> is  maximum.
       a  The atomized droplets,  being fine,
          rapidly accelerate to the velocity
          of the carrier gas; but even during
          this short period, relative velocities
          will be high and effective collision
          between droplet and particle will
          take place.

          It is during the period before the
          droplets attain the same velocity
          as the gas  stream that any relative
          velocity  between the droplets and
          particle  is obtained.  For example,
          a 100-micron droplet introduced
          into a gas stream moving at 100
          fps would accelerate to 90% of the
          gas velocity in 16 inches; a  20-
          micron droplet would reach 90%
          of the gas velocity in 2 inches.


 B Removal of the Dirty Droplets

    1  As the gas decelerates after passing
       through the throat, agglomeration of
       the particle-laden droplets takes
       place.

    2  The large agglomerates are readily
       removed by a cyclonic separator.
Ill  EFFICIENCY

 A Effect of Pressure Drop on Efficiency

    1  The higher the pressure drop,  the
       higher the removal efficiency of
       particles.  See Figures 2 and 3.

    2  Pressure drops across  the Venturi
       of 25-30 inches of water gage may be
       expected.

    3  Pressure drop can be increased (and
       hence efficiency can be  increased)
       simply by increasing the gas velocity
       and/or the water injection rate.  See
       Figure 4.
PA. C. pm. 68a. 5. 61

-------
Venturi Scrubbers
           CO«E BUSIER DISC
   TIIGmill C«S l»l!I
                           • HER     WTEI
                           OUTLET    INLET
  CONTABIIUTED   '/
  MS KLET
                    Cyclone scrubber
                                                                                     Orifice scrubber
                                 Venturi  scrubber
                                                     Figure 1

-------
                                                                           Venturi Scrubbers
     100
      95 •
10
                        20
30
40
Venturi Pressure Drop (in. w. g. )

    Curve A: Rotary iron powder kiln

           B: Lime kiln, asphalt plant
           C: Iron cupola

           D: Phosphoric acid plant (acid mist)

           E: Incinerator (sodium oxide fumes)
                 Figure 2
                          (5)
            10     20     30     40    50
            Venturi Pressure Drop (in.w. g. )

                Curve A:  Cupola gases

                       B:  Blast furnace gases
                 Figure
— , ^3
3
c 20
•H
£ 15
o
t-i
7 10
0)
$ s
to
CO
CU




/
/

^
^~



/
f


^
S~
	 '

.
/




S



J




s

.---






S








s


^^





^
*~


^*





s*



























£ 0 2 4 6 8 C
                           Water/Gas Ratio, Gal/1000 cu.  ft.

                           Relation Between Pressure- loss
                           and Water Usage  in Venturi
                           Scrubber

                                   Figure 4
                      4  When gas cleaning requirements change,
                         the only adjustment necessary to the
                         Venturi scrubber, in most cases, is
                         in the flow of scrubbing liquid to in-
                         crease the pressure drop.   Thus higher
                         cleaning efficiency is accomplished
                         without modification or addition.

                   B Effect of Particle Concentration on
                      Efficiency

                      1  If the number  of water droplets is held
                         constant and the number of  particles
                         (concentration) is increased, the number
                         of collisions would be expected  to in-
                         crease.  In other words, collection
                         efficiency should increase as loading
                         increases.

                      2  This increase, however, is due not
                         only to the increased chances of particle
                         collision with  droplets, but also due to
                         collisions between the particles
                         themselves.

                   C Size-Efficiency

                      1  The Venturi scrubber approaches
                         100% for all particles larger than 1. 5
                         to 2 microns.

                      2  Figure 5 shows a size-efficiency curve
                         for a Venturi scrubber(l).   Sizes above
                         2-microns were obtained on special

-------
•u
d

-------
                                                     Venturi Scrubbers
                          TABLE  1
TYPICAL PERFORMANCE DATA FOR VENTURI SCRUBBER
(5)
Source of Gas
IRON i. STEEL INDUSTRY
(jidv liim Cupola
U*VKI'" Slrd L'uiivei Itt
MtcM Upi'ii Hc.1t 111 1 urn-' r it U[n
SIcL'l Opin Ik'Jilli FJMUJCI
(Oxygen Lanced)
Blast Furnace (lion)
Electric Fuinace
Electric Furnace
Rotary Kiln— Iron Reduction
Crushing & Screening
CHEMICAL INDUSTRY
Acid— Humidified SO,
(a) Scrub with Water
(b) Scrub with 40% Acid
Acid Concentrator
Copperas Roasting Kiln
Chlorosulfonic Acid Plant
Dry Ice Plant
Wood Distillation Plant
TiCI, Plant, TiO, Dryer
Spray Dryers
Flash Dryer
Phosphoric Acid Plant
NON-FERROUS METALS INDUSTRY
Blast Furnace (Sec. Lead)
Reverberalory Lead Furnace
Ajax Furnace — Aluminum Alloy
Zinc Sintering
Reverberatory Brass Furnace
MINERAL PRODUCTS INDUSTRY
Lime Kiln
Lime Kiln
Asphalt Stone Dryer
Cement Kiln
PETROLEUM INDUSTRY
Catalytic Reformer
Acid Concentrator
TCC Catalyst Regenerator
FERTILIZER INDUSTRY
Fertilizer Dryer
Superphosphate Den & Mixer
PULP & PAPER INDUSTRY
Lime Kiln
Lime Kiln
Black Liquor Recovery Boiler
MISCELLANEOUS
Pickling Tanks
Boiler Flue Gas
Sodium Disposal Incinerator
A
Contaminants

Iron, Coke. Silica Dusl
lion Ou'U.1
Uon & /tin Ouilp
lion Oxide

lion Die t. Coke Dust
Ferro-Manganese Fume
Feiro-Silicon Dust
•ron, Carbon
Tacomte Iron Ore Dust

H,SO. Mist


H,SO, Mist
H.-SO, Mist
H^O, Mist
Amine Fog
Tar & Acetic Acid
TiOv-HCI Fumes
Detergents. Fume & Odor
Furfural Dust
H.PO, Mist

Lead Compounds
Lead & Tin Compounds
Aluminum Chloride
Zinc & Lead Oxide Dusts
Zinc Oxide Fume

Lime Dust
Soda Fume
Limestone 4 Rock Dust
Cement Oust

Catalyst Dust
H,SO, Mist
Oil Fumes

Ammonium Chloride Fumes
Fluorine Compounds

Lime Dust
Soda Fume
Salt Cake

HCI Fumes
Fly Ash
Sodium (hide Fumes
pproximat
Size Range
(Microns)

1 II!
b ?
OS 1
5-,'

.5-20
1-1
.1-1
.5-50
.5-100


	
	
—
—
	
—
—
.5-1
—
.1-1
—

.1-1
.1-.8
.1-.9
.1-1
.05-.5

1-50
.3-1
1-50
.5-55

.5-50
	
—

.05-1
—

.1-50
.1-2
—

—
.1-3
.3-.1
s Loading
(Grains/ cf)
Inlet Exit


8 10
5-1 5
1-6

3-24
10-12
1-5
3-10
5-25


303'
406'
136*
198'
756'
25'
1080'
1-5
—
1-1.5
192'

2-6
1-2
3-5
1-5
1-8

5-10
.2-5
5-15
1-2

.09
136'
756'

.1-.5
309'

5-10
2-5
4-6

25%
1-2
.5-1

05-15
05-08
03-06
.Ol-.O?

.008- Ob
.04-.03
Average
Removal
Efficiency (%)

yb
98 D
35
99

99
99
.1-.3 92
.1-.3
.005-.01


1.7'
2.8'
3.3'
2.0'
7.8'
2.0-
58.0'
.05-.!
—
.05-.08
3.8'

.05-.15
.12
.02-.05
.05-.!
.1-.5

.05-. 15
.OI-.05
.05-. 15
.05-.!

.005
3.3*
8.0'

.05
5V

.05-. 15
.01-.05
.4-.6

23'
.05-.08
.02
99
99.9


99.4
99.3
97.5
99
98.9
90+
95
95
95
95+
98+

99
91
95
98
95

99+
99
98+
97 +

95+
97.5
98+

85 +
98+

99+
99
90

90+
98
9S
* Milligrams per cubic ft
\«t<-: The efficiencies shown abare are average values for a particiitar

-------
Venturi Scrubbers
       silica  dust powders and those for the
       smaller sizes  on dispersed non-patho-
       genic bacteria.  This size-efficiency
       curve  suggests a very high efficiency
       for a comparatively simple piece of
       equipment.


D  Overall Efficiency


   Table I shows some efficiencies of collec-
   tion experienced by various installations.
IV   ENERGY USAGE


  A  Since pressure drops  of 30 inches water
     gage correspond to 120 kwh per million
     cubic feet of gas cleaned,  efforts have
     been made to reduce the pressure drop.
     However,  if pressure  drop is reduced,
     there is a tendency to  reduce the  efficiency
     also.

  B  Additional high energy usage results from
     the method of injecting water into the
     Venturi throat.  See Figure  6.
             THE TYPE  8-V VK1STTUBI

             The Chemico Type S-F Venturi Scrubber is particularly
             recommended for these hard-to-handle situations: re-
             moval of "sticky" solids from gases; recycling of
             heavy slurries where water supplies are limited; and
             recovery of process materials in concentrated form.

             In the S-F Venturi, scrubbing liquid is  introduced
             through troughs at the top  of the unit.  The liquid
             flows downwardly in a continuous film along the slop-
             ing walls to the deflecting lips, which direct it across
             the throat of the Venturi to be atomized by the force
             of the high velocity gas.
  TUB  TYPE P-A VENTUSU

  The Chemico Type P-A Venturi Scrubber is most ef-
  fective in  the very difficult applications requiring
  efficient removal of sub-micron dust, fume, and mist
  particles.
      Chemical Construction Corp.
                                                                          Figure 6

-------
                                                                         Venturi Scrubbers
V   PERFORMANCE DATA
          Gas flow	
          Gas velocity through throat'
          Pressure loss	
          Gas temperature	
          Overall efficiency	
          High efficiency on dusts with mass median
          size greater than	•	
          Humid air influence on efficiency-
          Water usage	
        -- 200 to over 145, 000 cfm
        --200-600 fps
        --up to 25-30 in. water gage
        — "unlimited"
        	usually high (97   99-<-%i
          -5-7 gpm o.f water per
           M cfm gas
 REKKRENCES

 1   Stairmand, C. J.  The Design and Per-
       formance of Modern Gas-Cleaning
       Equipn."nt.  Paper read before
       A.Inst. P.  London.  November,  1955.

 2   Nicklen,  G. T.  Some Recent Developments
       and Applications of Scrubbers in In-
       dustrial Gas Cleaning.   Proceedings
       APCA, V'nd Annual Meeting APCA.
       Los Angeles.  June,  1959.
3  Jones,  W. P.   Development of the Venturi
      Scrubber.  Ind. Eng. Chem. Nov.  1949.

4  Basse, B.  Gases Cleaned by the Use of
      Scrubbers.  Blast Furnace and Steel
      Plant.  Nov.  1956.

5  Chemico Gas Scrubbers for Industry.
      Bulletin M-104,  Chemical Construction
      Corporation, 525 West 43rd Street,
      N.Y.  36, N.Y.

6  Venturi Scrubbers for Industry,  Bulletin
      M-103A, Chemical Construction Cor-
      poration, 525 West 43rd St. N. Y.  36, N. Y.

-------
                                           17
 SECTION 17
Collectors with Self-Induced Sprays

-------
 COLLECTORS WITH SELF-INDUCED  SPRAYS
I  MECHANISM OF PARTICLE COLLECTION

A  In this equipment, the particle collection
   zone is a spray curtain which is induced
   by the gas flow itself through a specially
   designed orifice.  (The spray curtain is
   followed by a spray eliminator).

B  The Collection of  Particles

   1  Normal gas velocity of about 50 fps
      creates droplets about 320y.

   2  Collection of particles is mostly by
      Impingement on  the droplets during
      the free-falling period of the  droplets
      and also during the period of the accel-
      eration of the  droplets from rest (when
      high relative velocities are available).

IL  APPLICATION

A   Since there is an absence of ledges,
    moving parts,  and restricted passages,
    these units are especially adapted  to
    materials like:
   1  Magnesium and explosive  dusts

   2  Sticky or linty materials  like metallic
      buffing exhausts

III  PERFORMANCE DATA

Efficiency - See figure 2 (page  3)

Water usage - 10-40 gal/1000 cfm gas cleaned
              (Much or all of  this water may
              be recirculated).

Sensitivity - not particularly sensitive
              to cfm change  (at  least within
              + 25% of the design rate).

Concentration - high concentrations  (40 grains/
                ft3) (There  are  no fine clear-
                ances to  cause chokage)

Maintenance - The whole apparatus is well
              irrigated and  periodic hosing-
              down of the interior is easily
              done. There is an  absence of
              moving parts.  There may be cor-
              rosion difficulties.
 PA.C.pm.70.9.60

-------
Collectors With Self-Induced Sprays
                                             Figure  1

-------
CD
c
     n
     o
        100
          40
         20
                                           12
16
20
24
28
32
36
40
                                                                                                                                     o
                                                                                                                                     o
                                                                                                                                     :r


                                                                                                                                     CO

                                                                                                                                     ID
                                                                                                                                     3
                                                                                                                                     Q.

                                                                                                                                     C
                                                       Particle Size,  Microns

                                       Size-Efficiency Curve for Self-Induced Spray Collector

                                                             Figure 3
                                                                                                                                     CO

                                                                                                                                     ID

-------
Collectors with Self-Induced Sprays
                                            REFERENCES
   First, M., et. al., "Performance Character-
      istics of Wet Collectors," NYO - 1587
      Waste Disposal, Harvard University. 1953

   Stairmand, C.J., "Mist Collection by Im-
      pingement and Diffusion", paper read
      at the Inaugural Meeting of Midland
      Branch of A. Inst. P., Birmingham,
      England.  Oct. 14, 1950.

   Stairmand, C.J. "The Design and Performance
      of Modern Gas-Cleaning Equipment," paper
      read before the A. Inst. P., London.
      November, 1955.
Kane, J.M. "Operation, Application, and
   Effectiveness of Dust Collection Equip-
   ment," Heating and Ventilating. Aug.
   1952.

Nicklen, G.T. "Some Recent Developments
   and Applications of Scrubbers in In-
   dustrial Gas Cleaning," Proceedings,
   APCA, 52nd Annual Meeting, Los Angeles.
   June, 1959.

Magill, P.L. Air Pollution Handbook, Mc-
   Graw-Hill Book Co., Inc. 1956.

-------
SECTION 18




Impingement Type Scrubbing Tower

-------
                  IMPINGEMENT  TYPE SCRUBBING TOWER
 I  TYPES OF SCRUBBING TOWERS

 A There are two types of scrubbing towers
   commonly used:

   1  Those employing impingement target
      plates

   2  Those employing beds of spherical
      obstacles
 B Particle Concentration

    1  An important feature of this design is
       freedom from chokage  in spite of the
       small holes in the orifice plates.   This
       is due to:

       a  The very  violent circulation induced
          below the  targets by the air jets,
          and
II  TOWER WITH TARGET PLATES (Figure 1)

 A Construction and Operation

   1  This type of scrubber is a tower con-
      sisting of a vertical shell in which are
      mounted a large number of equally
      spaced,  circular, perforated (orifice)
      plates.

      a  At one side  of each orifice plate,  a
         conduit,  called a downspout, is pro-
         vided to  pass the liquid to the plate
         below.
      b  At the opposite side of the orifice
         plate,  a similar conduit feeds liquid
         from the plate above.

   2  Over each hole (about 3/1G" diameter)
      in the orifice plate,  a target plate is
      positioned.

      a  The  motion of the gas past the edge
         of the holes in the orifice plate re-
         sults in the formation of spray drop-
         lets  (about lOOp.).  These droplets
         are initially at rest and provide an
         effective relative velocity between
         particle and droplet for good
         impingement.

      b  The particle-laden gas passes through
         the holes in the plate and the particles
         impinge upon the  atomized droplets
         and on the target  plates.
       b  A preliminary spray zone  which
          helps to keep the orifice plate free
          from deposits.

    2  Concentrations of 40 grains/ft^ can
       readily be handled.


 C Efficiency

    1  An example of a size-efficiency curve
       is shown in Figure  2.

 D Pressure Drop

    1  Each plate imposes  a pressure drop of
       3 in. w. g.
Ill  TOWERS WITH BEDS OF SPHERES

 A  Construction and Operation

    (An example of a scrubbing tower with
    beds of spheres is shown in Figure 3)

    1  Large particles are removed by im-
       pingement on wet surfaces and contact
       with water spray in an area below the
       filter bed.

    2  Particle-laden gas then passes upward
       through a bed  of spheres.  In the inter-
       stices of the bed,  the particles are
       subjected to increased velocities which
       results  in their efficient impingement
       upon the surfaces  of the spheres.
 PA. C. pm. 73. 9. 60

-------
 Impingement Type Scrubbing Tower
Target plate
                                 Water
                                 level
                  Gas  flow
          ARRANGEMENT OF  "TARGET  PLATES"

          IN IMPINGEMENT  SCRUBBER
                                                               Water droplets atomized
                                                               at the edges of orifices
Downspout to
lower stage
                                                           MECHANISM OF  IMPINGEMENT SCRUBBER
                                             IMPINGEMENT
                                             BAFFLE STAGE
                                            AGGLOMERATING
                                              SLOT STAGE
                                                           Peabody Engineering Corp.
                                                                 Figure 1

-------
OJ
O
i-
o>
Q.
O
c:
CD
QJ


C
O
U
O)
   100
    20
                 14         6        8        10        12

                                  Particle size,  microns

                  Size-efficiency  curve for wet-impingement  scrubber

                                          Figure 2
14
                                                                                               CTQ
                                                                                               0>

                                                                                               3
                                                                                               rt>
                                                                                               3


                                                                                               H
                     en
                     o
                     a
                     
-------
   Impingement Type Scrubbing Tower
 Fan
                                  Transition
                                  Place
                                    Main Body
               C  Pressure Loss

                  1  Pressure loss is 4-6 in. w. g.


               D  Efficiency

                  1  Efficiency is high on two micron-sized
                     particles and above.
                                                    E  Water Consumption
                  1  Fresh water:
                     cleaned
                                                                             per 1000 cfm gas
                                       Sludge
                                       Ejector
Overflow
                  2  Recirculated water:  3 gpm per 1000
                     cfm gas cleaned (Scrubbing liquid can
                     have high solids content).


               F  Capacity

                  1  Units handle 500 to 40, 000 cfm.
                 Figure 3

       National Dust Collector Corp.


     3  The high gas velocity through the
        interstices of the packed spheres also
        results in pulling water upward with
        sufficient force to disintegrate the
        water streams into a turbulent mist in
        the zone above the filter bed.  Here,
        ultra- fine particles are trapped by  the
        mist and constantly flushed downward.

     4  Mist carried by the upward flowing
        cleaned gas is removed by passage
        through a bed packed with porcelain
        saddles.


  B  Particle Concentration

     1  Such units have self- cleaning action
        and there is freedom from build-up of
        solids and ease of cleaning.
     2  Concentrations of about 40 grains
        are readily handled.
_, Recirculation
   Pump
Ejector        REFERENCES
Pan
               1  First, M. et al.  Performance Character-
                     istics of Wet Collectors.  NYO-1587
                     Waste Disposal, Harvard University.
                     1953.

               2  Stairmand, C. J.  Mist Collection by Im-
                     pingement and Diffusion.  Paper read
                     at the Inaugural Meeting of Midland
                     Branch of A. Inst. P.  Birmingham,
                     England. October 14,  1950.

               3  Stairmand, C. J.   The Design and Per-
                     formance of Modern Gas-Cleaning
                     Equipment.  Paper read before the A.
                     Inst. P. London. November,  1955.

               4  Kane, J. M. Operation, Application,  and
                     Effectiveness of Dust Collection Equip-
                     ment. Heating and Ventilating. August,
                     1952.
               5  Nicklen, G. T.  Some Recent Development
                     and Applications of Scrubbers in In-
                     dustrial Gas Cleaning.  Proceedings
                     APCA, 52nd Annual Meeting,  Los
                     Angeles. June,  1959.

               6  Magill, P. L.  Air Pollution Handbook.
                     McGraw-Hill Book Co. , Inc.  1956.

-------
                                           19
SECTION  19
Wet Centrifugal  Collectors

-------
                        WET CENTRIFUGAL COLLECTORS
 I  TYPES OF WET CENTRIFUGAL
   COLLECTORS

 A Irrigated Types

   1  These rely upon the throwing of parti-
      Cjles against wetted collected surfaces,
      such as  wetted walls or impingement
      plates by centrifugal action.

 B Spray Chamber Types

   1  These depend upon impaction of the
      particles upon spray droplets and the
      subsequent precipitation of the  "dirty"
      spray droplets upon the wall of the
      unit by centrifugal action.
II   IRRIGATED TYPES (Figure 1)
        WET  CENTRIFUGAL COLLECTORS
   Clean  air outlet      G
   Entrainment separator H
   Water  inlet
   Impingement plates
   Dirty  air inlet       I
   Disintegrator
Inspection  door
Wet cyclone for
collecting  heavy
material
Water and sludge
drain
                Figure 1

         Ami-rir.-in Air Filter Co.
                        A  The efficiency of a centrifugal collector-
                           may be increased by irrigating its  walls,
                           if the attendant disadvantages of a wet
                           system can be tolerated.

                        B  Water distribution may be from low
                           pressure nozzles or gravity flow.

                        C  Performance Data

                        Water rates	3-5 gal/1000 cfm of
                                               gas treated

                        Draft loss	2\ to 6"

                        Draft loss sensitivity
                        to  cfm change	as (cfm) 2

                        High efficiency on
                        particles of mass
                        median greater than. . l-5(i
                        Efficiency sensitivity
                        to cfm change ....

                        Humid air influence
                        on efficiency ....

                        Gas temperature   . .
                        yes

                        none

                        "unlimited"
III  CYCLONE SPRAY CHAMBERS (Figure 2)

 A Operation

    1  The dust-laden gas enters tangentially
       at the bottom and spirals up through
       a spray of high velocity fine water
       droplets.

    2  The dust particles are  collected upon
       the fine spray droplets which are then
       hurled against the chamber wall by
       centrifugal action.

    3  An unsprayed section above the nozzles
       is provided so that the  liquid droplets
       containing the collected particles will
       have time  to reach the  walls of the
       chamber before the  gas stream exits.
       pr,i.

-------
Wet Centrifugal Collectors
   PLEASE-ANTHONY CYCLONIC  SPRAY SCRUBBER

                           cleaned gas
       core buster disc
         spray manifold
          tangential
          gas inlet
 swinging inlet
 damper
     gas  inlet
                                 water inlet
      water outlet
                 Figure 2

       Chemical Construction Corp.


B  Efficiency

   1  Efficiency of dust removal is given by:

                      3n rWH
         E
1 -e
                      2D0Q
      where:

         K  =  efficiency of collection

         r)    individual droplet efficiency

         r    radius of the cyclone (the
              length of the path of the droplet)

         W  =  volume rate of liquid through
              the nozzle

         D-., -  diameter of the droplets
        Q
volume rate of carrier gas
        II =  height of tower (The drops
              should not be made too small
              since entrainment may occur,
              requiring an increase in the
              height of the tower)
   2  Operating Conditions

Gas flow	500-more than 25, 000
                      cfm
Gas velocity into
cyclone	up to 200 fps

Separation factor .  .  . 50 to 300

Efficiency	97+% on dust above l(j.

High efficiency on
particles  of mass
median greater than  . 0. 5  5p-

Efficiency sensitive
to cfm change  .... yes

Draft loss	2-6"w. g.

Draft loss sensitivity
to cfm change  .... as (cfm)2

Water usage	3-10 gal/1000 cu ft of
                      gas cleaned

Humid air influence
on efficiency	none

Gas temperature .  .  .pre-cooling of high
                      temperature gases
                      necessary to prevent
                      rapid evaporation  of
                      fine droplets.

Power requirements  . 1 to 3 HP/1000 cfm of
                      gas
REFERENCES

I  First,  M. , etal.   Performance Char-
      acteristics of Wet Collectors.  NYO-
      1587 Waste Disposal.  Harvard
      University.   1953.

2  Stairmand, C. J.  Mist Collection by
      Impingement and Diffusion.  Paper
      read at the Inaugural Meeting of
      Midland Branch of A. Inst. P.  Birming-
      ham,  England.  October 14, 1950.

3  Stairmand, C. J.  The Design and Per-
      formance of Modern Gas-Cleaning
      Equipment.   Paper read before the A.
      Inst. P-  London.   November, 1955.

-------
                                                               Wet Centrifugal Collectors
Kane. .1. M.  Operation, Application,  and
   Effectiveness of Dust Collection Equip-
   ment.  Heating and Ventilating.
   August, 1952.

Nicklen, G. T.   Some Recent Developments
   and Applications of Scrubbers in
   Industrial Gas Cleaning.  Proceedings
   APCA, 52nd Annual Meeting.   Los
   Angeles.   June,  1959.
Mi gill,  P. L.  Air Pollution Handbook.
   M:Graw-Hill Book Co. ,  Inc.  1956.

-------
                                          20
SECTION 20
Wet Dynamic Precipitator

-------
                           WET DYNAMIC  PRECIPITATOR
  I  OPERATION (Figure 1)

  A  Wet dynamic precipitators combine the
     dynamic forces of a rotating fan to cause
     the particles to impinge upon numerous
     specially shaped blades.

  B  A film of water is maintained on the blades
     by spray nozzles.
                                                  High efficiency on
                                                  particles with mass
                                                  median greater than.
                                                  Efficiency sensitivity
                                                  to cfm change ....
                           .no
                                                  Water usage.
                            0. 5 to 1 gpm/1000
                            cfm gas
di rt and  water
discharged at
blade tips
         clean  air
         outlet
                               dirty air
                               inlet
                        water spray
                        nozzle
                    WET-TYPE DYNAMIC PRECIPITATOR
water and
sludge outlet
                Figure 1
                                                    REFERENCES
                                                  1  First,  M, , et al.  Performance Character-
                                                        istics of Wet Collectors.  NYO-1587
                                                        Waste Disposal.  Harvard University.
                                                        1953.

                                                  2  Stairmand, C. J.  Mist Collection by
                                                        pingement and Diffusion.  Paper read
                                                        at the Inaugural Meeting of Midland
                                                        Branch of A. Inst. P. ,  Birmingham,
                                                        England.  October 14,  1950.


                                                  3  Stairmand, C. J.  The Design and Per-
                                                        formance of Modern  Gas-Cleaning
                                                        Equipment.  Paper read before the A.
                                                        Inst. P.  London.  November,  1955.
  II  PERFORMANCE DATA

  PITSSU rc- drop. .  .  .
                        a function of mechan-
                        ical efficiency
                        Usually less than ]-
                        in.  w. g.
  Pressure drop
  sensitivity to cfm
  change	
            .  .  . '. .  .  . a function of mechan-
                        ical efficiency.

Particle concentration.  .less than 1 grain/ft .
                        (For heavy loading,
                        a pre-cleaner may
                        be used to lighten the
                        load on the unit).
4  Kane, J. M.  Operation,  Application, and
      Effectiveness .of Dust Collection
      Equipment.  Heating and Ventilating.
      August, 1952.


5  Nicklen, G. T.  Some Recent Developments
      and Applications of Scrubbers in In-
      dustrial Gas Cleaning.  Proceedings
      APCA,  52nd Annual Meeting,  Los
      Angeles.  June, 1959.


6  Magill,  P. L.  Air Pollution Handbook.
      M;Graw-Hill Book Co. ,  Inc.  1956.
  PA. C. pm. 71. 9. 60

-------
SECTION 21





Disintegrator Scrubbers

-------
                            DISINTEGRATOR SCRUBBERS
I  MECHANISM OF PARTICLE COLLECTION

A  Since for high collection efficiency there
   must be a small obstacle (D ) and a high
   relative velocity between the obstacle and
   particle (v  , ), attempt is made to ap-
   proach this^laeal by:

   1  Shooting water drops at the particles
      so that a high relative velocity (v  . )

      will be obtained  (even if such velocities
      are maintained for short periods) and
      arranging that this be done so that a
      very large number of impacts will be
      achieved.

B  Such action is incorporated in the disinte-
   grator scrubber (Figure 1).
II  OPERATION

A   A disintegrator scrubber consists  of  an
    outer casing containing alternate  rows  of
    stator and rotor bars,  the  relative velocity
    between adjacent bars  being of  the order
    of 200 - 300 fps.

B   Water is injected axially and is effectively
    atomized into fine droplets (say 25y)  by
    the rapidly rotating vanes.

C   The dust-laden gas also enter axially  and
    passes through the dense spray  zone where
    the particles are subjected to  intense
    bombardment by the water droplets.
                                           Water  inlets
                 Stators
             Dirty  gas  inlet
                                                               Rotors
               Clean gas
               Exit
                                                        Effluent
                                            Figure 1
PA.C.pm.69.9.60

-------
Disintegrator Scrubbers
III.   PERFORMANCE DATA
      Efficiency.
      Pressure drop.

      Energy usage.
      Water consumption.
 highly efficient.  See figure 2
 for a size-efficiency curve.

.less than 1-in.  w.g.

 high power requirements.  Total
 power consumption  may be  16—20
 HP per 1000 cfm  gas cleaned. This
 power is largely expended in atomizing
 and accelerating the  water.

.usually preceded by conventional
 collectors as  cyclones and scrubbers
 to insure that low concentrations of
 the order of % to  % grains per cu ft.
 are presented  to the  unit. These pre-
 cautions are necessary to avoid build-
 up in the disintegrator,  which, run-
 ning at high speed with fine clearance,
 is particularly  susceptible  to trouble
 if operated under  unsuitable conditions.

-------
                                                                         Disintegrator Scrubbers
100
 40
 20
                                      6789     10
                                       Particle Size,  Microns
           11
12   13
14    15
                           Size-Efficiency  Curve  for  Disintegrator  Scrubber

                                               Figure  2
  REFERENCES

  1   First, M.,  et  al.  Performance  Character-
       istics  of Wet  Collectors. NYO-1587
       Waste Disposal,  Harvard  University.  1953.

  2   Stalrmand,  C.J.  Mist  Collection  by  Im-
       pingement and  Diffusion. Paper read
       at  the  Inaugural Meeting of  Midland
       Branch  of A. Inst.  P.  Birmingham,
       England.  October 14, 1950.

  3   Stairmand,  C.J.  The Design and Performance
       of  Modern Gas-Cleaning Equipment. Paper
       read before  the  A.  Inst. P.  London.
       November,  1955.
Kane, J. M. Operation, Application,  and
  Effectiveness of Dust Collection Equip-
  ment. Heating and Ventilating. August,
  1952.

Nicklen, G.T. Some Recent Developments
  and Applications of Scrubbers in In-
  dustrial Gas Cleaning. Proceedings
  APCA, 52nd Annual Meeting. Los Angeles,
  June, 1959.

Magill, P.L. Air Pollution Handbook
  McGraw-Hill Book Co., Inc. 1956.

-------
                                      22
SECTION 22





Fabric Filtration

-------
                                FABRIC FILTRATION
                                       G. W. Walsh*
I  FABRIC FILTRATION   EQUIPMENT
   TYPES AND SIZES

A  Basic Unit

   In its simplest form, the industrial fabric
   filter (baghouse) consists of a woven or
   felted fabric through which dust laden
   gases are forced.  A combination of factors
   result in the collection of particles on the
   fabric fibers.  When woven fabrics are
   used, a dust cake eventually forms which,
   in turn,  acts predominantly as a sieving
   mechanism.   When felted fabrics are used,
   this dust cake is minimal or non-existent.
   Instead, the  main filtering mechanisms
   are a combination of inertial forces, elec-
   trostatic forces, impingement etc. ,  as
   related to individual particle  collection  on
   single fibers.  These are essentially the
   same mechanisms experienced in "Air
   Cleaning Type" filters.

   As particulates are  collected, pressure
   drop across  the filtering media increases.
   Because of fan limitations,  the filter must
   be cleaned.   This cleaning is accomplished
   "in-place" since the filter area is  usually
   too large and time between cleanings too
   short to allow for filter replacement or
   cleaning external to the baghouse.

   In order to meet the challenge of a variety
   of operating  conditions and applications, a
   multitude of proprietary designs exist.
   Essential differences are related to:

   1  Fabric

   2  Cleaning mechanism

   3  I^quipment geometry

   4  Mode of operation

   Depending on the above factors, equipment
   will follow one of three systems, as shown
   in Figure 1.1.
   Figure 1. la shows "bottom feed" units in
   which the dust-laden gas is brought through
   the baghouse hopper and then to the interior
   of the filter tube.  Obviously,  a portion of
   the dust is  removed in the  hopper and never
   reaches the fabric.

   Figure 1. Ib shows "top-feed" units,  in
   which the dust-laden gas enters the top
   of the filter tubes.

   Figure 1. Ic shows units wherein the  gas
   passes from the outside of the filters to
   the interior,  or clean-air side.   With this
   arrangement, the dust inlet can be located
   in many positions.  The fabric can be
   formed in a tubular shape, or it may be
   in an envelope form.

B  Baghouse Operation

   1  Intermittent  operation

     The fundamental principles of operation
     are embodied in the discontinuous-type
     or  intermittent  units.  For such filters
     the entire area  collects dust for a pre-
     set filtration time; at the end of this
     time  the unit is taken out of service and
     the whole area is cleaned of collected
     dust.  Examples of such collectors are
     shown in Figures 1.2,  1. 3,  and 1. 4.
     These collectors are primarily utilized
     for the control of small volume opera-
     tions such as grinding,  polishing,  etc. ,
     and for aerosols of a very coarse  nature.
     They are also used extensively for pilot-
     plant studies and research.   Many of
     these baghouses are  of the  so-called
     "unit" type,  in which the fan and filter
     are contained in a  single piece of
     equipment.

   2  Continuous operation

     For most air pollution control installa-
     tions and major dust control problems
     it is desirable to utilize collectors which
*Chief,  Air Pollution Training, Training Program,  SEC
  pa.c.pm.90.5.6fo

-------
                                                                                                     IP
                                                                                                      cr
                                                               \4
                                                                                                      o
                                                                                                      13
(a) Bottom Feed                  (b) Top Feed              (c) Exterior Filtration


                 Figure 1. 1   POSSIBLE FILTERING SYSTEMS

-------
                                                                                        Fabric  Filtration
MICMAMttM
                                                             aife 18
    (American Wheelabrator and Equipment Co.)
                   Figure 1. 2
(Pangborn Corporation)
     Figure  1. 3
                                         (W.  W. Sly Manufacturing Co.)
                                                  Figure 1. 4

-------
Fabric Filtration
      allow for continuous operation.  This
      end is accomplished by arranging
      several filter areas in a parallel flow
      system, and cleaning  one area at a time
      according to some pre-set mode of
      operation.  Examples of these control
      devices are shown in Figures  1.5,   1. 6,
      and 1.7.

      In a multicompartment baghouse the
      basic filter area is a compartment or
      section (see Figure 1.5).  Each section
      or  compartment is essentially the  Same
      as  a discontinuous unit.  In a reverse
      flow baghouse the basic filter is  one
      envelope or filter bag (Figure 1.6).
      Whereas a small portion of one filter
      tube is the basic filter area in a  re-
      verse jet baghouse (Figure 1. 7).

C  Filter Cleaning

   The heart of any fabric collector is the
   technique employed to remove dust  from
   the fabric.  There are two general types
   of  cleaning; the first involves flexing the
   fabric and the second involves a reverse-
   flow of clean air.   A breakdown of these
   types is as follows, according to com-
   monly accepted terminology:

   1   Fabric  flexing

      a  Mechanical shaking and rapping

        This type of cleaning generally in-
        volves the use of a  "rocker arm-
        lever assembly" to produce a motion
        to  the top of the filter tubes.  The
        motion may be generally horizontal
        (sometimes concave upwards, some-
        times concave downwards), vertical,
        or cover a 90° arc  from bottom to
        top of swing.  Vertical motion is
        sometimes accomplished by rapping.

      b  Sonic cleaning

        This type of cleaning utilizes sympa-
        thetic vibrations from sound waves
        to dislodge dutJ  from the cake.
        Sound waves are generated at  low
        frequency by means of an air horn.
      Figure 1. 8 shows a typical location
      of sound producing horns on the
      clean-air side of the filter tubes.
   c  Collapse cleaning
                                      (1.1)
      To clean filters by the "collapsing"
      technique small reversals in pres-
      sure are created,  such that AP
      from the dirty air side to the clean
      air side is slightly negative. This
      causes the filter tube to deflate, and
      hopefully, the dust cake is discharged.
      In some cases, the tube is slowly
      collapsed and "popped"  open.  If
      desired,  the bags  can be collapsed
      several times per cleaning period.
      Obviously, the baghouse must be
      equipped with suitable valves and
      ductwork to achieve AP reversal.

   d  Pressure-jet or pulse-jet cleaning

      For this method a "bubble" of com-
      pressed air is injected at  the top of
      the filter tube.  A schematic of the
      bubble as  it travels down the tube  is
      shown in Figure 1. 9, in combination
      with collapse cleaning.* '  ' Arrange-
      ment of the system when filtering  on
      the bag exterior is shown in Figure
      1. 10.

2  Reverse-air  cleaning

   a  Reverse-jet

      This mechanism employs  a high
      velocity (small volume)  jet of com-
      pressed air,  blown back through the
      fabric,  to dislodge collected dust.
      Figure  1. 7 shows the usual arrange-
      ment of mechanical cleaning devices.
      In the typical "reverse-jet" filter
      unit,  cleaning can be conducted con-
      tinuously, so that pressure differen-
      tial across the unit tends to remain
      constant.

   b  Reverse-flow cleaning

      Both filter tubes and envelope col-
      lectors can be cleaned by  a reverse

-------
                                                                   Fabric Filtration
 incoming gases
      Filtering
Fil tering
Fi 1 tering
                 J*LJ
to  fan
 All compartments filtering,  dampers open

  incoming gases
                           incoming gases
Shaking
Fi 1 tering
Fi 1 tering
                          to fan
                         One compartment shaking,  balance filtering

                          incoming gases
1

\
/N
Fi 1 tering
s/
_ *
h

Shaking
h
\

Fi 1 tering

>>



/
Fil
S
^
V
/ v
te
\)
t
h
r
>
''ing
/

/^
Fi 1 tering
\/
J*


Shaking
Y

to ffln                                   to fan
One compartment shaking,  balance filtering   One compartment shaking,  balance filtering
           Figure 1.5  TYPICAL PARALLEL FLOW SYSTEM FOR A
             CONVENTIONAL MULTICOMPARTMENT BAGHOUSE

-------
Fabric Filtration
                      INLfl
                                                HOTAHY DISCHARK
                                                    VALViS
             Figure 1.6   CONTINUOUS OPERATING ENVELOPE COLLECTOR

                     WITH REVERSE-FLOW CLEANING MANIFOLD

-------
                                        Fabric Filtration
  REVERSE JET



(Koppers Co. ,  Inc. )



    Figure 1. 7

-------
Fabric Filtration
               Figure 1. 8  AIR HORNS WHICH AID IN CLEANING BAGS
                                                             _TL
                           Figure 1. 9   PRESSURE JET CLEANING

-------
                                                                             Fabric Filtration
      TO
   IXHAUSTEH
A. FILTER CYLINDERS
B WIRE RETAINERS
C. COLLARS
D. TUIE SHEtT
[. VENTURI NOZZIE
F. NOZZLE Of ORIHCI
O. SOLENOID VAIVI
H. TIMER
J. AIR MANIFOLD
K. COUECTOH HOUSING
L. INLET
M. HOPPER
N. AIRLOCK
O. EXHAUST OUTLET
P. MANOMETER
O. UPPER PLENUM
                                     COMPRESSED
                                       AIR
                                     SUPPLT AT
                                     IOO P.S.I.O.

                                      I
                                     INDUCED n.OW
        MATERIAL DISCHARO4
                Figure 1. 10
         flow of clean air.   This would be at
         low or atmospheric pressures, and
         would utilize a much larger volume
         of air than the reverse-jet action.
         An arrangement used for envelope
         collectors is shown in Figure 1.6.
         In this case, the cleaning  manifold
         traverses the baghouse length,
         cleaning one row of envelopes at  a
         time.  Filtration, of course, is on
         the  outside of the envelopes.  An
         arrangement used for tubular bags
         is shown in Figure 1.11.  Note the
         rings which are used to maintain
         filter shape.

D  Baghouse Size

   Bag filter units are relatively large in so
   far as dust collection equipment is con-
   cerned.  Equipment size, therefore, is of
   importance to the buyer as a matter of
   economics and feasibility.
   For given manufacturers, the nominal
   filter velocity is a major factor in deter-
   mining equipment size.  This does not
   mean,  however,  that different types of
   units vary in size according to their rated
   nominal velocities.  Other factors, such
   as height limitations,  bag spacing,  and
   the length to diameter ratio  for tubular
   bags, are also important.  Table 1. 1 lists
   approximate size ranges for various cate-
   gories  of fabric collectors,  at nominal
   velocities of 1  fpm and at  nominal veloci-
   ties considered as normal for the particular
   collector.

E  Filter  Fabric

   1  Types of fabric

      Filter fabrics can be divided into the
      woven or felted classifications.  If
      felted fabrics are used, filter cleaning
      is limited to the pressure-jet and
      reverse-jet classifications.  When
      woven fabrics are employed any clean-
      ing  technique may be used.  In practice,
      however,  bag collectors  cleaned by
      the  reverse-jet technique operate at
      relatively high filter ratios.  To ensure
      high efficiency and to maintain low
      pressure  differentials these collectors
      usually employ felted fabrics.

      Woven fabrics can be sub-divided into
      the  following  classes:

      a Continuous-filament type, in which
         the filaments used to form the fabric
         strands are continuous in structure.
        Such a fabric is characterized by a
         smooth surface and absence of fibers
        or tendrils, and can be constructed
        only of synthetic materials.

      b Texturized strand, in which the
        fabric  strands  are mechanically
        degraded or broken at the surface
        to produce a fuzzy thread.  The
         texturized strand is usually woven
        in the fill direction.

      c Staple  strand,  in which the fabric
         strands are formed from short

-------
Fabric Filtration
                                                  INSPECTION DOOR
               DUST LADEN
                AIR INLET
         AIR REVERSAL VftlVE
         IN NORMAL FILTERING
            POSITION
                                                                  CLEAN AIR
                                                                    TO FAN
                THIS COMPARTMENT
                   FILTERING
                       PARTITION
      AIR REVERSAL VALVE
      IN BACK-WASH POSITION
                                                                     BACK-WASH AIR
THIS COMPARTMENT
IS BEING BACK WASHED
WITH CLEAN AIR.
ACCUMULATED DUST
DROPS  INTO HOPPER
                                                                        FILTER TUBES
                                                                     HOPPER
                          UNIFLOW-BACKWASH DUST COLLECTOR*

                                           Figure 1.11

              *The Ducon Company, 147 East Second Street, Mineola, New York.
10

-------
                                  Table 1.1  APPROXIMATE SIZE RANGES FOR FABRIC COLLECTORS
Collector
Reverse-Jet
(3)
Pressure- Jet
Conventional
tubular bags
Mechanical
Reverse
flow
Envelope
(fpm)
1.0
1.0


1.0

1.0
1.0
Collector volume
per 1,000 cfm
(ft3)
1,250
670


210 - 370

590
210 - 340
Collector
Floor-Area
per 1,000 cfm
(ft2)
57 - 294
111


26 - 50

30 •* 42
21 - 59
Uf(2)
' (fpm)
10
10


3

2
2
Collector volume
per 1,000 cfm
(ft3)
125
67


70 - 123

295
105 - 170
Collector
Floor-Area
per 1,000 cfm
(ft2)
5.7 - 29 . 4
11.1


8.7 - 16.9

15 - 21
10 . 5 - 29 . 5
(1)    Does not include dust hopper.




(2)    Common values for filter velocity.
                                              A
                                                                                        7&
(3)    As manufactured by Pulverizing-Machinery Company, N. J.

-------
 Fabric Filtration
         filaments.  This type of construction
         is necessary for the natural fibers.
         The fabric is characterized by its
         fuzzy appearance and forms a most
         efficient filter because of the fibers
         and tendrils which mat the  surface.

    2  Fabric properties

      a  Insofar as the materials of construc-
         tion are concerned,  prime factors
         of importance are temperature limi-
         tations,  and chemical  stability.
         Other factors which should be evalu-
         ated include "air permeability, "
         resistance to abrasion and shrinkage.
         Table 1. 2 lists some generally
         accepted properties for various
         materials commercially available
         and in use today. Since felted
         fabrics  are made from wool and
         orlon,  it can be seen that the range
   of applications for the reverse-jot
   units is limited to temperatures
   below 270°F and to dusts low in
   alkali content.

b  The actual weave patterns used for
   a fabric will influence the filtration
   process.  This will be discussed
   in greater detail in later sections.
   Some concept as to why  the weave
   pattern should influence filtration
   can be obtained by microscopic
   examination.  Figures 1.12, 1.13,
   and 1. 14 illustrate the fact that
   significant changes occur as both
   materials of construction and weave
   patterns are altered. *   '  Details
   of the fabrics shown are listed in
   Table 1.3.  These differences  be-
   come especially significant when a
   scale on the order-of-magnitude of
   particle diameters is used as a
   reference.
12

-------
                      Table 1. 2   PROPERTIES OF  FILTER FABRICS**
                                                                                       Fabric Filtration
FABRIC
Cotton
Wool
Nylon 6, 6 '"
HT-1 <"
Dacron '"
Orion ">
Creslan  American Cyanamid Reg. Trademark
•Temperatures recommended by Industrial Gas Cleaning Institute
                                                      131 Union Carbide Reg. Trademark        »> W, W. Crlswell Tradenime
   **W.W. Criswell Company,  Division of Wheelabrator Corporation, 800 Industrial
   Higheay,  Riverton, New Jersey.
                                                                                                        13

-------
FIBERGLASS  N2|  FIBERGLASS  N23 ~ti   FIBERGLASS IS|2|   FIBERGLASS N23
    REFLECTED  LIGHT
TRANSMITTED   LIGHT
                     20  X  MAGNIFICATION
                                                                               ^
                                                                               u
                                                                               7


                                                                               ~>

                                                                               ^
                                                                               o
                            Figure 1.12

-------
DACRON   A    DACRON    B
DACRON  A
DACRON   B
   REFLECTED  LIGHT
  TRANSMITTED   LIGHT
                  20  X  MAGNIFICATION
                       Figure 1. 13

-------
FIBERGLASS  NS|   DACRON    B
FIBERGLASS N*l   DACRON   B
    REFLECTED  LIGHT
   TRANSMITTED   LIGHT
                   20  X  MAGNIFICATION
                         Figure 1.14

-------
Table 1.3  CONSTRUCTION DETAILS OF TEST FABRICS
Filter Fabric
Air Permeability (cfm/ft.2 at 1/2" HO)
2
Weight (oz./yd. )
Yarn Count
Filament Diameter (in.)
Yarn Diameter (in.)
Filaments Per Strand
Strands Per Yarn
Twists Per Inch
Weave
Finish
Fiberglass
No. 1 No. 2
13.84 11.67
8.41 8.67
55 x 50 55 x 54
No. 3
7.86
9.06
55 x 58
0.00025
0.0097
408
2
3.8
3/1 Crowfoot
Silicone Oil










Dacron
A B
28.06 14.62
5.51 6.06
82 x 62 82 x 76
0.00113
—
50
1
3.5
3/1 Twill
Silicone Oil
                                                                                                  cr
                                                                                                  2

                                                                                                  o
                                                                                                  S-
                                                                                                  o'

-------
SECTION 23
Fabric Filtration-Basic Concepts

-------
                  FABRIC  FILTRATION  -  BASIC CONCEPTS
I   TERMS AND DEFINITIONS

As a beginning,  consider filtration in an
apparalus similar to that shown in Figure 1.
Such a test apparatus is admittedly far re-
moved from filtration in a commercial filter
unit.  I: will,  however, provide a sound basis
for the development of those concepts  and
equations  necessary to  understand full-scale
filtration.

As shown,  the device would consist of an
inlet duct  leading to an  expansion chamber
and the actual filter media.  Proper gaskets
and filter  media support would be provided.
The cleaned gas would be exhausted through
: ho fan to  almosphcrc-.  Pressure differential
 .•ross the filler media would be measured
by means  of a U-tube or inclined manometer
.•ad expressed in inches I^O.  The volumetric
flow role (Q),  the area  of the filter (Af) and
ihe mass of dusl filtered (W) are known or
arc measurable quantities.
The term "filter pressure differential" (AP)
has several synonyms which should be noted.
These are:

   1   Pressure  drop

   2   Pressure  differential

   3   Head loss

A second fundamental term is the superficial
face velocity (Uf) for the carrier gas stream.
By definition:

                Uf  = ~                 (1)

Synonyms for superficial face velocity are:

   1   Filter velocity

   2   Filter ratio
                                               Direction of gas  flow
                                               and dust feed
                                                 I
                             AP
                           T




^




filter
\


Clean gas
(suction


out
system)
U-tube manometer
              AP represents  the  filter pressure  differential  (inches  H 0)
                           1.  SCHEMATIC OF BASIC TEST APPARATUS

-------
Fabric Filtration   Basic Concepts
   3  Air-to-cloth ratio

   4  Nominal velocity

In developing the concepts of filtration it is
fundamental  to make the following assumptions:

   1  Flow through the filtering media and
      the dust cake is laminar in  nature.

   2  The aerodynamic properties of the
      filtering media and dust  cake  are con-
      stant for the filtering period.

   3  The weight of dust filtered per unit of
      time is constant.

Under these  conditions,  several plots of
pressure differential as a function of filtering
time might appear as shown in Figure  2.

At this point, it is necessary to evolve those
basic factors whi'ch will allow the curves of
Figure 2 to be evaluated in  terms of assump-
tions  1.  2 and 3.
                                                  First, it should be noted that the pressure
                                                  drop is not a fundamental characteristic of
                                                  the system.  The situation can be compared
                                                  to the pressure differential across an orifice.
                                                  For a given orifice the pressure drop can be
                                                  used as an indicator of the flow rate; for  two
                                                  dissimilar orifices, however,  the same pres-
                                                  sure drop would indicate different flow rates.

                                                  For a fixed laminar flow element, a direct
                                                  proportionality between pressure differential
                                                  and face velocity will exist.    That is:
                                                           AP
                                                         AP
                                                         	r
                                                         IT
                                                                                          (2)
                                                  In Figure 2 for example, the ratio AP/Ur. for
                                                  curves A,  B,  and C and at  time = 0 is equal
                                                  to 0.5 in. H2O/fpm.  This  ratio will hold true
                                                 "as long as assumptions 1 and 2 are valid;
                                                  that is,  either turbulent flow is not reached
                                                  or the filter media is not changed by elonga-
                                                  tion, compaction or other means.
      10
    O)
    +->
    ^ 4
                Curve
                 A
                 B
                 C
Fabric
type
                                      Tim*
                                 A P
                                t H20)
          0.75
          1 .5
          3.0
                                          U
                                         (fpm)
                                           1 .5
                                           3.0
                                           6.0
(grs  ft')
5
2.0
1.0
                                       End of cycle
         AP
                                  (in.
3.75
7.5
7.5
         Uf
       (fpm}
                     1 5
                     3.0
                     3.0
                       (g«  ft3)
5
2.0
2.0
                                                         _L
                 20        40        60        80        100
                                Filter  pressure differential, in  H20
                      Figure 2.  POSSIBLE PRESSURE  TIME  CURVES

-------
                                                       Fabric Filtration  Basic Concepts
          AP
The ratio —  is called "Filter Drag" and
           f
given the  symbol S.  In the equation form:

The rate at which dust is filtered can be cal-
culated from the equation:
        =  C
     t      p  A
                Q
C  • U,.     .
 p    f(avg.)
   (4)
In other words, assumption 3 is fulfilled if
the product (Cn •  UJ      is constant.
              Pf avg.
It is important to recognize the distinction
between assumption 3 and the often quoted
assumptions of "Constant Flow Rate" and
"Constant Dust Concentration. " Curves B
and C of Figure 2 reflect the difference.  For
curves  B and C assumption 3 was maintained
so that  the product (C_ • Uf)      remained
           ^         P   1 avg.
6. 0 (grains/ft ) per minute.  For Curve C,
however,  both Cp and U^ were varied.  The
result is the non-linear  relationship.  The
straight line relationship (Curves A and B,
Figure 2)  is the result of a constant feed
rate and constant flow rate.

It is at once obvious that changing the dust
feed rate  will alter the slope of the line,
although the basic characteristics of the dust
cake need not be altered.  This then, leads
directly to plotting Filter Drag (S) as a func-
tion of Filtered  Dust  Mass (W) in order to
delineate  the basic performance character-
istics of the system under study.  Replotting
the data of Figure 2 in these terms results
in a single line,  as shown  in Figure 3.   This
representation is known as a "Basic Per-
formance  Curve.
            II  INTRINSIC DUST CAKE PROPERTIES

             If assumptions 1,  2 and 3 are valid the basic
             performance curve will always be  a straight
             line; the slope of this line  will depend on
    3.0
  E
  CL
  . 2.0
  S-
  d)
    1.0
                           I
I
                          I
       I
I
                100       200        300        400       500 2     600
                                Filtered Dust Mass,  grains/ft
             Figure 3.  BASIC PERFORMANCE CURVE FOR DATA OF FIGURE 2

-------
Fabric Filtration - Basic Concepts
properties of the deposited dust cake.  An
indication of some of the basic  parameters
influencing the properties of the cake can be
obtained by examination of several equations
developed for laminar flow through packed
beds,  as  listed below:
   Chilton-Colburn
                   (2)
          AP

          Uf
JL  11
G   F
 c   a
          D
                          (5)
                           pe
   Fair-Hatch
              (3)
       — = —
       U,    G   KF&H
         f     c
                                9   A
   Hatch
         (4)
          AP
          U7
          (5)
                          (1)
AP

 f
          G   C,
           c   k
               1    (1 -ex)
                  H-     T~
                            D
                         2  A
             _£_
             V
              P
                          (8)
Historically,  each of these equations have
rosulled from an examination of D'Arcy's^
basic premise that resistance  to flow is pro-
portional to flow rate and depth of media.
In the terms so far developed,
          AS =
      = k X
                            AL
                          (9)
Because of assumption 2 it can be said that
the change in depth of the dust cake (AL) is
proportional to the change in weight of dust
collected (AW).   Therefore,
          AS
          or K
1 (AW)
 K
  AW
  AS~
                                         (10)
                                         (11)
which is the inverse-slope of the Basic Per-
formance Curve.  The proportionality constant,
K,  is given the term "Dust Permeability, "
and has the dimensions (grains  ft )l
(in. H2O/fpm).

From equations (5)  through (8),  it is  apparent
that K will be greatly influenced by particle
size and size distribution,  particle shape,
particle surface characteristics, manner of
cake formation,  and gas viscosity (viscosity
is discussed in detail in a latter section.)
Excepting viscosity, it is apparent that these
factors cannot be measured in a dust  cake.
Therefore,  dust permeability will be  a prime
factor for possible correlation with other
operating conditions.  The  terms shown in
equations (5) through (8) will  be  useful in
interpreting these results.  Generally
speaking, high values of dust permeability
imply a dust "easy" to filter, while a low
value of dust permeability implies design
problems in terms of pressure differential,
filtering time, and filter velocity.

Examples of the manner in which permeability
varies from dust  to dust are Shown in
Table l.<6>
                         /n\
According to the authors,    each dust listed
in Table 1 met the conditions of  laminar flow
and the cake properties were constant through-
out the experimental runs.  Several factors
should be noted,  however,  which prevent
extrapolation of the general conclusions to
other situations:

A  The dusts were deposited at a constant
   filter ratio of 10.0 fpm.  This, in  itself,
   is a high velocity that could force the
   deposition of a cake in an already compact
   condition.  In actual practice, velocity of
   deposition is a variable  quality,  as will
   be explained later.

B  The filter itself consisted  of  an unidentified
   piece of cloth "stretched tightly  to prevent
   sagging,  which was found to affect the
   resistance. "  The manner of  drawing the
   cloth is not  noted.  It can be  expected that
   these parameters,  (i. e.  , the cloth itself,
   the fact that the cloth could not stretch or
   change shape,  and the manner of cleaning)
   might influence dust permeability, and any
   relationship between filter drag  and weight
   of dust collected.

-------
                                                           Fabric Filtration   Basic Concepts
                   Table 1.  PERMEABILITY OF SEVERAL TEST DUSTS AS A
                        FUNCTION OF PARTICLE SIZE DISTRIBUTION*6*
Material
Granite
Foundry
Gypsum
Feldspar
Stone
Lamp black
Zinc oxide
Wood
Resin (cold)
Oats
Corn
Particle Size
Coarse
<20 Mesh
4,430
11,300


7,300


4,430
11,300
<140 Mesh
3, 180
4,430





11,300
<375 Mesh

1,850
1, 110
1, 110



4, 430
Medium(b)
< 90 ^




1, 110

1, 110
730
1,850
< 45 u







636
795
Fine
< 20u
363

370
257




<2M





148
446

278
(a)  Flocculated material not dispersed; size actually larger
(b)  Theoretical size based on density of silica; no correction made  for actual particle  density
   Table 1 also illustrates a difficulty that would
   arise in attempting to correlate dust cake pro-
   perties with particle properties.   Thus, dust
   permeability is listed as 4, 430 for Corn Dust
   passing a 375 mesh sieve (aperture equal to 39
   microns).   Note,  however, that dust  perme-
   ability of Corn Dust elul riated to sizes less
   lhan 4Ti microns is only 7!)5.  The experimental
   techniques, therefore, produced  two  entirely
   dil'ferenl dusts.  This discrepancy does no I
   negate Hie need for basic studies, but does
   indicate a need for caution in extrapolating
   such  results to operating conditions.
 Ill  FACTORS INFLUENCING DUST CAKE
     PROPERTIES
By the year 1955, the concepts of Williams,
el. al. ''J'  had been further investigated by
                 /7\              °      J
Snyder and Pringu' and utilized by  Hemeon.
                                             /Q\
                                             v  '
                                             .
  They were also published by Lapple^) in the
  3rd Edition of the Chemical Engineer's Hand-
  book.  As a result of their experiments,
                                                                             (7)
                                                   however, Snyder and Pringv ' questioned the
                                                   validity of constant  dust cake properties.

                                                   Two sets of data published by the authors
                                                   are shown in Figures 4 and 5.
Figure 4 shows the influence of the fabric
itself on rale of filter drag increase.  From
these data,  the authors concluded that for the
same particulate properties,  dust  cake
permeability increases as the  fiber surface
area (square feet per square feet of fabric)
increases.  To use an extreme example
illustrating this relation, consider the same
dust collected "on the surface" of a membrane
filter and also  collected by means  of a woven
cotton fabric.   In the first instance, the cake
formed  would not be influenced by  irregulari-
ties in the surface or fibers projecting beyond
the surface; it  would,  in fact,  be a pure dust
cake, with a clear dividing line between dust
and filter.  This would not be  the case when
using the  cotton fabric.  In fact, we would
logically expect a change in properties as the

-------
Fabric Filtration - Basic Concepts

       2.0
E
CL
O
 CM
:r
en
(O
i-
o
1.0
                                         A.   1-12 Fiberglas  fabric,  low fiber surface
                                             area per square foot of cloth.

                                         B.   Napped B-27 Orion,  medium fiber surface
                                             area.

                                         C.   B-26 staple  Orion fabric  napped both  sides,
                                             high fiber  surface  area.
                               200
                                  300
400
500
600
                                       Filtered  Dust  Mass,  grains/ft
                            Figure 4.   EFFECT OF FABRIC ON BASIC PERFORMANCE  CURVE
      2. or
o
 IN)
 cn
 (O
$
                                                        A.   High  twist  unnapped  Orion,  low
                                                             fiber surface  area per  square
                                                             foot  of  cloth.
                                                         B.   Fiberstock Orion, high  fiber
                                                             surface area.
                  TOO
                                       Filtered Dust Mass, grains/ft

                             Figure  5.  EFFECT  OF  FABRIC ON BASIC PERFORMANCE  CURVE

-------
                                                         Fabric Filtration   Basic gorurepts
dusl deposit first bridges the fabric pores,
then proceeds to fill  surface irregularities,
and eventually,  if sufficient quantities are
deposited, extends completely beyond the
range of fabric influences.  With a relatively
coarse dust the  depth of material per unit
increase in drag would undoubtedly be greater
than with fine materials.  Therefore, the
likelihood of a linear relationship between
drag and dust mass is greater with coarse
materials when  compared to fine dust.
Figure 6 should help  visualize the situation.

Figure 7 further supports the argument.  In
this case, different dusts are collected by the
same fabric.  The data do not negate  the
possibility of linear relationship for the finer
materials at higher dusl  loadings on the filter.

It should be noted that the test apparatus,
although similar to that  used by  Williams,
differs in several important aspects:

A  In the experiments of  Williams, the nature
   of the filtered dust was altered by  sieving
   and elutriation columns.  Snyder and Pring,
C
on the other hand,  passed the air-dust
mixture "through a settling chamber
(duplicating the action of the  inlet  chamber
of the Dustube collector)". It is not clear
if Snyder and Pring measured particle
properties (i. e. , grain loading and size
distribution) before or after the settling
chamber.  This, in turn,  makes it difficult
to generalize on the results obtained.

The  fabric is apparently unsupported  in
the apparatus used by Snyder and Pring.
Stretching of the cloth as drag forces
increased would undoubtedly lead to
changes in dust cake structure.

At the end of each run, the cloth was
cleaned by "shaking the fabric for  a timed
period in a prescribed manner. "  This
undoubtedly achieved better cleaning than
is experienced in practice, and would
necessitate the build-up of the entire  dust
cake with each run.  It has been our ex-
perience that this "over-cleaning" almost
always  results in non-linear filter-drag vs.
dust  mass curves,  especially in the
beginning of a run.
  Drag
                                   Depth of fabric  plus cake

             Figure 6.   DRAG AS A FUNCTION OF MASS OF COLLECTED DUST

-------
2.0
              1000
2000
     Filtered Dust  Mass,  grains/ft



3000        4000       5000       6000
                                                                                     7000
              100
200
            400
500
600
700
                                          Filtered Dust Mass, grains/ft
                                                                       8000
                                                            Wheelabrator Steel  Scale
                                                                               Dust
                                                            Moderately Coarse
                                                            2% less than
                                                            10 microns
                                                                Ground Limestone Dust
                                                                Coarsest,  32.33
                                                                minus 325  mesh
                                                                                                     0-5
                                                                       -   0.4
                                                                                                 -   03
                                                                                                 -   0.2
                                                                                                             O
                                                                                                              csi
                                                                                                             01
                                                                                                             
                                                                                                    fD
                                                                                                    •o
                Figure 7.  EFFECT OF DUST TYPE ON FILTER RESISTANCE  OF COTTON SATEEN  CLOTH

-------
                                                           Fabric Filtration   Basic Concepts
IV  PERFORMANCE ON COMMERCIAL-SIZE
    FILTER TUBES

  When the filtration process is extended from
  a single area, as previously described, to
  I he area of a commercial size filter tube,
  (he overall process becomes  less uniform
  from area lo area.  An evaluation as to the
  extent of these non-uniformities formed the
  basis for- research by  the Public Health
  Service which began in 1957.   The basic test
  unit for1 this research  is shown in Figure
  With this unit it  was possible to control I he
  volumetric flow  rale at a pro-set level,
  maintain constant  dust feed rates, vary shaking
  conditions relative to amplitude,  frequency,
  direction and duration, and vary tube diameter
  and  length.  The unit could be programmed
  for eonl inuous cycling, so that equilibrium
  could be approximated.  Point measurements
  of dusl  mass and filter velocity could also
  he made, using the probes shown in Figures
         As a result of studies on this unit,  it  lias
         been shown that dust  cake properties  vary
         over a filter area.  Typical  "profiles" illus-
         trating such variations are shown in Figures
         11  and 12.
                   (12)
       Further,  it. can be concluded
         that a non-uniform residual profile will exist
         at the start of every filtration period after
         the first, and that, as a result, filtration
         no longer occurs uniformly over the entire
         filter surface.  Rather,  different filter
         velocities,  different amounts of dust collected,
         and different  cake structure's will be found at
         various locations.

         In relation to dusl permeability, it  should be
         noted that  "when  considering larger areas such
         as that of an entire filter bag,  no longer is
         matrix structure solely of import,  but
         now,  the macroscopic structural features
         or 'topography1 of the  cake must be con-
         sidered.  In other words,  the shape of the
         mass and resistance profiles will have direct
         and significant bearing on the resultant effec-
         tive permeability of the dust collected on the
                   f-
                                       (*)
M L L T (j NI f I ( IS
 Nl VI ir, A I Hi nwf.HS
 LOWER HOU !-' i NG
UU'j f LNTRfllNME-Nl CHAMHER
 U'.T Ft" EDER
 iGU-SPELO  FAN
 QHPER
8  FiLUR TUUES
9  SHAKING MECHANISM
10 DUST CATCH BINS
  "IN PLACE"
I I  OUT.T CATCH tJINS DUR-
  ING FILT RAT lON
12 HOPPER DOOR
li DUST CATCH HOPPER
                         ©
                           Figure 8.  SCHEMATIC DRAWING OF TWO
                                         BAG TEST UNIT

-------
 Fabric Filtration - Basic Concepts
    Figure 9.
Pb21°-Bi21° MASS PROBE
Figure 10.  FILTER VELOCITY PROBE
 filter.  This seeming anomaly exists because
 of the physical configuration of the cake over
 the  whole filtering area and it demonstrates
 that data from laboratory bench-scale deter-
 minations of permeability  cannot be equated,
 for  design purposes, to the effective perme-
 ability of the same dust on full-sized tubes."

 Obviously, these same conclusions can be
 extended by considering possible variations
 from filter to filter in a single compartment.
V   SUMMARY

 A  Dust Collection Mechanisms

    1   Initial stages of filtration

       During the initial stages of a filtration
       cycle the collection of dust particles is
       accomplished by means of interception,
       intertial deposition,  diffusional impac-
       tion, and electrostatic forces,  as the
       dust laden gas is passed through a fabric
       of woven or felted structure.

    2   Dominant  collecting mechanism

       After some short interval of operation
       sufficient  dust will be retained to form
       a dust cake.   Once the dust  cake has
       been formed the dominant collecting
       mechanism is that of sieving.  Because
       the  sieving media  is composed of the
       material being collected pore sizes  in
       the  sieve will be on the order of particle
       diameters.  Therefore,  efficiency of
       collection is high,  and the fabric filter
       becomes a positive collection device.
                                     B  Properties of Dust Cakes

                                        1  Laminar flow elements

                                          The collected dust cakes, being com-
                                          posed of small particles and being
                                          subject to low superficial face velocities
                                          are considered as laminar flow elements.
                                          At any instant in time, the ratio of
                                          pressure drop across such an element
                                          to the superficial face velocity through
                                          the element is related to properties of
                                          the system not easily measured at this
                                          time.

                                     C  Residual Filter Drag

                                        Residual filter drag  is that drag value
                                        which exists after a  dust fabric combina-
                                        tion has been cleaned.   As such it is
                                        dependent on many interacting factors and
                                        relationships,  such as mode of cleaning,
                                        type fabric, the particular dust involved,
                                        velocities of filtration,  and,  time in
                                        service.  Experience has indicated that
                                        this residual resistance is a major portion
                                        of the total resistance across a fabric
                                        filter unit,  despite references to the
                                        contrary in the literature.  At the present
                                        time there  would appear to be no method
                                        whereby  this quantity can be estimated,
                                        although  for most installations it will
                                        probably lie within the range of 0. 5 to
                                        1. 5  in. H2O/fpm.

                                     D  Dust Permeability

                                        It can be seen from Figure 13 that after
                                        the initial stages of filtration a constant
 10

-------
                                                          Fabric Filtration  Basic Concepts
                         TYTY'Y  \
                          W0.1 W0.4 W0.5 W0.fe TO.IO W

-------
 Fabric Filtration  Basic Concepts
    relationship exists between changes in
    filter drag and increments of dust filtered.
    From equations (5) through (8) it can be
    surmized that the only factor changing the
    drag is the cake thickness,  L.  Since  it is
    impossible to measure L it becomes easier
    to consider cake  thickness as proportional
    to the weight of dust deposited.  On this
    basis,  it can be said that under particular
    conditions, the cake properties indicated
    in equations (5) through (8) can be expressed
    by one proportionality constant, the dust
    cake permeability, which is  calculated
    from the equation:
              K  =
AW
AS
(11)
   As discussed, it is important to know the
   conditions under which K has been deter-
   mined.  Of special significance is the
   filter area involved.

E  Terminal Filter Drag

   The terminal filter drag is that resistance
   of the dust fabric combination which exists
   immediately prior to filter cleaning.
12

-------
SECTION 24
Fabric Filtration Operations and
  Industrial  Applications

-------
                   FABRIC FILTRATION OPERATIONS
                                AND
                      INDUSTRIAL APPLICATIONS
  I.   INTRODUCTION

      A.   The Baghouse
      B.   Filtration Processes
      C.   Operational  Parameters

 II.   THE FILTERING MEDIA

      A.   Fiber Types
      B.   Yarn Construction
      C.   Weaving
      D.   Fabric Treatments
      E.   Bag Construction

III.   BAGHOUSE DESIGN

      A.   Cleaning Processes
      B.   Baghouse Construction
      C.   Ma i ntenance
      D.   Dust Disposal

 IV.   INDUSTRIAL APPLICATIONS

      A.   Cement Kilns
      B.   Foundry Cupolas
      C.   Steel  Furnaces
      D.   Nonferrous Metal  Furnaces
      E.   Carbon Black  Plants
      F.   Grain  Handling  Operations
      G.   Other  Applications
PA.C.pm.  103.4.73

-------
I.   INTRODUCTION

A.   The Baghouse

Filtration is the removal of solid particles
from a fluid by passing the fluid through
a filtering medium on which the particles
are deposited.  The industrial fabric filter
or baghouse consists of a woven or felted
fabric through which dust laden gas is
forced.  As particles accumulate, resistance
to gas flow increases and pressure drop
across the filter increases.  Accumulated
deposits are removed periodically by clean-
ing of the fabric in order to maintain
practical head loss.  Provision for cleaning
the fabric filters in place is a distinguish-
ing design characteristic of baghouses.

Baghouses are used to collect particles in
the size range of submicron fumes to powders
greater than 200 microns.  Fabric filtering
materials have been developed to handle gas
temperatures up to 550°F and can withstand
most chemical reactions.  Recovery of
collected material is also a characteristic
of baghouse operation.

Superficial filtration velocities (total
air volume filtered/total cloth area)
commonly called the air-to-cloth ratios are
a function of the quantity of ventilation
or process gas, the dust concentration, and
the flow resistance properties of the parti-
culate deposit.  The air-to-cloth ratio is
generally in the range of 1 to 15 feet per
minute and pressure drop is about 3 to 10
inches of water.

Commercial collectors are available in sizes
from a few square feet of cloth up to several
thousand square feet.  Smaller units may be
fabricated and assembled by production line
processes while larger units are usually
designed to meet the requirements of specific
applications and are assembled at the
installation site.

Costs for baghouses vary in proportion to
size and with respect to kind and arrange-
ment of fabric and cleaning apparatus.  A
range of 0.35 to 1.25 dollars per cubic foot
of gas per minute is typical of initial
collector costs although actual installed
costs run up to three times the base cost.
Where very high efficiency of collection of
small particles is required, baghouse oper-
ation costs are lower than other types of
collection equipment.  If fabric filters
are properly designed, installed, operated,
and properly maintained, collection efficien-
cies in excess of 99.9% may be obtained.

B.  Filtration Processes

Fabrics used for removing dust and fumes from
gas streams are usually woven with relatively
large open spaces up to 100 microns in size.
Since collection efficiencies for dust parti-
cles of 1 micron or less may be.greater than
90%, the filtering process obviously cannot
be simple sieving.  Dust particles are cap-
tured and retained on the fibers by means of
interception, impingement, diffusion,  gravi-
tational settling, electrostatic attraction,
and sieving.

Direct interception is possible because the
flow through fabric filters is usually lami-
nar.  A dust particle experiences direct
interception by the fabric filter when it
comes in contact with a fiber as the stream-
line passes by the fiber.  The particle adheres
to the fiber because of van der Waals forces.
This occurs in baghouses for particles less
than 1 micron in size as inertia effects
become dominant for particles larger than
that.

Larger particles do have appreciable inertia
and do not follow a streamline when the
streamline is deflected from a straight path.
The probability of a particle contacting the
surface of the obstruction depends on the
size of the obstruction and the size and
inertia of the particle.  Smaller fibers are
more effective as streamlines pass closer to
smaller obstructions than to larger obstruc-
tions.  Particles with greater inertia are
more likely to strike and impinge on a col-
lecting surface than a particle with less
inertia.

For very small particles with diameters less
than 0.1 or 0.2 microns, diffusion is the
prominent mechanism of collection.  Particles
as small as these no longer follow stream-
lines because collisions with gas molecules
occur, resulting is random Brownian motion
that increases the, chance of contact between
the particles and the collection surface.

Not a great deal of information is available
that describes the role electrostatics play
in the collection of particles by baghouses.

Electrostatics may  not  only assist  filtration
by  providing  an  attractive  force  between
dust  and  fabric,  but  also may  affect  agglom-
eration,  fabric  cleanability,  and  collection

-------
 efficiency.   Charging of particles is likely
 due  to  frictional effects.

 Once a  mat of dust or filter cake is accumu-
 lated,  further collection is accomplished by
 sieving as well as the other mechanisms.  The
 fabric  material serves as a supporting struc-
 ture for  the  filter cake.  After bag cleaning,
 some residual dust remains to aid in further
 filtering. A  short time after cleaning the
 dust cake acts as a seive and becomes the
 dominant  collecting mechanism. It is during
 this period that the extremely high efficienc-.
 ies  are reached.
              dust and  filter  combination  is  influenced by
              particle  size, size  distribution,  particle
              shape, surface characteristics,  the manner of
              dust cake formation,  and  the gas viscosity.
              In general, a high value  of  permeability
              implies a dust that  is  easy  to  filter.   A low
              value means that a high pressure drop may be
              expected along with  problems with  filtering
              time and filter velocity.

              Table  1  shows some  values of permeability
              for different dusts  and dust sizes.  Note that
              from the largest to  the smallest particle
              permeability may vary by  a factor  of 1000 or
              more.
C.  Operational Parameters

Clean filter air resistance to flow is depen-
dent on fiber structure and the weave of the
cloth.  A tight weave offers more resistance
to flow than a looser weave at the same rate.
As previously determined, the air flow
through a fabric is laminar and because of
this flow resistance or pressure drop will
vary directly with flow.  Air flow through a
fabric filter is usually described as the
superficial face velocity or air—to-cloth
ratio.  By definition:
                                      (1)
where Q is the volume flow rate of dirty air
through the filter and A  is the surface area
of the filtering media.  Since the pressure
drop, AP, is proportional to the superficial
face velocity, U , a constant known as the
filter drag, S, may be defined:
            AP
      S  =  —
            n
(2)
The filtering process of a fabric filter is
a function of the dust cake.   Factors that
influence the properties of the dust cake are
worth discussing.  A term commonly used with
regard to fabric filters is permeability.
Permeability is the openness  of a material to
the transmission of a fluid.   It is not deter-
mined theoretically for fabric filters, but
is measured using the relationship:
            AW
      K  =  	
            AS
(3)
where K is the permeability, AW is the change
in weight of dust collected during a change
in filter drag, AS.  The permeability of a
              A system of bags in a baghouse  is  similar  to
              a parallel electrical system.   An  analogy  may
              be drawn between total electrical  resistance
              and total flow resistance, S  ,  for a baghouse.
              Pressure drop may be determined across  each
              bag in a baghouse providing an  equivalent  or
              total pressure drop of
                          111           1
                  JL  =  	  +  	  +  	  +...  	   (4)
where n is the number of bags.

Figure 1 shows the basic performance curve of
the baghouse operation. Residual drag is that
drag value which exists after a dust - fabric
combination has been cleaned. It is dependent
on mode of cleaning, type of fabric, type of
dust, filtration velocities, and time in ser-
vice. Experience shows that residual drag is
a. major portion of the total resistance across
a. fabric filter. It is estimated to be about
0.5 to 1.5 inches of water per foot per min-
ute. Terminal filter drag is the resistance
of the dust - fabric combination immediately
prior to cleaning. The terminal resistance of
the fabric filter during the filtering opera-
tion may be estimated using:

                                                      (5)
where ST is the terminal resistance of the

filtering element. S* is determined by ex-
                    K
tending the linear portion of the Basic
Performance Curve to the zero intercept. C

represents the particulate concentration in
the dirty gas stream and t is the elaped time
of operation. The term U • C  • t is the
                        t   P
change in weight of dust collected per unit

-------
Material
Granite
Foundry
Gypsum
Feldspar
Stone
Lamp black
Zinc oxide
Wood
Resin (cold)
OatB
Corn
Particle SUc
Coarse

-------
of fabric area.  With this relationship  the
time between required cleanings  may be  de-
termined if the  terminal bag resistance is
set at some maximum level.  The face velocity
may be calculated and the dust concentration
is known or may  be sampled.  The  dust permea-
bility is determined from pilot  tests or pre-
vious experience.

For multi-compartment baghouses  one other
team must be defined. This is the equivalent
terminal drag of the baghouse system, S^. As
the number of bags or the number of compart-
ments increases  the equivalent terminal drag
effectively decreases for the same filtering
conditions. The  equivalent terminal drag may
be estimated by:
   et
                                         (6)
where m is a constant that is dependent on
the number of bags or compartments.  Table 2
shows the values of m as a function of the
number of compartments, n.
          Table 2.  VALUES FOR m AS
                FUNCTION OF n
1
2
3
4
5
6
7
8
9
10
1.0
0.78
0.72
0.68
0.65
0.63
0.61
0.60
0.59
0.58
11
12
13
14
15
16
17
18
19
20
0.57
0.56
—
0.55
--
—
0.54
--
—
0.53
II.  THE FILTERING MEDIA

A.  Fiber Types

The type of fabric selected for use in a bag-
house is an important consideration in a pro-
perly designed installation.   The material
must be compatable with the temperature of the
gas and the chemical constituents of the
effluent.  There are several types of fibers
used for filters in baghouses; each one has
characteristics which give it advantages  over
the others.

For many years cotton has been the  standard
fiber for most common dusts.  It is  inexpen-
sive, readily available, effective,  and rea-
sonably durable.  One major limitation is
temperature, as cotton cannot be used for gases
with temperatures greater than about 180°F.
Cotton is not recommended for effluent gases
of high acid or alkali content.  It  is used
extensively in applications for abrasive blast-
ing, rock crushing, and conveying.

Wool felt is used for many metallurgical
operations such as lead smelters and for
reverse-jet baghouses.  The temperature limit
for wool is about 220°F and wool can resist
the action of acid effluents reasonably well.
Wool has been mixed with asbestos for some
applications.

Nylon is relatively high in initial  cost, but.
has excellent resistance to abrasion and  flex-
ing, and a resistance to many chemicals.  Nylon
has a slick surface that allows for  easy
cleaning of the fabric surface. However,  other
synthetic fibers with similar properties  are
used more frequently than nylon because the
thermal properties of nylon allow only 220°F
for continuous operation.

Dynel is an acrylic fiber that has low moist-
ure absorption, good strength, resilience,
and resistance to many chemicals, mildew
and bacteria.  Dynel will not support combus-
tion and is used in applications where chem-
ical resistance is important.  The maximum
temperature gas recommended is about 175°F.

Orion and Dacron have similar properties  in
that both resist chemical reactions  and have
good heat resistance.  Orion is produced  only
in staple form while Dacron may be  purchased
in filament-type yarn.  A filament-type fabric
is more easily cleaned.  Because of  this  and
because Dacron is less expensive, many early
Orion users have switched to Dacron.  The
temperature limit for both is about  275°F.

Teflon has been used as a baghouse  material
for high temperature gases.  The tetrafluoro-
ethylene fiber can withstand continuous oper-
ation temperatures of about A50°F to 500°F
and is inert to most chemicals except fluorine
and chlorine.  The flex and abrasion strength
of Teflon bags is only fair and  teflon is
expensive.

Of all the materials available for  filtration,
glass fabrics have the highest resistance to
high temperatures and most  chemicals.  Glass

-------
                               Table 3.   PROPERTIES OF FILTER FABRICS**
FABRIC
Cotton
Wool
Nylon 6, 6 '"
Nomex
Docron '"
Orion '"
Creslan "'
Dyn.l '"
Polypropylene
Teflon '"
Hberglas
Flltron "'
Helling
Temperature
Decomposes at
302° F
Chars at 572° F
480° F
Chars at 700° F
482° F
482° F
Softens
475° F
Softens
325° F
Softens
333° F
Decomposes at
750° F
1470° F
505° F
Softens
fMlllWM
T.SSSS-
180° F
200° F
200° F
400° F
275° F
260° F
250° F
160° F
200° F
500° F; emits
toxic gas at
450° F
550° F
270° F
Acid
Resistance
Poor
Very Good
Fair
More resistant than
Nylon; inferior to Dae-
ron I Orion.
Good to most mineral
acids. Dissolves partial-
ly in concentrated
H.SO..
Good lo excellent in
mineral acids.
Good in mineral acids.
Little effect even In
high concentration.
Excellent
Inert except to fluorine.
Fair to Good
Good to excellent.
Mkili
Resistance
Very Good
Poor
Excellent
Not as resistant as
Nylon; superior to Dae
ron i Orion.
Good in weak alkali.
Fair in strong alkali
Fair to good in weak
alkalis.
Good in weak alkalis.
Little effect even in
high concentration.
Excellent
Inert except to chlorine,
tri-fluoride and molten
Alkaline metals.
Fair to Good
Good
Ffei
Abrasion
Very Good
Fair to Good
Excellent
Good
Very Good
Good
Good to Very Good
Fair to Good
Excellent
Fair
Fair
Good to Very Good
                <'> Du Pont Rag Tradamark     (!) American CyaMmid Reg Tradamark
                •Taniptralurti recommended by Industrial Gaa Cleaning Institute
                                                      111 Union Carbide Beg Trademark
                                                                           "> W. W. Criiwell Tredename
                  **W.W. Criawell Company, Division of Wheelabrator Corporation, 800 Industrial
                  Higheay, Riverton,  New Jersey.
 fibers,  however, have a low  resistance to
 abrasion and crushing, so special  precautions
 must  be  taken during the cleaning  cycle.
 Vigorous shaking is avoided  and  filtering
 velocity is usually less than  for  other
 fabrics  on the same dust.  Fiberglass  bags
 have  been applied to gases with  temperatures
 up  to  about 550°F.

 Table  3  shows most of the types of  fibers
 used  for filter bags in industrial applica-
 tions  and some of the characteristics  of  each.

 B.  Yarn Construction

The construction of the yarn for filter cloth
 is as  important as  the material used to make
the yarn.   The weave,  count, and finish are
characteristics that may be controlled  in
making the  filter cloth.   The  two  main  types
of yarns  used  in weaving are filament yarns
and staple  yarns.   Only synthetic  fibers  may
be made  into  filament yarns as synthetic
fibers are  manufactured by extruding material
through  a spinneret to form long individual
filaments.  These filaments may be twisted
together to form high tensile strength multi-
filament yarns.   Filament yarns have a slicker
surface  than  do  staple yarns.  Staple yarns
of synthetic  fibers are produced in a similar
manner except that filaments of staple yarns
are much shorter and finer.  The surface  of
the synthetic staple filaments is often
textured by using compressed air to rough up
the surface as the filament is extruded from
the spinneret.

Cotton staple fibers are cleaned and drawn
into parallel order by carding and are event-
ually twisted into yarns by a spinning process.
Some of  the properties of the spun yarn depend
on the spinning  and the amount of twist.  A
highly twisted yarn tends to resist penetration
of particles  into the interstices of the  yarn.

-------
Yarn number is a measure of linear density.
Direct yarn number is the mass  per unit length
of yarn, while the indirect yarn number or
yarn count is the length per unit mass  of  yarn.
Classification of yarns is different for cot-
tons and synthetics.   Cotton yarns have been
numbered by determining the weight in pounds
of 840-yard lengths or hanks, but more  fre-
quently by determining the number of 840-yard
hanks per pound.  For example,  if 840 yards
weigh one pound, the yarn count is Is;  if  30
such hanks weigh one pound, the count is 30s.
A heavy yarn would be Is, a medium yarn would
be 30s, while a light yarn may  be 160s.

Woolen system yarn is measured  by the number
of 300-yard hanks per pound, while worsted
system yarn is measured by the  number of
560-yard hanks per pound.  Man-made fibers
are usually measured using the  denier system.
The denier is equal to the weight in grams of
9,000 meters of yarn.  The cotton yarn  count
may be obtained by dividing 5,315 by the
denier number.

 C.  Weaving

 The production of a flexible material  for
 fabric filtration involves weaving. Most
 felts used in filtration are first woven  and
 then given further treatment.   Woven fabrics
 are formed by interlacing yarns at right
 angles on a loom, after which  the fabric  may
 be further treated.   The most  common patterns
 of interlacing for fabrics used for gaseous
 filtration are known as twill  and sateen  or
 satin.  Plain weave fabric is  sometimes used.

 The twill weave may  be recognized by the  diag-
 onal pattern formed  by the filling yarn inter-
 lacing more than one warp yarn.   The values
 of twill weave include its strength and drap-
 ability.   The diagonally arranged interlacings
 provide greater pliability and resilience than
 the plain weave.   Twill weaves are frequently
 tightly woven and will not get dirty or
 blinded as quickly as plain weave,  though
 twills are more difficult to clean when they
 do get soiled.   Twill weaves are used  where
 strong construction  is essential.

 Satin weave is  similar to twill  weave  in
 construction except  that a satin weave requires
 five to twelve  harnesses in construction
 while a twill  weave  requires no  more than
 four.   Satin weave differs in  appearance  from
 the twill because the diagonal of the  satin
 weave is  not visible.   Since more harnesses
 are required for  satin weave,  a  greater
 amount of fine  yarn  may be compressed  into
 a  given space  of  cloth.   This  compactness
 gives the fabric  more body as  well as  less
 porosity.
Plain weave is the simplest type of  con-
struction and is consequently most inexpen-
sive to produce.  On the loom the plain
weave requires only two harnesses as  each
filling yarn alternates over and under the
warp yarns.  If the yarns are close  together,
the plain weave has a high thread count and
the fabric is firm and will wear well.

Felts used in fabric filtration are also
woven in their early stages, but subsequent
steps change the character of the material
from a woven fabric.  Using a woven base
fabric called a scrim increases the strength
and stability of the fabric over a matted
felt fabric.  Woolen felt is produced by
mechanically working a woolen scrim in warm
water in the presence of certain lubricants'
and chemicals in order to shrink the material.
Napping the surface of a material produces
a felt-like fabric.  Needle punching  is a
method of combining two or more layers of
fabric into a felt-like fabric.  Usually one
layer is a scrim for strength, while  the
others may consist of fibers of almost any
description or combination.  In this way con-
siderable control over material properties is
possible.  Non-woven production methods
include resin bonding, wet bonding (paper-
like materials), spun bonding, heat bonding,
chemical bonding, spray bonding, and  stitch
bonding.  Nearly all fibers used in fabric
filtration can be used in non-woven fabrics
and non-woven fabrics can be produced more
rapidly than by weaving.

D.  Fabric Treatments

Dimensional stability is an important factor
in filter fabrics.  Cotton and wool  fabrics
must be preshrunk and synthetics are  given a
corresponding treatment called heat-setting.
This process contributes to a more even bal-
ence of warp and filling yarn tension, pro-
vides better surface smoothness, reduces yarn
slippage, controls porosity, and eliminates
shrinkage.  Man-made fibers frequently con-
tain one or more of the following additives:

  - plasticizers to reduce flow viscosity
   and improve low temperature flexibility,

  - solvents used in wet spinning and  in
   coating,

  - organic peroxides used as poly-
   merization initiators,

  - antioxidants  added  to  reduce
   oxidative deterioration during
   manufacture, processing, and  storage.

-------
2.0
                                          A.   1-12  Fiberglas  fabric,  low  fiber surface
                                              area  per square  foot  of cloth.

                                          B.   Napped  B-27  Orion, medium fiber surface
                                              area.

                                          C.   B-26  staple  Orion fabric napped both sides,
                                              high  fiber surface area.
                                 Filtered  Dust Mass, grains/ft'


                      Figure  1.   EFFECT OF FABRIC ON BASIC PERFORMANCE CURVE
 2. or
                                                         High  twist  unnapped Orion,  low
                                                         fiber surface  area per  square
                                                         foot  of cloth.
                                                         Flberstock  Orion, high  fiber
                                                         surface  area.
                                   Filtered Dust Mass, grains/ft'

                         Figure  3.   EFFECT OF FABRIC ON BASIC PERFORMANCE CURVE

-------
Other agents are added to fabrics such as
flow control agents, colorants,  flame retard-
ants, stabilizers, ultraviolet absorbers, and
antistatic agents.  Yarns are treated as
well as fibers.  Yarn treatments include
surface addition of lubricants,  antistatic
agents, and other mechanical operations.

Fabric finishing includes processes to improve
appearance or serviceability of  the fabric.
Cotton and wool fabrics are usually cleaned  and
bleached and are sometimes waterproofed and
treated to protect against mildew and fire.
Synthetic fabrics are sometimes  treated with
water repellants and antistatic  agents.  Glass
 fabrics  are usually treated with silicones,
 graphite,  and  other proprietary finishes to
 reduce fiber-fiber  abrasion during filter
 cleaning.

 E.   Bag  Construction

 The most common  shape of  filter elements used
 is  a simple, circular cross—section  tube.
 Commercial bags  are usually 5 or 6 inches in
 diameter and from 5 to  30  feet long.   Bag
 widths vary from the usual 5 or 6  inches to
 20  inches,  determined mainly by the  width of
 the material used to make  the bags.   There is
 no  standard length  to diameter ratio,  as,
 from a theoretical  point  of view,  the  length
 to  diameter ratio has no  effect on efficiency
 of  collection  of the bag.  The factor  limit-
 ing the  length of filter  bags is the  cleaning
 requirement.   Excessively long bags  are more
 difficult  to clean  and  create problems by
 rubbing  against  each other.  A practical limit
 for the  length to diameter ratio is  about 20
 to  1.

 The multiple tube bag is  a tube-type bag of
 oval cross  section  with vertical stitching
 that effectively divides  the bag into  cir-
 cular tubes.   The multiple tube bag  has  the
 advantage  of greater filtering area  for a
 given floor  space and helps break  its  own
 filter cake  when the blower is turned  off and
 the bag  returns  to  its  oval shape.   The bag
 requires a  special  mounting and is more expen-
 sive for the same filtering area than  a tub-
 ular bag.

 Envelope type  bags  are  nearly as common as
 tubular-type bags.   Filtering elements are
 flat panels of cloth stretched over  a frame.
 The panels  are usually  in.pairs.   Wear is
 increased  because of friction between the
 filter cloth and the wire frame  support.  One
 advantage  associated with envelope bags is
 that a greater filtering  area may  be installed
 in  a given size  volume  than for other designs.
Filter bags are available commercially over
a broad range of dimensions.  The  fabric  sur-
face per compartment required may  be  deter-
mined from information about the allowable
variation in gas flow with respect to process
ventilation, the availability of sizes of
commercial compartments or houses,  and the
expected frequency of maintenance.  Many
combinations of filter length, diameter, and
spacing are available in order to  obtain the
least expensive baghouse.  Maximum filter
packing and compact filter housing do not
necessarily give lowest cost baghouse.

Closely packed filter elements tend to wear
against one another, and make inspection and
maintenance difficult.  Taller units  give
lower cost per floor space than compact ones.

III.  BAGHOUSE DESIGN

A.  Cleaning Processes

As previously mentioned accumulated dust  tends
to increase the pressure loss through a fabric
filter until a desirable maximum value is
reached.  The filter must then be  cleaned.
Most of the development toward fabric filter
equipment has gone into improved methods of
removing the accumulated filter cake  from the
fabric.  As a result, a great variety of
cleaning mechanisms are available.  Four
objectives of cleaning, bags are: 1) to remove
the desired amount of deposit from the fabric
quickly; 2) leave enough residual  deposit to
improve collection efficiency at start-up for
woven fabrics; 3) avoid damaging cloth or
using too much power, either of which can be
a substantial part of operating costs; and
4) avoid excessive dispersal of removed dust
so that this dust would not have to  be refil-
tered.  There are two general types of clean-
ing; the first involves flexitig the fabric
and the second involves a reverse-flow of
clean air.

Mechanical shakers are the most common type
of baghouse cleaning equipment.  Electric
motors are used to oscillate the tops of  the
bags either vertically or horizontally.   This
is usually done under slight negative pressure
inside the bag so that more effective cleaning
may be accomplished.


Sonic cleaning utilizes sympathetic vibrations of
low frequency sound waves to vibrate  the  bag
frame work.  This method Is not successful
with difficult to remove filter cakes since
the total amount of energy imparted to  the
bags is low.

-------
                incoming gases
                    Filtering
Filtering
Fi Itering
                         Incoming gases
               to fan                                  to fan
                All compartments filtering, dampers open     One compartment shaking, balance  filtering
                 incoming gases
                        incoming gases


}
/N
Filtering
\/
_ A
*
)•

Shaking

*
<
Fill
s
- <«-*
1
/ s
er
1>
+
'J

>
ing
^

                            /N
                            Filtering

                            ^
                               t
                            Filtering
                                                                              Shaking
                                                                               y
               to fan                                to fan
               One compartment shaking, balance filtering   One compartment shaking, balance filtering
                          Figure 4.    TYPICAL PARALLEL FLOW SYSTEM FOR A
                           CONVENTIONAL MULTICOMPARTMENT BAGHOUSE
Bag collapse  Is  frequently used to improve
cleaning efficiency by permitting a small
flow of air to flow in the reverse direction
causing the bags  to collapse.  The process
may be repeated  several times for each clean-
ing.  The bag collapse method is often used
in conjunction with the other cleaning methods
to improve efficiency.

A variation of the  bag collapse method is the
pulse-jet cleaning.   For this method a "bub-
ble" of compressed  air is injected at the top
of the bag when  the bag is collapsed.  As the
Pulse of air  moves  down the bag the  filter
cake is flexed  and  the collected material
falls through the bag.

The reverse-jet  mechanism is an example  of
reverse-air cleaning.  A high velocity jet
of compressed air is blown back through  the
fabric to dislodge  the collected dust.   This
method is sometimes too efficient and does
                        not leave sufficient  residual for cleaning
                        at start-up.   In  a  typical reverse-jet filter
                        unit, cleaning may  be conducted continuously
                        in order to maintain  a constant pressure
                        differential across the unit.

                        Reverse air flow  at low or atmospheric pres-
                        sures is also  used  to clean baghouse filters.
                        This method is often  used for envelope-type
                        baghouses.  A  large volume of air is required
                        for this method which often requires a sepa-
                        rate air blower for cleaning.  Supporting
                        rings are occasionally used inside filter
                        bags so that they may maintain their shape
                        during cleaning.

                        B.  Baghouse Construction

                        The location of the blower on a baghouae deter-
                        mines the type of unit.   If the blower is
                        located on the clean  air side,  the baghouse
                        is referred to as a pullthrough baghouse.  In

-------
                        (
                                                   \
                    ^
                    ./
                    \^
                                                                             VJ
                (a) Bottom Feed                (b) Top Feed             (c) Exterior Filtration

                              Figure 5.   POSSIBLE FILTERING SYSTEMS
tliis position the blower is protected from
the dust or fume being handled.  However, a
relatively airtight housing is required of
the housing.  A pushthrough baghouse has the
blower located on the dirty air side, and the
sides of the housing may be left open.  This
configuration is often used for hot gases as
.1 greater degree of cooling may be obtained.
The dust loading that is handled by a push-
thi<>ii|;li b lower often causes substantial wear
.md frequent maintenance problems for the
1) lower.

The structure of the housing for a baghouse
must be able to withstand a pressure differ-
entia] of 8 inches of water or more.  This
c.lls for heavy gage metal and bracing of
wails for the housing and the hoppers.  Pull-
through baghouses are generally more of a
structural problem than pushthrough baghouses
as baghouse structures can withstand internal
pressure more easily than external pressure.

Bottom-feed baghouses are the most common
configuration used in industry.  Dirty air
enters the bottom of the bags and filtered
through the bag from the inside out.  The
clean air is on the outside of the bag.  A
tup-feed bag is similar to the bottom-feed
except the dirty air enters  the  top  of  the
bag.  Because of gravitational effects  a  top-
feed baghouse seems the more logical choice;
however, mechanical problems of  securing  bags
at both the top and bottom for top-feed bag-
houses makes the bottom-feed type easier  to
construct and, thus, the more common.   One
other possible configuration is  the  exterior
filtration type.  Dirty air  is brought  through
the bag from the outside of  the  bag  to  the
inside.  This type of arrangement requires an
inner-bag support structure  to keep  the inside
of the bag open.

Hopper size is dictated by the frequency  of
pick-up and disposal of collected material.
If the hopper does not have  adequate capacity,
dust already collected becomes reentrained
and increases the total dust load on the  filter
cloth.  This increases filter resistance  and
the performance of the baghouse  is affected.
Mechanical rappers or vibrators  are  sometimes
provided on hoppers to assist the collected
dust to flow freely from the discharge  gate.
Materials that tend to stick or  cake in the
hopper may be moved by rapping.

C.  Maintenance
10

-------
The hopper of a baghouse should be emptied at
least once a day.  Inspections of the equip-
ments should be conducted regularly at inter-
vals of a week, month, or quarterly, depending
upon the nature of the dust collected, the
quantity of dust, and the general severity of
service.  Moving parts must be greased and
serviced.  All bags should be examined once
a week to determine if any show wear.  Ripped
bags should be replaced immediately.

l!ag failures generally occur between one and
two years after installation.  After one year
frequent replacements are required meaning
that the baghouse equipment is down a great
deal of the possible operating time.  Many
baghouse operators replace all the bags in
a baghouse periodically before serious pro-
blems develop.  In this way bag replacement
may be scheduled to coincide with plant shut-
down periods.

Replacement of one or several bags in a large
baghouse is sometimes unavoidable.  The resis-
tance of a new bag in a baghouse during start-
up is very low and, as a result, the filtering
velocity through the new bag is many times in
excess of the normal rate.  This could result
in blinding of the new bag.  Blinding is a
plugging of the fabric pores to such an extent
that the resistance becomes excessively high
permanently.   One  solution  to  the problem  of
high filtering velocities  is  to precoat  the
bags with dust to  establish  a  dust  cake
immediately after  installation.  Some  author-
ities  require  all  bags  to  be  precoated after
each cleaning  cycle.

Li.   Uust Disposal

The  most common  means of disposing  of  dust
i.'i I Lee ted by ,1 bnghouse  is  to  transfer it
I  rum the hopper  into  a  truck  and take  it  to
,i dump.  Much  ol  the  dust  dropped from a hop-
per  to .1 truck wil.l escape  to  the air  if not
h.indlL'il propri iy,  defeating  I lie purpose  of
Lin-  bnghouse.  A sJeeve  of  canvas is  frequently
i njt.-ii led on the outlet  of  the hopper  to
eliminate this problem.  After the  dust  is in
Llie  truck it is  usually  wetted to keep it  in
UK  truck during the  trip  to  the dump.   This
l>nuedure is suitable for  installations where
i  lie dust is collected once  a day.  When  more
frequent collections are made, automatic or
seminutomatic methods are recommended.   Con-
veyors may be used to collect  the dust from
'-'ever.ii hoppers and discharge  the dust into
i covered tote box.

IV.  iNltUSTKlAL  APPLICATIONS
The collection of dust from rotary cement
kilns has been a difficult problem.  Large
volumes of gas are handled with high concen-
trations of very fine particles, high gas
temperatures, and for wet-process kilns, the
presence of large amounts of water vapor.
Often conventional cyclones are used to col-
lect the large particles in front of the
final cleanup collector.  Besides baghouses,
electrostatic precipitators are the only
devices available for cleaning the gases from
cement kilns.  Fabric filter applications
have obtained efficiencies of 99.5 percent,
outlet loading below 0.02 grains per standard
cubic foot, and plume opacities less than
10 percent.

B.  Foundry Cupolas

Exhaust temperatures from a grey iron cupola
range from 1000°F to 7500°F.  Effluent load-
ings are about 1.0 grain per cubic foot with
much of the emission in the form of a fine
metal oxide fume less than 0.5 micron in diam-
eter.  Gas cooling and high temperature fabric
filters are required.  Using evaporative cool-
ing off-gas temperatures are reduced to about
450°F before filtration through fiberglass
bags.  Typical filtering velocities are about
2.5 feet per minute and bags are about 11-j
inches in diameter and about 15^ feet long.

C.  Steel Furnaces

The exhaust  from the electric arc steel fur-
nace is characterized by a high percentage of
oxides of iron, highly variable gas tempera-
ture, variable dust  loadings, and highly
variable gas volumes during different process
cycles.  The use of  dilution air to provide
for gas cooling and  In-plant dust control
causes changes in the volume of gas to be
cleaned.  Stack temperatures may reach 750°F
and higher with closed hooded units.  Fiber-
glass bags are used  with air-to-cloth ratios
of about 1.5.  Orion bags are also used at
gas  temperatures of  around 200°F.  Orion bags
have about a five-year life.

Iron oxide fumes from oxygen lanced open
hearth furnaces may  be collected efficiently
by fiberglass bags.  Temperatures range  to
about 500°F  and the  filtering velocity is
about 2 ft/rain.  Reverse air flex cleaning
along with sonic horns are used.  Inlet  load-
ings may be  as high  as 20 grains per  cubic
foot and outlet loadings may be as low as
0.007 grains per cubic foot.

D.  Nonferrous Metal Furnaces

Fiberglass haghouses have been  applied  to
                                                                                                 11

-------
secondary lead smelters for fume collection
at temperatures higher than 400°F.   The high
temperatures eliminate the deposition of
organic tars on the bags.   Satisfactory
results have been obtained with filtering
velocities of about 1.2 feet per minute and
cleaning by shaking.   Fiberglass bags have
also been applied to primary copper and zinc
smelters with some success.

E.  Carbon Black Plants

Nearly all of the carbon black plants in the
United States are equipped with fiberglass
baghouses.  Gas temperatures range  from about
400°F to 500°F.  Gentle cleaning techniques
are used such as bag collapse and sonic horns.
Filtering velocity is about 1.5 feet per min-
ute and the average life is about a year.  The
high temperature of the gas is useful for
keeping the acid mist in the effluent above
the dewpoint.

 F.   Grain Handling Operations

 Dust emissions come  from  several different
 sources in grain handling operations:  clean-
 ing, rolling,  grinding, blending,  and  the
 loading of trucks, rail cars, and  ships.   Con-
 veying and storing grains also cause dust
 emissions.   Cyclones are  used to collect grain
 dusts larger than 10 microns.   Baghouses with
 mechanically shaken  woven cotton bags  remove
 99.9 percent of grain particles in the size
 range of  1 to  5 microns with air-to-cloth
 ratios of about 5 to 1.   Felted bags with
 reverse-jet cleaning may  handle air-to-cloth
 ratios up to 15 to 1.

 G.   Other Applications

 One full  scale bag filterhouse has been
 installed to filter  the entire exhaust efflu-
 ent of a  utility boiler.   Fiberglass filter
 bags are  used  with alkaline additives  to
 remove essentially all of the submicron parti-
 culate. matter  and a  large percentage of the
 sulfur trioxide.   Hie boiler in this case is
 oil-fired.
Baghouses have been  applied  in special cases
such as tobacco  steaming  processes,  asbestos
recovery in brake  lining  manufacture,  wood-
working facilities,  grinding wheel manufacture,
flour mills, and other  processes  that  may
include pulverizing,  grinding,  conveying, or
drying.  Small particles  and fumes from
industrial practices  may  be  collected  in fabric
filters.
REFERENCES

1.  Billings, C.E., et al, Handbook of Fabric
    Filter Technology.  Volume 1, Fabric
    Filter Systems Study, National Technical
    Information Service, Springfield, Va.,
    December, 1970, PB-200-648.

2.  U.S. Department of Health, Education, and
    Welfare, Control Techniques for Particu-
    late Air Pollutants, National Air Pollu-
    tion Control Administration Publication
    No. AP-51, 1970.

3.  U.S. Department of Health, Education, and
    Welfare.  Air Pollution Engineering Manual,
    by J.  A. Danielson, Public Health Service,
    Washington, D.C.: Government Printing
    Office, Pub. No. 999-AP-40.

4.  Sommerland, R.E.,  ''Baghouse Filters as
    Applied  to Power Plant Effluents," Heat
    and Fluid Dynamics Department, John
    Blizard  Research Laboratory, Foster
    Wheeler  Corporation, Carteret, N.J.,
    May, 1966.
12

-------
                                  Table 4.    APPROXIMATE  SIZE  RANGES  FOR FABRIC COLLECTORS

Collector
Reverse- Jet
Pressure-Jet^
Conventional
tubular bags
Mechanic*!
Reverse
flov
Envelope

(fpm)
1.0
1.0


1.0

1.0
1.0
Collector volume
per 1,000 cfm
(ft3)
1,250
670


210 - 370

590
210 - 340
Collector
Floor-Area
per 1,000 cfn
(ft2)
57 - 294
111


26-50

30-42
21-59

(fpn)
10
10


3

2
2
Collector volume
per 1,000 cfm
125
67


70 - 123

295
105 - 170
Collector
Floor-Area
per 1,000 cfo
(ft2)
5.7 - 29.4
11.1


8.7 - 16.9

15 - 21
10.5 - 29.5
(1)    Does not Include dust hopper.




(2)    Comon values for filter velocity.
C3)   As manufactured by Fulverizing-Machinery  Company, N.  J.

-------
                                           25
SECTION 25
Fabric  Filtration-Mathematics
  of Bag-Filter Operation

-------
IV.   FABRIC FILTRATION - MATHEMATICS OF BAG-FILTER
      OPERATION

      As  one moves  from concepts of filtration related to a single
      elemental area,  to filtration on commercial size filter tubes,
      to a parallel arrangement of tubes in a compartment, to the
      conventional multi-section baghouse, filtration becomes
      progressively  more dynamic and non-uniform.  Velocities of
      filtration are seen to be extremely variable and certainly not
      constant at  any particular point for a filtration period.  Meaningful
      design data, therefore, can only be gathered from actual operating
      installations or well designed pilot plants.  This data should include
      basic filtration parameters,  so that proper adjustments can be made
      to other operating conditions and other baghouse designs.  It is the
      purpose of this section to explore the basic mathematics of multi-
      compartmental operation in order to delineate these basic parameters
      and to evolve those equations necessary for baghouse design.  The   , ,  ,.
      materials contained in this section are based on the works of Stephan,
      Sargent,'    ' and Walsh,''*- ^) as modified and extended by the author.

      A.   Basic Equations and Assumptions

           As explained in previous sections, it is generally agreed that dust
           fabric combinations behave as laminar flow elements at any
           instant  in  time; that is:

                     =  constant = S                                (4. 1 )


           With certain limits, and depending on the nature of fabric and
           dust, it would also appear valid to consider that changes in
           filter drag (AS) are directly proportional to changes in the
           weight of dust collected per unit fabric area  (AW); that is:

               AS =  (constant) X  (  W ) =  ^    AW                  (4.2)


           For an  elemental  filter area,  then,  the following equation can
          be used to describe a Basic Performance Curve:


               ST  = SR  +                                         (4'3a)
                          U,  C  • t
          or  S   = S  +  — -  2 -                             (4.3b)
          PA.C.pm.  92.5.66

-------
    Where S* is an extension of the linear portion of the  Basic
    Performance Curve to the zero intercept.

    For convenience, Figure 4. 1  has been reproduced to summarize
    these terms and definitions.   It is important to note that in
    calculating S,  synonymous measurement of A P and U, must be
    made unless it is known that Uf is constant throughout the
    filtration period.

    Throughout this  text,  filtration period is the time during which
    an elemental area is on-stream; that is,  the time required to
    go from Sj^ to  Sn-,. Unless otherwise noted, the term S-^ will
    be considered  identical to SR,  for convenience and ease of
    analysis.

B.  Parallel Flow  System,  Discontinuous  Operation

    In any practical  filter unit (i. e. , a single filter tube, a
    collection of tubes in a  single compartment or section, or a
    multi-compartmented unit) an analogy can be drawn between
    an electrical circuit and the filtration system.  This  analogy
    has been illustrated and described by Stephan for filtration
    through a single  tube.   The illustration used is shown as
    Figure 4. 2.

    For the electrical circuit,  Ohm's Law applied, as follows:


        1=^-                                              (4.4)
              e

    Where:

         I-I^Iu.IJ.         x1               /„  e»
        -75— -  p- + T5 - + T5—  +....+ p-             (4.5)
         Ke     Rl   R2     R3              Rn

    By substituting like terms, it  follows that:

                                                            (4.6)
             S
              e
    Wherein Sg has the dimensions (in. H^O/cfm).   In filtration,
    however, it is  advantageous to relate resistance (filter drag)
    to one  square foot of filter area,  so that equation (4. 6) is
    usually written as:
                                                             4.2

-------
E
a
O
 CVl

X
cr
a

a:
LU

b
                        TOTAL CYCLE
                 REPEATED TO ATTAIN EQUILIBRIUM
    A
INTERVAL OF

CAKE REPAIR


       *


     

   Projected


   Residual Dra
             B
DEPOSITION OF HOMOGENEOUS

      DUST  MASS
             AW



          DUST PERMEABILITY
                                          AW
TERMINAL


  DRAG
     Figure 4. 1
                  FILTERED DUST MASS,W(GRAINS/FT.2 )

           - SCHEMATIC  REPRESENTATION  OF BASIC

                    PERFORMANCE PARAMETERS,

-------
    S  =  **                                          (4.7)
     6
and the units of S£ become (in.  P

As in the electrical  circuit,  the effective filter drag (Se)
is related to local filter drag (Sj, S2,  S3>  .....  Sn) as
follows:

 1    1  (aj)     I(a2)    1  (a3)                1 (aj
s; = ZJKJ +  3-P9 + -5^39- + ...-+ 3^9    <4-8a>

If each  elemental area is the same, then equation (4. 8a) can
be simplified to:
 _  = _                  .   .  .  .
Se   S2    SZ     b3                  n
                                                       (4. 8b)


                                         Af
Where n = the number of equal areas, or  - .
                                          ct


Consider, as an example, a single filter tube as shown in
Figure 4.2.  After cleaning, the distribution of  local filter
drags will not necessarily be uniform over the entire bag
area.  Such a non-uniform situation has been shown previously;
a further example is shown as  curve S  , in Figure 4. 3.   For
                                     °» •!•
this situation the average velocity was 1. 5 fpm and  A P     was
                                                     o, l
0.30 in.  H2O.  Using equation 4. 1,  then,  the velocity through an
elemental area, at 50% altitude  (i.e., half-way up the bag) was
approximately


          AP     0. 30      . , ,                        IA  o  \
                                                       (4. 9a)
while the velocity at 90% altitude was approximately
    uf = nre-  = 5-oofp™                              (4.9b)
The above are only approximate values  since the filter drag of
the fabric was not included in S   ,.  This drag, plus the drag
of the dust would determine actual velocity profiles.

                                                       4.4

-------
                                           "5
                                              "2
Figure 4.2   ELECTRICAL ANALOGY FOR FILTER FLOW  SYSTEM
     In reference to Figure 4.3, it can be  seen that the greater
     air flow in the low resistance areas brought correspondingly
     greater amounts  of dust to the filter surface per unit of time.
     This,  in turn, increased filter drag at a greater rate,  leading
     to a "flattening" of the profile with time.  As a result,  a  plot
     of Se vs W would be  inherently non-linear, despite a constant
     dust permeability.   The extent of this non-linearity  would
     depend on the uniformity of conditions after cleaning.   These
     same  concepts  can be applied to each filter tube in a compartment,
     using  the area of one tube as  a basis.  Likewise, they can be
     applied from compartment to compartment in a multi-sectioned
     baghouse using the entire filter area of one compartment as a
     basic  area.  The mathematics of such a such a system has
     been developed by Sargent.

     For the  sake of discussion, consider  again the  analogy shown
     in Figure  4. 2, and assume dust permeability is a constant.
     The rate of change of filter drag can then be written as follows:
        dS
                                                            (4. lOa)
         dS-
                K
                                                            (4. lOb)
                                                              4. 5

-------
           0.03 Q04 0.060.080.1     0.2  0.3 0.4  O.6 0.8 1.0

                   DUST MASS RESISTANCE (IN. H20/fpm)
                                                  2.0  3.0 4.O
Figure 4. 3   DEVELOPMENT OF DUST MASS RESISTANCE

     PROFILE THROUGH A FILTRATION PERIOD
    From equation (4. 1),
AP =
                                 = etc.
                                                                (4.11)
    Combining equations (4. 10) and (4. 11),
    or
  1  K
dS
                       dS
                          2  K

                              P
            dS
            _ ~"
            ~ar
                                                      (4. 12)
                                dS
                                _ f.
                                ~ar
                          = etc.
(4.13)
    Integrating equation (4. 13) from time t = o to time t = t,


         C2    2   C2    2   C2   2   .
         S  - s  = S  - s  = S   - s   etc.
                                                       (4.14)
                                                                 4.6

-------
Where  s.,  s?,  s^,  refer to filter drag values at time t = o.
Now if  the volumetric flow is Q, and the total area is A,,
then

     Q = A  (U   +U   + n   + .  .   .  .  + U,)           (4.15)
          1    1    12    *3                 n
Combining  equations  (4. 10) and (4. 15),

         KA   /dS    dS   dS                dS \
     Q = ^-    ( ,    +  .  + -,— +    .  .    +  .. n )       (4. I6a)
          p   \ t                               J

or,
     QC     U,.     ,-C:      dS.    dS9   dS7           dS
       P    f(avg. )   P   _    1,2,3
          =      K	--+- +
Integrating equation (4. 16) from time t = o to t = t


     Q- C (t)     n - n      n = n
                 n = 1       n = 1

In summary,  equation (4. 17)  says that the total of all drag changes
that occur during a filtration  period (i.e., ^ S    ifs ) can be
calculated on the basis of inlet conditions (i. e. ,
The total of all drag changes,  however, is not the  same as the
change  in effective  drag, A&e, since the sub-changes are
arranged in parallel.  This is  true if we consider a single filter
tube or several sections in a multi-section baghouse.  From this
we can  conclude that control over baghouse  AP will depend on the
distribution of filter drag as well as the total of all drag changes.
                                                          4. 7

-------
    As a check on equation (4. 17) assume that at time t = o all
    values of s = SR.  Then,  at time t = t all filter drag values
    will equal S-,  since filtration would have been at a uniform rate.
                = n(ST) -n(Sp)                             (4. 18a)
            f        i      *.

    or, rearranging terms,

        s,    .   -    r     -Vvs.)^-'
                                                           (4. 18c)


    which is identical to equation (4. 3a)


C.  Continuous Operation

    In Part A and B we considered basic equations and assumptions
    for a single incremental area and for a parallel arrangement
    of areas, respectively.  For section B it was necessary to assume
    some undefined values of filter  drag  at the beginning of filtration.
    In this section the equations will be extended to the continuous-
    operating, multi-section baghouse.  The same assumption of
    constant K will be maintained.

    For the multi-compartment baghouse, two operating schemes are
    possible, depending on the  number of compartments, filtration
    time, cleaning time,  and individual preference.  These are as
    follows:

    1.  One compartment maintained on a stand-by,  so that when a
        dirty section is taken off-stream for cleaning, a cleaned
        unit is placed on-stream.   The total number of compartments
        equal n + 1,  with n compartments always in service.

    2.  In most applications, all compartments are on-stream for
        some finite portion of the first part of a cycle; one com-
        partment is then taken off-stream,  cleaned,  and put back
                                                           4.8

-------
             in service, such that one less than the total number of
             compartments are filtering at the end of a cycle.  Actual
             performance will be similar to  (1),  provided cleaning time
             is a small fraction of filtration  time*.   The following
             treatment is based on operation (1); it is easily adopted to
             (Z) for many practical situations.

             As a beginning,  consider a baghouse with n + 1 = 5. Assume
             that each compartment  is cleaned to the same degree,  and
             that filter drag after  cleaning is equal for all sections  and
             is denoted as SR .  An electrical analogy for the 4 sections
             on-stream is shown  in  Figure 4.4.  Because of the cyclical
             nature of the operation  (i. e. , cleaning each section in
             sequence and to the same degree), a pattern of filter drag
             values will exist after several cycles of operation.
             Table  4. 1 illustrates the nature of the pattern.
      H
      co
      »— i
      CO
      W
      W
                                                            E ~ AP

R2~S2
R3~S3
R4
\
f
Q  =
                                              U
      Figure 4. 4   ELECTRICAL ANALOGY FOR FOUR COMPARTMENT
                                BAGHOUSE
* NOT El  For a multicompartment baghouse, the following terms will
be used:
     1.   Filtration time  (t)  - time required for a compartment to go
         from Sp to S™.
     Z.   Cleaning time - time a given compartment is off-stream for
         cleaning.
     3.   Cycle time (tc)  - time between cleanings of two  consecutive
         compartments.
                                                                  4. 9

-------
      Table 4.1   PATTERN OF FILTER DRAG VALUES IN
MULTI-COMPARTMENT BAGHOUSE AT EQUILIBRIUM. CONDITIONS

Cycle
1
20

7 1


22


23

24

25

Start
Finish

Compartment
1 I 2
Off- R
stream
Start SR

Finish . SA
Start SA

Finish S_
D
Start S^
JD

Finish , S_
Start S-.
1
Finish
Start
Finish
ST
Off-
stream
SA
SA

SB
SB

sc
sc

ST
Off-
stream
SR
SA
3 4
SA SB
SB
SB

sc
sc

ST
Off-

stream
SR
SA
SA
SB
sc
sc

ST
Off-

stream
SR

SA
SA
SB
SB
5
sc
ST
Off-

stream s
SR

SA
SA

SB
SB
sc
sc
SC ST
     = Residual filter drag = Drag after cleaning.

     - Terminal filter drag = Drag before  cleaning.

    , Sg, S_ = Intermediate drag values.
                                                            4. 10

-------
For cycle 21, and substituting in equation (4. 14),
             =  =   = 
Solving simultaneous equations:




     s2   s2 -  s2   s2
     SA - SR -  SA - SR



     s2   s2 -  s2   s2
     SA - SR ~  SB " SA



     s2   s2 -  s2   s2
     SA - SR -  SC - SB



     s2   s2 -  s2   s2
     SA - SR -  ST " SC
or,  4S^ - 4SR  =  S2  - S2                             (4. 20)







Rearranging,






      S"   ,4 r< "   o (T1                                     /^oi\
        =  4S.  -  3S                                     (4. 21)
      JL    2~\.    XV






For  a baghouse  with n compartments on-stream at all times


(total of n = 1  compartments with one always idle)


equation (4. 21) can be generalized to:







     S2 =  nSJ;   (n-l)S^                                (4.22a)





or;	


                       2
    ST =  N|nSA   (n-l)SR                              (4.22b)







Where S.  - filter  drag at end of cycle for most recently

cleaned compartment (i.e., compartment 2, cycle 20;


compartment 1, cycle 21; compartment 5,  cycle 22).
                                                      4. 11

-------
             2
Dividing by SR, and rearranging terms,


                   29
            SIC  /C \  _L /-r,  1 \
      A     (b „,/£>.£.)  T (n-JL;
         =  _Z	5^	                            (4. 23a)
                   n
In summary,  equation (4.23) allows for the calculation of filter
drag at the end of one cycle of operation for the compartment
just placed in service,  provided values for  S™,  Sp, and n are
known.   S.  is not a terminal drag.  It is an intermediate value
on the Basic Performance Curve for the compartment in question.
In like manner it can be shown that:
                                                       (4.23b)
                                                       (4.23c)
We can generalize  equations (4.23a, b, c) by letting x be the
designation of any compartment in a baghouse, such that

    x = 1  for most recently cleaned compartment

    x = n  for compartment at terminal conditions.

   Sx = Drag of any compartment at end of cycle.

Accordingly,
Repetitive solutions of equation (4.24) will give the pattern
of filter drags for each compartment at the end of a cycle. *""
Substitution of these values in equation (4. 8b) will then give S  .
                                                        4. 12

-------
The use of equation (4. 7) will give   P at the end of the cycle.
Further manipulation of equation (4. 8b) and (4. 7) will give
  P at the  start of the cycle.

By way of example, consider a baghouse under the following
conditions  of operation:

    U,,     .  = 3.0 fpm with n compartments on- stream

    S         =4.0 in.  H2O/fpm

    S_       - 0. 5 in.  H,O/fpm
    n
then,
     (S  /S )* + (n-1)     6
     — T  R - =     L     =11-5                (4.25)
           no                           v     '


        x = 1 (11. 5) - 0 = 11. 5          S1               (4.26a)
                                  and -—  =3.4
         = 2  (11.5) - 1  = 22             Q                (4. 26b)
                                       O~
                                  and  *  = 4. 7

        = 3 (11.5) - 2 = 32.5          S                (4. 26c)
                                  and _±- =  5.7
      - )= 4 (11.5) - 3 = 43            S                (4.26d)
      R/                         and -3 = 6. 6
                                     S
   /•s
   I -iN = 5 (11.5) - 4 = 53.5        S                 (4.26e)
     R/                         and    = 7.3
                                                       4. 13

-------
      i\ = 6  (11.5) - 5 = 64            S6               (4.26f)
      R/                         and     =8
                                       R

Using equation (4. 8b),
      eT
                                                      <4' 27a>
and


    S
      R  _ .294 + .213 + .175 + . 151 + .137 + . 125   1.095
      --- ----- -
      e
       T
        Se   =       = 2. 74 in. H0/fpm                (4. 27c)
Consequently, from equation (4. 7),
       P = (3.0 fpm)(2.74 in.H2O/fpm)                 (4.28)

         = 8. 22 in.H2O


At the start of a cycle we have physically substituted a
clean compartment for the dirtiest section,  so that in

                  S6                     SR
equation (4. 27a),  -5— has been deleted and -p— added.
                  SR                     SR

Thus,  equation (4. 27a) becomes,

    6S

          tt + Trr + Trr + iT5 + Tr3 + T!+  T 4     (4-29a)
       R
and .'. ,   Se   =         =  1'52 in>  H°/fPm           (4. 29b)
                                                      4. 14

-------
    Also,

            P = (3. 0)(1. 52) = 4. 56 in.  HO                  (4. 30)
    In summary,  the  six -compartment baghouse would cycle
    between 4.56 in.  H?O and 8.22 in. H?O,  assuming constant
    flow.  Flow variations could be taken into account provided
    the inlet dust mass flow  (grains/min. ) was constant.
D_  Approximate Solutions for Equation (4. 24)

    While  repetitive solution of equation (4. 24) will provide an
    exact solution to the problem of predicting maximum and
    minimum Ap's  of a multi- compartment baghouse, its use
    is cumbersone and time consuming.   Therefore,  a less
    involved technique is desirable, and is derived as follows:
         In equation (4. 24),  let

                 (S /S  )2 + (n-1)  - 1
             Z= — - - £— --                       (4.31)
         Equation (4.24) then reduces to:

            'S
              X ' = xZ  + 1                                   (4. 3 la)
    or,

        S
          X
=  JxZ + 1                                     (4. 31b)
    From equation (4. 8b)

         c      n = n
        ni>p      ~L  S  /c
         -*=    =  VSx                                (4.32)
    Utilizing equations (4. 31b) and (4. 32), tabular computations
    can be run for selected values of Z, as shown in Tables 4.2
    through 4. 7.  For any value of n,  in these tables,  column 3
                                                           4. 15

-------
represents S_/SR while column 6 represents Se /S_.  When
ST/SR is plotted as a function of Se /SR (FigureT4. 5) a family
01 curves is produced.  Each of these curves (excepting n = 1)
is slightly non-linear for the lower values of S^/S^ „   As can
be seen, the general slope of these curves tena toward  some
limiting values, with the most  significant changes  in general
slope occurring as n increases from 1 to 6.

If desired, a family of curves similar to those shown in
Figure 4. 5 could be used in place of equation 4. 24, with some
slight loss in accuracy.  To facilitate its use, a larger  chart
could certainly be prepared. An alternate procedure is
possible in situations wherein rapid calculation is  of prime
importance; the basis for this approach  is to approximate the
relationship between S—/S.- and Se.p/SR for  each value of n by
a straight line.  By way ofexample, we have chosen to  draw
this line through the  point where each curve  intersects ST/SR= 5. 0
The slope of each line can then be calculated (designated as  m
below) and an approximate relation between  SeT,/Sp and S^/Sp
developed, so that                                      IK

    SeT=SR+mAS                                  (4.33a)
                  (C  -Urt)
         = SR + m -^ -                            (4- 33b)
Values for m are listed in Table 4. 8.
                                                      4.16

-------
Table 4. 2.  SOLUTION TO EQUATIONS (4. 31b) and (4, 32)
                            for
                           Z  = 0. 5
1
X
1
2
0
4
5
6
7
8
9
10
11
12
13
14
15
2
xZ + 1
1. 5
2. 0
2. 5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7. 0
7 5
8.0
8.5
3

JxZ + 1
or
VSR
1.225
1. 414
1. 581
1. 732
1. 871
2.000
2. 121
2.236
2. 345
2.449
2. 550
2.646
2. 739
2. 828
2. 916
4
STJ/S
R' x
0. 816
0. 707
0. 633
0. 577
0. 534
0. 500
0. 471
0. 447
0.426
0.408
0. 392
0. 378
0. 365
: 0.354
0. 343
5
2 SD/S
R7 x
or
" SR/SeT
0. 816
1. 523
2. 156
2.733
3.267
3. 767
4.238
4.685
5. Ill
5. 519
5.911
6.289
6. 654
7. 008
7. 351
6
x i
^ sD/se
R' e-j'
or
VSR
1.225
1. 313
1. 391
1.464
1.530
1. 593
1.652
1.737
1. 761
1.812
1.861
1. 908
1.954
1.998
2.040
                                                             4. 17

-------
Table 4. 3.  SOLUTION TO EQUATIONS (4. 31b) and (4. 32)
                            for
                           Z =  1.0
1
X
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
2
xZ + 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
3

x(xZ + 1
or
Sx/SR
1.414
1.732
2.000
2.236
2.449
2.646
2. 828
3.000
3.162
3.317
3.464
3.605
3.742
3.873
4. 000
4. 123
4
VSx
.707
. 577
.500
.447
.408
.378
.354
. 333
.316
.301
.287
.277
.267
.258
.250
.242
5
SSR/Sx
or
HSR/SeT
.707
1.284
1.784
2.231
2.639
3.017
3.371
3.704
4.020
4.321
4.608
4.885
5. 152
5.410
5.660
5.902
6
X
nsR/seT
or
SeT/SR
1.414
1.558
1.682
1.793
1.895
1.989
2.076
2.160 !
2.239
2.314
2.387
2.456
2.523
2.588
2.650
2.711
                                                          4.18

-------
Table 4. 4.  SOLUTION TO EQUATIONS (4. 31b) and (4. 32)
                           for
                          Z = 2
1
:

X
1
2
; 3
I
4
5
6
7
8
9
10
11
12
13
14
15
16
2


xZ + 1
3
5
7

9
11
13
15
17
19
21
23
25
27
29
31
33
3

JxZ + 1
\l ./V*— I 1 X
or
Sx/SR
1.732
2. 236
2. 646

3.000
3. 317
3. 605
3.873
4. 123
4. 359
4. 582
4. 796
5. 000
5. 196
5.385
5. 568
5. 744
4


SR/Sx
. 577
.447
. 378

. 333
. 301
. 277
.258
. 242
.229
. 218
. 208
. 200
.192
. 186
. 180
. 174
5

S S /S
•" ~a/ v
Jtx X
or
. 577
1.024
1.402

1. 735
2. 036
2. 313
2. 571
2.813
3.042
3.260
3. 540
3.740
3. 932
4. 118
4.298
4. 472
1
6
i

X
or
seT/sR
1.732
1.953
2. 140

2.305
2.456
2.594
2.723
2.844
2.958
3.067
3. 107
3.208
3.306
3.400
3.490
3.578
                                                           4. 19

-------
Table 4. 5.  SOLUTION TO EQUATIONS (4. 31b) and (4. 32)
                           for
                          Z = 4
1

X


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
2

xZ + 1


5
9
13
17
21
25
29
33
37
41
45
49
53
57
61
65
3

JxZ + 1
or

Sx/SR
2.236
3.000
3.605
4. 123
4. 582
5.000
5. 385
5.744
6. 083
6.403
6.708
7. 000
7.280
7.550
7.810
8.062
4


O-|-| / S
K. x
.447
. 333
.277
.242
. 218
.200
. 186
. 174
. 164
.156
. 149
. 143
. 137
. 132
. 128
. 124
5
SSR/SX

or

.447
. 780
1.057
1.299
1. 517
1.717
1. 903
2.077
2.241
2.397
2. 546
2.689
2.826
2. 958
3. 086
3.210
6
X
IS /Se
K T
or
SeT/SR
2.236
2.564
2.838
3.079
3.296
3.494
3.678
3.852
4.016
4. 172
4.320
4.463
4.600
4.733
4.861
4.984
                                                          4.20

-------
Table 4. 6.  SOLUTION TO EQUATIONS (4, 31b) and (4. 32)
                            for
                           Z = 8
1
X
1
2
xZ + 1
9
2 17
3
4
25
33
5 ' 41
6 • 49
i
7 57
8
9
10
11
12
13
14
15
16
65
73
81
89
97
105
113
121
129
3

JxZ + 1
or
Sx/SR
3. 000
4. 123
5. 000
5.744
6.403
7. 000

7. 550
8. 062
8. 544
9. 000
9.434
9.849
10.247
10. 630
11. 000
11.358
4
Xx XI
. 333
.242
. 200
. 174
.156
. 143

. 132
. 124
. 117
. Ill
. 106
. 102
. 098
. 094
.091
.088
5
SS^/S
R' x
or
.333
.575
. 775
.949
1. 105
1.248

1.380
1. 504
1. 621
1. 732
1. 838
1.940
2. 038
2. 132
2. 223
2.311
6
x
or
3.000
3.478
3.871
4.215
4.525
4.808

5.072
5.319
5.552
5.774
5.985
6. 186
6.379
6.567
6.748
6.923
                                                            4.21

-------
Table 4. 7.  SOLUTION TO EQUATIONS (4. 31b) and (4. 32)
                           for
                         Z = 16.0
1
X
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
2
xZ + 1
17
33
49
65
81
97
113
129
145
161
177
193
209
225
241
257
3

x|xZ +1
or
s /sD
x' R
4. 123
5. 745
7. 000
8. 062
9.000
9. 849
10.630
11.358
12. 042
12.689
13. 304
13.892
14.457
15.000
15. 521
16. 031
4
STJ/S
R' x
0.242
0. 174
0. 143
0. 124
0. Ill
0. 102
0. 094
0. 088
0. 083
0.079
0.075
0. 072
o. 069
0. 067
0. 064
0. 062
5
S VSx
or
n VseT
0.242
0.416
0. 559
0.683
0.794
0.896
0. 990
1.078
1. 161
1.240
1.315
1.387
1.456
1. 523
1. 587
1. 549
6
X
"V^T
or
SeT/SR
4. 132
4.808
5.367
5.856
6.297
6.696
7.071
7.421
7.752
8.064
8.365
8.652
8.928
9-192
9- 582
10. 329
                                                          4.22

-------
Table 4. 8.   VALUES FOR m AS FUNCTION OF n
1 	
n
1
2
3
4
5
6
7
8
9
10
m
1.0
0.78
0.72
0.68
0. 65
0.63
0.61
0.60
0.59
0. 58
n
11
12
13
14
15
16
17
18
19
20
m
0. 57
0. 56
- -
0. 55
-
- -
0. 54
- -
- -
0. 53
    Consider then,  the problem previously stated,  wherein



         U,.,     . =  3.0 fpm with n compartments on-stream
          f(avg. )


         ST      =  4. 0 in.  H2O/fpm



         SR      =  0. 5 in.  H2O/fpm



         n        =6
                                                            4.23

-------
Substituting in equation (4. 33a)

     Se  = 0.5 + (0.63)(4. 0-0. 5)                         (4. 34a)

         = 0. 5 + 2. 2 = 2. 7 in. H2O fpm                  (4. 34b)

and, from equation (4. 1)

       P     = 2.7(3. 0) = 8. 1 in.  H7O                   (4.35)
        max                       £

as compared to 8.22 in.  H,O in previous calculations.

In view of other factors, the difference does not seem significant.
It can  be  said, then, that the maximum filter drag for a bag-
house  with n units on-stream at all times is,  within practical
limits, a direct function of the terminal and residual drags of
a single compartment, and a factor which is dependent on the
number of compartments.

An approximate design equation which can be used for all units
is:

                                                        (4.33)
Adjustments of Equation to Other Operating Conditions

As noted in the beginning of Section IV-C, the continuous operation
of multi-compartmented baghouses can take place in essentially
two ways:

1.  Operation with n compartments on-stream at all times,
    and a stand-by section available to replace a dirty unit.

2.  Operation with n compartments for the first part of  a
    cycle, and n-1 compartments for the final portion of a
    cycle.

Pressure differential readings from "the clean-air side  to the
dirty-air side" for a cycle  of operation in each of the above
cases can be represented by the schematics shown in Figure 4. 5.

For Figure 4. 5a,  Ap increases in a non-linear manner until
the end of a cycle. At this time, a clean compartment is
substituted for a dirty compartment, and  AP immediately
returns to the minimum value.


                                                         4.24

-------
5. 0
                                          1=1
                                                                   = 2
4.0
3 .0
 .oi—
i.o I-
                                                                     n= 4
                                                                   1
 L	..1	L	1	L	
1.0        2.0       3. 0         4. 0        5.0       6. 0        7. 0
                             ST/SR
Figure 4. 5   DESIGN CURVES  FOR MULTI-COMPARTMENT BAGHOUSE

-------
Q

a)
 CD
 CO
 CD
 h
a,
                  /
                  A
                /
                                                    -SeT'Uf
                     One cycle  of operation

                     r\ Compartments on-stream.
                               Time

       Figure 4. 5a  OPERATION ACCORDING TO CASE (1)
     SeT-Uf
Q

CD
co
CD
CD
h
0,
                     Compartments on-stream
                                             H - 1

                                             Compartments

                                             on-stream
                      One cycle of operation
                             Time


       Figure 4. 5b   OPERATION ACCORDING TO CASE (2)
                                                              4.26

-------
For Figure 4. 5b,  Ap again increases  in a non-linear manner.
Near the end of the cycle, however, a dirty compartment is
taken off-stream for cleaning.  At this time  AP jumps to a
higher level since  the filter area has been reduced (i. e. ,  average
velocity increased through compartments remaining on-stream).
Filtration  continues with n-1 compartments until the cleaned
unit is put on-stream. At that time AP falls to a minimum.
For this mode of operation, Se   will be approximately equal
to that of Figure 4. 5a, provided  cleaning time is not a significant
portion of  filtration time.  For example,  consider a four-
compartment baghouse with a cycle time  of 30 minutes and  a
cleaning time of 2  minutes.   In this case  each compartment
will collect dust for a period of (4)(30)  - 2 or 118 minutes.
Thus,  the  compartment will be off-stream  approximately 2%
of the time; this should not seriously alter  the pattern of filter
drags from that previously developed.

In comparing the two  modes of operation  some slight  confusion
is possible regarding the  (weight of dust)/(square foot of fabric)
collected in a given compartment.  For all cases this  can be
calculated as follows:
       W = C  • U,,     x-t                               (4. 36)
             p   f(avg.)                                 v     '
where
     t = filtration time;  i. e. ,  interval between the time a clean
        compartment is  put on-stream and the time taken off-
        stream.
Characterization of Filter Units

To complete this discussion of "Mathematics for Fabric
Filter Operation" it is necessary to develop those equations
which will allow for a field determination of critical
characteristics.  These characteristics are:

1.  Average residual drag of a single compartment (SR)

2.  Average terminal drag of a single  compartment (S_,)

3.  Residual velocity (Uf )„

                                                        4.27

-------
The importance of S-. and S — has been shown.  The significance
of Uf  will be demonstrated in latter sections.  At this time,
however, Ufp can be defined as that velocity which exists in a
cleaned  compartment immediately after it has been put in
service.  As such, it may be several times  greater than U^,    .
and, therefore, may be important in determining filter       °' '
efficiency, dust permeability, and filter blinding.  These
critical  operating characteristics may be determined from
pressure differential measurements taken across two points
common to the inlet and outlet sides of all compartments, if
the total flow through the system is known.  Such measurements
will produce a curve of the type shown in Figure 4. 6.  At time
t  on this figure a freshly cleaned compartment has come
on- stream, and during the  interval t -t, filtration proceeds with
a given number of compartments (n).  When t, is reached one
compartment is at its terminal condition and is taken out of
service for cleaning.   The  pressure differential immediately
increases from ^P  to  Ap    During this instant there is no
change in the drag 01 the remaining units as the flow through the
filters is laminar.  Filtration then proceeds with n-1 compart-
ments until the cleaned filters are back in service. At this
time  (tc) the pressure differential decreases from  ^PA to   Pi .
Again, there is no change in the drag of the units continually
filtering.  These instantaneous changes in pressure differential
are indirect measurements of terminal and residual drag  in a
particular  compartment and can be used with the relations which
follow to characterize a particular mode of operation.
Thus,  at time t^ and pressure differential ^F^' equation (4. 8b)
can be rewritten as follows:


                                                       (4. 37)
According to the mode of operation the compartment with
drag = S™ is now taken off -stream, so that at time = t,  and
pressure differential
                                                       (4.38)


Combining equations  (4. 37) and (4. 38)

                                                        4.28

-------
   o
     c^
   «
   CO
   W
   u
   3
   1—4
   H
   W
   W
   fa
   H
   tf
   w
   tf
   P-
                           TIME
Figure 4.6.  SCHEMATIC PRESSURE DIFFERENTIAL
 CURVE FOR MULTI-COMPARTMENT BAGHOUSE
    This then allows for calculation of S— from measurements  of
    Sp-r, and S« .
     er      e
   At time t  and pressure differential  A P,,
                                                          (4.40)
   Again,  according to the mode of operation, a compartment
   with filter drag So  is now brought on-stream,  so that at
   time t  and pressure  differential
         n
                         +
                R
                      R
(4.41)
                                                            4.29

-------
 Combining equations (4.40) and (4.41) produces an equation
 which allows for the calculation of Sj^, as follows:

      1     n      n"1-                                (4.42)
                     4
R     eR     "e
As an example, consider a baghouse with n = 6 and the following
values of AP and Uf:


      AP. =  5.32 in. H_O  Uf  =3.5 fpm
                             1
      AP2 =  8.22    "      Uf  =3.0  "

      AP3 =  8.60    "      Uf  =3.4  "

      AP. =  8.70    "      Uf  =3.4  "
        4                    4

Effective filter drags can then be calculated for the baghouse.
Thus,

     Se.  = Se  =  ^iH  = 1.52 in. H,O/fpm
       •L     R.    J • D             L*
                                   \\    \\
    Se  = ^4°  =2.53
      e3    3.4


    Se4=«0°=2.56        "    -
Using equation (4. 39)

     1     6       5
          2T74~  "  Z75T
                 = 0.22
                           and ST = 4. 5
Using equation (4. 42),

          1  -   6       5
             '  TT5T "
                           and S_ =0.5
                                XV
                                                      4.30

-------
In summary,   ST = 4. 5

               SR = 0. 5

              AS  = 4. 0
For the sake of discussion, assume that cycle time is equal
to 30  minutes for the above problem.   That is, filtration time
equals 6(30) or 180 minutes.  If, in an attempt to lower AP,
cycle time is reduced to 20 minutes, the following will result:

1.  Filtration time will be reduced to  6(20) or 120 minutes.

2.  From equation (4. 3b),

           „    ... 120   ., ,_
           s =  4-°       2-67
3.   From equation (4. 33b),  and Table 4. 8


         SeT = 0. 5  + 0. 63(2.67) = 1.68 in.  H2Q/fpm


4.   At U,  =3.0 fpm, therefore,
         A P?  =  l. 68(3. 0) = 5. 1 in.  H?O compared to the original
                            8. 22 in. H2O.

In a similar fashion the equation could be used to predict various
values of A P as n is  varied.

As mentioned in the beginning  of section IV-F, a characteristic
of bag filters which undoubtedly influences  filtration  is the
filter velocity  itself.  For any two units operating with the
same Uf,    ., entirely different velocities of filtration may
exist thrWjfn 'each compartment.  Each individual pattern
will be a function of S™, S^ , and n.
                      1   iv

The velocity in a compartment at any time  can be estimated,
provided the filter drag of the  compartment (S ) and  the effective
drag of the baghouse (S ) is known at that time.   Thus, for any
compartment,

     S  = *j£-                                         (4. 43)
      -A.    \J £
             X

                                                       4. 31

-------
   While for the entire baghouse,
                AP
         6    ^Vvg.)

   therefore, since AP is the same,
                                                               (4. 44)
                                                              (4.45)
   It would appear that a particularly critical value of D£
   would be the highest value reached in a  compartment, x
   This would be the velocity in a clean compartment as it is
   just put on-stream, and is designated Ur
                                             R
   Accordingly,

              S
               eR
                   U
                     f(avg. )
(4.46)
   An approximate  velocity pattern for a baghouse with n = 6 is
   shown in Figure 4.7.
                   AVERAOC VELOCITY : 3.0 lorn
                   RESIDUAL COMPARTMENT niTEH 'J«AG ! CW IN. H O/tpm
                   TERMINAL COMPARTMENT FII.TEH IRAS z.ow. Mjo/ipm
                                     COMPARTMENT
                                         6
                         30    40    50
                         TIME (MINUTES)
Figure 4.7   VELOCITY PATTERN IN 6-COMPARTMENT
       BAGHOUSE AS  FUNCTION  OF  TIME
                                                               4. 32

-------
    The fact that such velocity variations exist would appear to
    emphasize the need to gather data on filtration through actual
    field studies of multi-compartment units, or by studies of
    multi-compartment pilot plants.  The data obtained by techni-
    ques outlined in Sections I and II must certainly be used with
    caution.  This is  not to say,  however,  that a better understanding
    of filtration  cannot be achieved by such basic studies.
G.  Summary of Useful Equations and Tables.

                 AP
f
1
ST
1
5~ ~
**•
n
Se
T
n
s —
eR
n- 1
e3
n - 1

64
                                                           (4.1)


                                                           (4.2)
                  1  '• ^- + .   .  •   •  + g^-                  (4. 8b)
                   12                n
                                                            (4.24)
             = SD + mAS                                   (4. 33a)
           ~    t\.
          AW = C • U,,    .-t                               (4. 36)
                 P  f(avg.)                                      '


                                                            (4.39)
                                                            (4.42)
                S
         Uf  -  ---   U,,     .                               (4.46)
          fR    -SJ-   f(avg.)
                                                            4.33

-------
Table 4. 8.  APPROXIMATE VALUES FOR m AS FUNCTION OF n
n
1
m
1.0
2 1 0.78

3
4
5
6
7
8
9
10
,
0.72
0.68
0.65
0.63
0. 61
0.60
0.59
0.58
n
11
12

13
14
15
16
17
18
19
20
m
0.57
0. 56

- -
0. 55
- -
- -
0. 54
- -
- -
0. 53
                                                        4.34

-------
                                          26
SECTION 26
Effect of Changing Permeability,
  Varying Flow Rate,  and Non-Laminar
  Head Loss

-------
                     EFFECT OF CHANGING PERMEABILITY,
          VARYING FLOW RATE,  AND  NON-LAMINAR HEAD LOSS
Id*-Hi conditions for a baghouse include con-
stant inlet conditions,  a constant dust per-
meability, and the absence of non-laminar
head loss elements such as valves, duct
work, i'tc.   In this section we will explore
the  pffc.-i t of certain of these  parameters on
the  theories so-fax- advanced, especially in
relation to predictions of baghouse AP.

A  Graphical Solution to Design Equations

   To account for  a non-linear basic
   performance curve,  it will be necessary
   tn develop a graphical solution to the
   equations of Section IV.

   In Section IV,  it was shown that under
   ( onstant inlet conditions the total of all
   drag changes will be the same  from cycle
   to cycle (Equation  4. 17).  As discussed,
   !.hr changes in drag from  compartment to
   compartment will follow a definite pattern
   (j',jb]e '\. 1), with the greatest change
   occurring in the compartment just put
   on-stream.  The equation for  drag of any
   compartment at the end of a cycle was
   derived in Section  IV (Equation 4. 24vi; this
   drag also represents drag in the same
   compartment at the beginning  of the next
   cycle.  This concept is  shown in graphical
   form in Figure ,r>. 1.

   The vertical scale of Figure 5. 1 represents
   filter area  expressed as percent of the
   total.  Thus, using the example of Section
   iV-C, where n   G, each filter area
   represent.:--, 100 or  Hi. 7%  of the total.
   The horizontal scale represents the filter
   drag ratio S, /S|^ as defined by equation
   (I.Li'O.  if,  Ihcn,  V/L calculate SX/SR for
   each compartment at the end of a cycle,
   a profile of drag values within the bag-
   house can be  plotted.  This is shown for
   the example previously  cited, as the
   right-hand boundary of the cross-hatched
   sections ni Figure 5. 1.  A profile of drag
   values at the start of a cycle can also be
   represented.  Thus, for the compartment
   just cleaned, ST/SR=  1.0.  In addition,
   it is known that the drag of one compart-
   ment at the end of a cycle equals the
   drag of the adjacent compartment at the be-
   ginning of a cycle.  The  result is the left-
   hand boundary of the cross-hatched sections
   on Figure  5. 1.

   The cross-hatched areas represent the
   change in drag for each compartment of
   a multicompartment unit during one cycle
   of operation.  From cycle-to-cycle a
   given  compartment will proceed, in
   sequence,  through the incremental
   changes shown.
   Now,  if r; is increased beyond 6, the
   incremental changes per compartment
   will become smaller and smaller.  Ifr;
   is allowed to increase indefinitely, the
   cross-hatched areas will approximate a
   smooth curve.  This curve represents
   the Drag Profile in a multicompartment
   baghouse with an infinite number of filter
   areas.  For such an installation, the
   pressure drop would be constant.

   For convenience, equation (4. 24) has been
   solved for n - 10 and selected values  of
   ST/SR. The results are summarized in
   Table 5. 1.   By plotting this data and
   connecting the points with smooth curves,
   it is possible to generate a family of
   curves for r\  °°  as shown in Figure  5. 2.
   By reversing the "construction" process
   values of SX/SR can easily be determined
   giving the  solution to equation 4. 24 by
   graphical  means.

B  Effect of Varying Permeability

   As developed, the drag profiles shown
   in Figure 5. 2 represent conditions when
   K is constant.  If K is not constant, but
   assuming it  follows some regular pattern,
   a drag profile will still exist, but its
   shape will differ from those shown.   It

-------
   100	
    80
    60 ...	
g
o
M
0)
O.
a   40
Q)
M
M
I)
    20
               I     I     I     I     I     I     I     I     I     1
                                       13
                                       3

                                       3.
                                       3
                                      TO
                                                                                                                      s
                                                                                                                      fD
                                                                                                                      P
                                                                                                                      cr
         1.0       2.0        3.0       4.0        5.0        6J3        7.0
ao
                                 Filter Drag Ratio,  S   /

                                                      V SR
                Figure 5.1     Drag Profile with S  /   =   8.0 and n =  6

                                                    T/SR

-------
Table 5. 1 SUMMARY OF SOLUTION TO EQUATIONS (4. 24)
       SELECTED VALUES OF S  ,    AND r) =  10.
*/*
2
3
4
5
6
7
8
1
1. 140
1. 342
1. 581
1. 844
2. 121
2.408
2. 702
2
1. 265
1. 612
2.000
2.408
2. 828
3. 256
3. 688
3
1. 378
1. 844
2. 345
2. 864
3.391
3.924
4.461
4
1.483
2. 049
2. 646
3.256
3.873
4. 494
5. 119
5
1. 581
2. 236
2. 915
3. 606
4.301
5.000
5. 701
6
1. 673
2. 408
3. 162
3.924
4. 690
5.459
6.229
7
1. 761
2. 569
3. 391
4. 219
5.050
5. 882
6. 716
8
1. 844
2. 720
3. 606
4. 494
5. 385
6. 277
7. 169
9
1. 924
2. 864
3. 808
4. 754
5. 701
6. 648
7. 530
10
2.000
3. 000
9.000
5.000
6. 000
7. 000
8. 000

-------
      Figure  5.2    Ideal Drag Profiles  for r, = °°  and Selected Values  of S   /

                                                                                 V   i
d
OJ
a
m
ai
o.
       100
       80
       60
                    2.0        3.0       4.0       5.0        6.0        7.0        8.0 = S.
3      40
       20
                                                                                            n

                                                                                             0
                                                                                            3
                                                                                             q
                                                                                             3'
                                                                                            ffq


                                                                                            ft)

                                                                                             3

                                                                                             u
                                                                                            cr
          1.0
3.0       4.0        5.0
7.0         8.0

-------
                                                            Effect of Changing Permeability
 follows, then, that by assuming various
 shapes for the drag profile, different
 drag changes per compartment per cycle
 can be  graphically determined.  With this
 information,  velocities in each compart-
 ment at the beginning and end of a cycle
 can be  calculated and eventually the Basic
 Performance Curve for a given compart-
 ment cn.n be approximated.

 By way of example, assume the following:

 ST - 0. G   in. H2O/fpm

 SR   0. 2   in. H2O/fpm
 Uf -
4  fpm with n compartments
on-stream
 Kur-ther a.ssuine four possible variations
 in the Drag- Profile ranging from the ideal
 (Cast- A) to an extreme variation (Case D),
 as illustrated in Figure 5. 3.  By means of
 previously derived equations Serp/So and
 •VIJ/SR can be calculated; also, assuming
 ;i negligible but finite  cleaning time,
ij
           S(:./SR  can also be determined.
Since Uf i.s constant, values of AP can
then lie calculated.  These results are
summariy.ed in Table 5. 2,  wherein the
subscripts are the same  as used on
Figure  (4. (i).  Approximate AP curves
are shown in  Figure T>. 4.

From value:,  of Se,,,/,So    and  Sfp/Sp,.
and using equation (4.45),  filter velocities
c an be calculated for each  compartment
;it the beginning and end of a  cycle.  These
:irc shown in  Figure r>. T>,  with  approximate
connecting curves.  From  these curves,
nverage velocities can be determined for
 ;ich compartment allowing for a calcula-
lion of  AW.   liy plotting Sx as  a function
of AW  the Basir:  Performance  Curves
of Figure .5. >i  arc obtained.

It can be seen from the above illustrations
that as  the Drag Profile is skewed to the
right:
   1  Baghouse AP  generally increases
      and the magnitude of the jump from
      AP2 to APg  increases.

   2  Residual  velocity (UfR)  increases,
      resulting, perhaps, in increased
      bleeding.

   3  The Basic Performance Curve becomes
      further removed from the "ideal"
      straight line between SR and S'p.

   If desired,  any Basic Performance Curve
   can be matched to a  Drag Profile by
   "trial and error". The nature of the
   basic data and the accuracy required would
   determine the desirability of such  an
   approach.

C  Use of Equations (4.  38) and (4. 41)  when
   K is not Constant

   Of some importance  is the effect of a
   non-linear Basic Performance Curve on
   determination of S-p  and Sp^ from a bag-
   house AP  curve.  The original derivation
   was based solely on  changes in AP, as a
   compartment was taken off-stream and
   then returned to service.   The jumps in
   AP (Figure 4:6) were shown to be
   functions of ST and S^.

   To demonstrate the effect,  equations
   (4. 38) and (4.41)  have been solved using
   the data of Table 5. 2.  The results of
   these calculations are summarized in
   TabJe 5. 3,  indicating that the non-linear
   Basic Performance Curve has  little
   effect on determinations  of S-p  and  SR
   from baghouse AP.

   It should be noted, however, that the use
   of  such results to predict AP  at other
   operating conditions  would treat K  as a
   constant.  The predictions, therefore,
   would be on the low side.

-------
Effect of Changing Permeability
    c
    01
    CJ
    n
    
-------
                                                                Effect of Changing Permeability
                 Table  5. -  SUMMARY OF BAGHOUSF FIl.rER DRAG AND PRESSURE
                          DIFFERENTIALS FOR HYPOTHETICAL, CASES
   Case A
   Case R






Se2;
bR
2. 20
2.45

2.61
3e3/s
R
2. 06
2 . 36

o r -^


1 . 36
1. 48

1.5,
Po

1. 76
_^ __
1. 96

2.09
Po

2.06
2. 36

2. 53
i   L'ase D
 D Kffeet of Fan Characteristics on
   Baghouse Performance

   Situations wherein the rate of dust
   entering a baghouse (grains/ min. ) and
   the volumetric flow rate are variable,
   will,  of  course,  differ from  eases per-
   \ iou:,tv discussed.  Treatment of  the
   data in these  instances will  depend on
   tlii  specific installation.  A special case
   whii h  can be  analysed in a general
   1'ash'on,  however,  exists  when the rate
   •i( dust ma •;>  entering a baghouse  is
   constant and \-olumetric flow variations
   arc a result of tan characteristics.  For
   ;his situation, the  theories so far
   developed will apply relative to filter
   drae.  11 is emr intention, ho.vever, to
   demonstrate how fan characteristics
   •.oil influence baghouse AF and ^Tf-
Two extreme cases  era be visualized in
looking at a baghouse as part of an overall
process:

   1  The magnitude of energy  losses other
      than tliat associated with  the baghouse
      is large, so that  changes  in baghouse
      drag do not materially affect fan
      output.

   2  The magnitude of energy  losses other
      than that associated with  the baghouse
      is negligible,  so  that fan  performance
      is basically related to changes in
      filter  drag.

      It is the latter case that concerns us
      hert,   Consider, for example,  that
      a curve similar to  Figure ^4. G) has
      been predicted on the basis of  constant
      flow rate, and farther assume that a

-------
Effect of Changing Permeability
      o
       IN
            3.0
            2.0
             .0
                                 It
                                -O
Case A
                    iiii
                                      3.0
                                                             2.0
                                       .0
                                                           OQ


                                                          -O
Case B
                                                                     i     i     i
                         Time
                                                                          Time
      •rH

      T3
      a
      en
      en
      
-------
                                                     Effect of Changing Permeability
    Figure 5.5    Velocity  patterns in hypothetical  problem
EJ
o.
         8.0
         4.0
           0
         8.0
         4.0
          8.0
          4.0
A
                               Case  A
                                Case  B
                                Case C
            01234       5


                               Time
            01       2345



                              Time
            012345



                              Time
          8.0
          4.0
           0
                                Case  D
            012345



                             Time

-------
Effect of Changing Permeability
                       0.8
                       0.4
                                   Case A
                                                       12
                                                                 16
                                                                           20
              e
              o.
              00
              n
              n

              o
                       0.8
                       0.4
                       0.8
                       0.4
                                                       12
                                                                 16
                                                                          20
                                                       12
                                                                16
                                                                          20
                       0.8
                      0.4
                                                      12
                                                                16
                                                                          20
                                    Dust Mass  (  graines/ft  )
                   Figure  5.6     Basic performance  curves for hypothetical  problem
  10

-------
                                                        sffVci of Changing Permeability
 TABLE 5. 3  VALUES OF ST AND SR
  CALCULATED FROM FIGURE 5.4
  USING EQUATIONS (4. 38) and (4. 41)

Original Values
Cast.- A
Case B
Case C
Case D
ST
0. 60
0. GO
0. 58
0. 56
0. 59
SR
0. 20
0. 20
0. 19
0. 20
0. 20
 "backwardly inclined" fan has been
 chosen such that it will deliver its
 "rated capacity" at AP2-

 From the fan performance curve,
 (Figure 5. 7) it can be seen that as
 the static pressure at the  fan decreases,
 fan output increases.  Therefore,
 when a clean compartment is put on-
 stream and AP decreases, fan output
 will increase.  This, in turn,  will
 increase Uf  so that baghouse  AP will
 come to equilibrium at some value
 higher than predicted when considering
 constant flow rate conditions.
      Pressure to Fan Discharge is propor-
      tional to the effective drag of the
      baghouse.

      By means of this proportion, it is
      possible to estimate flow rates at
      various values of Se-  For example,
      consider Case D of the previous
      section,  wherein:
                                                                0.404 in. H2O/fpm

                                                                0. 558 in. H2O/fpm
      Further assume that Uf_   4 fpm and
      that the Minimum Design Flow Rate
      isAf  Uf2.

      When a cleaned compartment is put
      on-stream,  then the ratio Fan Static
      Pressure/Fan Discharge will equal:

               0. 404
               0. 558
                          0. 724
(5. 1)
      Entering Figure 5. 8 from the right.
      Fan Discharge is determined as 113. 5%
      of Minimum Design Flow Rate.
Therefore:

      u
                       (113. 5)
       fi  =  •±-u^'»   -Too-
           -  4. 55  fpm
                                       (5.2)
l''or convenience- in correcting for
inc reused flow rate,  we have first
replottrd the Static Pressure curve
of Kigure 5. 7 in generic terms,  such
lhat Kan Discharge is expressed as a
percent of Minimum Design Flow Rate
and Static: Pressure expressed as a
percent of Rated Static Pressure  (See
Figure 5. 8).   It is then possible to
construct a second curve with the ratio
of Static Pressure to Fan Discharge
plotted as a function of Fan Discharge
(also shown on Figure 5. 8).   For the
system  under consideration (i.e.,
essentially all head loss associated
with the baghouse) the ratio of Static
and:
also:
           =  0. 404   (4. 55)

             1. 84  in. H2O
                                      ( 5. 3)
              0  404
      UfR =  -—y- (4.55) = 9.2 fpm   (5.4)
Under conditions of constant velocity, it was
shown that:
and;
          AP    1. 63  in. H2O
                                                      UfR = 8. 2 fpm
                                                                                     11

-------
                              Figure 5.7     Fan performance curve
    10.0
                               10

                                ID

                                OQ

                                5'
                                crp
o
 CN

33
to
*j
CO
     8.0
     6.0.
     4.0
     2.0.
                                                                        Mechanical

                                                                          Efficiency
                                               Brake Horse Power
                                                                                              50
                                10           15          20           25          30


                                        Fan Delivery ( 1000  cfm )
35
W

rH
0)
                                                                                                    ,
                                                                                                    O
                                                                                                    01
          M
          cu

          I
          en

          Cl)
          CO
          M
          O
          X
                                                                                                     a
                                                                                                     H
                                                                                                     P5
                                rc
                                tu
                                cr

-------
                        Figure 5.(
Static pressure curve
120
ICO
 80
 60_
 40 .
 20 _
                 Siaiic Pressure
                            Static Pressure:  Discharge Ratio
                                                                                   1.0
                                                                                   0.8
                                                                                   0.4
                                                                                   0.2
                                                                                   0.
                                                  01
                                                  oc
             20
                       40
                                60
                                          80
                                                   100
                                                             120       140       160
                Fan Discharge ( Percent DF Minimum Design  Flow Rate )

-------
 Effect of Changing Permeability
   It can be concluded, then, that the use of
   a "backwardly-inclined" fan has the effect
   of increasing AP, and increasing Uf•  ,
   compared to constant flow-rate conditions.

   It should be noted that the increase in Uf
   occurs at a time when the system is sus-
   ceptible to bleeding.  This is especially
   significant when "velocity profiles" along
   the length of a filter bag are considered
   and if variations from filter to filter are
   considered.

 E Influence of Non-Laminar Head Loss

   In many fabric filter units, resistance to
   flow is predominantly associated with the
   dust-fabric combination per se.  In other
   units, significant energy loss can be
   associated with the entire baghouse, but
   it is still possible to determine that
   portion of the overall AP  associated with
   the dust-fabric combination because of
   common inlet and outlet chambers. In
   still other units, significant resistance
   to flow is caused by valves,  duct-work,
   elbows,  etc. in series with each compart-
   ment.  In these cases, the volumetric
   flow rate through a compartment will be
   influenced by the added elements.  The
   equations and concepts previously derived
   will not be  applicable since the energy
   loss associated with these components will
   be approximately proportional to (Uf)2
   instead of Uf.

   By way of example, consider one section
   of a baghouse,  in which the resistance of
   the dust-fabric combination is S and the
   resistance  of other components  is R.   The
   pressure differential across each compon-
   ent,  then,  is:
   AP  (dust-fabric)  r  Uf
(5.5)
   AP  (valves, elbows,  etc.) = Uf2 •  R (5.6)
Head loss across the entire compartment is:
   AP  total    AP   + A

                          2
               Uf S  + U/R
(5.7)

(5.8)
          and the resistance of the compartment
          becomes:
             Y  =
      AP total
        U.
=  S  + Uf R
(5.9)
          and is seen to be a function of U*.

             The addition of this resistance which is
             a junction of Uf would,  then,  act as an
             equalizer in the system.  It would, for
             instance,  result in lower values of
             which, in turn, would reduce the  rate
             of dust deposition in the most recently
             cleaned compartment.  The baghouse would
             reach equilibrium, but at a higher AP and
             showing less variation  from AP^  to APj.

             If desired, the non-laminar head  losses
             can be introduced in the derivations of
             Section III.  The  resulting equations,  how-
             ever, would, require a computer for their
             solution.

             To illustrate the  effect of non- laminar
             head loss (either in series with individual
             compartments, or in series with  the entire
             filter area),  as the average filter velocity
             (Uf avg. ) is increased or decreased,
             consider the following illustration:

             Assume a four compartment baghouse in
             which S for each  compartment is  1, 2, 4
             and 6 in. H2°/fpm, and Uf = 3. 0 fpm.
             Also assume R is the same for each
             compartment and that R - 0. 5
             in.H2O/(fpm)2.
For the entire baghouse, then:

                          R)
                                                 (5. 10)
Also,
   jl
   "§„
                                                     And,
                         1.917
                                   2.09
                                    (5.12)
From equation (5. 10),
Ye = 2.09 + (3. 0) (0. 5) = 3. 59 in. H2 O/fpm

                                    (5.11)
  14

-------
                                                             Effect of Changing Permeability
and,

   AP = 3. 59(3. 0)    10.77 in.  H2O   (5.12)


If Uf    is reduced to 1. 5 fpm, then

   Ye   2. 09 + (1. 5) (0. 5)   2. 84 in. H0O/fpm
                                     LJ
                                       (5. 13)

and,

   AP =2.84(1.5)  - 4. 26in. H2O.   (5.14)


   Thus,  in the example  cited above, reduc-
   ing the flow rate  by the ratio 1/2 has
   reduced AP by the ratio  1/2. 5.

F  Discussion

   In this Section we have attempted to show
   the effect of several parameters normally
   existing in any installation  that create
"non-ideal" conditions.  Each of these was
tested  idependently;  no attempt was made
to assess their combined impact on pre-
dicted  operating pressure differentials or
other potential problems.  Such an
analysis, of course,  would be specific to
a given installation.

Perhaps the area of greatest need is
further data on the extent of non-laminar
head loss for the "typical" bag filter unit.
In present designs  the head loss must be
at least 2   3 inlet  velocity heads, with
perhaps an average of 4   5.   Under such
conditions the "typical" baghouse  will
automatically have a  pressure drop of
say 3 inches H2O  without the  filter tubes
in place.  It would  seem that changes in
design could reduce this figure considerably.
                                                                                           15

-------
                                          27
SECTION 27


Electrostatic PreclpHators Operation
  and  Industrial  Applications

-------
       ELECTROSTATIC  PRECIPITATORS
OPERATION AND INDUSTRIAL APPLICATIONS

                      H. L. Engelbrecht
             I  Introduction
               A.  Particle Charging
               B.  Particle Transportation

            II  Influences on the Performance of an
               Electrostatic Precipitator
               A.  Migration Velocity
               B.  Specific Collecting Surface
               C.  Particle Size and Dust Concentration
               D.  Gas Velocity
               E.  Electrical Wind
               F.  Dust Resistivity

           III  Special Design Considerations
               A.  Discharge Electrodes
               B.  Collecting Electrodes
               C.  Rapping Cycle Control
               D.  Electrical System Control

            IV  Operation of an Electrostatic Precipitator
               A. Gas Conditioning
               B. Gas Distribution

             V  Typical Industrial Precipitator Applications
               A.  Precipitators for Flue Gases from
                   Power Stations
               B.  Precipitators for the Iron and  Steel
                   Industry
               C.  Precipitators for the Cement Industry
               D.  Precipitators for the Chemical  and
                   Non-Ferrous Metallurgical Industry
               E.  Precipitators for Waste Gases from
                   Incinerators

            VI  Summary

-------
                                  ELECTROSTATIC PRECIPITATORS
                             OPERATION AND INDUSTRIAL APPLICATIONS

                                    Heinz L. Englebrecht*
I  Introduction

   In general,  dust collectors can be class-
   ified  as dry process mechanical collectors
   (cyclones,  bag filters),  wet process
   mechanical  collectors (scrubbers), or dry
   or wet process precipitators (electrostatic
   precipitators).  Selection depends largely
   on the specific  dust collection problem,
   the economical situation, and the suitabil-
   ity of the  equipment for the user.

   Each one of these collectors has a prefer-
   red range of application as far as the
   required collecting efficiency is concern-
   ed.
      Figure 1.  COLLECTING EFFICIENCIES

                        Wet-Process
              vtttftM^ Precipitator
                     Dry-Process
                     precipitator
                 Wet-Process Mechanical Collector
  """""*
            I Dry-Process  Mechanical Collector
     Collecting Efficiency


   Since these ranges overlap at certain
   efficiencies, the selection of the equip-
   ment is limited by economical considera-
   tions .

      Total Cost
f////////Ah\\\\\\\\\\\\\VM Mechanical  Collector
7////////////////////K88&
Investment

     Cost
                       Electrostatic Precipitator
                       Operating  Cost
     Figure 2.  COST FOR DUST COLLECTING
                EQUIPMENT FOR EQUAL TOTAL
                COSTS
Other factors to be considered are code
requirements and the availability of
electric power, water, compressed air, and
waste treatment facilities.

Electrostatic precipitators have advantages
that in principle they can be designed for
any desired degree of collecting efficiency
and they will deal effectively with dust
or moisture particles in the sub-micron
range. The dust is recoverable in either
dry or wet state and flow resistance and
power consumption are usually less than
for all other methods of dust collection.

Industrial electrostatic precipitators are
normally single stage (charging and collect-
ing of the particles in the same stage or
electrical field)precipitators. The number
of fields in series or the number of elect-
rical sections in parallel will vary with
the size of the precipitator.

Electrostatic precipitators can be class-
ified according to the mode of operation
as either dry process precipitators or
wet process precipitators; according to
the collecting surface as either pipe type
precipitators or plate type precipitators;
or by direction of the inlet gas flow as
either horizontal gas flow precipitators
or vertical gas flow precipitators.

The predominant group is the dry process,
plate type,  horizontal gas flow precipi-
tator as shown in Figure 3.
                                                                       meters
Consulting Engineer, Wheelabrator Corporation
                                                          Figure 3.  DRY PROCESS, PLATE
                                                                     TYPE HORIZONTAL GAS
                                                                     FLOW PRECIPITATOR

-------
Electrostatic  Precipitators Operation and  Industrial Applications
  Electrostatic precipitation requires a
  discharge electrode (usually negatively
  charged) of small cross sectional area
   (such as a wire), and a collecting elect-
  rode  (usually positively charged and at
  ground potential) of large surface area
   (such as a plate or tube).

  In operation, a unidirectional (DC) high
  potential electrical field of the order
  of 10 to 70,000 volts Is set up across'the
  electrodes. The dirty gas stream is passed
  between these electrodes. A typical gas
  passage is shown in Figure A.
 Normally a negative charge is applied to
 the  discharge  electrode of an industrial
 electrostatic  precipitator.

 As the  ions move,  they attach themselves
 to the  neutral dust particles causing them
 to migrate toward  the appropriate electrode.
 Most of the dust particles pickup negative
 ions. This is  because the negative ions,
 moving  toward  the  positive collecting elect--
 rode, have a greater distance to travel than
 the  positive ions  which are  formed near and
 migrate toward the negative,  or discharge
 electrode.  As  the  charged particles contact
 the  electrodes,  they become  neutral after
 which it is relatively simple to remove them
 by rapping, washing or gravity flow.
    Figure 4.  TYPICAL GAS PASSAGE
               IN AN ELECTROSTATIC
               PRECIPITATOR
   At  certain critical voltages, ionization
   of  the air molecules (dissociation into
   negative and positive ions) takes place at,
   or  near, the surface of the discharge or
   negative electrode and is evidenced by a
   "corona" (no sparking). Figure 5 shows
   negative and positive corona discharges.
     Negative                   Positive


         Figure  5.   CORONA  DISCHARGE
 Figure 6.  CHARGING OF PARTICLES

Therefore, there are four steps involved
in electrostatic precipitation: (1) elect-
rically charging the particles by means of
ionization, (2) transporting the charged
particles to a collecting surface, by means
of the force exerted upon them by the
electrical field,  (3) neutralizing the
electrically charged particles precipitated
on the collecting  surface, and (4) removing
the precipitated particles from the collect-
ing surface to a receptable external to the
precipitator.

A.  Particle Charging

    If there is a  sufficient source of  ions,
    the number of  charges which a dielectric
    spherical particle can obtain is:
                                                                    E r*
                                                                     o
                                                                                              (1)
    where n  •  saturation number of charges
           8    on a particle

          Eo -  electrical field  strength

-------
                                Electrostatic Precipitators  Operation  and  Industrial  Applications
            r  =  particle radius

            e  =  electronic charge = 4.8 x
                  10l° esu

            K  =  dielectric constant (for
                  minerals, K = 1 to 10; for
                  water K = 80)

      A maximum of three times as many charges
      can be placed on a conducting particle
      compared with an insulating particle.
      However, since neither extreme is gen-
      erally found, the number of charges on
      closely sized particles of varying
      dielectric constant is about the same.

      For example, with a precipitation field
      of about 4,500-volts per cm, a 1-micron
      particle acquires a charge of the order
      of 250 electronic charges, while a 10
      -micron particle acquires about 25,000
      electronic charges.
     The electrostatic force on the 10-
     micron particle is about 300 times
     the gravitational force, assuming a
     unit density spherical particle.
     (See Figure 7.)

    The existence of these relatively
    large forces is what  accounts  for the
    great effectiveness of electrostatic
    precipitators to separate fine particles
    from gases. Because these forces are
    exerted directly on the particles
    themselves, there is  no need to exert
    the necessary inertial separating
    forces upon the particle by turning
    the stream-lines and  increasing the
    velocity of the carrier gas as is done
    in mechanical collectors.

B.  Particle Transportation

    The driving force Fs  separating charged
    particles from the gas is:
 •rl

 §
      While the maximum number of charges is
      time independent, the actual number
      depends upon the rate at which a part-
      icle is charged., A precipitator will
      create about 10  to 10° ions/cm  ,  in-
      dicating that particles are fully
      charged in 0.1 second or less.

      Charges also accumulate on particles
      due to thermal diffusion of ions in
      the gas. The rate depends on ion den-
      sity, ion mobility, and time, and  is
      significant only for particles less
      than 1 micron in diameter.
              F  = n  e E
               s    s    p
                                (2)
    10 XG
      10   _
      10   -
      10
 u      0. 1        1         10        100
            Particle  Diameter,  microns


Figure  7.   ELECTROSTATIC FORCliS ACTING ON
           PARTIULKS  UNDER TYPICAL PRECIPITATOR
           CONDITIONS.
    where nse is the total charge of the
    particle and Ep represents the collect-
    ion field intensity.  This force F  is
    counteracted by a resistance force Fj,
    which in the region of Stoke's law is:

              Fd = 6 IT r y W        (3)

    where: r = radius of particle, y =
    viscosity of the gas, and W = velocity
    of the particle.

    The particles are accelerated by the
    electrical force until the resistive
    force of the gas just equals the
    driving force, i.e. Fg-Fj.
             n  eE  = 6ir r y W
              s    p
                                                                                            (4)
    The dust then travels at essentially
    constant velocity to the collecting
    surface. The velocity of the particle
    then is:
                                                                   W
                  n e  E
                   s    P
                6  TT  r  \i
                                                                                            (5)
    For particles larger than 1 micron
    in diameter, the total charge is
          n e - D  E   r
           s        o

where:  D - 1 + 2 [ K - 1
                   K + 2
                                    (6)

-------
Electrostatic Precipitators  Operation  and  Industrial  Applications
        Therefore, the final migration velo-
        city is
                W = D Eo  Ep
                            (7)
                      6 TT  y

        where:  W = Migration velocity, cm/sec

                E = Peak effective electrical
                 0  charging field, esu/cm

                E = Average electrical field
                 p  at collecting electrode,
                    esu/cm

                r = Particle radius, cm

                y = Gas viscosity, poise

 II  INFLUENCES ON THE PERFORMANCE OF AN
    ELECTROSTATIC PRECIPITATOR

    The collecting efficiency of an electro-
    static precipitator follows an exponential
    function:            .
                       -fW
             E = 1 - e   ^              (8)
    where:  E = efficiency
            e = base of natural logarithm
            A = collecting surface
            Q = gas volume
            W = migration velocity
            ^
    Setting jr = f (specific collecting surface)

    we have:              „
              E = 1 - e-f W             (9)
    Therefore, the collecting efficiency is
    a function of the specific collecting
    surface f (or the collecting surface A
    and gas volume Q), and migration velocity
    W.

    A.  Migration Velocity

    The migration velocity of a given particle
    size is expressed by equation 7 as:
                    E   E
                  D   o   P  r
                     6  TT   y           (7)
W
    With EQ being a function of the discharge
    current i   , the migration velocity also
    becomes a function of the discharge cur-
    rent.

    With increased current the probability of
    the dust being charged increases and thus
    the collecting efficiency increases. After
    the particles reach their maximum pos-
    sible charge, a further increase in current
    will  not result in a consequent increase
    in efficiency. Furthermore, W is related
to the intensity of the electrical  field
Ep. The higher Ep, or  applied voltage,
the more the particles will be attracted
towards the collecting surface. The limit-
ation is set by the breakdown voltage of
the gas, so that actually the precipitator
voltage should always be lower than the
breakdown voltage. A third controlling
factor is the dielectric constant of the
dust, represented as D. D-values for in-
sulating materials vary between 1.0 and
2.5, while for water the value is 2.9.
This is one of the reasons than an  increase
in the dew point temperature normally
favors the electrostatic precipitation of
the suspended particles. The gas viscosity
y and the particle diameter r also  influence
the value of the dielectric constant.

Values of W determined from commercial in-
stallations do not agree with theoretical
values obtained. The experimental values
are generally less than half the theoret-
ical values for W. The theoretical  calcul-
ations do not consider the effects  of re-
entrainment, which occurs when plates are
rapped or when there is uneven gas  and
dust distribution.

For commercial design purposes, the value
of the migration velocity W must be deter-
mined experimentally or by using the data
of previous installations.

B.  Specific Collecting Surface

    The specific collecting surface is
    a constant for a specific precipitator
    operating with a certain constant gas
    volume. The greater the specific col-
    lecting surface, the higher the collect-
    ing efficiency; f increases with in-
    creasing collecting surface A and with
    decreasing gas flow Q.

    A number of variations to the basic
    efficiency equation can be derived.
    Consider a simple two plate precipi-
    tator of length L, height H,  spacing
    between plates d, and gas velocity v.


    The gas volume can be written as
           Q - H  d  v               (10)

    Substituting
                                                                                       d v

-------
                         Electrostatic Precipitators Operation and Industrial Applications
Gives
                         ,2 L W
                           d v
                                  (11)
Figure 8.
             SCHEMATIC SHOWING
             DIMENSIONS.
 If  precipitatdr  efficiency  is  to  be
 increased,  this  may  be  accomplished
 by  reducing spacing  between plates
 d;  increasing precipitator  length L;
 reducing  gas velocity v;  or increas-
 ing the drift velocity  W.

 Economic  considerations generally
 determine how these  parameters  are
 optimized.

 The  gas velocity v can  also  be  written
 as            L^
          v  = ~t
 where t is  the residence  time of  the
 gas  in the precipitator.  This gives
                                                   From practical considerations, it is
                                                   desirable to make a precipitator as
                                                   small as possible consistent with
                                                   obtaining a satisfactory cleaning
                                                   performance. The physical dimensions
                                                   can be reduced only if the drift
                                                   velocity can be increased to more
                                                   closely approach theoretical values.

                                                   To accomplish this, one must increase
                                                   the average field strength, Ep, and
                                                   improve plate rapping, power, and gas
                                                   distribution.

                                               C.  Particle Size and Dust Concentration

                                                   A typical particle size analysis be-
                                                   fore and after an electrostatic pre-
                                                   cipitator shows a distinct difference,
                                                   with the dust after the precipitator
                                                   being much finer (see Figure 9).
                                                            Range of Particle Size
                                                                    In Clean Gas
                                                  4-1
                                                  3
                                                 .
                                                 -rl
                                                  1-1
                                                                          Gas
                                                        Particle Size, microns

                                               Figure 9.   PARTICLE  SIZE DISTRIBUTION
                                                          IN DIRTY  AND CLEAN GAS.
                                                   The particle size analysis and the
                                                   overall collecting efficiency enable
                                                   the grade efficiency curve of the
                                                   precipitator to be determined. From
                                                   this,  the effective migration velocity
                                                   for a  given particle size can be
                                                   calculated.
             -(2 t W)
                               (12)
Obviously, the longer the gas is ex-
posed to the electrical forces,the
higher the efficiency. This is only
done, however, by decreasing the cap-
acity, since less gas will be handled
in a given precipitator.
                                                           E Average1
                                                    Fraction Size
                                                    Efficiency
                                                            W Average
                                                    Migration Velocity
JO

M

12

8


04-
                                                Particle Size, microns
                                                 Figure 10.  CURVES OF GRADE EFFICIENCY
                                                             AND MIGRATION VELOCITY.

-------
Electrostatic Precipitators  Operation and Industrial Applications
        This empirically determined relation
        of the W-value to the particle dia-
        meter is dependent on the dust content.
        If the dust content is small, the
        rated migration velocity has to be
        reduced, since the probability of
        the dust being charged, precipitated
        and collected in the hopper is smaller.
        On the other hand, when the dust
        content is very high, the flow of
        the current between the electrodes
        is reduced, and so is the current
        density, which also requires a low
        migration velocity. This space charge
        effect will increase both with the
        concentration of the dust and also
        with its degree of fineness.

    D.  Gas Velocity

        Depending on the design of the col-
        lecting surface, all dry-process
        electrostatic precipitators have a
        natural mechanical (inertial and
        gravitational) collection efficiency,
        which will increase with the gas
        velocity to a maximum and then de-
        crease due to higher re-entrainment
        losses. Therefore, each collecting
        plate has an optimum point for the
        gas velocity, which is inherent with
        its design and also with the pertinent
        characteristics of the dust and the
        carrier gas. The mechanical collecting
        efficiency can be as high as 40 to
        50 percent. The optimum gas velocity
        can practically be used when the
        specific application allows a high
        migration velocity to be used.

    E.  Electrical Wind

        It is well known that the electrical
        wind is an attendant phenomenon with
        corona discharge.  The glow point on
        the wire or on a sharp barb can be
        regarded as a tiny nozzle in which
        the gas molecules  are accelerated to
        a considerable velocity,owing to the
        high field strength,so that a high-
        pressure zone builds up in front of
        it and a low-pressure zone behind it.
        The hissing noise  which accompanies
        this discharge supports this concept.

        The measured wind  velocities attain-
        ed considerable values.  The curves
        for the barbs in two directions and
        in a single direction showed that
        the latter arrangement caused a
        considerably more  powerful flow of
        air in the direction to  the plate.

                                  -fe]
Figure 11.  ELECTRICAL WIND VELOCITY
            A.  BARBED WIRE IN ONE
                DIRECTION.
            B.  BARBED WIRE IN TWO
                DIRECTIONS.
            C.  STAR-SHAPED WIRE.
     Results obtained with dust-free gases
     show that the wind caused by pointed
     discharge electrodes had a velocity
     near the collecting surface comparable
     to the average gas velocities in in-
     dustrial electrostatic precipitators.
     Thus, an electrically charged particle,
     which follows the force lines towards
     the plate due to the eletrostatic
     field, receives an additional impulse
     in the same direction. The resultant
     of the two directions of motion, per-
     pendicular and parallel to the collect-
     ing surface, is a preferential direction
     of precipitation, which because of the
     direction of the gas leads to non-sym-
     metrical accumulations of dust on the
     plate.
 F.   Dust Resistivity

     Industrial precipitators normally
     operate in the temperature range of
     200 to 400°F,  with less favorable
     electrical conditions for precipitation
     towards the higher end of this range.
     With higher temperature, the break-
     down voltage decreases, the viscosity
     of the carrier gas increases, and in
     the temperature range of 200 to 320°F
     resistivity of most of the dusts in-
     creases.

     Specific  dust  resistivities range from
     104 to 1014 ohm-cm. From 104 to lOll
     no precipitation problems occur. Dust
     resistivities  higher than 10H ohm-cm
     cause the accumulated dust on the
     collecting surface to act as an in-
     sulator.

-------
                          Electrostatic Precipitators Operation and Industrial Applications
 Permanent Layer of Dust
 tu
 a
 a
 M-l
 M
 en
 oo
 •r-l
 -fcJ
 o
 a
Dust Layers of
various densities
in
3
Q
         Thickness
 Figure 12.  DUST DEPOSIT
 The electrical charges of the dust
 particles create a potential and thus
 an intense electrical field in this
 layer of dust, which at certain points,
 for example at enclosed pores of gas,
 forms an electrical ionization. A
 corona discharge, similar to the one
 at the discharge electrode.occurs.
 This so-called secondary ionization
 reduces the breakdown voltage con-
 siderably, with a simultaneous increase
 in current. The collection efficiency
 Is also reduced considerably.

 Practically all of the dusts have
 certain semi-conductor characteristics,
 i.e. an electrical resistance, varying
 with temperature and moisture content
 of the carrier gas. Normally in the
 lower temperature range, the dust
 resistivity is determined by the
 .surface conductivity, which  is a
 function of the adsorbed or  chemically
 combined moisture.
                                                 10
                                                  .13
                              12
                            10
                                                 10" •
                                                  10
                                                   Dew Point
                                                                          Gas Temperature
                                                           100
                                            200    300
400
                                                    Figure 13.
                                           EFFECT OF TEMPERATURE AND
                                           MOISTURE ON RESISTIVITY
                                           OF THE DUST
                                With increasing temperature the mois-
                                ture associated with the dust evapo-
                                rates, causing the dust resistivity
                                to increase.  With a further increase
                                in temperature the conductivity of the
                                dust particles increases and this causes
                                the dust resistivity to drop again.

                                If the dry dust is cooled in a dry
                                atmosphere the dust resistivity will
                                increase over the previously experienc-
                                ed maximum (shown by the dashed lines
                                in Figure 13). This shows the influence
                                of moisture content on dust resisti-
                                vity.
                                The presence  of 803 in the carrier gas
                                favors the electrostatic precipitation
                                process. For  a number of dusts, the
                                resistivity is reduced below the critical
                                limit of 10-*-   ohm-cm with very small
                                amounts of SO-j.

                                      Collecting Efficiency
                                              99%  - Constant
                                   180

                                   160

                                   140

                                   120

                                   100
                                     0 2 4  6  8 10 12 M 16 re 20 22 24
                                          S03, [PPM/VOL]

                                   Figure 14.  TEST OF COLLECTION
                                               EFFICIENCY WITH
                                               S03  INJECTION.

-------
Electrostatic Precipitators  Operation and Industrial Applications
        Tests at a pilot precipitator showed
        that under constant conditions the
        migration velocity doubled with an
        303 injection of 22 ppm. This essential
        improvement was due mainly to a re-
        duction in dust resistivity caused by
        the S03 injection. (See Figure 15).
                        No S0_ - Content
                                       III   SPECIAL  DESIGN CONSIDERATIONS

                                             An  electrostatic  precipitator, requires
                                             the following  basic  components: (1) Dis-
                                             charge system  with high voltage energizing
                                             set,  (2) Collecting  surface,  (3) Rapping
                                             system for  discharge and collecting
                                             electrodes  and (4) Suitable casing with
                                             means to receive  the precipitated dust.
                                     Roaster
                                     Gas
                     SO  Approximately 3mg/Nm

                  Gas Temperature °C
             100
      200    300    400    500
          Figure 15.
         EFFECT  OF  803 ON
         DUST  RESISTIVITY
        The presence of 803 is especially con-
        ducive for the precipitation of dust
        with oxide constituents,  such as Si02,
        MgO, and A1203, which react either
        slowly or not at all with water to
        form hydroxides

        Another conditioning system for dust
        particles is the injection of phos-
        phorous pentoxide (P20^)  or other
        similar agents.
  O
  •H
  O
  01
  G.
     10
     10
 700
                     156    350
                     Gas Temperature
 Figure 16.
DUST RESISTIVITY AND
BREAKDOWN VOLTAGE,
MEASURED IN THE WASTE
GAS OF A HOT-BLAST CUPOLA.
                                                       Figure 17.  ELECTROSTATIC PRECIPITATOR
                                             A.   Discharge Electrodes

                                                 By selecting an electrode system
                                                 especially suited to match the re-
                                                 quirements of a specific application,
                                                 an improvement in collecting effici-
                                                 ency can be obtained.

                                                 With the composition of dust in
                                                 earlier days, discharge electrodes
                                                 with spikes pointed toward the col-
                                                 lecting electrodes offered no ad-
                                                 vantage. Spiked discharge electrodes
                                                 increased the collecting efficiency
                                                 in one installation, but in another
                                                 installation gave an increase in
                                                 the precipitator current with the
                                                 same or even decreased efficiency,
                                                 making its use doubtful.
                                                          Figure 18.   DISCHARGE ELECTRODES

-------
                             Electrostatic Precipitators Operation  and  Industrial  Applications
       As  a  result  of intensive research
       work,  however, it has been establish-
       ed  that  with the barbs suitably
       arranged and of the right length,
       a corona discharge of high intensity
       can be produced. This corona dis-
       charge is capable of penetrating
       the ionized  dust space and to a
       certain  extent the deposited dust
       layer itself.  Thus, the two obstacles,
       high  electrical resistance and
       space ionization, can be sufficient-
       ly  overcome.
               \ \	FYt-clearing barbtd wire
               t  \ — Att*r-cl«3nng itar wwt
Figure 19.  MAX ACR-OVER CURRENT
            AS A FUNCTION OF PRECIPITATOR
            INLET TEMPERATURE. NORMAL
            GAS VOLUME.

       For example in a two field precipi-
       tator operating at nearly constant
       conditions, i.e. constant gas volume
       and raw gas dust content, both
       fields drew a considerably higher
       current after the first field was
       equipped with spiked discharge
       electrodes. The increased current
       was very distinct in the temperature
       range between 200 and 400°F. The
       dust resistivity in this case had
       a maximum of 5 x IflU ohm-cm at a
       gas temperature of 230°F and was
       well above lO^ ohm-cm between 170
       and 380°F.

   B.   Collecting Electrodes
       To achieve optimum efficiency the
       collecting electrodes should meet
       the following requirements:

       1.  An electrically smooth surface
           giving a high breakdown voltage.
       2.  A good mechanical collecting
           efficiency to prevent deposited
           dust from reentrainment into the
      gas flow during continuous op-
      eration.
  3.   A high absorption capacity to
      prevent reentrainment  of dust
      during or shortly after rapping.
  4.   Favorable vibration character-
      istics to dislodge the accumu-
      lated dust when the plate is
      rapped.

  Besides these basic requirements
  for  the operation  of the precipi-
  tator,  collecting  plates should
  combine a high physical strength  to
  withstand the influences of the
  environment,  have  enough rigidity
  to stay aligned, and have  a low
  weight  per area ratio for  easy trans-
  portation and handling.
Figure 20.  COLLECTING PLATES

  The development  in the  past ten
  years  included the step from the
  box type electrode (Figure  20a) to
  the folded electrode (Figure 20b)
  and finally to the roll formed
  electrode (Figure  20c).

  Comparing the characteristics of
  these  electrodes,  using the basic
  requirements as  criteria, it was
  found  that  the breakdown voltage of
  the roll formed  electrode is sub-
  stantially  higher  than  for  the box
  type electrode.  The  folded  electrode
  with its long flat center part  less
  the highest breakdown voltage,  but
  it  is  much  less  effective in pre-
  venting  reentrainment and has less
  adsorption capacity.
  The  box  type  electrode  is  only
  slightly better  in  terms of  reen-
  trainment  than the  roll formed
  electrodes, but  this  disadvantage
  of the  roll formed  electrodes  is
  more than  compensated for  by the
  better vibration characteristics.

-------
 Electrostatic Precipitators  Operation and Industrial Applications
           Collecting  plates  should  have  good
           dust  retaining  and absorption  char-
           acteristics.  A  picture  representing
           the gas  flow  in a  single  passage can
           be obtained by  using  a  water flow
           model.  (See Figure 21)  The quiescent
           zones formed  by the bends of the
           electrode outside  of  the  main  gas
           flow  serve  to collect the major por-
           tion  of'the dust and  shield it on
           the way  down  into  the hopper.
                                                                           4,4m -
Figure 21.   GAS  FLOW USING A WATER FLOW MODEL.

           The last  requirement  for a  collect-
           ing electrode  is  a  favorable vibra-
           tion  characteristic.  The impacts  of
           the rapping  system, which consists
           of a  hammer  arrangement on  a rotat-
           ing shaft  at the  lower  end  of the
           collecting surface, have to be trans-
           mitted over  the  total area.  This
           arrangement  has  the advantage in
           that  the  dead  weight  of the plates
           and the accumulated dust do not need
           to be lifted by  the rapper.
     Figure  23.   VERTICAL ACCELERATION
                 OF PLATE DURING RAPPING

        Furthermore it  leads  to  an accele-
        ration of  the plates  vertically to
        the surface, which  is of utmost
        importance for  the  dislodging of
        the accoumulated  dust. The amplitudes
        of this acceleration, measured
        vertically to the surface of  the
        collecting plates,  should be  at
        least 100  g (one  hundred times the
        acceleration due  to the  gravity).
        Otherwise  substantial accumulations
        of dust can occur.

        Collecting efficiencies  for differ-
        ent types  of collecting  surfaces can
        be compared by  plotting  the obtained
        migration  velocity  versus the pro-
        ;duct of migration velocity and
        specific collecting surface.  For a
        given value of  wf,  it can be  noted
        that the roll formed  electrodes give
        higher wf  - values  than  other pocket
        or flat plate electrodes.
                                                                  10 50   80 90  9i 97 90  99
                                                                     COLLECTING  EFFICIENCY
                               Figure 22.
                               RAPPER SHAFT
                               WITH BEARING
Figure 24.  COLLECTING EFFICIENCIES FOR
            DIFFERENT TYPES OF COLLECTING
            SURFACES.
 10

-------
                           Electrostatic Precipitators Operation and  Industrial  Applications
C.  Rapping Cycle Control


    Collecting electrodes engineered to
    give the maximum absorption capa-
    city, (maximum dust collection in
    pockets not subject to reentrainment)
    account for approximately 60 per
    cent of the dust. The remainder will
    collect on the outside of the pockets
    and can be reentrained under the
    influence of the rapping of the plates.


    From the  moment  the  particle  or  ag-
    glomerate falls  away from the  plate,
    it  is  subject  to the effect of grav-
    itational force  (settling velocity
    Vf)  and  the  velocity of  the gas  V.
    The  path  followed  by the dust  part-
    icle in  the  direction of the  hopper
    will thus correspond to  the result-
    ant  R shown  in Figure 25. If a line
    is  drawn parallel to R at the far
    end  of the precipitator  outlet,  one
    can find which part  of the dust
    collected on the collecting plates
    will drop into the hopper,  and which
    part will be carried out of the pre-
    cipitator by the moving  gas stream.
   f, . AREA OF LOSS
   H  RESULTANT
V • GAS VELOCITY
V, - SETTLING  VELOCITY
Figure 25.  TWO-STAGE ELECTROSTATIC
            PRECIPITATOR.


     The  area of loss can be expressed
     as a loss factor to become part of
     a revised W - value. It should be
     pointed  out that the area of loss
     only affects that part of the dust
     exposed  to the gas  stream and not
     the  part settling inside a hollow
     plate or In the dead air space of
     a profiled plate, in which case
     the  vector v is zero, since the
     dust travels down a quiescent path
     Into the hopper. The retaining cap-
     acity of an electrode is thus a
     factor of major importance.
                                        Referring again to the loss area,
                                        it is evident that the size of this
                                        area depends on the slope of  the
                                        settling line R. At constant gas
                                        velocity, as the settling velocity
                                        increases, the size of the loss area
                                        decreases and the collecting effi-
                                        ciency is improved. The heavier
                                        the agglomerates are,  the higher
                                        the probability that they will drop
                                        into the hopper before being carried
                                        out of the precipitator.  With respect
                                        to the efficiency of the  precipitator
                                        it would be favorable  if  the dust
                                        could be dislodged in  larger agglo-
                                        merates. This then requires a
                                        specific thickness of  the dust layer
                                        accumulated on the collecting plate
                                        and also a rapping intensity adjusted
                                        to this thickness.

                                        The results of a rapping  test using
                                        different time intervals  can be
                                        shown by plotting the  current of a
                                        photo-cell located in  the outlet of
                                        the precipitator as percentage of
                                        variation Against the  rapping cycle.
                                        Since the currents are inversely
                                        proportional to the dust  burden,
                                        the lowest dust content and highest
                                        collecting efficiency  respectively
                                        are obtained in the peak  of this
                                        curve.
              4O IOO BO 2OO ZiQ \ 3OO


                  RAPPING |
Figure 26.
                                                 RAPPING CYCLE  ["«*JT£jr

                                              RAPPING CYCLE  OF
                                              ELECTROSTATIC  PRECIPITATOR
                                              FOR CLEANING WET BOTTOM
                                              BOILER GASES.
                                   D.  Electrical System Control

                                       Electrostatic precipitators operate
                                       with high voltage direct current,
                                       normally between 25 and 60 KV, and
                                       in special cases up to 100KV. The
                                       current usually is in the milliamp
                                       range with exceptions up to one or
                                       two amps.

-------
Electrostatic Precipitators  Operation and Industrial Applications
          Basically an energizing set for an
          electrostatic precipitator has the
          following components:

          1.  high voltage transformer
          2.  rectifier
          3.  regulating unit
          4.  control unit

          High-voltage transformers do not
          present any special problems as far
          as  their use for an electrostatic
          precipitator is concerned.

          Various types of rectifiers have
          been  used starting with rotary syn-
          chronous mechanical rectifiers, which
          were  replaced by vacuum tube rect-
          ifiers in the early 1950's. Almost
          simultaneously suitable selenium
          rectifiers were developed. Today
          silicon rectifiers are used, which
          normally are installed in the trans-
          former tank, thus giving a compact
          and reliable equipment.
                Step-Up
                Transformer
               Rotary
               Synchronous
               Rectifier
      Figure 27.
ROTARY SYNCHRONOUS
MECHANICAL RECTIFIERS.
         When electrostatic precipitators
         were first put into operation, it
         became apparent that proper adjust-
         ment of the electrical input was
         required to obtain maximum perform-
         ance. At first these adjustments
         were made manually by personnel
         assigned to the task.

         In recent years automatic control
         devices have been developed to main-
         tain the desired operating point.


          Since with increasing voltage the
          corona discharge and the field in-
          tensity increase, the precipitator
          operation is highly a function of
          the applied voltage. The voltage is
          limited to the breakdown voltage
          between the two electrodes and has
          to be controlled to prevent exceed-
          ing the breakdown level. This level
                                             is  not constant and varies with dust
                                             concentration,  dust composition, gas
                                             composition,  gas temperature, pressure,
                                             mositure content,  process variations,
                                             etc.

                                             There  is no other  practical way to
                                             determine the breakdown level than
                                             to  reach this level In operation.
                                             The control unit of the energizing
                                             set therefore brings the operating
                                             voltage  to the  breakdown level, then
                                             reduces  it to a lower level and
                                             raises it again. The reduction of
                                             the voltage has to be done as fast
                                             as  possible whereas the rise is at
                                             a much slower rate. In this way the
                                             voltage  is controlled in a zigzig
                                             line along the  breakdown level.
                                    o a)
                                    -d oo
                                    .*! n)
                                    nj 4-1
                                    <0 i-l
                                    n o
                                    M >
                      Precipator
                      Voltage
                                                                                    Time
                                                         Figure  28.   AUTOMATIC VOLTAGE CONTROL
The rate at which the breakdown
level is reached greatly varies for
different .applications. The slope
of the rise of the voltage must
therefore be adjustable over a
broad range. In some installations
the breakdown level is reached more
than a hundred times in a minute,
whereas in others, it is only reached
once in several minutes. When a flash-
over occurs due to a voltage in
excess of the breakdown level, the
voltage in the precipitator drops
practically to zero and so does the
collecting efficiency. To reduce the
loss, the control has to extinguish
the flash as soon as possible and
bring the voltage in the precipita-
tor back up again.

In a single phase bridge current,
the arc-over extinguishes Itself
when the voltages pass through zero.
For this reason most of the pre-
cipitators are energized with single
phase circuits.

The regulating units for the trans-
12

-------
                              Electrostatic Preclpitators  Operation  and Industrial  Applications
       former rectifier began with a simple
       resistor in series, followed by a
       step-up transformer, an auto-trans-
       former, a saturable core reactor
       (transductor), and finally a thyr-
       istor. Either auto-trans formers
       or saturable core reactors combined
       with an automatic control are con-
       sidered standard equipment for an
       electrostatic precipitator. The
       only difference between the two is
       that the auto transformer controls
       the voltage, whereas the saturable
       core reactor controls the current.
       Although voltage control is generally
       preferred, it is of minor importance
       when operating close to the break-
       down voltage. A current controlling
       amplifier has the advantage of limit-
       ing the current in case of a short
       curcuit.

            Step-Up
            Transformer
AC
Supi
Saturable 8
Rector §

poooo
lv 1 I
0 0


1
' Rectifier
Precipitator
Control
Figure 29.   PRECIPITATOR POWER CONTROL
            BY USE OF SATURABLE REACTOR.
The block diagram shows a typical
precipitator control circuit.

Input power is applied through
suitable switching and protective
devices. These provide for compli-
ance with electrical code require-
ments and for safety and maintenance
considerations. Delivery of power
to the step-up transformer and
rectifier is modulated by a main
control element introduced in series
with the input to the transformer.
This element is usually a saturable
reactor, but electronic devices of
suitable rating such as thyristors
can also be used.

The action of the main control
element is determined by a signal
furnished by the amplifier section
of the equipment. The function of
this amplifier is to modify the
signals it receives by Increasing
their amplitude to make them suitable
for driving the main control element.

Direct manual control may be
accomplished by setting components
of the amplifier section or by
introduction of a manually-control-
led signal in place of the amplifier
section.

In an automatic system the signal
conditioning network compares the
Switching
13°wer and
Input Protective
Equipment
Direct
Manual
Control
Preset
Signals



— »-
— »•
Main
Control
Element


Step-Up Rectifier
Transformer (Tube or
j Semi- Con-
i ducted)
1
i I
! L

Amplifier


Signal
Condition-
ing
1
1
1
1
(Precipit
1
1
1
1
1
Feedback j
^ . J
Signals

ator

I

           Figure 30.  GENERALIZED BLOCK DIAGRAM - INDUSTRIAL PRECIPITATOR CONTROL.
                                                                                             13

-------
Electrostatic Precipitators  Operation and  Industrial Applications
          signals  from  the  actual  operation
          of  the precipitator  against preset
          signals  and the difference is  fed
          to  the amplifier.

IV  OPERATION OF AN ELECTROSTATIC  PRECIPITATOR

    The operation  of an electrostatic precip-
    itator, when properly designed and  in-
    stalled,  will  always follow some basic
    mechanical and electrical  laws. From the
    available know-how  it can  be predicted
    which conditions would  favor the operation
    of an electrostatic precipitator and how
    these conditions can be achieved.

    Once an electrostatic precipitator  is
    built, its design  and physical dimensions
    can only be changed with considerable
    efforts and expenses.  This leads  to the
    question of how the optimum operating con-
    ditions can be established and maintained.
    For a given installation this  problem can
    be split into four basic parts:

    (1) Installing and  operating a suitable
    gas and dust  conditioning  system if  re-
    quired, (2) achieving and  maintaining an
    effective gas  distribution ahead of  the
    electrostatic  precipitator, (3) establish-
    ing an optimum rapping  cycle for the
    collecting and discharge electrodes, and
    (A) operating  the precipitator with  an
    adequate  electrical control system  for
    the energizing sets.

    A.  Gas Conditioning

        Electrostatic precipitators for some
        applications,  such  as  the  cement in-
        dustry, steel mills, and some metallur-
        gical furnaces, will only  operate
        satisfactorily  If the  gas  temperature
        at the precipitator inlet  is  reduced
        to a specific level and the water dew
        point raised to a required level.
     L<5
       20
                 GOOD'
            100   no   120
                              POOR
                        130  140  150   160 °C
                           TEMPERATURE —
   Figure 31.   AREA OF  GOOD  OPERATION AS  A
               FUNCTION OF DEW POINT AND
               TEMPERATURE.
   This lowering of the gas temperature
   and simultaneous raising of the dew
   point can be effectively carried out
   by installing an evaporation cooler
   ahead of the dry process precipitator.

   The evaporation cooler uses the princi-
   ple of direct injection of a fine
   spray water.

   Different arrangements can be used,
   but the most suitable design consists
   of a standing cooling tower, using
   the concurrent flow principle, with
   the gas inlet and atomizers arranged
   at the top of the tower.
     1
 J77J
Concurrent Flow
                     Countercurrent Flow
  Figure  32.   GAS CONDITIONING SYTEMS

   The most  suitable  means  of injecting
   the water,  in view of  the large quan-
   tities  Involved, is by atomizers.
o
9-?
0)
O to
C 01
111 N
t-l i-l
t* to
P
O to
O 4J
O 01
• 0.
rH O
01 M
oi a
40.



30.


20-

10



\
I
1
1
I
\
V
N.^ Droplet Diameter »
                                                     01 U
                                                     > 01
                                                     •H *J
                                                     4J n]
                                                     <0 3
                                                     r-l
                                                     01 1W
  40

  30

  20

  10
                                                                  100  200  300   400  500/«
                                                                  100  200   300  400  500^
                                                              Figure 33.
                  SIZE DISTRIBUTION
                  OF WATER DROPLETS
 14

-------
                           Electrostatic I'recipitators Operation and Industrial Applications
   • -.o^ :MiS*0V?!  * '
                   O~s»  «W*
                     .<
  >    v-  °
       », o.  «
                      arvST;,*' •
                        5go
                      «   -v*'o.
   The  required treatment time for  the
   gas  is  given by the time it will take
   for  all the droplets to be evaporated.
   The  smaller the droplets, the faster
   they will evaporate, and this will be
   proportional to the square of the in-
   itial droplet diameter.

   A major aspect of the dimensional de-
   sign of an evaporation cooler is the
   determination of the overall height
   or required volumetric capacity. The
   volumetric capacity then corresponds
   to a required treatment time since
   each cooler of a given volume will
   give a  specific treatment time for a
   constant gas volume, regardless  of its
   diameter.
^1000
t
^ 800
<
i 6°°
»--
5( 400
                              i 1(5') 1.1 K)1
        „
        a.6 a?  as as
                           w  15 -
Figure  34.  CALCULATION OF THE HEIGHT
           OF AN EVAPORATION COOLER

   In addition to the  necessary uniform
   gas and water droplet  distribution
   over the cross section of  the cooling
   tower, a control system has to be pro-
   vided to ensure that a sufficient
   amount of water is  always sprayed in-
   to the system and is always completely
   evaporated. Inlet and  outlet gas
        temperatures at the evaporation  cooler
        as well  as  gas volume or heat content
        of the gas  can be used as parameters
        for the  control system.

    B.  Gas Distribution

        Depending on gas and dust conditions
        and the  required collecting efficiency,
        the gas  velocities in an industrial
        electrostatic precipitator are between
        2.5 and  8.0 fps. A uniform gas dis-
        tribution is of prime importance for
        the precipitator operation, and  it
        should be achieved with a minimum
        expenditure and pressure drop. This
        is not always easy, since gas veloci-
        ties in  the duct ahead of the pre-
        cipitator may be 30 to 100 fps in
        order to prevent dust build-ups. Com-
        paring the  cross-section of the inlet
        duct works  and precipitator, ratios
        up to 1:10  or 1:20 have to be coped with.

        Suitable distribution models have to
        be tested and the results related to
        the full size unit considering the
        laws of  dynamic similarities.

        The results of these tests can be
        plotted  on  coordinates in the open
        duct or  precipitator area. These so-
        called velocity profiles can now be
        compared and the most suitable one
        chosen for  the final installation.
        The results of these tests may be so
        close together, that it is quite dif-
        ficult to decide which arrangement to
        use. For this purpose, a calculation
        program  can be used, which sets  a
        scale by establishing comparable local
        collecting  efficiencies for different
        parts of the cross-sectional area of
        the precipitator.
                                              Measuring
                                              Plane  I
           uannuajaaa.
           luonnnnaa
           iJiauntuuuL
                                                                         tfr^-^
                                                                       II
Measuring
Plane I
                                                        jUDHaaanm i
                                                        LantmaaaaJ
                                                        LDDUOJBJUULI
                                                                         iJLHanjannnj
      Figure  35.  GAS DISTRIBUTION PROFILES
                 AFTER FLAP TYPE DEFLECTORS
                                                                                        15

-------
Electrostatic Precipitators Operation and Industrial Applications
 V  TYPICAL  INDUSTRIAL PRECIPITATOR APPLICATIONS

    A.   Precipitators for  Flue Gases From
        Power  Stations

        One  of the major applications for
        electrostatic precipitators is de-
        dusting of flue gases from coal fired
        boilers in thermal power stations.
        The  tightening of  regulations for the
        abatement of air pollution is confront-
        ing  many power stations with the task
        of installing high efficiency dust
        collectors. The trend is towards using
        low  grade coal with  a high ash content
        and  with the considerable increase in
        the  size of the boilers the problem
        becomes even more  significant. When
        boilers were relatively small compared
        to present standards, the dust emission
        did  not attain such  high proportions.

        Modern firing systems, in connection
        with the variety of  fuels used, con-
        front  the precipitator maufacturer
        with new problems  which require con-
        tinuous extensive  investigations.

        The  precipitator installation in Fig-
        ure  36 is designed for a boiler with
        a  total capacity of  770 tons of steam
        per  hour. The gas  volume from this dry
        bottom,  pressurized boiler is above
        500,000 scftn. The  collecting efficiency
of the two electrostatic precipitators
is 99.1 per cent with a guaranteed
residual of 0.044 gr/scf.


The two units in Figure 37  have  a
turbine power of 250 MW each.  Four
Babcock boilers with cyclone burners
and molten ash discharge generate
236,000 scfm into each one  of  the four
electrostatic precipitators. Tests
showed that the guaranteed  collecting
efficiency was met with a clean  gas
dust residual of 0.0326 gr/scf.

The introduction of boilers with wet
bottom ash discharge or with cyclone
burners necessitated dealing with very
small particles with up to  80  per cent
less than 10 microns. This dust  has
also a number of other unpleasant
characteristics. It is difficult  to
remove from the collecting surfaces
by rapping since it is very fine.  Be-
cause of the high temperatures in the
fire box, the dust containes a high
percentage of vaporized (sublimated)
minerals which stick to the plates and
have a high electrical insulating
effect.
                                    Figure 36.  POWER PLANT
16

-------
Electrostatic Precipltators Operation and Industrial Applications
      Figure 37.  POWER PLANT
                                                                17

-------
 Electrostatic Precipitators Operation and Industrial Applications
       This very fine dust also induces a
       space charge effect. Both the space
       charge and the insulating layer of dust
       on the collecting surfaces reduce the
       current flow. Thus the desired collect-
       ing efficiency can be attained only if
       the size of the collecting surface is
       correspondingly increased.

       Due to the very steep slope in the peak
       of the resistivity curve, changes of
       20 to 40°F in gas temperature can affect
       the operation of an electrostatic pre-
       cipitator considerably. (See Figure 39).
   5
   o
   f-
   i/1
   Co
   LJ
   CC
   U~l
   D
   D
          140     200         30O         4OO

              GAS  TEMPERATURE.  °F


          39.  COM?AR[SON OF DUST
              KI-.SISTIVITY IN CYCLONE
              BOILER AND SF.AG TAG 1501LER.
       Jn  general,  the  size  of  an  electrostatic
       p rt'fi p 1 tator will  largely depend  on
       1 ] ue  gu.s  volume,  the  required  efficiency,
       the firing  system  used,  the composition
       df  thf  coal  (sulfur,  moisture,  volatile
       matter, ash),  flue gas temperature,
       i-oke  content of  the dust, particle
       size, electrical resistance of  the dust,
       ;)nd the dust content  of  the flue  gas.

       Despite the  possibility  of  future com-
       plications,  the  requirements of the  air
       pollution codes  can be met  with a pro-
       perly sized  and  operated electrostatic
       precipitator.  Flue gas precipitators
       with  efficiencies  up  to  99.9 per  cent
       and more  are in  operation.

   li.   1'recipitntors  for  the Iron  and  Steel
       industry
       1 .   Sintering Plants
    Sintering has proved to be a highly
    effective method of agglomerating
    fine ores before they are charged
    into the blast furnace. There  are
    two main dust producing sources  in
    a sintering plant:  the sintering
    belt and the sinter crushing and
    conveying system.

    Modern sinter belts have capacities
    up to 10,000 tons per day and  the
    exhaust gas volume may be up to 1.2
    million cfm. Dust contents range
    from 0.22 to 1.3 gr/acf with 70
    per cent of the dust having a  part-
    icle size of less than 20 microns.
    Gas temperatures range from 210 to
    320°F with dew points ranging  from
    90 and 120°F. The dew point influ-
    ences the electrical resistivity
    of the dust and thus the collecting
    efficiency. In general low dew
    points complicate the operation.

    As an example of this application,
    a precipitator for a sinter belt
    with a capacity of 5,000 tons  per
    day is shown in Figure 40. The gas
    volume is approximately 600,000 cfm
    and the clean gas dust residual is
    less than 0.01 gr/acf.

    Precipitators for sinter crushing
    and conveying systems have to  deal
    with dust inlet loadings up to 11
    gr/acf. Electrical resistivity of
    the dust increases sharply with the
    air temperature and can reach
    critical values. Residuals down to
    0.01 gr/acf can be achieved.

2.  Coke Oven Plants

    The crude gas from coke oven plants
    contains tar, which is removed by
    condensing the gas and collecting
    the droplets with an electrostatic
    precipitator. Residual dust concen-
    trations in the clean gas after the
    precipitator can be as low as  0.0005
    gr/scf.

3.  Blast Furnace Plants

    The application shown in Figure 42,
    called the Venturion filter, is
    unusual in the U.S.A. The installa-
    tion is a combination of two diff-
    erent dust collecting processes
    giving an extremely high collecting
    efficiency.
18

-------
                                Electrostatic Precipitators Operation and Industrial Applications
                                   Figure  40.   SINTERING  PLANT
i BIBI^BBlMM
                                                         Figure 42.  LONGITUDINAL SECTION
                                                                     THROUGH PRESSURE TYPE
                                                                     VENTURION PRECIPITATOR
                                                                                                19

-------
 electrostatic  Precipitators Operation and  industrial AppHcatk
           The raw gas enters the Venturi
           washer and is greatly accelerated
           at its narrowest cross section.
           Water is injected simultaneously
           at this point to obtain an intensive
           mixing of gas and water.  The dust
           laden gas is washed to a large ex-
           tent and cooled. After a 180°  change
           of direction the gas enters the
           electrical fields of the electro-
           static precipitator. The cleaned
           gas is directed to the secondary
           Venturi stages installed behind  a
           partition on both sides of the
           electrical fields. The secondary
           Venturis are primarily used to cool
           the gas further. They also have  a
           slight scrubbing effect.

           Before the cleaned gas leaves  the
           Venturi filter, its pressure is  re-
           duced to the desired pressure.
           For example, in a pressurized
           blast  furnace  installation in South
           Africa designed for 225,000 scfm,
           the clean gas  residual is guaranteed
           for less than  0.0044 gr/scf.

       4.  Steel Works

           a.  Open Hearth Furnaces
    In this application  the  waste
    gas from the open  hearth fur-
    nace is cooled by  waste-heat
    boilers and/or water sprays.
    Dust loadings during an  oper-
    ating cycle vary widely,  rang-
    ing from as little as  0.1 g~f
    cf to as high as several  grains
    per cubic foot during  the
    oxygen blow period.  The mois-
    ture content of open hearth
    gas is about 7 to  8  per  cent
    by volume with a dew point in
    the neighborhood of  100°F.
    Because the extremely  fine
    dust (60 per cent  less than 5
    microns), care should be  taken
    to prevent re-entrainment.
    The dust is readily  precipitable,
    but the moisture content  must
    be maintained at a certain
    level.

b.  Basic Oxygen Processes

    It is general practice to  burn
    the CO gases with  slight  excess
    air in waste-heat boilers. It
    is immaterial whether a dry-
    operated precipitator with an
    evaporation cooler or a wet-
    operated precipitator with a
                                   Figure 43.   BLAST FURNACE
20

-------
                  Electrostatic  Precipitators  Operation  and  Industrial  Applications
saturator is installed. In
those instances where it ap-
pears to be of economic interest
to recover the dry dust (1.2
to 1.5 per cent of the pig
iron charge with an iron con-
tent of approximately 70 per
cent), and where the provision
of water is costly, it is log-
ical to use a dry operated
precipitator.

If, however, the plant should
already have a waste water
purification system, it is
advisable to install the smaller
wet operated electrostatic
precipitator.

Waste heat boilers entail con-
siderable capital outlay, and
the value of the steam generated
does not always warrant the
large investment.
C.
        For this reason, alternative
        solutions are sometimes favored:
        radiant heat boilers and jack-
        eted cooling stacks. Here too
        it is immaterial whether the
        subsequent electrostatic de-
        dusting is done by the dry or
        wpt process.

        For example, a dry process
        precipitator with an evaporation
        cooler for the waste gases of
        an LDAC converter will dedust
        the gas below the visibility
        limit of the so called brown
        smoke.  The capacity of the
        converter is 80 tons and the
        gas volume is between 88,000
        and 95,000 scfm.

Precipitators for the Cement Industry

In the cement industry electrostatic
precipitators can be used to dedust
                 Figure 44.  BASIC OXYCKN PROCESS
                                                                                 21

-------
 Electrostatic Precipitators Operation and Industrial  Applications
       waste air or waste  gases  from  raw  mat-
       erial drying, crushing  and  grinding
       processes, rotary kilns,  shaft  kilns,
       and cement mills.

       Normally dry process  plate-type pre-
       cipitators are used,  the  casing made
       either of steel or  concrete. Corrosive
       gases may require the use of aluminuir
       alloy internal parts.
The d imcMisions of  the precipitator  de-
pend on the required collecting  effi-
ciency, gas volume, and  the  process
used for the production  of the clinker,
especially the gas temperature and
moisture content of the  waste gases.
The operating temperatures of the
precipitator may range from  390  to
570°F. if the kilns are  charged  with
wet material and between 210 and 355°F
for the semi-wet process.
                                    Figure 45.  CEMENT PLANT
                                           Figure5 46.
22

-------
                              Electrostatic Preclpitators Operation and Industrial Applications
    If a dry process is used, the dust can
    be collected either at temperatures up
    to 660°F in relatively large precipi-
    tators or at temperatures from 195 to
    285°F in smaller precipitators by first
    cooling and humidifying the gases. In
    both cases the collected dust should
    be completely dry.

    Waste gases from dryers and grinding
    mills at temperatures of about 210°F
    can easily be dedusted with electro-
    static precipitators.

0.  Precipltator for the Chemcial and the
    Non-Ferrous Metallurgical Industry

    This application deals mostly with the
    recovery of valuable metallic oxides
    and salts from the waste gases from
    metallurgical processes, e.g. shaft
    furnaces, converters, stationary or
    rotating reverberatory furnaces, re-
    fining furnaces, electric furnaces,
    roasters (multiple hearth, rotary and
    fluidized bed furnaces), and sinter
    machines for lead and zinc ores.

    The gases from these applications are
    normally cooled in waste heat boilers,
    air coolers, evaporation coolers, or
Venturi scrubbers and then dedusted in
either pipe or plate type precipitators.

Normally a single stage cleaning system
is sufficient, but for gases with a
recoverable sulfur dioxide content
(e.g.  roaster gases), multiple stage
cleaning systems are used.

Since  most of these gases are highly
corrosive, precipitators operating
close  to the acid dew points of the
gases  have to be made out of corrosion
resistant materials, such as lead,
plastics or ceramics.

1.  Electrostatic Precipitators for
    Roaster Plants

    Pyrite (FeS2) is fed into a turbu-
    lent layer roaster.  The  exhaust
    gas S02 containing is normally
    cooled to 570 to 750°F in a waste
    heat boiler and followed by a dry
    process precipitator designed for
    the elevated gas temperatures. If
    the dust content is  too  high,
    cyclones are used ahead  of the pre-
    cipitator.
                           Figure 47.  COLLECTION OF SULFURIC ACID
                                       MIST FROM ROASTER PLANTS.
                                                                                             23

-------
                             ]ectrostatic Prccipltators Operation and Industrial Applications
Older incinerator installations, even
those using electrostatic precipitators,
received frequent complaints about burn-
ed or partially burned paper particles
which led to the addition of mechanical
collectors, such as arrestor plates after
the precipitator. Tests have, shown that
these particles are a prime subject for
reentrainment unless directly precipitat-
ed into a pocket or a dead air space of
the collecting surface. This required
the design of a special collecting plate
which eliminates this problem but must
be used with the proper gas velocity in
the precipitator. The gas velocity in
the electrical field has to be high
enough to cause an eddy current in the
pocket, but also low enough to prevent
re-entrainment from the flat parts of
the collecting surface.

The heat content of the gases exhausted
from the incinerator can be used in a
waste heat boiler to be converted into
steam either for heating or power genera--
ting purposes. The economics of such a
system will vary considerably and may
not be justified for smaller incinerators.

The flue gases leaving the incinerator
with temperatues up to 2200°F can be
cooled either directly or indirectly.
Indirect cooling using water or air re-
quires a substantial investment with
high maintenance costs.

1.  Municipal Incinerators with Waste-
    Heat Boilers

    The refuse incinerator installation
    includes eight electrostatic precip-
    itators for the waste gases of four
    large incinerators with waste-heat
    boilers. Burning municipal refuse
    with a lower heat content of 1500
    kcal/kg resulted in a dust content
    between 0.45 to 2.2 gr/scf. Gas
    temperatures at the inlet of the
    precipitator varied between 450 and

    510"F with a water dew point between
    115 and 145°F. When operating at de-
    sign conditions, i.e. 440 tons of
    refuse per day for each incinerator,
    the gas volume of each precipitator
    is approximately 141,500 cfm.

    Tests  showed clean gas  dust residuals
    between 0.0074 and 0.026 gr/scf with
    collecting efficiencies between 98
    and 99.5 per cent.  These tests  showed
    that  the operating conditions varied
    considerably with variations  in the
    refuse analysis.
                             Figure 49.  MUNICIPAL INCINERATOR
                                                                                           25

-------
Electrostatic. Precipitators  Operation and  Industrial Applications
                           Figure 50.   INCINERATOR COAL-FIRED BOILER
    2.  Combination of Incinerator and Coal-
        fired Boiler

        A combined central heating and power
        station is designed to burn either
        refuse or bituminous coal, or a vari-
        able mixture of both.  Both firing
        systems are located in separate cham-
        bers and the flue gases pass a common
        boiler and are subsequently dedusted
        in an electrostatic preclpitator.

        Test after commissioning showed com-
        bined dust collecting  efficiencies up
        to 99.9 per cent.

    3.  Industrial Incinerator with Waste-
        heat Boiler

        The industrial incinerator installation,
        which includes a waste heat boiler,
        is mainly fired with paper, wood,
        cardboard, and a substantial amount
        of rubber, artificial  leather and
        rubbish.

        Tested clean gas dust  residuals aver-
        aged 0.035 gr/scf or 0.0175 gr/acf.

        The water dew point was approximately
        104°F which is considerably lower
        than for municipal refuse. The effi-
        ciency of the electrostatic precipi-
tator is clearly related to the burn-
ing process. If for example carbon
black is formed, the efficiency will
decrease. It is important that enough
air is available to achieve complete
combustion.

Direct cooling can be done by simply
mixing false air with the gas but
unfortunately the gas volume increases
3.8 times when gases are cooled down
from 1800 to 570°F. All equipment
then has to be sized for this volume,
i.e. gas ducts, precipitator, fan and
stack.

Further it has to be considered that
In some air pollution codes the allow-
able stack discharge is specified for
either 12 per cent CO. or 50 per cent
excess air. Due to the dilution of
the air the C02 content in the gas
decreases, the actual allowable dust
residual decreases, too.

Direct  cooling ean also be  donfi by
injecting  water into  the gas  stream
ahead of the  precipitator.  This
water injection will  take place in
an evaporation cooler.  The  gas enters
the cooler with a  temperature of 1800
to 2200°F  and the  water is  sprayed
into it  by high pressure atomizers.
26

-------
                              Electrostatic Precipitators Operation  and  Industrial Applications
      The  amount of heat  required to eva-
      porate the water will reduce the
      gas  temperature to  approximately 570°F.
                                              A cooling system using water sprays
                                              to cool gases down from 1800°F to
                                              570°F will increase the standard gas
                                              volume by a factor of approximately
                                              1.4 but, the actual gas volume will
                                              decrease by a factor of approximately
                                              0.6.

                                              Since the installation costs for an
                                              evaporation cooler are quite high,
                                              sometimes in the range of the elect-
                                              rostatic precipitator itself, part of
                                              the incinerator can be used as a cool-
                                              ing zone.
Figure 51.
INDUSTRIAL INCINERATOR WITH
WASTE HEAT BOILER
  4.  Industrial Incinerator without Heat
      Recovery

      The gas cooling system ahead of the
      precipitator uses direct cooling by
      water injection into an evaporation
      cooler. Gas inlet temperatures are
      between 1800 and 2200°F; the outlet
      temperature is held constant at 570°F.
      An automatic control system for the
      water sprays is provided.
   Figure 52.   INDUSTRIE INCINERATOR
                WITHOUT HEAT RECOVERY
      The electrostatic precipitator has
      an outlet gas dust residual of less
      than 0.022 gr/scf.
                                      Figure iJ.  COOLING SYSTEM IN AN INCINERATOR
                                              Due to the limited size of LUC in-
                                              cinerator, cooling to 570°F cannot
                                              be achieved within the incinerator,
                                              but the gas can be cooled to approxi-
                                              mately 900 to 1100°F by water sprays
                                              installed at the end of the incinera-
                                              tor and further cooled with air to
                                              570°F. Thus the standard gas volume
                                              entering the precipitator will be
                                              approximately twice the gas volume
                                              leaving the incinerator.

                                              The moisture content of the gas at
                                              the inlet of the precipitator is
                                              approximately the same as at the outlet
                                              of the incinerator.
                                      VI  Summary

                                          The problem of air pollution is one which
                                          grows with out modern civilization and,
                                          generally, is a direct result of it. The
                                          separation of suspended particles from
                                          gases is one of the basic scientific and
                                          technical problems of the industrial era.
                                                                                             27

-------
 Electrostatic Precipitators Operation and Industrial Applications
     Electrostatic precipitators,  which can
     be used for a wide range of applications,
     can solve the problem of collecting even
     the finest particles. There is no funda-
     mental limit to the degree of cleaning
     attainable, and,  in practice, most pre-
     cipitator installations operate in the
     range of 90 to 99 per cent efficiency,
     with some as high as 99.99 per cent.
The high collecting efficiency,  the  low
flow resistance, the ability  to  treat
huge gas quantities at high gas  temperatures,
and the ability to cope sucessfully  with
corrosive atmospheres and particles  account
for the wide acceptance and diverse  appli-
cations of the electrostatic  precipitator
process.
28

-------
                                        28
SECTION 28





High Temperature Gas  Cleaning

-------
                     HIGH TEMPERATURE  GAS CLEANING
 I  COOLING HOT GASES

 A Fume collection problems ordinarily in-
   volve handling hot gases (above 600°F.)
   Thus, the problem of applying collection
   equipment to fume-producing operations
   is largely one of cooling the gases so as
   to prevent heat damage to the collector.

 B The following means of cooling are em-
   ployed,  either singly or in various
   combinations:

   1  Radiation cooling

   2  Heat recovery equipment

   3  Spray (evaporative) cooling

   4  Admission of tempering air

   5  Control of the manufacturing process
      can at times be used to control gas
      temperature


 C Radiation  Cooling

   1  U-tube  coolers

      a   Radiation cooling with conventional
         U-tube condensers is effective for
         high temperature gases when the
         temperature gradient through the
         tube walls is large.  (Heat transfer
         rates vary from 5BTU/hr/ft2/°F for
         HR stainless steel in the range
         AT = 1600 F to 0. 5 BTU/hr/ft2/°F
         for carbon at lower values of AT.)

      b   Disadvantages

         1)  U-tube coolers require a relative-
           ly large amount of ground space.
           Considerable lengths of heavy
           steel duct work from which heat
           is radiated at a low rate are
           necessary.
         2) They lack flexibility with respect
           to final temperature adjustment.
           For example, inadequate cooling
           may occur in hot weather and
           overcooling in the winter time.

         3) Usually a large capital investment
           is required.

D  Heat Recovery Equipment

   1  Waste heat boilers

      a  In these devices the hot polluted
         gases are passed over tubes through
         which water is flowing.  Hence, there
         is a transfer of heat from the gases
         to the water.  The result is a cooling
         of the gases and an economical use
         of the heat they release.

      b  Although they require a large capital
         investment, the financial return in
         the form of steam or power usually
         makes their use economically
         feasible.

      c  Process operations which are inter-
         mittent in nature cannot support
         waste heat boilers.

   2  Tubular heat exchangers

      a  These devices use cool outside air
         which is forced  around tubes carry-
         ing the hot polluted gases.

      b  They are somewhat economical on
         space requirements.

      c  They offer more flexibility in final
         temperature adjustment than do U-
         tubes or waste-heat boilers.
PA. C. pm. 26a. 1. 61

-------
High Temperature Gas Cleaning
   3  Cleaning, .kefQre_.h.e_at. recovery	
     equipment

     a  Advantages

        1) Fume deposition on the heat
           recovery surfaces of the boiler
           would be  minimized.  Thus there
           would be  improved heat transfer
           and reduced maintenance problems.

        2) Heat recovery equipment may be
           eliminated if economical to do so.

     b  Disadvantages

        1) Higher temperature resistant
           materials would be necessary.
           In  many cases, the problem is
           at  present insurmountable.

        2) The elevated temperature means
           increased viscosity of gas and
           consequently higher resistance.

        3) Increased temperature also means
           decreased gas density and con-
           sequent reduction in blower
           performance.

        4) Large gas volumes are involved
           at  the elevated temperatures.

   4  Cleaning after heat recovery equipment

     a  Advantages

        1) Lower gas volumes are  involved
           since temperature is reduced.

        2) Viscosity goes down, and density
           goes up.  Hence resistance is
           reduced.

        3) Heat recovery  equipment may
           act as an  agglomerator of particles
           which makes the particles easier
           to collect by inertial or filtration
           procedures.

        4)  Reduced particle concentration
           may result because of collection
           on  the heat recovery equipment
           surfaces.

        5)  Blower performance is improved.
         6)  Corrosion and temperature re-
            sistant materials to tolerate  the
            prevailing conditions are readily
            available.

      b  Disadvantages

         1)  Heat transfer is not good.


E  Spray Cooling

   1  Cooling involves the complete evapora-
      tion of water spray by the hot polluted
      gases.

   2  With proper controls,  initial gas
      temperatures above 2000°F may be
      reduced to 275°F by evaporative cooling.
      At the same time, the dew point may
      be maintained at less than 200°F.

   3  The additional volume of gas, due to
      the water vapor,  is nominal.

   4  Since a gas cleaning system installed
      for air pollution abatement is often a
      non-productive expenditure,  the equip-
      ment cost must be kept at a minimum.

      For this reason, spray cooling of hot gases
      is the most attractive of the available
      methods of gas cooling since it is reason-
      able in first cost, easily maintained, and
      results in only & nominal increase in the
      volume of gas to be cleaned.

F  Admission of Tempering Air
   1
      Cooling gases already high in water
      vapor content is frequently accomplished
      by the admission of outdoor air through
      bleed-in dampers.

   2  Precise regulation of gas temperature
      is readily accomplished.

   3  Disadvantages

      a  One disadvantage lies in the high
         resultant volume to be filtered,  thus
         requiring much larger collectors and
         accessory equipment.

      b  In some cases where gas being
         cleaned contains explosive gases,
         the admission of oxygen may be  ob-
         jectionable for reasons of safety.

-------
                                                           High Temperature Gas Cleaning
G  Comparison of Cooling Methods

   ]  The following hypothetical problem
     illustrates the difference in cooling
     methods. (2)

     a  In all cases,  it is assumed that
        30, 000 cfm of air at 600°F contain-
        ing 0. 01 Ib water vapor per Ib of
        dry air  is to be cooled to 275°F.
   An 80°F ambient temperature is
   assumed,  tempering air is taken
   at 80°F with 0. 01 Ib water vapor
   per Ib of dry air.  Water is avail-
   able at 70°F.  Extraneous heat
   losses are not taken into account.


b  Calculations

Wt. dry air (#/min)
\Vt, water vapor (#/min)
Wi. mixture (#/min)
j/'werfKje radiation loss (BTU/ft /hr/°F,)
KTU/miri to be removed
Kfg'd U-tube surface (fO
Length of 42" dia. pipe (ft)
1
Wt. 10 gauge U-tube, less hoppers (#)
Wt. tempering air (#/min)
**
Wt. cooling water (#/min)
Vol. tempering air at 275°F (cfm)
Vol. cooling water vapor at 275°F (cfm)
Voi . original mixture at 275 F (cfm)
Vol. to be cleaned at 275°F (cfm)
% clotii filter area required
Dt-w point
Cooling Method
U-tube
1096. 1
10.9
1107., 0
1.0
89, 443
15, 248
1,387
85,770




20, 629
20, 629
100
58°F
Temp. Air
1096. 1
10.9
1107.0

89,443



1,547.6

28,838

20,629
49,467
240
58°F
Water
1096.1
10.9
1107.0

89,443




90.30

2, 692
20, 629
23,322
113
125°F

-------
 High Temperature Gas Cleaning
II
         If spray cooling were used, from...
         operating viewpoint, the filtration
         temperature is a safe  155°F above
         the dew point.  With proper spray
         cooler design,  no difficulties would
         be encountered.
HIGH TEMPERATURE GAS CLEANING
WITH CLOTH FILTERS*2)
a  For example, when high tempera-
   atures are encountered, DuPont's
   Orion (R) acrylic fiber for temper-
   ature not exceeding 275°F has
   found application.

Where a high efficiency  of removal
of solids from hot gas is necessary,
filtration through cloth may provide
a solution.
 A Introduction
      The enactment of stringent atmospheric
      pollution codes has brought additional
      interest in cloth bag filtration as a
      means of reducing the mass rate of
      emission of pollutants to permissible
      levels.

      The development of new and higher
      temperature resistant fabrics has
      increased the usefulness of cloth fabrics
      in the high temperature gas cleaning
      field.
                                                     a However,  the success of such a
                                                       collector depends largely on the de-
                                                       gree of accuracy with which operat-
                                                       ing conditions can be  determined,
                                                       or estimated, and the pains taken
                                                       in the design of the  cloth filter and
                                                       accessory equipment  in order to
                                                       provide continuous operation with a
                                                       minimum of maintenance.

                                               B  Examples of Hot Filtration
                                450° F
                2000"F
                        Flue
                                 Electrical
                                .. Precipitator
                Furnace
                              Primary
                             Cooler
     275' f
                                                     , Secondary
                                                      Cooler
                                                                  .  Cloth Filter
                                       Figure 1
                             Carbon Black  Filtration System

-------
                                                              High Temperature Gas Cleaning
  Furnance
Rotary
  Dryer
                                           By-pass
                                                                                 Figure 2

                                                                            Asbestos Rock Drying
                                                                                Installation
  Electric
  Furriance
              Tempering  Air
              (Dust  Control  Point)
I                                                                              Figure  3

                                                                     ilectric Casting  Furnance
                                                                           Installation
                                                Fan
Canopy
lood
                    Water Jacket
                      Cooler
                watef
             4 water
^"~ ^X"
s
Tube Cooler
v

Air
°F
«
225'F
J

\AA/
s,
\
                                                                              Figure 4
                                                                   Reverberatory Brass Furnance
                                                                           Installation
         Furnance
                                 Cloth Filter

-------
High Temperature Gas Cleaning
         1950°F
               n
           Cupola
        .jpolas
                                    Water
                                    Supply
    t
                                 450'F
Quencher
                                                  By-Pass
MVMVK/
          Secondary      Fan
          Cooler
                                                             Cloth Filter   Stack
                                            Figure  5


                               Iron Foundry Cupola  Installation


                                                          rt  ft
     Secondary
     Cooling Air
                                                         ywv
                                                     )      Cloth  Filter
                                                                            Screw
                                                                          Conveyor
                                         Tempering
                                           Air

                                         •   Figure 6
                                  Rock Wool Cupola Installation
                                                         Stack
                                                                 Fan
                  -I     JJ    1
                             '   '      "
                           Spray Cooler
             Furnances
                 Figure 7


    Secondary Lead Smelting Installation
                                                       35O'F   ,.PM$
                                                               v
                                                                        Cloth Filter

-------
                                                                 High Temperature Gas Cleaning
III  HIGH TEMPERATURE GAS CLEANING
    WITH ELECTROSTATIC PRECIPITATOR


 A Introduction
                                    the collection of fine particles which
                                    may be less than 0. 001 microns and
                                    not greater than 3 microns in many
                                    cases.
       In dealing with the control of stack
       appearance, one is concerned with
                                    The collection of fines in this range
                                    may be efficiently done by electrostatic
                                    precipitators.
                  Powdered
                    Coal
                                Raw Feed
                                          Klin - 11'  x 175'
                                          Production  = 1800  BBLS/day
                                          K11n end temp.  = 1800°F
                                          Boiler exit temp.  • 430°F
                                          Cottrell =  Concrete shell with
                                             graded resistance electrodes
                                          Cottrell efficiency = 86%
                                          No dust returned
                 Clinker
                 Cooler
                         Waste
                         Heat
                         Boiler
                                                            Cottrell
                                                                              Stack
                                                Figure 8

                                        Dry Process Cement Plant.

                Kiln    11'x 360'
                Production
                Moisture  in Gases to CMP = 10 to 12%
                CMP  inlet temp.   400° to 700°F.
                Cottrell  = Steel Shell - Metallic Electrodes
                CMP  Efficiency  =95%
                All  Dust  returned to system
                                                               Slag  &
                                                               Lime rock
                                                                       Cottrell
              Powdered
                .Coal
               M/C
Clinker
Cooler
                                                                              Fan
                              T
                             To
                             Raw
                             Mill
                                                         Rotary  CMP
                                                         Dryer   Unit
                                                                                  Stack
                                              Figure  9


                                    Dry Cement Process  Using  Slag
                                           and Lime Rock

-------
High Temperature Gas Cleaning
                       Oyster
                        Shells
Fuel
       Slurry
       Storage
              Rotary
              Clinker
              Cooler
             fil
                         Kiln- 10' x 235'
                         Production - 1240 BBLS/day
                         Moisture in slurry 41%
                         Multiclone inlet temp.  - 800  -900°F.
                         Cottrell Inlet temp. * 130°F.
                         CottreU » Redwood Shell & Pipes
                         CottreU .efficiency * 87.8%
    800°-000°F.
       M/C

Kiln
End
Housing
                                                      Cottrell
                                                      Preclpitator
Scrubber
                                                         130°F.   Fan
                                                               Wood
                                                               Stack
                                        Figure 10

                         Wet Cement Process Using Oyster Shells
                                   DAM PIUS
                                                • EGINEHAIOB
                                        Figure 11

                            Flow Diagram for Open Hearth Gas
                                 Cleaning Installation

-------
                                                           High Temperature Gas  Cleaning
                                                         T)| cleaned gas
                                                            to boilers
                                       Figure 12

        Flow Diagram for Dry Cleaning Ferro-Manganese Blast Furnace Gas
                                                       AUTOMATIC STACK DAMPER
                           TRANSFORMER CABINET
                                          HEAT EXCHANGER

                            CONDITIONER        Two Point
SCHEMATIC DIAGRAM
        of
TYPICAL INSTALLATION
        on
 HOT BLAST CUPOLA
\—/              MAIN BLOWER

    Figure  13

-------
  High Temperature Gas Cleaning
IV  HIGH TEMPERATURE GAS CLEANING
    WITH VENTURI SCRUBBERS

  A Venturi scrubbers have been successfully
    employed in high temperature gas cleaning.
                         Waste gases can be handled up to 1600°F
                         with high efficiency of particulate removal.

                      B  Some applications are:
      Gas  Out~
      44,000  SCFM  (Dry Basis)

                          DAMPER
      Gas  at  125°  - 165°F. Sat'd.
      0.05 GR/SCF
S
                Gas in--
                44,000 SCFM
                at 500° - 1600°F.
                1 GR/SCF
    P-A VtNTURI

 _ _CYCJ.ONIC.
_^_ SEPARATOR
—»•
HIGH STATIC FAN

   "! THICKENER.
                                     RECIRC.
                                      am
                                      .at 40 PS]G
                     40~00~GPATM'A'K?i5p4WATER
                                                        TO
                                            Figure 1A

                                        Open Hearth Furnance
            CONVERTER
                                                                                lliMWLCflM.
                                                                               •oMIJ'F...
                                         RECKU UMK ,    •'- KCOVEtY,

                                            Figure 1'J

                                     Oxygen Steel Converter
10

-------
                                                            High Temperature Gas Cleaning
REFERENCES                                         Making Processes.  Air Repair 4,
                                                      No.  4,  189-196.  February,  1955.
1  Pring,  R. T.  Bag Filtration of Aerosols.
     Heating and Ventilating.  December,          4  O'Mara, R. et al.  Dust and Fume
     1952.                                            Problems in the Cement Industry.  Air
                                                      Repair  4, No. 4.  February, 1955.
2  Pring,  R. T.  Filtration of Hot Gases.
     Air Repair.  May,  1954.                     5  Richardson, H. L.  Scope of the Furnace
                                                      Fume Control Problem.   Iron and
3  Silverman,  L.  Technical Aspects of High            Steel Engineer.  January,  1956.
     Temperature Gas Cleaning for Steel
                                                                                         11

-------
                                          29
 SECTION 29




Sanitary Disposal  of Collected Material

-------
                                SANITARY DISPOSAL OF
                                COLLECTED MATERIAL
In considering the choice of control equip-
ment,  one must consider the  safe and sanitary
disposal of the material collected by the con-
trol equipment.  There are times when the
recovered pollutant  is a valuable byproduct
and can be marketed profitably.  But this is
not, always the case.  Usually, the storage or
disposal of the material is  a definite problem.
One that can add considerably to the cost of
the original control  equipment.

In any waste disposal problem the overall
effect of the process on the environment must
be considered.  If the solution of an air pol-
lution problem creates a problem in the
water pollution field, the problem is definitely
not solved. The reverse is also true.   Take
for example the practice  which exists in a
number of sewage treatment plants  --  that of
disposing of the gases resulting from the
digestion of sewage  sludge  by burning of the
gases in a flare.  Improperly designed  flare
mechanisms have a  tendency  to blow out,
thereby allowing unburned hydrogen sulfide
to escape to the surrounding environment,
causing a definite odor nuisance.  Suppose
the system is foolproof and all of the hydro-
gen sulfide is burned completely.   The  pro-
per combustion process is  now an efficient
producer of sulfur dioxide,  and as such
constitutes another potential air pollution
problem.

Years  ago, one of the methods of treating
spent plating wostes, was the acidification
of a batch  of spout wastes with sulfuric acid.
The cyanide fraction of the  waste  was con-
verted to hydrogen cyanide  gas and  vented to
thc> atmosphere.  Being a very toxic gas, the
potential hazard to the atmospheric environ-
ment was very great.  Disposal of plating
wastes to the waterways led to an equally
had problem  leading to numerous  fishkills.
Here, by correcting an air  pollution problem,
a water pollution problem resulted.

There are several avenues of  approach  used in
the disposal of air pollutants  removed  from
emissions;
A  Reuse of contaminant in original process.

B  Conversion or use of waste product as a
   saleable  item; e.g., recovery of SO9 and
   conversion into sulfuric acid; usage of
   flyash as fine aggregate in building blocks.

C  Storage of reclaimed product above
   ground.

D  Burial of contaminant in a sanitary landfill.

E  Burial of contaminant at sea e. g. ,  encase-
   ment of radioactive wastes in concrete
   containers and dumping at sea.

F  Discharge of materials to sewerage
   system.

G  Treatment of the  waterborne wastes re-
   sulting from control activities at the
   plant prior to discharging to a sewerage
   system.
It should be pointed out that although the con-
version of a waste product into a saleable  item
could be considered  as the most satisfactory
solution to the problem of disposal of the col-
lected waste, the demand for the product may
preclude any immediate sale, e.g.,  conversion
of sulfur dioxide into elemental sulfur.  One
plant at present has  a daily increase in its
storage area of ten tons of sulfur per day.
This company stored the sulfur above ground,
and was faced with the problem of curtailing the
emissions of sulfur dusts to the surrounding
environment due to wind errosion.

The  disposal of collected dusts by sanitary
landfill is in certain cases the safest method
of disposal.  A sketch of a landfill  operation
is shown in Figure 1.

The  landfill method in itself is relatively in-
expensive, assuming that land is available
for the  operation,  and assuming that the
collected  material is such that can be compacted
and buried. In considering the economics  of
PA.C.pm. 55a. 5.61

-------
Sanitary Disposal of Collected Material
                    	Undisturbed
                       earth
Source:  U. S. Government Printing Office.
"TB ENG 1 Sanitary Fill Method of Disposing
of Garbage and Refuse. "  1943.
Whether or not the effluent from a wet type
collecting device can be discharged directly
to a sewerage system or watercourse, de-
pends upon the effect of the waste on the
sewage treatment facilities or the  watercourse
considered.  There are numerous  contami-
nants that can effectively interfere with
anaerobic and aerobic treatment processes,
even in low concentrations, e.g.,  heavy
metals (chromium,  zinc, lead,  copper etc.);
acids both organic and  inorganic; and
alkalies.  Certain compounds, can create a
severe taste and odor problem in the water
supply if discharged to a watershed,  e.g.
aromatic compounds (phenols, naphthalene,
benzene, etc.)  Pretreatment of the wastes
at the plant may be called for.
this method of disposal, the cost of trans-
porting the collected material from the
collector site to the disposal site must be
considered.
Conclusion:  In any waste disposal problem,
the effect of the pollutant on the overall
environment,  the air and land and the water,
must be considered.

-------
                                          30
SECTION 30





Cost of Collection Equipment

-------
                       COST OF COLLECTION EQUIPMENT
 I  INTRODUCTION

 There are two questions management asks
 when they decide to act upon an air pollution
 problem.  These questions are:  Whal is the
 best way to  solve this problem and what is
 it going to cost?  Cost is of vital importance,
 for engineered control of a pollution  problem
 .seldom pays its way in recoverable product
 or by-product.   Rather, regardless of the
 value of the community good will gained,  the
 decrease in potential damage toHhe plant
 property,  or achievement of better surround-
 ings for employees, abatement  costs increases
 costs of production or service and, therefore,
 decreases profit.

 It is difficult to generalize about abatement
 costs with meaningful values because most
 pollution problems,  even those  in similar
 industries,  usually have attending  circum-
 stances that will vary the  cost from one
 specific application to another.  Rather,  what
 follows, is a list of items one might  consider
 in  attempt ing to estimate engineering costs in
 air pollution abatement.  Involved  are capital
 cosls,  i. <:. , I he costs of estimating require-
 ments, pur-chase:,  site1 preparation and inslal-
 Jalion and operating costs,  i. e. , the  costs of
 ul.ililie.s and maintenance.  Real property
 required in pollution  abatement consists of
 (I) the  abatement, device or .system itself,  and
 (2) those; accessories to the abatement device
 necessary for its operation. Such  accessories
 may be divided into those functions associated
 with:  (J) movement of air, (2) movement  of
 liquid,  (3) storage and disposal of  separated
 materials,  (4) construction and support,  and
 (5) control instrumentation.  These accessories
 n ••:: very important and should be considered
 i   ••stimating both capital and operating costs.
II  CAPITAL COSTS

 A Kaclors that must be considered in estimat-
   ing capital costs of a specific type of
   equipment, include:

   I  Capacity of the abalemen! equipmem or
      sy .<-: t e rn
   2  Accessories  to the abatement equipment

   3  Installation costs

   4  Special requirements or problems  to
      be solved

B  Capacity

   1  Inverse rule:  Abatement equipment
      prices per cfm handled will vary with
      the magnitude of the cfm involved.
      Usually the smaller the cfm the higher
      cost per  cfm.

   2  Break point:  In all designs there is a
      point where the price per cfm levels
      off, i.e., above this point,  regardless
      of  the magnitude of the cfm the cost
      per cfm remains the same. This "break
      point", however, varies with the design.

C  Accessories

   Some sellers include  all accessories,  such
   as blowers,  motors,  ductwork, etc.  Other
   sellers do not,  and in this case, these
   accessories must be obtained additionally
   by the purchasers.  To be considered  are:

   1  Air movement equipment

      a  fans and blowers
      b  electrical;  motors,  starters wire
         conduit, switches, etc.
      c  hoods, duct works,  gaskets,  dampers,
         etc.

   2  Liquid movement equipment (in wet
      abatement systems)

      a  pumps

      b  electrical:  motors,  starters, wire
         conduit, switches, etc.
      c  piping and valves
      d  set tling tanks
PA.C.pm. 49. 10. 50

-------
Cost of Collection Equipment
   3  Storage and disposal equipment
      a  dust storage hoppers
      b  sludge pits
      c  drag lines, track way,  road way, etc.
   4  Support construction
      a  structural steel work
      b  cement foundation, piers, etc.
      c  insulation  (thermal)
      d  vibration and/or anti wear materials
      e  protective  cover
   5  Instrumentation;  measurement and/or
      control of:
      a  air and/or liquid flow
      b  temperature and/or pressure
      c  operation and capacity
      d  power
      e  opacity of  flue gas (smoke meters,
        etc.)

D  Installation Cost
   Installation costs vary considerably with
   different types of equipment.
   1  Method of shipment and labor required
      at site.  This varies according to how
      the manufacture delivers the equipment
      and/or accessories.
      a  completely assembled
      b  sub assemblies
      c  completely knocked down
   2  Accessability:  Is is  necessary to:
      a  remove or relocate existing
        equipment,
      b  provide ladders and servicing plat-
        forms  for  maintenance?
     3  Utilities: Do these have to be extended
       or increased?
       a  electric power
       b  water
       c  sewerage: Will the liquid effluent
          create a stream pollution problem?
       d  steam lines

  E  Special Requirements:  Must the  system
     be designed for
     1  Resistance to corrosion,
     2  Resistance to abrasion or excessive
       wear,
     3  High or low process  temperature,
     4  Weather protection,
       a  wind
       b  flood
       c  temperature extremes

III   OPERATING COSTS
  A  Utilities
     1  Electric power
     2  Water
     3  Disposal
     4  Steam
  B  Maintenance
     1  Lubrication
     2  Surface protection; cleaning and
       painting
     3  Replacement parts or structure
       a  equipment wear or failure:  belt,
          bags, filters, etc.

-------
                                                                    Cost  of Collection Equipment
       b  abrasion

       c  corrosion


IV  COST COMPARISON

 To provide a comparison of various abate-
 ment systems Table I has been prepared.
  Conditions under which device operates are
  assumed similar (cfm,  temperature,  dust
  loading, size analysis,  etc.) unless other-
  wise  stated.   The values indicated are not
  intended to be used in a cost estimate for
  some particular equipment but rather to pro-
  vide an overall picture of comparison of
  cost.
 Table I -- Economic Comparison of Various Collection Systems (Reference 3)
                     3, 000 cfm,       68 F;       5  gr/cu. ft)

EQUIPMENT
Simple
cyclone
High eff.
cyclone
Irrigated
cyclone
Mull icyclone
Electrostatic
precjpitator
Irrigated
electrostatic
Conventional
fabric filte r
Reverse -jet
fabric filter
G ravitational
Spray tower
Wet impinge-
ment scrubber
Self- induced
Spray type
Venturi
•) -rubber
Lnsintegrator
Sc: rubber
>-,
o
c
V
o
£
w
65. 3
84.2
91.0
93, 8
94. 1
99. 0
99 9
99. 9
96 3
97. 9
93. 5
99. 7
98. 5
Capital Cost
($)
Total per
( 1) cfm
9, 240 14
17,640 .28
21,840 .36
19, 320 . 31
85,960 1.43
147, 840 2. 46
49,280 .81
47,600 78
51,240 84
28,840 48
24. 360 42
42,000 70
66, 640 1. 12
01
tc
3 C
S§-i
PH £ *
3. 7
4.9
3.9
4. 3
0.6
0. 6
4.0
5.0
1.4
6. 1
6. 1
22.0
--

s£
0 «»
0,
4, 732
6, 328
5, 634
5, 544
1, 736
3, 136
5. 264
11, 172
6, 650
8, 120
7, 896
29, 596
63, 560
r— t
0) ^
ta **--
QJ U
rt o
> 0
•* 0
T-l

--
4.0
--
...
2. 5
--
--
18. 0
3.0
0. 6
7. 0
5. 0
-&
o
o
. ^
«-£
a »
ES
--
--
1, 848
._
--
1, 232
--
--
9, 240
1, 540
308
3, 388
2, 380
Q)
g
? H
QJ ^"">
d •»
3
2
168
168
420
168
700
lr 120
8, 940
(2).
7. 560
fa)
840
840
560
840
560
i
v a
&8f.
13 w-C
s .s~
H15,4)
4, 900
6, 496
7, 952
5, 712
2, 436
5, 488
14, 168
18, 732
16, 730
10, 500
8, 764
33,824
66, 500
_ «
rt " t.
^j bo ^
-** ti ~?
g-rt^
°^5)
924
1, 764
2, 184
1, 932
8, 596
14, 784
4, 928
4, 700
5, 124
2, 884
2,436
4,200
6,664
Total
Cost
*/yr
5,824
8, 260
10, 136
7, 644
11,032
20, 272
19. 096
23,492
21, 854
13, 384
11, 200
38, 024
75, 104

C/1000cf
0.02
0. 029
0.034
0.027
0.038
0.070
0.066
0.082
0.075
0.047
0.038
0. 128
0.257
 ( 1) Includes accessories and erection

 (2) Includes complete change of bags once a year
(3) Includes complete change of bags twice a year

(4) Assuming 8000 hr/yr operation

(5) 10% of capital cost

-------
Cost of Collection Equipment
REFERENCES                                         PHS, U.S. Gov't Printing Office.
                                                      Washington, D. C.  1959.
1  Kane, J.M.  Operation, Application, and
      Effectiveness of Dust Collection Equip-       3  Stairmand,  C.J.  The Design and Fer-
      ment.  Heating and Ventilating.                    formance of Modern Gas-Cleaning
      August, 1952.                                    Equipment.  Paper read before the
                                                      Institute  of Chemical Engineers.
2  Lapple, C.E.  Engineering Control of                London.  November, 1955.
      Air Pollution.  Proc. National Con-
      ference on Air PoUution.  USDHEW,

-------
                                             31
SECTION 31
The Sylvan Chart

The class problems have been  specifically
prepared for the use of students attending the
Control of Particulate Emissions course and
should not be taken out of context.

-------
                                     SYLVAN CHART
The Sylvan Chart outlines many dust control
problems in terms of two important variables,
concentration and mean particle  size.  The
chart makes an ideal method of reporting
ranges  of dust conditions encountered from
typical  dust producing operations with the
number of operations reported limited only
by thr availability of the required field test
data.

The data provides a  guide for collector per-
formance and furnishes an approximation of
collection  efficiency and mass mean  particle
size of  effluent material.  P-rediction of mass
m^an particle  size of contaminant effluent is
based on the particle size distribution slope
for- a material being the  same on  the inlet
and discharge  from collectors operating  on
the impaction theory,  i. e. , centrifugals,
most wet collectors,  and fabric arresters.
For collectors in the electrostatic group,
deviation slopes differ and Sylvan Chart
cannot be used to approximate the mean
partii-le size of the effluent.

As  dusts found in practice  rarely consist of
particles of a given size, to define them  one
must specify not only the size but also the
rohilivo amounts  of each size.  In order  to
make a si/e and si'/,e distribution of a dusl,
it is necessary to make a particle size analy-
sis such as sieve, microscopic,  elutriation,
;irnl sedimentation.   For purposes of present-
ing, ' ompnring, analyzing, or extrapolating
pnrliele size aim lysis  in dust  control work,
logarithmic probability graph paper is norm-
ally used.   On this paper,  particle size is
plotted  on a logarithmic  scale against the
cumulative  weight percent  larger (or smaller)
than that size on an integrated probability
scale.   This plot  frequently gives straight
lines or lines of relative small curvature.
With this eurve, the  distribution function  can
be stated in terms of two parameters,  the
median size and standard geometric  deviation.
The latter  is expressed as  the 84. 13% size
divided  by  the 50% size.  Using these parame-
ters in  a mathematical function,  the deviation
lines shown on the chart have  been computed
for-  industrial dusts.
Example:  A suitable collector will be selected
for a lime kiln to illustrate the use of the
Sylvan Chart.  Referring to the chart, the
concentration and mean particle size of  the
material leaving the kiln can vary between
3 and 10 grains/cu. ft.  with 5 to 10 microns
range of mass mean particle size.  Assume
an inlet concentration of 7. 5 grains/cu. ft.
and inlet mean particle size of 9 microns.

Projection of this point vertically  downward
to the collection efficiency portion of the
chart will  indicate that a low resistance
cyclone will be  less than 50% efficient; a
high efficiency centrifugal will be  between
60 and 80% efficient and a wet collector,
fabric arrester and electrostatic preeipitator
will be 97% plus efficient.  The latter three
collectors are often preceded  by a precleaner
so a high efficiency centrifugal will be
selected.   Using the average line of this group,
group, the efficiency will be 70%.  Therefore,
the effluent from this collector will have a
concentration of 7. 5 (1. 00   .70)   2.25
grains/cu. ft.  Draw a  line through the initial
point witri  a slope parallel to the lines
marked "industrial dusts. "  Where deviation
is not known, the average of this group of
lines will normally be  sufficiently accurate
to predict  the mean particle size in the
collector effluent.  Where this line intersects
the horizontal line marked 2. 25 grains/cu. ft. ,
a vertical  line through  the point will indicate
the effluent mean particle size of fi. 0 microns.

A projection of triis point of collector effluent
vertically  downward shows that a second high
efficiency  centrifugal will be less  than 50%
efficient.   A wet collector, fauric arrester
and electrostatic preeipitator  will be not less
than 93% efficient.  Selection of a  good wet
collector will show an  efficiency of 98%.  The
effluent leaving this collector  will have a
concentration of 2. 25 ( 1. 00  . 98)   . 045
grains/cu. ft.  Using the line initially drawn,
at the point where  it intersects the line  of
. 045  grains/cu. ft. will indicate a  mean
particle size in the effluent of 1. 6 microns.
I'A. ('. pm. !). .r>. 57

-------
S) Ivan  Chart
                        RANGE OF PARTICLE SIZES, CONCENTRATION, & COLLECTOR PERFORMANCE

                 COMPILED BY S. SYLVAN APRIL 1952:     COPYRIGHT 1952   AMERICAN AIR FILTER CO. INC.
                 ACKNOWLEDGtMENTSOF PARTIAL SOURCES OF DATA REPORTED
                    FRANK W G  AMF.RICAN AIR FILTER - SI2E AND CHARACTER ISTICS OF AIR BORNE SOLIDS 1931
                    I JRST AN'.i DRINKER   ARCHIVES OF INDUSTRIAL HYGIENE AND OCCUPATIONAL MEDICINF APRIL 1952

-------
                                       32
SECTION 32





Class Problems

-------
Problem 1. Volumetric Flow Rate

     The volumetric flow rate from a process is a factor that affects
     the efficiency of a collection device.  A test yielded the following
     results:

              Flowing gas - dry air

              Temperature (t) = 300°F

              Static pressure (PB)   14 in. 1^0 vacuum

              Atmospheric pressure (Patra) ~ 29.6 in. Hg

              Average velocity head = 0.7 in. H,,0


              Duct dimensions » 20 in. by 15 in.

     Calculate the following:

          a.  Absolute temperature (T) in °R
                                                                          Answer: 760
          b.  Absolute  pressure  (P  ,  )  in  in.  Hg.   Specific  gravity  of
              mercury  is  13.6.
                                                                          Answer: 28.6
           c.  Has Density (p) in lbra/ft3.  To solve the problem, use  (1) the ideal gas
               law and (2) the equivalence of 359 ft'/lbm - mole at 32°F and 14.7 psia
               Molecular weight  (M) of air - 29  lbm/lbm - mole and gas constant
               (R)   21.83 (in.  Hg * ft1) / (lbm  mole * °R)


               (1)  [deal Gas Law

-------
    (2)  Equivalence
d.   Volumetric flow rate (Q)  in cfm.  Density  of  1^0 (PH 0^  " 62.4 Ib /ft
                                                                         Answer:  8559
    If the velocity doubles, what will be the velocity head?
e.  Volumetric flow rate (Q) in cfm at 250°F and -18 in. HO
                                                                        Answer:   8081

-------
Problem 2. Pressure Drop

    The pressure drops for most cyclone installations are said to range from
    2 to 7 in. w.g.  How would you measure the pressure drop across a cyclone?
    Is this static head loss or total head loss?  What causes this loss?  Is
    static head loss approximately equal to total head loss?   Explain:

-------
Problem 3. Reynold's Number

    A particle 70 microns (\i) in diameter settles at a rate of 1 tps in air
    at 300°F and 28.6 in. Hg absolute.  Determine:
        Gas Viscosity in ll^/ (ft * sec)
                                                            Answer:  1.55  *  10~5
    b.  Reynold's Number  (NRe>.  Refer to problem  1  for gas density.

        1  ft »  30.48 *  10%
        What is  tho settling  region of  the  particle?

-------
Optional Frob Jem. Reynold's Number

    Determine the largest particle size  that will  settle  in  ajr  at  70  ^
    under Stokes's Law for dusts with specific gravities  of  1, 2,  and  J
        What  Is the N   equation?
     b.   Solve for Dp' In terms of p  and fill in table below.  Use Stoke's
         extended region.

              pair   0.018 cp at 70°F

              ojU   i..';?S lbra/(t1 ai 70°F and 1A. 7 psia
N
Re
0. 1

9-8-
1
37

2
29

3
25


-------
Problem 4.  Molecular Weight  of Gas Mixture and Review



    An effuent  gas has  the following composition:   a
Component
C02
CO
0?
N;.
H;,0
Vol. %
11
0.5 i.
6
76
6.5
M''
44
28
32
28
18

-*,«/-
f,40
(,4?-
M, 
-------
c.   Stoke's settling velocity in fps for D  = 50u (s.g.=3).
    1 ft.=30.48 * 10V                   P                           Answer;
     What is the settling region for this particle?
     Was Stoke's law assumption correct?
                                                        Answer: Yes
 d.  Velocity head in in. H20 if v- 80 fps
                                                                    Answer: 0.94
                                                                                 n

-------
Problem 5.)  Cumulative and Frequency Distribution Curves

     Particle size analysis is an important consideration  in  evaluating
     control equipment.  The data can be represented  by plotting  a
     frequency distribution curve and/or a cumulative distribution
     curve or represented with the mean and standard  deviation.
     Particulate matter from a grinding operation is  known to have
     a log-normal distribution from previous analysis.  Because of
     this, only four size fractions were used for a particle  size
     analysis.  For the data given below:
                            Table  1.  Particle Size Analysis
,,-y
Particle Size
(microns)
0-10
10-20
20-40
+ 40
Total Weight in
Each Fraction
(%)
36.9
19.1
18.0
26.0
Cumulative
Z




                                                                                Jo
                                                                               f,;/
I       a.  Determine the geometric mean and geometric deviation.   To
           do this, construct a                                   on
           analysis (Use Figure 1).
                                graph paper for the particle size
                      Geometric Mean
                      Geometric Deviation •
        b.  Plot a. frequency distribution curve.  To do this, arbitrarily
            select size fractions desired.  Then, using the	
                             curve, determine the amount in each size
            fraction (Use Table 2 and Figure 2).
                 & <2!L

-------
                                          Figure 1    Cumulative distribution  curve
c
o
i.

     0.01       0.1 0.2       12     5     10     20   30  40  50  60  70   80     90   95     98   99     99.8  99.9     99.99



                                                  %  by weight less than stated size

-------
                   Table 2.   Frequency Distribution
Size
Fraction
M
0-1
1-2
2-3
3-4
4-6
6-8
3-10
10-12
12-14
14-16
16-18
18-20
20-30
30-40
40-50
>50


Lower
Size
0
2.5



24,2
31.0
36.9
41.8
46.0
50.0
53.0
56.0
67.0
74.0
78.8

% by Weight
Upper
Size
2.5
7.0



31.0
36.9
41.8
46.0
50.0
53.0
56.0
67.0
74.0
78.8
100.0

in Fraction
Difference
2.5
4.5



6.8
5.9
4.. 9
4.2
4.0
3.0
3.0
11.0
7.0
4.8



Per
Micron
2.5
4.5



3.4
2.95
2.45
2.1
2.0
1.5
1.5
1.1
0.7
0.5


Average
Size
P
0.5
1.5



7
9
11
13
15
17
19
25
35
45


/o

-------
6 iii
         -rrt-rr
                THi:
              ^
m
                                 ::~1

                                 :n±
                                      l- + -:f-
                                              "ti -iii t-
                                                nt.
     ,tiitH
     ittn
                     "irlt
                    -tfftT
                                 tir:
                                                           t Jii:
                                 n;t
                                                            ±nr
                                                                     ft
                                                                                  .lU-nU-
                                                                                        Jl.
                                                                                                     -I --•
                    MB
                                                                                                         ffi
                                                                                                   i
                                      12
                          16         20          24



                            Particle diameter, microns
28
36
40
                                          Figure  2     Frequency  distribution curve

-------
Problem 6.  Data Representation

     If a log-normal or normal distribution curve  exists,  the mean
     and standard deviation or geometric deviation are  sufficient to
     describe the data.

     Given the following data for log-normal distribution,  plot  the
     cumulative distribution curve on log-probability graph paper.
Source
Open Hearth
Fly Ash(b)
Cement Kiln*^
(d)
Gray-Iron Cupola
Fly-Ash, Cyclone
type Furnace

Mass Mean
Dia. 
-------
100
80
60
50

40


30


20









10
8

-------
  Problem 7.  Frequency Distributions



       a.  Sketch the frequency distribution curve for the following:
                50

        Log-probability curve
                                             c
                                             o>
                                             3
                                             cr
                                             01
                                             t-
         Log size

Frequency distribution curve
2.
                50

        Log-probability  curve
                                             u
                                             c.
                                             01
         Log  size

Frequency distribution curve
       b.  If you mix two different log-normal distributions with different

           means, what general curve will result on log-probability paper?
                                                    Mixture  of  1  and  2
              50
                                                          so
                           Log-probability curves

-------
        Problem 8:   Settling Chambers

             Two small heating plants, one using a traveling grate stoker
             and the other a spreader  stoker, desire to install a settling
             chamber.  The conditions  of operations are listed in Table 1.
                                  Table 1:   Optional Data
                                                    travel ing
                                                      grate
spreader
Chamber Width (ft.)
Chamber Height (ft.)
Chamber Length (ft.)
Volumetric Flow Rate (scf/sec^
Flue Gas Temperature (°F)
Flue Gas Pressure (in. Hg gage)
Dust Concentration (gr/scf)
Mass Mean Diameter »(u)
Geometric Deviation a
Particle Specific Gravity^

10.8
2.46
15.0
70.6
446.0
0.0
0.23
72.0
1.95
2.65
10.8
2.46
15.0
70.6
446.0
0.0
1.22
57.0
4.06
2.65
1
             Note:   Standard conditions - 32°F and 29.92 in. Hg
                  Estimate  the  overall  collection efficiency assuming actual
                  settling  velocity  • *5 theoretical Stoke' B settling velocity.
3 Particle size data from Anderson, D.H., et al.,
  Pure Air for Pennsylvania, Penn. Dept. of
  Health, November, 1961.

-------
(1)  In order to determine the overall efficiency we need a size
     efficiency curve.   Assuming vy = 2 vy(s)>  what is the size
     efficiency equation?
      Note that the chambers and  operational conditions are identical
      for both plants.   Will the  size efficiency curve be the same
      for both?  Yes           No
(2)   To  obtain the  size  efficiency curve we  need  to  solve  the
     equation in  (1).
     (a)   What  is  the  equation for v  ,    ?
       (b)   Substitute  (2a)  into the  size efficiency  equation.
     (c)  Gather all constants into the factor K.

-------
     (d)  Do we know or can we calculate all the values in K?
          If so, solve for K.  Do not forget to correct Q
          to operating conditions.
               L  •
               B
g -

<"

u  -

Q -
                   - P}
Since we wish to use p for D ,  place all conversion factors  into
K.                          P
                                                    Answer:  1.136 * 10

-------
(e)   By selecting arbitrary values  for  D  .  we  can  determine
     E  ,  or vice versa.   What  is  the  minimum particle  size
     in p
          that can be collected  at  100%  efficiency.
                                                        Answer:   94
 (f)  Fill in the table below and check the size efficiency
     curve in Figure 1.  If it does not coincide, check your
     caIculations.
                            8100
 _P
100

 92

-------
90
90
                                                                       Size  efficiency  curve
                                                                       for settling  chamber
                                      Particle size ( microns )
                                              Figure 1

-------
        »•-  '.  - .^r —-"-]"-T^^-'
o
        300
        200 t
    o
    s_
    CJ

        60 --—'-=±	i    ^	1=
          0.01       0.1  0.2       125     10     20   30  40  50  60  70    80    90    95     98  99


                                                     % less  than indicated size
99.6 99.9      99.99
                                               Figure 2    Cumulative Distribution curve

-------
(3)   Besides the size efficiency curve we need to know the particle
     size distribution of  the inlet  dust.

     (a)   Plot  the size distribution curve for the inlet
          dust  in Figure 2.

          traveling grate  stoker
                                                       Y
          V       H c V                        V      ™-
           84.13      50                       15.87    S
                   1.95X72p
                   140.3  y                            -36.9  y
          spreader  stoker
     (b)  Fill in the tables provided and calculate the overall
          efficiency.  Plot the outlet dust size distribution in
          Figure 2.
 b.   Kscircate the amount  and  size distribution of  material  escaping
     the .settling chamber.   See(a3b).

-------
Table 1.  Traveling Grate Stoker
Size
Fraction
p
0-20
20-30
30-40
40-50
50-60
60-70
70-80
80- 90
fr 44
Total
L "_"
InLet
%
2.7
6.9
9.4
10.5
10.5
9.5
7.0
9.5
34
100
gr/scf
0.0062
0.0159
0.0216
0.0242
0.0242
0.0218
0.0161
0.0218
0.0782
0.2300
'p
1.1
7.1
14.0
23.0
34.0
48.0
64.0
86.0
100
	
Hopper
Catch
gr/scf
0 . 0001
0.0011
0.0030
0,0056
0.0082
0.0105
0.0103
0.01.87
0.0782
0.135;

Outlet
gr/scf
0.0061
0.0148
0.0186
0.0186
0.0160
0.0113
0.0058
0.0031
0.0000
0.0943
%
6.5
15.7
19.7
19.7
17.1
12.0
6.2
3.3
0
100.1
Cum. %
6.5
22.2
41.9
61.6
78.6
90.6
96.8
100.1



      Inlet  -
inlet
                       CaLrli
                       Inlet'
      .1357
      .23
    .59

-------
Table 2,  Spreader Stoker
Size
Fraction
0-20
20-30
30-40
40-50
50-60
60-70
70-80
80-94
+ 94
TOTAL
Inlet
7.
23.0
9.5



4.5
3.5
4.5
36.0
100
gr/ecf
.2806
.1159



.0549
.0427
.0549
.4392
1.2200
E
P
%
1.1
7.1



48.0
64.0
86.0
100.0

Hopper
Catch
gr/acf
.0031
.0082



.0264
.0273
.0472
.4392
.6031
Outlet
gr/scf
.2775
.1077



.0285
.0154
.0077
0
.6169
%
45.0
17.5



4.6
2.5
1.2
0

%
45.0
62.5



96.3
98.8
100.0



-------
Problem 9.  Settling Chamber

     A settling chamber is 30 feet long, 6 feet high, and 12 feet wide.
     The air flowrate at 500°F and 29.92 in. Hg is 4,320 acfm and the
     entrained particles have a specific gravity of 2.5.  What is the
     diameter in microns of the smallest particle that can be removed
     with 100%.efficiency if the effective settling velocity in the
     chamber is h. the Stokes settling velocity?

-------
Problem 10.  Cyclone

     An efficiency teat on a cyclone produced the following data:

     Efficiency - 94.7%
     Specific gravity of dust - 2.75
     Inlet dust concentration -1.0 grains/ft
     Gas viscosity - 0.025 cp
     Dust Analysis:8
Size Fraction
M
0-5
5-10
10-15
15-20
20-25
25-30
30-35
35-40
40-45
> 45
Hopper Catch
wt. %
0.5
1.4
1.9
2.1
2.1
2.0
2.0
2.0
2.0
84.0
Outlet
wt. TL
76.0
12.9
4.5
2.1
1.5
0.7
0.5
0.4
0.3
1.1
      a C.A. Mau "Tho Elimination of Dust from Asphalt Plants"
        Air Repair  . 102-104 CNovember, 1953).

-------
                                                                                                                           005    001






OJDI
10       20    30   «   50   60    70    §0      M
 % by weight  less  than  indicated size
                                                                                                     M   99
                                                                                                                    99.8 99.9       99.99
                                             "•*-  re 3      Pa[    le

-------
s.   Draw the size-efficiency  curve  (Use Figure 2).
Size Fraction
u
0-5
5-10
10-15
15-20
20-25
25-30
30-35
35-40
40-45
> 45
TOTAL
Hopper
%
0.5
1.4
l,'l
3,>
1 '
tf >
2.0
2.0
2.0
2.0
84.0
100.0
gr/ft3
.0047
.0133
,01 SO
,d'^j1
0. / "1 1
.0189
.0189
.0189
.0189
.7955
.9469
Outlet
%
76.0
12.9
^
*>l
/^
0.7
0.5
0.4
0.3
1.1
100.0
gr/ft3
0.0403
0.0068
6.024-
6.VII
Ofirt
0003
0003
0002
0002
0006
0,OS3\
Inlet
gr/ft3
.0450
.0201
,,0304-
,02/3
'Ojo?
.0192
.0192
.0191
.0191
.7961
.9999
%
4.5
2.0
J,0
J,l
J-l
1.9
1.9
1.9
1.9
79.6
m
Eff .
%
10.4
66.2
#;;-
if ]
76,1
98.4
98.4
99.0
99.0
99.9
94.7
 What is the cut size?
c ./  If a geometrically similar cyclone,  one-half  the  diameter of the test cyclone,
/  is to be operated at three-quarters  of  the  inlet  velocity and a gas viscosity
    of 0.020 cp,  estimate the cut size.                                     /
             y
                     /      /
                    tf-  f*rr/fr
                                     X'
                                                                                  ,
                                                                                           ,•'_-

-------
d.  Draw the size efficiency curve for the condition in (c). Use Figure 2.
Efficiency
20
40
50
60
70
90
Test Cyclone
D
P






Geometrically Similar Cyclone
D
P







-------
   100
    80
c
cu
o
    40
   20
                           TT
10           15          20           25



                Particle  size,/i




        Figure 2    Size-efficiency  curve
                                                                               30
33
                                                                                                        40

-------
          Optional Problem:  Cyclone


               The following results were obtained from a cyclone servicing  a  hot-mix
               paving plant.(1)
         Vent
         Dryer
              3715 scfm
               200 °F
               743 Ib/hr dust
               22,050 scfm
                  430  F
                4,720 Ib/hr dust
Multiple cyclone
1525 Ib/hr dust
                           Table I.  Collection Efficiency Data
Dust Particle
Size u
0-5
5-10
10-20
20-50
> 50
Size Analysis, wt . %
Inlet
6.2
9.4
13.8
22.9
47.7
Outlet
19.3
31.9
31.6
15.1
2.1
                  The data was plotted on a log-probability paper as shown in Figure  1.


                  Draw the size efficiency curve.
(I)  Taken  from  Air  Pollution  Engineering Manual, PHS Publ. No. 999-AP-40, Cinn.  (1967)

-------
     100 99.99       99.9 99.g
                                                      Figure 1      Particle Size Distribution
                                   99    98      95     «0      80     70    £0    50   40   30    20       10
                                                                                                                      2     I    OJ   0.2  0.1  0.05     0.01
4;
o

          0.01    0.05 0.1  0.2    0.5   1     2
                                                       10
                                                               20    SO   40   M    SO    70
                                                                                                                                                              CO,
                                                                % Less Th*n

-------
Size Fraction
W
0 5
5 10
10 15
15 - 20
20 - 25
25 - 30
30 - 35
35 - 40
40 45
> 45
Total
Inlet
%











Ib/hr











Outlet
%




	





Ib/hr











Hopper
%











Ib/hr











Eff.
%












-------
    H--
      	L.T_4-J—I—1.-I—i '	'—.--
Eff.
 100



                                                                                 IT

                                                                                                         -1-4-
  80
  60
        •H-
         h-
  40
  20
10
15
20
                                                               25
                                               30
35
                                                           40
45 .

-------
Problem 11.  Cyclone
     A  single  cyclone  handling  6,000 cfm has  the following dimensions:
               Diameter, D

               Inlet width, W

               Inlet height, 1^

               Length  of cone.l
                  6            cone
               Height  of cylinder, 1

               Exit duct diameter, d
cyl
               Carrier gas density, Ib /ft3
                                      in

               There  is no entry vane
4 ft

1 ft

2 ft

8 ft

8 ft

2 ft

0.075
  In an attempt  to increase  efficiency,  a  group of new cyclones is to be
  designed  with  the same geometrical  proportions and same pressure drop
  as the single  cyclone.   Diameter  of the  small cyclone is to be 6 inches.
  Determine the  following:

      a..  What will be  the dimensions of the  new cyclones?

Diameter, D
Inlet width, W
Inlet height, 1
Length of cone, 1
cone
Height of cylinder, 1 ^
Exit duct diameter, d
o
Dimensions, ft
Old
4
1
2
8
8
2
New
0.5






-------
b.  How many of the small cyclones will be necessary to handle the original

    flow rate at the same resistance?



    .(1)  What is the equation for the pressure drop across a cyclone?
     (2)  If the resistance must be the same for both cyclones, the ratio

         of Ap     to Ap     must be equal to
               old     * new                  	
    (3)  What then is the relationship  between Ap     and Ap    ?
                                                * old      T new
    (4)   Solve for  the number  of   new  cyclones.
    What Is the ratio of the inlet velocities for the old and new cyclones?

-------
Problem 12.  Cyclone

   A cyclone 8 inches in diameter carrying a. gas (essentially air)
   at 340°F with an inlet velocity of 50 feet per second removes a
   5 micron particle (sp.g. » 2.5) with 50% efficiency.   If this
   cyclone were operated at 170°F with an inlet velocity of 25 feet
   per second, estimate the size of particle removed with 50%
   efficiency. Assuming that the equation for efficiency as a function
   of [Dp] cut and [Dp] mean (Eqn. 16 in the course manual) is valid
   for this case, and given a mean particle size ([Dp] mean) of 30
   microns, what are the values for the old and new efficiencies.    „/•
      - •/
      /
37

-------
     X


/<

-------
Problem 13.  Electrostatic Precipitator

     The following data represent results from the operation of  a pilot-plant
     electrostatic preclpitator on a 200,000 scfm power-plant effluent:
     Dust Loading

          Inlet = 3 grains/scf

          Outlet => 0.50 grains/scf
     Dust Analysis:
Size Fraction
(u)
0-10
10-20
20-44
+ 44
Inlet
(%)
30
20
12
38
Outlet
(%)
60
18
4
18
     If the final precipitator is made twice as long as the pilot plant:
     a.  Calculate the percent removed in each size range and  overall
         percent efficiency for both units.

         (l)Calculate the size and overall collection efficiencies  for  the
            given size fraction for the pilot plant (Use Table 1).
      Table 1.  Pilot Plant Precipitator
Size
Fraction
P
0-10
10-20
20-44
> 44
TOTALS
Inlet
%
30
20
12
38
Iff
gr/scf
0.90
0,£0
0,3 (>
/,(!-
? /-i
<.*> ' '--•
Outlet
%
60
18
4
18
foo
gr/scf
0.30
0,010
0,MO
6f^o
0,£
Hopper
gr/scf
0.60
as/
0.34-
1,05
J.f.
Efficiency
%
66.7
8S>0
7/,^

-------
  99.9
   99.8





   99.7



   99.6


   99.5


   99.4



   99.2


   99
        -  4-I--
                           n::
                          ft
                          -4

                                              4
                                                   .J4r
                                               -H+
                                                       	l_.l
                                                             Ml!
                                                              •itH
                                                       J- ; : -i

                                                       t|T|
                              3E
                                         f:
                                                   £
                                                                       •STri
                                                           .r."
                                                                             3^
                                             t£
                                                       m
     4P
     14
          =rM
          J.
      44-
                      ¥i
                                       T--E
                                                        W
                                                 mr
                                                 |-i4>4
                         rSEE
                               14
                                                       «
>
u
c

-------
(2)   What is the efficiency equation for the precipitator
     e = A *  °'058 * FX  * F2 * rD
         Q
      Noting the efficiency equation,  how does  doubling  the  precipitator

      length affect the factor e ?
 (3)   Calculate  the  size  and  overall  collection efficiencies for the given
      size  fraction  for the final  precipitator (Use Table 2).  A calculation
      aid is  provided  in  Figure  1.
                       Table 2.  Final Precipitator
Size
Fraction
V
0-10
10-20
20-44
-44
TOTALS
Conditions
Old
EP1
66.7
or O
-^C-1-/ LS
7A +
7<2>l

"i
1.10
/^
cZy'
C/,JJ

New
"2
2.20
J>,$
jf f:
.J. /

EP2
89.0
7jr^
Tlil
?7,*h-

Inlet
gr/scf
0.90
0,&d
QJ&
////-
J>,0
Hopper
gr/scf
.801
0,SB/
4,360
/,/?&
~3-&-$-
&i(^O
    (4)   Plot the size efficiency curve in Figure 2.
b.  Estimate the efficiency for final precipitator based on overall
    efficiency of pilot plant:

-------
                            -.-"f-
    80
.-Hi
    60
o
c
a*
•i—
o
   20
                              10
    15
 20           25


Particle size,
30
35
40
45
                                            Figure  2     Size efficiency curve

-------
Optional Problem.  Electrostatic Precipltator

     The following data represent results from the operations  of  a
     pilot-plant electrostatic precipitator on a power-plant effluent:0
    Dust Loading:
Inlet
Hopper
351 Ib /hr
210 lbm/hr
      m
    Particle  Size Distribution
             (microns)

           0-10

           10-20

           20-44

           + 44
                  %  In Fraction
              Inlet	  Hopper
               86

                8

                4.2

                1.8
                 93.3

                  2.6

                  1.5

                  2.6
   Calculate  the following data, assuming the final precipitator is
   made  twice as long as  the pilot plant:


   a.  Percent removed in  each  size range and overall - percent
       efficiency for both units.
                  Table  1.  Pilot Plant Precipitator
Size
Fraction
1.1
0-10
10-20
20-44
44
TOTALS
Inlet
%





Ib /hr
m





Hopper
%





Ib /hr
ra





Outlet
Ib /hr
m





Efficiency
%





    a C.R. Flodin and H.H. Haaland, "Some Factors Affecting Fly-Ash Collector
      Performance on Large Pulverized Fuel-fired Boilers,"  Air Repair 5,
      27-32 {May, 1955).

-------
                   Table  2.  Final Precipitator
Size
Fraction
u
0-10
10-20
20-44
> 44
TOTALS
Condition
Old
EP1





Bl





New
«2





V





Inlet
Ib /hr
m





Hopper
lbm/hr





%





Hopper
Actual
%
94.6
2.2
1.3
1.9
100
b.  Estimate the efficiency for  final precipitator based on  overall
    efficiency of pilot  plant:

-------
Optional Problem.  Electrostatic Preclpitator

If for the previous problem the particle  size distribution were to
change can estimate of overall efficiency be made?

-------
Problem 14.     Electrostatic Precipitator

An ele-etrostatlc precipitator operates at an overall efficiency of
95% when the gas temperature is 300°F.  Estimate the new overall
efficiency when the gas temperature is 400°F, the precipitator
handles the same mass flowrate of gas, and all other operating
parameters remain constant.

-------
Problem 15.   Electrostatic Precipitator
An industrial installation has two electrostatic precipitators each
designed to accomodate 72,000 cfm.   Given:

             Migration velocity:  0.4 feet/sec

             Collecting surface:  14,400 sq. ft per precipitator
           Calculate the collecting efficiency of each precipitator

                    (1) Write the efficiency equation.            ~IZZZL-~- X
                    (2) Solve for E

                                                                        Answer:  99.2
        b.   If  one  precipitator  is  shut  down  and  the  total  gas  volume  is
            treated in one  precipitator,  calculate  the  collecting efficiency
            (migration velocity  remains  constant).

            (1)  At this point a graphical method will  be illustrated. Note
                 that a plot of  efficiency (E)  vs.  collecting surface  per
                 unit gas volume (A/Q)  on semi-log  paper will yield a  family
                 of straight lines, each slope  depending on the migration
                 velocity  (v ).
                 Mathematically

                        E = 1-e -A/Q  «
                                       P

                        or (1-E) =  e - A/Q Vp

                        In (1-E) = - v   A
                                      P  Q

                 Draw the line for vp =° 0.4 ft/sec in Figure 1.

-------
 99.a
Eff.,
 99.5
                                                 -Hr
                                                            -rr
                                                                        Ft1,1
                                  l-T
                                                                                  r
                                                                                   t!  -
                                                                                     T:
 99
 98
tit
                                                       TT
                                                                        Mr
 95
                                                    TTtr
 90
                                                M i
                               ri
                                                  i i
                                                Hi
 80
 50
                                                                   I- t-
                                  10     12    14
                                   A/Q,  sees/ft
          16     18    20    22    24    26

-------
(2)  Calculate  A/Q  and  determine  the  efficiency  from the graph.
                                                                  Answer:  91.0
 Increased production results in an increased gas volume of 84,000 cfm
 per  precipitator.  New requirements call for a collecting efficiency
 of 99.84% when both precipitators are on line, and a minimum efficiency
 of 96.0% with all the gas going through one precipitator.  Assuming
 migration velocity to be constant at 0.4 ft/sec, for both conditions,
 what size precipitator should be used to meet both requirements?

                                                        Answer:  22,400  ft2

-------
          Optional  Problem.   Electrostatic Precipitator *


            Given:  A horizontal-flow, single stage electrical precipitator consisting of
                    two ducts formed by plates 8 ft wide by 12 ft high on 10 inch-centers,
                    handling 3,600 cfm with 2 grains fly ash (pulverized coal)/ft .
            Estimate  the  drift  velocity.   Find  the  collection efficiency and the
            outlet  dust emissions  in  Ib/hr for
             (a) Assuming  uniform  gas velocity.
                                                                            Answer:  5.42 Ib/hr
             (b)  Assuming the velocity through one of the duct
                 is  50% greater than the average and 50% lower in
                 other
                                                                             Answer:   9.27  Ib/hr
                          Table 1.  Typical Drift Velocities
Application
Pulverized coal (fly ash)
Paper mills
Open-hearth furnace
Secondary blast furnace (80% foundry iron)
Gypsum
Hot phosphorous
Acid mist (H2SOA)
Acid mist (Ti02)
Flash roaster
Multiple-hearth roaster
Portland cement manufacturing (wet process)
Portland cement manufacturing (dry process)
Catalyst dust
Gray iron cupola (iron-coke ratio-10)
Drift
velocity
ft/sec
0.33 to 0
0.25
0.19
0.41
0.52 to 0
0.19 to 0
0. 19 to 0
0.25
0.26
0.33 to 0
0.19 to 0
0.25
0.10 to 0
(V
44



64
25
25


37
23

12
* Taken from  Ai
r Pollution Engineering Manual, PHS Publ. No 999-AP-40, Cinn. (1967)
                                                                                               5"!

-------
Problem 16.  Venturi Scrubber



       A venturi scrubber has been operated in a phosphoric acid plant  to
       collect phosphoric acid mist.   Operating data include  the following


                   Throat Velocity,  vt = 218 fps (t = 68°F)

                   Water Injection Rate, L = 1.44 ft3 / 1000 ft3 (t - 86°F)

                   Particulate Loading

                       Inlet = 8.50 gr PZ DS/ ft3

                       Outlet = 0.123 gr P 0 / ft3

                   Particle Size Data

                       Inlet: mean = 1.61, std.  dev. =1.56

                       Outlet: mean = 0.85,  std.  dev. - 1.73

                   Mist Density,  P  = 116  lbm/  ft3
               From the above data,  the size-efficiency curve for the scrubber
               can be shown to be as  listed below:

                           Particle  Dia.                Efficiency
                             (H)                            %
                             0.2                          31.5
                             0-4                          64.0
                             0.6                          88.3
                             0.8                          96.0
                             1-0                          98.2
          a.  Compare the theoretical size-efficiency curve (K = 1.52) with the
              measured curve.
               (1)   Write down  the applicable equation.
                         E - 1 - e - e

                         e   KL ^

                         i(/   V  D 2 p
                                 P  kp
                             18 Dw  p

                              1.6 * IP1* + 28.5  L  l'5
      i, J.A., Contant, C.E., Ind. &Eng. Chem., v.50, n.8, 1958.

-------
      Where:  K •  constant, determined experiment^?ly


              L -  liquid injection rate, ft3/1000 ft3


              " -  throat velocity, fps


              D •  water droplet diameter, y


               "  impaction efficiency, dimensionless
(2)   Calculate the droplet size.
                                     2
 (3)   Calculate \Jj as a function of D

      Use D  in y.                  P

-------
  99.9
  99.8







  99.7




  99.6



  99.5


  99.4




  99.2




  "
       3
T1
               -::--±.
                                 -^
                                          J~
   l_
                                                            -r

o

c
o>


u
                                                        I T

                                                        1-1-
   r -i-

   T
                 -H-
          1
     . f;t+.
                               £E
                   Uib
                                                            m*
   96



   95


   9-:




   92



    0

                                                   t:4-:
                                                        b(-
                                       t

ftt
                                 ^-1
                                         w~
                                            f
                                                     -Hi
                             f'-
                                            T1-
   JO
              H-t
           --^-
           ri ti
T-I---
                til
  -•-hn-
            rr]U.
            'TnT
              iiiiiii
-l-i
4-r- -H-^-f
                      "trn;
                                                               r- -.

                 i
                    34



                     txponent,  f.

-------
   (4)  Calculate B as a function of D  in u .
                                      P
   (5)   Using Figure  1  fill  in the  table   below,  and plot  the data

        in Figure 2.
DP (y)
0.2
0.4
0.6
0.8
1.0
e





E (%)





b.  Compute a theoretical grade-efficiency curve for v  = 554

    (three times the original velocity)
     (1)  Calculate D   at v  - 654 fps.
                    w      t

-------
LT
        c
        O)
        o

        O)
        Q.
        u
        c
             90
             70
             60
             50
             40
             30
             20
             10
                                    I !
                                                                                             --U-
                                                                                             T
                                              0.5
1.0
1.5
                                                   Particle size  (  microns )
                                           Size efficiency curves  for Venturi scrubber


                                                           Figure 2

-------
2.   Calculate 1(1 as a function of D 2 where D  is in microns.
                                  P         P
 3.   Calculate z as function of  particle size  D  ,  where D  is
     in microns.                              P        ^
4.  Calculate  the  theoretical grade  efficiency curve from the
    relationship E -  1 -  e - e.
D
P
(Microns)
0.2
0.4
0.6
0.8
1.0
z





E
%






-------
Problem 17.  Venturi Scrubber

     A venturi scrubber is to be installed on four processes.
     All four applications will be using a venturi scrubber having
     a throat cross section of 10" x 4.0* and a correlation
     coefficient of 1.20.  Using the following equations developed
     by Ranz and Wong, and by Nukiyama and Tanawawa.
                       1 - e
                              ILL
                              7.38
                              v  D
                               18
                                           *   .488 * 10~2
                                                      1 .5
      Where:     L -  Liquid  injection rate  -  gal./lOOO  cf.

                D •  Average Particle Size  -  microns

                D •  Average Droplet Diameter -  microns
                 w
                v -  Throat  velocity of  gas -ft/sec

                v = Gas viscosity - centipoises

                p =  Particle density -  lb  /ft3
                 p                       m •

                K «=  Dimensionless correlation coefficient

                Q -  Flow Rate
        a.  Calculate the efficiency of impaction for each of the following:
Process
M.M.D. ,p
Q, cfm
L,gal/1000cf
P ,lb /ft3
p m
t, °F
1
2.00
50,000.00
8
120

68
2
1.00
50,000.00
8
120

68
3
1.00
70,710.00
8
120

68
4
1.00
50,000.00
16
120

68

-------
b.   Explain why the efficiency decreases with decrease in average  particle
    si ze.

-------
        What Is the effect of increasing velocity or liquid injection rate
        on efficiency?   Why?
loC

-------
Problem 18.  Venturl Scrubber
     In a venturi scrubber, the velocity at the throat is 328 ffeet per
     second.  The temperature of the carrier gas in 86°F.  The density
                                                  Ibm
     of the dust particles to be collected is 187 —- .   The liquid
                                                  ftJ
     injection rate is 2.0 ft3/1000 ft3 of air.  What is the minimum
     size particle in microns that can be removed with 98% efficiency?
     K for  the system is 1.52.

-------
        Problem  19.  Fabric Filter
             Pilot-plant tests have been run with a bottom feed, single-section
              (discontinuous operation)  baghouse on the exhaust gases from  an
             electric-arc steel furnace.  The fume concentration was measured
             ajLjS^JLgr/jfJ^ •  FUtEStifln—time (i.e., time between filter
             cleaning) was maintained at 90 minuses.  Volumetric flow thru  the
             baghouse was varied for a seTTe"s"™oF tests which produced the
             'following pressure drop data.3
uf
fpm
2
3
4
5
AP (in. H20)
APr
1.5
1.9
2. it
3.3
APt
2.7
3.5
4.6
6.6
              It  is desired to install a multi-section (continuous operation)
              baghouse consisting of six compartments.  The time between
              cleaning consecutive compartments (cycle time) is to be 30 minutes.
              Calculate the  maximum  pressure  drop  that will  be  anticipated  with all
              compartments on stream,  for  each velocity  level given in the
              table.
W.W  Campbell & R.W.  Fullerton,  "  Development of an Electric-
"urnace Dust-Control  System"   JAPCA 12,  574-577, 590  (Dec.,  1962).

-------
          a.   Write the applicable relationships.
              S   = S  + miS
               et    r
Where:  S  •=  Filter drag, in. H20/fpm


       AP  -  Pressure drop, in.H20


       Uf  -  Filter Velocity, fpm

                                              j
       C   -  Particulate concentration, gr/ft


       6   *  time, min,


       K   =  Dust permeability



Subscripts:


       t  =  terminal


       r  =  residual


       e  =  effective, when multiple compartments are used.
       o
        fvj
        CD
       -C
        0
AP3 = Se3 Uf3

AP2 = se2 uf2
                                                  AP1  =  Sel  Ufl
                       "t.     o
                        b     c

                       6, minutes

-------
(el
              b.   From  the experimental data done on a single section baghouse,
                  determine the dust  permeability (K) for ^uch of the filter
                  velocities.   Write the applicable equation below:
                   The figure below illustrates graphically  what  is
                   taking place for U, - 2 and  5.   Study  it,  then complete  the
                   table for Uf - 3 and 4:
                                                   200
                                         W, gr/ft'
                            300
                          2.7
1.35
                                            AP
                                            1.5
            r

         0.75
                                                               AS
                                                              0,._60
                                     150
                                            1,
                                   1,1$
                          6.6
1.32
3.3
0.66
                               rr
                                                              0^6-
3A1
               ;.   Plot K vs  Uf  on Figure  1.   Then  explain
                   why K is not  a  constant.

-------
CL



O
 
J-
Ol
O)
     200

     100
                             -ff
                                              T-|-rr
                                                    J±i
                                                                                  :.T:.:J-
                                             345


                                         . -  Nominal  face velocity ( fpm )
                Figure  1    Permeability of dust  cake  as  a function of face velocity

-------
        d.  Knowing K, write the equation for estimating Set  for a  six compartment
            baghouse.
         e.   Calculate AS for the compartments in new baghouse for each Uj.
(tic
I'f
2
3
4
5
AS




S
r




0.63 AS




set




APt




          f.   "loL  AP  vs.  U   and compare  with given data.

-------
                 Maximum pressure drop with all  compartments on steam as a function of  face velocity
o  6
                                                                                   -H~r
                                                     I I  ! I
                                                     r
                                                                                    H-l-
                                                                                    T i
                                                                      t±J

di
TT

-------