EPA-600/2-77-Q36
February 1977
Environmental Protection Technology Series
              EFFECTIVE  SAMPLING TECHNIQUES FOR
                       PARTICIPATE EMISSIONS FROM
                     ATYPICAL STATIONARY SOURCES
                                          Interim Report
                                  Environmental Sciences Research Laboratory
                                      Office of Research and Development
                                      U.S. Environmental Protection Agency
                                Research Triangle Park, North Carolina  27711

-------
                 RESEARCH REPORTING SERIES

 Research reports of the Office of Research and Development, U.S. Environmental
 Protection  Agency, have been grouped into five  series. These five broad
 categories were established to facilitate further development and application of
 environmental technology. Elimination of traditional  grouping was consciously
 planned to foster technology transfer and a maximum interface in related fields.
 The five series are:

     1.    Environmental Health Effects Research
     2.    Environmental Protection Technology
     3.    Ecological Research
     4.    Environmental Monitoring
     5.   Socioeconomic Environmental Studies

 This report has  been assigned  to  the ENVIRONMENTAL PROTECTION
 TECHNOLOGY series. This series describes research performed to develop and
 demonstrate instrumentation, equipment, and methodology to repair or prevent
 environmental degradation from point  and non-point sources of pollution. This
 work provides the new or improved  technology required for the control and
 treatment of pollution sources to meet  environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                         EPA-600/2-77-036
                                         February  1977
           EFFECTIVE  SAMPLING  TECHNIQUES
          FOR PARTICIPATE  EMISSIONS  FROM
            ATYPICAL  STATIONARY  SOURCES
                  Interim  Report
                        by
           H.  A.  Hanson and D.  P.  Saari
         FTuiDyne Engineering Corporation
            5900  Olson Memorial  Highway
           Minneapolis, Minnesota   55422
              Contract No.  68-02-1796
                  Project Officer

                  Thomas E.  Ward
Emissions Measurement and Characterization Division
    Environmental Sciences Research Laboratory
   Research Triangle Park, North Carolina  27711
    ENVIRONMENTAL SCIENCES RESEARCH LABORATORY
        OFFICE OF RESEARCH AND DEVELOPMENT
       U.S. ENVIRONMENTAL PROTECTION AGENCY
   RESEARCH TRIANGLE PARK, NORTH CAROLINA  27711

-------
                          DISCLAIMER
This report has been reviewed by the Environmental Sciences
Research Laboratory, U. S. Environmental Protection Agency,
and.approved for publication.  Approval does not signify
that the contents necessarily reflect the views and policies
of the U. S. Environmental Protection Agency, nor does mention
of trade names or commercial products constitute endorsement
or recommendation for use.

-------
                            ABSTRACT


     Techniques and instrumentation for sampling strategies to
measure particulate emissions from "atypical" stationary sources
were developed.  The four atypical source categories are low
effluent streams, extended dimensions, partially or totally
unconfined flow, and saturated gas streams or gas streams with
entrained liquid droplets.  The research program included litera-
ture surveys, laboratory model testing, and field testing of
several atypical stationary sources.  Techniques and instruments
were evaluated as to the degree of reliability of measured
emissions and applicability to general situations.

     Three specific sources — gravity roof ventilators, grain
dryers, and wet scrubbers -- were selected to provide the basis
for the research program of the four atypical source categories.
Basic characteristics of these sources were identified through
literature and personal contact surveys.  A program of model
testing and field testing Of roof ventilator emissions was
completed, and a similar program was undertaken for wet scrubbers.
The sampling strategy recommended for roof ventilator emission
measurement on the basis of the test program  includes a high
volume particulate sampler and a heated thermopile anemometer
deployed near the base of the ventilator.

     Future work will include the completion of test programs
related to wet scrubbers and grain dryers.  The implementation
of these test programs will be based on the information gathered
during the first year and presented in this interim report.
                               iii

-------
                          CONTENTS

                                                              Page

Abstract                                                       iii
Figures                                                        vii
Tables                                                           x
Acknowledgements                                                xi

I.   Introduction                                                1

II.  Classification and Evaluation of Emission Sources           3
     A.   Emission Sources with Low Velocity and/or
          Extended Dimensions                                    3
          1.   Partially Confined Emission Sources               3
               a.  Roof Ventilators                              3
               b.  Grain Drying and Handling                    13
               c.  Incinerators                                 17
          2.   Unconfined Emission Sources                      17
               a.  Stone Crushing and Asphalt Plants            17
               b.  Agricultural Burning                         20
               c.  Emission Measurements and Standards          20
     B,   Emission Sources with Saturated Gas Streams
          or Entrained Liquid Droplets                          21
          1.   Sulfuric Acid Plants                             21
          2.   Asphalt Plants                                   23
          3.   Industrial Dryers                                23
          4.   Wet Scrubbers                                    23
     C.   Selection of Sources for Test Prbgram                 24
          1.   Roof Ventilators                                 24
          2.   Grain Dryers                                     24
          3.   Wet Scrubbers                                    25

III. Roof Ventilator Sampling Techniques                        26
     A,   Review of Sampling Methodology                        26
     B.   Preliminary Field Tests                               30
     C.   Model Studies                                         41
          1.   Model Design and Fabrication                     41
          2.   Flow Velocity Studies                            49
          3.   Particulate Concentration Studies                53
     D.   Final Field Tests                                     56
     E.   Evaluation of Sampling Technique                      58
     F.   Applicability to Other Emission Sources               70

IV.  Wet Scrubber Sampling Techniques                           73
     A.   Review of Sampling Methodology                        73
     B.   Preliminary Field Tests                               78
     C.   Model Studies                                         89

V.   Future Work                                                91
     A.   Wet Scrubber Sampling Techniques                      91
     B.   Grain Dryer Sampling Techniques                       91

                              v

-------
                       CONTENTS  (Cont.)


VI.  Interim Conclusions and Recommendations                  92

References                                                    93

Appendix   Velocity Instrumentation for Low
           Velocity, Partially Confined
           Source Particulate Sampling                        97
                              VI

-------
                          FIGURES
Number                                                   Page
  1     Continuous Roof Ridge Ventilator                   7
  2     Roof "Monitor or Monitor Attachment"               7
                            ,'
  3     Rack and Column Type Grain Dryers  (Ref. 6)        15
  4     Unloading Grain from Box Car into a Deep
        Hopper  (Ref. 12)                                  18
  5     Unconfined Emission Sources at a Continuous
        Mix Asphalt Plant  (Ref. 5)                        18
  6     Cross Section of Roof Ventilator Showing
        Test Equipment  (Ref. 29)                          28
  7     Sampler System  (Ref. 30)                          28
  8     Emission Rate Measurements, Central Sampling
        Station  (Ref. 30)                                 29
  9     EPA Method 14 Sampling System                     29
  10    Roof Ventilators at Hitchcock Industries,
        Bloomington, Minnesota                            31
  11    Typical Roof Ventilator Exhaust at
        Hitchcock Industries Site                         31
  12    High Volume Sampler and Probe Assembly            33
  13    Hot Wire Anemometer and Protective Collar
        Assembly                                          34
  14    Sampling Locations for Preliminary Field
        Tests - Hitchcock  Industries Roof Ventilator      35
  15    High Volume Sampler Assembly Mounted in
        the Roof Ventilator at Hitchcock Industries
         (End View looking North)                          37
  16    High Volume Sampler Assembly Mounted in the
        Roof Ventilator at Hitchcock Industries
         (Top View Looking down into Exhaust Region)       37
  17    Preliminary Field  Test Concentration Measure-
        ments, Hitchcock Industries Roof Ventilator       38
  18    Preliminary Field  Test Velocity Survey 1,
        Hitchcock Industries Roof Ventilator              39
  19    Preliminary Field  Test Velocity Survey 2,
        Hitchcock Industries Roof Ventilator              40
  20    Roof Ventilator Model                             42
  21    Roof Ventilator Model Test Facility -
        Rosemount Energy Conversion Laboratory            43
                            vii

-------
                      FIGURES (Cont.)

Number                                                   Page

  22    Roof Ventilator Model Test Section                44

  23    Atomization Aerosol Generator and Hot Air
        Injection System - Rosemount Energy
        Conversion Laboratory                             45

  24    Roof Ventilator Model at Medicine Lake
        Aerodynamic Test Facility-Viewed from North       46

  25    Roof Ventilator Model and Ducting at Medicine
        Lake Aerodynamic Test Facility-Viewed from
        Northeast                                         46
  26    Atomization Aerosol Generator System with
        Hot Air Injection-Medicine Lake Aerodynamic
        Test Facility                                     47
  27    Roof Ventilator Model Sample Ports-
        Viewed from Northeast                             48

  28    Flow Direction Indicating Tufts in
        Roof Ventilator Model Exhaust                     50

  29    Smoke Generator Flow Indicator in Roof
        Ventilator Model Exhaust                          50

  30    Schematic of Observed Flow  (End View of
        Roof Ventilator .Model)                            51

  ,31    Typical Flow Velocities at Various Points
        in Roof Ventilator Model                          52

  32    Volumetric Flow Rate at Base of Roof
        Ventilator Model                                  54

  33    Particulate Concentration Measurements
        in Roof Ventilator Model                          55

  34    Particulate Concentration and Velocity
        Measurements at Base of Hitchcock
        Industries Roof Ventilator                        57

  35    Average Velocity and Particulate Concen-
        tration Measurements-Sample Port lf
        Hitchcock Industries Roof Ventilator              59

  36    Average Velocity and Particulate Concen-
        tration Measurements-Sample Port 2,
        Hitchcock Industries Roof Ventilator              60

  37    Average Velocity and Particulate Concen-
        tration Measurements-Sample Port 3,
        Hitchcock Industries Roof Ventilator              61
                           viii

-------
                      FIGURES  (Cont".)
Number                                                   Page

  38    Average Velocity and Particulate Concen-
        tration Measurements-Sample Port 4f
        Hitchcock Industries Roof Ventilator              62

  39    Average Velocity and Particulate Concen-
        tration Measurements ^-Sample Port 5,
        Hitchcock Industries Roof Ventilator              63

  40    Typical Velocity Distribution at Base of
        Hitchcock Industries Roof Ventilator              66

  41    Particulate .Concentration Profiles for
        Roof Ventilator Emission Example Problem          68

  42    Average Velocity Profiles for Roof
        Ventilator Emission Example Problem               69

  43    Inertial Impaction Liquid Droplet Separator
        Used by a Wet Scrubber Manufacturers Test Group   75

  44    Method Used to Determine Axial Component in
        a Single Vortex Cyclonic Flow                     76

  45    Velocity Error with Yaw Angle  (3/8" S-tube)       77

  46    Several Types of Directional Pitot Tubes
         (Ref. 43)                                         79

  47    Pressure Distribution over a Cylinder in
        Cross Flow                                        80

  48    Typical Fecheimer Probe and Pressure
        Monitoring System  (Ref. 44)                       80

  49    Fecheimer Probe Built into Filter Holder
         (Ref. 45)                                         81

  50    Connecticut State Department of Environmental
        Protection Probe  (Ref. 46)                        82

  51    Wet Scrubber Exhaust Ducts at Seneca Waste-
        water Treatment Plant, Eagan, Minnesota           84

  52    Filter Samples Taken during Preliminary Field
        Tests at Seneca Wastewater Treatment Plant        85

  53    Inertial Separation Precutter used in Pre-
        liminary Field Tests                              87

  54    Internal View of Inertial Separation Precutter    87

  55    Typical Velocity Profile Obtained During
        Preliminary Field Tests at Seneca Wastewater
        Treatment Plant                                   88

  56    Proposed Cyclonic Flow and Entrained Liquid
        Droplet Test System                               90
                            ix

-------
                          TABLES

Number                                                   Pagj

  1     Classification of Sources with Respect to
        Emission Measurement Parameters                   4
  2     Partially Confined Emission Sources               6
  3     Performance Table for Gravity Roof
        Ventilators (Ref. 9)                              9
  4     Industry Reported Emission Data for
        Primary Aluminum Facilities (Ref. 11)            10
  5.    Summary of Emission Data - Primary
        Aluminum Industry (Ref. 11)                      12
  6     Emission Factors for a Vertical Stud Soder-
        berg Potline - EPA Test Results  (Ref. 11)        14
  7     Summary of Grain Dryer Emissions at Grain
        Elevators (Extracted from Ref. 6)                16
  8     Unconfined Emission Sources                      19
  9     Emission Sources with Saturated Gas Streams
        or Entrained Liquid Droplets                      22
  10    Average Velocity Determined from Widthwise
        Velocity Surveys, Based on Fig. 40                67
  11    Calculated Emissions for Example Problem          71

-------
                      ACKNOWLEDGEMENTS

     The authors wish to express their appreciation to the
Project Officer, Mr. Thomas E. Ward, for his guidance and
assistance during the course of this project.

     Special thanks are due to both Hitchcock Industries,
Bloomington, Minnesota, and the Minneapolis-St. Paul area
Metropolitan Waste Commission.  The cooperation and assistance
of all members of these organizations in allowing the
completion of field test programs is greatly appreciated.

     Grateful acknowledgements are also expressed to a great
number of regulatory agencies, environmental management and
testing organizations, pollution equipment manufacturers,
and others who provided valuable information for use in the
project.

     The authors also wish to extend thanks to Mr. Alfred A.
Iversen and to the many members of the staff of FluiDyne
Engineering Corporation for their parts in the accomplishment
of this study.
                              xi

-------
                         I.   INTRODUCTION

     The capability to determine emission levels is of primary
importance in any program intended to control emissions for two
reasons.  First, reliable emission levels must be established in
order to assess the severity of an emission problem and, if
necessary, set standards for its control.  Second, in order to
enforce compliance with established standards, emissions from a
given source must be accurately determined.  Procedures for
selection of sampling sites, velocity measurements, and deter-
mination of particulate concentration have been established by
the United States Environmental Protection Agency  (Ref. 1) for
emission sources characterized by well defined and constrained
flow fields with velocities greater than 1.5 to 2 m/sec.  Studies
such as Ref. 2 have examined the reliability of emission measure-
ments in sources, such as exhaust stacks and ductwork in large
power plants, which fall into this category.

     As programs to control emissions have met with success, more
interest has been focused on emission sources in which the velo-
city, flow fields, and emissions are not well defined or the
effluent gases are not confined.  Emission sources of this nature
are generally classified as "atypical".  The intent of this study
is to examine the emissions from sources characterized by one or
more of the following:

     1.   low velocity (less than 2 m/sec);
     2.   one or more extended dimensions;
     3.   partially or totally unconfined flow;
     4.   saturated gas streams or gas streams with entrained
          liquid droplets.

Sampling techniques and strategies will be discussed for flow and
particulate concentration measurement related to several specific
atypical emission sources.  The methods will be evaluated in
regard to their usefulness for these specific applications as
well as their applicability to other atypical sources.

     The particular applications which will be discussed are
roof ventilators  (representing a low velocity/extended dimen-
sion source), grain dryers  (partially unconfined flow) and wet
scrubbers (entrained liquid droplets).  Although both roof
ventilators and grain dryers may be classifed as having charac-
teristics under Items 1, 2 and 3, these sources were chosen as
primarily representative of the indicated characteristics for
determination 'of sampling techniques.

     Having identified these specific applications as represent-
ative of the various atypical source categories, the basic
procedure followed in investigating each emission  source was as
follows:

-------
     1.    Identification of typical characteristics of the
          source through preliminary field testing;

     2.    Design and fabrication of a representative model of
          the source;

     3.    Testing of sampling techniques in the controlled
          model environment, identification of a useful sampling
          technique, and evaluation of accuracy;

     4.    Final field tests, if necessary, to confirm the
          applicability of the sampling technique to actual
          field situations;

     5.    Assessment of the applicability of the selected
          sampling technique to other atypical sources.

An overall evaluation of atypical emission sources will be
presented first, followed by a discussion of the implementation
of the above program for each of the selected sources.
                              - 2' -

-------
              II.   CLASSIFICATION AND EVALUATION OF
                     ATYPICAL EMISSION SOURCES

     A survey of a number of atypical emission sources was made
in order to determine which of the characteristics of interest
apply to the various sources and, furthermore, to attempt to
rank these sources in importance for consideration in this
study.  Information was extracted from the general literature
as well as from contacts with instrumentation manufacturers,
commercial firms engaged in emission measurements, and eight
state pollution agencies in eight different EPA regions.

     Table 1 includes a list of several atypical emission
sources and classifies them with respect to the various char-
acteristics.  The term "low velocity" here refers to sources
having emission stream velocities which may often be less than
2 m/sec.  Sources such as roof ventilators and louvered panels
confine the flow for a short path length and are characterized
here as "partially confined".  The distinction is drawn between
partially confined sources and unconfined sources, such as open
field burning or agricultural tilling.  Brief descriptions of
some of the sources outlined in Table 1 and evaluations of the
degree of importance of the various sources are included in the
following.

A.   Emission Sources with Low Velocity and/or
     Extended Dimensions

     1.   Partially Confined Emission Sources

     Several partially confined emission sources are listed in
Table 2 together with information extracted from Refs. 3
through 8 concerning emission characteristics of these sources.
Such sources are often also characterized by one or more extend-
ed dimensions and low velocity effluent streams, as are the
sources listed in Table 2.  The particular configuration of
each source type, however, will usually establish one or two of
these characteristics as the primary factor or factors in
defining a sampling technique.

          a.   Roof Ventilators

     Roof ventilators as emission sources are found in a number
of industrial applications, some of which are indicated in
Table 2.  Continuous gravity ventilators are typically utilized
in situations requiring the removal of large heat loads.
Basically, two styles of gravity ventilators are marketed
commercially.  Figures 1 and 2 illustrate the two basic types.
The type with a single emission plane  (Fig. 1) may extend along
the entire length of a building and is generally referred to as
a "continuous roof ridge ventilator", while the two-channel
                              - 3 -

-------
              TABLE 1
CLASSIFICATION OF SOURCES WITH RESPECT TO
     EMISSION MEASUREMENT PARAMETERS
Low
Source Velocity
Roof ventilators
Grain dryers
Wigwam incinerators
Unload grain box
cars-trailer trucks
into deep hoppers
Agricultural tilling
Unpaved road dust
Open field burning
Construction and
land excavation
Open baghouses
Open incinerators
Mining and quarry
activities
Aggregate stockpiles
Handling & transfer
at:
Grain elevators
Cotton gins
Cement plants
Asphalt plants
X
X
X
X
X
X
X
X
X
X
X
X

X
X
X
X
Extended Confinement Saturated
Dimensions Duct Partial None Gas
x x
X X
X
X X
X X
X X
X X
X X
x x
x x
x x
x x

x x
x x
x x
x x
Liquid Weather
Droplets Sensitive
x
X
X
X
X
X
X
X
X
X
X
X

X
X
X
X

-------
                                                 TABLE 1 (CONT'D)
    Source
                      Low        Extended        Confinement        Saturated   Liquid      Weather
                    Velocity    Dimensions    Duct  Partial  None      Gas      Droplets    Sensitive
I
Ui
I
Open vats

Wet scrubber outlets
Absorber exhaust
  sulfuric acid plants
Saturators and
  blowers-asphalt
  roofing mfg.

Paint booths
Industrial dryers,
  e.g. plywood
  veneer dryer

Ventilators emitting
  oil mist, e.g.,
  machine shop,
  underground
  construction
                                                    x

                                                    X
                                                    X

                                                    X
X

X

X
X

X
X

X
X

X

-------
                                                    TABLE 2

                                      PARTIALLY CONFINED EMISSION SOURCES
Source

Roof Ventilators**

1) Aluminum mill
   reduction  cells

   a) H.S.Soderberg
   b) V.S.Soderberg
   c) Prebake

2) Open hearth  steel
   furnaces
   a) No oxygen lancing
   b) With  oxygen lancing

3) Electric arc steel
   furnaces

   a) No oxygen lancing
   b) With  oxygen lancing

4) Iron Foundary furnaces

Grain Elevators

1) Grain Dryers

2) Grain loading/
     unloading
                         Velocity at
                      Plane of Emission
                      	m/sec.	

                           0.5 - 5
                                     Avg. Particulate
                                       . Loading
                                            0.07-4.6
Particulate*
   Size
Emission Rate
   Factor
gm  part per
Kg of product
                                                        Range down
                                                        to submicron
                  72
                  42
                  32
0.9
3.4
0.2-5
2.3-25
2.3-7
9-80
0.25
50% < 5 u
45% < 5 \i
60% < 5 {A
60% < 5 u
10% < 5 p.
Geo . means
Oats 3.1 y
Wheat 2.1 y
4
11
5
5
8
3-3.5
1.5
 Estimated Total
Annual Emissions
  metric tons
                    32,000
                     9,100
                    18,200

                   306,000
                                                                                              16,500
                                                                                             100,000
                                                                                               35,000
                                                                                             110,000
Wig-Warn incinerators
                                        0.39
24% < 2
  * Size distribution information based on % by weight less than stated size.
 **
                   120,000
^ .i.^C U^£> L* i. J-U U- (•• -tWli .fciUJt- vs «• u*c* b. j.w** i^**v* »••»* v*» tv *" J ™ •^-»-^-" ••• —••»-—;—• —•-- ——— — — — —— —	
The figures tabulated represent total emission quantities through all controls and exhausts  for  the
activities listed and are included in order to establish their relative importance.

-------
     FIGURE  1.   CONTINUOUS ROOF RIDGE VENTILATOR
FIGURE 2.  ROOF  "MONITOR"  OR "MONITOR ATTACHMENT"
                          - 7 -

-------
exhaust style  (Fig. 2) is usually denoted by the terms "moni-
tor" or "monitor attachment".  The term "roof monitor" however,
is often used as a description of either style of gravity
ventilator or even certain types of powered ventilators.

     Gravity ventilators find application in steel and aluminum
mills, foundries, forge shops and glass plants, as well as
warehouses, zoos, and other locations where air conditioning
units are not economically practical due to large heat loads
and/or large interior spaces.  Both types of pre-built venti-
lators, continuous roof ridge ventilators and roof monitors,
are commercially available in 10 ft. lengths which may be
spliced together to form runs of any desired length.  The
continuous roof ridge ventilator is available in sizes having
throat widths ranging from 4 in. to 15 ft. and heights of less
than 1 ft. to 18 ft. above the roof peak.  The monitor type is
available with openings ranging from 3 ft. to 15 ft. Both types
are equipped with dampers which can be opened or closed to
exhaust or retain building heat as required.

     Roof ventilators often extend along the entire length of
a roof ridge.  Heat, fumes and particulate matter which evade
primary control devices are propelled upward with thermal
currents and may be emitted through the roof ventilator, which
may often be considered a line source due to the large length
dimension involved.  The volume of emissions is dependent upon
wind speed, temperature, building design, louver adjustment,
and level of activity in the ventilated area.  Table 3 is a
sample from a manufacturer's bulletin  (Ref. 9) which illustra-
tes the effects of temperature, height, and wind speed on flow
capacity through a roof ventilator.  Such tables are generally
calculated on the assumption of wind direction perpendicular to
the ventilator length, based on the negative pressure which
develops as the flow passes over the ventilator and creates an
aspirating effect.  Winds directed along the length of a venti-
lator will have much smaller effects than indicated in Table 3.

     Certain general principles for roof ventilator design and
performance calculations are specified in handbooks, such as
Ref. 10.  Detailed analyses of the flow are not available,
however.  In general, it may be said that flow rate and par-
ticulate concentration in roof ventilator exhausts are poorly
understood, due to the large number of factors which affect
performance.

     An indication of the importance of roof ventilators as an
emission source is given by Tables 4 and 5.  This information
was extracted from Ref. 11, which is a study of only one of the
industrial applications listed in Table 2, the primary  aluminum
industry.  This data is based on a survey of aluminum production
facilities throughout the United States.  As indicated  in
Tables 4 and 5, since the primary collection efficiency of most

                             -  8  -

-------
          TABLE 3



PERFORMANCE TABLE FOR GRAVITY

  ROOF VENTILATORS (Ref. 9)
            C.F.M. Per Square Feet of
             Ventilator  Throat Opening
Temp . Stack
Diff. Height
10 ft.
20
10 °F 30
40
50
10 ft.
20
20°F 30
40
50
10 ft.
20
30°F 30
40
50
2 MPH
193
236
271
298
323
236
298
345
385
414
271
345
403
451
494
At Wind Velocities of:
4 MPH 8 MPH 10 MPH
281
324
359
386
411
324
386
433
473
502
359
433
491
539
582
395
438
473
500
525
438
500
547
587
616
473
547
605
653
696
457
500
535
562
587
500
562
609
642
678
535
609
667
715
758
              1 foot =  .3048 meters
          -  9 -

-------
              TABLE 4
 INDUSTRY REPORTED EMISSION DATA FOR
PRIMARY ALUMINUM FACILITIES (Ref. 11)
% Capacity
Reporting
Prebake Potlines
Total effluent
Primary collection
Secondary collection
Primary emission
Secondary emission
Total emission

60
83
65
76
65
65
Kg Particulate/Metric Ton Aluminum
high

88.6
84.5
7.8
12.5
7.8
16.3
Overall control efficiency
Vertical Stud Soderberg
Total effluent
Primary collection
Secondary collection
Primary emission
Secondary emission
Total emission

—
—
—
89
—
—

	
	
	
6.5
	
	
Overall control efficiency
Horizontal Stud Soderberg
Total effluent ,
Primary collection
Secondary collection
Primary emission
Secondary emission
Total emission

93
93
93
93
93
93

52.0
42.0
10.4
10.1
10.4
20.5
Overall control efficiency
average

47.2
43.8
4.0
4.6
4.0
8.1
81%

39.2
22.0
11.0
4.4
6.7
11.7
70%

49.2
39.1
10.1
8.9
10.1
19.0
61%
low

22.5
21.6
2.0
1.1
2.0
3.5


	
__^_
	
2.2
	
	


41.8
31.4
10.0
8.5
10.0
18.5

             -  10 -

-------
                              TABLE 4 CONT'D
                        % Capacity   Kg Particulate/Metric Ton Aluminum
                         Reporting
All types
   Total effluent           63
   Primary collection       82
   Secondary collection     71
   Primary emission         82
   Secondary emission       71
   Total emission           71
   Overall control efficiency
high
88.6
84.5
7.8
24.4
7.8
23.5

average
47.7
40.3
6.9
5.9
6.4
12.3
73%
low
22.2
16.4
2.0
1.1
2.0
3.5

                                 -  11 -

-------
                                  TABLE 5
                         SUMMARY OF EMISSION DATA-
                    PRIMARY ALUMINUM INDUSTRY (Ref.  11)
                                              Emission Factor
                                      Kg Particulate/Metric Ton Aluminum^
                           Prebake      V. S.      H. S.     All types
                                  H-
Total effluent              47.2        39.2       49.2      47.7
Primary emission             4.6         4.4        8.9       5.9
Secondary emission           4.0         6.7       10.1       6.4
Secondary collection         4.0        11.2       10.1       6.9
Portion of total effluent
   emitted through roof
   monitor                   8.5%       17.1%      20.5%     13.4%
Efficiency of secondary
   (roof monitor) control   50%         63%        50%       52%
                                - 12  -

-------
aluminum process facilities is quite high and secondary collec-
tion is either very low in efficiency or nonexistent, the
particulate emission through roof ventilators generally equals
or exceeds the total emissions from primary particulate control
systems.  Table 6 includes data/ extracted from Ref. 11, related
to EPA tests of a single vertical stud Soderberg potline, which
indicates similar conclusions.

          b.   Grain Drying and Handling

     A number of partially confined emission sources are found
in the grain and feed industry, one of which is grain drying.
Grain drying is required when the moisture content of grain
received at an elevator is too high for the grain to be safely
stored.  Drying is normally required when handling corn, althoug:
many other grains may require drying under certain conditions.
Two common types of grain dryers, rack and column dryers, are
illustrated in Fig. 3.  Heated air is generally used as the
drying medium, and newer grain dryer designs often incorporate
continuous recirculation of a portion of this heated air for
higher drying efficiency.

     A survey of grain elevators in 1973 indicated that roughly
1.6 million bushels  of grain are dried annually  (Ref. 6).
Contacts with agricultural associations and pollution control
agencies indicated that virtually all corn elevators use grain
dryers, while dryers are not as prevalent in handling wheat and
other types of grain.  Overall, grain dryers are found to have a
large degree of application.

     Emission levels for grain dryers, extracted from Ref. 6,
are indicated in Table 7.  Several other emission sources in
the grain and feed industry, such as headhouses, are indicated
in Ref. 8 as having higher total annual emissions than grain
dryers.  However, grain dryers present unique problems in
emission measurement due to the large surface area of the
emission plane, low velocities, and large particulate size
 (Ref. 6).  While reliable emission measurements have been
obtained for most emission sources in the grain and feed
industry using Method 5 equipment and techniques, emission data
for grain dryers has been shown to be very unreliable due to
these sampling problems.

     In additon to hulls, cracked grain, weed seeds, and field
dust, grain dryer emissions include other large, lightweight
particles denoted as "chaff", or "beeswing" in the case of
corn, which break off from the grain during drying and handling.
The low specific gravity of emitted particulate matter inten-
sifies the sampling and collection problems associated with the
partially unconfined nature of grain dryers as an emission
source.  The emission level is affected by a number of factors,
                               ',
                              - 13 -

-------
                           TABLE 6
       EMISSION FACTORS FOR A VERTICAL STUD SODERBERG
            POTLINE - EPA TEST RESULTS (Ref. 11)
                                  Emission Factor
                              kg Particulate/Metric Ton Aluminum
                                   (1)        (2)
Primary collection                45.63      33.80
Secondary collection              13.56      13.34
Primary emission                   0.06       0.06
Secondary emission                 4.77       2.92
Total emission                     4.83       2.97
Primary efficiency                99.86%     99.84%
Secondary efficiency              64.85%     78.15%
Overall efficiency                91.85%     93.70%
     (1)   Average of 3 tests.
     (2)   Average of 2 tests; 1 test deleted due
          to stud blow during test.
                          - 14 -

-------
                                      MOISTURE
                                        LADEN
                                         AIR
                                         OUT
                                                    Dryer Section
         (a)  Rack Dryer
                                                         Cooler
                                                         Section
                                                                      Grain
                                                                     Receiving
                                                                      Garner
                                                                        n
                                                                        a
                                                                        §-
                                                                        n
                                                              Variable  Speed Discharge

                                                                    (b) Column Dryer
FIGURE 3.   RACK AND  COLUMN TYPE GRAIN DRYERS  (FROM REF.  6)

-------
                       TABLE 7
      SUMMARY OF GRAIN DRYER EMISSIONS AT GRAIN
          ELEVATORS (Extracted from Ref. 6)
                          Total Emissions  (1971)
                             (metric tons/yr)
Country Elevators             2.14 x 104
                                       o
Terminal Elevators            2.53 x 10
Export Elevators              3.82 x 102
                       - 16 -

-------
including the type of dryer, the type and hardness of the
grain, the moisture content of the grain  (typically 10-20%) and
the amount of foreign material in the grain.

     Another partially confined source related to grain elevators
is the deep receiving hopper.  Grain is dropped from boxcars or
trucks in surges from a height of 1 to 5 m  (Fig. 4).  The
falling grain particles disperse as they accelerate and induce
into motion a column of air moving in the same direction.  When
the mass of particles strikes the hopper bottom, the dissipated
energy results in extreme turbulence as well as abrasion and
dispersion of the particles.  The dust generated may form a
plume of 100% opacity and sufficient volume to envelope an
entire boxcar.  Sampling of such sources is usually accomplished
by assessment of the ambient conditions.

          c.   Incinerators

     Incineration of sawmill wood and bark wastes is a primary
air pollution source in the lumber industry.  Effluent gas
velocities from wigwam incinerators may be on the order of 3
m/sec.  Air quality problems may often be reduced by improvements
in combustion efficiency.
     2.   Unconfined Emission Sources

     Emission sources from which the pollutants are not con-
strained by process streams are classified as unconfined
sources.  Unconfined, or fugitive, emissions are generated by
natural means as well as industrial processes.  Dust arising
from wind erosion has been estimated to comprise 20% of the
total annual worldwide aerosol production  (Ref. 13).  Fugitive
dust, sources are now being recognized as possible causes for
regional noncompliance with air quality standards.

     Several activities which may generate unconfined emissions
are summarized in Table 8.  The data presented in this table
was extracted from Refs. 3 and 5.  For many unconfined emission
sources, published data is either of a qualitative nature or is
unavailable.

          a.   Stone Crushing and Asphalt Plants

     Fugitive emissions in the stone crushing industry can be
generated by activities such as drilling, blasting, crushing,
conveying, removal of fines, storage and loading.  Most emis-
sions are of heavy particles that settle near the source.
Crushing is done in three stages, and the third crushing and
screening processes contribute the major portion of the emis-
sion.  The degree of enclosure of the processing and transfer
                             - 17 -

-------
                                ,-OUST
                               X PLUME
   FIGURE 4.  UNLOADING GRAIN FROM BOX-CAR
               INTO A DEEP HOPPER (REF.  12)
       1ndtc«t«i lugifir* dusl 1ot»t«
       If no! «•>! •nelet«4
FIGURE  5.   UNCONFINED EMISSION SOURCES AT A
            CONTINUOUS MIX ASPHALT  PLANT  (REF. 5)

                   - 18  -

-------
                                             TABLE  8

                                    UNCONFINED EMISSION  SOURCES
10
I
Source

Crushing

   a)  Stone
   b)  Sand & Gravel

Asphalt Plants

Agricultural Field
  Burning

Cement Plants

Cotton Gins

Aggregate Stockpiles
                                  Average
                                  Particulate
                                 Concentration
                                    gm/m3	
                                    13-87
.01-.06

2-35

.004-1.25
            Particle
              Size
          Emission Rate
            Factor
          gm Particulate
          per kg Product
             10-50|i*      8
             3.5-9.4u*    0.05

             40%<10u***   4
0.5n*

30%<5ji***
8.5

26

5kg/bale

0.1-0.5%**
                Estimated Total
               Annual Emissions
                 metric tons
                              4,150,000
                                 42,000

                                 37,000
2,200,000

  172,000

   41,000

6,000,000
      * Typical mass  median  diameter.

     ** Stockpile  losses  due to wind erosion.

    *** Size  distribution information based on  % by weight less than stated size.

-------
areas and the degree to which control equipment is used in the
various processes determine the quantity and concentration of
emissions.  Possible fugitive emission sources at asphalt
plants are indicated in Fig. 5.  Dust is generated at elevator
transfer points, hot screens, and hot and cold aggregate bins.

          b.   Agricultural Burning

     Smoke from agricultural field burning is composed primarily
of carbon particles, ash, and various gases.  These emissions
cause reduction of visibility and odors.  Emission plumes
consist mostly of particles less than 1.3 microns in diameter,
identified as medium-and large-molecule hydrocarbons which
provide a serious potential for lung irritation CRef. 14).
Emissions from field burning are highly dependent on wind,
ambient temperature and type of fuel being consumed.  Slash
burning, primarily practiced in the Western United States to
reduce the flammability of heavy slash concentrations after
timber cutting, has been estimated by the United States Depart-
ment of Agriculture to produce 5.5 million metric tons of
particulate each year.

          c.   Emission Measurements and Standards

     A number of studies have been made concerning the predic-
tion of plume heights and concentrations related to unconfined
sources using various dispersion models.  Actual measurement of
emissions from unconfined sources is poorly defined, however.
Most measurement studies have utilized ambient monitoring
equipment located in the vicinity of an emission source, often
upstream or downstream of the source in the direction of the
prevailing wind.

     Ref. 15 includes empirical equations for emission factors
of four categories of unconfined sources  (agricultural tilling,
unpaved roads and airstrips, heavy construction activities and
aggregate storage piles) which are based on an extensive field
testing program.  The testing reported in Ref. 15 included
isokinetic dust exposure with specially designed sampling
equipment, conventional high volume sampler measurements and
particle size measurements with high volume cascade impactors.
Emission factors were related to meteorological and source
parameters including properties of the emitting surface and
characteristics of the vehicle or implement which caused the
emission, and correction factors were developed to reflect
regional differences in climate and .surface properties.

     Emission standards for unconfined sources are generally  of
a qualitative nature.  Regulatory agencies have normally adopted
regulations prohibiting airborne particulate matter from cross-
ing property lines.  Enforcement of such regulations is gener-
aly handled on a complaint basis.
                             ,- 20 -

-------
B.   Emission Sources with Saturated Gas Streams
     or Entrained Liquid Droplets

     Effluent streams from certain industrial processes are
characterized by saturation with, various types of vapors or by
the presence of entrained liquid droplets.  Several emission
sources of this type are listed in Table 9 along with emission
data extracted from Refs. 3f 5, 8, 16, 17 and contacts with
various information sources.

     Cooling of exhaust gases from evaporators, boilers and
stills below the dew point of any gaseous components, such as
water or acid, can generate large quantities of liquid droplets.
Wet scrubber control systems also introduce droplets into
exhaust gases.  Droplets formed by tearing of liquid sheets and
ligaments and by splashing of liquid drops are generally 200 y
and larger.  The mist formed by condensation of saturated vapor
contains submicron droplets.  Entrained liquid droplets may
cause excessive corrosion, decrease in equipment performance,
or loss of processed material in addition to the air pollution
problems.

     1.   Sulfuric Acid Plants

     The exhaust gases from sulfuric acid manufacturing plants
are common sources of acid mist.  Sulfuric acid is manufactured
by either the chamber process or the contact process (Ref. 16).
The primary emission source in the chamber process is the Gay
Lussac Tower, whose function is to recover nitrogen oxides
released in the manufacturing-process.  Aerosol concentrations
may range from 0.2 - 1.2 gm/m .  Over 90% of the particles, by
mass, are larger than 3 p in diameter.

     The major emission source in sulfuric acid contact plants
is the stack from the absorption tower, which removes most of
the sulfur trioxide.  Small amounts of unabsorbed sulfur triox-
ide and a larger quantity of acid mist are emitted to the
atmosphere.  The presence of a substantial number of particles
smaller than 3 u is evidenced by the appearance of a dense,
white plume at the stack exit which does not necessarily reflect
the mass concentration of sulfuric acid mist.

     Several types of mist eliminators are used with varying
degrees of success in removing the entrained droplets.  Electro-
static precipitators and glass-fiber mist eliminators can
achieve 92-99.9% collection efficiency over the entire particle
size range.  Stainless steel wire-mesh mist eliminators have
low initial cost but have low efficiency during the production
of oleum, a solution of free, uncombined sulfur trioxide, when
the proportion of small particles is sharply increased.  Tall
stacks, on the order of 50 m or larger, have been shown to be
effective in reducing acid spray emissions, since large particles

                              -  21  -

-------
                                                       TABLE  9
                       EMISSION SOURCES WITH SATURATED GAS  STREAMS OR ENTRAINED LIQUID DROPLETS
    Source
    Sulfuric acid plants
       Chamber process
       Contact process
       Spent acid concen-
             trators
1   Asphalt Roofing Mfg.
IM      Blowers & Saturators
i
    Plywood Veneer Dryers
    Wet Scrubber
       Downstream of Thermal
       Coal Dryer
    Sewage Sludge Inciner-
            ators
                                Flow Rate
600-5500
850
              Avg.  Particulate
                               Emission Rate
                                   Factor
Estimated Total
Nm3. .
/min
55-370
140-1750
1700
280-560
_
Concentrations
0,2-1,2
0.04-1.75-
2-4
1.0-1.8
n_[
Particle
Size
10% < 3 y**
64% < 3 y**
-
^ ly
— ,
gm part per
Kg of Product
2,5
1
15
2
3Kg/1000 m2
Annual Emissions
Metric Tons
1,800
3,600
7,200
15,500
3,600
.070*
0.3*
  94,000*
       * Estimate includes solid particulate.
      ** Size distribution information based on % by weight  less  than stated size.

-------
tend to collect on the stack wall.  Indications are that
approximately 90% by mass of the total acid mist and spray from
an absorber outlet has been collected in an 80 m stack.

     2.   Asphalt Plants

     Another source of vapors and mists are the effluents from
"blowing", or oxidizing of stills and saturators used in the
manufacture of asphalt roofing.  Blowing is accomplished by
bubbling air through the liquid asphalt at a temperature of
about 230-260°C for 8 to 16 hours.  After blowing, asphalt is
transferred to a saturation tank or spray area where it is
sprayed onto one side of felt that is continuously fed from
rollers.  This is done in order to remove moisture from the
felt which could cause blisters when? the felt is saturated by
passing it through a tank of molten asphalt.

    - The high application temperature causes vaporization of
those asphalt components having lower boiling points.  Moisture
in the felt is also vaporized, which contributes to the formation
of a mist of high opacity.  Additional mists and vapors are
emitted as the saturated felt is cooled.  Particulate emitted
from saturators is formed by condensation and, therefore, is
likely to have a size on the order of 1 p. Various hood and
exhaust configurations have been devised for exhausting blower
and saturator emissions which require volumetric capacities of
280-560 Nm /min.

     3.   Industrial Dryers

     Industrial dryers are another source of vapor and liquid
particulate emissions.  Plywood veneer dryers are used to
reduce the moisture content of green veneer panels from about
50% to less than 10% in long, heated, enclosed chambers.  Dryer
design varies greatly.  Emission plumes are saturated with
water vapor and other condensibles including wood resins, resin
acids and wood sugars which form a blue haze when cooled.
Volatile components include terpenes as well as unburned methane
when gas-fired dryers are used.  Emission factors are generally
expressed in terms of pollutant weight per unit of surface area
of I cm  (or 3/8 in) plywood produced.  Condensible compounds
are estimated by Ref. 8 to contribute 63% of the emitted
pollutant, while volatile compounds make up 37%.

     4.   Wet Scrubbers

     Wet scrubbers have found application in a wide range of
effluent control problems.  There are many varieties of wet
collectors, but they may generally be classifed as low energy
or high energy scrubbers.  Simple spray towers, packed towers,
impingement tray scrubbers, and other configurations which
                              -  23 -

-------
operate with low pressure drop may be classified as low energy
scrubbers.  Efficiences can exceed 90-95%.  Air pollution
control systems for incinerators, fertilizer manufacturing
plants, lime kilns, iron foundries, stone crushing operations,
and clay product production commonly incorporate low energy
scrubbers.

     High energy, or Venturi, scrubbers incorporate a converging-
diverging duct section to attain high gas stream velocity while
scrubbing liquid is injected.  Higher collection efficiency is
realized, but the higher efficiency is accompanied by higher
pressure drops requiring high power draft fans.  Common appli-
cations for Venturi scrubbers include steel furnaces, pulp
mills and foundry cupolas.

     The three situations indicated in Table 9 are not repre-
sentative of the large number of scrubber applications, but
they do indicate the broad range of conditions which may be
encountered in scrubber exhaust gases.

c-   Selection of Sources for Test Program

     After evaluating a number of emission sources, including
the information presented in the previous sections as well as
information found in Refs. 17-28, three specific sources were
chosen for the test program.  By limiting the actual test
program to these specific emission sources, it was felt that a
reasonable amount of useful information about each of the
selected sources could be obtained within the alloted limits of
the program.  If the selected sources are sufficiently represen-
tative of the atypical source categories of interest, the test
program should provide insight into sampling techniques for
general emission sources possessing the various atypical source
characteristics.  The selected sources are discussed briefly in
the following.

     1.   Roof Ventilators

     Roof ventilators were selected as a representative example
of two of the atypical source categories, emission sources
characterized by (1)  low velocity and  (21 extended dimensions.
The bases for this selection are the wide degree of application
of roof ventilators, the significant emission levels associated
with -them and the clearly defined nature of the characteristics
of the source (i.e., low velocity and extended dimension are
clearly the outstanding features of roof ventilators}..

     2.   Grain Dryers

     Grain dryers were selected to represent the category of
partially or totally unconfined flow.  The choice of a partially
                               -  24  -

-------
confined source, rather than a totally unconfined source, was
made in view of the overall philosophy of the program in regard
to model studies.  Modeling a partially confined emission
source, such as a grain dryer, constitutes a reasonable extension
of the sampling problem into the realm of this atypical source
category without attempting to complicate matters to a degree
which would very likely preclude any concrete results within
the scope of the present study.  The choice of grain dryers as
the specific source was based on their common usage as well as
the previously noted sampling difficulties cited by other
investigators of grain dryer emissions.

     3.   Wet Scrubbers

     The exhaust flow from wet scrubbers was chosen as the
representative example of emission sources with saturated gas
streams or gas streams with entrained liquid droplets.  This
choice was based on the popularity of scrubbers in pollution
control systems and the resulting large number of applications
in various industrial situations.  In addition, the exhaust of
a scrubber is generally a duct or stack, which lends itself
well to modeling and laboratory testing.
                              - 25 -

-------
            III.   ROOF VENTILATOR SAMPLING TECHNIQUES

     Continuous gravity roof ventilators, in the typical appli-
cation, are primarily characterized by low velocity and extended
dimensions.  These two characteristics make emission measurement
following EPA Method 5 sampling techniques inadequate.  Under
the present study, sampling methodology which has been applied
to roof ventilators was reviewed, and techniques and instrument-
ation for measurement of roof ventilator emissions were evalu-
ated following a test program as outlined in the Introduction.

A.   Review of Sampling Methodology

     Contacts with a number of commercial firms which have
performed emission measurements in roof ventilator applications
revealed that specific procedures for such situations have not
been established.  In general, sampling techniques and instrumen-
tation have been selected for a given application on the basis
of engineering judgment.  Sampling techniques have often employed
temporary collection hoods of some type, and generally the use
of high volume samplers has been prescribed for particulate
collection.  The most frequently used velocity instruments were
vane or propeller anemometers, followed in preference by hot
wire anemometers.  Reliability of emission measurements has
generally been considered poor due to the low velocity, fluctu-
ating, often circulatory flow characteristic of roof ventila-
tors as well as the intermittent nature and low concentration
of the particulate emissions from most roof ventilators.

     A study of roof ventilator emissions at electric furnace
operations led to the development of test methodology by Kreichelt
and Keller  (Ref. 29).  Particulate concentration measurements
were accomplished using standard high volume samplers suspended
inside the roof ventilators, and velocity measurements were
obtained using both hot wire anemometers and rotating vane
anemometers.  In this study, emission measurements were deter-
mined for a 60 m long roof ventilator both during periods of
changing and tapping of the steel furnaces and during periods
of so called "background" emission.  Several high volume samp-
lers spaced at 8.5 m intervals along the ventilator length were
operated simultaneously;  during charging and tapping operations,
the samples were fixed at the midpoint of the monitor width due
tp the short duration of these operations, while the  samplers
were traversed across the width during background measurements.
Velocity traverses were made independently of particulate
concentrations during charging and tapping operations, the
total number of traverse points being dictated by the time
duration of the operations.  Background period velocity measure-
ments consisted of 6 point traverses at each of the 7 sample
locations.
                              «  26 T

-------
     Figure 6 indicates the sampling technique discussed in
Ref. 29.  No attempt was made to sample isokinetically due to
the large velocity fluctuations observed.  Variations as large
as a factor of 6 were observed from point to point, yet overall
average velocities from various surveys during the background
periods agreed well.   (The average sampling velocity did agree
fairly well with the average effluent velocity.1  An overall
assessment of the test methodology in Ref. 29 concluded that it
provided a technically feasible, but costly, method for deter-
mination of emissions from roof ventilators;  the total time
and cost estimate given in Ref. 29 was 4 months or more and
$10,000-$20,000 to complete a single emission study.

     Souka, et al  (Ref. 30), described a sampling technique
used to evaluate emissions from a 1.22 m x 91.5 m roof venti-
lator at a graphitizing facility.  Three sampling systems  (Fig.
7), each including a high volume sampler and an electric anemom-
eter, were located at" the midpoints of equal sections of the
ventilator outlet.  Baffle plates, roughly 3 m in length, were
located about 3 m on each side of each sampling station to
minimize wind crossflow effects.

     The testing period reported in Ref. 30 was 160 hours of
almost continuous sampling.  Particulate collection filters
were changed at times dictated by observing the decreasing flow
rate through the samples.  Velocity readings were obtained
every 1.2 minutes and averaged over the appropriate sampling
periods.  Since isokinetic sampling could not be achieved due
to  the fluctuating effluent velocity, sampling rates were kept
well below isokinetic rates in order to achieve conservatively
high emission rate measurements.  Emission rate determined by
this method is indicated in Fig. 8 for one of the three sampling
stations.  The conclusion drawn from Ref. 30 is that the wide
emission rate fluctuation observed dictates long sampling
periods to obtain a true picture of actual emissions.  Total
cos.t for such an emission test were estimated to be on the
order of $15,000-$25,000  (Ref. 31).

     Federal regulations in regard to roof ventilator sampling
were formulated in conjunction with emission standards for the
primary aluminum industry in the form of EPA Method 14  (Ref.
32) . This method substantialy follows the sampling method
developed by Alcoa for roof ventilator sampling  (Ref. 33).
Particulate concentration measurements, following Method 14,
are to be accomplished by means of a sampling network as
outlined in Fig. 9. In this system, effluent gas is drawn
through 8 nozzles, mounted near the center of the roof ventila-
tor, to a manifold.  The velocity through each nozzle, as mea-
sured with an S-tube inserted through the calibration holes
shown in Fig. 9, is adjusted to the same value by means of
blast gates or valves before emission measurements are made.
     •'                         ^ 27 -

-------
  FIGURE 6.  CROSS SECTION OF ROOF VENTILATOR
              SHOWING  TEST EQUIPMENT  (REF.  29)
 Monitor top
  Electric
anemometer
-/ft
-  I!
                                   High
                                  • volume
                                   sampler
                               Sampling
                               platform
 FIGURE 7.   SAMPLER SYSTEM (REF.  30)


                 - 28  -

-------
 10
Is
I
                                            FTh
  12M
        12M
12M
12M     12M
   Time
                                  12M
                                         12M
                                               12M
FIGURE  8.   EMISSION RATE MEASUREMENTS, CENTRAL
            SAMPLING STATION (REF. 30)
 FIGURE 9.  EPA METHOD 14 SAMPLING SYSTEM

                     - 29 -

-------
Particulate sampling is then performed in the duct exhausting
from the manifold.  The length of the sampling manifold is to
be  85 m or 8% of the total ventilator length., whichever is
greater.

     Velocity is to be determined by means of propeller or vane
anemometers with electric output, appropriately protected from
a dusty and corrosive atmosphere.  One anemometer is required
for each 85 ra of ventilator length, located at the center of
equal length segments and permanently mounted at a point of
average velocity as determined by a widthwise velocity traverse.
Velocity readings are to be recorded at maximum time intervals
of  15 minutes during a test;  the sample rate during a test is
to  be adjusted  such that the velocity inlet to the sampling
nozzles equals  the average velocity determined for the:24 hours
preceding the test.

     All of these methods described in the literature emphasize
particulate concentration measurement through the collection
of  relatively large sample volumes  (compared to EPA Method 5),
employing standard high volume samplers, modified high volume
samplers, or specially constructed high volume nozzles and
manifolds.  Volumetric flow rates determined from single point
measurements or widthwise velocity traverses at a small number
of  locations along the roof ventilator length together with the
concentration measurements have been used in these methods to
determine total emission rates.  With the exception of the
baffle plates in Ref. 30, artificial collection apparatus is
not recommended in the literature, in contrast to the general
impression received through commercial firm contacts.

B.   Preliminary Field Tests

     The site selected for field testing was Hitchcock Indus-
tries in Bloomington, Minnesota, a secondary aluminum foundry.
This facility has a number of roof ventilators, both of the
powered and natural draft type.  The particular roof ventilators
chosen for initial field testing are pictured in Fig. 10.  Each
of  these ventilators is a 15.2 m  (50 foot) length of a configu-
ration known as the Swartwout  C42 inch) Heat Valve.  The venti-
lator height is approximately 2 m; the base width is roughly  2
m and the exhaust plane at the top is about 1.5 m wide.  The
ventilators shown in Fig. 10 are located above a room contain-
ing six gas-fired furnaces which are used to melt the aluminum
stock.  The roof ventilator exhaust configuration, which has  two
exhaust planes, is shown in Fig. 11.

     Due to the expected low concentrations to be encountered,
a high volume particulate sampler was chosen for the basic
concentration measurement equipment.  A Staplex Model TF1A high
volume sampler  was fitted with an extension probe and a 20.3  cm
x 25.4 cm (8 inch x 10 inch) filter holder designed and fabri-
                              ^ 30 r-

-------
FIGURE 10.
ROOF VENTILATORS  AT HITCHCOCK INDUSTRIES,
BLOOMINGTON, MINNESOTA,  VIEWED FROM THE
NORTHWEST.
                                                 ^
                '
 FIGURE 11.
 TYPICAL ROOF VENTILATOR EXHAUST AT HITCHCOCK
 INDUSTRIES SITE.

                    t -

-------
cated at FluiDyne facilities.  The sampling probe is used with
standard Gelman Type A glass fiber filters.  The entire assembly
is illustrated in Fig. 12.  This sampling probe has a flow rate
capacity as high as 2 m /min C70 ft /minl_.

     The velocity instrumentation selected for the initial
field tests consisted of a Thermo Systems, Inc. Model 1610 hot
wire anemometer and a Flowrite Model MRF vane anemometer.  The
hot wire anemometer was fitted with a protective collar designed
and built by FluiDyne  (Fig- 131 for a previous project in which
velocities in the range of 0.03 -4.57 m/sec CQ-1 - 15 ft/sec)
had been successfully measured.  The vane anemometer was a
direct reading instrument with a range of 0.3 - 15.2 m/sec
 (50 - 3000 ft/min).  Both instruments were mounted on extensions
to be hand held for the velocity measurements.  The hot wire
anemometer was also fitted with an iron-constantan thermocouple
for temperature measurement.

     During two days of testing at the Hitchcock Industries
site, particulate concentration and velocity measurements were
accomplished at several locations in one of the roof ventilators
shown in Fig. 10.  The sampling planes and locations are indicat-
ed in Fig. 14.  Particulate concentration measurements were
made at stations 2, 3 and 4, in both the east and west sample
planes, and velocity surveys included measurements at all 5
sampling stations in both sample planes.

     During the field tests, the operations within the foundry
were also observed.  The furnaces were loaded with aluminum of
various degrees of quality, depending on the specifications of
the various products being manufactured, to be melted down for
pouring into molds.  Scrap aluminum was sometimes melted, which
often included foreign material such as steel wool.  The furnace
doors were often opened for brief periods during the melting
process.  The molten aluminum flowed from the furnaces into
large vats.  When the vats were full, the furnace doors were
opened and the vats removed.  In general, all six furnaces were
not in use at one time;  several furnaces often "idled" with
the furnace doors open.  The turnaround time for each furnace
appeared to be roughly the same, thus keeping the overall level
of activity generally fairly constant.

     Visual observations of occasional plumes rising from the
furnaces indicated that the flow was sometimes very turbulent,
becoming well mixed by the time the roof ventilator was reached,
and sometimes appeared much more uniform, rising steadily in a
single column.  In general, it was difficult to characterize
the activities within the foundry by specific, scheduled opera-
tions, such as the charging and tapping operations described in
Ref.  29.  Overall, it was felt that the Hitchcock Industries
site provided a good example of the type of problems to be
encountered in studying roof ventilator emissions.
  ,                            -  32  -

-------
FIGURE 12.  HIGH  VOLUME SAMPLER AND PROBE ASSEMBLY
                       - 33 -

-------
      Probe Extension  1.53 m  (5 ft.)
                                                        ::©
                                         Thermocouple
                                         mounted  here
               TOP VIEW
Temperature Compensation
   Electronics Inside
        Housing
                             I. ____
              SIDE VIEW
            Flow
          Direction
Hot Wire
Anemometer
Sensor Head
7
   FIGURE 13.  HOT WIRE ANEMOMETER AND PROTECTIVE COLLAR ASSEMBLY

                           -  34 -

-------
 Sample Planes
                  0.534 m
End View Gut Away
Station-
   5
                                Station
                                   4   "
                                Station
                                   3   -
                                Station.
                                   2
                                Station.
                                   1
                                           2.44 m
                                                         3.05 m
                                                         3.05 m
                                                         3.05 m
                                                         3.05 m
                                                         1.52 m
                   FIGURE 14

SAMPLING LOCATIONS FOR PRELIMINARY FIELD TESTS
     HITCHCOCK INDUSTRIES ROOF VENTILATOR
                                                            N
                   - 35  -

-------
     The sample planes were chosen for maximum ease of alignment
of the particulate sampling probe, as illustrated in Figs. 15
and 16.  Samples were extracted for 90 minutes at a constant
flow rate indicated by a rotameter mounted on the high volume
sampler.  The rotameter was calibrated by measuring the pressure
drop across a series of orifice metering plates attached to the
inlet of the probe nozzle in order to determine the actual flow
rate through the sampling probe.  Isokinetic sampling was
attempted, but the fluctuating nature of the velocity made this
impossible.  Later review of the data indicated that sampling
rates were typically in the range of 15Q% to 23Q% of the iso-
kinetic rate.

     Particulate concentration measurements obtained during the
preliminary field tests are shown in Fig. 17.  Since the measure-
ments were extracted at various times during the two days of
testing, the results shown in Fig. 17 are representative of
several different activities in the furnace room below the
ventilator.  Thus, these results should be interpreted only as
indications of the magnitude of concentration levels in the roof
ventilator emission stream, and not as indications of possible
concentration gradients which may exist at any given time.

     Several velocity traverses were also made during the pre-
liminary field tests.  The measurement planes were the same as
for the particulate concentration measurements.  The hot wire
anemometer proved to be unsuitable for the awkward conditions
encountered; despite the special protective collar, the fragile
wire tip was soon damaged, rendering the instrument inoperable.
Thus, the bulk of the velocity measurements during the prelimin-
ary field tests were obtained with the vane anemometer.

     The results of two velocity surveys are shown in Figs. 18
and 19.  Although the velocities shown in Figs. 18 and 19
appear to be quite similar in the two surveys, the actual
velocity fluctuated considerably about these plotted values
determined from the velocity instruments, and the magnitude of
the average velocity differed by a factor of 1.7.  Different
wind conditions were observed during these two velocity surveys,
the component of wind normal to the length of the ventilator
being 14 times larger in the case of the larger measured velo-
cities.  Other factors which could affect the exhaust velocity
are the ambient air temperature and the temperature distribution
inside the furnace room, which depends on the particular activ-
ities inside the foundry.  Therefore, the velocity data shown in
Figs. 17 and 18 must also be considered only as an indication of
typical magnitudes and not definitive characteristics of the
roof ventilator exhaust.

   :                           - 36 -

-------
FIGURE 15.
HIGH VOLUME  SAMPLER ASSEMBLY MOUNTED  IN  THE ROOF
VENTILATOR AT  HITCHCOCK INDUSTRIES  (END  VIEW
LOOKING NORTH)
FIGURE 16.
HIGH VOLUME  SAMPLER ASSEMBLY MOUNTED  IN  THE ROOF
VENTILATOR AT  HITCHCOCK INDUSTRIES  (TOP  VIEW
LOOKING DOWN INTO EXHAUST REGION)
                - 37 -

-------
                      East Exhaust Plane
Partieulate
Concentration
(mg/m )

0.5-
0.4-
0.3-
0.2'
0.1"
f\
D
A
<0
i
i
1 1 1 	 1 	 1 	 1 	 1 	 h
                  2    4     6    8    10   12   14
                     Roof Ventilator Length (m)
16
                       West Exhaust Plane


Particulate
Concentration
3
(mg/m )



0.5 '
0.4'

0.3-

0.2-
O.I1
n
i
»

»
D ^
o
•
	 1 	 1 	 1 	 4 . i n4 	 1 	 1 1
                  2    4    6    8    10   12   14
                     Roof Ventilator Length (m)
16
                    FIGURE 17
PRELIMINARY FIELD TEST CONCENTRATION MEASUREMENTS
       HITCHCOCK INDUSTRIES ROOF VENTILATOR
                   - 38  -

-------
Distance from East Edge (m) Distance from West Edge (m)
0 0.1 0.2 0.3 0.4 0.5 0.5 0.4 0.3 0.2 0.1 0
3 -j


2 -
1 -
0 -
3 -

2 .
1-


3 -

2 _
"o
0)
tn 1 _
^
£ oJ
o
o
H 3 -
>
2 -
1 -

o -
3 -
1
2 -
1 «•
o J
i I i i i

O


o
n
^
A


A
A

0




o


e. 	
*

D

D





O
0
n
'•••'


Station
5



Station
4

'


Station
3






Station
2




Station
1


i i i i i

O


o
e-
-

A

A
A



O


o
o





D

D




O
o
n

-3


-2
-1
-0
-3

-2
-i


-3

-2

.1

-o

-3

-2
.1

•-o
-3

•2
-1
0
FIGURE 18
PRELIMINARY FIELD TEST VELOCITY SURVEY 1,
HITCHCOCK INDUSTRIES ROOF VENTILATOR
	 aa 	

-------
Distance From East Edge (m) Distance From West Edge (m)
0 0.1 0.2 0.3 0.4 0.5 0.5 0.4 0.3 0.2 0.1 0
0
J —
f\ _

1 -

o
V/ **l
3 -

2 -

1 -

0
'o' o
0) J -i
co
"a
. "-" 2 -
4-1
•H
U -I -
OX
rH
(U
> 0 -
3 -
2 -
..
1 -

0 -J
3-

2 -
1 "

0 -
iii*i i i § » i
O

O

«••>>




A

A
A






O
0 0
V



D
D




O

O
9_.

Station
5
^/



Station
4






Station
3
ij



Station
2




Station
1





0
O
(~*i




A

A
— A 	





O

o
s\
>r



D
D

e,,
_____________________


O

O
e'
	
.3
•J
-2

"1

n
-3

-2
'.
-l

.0
-3


-2

•1
-0
-3
~2
£*
-1

U
•3

"2
-1

J-Q
FIGURE 19
PRELIMINARY FIELD TEST VELOCITY SURVEY 2
HITCHCOCK INDUSTRIES ROOF VENTILATOR
- 40 -

-------
     The exhaust stream temperature was also found to vary
considerably, ranging from 26°C to 49°C, the average temperature
being approximately 34°C.  The temperature was seen to vary with
time and with position along the roof ventilator/ although no
attempt at correlation was made.

C.   Model Studies

     The model studies phase of the test program was developed
on the basis of these preliminary observations and measurements.
A laboratory model representative of a typical roof ventilator
was designed and fabricated, and this model was used to evaluate
techniques and instrumentation for use in roof ventilator emis-
sion measurement.

     1.   Model Design and Fabrication

     Manufacturer's literature  (Ref. 34), and dimensional measure-
ments taken at the Hitchcock site were used to fabricate a model
of a section of roof ventilator  (Fig. 20).  This roof ventilator
model was attached to a test section and connected to,an exist-
ing duct network, used in a previous FluiDyne study at the
firm's Rosemount Energy Conversion Laboratory.  The model was
oriented on the roof in such a way as to maximize the effect of
the prevailing wind.  The model test facility is shown sche-
matically in Fig. 21 and the roof ventilator model atop the
Energy Conversion Laboratory is illustrated in Fig. 22.

     In order to provide the capability of introducing particu-
late matter into the model exhaust, an aerosol generation system
was added to the existing ductwork  (Fig. 23).  The concept of
the aerosol generation system was to atomize a weak saline
solution by means of a spray nozzle.  The heated air then induced
evaporation of the water, leaving small particles of salt sus-
pended in the flow stream.

     Unfortunately, during the course of the test program with
the roof ventilator model, FluiDyne's Energy Conversion Labor-
atory facility was destroyed by a fire, as was the entire roof
ventilator test facility.  This necessitated the construction of
a second model at FluiDyne's Medicine Lake Aerodynamic Testing
Laboratory.  This model, illustrated in Figs. 24 and 25, fol-
lowed essentially the same design as the original roof venti-
lator model.  The aerosol generation system  for the second roof
ventilator model, illustrated in Fig. 26, also followed the same
concept as in the original model facility.

     Four sample ports were located at the base of the second
roof model  (Fig. 27) for evaluation of a second sampling plane
in addition to the ventilator exhaust.  Testing on both roof
ventilator models included flow visualization studies, velocity
measurements, and particulate concentration measurements.
                             - 41 <-

-------
FIGURE 20.  ROOF  VENTILATOR MODEL
             - 42 -

-------
              Roof
   Ventilator Model
         Roof Model
         Test Section
                                                     est Stand
                                                                                               32"  I.D.  Duct
   Up to 9000 CFM
  with Heat
  wo. uu nca i_  f—
1  InjtiuLiuft-  W
  Atomization
   Aerosol
  Generator
s	j:	^, Pressurized Air
I           for Atomizer
                          3
                            u
                        Liquid Droplet
                         Test Ports
                              j-<:--Test Stand
   Early
Electrostatic
 Precipitator
   Model
                                                                            Sample Port
                                                                                          '/ s ,'r,- r 7
                                 FIGURE 21.  ROOF VENTILATOR MODEL TEST FACILITY —
                                             ROSEMOUNT ENERGY CONVERSION LABORATORY

-------
FIGURE 22.  ROOF  VENTILATOR MODEL TEST SECTION
                 - 44 -

-------
Ul
I
    24" I.D.
Viewing Port
                                     Pressure
                                    Regulator
                               Pressure
                                Gauge
                                                         100 psig
                                                                      /•
                                                                                       Adjustable
                                                                                      Heated  Air
                                                                                      Injection
                                                             Adjustment Port
                    Liquid
                    Flow
                    From
                 Reservoir
                               -I'
                                         O| Rotameter
                                                                                       Flow
   FIGURE 23.  ATOMIZATION AEROSOL GENERATOR AND HOT AIR INJECTION SYSTEM —
                  ROSEMOUNT ENERGY CONVERSION LABORATORY

-------
FIGURE 24.   ROOF VENTILATOR MODEL AT MEDICINE LAKE
             AERODYNAMIC TEST FACILITY   VIEWED FROM
             NORTH
FIGURE 25.
ROOF VENTILATOR  MODEL AND DUCTING AT
MEDICINE LAKE AERODYNAMIC TEST FACILITY
- VIEWED FROM NORTHEAST

          - 46 -

-------
                               AIM
                                           *i.o. fife (*ta. JH i-o.)
                                          II 6A. (.IIO) MM. WALL. YHICKMCM
                                          CCMTtMO IU la'BUCT
   -eMROMALOX CCH-S
    * KW AIM CIKCULATiOM HCATM
!• VCMTiLATiOM
(cxitTiuo)
  • GALLON
LIQdO tOLUTIOM
COMTAIMCR
   FIGURE 26.   ATOMIZATION  AEROSOL GENERATOR  SYSTEM WITH HOT AIR INJECTION
                        -  MEDICINE  LAKE  AERODYNAMIC TEST FACILITY

-------
FIGURE 27.
ROOF VENTILATOR MODEL SAMPLE PORTS
VIEWED FROM NORTHEAST
                      - 48 -

-------
     2.   Flow Velocity Studies

     The first objective of the model testing program was to
evaluate the capability to determine total volumetric flow
through the roof model.  Early observations indicated that the
exhaust from the roof ventilator model often exhibited very
complex flow patterns which could not be readily determined with
conventional velocity instrumentation.  Therefore, flow visuali-
zation studies were conducted using lightweight flow direction
tufts as well as smoke injection studies to examine the fluid
motion through the roof ventilator  (Figs. 28 and 29) .

     A circulatory flow pattern, illustrated schematically in
Fig. 30, was observed at the ventilator exhaust.  The severity
of the exit velocity profile as well as the depth in the roof
ventilator to which circulatory flow was evident were seen to
depend strongly on the atmospheric wind conditions.  This fact,
together with the inability of conventional velocity instru-
mentation to give reliable values for total volumetric flow
rate, led to the conclusion that velocity measurements should be
made in a plane other than the exit from the ventilator.

     A plane near the base of the roof ventilator was felt to
provide a better sampling location.  The sample ports shown in
Fig. 27 were used to study the flow behavior near the base of
the roof model.  A certain degree of circulatory flow was still
evident, but appeared to be of a steady state nature, unaffected
by wind gusts and other external disturbances.  It was deter-
mined that this circulatory flow was caused by the rapid expan-
sion of the test section which supplied the flow to the roof
model  (see Figs. 24 and 25) and that insertion of a flow straight-
ening baffle upstream of the roof model eliminated this circula-
tory flow near the base of the model.  A typical flow profile
observed with this flow straightening baffle in place is illus-
trated in Fig. 31.  A useful feature of this baffle was the
capability to produce circulatory flow or uniform flow at the
base of the model for the purpose of evaluating the usefulness
of various velocity instruments.

     Having established the basic characteristics of the flow
through the roof ventilator model, the remaining task was to
evaluate the suitability of various commercially available in-
struments for measurement of the volumetric flow rate.  A large
number of instruments were considered for this application, and
their suitability was evaluated through extensive testing both
in the roof ventilator model and at the Hitchcock Industries
site.  A detailed description of this evaluation program is
given in the Appendix.  The description of this program was
arranged in the form of a self-contained dissertation entitled
"Velocity Instrumentation for Low Velocity, Partially Confined
Source Particulate Sampling" due to the general importance and
usefulness of this material with regard to the problems of
atypical emission sources.  To summarize the results of the
                              - 49 -

-------
FIGURE  28.
FLOW  DIRECTION INDICATING TUFTS IN ROOF
        VENTILATOR MODEL EXHAUST
FIGURE 29.
SMOKE GENERATOR FLOW INDICATOR IN ROOF
       VENTILATOR  MODEL EXHAUST
                         - 50 -

-------
                                   Recirculating Flow
                               Flow Direction
                               Indicating Tufts
                  Exhaust  Flow
FIGURE 30.
SCHEMATIC OF OBSERVED FLOW
 (END VIEW OF ROOF MODEL)
                    - 51 -

-------
               Exhaust
              Sample Plane
Base
Sample
Plane
                                Velocities in

                                Parentheses

                                in meters/sec
   f        A       A        A       i
— ©	©	® — -  ®  - ;— ®	
 (0.71)   (0.62)   (0.44)   (0.32)   (0.23)    (0.22)
     FIGURE 31.  TYPICAL FLOW VELOCITIES AT VARIOUS  POINTS
                      IN ROOF VENTILATOR MODEL
                           - 52 -

-------
investigation into velocity instrumentation/ it was found that a
heated thermopile anemometer, the Hastings-Raydist PCI-30, was
capable of accurate determination of the total volumetric flow
rate when used at the base sampling plane under two conditions:
1)  when the uniform flow baffle was used to provide a non-
circulatory flow pattern, and 2)  when a secondary observation
tool, a lightweight flow direction indicator, was used in conjunc-
tion with the instrument in a circulatory flow field with the
baffle absent.  These results are indicated in Pig. 32.  None of
the velocity instruments evaluated were capable of providing
accurate flow rate data when used at the exhaust plane of the
roof ventilator model.

     3.   Particulate Concentration Studies

     The reliability of particulate concentration measurements
using the high volume sampler  (Fig. 12) was also evaluated
through roof ventilator model studies.  The procedure utilized
for this evaluation was to induce particulate into the model
ductwork by means of the aerosol generation system and attempt
to measure the particulate concentration at the roof ventilator
model exhaust or at the base of the roof model through the
sample ports.  The actual concentration was monitored in the
ductwork upstream of the roof ventilator model using EPA Method
5 sampling techniques, collecting particulate on 47 mm Gelman
Type A glass fiber filters.

     The results of these measurements are shown in Pig. 33.
The measurements at the exhaust plane consist of an average of
several measurements at different points, while those at the
base plane represent measurement at a single point near the
center of the roof ventilator model.  A great variety of atmos-
pheric conditions was encountered in the course of the particu-
late concentration testing.  Isokinetic sampling with the high
volume sample probe was rarely achieved.  Sampling rates were
typically in the range of 150% to 250% of the isokinetic rate.
No obvious correlation was seen between sampling rate or atmos-
pheric conditions and degree of accuracy of the results.

     The measured particulate concentration is generally seen
from Fig. 33 to agree rather poorly with the actual values.
However, an overall reliability of +_ 25% is indicated over a
large range of particulate concentrations for samples taken at
the base of the roof ventilator model.   In all but one test,
the measurements at the base plane indicate a higher than actual
concentration.  At very low concentrations, on the order of
those observed in the preliminary field tests, accuracy is seen
to be as poor as +77% in one case.  The disagreement may be
exaggerated, however, due to the small weight changes of the
standard filters at these low concentration levels.
                             -  53 -

-------
Actual Volumetric Flow
Rate (m^/min)
  350  •
   300 -
   250 •*
   200
   150
   100
    50

Without Uniform
Flow Baffle
With Uniform
Flow Baffle
With/

©
m

Without/

o
a

                                                              Flow Direction
                                                              Indicator
                                                             100% Accuracy
                                                                         .+50%
               50
100
150
200
250
300
350
                                              Indicated Volumetric Flow
                                              Rate (
         FIGURE 32.   VOLUMETRIC  FLOW RATE AT BASE OF ROOF VENTILATOR
                     MODEL  DETERMINED WITH  HASTINGS-RAYDIST PCI-30
                     HOT THERMOPILE  ANEMOMETER
                                    - 54  -

-------
Actual Concentration
      (mg/m3)
 10-r
                                                 r-. Sample Plane at Exhaust
                                                   (Average of Several
                                                         Measurements)
                                                 O Sample Plane at Base
                                                   (Single point measurements)
                                                 ^	1—i
                                                                10
                                                Measured Concentration
                                                       (mg/m3)
                                FIGURE 33

                PARTICIPATE CONCENTRATION MEASUREMENT IN
                          ROOF VENTILATOR MODEL

-------
     In general, the high volume sampler method used in a sam-
pling plane at the base of the roof ventilator model gave conser-
vatively  higher particulate concentration results regardless of
sampling rate.  Further evaluation of the sampling techniques
was precluded by time and budget limitations, due in part to
difficulties in operation of the aerosol generation system.
Additional study at a future time would be desirable in order to
more precisely evaluate the reliability of the method.

D.   Final Field Tests

     In order to evaluate the usefulness of the sampling tech-
nique described above under actual field conditions and to
examine the effects of the extended length dimension on the roof
ventilator problem, which was not possible in the model studies,
a final series of field tests was carried out at the Hitchcock
Industries site.  Five .sample ports were installed near the base
of the roof ventilator pictured in Fig. 10 during a scheduled
shutdown of plant activities.  The ports were located at the
midpoints of five equal area sections of the ventilator, i.e.,
at locations corresponding to Stations 1-5 shown in Fig. 14.
These sample ports provided access to the more favorable samp-
ling plane identified by the model studies, as shown in Fig. 34.

     Measurements of particulate concentration, velocity, and
temperature were made using the high volume sample probe, the
heated thermopile anemometer and a calibrated iron-constantan„-
thermocouple, all inserted into the roof ventilator effluent
stream through the sample ports  (Fig. 34).

     Particulate concentration measurements were conducted
throughout each day of testing.  A total of five 1.5 hour sam-
ples (except one test which was concluded after only 45 minutes
at the "first sample port) were taken at each of the five sample
ports in the hope that the data would provide insight into the
timewise and lengthwise variations of particulate concentration
and provide a basis for evaluating the sampling technique with
respect to applicability to other extended dimension emission
sources.
           >'
     Velocity measurements were made at various times during
eaqh day.  During all five days of testing, the flow was ob-
served to be directed relatively uniformly upward, as indicated
by a lightweight flow direction tuft, in contrast to the sup-
posed occasional circulatory flow based on visual observations
of rising plumes from inside the foundry.  Velocity surveys
including measurements at all five sample ports were possible
before and -after the concentration measurements, but only  four
sample ports were used for velocity surveys during the concen-
tration measurements since ,the sample probe and velocity probe
could not be inserted through the sample ports simultaneously.
                             -.56 -

-------
FIGURE 34.  PARTICULATE CONCENTRATION AND VELOCITY
            MEASUREMENTS AT BASE OF HITCHCOCK
            INDUSTRIES ROOF VENTILATOR
                     -  57  -

-------
     The measurements made during the final field tests are
summarized in Figs. 35 through 39.  The average velocity at each
sample port, determined from a six-point widthwise velocity
survey, and the particulate concentration are plotted vs. the
time of day at which each measurement was made.  All five days
of field testing are compressed onto a single scale, represent-
ing the normal work day, and each data point is identified as to
the day it was taken.

     Several observations can be made from the roof ventilator
measurements shown in Figs. 35 through 39.  The average velocity
at a given sampling station often varied considerably with time
on a given day and also varied from day to day at a given time.
General trends can be seen, however. The curves shown on these
Figures represent intuitive estimates of the "average" data and
are not based on curve-fitting or regression formulas.  It should
be emphasized that the "point of average velocity" was not the
same in all cases, as would be expected for flow in a pipe or
duct, but was seen to vary over the entire width of the roof
ventilator as the various velocity surveys were conducted.

     The particulate concentration also varied during each day
of testing.  In general, concentration levels were seen to be
higher near the midpoint of the roof ventilator than at the
ends.  Since the concentration measurements at each sample
location were conducted on different days, it cannot be definitely
determined that this variation is due to the position along the
ventilator.  However, observations of the foundry activities
from day to day revealed no obvious procedural differences which
would explain overall concentration variations of the magnitude
observed in these measurements.  Therefore, the observed varia-
tions of concentration with position in the roof ventilator are
assumed to be indicative of the actual situation.

     Investigations into a possible relationship between the
effluent stream temperature and the velocity and/or concentra-
tion indicated no useful correlation.  Since the flow through
the roof ventilator is governed by a complex interaction of the
inside temperature, outside temperature, wind conditions, and
many other factors, this observation is not surprising.  Often,
higher temperatures accompanied higher velocities, but this was
not always the case.  The temperature occasionally appeared to
vary cyclically over a period of several hours, but more often
appeared quite random both with time and position, ranging
between 35°C and 78°C over the five day test period.  The aver-
age temperature was on the order of 55°C.

E.   Evaluation of Sampling Technique

     The information obtained during the roof ventilator  test
program led to several conclusions concerning basic measurement
                             V 58 -

-------
5
O Day i
1.41
-
1.2-

1.0-
Average n 8 -
Velocity
J
(m/sec)
0.6"

0.4"
0.2-
0_
D Day 2
O Day 3
A Day 4 <>
0 Day 5

0
ADO 0
^ """"\D
s^^W\ %> A ^\^
O A
a A

^ * i i i i i i i i i i •
7 8 910 11 12 12 34 56
AM PM
0.6-
0.5-
Particulate _ .
0 4 ""*
Concentration "
(mg/m3)
0.3^
0.2^
O.-,-! 	 	 	 	 	






9... ....

r\ e^.
— i i Q i
1 I 1 1 * 1 I 1 I J 1
78 9 10 11 12 1 2 345 6
AM . _ PM
Time of Day
FIGURE 35
AVERAGE VELOCITY AND PARTICULATE CONCENTRATION
MEASUREMENTS - SAMPLE PORT 1
HITCHCOCK INDUSTRIES ROOF VENTILATOR
- 59 -

-------
1.4 "
1.2 -
i.o -
Average
Velocity 0.8 -
(m/sec)
0.6 '
0.4 -
0.2 "
0
0.6
0.5 ~
Particulate
Concentration 0.4
(mg/m3)
0.3 "
0.2 "
0.1 -
0
O Day 1
D Day 2
O Day 3
O O A Day 4
O Day 5
0 o 0
oo __^
A^T ^^
A ]P
A t-i
A A

	 i i i "i i i i < i >
' 8 9 10 11 12 1 2 3 45 6
AM PM



B.
r-i 1 	 H 	 1
n
n u
LJ
i i i l 1 i it
' 8 9 10 11 12 1 2 3 4 56
AM PM
Time of Day
FIGURE 36
AVERAGE VELOCITY AND PARTICULATE CONCENTRATION
MEASUREMENTS - SAMPLE PORT 2
HITCHCOCK INDUSTRIES ROOF VENTILATOR
- 60 -

-------
 Average
 Velocity
  (m/sec)
1.4-

1.2-

i.o-

0.8

0.6

0.4-

0.2-

 0
                0.6 -

                0.5-
 Particulate
Concentration   0.4
   (mg/m3)

                0.3
                 0.2

                 0.1-

                  o
       O
O
O Day  1
O Day  2
O Day  3
A Day  4
O Day  5
                         8    9   10    11    12   1
                                   AM
                       3    4
                        PM



/\
V




/s
V

                                  10   11   12
                                    AM
                                          Time of Day
                       3   4
                        PM
                                FIGURE  37

             AVERAGE VELOCITY AND PARTICULATE CONCENTRATION

                     MEASUREMENTS -  SAMPLE PORT 3

                 HITCHCOCK  INDUSTRIES  ROOF VENTILATOR



                               - 61  -

-------
1.4-
1.2-
Average
Velocity 1.0~
(m/sec)
0.8"
0.6-
0.4-
0.2-
0 -
0.6-
Particulate
Concentration
(rag/m3) 0.4"
0.3-
0.2-
o.i-
0 -
o Day i
O Day 2
O Day 3
A Day 4
O Day 5
0 0 ° D
\o C^x^ — -^^
° O o &
a
7 .;8 9 10 11 12 1 23 4 56


A '
A
*-* jA A A,

i > i i i i i i i
7 8 9 10 11 12 1 2 3 45 6
AM PM
Time of Day
FIGURE 38
AVERAGE VELOCITY AND PARTICULATE CONCENTRATION
MEASUREMENTS - SAMPE PORT 4
HITCHCOCK INDUSTRIES ROOF VENTILATOR
- 62 -

-------
O Day 1

1.2~
Average
Velocity 1.0-
(m/sec)
0.8-
0.6-
0.4~

0.2-
0 -
D Day 2
O O Day 3
A Day 4
O Day 5

o n
A\n A____ __ ___° o
A ^ 	 Ch^~^A
DO D °
o
1 8 9 10 11 12 1 2 345 6
AM PM
0.6~
0.5-
Particulate
Concentration
(mg/m3) 0.4
0.3-
0.2-
0 -





er f^ 	 1 f^\
U \^
6-

1 89 10 11 12 12 3 4 56
AM PM
TIME OF DAY
FIGURE 39
AVERAGE VELOCITY AND PARTICULATE CONCENTRATION
MEASUREMENTS - SAMPLE PORT 5
HITCHCOCK INDUSTRIES ROOF VENTILATOR
- 63 -

-------
techniques for roof ventilator emissions.  The complex flow
patterns observed at the roof ventilator exhaust dictate the
choice of a sampling plane near the base of the roof ventilator,
where the velocity is more uniform and less susceptible to
atmospheric winds.  The volumetric flow rate studies showed that
transverse velocity surveys with the heated thermopile anemometer
provided very good indications (within +_ 10%) of the total
volumetric flow rate through the roof ventilator model.  The
particulate concentration studies in the roof ventilator model
indicated that measurements with the high volume sample probe,
while lacking a high degree of accuracy, provided reasonable,
conservatively high estimates  (typically on the order of +25% as
seen in Fig. 33) of actual concentration regardless of sampling
rate.

     These observations defined a sampling technique and a
degree of reliability for emission measurements in the .model
section of a roof ventilator.  The questions which remained
unanswered concern the number of sampling locations required for
a given full-size roof ventilator.  The choice of five sampling
stations for the 15.24 m roof ventilator studied in the final
field tests was based on a desire to obtain the maximum amount
of information possible within the scope of the present project.
For maximum reliability, particulate samples and velocity measure-
ments should be obtained simultaneously and continuously at all
sampling stations throughout the period of emission measurement.

     A test program of this magnitude on a roof ventilator in
the primary aluminum industry, for example, where ventilators
may often be 200 m or more in length, would encounter serious
problems with regard to economic feasibility.  Thus, the number
of sample stations required in a given application is of vital
importance.  The single application of a secondary aluminum
foundry furnace room roof ventilator, studied under the present
program, cannot be expected to provide the solution to this
problem, since the type of flow and range of particulate con-
centration is heavily dependent on the particular activity which
generates the emissions.  In other words, the number of sample
stations required for any specific application will undoubtedly
depend on characteristics of that application and must be de-
termined for each individual case.

     The long sampling time required in order to obtain a
weighable collection of particulate tends to preclude the use of
widthwise particulate concentration traverses at each selected
sampling station.  Single point concentration measurements were,
therefore, chosen on the basis of practicality.  Widthwise
velocity surveys, however, may be made with considerably less
difficulty.  In order to demonstrate the importance of the
number of points chosen for a widthwise velocity survey, an
example based on field test measurements was considered.
                            -  64 -

-------
     A typical velocity distribution, measured at approximately
3:30 p.m. on the second day of final  field testing, was analyzed
(see Fig. 40) .  The average velocity  which would have been
determined from 1, 2, 4, and 5 point  widthwise surveys of the
velocity distribution shown in Fig. 40 was computed at each
sample station and compared with the  average velocity from the
6 point surveys actually made during  the  field tests.  The
results of this comparison are shown  in Table 10.  For this
particular velocity distribution,  it  can  be seen that at least 4
or 5 point surveys are required to determine the average velocity
at each station to within + 10%.   It  must be emphasized that,
while this velocity distribution is typical of those observed
throughout the field test period,  not only the average velocity
but also the points of average velocity continuously varied with
time as the basic velocity distribution varied with atmospheric
conditions and activities within the  foundry.

     As an illustration of the importance of the number of
sampling stations, a second example based on the final field
test measurements may be of value.  Consider a six-hour period,
from 10:00 a.m. to 4:00 p.m., of emission from the 15.24 m roof
ventilator.  Based on the field test  data, typical particulate
concentration and average velocity profiles can be constructed
as shown in Figs. 41 and 42 for three time instants during this
six hour interval.  It must be emphasized that these profiles
were constructed from data taken over a five day period.  The
profiles shown in Figs. 41 and 42  are intended to represent a
typical example only and should not be construed as actual
emission measurements during a single six hour period.

     The total emission during the six hour period was calculated
following a number of simple methods.  The basic emission cal-
culation procedure followed the formula:
                M
                      (C-VA  )t
                        3331
Where:    E =  total  emission  (gm)             3
          C =  particulate  concentration  (gm/m )
          N =  number of  time  increments
          M =  number of  sample  stations
          V =  average velocity  (m/sec),
          A =  cross  sectional area  (m  )
          t =  time interval  (sec)

The first calculation was  made  assuming  that  the profiles  shown
in Figs. 41 a,b,c and 42 a,b,c  each  represented the average
conditions for a two-hour  period  (4 la  and  42a represent the
period from 10:00 a.m. to  12:00 p.m.,  etc.).  Several emission
calculations using various combinations  of the data from Figs.
•41 and 42 were then  compared  with the  first calculation.
                            .U 65  r-

-------
V(m/sec)
V(m/sec)
V(m/sec)
V(m/sec)
V(m/sec)
1.5 "

i.o ••

0.5 ••

 0   L
1.5

1.0

0.5

 0

1.5

i.o ••

0.5

 0
1.5 "

i.o ••

0.5 •'

 0   L
          1.5 +
                                                               Station 5
                                                               Ave. Velocity
                                                               =.56 m/sec
                                                     Station 4
                                                     Ave. Velocity
                                                     =.81 m/sec
                                                     Station 3
                                                     Ave. Velocity
                                                     =.44 m/sec
                                                     Station 2
                                                     Ave. Velocity
                                                     = .56 m/sec
                                                               Station 1
                                                               Ave, Velocity
                                                               =  .64 m/sec
                       0.4      0.8       1.2      1.6       2.0

                       Distance from Edge  of Roof Ventilator  (m)

                               FIGURE 40

                 TYPICAL VELOCITY DISTRIBUTION AT BASE

                OF HITCHCOCK INDUSTRIES ROOF VENTILATOR

                               - 66  -

-------
                         TABLE 10



AVERAGE VELOCITY DETERMINED FROM WIDTHWISE VELOCITY SURVEYS,



                       BASED ON FIG. 40
Sample
Station
1
2
3
4
5
No. of Points in Survey
1
Avg. Vel,
(m/sec)
.82
.97
.42
.93-
.43
% Diff.
+28.1
+73.2
-4.5
+14.8
-23.2
2
Avg. Vel,
(m/sec)
.61
.39
.37
.85
.45
% Diff.
-4.7
-29,5
-17.0
+4.9
-20.5
4
Avg, Vel.
(m/sec)
.63
.55
,48
.79
.67
% Diff.
-2,3
-2,2
+8.0
-2.2
+19.2
5
Avg . Vel .
(m/sec)
.61
,53
.45
,80
.61
% Diff.
-5.0
-5.4
+2.7
-1.7
+8.2
6
Avg. Vel.
(m/sec)
.64
.56
.44
.81
.56

-------
 Particulate
Concentration
  (mg/m3)
 Particulate
Cone en t ra t ion
   (mg/m3)
 Particulate
Concentration
  (mg/m3)
                  0
0.5 t (a) 11 AM

0.4

0.3

0.2

0.1
                                             -^—e-
        4-
-j—I—I—I—1—h
 4     68     10   12    14
    0    2   4     68    10  12   14
       Roof Ventilator  Length (m)


0.5 f (b) 1 PM

0.4
                 0.3

                 0.2 +
                 0.1'•
                          I	A,
                                   •e-
                       	j	,	,	,	|	j	h.
                     0    24    6    8   10   12   14
                        Roof Ventilator Length (m)

0.5 -
0.4 '
0.3 •
0.2"
0.1 •
0

. (C)3 PM

•
'

	 1 —



^S
V








A
£_i 1 /~\
\_/
— i 	 1 	 1 	 1 	 1 	 1 —
                     02     4   6    8   10   12   14
                        Roof Ventilator Length (m)
                           FIGURE 41

              PARTICULATE CONCENTRATION PROFILES
         FOR ROOF VENTILATOR EMISSION EXAMPLE PROBLEM
                           - 68  -

-------
            1.0-- (a)  11 AM
            0.8
Average
Velocity    0.6
 (m/sec)

            0.4'
            0.2
             0
        -I
            0.8
 Average
  Velocity  A. ,--
    (I   s    U. O
  m/sec)

            0.4--
            0.2'
                    2    46    8    10  12   14
                   Roof Ventilator Length (m)
                   (b)  1 PM
                    2    4    6   8   10   12
                   Roof Ventilator Length (m)
 Average
 Velocity
   (m/sec)
1.0--  (c)  3 PM

0.8-'


0.6


0.4

0.2+
                                    14
                  	1	1	1	1	,	

                02    4    6   8    10  12  14
                   Roof Ventilator Length  (m)

                      FIGURE 42
              AVERAGE VELOCITY PROFILES
    FOR ROOF VENTILATOR EMISSION EXAMPLE PROBLEM
                     -  69 -

-------
     The results of these example calculations are shown in
Table 11. It can be seen that consideration of the actual varia-
tion of concentration and velocity with time and location is not
important, since averaging all concentration and velocity data
from Figs. 41 and 42 to single values resulted in only a 3.1%
difference from the multistep calculation.  This result tends to
support the applicability of the concept of a manifold system,
such as specified in EPA method 14, which results in average
concentration measurements from a number of sampling locations.

     Considering only averaged data from 3 sample stations, the
center and both ends of the monitor, also had little effect on
the calculated total emission, resulting in a 5.8% difference.
However, when only data from a single port was used to calculate
the emission from the entire 15.24 m length of roof ventilator,
the "error" ranged as high as 74.4%.  This result might cause
one to question the wisdom of the velocity measurement specifi-
cation of EPA Method 14, which requires only a single measure-
ment point for every 85 m of length.  The flow through aluminum
potroom roof ventilators may be more uniform than that encounter-
ed at the Hitchcock Industries field test site, however.

     To summarize, the sampling technique demonstrated through
the test program described above consists of velocity and partic-
ulate concentrations at an unspecified number of sampling sta-
tions near the base of the roof ventilator, using widthwise
velocity surveys with heated thermopile anemometers to determine
average velocities and constant flow rate sampling with high
volume samplers to determine particulate concentrations. An as yet
unanswered question is the required number of simultaneous
sampling locations over the large emission surface of a given roof
ventilator.  Engineering judgment will be required with regard to
the expected degree of variation of effluent stream characteristics
with position along the length of the roof ventilator and various
economic factors involved in conducting the tests.

F.   Applicability to Other Emission Sources

     The sampling technique developed for this example of a low
velocity, extended dimension emission source can be generalized
to other low velocity, extended dimension sources in terms of
three basic choices:

     1.   Selection of suitable velocity instrumentation having
          sufficient accuracy in the;expected velocity range,
          yet sufficient durability £o withstand field test
          conditions.             :

     2.   Selection of a high volume particulate sampler to
          ensure sufficient volume of gas sampled to obtain
          useful results.

                            T 70 -

-------
                                    TABLE  11

                  CALCULATED EMISSIONS FOR EXAMPLE PROBLEM
Total Emission,  E =
N
£
               M
               Z
                                (C
3


3


1

3
      M

      WV^M^—

      5


      1


      1


      3
      C.
    (gm/m )

Individual values
From Fig. 41a,b,c

Average values
From Fig.41a,b,c
Average value
From Fig. 41
Individual values
From .Fig. 41a, b, c
at Stations 1,3,5
Average value
From Fig. 41 at
Stations 1,3,5
Individual values
From Fig. 41 at
Station  3
Average value
From Fig. 41
at Station 3
Average value from
Fig. 41 at
Station  5
                    j
                  (m/sec)
            *mi***iiiiiiiim^mmii**i^*iiimm*iii**~*~*^^^^^~l^^**ll^*~^mv*~~*f^~f***~
            Individual values
            From Fig,42a,b,c

            Average values
            From Fig.42a,b,c

            Average value
            From Fig. 42.

            Individual values
            From Fig.42a,b,c,
            at Stations 1,3,5

            Average value
            From Fig. 42 at
            Stations 1,3,5

            Individual values
            From Fig 42 at
            Station 3

            Average value
            From Fig. 42
            at Station 3

            Average value from
            Fig. 42 at
            Station 5

A.
5,58
27.9
27.9
9.3

t.
i
(sec)
7200
7200
21,600
7200 .
S
E
(gm)
79,8
82,7
82.3
81.3

% Difference
-
+3.6
+3.1
+1.9
                                                           27.9    21,600    84.4
                                                           27.9     7200    133.4
                                                           27.9    21,600   139.2
                                                           27.9     21,600    49.2
                                                                                           +5.8
                                                                                           +67.2
                                                                                           +74.4
                                                                                           -38.3

-------
     3.   Selection of appropriate sampling locations where
          external disturbances are minimized and sufficient
          information can be obtained to assess the total
          emissions.

The general concept should be applicable to any low velocity,
extended dimension emission source where satisfactory choices
can be made.

     In certain casest the heated thermopile type velocity
instrument may be unsuitable, such as in basic oxygen furnace
(EOF) shop ventilators, where the effluent streams reportedly
are characterized by temperatures above the operating limit of
this instrument.  Several velocity instruments are discussed iri
the Appendix, which can be used as a guideline for selection of
suitable velocity instrumentation.

     The selection of sampling locations presents the principal
problem in adapting the method to other sources, since a clear
representation of this aspect was not possible within the scope
of the present study.  Again, at this point in time, engineering
judgment must play a large role in determining the number of
sampling points needed.
                             - 72 r

-------
              IV.   WET SCRUBBER SAMPLING TECHNIQUES

     The principal characteristic of the flow downstream of a
scrubber is the presence of saturated gas and/or entrained
liquid droplets.  It was determined that various problems assoc-
iated with this characteristic, and with the various devices
used to remove liquid droplets from the flow  (mist eliminators) ,
generally cause the standard Method 5 sampling train to be
inadequate for emission measurements.  In order to examine
possible alternative sampling techniques, a program of method-
ology review, field testing and model testing, following the
same basic approach taken for the study of roof ventilator
sampling techniques, was undertaken.

A.   Review of Sampling Methodology

     A wide variety of sampling equipment has found application
in sampling scrubber exhaust streams, including such devices as
wet and dry impingers, impactors with various substrates, cyclone
precutters, and fabric filters.  This is due in part to the wide
variety of pollutants, both gaseous and particulate, which are
found in scrubber applications.  A partial list of particulate
sampling trains used by various organizations may be found in
Ref. 35.

     Several aspects which must be considered in scrubber emis-
sion measurements are specified in Refs. 36 and 37.  When samp-
ling a wet scrubber system, particulate concentration should be
analyzed on a dry gas basis so that inlet and outlet conditions
can be compared for a realistic collection efficiency computa-
tion.  Thus, the dry gas flow rate and the volume of entrained
liquid must be measured.  In addition, isokinetic sampling is
very important since droplet sizes may have a wide range depend-
ing on the particular entrainment separator used.

     Scrubber sampling methods have usually specified that the
sample probe be heated  (Refs. 16,22,35).  The required tempera-
tures may be critical depending on both the chemical makeup of
the exhaust gas and the sampling train.  Temperatures should
generally remain above the scrubber fluid dewpoint, but not so
high as to vaporize Various liquid pollutants of interest.  As
in the case of roof ventilator sampling, engineering judgment
plays a large role in the selection of sampling equipment and
techniques  (Refs. 17, 36).
               i
     A great deal of information concerning sampling techniques
and problems was obtained through contacts with twelve manufac-
turers of wet scrubber systems.  Since scrubber manufacturers
often tend to specialize in wet scrubber systems for a particu-
lar industry, the sampling techniques used vary according to the
type of pollutants encountered in these various industries.  The
most common practice was found to be use of the basic Method 5
                             -  73  -

-------
 sampling train with modifications required to overcome specific
 problems encountered.  Typical modifications include various
 placements of the filtration assembly relative to the other
 sampling train components and the use of liquid droplet precutters
 at the  sampling probe inlets.  Precutters are usually of the
 cyclone type, several of which are described in Ref. 37, or
 inertial impaction separators.  One type of precutter is illus-
 trated  in Fig. 43.  Occasionally, totally different sampling
 techniques have been employed in particularly troublesome situa-
 tions .

     Other difficulties discussed in these manufacturer contacts
 included the variation of particulate concentrations in scrubber
 system  ductwork due to gravitational settling and wall impinge-
 ment and problems associated with sustained high temperature of
 heated  sampling probes.  In certain cases, the probe temperature
 of 121°C  (250°F) specified by Method 5 was suspected to cause
 breakdown of various particulates.  Several manufacturers also
 have encountered unusually large degrees of droplet carryover
 from the mist eliminators in certain systems, usually in small
 scrubber systems or systems with high velocities.  One manu-
 facturer observed that liquid entrainment became very severe
 above a critical velocity on the order of 11.7 to 12.7 m/sec
 (2300 to 2500 ft/min).

     Possibly the most common difficulty reported was the pres-
 ence of cyclonic flow in the stream to be sampled, often induced
 by the  mist eliminator.  This problem was observed by many
 manufacturers. In addition, seven of nine scrubber systems
 discussed in Ref. 21 had swirling or cyclonic flow.  Sampling
 problems reported in Ref. 21 were handled either by orienting
 the sample probe in the direction of maximum velocity or by
 adding  permanent "egg crate" or vane-type flow straighteners.
 The use of flow straighteners ranging from bundles of stove
 pipes to sophisticated vane arrangements were reported by various
 manufacturers.  Concern was expressed, however, that overall
 flow rate and particulate concentration may be altered when  such
 flow straightening devices were introduced.

     Cyclonic flow presents problems in determination of the
 volumetric flow rate since the axial component of velocity must
 be known at all points in a given cross section in order to
 calculate the volumetric flow rate through that section.  A
 hypothetical cyclonic flow is illustrated in Fig. 44 showing  the
 normal  method for finding this axial component, which requires
 the ability to measure the yaw angle.  i;'f an S-type pitot tube
were used to attempt direct measurement of the axial component
of velocity, the measurement would be in error since the output
of a pitot tube does not correspond to the cosine of the yaw
angle,  as indicated by the data shown in Fig. 45   (extracted
from Refs.  2 and 38).
                            -  74  -

-------
              Probe Inlet
0)
•l-l
(U
'O
•H
W
                                       Water Trap
                S-tube
    FIGURE 43,  INERTIAL IMPACTION LIQUID DROPLET SEPARATOR

                USED BY A WET SCRUBBER MANUFACTURER'S TEST  GROUP

-------
                                     U

                                     U
                                     U
                                            — S-tube oriented
                                             along this line
                             =  Actual Duct Velocity -

                             =  Axial Velocity
                                 Component

                             =  Tangential Velocity
                                 Component

                             -  Angle between axial
                                Direction and the Flow
                                 Direction at the
                                 Sample Point

                             =  U cos 0
FIGURE 44.
METHOD USED TO DETERMINE AXIAL COMPONENT
IN A SINGLE VORTEX CYCLOKKC FLOW
                         -  76  -

-------
           40       -20
            6,  degrees
                                 +10%
                                                  +20
                                                  6, degrees
-J
1
                               ---10%
                                 Velocity
                                  Error
                               ---20%
                                --30%
               4,57 m/sec (15 ft/sec) - Model Test Section
               15.24 m/sec (50 ft/sec)- Model Test Section
               9.14 m/sec (30 ft/sec) - 20.3 cm Dia.- Free Jet
                                I
FIGURE 45.   VELOCITY ERROR WITH YAW ANGLE  (3/8" S-TUBE)

-------
     Several studies have been made concerning sampling within
vortices or cyclonic flow regions  (e.g.* Refs. 39-42).  It has
been observed that in a flow pattern consisting of a vortex
motion superimposed on an axial motion, or a single vortex
cyclonic flow, the axial velocity profile varies considerably as
the swirling velocity increases.  The axial velocity component
at the duct center decreases relative to the axial velocity
component at larger radius values as the swirling velocity
increases.  Another measurement consideration is the possible
effect of the inserted probe size on a cyclonic flow  (Ref. 39).
Insertion of a large probe in a small duct can readily disturb a
cyclonic flow.

     Various types of instruments have been used to determine
volumetric flow in cyclonic flow fields.  Several directionally
sensitive pitot tubes are shown in Fig. 46. One of the most
widely used is commonly referred to as the Fecheimer probe.  The
principle of operation of the Fecheimer probe is related to the
pressure distribution around a cylinder in cross flow, as shown
in Fig. 47.  The static pressure angle is quite uniform at
39j.25° for values of Reynold's number in the range 10  - 2 x
10 .  By arranging static pressure taps at the two locations and
a total pressure tap as indicated in Fig. 48  (from Ref. 44), the
yaw angle can be found by rotating the probe until a null point
is reached between the two static pressure taps, and the dynamic
pressure is then the difference between the total and static
pressure taps.  Variations of this concept are illustrated in
Figs. 49 and 50, extracted from Refs. 45 and 46.

     Other directional sensing velocity instruments appearing in
the literature include spherical head direction pitot tubes
(similar to Fig. 45), and single or multiple element hot wire or
hot thermopile anemometers.  These devices have not been used as
extensively in large duct cyclonic flow as the Fecheimer probe
or its variations.

     To summarize, the primary problems related to particulate
sampling at the exhaust of wet scrubbers revealed through a
survey of literature and contacts with scrubber manufacturers
were: 1)  various difficulties in recovering the pollutant,
either in dissolved form or in an undissolved mixture, from the
liquid droplets, and  2)  difficulties caused by cyclonic flow
often resulting from mist eliminators intended to remove the
droplets from the exhaust stream.

B.   Preliminary Field Test$

     The Seneca Waste Water Treatment Plant in Eagan, Minnesota,
was selected as the site for initial field testing.   The air
pollution control system at this site includes two Peabody
impingement tray scrubbers downstream of a twin incinerator
system used to burn the sewage sludge obtained from the  final
                             -  78  -

-------
                      of rotation

                                          (e)
                                          (*)
                    (W
                      (e)

                                          (9)
FIGURE 46.  SEVERAL TYPES OF DIRECTIONAL PITOT TUBES
                       (REF. 43)
                         r-. 79 i-

-------
             State pressure
                 angle  '
             Ftow	»
                            V..--'
FIGURE 47.   PRESSURE  DISTRIBUTION OVER  A
               CYLINDER  IN  CROSS FLOW
                                  Total Pressure Tap
          Total Pressure Tap
            Connection
          Static Pressure Tap
            Connections
                                           Static Pressure Taps
                                       Total Pressure Tap

                                       Static Pressure Taps
                            Null Balance
. Null Balance
FIGURE 48.   TYPICAL  FECHEIMER  PROBE AND  PRESSURE
               MONITORING  SYSTEM  (REF, 44)

                            -  80  -

-------
                                                        r- / 0-2
                                                         to / VELOCITY
                                                             HEAD
                                                            IN. W G.
FASTENING SCfiEWS (4| 	",

               -£
                      i. LOCATING STOOS !3

               (Zl Z-GHANITE GASKET   PER?OrtATCD CYUNOEH
                                                  V-
     FIGURE 49,
FECHEIMER  PROBE BUILT  INTO FILTER HOLDER
            (REF.  45)

            - 81 -

-------
     Directional Pressure Taps
c
Pressure
                  \

                                  \
                                                Probe Tip
                                           •Static Pressure Tap
       FIGURE  50,   CONNECTICUT STATE DEPARTMENT OF
                   ENVIRONMENTAL PROTECTION PROBE
                          (REF.  46 )

                           -  82  -

-------
process at the treatment plant.  The  flow rate capacity of these
scrubbers is 340 m /min  (12,000  ft  /rain) .   Impingement tray
scrubbers of this size are commonly found within the industrial
community, particularly in the collection of particulate matter
from incinerators.  A diagram of the  wet scrubber system and the
exhaust ducting is shown in Pig. 51.  Each  scrubber has a fixed
vane type centrifugal mist eliminator at the exhaust end.

     The sampling train assembled for preliminary field testing
followed EPA Method 5 specifications, including a glass lined
heated probe and attached S-tube, a filter  holder, two wet and
two dry impingers or a condenser apparatus, a container of
silica gel desiccant, a vacuum pump,  a dry  gas meter and a
rotameter.  A Fecheimer probe was selected  to measure flow
angularity.  After the angle of  flow  was determined at a given
location, the sample probe and S-tube were  oriented in the flow
direction as determined by the Pecheimer probe.  Velocity was
then measured with the S-tube, and  isokinetic sampling was
achieved by adjusting the flow to the S-tube measured velocity.

     The first sampling  in the series of preliminary field tests
was accomplished at points upstream and downstream of the I.D.
fan, at the sample ports shown in Fig. 51.  A number of obser-
vations were made concerning the exhaust stream characteristics
at these two locations.

     A large degree of cyclonic  flow  was present at the upstream
sampling location, with  yaw angles  as large as 60° being measured.
The measured velocities  ranged from 12.2 m/sec to 17.0 m/sec.
The duct walls were moist at this location, but only a small
amount of liquid droplets were detected  in  the sampling train.
The flow downstream of the I.D.  fan,  however, was much more
uniform, and liquid droplets were present to such a degree as to
thoroughly wet the sampling probe.  Large pressure drops occurred
quickly as droplets built up on  the filter  mats and the vacuum
pump was not able to maintain isokinetic sampling rate, forcing
sampling times to be kept very short.  The  average particulate
concentration measured at the two locations also differed, being
97 mg/Nm  upstream and 61 mg/Nm  downstream of the fan.

     An illustration of  these differences was provided by placing
the filter holder immediately after the  sample probe nozzle.
Filter samples taken upstream and downstream of the fan using
this arrangement are shown in Fig.  52.   It  can be seen that  the
upstream filter is much  more uniform  in  density of collected
particulate than the downstream  filter.  Whether or not the
presence of liquid droplets was  the cause for the rather uneven
collection of particulate at the downstream location was not
determined.

     Another aspect of the first field tests was an investiga-
tion of the degree of particulate lost on the walls of the glass

                            - 83 -

-------
   »7
00
                   I.D. Fan
    Flow Direction
    Indicated by
    Arrows

        FRONT VIEW
                                              \    /
                                                           Sample
                                                           Ports
                                                           \
\
          Wet Scrubber
           (Extends Down
           Two Floors)

           SIDE VIEW
              FIGURE 51. WET SCRUBBER AND EXHAUST DUCTS AT THE SENECA WASTEWATER TREATMENT PLAN1
                                            EAGAN, MINNESOTA

-------
FIGURE 52.
FILTER SAMPLES  TAKEN AT SENECA
WASTEWATER TREATMENT PLANT
DURING PRELIMINARY FIELD TESTS

 A = UPSTREAM OF I.D. FAN

 B = DOWNSTREAM OF I.D. FAN
                    - 85  -

-------
lined probe due to impingement and gravimetric settling of
liquid droplets.  The probe was sealed after a series of tests
and returned to the FluiDyne Laboratories, where it was cleaned
and all particulate matter deposited within the probe was collect-
ed.  It was found that the losses within the probe amounted to
13% of the total collected particulate.

     Following the first series of field tests, sampling ports
were installed at point A in Fig. 51, just downstream of the
scrubber.  Installation costs were shared by the Metropolitan
Waste Commission, whose members were very helpful and cooperative
during the field test study.  In addition to the added sample
ports, a modification of the sample train was made in order to
alleviate the problems observed in the liquid droplet environ-
ment.  This consisted Of adding an inertia! precutter to the
sampling train, followed directly by the filter holder, as
illustrated in Figs. 53 and 54.  The precutter was designed and
built by FluiDyne from stainless steel.  The design particle cut
size is 10yi\; larger particles are impacted on the cylinder wall
and collected within the container, to be removed after each
sampling run for analysis.

     After these modifications to the ductwork and the sampling
train had been completed, a second series of preliminary field
tests was made.  The tests included 48-point velocity surveys and
12-point particulate concentraton surveys at the new sample
location  (point A in Fig. 50) and downstream of the I.D. fan.
Again, the Fecheimer probe was first used to determine the yaw
angle of the flow at each sample point.  A flow direction tuft
was then used to visually determine the yaw angle as a check,
since the Fecheimer probe was rather large  (  2.54 cm or 1 in)
for the ducting being sampled.  The visually determined angle in
all cases confirmed the indications of the Fecheimer probe.  The
sample probe and S-tube were then oriented in the measured flow
direction, the velocity was measured, and an isokinetic exhaust
gas sample was extracted.

     Typical total velocity and axial velocity component profiles
are shown in Fig. 55-  When the axial velocity component was
used to calculate volumetric flow rates at the scrubber exhaust
(point A) and downstream of the I.D. fan, the results agreed to
within 6%, indicating that this method of velocity determination
is useful in cyclonic flow of the type observed in this field
test situation.  The degree of cyclonic flow at the scrubber
exhaust was severe, with yaw angles as }.ar$e as 75°, while the
flow downstream of the fan had a very small degree of cyclonic
motion.                                        '

     Average particulate concentration was aga|n seen to vary
from 100 mg/Nm  at the scrubber exhaust to 41 mg/Nm  downstream
of the fan on one day of testing, in relatively close agreement
                            r«.  86 *-

-------
FIGURE 53.
INERTIAL SEPARATION PRECUTTER USED IN
PRELIMINARY FIELD TESTS AT  SENECA WASTE-
WATER TREATMENT PLANT
FIGURE  54.   INTERNAL VIEW OF INERTIAL SEPARATION PRECUTTER
                        - 87 -

-------
o
o

rH

II
0)
4J
c

8
0)
o

10
4J
to
•H
Q


1
             Swirl Velocity (U)
                                     Axial Velocity  (U  )
      1.00
0.80 -
0.60 -
0.40 -
    0.20  -
                      Indeterminant

                 / near the  cAnter due
                      to turbulence
                     10     15     20     0


                        Velocity  (m/sec)
                                               10    15    20
  FIGURE  55.  TYPICAL VELOCITY PROFILE  OBTAINED

              DURING PRELIMINARY FIELD TESTS

               AT SENECA WASTEWATER TREATMENT PLANT

                               -  88 -

-------
to the first measurements.  Measurements made on  the  following
day indicated a concentration of  56 mg/Nm   at the scrubber
exhaust, however.

     The inertial precutter appeared  to eliminate the problem
of rapid increase in pressure drop due to buildup of droplets on
the filters.  When the precutter  was  opened after sampling,
however, no liquid was found.  When the water vapor content was
measured with an Alnor dewpoint indicator,  the exhaust stream
was found to have only a  79% relative humidity.   Thus, it was
concluded that the liquid droplets impacted on the wall of the
precutter were evaporated due to  the  non-saturated conditions.
This interesting finding  leads to the conclusion  that one should
not assume a saturated stream based on the  presence of liquid
droplets as suggested in  Method 5 and that  an accurate hygrometer
or Wet/dry bulb thermometer readings  should be used to calculate
water vapor content.  Large discrepancies in total water vapor
content calculated from dewpoint  measurement and  by the Method 5
calculation were also observed in Ref. 47.

     The variation in liquid droplet  concentration is probably
explained by the cyclonic flow resulting from the mist elimina-
tor.  Droplets which pass through the mist  eliminator are driven
to the duct walls by the  cyclonic flow and  pass along the walls
into the I.D. fan, where  they are evenly distributed and re-
entrained by the flow leaving the fan.

     In summary, the preliminary  field tests of a scrubber
exhaust stream provided a number  of observations.  The described
method for evaluation of  the cyclonic flow  field  appears to have
merit.  Some method for reducing  the  amount of liquid droplet
buildup on the filters is necessary.  The inertial precutter did
appear to accomplish this goal, although the subsequent evapor-
ation of the liquid made  recovery of  the particulate carried by
the impacted droplets after each  sample difficult.  Determina-
tion of liquid content by means of an accurate instrument is
necessary.

C.   Model Studies
     - 	 -   ~   '"~                     *,
     Following the overall testing plan outlined  for  the comple-
tion of this study program, a laboratory model was designed to
simulate typical flow situations  occurring  in the exhaust streams
of wet scrubbers.  The overall design concept .is  illustrated in
Fig. 56.  The model is to use the same blower as  used to power
the roof ventilator model.  A fixed vane type cyclonic mist
eliminator was purchased  from Peabody Engineering Corp., Stam-
ford, Connecticut, to be  placed in the model in order to simu-
late actual flow profiles.  Liquid droplets and particulate
matter will be introduced into the model by means of  a spray
nozzle.  At the present writing of this report, the model is
under construction at the FluiDyne Medicine Lake  Aerodynamic
Test Facility.
                             - 89  -

-------
o
1
             1 meter
FIGURE 56.
                         PROPOSED CYCLONIC  FLOW AND ENTRAINED LIQUID DROPLET TEST SYSTEM
                         LOCATED AT THE MEDICINE  LAKE AERODYNAMIC TEST FACILITY
                         AS VIEWED FROM THE SOUTHWEST

-------
                         V.   FUTURE WORK

A.   Wet Scrubber Sampling Techniques

     Future efforts will be directed toward  evaluation of sampling
techniques in the flow model  described  in  Section  IV-C.  Initial
studies will concentrate on the  selection  of appropriate velocity
instrumentation  for proper computation  of  volumetric  flow rate
in the cyclonic  flow  field produced by  the mist  eliminator.  The
second phase of  model testing will  attempt to evaluate the
accuracy of selected  sampling trains in measuring  the concentra-
tion of particulate introduced  in the flow.   The particular
sampling train configuration  determined to best  fit the require-
ments of the scrubber application will  then  be evaluated in a
final field test, and conclusions and recommendations will be
made as to its usefulness  in  emission measurement  in  other
emission sources with saturated  gas streams  or entrained liquid
droplets.

B.   Grain Dryer Sampling  Techniques

     A program following the  same plan  as  used in  evaluating
roof ventilators and  wet scrubbers  will be undertaken for the
study of grain dryer  emission measurement.  A review  of presently
used sampling methodology  will  be followed by a  preliminary
field test in order to determine typical  sampling  problems and
conditions.  A  laboratory  model will then  be constructed using
the existing blower and ductwork at the Medicine Lake Aerodynamic
Test Facility, and  a  model test program will be  performed to
evaluate techniques and instrumentation for  measurement of grain
dryer emissions.  The final  sampling technique will be further
tested in  the  field,  if necessary,  and  evaluated as to its
usefulness with  other partially confined  emission  sources.
                             « 91 -

-------
           VI.   INTERIM CONCLUSIONS AND RECOMMENDATIONS

     The test program for evaluating roof ventilator emissions
led to the conclusion that a combination of a high volume parti-
culate sampler and a heated thermopile anemometer provide a
reasonable, conservative estimate of total particulate emission
through a cross section of a roof ventilator when deployed at a
sampling location near the base of the ventilator.  Widthwise
velocity surveys of on the order of 4 to 6 points at each
sampling location were required to determine the average
velocity.  The number of sampling locations required to obtain a
given emission measurement accuracy for a given length of roof
ventilator must be determined for each specific application;  an
example based on field tests at a secondary aluminum foundry
showed that three sampling locations were needed for a 15.24 m
ventilator.  Engineering judgment should be used to determine a
balance between economic considerations and the maximum amount
of information obtainable in selecting a number of sampling
locations when making emission measurements at a given location.
                            - 92 -

-------
                             REFERENCES


 1.    "Standards of Performance for New Stationary Sources,"
      Federal Register, Volume 36, No. 247, December 23, 1971.

 2.    Hanson, H. A., et al, Particulate Sampling Strategies
      for Large Power Plants Including Non-Uniform Plow,
      EPA-600/2-76-170, June 1976.              '	'	

 3.    Vandegrift, A. E., et al, Particulate Pollutant System
      Study, Volume I - Mass Emissions, EPA Report APTD-0743,
      May 197IT                    ~~

 4.    Shannon, L.  J., et al, Particulate Pollutant System
      Study, Volume II - Fine Particle Emissions, EPA Report
      APTD-0744, August 1971.

 5.    Vandegrift, A. E., et al, Particulate Pollutant System
      Study, Volume II - Handbook of Emission Properties, EPA
      Report APTD-0745, May 1971.

 6.    Shannon, L. J., et al, Emissions Control in the Grain and
      Feed Industry;  Volume I - Engineering and Cost Study,
      EPA-450/3-73-003a, December 1973.

 7.    Shannon, L. J., et al, Emissions Control in the Grain and
      Feed Industry;  Volume II - Emission Inventory,  EPA-
      450/3-73-003b, September 1974.

 8.    Compilation of Air Pollutant Emission Factors, EPA Report
      AP-42, April 1973.

 9.    Engineering Bulletin PAR<-64, Penn Ventilator Company, Inc.

10.    ASHRAE Handbook of Fundamentals, American Society of Heating,
      Refrigerating, and Air-Conditioning Engineers, George
      Banta Co., Inc., 1967.
11.
      Singmaster and Breyer, Air Pollution Control in the Primary
      Aluminum Industry, Volume I, EPA-450/3~73-004a, July 1973.

12.   Danielson, J. A. ed., Air Pollution Engineering Manual,
      U. S. Department of Health, Education, and Welfare, Public
      Health Service Publication No. 999-AP-40, 1967.

13.   Hidy, G, M. and Brock, J. R,, "An Assessment of the Global
      Sources of Tropospheric Aerosols," Proceedings of the
      Second Clearj Air Congress, Washington, D.C., December, 1970.

14    Carroll, John J. ,  "Determination of Temperature, Winds
      and Particulate Concentrations in Connection with Open
      Field Burning," Air Resources Board, State of California,
      November 1973.           ,_,     _

-------
15.   Cowherd, Chatten, Jr., et al, Development of Emission
      Factors for Fugitive Dust Sources, EPA-450/3-74-037,
      June 1974.

16.   Atmospheric Emissions from Sulfuric Acid Manufacturing
      Processes, U. S. Department of Health, Education and
      Welfare, Public Health Service, Publication No. 999-AP-
      13, 1965.

17.   Calvert, Seymour, et al, Wet Scrubber System Study,
      Volume I, Scrubber Handbook,EPA-R2,72-118a, August
      1972.

18.   Background Information for Proposed New Source Per-
      formance Standards,;  Steam Generators, Incinerators,
      Portland Cement Plants, Nitric Acid Plants, Sulfuric
      Acid Plants, U. S. Environmental Protection Agency,
      APTD 0711, August 1971,

19.   Background Information for Proposed New Source Per-
      formance Standards*  Asphalt Concrete Plants, Petrole-
      um Refineries, Storage Vessels, Secondary Lead Smelters
      and Refineries, Brass or Bronze Ingot Production Plants,
      Iron and Steel Plants, Sewage Treatment Plants, Volume
      !_, U. S. Environmental Protection Agency, APTD 1352a,
      June 1973.

20,   Background Information for Standards of Performance;
      Electric Submerged Arc Furnaces for Production of
      Ferroalloys, Volume 2.  Test Data Summary, EPA-450/2-
      74-0186, October 1974.

21,   Background Information for Standards of Performance;
      Coal PreparationPlants, Volume 2y  Summary and Test
      Data, EPA 450/2-74-0216, October 1974.

22.   Atmospheric Emissions from Nitric Acid Manufacting
      Processes, U. S, Department of Health, Education and
      Welfare, Public Health Service, Publication No. 999-
      AP-27, 1966.

23.   Kreicheltf Thomas E,, et al, Atmospheric Emissions  from
      the Manufacture of Portland Cement, U. S, Department of
      Health, Education, and Welfare, Public Health  Service,
      Publication No, 999-AP-17, 1967,

24,   Control and Disposal of Cotton-Ginning Wastes, U.. S,
      Department of Health, Education and Welfare, Public
      Health Service; Publication 999-AP-31, 1967,

                             r 94 -

-------
25.   Evans, Robert J., "Methods and Costs of Dust Control in
      Stone Crushing Operations," U. S, Department of the
      Interior, Bureau of Mines, Information Circular 8669.

26.   Stear, James R. Municipal Incineration;  A Review of
      Literature, U. Sf Environmental Protection Agency
      Publication No. AP-^79.

27.   Noll, K. E. , et al, State of the Art of Air Pollution
      Control Techniques for  Industrial Processes and
      Power Generation, Southern Section Air Pollution Control
      Association and University of Tennessee, April 1972.

28.   PEDCO - Environmental Specialists, Inc., "Investigation
      of Fugitive Dust - Sources, Emissions and Control,"
      U. S. Environmental Protection Agency, APTD 1582,
      May 1973.

29.   Kreichelt,  Thomas E. and Keller, Thomas G., "Roof
      Monitor Emissions: Test Methodology," Journal of the
      Air Pollution Control Association, Volume 22, No. 8,
      August 1972, pp. 640-645.

30.   Souka, A.,  et al, "A New Approach to Roof Monitor Parti-
      culate Sampling," Journal of the Air Pollution Control
      Association, Volume 25, No. 4, April 1975, pp. 397-398.

31.   Souka, A.,  "Continuous  Roof Monitor Emission Tests,"
      presented at the symposium on Fugitive Emissions:
      Measurement and Control, Hartford, Conn. May 17-19,1976.

32.   "Performance Standards  for New Stationary Sources;
      Primary Aluminum Industry,"  Federal Register, Vol.  41,
      No. 17, January 26, 1976.

33.   Colpitts, J, W. , Automation of Monitor Sampling, Confi-
      dential  Internal Report, ALCOA Badin Smelting Works,
      July  1968.

34    Gravity Ventilation,  Zurn Industries, Inc. Air Services
      Division, Bulletin PMB-2, p. 4,

35.   Cooper, H.B.H., Jr. and A. J. Rosand, Jr., Source
      Testing  for Air Pollution Control, Environmental Science
      Service Division, 1971.

36.   Calvert, S., Entrainment  Separators  for  Scrubbers,  EPA
      650/2-74-119a, October  1974.


                              ^  95^

-------
37.   Calvert, S., Fine Particulate Scrubber Performance
      Tests, EPA 650/2-74-093, 1974.

38.   Grove, D. J., and W. S. Smith, "Pitot Tube Errors Due
      to Misalignment and Non-streamlined Flow," Stack
      Sampling News, November 1973.

39.   Lea, J. F. and D. C. Price, "Mean Velocity Measure*
      ments in Swirling Flow in a Pipe,"  Flow:  Its
      Measurement and Control in Science and Industry,
      Vblume I, Part 1, Flow Characteristics, Instrument
      Society of America, 1974, pp. 313-317.

40.   N. A. Chigier, "Velocity Measurement in Vortex Flows,"
      Flow; Its Measurement and Control in Science and In-
      dustry, Volume I, Part 1, Flow Characteristics,
      Instrument Society of America, 1974, pp, 399-408.

41.   Orloff, K. L. and H. H. Bossel, Laser Doppler Velocity
      Measurements of Swirling Flows with Upstream Influence,
      NASA CR 2284, July 1973.

42.   Odom, J. L., "Testing Stacks with Cyclonic Flow,"
      Stack Sampling News, Volume 3, No. 3, September 1975.

43,   Ower, E. and R, C. Pankhurst, The Measurement of Air
      Flow, Pergamon Press, 1966.

44.   Steam, Its Generation and Use, 38th Edition, Babcock
      and Wilcox Company, 1972.

45.   Madowski, E. R., "Stack Particulate Sampling,"
      Mechanical Engineering, October 1974,

46.   Method 8, FE 409, Connecticut Department of Environmental
      Protection.

47.   Gilardi, E, F, and H. F, Schiff, "Comparative Results
      of Sampling Procedures used During Testing of Prototype
      Air Pollution Devices at New York City Municipal In-
      cinerator," Stack Sampling News, Volume I, No. 4,
      October 1973.
                              ~ 96 -

-------
             APPENDIX
 VELOCITY INSTRUMENTATION FOR LOW
VELOCITY, PARTIALLY CONFINED SOURCE
       PARTiCULATE SAMPLING
                By

           H. A. Hanson

            D. P. Saari
             -  97 -

-------
                        CONTENTS








Al.0      Introduction                           99






A2.0      Types of Velocity Instruments         100





A3.0      Survey of Applications                1°5





A4.0      Evaluation of Selected Instruments    106





A5.0      Conclusions and Recommendations





A6.0      References
                          - 98 -

-------
                        Al.Q  INTRODUCTION

     Procedures and instrumentation for the determination of
velocity and particulate concentration have been established
and studied for emission sources having well-defined flow
fields and velocities above 1.5 to 2 in/sec  Otef. All.
Little information is available, however, concerning procedures
and instrumentation applicable to emission sources for which
the flow is not well defined or the effluent  stream is not
confined.  An example of such an emission source is the roof
monitor or roof ridge ventilator.  This type  of emission
source is characterized by low velocities and complex, often
circulatory, flow fields.

     Knowledge of the velocity of an emission stream is
necessary for determination of the volumetric flow rate,
which must be! known in order to establish the total particulate
emission level.  A further requirement for establishing the
velocity is dictated by the need to establish isokinetic
sampling rates, although this is typically not of great
importance in the study of roof ventilators due to the small
size of particulate generally emitted by such sources.

     Measurement of the velocity and determination of the
volumetric flow rate are complicated by low velocities and
circulatory flow, as opposed to the situation existing in
power plant ducting and exhaust stacks, for example.  In
such a situation, it has been shown  (Ref. A2) that a small
number of measurements with an S-tube over the cross section
of a duct or stack can often provide an accurate determination
of the volumetric flow rate.  At low velocities  (approximately
1 m/sec or less), however, the S-tube or pitot static tube
introduce a high degree of uncertainty to velocity measurements
due to the extremely small magnitude of the dynamic pressure
of the gas stream.  In addition, the possibility of regions
of reverse flow and large  degrees of nonuniform flow angularity
in unconfined source emissions requires the determination of
the flow direction as well as velocity magnitude in order to
establish the net volumetric flow rate.

     Thus, the selection of suitable velocity instrumentation
for use in connection with low velocity, unconfined sources
is an important consideration.  Therefore, the suitability of
a number of velocity instruments was evaluated for this
purpose.

                               r 99 r^

-------
                A2.Q  TYPES OF VELOCITY INSTRUMENTS

     Several types of velocity measuring instruments, are
 listed in Table Al along with some brief comments, concerning
 the characteristics of each type.  There ake generally
 several manufacturers for each type of instrument, but those
 listed in Table Al proved a representative sample.  A brief
 description of the principle of operation and an evaluation of
 the possible usefulness of each type of instrument is included
 in the following.

     A2.1 Pitot static Tubes or S-Tubes

          The principle of the pitot static tube or S-tube is
 to measure, be means of a probe inserted in the flow, the
 difference between total pressure and static pressure  (or
 nearly static pressure, in the case of the S-tubej.  Thus, the
 dynamic pressure  (or a quantity proportional to the dynamic
 pressure) is obtained directly from the measurement, and the
 velocity may be computed if the fluid density is known.  This
 type of instrument is widely used for velocity measurement, and
 the S-tube is the standard EPA-approved instrument for sampling
 in ducts and stacks.  However, since the dynamic pressure is
 extremely small in low velocity gas flows, accuracy of measure-
 ments becomes poor.  In addition, determination of flow direc-
 tion is not a suitable application for a standard pitot static
 tube or an S-tube.

          The useful velocity range of the S-tube can be extend-
 ed to some extent with the use of a purge gas flow through the
 S-tube and a fluidic amplifier.  The principle of such an
 instrument involves a pneumatic bridge between the two S-tube
 lines, which is balanced in the no-flow condition.  As a pressure
 differential develops between the two S-tube ports, a secondary
 flow of purge gas is induced in the pneumatic bridge which
 is proportional to the fluid velocity.  This arrangement not
 only extends the low velocity range of the S-tube to approx-
 mately 15 m/min, but also prevents fouling of the tube in a
 particulate stream, since no fluid actually enters the S-tube.
 |!owever, the accuracy of this instrument is poor in the very low
 velocity range.

     A2.2 Vane or Propeller Anemometers

          A vane or propeller anemometer is positioned normal
 to the flow velocity.  The flow through the anemometer imparts
 aerodynamic force to the vanes, or propellerr which then rotate
 at a rate proportional to the fluid velocity.  In a mechanical
 anemometer,  the rotation of the instrument is recorded on a
dial as a cumulative linear distance, which provides the fluid

                              >  100  -

-------
INSTRUMENT
Pitot Tubes or S-Tubes
l.with Inclined Nanometer
2. with Micromanometer
3. S- tube with purge flow




Vane or Propeller
Anemometer*
4. Vane Anemometer


5. Rotating Vane(mech)

6. Rotating Vane (elect)

7. Propeller -Vane
8. Propeller with purge
gas cleaning


Heated Element
Anemometers
9. Hot Hire

10. Hot Film


11. Hedge hot film



12. Hot thermopile

Fluidic Anemometers
13. parallel Jet

14 . Perpendicular Jet

15. Heated tube-
purge flow cooled
VELOCITY MAX.OPER.
RANGE TEMP.
(ft/min) (°F)

600+ 800+
400+ 800+
50-1500 800+





*
50-5000 250


30-5000 250

25-5000 250

25-5000. 220

25-5000 220




10-6000 800

10-6000 200

.
500



10-3000 200


30-8000

12-3000 400
V
60-6000

LOW VHUOCirr IMS'
FUNCTIONING IN
PARTICULATE
STREAM

Jonly limited
iby probe
(blockage
Good






Fair


Fair

Fair

Fair

Good




Needs occa-
sional Cleaning
Needs occa-
sional Cleaning

Needs occa-
sional Cleaning


Needs occa-
sional Cleaning

Tested to
25 g/m3
4,5 g/m3

Good

rKUMKNTATlOM
ACCURACY
AT
1 ft/sec

N/A
N/A *e
+50%






+20% (eat}


+20% test)

+20% (eat)

+20% Cestl

+20% (eat)




+20%

+20%






+6%






+50%

DIRECTIONAL
SENSITIVITY


,s (S-tube)
Poor






No


!f
direction
of
alignment
Yes






NO

to 0°
or 180°

No



to 0°
or 180°

Possible

Possible

Possible

RUGGED-
NESS

Good
Good
Good






Fair-
Good

Fair-
Good
Fair-
Good
Fair-
Good





Poor

Fair


Good



Good


Good

Good

Good

TYPICAL I
MANUFACTURER

Dwyer
Dwyer
Hasting*-
^ * k •_
Raydist





Flowrite

~
Davie

Davis,
Gill
QJ11

Gill




Tfiemo-
Systems
Thermo*-
Systems

Thermo-
Systems


Hastings-
Raydist

Bowles

FiuiDynamic

Hastings-
Raydist

APPROX
MODEL COST

$100
$200
TnSr $1900
ID3K





MRF $100


$200

$800

$500

$200
+purge
system


1610 $1000

1650 $500


1234H $1300
, COKMENTS


Affected by
Vibration
"Good for uni-
form flow; aux-
iliary readout
needed below
- 100 ft/min


Direct
* Velocity
indication
"Cumulative
reading
Continuous
reading



gas



Tenp. range
may vary
Response time
may be too
fast
~~Hot temp.
w/readout comp. (+ $100


PCI-30 $700


$200

$1000

AFI-6K $1500

for temp, comp
model.)
~~Bood field
instrument .



Temp. range
may vary



-------
velocity when divided by the measurement time, or as a direct
velocity indication.  Electronic instruments provide a continu-
ous signal which, can be recorded.  The sigkal polarity can he
proportional to the flow direction.  A direct velocity indicat-
ing vane anemometer does not respond correctly to a reversed
flow.  A rotating vane anemometer with a cumulative distance
indication allows the determination of a net velocity over
the measurement time since a reversed flow will cause dis-
tance to be subtracted from the cumulative total.  When the
indicated cumulative distance is then divided by the measurement
time, the resulting velocity is the net velocity which has
passed through the anemometer.  A propeller vane anemometer
aligns itself with the flow, thus providing an indication of
flow direction as well as velocity.

     These instruments allow measurement of lower velocities
than do pitot static or S-tubes.  However, the rate of rotation
is influenced by bearing friction; thus, this type of instru-
ment must be calibrated frequently with wear, and the calibra-
tion may be affected by particulate matter.  The use of a purge
gas flow to clean the bearings is a common means to prevent
fouling of the instrument in a particulate stream.  In addition,
calibrations are nonlinear near zero velocity, causing poor
accuracy at velocities near the stated minimum Capproximately
7.5-15 m/min) for this type of instrument.

     A2.3 Heated Element Anemometers

          The principle of operation of hot wire and hot film
anemometers is essentially the same.  The sensing element
(usually platinum wire or film) is heated, either with a constant
electric current or to a constant temperature.  When the element
is introduced in a fluid stream, the heat transfer from the
element is proportional to the fluid velocity.  The output from
these instruments is then either the resistance of the element,
with constant current, or the electric current necessary to
maintain the heated element at constant temperature, which is
proportional to the fluid velocity.  Comercially available
instruments generally incorporate temperature compensation,  so
that calibrations are not affected by fluid temperature.  Since
the heat transfer from the sensing element will change with  a
significant buildup of particulate matter, this type of instru-
ment will need occasional cleaning for use in a particulate-
laden stream, although the hot film type is less susceptible
to fouling.  Frequent calibration is also advisable.

          A heated thermocouple  (or thermopile) anemometer also
provides output due to the heat transfer from heated elements
in a flowing stream.  In this type of instrument, the sensing
elements are hot junctions of a thermopile, while alternate
junctions are not heated to provide for temperature compensation.
                              - 102 -

-------
The reduced thermocouple current when the hot junctions are
cooled by a flowing gas stream .is proportional to the velocity.
As in the case of hot wire or hot film anemometers, occasional
cleaning is necessary in a parttculate stream.

          Directional sensitivity is not a basic feature of
heated element anemometers.  However, protective caps designed
for some instruments of this type provide the additional
capability to determine yaw angle in a flow.

     A2.4 Fluidic Anemometers

          One type of fluidic velocity measurement instrument
incorporates a jet of fluid, compatible with the fluid whose
velocity is being measured, which is introduced into the flow,
either parallel or perpendicular to the stream.  This fluid jet
impinges on two sensing ports situated such that the pressure
at these two ports is equal for no flow. 'When introduced into
a flowing stream the jet is deflected, causing a pressure
differential between the two sensing ports.  This differential
pressure is the output of the sensor and can be made linear
with the velocity depending on the supply pressure for the jet.
The parallel jet type sensor requires lower supply pressures
than the perpendicular jet type  (Ref. A3).  The operational
principle does not preclude use in moderate particulate
concentrations as long as the tubing systems do not become
blocked.  Two or three sensors could be arranged to determine
orthogonal velocity components.  If positive and negative
pressures can be distinguished, reverse flow should also be
detectable with fluidic sensors.

          Another type of velocity measurement instrument
which employs fluidic principles involves a tube through
which purge gas is continually injected into the flow-  A
pressure differential caused by the flow at two purge gas
outlets induces a secondary flow of purge gas through a
heated tube inside the instrument.  This secondary flow cools
the heated tubing and an electrical output results following
the same principle as a heated element anemometer.  Since the
cooling flow is clean purge gas and not the actual flow being
measured, this type of instrument would function well regardless
of particulate concentration.  However, the output is very
nonlinear and the minimum velocity is about 0,3 m/sec.

     A2.5 Vortex Shedding Anemometer

          A new velocity measurement  instrument under develop-
ment employs vortices generated by a  strut and counted by an
ultra-sonic sensor  CRef. A4J.  The frequency of vortex genera-
tion is directly proportional to the velocity  if the Strouhal

-------
 number  of  the obstruction  is known, thus providing a calibra-
 tion.   The instrument reportedly has great accuracy at veloc-
 ities as low as  0.5 m/sec, and based on the assumption of a
 constant Strouhal  number over a wide Reynold's number range,
 should  be  insensitive to variations in fluid density, tempera-
 ture and pressure.  However, the characteristics of the instru-
 ment have  not been sufficiently established to make a judgment
 as  to its  applicability in particulate streams.  A concept for
 development of a direction-sensitive instrument employing this
 concept has also been reported  CRef. A4).

     A2.6  Ion Deflection Anemometer

           This type of instrument utilizes an ion-emitting
 source  in  a tube aligned parallel to the flow.  The ions are
 projected  radially from the source toward the walls of the
 tube.   A flow through the  tube then causes a deflection of the
 ion stream which is directly proportional to the total mass
 flow through the tube, from which the velocity is deduced.
 Flow direction through the tube is also indicated by this type
 of  instrument making it useful in reversed flow situations.
 The presence of  particulate matter in a flow, however, may
 cause the  velocity calibration to be in error.  Since the
 instrument actually senses the flow of mass, a constant veloc-
 ity flow should  give different velocity indications for differ-
 ent particulate  concentrations.  In addition, the tube would
 require occasional cleaning, since buildup of particulate could
 interfere  with the ion stream and collector.

     A2.7  Sonic  Pulse and  Laser Doppler Instruments

           The sonic pulse  type instrument employs the difference
 in  time for oppositely-directed sound waves to travel a fixed
 distance to deduce the flow velocity.  Two-or three-axis systems
 allow the  measurement of velocity components, giving the inform-
 ation needed for volumetric flow evaluation even in highly
 turbulent  flows.   Particulate matter should have little effect
 on  such an instrument.  The laser Doppler type instrument
 determines the shift in frequency caused by scattering of a
 light from a monochromatic source  (laser}.  The scattering is
 accomplished by  particles  introduced in the flow, assumed to
 move with  the velocity of  the fluid? thus, the velocity assoc-
 iated with the measured Doppler shift is the velocity of the
 fluid.  In a flow  containing various sized particulates, some
 of which, do  not  move with  the fluid velocity, the instrument
 could be affected  by random light scattering by the particulate
matter.  The major obstacle in consideration of both, of these
 types of instruments, however, is the prohibitively high cost.
 In addition, the laser Doppler method has been used only as
a laboratory instrument at this time.

                              - 104 ~

-------
                   A3.Q  SURVEY OF APPLICATIONS,

     Contacts were made with a number of commercial firms
which have had experience in sampling emissions  from low
velocity sources.  This survey revealed that propeller or
vane anemometers and rotating vane anemometers find the
greatest popularity for low velocity sources.  A few investigators also
indicated the use of hot wire anemometers.

     The popularity of these types of instruments is also
reflected in the literature with regard to roof  ventilators.
Reference A5 cites the use  of both hot wire  CAnemotherm) and
rotating vane anemometers and states that the latter type,
either mechanical or electric models, are more convenient for
general field use.  Reference A6 reports a study in which propel-
ler anemometers with electric output  CGill Model 27100) were
successfully used for velocity measurements in roof ventila-
tors .

     EPA Method 14  (Ref. A7) specifies the use of vane or
propeller anemometers with  an electric signal output for
velocity measurement in roof ventilator sampling.  A single
anemometer is to be installed at a point of average velocity,
based on a widthwise velocity traverse, for each 85 m of
roof ventilator length.  Method 14 also specifies that the
anemometers be able to withstand dusty and corrosive environ-
ments .

     Pitot tubes were also  used by some investigators for
low
velocity measurements.  The use of a purge gas system with
an S-tube was reported to be a useful means of approaching the
problem, while an S-tube with a micromanometer reportedly
was not successful for very low veloctiy testing.

     In many cases, testing with the above-mentioned instru-
ments has been done in conjunction with various  types of
protective screens, collection hoods, or flow straighteners.
In addition, assessments of the accuracy of the  instrumenta-
tion in these measurement applications are not available,
since the actual flow rate  and velocity distribution are not
known in field applications.  In short, the overall situation
pointed to the need for further evaluation of the potential
use of a number of instruments in low velocity applications.
                          ^ 105 **

-------
             A4.0  EVALUATION OF SELECTED INSTRUMENTS

     A number of velocity measurement instruments were selected
from the basic list given in Section A2.0 and evaluated as to
their suitability for low velocity, unconfined source particu-  ,
late sampling.  Testing of these instruments was accomplished
either in a model facility of a roof ventilator  (Figs. Al and
A2) or in the field, measuring the flow from roof ridge ventila-
tors atop a secondary aluminum foundry  (Figs. A3 and A4).
While field test evaluations provide a realistic view of actual
conditions to be encountered in using the velocity instrumenta-
tion, it is not possible to determine the accuracy of an instru-*
ment in the field, as discussed previously.  Thus, the roof
ventilator model was used to determine the degree of reliabil-
ity of the various instruments under conditions which simulated
those existing in field applications.  A summary of the tests
of selected instruments follows.

     A4.1 Vane Anemometer - Flowrite Model MRF

          This instrument was tested both in the field and with
the roof ventilator model.  In the field test of the vane
anemometer the instrument was fitted with an extension arm and
hand held at the exhaust from the foundry roof ventilators.
The instrument proved to be sufficiently rugged and easy to
use under the awkward field test conditions, and velocity data
obtained with the vane anemometer appeared to agree reasonably
with expected results.

          When tested in the same manner with the roof ventilator
model, velocities were determined at 30 points in a plane at
the exhaust from the model.  The volumetric flow rate determin-
ed from these measurements was more than 100% greater than the
actual flow rate supplied to the model.  This large discrepancy
was ascribed to the observed unsteady and circulatory flow
emitting from the exhaust; since the vane anemometer cannot
detect a reversed flow, the total velocity indicated by the
anemometer is expected to be too large.  Thus, this type of
instrument was deemed unsatisfactory for use in  circulatory
flow patterns.

     A4.2 Rotating Vane Anemometer - Davis Mechanical
          Vane Anemometer  (Figs. A5 and A6)

          This instrument was tested with' the roof ventilator
model, and volumetric flow rates determi,ne4 from: 24 point
velocity surveys were compared with the actual flow rate

                         r- 106 "<•

-------
 FIGURE A-l.ROOF MODEL LOCATED AT THE MEDICINE LAKE
            AERODYNAMIC TEST FACILITY - LOOKING SOUTH
FIGURE A-2.ROOF MODEL AND DUCTING LOCATED AT THE
           MEDICINE LAKE AERODYNAMIC TEST FACILITY
           - LOOKING SOUTHWEST
                  " 107 ~

-------

FIGURE A-3,   ROOF VENTILATORS AT  HITCHCOCK INDUSTRIES, BLOOMINGTON
              MINNESOTA, VIEWED FROM THE NORTHWEST
FIGURE A-4.ROOF VENTILATORS AT HITCHCOCK INDUSTRIES,  BLOOMINGTON
           MINNESOTA,  VIEWED FROM THE EAST
                        - 108  -

-------
 FIGURE A-5.
 DAVIS MECHANICAL ROTATING VANE ANEMOMETER
 WITH BALL SWIVEL ATTACHMENT
FIGURE A-6.
DAVIS MECHANICAL ROTATING VANE ANEMOMETER
WITH EXTENSION

           -  109  -

-------
supplied to the model.  The anemometer was fitted with an
extension and hand held at the exhaust plane pf the. model.
With thi§ method, at flow rates ranging.; frQro about 75m /rain
to 2QO m /min, the flow rate indicated b,y velocity measure-
ments with the rotating vane anemometer and the actual flow
rate differed by amounts varying from -25% to +75%.  This lack
of consistency was judged unacceptable.

          A second measurement strategy was also evaluated with
this instrument.  In this method, the anemometer was inserted
in the flow through sampling ports located at the base of the
roof model.  This location was deemed more appropriate for
particulate sampling since the flow at the base of the model
was observed to be more uniform and less dependent on wind condi-
tions.  A certain degree of circulatory flow was still present
at the base of the model, however, due to the expansion of the
inlet to the model downstream of the round ductwork which
supplied the flow (Pig. A2).  The flow at the base of the roof
model could be made more uniform by insertion of a baffle in
this diffuser section.  Therefore, the flow entering the model
at the base could be very uniform, with the baffle inserted, or
somewhat circulatory, with the baffle removed.  Both types of
flow were observed at the field test site, related to the level
of activity within the foundry, and the removable baffle thus
allowed the simulation of both types of flow entering the roof
ventilator model.

          Volumetric flow rate, determined from 24-point veloc-
ity surveys at the base of the model, made with the Davis
anemometer, are shown in Fig. A7.  As can be seen, the accuracy
of the measurements is quite poor, although more consistent
than the surveys made at the model exhaust.  The results were
somewhat better when the baffle was inserted, i.e., when the
flow was more uniform, but in both cases the results were
unacceptable.

     A4.3 Hot Wire Anemometer - Thermo-Systerns Model 1610

          This instrument was fitted with a special protective
collar used successfully in previous studies in which veloci-
ties from Q.03 to 4.5 m/sec were measured.  However, the  instru-
ment still proved to be too fragile for use in the field.  When
tested at the aluminum foundry, the measurement probe ,was
quickly damaged and the instrument rendered inoperable.   There-
fore,  the hot-wire type anemometer was judged inadequate  for
field work in roof ventilator sampling.
                           - 110 -

-------
Actual Volumetric Flow
Rate (m3/min)
    350 .
    300
    250
    200
    150
    100  .
     50  .
                       O  Without Uniform
                          Flow Baffle
                                                            100% Accuracy
                                    150
200
250
300
                                                                        +25%
350
                                               Indicated Volumetric Flow
                                               Rate  (
         FIGURE A-7.VOLUMETRIC FLOW RATE AT BASE OF ROOF VENTILATOR
                    MODEL DETERMINED WITH DAVIS MECHANICAL' ROTATING
                    VANE ANEMOMETER
                               - Ill  -

-------
     A4.4 Hot Thermopile Anemometer - Hastings-Raydist
          Model PCI-30  (Fig. A8)

          This instrument was tested extensively with the
roof ventilator model, including velocity surveys at the
exhaust and at the base of the model.  Volumetric flow rates
determined from velocity measurements at the exhaust, as well
as the base when the flow-uniformity baffle was absent (Fig. A9),
differed from the actual by amounts on the order of 60%.   This
may be attributed to the inability of the instrument to dis-
tinguish between velocities 180° apart.  This is substantiated
by observing the comparison of measured and actual flow rates
with the baffle in place (Fig. A9), in which case the agreement
was excellent.

          In order to overcome this problem, a light weight
flow direction tuft was attached to the instrument.  Several
24-point velocity surveys were then made at the base of the
roof model, using the flow direction tuft to align the probe
with the flow and estimate the flow angle.  The flow was
generally found to be vertically upward or downward, and
incorporating the uniform flow baffle eliminated the regions
of downward flow.  As indicated by Fig. A9, these velocity
surveys provided a total volumetric flow rate which agreed
very well with the actual flow rate even in the absence of
the baffle.

          The directional sensitivity of the instrument was also
examined (Figs.AlO and All), and it can be seen that flow
direction (to 0° or 180°) can be determined by observing the
output of the anemometer if the protective cap is used.
Therefore, this instrument could be used to determine flow
direction without a tuft, if a means were available to
distinguish between flows 180° apart.  Pulsating readings
make alignment of the probe without a tuft somewhat difficult,
however, due to the high sensitivity of the instrument.

     A4.5 Hot Film Anemometer - Thermo-Systems Model 1650
          (Fig. Ally

          This instrument was found to be somewhat more
rugged than the hot wire type anemometer and was lightweight
and easy to handle.  It was observed that in regions of
pulsating or reversed flow, the response time of the instrument
was very fast, making reading of the velocity difficult.   In
two tests of the instrument at the base of the roof ventilator
model,  with the uniform flow baffle in place, volumetric
flow rate determined with the hot film anemometer was less
than the actual flow rate by an average of only 3%.  This
instrument would also need an additional means for determining
flow direction in a circulatory flow field.
                          - 112 -

-------
FIGURE A-8
HASTINGS/RAYDIST PCI-30 HOT
THERMOPILE  ANEMOMETER WITH
METER AND EXTENSION
                    - 113  -

-------
Actual Volumetric Flow
Rate
  350  .
   300
   250
   200
   150
   100
    50

Without Uniform
Flow Baffle
With Uniform
Flow Baffle
With/

©
B

Without/

0
D

                                                             Flow Direction
                                                             Indicator
                                                            100% Accuracy
                                                                        .+50%
               50
100
150
200
250
300
350
                                              Indicated Volumetric Flow
                                              Rate (m3/min)
       FIGURE A-9.VOLUMETRIC FLOW RATE! AT BASE OF ROOF VENTILATOR
                  MODEL DETERMINED WITH HASTINGS-RAYDIST PCI-30
                  HOT THERMOPILE ANEMOMETER

                             - 114 -

-------
         % Error in Velocity
            +8 *
            +6 ,
            +4
            +2 T
            -2
            -6
                             -©-
                   10   20   30   40
                        -©-
                         50
C9(deg)
                                    Plow
    Results Symmetrical

       For -
FIGURE A-10.
VELOCITY ERROR WITH YAW ANGLE FOR HASTINGS-RAYDIST

PCI-30 HOT THERMOPILE ANEMOMETER - WITHOUT

PROTECTIVE CAP

         - 115 -

-------
                          30   40   50   60  70   80   90
         -100

     %  Error in Velocity
                                    Flow


-------
FIGURE A-12.
THERMO-SYSTEMS,  INC.  MODEL 1650
HOT FILM ANEMOMETER
                  - 117 -

-------
               A5.0  CONCLUSIONS AND RECOMMENDATIONS

     The most immediate observation that can be made from
the above evaluation of velocity instruments is that accurate
measurement of the velocity in the complex flow fields
related to such low velocity, partially confined sources as
roof ventilators is a very difficult matter.  Several factors
must be considered in selecting an instrument for such an
application, among them:

     1)   versatility and ease of handling;
     2)   ruggedness;
     3)   ability to determine velocity direction
          as well as magnitude;
     4)   reasonable accuracy at the very low
          velocity ranges encountered;
     5)   ability to operate over ranges of temperature
          and• pi-articulate concentration.
                   V <•
Additional limitations may be placed on the selection of
velocity instrumentation by cost considerations.

     Considering these overall criteria, and on the basis
of the evaluations described above, the most likely candidate
for general velocity measurement in low velocity, partially
confined sources is the hot thermopile anemometer  (Hastings-
Raydist Model PCI-30).  This instrument is relatively light-
weight and easy to handle, is very rugged, showed excellent
results when used in relatively uniform flow and when used
with a flow direction tuft, and operates reasonably well
over a fairly broad range of temperature and particulate
loading.  In addition, the ability to determine flow direction
could be utilized in circulatory flow if a means were available
to distinguish between flows 180° apart.

     The instrument may not be applicable to certain specific
applications, however, such as low velocity sources with
very high gas temperatures.  The overall list of instruments
shown in Table AI may be used as a guideline for selection of
more appropriate instrumentation in such cases..
                           - 118 -

-------
                    A6.0  REFERENCES
Al.  "Standards of Performance  for New Stationary Sources,"
     Federal Register, Volume 36, No. 247, December  23,
     1971.

A2.  Hanson, H. A.,  et al,  Particulate Sampling  Strategies
     for Large Power Plants: Including Nonuniform Flow,
     EPA-600/2-76-17G, June 1976.

A3.  Neradka, V. F.,  "Fluidic Wind Sensors,"   Flow;   Its
     Measurement and Control in Science  and  Industry,
     Volume  I, Part  2, Flow Measuring Devices, R. E.  Wendt,  Jr.
     ed.,  Instrument Society of America,  Pittsburgh,  1974,
     pp. 985-988.

A4.  Klass,  Philip  J.,  "Airspeed Sensor  Accurate to  1%, 1
     kt",  Aviation  Week  and Space Technology,, November 6,
     1972, pp.  43-47.

A5.  Kreichelt, Thomas E.  and Keller, Thomas G., "Roof
     Monitor Emissions:   Test Methodology,"  Journal  of The
     Air Pollution  Control Association,  Volume 22, No. 9,
     August  1972, pp.  640-645.

A6.  Souka,  A., Marek,  R., and Gnan  L.,  "A New Approach to
     Roof  Monitor  Particulate Sampling"  Journal  of the Air
     Pollution Control Association,  Volume 25, No. 4, April

A7.  "Performance  Standards for New Stationary Sources:
     Primary Aluminum Industry," Federal Register, Volume 41,
     No.  17, January 26, 1976.
                         - 119 -

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
 1. REPORT NO.
              EPA-600/2-77-036
                              2.
                                                           3. RECIPIENT'S ACCESSION NO.
 4. TITLE AND SUBTITLE
  EFFECTIVE  SAMPLING TECHNIQUES FOR PARTICULATE
  EMISSIONS  FBOM ATYPICAL STATIONARY SOURCES
  Interim  Report                   	
                                5. REPORT DATE
                                  February 1977
                                6. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)
                                                           8. PERFORMING ORGANIZATION REPORT NO.
  H. A.  Hanson and D.  P. Saari
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
  FluiDyne Engineering Corporation
  5900  Olsaon Memorial Highway
  Minneapolis, Minnesota  55422
                                                            10. PROGRAM ELEMENT NO.
                                   1AD712
                                 11. CONTRACT/GRANT NO.

                                   68-02-1796
  12. SPONSORING AGENCY NAME AND ADDRESS
   Environmental  Sciences Research Laboratory  -  RTP,  NC
   Office of Research and Development
   U.S.  Environmental Protection Agency
   Research Triangle Park, NC  27711	'
                                 13. TYPE OF REPORT AND PERIOD COVERED
                                   Interim 6/75 - 9/76	
                                 14. SPONSORING AGENCY CODE


                                   EPA/600/09
  15. SUPPLEMENTARY NOTES
  16. ABSTRACT
      Techniques and instrumentation for  sampling  strategies to measure particulate
      emissions from "atypical" stationary  sources were developed.  The four atypical
      source categories are low effluent  streams,  extended dimensions, partially  or
      totally unconfined flow, and saturated  gas  streams or gas streams with entrained
      liquid droplets.   The research program  included literature surveys, laboratory
      model  testing, and field testing of several  atypical stationary sources.  Tech-
      niques and instruments were evaluated as  to  the degree of reliability of measured
      emissions and applicability to general  situations.

      Three  specific sources—gravity roof  ventilators, grain dryers, and wet scrubbers-
      were selected to  provide the basis  for  the  research program of the four atypical
      source categories.  Basic characteristics of these sources were identified
      through literature and personal contact surveys.   A program of model testing and
      field  testing of  roof ventilator emissions was completed, and a similar program
      was undertaken for wet scrubbers.   The  sampling strategy recommended for roof
      ventilator emission measurement on  the  basis of the test program includes a high
      volume particulate sampler and a heated thermopile anemometer deployed near the
      base of the ventilator.
 7.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                   b.lDENTIFIERS/OPEN ENDED TERMS
c. cos AT I Field/Group
    Air Pollution
    Particles
    Industrial  Plants
    Nonuniform  Flow
    Sampling
    Systems
    Evaluation
* Field Tests
     13B
     131
     20D
     14B
 8. DISTRIBUTION STATEMENT

    RELEASE TO PUBLIC
                    19. SECURITY CLASS (ThisReport)

                      UNCLASSIFIED
21. NO. OF PAGES
     130
                                              20. SECURITY CLASS (Thispage)

                                                 IINfl ASSTFTFD	
                                              22. PRICE
EPA Form 2220-1 (9-73)
                                            -120-

-------