United Stataa
Environmental Protection
A(jency
Environmental Sciences Research EPA-600/2-80-034
Laboratory January 1980
Research Triangle Park NC 27711
neaearch and Development
Effective Sampling
Techniques for
Particulate Emissions
from Atypical
Stationary Sources
Final Report
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental 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-600/2-80-034
January 1980
EFFECTIVE SAMPLING TECHNIQUES
FOR PARTICULATE EMISSIONS FROM
ATYPICAL STATIONARY SOURCES
Final Report
By
D. P. Saari and H. A. Hanson
FluiDyne 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 use in measurement
of particulate emissions from "atypical" stationary sources
were developed and evaluated. The four atypical source
categories studied are low velocity effluent streams,
emission planes having one or more extended dimensions,
partially or totally unconfined flow, and saturated gas
streams or streams with entrained liquid droplets. Basic
characteristics of these source categories were identified
through literature and personal contact surveys. The
research program also included laboratory model testing and
field testing of several atypical stationary sources.
Three specific sources - gravity roof ventilators,
grain dryers, and wet scrubbers - were selected to provide
the basis for the development program for the four source
categories. The sampling technique recommended for roof
ventilator emission measurements includes a high volume
particulate sampler and a heated thermopile anemometer
deployed near the base of the ventilator. The same instru-
ments, deployed at the louver exhaust, are recommended for
grain dryer emission measurements. An EPA Method 5 type
sampling train, an inertial droplet separator, and a Fecheimer
probe are recommended for use in wet scrubber emission
measurements. These techniques and instruments were evaluated
as to the degree of reliability of emission measurements in
the specific sources as well as applicability to general
atypical stationary sources.
This report was submitted in fulfillment of Contract
No. 68-02-1796 by FluiDyne Engineering Corporation under the
sponsorship of the U. S. Environmental Protection Agency.
111
-------
CONTENTS
Abstract iii
Figures vi
Tables xiii
Acknowledgements x^v
1. Introduction 1
2. Conclusions and Recommendations 3
Roof ventilator emissions 3
Grain dryer emissions 3
Wet scrubber emissions 3
General- atypical emission sources 4
Further study 4
3. Classification and Evaluation of Emission
Sources 5
Emission sources with low velocity
and/or extended dimensions 5
Emission sources with saturated
gas streams or entrained liquid droplets . .23
Selection of sources for test program .... 26
4. Roof Ventilator Sampling Techniques 28
Review of sampling methodology 28
Preliminary field tests 32
Model studies 43
Final field tests 58
Evaluation of sampling technique 66
Applicability to other emission sources ... 74
5. Grain Dryer Sampling Techniques 75
Review of sampling methodology 75
Preliminary information 77
Model studies 78
Evaluation of sampling technique 86
Applicability to other emission sources ... 86
6. Wet Scrubber Sampling Techniques 90
Review of sampling methodology 90
Preliminary field tests 100
Model studies 106
Evaluation of sampling technique 150
Applicability to other emission sources . . .152
References 153
Appendix Velocity instrumentation for low velocity, .159
partially confined source particulate
sampling
v
-------
FIGURES
Number Page
1 Continuous roof ridge ventilator 10
2 Roof "monitor" or "monitor attachment" . . 10
3 Rack and column type grain dryers
(Ref. 6) 13
4 Unloading grain from box car into a
deep hopper (Ref. 12) 20
5 Unconfined emission sources at a
continuous mix asphalt plant (Ref. 5) ... 20
6 Cross section of roof ventilator showing
test equipment (Ref. 29) 30
7 Sampler system (Ref. 30) 30
8 Emission rate measurements, central
sampling station (Ref. 30) 31
9 EPA Method 14 Sampling System 31
10 Roof ventilators at Hitchcock Industries,
Bloomington, Minnesota 33
11 Typical roof ventilator exhaust at
Hitchcock Industries site 33
12 High volume sampler and probe assembly . . 35
13 Hot wire anemometer and protective collar
assembly . . 36
14 Sampling locations for preliminary field
tests - Hitchcock Industries roof venti-
lator 37
15 High volume sampler assembly mounted in
the roof ventilator at Hitchcock Indus-
tries (end view looking north) 39
VI
-------
Number FIGURES (Cont.) Page
16 High volume sampler assembly mounted
in the roof ventilator at Hitchcock
Industries (top view looking down into
exhaust region) 39
17 Preliminary field test concentration measure-
ments, Hitchcock Industries roof ventilator. . 40
18 Preliminary field test velocity survey 1,
Hitchcock Industries roof ventilator 41
19 Preliminary field test velocity survey 2,
Hitchcock Industries roof ventilator 42
20 Roof ventilator model 44
21 Roof ventilator model test facility
Rosemount Energy Conversion Laboratory .... 45
22 Roof ventilator model test section 46
23 Atomization aerosol generator and hot air
injection system - Rosemount Energy
Conversion Laboratory 47
24 Roof ventilator model at Medicine Lake Aero-
dynamic Test Facility - looking south 49
25 Roof ventilator model and ducting at Medicine
Lake Aerodynamic Test Facility - looking
southwest 49
26 Atomization aerosol generator system with hot
air injection - Medicine Lake Aerodynamic
Test Facility 50
27 Roof ventilator model sample ports -
viewed from northeast 51
28 Flow direction indicating tufts in roof venti-
lator model exhaust 52
29 Smoke generator flow indicator in roof venti-
lator model exhaust 52
30 Schematic of observed flow (end view of roof
ventilator model) 53
VII
-------
Number FIGURES (Cont.) Page
31 Typical flow velocities at various points
in roof ventilator model ............ 55
32 Volumetric flow rate at base of roof
ventilator model .......... ...... 56
33 Particulate concentration measurements in
roof ventilator model .......... ... 57
34 Particulate concentration and velocity mea-
surements at base of Hitchcock Industries
roof ventilator ................ 59
35 Average velocity and particulate concentra-
tion measurements - Sample Port 1, Hitch-
cock Industries roof ventilator ........ 61
36 Average velocity and particulate concentra-
tion measurements - Sample Port 2, Hitch-
cock Industries roof ventilator ........ 62
37 Average velocity and particulate concentra-
tion measurements - Sample Port 3, Hitch-
cock Industries roof ventilator ........ 63
38 Average velocity and particulate concen-
tration Measurements - Sample Port 4,
Hitchcock Industries roof ventilator ...... 64
39 Average velocity and particulate concen-
tration measurements - Sample Port 5,
Hitchcock Industries roof ventilator ...... 65
40 Typical velocity distribution at base of
Hitchcock Industries Roof ventilator. ..... 68
41 Particulate concentration profiles for
roof ventilator emission example problem. ... 70
42 Average velocity profiles for roof venti-
lator emission example problem ......... 71
43 Particulate Collection Device used for
grain dryer emissions measurement
(Refs. 37, 38) ............. .... 76
44 Grain dryer model constructed with
existing air supply at Medicine Lake
Laboratory ................... 79
Vlll
-------
Number FIGURES (Cont') Page
45 Cutaway of grain dryer model showing air
flow passage on one side of model 80
46 Grain dryer model with exhaust louvers .... 81
47 Heated thermopile anemometer deployed at
surface of exhaust louvers 83
48 Sampling point matrices used for velocity
traverses of grain dryer model 83
49 Volumetric flow rate determined with heated
thermopile anemometer at exhaust louvers of
grain dryer model 84
50 High volume sampler deployed at surface of
exhaust louvers for particulate traverses. . . 85
51 Sampling point matrices used for particu-
late traverses of grain dryer model 85
52 Particulate concentration measured with
high volume sampler at exhaust louvers
of grain dryer model 87
53 Concept for particulate sampling of
grain dryers 88
54 Inertial impaction liquid droplet separator
used by a wet scrubber manufacturer's test
group 92
55 Method used to determine axial component
in a single vortex cyclonic flow 93
56 Velocity error with yaw angle (3/8 in S-tube) . 95
57 . Several types of Pitot Tubes (Ref. 51) .... 96
58 Pressure distribution over a cylinder in
cross flow 97
59 Typical Fecheimer Probe and Pressure
Monitoring System (Ref. 52) 97
60 Fecheimer Probe built into filter holder
(Ref. 53) 98
IX
-------
"11TnwQT. FIGURES (Cont.)
Number page
61 Connecticut State Department of Environ-
mental Protection Probe (Ref. 54) 99
62 Wet scrubber exhaust ducts at Seneca Waste-
water Treatment Plant, Eagan, Minnesota. ... 101
63 Typical velocity profile obtained during
preliminary field tests at Seneca Waste-
water Treatment Plant 102
64 Filter samples taken during preliminary
field tests at Seneca Wastewater Treat-
ment Plant 104
65 Inertial Separation Precutter used in pre-
liminary field tests at Seneca Wastewater
Treatment Plant 105
66 Internal view of Inertial Separation
Precutter 105
67 Schematic of scrubber exhaust model 108
68 Scrubber exhaust model and test platform
at Medicine Lake Laboratory 110
69 Droplet injection system for scrubber
exhaust model - external hardware Ill
70 Spray nozzles for droplet injection
system Ill
71 Droplet injection system spray nozzles
producing low droplet .density mist 112
72 Droplet injection system spray nozzles
producing high droplet density mist 112
73 Volumetric flow rate error for 12-point
velocity surveys in scrubber exhaust
model with S-tube parallel to duct
centerline 113
74 United Sensor DC-125 3-dimensional
Directional Probe and Traverse Unit 115
75 Sensing Head of United Sensor DC-125
Directional Probe 115
-------
FIGURES (Cont.)
Number Page
76 United Sensor DC-125 Directional Probe
mounted in scrubber exhaust model for
velocity traverses 116
77 Normalized velocity components measured
0.75 duct diameters downstream of mist
eliminator - United Sensor Probe 117
78 Normalized velocity components measured
2.25 duct diameters downstream of mist
eliminator - United Sensor Probe 118
79 Normalized velocity components measured
3.75 duct diameters downstream of mist
eliminator - United Sensor Probe 119
80 Normalized velocity components measured
5.25 duct diameters downstream of mist
eliminator - United Sensor Probe 120
81 Fecheimer Probe used in velocity survey
of scrubber exhaust model 121
82 Normalized velocity components measured
0.75 duct diameters downstream of mist
eliminator - Fecheimer Probe 122
83 Normalized velocity components measured
2.25 duct diameters downstream of mist
eliminator - Fecheimer Probe 123
84 Normalized velocity components measured
3.75 duct diameters downstream of mist
eliminator - Fecheimer Probe 124
85 Normalized velocity components measured
5.25 duct diameters downstream of mist
eliminator - Fecheimer Probe 125
86 Ratio of measured and actual volumetric
flow rate for 24-point velocity surveys
with Fecheimer Probe downstream of mist
eliminator 127
87 Error in isokinetic sampling rate due to
assumption that droplet-laden stream is
saturated 128
x.i
-------
M^K^V-C, FIGURES (Cont.)
Numbers ' Page
88 Approximations to velocity components
measured in scrubber exhaust model . 131
89 Number of duct diameters downstream of
mist eliminator where water droplets
are impacted on duct wall, scrubber
exhaust model 133
90 Trajectories of 10 urn water droplets in
scrubber exhaust model 135
91 Standard sampling nozzle for use as
inertial droplet eliminator nozzle in
two-sample method 137
92 Standard nozzle with shield for use as
inertial droplet eliminator nozzle in
two-sample method 138
93 Inertial droplet eliminator nozzle with
suction slot for liquid film removal 139
94 Inertial droplet eliminator nozzle with
suction slot and aerodynamic tail 140
95 Inertial droplet eliminator nozzle with
suction slot oriented in upstream
direction 141
96 Standard sample probe collecting isokinetic
sample in scrubber exhaust model 143
97 Standard sample nozzle oriented in down-
stream direction as inertial droplet
separator in scrubber exhaust model 144
98 Standard sample nozzle with droplet shield
as inertial droplet separator in scrubber
exhaust model 145
99 Inertial droplet separator nozzle shown in
Figure 94 in scrubber exhaust model 146
100 Inertial droplet separator nozzle shown in
Figure 95 in scrubber exhaust model 147
101 Droplet size distribution at spray injec-
tion point 151
XII
-------
TABLES
Number Pac
^^^^^». **
1 Classification of Sources with Respect to
Emission Measurement Parameters 6
2 Partially Confined Emission Sources 8
3 Performance Table for Gravity Roof
Ventilators (Ref. 9) 11
4 Industry Reported Emission Data for
Primary Aluminum Facilities (Ref. 11) 14
5 Summary of Emission Data - Primary
Aluminum Industry (Ref. 11) 16
6 Emission Factors for a Vertical Stud
Soderberg Potline - EPA Test Results
(Ref. 11) 17
7 Summary of Grain Dryer Emissions at
Grain Elevators (Extracted from Ref. 6) 18
8 Unconfined Emission Sources 22
9 Emission Sources with Saturated Gas
Streams or Entrained Liquid Droplets 24
10 Average Velocity Determined from
Widthwise Velocity Surveys (based on
Figure 40) 69
11 Calculated Emissions for Example Problem .... 73
12 Particulate Concentration Measurements
from Preliminary Field Tests 107
13 Comparison of Measured Mass Flow Rate of
Water Vapor and Droplets with Actual
Mass Flow Rate 148
14 Performance of Inertial Droplet
Separator Nozzle 149
Kill
-------
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. Installation
costs for additional sample ports at the Seneca Wastewater
Treatment Plant were shared by the Metropolitan Waste Commis-
sion.
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, Mr. Frederick A. Bezat, Mr. Leon Zacho and the many
members of the staff of FluiDyne Engineering Corporation for
their parts in the accomplishment of this study.
xiv
-------
SECTION 1
INTRODUCTION
The capability to determine particulate emission levels
from stationary sources is of primary importance in programs
intended to control air pollution for two reasons. First,
reliable emission levels must be established in order to assess
the severity of a pollution problem and, if necessary, set
standards for its control. Second, in order to enforce compli-
ance with established standards, emissions from a given source
must be accurately determined. Procedures for selection of
sampling sites, velocity measurements, and determination 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 2 to 4 m/sec. Studies such
as Ref. 2 have examined the reliability of emission measurements
in stationary 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 stationary emission sources in
which the velocity, flow fields, and particulate concentrations
are not well defined or the effluent gases are not confined by
an accessible exhaust stack. Emission sources of this nature
are generally classified as "atypical". The intent of this
study is to examine the particulate emissions from stationary
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.
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.
Sampling techniques and strategies will be discussed for flow
and particulate concentration measurement related to several
specific atypical emission sources. The methods will be dis-
cussed with 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 dimension
source), grain dryers (partially unconfined flow) and wet scrub-
bers (entrained liquid droplets). Roof ventilators and grain
dryers were chosen as primarily representative of these parti-
cular characteristics for determination of sampling techniques,
although both roof ventilators and grain dryers may be classi-
fied as having characteristics under Items 1, 2 and 3.
Having identified these specific applications as represent-
ative of the various atypical source categories, the basic pro-
cedure 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 laboratory
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 appli-
cability of the sampling technique to actual field
situations;
5. Assessment of the applicability of the selected
sampling technique to other atypical sources.
-------
SECTION 2
CONCLUSIONS AND RECOMMENDATIONS
Conclusions and recommendations based on the program for
evaluation of techniques and instrumentation for particulate
emissions from atypical emission sources are presented in the
following.
ROOF VENTILATOR EMISSIONS
The recommended sampling technique utilizes a heated
thermopile anemometer and a high volume particulate matter
sampler deployed at sampling locations near the base of the roof
ventilator. Widthwise velocity surveys of on the order of 4 to
6 points at each sampling location were required to determine
the average velocity at any given time. The number of sampling
locations for a given length of roof ventilator must be deter-
mined for each application; an example based on field tests at a
secondary aluminum foundry showed that three sampling locations
were required for a 15.24 m ventilator. Engineering judgment
should be used to determine a balance between economic consider-
ations and the maximum amount of information obtainable in
selecting the number of sampling locations when making emission
measurements.
GRAIN DRYER EMISSIONS
The recommended sampling technique involves the use of the
same instrumentation as used for roof ventilator emissions, in
this case to be deployed at the surface of the exhaust louvers.
A balanced matrix sampling strategy should be adopted. The
total number of sampling points should be chosen to assure that
the entire traverse can be completed during a period of rela-
tively steady operation of the grain dryer.
WET SCRUBBER EMISSIONS
The recommended sampling technique includes velocity
surveys with a Fecheimer type probe, humidity measurement with
an inertial droplet separator type probe, and particulate
concentration measurement with a Method 5 type sampling train.
The number of sampling points should be determined from EPA
Method 1.
-------
GENERAL ATYPICAL EMISSION SOURCES
The basic characteristics of a general atypical emission
source should be assessed. The techniques discussed above
should then be selected according to the characteristics of the
emission source which categorize the source in one of the areas
considered in this study.
FURTHER STUDY
The limited scope of the present study did not allow
complete answers to some of the questions in the extremely broad
area of atypical emissions. Two areas toward which further
study should be directed are:
1. Collection and evaluation of more data on the required
number of sampling points for sources with extended
dimensions, and
2. development of the reliability of inertial droplet
separators.
-------
SECTION 3
CLASSIFICATION AND EVALUATION OF
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, com-
mercial firms engaged in emission measurements, eight state
pollution agencies in eight different EPA regions, and EPA
personnel.
Table 1 includes a list of several atypical emission
sources and classifies them with respect to the various charac-
teristics. 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.
EMISSION SOURCES WITH LOW VELOCITY AND/OR EXTENDED DIMENSIONS
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. Par-
tially confined sources are often also characterized by one or
more extended 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.
-------
TABLE 1. CLASSIFICATION OF SOURCES WITH RESPECT TO
EMISSION MEASUREMENT PARAMETERS
CTi
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
Dimensions
x
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Confinement
Duct Partial None
x
x
: x
X
X
X
X
X
X
X
X
X
X
X
X
X
Saturated Liquid
Gas Droplets
Weather
Sensitive
x
X
X
X
x
x
x
X
X
X
X
X
X
X
X
-------
TABLE 1 (CONT'D)
Source
Low
Velocity
Extended
Dimensions
Confinement
Duct Partial None
Saturated
Gas
Liquid
Droplets
Weather
Sensitive
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
X
X
X
X
-------
TABLE 2. PARTIALLY CONFINED EMISSION SOURCES
CO
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
2.3-7
9-80
0.25
Particulate*
Size
Emission Rate
Factor
gnt part, per
Kg of product
Range down
to submicron
10% < 5 |lm
Geo, means
Oats 3.1 y m
Wheat 2.1ym
72
42
32
0.9
3.4
0.2-5
2.3-25
50% < 5 l-tm
45% < 5 nm
60% < 5 urn
60% < 5 urn
4
11
5
5
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
Vi<-v^ incinerators
0.39
24% < 2
m
120,000
>i.-e distribution information based on % by weight less than stated size.
.':-.c iigures tabulated represent total emission quantities through all controls and exhausts for the
. .. rivicies listed and are included in order to establish their relative importance.
-------
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 j,n
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 "contin-
uous roof ridge ventilator", while the two-channel exhaust style
(Fig. 2) is usually denoted by the terms "monitor" 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 prefabricated
ventilators, continuous roof ridge ventilators and roof moni-
tors, are commercially available in 10 ft. lengths which may be
spliced together to form runs of any desired length. The con-
tinuous 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 cur-
rents and may be emitted through the roof ventilator, which may
often be considered a line source due to the large length
dimension involved. Particulate emissions from roof ventilators
typically consist of fine particles which closely follow the gas
streamlines. This is due in part to the relatively small par-
ticulate emitted from typical industrial situations in which
roof ventilators find application, as seen in Table 2, and in
part to the fact that larger particulate tends to settle out
before reaching the roof line or to be collected by primary
control devices, when used. The volume of emissions is de-
pendent 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
illustrates 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 per-
pendicular to the ventilator length, based on the negative
pressure which develops as the flow passes over the ventilator
-------
Figure 1. Continuous roof ridge ventilator
Figure 2. Roof "monitor" or "monitor attachment"
10
-------
TABLE 3. PERFORMANCE TABLE FOR GRAVITY
ROOF VENTILATORS (Ref. 9)
C.F.M. Per Square Feet of
Temp.
Diff .
10°F
20°F
30°F
Height
Above
Air Inlet
10 ft.
20
30
40
50
10 ft.
20
30
40
50
10 ft.
20
30
40
50
2 MPH
193
236
271
298
323
236
298
345
385
414
271
345
403
451
494
w ^^ ft * ^f **m «*ta *** *^ *^ *** * * * *"p* -*^ »« *^ ^^
At Wind Velocities
4 MPH 8 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
of:
10 MPH
457
500
535
562
587
500
562
609
642
678
535
609
667
715
758
1 foot = .3048 meters
11
-------
and creates an aspirating effect. Winds directed along the
length of a ventilator 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 parti-
culate 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
emission sources 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 alumi-
num process facilities is quite high and secondary collection 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 indi-
cates similar conclusions.
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, al-
though many other grains may require drying under certain con-
ditions. Two common types of grain dryers, rack and column
dryers, are illustrated in Fig. 3. Heated air or combustion
products are 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
12
-------
(a) Rack Dryer
Air
Sea
Dryer Section
MOISTURE
LADEN
AIR
OUT
^
1
J
Cooler
Section
,
«*M
*
*
^
^
.*.
^M.
.«_
V.
RS
fi
^
J«
j
*
u
|:
il
!
1*
L.
^
Grain
teceivin
Garner
/»s
5^ >ii
~» <~-
~* ~
>. ^~
^
I ( (
Chamber
' L (
~ » ^
^ «»^HB
V^ «A^^
^ 1
1
1
1
g
h
N
ri
ifi
«f"
ta«
X
t"
**
*
u
fe
1
*-
IH*.
_».
_»
i"^**
>
-»»
J)
Figure 3. Rack and column type grain dryers (from Ref. 6)
Variable Speed Discharge
(b) Column Dryer
-------
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
Overall control efficier
93
93
93
93
93
93
icy
52.0
42.0
10.4
10.1
10.4
20.5
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
14
-------
TABLE 4 CONT'D
% Capacity Kg Particulate/Metric Ton Aluminum
Reporting high averaae low
All types
Total effluent
Primary collection
Secondary collection
Primary emission
Secondary emission
Total emission
Overall control efficiency
average
63
82
71
82
71
71
y
88.6
84.5
7.8
24.4
7.8
23.5
47.7
40.3
6.9
5.9
6.4
12.3
73%
22.2
16.4
2.0
1.1
2.0
3.5
15
-------
TABLE 5. SUMMARY OF EMISSION DATA -
PRIMARY ALUMINUM INDUSTRY
(Ref. 11)
Emission Factor
kg Particulate/Metric Ton Aluminum
Vertical Horizontal
Prebake Stiid. Stud All Types
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%
16
-------
TABLE 6. EMISSION FACTORS FOR A VERTICAL
STUD SODERBERG POTLINE
EPA TEST RESULTS (Ref. 11)
Primary collection
Secondary collection
Primary emission
Secondary emission
Total emission
Primary efficiency
Secondary efficiency
Overall efficiency
Emission Factor
kg Particulate/Metric Ton Aluminum
(1)
45.63
13.56
0-06
4.77 ..
4.83
99.86%
64.85%
91.85%
(2)
33.80
13.34
0.06
2.92
2.97
99-84%
78.15%
93.70%
(1) Average of 3 tests.
(2) Average of 2 tests; 1 test deleted due
to stud blow during test.
17
-------
TABLE 7. SUMMARY OF GRAIN DRYER EMISSIONS AT
ELEVATORS (Extracted from Ref. 6)
Total Emissions (1971)
(metric tons/yr)
Country Elevators 2.14 x 104
Terminal Elevators 2.53 x 103
Export Elevators 3.82 x 102
18
-------
in Ref. 8 as having higher total annual emissions than grain
dryers. However, grain dryers present unique problems in emis-
sion 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 intensifies
the sampling and collection problems associated with the parti-
ally unconfined nature of grain dryers as an emission source.
The emission level is affected by a number of factors, 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 eleva-
tors is the deep receiving hopper. Grain is dropped from box-
cars 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 dissi-
pated 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 envelop an entire
boxcar. Sampling of such sources is usually accomplished by
assessment of the ambient conditions.
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 improve-
ments in combustion efficiency.
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 non-
compliance with air quality standards.
19
-------
Figure 4. Unloading grain from box-car
into a deep hopper (Ref. 12)
?lndkot«i fugitive dutl !«««*
If Ml trail
-------
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.
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
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.
Agricultural Burning
Smoke from agricultural field burning is composed primarily
of carbon particles, ash, and various gases. These emissions
cause odors and reduction of visibility. Emission plumes consist
mostly of particles less than 1.3 microns in diameter, identi-
fied as medium-and large-molecule hydrocarbons which provide a
serious potential for lung irritation (Ref. 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 t,imber cutting, has been
estimated by the United States Department of Agriculture to
produce 5.5 million metric tons of particulate matter each year.
Emission Measurements and Standards
A number of studies have been made concerning the prediction
of plume heights and concentrations related to unconfined sources
using various dispersion models. Actual measurement of emis-
sions 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
and 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
21
-------
TABLE 8. UNCONFINED EMISSION SOURCES
ro
Average Emission Rate
Particulate Factor Estimated Total
Concentration Particle gm Particulate Annual Emissions
gm/m
Size per kg Product metric tons
Source
Crushing
a) Stone
b) Sand & Gravel
Asphalt Plants
Agricultural Field
Burning
Cement Plants
Cotton Gins
Aggregate Stockpiles
* Typical mass median diameter
** Stockpile losses due to wind erosion
*** Size distribution information based on % by weight less than stated size
13-87
-
.01-. 06
2-35
.004-1.25
-
10-50ym*
3.5-9.4ym*
40% lOyjn***
0.5]Jm*
30% 5ym***
-
-
8
0.05
4
8.5
26
5kg/bale
0.0-0.5%**
4,150,000
42,000
37,000
2,200,000
172,000
41,000
6,000,000
-------
isokinetic sampling of dust with specially designed 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 differ-
ences 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 generally
handled on a complaint basis.
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. 3, 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 ex-
haust gases. Droplets formed by tearing of liquid sheets and
ligaments and by splashing of liquid drops are generally 200 ym
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.
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 ym 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 tri-
oxide and a larger quantity of acid mist are emitted to the
23
-------
to
TABLE 9. EMISSION SOURCES WITH SATURATED GAS STREAMS OR
ENTRAINED LIQUID DROPLETS
Source
Sulfuric acid plants
Chamber process
Contact process"
Spent acid concen-
trators
Asphalt Roofing Mfg.
Blowers s Saturators
Plywood Veneer Dryers
Wet Scrubber
Downstream of Thermal
Coal Dryer
Sewage Sludge Inciner-
ators
Emission Rate
Flow Rate
Nm . .
/mm
55-370
140-1750
1700
280-560
_
Avg. Mass
Concentrations
gm/m
0.2-1.2
0.04-1.75
2-4
1.0-1.8
«
Factor
Particle gm part per
Size kg of Product
10% < 3ym** 2,5
64% < 3ym** I
15
^lym 2
3Kg/1000 m2
Estimated Total
Annual Emissions
Metric Tons
1,800
3,600
7,200
15,500
3,600
600-5500 .070*
94,000*
850
0.3*
* Estimate includes solid particulate.
** Size distribution information based on % by weight less than stated size.
-------
atmosphere. The presence of a substantial number of particles
smaller than 3 ym 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 under certain plant opera- <
tions when the proportion of small particles is sharply in-
creased. This occurs, for example, during production of oleum,
a solution of free, uncombined sulfur trioxide. Tall stacks, on
the order of 50 m or larger, have been shown to be effective in
reducing acid spray emissions, since large particles 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.
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 forma-
tion 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 ym. Various hood and
exhaust configurations have been devised for exhausting blower
and saturator emissions which require volumetric capacities of
280-560 NmVmin.
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
25
-------
wood sugars which form a blue haze when cooled. Volatile com-
ponents 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 1 cm
(or 3/8 in) plywood produced. Condensible compounds are esti-
mated by Ref. 8 to contribute 63% of the emitted pollutant,
while volatile compounds make up 37%.
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
operate with low pressure drop may be classified as low energy
scrubbers. Efficiences can exceed 90-95%. Air pollution con-
trol 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 applica-
tions 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 encount-
ered in scrubber exhaust gases.
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 pro-
gram to these specific emission sources, a reasonable amount of
useful information about each of the selected sources was ob-
tained within the alloted limits of the program. The selected
sources are sufficiently representative of the atypical source
categories of interest; therefore, the test program provided
insight into sampling techniques for general emission sources
possessing the various atypical source characteristics. The
selected sources are discussed briefly in the following.
Roof Ventilators
Roof ventilators were selected as a representative example
26
-------
of two of the atypical source categories, emission sources
characterized by (1) low velocity and (2) 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).
Grain Dryers
Grain dryers were selected to represent the category of
partially or totally unconfined flow. The choice of a partially
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.
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.
27
-------
SECTION 4
ROOF VENTILATOR SAMPLING TECHNIQUES
Continuous gravity roof ventilators, in typical appli-
cations, are primarily characterized by low velocity and
extended dimensions. These two characteristics make emission
measurement following EPA Method 5 sampling techniques inade-
quate. Under the present study, sampling methodology which has
been applied to roof ventilators was reviewed, and techniques
and instrumentation for measurement of roof ventilator emis-
sions were evaluated following a test program as outlined in
the Introduction.
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 instru-
mentation 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 parti-
culate collection. The most frequently used velocity instru-
ments 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,
fluctuating, often circulatory flow characteristic of roof
ventilators as well as the intermittent nature and low concen-
tration 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 measure-
ments were determined for a 60 m long roof ventilator both
during periods of charging and tapping of the steel furnaces
and during periods of so called "background" emission. Several
high volume samplers spaced at 8.5 m intervals along the venti-
lator length were operated simultaneously; during charging and
tapping operations, the samples were fixed at the midpoint of
28
-------
the monitor width due to the short duration of these operations,
while the samplers were traversed across the width during
background measurements. Velocity traverses were made indepen-
dently of particulate concentrations measurements during
charging and tapping operations, the total number of traverse
points being dictated by the time duration of the operations.
Velocity measurements during background periods consisted of 6-
point traverses at each of 7 locations along the length of the
roof ventilator.
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.) 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.
f
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 anem-
ometer, 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
cost 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
29
-------
Volt
Figure 6. Cross section of roof ventilator
showing test equipment (Ref. 29)
Recorder
Monitor top
A
4" U-tube
Electric-/ I I
anemometer
Sampling
platform
Figure 7. Sample system (Ref. 30)
30
-------
12M 12M
12M 12M
Figure 8. Emission rate measurements, central
sampling station (Ref. 30)
Figure 9. EPA Method 14 sampling system
31
-------
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
measured 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.
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 environment. One anemometer is required
for each 85 m 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 trav-
erse. 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 testing firm contacts.
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 ft ) length of a configura-
tion known as the Swartwout (42 inch) Heat Valve. The ventilator
height is approximately 2 m; the base width is roughly 2 m and
the exhaust plane at the top is about 1.5m wide. The ventila-
tors shown in Fig. 10 are located above a room containing six
32
-------
Figure 10. Roof ventilators at Hitchcock Industries,
Bloomington, Minnesota, viewed from the
northwest
Figure 11.
Typical roof ventilator exhaust at Hitchcock
Industries site
33
-------
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 low concentrations expected 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-
cated by FluiDyne. The sampling probe is used with standard
Gelman Type A glass fiber filters. The entire assembly is
illustrated in Fig. 12.3 This sampling probe has a flow rate
capacity as high as 2 m /min (70 ft /min).
Two velocity instruments were selected for the preliminary
field tests, a Flowrite Model MRF vane anemometer and a Thermo-
Systems, Inc. Model 1610 hot film anemometer. The hot film
anemometer was a special high-temperature (160°C) version of
the standard hot wire anemometer Model 1610. The hot wire
anemometer was fitted with a protective collar designed and
built by FluiDyne (Fig. 13) for a previous project in which
velocities in the range of 0.03 -4.57 m/sec (0.1 - 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 indi-
cated 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.
34
-------
Figure 12. High volume sampler and probe assembly
35
-------
Probe Extension 1.53 ra (5 ft.)
S
TOP VIEW
Thermocouple
mounted here
Temperature Compensation
Electronics Inside \
Housing
L
SIDE VIEW
Flow
Direction
--. _.--.
1
J
)
T
} .
*
f
t
Hot Wire _J
Anemometer
Sensor Head
Figure 13. Hot wire anemometer and protective collar assembly
36
-------
Sample Planes
0.534 m
End View Cut Away
Station-
5
Station
4 "
Station
3 ~
Station..
2
Station.
1
2.44 m
Figure 14. Sampling locations for preliminary field
tests - Hitchcock Industries roof venti-
lator
3.05 m
3.05 m
3.05 m
3.05 m
1.52 m
N
37
-------
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, the Hitchcock Industries site provided a
good example of the type of problems to be encountered in
studying roof ventilator emissions.
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 150% to 230% of the iso-
kinetic rate.
Particulate concentration measurements obtained during the
preliminary field tests are shown in Fig. 17. Since the
measurements 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
preliminary 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 condi-
tions 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 preliminary field tests were obtained with the vane ane-
mometer.
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,
38
-------
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)
39
-------
East Exhaust Plane
Particulate
Concentration
3
(mg/m )
0.5-
0.4-
0.3-
0.2-
0.1'
A
^
D
A
O
1 1 1 1 1 1 1 1-
2 4 6 8 10 12 14
Roof Ventilator Length (m)
16
Particulate
Concentration
(mg/ra )
0.5T
0.4
0.3-.
0.2-
0.1
West Exhaust Plane
D
O
-f-
4-
4-
-4-
-4-
2 4 6 8 10 12 14
Roof Ventilator Length (m)
16
Figure 17. Preliminary field test concentration measurements
Hitchcock Industries roof ventilator
40
-------
Distance from East Edge (in)
0 O.L 0.2 0.3 0.4 0.5
Distance from West Edge (m)
0.5 0-4 0.3 0.2 O.L 0
u
Ci
0
0
J -
2 '
1 -
0 -
3 -
2 .
1 -
0_
3 -
2
1 -
Oj
3 -
2 -
1 -
o -
3 -
2 -
1-
n -
0
o
^
A
A
A
O
o
/\
^
D
C
B.
O
o
- .£X
Station
5
Station
4
Station
3
Station
2
Station
1
O
0
r^
- J
-2
-1
n
o
A
A
A
0
0
0
D
D
-3
2
.1
-3
-2
.1
..
-3
"2
-1
r%
B, .
u
O
o
0
-3
-2
-1
n
Figure 18.
Preliminary field test velocity survey I,
cock Industries roof ventilator
41
-------
Distance From East Edge (m)
0 0.1 0.2 0.3 0.4 0.5
Distance From West Edge (m)
0.5 0.4 0.3 0.2 0.1 0
y-N
O
03
&
5
i-t
u
o
0)
J -
2 -
1 -
n
3 -
2 -
1 -
0
3 -
2 -
1 -
0 -
3 -
2 -
1 "
o -
3 -
2-
1 -
o -I
O
O
(**
A
4
A
O
0 0
V
a
D
n
_j
/^
VJ
o
S-\
^
Station
5
Station
4
Station
3
Station
Station
1
H-,_ -o-'-i-i.-a-i-ii.--.. UIH-'-UL-U- , 1, ' '" ' "~
0
o
f*.
^
A
A
.A
W..B.IIIt
O
o
XV
-2
1
-0
-3
-2
-1
n
-3
-2
"1
n
v -
D
D
e.
3.
"2
-1
o
0
o
3
£
-1
8,
- -o
Figure 19
Preliminary field test velocity survey 2
Hitchcock Industries roof ventilator
42
-------
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 distribu-
tion inside the furnace room, which depends on the particular
activities 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 character-
istics of the roof ventilator exhaust.
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.
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
emission measurement.
Model Design and Fabrication
Manufacturer's literature (Ref. 34), and dimensional
measurements 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 existing duct network, used in a previous FluiDyne study
at the firm1s 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 schematically 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. Heated air then
caused evaporation of the water, leaving small particles of
salt suspended in the flow stream. The particulate produced by
the aerosol generator was estimated to have a mass median
diameter in the range of about 1 to 4 ym, varying over this
range as the particulate injection mass flow rate was varied to
produce different concentration levels in the roof ventilator
model. This aerosol generator provided a good simulation of
typical particles emitted from a roof ventilator.
43
-------
Figure 20. Roof ventilator model
44
-------
Roof
Ventilator Model
Roof Model
Test Section
Test Stand
en
Up to 9000 CPM
with Heat
Inj ection
*-
Atomization
Aerosol
Generator
_ r _
T
Pressurized Air
for Atomizer
u I u
o
TT
Liquid Droplet
Test Ports
-*Test Stand
FIGURE 21. Roof ventilator model test facility -
Rosemount Energy Conversion Laboratory
Early
Electrostatic
Precipitator
Model
\
Sample
Ports
-------
Figure 22. Roof ventilator model test section
46
-------
Pressure
Regulator
Pressure A"">
Gauge *
100 psig
Liquid
Flow
From
Reservoir
Adjustable
Heated Air
Injection
Flow
Figure 23. Atomization aerosol generator and hot air injection system
Rosemount Energy Conversion Laboratory
-------
Unfortunately, during the course of the roof ventilator
model test program, FluiDyne's Energy Conversion Laboratory
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 Laboratory. This
model, illustrated in Figs. 24 and 25, followed essentially the
same design as the original roof ventilator model. The aerosol
generation system for the second roof ventilator model, illus-
trated in Fig. 26, also followed the same concept and produced
similarly sized particles as the original model facility.
Four sample ports were located at the base of the second
model (Fig. 27) to allow sampling in a plane at the base of the
ventilator as well as in the ventilator exhaust. Testing on
both roof ventilator models included flow visualization studies,
velocity measurements, and particulate concentration measure-
ments .
Flow and Velocity Studies
The first objective of the model testing program was to
evaluate the capability to determine total volumetric flow
through the roof ventilator model. Early observations indi-
cated that the exhaust from the model often exhibited very
complex flow patterns which could not be readily determined
with conventional velocity instrumentation. Therefore, flow
visualization studies were conducted using lightweight flow
direction tufts as well as smoke injection 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 provided
a better sampling location. The sample ports shown in Fig. 27
were used to study the flow behavior near the base of the
ventilator 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 determined that this circulatory flow was caused by the
rapid expansion of the test section which supplied the flow to
the model (see Figs. 24 and 25) and that insertion of a flow
straightening baffle upstream of the model eliminated this
circulatory flow near the base of the ventilator. A typical
flow profile observed with this flow straightening baffle in
48
-------
Figure 24. Roof model located at the Medicine Lake
Aerodynamic Test Facility - looking south
Figure 25.
Roof model and ducting located at the
Medicine Lake Aerodynamic Test Facility
- looking southwest
49
-------
100 PSIG Air
ui
o
Chromalox 6 Ch-3
3 KW Air Circulation
Heater
4"I.D. Pipe (Min. 3-7/8"I.D.)
11 GA.(.120) Min. Wall Thickness
Centered in 18" Duct
18" Ventilation
Ducting
(Existing)
Pipe
Supports
(l/4"xl" Strip)
1/4 JN-F4
Press
Gauge
3/4" Pipe "J-
Transition-
4"to3/4 NPT
Inspection
Plug & Coupling
(172"NFT)
1/4" Copper
IJL
Press
Gauge
Valve
Existing Valve
Rotameter
100 PSIG Air
8 Gallon
Liquid
Solution
Container
Blower
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
51
-------
Figure 28. Flow direction indicating tufts in roof
ventilator model exhaust
Figure 29.
Smoke generator flow indicator in roof ventila-
tor model exhaust
52
-------
Recirculating Flow
Flow Direction
Indicating Tufts
Exhaust Flow
Figure 30.
Schematic of observed flow
(end view of roof model)
53
-------
place is illustrated 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
instruments for measurement of the volumetric flow rate. A
large number of instruments were considered for this applica-
tion, 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 is 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 use-
fulness of this material with regard to the problems of atypical
emission sources. To summarize the results of the 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-circu-
latory flow pattern, and 2) when a secondary observation tool,
a lightweight flow direction indicator, was used in conjunction
with the instrument in a circulatory flow field with the baffle
absent. These results are indicated in Fig. 32. None of
the velocity instruments evaluated was capable of providing
accurate flow rate data when used at the exhaust plane of the
roof ventilator model.
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 introduce particulate matter 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 model
through the sample ports. The actual concentration was moni-
tored in the ductwork upstream of the roof ventilator model
using EPA Method 5 sampling techniques, collecting particulate
matter on 47 mm Gelman Type A glass fiber filters.
The results of these measurements are shown in Fig. 33.
The measurements at the exhaust plane consist of single point
measurements at various points, while those at the base plane
were taken at a single point near the center of the roof
54
-------
Exhaust
Sample Plane
Velocities in
parentheses
in meters/sec
ase
imple
Lane
© © 0---© © 0 - -
(0.71) (0.62) (0-44) (0.32) (0.23) (0-22)
Figure 31.
Typical flow velocities at various points
in roof ventilator model
55
-------
Actual Volumetric Flow
Rate (mVmin)
350 i
Flow Direction Indicator
Yes
No
Without Uniform;
r .LOW odJ.j.jLe
With Uniform
Flow Baffle
*y
II
W
a
100% Accuracy
X
+50%
50
100
150 200
250
300 350
Indicated Volumetric Flow
Rate
Figure 32. Volumetric flow rate at base of roof ven-
tilator model determined with Hastings-
Raydist PCI-30 hot thermopile anemometer
56
-------
20
10
8
6
4
I '
6
I
c 0.8
0)
o
S 0.6
0.4
0.2
0.1
-25% X,
D Sample Plane
at Exhaust
O Sample Plane
at Base
f Small Filter
Loadings
0.1 0:2 0.4 0.6 0.8 1 2
Measured Concentration
mg/m
8 10
20
Figure 33. Particulate concentration measurements
in roof ventilator model
57
-------
ventilator model. A great variety of atmospheric conditions
was encountered in the course of the particulate 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 atmospheric
conditions and degree of accuracy of the results. This obser-
vation is in agreement with Ref. 35, which includes data on the
effect of anisokinetic sampling on accuracy of particulate
concentration measurements. Using the correlation parameters
of Ref. 35 for the injected salt particles, the high volume
sampling probe, and the velocity of the model exhaust indicates
virtually no detectable error due to anisokinetic sampling over
a large range in sampling rates.
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, with the notable ex^ept-
ion of two points near an actual concentration of 0.2 mg/m .
The accuracy of these two measurements is doubtful, however,
since the tests were of insufficient duration and resulted in
extremely small loadings on the standard filters. In general,
the measurements at the base plane indicate a slightly higher
than actual concentration regardless of sample rate.
FINAL FIELD TESTS
In order to evaluate the usefulness of the sampling
technique 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 over a five day period. 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 sampling 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
samples (except one test which was concluded after only 45
58
-------
Figure 34.
Particulate concentration and velocity
measurements at base of Hitchcock
Industries roof ventilator
59
-------
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 particu-
late 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
each 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
supposed occasional circulatory flow based on visual observa-
tions 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 concentration measurements since the sample probe and
velocity probe could not be inserted through the sample ports
simultaneously.
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 defin-
itely 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 variations of concentration with position in the
60
-------
1.4-
1.2-
1.0-
Average _ _
Velocity
(m/sec)
0.6
0.4
0.2
o
0.6-
0.5-
Particulate _ .
0 A *"
Concentration
(mg/m3)
0.3-
0.2-
o.i-
o -
O Day i
Q Day 2
O Day 3
A Day 4 O
O Day 5
° ri
A aO_D _^^
/^^°i ^^^X
°A ^
D A
7 8 9 10 11 12 1 2 3 4 56
AM PM
9__,.
9/-\
© ' ^s dL.
1 1 1 1 1 1 1 1 1 1 1
78 9 10 11 12 1 2 345 6
AM
Time of Day
Figure 35. Average velocity and particulate concentration
measurements - Sample Port 1 - Hitchcock
Industries roof ventilation
61
-------
O Day 1
1.4 '
1.2 '
1.0 "
Average
Velocity 0.8 -
(m/sec)
0.6 -
0.4 -
0.2 -
0
/
D Day 2
O Day 3
O O A Day 4
O Day 5
°0 0 0
O 0 __^
cp -"^ ~~ -^^
A^£Q oA
A ^
A A
1 ) j 1 1 i i i i i r
' 8 9 10 11 12 1 2 3 45 6
AM PM
0.6 1
0.5 -
Particulate
Concentration 0.4 ~
(mg/m3)
0.3 ~
0.2 ~
0.1 -
n
w
r-, I H .
i B [ -
n
i T ^ r^ i r r~^ i r 1 ^nr
7 8 9 10 11 12 1 2 3 4 56
AM . PM
Figure 36.
Average velocity and particulate concen-
tration measurements - Sample Port 2
Hitchcock Industries roof ventilator
62
-------
1.4-
1.2-
Average
Velocity I.Q-
(m/sec)
0.8~
O
l
0.6"
0.4-
0.2-
o
O
J
£-£(
O
D
^TTT"
-4 1 *+**
D
O Day 1
D Day 2
O Day 3
A Day 4
O Day 5
A A A ^-^X
i |
789
0.6 -
0.5-
Particulate
Concent rat ion 0.4"
(mg/m^)
0.3 J
0.2*
o.i-
V
0 " '
789
A
i i
A
i
10 11 12
AM
'\^S
f
/\
\/
i i
10 11 12
AM
m *
\/
i i i i i i
1 23456
PM
Ot -
i
< 1 1 I I 1
123456
PM
Figure 37. Average velocity and particulate concentra-
tion measurements - Sample Port 3
Hitchcock Industries roof ventilator
63
-------
1.4-
1.2-
Average
Velocity 1.0~
(m/sec)
0.8-
0.6-
0.4-
0.2-
0 -
O Day 1
D Day 2
O Day 3
A Day 4
O Day 5
0 0 ° D
°\?J^x^ ^^X
o o &
D
i i ; i i i i f i i <
1 8 9 10 11 12 1 2 3 4 5 6
AM PM
0.6-
Particulate
Concentration
(mg/m3) 0.4"
0.3-
0.2-
o.i-
0 -
f
A
A j- i A i
* ' i ts-' i
7 8 9 10 11 12 1 2 3 45 6
AM * c ^ PM
Figure 38. Average velocity and particulate concentration
measurements - Sample Port 4
Hitchcock Industries roof ventilator
64
-------
Average
Velocity
(m/sec)
1.4-,
1.2~
1.0-
0.8-
0.6-
0.4~
0.2-
0 -
o Day i
D Day 2
O O Day 3
A Day 4
O Day 5
0 Q
DO AD ^
O
i i i i > r i i ill
1 8 9 10 11 12 1 2 3 45 6
AM PM
0.6
0.5
Particulate
Concentration
(mg/m3)
0.4
0.3~
0.2-
0.1
0_
7
^ n
O
w'
.
i
11 i i i i i i i i i
89 10 11 12 12 3 4 56
AM PM
TIME OF DAY
Figure 39. Average velocity and particulage concentration
measurements - Sample Port 5
Hitchcock Industries roof ventilator
65
-------
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 severalhours, 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 average temperature was on the order of 55°C.
EVALUATION OF SAMPLING TECHNIQUE
The information obtained during the roof ventilator test
program led to several conclusions concerning basic measurement
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 re-
asonable, estimates (typically within + 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 measurements 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
66
-------
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 particul-
ate concentration 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 determined for each individual case.
The long sampling time required in order to obtain a
weighable collection of particulate matter 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 practi-
cality. 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.
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 emphas-
ized 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 for total emission measurements, 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 inter-
val. 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 measure-
ments during a single six hour period.
67
-------
V(m/sec)
V(m/sec)
V(m/sec)
V(m/sec)
1.5 "
i.o -
0.5 -
0
1.5
1.0
0.5
0
1.5
1.0
0,5
0
1.5 "
i.o
0.5
0
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
1.5 4-
V(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 .
68
-------
TABLE 10. AVERAGE VELOCITY DETERMINED FROM WIDTHWISE VELOCITY
SURVEYS, BASED ON FIGURE 40
No. of Points in Survey
ON
vo
Sample
Station
1
2
3
4
5
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 .
(ra/sec)
.63
.55
,48
.79
.67
% Diff.
-2,3
-2.2
+8.0
-2.2
+19.2
ample
tation
1
2
3
4
5
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
-------
ParticulaCe
Concentration
(mg/m3)
Particulate
Concentration
(mg/m3)
0.5 t U) 11
0.4
0.3
-e-
-G-
-0-
A
-e-
0.2
0.1
0
0 2 4 68 10 12 14
Roof Ventilator Length (m)
0.5 -j- (b) 1 PM
0.4
0.3
0.2--
0.1 "
-B-
I A-
0 24 6 8 10 12 14
Roof Ventilator Length (m)
O.5.. (03 PM
Particulate
Concentration 0.4 i"
(mg/m3)
0.3
0.2
0.1
0
-3-
H »-
02 4 6 8 10 12 14
Roof Ventilator Length (m)
Figure 41. Particulate concentration profiles for roof
ventilator emission example problem
70
-------
Average
Velocity
(m/sec)
1.0-
0.8-
.6
0.4"
0.2'
n
- (a) 11 AM
m
"~L .
1
t
*
1 i 1 -4 1 1 J 1
2 46 8 10 12
Roof Ventilator Length (m)
Average
Velocity
(m/sec)
1.0"
0.8"
0.6"
0.4-
0.2"
n
(b) 1 PM
1
1
to
l»
Jt J 1 1 1 1 1
0 2 4 6 8 10 12
Roof Ventilator Length (m)
l.Ot (c) 3 PM
0.8"
Average
Velocity
(m/sec)
0.6
0.4"
0.2+
0
14
Pigure 42.
02 4 6 8 10 12 14
Roof Ventilator Length (m)
Average velocity profiles for roof ventilator
emission example problem
71
-------
The total emission during the six hour period was calcu-
lated following a number of simple methods. The basic emission
calculation procedure followed the formula:
N M
E =
Where: E = total emission (gm) 2
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.
The results of these example calculations are shown in
Table 11. It can be seen that consideration of the actual
variation 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 theoretically
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 ventilator, also had little effect
on the calculated total emission, resulting in a 5.8% differ-
ence. 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
indicates the need for no less than three velocity measure-
ment stations in the roof ventilator system covered by the
present velocity measurement specification of EPA Method 14,
which requires only a single measurement point for every 85 m
of length. The flow through aluminum potroom roof ventilators
may be more uniform than that encountered at the Hitchcock
Industries field test site, however.
To summarize, the sampling technique demonstrated through
the test program described above consists of velocity and
particulate concentrations at an unspecified number of sampling
stations near the base of the roof ventilator, using widthwise
72
-------
TABLE 11. CALCULATED EMISSIONS FOR EXAMPLE PROBLEM
10
Total Emission, E = £ Z (C . V.A.Jt.
. i=i j=i 3331
N = Number of Time Increments
M = Number of Sample Locations
N
3
3
Ci 3 Vi
(gm/m ) (m/sec)
5 Individual values Individual values
From Fig. 41a,b,c From Fig.42a,b,c
1 Average values Average values
Aj fci E % Difference
(m } (sec) (gm)
5.58 7200 79.8
27.9 7200 82,7 +3.6
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
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
27.9 21,600 82.3
9.3 7200
81.3
27.9 21,600 84.4
27.9
7200 133.4
27.9 21,600 139.2
27.9 21,600 49.2
+3.1
+1.9
+5.8
+67.2
+74.4
-38.3
-------
velocity surveys with heated thermopile anemometers to deter-
mine 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 simulta-
neous 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 con-
ducting the tests.
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
such as that listed in Table Al having sufficient
accuracy in the expected velocity range, yet suffi-
cient durability to withstand field test conditions.
2. Selection of a high volume particulate sampler to
ensure sufficient volume of gas sampled to obtain
useful results.
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 cases, 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 in
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, engineeri-
ng judgment must play a large role in determining the number of
sampling points needed.
74
-------
SECTION 5
GRAIN DRYER SAMPLING TECHNIQUES
Grain dryers are characterized by partially unconfined
effluent streams, typically emanating from one or more sections
of louvers which direct the flow for a short path length.
Grain dryer emissions may also be characterized by extended
dimensions, due to the large surface area of the louvered
sections, and by low velocity streams. In the present study,
grain dryer sampling methodology was reviewed, and a program
for evaluation of techniques and instrumentation for measure-
ment of grain dryer emissions was undertaken.
REVIEW OF SAMPLING METHODOLOGY
As in the case of roof ventilator emissions, contacts with
commercial firms indicated that standard techniques for grain
dryer emission measurements are not well, established. In
general, though, equipment specified for grain dryer applica-
tions have followed EPA Reference Methods or ASME Power Test
Codes (Ref. 36). The sampling equipment has been deployed in
various ways, generally employing some sort of collection stack
but often merely inserted in the effluent stream.
Meiering, et al (Refs. 37, 38) developed a particulate
collection box, shown schematically in Fig. 43, which could be
attached directly to the surface of a grain drying column. A
large number of these devices were used over extended periods
to study the emission of particulate matter from the actual
drying process in order to examine methods for reduction of
emissions. No measurements of particulate matter emitted from
the actual grain dryer structure into the atmosphere were
attempted, however.
Reference 39 reports on grain dryer particulate emissions
tests performed by a state regulatory agency. The test method-
ology adopted in this case was to attach a rectangular duct to
a portion of the exhaust louvers and use EPA Method 5 to sample
the stream flowing through this sampling duct. The duct ex-
tended 3.05 m (10 ft) from the surface of the louvers, and the
cross sectional area of the duct amounted to 10% of the total
louvered area. Only the louvers on one side of the grain dryer
were sampled, and the emission measurements were noticeably
affected by external wind conditions.
75
-------
Grain Column
Collected
Particulate
Drying
Air Flow
Figure 43. Particulate collection device used for
grain dryer emissions measurement
(Refs. 37, 38)
76
-------
Tests conducted by the EPA on a grain dryer with a screen
filter type emission control device are reported in Ref. 40.
In this instance, the exhaust louvers were removed from the
emission surface before the measurements were made. Sampling
points were established at the midpoints of 12 equal area
segments over the 2.44 m x 7.32 m exhaust surface. A Method 5
type sampling probe was inserted in a 0.46 m (1.5 ft) "stove-
pipe" flow straightener and traverses of the exhaust surface
were performed. Pitot tube and hot wire anemometer measure-
ments were made, but the results were considered unreliable,
and therefore manufacturers specifications rather than actual
velocity measurements were used to determine the sampling rate.
Tests with a high volume sampling train are also reported
in Reference 40; correlation between the high volume and Method
5 trains were generally poor in these grain dryer emission
tests, however.
The variety of approaches to grain dryer emission measur-
ement is indicative of the general lack of standard test
methodology. As in the case of roof ventilator emission
measurements, assessment of the accuracy of the various measure-
ment techniques reported here was not possible. Engineering
judgment as to the feasibility and accuracy of the various
methods was the basic criterion for selection of these tech-
niques.
PRELIMINARY INFORMATION
A potential field test site was located at a Minneapolis
area manufacturing plant, Honeymead Products Company. This
facility includes a column type grain dryer, Carter-Day model
HC-66, used primarily for drying sunflower seeds used in the
production of sunflower oil. Preliminary arrangements were
made to conduct a field test program at the Honeymead grain
dryer site. However, this field test program was not completed
due to inactivity of the grain dryer, first due to a labor
strike and then due to decreased demand for sunflower oil.
Several detailed inspections of the grain dryer, including the
exterior exhaust louvers as well as the interior flow passages,
fans, burners, grain columns, screens, etc., were made, how-
ever.
In lieu of a preliminary field test program, the grain
dryer manufacturer was contacted for information concerning
expected flow rates, particulate concentrations, and diffi-
culties to be encountered in grain dryer emission measurements.
Information was gathered on the Honeymead grain dryer as well
as several other models used in various parts of the country.
77
-------
MODEL STUDIES
The laboratory model studies program was developed on the
basis of the information presented above. A laboratory model
representative of the emission plane of a typical grain dryer
was designed and fabricated, and this model was used to evalu-
ate techniques and instrumentation for use in measurement of
particulate emissions from grain dryers.
Model Design and Fabrication
A model of the upper portion of a typical grain dryer was
fabricated and attached to the existing ductwork and diffuser
section constructed for the roof ventilator model studies
(Figure 44). The primary intent of the model design was to
simulate the interaction of the internal flow of a grain dryer
with the external atmospheric conditions. A schematic diagram
of the flow through the grain dryer model is shown in Figure
45. This flow passage simulates the flow through a typical
Carter-Day column type grain dryer.
Two sets of exhaust louvers were included in the grain
dryer model. One set of louvers is shown in Figs. 44 and 46,
the other being directly opposite this set on the other long
side of the grain dryer model. The total exhaust louver sur-
face area of the grain dryer model amounted to 1.24 m , or
about 6% of a typical column type grain dryer. The length and
width dimensions of the model were scaled approximately to
match the Honeymead dryer.
The same aerosol generation system used in the roof
ventilator tests, illustrated in Fig. 26, was used to introduce
particulate matter into the grain dryer model exhaust. This
model aerosol had a mass median diameter of approximately 2 ym.
Typical grain dryer particulate emissions are generally expected
to include much larger particles. However, the relatively high
density of the model salt particles as compared to typical
grain particles resulted in an aerodynamic particle size which
was a reasonable model of grain dryer particulate for the
purpose of the laboratory model testing.
Flow Velocity Studies
Based on the available information on flow rates through
various grain dryers, velocities in the range of 1-4 m/sec are
expected in grain dryer exhaust streams. Based on the exten-
sive evaluation of velocity instrumentation shown in the Appen-
dix, the instrument selected as having the greatest potential
for use in grain dryer sampling was the heated thermopile
anemometer. The Haystings Raydist PCI-30 anemometer, the same
instrument selected for roof ventilator testing, was chosen for
evaluation on the basis of an overall judgment of accuracy in
78
-------
Figure 44.
Grain dryer model constructed with existing air
supply at Medicine Lake Laboratory
79
-------
Exhaust
Louvers
Air Supply
From Plenum
Figure 45.
Cutaway of grain dryer model showing air flow
passage on one side of model
80
-------
Figure 46. Grain dryer model with exhaust louvers
-------
the expected velocity range, ease of handling, and ruggedness.
The basic technique chosen for volumetric flow measurement
was to place the anemometer directly at the outlet of the
exhaust louvers as shown in Fig. 47. This technique was
chosen in order to minimize potential interference with the
exhaust flow field which might be introduced by a collection
duct or hood and to eliminate any modifications of the dryer
itself, which might be required, such as removal of the louvers
as reported in Ref. 40. The sensing head of the anemometer was
aligned with the angle of the louvers, in this case 45°.
A number of velocity traverses were made with the heated
thermopile anemometer, employing two matrices of sampling
points as illustrated in Fig. 48. Atmospheric winds were also
measured during these tests and were found to vary from a
minimum of 0.15 m/sec to a maximum of 5.1 m/sec. The average
wind velocity was roughly 1.5 m/sec. The wind direction varied
from parallel to the exhaust through the louvers to perpendic-
ular to the exhaust. In order to simulate the result of an
obstruction to the exhuast flow path, such as proximity to a.
building, or an internal blockage within a grain dryer, approx-
imately 45% of the louver area was blocked off during several
of the velocity measurements.
Results of these tests are presented in Fig. 49, which
compares the volumetric flow rate determined from the velocity
traverses of the two exhaust louver faces with the actual flow
rate measured in the supply ductwork upstream of the grain
dryer model. As can be seen from this Figure, all data lies
within an accuracy range of +15% regardless of the wind veloci-
ty, wind direction, or degree of blockage of the exhaust
louvers. These results indicate the usefulness of the basic
measurement technique of simply deploying the anemometer at the
outlet of the louvers.
Particulate Concentration Studies
The sampling instrument chosen for particulate measurement
studies was the high volume sampler and probe developed for use
in roof ventilator particulate emission measurements. This
probe was deployed at the surface of the exhaust louvers (Fig.
50) for the same reasons as the velocity instrument, i.e., in
order to minimize flow disturbance and required modifications.
A number of concentration measurements were made by
traversing the surface of the exhaust louvers. Two sampling
matrices were employed, as shown in Fig. 51. Wind conditions
were comparable to those encountered during the velocity
testing. A constant sampling rate was used at all of the
sampling points on each of the two sets of exhaust louvers.
The selected rate was based on isokinetic sampling at the
82
-------
CO
OJ
2X9
3X4
Figure 47.
Heated thermopile
anemometer deployed
at surface of exhaust
louvers
Figure 48.
Sampling point matrices
used for velocity
traverses of grain
dryer model
-------
200
150 -
c
H
e
o
iH
fc
u
H
-P
0)
id
D
.p
100 -
2x9 Sample Matrix
3x4 Sample Matrix
45% of Louver Area
Blocked
2x9 Sample Matrix
50
100
150
200
Measured Volumetric Flow Rate (m /min)
Figure 4 9. Volumetric flow rate determined with heated
thermopile anemometer at exhaust louvers of
grain dryer model
84
-------
00
U1
Figure 50. High volume sampler
deployed at surface of
exhaust louvers for
particulate traverses
2X3
3X3
Figure 51.
Sampling point matrices
used for particulate
traverses of grain
dryer model
-------
average velocity determined from velocity surveys with the
heated thermopile anemometer, rather than attempting to attain
isokinetic sampling at each individual point.
The results of the particulate concentration measurements
are shown in Fig. 52. The actual concentration was determined
from concentration traverses in the ducting upstream of the
grain dryer model. The results indicate that the measured
concentration was generally within + 25% of the actual.
Measurements made with a 3 x 3 sampling matrix, or a "balanced"
matrix as proposed for rectangular ducts in Ref. 2, appeared to
be more reliable than those made with a 2 x 3 matrix. Again,
the results indicate that simply deploying the sampling probe
at the surface of the exhaust louvers provides reasonable
results.
EVALUATION OF SAMPLING TECHNIQUE
The basic sampling technique developed for grain dryer
emission measurements involves the use of a heated thermopile
anemometer for velocity measurements and a high volume particu-
late probe for particulate concentration measurements, both of
these instruments to be traversed across the surface of the
exhaust louvers. Based on the limited data presented here, a
balanced matrix scheme of sampling points is recommended. The
total number of sampling points should be chosen such that the
total time required for traversing the louver surface is com-
patible with a period of relatively steady operation of the
grain dryer.
The principal problem in employing this method on full
scale grain dryers is gaining access to the exhaust louvers,
which are usually situated a great distance above the ground.
Therefore, some support scaffolding will generally be required.
An alternative to the use of support scaffolding, however,
might be to situate the basic collection mechanism on the roof
of the grain dryer and use a flexible sampling probe to tra-
verse the exhaust louvers. A long handled velocity probe with
the sensing head inclined at the appropriate angle of the
exhaust louvers could be employed in a similar manner. This
concept is illustrated in Fig. 53.
APPLICABILITY TO OTHER EMISSION SOURCES
The sampling technique developed for this example of a
partially unconfined particulate emission source should be
applicable to other partially unconfined sources for which a
suitable choice of a measurement plane can be made. In the
case of grain dryer emissions, the louvers confine the flow for
a sufficient length so as to provide a flow field which is
relatively unaffected by the external conditions near the
exhaust surface. In the case of roof ventilators, in contrast,
86
-------
14
12
10
0
H
4J
2
4J
C
0)
U
8
-------
Filter
Holder
Pump,
Condensejfr
Etc.
Flexible
Probe
Figure 53. Concept for particulate sampling of grain dryers
88
-------
the low velocity, circulatory flow field emanating from the
exhaust surface was shown to be unsuitable for emission measure-
ments due to the severe interaction with external conditions.
It is evident, therefore, that emission measurements can
be made externally only if the partial confinement results in
a suitable flow field. For partially unconfined emission
sources which can not readily be classified in this matter by
engineering judgment, some simple flow visualizaiton studies
may provide the means to decide if measurements can be made
externally, as with grain dryers, or if an internal site must
be identified, as with roof ventilators.
This method is very likely not applicable for totally
unconfined emission sources. Particulate emissions from this
type of emission source should probably be measured with a
sampler such as the "exposure profiler" developed by Cowherd,
et al (Refs. 15, 41) or the fugitive emissions sampler reported
by Jorgenson, et al (Ref. 42).
89
-------
SECTION 6
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
associated 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
methodology review, field testing and model testing, following
the same basic approach taken for the study of roof ventilator
and grain dryer sampling techniques, was undertaken.
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. 43.
Several aspects which must be considered in scrubber
emission measurements are specified in Refs. 44 and 45. When
sampling a wet scrubber system, particulate concentration
should be analyzed on a dry gas basis. 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 depending on the particular en-
trainment separator used.
Scrubber sampling methods have usually specified that the
sample probe be heated (Refs. 16,22,43). 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 exhaust gas dewpoint, but not so
high as to vaporize various liquid pollutants of interest. As
in the case of roof ventilator and grain dryers sampling,
engineering judgment plays a large role in the selection'of
sampling equipment and techniques (Refs. 17, 44).
90
-------
A great deal of information concerning sampling techniques
and problems was obtained through contacts with twelve manu-
facturers of wet scrubber systems. Since scrubber manufacturers
often tend to specialize in wet scrubber systems for a particu-
lar industry, the reported sampling techniques varied according
to the type of pollutants encountered in these various indus-
tries. The most common practice was found to be the use of the
basic Method 5 sampling train with modifications required to
overcome specific problems encountered. Typical modifications
included 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. 45, or inertial impaction separators. One type of
precutter is illustrated in Fig. 54. Occasionally, totally
different sampling techniques have been employed in particu-
larly troublesome situations.
Other difficulties discussed in these manufacturer con-
tacts included the variation of particulate concentrations in
scrubber system ductwork due to gravitational settling and wall
impingement and problems associated with sustained high tempera-
ture of heated sampling probes. In certain cases, the probe
temperature of 121°C (250°F) specified by Method 5 was sus-
pected 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 veloci*-
ties. One manufacturer 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
presence 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. 55 showing
the normal method for finding this axial component, which
91
-------
Top
View
Probe Inlet
to
Side
View ,,
Water Trap
S-tube
L~
^> -p~~-""; -
Figure 54. Inertial impaction liquid droplet separator used by
a wet scrubber manufacturer's test group
-------
S-tube oriented
along this line
V = Actual Duct Velocity
Vz = Axial Velocity
Component
Vm = Tangential Velocity
Component
Q = Angle between Axial
Direction and the
Flow Direction at
the Sa'mple Point
V,, = V cos 0
Figure 55.
Method used to determine axial component
in a single vortex cyclonic flow
93
-------
requires the ability to measure the yaw angle. If an S-type
pitot tube were aligned in the axial direction and used to
attempt direct measurement of the axial velocity component, the
measurement would be in error since the output of a pitot tube
does not correspond to the cosine of the yaw angle, as indi-
cated by the data shown in Fig. 56 (extracted from Refs. 2 and
46) .
Several studies have been made concerning sampling within
vortices or cyclonic flow regions (e.g., Refs. 47-50). It has
been observed that in a flow pattern consisting of a vortex
motion superimposed on an axial motion, i.e. a single vortex
cyclonic flow, the axial velocity profile varies considerably
as the swirling velocity increases. The axial velocity com-
ponent 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. 47). 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. 57. 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. 58. The static pressure angle is quite uniform
at 39.25° for values of the Reynolds number in the range 10,000
- 200,000. By arranging static pressure taps at the two loca-
tions and a total pressure tap as indicated in Fig. 59 (from
Ref. 52), 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. 60 and 61, extracted from Refs. 53 and
54.
Other directional sensing velocity instruments appearing
in the literature include spherical head direction pitot tubes
(similar to Fig. 57), 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 Fech-
eimer probe or its variations.
As in the case of other atypical particulate emission
sources, a review of sampling methodology revealed the lack
of an accepted standard sampling technique for emissions from
wet scrubbers. The primary problems cited in relation to wet
scrubber sampling were related to:
94
-------
40 -20
6 f degrees
+20 +40
6, degrees
Velocity
Error
O- 4.57 m/sec (15 ft/sec) - Model Test Section
O- 15.24 m/sec (50 ft/sec)- Model Test Section
A- 9.14 m/sec (30 ft/sec) - 20.3 cm Dia.- Free Jet
Figure 56. Velocity error with yaw angle (3/8" S-Tube)
-------
rotation
Cf\ _,._. _, 3Z.3H
(d)
' ^~ " -
(e)
(b)
(c)
(g)
Figure 57. Several types of directional pitot tubes
(Ref. 51)
96
-------
Stotie pressure
angle
Flow,
-------
FASTENING SCREWS (4)
MftCNCMUlC
PANEL BOAR
- PI^E couniNG (AoomenAi pi»e AS
ACOUinCOIDOftTM
MCCESSAAV P^OtE Ij)
BASiC SAMPLER
PERFOHATE2
eMAWITC CASKCT
T%PE*V FILTER PAPE4
~^*W /
L. FECHEIMEH INLET
Ikl «T NOZ7LE Jk 'WTS / ^- tOCATlSa STUDS 121
WLfefNOza* _ '-OSANITEOASKET WWOBATCO CVUNWH
Figure 60. Fecheimer Probe built into filter holder
(Ref. 53)
98
-------
Directional Pressure Taps
Total Pressure Tap
V *
/_ » _
*
>
1 -
\
Probe Tip
Static Pressure Tap
Figure 61. Connecticut State Department of Environmental
Protection probe (Ref. 54)
99
-------
1. Properly accounting for the presence of cyclonic
flow, and
2. collecting and accounting for the liquid present in
droplet form.
Development of techniques and instrumentation which address
these two problems was the basic goal of the field testing and
laboratory model testing programs outlined in the following.
PRELIMINARY FIELD TESTS
The site chosen for initial field testing was the Seneca
Wastewater Treatment Plant in Eagan, Minnesota. A twin incin-
erator system is used at this plant to burn sewage sludge
collected through water treatment processes. Two Peabody
impingement tray scrubbers are used for particulate emission
control from the sludge incinerators. Each scrubber has a
capacity of 340 m /min (12,000 ft /min) and is equipped with a
fixed vane type centrifugal mist eliminator at the exhaust end.
A diagram of the wet scrubber system and the exhaust ducting is
shown in Fig. 62. Impingement tray type scrubbers of this size
are very common in industry, particularly in particulate collect-
ion from incinerators.
A total of 5 days were spent in performing the preliminary
field tests. Velocity and particulate concentration measure-
ments were made through existing sample ports located at points
B and C in Fig. 62 as well as in additional sample ports which!
were added at point A.
Velocity Measurements - Cyclonic Flow
Flow angularity was measured with a Fecheimer probe, while
actual velocity measurements were made with an S-tube. The
procedure used for velocity measurements was to first conduct a
traverse of the duct with the Fecheimer probe to determine the
flow angle at each sample point and then traverse the duct with
the S-tube, orienting the probe according to the angle deter-
mined from the Fecheimer probe traverse. A flow direction tuft
was also used to visually check the yaw angle measured with the
Fecheimer probe. In all cases the angle observed with the flow
direction tuft confirmed the Fecheimer indications.
Several velocity traverses were performed in this manner
at each of the three sampling locations indicated in Fig. 62.
Pronounced cyclonic flow was observed at point A, just down-
stream of the scrubber exhaust. Yaw angles as large as 75° and
swirl velocities as high as 20 m/sec were observed at this
location. Typical profiles of the velocity and the axial
velocity component measured at point A are shown in Fig. 63.
The degree of cyclonic flow was somewhat reduced at point B,
100
-------
Flow Direction
Indicated by
Arrows
FRONT VIEW
\
\ /
Wet Scrubber
(Extends down
two floors)
SIDE VIEW
\
\ \
Figure 62. Wet scrubber and exhaust ducts at the Seneca Wastewater
Treatment Plant - Eagan, Minnesota
-------
Swirl Velocity (v)
o
n
(D
-U
C
0)
u
o
o
C
n)
4J
(0
-H
Q
a
a>
N
o
z
Axial Velocity (y )
z
1.00-
0.80 -|
0.60 H
0.40 H
0.20 H
Indeterminant
near the center due
to turbulence
10 15 20 0
Velocity (in/sec)
10
15
20
Figure 63,
Typical velocity profile obtained during pre-
liminary field tests at Seneca Wastewater
Treatment Plant
102
-------
the maximum yaw angle and velocity being approximately 60° and
17 m/sec. The flow downstream of the ID fan, at point C in
Fig. 62, was nearly uniform in direction. When the axial
component of velocity was used to compute the volumetric flow
rate through the scrubber exhaust ducting, the measured flow
rates at points A and C agreed to within 6%, indicating the
basic validity of this method of velocity determination for
cyclonic flow of the type observed in this field application.
Particulate Sampling - Liquid Droplets
The sampling equipment assembled for the preliminary field
tests included the basic components of a standard EPA Method 5
sampling train: a glass lined, heated probe and attached S-
tube, a filter holder, two wet and two dry impingers, a conden-
ser, silica gel dessicant, a vacuum pump, a dry gas meter, and
a rotameter. Particulate concentration measurements were made
at all three of the sample port locations indicated in Fig. 62.
The sample probe was oriented at the yaw angle determined by
the Fecheimer probe at each traverse point, and an isokinetic
sampling rate was set according to the velocity indicated by
the S-tube.
Although the duct walls were wetted at locations A and B,
few droplets were visually observed in the sampling train. At
point C, however, liquid droplets were present to such a degree
as to thoroughly wet the sampling probe. Large pressure drops
in the sampling train developed quickly as droplets built up on
the filter mats and the vacuum pump was not able to maintain
the required isokinetic sampling rate. This forced sampling
times to be kept very short.
An illustration of the difference in droplet concentra-
tions at points B and C was obtained by using an in-stack
filter to collect particulate samples at each location.
Filters inserted at each location are shown in Fig. 64. It can
be seen that the filter deployed upstream of the ID fan is much
more uniform in density of collected particulate than the
downstream filter. The uneven appearance of the downstream
filter may be due to the presence of particulate contained in
or on the surface of liquid droplets. The variation in droplet
concentration is likely due to the cyclonic flow field; en-
trained droplets leaving the scrubber are driven to the duct
walls and pass in a film near the walls into the ID fan. The
droplets are then evenly distributed and reentrained by the
flow leaving the fan.
A liquid droplet precutter'designed and fabricated by
FluiDyne was added to the sampling train in an attempt to
collect the sampled liquid droplets. The precutter is shown in
Figs. 65 and 66. The designed droplet cut diameter for this
inertial separator was 10 ym. Use of the precutter in further
103
-------
Figure 64.
Filter samples taken at the Seneca
Wastewater Treatment Plant during
preliminary field tests
B = Upstream of I.D. fan
C = Downstream of I.D. fan
104
-------
Figure 65. Inertial separation precutter used in prelim-
inary field tests at Seneca Wastewater
Treatment Plant
4
I I I I I I
Figure 66. Internal view of inertial separation precutter
105
-------
particulate sampling eliminated the pressure buildup due to
droplets on the filters. When the precutter was opened after
sampling, however, no liquid was found, although some particu-
late matter was present.
When the water vapor content of the scrubber exhaust
stream was measured with a dewpoint indicator, (Alnor Model
7000 V) the stream was found to have a relative humidity of
only 79%. Thus, it was concluded that liquid droplets which
were collected in the precutter were evaporated by the sample
flow, since the gas stream was not saturated. This interesting
finding leads to the conclusion that one should not assume a
saturated stream on the basis of observed liquid droplets, as
is suggested in EPA Method 5. Large discrepancies in total
water vapor content calculated from dewpoint measurements and
with the Method 5 calculation were also observed in Ref. 55.
In addition to the various problems described above, the
particulate concentrations measured in the preliminary field
tests varied considerably. These measurements are summarized
in Table 12. The wide variation emphasizes the basic diffi-
culty in obtaining reliable particulate concentration measure-
ments in the scrubber exhaust stream. Some of the unreliability
may be related to the amount of particulate matter lost in the
sample probe due to impingement and settling of water droplets,
even though the probe was cleaned according to Method 5 pro-
cedures. After one series of tests, the particulate matter
collected in probe washing amounted to 13% of the total parti-
culate matter collected.
MODEL STUDIES
A model studies program was developed on the basis of the
information obtained through the preliminary field tests. A
laboratory model of the exhaust of a typical wet scrubber was
designed and fabricated. This model was then used to evaluate
techniques and instrumentation for use in wet scrubber emission
measurements.
Model Design and Fabrication
The laboratory model, shown schematically in Fig. 67, was
constructed at FluiDyne's Medicine Lake Laboratory. A liquid
droplet injection system located in the horizontal air supply
ductwork provided the simulation of entrained liquid droplets
in the flow stream. A fixed vane type centrifugal mist elimi-
nator located just upstream of the test section produced a
cyclonic flow field. The mist eliminator was purchased from
Peabody Engineering Corporation, the manufacturer of the scrub-
bers used at the preliminary field test site. The mist elimi-
nator is a scaled-down version of those used for separation of
entrained droplets in the scrubber exhaust at the Seneca
106
-------
TABLE 12. PARTICULATE CONCENTRATION MEASUREMENTS FROM
PRELIMINARY FIELD TESTS
Sample Location (See Figure 62) ABC
Particulate Concentration Measurements
without Droplet Precutter in Sample
Train (mg/Nm3) - 9? 61
Particulate Concentration Measurements
with Droplet Precutter in Sample 100 - 41
Train (mg/Nm3) 56
107
-------
o
oo
West Test Port
5.25 D
South Test Port
3.75
-AIR SHUT-OfT VALVE
2.25 D
0.75 I
Pitot
Tube
Traverse
Test Port
for Droplet
Eliminator
Nozzles
I U-T£3T "PORT
Test
Port
Levels
Mist
Eliminator
//////////// / i / /////////
Figure 67. Schematic of scrubber exhaust model
-------
Treatment Plant. Two sample ports were installed at each of
four levels downstream of the mist eliminator in the test
section as indicated in Fig. 67. A platform was constructed to
provide access to the test section sample ports, as shown in
Fig. 68.
The droplet injection system is pictured in Figs. 69-72.
Three spray nozzles inject water into the flowing air stream at
a rate which can be controlled by adjusting the water pressure,
the atomizing air pressure, and needle valves inserted in the
spray nozzles. The nozzles were directed such as to fill the
duct uniformly with water droplets, as determined by visual
observation.
Flow Velocity Studies
The main objective of the velocity studies in the scrubber
exhaust model was to assess the ability of various instruments
to determine the volumetric flow rate and the flow angularity
in the cyclonic flow field. Three instruments were used to
make velocity measurements in the model; a standard S-tube, a
three dimensional directional pitot probe, and a Fecheimer type
probe.
The S-tube, attached to a standard EPA Method 5 type
sampling probe, was first used to assess the error level which
would exist if this instrument were simply oriented parallel to
the duct centerline without accounting for the cyclonic nature
of the flow field. Several 12 point velocity surveys were made
with the S-tube, traversing the duct as if the flow were uni-
formly parallel to the duct centerline, and the volumetric flow
rate determined from these traverses was compared with the
actual volumetric flow rate through the model as determined
from pitot tube traverses in the horizontal air supply ductwork
upstream of the test section. The results of these surveys,
shown in Fig. 73, show that the error is on the order of 60% to
115% of the actual flow rate and that the error actually in-
creases from 2 to 6 duct diameters downstream of the mist
eliminator, after an initial decrease.
Attempts were made to determine the yaw angle of the flow
by using the demonstrated directional sensitivity of an S-tube
as indicated in Fig. 56. However, the turbulent fluctuations
in the flow caused fluctuations in the S-tube output of such
magnitude as to completely overshadow the capability of this
instrument to locate the angle of maximum velocity. Thus, the
S-tube could not be used with any degree of reliability to
determine the yaw angle in cyclonic flow in this manner. A
potential method for use with the S-tube which was not evalua-
ted is the so called "null method," in which the S-tube is
rotated until the pressure sensed by the two pressure taps is
equal. The instrument should then be aligned at 90° to the
109
-------
Figure 68. Scrubber exhaust model and test platform at
Medicine Lake Laboratory
110
-------
Figure 69.
Droplet injection system for scrubber exhaust
model-external hardware
Figure 70.
Spray nozzles for droplet injection system (re-
moved from ductwork of scrubber exhaust model)
111
-------
Figure 71.
Droplet injection system spray nozzles producing
low droplet density mist
Figure 72.
Droplet injection system spray nozzles producing
high droplet density mist
112
-------
V
m
V
V = measured volumetric flow rate
in
V = actual volumetric flow rate
3.
2.2
2.0-
1.8'
1.6-
1-4.
1". 2 -
l.Q.
.x
0.8-
0.&-
0.4-
0.2,.
0 -
O
D
a
o
o
a
O - V = 115 m /min
a,
D - V = 68 m /min
a.
2
Distance Downstream of Mist Eliminator,
(Duct Diameters)
Figure 73.
Volumetric flow rate error for 12-point
velocity surveys - in scrubber exhaust
model with S-tube parallel to duct
centerline
113
-------
velocity and can be rotated so as to be properly aligned for a
velocity measurement.
Extensive surveys of the cyclonic flow field were made
with a directional velocity probe capable of accurately measur-
ing both pitch and yaw angles as well as the flow velocity.
This instrument, the United Sensor DC-125 directional velocity
probe, is shown in Figs. 74 and 75 together with a duct traverse
unit which was used to accurately position the probe in the
test section of the scrubber exhaust model. As can be seen in
Fig. 75, the pressure taps on this probe are very small, thus
making use of the instrument impractical in particulate-laden
streams. It was used in the model, however, to obtain an
accurate map of the flow field for evaluation of other velocity
instruments. This probe mounted in the scrubber exhaust model
is shown in Fig. 76. Manufacturer's calibration data were used
to reduce the probe measurements to velocites.
The results of the velocity measurements with the United
Sensor probe are shown in Figs. 77-80. The definition of the
coordinates used in presenting the data is shown in Fig. 67.
The actual flow rate, Q , used to nondimensionalize the velocity
terms in Figs. 77-80, was determined from pitot tube traverses
in the horizontal air supply ductwork upstream of the test
section. The curves indicated in Figs. 77-80 represent intui-
tive estimates of the "average" data and are not based on
curve-fitting or regression formulas. The most interesting
observation made from these velocity surveys is the fact that
the average radial velocity component is very small, particu-
larly as the distance downstream of the mist eliminator in-
creases.
Extensive velocity surveys were also made with a Fecheimer
type probe, shown in Fig. 81. The traverse unit shown in Figs.
74 and 76 was used to position the Fecheimer probe for these
measurements. The same points were measured with the Fecheimer
probe as with the United Sensor probe. The Fecheimer probe was
calibrated against a standard pitot tube and was found to have
a velocity coefficient of 1.06.
Velocity components determined from the yaw angle and
total velocity measured with the Fecheimer probe are shown in
Figs. 82-85. The same dimensionless presentation is used as
for the United Sensor probe. While the United Sensor probe
appears to provide somewhat more repeatable data, comparison of
the measurements indicates that the average velocity components
determined with the two instruments agree quite closely, with
the exception of the radial component which can not be measured
with the two-dimensional Fecheimer probe. Since the radial
components indicated in Figs. 77-80 are essentially negligible,
however, the approximation made in using the two-dimensional
instrument seems to be acceptable.
114
-------
Figure 74. United sensor DC-125 3-dimentional directional
probe and traverse unit
Figure 75.
Sensing head of united sensor DC-125 directional
probe
115
-------
INTERNAL
VIEW
EXTERNAL
VIEW
Figure 76. United sensor DC-125 directional probe mounted
in scrubber exhaust model for velocity traverses
116
-------
Radial
Component
17 *
v =
O Q
a o
O e
A e
o
90
180
270
1
0
1
3 -
Circumferential
Component ^
V
1 -
Axial
Component
v
2
1
0
-l1-
Total
Velocity
v*
Duc-t
Centerline
Figure 77. Normalized velocity components measured 0.75 duct
diameters downstream of mist eliminator -
United sensor probe
117
-------
Radial
Component
vr*
0
-1
Ttr,
o e
O e
O e
A e
= o
= 90
= 180
= 270
4 r
Circumferential
Component 3
IT *
2 -
1 -
Axial
Component
2 _
1
0
-1
Total
Velocity.
v*
0
Duct
Centerline
0.6 . 0.8
r/ro
1.0
Duct
Wall
Figure 78. Normalized velocity components measured 2.25 duct
diameters downstream of mist eliminator -
United sensor probe
118
-------
v*
7trf
Radial
Component 0
r -1
O
D
O
A
e
e
e
o
90
180
0 » 270
^
.1^.
\f
J-&
Circumferential
Component
V
Axial
Component
v *
z
2
1
0
-1
O O
Total
Velocity
v*
0.2
Duct
Centerline
I
0.4 0.6 0.8
r/r_
1.0
Duct
Wall
Figure 79. Normalized velocity components measured 3.75 duct
diameters downstream of mist eliminator -
United sensor probe
119
-------
Radial
Component
v* =
Ttr,
0
_n
A
O
D
6 = 0
6 =» 90
9 = 180
6 = 270
A
Circumferential
Component
-e*
Axial
Component
1
0
Total
Velocity
A r
3
2
1
0
0
0.2
Duct
Centerline
0.4
0.6
r/r
0.8
1.0
Duct
Wall
Figure 80. Normalized velocity components measured 5.25 duct
diameters downstream of mist eliminator -
United sensor probe
120
-------
Figure 81. Fecheimer probe used in velocity
surveys of scrubber exhaust model
121
-------
7* =
?cr
o
D
O
A
9
e
e
o
90
180
e = 270
Circumfer-
ential
Component
V
Axial
Component
v *
z
2
I
0
-1
Total
Velocity
4
3
1
0
0
Duct
Centerline
0.2
0.4
0.6 , 0.8
r/ro
i.o
Duct
Wall
Figure 82.
Normalized velocity components measured 0.75 duct
diameters downstream of mist eliminator -
Fecheimer probe
122
-------
+ Ttr
v* o
o e
D e
O e
A e
o
90
180
270
4r-
Circumferential 3
Component
V 2
Axial
Component
2
1
0
-1
Total
Velocity
4
3
2
1
0.2
Duct
Centerline
0.4 0.6 . 0.8
r/r0
1.0
Duct
Wall
Figure 83. Normalized velocity components measured 2.25 duct
diameters downstream of mist eliminator -
Pecheimer probe
123
-------
Tcr
7* =
o
D
O
A
9
9
9
0
90 \
180
9 = 270
Circumferential
Component
V
2_
Axial
Component
v *
z
Total
Velocity
0
Duct
Centerline
Figure 84. Normalized velocity components measured 3.75 duct
diameters downstream of mist eliminator -
Fecheimer probe
124
-------
v* =
Qa
O
D
O
A
e
e
e
o
90
180
e = 270
Circumferential
Component
V
Axial
Component
2
1
0
-1
Total
Velocity
0
0.2
Duct
Centerline
0.4
0.6
0.8
r/r
1.0
Duct
Wall
Figure 85. Normalized velocity components measured 5.25 duct
diameters downstream of mist eliminator -
Fecheimer probe
125
-------
The degree of accuracy of the Fecheimer probe in deter-
mining total volumetric flow rate was also examined. Compari-
sons of flow rates determined from the axial velocity component
in several 24-point surveys and the actual flow rate are shown
in Fig. 86. It can be seen that the reliability of the volu-
metric flow rate surveys increased greatly with distance down-
stream of the mist eliminator. At the maximum distance achiev-
ed in the scrubber exhaust model, 5.25 duct diameters, the 24-
point surveys provided accuracy within 5% of the actual value.
At locations closer to the mist eliminator, a greater number of
measurement points would be required to achieve sufficient
accuracy and repeatability.
On the basis of the information presented above, the
Fecheimer probe appears to be a suitable instrument for deter-
mination of both volumetric flow rate and flow angularity in
the cyclonic flow typical of wet scrubber outlets.
Liquid Droplet Studies
As outlined previously, the principal difficulty in
sampling a droplet-laden stream is to properly collect and
account for the liquid droplets. This is important for a
number of reasons. First, the collection of liquid droplets by
the sample train is critical since particulate may generally be
present in the droplets, either in solution or merely adhering
to the droplet surfaces. Thus in order to adequately measure
the total particulate concentration in the gas stream, droplets
must be collected. Second, the humidity of a gas stream must
be known in order to calculate the proper isokinetic sampling
rate. Since humidity measurements in a droplet-laden stream are
difficult to make, it is usually assumed that such gas streams
are saturated. As observed in the preliminary field tests,
however, this is not necessarily the case. The potential error
in isokinetic rate arising from this assumption is illustrated
in Fig. 87. Finally, in some cases a droplet concentration
gradient may exist in a given flow field, also as observed in
the preliminary field tests, which results in a difference in
the amount of liquid present at various sampling points. In
order to understand the potential effects of such droplet
concentration gradients, being able to account for the mass of
liquid droplets at a given sampling point would be very useful.
Analysis of Droplet Motion
Analysis of the motion of entrained liquid droplets in
cyclonic flow indicates that the characteristic situation
consists of a core region of the duct which is free of liquid
droplets and an ever increasing concentration of droplets
nearer to the duct wall, as observed at the Seneca plant. This
was demonstrated by an analysis of droplet motion based on the
velocity field measured in the scrubber exhaust model. In the
following, the coordinate directions are defined as shown in
126
-------
Qm
Qa
1.3
1.2
1.1
1.0
0.9
0.8
0
T
I
I
I
I
I
01 23 456
Duct Diameters Downstream of Mist Eliminator
Figure 86. Ratio of measured and actual volumetric flow rate
for 24-point velocity surveys with Fecheimer probe
downstream of mist eliminator
127
-------
RH = Actual Relative Humidity
of Droplet-Laden Stream
0)
4-1
&
tr>
c
-H
1
Q
H
4J
C
*H
.X
0
tn
H
q
H
0)
o
OJ
c
H
j^
0
w
200
180
160
140
120
100
80
60
40
20
0
RH = 50%
100%
Figure 87. Error in isokinetic sampling rate due to assumption
that droplet-laden stream is saturated
128
-------
Fig. 67.
Certain simplifying assumptions were made in order to
arrive at some useful results. These assumptions include:
1. The radial component of the gas velocity is negli-
gible .
2. The circumferential velocity component of a droplet
at any point is the same as the circumferential
velocity component of the gas.
3. The droplets are spherical.
4. The motion of an individual droplet is not affected
by the presence of neighboring droplets.
5. The radial and axial components of drag force on a
droplet are given by Stokes1 Law.
These assumptions are similar in nature to the usual assump-
tions applied to the analysis of particle motion in cyclone
type dust collectors (Refs. 56, 57).
Equations of motion in the radial and axial directions may
then be written:
dt
&- - V V
dt VZ)
2 (2)
where *C and d^ are the mass and diameter of a given
droplet? respectively . These may be reduced to:
2r . 18y dr _ V02 = 0 (3)
~
dr
2Z A 18y /dz . vV g - 0
72 + 7TT" Vdt z)
d Z
dt
129
-------
where PD is the liquid density and u is the gas viscosity.
Since the radial and axial accelerations are relatively small,
it is convenient to simply neglect the second order derivatives
in order to arrive at a system of equations which can be solved
analytically. This approximation, while obviously not correct,
has been shown to be of value in the analysis of cyclones
(Refs. 56 and 57). For a particular form of the circumferen-
tial velocity term VQ as used in Ref. 57, this approximation
was shown to give a solution which nearly matched the solution
of the actual radial equation of motion after a certain elapsed
time. Again, this approximation is used routinely in cyclone
analyses.
Neglecting the second order derivatives and combining
equations (3) and (4) then gives:
(5)
Solutions to Equation (5) will then provide approximations to
the trajectories of droplets entrained in the swirling flow.
Simple polynomial functions were derived to approximate the
velocity components of the gas, based on the measured velocity
data presented in Figs. 82-85. These functions are illustrated
in Fig. 88. By introducing these polynomial functions into
Equation (5), the solution may be found by integration. In
this manner, the trajectory of a droplet may be determined; a
droplet initially located at point (r,, z,) moves to a point
z,) moves to a point r~, z~) , where
(continued on page 132)
130
-------
V*
Radial
Component 0
r -1
Circumferential
Component
V
V
12 () -10 ()
ro' vro'
Axial
Component
= 5.2(^-)2 - 4£-
ro
r ^3
ro'
Total
Velocity
v*
Duct
Centerline
r/ro Duct
Wall
Figure 88.
Approximations to velocity components measured
in scrubber exhaust model
131
-------
2 6
VZ2=
144 Q "
u
(12-10 Q (12-10 fa
+ In
(.792) in / 12-10 r.
12-10 ^
o
(6)
If r, is taken as the position of a given droplet at z, = 0
and r- is the radius of the duct (r_ = rQ), the distance
downstream of the mist eliminator (z = 0; at which the droplet
is impacted on the duct wall (in duct diameters) is given by:
a
Vo\ [
50 (r-L/r^l)
12-10 rI/ro
. 72 ( ~ - ij -
+ In (6ro/r1
-(ft--),,
12-10 ^i
-0
.792)1
(7)
This impaction distance is shown for various droplet sizes and
initial radial positions in Fig. 89, based on average condi-
tions from the cyclonic flow velocity tests and an assumed
droplet mass density of 1000 kg/m . As can be seen from Fig.
89, all droplets 30 urn or larger in diameter should be removed
from the flow before reaching the first test port level in the
cyclonic flow model. Some droplets on the order of 10 ym and
smaller, however, should remain entrained in the gas stream
throughout the entire test section of the cyclonic flow model.
132
-------
Zimp
2r
7 *-
6 _
3 -
1 -
Duct
Centerline
Droplet Size
10/MD
Droplets Injected-^
Duct
Wall
Figure 89.
Number of duct diameters downstream of mist elimi-
nator where water droplets are impacted on duct
wall, scrubber exhaust model
133
-------
Trajectories of 10 ym droplets as given by Equation (6)
are plotted in Fig. 90. These results show that the central
area of the duct should have no liquid droplets, and that the
concentration of droplets near the duct walls should increase
continually with distance downstream of the mist eliminator,
less those droplets actually deposited on the walls.
The results of this approximate analysis agree qualita-
tively with the observations made at the Seneca Treatment
plant. In addition, measurements of droplet size made by
another investigator at the Seneca Treatment Plant at location
B in Fig. 62 were reported to indicate a mass median diameter
of approximately 5 ym (Ref. 58) in close agreement with the
results of the analysis. The predicted behavior of liquid
droplets in cyclonic flow again points to the need to be able
to account for the mass of liquid in droplet form at various
points in the flow stream.
Liquid Droplet Testing--
A sampling method was selected for evaluation in the
scrubber exhaust model on the basis of the information gained
through the methodology review and the preliminary field test
program. The method involves collecting two gas samples with a
heated probe in order to vaporize all the collected liquid.
The sampled liquid in droplets, as well as the liquid present
in vapor form at the sample point, is then collected in a
condenser apparatus after passing through a particulate filter.
The first sample is collected isokinetically with a standard
sampling nozzle directed into the gas stream. Moisture collect-
ed in this first gas sample will consist of the liquid in both
vapor and droplet form. The second gas sample is collected
with a sampling nozzle which inertially separates the droplets
from the collected gas sample. Moisture collected in this
sample should consist of only the vapor phase. Comparing the
collected moisture in the two gas samples then will provide the
mass of liquid in droplet form.
This method was chosen in view of the difficulties re-
portedly encountered with other possible droplet collection
schemes. Precutters, such as that used in the preliminary
field tests, allow collected water to be reevaporated by the
sample stream. This difficulty also exists when filters are
used to remove droplets from a sample stream. The various
problems associated with humidity measurements in a droplet-
laden stream are discussed in Ref. 59, which also recommends
the use of an inertial separation type probe. These difficul-
ties were also confirmed by experiments in the scrubber exhaust
model with entrained liquid droplets in which various types of
hygrometers and dew point indicators were used with and without
filters in attempting to measure the humidity of the gas stream.
Repeatable results could not be achieved, and these methods
were abandoned.
134
-------
7 r
2r,
6 ~
5 -
3 -
2 _
0
r/r
Figure 90. Trajectories of 10 ym water droplets in scrubber
exhaust model
135
-------
Several nozzles designed to remove liquid droplets are
shown in Figs. 91-95, together with a standard sample nozzle to
be used for isokinetic sampling. The simplest concept for
inertial separation is illustrated by Fig. 91, in which a
standard sampling nozzle is simply oriented in a downstream
direction and a sample is collected at a low flow rate.
Figure 92 illustrates a modification to this concept in which
a shield is added to the downstream-facing standard nozzle to
inhibit growth of a liquid film on the probe surface and to
impart a radially outward motion to the droplets. A suction
slot, through which the liquid film is drawn off, and a separate
sampling area are employed in the nozzles shown in Figs. 93 and
94. A similar concept was followed in the design of the nozzle
illustrated in Fig. 95, but this nozzle faces in the direction
of the flow in order to avoid potential disturbances from
locating the sampling slot in the wake of the sample probe.
The sample is collected through a slot which forces the flow to
pass around a sharp obtuse angle in the probes shown in Figs.
94 and 95.
It was originally intended to evaluate the two-sample
method in the cyclonic flow downstream of the mist eliminator
in the scrubber exhaust model. However, after a number of
initial tests were made in the cyclonic flow test section this
test plan was abandoned because of difficulties in accounting
for the total mass of water removal from the flow stream due to
impingement on the turning elbow upstream of the mist elimi-
nator and on the face of the mist eliminator itself (see Fig.
67). This made evaluation of the performance of the nozzles
with an acceptable degree of accuracy impossible.
Testing of the various two-sample nozzle arrangements was
then moved to a position in the horizontal air supply ductwork
upstream of the mist eliminator as indicated in Fig. 67.
Although this point is not a sufficient number of duct dia-
meters upstream of the elbow as specified by EPA Method 1, this
site was selected to ensure that the injected droplets were
sufficiently well-mixed in the flow and that a steady state
liquid-vapor mixture was reached before the test point.
Samples of the droplet-laden stream were collected at the
center of the duct at the test location with each of the pro-
posed inertial droplet separator nozzles and passed through a
heated probe which evaporated any collected liquid. The mois-
ture ratio of this extracted sample as determined by collection
of water in a condenser following EPA Method 5 procedures was
then compared with an estimated humidity ratio of the gas
stream based on the rate of injected water droplets and the
thermodynamic state of the air in the duct upstream of the
water injection point. An evaluation of the effectiveness of
the proposed nozzles in removing droplets from the stream could
then be made on the basis of this comparison.
136
-------
Figure 91.
Standard sampling nozzle for use as inertial drop-
let eliminator nozzle in two-sample method
137
-------
Figure 92.
Standard nozzle with shield for use as inertial
droplet eliminator nozzle in two-sample method
138
-------
Figure 93.
Inertial droplet eliminator nozzle with suction
slot for liquid film removal
139
-------
- -o
Figure 94. Inertial droplet eliminator nozzle with suction
slot and aerodynamic tail
140
-------
o
Figure 95.
Inertial droplet eliminator nozzle with suction
slot oriented in upstream direction
141
-------
A total of 64 such screening tests were made with various
proposed inertial droplet separator nozzles. The results of
these tests showed that the separator nozzle illustrated in
Fig. 95 had the greatest potential for use in the two-sample
method. The standard sampling nozzle oriented in a downstream
direction (Fig. 91), even when used with a liquid film shield
such as shown in Fig. 92, or with other more aerodynamically
designed shields and droplet deflectors not pictured here,
consistently indicated humidity ratios far in excess of the
estimated duct humidity. In certain tests the indicated humi-
dity ratio actually exceeded that possible by collecting all of
the droplets passing the sample point. This poor performance
was attributed to the presence of liquid film collecting on the
surface of the nozzles and to large droplets which collected on
the portions of the nozzle upstream of the sample point and on
the heated probe (Figs. 96 and 97). These droplets were perio-
dically sheared from the surfaces on which they collected into
the flow stream and conceivably could have been collected by
the nozzle. The performance of the nozzles shown in Figs. 93
and 94 was somewhat better in general, due to the use of the
suction slot to remove liquid film buildup, but was found to be
very erratic. This is very likely due to the droplet shearing
problem described above, which is inherently present when the
sampling point is downstream of other surfaces on which drop-
lets can collect (see Fig. 98). This problem is avoided by the
design of the nozzle illustrated in Fig. 95 as can be seen in
Fig.100. This nozzle indicated the best performance in the
screening tests.
A careful evaluation of the two-sample method was then
made on the basis of these screening tests using a standard
nozzle to sample the liquid droplets and vapor and the inertial
droplet separator nozzle shown in Fig. 95 to sample for vapor
content only. The elbow was removed from the air supply duct-
work in order to eliminate any flow disturbances. The thermo-
dynamic state of the air ahead of the droplet injection point
was measured, and all water deposited on the walls of the duct
was collected and measured. This allowed an accurate assess-
ment of the nozzle performance to be made.
A standard 6.35 mm (1/4 in) sampling nozzle was used to
collect the water droplets isokinetically. A summary of the
standard nozzle tests is shown in Table 13. The good agreement
between the mass concentration of water collected by the nozzle
when compared with the water injected and the water collected
on the duct wall indicates that the standard isokinetic sampling
technique is sufficient to collect the water droplets.
The results of the tests with the inertial droplet separa-
tor nozzle are shown in Table 14. The "actual duct humidity"
shown in Table 14 is determined from a computer model of the
evaporation of the water droplets. This model used a finite
142
-------
Figure 96.
Standard sample probe collecting isokinetic sample
in scrubber exhaust model
143
-------
Figure 97.
Standard sample nozzle oriented in downstream
direction as inertial droplet separator in
scrubber exhaust model
144
-------
Figure 98.
Standard sample nozzle with droplet shield as
inertial droplet separator in scrubber exhaust
model
145
-------
Figure 99.
Inertial droplet separator nozzle shown in
Figure 94 in scrubber exhaust model
146
-------
Figure 100. Inertial droplet separator nozzle shown in
Figure 95 in scrubber exhaust model
147
-------
TABLE 13. COMPARISON OF MEASURED MASS FLOW RATE OF WATER VAPOR AND DROPLETS
WITH ACTUAL MASS FLOW RATE
Measured
Actual
00
Run
No.
106
107
108
109
110
W
meas
m .
air
W
meas.
m .
gm air
water
kgair (kg/h;
16.36 4893
16.50 4867
16.27 4840
16.48 4874
16.72 4867
= Concentration of
collected sample
= Flow rate of air
A Ambient Water Injected
water Vapor in Droplets
Air
r) (kg/hr) (kg/hr) (kg/hr)
80.10
80.31
78.75
80.32
81.38
water in
in model
23.19 75.71
23.12 75.71
22.51 75.71
24.42 75.71
24.38 75.71
m . = Ambient
water , n ,
model +
Deposited
m
On Wall water
(kg/hr) (kg/hr)
21.39 77.51
20.35 78.48
23.19 75.03
24.13 76.00
24.37 75.72
vapor entering
injected drop-
measured
actual
1.03
1.02
1.05
1.06
1.07
lets - water deposited
on duct wall
m
water
W
meas
m
air
-------
TABLE 14. PERFORMANCE OF INERTIAL DROPLET SEPARATOR NOZZLE
Run
No.
Ill
112
114
115
116
W
measured*
gro
water
kg .
air
9.31
9.91
10.72
12.28
13.02
W
actual**
am
3 water
kg .
air
8.25
8.17
8.68
8.28
8.78
Mass of
Droplets
at Sample
Point
gm
water
kg .
air
7.39
7.57
8.40
10.17
9.54
Mass of
Droplets
Removed
by Nozzle
gm
water
kg .
^air
6.31
5.82
6.35
6.15
5.28
Nozzle
Efficiency
i
.85
.77
.76
.60
.55
Avg.
.71
* Concentration of water vapor indicated by collected sample
** Concentration of water vapor calculated with computer model
-------
element method to analyze the evaporation of individual drop-
lets of diameter dD by the equation
= 12 D p
* v a m
(8)
where & (d_) = mass rate of evaporation of droplets of
size d_.
D = binary vapor diffusivity
p= = mass density of dry air
a
p = mass density of water
p = static pressure in duct
p = vapor pressure of water at droplet surface
v temperature
w, = average humidity ratio in duct
m(dD)= mass of droplets in a computational bandwidth
characterized by diameter dD
A droplet size distribution at the spray nozzle injection point
was determined from manufacturer's information (Fig. 101).
Deposition of droplets on the duct wall was also computed by
the model following methods presented in Refs. 60 and 61 and
was found to agree well with the water collected on the duct
walls in the evaluation tests.
The results shown in Table 14 indicate that the inertial
droplet separator nozzle removed roughly 70% of the droplets
existing at the sample point. Therefore, this nozzle could be
used to determine the humidity of the droplet-laden stream by
using a calibration factor as determined by the evaluation
tests. The removal efficiency of the droplet separator is very
likely dependent on the size distribution of the droplets,
however, and further evaluation of the separator with different
sized droplets would be desirable.
EVALUATION OF SAMPLING TECHNIQUE
The basic sampling technique developed for wet scrubber
150
-------
.06
.05
.04
.03
.02
.01
m
0246 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42
d (Mm)
Figure 101. Droplet size distribution at spray injection point
-------
particulate emission sampling involves the use of a Fechiemer
probe for use in volumetric flow measurement and to determine
proper angular positioning for the sample probe, an inertial
droplet separator nozzle for use in humidity measurement, and
a Method 5 type sampling probe to collect the particulate.
Based on the information obtained in the model studies program,
this approach should provide satisfactory emission measurements
in a typical wet scrubber exhaust.
Further development and evaluation of the inertial droplet
separator nozzle is desirable. If such a nozzle could be
developed to eliminate all of the droplets at a given sample
point, then the use of a heated probe and condenser in con-
junction with the droplet eliminator would be unnecessary. In
such a case, the sampled gas could simply be analyzed for
moisture content with any of various available types of humidity
measurement devices.
The question of the required number of sample points was
not addressed in this study. Since the basic particulate
concentration measurement follows EPA Method 5, the required
number of sampling points for this method is recommended for
use in scrubber exhaust measurements.
APPLICABILITY TO OTHER EMISSION SOURCES
The particulate sampling method developed for wet scrubbers
should be generally applicable to any confined stream with
entrained liquid droplets. The Fecheimer probe is not required,
of course, unless cyclonic flow is present. This probe also
should not be used if multiple vortex cyclonic flow is present
(see Ref. 2).
152
-------
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 Flow,
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 1971.
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-45Q/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.
153
-------
REFERENCES (Cont.)
23. Kreichelt, Thomas E.f 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.
25. Evans, Robert J., "Methods and Costs of Dust
Control in Stone Crushing Operations," U. S. De-
partment of the Interior, Bureau of Mines, Informa-
tion Circular 6669.
26. Stear, James R., Municipal Incineration: A Review
of Literature, U. S. 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., "Investi-
gation 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
Particulate 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.
154
-------
REFERENCES (Cont.)
13. Hidy, G. M. and Brock, J. R., "An Assessment of
the Global Sources of Tropospheric Aerosols,"
Proceedings of the Second Clean 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
450/3-74-047, June 1974.
16. Atmospheric Emissions from Sulfuric Acid Manufactur-
ing 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
Performance Standards: Steam Generators, Incinera-
tors, 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, Petro-
leum Refineries, Storage Vessels, Secondary Lead
Smelters and Refineries, Brass or Bronze Ingot
Production Plants, Iron and Steel Plants, Sewage
Treatment Plants, Volume I, 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 Preparation Plants, Volume 2. Summary and Test
Data, EPA 450/2-74-0216, October 1974.
22. Atmospheric Emissions from Nitric Acid Manufacturing
Processes, U. S. Department of Health, Education and
Welfare, Public Health Service, Publication No. 999-
AP-27, 1966.
155
-------
REFERENCES (Cont.)
33. Colpitts, J. W., Automation of Monitor Sampling,
Confidential Internal Report, ALCOA Badin Smelting
Works, July 1968.
34. Gravity Ventilation, Zurn Industries, Inc. Air
Services Division, Bulletin PMB-2, p. 4.
35. Lundgren, D. and S. Calvert, "Aerosol Sampling
with a Side Port Probe," American Industrial Hygiene
Association Journal, No. 28, 1967, pp. 208-215
36. "Determining Dust Concentration in a Gas Stream,"
PTC 27-1957, American Society of Mechanical Engineers,
New York 1957.
37. Meiering, A. G. and H. J. Hoefkes, "Dust Emission
in Corn Drying," presented at the Annual Meeting of
Canadian Society of Agricultural Engineering,
Laval University, Ste. Troy, Quebec, August 4-8,
1974.
38. Meiering, A. G., H. J. Hoefkes, and L. Otten,
"Particulate Emission in Corn Drying," Journal
of the Air Pollution Control Association, Volume 27,
Number 6, June 1977, pp. 548-552.
39. Source Test Section, Kentucky Division of Air Pol-
lution, "Particulate Emission Tests, Wabash Elevator
Company, Uniontown, Kentucky," November 1974.
40. Ward, Thomas E., "Emissions from the Grain Drying
Facility at Quaker Oats Company, Chatanooga, Tennessee,
Emission Testing Report 73-GRN-4, Office of Air
Quality Planning and Standards, Environmental Pro-
tection Agency.
41. Cowherd, C., "Measurement of Fugitive Particulate,"
presented at Second Symposium on Fugitive Emissions:
Measurement and Control, Houston, Texas, May 23-25,
1977.
42. Jorgenson, G. V., J. P. Pilney and E. E. Erickson,
"Air Emissions in Iron Ore Mining and Enrichment,"
presented at Second Symposium on Fugitive Emissions:
Measurement and Control, Houston, Texas, May 23-25,
1977.
156
-------
REFERENCES (Cont.)
43. Cooper, H.B.H., Jr. and A. J. Rosand, Jr., Source
Testing for Air Pollution Control, Environmental
Science Service Division, 1971.
44. Calvert, S., Entrainment Separators for Scrubbers,
EPA 650/2-74-119a, October 1974.
45. Calvert, S., Fine Particulate Scrubber Performance
Tests, EPA 650/2-74-093, 1974.~
46. Grove, D. J., and W. S. Smith, "Pitot Tube Errors
Due to Misalignment and Non-Streamlined Flow,"
Stack Sampling News, November 1973.
47. Lea, J. F. and D. C. Price, "Mean Velocity
Measurements in Swirling Flow in a Pipe," Flow;
Its Measurement and Control in Science and Industry,
Volume If Part 1, Flow Characteristics, Instrument
Society of America, 1974, pp.313-317.
48. Chigier, N.A., "Velocity Measurement in Vortex Flows,"
Flow; Its Measurement and Control in Science and
Industry, Volume I, Part 1, Flow Characteristics,
Instrument Society of America, 1974, pp. 399-408.
49. Orloff, K. L. and H. H. Bossel, Laser Doppler
Velocity Measurements of Swirling Flows with
Upstream Influence, NASA CR 2284, July 1973.
50. Odom, J. L., "Testing Stacks with Cyclonic Flow,"
Stack Sampling News, Volume 3, No. 3, September 1975.
51. Ower, E. and R. C. Pankhurst, The Measurement of Air
Flow, Pergamon Press, 1966.
52. Steam, Its Generation and Use, 38th Edition, Babcock
and Wilcox Company, 1972-.
53. Madowski, E. R., "Stack Particulate Sampling,"
Mechanical Engineering, October 1974.
54. Method 8, FE 409, Connecticut Department of
Environmental Protection.
55. Gilardi, E. F. and H. F. Schiff, "Comparative Re-
sults of Sampling Procedures used During Testing
of Prototype Air Pollution Devices at New York City
Municipal Incinerator."'Stack Sampling News,
Volume I, No. 4, October 1973.
157
-------
REFERENCES (Cont.)
56. Strauss, W., Industrial Gas Cleaning, Pergamon
Press, New York, 1966.
57. Leith, D. and Licht, W., "The Collection Effi-
ciency of Cyclone Type Particle Collectors - A
New Theoretical Approach," AIChE Symposium
Series, Vol. 68, No. 126.
58. Tomaides, M., Private Communication.
59. Dean, Robert C., Aerodynamic Measurements, Gas
Turbine Laboratory, Massachusetts Institute of
Technology, 1953.
60. Friedlander, S. K. and Johnstone, H. F.,
"Deposition of Suspended Particles from Tubulent
Gas Streams," Industrial and Engineering Chemistry,
Vol. 49, No. 7, July 1957, pp. 1151-1156.
61. Beal, S. K., "Deposition of Particles in
Turbulent Flow on Channel or Pipe Walls,"
Nuclear Science and Engineering; 40,
1-11 (1970), pp. 1-11.
158
-------
APPENDIX
VELOCITY INSTRUMENTATION FOR LOW
VELOCITY, PARTIALLY CONFINED
SOURCE PARTICULATE SAMPLING
By
D. P. Saari
H. A. Hanson
F. A. Bezat
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 m/sec (Ref. Al). 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 con-
fined. 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 circula-
tory, flow fields.
Knowledge of the velocity of an emission stream is neces-
sary 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 gen-
erally 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 volu-
metric 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
159
-------
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 uncon-
fined 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.
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 are generally several
manufacturers for each type of instrument, but those listed in
Table Al provide 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.
Pitot Static Tubes or S-tubes
The principle of the pitot static tube or S-tube is to
measure, by means of a probe inserted in the flow, the differ-
ence between total pressure and static pressure (or nearly
static pressure, in the case of the S-tube). 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 in-
strument 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 direct-
ion 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 extended 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 differen-
tial develops between the two S-tube ports, a secondary flow of
purge gas is induced in the pneumatic bridge which is propor-
tional to the fluid velocity. This arrangement extends the low
velocity range of the S-tube to approximately 15 m/min and
prevents fouling of the tube in a particulate stream, since no
fluid actually enters the S-tube. However, the accuracy of this
instrument is poor in the very low velocity range.
160
-------
TABLE Al. LOW VELOCITY INSTRUMENTATION
1» - .3048 ft.
°C - 5/9 (»F -32)
CTl
inaiKUHEin:
Pitot Tube* or 8-Tubes
l.with Inclined Manometer
2. with Micromanometar
3. S- tuba with purge flow
Vane or Propeller
Anemometers
4. Vane Anemometer
5. Rotating Vane(mech)
6. Rota ting 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
16. Vortex Shedding
Anemometer
li. Ion Deflection
18. Sonic Pulse
19. Laser Doppler
VELOCITY
RANGE
(ft/min)
600+
400+
50-1500
50-6000
30-5000
25-5000
25-5000.
25-5000
10-6000
10-6000
10-3000
30-8000
12-3000
60-6000
100+
1-12,000
20-2000
.03-
2xl05
MAX.OFER. FUNCTIONING IN
TEMP. PARTICULAR
(°F) STREAM
800+ (only limited
BOO+ lby probe
800+ (blockage
800+ Good
250 Fair
250 Fair
250 Fair
220 Fair
220 Good
BOO Needs occa-
sional Cleaning
200 Needs occa-
sional Cleaning
500 Needs occa-
sional Cleaning
200 Needs occa-
sional Cleaning
Tested to
25 g/m3
400 4,5 g/n3
Good
140 Needs SOB*
Cleaning _,_
120 Relatively
Unaffected
uniimitea "Gooa ror
Small
P articulate
ACCURACY
AT
1 ft/ceo
N/A
N/A *'
+50%
+20% (estl
+20% (estl
+20% (estl
+20% (estl
+20% (eat)
+20%
+20%
+6%
+50%
N/A
44%
+30%
DIRECTIONAL
SENSITIVITY
BS (8-tube)
Poor
No
(± in
I direction
lo£
(alignment
Yes
NO
to 0°
or 180°
No
to 0°
or 180°
Possible
Possible
Possible
No
Yes
Yes
Yes
RUGGED-
MESS
Good
Good
Good
Fair-
Good
Fair-
Good
Fair-
Good
Fair-
Good
Poor
Fair
Fair
Good
Good
Good
Good
Good
Good
Poor
Poor
TYPICAL IJWTROMBHT
APFROX. COMMENTS
MANUFACTURER MODEL COST
Dwyer
Dwyer
Hastinga-
Raydist
Flowrite
Davis-
Davis,
Gill
Gill
Gill
Thenno-
Sys terns-
Thermo-
Systems
Thermo-
Systems
Rastlngs-
Raydist
Bowles
FluiDynamic
Bastings-
Ravdist
J-Teo
Thermo*
Systems
E.G.SG.
Brown-
Beveri,
DI6A
$100
$200
KRF $100
$200
$800
$500
$200
+purge
system
1610 $1000
1650 $500
1234H $1300
w/reade
PCI-30 $700
$200
$1000
AFI-6K 11500
$1000
$2000
$7000
$10 K
20K
Affected by
Vibration
Good for uni-
form flow; aux-
iliary readout
needed below
- 100 ft/min
Direct
' velocity
Indication
Cumulative
reading
Continuous
reading
gas
Tamp, rang*
may vary
Response time
may be too
fast '
not temp.
iut eomp. (,+ $100
for temp, comp
model.)
Good field
Instrument.
Temp. range
may vary
Under
Development
Under
Development
- Not a field
instrument
at this time.
-------
Vane or Propeller Anemometers
A vane or propeller anemometer is positioned normal to the
flow velocity. The flow through the anemometer imparts aero-
dynamic force to the vanes, or propeller, 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
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 signal polarity can be
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 distance 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 (approximately
7.5-15 m/min) for this type of instrument.
Heated Element Anemometers
The principle of operation of hot wire and hot film ane-
mometers 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. Commercially 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
162
-------
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 compensa-
tion. 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 particulate 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.
Fluidic Anemometers
One type of fluidic velocity measurement instrument in-
corporates 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 concen-
trations 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 differ-
ential 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 con-
centration. However, the output is very nonlinear and the
minimum velocity is about 0.3 m/sec.
163
-------
Vortex Shedding Anemometer
A new velocity measurement instrument under development
employs vortices generated by a strut and counted by an ultra-
sonic sensor (Ref. A4). The frequency of vortex generation is
directly proportional to the velocity if the Strouhal number
of the obstruction is known, thus providing a calibration.
The instrument reportedly has great accuracy at velocities 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, temperature and
pressure. However, the characteristics of the instrument 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 (Ref. A4).
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. In addition, the tube would
require occasional cleaning, since buildup of particulate
could interfere with the ion stream and collector.
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 information needed for volumetric flow evaluation even in
highly turbulent flows. Particulate matter should have little
effect on such an instrument. The laser Doppler type instru-
ment 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
associated 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 instru-
ment 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
164
-------
only as a laboratory instrument at this time.
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 (Anemotherm) 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
propeller anemometers with electric output (Gill Model 27100)
were successfully used for velocity measurements in roof
ventilators.
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 ane-
mometers 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 velocity 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.
EVALUATION OF SELECTED INSTRUMENTS
A number of velocity measurement instruments were selected
from the basic list given in Table Al and evaluated as to
their suitability for low velocity, unconfined source particu-
late sampling. Testing of these instruments was accomplished
165
-------
either in a model facility of a roof ventilator (Figs. Al and
A2) or in the field, measuring the flow from roof ventilators
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 instru-
mentation, it is not possible to determine the accuracy of an
instrument in the field, as discussed previously. Thus, the
roof ventilator model was used to determine the degree of
reliability of the various instruments under conditions which
simulated those existing in field applications. A summary of
the tests of selected instruments follows.
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 instru-
ment 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 deter-
mined 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 circula-
tory 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.
Rotating Vane Anemometer - Davis Mechanical Vane Anemometer
(Figs. A5 and A6)
This instrument was tested with the roof ventilator
model, and volumetric flow rates determined from 24 point
velocity surveys were compared with the actual flow rate
supplied to the model. The anemometer was fitted with an
extension and hand held at the exhaust plane of the model.
With thi^ method, at flow rates ranging from about 75m /min
to 200 m /min, the flow rate indicated by 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
166
-------
Figure Al. Roof model located at the Medicine Lake
Laboratory - looking south
Figure A2.
Roof model and ducting located at the
Medicine Lake Laboratory - looking southwest
167
-------
Figure A3. Roof ventilators at Hitchcock Industries, Bloom-
ington, Minnesota, viewed from the northwest
Figure A4.
Roof ventilators at Hitchcock Industries, Bloom-
ington, Minnesota, viewed from the east
168
-------
Figure A5.
Davis mechanical rotating vane anemometer with
ball swivel attachment
Figure A6.
Davis mechanical rotating vane anemometer with
extension
169
-------
ventilator 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 conditions. 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 circular
ductwork which supplied the flow (Fig. A2). The flow at the
base of the ventilator 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 velocity
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.
Hot Film Anemometer - Thermo-Systems Model 1610
This instrument was a special version of the standard hot
wire anemometer Model 1610 designed to withstand high tempera-
tures (160°C). The instrument was fitted with a special
protective collar used successfully in previous studies in
which velocities from 0.03 to 4.5 m/sec were measured. How-
ever, the instrument 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.
Therefore, the hot-wire type anemometer was judged inadequate
for the type of field work performed in this study in roof
ventilator sampling.
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 deter-
mined 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
distinguish between velocities 180° apart. This is substan-
tiated by observing the comparison of measured and actual flow
rates with the baffle in place (Fig. A9), in which case the
agreement was excellent.
170
-------
Actual Volumetric Flow
Rate (m3/min)
350 .
300
250 -
200 .,
150
100 .,
50 -
O Without Uniform
Flow Baffle
D
With Uniform
Flow Baffle
100% Accuracy
100
150 200 250
,+25%
50%
300 350
Indicated Volumetric Flow
Rate (
Figure A7
Volumetric flow rate at base of roof ventilator
model determined with Davis mechanical rotating
vane anemometer
171
-------
Figure A8. Hastings/Raydist PCI-30 hot thermopile anemometer
with meter and extension
172
-------
Actual Volumetric Flow
Rate (m3/min)
350
Flow Direction Indicator
300
250
200
150
100
50 '
Without Uniform
Flow Baffle
With Uniform
Flow Baffle
Yes
B
No
0
a
100% Accuracy
.+50%
0
50
100
150
200
250
300
350
Figure A9.
Indicated Volumetric F]ow
Rate (m^/min)
Volumetric flow rate at base of roof ventilator
model determined with Hastings-Raydist PCI-30
hot thermopile anemometer
173
-------
In order to overcome this problem, a lightweight flow
direction tuft was attached to the instrument. Several 24-
point velocity surveys were then made at the base of the
ventilator model, using the flow direction tuft to align the
probe with the flow and estimate the flow angle. The volu-
metric flow rate was then computed from the proper component
of the measured velocity as determined from the estimated flow
angles. The flow was generally found to be vertically upward
or downward, and incorporating the uniform flow baffle elimi-
nated 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. AID 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. There-
fore, 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.
Hot Film Anemometer - Thermo-Systems Model 1650
(Fig. A12)
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.
Wedge Type Hot Film Anemometer - Thermo-Systems Model 1234 H
(Figs. A13-A16) ~~
An extensive evaluation of this instrument for potential
use in roof ventilator emission measurement was conducted in
the roof ventilator model. Details of this evaluation are
presented in Ref. A8, which was prepared for the Emission
Measurement Branch, Emission Standards and Engineering Division,
Office of Air Quality Planning and Standards, of EPA. The
particular system which was evaluated consisted of four velocity
sensing probes, each with an associated temperature compensating
probe. One pair of sensing probes and the various output
signal equipment, cables, and switches are shown in Fig. A13.
174
-------
Error in Velocity
+8 T
+6
\ fj\
tf ^&
10
0
20
30
/ 1
m m i ill
40 50 60 70 80 |90
<3?(deg)
-6 .
Results Symmetrical
For - (
Flow
Figure AlO. Velocity error with yaw angle for Hastings-Raydist
PCI-30 hot thermopile anemometer - without
protective cap
175
-------
(deg)
0 10 20 30 40 50 60 70 80 90
0
-10 -I
-20
-30 4
-40 '
-50 ..
-60 -
i
f
1
-70 -
r
-80 T
j
-90 -
-100
Error in Velocity
Flow
RESULTS SYMMETRICAL FOR -
Figure All. Velocity error with yaw angle for Hastings-Raydist
PCI-30 hot thermopile anemometer - with protective
cap
176
-------
Figure A12.
Thermo-Systems, Inc. Model 1650
Hot Film Anemometer
177
-------
Figure A13.
Thermo-Systems, Inc. Model 1234H Hot Film
Anemometer System as furnished by manufacturer
178
-------
Figure A14.
Thermo-Systems, Inc. Model 1234H Hot Film
Anemometer System as modified and tested
by FluiDyne
179
-------
Figure A15.
Sensor element and temperature compensating
element mounted on rigid tubing
Figure A16.
Temperature compensating trim resistor box
and 15 CM cable insert modification made by
FluiDyne
180
-------
Actual Volumetric
Flow Rate (m3/min)
150 -
-25%
125 -
100 .
+25%
V
O Probe #7421
Q Probe #7423
25
50
75
100
125
150
Measured Volumetric Flow Rate (m /min)
175
Figure A17.
Volumetric flow rate at base of roof ventilator
model determined with Thermo-Systems, Inc. Model
1234H hot film anemometer (single probe tests)
181
-------
Certain modifications were made before the evaluation
tests were conducted, as illustrated in Figs. A14, A15, and
A16. The velocity sensing and temperature compensating sen-
sors were mounted on lightweight, rigid support tubes to
facilitate proper placement and alignment in the roof ventilator
model and to prevent sagging and oscillation of the sensing
elements in the flow field. Despite these modifications and
very careful handling of the equipment, two of the fragile
sensing elements were damaged during the course of testing.
Several 24 point velocity surveys were made at the base
of the roof ventilator model with the two remaining sensor
probes. All measurements were made with the flow straightening
baffle in place. The results of these volumetric flow measure-
ments are shown in Fig. A17. This Figure indicates that, when
the probes are properly aligned and protected, the accuracy of
measurements made with the wedge type hot film anemometer is
acceptable. It is not known why the error for the two probes
shown in Fig. A17 was different; contributing factors could
have been misalignment of the hot film elements in the velocity
tests, incorrect reading of the voltage output, or improper
calibration by the manufacturer.
A major advantage of this instrument is its temperature
rating, which is somewhat higher than other low velocity
instruments, as seen in Table Al. The major disadvantage is
the very fragile nature of the sensor elements. The instru-
ment also showed a slight degree of yaw sensitivity which
might be helpful for aligning the probes in smoother flow
streams.
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 applica-
tion, 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 particulate concentration.
182
-------
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 and as considered in these
studies, 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 re-
latively 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 Al may be used as a guideline for selection of more
appropriate instrumentation in such cases.
183
-------
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-
600/2-76-170, 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, RlpT 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
1975, pp. 397-398.
A7. "Performance Standards for New Stationary Sources:
Primary Aluminum Industry," Federal Register, Volume 41,
No. 17, January 26, 1976.
A8. Bezat, Frederick A. and Saari, David P., "Evaluation of
Thermo-Systems, Inc. Model 1234H Wedge Type Hot Film
Anemometer for Roof Ventilator Velocity Measurements,"
FluiDyne Engineering Corporation, April, 1977.
184
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
. REPORT NO.
EPA-600/2-80-034
.TITLE AND SUBTITLE
EFFECTIVE SAMPLING TECHNIQUES FOR PARTICIPATE
EMISSIONS FROM ATYPICAL SOURCES
Final Report
. AUTHOR(S)
5. REPORT DATE
January 1980
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPOHT NO.
D. P. Saari, H. A. Hanson
3. RECIPIENT'S ACCESSION NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Fluidyne Engineering Corporation
5900 Olson Memorial Highway
Minneapolis, Minnesota 55422
10. PROGRAM ELEMENT NO.
1AD712 BA-031 (FY-77)
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, N.C. 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 2/77-1/79
14. SPONSORING AGENCY CODE
EPA/600/09
15. SUPPLEMENTARY NOTES
Interim Report: EPA-600/2-77-036, February 1977
16. ABSTRACT
Sampling techniques for measuring particulate emissions from four "atypical" stationary
source categories were developed and evaluated. The categories include low effluent
velocity streams, extended dimensions, partially or totally unconfined flow, and
saturated gas streams or gas streams with entrained liquid droplets. Research
included literature surveys, laboratory model testing, and field testing of instru-
ments and techniques on three specific sources -- gravity roof ventilators, grain
dryers, and wet scrubbers. These three sources served as representative examples of
the four atypical source categories.
The sampling techniques recommended for roof ventilator emission measurements include
a high volume particulate sampler and a heated thermopile anemometer deployed near
the base of the ventilator. The same instruments, deployed at the lower exhaust, are
recommended for grain dryer emission measurements. An EPA Method 5 type sampling train
an inertial droplet separator, and a Fecheimer probe are recommended for use in wet
scrubber emission measurements.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
* Air pollution
* Particles
Industrial plants
* Nonuniform flow
* Sampling
* Systems
* Evaluation
Tests
Field tests
13B
131
20D
14B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
199
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
PREVIOUS EDITION IS OBSOLETE
185
------- |