EPA-600/2-76-284
December 1976
Protection T^ctmology Series
DEVELOPMENT OF PROCEDURES FOR THE
MEASUREMENT OF FUGITIVE EMISSIONS
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Tri"r?$8 ParK, North Carolina 27711
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and non-point sources of pollution. This
work provides the new or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-76-284
December 1976
DEVELOPMENT OF PROCEDURES
FOR THE MEASUREMENT OF
FUGITIVE EMISSIONS
by
P.W. Kalika, R.E. Kenson, and P. T. Bartlett
TRC-The Research Corporation of New England
125 Silas Deane Highway
Wethersfield, Connecticut 06109
Contract No. 68-02-1815
ROAP No. 21AUZ-004 '
Program Element No. 1AB015
EPA Project Officer: Robert M. Statnick
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
FOREWORD
This document constitutes the final report for the work accomplished
on EPA Contract No. 68-02-1815, Methods of Measurement of Fugitive
Emissions.
The Environmental Technology Division of TRC - The Research Corp-
oration of New England was responsible for the work performed in this
program. The work was conducted under the technical and administrative
direction of Dr. R. M. Statnick of the Industrial Environmental Research
Laboratory, Process Measurements Branch of the Office of Research and
Development of the Environmental Protection Agency, which is located at
Research Triangle Park, North Carolina. Mr. John E. Yocom, Chief Engi-
neer of TRC, was Program Director and Mr. Peter W. Kalika, Engineering
Manager, was project manager. Dr. Robert E. Kenson and Mr. Paul T. Bart-
lett served as task managers. Major technical contributions were pro-
vided by Mr. Don L. Shearer, Mr. William A. Marrone, Mrs. Roberta Huston,
Mr. Kevin C. Tower, Miss Nicola F. Grappone, and Miss Jo Anne Marchese.
ii
-------
TABLE OF CONTENTS
SECTION PAGE
1.0 SUMMARY 1
2.0 INTRODUCTION 4
2.1 The Nature of the Fugitive Emission 4
Problem
2.2 Measurement Techniques 6
2.3 Scope of Work Undertaken 7
3.0 CONCLUSIONS 9
3.1 Industrial Sources 9
3.2 Sampling Strategies 9
3.3 New Methods 9
3.4 Cost-Effectiveness 12
3.5 Quasi-Stack Strategy 12
3.6 Alternative Strategy 13
3.7 Method Limitations 13
4.0 RECOMMENDATIONS 14
4.1 Industry Studies 14
4.2 Strategy Tests 14
4.3 New Methods 15
4.4 Documentation 15
4.5 Package Development 15
4.6 Impact Determination 15
4.7 Refinement of Quasi-Stack 16
5.0 DISCUSSION 17
5.1 Task I - Identify the Sources of 17
Fugitive Emissions
5.1.1 Literature Search for Sources 17
5.1.2 Trade Association/Industry/Govern- 18
ment Agency Contacts
5.1.3 Matrix Charts for Identification of .... 21
Selected Industries
5.1.4 Fugitive Emission Sampling Techniques .... 23
and Strategies
5.2 Task II - Evaluating Various Sampling 27
Strategies
5.2.1 Evaluating 18 Selected Industries 27
5.2.2 Suitability of Sampling Strategies 28
5.2.3 Evaluate Sampling/Analysis Techniques .... 28
5.2.4 Selection of Sampling/Industry Com- .... 32
binations for Further Evaluation
5.2.5 Cost-Effectiveness of Sampling 34
Strategies
5.3 Task III - Prepare a Technical Manual for the . 36
Measurement of Fugitive Emissions
iii
-------
5.4 Task IV - Field Test of a Fugitive Emis- ... 36
sion Technical Manual
5.4.1 Summary of Procedures and Analysis .... 40
5.4.1.1 Description of the Mold Pouring .... 40
Hood Arrangement
5.4.1.2 Description of Sampling Port Loca- .... 44
tions
5.4.1.3 Field-operated Instrument and Chem- ... 46
ical Laboratory
5.4.1.4 EPA Method 5 Particulate and Conden- ... 48
sible Train
5.4.1.5 G.I. Cascade Impactor and Condensi- ... 51
ble Train for Particulates
5.4.1.6 IKOR Model 206 Continuous Particu- ... 54
late Monitor
5.4.1.7 Beckman Model 109 Flame lonization ... 56
Detector for Total Hydrocarbons
5.4.1.8 Intertech Model Infra-2 Non-disper- ... 56
sive Infrared Analyzer for Carbon
Monoxide
5.4.1.9 High-volume Filter Sampler for Am- .... 58
bient Particulates
5.4.2 Analytical Procedures 60
5.4.2.1 Sample Recovery and Preparation in .... 60
the Field
5.4.2.2 Analytical Laboratory Procedures 62
5.4.3 Evaluation of Particulate Concentra- .... 65
tion Data
5.4.3.1 Background Tests 70
5.4.3.2 EPA Method 5 Test Comparison .70
5.4.3.3 Cascade Impactor Test Comparison 74
5.4.3.4 Continuous Monitor Test Comparison .... 76
5.4.3.5 Test Setup Limitations 80
5.4.3.6 Critique of Sampling Methods 80
5.4.3.7 Presentation of Particle Sizing Data ... 81
5.4.3.8 Particulate Composition 84
5.4.4 Evaluation of Hydrocarbon Concentra- .... 85
tion Data
5.4.5 Evaluation of Carbon Monoxide Concen- ... 89
tration Data
5.5 Task V - Issue a Technical Manual 91
5.6 Units of Measure 93
APPENDIX
A LITERATURE REFERENCES TO FUGITIVE EMISSIONS
B SUMMARY OF TRADE ASSOCIATION/INDUSTRY/
POLLUTION CONTROL AGENCY CONTACTS ON
FUGITIVE EMISSIONS
C MATRIX CHARTS USED TO IDENTIFY AND CHARACTER-
IZE FUGITIVE EMISSION SOURCES
iv
-------
D MEASUREMENT OF FUGITIVE EMISSIONS - LITERATURE
REFERENCES TO METHODS
E SAMPLE CRITIQUE OF SAMPLING METHODS
F EPA METHOD 5 DATA SHEET
G CASCADE IMPACTOR DATA SHEET
H IKOR CONTINUOUS MONITOR DATA SHEET
I HI-VOL FILTER TEST DATA SHEET
J LABORATORY WORKSHEET FOR FIELD SAMPLES
K CONVERTING UNITS OF MEASURE
-------
LIST OF FIGURES
FIGURE PAGE
5-1 Simplified Flow Sheet for Gray Iron Foundry 31
5-2a Cost-effectiveness of Four Fugitive Emission .... 35
Sources - Sampling Strategies Applied to Different
Sources
5-2b Cost-effectiveness of Four Fugitive Emission .... 35
Sources - Sampling Strategies Applied to the Same
Source
5-3 Illustration of Test Set-up for Measuring Fugitive . 39
Emissions from Mold Pouring in a Gray Iron Foundry
5-4 Schematic Diagram of Sample Port Locations 45
5-5 EPA Method 5 Particulate Sampling Train Including . . 49
t Condensibles
5-6 G.I. Cascade Impactor Assembly 52
5-7 IKOR Continuous Particulate Monitoring System .... 55
5-8 Schematic Diagram of Laboratory Trailer Setup .... 57
Showing Hydrocarbon and Carbon Monoxide Analyzers
5-9 High Volume Sampling Station for Ambient Particu- . . 59
lates
5-10 Test 17: Continuous Monitor Trace 77
5-11 Test 21: Continuous Monitor Trace 78
5-12 Layout of Molds 86
5-13 Plot of Mean Half-Hour Hydrocarbon Concentrations . . 88
for Period April 7 - 21, 1975
5-14 Plot of Mean Half-Hour Carbon Monoxide Concentra- . . 92
for Period April 7-21, 1975
vi
-------
LIST OF TABLES
TABLE PAGE
3-1 Classification of Industries for Possible Future ... 10
Studies of Fugitive Emissions
3-2 Evaluation of Candidiate Strategies for Measuring . . 11
Fugitive Emissions
5-1 Summary of TRC Studies Relevant to Fugitive Emis- . . 19
sions Measurement
5-2 Matrix Chart of Fugitive Emission Sources in Copper . 20
Smelting
5-3 Classification of Industries for Possible Future ... 22
Studies of Fugitive Emissions
5-4 Feasible Sampling Strategies for Industries/ 25
Processes Selected for Possible Fugitive
Emissions Studies
5-5 Method of Rating Hazard Potential of Fugitive Emis- . 29
sions of 18 Selected Industries
5-6 Matrix of Iron and Steel Foundries Fugitive Emission . 30
Source/Sampling Strategy Combinations
5-7 Matrix of Final Selections of Fugitive Emissions ... 33
Sampling Strategy/Industry Source Combinations
5-8 Summary of Tests - Mold Pouring . 41
5-9 Foundry Activity During Testing . . 42
5-10 Particulate Concentration Data 67
5-11 Summary of Statistical Analysis-Total of All Iso- . . 68
kinetic Tests for Each Method
5-12 Summary of Statistical Analysis-Total of 10 Tests: . . 69
All Methods Iso-kinetic and All Methods were Run
Concurrently
5-13a Differences in Test Method Performances-Total .... 71
Particulate Concentration, GR/DSCF
5-13b Differences in Test Method Performances-Total .... 72
Condensibles Concentration, GR/DSCF
vii
-------
5-13c Differences In Test Method Performances; Partlculate . 73
on Filter(s) Concentration, GR/DSCF
5-14 Particle Size Distribution (% by Count): EPA .... 82
Method 5
5-15 Particle Size Distribution (% by Count): Mean .... 83
Values and Standard Deviation
5-16 Particle Size Distribution (% by Count): Statis- . . 83
tical Differences Between Process Emissions
and Background Condition
5-17 Total Hydrocarbon Concentration 87
5-18 Carbon Monoxide Concentration Data 90
viii
-------
1.0 SUMMARY
This document is a final report of EPA Contract No. 68-021815
"Methods of Measurement of Fugitive Emissions." TRC - The Research
Corporation of New England conducted the studies undertaken in this
program under the supervision and sponsorship of the Industrial Environ-
mental Research Laboratory, Process Measurements Branch of the Office of
Research and Development of the Environmental Protection Agency, which
is located at Research Triangle Park, North Carolina.
It has been recognized that fugitive (non-stack) emissions from
industrial processes are a potentially significant portion of the total
air emissions in many industries. The purpose of the program undertaken
was to develop basic procedures which could be used to sample and analyze
these fugitive emissions. The TRC studies were conducted over a 12-month
period between July, 1974, and June, 1975, and divided into the following
tasks:
1. Task I - Identify the Sources of Fugitive Emissions. This
task consisted of a literature search, a review of TRC exper-
ience in fugitive emission measurements, and contacts with
industries, regulatory agencies, and industry trade associa-
tions. The results are presented in the form of a matrix
showing industries, processes, fugitive emission sources, and
type of pollutants emitted.
2. Task II - Evaluate Various Sampling Strategies. In this task,
TRC considered the application of three broad measurement ap-
proaches to the sources listed in the matrix from Task I. Cost
and accuracy estimates were made for each of the methods. These
approaches are:
A. Quasi-stack. Here the fugitive emission is temporarily
hooded or enclosed and a temporary duct or stack and fans
are installed to permit sampling by means of standard stack
sampling methods.
B. Roof Monitor. The term "roof monitor" encompasses a var-
iety of similar situations in which fugitive emissions
-1-
-------
leave a building or enclosure through roof monitors, roof
ventilators, and windows and doors.
C. Upwind-downwind. A carefully designed upwind-downwind
sampling network is used under selected meteorological
conditions to show the contribution of fugitive sources
to ambient air quality. Three dimensional arrays may be
needed to permit measurement of pollutant flux past a
sampling grid. Mathematical models and tracer studies
may be required to describe complex situations.
3. Task III - Prepare a Technical Manual for the Measurement of
Fugitive Emissions. Based on Task I and Task II and discussions
with EPA, the quasi-stack method was selected for preparation of a
test procedures document. A draft was prepared which used as an
example case the measurement of fugitive emissions from mold
pouring at a gray iron foundry.
4. Task IV - Field Test of a Fugitive Emission Technical Manual.
In this task, TRC conducted a field test of the quasi-stack
method for mold pouring at a gray iron foundry in southern
Connecticut.
5. Task V - Issuance of a Technical Manual. At the completion
of the field test program and after all data had been analyzed,
the Field Test Procedures Document was thoroughly evaluated by
TRC and EPA and modified as appropriate. The approved document
was issued as Volume 2 of the final report.
The basic conclusions developed by TRC in this study were:
7
1. Eighteen industries/processes were believed to have potentially
significant fugitive emissions. These were selected from all
the industries/processes studied based on the potential of their
fugitive emissions to be hazardous and to have high emission
rates.
2. The industry or process step/sampling strategy combinations
best able to define and measure the fugitive emissions of these
18 industries/processes were identified and studied. This re-
sulted in the specification of three basic sampling strategies
(quasi-stack, roof monitor, and upwind-downwind) which could
be used to define and measure most of the potentially signifi-
cant fugitive emissions identified in this program.
3. For the field test portion of this program (Task IV), a quasi-
stack sampling strategy was identified as capable of measuring
a significant number of fugitive emission sources and as being
amenable for implementation within reasonable time and cost
-2-
-------
constraints. Mold pouring in a gray iron foundry in Southern
Connecticut was chosen as the fugitive emission source for ap-
plication of the quasi-stack strategy.
4. A quasi-stack sampling strategy was found to be a valid method
for studying mold pouring fugitive emissions provided that con-
tinuous analysis methods are used and that background pollutant
concentrations are separable from those from the operation
studied.
As a result of this study, TRC makes the following basic recommen-
dations:
1. EPA should consider further studies of the fugitive emissions
of six specific industries.
2. EPA should fund the study of roof monitor and upwind-downwind
sampling strategies.
3. EPA should develop specific instrument packages for fugitive
emissions.
This report consists of two volumes. One, which is titled, "Methods
of Measurement of Fugitive Emissions — Industrial Fugitive Emissions
Sources and Sampling Strategies," documents Tasks I, II, and IV of this
study. A second, which is titled, "Technical Manual for the Measurement
of Fugitive Emissions — Quasi-Stack Sampling Method for Industrial Fugitive
Emissions," documents Tasks III and V.
-3-
-------
2.0 INTRODUCTION
The program was undertaken over a 12-month period between July,
1974, and June, 1975, for the purpose of development of sampling proce-
dures to determine the emission rate and chemical and physical charac-
teristics of fugitive emissions.
It has been recognized that as the emissions from industrial pro-
cesses through stacks or vents are controlled, by 90 percent, for exam-
ple, overall process emissions may not be reduced 90 percent. Non-stack
or fugitive emissions, which can be produced by almost every industrial
process, could become significant portions of the total process emis-
sions. For the purpose of this program, fugitive emissions were defined
as those emissions of air pollutants not directed through ducts or stacks
and not amenable to measurement by established source sampling methods.
2.1 The Nature of the Fugitive Emission Problem
Almost every industrial operation capable of emitting air pollu-
tants exhibits some fugitive emissions. These emissions vary with the
basic process and the manner in which a given process is carried out,
e.g., a new cement plant with enclosed raw materials handling and good
housekeeping will have far less fugitive dust emissions than an old,
poorly operated plant with open storage.
The mechanism for release also varies widely. Fugitive emissions
of gaseous pollutants result from:
1. Leaks in piping, packing glands, and other components of pro-
cess equipment.
2. Exposure of liquids containing a component of appreciable va-
por pressure.
-4-
-------
3. Exposure of unstable solids, one of whose breakdown products
is a gaseous pollutant.
4. Generalized emissions from a building roof monitor or roof
fan where many smaller fugitive emissions accumulate within
and are vented from the building.
5. Operations carried out outdoors or partially outdoors (e.g.,
coke charging and pushing).
Gaseous emissions, once they have left the confines of a process
and when they are not exhausted through a duct or stack where emission
rates can be measured, are fugitive emissions. On the other hand, par-
ticulate fugitive emissions, because of gravity and inertial effects,
may or may not become fugitive emissions once they have left a process.
Fugitive particulate emissions behave differently from gaseous fu-
gitive emissions. This difference is principally a function of parti-
cle size. Particles below about five to 10 microns behave similarly
to gases because of their slow settling rate. Above this range of par-
ticle size, the settling rate will influence how far a particle will
travel and how easily it will become reentrained. Thus, measurements
of fugitive particulate emissions must include particle size determina-
tions. In addition, meteorological conditions and the geometry of the
release point produce complications in describing the behavior of par-
ticulate as compared to gaseous fugitive emissions.
Typical mechanisms for the release of particulate fugitive emis-
sions include:
1. Operations carried on outdoors (e.g. , material handling, coke
charging and pushing).
2. Participates formed from the reaction of gaseous fugitive emis-
sions with the atmosphere, moisture, or other gases (e.g.,
?2®5 from phosphorous emissions, and NH^Cl from adjacent
NH3 and Cl£ sources).
-5-
-------
3. Leaks from confined process streams carrying particulates or
condensible vapors (e.g., coke oven door leakages).
4. Outdoor storage of solid materials (e.g., raw material piles,
dried sludge beds).
5. Accumulations of dust from inadequate housekeeping which are
subject to reentrainment.
6. Traffic on dusty plant roads.
Another factor which complicates the assessment of fugitive parti-
culate matter from an industrial process is the ubiquity of airborne
particulate matter as a background component of generalized air qual-
ity. Unless the particulate fugitive emission is unique chemically
and/or physically and has not already influenced the characteristics
of surrounding background particulate matter, it may be difficult to
assess the impact of the fugitive emission in question.
2.2 Measurement Techniques
The techniques presently available for measurement of fugitive
emissions are:
1. Material Balances
2. Quasi-Stack Sampling
3. Roof Monitor Sampling
4. Upwind-downwind Sampling
Material balance techniques involve no actual tests, but simply
determine what enters and exits the process. The material unaccounted
for by product or measured emissions is assumed to be the fugitive
emission.
The quasi-stack method involves the temporary enclosing, hooding,
and ducting of the emission source such that well-established source
sampling and analysis techniques can be applied.
-6-
-------
The roof monitor method refers to the specialized sampling of fugi-
tive emissions from well-defined building openings such as roof monitors,
skylights, ceiling vents, windows, etc.
The upwind-downwind method is generally applied to large outdoor
sources and involves carefully designed arrays of samplers upwind and
downwind of the suspected fugitive emission source. These samplers can
be located both at ground level and at elevated positions. The method
involves meteorological measurements which must be related to the sam-
ples taken. Diffusion modeling and tracer techniques may also be in-
volved to assist in identifying individual sources of fugitive emissions.
2.3 Scope of Work Undertaken
The program to develop procedures for the measurement of fugitive
emissions was divided into the following tasks:
Task I - Identify the sources of fugitive emissions (15 percent
of total program effort)
Task II - Evaluate various sampling strategies (15 percent)
Task III - Prepare field test procedures document (10 percent)
Task IV - Field test of fugitive emission field test procedures
document (45 percent)
Task V - Issuance of final field test procedures document (15
percent)
This report (Volume I) summarizes the results of Tasks I, II, and
IV, as follows:
Section 3.0 - Conclusions based upon the program undertaken
Section 4.0 - Recommendations for further studies
Section 5.0 - Discussion of the procedures and results of the work
performed
Appendices - Detailed descriptions of specific methods, refer-
ences, and detailed matrix charts.
-7-
-------
The studies of the "Methods of Measurement of Fugitive Emissions --
Quasi-Stack Sampling Method for Industrial Fugitive Emissions" (Tasks
III and V), comprise a separate volume (Volume 2) of this report.
-------
3.0 CONCLUSIONS
The study of industrial sources of fugitive emissions and fugitive
emission measurement methods, as well as the field test of a quasi-stack
sampling strategy, has led to the following conclusions:
3.1 Industrial Sources
There were 18 industries/processes believed to have potentially
significant fugitive emissions. These 18 (Table 3-1) were selected
from all those studied based on the potential for their fugitive emis-
sions to be hazardous and to have high emission rates. Other criteria
were also used to rank these 18.
3.2 Sampling Strategies
There are three basic sampling strategies suitable for measurement
of the fugitive emissions from most potentially significant sources in
all the industries studied. These three, the quasi-stack, the roof mon-
itor/enclosure vent, and the upwind-downwind sampling strategies, are
described briefly in Table 3-2, including their limitations and general
accuracy ranges.
3.3 New Methods
For some significant fugitive emission sources, present sampling
strategies may be inadequate. An example was the copper smelting indus-
try, where openness of building structures limits present sampling strat-
egy application. New developments in remote sensing were seen as po-
tentially useful tools in this application.
-9-
-------
TABLE 3-1
CLASSIFICATION OF INDUSTRIES FOR
POSSIBLE FUTURE STUDIES OF FUGITIVE EMISSIONS
Category
Metallurgical
Energy/Fuels
Chemical Products
Rock Products
Agricultural Opera-
tions
Industry/Process
Primary Aluminum
Primary Copper
Electric Furnace Steel
Iron & Steel Foundries
Coke Making
Coal Mining
Coal Gasification
COG (Char-Oil-Gas)
Shale Oil
Petroleum Refining
Oil Production
Plastics
Tire & Rubber
Phosphate Fertilizer
Lime
Sand & Gravel
Asphalt Batching
Cattle Feedlots
Soil Tilling
Grain Harvesting
-10-
-------
TABLE 3-2
EVALUATION OF CANDIDATE STRATEGIES FOR MEASURING FUGITIVE EMISSIONS
Candidate .Strategy.
1. Roof-Monitor/Enclosure
Vent Strategy (Use of
Available Enclosures
such as Buildings
Quasi-Stack Sampling
Strategy (Construc-
tion of special en-
closures with off
take duct for
sampling.)
3a
Upwind-Downwind
Strategy (Simul
taneous Upwind-
Downwind sampling.)
3b
Upwind-Downwind
Strategy (Upwind™
Downwind sampling
together with dif-
fusion modeling and
tracer studies.)
Requirements
Openings must be regular and
amenable to sampling for
concentration and air flow.
Impractical if too many
openings.
Ventilation (natural or
power) must be such that a
material blance can be
established.
3. If natural ventilation 3.
exhausts the fugitive
emissions, enough data must
be taken to establish emissions
under a variety of
meteorological conditions.
1. Enclosure must capture essentially 1.
100% of fugitive emissions.
2. Enclosure must not affect emission 2.
rate or physical and chemical
characteristics of emissions.
3. The scale of the fugitive emissions 3.
may be too great for construction
of a reasonably sized enclosure.
1. Fugitive emissions must be unique 1.
and identifiable in relation to,
or not influenced by, other
plant emissions.
2. Sampling array must encompass the 2.
fugitive emissions cloud and
provide enough detail to establish
concentrations across the entire
cloud.
3. Wind flux measurements are needed
at several points both laterally
and longitudinally in the cloud
path.
1. Model must be applicable to i.
fugitive emission points.
Applicability
Buildings housing multiple
units which have fugitive
emissions - Cl caustic cells,
aluminum pots, electric
furnaces, etc.
Buildings containing materials
storage which emits pollutants
while stored, e.g., synthetic
fertilizer, odors from render-
ing.
Buildings containing a variety
of materials handling
and processing operations.
Can be applied to a wide range
of relatively small operations
such as welding, solvent use.
Can be used for measuring
leaks from piping, packing
glands, etc.
This method where applicable
can provide quite accurate
results.
Must be applied when two
previous strategies cannot be
used.
Applicable where fugitive
emissions of the same type
are emitted from several areas
of a plant.
- 25%-50%
- 25%
Applicable where the preceeding
three strategies cannot be
used alone.
- 50%
- 200%
2. Detailed micro-meteorological
data are required.
3. Field measurements must be carried
out under specified meteorological
conditions.
Applicable where various types
of fugitive emissions are
present and where background
pollutants influence
meas ur emen t s
-11-
-------
3.4 Cost-Effectiveness
The cost-effectiveness of each sampling strategy when applied to a
specific source varied considerably. For different sources, different
optimum sampling strategies could be found. It was found that cost-
benefit tradeoffs could be examined for each of the three fugitive emis-
sion sampling strategies by developing error limits vs. cost curves for
selected fugitive emission sources.
3.5 Quasi-Stack Strategy
The quasi-stack sampling strategy was identified as being widely
applicable to many significant industrial fugitive emission sources,
and as an excellent candidate for field tests. It has generally accept-
able accuracy limits and in many cases can be made cost-effective. It
was found to be a valid method for the study of mold pouring emissions
provided that continuous analysis methods are used and certain other
criteria are met. These other criteria include:
1. Background pollutant concentrations are separable from the
operation studied.
2. Physical location of sampling allows proper positioning of re-
quired equipment and isolation of the test area from other
plant operations.
3. Sampling times where significant fugitive emissions will occur
are compatible with instrument response times.
4. Sample concentrations are within the analysis capability of
the instruments.
5. The quasi-stack configuration can be designed to capture all
of the fugitive emissions to be studied without serving as a
general ventilation hood for all in-plant pollutants, or other-
wise biasing pollutant concentrations or characteristics.
-12-
-------
3.6 Alternative Strategy
As an alternative to the quasi-stack sampling strategy in an iron
foundry or similar situation, a roof monitor sampling approach would be
applicable under certain conditions, which include:
1. High background pollutant concentrations prevent separation
of any single pouring or other operation.
2. Physical limitations prevent setup or isolation of a quasi-
stack sampling area.
3. Response times of instruments prevent capture or display of
single event fugitive emissions.
4. Sample concentrations from single events are below the instru-
ment threshold.
5. The quasi-stack configuration cannot be used for emission cap-
ture and the building structure is well enclosed and has well
defined vents.
3.7 Method Limitations
Standard particulate measurement techniques (time-averaging meth-
ods such as filter collection) may be unsuitable for fugitive emission
studies, especially where emissions are transient in nature and release
is over a short time duration.
-------
4.0 RECOMMENDATIONS
TRC, on the basis of its studies of fugitive emissions under EPA
Contract 68-02-1815, makes the following recommendations:
4.1 Industry Studies
EPA should consider further studies of the fugitive emissions from
some of the industries listed in Table 3-1. In particular, the indus-
tries most in need of study would be:
1. Primary copper
2. Coke making
3. Electric furnace steel
4. Petroleum refining
5. Coal gasification
6. Iron and steel foundries (continue and expand work begun in
this study)
This listing is based upon both the mass rate and the hazardous
nature of their fugitive emissions. In particular, the fugitive emis-
sion potential of a significant new energy process, coal gasification,
should be examined in detail. The impact of this industry on air qual-
ity in the future could be significant, and an important part of this
impact could be the fugitive emissions.
4.2 Strategy Tests
EPA should fund the study of the roof monitor and upwind-downwind
sampling strategies for fugitive emission measurement. A comparison
of all three major strategies could then be made after field test of
these fugitive emission sampling methods.
-14-
-------
4.3 New Methods
EPA should consider funding of research in the application of re-
mote sensors for the study of fugitive emission sources. For some po-
tentially significant sources, the three major strategies are inade-
quate. This research might be conducted in parallel with field tests
of either the roof-monitor or upwind-downwind strategies to attempt
correlation.
4.4 Documentation
EPA should develop a fugitive emissions measurement handbook. This
would outline sampling strategies available to the investigator, includ-
ing a basic methodology and procedures. By use of a multi-volume for-
mat, annual updates would be possible as technology advances.
4.5 Package Development
EPA should develop instrumentation packages designed to measure
fugitive emissions. Although they may not be adopted as standard EPA
methods, they could be presented as EPA-recommended methods. This would
help standardize fugitive emission measurements and aid in the develop-
ment of correlations between investigators, including better estimates
of relative and absolute errors. These may include remote sensing methods.
4.6 Impact Determination
EPA should consider the special needs of fugitive emission impact
estimation procedures. Since present stack plume models may need modification
-15-
-------
to be used for the simulation of the impact on ambient air of fugitive
emissions, a comprehensive study program may be required in this area.
TRC studies of the impact of fugitive sources on ambient air quality
has shown that estimation of the source strength, source height, and
diffusion rate of the fugitive pollutants from open sources, open vent/
roof monitor sources, or even short stack sources can be considerably
in error unless examined in some detail and appropriate modifications
made in the diffusion models. This would aid EPA efforts in many areas
of study, where fugitive emission impacts up to now have not been con-
sidered.
4.7 Refinement of Quasi-Stack
EPA should sponsor research to refine quasi-stack fugitive emis-
sion measurement methods. The limitations found by TRC in its field
studies, especially with regard to time intervals of fugitive emis-
sions and their variability and isolability, point out areas for fur-
ther development. Extensive testing of various continuous monitors,
especially for particulate matter mass rate and particle size, may be
required.
-16-
-------
5.0 DISCUSSION
The overall objective of the program was to develop methods for the
measurement of fugitive emissions of air pollutants from several indus-
trial processes. Fugitive emissions, for the purposes of this study,-
were defined as those emissions of air pollutants which are not directed
through ducts or stacks and which are not amenable to measurement by es-
tablished source sampling methods. The methods developed under this pro-
gram were to be adequate for quantifying fugitive emissions within rea-
sonable and calculable error limits and capable of use for assess-
ing measures designed to control fugitive emissions.
A five task program to accomplish this was conducted as follows:
5.1 Task I - Identify the Sources of Fugitive Emissions
In order to identify possible industrial sources of fugitive emis-
sions, a four part program was conducted with the following results:
5.1.1 Literature Search for Sources
A bibliography of references specifically directed to fugitive emis-
sions was developed and is presented in Appendix A. The key documents
used to develop the fugitive emission data for industries believed to
have significant fugitive emission potential are included. A substan-
tial number of the key documents were EPA research project reports
wherein overall process emissions and in some cases fugitive emissions
were quantified. Only recently have any significant number of fugitive
emission studies been reported in the open literature.
-17-
-------
Another source of fugitive emission information was TRC in-house in-
formation developed during studies for industrial and governmental clients.
Table 5-1 summarizes these. Because of their confidential nature (with
a few exceptions), it was not possible to cite these in the bibliography.
They are a substantial body of information, especially in regard to fu-
gitive emission measurement methods.
5.1.2 Trade Association/Industry/Government Agency Contacts
Six trade associations and two companies were contacted to obtain
information concerning fugitive emissions in industries of potential in-
terest. Four of these contacts have supplied data which will be useful
in identifying and quantifying fugitive emissions. Six air pollution
control agencies were contacted but no "hard" data were obtained from
them. Appendix B summarizes all the contacts which were made. The con-
tacts made did have an appreciation for the fugitive emission potential
of the industrial processes with which they were familiar. However, in
common with many others, they had for the most part not conducted studies of
fugitive emissions because of the priorities which have had to be given
to the characterization and control of stack emissions. One exception
to this was the contact with Dr. Fred Templeton, Assistant to the Presi-
dent of the Metal Mining Division of Kennecott Copper. Dr. Templeton sup-
plied EPA and TRC with fugitive emission data developed by Dr. John
Heaney of the Utah Copper Division of Kennecott Copper. These data, in
conjunction with those available from other sources, were used to develop
a matrix chart which describes most of the fugitive emission factors
for the copper smelting industry. Table 5-2 summarizes these emission
-18-
-------
TABLE 5-1
SUMMARY OF TRC STUDIES RELEVANT TO
FUGITIVE EMISSIONS MEASUREMENT
Client Category
1. Iron "and Steel Works
2. Inorganic Chemical
3. Steel Company
Coke Works
4. Municipal Health Department
(Report on Proposed Iron
& Steel Works)
5. Inorganic Chemical Company
6. ASHRAE
(Research Project TRP-93)
7. USPHS, DHEW, NAPCA
Contract CPA-22-69-14
8. Assembly Plant
Electric Components
9. EPA
Contract CPE-69-5
10. Cement Company
11. Cement Company
12. Sinter Plant in
Steel Mill
13. Iron Foundry
14. Rubber Manufacturing
15. Turbine Manufacturing
16. Rubber Manufacturing
17. Parking Garage
18. Gravel Pit
19. Coke Works
20. EPA
Contract 68-02-0047
Fugitive Emission Studies Done
Welding operations gas and particulate
emissions
Emissions of dust from ground, materials
handling and storage piles in plant.
Smoke emissions from leaking doors and coke
pushing and charging
Contribution of dust emissions from proposed
plant on local air quality. Similar plant
emission data and diffusion model used
for prediction
Dust emissions from uncontrolled vents and
from open windows of dry bleach plant
Use of air conditioning equipment to reduce
outdoor air pollutants drawn into buildings
Leakage of emissions inand out of structures
such as houses. Measurement of fugitive
emissions from stoves, etc.
Solvent emissions from coating operations
where concentrations v ary.
Study of emissions from disposal of solid
wastes
Upwind and downwind particulate emission
measurement to determine plant contribution
Upwind and downwind particulate emission
measurement plus diffusion modeling
Analysis of local air quality and contribu-
tion of sinter plant emissions to this.
Used up and downwind measurements.
Contribution to particulate emissions of
unpaved parking lot
Uncontrolled curing press emissions
measured
Fuel and Solvent transfer system emissions
Emissions inventory of uncontrolled sources
in tire plant
Dispersion modeling to determine CO con-
centrations at peak times
Air quality measurements of particulates
from fugitive and other sources
Air quality measurements in plant including
those near fugitive sources
Evaluation of particulate control strategies
which include fugitive emissions from
demolition
-19-
-------
TABLE 5-2
MATRIX CHART OF FUGITIVE EMISSION SOURCES
IN COPPER SMELTING
Process Step
1. Ore Concentrate
Storage and
Handling
2. Limestone Storage
and Handling
3. Slag Handling
4. Dust Collection
and Transfer
5. Roaster toading
and Operation
6. Calcine Transfer
7. Revcrberatory
Furnace Loading
and Operations
8. Matte Transfer
9. Converter
Loading and
Blowing
10. Blister Copper
Transfer
Particulate Fugitive Lmissions
Hnss Kate
1
10 lbs/ton(3)
Copper
1
i
1
1
2.0 Tons/
DayU)
1
*~5 ft
Dayu
,
T>
Composition
(Annroximate)
Cu 28%(3) Si02 117.
Te 24% Other 5%
S 32%
CaO ~ 60%
Cu 0.5%^ Si02 38%
Fe 40% S 1.5%
Unknown
Cu 5%(4) Pb 18%
Zn 16% As 60%
Cd 0.5X 70%<2um
Unknown
Cu 5%(A) Pb 18%
Zn 16% As 602
Cd 0.5% 70%<2um
Cu 42%^ Si02 1%
Fe 32% S 25%
Cu 1%(A) Pb 50%
Zn 8% As 37%
Cd 4% 10%<2vra
Gaseous Fugitive Kralssions
Mass Rate
Unknown
Unknown
0.4 Ibs.S^
/Ton Slag
Unknown
i ,
5.2 Tons/
Day S02 W
)
/8.3 lbs.S(2)
1 /Ton of
| Matte
* (1)
35-40 Tons/
Day S02
/"Olbs.S^
1 /Ton Cu
Compos i tion
^Approximate)
Unknown
Unknown
Unknown
Unknown
SOx-5-6*/^ ,
SOx-5-15%k ' ;
Unknown
SOx-5-6%^)
SOx-1-3%^ ' '
Unknown
SOx-l-5-4%^J
SOx-14-19%l'J'
Unknown
Utah Copper Division Data Supplied by Kennecott Copper Corporation.
(2)
(3)
(4)
(5)
Noranda Mines Ltd. Data Supplied by Kennecott Copper Corporation.
"Systems Study for Control of Emissions - Primary Non-Ferrous Smelting Industry"
Arthur C. HcKce and Company NTIS Report No. PB 184 884.
Measurement of Sulfur Dioxide, Particulate and Trace Elements in Copper Smelter
Converter and Ronster Rcverberatory Gas Streams, KPA Report No. 650/2-74-11 by
Couljol Sy^tcnis Laboratory.
Air Pollution Control Technology and Costs in Nine Selected Areas, IGCI Report on
KPA Com rnrt 68-0? 0101
(6) Emission includes all these operations.
-20-
-------
factors.
Mr. Tom Keller of Bethlehem Steel contributed information on roof
monitor emission measurement techniques. Although not used as a samp-
ling strategy in the field test program, this information offered val-
uable guidelines for conducting such tests in the future.
5.1.3 Matrix Charts for Identification of Selected Industries
A list of possible industries and processes which might have fugi-
tive emissions was developed. To identify where these occur, a matrix
chart was evolved which lists vertically the industry or process studied
and horizontally the operations which make up these industries or pro-
cesses. These were provisions on the charts for symbols to characterize
the type of fugitive emission encountered. Appendix C shows these charts.
The completed matrix charts were used as a basis for selection of
industries which might involve significant fugitive emissions. A list
of 18 such industries was developed and the matrix charts of these indus-
tries were consolidated. These industries were reviewed in more depth
so that the list would be refined to identify the most serious candidate
industries for fugitive emissions testing. Tentative sampling strategies
concepts were developed for these 18. Table 5-3 identifies the indus-
tries and their major potential fugitive emission problems.
In the absence of "hard" data, mass rates of fugitive emissions
were estimated not in terms of absolute numbers but in terms of their
potential fraction of total air emissions. Where the potential frac-
tion was high (> 10 percent), the industry or process might be worthy
of further study. Industries with potential for release of low TLV and/
-21-
-------
TABLE 5-3
CLASSIFICATION OF INDUSTRIES FOR POSSIBLE FUTURE
STUDIES OF FUGITIVE EMISSIONS
Category
Metallurgical
Energy/Fuels
Chemical Products
Rock Products
Industry/Process
Primary Aluminum
Primary Copper
Electric Furnace Steel
Iron & Steel Foundries
Coke Making
Coal
Coal Gasification
COG (Char-nil-Gas)
Shale Oil
Petroleum Refining
Oil Production
Plastics
Tire & Rubber
Phosphate Fertilizer
Lime
Sand & Gravel
Asphalt Batching
Agricultural Opera-
tions
Major Fugitive Emissions
Fume, Fluorides, PNA's
Fume, S02» Dust
Fume, Kish, CO, Odors
Fume, Odors
Hydrocarbon Fumes, Odors,
CO, PNA's
Dust
Hydrocarbon Fumes, CO, PNA's
Hydrocarbon Fumes & Vapors,
CO, PNA's
Dust, Hydrocarbon Fumes
& Vapors, CO, PNA's
Hydrocarbon Vapors, Odors,
PNA's
Hydrocarbon Vapors, Odors
Hydrocarbon Vapors, Odors
Hydrocarbon Vapors, Odors
Dust, Fluorides, S02
Dust
Dust
Dust, Odors, PNA's
Dust, Odors
-22-
-------
or known carcinogenic materials were considered for further study. Highly
reactive or toxic fugitive emissions also help to place an industry on
this list.
The 18 industries selected for possible study could be reduced to
four categories (see Table 5-8): metallurgical, fuels/energy, chemical
products, and rock products. Only one industry escaped classification
in Table 5-3: agricultural operations. It was important, however, that
this category be studied, since its fugitive dust problems had signifi-
cant impact on measurements of particulate emissions in other industries
such as rock products. The state of Colorado, for example, cannot meet
particulate matter primary air quality standards because of agricultural
contributions.
5.1.4 Fugitive Emission Sampling Techniques and Strategies
A scanning of the literature on sampling techniques showed that few
new techniques have been developed for measurement of fugitive-type emis-
sions. Appendix D summarizes specific references found. One advance
was the development of a procedure for sampling emissions from roof moni-
tors. This procedure is more of a refinement of a technique previously
used rather than a new sampling technology. Much the same can be said
about the other measurements mentioned in the literature.
Our new development indicated that lidar (light detection and rang-
ing/laser radar) might be applicable to measurement of fugitive dust emis-
sions. It has two possible advantages: readings are instantaneous and a
large number of accurate mappings of fugitive dust cloud might be obtain-
able in the time needed for one such set of data from a standard particulate
sampling test array. There are two possible disadvantages: commercial
-23-
-------
models may not be readily available and we are not aware of any present
studies which seek to confirm direct applicability to the type or problems
encountered in fugitive dust measurements (most work has been done with
lidar used in stack plume mapping).
The 18 industries/processes of interest for possible fugitive emis-
sions studies were closely examined to see which sampling strategy might
be applied to each of the suspected fugitive emission sources. If more
than one strategy or a combination of strategies might be appropriate,
this was noted. An attempt was also made to rate the strategy as to its
applicability.
The sampling strategy/industry or process combinations considered
to be feasible are given in Table 5-4.
In all cases, a feasible sampling strategy was found for each in-
dustry/process of possible future interest for fugitive emission studies.
In some cases, however, there were serious limitations to the applicabil-
ity of the proposed sampling strategy and it is apparent that alternate
new or new combinations of older sampling strategies should be investi-
gated.
The most universally applicable sampling strategy for outdoor sources
was found to be the upwind-downwind technique. In 16 of the 18 indus-
tries/processes of interest, this strategy would adequately quantify most
of the major outdoor fugitive emissions encountered. There are limita-
tions in that tracers and/or diffusion modeling might have to be done in
conjunction with sampling in order to obtain accurate results. Elaborate
sampling systems may be required and considerable total sampling time may
be required for the definition of each fugitive source, which can be a
-24-
-------
TABLE 5-4
FEASIBLE SAMPLING STRATEGIES FOR INDUSTRIES/PROCESSES
SELECTED FOR POSSIBLE FUGITIVE EMISSIONS STUDIES
Industry/Process
Primary Aluminum
Primary Copper
Electric Furnace
Steel
Iron A steel
Foundries
Coke Making
Coal
Coal Gasification
COG
(Char-Oil-Gas)
Shale Oil
Petroleum Refining
Oil Production
Plastics
Tires & Rubber
Phosphate Fertilizer
Lime
Sand & Gravel
Asphalt Batching
Agricultural
Operations
Critique of Sampling Indoor
_Fu_gi tiye Emissigns
Roof'Monitor/Enclosure Vent
is ideal.
Roof Monitor/Enclosure Vent
limited in accuracy because
buildings too open. Quasi-
Stack may he used.
Roof Monitor/Enclosure Vent
plus Ouasi-^tack are ideal.
Roof Monitor/Enclosure "ent
plus Quasi-Stack are ideal.
None
None
Maybe indoors to some
extent, then Roof Monitor/
Enclosure Vent or Ouasi-
Stack would be applicable.
Maybe indoors to some
extent, then Roof Monitor/
Enclosure Vent or Quasi-
Stack would be applicable.
Maybe indoors to some
extent, then Roof Monitor/
Enclosure Vent or Quasi-
Stack would be apolicable.
None
None
Strategy For: Outdoor
Fugitive Emissions
Upwind-Downwind Acceptable
Upwind-Downwind requires
tracers and complicated
sampling.
Slaa expansion might require
nuasi-Stack. Upwind-Down-
wind acceptable.
upwind-Downwind has problems
but acceptable for open
buildings.
Dpwind-Pownwind requires
tracers and complicated
sampling. Quasi-Stack
may isolate individual
sources but there are
many of them.
Upwind-Downwind useful,
but fires in waste piles may
defy analysis. Quasi-Stack
may also be used to point
sources.
Same problems as Coke
Making.
Same problems as Coke
Making.
Same problems as Coke
Making.
Upwind-Downwind would aopear
to be feasible. Would be
difficult to isolate
individual sources, however
Emissions Inventory can be
used also. Ouasi-Stack can
apply to point sources.
Upwind-Downwind would
appear to be feasible.
Would be difficult to
isolate individual sources*
however. Emissions Inventory
can be used also.
Roof Monitor/Enclosure Vent Upwind-Downwind acceptable.
plus Emission Inventory would
be ideal combination.
Roof Monitor/Enclosure Vent Upwind-Downwind acceptable.
plus Emission Inventory would
be ideal combination.
Quasi-stack may be used
inside.
None
None
None
None
Upwind-Downwind limited
because of number of
sources involved. Consider
remote sensing.
Upwind-Downwind limited
because of number of
sources involved.
Consider remote sensing.
Upwind-Downwind limited
because of number of
sources involved. Consider
remote sensing.
Upwind-Downwind limited
because of number of sources
involved. Consider
remote sensing. Quasi-
stack for hydrocarbons.
Only methods are Upwind-
Downwind and Quasi-Stack.
-25-
-------
problem if many are present in one industry/process, and if operating
cycles are short.
The possibility of remote sensing which would allow short time in-
terval mapping of fugitive dust clouds or gas emissions should be con-
sidered as an alternate sample strategy. One example where this might
be a good choice would be at a copper smelter. Lidar could be used to
map the particulate matter fugitive dust clouds from storage piles and
building openings and a correlation spectrometer could be used to map
fugitive S(>2 emissions.
The other two strategies which have been considered have about equi-
valent applicability to fugitive emission sources. The quasi-stack and
roof monitor/enclosure vent strategies measure fugitive emissions which
can be directed into and vented through specific openings or enclosed
areas. This makes them ineffective in determining fugitive emissions from
outdoor sources, such as storage piles or process steps such as crushing
and grinding of gravel, lime, etc., where the operation is accomplished
in an open-sided structure or entirely in the absence of an enclosing
structure. However, for many metallurgical processes which occur inside
buildings, such as aluminum electrolysis or electric furnace steel mak-
ing, these techniques are ideal for measurement of fugitive emissions.
It can be concluded that, although adequate concepts of sampling
systems and analytical methods for fugitive emissions measurements can
be specified for all the industries currently of potential interest on
this project, there is obviously room for improvement. It appeared that
it would be advantageous to continue through Task II some additional lit-
erature search effort directed toward finding new fugitive emission sam-
-26-
-------
pling methods. It also appeared at this juncture that any of the three
strategies mentioned above would be a good choice for most significant
cases of fugitive emissions.
5.2 Task II - Evaluating Various Sampling Strategies
Task II was divided into a series of subtasks. These subtasks,
which followed in logical sequence from Task I, can be described as fol-
lows:
5.2.1 Evaluating 18 Selected Industries
The criteria used for development of the original 18 industry list
were:
1. Large number of fugitive emission sources
2. Large mass rate of fugitive emissions
3. Large number of plants extant in the United States
4. Industry found near urban areas.
The list was reviewed for the purpose of rating the industries for
overall fugitive emissions potential and also the steps within them which
might exhibit major fugitive emissions so that the list could be further
culled. The criteria applied were the above four plus the following:
5. Hazardous nature of gaseous fugitive emissions
6. Particle size of particulate fugitive emissions (respirability)
7. Hazardous nature of particulate fugitive emissions
8. Probability of control of fugitive emissions by new technology
in the immediate future.
A rating system was therefore used as a guide to aid in the selec-
tion of the final candidate industries, but this was not the only process
used to select these industries. It was not possible to rank the indus-
-27-
-------
tries/processes in direct numerical order of the seriousness of their
fugitive emissions. This was dictated by the nature of the fugitive emis-
sion sources, which makes it difficult to quantify the emission rate
and identify the components of that emission. Table 5-5 shows how the
ratings were arrived at for the 18 industries.
5.2.2 Suitability of Sampling Strategies
Another subtask involved evaluation of the suitability of the pro-
posed basic sampling strategies to measure fugitive emissions of the pro-
cess steps of the 18 industries where fugitive emissions were considered
significant. This was done by developing a matrix (Table 5-6, for exam-
ple), listing for each of the industries/processes the process steps be-
lieved to have significant fugitive emissions, the nature of the gaseous
and particulate fugitive emissions, and the basic sampling strategy most
applicable to the fugitive emissions of the step. A process flowsheet
was also developed or extracted from the literature. Figure 5-1 is an
example.
5.2.3 Evaluate Sampling Analysis Techniques
The procedures of subtasks 1 and 2 pointed out there existed sub-
stantial commonality in sampling and analysis requirements from industry
to industry. There was an obvious need to specify the best sampling/
analysis techniques which might be applied to the measurements of fugi-
tive emissions, but these techniques would also need to be compatible
with the technical requirements and the financial resources of each sam-
pling strategy to which they would be applied.
-28-
-------
TABLE 5-5
METHOD OF RATING HAZARD POTENTIAL OF
FUGITIVE EMISSIONS OF 18 SELECTED INDUSTRIES
Industry or
Process
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
Primary aluminum
Primary copper
Electric furnace
steel
Iron and steel
foundries
Coke making
Coal
Coal gasification
COG
Shale oil
Petroleum refining
Oil production
Plastics
Tire and rubber
Phosphate ferti-
lizer
Lime
Sand and gravel
Asphalt batching
Agricultural op-
erations
A
4
5
3
4
5
5
5
5
5
5
3
4
4
4
2
2
2
5
B
4
5
4
4
5
4
4
4
4
4
3
3
3
4
5
5
3
2
C
4
3
5
5
3
4
(4)
(2)
(2)
3
3
5
3
3
5
5
5
5
Criterion
D E
3
3
5
5
5
1
2
2
2
4
2
5
5
3
3
3
3
1
4
4
3
4
5
1
5
5
5
3
3
4
4
3
2
2
4
1
F
4
4
4
4
5
3
5
5
5
4
4
3
3
4
2
2
3
2
G
4
5
4
3
5
3
5
5
5
4
4
3
3
3
2
2
5
2
H
2
4
3
3
3
3
3
3
3
4
4
3
3
2
4
4
3
5
Average
3.6
4.1
3.9
4.0
4.5
3.0
4.1
3.9
3.9
3.9
3.3
3.8
3.5
3.3
3.1
3.1
3.5
2.9
( ) = Future Estimates
A Large number of fugitive emission sources
B Large mass rate of fugitive emissions
C Large number of plants extant in the United States
D Industry found near urban areas
E Hazardous nature of gaseous fugitive emissions
F Particle size (respirability) of particulate fugitive emission
G Hazardous nature of particulate fugitive emission
H Probability of control of fugitive emissions by new technology
in the immediate future
(Highest number rates worst)
-29-
-------
TABLE 5-6
MATRIX OF IRON AND STEEL FOUNDRIES
FUGUTIVE EMISSION SOURCE/SAMPLING
STRATEGY COMBINATIONS
Operation
Coke, silica, sin-
ter, scrap deliv-
ery and storage
Furnace or cupola*
charging
felting
Tap and pour
slagging
fold pouring*
'roduct finishing
'ollution control
equipment dust
transfer and
storage
Particulate
Dust**
Fume**
Carbon-dust**
Smoke (oil)**
Fume**
Dust**
Fume**
Dust**
Dus t**
Dust**
Gas and Vapor
Emissions
CO**
Hydrocarbons
CO**
Hydrocarbons**
(odors)
CO,** S02,
H2S
Hydrocarbons**
(odors)
CO,** PNA**
S02, CO
Sampling
Strategy
Upwind-downwind
or Quasi-stack
Quasi-stack or
Roof Monitor
Quasi-stack or
Roof Monitor
Quasi-stack
Quasi-stack
Quasi-stack
Upwind-downwind
or Quasi-stack
*Most critical steps for fugitive emissions
**Most significant fugitive emissions
-30-
-------
Binder
Dust and gases
Return
sand
participate
emissions
Y?
Cores W i
Dust
FINISHING
AND
SHIPPING
CASTING
SHAKEOUT
COOLING AND
CLEANING
Core sand
and binder
SAND
PREPARATION
CORE
MAKING
Figure 5-1: Simplified Flow Sheet for
Gray Iron Foundry
-------
An in-depth evaluation or critique of certain sampling/analysis tech-
niques was therefore undertaken in the areas of:
1. Hydrocarbon analysis (gaseous and particulate)
2. Fluorides analysis (gaseous and particulate)
3. Total suspended particulates analysis (size distribution,
chemical nature, continuous analysis)
4. Sulfur compounds analysis (identification, continuous analysis)
5. Remote sensing (particulates or gas, continuous analysis)
One of these critiques is included as Appendix E. These critiques
aided in evaluation of sampling strategies and they were also very im-
portant in the preparation of a suitable field test procedure in Task
III.
5.2.4 Selection of Sampling Strategy/Industry Combinations for
Further Evaluation
An evaluation was made to establish the most common basic sampling
strategy/industry source combinations in the 18 industry list. This
was accomplished by first establishing a matrix of 23 strategy/source
combinations which would cover all or most of the significant fugitive
emission sources found on the 18 industry list. From this a determina-
tion was made of how many of these were related well enough to list
common combinations. This final matrix of four general combinations
which also provides the first step of Task III, is shown in Table 5-7.
It includes some additional elements which relate to Task III.
This list of four was jointly reviewed by TRC and EPA, and led to
the choice for preparation of a test procedure and field demonstration
of a quasi-stack sampling strategy.
-32-
-------
TABLE 5-7
MATRIX OF FINAL SELECTIONS OF FUGITIVE EMISSIONS
SAMPLING STRATEGY/INDUSTRY SOURCE COMBINATIONS
I
u>
Fugitive Source
Sampling Strategy
1 , Upwind-Downwind
2. Roof Monitor
Industry Process Step Other Industries
Applied To Involved Where Applicable
Specific
Coal Reactors Charge Coal Gasification
Gasification & Discharge Petroleum Refining
& Storage Oil Production
Of Coal Shale Oil Product
Char-Oil-Gas Prodn.
Asphalt Batching
General
Coal Preparation
Phosphate Fert.
Lime
Sand & Gravel
Primary Potroom " „ sPeciflc
/luminum Primary Aluminum
Primary Copper
Elec. Furnace Steel
Iron & Steel Found.
General
Phosphate Fert .
Plastics
Tire & Rubber
Has:
Med.
Med.
Med.
Med.
Med.
High
Med.
High
Med.
Med.
Fugitive
s Rate
to High
to High
to High
to High
to High
to High
to High
to High
Emissions
Hazardous Nature
PNA's. Reactive Hydrocarbons
PNA's "
"
1
i ii
PNA's only
Fluorides, PNA' st Small Metallic Particulate
SO , Small Metallic Particulate
Fluorides, Small Metallic Particulate
PNA.s, Small Metallic Particulate
3. Quasi-Stack
Iron & Steel
Foundries
Cupola
„ ,
Found'
EUc_ Furnace gteel
Primary Copper
Primary Aluminum
General
Med. to High ReactiveLjHydrocarbons^ Metal Parji^cuJ.ate
Med. to High " " " "
Very High S0?, Metal Particulate of Small Size
Med. to High PNA's, Metal Particulate of Small Size
4. Quasi-Stack
Iron & Steel
Foundries
Mold
Pouring
Asphalt Batching
Phosphate Fert.
Plastics
Tire & Rubber
Lime
Coal Gasification
Coke Making
Shale Oil Prodn.
Specific
Iron & Steel Found.
Elec. Furnace Steel
Primary Copper
Primary Aluminum
General
Asphalt Batching
Phosphate Fert.
Plastics
Tire & Rubber
Lime
Coal Gasification
Coke Making
Shale Oil Prodn.
Medium PNA' s, Hydrocarbons L Metal Parjticula^tes^ o£ Small Size
Medium Metal Particulates of Small Size
Medium S0_ Metal Particulates of Small Size
Medium PNS's, Metal Particulates of Small Size
Nature of Fugitive
Em is sions Measured
Particulates:
A. Size Dist.
B. Mass
C . Compos it ion
D. Organics (PNA)
Hydrocarbons
CO
Sulfur Gases
Phenol s
Pyridines
Particulates:
A. Size Dist.
B. Mass
C. Composition
E. Organics (PNA)
Hydrocarbons
CO
Sulfur Gases
Fluorides
A. Gas
B. Particulate
Particulates:
A. Size Dist.
B. Mass
C. Composition
D. Organics (PNA)
Hydrocarbons
CO
Sulfur Gases
Fluorides
A. Gas
B. Particulate
Particulates:
A. Size Dist.
B. Mass
C. Composition
D. Organics (PNA)
Hydrocarbons
CO
Sulfur Gases
Fluorides
A. Gas
B. Particulate
Odors
-------
5.2.5 Cost-Effectiveness of Sampling Strategies
The four final matrix sampling strategy/industry source combinations
were examined to estimate their cost-effectiveness. The basic procedure
was:
1. Determine accuracy and reproducibility of the techniques.
2. Determine minimum experiments required to achieve the above.
3. Compare the four techniques under the best conditions for which
they can be employed.
4. Establish matrix of strategy, industry in which employed, fugi-
tive emissions measured, costs, accuracy, and reproducibility.
5. Develop cost curves for each strategy and plot against the rela-
tive error limits (error limit depends on the number of experi-
ments run) .
Figure 5-2a shows such a cost curve. It was developed from the ba-
sic costs of experiments plotted against the relative error. It was as-
sumed that the two quasi-stack sampling strategies were equivalent in
costs.
This procedure results in relative cost-effectiveness curves for
different sources of fugitive emissions, whereas to compare each of the
strategies to'each other, either a weighting factor for the importance
of the fugitive emissions measured needs to be applied, or another pro-
cedure must be used to determine the cost-effectiveness for the same fu-
gitive emission source.
The following alternate procedure for evaluating all strategies on
the same basis was used:
1. Determine accuracy and reproducibility of the techniques.
2. Determine minimum experiments required to achieve this.
-34-
-------
300
i
u>
Ui
O
O
(O
JO
2
'5
•o
S
a
0)
cc
200
I
I
I
O
I
I
I
I I
I I
"\ I
\ I
I
I
A
1
I
Best effort
— Least cost
\
O
Upwind - downwind
(Coal gasification)
I
^T— Quasi - stack
• Q (Iron and
• I ctaal frtunrli*
100
L I
\
I
\
o
\
i
steel foundry)
Pouring or charging
Roof - monitor
(Aluminum pot room)
I
I
Best effort
— — Least cost
Roof
Monitor
Quasi - stack
(Pouring and charging)
I
I
50,000 100,000 150,000 200,000 250,000
Dollar cost
Cost-effectiveness of four fugitive emission source
sampling strategies applied to different sources.
50,000 100,000 150,000 200,000 250,000
Dollar cost
B
Cost-effectiveness of four fugitive emission source
sampling strategies applied to the same source.
(2 converters in one copper smelter building).
Figure 5-2
-------
3. Compare results of use of all four strategies in same industry
or process (e.g., copper smelter converter house).
4. Determine cost vs. accuracy and reproducibility for each.
5. Plot cost vs. reproducibility for each strategy (Figure 5-2b)
by varying the number of experiments and thereby changing the
relative error limits of tests and the costs.
This alternative procedure clearly shows that the cost-effectiveness
of a fugitive emission source sampling strategy is dependent on the source
to be tested. It can, for instance, be shown that a quasi-stack strategy
is most cost effective when a single source such as iron cupola charging
is studied, whereas in a copper smelter converter building (multiple en-
closed sources) a roof monitor strategy might be preferred. In a coal
gasification plant, the upwind-downwind strategy would probably be the
choice.
5.3 Task III - Prepare a Technical Manual for the Measurement of Fugitive
Emissions
A draft of a field test procedure for mold pouring in an iron foun-
dry was prepared.
5.4 Task IV - Field Test of a Fugitive Emission Technical Manual
Task IV consisted of the field evaluation of a procedure for the
quasi-stack sampling method for industrial fugitive emissions, which
was developed during Task III of the program. The procedures document
presents general guidelines for the design and conduct of a field test
program utilizing the quasi-stack method. The field evaluation was in-
tended to gather relevant information for refinement of the procedures
document based on actual field operating experience.
-36-
-------
As a result of an extensive literature study executed during Tasks
I and II which dealt with the identification of sources of fugitive
emissions and the evaluation of sampling strategies, and further re-
view by TRC and EPA, it was decided to evaluate the quasi-stack samp-
ling method on fugitive emissions. As an example which was considered
feasible, mold pouring in an iron foundry was chosen for the field study.
Five foundries which were cooperative with TRC upon initial contact
were inspected for suitability for these tests. A foundry located in
southern Connecticut was selected.
This foundry operates one cupola furnace which has a melting rate
of about 12 tons (109 kgs) per hour. Most of the production occurs in a
large main bay area, where molds are assembled in place and remain sta-
tionary during pouring and cooling. Most of these molds contain sand
cores. An overhead crane is used to transport the pouring ladle around
the bay and to deliver the molds to the sand shakeout operation.
It was possible for the field test setup to be located at one end
of the main bay where it was isolated from much of the foundry activ-
ity. Operation of the cupola, manufacture of the sand cores, mold
shakeout, product cleaning, and a large part of the mold assembly ac-
tivity occurred either at the other end of the bay or in another part
of the foundry. A ventilation system was designed and installed to
capture emissions during the initial pouring of hot metal in the mold
and the subsequent cooling period. The system consisted of an enclo-
sure or hood for a test mold, a fan, and associated ductwork.
Pollutant emission measurements were made in ports located in a
-37-
-------
straight run of the ductwork. The following listing summarizes the
specific equipment employed in the test program along with the types
of pollutants measured:
1. IKOR Model 206 Continuous Particulate Monitor; dry filterable
particulate.
2. EPA Method 5 Particulate Train per Federal Register, August
17, 1971: dry filterable particulate, organic and inorganic
condensibles particle size, and particulate composition
(i.e., iron, silica, nitrate, and sulfate).
3. Glass Innovation (G.I.) Cascade Impactor with EPA Method 5
condensible train; dry filterable particulate, organic and
inorganic condensibles, and particle size distribution (aero-
dynamic diameter).
4. Beckman Model 109 Flame lonization Detector; total hydrocar-
bons (methane basis).
5. Intertech Model Infra-2 Non-dispersive Infra-red Analyzer;
carbon monoxide.
A trailer was located on the foundry bay next to the test setup.
This trailer served as an instrument and chemical field laboratory (ref-
erence Figure 5-3). In addition, high volume air sampling stations were
used to measure ambient air particulate concentrations in the foundry
main bay.
The test program for Task IV was conducted during the period from
March 7 to March 21, 1975. Twenty-three (23) individual mold pouring
tests were run as well as four background tests. While a wide variety
of mold types were encountered, the mix was considered representative
of a normal production schedule for this foundry.
This report presents the results of the measurements and discusses
their significance, both with respect to the foundry operation and the
quasi-stack measurement method employed. The measurement and analysis
procedures are also presented in detail.
The purpose of the field test program was to evaluate the quasi-
-38-
-------
vo
I
Particulate Measurement Devices
IKOR EPA CASCADE
IMPACTOR
Capture
hood
HC and CO line
Instruments
Figure 5-3: Illustration of test set-up for measuring fugitive emissions from
mold pouring in a gray iron foundry. (EPA Contract 68-02-1815)
-------
stack sampling strategy for measuring fugitive emissions (in this case,
mold pouring in an iron foundry) and to provide field experience for im-
proving the strategy's test procedures. Emission concentrations were
obtained during the evaluation and they were related to the process con-
ditions at the time of the test.
Table 5-8 presents a summary listing of the tests that were con-
ducted, the associated casting/core weights of the molds that were poured,
and the test methods that were used to measure the mold pouring emis-
sions. Table 5-9 summarizes foundry activity during tests.
5.4.1 Summary of Procedures and Analysis
5.4.1.1 Description of the Mold Pouring Hood Arrangement
(Refer to Figure 5-3)
A hood was designed to enclose a single mold and capture most of
the emissions given off from the mold and ladle during the initial pour
and after the pour while the mold was cooling. After witnessing several
mold pouring operations on preliminary inspections, it was observed
that there was a significant evolution of hot exhaust gas from the
ladle alone. The enclosure design, therefore, had to take into account
both the ladle and mold configurations. Guidelines established by the
American Conference of Governmental Industrial Hygienists were used
to develop initial plans. Modifications to the design were made to
incorporate features dictated by specific plant operating practices.
Since mold and ladle transfer to and from the hood would be by over-
head crane, it was necessary to provide hinged side and top panels for
better access. This also provided a greater degree of operator safety
while handling the ladle during the pour. The final hood design, there-
-40-
-------
TABLE 5-8
SUMMARY OF TESTS - MOLD POURING
Test
No.
1
2
3
4
5
6
Bl
7
B2
8
9
10
11
12
13
14
15
16
17
B3
18
19
20
B4
21
22
23
Test
Date
4/7/75
4/7
4/7
4/8
4/8
4/8
4/9
4/9
4/9
4/10
4/10
4/10
4/14
4/14
4/14
4/15
4/15
4/16
4/16
4/17
4/17
4/17
4/17
4/21
4/21
4/21
4/21
Process Conditions
Casting Core
Wt. , LBS. ¥t., LBS.
800 250
300 275
1000 900
600 500
600 500
300 150
(Background Test-A.M.)
600 500
(Background Test-P.M.)
600 500
600 500
600 500
300 150
300 50
650 0
600 500
300 125
300 175
250 25
(Background Test-A.M.)
250 30
250 30
250 30
(Background Test-A.M.)
600 300
250 125
250 125
Test Methods
Particulate
EPA Cascade
Method 5 Impactor
j
/
/
T/
,/
/
/
/
/
y
/
T/
/
/
/
/
/
/
/
/
/
/
/
/
'
,
/
/
/
/
/
/
/
/
/
/
/
/
J
/
17
Continuous
Monitor
/
/
/
/
/
•/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
J
/
/
Other
)
> Hi-Vol1
j
J Ditto
)
)
> Ditto
j
Ditto
Ditto
T^-T *- t-n
t Ditto
-
-
Hi-Vol1
Tli ^t*n
Is ± L> i~D
Total HC
(CH4 Basis)
7
/
/
,/
/
/
/
/
'/
/
/
/
/
/
/
y
^
/
/
/
/
/
/
/
/
/
CO
y
/
J
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
/
^
Process Conditions
Casting
(kgs)
363
136
454
272
272
136
272
272
272
272
136
136
295
272
136
136
114
114
114
114
272
114
114
Core
(kgs)
114
125
408
227
227
68
227
227
227
227
68
23
0
227
56
80
11
14
14
14
136
56
56
1 Background tests run by Hi-Volume Sampler near test booth and in production area.
-------
TABLE 5-9
FOUNDRY ACTIVITY DURING TESTING
Test
No.
1
2,3
4
5
6
Bl
7
B2
8
9
10
11
12
13
14
15
16
17
B3
18
19
20
B4
21
22
23
Test
Pate
4/7/75
A/7/75
4/8/75
4/8/75
4/8/75
4/9/75
ft/9/75
4/9/75
4/10/75
4/10/75
4/10/75
4/14/75
4/14/75
4/14/75
4/15/75
4/15/75
4/16/75
4/16/75
4/17/75
4/17/75
4/17/75
4/17/75
4/21/75
4/21/75
It/21175
4/21/75
Test
Start
Time
11:58
1:12,2:18
12:07
1:29
2:49
10:25
12:06
2:11
12:05
12:49
1:45
11:57
1:05
2:01
11:59
1:25
12:08
2:06
10:08
12:44
1:49
2:37
12:12
12:05
1:09
2:10
Description of Foundry Activity During Testing
•
i 15 molds brought into booth area during tests.
Pouring and positioning molds near booth during
test.
Positioning molds near booth, no pouring.
Hold from test f!5 dropped 10 ft in front of
booth upon removal; sat smoking during test
#6, caught somewhat by draft of booth.
Booth background: no mold in booth, no produc-
tion or molds at trailer end of main bay. Ac-
tivity at Herman press.
Molds transferred and assembled in bay area in
front of booth during last half of test.
Pouring completed ^ 2:30. 16 molds poured near
booth .
Mold pouring on bay in front of booth 12:10-12:20.
Mold pouring on bay cupola Herman.
Flask transfer to Herman during run.
Pouring of molds near cupola during test. Opera-
tors and mold assemblers at lunch during test per-
iod - emissions from mold alone. Booth top in full
raised position during pour.
Mold assembly in front of booth during test. Both
roof ventilators over booth on during test. Pour-
ing at other end of bay.
Pouring molds in front of booth 2:00-2:10. No
other activity. Ventilators on, room clear.
Pouring on main bay from Herman to cupola, 12:00-
1:00. No mold pouring near booth 12:00-1:00. Roof
fans closed during first 15 minutes of test, then
opened. Mold assembly in front of booth 12:35-1:00.
Ventilators on over booth. Flasks next to booth
being transferred to Herman 1:20-1:45. Mold assem-
bly in front of booth 1:20-1:50.
Roof monitors on, mold assembler on break 12:00-
12:30. Pouring on main bay cupola to'Herman 12:10-
12:30, 12:30-1:00.
Molds in front of booth poured. No assembly occur-
ring. Roof fans on.
Mold preparation between Herman and cupola. Prepa-
ration near booth at 10:30. Roof fan on during last
15 minutes of test.
Roof monitors on. Mold assembly near booth during
test. Pouring at cupola end of bay 12:30-1:40.
Molds in front of Herman poured 1/2 hour before
test. Room very hazy ^ 1:15. 1:20 - cleaning Her-
man with air gun.
Mold pouring near booth started near end of test.
Roof vents on. No activity in this end of bay 1:45-
2:15. Pouring at other end of bay 1:45-. Molds
from this afternoon still smoking near booth.
Mold pouring in front of booth fron end of Test 19,
continuing until 2:45. Molds noticeably smoking
during latter part of test (on bay floor).
Mold assembly between cupola and Herman. Flask
transfer to Herman. Core cleaning with air hose in
front of booth ^ 10:25.
Roof vents on. Mold assembly break. Mold pouring
in front of Herman and booth 12:05-12:25. Pouring
at other end of bay, 12:25-test end.
Roof vents on. Mold assembly in front of booth.
One row of molds smoking from pour 12:05-12:25.
Flask transfer to Herman next to booth during test.
Atmosphere very hazy. 2:10 mold blowout next to
trailer. Mold pouring of second rwo in front of
booth 2:10-.
-42-
-------
fore, was functional from the standpoint of plant operating practices,
provided ventilation characteristics sufficient to capture the mold
pouring emissions, and incorporated physical dimensions dictated by
the sizes and configurations of the mold and ladle.
A two-foot (0.65 meter) diameter duct served as the exhaust duct-
work for the hood and as the sampling test section. This duct discharged
from the top of the hood and, by the use of three 90° elbows, was ar-
ranged to run horizontally, three feet above the floor. The horizontal
section which incorporated a continuous run of about 22 feet (7 meters),
contained the sample ports and test equipment probes.
The front opening of the hood was designed for a face velocity of
150 feet (46 meters) per minute (fpm). This ensured capture of most of
the exhaust gases from mold pouring, particularly from the ladle when
in the pour position. Based on an open face area of 57 square feet (5.4
square meters) for the hood and a design capture velocity of 150 fpm
(46 meters/minute), the required exhaust volume flow is 8500 cubic feet
(240 cubic meters) per minute (cfm). This exhaust flow in the two-foot
(0.65 meter) diamter duct represents a velocity of 2500-3000 fpm (770 925
meters/minute) which is considered adequate to insure that deposition
of particulate by settling will not occur in the duct. The fan for this
system was located at the end of the horizontal duct run. The exhaust
flow was returned directly to the bay. This was considered necessary
to prevent the hood enclosure from causing room air changeover and
thereby affecting the background concentrations of contaminants which
may have resulted if discharge was to the outside of the foundry struc-
ture. (The location of this hood in the foundry and the quantity of
-43-
-------
air moved would have made this system one of the largest exhaust sys-
tems in the bay area.) Figure 5-3 schematically depicts the mold pour-
ing hood arrangement, including associated ductwork, sample equipment
locations, and field laboratory trailer. More detailed descriptions of
the trailer and sampling equipment are presented in other sections.
5.4.1.2 Description of Sample Port Locations
The location of four sample test ports in the horizontal duct is
shown in Figure 5-4. Pertinent dimensions and the arrangement of the
sample equipment are also depicted in this figure.
The placement of the three side ports was chosen to minimize tur-
bulence effects. Disturbances to the normal flow pattern were expected
on the downstream side of each probe; therefore the actual port loca-
tions had to be arranged so that any one probe was not in the turbu-
lent wake of the preceding probe.
The IKOR continuous particulate monitor probe was located in the
furthest upstream port. This probe was positioned at a point nine
inches (23 cm) from the far side, which was determined from a velo-
city traverse to represent a point of average velocity. The IKOR probe
was not traversed during the test program.
The probe for an EPA Method 5 particulate train including a con-
densible collection section was located in the middle port. A six
point horizontal traverse was conducted for each test. Figure 5-4
gives the position of the traverse points with respect to the duct
diameter.
A Glass Innovation (G.I.) Cascade Impactor System coupled to a
-44-
-------
12"
«-~| A
Cascade Impactor
train
Gas sample probe
TOP VIEW
IKOR
train
EPA
train
Fan
3'
2'-.
12'
Elbow
FRONT VIEW
EPA traverse
point location
123 4 5 6
SECTION A-A
Distance from
Point back wall (inches)
1 1
2 3.5
3 7
4 17
5 20.5
6 23
Figure 5-4: Schematic Diagram of
Sample Port Locations
-45-
-------
condensible train Identical to the EPA Method 5 train was located in
the third port. As in the case of the IKOR, the impactor probe was
positioned at a point nine inches (23 cm) from the rear wall and was
not traversed during the tests. Although downstream of other particu-
late measurement devices, no significant effect on the mass emissions
as measured by the impactor was seen from its placement in the wake of
other instrument probes.
A gas sampling probe was placed in the vertical port situated at
the same location as the G.I. cascade impactor and inserted a distance
of about nine inches (23 cm). Twenty-five feet (7.7 meters) of tygon
sample line carried a sample flow from this probe to a fan-manifold
system contained in the laboratory trailer. This gas flow provided gas
samples for the Beckman hydrocarbon analyzer and the Intertech carbon
monoxide analyzer which were operated in the trailer.
The placement of the three sampling train probes and the one gas-
eous sampling probe in the same duct section was designed so that all
equipment could be operated simultaneously during each test. This al-
lowed for comparisons among sampling techniques.
5.4.1.3 Field-operated Instrument and Chemical Laboratory
A utility trailer was located inside the foundry close to the
hood ducting near the fan discharge. It was employed to provide a
clean environment within the foundry for sample preparation and re-
covery and for the operation of the hydrocarbon and carbon monoxide in-
strumentation and chart recorders.
A manifold system consisting of two-inch (5 cm) diameter glass
piping and an exhaust blower was used to collect the gaseous sample
-46-
-------
from the vertical sampling probe and tygon tubing. This sytem ex-
hausted excess flow out of the trailer through flexible ducting on the
blower discharge, while the hydrocarbon and carbon monoxide analyzers
extracted independent samples on a continuous basis from the glass pip-
ing. Each instrument extracted a small portion of the total sample
flow and only as much as it needed for analysis.
A strip chart recorder with a three channel output was set up in
the trailer to present continuous traces for the hydrocarbon analyzer,
carbon monoxide analyzer, and for the IKOR continuous particulate moni-
tor. Although the IKOR output was continuously monitored on the strip
chart recorder in the trailer, the instrument itself was operated at
the test port location, as mentioned previously.
The trailer, in addition to serving as an instrument laboratory,
was important in functioning as a field chemical laboratory. Having
the availability of a relatively clean environment located close to
the test setup increased assurance that the integrity of the samples
would be maintained. A procedure was adopted and followed (discussed
in more detail in Section 5.4.2) for handling samples in the field.
All sample mediums, i.e., filters, were handled and assembled in the
trailer prior to each test. At the end of each test all collected
samples were again handled and recovered inside the trailer. Samples
Were placed in suitable containers and logged properly. Some sample
preparation steps were done in the laboratory trailer that would nor-
mally have been done at TRC's permanent facilities. Filters were stored
in a dessicator after removal from their holders and chemical extractions
of the liquid medium in the condensible trains were performed, yield-
ing separate organic and inorganic fractions. These steps aided in
shortening the overall turnaround time for sample analysis and there-
-------
by permitting the test team to evaluate data at least on a preliminary
basis within only a few days. This procedure would indicate or suggest
the need for modification to planned procedures as the program pro-
gressed.
At the end of each test day, all samples were assembled in the
trailer, prepared for transport to TRC's Wethersfield, Connecticut,
permanent laboratory, and logged. As will be discussed later, samples
were delivered to the office laboratory on the day of their respective
tests and were immediately prepared for subsequent analysis.
5.4.1.4 EPA Method 5 Particulate and Condensible Train
Particulate sampling was accomplished by means of the EPA collec-
tion train, Method 5, as described in the August 17, 1971, edition of
the Federal Register. This method permits the measurement of both dry
filterable particulate and condensible particulate material. It is
shown schematically in Figure 5-5 and consists of a nozzle, glass-lined
probe, filter, four impingers, vacuum pump, dry gas meter, and an ori-
fice flow meter. Following the probe, the gas stream impacts directly
on a 2-1/2 inch Millipore membrane filter supported on a coarse fritter-
glass disc in a glass filter holder. A membrane type filter was used
to permit microscopic particle size analysis and particulate composi-
tion determinations to be performed. An ice bath containing four im-
pingers was attached to the back end of the filter via a short section
of glass connectors. All condensible material in the sample stream was
collected here. The stack velocity pressure was measured using a cali-
brated type-S pitot tube and inclined manometer, while the stack temper-
ature was monitored by a thermocouple-potentiometer arrangement. A nom-
-48-
-------
Potentiometer
Duct wall
Thermocouple
Nozzle
Calibrated
orifice
oo
Main
Bypass valve
1
o
/
Leakless
pump
Figure 5-5: EPA Method 5 Particulate Sampling
Train Including Condensibles
-49-
-------
ograph rapidly determined the orifice pressure drop which is representa-
tive of the sampling flow rate required for any pitot velocity pressure
in order to maintain isokinetic sampling conditions. It is believed that
particle pick-up would be weight-biased when either sub or super-isokin-
etic sampling rates are employed. Generally, a range of within + 10
percent of isokinetic was an acceptable test. Sampling flow was adjusted
by means of the needle valve and bypass valve arrangement on the leakless
pump.
Test data recorded included test time, sample point location,
sample duration at each traverse point, velocity pitot pressure, stack
temperature and pressure, meter volume and temperature, and orifice
pressure drop (see Appendix F for an example data sheet) .
At the end of each test, the probe, filter, and impinger assembly
were disconnected and transferred intact to the field laboratory.
Samples were recovered and placed in the following containers:
Container #1 (plastic petri dish) - 2-1/2" (6.3 cm) membrane fil-
ter from glass holder.
Container //2 (glass jar) - acetone washings of probe and front
half of filter holder.
Container #3 (glass jar) - total impinger solution (after volume
recorded) and water wash of impinger.
Container /M (plastic bottle) - silica gel from the fourth impin-
ger.
Container #5 (glass jar) - acetone prewash of the probe before
start of next test.
A more detailed description of sample recovery, handling, and
analysis is presented in Section 5.4.2. A computer program developed
by TRC was used to calculate the duct emission levels of particulate
material in grains per actual and standard cubic foot, and the duct
-50-
-------
volume flow. Reference is made to the Federal Register, August 17,
1971, for the formulas used in performing the required computations.
5.4.1.5 G.I. Cascade Impactor and Condensible Train for Particulates
A Glass Innovation Model 226 multi-stage cascade impactor was op-
erated in its in-stack mode and attached directly to the front end of a
standard EPA Method 5 condensible train with only slight modifications.
A specially designed type-S pitot was used which allowed for a velocity
pressure reading to be taken at the impactor nozzle location. Sampling
was performed isokinetically and data was recorded in much the same
manner as for the EPA Method 5 train with the noteable exception that
the cascade train was not traversed during the course of the test. Ap-
pendix G presents an example data sheet for the cascade impactor.
The impactor unit, which is illustrated in Figure 5-6, consists of
a cone-shaped nozzle and six impactor stages followed by a built-in
backup filter stage. All parts are constructed of stainless steel.
The unit is designed to hold seven radially slotted glass fiber fil-
ters and one glass fiber backup filter as follows:
Filter Type Location
1
2
3
4
5
6
7
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Slotted
Top of Stage 0
Top of Stage 1
Top of Stage 2
Top of Stage 3
Top of Stage 4
Top of Stage 5
Bottom of Stage 5
8 Whole On backup filter support
The use of an impactor with filters for particle collection on
-51-
-------
EPA probe
Elbow adapter
\
Back-up
filter support
Stage
Nozzle
Type S pitot
Figure 5-6: GI Cascade Impactor Assembly
-52-
-------
each stage was chosen since it offered a lower tare weight (as opposed
to steel plate collection) and enhanced particle collection by minimiz-
ing reentrainment. The cascade impactor determines the aerodynamic
particle size distribution with particle size cutoffs ranging from .5
to 15 microns. The actual cutoffs obtained depend on the operating
conditions during a test as compared to the unit's calibration data.
Since this impactor is designed to capture all particulates, it is also
used to determine total filterable particulate emissions. As mentioned
previously, this train contains a condensible section identical to the
EPA Method 5 train so that condensible particulates are also measured.
At the end of each test the sampling train was partially discon-
nected with the impactor and impingers transferred intact to the field
laboratory. Samples were assigned containers as follows:
Containers #1-8 (plastic petri dishes) - eight impactor stage fil-
ters from the cascade device.
Container #9 (glass jar) - total impinger solution (after volume
recorded) and water wash of impingers.
Container #10 (plastic bottle) - silica gel from the fourth impin-
ger.
Probe wash and prewash procedures were not followed on the cascade
impactor train since the impactor (filter) was located at the front
of the sampling ahead of the probe. Section 5.4.2 presents additional
detail on the sample recovery, handling, and analysis procedures. The
duct emission of particulate in terms of grains per actual and
standard cubic foot was determined by TRC's computer program as
discussed in Section 5.4.1.4. The aerodynamic particle size distri-
-53-
-------
bution was obtained by applying the weight percentage of material cap-
tured on each stage to the mean particle diameter representative of
that stage.
5.4.1.6 IKOR Model 206 Continuous Particulate Monitor
An IKOR continuous particulate monitor, which was supplied by EPA's
Industrial Environmental Research Laboratory, was used to provide a quanti-
tative measurement and a real-time continuous display of mold pouring par-
ticulate emissions. The system, shown schematically in Figure 5-7, con-
sists of a probe, sensing unit, gravimetric filter unit, and a control
unit with signal integrator.
Particulate matter in the gas stream flowing through the probe col-
lides with an inline sensing element resulting in a transfer of charge be-
tween the particle and sensor. An electrical signal is generated which
is believed to be in response to the mass rate of particulate emissions
when the characteristics of the particles do not vary with time. A cali-
bration factor is obtained by developing a ratio between a simultaneous
gravimetric sample collected in the filter unit (142 mm glass fiber fil-
ter) and an integrated signal of the instantaneous output response.
This factor may then be used to convert subsequent integrated signals
into mass particulate emissions.
Sampling is conducted isokinetically through the one inch I.D.
probe. A specially designed venturi section along the probe length
provides a means for measuring the sample flow rate. Isokinetic sam-
pling conditions are attained by matching the pressure drops across the
venturi and a type-S pitot monitoring the duct velocity. The sample flow
is adjusted by a knob on the control unit which regulates a variable
-54-
-------
Signal integrator unit
Blower
discharge
Sensor unit
(contains sensor element,
gravimetric filter, and
blower unit)
Fig. 5-7 :IKOR continuous paniculate monitoring system.
-55-
-------
speed blower located downstream of the sensor element and gravimetric
filter unit.
Recorded data included the test time, sample velocity pressure,
stack velocity pressure, stack temperature, sensor temperature, inte-
grator parameters, and the integrated signal. An example data sheet
can be found in Appendix H. For those tests which include a filter
run (for calibration purposes) , the filter cartridge was removed and
transferred to the field laboratory. The recovered filter was placed
in a plastic petri dish. Handling and analysis of the filter is dis-
cussed in Section 5.4.2. Duct particulate emission levels in terms of
grains per actual and standard cubic foot were obtained by converting
integrated signal data through application of the calibration factor.
5.4.1.7 Beckman Model 109 Flame lonization Detector for Total
Hydrocarbons
Figure 5-8 shows the setup of the hydrocarbon flame ionization de-
tector which was located in the laboratory trailer. A continuous dis-
play of the output was provided by means of a chart recorder. The hy-
drocarbon analyzer abstracted a small portion of the total sample flow
from the gas sampling manifold system. Calibration of the hydrocarbon
analyzer was performed with a known concentration of methane gas in
nitrogen. The data output is therefore presented in terms of the cali-
brating methane gas (as parts per million
5.4.1.8 Intertech Model Infra-2 Non-dispersive Infrared
Analyzer for Carbon Monoxide
The carbon monoxide analyzer is shown schematically in Figure 5-8.
The measurement of carbon monoxide by non-dispersive infrared is the
-56-
-------
Blower
Sample manifold
(glass)
Supply air
Fuel supply
Intertech NDIR
carbon monoxide analyzer
Beckman Flame
lonization Detector
3 channel
chart recorder
Fig. 5-8 :Schematic diagram of laboratory trailer set-up showing hydro-
carbon and carbon monoxide analyzers.
-57-
-------
reference method acknowledged by EPA. Calibration of the carbon monox-
ide analyzer was made with a gas of a known carbon monoxide concentra-
tion. As in the case of the hydrocarbon analyzer, a sample was obtained
continuously from a glass sampling manifold and the analyzer output is
presented on a strip chart recorder.
5.4.1.9 High-volume Filter Sampler for Ambient Particulate
The measurement of ambient particulate concentrations at two loca-
tions on the foundry main bay was accomplished with high volume sampling
stations as illustrated in Figure 5-9- One sampling location was near
the booth enclosure while the other was in a production area at the op-
posite end of the bay, near the cupola.
A metal shelter encloses a filter-blower assembly in such a manner
that a measurement of suspended particulates is obtained. A high vol-
ume blower with a flow range of 20 to 60 cubic feet (0.5-1.7 cubic
meters) per minute draws sample into the shelter hood through slots at
the top. Particulates are captured on an 8 inch by 10 inch (20.3 by
25.4 cm) glass fiber filter attached to the front of the blower.
Sample flow through the unit is monitored by a calibrated rotameter
attached to the blower. Data collected during each test included sam-
ple time, duration of test and the initial and final sample flows. Ap-
pendix I contains a copy of a data sheet for the hi-vol ambient sampling
procedure. Filters were removed from the shelter at the end of each
test, folded, and placed in a manila envelope. Section 5.4.2 discusses
the procedures for handling and analyzing the filter samples.
-58-
-------
Filter
Hi-vol blower
Calibrated.
flow meter
Fig. 5-9: High volume sampling station for ambient particulates.
-59-
-------
5.4.2 Analytical Procedures
A quality assurance procedure was developed which was primarily
concerned with the proper methods for sample handling and analysis. A
prescribed chain of custody was assigned for samples taken in the field
and delivered to TRC's permanent laboratory facility in Wethersfield,
Connecticut. The following sections describe all aspects of the quality
assurance program including sample recovery and preparations in the
field, transport of samples form the field to the permanent laboratory,
analytical laboratory procedures, and data reporting.
5.4.2.1 Sample Recovery and Preparation in the Field
All sample media were handled in such a way that initial loading
and final recovery were conducted inside the field trailer, except for
the high-volume filters, which were loaded and recovered at the sampling
site. A chemical technician operating inside the trailer was responsi-
ble for receiving the samples from the test team, for placing the recov-
ered samples in assigned containers, for labeling and logging all sam-
ples, for conducting in-field analytical procedures, and for assembling
samples in a suitable manner for transport to TRC's laboratory. It was
considered essential that these activities be conducted in an enclosed
structure such as the trailer. During the handling procedure, the test
team remained outside the field laboratory, thereby minimizing undesir-
able interferences with the collected samples.
The filters from the EPA, cascade impactor, and IKOR trains were
removed from their respective holders and placed in plastic petri dishes.
The hi-vol filters were removed at the sampling station, folded, and
-60-
-------
transferred to a manila envelope. All filter samples were stored in s.
desiccator in the trailer during the day until preparations were made
for transport to the perinament laboratory. Desiccation in the field
shortened the time required in the laboratory for the filters to achieve
a. constant weight.
Silica gel samples were removed from the fourth impinger of both
the EPA and cascade trains and placed in their respective PVC plastic
containers. The silica gel samples from each test were transferred to
their original containers. All sample containers were tightly sealed
and put aside until they were transported.
The liquid fraction from the first and second impingers of the con-
densible train (EPA and cascade) was handled as follows:
1. Total liquid volume was measured at the end of the test in a
graduated cylinder. This volume was compared to the original
200 milliliters of distilled water placed in the impingers and
the net gain in volume recorded on the field test data sheet.
2. The liquid was transferred to a glass sample container. Rins-
ings of the impingers and graduated cylinder with distilled
water were added to the container.
Extractions were performed in the field on the liquid samples af-
ter recovery was complete. The samples were first extracted three times
with 25 milliliters of chloroform and then three times with 25 milli-
liters of ethyl ether. The chloroform and ether layers were combined
and transferred to a clean glass container. The container was then
stored on ice. The remaining water layer from the extraction was trans-
ferred to another glass container and set aside until assembled for
transport.
The EPA Method 5 probe was washed with acetone to rinse out any
-61-
-------
particulate matter adhering to the glass lining of the probe. One wash
(labeled a pre-wash) was conducted before each test and recorded on the
previous test's data sheet. A second probe wash was completed after
each test. The probe wash samples were collected in glass containers
and stored on ice.
Portions of the distilled water used in the condensible trains and
the acetone for the probe washes were taken daily and transferred to
glass containers. These samples served as blanks in the chemical analy-
sis.
At the end of each test day all samples were gathered together by
the chemical technician and prepared for transport to TRC permanent lab-
atory facility. The filters were removed from the desiccator and
sealed in the plastic petri dishes. Sample containers from the acetone
probe washings and chloroform-ether extractions were placed in an ice
chest. A log was prepared which identified the batch of samples being
transported. An example of this log appears in Appendix J. Samples
were delivered to the permanent laboratory by a member of the test team
on the day of their respective tests.
5.4.2.2 Analytical Laboratory Procedures
Immediately upon delivery to the analytical laboratory the samples
were unpacked and prepared by the test team technician for subsequent
analysis. The lids from the petri dishes were removed and the filters
(in the bottom half of the petri dish) were placed in a desiccator.
The high-volume filters were placed on trays in the desiccator after
being removed from the manila envelopes. Sample probe washes and con-
-62-
-------
densible extractions were transferred from their respective glass con-
tainers to tared dishes. These dishes were placed in an evaporator hood
to promote drying. The beaker containing the condensible water layer
was placed on top of a hot plate (temperature ^ 100°F [ 38°C] to induce
evaporation and shorten the drying time. A log was maintained to identify
samples for the analytical chemist performing the sample analysis.
The following is an outline of the procedures used by the analytical
laboratory:
Filter weight gain - filters were weighed on an analytical balance
twice a day and until a constant weight was attained. Weights were
reported to the nearest .01 milligrams.
Silica gel weight gain - silica gel was removed from the plastic
container and weighed on an analytical balance. The amount of
moisture collected was represented by the difference between this
value and an original silica gel weight of 200 grams. The results
were reported to the nearest .5 grams.
Particulate in probe wash - acetone was evaporated from the tared
beaker until a completely dry residue remained. The weight gain
of particulate in the probe wash, reported to the nearest .01 mil-
ligrams, was obtained from the difference in the final and tare
weights of the beaker as measured on an analytical balance.
Organic and inorganic condensible - the chloroform-ether and water
extraction beakers were evaporated until a dry residue remained.
The organic condensible particulate is represented by the residue
from the chloroform-ether extraction while the inorganic condensi-
ble particulate is residue from the water layer. The results, ob-
tained by a measured weight gain in the tared beakers, are reported
to the nearest .01 milligrams for both organic and inorganic par-
ticulate.
Microscopic particle sizing of EPA membrane filters - by assuming
an even distribution of particulate on the membrane filter, a pie-
shaped section was cut away. This piece was mounted on a viewing
slide and made transparent by a mounting solution. Dark phase mi-
croscopy was used to count the particles. Results were reported
in terms of the percentage of particulates by count in the ranges
of 3y.
Iron content of particulate on EPA membrane filter - a portion of
the membrane filter was dissolved in nitric acid to dissolve the
-63-
-------
iron. The distribution of iron and total particulate was assumed
the same for this portion as for the whole filter. The iron con-
centration of the known solution volume was determined by atomic
absorption. The results were presented in terms of the percent
iron content of the total particulate.
Quartz content of particulate on EPA membrane filter - nitric acid
was used, as for iron, to dissolve a portion of the filter. Hydro-
chloric acid was added to eliminate any metallic interferences.
Silicates are removed by a hot phosphoric acid digestion. The re-
maining quartz was dissolved with hydrofluoric acid. A color change
produced when ammonium molybdate was added was compared with stan-
dards of known concentration on a spectrophotometer. Results were
reported in terms of the percent quartz in the total particulate
collected.
Nitrate and sulfate content of particulate on EPA membrane filter -
A portion of the filter was refluxed for 90 minutes with water. The
sulfate concentration of an aliquot was determined by the turbidi-
metric method. Treatment of the sample with barium chloride froms
a barium sulfate suspension. The brucine method was used for nitrate
determination. A yellow color change occurs when brucine reacts
with nitrate while applying controlled heat. A spectrophotometer
was used to relate the color change to a nitrate concentration as
compared to standards.
After analysis was completed, all filters (except those destroyed
in iron, quartz, sulfate, and nitrate determinations), extraction resi-
dues, and probe wash residues were assembled together and prepared for
storage. Filters were returned to their respective petri dishes and
sealed. Sample residues were redissolved in the original solvent (i.e.,
water acetone, chloroform-ether) and transferred to clean glass contain-
ers. The solvent was permitted to evaporate again leaving the sample
residue. These containers were then sealed. Samples were assembled in
a large box and assigned a storage area. These samples will be retained
for possible future analyses until such time as their disposition is spe-
cified by EPA.
-64-
-------
5.4.3 Evaluation of Particulate Concentration Data
The particulate concentration data obtained in the gray iron
foundry tests were analyzed in detail. The low concentrations and the
similarity of the results obtained using the three (EPA Method 5, Cas-
cade Impactor, IKOR Continuous Monitor) different sampling methods made
it imperative that a complete statistical analysis of the data be done
to determine if there was:
A. Any significant difference between the results using the three
sampling methods.
B. Any significant difference between these results for mold pour-
ing fugitive emissions and background tests in the gray iron
foundry.
The data comparison can be summarized as follows:
A. None of the three particulate sampling methods gave statis-
tically significant difference in results when compared to in-
plant particulate background as measured by hi-vol filters.
B. The instantaneous particulate fugitive emissions can be in-
ferred from the sample traces from the IKOR online contin-
uous sampling. This is approximately an order of magnitude
higher than the average over the entire test period and the
general background concentration.
C. The period of high fugitive particulate concentration comes
at the beginning of the test and is of short duration. This
results in a low contribution of these actual fugitive emis-
sions to the total particulate concentration over the test
period (total = fugitive from pour + background).
D. Based upon C, it was considered that only gross estimates of
fugitive particulate concentrations could be obtained from the
data taken, since concentrations were low and near error limits
of measurement.
E. Based upon the quasi-stack configuration used, only the IKOR
continuous monitor was capable of sensing and measuring par-
ticulate fugitive emissions from the mold pouring operation.
This could be accomplished without a filter on the IKOR after
an emission factor has been established, or by accumulating
on a single filter a large number of short interval mold pour-
-65-
-------
ing tests (run IKOR only during peak particulate periods) in
the gray iron foundry and using the accumulated sample as an
average result.
Table 5-10 presents the particulate concentration data obtained us-
ing the three comparative methods:
1. EPA Method 5, including condensibles.
2. Cascade Impactor, including condensibles.
3. IKOR continuous mass monitor.
The duct background test results are included in the table and am-
bient background data are presented from hi-vol tests that were conducted
near the booth during mold pouring.
Mean values and standard deviations were calculated for all the tests
which were within 90 to 110 percent of true isokinetic, including back-
ground tests. Table 5-11 summarizes these calculations. These results
are for all isokinetic tests (all methods not necessarily run concurrent-
ly) . The same calculation was repeated for isokinetic tests and back-
ground tests when all of the tests methods were run concurrently. Table
5-12 summarizes the latter calculation.
One aspect of reviewing the sampling strategy was to compare the var-
ious test procedures for accurately monitoring fugitive emissions. This
was done statistically by conducting "t" and "f" tests on the mean val-
ues and standard deviations (Table 5-112) of the particulate results.
The "t" tests are presented as a confidence interval for the true dif-
ference between the averages of the various test methods for both process
emissions and background conditions. The "f" tests were conducted to con-
firm the results of the "t" tests. Intervals are reported for 95 or 99
percent confidence. If a 95 percent confidence interval is reported, no
comment regarding a significant difference can be made with a confidence
-66-
-------
TABLE 5-10
PARTICULATE CONCENTRATION DATA
GR/DSCF
Test
No.
1
2
3
4
5
6
Bl
7
B2
8
9
10
11
12
13
14
15
16
17
B3
18
19
20
B4
21
22
23
Test
Date
4/7/74
4/7
4/7
4/8
4/8
4/8
4/9
4/9
4/9
4/10
4/10
4/10
4/14
4/14
4/14
4/15
4/15
4/16
4/16
4/17
4/17
4/17
4/17
4/21
4/21
it/21
it/21
Process conditions
Casting Core wt.,
Wt., Ibs Ibs
800 250
300 275
1000 900
600 500 :
600 500
300 150
(Background Test-a.m.)
600 500
(Background Test-p.m.)
600 500
600 500
600 500
300 150 i
300 50
650 0 '
600 500
300 125
300 175
250 25
(Background Test-a.m.)
250 30
250 30
250 30
(Background Test-a.m.)
600 300
250 125
250 125
Test Method
EPA Method 5
Total
Part.
.020
.014
.011
.009
.016
.008
—
.014
.018
.006
.005
.006
.006
.017
.005
.007
.008
.008
.006
.008
.008
.005
.002
.007
.009
.003
.008
Condensibles
Total
.009
.005
.002
.006
.004
.007
—
.007
.013
.003
.004
.004
.004
.015
.004
.004
.007
.006
.004
.004
.002
.001
.002
.003
.005
.002
.003
Inorg
.003
.001
.001
.002
0
.002
—
.004
.005
.001
.002
.002
.002
.012
.002
.002
.005
.004
.002
.002
.001
.001
.002
0
0
0
0
Org
.006
.004
.001
.004
.004
.005
—
.003
.008
.002
.002
.002
.002
.003
.002
.002
.002
.002
.002
.002
.001
0
0
.003
.005
.002
.003
Probe
Hash
.006
.004
.004
.003
.009
.001
—
.004
.002
.003
0
.001
0
0
0
.002
.001
.001
0
.002
.005
.002
0
.001
.003
.001
.004
Filter
.005
.004
.005
0
.003
0
—
.003
.003
0
0
0
.002
.002
.001
.001
0
.001
.001
.001
.001
.002
0
.003
.001
0
.001
% Iso-
kinetic
68
60
62
61
62
61
—
58
61
101
99
98
100
101
102
104
100
102
100
100
100
103
104
99
103
102
103
Cascade Impactor
Total
Part.
.003
.006
.005
.012
.004
.010
—
.010
—
—
—
—
—
—
—
.002
.003
.003
.003
—
.001
.002
.001
—
.004
.007
.005
Condensibles
Total
.003
.006
.004
.008
.004
.009
—
.010
—
—
—
—
—
—
—
.002
.002
.003
.002
—
.001
.002
.001
—
.004
.005
.003
Inorg
0
.001
.001
.003
0
.005
—
.004
—
—
—
—
—
—
—
.001
.001
.002
.001
—
.001
.001
.001
—
.002
.002
0
Org
.003
.005
.003
.005
.004
.004
—
.006
—
—
—
—
—
—
—
.001
.001
.001
.001
—
0
.001
0
—
.002
.003
.003
Filter
0
0
.001
.004
0
.001
—
0
—
—
—
—
—
—
—
0
.001
0
.001
—
0
0
0
—
0
.001
.002
% Iso-
kinetic
102
98
102
67
58
67
—
33
—
—
—
—
—
—
—
95
97
97
103 ,
—
98
106
101
—
96
99
99
IKOR
Monitor:
Filter
—
—
.001
.001
.001
.0006
.002
.003
.0007
.0015
.002
.0007
.0012
.0013
.003
.002
.002
.0023
.0034
.0033
.0012
.0011
—
.0011
.0023
.0035
Hi-vol
Filter
(@ booth)
\
[ .003
)
.001
.0001
.0002
.0002
.0021
)
} .0008
j
.0024
—
—
—
\
> .0025
)
.0035
.0023
-------
TABLE- 5-U
SUMMARY OF STATISTICAL ANALYSIS
TOTAL OF ALL ISO-KINETIC TESTS FOR EACH METHOD
Test Method
EPA Method 5
(16 Tests)
Cascade Impactor
(13 Tests)
IKOR Monitor
(20 Tests)
Duct Background
EPA Method 5
(2 Tests only)
Duct Background
IKOR Monitor
(4 Tests)
Room Background
Hi-Vol Near Booth
(11 Tests)
Room Background
Hi-Vol in Production
Area. (7 Tests)
Particulate Concentration, Gr/DSCF
Total Part., Incl. Cond.
Mean, y
.0068
.0035
—
.0075
^^
— —
Std . Dev . , o
.0033
.0019
—
.0007
^_
~~
Total Condensibles
Mean . y
.0044
.0029
—
.0035
~~^
•••
Std. Dev. , 0
.0032
.0015
—
.0007
"
"
Part, on Filter (s)
Mean, y
.00231
.0005
.0018
.00351
.0025
.0019
.0036
Std. Dev.j^ a
.0018
.0007
.0009
.0007
.0013
.0011
.0025
CO
I
1 Includes Probe Wash. (Probe wash does not apply to other test methods.)
-------
TABLE 5-X2
SUMMARY OF STATISTICAL ANALYSIS
TOTAL OF 10 TESTS: ALL METHODS ISO-KINETIC AND
ALL METHODS WERE RUN CONCURRENTLY
Test Method
EPA Method 5
Cascade Impactor
IKOR Monitor
Duct Background
EPA Method 5 (2 tests
only)
Duct Background
IKOR Monitor (4 tests)
Room Background
Hi-Vol near booth (4
tests)
Room Background
Hi-Vol in production
area (3 tests only)
Particulate Concentration, Gr/DSCF
Total Part., Incl. Cond.
Mean , p
.0064
.0031
.0075
—
"™^
"
Std. Dev. , a
.0024
.0019
.0007
—
^_»
"
Total Condensibles
Mean , y
.0036
.0025
.0035
—
"•"""
Jim"1^
Std. Dev. , cf
.0020
.0013
.0007
—
__
-» —
Part, on Filters
Mean , y
.00271
.0005
.0023
.00351
.0025
.0027
.0052
Std. Dev. , a
.0020
.0007
.0010
.0007
.0013
.0006
.0029
vo
Includes probe wash. (Probe wash does not apply to other test methods.)
-------
greater than 95 percent. This statistical analysis is presented in Ta-
ble 5-13a for total particulate concentration, in Table 5-13b for total
condensible concentration, and in Table 5-13c for the solid particulate
concentration collected on the filter(s).
5.4.3.1 Background Tests
The three comparative test methods were run concurrently at isokin-
etic conditions for 10 tests. These 10 tests were conducted over a four
day period. During those four days a total of four hi-vol tests near the
booth for background testing spanned the production duration for mold
pouring. Several mold pouring emission tests were often conducted dur-
ing one hi-vol test. Duct background data was provided by two isokin-
etlc tests with the EPA Method 5 train. These tests occurred in the
morning prior to mold pouring.
5.4.3.2 EPA Method 5 Test Comparison
From Table 5-12 it can be observed that there was no statistically
significant difference between the duct tests by SPA Method 5 and the:
1. Duct background "reference" that was determined by EPA Method
5 and the continuous monitor in the duct.
2. Continuous monitor tests ("particulate on filter").
3. Room background "reference" that was established by the hi-vol
tests near the booth.
The EPA Method 5 duct results were higher than the cascade im-
pactor results. Compared to the cascade impactor, the following confi-
dence intervals were obtained for the difference in results:
-70-
-------
TABLE 5-13 (a)
DIFFERENCES IN TEST METHOD PERFORMANCES
TOTAL PARTICULATE CONCENTRATION, GR/DSCF
Comparison of Methods
Cascade Impact or
EPA
Method 5
to-
Duct Backgrounds
by EPA Method 5
Cascade Impactor
to-
Duct Background
by EPA Method 5
Statistical Analysis
"t" Tests
Interval
0.003310.0029
None
0.004410.0026
Confidence
99%
99%
99%
"f" Tests
"f" Ratio
12.0
0.6
10.1
Critical "f"
& % Confidence
8.3; 99%
10.0; 99%
10.0; 99%
Remarks
Significant dif-
ference confirmed
by "f" test.
No statistical
difference con-
firmed by "f"
test.
Significant dif-
ference of lower
Cascade results
confirmed by "f"
test.
-------
TABLE 5-13 (b)
DIFFERENCES IN TEST METHOD PERFORMANCES
TOTAL CONDENSIBLES CONCENTRATION, GR/DSCF
i
-o
N>
Comparison of Methods
Cascade Impactor
EPA Method 5
to -
Duct Background
by EPA Method 5
Cascade Impactor
to -
Duct Background
by EPA Method 5
Statistical Analysis
"t" Tests
Interval
None
None
None
Confidence
997.
99%
99%
"f" Tests
"f" Ratio
2.3
0.3
1.3
Critical "f"
& % Confidence
8.3; 99%
10.0; 99%
10.0; 99%
Remarks
No statistical
difference con-
firmed by "f"
test.
No statistical
difference con-
firmed by "f"
test.
No statistical
difference con-
firmed by test.
-------
TABLE 5-13 (c)
DIFFERENCES IN TEST METHOD PERFORMANCES;
PARTICIPATE ON FILTER(S) CONCENTRATION, GR/DSCF
Comparison of Methods
EPA ° Cascade Impactor
Method 5 (10 Tests)
(10 Tests)
to -
° Continuous Monitor
(10 Tests)
° Duct Background,
By EPA Method 5
(2 Tests Only)
0 Duct Background,
By Continuous Moni-
tor (4 Tests only)
0 Room Background
At Booth, By Hi-Vol.
(4 Tests only)
0 Continuous Monitor
Cascade
Impactor
to -
* Duct Background
By Method 5
" Duct Background
By Continuous Moni-
tor
° Room Background
Statistical Analysis
"t" Tests
Interval
0.0022+0.0020
None
None
None
None
0.0018+0.0011
0.0030+0.0023
0.0020+0.0019
Confidence
99%
99%
99%
99%
99%
99%
99%
95%
(
0.0022+0.0012
At Booth, By Hi-
Vol.
0 Duct Background
By EPA Method 5
Continuous
Monitor
to - ° Duct Background,
By Continuous Moni-
tor
° Room Background
At Booth, By Hi-
Vol.
Duct ° Duct Background,
Back- By Continuous Moni-
ground , tor .
By EPA
Method 5
to - ° Room Background
At Booth, By Hi-
Vol.
Duct
Back-
ground ,
By Con-
tinuous. ^om Background
Monitor Afc Booth> By Hi_
to * Vol.
None
None
None
None
None
99%
99%
99%
99%
99%
99%
None
99%
"{" Tests
"f" ratio
10.8
0.3
0.6
0.03
0.01
2.16
21.6
14.3
29.8
2.5
0.1
0.6
0.5
0.6
0.2
Critical "f"
5. % confidence
8.3; 99%
8.3; 99%
10.0; 99%
9.3; 99%
9.3; 99%
8.3; 99%
8.3; 99%
4.8; 95%
9.3; 99%
10.0; 99%
9.3: 99%
9.3; 99%
21.2; 99%
21.2; 99%
13.8; 99%
—
Remarks
Significant
difference c ^--
firmed by "f" tes-. .
No statistical
difference con-
firmed by "f" test.
Ditto
Ditto
Ditto
i
Significant
difference of lower
Cascade results
confirmed by "f" test.
Ditto
Ditto
Ditto
No statistical
difference con-
firmed by "f" test.
Ditto
Ditto
No statistical
difference con-
firmed by "f" test.
Ditto
No statistical
difference con-
firmed by "f" test.
,
!
-73-
-------
1. Total particulate (99 percent confidence interval) - 0.0033 +
0.0029 gr/dscf
2. Total condensibles (99 percent confidence interval) - no sta-
tistically significant difference.
3. Particulate on filter(s) (99 percent confidence interval) -
0.0022 + 0.0020 gr/dscf
Thus, the most similar results between the two methods were the
total condensibles. From the mean values of Table 5-15, total condensi-
1
bles represented 80 percent of the total sample collected by the cascade
impactor train and 56 percent of the total sample collected by the EPA
Method 5 train. The higher percentage of condensibles in the cascade
train can be attributed to collecting insufficient dry particulate on
each filter stage (as will be discussed in Section 5.4.3.3).
In summary, process emissions in excess of the background concentra-
tion were not detected by the EPA Method 5 train. Statistically, there
was no significant difference between the EPA Method 5 train results and
the continuous monitor results. The EPA Method 5 results were statis-
tically higher than the cascade impactor results, with the exception of
total condensibles.
5.4.3.3 Cascade Impactor Test Comparison
Particulate results from the continuous monitor are limited to a
dry particulate (listed as "particulate on filter" in Table 5-13c) since
there is no condensibles train presently designed for the unit. The re-
sults obtained showed no statistically significant difference when com-
pared to the:
-74-
-------
1. EPA Method 5 dry particulate duct test results..
2. Background results obtained in the duct by:
A. EPA Method 5
B. Continuous monitor
3. Background results obtained in the room by hi-vol tests near
the booth.
The total particulate 99 percent confidence interval for the true
difference between the averages of the methods was:
1. 0.0033 + 0.0029 Gr/DSCF when compared to EPA Method 5
2. 0.0018 + 0.0011 Gr/DSCF when compared to the continuous
monitor ("Particulate on Filter").
Several conditions must be discussed pertaining to the cascade im-
pactor results. First, no traversing was done. This may result in lower
concentrations than EPA Method 5.
Secondly, the particulate catch on individual filters was frequently
not detectable or lower than the amount that is recommended by the im-
pactor's manufacturer (minimum of 0.2 to 0.5 mg per filter) for reliable
results. A longer test time to collect more particulate on the filters
may be required, but this was prevented by the constraints of test tim-
ing. The range of total dry particulate collected for the impactor tests
was only 0.5 to 2.75 mg, whereas the total tare weight of cascade fil-
ters was approximately 850 mg. Thus the greatest collection rate only
represented about 0.3 percent of the total tare weight. The amount col-
lected and the accuracy of detecting that amount leads to the conclusion
that no meaningful particle size distribution results were obtained and
that the total dry particulate results may be erroneous.
Thus, in applying the cascade impactor procedure to the quasi-stack
-75-
-------
strategy, a process emission concentration and duration must be substan-
tially greater than this particular mold pouring operation to provide
meaningful results.
5.4.3.4 Continuous Monitor Test Comparison
Particulate results from the continuous monitor are limited to a
dry particulate (listed as "Particulate on Filter" in Table 5-13c) since
there is no cohdensibles train presently designed for the unit.
The results obtained showed no statistically significant difference
when compared to the:
1. EPA Method 5 dry particulate duct test results
2. Background results obtained in the duct by:
A. EPA Method 5
B. Cascade Impactor
3. Background results obtained in the room by Hi-vol tests
near the booth.
The continuous monitor results were higher than the cascade impac-
tor results (a 99 percent confidence interval of 0.0018 + 0.0011 gr/dscf
was obtained).
From the continuous monitor trace it was found that substantial mold
pouring particulate emissions occurred at the time of pouring, and, in
a matter of one or two minutes, the emissions returned to the genera
background levels for the remainder of the test (30 minutes total for
participates). This is clearly illustrated by Figures 5-10 and 5-11. Of
all the particulate emission testing procedures that were used during
this program, the continuous monitor was the only procedure that was ca-
-76-
-------
3
CO
H
8
O
CO
I
5
-------
I
?
CO
H
n
o
M
CO Ln
-------
pable of defining (either quantitatively or qualitatively) this brief,
large emission at the time of mold pouring. The EPA Method 5 and cas-
cade impactor procedures averaged that short emission excursion over a
30 minute sampling period, in this case and thus reduced the detectabil-
ity of this emission.
From Table 5-11, the mean concentration that was determined by con-
tinuous monitor was 0.0018 gr/dscf with a standard deviation of 0.0009
(20 tests). Figures 5-10 and 5-11 show that the concentration due to the
emissions at the time of pouring may be an order of magnitude greater
than the general background level, or approximately 0.02 gr/dscf (at
8500 scfm). This brief emission excursion for one mold is average by
EPA Method 5 and cascade impactor procedures over the 30 minute test per-
iod. This emission excursion has greater significance when it is con-
sidered as an accumulative emission for the large number of molds that
are typically poured per day in this foundry.
Thus, for a mold pouring operation such as this where the molds are
stationary at all times and the particulate emissions only occur briefly,
the EPA Method 5 and cascade impactor procedures average the emissions
over a long period of time and essentially monitor the background con-
centration. A continuous monitor is required to define the emission ex-
cursion.
The EPA Method 5 and cascade impactor procedures may be suitable to
mold pouring operations where numerous molds are conveyed to a station-
ary pouring station in a short period of time. Monitoring particulate
concentrations in excess of the background condition would be more likely
to occur than in the case tested for this program, and there would be
-79-
-------
more of a particulate catch for particle size distribution and composi-
tion analysis.
5.4.3.5 Test Setup Limitations
As discussed earlier, the nature of the mold pouring operation was
such that a large booth and large flow volume was required to capture
the fugitive emissions. The large flow volume probably induced a con-
vective air current at a considerable distance from the booth. Any
activity adjacent to the booth such as core cleaning, sand compacting,
or vehicle traffic (all of which occurred regularly during the testing),
could generate local fugitive emissions that might enter the booth.
Thus, a general activity background level would be monitored by the sam-
pling apparatus in the duct. That activity emission level may be as sig-
nificant as the 30 minute averaged emissions from the mold pouring and
cooling operation itself. This is supported by the statistical compari-
sons discussed earlier.
5.4.3.6 Critique of Sampling Methods
In Section 5.4.3.4 a discussion was given of the significance of
the emission excursion above general background concentrations at the
time of mold pouring. The continuous monitor was the only method that
defined that excursion. Unfortunately, the continuous monitor did not
have the provision for monitoring total condensibles, which comprised the
major portion of the particulate fugitive emissions (Section 5.4.3.2).
A condensibles train designed specifically for the continuous monitor
would have to be larger than the EPA Method 5 train since the sample flow
-80-
-------
volume is much greater. This would constitute a hybrid monitor that gave
continuous dry particulate emissions and integrated condensible material
emissions.
For this application, the cascade impactor results were question-
able. For more concentrated emissions and longer sampling times, this
method may be a suitable approach for providing particle size distribu-
tion and total particulates, including condensibles. The short sampling
times and generally low emissions encountered during this program led
to the conclusion that the EPA Method 5 train is probably the best all-
purpose method for particle size distribution (by microscopy, which will
be discussed in Section 5.4.3.7), total particulates including condensi-
bles, and particle composition (which will be discussed in Section 5.4.3.8.
The major limitation of the EPA Method 5 for this application was the in-
ability to determine brief excursion emissions, as described above.
5.4.3.7 Presentation of Particle Sizing Data
The cascade impactor test results were eliminated from considera-
tion for particle size analysis for the reasons outlined in Section
5.4.3.3. As a backup for the cascade impactor, a membrane filter was
used in the EPA Method 5 train setup. Particle size distribution by count
was determined by microscopy. Table 5-14 presents the results of that
analysis.
The statistical analyses of the 16 isokinetic emission tests and
the two EPA Method 5 duct background tests (also isokinetic) are presented
in Tables 5-15 and 5-16. Mean values and standard deviations of the par-
ticle size distribution by percent count are presented in Table 5-14.
-81-
-------
TABLE 5-14
PARTICLE SIZE DISTRIBUTION (% BY COUNT):
EPA METHOD 5
Test No.
1
2
3
4
5
6
Bl
7
B2
8
9
10
11
12
13
14
15
16
17
B3
18
19
20
B4
21
22
23
% Isokinetic
68
60
62
61
62
61
—
58
61
101
99
98
100
101
102
104
100
102
100
100
100
103
104
99
103
102
103
Distribution
% 3p
6.2
3.4
4.6
—
3.9
2.5
—
1.9
2.3
2.4
2.2
3.2
1.8
3.0
2.0
3.0
3.0
3.0
1.0
5.0
3.0
2.0
1.0
6.0
3.0
2.0
2.0
-82-
-------
TJSBLE 5-15
PARTICLE SIZE DISTRIBUTION (% BY COUNT):
MEAN VALUES AND STANDARD DEVIATION
Tests
16 Isokinetic Emission
Tests (p.m.)
2 Isokinetic Duct Back-
ground Tests (a.m.)
1 Distribution
3y
3y
y,
Mean
67.03%
30.00
2.35
52.00%
42.50
5.50
o,
Std. Deviation
6.91%
5.53
0.72
2.83%
2.12
0.71
TABLE 5-16
PARTICLE SIZE DISTRIBUTION (% BY COUNT):
STATISTICAL DIFFERENCES BETWEEN PROCESS EMISSIONS
AND BACKGROUND CONDITION
Distri-
bution
3y
Statistical Analysis
"t" Tests
interval
15.03% +9.81%
12.50% + 7.14%
3.15% + 2.30%
Confi-
dence
99%
99%
95%
"f" Tests
u.pii
ratio
8.91
9.60
34.13
Critical "f"
& % confidence
8.53; 99%
8.53; 99%
4.49; 95%
Remarks
Statistical
difference con-
firmed by "f"
test
Ditto
Ditto
-83-
-------
Table 5-15 presents the "t" test results as a confidence interval, as
discussed in Section 5.4.3, and the "f" test results to confirm that a
statistical difference between the sets of data existed.
From the analysis it can be concluded that there is a statistical
difference between the process emissions (monitored in the afternoon)
and the background condition (monitored in the morning). Approximately
two thirds of the process emission particles were less than one micron
in size, and approximately one third were in the one to three micron size
range. This compares to the duct background condition of approximately
half of the particles at less than one micron and about 40 percent between
one and three microns in size.
As discussed above, there is a statistical difference between the
process and background data. It can be concluded that a process emis-
sion particle characteristic, not a general background particle charac-
teristic, was monitored; but, as discussed in Section 5.4.3.1, the concen-
tration that was monitored by EPA Method 5 did not exceed the general
background concentration.
5.4.3.8 Particulate Composition
The average particulate catch on the EPA Method 5 membrane filters
was only about 1 mg for the isokinetic tests. For most laboratories and
standard analytical methods this is an insufficient particulate sample
for analysis of composition by percent weight. Longer sampling time to
obtain more of a particulate catch was ruled out since most of the emis-
sions occurred in a one to two minute period at the time of pouring, as
discussed in Section 5.4.3.4.
-84-
-------
Analytical methods described in Section 5.4.2 were employed despite
the insufficient sample. The accuracy of the results prevents reporting
meaningful data for composition by percent weight, but it can be reported
that traces of Si02, Fe, SO^" and N0~3 were detected on most of the fil-
ters.
To improve particulate composition results for operations such as
the one tested, entire filters rather than filter fractions must.be
used for each analysis. For operations with molds conveyed to a station-
ary pouring station, both entire filters and longer sampling periods
should be used to improve particulate composition results.
5.4.4 Evaluation of Hydrocarbon Concentration Data
The layout of molds on the foundry floor included double rows of
molds along the length of the bay. Pouring began, as shown in Figure
5-12, at approximately noon each day at the cupola area opposite the test
booth. Pouring progressed along one row towards the booth, then started
on the second row, and again progressed toward the booth. Pouring was
usually completed by 2:30 or 3:00 p.m. each day. Table 5-9 presents a
detailed description of foundry activities during each test.
Table 5-17 presents the total hydrocarbon (methane basis) concentra-
tion data that was obtained as outlined in Section 5.4.4. Concentrations
are reported from the daily plots at 30 minute intervals between noon
and 3:00 p.m. for each of the nine test days. The mean concentration
and standard deviation was calculated for each time, tabulated in Table
5-17, and presented graphically in Figure 5-13.
Normally three tests (mold pourings) were conducted at the test booth
-85-
-------
Pouring
direction
Holding
furnace Cupo|a
DO
n
nn
an
a
Molds
an
a
nn
n
n
nn
1st 2nd
row row
nnnn
Herman
press
fJ
Test
booth
Test
duct
Fan
Laboratory
trailer
Foundry
bay
Fig. 5-12. Layout of molds.
-86-
-------
TABLE 5-17
TOTAL HYDROCARBON CONCENTRATION
(ppm as
Day
Time
12:00
12:30
1:00
1:30
2:00
2:30
3:00
Total Hydrocarbons, ppm (as CH^)
4/7/75
48
46
34
30
28
58
> 64
4/8/75
5
14
25
29
29
62
47
4/9/75
7
10
38
20
16
30
50
4/10/75
4
17
231
70
48
46
86
4/14/75
22
96
62
26
70
39
96
4/15/75
20
22
c.200
56
41
54
—
4/16/75
12
12
c.200
50
20
72
50
4/17/75
—
—
48
56
48
70
72
4/21/72
82
60
46
66
76
72
68
Statistical Analysis
Mean HC Cone . ,
ppm (as CH^.)
25
34.6
98.2
44.8
41.8
55.9
66.6
Std. Dev., HC
ppm (as CHtt)
27
30.6
85.2
18.7
21.0
15.1
17.8
I
00
-4
I
-------
zoo
180
160
140
120
O
|ioo
80
60
40
20
«••
—
/
/
/
i
••
8
_
<
;
*
f ***
Startof
1st test
i
I I
/
i
4
•
9 -No. of test days
/T/ Standard deviation
^
\
\ 4
V/v Plot of mean values 1 1
* T i^
\ *• • ^***^l
^™ 1
-*--••
y^-Duct
X^ background
X test
Startof * startof
2nd test 3rd test
^ t
12--00 1:00 2:00 3=00
Time of day
Fig. 5-13. Plot of mean half-hour hydrocarbon
concentrations for period April 7-21,1975.
-88-
-------
each day. The average start time for the three tests is shown in Figure
5-13. One background test was conducted in the duct when no mold wss in
the booth. This test is also presented in Figure 5-13, and It generally
follows the pattern of the mold pouring tests.
It can be observed from Figure 5-13 that two peaks occurred during
the day. Daily observations determined that these peaks occurred about
50 minutes after pouring activity in the vicinity of the booth or the
adjacent Herman press—not necessarily after a mold was poured in the
test booth. Entries in Table 5-17 show that pouring occurred near the
booth (first row pouring) and Herman press area between 12:10 and 12:30
on April 8, 10, and 21 (booth) and April 15 and 16 (Herman press). Fig-
ure 5-13 illustrates the first peak occurring from this pouring activity
about 1:00 p.m. The second row near the booth was poured between 2:00
and 2:45 on April 9, 14, 16, 17, and 21, as recorded in Table 5-17. The
second peak of Figure 5-13 appears at 3:00 p.m. or later, after pouring
had been completed.
It can thus be concluded that HC monitoring essentially defined a
background or general foundry activity condition and isolation of a spe-
cific mold or the emissions contribution of one mold was not distinguished
from general background concentrations.
5.4.5 Evaluation of Carbon Monoxide Concentration Data
Table 5-18 presents the carbon monoxide concentration data that was
obtained as outlined in Section 5.4.1.8. Concentrations are presented from
the daily plots at 30 minute intervals between noon and 3:00 p.m. The
mean concentration and standard deviation was calculated for each time,
-89-
-------
TABLE :5-18
CARBON MONOXIDE CONCENTRATION DATA,
PPM
Day
12:00
12:30
1:00
Time 1 : 30
2:00
2:30
3:00
4-7-75
27
28
28
30
32
66
66
Carbon Monoxide Concentration, ppm
4-8-75
5
31
24
25
25
85
60
4-9-75
5
26
28
20
20
55
36
4-10-75
4
25
44
35
35
46
48
4-14-75
9
50
31
19
34
56
43
4-16-75
5
21
27
20
15
65
38
4-17-75
—
—
31
24
23
45
70
4-21-75
17
14
12
12
18
20
22
Statistical Analysis
Mean CO
Concentration, ppm
10.3
27.9
28.1
23.1
25.3
54.8
47.9
Standard
Deviation
8.7
11.2
8.8
7.1
7.6
19.0
16.6
O
-------
tabulated In Table 5-18, and presented graphically in Figure 5-14. A back-
ground test was also plotted on Figure 5-14.
From Figure 5-14 it can be observed that as in the case of total hy-
drocarbons, there are two peaks again appearing. From the pouring ac-
tivity description given in Table 5-18 and Table 6-9, it was noted
that CO concentrations increased almost immediately after pouring occurred
near the booth and in the adjacent Herman press area. The one duct back-
ground test also closely followed the pattern of Figure 5-14.
Thus, it can be concluded that a general background or activity con-
dition was monitored during the CO testing. Isolation of a specific mold
or the emissions contribution of one mold was not distinguishable from
general background concentrations.
As in the case of the HC measurements, this procedure for CO moni-
toring may be more suited to the foundry with a central pouring station(s),
If quasi-stack procedures are to be applied to the type of foundry opera-
tion that was tested, the test mold will have to be isolated better from
general room activity. If this is not possible, a broader approach for
defining emission concentrations will have to be employed, such as roof
monitoring or traditional industrial hygiene measurement techniques.
5.5 Task V - Issue a Technical Manual
After revision and thorough testing in the field, a final draft of
the field test procedure was issued.
-91-
-------
80
70
60
uw
50
8
| 40
30
20
10
0
^hM
8
•
M»
4
\
I
7- No. of test days jl
T T //
Standard deviation a / f "
1
i
8
1
A
'~"V
XV 1
\N-
\
\
" - f T/7 A
Plot of mean values j ,-'*" ^ '-.^ / ,' Duct
\^X ^^^ XJ ' M background
/ M '" I test
-7l 1
Average Average Average
. start of start of start of
A 1st test A.2ndtest 1 3rd test
.. T T
i • i i
1 11
12:00 1:00 2:00 3:00
Time of day
Fig. 5-14. Plot of mean half-hour carbon monoxide
concentrations for period April 7-21,1975.
-92-
-------
5.6 Units of Measure
Although it is the policy of EPA to use the metric system for qual-
itative descriptions, the British system is generally used in this re-
port. Appendix K contains a conversion table to aid readers who wish
to use metric units where only British units are cited.
-93-
-------
APPENDIX A
LITERATURE REFERENCES TO FUGITIVE EMISSIONS
-94-
-------
1. "^ocessing Emissions and Occupational Health in the Ferrous Foundry
Industry, C. E. Bates, L. D. Scheel, AIHA Journal, 452, August,
2. "Systems Analysis of Emissions and Emission Control in the Iron
Foundry Industry," A. T. Kearney & Co., Chicago, Illinois, Report
on EPA Contract CPA-22-69-106, February, 1971.
3. "Environmental Surveys of Aluminum Reduction Plants," P. J. Shuler,
P. J. Bierman, HEW Publs., (HIOSH) 74-101, April, 1974.
4. "Air Pollution Control in the Primary Aluminum Industry," Singmas-
ter & Breyer, New York, New York, Report on EPA Contract CPA 70-71,
March, 1972.
5. "Investigation of Fugitive Dust - Sources, Emissions, and Control,"
PEDCO Environmental Specialists, Cincinatti, Ohio, Report on EPA
Contract 68-02-0044, June, 1974.
6. "Soil Conditions that Influence Wind Erosion," W. S. Chepil, U. S.
Dept. of Agr. Tech. Bull. 1185, 1958.
7. "Asphalt Hot-Mix Emission Study," The Asphalt Institute, College
Park, Maryland, Pub. RR-75-1, 1975.
8. "Atmospheric Emissions from the Asphalt Industry," L. Laster, EPA
Publ. EPA-6509-73-046.
9. "Air Pollution from Coal Refuse Disposal Areas," V. M. Sussman and
J. J. Mulhern, JAPCA 14, 279, July, 1974.
10. "Control of Coke Oven Emissions," Battelle Columbus Laboratories
Report to AISI, December 31, 1971.
11. "Control of Coke Oven Emissions," T. Dancy, AISI Yearbook, p. 181,
1970.
12. "Systems Study for Control of Emissions - Primary Non-ferrous Smelt-
ing Industry," A. G. McKee & Co., Cleveland, Ohio. Report on NAPCA
Contract PM86-6585, June, 1969.
13 "Retrofit S02 Techniques for Existing Equipment at Arizona Smelters,"
I. J. Weisenberg, P. Wondra, 68th Annual APCA Meeting, Boston, Mas-
sachusetts, Paper 75-23.1, June, 1975.
14. "Compilation of Air Pollutant Emission Factors," EPA Pub. AP-42,
second edition, April, 1973.
15. "Solutions for Feedlot Odor Control Problems: A Critical Review,"
R. M. Bethea, JAPCA 22, 765, October, 1972.
16. "The Problem of Fugitive Dust in the Highway Construction Industry,"
-95-
-------
F. A. Rennlnger, 46th Annual Meeting Highway Research Board, Wash-
ington D. C., January, 1967.
17. "Hydrocarbon Pollutant Systems Study," MSA Research Corporation,
Homer City, Pennsylvania, Report to EPA, October 20, 1972.
18. "The Identification of Effluents from Rubber Vulcanization," S. M.
Rappaport, presented at the EPA Conference on Environmental Aspects
of Chemical Use in the Rubber Processing Operations, Akron, Ohio,
March 13, 1975.
19. "The Contribution of Open Burning of Land Clearing Debris to Air
Pollution," M. Feldstein, JAPCA 1JJ, 542, November, 1963.
20. "Control of Particulate Emissions from Lime Plants - A survey,"
L. J. Minnick, JAPCA, 195, April, 1971.
21. "Particulate Pollutant System Study," Midwest Research Institute,
Kansas City, Missouri, Report on EPA Contract CPA 22-69-104, May,
1971.
22. "Source Control of Air Emissions," H. F. Elkin, R. A. Constable,
Hydrocarbon Processing, 51, 113, October, 1972.
23. "Fluorine Emissions from Wet Process Phosphoric Acid Plant Process
Water Ponds," W. R. King and J. K. Ferrell, 68th Annual APCA meet-
ing, Boston, Massachusetts, paper 75-25.7, June, 1975.
24. "Air Pollution Control Technology and Costs in Seven Selected
Areas," IGCI, Stamford, Connecticut, Report to EPA on Contract 68-02-
0208, December, 1973.
25. "Odors and Air Pollution from the Treatment of Municipal Waste
Water," G. P. Sutton, 64th Annual Meeting of the Air Pollution
Control Assn., Atlantic City, N. J., June, 1971.
26. "An Evaluation of Charging and Tapping Emissions for the Basic
Oxygen Process," R. P. Mattis, 68th Annual APCA Meeting, Boston,
Massachusetts, Paper 75-75.1, June, 1975.
27. "Control of H2S Emissions During Slag Quenching," F. H. Rehmus,
D. P. Manka, E. A. Upton, JAPCA 23_, 864, October, 1973.
28. "A Systems Analysis Study of the Integrated Iron and Steel Indus-
try," H. W. Lownie, J. Varga, Battelle Memorial Institute, Colum-
bus, Ohio, Report to NAPCA on Contract PH-22-68-65, May, 1969.
29. "Exposure to Coal Tar Pitch Volatiles at Coke Ovens," N. Fannick,
L. T. Gonskor, J. Shockley Jr., AIHA Journal, J33, 461, July, 1972.
-96-
-------
APPENDIX B
SUMMARY OF TRADE ASSOCIATION/INDUSTRY/POLLUTION
CONTROL AGENCY CONTACTS ON FUGITIVE EMISSIONS
-97-
-------
SUMMARY OF TRADE ASSOCIATION/INDUSTRY/POLLUTION CONTROL AGENCY
CONTACTS ON FUGITIVE EMISSIONS
Organization Contact
1. National Crushed Mr. Rick Renninger
Stone Assn. V.P. Operations
2. American Iron &
Steel Inst.
3. Aluminum Assn.
Mr. Ed Lally
Dr. William Balgord
Manager Env. Services
4. Asphalt Institute Mr. V. Piuzinauskas
Chem Engineer
5. American Foundry-
men's Assoc.
6. Kennecott Copper
Corp.
7. Bethlehem Steel
Co.
8. Rubber Manufac-
turers
9. EPA Region III
(Philadelphia,
Pa.)
10. Arizona Div. of
Air Pollution
Control
11. Texas Air Control
Board - El Paso
llr. Hulsen
Dr. Fred Templeton
Mr. Tom Keller
Mr. Frank T. Ryan
Director, Env. Health
Safety Affairs
Mr. Lou Fallison
Mr. Mead Sterling
Mr. Jim Shoults
12. Pa. Dept. of Env. Mr. Feigenbaum
Resources Alle-
gheny Cty.
13. Bay Area Air Pol- Mr. Warren Grouse
lution Control
District-Californ-
ia
14. Los Angeles Air Mr. William Krenz
Pollution Control
District-Californ-
ia
Results
Contributed reports, reprints, EPA proposal,
all on fugitive dust in the stone industry.
Discussed subject in person at office.
Referred TRC to AISI environmental committee.
Contributed coke oven emission article, steel
making flowsheets on office visit.
Requested letter to environmental committee.
Suggested NIOSH data as being useful to us.
Visited office for discussion.
Discussed report on hydrocarbon emissions done
for them by Esso. Visit was made and data was
discussed before report issued. We will get
copy.
Discussed problems of cupola charging and pour-
ing. Suggested other contacts. Also dis-
cussed emissions from mold making and pouring.
Will contribute their data on fugitive emis-
sions at Utah smelter. Would like meeting
with TRC/EPA.
Discussed Roof Monitor Sampling Methods.
TRC may make presentations to environmental
committee. Other contacts suggested who may
supply data.
Suggested other contacts. Thinks dust blowing
off open hearth roofs is fugitive. EOF Tap
and charge emissions not well controlled.
Smelters have 10-25% of S02 as fugitive emis-
sion. Furnace leakage a problem. Storage
pile fugitive dust.
Fugitive S02 at smelters 5-10% of total.
Storage piles a problem — need to build
structures over them.
Most coke oven emissions are fugitive. They
have no data, but air quality near coke plants
known.
Uses API data as hydrocarbon emission guide-
lines. No new data known.
Uses API data as hydrocarbon emission guide-
lines. No new data known.
-98-
-------
APPENDIX C
MATRIX CHARTS USED TO IDENTIFY AND CHARACTERIZE
FUGITIVE EMISSION SOURCES
-99-
-------
KEY TO PROCESS STEP HEADINGS ON MATRIX CHARTS
List process or industry, whichever provides best description. For each
process step (described below), enter symbol (from attached list) de-
scribing character of emissions.
1. Material Delivery - Emissions from transfer of raw material from
transport mode to plant storage area (e.g., carbon black delivery
by railroad car to tire plant).
2. Material Storage - Raw material emission from process storage areas
(e.g., sand blown from cement plant pile, evaporation of hydrocarbons
from oil storage tanks).
3. Material Transfer - Emissions in getting materials from storage to
process use point (e.g., coal dust from conveyors in coal washing
plant).
4. Mixing and Grinding - Emission from putting ingredients together. Also
sizing and classifying emissions in this category (e.g., mixing of lime
and silica for cement manufacturing).
5. Ingredient Preparation - Emissions from special steps needed to make a
prime ingredient of the process (e.g., drying of stone for use in
asphalt paving, mold making for iron casting).
6. Reactor Charging - Emissions during opening of reactor to add raw
materials (e.g., blast furnace loading with iron, limestone and coke).
7. Process Leakage - Leaks of main process system where emissions are involved
(e.g., coke oven door seal leakage).
8. Enclosure Leakage - Open building emissions not treated by pollution
control air handling systems (e.g., dust blown out of windows or
doors in foundry due to insufficient hooding).
9. Reactor Discharge - Emissions when product is removed from main reactor
(e.g., tapping of a blast furnace).
10. Final Product Preparation - Any emissions from size reduction or
final forming needed to make usable product (e.g., cement clinker grinding
to a powder)
11. Solvent Evaporation - Emissions from uncontrolled solvent removal
operations (e.g., drying of ink in printing).
12. Ground Dust - Dust emissions not primarily from the process but made
airborne by wind or traffic near or in the plant (e.g., dust from
truck deliveries on dirt roads).
-100-
-------
13. Waste Transfer - Emissions from transfer of waste from process to
storage point. This includes waste captured by all forms of pollution
control equipment (e.g., conveying of slag to slag pit from blast
furnace).
14. Waste Storage - Emission from waste in a disposal area (e.g., vapors
from wastewater settling pond).
15. Product Transfer - Emission during transfer of product from process
line to storage area or transport (e.g., conveying of grain to
storage silos).
16. Product Storage - Emissions from product during storage before delivery
to user (e.g., vapors from creosoted telephone pole piles).
17. Product Packaging - Emissions in going from process or storage to
shipping or final retailing or wholesaling containers (e.g., fertilizer
bagging).
18. Mining and Blasting - Emissions from excavation, material breaking
and explosive blasting within mining type operations (e.g., dust plumes
from blasting).
-101-
-------
EXPLANATION OF SYMBOLS USED ON MATRIX CHARTS
TO DESCRIBE EMISSIONS CHARACTER
Symbol Meaning Particle Size
M Mining >2y
m Mining Dust <2y
S Stockpile Dust >2y
s Stockpile Dust <2y
T Truckload and Road Dust* >2y
t Truckload and Road Dust* <2y
C Crushing and Grinding Dust >2y
c Crushing and Grinding Dust <2y
D Transfer Material Dust >2y
d Transfer Material Dust <2y
0 Organic Particulate (Solid) >2y
o Organic Particulate (Solid) <2y
P Metallic Particulate >2y
p Metallic Particulate <2y
W Aqueous Mist >2y
w Aqueous Mist <2y
L Organic Particulate (Liquid) >2y
1 Organic Particulate (Liquid) <2y
G Odoriferous Gas
g Gas
V Odoriferous Vapor
v Vapo r
R Reactive Gas Emission
r Reactive Vapor Emission
F Agricultural Field Dust >2y
f Agricultural Field Dust <2y
B Smoke from Burning
E Dust from Pollution Control Equipment >2y
e Dust from Pollution Control Equipment <2y
Y Fume or Process Dust from Reactors >2y
y Fume or Process Dust from Reactors <2y
K Livestock Generated Soil Erosion >2y
k Livestock Generated Soil Erosion <2y
M Mining & Blasting Dust >2y
m Mining & Blasting Dust <2y
* Within Plant Confines
-102-
-------
MATRIX CHART - FUGITIVE EMISSIONS SOURCES IDENTIFICATION AND CHARACTERIZATION
I
H1
O
I
COMMENTS ON COMMENTS ON
FUGITIVE CANDIDATE
EMISSIONS MEASUREMENT
-------
APPENDIX D
MEASUREMENT OF FUGITIVE EMISSIONS
LITERATURE REFERENCES TO METHODS
-104-
-------
1. "Fugitive Dust Measures by the Rotary Aabient Adhesive Impac-
G" HarmS " '
2-
3. "A Microscopal Classification of Settled Particles Found in the
™nl^ ^V Coke-^king Unit," R. A. Herrick and L. G. Benedict,
JAPCA 19 (5), 325, May, 1969.
4. "Device for the Continuous Determination of the Dust Flow in Flow-
ing Gases," J. Bosch, Staub 32_, 8, November, 1972.
5. "Roof Monitor Emissions - Test Methodology," T. E. Kreichelt and
T. G. Keller, JAPCA 22_, 640, August, 1972.
6. "The Collection and Analysis of Inorganic Dust Downwind of Source
Effluents, "M. L. Feldstein, B. Potter, A. E. Alcocer, and H. Moore,
APCA 61st Annual Meeting, St. Paul, Minnesota, June, 1968.
7. "Automation of the Conventional Gravimetric Dust Measuring Method
for Quasi Continuous Following of the Courses of Processes," D.
Eickelpasch, Proc. Second Intl. Clean Air Congress, Washington,
D. C. , December, 1970, p. 446, Academic Press, New York (1971).
8. "The Use of Multiple Tracers in Determining Source Contributions
from a Smelter," S. L. Kiel, R. E. Stephenson, L. D. Smoot, and
D. H. Barker, AIChE 78th National Meeting, Salt Lake City, August,
1974.
9. "A Monitoring System for the Detection and Control of Airborne
Dust," F. A. Renninger, Dust Topics Magazine, October, 1966.
10. "A Comparative Study of Particulate Loading in Plumes Using Multi-
ple Sampling Devices," B. R. Meland, JAPCA, 1J3 (8), 589, August,
1968.
11. "Personal and High-Volume Air Sampling Correlation Particulates,"
P. M. Duvall and R. C. Bourke, Environmental Science and Technology
jj"(8), 765, August, 1974.
12 "Development of Methods for Sampling and Analysis of Particulate
and Gaseous Fluoride from Stationary Sources,1 A. D. Little, Inc.,
Report to EPA on Contract R2-72-126.
13 "Field Evaluation of the High Volume Particle Fractionating Cas-
cade Impactor," R. M. Burton, J. N. Howard R. L. Penly, P. A. Ram-
say, T. A. Clark, JAPCA 23, 277, April, 1973.
-105-
-------
15. "Comparison of 'Light Scattering Diameter1 Based on Forward Scat-
tering Measurements and Aerodynamic Diameter of Aerosol Particles,"
C. J. P. van Buitjenen, F. Oeseburn, Atmospheric Environment J3,
855, 1974.
16. "Developments in Sampling and Analysis Instrumentation for Station-
ary Sources," J. S. Nader, JAPCA 23, 587, July, 1973.
17. "Identification and Quantitative Analysis of Particulate Air Con-
taminants by X-Ray Diffraction Spectrometry," P. 0. Warner, L.
Sand, and J. 0. Jackson, JAPCA 22_, 887, November, 1972.
18. "A Portable Battery-Operated Immediate-Readout Dust Particle Analy-
zer," W. Withstandley, W. J. Moroz, G. W. Anderson, JAPCA 21_, 571,
September, 1971.
19. "Size Distribution Measurement of Airborne Coal Dust by Optical
Particle Counters," B. Y. H. Lin, V. A. Marple, K. T. Whitley,
N. J. Barsie, Am. Ind. Hyg. Assn. Jour. 35, 443, August, 1974.
20. "A New Approach to Roof Monitor Particulate Sampling," A. Sinka,
R. Marek, L. Gnan, JAPCA, 25, 397, April, 1975.
21. "Lidar Effort," E. W. Burgess, Lt. W. M. Pekny, R. E. Meyers, Des-
eret Test Center, Fort Douglas, Utah, July 21, 1971.
22. "Lidar-Traces Atmospheric Diffusion Measurement System," Ross, Des-
eret Test Center, Fort Douglas, Utah, August, 1971.
23. "Laser Radar Technology," Lt. W. M. Pekney, Deseret Test Center,
Fort Douglas, Utah, October, 1971.
24. "Composition and Concentration Measurements of Atmospheric Pollu-
tants by Remote Probing (Lidar)," Lt. W. M. Pekney, December, 1971.
25. "Lidar Study of the Keystone Stack Plume," W. B. Johnson, E. E.
Uthe, Stanford Research Institute, Menlo Park, California, April 17,
1971.
26. "Lidar," R. T. Collis, Applied Optics, August, 1970.
27. "Remote Measurement of Smoke Plume Transmittance Using Lidar," C.
S. Cook, G. W. Bethke, W. D. Conner, Applied Optics, August, 1972.
28. "Long-Path Monitoring of Atmospheric Carbon Monoxide with a Tunable
Diode Laser System," R. T. Ku, E. D. Hinkley, J. S. Sample, Lin-
coln Laboratory, Massachusetts Institute of Technology, Lexington,
Massachusetts, March, 1974.
29. "Bistatic Monitoring of Gaseous Pollutants with Tunable Semiconduc-
tor Lasers," E. D. Hinkely, Lincoln Laboratory, Massachusetts Insti-
tute of Technology, Lexington, Massachusetts, March, 1974.
-106-
-------
APPENDIX E
SAMPLE CRITIQUE OF SAMPLING METHODS
-107-
-------
Two important parameters which are of prime interest in particulate
emission measurements are particle size and material identification.
Both are obtainable with a reasonable degree of accuracy under certain
test conditions. Another parameter not readily obtained, however, is
the instantaneous particulate emission rate from a fugitive source,
Some progress has been made toward this goal in present techniques.
The methods of particulate fugitive emission measurement available
involve seven basic techniques:
1. Standard filter impaction, collection, and weighing.
2. Piezoelectric mass detection.
3. Light scattering (Nephelometry of Lidar).
4. Radiation scattering (B gauges).
5. Electrostatic precipitation with controlled voltage.
6. Size selective impaction and weighing.
7. Adhesive impaction.
Filter Impaction. This technique collects most suspended particu-
late matter with the exception of fine aerosols and particles (< 0.1 ym)
1
unless special filters are used which collect fine particles well at
lower sample flow rates. The total particulate can be weighed with some
degree of accuracy and analysis by instrumental or chemical techniques
can readily be done. Some size distributions can be determined by use
of a cyclone before the filter or two different porosity filters in a
series. The filter catch can be examined under microscopes (optical and
electronic) for determination of particle size. As a routine measure,
however, this would be prohibitively expensive. Another problem of this
technique is that it also gives only the average particulate concentra-
tion over the sampling period. If fugitive emission rates are time de-
pendent, a lot of detail of the emission characteristics can be lost.
-108-
-------
Piezoelectric. This technique accurately determines total parti-
culate loadings of samples in very short time periods (< 1 second). It
is, therefore, essentially on online continuous particulate matter moni-
tor. There are two problems in that size distribution and material an-
alysis are not possible. These are major disadvantages, since these are
the critical criteria for fugitive particulate emissions. Some size dis-
crimination could be done by suitable air sample prefiltration/separation,
but this has limited applicability unless multiple instruments are used.
Light Scattering. Use of a Nephelometer allows the direct measure-
ment of total particulate concentration and also the size distribution
in a continuous online monitoring system. There would be no question
that it met two important criteria except that these conclusions must
be qualified. Experience has shown that this instrument must be precal-
ibrated with the particle size and type which is to be examined. There
are serious problems unless this is done. Also, dilution is required
for measurements of concentrated particulate bearing gas streams, which
introduces a possible error. The use of the Nephelometer is a question
of how much pretesting can be tolerated to get good field data. Also,
a material analysis is not obtained by this technique and indeed must be
obtained before measurements can be made with the Nephelometer.
Another light scattering based system is Lidar, which remotely
scans a sector of the atmosphere, and the amount of scattered beam de-
tects the particulate concentration in that sector. The analysis of
dust from fugitive sources can be made instantly, and entire areas can
be "particulate mapped" over short time periods. Correlations with
other methods have not been well-proven at the present, and commercial
-109-
-------
availability of equipment in in doubt.
Radiation scattering. The primary technique is to measure scatter-
ing of 3 radiation from a radioactive source by collected participates.
Since only small amounts of sample need to be collected, data acquisi-
tion is semi-continueus. Two different sample collection techniques
have been employed. One uses a moving paper tape upon which the sample
impacts and is then read with the g-scatter. The other uses a condensa-
tion chamber (water droplets trap the particulate) to concentrate the
particulate on a glass surface, which is then read with the 3-scatter.
Both are readily adaptable for automatic sampling/readout. Hand port-
able instruments can be used so that areas of fugitive emission could
be mapped by an operator walking through a specified sector. The method
does require calibration with dust of the same type as being measured
and also calibration of the individual instruments themselves.
Electrostatic precipitator. This is basically a small-scale ver-
sion of an electrostatic precipitator and can be used to collect a re-
presentative sample on a suitable medium. Particle size and mass weight
can be determined as easily as with a filter. Efficiency for fine par-
ticulate is better than a filter; a semi-continuous version is avail-
able for submicron particle sizing. A change in collector rod voltage
is used to discriminate size. The techniques seem ideal for inorganic
dust and aerosols, but no data on organics such as tars has been found.
These techniques may be applicable to specific situations where size
distribution of inorganic particulate needs to be measured.
Size selective impaction. A cascade impactor such as the Andersen
Sampler can be used. It is reproducible and has been correlated to
-110-
-------
other techniques and also with known standards. It offers a reasonable
range of size detection for studies of hazardous particulates. The
problem is that it requires a large sample (plates are weighed) for ac-
curate results and that this is only an average over the sample time
period. It is, however, one of the easiest size discriminators to use
under field test conditions.
Adhesive impaction. For size detection with low sample concentra-
tions and/or short sampling times, these are ideal. They can be acti-
vated whenever important sampling is needed and many samples can be
taken during a day. The problem is the associated laboratory time spent
counting particles is large, and the analysis still is an average of
the time interval (although short, results are not instantaneous). It
would probably be advantageous to have some available for special tests
of localized fugitive emissions.
The most viable combination depends upon the strategy needed for
measurement of the fugitive emissions. The best recommendations are as
follows:
1. Pseudo duct strategy
A. Piezoelectric mass monitor
B. Andersen impactor
C. EPA train as backup and confirmation of mass monitor
(also used for organic particulates).
2. Roof monitor strategy
A. Multiple hi-vols for TRC 4" filters
B. Multiple Andersen impactors
C. Piezoelectric mass monitor for special runs
3. Upwind-downwind strategy
A. Multiple hi-vols or TRC 4" filters
B. Multiple Andersen impactors
-111-
-------
C. Piezoelectric mass monitor for special runs
D. Rotorod Adhesive Impactor for FP tracer tests and
other special situations
E. Consider experimental Lidar test
-112-
-------
APPENDIX F
EPA METHOD 5 DATA SHEET
-113-
-------
Firm Name
Plant Location
Test No.
FIELD DATA SHEET
(EPA Method 5 Sampling Procedure)
Pump No.
^O Ambient Temp.
Sampling Location Ho\dL Penxviv-»Q. Bar Press.
Date 1-IQ05 Probe Id en
Teste* W.M. ? c».p. Filter Ide
2-3 Probe Diameter VlG>
in.
OH- £>H = \.2
t. No. 40 CP -871 Probe Heater Setting —
nt. No. IO7 Box Temp. Setting
min
min
amp
F
Port
i-j
/M
4>
, 1
Point
1
Z
3
H
5
fc
3
4
5;
Time
Min.
5
5
5
5
S
5,
6J7
8J9 1
.0
Velocity
P
in H20
^
£>
26
27 .
28 2
9 bo
PSTACK
|
1
1
1
O
0
•*>
7
2
0
°l
1
O
5
O
0
1
<
2
33 :
)4 :
15
TSTACK
°l
t>
d
fe
(t>
t
1
8
7
«
H
5
ko
f
41
42
43 '
w
43
.. ..Initial
tfeter volume
cu. ft.
^
H
4
H
M
4
4
1
1
1
3
8
$
?
\
4
7
0
3
<4
&
4
H
7
7
s
0
z.
1
5
7
5
fe>
P
?
50
\
|
)
1
1
1
\
1
Pump
Vac.
in Hg.
Remarks
F'^?\ .
voiwiwe
-------
APPENDIX G
CASCADE IMPACTOR DATA SHEET
-115-
-------
CASCADE IMPACTOR DATA SHEET
Test Location
Test No. \S
Date 4 - I
Test Time
Operators
Test Point Location Hot A
Start of Test (Time) 17.
Initial Meter Volume (Cu. Ft.)
Initial Average Meter Temp (°F)
Filters:
Collection Paper 1
Collection Paper 2
Collection Paper 3
Collection Paper 4
Collection Paper 5
Collection Paper 6
Collection Paper 7
Back-up Filter
1OS.QS
Filter No.
IO3
lOl
I OO
T '
Velocity Pressure iP Cin H20) O-M-O
Orifice 4H C
-------
APPENDIX H
IKOR CONTINUOUS MONITOR DATA SHEET
-117-
-------
Test Location HO\A p€>U.
«. t
Time Constant
(Sec.)
IOO
••
ii
t •
ti
Sensitivity
JO
«•
'•
M
Instantaneous
Mass Flow MS
2.
1
.5
-
Integrated
Mass. Read.
Mi
O
M
to
m-s
Back-Up
Filter
No.
^
•%
»*
II
H
-------
APPENDIX I
HI-VOL FILTER TEST DATA SHEET
-119-
-------
HIGH VOLUME FILTER TEST
DATA SHEET
; EQUIPMENT
IDENTIFICATION
V4V-2.
UM-3
i
i
i
1
i
;
I
LOCATION
ooorK
P Y"O(JLutA •
Area-
FILTER
NUMBER
UIS3
...»
TIME
ST*.r.T
i
1
U-.13
•r«
INITIAL
VOLUME FLOW
•40
Bo-U fcosz
C-oto SSV*
*
i
TIME
END
y.*i
V.M
j
i
FINAL
VOLUME FLO™
i4
Mt.
-------
APPENDIX J
LABORATORY WORKSHEET FOR FIELD SAMPLES
-121-
-------
LABORATORY WORKSHEET
FOR
EPA FUGITIVE EMISSION STUDY
PROJECT 32397-04
Test Location
Test Date
Sent By
Reported To
KJ.
\\) .
Identification
Test No.
Sample No
Sample Description
Ana lysis Required
Special Comments
5
(p
Ikor
-122-
vot,
lot.
iot.
Ujt.
iot.
tot
GPA
-------
APPENDIX K
CONVERTING UNITS OF MEASURE
-123-
-------
EPA policy is to express all measurements in metric units. When
implementing this practice will result in undue cost or lack of clarity,
conversion factors are provided for the non-metric units. Generally,
this report uses British units of measure. For conversion to the metric
system, use the following conversions:
To convert from:
to:
Multiply by;
acfm
°F
ft
ft/sec
gal /me f
gpm
gpm/ft2
gr/scf
in
in H20
Ib moles
Ib moles/hr
Ib moles/min
tons
nm3/hr
°C
m
m/sec
1/m3
1/m
1/min/m2
gm/m3
cm
mm Hg
gm moles
gm moles/min
gm moles/sec
kg
1.70
Subtract
divide
0.305
0.305
0.134
3.79
40.8
2.29
2.54
1.87
454
7.56
7.56
907.2
32, then
by 1.8
-124-
-------
TECHNICAL REPORT DATA 1
(Please read Instruction s on the reverse before completing!
1 REPORT NO. 2.
EPA-600/2-76-284
4. TITLE AND SUBTITLE
Development of Procedures for the Measurement of
Fugitive Emissions
7.AUTHOB
------- |