MIDWEST RESEARCH INSTITUTE
OCCUPATIONAL EXPOSURES FROM BAGGING AND DRUMMING OPERATIONS
FINAL REPORT
EPA Prime Contract No. 68-02-4252
Work Assignment No. 10
MRI Project No. 8810-A(01)
September 9, 1987
For
U.S. Environmental Protection Agency
Office of Toxic Substances
Field Studies Branch, TS-798
Washington, D.C. 20460
Attn: Mr. Michael R. Kalinoski, Work Assignment Manager
Dr. Kin F. Wong, Work Assignment Manager
Dr. Joseph J. Breen, Project Officer
Mr. William M. Burch, Project Officer
MIDWEST RESEARCH INSTITUTE
-------
OCCUPATIONAL EXPOSURES FROM BAGGING AND DRUMMING OPERATIONS
FINAL REPORT
EPA Prime Contract No. 68-02-4252
Work Assignment No. 10
MRI Project No. 8810-A(01)
September 9, 1987
Attn:
For
U.S. Environmental Protection Agency
Office of Toxic Substances
Field Studies Branch, TS-798
Washington, D.C. 20460
Mr. Michael R. Kalinoski, Work Assignment Manager
Dr. Kin F. Wong, Work Assignment Manager
Dr. Joseph J. Breen, Project Officer
Mr. William M. Burch, Project Officer
MIDWEST RESEARCH INSTITUTE 425 VOLKER BOULEVARD, KANSAS CITY, MISSOURI 64110 • 816 753-7600
-------
This document has been reviewed and approved for publication by
the Office of Toxic Substances, Office; of-'Pesticides and Toxic Substances,
U.S. Environmental Protection fA"gency? -; The^use^df-^lkraae names or commercial
products does not constitu.tSe^Agencyfendbrsemerit.or -recommendation for use.
-------
PREFACE
This report describes the study performed by Midwest Research Insti-
tute (MRI) under Work Assignment 10 of Environmental Protection Agency (EPA)
Contract No. 68-02-3938. The objective of this work was the development and
validation of predictive models to be used in estimating the inhalation expo-
sure of workers involved in the bagging or drum filling of solids or in the
drumming of liquids. The EPA work assignment managers were Dr. Kin Wong and
Michael Kalinoski.
The task leader for Midwest Research Institute was Dr. Chatten
Cowherd, Director, Environmental Systems Department. The authors of this
report were Dr. Chatten Cowherd, Mary Ann Grelinger, and Dr. Gregory Muleski.
Mrs. Grelinger coordinated the laboratory studies and assisted in the plant
testing and experimental data analysis. Dr. Muleski was responsible for the
exposure model development. Mr. Fred Bergman designed the low-flow wind tun-
nel and coordinated the plant testing and the chemical analysis of field and
laboratory samples. Mr. Steve Cummins assisted in the collection of field
and laboratory samples and in data analysis. Other MRI staff members who con-
tributed to this study were Fred Hopkins, Bonnie Carson, Mark Golembiewski,
Dr. George Scheil, David Steele, Frank Pendleton, and Dr. Robert Hegarty.
MIDWEST RESEARCH INSTITUTE
Chatten Cowherd, Director
Environmental Systems Department
Approved:
r"
o4^ c&kZL*^ — if .
Jack Balsinger
Quality Control Coordinator
*aul C. Constant
Program Manager
-------
TABLE OF CONTENTS
I. Summary .......................... 1
II. Introduction ........................ 5
A. Background and Purpose .............. 5
B. Industry Characterization ............ 5
C. Prior Modeling Approaches ............ 6
D. Study Design ................... 7
E. Quality Assurance ................ 8
III. Conclusions ........................ g
A. Plant Studies .................. 9
B. Laboratory Drumming Study ............ 9
C. Model Development ................ 10
IV. Plant Scale Studies .................... 11
A. Drumming Operation (Plant A) ........... 11
B. Bagging Operation (Plant B) ........... 22
V. Laboratory Drumming Study ................. 51
A. Emission Tests .................. 51
B. Dispersion Tests ................. 63
C. Exposure Tests .................. 72
VI. Model Development ..................... 91
A. Modeling Alternatives .............. 91
B. Modeling Approach ................ 93
C. Model Validation ................. 97
References .............................. 101
Appendix A - Industry Survey
Appendix B - Reports on Additional Plant Surveys
Appendix C - Laboratory Pilot Study
Appendix D - Quality Control Data
Appendix E - Data from Laboratory Drumming Tests
-------
LIST OF FIGURES
Number Title Page
IV-1 Plant A drumming line 12
IV-2 Sampling element for collection of hydraulic fluid mist
and vapors 14
IV-3 Sampler deployment around the Plant A drumming operation. . 15
IV-4 Overhead view of airflow at 1.5 m above drumming platform
for Test A-l (Fluid 1) 20
IV-5 Overhead view of airflow at 1.5 m above drumming platform
for Test A-2 (Fluid 2) 21
IV-6 Vapor concentrations (ug/m3) of Fluid 2 observed during
Test A-2 23
IV-7 Mist concentrations (ug/m3) of Fluid 2 observed during
Test A-2 24
IV-8 Plant B bagging facility 27
IV-9 Sampling elements for collection of sodium tripoly-
phosphate particles 29
IV-10 Sampler deployment around the Plant B bagging operation . . 30
IV-11 Overhead view of airflow at operator breathing height
during Test B-l 35
IV-12 Overhead view of airflow at operator breathing height
during Test B-2 36
IV-13 Overhead view of airflow at operator breathing height
during Test B-3 37
IV-14a Total suspended particulate concentrations of STP (ug/m3)
for Test B-l 41
IV-14b Respirable particulate concentrations of STP (ug/m3)
for Test B-l 42
IV-15a Total suspended particulate concentrations of STP (ug/m3)
for Test B-2 43
IV-15b Respirable particulate concentrations of STP (ug/m3)
for Test B-2 44
IV-16a Total suspended particulate concentrations of STP (ug/m3)
for Test B-3 45
IV-16b Respirable particulate concentrations of STP (ug/m3)
for Test B-3 46
IV-17 Airborne particle size distributions obtained using
cyclone/cascade impactor 47
V-l Overhead view of pull-through flow tunnel 53
V-2 Full scale 55-gal drumming operation in flow tunnel .... 54
V-3 Comparison of emissions from top versus bottom drum
filling of perc at nominal 25 gpm 60
V-4 Profile of methanol concentrations at 3 m sampling plane
(200 fpm) 65
V-5 Profile of wind speeds at 3 m sampling plane for methanol
test at 200 fpm 66
V-6 Comparison of vertical profiles of methanol concentrations
from a drum side release with wind speed of 200 fpm ... 67
V-7 Comparison of vertical profiles of methanol concentrations
from a drum side release with wind speed of 100 fpm ... 68
vn
-------
LIST OF FIGURES (continued)
Number Title Page
V-8 Cross-sectional view of methanol concentrations (ppm)
at 1 m downwind for a drum side release at 100 fpm,
compared to downwind facing worker exposure (ppm) .... 69
V-9 Cross-sectional view of methanol concentrations (ppm)
at 3 m downwind for a drum side release at 100 fpm,
compared to downwind facing worker exposure (ppm) .... 70
V-10 Sampling locations for upwind bung position 73
V-ll Sampling locations for side bung position 74
V-12 Example exposure testing 75
V-13 Effect of breathing height and upwind distance on exposure
to perc concentrations (ppm) 80
V-14 Effect of breathing height and upwind distance on exposure
to perc drumming emissions 81
V-15 Effect of wind speed and upwind distance on exposure to
methanol drumming emissions at 1.5 m breathing height . . 82
V-16 K contours for methanol (M < 100), 100 fpm air velocity
and downwind worker orientation 83
V-17 K contours for methanol (M < 100), 200 fpm air velocity
and downwind worker orientation 84
V-18 K contours for methanol (M < 100), 100 fpm air velocity
and non-downwind worker orientation 85
V-19 K contours for methanol (M < 100), 200 fpm air velocity
and non-downwind worker orientation 86
V-20 K contours for methanol (M > 100), 100 fpm air velocity
and downwind worker orientation 87
V-21 K contours for methanol (M > 100), 200 fpm air velocity
and downwind worker orientation 88
V-22 K contours for methanol (M > 100), 100 fpm air velocity
and non-downwind worker orientation 89
V-22 K contours for methanol (M > 100), 200 fpm air velocity
and non-downwind worker orientation 90
VI-1 The box (uniform mixing) model 92
vm
-------
LIST OF TABLES
Number Title Page
IV-1 Plant Tests of Drumming Operation 18
IV-2 Representative Drum Fill Cycles 18
IV-3 Plant A Ventilation 19
IV-4 Mass Concentrations Observed During Test A-2 (185-min
test period) 25
IV-5 Mass Concentrations Observed During Test A-2 (98-min
test period) 26
IV-6 Sodium Tripolyphosphate Analysis Protocol 32
IV-7 Plant Tests of Bagging Operation 33
IV-8 Plant B Ventilation 34
IV-9 Mass Concentrations Observed During Test B-l 38
IV-10 Mass Concentrations Observed During Test B-2 39
IV-11 Mass Concentrations Observed During Test B-3 40
IV-12 Bulk STP Particle Size Distribution 48
V-l Chemicals Used in Drumming Simulations 52
V-2 Emissions Test Data - Perchloroethylene 57
V-3 Emissions Test Data - Ethylene Glycol 58
V-4 Rates of Attainment of Maximum Drum Exit Concentration. . . 61
V-5 Drumming Emission Factors 62
V-5a Comparison of Saturation Factors 63
V-6 Dispersion Test Log 71
V-7 Exposure Test Leg 76
V-8 Drum Emissions for Tests of Dispersion and Exposure .... 76
V-9 Special Exposure Studies 78
V-10 Effect of Worker Movement on Exposure 79
VI-1 Default Values for Equation 1 93
VI-2 Default K Values 95
ix
-------
I. SUMMARY
Under the Toxic Substances Control Act, the potential exposure to
workers involved with the manufacture and processing of a new chemical must
be estimated. A significant portion of worker exposure is thought to occur
during packaging operations, in particular, drumming and bagging of liquids
and solids. Currently, crude techniques are used to predict worker exposure
to emissions from the drumming of liquids and the bagging or drumming of
solids. Although mathematical models have been developed previously to im-
prove these estimates, available data are generally inadequate to verify the
models.
The purpose of this work assignment was to develop and validate
predictive models used to determine the inhalation exposure levels of workers
involved in the drumming of liquids and the bagging or drumming of solids.
The study consisted of four tasks. An industry survey was conducted to char-
acterize industrial drumming and bagging operations and to select chemicals
and facilities for study. An initial laboratory pilot study was performed
to test emission models, dispersion theories, sampling strategies, and ana-
lytical methods. A plant-scale occupational exposure study was directed to
the collection of specific data on worker breathing zone concentrations re-
sulting from actual bag and drum filling operations. Finally, a second labo-
ratory pilot study was performed using a drum filling station to develop a
predictive model of inhalation exposure, which was tested against data from
the plant-scale occupational exposure study.
In the first laboratory pilot study, a walk-in laboratory flow tun-
nel was used to characterize the low-flow (i 2 m/s) dispersion of methane gas
and ammonium chloride particles at distances of 2 to 6 m from the point of
release. Several release configurations were included: single- and double-
point releases into an unobstructed flow; point release near a wall; point
release past a side-draft hood; and bung release from a 55-gal drum (with and
without a close fitting hood). A two-dimensional array of samplers was used
to characterize dispersion patterns at various distances downstream. The ob-
served dispersion patterns were compared to existing dispersion theories and
then used as the basis for a modified dispersion algorithm. The predictive
capability of the resulting dispersion model was found to be significantly
better than the other models tested against the laboratory data. The first
laboratory pilot study is presented in Appendix C of this report.
The plant-scale occupational exposure study was performed on full
scale packaging systems at two industrial plants. At Plant A, low-volatility
hydraulic fluids were packaged in 55-gal drums; and at Plant B, granular so-
dium tripolyphosphate (STP) was packaged in 50- and 100-Ib bags. Integrating
samplers were deployed at fixed points surrounding each work station. Pairs
-------
of samplers were colocated to determine measurement precision. Personal sam-
plers were operated simultaneously by plant personnel. Additional data col-
lected during each test included ventilation flow speeds and directions
across the work station and characterization of worker activity (time and
motion). Although hydraulic fluid vapors were found in low concentration at
the drumming station (^ 100 |jg/m3), comparable amounts of mist were also
present.
The concentrations of particulate STP in the vicinity of the drum-
ming were much higher (•>» 1,000 ug/m3) with about 20% in the respirable size
range (i 10 umA). Moreover, respirable STP concentrations measured by the
personal sampler were considerably higher than the respirable particulate
concentrations measured at nearby sampling points. This is believed to result
from the airflow which was directed toward the back of the bagging operator,
and which produced a relatively stagnant air region immediately in front of
the operator. Gently circulating air in this region is likely to have trans-
ported emissions from the bagging operation directly to the breathing zone.
Subsequently a second laboratory pilot study was conducted using a
drumming station and encompassed the measurement of emissions, dispersion,
and operator exposure resulting from a full-scale 55-gal drumming operation
in the low-flow laboratory wind tunnel. The second laboratory pilot study is
described in Section V. A 55-gal drum with an added bottom drain was posi-
tioned on a platform within the flow tunnel. A 300-gal reservoir was in-
stalled outside the tunnel at a height of 2.13 m (7 ft) above the floor. The
overhead tank was fitted with chemically resistant hosing and a pump for
transfer of the test liquid to the 55-gal drum at a fixed rate controlled by
a ball valve at the hose nozzle. After each drum filling, the contents of
the drum were pumped back to the overhead tank. The test liquids (methanol,
perchloroethylene, and ethylene glycol) spanned a vapor pressure range from
about 0.1 to 125 mm Hg at 25°C.
Drumming emissions were measured by sampling and continuously ana-
lyzing vapor effluent from the bung opening as the drum was being filled.
Drumming emissions were characterized for top versus bottom filling and as a
function of drum fill rate. Top fill emissions exceeded bottom fill emissions
by a factor of two or more. Because of the tendency of perchloroethylene to
froth, top fill emissions were found to exceed the equivalent of saturated
vapor.
Distributions of airborne concentration and personal exposure for
fixed worker positions were measured in the vicinity of the drumming opera-
tion. The presence of the worker enhanced the exposure by drawing the plume
upward on the lee side of the worker. The highest exposure occurred with the
source in front of a worker and the ventilation airflow approaching from be-
hind the worker.
A semi empirical model of drumming exposures was developed using
(a) contours of breathing height concentrations developed from the second
laboratory pilot study and (b) worker time/motion data from plant-scale oper-
ations. The model separates worker exposure into two components—background
-------
and plume contribution. The background value is calculated by assuming com-
plete mixing of the plume in the plant ventilation air. The plume contribu-
tion depends on emission rate and worker positioning in relation to the
source and the ventilation flow direction. The validity of this modeling
approach was checked against the test data from the pi ant-scale drumming oper-
ation, with favorable results.
-------
II. INTRODUCTION
A. Background and Purpose
Under the Toxic Substances Control Act (TSCA), a manufacturer or
processor must submit to the EPA a premanufacturing notice (PMN) for every
new chemical prior to commercial production. The Agency has 90 days to re-
view the PMN and to take any regulatory actions deemed necessary. As part of
this review, the potential workplace exposure associated with the manufacture
and processing of the new chemical must be estimated.
Monitoring data on similar substances, structural analogs, or mathe-
matical models compiled from experience with the specific industrial opera-
tions are often utilized to evaluate the level of worker exposure. The common
models generally are based on material balances and diffusion transport con-
siderations. However, data from the open literature are frequently not avail-
able to verify the accuracy and precision of these models.
Currently under the PMN review process, a simple mass balance model
is used to estimate worker exposure to vaporous emissions (Berman 1982, USEPA
1984). The model is based on complete mixing corrected by a mixing factor to
account for incomplete mixing. For the packaging of solids, no dispersion
model is currently used. Instead the OSHA permissible exposure limit for
nuisance dust (15 mg/m3 time-weighted average, or TWA) is used as a conserva-
tive estimate in the assessment.
The purpose of this work assignment is to develop and validate pre-
dictive models used to determine the inhalation exposure levels of workers
involved in the drumming of liquids and the bag or drum filling of solids.
Such models must address both the release and the dispersion of emissions in
the immediate vicinity of the packaging equipment.
B. Industry Characterization
The bagging and drumming of PMN chemicals takes place under a wide
range of conditions that influence worker exposure during packaging.* These
variables include degree of automation; employee work practices; level of
occupational health awareness (managers and employees); filling equipment
design, age, and frequency of maintenance; container and closure design;
plant layout and ventilation; effectiveness of emission control devices, if
present; employee use of personal protective gear; unit value of product
(lower product loss rates, including packaging emissions, will be permitted
*Bomberger et al. (1983) provides a good, up-to-date source of information
describing typical filling technology and available literature. The reader
is referred to that for additional references.
-------
if the product value is high); and product flammability. Some volatile liq-
uids and dust-prone solids are carefully controlled because of the fire haz-
ard they represent in the absence of such controls. In an effort to assemble
a broadly applicable data base and methodology for estimating worker exposure
in the bagging and drumming of PMN chemicals, only the most relevant of these
variables could be studied.
1. Drumming of Liquids
In the drumming of PMN liquids, it is expected that both fully
automated and semiautomated filling practices would be commonly used. The
product will be either top filled (dispenser stays near the top of the bar-
rel, liquid splashes freely during filling) or bottom filled (dispenser re-
mains submerged inside the barrel, as much as possible, to minimize volatile
emissions). Emission controls may or may not be present. An operator may be
required to manually screw a cap on each barrel or wipe droplets off the tops
of the barrels (Bomberger et al. 1983).
In addition to work practices and controls, exposure to a specific
chemical is dependent on its inherent properties. Vapor pressure is a key
physical/chemical parameter for any analysis of emissions from the drumming
of liquids. PMN chemicals have been found to exhibit a considerable range of
vapor pressures common to industrial chemicals in general.
2. Bagging of Solids
In the bagging of PMN solids it is unlikely that production volumes
will justify fully automated packaging lines in many instances (Bomberger
et al. 1983). However, the use of automated or semiautomated equipment re-
mains a key variable in predicting worker exposure. The product will typi-
cally be placed into bags (open top or valve) by gravity feed, auger feed,
or vibrator feed equipment. Emission controls may or may not be present.
Frequently, an operator may be required to manually staple or otherwise seal
the bag tops.
C. Prior Modeling Approaches
The models available for modeling contaminant concentrations in an
industrial environment may be grouped into three types based on their theo-
retical basis: mass balance; Gaussian diffusion; and aerosol mechanics.
Mass balance models are based on mathematical accounting schemes
whereby all fractions of the material passing through the process under study
are quantified, and the total outputs are always equal to the total inputs.
These approaches are most applicable for gaseous emissions because vapors
follow air currents freely and are not influenced by gravity. Particulate
emissions consisting of fine particles—smaller than 10 to 20 |jm in diameter
according to Clement Associates (1981) and generally smaller than 5 urn in
diameter according to Bomberger et al. (1983)—may also be modeled effec-
tively using mass balance equations.
Gaussian dispersion models are not commonly employed in an indoor
setting. However, Schroy (1979) has discussed the use of this type of model.
-------
Such models are based on the Gaussian diffusion of particles as they move
away from a source, and they predict the maximum downwind concentration of
the contaminant as a function of distance from the source.
The third model type, based on aerosol mechanics, is exemplified by
Davis (1971). Here, the modeling approach is based on the properties and
mechanics of airborne particles, such as particle generation, coagulation,
and settling. Because the use of this method involves solutions of compli-
cated equations, use of a computer is recommended.
The discussion section at the conclusion of the Clement Associates
(1981) report notes that mass balance equations are both the simplest and most
commonly used approach for estimating exposure to a substance in an enclosed
space. However, there is no widely accepted methodology for this type of
analysis.
Berman (1982) provides methodology for estimating workplace expo-
sure to PMN substances. His approach is based on a collection of typical and
worst case approximations combined with easily understood equations and nomo-
graphs. For bagging operations, a typical concentration of 2 mg/m3 of mate-
rial (30% of which is respirable) is assumed in the air surrounding the bag
packer and, if warranted, a worst case value of 30 mg/m3 (also 30% respirable)
is assumed for especially dusty conditions. For drum filling of liquids, a
plug of saturated air equivalent to the barrel volume is assumed to be gener-
ated for every container filled. The concentration and total volume of con-
taminated air so released is calculated using an equation which accounts for
vapor pressure, temperature, and gram-molecular weight of the liquid, as well
as the filling rate and filling mode (top filling versus bottom filling).
Evaporation from spillage is ignored. A typical ventilation rate for bagging
and drumming areas is assumed to be 85 mVmin (3,000 ft3/min) and a worst case
value is 14.2 mVmin (500 ft3/min). For outdoor operations with only minimal
structural protection, the average wind velocity of North America, 14.5 km/h
(9 mph), is assumed in place of a building ventilation rate.
The validity of this methodology remains unknown, because insuffi-
cient data are available for verification. However, the Berman report repre-
sents the most complete set of guidelines presently available for estimating
workplace exposure to PMN chemicals during bagging and drumming activities.
D. Study Design
The study design was comprised of four tasks. An industry survey
(Task A) was conducted to characterize industrial drumming and bagging opera-
tions and to select chemicals and facilities for study. An initial series of
laboratory pilot studies (Task B) were performed to test emission models, dis-
persion theories, sampling strategies, and analytical methods. The plant-
scale occupational exposure study and the laboratory drumming study (Task C)
were directed to the collection of specific data on worker breathing zone con-
centrations resulting from actual bag and drum filling operations. Finally,
the results of the laboratory drumming studies were used to develop a predic-
tive model of inhalation exposure, which was tested for validity against data
from the plant-scale occupational exposure study (Task D).
-------
E. Quality Assurance
Consistent with the Quality Assurance Program Plan, specific qual-
ity control checks were used throughout the experimental phases of this study.
These are reported separately in Appendix D.
-------
I. CONCLUSIONS
The conclusions derived from the plant scale bagging and drumming
studies and from the laboratory drumming study are presented separately below.
A. Plant Studies
1. Drumming of Low Volatility Hydraulic Fluids
• Low but detectable concentration levels (^ 100 ug/m3) were
observed only when the side draft hood was not operated.
• Mist, rather than vapor concentrations, predominated at nearly
all sampling locations.
• Actual drumming operator exposure was limited by the small
percentage of time spent at the drum filling apparatus.
2. Bagging of Granular Sodium Tripolyphosphate (STP)
• Airborne suspended particulate concentrations of STP were
fairly uniform with 20% or less in the respirable size range
(< 10 urn aerodynamic diameter).
• Silt fraction of bulk STP (i.e., portion of dry material
passing a 200-mesh screen) was found to be good determinant of
relative emission rates of STP.
• Enhanced operator exposure resulted from stationary operator
position at the bagging apparatus and from ventilation air
backflow.
B. Laboratory Drumming Study
1. Emissions
• Emissions increase with vapor pressure of the drummed liquid.
• Emissions of the drummed chemical (vapor plus mist) may exceed
the value calculated as saturated vapor.
• For the chemicals tested, topfill emissions were found to ex-
ceed bottomfill emissions by a factor of 2 or more.
-------
2. Dispersion
• Plumes of vapor/mist were found to sink because of: (a) en-
trainment into drum wake with downward momentum; and (b) higher
density caused by higher molecular weight or lower temperature
in relation to ambient air.
• The bung position has little effect on downwind dispersion of
plume.
3. Worker Exposure
• Highest worker exposure was found to occur with source in
front of the worker and ventilation airflow from behind.
• Cross ventilation of drumming operation produced significantly
lower exposures than rear ventilation.
C. Model Development
• Previously developed models are inadequate because either:
(a) they do not account for spatial variability of exposure;
or (b) they require such detailed physical/chemical data and
plant-specific input data as to be impractical for assessing
exposure to new chemicals.
• The proposed model accounts for both spacial variability and
the previously uncharacterized effect of worker orientation
while requiring limited physical data.
• The proposed model was shown to successfully predict the plume
contribution to worker exposure at the plant-scale drumming
facility, using an emission factor developed from the labora-
tory study of drumming emissions in the low-flow wind tunnel.
10
-------
IV. PLANT SCALE STUDIES
The plant-scale occupational exposure study was performed on full
scale packaging systems at two plants. At Plant A, low-volatility hydraulic
fluids were packaged in 55-gal drums; and at Plant B, granular sodium tri-
polyphosphate (STP) was packaged in 50- and 100-lb bags. Integrating samplers
were deployed at fixed points surrounding each work station. Pairs of sam-
plers were collocated to provide for determination of measurement precision.
Personal samplers were operated simultaneously by plant personnel. Additional
data collected during each test included ventilation flow speeds/directions
across the work station, and characterization of worker activity (time and
motion). The studies performed at Plant A and Plant B are described in detail
in the sections immediately following.
A. Drumming Operation (Plant A)
1. Introduction
The first plant-scale study was directed to the collection of spe-
cific data on worker breathing zone concentrations resulting from the drumming
of liquids. These data will be used to develop and refine models for the
assessment of inhalation exposure for workers involved in such packaging oper-
ations.
The testing was conducted at a plant facility where low volatility
hydraulic fluids are packaged in 55-gal drums by means of an automated bottom-
fill system. This facility was judged to be representative of full-scale
drumming operations and associated emission controls. The testing took place
in April 1985.
The layout of the central drumming facility is shown in Figure IV-1.
Typically 160 to 180 drums (55 gal) are filled in a period of 4 to 5 h during
the day shift. The drumming operator spends most of the time inspecting and
labeling drums prior to filling. When the drum filling apparatus is activated
by the operator, the filling lance is automatically inserted and lowered into
the drum. As the drum is filled, the lance is automatically retracted such
that the discharge end of the lance remains under the liquid level. The ac-
tual drum fill time is approximately 1 min. After the drum is filled, a bung
cap and cover are installed by the worker using self-powered crimping tools.
The manual filling station for 5-gal drums in the central drumming
facility is operated on an irregular schedule. The operator spends about 50%
of his time filling the drum. The remainder of the time is spent getting
empty drums and placing the filled drums on a pallet. Filling of each 5-gal
drum requires approximately 20 s.
The ventilation pattern in the drumming area is influenced by wall
fans which blow air across the packaging line from the rear. When operating,
11
-------
N
Manual
Fill
0 1 2
m
Heater
OQCDO
Drumming Platform
Overhead •
Fans
Wall
Figure IV-1. Plant A drumming line.
-------
the overhead heater fan blows air laterally in the direction shown on Fig-
ure IV-1. In addition, both drum filling stations are equipped with a local
side draft hood, which pulls air from around the bung opening, thereby cap-
turing vapors discharged by each drum as it is filled.
The hydraulic fluids drummed at this facility span a range of vola-
tility (vapor pressures at 22°C from 5 mm Hg to < 0.01 mm Hg). In order to
provide for greater analytical sensitivity, the decision was made to test the
two most volatile liquids (referred to as Fluid 1 and Fluid 2) having vapor
pressures at 25°C of 5.4 mm Hg and < 2 mm Hg, respectively.
Preliminary sampling of organic vapors generated by the drumming of
Fluid 2 was conducted by MRI on February 7, 1985. Integrated samples were
collected on glass fiber filters followed by silica gel tubes (as described
below) over a period of 281 min, during which 55 drums (55-gal capacity) of
Fluid 2 and 162 drums (5-gal capacity) of Fluid 1 were filled. Analysis of
the four samples by gas chromatography yielded total mist and vapor concen-
trations of 0.060 to 0.090 mg/m3. The three samples collected very near
(< 1 m) the point of vapor release from the 55-gal drums produced nearly the
same concentration as the "background" sample collected 4 m away. This uni-
formity of concentrations was judged to be the result of secondary emissions
associated with minor spillage in the general area and/or the operation of
the 5-gal drumming station nearby.
2. Experimental Methods
This section describes the sampling and analysis procedures used in
the primary testing of the 55-gal drumming station during April of 1985.
a. Sampling Procedure
A network of 20 samplers was deployed at fixed points in the
vicinity of the 55-gal drumming operations. As shown in Figure IV-2, each
sampler consisted of a filter holder with a 20-mm glass fiber filter followed
by a tube containing activated 42/60 mesh silica gel. The tube contained
50 mg of silica gel in both a front and back section. Each silica gel tube
consisted of a front or main section and a backup section. The front and back
sections were separated by an inert divider. The two fractions were recovered
separately. A test to demonstrate the retention of the analyte by the silica
gel tube is described in Appendix D. Analysis of randomly selected backup
portions of two tubes for each analyte indicated levels beneath the limit of
detection. The absence of detectable amounts of analyte in the back half was
taken as an indication that breakthrough had not occurred. Air was drawn
through each sampler by a suction manifold connected to a vacuum pump. The
use of 25-ga. x 5/8-in. hypodermic needles as critical orifices in each line
ensured a standard flow rate of 0.30 L/min (±5%) through each sampler.
Most of the samplers were positioned at breathing height (1.5 m
above the work station platform, along the front and rear boundaries, as
shown in Figure IV-3. Four sets of samplers were colocated to provide for
determination of measurement precision. Samplers 3 through 8 were positioned
to profile for location of the plume centerline downwind of the point of
vapor release from drum filling. An operator personal sampler and additional
13
-------
Critical
Orifice
Figure IV-2.
Sampling element for collection of hydraulic
fluid mist and vapors.
14
-------
A Sampler (Prefilter
+ Silica Gel Tube)
Figure IV-3. Sampler deployment around the Plant A drumming operation.
-------
colocated samplers (of the same type as described above) were operated by
plant personnel during each test.
All samplers were operated simultaneously durina periods of
actual drumming rather than the full 8-hr work shift. Even though the mea-
sured concentrations would not represent normal workday values, they would be
substantially more useful for purposes of exposure model development and
validation.
Because the ventilation patterns and secondary source activi-
ties might strongly influence the resulting concentration field in the worker
zone, these conditions were defined with plant personnel to the extent pos-
sible before each test. It was desirable to minimize secondary source activi-
ties (e.g., operation of the 5-gal drumming line) because of the resulting
interferences with model development.
During each sampling period, a warm-wire anemometer (Hastings-
Raydist air meter) and smoke traces were used to determine the ventilation
flow pattern in the vicinity of the work station. Also the volumetric flow
rates at all major ventilation inlets and outlets of the central drumming fa-
cility were determined to provide for calculation of overall air exchange
rate.
Worker positioning and movement during each sampling period
was documented by determining the relative amounts of time spent in each work
zone. Worker activity was keyed to the number of drums filled during each
time increment, which was monitored separately. Preliminary analysis of
worker positioning was used to define the work zone, for purposes of sampler
placement as described above.
b. Analysis Procedure
After a test, each exposed sample collection medium was removed
from the sampling system, capped, and placed in a refrigerated container for
return to the laboratory for analysis. After arrival at the laboratory, the
silica gel tubes were broken and the front and backup sections transferred to
1-dram vials. The filters, already transferred to autosampler vials at the
end of each test, did not require transfer.
Samples were extracted by adding 1.0 ml of acetone, accurately
measured, to each vial and allowing the vials to sit for 30 min with occa-
sional swirling. Aliquots of the extracts from the tubes were transferred to
autosampler vials for analysis by gas chromatography. The filters were re-
moved from the extracts prior to analysis.
The silica gel and filter extracts were analyzed separately by
gas chromatography. The analytical parameters of the method were as follows.
16
-------
Gas Chromatographic System
Instrument: Varian 3700 gas chromatograph
Detector: Flame ionization
Column: DB-5 capillary, 15-m x 0.25-mm ID, fused silica
Injection mode: Splitless for 45 s
Carrier gas: Helium
Carrier gas linear velocity: 30 cm/s
Splitter flow rate: 100 cc/min
Inlet temperature: 250°C
Detector temperature: 300°C
Column oven temperature program: 40°C for 1 min, then 40 to 250°C
at 10°C/min
Data handling: Nelson analytical data system
Analyte concentrations in the extracts were determined by com-
parison of the electronically integrated peak areas of chromatograms obtained
from injection of standard solutions with peak areas obtained from injection
of the sample extracts. Response factors obtained from injection of standard
solutions of each analyte (1-100 ug/mL) were found to vary ±20% (RSD) for
Fluid 1 and ±25% (RSD) for Fluid 2 over the range of 1 to 100 |jg/mL. These
response factors were considered invariant, and the averages were used to
calculate analyte levels in the samples. The method has shown a limit of
detection (LOD) of 0.5 ug/sample (based on a 1-mL extraction volume) which
corresponds to a discernible peak above the noise level of the instrument.
The limit of quantisation (LOQ) has been set at 2 ug/sample (based on a 1-mL
extraction volume) which is 4 x LOD.
Masses found on the filter and in the silica gel were combined
to determine total mass sampled. Concentration values were calculated by:
Concentration fufl/m3} - Total mass collected (uq) x 1.000 (L/m3)
Loncentrat10n (ug/m ) - Sam1
Samp11ng rate (L/min) x Samp1ing time (min)
3. Test Results
Two tests of drumming were performed as summarized in Table IV-1.
Limitations in the drumming frequency of the two more volatile hydraulic
fluids prevented additional testing of these fluids.
As indicated in Table IV-2, the automated drum fill time for either
test fluid was approximately 1 min. Over 99% of the weight of fluid was
added in the fast fill mode at a rate of approximately 60 gpm. The lance re-
ceded from the drum during filling such that the discharge (bottom) end re-
mained below the fluid level in the drum.
During these tests of drum filling, the operator was observed to
spend about two- thirds of the time at the south end of the work platform
labeling the drums and placing loose caps on them. Approximately 30% of the
time was spent crimping the caps on the drum openings after filling. Only
about 3% of the time was spent in front of the drum filling apparatus, posi-
tioning the drum and activating the automated filling operation.
17
-------
Table IV-1. Plant Tests of Drumming Operation
No. of Test
Test Date Substance 55-gal drums duration
drummed filled (h:min)
A-l 4/23/85 Fluid 1 100 drums 3:05
A-2 4/25/85 Fluid 2 59 drums 1:38
Air
temperature
(°F)
82
70
Table IV-2. Representative Drum Fill Cycles
Fluid 1 Fluid 2
Downward. movement of lance begins
Fill starts
Fast fill ends
Slow fill ends
Lance retracted from drum
Total fill weight
Fast fill total
Slow fill total
0:00
0:03
0:54
0:57
1:00
475 Ib
472 Ib
3 lba
0:00
0:04
1:06
1:14
1:16
454 Ib
451 Ib
3 lba
In half-pound increments.
18
-------
a. Ventilation Patterns
The air velocity distribution at breathing height (1.5 m above
work platform) is shown in Figures IV-4 and IV-5 for Tests A-l and A-2, re-
spectively. Air speeds at worker breathing heights (1.5 m above the drumming
platform) were generally low (< 50 fpm) with a forward component across the
work platform. The characteristic upward component was due to (a) the higher
velocity air stream above the work station created by the overhead wall fans,
and (b) the location of primary air exhaust vents in the ceiling of the
building. The lateral (northerly) component of the air motion at the south
end of the work platform was caused by the southerly ambient wind entering
door openings along the south wall of the building.
A study of total workroom ventilation indicated good agreement
between the total air inflow and outflow of about 12,000 cfm. As shown in
Table IV-3, this volumetric flow rate produced about 26 air changes per hour
within the drumming facility.
Table IV-3. Plant A Ventilation3
Inflow total 12,711 ftVmin
Outflow total 10,789 ftVmin
Ventilation rate 11,750 ftVmin
(inflow-outflow average)
Total room volume 29,900 ft3
Free room volume 26,910 ft3
(90% of total room volume)
Room ventilation turnover rate 26.2 air changes
per hour
aVentilation study conducted on April 26, 1985.
b. Concentration Patterns
All of the Fluid 1 analyte masses collected on the prefilters
(mist) and in the silica gel traps (vapor) during Test A-l were below the 2-pg
limit of detection. Based on a sampling rate of 0.30 L/min and a sampling
duration of 185 min, the minimum detectable concentration corresponding to a
2-ug sample mass is 36 pg/m3. Thus, if the analyte mass was equally present
as mist and vapor, the total analyte concentration could have been as high as
72 pg/m3. The total analyte concentration measured by the plant laboratory
on the operator personal sampler for Test A-l was 150 pg/m3.
19
-------
Sampler (Prefilter
+ Silica Gel Tube)
••
ro
o
L
I
Wind Vector
= 50 fpm
Down,
« SL
12
*— I anr^o F
i
-------
ro
Sampler (Prefilter
+ Silica Gel Tube)
t
Wind Vector
= 50 fpm
I
Up
i@ 180 fpm t f
•• •' '•
Up
Up
Variable
QQOQQ
Lance Fill Position
Variable
10 fpm
Up
I
Up
Up
Up
' J
Note: Strong flow of air entering
drumming room from south door
(upper right corner of diagram).
r
Overhead
Fans
Figure IV-5. Overhead view of airflow at 1.5 m above drumming platform
for Test A-2 (Fluid 2).
-------
Unlike the case for Fluid 1, detectable concentrations of
Fluid 2 analyte were measured at several sampling locations in Test A-2. The
distributions of analyte concentrations in the form of vapor and mist are
shown in Figures IV-6 and IV-7, respectively. Analyte concentrations of mist
exceeded concentrations of vapor at nearly all of the sampling locations.
The highest mist concentrations occurred at sampling locations downwind of
the drum filling apparatus, as would be expected. Except for the sampling
points near the drum filling apparatus, the vapor concentrations were below
the detectable minimum of about 76 ug/m3 (assuming a 0.30 L/min sampling rate).
Although the limit of detection for the Fluid 2 analyte mass was also 2 ug,
the minimum detectable concentrations were larger than those for Fluid 1
(Test A-l) because of the shorter sampling duration (98 min) for Test A-2.
The total analyte concentration measured by the plant laboratory on the oper-
ator personal sampler for Test A-2 was 240 ug/m3.
The higher Fluid 2 concentrations measured during Test A-2 are
thought to be due primarily to the fact that the local side draft hood at the
drumming station was inadvertently not operated by plant personnel. Even
though the 5-gal drumming station was active during Test A-l along with the
55-gal drumming line, both side-draft hoods were in operation throughout this
earlier test. Analyte masses and concentrations for Tests A-l and A-2 (both
prefilter and silica gel trap values) are given in Tables IV-4 and IV-5.
B. Bagging Operation (Plant B)
1. Introduction
The second plant-scale study described herein was directed to the
collection of specific data on worker breathing zone concentrations resulting
from the bagging of solids. These data will be used to develop and refine
models for the assessment of inhalation exposure for workers involved in such
packaging operations.
The testing was conducted at a plant facility where granular sodium
tripolyphosphate is packaged in 50 and 100 Ib bags. This facility was judged
to be representative of full-scale bagging operations and associated emission
controls. The testing took place in April of 1985.
Sodium tripolyphosphate (STP) is incorporated into synthetic deter-
gent granular formulations as a "builder" or sequestering agent. It serves
to eliminate interference with the detergent action by the calcium and mag-
nesium ions (hardness) in the water used in the wash solution. STP may be
used in powder, prill, or granule form.
At Plant B, 40,000 to 60,000 Ib of tripolyphosphate are bagged on a
two-spout autopacker (see Figure IV-8) in a period of about 6 hr during the
day shift. Two grades of product are packaged—a food grade and a technical
grade. Data on the product texture as provided by plant personnel indicate
that these powders have silt contents (i.e., portions passing a 200-mesh
screen) of only 1 to 2%. According to plant personnel, worker exposure data
based on periodically collected personal samples indicate maximum exposure
levels of 2 to 3 mg/m3.
22
-------
oo
• Sampler (Prellller
-» Silica Gel Tube)
<69
<80
<74
Figure IV-6. Vapor concentrations (yg/m3) of Fluid 2 observed
during Test A-2.
-------
• Sampler (Pretlller
+ Silica Gel Tube)
169
108
ro
Trap Only (Vapor and Mist)
Figure IV-7.
Mist concentrations (yg/m3) of Fluid 2 observed
during Test A-2.
-------
Table IV-4. Mass Concentrations Observed During
Test A-l (185-min test period)
Sampler
location
ID
1
2a
2b
3
4
5
6
7
8a
8b
9a
9b
10
11
12
13
14
15a
15b
16
17
Sampler
flow rate
(L/min)
0.301
0.300
0.307
0.299
0.271
0.274
0.276
0.279
0.271
0.300
0.285
0.309
0.276
0.258
0.275
0.288
0.300
0.275
0.273
0.307
0.280
Total analyte
(UQ)
Mist
(filter)
< 2a
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
Vapor
(trap)
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
Mass
concentration
(ug/rn3)
Mist
< 36
< 36
< 35
< 36
< 40
< 40
< 39
< 39
< 40
150b
< 38
180b
< 39
< 42
< 39
< 38
< 36
< 39
< 40
< 35
< 39
Vapor
< 36
< 36
< 35
< 36
< 40
< 40
< 39
< 39
< 40
< 38
< 39
< 42
< 39
< 38
< 36
< 39
< 40
< 35
< 39
bAnalyte masses of < 2 pg are below the LOQ.
Plant laboratory analysis for combined filter and trap.
25
-------
Table IV-5. Mass Concentrations Observed During
Test A-2 (98-min test period)
Sampler
location
ID
1
2a
2b
3
4
5
6
7
8a
8b
9a
9b
10
11
12
13
14
15a
15b
16
17
Sampler
flow rate
(L/min)
0.296
0.272
0.294
0.296
0.260
0.264
0.270
0.286
0.264
0.296
0.274
0.294
0.272
0.252
0.265
0.272
0.286
0.264
0.264
0.300
0.274
Total analyte
(UQ)
Mist
(filter)
4.9
2.9
4.0
2.3
4.6
3.5
< 2
a
2.2
3.8
2.9
< 2
< 2
2.5
< 2
< 2
3.1
3.2
2.9
Vapor
(trap)
< 2a
< 2
< 2
2.4
2.1
< 2
2.6
3.2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
< 2
Mass
concentration
(ug/m3)
Mist
170
110
140
79
180
140
< 76
b
85
< 210b
140
240C
110
< 81
< 77
94
< 71
< 77
120
110
110
Vapor
< 69
< 75
< 69
83
82
< 77
98
110
< 77
< 74
< 75
< 80
< 77
< 75
< 71
< 77
< 77
< 68
< 74
.Analyte masses of < 2 pg are below the LOQ.
Silica gel trap only.
Plant laboratory analysis for combined filter and trap.
26
-------
ro
Fine
Particle
Collection
Barrel
Coarse
Particle
Collection
Barrel
Control
Panel
Pallet
Conveyor
28 Inch
Overhead
Fan
J
1
m
Figure IV-8. Plant B bagging facility.
-------
The operator, seated in front of the autopacker during its opera-
tion, slides a bag onto one of two spouts and then activates the flow of
powder. Some visible emissions can be observed as the bag is filled. Addi-
tional emissions occur when each self-sealing bag is tipped from the spout
onto the conveyor below.
The ventilation airflow across the packaging equipment is strongly
influenced by the 28-in. overhead fan to the rear of the operator seat. The
local side-draft hood behind the spouts for capture of dust emissions does
not induce a significant air draft in front of the spouts near the worker
breathing zone.
2. Experimental Methods
This section describes the sampling and analysis procedures used in
the testing of the bagging station in April of 1985.
a. Sampling Procedure
A network of 24 samplers (12 pairs) was deployed at fixed
points in the vicinity of the bagging operation. As shown in Figure IV-9,
both elements of the sampling pair collected particulate on 37-mm Millipore®
filters (0.2 pm pores) in cassettes. One sampler in each pair was fitted
with a Dorr-Oliver 10-mm nylon cyclone preseparator. At a flow rate of
2 L/min, this cyclone has a 50% cut point of about 3.5 pm aerodynamic diameter
for measurement of respirable particulate concentrations. The other sampler
in each pair was used to measure total suspended particulate concentrations.
A standard flow rate of 2.0 L/min (±5%) through each filter
cartridge was ensured throughout substrate buildup by maintaining sonic flow
through a critical orifice. A 20 ga. x 1 in. hypodermic needle served as the
critical orifice in each line connecting a sampler to the suction mainfold.
Most of the samplers were positioned at breathing height along
the boundaries of the bagging equipment, as shown in Figure IV-10. Two sets of
samplers were colocated to provide for determination of measurement precision.
Samplers 3a and 3b were positioned for location of the plume centerline down-
wind of the points of particulate release. An operator personal sampler
fitted with a 10 mm cyclone (of the same type as described above) was op-
erated by plant personnel during each test.
Further particle size characterization was provided by a Sierra
Instruments Model 230 high-volume (20 cfm) five-stage cascade impactor with a
cyclone preseparator. This equipment was operated near the point of release
to measure particle size distribution in the range of 1 to 15 umA. In this
study, the particle size fractions < 15 pmA are referred to as inhalable par-
ticulate. Finally, a RAM-1 aerosol analyzer was used periodically to monitor
the time dependence of particle concentrations near the bagging operation.
All integrating samplers were operated simultaneously during
periods of actual bagging rather than the full 8-hr work shift. Even though
the measured STP concentrations may not represent normal workday values, they
are substantially more useful for purposes of model refinement and validation.
28
-------
Respirabje Dust|Sampler
a
• Nylon Cyclone Presepafator
Figure IV-9.
Sampling elements for collection of sodium
tripolyphosphate particles.
29
-------
Cassette Sampler Pair
O Cyclone/Cascade Impactor
Figure IV-10.
Sampler deployment around the Plant B bagging
operation.
30
-------
During each sampling period, warm-wire anemometers and smoke
traces were used to determine the ventilation flow pattern in the vicinity of
the work station. Also, the volumetric flow rates at all major building
ventilation inlets and outlets were determined to provide for calculation of
overall air exchange rate.
b. Analysis Procedure
At the end of each particulate sampling period, the cartridges
with the Millipore® filters were removed from the sampling system. Plastic
stoppers were replaced on the cartridge inlets and outlets, and the cartridges
were placed in protective containers for shipment to the laboratory. After
each day of testing, the cyclones were washed with distilled water and air
dried.
The exposed Millipore® filters in their respective cassettes
were returned to the MRI laboratories to be analyzed for phosphate ion with a
Technicon AutoAnalyzer system. The specific reagent preparation and analysis
procedures were adapted from those found in EPA Method 365.1 for phosphorous
in water and wastes (EPA-600/4-79-020).
In the Method 365.1 analysis scheme, sulfuric acid hydrolysis
converts the polyphosphates to the orthophosphate form for reaction with
ammonium molybdate and antimony potassium tartrate. The antimony-
phosphomolybdate complex formed by this reaction is reduced to an intensely
blue-colored complex by ascorbic acid. The color is proportional to the
phosphorous concentration and is measured with the Technicon AutoAnalyzer.
The following modifications to Method 365.1 were incorporated:
1. The system was standardized with sodium tripolyphosphate (STP)
from the fill line. The STP was hydrolized following steps 8.2
through 8.2.3 of Method 265.1. Standardization with STP from
the fill line permitted the results to be presented directly
as ug of STP.
2. Brig-35 (nonionic wetting agent) was not used in the system
because it interfered with the analysis.
3. Acid wash water was not used because it interfered with the
analysis.
4. The stable solution (Note 1: under step 7.5) was not used be-
cause it proved unsatisfactory. The single mixed reagent was
used.
5. The sample rate (Figure 3 of method) was changed to 0.6 mL/min.
6. The mixed reagent rate was changed to 0.4 mL/min.
7. The waste rate was changed to 0.8 mL/min.
8. The delay coil in the heating bath was changed from 13 turns to
22 turns to increase the development time.
31
-------
The phosphate analysis of each participate sample was initi-
ated by removing the filter from the cartridge and placing the filter (ex-
posed side up) in a bottle with a diameter slightly larger than the filter
diameter. After 30 mL of deionized water was added, the solution was boiled
for 10 to 15 min to a volume of 20 mL. The filter was removed from the solu-
tion and washed three times with deionized water, with the wash water added
to the sample. After 1 mL of 11 N sulfuric acid solution was added to the
sample solution, it was gently boiled for 30 to 40 min to obtain a volume of
less than 10 ml. Then the solution was cooled and made to 10 mL in a volu-
metric flask and then refrigerated. This procedure is detailed in Table IV-6.
Table IV-6. Sodium Tripolyphosphate Analysis Protocol
(based on EPA Method 365.1 for phosphorous)
1. Place plugs in both ends of filter cassette and air blow off any par-
ticulate from outside of cassette. Do this in an area which cannot be
contaminated.
2. Open cassette and observe filter condition.
a. If filter is torn, then place both backup filter and Millipore®
filter in ~ 60 mL Pyrex® bottle. Otherwise place only Millipore®
filter in each uniquely labeled filter bottle.
b. Rinse front half of cassette interior with DI water (three small
rinses) and add rinse water to filter bottle. Add DI H20 as
necessary to attain approximately 30 mL total volume for each
filter sample.
3. Take water blanks from unused cassettes to check handling and DI water
contamination. (Minimum of one DI water blank per day in addition to
field blanks.)
4. Boil the solutions in Pyrex® bottles (containing the filter/rinse water)
for about 10-15 min to reduce volume to ~ 20 mL.
5. Remove filter from bottle and rinse filter with DI water using funnel.
Rinse volume should be added to bottle to total approximately 30-40 mL
volume. Discard filter.
6. Add 1 mL of 11N H2S04 to the 30-40 mL sample volume.
7. Again, distill off water to obtain a sample volume of less than 10 mL.
Allow the sample to cool to room temperature. Pour the concentrate into
a 10 mL volumetric flask. Dilute the sample to 10.0 mL volume. Transfer
the sample back to the original Pyrex® container, cap, and store under
refrigeration for subsequent analysis using the Technicon AutoAnalyzer.
aAll glassware should be washed with hot 1:1 HC1 and triple rinsed with
distilled H20.
32
-------
The sample analysis was initiated by bringing the samples to
room temperature. The sample and container were weighed and recorded as value
(A). One drop of phenolphthalein indicator was added to each sample solution,
and the solution taken to the end point with IN sodium hydroxide solution.
The container was then reweighed with the value recorded as (B), and the solu-
tion transferred to a clean container. The original container was dried and
weighed and recorded as value (C).
The total mass of sodium tripolyphosphate in the original sam-
ple was calculated as follows:
Total ug STP =
_ B-C
A-C
ug/mL value from Technicon • 10 mL
Mass concentrations were calculated by:
r«n^Qn+M+,-«n f„f,/m3^ - Total mass collected (pq) x 1,000 (L/m3)
Concentration (ug/m3) - Sampling rate (L/min) x sampling time (min)
3. Test Results
Three tests of drumming were performed as summarized in Table IV-7.
The first two tests (B-l and B-2) involved the packaging of technical grade
STP, and the third test (B-3) involved food grade STP.
Table IV-7. Plant Tests of Bagging Operation
Test date
4/23/85
Packaged
substance
Sodium tri poly-
Packaging
operation
80 100- Ib bags
Test
duration
(hr:min)
2:11
Air
temperature
77
4/24/85
4/25/85
phosphate (tech-
nical grade)
Sodium tripoly-
phosphate (tech-
nical grade)
Sodium tripoly-
phosphate (food
grade)
and 200 50-Ib
bags
514 50-Ib bags
736 50-Ib bags
3:19
2:54
63
78
33
-------
a. Ventilation Patterns
The air velocity distribution at operator beathing height is
shown in Figures IV-11 through IV-13 for Tests B-l, B-2, and B-3, respectively.
The overhead fan at the rear of the operator produced the primary flow pattern
with speeds of several hundred feet per minute. This primary flow stream drew
air from either side and blew it across the packaging equipment as shown.
A study of total workroom ventilation indicated good agreement
between the total air inflow and outflow of about 23,000 cfm. As shown in
Table IV-8, this volumetric flow rate produced about 15 air changes per hour
within the packaging facility.
Table IV-8. Plant B Ventilation3
Inflow total 21,583 ftVmin
Outflow total 24,200 ftVmin
Ventilation rate 22,892 ftVmin
(inflow-outflow average)
Total room volume 98,925 ft3
Free room volume 89,032 ft3
(90% of total room volume)
Room ventilation turnover rate = 15.4 room changes/hr
aVentilation study done April 26, 1985.
b. Concentration Patterns
Analyte masses and calculated concentrations for Tests B-l,
B-2, and B-3 are given in Tables IV-9, IV-10, and IV-11, respectively.
Analysis of the samples obtained at location 3b and the personal samples was
performed by the plant laboratory.
For each test, the concentration pattern was found to be fairly
uniform (as shown in Figures IV-14 through IV-16), with only a small fraction
of suspended particulate concentration consisting of STP particles in the
respirable size range (^ 10 umA). The airborne particle size distributions
for the three tests, as determined from the high-volume cyclone/cascade im-
pactor data, are plotted in Figure IV-17.
34
-------
Wind Vector
= 200 fpm
a
t
! 900 fpm
Up
t
• Filling •
Spouts
Operatoc/Seat
Variable
Cassette Sampler Pair
Cyclone/Cascade Impactor
A
28 Inch
Overhead
Fan
Figure IV-11. Overhead view of airflow at operator
breathing height during Test B-l.
35
-------
Wind Vector
= 200 fpm
^ r-
c
N
vup
•
rn
<
i
i
4
i
k
Down
i
i
Filling
Sn/*\i i+o
-
\
Up
•
| —
"1
Up @ 60 fpm
• Cassette Sampler Pair
Cyclone/Cascade Impactor
Operator Seat
t
illo:
Up;Variable
A
28 Inch
Overhead
Fan
Figure IV-12. Overhead view of airflow at operator
breathing height during Test B-2.
36
-------
Wind Vector
= 200 fpm
a
i
\
500 fpm
Down
• Filling •
Spouts
Operator Seat
UP
UP
T
Cassette Sampler Pair
Cyclone/Cascade Impactor
A
28 Inch
Overhead
Fan
Figure IV-13. Overhead view of airflow at operator
breathing height during Test B-3.
37
-------
Table IV-9. Mass Concentrations Observed During Test B-l
(131-min sampling period)
Sampler
ID
1
1- Cyclone
2
2-Cyclone
3a
3a-Cyclone
3b
3b-Cyclone
4
4-Cy clone
5
5-Cyclone-
6
6-Cyclone
7a
7a-Cyclone
7b
7b-Cyclone
8
8-Cyclone
9
9-Cyclone
« -, d
Cyclone-
Cascade Impactor6
Sampler
flow rate
(L/min)
2.00
2.0a
^l
2.0a
1.90
2.0a
2.00
2.0
1.95
2.0a
1.90
2.00
2.00
1.95
2.10
2.05
2.00
2.05
1.95
2.05
1.95-
2.20
566*
566
Total analyte
as Na5P307 (pg)
386
19
239
131
328
b
114c
< 35C
258
55
697
98
421
23
158
45
301
16
370
30
442
< 3
55; 700,
60,500
STP
concentration
(M9/m3)
1,500
70
940
500
1,300
b
440
< 130
1,000
210
2,800
370
1,600
90
570
170
1,200
60
1,400
110
1,700
< 10
740
800
a
Estimated value.
"Sample destroyed.
.Plant laboratory analysis.
Particles > 15 pmA; 134-min sampling time.
Particles < 15 pmA including background; 134-min sampling time.
38
-------
Table IV-10. Mass Concentrations Observed During Test B-2
(199-min sampling period)
Sampler
ID
1
1- Cyclone
2
2-Cyclone
3a
3a-Cyclone
3b
3b-Cyclone
4
4-Cyclone
5
5-Cyclone
6
6-Cyclone
7a
7a-Cyclone
7b
7b-Cyclone
8
8-Cyclone
9
9-Cyclone
« , b
Cyclone
Cascade Impactor0
Sampler
flow rate
(L/min)
2.00
1.95
1.95
2.05
1.95
1.95
2.00
2.10
2.00
1.95
1.95
2.20
2.00
1.95
2.00
2.05
2.05
2.05
1.95
2.05
2.05
1.95
566
566
Total analyte
as Na5P307 (pg)
513
86
286
67
303
44
298a
< 35a
598
23
845
< 3
687
51
713
16
566
49
7,480
61
1,400
201
168,000
78,400
STP
concentration
(|jg/m3)
1,300
220
740
160
780
110
750
< 80
1,500
60
2,200
< 10
1,700
130
1,800
40
1,400
19,000
150
3,400
520
1,600
750
.Plant laboratory analysis.
Particles > 15 jjmA; 184-min sampling time.
Particles < 15 umA including backbround; 184-min sampling time.
39
-------
Table IV-11. Mass Concentrations Observed During Test B-3
(174-min sampling period)
Sampler
ID
1
1-Cyclone
2
2- Cyclone
3a
3a-Cyclone
3b
3b-Cyclone
4
4- Cyclone
5
5-Cyclone
6
6- Cyclone
7a
7a-Cyclone
7b
7b-Cyclone
8
8- Cyclone
9
9-Cyclone
r* , fa
Cyclone
Cascade Impactorc
Flow rate
(L/min)
2.10
2.00
2.10
2.10
1.90
1.90
2.00
2.20
1.95
2.00
1.95
2.05
2.05
1.95
1.95
2.00
2.00
2.05
1.95
2.05
1.95
2.05
566
566
Total analyte
as Na5P307 (ug)
794
156
2,520
65
722
247
483a
< 35a
927
174
2,800
188
1,530
145
693
105
997
307
1,020
71
1,050
110
64,800
61,800
STP
concentration
(ug/m3)
2,200
450
6,900
180
2,200
750
1,400
< 90
2,700
500
8,200
530
4,300
430
2,000
300
2,900
860
3,000
200
3,100
310
1,300
1,200
.Plant laboratory analysis.
Particles > 15 umA; 88-min sampling time.
Particles < 15 umA including background; 88-min sampling time.
-------
1450
1470.
940
1010
1730
1530
Bagging Spout
• Cassette Sampler
O Cyclone/Cascade Impactor
Figure IV-14a. Total suspended particulate concentrations of
STP (jjg/m3) for Test B-l.
41
-------
110
210
7401
Bagging Spout
• Cassette Sampler With
Cyclone Precollector
O Cyclone/Cascade Impactor
aSample destroyed
blnhalable Particulate,
Figure IV-14b. Respirable particulate concentrations of STP
(ug/m3) for Test B-l.
42
-------
1500
• Cassette Sampler Pair
O Cyclone/Cascade Impactor
aSample Contaminated
Figure IV-15a. Total suspended particulate concentrations of STP
(ug/m3) for Test B-2.
43
-------
150
60
750'
Bagging Spout
Cassette Sampler With
Cyclone Precollector
O Cyclone/Cascade Impactor
alnhalable Particuiate
Figure IV-15b. Respirable particulate concentrations of STP
(|jg/m3) for Test B-2.
-------
fin
2170.
2730!
3100
3020
2540
Bagging Spout
• Cassette Sampler
O Cyclone/Cascade Impactor
Figure IV-16a. Total suspended particulate concentrations of STP
(jjg/m3) for Test B-3.
45
-------
200
450
310
180
500
1240*
Bagging Spout
• Cassette Sampler With
Cyclone Precollector
O Cyclone/Cascade Impactor
alnhalable Paniculate
Figure IV-16b. Respirable particulate concentrations of STP
(|jg/ni3) for Test B-3.
46
-------
WEIGHT % GREATER THAN STATED SIZE
999 99.8 99 98 95 90 80 70 60 50 40 30 20 10 52 1 050201
100
50
20
^ 10
(0
O
K
o
S S
oc
in
<
5
UJ y
PARTICL
05
0.2
0.1
I
|
Jb
J
i
B-2-
y
/
/
/ ^
/
/
-B-3
•M
/
/
u
f
r
t
/
//
/
y
//.
f
V
«-
p
-I
3-1
10
50
10
2
05
U £
01
WEIGHT % LESS THAN STATED SIZE
Figure IV-17. Airborne particle size distributions obtained using
cyclone/cascade impactor.
47
-------
In the relatively stagnant flow region immediately in front of the operator,
it is likely that gently circulating air transported emissions from the bag-
ging operation directly to the breathing zone of the operator.
The concentrations observed during Test B-3 are significantly
higher than the values measured for Tests B-l and B-2. This resulted from
the higher fraction of fines in food grade STP compared to technical grade
STP, as detailed below.
c. Bulk Particle Size Distributions
Three representative samples of the bagged STP were obtained
on the 3 days of testing at Plant B, to be analyzed for particle size distri-
bution. These samples were dry sieved, yielding the results shown in Table
IV-12. Also shown are dry sieving results from plant laboratory analysis of
earlier samples of food grade and technical grade STP.
Table IV-12. Bulk STP Particle Size Distribution
Cumulative percent
Mesh
No.
10
16
20
30
40
50
60
80
100
140
200
Particle
diameter
(urn)
> 2,000
> 1,190
> 850
> 590
> 425
> 297
> 250
> 180
> 150
> 106
> 75
< 150
< 75
Plant
Food
grade
0
0.13
3.6
17
-
57
-
79
87
-
~
13
""
B analyses
Technical
grade
0.002
0.016
4.5
21
-
67
-
86
91
-
~
9
™
Technical
grade
(Test B-l)
-
-
2.6
32
61
-
89
95
97a
99
99
2.9a
0.71
MRI analyses
Technical
grade
(Test B-2)
-
-
-
30
56
-
85
93
96a
98
99
4.1a
0.79
Food
grade
(Test B-3)
_
-
-
17
38
-
72
84
89a
95
97
II3
3.1
Value obtained by interpolation using log normal distribution.
48
-------
As indicated in Table IV-12, the MRI analysis of the technical
grade STP yielded a smaller weight fraction of finer particles (particles
smaller than 150 |jm in physical diameter). If the process for the manufacture
of technical grade STP is not as closely controlled as for food grade STP,
significant differences in bulk size distributions could occur. The MRI and
plant analyses of the food grade STP were comparable for the < 150 pm size
fraction.
49
-------
V. LABORATORY DRUMMING STUDY
In the first laboratory pilot study, a walk-in flow tunnel was used
to characterize the low-flow (^ 2 m/s) dispersion of methane gas and ammonium
chloride particles at distances of 2 to 6 m from the point of release. The
release configurations included: single- and double-point releases into an
unobstructed flow, point release near a wall, point release past a side-draft
hood, and bung release from a 55-gal drum (with and without a close fitting
hood). A two-dimensional array of samplers was used to characterize disper-
sion patterns at each downstream distance. The observed dispersion patterns
were compared to existing dispersion theories and then used as the basis for
a modified dispersion algorithm which balanced realism against ease of appli-
cation. The predictive capability of the resulting dispersion model was
found to be significantly better than the other models tested against the
laboratory data. The first laboratory pilot study is not further discussed
here. (For details refer to Appendix C.)
The second laboratory pilot study was conducted using a drumming
station at MRI and encompassed the measurement of emissions, dispersion, and
operator exposure resulting from a full-scale 55-gal drumming operation in
the low-flow wind tunnel. A 55-gal drum with an added bottom drain was posi-
tioned on a platform within the flow tunnel. A 300-gal reservoir was in-
stalled outside the tunnel at a height of 2.13 m (7 ft) above the floor. The
overhead tank was fitted with polyethylene hosing and a pump for transfer of
the test liquid to the 55-gal drum at a fixed rate controlled by the hose
nozzle. After each drum filling, the contents of the drum were pumped back
to the overhead tank. The test liquids (ethylene glycol, perchloroethylene,
and methanol) spanned a vapor pressure range from about 0.1 to 125 mm Hg at
25°C, as indicated in Table V-l. The portions of the second laboratory study
having to do with emissions, dispersion, and exposure are described separately
in the sections immediately following.
A. Emission Tests
1. Experimental Apparatus
The test facility used for this study was the pull-through flow
tunnel constructed in MRI Field Station Building No. 1. As shown in Fig-
ure V-l, the cross section of the tunnel was 2.7 m wide by 2.1 m high. The
centerline airflow velocity was adjustable over a range of mean values from
0.5 to 2.2 m/s (0.9 to 4.9 mph). The flow tunnel is described further in
Appendix C.
A standard 55-gal steel drum was positioned on a platform within
the flow tunnel, as illustrated in Figure V-2, such that the top of the drum
was 1 m above the tunnel floor. The drum was equipped with a bottom outlet
consisting of 1-1/4 in. cast iron pipe connected by a flange to the center of
the drum bottom. A lever-lock removable lid provided for easy access to the
interior of the drum.
51
-------
Table V-l. Chemicals Used in Drumming Simulations
en
ro
Chemical
Name
Perchloroethylene
Methanol
Ethylene glycol
Formula
C2C14
CH40
C2H602
Molecular
weight
166
32.0
62.1
Surface tension
(dynes/cm @ 20°C)
31. 7b
22. 6C
48.4
Viscosity
(cP @ 20°C)
0.798 @ 30°C
0.55
19.83
0.0198
Vapor pressure3
(Torr @ 25°
18.5
125
0.098
C)
Vapor pressures at other temperatures may be calculated using the Antoine Equation:
logloP = A - B/(C + t)
where P is the vapor pressure of the compound in mm of mercury (Torr) and where t is the
temperature in degrees centigrade.
Perchloroethylene
Methanol
Ethylene glycol
Antoine equation constants
ABC
7.02003 1415.49 221.0
7.87863 1473.11 230.0
8.2621 2197.0 212.0
Temp.
range (°C)
-20 to +140
25 to 112
In contact with vapor.
In contact with air.
-------
OVERHEAD VIEW OF PULL—THROUGH
FLOW TUNNEL
cn
co
-4m-
•8 m
o
55-Gal Drum
Wind Flow
Figure V-l. Overhead view of pull-through flow tunnel.
-------
Vent:
in
-£>
Figure V-2. Full scale 55-gal drumming operation in flow tunnel
-------
A 300-gal steel reservoir was installed outside the tunnel at a
height of 2.13 m (7 ft) above the floor. This overhead tank was fitted with
a cross-linked polyethylene hose for transfer of the test liquid to the
55-gal drum. A model 26 Gasboy vane type pump, with direct drive connection
to a 115 V AC explosion proof motor, provided for liquid delivery to the
55-gal drum at up to 30 gal/min. The pump was also used to return the con-
tents of the drum back to the overhead tank. A series of stainless steel
full flow ball valves and innerconnecting chemically resistant hose was used
to direct the liquid flow and control its rate of flow. The internal diameter
of the hosing was 3.18 cm (1.25 in.) upstream of the pump and 2.54 cm
(1.0 in.) downstream of the pump.
The liquid delivery apparatus provided for either top filling or
bottom filling of the drum. The top fill nozzle consisted of a 6-in. length
of 1-in. pipe connected by a 90-degree elbow to a section of hose fitted with
a ball valve for control of flow rate. The bottom fill nozzle consisted of a
36-in. length of pipe connected in the same manner. During drum filling,
liquid volume was monitored by the totalizing flowmeter connected to the
pump. A thermometer was installed in the flow line at the pump outlet for
monitoring of liquid temperature.
The primary analytical device for these studies was the Beckman 402
total hydrocarbon analyzer. To measure emissions, a Teflon sampling line
(1/8-in. I.D. by 13 to 20 ft in length) drew vapor effluent from the bung
opening surrounding the fill nozzle as the drum was being filled. A 3-in.
long pipe was screwed into the bung to provide an extension for prevention
of air infiltration into the bung effluent. The sample was delivered to the
analyzer at a rate of 2.2 to 3.6 L/min such that the residence time in the
sampling line was between 0.5 and 1.5 sec.
The output of the analyzer was fed to an EPSON HX-20 microcomputer
for sequential calculation of 5-sec average analyte concentrations. A Heath
continuous recorder with a chart speed of 1 in/min produced a time concentra-
tion profile of the emissions.
During an emissions test, the drum fill rate was held constant
over the entire filling period. The shut-off time corresponded either to a
certain liquid height on the sight tube or a reading of 53.6 gal on the volume
meter, which was reset to zero prior to each test. The fill time was recorded
with a stop watch, and the liquid temperature in the fill line was recorded
at the beginning and end of the fill.
In the course of the experiments, tests of emissions from an ini-
tially "dry" drum, and from a just emptied ("wet") drum were performed. Prior
to the tests of perchloroethylene emissions from a dry drum, the drum was
cleaned using a cotton mop and a stream of compressed air. Then a rubber
stopper (to which was attached a long wire for removal purposes) was placed
in the drum drain opening. All tests of ethylene glycol emissions were per-
formed with a dry drum created by inserting a clean polypropylene bag into
the drum. The bag was pressurized to set against the drum walls, and then
the open end of the bag was wrapped over the rim of the drum and the lid
affixed.
55
-------
2. Test Results
Emission tests were performed for two chemicals, perchloroethylene
(perc) and ethylene glycol ; methanol was not tested because of the potential
explosion hazard. Top and bottom fill modes and three fill rates ranging from
10 gal/min to 27 gal/min were included in the experimental design. In addi-
tion, both the filling of a dry drum and a wet drum were tested. A summary
of drumming emission tests is shown in Tables V-2 and V-3 for perchloroethyl-
ene and ethylene glycol, respectively.
For each drum filling test, the emission factor (mass loss per drum
fill) was determined as follows. After the 5-s integrated concentrations from
the Epson microcomputer were averaged over the total fill time, then the re-
sulting mean volumetric concentration (ppm) was corrected for decay in sensi-
tivity of the Beckman 402 over time. Next the adjusted mean concentration
value was converted to a mass concentration value (mg/m3), and then combined
with the volume of air displaced from the drum by the fluid. Finally, the
total chemical loss per drum fill was calculated as follows:
L =
CV V0 MW)/1Q6 VM
where: L = chemical mass loss per fill operation, g
Cy = mean of 5-s integrated chemical concentrations observed
in air drawn from drum during fill operation, ppm (or L of
chemical vapor in 106 L of air)
VQ = drum volume filled (equal to the volume of displaced air
from drum), gal
1^ = molecular weight of drummed chemical, g/mole
VM = molar volume of chemical vapor emitted from drum at ambient
temperature and pressure, L/mole
and k = 3.785 L/gal
During the top filling of perc, a considerable amount of froth was
observed to collect on the liquid surface. This was also present during bot-
tom filling, but to a lesser extent. Observed perc concentrations were above
the saturation concentration of about 25,000 ppm (volumetric) at 25°C. A
sampling line tip with internal diameter of 3 mm, 7 mm, or 11 mm was selected
to provide for near-isokinetic sampling of perc at the respective fill rate.
The production of mist from the perc froth is believed to account for concen-
trations as high as 38,000 ppm at the drum bung opening during fill cycles.
Ethylene glycol produced much less froth and presumably less mist.
For purposes of separating vapor emissions from mist emissions, a
10 mm Dorr-Oliver nylon cyclone precollector with 50% cutpoint of about
3.5 umA was attached to the inlet of the sampling line for about half of the
ethylene glycol tests.
56
-------
Table V-2. Emissions Test Data - Perch!oroethylene
Date
8/1/85
8/1/85
8/1/85
8/2/85
8/2/85
8/2/85
8/2/85
8/2/85
8/2/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/5/85
8/6/85
8/5/85
8/5/86
8/6/85
V6/85
8/6/85
8/9/85
8/9/85
8/9/85
8/9/85
8/9/85
Run
no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Initial
drum
condition
Dry
Dry
Dry
Dry
Dry
Dry
Wet
Wet
Wet
Dry
Dry
Dry
Wet
Wet
Wet
Wet
Wet
Wet
Dry
Dry
Dry
Dry
Dry
Dry
Wet
Wet
Wet
Dry
Dry
Dry
Dry
Dry
Filling
mode
Top
Bottom
Top
Top
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Top
Top
Top
Top
Top
Top
Top
Top
Top
Top
Top
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Top
Top
Top
Bottom
Bottom
Fill rate
(gal /mi n)
24.8
25.7
25.5
25.5
25.5
25.5
24.4
24.4
25.5
25.5
25.5
25.5
25.6
24.4
24.4
9.8
10.3
9.1
11.2
11.2
10.3
11.9
11.4
11.2
11.2
11.7
11.4
22.4
16.8
16.8
16.3
16.7
Fluid
temp. (°C)
21
22
22
25
26
26
27
27
27
23
24
24
25
26
23
27
28
29
29
30
31
27
32
33
28
29
29
29
29
30
31
32
Total
mass
loss
(g)
42.0
18.5
46.5
45.0
19.1
19.5
30.8
31.1
32.4
25.2
42.2
41.8
47.8
46.2
47.4
50.0
51.2
50.0
52.4
52.8
54.7
16.4
13.8
11.5
31.8
34.7
31.7
43.6
44.9
46.4
22. 0
25.0
Drum
emission
rate3
(mg/sec)
324
148
369
357
151
155
234
236
257
200
335
331
380
351
360
152
164
141
182
184
175
60.7
48.9
40.0
111
126
112
304
235
242
112
130
Based on 53.6-gal total fill volume.
57
-------
Table V-3. Emissions Test Data - Ethylene Glycol
en
00
Date
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/11/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
9/13/85
Test
no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Initial
drum
condition
Dry
Wet
Wet
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Dry
Filling
mode
Top
Top
Top
Top
Top
Top
Top
Top
Top
Top
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Bottom
Top
Top
Top
Top
Top
Top
Fill rate
(gal/min)
27.1
27.0
27.2
26.1
27.2
27.5
26.5
26.3
26.3
26.5
26.5
26.1
26.0
26.0
9.7
10.5
9.8
9.6
10.6
9.7
10.2
11.3
10.4
22.4
20.9
Fill
volume
(gal)
53.8
53.9
53.9
53.6
53.9
54.0
53.9
53.9
52.1
52.6
53.9
53.9
52.4
52.8
53.9
53.9
52.8
52.6
53.9
53.9
53.9
52.6
52.4
53.9
53.9
Fluid
temp. (°C)
22
22
22
22
22
23
23
23
23
23
23
24
24
24
17
18
18
20
20
20
21
21
21
21
21
Cyclone
(Y/N)
N
N
N
N
N
Y
Y
N
Y
Y
N
N
Y
Y
N
N
Y
Y
N
N
N
Y
Y
N
N
Total mass
loss
(g/drum)
a
a
a
2.0
2.2
b
b
b
1.9
1.9
0.5
0.39
0.40
0.33
0.31
0.22
0.21
0.22
0.20
1.1
1.1
1.0
1.0
1.1
1.1
Drumming
emission rate
(mg/s)
a
a
a
16.2
18.5
b
b
b
16.0
16.0
4.3
3.1
3.3
2.7
0.93
0.71
0.65
0.67
0.66
3.3
3.5
3.6
3.3
7.6
7.1
.Test not used in summary; no bag in drum.
Sampling inlet with cyclone "contaminated"
with fresh air; very irregular chart trace.
-------
The mist generated in both perc and ethylene glycol tests is be-
lieved to have consisted mostly of particles greater than 50 |jm. These
particles would have a tendency to settle out near the source.
Continuously recorded concentration-time profiles of the dry drum
emission tests are presented in Appendix D together with information on ob-
served maximum 5-sec mean concentrations and the integrated mass lost during
each drum fill. Figure V-3 presents a characteristic comparison of profiles
for top fill versus bottom fill of perc. The top filling of a drum causes
exit stream concentrations to rise quickly with time as the nozzle sprays
liquid into the drum, resulting in considerable splashing and mixing with
air. During a bottom fill, the spray of liquid produced by the lance at the
bottom of the drum is quickly submerged causing concentrations to rise more
slowly with time.
To determine whether the concentration-time profiles of ethylene
glycol were similar to those of perc over similar fill times, the 5-s con-
centration values at different points into the fill cycle were normalized to
the respective maximum concentration values. A summary of this analysis is
presented in Table V-4. For the top-fill tests of perc at the fast (25 gal/
min) fill rate, emissions had reached 91.4% of the maximum recorded 5-s con-
centration value at 25% total fill time (about 30-35 s). In the case of
ethylene glycol, however, only 33% of the concentration maximum had been
reached at 25% of the fill time. The opposite occurred during the fast
bottom-fill tests which showed a much more pronounced concentration increase
for the ethylene glycol than for perc. Similar effects were also observed in
the slow fill (11 gal/min) tests. The bursting of perc bubbles from the froth
and consequent injection of mist into the upper part of the drum is believed
to have caused the quick increase of concentrations observed throughout perc
top filling tests. During perc bottom filling, however, less froth is created;
fewer bubbles are available to burst causing mist; and consequently, much slower
rates of concentration increases are seen in the early stages of a fill cycle.
Table V-5 presents mean emission rates and emission factors for 39
tests of dry drum filling. A comparison of top versus bottom filling of perc
shows that top filling produces from 2 to 4 times the emissions than from
bottom filling. The emission factor decreases as the drum fill rate increases.
However, for the bottom filling of perc, emission factors for the two faster
fill rates exceeded that of the slowest fill rate. As indicated in Table
V-5a, the percent of vapor saturation values for top and bottom fill with perc
agrees favorably with values presented by EPA (1985) for cargo tanks. How-
ever, the values for larger vessels presented by EPA (1985) and Schrag (1985)
are substantially smaller.
The ethylene glycol emission factors are believed to have been in-
flated because of a higher volatility contaminant in the liquid. This un-
identified contaminant is believed to have been added to the original two
55-gal drums of ethylene glycol through contaminated transfer hoses. Maximum
concentrations in the range of 6,000 to 8,000 ppm v/v were observed even
though the saturation vapor concentration of ethylene glycol is known to be
less than 135 ppm at 25°C. As can be seen in Table V-5, Tests 4 and 5 of
ethylene glycol produced raw emission factors of 2.04 g and 2.15 g, respec-
tively, while tests 24 and 25 under similar conditions produced emission fac-
tors of 1.06 g and 1.14 g, respectively. This reflected the loss of volatile
contaminant as the ethylene glycol was repeatedly handled during the testing.
59
-------
30000
a
a
20000
-------
Table V-4. Rates of Attainment of Maximum Drum Exit Concentration
Percentages of
maximum observed concentrations
at given percents of total fill time3
25% 50% 75%
A. Top fill at 25 gal/min
Perchloroethylene 91.4 (3.6) 96.7 (1.7) 95.4 (4.1)
Ethylene glycol 33.4 (13.1) 57.0 (16.3) 83.7 (12.6)
Both 56.6 (30.2) 71.9 (23.2) 88.1 (11.7)
B. Bottom fill at 25 gal/min
Perchloroethylene 16.5 (10.8) 46.9 (10.4) 80.6 (7.5)
Ethylene glycol 52.8 (6.3) 65.6 (9.4) 83.6 (9.8)
Both 34.6 (20.2) 56.2 (13.6) 82.1 (8.8)
C. Top fill at 10 gal/min
Perchloroethylene 91.9 (2.6) 96.6 (1.0) 97.9 (0.6)
Ethylene glycol 48.5 (1.5) 73.7 (2.5) 90.0 (2.8)
Both 67.1 (21.6) 83.5 (11.5) 93.4 (4.5)
D. Bottom fill at 10 gal/min
Perchloroethylene 0.8 (0.3) 13.5 (11.3) 57.4 (15.4)
Ethylene glycol 42.8 (22.0) 49.5 (9.2) 67.5 (7.0)
Both 27.0 (26.8) 36.0 (20.1) 63.7 (12.0)
E. Top fill at 17 gal/min
Perchloroethylene 96.2 (0.9) 98.9 (0.1) 99.3 (0.2)
F. Bottom fill at 17 gal/min
Perchloroethylene 16.0 (1.0) 44.5 (6.9) 78.9 (3.2)
25%, 50%, 75% values are from the nearest 5 s mean concentration value.
Values in parentheses are standard deviations.
61
-------
Table V-5. Drumming Emission Factors1
Chemical name
Perchl oroethyl ene
Ethyl ene glycol
Filling Nominal
mode' "fill
rate (gpm)
Top 11
17
25
Bottom 11
17
25
Top 11
25
Bottom 11
25
No. of
tests
3
2
6
3
2
4
2
2
.2
2b (4)
2
3
2
2
Mean emission
rate (mg/s)
180
238
337
50
121
164
3.4
3.4
16.0.
17. 4b
0.66
0.77
3.0
3.7
Mean emission
factor
(g/drum fill)
53.3
45.6
43.6
13.9
23.5
20.6
1.01
1.11
1.88.
2.10b
0.22
0.24
0.36
0.46
Corrected for
contaminants
(g/drum fill)
0.95
0.96
1.28
1.10
0.18
0.20
0.28
0.34
Vapor
and
mist
X
X
X
X
X
X
X
X
X
X
Vapor
X
X
X
X
a
Dry drums only.
Tests 4 and 5 only; tests 24 and 25 show significantly less emissions.
See text.
-------
Table V-5a. Comparison of Saturation Factors
Operation
Present
study
(drums)
Section 4.4 of
AP-42
Cargo. Marine
Tanks0 Vessel sc
Shroy (1985)
Barges
Top (Splash)
Fill 1.20 1.45
Bottom (Submerged)
Fill 0.46 0.50 0.2 - 0.5 0.06 - 0.23
? Geometric means for perchlorethylene tests (dry drums only).
Suggested values for loading petroleum liquids into clean cargo tanks.
c Suggested values for loading petroleum liquids other than gasoline and
crude oil.
Filling of clean barges with acetone and benzene. Values in table represent
maximum saturation factor reported for fill operation.
A separate column in Table V-5 presents ethylene glycol emission
factors adjusted for the loss of contaminant. The mean emission factor for
tests 4 and 5, for example, was reduced to 1.10 g, the same value as the mean
emission factor for tests 24 and 25. Intermediate test values were also re-
duced based on the assumption that top filling produced four times the con-
taminant loss of bottom filling. With this adjustment, test-to-test compari-
sons can be made.
Top filling emissions for ethylene glycol are shown to be from 3 to
5 times the emissions from bottom filling, a slightly higher ratio than for
the perc tests. In three of the four sets of tests with a 10 mm nylon cyclone
precollector (50% cutpoint of 3.5 urn AED), vaporous emission factors were
slightly lower than comparable factors for vapor plus mist.
B. Dispersion Tests
1. Experimental Apparatus
For the tests of plume dispersion of emissions from the drum, a
steady state emission rate was created by bubbling air through a half-filled
drum. The bottom fill lance (nozzle) from the emission tests, with wire
screen covering the discharge end, was used to introduce air into the liquid.
A Bell and Gossett 1/2 hp vacuum pump forced air through a 1 gal surge tank
followed by an American Charcoal V-l test meter and into the lance. The bub-
bling rate of 3.0 cfm corresponded to a drum fill time of 2 min and 23 sec,
which was close to the fast fill rate used in the emission tests. The airflow
rate was determined from successive dry gas meter readings separated by a
fixed time interval.
63
-------
The wind tunnel flow rate was adjusted to nominal values of 100 fpm
or 200 fpm, by opening windows to reduce the vacuum in the building which
housed the tunnel. A Hastings warm wire anemometer was used to measure air
velocity profiles during the dispersion tests. Before or after each disper-
sion test, the anemometer was used to check the centerline airflow at 1, 2,
3, 5, 6, 7 and 8 m from the tunnel inlet.
The drum center was located at a distance of 4 m from the tunnel
inlet. Two bung positions were used in the testing—upwind (facing the tun-
nel inlet) and side (facing the tunnel wall).
A rectangular array of 72 sampling points with a 0.25 m spacing was
used to determine the analyte concentration patterns at distances of 1 and
3 m downwind from the center of the drum. The inlet to the air sampling line
was held in succession at each grid point for 5 to 18 readings of 5-s concen-
trations following a 15-s equilibration period. Background concentration was
determined by sampling at an upwind position I m above the floor and 2 m down-
stream from the tunnel inlet.
The temperature of the liquid was measured at least once during
each dispersion test. For the tests of perchloroethylene, the latent heat of
the inlet air (heated by the pump) tended to balance the cooling effect of
evaporation so that the liquid temperature remained closed to the air tem-
perature in the tunnel. However, in the case of methanol, the substantially
greater heat absorbed by evaporation caused the liquid to cool several degrees
centigrade below the ambient temperature.
2. Test Results
A summary of the 18 dispersion tests using perc and methanol is
shown in Table V-6. The propose of these tests was to characterize the emis-
sion plume shape as a function of test chemical properties, tunnel flow rate,
and bung position.
An example array of concentrations observed at a downwind distance
of 3 m during the side bung release of methanol emissions is shown in Fig-
ure V-4. Corresponding wind speeds are given in Figure V-5. For this test
(Test 14) with a mean 199 fpm tunnel centerline flow rate, concentration ranged
up to 500 ppm with an observed background level of 6 ppm. The center of the
plume was near the tunnel centerline about 0.75 m above the floor. A slight
drift to the right side (looking downwind) was observed only at the 0.75 m
height; otherwise the horizontal concentration pattern was symmetrically dis-
tributed reflecting the dominant effect of the drum wake on plume dispersion.
Both perc and, to a lesser extent, methanol emissions were found to
sink to the tunnel floor, as illustrated in Figures V-6 and V-7 which compare
vertical profiles of methanol concentration from side releases at nominal
tunnel flows of 200 fpm and 100 fpm. The cross-sections of the plume for a
100-fpm flow are shown in Figures V-8 and V-9, for downwind distances of 1 m
and 3 m, respectively.
64
-------
Chemical: Nctlianol
, Tunnel Ceiling
-
:>„
;.>
^x}
x£
S
^6
Left ^
Tunnel -'^4
Wall >£
Xx3
-:2
.' i
:-;
f
1 meter ,
7
15 7 9
18 70 27
14 149 391 127
9 497 302
11 16 46 225 424 286
10 39 42 260 423 260
13 38 108 217 299 244
215
32 73 210 250 251
1 i i 1 i 1 1
p 1-2 3 4 ; ; 5- --- 6 ;
^ • ^ s~ s _' UUIIK ruaEI &UII I «BftUU
Sampling Plane: 3 met
Test Date: 8/29/85
Concentration (ppm)
.
... |
18 ^ R'9ht
• • • x-^x Tunnel
/> Wall
44 ^x
177 57 14 ^>
164 84 -x^-
196 56 46 '.-'^
s'_
140 27 T-''-'
L i A
> 7 -'--8 9 - .
Tunnel Floor
Figure V-4. Profile of methanol concentrations at 3 m
sampling plane (200 fpm).
-------
cn
O1
- 2 ' . ' ' 3
4 ;>j>x^ --6/-'>7
funnel Floor
Chemicali Hcthanul
Bung Position: Side
Sampling Plane: 3 meters
Test Date: 8/29/85
Wind speed (m/s)
Figure M-5. Profile of wind speeds at 3 m sampling
plane for methanol test at 200 fpm.
-------
o
UJ
3m Downwind
0.4 0.6
C (h)/Cmax
(along tunnel
Figure V-6. Comparison of vertical profiles of methanol concentrations from a drum
side release with wind speed of 200 fpm.
67
-------
CD
ui
1m Downwind
3m Downwind
0.2
0.4 0.6
C(h)/Cmax
(along tunnel
0.8
Figure V-7. Comparison of vertical profiles of methanol concentrations from a drum
side release with wind speed of 100 fpm.
68
-------
Left
Tunnel -^ 4
Wall
Worker Exposures
•
330 138 14
Right
Tunnel
Wall
Tunnel Floor
Figure V-8. Cross-sectional view of methanol concentrations (ppm) at 1m downwind for a drum side
release at 100 fpm, compared to downwind facing worker exposure (ppm).
-------
Worker Exposure
• •
168 73
//Right
Tunnel
Wall
.< "4'
-------
Table V-6. Dispersion Test Log
Date
8/20/85
8/21/85
8/22/85
8/23/85
8/28/85
8/29/85
8/30/85
Test
no.
1
2
3a
4a
5
6
7b
8b
9
10
11
12
13
14
15
16
17
18
Chemical
name
Perc
Perc
Perc
Perc
Methanol
Methanol
Methanol
Mean
fluid
temp.
(°C)
20
21
22
24
25
23
20
23
22
21
25
19
24
21
27
23
23
20
Nominal Tunnel
centerl ine
flow (fpm)
flow (fpm)
200
100
100
200
100
200
200
100
Bung
position
Upwi nd
Upwi nd
Upwi nd
Side
Side
Upwi nd
Side
Side
Upwi nd
Downwi nd
sampling
distance
(m)d
1
2
3
1
1
1
3
1
1
3
1
3
1
3
1
3
1
3
.Test aborted and pump moved out of tunnel.
Exit airflow from drum of 2 ftVmin; on all
cairflow was 3 ftVmin.
As measured downwind from the drum center.
other dispersion tests, exit
This plume sinking behavior is compatible with the high molecular
weight of perc (166) but not with the molecular weight of methanol (32).
However, the relatively high rate of vaporization of methanol resulted in the
cooling of the emission stream released from the drum, which produced a
slightly negative plume buoyancy. As evidence of the cooling, the lower half
of the drum exterior was observed to collect condensation as methanol testing
proceeded. This effect was not observed with either perc or ethylene glycol.
With about 2 gal of methanol vaporized during a typical evening of tests, the
total vaporation heat loss approximated 1,400 kilocalories. This loss was
somewhat offset by the latent heat of the air supply (29° to 35°C) to the
bottom of the test drum in relation to the tunnel air temperature. The lowest
methanol temperature observed during the dispersion and exposure experiments
was at the end of Test 12 where 17°C was measured. The corresponding tunnel
air temperature was 22°C.
71
-------
Concentration and wind speed values measured at the 3 m distance
were used to calculate mass flux values for comparison with the steady state
emission rates. These data are presented in Appendix D which discusses
quality control data.
Appendix E contains plotted concentrations of perc and methanol at
1 m and 3 m downwind of the drum center. In addition, for each concentration
distribution, there is an accompanying plot of array wind speeds.
C. Exposure Tests
1. Experimental Apparatus
For the tests of worker exposure, a steady state emission rate was
created in exactly the same manner as for the dispersion tests described above.
Moreover, the drum was located at the same position, and the same two bung
positions were used.
Unlike the dispersion tests, concentrations for the exposure tests
were measured only at a breathing height of 1.57 m for a standard worker
height of 1.73 m. Two arrays of sampling points (worker positions) were uti-
lized, one corresponding to the upwind bung position and the other to the side
bung position. These are illustrated in Figures V-10 and V-ll.
The two bung positions, upwind and side, simulated two airflow di-
rection scenarios. The upwind bung position results when the wind flow is
directed toward the worker's back at the drumming station, and the side bung
position results when the wind flow is across the drumming operation from the
s i de.
2. Test Results
The 14 exposure tests are summarized in Table V-7. Exposure tests
differed from dispersion tests in that the sampling line inlet was placed in
the breathing zone of a "worker." In Figure V-12, the worker positioned him-
self near the drum with the wind flow directed to his back (upwind bung posi-
tion). In this photo, the air hose used to supply fresh air to the worker
during periods of testing is not shown.
The relatively constant emission rate from the drum enabled steady
state exposure tests to be efficiently done. However, it is necessary to re-
late these results to the nonsteady state emissions tests described in Sec-
tion A. A summary of emission rates used both in tests of dispersion and
exposure is given in Table V-8.
The emissions from bubbling air through perc at 25°C produced con-
centrations of about 35,000 ppm at the bung hole, which corresponds closely
to the maximum concentrations observed in top filling at the fast fill rate.
Since top filling produced fairly uniform concentrations (~ 30,000 to 37,000
ppm) over the latter 75% of the drum fill cycle, the concentrations measured
in the perc exposure tests represent well the concentrations expected to be
inhaled by a worker. For the remaining 25% of the top filling cycle, for
bottom filling, and for ethylene glycol tests, it is necessary to apportion
observed exposure concentration values by referring to emission rates in
Tables V-3 and V-5.
72
-------
Wind
Flow
50 Centimeters
Figure V-10. Sampling locations for upwind bung position.
73
-------
oo
Wind
Flow
e • •
-l 1-
1 Meter
Figure V-ll. Sampling locations for side bung position.
74
-------
Figure V-12. Example exposure testing.
75
-------
Table V-7. Exposure Test Log
Date
8/21/85
8/23/85
8/29/85
9/2/85
9/13/85
Date
8/13/85
9/2/85
9/13/85
Test
no.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Table V-8.
Time
1503
1813
1820
2336
1830
1834
Chemical name
Perc
Perc
Methanol
Methanol
Ethylene glycol
Drum Emissions for
Chemical
Perc
Methanol
Methanol
Methanol
Ethylene glycol .
Ethylene glycol
Fluid
temp (°C)
25
24
23
21
26
24
20
28
23
21
21
21
21
21
Tests of Di
Fluid
temp.
(°C)
25C
28
28
20
21
21
Nominal tunnel
centerline
Bung
flow (fpm) positions
100
200
200
100
200
200
200
100
100
200
200
200
100
100
spersion and Exposure
Mean 5-s
concentration (ppm)
Raw Corrected
34534 34819
126636 120937
117686 120628
100958 101968
1381 1519
13626 1505
Upwind
Upwi nd
Side
Side
Side
Upwi nd
Upwi nd
Upwi nd
Side
Upwi nd
Upwi nd
Side
Side
Upwi nd
Emission
rate0
(mg/s)
326
216
216
187
5.40
5.35
.Based on checks of the Beckman THC analyzer with bag standards.
Based on 3.0 ft3/min air supply to drum bottom.
.Estimated value.
Believed to be contaminated with volatiles.
ewith cyclone on sampler inlet.
76
-------
Special studies were conducted within the 14 exposure tests to mea-
sure the effect on exposure of: (a) worker orientation to wind flow; (b)
worker movement; and (c) breathing height. Table V-9 presents mean 5 s chem-
ical concentrations observed in studies conducted in an upwind bung situation
with the "worker" always facing downwind. By moving only slightly while fill-
ing a drum, the worker can lower his exposure by a factor of from 3 to 45 as
illustrated in Table V-10 for methanol. This effect results from the mixing
of the contaminant stream with clean background air.
As indicated in Figures V-13 and V-14, worker height between 1.4
m (5 ft 2 in.) and 1.5 m (5 ft 8 in.) appears to have little effect on perc
exposure near the drum, but the exposure is sensitive to wind speed. At dis-
tance of about 0.5 m from the bung, the effect of breathing height is more
pronounced. The effect of worker distance and wind speed on methanol expo-
sure is shown in Figure V-15.
The major conclusion from the studies of breathing zone concentra-
tions is that worker orientation to the wind flow is a significant parameter
in predicting worker exposure. Much higher analyte concentrations were ob-
served in the breathing zone when the worker faced downwind than when the
worker faced upwind or faced crosswind. This unexpected test result is dis-
cussed further in the next section on model development.
For purposes of exposure model development to be discussed in
Section VI, isopleths of normalized breathing zone concentrations for worker
positions near the drum during filling operations have been prepared. A
total of eight figures (Figures V-16 through V-23) are provided, correspond-
ing to two simulants (perchloroethylene and methanol), two airspeeds (100
and 200 fpm), and two worker orientations (facing downwind and facing either
upwind or crosswind).
The units for the normalized isopleth values are seconds per cubic
meter (s/m3). The value 0.1 s/m3 is used to represent the edge of the plume.
If these values are multiplied by the emission rate, then the isopleths rep-
resent lines of constant concentration. For example, if the emission rate is
5 mg/s, then the curves shown in Figure V-16 correspond to 0.5, 5, and 50 mg/m3.
All airflows are from the top of the page to the bottom. It is
recommended that the 100-fpm curves be used for most indoor operations.
As a rule of thumb, it is suggested that the methanol curves be
used for drumming of liquids with molecular weights of less than 100 and that
the perchloroethylene results be used for larger molecular weights. This
recommendation is based on the observed settling of denser vapors away from
the breathing zone.
Finally, it is important to note that worker motion may often be as
important as orientation in determining exposure. The results presented in
these graphs are based on motionless workers and thus represent highest poten-
tial exposure. As noted in the previous section, slight motion of the worker
may reduce observed concentrations by as much as an order of magnitude.
77
-------
Table V-9. Special Exposure Studies
00
Date
8/21/85
8/29/85
8/29/85
9/2/85
Time
2055
2120
2126
2135
2316
2318
2319
2039
2040
Tests were of upwi
20 cm directly upwi
Chemical
Perc
Perc
Perc
Perc
Methanol
Methanol
Methanol
Methanol
Methanol
Temp.
24
24
23
23
20
20
20
~ 24
~ 24
nd bung emissions at
nd from bung opening.
Mean 5-s Situation
concentration (ppm)
Raw
695
683
720
128
5209
382
1859
4147
90
mean tunnel
Corrected
806
792
835
148
5339
392
1905
4396
95
centerl ine
Facing downwind and standing @ 0, 20a
Facing downwind and standing @ 0, 20a
Facing downwind and sitting @ 0, 20a
Facing downwind with nostrils at
drum level @ 0, 20
Facing downwind @ 0, 30 and stationary
Moving between -40, 30, and +40, 30
with large steps while facing downwind
Moving between -40, 30, and +40, 30
with small steps while facing downwind
Facing downwind and standing @ 0, 30C
Moving between -40, 30C and +40, 30
flow of 209 fpm. Position 0, 20 refers to
Tests were of upwind bung emissions at mean tunnel centerl ine flow of 188 fpm.
£o 30 cm directly upwind of drum and ±40 cm to side.
Tests were of upwind bung emissions at mean tunnel centerl ine flow of 104 fpm.
to 30 cm directly upwind of drum and -40 cm to side.
Position ±40, 30 refer
Position -40, 30 refers
-------
Table V-10. Effect of Worker Movement on Exposure
(30 cm upwind, methanol)
Worker movement
(side-to-side)
Stationary
Slow
Exposure
Wind speed: 100 fpm
7,980
-
(ppm)
200 fpm
5,410
1,740
Rapid
Continuous movement 90
Large steps - 380
79
-------
2
•1.5m—•
• 1.4m— •
2
Breathing
Height
Wind Flow
200 fpm
22 330 480 690
15 75 90 720
2 2 2 2 140
•1.0m—• • • • •
Release
Point
55 gal Drum
Perchlorethylene
100 80 60 40 20 0
Dlstahce.UpwInd from Source .(cm)
Figure V-13. Effect of breathing height and upwind
distance-on exposure to perc
concentrations (ppm).
80
-------
T800
100
80 60 40 20
Distance Upwind from Drum (cm)
Figure V-14. Effect of breathing height, wind speed and
upwind distance on exposures to perc
drumming emissions.
81
-------
-I 8000
60, 30
Distance Upwind from Drum (cm)
Figure-V-15.
Effect-of wind speed and upwind
-------
Figure V-16.
K contours for methanol (M<100), 100 fpm air velocity and downwind
worker orientation.
83
-------
Figure-V-l-7. K-contours-for methanol (M<100), 200 fpm air velocity and downwind
worker orientation.
84
-------
A* r-
F/o*/
Figure V-18. K contours for methanol (M<100), 100 fpm air velocity and non-downwind
worker orientation.
85
-------
'fir
Figure V-19.
K contours for methanol (M<100), 200 fpm air velocity and non-downwind
worker orientation.
86
-------
Figure V-20. K contours for perchlorethylene (Mi 100), 100 fpm air velocity and
downwind worker orientation.
87
-------
A
s n €.
2.OO fpm Air
Air
f?/o*/
Figure V-21. K-'contours for- perch!orethylene (M£100), 200 fpm-air velocity and
downwind worker orientation.
88
-------
/ /<
-------
I //7J
/7
Ai
ir
/4/y
/e/c»«v
Figure \l-23. K contours for perch!orethylene (Mi 100), 200 fpm air velocity and
non'-downwind- worker orientation.
90
-------
VI. MODEL DEVELOPMENT
A. Modeling Alternatives
Several models have been proposed for the purpose of assessing
worker exposure to bagging and drumming operations. These have ranged from
models which are very simple to apply to those which are not only complicated
but also restricted to the actual ventilation pattern considered. Most models
currently used to estimate workplace concentrations are based on the concept
of mass balance (Figure VI-1) possibly with additional features such as in-
complete mixing or time-dependent release rates.
For the drumming of liquids at room temperatures, the following
equation has been suggested for use (Office of Toxic Substances 1984):
r _ 0.6 f V P° r
Cv T Ec»- I
where: GV = concentration, ppm
f = saturation factor for loading, dimensionless
k = mixing factor, dimensionless
P° = vapor pressure, atm.
Q = ventilation rate, ftVmin
r = loading rate, drums/hr
V = volume of container, cm3
The values used in the above formula are taken from Table VI-1,
unless other values are known. The values in the table are representative of
worst case and typical case. Equation 1 assumes ideal gas behavior and the
absence of chemical reaction.
The worst case represents a situation of general ventilation at low
rates in the workplace. The typical case takes into account higher ventila-
tion rates and more complete mixing. The worst case is assumed when no infor-
mation is known about the workplace controls (Office of Toxic Substances 1984).
However, the primary limitations of any mass balance model result
from the neglect of spatial variation of concentration and the influence of
worker position, orientation, and motion on exposure. It is important to note
that even though other models take spatial variation into account, the fact
that a worker presents a substantial obstruction to the generally low venti-
lation flows requires that worker location and orientation be considered in
91
-------
BOX MODEL
Room Volume
V
Volumetric
Air Flow
Mass Emission
Rate
Source
^r = G-CQ
at.
Figure VI-1. The box (uniform mixing) model
92
-------
Table VI-1. Default Values for Equation
Term
V
f
r
k
QC
Drums
2 x 105 cc (55 gal)
Worst
case
1.0
30/h
0.1
500 cfm
Typical
case
1.0
20/h
0.5
3000 cfm
Tank
1.9 x 107
Worst
case
1.0
2/h
0.1
3000 cfm
trucks
cc (5000 gal)
Typical
case
1.0
2/h
0.5
d
Tank
7.6 x 107 cc
Worst
case
1.0
1/h
0.1
3000 cfm
cars
(20,000 gal)
Typical
case
1.0
1/h
0.5
d
Adopted from "A Manual for the Preparation of Engineering Assessments." a
manual prepared by the Chemical Engineering Branch, Economics and Technology
Division, Office of Toxic Substances, EPA, Washington, D.C. 20460. August
b!984.
Cans and bottles are assumed to be similar to drums except for the volume.
cUse 19,000 cm3 (5 gal) for volume of a can.
If Q is known for a particular scenario, substitute it into the formulae
.above.
Typical operations take place outdoors. Berman (1982) suggests Q = 23,400 V,
where V is wind speed in mph.
assessing inhalation exposure. Because it is almost impossible to standardize
worker motions, the model presented below yields exposure estimates correspond-
ing to an absolutely motionless worker. As noted earlier, slight motions may
reduce breathing height concentrations substantially.
B. Modeling Approach
The modeling approach adopted in this study attempts to incorporate
the relative simplicity of the mass balance models together with the features
of spatial variation and worker position necessary to provide more reliable
estimates of exposure. Results of the exposure tests described in Section V.C
have been combined in the K contours given in that section. These isopleths
(in units of s/m3) correspond to breathing zone concentrations resulting from
a unit emission rate of 1 mg/s. Thus, when the values are multiplied by the
emission rate (in mg/s) for a particular application, the isopleths become
lines of constant concentration (in mg/m3) for that application.
93
-------
The basic steps in the modeling procedure are as follows:
1. Estimate mass emission rate, G (mass/time), by
No. of drums x loss per drum
Averaging time period
2. Estimate ventilation rate, Q (air volume/time).
3. Apply the uniform mixing model (Figure VI-1) to obtain the
steady state concentration, C = G/Q. This value represents
the background concentration-to which is added the plume con-
tribution due to filling.
4. Select the appropriate set of K contours, using the criteria
below:
Molecular weight, M - Use methanol contours if M is less than
100 and perchloroethylene results otherwise.
Airflow rate - The use of the 100-fpm. contours is recommended
for most indoor applications. The 200-fpm results may be
used for outdoor operations or-in cases of indoor filling,
with ambient winds aiding ventilation.(as might be the
case during the summer months).
5. Conduct a time-motion study of the worker noting time spent in
each location (as well as orientation to drum) during fill cycle.
Average over several cycles.
6. Using the appropriate K contours, develop a weighted K factor,
K, by
n
i-i *i K'
K= Ji-^-
i-i' V
where: t.. = time spent in i-th location
K.J = K factor for specific orientation in i-th
location
94
-------
7. Determine the plume contribution resulting from drumming or
bagging by multiplying K (in s/m3) by G (in mg/s). Call this
result C ..
d
8. Add C to C . to obtain the total estimate of breathing zone
concentration C*.
9. Multiple C* by 1.25 m3/h to estimate inhalation exposure.
Because PMN assessments are often made before any drumming line has
been designed, default values for K have been provided for typical filling
operations in Table VI-2. The modeling procedure using these default values
follows the same steps as above, except that steps 4 through 6 are no longer
needed.
Table VI-2. Default K Values
Filling operation3
Automated
Manual
Airflow
Side
Rear
Side
Rear
K factors (s/m3)c
M < 100d M £ 100d
0.2 0.1
0.6 0.6
0.3 0.1
3 3
a.,, . .
• All indoors.
lt«^IU\*IVW WW YV\J I I\C I .
As noted in the text, the resulting breathing zone estimates should be
considered as "worst case" in that no worker motion is considered. See
ddiscussion in text.
M represents molecular weight.
95
-------
The default values were developed using time-motion data gathered
during this program. These data are given below:
Percentage of time at station
Automated Manual
Station fill fin
Drum fill 5 25
Drum close 25 30
Drum setup/label 65 40
Five percent of total time spent away from drum
line.
To derive the default values, typical work station locations for
the two types of fills were placed on the K contour plots given earlier. In
addition, two fill line orientations were considered—airflow parallel to the
line (i.e., worker not facing downwind) and airflow perpendicular to the line
(i.e., flow from behind the worker). The default values for R were then ob-
tained for the four cases using the procedure presented earlier as steps 5
through 7.
As an example of the use of the default R values, consider a hypo-
thetical drumming line which fills 290 drums with a liquid of molecular weiqht
of 200 over an 8-h period.
1. If an emission factor of 0.3 g/drum (mean value for ethyl-
ene glycol) is used, then the average emission rate over the 8 h is
given by
r _ 290 drums n 0 , . _ ,
G -- g-^ - x 0.3 g/drum = 3 mg/s
Step 2. Current ventilation plans call for Q = 12,000 cfm (5.7 m3/s)
Step 3. Using the uniform mixing model,
Css • £r%k - "I MWV
Steps 4 to 6. The use of an automated fill line with airflow from
behind the worker is planned. The appropriate K value (from
Table VI-2) is 0.6 s/m3.
Step 7. The additional contribution due to drumming is
Cd = KG = 1.8 mg/m3
8. The estimate of total breathing zone concentration is
C* = Css + Cd = 2.3 mg/m3
96
-------
C. Model Validation
Validation of the proposed modeling technique requires a complete
set of independent data obtained from a plant-scale bagging or drumming oper-
ation. Because no such set is found in the literature, the plant-scale test-
ing reported earlier in this document was done in order to provide data for
model validation. Both bagging and drumming operations were tested as speci-
fied by the scope of work, with the idea that emission factors for these types
of operations could be developed independently. Since the literature search
yielded few (if any) appropriate emission factors for these operations, it be-
came necessary to develop a limited set of experimental values for this study.
Because of the need to limit the cost of this phase of the experimental pro-
gram, it was decided to develop emission factors for the drumming operation
only.
1. Plant A Estimates
Data from the drumming operation at plant A are suitable for a
limited validation of the modeling approach. Note that data from run A-l are
not suitable because all sample masses were below the 2 ug limit of detection.
No reliable estimate for the vapor pressure for fluid 2 could be
provided. As a result, it was not possible to use the vapor pressure to es-
timate an emission level for run A-2. Because of the low volatility of the
fluid, the mean emission factor measured for ethylene glycol (0.3 g/drum) was
used in the validation process.
Using this emission factor, the emission rate during test A-2 was
found as
G = 59 drums x 0.3 g/drum -=• 98 min = 3 mg/s
As stated in Section IV-A, the worker_spent only 3% of the time in front of
the fill station. Thus, the weighted K value is
K = 0.97 x (0 s/m3) + 0.03 x (1 s/m3) = 0.03 s/m3
Note that a multiplier of 1 s/m3 (rather than 10 s/m3) has been used to ac-
count for the fact that only about 3 s per cycle was spent at the fill loca-
tion, and thus no strong buildup of vapors was possible. Also, because this
discussion is meant to partially validate the model using plant data, a value
of 0 s/m3 is used rather than 0.1 s/m3 for locations other than the fill
station.
The plume contribution resulting from drumming is then given by
Cd = KG = 0.03 s/m3 x 3 mg/s = 90 ug/m3
As noted in the text, an operator exposure value of 240 ug/m3 was measured by
a personnel sampler. Because samplers 15a through 17 in Table IV-5 can be
considered as measuring upwind concentrations, it is seen that the background
level is about 110 to 180 ug/m3. Thus, the estimated drumming contribution
of 90 ug/m3 compares very favorably with the measured value of 60 to 130 ug/m3.
97
-------
The background estimate for test A-2 using the modeling procedure
would be
Css = G/Q
= 3 mg/s -r 11,750 cfm
= 540 ug/m3
which is approximately 3 to 5 times higher than observed. Note that results
from samplers 3 through 7 suggest some stratification of concentration with
height, presumably because of the relatively high density of the hydraulic
fluid vapors. As a result, use of the uniform mixing model to estimate back-
ground levels may be conservatively high when applied to the drumming of
heavier liquids.
2. Plant B Estimates
A previously determined emission factor of 0.095 kg/Mg (0.19 Ib/ton)
for bagging of urea (USEPA, 1981) was used to estimate the uncontrolled emis-
sion rate at plant B. Furthermore, a collection efficiency of 80% was as-
signed to the local hood, resulting in an uncontrolled emission factor of
0.019 kg/Mg (0.03 Ib/ton). Finally, based on particle size ratios measured
by both the cascade impactor and cassette samplers, 14% of the controlled
emissions was assumed to consist of respirable particulate (^ 3.5 urn in aero-
dynamic diameter).
The following estimated emission rates, G, are found using the data
of Table-IV-7:
Controlled
emission rate (mg/s)
Run
B-l
B-2
B-3
Suspended
particulate
20
18
30
Respirable
particulate
2.8
2.6
4.2
Time and motion studies conducted at the plant indicate that roughly
60% of the operator's time is spent at the autopacker. Thus, the weighted K
value is
K = 0.4 x (0 s/m3) + 0.6 x (1 s/m3) = 0.6 s/m3
The multipliers (0 and 1 s/m3) are the same as those used in the drumming
estimates.
98
-------
The predicted plume contribution arising from bagging is given by
KG. For the three runs,
Run
B-l
B-2
B-3
The background concentration
model are given by G/Q, where
Predicted plume
contribution (ug/m3)
Suspended Respirable
particulate particulate
12,000 1,700
11,000 1,600
18,000 2,500
estimates calculated from the uniform mixing
Q = 22,900 cfm:
Predicted background (ug/m3)
Measured background
from Samplers 6, 7a
Run
B-l
B-2
B-3
Suspended Respirable
particulate particulate
1,800 260
1,700 230
2,800 390
concentrations were found by averaging the concentrations
, 7b, and 8:
Run
B-l
B-2
B-3
Measured background Cug/m3!
Suspended Respirable
particulate particulate
1,200 110
1,600 110
3,000 450
The background values predicted by the uniform mixing model agree well with
the measured values for the two particle size fractions.
99
-------
To assess the model's performance in predicting the plume contribu-
tion to worker exposure, only respirable particulate concentrations are avail-
able. Respirable particulate concentrations of 1,890 and 1,200 ug/m3 were
measured with the operator personal samplers for Runs B-l and B-2, respec-
tively. (No value was reported for Test B-3.)
The plume contributions of 1,700 and 1,600 ug/m3 predicted by the
modeling technique compare well with the measured contributions of 1,800 and
1,100 ug/m3, after subtracting measured background concentrations.
100
-------
REFERENCES
Berman DW. 1982. Methods for estimating workplace exposure to PMN sub-
stances. Clement Associates, Inc., Arlington, VA. Prepared for U.S.
Environmental Protection Agency, Economics and Technology Division,
Washington, DC.
Bomberger DC, Boughton R, Endlich R, Fowler D, Gikis B, Ludwig F, Wilhelm D,
Witham C. 1983. Industrial process profiles to support PMN review: filling
of drums and bags, draft. SRI International, Menlo Park, CA. Prepared for
U.S. Environmental Protection Agency, Office of Toxic Substances, Washington,
DC. Contract No. 68-01-6016, Technical Directive 69.
Clement Associates. 1981. Mathematical models for estimating workplace
concentration levels: a literature review. Clement Associates, Inc.,
Washington, DC. Prepared for U.S. Environmental Protection Agency, Economics
and Technology Division, Washington, DC.
Davis RJ. 1971. A simple model for the estimation of aerosol concentration
in a closed vessel. Am Ind Hyg Assoc J 32:603-609.
USEPA. 1984. A manual for the preparation of engineering assessments. Pre-
pared by the Chemical Engineering Branch, Economics and Technology Division,
Office of Toxic Substances, Washington, DC.
Schroy JM. 1981. Prediction of workplace contaminant levels, pp. 190-206.
Symposium Proceedings on Control Technology in the Plastics and Resins Indus-
try, Atlanta, GA, February 27-28. Cincinnati, OH: National Institute for
Occupational Health and Safety. NIOSH Publication NO. 81-107.
Schroy JM. 1985. Dynamic conditions in filling drums. Presented at the
1985 Summer National Meeting of the American Institute of Chemical Engineers
Seattle, WA.
USEPA. 1981. Urea manufacturing industry - technical document. EPA-450/
3-81-001. Office of Air Quality Planning and Standards, Research Triangle
Park, NC.
USEPA. 1985. Compilation of air pollutant emission factors (AP-42). Vol I,
Section 4.3. Research Triangle Park, NC.
101
-------
APPENDIX A
INDUSTRY SURVEY
TABLE OF CONTENTS
Page
I. Introduction A-l
II. Industry Survey A-l
A. Information Search Methodology A-l
B. Significant Documents Identified A-2
C. Industry Characterization Summary A-12
III. Arrangements for Plant Visits A-13
A. Industry Contacts A-17
B. Obstacles Encountered in Plant Visit
Arrangements A-17
IV. References A-19
V. Annotated Bibliography of References on File A-20
VI. Packaging Association Phone Survey Approach and Results . . A-39
-------
I. INTRODUCTION
Under the Toxic Substances Control Act of 1976 (TSCA), a manufac-
turer or processor must submit to the U.S. Environmental Protection Agency
(EPA) a premanufacturing notice (PMN) prior to commercial production. The
EPA has 90 days to review the PMN and to take any necessary regulatory actions
deemed necessary. During this review, the potential workplace exposure asso-
ciated with the manufacture and processing of the new chemical must be esti-
mated.
This project seeks to develop and validate predictive models which
may be used in the estimation of inhalation exposure of workers involved in
the drumming of liquids and the bag or drum filling of solids. The study de-
sign is comprised of four tasks, the first of which is an industry survey to
characterize drumming and bagging operations and to select the chemicals and
facilities for subsequent study.
This appendix provides results of the literature-based industry
survey and efforts to identify and gain access to plant sites suitable for
testing. Two related items are included here: an annotated bibliography of
the references on file (Section V) and information on a packaging association
survey (Section VI).
II. INDUSTRY SURVEY
The industry survey involved literature search and literature re-
view activities as well as personal contacts with agencies, companies, and
associations.
A. Information Search Methodology
Computerized literature searches of National Technical Information
Service (NTIS), Chemical Abstracts, and a few other data bases were conducted.
Hard copies of all promising references were obtained and evaluated. See
Section V for an annotated bibliography of all references on file.
A manual review of the Applied Science and Technology Index for 1981
and succeeding years was also conducted to check for relevant data in trade
publications. All promising references were obtained and evaluated, and those
found to be useful were placed on file.
Contacts with knowledgeable persons and groups were conducted by
phone and, in a few cases, in person. Individuals and groups contacted in-
cluded Thomas Cooper and others at NIOSH; Jon Volkwein at the Bureau of Mines;
Garrity Baker, Thomas Goldberg, and Doug Fratz at the Chemical Manufacturers
Association, the Synthetic Organic Chemical Manufacturers Association, and
the Chemical Specialties Manufacturers Association, respectively; and repre-
sentatives of 13 packaging associations. Section VI describes the packaging
association phone survey approach and results.
A-l
-------
B. Significant Documents Identified
Only a few publications which discuss or quantify emission levels
and worker exposure in drumming and bagging activities were located. These
publications were products of federally funded research for the EPA, NIOSH,
and the Bureau of Mines.
1. EPA Studies
Several useful EPA reports have been obtained and reviewed. These
are described below.
Bomberger et al. (1983) discuss the filling technology commonly em-
ployed in both the bagging and drumming of solids and the drumming of liquids.
Typical equipment and modes of operation are described in detail (i.e., bag
and drum designs, container closure methods, conveying and filling equipment
configurations, and emission control systems). However, these authors empha-
size that it is difficult to make general statements on industry-wide packag-
ing practices by chemical manufacturers, especially emission control practices.
According to Bomberger, equipment purchased for the packaging of
industrial chemicals represents a minor fraction of the packaging equipment
market. The bulk of this machinery is designed and sold for use with foods,
cosmetics, and Pharmaceuticals. Hence, the original design of such equipment
usually does not include toxic emission controls. The frequency and effec-
tiveness of on-site machinery modification to achieve dust or volatile emis-
sion control by chemical manufacturers are, therefore, dependent on in-plant
recognition of potential industrial hygiene problems and on the ability of
plant personnel to install and maintain emission control devices.
Bomberger also discusses estimated emission rates and the resulting
workplace air concentrations during bag and drum filling of chemicals. Bag
filling emissions have not been well characterized to date. Existing studies
quantify workplace concentrations or emission rates at individual locations
for a very limited group of materials, and only certain parameters (and not
of a proprietary nature) are reported. Consequently, it is difficult to
determine how representative the workplace concentrations and emission rates
which appear in the literature may be, relative to the wide range of packaging
environments which occur in the chemical manufacturing industry. Bomberger
concludes that there are insufficient data to expand or revise the estimates
in the previous EPA literature which assume levels of 2 mg/m3 around typical
semi-automatic filling stations without sophisticated dust control mechanisms
in place and 30 mg/m3 around very dirty operations—both of which also assume
that 30% of the airborne material is respirable (Berman 1983).
Bomberger's discussion of volatile emissions from the drumming of
liquids is based largely on data from the petroleum refining industry because
no other data were available. Table A-l provides estimated emission rates for
drum filling of a light liquid with a relatively high vapor pressure (benzene)
and a heavy liquid with a relatively low vapor pressure (cyclohexanol) under
A-2
-------
Table A-l. Estimated Drumming Emissions From Light and Heavy Liquids'
Material
Emission rate to workplace environment (q/h)
High1
Moderate
Low"
Light liquid (benzene)
drum vent
fitting6 f
chronic wet spot
Total
2500 (85%)
390 (13%)
34 ( 2%)
2924
125 (41%)
144 (47%)
34 (12%)
303
42 (19%)
144 (65%)
34 (16%)
220
Heavy liquid (cyclohexanol)
drum vent
fittings
chronic wet spot^
Total
33 (35%)
60 (64%)
1 ( 1%)
94
2 ( 4%)
51 (94%)
1 ( 2%)
54
1 ( 2%)
51 (96%)
1 ( 2%)
53
b$ource: Bomberger et al. 1983.
Top filling with no collection of vapors. No routine maintenance of
fittings, which are not inside a hood. 200-cm2 wet spot, which is not
inside a hood.
Top filling with 95% collection of vapors. Routine maintenance of
fittings, which are not inside a hood. 200-cm2 wet spot, which is not
.inside a hood.
Bottom filling of drum with 95% collection of vapors. Routine main-
tenance of fittings, which are not inside a hood. 200 cm2 wet spot,
which is not inside a hood.
fThree valves, one pump, two rotating or swivel fittings.
A mass transport coefficient of k = 0.12 representing quiescent
evaporation was used.
9A mass transport coefficient of k = 0.10 representing quiescent
evaporation was used.
A-3
-------
three sets of filling conditions: a high rate of emissions (top filling with
no collection of vapors and no routine maintenance of fittings which are not
inside a hood), a moderate rate of emissions (top filling with 95% collection
of vapors), and a low rate of emissions (bottom filling with 95% collection
of vapors). The range of estimated emission rates was 220 to 2,924 g/h for
the light liquid and 53 to 94 g/h for the heavy liquid.
Although Bomberger discusses methods for modeling workplace concen-
trations, this topic is covered more thoroughly by Clement Associates (1981).
The models available for modeling contaminant concentrations in an industrial
environment may be grouped into three types based on their theoretical basis:
mass balance, Gaussian diffusion, and aerosol mechanics.
Mass balance models are based on mathematical accounting schemes
whereby all fractions of the material passing through the process under study
are quantified, and the total outputs are always equal to the total inputs.
These approaches are most applicable for gaseous emissions because vapors fol-
low air currents freely and are not influenced by gravity. Particulate emis-
sions consisting of fine particles—smaller than 10 to 20 urn in diameter ac-
cording to Clement Associates (1981) and generally smaller than 5 urn in diam-
eter according to Bomberger et al. (1983)--may also be modeled effectively
using mass balance equations.
Gaussian dispersion models are not commonly employed in an indoor
setting. However, Clement Associates (1981) offer a notable example, Schroy
(1979), demonstrating the feasibility of such usage. Such models are based
on the Gaussian diffusion of particles as they move away from a source, and
they predict the maximum downwind concentration of the contaminant as a func-
tion of distance from the source.
The third model type, based on aerosol mechanics, is exemplified by
Davis (1971). Here, the modeling approach is based on the properties and
mechanics of airborne particles, such as particle generation, coagulation,
and settling. Because the use of this method involves solutions of compli-
cated equations, use of a computer is recommended.
The discussion section at the conclusion of the Clement Associates
report notes that mass balance equations are both the simplest and most com-
monly used approach for estimating exposure to a substance in an enclosed
space. However, there is no widely accepted methodology for this type of
analysis.
Berman (1982) provides methodology for estimating workplace exposure
to PMN substances. His approach is based on a collection of typical and worst
case approximations combined with easily understood equations and nomographs.
For bagging operations, a typical concentration of 2 mg/m3 of material (30%
of which is respirable) is assumed in the air surrounding the bag packer and,
if warranted, a worst case value of 30 mg/m3 (also 30% respirable) is assumed
for especially dusty conditions. For drum filling of liquids, a saturated
plug of contaminated air equivalent to the barrel volume is assumed to be
generated for every container filled. The concentration and total volume of
A-4
-------
contaminated air so released is calculated using an equation which accounts
for vapor pressure, temperature, and gram-molecular weight of the liquid, as
well as the filling rate and filling mode (top filling versus bottom filling).
Contribution from spillage is ignored. A typical ventilation rate for bag-
ging and drumming areas is assumed to be 85 m3/min (3,000 ftVmin) and a worst
case value is 14.2 mVmin (500 ftVmin). For outdoor operations with only
minimal structural protection, the average wind velocity of North America,
14.5 km/h (9 mph), is assumed in place of a building ventilation rate.
The validity of this methodology remains unknown, because insuffi-
cient data are available for verification. However, the Berman report repre-
sents the most complete set of guidelines presently available for estimating
workplace exposure to PMN chemicals during bagging and drumming activities.
2. NIOSH Studies
A visit to the NIOSH offices in Cincinnati revealed the existence
of two NIOSH studies which were highly relevant to the current study. These
were entitled "Health Hazard Control Technology Assessment of the Silica Flour
Milling Industry" and "Control Technology Assessment of Dry Chemical Bagging
and Filling Operations." A set of three in-depth survey reports was obtained
for each study (Caplan et al. 1981a; Caplan et al. 1981b; Caplan et al. 1981c;
Cooper 1984; Cooper 1983a; and Cooper 1983b). The silica flour industry study
also included a summary report.
The purpose of the silica flour processing plant study (Caplan et al.
1981a, 1981b, 1981c) was to evaluate the effectiveness of health hazard con-
trol procedures that are presently being used or being developed for bagging,
handling, and shipping operations. The specific evaluation of bag filling
operations comprised only a small portion of the overall study.
The authors noted that bag filling machines used in silica flour
plants vary in design and operation. They found the most common type to be
a pneumatically operated, forced-flow filling system with one to four spouts.
For very fine product flour (~ 5 urn), a single-spout, auger-feed type machine
is normally used. Each bagging machine is usually equipped with a local ex-
haust ventilation arrangement that is connected to a main exhaust system.
The bagging machines are either manual or semiautomatic systems, and fill
multi-ply, valve-type paper bags.
Air measurements of respirable dust contamination were made near
potential dust sources and at general work areas using MSA Gravimetric Dust
Samplers with cyclone separators. These samples were analyzed quantitatively
for respirable dust and respirable silica dust. At the operation where the
main product flour was bagged (at all three plants surveyed), the average
respirable silica dust concentration was 84 ug/m3 (56% silica content). At
two locations where fine product flour (< 38 urn) was being packed intermit-
tently, the respirable silica dust concentrations averaged 451 ug/m3 (76%
silica content). Table A-2 describes and summarizes the ventilation control
systems at the various bag filling operations.
A-5
-------
Table A-2. Ventilation Control Systems Used in Silica Flour Packaging
Location/operation
Description of system
Velocity measurement
m/min (fpm)
Respirable
dust concentration
Silica
Total Silica content
((ig/m3) (Mg/mJ) (X)
Remarks
1. 4-spout main product
packer, packing 100-Ib
bags
a. behind packer
machine
b. directly behind
packer operator
1.5 x 1.2 m (5 x 4 ft) hood
directly behind 4 spouts of
machine
1.2 x 10 cm (4 ft x 4 in.)
slot pulls air from con-
veyor belt
at hoodface - 14 (45) av.; 103
30 cm (1 ft) from face -
6.1 (20) av.; at slot -
305 (1,000) av.
Fair control - air velocity
fairly uniform around spouts.
Slot behind operator pulls in
96 93 opposition to packer machine
hood. It could be more stra-
tegically located. Host dust
emanates from packer.
cn
2. 4-spout main product
packer, packing 100-Ib
bags behind packer
machine
1.3 x 8.4 m (53 x 33 in.)
lateral hood, plus area fan
above head of operator
at hoodface (av.):
top = 40.8 (134)
middle = 35.3 (116)
bottom = 8.5 (28)
Good control; air moves from
145 63 43 above and back of operator
downward and to rear of packer
machine. Host dust emanates
from general environment.
4-spout main product
packer, packing 100-Ib
bags
a. behind packer
machine
b. below and behind
packer operator
1.5 x 1.1 m (60 x 45 in.)
lateral hood, directly be-
hind packer machine, moves
air laterally away from
operator
76 x 10 cm (30 x 4 in.)
lateral slot moves air
down and away from operator
at hoodface (av.):
top = 105 (343)
middle = 109 (356)
bottom - 88 (288)
rt end = 366 (1,200)
control = 655 (2,150)
left end = 335 (1,100)
212
112
53
Good control; air moves from
above and back of operator
downward and to rear of packer
machine. Most dust emanates
from general environment.
(continued)
-------
Table A-2. (Concluded)
Location/operation
Description of system
Velocity measurement
m/min (fpm)
Respirable
dust concentration
Silica
Total Silica content
(ug/m1) (pg/m») (X)
Remarks
4. 2-spout fine product
packer, packing and
palletizing
a. left side of No. 1
fill spout
b. behind No. 2 fill
spout
(56 x 36 cm) 22 x 14 in.)
side hood, adjacent to
No. 1 spout
18 x 11 cm (7 x 4-1/2 in.)
side
at hoodface (av.):
top = 55 (180)
middle = 30 (100)
bottom = 12 (40)
at slot = 1,460 (4,800)
at spout = 274 (900)
686
578
84
Uneven airflow, good control
control at No. 1 packer; poor
at No. 2 packer, cross drafts
and palletizing permit dust
dispersion. Most dust ema-
nates from packer spouts and
bag handling.
i
-j
5. 1-spout fine product
packer, packing and
palletizing
47 x 29 cm (18-1/2 x
11-1/2 in.) hood behind and
below spout
at hoodface (av.):
top = 61 (200)
middle = 75 (247)
bottom = 32 (105)
Good air velocity, but open
344 119 58 sides and remoteness of hood
from spout, spills on floor
and cross drafts reduce ef-
ficiency. Dust from packer
and general environment.
Source: Caplan et al. (1982).
-------
Local exhaust ventilation systems which were observed in these fa-
cilities included exhaust vents behind the bag packer fill nozzles (a system
of this nature is described and shown in figure form in the discussion of
Bureau of Mines studies) and beneath the operator's seat. The use of an over-
head fan blowing down from above and behind an operator was also observed in
one location in combination with local exhaust dust vents. Figure A-l shows
a typical relationship between the operator's location, the open side of the
bag packer hood, and the dispensing nozzles. The ventilation systems observed
in these facilities were capable of maintaining silica dust concentrations
close to or below the MSHA TLV of 100 (jg/m3, when main product silica flour
was being packed. However, they were not capable of reducing dust levels to
the NIOSH recommended level of 50 ug/m3.
The three in-depth surveys of dry chemical bagging operations were
conducted at plants which produced calcined diatomite, an asbestos product,
and a herbicide (Cooper 1983b; Cooper 1984; and Cooper 1983a).
An in-depth survey of three high volume, valve-type bagging opera-
tions of calcined (product with particle diameters predominantly < 40 urn) and
fluxed calcined (product with particle diameters predominantly < 125 pm) diat-
omite was conducted at Manville Products Corporation in Lompoc, California.
Two of the systems were high volume, manual bagging operations which involved
filling hand-tucked-valve bags. The third system consisted of a completely
automatic, high volume system for filling pasted-valve bags. Factors which
served to minimize exposure included automation, ventilated capture hoods at
the packers and in the palletizing areas, and a combination of ventilated and
non-ventilated hoppers and hoods beneath the packers and conveyor belts.
Area, source, and personal respirable dust samples were collected
and analyzed. The control technology systems at this plant were capable of
maintaining average respirable "free" crystalline silica dust (cristobalite)
concentrations below 0.03 mg/m3 at the two manual packaging stations being
evaluated. At the automated packaging station, these concentrations were
maintained below 0.02 mg/m3.
Air velocity, volumetric flow rates, and flow patterns were also
obtained and evaluated for the bagging and palletizing areas. Air velocities
entering the emission control hoods on the bagging machines ranged from 110
to 910 m/min (370 to 3,000 fpm).
The NIOSH work on asbestos bagging is not directly relevant to the
current study, because the fibers per cm3 unit used to describe air concen-
trations cannot be readily compared with the mg/m3 unit more commonly used
to describe air concentrations of other non-fibrous particulates which are
discussed in the literature and modeled in the current laboratory study.
The in-depth survey of a continuous, open mouth bagging operation
was conducted at Monsanto in Muscatine, Iowa. The herbicide being bagged dur-
ing the survey was AVADEX BW (particle size - 24 to 40 mesh). The bagging
system consisted of a fully automatic open mouth bagger, the Sackmatic Model
207 manufactured by Quachita Machine Works and ancillary automated equipment.
A-8
-------
Bag Packer
Mood with
Dust Exhaust
to the Rear
Operator's Seat
\| ' Indicates direction of exhaust ventilation air flow
Source- Adapted from Caolan et al (1982).
Figure A-l. Operator location and exhause ventilation
patterns for a silica sand bag packer.
A-9
-------
Controls included an exhaust hood over the bagger, local ventilation, general
ventilation, and system automation.
Area and personal samples were taken for total and respirable dust.
Total respirable dust in the bagging room ranged from 0.1 to 1.5 mg/m3 during
testing. Samples were also taken inside the hood. Ventilation velocity and
flow rate were measured for the bagging system. Air velocities entering the
emission control hoods on the bagging machines ranged from 12 to 18 m/min (40
to 60 fpm).
3. Bureau of Mines Studies
The Bureau of Mines (BOM) has performed research on dust control
during the bagging of silica sand and silica flour. One of the BOM studies
(Volkwein, 1978) includes emissions and bagging area concentrations, as well
as a dust control hood design suitable for bagging equipment common to the
silica sand industry. Figure A-2 shows a typical hood design. Bagging emis-
sions were of 17.5 to 18.3 mg of respirable dust per bag loaded. These emis-
sion rates were measured in a room housing a four-nozzle fluidized valve pack-
ing machine. Substantial background dust drifting into the bagging area from
silica flour grinding mills had to be accounted for by the sampling strategy.
The production rate during the sampling period varied from 1,170 bags (a mix-
ture of 50- and 100-lb bags) to 2,100 bags per work shift. Respirable dust
levels at the worker station ranged from 0.118 to 0.607 mg/m3 during the
sampling.
Another BOM study (Vinson et al. 1981) reports on the effectiveness
of bagging machine hood enclosures at controlling emissions in three different
facilities approximately 3 years after they were installed. In each of these
facilities, sulfur hexafluoride was released inside the hoods at a point ad-
jacent to the filling nozzle. Air samples were then taken to see if the
tracer gas was removed by the hood exhaust system.
In the first of the three facilities, the average velocity of the
air entering the hoods on a fluidized-type four nozzle bagging machine was
37 m/min (120 fpm). Jets of ventilation makeup air entered the bagging area
(a plywood enclosure with two openings for a bag conveyor) through grillwork
in the roof and produced turbulence on the open side of the hoods (each of
the four nozzles had its own hood). This turbulence caused some dust to es-
cape the hoods, and the tracer gas also exhibited this pattern.
In the second facility a three nozzle machine was enclosed in a room
with its own ventilation system. Average air velocity entering the hoods was
29 m/min (95 fpm). Tracer gas studies were more encouraging, as very little
gas could be detected outside the hoods. BOM personnel attributed the improved
performance, compared to the first facility, to the lack of jets of air from
the ventilation makeup.
The third facility contained a bagging machine in an open area at
one end of a large building used for loading box cars and trucks. The bag-
ging hoods were particularly well constructed and maintained and the perimeter
A-10
-------
,Ventitaflori hood
Intokt or
Sag clamp
Note: '/g-incft clearance
between plate
and hood
Deflector
plate
x Spill funnel
Figure A-2. BOM dust ventilation hood design.
(Source: Volkwein 1978.)
A-ll
-------
of the hood configuration was flared to reduce the pressure loss of air enter-
ing the hoods. The tracer gas studies indicated that the hood system was
working very well. The intake air velocity of these hoods was relatively high,
65 m/min (213 fpm).
Other BOM studies, which are not as directly relevant for this pro-
gram, have also been performed on silica bagging related to dust suppression
techniques, e.g., use of water sprays and foams (Volkwein et al. 1982; Volkwein
et al. 1983). It is interesting to note that average respirable dust concen-
tration beside bagging machines without water spray dust suppression averaged
1.21 mg/m3 and 1.76 mg/m3 at two locations (Volkwein et al. 1983).
BOM findings suggest that over half the worker exposure to respir-
able alpha quartz dust in bagging areas may consist of background dust. This
situation demonstrates the importance of isolating bagging areas from other
industry operations and providing clean incoming ventilation air.
C. Industry Characterization Summary
The bagging and drumming of PMN chemicals takes place under a wide
range of conditions which influence worker exposure during packaging.* These
variables include degree of automation; employee work practices; level of
occupational health awareness (managers and employees); filling equipment de-
sign, age, and frequence of maintenance; container and closure design; plant
layout and ventilation; effectiveness of emission control devices, if present;
employee use of personal protective gear; unit value of product (lower product
loss rates, including packaging emissions, will be permitted if the product
value is high); and product flammability. Some volatile liquids and dust-prone
solids are carefully controlled because of the fire hazard they represent in
the absence of such controls. In an effort to assemble a broadly applicable
data base and methodology for estimating worker exposure in the bagging and
drumming of PMN chemicals, only the most relevant of these variables may be
accounted for.
1. Bagging of Solids
In the bagging of PMN solids it is unlikely that production volumes
will justify fully automated packaging lines in many instances (Bomberger
et al. 1983). However, the use of automated or semiautomated equipment re-
mains a key variable in predicting worker exposure. The product will typi-
cally be placed into bags (open top or valve) by gravity feed, auger feed,
or vibrator feed equipment. Emission controls may or may not be present.
Frequently, an operator may be required to manually staple or otherwise seal
the bag tops.
The literature is inadequate to permit an exhaustive review of bagging and
drumming practices for all types of chemical products. However, as previously
noted, Bomberger et al. (1983) is a good, up-to-date source of information
describing typical filling technology. No attempt is made in this document
to improve on that compilation.
A-12
-------
The matrix shown in Figure A-3 is based on these primary equipment
variables combined with a gross segregation of solids into coarse (particle
diameters predominantly > 100 urn) and fine (particle diameters predominantly
< 100 urn). The resulting 24 categories represent different packaging environ-
ments which the MRI sampling team would like to characterize. Realistically
only a few of these equipment-material combinations will be examined. However,
during the field work, an effort will be made to sample as many combinations
as possible.
2. Drumming of Liquids
In the drumming of PMN liquids it is expected that both fully auto-
mated and semiautomated filling would be commonly practiced. The product will
be either top filled (dispenser stays near the top of the barrel, liquid
splashes freely during filling) or bottom filled (dispenser remains submerged
inside the barrel, as much as possible, to minimize volatile emissions).
Emission controls may or may not be present. An operator may be required to
manually screw a cap on each barrel or wipe droplets off the tops of the bar-
rels (Bomberger et al. 1983).
The matrix shown in Figure A-4 is based on these primary equipment
variables combined with a general subdivision of liquids into heavier liquids
with relatively low vapor pressures, g 0.3 kPa (2.25 mm Hg) at 20°C, and
lighter liquids with relatively high vapor pressures, > 0.3 kPa at 20°C.
This criterion follows a precedent established by Bomberger et al. (1983) in
a similar analysis. The resulting 16 categories represent different packag-
ing environments which MRI would like to characterize. As with the bagging
matrix, complete coverage of all categories in the drumming matrix was not
possible within the scope of the current contract. However, it served as a
guideline for sampling as many equipment-material categories as possible.
For the sampling of liquid drumming lines during the current study,
emphasis should be placed on sampling 55-gal steel drum filling lines wherever
possible. The study will focus on one common container and facilitate the
comparison of the effects of other variables on drumming emissions.
Vapor pressure is a key parameter for any analysis of emissions from
the drumming of liquids. However, it was not possible for MRI to restrict its
sampling effort to materials having a limited range of vapor pressures, because
PMN chemicals are likely to exhibit the considerable range of vapor pressures
common to industrial chemicals in general. Table A-3 lists vapor pressure and
a number of other useful properties for a group of organic liquids representing
a broad range of vapor pressure.
III. ARRANGEMENTS FOR PLANT VISITS
Substantial effort was devoted to the task of arranging for on-site
sampling of bagging and drumming operations. Initially, the response to re-
quests for industry data and permission to sample in packaging facilities was
disappointing. However, subsequent to the first laboratory study described
in Appendix B, sampling was conducted at two plants with bagging and drumming
operations.
A-13
-------
Range of Respirable Dust Levels in the Air of the
Bagging Area During Operation
(mg/m3)
Coarse
Materials
whose
Particle
Diameters are
Predominantly
> 1 00/tfli
Fine
Materials
whose
Particle
Diameters are
Predominantly
< 100/zm
Gravity Feed
Fully
Automated
With
Dust
Control
Without
Dust
Control
Semi-
Automated
With
Dust
Control
Without
Dust
Control
Auger Feed
Fully
Automated
With
Dust
Control
Without
Dust
Control
Semi-
Automated
With
Dust
Control
Without
Dust
Control
Vibrator Feed
Fully
Automated
With
Dust
Control
Without
Dust
Control
Semi-
Automated
With
Dust
Control
Without
Dust
Control
Figure A-3. Model sampling matrix for banqinn solids.
A-14
-------
Range of Vapor Concentrations in the Air of the
Drumming Area During Operation
(mg/m3)
Liquids
with
Vapor
Pressures
10.3KPa
at20°C
Liquids
with
Vapor
Pressures
>0.3KPa
at20°C
Top Fill
Fully Automated
With
Emission
Controls
Without
Emission
Controls
Semi -Automated
With
Emission
Controls
Without
Emission
Controls
Bottom Fill
Fully Automated
With
Emission
Controls
Without
Emission
Controls
Semi -Automated
With
Emission
Controls
Without
Emission
Controls
Figure A-4. Model sampling matrix for drumming liauids.
A-15
-------
Table A-3. Vapor Pressures and Other Useful Properties for a Group of Organic Liquids
Representing a Broad Range of Vapor Pressures
i
en
Vapor pressure
ftPa at (mm Hg
STP)* at STP)*
Ethyl ethera'b'c
Methlyene chlorideb'd
Acetone3 lb'e
n-Hexanebff
Isopropyl ether3'6
Methyl alcohol9'*1'6
Isopropyl acetate3'6
2-Butanone (MEK)3'b'e
Isopropyl alcohol3'6'*1'6
Methyl cyclohexane3'6'6
n-Propyl acetate3'6
Toluene1 'b'h'e
Propyl alcohola'b'h'e
Isobutyl acetate3'6
Methyl methacrylate3'6
Isobutyl alcohol3'6
Xylenei'b'h'6
(mixed, o-, m-, p-)
Isoamyl acetate-1'6
Styrene monomer1' 'e
Turpentine1'11'1
Vinyl toluene1'1
Oiethylphthalate3
59
52
30
20
16
13
10
9.3
5.9
5.7
4.7
4.0
2.8
2.7
1.7
1.6
1.3
0.08
0.67
0.67
0.13
0.07
440
390
227
150
119
95
73
70
44
43
35
30
21
20
13
12
10
6
5
5
1
0.5
Boiling
point
<°C)
35
40
57
69
69
64
84
79
83
101
102
111
97
118
100
107
140
143
146
149
171
296
Surface OSIIA
tension (mg/m3
(dynes/
cm)
17 1,200
1,800
24 2,400
1.800
2.100
260
950
25 590
980
2.000
840
28 750
500
700
410
300
~ 30 435
525
420
560
480
-
P.E.L.
) (ppm)
400
500
1,000
500
500
200
250
200
400
500
200
200
200
150
100
100
100
100
100
100
100
none
Flash
point
-45
1
-18
-22
-28
11
4
-6
12
-4
14
4
25
18
10
28
29
25
32
35
54
163
Lower explosive
limit (% by vol.)
1.
15.
2.
1.
1
7.
1
1
2.
1.
2.
1.
2.
2.
1.
1 2 (at
1.
1.
1.
0
.
-
9
5
6
1
4
3
8
8
0
2
0
2
1
4
7
212°F)
1
0
1
8
Upper explosive
limit (X by vnl
36
66
12.
7.
7.
36
8
10
12
6.
8
7.
13.
10.
8.
10.
7.
7
6.
-
-
8
5
9
7
0
1
5
5
2
9
0
5
1
Specific
. ) grav i ly
0 71
1.34
0.791
0.660
0.723
0.792
0.877
0 805
0. 785
0.770
0.887
0 866
0 804
0 871
0.940
0.806
0.87
0 876
0.905
0.857
0.890
1.12
Source: Fundamentals of Industrial Hygiene, 1979 and Kirk-Othmer, 1983.
*STP = standard temperature and pressure, i.e., 0°C and 760 mm of Hg.
tNonflammable in oxygen.
Hazards
.Mild irritation - eye, nose, throat
Narcosis.
dFire (severe)
Chronic systemic toxicity
6F»re
Polyneuritis
?Acute toxicity (narcosis, eye damage)
.Dermal absorption
Moderate irritation - eye, nose, throat
^Moderate irritation - upper respiratory tract
^Cumulative kidney damage
Fire (moderate)
-------
A. Industry Contacts
Initial attempts were made to gain access to plants through contacts
with national chemical manufacturers associations and by contacting local
firms in the Kansas City area.
1. National Associations
Contacts were established with the Chemical Manufacturers Associa-
tion (CMA), the Synthetic Organic Chemical Manufacturers Association (SOCMA),
and the Chemical Specialties Manufacturers Association (CSMA) in an effort to
locate facilities willing to cooperate. These three associations were thought
to represent a significant fraction of the companies involved in the bagging
and drumming of PMN chemicals.
2. Local Manufacturers
A review of several Kansas City business reference books and chamber
of commerce directories was completed to compile an expanded list of local
industries. Table A-4 lists the local industries considered for contact to
gain access to bag loading or drum filling process operations. Six telephone
contacts have been made from Table A-4 but no substantial leads have been
achieved in the initial attempt to find suitable test sites. However, two
referrals to other facilities were received.
B. Obstacles Encountered in Plant Visit Arrangements
One of the obstacles confronting MRI in the effort to gain entry to
packaging lines, is the inability to offer an incentive to cooperating plants.
The only rationale available to persuade industry personnel to cooperate is a
statement of the mutual benefit which both industry and EPA enjoy when EPA
rule making is based on sound data. Thus far, this line of reasoning has not
been effective.
As long as there is no compelling reason, industry personnel are
not inclined to allow outsiders into their facilities to examine their opera-
tions and gather data that may eventually be used to support regulations.
Given the present circumstances, plant managers and corporate officers appar-
ently feel that cooperation offers the prospect of negligible benefits and
potentially significant risks. However, we will continue to pursue any leads
which offer the possibility of plant visits and/or the acquisition of relevant
data.
A-17
-------
Table A-4. Potential Sampling Opportunities in Greater Kansas City
Along Bagging or Drumming Lines
Adhesives
H. B. Fuller Co.
National Starch & Chemical
Corp.
Agricultural Chemicals
Chem-Trol, Inc.
Farmland Industries
Mobay Chemical Corp.
Pfizer, Inc.
Building Materials
Aylward Products Co.
GAF Corp.
Owens-Corning Fiberglass Corp.
Sawyer Material and Sand Co.
Chemicals. General
Chemcentral-Kansas City
Dow Chemical Co.
Reichold Chemical, Inc.
Thompson-Hayward Chemical Co.
Union Chemicals
Dairy Products
Fairmont Foods
Zarda Brothers Dairy
Emulsifiers
Grindsted Products, Inc.
Feed
Carnation-Milling Division
Columbian Hog and Cattle
Powder Co.
National Alfalfa
Ralston Purina Co.
Industrial Chemicals
Century Laboratories, Inc.
Chemtech Industries, Inc.
Liquid Products, Inc.
Protective Coatings
Epic Manufacturing Co.
Outgo, Inc.
Inc.
A-18
-------
IV. REFERENCES
Berman DW. 1982. Methods for estimating workplace exposure to PMN substances
Clement Associates, Inc., Arlington, VA. Prepared for U.S. Environmental Pro-
tection Agency, Economics and Technology Division, Washington, D.C.
Bomberger DC, Boughton R, Endlich R, Fowler D, Gikis B, Ludwig F, Wilhelm D,
Witham C. 1983. Industrial process profiles to support PMN review: filling
of drums and bags, draft. SRI International, Menlo Park, CA. Prepared for
U.S. Environmental Protection Agency, Office of Toxic Substances, Washington
D.C. Contract No. 68-01-6016, Technical Directive 69.
Bureau of Mines. 1978. Dust control hood for bag-filling machines. Techno!-
logy News No. 54, Technology Transfer Group. Washington, DC: Bureau of Mines
U.S. Department of the Interior. Revised June 1983.
Caplan PE, Reed LD, Amendola AA, Cooper TC. 1981a. Report on an in-depth
survey of silica flour dust during packing, transfer and shipping at
Pennsylvania Glass Sand Corporation, Berkeley Springs, West Virginia. Draft
report. Cincinnati, OH: National Institute for Occupational Safety and Health.
Caplan PE, Reed LD, Amendola AA, Cooper TC. 1981b. Report on an in-depth
survey of silica flour dust during packing, transfer and shipping at the
Central Silica Company, Glass Rock, Ohio. Draft report. Cincinnati, OH:
National Institute for Occupational Safety and Health.
Caplan PE, Reed LD, Amendola AA, Cooper TC. 1981c. Report on an in-depth
survey of silica flour dust during packing, transfer and shipping at Ottawa
Silica Company, Ottawa, Illinois. Draft report. Cincinnati, OH: National
Institute for Occupational Safety and Health.
Clement Associates. 1981. Mathematical models for estimating workplace
concentration levels: a literature review. Clement Associates, Inc.,
Washington, D.C. Prepared for U.S. Environmental Protection Agency, Eco-
nomics and Technology Division, Washington, D.C.
Cooper TC. 1984. In-depth survey report: control technology for bag filling
operations at Union Carbide Corporation, King City, California. Draft report.
Cincinnati, OH: National Institute for Occupational Safety and Health. Report
No. 112.21b.
Cooper TC. 1983a. In-depth survey report control technology for a dry chem-
ical bagging and filling operation at Monsanto (Monsanto Agricultural Products
Company), Muscatine, Iowa. Draft report. Cincinnati, OH: National Institute
for Occupational Safety and Health.
Cooper TC. 1983b. In-depth survey report: control technology for bag filling
operations at Manville Products Corporation, Lompoc, California. Draft report.
Cincinnati, OH: National Institute for Occupational Safety and Health. Report
No. 112,23.
Davis RJ. 1971. A simple model for the estimation of aerosol concentration
in a closed vessel. Am Ind Hyg Assoc J 32:603-609.
A-19
-------
Fundamentals of Industrial Hygiene. 1979. Olishifski JB, ed. 2nd ed.
National Safety Council, Chicago, 111.
Kirk-Othmer. 1978-1984. Encyclopedia of chemical technology, 3rd ed.
New York: John Wiley and Sons.
Schroy JM. 1981. Prediction of workplace contaminant levels, pp. 190-206.
Symposium Proceedings on Control Technology in the Plastics and Resins Indus-
try, Atlanta, GA, February 27-28. Cincinnati, OH: National Institute for
Occupational Health and Safety. NIOSH Publication No. 81-107.
Vinson RP, Volkwein JC, Thimons ED. 1981. SF6 tracer gas tests of bagging
machine hood enclosures. Washington, DC: U.S. Department of the Interior.
Report of Investigations, Bureau of Mines, 8527.
Volkwein JC, Thimons ED. 1983. Ventilation hoods to reduce dust during bag
filling. Mining Equipment International.
Volkwein JC, Vinson RP, Thimons ED. 1982. Effectiveness of three water spray
methods used to control dust during bagging. Washington, DC: U.S. Department
of the Interior. Report of Investigations, Bureau of Mines, 8614.
Volkwein JC. 1979. Dust control in bagging operations, pp. 51-57. ACGIH
Symposium Proceedings on Industrial Hygiene for Mining and Tunneling, November
6-7, 1978. Cincinnati, OH: American Conference of Governmental Industrial
Hygienist.
V. ANNOTATED BIBLIOGRAPHY OF REFERENCES ON FILE
Table A-5 summarizes references on file at MRI. Each document in
the file is summarized below by document number.
P101
Bomberger DC, Boughton R, Endlich R, Fowler D, Gikis B, Ludwig F, Wilhelm D,
Witham C. 1983. Industrial process profiles to support PMN review: filling
of drums and bags, draft. SRI International, Menlo Park, CA. Prepared for
U.S. Environmental Protection Agency, Office of Toxic Substances, Washington,
D.C. Contract No. 68-01-6016, Technical Directive 69.
Technologies commonly, used for bagging>and -drumming solids and for
drumming liquids are discussed.' Methodologies for modelling workplace con-
centrations are also discussed.
Typical equipment and modes of operation are described in detail
(i.e., bag and drum designs, container closure methods, conveying and filling
equipment configurations, and emission control systems). Vapor and particu-
late emission rates and the resulting workplace air concentrations during bag
and drum filling of chemicals are estimated.
A-20
-------
Table A-5 Summary of References on File at MRI
I
PO
Doc.
No.
P101
P102
P103
P104
P105
P106
P107
P108
P109
P110
Pill
P112
P113
P114
P115
P116
P117
P118
P119
P120
P123
P124
P125
P126
P127
P128
P129
P130
P131
P132
P133
P134
P135
P136
P137
P138
P139
P140
P141
P142
Reference
Bomberger et al. (1983)
SRI (1980)
Clement Assoc. (1982)
Clement Assoc. (1981)
Caplan et al. (Sep. 1981)
Caplan et al. (Nov. 1981)
Caplan et al. (Aug 1981)
Cooper (Jan. 1984)
Cooper (Oct. 1983)
Wang (1981)
Wolfe et al. (1978)
Winegar (1983)
Clarke (1983)
Buyers Guide & Dir. (1982)
EPA (1983)
NIOSII (July 1978)
Caplan et al. (June 1982)
Becker et al. (1979)
Cooper (Mar. 1983)
Wechsler et al. (1982)
Hertle (1981)
Holmgren (1982)
Cahners (1982)
Checkoway & Williams (1982)
Davies et al. (1982)
Volkwein et al. (1982)
Vinson et al. (1981)
Brest in & Niewicdomski (1982)
Ward & Campbell (1983)
Phillips & Jones (1978)
Thomson (1977)
Midwest Int. (1981)
ACS (1983)
ACS (1982)
Freed et al. (1983
Adkins et al. (1983)
Dixon et al. (1983)
Versar (Apr. 9. 1982)
Versar (Apr. 7, 1982)
Flores (1983)
Specific
chemicals
X
X
X
X
X
asbestos
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Physical Workplace
properties filling
of chemicals operations*
X X D,B
X X 0,8
X D.B
X X B
X X B
X X B
X B
X B
X X B
X B
X D.B
X
B
X B
X B
X X
X B
X
B
B
B
X
X
X X
Emissions Workplace
controls
X X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X X
X
X
X
X
X
Workplace
concentrations
or exposures
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Dispersion
mode 1 s
X
X
X
X
X
(continued)
-------
Table A-5. (Concluded)
Doc.
No.
P143
P144
P145
P146
P147
P148
P150
P151
P152
P154
P155
P156
An J--
Reference
r \ -
Dunn et al. (1983)
Tomb & Treattis (1983)
Peach & Myers (1981)
Schroy (1981)
Volkwein (1979)
Bureau of Mines (1983)
Volkwein et al. (1983)
Volkwein & Thimons (1983)
Bureau of Mines (undated)
O'Leary et al. (1983)
Callahan (1978)
Cahners (1983)
Specific
chemicals
X
X
X
X
X
X
Physical
properties
of chemicals
X
X
Workplace
filling
operations*
X
B
B
0
B
B
X
X
Emissions Workplace
controls
X
?
X X
X
X X
x'
X
X
X
X
X X
Workplace
concentrations
or exposures
X
X
X
Dispersion
models
x
(\>
ro
-------
The discussion on volatile emissions from drumming liquids is based
largely on data from the petroleum refining industry, because no other data
were available. Emission rates for drum filling of a light liquid with a
relatively high vapor pressure (benzene) and for a heavy liquid with a rela-
tively low vapor pressure (cyclohexanol) are estimated under three sets of
filling conditions: a high rate of emissions (top filling with no collection
of vapors and no routine maintenance of fittings which are not inside a hood),
a moderate rate of emissions (top filling with 95% collection of vapors), and
a low rate of emissions (bottom filling with 95% collection of vapors).
The authors conclude that it is difficult to make general statements
on packaging practices, especially on emission control practices, in the
chemical manufacturing industry. The bulk of packaging machinery is designed
and sold for the packaging of foods, cosmetics, and Pharmaceuticals. Thus,
emission controls, if present at all on original equipment, are generally for
control of nuisance dusts.
P102
SRI. 1980. Control technology assessment of the pesticides manufacturing
and formulating industry, draft. SRI International, Menlo Park, CA. Pre-
pared for National Institute for Occupational Safety and Health, Cincinnati,
OH. Contract No. 210-77-0093.
The 50-page product packaging section of this 652-page report on
pesticide production covers methods used to reduce operator exposure during
13 specific bagging and liquid filling operations. Most dry pesticide is
packaged in bags and most bags are open-top rather than valve-top or bags
formed of heat-sealable films. Drum filling is generally semi automated.
P103
Clement Associates. 1982. Methods for estimating workplace exposure to PMN
substances. Clement Associates, Inc., Arlington, VA. Prepared for U.S.
Environmental Protection Agency, Economics and Technology Division, Washington,
D.C.
Representative scenarios were developed for operations associated
with processing chemicals. Operations considered were drumming of liquids,
drumming and bagging of solids, cleaning and maintenance, and sampling and
analysis. Rates of generating emissions during filling of drums with liquids
depend principally on drum size and filling rate (drums/h). For each drum
filled, a plug of chemical-saturated air is assumed to be emitted. The com-
plex dependence of dust formation during drumming and bagging on properties
of solids gives a range of values for dust generation from a variety of re-
lated processes. Models are applied to the data developed for the process
operations to estimate workplace exposures. Airborne particulate concentra-
tions associated with handling a few specific solid chemicals were compiled
from the literature. For example, total particulate concentration in air
while bagging azelaic acid was 0.2 to 6.2 mg/m3 (30% respirable) and in drum-
ming azo dyes was 13.1 mg/m3.
A-23
-------
P104
Clement Associates. 1981. Mathematical models for estimating workplace con-
centration levels: a literature review. Clement Associates, Inc., Washington,
D.C. Prepared for U.S. Environmental Protection Agency, Economics and Technol-
ogy Division, Washington, D.C.
Mathematical models for estimating workplace concentrations of chem-
icals include mass balance models (simple models, simple equilibrium models,
and complex models); the Schroy model; and the Davis model. Measured work-
place concentrations for benzene, hexane, and toluene are compared with con-
centrations predicted by use of the equation for the simple mass balance
equilibrium model: Cgq = G/kQ, where C is the equilibrium concentration,
G is the generation rate, k is the mixing factor, and Q is the ventilation
flow rate. The Schroy model [see Schroy (1981)], which uses both emission
rate and Gaussian diffusion, predicts maximum downwind concentration as a
function of distance from the source. The complex Davis model calculates
aerosol concentration in a closed vessel.
P105
Caplan PE, Reed LD, Amendola AA, Cooper TC. 1981. Report on an in-depth
survey of silica flour dust during packing, transfer, and shipping at
Pennsylvania Glass Sand Corporation, Berkeley Springs, West Virginia. Draft
report. Cincinnati, OH: National Institute for Occupational Safety and
Health.
Ventilation and other dust control measures during bagging of silica
flour into 50- or 100-Ib bags and bulk loading of railroad cars and trucks are
described in detail with diagrams of plant layout and machinery. The average
silica dust concentration was 578 ug/m3 in the vicinities of the one-spout
packer (bagging) machines for the 10-15-30 micron product. High silica dust
concentrations were probably caused by bag breakage, ineffective ventilation
at packer spouts, and poor housekeeping. The ventilation was better at the
four-spout packer and at the packer for the 5-micron product.
P106
Caplan PE, Reed LD, Amendola AA, Cooper TC. 1981. Report on an in-depth
survey of silica flour dust during packing, transfer, and shipping at the
Central Silica Company, Glass Rock, Ohio. Draft report. Cincinnati, OH:
National Institute for Occupational Safety and Health.
Atmospheric dust concentrations and ventilation control systems were
evaluated for packing (filling of 100-Ib bags with a four-spout packer) and
transfer operations. The bag handling workers rotate from packing to loading
to stacking opeations in each shift. The average silica dust concentrations
were 112 ug/m3 at the packer area and 165 ug/m3 at the bag handling area.
A-24
-------
The ventilation was so good that dust concentrations were practically the
same whether bagging and truck loading were occurring or not. The high con-
centrations (264 ug/m3 during loading compared to < 28 ug/m3 during no load-
ing) at the conveyor discharge, where bags were off-loaded from the conveyor
and stacked on pallets, apparently arose from the bags as leakage, spills,
or surface contamination.
P107
Caplan PE, Reed LD, Amendola AA, Cooper TC. 1981. Report on an in-depth
survey of silica flour dust during packing, transfer and shipping at Ottawa
Silica Company, Ottawa, Illinois. Draft report. Cincinnati, OH: National
Institute for Occupational Safety and Health.
Dust concentrations in the air, ventilation control systems, and
other dust control techniques were evaluated in the packing area (where 100-lb
paper bags were filled with silica flour), in the bag handling area, and in
the boxcar loading area. Bag handling workers normally rotate among the jobs
of operating the packer machine, transferring bags, and stacking bags in box-
cars or trucks. Two packer machines had two fill-spouts and three had four
spill-spouts each. Water injection into the product and water spray onto bag
surfaces reduced atmospheric dust concentrations by 60 to 80% (80% in the bag
handling areas). Local exhaust, the water control methods, and housekeeping
reduced packing station dust concentrations to 50 ug/m3 air.
P108
Cooper TC. 1984. In-depth survey report: control technology for bag filling
operations at Union Carbide Corporation, King City, California. Draft report.
Cincinnati, OH: National Institute for Occupational Safety and Heatlh. Report
No. 112.21b.
Exemplary controls used in packaging chrysotile asbestos fibers at
the title plant maintained bagging area atmospheric concentrations at < 0.9
fiber/cc. Bag filling and bag flattening operations are enclosed, and the
enclosures are equipped with exhaust ventilation. Pellets are shrink wrapped
and stretch wrapped; pelletizing is automated. Bagger rooms are isolated.
Bagging equipment used includes three single-spout force-flow packers and one
single-spout auger packer. Filled bag weights are 10 to 82 Ib; packaged den-
sities range from 5 to 45 lb/ft3 (pellets).
P109
Cooper TC. 1983. In-depth survey report: control technology for bag filling
operations at Manville Products Corporation, Lompoc, California. Draft report.
Cincinnati, OH: National Institute for Occupational Safety and Health. Report
No. 112,23.
The control technology used at the title plant for three high-volume
operations bagging calcined and fluxed calcined diatomite maintained average
respirable "free" crystalline silica dust (cristobalite) at concentrations
A-25
-------
< 0.03 mg/m3 air at the two manual packaging stations and at < 0.02 mg/m3 at
the automated packaging station. The manual bagging operations filled hand-
tucked-valve bags; the completely automatic system filled pasted-valve bags.
Ventilated capture hoods were used at the packers and in the pelletizing areas,
and both ventilated and nonventilated hoppers and hoods were used beneath the
packers and conveyer belts. The relations between occupational atmospheric
dust exposures (as determined by analysis of area, source, and personal
respirable dust) and control systems were evaluated.
P110
Wang CCK. 1981. Walk through survey report of the pesticide/herbicide bag-
ging operation at Monsanto Agricultural Products Company, Muscatine, Iowa.
Cincinnati, OH: National Institute for Occupational Safety and Health. NTIS
No. PB83-243840.
The title bagging operation for filling 50-Ib paper bags with gran-
ulated (40 to 50 mesh) pesticide/herbicide formulations and the control tech-
nology are described in detail. Several figures depict the machinery used.
The strict industrial hygiene policies are also discussed. The control tech-
nology may be successfully adapted to control dust problems in other bag fill-
ing operations.
Pill
Wolfe HR, Staiff DC, Armstrong JF. 1978. Exposure of pesticide formulating
plant workers to parathion. Research Triangle Park, NC: U.S. Environmental
Protection Agency, Office of Research and Development. EPA-600/J-78-042,
NTIS No. PB-287 134.
Worker respirator and dermal exposure potential was determined near
certain work stations at a parathion formulating plant. Ordinary work cloth-
ing was assumed to offer minimum protection. Mean parathion exposure was
67.3 mg/h of work activity; for inhalation exposure, the value was 0.62 mg/h.
Bagging machine workers were exposed to more parathion than were workers who
operated the bag/closing machine or handled the filled bags for storage or
shipment. Calculated inhalation exposure for full-time baggers was 0.69 ±
0.77 mg/h and for baggers who were also mixers, the inhalation exposure was
0.93 ± 0.51 mg/h. No information is given about the bagging or control ma-
chinery, and the plant is not identified.
P112
Winegar DL. 1983. Machinery-filling, dry products. Package Engineering
28(4):229-231.
Gravimetric and volumetric filling equipment for dry materials are
described with schematics. The two types of gravimetric equipment are net
weighing and gross weighing fillers. Volumetric filler types include timed
vibrator, auger, vacuum-filled container, gravity-filled pocket, and vacuum-
filled pocket fillers. (They are listed in order from least to most accurate.)
A-26
-------
Types of materials, containers, and container size suitable for use with each
filler type are mentioned.
P114
Buyers Guide and Directory. 1982. Package Engineering 27(11).
This directory issue of Package Engineering lists the names of
manufacturers of machinery and containers, contract packagers, and research
and testing laboratories.
P115
EPA. 1983. Premanufacture notification requirements and review procedures.
Washington, DC: U.S. Environmental Protection Agency. 40CFR720.
Premanufacture notification (PMN) requirements and review procedures
through 48FR41132, September 13, 1983, include specifications for procedures
for reporting new chemical substances by manufacturers and importers under
Section 5 of the Toxic Substances Control Act, applicability, notice forms,
disposition of notices, confidentiality and public access to information,
commencement of manufacture or import, and compliance and inspection.
P116
NIOSH. 1978. Criteria for a recommended standard occupational exposure
during the manufacture and formulation of pesticides. Cincinnati, OH: National
Institute for Occupational Safety and Health. Publication No. 78-174.
Packaging operations in the pesticide industry may be done remotely
(automatic dispensing, weighing, and sealing) but is more frequently done with
a local exhaust system or with no specific controls. Pesticides may be pack-
aged in the form of powders, liquids, dust, granules, or pressurized aerosols.
Containers vary from small glass bottles to drums and railroad tank cars.
Bagging or mixing stations are often work sites with the highest exposure po-
tential. At a plant formulating 4 and 50% carbaryl dusts, workers at such
stations had a mean respiratory potential exposure of 1.1 mg/h work activity.
In the formulation of a 50% DDT wettable powder, the baggers who filled 50-Ib
and 4- or 5-lb bags had potential exposures of 160 and 539 mg/h, respectively.
Workers in liquid filling operations may be exposed to the vapors before cap-
ping or to the pesticide by splashing. Remote control filling is recommended,
but bottom filling of drums instead of top filling would minimize splashing.
Properly designed local exhaust ventilation is recommended for dust collection
during solids packaging operations.
P117
Caplan PE, Reed LD, Amendola AA, Cooper TC. 1982. Health hazard control
technology assessment of the silica flour milling industry. Draft report.
Cincinnati, OH: National Institute for Occupational Safety and Health.
A-27
-------
Three silica flour mills were evaluated in depth. The primary oper-
ations investigated were bag packing, conveyor transfer of bags, palletizing,
loading of bags in boxcars and trucks, and bulk loading of hopper cars and
trucks. Several innovative dust control procedures were found to be effective.
These included: (1) the injection of agglomerating agents such as water and
Deter Micro Foam® into products; (2) water spraying of outer surfaces or
product-filled bags during conveyance; and (3) bulk loading of silica flour
into enclosed hopper trucks and railroad hopper cars under local exhaust ven-
tilation control. Plant personnel have estimated that half of the environ-
mental silica dust problems may be effectively controlled by good work prac-
tices and effective housekeeping procedures.
P118
Becker D, Fochtman E, Gray A, Jacobius T. 1979. Methodology for estimating
direct exposure to new chemical substances. Washington, DC: U.S. Environ-
mental Protection Agency, Office of Toxic Substances. EPA-560/13-79-008, NTIS
No. PB80-102262.
The work reported was directed toward the development of a procedure
for the orderly and rapid prediction of direct human exposure which might re-
sult from the manufacture of new chemical substances. The procedure developed
involves the following steps: (1) prediction of unavailable physical and chem-
ical properties from analogs and general chemical knowledge, (2) prediction of
production volume based upon company size, current markets and total market
volume, (3) prediction of chemical operator exposure and exposures in the
vicinity of the plant based upon fugitive emissions, and (4) prediction of
consumer exposure based upon active use and passive use of the chemical.
P119
Cooper TC. 1983. In-depth survey report control technology for a dry chem-
ical bagging and filling operation at Monsanto (Monsanto Agricultural Products
Company), Muscatine, Iowa. Draft report. Cincinnati, OH: National Institute
for Occupational Safety and Health.
Controls for packaging the granular herbicide AVADEX BW in 50-lb
pinch bottom bags by a fully automatic open-mouth bagger included an exhaust
hood over the bagger, local and general ventilation, and system automation.
Relations among ventilation velocity, flow rate, and total and respirable
dust from personal, area, and hood interior samples were evaluated. Average
atmospheric concentrations over a 2-day period were 2.9 mg/m3 for total dust
inside the hood, 1.9 mg/m3 for total dust on the hood face, 1.5 mg/m3 for
total dust in the bagging room, 0.2 mg/m2 for respirable dust in the bagging
room, 1.1 mg/m3 for concentration of total dust to which two operators were
exposed, and 0.2 mg/m2 for the background level.
A-28
-------
P120
Wechsler AE, Eschenroeder AQ, Gilbert D, Loos K, Poston P, Stevens JM. 1982.
Feasibility of developing a comprehensive methodology for source identification
and environmental loading (materials balance). Athens, GA: U.S. Environmental
Protection Agency, Environmental Research Lab. EPA-600/3-82-047, NTIS No.
PB82-239286.
Materials balance methodologies are discussed from extraction of the
raw materials through manufacturing/processing, use, and disposal. The focus
is on releases to the outdoor environment. Fugitive processing emissions are
mentioned, but methodologies for estimating releases to indoor air are not
covered.
P123
Mertle B. 1981. Bulk packaging—from bauxite to bag, Reynolds makes easy
work of it. Material Handling Engineering 36:114-119.
Automatic packaging of alumina in 50- and 100-Ib bags is described.
Ten auger-type bag packing machines are used for hydrates and ground calcines,
and three air-type packers are used for unground calcined alumina.
P124
Holmgren RB. 1982. Electronics, skilled hands keep bag line's output high.
Package Engineering 27:69-72.
Automated packaging of powdered milk in 50-Ib bags is described.
Weighing, filling, and sealing are automated, but five men are required, one
each to attend the bagger, to hand-tie each bag liner, to oversee the sealer,
to stack the bags, and to operate the forklift truck. The first four fre-
quently alternate their jobs. Vacuum tubes at the point of filling, air
evacuation, and a large hood above the point of bag tying control dust.
P125
Cahners. 1982. Bag's valve, unitizing cut hazardous leakage. Cahners
Publishing Company, Denver, CO. Package Engineering 27:78-9.
Bag leakage in packaging silica flour was solved by switching to a
bag with a polyethylene valve and by shrink wrapping pallets before shipping.
Formerly, use of standard kraft bags (50- or 100-lb) with tuck-in sleeve and
pasted valve allowed a small amount of silica flour to force its way back out
through the bag's valve. Once the new bags are filled, the valve overlaps
itself to form a seal.
P126
Checkoway H, Williams TM. 1982. A hematology survey of workers at a styrene-
butadiene synthetic rubber manufacturing plant. American Industrial Hygiene
Association Journal 43(3):164-169.
A-29
-------
Of the workers at the title plant, workers in shipping and receiving,
who transport (packaging not explicity mentioned) the finished latex products,
were exposed to the lowest concentrations of butadiene (0.08 ppm), styrene
(0.13 ppm), and benzene (benzene is an impurity of styrene and toluene) (0
ppm) and the second lowest concentration of toluene (a styrene impurity, also
used as a tank cleaning agent) (0.05 ppm).
P127
Davies JM, Strunin L, Craig DB. 1982. Leakage of volatile anaesthetics from
agent-specific keyed vapourizer filling devices. Canadian Anaesthetics Society
Journal 29(5):473-476.
Agent-specific keyed vaporizer filling devices (designed to ensure
that an anaesthetic vaporizer is filled with the correct agent) should be left
on the bottle of volatile agent between fillings. Operating room pollution
is greater when the device is replaced by the screw-on cap between use.
P128
Volkwein JC, Vinson RP, Thimons ED. 1982. Effectiveness of three water spray
methods used to control dust during bagging. Washington, DC: U.S. Department
of the Interior. Report of Investigations, Bureau of Mines, 8614.
Three methods of water application to suppress dust during the bag-
ging of industrial sand were evaluated: injecting water into the bag while it
filled, atomized spraying the outside of the bag during filling and conveying,
and wetting the outside of the bag with a plain water mist during conveying.
Water injection and external wetting plus injection reduced average dust con-
centrations beside the bagging machine from 1.21 mg/m3 and dry conditions to
0.61 and 0.87 mg/m3, respectively.
P129
Vinson RP, Volkwein JC, Thimons ED. 1981. SF6 tracer gas tests of bagging
machine hood enclosures. Washington, DC: U.S. Department of the Interior.
Report of Investigations, Bureau of Mines, 8527.
SF6 tracer gas tests in specially designed hoods of machines for
bagging silica indicated that makeup air must be evenly dispersed, hood enclo-
sures and duct systems must be airtight and must be properly maintained, and
the average intake air velocity of the hood must be £ 200 fpm. The effective-
ness of the hoods during bagging at three facilities varied from no tracer
gas found outside the hood at one plant to significant SF6 concentrations in
the breathing zone of the operator at another plant.
P130
Breslin JA, Niewiadomski GE. 1982. Improving dust control technology for
U.S. mines: the bureau of mines respirable dust research program, 1969-82.
Washington, DC: Bureau of Mines, U.S. Department of the Interior.
A-30
-------
The title program has studied new and improved techniques and equip-
ment to control dust during mining operations such as cutting. Control of
coal mine dust by ventilation, conveyer screens, water sprays, dust collectors
or scrubbers is described. Silica dust emissions during silica bagging opera-
tions are controlled by equipping bagging machine with exhaust ventilation
hoods. Sampling and analysis methods are discussed.
P131
Ward P, Campbell D. 1983. Requirements definition for process control systems,
pp. 585-600. Proceedings of the ISA International Conference and Exhibit on
Advances in Instrumentation, Vol. 38, Part 1, Houston, TX, October 10-13, 1983.
Instrumental Society of America.
"Structured analysis" provides a well-developed framework for de-
scribing requirements for typical data processing systems. This paper de-
scribes an extension of structured analysis to describe process control system
requirements. The use of the technique is illustrated on a bottling system.
[Adapted from the author abstract.]
P132
Phillips CF, Jones RK. 1978. Gasoline vapor exposure during bulk handling
operations. American Industrial Hygiene Association Journal 39(2):118-128.
Gasoline vapor concentrations were determined in the breathing zone
of employees during bulk loading of gasoline tank trucks with the use of var-
ious types of loading and control systems. Top and bottom loading operations
with and without vapor recovery are described. Employee exposure to gasoline
vapor was lowest at the plant with top loading and vapor recovery (98% of the
samples contained £ 25 ppm gasoline). Full-time loaders are probably exposed
to ^ 85 ppm gasoline on 98% of the days during top loading without vapor
recovery.
P133
Thomson FM. 1977. Storage, feeding, and metering of fine powders, pp. 123-
129. Proceedings of the Workshop on Particle Technology-Research Needs,
Opportunities, and Priorities, Philadelphia, August 1975. Washington, DC:
National Science Foundation.
State-of-the-art design for powder storage bins and hoppers is dis-
cussed. Several areas needing additional research are pointed out. For
example, poor weight control and operator exposure to dust during package
filling are ascribed to a lack of knowledge about the effect of the gaseous
phase (air) on dry powder flow behavior.
P134
Midwest International. 1981. Dust eliminated during filling operation.
Midwest International, Inc., Charlevoix, MI. Chemical Engineering 88(13):
55-56.
A-31
-------
A new system for filling semibulk flexible containers improves oper-
ating rates, reduces manpower needs, and controls dust without separate pollu-
tion equipment. The system was designed for Pfizer, Inc.'s, Minerals, Pigments,
and Metals Division for filling containers with approximately 1,400 Ib precipi-
tated calcium carbonate powder of bulk density 17 to 25 lb/ft3. The system
comprises a cyclone separator and aerated storage bin, a dust suppression unit,
a hydraulic container support and lifting mechanism, and a scale with digital
readout.
P135
ACS. 1983. Production by the U.S. Chemical Industry. American Chemical
Society, Washington, DC. Chemical and Engineering News 61(24):28-34.
Statistics are given for U.S. production in 1982 and previous
years of the top 50 chemicals, fertilizer raw materials, and synthetics
(rubber, plastics, and fibers).
P136
Plant AF, ed. 1982. ACS. American Chemical Society, Washington, DC.
Chemcyclopedia 1982-83: Volume 1.
Thousands of commercially available chemicals are tabulated along
with information including physical forms in which the chemicals are available,
shipping quantities, and shipping containers. This version is more useful for
quick identification and enumeration of chemicals that are commonly drummed
or bagged than the 1983 edition, which is formatted in paragraphs for each
chemical.
P137
Freed JR, Chambers T, Christie WN, Carpenter CE. 1983. Methods for assessing
exposure to chemical substances, volume 2: methods for assessing exposure to
chemical substances in the ambient environment. Washington, DC: U.S. Environ-
mental Protection Agency, Office of Toxic Substances. EPA 560/5-83-015.
One of nine volumes, this volume describes environmental pathways,
sources of monitoring data, and the calculation of exposures by different
routes. (Models used to calculate inhalation exposures are discussed.) The
bulk of the volume comprises three appendices: (a) a compilation of pertinent
computerized data bases and printed sources; (b) lists of chemicals regulated
by EPA under the Resource, Conservation, and Recycling Act (RCRA), the Clean
Air Act, and the Clean Water Act and by OSHA, and lists of chemicals studied
by these agencies and NIOSH; and (c) emission characteristics of the organic
chemicals industry, the plastics industry, and the lubrication and hydraulic
fluids industry.
A-32
-------
P138
Adkins LC, Doria JJ, Christopher MT. 1983. Methods for assessing exposure
to chemical substances, volume 3: methods for assessing exposure from dis-
posal of chemical substances. Washington, DC: U.S. Environmental Protection
Agency, Office of Toxic Substances. EPA 560/5-83-016.
Environmental releases from chemical disposal are estimated in a
five-step process: (1) estimate releases to disposal; (2) characterize waste
stream releases and concentrations; (3) allocate waste streams to disposal
practices; (4) allocate waste streams to individual disposal sites; and (5)
estimate environmental releases from disposal sites. Disposal practices con-
sidered are landfilling, land treatment, holding in surface impoundments,
sewage treatment, incineration, and deep-well injection. Besides appendices
E through J containing auxiliary information on the disposal practices, other
appendices cover: (a) useful models and data bases; (b) information collected
by state solid waste agencies on waste generation, disposal practices, and
disposal facilities; (c) waste disposal practices of the petroleum refining
and organic chemicals industries; and (d) various kinds of information needed
for step (3) above.
P139
Dixon DA, Hendrickson GL, Jennings P, Chambers T, Borenstein A, Doria J,
Faha T. 1983. Methods for assessing exposure to chemical substances, volume
9: methods for enumerating and characterizing populations exposed to chemical
substances. Washington, DC: U.S. Environmental Protection Agency, Office of
Toxic Substances. EPA 560/5-83-022.
Methods for identifying, enumerating, and characterizing populations
exposed to chemicals are described. Populations considered are those exposed
in the ambient environment, in the occupational environment, via food, via
consumer products, and via drinking water. Appendix A applies the methods to
example problems for each exposure pathway. Appendix B gives examples of data
bases used in the occupational populations methods section. (No information
is presented on enumerating subgroups of a particular occupational population
such as packagers.)
P140
Versar. 1982. Appendix A, processes and exposure potential, methodology for
assessing occupational exposure. Versar, Inc., Springfield, VA. Prepared for
U.S. Environmental Protection Agency, Office of Toxic Substances. EPA Contract
No. 68-01-6271, Task 10.
Multimedia emission characteristics by unit process and by unit
operations are identified. Twenty-three large-volume SOCMI (Synthetic Organic
Chemicals Manufacturing Industry) unit process components are considered.
(These are basic synthesis steps such as halogenation and alkylation.) Each
unit process is described with available release data. Data are presented on
controlled and uncontrolled releases from process, storage and handling, and
fugitive sources to air, land, and water.
A-33
-------
P141
Versar. 1982. Appendix B, information resource matrix, methodology for
assessing occupational exposure to chemical substances. Versar, Inc.,
Springfield, VA. Prepared for U.S. Environmental Protection Agency, Office
of Toxic Substances, Washington, DC. EPA Contract No. 68-01-6271, Task 10.
This matrix is similar to the one in Appendix B of Freed et al.
(September 1983). Information resources included are numeric data bases,
bibliographic retrieval data banks, standard reference manuals, encyclopedias,
journals, and books. Photocopies of pages of source documents and of litera-
ture from the data base producers comprise much of the contents of this
document.
P142
Flores GH. 1983. Controlling exposure to alkylene oxides. Chemical Engi-
neering Progress 79(3):39-43.
Peak exposure levels are given and control measures are described
for reducing ethylene oxide and propylene oxide exposure during the following
activities: process sampling, tank car sampling, and tank car disconnect in
process areas; laboratory analysis and cylinder transfer; and maintenance.
P143
Dunn DW, Johnson ML, Hedley WH, Pate JB, Barrett GJ, McKinnery WN, Jr. 1983.
Control practices at formaldehyde production plants. Chemical Engineering
Progress 79(3):35-38.
Preliminary industrial hygiene/control technology surveys were done
at 11 formaldehyde production plants. The plants represent eight companies
and a cross-section of the industry on the basis of size, geography, and pro-
cess (catalyst type). Walkthrough surveys began with the methanol receiving
area, continued through production, and concluded with the formaldehyde stor-
age and shipment areas. Personal sampling revealed exposures from 0.01 to
2.4 ppm (|jg/m3) formaldehyde. Personal protective equipment is generally used
during the loading of formaldehyde--a top-fill operation using a dip tube into
the tank truck. Other plants use ventilation and scrubbing to protect the
workers during loading. Automatic loading devices are common.
P144
Tomb TF, Treattis HN. 1983. Review of published experimental calibrations
performed on the 10-millimeter nylon cyclone. Pittsburgh, PA: Mining En-
forcement and Safety Administration, Pittsburgh Technical Support Center.
MESA-IR-1040.
Research is reviewed on calibration of the 10-mm Dorr-Oliver nylon
cyclone to recommend an operational flow rate that will give size-selection
characteristics most closely approximating the Atomic Energy Commission (AEC)
A-34
-------
criterion for respirable dust. The experimental scatter precludes recommend-
ing any flow rate within a tolerance > 0.2 L/min. A flow rate of 2.0 L/min
during operation of the cyclone will closely approximate the AEC criterion.
P145
Peach MJ, Myers WR. 1981. Field research report on the performance of MSA
powered air purifying respirators on silica flour baggers at Illinois Mineral
Company, Elco, Illinois. Morgantown, WV: National Institute for Occupational
Safety and Health. NTIS No. PB83-258012.
The Mine Safety Appliances (MSA) Powered Air-Purifying Respirator
(PAPR) was evaluated in the laboratory and in the field when worn by two bag-
gers at the two-spout bagger and when placed in a fixed location minus the
facepiece. Size distributions of suspended particulates from the two-spout
bagger were determined in an area 4 to 8 ft from the baggers. Silica concen-
trations in air outside the facepiece of the respirators worn by the two bag-
gers ranged from 4.365 to 41.68/mg/m3 over a 3-day period. The maximum con-
centration observed in air inside the facepiece was 1.479 mg/m3; 5 of 11 air
samples from inside the mask exceeded the calculated threshold limit value of
0.291 mg/m3. The Illinois Mineral Company was advised to discontinue use of
the MSA PAPR units until the design problems allowing particulate penetration
inside the mask are corrected.
P146
Schroy JM. 1981. Prediction of workplace contaminant levels, pp. 190-206.
Symposium Proceedings on Control Technology in the Plastics and Resins Indus-
try, Atlanta, GA, February 27-28, 1979. Cincinnati, OH: National Institute
for Occupational Health and Safety. NIOSH Publication No. 81-107.
Vapor emission rate data are used to estimate volatile losses from
equipment, piping, and open pools of pure substances when monitoring data are
unavailable. The maximum concentration of a volatile chemical at some dis-
tance downwind from the leaking source is related to the emission rate, the
ventilation rate, diffusion coefficients, and the molecular weight of the
chemical of concern. The relations of fugitive emission rates of acryloni-
trile and vinyl chloride to workplace air concentrations are graphically
presented.
P147
Volkwein JC. 1979. Dust control in bagging operations, pp. 51-57. ACGIH
Symposium Proceedings on Industrial Hygiene for Mining and Tunneling, November
6-7, 1978. Cincinnati, OH: American Conference of Governmental Industrial
Hygienists.
The Bureau of Mines identified emission sources from fluidized-type
bag filling machines at silica sand processing mills and recommended control
measures. Dust was emitted while filling the bag. The Bureau of Mines devel-
oped an improved ventilation hood that encloses the bag and nozzle area and
A-35
-------
exhausts the dusty air to a dust collector. Dust from spillage of material
from nozzle and bag sleeve as the bag leaves the machine was controlled by
delaying release of the bag and then forcing material remaining in the nozzle
into the bag by a low-volume blast of compressed air. Other improvements were
needed. Dust adhering to the bag surface (which can become airborne during
conveying and loading) was not satisfactorily controlled by air brushing.
Dust emissions due to poor bag quality (inadequate closure, broken bags, and
leaky seams) could be improved by administrative control over bag quality.
Background dust levels from other dust operations were also important sources
of respirable dust exposure. The bagging operation should be isolated, lo-
cated far from other dusty operations, or fresh makeup air should be supplied
to the baggers.
P148
Bureau of Mines. 1978. Dust control hood for bag-filling machines. Technol-
ogy News No. 54, Technology Transfer Group. Washington, DC: Bureau of Mines,
U.S. Department of the Interior. Revised June 1983.
The Bureau of Mines hood for controlling silica sand dust during
bagging is described. A typical hood has 4 ft2 of open surface area and re-
quires an air velocity of 800 ftVmin. Preliminary tests showed dust concen-
trations by up to 90%; tests with SF6 tracer gas showed that the hoods can be
operated with 100% efficiency in reducing air contamination.
P150
Volkwein JC, Cecala AB, Thimons ED. 1983. Use of foam for dust control in
minerals processing. Washington, DC: U.S. Department of the Interior. Re-
port of Investigations, Bureau of Mines, 8808.
Total respirable dust reductions of 80 to 90% were achieved by mix-
ing foam (water plus surfactant) generated by compressed air with whole-grain
silica sand. Samples were taken under the screw conveyor, at the top of the
elevator, at the chute to the conveyer, and at the feeder.
P151
Volkwein JC, Thimons ED. 1983. Ventilation hoods to reduce dust during bag-
filling. Mining Equipment International.
The Bureau of Mines developed an improved ventilation hood for
fluidized valve-type bagging machines to capture silica sand dust close to
the generation source and to minimize the amount of air needed. The same
cut away diagram of the hood that appears in Bureau of Mines (1983) is ac-
companied by an exploded view of the hood in quarter sections.
P152
Bureau of Mines. Undated. Technical information for dust control ventilation
hood for bag picking machines. Draft report. Washington, DC: Bureau of Mines,
U.S. Department of the Interior.
A-36
-------
The Bureau of Mines hood developed for silica sand dust control dur-
ing bagging surrounds the filling bag so that exhaust ventilation air is drawn
through the open front of the hood, over the bag surface, and out through duct
work at the rear of the hood to a dust collection system. This reference con-
tains much of the same information as in Volkwein and Thimons (1983). Diagrams
show a typical dust hood installation and duct placement for a split type
weighing arm.
P154
O'Leary DT, Richter KM, Hi 11 is PA, Wood PH, Campbell SE. 1983. Methodology
for estimating environmental loadings from manufacture of synthetic organic
chemicals. Athens, GA: U.S. Environmental Protection Agency, Environmental
Research Laboratory. EPA-600/3-83-064, NTIS No. PB83-241331.
A methodology is presented for estimating the multimedia environ-
mental loadings for a "new" chemical, in the absence of manufacturing plant
emission data. The methodology draws on an environmental release data base
that contains multimedia environmental loadings for structurally similar com-
pounds that undergo similar process (physical and chemical) unit operations.
The data base is integrated with other pertinent available data on the manu-
facturing process of the new chemical such as (a) physical and chemical prop-
erties of process feedstock, products, and by-products; (b) reaction stoi-
chiometry, thermodynamics, and reaction kinetics; (c) process flow diagram
and process mass balance; (d) location and composition of environmental re-
leases and method of disposal; (e) process environmental control technology
including performance; (f) process storage and handling requirements; and (g)
plant equipment components (in numbers and classes). In practice, sufficient
direct data are rarely available for estimating the environmental loadings of
the compound under review; the methodology has been designed with this reality
in mind. In every case, where data deficiencies are likely to occur, alterna-
tive means are suggested for filling the data gaps. The methodology integrates
all pertinent data to enable the user to estimate multimedia (controlled and
uncontrolled) environmental loadings under the classifications of storage and
handling, process and fugitive emissions, respectively. An example is pro-
vided to demonstrate the methodology's applicability. [Author abstract]
P155
Callahan JM. 1978. Refueling emissions-oil versus auto co.'s. Automotive
Industries 158:14-15.
EPA was expected to require control of automotive refueling emis-
sions by fume collection onboard the car or at the gas station. For each
gallon of gasoline, 5.3 g vaporized hydrocarbons are released during refuel-
ing. The collector in the car would be a large canister containing charcoal.
Three oil companies have developed prototype automobile collectors. Fumes
collected at the gas station could be done by the method used in California.
Vapor is returned to the station's storage tanks through special pump nozzles
and hoses. The emissions are recycled by forcing them from the storage tank
to a tanker truck while it replenishes the station.
A-37
-------
P156
Cahners. 1983. FMC gains accuracy, speed with new drum, bagline. Cahners
Publishing Company, Denver, CO. Packaging Engineering 29:92-94.
Automated bag and drum filling of microcrystalline cellulose at FMC's
food and pharmaceutical products division is described.
P157
Schroy JM. 1985. Dynamic conditions in filling drums. Presented at the
1985 Summer National Meeting of the American Institute of Chemical Engineers,
August 1985, Seattle, WA.
Drum filling is described as a non-steady state process involving
both vapor diffusion and convective gas flow. The process of vessel filling
and consequent emission rate is defined in terms of the vaporization and
emission processes and their time dependent relationships. Large barge load-
ing tests are used to check a model for organic vapor losses from clean drum
filling.
P158
EPA. 1984. A manual for the preparation of engineering assessment.
Washington, DC: U.S. Environmental Protection Agency, Office of Toxic
Substances.
This manual is an aid for the chemical engineer in the Chemical
Engineering Branch who is responsible for evaluating occupational exposures
and environmental releases under the Toxic Substances Control Act.
P159
Versar, Inc. 1982. Methodology for assessing occupational exposure to chem-
ical substances. Draft final. Versar, Inc., Springfield, VA. Prepared for
U.S. Environmental Protection Agency, Office of Toxic Substances. Contract
No. 68-01-6271, Task 10.
This report describes methods of estimating,.workplace.concentrations
and identifies exposed populations.
P160
JRB Associates. 1984. Draft data element list for the comprehensive assess-
ment-information rule. JRB Associates, McLean, VA. Prepared for U.S. Environ-
mental Protection Agency, Washington, DC. JRB Contract 2-813-975-13, EPA
Contract 68-01-67C9.
This report identified federal information collection instruments
used in conducting chemical assessments. The purpose was to develop a com-
prehensive list of the data elements required to conduct chemical assessments
at EPA and other federal agencies.
A-38
-------
P161
Norwood SK. 1984. Estimating employee exposures from continuous air moni-
toring data. Dissertation submitted to the University of Oklahoma.
This study investigates the relationship between exposure estimates
based on personal samples and exposure estimates based on area samples taken
by a continuous air monitor (CAM).
VI. PACKAGING ASSOCIATION PHONE SURVEY APPROACH AND RESULTS
Various trade associations related to the packaging industry were
contacted for the packaging characterization study. The following type of
information was requested.
• Types of containers and sizes commonly used in chemical packaging
industry (packaging of semi bulk chemicals).
• Data on the types and sizes of containers produced in this coun-
try, and the percentage used in the chemical industry.
• Product loss during packaging operations using various types of
containers (this includes steel drums, fiber drums, plastic
pails, bags, etc.).
• Any information on emission factors from packaging processes
(same question as above).
• Other sources of data.
From this telephone survey, it was concluded that the information
sought was not readily available. The information which is available is
either strictly for association members, or it is not specific enough to
characterize packaging operations. In general, 55 gal steel, plastic, and
fiber drums; 5 gal plastic and steel pails; and 50 Ib multilayer kraft and
plastic bags comprise the majority of the packaging in the chemical industry.
No specific breakdown was found of what percentage of these packages are used
in the chemical industry or any other breakdown relating to the types of
chemicals (e.g., volatile, liquid, granular, fine powder, etc.). Product
loss in packaging operations varies from 0.1% to 1% of the production depend-
ing upon product value and other characteristics. If the product is expen-
sive, then more effort is put in to control the product loss during packaging
and in other operations. Table A-6 provides a summary of contacts and infor-
mation received.
A-39
-------
Table A-b. Summary of Contacts and Information Received
Name of association
contact person
phone number
Summary of conversation
*Fiber Drum Technical Council
Mr. Lawrence W. Bierlein
(202) 342-5250
Continental Fiber Drum
Mr. P. White, Sales
(203) 964-6717
Stanford, Connecticut
'"Packaging Institute
Ms. Carol Boyle
(212) 687-8874
Mr. Bob Temper
Mallinckrodt, St. Louis, Missouri
(314) 895-2057
He is a one man operation. They only handled technical aspects of
fiber drum industry. They do not keep any statistics. He referred
me to two companies which make up the largest segment of the fiber
drum industry. These companies are Continental Fiber Drum and
Grief Brothers.
Mr. White, Head of Sales, was not available. In talking with
staff members it was found that the information inquired is
very sensitive. They cannot disclose, especially considering a
very limited number of companies make up the whole industry.
The packages prevalently used in chemical industry are 55, 40,
and 30 gal. steel drums, 5, and 15 gal. steel and plastic pails,
100, 80, and 50 Ib multilayer craft and plastic bags, 18 kg
pails, 25 kg bags, 200 and 50 kg drums.
Ms. Carol Boyle was not available. I talked to a staff member.
They do not have the inquired information. He referred me to
Mr. Bob Temper of Mallinckrodt, who is the Chairman of the Chem-
ical Packaging Committee for the Packaging Institute.
He does not have the information and referred me to the Chairman
of the Bag Committee/Plastic Drum Institute and the Steel Shipping
Container Institute. According to him, product loss information
during filling of bagging operations would be hard to find. This
type of information is kept within the plant and only line super-
visors might keep records of this. He thinks maybe the Chemical
Manufacturer Association can help us.
(continued)
-------
Table A-6. (continued)
Name of association
contact person
phone number
Summary of conversation
Packaging Institute
Mr. Warner 0. Fleming (Du Pont)
Chairman - Bag Committee
(302) 774-0188
*Paper Shipping Sack Manufacturing Assoc.
Mr. Brent C. Dixon
(914) 723-6440
Packaging Institute
Mr. Dan Barber (Container Corporation
of America)
Chairman
Plastic Drums Committee
(302) 573-2605
*Steel Shipping Containers Institute
Mr. Mai Anderson
(201) 688-8750
His committee did not have the information. He refered me to the
•Paper Shipping Sack Manufacturing Assoc. According to him product
loss or "giveaway" during filling and bagging operations can run
from 0.1 to 1% depending upon the product. It will be very hard
to find any numbers on product loss and its relationship to sizes
and types of containers. Also, the more expensive the product,
the less product loss.
Their association keeps the types of statistics we are interested
in but they do not distribute this information outside of the
association, it is for their members use only. In general, he
said, valve bags are used more than open mouth bags and typical
sizes are 50 and 25 Ib bags. Again, weight depends upon the
density of the products.
He sent us his company's product literature. This covers a
majority of the type of containers used in the chemical industry.
His product manager will call me to discuss other information.
According to Mr. Anderson, 55 gal. drums are used largely in the
chemical industry. Other sizes in use are 30 and 16 gal. drums,
and 5 gal. pails. Department of Commerce, Bureau of Census, com-
piles various statistics on steel shipping containers using pro-
duction information from them into a M 34K report. The 1981
report shows about 17 million steel drums out of 41 million pro-
duced in this country were used in the chemical industry. This
number includes all sizes and varieties. For more information
he refered me to Mr. Malcolm Burnhart of Bureau of Census.
(continued)
-------
Table A-6. (concluded)
Name of association
contact person
phone number
Summary of conversation
U.S. Dept. of Commerce
Bureau of the Census
Mr. Malcolm Burnhart
(301) 763-2518
"Plastic Shipping Container Institute
(Plastic Pail*)
Mr. Robert Hultquist
(312) 368-4500
j^ Society of Plastic Industry
M Mr. Robert Sherman
(312) 297-6150
"Center for Packaging Education
Mr. Bob Goldburg
(212) 624-6157
"National Barrel and Drum Association
Ms. Pam Terry
(202) 296-8028
"Packaging Machinery Manufacturers Institute
Mr. Claude Breeden
(202) 347-3838
Bureau of the Census does not keep the specific information which
we are seeking. They put out. reports on total shipments of steel
pails and steel drums. Ho will send us the 1982 and 1983 reports.
These reports do not have any useful information for our purposes.
In his view, no governmental agency keeps the type of information
In which we are interested.
They do not keep the type of information we are looking for. He
refered me to Mr. Robert Sherman of the Society of Plastic
Industry.
The type of information which we seek is not kept by them.
They do not have the information but they can collect it for us
on a fee basis. He thinks the study may take about 6 months to 1 year.
Their members are recyclers of used steel drums. They usually
recondition about 45 to 50 million 55 gal. drums (about 2.5 times
the new steel drum production). According to her, based on Census
Bureau's 1980 statistics about 39% of steel drums are used in
the chemical industry.
Their members build machinery for the Packaging Industry. They
do not have any information on packaging containers emission
factors related to individual types of machinery.
"These names were obtained from the "Encyclopedia of Associations."
-------
APPENDIX B
REPORTS ON ADDITIONAL PLANT SURVEYS
-------
INTRODUCTION
Efforts to gain plant access for characterization of industrial
drumming and bagging operations were successful in gaining approvals to visit
five plants in the Kansas City area. At these facilities, the following
operations were observed by MRI personnel:
1. Manual top filling of 55-gal drums with industrial solvents.
2. Manual top filling of 5- to 55-gal drums with automotive
chemicals.
3. Bagging of detergent powders into sacks, cartons and drums.
4. Bagging of pharmaceutical feed mixes into 50-lb bags.
5. Manual top filling of open 55-gal drums with adhesives.
6. Bagging of Portland cement into 94-Ib bags.
Details of the five plant visits are presented below.
Date of Survey: November 5, 1984
Packaging Operation: Drumming and filling of automotive chemicals, fuel and
oil additives, cooling system cleaners, conditioners and sealers, and cleaning
compounds.
Filling Apparatus: Manual top filling of 5-, 30- and 55-gal drums from inter-
ior 250-gal tanks and exterior 6,000-gal tanks. Gravity filling and pump
filling through a 1-1/2 in. hose. Rate estimated at 20 gal/min. Detergent
powders filled in 1-lb sacks, 50-lb cartons and 400-Ib drums from overhead
auger into hopper for sacks or down chute for cartons and drums.
Worker Locations/Movement: Liquid filling of 6 cans at a time (up to 1 gal).
Lids put on by hand or moved to machine for sealing. Worker uses finger to
test when 55-gal drum filled. Worker monitors scale for filling of powder
into boxes or drums. Automatic operation of 1-lb bagger allows worker to
move away from bagging area.
Cycle Times: Liquids into 55-gal drums at *• 20 gal/min. Small batches of
< 20 drums. Powders boxed or drummed at ^ 50 Ib/min for large containers.
Ventilation/Control Devices: None for liquid filling. Problems exist mostly
because of no open doors or windows and because of the tank cleaning opera-
tion. Humidity is controlled in solid filling area. Filters used in air
circulation of powder filling room.
General Comments: Little ventilation of liquid filling area. Cleaning of
previously used tank (for 1 pt filling) caused heavy fumes.
B-l
-------
Date of Survey: November 5, 1984
Packaging Operation: Fifty-five gallon filling (also truck filling) of min-
eral spirits, toluene, xylene, isopropanol, acetone, methylethyl ketone in
well-ventilated open building (to be partly enclosed for winter).
Filling Apparatus: A 2 in. hose is connected to one of many faucets coming
from large outside tanks and is first flushed of previous chemical. Barrel
is moved into position, ground cable attached, scale zeroed, pump on, valve
on, barrel filled, previous barrel sealed, pump off, valve off.
Worker Locations/Movement: Single worker moves filling hose from one barrel
to next after shoving filled barrel down rollers. When pump is turned on and
hose valve is opened, worker seals previously filled barrel. Ground cable
attached to barrel rim during filling.
Cycle Times: One barrel filled every 45-47 s. (Toluene filled at one
barre 1/min.)
Ventilation/Control Devices: A 5 in. x 10 in. inlet draws air from the bung
area. The bottom edge of inlet is about 1-1/2 in. above the bung. Vapors
exhausted to atmosphere above building.
General Comments: When chemical is changed, hose is flushed into chute with
open lid. Hose end is placed here to drain between filling operations. In
winter, hot air draft is projected toward worker area from behind right rear
shoulder of worker; air inlet is on heater bottom.
Dates of Survey: November 6, 1984 and November 9, 1984
Packaging Operation: Bagging of pharmaceutical feed mixes into 50-lb bags.
Four or five bagging machines (one had no hood). Worming medication bagged;
also bagging of antibiotic products in rice hull or other carriers.
Filling Apparatus: Hopper discharges solid material down chute into bag
which drops slightly while filling, emitting puff of particulate. Bag moved
on conveyor to sealing operation.
Worker Locations/Movement: Worker sets filling time, trickle time for correct
weight, checks first bag, takes samples for analysis. Glass shields worker's
face from chute area.
Cycle Times: 7 to 8 bag/min (50-lb bags).
Ventilation/Control Devices: Horizontal inlet (rectangularly shaped) posi-
tioned to rear of drop chute.
General Comments: Observed bagging of 40% (by wt) antibiotic (dusty). Dump-
ing of material for blending is a dusty operation, done in three-sided enclo-
sure with overhead hood. Oil is blended into mixtures to control dust. Plant
has personnel dust sample data taken at least twice per year.
B-2
-------
Date of Survey: November 13, 1984
Packaging Operation: Manual top filling of creamy white liquid adhesive into
open 55-gal drums. Viscosity of 1,500 cp.
Filling Apparatus: Large 65-barrel batch is cooked, then gravity fed through
3 in. hose to pump system with two long bag filters, then through 2 in. hose
to open 55-gal drum lined with poly bag. Approximately 15 drums arranged in
semicircle around pump. No splattering or barrel bubbling seen.
Worker Locations/Movement: Worker moves hose from barrel to barrel. New
drum is tipped under rim of drum currently being filled before nozzle is
shifted. No spillage seen even though pump not shut off between barrels.
No shutoff valve on nozzle end.
Cycle Times: About 85 s to fill one 55-gal drum. While drum is being filled,
previous drum is covered with plastic, lidded and sealed using a rim ring
which is tightened with its own lever (60-drum batches).
Ventilation/Control Devices: No windows or door seen. Fans (exhaust?) are
located over each large heated mixing tank.
General Comments: When changing adhesives, large tank and pump are cleaned
with hot water flush. When pump flow decreases, worker unbolts pump filter,
removes long filter bags and discards in open cardboard barrel. Another
worker removes filled barrels with a hand truck and replaces with empty bar-
rels. Only slight smell discernable.
Date of Survey: November 13, 1984
Packaging Operation: Four-spout St. Regis packer for 94-Ib bags of Portland
cement.
Filling Apparatus: Cement is gravity-fed from a hopper into paper-plastic
bag after bag inserted onto outlet. When bag is filled, worker releases bag
which falls back on large mesh conveyor for delivery to automatic palletizer.
Worker Locations/Movement: Worker sits on horizontally movable bench and
places bags in succession on four hopper spouts. Button above each outlet
controls start of cement flow. Foot control releases bag to fall back on
conveyor. Valved bag is self sealing.
Cycle Times: Estimated at about 10-15 bags/min.
Ventilation/Control Devices: Hood on packer to bag house controls cement
dust released during bag filling, but does not control conveyor emissions.
Filling operator wears space helment connected to air supply.
General Comments: Conveyor is meshed, allowing spilled cement to be collected
underneath by screw conveyor. Major points of emissions are conveyor drop
points and where conveyor moves around corners. Lots of doors open in summer.
One large door open during survey.
B-3
-------
APPENDIX C
LABORATORY PILOT STUDY
-------
APPENDIX C
TABLE OF CONTENTS
Page
I. Introduction C-l
II. Experimental Facilities and Apparatus C-l
A. Flow Tunnel C-l
B. Plume Generation Apparatus C-3
C. Sample Collection Apparatus C-3
D. Wind Speed Monitors C-ll
E. Photography C-ll
III. Experimental Methods C-ll
A. Sampling Procedures C-ll
B. Analysis Procedures C-14
C. Data Reduction and Analysis Procedure C-15
IV. Results C-16
A. Methane Experiments C-16
B. Ammonium Chloride Experiments C-16
V. Model Development C-16
A. Dispersion Coefficients Calculated from Methane
Experiments C-19
B. Particulate Dispersion Results C-30
VI. Procedures for Calculation of Maximum Concentrations. . . . C-32
VII. References C-38
C-i
-------
APPENDIX C
LIST OF FIGURES
Figure Title Page
II-l Overhead view of pull-through flow tunnel C-2
II-2 Overhead view of release configurations C-4
II-3 Bung releases from 55-gal drums (Configurations 2 and 3). . C-5
II-4 Ammonium chloride plume generation apparatus C-6
II-5 Sampling arrays for methane dispersion experiments C-8
II-6 Sampling arrays for ammonium chloride dispersion experi-
ments—tunnel centerline releases C-9
I1-7 Sampling arrays for ammonium chloride dispersion experi-
ments—tunnel side releases C-10
II-8 30-Second time exposure of ammonium chloride dispersion
at 0.5 m/s (1.1 mph) C-12
II-9 30-Second time exposure of ammonium chloride dispersion
at 1.2 m/s (2.6 mph) C-13
V-l Comparison of measured versus circular normal modeled
concentrations for Experiment 7 C-21
V-2 Comparison of measured versus circular normal modeled
concentrations for Experiment 8 C-22
V-3 Comparison of measured versus circular normal modeled
concentrations for Experiment 9 C-23
V-4 Comparison of measured versus single-point circular-
normal concentrations for Test 22-A (4.8 mph wind
speed and 2 m downwind from the source) C-24
V-5 Comparison of measured versus circular normal modeled
concentrations for Test 22-B (4.7 mph wind speed and
4 m downwind from the source) C-25
V-6 Comparison of measured versus circular normal modeled
concentrations for Test 22-C (4.7 mph wind speed and
4 m downwind from the source) C-26
V-7 Comparison of measured versus modeled concentrations at
1.05 m above tunnel floor for Experiment 22 (4.8 mph) . . C-27
V-8 Comparison of measured maximum concentrations (hooded
drum tests) with two models for the maximum values. . . . C-31
V-9 Comparison of photographically determined vertical spread
of ammonium chloride plume with Turner dispersion
coefficients C-33
VI-1 Methane concentrations above background level and with
blank criterion applied C-37
C-iii
-------
APPENDIX C
LIST OF TABLES
Table Title Page
II-l Ammonium Chloride Particle Size Fractions C-7
IV-1 Experimental Conditions for Laboratory Pilot Study of
Methane Dispersion C-17
IV-2 Experimental Conditions for Laboratory Pilot Study of
Ammonium Chloride C-18
V-l Empirically Determined Dispersion Coefficients Determined
from Methane Experiments C-20
V-2 Comparison of Measured and Modeled Concentrations of
Methane C-29
V-3 Comparison of Measured and Modeled Maximum Concentrations
of Ammonium Chloride C-32
VI-1 Diffusion Function Reported by Schroy (1981) C-36
C-v
-------
I. INTRODUCTION
This section describes the initial laboratory pilot study performed
to test dispersion theories and to develop efficient sampling strategies for
use in the plant scale occupational exposure study. The laboratory pilot
study utilized a walk-in, low-speed flow tunnel which was constructed at MRI's
Deramus Field Station.
Dispersion patterns of gases and suspended solids were studied for
single and multipoint sources released into free flow and obstructed flow en-
vironments. Twenty-five experiments of gaseous dispersion and four experi-
ments of particulate dispersion were conducted. The observed dispersion pat-
terns were compared with existing theories and then used as the basis for a
modified dispersion algorithm.
The presentation is organized by subject area as follows:
• Section II describes the experimental facilities and apparatus
used to characterize dispersion of packaging emission simulants
at low flow speeds.
• Section III presents the experimental procedures for emission
generation and for sample collection and analysis. Section III
also describes techniques for data reduction and analysis to
quantify tunnel flow patterns, simulant concentration distribu-
tions, and empirical dispersion coefficients.
• Section IV summarizes the dispersion experiments and presents
the experimental results.
• Section V describes the preliminary dispersion model developed
in this study and compares measured concentrations with the
predictions of this model and of other available models.
• Section VI presents calculation procedures for the simple mass
balance model, the Schroy (1981) model, and the circular normal
model.
II. EXPERIMENTAL FACILITIES AND APPARATUS
This section describes the experimental facilities and apparatus
used to study dispersion of gases and suspended solids.
A. Flow Tunnel (see Figure II-l)
A pull-through flow tunnel (2.1 m in height x 2.7 m in width x
9.8 m in length) was constructed in MRI Field Station Building No. 1 to
gather basic data on dispersion patterns as a function of release point con-
figuration and flow velocity approaching the release point(s). Building
No. 1 is 8.5 m x 15.2 m in length with a 3 m wall height. A 2.1 m x 2.7 m
garage door opening centered on the west wall of the building was used for
the wind tunnel entrance. The building exhaust fan drew air through the tun-
nel at mean flow velocities from 0.5 to 2.2 m/s (0.9 to 4.9 mph).
C-l
-------
I
ro
Reducing section at tunnel exit was-
added after methane experiments
•1 m
1.3m
Figure II-l. Overhead view of pull-through flow tunnel.
-------
One side wall and the tunnel ceiling were constructed of 6 mil
plastic sheeting mounted on a 2 in. x 4 in. wood framework. The other wall
of plyboard was painted flat black for both visual and photographic monitor-
ing of particulate plume dispersion. In addition, a plexiglass window was
framed in the tunnel wall facing the black plyboard to allow clear viewing of
the tunnel interior during an experiment. For photographic work, No. 8 white
thread was vertically suspended down the tunnel centerline at 10 cm dis-
tances.
In order to reduce the effects of external wind on the flow through
the tunnel, a special tunnel entrance and exit were built. A 2.1 m x 2.7 m x
0.6 m flow straightener constrained ambient winds entering the tunnel. In
addition, a lattice work was placed in front of the straightener to improve
the uniformity of low wind flows by reducing the propagation of external at-
mospheric turbulence into the tunnel. Also a reducing section was attached
to the exit of the tunnel so that the flow was drawn through a centrally lo-
cated exhaust area of size 1.1 m x 1.4 m.
B. Plume Generation Apparatus
For gaseous dispersion experiments, natural gas was released at a
monitored rate (Matheson 602 rotameter) from a Fisher burner to enable thor-
ough mixing of the gas with air. The single point release configurations in-
cluded: tunnel centerline point release into unobstructed flow, point re-
lease near a wall, point release past a sidedraft hood, and bung release from
a 55 gal drum with and without a close-fitting hood. Figure II-2 shows an
overhead view of the eight release configurations. Figure II-3 illustrates
the two bung releases (Configurations 2 and 3 in Figure II-2). The double
point releases were used to check on the additivity of plumes.
Ammonium chloride was the simulant for particulate dispersion
modeling. This compound was formed by bubbling air through two Smith-
Greenburg impingers containing concentrated HC1 and NH4OH solutions, and by
combining the resultant vapor flow streams. The two nozzles which discharged
vapors from the chemical solutions into the tunnel were positioned 3/8 in.
apart and faced each other for good mixing of the vapors and consequent pro-
duction of NH4C1 (see Figure II-4). The airflow rates through the impingers
were monitored with Matheson (602 and 604) rotameters. The white ammonium
chloride particles could easily be seen to disperse downwind in the tunnel.
Particles of ammonium chloride formed in this manner were mostly smaller than
10 urn in aerodynamic diameter, considered the upper limit for respirable par-
ticles. Table II-l presents the size distribution of ammonium chloride par-
ticles as measured by a Climet particle analyzer. Approximately 63% of the
mass consisted of particles less than 0.5 pm diameter. Particles of this
size range disperse in much the same manner as gases.
C. Sample Collection Apparatus
1. Natural Gas
For the sampling of natural gas in air, polypropylene sample lines
were located at several fixed positions on a movable two-dimensional sampling
C-3
-------
Tunnel'
Wall '
O
'Side Drafti
"Booth ;
.-** _«v-. . . • ^,-*i
¥
Figure II-2. Overhead view of release configurations.
C-4
-------
Gas Source
Free Standing Drum
Gas Source
Drum with "Local Hood"
Figure II-3. Bung releases from 55-gal drums
(Configurations 2 and 3).
C-5
-------
Figure I1-4. Ammonium chloride plume generation apparatus.
C-6
-------
Table II-l. Ammonium Chloride Particle
Size Fractions
Average no. Percent .
Particle of particles total mass
diameter (urn) samples
0.3
0.5
1.0
3.0
5.0
10.0
- 0.5
- 1.0
- 3.0
- 5.0
- 10.0
- > 10.0
272,000
10,700
3,080
374
96
31
62.6
6.8
7.9
8.6
6.1
7.9
Determined by Climet particle analyzer at
.7.1 L/min.
Calculated based on assumed spherical shape
of particles.
array (Figure II-5) and routed to a common manifold connected to the analyti-
cal instrument. During sampling, each line was opened in sequence by a valv-
ing system controlled by a 24 pole rotary switch connected to a 30 s timer.
This sampling method allowed for rapid switching from sample point to sample
point every 30 s.
A Beckman 402 total hydrocarbon analyzer was used to determine nat-
ural gas concentrations in air. This unit utilizes a flame ionization detec-
tor (FID) to quantify total hydrocarbon (THC) levels down to 1 ppm. The sam-
pling rate was approximately 3 L/min. An EPSON HX-20 microcomputer and an
interface built around a Wintek microprocessor produced printouts of inte-
grated 10 s methane concentrations.
2. Ammonium Chloride
Respirable particles of ammonium chloride were collected using a
network of standard respirable dust samplers (37 mm filter cassettes with
Dorr-Oliver 10 mm nylon cyclone preseparators). The cyclones trapped parti-
cles larger than about 10 urn aerodynamic diameter. Each particulate sample
was collected on a Millipore® filter with 0.2 urn pores. A standard flow rate
of 2.0 L/min through the filter cartridge was ensured throughout substrate
buildup by maintaining sonic flow through a critical orifice. A pressure drop
exceeding 0.53 x 760 mm Hg through a 20 gauge needle established the required
flow. Separate sampling arrays were used for center and side releases, as
shown in Figures I1-6 and I1-7.
C-7
-------
o
00
• Source
O Sampler
^> Wind Flow
-f- Anemomaler
Figure II-5. Sampling arrays for methane dispersion experiments.
-------
%-V Wind Flow
I Anemometer
Figure II-6. Sampling arrays for ammonium chloride dispersion experiments--
tunnel centerline releases.
-------
o
o
ler
Wind Flow
~]~ Anemometer
Figure II-7. Sampling arrays for ammonium chloride dispersion experiments—
tunnel side releases.
-------
D. Wind Speed Monitors
Kurz warm wire anemometers were used to monitor the tunnel flow.
Tunnel centerline wind speeds were monitored at 1 m downwind from the en-
trance of the tunnel and 0.46 m upwind of the sampling plane for both gaseous
and particulate dispersion tests. Previously established wind speed profiles
were used to construct flow distributions from the mean centerline velocity
measured during a given test. Data from the anemometers were presented on
both the Epson HX-20 and the Kurz digital wind speed display panel. The
Epson computer also calculated and listed mean wind speeds for user-selected
time intervals.
E. Photography
Ammonium chloride plume dispersion patterns were photographically
recorded with a single lens reflex camera at different wind speeds and for
different release arrays. Time exposures of 35 mm black and white film (ASA
32) effectively documented the time-averaged plume shape for a given set of
experimental conditions (Figures II-8 and II-9).
III. EXPERIMENTAL METHODS
A. Sampling Procedures
1. Methane
Prior to each day of natural gas dispersion experiments, the
Beckman 402 was heated for 1 h, then calibrated with a reference gas. In the
absence of a 0 ppm THC gas, the zero end of scale was determined by blocking
all input to the Beckman. The Fisher burner was adjusted for maximum mixing
of natural gas with air. A rotameter between the gas source and the Fisher
burner was used to adjust and monitor the gas release rate.
Before each test the tunnel flow was set for the desired centerline
wind speed. Wind speeds from the warm wire anemometers were monitored on
Kurz digital displays and simultaneously on the Epson HX-20. Once the sam-
pling array was in the proper position, testing began with activation of the
rotary switching device and starting of the Epson HX-20 "TEST" program.
Three separate tests, each with the sampling plane at different distances
downwind of the source (5, 7, and 9 m from the tunnel entrance), constituted
one dispersion experiment.
During testing, each sampling tube on the array supplied a sampled
air/methane mixture to the Beckman 402 analyzer for a period of 30 s. The
THC concentration from the first 10 s, as the sampling tube was flushed and
the Beckman adjusted to the new THC value, was discarded. Two 10 s readings
(in ppm v/v) were then obtained before the unit automatically switched to a
new sampling position.
C-ll
-------
o
I
ro
Figure II-8. 30-Second time exposure of ammonium chloride dispersion at
0.5 m/s (1.1 mph).
-------
o
OJ
Figure II-9. 30-Second time exposure of ammonium chloride dispersion at
1.2 m/s (2.6 mph).
-------
2. Ammonium Chloride
Before each test the sampler holders were fitted with new filter
cartridges. With the vacuum pump turned on momentarily, each cartridge was
checked for a 2.0 L/min flow rate using a GCA RDM-101 rotameter. If a car-
tridge failed to meet a ±10% flow rate tolerance, that filter (usually in a
cracked cartridge) was designated a blank, and a new cartridge assembly in-
stalled. (Other blanks were also selected, as necessary, for test quality
assurance.) Then a leak check was performed on the entire system using a
mercury manometer.
Before each test of approximately 1 to 1-1/2 h duration, fresh
solutions of NH4OH and HC1 were cooled in an ice bath surrounding the two
Smith-Greenburg impingers. The production rate of NH4C1 was adjusted using a
bypass valve on the air pump supplying the impingers and by separately ad-
justing the airflow to each impinger. After the tunnel flow was set to the
desired centerline wind speed, testing was initiated with the activation of
the air pump. Three separate tests, with the sampling planes at 5, 6.5, and
8.0 m from the tunnel entrance, constituted one dispersion experiment.
During a test, the generation rate of ammonium chloride tended to
decrease with time. Consequently, the bypass valve on the air pump supplying
the impingers was adjusted to allow for a relatively constant mass rate of
particulate formed. The production of heat from ammonium chloride formation
was also diminished at lower generation rates. As a test progressed and the
chemicals in the impingers became less concentrated, the pump bypass valve
was opened to increase ammonium chloride production.
During a test, the ammonium chloride which agglomerated on the noz-
zles was removed at regular intervals using a water spray. The downwind
anemometer, which was exposed to NH4C1 particulate, was also washed at regu-
lar intervals with distilled water.
A test was aborted if tunnel wind speeds were subject to signifi-
cant ambient wind variability. Consequently, all particulate testing was
done during evening hours when calm ambient winds prevailed.
Black and white photographs of plume dispersion were taken with a
35 mm SLR camera positioned on the side wall at 1.06 m above the floor and
1 m downwind from the release point. The field of view from this point was
about 1 m at the tunnel centerline. Later a Plexiglas window was constructed,
and the camera position moved both away from the tunnel and downwind to allow
a field of view of 1.5 m at the tunnel centerline.
B. Analysis Procedures
1. Methane
Methane concentrations (10 s averages) measured by the Beckman 402
instrument were printed by the Epson HX-20 during each test. Methane concen-
trations were used directly for calculation of gaseous dispersion coeffi-
cients after subtraction of background values and elimination of values not
meeting a blank criterion.
C-14
-------
2. Ammonium Chloride
At the end of each participate dispersion test, the cartridges with
the Millipore® filters were removed from the sampling array, stoppers re-
placed on both the inlets and outlets, and the cartridges returned to the
laboratory in a protective container. After every second test, the cyclones
were washed with distilled water and air dried.
The chloride analysis of each particulate sample was initiated by
removing the filter from the cartridge and placing the filter (exposed side
up) in a bottle with a diameter slightly larger than the filter. Ten milli-
liters of H20 was added. The bottle was capped and then placed in an ultra-
sonic bath for 5 min. Two milliliters from each bottle was transferred to a
disposable sample cap for analysis using the Technicon Autoanalyzer. A sec-
ond 2 ml aliquot was taken from each filter bottle for replicate analysis.
After analysis of every tenth sample, a distilled H20 blank was
analyzed. Every 2 h standard chloride solutions (100 and 200 ug/mL) were
analyzed. If high concentrations forced the analyzer off scale, a diluted
sample of the filter solution was prepared and analyzed. The technique de-
scribed above is commonly used for water quality analysis and will detect
chloride ion down to 4 ug.
C. Data Reduction and Analysis Procedure
1. Tunnel Flow Patterns
Ten-second wind speed averages from the upwind anemometer were av-
eraged to characterize the overall centerline tunnel flow for each natural
gas test. One-minute wind speed averages were averaged over an entire par-
ticulate dispersion test period. This single test wind speed value was ad-
justed to sampling plane wind speeds using wind speed traverse factors.
These factors correlated centerline wind speeds measured at a point 1 or 2 m
from the tunnel entrance to wind speed values at each of nine points in each
sampling plane.
2. THC Concentrations
THC concentrations were input into a BASIC computer program for av-
eraging of the 4 to 12 replicate THC values obtained at each sampling point
during a test. These final THC concentrations were then plotted by the com-
puter on a representation of the array grid. The THC concentration data
plotted for each test included raw THC averages, standard deviations, and
numbers of values. Three arrays of THC averages were plotted—raw values,
values with a blank correction applied to them, and values from which the
background had been removed. The blank correction consisted of excluding THC
values of less than 5 times the standard deviation of blank values.
3. Chloride Concentrations
The Technicon Autoanalyzer was used to determine NH4C1 (as Cl")
present in the filter solution. The NH4C1 concentrations were plotted on a
representation of the sampling array for analysis of the plume shape.
C-15
-------
IV. RESULTS
This section identifies and describes the dispersion tests per-
formed both with methane gas and ammonium chloride particulate.
A. Methane Experiments
Table IV-1 lists the testing conditions for Experiments 1 through
26. Methane was the dispersion simulant for all except the last experiment,
which was performed with propane. Figure II-2 illustrates the release con-
figurations described in Table IV-1. Tunnel centerline wind speeds measured
1 m from the tunnel entrance ranged from 0.42 to 2.2 m/s (0.93 to 4.9 mph).
B. Ammonium Chloride Experiments
Ammonium chloride particulate was released from a single point
source in Experiments 27 through 30. Each of the three tests comprising a
single experiment are listed in Table IV-2 together with respective experi-
mental conditions. Dispersion was studied for wind speeds in the range of
0.50 to 1.6 m/s (1.1 to 3.7 mph) and for both tunnel centerline releases and
tunnel side releases.
V. MODEL DEVELOPMENT
The model proposed for the simplest release patterns in the current
program is based on the bivariate Gaussian distribution of concentration that
is commonly used in modeling of outdoor air pollution sources. The general
form of this distribution is fairly complicated, but several simplifying as-
sumptions have been made for the present application. First, it is assumed
that the distribution may be written as the product of a function of the lat-
eral crosswind distance times a function of the vertical crosswind distance.
This is the form taken by the distribution when applied to outdoor sources.
The second assumption is based upon the configuration of the test
chamber. Because the chamber has roughly a square cross-section, it was as-
sumed that the lateral and vertical dispersion coefficients are identical.
Under these assumptions, the model takes the form:
C(y,z) = A exp {- £ [(y-yQ)2 + (z-zQ)2]} (1)
where: C = modeled concentration (ppm v/v)
A = maximum modeled concentration (ppm v/v)
a = dispersion coefficient (m)
y,z = lateral and vertical distances from chamber centerline (m)
V0»Z0 = location of maximum modeled concentration (m)
C-16
-------
Table IV-1. Experimental Conditions for Laboratory Pilot Study
of Methane Dispersion
Date Experiment Release Nominal tunnel
no. configuration centerline velocity
5/30/84
5/29/84
5/29/84
5/30/84
5/30/84
6/4/84
6/5/84
6/5/84
6/5/84
6/5/84
6/5/84
6/5/84
5/30/84
5/30/84
5/30/84
6/5/84
6/4/84
6/4/84
6/4/84
6/4/84
6/4/84
6/6/84
6/6/84
6/6/84
6/27/84
1
2
3
4
5
6
7
8
9
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26a
Single point centerline release
Single point centerline release
Single point centerline release
Drum centerline release
Drum centerline release
Drum centerline release
Stacked drum centerline release
Stacked drum centerline release
Stacked drum centerline release
Single point release with in-
serted side-draft hood
Single point release with in-
serted side-draft hood
Single point release with in-
serted side-draft hood
Single point side release
Single point side release
Single point side release
Double point centerline release
Double point centerline release
Double point centerline release
Double point side-by-side release
Double point side-by-side release
Double point side-by-side release
Single point release with flush
side-draft hood
Single point release with flush
side- draft hood
Single point release with flush
side-draft hood
Single point centerline release
(m/s)
0.42
0.99
1.7
0.44
0.77
2.2
0.50
0.87
2.1
0.50
0.87
2.1
0.43
0.79
1.6
0.52
1.0
2.1
0.53
1.1
2.1
0.50
0.90
2.0
0.8
(mph)
0.93
2.2
3.8
0.99
1.7
4.9
1.1
1.9
4.7
1.1
1.9
4.7
0.97
1.8
3.6
1.2
2.3
4.6
1.2
2.4
4.7
1.1
2.0
4.5
1.8
Test aborted; wind speeds ranged from 1.3 to 2.1 mph.
C-17
-------
Table IV-2. Experimental Conditions for Laboratory Pilot Study
of Ammonium Chloride
Date
Experiment Release
no." configuration
7/18/84
7/18/84
7/24/84
7/19/84
7/23/84
7/19/84
8/2/84
7/31/84
7/31/84
7/30/84
7/30/84
7/30/84
27A
27B
27C
28A
28B
28C
29A
298
29C
30A
30B
30C
Single point center line release
Single point centerline release
Single point centerline release
Single point centerline release
Single point centerline release
Single point centerline release
Single point side release
Single point side release
Single point side release
Single point side release
Single point side release
Single point side release
Nominal tunnel
centerline
(m/s)
. 0.65
0.63
0.50
1.6
1.5
1.4
0.55
0.56
0.52
1.6
1.6
1.6
velocity
(mph)
1.4
1.4
1.1
3.6
3.4
3.2
1.2
1.2
1.2
3.5
3.7
3.7
Position A was 5 m into tunnel; B was 6.5 m into tunnel; C was 8 m into
tunnel.
Other models are also currently under consideration; especially, for
more complicated release configurations (such as the side draft hood). These
models include a Gaussian distribution similar to the above, but w'ith differ-
ent dispersion characteristics in the lateral and vertical directions (which
has been found useful in the preliminary analysis of the particulate tests),
and a distribution represented by an elliptical paraboloid.
The nonlinear normal equations resulting from a least-squares cri-
terion applied to Eq. 1 (after taking the logarithm of each side) can be
linearized using the following change of variables:
In A =
I/a2 =
4a,
a2
a3/2a2
a4/2a2
C-18
-------
With this change of variables, the normal equations become:
Nr I(y2+z2) Sy iz
I(y2+Z2)2 Z(y3+Z2y) Z(y2Z+Z3)
Iy2 lyz
"Symmetric iz2
where: N = number of data points
y,z = lateral and vertical distances (m) from tunnel centerline,
respectively
C = measured concentration (ppm v/v) at the point y,z
The modified normal equations are easily solved for each location and each
experiment using a Gauss-Jordan elimination algorithm for simultaneous linear
equations. Results are presented in the following sections.
A. Dispersion Coefficients Calculated from Methane Experiments
Experimentally determined concentrations were input into a BASIC
computer program for determination of dispersion coefficients at 2, 4, and
6 m from the source(s).
The circular bivariate normal form was applied to the symmetric re-
lease configurations (such as the single point source and the drum with a
hood). Experimentally determined dispersion coefficients for the methane
tests are reported in Table V-l. Figures V-l through V-3 compare the model
results for Experiments 7 through 9 (involving the hooded drum) with the mea-
sured concentrations.
Results from the simple release configurations can be used to es-
timate the concentration profile for more complicated configurations. For
example, Figures V-4 through V-6 compare the measured concentrations for Ex-
periment 22 (double point release) with the superposition of the single point
model developed on the basis of Experiment 3. Figure V-7 presents a view of
concentrations measured in Experiment 22.
Because the filling of 55 gal drums with a close fitting hood/fill
mechanism may be fairly common for liquids emitting organic vapors, the fol-
lowing discussion will be limited to the model results for Experiments 7
through 9.
In order to estimate the spatial variation of methane concentration
at downwind stations other than 2, 4, and 6 m (see Figures V-l through V-3),
the regression coefficients obtained for each downwind array were themselves
fit to power law models with downwind distance (x) and wind speed (U) as the
independent variables.
C-19
-------
Table V-l.
Empirically Determined Dispersion Coefficients
Determined from Methane Experiments
Expt.
1
2
3
4
5
6
7
8
9
17
18
19
Distance
from
source
(m)
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
Maximum concentration,
Cmax (pPm v/v)
Measured
100
44
21
55
19
15
53
20
11
55
24
14
55
22
17
18
29
-
20
7.8
7.2
18
7.6
4.5
15
5.1
5.1
88
62
31
110
33
18
88
27
16
Modeled
—
28
13
36
16
-
47
16
7.5
81
*
*
-
27
*
21
*
-
15
7.7
6.4
13.5
5.7
4.4
18.5
7.1
4.8
40.7
49.1
28.3
100
7.3
13.3
70.1
10.4
13.2
I/O2
(m-2)
.
2.4
0.41
11
1.3
-
17
3.9
1.3
9.1
ik
*
-
7.6
*
6.8
*
-
1.6
1.3
0.19
3.4
2.0
1.2
8.8
4.4
2.3
1.8
4.2
1.6
35
3.8
4.4
14
7.7
7.2
Location of max .
concentration (m)
yo
.
-0.063
-0.18
-0.032
-0.10
-
-0.036
-0.15
-0.15
-0.12
*
*
-
-0.055
*
-0.20
*
-
-0.095
-0.15
-0.48
-0.058
0.0063
0.047
-0.12
-0.21
-0.035
0.056
-0.012
0.037
-0.037
0.083
0.19
-0.001
-0.10
-0.17
Z0
.
0.15
-0.12
-0.040
0.076
-
0.065
0.074
0.13
-0.55
*
*
-
-0.70
*
-0.52
*
-
-0.15
-0.58
-1.6
-0.12
-0.22
-0.33
-0.29
-0.26
-0.38
-0.15
0.19
-0.11
0.020
0.039
-0.038
0.22
-0.013
0.041
Imaginary dispersion coefficent.
- Not enough valid data to model.
Using the bivariate Gaussian model described in Section VI for the
.simple release patterns.
Maximum concentration locations are offset distances from tunnel
centerline.
C-20
-------
2m
2m
2m
Drum
and Hood
2.7m
10 20
Plan View
0 10 20
ppm
0 Measured Concentrations
— Circular Normal Modeled
Concentrations
Figure V-l. Comparison of measured versus circular normal modeled
concentrations for Experiment 7.
C-21
-------
2 m:
m
•2 m v-
Drum
and Hood
2.7 ni'
et --
o
, O
Plan View
'A J!°L
ppm]
0 Measured Concentrations
— Circular Normal Modeled
Concentrations.
Figure V-2. Comparison of measured versus circular normal modeled
concentrations for Experiment 8.
C-22
-------
2m
2m \-
2m
Drum
and Hood
2.7 m
Plan View
0 5 10
ppm
Measured Concentrations
Circular Normal Modeled
Concentrations
Figure V-3. Comparison of measured versus circular normal modeled
concentrations for Experiment 9.
C-23
-------
Wall:
-Wall'
t
——— Measured Concentration
Single -Point Modeled Concentration
Figure V-4. Comparison of measured versus single-point circular-normal
concentrations for Test 22-A (4.8 mph wind speed and
2 m downwind from the source).
C-24
-------
-------
-------
Wall-
2 m Downwind
4 m
• Wall
I _t
Release Points
Figure V-7. Comparison of measured versus modeled concentrations
at 1.05 m above tunnel floor for Experiment 22. (4.8 mph)
C-27
-------
The correlation matrices are:
Ax U
A 1 -0.95 -0.036
x 10
U 1
for the maximum modeled concentration A, and:
a
X
U
a
1
X
0.58
1
U
-0.72
0
1
for the dispersion coefficient a. From these matrices, the downwind distance
appears to have the greater influence on maximum concentration while wind
speed is apparently more important in determining the dispersion. The power
fits were found to be:
A = 32 x"1-0
a = 0.48 x°-65 u"°-55
where the caret indicates the estimated value. Thus, Eq. (1) as applied to
the hooded drum Experiments 7 through 9 may then be written as:
C = 32 x"1-0 exp
'— V * i')'
(2)
where: C = modeled concentration (ppm v/v)
x = downwind distance (m)
U = wind speed (mph)
y,z = horizontal and vertical distances (m) from plume centerline,
respectively
Table V-2 compares the maximum concentrations measured during the
simplest methane release tests with: (a) the maximum values modeled by
Eq. (1); (b) the maximum values estimated by the Schroy (1981) model; and
(c) the steady-state concentrations assuming the common mass balance model,
C-28
-------
Table V-2. Comparison of Measured and Modeled
Concentrations of Methane
Expt.
1
2
3
4
5
6
7
8
9
Distance
from
source
(m)
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
Maximum concentration, C (ppm v/v)
(TlaX
Meas.a
100
44
21
55
19
15
53
20
11
55
24
14
55
22
17
18
29
NA
20
7.8
7.2
18
7.6
4.5
15
5.1
5.1
Modeledb
NA*
28
13
36
16
NA
47
16
7.5
81
NA
NA
NA
27
NA
21
NA
NA
15
7.7
6.4
14
5.7
4.4
18
7.1
4.8
Schroy model
8.4
5.8
4.8
3.6
2.4
2.0
2.1
1.5
1.1
8.0
5.6
4.3
4.6
3.2
2.5
1.6
1.1
0.90
7.2
5.0
4.0
4.0
2.9
2.3
1.7
1.2
0.94
Uniform mixing
model
16
16
16
6.8
6.8
6.8
3.6
3.6
3.6
14
14
14
8.7
8.7
8.7
3.1
3.1
3.1
14
14
14
7.7
7.7
7.7
3.2
3.2
3.2
= not available.
.Maximum concentrations are not necessarily on tunnel centerline.
Using Equation 1 (circular normal model).
C-29
-------
as described by Clement Associates (1981). The Appendix presents example
calculations for these models.
It is clear from Table V-2 that the simple mass balance (uniform
mixing) model, in addition to being incapable of describing any spatial vari-
ation, grossly underestimates the measured maximum concentration at the 2 m
distance. Although the uniform mixing model more closely estimates the mea-
sured maximum further downwind, this is the result of the small mixing cross-
section of the flow tunnel. Because full-scale packaging facilities would be
located in much larger room volumes, the large underestimation of concentra-
tion would persist throughout the area occupied by packaging equipment opera-
tors.
On the basis of Table V-2, the Schroy (1981) model also appears to
substantially underestimate the maximum concentrations. The degree of under-
estimation lessens as downwind distance increases, as the flow rate de-
reases, or as more obstructions are introduced into the flow. As an illus-
tration of the last point, note that in Table V-2 the agreement between the
Schroy model and measured maximum values are better for the hooded drum (Ex-
periments 7 through 9) than for the simple point release (Experiments 1
through 3). Figure V-8 compares the maximum concentration values measured
for the hooded drum tests with estimates of the maximum from both the Schroy
model and Eq. (2).
Schroy's model is based primarily upon the Gaussian dispersion
model developed for outdoor diffusion. In addition, because the source and
receptor can be very close to each other in a workplace setting, Schroy ap-
peals to dispersion coefficients presented by Bowne (1974) which again only
concern outdoor results and are basically extensions back from the minimum
100 m source-receptor separation of the Pasquill-Gifford curves (Turner,
1970). From Schroy1s paper, it is not clear whether rural or urban disper-
sion coefficients are included in the workplace model, nor is any substanti-
ation offered for his claim that "at present Gaussian models represented
[sic] the most practical method of analyzing the downwind impact of an emis-
sion to an [indoor] workplace." Finally, no experimental evidence is pre-
sented in the paper to justify the approach.
Thus, on the basis of the methane tests conducted to date, it would
appear that the models commonly employed to assess worker exposure (a) tend
to substantially underestimate the maximum concentration (especially close to
the source), and (b) are not capable of describing the spatial variation in
concentration.
B. Particulate Dispersion Results
Table V-3 compares the measured maximum concentrations of ammonium
chloride (single-point center release) with the values calculated using the
circular normal model (Eq. (2)). The model predictions and the measured
values compare favorably.
C-30
-------
30
25
20
o
8 15
o
O
|
i
S 10
5
I I
Experiment No.
\
Schroy
Model
(1 mph)
Schroy
Model
(2 mph)
7
8
9
• 1_mph
O 2 mph
• 5 mph
•Maximum Estimated
MRI Model
i
246
Downwind Distance (m)
8
Figure V-8. Comparison of measured maximum concentrations
(hooded drum tests) with two models for
the maximum values.
C-31
-------
Table V-3. Comparison of Measured and Modeled Maximum Concentrations
of Ammonium Chloride
Expt.
27
28
Downwi nd
distance
(m)
2
3.5
5
2
3.5
5
Measured
maximum
(mg/m3)
75
12
9.2
110
48
14
Modeled
maximum
(mg/m3 )
70
11
11
130
77
5.0
Measured
dispersion
coefficients
ay
0.26
0.49
0.40
0.60
0.23
0.53
0
0
0
0
0
0
(m)
CT2
.31
.51
.43
.16
.22
.49
Locations
of
concentrations
>
0.
-0.
-0.
0.
0.
-0.
r
0080
064
093
27
18
0022
0
-0
-0
0
-0
-0
max
(m)
zo
.082
.033
.31
d
.0036
.14
.040
Maximum concentration locations are offset distances from tunnel centerline
The vertical plume spread was photographed at 1 m downwind for a
single point tunnel centerline release of ammonium chloride under different
wind speed conditions. The plume height was visually estimated for several
different exposures at the same wind speed, averaged, and plotted in Figure
V-9. These visible plume spreads were assumed to represent 4 a (4.3 ay would
encompass ~ 98% of the plume). Turner (1970) indicates that the 4 oy disper-
sion coefficients for stability classes D, E, and F range from 0.2 to 0.4 m
(if Turner's curves are extended on log-log paper to 1 m from the source).
These figures compare well with results from the photographs (e.g., 0.18 to
0.54 m).
VI. PROCEDURES FOR CALCULATION OF MAXIMUM CONCENTRATIONS
Three models which have been proposed for estimating worker exposure
to gaseous emissions are described in this section. The first two models ad-
dress only maximum concentrations downwind of the source; the third (based on
the pilot laboratory study test results) provides a description of the spatial
variation in concentration. The other major differences between all models
in this section and that presented in Volume I of this report lie in the omis-
sion of flow obstacles and worker position/orientation in the estimation of
exposure.
1. Mass Balance Model
Assuming plug flow through the chamber with velocity U (in mph),
the volumetric flow Q is:
Q = (2.68 m) (2.12 m) (U (mph))
= 9,140 U
where Q is in m3/h and U is in mph.
(D
(2)
C-32
-------
o
I
CO
oo
L1.0I
"^"
•^^
O
i
3
O.
"5
XI
Q.
"5
0)
>
^^
;o.8J
'e
'I
in ' »
' 5 • 0'6
co j •«"-•!
"»!
QCi
E;
* , .
^M I 1 J
^^ (
:E]
^j
i QJ - i
n o
L- - v.f.
^
—
_
Is" = 4l
\
\
X
>v
N|n = 4 ; Extrapolated from Turner's ;
X Workbook for Stability
\Tn = 2' Classes D1 E and F.. '
•^^^^^^ '
* ^ ^! (
^^~^~~-~-JLl • '
i i 1 I 1 I 1
0 1 (mph) 2 3
I | i I i i i i
b| 0.2,! 0.4; 0.6j(m/s) 0.8 .1.01 '1.2; '1.4
Wind Speed
Figure \J-9. Comparison of photographically determined vertical spread
of ammonium chloride plume with Turner dispersion
coefficients.
-------
The steady-state concentration C estimated by simple mass balance
model (e.g., Eq. (7) on page 9 ofClement Associates, 1981) is:
Css = G/Q (3)
where: G = mass emission rate (mass/time)
Q = ventilation flow rate (volume/time)
As an example, for Experiment 3 with an emission rate of 82.5 g/h
and a tunnel velocity of 3.8 mph:
Q = 34,700 mVh
and, from (3),
r = 82.5 g/h
Ss 34,700 mVh
=2.4 mg/m3
For methane, 1 ppm = 656 |jg/m3 and:
Css =3.6 ppm
2. Schroy Model
The Schroy (1981) model for methane is:
max
where: C = estimated maximum concentration at a downwind
md* distance d (ppm)
G = mass release rate (g/h)
U = velocity of air (fpm)
-y- = function of distance d given in tabular form
For the same* example given above:
G = 82.5 g/h
U = 3.8 mph = 334 fpm
and, from Table VI-1, the following dispersion parameters are found:
1
d (27tayoz)
2 m 0.1020 m~2
4 m 0.07074 m"2
6 m 0.05507 m 2
C-34
-------
and substituted into Eq. (4).
as follows:
Thus, the estimated maximum concentrations are
2 m
4 m
6 m
3. Circular Normal Model
max
2.1 ppm
1.5 ppm
1.1 ppm
As an example of the circular normal model, consider the 2 m down-
wind station for Experiment 3. Figure VI-1 shows the concentration values
used for modeling purposes. Values of -9.99 are less than background concen-
tration levels or do not meet a blank criterion. Because the spacing between
samplers on the y (lateral) axis is 0.383 m and 0.303 m on the z (vertical),
the normal equations (see Section IV) in this example are:
10.8
.793
-0.354
0.396
-5
0.477
0
-n
0.477
0.0599
0
n
0
0
0.293
n
o i
0
0
n i on -
a2(
a3j
Note that zero or negative values of net concentration are not
used. In this example, only five data points can be used. The solution
vector is:
3i = 3.75
a2 = -16.6
a3 = -1.20
a4 = 2.16
Thus, with a change of variables discussed in Section V, the fol-
lowing dispersion parameters are found:
A = 47
I/a2 = 17
yQ = -0.036
ZQ = +0.065
These are the values given for the 2 m downwind station of Experi
ment 3 on Table V-l. In practice, a BASIC language program was used to gen-
erate the dispersion parameters reported here.
C-35
-------
Table VI-1. Diffusion Function Reported by Schroy (1981)
Distance,
d i
ft m
1.5
1.64 0.5
2.0
3.0
3.28 1.0
4.0
5.0
6.0
6.56 2.0
7.0
8.0
9.0
9.84 3.0
10.0
Value of
U/(2nayoz))
0.1458
0.1447
0.1420
0.1347
0.1326
0.1259
0.1166
0.1072
0.1020
0.09929
0.09288
0.08653
0.08120
0.08069
Distance,
d
ft m
12.0
13.12 4.0
14.0
16.0
16.41 5.0
18.0
19.69 6.0
20.0
25.0
26.25 8.0
29.53 9.0
30.0
32.80 10.0
35.0
Value of
(l/(27toyaz);
0.07431
0.07074
0.06844
0.06324
0.06217
0.05873
0.05507
0.05451
0.04912
0.04528
0.04188
0.04131
0.03789
0.03521
Distance,
) d
ft
65.6
98.4
131.0
164.0
196.8
229.7
262.5
295.3
328.1
m
20
30
40
50
60
70
80
90
100
Value of
(l/(2noyo-z))
0.01989
0.01255
0.00938
0.00793
0.00642
0.00553
0.00463
0.00426
0.00372
C-36
-------
EXRERIMENT 3 LOCATION A
MEASUREMENTS < EXCLUDING BACKGROUND)
N
MEAN THC VEL
STANDARD DEU.
04 94
•1.99 3.90 -9.99 3.42
•f.9? D.20 -9.99 O.U
0
-9.99
-9.99
0
-9.99
-9.99
2
13.19
1.99
66 6
7.10 3.7* 53.34
4.24 0.17 15.54
3
3.57
3.10
0
-9.99
-9.99
0
-9.99
-9.99
4
3.91
0.43
4
3.70
0.20
1
*
3.83
0.10
4 44 04 0 4
3.44 2.82 3.95 -9.99 3.85 -9.99 3.S7
0.20 1.2? 0.34 -9.99 0.18 -9.99 ft.24
4
3.31
0.23
4
3.84
0.22
4
3.77
0.24
Figure VI-1. Methane concentrations above background level and
with blank criterion applied.
C-37
-------
VII. REFERENCES
Bowne NE. 1974. Diffusion rates. Air Pollution Control Association J
24(9): 832-835.
Clement Associates. 1981. Mathematical models for estimating workplace
concentration levels: a literature review. Clement Associates, Inc.,
Washington, D.C. Prepared for U.S. Environmental Protection Agency, Eco-
nomics and Technology Division, Washington, D.C.
Schroy JM. 1981. Prediction of workplace contaminant levels, pp. 190-206.
Symposium Proceedings on Control Technology in the Plastics and Resins Indus-
try, Atlanta, Georgia, February 27-28. Cincinnati, Ohio: National Institute
for Occupational Health and Safety. NIOSH Publication No. 81-107.
Turner DB. 1970 (Rev.). Workbook of atmospheric dispersion estimates.
AP-26. Environmental Protection Agency, Office of Air Programs, Research
Triangle Park, North Carolina.
C-38
-------
APPENDIX D
QUALITY CONTROL DATA
-------
Quality control data are presented in this appendix for the two
plant studies and for the laboratory drumming study.
A. PLANT A DRUMMING TESTS
The internal quality control checks performed in this test program
included: analyses of field blanks, spiked blank silica gel tubes, and
analysis of duplicate (colocated) samples by separate laboratories. In addi-
tion, a Century continuous hydrocarbon analyzer was used to verify the very
low levels of air concentrations of Fluid 1 during Test A-l in comparison to
the background hydrocarbon concentration of 12 ppm as propane. The concen-
tration of analyte on the field blank was found to be below the limit of
detection. Recovery of analyte from blank tubes spiked with 5 ug of analyte
was 128% for Fluid 1 and 123% for Fluid 2. This indicated that good recov-
eries could be obtained from silica gel. Analysis of randomly selected back-
up portions of two tubes for each analyte indicated levels beneath the limit
of quantisation of the method.
The results of the analysis of colocated samples by MRI and by the
plant laboratory are compared in Table D-l. Total analyte concentrations from
duplicate (colocated) samples agreed well except in the cross laboratory com-
parisons for Test A-l. However, the higher concentrations determined by the
plant laboratory appear to be inconsistent with the consistently lower con-
centrations obtained by MRI at all sampling locations for Test A-l.
Table D-l. Analysis of Duplicate (Colocated) Samples
Analyte concentration (ug/m3)
Plant
MRI laboratory
Test-sampling location analysis analysis
Al-2a,b
Al-8a,b
Al-9a,b
Al-15a,b
A2-2a,b
A2-8a,b
A2-9a,b
A2-15a,b
< 72,
< 80
< 76
< 78,
< 184
< 162
< 216
< 154
< 70
< 80
, < 208
, < 197
150
180
< 210
240
Two external quality control performance audit samples were per-
formed using Fluid 2 supplied by the QCC. The first sample showed that the
actual concentration of Fluid 2 in acetone (40 ug/mL) compared well (103%
recovery) with the reported concentration (41 ug/mL). The second sample of
25.4 ug/mL produced a reported value of 15.2 ug/mL, a 60% recovery.
D-l
-------
8. PLANT B BAGGING TESTS
The internal quality control checks performed in this test program
included: analysis of field and laboratory blanks and analysis of duplicate
(colocated) samples by separate laboratories.
The analyte masses found on five field blanks were each less than
the detection limit of 2 ug. Lab blanks also yielded less than 2 ug for each
of five blanks, and 5 ug for the sixth lab blank.
The results of the analysis of colocated samples by MRI and by the
plant laboratory may be compared in Tables IV-9 through IV-11 in Volume I of
this report. The location 3b samples (suspended particulate and respirable
particulate fractions) from each test were turned over the the plant labora-
tory for analysis; the colocated samples collected at locations 3a, 7a, and
7b were analyzed by MRI.
The external quality control performance audit samples consisted of
blind performance samples prepared by the QCC. Reported phosphate concentra-
tions compare favorably with the exception of the low level sample (2.5 ug/mL)
near the limit of detection. The actual concentrations of the four blind
performance audit samples are shown below.
Phosphate concentrations (ug/mL)
Recovery
Actual Found (%)
0.0
2.5
5.0
10.0
< 2
1.6
4.7
9.3
_
64
94
93
C. LABORATORY DRUMMING TESTS
Two primary instruments were used to characterize analyte concen-
trations carrier gas (air speeds during the laboratory drumming tests. The
Beckman 402 Total Hydrocarbon Analyzer detected total hydrocarbons emitted
from the drum. Hastings air meter was used to measure wind speeds in the
flow tunnel. The quality control data obtained for these - instruments are
presented below.
1. Beckman 402 THC Analyzer
The Beckman 402 THC analyzer was calibrated each day with freshly
prepared ~ 8 L bag samples of chemical vapors. This entailed the use of inert
dry gas, appropriately-sized and calibrated syringes, 10-L capacity Tedlar
bags, and a calibrated dry gas meter with temperature gauge. The gas standard
for the Beckman calibration was prepared using a known amount of the chemical
of interest which was injected into a known amount of hydrocarbon-free gas
D-2
-------
(N2). The bag contents were heated and then mixed by alternately squeezing
each side of the bag from 25-50 times. The bag contents were then sampled in
order to calibrate the Beckman 402 at the initial span reading. Subsequent
checks of this standard and other prepared bag standards established a sensi-
tivity decay function for each day of testing. Hydrocarbon free (zero) air
was used to calibrate the Beckman at the low end of the scale. Table D-2 pre-
sents the calibration and calibration check data used in the drumming tests.
2. Hastings Air Meter
Calibration of the Hastings air meter under low air flows U 250 fpm)
required that the probe be inserted through a side orifice into the center of
a 59.5 cm long (5.1 cm I.D.) PVC tube. The outlet of the PVC tube was con-
nected to the inlet of a Roots positive displacement meter, allowing the Roots
meter to adjust the flow through the PVC tube. The Roots meter is known to
read low by 1£ at 2.5% of maximum flow (125 ftVmin). The following Table D-3
presents the calibration data taken on 8/13/85 with a temperature of 292 K and
a barometric pressure of 736.3 mm Hg.
3. Comparison of Integrated Mass Flux to Drumming Emissions
The concentrations measured in the sampling plane at 3 m downwind
from the drum center were combined with corresponding wind speeds to produce
point mass flux values. These were integrated across the sampling plane to
produce a comparison of total mass fluxes with drumming emission rates. The
results of these comparisons are presented in Table D-4.
The emission rate of perchloroethylene was measured at 326 mg/s and
is higher than each of the three mass flux values of 211, 199, and 255 mg/s.
The mass flux values are believed to be low because perc emissions consist of
a large fraction of mist. Significant perc mass is suspected to have settled
to the floor before reaching the 3 m downwind sampling plane.
The methanol 3 m integrated mass fluxes are all higher than the
measured drumming emission rate of 206 mg/s. As noted in the dispersion and
exposure test results, the bubbling of air through a half-filled drum cooled
the methanol to several degrees Centigrade less than the temperature of the
ambient air. Methanol vapors were consequently shown to settle to the tunnel
floor downwind of the drum. The concentration values measured very close to
the floor at the 3 m position were combined with fairly high wind speeds and
thus produced significant, yet probably unrealistic, contributions to the
total mass flux. In addition, the cooling of the drum contents may have pro-
duced a lower molar volume for the drum exit gas than was calculated using
the methanol liquid temperature. Both of these factors would tend to increase
the calculated mass flux of methanol.
Two other factors are known to influence the vapor emission rate -
vapor pressure and molecular weight. When only these two parameters are
considered, vapor emissions of methanol should exceed that of perc by a factor
of 1.30.
D-3
-------
Table D-2. Daily Calibration of Beckman 402
Date
Time
Beckman 402
Attenuation Concentration
reading (ppm)
8/1/85
8/2/85
8/5/85
8/6/85
8/9/85
8/13/85
8/20/85
8/21/85
8/22/85
8/23/85
8/28/85
8/29/85
8/30/85
9/2/85
9/3/85
1109
1423
1632
1658
1046
1146
1635
1223
1331
1316
1448
1634
1011
1156
1437
1713
1735
2159
2226
1809
2048
1811
2345
1927
2147
2341
1606
1918
2244
1643
1950
1800
1810
1828
1950
2300
2331
1906
2223
2223
2341
X1000
X1000
X1000
X1000
X1000
X1000
X1000
X1000
X1000
X1000
X1000
X10
X10
X10
X10
X10
X50
X50
X50
X50
X50
X50
X50
X50
X50
X50
X50
X50
X1000
X100
X100
X5000
X5000
X5000
X500
X500
X5000
X10
X50
X50
X100
22941
21154
18338
18406
21044
19815 •
18451
20885
19449
18521
21411
107
87
84
87
81.5
103
84
104
104
95
101
79.3
889
779
816
1028
954
33423
1208
1208
104755
108870
107788
5238
4246
94728
50
29
29.8
29
Bag concen-
tration (ppm)
22941
22941
18338
18338
21044
21044
21044
20885
20885
18521
21411
107
87
87
87
87
103
103
103
104
104
101
101
889
889
889
1028
1028
33423
1208
1129
104755
104755
104755
5238
5238
94728
50
29
29
29
Corr. factor3
1.00
1.08
1.00
1.00
1.00
1.06
1.14
1.00
1.07
1.033/hr est.
1.033/hr est.
1.033/hr est.
1.00
1.04
1.00
1.07
1.00
1.27
0.99
1.00
1.09
1.00
1.27
1.00
1.14
1.09
1.00
1.08
1.033/hr est.
1.00
1.07
1.00
0.96
0.97
1.00
1.23
1.033/hr est.
1.033/hr est.
1.00
0.96
1.033/hr est.
D-4
-------
Table D-2. (Concluded)
Date Time Beckman 402 Bag concen- Corr. factor3
Attenuation Concentration tration (ppm)
reading (ppm)
9/11/85
9/11/85
9/13/85
1200
1203
1420
1421
1422
1137
1448
1449
1606
1607
2044
2045
X500
X500
X500
X500
X500
X500
X500
X500
X500
X500
X500
X500
57.9
43
1007
999
349
1000
939
321
319
946
885
273
57.9
57.9.
1000°
1000°
350 h
1000°
1000°
350^
350 c.
1000°
1000°
350C
1.00
-
1.01
1.00
1.00
1.00
1.06
1.09
1.10
1.06
1.13
1.28
Bag concentration -=- Beckman 402 concentration reading; values of 1.00
bsignify initial span calibration was done.
Empty wet drum "A" of ethylene glycol used to reproduce 9/11/85 calibra-
ction on 9/13/85.
Empty wet drum "B" of ethylene glycol used to check 9/13/85 calibration
against 9/11/85 calibration.
D-5
-------
Table D-3. Calibration Data of Hastings Air Meter
Roots meter flow rate
(m3/min)
Actual
0.043
0.075
0.140
Corrected (STP)
0.0393
0.686
0.1281
Tube air speed (fpm)
At STP
63.5
111
207
+1% Correction
64.1
112
209
Indicated
air speed (fpm)
Hastings
72
115
200
Indicated/
actual
1.12
1.03
0.96
Table D-4. Comparison of Integrated Mass Flux With Drumming Emissions
Nominal
tunnel
centerline
velocity
Date
8/20/85
8/22/85
8/23/85
8/28/85
8/29/85
8/30/85
8/30/85
Times
1716-1822
1942-2047
2130-2252
2155-2338
2054-2209
1834-1950
2317-0018
(fpm)
200
200
100
200
200
100
100
Chemical
Perc
Perc
Perc
Methanol
Methanol
Methanol
Methanol
Fluid
temp.
(°C)
22
20
21
20
21
23
20
Mass
Raw
196
186
213
266
303
329
326
flux (mg/s)
Corrected
211
199 222
255
293
345 0/M
347 341
378
Emission
rate
(mg/s)
326
206
Based on checks of Beckman THC analyzer with bag standards.
D-6
-------
APPENDIX E
DATA FROM LABORATORY DRUMMING TESTS
-------
APPENDIX E
TABLE OF CONTENTS
Pages
Dispersion Tests
Perch!oroethylene E-l thru E-16
Methanol E-17 thru E-31
Emissions Tests
Perchloroethylene E-32 thru E-47
Ethylene glycol E-48 thru E-70
E-i
-------
.Tunnel Celling.
Chemical: Pcrchlorcthy-
lene
^
x;
^
>*
^
Left ;X
Tunnel^ 4
Wall X^
^
P
f?
^ ^ ' ^ -^ ^ ,
1 meter , c
i
2
3 8 14 9 2
2 10 60 231 87 «
81 108
10 29 260 45 6
2 10 49 275 79 7 2
5 13 60 24 63
6 11 45 5 2
6 10 25 3 2
i 1 1 L L. . \ . \ „ , I . 1 1 ^ ^ J
>ung Post 1 ion: Side
ianpllng Plane: 1 KM; for
real Date: 8/22/85
Concentration (ppn)
•
^
^X Right
X- Tunnel
>^Wall
0
^
^
-2-, -.--a
Tunnel Floor
-------
ro
^
Left ^
Tunnel ^4
Wall 9S • l'Q\ • K08 • •
.97 .76 .56 .97 ^ .97^
.56 .18 .94 0
-1 * ,-* 1 * » * 1 • > - k > ^ -
Chemicali Perchlorothy-
Icne
PasItIant Side
Sampling Planet 1 meter
Test Datei 8/22/85
Hind speed (m/a)
Tunnel Floor
-------
Chemical: Pcrchlorethy-
lenc
Left
Tunnel -^ 4
Wall
1 meter ,
0
1 8
53
2 21 73
393
1 24 101 369
1 19 157 2'J7
1 34 111 223
1 26 128 223
!
<
1
1
10
137
101 1
128
". I I
105
79 6 0
108 9 1
45 5 1
1 i 1 I
Bung Position: Side
Sampling Plnnci 1 meter
(2 cfra fill rate)
Test Date: 8/23/85
Concentration (pptn)
Right
-<>- Tunnel
Tunnel Floor
-------
Chemicali Perchlorethy-
lenc
0
^
0
1
Left ^
Tunnel
^
x-1
Xfj
i
. 1 mater ,
I I
i
.46 ^ .41^ .41
.18
• -5l • '°*« •**•
.04
.38 .04 .36
1111111144
Sampling Planet 1 meter
(2 cfm fill rate)
feat Datet 8/23/8S
Hind apeed (m/a)
%
^ Right
-^X Tunnel
X^X tJUnll
x^*" wail
0
%
^
^
Tunnel Floor
-------
cn
Chemical i Perchloretliy-
lene
Bung Position: Side
Sampling Planet 1 meter
(3 cfo fill rate)
Test Date: 8/23/05
Concentration (ppn)
Tunnel Floor
-------
CT>
Chemicalt Porchlarethy-
lene
Bung Positioni Side
Sampling Planet 1 meter
(3 eCn fill rate)
Test Datei 8/23/85
Hind speed (m/s)
unnel
unnel Floor
-------
I
•vj
Left
Tunnel
Wall
.Tunnel Celling)
, 1 meter ,
i
l
13
1 8 16 15
60
1 6 13 31 48
43 84
5 19 26 52 86
3 32 66 90 111
13 60 59 131 135
1 i i i k U^-
— • — r s , scssrrrff Bu,
San
°.
0 1
2 1
32
22 2 1
92
41 5 0 1^
94
113 41 15 3
114 59 32 22
140 107 97 43
1111
Chemlealt Perchlorethy-
lene
Bung Posltlont Side
Sampling Plane: 3 meters
Test Datei 8/23/85
Concentration (ppn)
^X Right
x> Tunnel
Tunnel Floor
-------
Chemicali Perchlorethy-
Icne
00
1
^
2
0
Left ^
Tunnel ^4
Wall ^
Ij.
-Xx
__ — — — — — — — rrrrrrrr
1 meter ,
•
Bung FOB it ton t Side
Sampling Plane i 3 meters
Fest Date: 8/23/85
Wind speed (n/s)
v •••••••••
• ••••••0*
.46 .51 .46
.43 .43 .46 .41 .43 .51
.38 .36 .30 .36 .41
1111111111
^
^ Right
-^ Tunnel
'^^^ Ufa II
^^^^ Wflll
^
^
Tunnel Floor
-------
Chemical: Perchlorethy-
Icnc
Bung PositIani Side
Sampling Planet 3 meters
Test Datei 8/22/8S
Concentration (ppm)
Right
-Xx Tunnel
Tunnel Floor
-------
m
i
Left
, Tunnel
Wall
Chemicali Perchlorcthy-
lene
Bung Posit loot Side
Sampling Plane: 3 meters
Test Datei 8/22/85
Wind speed (D/S)
Right
/x Tunnel
-------
Lelt „
Tunnel o 4
Wall
Chemicali Perchlorethy-
lenc
Bung Posltloni Center
Sampling Planet 1 meter
Test Datei 8/20/B5
Concentration (ppm)
•
Right
X^ Tunnel
Tunnel Floor
-------
ro
Tunnel Celling'.
Chenlcalt Pcrchlorethy-
i';^$$$^^
Bung Poaitloni Center
Sampling Planet 1 meter
Test Datei 8/20/85
Wind speed (in/a)
Right
unnel
Tunnel Floor
-------
Chemicali Perchlorethy-
lene
-^x
xjS
^
^x
^X
x;6
x^
Left ;X
Tunnel ^4
Wall - Wall
^x
^
^
^
^
^x
Xx
^^xjjx;;
Tunnel Floor
-------
emnj, ^
^
^
%
^
Led ;X
Tunnel ^4
Wall '^x
g
^5
^
. . Background 1, 1
1 meter , sai
Te
1.00 l.OS 1.00
• • • • . • • • • •
• ••••••••
1.00 .92 1.05
• • • ••• • • • •
.34 .79
• ••••••••
.67 .67 .94
• ••a*****
• ••••••••
.34 .41 .64
.1 i._ i * i . .* . ,. 1 J , A . , * . ., j ., ; .. '
Chenlea1i Perchlorethy-
lene
Bung PositIoni Center
Sampling Planei 2 metera
Teat Datei 8/20/8S
Wind speed (m/s)
Tunnel Floor
-------
Left
Tunnel
Wall
1 meter
^^
Chemical: Pcrchlorcthy-
lene
Bung Position: Center
Sampling Plane: 3 meters
Test Date: 8/20/85
Concentration (ppo)
•
* . 3 . | . ' .
6 21 27 19 6
2 6 26 32 55 44 21 S(
22 27 38 38 19
40
6 21 36 26
6 24 40 32
1 . . A . . , .i * v s± ; ; j > ; ^
I
.
3.
.
4
4
.1 . ,
< Right
Xx Tunnel
^
Tunnel Floor
-------
VZ&&&2f2#lf&
Lett
Tunnel<4
Wall
1 meter
\ZZ%Z2ZZ%
Chemical: Perchlorethy-
lenc
Bung PositIont Center
Sampling Planet 3 meters
Test Datei 8/70/85
Wind speed (m/a)
Right
Tunnel
Tunnel Floor
1 "v *
-------
i
^j
Left
Tunnel
Wall
XXCXX^''X^XCXx^XXx'x''S-'S»-
1 meter
6 10 12
7 9 11
6 9 109
19 IB 188
6 IB 227
7 21 173
6 72 211
8 128 253
1 i 1 i i
1
11
9
14
1367
1417
1322
1791
601
429
453
1
X S • ^X*- •r'JX/'X^''^XXX'^^.x'
s
Ti
• • • •
11 9
39 11
726 8 8
419 58 14
344 39 7
398 68 6
134
213 14 8
54
103 9 4
1 1 1 1
Chemicali Hcthanol
Bung Positiont Side
Sampling Planet 1 meter
Toot Date: 8/29/85
Concentration (ppo)
Xx Right
Xx Tunnel
Tunnel Floor
-------
CO
Left
Tunn8\^4
Wall
, 1 meter
5
7 4
2 S 11^
29 13
6 2 326
6 3 105 165
2 3 332 837
4 44 743 1564
15 317 1260 1944
1 . . i . . i . . i i.
1
3 6
23 68
2273 178g
4951 402
4262 163
2589 655
I860 661
1303
2180 559
1 i
So
Tc
12
4 ^ t
3 6
3 3
182 31
70 13
Chemicali Methanol
Bung Positioni Side
Sampling Planei 1 meter
Teat Datei 8/30/85
Concentration (ppn)
Tunnel Floor
-------
^^^^^e^^^.C.ftnj^.;^^-^^
,—"/
Left
Tunnel
Wall
Chemical: Melhanol
Bung Position] Side
Sanpllng Planei 1 meter
Test Date: 8/30/85
Wind speed (m/a)
Tunnel Floor
-------
ro
o
Left ^
Tunnel^
Wall
1 meter ,
7
15 7 9
18 70 27
14 149 391 127 18
9 497 302 44
11 16 46 225 424 286 177 57
10 39 42 260 423 260 164 84
13 38 108 217 299 244^ 196 56
32 73 210 250 251 140
1 . A A .. i . * ,. , * , . * * .
Sai
Tc
•
.
.
14
•
46
27
Chemicalt Mcthanol
x^x^Tunnel Cejlngx^".%%S&si2222, B™B Position, side
Saiapllng Plane i 3 metera
Test Datei 8/29/85
Concentration (ppm)
Tunnel Floor
-------
ro
Left
Tunnel
-------
Ctiemlca 11 He thnno 1
Bung Positions Side
ro
^
^
^J
xl
1
XJ/-
Left ;x:
Tunnel^; 4
^
0
^
P
•^x
1 meter ,
2
' ,2
5 55
20 464
4 250 583
835
112 226 -328 701
516
398 573 1006
774 578 1092 1390
\ 11111
^0'<.'"' 4'^x^ 5x^
Sanpllng Planet 3 nctc
Test Dote. 8/30/85
Concentration (ppn)
5 1
35 45 ^ ^X"
66 3 ^ Right
• • • • -^> Tunnel
145 ^<
"• . "4 . ^
275 . 275 85. ^
631 795 883 . ^
1359 1171 582 ^
>^ 6 -^' 8 9 x-^^^x-
Tunnel Floor
-------
Chemicali Hethanol
Sampling Planet 3 meters
Teat Date: 8/30/8S
Wind apccd (in/a)
Right
x^x- Tunnel
Tunnel Floor
-------
ro
Left
Tunnel-^4
Wall
Celn
1 meter ,
. 8 3 4
360
3 8 226 £ %
2893
12 391 3440 9 3
• ••••••
4532
13 35 1345 4974 632 20
• • • • - • • •
3498
7 35 1416 3862 1100 63
2331 161
16 697 1581 2391 1822 496
2071
9 59 200 1915 2913 2249 195
1 i i 1 *,,*,,!,*,
Sa
Tc
• •
• .
*
7
I .
7
6 6
Tj
19 12
,1,. 1 , ,^
Chemicali Hcthanol
Bung Position, Center
Sampling Planet 1 meter
Test Dotei 8/30/85
Concentration (ppm)
Right
-X- Tunnel
•V^-" x
3^'
' .S s' .•
Tunnel Floor
-------
I
INS
Ul
Left
Tunnel
Chenleali Hcthanol
Bung Position: Center
Sampling Planei 1 meter
Test Date: 8/30/85
Wind speed (m/s)
Right
xx Tunnel
X^Wall
Tunnel Floor
-------
Left
Tunnel
Wall
rel CeHIng>>''
, 1 meter
8 10
10 31
60 819
8 9 29 661
1034
37 - 1063
5 21 SOS
5 108 162
8 108 242
1— A J 1 i-
1
5
27
1936
1552^
1830
1960
2138
1087
495
307
1
.
14
IS
263
426
440
531
384
255
113
1
Background 5, 7, 4 s
T<
i
11
7
" •
54 14 11
69 6
25 5
10 4
1 1 1
'' '
Tunnel Floor
Chemical: Hethanol
Bung Position, Center
Sampling Plane, 1 meter
Tent Date: 8/28/85
Concentration (ppm)
Right
x>- Tunnel
-------
m
^
!
/J
Left ;X
Tunnel ^4
Wall -^
^
^
I
XQ
Background 5, 7, 4 S
, 1 meter ,
1 1 T
• ••••••••
.89 .86 .91
• ••••••••
• ••••••••
.94 .51 .46 .51 .94
• ••••••••
1.02
.08
1.02 .97 .08 .61
.61 .46 .41
. 1 1 1 1 A 1 1 i _ 1
Chemical: Hethanol
Bung Positioni Center
Sampling Planei 1 meter
Test Datei 8/28/85
Wind speed (n/s)
Tunnel Floor
-------
^Tunnel Celling x^
Chemicali Hcthanol
Bung Positioni Center
^
f
0
%
Left x^
71 Tunnel^
co W«» ;xj
x J
x?
X"
^
Sampling Planet 3 met
1 meter j
.' -7 2-
. 8-° ? ^
3 49 129 SO J .
7 30 302 290 IK 126 26
• • • ••• • • • *
7 72 268 400 S9Q 309^ 432 224^ 129^
216 524 746 784 559^ ^ 657 682^
626 833 889 1210 1239 1101 1371 1119 909
, i i j 1 Aii*^A ••
at Date t 8/30/85
Concentration (ppo)
0
^/ Right
.Xx- Tunnel
^
0
^
^
Tunnel Floor
-------
i i A A - . 1
Chenleal: Hethanol
. Bung Position: Center
Sampling Plane: 3 meters
Test Datei 8/30/85
Wind speed (a/a)
Wow
Xx- Tunnel
Tunnel Floor
-------
0
^
Xx
>5
,Lefl ^
'T1 1 Tunnel -^, 4
g [Wall ^
x^^x
>3'
0
^
X* rt
j- U
. 1 meter
11 12
10 8 51
68 139
10 28 36 127
7 20 87 167
8 44 125 204
25 36 101 224
13 69 133 252
1 i 1 1 1
i
7
23
52
235
168
338
337
326
3'2«
307
1
6
897
25 23 7
109
81 37 21
199 96 44
268 166 79
296 151 101^
319 226 199
111
r
s
Ti
•
.
.
10
28
34
'•
54
1
Chemicali Hothanol
Bung Posltloni Center
Sampling Planet 3 meters
Test Datet 8/28/85
Concentration (ppm)
Tunnel Floor
-------
I
CO
Left
Tunnel
Wall
.97
.89
.56
.89
.86
.91
.43
1 meter .
.99
.76
.76
.89
.76
.76
.51 .56
1111
Tunnel Floor
-91
.89
.56
Chemicali Methanol
Bung PasIt ioni Center
Sampling Planet 3 mete
Teat Date: 8/28/85
Hind speed (m/a)
^ Right
/Zs Tunnel
-------
CHEMICAL NAME
TEST NO. 2-
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-32
-------
CHEMICAL NAME
TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-33
-------
CHEMICAL NAME
TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-34
-------
CHEMICAL NAME
TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
32-100
TOTAL MASS LOSS /?,/
E-35
-------
CHEMICAL NAME %rcJ,/o
TEST NO.
FILL MODE
onJ
MAX.
CONCENTRATION
ATTENUATION
11
TOTAL MASS LOSS
E-36
-------
CHEMICAL NAME
TEST NO.
FILL MODE
MAX.
CONCENTRATION
3/300
ATTENUATION
TOTAL MASS LOSS
E-37
-------
CHEMICAL NAME
TEST NO. //_
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-38
-------
CHEMICAL NAME
TEST NO. /Z
FILL MODE 700
~
MAX
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
c/
E-39
-------
CHEMICAL NAME fflrdi/oree
TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-40
-------
CHEMICAL NAME
TEST NO. 23
FILL MODE
j%>#
o»t
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-41
-------
CHEMICAL NAME
TEST NO.
FILL MODE tforfart @ //
MAX.
CONCENTRATION 36>#Q0
^//
ATTENUATION
TOTAL MASS LOSS
E-42
-------
CHEMICAL NAME
TEST NO.
FILL MODE
/
CONCENTRATION
ATTENUATION
22.
X/MO
TOTAL MASS LOSS
E-43
-------
CHEMICAL NAME
TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-44
-------
CHEMICAL NAME
TEST NO. 30
FILL MODE
MAX.
CONCENTRATION 37000
ATTENUATION
TOTAL MASS LOSS
E-45
-------
CHEMICAL NAME
TEST NO.
FILL MODE 8*#
e»,
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-46
-------
CHEMICAL NAME %rMrt>£
TEST NO. 32
FILL MODE ffefo* &. /&, 7 a
MAX.
CONCENTRATION 372O0
ATTENUATION
l/QQQ
TOTAL MASS LOSS
E-47
-------
CHEMICAL NAME
TEST NO. /
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TI
TOTAL MASS LOSS
E-48
-------
CHEMICAL NAME
TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
27.0
TOTAL MASS LOSS /. 0
E-49
-------
CHEMICAL NAME
TEST NO.
¥•_
FILL MODE ~7d0 @ 2&,/
MAX
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-50
-------
CHEMICAL NAME
TEST NO.
FILL HODE
MAX.
CONCENTRATION
ATTENUATION
37.2
TOTAL MASS LOSS
E-51
-------
CHEMICAL NAME
TEST NO.
FILL MODE 72?
~
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS /.2
E-52
-------
CHEMICAL NAME
TEST NO. "7
FILL MODE 0
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS /
E-53
-------
CHEMICAL NAME
^
TEST NO. 7
FILL MODE 700 @ 2£.3
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-54
-------
CHEMICAL NAME
TEST NO.
FILL MODE 700$
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS /. ##
E-55
-------
CHEMICAL NAME
TEST NO. //
MODE
MAX.
CONCENTRATION
AHENUATION
TOTAL MASS LOSS
E-56
-------
CHEMICAL NAME
TEST NO.
FILL MODE
ATTENUATION
MAX.
CONCENTRATION //7# W»
TOTAL MASS LOSS
E-57
-------
CHEMICAL NAME
TEST NC.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-58
-------
CHEMICAL NAME
TEST NO.
FILL ,'WDE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-59
-------
CHEMICAL NAME £f6tf/e#e
TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-60
-------
CHEMICAL NAME
TEST NO.
FILL MODE
MAX.
CONCENTRATION
AHENUATION
TOTAL MASS LOSS
E-61
-------
CHEMICAL NAME
TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-62
-------
CHEMICAL NAME
TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS 0,22
E-63
-------
CHEMICAL NAME £f6u/e»e
C7 <
TEST NO. /?
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-64
-------
CHEMICAL NAME
TEST NO. 20
FILL MODE 7^7
~~
MAX.
CONCENTRATION
ATTENUATION
/ r
TOTAL MASS LOSS // e,
E-65
-------
CHEMICAL NAME £Mtt/e»f_
~J
TEST NO. 2/
FILL MODE
/0.2
MAX.
CONCENTRATION
ATTENUATION
/ /
TOTAL MASS LOSS
E-66
-------
CHEMICAL NAME Ejj
v7
TEST NO. 22.
FILL MODE
ft //,2
MAX.
CONCENTRATION
ATTENUATION
fT
TOTAL MASS LOSS
E-67
-------
CHEMICAL NAME
TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL MASS LOSS
E-68
-------
CHEMICAL NAME £#r«/e»
-------
CHEMICAL NAME £*Atf/exe
TEST NO.
FILL MODE
MAX.
CONCENTRATION
ATTENUATION
TOTAL flASS LOSS
E-70
------- |