»EPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
EMB Report 80-VNC-18
July 1982
Air
Industrial Surface
Coating
Emission Test Report
General Tire and
Rubber Company
Reading, Massachusetts
Test Series 2
-------
PROCESS EMISSION TESTS AT THE
GENERAL TIRE AND RUBBER COMPANY
VINYL-COATED FABRIC PLANT IN
READING, MASSACHUSETTS
MARCH, 1981
Environmental
Consultants, Inc
Thomas M. Bibb Samuel S. Cha
EPA Project Officer Work Assignment Manager
Nancy D. McLaughlin Eric A. Pearson
EPA Task Manager Project Scientist
EMB Project 80-VNC-1B S. Dexter Pierce
ESED Project 80/19 Environmental Engineer
EPA Contract 68-02-3543
Work Assignment 2 Peter w. Kalika
TRC Project 1473-E80 Program Manager
July 31, 1982
800 Connecticut Blvd.
East Hartford, CT 06108
(203) 289-8631
-------
PREFACE
The work described herein was conducted by personnel from TRC
Environmental Consultants, Inc. (TRC), the Radian Corporation/ Engineering
Science, General Tire and Rubber Company (GTR) in Reading, Massachusetts, and
the United States Environmental Protection Agency (EPA).
The scope of work was issued under EPA Contract No. 68-02-3543, Work
Assignment 2. The work was performed under the supervision of the TRC Work
Assignment Manager Mr. Samuel S. Cha.
Radian personnel responsible for monitoring process operations during the
testing program included Mr. Hal Laube, Senior Engineer, and Ms. Nancy Krohn,
Chemical Engineer. Engineering Science was responsible for fabric residual
solvent analyses. Personnel of the GTR Reading, Massachusetts, plant whose
assistance and guidance contributed greatly to the success of the test program
include Mr. Henry Waldron, Senior Process Engineer.
Ms. Nancy D. McLaughlin, Office of Air Quality Planning and Standards,
Emission Measurement Branch, EPA served as Task Manager and was responsible
for coordinating the test program.
-ii-
-------
TABLE OP CONTENTS
SECTION PAGE
PREFACE ........ ii
1.0 INTRODUCTION 1
1.1 Background 1
1.2 Brief Process Description 2
1.3 Measurement Program ,. 5
1.4 Description of Report Sections 6
2.0 SUMMARY AND DISCUSSION OF RESULTS 7
2.1 Summary of Results » 7
2.2 Duct Flowrate Measurements 8
2.3 VOC Measurements 10
2.3.1 FID Sampling 10
2.3.2 NMO Sampling with Method 25 ' 19
2.4 Print Room Ambient Air Measurements 20
2.4.1 Ambient Air VOC Measurements 20
2.4.2 VOC Measurements at the Embosser 22
2.4.3 Doorway Flowrates 26
2.4.4 Eight-Hour Exposure Sampling 27
2.5 Carbon Bed Wastewater Samples 28
2.6 Fabric Solvent Residue 32
2.7 FID Analyzer Audit Results 38
3.0 PROCESS DESCRIPTION ..... 42
3.1 General Description 42
3.2 Printing Operation Description 45
3.3 Printing Operation Emission Controls 47
3.4 Print-Line Building Air Circulation 48
3.5 Operation Parameters Controlled During this
Testing Program 48
. 3.6 Monitoring of Process Operations 50
4.0 DESCRIPTION OF SAMPLING LOCATIONS 51
4.1 Carbon Adsorption (CA) Unit Inlet 51
4.1.1 Flowrate Measurements 51
4.1.2 VOC Sampling 51
4.2 CA Unit Outlets 53
4.2.1 Flowrate Measurements 53
4.2.2 VOC Measurements 56
4.3 Embosser Electrostatic Precipitator (ESP) Inlet . . 56
4.3.1 Flowrate Measurements 56
4.3.2 VOC Sampling 56
4.4 Wall Fan Exhaust Duct 58
4.4.1 Flowrate Measurements 58
4.4.2 VOC Sampling 58
4.5 Print-Line Building Ambient Air Measurements ... 58
4.6 Wastewater Sampling Locations 60
4.7 Wallcovering Product Sampling Locations 60
-iii-
-------
TABLE OF CONTENTS (Continued)
SECTION
PAGE
5.0 SAMPLING AND ANALYSIS METHODS 63
5.1 EPA Reference Methods Used During this Program . . 63
5.2 Duct Flowrate Measurements . 63
5.3 VOC Measurements with FID Analyzers 65
5.3.1 Sampling with FID Analyzers 65
5.3.2 Calibration of the FID Analyzers ........ 67
5.3.3 Audit Sample Analysis of FID Analyzers 70
5.3.4 Data Reduction and Calculations for FID
Analyzers 71
5.4 MHO Sampling with Method 25 75
5.4.1 Preparation for Method 25 75
5.4.2 Sampling for Method 25 75
5.4.3 Analysis for Method 25 77
5.5 Print Room Ambient Air Measurements 78
5.6 Wastewater Sampling 80
5.7 Fabric Solvent Residue ...... 81
5.8 Effects of Process Operations on VOC Emission
Measurements .................. 81
REFERENCES ........ 84
APPENDICES
A
B
E
F
G
B.I
B.2
C.I
C.2
C.3
D.I
D.2
E.I
E.2
E.3
G.I
G.2
VELOCITY TRAVERSE DATA FORMS
FID VOC DATA
VOC Work Sheets
FID Strip Charts
FID ANALYZER OPERATIONS
FID Calibration Procedures
Data Reduction Procedures
FID Calibration Data, EPA Audit Results
METHOD 25 NMO SAMPLING
NMO Sampling and Analysis Procedures
1010 Data Forms and Analysis Results
PRINT ROOM AMBIENT AIR MEASUREMENTS
Ambient Air Data Forms
Instrument Calibrations
Charcoal Tube Analyses
WASTEWATER SAMPLING
FABRIC SOLVENT RESIDUE
Method for Determination of Residual
Solvent in Fabric
Field Data Sheets
H
PROCESS OPERATIONS
-iv-
-------
LIST OF TABLES
TABLE PAGE
2-1 Flowcate Measurements at the Printing Operation 9
2-2 Summary of FID VOC Emissions from Printing Operations ... 12
2-3 Carbon Adsorption Unit Control Efficiencies 18
2-4 Summary of NMO Method 25 Analysis Results 21
2-5 Summary of Print Room Ambient Air Surveys 23
2-6 Duct and Ambient Air VOC Measurements at the Embosser ... 25
2-7 Eight-Hour Sampling Data in the Print Room 29
2-8 ' Carbon Adsorption Unit Wastewater Analysis Results .... 33
2-9 Summary of Wallcovering Solvent Residues 34
2-10 Summary of Time Lag and Pattern Duplication of Wallcovering
Samples 37
2-11 Audit Sample Analysis Results 40
5-1 FID Analyzer Calibration Concentrations Used During the
VOC Measurement Program 69
5-2 Propane and MEK Calibration Equations used to Establish
Propane-to-MEK Conversion Equations 72
-vi-
-------
LIST OP FIGURES
FIGURE PAGE
1-1 Overhead Diagram of Printing Operation Facilities 3
1-2 Schematic of Print-Line Operation 4
3-1 Overhead Diagram of Printing Operation Facilities 43
3-2 Schematic of Print-Line Operation ........ 44
4-1 Carbon Adsorption Unit Inlet Velocity Traverse Location . . 52
4-2 Carbon Adsorption Unit Inlet VOC Sampling Location .... 54
4-3 Carbon Adsorption Unit Outlet Sampling Locations 55
4-4 Embosser Electrostatic Precipitator Inlet Sampling
Location 57
4-5 Wall, Fan Exhaust Sampling Location 59
4-6 Print-Line Building Ambient Air Measurement Locations ... 61
4-7 Ambient Air Measurement Locations Along the Embosser ... 62
5-1 Flame lonization Detection Sampling System . 66
5-2 Method 25 Sampling Train 76
5-3 Residual Solvent in Fabric Purging System .. 82
-v-
-------
1.0 INTRODUCTION
1.1 Background
Section 111 of the Clean Air Act of 1970 charges the Administrator of the
U.S. Environmental Protection Agency (EPA) with the responsibility of estab-
lishing Federal standards of performance for new stationary sources which may
significantly contribute to air pollution. When promulgated, these standards
of performance for new stationary sources are to reflect the degree of emis-
sion limitation achievable through application of the best demonstrated emis-
sion control technology. EPA utilizes emission data, obtained from controlled
sources in the particular industry under consideration, as a partial basis for
new source performance standards.
The EPA Office of Air Quality Planning and Standards selected the General
Tire and Rubber Company (GTR) vinyl-coated fabric plant in Reading, Massachu-
setts, as a site for an emission testing program. This plant produces wall-
coverings and is considered to employ process and emission control equipment
representative of the state-of-the-art in the vinyl coating industry. The
test program was designed to provide a portion of the emission data base re-
quired for the vinyl-coating industry new source performance standards.
EPA engaged TRC to measure volatile organic compound (VOC) emissions from
the printing operation at the GTR plant. Most measurements were conducted
during times of normal operation of the wallcovering printing process and as-
sociated emission control equipment. Some special operating conditions were
established for the purposes of this measurement program. The measurement
program was conducted during March 16-27, 1981, and was the second EPA test
conducted at this plant. The first test was performed by TRC in September and
October 1980 (1), but because of printing process and emission control opera-
tion problems a second test was needed.
-1-
-------
1.2 Brief Process Description
Figure 1-1 presents an overhead view of the plant facilities associated
with the printing operation, and Figure 1-2 presents a schematic of the print-
ing operation. This process is described very basically in the following par-
agraphs.
The printing operation consists of a Baker-Perkins rotogravure printing
machine utilizing six printing heads. The vinyl-coated substrate is fed
through a preliminary dryer, the six print heads, and an embossing unit. Pre-
mixed ink is supplied to each print head from a pump tank located next to each
print head. Ink is pumped from the pump tank to a tray within the print head
where a print roller, half-submerged in the tray, transfers ink from the tray
to the substrate. The inked substrate is dried in an oven contained within
each print head. Excess ink is gravity fed back to the pump tank. During a
print run, solvent or ink base is occasionally added manually to the pump
tanks to maintain the required ink viscosity. The solvent used in the inks is
primarily methyl ethyl ketone (MEK) with some methyl isobutyl ketone (MIBK)
and toluene.
Emissions from the preliminary dryer and print head ovens are manifolded
and ducted to a carbon adsorption (CA) unit before being released to the at-
mosphere. The CA unit has three carbon beds, but only beds 1 and 3 were used
during the test program. Emissions from the embosser are controlled with an
electrostatic precipitator (ESP). Fugitive emissions within the print-line
building are vented to the atmosphere through a pair of wall exhaust fans.
Air is supplied to the print-line building by a make-up fan on the roof and
from seven doors that open to the outside and to other areas of the plant.
During the test program only one wall fan was operated and then only briefly.
The make-up fan was off at all times. All doors but one were closed.
-2-
-------
EMBOSSER
ELECTROSTATIC
PRECIPITATOR
WALL FAN
EXHAUST
DOOR
F
PRINT LINE E
BUILDING
MAKE-UP
AIR INTAKE
WALL FAN /
EXHAUST'
EXTENSION
OVERHEAD
DOOR
J
EXHAUST DUCT FROM
PRINT HEADS
N
DOOf
DOORS
MAIN PLANT BUILDING
CA INLET VELOCITY
TRAVERSES
CA INLET""
VOC SAMPLING
HUMIDITY
CONTROLLERS
CA UNIT
CONTROL ROOM
CARBON BEDS
SAMPLING LOCATIONS
CA = CARBON ADSORPTION
CA OUTLETS
TESTED
NOT TO SCALE
FIGURE 1-1:
OVERHEAD DIAGRAM OF PRINTING OPERATION FACILITIES AT
GENERAL TIRE AND RUBBER COMPANY, READING, MASSACHUSETTS
-3-
-------
WALL FAN
EXHAUST DUCT V
TO CARBON
ADSORPTION UNIT
WALL FAN
oo
MAKE-UP
AIR UNIT
oo
ELECTROSTATIC
PRECIPITATOR
.WALL FAN
(NOT USED DURING
TESTING)
*>.
i
PRINT
HEAD ,
FANS j
^cnTOTffy
I
a
I
a
x^
0
VINYL PRELIMINARY
COATED DRYER
SUBSTRATE
PRINTING HEADS/OVENS
EMBOSSER
FINISHED
PRODUCT
FIGURE 1-2:
NOT TO SCALE
SCHEMATIC OF PRINT-LINE OPERATION
AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTES
-------
1.3 Measurement Program
The measurement program was conducted at the GTR vinyl-coated fabric plant
in Reading, Massachusetts during March 16-27, 1981. The emission tests were
designed to quantify controlled and uncontrolled VOC emissions from the print-
ing process. General sampling locations are shown in Figure 1-1. The mea-
surement program consisted of the following:
Flowrate Measurements
Velocity traverses were performed at the embosser ESP inlet, wall fan
exhaust duct, CA unit inlet and CA. unit outlets. Plowrates were cal-
culated from the velocity head and temperature data obtained during
these traverses. Measurements were performed in accordance with EPA
Methods 1 and 2.
VOC Measurements by FID Analyzer
VOC concentrations were continuously monitored at the embosser ESP
inlet, wall fan exhaust duct, CA unit inlet and CA unit outlets using
flame ionization detection (FID) analyzers. Sampling and calibration
were performed in accordance with proposed EPA Methods 25A and 110.
HMO Measurement by Method 25
Integrated gas samples were periodically drawn at the four sampling
locations during the FID continuous monitoring. This sampling was
performed in accordance with EPA Method 25, and samples were analyzed
for non-methane organics (NMO).
Print Room Ambient Air Measurements
VOC concentrations in the ambient air throughout the print room were
periodically monitored with a portable hydrocarbon analyzer, and
flowrates through open doorways were measured with a portable hot-
wire anemometer and a vane anemometer. Eight-hour exposure sampling
was performed at selected points in the print room using charcoal
tubes purged continuously with ambient air.
Carbon Adsorption Onit Wastewater Sampling
Samples of wastewater from the CA unit were collected during two days
of the testing program. These samples were analyzed for solvent, and
total organic carbon content.
Fabric Solvent Residue Measurements
Samples of the finished and unfinished wallcovering product were col-
lected during two days of the testing program and were analyzed for
residual solvents.
-5-
-------
1.4 Description of Report Sections
The remaining sections of this report present a summary and discussion of
test results (Section 2), description of the printing operation (Section 3),
description of the sampling locations (Section 4), and a discussion of the
sampling and analysis methods (Section 5). Field data sheets and laboratory
analysis data are presented in the various appendices, as noted in the Table
of Contents.
-6-
-------
2.0 SUMMARY AND DISCUSSION OF RESULTS
This section presents the results of the VOC emission tests conducted dur-
ing March 1981 at the GTR vinyl-coated fabric plant in Reading, Massachu-
setts. The purpose of these tests was to measure the controlled and uncon-
trolled VOC emissions from the wallcovering printing and embossing operations.
VOC measurements were performed with flame ionization detection (FID) an-
alyzers at five ducted locations: carbon adsorption (CA) unit inlet, CA out-
lets to beds 1 and 3, embosser electrostatic precipitator (ESP) inlet, and
wall fan exhaust. In addition, ambient air VOC sampling was performed in the
print-line building with a portable photoionizer hydrocarbon analyzer.,
VOC sampling was performed at the embosser ESP inlet, wall fan exhaust
duct and CA unit inlet on March 18, 19, 20, and 23, 1981. These results char-
acterized the emissions at these locations, and were used to determine the
capture efficiency of the print-line hoods. VOC sampling was performed at the
CA unit inlet and outlets on March 25 and 26, 1981, to determine the control
efficiency of the CA unit. No measurement work was performed on March 24 be-
cause no wallcovering was printed that day.
2.1 Summary of Results
VOC concentrations and air flowrates were measured at the embosser ESP
inlet, wall fan exhaust duct, CA unit inlet and CA unit outlets during print-
line operations. Ambient air VOC measurements were made inside the print-line
building (print room). The results of this measurement program showed that:
1. Under the operating conditions of this measurement program
(make-up air fan and wall exhaust fans off), the majority of
print-line VOC emissions is ducted to the CA inlet.
2. The VOC ducted to the embosser ESP inlet is a combination of em-
bosser-generated VOC and ambient print room VOC.
-7-
-------
3. The amount of print room ambient air VOC that is ducted to the
embosser ESP inlet appears to be small, but can only be estimated
from the results of this program.
4. The control efficiency of the CA unit carbon beds is approxi-
mately 98 percent averaged over times when the print-line was
operating and not operating, and approximately 99 percent aver-
aged over just those times when the print-line was operating.
2.2 Duct Flowrate Measurements
Velocity traverses were performed periodically at the VOC sampling loca-
tions during each measurement day. The flowrates measured at the embosser ESP
inlet, wall fan exhaust duct, CA unit inlet and CA unit outlets (carbon beds 1
and 3) are shown in Table 2-1. No flow existed in the wall fan exhaust duct
except during the afternoon of March 19, 1981. Prior to about 1330 on March
19, the fan motor was on but the fan belt was slipping. When the belt was
tightened flow in the duct was about 10000 SCFM. After March 19 the wall fan
was kept off in order to maximize the VOC loading to the CA unit.
VOC measurements were discontinued at the embosser ESP inlet after
March 23, 1981. However, velocity traverses here were continued on March 25
and 26. No velocity traverses were performed at the CA unit outlets until
March 25, when the VOC sampling was begun at this location.
The moisture content of the duct gases at the embosser, wall fan and CA
inlet was estimated from ambient air temperature measurements made in the
print room. These measurements indicated a very low moisture content (less
than 0.5 percent) and the gases in these ducts were considered dry. A mois-
ture content of 5 percent was assumed for the CA outlet, and Table 2-1 shows
CA outlet flowrates corrected and uncorrected for moisture content.
Dry flowrates at the CA outlets exceed those at the CA inlet by 3 to 17
percent. According to plant personnel this is expected because the humidity
controllers do leak, and because excess air pressure in the water/solvent dis-
-8-
-------
TABLE 2-1
PLOWRATE MEASUREMENTS AT THE PRINTING OPERATION
AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Embosser
Wall Fan
CA Inlet
CA Outlet
Date
3-18-81
Time
1030
1500
1600
Plowrate
3440
3200 3350*
3410
Time Flowrate
No
Measurable
Flow
Time
0900
1115
1515
Flowrate
10800
8310 8383*
8450
Time Flowrate
No
Meaisurments
Until 3-25
3-19-81
0930
1100
1645
3270
3330 3280*
3250
1415 10100 1000
1530 10000 1005°* 1100
1440
1508
1611
3-20-81 1015 3370
1340 3430 340°*
1515 4340
3-23-81 1020 3390
1225 3260 333°*
1700 2790
3-25-81 1135 3340
3-26-81 1050 2840
1345 3200
1500 3230 322°*
Wall Fan
Off
1015
1050
1310
0955
1300
1410
1725
1000
1200
1010
1225
1420
1445
1600
8330
8190
7500
7700 7600*
6830
8690
8150 8550*
8810
8650
8560
8780 "20*
8900
8140
7600
8410
Wet Dry**
1055 $1000 8550
1150 9970 9470
1020 10000
8840 1040 10400 9660*
8890 8940* 1150 10100
9080 1237 9630
8330 1400 9900
1455 9870 9290*
1545 9700
* Average of indicated flowrates.
** Estimated 5% moisture in CA outlet. Dry » 0.95 wet. Duct gases at em-
bosser, wall fan and CA inlet were considered dry. Dry flowrates were
needed in association with the Method 25 VOC test results which were on a
dry basis.
-9-
-------
tillation column is vented to a point just upstream of the CA unit fan (be-
tween the humidity controller and the fan).
All flowrates shown in Table 2-1 were used with measured VOC concentra-
tions in order to compute VOC mass emissions, as described in Section 2.3.
Since flowrate measurements were not made continuously, some judgements were
made to estimate the occurrence and significance of flowrate changes. Average
flowrates were calculated over two or more consecutive flowrate measurements
if flowrate changes were within approximately ±10 percent. This represents a
reasonable estimate of the accuracy of the flowrate measurement technique (EPA
Method 2 with standard pitot tubes) and a reasonable estimate of random, nor-
mal flowrate fluctuations within each duct (unaccompanied by significant pro-
cess changes). The process log (Section 3.6) and the consistency of the FID
analyzer strip chart traces were reviewed to determine if process changes were
occurring which could account for any indicated flowrate changes.
2.3 VOC Measurements
2.3.1 FID Sampling
Continuous monitoring of VOC concentrations was performed at each sampling
location with FID analyzers calibrated with propane and MEK standards. The
purpose of the monitoring was to determine the distribution of print-line VOC
emissions to each sampling location, the control efficiency of the carbon
beds, and the capture efficiency of the print room air management system.
The calculated VOC emissions at each sampling location are shown in
Table 2-2. The mass of VOC (pounds MEK) passing a sampling location during a
given time interval was calculated by multiplying the VOC concentration as MEK
during that time interval by the flowrate (measured as described in Section
2.2). As noted in Section 2.2, there was no flow in the wall fan exhaust duct
-10-
-------
except during the*afternoon of March 19, 1981, and except during a few brief
periods when the print-line was down. VOC monitoring at the CA unit outlets
was performed on March 25 and 26 only. Throughout the measurement program
only carbon beds 1 and 3 were operating.
"Total Run Time" includes color matching and other preparatory activities
in addition to the time necessary to print the product; "Total Print Time"
includes just the time spent to print the product. Both of these time periods
include times when the print-line was not running (for various reasons, in-
cluding maintenance and adjustments). The "1000 Yards Printing" consists of
the continuous time periods in which 1000 yards of product was printed.
Table 2-2 presents the calculated control efficiencies of the carbon ad-
sorption unit for each of the indicated time periods on March 25 and 26. Dur-
ing the VOC monitoring on these two days the two operating carbon beds were
continually cycling from an adsorbing mode to a desorbing mode. The FID anal-
yzer probe was always moved to the adsorbing bed outlet. One cycle lasted for
about 120 minutes: 60 minutes for adsorption and 60 minutes for desorption.
Since the cycling times of each bed did not always coincide with the print-
line operation time intervals in Table 2-2, the indicated control efficiencies
are essentially composite efficiencies for both beds. Control efficiences for
each individual bed, based on the observed cycling times, .are presented in
Table 2-3.
The control efficiencies for "Total Print Time" and "Total Run Time" in
Table 2-2 are slightly higher than the average control efficiencies in Table
2-3. This is because the data in Table 2-3 include time intervals in which
control efficiencies were noticeably low. For example, at 1108-1217 on
March 25 the VOC loading at the CA inlet was relatively low (7.69 pounds),
because the print-line was down in this interval and because the wall fan was
-11-
-------
TABLE 2-2
SUMMARY OP FID VOC EMISSIONS FROM PRINTING OPERATIONS
AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Production
Date Order Nuaber Process Operations
3-18-81 T-145B2
NHi Not neaeured -
a Start tine for
Preparation
Leader Threading
Color Matching
1000 Yards Printing
1000 Yards Printing
1000 Yards Printing
1000 Yards Printing
1000 Yards Printing
Completion of Run
Threading New Leader
Clean Up
Clean Up
TOTAL PRINT TIME
TOTAL RUN TIME
Tine Interval Total
Start"
0915
1035
1043
1401
1423
1445
1507
1S29
1551
1607
1613
1618
1401
1043
analyser problems or calibrations
the Initial tine Interval
Is the
End
1035
1043
1401
1423
1445
1507
1529
1551
1607
1613
1618
1640
1607
1613
Minutes
80
8
198
22
22
22
22
22
16
6
5
22
126
330
VOC
Embosser
NM
0.15
39.20
2.74
2.55
2.32
2.33
2.49
1.73
0.69
0.58
2.23
14.2
54.0
Ealsslona (Pounds as HER)
Hall Fan CA Inlet
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2.57
0.60
76.5
34.8
23.5
24.4
23.8
23.2
16.5
4.27
3.56
NM
146.2
230.5
Total
__
0.75
115.7
37.5
26.0
26.7
26.1
25.7
18.2
4.96
4.14
160.4
281.0
In progress.
tine
when FID BK
>nltorlng
began that
day.
-------
TABLE 2-2 (Continued)
SUMMARY OF FID VOC EMISSIONS FROM PRINTING OPERATIONS
AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
I
M
U)
Date
3-19-81
KM I Not
Production
Order Nuaber Procesa Operations
T-1S626 Printing in Progress
Printing
Stop and Start
1000 Yards Printing
Stop and Start
1000 yards Printing
TOTAL PRINT TIME
TOTAL RUN TINE
T-15523 Preparations Cor Nest Run
Color Matching
Eabosser Repairs. Hall
Fan On
1000 Yards Printing
1000 Yards Printing
1000 Yards Printing
1000 Yards Printing
Run Completed
Clean Up
Clean Up
Clean Up
TOTAL PRINT TIME
TOTAL RUN TIME
Tine Interval Total
Start0
0734
0848
08S4
0908
0930
0945
0848
0848
1007
1230
1332
1420
1442
1504
1526
1550
1610
1628
1632
1420
1230
easured - analyser problems or calibrations
a Start tl*« for the Initial tlmo Interval la the
End
0848
0854
0908
0930
0945
1007
1007
1007
1230
1332
1420
1442
1504
1526
1550
1610
1628
1632
1634
1610
1610
VOC
Ealsslons (Pounds as MEK)
Minutes BBbosser Hall Fan CA Inlet
74
6
14
22
15
22
79
79
143
62
48
22
22
22
24
20
18
4
2
110
220
8.53
0.902
1.50
2.65
0.781
2.47
8.30
8.30
2.49
4.14
4.41
4.05
3.85
3.99
4.25
3.58
1.14
0.18
0.09
19.7
28.3
0
0
0
0
0
0
0
0
0
0
6.24
2.43
2.96
3.20
3.49
3.14
2.25
0.48
NM
15.2
21.5
NM
4.09
6.22
8.36
3.04
11.6
33.3
33.3
7.28
10.6
15.1
9.22
10.3
10.0
10.9
7.09
10.2
NM
NM
47.5
73.2
Total
_..
4.99
7.72
11.0
13.8
14.1
41.6
41.6
9.77
14.7
25.7
15.7
17.2
17.2
18.7
13. tt
13.6
82.4
123.0
in progress.
tine
when FID
onltoring
began that day.
-------
TABLE 2-2 (Continued)
SUMMARY OF FID VOC EMISSIONS FROM PRINTING OPERATIONS
AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Production
Date Order Number Process Operations
3-20-81 T-1SS21 Completing Previous Run
Completing Previous Run
Preparation Cor T-15521
Color Matching
Printing Start/Stop
(or Repairs
Printing, Embosser On
1000 Yards Printing
Printing Start/Stop
for Repairs
1000 Yards Printing
1000 Yards Printing
1000 Yards Printing
Run Completed
Cleaning Print Heads
Clean Up
TOTAL PRINT TIME
TOTAL RUN TIME
Time Interval
Start*
0740
0744
0814
0958
1019
1148
1150
1212
1256
1318
1340
1356
1402
1426
1019
0958
End
0744
0814
0958
1019
1148
1150
1212
1256
1318
1340
1356
1402
1426
1532
1410
1410
Total
Minutes
4
30
104
21
89
2
22
44
22
22
16
6
24
66
231
252
VOC Emissions (Pounds as HER)
Embosser
0.67
3.42
4.56
1.91
7.97
0.233
4.32
6.76
4.27
4.25
3.22
1.15
3.35
6.06
33.4
35.3
Mall Fan
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
CA Inlet
NM
19.4
14.7
31.2
35.6
1.16
18.2
29.5
17.27
17.6
13.6
5.30
13.10
NM
143.8
175.0
Total
H
22.8
19.3
33.1
43.6
1.39
22.5
36.3
21.5
21.8
16.8
6.45
16.45
177.2
210.3
NMi Not measured - analyser problems or calibrations In progress.
Start time (or the Initial time Interval Is the time when FID monitoring began that day.
-------
TABLE 2-2 (Continued)
SUMMARY OF FID VOC EMISSIONS FROM PRINTING OPERATIONS
AT GENERAL TIRE AND ROBBER COMPANY
READING, MASSACHUSETTS
U1
I
Production
Date Order Number Process Operations
3-23-81 T-15516 Printing In Progress
1000 Yards Printing
1000 Yards Printing
1000 Yards Printing
Run Completed
TOTAL PRINT TIMB
TOTAL RUN TIMB
T-15519 Threading Leader
Cleaning. PU Fans off.
Hall Fan on
Color Matching,
Heb Alignment
Hall Fan Off.
Color Matching
Printing Line Down Once
1000 Yards Printing
Line Up and Down.
Trimming Problems
Problems Persist.
Run Ended
Repairs
Repairs
TOTAL PRINT TIMB
TOTAL RUN TIMB
Time Interval
Start"
0850
0909
0931
0953
1015
0850
0850
1025
1037
1239
1244
1324
1351
1413
f
1605
1628
1633
1324
1239
End
0909
0931
0953
1015
1025
1025
1025
1037
1239
1244
1324
1351
1413
1605
1628
1633
1636
1605
1628
Total
Minutes
19
22
22
22
10
95
95
12
122
5
40
27
22
112
23
5
3
161
229
VOC
Emissions (Pounds as MEK)
Embosser Hall Fan CA Inlet
1.92
2.60
2.78
3.12
1.48
11.9
11.9
1.41
5.25
0.070
0.85
1.88
1.56
6.71
1.14
0.27
NM
10.2
12.2
Sampling 10.51
Discontinued 13.06
14.70
14.20
7.15
59.6
59.6
24.60
4.17
0.180
10.3
9.90
9.02
40.0
10.5
1.16
1.93
58.9
69.6
Total
12.43
15.7
17.5
17.3
8.63
71.5
71.5
26.0
9.42
0.250
11.2
11.8
10.6
46.7
11.6
1.43
69.1
81.8
NMi Not measured - analyser problems or calibrations In progress.
Start time for the Initial time Interval Is the time when FID monitoring began that day.
-------
TABLE 2-2 (Continued)
SUHHABY OF flD VOC EMISSIONS FROM PRINTING OPERATIONS
AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
VOC Emissions
Production TJMe Interval
Date Order No. Process Operations Start* End
Total
Minutes
(Pounds
Cft Inlet
As HER)
Ca Outlet
CA Unit
Control
Efficiency (%)b
Carbon Bed
In Operation
(Adsorbing)
Tine of Bed
Switch0
3-24-81 Eabossing Entire Day
3-25-81 T-15511 Color Matching 0859 0900
*
**
a
b
c
MM!
Printing in Progress 0900 0922
Printing 0922 0942
1000 Yards Printing 0942 1003
1000 Yards -Printing 1003 1020
1000 Yards Printing 1020 1037
Run Completed 1037 1047
Leader Threading 1047 1108
Nail Fan on Preparation
For Next Run 1108 1217
Hall Fan on Preparation
For Next Run 1217 1230
TOTAL PRINT TIME 0922 1047
TOTAL RUN TIME 0922 1108
1
22
20
21
17
17
10
21
69
13
85
106
0.66
20.1
21.8
23.5
21.2
22.7
10.7
15.7
7.69
NM
130
146
NM
NM
0.048
0.071
0.067
0.093
0.065
0.114
0.317
0.053
0.344
0.458
..
99.8
99.9
99.7
99.6
99.4
99.3
95.9
99.7
99.7
3
3
1
1
1
3
3
3
3/1
3
0922*
1117**
Beginning tlaa of 0922 was estimated, based on observed end ti«e of 1020
Bed No. 1 began adsorbing 1117 and continued to 1217.
Start time for the initial tine Interval is the time when FID i
100 ( 1- (OUTLET/INLET)!
Nominal bed cycle (adsorptlon/desorptlon) is about 120 Minutes,
Not Measured - analyzer problems or calibrations in progress.
onltorlng
began that
day.
-------
TABLE 2-2 (Continued)
SUMMARY OF FID VOC EMISSIONS FROM PRINTING OPERATIONS
AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Production rime interval
Date Order No. Process Operations Start* End
3-26-81 T-15508 Preparation 0856
Preparation 0936
Color Matching 0939
Printing 1059
1000 Yards Printing 1126
1000 Yards Printing 1151
Run Completed 1216
Embosser oft. Clean up 1222
TOTAL PRINT TIME 1059
TOTAL RUN TIME 0939
T-15507 Preparation 1229
Color Matching 1326
Printing. Line Down
Once 1420
1000 Yards Printing 1439
1000 Yards Printing 1501
Run Completed 1523
Line Down. Preparation
for next run 1540
Line Down. Preparation
for next run 1612
TOTAL PRINT TIME 1420
TOTAL RUN TIME 1326
0936
0939
1059
1126
1151
1216
1222
1229
1222
1229
1326
1420
1439
1501
1523
1540
1612
1614
1540
1540
Total
Minutes
40
3
80
27
25
25
4
7
83
163
57
54 .
19
22
22
17
32
2
80
134
VOC Emissions
(Pounds As MEK)
CA Inlet
NM
0.880
37.0
21.1
20.7
21.6
5.15
5.11
72.6
110.7
28.8
31.6
13.8
16.9
17.7
12.8
15.0
NM
61.2
92.8
Ca Outlet
0.194
0.036
1.26
0.356
0.313
0.313
0.072
0.156
1.05
2.47
0.588
0.438
0.158
0.211
0.160
0.116
0.217
0.014
0.645
1.08
CA Unit
Control
Efficiency (%)b
95.9
96.6
98.3
98.5
98.6
98.6
96.9
98.6
97.8
98.0
98.6
98.9
96.8
99.1
99.1
98.6
98.9
98.8
Carbon Bed
In Operation
(Adsorbing)
1/3
3
3/1
1
1/3
3
3
3
3/1
1/3
3/1
1
1
1/3
3
3
Time of Bed
Switch0
0928
1028
1128
1233
1334
1435
1534
8 Start tine for the initial time Interval is the time when FID monitoring began that day.
b 100 I1-(OUTLET/INLET)I
c Noalnal bed cycle (adsorptlon/desorptlon) is about 120 minutes.
-------
TABLE 2-3
CARBON ADSORPTION UNIT CONTROL EFFICIENCIES
AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
VOC Measurement Carbon
Adsorption Time at Both Bed in
VOC Emissions
(pounds as MEK)
Date Time Inlet and Outlet Operation CA Inlet CA
3-25-81 0922*-1020 same
1020 -1117 same
1117 -1217 same
3-26-81 0928-1028 0936-1028
1028-1128 same
1128-1233 same
1233-1344 same
1344-1435 ° ' same
1435-1534 same
1534-1635* 1534-1612
1
3
1
3
1
3
1
3"
1
3
96.5
50.8
6.02
average
21.6
38.8
54.1
34.4
34.7
45.3
19.2
Outlet
0.186
0.321
0.268
0.771
0.910
0.805
0.752
0.423
0.479
0.258
average:
Control
Efficiency (%)a
99.8
99.4
95.5
98.2 -
96.4
97.7
98.5
97.8
98.8
98.9
98.7
98.1
a [1-(outlet/inlet)] 100
* estimated times
-18-
-------
turned on at 1108. The flowrate at the CA inlet did decrease from 8140 SCFH
at 1000 to 7600 SCFH at 1200 (see Table 2-1), and the measured VOC concentra-
tions in the inlet decreased from about 310 ppra (as MEK) at 1101 to about 110
ppm at 1109 (see Appendix B). The VOC loading at the CA outlet in this time
interval is relatively large (0.317 pounds), but only because the time inter-
val itself is much larger (69 minutes) than other time intervals on this day.
The relatively low control efficiency from 0936 to 1059 on March 26 is due
£
only to the fact that the VOC concentrations in the CA inlet at this time were
relatively low (190-370 ppm) compared to subsequent time intervals (420-540
ppm) (see Appendix B). Reduced control efficiency as a result of reduced in-
let concentrations is a general characteristic of carbon bed adsorption units.
2.3.2 NMO Sampling with Method 25
Sampling for non-methane organic (NMO) compounds was performed concur-
rently with the continuous FID monitoring at the embosser ESP inlet, wall fan
exhaust duct, CA unit inlet and CA unit outlets. Integrated gas samples were
collected once each monitoring day (except March 18) over periods of time
ranging from 18 minutes to 56 minutes using an evacuated tank and condensible
trap in accordance with EPA Method 25. Tank and trap contents were then anal-
yzed at the TRC laboratory for NMO expressed as carbon. The results of these
analyses are presented in Table 2-4. Details of the results are presented in
Appendix D.2.
There was relatively good agreement between the NMO duplicate sample re-
sults for combined traps and tanks. However, there appears to be no correla-
tion between the results obtained from the duplicate traps alone, or between
the results obtained from the duplicate tanks alone. The difference between
the paired sample collection components may reflect the effects of the ambient
-19-
-------
temperature trap purge procedure. The volatility of the solvents is such that
at room temperature some of them may be purged into the sample tank. The
quantity purged will vary, depending on factors such as trap temperature
before insertion into the bath and how long the trap remained in the bath
Before being returned to room air.
There is poor correlation also between NMO Method 25 results and FID
Method 25A results. There is no definite explanation for this poor correla-
tion, though it may be due to problems in the NMO Method 25 analytical pro-
cedures. The FID results are considered more reliable since the FID analyzers
were calibrated and audited daily.
2.4 Print Room Ambient Air Measurements
Measurements of ambient air VOC concentrations and doorway air flowrates
were performed in the print room periodically each day. The results of these
various measurements are discussed in the following subsections.
2.4.1 Ambient Air VOC Measurements
VOC measurements were made in the immediate vicinity (within 5 feet) of
the print heads and the embosser, and at locations throughout the print room,
using a portable photoionizer hydrocarbon analyzer sampling at approximately
breathing level. A summary of these measurements is shown in Table 2-5. All
measured concentration data are presented in Appendix £.
The "Print Room" MEK concentrations shown in Table 2-5 represent averages
of measurements taken at 10 or 12 locations throughout the print room. The
"Print-line" MEK concentrations represent averages of measurements taken with-
in 5 feet of the print heads. These measurements were taken at times between
print heads, at times approximately 2 feet away from the print heads, and at
-20-
-------
TABLE 2-4
SUMMARY OF NMO METHOD 25 ANALYSIS RESULTS
GENERAL TIRE AND RUBBER COMPANY, READING, MASSACHUSETTS
Sampling Time
Total Concentration
Date Location Start End Minutes Sample A
3-19-81 Wall Fan 1449 1534
Embosser 1449 1534
CA Inlet 1449 1534
3-20-81 Embosser 1301 1346
CA Inlet 1301 1346
3-23-81 Embosser 1334 1337
1410 1423
1554 1557
CA Inlet 1334 1337
1410 1423
1555 1557
3-25-81 CA Outlet 1022 1049
CA Inlet 1022 1049
3-26-81 Embosser 1105 1150
CA Outlet 1105 1119
1123 1205
1205 1231
1249 1312
CA Inlet 1105 1119
1130 1201
* Duplicate sample
f Filter in sampling probe
** Time-weighted average
NM Not measured
45 380
45 918 £
45 1,492
45 913 f
45 1,104
3
13
_3
19 642
3
13
2
18 453
27 210
27 2,113
45 68
14
42_
56 880 a
26
23
49
14
31
45 1,226
(ppm as carbon)
Sample B*
NM
833
2,187
923
1,973
279
317
492
2,314
NM
719 a
1,275
FID**
Concentration
(ppm as MEK)
75
291
322
310
501
109
256
4
781
NM
7
506
Two seperate tanks were used for this run for each sample - one tank from
1105 to 1205, a second tank from 1205 to 1312.
-------
times approximately 5 feet away from the print heads. The "Embosser" MEK con-
centrations represent averages of measurements made approximately 2 feet in
front of the cage surrounding the embosser, at five locations along the cage.
The embosser measurements are discussed further in the following sub-section.
2.4.2 VOC Measurements at the Embosser
The primary objective of performing ambient air VOC measurements near the
embosser was to compare the VOC concentration of air near the embosser to the
VOC concentration measured by FID analyzer in the embosser ESP inlet. This
comparison is presented in Table 2-6.
These data indicate that VOC concentrations in the embosser ESP inlet duct
are always higher than VOC concentrations in the print room ambient air near
the embosser, except perhaps at times when the embosser itself is off. The
embosser ESP fan was on at all times during the measurement program.
Continuous VOC monitoring in the ESP inlet was discontinued after March
23, 1981. Ambient air measurements at the embosser were continued on March 25
and 26, as shown on Table 2-5. On March 25 at 0910 and 0955, with the em-
bosser off, ambient air VOC concentrations at the embosser were 170 ppm and
125 ppm, respectively. At 0925 however, a measurement was made at the center
of the ESP inlet duct with the portable photoionizer hydrocarbon analyzer.
Over approximately a one-minute time interval concentration readings ranged
from 1 to 20 ppm, averaging approximately 3 or 4 ppm. These duct concentra-
tions are inconsistent with the print room ambient air concentrations, if one
assumes that the ambient air near the embosser is being drawn into the ESP
inlet through the hood that hangs over the embosser.
Once on March 25 (1046) and once on March 26 (1159) ambient air flow near
the embosser was monitored with a sensitive vane anemometer. No air motion
-22-
-------
TABLE 2-5
SUMMARY OF PRINT ROOM AMBIENT AIR SURVEYS AT
GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
to
LJ
Print-Line Operations
Date
3-I8-B1
3-19-81
3-20-81
3-23-81
Time
1440-1510
1520-1540
0850-091S
0916-0926
1401-141S
1458-1513
1037
1045-1055
1115
1130
1155
1215-1225
1240
1315
1322-1332
1355
1400-1406
1447
0923
0925
1015
1022-1028
1340
1417
1420-1430
1548
1610
1630-1640
Order
No.
T-14582
T-14582
T-15626
T-15626
T-15523
T-15523
T-15521
T-15521
T-15521
T-15521
T-15521
T-15521
T-15521
T-15521
T-1S521
T-15521
T-1S521
T-15516
T- 15516
T-15516
T-15516
T-15519
T-15519
T-15519
T-15519
No. Print
Heads
5
5
1
1
4
4
4
4
4
4
4
4
4
4
4
4
4
5
5
5
5
3
3
3
3
Embosser
ON
ON
ON
ON
ON
ON
OFF
OFF
OFF
OFF
ON
ON
ON
ON
ON
ON
ON
OFF
ON
ON
ON
OFF
ON
ON
ON
ON
General
Printing
Printing
Printing
Printing
Printing.
Hall fan on
Printing.
Wall Can on
Color Batching
Color Batching
Color Batching
Color Batching
Printing
Printing
Printing
Printing
Printing
Printing
Printing
Line down.
Cleaning heads
Printing
Printing
Printing
Line down
briefly
Printing
Printing
Line up and down
Trinalnq problems
Line down
Printing.
Testing leader
Average MBK Concentration (ppm)
Print ROOD" Print-Line0 Embosser15
120 280 140
90 210 105
50 95
130
95
150
125
140 250 150
130
130 170
135
140
105
130 135
70
100
90
95 210
Doorway B KEK Concentration
Flowrate Outside Doorway B
(SCFM) (PP»)
5680
3860 (D)
»*2930 (B)
12470 (Total)
6080
20100
2430d
8510
9120
6690e
8510
9420
7300
8816
4
70 (D)
20 (B)
2
5
4
2
3.5
3
3
2
3
3
MEK Mass Flowrate
Into Print Room
(Pounds/Hour)
0.26
3.04
0.66
3.96
0.13
1.12
0.11
0.19
0.36
0.22
0.29
0.21
0.25
0.30
(D)
(B)
a Average of Beasureaents Bade at 5 feet above the floor, throughout printrooa. Does not include neasureaents Bade near open doorways.
D Average of measurements Bade 2 to 5 feet fron print heads at 5 feet above the floor.
° Average of measurements Bade about 2 feet In front of the embosser at 5 feet above the floor.
d OH door open about 10 Inches.
* Print head fans were off during these doorway measurements
* Measured in Doorway D.
Measured in Doorway B.
-------
TABLE 2-5 (Continued)
SUMMARY OF PRINT ROOM AMBIENT AIR SURVEYS AT
GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Print-Line Operations
Date Tine
3-24-81
3-25-81 0910
0950
0955
0958-1006
1046
1105
1240-1257
1422
1450
3-26-81 0950-1024
0952
1023
1026
1115-1125
1138
1159
1355
1426
1445-1456
1532
Order
No.
T-15511
T-15511
T-15511
T-15511
T-15511
T-15511
T-15508
T-15508
T-15508
T-15S08
T-15508
T-15508
T-15508
T-15507
T-15507
T--15507
T-15507
No. Print
Heads
4
4
. 4
4
4
4
5
5
5
5
5
5
5
5
5
5
5
EBboaser
ON
Off
OFF
OFF
OFF
ON
OFF
OFF
ON
ON
ON
ON
ON
ON
ON
ON
ON
OFF
ON
ON
ON
General
Eobosalng
all day
Printing
Printing
Printing
Printing
Printing
Line down
Line down. '
Floor cleaning
with HER
Embossing only
Embossing.
Wall fan on
Color Batching
and printing
Color natchlng
Printing
Printing
Line up and down
Printing
Printing
Line down
Printing
Printing
Printing
; Doorway B MEK Concentration MEK Mass Flowrate
Average MEK Concentration (pp») Flowrate Outside Doorway B Into Print Room
Print Room" Print-Line0 Embosser0 (8CFM) (ppn) (Pounds/Hour)
170
6690 2 0.15
125
110 225
150
130
SO 60 170
32
21300 2.5 0.60
65 250
200
190
10000 2 0.22
120 175 95 10300 2 0.23
120
100
65
125
110 200 110
80
a Average of neaaurementa Bade at 5 feet above the floor, throughout print room. Does not include neaaurementa Bade near open doorways.
b Average of measurements Bade 2 to 5 feet fro» print heads at 5 feet above the floor.
c Average of Beasurenenta Bade about 2 feet in front of the enbosser at 5 feet above the floor.
-------
TABLE 2-6
DUCT AND AMBIENT AIR VOC MEASUREMENTS AT THE EMBOSSER
AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Date
3-18-81
3-19-81
3-20-81
3-23-81
Time
1506
0853
1037
1115
1155
1215
1315
1355
0923
1026
1417
1548
Embosser
Operation
On
On
Off
Off
On
On
On
On
On
Off
On
On
VOC Concentrations
(ppm as MEK)
Ambient
140
105
130
150
125
150
130
135
105
135
70
90
Duct3
173
245
168
118
309
320
309
316
198
178
110
121
Duct Flowrate
Measurements'3
Time SCPM
1500 3200
0930 3270
1015 3370
1340 3430
1020 3390
1225 3260
1700 2790
Measurements made in embosser ESP inlet. No duct measurements at this
location were made 3-25 and 3-26.
See Table 2-1.
-25-
-------
was observed either horizontally or vertically up to 8 feet above the floor
and within the cage surrounding the embosser. At both times the embosser was
operating. These vane anemometer readings and the low in-duct VOC concentra-
tions observed at 0925 on March 25 indicated that air flow into the embosser
hood is localized at the level of the hood. At this height VOC concentrations
may be lower than at 5 feet above the floor; the VOC concentrations measured
at 1115 on March 20 support this hypothesis (Table 2-6).
2.4.3 Doorway Flowrates
Air speed through open print room doorways were measured with a hot-wire
anemometer and a vane anemometer. Average speed was multiplied by the doorway
area to obtain flowrate. The results of the air flow measurements are shown
in Table 2-5.
Doorway E (at the northeast corner of the print room, connecting the print
room to the main plant) was fully open during the entire measurement program.
All other doorways were closed, except for doorways B and D on March 18 which
were partially open during that day. The wall fan was off except for the
afternoon of March 19, and except for occassional brief periods when the
print-line was down and the print room needed ventilation.
Air flowrates through doorway E ranged from about 2430 SCPM to 21300
SCFM. Excluding the flowrate measurement on March 18 (when doorways B and D
were also open), the low measurement of 2430 SCFM (when the overhead door was
open 10 inches), and the flowrates measured when the wall fan was on (20100
SCFM and 21300 SCFM), the air flowrate through doorway E ranged from 6080 SCFM
to 10300 SCFM. During all measurement periods the direction of air flow was
into the print room.
Considering measured flows into and out of the print room, with the wall
fan both on and off, a general flow balance does exist (see Tables 2-1 and
-26-
-------
2-5). During the times the wall fan was on and doorway £ flow measurements
were made (1400 on March 19 and 1450 on March 25), measured flow into the
print room was about 21000 SCFM. Measured flow out of the print room at this
time was about 3300 SCFM at the embosser plus about 10000 SCFM at the wall fan
plus about 7500 SCFM at the CA inlet, or a total of about 20800 SCFM. During
the times when the wall fan was off, flow at the embosser still averaged about
3300 SCFM and flow at the CA inlet averaged about 8500, for a total flow of
approximately 11800 SCFM out of the print room. Air flow into the print room
when the wall fan was off ranged from 6080 SCFM to 10300 SCFM, as noted
above. These latter flowrates indicate that miscellaneous air flowst into the
print room (around other closed doors and general building leakage) are sig-
nificant when the wall fan is off.
During each flow measurement period, the ambient air VOC concentration at
approximately 5 feet outside doorway E (in doorways E, B and D on March 18)
was measured about 5 feet above the floor with the portable hydrocarbon analy-
zer. These concentrations ranged from 2 to 5 ppm (as MEK) in doorway E.
Higher concentrations were measured in doorways B and D probably because these
doorways connect the print room directly to areas of the main plant where inks
are mixed. The VOC mass flowrate into the print room through open doorways
was estimated by multiplying these doorway concentrations by the measured
doorway flowrates. These mass flowrates are shown in Table 2-5.
Xn general a steady north-to-south air flow of about 300-400 feet per
minute existed just outside doorway E. Air drawn into the print room had to
make a sharp right turn, so the highest air speeds were usually measured at
the south end of doorway E.
2.4.4 Eight-Hour Exposure Sampling
Continuous eight-hour VOC sampling was performed at four locations near
-27-
-------
the print-line using battery-operated personal samplers and charcoal tubes.
The four sampling locations were:
1. on the cage in front of the embosser, about 5 feet above the
floor and 3 feet from the embosser;
2. same as location 1 but in back of the embosser;
3. in front of print head No. 6, on the side of the stairs leading
to a catwalk above the print-line, about 6 feet above the floor;
4. above print head No. 2, about 2 feet below the fan air intake,
about 17 feet above the floor.
Eight-hour samples were taken on March 23, 25 and 26 at these four locations.
During each eight-hour period, charcoal tubes were replaced every two hours.
The results of the charcoal tube solvent analyses are shown in Table 2-7.
Sampler malfunctions on March 23 led to poor data recovery on that day, so
only the samples from March 25 and 26 were analyzed.
The results from the carbon tube analyses indicate generally lower concen-
trations than the room air survey results made with the portable hydrocarbon
analyzer. There is no definitive explanation for this difference. However,
the carbon tube samples were taken over a continuous eight-hour period. The
print-line was not always operating during these times, so lower carbon tube
concentration could be expected. Also, VOC was found in most back-up sections
of the carbon tubes. This indicates that VOC break-through may have occurred
with the result that the reported data may be lower than actual values.
2.5 Carbon Bed Wastewater Samples
Wastewater samples from the carbon absorption unit were collected period-
ically on March 25 and 26 during the continuous VOC monitoring at the CA unit
inlet and outlets. Nine samples were taken from a common drain at the base of
the carbon beds (bed condensation samples) and 12 samples were taken from a
-28-
-------
TABLE 2-7
BIGHT-HOUR SAMPLING DATA IN TUB PRINT ROOM AT
GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Sampling Tina Total Tine Volume of Ale HEK MIBK Toluene
Date Location Start End (minutes) Sampled (liters) mq ppm(v/v) ag ppm(v/v) ng ppa(v/v)
3-23-81 Front Embosser 0907 1107 120
1107 1307 120
1307 1507 NM
1507 1707 NM
0907 1307 240
Back Embosser 0909 NM
Ladder 0903 1103 120 SAMPLES NOT ANALYIBD
1103 1304 121
1304 1507 123
1507 1730 143
0903 1730 527
Above PH 2 0911 1111 120
1111 1311 120
1311 1511 120
I 1511 1732 141
to
f 0911 1732 521
NM - No measurement due to sampler malfunction.
-------
TADLB 2-7 (Continued)
BIGHT-HOUR SAMPLING DATA IN THE PRINT ROOM AT
GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
CO
O
SaBpllng Time Total Tine
Date Location Start
3-25-81 Front Embosser 0841
1037
1237
1437
0841
Back Embosser 0845
1039
1239
1439
0845
Ladder 0834
1030
1230
1430
0834
Above pH 2 0830
1033
1233
1433
0830
End (alnutes)
1037
1237
1437
1637
1637
1039
1239
1439
1639
1639
1030
1230
1430
1630
1630
1033
1233
1433
1633
1633
116
120
120
120
476
114
120
120
120
474
116
120
120
120
476
123
120
120
120
483
Volute of Air
MEK
Sampled (liters) mg ppm(v/v)*
91.6
94.7
97.9
82.8
367.0
97.8
116.1
116.1
109.6
439.6
140.6
156.5
156.5
151.0
604.6
41.0
40.0
40. Q
40.0
161.0
1.39
8.48
11.86
6.30
28.03
17.41
14.91
1.56
3.97
37.85
14.54
6.97
10.12
3.77
37.40
3.77
2.18
1.02
0.68
11.42
5.16
30.47
41.33
25.89
25.99
60.58
43.70
4.57
12.23
29.30
35.19
19.50
22.00
8.50
21.05
31.29
18.55
8.68
5.78
24.14
MIBK
ng ppm(v/v)*
0.50
0.71
0.69
1.44
3.34
1.79
1.79
0.62
1.32
5.52
2.04
0.94
0.57
0.85
4.40
0.85
0.42
0.38
0.38
2.03
1.34
1.84
1.73
4.26
2.23
4.48
3.78
1.31
2.95
3.07
3.55
1.47
0.89
1.37
1.78
5.08
2.57
2.32
2.32
3.09
Toluene
og ppm(v/v)*
0.42
0.58
0.54
1.13
2.67
1.30
1.35
0.49
1.05
4.19
1.41
0.85
0.60
0.54
3.40
0.55
0.50
0.47
0.47
1.99
1.22
1.63
1.47
3.63
1.94
3.54
3.10
1.12
2.55
2.54
2.67
1.45
1.02
0.95
1.50
3.57
3.33
3.13
3.13
3.29
* 25C, 760
-------
TABLE 2-7 (Continued)
BIGHT-HOUR SAMPLING DATA IN THE PRINT ROOM AT
GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Sampling Tl»a Total TlM
Date Location Start
3-26-81 Front Eobosser 1007
1212
1404
1604
1007
Back Embosser 1011
1214
1406
160S
1011
Ladder 1001
1207
1409
1610
1001
Above PH 2 1004
1210
1412
1615
1004
End (i
1212
1403
1602
1803
1803
1214
1406
1608
1806
1806
1207
1409
1610
1808
1808
1210
1412
1615
1815 i
1815
ilnutes)
125
111
118
119
473
123
112
122
118
475
126
122
121
118
487
126
122
123
120
491
Volume of Air
HER
Sampled (liters) ng
116.
95.
93.
82.
387.
115.
' 96.
104.
101.
417.
189.
176.
170.
162.
698.
42.
40.
41.
40.
163.
3
6
5
1
5
0
1
7
2
0
0
0
6
8
4
0
6
0
0
6
21.
14.
10.
5.
51.
5.
10.
10.
8.
35.
11.
2.
9.
1.
24.
2.
2.
3.
0.
9.
pp»(v/v)*
18
01
75
46
40
84
26
98
80
88
03
61
33
78
75
74
48
43
99
64
61.92
49.87
39.12 .
22.63
45.10
17.28
36.33
35.69
29.59
29.28
19.86
5.05
18.61
3.72
12.04
22.20
20.79
28.47
8.42
20.05
MIBK
g ppm(v/v)»
1.54
1.07
0.06
0.31
2.98
1.05
1.57
2.69
1.10
6.41
0.58
0.76
1.23
0.76
3.33
0.32
0.24
0.39
0.14
1.09
3.24
2.74
0.16
0.93
1.88
2.24
4.00
6.29
2.93
3.77
0.75
1.06
1.76
1.14
1.17
1.86
1.45
2.33
0.85
1.63
Toluene
ng ppm(v/v)*
1.00
0.68
0.04
0.18
1.90
0.81
1.23
1.26
0.63
3.93
0.40
0.47
0.90
0.46
2.23
0.28
0.20
0.33
0.13
0.94
2.29
1.89
0.11
0.58
1.31
1.88
3.41
3.49
1.66
2.51
0.56
0.71
1.40
0.75
0.85
1.78
1.31
2.14
0.87
1.53
25 C. 760 mnllg
-------
drain at the base of the water/solvent distillation column (bottom product
samples). Under the direction of the EPA task manager/ twelve representative
samples were selected out of the 21 samples and were analyzed for MEK, MIBK,
toluene, and total organic carbon (TOG) content. The results of these analy-
ses are shown in Table 2-8.
The TOC results are always greater than the sum of the individual compon-
ent results (MEK, MIBK and toluene). This difference may be because two dif-
ferent analysis procedures were used and because the solvents may begin to
break down during the CA steam regeneration and distillation processes.
2.6 Fabric Solvent Residue
Four different wallcoverings were sampled from March 23 to March 25,
1981. The fabric solvent residue analysis results are summarized in Table
2-9. The time lag from sample collection to start of recovery, pattern
duplication information and paired sample identification information are
presented in Table 2-10.
The purpose of this part of the test program was to determine the embosser
emissions indirectly from the product samples collected before and after being
embossed. The analysis procedure used was a preliminary procedure still in
the development stage. The results do not compare well with the embosser
emissions measured in the embosser ESP inlet duct with Methods 2 and 25A, and
were therefore not used to estimate emissions.
In general, the weights are fairly consistent and reproducibility within a
given set of samples is good. Only two samples (29 and 30) had any
bleedthrough to the backup tube. The first set of samples (order no. 15516)
does not exhibit substantial differences before and after embossing. This may
be because the "before" samples were run through the heated embosser area
-32-
-------
TABLE 2-8
CARBON ADSORPTION UNIT WASTEWATER ANALYSIS RESULTS AT
GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Sample
Date No. Time
3-25-81 1
2
3
4
5
6
7
8
9
10
3-26-81 11
12
13
14
15
16
17
18
19
20
21
1021
1029
1030
1035
1050
1052
1102
1106
1214
1421
1030
1105
1135
1140
1201
1203
1240
1240
1320
1328
1525
Bed Condensation Samples
Carbon Bed (mq/2.)
Adsorbing MEK MIBK Toluene TOC
3 - - -
3
3
3
3
3 _ _
3 - -
3
1
3
1 1.3 0.2 n.d. 7.2
1 n.d. 0.1 n.d. 10.9
3 2.5 0.1 n.d. 17.6
3
3 - - _
3
1 4.8 0.2 n.d. 11.6
1
1 6.2 0.2 n.d. 13.9
1
1
Bottom Product Samples
(mq/Z)
MEK MIBK Toluene TOC
41.2 n.d. n.d. 65.4
_ - _
_ - -
_ -
_ -
15.3 0.3 n.d. 68.4
- _ _ _
..
14.5 0.1 n.d. 55.3
- -
. _ - -
38.7 n.d. n.d. 63.9
* . "
n.d. = not detectable
- a sample not analyzed
-33-
-------
TABLE 2-9 (Continued)
SUMMARY OF WALLCOVERING SOLVENT RESIDUES
GENERAL TIRE AND RUBBER COMPANY, READING, MASSACHUSETTS
MARCH 23-25, 1981
Sample
No.
26Ab
26B
29Ac>d
29B
30A°«d
30B
33A
33B
34A
34B
35A
3 SB
36A
36B
37A
37B
3BA
38B
39A
40A
Order
No.
14796
15511
15511
15516
15516
15511
15511
14796
14796
15519
15519
Condition
After embosser
Before embosser
Before embosser
Unprinted
Unprlnted
Unprlnted
Unprinted
Unprinted
Unprinted
Unprlnted
Unprlnted
Sample Size
(sq. inches)
28.41
28.16
31.30
36
36
36
36
36
36
36
36
mg MEK
Sample
0.40
0
27.02
10.21
35.83
7.10
3.49
2.12
1.54
1.62
1.60
1.80
0.04
0.03
mg MIBK
Sample
0.10
0
3.55
0
2.25
0
0.32
0.32
0.12
0.19
0.11
0.11
0.005
0.004
mg Toluene
Sample
0.07
0
2.20
0
1.37
0
0.26
0.11
0.12
0.14
0.10
0.11
0.002
0.002
lb MEK
(sq. yd.)
4.02 ,
10-5
3.78
10" 3
3.92 x
10-3"
2.77 x
io-4
1.68 x
1.22 »
10~4
1.29,
10~4
1.27
10"4
*« «
10"4
3.17 x
10-'
2.38 ,
NT6
lb MIBK
(sq. yd.)
1.01 x
ID"5
3.60 ,
ID"4
2.05 x
io-4
2-54 x
10"5
2.54 x
10~5
9.50 x
1.51.x
IO-5
8.73 x
10-6
8.70 x
3.97 ,
10"'
3.18 x
10-'
lb Toluene
(sq. yd.)
7.00 x
2.23 x
io-4
1.25 x
io-4
2.06 x
10~5
8.70 x
10~6
9.50 x
10-6*
1-11 x
ID"5
7.90 x
IO-6
8.70 x
ID"6
1.59 x
10"'
1-59 x
10-'
a Samples 1 and 2, "before embosser" samples, were not embossed, but were run through the heated embosser area. These two samples
were brought back to the recovery room, cut and loaded within ten minutes.
b These samples were placed in Hheaton Jars while awaiting the availability of the drying tubes.
0 These samples were collected, cut and loaded directly into the drying tubes.
samples have very high area counts and they are fairly consistent with each other. Although the wallcovering does not have
especially heavy coverage. It Is possible that the inks used on this covering (order no. 15511) have a particularly high solvent
content. There are no "after embosser" samples. This was the last day of ES's field testing and GTR was only printing that day,
intending to emboss that night.
-------
TABLE 2-9 (Continued)
SUMMARY OF WALLCOVERING SOLVENT RESIDUES
GENERAL TIRE AND RUBBER COMPANY, READING, MASSACHUSETTS
MARCH 23-25, 1981
Sample Order Sample Size
No. No. Condition (sq. inches)
41A 15511 Unprinted 36
42A 15511 Unprinted 36
a Samples 1 and 2, "before embosser" samples, were not
were brought back to the recovery room, cut and loaded
mg HER mg MIBK
Sample Sample
0.01 0.005
0.03 p. 001
mg Toluene
Sample
0.001
0.001
embossed, but were run through the
within ten minutes.
b These samples were placed in Hheaton jars while awaiting the availability of
c These samples were collected, cut and loaded directly
into the drying tubes.
Ib HER Ib MIBK
(sq. yd.) (sq. yd.)
7.94
2.3B
10~6
heated
X 3.97 x
10~7
x 7-948x
embosser area. These
Ib Toluene
(sq. yd.)
7.94 x
IO-8
10"8
two samples
the drying tubes.
&1 t-kmft
especially heavy coverage, it is possible that the inks used on this covering (order no. 15511) have a particularly high solvent
content. There are no 'after embosser* samples. This was the last day of ES's field testing and GTR was only printing that day,
intending to emboss that night.
I
U>
-------
TABLE 2-9
SUMMARY OP WALLCOVERING SOLVENT RESIDUES
GENERAL TIRE AND RUBBER COMPANY, READING, MASSACHUSETTS
MARCH 23-25, 1981
Sample
No.
IB
2Aa
2B
5Ab
5B
6Ab
6B
9Ab
9B
10Ab
10B
13B
14AC
14B
19B
20AC
20B
25Ab
25B
Order
No.
15516
15516
15516
15516
15519
15519
15519
15519
14796
14796
14796
Condition
Before embosser
Before embosser
After embosser
After embosser
Before embosser
Before embosser
After embosser
After embosser
Before embosser
Before embosser
After embosser
Sample Size
(sq. Inches)
36
36
36
36
33
36.56
38.88
40.06
32.87
27.25
23.51
mg HER
Sample
3.57
0
3.37
0
3.11
0
2.75
0
4.35
0
5.00
0
1.25
0
1.53
0
0.57
0
0.26
0
0.26
0
mg MIBK
Sample
1.16
0
0.89
0
0.99
0
0.90
0
1.57
0
1.88
0
0.44
0
0.46
0
0.41
0
0.45
0
0.18
0
mg Toluene
Sample
0.48
0
0.38
0
0.42
0
0.37
0
0.68
0
0.81
0
0.02
0
0.19
0
0.26
0
0.20
0
0.12
0
Ib HER
(sq. yd.)
2.83 x
2.68 ,
10~4
2-47 x
10~4
2.18 x
3.77 x
3.91 x
4"
*
9.19 x
1<0V
4.95 x
10-5*
2.73 x
10-5
3.16,
lO"5
Ib MIBR
(sq. yd.)
9.21 x
10-5*
7.06 x
ID"5
7.86 x
10~5
7.14 x
10-5*
1-36 ,
10~4
^o-**
3.23 x
10-5
3.28 x
10-5
3.56 x
10-5
4-72 ,
ID'5
2.19 x
10-5*
Ib Toluene
(sq. yd.)
3.81 x
C
ID"5
3-02 x
ID'5
3.33 x
10"5
2.94..X
5.89 x
10-5*
6.33 .
10-5"
1-50 .
10"6
1-36 ,
ID"5
2.26.
ID"5
2.10 x
10" 5
1.46 x
C
ID"5
a Samples 1 and 2, "before embosser" samples, were not embossed, but were run through the heated embosser area. These two samples
were brought back to the recovery room, cut and loaded within ten minutes.
b These samples were placed in Mheaton jars while awaiting the availability of the drying tubes.
c These samples were collected, cut and loaded directly into the drying tubes.
d These samples have very high area counts and they are fairly consistent with each other. Although the wallcovering does not have
especially heavy coverage, it is possible that the inks used on this covering (order no. 15511) have a particularly high solvent
content. There are no "after embosser" samples. This was the last day of ES's field testing and GTR was only printing that day,
intending to emboss that night.
-------
TABLE 2-10
SUMMARY OF TIME LAG AND PATTERN
DUPLICATION OF WALLCOVERING SAMPLES
GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Sample No.a
1A, fib
2A, Bb
5A, Bj
6A, Bb
9A, Bc
IDA, Bc
13A, Bc
14A, Bc
19A, Bc
20A, Bd
25A, B^
26A, Ba
29A, Bc
30A, Bc
33A, B
34A, B
35A, B
36A, B
37A, B
38A, B
39
40
41
42
Date
3/23
3/23
3/23
3/23
3/23
3/23
3/23
3/23
3/24
3/24
3/24
3/24
3/25 _
3/25
3/27
3/27
3/27
3/27
3/27
3/27
4/21
4/21
4/21
4/21
Time Lapse From Collection
to Start of Recovery
10
10
49
49
2
2
4
4
2
2
minutes
minutes
minutes
minutes
minutes
minutes
minutes
minutes
minutes
minutes
2 hours, 23 minutes
2 hours, 28 minutes
_ _ .1
1
minute
minute
Samples 33-42 were re-
covered in the ES lab
a Every two numbers represents a pair of samples which were recovered
simultaneously.
k These samples were cut attempting to duplicate the patterns.
Samples numbers 1 and 2 are a pair, and 5 and 6 are a pair.
c These samples were loaded directly into the drying tubes in the
print room in less than two minutes.
d These samples were placed in the Wheaton jars immediately after
being cut in the print room, while awaiting the drying tubes.
-37-
-------
causing some solvent to be driven off. The next set of samples (order no.
15519) shows a considerable difference between the before and after embosser
samples. The third set (order no. 14796) was a different wallcovering from
the others and was know to be far less absorbent. The resulting data reflect
this. The fourth set (order no. 15511) contains only before embosser samples.
During the field test, there was considerable "down" time in the print
room. Also during these three days, GTR experienced problems with some
wallcovering which was expanding with the heat of the process and, therefore,
could not be trimmed properly. When this occur ed, GTR printed the
wallcovering and waited until later in the day or at night to the embossing.
These factors account for the small number of samples collected.
The data indicate that some solvent is driven off by the heat of the
embosser and some remains on the wallcovering and is going out with the
-4
product. For example, sample 5A retained 2.47 x 10 pounds of MEK per
square yard. A hypothetical run of 5,000 square yards would then retain 1.24
pounds (0.7 liter) of MEK.
The unprinted wallcovering samples contained a significant amount of
solvent material. This may be because one or more of the solvents may be used
in the manufacturing of the vinyl or the fabric web. To check for sample
container contamination, samples of unprinted wallcovering were stored at room
temperature in a manila folder and analyzed one month after the original
analyses. Small amounts of solvent were detected in these delayed analyses.
2.7 FID Analyzer Audit Results
Two standard concentrations of propane and one standard concentration of
MEK were supplied by EPA and analyzed by TRC in order to assess the accuracy
of the FID analyzers. These standard concentrations (audit samples) were an-
-38-
-------
alyzed in the TRC laboratory before and after the measurement program and in
the field periodically during the program. The results of these audit sample
analyses are shown in Table 2-11. Details are presented in Appendix C.3.
The average error for each analyzer was within the specified criterion of
+10 percent for both propane and MEK. The analysis results for the one MEK
audit sample exhibit error considerably larger than the propane audit error.
The HER results for the analyzer used at the CA inlet were consistently high
(averaging +5.2%) while the MEK results for the wall fan/CA outlet and em-
bosser were consistently low (averaging -5.7% and -7.0%, respectively). The
reason for the difference between the CA inlet analyzer and the other two an-
alyzers is not evident. This difference may be attributable to the unique
response characteristics of each analyzer, since a similar though less pro-
nounced pattern is evident in the propane results.
The larger errors in the MEK results compared to the propane results are
probably due to the NEK bag standards used to calibrate the analyzers. These
bag standards were prepared as described by EPA Method 110. Very small
amounts of liquid MEK (0.5 microliters to 100 microliters) were vaporized in
1 cubic foot of air and contained in 30-liter Tedlar bags. The difficulties
inherent in measuring .out those small amounts of MEK with syringes, and in
accounting for other variables such as the volume, temperature and pressure of
the diluting air, contribute significantly to error. The potential adsorption
and desorption of MEK onto and from the bag walls, while controlled through
the use of pre-conditioned bags, can also contribute to error.
MEK calibrations were not performed on March 17 and March 23 (except for
the CA inlet analyzer), so propane/MEK relationships specific for these two
days could not be calculated. On these two days analyzer responses to the MEK
audit sample were converted to propane concentrations (since propane calibra-
-39-
-------
TABLE 2-11
AUDIT SAMPLE ANALYSIS RESULTS AT
GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Audit Sample
Sampling MEK
Analyzer Location Date 38.7jpm
Bendix 1 TRC Lab 3-11-81
CA Inlet 3-17-81
3-18-81
3-23-81
3-24-81
3-25-81
3-26-81
TRC Lab 3-31-81
Average
Error (%)
Bendix 2 TRC Lab 3-11-81
Wall Pan 3-17-81
CA Outlet 3-25-81
3-26-81
TRC Lab 3-31-81
Average
Error (%)
Scott 1 TRC Lab 3-11-81
Embosser 3-17-81
3-23-81
TRC Lab 3-31-81
Average
Error (%)
37.7
39.3
NA
49. la
41.7
34.8
42.5
40.1
40.7
5.2
33.8
37. 7a
33.7
37.8
39.7
36.5
-5.7
33.2
37. 6a
35. 8a
37.3
36.0
-7.0
Propane
309 ppm
297
302
298
301
NA
301
303
291
299
-3.2
283
294
292
303
303
295
-4.5
304
298
294
294
298
-3.6
Propane
10.01 ppm
10.2
*
10.0
10.3
NA
9.6
9.9
10.3
10.0
0.0
9.7
9.9
10.0
10.3
10.0
10.0
0.0
10.7
10.0
9.1
9.6
9.8
-2.0
Weighted Average Error (%) -1.3
-3.7
-0.5
* Analyzer attenuation was not low enough to yield proper
resolution.
a Measured as propane and converted to MEK using propane/MEK
relationships developed on nearest days.
-40-
-------
tions were performed daily), and then converted to MEK using propane/MEK rela-
tionships developed on the nearest days. All FID analyzer calibration pro-
cedures are discussed further in Section 5.3 and in Appendix C.
-41-
-------
3.0 PROCESS DESCRIPTION
Emission tests were conducted at the GTR vinyl-coated fabric plant in
Reading, Massachusetts during March 1981. The tests were designed to quantify
controlled and uncontrolled VOC emissions from the printing of vinyl
wallcoverings. An overhead diagram of the printing operation facilities is
shown in Figure 3-1. A schematic of the printing operation is shown in Figure
3-2.
, The print-line tested by EPA was housed in a room separate from the main
plant. The print room ventilation system consisted of a room air exhaust fan,
a room air.supply fan, a carbon adsorption unit inlet fan, an embosser exhaust
fan, and several doors. The print room itself and the room air supply fan are
designed to accomodate three print-lines. At the time of these tests only one
print-line, PE-3, was in the print room.
3.1 General Description
The vinyl wallcovering manufacturing process at this GTR plant consists of
three operations: the calendering operation (preparation of the substrate),
the ink mixing operation and the printing operation. The calendering opera-
tion works five days per week, eight hours per day, with an occasional shift
on Saturday and Sunday. The printing operation (print-line) works as a batch
operation five days per week, 24-hours per day.
The substrate is prepared by mixing plasticizers in three banbuiries. The
mixed plasticizers are fed into a feed mill for further mixing before being
transported to the calendering operation. Emissions from the three banburies
are manifolded and exhausted through a roof fan to the atmosphere. The feed
mill is vented by a canopy hood which is exhausted through a roof fan to the
atmosphere.
-42-
-------
EMBOSSER
ELECTROSTATIC
PRECIPITATOR
WALL FAN
EXHAUST
DOOR
F
PRINT LINE E
BUILDING
MAKE-UP
AIR INTAKE
"WALL FANf
EXHAUST'
EXTENSION
OVERHEAD
DOOR
EXHAUST DUCT FROM
PRINT HEADS
N
DOOf
DOORS
MAIN PLANT BUILDING
CA INLET VELOCITY
TRAVERSES
CA UNIT
CONTROL ROOM
CA INLET -
VOC SAMPLING
HUMIDITY
CONTROLLERS
SAMPLING LOCATIONS
CA = CARBON ADSORPTION
CA OUTLETS
TESTED
NOT TO SCALE
FIGURE 3-1:
OVERHEAD DIAGRAM OF PRINTING OPERATION FACILITIES AT
GENERAL TIRE AND RUBBER COMPANY, READING, MASSACHUSETTS
-43-
-------
WALL FAN v
EXHAUST DUCT V
MAKE-UP
AIR UNIT
oo
ELECTROSTATIC
PRECIPITATOR
TO CARBON
ADSORPTION UNIT
WALL FAN
.WALL FAN
(NOT USED DURING
TESTING)
PRINT
HEAD ,
FANS {
0
VINYL PRELIMINARY
COATED DRYER
SUBSTRATE
PRINTING HEADS/OVENS
EMBOSSER
FINISHED
PRODUCT
NOT TO SCALE
FIGURE 3^2; SCHEMATIC OF PRINT-LINE OPERATION
' . .'; AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTES
-------
In the calendering operation, the mixed plasticizers from the feed mill
are coated onto a fabric web. The heated rollers of the calender volatilize
the plasticizers, producing a light smoke over the calender stand. This smoke
is vented by a canopy hood and is exhausted directly to atmosphere through two
exhaust fans located on the roof. The finished product from the calendering
operation is rolled and transported to the print-line.
Inks for the printing operation are prepared by mixing ink base, color
pigments and solvents. Inks are transported to the print-line in 55-gallon
drums. During most times of the year, methyl ethyl ketone (MEK) makes up
approximately 95 percent of the solvents used to prepare the inks. Emissions
from the substrate preparation and ink-mixing operations were not investigated
in this test program.
3.2 Printing Operation Description
The print-line is located in a new building, separated from the calender-
ing area and ink mixing area. A Baker-Perkins rotogravure printing machine
with an in-line embossing unit utilizes six printing heads capable of printing
up to six different colors on the substrate. Depending on the type of print-
ing required for the substrate, one to six print heads can be utilized at any
one time.
The pre-mixed ink is transferred manually from drums to a pump tank loca-
ted next to each print head. When filled, a pump tank is connected to the
print head, and ink is pumped from the pump tank to a tray within the print
head. A print roller, half-submerged in the tray, transfers ink from the tray
to the substrate. Excess ink is gravity fed back to the pump tank.
After ink charging, the print-line is ready for the color matching pro-
cedure whereby the operator compares the colors of the new product with a
-45-
-------
standard sample. This is done by threading a leader substrate through the
entire print-line with a 4-foot square sample of the actual substrates taped to
the printed surface. This avoids wasting large quantities of new substrate
during the matching process. After the sample has passed through all the
print heads, it is removed from the'leader and compared to the standard. If
the test sample fails to match the standard, inks, ink base or solvent are
added to the pump tanks as needed. The sample printing is repeated until all
colors adequately match the standards. At the conclusion of color matching,
the leader substrate is drawn through the print-line and the actual product
substrate is threaded behind it to start the print run.
During the print run, operators will occasionally add NEK solvent or ink
base to the pump tanks to make-up liquid that has evaporated in order to main-
tain the required ink viscosity. Each addition is made manually using five-
gallon pails which are filled from 55 gallon drums located near the print-line.
During a printing operation a roll of substrate is fed through a prelimi-
nary dryer, the six print heads, and an embosser. The inked substrate is
dried in an oven contained within each print head and is further heated within
the embosser. The final product is re-rolled as it emerges from the embosser.
Periodic line.shuts-downs are common during each run due to problems with
color matching, registration, the embosser, printers and ink circulation sys-
tems. During the down periods, print rollers are often cleaned with small
quantities of tetrahydrofuran (THF) and a stiff brush.
To minimize the number of times the pump tank and ink circulation system
must be cleaned with solvents, operators will often schedule a series of runs
using similar colors. When this occurs, a new color is added to a pump tank
without removing the residual ink from the previous run. However, solvent
cleaning is necessary if the new color contrasts significantly with the pre-
-46-
-------
vious color. Cleaning is done by disassembling all ink circulation system
components and rinsing them with MEK solvent. The resulting solvent/ink waste
mixture is placed into the original ink mixing drum and removed. The rinsed
circulation components, including the used pump tank, are sent out for addi-
tional cleaning and a new pump tank and circulation system is installed.
3.3 Printing Operation Emission Controls
VOC emissions are generated by the vaporization of solvents used in the
print-line. The print-line emissions are captured by a hooding system that
allows the captured emissions to be drawn into the individual print hood
ovens. The print-line gaseous emissions which are not captured enter the
print room atmosphere and are removed through the other ventilation equipment.
Emissions from the preliminary dryer and the forced-air drying ovens with-
in each print head are manifolded and ducted to a carbon adsorption (CA) unit
before being released to the atmosphere. The CA unit is located approximately
300 feet from the print-line building. The unit was designed and installed by
Sutcliffe-Speakman, Leigh of Lancashire, England.
There are now three carbon beds associated with the carbon adsorption sys-
tem. As one carbon column is removing the organic compounds from the air/va-
por mixture (adsorption), another is being steam regenerated (desorption).
The adsorption/desorption cycle lasts for approximately 120 minutes. The sol-
vent/water mixture recovered from the desorption process is treated by azeo-
tropic distillation and dehydration with calcium chloride. During this mea-
surement program only carbon beds 1 and 3 were operating.
After leaving the sixth print head, the printed substrate passes through a
heated embossing unit. The embossing operation further volatilizes the plas-
ticizer, resin, and solvents. Some portion of these plasticizers condense and
-47-
-------
form an aerosol. VOC emissions are also generated during embossing. Embosser
generated emissions as well as print room ambient emissions are exhausted by
the embosser fan. The embosser emissions are exhausted to an electrostatic
precipitator (ESP) manufactured by United Air Specialists, Cincinnati, Ohio.
The ESP removes the aerosols but does not control the VOC emissions.
3.4 Print-Line Building Air Circulation
Air circulation within the print-line building is affected by four factors:
1. Two wall exhaust fans on the west wall;
2. A make-up air fan on the roof;
3. An overhead door and two small doors opening to the outside; and
4. Four sliding doors connecting the print-line building to the
. adjacent plant building.
The wall fans are designed to exhaust the fugitive emissions generated by the
print-line to the atmosphere. The overhead door is generally open during the
summer and closed during the winter. The make-up air fan supplies outside air
to the print-line building at the rate of approximately 35000 SCFM, as esti-
mated by plant personnel.
3.5 Operation Parameters Controlled During this Testing Program
The make-up fan remained off during this testing program primarily because
plant personnel had determined that the turbulence it generates adversely af-
fects drying within the print head ovens and the embosser. Since it intro-
duces air from a single point about 20 feet above the floor,, air distribution
is nonhomogeneous along the print-line. In addition, it was shown in the pre-
vious EPA test program at this GTR plant that with the make-up fan on, there
is a net air flow out of the print-line building through open doorways (1).
To maximize the VOC loading to the CA unit, the make-up fan was kept off.
-48-
-------
A duct to contain exhaust air from one of the two wall fans was installed
on the roof of the print-line building prior to the first EPA testing program
in order to facilitate monitoring of wall fan exhausts. During this second
test program both wall fans remained off except during the afternoon of
March 19, 1981. Prior to this time the fan belt on the ducted fan had been
slipping and no flow was detectable in the duct even though the fan motor was
on. At approximately 1330 on March 19 the belt was tightened and this one fan
remained on the rest of the afternoon. Subsequently the wall fan remained off
except for brief periods when the print-line was down. During these times the
ducted wall fan was turned on to ventilate the print room.
During the color matching and actual printing process steps, the print
head fans were all on to remove the VOC emissions from the ovens and ink
wells. Often during the switching and clean up steps, the print head fans
were turned off to reduce noise levels in the area. When the print head fans
are off, emissions around the print-line are reduced because of lower air cir-
culation around the print-line and because no printing is occurring.
Generally, all but one print room doorway remained closed during the test-
ing program in order to control the amount of air flowing into the print
room. The open door was doorway E at the northeast corner of the print room,
opening to the main plant. An exception to this occurred on March 18, 1981,
when doorways D and B, also opening to the main plant, remained partially
open. During the remainder of the testing program other doorways were occa-
sionally opened and usually closed immediately.
Immediately prior to this test program, the carbon adsorption system had
been down for repairs and a third carbon bed had just been added as part of a
planned expansion. Trained operators were not yet available for night shift
operation so the system was operated only on the day shift. At the end of
each day the carbon adsorption system was completely shutdown. The carbon
-49-
-------
beds were regenerated two times to prevent any remaining solvent from causing
fires. This extra regeneration of the beds each evening would make the carbon
bed system somewhat more efficient the next day, compared with normal 24-hour
per day operation.
3.6 Monitoring of Process Operations
Print-line operations were continuously monitored and recorded by Radian
personnel during the testing program. A copy of this process log is shown in
Appendix H.
-50-
-------
4.0 DESCRIPTION OF SAMPLING LOCATIONS
This section presents descriptions of the sampling locations used during
the VOC emission testing program conducted at the GTR vinyl-coated, fabric
plant in Reading, Massachusetts, during March 1981. An overhead diagram of
the printing operation facilities and approximate sampling locations is shown
in Figure 1-1.
4.1 Carbon Adsorption (CA) Unit Inlet
4.1.1 Flowrate Measurements
The CA unit inlet flowrate sampling location was located in a 40-inch i.d.
horizontal section of metal duct. A schematic of this location is shown in
Figure 4-1. This sampling location was approximately 20 feet, above the ground
and was reached by erecting temporary scaffolding.
Two 2-inch capped sampling ports were located approximately 40 feet up-
stream from the carbon unit. The ports were positioned 90 degrees apart in a
vertical plane, each 45 degrees on either side of the top of the duct. The
ports were approximately 7 feet (2.1 duct diameters) upstream and 27 feet (8.1
duct diameter) downstream from bends in the duct. Since this location did
meet the eight-and-two diameters criteria of EPA Method 1, six sampling points
were used on each traverse axis/ for a total of 12 sampling points.
4.1.2 VOC Sampling
The CA unit inlet VOC sampling location was located at the center of a 90
degree bend in a 24-inch i.d. section of metal duct. The continuous FID sam-
ples and NMO Method 25 samples were drawn from the center of the duct using a
stainless steel probe inserted through one 2-inch port located approximately
-51-
-------
VERTICAL
A
PRINT-LINE
BUILDING
40"
TO CARBW
ADSORPTION
UNIT
VELOCITY
TRAVERSE
PORTS
OVERHEAD VIEW
TRAVERSE POINT
NUMBER
1
2
3
4
5
6
TRAVERSE POINT
LOCATION FROM
DUCT WALL
(INCHES)
1.8
5.8
11.8
28.2
34.2
38.2
FIGURE 4-1:
CARBON ADSORPTION UNIT INLET VELOCITY
TRAVERSE LOCATION AT GENERAL TIRE AND
RUBBER COMPANY, READING, MASS.
-52-
-------
36 inches upstream from the newer of the two CA unit humidity controllers.
The FID samples were drawn through 20 feet of heated 1/4-inch Teflon tubing to
the FID analyzer. This sampling location is shown in Figure 4-2.
The VOC samples were not taken at the flowrate measurement sampling loca-
tion for the following reasons. EPA Method 1 criteria for sampling site se-
lection are not applicable and not required for VOC sampling. The flowrate
measurement sampling location was inconvenient because it was about 20 feet
above ground and 60 feet from the sheltered CA unit control room where the FID
analyzer was located. Instead, the VOC sampling location provided representa-
tive samples and was more accessible. The VOC sampling site was only 5 feet
above ground and within 20 feet of the FID, thus less heated sampling line was
heeded.
4.2 CA Unit Outlets
4.2.1 Flowrate Measurements
The CA unit outlet flowrate sampling locations were located in 24-inch
i.d. vertical metal ducts. A schematic of these locations is shown in
Figure 4-3.
On each of the two carbon bed outlet stacks (carbon beds 1 and 3) / two
2-inch sampling ports were positioned 90 degrees apart in a horizontal plane.
The ports were located approximately 4 feet upstream from the top of the
stack, and approximately 16.5 feet downstream from where a 6-inch steam bypass'
duct attaches to the stack. Since these sampling locations did meet the
eight-and-two-diameters criteria of EPA Method 1, six sampling points were
used on each traverse axis, for a total of 12 sampling points.
-53-
-------
LOOKING SOUTH
CA UNIT
CONTROL
ROOM
TO CARBON BEDS
FROM PRINr
HEADS
HUMIDITY
CONTROLLERS
SAMPLING
PORT
LOOKING WEST
0--
TO CARBON
BEDS
FAN
HUMIDITY
CONTROLLER
FROM PRINT HEADS
24,, ~*-7£ SAMPLING PORT
3'-
FIGURE 4-2: CARBON ADSORPTION UNIT INLET VOC SAMPLING LOCATION
AT GENERAL TIRE AND RUBBER COMPANY, READING, MASSACHUSETTS
-54-
-------
PLATFORM-
HUMIDITY'
NTROLLER
CARBON
BED
LOOKING
NORTH
c
. lc
1
o
2'
4 \
Q
L,
f
1 " _J/ -
if 4'
VELOCITY
TRAVERSE
PORTS
16.5'
PORT 2
STEAM
BY-PASS
VOC SAMPLING
PORT
TRAVERSE
POINT
NUMBER
1
2
.3
.4
5
6
DISTANCE
FROM
DUCT WALL
(INCHES)
1.1
3.5
7.1
16.9
20.5
22.9
FIGURE 4-3: CARBON ADSORPTION UNIT OUTLET SAMPLING LOCATIONS
(CARBON BED 3) AT GENERAL TIRE AND RUBBER COMPANY,
READING, MASSACHUSETTS.
-55-
-------
4.2.2 VOC Measurements
The CA unit outlet sampling locations were located approximately 20 inches
downstream from where the outlet stacks emerge from the base of the carbon
beds. The VOC samples (continuous FID and NMO Method 25) were drawn from the
center of the duct using a stainless steel probe inserted through a 2-inch
port. The continuous FID samples were drawn through 40 feet of heated 1/4-
inch Teflon tubing to the FID analyzer. This sampling location is shown in
Figure 4-3. The VOC sampling location was different from the flowrate mea-
surement location because the VOC sampling location was more convenient and
EPA Method 1 criteria are not required for VOC sampling.
4.3 Embosser Electrostatic Precipitator (ESP) Inlet
4.3.1 Flowrate Measurements
The embosser ESP inlet sampling location was in a 20-inch i.d. horizontal
section of metal duct. A schematic of this location is shown in Figure 4-4.
Two 1.5-inch capped ports were positioned 90 degrees apart in a vertical
plane. The ports were located 24 inches (1.2 duct diameters) upstream from
the ESP and 143 inches (7.2 duct diameters) downstream from a 90 degree bend.
Since this location did not meet the eight-and-two diameters criteria of
Method 1, eight sampling points were used on each traverse axis, for a total
of 16 sampling points.
4.3.2 VOC Sampling
MMWMMM^^M^kBM^MM** j
FID and NMO sampling was performed at the same location used for the flow-
rate measurements. VOC samples were drawn from the center of the duct with a
stainless steel probe inserted through one port. The continuous FID samples
were drawn through 60 feet of heated 1/4-inch Teflon tubing to the FID analy-
zer.
-56-
-------
. LOOKING WEST
EXHAUST
SAMPLING
PORTS
O
r
48"
1
/
f
<*~ EMBOSSER 20" ; \)
X^"l 1/1?" >
'
ELECTROSTATIC
PRECIPITATOR ,
- ?/i"__..»J\
fct *l X
PRINT-LINE BUILDING ROOF
(
FAN
PORT 2
PORT 1
TRAVERSE POINT
NUMBER
1
2
3
4
5
6
7
8
TRAVERSE POINT
LOCATION FROM
DUCT WALL
(INCHES)
0.6
2.1
3.9
6.5
13.5
16.1
17.9
19.4
FIGURE 4-4: EMBOSSER ELECTROSTATIC PRECIPITATOR INLET SAMPLING LOCATION
AT GENERAL TIRE AND RUBBER COMPANY, READING, MASSACHUSETTS.
-57-
-------
4.4 Wall Fan Exhaust Duct
4.4.1 Flowrate Measurements
The print-line building wall fan exhaust sampling location was located in
a 48-inch i.d. square horizontal section of metal duct. A schematic of this
location is shown in Figure 4-5. The exhaust extension was applied to only
one of the two wall fans. The other fan was not operated during the testing
program.
Three 3-inch capped sampling ports were positioned on the north side of
the duct in a vertical plane. The ports were located 96 inches (2.0 equiva-
lent duct diameters) upstream from the end of the duct and 384 inches (8.0
equivalent duct diameters) downstream from a 180 degree bend. Since this lo-
cation did meet the eight-and-two-diameters criteria of EPA Method 1, four
sampling points were used on each of the three traverse axes, for a total of
12 sampling points.
4.4.2 VOC Sampling
FID and NMO sampling was performed at the same location used for the flow-
rate measurements. VOC samples were drawn from the center of the duct with a
stainless steel probe inserted through the center port (Port B). The contin-
uous FID samples were drawn through 20 feet of heated 1/4-inch Teflon tubing
to the FID analyzer.
4.5 Print-Line Building Ambient Air Measurements
Ambient air VOC measurements were made with a portable hydrocarbon analy-
zer in the immediate vicinity of the print heads and embosser and at locations
throughout the print-line building. Near the print heads measurements were
made at the center of the print-line, 2 feet from the print-line and 5 feet
from the print-line. Near the embosser measurements were made approximately
-58-
-------
LOOKING SOUTH
I
Ul
EXHAUST EXTENSION
^ to
rf 12"
< 30" >
< 18" -H
-6'Lj |
>J
^
+ + +4
*- -« + -f
i
F
+ *!! -1- . -
PRINT LINE BU
>
A
1
J
B
P^F^
c '
T 24"
JIT ,
8"
i 1 ^
4
0"
' i
t
8"
r
WALL
EXHAUST
NOT TO SCALE
/ / / // 7
FIGURE 4-5:
WALL FAN EXHAUST SAMPLING LOCATION AT
GENERAL TIRE AND RUBBER COMPANY,
READING, MASSACHUSETTS.
-------
2 feet from the cage surrounding the embosser. During the first two ambient
surveys, sampling was done at three levels: 1, 5 and 8 feet above the floor.
All later sampling was done at only one level, approximately 5 feet above the
floor (breathing level). The VOC measurement positions shown in Figure 4-6
represent approximate locations and were not precisely dimensioned. Figure
4-7 shows the VOC measurement site around the embosser.
Air flow measurements were made in open doorways with a hand-held hot-wire
anemometer and a vane anemometer. These approximate measurement locations are
shown in Figure 4-6. Nearly all air flow measurements were made in doorway £
which was always completely open. Nine measurement points were used in door-
way E. Six measurement points were used (three rows of two) in doorway 0 and
nine measurement points were used in doorway B when these two doorways were
partially open on March 18.
4.6 Wastewater Sampling Locations
Wastewater samples from the carbon adsorbtion unit were collected from a
common drain at the base of the carbon beds and from a drain at the base of
the water/solvent distillation column.
4.7 Wallcovering Product Sampling Locations
i
Wallcovering samples were collected from the print-line immediately before
and after the embosser.
-60-
-------
1
*
4
7
*
2
*
5
8
*
3
*
6
9
*
DOORWAY
FLOWRATE
MEASUREMENT
LOCATIONS
AMBIENT AIR VOC MEASUREMENT POINTS (BREATHING LEVEL)
OVERHEAD
DOOR
ON
loo I
OFF
loo]
PRINT HEADS (STAGES)
CENTER
SUBSTRATE
DRYER
t
1
EMBOSSER
PRODUCT
]
00
OVERHEAD
MAKE-UP
AIR FAN
DESKS
MIXING
AREA
SWITCH
ROOM
OFFICE
D
NOT TO SCALE
FIGURE 4-6: PRINT-LINE BUILDING AMBIENT AIR MEASUREMENT LOCATIONS
AT GENERAL TIRE AND RUBBER COMPANY, READING, MASSACHUSETTS
-------
TO ESP
to
i
FROM
PRINT
HEADS
EMBOSSING
ROLLERS
TO
TRIMMERS
APPROXIMATE MEASUREMENT LOCATIONS
APPROXIMATE SCALE
0
10 FEET
FIGURE 4-7: AMBIENT AIR MEASUREMENT LOCATIONS ALONG THE EMBOSSER AT
GENERAL TIRE AND RUBBER COMPANY, READING, MASSACHUSETTS.
-------
5.0 SAMPLING AND ANALYSIS METHODS
This section presents descriptions of the sampling and analysis procedures
used during the VOC emission testing program conducted at the GTR vinyl-coated
fabric plant in Reading, Massachusetts during March 1981. . . .
5.1 EPA Reference Methods Used During this Program , L
Method 1 - Sample and Velocity Traverses for Stationary Sources (2)
This method specifies the number and location of sampling points
within a duct, taking into account duct size and shape and local flow
disturbances. : : .
Method 2 - Determination of Stack Gas Velocity and Volumetric Flow-
rate (2) : .
. This method specifies the measurement of gas velocity and flowrate
using a pitot tube, manometer, and temperature sensor. The physical
dimensions of the pitot tube and its spatial relationship to the tem-
perature sensor and any sample probe are also specified.
Proposed Method 25A - Determination of Total Gaseous Organic Concen-
tration Using a Flame lonization Analyzer (3) . ' .
- - - This method describes how VOC are continuously sampled and. analyzed. ..
using an FID analyzer. \.
Method 25 - Determination of Total Gaseous Non-Methane Organic Emis-
sions as Carbon (4) i^_ : _..
-. i. This: method describes how gaseous non-methane organic compounds are- -
sampled and analyzed. An emission sample is drawn through a ccnden- .
:--... sate .trap and into an evacuated tank. Trap and tank contents are
oxidized to C02, reduced to methane and analyzed by EID. . - . . : :;....:;
. Proposed Method 110 - Determination of Benzene from Stationary j
Sources (5)
This:: method describes how to prepare standard gas mixtures., (benzene.
or other solvents) in Tedlar bags. ...- ;:.-.-:: ......
5.2 Duct Flowrate Measurements . _._._; _,.
Velocity traverses were performed in accordance with EPA Methods 1 and 2
at the embosser ESP inlet, wall fan exhaust duct, CA unit inlet and CA unit
outlets periodically during each testing day. The primary purpose of these
-63-
-------
measurements was to determine representative flowrates through these ducts
when the print-line was operating.
Traverses were performed at each location with a standard pitot tube for
velocity head measurements and a Thermo-Electric Digimite model 31160 elec-
tronic thermocouple for temperature measurements. Some temperature measure-
ments were made with an ASTM thermometer when one of the two Oigimites brought
to the field broke. A 5-foot standard pitot with a 0-10-inch water manometer
was used at the CA inlet. A 3-foot standard pitot was used with a 0-10-inch
manometer at the embosser ESP inlet and the CA outlets. A 5-foot standard
pitot was used with a 0-0.25-inch manometer at the wall fan exhaust duct. A
more sensitive manometer was used at the wall fan exhaust duct because of the
very small velocity heads expected at this location, based on the results of
the previous measurement program at 6TR (1).
A measurable flow (about 5000 SCFM) was expected in the wall fan exhaust
duct with the one ducted wall fan on. However, velocity traverses performed
on March 17 and 18 indicated no flow in this duct, even though the fan motor
was on. Plant personnel discovered on March 19 that the fan belt was slip-
ping, and when this was corrected at about 1330 on March 19, flow in this duct
was about 10000 SCFM. As a result, air flow in the CA inlet duct decreased by
about 7 percent (from about 8200 SCFM to about 7600 SCFM), air flow in the
embosser ESP inlet remained unchanged, and air flow into the print room
through doorway E approximately tripled (from about 6100 SCFM to about 20000
SCFM). In order to maximize the VOC loading to the CA unit, Radian personnel
decided to continue the measurement program after March 19 with the wall fan
off when the print-line was operating.
VOC measurements were made at the CA unit outlets on March 25 and 26 only,
so flowrate measurements at the CA outlets were made only on those two days.
At any given time, VOC and flowrate measurements were made at either carbon
-64-
-------
bed 1 outlet or carbon bed 3 outlet, depending on which bed was in operation
(adsorbing).
To calculate flowrate (SCFM) from the velocity head.and temperature data,
moisture contents were estimated. Wet bulb and dry bulb temperature measure-
ments were made daily in the print room with a Bendix. Psychron. These mea-
surements indicated a moisture content in the print room of between 0.2 per-
cent and 0.5 percent. Air in the CA inlet, embosser ESP inlet and wall fan
exhaust ducts was assumed to be dry (molecular weight 28.8). The moisture
content in the CA outlet ducts was assumed to be 5 percent (molecular weight
28.3). This assumption is based on the fact that air with a moisture content
of 5 to 6 percent is nearly saturated at the CA outlet stack temperatures
(about 110 F). Barometric pressures were obtained daily from Logan Airport
in Boston. .
5.3 VOC Measurements with FID Analyzers ."/ _
5.3.1 Sampling with FID Analyzers :
VCX: concentrations at the CA unit inlet, embosser ESP inlet and wall fan
exhaust-duct were measured continuously during each; day. (daytime shift) from
March 18 to March 20, 1981. Since there was no flow, in the wall fan exhaust
duct, monitoring at this location was discontinued on. March 23 (Monday). No
sampling was performed on March 24 because no printing, was done this day. A
large backlog of printed but un-embossed wallcovering, was accumulated on March
23 : because of embosser problems this day. Consequently, March 24 was
dedicated -to ^embossing only. VOC concentrations were ;measured continuously
only at the CA unit inlet and CA unit outlets on March 25 and. March 26.
The FID analyzers were operated concurrently at each, sampling location. A
schematic of the sampling equipment is shown in Figure 5-1. A stainless steel
probe was inserted through one port and samples were drawn from the center of
-65-
-------
FID
ANALYZER
HEATED TEFLON TUBING
STAINLESS STEEL
TUBING
FILTER
HOLDER
RECORDER
DUCT
WALL
FIGURE 5-1: FLAME IONIZATION DETECTION SAMPLING SYSTEM
-------
the duct. A 2-inch glass fiber filter was positioned between the probe and a
heated Teflon sample line to remove particulate matter. This filter was re-
placed each day, though no significant particulate accumulation was ever ob-
served. The sample line was heated to at least 10°F above stack temperature
to prevent condensation of water and organics within the line, and the line
was attached directly to the FID analyzer. The continuous analyzer output was
recorded on a strip chart recorder.
The analyzers monitoring emissions at the embosser ESP inlet ami wall fan
exhaust duct were housed in heated fiberglass shelter atop the print-line
building. The analyzers monitoring emissions at the CA unit inlet and outlets
were housed in the CA unit control room.
Each monitoring day the analyzers were turned on at about 0700, allowed to
warm up for about one hour and then calibrated. Sampling continued until the
end of the work day or earlier if printing operations ended before the end of
the day. Calibrations were repeated at the end of each monitoring day.
Three FIO analyzers were used during this measurement program. The sam-
pling locations for each analyzer were as follows:
FID ANALYZER SAMPLING LOCATION DATES
Bendix 8402, TRC-1 CA unit inlet March 18-20,23,25,26
Bendix 8402, TRC-2 Hall Fan Exhaust March 18-20
CA unit outlets March 25-26
Scott 415, TRC-1 Embosser March 18-20,23
5.3.2 Calibration of the FID Analyzers
During the week prior to and the week following the field work, the three
FID analyzers were calibrated at the TRC laboratory with both propane and MEK
standards. The primary purpose of these calibrations was to determine that
the analyzers were operating satisfactorily. In addition, a preliminary
-67-
-------
relationship between the response to propane and the response to MEK was es-
tablished for each analyzer. With this relationship, the total hydrocarbon
(THC) field measurements (based on propane calibrations) could be converted to
THC measurements expressed as MEK. Calibrations were performed with six pro-
pane compressed gas mixtures and eight MEK bag standards. The MEK standards
were prepared by diluting known amounts of MEK in a known volume of hydro-
carbon-free air, following the general procedure presented in EPA proposed
Method 110. The propane and MEK standards used during these calibrations are
shown in Table 5-1.
In the field each analyzer was. calibrated at the beginning and end of each
monitoring day with propane standards. Because of the difficulty in carrying
the compressed gas cylinders onto the roof of the print-line building, propane
bag standards were prepared every other day and were used to calibrate the
analyzer at the embosser ESP inlet and wall fan exhaust duct. The propane bag
standards "were prepared by half-filling 30-liter Tedlar bags with the propane
mixtures. When a bag was filled for the first time/ it was labelled and set
aside for an hour to allow for adsorption onto the bag wall. The bag was then
emptied and immediately refilled with the same propane mixture. Subsequently
a given bag was refilled with the same propane mixture previously used in it.
In addition to the calibrations performed with several propane concentra-
tions at the beginning and end of each day/ a one-point calibration was per-
formed one or more times each day as a check on the response of the analyzer.
A propane concentration near the highest expected concentration was used for
this one-point (span) check. The propane standards used at each sampling lo-
cation during the field program are shown in Table 5-1. To save time, all six
propane standards were not always used with each analyzer. Four or five stan-
dards were identified as the base standards for each analyzer, depending on
the concentration range expected at each sampling location.
-68-
-------
TABLE 5-1
PID ANALYZER CALIBRATION CONCENTRATIONS
USED DURING THE VOC MEASUREMENT PROGRAM AT
GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Measured
Concentration Range
Sampling Location (ppm as propane) Propane Standards (ppm) MEK Standards (ppm)*
eneral 0, 5.4, 9.7, 75.1, 274, 0, 4.7, 9.5, 18.9, 47.3,
499, 949 94.5, 189, 473, 945
KA Inlet 7.5 - 820 0, 9.7, 75.1, 274, 499, **
949
ftA Outlet
Embosser ESP
Wall Fan
1
15
1
.3 -
.2 -
.1 -
13.0
362
146
0,
0,
0,
5.4,
5.4,
5.4,
9
9
9
.7,
.7,
.7,
75
75
75
.1,
.1,
.1,
274
274
274
**
**
**
* These MEK concentrations are approximate; exact concentrations varied each time MEK bag
standards were prepared, depending on temperature and pressure.
** The same MEK bag standards were used at all sampling locations.
-69-
-------
Each analyzer was calibrated with MEK standards periodically during the
measurement program. These calibrations were performed with one of the
twice-daily propane calibrations in order to establish propane-to-MEK response
relationships. MEK bags were prepared every day or every other day, and the
set of seven or eight bags so prepared were used at all sampling locations.
Each bag was labeled and reused with the same approximate MEK concentrations.
The approximate MEK concentrations used in the field are the same as those
used during the laboratory calibrations, as shown in Table 5-1. The exact
concentrations varied from day to day depending on temperature and pressure of
the dilution air as it was metered into the bags.
When a given calibration gas was introduced into an analyzer, the analyzer
*
response was recorded on as many attenuations as would fit on the strip chart
or as would give a readable response. In this way a. calibration curve could
be prepared for each attenuation.
5.3.3 Audit Sample Analysis of FID Analyzers
Three audit samples (one MEK and two propane) were provided by EPA. These
compressed gas mixtures were introduced to each FID analyzer during the cali-
brations before and after the field program, and periodically during the field
program. As with the propane standards, Tedlar bags were filled with the
audit gases in order to avoid carrying the gas cylinders to the print-line
building roof. The response (in chart divisions) of each analyzer to the
audit samples on a given day was converted to concentration through the
propane and MEK calibration curves calculated from the propane and MEK cal-
ibration data of that day. Details of the audit results are presented in
Appendix C.3.
-70-
-------
5.3.4 Data Reduction and Calculations for FID Analyzers
Calibration equations relating analyzer response (chart divisions) to
total hydrocarbon (THC) concentration were calculated for each day that cali-
brations were performed. Calibration equations were calculated for each anal-
yzer and for each attenuation used for the measured duct concentrations and
audit sample analyses.
Propane and MEK calibration equations were calculated for each day that
MEK calibrations were performed. For each analyzer and attenuation the two
calculated relationships were:
CD - MpXp + Bp (5-1)
CD
Where: CO = chart divisions measured from zero offset
M a slope of propane response line
= slope of MEK response line
a total hydrocarbon propane concentration (ppm)
a total hydrocarbon HER concentration (ppm)
a intercept of propane response curve
a intercept of MEK response curve
By combining equations (1) and (2) , the MEK concentration equivalent to a
given propane concentration was expressed as follows:
XMEK "p 'VMEK' * K <5-3>
where: K is a constant equal to (Bp -
A summary of equations 5-1, 5-2 and 5-3 for each calibration day is shown
in Table 5-2. These equations were used only to produce the propane/MEK rela-
tionships. A propane/MEK relationship calculated for a given day was also
used for the nearest days for which no MEK calibrations were performed. The
pre-field-program and post-field-program calibration data were not used to
reduce field data.
-71-
-------
TABLE 5-2
PROPANE AND MEK CALIBRATION EQUATIONS 08ED TO
ESTABLISH PROPANB-TO-MEK CONVERSION EQUATIONS
GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
I
~J
10
Analyser
Bendlx 8402
THC-1
Sampling Calibration Dates
Location Date* Usedb
CA Inlet 3-20
3-25
3-20
3-24
3-25
3-26
3-20
3-24
3-26
3-24
3-20,19,18
3-25
3-20, 19
3-23
3-25
3-26
3-20.19,18
3-23
3-26
3-23
Propane0
Attenuation
10000
ioooo
3000
3000
3000
3000
1000
1000
1000
300
r
0.9999
0.9908
0.9999
0.9999
0.9999
0.9978
0.9999
0.9993
0.9983
0.9982
M
0.032
0.033
0.098
0.096
0.096
0.095
0.292
0.269
0.282
0.793
B
-0.50
-0.48
-0.3
-0.9
-0.3
-0.4
-1.1
-1.1
-1.1
-1.8
MBKC
r
0.9999
0.9999
0.9976
0.9990
0.9957
0.9991
0.9957
0.9970
0.9992
0.9986
H
0.031
0.030
0.106
0.105
0.097
0.099
0.317
0.304
0.292
0.927
B
0.06
0.26
-0.6
-0.7
1.0
-0.5
-2.0
-1.5
-1.4
-2.7
Propane/MEK
Relationship**
A K
1.03
1.10
0.92
0.91
0.99
0.96
0.92
0.89
0.97
0.86
-IB. 7
-24.7
2.8
-1.9
-13.7
0.7
2.8
1.3
1.0
1.0
Day when MBK and propane calibrations were perforated.
b Days when the Indicated propane/MEK relationship was used (because MEK/propane calibrations were not performed
every day).
0 CD « MX + B where CD chart divisions, H » slope, X concentration (ppn), B - Intercept, r correlation
coefficient.
d MEK concentration (ppm) - |A * propane concentration (ppai) I + K. This equation Is calculated by combining the
propane and KEK calibration equations. See Section 5.3.4.
-------
TABLE 5-2 (Continued)
PROPANE AMD MEK CALIBRATION EQUATIONS USED TO
ESTABLISH FROPANB-TO-MBK CONVERSION EQUATIONS
GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
I
-J
Analyser
BendlB 8402
TRC-2
Scott 415
TRC-1
Sampling Calibration Dates
Location Date* Uaedb
Hall Fan
Nail Fan
Mall Fan
CA Outlet
Nail Fan
CA Outlet
Babosser
3-20
3-20
3-20
3-26
3-20
3-26
3-20
3-24
3-20
3-20,19.18
3-20,19,16
3-20,19.18
3-26,25
3-20,19,18
3-26,25
3-20,19,18
3-23
3-20,19,18
Propane0
Attenuation
1000
300
100
100
30
30
1000
1000
i
500
r
0.998S
0.9999
0.9808
0.9921
1
0.9928
0.9980
0.9997
0.9968
M
0.324
0.969
3.033
2.219
7.037
7.416
0.245
0.248
0.463
B
-1.1
0.3
-1.5
-0.7
0
-2.2
-0.1
-0.3
1.4
HEKC
r
0.9964
0.9983
0.9988
0.9981
1
0.9904
0.9975
0.9999
0.9963
M B
0.341 -1.9
0.976 1.1
3.066 0.4
2.377 1.1
11.064 0
8.838 3.3
0.276 -0.7
0.274 -0.1
0.497 0.5
Propane/MEK
Relationship*1
A K
0.95
0.99
0.99
0.93
0.64
0.84
0.89
0.91
0.93
2.3
-0.8
-0.6
-0.8
0*
-0.6
2.0
-0.7
1.8
Only one non-zero calibration point could be used for the propane and MEK equations.
i
Day when MEK and propane calibrations were performed.
.. i
b Days when the indicated propane/MEK relationship was used (because HBK/propane calibrations were not performed
every day). j
0 CD » MX + B where CD chart divisions, M slope, X concentration
-------
Strip charts were divided into time intervals according to the consistency
of the trace and according to times corresponding to significant process oper-
ations. An average VOC-as-propane (X ) concentration was calculated for
each time interval, using a cursor to determine chart divisions (CO) and then
using a daily propane calibration equation, calculated from the combined
beginning-of-day and end-of-day calibration data, to determine X . These
daily propane calibration equations (calculated for each attenuation) have the
same form as Equation 5-1, and are shown in Appendix C.3. With X and the
measured flowrates the following parameters were then calculated:
VOC-as-MEK concentration (%ER) ' using equation (5-3) ;
mass as MEK emitted during' each time interval (Ib) ;
mass emission rate during each time interval (Ib/hour) ; and
total mass emitted during each print run (Ib) .
The mass as MEK emitted during each time interval was calculated by multi-
plying the VOC concentration for that time interval by the average flowrate
for the run and by the length of the tine interval:
MassMEK " XMEK Q T p (5-4)
Where: %EK * total hydrocarbon concentration as MEK (ppm)
Q m average flowrate for the run (SCFM)
T « time interval (minutes)
F - units conversion factor
Details of these calculation procedures are presented in Appendix C.
Since flowrates were not measured continuously, judgements were made on
what flowrates should be assigned to which time intervals, as discussed in
Section 2.2. When two or more consecutive measured flowrates did not change
significantly (approximately ± 10%) , these flowrates were averaged. Using the
process log and the consistency of the strip chart traces as guides, flowrates
were then assigned to time intervals. The flowrate assignments are shown on
the work sheets in Appendix B.
-74-
-------
5.4 NMO Sampling with Method 25
5.4.1 Preparation for Method 25
The NMO sampling train condensate traps and evacuated tanks were prepared
at TRC in accordance with Method 25 guidelines. Detailed preparation pro-
cedures are shown in Appendix 0.
Each uniquely identified trap was connected to a source of hydrocarbon-
free (HCF) air and a C02 analyzer and placed in a furnace. While heated to
600°C the trap was purged with HCF air until CO. concentrations were below
10 ppm.
Each uniquely identified tank (each with a volume of approximately
4800 cc) was purged with HCF air and evacuated three times. Twenty percent of
the tanks were then analyzed to ensure they contained less than 10 ppm hydro-
carbons.
Six flow control assemblies were adjusted to a nominal flowrate of approx-
imately 50 cc/mihute. ""The" evacuated tank was connected to the flovr assembly
and the on/off value was then opened. Flow into the tank was adjusted with
the flow control valve, and the flow assembly was then sealed.
5.4.2 Sampling for Method 25
Prior to assembly of the sampling train, tank vacuum pressure was checked
with a mercury U-tube manometer to ensure no change had occurred since labora-
tory preparation. At each sampling location two Method 25 sampling trains
were assembled, as shown in Figure 5-2, except without the glass fiber filter,
and each train was leak checked. The sampling plan was to collect duplicate
samples at each location only while the print-line was operating and concur-
rently with the continuously operating FID analyzers. At the embosser ESP
-75-
-------
IN-STACK
FILTER
SWAGELOK
CONNECTORS
PROBE
i
WAL
CONDENSATE TRAP
VACUUM
GAUGE
FLOW
RATE
CONTROLLER
ON/OFF FLOW
VALVE
QUICK rn
ONNECT O
EVACUATED
SAMPLE
TANK
FIGURE 5-2: METHOD 25 SAMPLING TRAIN
-------
inlet on March 19 and 20, one train was assembled with a glass fiber in-stack
filter in order to assess the effects of the filter on the collected sample.
The probes were inserted through the sampling port to the center of the
duct. If conditions were favorable (i.e., the print-line was operating) the
on/off valves were opened on all sampling trains at a pre-arranged time and
sampling continued for 45 minutes. If the print-line went down before 45
minutes had elapsed, sampling was stopped and then re-started when the print-
line came back up. All stop and start times were recorded. Because of the
unpredictable operation of the print-line, 45 minutes of sampling was not
always achieved. Actual sampling times ranged from 18 to 56 minutes;. An ad-
ditional sampling tank was used at the CA outlets on March 26 on both sampling
trains to draw a larger sample because of the low duct concentrations.
During sampling, tank vacuum pressure was recorded every five or ten
minutes. At the conclusion of sampling the final tank vacuum pressure was
recorded, the probe was removed from the duct and the probe tip was plugged.
The trains were then leak checked and dissassembled. The two ends of the con-
densate trap were sealed and the trap was kept packed in dry ice until analy-
sis at the TRC laboratory.
5.4.3 Analysis for Method 25
Tanks and traps were analyzed at the TRC laboratory within 30 days of sam-
ple collection, following the EPA Method 25 procedures modified as described
in the February 1981 TRC report "EPA Method 25 Collaborative Report* (6).
Details of the analysis procedures are presented in Appendix D.
Each trap was first purged with nitrogen at ambient temperature to sweep
any C02 in the trap into its associated tank and to bring the tank pressure
from vacuum to positive (15-20 inches Eg). The tank contents were then passed
-77-
-------
through a Varian model 2800 gas chroma tog raph and flame ionization detector
(GC/FID) . When CO, CH4 and CO2 eluted, the GC column was backflushed and
all other organic compounds in the tank were passed through oxidation cata-
lysts (converting everything to C0_) and then through reduction catalysts
(converting everything to CH ) and finally through the FID for measurement
of non-methane organics (NMOha..!.) as ^4*
The trap was then heated and purged with air or oxygen to volatilize all
heavy (condensable) organics. These organics were passed through the oxida-
tion catalyst and collected as CO- in a clean, evacuated intermediate tank.
The intermediate tank contents were then passed through a reduction catalyst
and analyzed in the FID as CH (NMO. ) . The sum of NMO plus
4 trap tank
NMO equals the total non-methane organics sampled by the Method 25 train.
The GC/FID was calibrated with propane standards that were first passed
through the oxidation and reduction catalysts and then measured as methane.
5.5 Print Room Ambient Air Measurements
Measurements of ambient air VOC concentrations were made in the immediate
vicinity of the print-line and throughout the print room periodically during
each day of the measurement program. In addition, air flowrate measurements
were made in the open doorways of the print room. The purpose of these mea-
surements was to determine typical VOC concentrations in the print room during
times of print-line operation, and to estimate the amount of VOC drawn into
the print room from other plant areas through the open doorways.
VOC measurements were made with a portable H-NU model PI 101 Photoionizer
hydrocarbon analyzer. The operating principle of this analyzer is that sam-
pled organics are ionized with ultra-violet light within an electric field,
producing a current proportional to concentration. At each location in the
-78-
-------
print room the analyzer probe was held at breathing level (about 5 ifeet above
the floor) until an average concentration at that location could be estimated
on the meter readout (approximately 10-15 seconds). Measurements were made at
pre-determined representative locations throughout the print room, at the
center of the print-line (in the spaces between the print heads) and at loca-
tions 2 feet and 5 feet from two print heads. Frequent measurements were also
made directly in front of the embosser (approximately 2 feet from the cage
surrounding the embosser). The purpose of these latter measurements was to
estimate the VOC concentration of air being drawn into the embosser exhaust
hood intake.
The H-NU photoionizer was calibrated daily with the same MEK bag standards
used to calibrate the PIO analyzers. These bags were prepared every one or
two days.
Eight-hour exposure sampling was performed March 23, 25 and 26 at four
locations in the print room. Sampling was performed with Bendix model BDX 44
personal samplers using 7-cm glass charcoal tubes. All four samplers were set
out, turned on and turned off at the same time. Charcoal tubes were changed
every two hours. Sampler flowrates were approximately one liter per minute
and flowmeter readings were recorded each time the samplers were attended.
All exposed charcoal tubes were refrigerated until analysis at the TRC labora-
tory, in order to minimize any possible breakdown of captured MEK at room tem-
peratures.
The tubes were analyzed at the TRC laboratory within 30 days after the end
of the measurement program. Analysis consisted of extraction in Ccirbon dis-
ulfide and then injection into a Perkin-Blmer Sigma IB gas chromatograph
(GC). The GC was calibrated with standard solutions of MEK, MIBK and toluene.
-79-
-------
Air speed measurements were made in open print room doorways with a por-
table TSI model 1650 hot-wire anemometer and a Davis model 23B vane anemo-
meter. The only open doorway was doorway £ (opening into the main plant at
the northeast corner of the print room), except on March 18 when in addition
doorways B and D were partially open. Measurements were made at six to nine
equally spaced points in each doorway, with air speed averaged over 30 seconds
at each point. An average air speed for the entire doorway was then calcu-
lated and multiplied by the area of the doorway to determine flowrate.
The TSI hot-wire and the Davis vane anemometer were calibrated prior to
the field program in the TRC wind tunnel. Further details of the ambient air
measurements and instrument calibrations are presented in Appendix F.
5.6 Wastewater Sampling
Wastewater samples from the carbon adsorption unit were collected period-
ically during March 25 and 26, 1981. Nine samples were collected from the
common drain at the base of the carbon beds and twelve samples were collected
from a drain at the base of the water/solvent distillation column (bottom
product). Each sample was collected in a 400-ml glass jar with Teflon-lined
cap.
The nine samples from the carbon beds were collected during each phase of
the carbon bed cycle (desorption, drying, adsorption). The duration of this
cycle for each bed is approximately 120 minutes. The distillation column
operates independently from the carbon beds, and samples from this column were
collected throughout each day as time allowed.
The water samples were analyzed for MEK, MIBK, toluene and total organic
carbon (TOC) at the TRC laboratory within 30 days of sample collection. MEK,
MIBK and toluene analyses were performed with the purge-and-trap method, using
a Tekmar model LSC-2 liquid sample concentration. Nitrogen carrier gas was
-80-
-------
bubbled through a small aliquot of each sample for about 10 minutes. The car-
rier picked up the organic solvents in the sample and then passed into a gas
chromatograph for solvent analysis. The GC was calibrated with prepared stan-
dards containing known amounts of MEK, HIBR and toluene. TOC analyses were
performed using an Ionics model 445 Total Organic Carbon analyzer. Details of
the wastewater sample analyses are presented in Appendix F.
5.7 Fabric Solvent Residue
The embossing operation heats the wallcovering and may drive off solvents
remaining from the printing operation. Printed wallcovering samples were
collected before and after the embosser to estimate solvent evaporation in the
heated embosser area. Unprinted samples were also collected to determine the
background solvent content of the wallcovering.
The sampling and analytical procedure is summarized as follows:
1) Pieces of printed and unprinted wallcovering were cut and the
area of each sample was measured;
2) the samples were placed in drying tubes and heated in a small
oven for a predetermined amount of time; (
3) while heating', the tubes were purged with hydrocarbon-free air to
help remove residual solvent;
4) solvents in the air stream were collected in charcoal tubes,
desorbed with carbon disulfide, and analyzed on a gas
chromatograph.
Details of this procedure are presented in Appendix G. The sample purging
system is shown in Figure 5-3.
5.8 Effects of Process Operations on VOG Emission Measurements
The amount and distribution of VOC emitted from the printing operations at
this plant, as measured at the sampling locations used during this program, is
-81-
-------
ROTOMETER
HYDROCARBON-FREE AIR
T OVEN -7 DRYING
TUBE
CHARCOAL
COLLECTION
TUBE
BACKUP
CHARCOAL
COLLECTION
TUBE
Figure 5-3. Residual solvent in fabric purging system.
-------
affected by several factors related directly and indirectly to the process.
Factors directly related to the process include the number of print heads
operating, line speed, amount of ink used in the pattern, whether the embosser
is on or off, whether the door-hoods on each print head are open or closed,
and the frequency and extent of equipment and floor cleaning with MEK. In-
direct factors include the number of print head fans operating, whether the
make-up fan is .operating, whether the wall fans are operating, and how many
print room doors are open.
As noted in Section 3.5, the indirect factors were considered prior to the
beginning of this measurement program. The make-up fan remained off because
of its effects on substrate drying within the print heads and its effects on
air flow to the CA unit and into the print room. The adjusted wall fan (fan
belt tightened) was found to draw more air out of the print room than either
the embosser ESP or the CA unit collection system. The air flow to the CA
unit from the print room is affected by the number of print head fans opera-
ting. In order to maximize the VOC loading to the CA unit the wall fan was
therefore kept off, and all six print head fans were kept, on whenever pos-
sible. Only one of the print room doors was kept open during the measurement
program in order to control the amount of VOC entering the print room from
other areas of the plant.
Control of the direct factors by the measurement team was, of course, not
possible. Changes in the operation of the print-line were recorded in the
process log, shown in Appendix H.
-83-
-------
REFERENCES
(1) "Process Emission Tests at the General Tire and Rubber Company Vinyl-
Coated Fabric Plant in Reading, Massachusetts. September and October,
1980." EPA report 80-VNC-1A. Prepared by TRC Environmental Consultants,
Inc., under EPA Contract 68-02-3543, Work Assignment 2.
(2) CFR 40, July 1, 1980, Part 60, "Standards of Performance for New
Stationary Sources," Appendix A, pp. 183 ff.
(3) Federal Register, Volume 45, No. 194, Wednesday, December 17, 1980, pp.
83149ff.
(4) Federal Register, Volume 45, No. 194, Friday, October 3, 1980, pp. 65959
ff.
(5) Federal Register, Volume 45, No. 77, Friday, April 18, 1980, pp. 26677 ff.
(6) "EPA Method 25 Collaborative Study", dated February 1981. Prepared by
TRC Environmental Consultants, Inc., for Midwest Research Institute. TRC
project 1503-K80.
-84-
-------
&EPA
United States
Environmental Protection
Agency
Office of Air Quality
Planning and Standards
Research Triangle Park NC 27711
Air
EMB Report 80-VNC-18
July 1982
i
-Revised pages-Sept
Industrial Surface
Coating
Emission Test Report
General Tire and
Rubber Company
Reading, Massachusetts
Test Series 2
-------
1.2 Brief Process Description
Figure 1-1 presents an overhead view of the plant facilities associated
with the printing operation, and'Figure 1-2 presents a schematic of the print-
ing operation. This process is described very basically in the following par-
agraphs.
The printing operation consists of a Baker-Perkins rotogravure printing
machine utilizing six printing heads. The vinyl-coated substrate is fed
through a preliminary dryer, the six print heads, and an embossing unit. Pre-
mixed ink is supplied to each print head from a pump tank located next to each
print head. Ink is pumped from the pump tank to a tray within the print head
where a print roller, half-submerged in the tray, transfers ink from the tray
to the substrate. The inked substrate is dried in an oven contained within
each print head. Excess ink is gravity fed back to the pump tank. During a
print run, solvent or ink base is occasionally added manually to the pump
tanks to maintain the required ink viscosity. The" solvent usedTn the"inks is
primarily methyl, ethyl ketone (HEX) with some methyl isobutyl ketone (MIBK)
and toluene.
Emissions from the preliminary dryer and print head ovens are manifolded
and ducted to a carbon adsorption (CA) unit before being released to the at-
mosphere. The CA unit has three carbon beds, but only beds 1 and 3 were used
during the test program. Emissions from the embosser are controlled with an
electrostatic precipitator (ESP). Fugitive emissions within the. print-line
building are vented to the atmosphere through a pair of wall exhaust fans
and through the embosser exhaust system.
Air is supplied to the print-line building by a make-up fan .on the roof and
from seven doors that open to the outside and to other areas of the plant.
During the test program only one wall fan was operated and then only briefly.
The make-up fan was off at all times. All doors but one were closed.
-2-(Revised September 1982)
-------
2.0 SUMMARY AND DISCUSSION OF RESULTS
This section presents the results of the VOC emission tests conducted dur-
ing March 1981 at the GTR vinyl-coated fabric plant in Reading, Massachu-
setts. The purpose of these tests was to measure the controlled and uncon-
trolled VCX; emissions from the wallcovering printing and embossing operations.
VOC measurements were performed with flame ionization detection (FID) an-
alyzers at five ducted locations: carbon adsorption (CA) unit inleit, CA out-
lets to beds 1 and 3, embosser electrostatic precipitator (ESP) inlet, and
wall fan exhaust. In addition, ambient air VOC sampling was performed1 in the
print-line building with a portable photoionizer hydrocarbon analyzer.
VOC sampling was performed at the embosser ESP inlet, wall iian exhaust
duct and. CA unit inlet on March 18, 19, 20, and 23, 1981, Surveys of the ambient
VOC concentrations in the print room, near the print line, and near the embosser
hood intake were conducted each testing day. v00 samPlin(3 was performed at the
CA unit inlet and outlets on March 25 and 26, 1981, to determine the control
efficiency of the CA unit. No measurement work was performed on March 24 be-
cause no wallcovering was printed that day.
«
2.1 Summary of Results
VOC concentrations and air flowrates were measured at'the embosser ESP
inlet, wall fan exhaust duct, CA unit inlet and CA unit outlets during print-
line operations. Ambient air VOC measurements were made inside the print-line
building (print room). The results of this measurement program showed that:
1. Under the operating conditions of this measurement program
(make-up air fan and' wall- exhaust fans off), the majority of
print-line VOC emissions 'is ducted to the CA inlet.
2. The VOC ducted to the embosser ESP inlet is a combination of em-
bosser-generated VOC and ambient print room VOC.
_7_ (Revised September 1982)
-------
TABLE 2-2 (Continued)
SUMMARY OF FID VOC EMISSIONS FROM PRINTING OPERATIONS
AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Production
Date Order Number Process Operations
3-20-81 T-15521 Completing Previous Run
Completing Previous Run
Preparation for T-15521
Color Matching
Printing Start/Stop
for Repairs
Printing, Embosser On
1000 Yards Printing
Printing Start/Stop
for Repairs
1000 Yards Printing
1000 Yards Printing
1000 Yards Printing
Run Completed
Cleaning Print Heads
Clean Op
TOTAL PRINT TIME
TOTAL RUN TIME
Time Interval
Start9 End
0740.
0744'
0814
0958
1019
1146
1150
1212
1256
1318
1340
1402
1410
1426
1019
0958
0744
0814
0958
1019
1148
1150
1212
1256
1318
1340
1402
1410
1426
1532
1410
1410
Total
Minutes
4
30
104
21
89
2
22
44
22
22
22
8
16
66
231
252
VOC Emissions (Pounds as MEK)
Embosser
0.67
3.42
4.56
1.91
7.97
0.233
4.32
6.76
4.27
4.25
4.37
1.26
2.09
6.06
33.4
35.3
Hall Fan
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
CA Inlet
NM
19.4
14.7
31.2
35.6
1.16
18.2
29.5
17.27
17.6
18.93
5.61
7.49
NM
143.8
175.0
Total
__
22.8
19.3
33.1
43.6
1.39
22.5
36.3
21.5
21.8
23.3
6.87
9.58
177.2
210.3
NM: Not measured - analyzer problems or calibrations in progress.
a Start time for the Initial time Interval Is the time when FID monitoring began that day.
H-
tn
(D
O.
<£>
0
ft
ffi
n>
h
ft
-------
TABLE 2-2 (Continued)
SUMMARY OF FID VOC EMISSIONS FROM PRINTING OPERATIONS
AT GENERAL TIRE AND RUBBER COMPANY
READING, MASSACHUSETTS
Ul
I
Production
Date Order Number Process Operations
3-23-81 T-15516 Printing in Progress
1000 Yards Printing
1000 Yards Printing
1000 Yards Printing
Run Completed
TOTAL PRINT TIME
TOTAL RUN TIME
T-15519 Threading Leader
Cleaning. PH Fans off.
Hall Fan on
Color Matching,
Heb Alignment
Hall Fan Off.
Color Matching
Printing Line Down Once
1000 Yards Printing
Printing
Line Up and Down.
Trimming Problems
Problems Persist.
Run Ended
Repairs
Repairs
TOTAL PRINT TIME
TOTAL RUN TIME
Time Interval
Start8 End
0850
0909
0931
0953
1015
0850
0850
1025
1037
1239
1244
1324
1351
1413
1423
1605
1628
1633
1324
1239
0909
0931
0953
1015
1025
1025
1025
1037
1239
1244
1324
1351
1413
1423
1605
1628
1633
1636
1605
1628
Total
Minutes
19
22
22
22
10
95
95
12
122
5
40
27
22
10
102
23
5
3
161
229
VOC
Embosser
1.92
2.60
2.78
3.12
1.48
11.9
11.9
1.41
5.25
0.070
0.85
1.88
1.56
0.64
6.07
1.14
0.27
KM
10.2
12.2
Emissions (Pounds as HER)
Hall Fan CA Inlet
Sampling 10.51
Discontinued 13.06
14.70
14.20
7.15
59.6
59.6
24.60
4.17
0.180
10.3
9.90
9.02
4.17
35.8
10.5
1.16
1.93
58.9
69.6
Total
12.43
15.7
17.5
17.3
8.63
71.5
71.5
26.0
9.42
0.250
11.2
11.8
10.6
4.81
41.9
11.6
1.43
69.1
81.8
NMt Not measured - analyzer problems or calibrations in progress.
a Start time for the initial time Interval Is the time when FID monitoring began that day.
------- |