xvEPA
United States ( Office of Air Quality
Environmental Protection -
Agency
Planning and Standards
Research Triangle Park NC
27711
EPA-450/3-79-005b
September 1980
Air
Glass Manufacturing EIS
Plants
Background Information
for Promulgated
Standards of Performance
MtHtasal
-------
'
-------
EPA-450/3-7a-005b
Glass Manufacturing Plants
Background Information for
Promulgated Standards
of Performance
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air, Noise, and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
September 1980
-------
This report has been reviewed by the Emission Standards and Engineering
Division, Office of Air Quality Planning and Standards, Office of Air, Noise,
and Radiation, Environmental Protection Agency, and approved for publica-
tion. Mention of company or product names does not constitute endorsement
by EPA. Copies are available free of charge to Federal employees, current
contractors and grantees, and non-profit organizations - as supplies permit
from the Library Services Office, MD-35, Environmental Protection Agency,
Research Triangle Park, NC 27711; or may be obtained, for a fee, from the
National Technical Information Service, 5285 Port Royal Road, Springfield,
VA 22161.
Publication No. EPA-450/3-79-005b
11
-------
ENVIRONMENTAL PROTECTION AGENCY
Background Information
and Final
Environmental Impact Statement
for Glass Manufacturing Plants
Prepared by:
Don R. Goodww
Director, Emission Standards and Engineering Division
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
(Date)
1. The promulgated standards of performance will limit particulate
matter emissions from new, modified, and reconstructed glass
manufacturing plants. Section 111 of the Clean Air Act
(42 U.S.C. 7411), as amended, directs the Administrator to establish
standards of performance for any category of new stationary
source of air pollution that ". . . causes or contributes signi-
ficantly to air pollution which may reasonably be anticipated to -.-
endanger public health or welfare." Approximately 17 States
located in all areas of the nation will be affected by these
standards.
2. Copies of this document have been sent to the following Federal
Departments: Labor, Health and Human Services, Defense,
Transportation, Agriculture, Commerce, Interior, and Energy; the
National Science Foundation; the Council on Environmental Quality;
members of the State and Territorial Air Pollution Program Adminis-
trators; the Association of Local Air Pollution Control Officials;
EPA Regional Administrators; and other interested parties.
3. For additional information contact:
Ms. Susan R. Wyatt
Standards Development Branch (MD-13)
-U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
telephone: (919) 541-5421
4. Copies of this document may be obtained from:
U.S. EPA Library (MD-35)
Research Triangle Park, NC 27711
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
i i 1
-------
-------
TABLE OF CONTENTS
1.0 SUMMARY 1-1
1.1 Summary of Changes Since Proposal 1-2
1.2 Summary of Impacts of the Promulgated Action .... 1-4
1.2.1 Environmental Impacts of the Promulgated Action . . 1-4
1.2.2 Energy Impacts of the Promulgated Action 1-4
1.2.3 Economic Impacts of the Promulgated Action .... 1-6
2.0 SUMMARY OF PUBLIC COMMENTS ' 2-1
2.1 Need for Standards 2-1
2.2 Emission Control Technology 2-15
2.3 Modification, Reconstruction, and Other
Considerations 2-52
2.4 General Issues 2-61
2.5 Environmental Impact 2-69
2.6 Economic Impact 2-77
2.7 Energy Impact 2-99
2.8 Test Methods and Monitoring 2-104
2.9 Clarifications 2-107
Appendix A A-l
Appendix B B-l
Appendix C C-l
Addendum to the Glass Manufacturing Plants Background
Information;proposed Standards of Performance. D-l
LIST OF TABLES
Page
1-1 Matrix of Environmental and Economic Impacts of
the Promulgated Action 1-5
2-1 List of Commenters on the Proposed Standards of
Performance for the Glass Manufacturing Industry 2-2
LIST OF FIGURES
/Page
2-1 Economic Impact Decision Flow Chart 2-81
-------
-------
1. SUMMARY
On July 20, 1977, a notice of intent to develop standards of
performance for glass manufacturing plants was published in the Federal
Register (42 FR 37213). Prior to proposal of the standards, interested
parties were advised by public notice in the Federal Register
(43 FR 11259, March 17, 1978) of a meeting of the National Air Pollution
Control Techniques Advisory Committee to discuss the glass manufacturing
plant standards recommended for proposal. This meeting occurred on
April 5-6, 1978. The meeting was open to the public and each attendee
was given an opportunity to comment on the standards recommended for
proposal. As a result of this meeting several changes were made to
the recommended standards. On June 14, 1979, glass manufacturing
plants was added to the list of categories of stationary sources which
the Administrator has determined may contribute significantly to air
pollution which causes or contributes to the endangerment of public
health or welfare (44 FR 34193).
On June 15, 1979, the Environmental Protection Agency (EPA)
proposed standards of performance for glass manufacturing plants
(44 FR 34840) under authority of Section 111 of the Clean Air Act.
Public comments were requrested on the proposal and on the listing of
glass manufacturing plants in the Federal Register publication.
Thirty-three comment letters Were received and 11 interested parties
testified at the public hearing. These comments were made by glass
manufacturers; an ad hoc industry group; a trade association; local,
State and Federal government offices; and an environmental group. The
comments that were submitted, along with the responses to these comments,
are summarized in this document. The summary of comments and responses
serves as the basis for the revisions that have been made to the
standards between proposal and promulgation.
1-1
-------
1.1 SUMMARY OF CHANGES SINCE PROPOSAL
A number of changes of varying importance have been made since
proposal of these standards. One of the most significant of these is
the change made to the regulation in response to comments that the
establishment of the pressed and blown (other-than soda-lime) category
had been baseJ on an incorrect evaluation of the data. The pressed
and blown glass melting standards are now set out in three subcategories:
(1) borosilicate, 1.0 Ib/ton (2) soda-lime and lead, 0.2 Ib/ton and
(3) other-than borosilicate, soda-lime, and lead, 0.5 Ib/ton. These
standards, as well as the others, are based on data which show their
achievability. The definitions of the terms "pressed and blown glass,"
"lead recipe," and "borosilicate recipe" in Section 60.291 of the
regulation have also been redrafted or added as a result of these
changes. Further study and analysis also showed that the limits
promulgated for the flat glass and wool fiberglass categories should
be 0.45 Ib/ton and 0.5 Ib/ton, respectively. Where appropriate, the
applicable impacts have been adjusted to reflect these uncontrolled
emission rate and limitational changes. Refer to the Emission Control
Technology section (2.2) of this document.
An adjustment was provided for in the regulation for the calculation
of the furnace emission rate in order to take into consideration the
fact that there are emissions at times of no production due to the
maintenance of the molten glass at the proper temperature. Refer to
the Emission Control Technology section (2.2) of this document.
A change was made to the proposed test method. The promulgated
standards require that the filter box temperature for EPA Method 5
tests be maintained at up to 350°F as opposed to the proposed tempera-
ture of 250°F. Section 60.296(b) was redrafted accordingly. Refer to
the Test Methods and Monitoring section (2.8) of this document.
Another change was made in response to comments that the definition
of the affected facility was ambiguous. Section 60.290 of the Regulation
was changed to clarify the definition of the affected facility. Refer
to the Clarifications section (2.9) of this document.
The introductory sentence of Section 60.291 of the regulation was
changed to better define the meanings of the Section's defined terms.
1-2
-------
Table 2-1. LIST OF COMMENTERS ON THE
PROPOSED STANDARDS OF PERFORMANCE FOR THE
GLASS MANUFACTURING INDUSTRY (CONTINUED)
Commenter*
D-23
D-26
D-27
D-29, D-33
D-30, F-l
D-31
Affiliation
Ronald A. Friesen, Chief
Industrial Project Evaluation and
.Control Safety Development Branch
State of California Air Resources Board
1-102 Q Street
Post Office Box 2815
Sacramento, California 95812
William V. Skidmore
Acting Deputy General Counsel
General Counsel of the United States
Department of Commerce
Washington, D.C. 20230
Daniel J. Goodwin, Manager
Division of Air Pollution Control
Illinois Environmental Protection Agency
2200 Churchill Road
Springfield, Illinois 62706
Joseph V. Saliga, President
Society for Glass Science and Practices
Post Office Box 166
Clarksburg, West Virginia 26301
Charles N. Frantz
Acting Manager
Environmental Control Department
Owens-Illinois
Post Office Box 1035
Toledo, Ohio 43666
Jeff Jacobson
Plant Engineer
Guardian Industries Corporation
11535 East Mountain View
Kingsburg, California 93631
2-5
-------
Table 2-1. LIST OF COMMENTERS ON THE
PROPOSED STANDARDS OF PERFORMANCE FOR THE
GLASS MANUFACTURING INDUSTRY (CONCLUDED)
Commenter*
F-l, 6-3, 6-5
F-l, 6-4
F-l
F-l
Affiliation
Roger Strelow
E. Donald Elliot
Ad Hoc Glass industry Air Quality Group
Leva, Hawes, Symington, Martin & Oppenheimer
815 Connecticut Avenue, N.W.
Washington, D.C. 20006
Ronald Moore
Glass Container Division
Engineering & Research Department
Owens-Illinois
Post Office Box 1035
Toledo, Ohio 43666
Larry Sculley, Manager
Environmental and Natural Resources Practice
Peat, Marwick & Mitchell & Company
Washington, D.C.
Domhnall OBroin
Society for Glass Science and Practices
Post Office Box 166
Clarksburg, West Virginia 26301
2-6
-------
92930
A~asji3[> M3N j.
*3iuA"g 'i UBpuaug
ZTTfrfr
A"uediiioo
pue qoueasay
90UE
uoi.suaq.X3
luaiudo
uqop
'uia|jes-UOisuLM
>|aaa3 SB us £008
.o UOLSLALQ
A"q.i.unuiuioo pus saounosay
[Bq.uaaiuojLAU3 -ap *
6ui.uaauL6u3
-\
9000Z 'O'Q '
M'
aq.nq.iq.sui
008T
60EZZ
pjiBAainog uo§[l-M 9TSI
uoj.q.BLoossv SBQ ueouaui\/
q.uapisaJd
'opa.01
69991;
ILOLO
uoiq.BJodaoo
S30LAU3S
660t
u^sw
331A
OZ-Q
6T-G
8T-G
I-d *
AyisnaNi
3Hi yod 30NVwyojy3d do
3Hi NO SiGBGWWQO JO 1SI1 'T'Z
-------
E-Z
peoy J-LOM 09
uoi.q.BAaasuoo [ieq.U9iuuoai.AU3
j.o q.uauiq.aBdaQ aq.Bq.$ >|aoA Maj\]
JLty J.O UOLSLALQ 'aoq.OaaLQ
"3'd c>a[>"
9ZI817
aug
pjioj
'd
9ZI817
BuuaauLBug
aug
puoj
PUB LBq.uauiuojLAU3
oq.oaa.LQ 'UBiussns 'H*A
ZIZ08
_
Leq.uauiuoa.LAU3 A"q.i.uniuuioo
* pa Laj. 631.5 'NT
92980 A^asaap Ma^j 'uoq.uaai
'ezBLd M^l-d "Mop
AU3 j.o UOLSLALQ
uo.Lq.oaq.oad L^1U91UU0^.LAU3
j.0 q.uauiq.aBdaa*aq.Bq.s A'asaap
aoq.oaa.LQ 'a
TOfrOS
A"uBduioo
euapL00
j.0 aoq.oaai.Q
SLM31
Z6ZZZ
qq.aoN 'uoq.6ui.xaT
xog 30J4J.O q.sod
paq.Baodaooui *sai.aq.snpui gdd
uo.Lq.BnLBA3 ssaooad SULUUO-I
q.s.Luiai)o aoLuas ''
azi-n
ZT-Q
T-d 'TT-Q
OT-Q
6-Q
8-Q
AyisnaNi SNiyniovdnwvw
3H1 yOd 30NWHOdH3c! dO SQy\/QNVlS
3H1 NO SH31N3WW03 dO 1SH "T'Z
-------
Sections 60.293 through 60.295 are now reserved, and proposed
Section 60.293 (Test Methods and Procedures) is now Section 60.296.
Refer to the Clarifications section (2.9) of this document for details.
The term "glass manufacturing plant" was deleted from Section 60.291.
This term was determined to be unnecessary.
The term "glass production" has been replaced in Section 60.291
by the term "glass produced" to better describe the basis upon which
compliance is to be determined.
The term "pot furnace" has been deleted from Section 60.291.
This was done partially in response to a comment from the industry
that the glass melting furnace described in the proposed definition is
a "day tank." It was also pointed out that the designed production
capacity for day tanks is more properly represented by 4,550 kilograms
of glass produced per day. The definition was deleted as being
unnecessary and the exemption was changed. Refer to the Modification,
Reconstruction, and Other Considerations section (2.3) and Clarifications
section (2.9) of this document.
The term "hand glass melting furnace" was added to Section 60.291
in order to define the new category of glass melting furnace being
exempted from compliance with the standards. Refer to the Modifications,
Reconstruction, and Other Considerations section (2.3) of this document.
The term "recipe" was deleted from Section 60.291 and sodium
sulfate was included as a miscellaneous material for the "soda-lime
recipe" definition.
Section 60.292(b) of the proposed regulation has been redrafted
to more precisely describe the process by which the heating value for
liquid and gaseous fuels are to be determined"and has been changed to
Section 60.296(f).
In response to additional comments from industry, the uncontrolled
emission rates used in the impact analyses of the proposed standards
were changed for both the container glass and flat glass categories.
The representative uncontrolled emission rate used for proposal for
container glass melting furnaces was 1.5 Ib/ton, rather than 2.5 Ib/ton.
The representative uncontrolled emission rate used for proposal for
flat glass melting furnaces was 2.0 Ib/ton, rather than 3.0 Ib/ton.
Refer to the General Issues section (2.4) of this document.
1-3
-------
1.2 SUMMARY OF IMPACTS OF THE PROMULGATED ACTION
1.2.1 Environmental Impacts of the Promulgated Action
The promulgated standards will reduce projected 1984 emissions
from new uncontrolled glass melting furnaces from about 4,890 megagrams
per year [Mg/yr] (5,390 tons/yr) to about 550 Mg/yr (610 tons/year).
This is a reduction of about 90 percent of uncontrolled emissions.
Meeting a typical State Implementation Plan (SIP), however, will
reduce emissions from new uncontrolled furnaces by about 3,150 Mg/yr
(3,475 tons/yr). The promulgated standards will exceed the reduction
achieved under a typical SIP by about 1,190 Mg/yr (1,310 tons/yr).
This reduction in emissions will result in a reduction of ambient air
concentrations of particulate matter in the vicinity of new glass
manufacturing plants.
The promulgated standards are based on the use of electrostatic
precipitators (ESPs) and fabric filters, which are dry control tech-
niques; therefore, no water discharge will be generated and there will
be no adverse water pollution impact.
The solid waste impact of the promulgated standards will be
minimal. Less than 2 Mg (2.2 tons) of particulate will be collected
for every 1,000 Mg (1,102 tons) of glass produced. In some cases,
this material can be recycled, or it can be landfilled if recycling
proves unattractive. The current solid waste disposal practice among
most controlled plants surveyed is landfill ing. Since landfill opera-
tions are subject to State regulation, this disposal method is not
expected to have an adverse environmental impact. The additional
solid material collected under the promulgated standards will not
differ chemically from the material collected under a typical SIP
regulation; therefore, adverse impact from landfilling will be minimal.
Also, recycling of the solids has no adverse environmental impact.
The environmental impacts of the promulgated, standards are summarized
in Table 1-1.
1.2.2 Energy Impacts of the Promulgated Action
For model plants in the glass manufacturing industry, the total
increased energy consumption that will result from the promulgated
standards, including the amount attributable to SIP, ranges from about
1-4
-------
TO
E
O
03
LU
O
z
o
UJ
Q
«=c
-a oj
i- +J
t V)
o -a
c in s-
o o TCS
»-* o c
CL s- ro
O CL. 4->
X
X
X
X
X
X
X
X
X
X
+J
a.
o
t/>
a
m to
"O "O
s- c
(« re
a +J
ra o
to >
O n3
a>
a
u
Q. 03
o
ra
»-> a.
o E
-
E s-
4-> (U
$-
o c s-
JT o s-
XXX
X X
X
o
rtj »->
a. o
E 10
« i -M O-4->
O E o
t 0. Qt.
o J3 E
Cn ro -O 5-
o a> E o «a
O «-H CJ CO
O
.
E o
i ia
O.
5a
u a
r- 10
t- i-
O)
QJ -O
oo«a;
i
+ I
-------
0.1 to 2 percent of the energy consumed to produce glass in new plants.
The energy required in excess of that required by a typical SIP regulation
to control all new glass melting furnaces constructed by 1984 to the
level of the promulgated standards will be about 9.13 x 10
kilowatt-hours per year in the fifth year and is not considered significant.
Thus, the prorulgated standards will have a minimal impact on national
energy consumption.
1.2.3 Economic Impacts of the Promulgated Action
Compliance with the standards will result in annualized costs in
the glass manufacturing industry of about $8.5 million by 1984.
Cumulative capital costs of complying with the promulgated standards
for the glass manufacturing industry as a whole will amount to about
$28 million between 1979 through 1984. The percent price increase for
products from new plants necessary to offset costs of compliance with
the promulgated standards will range from about 0.3 percent in the
wool fiberglass category to about 1.8 percent in the container glass
category. Industry-wide, the price increase for products from new
plants will amount to about 0.7 percent. The economic impacts of the
promulgated standards are summarized in Table 1-1. These economic
impacts are reasonable.
1-6
-------
2.0 SUMMARY OF PUBLIC COMMENTS
The list of commenters and their affiliations is shown in Table 2-1.
Thirty-three letters contained comments and 11 people testified at the
public hearing relative to the proposed standards and Volume I of the
Background Information Document. The significant comments have been
combined into the following nine major areas:
1. Need for Standards
2. Emission Control Technology
3. Modification, Reconstruction, and Other Considerations
4. General Issues
5. Environmental Impacts
6. Economic Impacts
7. Energy Impacts
8. Test Methods and Monitoring
9. Clarifications
The comments and issues and the responses to them are discussed
in the following Section of this chapter. A summary of the changes to
the standards is included in Section 1.1.
2.1 NEED FOR STANDARDS
Several commenters questioned the need for standards of performance
for the glass manufacutring industry. Standards of performance are
promulgated under Section 111 of the Clean Air Act. Section lll(b)(l)(A)
requires that the Administrator establish standards of performance for
categories of new, modified, or reconstructed stationary sources which
in his judgment cause or contribute significantly to air pollution
which may reasonably be anticipated to endanger public health or
welfare. The overriding purpose of standards of performance is to
prevent new air pollution problems from developing by requiring the
2-1
-------
Table 2-1. LIST OF COMMENTERS ON THE
PROPOSED STANDARDS OF PERFORMANCE FOR THE
GLASS MANUFACTURING INDUSTRY
Commenter*
D-2B, D-15, F-l, G-l
D-4B
D-5
D-6
D-6, D-7, F-l, G-2
D-7A
Affiliation
Werner Ganz
Director of Engineering - Facilities
Libbey-Owens-Ford Company
Technical Center
1701 East Broadway
Toledo, Ohio 43605
Fresno County Air Pollution Control District
1246 "L" Street
Fresno, California 93721
George B. Zurheide, Vice President
Environmental Engineering
CertainTeed Corporation
Post Office Box 1100
Blue Bell, Pennsylvania 19422
Daniel S. Welebir, R.S., MPH
Environmental Health Director
San Joaquin Local Health District
1601 East Hazel ton Avenue
Post Office Box 2009
Stockton, California 95201
J.T. Destefano, Director
Safety, Health and Environmental Affairs
PPG Industries, Incorporated
1000 RIDC Plaza
Box 2811
Pittsburgh, Pennsylvania 15230
G.H. Mosely, Manager
Environmental Control
Corning Glass Works
Corning, New York 14830
*These designators represent docket entry numbers for Docket OAQPS 79-2.
These docket entries are available for public inspection at:
U.S. Environmental Protection Agency, Central Docket Section
Room 2902
Waterside Mall
401 M Street, S.W.
Washington, D.C. 20460
2-2
-------
application of the best technological system of continuous emission
reduction which the Administrator determines to be adequately demon-
strated. The 1977 Amendments to the Clean Air Act added the words,
"in the Administrator's judgment," and the words, "may reasonably be
anticipated," to the statutory test. The legislative history for
these changes stresses two points: (1) the Act is preventive, and
regulatory action should be taken to prevent harm before it occurs;
and (2) standards should consider the cumulative impact of sources and
not just the risk from a single class of sources.
The 1977 Amendments to the Clean Air Act also required that the
Administrator promulgate a priority list of source categories for
which standards of performance are to be promulgated. The Priority
List, 40 CFR 60.18, was proposed in the Federal Register on August 31,
1978 (43 FR 38872). Glass manufacturing was ranked thirty-eighth on
that list. On June 14, 1979, the Administrator listed glass manufac-
turing (44 FR 34193) among the categories of stationary sources which
contribute significantly to air pollution which causes or contributes
to-the endangerment of public health or welfare. Even though glass
manufacturing had been included on the proposed priority list, it was
listed because the priority list had not been finalized.
Commenters questioned the ranking of glass manufacturing as
thirty-eighth on the proposed priority list and questioned basing the
decision to add glass manufacturing to the list of significant source
categories on the proposed priority list. Development of the priority
list was initiated by compiling data on a large number of source
categories from literature resources. Major stationary source categories
were then subjected to a priority ranking procedure using the three
criteria specified in Section lll(f) of the Act. The procedure ranks
source categories on a pollutant-by-pollutant basis. In this ranking,
first priority was given to the quantity of emissions, second priority
was given to the potential impact on health or welfare, and third
priority was given to mobility. This procedure resulted in glass
manufacturing being ranked thirty-eighth on the proposed priority
1 i st.
2-7
-------
The ranking of glass manufacturing on the proposed priority list
was reviewed when deciding to establish new source performance standards
for glass manufacturing plants. However, this review was made only
for comparison with other actions that EPA was considering. The
Priority List was only used for comparison purposes because the Priority
List had not been finalized.
Another study was conducted to investigate the glass manufacturing
industry in more detail. This study resulted in the development of a
Background Information Document (BID), Volume I which specifically
addressed the industry in terms of its structure, processes, and
emission control techniques. The BID also described modification and
reconstruction, alternative regulatory options, and the environmental,
economic, and energy impacts that would be associated with the
implementation of the various regulatory options. ^The decision to
develop these standards was based on the BID study but weighed factors
similar to those considered in the development of EPA's Priority List.
National glass production is estimated to grow annually at a rate of
up to 7 percent through 1984; the industry, while concentrated in
17 States, is not geographically tied to either markets or resources,
thereby having relative mobility; and the industry is a significant
contributor to air pollution, as well as having a high ranking with
regard to potential emission reduction. These factors, as discussed
in the preamble (44 FR 34841) to the proposed standards, led to the
development of these standards of performance. Each of these factors
was questioned by commenters.
Commenters questioned the glass manufacturing industry's ability
to locate its plants in order to avoid stringent SIP regulations. In
adding glass manufacturing to the list of stationary source categories,
EPA also explained that new glass manufacturing operations could be
located in States which have less restrictive SIP regulations. Industry
commenters explained that raw material, customer, and financial con-
siderations were much more important in determining plant location
than the stringency of a particular State's environmental regulatory
scheme.
2-8
-------
All of these factors need to be taken into consideration in
deciding where to construct a new facility. . What was meant to be
emphasized was the relative flexibility in location that a glass
manufacturer has in locating a new plant. Manufacturers who have the
freedom to locate a new plant with only minor restrictions caused by
raw material suppliers and product market are considered to have
mobility. Glass manufacturing plants are not restricted to locating
in a particular region of the country as would a coal mine or a stone
quarry. For this industry, raw materials and glass products can be
and are shipped across the country.
The glass industry, due to its relative mobility, could readily
locate in States with less stringent standards or compliance deadlines.
This has in fact occurred in at least one State where, at a public
hearing, a glass industry representative specifically suggested that
his company would relocate and construct new plants in another State
to avoid having to "spend multi-mi 11 ion dollars for air pollution
control equipment." This was shown to be somewhat of a trend by the
State involved when it was found that in the past five years in excess
of 10 percent of the State's glass furnances have been shutdown and no
new ones constructed (docket entry OAQPS 77/1-IV-D-10). This is
especially significant while considering the glass industry's nationwide
production increases in the past several years. One purpose of these
nationally applicable standards is to avoid situations in which
industries could be lured to one State from another just by virtue of
there being a less stringent regulation in effect.
Commenters suggested that particulate emissions from glass
manufacturing plants do not contribute significantly to air pollution.
These commenters explained that the estimated reduction of 1,620 tons/yr
of particulate emissions from glass manufacturing plants is small in
comparison to the total quantity of nationwide particulate emissions
and to the quantity of emissions from other industries.
Almost any industry by itself accounts for a small portion of the
Nation's total emissions. The 1,620 tons/yr. estimate of emissions
reduced by the proposed standards was the quantity attributable to the
proposed standards and neglected the emission reduction attributable
to SIP regulations. In addition, this quantity only applied to glass
2-9
-------
manufacturing plants estimated to be built within a 5 year period,
i.e., new sources. The total emissions reduced by the promulgated
standards for new sources, including the emissions reduced by SIP
regulation, is about 4,800 tons/yr. in the fifth-year. Also, these
emissions only originate in about 20 States.
Specifically referred to by the commenter was the National Asphalt
Pavement Association case [539 F.2d 775 (1976)]. The commenter implied
that EPA, in that case, relied for its "significance" determination on
a 1967 study estimating controlled emissions from the subject industry
to increase from 243,000 tons/year to 403,000 tons/year in 1977. That
would be the fifth year annual particulate emissions increase of
160,000 tons. The commenter went on to compare the study's increase
to the increase in the amount of particulate emissions estimated by
EPA to be emitted in 1983 by glass manufacturing plants having to meet
typical SIP requirements (1,620 tons/yr), noting the difference between
160,000 tons/yr and 1,620 tons/yr.
What the commenter failed to note was the Court's determination
in "National Asphalt," in the paragraph following the discussion of
the study, that the Administrator never relied on the disputed 1967
study to estimate the industry's level of controlled emissions. The
Court instead found that the Administrator based his decision to
develop standards on uncontrolled emissions, stringencies of SIP's,
the number of existing plants, and the expected rate of growth in the
number of plants. At no time in the "National Asphalt" opinion did
the Court find a specific quantity of emissions to be "significant" as
alleged by the commenter.
A figure that was offered to the Court, in "National Asphalt,"
for consideration was an estimate, made by the petitioning company, of
the industry's total annual particulate emissions for 1972 amounting
to 40,000 tons. Noting this figure, it is interesting to compare an
estimation made by EPA that in 1976 the total annual particulate
emissions for the glass manufacturing industry were approximately
20,000 tons.
Regardless of the quantity of emissions estimated to be
attributable to an industry, the Administrator is called upon in the
Clean Air Act to evaluate an industry's contribution to air pollution
2-10
-------
and make a determination as to the significance of the subject industry's
emission contribution. A firm definition of "significant" contribution
cannot be applied nationwide due to variations in topography, emission
source distribution, and heights from which the pollutant is emitted.
Instead it is necessary that the Administrator make a determination of
"significance" on an industry-by-industry basis. However, general
criteria are presently being used to develop standards for sources on
the priority list. If a source category includes sources that may
emit about 110 Mg per year and show growth potential, then standards
are being developed. In addition, as noted in the preamble (44 FR 49223)
to the final priority list (40 CFR 60.16); the Administrator may
develop standards for sources not on the priority list, especially
certain minor sources.
In the case of the glass manufacturing industry, the Administrator
has determined that particulate emissions from new or modified glass
manufacturing plants will contribute significantly to air pollution,
even though the total amount of emissions is a small portion of the
Nation's total particulate emissions and this industry has been
determined to significantly contribute to the Nation's total emissions.
The quantity of emissions for an industrial category is not the
only criterion considered. With regard to public health and welfare,
the submicron size of most glass melting furnance-generated particu-
lates, among other factors, requires consideration, (docket entry
OAQPS-77/1 IV-D-10.) Of special concern is the capability of these
submicron particles to by-pass the body's respiratory filters and
penetrate deeply into the lungs. In excess of 30 percent of the
particles less than 1 micrometer in size that penetrate the pulmonary
system are deposited there. These particulates also have fairly long
lives in the atmosphere and can absorb toxic gases, thus leading to
potentially severe synergistic effects when inhaled.
A report prepared by the National Academy of Sciences discussed
that while population and regional variables such as temperature,
relative humidity, nutritional state, exercise, and coexistent
pulmonary and circulatory disease should be taken into account
in analyzing the effects of inhaled particles, particles that are
2-11
-------
soluble in the respiratory tract fluid, such as the ones emitted by
glass melting furances, systemic uptake may be relatively complete
for all deposition patterns. It was emphasized that as a result of
this exposure local toxic and/or irritant effects may result.
Commenters suggested that particulate emissions from glass
manufacturing plants do not contribute significantly to air pollution
because Class I Prevention of Significant Deterioration (PSD) incre-
ments were not exceeded. The fact that emissions from a single plant
would be less than the Class I PSD increment does not show that the
category should not be listed. First, the test is whether the category,
not an individual plant, contributes significantly. Second, although
a single plant might not exceed a Class I increment, it could contribute
significantly to total level of emissions in excess of the increment.
Most importantly, the major purpose of Section 111 is to "prevent new
air pollution problems." National Asphalt Pavement Association v. Train.
539 F.2d 775, 783 (D.C. Circ., 1976). That is, new source performance
standards should prevent standards of PSD increments from being threatened
by requiring maximum control of new sources. It is, therefore, not
necessary to show that individual sources in the category would violate
an increment.
Commenters questioned the establishment of standards for glass
manufacturing plants before the establishment of standards for source
categories with a higher priority. The priority ranking in 40 CFR 60.16
is indicated by the number to the left of each source category and is
used to decide the order in which new projects are initiated. However,
this is not necessarily an indicator of the order in which projects
will be completed. The establishment of these standards began before
the priority list was proposed, as indicated in the notice of intent
to develop an NSPS published in the Fgderal Register on July 20, 1977
(42 FR 37213). Therefore, because the establishment of these standards
began before the priority list was promulgated, establishment of glass
manufacturing standards before establishment of standards for source>
categories with a higher priority is appropriate. It would also.be
pointless and wasteful to postpone the development of a source category
2-12
-------
project already started simply because the source category is lower on
the priority list than other source categories.
Based on the judgment that particulate air pollutants from glass
melting furnaces contribute significantly to air pollution, which may
be reasonably anticipated to endanger public health or welfare, the
EPA listed glass manufacturing plants as a source category necessitating
the establishment of new source performance standards. Comments, as
discussed above, have not led to a change in this decision.
Several commenters suggested that standards of performance should
have been developed for NO and S0? emissions, in addition to standards
f\ *
for particulate emissions. In deciding to regulate particulate emis-
sions only, consideration was given to the possible, regulation of N0x
and S02 emissions. The relative contributions to air pollution that
NO and S0? emissions from glass manufacturing plants present is
/\ ^
recognized and has been considered. However, the analysis of the
glass manufacturing industry did not find control techniques for N0x
and S02 adequately demonstrated. Therefore, new source performance
standards were not proposed for the control of NOX and S02.
Several States have enacted or are in the process of enacting
regulations mandating the control of NO and S02 emissions from manu-
facturing processes such as glass manufacturing. Means of control are
being developed by industry and the Office of Research and Development
and applied for both NO and S09 in an attempt to comply with these
X £
State regulations. Control techniques for N0x and S02 may be demonstrated;
however, the presently available analysis has not evaluated these
techniques fully. If, at the time of the fourth-year regulatory
review, control techniques are found to be adequately demonstrated for
the control of NO and SO,, emissions from glass manufacturing plants,
X £
an in-depth review of standard of performance for N0x and S02 from
glass manufacturing plants could be undertaken.
Commenters suggested that standards of performance should have
been developed for S02 rather than standards for particulate matter.
These commenters pointed out that the ranking of pollutants used to
establish EPA's proposed priority list was in the following order for
glass manufacturing: S02, particulate matter, and NOX. They explained
2-13
-------
that this ranking, which assigns a relative priority to pollutants
based upon the potential impact of NSPS, should lead the EPA to develop
S02 standards before particulate matter standards. However, the EPA
considers it appropriate to base the listing decision partially on one
form of emissions (i.e., S02), and yet propose a standard for one of
the others (i.e., NO or particulate matter).
J\
Again, as stated previously, the analysis of the glass manufacturing
industry did not find control techniques for emissions other than
particulates to be adequately demonstrated. Due to limited resources
and the immediate feasibility of setting standards for the control of
particulate emissions, it was decided to proceed with the development
of these regulations. It was considered to be impractical and wasteful
to delay the progress of these standards to wait for the possibility
that some means of control for the other pollutants (S0"2 and N0x)
could be immediately investigated.
Several commenters mentioned that certain emissions, such as
fluoride, boron, and lead emissions, were not specifically regulated.
These emissions are unique to glass manufacturing plants in a small
portion of the glass manufacturing industry and are, therefore, small
in amount. The development of these promulgated standards of per-
formance was based on industry-wide considerations. Thus, the particular
concern was concentrated on particulates, S02, and NOX> that is,
pollutants common to all the glass manufacturing industry. Review of
these emissions could occur at the fourth-year regulatory review.
One commenter suggested that the EPA did not seriously consider
the alternative of rejecting New Jersey's Governor Byrne's petition
that standards of performance be established for glass manufacturing
plants. As a result of the Governor's petition, an evaluation of his
claims was made. This petition was submitted with the same intent
that an application of the Governor of a State under Section lll(g)(2)
would today be submitted. A notice of intent to develop the standards
of performance for glass melting furnaces was published in the Federal
Register (42 FR 37213) on July 20, 1977.
In the performance of the analysis, all relevant factors were
taken into consideration in determining whether to proceed with or
reject the Governor's petition that new source performance standards
2-14
-------
be developed for the glass manufacturing industry. In this case, the
Administrator has determined that glass manufacturing plants con-
tribute significantly to air pollution which may reasonably be
anticipated to endanger public health or welfare. Therefore, the
Governor's petition was evaluated, found to have merit, and resulted
in the proposal of these standards in June 1979.
2.2 EMISSION CONTROL TECHNOLOGY
One commenter suggested that the limitations imposed .by the
standards of performance invite borderline compliance status in all of
the four major categories of glass manufacturing plants. This com-
menter stated that not providing a sufficient regulatory cushion does
not follow in the intended spirit of the development of these standards.
Other conmenters questioned the ability of the glass manufacturing
industry to achieve compliance with the standards.
Upon reviewing data submitted during this rulemaking, some of the
standards have been changed to more accurately reflect the emission
control capabilities of the four categories of glass production. The
promulgated standards of performance are based on test results conducted
in accordance with EPA Method 5 and the Los Angeles Air Pollution
Control District (LAAPCD) method, as discussed in the preamble to the
proposed standards. The standards are based on emission data and
detailed engineering and cost analyses and were not developed to
invite borderline compliance, as suggested by the commenter. The
promulgated standards reflect, for each individual category of glass
manufacturing plant, the best system of continuous emission reduction,
which the EPA has determined to be adequately demonstrated taking into
consideration the costs, and nonair quality health and environmental,
and energy impacts associated with their attainment.
The container glass and pressed and blown (soda-lime) glass
standards are the only ones to remain as they were proposed. The
standards for the wool fiberglass and flat glass categories, as well
as most of standards for the pressed and blown glass sub-categories,
were changed. In addition, pressed and blown glass is now separated
into three sub-categories: borosilicate; soda-lime and lead; and,
2-15
-------
other-than borosilicate, soda-lime, and lead. Refer to the discussions
in the specific subsections in this section for an explanation of the
changes to these standards.
Several commenters were concerned that technology transfer had
been used in setting some of the standards. Specifically referred to
were the wool fiberglass and flat glass standards.
The promulgated standards have been based on verified test results.
These tests were conducted on glass melting furnaces for each category
and sub-category of glass formulation that are covered by these standards.
However, this is not to say that technology transfer is not a valid
means for determining an available control technology appropriate for
the development of new source performance standards. The Clean Air
Act does not require that the best system of continuous emission
reduction for a particular source category must have been actually
applied to sources in that category but rather that the system should
be adequately demonstrated.
The decision to regulate the glass manufacturing industry as four
categories of production was made based on technological information
received and collected prior to and subsequent to proposal as well as
regulatory simplification, as mandated by Executive Order 12044. In
assessing the entire glass manufacturing industry it was determined
that the source to be regulated, the glass melting furnace, varied
technologically in principally four areas of production (container
glass, pressed and blown glass, fiberglass, and flat glass). In the
process of determining the major categories of glass production it was
found that the pressed and blown glass category had, within itself,
areas of production that were individually unique as to their potential
for particulate emission control. As a result, the pressed and blown
category was divided into three subcategories: borosilicate, soda-lime
and lead, and other-than borosilicate, soda-lime, and lead.
It was not practically possible to test glass manufacturing
plants melting all types of batch formulations. The Standard Industrial
Classification Manual lists in excess of 80 final glass products.
Each of these glass products is liable to have several glass formulations
depending upon the final use of the product, the color of the final
product, or the manufacturer of the product.
2-16
-------
Despite the numerous formulations utilized throughout the industry
it was found, through data and information received, that the four
major categories and the three subcategbries for pressed and blown
glass selected for these standards adequately represent the emission
reduction levels achievable for the melting of all glass formulations.
There is no reason to believe that any affected facility, as defined
in the regulation, will not be able to comply with the standards. The
standards set out in the regulation represent levels of control typically
achievable by manufacturers of all types of glass.
Control device design considerations are influenced by process
factors such as; glass type; production rate; furnace size; volume
flow, temperature, and moisture of the gas stream; amount, type, and
size of particulate; the resistivity of the particulate and gas stream;
and the presence of other pollutants. These process factors have been
taken into account by dividing the industry into four major categories
with certain subcategories. Each category and subcategory represents
a set of process factors that typify that category or subcategory.
Thus, by segmenting the glass manufacturing industry into categories
and subcategories, where data showed this necessary, process factors
that influence control device de'sign considerations have been considered.
Also, by segmenting the glass manufacturing industry into these categories
and subcategories, the influence of the process factors has been
minimized. Therefore, collecting pertinent information and then
establishing standards by categories has allowed the development of
achievable standards for the glass manufacturing industry. Thus,
variations in these process factors presented by the manufacture of
the 80 or so glass types do not preclude the achievability of the
standards.
Process factors, as mentioned above, are generally known before
the design of a control device. In the glass manufacturing industry,
these factors have been characterized by EPA, by individuals and by
industry personnel. During the design of a control device used to
bring a glass melting furnace into compliance with the standards, an
2-17
-------
air pollution control engineer would review available information and
appropriately size the control device. Design procedures, while
generally the same from design to design, require different information
for each design. Information on the volume flow, temperature, and
moisture of the gas stream can be readily determined from design
considerations for the glass melting furnace. In addition, the amount
of particulate, the type of particulate, and the resistivity of the
particulate and gas stream can be determined from glass melting furnace
design considerations, including the glass type, the size of the
furnace, and the production rate. The size of particulates from glass
melting furances has been characterized. Such factors were considered
in establishing the promulgated standards of performance.
A commenter suggested that EPA, in promulgating the standards as
proposed, will not allow industry to choose its method of compliance
from a wide range of methods available to it. The proposed standards
of performance were based on the criteria set forth in.Section 111 of
the Clean Air Act for the best available continuous method of emission
reduction that has been adequately demonstrated. The promulgated
standards are based on the emission limitations that are achievable
and are not meant to exclude any one method of control. Many forms of
control have been investigated in the development of these standards.
However, not all forms of control are capable of achieving the degree
of control necessary to comply with the standards; this conclusion is
based on presently available information. For example, scrubbers have
been investigated in the development of these standards and in virtually
all cases were found not to achieve the new standards. This does not
mean that scrubbers cannot be designed to effectively control glass
plant particulate emissions and to achieve compliance with the standards.
Means of emission reduction, not presently able to meet the standards,
could possibly be designed to meet them at a later date.
During the public comment period, comments were received concerning
the use of process modifications as a method of reducing particulate
emissions from the glass melting furnace. Many of the comments indi-
cated that during the.development of the Background Information Document,
2-18
-------
(BID) Volume I, EPA did not perform a thorough investigation into the
use of process modifications as a continuous emission reduction technique.
These commenters also stated that process modifications are effective
methods of emission control, and, therefore should be considered as
alternatives to the add-on control devices as the basis for these
standards of performance.
In light of these comments and the need to resolve this issue, a
re-examination into the use of process modifications was performed.
Before attempting to address these comments, a review of the material
used in the development of the section on process modification in the
BID, as well as a review of the BID itself, was performed. Throughout
this review each industry segment was dealt with separately.
Background Material. A review of the docket material as it
relates to the use of process modifications in the flat glass manu-
facturing industry reveals that the types of process changes being
employed by this segment of the industry to control air emissions are
primarily designed to reduce both the entrainment of dust in the
combustion gases and the volatilization of the melt. In Source Assessment:
Flat Glass Manufacturing Plants, EPA 600/2-76-0326 (docket entry
OAQPS 77/1-II-A-3), the elimination of less than 44 micrometer (minus
325-mesh) particles in the feed material and the addition of water to
the glass batch are two techniques identified as process modifications
that will minimize dust entrainment. The flat glass source assessment
report further states that the volatilization of the melt can be
reduced by controlling the feed material, by designing the furnace
properly, by lowering the furnace temperature with electric melting,
and by reducing salt cake (sodium sulfate)..
Salt cake or another sulfate is a necessary flux that prevents
scum formation in the melting furnace and aids in the melting process.
Manufacturers reduce salt cake by eliminating excess sodium sulfate
through glass formulation changes. However, exact details are con-
sidered proprietary. The flat glass source assessment report states
that by improving the overall furnace efficiency with techniques, such
as:
(1) improving the refractories for corrosion resistance and
better insulation;
2-19
-------
(2) increasing the checker volume for better heat recovery;
(3) controlling combustion to produce large luminous flames that
eliminate hot spots in the furnace and provide better heat transfer to
the melt and applied instrumentation to regulate air/fuel mixtures;
(4) monitor furnace temperature and stack gas composition; and
(5) automaticeMy charge the batch into the furnace and reverse
the air flow through the regenerative checkers;
emissions from fluxing agents will be lowered. As a result of these
process changes, the flat glass source assessment report states that
fuel consumption rates in the flat glass manufacturing industry have
decreased significantly.
In addition, the flat glass source assessment report identified
furnace temperature as having a profound influence on the particulate
emission rate. The report referred to two studies which indicated
that emission rates increased exponentially with temperature. The
furnace temperature can be lowered by improving furnace efficiencies,
by decreasing the production rate, and by utilizing supplemental
electric heating.
Although the flat glass source assessment report identified
electric boosting as a means of lowering furnace temperature, thereby
reducing emissions, it also stated that due to the rather large production
rates typical of that segment of the industry, electric boosting has
been found to be very difficult to employ in flat glass production.
The difficulty of utilizing electric boosting in the flat glass manu-
facturing industry was further illustrated in a study supplied by PPG
Industries, Incorporated (PPG) [docket entry OAQPS 77/1-II-I-67]. It
is PPG's belief that electric boosting is not yet proven technologically
on the scale required by PPG; but on a trial basis, (utilizing small-sized
furnaces) electric boosting has reduced emissions. PPG's study also
provided information on experimentation with batch pelletizing and
salt cake reduction in which a 30 percent reduction in particulate
emissions was reported. PPG also indicated that further work must be
performed in these areas before they can be used on a large scale.
A review of the docket material with respect to the container
glass industry reveals that process modifications are used extensively
2-20
-------
by this segment of the industry. Glass formulation changes have
reduced the use of materials, such as sulfates,, fluorides, and
selenium, and the use of arsenic has virtually been eliminated (docket
entry OAQPS 77/1-11-1-41).
The container glass industry has performed extensive work in
modifying furnace design to increase fuel efficiency which can in turn
lead to a decrease in combustion products, a decrease in dust entrainment
by hot combustion gases over the melting glass batch, and a possible
decrease in furnace temperature (docket entry OAQPS 77/1-II-A-5). The
methods currently in practice to improve furnace efficiency are virtually
the same as the techniques used by the flat glass manufacturing industry,
except that electric melting and electric boosting have been used to a
much greater extent. Innovative techniques, such as batch preheating
and agglomeration, have been researched by a joint EPA and Department
of Energy (DOE) effort but are still considered to be in the developmental
stages (docket entry OAQPS 77/1-II-B-248).
The use of all-electric melting and electric boosting in the
container glass industry has been wide-spread. Approximately half of
the container glass manufacturers in the United States have electric
boosters; and at least one hundred all-electric furnaces, ranging in
size from 4 to 140 tons/day, are in operation throughout the world
(docket entry OAQPS 77/1-II-1-40).
The docket contains several examples of the type of emission
reductions which can be attained by employing electric boosting. In a
letter from Mr. K.B. Tanner Jr. to Mr. D.R. Goodwin (EPA), dated
October 12, 1977 (docket entry OAQPS 77/1-II-D-179), Mr. Tanner presented
emission data for one of Brockway Glass Company's plants in Pomona,
California, in which particulate emissions were reduced through the
use of electric boosting to a level of 0.68 Ib/ton. Mr. H.R. Carroll
of the Glass Container Corporation in a letter to Mr. Herring (EPA)
dated October 14, 1977 (docket entry OAQPS 77/1-1I-D-183), presented
test data for the Knox, Pennsylvania, plant which indicated a particulate
emission level of 1.2 Ibs/ton. In still another letter from Mr. I.E.
Cruser of the Ball Corporation, to Mr. Ronald Boone (State of North
Carolina), dated January 20, 1977 (docket entry OAQPS 77/1-II-D-117),
2-21
-------
test data were presented which indicated a participate emission level
of 1.5 Ibs/ton.
Also contained in the docket is a letter from Mr. Robert Drake of
the Glass Packaging Institute to Mr. D.R. Goodwin (EPA), dated
September 19, 1978 (docket entry OAQPS 77/1-1I-D-248). Mr. Drake
claimed that new melting furnaces that would be subject to the New
Source Performance Standards could generally achieve a particulate
emission limitation of 0.8 Ib/ton of glass produced by employing
available process modifications, including electric boosting.
The docket contains a letter from Mr. John F. Blumenfeld, of the
Hartford Division of Emhart Industries, Inc. to Ms. Margaret A. Timothy,
(JACA, Inc.) dated May 19, 1978 (docket entry OAQPS 77/1-H-I-65,
exhibit 3), in which a different view of electric boosting is presented.
In this letter, Mr. Blumenfeld states that electric boosting shortens
the life-span of a furnace, since it results in higher temperature
molten glass and usually is used to increase the pull .rate. Both of
these factors increase the furnace's rate of erosion. Mr. Blumenfeld
is also of the opinion that fuel firing rates in a boosted furnace are
generally higher than in a non-boosted furnace. However, the letter
continued by stating that although the tons per furnace life are
greater with an electric boosted furnace than for the non-boosted
furnace (25 percent increase in factory output), this increase is
often off-set by the cost of electricity.
In attempting to assess the benefits of all-electric melting in
the container glass industry, the docket contains an article entitled,
"Energy Use and Air Pollution Control in New Process Technology"
(docket entry OAQPS 77/1-1I-1-40). The article states that the cold
top electric melter has many benefits. Since there is no fossil fuel
being fired, S02 emissions are reduced significantly. In addition,
NO is not created because there is no combustion taking place in the
air's atmosphere above the melt. The only air emissions are from the
decomposition of carbonates, sulfates, nitrates, etc. in the glass
batch. There is also no dusting due to the entrainment of batch
ingredients as occurs when the high velocity flames of fossil
fuel-fired melting tanks are in use. The exhaust is almost entirely
2-22
-------
C02 plus a quantity of SOp. The all-electric melter also provides no
water effluent streams.
The benefits of electric melting are also contained in an article
entitled, "Practical Data for Electric Melting" by Robert E. Loesels
(docket entry OAQPS 77/1-II-I-23). The article states that electric
melting utilizes 65 to 88 percent of the direct heating energy. It
also states that the cost of an electric furnace is less due to the
fact that there is no need for regenerator chambers, port racks,
checkers, flues, reversing valves, and, in most cases, stacks are
eliminated.
The problems associated with electric melting in the container
glass industry are illustrated in an article entitled, "Pollution
Control" by Roy S. Arrandale (docket entry OAQPS 77/1-II-I-16, December
1974). Mr. Arrandale is of the opinion that electric melting shortens
the life-span of the furnace, due to its severe attack upon the
refractories. Mr. Arrandale continues in his article by stating that
the electric furnaces now being built are generally of the 100 tons/day
or less production capacity. It is Mr. Arrandale's opinion that these
furnaces are too small to adequately supply molten glass to the container
forming process and, therefore, are uneconomical production-wise for
the container glass industry. Also, one of the major problems associated
with electric melting is that the furnaces are restricted to glasses
having suitable electrical conductivity characteristics and mild
chemical attack on molybdenum or graphite electrodes at molten glass
temperature. Contrary to Mr. Loesel's statements, Mr. Arrandale is of
the opinion that the cost of an electric furnace will be very high,
due to the fact that electric power is becoming shorter in supply and
the energy cost for the furnace is doubled. Mr. Arrandale's article
also stated that the electric melter is not thermodynamically efficient.
Also contained in the docket is an engineering study program
entitled, "Glass Furnace Emissions Abatement: Particulate Control
Through Process Modification," prepared for the Glass Container
Manufacturers' Institute (GCMI) by TRW Systems Group (docket entry
OAQPS 77/1-II-I-9A). This is a parametric analysis of a number of
variables that were performed utilizing a generalized computer program.
2-23
-------
The process variables that were evaluated were melt temperature and
stack gas temperature. The chemistry variables evaluated were salt
cake input concentration, water concentration in the gas-melt
equilibrium zone, steam injection into process effluent, calcium
carbonate injection into the process effluent, and the utilization of
fuel oil with varying sulfur contents.
The study concluded that the most effective process modifications,
identified as resulting in 60 percent to 70 percent particulate emission
reductions, are the reduction of the glass melting temperature to
below 2400°F and the use of a dry air curtain between the high water
vapor concentration flame combustion zone and the melt surface. The
study also added that neither of these process modifications may be
economically attractive because of reduced furnace capacity. In a
review of this report, GCMI agreed that the reduction of glass
temperature to below 2400°F is not practical. As for the method of
using a dry air curtain between the high water vapor concentration
flame combustion zone and the melt surface, GCMI was of the opinion
that further study is needed.
A review of the docket material as it pertains to process
modifications in the pressed and blown glass industry reveals that the
modifications of the feed material and the furnace design are virtually
identical to the types of modifications employed in the container
glass industry. The only real difference between the two categories
is that electric melting is utilized to a much greater extent in the
manufacture of pressed and blown glassware than in container glass
production (docket entry OAQPS 77/1-II-A-7). The pressed and blown
glass manufacturing industry is better adapted for the use of all-
electric melting due to the small-sized furnaces employed by this
segment of the industry.
In a position paper entitled, "Proposed Regulation Change for
Pressed, Blown, or Spun Soda-Lime Glass Melting Furnaces," submitted
by the Carr-Lowrey Glass Company (docket entry OAQPS 77/1-II-I-20),
the point is raised that due to the research that has been done over
the past few years, such as with the removal of volatile materials-
from the batch (i.e., sulfur and fluorides), particulate emissions
2-24
-------
have been reduced by as much as 50 percent. The paper further states
that their Anchor-Hocking plant has not been able to develop a new
batch formulation which would provide the necessary glass quality for
Carr-Lowrey production requirements. The major contributing factor to
this problem is the fact that the speciality glass requires very close
control of the glass constituency in order to achieve the intricate
forming that is characteristic of much of the production. Removal of
the volatile components beyond a limited amount or other batch modifi-
cations would be a detriment to the forming characteristics. Additional
batch ingredients would have to be added to the formulation, resulting
in no net emission reductions. This position paper concluded that it
would not be possible to make the significant process changes, such as
decreasing the melter temperature, which would have a positive effect
on emission reduction. Electric boosting in this industry does have a
positive effect upon emission reduction, but the related costs are
very high.
In a letter from Mr. J.T. Harrsen (G.E. Co., Inc.) to Ms. Margaret A.
Timothy (JACA, Inc.), dated May 23, 1978 (docket entry OAQPS 77/1-II-I-65,
exhibit 4), Mr. Harrsen was of the opinion that electric boosting in
the pressed and blown glass industry has not been in use long enough
to make a determination as to its effects on furnace life. In addition,
Mr. Harrsen was of the opinion that all-electric melting is still
considered to be in the experimental stages.
The topic of all-electric melting in the pressed and blown glass
industry was also mentioned in an EPA trip report (docket entry
OAQPS 77/1-II-E-21). Corning Glass representatives characterized cold
crown vertical melting as follows:
"no emissions; very expensive, due to high electric power costs;
cannot melt all types of glass; unforgivingtank control is
critical; can be used for fluoride, opal and Pyrex seal beam
headlights."
A review of the docket material as it pertains to process
modifications in the wool fiberglass industry indicates that the
electric furnace is used by individual fiberglass manufacturers and
relatively low emission levels are reported. The future installation
2-25
-------
of electric furnaces in new fiberglass plants is uncertain due to the
unclear energy situation production limitations, and the problem of
the space requirement in converting from a liquid or gaseous fuel-firing
furnace to an all-electric furnace (docket entry OAQPS 77/1-II-A-l).
Background Information Document. An examination of the section
on process modifications in the BID reveals that many of the process
changes discussed earlier are also contained in this section. The BID
states that reducing the amounts of certain'materials in the feed,
increasing the use of cullet, installing sensing and controlling
equipment, modifying the burner design and firing pattern, and utilizing
electric boosting are examples of the types of process modifications
employed by the glass manufacturing industry.
The BID also points out the benefits of some of the process
changes. Certain process changes have caused the elimination of
arsenic from the feed material in the container glass industry. In
addition, the amounts of soda, fluorides, and selenium .fed to the
furnace have been minimized. Many process modifications offer the
double benefit of lowering pollutant emission rates as well as lowering
fossil fuel consumption rates. The BID further states that emission
tests were not available to document the lowering of particulate
emissions by the use of process modifications.
The BID also contains a discussion on electric boosting, in which
it indicates that this technique decreases the required bridge wall
temperature, decreasing the fuel consumption rate, which therefore
decreases the particulate'and gaseous pollutant levels. This discus-
sion was also accompanied by references to indicate the type of emission
reductions which have been associated with electric boosting. In a
Glass Packaging Institute "Issue Paper," (docket entry OAQPS 77/1-II-D-ll)
it was reported that electric boosting has reduced particulate emis-
sions to a level of between 0.68 to 1.76 Ibs/ton, a range which is
described as a rough estimate, due to the fact that some of these
emission tests were not performed according to the specified EPA
Method 5 procedures. Another reference is to a gas-fired/electric
boost container glass-furnace in which the particulate emission per
kilogram of glass produced dropped 55 percent, along with the
2-26
-------
consumption of less energy (docket entry OAQPS 77/1-II-I-20). .The BID
concluded the discussion by stating that, in. general, the levels of
particulate emissions from glass melting furnaces using process
modifications are indistinguishable from the uncontrolled emission
levels.
The BID also contained a section on the use of all-electric
melting in the glass manufacturing industry. A review of this section
reveals that because the surface of the melter in a cold top electric
furnace is maintained at ambient temperature and fresh raw batch
materials are fed continuously over the entire surface, the emissions
are greatly reduced. The gases discharged largely consist of carbon
dioxide and water vapor. Construction costs are generally less since
there are no regenerator chambers, port necks, checkers, flues, or
reversing valves, and in most cases stacks are eliminated.
Because many of these melters do not have stacks, the level of
emission control cannot be soundly documented. However, from the
nature of the melting process, potential emissions can be deduced and
relative amounts of emissions can be estimated. In general, all-electric
melters have particulate emission levels of approximately 0.2 Ib/ton
in the production of soda-lime and borosilicate glasses. In addition,
this can be accomplished with no changes in the solid waste or water
pollution impacts.
All-electric melting is a relatively new technology in the glass
manufacturing industry. Therefore, there are several limitations as
to the application of this technique throughout the industry. Not all
glasses possess the electrical properties required for successful
all-electric melting. Additionally, the all-electric technology is
not far enough advanced to satisfactorily produce glass in large
quantities.
After reviewing the section on process modifications in the BID,
along with the material used to develop that section, several conclu-
sions can be reached. An attempt was made by EPA to formulate a
comprehensive information base with respect to the use of process
modifications in the glass manufacturing industry. Information was
gathered from a wide variety of sources, including technical journals,
2-27
-------
private contractors' reports, and industry comments. In addition,
industry was requested to supply any additional information which
could possibly aid in the evaluation of this method of emission control
(i.e., Section 114 letters).
The information that was compiled enabled EPA to present a rather
accurate account of the types of process modifications employed by
industry in controlling particulate emissions. The material also gave
an indication as to the possible benefits, as well as the potential
problems, that could occur when employing certain process changes. In
addition, the material provided an indication as to the possible
levels of particulate emissions that could be attained by certain
segments of the industry when employing certain process changes.
However, this study was by no means complete. Many issues concerning
the use of process modifications were left unresolved. There was a
general lack of evidence substantiating the efficiency of many of the
process changes employed by industry to control emissions. This
problem was compounded by the fact that several glass manufacturers
considered the exact details of their process modifications as
proprietary information.
Be that as it may, the lack of information does not indicate that
EPA performed an inadequate investigation into the use of process
modifications, nor does it indicate that the contents are in any way
misleading or incorrect. What this does indicate is that a consider-
able amount of uncertainty exists with regard to the ability of process
modifications to control particulate emissions in the glass manufacturing
industry. This point is further illustrated when reviewing the docket
material.
The docket material clearly indicates a diversity of opinion as
to the performance capabilities associated with the use of process
modifications as a method of continuous emission abatement. A process
modification technique considered beneficial by one individual was
also criticized as being a possible hazard by another individual. The
docket also contains reports by industry representatives indicating
that many forms of proces's modifications are still considered to be in
their experimental stages. There are also reports which indicate that
2-28
-------
certain process changes can only be employed in a particular segment
of the industry or at a particular production rate.
In view of the fact that, at the time of the study on the use of
process modifications in the glass manufacturing industry, a consider-
able amount of uncertainty existed as to the use and the capabilities
of this method as a continuous means of emission abatement, it can be
concluded that the decision to base the NSPS on add-on control devices
of demonstrated effectiveness was justified.
Comments Received After Proposal. The remainder of this discussion
deals specifically with the comments on the use of process modifications
in the glass manufacturing industry that were received after the
proposal of the standards. While reviewing those comments, it became
apparent that many of the commenters raised issues which have already
been discussed in the Background Material and Background Information
Document sections of this document.
The docket contains five comments concerning the use of process
modifications in the flat glass industry. In a letter from Mr. J.T.
Destefano (PPG, Inc.) to Mr. D.R. Goodwin (EPA) dated August 10, 1979
(docket entry OAQPS 77/1-IV-D-7), Mr. Destefano stated that PPG believes
that process modifications are a preferable form of emission control
and reduction for a flat glass furnace, rather than add-on controls.
It is Mr. Destefano!s opinion that certain process modifications are
the best available control technology because they have been demon-
strated, are a more reasonable means of achieving the goals of emission
reduction, control all emissions rather than merely particulates, and
are more reliable. However, Mr. Destefano was of the opinion that the
ultimate level of control is uncertain at the present time. Mr. Destefano
also attached eight test results for three of the PPG Industries, Inc.
plants that have attempted process modifications. The particulate
emissions ranged from 2.01 Ibs/ton to 1.08 Ibs/ton, with an average of
approximately 1.41 Ibs/ton. Mr. Destefano explained that three of the
furnaces at these three plants employed salt cake reduction techniques,
while the remaining furnaces used other raw material modifications.
It has been PPG's position throughout the development of these standards
that the exact details concerning several raw material modifications
2-29
-------
are considered proprietary (refer to Summary of Meeting Held Between
EPA and PP6 Industries Inc., dated August 21, 1979, docket entry
OAQPS 77/1-IV-E-3).
Also included in the docket is a letter from Mr. V.S. Sussman
of the Ford Motor Company, dated August 13, 1979 (docket entry
OAQPS 77/1-IV-D-12). In this letter Mr. Sussman states that batch
modifications (i.e., pelletized batch feed system and raw materials
alteration) have decreased particulate emissions at their facilities
by 59 percent.on the average (assuming an uncontrolled emission rate
of 3 Ibs/ton with no energy penalties, no adverse environmental side
effects, and with a relatively minimal cost). This emission reduction
claim was also accompanied by test results from 22 furnaces, with only
5 of the 22 test results accompanied by emission data (docket entry
OAQPS 77/1-1V-D-12B). This data does indicate that particulate emission
levels are moving downward from 1.73 Ibs/ton in 1974 to 1.05 Ibs/ton
in 1977. This translates to a particulate emission reduction of
approximately 39 percent (15 percent assuming an average emission rate
of 2.0 Ibs/ton). Mr. Sussman also stated that an additional benefit
of process modifications has been the reduction of SOg emissions by as
much as 70 percent. The letter concludes by suggesting that process
modifications should be a control option to achieve compliance with
the standard set at 1.1 Ibs/ton.
Mr. Frank Partee of the Ford Motor Company stated during the
public hearing on the proposed standards of performance for new glass
melting furnaces under Section 111 (docket entry OAQPS 77/1-IV-F-l)
that a meaningful Option II should be established at a level which
could be achieved by process changes, such as batch formula modifi-
cations. Mr. Partee is also of the opinion that a reasonable standard
for flat glass production would be 1.1 Ibs/ton.
Mr. Werner Sanz of the Libbey-Owens Ford Company in a letter to
the Central Docket Section, dated September 5, 1979 (docket entry
OAQPS 77/1-IV-D-15), stated that at their Lathrop, California, facility
process modifications have reduced particulate emissions from
1.27 Ibs/ton to 0.65 Ib/ton. This translates to a 49 percent reduction
in particulate emissions. Mr. Ganz also stated that process modifications
2-30
-------
reduce not only particulate matter emissions but other pollutant
emissions as well. In addition, he pointed out that process modifi-
cations can eliminate any adverse environmental effects because disposal
of residuals is reduced.
Again, the comments concerning the use of process modifications
in the container glass industry were very similar to the type of
comments presented earlier in the Background Material and Background
Information Document sections of this document. The only comment that
supplied additional information was a report prepared by Mr. Roger
Strelow for the Glass Packaging Institute, dated September 14, 1979
(docket entry OAQPS 77/1-1V-D-19). Mr. Strelow presented stack test
results from 10 container glass furnaces that have employed electric
boosting. The results indicate an average particulate emission level
of 0.44 Ib/ton. It must be pointed out that although the standard may
at times be achievable with the use of process modifications that
achievability is not assured.
Also accompanying Mr. Strelow's comment was a report entitled,
"Optimizing Operating Conditions to Reduce Stack Emissions From a
Glass Container Furnace," by K.B. Tanner, Jr. (1975) In the report
Mr. Tanner stated that increasing the production rate, and therefore
the temperature, increases the particulate emission rate. If the
bridge wall operating temperature is decreased 100°F, the result will
be a 50 percent decrease in the emission rate. The report further
indicated that a reduction of SO,, emissions can be accomplished by
careful furnace operation and selection of batch composition. Addi-
tionally, the formation of oxides of nitrogen, it was claimed, can be
controlled by the flame temperature and availability of oxygen in the
combustion zone. Mr. Tanner also reports that the use of carbon as a
minor ingredient in a flint glass batch has reduced particulate emissions
by a factor of two.
In addition to this report, Mr. Strelow's comments were accompanied
by another report entitled, "Control of Fine Particulates From Continuous
Melting Regenerative Container Glass Furnaces," by H. Simon and J.E.
Williamson (1975). The report indicated that the future looks brighter
for all-electric furnaces (which are usually too small and electric
2-31
-------
energy rates are too expensive to seriously compete with the larger
fuel-fired furnace) in the event of gas use curtailment. Also, rate
increases narrow the advantage now enjoyed by gas-fired equipment.
The report further indicated that electric boosting is a proven aid in
reducing air contaminant emissions from fuel-fired furnaces. Accom-
panying this statement were six test results of furnaces employing
electric boosting as an emission control technique. The results
indicated an average emission level of 1.47 Ibs/ton, with a low of
1.25 Ibs/ton (1972). The report also stated that with the use of
minor operational variations in air-fuel ratio, batch moisture, and
pull rates the average particulate emission level for five furnaces
was 0.65 Ib/ton. In addition to these results, the report stated that
four tests were conducted on furnaces using operational variations and
electric boosting. The test results indicated an average particulate
emission level of 0.58 Ib/ton. The report also contained test results
from five all-electric melters which indicated a particulate emission
level of 0.23 Ib/ton. The report concluded with the statement that
electric boosting and modification of the operating variables may
allow some fuel-fired furnaces to operate with very low emission
rates. It must be pointed out again that although the standard may at
times be achievable with the use of process modifications that
achievability is not assured.
Mr. Strelow's comments were also accompanied by another report
entitled, "Use of Electric Boost to Reduce Glass Furnace Emissions,"
by R.J. Ryder. The purpose of Mr. Ryder's report was to describe
methods that have been developed which permit acceptably accurate
estimations to be made of emissions of particulates and NO from glass
/\
melting operations when electric boosting is used to provide temperature
and emission reduction. The report, supplemented with test results,
indicated that the use of electric boosting coupled with a decrease in
above-the-melt temperatures were quite successful in reducing the
emissions of particulates to approximately 0.58 Ib/ton. In addition,
emissions of NO were greatly reduced.
X
There were three comments received concerning the use of process
modifications in the pressed and blown glass industry. In a letter
2-32
-------
from Mr. 6.H. Mosely of Corning Glass Works, Inc. to Mr. D.R. Goodwin
(EPA), dated December 15, 1978 (docket entry OAQPS 77/1-IV-D-7A,
reference 7), Mr. Mosely states that process modifications are much
more effective than add-on control devices in terms of energy utili-
zation, initial capital cost, and operating and maintenance costs. In
addition, Mr. Mosely believes that process modifications offer the
benefit of eliminating any potentially hazardous solid waste. Mr. Mosely
concluded by stating that process modifications should be considered
the best demonstrated control technology for borosilicate-type glass.
Mr. Mosely also attached test results from furnaces using a combination
of emission reduction techniques such as batch composition, tank
operation, and electric boosting. The results indicated that for
borosilicate type glass, particulate emissions were approximately
1.5 Ibs/ton. The particulate emission level for lead-type glass
averaged approximately 1.8 Ibs/ton and the particulate emission level
for "non-soda-lime"-type glass averaged approximately 1.7 Ibs/ton.
Also contained in this letter were estimates of the cost of particulate
removal for the process changes being employed by Corning Glass Works.
The results indicated that for borosilicate-type glass, an abatement
cost of approximately $1.13/lb of particulate removed can be accomplished.
In addition, lead-type glass would attain an abatement cost of $.39/lb
of particulate removed. The letter concluded with the statement that
the pressed and blown glass industry should be allowed to use the
option of process modifications as a means of meeting the limits set
for new sources.
Another comment concerning the use of process modifications in
the pressed and blown glass industry was included in a letter from
Mr. C.M. LeCroy of PPG Industries, Inc. to the Central Docket Section
(docket entry OAQPS 77/1-IV-D-8). Mr. LeCroy was of the opinion that
process modifications currently under study by PPG, could possibly
allow PPG to operate a 100 tons/day furnace and meet the New Jersey
standards without the use of add-on control equipment.
The post public hearing docket contains one comment concerning
the use of process modifications in the wool fiberglass industry. In
a report from Mr. E.D. Switala of Owens-Corning Fiberglas Corporation
2-33
-------
to the Central Docket Section, dated September 12, 1979 (docket entry
OAQPS 77/1-IV-D-17), Mr. Switala stated that process modifications
offer a variety of advantages over add-on controls, such as lower
capital costs, lower fuel oil and natural gas consumption, reduction
of furnace emissions, and reduction of equipment malfunction downtime.
Mr. Switala concluded his comments by stating that process modifica-
tions can be used to meet a particulate emission level of 0.8 Ib/ton.
Summary. After reviewing the section in the BID on process
modifications, along with the materials used to develop that section,
it is apparent that the use of process modifications in the glass
manufacturing industry was taken into consideration during the development
of these standards of performance. The types of process change employed
by industry, along with the possible benefits and potential problems
associated with these techniques, were presented accurately. However,
it is clearly evident that many issues concerning these methods were
left unresolved.
These issues were left unresolved because the information on this
area of emission control was minimal and often lacked substantiating
evidence. The information on emission reduction indicated that emission
reduction by process modifications is uncertain with respect to the
effectiveness of the technique. This uncertainty led to the decision
to base the NSPS on an add-on control device of known and proven
effectiveness.
Since proposal of the standards of performance for the glass
manufacturing industry, additional information has been made available
concerning the use of process modifications. 'This information has
indicated that progress is being made by several glass manufacturers
in reducing emissions by the use of certain process modification
techniques. However, several problems arise in attempting to evaluate
this progress. A major problem which has been experienced throughout
the compilation of these standards is that there has been a general
lack of quantifiable emission data accompanying the reported emission
levels. This limitation has been primarily due to industry's insis-
tence that the exact details of many of the process modifications be
considered highly confidential.
2-34
-------
An additionally important factor in considering the use of these
process modification techniques is that from the information submitted,
it appears that a variety of achievable emission levels exist. Only
in certain instances has the data indicated that the particulate emis-
sion reductions attributable to process modifications approach the
levels required by the promulgated standards. The majority of the
data indicates particulate emission levels slightly lower than, or in
some cases higher than, the uncontrolled emission rates.
It appears from the data submitted that the most effective form
of process modifications is the all-electric melter. However,
all-electric melting has several limitations. At present, the
all-electric melter is restricted in most cases to a furnace capacity
of less than 100 tons/day and glass types with specific electrical
characteristics. In addition, all-electric melting has reportedly
substantially reduced the campaign life of these furnaces. The second
most effective form of process modification technique appears to be
electric boosting. But, it appears that there are also several limitations
associated with this type of process change. As indicated from the
information submitted, there is no clear indication as to the levels
of emission reduction potentially achievable with electric boosting or
if these levels can be reached on a continuous basis. Another limitation
associated with electric boosting is that electric boost is also
restricted to melting a glass type with specific electrical charac-
teristics. Also, it has reportedly substantially reduced the campaign
life of these furnaces. It is for these reasons that the EPA considers
all-electric melting and electric boosting to be inadequately demonstrated
means of continuous emission reduction for the industry. Other process
changes such as furnace temperature reduction, limitations on production
rates, and glass formulation changes are not considered to be appropriate
because they are simply not practical for the industry; the techniques
are not well defined; the techniques presently only represent a
non-continuous means of emission reduction.
It is quite evident from the responses to the request by EPA to
supply any information concerning the use of process modifications
2-35
-------
that only major manufacturers have the research and developmental
resources to attain the lower emission levels through the use of
process modifications. It must be pointed out that only the manufac-
turers who have performed extensive experimentation in this area of
emission control have indicated the desire to base the NSPS on the use
of process modifications; yet these manufacturers have also indicated
that it would be impossible to disclose the exact details of many of
the techniques to other manufacturers for competitive reasons.
The re-examination into the use of process modifications in the
glass manufacturing industry has led to the conclusion by EPA that
process modifications are still in the research and development stages;
the achievable levels of emission reduction are not well defined; the
emission reductions may not be continuous; the reported emission
reductions are based on proprietary information, and; only the large
manufacturers have the ability to reduce emissions with process
modifications. Therefore, it has been decided to base promulgated
standards on an add-on control device of known and proven effectiveness,
It should be pointed out that Section lll(j) of the Clean Air Act
provides a means by which an industry source subject to new source
performance standards can request the EPA for one or more waivers from
the requirements of Section 111 with respect to any pollutant to
encourage the use of an innovative technological system of continuous
emission reduction. The purpose of this Section of the Act is to
allow and encourage industry to develop new means of control, such as
process modifications, subject to certain restrictions.
If the source can adequately show that:
(1) the proposed system or systems have not been adequately
demonstrated;
(2) the proposed system or systems-will operate effectively and
there is substantial likelihood that such system or systems will
achieve greater continuous emission reduction than that required to be
achieved under the standards of performance which would otherwise
apply, or achieve at least an equivalent reduction at lower cost in
terms" of energy, economic, or nonair quality environmental impact;
2-36
-------
(3) the owner or operator of the proposed source has demonstrated
to the satisfaction of the EPA that the proposed system will not cause
or contribute to an unreasonable risk to public health, welfare, or
safety in its operation, function, or malfunction; and
(4) the number of waivers granted with respect to a proposed
technological system of continuous emission reduction does not exceed
such number as the EPA finds necessary to ascertain whether or not
such system will achieve the conditions specified in (2) and (3), the
EPA with the consent of the Governor of the State in which the source
is to be located, and after notice and opportunity for a public hearing
may grant a waiver from the requirement from Section 111. There are
additional factors and limitations that the EPA is required to consider
in making this determination, and they are found in Section lll(j) of
the Clean Air Act.
Until such time that process modifications can be shown to be an
effective means of continuous emission reduction able to achieve the
limitations imposed by these standards, industry has at its disposal
on an individual basis, and subject to the terms of Section lll(j), a
means for developing and perfecting these methods of control.
Another commenter pointed out that there was no provision in the
standards requiring the installation of standby equipment in case of
failure. The -standards require that the method of control utilized by
the affected facility be able to achieve compliance with the standards
continuously.- This would include the maintenance of such method in
accordance with best engineering practices. Specifically addressing
this point is an EPA regulation [40 CFR 60.11(d)] which reads, in
part, as follows:
At all times, including periods of startup, shutdown, and malfunction,
owners and operators shall, to the extent practicable maintain
and operate any affected facility including associated air pollution
control equipment, in a manner consistent with good air pollution
control practice for minimizing emissions. (Emphasis added.)
The cost of such maintenance was taken into account in performing the
economic analyis for these standards.
Other commenters suggested that a linearly related production
rate mass particulate standard is unfair to those furnaces operating
2-37
-------
at low production rates due to such things as non-production incidents
and holidays. It was claimed that plants operating at such lower
production rates would be required to meet inordinately stringent
limitations.
Commenters suggested that a linearly related production rate mass
standard is unfair to those furnaces operating at low production rates
due to such things as non-production incidents and holidays. A related
comment raised by several commenters suggested that the proposed
standards would prove to be unfair to those furnaces operating at
other than "normal" levels of production. Specifically of concern to
these commenters was the inability of glass furnaces to achieve a zero
emission rate at times when the production rate approaches zero. It
was emphasized by the commenters that even when the production rate of
a glass melting furnace is zero there would be associated emissions
due to the maintenance of the molten glass at the proper temperature.
In an attempt to resolve this issue it was suggested by a commenter
that a lowest level emission limit be set at either 227 g/hr or 454 g/hr.
This commenter explained that, based on the industry-wide estimation
that emission levels at zero production rate are roughly 20 percent of
those at normal production rates, a lowest level emission limit would
have to be incorporated in the standards in order for furnaces operating
at the lower end of their operational ranges to be able to comply with
the standards. Due to the concerns expressed by these commenters, the
method for the calculation of the furnace emission rate was changed in
order to correct for the fact that emissions are generated at zero
production rate.
Correction factors were developed after reviewing comments on
this issue. Only one commenter offered a solution to this issue.
This commenter suggested that a lowest level emission limit be set at
either 227 g/hr or 454 g/hr. In comparing these figures with the
controlled emission rates using the 20 percent figure it was determined
that a correction of 227 g/hr should be applied to the container,
pressed and blown (soda-lime and lead), and pressed and blown (other-
than borosilicate, so'da-lime, and lead) glass categories and
subcategories; and an adjustment of 454 g/hr should be applied to the.
2-38
-------
pressed and blown (borosilicate), wool fiberglass, and flat glass
categories and subcategory.
The mechanism for providing the correction factors is to subtract
this predetermined amount (g/hr) from the particulate emission rate
(g/hr) determined in the procedure using EPA's Method 5. That amount
is consequently applied to the rate of glass production (kg/hr) which
is ultimately used to determine the furnace emission rate (g/kg). By
using these correction factors, the furnace emission rate will approach
zero as the production rate approaches zero, thereby making the standards
slightly easier to achieve.
For the purposes of these standards the furnace emission rate
will be computed as follows:
Where:
(1) R is the furnace emission rate (g/kg);
(2) E is the particulate emission rate (g/hr);
(3) A is the zero production rate adjustment
A is [227g/hr for container glass, pressed and blown (soda-
lime and lead) glass, and pressed and blown (other-than
borosilicate, soda-lime, and lead) glass;
A is 454 g/hr for pressed and blown (borosilicate) glass,
wool fiberglass, and flat glass]
(4) P is the rate of glass production (kg/hr).
Although the standards will be slightly easier to achieve, the
impacts of the standards will not be substantially affected. This
correction factor should not lead to the design of control devices any
less efficient than those considered appropriately designed to achieve
the standards. This is due to the fact that as the production rate
increases from zero, the particulate emission increases and outweighs
the zero production rate correction factors. Thus, emission reduction
and cost impacts will not be substantially changed.
2-39
-------
Container Glass. One commenter referred to a hearing held in New
Jersey in 1975 before the New Jersey Department of Environmental
Protection. It was the commenter's contention that based on the
testimony presented at that hearing, manufacturers of electrostatic
precipitators (ESP) could not guarantee the achievement of the State
imposed limits. Other commenters also questioned the achievability of
the proposed container glass standard of performance for container
glass manufacturing plants.
The proposed standards of performance were based on test results
conducted in accordance with EPA Method 5 and the LAAPCD method, as
discussed in the preamble to the proposed standards. Emission tests
(using EPA Method 5) on three container glass furnaces equipped with
ESPs indicate an average particulate emission of.0.06 g/kg (0.12 Ib/ton)
of glass pulled. These tests show the achievability of the proposed
standards through the use of ESPs. These glass melting furances are
typical of glass melting furnaces for container glass manufacturing
plants. Process factors, such as production rates and glass type, of
these plants are representative of container glass melting furnaces
likely to be built.
The least representative aspect of these furnaces is that none of
them used fuel oil as the fuel of combustion. Use of natural gas does
not generate sulfur oxides in addition to those resulting from the raw
materials. Use of fuel oils containing sulfur most likely increases
the generation of sulfur oxides. In some cases, these sulfur oxides
may require removal. One reported approach is to introduce a chemical
absorbent and collect the sulfur oxides in a dry particulate. As one
commenter noted, these particulates increase the quantity of total
particulates that must be removed in order to comply with the proposed
standard.
The quantity of particulates added as a result of using a chemical
absorbent to remove sulfur oxides does not preclude the achievability
of the proposed standards. If a plant must remove sulfur oxides
through absorption in addition to particulates, the total quantity of
particulates would be known. With this factor known, design of an ESP
that could achieve the proposed standard is possible. Also, the size
2-40
-------
and therefore the cost of an ESP would change only minimally because
the size of an ESP is primarily influenced by the volumetric gas flow
rate and minimally influenced by the quantity of the particulates to
remove. Consequently, this addition of particulates does not preclude
the achievability of the proposed standards. This was seen in test
results submitted by one commenter. Thus, this aspect of the data
base does not indicate that the proposed standards are not achievable.
Emission test data for container glass furnaces equipped with
fabric filters are not available. However, emission test results for
a pressed and blown glass furnace melting a soda-lime formulation
essentially identical to that used for container glass indicate that
emissions can be reduced to 0.12 g/kg (0.24 Ib/ton) of- glass pulled
with a fabric filter. This fabric filter installation was tested with
the Los Angeles Air Pollution Control District particulate matter test
method (LAAPCD Method), which considers the combined weight of the
particulate matter collected in water-filled impingers and of that
collected on a filter. EPA Method 5 also uses impingers and a filter,
but considers only the weight of the particulate matter collected on
the filter. The LAAPCD Method collects a larger amount of particulate
matter than does EPA Method 5, and, consequently, greater mass emissions
would be reported for comparable tests. Using only the "dry" data
(front-half data) compiled in this test report an emission rate of
.076 Ib/ton is calculated. Therefore, it can be readily assumed that
an emission level of 0.1 g/kg (0.2 Ib/ton), as determined by EPA Method 5,
could be achieved by a container glass furnace equipped with a properly
designed and operated fabric filter.
Even though this furnace melted a soda-lime formulation essentially
identical to that used for container glass, its glass production rate
is low compared to typical container glass melting furnaces and it
also burned natural gas. However, design of a fabric filter could be
completed that would achieve the proposed standards for container
glass melting furnaces because the factors needed for the design would
be known.
Based on data compiled during the development of this rulemaking,
an emission level of 0.1 g/kg (0.2 Ib/ton) of glass pulled from container
2-41
-------
glass furnaces firing natural gas that can be achieved with either
electrostatic precipitation and fabric filters. The promulgated
standards reflect the degree of continuous emission reduction which
the EPA has determined to be adequately demonstrated after taking into
consideration the cost of achieving such emission reduction, nonair
quality health and environmental impacts, and energy requirements for
each category of glass manufacturing plants. These considerations are
discussed in sections Environmental Impacts, Economic Impacts, and
Energy Impacts.
Pressed and Blown Glass. One commenter questioned the validity
of referenced test results cited in Chapter 4 of the Background
Information Document, Volume I. Specifically, test Reference Numbers 25
and 41 through 44 were mentioned as being either inaccurate or misreferenced.
As explained below, test Reference Numbers 41 and 44 were incorrectly
reported in BID, Volume I. These incorrectly reported tests were used
in part, as the basis for the proposed standards for the pressed and
blown glass (other than soda-lime) subcategory. Thus the change in
the standards, discussed below, was required.
The data for test reference No. 25 were verified to be correct;
however, the source was misreferenced as being from an Owens-Illinois
plant. The data were actually from a soda-lead borosilicate glassware
Corning Glass Works plant and should have been referenced to Reference 22.
In Table 4-4, page 4-23 of the Background Information Document,
Volume I, the omission of the particulate emissions from the precipitator
outlet for Test No. 41 was an oversight. Upon reviewing the test
data, two sets of data were found from the same referenced company.
These tests were of borosilicate furnaces controlled by electrostatic
precipitators.
The first test, test No. 41a, was of a 25 tons/day borosilicate
furnace. This test was conducted February 25-26, 1976, using EPA
Method 5. Based on the input data, the average emission rate was
0.09 Ib/ton. When correcting to an estimated 85 percent pull rate,
the new emission rate is 0.11 Ib/ton. During the testing period, the
ESP was processing about 42 percent of the design volume flow rate
(scfm).
2-42
-------
The second emission test, designated as 41b, was for a
95.6 tons/day borosilicate furnace. This test was conducted
September 28-30, 1977, using EPA Method 5. Based on fill rates, the
average emission rate was 1.42 Ibs/ton. The corrected emission rate
is 1.67 Ibs/ton for 85 percent pull rate. During the test, the ESP
unit was reported to be operating at about 61 percent over the design
flow rate (scfm), and the efficiency of the precipitator was about
83 percent.
These two emission tests, 41a and 41b, yielded emission rates
from ESPs operating at about 60 percent under design and 60 percent
over design capacity, respectively. These test results are also from
furnaces differing in size. Thus, a conservative assumption would be
that an average of the two test results is representative of the
actual emission rates. Accordingly, an average emission rate of
0.89 Ib/ton will be used for emission test Reference 41.
Referenced tests 42 and 43 were performed at the cited facilities
and show levels of control achievable by ESPs on a continuous basis.
Therefore, the origin of information was correct and not incorrect as
the commenter suggested test No. 42 measured an emission rate of
1.14 Ib/ton and test No. 43 measured an emission rate of 0.96 Ib/ton.
However, emission test Reference 44 contained erroneous mass
emission rates. From the April 28, 1976, test performed on the
ESP-controlled furnace using EPA Method 5, an emission rate of 0.95 Ib/ton
was calculated. The furnace for this test was fired with No. 5 fuel
oil with electric boost and was melting borosilicate glass.
Several commenters suggested that various glass products.be taken
out of their generalized categories and be set apart individually as
additional glass manufacturing categories. The necessity for this
action, they claimed, is the individually unique processing character-
istics of particular glass products and in their unique emission
characteristics. They claim that specific glass products, such as
textile and continuous-strand fiberglass [included in the proposed
pressed and blown (other-than soda-lime) category], borosilicate
2-43
-------
pressed and blown glass [included in the proposed pressed and blown
(other-than soda-lime) category], and hand blown glass [included in
the proposed pressed and blown (other-than soda-lime) category],
contain either unique processing characteristics or emission
characteristics, or both.
However, the excision to divide the glass manufacturing industry
into four categories, as discussed earlier, was made based on technical
and economic considerations. Rather than promulgate an individual
standard for each individual type of glass item produced, which number
in excess of 80, four categories were chosen based on engineering
judgment that was based on the similarities of the process factors
including characteristics of glass melting furnaces and types of glass
produced. As discussed below, further division of the pressed and
blown glass category was based on technical considerations to make
certain the standards are achievable. Taking into consideration the
basic technical and economic characteristics of all the types of glass
manufactured, four major categories were finally chosen to represent
the industry: container glass, pressed and blown glass (borosilicate;
soda-lime and lead; and other-than borosilicate, soda-lime, and lead),
wool fiberglass, and flat glass.
The further division of the pressed and blown glass manufacturing
category has not altered the predicted impacts of the promulgated
standards. The new plants, predicted to be constructed within the
impact analysis period for the two original subcategories, were allotted
to the new subcategories, the total number of new plants remaining the
same. In addition, as discussed in the Economic Impacts section of
this document, the cost and economic analyses do not require substantial
changes and indicate the same impacts as presented when the standards
were proposed; i.e., the cost and economic impacts are reasonable
because the standards are clearly affordable and the product pricing
is less than 1 percent.
The decision to subdivide the pressed and blown glass category
into three subcategories was based on test data and information gathered
throughout the development of these standards. In studying the data
and information it was found that borosilicate-type glass emissions
2-44
-------
were uniformly the most difficult to control, with soda-lime and lead
glass emissions being relatively more controllable.. With these two
extremes in potential particulate emission control, the balance of the
pressed and blown glass formulations (other-than borosilicate, soda-lime,
and lead) were found to be controlled, at least, at a relatively
median level of control.
As a result of comments received on the proposed standards, a
review of the achievability of standards for pressed and blown
borosilicate glass melting furnaces was performed to evaluate industry's
contentions that the proposed standard for pressed and blown boro-
silicate glass melting furnaces was "unachievable. It was alleged that
these furnaces are up to four times more difficult to control than
other types of furnaces. In addition, as explained above, the tests
had been incorrectly reported in the Background Information Document,
Volume 1. These tests dealt with the pressed and blown borosilicate
subcategory. Also, a commenter contended that pressed and blown lead
glass manufacturing plants could achieve a more restrictive standard.
Emission tests using EPA Method 5 on four furnaces melting pressed
and blown glass with borosilicate formulations and equipped with ESPs
yeilded a representative emission rate of about 1.0 Ib/ton of glass
produced. All of these test results except one were less than 1.0 Ib/ton.
The test result greater than 1.0 Ib/ton was collected at an ESP with a
2
specific collection area of 0.65 ft /scfm and found an emission rate
of 1.14 Ibs/ton. For the other tests, the specific collection area
2 2
was greater than 0.85 ft /scfm and averaged about 1 ft /scfm. The
2
specific collection area (ft of ESP plat collection area 1 standard
cubic foot per minute of gas flow) directly influences the collection
efficiency of an ESP. Thus, the test result showing 1.14,Ibs/ton only
indicates that an ESP operated less efficiently would not reduce the
emission rate to less than 1.0 Ib/ton. In evaluating the size of an
ESP needed to meet the proposed standard, a specific collection area
o
of 1.0 ft /scfm was used. Therefore, because the emission tests did
not confirm the achievability of the proposed standard of 0.5 Ib/ton
and the proposed standard needs changed, a review of the test data was
needed. These emission test data indicate that a limit of 1.0 Ib/ton
2-45
-------
would be achievable for furnaces melting borosilicate type glasses
controlled by an electrostatic precipitator. This limitation of
1.0 Ib/ton is promulgated as the standard for pressed and blown
(borosilicate) glass.
Emission test results for pressed and blown lead glass plant
average 0.23 Ib/ton. Two of six test results are higher in value than
this average. Reviewing the information on these plants indicates
that the specific collection area for one of these two plants was less
than the specific collection area for the remainder of the plants.
However, other process variables were similar. The information on the
other plant did not include the specific collection area. The average
of the test results for the remaining plants is 0.13 Ib/ton. Thus,
because the information on all the tests shows that the larger values
are most likely a result of ESPs designed for less efficiency than the
remainder of the values, these emission tests indicate that furnaces
melting lead glass can achieve a limit of 0.2 Ib/ton, equal to that
proposed for soda-lime pressed and blown glass controlled by an
electrostatic precipitator or possibly a fabric filter. As a result,
a second subcategory for pressed and blown (soda-lime and lead) glass
is promulgated at 0.2 Ib/ton.
Finally, glasses being produced that are of other-than borosilicate,
soda-lime, and lead recipes have been found, as discussed in the
preamble to the proposed standards, through referenced tests and data
set out in Tables 4-2 and 4-4 of the Background Information Document,
Volume I, to be able to achieve a 0.5 Ib/ton standard with either an
electrostatic precipitator or a fabric filter. As a result, a third
subcategory for pressed and blown (other-than borosilicate, soda-lime,
and lead) glass is promulgated at 0.5 Ib/ton.
Another commenter stated that the average controlled emissions
from a borosilicate furnace controlled by an electrostatic precipitator
is 1.56 Ibs/ton. It was his opinion that the standard should be
1.56 Ibs/ton.
Standards of performance are not based on controlled emissions
averaged over a number of plants. These standards are based on the
application of the best technological system of continuous emission
2-46
-------
reduction which (taking into consideration the cost of achieving such
emission reduction, any non-air quality health and environmental
impact and energy requirements) the Administrator has determined to be
adequately demonstrated. In choosing a limitation indicative of the
best technological system of continuous emission reduction, emission
test data from the better controlled, representative sources are
reviewed and a limitation indicative of these emission tests data is
chosen as the standard.
Wool Fiberglass. A commenter pointed out that the three emission
tests for wool fiberglass furnaces controlled by ESPs referred to in
the preamble for proposal and the Background Information Document,
Volume I, were actually three runs taken in sequence on the same
furnace and were not, as had been suggested in the BID, Volume I, from
three separate furnaces. The three referenced tests were actually
three runs on the same furnace as pointed out by the commenter. This
test of three runs averaged to 0.36 Ib/ton. A series of five tests
was actually conducted. The first three tests were run with all three
fields of the electrostatic precipitator (ESP) energized. The last
two tests however, were run with less than the total number of electrical
fields.
The first test, conducted after putting the ESP on the line,
yielded an emission rate of 0.72 Ib/ton. This result is unusual in
that the test was conducted in what appears to have been similar to
startup conditions. Just prior to commencement of the testing the ESP
was taken off line, inspected, and cleaned. This test, however,
cannot be considered non-representative for this unit, but is considered
to be highly suspect after comparing it with the tests conducted
immediately following it.
The next two tests were run at substantially identical conditions
and yielded results of 0.19 Ib/ton and 0.17 Ib/ton, respectively. The
fourth test, conducted on the same day as the third, was unlike the
three prior tests in that the ESP was only operated with two of the
three electrical fields energized. Despite this reduction in collection
potential the emission rate was 0.2 Ib/ton, substantially lower than
the first test and noticably similar to the second and third tests.
The fifth test was conducted with only one of the three electrical
fields operating and yielded an emission rate of 2.25 Ibs/ton. This
2-47
-------
result shows the effects of operating a control device at less than
designed operating levels. This test should obviously be discounted
as being unrepresentative of normal operating conditions for an ESP
controlled glass melting furnace.
Emission tests conducted on wool fiberglass furnaces controlled
by ESPs and fabric filters indicate that a level of 0.5 Ib/ton can be
achieved. Emission tests indicate a level of 0.36 Ib/ton is achievable,
even using a worst-case interpretation of the data, by an ESP and a
level of 0.4 Ib/ton is achievable by fabric filters. Also, an emission
test using the LAAPCD method on a wool fiberglass plant using a fabric
filter control device shows a level of 0.52 Ib/ton which indicates
that a standard more restrictive than 0.52 Ib/ton would be attainable.
This is a reasonable assumption because of differences between EPA's
Method 5 and the LAAPCD's testing method. Test results for a furnace
tested according to the LAAPCD's method are higher than when tested
according to EPA's Method 5. Based on this, it can be assumed that
this standard can be met by using an electrostatic precipitator or a
fabric filter. Because the data that was the basis for the proposed
standard was three runs of one test and not three tests, the standard
has been changed from 0.4 Ib/ton to 0.5 Ib/ton. The impacts associated
with the change in the standard from 0.4 Ib/ton to 0.5 Ib/ton will be
minimal and are explained in the Environmental, Economic, and Energy
Impact sections of this document.
One commenter suggested that the wool fiberglass industry utilizes
a unique process separate from the other categories and should not
have its standard based on a technology transfer. The'opinion of this
commenter was that the wool fiberglass sector of the glass manufacturing
industry had not been sufficiently investigated.
The standard is based solely on test results conducted at wool
fiberglass plants with emissions typical of wool fiberglass plants.
Technology transfer was not used in developing the standard for wool
fiberglass manufacturing plants. However, as discussed above, the
design of a control device to bring a wool fiberglass manufacturing
plant into compliance with- this standard would require consideration
of plant-specific process factors. If a wool fiberglass plant, for
2-48
-------
instance, uses a recuperative heat recovery system rather than a
regenerative system, then the ordinary design of the control device
must take into consideration the uniqueness that may result from this
difference. These differences were considered in proposing the standard
for wool fiberglass plants. In addition, a plant was tested and a
level of 0.52 Ib/ton was measured using the LAAPCD's method. Emission
test results were collected, reviewed, and analyzed for wool fiberglass
plants having glass melting furnaces controlled by fabric filters and
ESPs, and plants utilizing all-electric glass melting furnaces. These
results reflect the ability of this category's glass melting furnaces
to meet the standard with ESPs and fabric filters.
Flat Glass. Several commenters suggested that control technology
in the flat glass industry has not been adequately demonstrated and
that the proposed standard of 0.3 Ib/ton was too stringent. Commenters
also disagreed with the transfer of control technology from container
glass melting furnaces to flat glass melting furnaces. Commenters did
not consider the proposal of a standard based on technology transfer
to be reasonable.
The basis for using technology transfer was the similarity in the
soda-lime formulations used by both container and flat glass manufac-
turers and the chemical composition and the physical characteristics
of the particulate emissions. The conclusion that the percentage
reduction in particulate emissions achieved by the control of container
glass furnaces could also be achieved by the control of flat glass
furnaces, thereby lending credence to the transfer of technology
theory, was supported by an ESP manufacturer's performance guarantee.
Since proposal, an emission test performed by EPA and an ESP
manufacturer has become available for an oil-fired flat glass plant
with an ESP. This emission test does not clearly confirm that the
proposed emission limit of 0.3 Ib/ton is achievable. This test was
performed at a 1-year-old flat glass plant. To maintain confiden-
tiality, as requested by the plant, the production rate must be withheld
but a particulate emission rate was measured to be about 5 Ibs/hr.
Based on these factors and an allowance to ensure the achievability of
the standard, the promulgated standard for flat glass melting furnaces
2-49
-------
is set at 0.45 Ib/ton. This standard is clearly supported by the
emission test performed by EPA.
The promulgated numerical emission limit includes an allowance
for variations in exact testing procedures, age of control devices,
and the limited data upon which the standard is based. The allowance
is primarily a means of assuring that flat glass plants can achieve
the promulgated standards especially in light of the fact that one
emission test is the basis upon which the standard is set. The flat
glass standard does not have a fuel-oil increment, as in the proposed
standards, because the promulgated standard is based on emission tests
performed at an oil-fired flat glass manufacturing plant. These tests
were conducted in accordance with modified EPA Method 5 procedures. A
temperature of 350°F was maintained in the filter box. Refer to the
Test Methods and Monitoring section of this document. .
Commenters suggested that process modifications were the only
demonstrated techniques available for particulate emission reduction
from flat glass manufacturing plants. Process modifications were
considered in the development of the promulgated standard. However,
test data available before proposal indicated that emission reduction
by process modifications is indefinite with respect to the effective-
ness of such techniques. Refer to the general discussion on process
modifications located in this section of this document. The selection
of the best demonstrated system of continuous emission control was
based on analyses performed determining the technological effectiveness
and the environmental, economic, and energy impacts associated with
its selection. These analyses led EPA to the conclusion that a level
of control more stringent than that that may be consistently achieved
by process modifications represents the best demonstrated system of
continuous emission reduction for this industry.
The members of the flat glass manufacturing industry that support
basing the standard on process modifications claim that they can
consistently maintain emissions at a rate of approximately 1.1 Ibs/ton.
However, an emission test was performed in an attempt to confirm
compliance with a State's regulation by using process modifications.
The results of those emission tests yielded an average controlled
2-50
-------
emission rate of 1.4 Ibs/ton, which is higher than the industry suggested
process modification rate of 1.1 Ibs/ton. As shown by this example
and other emission tests, an emission rate of 1.1 Ibs/ton achieved
through process modifications is not certain. Only an average of
approximately a 30 percent emission reduction, or 1.4 Ibs/ton could be
achieved by process modifications if an uncontrolled emission rate of
2.0 Ibs/ton is assumed.
An additionally important factor in considering the industry's
suggested use of process modification techniques is that the technology
involved in achieving this rate is considered confidential by those
manufacturers who recommend the 1.1 Ibs/ton rate. It seems that only
the major manufacturers of flat glass have the research and development
resources at their disposal to experiment with process modifications
in the form of altered batch formulations and forms of innovative
technology. And many of these batch formulations and technological
innovations are in experimental development. In fact, comments have
indicated that product quality may be affected by process modifications.
In these instances, suspension of the use of process modifications,
necessarily resulting in the exceedance of the standards, may be
necessary to expedite the attainment of product quality.
Data made available since proposal have not added any certainty
to the question of the effectiveness of process modification techniques
for flat glass manufacturing. From the test results received, it is
apparent that no consistency in emission reduction can be achieved by
any one form of process modification. Results range from 0.5 Ib/ton
to greater than 2.0 Ibs/ton for flat glass melting furnaces
experimenting with process modifications.
Several commenters also contended that various SIP limitations
were being met by flat glass manufacturers by producing without add-on
control devices and using process modifications. Using the baseline
as an indicator of what is typically required of new flat glass plants
to meet, it can be seen that the industry's suggested uncontrolled
emission rate achieved through the use of confidential process
modifications (1.12 Ibs/ton) cannot comply with a majority of the
SIPs, although some do. Flat'glass plants located in California and
2-51
-------
New Jersey have been equipped with add-on controls to comply with
SIPs. Industry's ability to comply with State limitations by using
process modifications and without the use of add-on control equipment
has been accomplished, in some cases, through reduction in production
rates.
The flat glass industry's suggestion that the standard should be
based on an apparently experimental and unproven means of emission
reduction is not convincing. EPA considers the use of process
modifications in the manufacture of flat glass to be an inadequately
demonstrated means of emission reduction. Thus, even though the
intent of the standards is not to preclude the use of process modifi-
cations, the selection of the best system of continuous emission
reduction is based on add-on emission reduction techniques of known
and proven effectiveness that have been considered adequately demonstrated.
The available information concerning the methods of emission
reduction which industry has suggested for the basis of this standard
has not led the EPA to change the decision that add-on control technology
should be the basis of this standard of performance. However, EPA has
information which indicates that the standard could be achieved using
process modification techniques in some instances. This information
is unique to one manufacturer, is considered proprietary by this
manufacturer, and does not, as stated before, indicate that process
modification techniques are demonstrated methods of control for the
glass manufacturing industry.
2.3 MODIFICATION, RECONSTRUCTION, AND OTHER CONSIDERATIONS
Most of the comments concerning the exemption of all-electric melters
[40 CFR 60.292(e)], the fuel conversion exemptions [40 CFR 60.292(c)],
the exemption for plants producing less than 2.0 tons of glass per day
[40 CFR 60.292(e)], and the rebricking exemption [40 CFR 60.292(d)]
supported them as being necessary for the future development of the
glass manufacturing industry. However, two commenters suggested that
the all-electric melter exemption and the liquid fuel increment are
inappropriate in light of the energy constraints under which the
Nation is presently operating. One commenter expressed the opinion
that to encourage the use of electric power and liquid fuels, such as
2-52
-------
fuel oil, as opposed to natural gas was contrary to national energy
policies.
During the course of the development of these standards, it was
found that, due to production constraints, all-electric melters are
only feasible for furnaces of relatively low pull rates. This being
the case, only a fraction of the projected new plants predicted to be
constructed over the 5-year period following the promulgation of these
standards can possibly use electric power for melting exclusively.
Thus, the exemption for all-electric melters is not expected to shift
the energy use of this industry to any great extent. Another energy
factor taken into consideration is the fact that electricity will be
increasingly produced by the firing of coal. This is consistent with
the Nation's goals of attempting to utilize its coal resources rather
than importing other fossil fuels.
Available information on well-operated and maintained all-electric
furnaces indicated that particulate emissions are only slightly higher
than fossil fuel-fired furnaces controlled to meet the promulgated
standards. Most of these all-electric furnaces are open to the atmosphere
and do not have stacks. In order to test all-electric melters, a
stack or concentrated emission outlet would have to be designed and
constructed. In light of the minimal emissions produced by the
all-electric melter melting process, the cost and inconvenience were
determined to outweigh any benefits that might accrue. Therefore,
all-electric melting furnaces are not regulated by the promulgated
standards.
One commenter suggested that the all-electric glass melting
furnace exemption would encourage all new furnaces, estimated to be
constructed through 1984, to use electricity as their only means of
heat production. It was the commenter's opinion that to have this
occur would precipitate a secondary environmental impact at the electric
generating plants that would more than offset the emission reduction
benefits realized at the glass manufacturing plants.
A major element upon which the commenter based his conclusion is
incorrect. It was assumed by the commenter that all of the estimated
production growth through 1983 would be attributed to all-electric
2-53
-------
glass melting furnaces. All-electric melting technology has several
key limitations which restrict its applicability throughout all
categories of the glass industry. Not all glasses possess the electrical
properties required for successful all-electric furnace operation;
certain glass formulations attack the electrodes presently used in
all-electric furnaces; and all-electric melter technology is not far
enough advanced to satisfactorily produce glass in large quantities.
Consequently, all-electric furnaces are presently limited to portions
of the container, pressed and blown, and wool fiberglass categories.
The commenter estimated the secondary environmental impact
attributable to these standards to be approximately 50,000 tons of
particulate emissions per year. It is uncertain as to the commenter's
source for his air pollution per BID figures, but it appears to have
been some form of uncontrolled emission rate for a coal-fired utility
boiler. Using AP-42's emission factor for an uncontrolled utility
boiler fired with pulverized bituminous coal, the coal having a heating
value of approximately 13,000 BTUs per pound, an emission rate of
approximately 0.62 pounds of particulate per million BTUs is calcuated.
This would result in a net reduction in emissions of approximately
37 percent. It must be noted, however, that as of 1971 new coal-fired
utility boilers have had to comply with a new source performance
standard (NSPS) limiting this source's emissions to 0.1 pound of
particulate per million BTUs. Using this NSPS limitation, a net
reduction of approximately 82 percent is achieved. Additionally, it
should be noted that as of 1979 new coal-fired utility boilers will
have to comply with a recently revised NSPS limitation of 0.03 pound
of particulate per million BTUs. The implementation of this newest
limitation on coal-fired utility boilers will result in an net emission
reduction of approximately 87 percent.
Based on the calculations presented above for a 100 ton per day
furnace operating 365 days per year, with referenced power requirements
of 850 kWh/per ton of glass produced and 10,000 BTUs per kWh required,
the estimated minimal impacts and the benefits expected to accrue from
the implementation of this technology led to the all-electric melter
exemption being retained.
2-54
-------
A few commenters were of the opinion that implementation of the
liquid fuel increment allowed for the firing of fuel oil would encourage
the burning of fuel oil at glass manufacturing plants. It should be
recognized that, barring substantial supply curtailments, glass
manufacturers will use natural gas rather than fuel oil. Natural gas
is preferred because fuel oil burns less cleanly and is less combustion
efficient. For these reasons, as well as others of less significance,
natural gas is perferred to fuel oil. The ultimate influence will be
economics and not the liquid fuel increment. Without a disproportionate
imbalance in price and availability the use of fuel oil will not be
encouraged by this liquid fuel increment.
Several commenters questioned the 15 percent allowance in the
liquid fuel increment in the proposed standards for the firing of fuel
oil as opposed to natural gas. They were of the belief that a
substantially larger increment would more accurately reflect the
actual increase in particulate matter emissions attributable to the
firing of fuel oil in glass melting furnaces.
These standards of performance are based predominantly on test
results from glass melting furnaces firing natural gas. Realizing the
increased particulate emissions experienced when firing fuel oil as
opposed to natural gas, an increment was developed for the proposed
standards that was considered representative. This increment was set
at 15 percent as a result of an analysis performed with the data that
were available at that time.
Since proposal, additional data have been received that, when
considered with data already in EPA's possession, suggest a larger
percentage increment might be more representative for the industry.
The data considered were submitted by both industrial and governmental
sources.
Data in EPA's possession prior to proposal specifically comparing
one furnace's particulate matter emissions while firing oil and then
natural gas were compiled by EPA's Industrial Environmental Research
Laboratory (IERL). These tests were part of an IERL test program
conducted on uncontrolled glass melting furnaces in 1976. The four
2-55
-------
tests considered in setting the increment for the proposed standards
were conducted on a pressed and blown (borosilicate) glass melting
furnace. Refer to docket entries OAQPS-77/1-II-B-81 and II-B-101.
Two of the four tests were conducted on this furnace while firing
fuel oil of an unknown sulfur content. The remaining tests were
conducted while firing natural gas. The natural gas firing tests
revealed parti ail ate emissions averaging approximately 6.5 Ibs/hr
while the fuel-oil firing tests averaged approximately 7.3 Ibs/hr. In
comparing these emission rates, as to the amount of glass produced at
the times of the tests (24.6 tons/day and 25.1,tons/day, respectively),
figures of 6.34 Ibs/ton of glass for natural gas firing and 7.00 Ibs/ton
of glass for fuel-oil firing were calculated. Therefore, the difference
in emissions attributable to the firing of these two fuels was
approximately 10 percent.
Since proposal, Ford Motor Company (Ford) has submitted test data
(docket entry OAQPS 77/1-IV-D-12) comparing particulate emissions from
its Nashville Number 2 furnace. These test results were compiled over
a span of approximately five years on an uncontrolled flat glass
melting furnace. The initial tests, conducted while firing natural
gas, ranged from a high of 2.81 Ibs/ton in March 1973 to a low of
1.05 Ibs/ton in April 1977. The firing of fuel oil, on the other
hand, resulted in a range of emissions from 1.40 Ibs/ton in April 1978
to 1.69 Ibs/ton in May 1978.
For purposes of this analysis, the natural gas-firing test from
March 1973 was considered, but was not taken into account, as it was
inordinately high with respect to the other test results. A possible
reason for this high test result could be attributable to the relatively
recent advances made by the industry in salt cake reduction in batch
formulations. Averaging the natural gas-firing test results covering
a period of less than two years (December 1975 to October 1977), an
emission rate of 1.41 Ibs/ton was calculated. This rate, when compared
to the average emission rate calculated for firing fuel oils of varying
sulfur contents of from 0.73 percent to 0.98 percent, (1.55 Ibs/ton),
revealed a difference of approximately 10 percent.
2-56
-------
It was the commenter's suggestion that the last two test results
for fuel oil-firing should be used when comparing.the emissions attri-
butable to fuel oil-firing as opposed to natural gas-firing. This
average emission rate was 1.14 Ibs/ton and when compared to the average
rate for fuel oil-firing, 1.55 Ibs/ton, resulted in a 36 percent
increase in particulate emissions attributable to fuel oil-firing.
However, a more accurate comparison can be made by using the test
results compiled by Ford while firing natural gas in October 1977
(1.23 Ibs/ton) and then firing Number 6 fuel oil in November 1977
(1.56 Ibs/ton) on the same furance. This 1-month testing interval
tends to minimize the effects on the test results that may be attri-
butable to batch modifications and represents fuel oil-firing for an
oil with 0.85 percent sulfur content. This comparison reveals a
difference in emissions of approximately 27 percent.
Owens-Illinois submitted test data for its Mansfield, Massachusetts,
container glass plant (docket entry OAQPS 77/1-II-D-238). These tests
were conducted in April 1978 to measure the control efficiency of a
electrostatic precipitator (ESP) controlling two fuel oil-fired furnaces.
The ESP is also operated in conjunction with a sodium carbonate
pretreatment spray system.
It was explained by Owens-Illinois that the installation of this
pretreatment system, located before the electrostatic precipitator,
was necessitated by difficulties encountered in the removal of the
particulate from the ESP's collector plates. It was pointed out by
Owens-Illinois that sulfur in the fuel oil, upon combustion, reacts
with the particulates and significantly contributes to the accelerated
deterioration of these plates. In an effort to minimize this deterio-
ration, Owens-Illinois developed a sodium carbonate exhaust gas
pretreatment system. It is the purpose of this system to convert the
sulfur trioxide formed in the combustion of the fuel oil to an alkaline
sulfate particulate. This particulate matter, it was explained, is
easier to remove from the ESP collector plates and minimizes plate
deterioration.
The test results from three runs taken at the outlet of this
2-year-old electrostatic precipitator ranged from 0.17 Ib/ton to
0.31 Ib/ton. The average emission rate for the three runs was
2-57
-------
0.26 Ib/ton while firing 0.9 percent sulfur Number 6 fuel oil. It is
important to note that the successive runs did not follow an increasing
or decreasing pattern; the last run presented the lowest emission
rate. It should be additionally noted that the promulgated standard
is 0.26 Ib/ton for container furnaces firing fuel oil. This Owens-Illinois
test demonstrates the ability of two glass container furnaces controlled
by an ESP, in operation in excess of two years, to meet the promulgated
standard with the introduction of added pretreatment material to the
gas stream.
Judging from the data collected, the percentage of sulfur contained
in No. 6 oil used to fire a glass melting furnace influences the
amount of particulate emissions collected while conducting source
sampling. Although the reason for this phenomenon is not fully understood,
it is considered a possibility that EPA's Test Method 5. requirement
that the filter temperature be maintained at 250°F may effect these
results. In response to comments received and tests conducted subsequent
to proposal, the Method 5 filter temperature has been raised to up to
350°F. Data indicate that Method 5 collects sulfuric acid mist at a
filter box temperature of 250° F but not at 350°F. By changing the
filter box temperature to 350° F, the influence of the sulfur in No. 6
oil as well as in the raw materials should be diminished. Refer to
the Test Methods and Monitoring section (2.8) of this document for
details on the issue concerning the filter box temperature.
It should also be pointed out that comments offered by Mr. Strelow
in behalf of the Glass Packaging Institute (docket entry OAQPS 77/1-IV-D-19,
at 78-82) inadvertently misinterpreted a study prepared by the New
Jersey Bureau of'Air Pollution Control on Fuel Oil Particulate Emissions
from Direct Fired Combustion Sources (docket entry OAQPS-77/1-II-I-22).
It was Mr. Strelow's contention that the calculations presented on
page 14 of the study (Glass Furnaces Data Sheet) represented increases
in emissions attributable to the firing of fuel oil over natural gas.
The numbers cited in Mr. Strelow's comments (30.60 percent and
89.95 percent) do not represent increases in particulate emissions but
instead represent theoretical calculations based on emission factors
developed by New Jersey's Bureau of Air Pollution Control estimating
2-58
-------
the percent of total particulate emissions that could be attributable
to the firing of fuel oil. There is no mention of emissions attributable
to the firing of natural gas; therefore, no comparison of the two
based on these data can be made.
As a result of data received, comments offered, and a review of
the issue the increment for firing fuel oil was increased to 30 percent.
In choosing the 30 percent increment, the theoretical analysis was
weighed against the data concerning this issue. The data collected
through testing were used to establish the increment because the
theoretical analysis is questionable. The theoretical analysis is
particularly questionable after considering the effects of the change
in the filter box temperature and its influence on the amount of
particulate indicated by the original Method 5 tests that support the
theoretical analysis. In addition, data collected through testing
would include variables not possible to include in the theoretical
analysis, such as the influence of the requirement for sulfur as part
of the glass product.
The liquid fuel increment is needed to compensate for the additional
particulate that results from burning a fuel that contains ash. The
standards were based mainly on data collected at gas-fired furnaces.
Therefore, an increment such as has been implemented is appropriate.
Commenters suggested that the proposed exemption for "day pot
furnaces," or more accurately termed day tanks, week tanks, or pot
furnaces, does not adequately cover small production glass furnaces.
Specifically in question was whether the proposed two tons of glass
produced per day (TPD) criterion as an exemption for "day pot furnaces"
was an adequate representation of the typical design capacity of a
newly constructed small production glass furnace.
The primary reason for the provision of this exemption was economic.
These small glass melting furnaces constitute an extremely small
percentage of the Nation's total glass production. Commenters stated
that the cost to control these small volume furnaces in some cases was
as much as 10 times the cost to construct a new furnace. As stated in
preamble to the proposed standards, their control is considered
economically unreasonable. It is the intent of this exemption to
2-59
-------
exclude small glass manufacturing plants from having to comply with a
standard that would effectively put them out of business.
Data submitted by industry representatives indicated that a
greater production rate should be used as a cutoff for this exemption.
It was pointed out that the approximate average daily pull rate for
day tanks is presently 0.989 tons and for week tanks is presently
4.1 tons. A commenter suggested that tanks of a production capacity
of equal to or less than 5 TPD, the largest daily pull rate for any
one furnace in either category, should be exempted from the standards.
Industry statistics show that this daily production rate (5 TPD) would
be a maximum for pot furnaces, day tanks, and week tanks that are
likely to be constructed. Thus, the exemption was extended from 2 TPD
to 5 TPD.
In developing the promulgated standards it was ultimately decided
to exempt all hand glass melting furnaces from having to comply with
these standards. This decision was based on a further analysis of the
industry precipitated by comments received. Hand glass furnaces would
not likely be able to survive the associated economic impact. Thus,
hand glass melting furnaces were exempted from having to comply with
these standards of performance. This is not to say that standards
will never be developed for these furnaces; standards may, at some
later date, be developed if a subsequent analysis of this segment of
the industry shows that they are necessary arid viable. There would be
no associated impacts as a result of this change as there were no hand
glass furnaces expected to be built within the next five years.
Other commenters suggested that the cutoff be set at as high as
15 TPD. All known furnaces producing greater than 5 TPD are continuous
production furnaces. The proposed standards addressed continuous
production furnaces and evaluated the smallest continuous production
furnaces at an average operational capacity of 50 TPD. No comments
were received that indicated this size furnace was not representative
of new small continuous production furnaces. In any case, the intent
of these standards is to cover continuous production furnances and
therefore, any new glass melting furnace with a production capacity in
excess of 5 TPD, not specifically exempted, is subject to the standards.
2-50
-------
In addition, in gathering data to support the standards, information
and data on existing glass melting furnaces producing less than 50 TPD
were reviewed and considered in establishing the standards. All
indications were that new continuous glass melting furnaces that are
being constructed are 50 TPD or more. Thus, the exemption was extended
only to include all glass melting furnaces designed to produce five tons
or less of glass per day.
2.4 GENERAL ISSUES
Several commenters stated that the State Implementation Plan
(SIP) used for comparison of the impacts of the proposed standards was
not typical. They stated that a SIP such as New Jersey's should not
be considered typical for the industry due to the relatively few glass
manufacturing plants located in that State. Prior to the proposal of
these standards, an analysis was made of this very issue and was
included in the docket (docket entry OAQPS 77/1-II-A-16).
The analysis explained that to visually compare the baseline SIP
profile on a graph to other SIP profiles would not accurately reflect
the typical ness of the baseline SIP. The typical SIP for this industry
was found to be one similar to both New Jersey's and Pennsylvania's.
To view New Jersey's SIP on a graph, illustrating its regulatory
restrictiveness versus pull rate relationship in comparison to other
states having a ma-jor share of the Nation's glass manufacturing plants
[California, Illinois, and Tennessee (new and existing source rules),
Indiana, New York, Ohio, Oklahoma, Pennsylvania, Texas, and West
Virginia], New Jersey's SIP may not appear to be typical. But, to
accurately reflect the status of the air pollution control requirements
that the glass manufacturing industry would experience if no NSPS were
established, this graphical relationship was considered jointly with a
relationship exhibiting the relative share of the number of glass
manufacturing plants located within each of the States.
The container glass industry is highly concentrated within the
States of California, Pennsylvania, Illinois, and New Jersey. Most
plants of the pressed and blown category are located in West Virginia,
Ohio, and Pennsylvania. Flat glass is for the most part spread evenly
2-61
-------
among Tennessee, Pennsylvania, California, Ohio, and Michigan. And
finally, a large share of the wool fiberglass is produced in Kansas,
Ohio, New Jersey, and California.
The major portion of the container glass industry is required to
comply with SIPs having a restrictiveness comparable to New Jersey's.
It is readily apparent by comparing the graphical relationships with
the number of plants located in each State that the four States with
the largest volume of container glass production (California,
Pennsylvania, Illinois, and New Jersey) also have the strictest standards,
It is especially important to note that the container glass industry
represents approximately 70 percent of the nationwide total glass
production. This gives added weight to considering New Jersey's SIP
typical, especially when comparing the four major container glass
States' SIPs.
The pressed and blown industry plants, on the other hand, are
relatively small in size and production output as compared to the
other categories. Knowing this, it can be seen that West Virginia's
emission limitations are the second most restrictive for the sources
with pull rates of less than approximately 60 tons per day. The State
of Ohio's limitations have not been representative of those by which
industry in general has to abide. Pennsylvania, the third major
pressed and blown industry State, has emission limitations much like
New Jersey's; actually, it's restrictions are more stringent for the
lower production rate glass manufacturing plants than New Jersey's.
Wool fiberglass production is led jointly by California (with
emission limitations very similar to New Jersey's) and Ohio. These
two States are followed by New Jersey, Kansas (median in nature),
Georgia (median in nature), Indiana (median in nature), and Texas
(relatively less restrictive).
The flat glass category is led by Tennessee, a State with a SIP
that is graphically in the middle of the restrictiveness scheme.
Following Tennessee are Pennsylvania (the standard after which New
Jersey's SIP was modeled), California (with emission limitations very
similar to New Jersey's), Ohio and Michigan (which has standards much
the same as New York's, median in nature).
2-62
-------
Choosing New Jersey's SIP as being generally typical throughout
the glass manufacturing industry is reasonable. A large portion of
the industry is located in New Jersey and Pennsylvania arid in States
with similar SIP profiles. Considering the nationwide glass production
represented by these States, the variety of categories of glass produced
within their bounds, and the relative restrictiveness of each of the
SIPs, a SIP similar to New Jersey's is the best choice for the purposes
of the study performed. To select the emission limitations set by New
Jersey's SIP as being typical provides for a representative view of
the industry-wide emission limitations imposed on glass manufacturing
plants as they exist today.
Despite the conclusions of this analysis, an effort was made to
compare impacts using New Jersey's SIP with the impacts using Tennessee's
SIP, a "visually typical" SIP. (docket entry OAQPS 77/1-1I-A-16.)
The cost impacts associated with a baseline based on Tennessee's SIP
yields, as with a baseline based on New Jersey's SIP, a minimal impact
to the consumer. The largest consumer price percent increase attri-
butable to the standards (represented by the container glass category)
based on the selection of New Jersey's SIP is approximately 1.8 percent
and based on the selection of Tennessee's SIP is approximately 2.5 percent.
This analysis overestimated the consumer price percent increase by
choosing Tennessee's SIP but, despite this fact, the consumer price
percent increase of 2.5 percent would not have indicated a prohibitive
impact on the container glass manufacturing industry.
In summary, the suggestion that New Jersey's SIP should not be
considered typical for the glass manufacturing industry does not
warrant the re-evaluation of the impacts associated with the promul-
gation of these standards. Taking into consideration all of the
points just explained, a SIP similar to New Jersey's is considered
typical for the glass manufacturing industry.
The impacts of the standards have been adjusted to the five year
period from 1979 to 1984 because the actual impact of the standards
will begin for plants built after June 15, 1979. However, the dollars
are still on a January 1978 basis.
2-63
-------
A commenter suggested that a different format should be used in
setting the standards. It was commented that different means of
determining compliance should be used. It was specifically suggested
that such criteria as the amount of particulate emitted annually and
the amount of particulate emitted per hour of production be used. It
was additionally s-ggested that the standards be based on fill rate
rather than pull rate, as well as, to develop concentration standards as
opposed to mass standards.
Two alternative formats were considered for the proposed standards:
mass standards, which limit emissions per unit of feed to the glass
furnace or per unit of glass produced by the glass furnace; and
concentration standards, which limit emissions per unit volume of
exhaust gases discharged to the atmosphere.
Enforcement of concentration standards requires a minimum of data
and information, decreasing the costs of enforcement and reducing
chances of error. Furthermore, vendors of emission control equipment
usually guarantee equipment performance in terms of the pollutant
concentration in the discharge gas stream.
There is a potential for circumventing concentration standards by
diluting the exhaust gases discharged to the atmosphere with excess
air, thus lowering the concentration of pollutants emitted but not the
total mass emitted. This problem can be overcome, however, by correcting
the concentration measured in the gas stream to a reference condition,
such as a specified oxygen percentage in the gas stream.
Concentration standards would penalize energy-efficient furnaces
since a decrease in the amount of fuel required to melt glass decreases
the volume of gases released but not the quantity of particulate
matter emitted. As a result, the concentration of particulate matter
in the exhaust gas stream would be increased even though the total
mass emitted remained the same. Even if a concentration standard were
corrected to a specified oxygen content in the gas stream, this
penalizing effect of the concentration would not be overcome.
Primary disadvantages of mass standards, as compared to
concentration standards, are that their enforcement is more costly and
2-64
-------
that they require more numerous calculations and increase the opportunities
for error. Determining mass emissions requires the development of a
material balance on process data concerning the operation of the
plant, whether it be input flow rates or production flow rates.
Development of this balance depends on the availability and reliability
of production figures supplied by the plant. Gathering of these data
increases the testing or monitoring necessary, the time involved, and
consequently, the costs. Manipulation of these data increases the
number of calculations necessary, e.g., the conversion of volumetric
flow rates to mass flow rates, thus compounding errors inherent in the
data and increasing the degree of inaccuracy.
As explained in the preamble to the proposed standards, even
though concentration standards involve lower resource requirements for
testing than mass standards, mass standards are more suitable for
regulation of particulate emissions from glass melting furnaces because
of their flexibility to accommodate process improvements and their
direct relationship to the quantity of particulate emitted to the
atmosphere. These advantages outweigh the drawbacks associated with
creating and manipulating a data base. Consequently, mass standards
are selected as the format for expressing standards of performance for
glass melting furnaces.
Standards based on limiting the emissions from glass melting
furnaces in units of grams of particulate per kilogram of glass produced
were selected rather than a mass of particulate collected per unit of
time in order to take into account the vast range of production rates
in the glass manufacturing industry. In order to develop cohesive
emission standards to be applied equitably to furnaces of all sizes
within the particular categories of glass production, standards based
on emission mass per unit of mass production were selected. Another
factor considered is the energy conservation attributable to mass
standards as opposed to standards based on exhaust gas flows and
volumes. Basing a standard on other than the relationship between
emissions and production rate could possibly encourage the use of
additional energy to manipulate the test results.
2-65
-------
The promulgated standards express allowable particulate emissions
in grams of participates per kilogram of glass pulled. While emissions
data referring to raw material input as well as data referring to
glass pulled were used in the development of the standards, an examination
of the several sectors of the glass manufacturing industry indicated
that an emission rate based on quantity of glass pulled would be more
typical of industry measurment practices. Accordingly, the mass of
glass pulled is used as the denominator in the standards. Raw material
input data could possibly be employed to aid in the estimate of glass
pulled from a furnace if a quantitative relationship between raw
material input and glass pulled were developed following good engi-
neering methods. The better and most used means of calculation of the
production rate is the actual pull rate of the furnace, calculated by
taking representative samples and analyzing, at regular, intervals,
glass as it is produced. Refer to Appendix C of this document.
One commenter suggested that improved installation and burner
systems will result in more efficient furnaces and less unnamed emis-
sions. It was further suggested that the emission reductions resulting
from these installation and burner improvements will more than offset
the emissions created by any new furnaces that will be put into operation.
To date, no such technology has been demonstrated and no data have
been submitted that indicate that these technological advancements are
readily forthcoming for the control of particulates. Standards of
performance are based on demonstrated means of control. This being
the case, the suggested installation and burner system advancements,
have not been considered as the basis for these standards of performance.
Container Glass. Several commenters suggested that the uncontrolled
particulate matter emission rate for container glass should be 1.5 Ibs/ton
or 1.12 Ibs/ton, not 2.5 Ibs/ton. These commenters stated that the
emission tests used by EPA in selecting its uncontrolled emission rate
were not representative of the industry's emissions as they are today.
The uncontrolled emission rate used in the analysis of the impacts
of the proposed standard for container glass was based on emission
tests conducted by an. EPA contractor covering a wide spectrum of the
glass manufacturing industry. These results were used in response to
2-66
-------
comments made by industry at the National Air Pollution Control Techniques
Advisory Committee (NAPCTAC) meeting. It was in response to those
comments, expressing concern with the basis of the uncontrolled emission
rate used prior to NAPCTAC (1.5 Ibs/ton), that the rate for analyzing
the impacts of the proposed standard was changed to 2.5 Ibs/ton.
Some industry representatives suggested that the uncontrolled
emission rate for container glass is as low as 1.12 Ibs/ton. This
figure is based on the utilization of process modifications and are
not considered to be representative of the container glass industry on
the whole. Refer to the Emission Control Technology section of this
document for further details on this issue. As this suggested rate is
not presently prevalent throughout the industry, it was not used to
develop the standard's impacts.
Following the proposal of the standard, however, industry submitted
updated test results that show the uncontrolled emission rate for
container glass production to be more properly represented by 1.5 Ibs/ton.
These tests were conducted on a number of different container glass
furnaces and reflect emissions from these types of furances as they
presently operate. Even though the 2.5 Ibs/ton rate used for the
proposed standard's evaluation is realistically possible, the uncon-
trolled emission rate of 1.5 Ibs/ton was selected as the rate most
representative of the industry as it exists today.
The impacts associated with this change in uncontrolled emission
rate from 2.5 to 1.5 Ibs/ton are minimal and have not affected the
decision to regulate this category of the industry. However, the
re-evaluation of the impacts is discussed in the Environmental, Economic,
and Energy Impact sections of this document.
Wool Fiberglass. It was one commenter's contention that the
uncontrolled emission rate for the wool fiberglass category would be
more properly represented by a rate of from 22.0 Ibs/ton to 30,0 Ibs/ton.
Other comments have been received that suggest the uncontrolled emission
rate of 10.0 Ibs/ton, used in the development of the impacts attributed
to this standard of performance, was accurate for the wool fiberglass
industry. "
2-67
-------
The source tests used to determine the uncontrolled emission rate
and used to develop the impacts of the standard were conducted at
representative plants and were performed in accordance with EPA Method 5
and LAAPCD's source test method. Uncontrolled emission rates are used
solely to develop environmental, economic, and energy impacts associated
with the promulgation of these standards. A review of test data shows
that the uncontrolled emission rate ranges from less than 10.0 Ibs/ton
to greater than 20.0 Ibs/ton. However, the uncontrolled emission rate
representative of new wool fiberglass plants is best indicated by
10 Ibs/ton. EPA believes, after reviewing the test data in its possession,
that the 10 Ibs/ton figure reasonably reflects the uncontrolled emission
rate for the wool fiberglass industry as it presently operates.
Flat Glass. Some commenters questioned the uncontrolled particulate
matter emission rate for flat glass manufacturing plants of 3.0 pounds
per ton of glass produced (3.0 Ibs/ton). These commenters considered
the 3.0 Ibs/ton rate large, especially in light of industry technological
advancements, such as process modifications, that have evolved in the
industry over the past several years. One commenter explained that
3.0 Ibs/ton may have been representative of the uncontrolled emission
rate for the flat glass manufacturing industry five to seven years
ago.
EPA selected 3.0 Ibs/ton as a conservative estimate based on
information received concerning flat glass manufacturing plants.
Historically, flat glass manufacturing plant uncontrolled particulate
emission rates have been as high as 4.0 Ibs/ton or more. However, in
the past five to seven years, the industry has reduced its uncontrolled
emissions.
Industry commenters suggested that the industry-wide uncontrolled
emission rate should be in the vicinity of 1.12 Ibs/ton. This figure
was selected by industry as a result of the industry's survey of the
four major flat glass manufacturers in this country. However, review
of available data, which includes values greater than 2.0 Ibs/ton,
several values greater than 1.5 Ibs/ton and a few values less than
1.0 Ib/ton, indicates that the 1.12 Ibs/ton emission rate is a rate
2-68
-------
more typical of flat glass melting furnaces that might have the
capability to utilize advanced process modification techniques.
However, as discussed in detail in the Emission Control Technology
section of this document, process modifications are not readily available
to the industry on the whole and are not considered representative.
It is for this reason that the 1.12 Ibs/ton figure was not used to
evaluate the flat glass industry's regulatory impacts. As a result of
the comments mentioned and test data presently available, a decision
was made to evaluate the .impacts of the flat glass standard based on
an uncontrolled emission rate of 2.0 Ibs/ton. A re-evaluation of the
impacts is discussed in the Environmental, Economic, and Energy Impact
sections of this document.
2.5 ENVIRONMENTAL IMPACT
One commenter stated that there has been no experience to his
knowledge in the recycling of particulates collected from glass furnace
exhaust gases. Depending on the category of glass production, the
collected particulate may be recycled as a raw material. In certain
cases, recycling collected particulate requires changes in the batch
formulation. In other cases, chemical constituents of the collected
particulate limit the recycling of the collected particulate as a raw
material. For the categories that follow this practice of recycling,
the solid waste disposal impact will be essentially zero.
Industry representatives were concerned that certain emissions,
such as fluoride, boron, and lead, may adversely affect the environment
if landfilled in the form of collected particulate. No data were
submitted nor does EPA have data to substantiate the claims of adverse
environmental effects resulting from the landfill ing of particulates
containing these elements. However, should these landfilled participates,
at a later date, be determined to adversely affect the public health
and welfare, these landfill ing operations could be covered by the
Resource Conservation and Recovery Act (RCRA) [42 U.S.C. 6901, ejt
seq.].
One commenter suggested that solid waste disposal of the collected
particulates could possibly endanger the public health or welfare.
Industry has suggested that the disposal of particulates might interfere
2-69
-------
with drinking water or stream quality standards. Less than 2 Mg
(2.2 tons) of participate will be collected for every 1,000 Mg
(1,102 tons) of glass produced. There is no indication that landfill ing,
the commonly practiced method of solid waste disposal, will create
such a problem. As landfill operations are subject to State regulation
and the particulates collected as a result of the promulgation of
these standards do not differ chemically from the material collected
under a typical SIP regulation, there is a minimal adverse impact on
the environment. Therefore, current practices in landfill ing are
expected to continue throughout the industry and the waste impact of
these standards is considered to be minimal.
One commenter suggested that the proposed standards of performance
were not based on EPA's ambient air analysis. The establishment of
standards of performance for new, modified, or reconstructed stationary
sources is based on the best technological system of continuous emission
reduction which (taking into consideration the cost of achieving such
emission reduction, any nonair quality health and environmental impact
and energy requirements) the EPA determines has been adequately
demonstrated. Standards of performance are based on air quality
concerns, but additionally require that the best system of demonstrated
continuous emission reduction be applied. Standards of performance
are not based on ambient air analyses resulting from dispersion modeling.
The determination of significance is also not made based on the
annual average contributions of particulate matter per plant as is
done for PSD purposes. The issue of significance is addressed in the
Need For Standards section (2.1) of this document.
One commenter claimed that the Industrial Source (ISCST) Model
was neither EPA approved nor available to the public. The commenter
further implied that the Single Source (CRSTER) Model probably should
have been used and questioned the addition of downwash as an option.
The report on the dispersion modeling performed by H.E. Cramer
Company, Incorporated dated August 1978 clearly states:
"For the model features used in this study the ISCST
Program corresponds to the Single Source (CRSTER) Model,
modified to incorporate the Huber (1977) procedures for
quantifying the effects of aerodynamic downwash on plume
dispersion."
2-70
-------
Thus, the ISCST Program is compatible with the CRSTER Model which is
both approved and available.
Downwash is calculated in order to show the effects of air currents,
affected by stack heights, geography, and structures that influence
the plume emanating from the glass melting furnace stack. The addition
of downwash as an option also has merit in that the Huber procedures
have been technically reviewed and have appeared in papers presented
to the American Meteorological Society and the Air Pollution Control
Association.
The commenter also criticized the fact that downwash was.considered
in the analysis. In reviewing the August 1978 dispersion study it has
been found that the procedures used to determine whether downwash
should be considered are consistent with state-of-the-art downwash
calculations. This procedure involves taking the ratio of the sum of
the physical stack height and momentum plume rise to the building
height and comparing the value to 2.5. Ratios less than 2.5 have been
shown, generally, to indicate possible downwash. A stack height of
21 meters (70 feet) was used in thi.s analysis. Based on the above
ratio test, downwash was included. If 30 meter (100 foot) stacks had
been included in the analysis as the commenter suggested, downwash
still should have been included in the analysis for most cases.
Hence, the inclusion of downwash in the glass manufacturing analysis
was justified.
One commenter was concerned about the possible national health
and welfare effect of the plumes of charged particles emitted from
ESPs. Of specific concern was the possible influence these particles
will have on the climate in the immediate vicinity of the regulated
glass manufacturing plants and on the climate on a global basis.
The comment specifically referred to a preprint of a report
presented at a technical meeting by the representatives of the National
Oceanic and Atmospheric Administration dealing with electric field
measurements in coal-fired power plant plumes. However, the report
itself came to no conclusions as to global climatic effects. And,
electric field measurements indicated similar conditions downwind from
wet scrubbers. At present, the actual phenomenon is not clearly
2-71
-------
understood and, as recommended in the report, requires further review.
Thus, we have no reason to believe that this phenomenon would result
in any significant impact on the environment.
Container Glass. The promulgated standard will reduce particulate
emissions from a new uncontrolled container glass melting furnace from
about 58 Mg/year (6* tons/year) to about 8 Mg/year (9 tons/year).
This is a reduction of about 86 percent of the uncontrolled emissions.
Meeting a typical SIP, however, would reduce paniculate emissions
from a new uncontrolled furnace by about 21 Mg/year (23 tons/year).
The promulgated standard will result in a reduction in the level of
emissions achieved under a typical SIP by about 29 Mg/year (32 tons/year).
Twenty-five new container glass melting furnaces are expected to
be built by 1984. The total particulate emissions expected to be
emitted by 1984 for the container glass manufacturing plants subject
to this standard is 1,600 tons/year. Meeting a typical SIP would
reduce particulate emissions from 25 new container glass furnaces to
about 1,025 tons/year. Achievement of the standard will reduce the
particulate emissions for these 25 container glass furnaces to about
225 tons/year.
Pressed and Blown Glass (Borosilicate). The promulgated standard
will reduce particulate emissions from a new uncontrolled pressed and
blown (borosilicate) glass melting furnace designed to operate at
100 tons/day from about 156 Mg/year (172 tons/year) to about 15 Mg/year
(17 tons/year). This is a reduction of about 90 percent of the uncon-
trolled emissions. Meeting a typical SIP, however, would reduce
particulate emissions from a new uncontrolled furnace by about 130 Mg/year
(143 tons/year). The promulgated standard will result in a reduction
in the level of emissions achieved under a typical SIP by about 11 Mg/year
(12 tons/year). This reduction in emissions will result in a reduction
of ambient air concentrations of particulate matter in the vicinity of
new pressed and blown (borosilicate) glass manufacturing plants.
Two new pressed and blown (borosilicate) glass melting furnaces
designed to operate at 100 tons/day are expected to be built by 1984.
The total particulate emissions expected to be emitted by 1984 for the
pressed and blown .(borosilicate) glass manufacturing plants designed
2-72
-------
to operate at 100 tons/day subject to this standard is 344 tons/year.
Meeting a typical SIP would reduce particulate emissions from the two
pressed and blown (borosilicate) glass melting furnaces operating at
100 tons/day expected to be subject to this standard to about
58 tons/year. Achievement of the standard will reduce the particulate
emissions for these two affected facilities to about 34 tons/year.
The promulgated standard will reduce particulate emissions from a
new uncontrolled pressed and blown (borosilicate) glass melting furnace
designed to operate at 50 tons/day from about 78 Mg/year (86 tons/year)
to about 8.2 Mg/year (9 tons/year). This is a reduction of about
89 percent of the uncontrolled emissions. Meeting a typical SIP,
however, would reduce particulate emissions from a new uncontrolled
furnace by about 55 Mg/year (61 tons/year). The promulgated standard
will result in a reduction in the level of emissions achieved under a
typical SIP by about 19 Mg/year (21 tons/year). This reduction in
emissions will result in a reduction of ambient air concentrations of
particulate matter in the vicinity of new pressed and blown (borosilicate)
glass manufacturing plants.
No new pressed and blown (borosilicate) glass melting furnaces
designed to operate at 50 tons/day are expected to be built by 1984.
The total particulate emissions expected to be emitted by 1984 for the
pressed and blown, (borosilicate) glass manufacturing plants designed
to operate at 50 tons/day subject to this standard is therefore zero.
However, meeting a typical SIP would reduce uncontrolled particulate
emissions (86 tons/yr) from one pressed and blown (borosilicate) glass
melting furnace designed to operate at 50 tons/day expected to be
subject to this standard to about 25 tons/year. Achievement of the
standard will reduce the particulate emissions for an affected facility
to about 9 tons/year.
Pressed and Blown (Soda-lime and Lead). The promulgated standard
will reduce particulate emissions from a new uncontrolled pressed and
blown (soda-lime and lead) glass melting furnace designed to operate
at 100 tons/day from about 39 Mg/year (43 tons/year) to about 2.7 Mg/year
(3 tons/year). This is a reduction of about 93 percent of the uncon-
trolled emissions. Meeting a typical SIP, however, would reduce
2-73
-------
particulate emissions from a new uncontrolled furnace by about 13 Mg/year
(14 tons/year). The promulgated standard will result in a reduction
in the level of emissions achieved under a typical SIP by about 24 Mg/year
(26 tons/year). This reduction in emissions will result in a reduction
of ambient air concentrations of particulate matter in the vicinity of
new pressed and blown (soda-lime and lead) glass manufacturing plants.
Six new pressed and blown (soda-lime and lead) glass melting
furnaces designed to operate at 100 tons/day are expected to be built
by 1984. The total particulate emissions expected to be emitted by
1984 for the pressed and blown (soda-lime and lead) glass manufacturing,
plants designed to operate at 100 tons/day subject to this standard is
258 tons/year. Meeting a typical SIP would reduce particulate emissions
from the six pressed and blown (soda-lime and lead) glass melting
furnaces designed to operate at 100 tons/day expected to be subject to
this standard to about 174 tons/day. Achievement of the standard will
reduce the particulate emissions from these six affected facilities to
about 18 tons/year.
The promulgated standard will reduce particulate emissions from a
new uncontrolled pressed and blown (soda-«lime and lead) glass melting
furnace designed to operate at 50 tons/day from about 19 Mg/year
(21 tons/year) to about 1.8 Mg/year (2 tons/year). This is a reduction
of about 90 percent of the uncontrolled emissions. As the typical SIP
limitation exceeds the uncontrolled emission rate, the total emission
reduction is attributable to the standard. This reduction in emissions
will result in a reduction of ambient air concentrations of particulate
matter in the vicinity of new pressed and blown (soda-lime and lead)
glass manufacturing plants.
Four new pressed and blown (soda-lime and lead) glass melting
furnaces designed to operate at 50 tons/day are expected to be built
by 1984. The total particulate emissions expected to be emitted by
1984 for the pressed and blown (soda-lime and lead) glass manufacturing
plants designed to operate at 50 tons/day subject to this standard is
84 tons/year. Achievement of the standard will reduce the particulate
emissions for these four affected facilities to about 8 tons/year.
2-74
-------
Pressed and Blown Glass (Other-Than Borosilicate, Soda-Lime, and
Lead). The promulgated standard will reduce particulate emissions
from a new uncontrolled pressed and blown (other-than borosilicate,
soda-lime and lead) glass melting furnace designed to operate at
100 tons/day from about 156 Mg/yr (172 tons/yr) to about 8.2 Mg/yr
(9 tons/yr). This is a reduction of about 95 percent of the uncon-
trolled emissions. Meeting a typical SIP, however, would reduce
particulate emissions from a new uncontrolled furnace by about 130 Mg/yr
(143 tons/yr). The promulgated standard will result in a reduction in
the level of emissions achieved under a typical standard by about
18 Mg/yr (20 tons/yr). This reduction in emissions will result in a
reduction of ambient air concentrations of particulate matter in the
vicinity of new pressed and blown (other-than borosilicate, soda-lime
and lead) glass manufacturing plants.
No new pressed and blown (other-than borosilicate, soda-lime and
lead) glass melting furnaces designed to operate at 100 tons/day are
expected to be built by 1984. Therefore, there is no predicted
environmental impact associated with this subcategory of glass melting
furnace.
The promulgated standard will reduce particulate emissions from a
new uncontrolled pressed and blown (other-than borosilicate, soda-lime
and lead) glass melting furnace designed to operate at 50 tons/day
from about 78 Mg/yr (86 tons/yr) to about 3.6 Mg/yr (4 tons/yr). This
is a reduction of about 95 percent of the uncontrolled emissions.
Meeting a typical SIP, however, would reduce particulate emissions
from a new uncontrolled furnace by about 55 Mg/yr (61 tons/yr). The
promulgated standard will result in a reduction in the level of emissions
achieved under a typical standard by about 19 Mg/yr (21 tons/yr).
This reduction in emissions will result in a reduction of ambient air
concentrations of particulate matter in the vicinity of new pressed
and blown (other-than borosilicate, soda-lime and lead) glass
manufacturing plants.
One new pressed and blown (other-than borosilicate, soda-lime and
lead) glass melting furnace designed to operate at 50 tons/day is
expected to be built by 1984. The total particulate emissions expected
2-75
-------
to be emitted by 1984 for the pressed and blown (other-than borosilicate,
soda-lime and lead) glass manufacturing plant designed to operate at
50 tons/day subject to this standard is 86 tons/yr. Meeting a typical
SIP would reduce particulate emissions from the pressed and blown
(other-than borosilicate, soda-lime and lead) glass furnace operating
at 50 tons/day expected to be subject to this standard to about 25 tons/yr.
Achievement of the standard will reduce the particuTate emissions from
this affected facility to about 4 tons/yr. This reduction in emissions
will result in a reduction of ambient air concentrations of particulate
matter in the vicinity of new pressed and blown (other-than borosilicate,
soda-lime and lead) glass manufacturing plants.
Hool Fiberglass. The promulgated standard will reduce particulate
emissions from a new uncontrolled wool fiberglass melting furnace from
about 312 Mg/year (343 tons/year) to about 15 Mg/year (17 tons/year).
This is a reduction of about 95 percent of the uncontrolled emissions.
Meeting a typical SIP, however, would reduce particulate emissions
from a new uncontrolled furnace by about 278 Mg/year (306 tons/year).
The promulgated standard will result in a reduction in the level of
emissions achieved under a typical SIP by about 18 Mg/year (20 tons/year).
This reduction in emissions will result in a reduction of ambient air
concentrations of particulate matter in the vicinity of new wool
fiberglass manufacturing plants.
Six new wool fiberglass melting furnaces are expected to be built
by 1984. The total particulate emissions expected to be emitted by
1984 for the wool fiberglass manufacturing plants subject to this
standard is 2,058 tons/year. Meeting a typical SIP would reduce
particulate emissions from the six wool fiberglass furnaces expected
to be subject to this standard to about 222 tons/year. Achievement of
the standard will reduce the particulate emissions for these six
affected facilities to about 102 tons/year.
Flat Glass. The promulgated standard will reduce particulate
emissions from a new uncontrolled flat glass melting furnace from
about 218 Mg/year (240 tons/year) to about 49 Mg/year (54 tons/year).
This is a reduction of about 77 percent of the uncontrolled emissions.
Meeting a typical SIP, however, would reduce particulate emissions
2-76
-------
from an uncontrolled furnace by about 147 Mg/year (162 tons/year).
The promulgated standard will result in a reduction in the level of
emissions achieved under a typical SIP by about 22 Mg/year (24 tons/year).
This reduction in emissions will result in a reduction of ambient air
concentrations of particulate matter in the vicinity of new flat glass
manufacturing plants.
Four new flat glass melting furnaces are expected to be built by
1984. The total particulate emissions expected to be emitted by 1984
for the flat glass manufacturing plants subject to this standard is
960 tons/year. Meeting a typical SIP would reduce particulate emissions
from the four flat glass furnaces expected to be subject to this
standard to about 312 tons/year. Achievement of the- standard will
reduce the particulate emissions for these four affected facilities to
about 216 tons/year.
2.6 ECONOMIC IMPACT
Several commenters suggested that the cost-effectiveness of the
proposed standards, i.e., high costs per unit of pollutant removed,
does not warrant the promulgation of standards. These commenters
appear to claim that the cost of removing one kilogram of particulate
exceeds the benefits derived from its removal.
The cost and benefit to public health and welfare associated with
the reduction of air pollutant emissions is difficult, if not impossible,
to quantify. In general, it is much easier to quantify the cost of
emission reduction than to quantify the benefit. Thus, the cost is
usually the subject of much discussion whereas the benefit is not.
Given the inability to quantify both the cost and the benefit, it is
not possible to determine whether the cost of control exceeds the
benefits associated with control. Thus, it would be inappropriate to
make a regulatory decision based on a cost-benefit analysis.
Cost-effectiveness calculations are certainly useful in regulatory
analyses for choosing among competing regulatory alternatives which
achieve the same level of control. On the other hand, cost-effective-
ness has too many limitations to be used as the major decision-making
factor in setting standards of performance under Section 111. First,
the most cost-effective controls for a source are often at the low end
2-77
-------
of the spectrum of pollutant reduction where the cost per unit of
pollutant removed is typically low. The level of control associated
with standards of performance under Section 111 of the Clean Air Act
would typically not correspond to the most cost-effective control.
Second, it is not practical to identify a numerical criterion which
represents an upper limit in cost per unit of pollutant removed.
Technological differences among industries cause control costs for any
given pollutant to vary considerably. In the case of glass manufacturing,
this is illustrated by the fact that among several segments there are
considerable differences in cost per unit of pollutant removed. There
are also segments where little difference in costs between SIPs and
NSPS is evident, while in other segments there are distinct differences.
Third, the economic impact analysis employed in this instance used the
most costly controls to determine worst-case effects. The other less
costly alternatives that achieve equivalent control levels are also
available to the source.
In reaching the conclusion that the promulgated standards would
have no significant economic impact on the glass manufacturing industry,
other factors besides cost-effectiveness, were taken into consideration.
The costs associated with the achievement of these promulgated standards
were considered in the context of the cost structure of the industry
by means of an economic analysis including, where necessary, a discounted
cash flow model. Upon making these considerations, the economic
impacts of the proposed standards were determined to be reasonable.
These impacts are still considered reasonable for the promulgated
standards of performance.
Commenters claimed that the cost of particulate control should be
totally attributed to the standards of performance. The cost of the
standards of performance are analyzed based on the assumption that SIP
regulations require control where uncontrolled emissions are greater
than SIP allowed emissions. As discussed in the General Issues section
of this document, SIPs similar to New Jersey's and Pennsylvania's are
considered typical of what glass manufacturing plants have been required
to comply with. These SIPs require, in most cases, emission reduction
2-78
-------
through add-on control technology. It would be unrealistic not to
delete the cost that a new plant would incur without the establishment
of standards of performance. Therefore, it is realistic to estimate
the added or incremental cost that would be incurred if a standard of
performance control level greater than that required by SIP is
established.
These commenters argued that uncontrolled emission rates and SIP
regulations do not necessitate the use of add-on controls as indicated
in the economic analysis. However, as explained in the Emission
Control Technology section of this document, typical glass melting
furnaces, with the exception of the 50 tons/day pressed and blown
(soda-lime and lead) glass furnaces, located in States with
representative SIP regulations require add-on controls.
Other commenters stated that process modifications were more
cost-effective than the controls analyzed and that process modifica-
tions should be the basis upon which the standards are set. (Refer to
the discussion on process modifications in the Emission Control
Technology section of this document.) Process modifications are not
precluded from use as a method of compliance with a NSPS. If they are
indeed more cost-effective than those regulatory alternatives analyzed
then those sources that can utilize process modifications will
experience lesser impacts than those estimated in the regulatory
analysis.
One commenter suggested that more weight should be given to
economic costs than a mere determination that the industry can bear
the costs of a control technique and still survive. The comment
appears to refer to the focus of the discounted cash flow (DCF) technique
on whether or not controls will inhibit the building of new glass
manufacturing plants. The use of the DCF was limited to special cases
where survival might have been in question. However, for most of the
glass manufacturing categories the DCF analysis was not needed because
the impact of the standards does not approach the point where survival
might be in question. For example, the industry average price increase
is less than 1 percent, a price increase that would not deter the
construction of any new plants by itself.
2-79
-------
A review of the economic impact analysis may be helpful. Figure 1
is a flow chart description of the decision-making process used in
evaluating the Economic Impact Assessment of the New Source Performance
Standards (NSPS) for glass manufacturing plants. The first step
called for a scaling of certain model plant control costs to obtain
control costs for other plant sizes suggested by industry as not being
covered by the model plant sizes. The pext step in the calculations
?*,* " * « r
was to determine the percent of plant revenue represented by the
*,*.«*: - ' - - -t r - P - , '
annualized control cost for each model and the other sized plants.
The control costs selected for these calculations were for the highest-
. 7 - "i « -J #**»!' «* '* w »* * ' "" '
cost for a regulatory alternative. The purpose of this step was to
determine the price increase potential of the controls if costs are
passed on.to the consumer, or the profit reduction that would occur
with np Pfgcfqct price increases. Revenue was determined on the basis
of the capacity, capacity utilization rate, and price per unit of
output determined, for that particular industry category.
A decision point followed those calculations. If the percentage
of cost to plant reveriue was less than 1 percent, the cost was not
considered to have a significant impact. For those plants incurring a
control cpst greater than 1 percent, a further test was applied to
determine if the decrease in return on investment (ROI) after pollution
controls was greater than 10 percent. If the decrease was less than
10 percent, then that size plant was estimated not to experience a
significant economic impact. If the percent decrease was greater than
10, that size plant was subjected to a further set of calculations to
determine if the impact was significant. It must be noted that this
10 percent limit was not considered as the point of significant impact
as was implied by some commenters. Ten percent was considered as a
screening criterion to determine which plants would receive further
analysis by a more sophisticated financial technique.
For the remaining plants a discounted cash flow (DCF) analysis
(refer to Appendix B) was performed to determine if significant economic
impacts were present for these glass manufacturing plants. If the net
present value (NPV), as determined from the discounted cash flow
2-80
-------
SCALE CONTROL COSTS
TO REVISED MODEL
SIZE PLANTS (APPLI-
CABLE TO 3 MODELS)
CALCULATE PERCENT
OF MODEL SIZE PLANT
REVENUE REPRESENTED
BY ANNUALIZED COST
IMPACT
NOT
SIGNIFICANT
IS
PERCENT
GREATER
THAN
1?
YES
CALCULATE PERCENT
CHANGE IN ROI FOR
MODEL SIZE PLANTS
IMPACT
NOT
SIGNIFICANT
IS
PERCENT
GREATER
THAN
-10?
YES
PERFORM DISCOUNTED
CASH FLOW ANALYSIS
IMPACT
NOT
SIGNIFICANT
IS
NPV WITH
CONTROLS
WAS
NPV
WITHOUT
CONTROLS
IMPACT
NOT
SIGNIFICANT
IMPACT
SIGNIFICANT
NOTE: ROI-RETURN ON INVESTMENT
NPVNET PRESENT VALUE
DETERMINED FROM DISCOUNTED CASH FLOW ANALYSIS
FIGURE 1. JMPACT DECISION FLOW CHART
2-81
-------
calculations, was greater than zero then the controls were estimated
to not have a significant impact. If the NPV was less than zero, and
it would have been greater than zero in the absence of NSPS, the plant
was considered to be significantly impacted. The minimum acceptable
rate of return for the industry was used to discount future cash flows
for computing NPV.
By following this decision process, two plants were found to be
in the category requiring a DCF analysis to determine if significant
impact was present. (Refer to BID, Volume I and Appendix B of this
document.) The 50-tons per day handmade consumerware plant was subjected
to a DCF analysis as was the 500 tons per day container glass manufac-
turing plant. The results of the DCF analyses for each of those
plants indicated that the incremental controls of the most costly NSPS
would not produce a NPV less than zero. It was, therefore, concluded
that the construction of new facilities of these types would not be
inhibited by the glass manufacturing NSPS.
Thus, the impact analysis which has been performed, does indeed
go beyond the mere determination that the industry can survive. The
analysis yields figures which indicate the specific (quantitative)
impact on profits, on ROI, and on price increases. It should be noted
that, for most of the glass manufacturing categories the test for
survival fell far short of the survival point (yielding less than the
minimum acceptable rate of return), indicating no significant impact.
A commenter explained that because the starting point of the
economic impact analysis is the "minimum" acceptable return, and
additional expenses are then deducted, by definition the result will
come out below the minimum acceptable rate of return. The comment
appears to refer to the DCF analysis where the pre-NSPS control profit
rate was selected as the minimum acceptable rate to the industry.
Thus, any deterioration in that rate would be unacceptable.
The internal logic of the comment is correct. Given any return
(whether or not "minimum" acceptable), the imposition of additional
expenses will, by definition, further reduce that return. However, in
the analysis, rate of return was not used as the decisive parameter
for determining impact significance. Rate of return was used only as
2-82
-------
one ingredient in the total analysis. Specifically, it was used as
the discount rate of projected cash flows of the new source to deter-
mine if the NPV of such cash flows, after discount, exceeded the
initial investment.
However, this comment caused other parts of the economic analysis
to be re-examined. This led to the conclusion that the same minimum
acceptable return under discussion can, in fact, be considered to be
inappropriate for one of the procedures that was followed. That
procedure also involved using the-minimum threshold rate to calculate
yearly pre-NSPS profit levels in the DCF analysis. This may have been
due to a misunderstanding by EPA as to the intended use of this minimum
threshold rate when supplied by the Glass Packaging Institute.
Taking the analysis performed by Peat, Warwick & Mitchell (PMM)
in their report (docket entry OAQPS 77/1-IV-D-19), instead of the BID,
Volume I, Chapter 8 calculation on this point, it is noted that the
ultimate conclusion remains unchanged. That is, the change in rate of
return due to the incremental new source performance standards' costs
is not significant enough for the project to be rejected. The PMM
calculations even overestimate the impact shown in EPA's regulatory
analysis because both SIP and incremental NSPS costs were used in
PMM's calculations.
The Peat, Marwick & Mitchell DCF analysis was similar in other
respects to EPA's DCF analysis with the following exceptions: 1) the
regulatory analysis' use of a 15 percent pre-tax profit was reduced to
8 percent, 2) the full amount of interest was added back as cash flow
(as opposed to interest times the tax rate which was used in EPA's
analysis and which EPA still maintains is more proper), 3) the use of
full control costs to meet the NSPS as opposed to EPA's use of incremental
cost between the SIP and the NSPS levels. Even under this analysis,
the economic impact of the standards is reasonable.
It was commented that the economic analysis indicates that the
impact of the new source performance standards on a new grassroots
plant would be to change the projected rate of return from 14 percent
to 12 percent. The commenter explained that this is a 15 percent
change in the expected rate of return, it is a significant change even
2-83
-------
by the criteria used in the Economic Impact Analysis (Chapter 8 of the
BID, Volume I), and is hardly an incentive to construct. The commenter
also stated that the total air pollution control cost should be used
to calculate the change in the projected rate of return.
The comment appears to assume that a 10 percent change in rate of
return was considered to be the level of significant impact. As
explained earlier, the economic impact of a new source performance
standard is the impact of going beyond SIP control levels to the more
stringent NSPS levels. EPA's economic impact analysis uses the
incremental cost in going from one stage to the other, recognizing .
that in the absence of a NSPS the new source would have to attain SIP
control level.
If the full-cost assumption were to be followed, the commenter's
figures would be correct. On the other hand, using the assumption
that only incremental new source performance standard costs are relevant,
calculations show a change from 14 percent to 13.1 percent in the
projected rate of return period. This is a decrease of 7 percent, as
opposed to the 15 percent decrease calculated by the commenter.
However, it is important to note that the 10 percent limit,
referred to in the BID and as explained earlier, was not a figure
beyond which the development of new source performance standards would
be ruled out, but was a figure beyond which other analyses would be
brought into play. As a result of the DCF anaylsis, it was concluded
that the construction of new facilities for all glass types would not
be inhibited by the glass manufacturing plants NSPS.
Two commenters made comments with respect to the 10 percent rate
of interest used in the analysis. It was suggested that with the
prime rate at 13 percent as of September 1979 and little indication of
a downturn in the short or medium term, use of a 10 percent interest
rate understates cost.
When the annual!zed cost calculations were made, the prime interest
rate was less than the 10 percent figure used. The 10 percent figure
used was not based on an attempt to track a highly variable prime (The
prime rate has risen from 10 to 20 percent and fallen to 11 percent in
the first 6 months of 1980). It is a conservative figure generally
2-84
-------
employed in analyses to reflect cost of capital under applicable
financing instruments. It is adjusted periodically to reflect any
persistent trend.
Pollution control expenditures are generally not financed at
prime rates, but at lower rates reflecting the longer loan period and
financing instruments available. For example, even at the recent
unusually high prime rate (15 1/4 percent on February 4, 1980) an
A-grade (Moody's) long-term industrial bond yield was 12.21 percent.
Yields on Aa and Aaa rated bonds were lower.
Annualized cost calculations in the glass manufacturing industry
are not highly sensitive to interest rate changes. For example, if
the 10 percent rate used was adjusted to the 13 percent rate cited in
the comment, an increase of 30 percent, the estimated annualized cost
would increase by 7.7 percent from $351,000 annually to $378,000
annually for a fabric filter designed to achieve the standard for
container glass. This would not appreciably affect findings in this
industry with regard to new plant construction. The highest potential
price increase of any of the glass segments studies would only change
by 0.1 percent.
Further, if current interest rates were used in an analysis other
financial and economic data used in the DCF analysis, would also
require updating.. These other financial and economic data used in the
analysis were all tied to conditions at a point in time. If the
interest rates are updated then these other data must be updated.
Once this would be done, the DCF analysis would most likely, indicate
the same result. The purpose in selecting one point in time to base
the analysis around is to provide a common basis for all the data.
A commenter claimed that EPA cannot analyze the effect of standards
on a number of new plants constructed in a time period by simply
estimating a rate of return expected from a typical plant, under price
and cost assumptions like the ones used in the Background Information
Document, Volume I. The comment appears to question the ability to
extrapolate the results of a model plant analysis, using one set of
data, on the industry.
2-85
-------
Using typical or model plant parameters is the most prevalent
technique utilized in economic impact studies involving standards of
performance (and other regulations) developed by or for EPA, and is
widely accepted by various industrial segments previously affected by
EPA regulations. Model plant analysis is a technique which is capable
of addressing regulatory impacts on projected new plants which typically
face financial parameters, including costs, which are different from
those faced by existing plants.
A model plant analysis begins with an assessment of the range of
plant sizes expected to be built in the future. This is done to make
certain that the analysis is reasonably representative of what should
occur once the standards are put into effect. Typically, this range
is different from the range of existing plant sizes. Once a future
plant size range is determined, a model plant is selected at the low
end of this range for economic analysis. This is done to ensure a
conservative analytical framework, due to economies of- scale which
typically prevail with respect to production and pollution control
costs.
A commenter suggested that reasons why a company decides to build
a new plant should be addressed before the effects of the regulation
on new plant construction are assessed. This commenter suggested that
the approach used in the BID, Volume I, would lead one to conclude
that either all plants will be built or no plants will be built. The
comment appears to, again, refer to the limitations of using one model
plant to make generalizations to an industry segment. The model plant
analysis used in the economic analysis has general application to the
categories considered since it utilizes worst-case conditions to
evaluate impacts.
The comment also appears to suggest that other criteria for plant
construction should be used in the analysis. Concerning other criteria
for building new plants, EPA concurs that other criteria such as
maintenance of market share or technological advances may enter into
deciding whether a new plant would be built or not. However, these
criteria are not subject to quantification and the application of a
specific industry-wide criterion. Furthermore, regardless of the
2-86
-------
"reason" for construction, new plant construction limited by other
criteria, would not allow a review of the impacts of the standards.
Use of the model plant approach allows a review of the effects of
the regulation on new plant construction. In order to implement this
approach a large amount of data must be compiled and analyzed. For
the. development of these standards the data were compiled and analyzed
and, when finalized, were used to efficiently develop model plants and
compare the impacts these standards will present environmentally,
economically, and energy-wise. The model plant approach is a sufficiently
informative approach and efficiently determines the impacts of the
standards on the construction of new plants.
It was suggested by a commenter that an estimate of 1982 shipments
cannot be used directly to estimate the expected number of new glass
manufacturing plants to be constructed by 1982. In performing the
analysis, it was assumed that plant operation for the newly constructed
glass plants would be at 100 percent capacity, rather than at some
lesser rate. Reduction in this capacity utilization rate would indeed
require construction of additional new plants, with a consequent
increase in industry costs.
Experience in the industry shows a general production rate capacity
utilization of 90 percent or more (Background Information Document,
Volume I, Section 8.1.3.5 - Supply Considerations). By adjusting the
calculations to take this 90 percent production rate into account, the
total industry control costs would increase by approximately 11 percent.
However, the assumption that new plants would be operated at 100 percent
of capacity is not unreasonable, particularly in an industry where
plants are operated at 90 percent or more and in some cases greater
than 100 percent of design capacity. Therefore, 100 percent is
considered to be a correct capacity utilization figure. This does not
significantly affect the economic impacts associated with the
promulgation of these standards.
In addition, as explained below, during preparation of the
Background Information Document, Volume 1, two estimations of the
number of new plants were made. In order to maximize industry wide
cost estimates, the largest estimation of the number of plants was
2-87
-------
chosen as the basis for estimating industry cost. Adjusting the
calculations by 11 percent would not be appropriate because the maximum .
industry-wide costs have been calculated.
One commenter stated that the estimate (industry-wide annualized
costs) of $5 million a year for new sources to meet the standards is
only a fraction of the total, once the entire industry is covered by
these standards (i.e.* those built over the next 5 years). The commenter
added that once all industry capacity is subject to the standards, the
total cost would be $33.9 million per year. This would occur after
all existing industry would become subject to the standrds. This is
likely to take many years.
Concerning the number of years, the 5-year period used is consistent
with other considerations. Performance standards are reviewed on a
4-year cycle * the additional year allowing for a certain amount of
rulemaking time. Secondly, a "period" is specified in Executive
Order 12044 [Section 4(f)] for applying a dollar criteria to
industry-wide costs for each year within the total time span. To
predict impacts beyond five years would not tend to change the validity
of the standards and their impacts. However, in order to be consistent
with Executive Order 12044, a 5-year period is necessary. Thus, in
the case of new source performance standards for glass manufacturing
plants, it is appropriate and reasonable that a fifth-year review be
used.
Concerning the $33.9 million/year, this figure does appear to be
correct, assuming that there is a point in the future at which all
present capacity has become subject to NSPS. Although not expected to
occur soon, this may eventuate through such events as the replacement
of existing plants with new plants. However, the $33.9 million/year
figure still does not affect the conclusion of the economic impact
analysis. First, the analysis of the impact on individual plants is
unaffected, regardless of whether the total costs for all plants are
taken to be $5 million or $33.9 million/year. Second, with an annualized
cost of control estimated to be $33.9 million, a proportionate increase
2-88
-------
in emission reduction would occur. Thus, because the impacts would
not differ in perspective, if the entire industry^were covered rather
than just the new plants built within five years, the $33.9 million/year
would not affect the conclusion of the economic impact.
Container Glass. One commenter said that the scaling conversions
from a 250 tons/day container glass plant to a 500 tons/day container
glass plant are not explained.
Use of the scaling conversions was needed to adjust capital and
annualized costs for a 250 tons/day to a 500 tons/day container glass
plant. This adjustment was made so that information supplied by the
Glass Packaging Institute could be used in the economic analysis.
This adjustment resulted in an overstatement of costs in the economic
impacts section of the Background Information Document, Volume I. The
scaling method used to estimate control costs for the 500 tons/day
plant was to less-than-double the incremental costs for the 250 tons/day
container glass plant. This method was predicated on the tendency for
economies of scale to prevail over control equipment. However, by
using the scaling formula utilized by EPA in Section 8.2 of the
Background Information Document, Volume I, for the 250 tons/day plant,
the incremental capital cost would be $120,000 less than that postulated
on a proportional basis for the 500 tons/day plant, whereas the
annualized cost would be $56,000 higher. These adjustments are required
after recomputing the impact on the rate of return. The rate of
return after controls would be reduced by about 0.1 percent.
One commenter suggested that erroneous tax assumptions, when
corrected.,, result in figures showing that new container glass plants
subject to the standards would not profitably repay the original
investment necessary to build them. There are two tax assumptions
about which questions have been raised by this comment. One involves
treatment of interest and the other involves the allocation of the
investment tax credit over different years.
The commenter asserted that interest in"the DCF analysis was
treated as a tax credit and not as a deduction, as interest is normally
treated. In fact, however, interest was deducted as an expense in
computing profits in the DCF analysis and then added back to cash
2-89
-------
flow. The reason for this add-back is to account for the tax shelter
of interest for borrowed money. The treatment of interest in this
manner follows the recommendations of well-known authors in the financial
field, including the ones cited by PMM in their comments.
The use of the investment tax credit in the DCF analysis was
improper, as suggested by the commenter. The investment tax credit
was incorrectly used in full in the first year calculation of the DCF.
Under the Internal Revenue Code, only one-half of the amount of taxes
payable in any one year may be used as the investment tax credit.
Unused portions of an investment tax credit in any one year can be
distributed over a certain number of subsequent years.
As a result of this error, the NPV of the tax credit is less than
it would have been had it been used fully in the first year. The ,
resultant correction of the DCF amount is approximately 0.1 percent.
This percentage is estimated by taking the increase in model plant
control costs attributable to this error ($56,000), recognizing that
it represents a $0.34/ton increase in control costs, and applying it
to the relationship supplied by the Glass Packaging Institute's comment.s
on EPA's Cash How Table (a $7.I/ton increase in control costs results
in a 2.0 percent change in R.O.I.). This discrepancy does not alter
the conclusion that the economic impact on the container industry is
not significant.
Commenters suggested that the EPA failed to consider important
potential economic impacts attributable to the proposed standard.
Specifically of concern were the effects of the standard on small
container glass manufacturers and the effect of competition-fw
substitute products.
In the initial stages of its analysis, EPA utilized a 250 tons/day
model plant. Based upon discussions with the Glass Packaging Institute,
EPA was persuaded to perform the economic analysis on a 500 tons/day
container glass plant, a production level twice that of the previous
plant being used in the analysis. EPA purposely selects small plants
for its analyses, not to establish a separate treatment for them, but
rather to maintain an across-the-board conservative treatment. Small
2-90
-------
plants are typically deprived of the economies of scale which are
available to large plants. Therefore, an analysis of small plants
will tend to state the costs incurred by the industry more conser-
vatively (i.e., higher than if a larger sized plant were to be used in
the analysis).
For container glass plants, a small plant at 250 tons/day was
analyzed and a 1.8 percent increase in the consumer price was estimated
to be the result of implementation of the standard. A larger plant
(500 tons/day) was analyzed and the increase in the consumer price was
estimated to be 1.3 percent. However, larger plants are more centralized
and therefore, would experience higher cost for transportation of raw
materials and finished products. In any case, the price increase for
either sized plant would not be prohibitive in this industry where the
price of consumer goods has increased by approximately 8 to 10 percent
per year over the past 10 years.
As for competition from .substitute products, this was taken into
account in the analyses. Part of the analysis was performed on the
assumption that in the short run there would be no price increase in
glass products due to the severity of competition from other products
not similarly impacted. This means that its current competitive
position relative to substitute products would be unaffected. The
effect of this cost absorption approach is to reduce the potential
rate of return of new plants. The use of the DCF analysis was to
determine if that reduction would cross the threshold at which new
plant construction would be. inhibited. The results of this analysis
indicated that this threshold had not been crossed.
A commenter suggested that the economic analysis should include
the effects on plants of various sizes and on the production of
containers for which largfe, centralized plants may be relatively less
practical due to prohibitive transportation costs.
EPA believes that its analysis does include the effects on plants
of various sizes. The economic impact analysis was conducted for the
smaller plants in the industry, based upon information supplied by the
Glass Packaging Institute. The analysis also takes into consideration
2-91
-------
the fact that larger plants would have greater economies of scale and
be in a better position to absorb pollution control costs than the
smaller plants. Thus, if the economic impact is not significant for
the small plants, it would not be significant for larger plants.
It was suggested by one commenter that prices in the glass
manufacturing industry, particularly the container segment, should not
be considered to be fixed for rate of return calculation.
Glass products are not commodities which have exhibited wide
price fluctuations in the past, in contrast with such commodities as
copper, hogs, or soybeans. Due to the acquisition and operation of
pollution controls required by-the standards, plant costs will inevitably
increase and,this will result in an upward pressure on prices. However,
as shown in the Background Information Document, Volume I's economic
analysis, the control costs attributable to the standards in certain
instances may not be passed through to the consumer in the form of
higher product prices. In these instances this would be due to the
price elasticity effects of competition from substitute products.
No indications of circumstances resulting in a downward price
trend, such as a major technological breakthrough or a significant
increase in the supply of substitute products or demand shifts, have
been found to exist or have been alleged to be forthcoming. Therefore,
under the steady price scenario, the DCF analysis uses fixed (constant)
prices which are consistent with the real (constant) dollar terms used
throughout the analysis. In light of the above, a steady price scenario
avoids underestimating the regulatory impact.
The incremental installed cost (cost in excess of a typical SIP
regulation cost) in January 1978 dollars associated with the standard
for controlling particulate emissions from a 225 Mg/day container
glass furnace will be about $700 thousand[..for an ESP and about
$1.2 million for a fabric filter. Incremental annualized costs
associated with the standard for a 225 Mg/day furnace will, be about
$200 thousand/year and, about $350 thousand/year for an ESP and a
fabric filter, respectively.
Incremental cumulative capital costs for the 25 new 225 Mg/day
container glass furnaces during the 1979 through 1984 period associated
2-92
-------
with the standard will be about $17 million if ESPs are used. Fifth-year
annualized costs for controlling the 25 new glass melting furnaces to
comply with the standard will be about $5 million/year.
Based on the use of control equipment with the highest annualized
cost (worst-case conditions), a price increase of about 1.8 percent
will be necessary to offset the cost of installing control equipment
on a 225 Mg/day container glass furnace to meet the emissions limit of
the standard.
Pressed and Blown Glass. It was suggested that the annualized
cost for solid waste disposal in the pressed and blown category was
calculated incorrectly. It was suggested to be $80.00/ton, not $5.30/ton.
The cost of $5.30/ton, presented in Section 8.2 of the Background
Information Document, Volume I, is based on the authoritative reference
numbers 10 and 11 as shown on pages 8-64 and 8-88. Recent studies
indicate that disposal costs in the range of $5.00/ton are typical,
with some costs as high as $40.00/ton having been reported. Although
it is possible that a cost of $80.00/ton could occur for a specific
facility, such a cost appears to be abnormal. Even if a disposal cost
of $80.00/ton were to occur, the total annualized control cost would
Increase only 6.0 percent from the total annualized control cost based
on the $5.30/ton cost used in Section 8.2 of the Background Information
Document, Volume I. This would still not inhibit the construction of
new facilities for the production of any type of glass. However, use
of the $5.00/ton is considered reasonable.
One commenter was of the opinion that the economic impacts outlined
in the proposed rules do not take hand glass plants into consideration.
It was explained that these plants are very small with low daily pull
Crates, and many of the plants are marginally surviving already. It
was this commenter's opinion that the proposed standard would be
disastrous for this sector of the industry.
Hand glass-plants, as discussed in the Emission Control Technology
section (2.2) of this document, have been taken into consideration in the
promulgation of these standards in the form of an exemption.
One commenter pointed out that the growth rate figures for the
pressed and blown glass categories varied within the Background
2-93
-------
Information Document, Volume I. It was noted that Chapter 7 used
3.5 percent and Chapter 8 used 4.0 percent as their respective growth
rates.
The use of these two percentages was a mistake. The use of
3.5 percent as an annual growth rate in Chapter 7 (Environmental
Impact) would favor industry in revealing less projected emissions
(new sources). The use of 4.0 percent as an annual growth rate in
Chapter 8 (Economic Impact), on the other hand, would also tend to
favor industry showing higher costs attributable to the promulgation
of the standards. In selecting a .figure to be representative for the
industry, the 4 percent estimate was chosen. Four percent is considered
to be more accurate due to comments received during the comment period
that the industry expansion figures were underestimated. No correction
of the impact analysis appears to be required because'the 4 percent
growth rate was used to estimate the impacts of the proposed standards.
However, it is noted that a 4 percent growth rate figure constitutes
an assumption which is more conservative than 3.5 percent would have
been. .
Incremental installed costs in January 1978 dollars associated
with the standard for controlling particulate emissions from a 90 Mg/day
pressed and blown glass furnace melting borosilicate formulations will
be about $800 thousand for an ESP and about $260 thousand for a fabric
filter. Incremental annualized costs for a 90 Mg/day furnace associated
with the standard will be about $245 thousand per year and about
$85 thousand per year for an ESP and a fabric filter, respectively.
Incremental cumulative capital costs for the 1979 through 1984
period associated with the standard for the two new 90 Mg/day furnaces,^
will be about $500 thousand if fabric filters are used. Fifth-year"
annualized costs for controlling the two new glass melting, furnaces to
comply with the standard will be about $160 thousand. ^
Based on the use of control equipment with the,highest annualized
costs, a price increase of about 0.8 percent wMl be necessary to
offset the costs of installing control equipment on the 90 Mg/pressed
and blown glass furnace melting borosilicate formulations to meet the
emission limits of the standard.
2-94
-------
Incremental installed costs in January 1978 dollars associated
with the standard for controlling participate emissions from a 45 Mg/day
pressed and blown glass furnace melting borosilic&te formulations will
be about $760 thousand for an ESP and about $235 thousand for a fabric
filter. Incremental annualized costs for a 45 Mg/day furnace associated
with the standard will be about $230 thousand/year for an ESP and
about $70 thousand/year for a fabric filters.
Incremental cumulative capital costs for the 1979 through 1984
period associated with the standard for a 45 Mg/day furnace will be
about $235 thousand if an ESP is used. Fifth-year annualized costs
for controlling a new glass melting furnace in this category to comply
with the standard will be about $70 thousand.
Based on the use of control equipment with the highest annualized
costs (worst-case conditions), a price increase of about 0.4 percent
will be necessary to offset the costs of installing control equipment
on a 45 Mg/day pressed and blown glass furnace melting borosilicate
formulations to meet the emission limits of thelstandard.
The incremental installed costs in January 1978 dollars associated
with the standard for controlling particulate emissions from a 90 Mg/day
pressed and blown glass furnace melting soda-lime and lead formulations
will be about $615 thousand for an ESP and about $770 thousand for a
fabric filter. Incremental annualized costs for a 90 Mg/day furnace
associated with the standard will be about $175 thousand/year and
about $235 thousand/year for an ESP and a fabric filter, respectively.
Incremental cumulative capital costs for the 1979 through 1984
period associated with the standard for the six:new 90 Mg/day furnaces
will be about $3.7 million if ESPs are used. Fifth-year annualized
costs for controlling these six new glass melting furnaces to comply
with the standard will be about $1.1 million.
Based on the use of control equipment with;the highest annualized
costs, a price increase of about 0.6 percent wi'll be necessary to
offset the costs of installing control equipment on the 90 Mg/day
pressed and blown glass furnace melting soda-lime and lead formulations
to meet the emission limits of the standard.
2-95
-------
The incremental installed costs in January 1978 dollars associated
with the standard for controlling participate emissions from a 45 Mg/day
pressed and blown glass furnace melting soda-lime and lead formulations
will be about $740 thousand for an ESP and about $710 thousand for a
fabric filter. Incremental annualized costs for a 45 Mg/day furnace
associated wixn the standard will be about $230 thousand/year for both
ESPs and fabric filters.
Incremental cumulative capital costs for the 1979 through 1984
period associated with the standard for the four new 45 Mg/day furnaces
will be about $2.8 million if a fabric filter is used. Fifth-year
annualized costs for controlling the four new glass melting furnaces
to comply with the standard will be about $910 thousand.
Based on the use of control equipment with the highest annualized
costs (worst case conditions), a price increase of about 0.6 percent
will be necessary to offset the costs of installing control equipment
on a 45 Mg/day pressed and blown glass furnace melting soda-lime and
lead formulations to meet the emission limits of the standard.
Incremental installed costs in January 1978 dollars associated
with the standard for controlling participate emissions from a 90 Mg/day
pressed and blown glass furnace melting other-than borosilicate,
soda-lime, and lead formulations will be about $800 thousand for an ESP
and about $260 thousand for a fabric filter. Incremental annualized
costs for a 90 Mg/day furnace associated with the standard will be
about $245 thousand per year and about $85 thousand per year for an
ESP and a fabric filter, respectively.
Incremental cumulative capital costs for the 1979 through 1984
period associated with the standard for a new 90 Mg/day furnace will
be about $250 thousand if a fabric filter is used. Fifth-year
annualized costs for controlling a" new glass metlting furance to
comply with the standard will be about $80 thousand.
Based on the use of control equipment with the highest annualized
costs, a price increase of about 0.4 percent will be necessary to
offset the costs of installing control equipment on a 90 Mg/day pressed
and blown glass furnace melting formulations other-than borosilicate,
soda-lime, and lead to meet the emission limits of the standard.
2-96
-------
'"X
Incremental installed costs in January 1978 dollars associated
with the standard for controlling participate emissions from a 45 Mg/day
pressed and blown glass furnace melting other-than borosilicate,
soda-lime, and lead formulations will be about $760 thousand for an ESP
and about $235 thousand for a fabric filter. Incremental annualized
costs for a 45 Mg/day furnace associated with the standard will be
about $230 thousand/year for an ESP and about $70 thousand/year for a
fabric filter.
Incremental cumulative capital costs for the 1979 through 1984
period associated with the standard for the 45 Mg/day furnace will be
about $235 thousand if an ESP is used. Fifth-year annualized costs
for controlling the new glass melting furnace in this category to
comply with the standard will be about $70 thousand.
Based on the use of control equipment with the highest annualized
costs (worst case conditions), a price increase of about 0.4 percent
will be necessary to offset the costs of installing control equipment
on a 45 Mg/day pressed and blown glass furnace melting other-than
borosilicate, soda-lime, and lead formulations to meet the emission
limits of the standard.
The economic impact for the new pressed and blown glass plants to
be built in the near future was based on anaylses of the various
segments of this category's three subcategories. It was through these
analyses that EPA has determined the economic impacts of these standards
of performance to be reasonable.
Wool Fiberglass. One commenter suggested that the wool fiberglass
standard would pose a monumental financial burden to the industry in
capital and operating costs.
For wool fiberglass, the new source performance standard operating
costs ar£ estimated to be 0.3 percent of sales for the model plant
analyzed. Just,.as for all segments of the industry, so also for wool
fiberglass, it was assumed that this would lead to no increase in
product price, but would be absorbed entirely out of profit. On that
basis, these operating costs fell below the criteria for significant
impact reconsideration, which were set at 1 percent of product price
and 10 percent of rate of return.
2-97
-------
It might be added that the assumption of no product price increase
is considered to be conservative for this industry segment. As pointed
out in Section 8.1 of the Background Information Document, Volume I,
projected demand in this segment appears to outstrip general demand
and shipments have increased in the face of rising prices. This
suggests some possibility of absorbing a 0.3 percent cost increase by
means of increased price rather than reduced profit.
As for capital requirements, the new source performance standard
capital cost was estimated to be $504,000. This is 0.9 percent of the
total capital required for construction of a new model plant. For the
companies in this industry segment, this kind of increase in capital
requirements would not put them out of the reach of the capital markets.
Incremental installed costs in January 1978 dollars associated
with the standard for controlling particulate emissions from a 180 Mg/day
wool fiberglass furnace will be about $500 thousand for an ESP and
about $70 thousand for a fabric filter. Incremental annualized costs
associated with the standard for a 180 Mg/day wool fiberglass furnace
will be about $155 thousand/year and about $20 thousand/year for an
ESP and a fabric filter, respectively.
Incremental cumulative capital costs for the six new 180 Mg/day
wool fiberglass furnaces during the 1979 through 1984 period associated
with the standard will be about $3 million if ESPs are used. Fifth-year
annual!zed costs for controlling wool fiberglass furnaces complying
with the standard will be about $930 thousand.
Based on the use of control equipment with the highest annualized
costs (worst case conditions), a price increase of about 0.3 percent
will be necessary to offset the costs of installing control equipment ^
on a 180 Mg/day wool fiberglass furnace to meet the emission limits^of
the standard.
Flat Glass. One commenter suggested that the number of flat
glass plants that will be affected by the standard by 1984 is four.
Due to an apparent computational error in a^prior analysis, the
number for flat glass plants to be built by 1984 was stated as one.
The correct number should have been four. The effect of this correction
in the number of new plants is to quadruple the total annualized
2-98
-------
control costs for this segment in the fifth year from $190,000 to
$760,000. No change in the impact study's conclusions results from
the correction of this computational error.
Annualized costs associated with the promulgated flat glass
manufacturing standard will be about $0.33 million for each of the
four new plants expected by 1984. Compliance with the standard will
result in annualized costs in the flat glass manufacturing industry of
about $1.32 million by 1984. For typical plants constructed between
1979 through 1984, capital costs associated with the promulgated
standard will be about $1.2 million. Cumulative capital costs of
complying with the promulgated standard for the estimated four additional
flat glass manufacturing plants will amount to about $4.8 million
between 1979 and 1984. The percent price increase necessary to offset
costs of compliance with the promulgated standard will be about
0.8 percent.
2.7 ENERGY IMPACT
One commenter suggested that one of the benefits of implementing
process modification techniques is the conservation of energy. As
discussed in the Emission Control Technology Section (2.2) of this
document, these techniques have not been adequately demonstrated.
Therefore, the EPA has not based the standards on these techniques.
However, application of process modifications should not conflict with
achieving the standards, and thus energy conservation attributable to
process modifications should be possible with or without the promulgation
of the standards.
Several commenters suggested that the national energy problem was
not being adequately taken into consideration if ESPs are installed on
glass melting furnaces that are regulated by the promulgated standards.
A detailed energy analysis was made prior to the decision to
recommend the use of add-on controls as a means of particulate emission
reduction from glass manufacturing plants. This analysis showed that
the energy use attributed to the installation of add-on controls to
comply with the promulgated standards will be minimal. Refer to
Section 1.2.2 of this document for details of these energy impacts
associated with the promulgation of these standards.
2-99
\
-------
An additional factor illustrating EPA's concern for the Nation's
energy policies is the incremental allowances provided to those glass
manufacturers that use fuel oil for their melting operations. This
allowance, provided for each of the glass manufacturing categories
with the exception of flat glass, effects an equalization in a company's
decision of whether to use fuel oil or natural gas to melt its glass.
The decision of what fuel to fire by a particular plant would not be
based on the respective emission potentials of each type of fuel but
would be based on the economics of fuel supply.
Several commenters suggested that process modifications use less
energy than add-on control devices. It was their opinion that the
installation and operation of add-on controls does not represent a
concerted effort on the part of EPA to consider the energy impacts
attributable to setting the standards as they are promulgated.
Process modifications, not yet demonstrated as a best system of
continuous emission reduction, do require the consumption of less
energy by glass manufacturing plants in certain instances (e.g., the
reduction of production reduces fuel consumption). In other instances,
they require the consumption of more energy by glass manufacturing
plants. Although this may be so, energy impacts are not the sole
criterion upon which the establishment of NSPS' are based. Energy
impacts are one of several factors taken into consideration in setting
standards that represent a demonstrated best system of continuous
emission reduction. In deciding what standards to establish, such
factors as economic, energy, emission reduction, production, and
others were considered. The impacts drawn from this analysis were
taken into account in the development of these standards.
As previously mentioned, if all the sources projected to be
subject to the standards find it necessary to install add-on control
devices the energy consumption attributable to the standards will be
approximately 62.8 million kWh/yr in 1984. Compared to the energy
requirements of the whole plant, this figure is considered small.
Container Glass. The energy required to control particulate
emissions from the 25 new container glass furnaces will be about
40.4 million kWh (22 thousand barrels of oil/year) for a typical SIP
2-100
-------
regulation for the new furnaces equipped with ESPs. This required
energy will be about 0.2 percent of the total energy use of these
container glass melting furnaces. The energy impact associated with
the promulgation of this standard will not be significant because the
energy required to operate an ESP for this standard (40.8 million kWh/yr
in 1984) is essentially the same as the energy required to operate an
ESP for a typical SIP regulation.
Pressed and Blown Glass (Borosilicate). Control to the level
required by a typical SIP regulation of the two new 90 Mg/day pressed
and blown glass furnaces melting borosilicate formulations projected
to come on-stream during the 1979 through 1984 period will require
about 6.6 million kWh (3,700 barrels of oil/year) if an ESP is used.
The energy requirements to achieve the standard's emission limit will
be essentially the same as the requirements for meeting a typical SIP
regulation (6.6 million kWh).
Although there are none projected to be built, the energy required
to control particulate emissions from a new 45 Mg/day pressed and
blown glass furnace melting borosilicate formulations to the level
required by the typical SIP regulation would be about 2.7 million kWh
(1,500 barrels of oil/year). The energy required to comply with the
standard's emissions limit would be essentially the same as that
required for meeting a typical SIP regulation (2.7 million kWh).
The energy required to comply with the emission limit of the
standard will be about 0.1 percent of total energy use of the affected
pressed and blown (borosilicate) glass melting furnaces in this glass
manufacturing category. Considering the small amounts of additional
oil and electricity required and the slight increase in total energy
use in this sector, the energy impacts of the standard will be negligible.
Pressed and Blown Glass (Soda-Lime and Lead). The energy required
to control particulate emissions from the six new pressed and blown
(soda-lime and lead) 90 Mg/day furnaces would be 4.1 million kWh
(2,500 barrels of oil/year) for a typical SIP regulation, or the
standard if ESPs are installed. There will be no energy impact
associated with the standard for the new 90 Mg/day furnaces beyond the
impact associated with the requirements to meet a typical SIP regulation
as the energy requirements are essentially equivalent (4.1 million kWh).
A2-101
1
\
-------
There will be no associated energy requirement for SIP compliance
since the four new pressed and blown (soda-lime and lead) 45 Mg/day
furnaces will be in compliance with a typical SIP regulation without
add-on controls. The estimated energy required to control particulate
emissions from the four new 45 Mg/day furnaces projected to come
on-stream in the 1979 through 1984 period to the level required by the
standard Will be about 1.5 million kWh (900 barrels of oil/year).
The energy required for a pressed and blown (soda-lime artd lead)
glass melting furnace to comply with the emission limits of the standard
will be about 0.5 percent of the total energy use of the affected
pressed and blown (soda-lime and lead) melting furnaces in this glass
manufacturing category. Considering the small amounts df additional
oil and electricity required and the slight increase in total energy
use in this sector, the energy impacts of the standard are considered
reasonable.
Pressed and Blown Glass (Other-than Bordsilicate. Soda-Linte,
and Lead). Control to the level required by a typical SIP regulatidh
of the two new 90 Mg/day pressed and blown glass furnaces melting
glass formulations other-than bdirdsilicate, soda-lime» and lead
projected to come on-strgaiti durifig the 1979 thrdugh 1984 period will
require about 2.6 million kWh (1,460 barrels of oil/year) if an ESP is
used. The energy requirements to achieve the standard's emission
limit will be essentially the same as the requirements for meeting a
typical SIP regulation (2.6 million kWh).
Although there are none projected to be built, the energy required
to control particulate emissions from a new 45 Mg/day pressed and
blown glass furnace melting other-than borosilicate* soda-lime, and
lead formulations to the level required by the typical SIP regulation
would be about 0.6 million kWh (350 barrels df dil/year). The energy
required to comply with the standard's emissions limit would be
essentially the same as that required for meeting a typical SIP
regulation (0.6 million kWh).
The energy required to comply with the emission limit of the
standard will be about 0.1 percent of total'energy use of the affected
pressed and blown (other-than borosilicate, soda-lime,,and lead)
2-102
-------
melting furnaces in this glass manufacturing category. Considering
the small amounts of additional oil and electricity required and the
slight increase in total energy use in this sector, the energy impacts
of the standard are considered reasonable.
Wool Fiberglass. The estimated energy required to comply with a
typical SIP regulation for the six new wool fiberglass furnaces expected
to come on-stream in the 1979 through 1984 period will be about
6.8 million kWh (3,850 barrels of oil/year) if ESPs are used. This
required energy will be about 0.3 percent of the total energy use of
the affected wool fiberglass melting furnaces. Complying with the
emission limits of the standard with ESPs will require about 6.9
million kWh (3,900 barrels of oil/year). The energy impacts of the
standard are considered reasonable (only about 50 barrels of oil/year).
Flat Glass. Because changes in the uncontrolled emission rate
and controlled emission rate were made for the flat glass category,
the energy impacts were re-evaluated.
The amount of electrical energy required of the ESP to meet the
typical SIP regulation is about 31,800 Btu/ton (1.03 x 10°2 kWh/Kg) of
glass. However, to meet the level attributable to the promulgated
standard, the ESP will consume approximately 32,200 Btu/ton
(1.04 x 10" kWh/Kg), which is only an additional energy requirement
of approximately 500 Btu/ton (1.5 x 10"4 kWh/Kg), and is considered
minimal.
Industry-wide, the addition of four new flat glass melting furnaces
by 1984 will require the consumption of approximately 33,600 million
Btu (9.839 million kWh) per year to meet the typical SIP regulation,
which is equivalent to 5,600 barrels of oil-per year. The energy
required to control these furnaces to the level of the promulgated
standard is. approximately 34,150 million Btu (9.97 million kWh) per
year, which is an additional consumption of 550 million Btu
(0.131 million kWh) per year. This is considered to be reasonable.
The secondary air quality impact associated with the promulgated
flat glass standard is not considered to be significant. To meet the
typical SIP regulation, the four new flat glass manufacturing plants
will cause affected electric utility plants to emit approximately
2-103
-------
4.9 tons (4,500 Kg) per year of participate through 1984. However, to
control these furnaces to the level of the promulgated standard, a
negligible amount of additional particulate will be emitted by the
electric utility plants.
For each of the new flat glass plants, the increased energy
consumption that would result from the promulgated standard is about
4 percent of the electrical energy consUWed to produce flat glass.
This is abtftit 0.2 percent of the total energy consumed to produce flat
glass arid is considered reasonable.
Because the additional energy requirement attributable to the
promulgated standards has no adverse impact oh national energy consumption,
and the §§c6ndary Sir quality impact is considered negligible, the
energy impacts associated with the promulgated, standards are considered
reasonable.
2.8 TEST METHODS AND MONITORING
Commenters stated that EPA Method 5 contains several sources of
error when applied in sampling emissions from soda-lime glass melting
furnaces.
First, commenters stated that misclassification of particulate
and gaseous species and inflated particulate emission values are
errors which can be caused by the use of filter temperatures below the
sulfur trioxide dew point. These comments are related to the presence
of sulfur trioxide (S03) in the exhaust stream from the glass melting
furnace. When particulate matter is filtered at about 250°F, sulfur
trioxide that is present and which condenses will be collected on the
filter as sulfuric acid. The measurement of this sulfuric acid by
Method 5 does not constitute an error in the method because Method 5
normally considers sulfuric acid as particulate matter. However, for
the glass manufacturing industry, the decision was made to define
particulate matter to exclude sulfuric acid for the following reasons:
(1) the variability of the S03 content in the stack gas with the
sulfur content of the fuel was not considered in developing the standards,
and (2) the technologies considered in establishing the standards do
not control S03. Therefore, the method was modified, for glass manufacturing
plants, to prevent collection of sulfur trioxide, as sulfuric acid, by
.-'"I
2-104
-------
allowing the filter and the probe to operate at temperatures of up to
350°F, which is above the acid dew point or condensation point.
Second, commenters stated that sulfuric acid will react with the
sodium sulfate particulate matter collected on the filter. To account
for this reaction, the commenters suggested that the Method 5 analytical
procedure be modified by taking an aliquot from the probe wash and
water extract of the filter, after drying and weighing the filter, and
titrating to pH 6 with 0.1 N sodium hydroxide. Then the amount of
sulfuric acid that reacted can be calculated and subtracted to determine
the particulate emission value.
The data that have been submitted indicate that sulfuric acid
will react with the particulate matter collected on the filter. EPA
evaluated the suggestion as follows: (1) the reaction of sulfuric acid
with sodium sulfate particulate will go to completion even at the low
levels of sulfuric acid present in gas-fired furnaces and, therefore,
uniformly influence particulate emission values whether from gas-fired
or fuel oil-fired furnaces; (2) the particulate emission values obtained
during data gathering were not corrected for the sulfuric acid reaction;
and, (3) if the particulate emission values are adjusted to account
for the reaction, then the level of the standard must be appropriately
adjusted. Thus, to be consistent with the manner in which the standards
were set, the method was not revised to allow that reacted sulfuric
acid be subtracted.
Third, commenters remarked that sulfur dioxide (S02) and SO, can
react with the alkali in the Method 5 filter and cause higher than
true particulate emission values.
An EPA report indicates that S02 or S03 reacts with some glass
fiber filters resulting in a significant weight gain. The report also
shows that this potential weight gain can be avoided by choosing a
source of filter material demonstrated to be nonreactive to S02 and
SOg. The degree to which this reaction occurs is apparently related
to the final rinse step of filter production, which varies according
to the supplier. In addition, this potential weight gain is not
significant when sampling high concentration emissions for short
sampling periods and when the gas contains no S02 or S03. The use of
2-105
-------
nonreactive filters has always been an option in Method 5; however,
the use of nonreactive filters would eliminate any such weight gain.
The filters used in collecting the data used as the basis of this
standard may have been reactive or may have been nonreactive. Therefore,
EPA is revising Method 5 to require the use of nonreactive filters in
testing sources whose gas streams contain S02 or S03.
Fourth, commenters indicated that EPA Method 5 contains a source
of error by including extraneous water vapor.
This comment is related to the fact that sulfuric acid is hygroscopic
and retains combined water after desiccation. Experiments have shown
that sulfuric acid desiccated with calcium sulfate retains 2 molecules
of water per molecule of acid after reaching equilibrium. If samples
are desiccated to equilibrium, the amount of combined water remaining
would be proportional to the amount of sulfuric acid present in the
collected particulate matter; while maintaining the samples at
equilibrium during weighing can present some problems,.it cart be
achieved with humidity controlled weighing rooms and careful techniques.
This potential source of error is only a problem if sulfuric acid is
collected by the filter. Because the method of emission measurement
for these standards was modified to prevent sulfuric acid mist from
being collected by the filter* this comment is ho longer an issue.
Last, commenters also suggested that a quartz or glass nozzle be
used on the probe, that the test method should allow a smaller minimum
sample volume, and that the two most consistent runs of the required
three runs should be used in determining the emission values*
A quartz or glass nozzle is allowed by Method 5. The minimum
sample requirement was modified to allow this option of lower sampling
volumes provided that a minimum of 50 milligrams of sample is collected.
Determination of compliance, as set forth in 40 CFR 60.8, is based on
the arithmetic average of three runs. The standards promulgated for
glass manufacturing are based on such an average. Therefore, a change
in this average would be inconsistent with the data used to establish
the standards. However, as set forth in 40 CFR 60.8, if certain
conditions exist, compliance may be determined, upon approval by EPA,
by using the average of two runs.
2-106
-------
2.9 CLARIFICATIONS
Commenters expressed concern with the possible confusion of
whether an entire glass manufacturing plant might be considered to be
an affected facility if one of its glass melting furnaces was to be
modified or reconstructed and thereby subject to these new source
performance standards. This confusion was remedied by redrafting the
description of the affected facility to delete glass manufacturing
plants as part of the affected facility. The affected facility is now
limited to the glass melting furnace as defined in the regulation.
Also suggested was a provision to specifically exclude the float
bath used in the flat glass category from being regulated as a part of
the furnace (affected facility). The float bath is considered to be
part of the forming process, not the melting process, and is therefore
not regulated by these NSPSs. To remedy this possible area of confusion,
the regulation has been rewritten as suggested.
One commenter corrected EPA in its use of the word "day pot
furnace" to describe the 2 tons per day glass melting furnaces exempted
under the proposed regulations. It was pointed out that the industry-wide
term most used to refer to small glass melting furnaces is "pot furnace."
It was suggested that instead of the terms "day pot" and "day pot furnace,"
as used in the proposed regulatory definitions [Section 60.291(c)] and
standards for particulate matter [Section 60.292(d)] and the proposed
preamble (44 FR 34842), respectively, the term "pot furnace" be used.
This term was deleted from the regulation as a result of the exemption
of hand glass melting furnaces from compliance with these standards.
The term "glass manufacturing plant" was removed from Section 60.291
Definitions of the regulation because it was not needed.
The recipe definitions were also changed where appropriate to
describe the specialized batch formulations found in the pressed and
blown glass category. Detailed recipes for borosilicate, soda-lime
and lead, and other than borosilicate, soda-lime and lead were included
in Section 60.291 Definitions of the regulation.
Sections 60.293 to 60.295 are reserved in the event additional
provisions will be necessary for clarity or other reasons.
L
2-107
-------
The term "glass production" in the proposed regulation's
Section 60.291 was changed to "glass produced" to better state the
basis upon which the standards are determined to be met.
The term "hand glass melting furnace" was added to the regulation in
Section 60.291 as a result of the exemption of the furnaces typically
operated in this sector of the industry.
2-108
-------
APPENDIX A
A-l
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Run
Date
Production Rate*
(Ib/hr)
Stack Flow Rate,
(scfm)
Temperature, (°F)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
None
Owens-Corning Fiberglas
Toledo, Ohio
FG-2, All-electric
Wool Fiberglass
4-16
II-B-77, 3/16/78
TEST DATA
1
4/18/73
Confidential
14*300
110
0.005
0.59
A-2
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
None
Owens-Corning Fiberglas
Toledo, Ohio
F6-2, All-electric .
Wool Fiberglass
4-17
II-B-77, 3/16/78
TEST DATA
Run
Date
Production Rate,
(Ib/hr)
Stack Flow Rate,
(scfm)
Temperature, (°F)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
3
4/19/73
Confidential
9,680
126
0.01
0.82
A-3
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I ftef.
EMB Memo Ref*
None
Owens-Corning Fiberglas
Toledo, Ohio
#70, All-electric
Wool Fiberglass
4-18
II-B-77, 3/16/78
TEST DATA
Run
Date
Pull Rate,
(Ib/hr)
Stack Flow Rate,
(scfm)
Temperature, (°F)
Grain Loading*
(Gr/dscf)
Mass Emissions
(Ib/hr)
4
3/15/76
Confidential
8,613
116
0.013454
0.993
A-4
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
None
Ball Corporation
El Monte, California
No. 2, All-electric
Container
4-19
None (LAAPCD Method)
TEST DATA
Run
Date
Process Weight Rate,
(Ib/hr)
Stack Flow Rate,
(dscf)
Temperature, (°F)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
1
12/12/74
7774
400
1,040
0.2
0.7
2
1/30/75
8107
568
930
0.2
0.9
3
2/11/75
8060
576
1,090
0.2
1.0
A-5
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Fabric Filter
Owens-Illinois
City of Industry, California
B
Pressed and Blown (soda-lime)
4-24
II-B-55, 12/21/77
TEST DATA
Run 1
Date 11/29/73
Process Weight Rate,
(Ib/hr) 6,827
Flow Rate Thru
Control Device, (dscf) 4,600
Temperature, (°F) 650
Grain Loading,
(Gr/dscf) 0.066
Mass Emissions,
(Ib/hr) 2.6
A-6
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Fabric Filter
Corning Glass Works
Central Falls, Rhode Island
# 08
Pressed and Blown; Soda-Lead-Borosilicate, Code 7720
4-25
II-B-63, 1/09/78
TEST DATA
Run
Date * '
Fill Rate,
(Ib/hr)
Flow Rate Thru
Control Device,
Temperature, (°F)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
1
1/14/77
2,263
(dscf) 9,197
295
0.005
0.41
2
1/14/77
2,763
9,255
292
0.004
0. 31
3
1/14/77
2,263
. 9,228
292,
0. 003
0.27
A-7
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref,
Fabric Filter
Owens Corning Fiberglass
Toledo, Ohio
T and P20
Wool Fiberglass
4-26
IL.B-77, 3/16/78
TEST DATA
1
6/04/74
Confidential
Run
Date
Production Rate,
(Ib/hr)
Flow Rate Thru
Control Device, (scfm) 21,200
Temperature, (°F) 220
Grain Loading,
(Gr/dscf) 0.012
Mass Emissions,
(Ib/hr) 2.1
A-8
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Fabric Filter
Owens Corning Fiberglas, Inc.
Toledo, Ohio
K Regenerative
Wool Fiberglass
4-27
II-B-77, 3/16/78
TEST DATA
Run
Date
Production Rate,
(Ib/hr)
Flow Rate Thru
Control Device, (acfm) 14,400
Temperature, (°F)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
6/19/74
Confidential
336
0.072
8.9
A-9
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Fabric Filter
Owens Corning Fiberglas
Toledo, Ohio
T and P20
Wool Fiberglass
4-28
II-B-77, 3/16/78
TEST DATA
3
6/04/74
Confidential
Run
Date
Production Rate,
(Ib/hr)
Flow Rate Thru
Control Device, (scfm) 21,200
Temperature, (°F)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
224
0.017
3.1
A-10
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Venturi Scrubber
Glass Containers Corporation
Dayville, Connecticut
#1 #5 Fuel Oil
Container
4-31
II-B-61, 1/05/78
TEST DATA
Run
Date
Pull Rate,
(Ib/hr)
Flow Rate Thru
Control Device,
Temperature, (°F)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
1
11/11/75
18,833
(dscf) 19,174.2
140
0.0507
8.33
2
11/12/75
18,833
19,691.5
140
0.0361
6.09
3
11/12/75
18,833
18,732.4
140
0. 0404
6. 49
A-ll
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Venturi Scrubber
Glass Containers Corporation
Vernon, California
No. 1
Container
4-32
None
TEST DATA
Run 1
Date 08/21/74
Process Weight Rate*
(Ib/hr) 15,720
Flow Rate Thru
Control Device, (dscf)
inlet/outlet 8,470/8,890
Temperature, (°F) - 160
Grain Loading,
(Gr/dscf) 0.0225
Mass Emissions,
(Ib/hr) 1.6
A-12
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Venturi Scrubber
Brockway Glass Company
Oakland, California
No. 21
Container
4-33
II-B-75, 3/13/78
TEST DATA
Run 1
Date 04/06/77
Production Rate,
(Ib/hr) 16,333
Flow Rate Thru
Control Device,
. (dscf) 9,400
Temperature, (°F) 150
Grain Loading,
(Gr/dscf) 0.00995
Mass Emissions,
(Ib/hr) 2.3
A-13
-------
Control Device
Company
Furnace
, 61 ass Type
BID, Vol. I Ref.
EMB Memo Ref.
Venturi Scrubber
Glass Containers Corporation
Vernon, California
No. 1, Natural Gas (runs I and 2) Fuel Oil (run 3)
Container
4-34
II-B-41, 10/28/77
TEST DATA
Run
Date
Process Weight Rate,
(Ib/hr)
1
01/24/75
15,200
2
01/29/75
15,700
3
01/31/75
15,700
Flow Rate Thru
Control Device, (dscf)
Scrubber inlet/Knockout
Tower outlet 8,460/8,460
Temperature, (°F)
Packed Tower Inlet/
Knockout Tower Outlet 960/160
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
0.031
2.2
6,400/8,740
12,100/12,300
970/160
0.034
2.5
970/150
0.335
35.3
A-14
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Thatcher Glass Manufacturing Company
Saugus, California
No. 3
Container
4^38
None (LAAPCD Method)
TEST DATA
Run
Date
Process Weight Rate,
(Ib/hr)
Flow Rate Thru
Control Device, (dscf)
(inlet/outlet) 23
Temperature, (°F)
(inlet/outlet)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
1
05/02/75
26,500
,200/30,100
670/500
0.005
1.3
2
05/07/75
25,827
24,000/31,300
650/500
0. 0034
.0.9
. 3
05/09/75
23,800
27,000/31,000
640/520
0.0059
1.6
A-15
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Owens-Illinois
Los Angeles, California
23 A
Container
4-39
None (LAAPGD Method)
TEST DATA
Run
Date
Process Height Rate,
(Ib/hr)
Flow Rate Thru
Control Deyice?
Temperature,, (°F)
Grain Loading.
(Gr/d,scf)
Mass Emissions,
Ob/hr)
01/p7/7§
18,816
13,509
610
0. 00795
0.92
01/28/75
17,640.7
13?9QP
§60
0. 00740
0,825
02/27/75
15,765.5
13,300
620
0.00757
0.9
A-16
-------
">
f
(-3
-o '
C
O
to to CQ
4J r**
r- C *
Q. i. «£
i- O
O «t- «/>
CO - +J
Q. 'to *C
O 33
i- '5
-*--> C"» (/» f/K
T^ b» VJ VJ
4-» i^ r- C§ S-
CO r CD CO
0 I-H cn f c
S- 1 C * -
+-> co «=c s_ re
O C 3 +J
CO CO t/> *-* C
i S O
(U
o
r-
>
CO a)
***** **"*
>>
r >j CO H-
O C U
im
CO
CM
^s^
^"
CO
CM
^~
CO
CM
^*
vo
CM
^^^
**
A
CO
4-> 3
IQ ^.
*^ f^
^""^ 1"^
r^> c
r^ o >~» co
en *r* &_ 4^
«-H +* J= «
~ o «>» a;
0) -0 i S
4-> 0 ^^ 0
1T3 S- r
O 0. Ll-
CS
O
en
f^
0
o
CO
A
LO
O
0
CO
M
LO
"=*
0
§
«t
CO
o
o
co
t
CM-
o
o
CM
A
*
O
o
A
^
O
O
LO
41
p*^
^
o
o
vo
>
«3-
O
0
A
O^
§
IO
«*
0
o
M
en
o
A
LO
CO
U
r-
>
CO
a
pH
O ^""^
i ^
c o
StO
^^
0
co
^j-
o
en
o
^j»
o
CM
LO
O
CT>
O
en
LO
O
en
LO
0
o
LO
O
CO
LO
o
o
LO
o
CO
o
LO
U.
0
^^
A
CO
^
4^
s_
CO
Q.
t
A
o>
c
r-
o
<0
o
J
c
r-
CD
CO
O
O
0
t 1
I <
o
CD
*5t"
o
o
o
r-H
O
0
o
f«s^
11
O
»
0
l*^fc
o
0
C3
VO
p*" 4
o
CM
i I
O
O
^^
o
0
o
CM
O
O
o
vo
0
o
o
CO
0
o
*
o
en
0
o
_
o
*^ (/)
l^ c
U O
O co
c_ ,^
t5 E
V)
CO
^~
o
CM
,
CM
en
10
0
U3
CM
,
O
O
1 1
.
CO
1-4
CO
r-l
O
CO
f«^
t 1
,
«M
i 1
CO
o
el-
m
O
t-4
t-l
1 1
CO
LO
O
O
co
^^
s_
^s^
.a
J3.
A-17
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
General Electric Company
Jackson, Mississippi
Natural Gas
Pressed and Blown (Borosilicate)
4-41(a)
II-B-47, 11/11/77
TEST DATA
Run
Date
Process Weight Rate,
(Ib/hr)
Flow Rate Thru
Control Device, (acfm)
Temperature, (°F)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
f
02/25/76
2,170
9,501
323
0.0020
0.1025
2
02/26/76
2,170
9,754
326
0. 0020
0. 1040
3
02/26/76
2,245
9,588
325
0. 0020
0. 1001
A-18
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
General Electric Company
Niles - Mahoning, Ohio
Pressed and Blown (Borosilicate)
4-41(b)
II-B-76, 3/13/78
TEST DATA
Run 1
Date 09/28/77
Process Weight Rate,
(Ib/hr)
Flow Rate Thru
Control Device, (dscf) 31,021
Temperature, (°F) 504
Grain Loading,
(Gr/dscf) 0.02857
Mass Emissions,
(Ib/hr) 6.94
09/28/77
Confidential
09/28/77
31,531
483
0. 02594
6.35
29,022
494
0.02916
6.69
A-19
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator (2)
Owens-Illinois
Vine!and, New Jersey
J, K, L, M, and R.
Pressed and Blown (Borosilicate)
4-42
II-B-5, 12/22/77; II-B-60, 1/3/78; II-B-99, 12/3/78
TEST DATA
12/05/73
13,321
Run
Date
Process Weight Rate,
(Ib/hr)
Flow Rate Thru
Control Device, (dscf) 47,810
Temperature, (°F) 414
Grain Loading,
(Gr/dscf) 0.020
Mass Emissions,
(lb/hr) (None presented)
12/05/73
13,189
48,136
410
0.018
12/06/73
13,694
52,601
426
0.017
A-20
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Corning Glass Company
Fall brook, New York
Tank 41 Natural Gas with Excess Air
Pressed and Blown (Borosilicate)
4-43
II-B-63, 1/09/78
TEST DATA
Run 1
Date 08/06/75
Fill Rate,
(Ib/hr) 1,046
Stack Flow Rate, (dscf) 8,420
Temperature, (°F) 295
Grain Loading,
(Gr/dscf) 0.005
Mass Emissions,
(Ib/hr) 0.38
08/06/75
1,046
8,730
298
0.007
0.48
08/06/75
1,046
8,295
296
0.009
0.66
A-21
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
General Electric Company
Somerset, Kentucky
# 5, Fuel Oil
Pressed and Blown (Borosilicate)
4-44
II-B-48, 11/14/77
TEST DATA
Run
Date
Production Rate,
(Ib/hr)
Flow Rate Thru
Control Device,
Temperature, (°F)
Grain Loading,
(Gr/scf)
Mass Emissions,
(Ib/hr)
1
04/28/76
8,750
(acfm) 29,927
458
0.0270
3.61
2
_ 04/28/76
8,750
30,646
462
0. 0329
4.48
3
04/28/76
8, 750
30,400
458
0.0322
4.4
A-22
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Corning Glass Works
Charleroi, Pennsylvania
Tank 66, Natural Gas
Pressed and Blown (Fluoride/Opal)
4-45
II-B-63, 1/09/78
TEST DATA
Run 1
Date 03/19/76
Fill Rate,
(lb/hr) 8,000
Flow Rate Thru
Control Device, (dscf) 19,165
Temperature, (°F) 364
Grain Loading,
(Gr/dscf) 0.006
Mass Emissions,
(lb/hr) 1.05
03/19/76
8,000
18,690
377
0.008
1.25
A-23
-------
Control Device
Company
Furnace
Glass Type
BID, VoU I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Corning Glass Works
State College, Pennsylvania
Tank 222, Natural Gas with Excess Air
Pressed and Blown (lead)
4-46
II-B-63, 1/09/78; II-B-54, 12/09/77
TEST DATA
Run i
Date 04/30/75
Fill Rate,
(Ib/hr) ; 5,500
Flow Rate Thru
Control Device, (dscf) 14,650
Temperature, (°F) 396
Grain Loading,
(Gr/dscf) 0.0024
Mass Emissions*
(Ib/hf) 0.31
A-24
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
General Electric Company
Logan, Ohio
Natural Gas
Pressed and Blown (lead)
4-47
II-B-49, 11/14/77
TEST DATA
Run 1
Date 04/21/76
Fill Rate,
(Ib/hr) . 3,760
Flow Rate Thru
Control Device, (acfm)
Temperature, (°F)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr) 0.278
2
04/21/76
3,760
17,600
365
0.0040
0. 347
3
04/22/76
3,760
17,789
370
0.0028
0.247
A-25
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Corning Glass Works
State College, Pennsylvania
Tank 221, Natural Gas
Pressed and Blown (Lead)
4-48
II-B-63, 1/09/78
TEST DATA
(Dry
Particulate)
Run 1
Date 11/04/75
Fill Rate,
(Ib/hr) 15,167
Flow Rate Thru
Control Device, (dscf) 35*370
Temperature, (°F) 422
Grain Loading,
(Gr/dscf) 0.0013
Mass Emissions,
(Ib/hr) 0.4
(Dry
Particulars)
2
11/04/75
15,167
35,390
419'
0/0016
0.49
(Total
Particulate)
3
11/04/75
15,167
33,910
421
0.0052
1.5 (Total)
0.5 (Dry)
A-26
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Corning Glass Works
Danville, Kentucky
Tank 122; Natural Gas with Excess Air
Pressed and Blown (lead)
4-49
II-B-63, 1/9/78
TEST DATA
Run
Date
Fill Rate,
(Ib/hr)
Flow Rate Thru
Control Device,
(dscf)
Temperature, (°F)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
1
11/18/75
6,374
16,170
331
0. 0044
0.61
2
11/18/75
6,374
15,700
338
0.0021
0.28
3
11/19/75
6,049
15,900
325
0. 0034
0.46
4
11/19/75
6,049
16,100
308
0.0032
0.44
5
11/19/75
6,049
16,600/
304
0.0025
0.36
A-27
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Owens-Illinois
Vine!and, New Jersey
H
Pressed and Blown (lead)
4-50
None
TEST DATA
Run
Date
Process Weight Rate,
(Ib/hr)
Flow Rate Thru
Control Device* (dscf)
Tempera ture* (°F)
Grain Loading i
(Gr/dscf)
Mass Emissions,
(Ib/hr)
1
10/08/74
2,494
11,135
325
0.004
0*3
2
10/09/74
2,347
10,045
325
0.004
0.4
3
10/09/74
2,347
11,270
330
0.005
0.4
A-28
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
GTE Sylvania
Versailles, Kentucky
Propane and #6 Fuel Oil
Pressed and Blown (lead)
4-51
IV-B-14, 6/18/80
TEST DATA
Run
Date
Process Weight Rate,
(Ib/hr)
Flow Rate Thru
Control Device, (dscf)
Temperature, (°F)
Grain Loading,
(Gr/dscf)
Mass Emissions,
(Ib/hr)
1
08/24/76
5,200
11,211
419
0.015
1.4
2
08/24/76
5,200
11,334
430
0.015
1.42
3
08/25/76
5,200
11,265
451
0.015
1.45
A-29
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostati c.Preci pi tator
Owens-Illinois
Los Angeles, California
#23A
Pressed and Blown (Potash-Soda-Lead)
4-52
None (LAAPCD Method)
TEST DATA
Run
Date
Process Weight Plate,
(Ib/hr)
Flow Rate Thru
Control Device, (dscf)
Temperature, (°F)
Grain Loadingj
(Gr/dscf)
Mass EmissiortSj
(Ib/hr)
North ESP
07/09/74
21,524S4
South ESP
07/09/74
7i650
670
0*022
1.4
7*390
630
0. 00991
0,6
A-30
-------
Control Device
Company
Furnace
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Johns-Manville Sales Corporation
Parkersburg, West Virginia
#402
4-53
II-B-92, 06/08/78
TEST DATA
Run 1
Date 11/09/77
Production Rate,
(Ib/hr) 10,000
Flow Rate Thru
Control Device, (dscf) 23,810
Temperature, (°F)
ESP Inlet/Outlet 550/510
Grain Loading,
(Gr/dscf) 0.0175
Mass Emissions,
(Ib/hr) 3.58
A-31
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Johns-Manville Sales Corporation
Parkersburg, West Virginia
#402
Wool Fiberglass
4-54
II-B-92i 06/08/78
TEST DATA
Run 2
Date 11/09/77
Production Plate,
(lb/hr) 16,000
Flow Rate Thru
Control Device* (dscf) 23,582
Temperaturej (°F)
ESP Inlet/Outlet 550/505
Grain Loading,
(Gr/dscf) 0.0047
Mass Emissions,
(lb/hr) 0.961
A-32
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Johns-Manville Sales Corporation
Parkersburg, West Virginia
#402
Wool Fiberglass
4-55
II^B-92, 06/08/78
TEST DATA
Run 3
Date 11/10/77
Production Rate,
(Ib/hr) 10,000
Flow Rate Thru
Control Device, (dscf) 24,277
Temperature, (°F)
ESP Inlet/Outlet 550/495
Grain Loading,
(Gr/dscf) 0.0041
Mass Emissions,
(Ib/hr) 0.861
A-33
-------
Control Device
Company
Furnace
Glass Type
BID, Vol. I Ref.
EMB Memo Ref.
Electrostatic Precipitator
Guardian Industries, Inc.
Kingsburg, California
Oil-Fired
Flat
None
None - Final Report No. 2600-01-1079
EPA Contract No. 68-02-2813
Work Assignment No. 38
TEST DATA
Run 2
Date 9/18/79
Production Rate,
(Ib/hr)
Gas Sample Volume
At Std. Conditions,
(Ft3) 54.137
3
9/19/79
Confidential
51.650
4
9/19/79
51.855
Stack Gas Velocity,
(Ft/sec) 57.95
Temperature, (°F) 614
Grain Loading,
(Gr/dscf) 0.0156
Mass Emissions,
(Ib/hr) 5.22
57.61
620
0.0143
4.70
57.35
620
0.0230
7.59
A-34
-------
APPENDIX B
B-l
-------
APPENDIX B
An Explanation of the Discounted Cash Flow Methodology
The use of the discounted cash flow (DCF) methodology has several
purposes: 1) it helps to overcome the disadvantages of a static tool
such as ROI, Return on Investment, which is computed on the basis of
one year's figures and 2) it can take into consideration all of the
various cash inflows and outflows that occur on a random or changing
basis over the life of a project. Concerning variation in cash flows,
an investment tax credit is one example of a cash flow that only
occurs in the first or early years of the project life. Other examples
giving rise to varying cash flows are differences in financing and
depreciation periods over the useful life of a project for both the
basic process equipment as well as pollution control equipment. A
static analysis which focuses on the data of only one year cannot
adequately take all of these changing variables into account.
The DCF technique estimates and compares cash inflows over the
life of any project (e.g., newly constructed plants). The changing
value of money over time is considered in the comparisions by discounting
those cash flows to the present time. The discount rate used is the
firm's cost of capital. The effect of inflation over a project life
can also be taken into consideration by choosing the firm's real cost
of capital or a constant or current dollars cost of capital.
' If the present value of the discounted cash inflows is greater
than the present value of the discounted cash outflows, the project is
economically justified. The difference between the discounted cash
inflows and outflows is the net present value. If the net present
value of a project with NSPS controls yields a net present value
greater than zero, then it can be concluded that the impact of NSPS
controls would not cause an inhibition of the project.
The DCF technique is considered appropriate for decision making
on a profit maximizing basis, and has the capacity to address all the
important economic variables involved in such a decision contextl It
B-2
-------
is recognized that factors other than profit maximization may exert
considerable influence in individual plant investment decisions
(maintenance or enhancement of market shares is one example); however,
such factors are generally not amenable to objective analysis.
The cash inflows of the project generally consist of the firm's
after-tax profit, plus the various sources of depreciation and, in
this instance, after-tax interest. The cash outflows are dependent on
the method of project financing and selection of discount rate. Two
general approaches can be used. If the weighted average cost of
capital is the discount rate, the cash outflows consist of the
initial investment and any sustaining capital expenditures that would
be necessary to maintain the useful life of the project. In this
case, sustaining capital expenditures were assumed to be zero. In
contrast, if the cost of equity is the discount rate, the cash outflows
consist of the initial equity portion of the total investment plus the
debt repayments that occur over the financing period.
Concerning the specific cash inflows and outflows, Table 8-28 of
the Background Information Document, Vol. 1 for the New Source
Performance Standards for glass manufacturing plants presented the
formulae and the sources of data for the actual tables that followed
(8-29 through 8-31). Table 8-28 provides information for either the
weighted average cost of capital or the cost of equity at discount
rates. Tables 8-29 and 8-30 employ the cost of equity discount method,
whereas Table 8-31 employs the weighted average cost of capital.
As they now appear in the Background Information Document, Vol. 1,
there are some changes required in the lines for Tables 8-29 through 8-31
concerning cash inflows and outflows. Tables 8-29 and 8-30 mistakenly
include interest time tax as a cash inflow, whereas Table 8-31 mistakenly
includes debt repayments as a cash outflow. These lines in those
three tables should be eliminated and the subsequent yearly cash flows
corrected by subtracting the amount in these lines from the yearly
cash flows.
The information on the left-hand side of Table 8-28 is for both
the container and handmade consumerware DCF. All the calculations of
Tables 8-29 through 8-31 were computed on a per-ton-of-glass-produced
basis.
B-3
-------
The first seven lines of Table 8-28 are all connected to the
derivation of the cash inflow of line 6, i.e. profit. The first line
indicates that the initial cash inflow is the revenue received from
the sale of product, which is the average selling price per ton. To
take cost into account, that average selling price is multiplied by a
profit rate before taxes and before pollution control costs. The
resultant figure is profit before taxes. The next figure subtracted
from that profit level is the annualized pollution control cost, taken
from the tables in Section 8.2 and divided by the average tons produced
so that the utilization rate or yield of the plant is taken into
consideration. The annualized control costs in Section 8.2 were
computed on the basis of full capacity and annual costs. The resultant
cash flow at this point is profit before taxes and after pollution
control costs. This amount is then multiplied by the tax rate, which
at the time of the analysis was 48 percent. This calculation yields
taxes payable, which when deducted from the previous line results in
the next line, or profits. It should here be noted that line 7 should
read, "Profit After Taxes and After Pollution Control," not "Profit
Before Taxes and After Pollution Control."
The next cash inflow that must be included is the investment tax
credit which is available for the purchase of equipment, including
pollution control equipment. The Internal Revenue Code provides that
10 percent of the equipment cost can be taken as a credit against
taxes payable, beginning in the first year of the project life. There
are rules applicable to the limit of investment tax credit that can be
taken in any one year. The limit at the time that these calculations
were performed was 50 percent of the taxes payable (line 6).
The next cash inflow included is the depreciation from the buildings
and equipment, including pollution control equipment. The depreciation
periods for equipment were assumed to be 15 years for both the container
and handmade consumerware industries. The depreciation period for a
building was assumed to be 40 years in the case of container glass
plants and 33 years in the case of handmade consumerware plants.
Equipment depreciation was derived by taking the total equipment cost
for the new investment, including baseline control, and dividing that
B-4
-------
cost by 15 years, and then further dividing that yearly figure by the
number of tons of production capacity times the yield of the plant in
any one year. In the case of the three DCF calculations presented in
Tables 8-29 through 8-31 the yields were less in the first year than
in the second year. This is why the costs in 1978 differ from those
of 1979. To obtain building depreciation a similar procedure is
followed as that for the equipment depreciation.
In the weighted average cost of capital discounting approach, the
next cash flow to enter is interest times the tax rate of 48 percent.
The purpose of including interest times tax as a cash flow is to
recognize the tax savings that were generated by the use of someone
else's borrowed money in the business. The interest rate utilized was
10 percent which was multiplied in the first year by the amount of the
total investment in the project for which debt was incurred and by the
tax rate. Subsequent years' interest levels are lower due to yearly
debt payments. In the case of the 500 ton per day container glass
plant, the debt incurred was 85 percent of the investment. The resultant
interest calculation yields the yearly amount of after-tax interest
which then must be divided by the number of tons of container glass
produced in a given year.
Under the cost of equity DCF approach, the cash flow for debt
repayments must be subtracted as a cash flow. The amount of debt is
determined by multiplying the total investment by the percent that is
funded by debt and dividing by the yearly production capacity of the
plant multiplied by the yearly production yield. The assumptions
(from industry sources) were that the container glass plant would be
85 percent debt financed and that the handmade consumerware plant would
be 50 percent debt financed.
The net cash flow line represents the net sum of the cash inflows
minus the cash outflows. For the weighted average cost of capital DCF
method, this represents profit plus the investment tax credit plus
depreciation plus interes't multiplied by tax. For the cost of equity
method, the net cash flow is the sum of profit plus investment tax
credit plus depreciation minus debt repayment. Each year's net cash
flow is multiplied by a discount rate which takes into account the
B-5
-------
value of money to the firm, to obtain the discounted cash flow. For
the weighted average cost of capital method, the discount rate was
8 percent, whereas it was 15 percent under the cost of equity method.
(Figures supplied by the industry.) The present value of the discounted
cash flow for 20 years was then compared to the initial equity investment
under the cost of equity discounting method (Tables 8-29 and 8-30) and
to the total investment under the weighted average cost of capital
method (T§M§ 8*31). Jn all cases the net present value was greater
than thpse investment figures, signifying that the project returned
more than the minimum acceptable rate of return and thus justified the
prpjept? iffi-'» tne building pf the new plant.
B-6
-------
APPENDIX C
C-l
-------
APPENDIX C
Startup Considerations
Introduction
The promulgated standards of performance for glass manufacturing
plants apply to new, modified, and reconstruct-glass melting furnaces.
A glass melting furnace is a unit comprising a refractory vessel in
which raw materials are charged, melted at high temperatures, refined,
and conditioned to produce molten glass. Glass melting furnaces with a
production capacity of less than 5 tons/day and all-electric melting are
exempt from the promulgated standards of performance. In addition, an
increment of 30 percent is provided for oil-fired glass melting furnaces,
except for flat glass melting furances.
Specifically, the standards of performance, as they apply to gas-fired
glass melting furnaces would limit particulate exhaust emissions to:
(1) 0.2 Ib/ton of container glass produced, (2) 1.0 Ib/ton of pressed
and blown (borosilicate) glass produced, (3) 0.2 Ib/ton of pressed and
blown (soda-lime and lead) glass produced, (4) 0.5 Ib/ton of pressed and
blown (other-than borosilicate, soda-lime and lead) glass produced,
(5) 0.5 Ib/ton of wool fiberglass produced, and (6) 0.45 Ib/ton of flat
glass produced. Sources constructed, reconstructed,, or modified after
June 15, 1979 would be subject to the regulation.
Process Description
There are four basic processes involved in glass manufacturing:
1) raw materials handling and mixing, 2) glass melting, 3) forming and
finishing, and 4) packaging. As mentioned above, standards of performance
for the glass manufacturing industry apply only to the glass melting
process. A detailed description of the glass melting process is included
in Volume I, "Glass Manufacturing Plants, Background Information: Proposed
Standards of Performance," EPA-450/3-79-005a (refer to pages 3-6 to 3-8).
C-2
-------
Glass Melting Furnace
The glass melting furnace consists of a foundation, superstructure,
and retaining walls, raw materials charging system, heat exchangers,
melter, cooling system, exhaust boosting equipment, integral control
systems, and instrumentation, and appendages for conditioning and distributing
molten glass to forming apparatuses.
There are basically four types of glass melting furnaces in use
today by the glass manufacturing industry. As mentioned earlier, glass
melting furnaces designed to produce 5 tons or less of glass per day and
all-electric furnaces are exempt from the standards of performance and
therefore will not be discussed. The remaining two types of melting
furnaces are direct-fired and continuous and are referred to as regenerative
and recuperative furnaces. Many of these furnaces have added electric
induction systems called "boosters" to increase capacity. The furnaces
are fired either by natural gas or fuel oil. A detailed description of
regenerative and recuperative furnaces is included in Volume I, "Glass
Manufacturing Plants, Background Information: Proposed Standards of
Performance," EPA-450/3-79-005a (refer to pages 3-6 to 3-8).
Mixed raw materials are fed into the furnace by a charging system
and float on the bed of molten glass until it melts. Here, in the
melter section at a temperature of approximately 2,700°F, diffusion and
chemical reaction transform the batch into molten glass. Glass flows
from the melter into a second compartment, commonly referred to as the
refiner, where it is mixed for homogeneity and heat conditioned to
eliminate bubbles and stones. Glass temperature is gradually lowered to
about 2,200°F before exiting the furnace for forming.
Air pollutants generated in the melting of glass arise from the
vaporization of raw materials and the combustion of fuel. Pollutants
emitted from fossil fuel-fired furnaces producing soda-lime glass are
oxides of nitrogen, oxides of sulfur, carbon monoxide, hydrocarbons, and
submicron-sized particulates. Other pollutants such as, arsenic, borates,
fluorides and lead, which are emitted in the exhaust of pressed and
blown glass furnaces make significant contributions to the total emissions
from the glass industry. Vents located opposite the refiner section of
the furnace exhaust all air pollutants generated during the melting
process to a stack. Exhaust gases are either ducted to an air pollution
C-3
-------
control device before exiting through the stack or directly to the
atmosphere.
Pre-Startup Operations
Although the glass manufacturing industry has been classified into
four major categories depending upon the type of raw materials used
and/or final product, the procedures involved in furnace pre-startup and
startup are virtually the same. During the pre-startup phase of a glass
melting furnace, all ancillary processes such as raw material handling
and transport, forming and finishing, conveying, packaging, and fuel
storage must be tested under operating conditions prior to the introduction
of batch into the furnace. Electrical systems and mechanical equipment
are tested initially on an individual basis and then as integral parts
of the process operation. Instrumentation and control panels are also
debugged during this phase. In addition, various calculations are
performed to determine the exact position to place the forming machines
to ensure proper alignment after furnace expansion has occurred.
Preliminary testing and debugging may last from 1 to 3 months.
Once "pre-startup shakedown" has been completed the "seasoning" or
drying out of the furnace begins. New refractory consisting of a high
alumina, zerconia, and silica composition contains residual amounts of
water that must be driven off before batch can be fed into the furnace.
Seasoning requires the placement of supplemental natural gas burners at
various locations along the refractory wall. The drying phase is a
gradual warming of the bricks lasting about 2 to 3 days at a temperature
between 100 to 150°F. The furnace is constantly being monitored during
the drying phase and throughout the entire operation to ensure that the
refractory temperature is increasing uniformly. If the refractory heats
up too fast or a situation develops where there is a nonuniform heat up,
internal stresses can loosen or crack the refractory wall.
As the drying phase nears completion, the temperature is gradually
increased to between 2,000 to 2,700°F. During this heatup phase two
critical movements occur within the refractory lining. At a temperature
of about 1,400°F the first expansion takes place and at approximately
2,000°F the second expansion occurs. The two movements result in a
total expansion of the furnace of approximately six inches. Horizontal
and vertical tie rods located within the network of the refractory
lining may have to be adjusted during the heatup phase to allow for the
UT1
-------
expansion. Approximately 48 hours are required to successfully "heatup"
the furnace.
Once the furnace has reached the desired temperature (approximately
2,700°F) and if all equipment checks are satisfactory the supplemental
burners are removed, a fine layer of cullet is added to the furnace and
the furnace burners are brought on line. Gullet is used as the initial
ingredient because it requires a relatively minimal amount of energy to
get it to a molten state.
Once the molten cullet reaches a predetermined level in the furnace,
mixed raw ingredients are fed to the furnace. The molten glass level
rises to a point and then flows through a drainage bushing where the
glass is returned to the system as cullet. This process lasts several
days while glass quality checks and equipment adjustments are performed.
When the glass has met the desired quality specifications and if all
auxiliary equipment checks out, the drainage bushing is closed and
molten glass is routed to the forming operation.
Startup Operation
Startup for glass manufacturing operations is generally considered
to be the point at which glass is produced (refer to Section 60.296,
Test Methods and Procedures of the promulgated Regulation). Although
the desired production rate may take several weeks or even months to
attain, it is during this period that substantial quantities of particulate
matter is being generated.
Once startup has begun, an intensive period of equipment shakedown
takes place. Process operations must be synchronized, conveyor and feed
systems are adjusted, instrumentation is rechecked under actual operating
conditions and air pollution control equipment is monitored. Depending
upon the extent of the equipment debugging, fine tuning, and the designated
percent of capacity at which the line is to operate at, production for
market may take several weeks or possibly months to attain. Extended
delays would result from malfunction or failure of auxiliary equipment
such as conveying systems, material and fuel feeding systems, and
instrumentation. Unless major process equipment malfunctions or fails,
a 180 day startup period, allowing for equipment shakedown prior to the
required performance tests, provides sufficient time to reach desired
product quality and production rates.
C-5
-------
Glass Produced
Introduction
For the complete definition of "glass produced" refer to Section 60.291,
entitled, "Definitions" of the promulgated Regulation. The following are
possible guidelines which may be used when determining the amount of glass
produced. Each method discussed is considered to be accurate and currently
in use throughout each category of the glass manufacturing industry.
Flat Glass
The method employed by the flat glass industry to determine the
amount of glass produced involves removing and inspecting sections of
glass taken from the glass ribbon immediately after cooling. From this
section density calculations and thickness measurements are made. These
measurements, together with the known width of the ribbon and the speed
at which the line is moving provides an accurate accounting of the amount
of glass produced.
Fiberglass
The fiberglass industry uses what is commonly referred to as a
"dipper" to determine the amount of glass produced. The procedure
involves removing a spinner from the forming process and inserting the
dipper under the bushing. The dipper, which is a rectangular cup of a
known volume captures the molten glass flowing out of the bushing. The
time required to fill the dipper is recorded as is the weight of the
glass in the dipper. This procedure is used to determine the glass flow
rate out of the bushing. The flow rate is then multiplyed by the number
of bushings per melting furnace to determine the total amount of glass
produced per furnace per unit of time.
Container Glass
In the container glass industry, the amount of glass produced per
melting furnace is determined by removing a representative sample container
from the line, weighting it, then multiplying that by the total number
of molten glass "globs" entering each forming unit per unit of time.
Pressed and Blown Glass
The method employed by the pressed and blown glass industry to
determine the amount of glass produced per melting furnace is essentially
idenitcal to the method used by the container glass industry. A
/-
-------
representative sample is removed from the line, weighed and then
multiplied by the total number of globs entering each forming unit per
unit of time.
C-7
-------
-------
Addendum to the Glass Manufacturing
Plants Background Information: Proposed Standards
of Performance (EPA-450/3-79-005a)
D-l
-------
Addendum to the Glass Manufacturing
Plants Background Information: Proposed Standards
of Performance (EPA-450/3-79-005a)
Table 4-1 of the Background Information Document, Volume I has
been revised for emission test reference numbers (ETRN) 16, 17, 18,
and 19. The mass emissions for ETRNs 18 and 19 were changed as well
as the particulate concentrations for ETRNs 16, 17, 18, and 19.
Table 4-2 of the Background Information Document, Volume I has
been revised for ETRNs 24, 25, 26, 27, and 28. The mass emissions and
the particulate concentrations for these ETRNs were changed.
Table 4-3 of the Background Information Document, Volume I has
been revised for ETRNs 32, 33, and 34. The mass emissions for ETRNs
32 and 34 were changed as well as the particulate concentrations for
ETRNs 32, 33, and 34.
Table 4-4 of the Background Information Document, Volume I has
been revised for ETRNs 38 through 55, inclusive. The mass emissions
for ETRNs 38, 39, 40, 41, 44, 45, 46, 48, 50, 51, and 52 were changed.
The particulate concentrations for all of the ETRNs except 43 and 50
were changed. ETRN 41 was changed to 41a and 41b. and ETRN 56 was
added.
There are no additional associated impacts as a result of these
ETRN revisions.
Table 8-7 of the Background Information Document, Volume I has
been revised to" properly present the flow rates for the Pressed and
Blown (Borosilicate, Opal, and Lead) industry segment. The flow rates
for the 100 ton per day model plant is corrected to 906 m min.(32,000
acfm) and the flow rate for the 50 ton per day model plant is corrected
to 453 m3 min.(16,00 acfm). These flow rates are identical for the
respective model plant sizes for the three new Pressed and Blown
sub-categories (Borosilicate, Soda-Lime and Lead, and Other-Than
Borosilicate, Soda-Lime, and Lead) used in the promulgated regulation.
The effects of this correction in flow rates will be a slight
decrease in costs. This will result in a minimal economic impact to
the industry.
D-2
-------
to
H-
CO
LU
I
to
o
II
to
Lu
5
O
ii
a;
a.
LU
o
Zi
C£
LU
to
to
a:
CJ
LU
LU
i
cu
.a
to
to
£=
o
r
to
to
r-
LU
CU
-U
to
r
O
r
-M
fO
O_
cu
o
10
S-
3
u_
o
^*
S-
o
cu
LU
1
i
r~ d
I-5 CU
£- 0
(Q 23
Q- 0
O
to
o
r-
to
to
E
LU
to
t/)
tO
s:
*^-*^
U-
C_3
OO
CD
S-
cn
**+
H-
^
" *
O)
J^
^^^
O)
cu
Q.
1/7
to
to
>> >>
to +-> o
to in en
to 3 cu
i T3 !->
cc .c id
l H C-3
CU
E O *
0 E S.
i- +J CU CU
to to s- J3
in CU CU £
r-l4-3
cu -z.
LU Q;
If)
o
o
0
*~;
LO
o
*
o
cu
4-3
«
CU 0
I !-
i to
(O O
-o s-
O 0
tO DO
to
to
03
O)
r- CO
O J3
0 T-
3 u-
vo
f«
o
^J-
1
t^
O
CU
4J
n3
CU U
_J -i-
i to
to o
a s-
O 0
tO CQ
to
to
03
r~~
S-
r CU
0 -Q
O -i-
2 Lu
r^
oo
o
en
r
O
.
CU
to
cu o
1 !-
1 (/>
to o
-o s-
o-o
tO CQ
to
to
CD
i CO
o .a
OT-
12 Ll_
CO
!"
CM
CM
CM
__
^.v
,
cu
'11
to
-a
0
to
s_
cu
E
T
res
0
<7>
s-
cu
+J
a.
03
o
cu
-C
^-
o
c:
cu
cu
-C
10
a
cu
cu
S-
to
cu
o
c
cu
I
cu
cc
*
D-3
-------
to
Ul
CJ
ce:
co
UJ
E
CO
CD
O
CO.
UJ CO
DC o:
UJ
UJ 1
H-l Qi
CO CO
CO UJ
co a.
t-t D-
O, UJ
CM
(U
(O
E
O
i
r
UJ
CO
to
r»
3
U
r-
S-
o.
CO
4J
3
O
S-
0)
JJ
tZ
o
r
S-
(O
u.
c
CO 0
4-5 V-
tO 4-5
i tO
3 S-
O 4-5
r- E
4-5 CO
S- 0
tO E
0- O
o
E
O
I
to
r
1§
to
co
* ^
u.
co
Q
^^
s-
01
^^*
r
£3-
j?
O)
* '
CO
tor 04J"
r td E E
3 > CO CO
0 0-r- 0
r- S 0 $-
4-> CO « CO
yj * CM CM r-^
op
CO r r <£) CM
» "
CM CO
r**» *
^3r
0.
UJ
.G o
s- -P-I-
i- O 4->
«=C t (C
IT)
vo
LO
CO
LO
o
CO
+J
CD
+J
(1)
to
to co
tO O-
01-
>> >>
to 4-» O
c/1 tO O)
CO 3 O)
r--0 4-5
OJ E fO
HH 0
CO
E O *
^ r- t
CO
E
r
1
1
fO
-o
o
CO
.. -a
E tO
O r
CQ "O
E
-o to
c
CO CO
-a T-
O> r
tO 1
to CO
cu -a
s_ o
Q- CO
I CD
-a o
CO -r-
CO r
i to
CO O
"O &-
O O
CO JD
m m
E
0
DO
CU
"O 4-5
E fO
CO CJ
T3 r
CO !-
t/) 10
t/J O
CO 5-
S- 0
O. CQ
CD
O
r
J~
O
O
CQ
W
CO
CO
r~
s-
r" I
I 10
fO O
a s-
O 0
CO CQ
to
to
CO
r
<~
CO
-Q
r-
U-
"o
o
3
»«/
CU U
E "r"
""" J
1 (/)
(0 O
a s-
o o
CO CQ
to
^0
-------
co
a:
UJ
oa
ca
ZD
a;
o
co
UJ
a
UJ
a.
a.
co-
co
UJ
CJ>
o:
co
co
=c
CD
co
co
UJ
co
UJ
CO
CO
UJ
a.'
co
a>
CO
E
O
1
CO
CO
E
UJ
cu
4~*
to
o
S-
tO
CL.
CU
4->
O
S-
cu
>->
S-
0
co
1
I
s-
3
E
CU
cu o
4-> -r-
fO 4-3
to
3 i-
0 4->
r- E
4-> CU
S- O
fO E
CL. O
CJ
CO
o
CO
CO
'r^
UJ
CO
CO
to
s:
^-^>
u_
0
co
^s^
s-
a>
S^. f
t
r~
(~)
r«-
*
O5
**s.
"**^
O)
cu
(0 i O+j"
Of E E
3 > CU CU
O O -i- O
r- £ 0 S-
4J CU < CU
S- DC Jj-,0^,
CL. UJ
CO
CO CU
to o.
CD i
J_ £_
CO 4-> O
co co en
tO 3 CU
i -a 4->
CD E to
i i C_>
cu
E O *
O 4-> E S-
r- CO tt) CU
CO CU S- .Q
co 1 cu E
,. 1 Ij^, ^
E CU 2:
UJ DC
OJ CO O CO
^t- CM i CO
O O O O
.
<* O CO i
r^. CM CM co
1*^ O *si" ^O
CO r t «
LO CO
»
CM cy>
co r--
CU CU CU CU
t r r r
1 1 1 1
to to to to
-o -a -a -a
o o o o
co co co co
S-. S- S- S-
CU CD CU CU
E E E E
r -I 'I1 »r
fO to (O tO
4-* ) ^ | ^ | ^
E E E E
0000
CJ C_> CJ O
i CM CO *
CO CO CO CO
i
S-
-------
LU
to cc
to o
r-tO
too
o 3
>-«C3'
j_UJ
<
CU LU
.1
c c
D O) O
I/I CO -P O
c/) r (O to
T- 3 S- Q
E 0-P >
LU -r C S-
.p 01 CD
0) I- CJ «
4-> 10 C
IO O- O
r 0
U
S- 1
CO
Q.
OJ tn
r C 1
P O **
3 - .n
o tn i
10
I. f-
O E
4J LU
JS « o>
,fn, yj _Vf
^L (O ***
i- £ CD
o
01
s_
o_
01
(O r~~ O "P
i ra c c
3 > Ql Ol
CJ Q »- O
i- S U S-
IO
a o
e cn
cn id
tn o
CO
r--
CO
P
tn
01 Ol
( cj *
c s-
C CO CO
O i~ ^3
1- Ol E
«>)- S
I/I O) Z
r-o:
LO co r^
CJ O O
CJ O O
0 0 *
LO LO t^
o o o
^_
Ol
CO tO
CO i
tD r
* ^ c.
00 1^-
je CO CO
* r- CM
1 i 1
T i i
T3 T3 13
o o o
CO to VI
S. S- S-
01 01 01
c c c
10 10 10
'c c c
o o o
000
co cn o
CO CO «=T
CM
o
o
cn
o
in
0
0)
10
CJ
r-
t/1
O
S-
o
CO
cr
o
CO
01
-o (->
cz 10
ca u
Pressed
Borosili
IO
CO
CM
o
CM
-
01
p
id
u
U)
g
o
ca
c
o
ca
01
"O -P
c: id
ro u
Pressed
Borosili
5
co
o
2
i
cn
CO
LcT
to
o
CO
CO
01
rs
u
tn
o
o
CO
c
3:
o
E IO
id u
a i^-
Ol -i-
tn tn
tn o
01 S-
S- 0
a. ca
CM
o
o
to
cn
CO
CO
CO
^.
o
cn
CM
01
S
f
tn
O
s_
o
ca
2~
o
"a -p
c 10
ro u
Pressed
Borosili
CO
3-
CO
o
in
§
CU
id
CJ
o
o
CO
c
o
ca
01
-o -P
c ro
10 u
Pressed
Borosili
3-
r
o
o
cn
CM
in
CO
cn
r
cn
CO
a.
o
01
si
o
LU
C
o
ca
T3
c
Id
T3
01 f.
tn 01
tn s:
01 -P
S- O
O--
LO
a-
D-6
,
CM
O
O
^
to
o
CO
LO
cn
o
^
CO
CO
CM
"O
id
01
sr id
3 01
O 1
CO XI
c
xi 10
10 01
XI *r
01 i
tn i
i/} ro
£"§
a. in
3
CO
0
o
to
CO
0
1 ,
cn
in
^2-
CO
CO
IO
01
- c id
3 01
O 1
ca xi
xj 10
id 01
XJ T-
01 1
tn i
tn ro
Ol XI
S- O
o. to
3-
CO
to
o
CO
o
XI
ro
c
1
c 10
O _J
ca xi
XI 10
c
id ai
E
t/l 1
in ro
01 XI
s- o
a. to
co
CO
o
0
2:
0
C1
to
CO
^,
CO
CO
XI
id
01
fro
CU
1
CO XI
XI IO
c
ra 01
XI !-
01 r
tn I
tn ro
01 XI
S- 0
a. co
cn
=!
-------
OO
oo ee
oo o
- _
LU
oo az
t'CL.
;s
003:
-> I
LiJ O
H- LU
30.
a.
,
=t OO
a. LU
.1
a- c£.
i =>
«3- U-
to
C c
o (J
to i <8 00
f| 3 i. 0
UJ -i- C S.
M 01 en
0) s- o ^^
-W It) C
U
It)
Q.
.01 in ^~~
*J o ^.
3 V- .Q
o to i
0) v_-
M LU
*-> to 'S?
r- to -i£
Q. m ^
c s s
a) a>
O O i- t- a>
s-a:<*- a,
10 >!-
a. LU
O S_> O)
oo 1
c c en
0) D) C
O !-!-
S- to S-
a) a> 3
o. o o
U.
CJ
ro oo
Sc\T
U Lu
ai o to
a. a)
OO i CO
o!
0
OJ
a>
to
g,
o o
£= cn
I-H a)
*j
to 10
to
It)
to
0) (!)
1 (J *
E 01 01
0 S- J3
f- 0) E
t/J *4- 3
to OJ S
r- Q£
J
«* LO tD CO LO ^ CO
O t r O O r
O O O O O O O
r-* LO Cn C\l Cn t*^ LO
CO LO r r*. r r CO
VO CO O - CO
r- CM i CO O O r-
, C3 CM
CO CM
CM CM «^- LO CM
O^ i CD CD CO
*-' --- - ~ -
LO F"^ CO CM VD
cn co- CM CM i
r CM CM CM CM
CO
CU
1 O) CU CU
O U U O CU
in t/t t/j c/i i
"O "D ro O O O (O
as ro 4-> s- s- s- -a
cu cu o o o o o
_J -J CL, CQ CQ CQ CO
.. -a .. -a -a
' c fO a to c to
3 cu s i t/i i
cu -a cu -a -cu-a *© "o 'o ro
S-O S-O i-O O O Or
Q_CO D-CO Q_tO 3 3 3 U_
CO i CM OO ^f LO (O
LO LO LO LO LO LO LO
a
IO
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO
EPA 450/3-79-005b
2.
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Glass Manufacturing Plants - Background Information for
Promulgated Standards of Performance
5. REPORT DATE
September 1980
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO.
9= PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Protection Agency
Office of Air Quality Planning and Stnadards
Emission Standards and Engineering Division
Research Triangle Park, N.C. 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
12.SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
DAA for Air Quality Planning and Staridards
Office of Air, Noise, and Radiation
U,S, Environmental'Protection Agency
Reasearch Triangle Park, N.C. 27711
14. SPONSORING AGENCY CODE
EPA/200/04
16. SUPPLEMENTARY NOTES
This document is the second volume of EPA 450/3-79-005 series. The first volume
discussed the proposed standards and the resulting environmental and economic impacts,
10. ABSTRACT
Standards of performance are being promulgated under Section 111 of the Clean
Air Act to control particulate matter emissions from new, modified, and reconstructed
glass manufacturing plants. This document contains a detailed summary of the public
comments on the proposed standards (44 FR 34840), responses to these comments and
a summary of the changes to the proposed standards.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COSATI Field/Group
Air Pollution
Glass Manufacturing and Processing
Emission Standards
Air Pollution Control
13 B
18, DISTRIBUTION STATEMENT
Unlimited - Avaialble to the public free
of charge from : U.S. EPA Library (MD/35)
THangTp Park. N.T- 97711
19. SECURITY CLASS (ThisReport}
unclassified
21. NO. OF PAGES
175
20. SECURITY CLASS (Thispage)
unclassified
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION is OBSOLETE
-------
-------
8I.S
.?£.
III
!§§
fii
CD O O
D) O (D
§.3,3.
Ifi
CD 0)'
------- |