United States Office of Air Quality-
Environmental Protection Planning and Standards
Agency Research Triangle Park NC 27711
Air ~~~
EPA-45O/3-82-O2Ob
August 1984 •
-------
-------
EPA-450/3-82-020b
Electric Arc Furnaces and Argon-Oxygen
Decarburization Vessels In Steel Plants-
Background Information for Promulgated
Standards
Emission Standards and Engineering Division
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Radiation
Office of Air Quality Planning and Standards
Research Triangle Park, North Carolina 27711
August 1984
-------
This report has been reviewed by the Emission Standards and Engineering Division of the Office of Air
Quality Planning and Standards, EPA, and approved for publication. Mention of trade names or
commercial products is not intended to constitute endorsement or recommendation for use. Copies of
this report are available through the Library Services Office (MD-35), U.S. Environmental Protection
Agency, Research Triangle Park, N.C. 27711, or from the National Technical Information Services,
5285 Port Royal Road, Springfield, Virginia 22161.
-------
ENVIRONMENTAL PROTECTION AGENCY
Background Information
and Final
Environmental Impact Statement
for Electric Arc Furnaces and
Argon-Oxygen Decarburization Vessels in Steel Plants
Prepared by:
R. Farmer
Director, Emission Standards and Engineering Division
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
/ (rate)
1. The promulgated revised standards of performance continue to limit
particulate matter emissions from an electric arc furnace and argon-
oxygen decarburization vessels. The visible emission standard is less
than 3 percent from a control device, less than 6 percent from the
shop, and less than 10 percent from the dust-handling system.
• 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 significantly to air pollution which may reasonably be
anticipated to endanger public health or welfare." Stee.l plants are
located in all areas of the nation.
2. Copies of this document have been sent to the following Federal
Departments: Labor, Health and Human Services, Defense, Transporta-
tion, Agriculture, Commerce, Interior, and Energy; the National Science
Foundation; the Council on Environmental Quality; members of the State
and Territorial Air Pollution Program Administrators; and Association
of Local Air Pollution Control Officials; EPA Regional Administrators;
and other interested parties.
3. For additional information contact:
Mr. C. Douglas Bell .
Standards Development Branch (MD-13)
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Telephone: (919) 541-5578
4. Copies of this document may be obtained from:
U. .S. EPA Library (MD-35)
Research Triangle Park, North Carolina 27711
Telephone: (919) 541-2777
National Technical Information Service
5285 Port Royal Road
Springfield, Virginia 22161
111
-------
-------
TABLE OF CONTENTS
Section Page
1 SUMMARY 1-1
1.1 Summary of Changes Since Proposal 1-2
1.2 Summary of Impacts of Promulgated Amendments 1-3
1.2.1 Alternatives to the Promulgated Action 1-3
1.2.2 Environmental Impacts of the Promulgated
Action 1-3
1.2.3 Energy and Economic Impacts of the Promulgated
Action 1-3
1.2.4 Other Considerations 1-3
1.2.4.1 Irreversible and Irretrievable
Commitment of Resources • . . 1-3
1.2.4.2 Environmental and Energy Impacts of
Delayed Standards 1-4
1.2.4.3 Urban and Community Impacts 1-4
2 SUMMARY OF PUBLIC COMMENTS 2-1
2.1 Test Methodology 2-1
2.1.1 Mass Emissions 2-1
2.1.2 Visible Emissions 2-11
2.2 Emission Limits 2-17
2.2.1 Mass Emission Standard 2-17
2.2.2 Visible Emission Standard 2-19
2.3 Costs of Testing 2-26
2.4 Miscellaneous 2-27
IV
-------
-------
LIST OF TABLES
Table
2-1 List of Commenters
Page
2-2
-------
-------
1. SUMMARY
On October 21, 1974 (39 FR 37466), standards of performance were
proposed under Section 111 of the Clean Air Act to control particulate
matter emissions from electric arc furnaces (EAF's) in the steel industry.
Standards of performance were promulgated on September 23, 1975
(40 FR 43850), and apply to any facility constructed, modified, or
reconstructed after October 21, 1974. Under the Clean Air Act amendments
of 1977, standards of performance must be reviewed every 4 years and
revised, if appropriate. On April 21, 1980, a notice was published in
the Federal Register (45 FR 26910) announcing such a review of the
standards of performance for EAF's in the steel industry. The review
found that fugitive emissions capture technology had improved since
promulgation of the existing standards of performance for EAF's. Another
finding was that argon-oxygen decarburization (AOD) vessels are a
significant source of particulate matter emissions in specialty steel
shops. As a result of these findings, additional data were collected on
the controlled emission levels from EAF's and AOD vessels to determine
how the standards should be revised.
Revised standards of performance were proposed on August 17-, 1983.
The period for public comment extended through October 31, 1983, and
seven written comments were received. A public hearing was requested by
one individual, but this request was later withdrawn, and no public
hearing was held. The comments on the proposed standards, together with
responses to each comment, are presented in this document. The comments
and responses serve as the basis for the revisions that have been made
to the proposed standards.
1-1
-------
1.1 SUMMARY OF CHANGES SINCE PROPOSAL
In response to public comments, certain changes have been made in
the proposed standards, and the more significant changes' are summarized
below. All changes that have been made to the regulation are explained
fully in the responses to the comments.
For sources built between October 21, 1974, and August 17, 1983,
Section 272(a)(3)(iii), which was in the original regulation but was
changed in the proposed revisions, is reinstated in the promulgated
regulation. The related sections 274(a)(3), (a)(4), (b), (c), (e), and
(f) of the original regulation are also reinstated. Sections 274(b) and
(c) have been revised, and Section 274(e), (f), and (g) have been
redesignated (f), (g), and (h). These sections require continuous
monitoring of the flow rate through each capture hood and-the pressure
in the free space inside the furnace and require that the monitored flow
rates and pressure be maintained at baseline levels established during
the most recent performance or compliance test.
Modular, multiple-stack, negative-pressure baghouses have been
included with 'positive-pressure baghouses as control devices that may be
monitored by Reference Method 9 observations in lieu of transmissometers.
Sections 275(i) and 275a(c) have been revised to make it clear
that, where it is possible to determine that visible emissions at
multiple sites are attributable to only one incident of the visibl-e
emissions, one set of Reference Method 9 observations from the point of
highest opacity that directly relates to the cause (or location) of the
incident will be sufficient.
In addition, because of the Agency's continued standards development
work, several other changes have been made in the standards. Both
Subparts AA and AAa are revised to permit either periodic monitoring and
recording of fan motor amperage and damper position or continuous moni-
toring and periodic recording of flow rates. In Subpart AA, if fan motor
amperage/damper position monitoring is the chosen alternative, the monthly
operational status inspections that were proposed will be required.
Sections 275(a)(l) and 275a(a)(4) have been revised to make it clear
that only Reference Method 5 is to be used on negative-pressure fabric
1-2
-------
added to Subpart AA. This section requires that when the "baseline"
monitored values are outside of acceptable ranges, these values must be
reported semiannually. To be consistent with Subpart AA, Subpart AAa
has been revised to require establishment of these same "baseline" values.
Semiannual reporting of values outside of the specified ranges is also
required for Subpart AAa. Both Subparts AA and AAa have had a provision
added to clarify the requirements in Sections 275(g)(2) and 275a(h)(2)
of acceptance by the Administrator. When utilizing a performance test
method that compensates for the emissions from the facilities not subject
to the provisions of the standards, the Administrator must be notified
of the method to be used 30 days prior to the performance test and must
approve the method.
1.2 SUMMARY OF IMPACTS OF PROMULGATED AMENDMENTS
1.2.1 Alternatives to the Promulgated Action
The regulatory alternatives are discussed in Chapter 6 of Volume I
of the background information document (BID) for the revised standards
(EPA-450/3-82-020a). These regulatory-alternatives reflect the different
levels of emission control that were analyzed in determining best
demonstrated technology, considering costs, nonair quality health,
environmental, and economic impacts for EAF's and AOD vessels in steel
plants. These alternatives remain the same.
1.2.2 Environmental Impacts of the Promulgated Action
The environmental impacts resulting.from the revised standards are
described in Chapter 7 of Volume I of the BID. These impacts remain the
same.
1-2.3 Energy and Economic Impacts of the Promulgated Action
Energy and economic impacts resulting from the standard are
discussed in Chapters 7 and 9, respectively, of Volume I of the BID. No
changes in these impacts have occurred since the standards were proposed.
1.2.4 Other Considerations
1-2-4.1 Irreversible and Irretrievable Commitment of Resources.
The regulatory alternatives defined in Chapter 6 of Volume I of the BID
would not preclude the development of future control options nor would
they curtail any beneficial use of resources. The alternatives do not
involve short-term environmental gains at the expense of long-term
1-3
-------
environmental losses, and the alternatives yield successively greater
short- and long-term environmental benefits. Further, none of the
alternatives result in the irreversible and irretrievable commitment of
resources. No change in these considerations has resulted since proposal
of the standards.
1.2.4.2 Environmental and Energy Impacts of Delayed Standards. As
discussed in Chapter 7 of Volume I of the BID, delay in the revised
standards would cause a similar delay in realizing the beneficial impacts
associated with the standard. No changes in the potential effects of
delaying the standards have occurred since proposal of the revised
standards.
1.2.4.3 Urban and Community .Impacts. Urban and community impacts
of the standards are considered under economic impacts in Chapter 9 of
Volume I of the BID. No changes in these impacts have occurred since
the standards were proposed.
1-4
-------
2. SUMMARY OF PUBLIC COMMENTS
A list of commenters and their affiliations is presented in Table 2-1.
Seven individuals representing two steel companies, two local government
agencies, two trade associations, and one private citizen submitted
written comments during the public comment period.
The comment letters often contained several comments. Each comment
is addressed separately, and the commenter is identified'by the appropriate
docket number.
2.1 TEST METHODOLOGY
2.1.1 Mass Emissions
Comment: One commenter (IV-D-6) states that Reference Method 50 is
the most practical test method under the circumstances.
Response: No response is necessary.
Comment: Commenter IV-D-4 considers Reference Method 5D to be
unacceptable because the Environmental Protection Agency (EPA) has only
shown that Method 5D yields results,similar to those obtained by Method 5;
however, the EPA has not shown that Reference Method 5D is accurate and
precise. The commenter cites Portland Cement v. Ruckelshaus (486 F.2d 375)
and Chrysler v. DOT (472 F.2d 659) for the requirement that the test
method used should be objective, repeatable, and precise.
Another commenter (IV-D-7) also believes that the accuracy and
precision of any proposed compliance sampling method should be demonstrated.
Response: Method 5D is an adaptation of Method 5 for the emission
testing of positive-pressure fabric filters. The sample collection and
analysis procedures specified in Method 5D are identical to those in
Method 5, and, therefore, the Agency expects the precision of Method 5D
2-1
-------
6
7
TABLE 2-1. LIST OF COMMENTERS FOR THE REVISION TO THE
NSPS FOR EAF'S AND AOD VESSELS IN STEEL PLANTS3
No.
1
2
3
4
Commenter
Steel Bar Mills Assoc.
(0. A. Dulle, Jr.)
C F & I Steel Corp.
(B. D. Egley)
National Steel Corp.
(J. G. Manda)
American Iron and Steel
Date of comment
9/2/83
10/14/83
10/28/83
10/27/83
Docket entry No.
IV-D-1
IV-D-2
IV-D-3
IV-D-4
Institute
(E. F. Young, Jr.)
Allegheny County (PA) 10/24/83
Bureau of Air Pollution
Control
(R. J. Chleboski)-
Donald L. Shepherd 10/28/83
Illinois Environmental 10/28/83
Protection Agency
(D. J. Goodwin)
IV-D-5
IV-D-6
IV-D-7
aCopies of all correspondence received from commenters appear in
Docket A-79-33.
2-2
-------
to be similar to that of Method 5. The absolute accuracy of a particulate
emission sampling method is undeterminable, but the precision of Method 5
results have been demonstrated over many years of testing. Method 50 is
discussed in greater detail on page 2-5.
The issues of Reference Method 5D repeatability, objectivity, and
precision were resolved during development of the emission limit for
this regulation. Method 50 was used as the-field test method for acquiring
the data for this regulation, and this method was applied to several
different fabric filter configurations. (See BID, Vol. I, Appendix C for
descriptions.) The precision of the results of these tests was considered
in developing the emission limit for this standard.
Comment: Commenter IV-D-4 believes that Reference Methods 5 and 5D
are less appropriate than high-volume sampling for the testing of positive-
pressure fabric filters. This commenter states that the high-volume
sampling method has been used predominantly to measure the emissions
from EAF fabric filters. However, the EPA has proposed that Reference
Method 5 or 5D should be used on positive-pressure fabric filters. The
EPA supported this change by presenting the results from two tests using
Reference Method 50 that show higher particulate matter concentrations
than those obtained with the high-volume sampling method. The commenter
states that the EPA has not clearly shown which of these methods is correct.
The commenter believes that the high-volume sampling method is better
suited to positive-pressure fabric filter exhaust conditions because it
is designed to operate at low air velocities, and it collects more
sample weight than is collected by Method 50.
Two other commenters (IV-D-2 and IV-D-5) believe that high-volume
sampling is an appropriate test methodology for testing positive-pressure
fabric filters.
Commenter (IV-D-5) believes that, for pressurized fabric filters
with gas flow rates under 100,000 acfm, high-volume sampling results in
emission data that are about 20 percent lower than Method 5 sampling
results. At a flow rate of 100,000 acfm or greater, direct testing of
exhaust stacks becomes economical.
Response: The Agency proposed Method 5D instead of high-volume
methods for sampling of positive-pressure fabric filters for two reasons.
2-3
-------
First, the Agency conducted simultaneous comparison tests using both
Method 5 equipment and high-volume samplers. The data obtained from
these tests show that the high-volume particulate concentration results
were 70 to 85 percent lower than those indicated by the Method 5 equipment
on emissions from a positive-pressure fabric filter. Results of other
comparisons between the two methods, both direct and indirect, also show
that high-volume sampling methods produce results lower than Method 5 or
Method 5D (docket entry IV-A-1).
A subsequent conversation with Commenter IV-D-5, discussing a
20 percent difference between high-volume sampling and Method 5, dis-
closed that this was a subjective estimate and was not based on any
direct comparisons (docket entry IV-E-4).
Second, as is discussed further on page 2-5, the Agency has determined
that it is necessary to use demonstrably reliable equipment and multipoint
sampling to assure a representative collection of particulate emissions
from most emission sources, including fabric filters. Method 5D incorporates
the multipoint sampling requirements with the use of reliable Method 5
equipment to provide a practical method for testing positive-pressure
fabric filters."
Comment: One commenter (IV-D-2) states that the diagrams presented
on'pages 37356 and 37357 of the Federal Register proposal notice are
oversimplified and not typical of the fabric filter at .the commenter1s
facility, which has 32-compartments and does not have a ridge vent roof.
The commenter suggests using high-volume sampling as an alternative
to Reference Method 5D, thus reducing the number of sampling stations
(in the case of his fabric filter, from 32 to 2). The commenter believes
that the costs of performance testing would range from about $10,000 to
$24,000, for a Reference Method 5 series of three test runs. He states
that high-volume sampling would cost $5,000 to $8,000.
Another commenter (IV-D-4) expresses concern that Reference Method 5D
could be used by enforcement personnel for any roof monitor exhaust for
which other test methods have not been specified, i.e., shop roof exhausts.
Therefore, the commenter recommends the high-volume sampling method;
alternatively, Reference Method 5D should be forbidden under
40 CFR 52.12(c)(l) for State implementation plan (SIP) enforcement.
2-4
-------
Response: As noted in the preamble to the proposed revisions
(48 FR 37348-9), because the General Provisions (40 CFR 60.8[e]) require
that all control devices be testable, some States have been requiring
affected facilities controlled with positive-pressure fabric filters to
undertake the expensive retrofit of stacks or stack extensions onto the
fabric filter for testing purposes. This situation was brought to the
Agency's attention during development of this revised NSPS. Reference
Method 5D was developed in response to this situation to ensure the
availability of uniform test procedures for positive-pressure fabric
filters.
Method 5D was proposed because the Agency recognized the
complications involved in testing the many different configurations of
positive-pressure fabric filters. The procedure section of Method 5D
addresses most of the sampling issues for these tests including examples
of the application of the method to multicompartment fabric filters.
The requirement for multipoint sampling.for particulate concentration is
necessary to achieve representative results because of the possibility
of stratification of the gas stream occurring within a positive-pressure
fabric filter. This stratification is caused by incomplete mixing of
the gas in the fabric filter and results in varying particulate
concentrations across the exhaust gas profile of the fabric filter.
Method 5D represents a reasonable balance between the multipoint sampling
requirements necessary to achieve representative results and a
recognition of the practical problems of testing the many different
outlet configurations found on positive-pressure fabric filters.
Method 50 is a procedure based on the EPA Reference Method ..5 to
accommodate testing of emission sites that do not conform to conventional
exhaust configurations and is appropriate for testing positive-pressure
fabric filters. The commenter presumes that application of the high-
volume method would reduce the number of sampling locations required for
a test and concludes that the costs of testing would be greatly reduced
with the use of high-volume methods. The Agency has determined that
there is a need for multipoint sampling and for testing a representative
number of sampling locations. Use of high-volume sampling equipment
would not preclude these requirements. Applying the multipoint sampling
2-5
-------
requirements of Method 5D to high-volume sampling may actually be more
expensive than testing with Method 5 equipment under the same require-
ments. A greater number of samples might have to be analyzed because
more than one filter catch might be necessary for each-sample location
when using a high-volume sampling train. Additional equipment, such as
a greater number of high-volume samplers and volume measurement devices,
might also be necessary.
The commenter (IV-D-2) who cites the example of a 32-compartment
fabric filter with two end-vents apparently assumes that such a fabric
filter would require 32 sample locations using Method 5D. Section 4.2.3
of Method 5 or 5D indicates that, if all 32 compartments are potential
measurement sites (i.e., no roof monitor or stack exists), 12 compartments
or 50 percent of the total number of compartments, whichever is greater,
must be tested. For the commenter1s example, a maximum of 16 sites, not
32, need to be tested following Method 50. However, six measurement
sites would need to be sampled per test run, in keeping with the require-
ment that the same number of sites be sampled for each test run. It
should.be noted that, if the two end-vents comply with the stack con-
figuration requirements in Method 50; the number of test sites could be
reduced from 16 to 2.
The basis of the commenter's estimated costs of a performance test
are not explained, and those costs presented in the proposal's preamble
are believed to be correct. The costs to conduct a performance test
according to Reference Method 50 are explained in detail in Appendix 0
of Volume I of the BID. These costs are based on the assumption that a
test may require between 4 to 8 person-days of field work. The costs of
$5,000 to $8,000 include the test report.
The applicability of Reference Method 50 was explained at proposal.
Uses of Method 50 other than on the exhaust of the positive-pressure fabric
filter were not endorsed and no data exist to support use of the Method
on other sources.
Comment. A commenter (IV-D-2) states that determining a velocity
profile, as required by Reference Method 2, for even a small fabric
filter such as the five-compartment fabric filter depicted in the proposal
2-6
-------
preamble, would be difficult because of the large area over the filter
compartments.
Another commenter (IV-D-7) believes that Reference Method 2 yields
inaccurate measurements of the gas stream when gas velocities are very
low.
Response: Commenter IV-D-7 is correct about measurements of
velocities outside the range recommended by Method 2. Section 4.3 of
Method 5D includes a provision for measuring inlet gas flow rates and
calculating average outlet gas velocities (using the known outlet area)
for use in isokinetic determinations for those sources whose outlet
velocities cannot be accurately measured using Method 2. It is the
Agency's intent that velocity measurements be conducted at each outlet
sampling point if the velocity head exceeds 0.05 inch of water column
(in. w.c.). However, the inlet sampling option is a reasonable
alternative when lower velocity conditions exist. Other approaches to
measuring outlet gas velocity, such as micromanometers or special low
velocity Pi tot tubes, may also be applied subject to advance approval by
the Administrator, as described at 40 CFR 60.8(b).
Comment: One commenter (IV-D-5) believes that there could be
problems calculating the gas velocity required for Reference Method 5D
sampling because leakage could occur between the inlet and outlet of the
fabric filter. The commenter states that using the average gas velocity
at the inlet, corrected to outlet conditions as specified in Method 50,
is not as accurate as using velocity measurements at each outlet site.
Another commenter (IV-D-7) questions whether it is appropriate to
accept inlet flow rate measurements to determine total volumetric flow.
Commenter IV-D-2 states that Reference Method 5D is not appropriate
for emission testing of positive-pressure fabric filters because of
possible ambient air infiltration into the fabric filter and the high
costs of sampling each compartment. As an example of problems of ambient
air infiltration, the commenter describes a basic oxygen furnace secondary
emission fabric filter at his facility where the entry level at the
filter floor is open to the atmosphere. The commenter states that, as a
result, significant dilution of the exhaust stream occurs at either the
top of the bags or at the exhaust portal. On the other hand, the commenter
2-7
-------
also describes an EAF fabric filter with man-sized doors near the ports
where a "flue" type duct situation near the doors is unlikely to occur.
Response: There is no technical difficulty in using a flow balance
approach to determine fabric filter outlet air velocity, provided that
leaks of air into and out of the fabric filter structure are identified
and sealed. The Agency addresses the issue of air leakage into or out
of the fabric filter with the note in Section 4.3 of Method 50. This
note states: "All sources of gas leakage into or out of the fabric
filter housing between the inlet measurement site and the outlet measure-
ment site must be blocked and made leak-tight." This requirement would
apply when measurements at the inlet to the fabric filter are used to
calculate average gas velocity for isokinetic determinations and when
sources of ambient air in-leakage or emission leakage exist in the
housing.
In using the words "leak-tight," it is the EPA's intent that major
air leakage sites, such as open grate floors and large access doors be
sealed adequately with relatively inexpensive plastic or fiberglass
materials. This procedure .for preventing air in-leakage has been demon-
strated during the data collection process for this regulation. The
positive-pressure characteristic of the fabric filters used in this
industry aids in identifying the location of most small air leaks.
Those leaks that can be located should be sealed prior to, and for the
duration of, the test period.
In promulgating Method 5D, the note in Section 4.3 has been moved
to the test procedures section to give it more visibility. Also, the
note has been revised to state that the steps to block the leakages
should be taken prior to the emission measurements.
Comment: One commenter (IV-D-2) states that probe supports would
be required at each sampling port because of the probe length required
to span even the smallest fabric filter module.
Another commenter (IV-D-7) also believes that problems of probe
supports or probe rigidity should be considered in the test method.
Response: Rather than specify solutions in Method 5D to each
testing problem that might arise,.site specific testing problems are
usually left to testing and shop personnel to resolve. There are a
2-8
-------
number of approaches to sampling over long traverses; for example,
diametrically opposed sampling ports will reduce probe length by half,
internally mounted cables or bars for probe supports will eliminate need
for external structures, and sampling from a roof monitor or vent
location will greatly reduce the traverse lengths. Probe length require-
ments did not pose problems in using Method 5D to acquire the data
necessary to establish the NSPS for EAF's and ADD vessels in steel
plants.
. Comment: One commenter (IV-D-5) states that it is not necessary
for the test method to require both a volume of 160 dry standard cubic
feet (dscf) and a sample weight of 50 milligrams (mg). The commenter
points out that emissions from some sources are so low that it could
become necessary to test for up to 8 hours to collect a 50 mg sample.
The commenter believes this amount of time is unreasonable, especially
for a clean source. The commenter believes that the minimum sample
volume of 160 dscf is adequate and recommends that the minimum sample
weight of 50 mg be deleted.
Response: The Agency agrees with this comment. Proposed
paragraph 60.275 of Subpart AA reflects this decision, but proposed
paragraph 60.275a of Subpart AAa mistakenly included the 50 mg minimum
catch requirement. This has been corrected in the final rulemaking by
amending paragraph 60.275a to delete the 50 mg requirement.
Comment: Commenter IV-D-7 questions whether isokinetic sampling of
the exhaust gas stream from a fabric filter i's necessary if particulate
matter emissions are in the submicron range.
Response: Particle size distribution tests on fabric filter exhausts
have shown that the mass median diameter of the outlet particles is some-
what smaller than that of the inlet particles, but it is not necessarily
a submicron particle size. For emissions from EAF operations, it is
expected that the change in mass median diameter from inlet to outlet
would be very slight. Therefore, unless the inlet particle size
distribution is submicron, which is an unlikely occurrence for EAF and
AOD vessel emissions, the outlet particle size distribution will probably
not be skewed toward the submicron range. Isokinetic sampling must be
maintained to assure a representative collection of particles greater
than 2 microns.
2-9
-------
Comment: Commenter IV-D-7 questions whether the sampling probe is
heated and asks how to accomplish such heating, if necessary.
Response: Method 5D specifies sampling at or above the stack
temperature up to a nominal 248°F. If Method 17 equipment is used, no
filter or probe heating is required. Heating of the probe and filter
for Method 5 equipment is as specified in Method 5. Heating techniques
are outlined in Method 5 reference material and include rheostat- or
thermostat-controlled nichrome wire for glass-lined probes and insulated
heating wire for metal probes.
Comment: Commenter IV-D-7 questions whether stack testing apparatus
can operate properly for the duration of a test, which is up to 4 hours.
Response: Properly designed and constructed Method 5 sampling
equipment will endure test periods much longer than 4 hours. The parts
(usually industrial grade) that constitute the sampling train are
commonly available and, with normal operation and care, should operate
satisfactorily for indefinite periods. The sampling periods for tests
in support of this regulation often exceeded 8 hours with few or no
equipment problems.-
Comment: Commenter IV-D-7 believes that design standards for
positive-pressure fabric filters should be considered in lieu of emission
testing requirements.
Response: Sections lll(h)(l) and (2) of the Clean Air Act require
that emission limits be established if it is "feasible to prescribe or
enforce a standard of performance." This phrase means "any situation in
which the Administrator determines that (A) a pollutant or pollutants
cannot be emitted through a conveyance designed and constructed .to emit
or capture such pollutant, or that any requirement for, or use of, such
a conveyance would be inconsistent with any Federal, State, or local
law, or (B) the application of measurement methodology to a particular
class of sources is not practicable due to technological or economic
limitations" (Section lll[h][2]). For this industry, feasible control
technology and test methods exist, and, thus, design standards are not
permitted in lieu of emission limits.
2-10
-------
Moreover, in most cases, it would not be possible to evaluate
design parameters alone to predict compliance with the emission limit.
This is because there are operational parameters, such as the timing and
duration of cleaning cycles or compartment air flow distribution, that
can affect fabric filter performance.
Method 5D was proposed as a practical emission testing method. It
is appropriate for measuring the emission- levels from positive-pressure
fabric filters, and the costs of testing are reasonable.
2.1.2 Visible Emissions
Comment: One commenter (IV-D-4) endorses the use of Reference
Method 9 as an alternative to transmissometers for continuous monitoring
of positive-pressure fabric filters. At the same time, the commenter
believes that continuous monitors should not be required on modular,
negative-pressure fabric filters that have multiple stacks. Such fabric
filters would require multiple monitors, which would significantly
increase the capital and operating costs. Therefore, the commenter
recommends that Reference Method 9 be allowed on both modular, negative-
pressure fabric filters and positive-pressure fabric filters as an
alternate method of continuous monitoring.
•Response: Two fabric filter vendors were contacted (docket
entries IV-E-1 and IV-E-2) for information about current installations
and trends in the use of modular, multiple-stack, negative-pressure
fabric filters. The vendors confirmed the EPA information that the
industry trend is toward positive-pressure fabric filters.
The commenter, a representative of a trade association, was also
contacted (docket entry IV-E-3), and he explained that a staff member's
concern about possible future use of modular, multiple-stack, negative-
pressure fabric filters prompted the comment. The commenter was not
aware of any installations of such fabric filters or of plans for their
use in the future.
Although it is unlikely that modular, multiple-stack, negative-pressure
fabric filters will be used extensively by the industry, the Agency is
aware of three such fabric filters in use to control emissions from
EAF's. Therefore, it is appropriate to permit Reference Method 9 visible
emission observations by a certified observer in lieu of a transmissometer
2-11
-------
to monitor visible emissions from such units, and Sections 273(c),
275(i), 273a(c), and 275a(c) of the regulations have been changed to
reflect this position.
Comment: One commenter (IV-D-5) supports the use of Reference
Method 9 visible emission observations as an alternative to transmisso-
meters. However, the commenter does not believe that these observations
are necessary 5 days per week. The commenter recommends making the
observations every third operating day because he believes that this
frequency is sufficient to ensure proper operation and maintenance of
positive-pressure fabric filters.
Response: The decision to permit Reference Method 9 visible emission
observations to monitor opacity from positive-pressure fabric filters
and modular, negative-pressure fabric filters with multiple stacks was
made because studies indicate that there are difficulties associated
with the use of one transmissometer to monitor multiple stacks or very
long path lengths (docket entry II-I-86). The capital and operating
costs for installing multiple transmissometers, which may be necessary
in some cases, are considered to be unreasonable. However, the Agency
continues to believe that, in general, continuous monitoring of visible
emissions from control device stacks provides the best indication of the
status of operation and maintenance of the control device. The require-
ment to observe visible emissions 5 days per week is based on our
engineering judgment that at least daily observations are necessary to
detect bag failure and to prevent associated excess emissions. In view
of these considerations, requiring Reference Method 9 visible emission
observations once per day of operation for 5 days per week is reasonable
and appropriate.
Comment: One commenter (IV-D-4) states that the regulations allow
the use of Reference Method 9 in lieu of transmissometers for continuous
monitoring of positive-pressure fabric filters. However, the preamble
(in two places) appears to allow no choice between visual opacity
monitoring by Reference Method 9 and continuous monitoring by a trans-
missometer on positive-pressure fabric filters. Therefore, the commenter
2-12
-------
recommends that the preamble be corrected to reflect the regulation,
which is the appropriate position.
Response: The commenter is correct; there is an inconsistency
between the proposal preamble and the regulation. The promulgation
preamble will explain that, for positive-pressure and modular, multiple-
stack, negative-pressure fabric filters, it is permissible to have a
certified visible emissions observer monitor the opacity of visible
emissions in lieu of installing a transmissometer. Reference Method 9
visible emission observations must be used to monitor the emissions from
these units if a transmissometer is not installed.
Comment: One commenter (IV-D-4) states that the EPA has not evaluated
the accuracy or precision of Method 9 for monitoring visible emissions
from area'sources such as shop roof monitors. The commenter states that
it is essential for the EPA to establish the accuracy and precision of
Method 9 for this type of source because (1) the method stipulates that
the accuracy of the method must be taken into account when determining
compliance; (2) the method provides information regarding its accuracy
and precision only for sources of continuous emissions being discharged
-.
through a stack; and (3) many variables exist which affect the accuracy
and precision of EAF shop roof monitor opacity readings, and these
variables were not evaluated adequately when the EPA developed and
adopted Method 9. Some of the variables that must be evaluated include
roof monitor size and geometry, air flow and velocity, orientation with
respect to the sun, various meteorological factors, fluctuations in the
appearance of emissions, and scheduling of activities in the shop.
Therefore, the commenter recommends that the EPA include an alternative
mass emission limitation in the regulation for process fugitives as an
option available in the event that the opacity limit cannot be met.
The commenter notes that there is a published method for measuring
shop roof mass emissions: D. Trozzo and J. Turnage, "Method for Deter-
mining Mass Particulate Emissions from Roof Monitors," Journal of the
Air Pollution Control Association, October 1981, Volume 31, Number 10.'
2-13
-------
Response: The "EPA Response to Remand Ordered by U.S. Court of
Appeals for the District of Columbia in Port!and-Cement Association v.
Ruckelshaus (486 F.2d 375, June 29, 1973)", discusses in detail the
reliability and accuracy of Reference Method 9 and accompanying certifi-
cation techniques for determining compliance with visible emission
standards. On the basis of this response, the visible emission standard
included in the NSPS for portland cement plants was affirmed by the
Court on appeal in Portland Cement Association v. Train, 513 F.2d 506.
The data gathered in responding to the remand for portland cement plants
convincingly demonstrate that individual visible emission observers can,
for single runs, read the opacity of visible emissions within an acceptable
level of precision. The accuracy of the Method is taken into account in
the enforcement process, as provided explicitly by Reference Method 9.
Reference Method 9 is applicable "... for the determination of
the opacity of [visible] emissions from stationary sources pursuant to . •
Section 60.11(b) ..." (40 CFR 60 Appendix A, Method 9). As stated in
the method: "Many stationary sources discharge visible emissions into
the atmosphere; these emissions are usually in the shape of a plume.
This method involves the determination of plume opacity by qualified
observers." The major factors influencing plume opacity are: particle
characteristics (particle size distribution, particle density, refractive
index), particulate concentration, the background against which the
emissions are viewed, the observer's position relative to the sun, and
the light path length through the emission plume. Particle characteristics
and particulate concentration are determined by the process operation
and the emission control technology.
In the steel industry, plumes are released at elevated points such
as roof monitors. As a result, the background for reading the opacity
of visible emissions is the same for both types of sources, generally
consisting of sky, horizon, or other structures. Furthermore, Method 9
indicates how various meteorological conditions are to be taken into
account. Thus, the ability to read the opacity of visible emissions
from roof monitors is not influenced by background anymore than is the
ability to read visible emissons from stacks.
2-14
-------
The geometry of stacks and roof monitors do differ. Stacks are
generally circular, and as a result, the path length through the plume
is essentially the same in all directions. Roof monitors tend to be
rectangular with a long and short dimension. The light path length
through the plume, therefore, is different depending on whether the
observer sights along the long dimension or the short dimension. When
the opacity of visible emissions from roof monitors is read, Reference
Method 9 specifically requires that observations be taken approximately
perpendicular to the long dimension of the roof monitor (i.e., across
the short dimension), which ensures that observed opacity is minimized.
When the visible emission standards for this industry were developed,
visible emission observations were taken from existing furnace shops
with typical roof monitor designs. Thus, the effect of path length on
opacity is taken into account during development of the standards, and
Reference Method 9 ensures that compliance with the standards is determined
by reading plumes across the shorter path length.
- The ability to read the opacity of visible emissions, therefore,,
does not depend on whether these emissions are released from stacks or
roof monitors. Reference Method 9 is applicable to plumes from stacks,
roof monitors, and other points of release.
The test program for each specific source type is designed to
account for the variables that affect the visible emissions of that
specific source. In the test program for the NSPS that applies to EAF's
and ADD vessels in steel plants, the tests covered the entire heat
cycle, thus accounting for fluctuations in the appearance of emissions
and the scheduling of relevant activities in the shop. In addition, the
various facilities tested covered the range of roof monitor sizes and
configurations and of air flows and velocities that are expected in the
industry (see BID, Vol. 1, Appendix C.).
The use of visible emission standards is technically sound and
provides the most practical and inexpensive means to ensure that affected
facilities are properly maintained and operated. The opacity of visible
emissions exiting the shop roof monitor is a good indicator of the
performance of the process and fugitive emissions capture systems.
2-15
-------
Therefore, shop roof monitor visible emission opacity limits were selected
as the format for this standard. Practical methodology does not exist
to obtain measurements of mass emissions discharged from shop roof
monitors. The shop roof mass emission testing procedures suggested by
the commenter would be more expensive than Reference Method 9 observations
and, in some cases (where there is not only low exit velocity but possibly
negative pressure and fluctuating mass concentrations), would be highly
inaccurate and unreliable. In addition, this shop roof mass emission
measurement technique has not yet been adequately tested and evaluated.
Therefore, a mass emission limit for fugitive emissions from the shop
roof would not be consistent with the requirements of the Clean Air Act.
Comment: One commenter (IV-D-2) states that it is not practical to
use Reference Method 9 observations to monitor fugitive emissions from
the dust handling equipment because of the difficulty of determining at
what point in the emission plume the opacity should be read. The commenter
suggests that it would be more appropriate to require proper operation
of'the dust handling equipment, which could include routine operational
status- inspections.
Response: Reference Method 9, Section 2.3, specifies that opacity
observations must be made at the point of greatest opacity in that
portion of the plume where condensed water vapor is not present. The
plumes that result from fugitive emissions from the dust-handling equipment
associated with EAF's in the steel industry would not be expected to
contain condensed water vapor because the temperatures of such plumes
are typically about 120° to 130°F. Thus, there should be no difficulty
in determining at what point in the visible fugitive emission plume the
opacity should be read because a certified observer only needs to look
for the point of greatest opacity.
Comment: One commenter (IV-D-2) states that, in some cases, it
could be necessary to perform three Reference Method 9 opacity observations
at each site of visible emissions from a fabric filter to comply with
40 CFR 60.275a(c). The commenter cites his equipment as an example: a
positive-pressure fabric filter with 32 compartments, each of which is
discharged into a common outlet plenum that is open to the atmosphere at
2-16
-------
each end of the fabric filter. In addition, a horizontal slot is located
on the front, bottom side of each compartment. Thus, visible emissions
resulting from a broken bag in any one compartment could be seen at
three locations. The commenter concludes that Section 275a(c) would
require 54 minutes of Reference Method 9 observations for the one incident.
Response: It is not the Agency's intent to create unnecessary work
for owners or operators of affected facilities. Thus, Sections 275(i)
and 275a(c) have been revised to make it clear that, where it is possible
to determine that visible emissions at multiple sites are attributable
to only one incident of the visible emissions, one set of Reference
Method 9 observations from the point of highest opacity that directly
relates to the cause (or location) of the incident will be sufficient.
'2.2 EMISSION LIMITS
2.2.1 Mass Emission Standard
Comment: One commenter (IV-D-5) states that, in his experience,
well-designed and -maintained positive-pressure fabric filters can
control emissions far below 0.0052 gr/dscf; however, at this time, the
commenter does not recommend a more stringent standard.
Another commenter (IV-D-6) believes that it is the EPA's policy to
base NSPS emission limits on the worst performance of any of the sources
selected for testing without regard to the condition of that worst
source and any other mitigating factors. The commenter further believes
that this policy rewards sources demonstrating a poor performance record.
The commenter recommends that the Agency establish standards that
encourage development of control strategies that reduce emissions as
much as possible. The commenter notes that Figure 4-8 of the BID
illustrates that only one fabric filter test on an EAF exceeded
0.0030 gr/dscf, and he concludes that this emission level represents the
"degree of emission limitation . . . achievable through application of
the best technological system of continuous emission reduction . . ."as
required by the Clean Air Act. The commenter states that lowering the
mass emission standard for EAF's and AOD vessels in steel plants from
0.0052 to 0.0030 gr/dscf would reduce particulate matter emissions by
2-17
-------
about 35 percent; therefore, the commenter recommends a mass emission
standard of 0.0030 gr/dscf for these sources.
Response: In establishing NSPS emission limits, the condition of
sources selected for testing is not disregarded, as commenter IV-D-6
suggests. Rather, sources that are believed to be representative of
modern practice in one or more segments of the industry to be regulated
are selected for emission testing. Moreover, the operational status of
the source is documented and monitored continuously during testing to
ensure that the test data reflect an emission control level representative
of best demonstrated control technology (BDT). Standards are set so
that well-designed, -operated, and -maintained plants, using BDT, can
achieve the standards over the range of conditions likely to recur in
the industry.
In general, the data available on well-designed and well-operated
EAF fabric filters do indicate that such fabric filters can perform to
reduce emissions below 0.0052 gr/dscf. In fact, except for one test run
at one facility, the data collected during the revision of this standard
demonstrated that fabric filters on EAF's can achieve an emission level
of less than 0.0031 gr/dscf. However, the Agency agrees with the recom-
mendations made by Commenter IV-D-5 that the mass standard should not be
lowered. This is because it was determined that, to guarantee fabric
filter compliance with a 0.0031 gr/dscf standard, vendors might increase
capital costs of fabric filters as much as 25 percent (docket nos. II-E-56,
II-E-57, II-E-58, II-E-60). This increase in costs would result from
the increased air-to-cloth ratio and other design factors needed to
assure continuous compliance with the more stringent emission limit.
Thus, the incremental cost effectiveness of the more stringent standard
could be as much as $8,000/ton, which is considered to be unreasonable.
According to the Clean Air Act amendments of 1977, Section lll(a)(l),
"a standard of performance shall reflect the degree of emission limitation
and the percentage reduction achievable through application of 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
2-18
-------
Administrator determines has been adequately demonstrated." The
0.0052 gr/dscf limit is based on the data available from well-controlled
and -operated facilities, and it takes into account the costs of complying
with the standards.
2.2.2 Visible Emission Standard
Comment: One commenter (IV-D-7) states that the proposal preamble
explains neither the increase from 0 to 6 percent opacity in the visible
emission limit for the roof monitor, nor the economic, energy, or environ-
mental impacts associated with this increase. The commenter believes
that, because the original standard allowed up to 20 percent opacity
during charging "and up to 40 percent opacity during tapping, this standard
effectively applied only during melting. He concludes that the rationale
for establishing an overall 6 percent visible emission limit presented
in the proposal preamble is not applicable. The commenter draws this
conclusion because the maximum opacity varies from 0 to 5 percent,
although the average opacity from the tested plants only varies from 0
to 0.2 percent. The commenter notes that available control technology
can achieve 3.3.percent opacity as an upper limit. The commenter
recognizes the practical difficulties in implementing the 0 percent
visible emission standard, and he recommends that the standard be set at
4 percent opacity and an exception be developed to address any infrequent
exceedance. Furthermore, he believes that Section 60.11(c) of the
General Provisions already accommodates those exceedances resulting from
start-ups, shut-downs, or malfunctions. The commenter states that he
supports a simple regulation that provides a single emission limit, and
he suggests that it might be appropriate to consider different visible
emission limits for carbon steel EAF, specialty steel EAF and AOD vessels.
Another commenter (IV-D-6) notes that Table 4-5 of the BID
demonstrates that, except for one mill melting dirty scrap, all tested
plants could comply with a 5 percent visible emission limit on all
operations. Thus, the commenter concludes that a 5 percent visible
emission limit for all operations is achievable and required by the
Clean Air Act.
Response: We agree with commenter IV-D-7 that a simple regulation
with a single emission limit is appropriate for this industry. By
2-19
-------
setting the level of the standard to include all the data acquired
during entire heat cycles, provisions for exceedances are not necessary.
As was explained in the proposal preamble (48 FR 37347), the visible
emission limits were selected based on the performance of the capture
and control technologies that served as the basis for Regulatory Alternative B.
The economic, energy, and environmental impacts associated with the
revised standard were presented in the proposal preamble.
Although the impacts associated with Regulatory Alternative C are'
considered reasonable, this alternative was not considered suitable as
the basis for national standards of performance because it is based on a
closed roof configuration which may aggravate worker and equipment heat
stress problems. Operating experience with this roof configuration is
limited in areas of the country where ambient temperatures and humidity
are high. Because the effects of heat stress cannot be fully evaluated
at this time, Regulatory Alternative B was selected as the basis for the
proposed revised standards.
Twenty-seven hours of opacity observations were made of shop roof
visible emissions at two shops that utilized the capture systems upon
which Regulatory Alternative B is based. These observations show that
the maximum opacity of shop roof visible emissions is 5 percent. Visible
emission limits for NSPS are based on achieved levels at well-operated
and -maintained facilities that have installed what is considered to be
the best demonstrated control technology. Visible emission limits are
not based on opacity values averaged for the period of the test (see
Portland Cement v. Train, supra). Thus, the visible emission level for
this industry was set at 6 percent, which includes the highest Reference
Method 9 observation plus a reasonable margin of safety. This methodology
was approved by the Court in Portland Cement v. Train, supra. Under
Regulatory Alternative B, exceptions for exceedances are unnecessary
because the maximum opacity recorded was 5 percent throughout the heat
cycle. As the commenter notes, 40 CFR 60.11(c) provides an exception
for nonnormal operation. It is not appropriate to develop separate
visible emission limits for carbon steel and specialty steel shops
because the EAF's in both types of shops are operated and emit particulate
matter in a similar manner, and, also, it is possible that the EAF will
be operated alone from time to time in the specialty shop.
2-20
-------
The plant that commenter IV-D-6 refers to appears to be Plant H.
This plant was among those plants tested in developing the original
standards. Control technology for EAF's, particularly for their fugitive
emissions, has improved since the original test program. Thus, this
plant was not used as a basis for the revised standards. For the reasons
enumerated above, we believe that 6 percent opacity rather than 5 percent
is the appropriate level for the visible emission standard for this
industry. The Clean Air Act requires that the standard be achievable
and representative of the best demonstrated control technology, taking
into consideration other factors such as reasonable cost. The Court
decision in National Lime Association-v. EPA (627 F.2d 416 [1980])
requires that the standard must be achievable under all conditions
expected to recur, and, in this industry, this includes the use of dirty
scrap.
Comment: One commenter (IV-D-4) states that the EPA has not adequately
documented that EAF and AOD vessel facilities subject to the NSPS will
be able to comply with the 6 percent visible emission limit from shop
roofs during charging and tapping (processes likely to cause the greatest
fugitive emissions). The commenter states that only two out of seven
plants tested were representative of the recommended emission capture
system that is the basis for the standard (closed roof over the furnace
and open elsewhere); four plants have totally closed roofs (representa-
tive of a more effective capture system); and one plant has a completely
open roof monitor (representative of a less effective capture system).
Forty-five 6-minute opacity readings were taken during charging and 24
6-minute readings were taken during tapping at the two plants that were
representative. This means that the standard is based on less than
7 hours of relevant observations. The" commenter believes that the
number of observations made on shops with a closed roof over the furnace
is limited, and, therefore, these data should not be used as a basis for
the regulatory limit. The commenter cites National Lime Association v.
EPA, supra, as an example of a case in which the EPA failed to establish
the representativeness of the tested plants and the achievability of the
standard under adverse conditions. The commenter recommends that excep-
tions to the opacity limits be allowed during charging and tapping as
2-21
-------
are allowed in the original NSPS or that the preamble and regulation
address emission problems that occur because of equipment malfunctions.
Furthermore, the commenter recommends that any source meeting the EPA's
suggested equipment requirement but failing the opacity standard be
allowed to demonstrate compliance based on alternate mass emission
limitations.
The commenter also questions the validity of the correlation between
the shop roof opacity and mass emission rates. Because these estimated
mass emission rates were used to determine the ambient air quality
impacts, the commenter doubts the accuracy of the air quality impacts
upon which the need for the proposed regulatory limit is based.
The commenter notes that Section 60.11(e) of the General Provisions
allows a source to exceed the visible emissions standard if the mass
standard is being met.
Response: The commenter is correct that the data base for the
control configuration recommended for the NSPS contains tests at two
facilities (Plants J and N) that "are representative of the suggested
technology (closed roof monitors over furnace only)" [Regulatory
Alternative B]. The commenter also points out that approximately 7 hours
of Reference Method 9 observations were made for the charging and tapping
portions of the heat cycle. The total data base includes 27 hours of
Reference Method 9 observations during the course of entire heat cycles
at plants representative of the recommended technology. This amount of
visible emission data acquired at representative plants is, in the
EPA's judgment, an adequate data base upon which to set a standard.
National Lime Association v. EPA, which the commenter cites, does require
that the data be from representative facilities and that the standard be
achievable. However, the Court did not specify any quantity of data that
must be acquired before a standard can be set, and the Agency believes
that the data are sufficient to demonstrate the achievability of the
standard because worst-case conditions for this industry were included
in the test program. The questions of achievability of the standard and
limited data were raised by the American Iron and Steel Institute at the
National Air Pollution Control Techniques Advisory Committee meeting in
July 1982, prior to proposal of the revised standards. In response to
2-22
-------
these concerns, Plant N was visited and tested. Even during furnace
upset conditions, when the fugitive emission capture system was receiving
furnace emissions at a rate estimated to be,almost 10 times higher than
it would during normal furnace operation, Plant N achieved the standard.
The maximum 6-minute average visible emission reading over a 2-day
period that covered many entire heat cycles was 3.3 percent. All of the
data for Alternative B demonstrate that the visible emission limit of
6 percent is achievable.
As noted in the proposal preamble (48 FR 37346), although most of
the shops in the industry are closed roof or are changing to closed
roof, and new sources are expected to be located in closed roof shops,
Alternative B was recommended because the effects of heat stress on
workers and equipment in closed roof shops in some areas of the country
were unknown. The Agency did not want to risk causing any facility to
incur problems with heat stress to achieve compliance with the standards.
The commenter represents the major industry trade association, and this
association's comments (II-D-67 and II-E-54) about possible heat stress
problems in closed roof shops persuaded the Agency to conclude that the
standard should allow the less stringent Regulatory Alternative B. As
both the trade association and the Agency recognized, there were few
partially open shops in existence, and, thus, only limited data could be
acquired; however, these data are considered to be sufficient to set
standards based on Regulatory Alternative B which is "worst case" in
terms of emissions to the atmosphere.
Because the 27 hours of data acquired during charging, melting, and
tapping demonstrate that the 6 percent visible emission limit can be
achieved with best demonstrated control technology, the Agency no longer
believes that exceptions to the standard are appropriate for the charging
and tapping portions of the EAF heat cycle. In addition, it is not
necessary to include an exception for equipment malfunctions in this
standard because the General Provisions (40 CFR 60.11[c]) already provide
relief for exceeding the standard during start-up, shutdown, and
malfunctions.
Finally, Section 60.11(e) is not a provision that allows
substitution of a mass emission standard for a visible emission standard.
2-23
-------
Rather, Section 60.11(e) provides that owners or operators of any
affected facility from any source category that meets the mass emission
standard but does not meet the visible emission standard may apply for
an individual visible emission standard tailored to the unique circum-
stances of their facility. This individual standard applies for the
life of the affected facility and is automatically approved upon
demonstration that: (1) the facility is in compliance with the mass
emission standard; (2) the facility and associated air pollution control
equipment were operated and maintained in a manner to minimize the
opacity of emissions during the performance tests; (3) the performance
tests were performed under the conditions established by the Administrator;
and (4) the facility and associated air pollution control equipment were
incapable of being adjusted or operated to meet the applicable opacity
standard. For the reasons discussed^in Section 2.1.2, page 2-14, a mass
emission standard for fugitive emissions from shop roofs is not practical
at this time. Therefore, Section 60.11(e> is not applicable to the shop
roof standard.
As the commenter stated, the BID, Vol. I, page 6-16, states both
that estimated fugitive emission reduction efficiencies were based on a
review of the literature (e.g., II-A-8, II-I-68, and II-I-94),
observation of furnace capture technologies at the facilities tested,
and engineering judgment, and that this methodology is reasonable. The
Agency did riot, as the commenter concludes, attempt to correlate observed
EAF shop roof opacities with mass emission rates. The visible emission
standard and the mass emission standard are based on achieved emission
rates. The estimated efficiencies were used to obtain estimated emissions
reductions required to calculate cost effectiveness values of the various
regulatory alternatives. Dispersion modeling was used to predict the
contribution of emissions from EAF's and ADD vessels to the ambient
particulate concentration. There was no attempt to establish any
relationship between the opacity of visible emissions and ambient air
concentrations.
Comment: One commenter (IV-D-4) states that the deletion of
Section 272(a)(3)(iii) for sources built between October 21, 1974, and
August 17, 1983, was not explained at proposal and is inappropriate.
2-24
-------
This subsection required compliance with the shop roof opacity standard
only when the flow rate through each capture hood and the pressure in
the free space inside the furnace were being measured during a perfor-
mance test. The flow rates and pressure established at this time became
"baseline." At all other times, these operating conditions were required
to be maintained at the baseline values or better. The commenter
believes that the deletion of this paragraph results in the imposition
of a new and more stringent emission limit on shops built to comply with
the original NSPS because these shops will now have to meet the shop
opacity standards during all routine EAF operations. The commenter
believes that this is retroactive regulation of existing sources and
exceeds the EPA's authority under Section 111,of the Clean Air Act. The
commenter recommends reinstatement of the paragraph.
Response: The deletion of Section 272(a)(3)(iii) from the standards
is not considered to be more stringent regulation and occurred because
it was believed that not having to continuously monitor the flow rate
and pressure would relieve some of the monitoring burden on owners or
operators of affected facilities. Deletion of this section is less
expensive for, and more convenient to, owners or operators; however,
because of the concern expressed by the commenter, the regulation has
been amended. Sources built between October 21, 1974, and August 17,
1983, will again be responsible for continuously monitoring, and main-
taining at baseline values, the flow rate through each capture hood and
the pressure in the free space inside the furnace. The shop roof visible
emission standard will apply only during performance and compliance
tests. Section 272(a)(3)(iii) and related Sections 274(a)(3), (a.)(4),
(b), (c), (e), and (f) of the original regulation will be reinstated.
Sections 274(b) and (c) have been revised, and Sections 274(e), (f),
and (g) have been redesignated (f), (g), and (h).
Comment: One commenter.(IV-D-7) states that he generally agrees
with the basis (improvements in fugitive emissions capture technology
and increase in use of positive-pressure fabric filters) for the
amendments (the shop roof visible emission standard and permitting
Reference Method 9 observations in lieu of a transmissometer for
2-25
-------
positive-pressure fabric filters) to the NSPS for EAF's in steel plants,
and, thus supports the revisions to the standards.
Response: No response is necessary.
2.3 COSTS OF TESTING
Comment: One commenter (IV-D-2) states that compliance tests and
procedures required by 40 CFR 60.275a(b) should not be required for
facilities that exhibit no visible emissions (zero opacity). The
commenter cites 48 FR 37344, Table 3, as evidence that the average
concentration of the mass emissions is lower than the standard even when
visible emissions exhibit opacities that are as high as 2.8 percent.
Thus, the commenter believes that the expense of installing sampling
ports and platforms should only be incurred if visible emissions are
observed. The commenter states that sampling ports are just another
source for leakage in a positive-pressure fabric filter and require
additional maintenance.
Response: Visible emissions of zero opacity do not necessarily
mean that the mass emission standard is being achieved by a-control
device because visible emissions do not relate directly to mass emissions
but rather to mass concentration. This is one of the reasons that
40 CFR 60.8(a) of the General Provisions requires all new facilities to
conduct performance test(s) and to furnish the Administrator a written
report of the results of such performance test(s). Section 60.8(e)
requires that the owner or operator of an affected facility provide
testing facilities such as sampling ports, sampling platforms, safe
access to platforms, and utilities for sampling equipment. Because
these sampling requirements are necessary for any new facility required
to comply with any standard, these requirements are not an additional
burden to .new EAF and AOD vessel facilities as a result of this NSPS.
Sampling ports on the fabric filter inlet side can be sealed effectively
against air leakage and ports on the outlet side cause no leakage
problems because they are downstream of the bags. Sampling ports do not
require significant maintenance.
Comment: One commenter (IV-D-2) states that monitoring fabric
filters during a compliance test would be time consuming and expensive,
2-26
-------
especially in the case of his 32-compartment fabric filter where each
compartment would have to be checked for its sequence in the cleaning
cycle, damper positions, broken bags, etc.
Response: As explained in Section 2.1.1, the maximum number of
sampling sites would be 16 for this fabric filter. Acquiring information
on control device conditions may be time consuming; however, this is
part of any performance test and is essential in evaluating the test
results. The Agency is aware of the time and expense involved in testing
the fabric filter, and, therefore, every attempt was made to simplify
testing procedures for a multi-compartment fabric filter. The costs of
testing were included in the economic analyses of the standard.
Comment: One commenter (IV-0-3) states that the test log required
by 40 CFR 60 Sections 275(c) and 275a(d), Item 23, repeats all of the
'information required by Items 1 through 22 in these sections and, thus,
is unnecessary. Furthermore, the commenter believes that maintaining
the log during testing could require the use of an additional person,
thereby increasing the cost of testing.
Response: There is sufficient information to evaluate the testing
procedure and accuracy in Items 1 through 22 to be included in the test
reports required under Sections 60.275 (c) and 60.275a(d). Therefore, -
item 23 (test log) is not necessary and has been deleted. Item 7 has
also been changed to include test times, as well as test dates.
2.4 MISCELLANEOUS
Comment: One commenter (IV-D-4) states that the requirement to
monitor the pressure in the free space within the EAF is unnecessary.
The commenter states that the quality of steel produced is dependent
upon furnace pressure, and the commenter does not believe that the EPA
has authority to require monitoring of operating parameters that do not
relate directly to control devices. The commenter believes that it is
the owner or operator's obligation to control emissions to the EPA-
specified limits and that he should be allowed to choose how to adjust
process operations to achieve this objective. Therefore, the commenter
recommends that the EPA delete the requirement that EAF free space pres-
sure be monitored/recorded.
2-27
-------
The commenter also questions what improvement in air quality can be
expected as a result of other monitoring requirements, such as those
stated in Section 60.274a(c) that requires the owner or operator to
perform monthly operational inspections of capture equipment and record
the deficiencies noted. The commenter believes that the owner or
operator should be allowed to determine the frequency and type of
inspection and maintenance program that is most efficient for his
operation.
Response: Electric arc furnace facilities use pressure monitors
inside the furnace to measure the pressure in the free space because
this pressure is a critical parameter in steel production and, thus, is
routinely monitored and recorded at every facility. This regulation
does not specify any pressure to be maintained, only that the pressure
be monitored and recorded. The Agency may require monitoring of
parameters that indicate proper operation and maintenance, whether or
not they relate directly to control devices. Records of the pressure
enable enforcement personnel to compare the pressure with the baseline-
pressure established during the performance test. This parameter is an
indicator of the proper operation and maintenance and relates to
effectiveness of emission capture.
The requirement that the capture equipment be visually inspected
and that all deficiencies be noted and corrected is critical to ensure
proper operation of the capture equipment at the levels observed during
the performance test. The once per month frequency for inspections has
been determined to be reasonable. The owner or operator may, of course,
increase the frequency of inspections.
Comment: One commenter (IV-D-4) states that the definitions of an
EAF affected facility in the BID, Vol. I, and in Section 271a(a) of the
regulation are not the same. Also, the definitions presented do not
include all of the necessary equipment (such as cranes, ladles, additive
systems, etc.) for the production of molten steel. The commenter recom-
mends that an EAF shop be defined as one which consists of all equipment
that must be required to produce molten steel. The commenter states
that the BID does not make it clear that no one particular equipment
change would necessarily constitute reconstruction. Thus, the commenter
recommends deletion of the specific equipment listing.
2-28
-------
Response: The affected facility definition in Section 271a(a) and
the identification of components (48 FR 37349 and BID Vol. I, page 5-3)
that would be examined in determinations of modification and
reconstruction are not the same. Section 271a(a) defines the EAF to
which the proposed regulation is applicable. On the other hand, the
proposal preamble and the BID identify the components that can affect
the rate of emissions and, thus, compliance. These definitions provide
information appropriate to the specific context, and there is no incon-
sistency. In general, modifications are "any physical or operational
change to an existing facility which results in an increase in the
emission rate to the atmosphere of any pollutant to which a standard
applies ..." (40 CFR 60.14[a]). Reconstruction means the replacement
of components of an existing facility to such an extent that: (1) the
fixed capital cost of the new components exceeds 50 percent of the fixed
capital cost that would be required to construct a comparable entirely
new facility, and (2) it is technologically and economically feasible to
meet the applicable standards set forth in this part (see 40 CFR
60.15[a] and [b]). Because it can change the emission rate or costs of
control, for purposes of modification and reconstruction determinations, '
the emission control equipment is included in the definition relating to
these determinations. However, the control equipment does not itself
create emissions and, thus, is not considered to be part of the affected
facility for purposes of the regulation.
The commenter requests that the definition of the affected facility
be expanded to include items in a shop such as ladles, cranes, and
additive systems. Available information suggests that ancillary equipment
does not cause or contribute to emissions (see 48 FR 37340) or change
the emission rate from an EAF and, thus, is not appropriate for inclusion
in the definition of an EAF affected facility. If the commenter1s
additions were identified as affected facilities, they would be separate
affected facilities. The EAF is the practical integral emission source
and is, therefore, the unit to which the proposed regulation applies. '
The Agency believes that the terms EAF and shop in Section 271a(a)
should remain as defined.
2-29
-------
With respect to deleting the paragraph containing examples of
equipment that might be examined in reconstruction considerations
(48 FR 37349), the preamble clearly states that no one of the changes
would constitute reconstruction, and, thus, no ambiguity exists that
would create the enforcement problem the commenter anticipates. The
equipment list merely suggests items to consider in making a
reconstruction determination.
Comment: One commenter (IV-D-2) states that there may be a carbon
monoxide (CO) hazard that would require use of airpacks by testing
personnel during either in-stack or high-volume testing.
Response: Of the 19 plants providing the test data used in
developing the data base to support the revised NSPS, 8 plants were
tested for CO; of those 8, only 4 had measurable CO in the gas downstream
of the collectors. In none of the tests was it necessary that test
personnel use airpacks. However, if a situation arises where CO is a
test hazard, the test personnel would be expected to use appropriate
equipment for which the costs would be negligible.
Comment: One commenter (IV-D-5) notes that jet pulse fabric filters
have not been given consideration in developing the revised standards
for EAF's and AOD's in steel plants. The commenter states that the use
of these fabric filters is popular, and, therefore, he believes they
should be investigated.
Response: "Jet pulse" refers to a pulse-jet fabric filter cleaning
method. At present, fabric filters using a reverse air flow cleaning
mechanism are typical on EAF's and AOD vessels in steel plants. Of 21
plants visited during development of the original and the revised NSPS,
14 had fabric filters with a reverse air flow cleaning mechanism, 6 used
a shaker type mechanism, and only 1 used a pulse-jet cleaning mechanism.
The one fabric filter utilizing a pulse-jet cleaning mechanism was
tested during development of the revised standards and achieved the
standard even during furnace upset conditions.
In summary, the pulse-jet fabric filter has been investigated and
can be used to comply with the proposed particulate mass emission limit.
This standard of performance is expressed as an emission limit to permit
the EAF owner or operator to select a control device of his own choosing
that will comply with the standard.
2-30
-------
Comment: One commenter (IV-D-3) requests that the word "daily" 'be
deleted from 40 CFR 60 Sections 273(c) and 273a(c) and that the words
"Method 9" in these sections be replaced with "Section 275(i) [275a(c)]
of this subpart." The commenter recommends these changes to make it
clear that Reference Method 9 visible emission observations are required
for not more than 5 days per week.
Response: Use of "daily" is confusing and, thus, the commenter's
suggested change to the regulation has been made.
2-31
-------
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-450/3-82-020b
3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
Electric Arc Furnaces and Argon-Oxygen Decarburization
Vessels in Steel Plants—Background Information for
Promul-gated Standards of Performance
5. REPORT DATE
August 1984
6. PERFORMING ORGANIZATION CODE
7, AUTHOR(S)
8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Office of Air Quality Planning and Standards
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3059
12. SPONSORING AGENCY NAME AND ADDRESS
Director for Air Quality Planning and Standards
Office of Air and Radiation
U. S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/200/04
15. SUPPLEMENTARY NOTES
16. A8ST1
\CT
Standards o.f performance for the control of particulate matter emissions from
electric arc furnaces and argon-oxygen decarburization vessels at new, modified, or
; reconstructed steel plants are being promulgated under the authority of Sections 111,
114, and 301(a) of the Clean Air Act, as''amended. These standards would apply to
those affected facilities that commence construction on or after August 17, 1983, the
date of original proposal. This document contains a summary of the public comments
on the proposed revised standards and the EPA's responses, as well as summary economic
and environmental impact statements.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTlFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Air Pollution
Pollution Control
Standards of Performance
Steel Plants
Particulate Matter
Air Pollution Control
8. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
38
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220—1 (Rev. 4—77) PREVIOUS EDITION is OBSOLETE
-------
-------
-------
r
COO) £*
O 3 —
QQJ Ot
£T O
(0 o>
5«
cu w
CD o o
QJ Co CD
S' Q 0)
tfa
"-3
_* a.
CO
------- |