Summary of Comments and Responses for the Primary Aluminum Reduction
Plants

Source Category:  Risk and Technology Review, Phase II, Group 2

Advance Notice of Proposed Rulemaking (72 FR 14734, March 29, 2007) 

Metals and Minerals Group

Sector Policies and Program Division

Office of Air Quality Planning and Standards

U.S. Environmental Protection Agency

November 16, 2009

INTRODUCTION

This document provides a summary of comments received for the Advance
Notice of Proposed Rulemaking (ANPR) for the Risk and Technology Review
for the Primary Aluminum Reduction Plants (72 FR 14734, March 29, 2007).
 It also includes comments on POM emissions and speciation solicited by
EPA from industry after the ANPR that was needed for the emissions
database to be used for future risk modeling for the proposed rule.  The
EPA responses to the comments follow the comments below, as appropriate.
 All commenters are listed in Table 1.  The majority of comments focused
on corrections to the National Emissions Inventory (NEI) database for
specific primary aluminum reduction plants.

Table 1.  List of Commenters for Primary Aluminum NESHAP ANPR and 

Revised Emissions Modeling Database 

ANPR Commenters

Commenter Number*-	Commenter	Affiliation

0013	Brian Galley	Pechiney Plastic Packaging, Inc.

0014	Tim Adamowicz	Ormet Primary Aluminum

0015, 0022	Sean Lane	Century Aluminum

0016	Jamie Coomes	ALCAN Primary Products

0017	Robert Strieter	The Aluminum Association

0018	Keith Overcash	North Carolina Department of Environment and Natural
Resources

0019	Brandy Albertson	Oregon Department of Environmental Quality

0020	David Shaw	New York State Department of Environmental Conservation

0021	Michael Palazzolo	ALCOA

0023	Michael Palazzolo	ALCOA

Commenters on Revised Emissions Modeling Database

0024**	Michael Palazzolo	ALCOA

0025**	Michael Palazzolo	ALCOA

0026**	Michael Palazzolo	ALCOA

0027	Michael Palazzolo	ALCOA

0028	Robert Strieter	The Aluminum Association

0029	Steve Wright	Columbia Falls Aluminum Company

0030	Michael Palazzolo	ALCOA

*  Docket numbers for these comments use the above 4 digits with the
prefix:  EPA-HQ-QAR-2002-0031-XXXX.

**  Test reports for sampling and analysis of polycyclic aromatic
hydrocarbons.

NATIONAL EMISSION INVENTORY (NEI) COMMENTS

General Inventory Comments

Comment:  All of the commenters shown in Table 1 above provided changes
to the NEI database, and the most extensive changes were requested by
Commenter 0021, whose company owns or operates 10 of the 16 primary
aluminum plants in the U.S.  

Examples of changes provided by Commenter 0021 for the NEI database
include:

Identifying plants, potlines, and other operations that have shut down;

Removing emission points and emissions associated with secondary
aluminum production; and

Correcting stack gas parameters, stack diameters and heights, stack
latitudes and longitudes.

Response:  The changes provided by the commenters were reviewed, and
changes were made to update the NEI based on the corrected site-specific
data from the commenters, as detailed in Appendix A.  We also
investigated other deficiencies in the NEI database identified by the
commenters and gathered information to fill in data gaps.  A listing of
these other improvements made to the NEI database, in large part due to
observations of reviewers and commenters, are also noted in Appendix A.

Comment:  Commenters 0017 and 0021 claimed that the preliminary risk
review incorrectly identified HCl and mercury as key pollutants causing
potential health risks from primary aluminum plants.  The commenters
noted that emissions of HCl are more commonly associated with secondary
aluminum processes, and emissions of mercury are associated with
coal-fired power plants.

Response:  We have confirmed that the HCl emissions in the preliminary
analysis were from secondary aluminum production, and we have removed
secondary aluminum emissions from the primary aluminum database.  We
also confirmed that a majority of the mercury emissions were from a
coal-fired power plant co-located with a primary aluminum facility;
these mislabeled emissions have been removed from the modeling database.
 

Comment:   Two commenters (0027, 0028) recommended using 100 percent
utilization of capacity to estimate both “actual” and
“allowable” emissions.  The commenters explained that although a few
plants are operating at less than 100 percent due to current economic
conditions, many are operating at 100 percent, and those that are not
will increase to 100 percent in future years as economic conditions
improve.  

Response: EPA agrees with the commenters, since using an industry-wide
average of 76 percent utilization could underestimate emissions for many
plants.  Therefore, POM and other emission estimates will be based 100
percent utilization of capacity.  

Comment:  Commenter 0021 stated that even with the extended comment
period, there was not enough time to fully review and update the NEI
database.  The commenter asked that EPA not preclude the submission of
supplemental data after the end of the comment period.

Response:  We continued to work with this commenter, as well as other
industry representatives and the trade association, after the end of the
comment period.  The commenter continued to provide additional data from
his company’s environmental professionals, and we have incorporated
these improvements into the database.  As noted below, additional
opportunities were provided for industry representatives to review the
changes made to the database with a particular emphasis on filling in
data gaps for emissions of POM species.

Polycyclic Organic Matter (POM) Comments

Comment:  Commenter 0021 provided speciated data for polycyclic organic
matter (POM), corrected emission estimates, added emission estimates
that were missing, and separated total emissions to identify emissions
from individual potlines and roof vents.

Comment:  Commenter 0020 said that POM emissions were not included for
two plants in their State and asked that the POM emission estimates be
added.

Response:  The data provided by the commenters, above, indicated that
many plants had not included emissions of POM species.  EPA compiled
data for POM species for the different types of processes, developed
emission factors, and applied these factors to estimate emissions for
POM species.  The methodology and estimates were sent to industry
representatives and the trade association for their review.  Revisions
were made by EPA based on their review comments, and the POM database
was sent for a second review.  We incorporated the few minor additional
comments received from one company during the second review (0030).

Comment:  One commenter (0027) submitted test results to estimate POM
emissions from horizontal stud Soderberg (HSS) plants.  (See Docket Item
0025.)  

Response:  EPA has replaced the estimated POM emissions from the HSS
plant based on the comments, since the commenter (ALCOA) has the only
existing HSS plant, and the test results are specifically for that
plant.  The data used previously in the EPA estimates were from a plant
that has permanently closed. 

Comment:  One commenter (0029) submitted POM estimates for vertical stud
Soderberg (VSS) plants.  This facility is the only VSS plant currently
operating.  The data were based on several years of site-specific
testing results submitted by the facility for methylene chloride soluble
organics (MCSO), which are the surrogates used for POM.

Response:  EPA adjusted the POM estimates for vertical stud Soderberg
(VSS) plants, since this facility is the only VSS plant currently
operating.  

MODELLING COMMENTS

Comment:  Commenters 0017 and 0021 said that EPA’s use of AERMOD to
evaluate risk from potline fugitive emissions is inappropriate because
it does not contain or make use of the Buoyant Line and Point (BLP)
source algorithms required to properly model 300-m long roof vents with
multiple emission points.  The commenters recommended using the BLP or
CALPUF models.

Response:  These initial ANPR results are a screening assessment.  If
needed, EPA will look at some facilities on a more refined level, such
as using BLP, before we get to the final rulemaking phase of the
residual risk analysis.  The EPA took a similar approach for the coke
ovens rulemaking to use the BLP to predict plume rise for coke
batteries.  Also, EPA has some strategies for entering data from line
sources so that the air model represents them accurately:  split the
emissions into a group of points so that the plume rise from the hot
plumes can be captured; and/or modify the NEI emission inventory input
file with a change to the model source inputs and make it as multiple
area or point source release points to better represent a line source
emission.

RISK ASSESSMENT COMMENTS

Comment:  Commenters 0017 and 0021 stated that it was not appropriate to
multiply the average emission rate by a factor of 10 to evaluate acute
(1-hr) exposures.  The commenters said that potroom control devices,
paste plants and anode bake furnaces are steady state operations,
emissions that do not vary much from average values, and no safety
factor should be applied.  However, the commenter also stated that
emissions increase from potroom vents during carbon setting and metal
tapping and recommended a factor of 4 to 6 times the average to estimate
hourly roof emissions.

Response:  The suggestions for acute exposure factors suggested by the
commenters will be used by the EPA to develop the factors to be used in
the risk analysis performed for the proposed rule.  The initial factor
used by EPA for the ANPR risk analysis was a default that was considered
reasonable in the absence of industry-specific information.  The public
will have additional opportunity to comment on the factors used by EPA
in the proposal comment period.

Comment:  Commenter 0020 recommended that the risk evaluation be based
on the facility’s actual maximum emissions if not potential emissions.
 The commenter cited an example of one plant that reported carbonyl
sulfide emissions of 173 tons per year in the 2002 NEI whereas the
Toxics Release Inventory (TRI) shows 285 tons per year.  The commenter
also said that another plant had similar stack emissions reported in
both the NEI and TRI; however, the fugitive emissions in the TRI were 20
times greater than those reported in the NEI.

Response:  We plan to perform more refined modeling based on both
“actual” emissions as reported by the companies and States, and
“allowable” emissions, which are the maximum emissions permitted
under the current regulations.  The maximum or “potential” emissions
recommended by the commenter are comparable to the allowable emission
estimates.Appendix A  

Details of Revisions to the NEI Database for Primary Aluminum 

Based on ANPR Comments and Subsequent Industry Review and Comment 

on the Modeling Database

Incorporated changes in emission points, stack characteristics,
locations, etc., according to facility comments.

Corrected facility names, facility operating status, and operating
parameters:

Deleted Reynolds (Troutdale) and Northwest Aluminum because they have
permanently closed.  

Deleted Alcan Packaging because it is not an aluminum plant.

Updated name changes for mostly ALCOA acquisitions, and Century
Aluminum.

Adjusted the capacity of the ALCOA plant in Badin, NC to 60,000 metric
tons/year because one of the potlines has been permanently shut down.

Revised (upwards) the aluminum and paste plant capacities for Columbia
Falls based on their comments.

Changed the tons anode per ton aluminum ratio from 0.5 to 0.57 tons
anode/ton aluminum based on comments from the Aluminum Association and
ALCOA.

Checked emission points for completeness, e.g., most were missing the
paste plant.  Added emission units that were missing.  

Information for three plants was particularly incomplete:  Columbia
Falls (only one emission point entry, unidentifiable); Century Aluminum
in WV; and Lake Charles Carbon.  EPA obtained their Title V permits and
entered emission point data that were available. 

Removed emissions for a paste plant at ALCOA (West) Massena, NY, because
it is no longer operating since the facility now purchases anodes from a
company in Canada.

Reviewed and corrected the default stack characteristics for
reasonableness: 

Revised several default entries that had stack diameters of 0.003 ft and
5,060 ft, which did not seem reasonable.  

Compiled stack characteristics by type of process and averaged them. 
Filled in any missing stack characteristics with the overall averages
for height, temperature, diameter, and velocity.  Calculated volumetric
flow from diameter and linear velocity to be consistent.  

Commenter 0030 provided latitude and longitude data for potroom roof
vents for several ALCOA plants.

Developed emission factors and estimates of POM species for the
different types of processes.  Industry representatives and the trade
association provided review.  The following changes were made based on
the review comments:

Revised the POM emission estimates on 100% utilization of capacity.

Used test results submitted by ALCOA to estimate POM emissions from
horizontal stud Soderberg (HSS) plants.

Adjusted the POM estimates for Columbia Falls Aluminum’s VSS facility
based on several years of site-specific testing results provided by the
facility.

 PAGE   

 PAGE   5 

 PAGE   

 PAGE   7 

